SWE-Dev/SWE-Dev-3 · Datasets at Hugging Face

repo stringlengths 7 59	instance_id stringlengths 11 63	base_commit stringlengths 40 40	patch stringlengths 167 798k	problem_statement stringlengths 20 65.2k	hints_text stringlengths 0 142k	created_at timestamp[ns]date 2015-08-30 10:31:05 2024-12-13 16:08:19	FAIL_TO_PASS sequencelengths 0 0	PASS_TO_PASS sequencelengths 0 0
GAA-UAM/scikit-fda	GAA-UAM__scikit-fda-510	e69154b77d9e0fc260dd705de44c01fdf468e202	diff --git a/skfda/preprocessing/dim_reduction/_fpca.py b/skfda/preprocessing/dim_reduction/_fpca.py index d82debbaa..fba44acd3 100644 --- a/skfda/preprocessing/dim_reduction/_fpca.py +++ b/skfda/preprocessing/dim_reduction/_fpca.py @@ -1,8 +1,8 @@ """Functional Principal Component Analysis Module.""" - from __future__ import annotations -from typing import Callable, Optional, TypeVar, Union +import warnings +from typing import Callable, TypeVar import numpy as np import scipy.integrate @@ -34,7 +34,13 @@ class FPCA( # noqa: WPS230 (too many public attributes) Parameters: n_components: Number of principal components to keep from - functional principal component analysis. Defaults to 3. + functional principal component analysis. + + .. versionchanged:: 0.9 + In future versions, it will default to the maximum number + of components that can be extracted. + Currently, it still defaults to 3 but do not assume this + behavior as it will change. centering: Set to ``False`` when the functional data is already known to be centered and there is no need to center it. Otherwise, the mean of the functional data object is calculated and the data @@ -86,13 +92,23 @@ class FPCA( # noqa: WPS230 (too many public attributes) def __init__( self, - n_components: int = 3, + n_components: int \| None = None, *, centering: bool = True, - regularization: Optional[L2Regularization[FData]] = None, - components_basis: Optional[Basis] = None, - _weights: Optional[Union[ArrayLike, WeightsCallable]] = None, + regularization: L2Regularization[FData] \| None = None, + components_basis: Basis \| None = None, + _weights: ArrayLike \| WeightsCallable \| None = None, ) -> None: + + if n_components is None: + warnings.warn( + "The default value of n_components will change in a future " + "version to the maximum number of components that can be " + "extracted. Update your code to specify explicitly the " + "number of components to avoid this warning.", + DeprecationWarning, + ) + n_components = 3 self.n_components = n_components self.centering = centering self.regularization = regularization	FPCA default number of components is wrong Describe the bug By default, it is natural that FPCA returns as many components as possible. The current behaviour is to return only 3, which is not a good default. We should deprecate it and change it in a future version.		2023-02-01T15:45:29	[]	[]
GAA-UAM/scikit-fda	GAA-UAM__scikit-fda-502	0b578efd1dc78a56124a7e310a35d30116398429	diff --git a/skfda/misc/_math.py b/skfda/misc/_math.py index 0cb71737b..9b5f0b5a8 100644 --- a/skfda/misc/_math.py +++ b/skfda/misc/_math.py @@ -468,7 +468,7 @@ def _inner_product_integrate( len_arg2 = len(arg2) else: # If the arguments are callables, we need to pass the domain range - # explicitly. This is used internally for computing the gramian + # explicitly. This is used internally for computing the gram # matrix of operators. assert _domain_range is not None domain_range = _domain_range diff --git a/skfda/misc/operators/__init__.py b/skfda/misc/operators/__init__.py index 605b3a933..482a3a5cb 100644 --- a/skfda/misc/operators/__init__.py +++ b/skfda/misc/operators/__init__.py @@ -12,8 +12,8 @@ "_operators": [ "MatrixOperator", "Operator", - "gramian_matrix", - "gramian_matrix_optimization", + "gram_matrix", + "gram_matrix_optimization", ], "_srvf": ["SRSF"], }, @@ -28,7 +28,7 @@ from ._operators import ( MatrixOperator as MatrixOperator, Operator as Operator, - gramian_matrix as gramian_matrix, - gramian_matrix_optimization as gramian_matrix_optimization, + gram_matrix as gram_matrix, + gram_matrix_optimization as gram_matrix_optimization, ) from ._srvf import SRSF as SRSF diff --git a/skfda/misc/operators/_identity.py b/skfda/misc/operators/_identity.py index 11f2084a2..43be259cf 100644 --- a/skfda/misc/operators/_identity.py +++ b/skfda/misc/operators/_identity.py @@ -7,7 +7,7 @@ from ...representation import FDataGrid from ...representation.basis import Basis from ...typing._numpy import NDArrayFloat -from ._operators import InputType, Operator, gramian_matrix_optimization +from ._operators import InputType, Operator, gram_matrix_optimization T = TypeVar("T", bound=InputType) @@ -28,7 +28,7 @@ def __call__(self, f: T) -> T: # noqa: D102 return f -@gramian_matrix_optimization.register +@gram_matrix_optimization.register def basis_penalty_matrix_optimized( linear_operator: Identity[Any], basis: Basis, @@ -37,7 +37,7 @@ def basis_penalty_matrix_optimized( return basis.gram_matrix() -@gramian_matrix_optimization.register +@gram_matrix_optimization.register def fdatagrid_penalty_matrix_optimized( linear_operator: Identity[Any], basis: FDataGrid, diff --git a/skfda/misc/operators/_linear_differential_operator.py b/skfda/misc/operators/_linear_differential_operator.py index 02b6f190e..776b4aa72 100644 --- a/skfda/misc/operators/_linear_differential_operator.py +++ b/skfda/misc/operators/_linear_differential_operator.py @@ -18,7 +18,7 @@ ) from ...typing._base import DomainRangeLike from ...typing._numpy import NDArrayFloat -from ._operators import Operator, gramian_matrix_optimization +from ._operators import Operator, gram_matrix_optimization Order = int @@ -218,12 +218,12 @@ def applied_linear_diff_op( ############################################################# # -# Optimized implementations of gramian matrix for each basis. +# Optimized implementations of gram matrix for each basis. # ############################################################# -@gramian_matrix_optimization.register +@gram_matrix_optimization.register def constant_penalty_matrix_optimized( linear_operator: LinearDifferentialOperator, basis: ConstantBasis, @@ -299,7 +299,7 @@ def _monomial_evaluate_constant_linear_diff_op( return polynomials # type: ignore[no-any-return] -@gramian_matrix_optimization.register +@gram_matrix_optimization.register def monomial_penalty_matrix_optimized( linear_operator: LinearDifferentialOperator, basis: MonomialBasis, @@ -427,7 +427,7 @@ def _fourier_penalty_matrix_optimized_orthonormal( return penalty_matrix -@gramian_matrix_optimization.register +@gram_matrix_optimization.register def fourier_penalty_matrix_optimized( linear_operator: LinearDifferentialOperator, basis: FourierBasis, @@ -445,7 +445,7 @@ def fourier_penalty_matrix_optimized( return _fourier_penalty_matrix_optimized_orthonormal(basis, weights) -@gramian_matrix_optimization.register +@gram_matrix_optimization.register def bspline_penalty_matrix_optimized( linear_operator: LinearDifferentialOperator, basis: BSplineBasis, @@ -576,7 +576,7 @@ def bspline_penalty_matrix_optimized( return penalty_matrix -@gramian_matrix_optimization.register +@gram_matrix_optimization.register def fdatagrid_penalty_matrix_optimized( linear_operator: LinearDifferentialOperator, basis: FDataGrid, diff --git a/skfda/misc/operators/_operators.py b/skfda/misc/operators/_operators.py index 770401d7a..221c4ac89 100644 --- a/skfda/misc/operators/_operators.py +++ b/skfda/misc/operators/_operators.py @@ -37,26 +37,26 @@ def __call__(self, vector: OperatorInput) -> OperatorOutput: @multimethod.multidispatch -def gramian_matrix_optimization( +def gram_matrix_optimization( linear_operator: Any, basis: OperatorInput, ) -> NDArrayFloat: """ - Efficient implementation of gramian_matrix. + Efficient implementation of gram_matrix. Generic function that can be subclassed for different combinations of operator and basis in order to provide a more efficient implementation - for the gramian matrix. + for the gram matrix. """ return NotImplemented -def gramian_matrix_numerical( +def gram_matrix_numerical( linear_operator: Operator[OperatorInput, OutputType], basis: OperatorInput, ) -> NDArrayFloat: """ - Return the gramian matrix given a basis, computed numerically. + Return the gram matrix given a basis, computed numerically. This method should work for every linear operator. @@ -70,19 +70,19 @@ def gramian_matrix_numerical( return inner_product_matrix(evaluated_basis, _domain_range=domain_range) -def gramian_matrix( +def gram_matrix( linear_operator: Operator[OperatorInput, OutputType], basis: OperatorInput, ) -> NDArrayFloat: r""" - Return the gramian matrix given a basis. + Return the gram matrix given a basis. - The gramian operator of a linear operator :math:`\Gamma` is + The gram operator of a linear operator :math:`\Gamma` is .. math:: G = \Gamma\Gamma - This method evaluates that gramian operator in a given basis, + This method evaluates that gram operator in a given basis, which is necessary for performing Tikhonov regularization, among other things. @@ -91,11 +91,11 @@ def gramian_matrix( """ # Try to use a more efficient implementation - matrix = gramian_matrix_optimization(linear_operator, basis) + matrix = gram_matrix_optimization(linear_operator, basis) if matrix is not NotImplemented: return matrix - return gramian_matrix_numerical(linear_operator, basis) + return gram_matrix_numerical(linear_operator, basis) class MatrixOperator(Operator[NDArrayFloat, NDArrayFloat]): diff --git a/skfda/misc/regularization/_regularization.py b/skfda/misc/regularization/_regularization.py index 47c3a5432..d8ca23788 100644 --- a/skfda/misc/regularization/_regularization.py +++ b/skfda/misc/regularization/_regularization.py @@ -11,7 +11,7 @@ from ...representation import FData from ...representation.basis import Basis from ...typing._numpy import NDArrayFloat -from ..operators import Identity, Operator, gramian_matrix +from ..operators import Identity, Operator, gram_matrix from ..operators._operators import OperatorInput @@ -101,7 +101,7 @@ def penalty_matrix( else self.linear_operator ) - return self.regularization_parameter gramian_matrix( + return self.regularization_parameter * gram_matrix( linear_operator, basis, )	Rename Gramian matrix to Gram matrix. In some places we use the terminology Gramian matrix (even in function names). We should use always the name Gram matrix for consistency.	I'd like to take this up. I would really appreciate some clarification on the same Sure! There are some places in which we used the terminology "Gramian matrix" instead of the more common "Gram matrix": https://github.com/GAA-UAM/scikit-fda/search?q=gramian In particular, the functions `gramian_matrix`, `gramian_matrix_numerical` and `gramian_matrix_optimization` should be renamed, as well as the related comments. As far as I know, nobody is using these functions in other projects (although `gramian_matrix_optimization` was exposed to be able to provide optimized implementations for particular combinations of basis and operator), so no deprecation would be needed. I assign you to this issue, please tell me in case you have any other doubt. So Basically we have to change Gramian_matrix,gramian_matrix_numerical to Gram matrix to be more consistent. Wherein gramian _matrix_optimization is able to provide optimized implementation so it doesn't need to be changed right?	2022-12-24T11:07:24	[]	[]
GAA-UAM/scikit-fda	GAA-UAM__scikit-fda-436	ebab51dbf57f3f76c0df615d830868425366f425	diff --git a/skfda/misc/hat_matrix.py b/skfda/misc/hat_matrix.py index fc94adc6b..81694bd47 100644 --- a/skfda/misc/hat_matrix.py +++ b/skfda/misc/hat_matrix.py @@ -11,20 +11,21 @@ import abc import math -from typing import Callable, TypeVar, Union, overload +from typing import Callable, Final, TypeVar, Union, overload import numpy as np from .._utils._sklearn_adapter import BaseEstimator from ..representation._functional_data import FData from ..representation.basis import FDataBasis -from ..typing._base import GridPointsLike from ..typing._numpy import NDArrayFloat from . import kernels -Input = TypeVar("Input", bound=Union[FData, GridPointsLike]) +Input = TypeVar("Input", bound=Union[FData, NDArrayFloat]) Prediction = TypeVar("Prediction", bound=Union[NDArrayFloat, FData]) +DEFAULT_BANDWIDTH_PERCENTILE: Final = 15 + class HatMatrix( BaseEstimator, @@ -52,8 +53,8 @@ def __call__( self, , delta_x: NDArrayFloat, - X_train: Input \| None = None, - X: Input \| None = None, + X_train: Input, + X: Input, y_train: None = None, weights: NDArrayFloat \| None = None, _cv: bool = False, @@ -65,8 +66,8 @@ def __call__( self, , delta_x: NDArrayFloat, - X_train: Input \| None = None, - X: Input \| None = None, + X_train: Input, + X: Input, y_train: Prediction \| None = None, weights: NDArrayFloat \| None = None, _cv: bool = False, @@ -77,8 +78,8 @@ def __call__( self, , delta_x: NDArrayFloat, - X_train: Input \| None = None, - X: Input \| None = None, + X_train: Input, + X: Input, y_train: NDArrayFloat \| FData \| None = None, weights: NDArrayFloat \| None = None, _cv: bool = False, @@ -186,7 +187,7 @@ def _hat_matrix_function_not_normalized( ) -> NDArrayFloat: bandwidth = ( - np.percentile(np.abs(delta_x), 15) + float(np.percentile(delta_x, DEFAULT_BANDWIDTH_PERCENTILE)) if self.bandwidth is None else self.bandwidth ) @@ -278,8 +279,8 @@ def __call__( self, , delta_x: NDArrayFloat, - X_train: FData \| GridPointsLike \| None = None, - X: FData \| GridPointsLike \| None = None, + X_train: FData \| NDArrayFloat, + X: FData \| NDArrayFloat, y_train: None = None, weights: NDArrayFloat \| None = None, _cv: bool = False, @@ -291,8 +292,8 @@ def __call__( self, , delta_x: NDArrayFloat, - X_train: FData \| GridPointsLike \| None = None, - X: FData \| GridPointsLike \| None = None, + X_train: FData \| NDArrayFloat, + X: FData \| NDArrayFloat, y_train: Prediction \| None = None, weights: NDArrayFloat \| None = None, _cv: bool = False, @@ -303,19 +304,37 @@ def __call__( # noqa: D102 self, , delta_x: NDArrayFloat, - X_train: FData \| GridPointsLike \| None = None, - X: FData \| GridPointsLike \| None = None, + X_train: FData \| NDArrayFloat, + X: FData \| NDArrayFloat, y_train: NDArrayFloat \| FData \| None = None, weights: NDArrayFloat \| None = None, _cv: bool = False, ) -> NDArrayFloat \| FData: bandwidth = ( - np.percentile(np.abs(delta_x), 15) + float(np.percentile(delta_x, DEFAULT_BANDWIDTH_PERCENTILE)) if self.bandwidth is None else self.bandwidth ) + # Smoothing for functions of one variable + if not isinstance(X_train, FDataBasis) and X_train[0].shape[0] == 1: + delta_x = np.subtract.outer( + X.flatten(), + X_train.flatten(), + ) + + assert y_train is None + + return super().__call__( # noqa: WPS503 + delta_x=delta_x, + X_train=X_train, + X=X, + y_train=y_train, + weights=weights, + _cv=_cv, + ) + # Regression if isinstance(X_train, FData): assert isinstance(X, FData) @@ -326,44 +345,41 @@ def __call__( # noqa: D102 ): raise ValueError("Only FDataBasis is supported for now.") + m1 = X_train.coefficients + m2 = X.coefficients + if y_train is None: y_train = np.identity(X_train.n_samples) - m1 = X_train.coefficients - m2 = X.coefficients + else: + m1 = X_train + m2 = X + + if y_train is None: + y_train = np.identity(X_train.shape[0]) - # Subtract previous matrices obtaining a 3D matrix - # The i-th element contains the matrix X_train - X[i] - C = m1 - m2[:, np.newaxis] + # Subtract previous matrices obtaining a 3D matrix + # The i-th element contains the matrix X_train - X[i] + C = m1 - m2[:, np.newaxis] + # Inner product matrix only is applicable in regression + if isinstance(X_train, FDataBasis): inner_product_matrix = X_train.basis.inner_product_matrix() # Calculate new coefficients taking into account cross-products # if the basis is orthonormal, C would not change C = C @ inner_product_matrix # noqa: WPS350 - # Adding a column of ones in the first position of all matrices - dims = (C.shape[0], C.shape[1], 1) - C = np.concatenate((np.ones(dims), C), axis=-1) - - return self._solve_least_squares( - delta_x=delta_x, - coefs=C, - y_train=y_train, - bandwidth=bandwidth, - ) - - # Smoothing - else: + # Adding a column of ones in the first position of all matrices + dims = (C.shape[0], C.shape[1], 1) + C = np.concatenate((np.ones(dims), C), axis=-1) - return super().__call__( # type: ignore[misc, type-var] # noqa: WPS503 - delta_x=delta_x, - X_train=X_train, - X=X, - y_train=y_train, # type: ignore[arg-type] - weights=weights, - _cv=_cv, - ) + return self._solve_least_squares( + delta_x=delta_x, + coefs=C, + y_train=y_train, + bandwidth=bandwidth, + ) def _solve_least_squares( self, @@ -485,7 +501,7 @@ def _hat_matrix_function_not_normalized( # For each row in the distances matrix, it calculates the furthest # point within the k nearest neighbours vec = np.sort( - np.abs(delta_x), + delta_x, axis=1, )[:, n_neighbors - 1] + tol diff --git a/skfda/preprocessing/smoothing/_kernel_smoothers.py b/skfda/preprocessing/smoothing/_kernel_smoothers.py index c07489235..e9b56f44d 100644 --- a/skfda/preprocessing/smoothing/_kernel_smoothers.py +++ b/skfda/preprocessing/smoothing/_kernel_smoothers.py @@ -9,9 +9,11 @@ import numpy as np -from ..._utils._utils import _to_grid_points +from ..._utils._utils import _cartesian_product, _to_grid_points from ...misc.hat_matrix import HatMatrix, NadarayaWatsonHatMatrix +from ...misc.metrics import PairwiseMetric, l2_distance from ...typing._base import GridPointsLike +from ...typing._metric import Metric from ...typing._numpy import NDArrayFloat from ._linear import _LinearSmoother @@ -114,10 +116,12 @@ def __init__( , weights: Optional[NDArrayFloat] = None, output_points: Optional[GridPointsLike] = None, + metric: Metric[NDArrayFloat] = l2_distance, ): self.kernel_estimator = kernel_estimator self.weights = weights self.output_points = output_points + self.metric = metric self._cv = False # For testing purposes only def _hat_matrix( @@ -126,18 +130,18 @@ def _hat_matrix( output_points: GridPointsLike, ) -> NDArrayFloat: - input_points = _to_grid_points(input_points) - output_points = _to_grid_points(output_points) + input_points = _cartesian_product(_to_grid_points(input_points)) + output_points = _cartesian_product(_to_grid_points(output_points)) if self.kernel_estimator is None: self.kernel_estimator = NadarayaWatsonHatMatrix() - delta_x = np.subtract.outer(output_points[0], input_points[0]) + delta_x = PairwiseMetric(self.metric)(output_points, input_points) return self.kernel_estimator( delta_x=delta_x, weights=self.weights, - X_train=input_points, - X=output_points, + X_train=np.array(input_points), + X=np.array(output_points), _cv=self._cv, ) diff --git a/skfda/preprocessing/smoothing/_linear.py b/skfda/preprocessing/smoothing/_linear.py index 8759704c9..09c6fadcc 100644 --- a/skfda/preprocessing/smoothing/_linear.py +++ b/skfda/preprocessing/smoothing/_linear.py @@ -28,6 +28,7 @@ class _LinearSmoother( ``hat_matrix`` to define the smoothing or 'hat' matrix. """ + input_points_: GridPoints output_points_: GridPoints @@ -116,9 +117,25 @@ def transform( ) ) + dims = [len(e) for e in self.output_points_] + dims += [len(e) for e in self.input_points_] + + hat_matrix = np.reshape( + self.hat_matrix_, + dims, + ) + + data_matrix = np.einsum( + hat_matrix, + [Ellipsis, range(1, len(self.output_points_) + 1)], + X.data_matrix, + [0, *range(1, len(self.output_points_) + 2)], + [0, Ellipsis, len(self.output_points_) + 1], + ) + # The matrix is cached return X.copy( - data_matrix=self.hat_matrix_ @ X.data_matrix, + data_matrix=data_matrix, grid_points=self.output_points_, )	Smoothing in several dimensions The current smoothers only work for one-dimensional functions. It should be reasonably easy to extend them to several dimensions.		2022-03-28T15:25:31	[]	[]
GAA-UAM/scikit-fda	GAA-UAM__scikit-fda-430	c47ec786f903dc35dea466c341a6e947a5e7faa6	diff --git a/examples/plot_kernel_regression.py b/examples/plot_kernel_regression.py index 39273791e..d3cb895bb 100644 --- a/examples/plot_kernel_regression.py +++ b/examples/plot_kernel_regression.py @@ -109,8 +109,8 @@ print('Score NW:', nw_res) ############################################################################## -# For Local Linear Regression, FDataBasis representation with an orthonormal -# basis should be used (for the previous cases it is possible to use either +# For Local Linear Regression, FDataBasis representation with a basis should be +# used (for the previous cases it is possible to use either # FDataGrid or FDataBasis). # # For basis, Fourier basis with 10 elements has been selected. Note that the diff --git a/skfda/misc/hat_matrix.py b/skfda/misc/hat_matrix.py index ae0e9e693..67f51810e 100644 --- a/skfda/misc/hat_matrix.py +++ b/skfda/misc/hat_matrix.py @@ -14,10 +14,9 @@ import numpy as np from sklearn.base import BaseEstimator, RegressorMixin -from skfda.representation._functional_data import FData -from skfda.representation.basis import FDataBasis - -from ..representation._typing import GridPoints, GridPointsLike +from ..representation._functional_data import FData +from ..representation._typing import GridPoints, GridPointsLike, NDArrayFloat +from ..representation.basis import FDataBasis from . import kernels @@ -36,7 +35,7 @@ def __init__( self, , bandwidth: Optional[float] = None, - kernel: Callable[[np.ndarray], np.ndarray] = kernels.normal, + kernel: Callable[[NDArrayFloat], NDArrayFloat] = kernels.normal, ): self.bandwidth = bandwidth self.kernel = kernel @@ -44,13 +43,13 @@ def __init__( def __call__( self, , - delta_x: np.ndarray, + delta_x: NDArrayFloat, X_train: Optional[Union[FData, GridPointsLike]] = None, X: Optional[Union[FData, GridPointsLike]] = None, - y_train: Optional[np.ndarray] = None, - weights: Optional[np.ndarray] = None, + y_train: Optional[NDArrayFloat] = None, + weights: Optional[NDArrayFloat] = None, _cv: bool = False, - ) -> np.ndarray: + ) -> NDArrayFloat: r""" Calculate the hat matrix or the prediction. @@ -99,8 +98,8 @@ def __call__( def _hat_matrix_function_not_normalized( self, , - delta_x: np.ndarray, - ) -> np.ndarray: + delta_x: NDArrayFloat, + ) -> NDArrayFloat: pass @@ -141,8 +140,8 @@ class NadarayaWatsonHatMatrix(HatMatrix): def _hat_matrix_function_not_normalized( self, , - delta_x: np.ndarray, - ) -> np.ndarray: + delta_x: NDArrayFloat, + ) -> NDArrayFloat: if self.bandwidth is None: percentage = 15 @@ -185,7 +184,7 @@ class LocalLinearRegressionHatMatrix(HatMatrix): For kernel regression algorithm: Given functional data, :math:`(X_1, X_2, ..., X_n)` where each function - is expressed in a orthonormal basis with :math:`J` elements and scalar + is expressed in a basis with :math:`J` elements and scalar response :math:`Y = (y_1, y_2, ..., y_n)`. It is desired to estimate the values @@ -222,13 +221,13 @@ class LocalLinearRegressionHatMatrix(HatMatrix): def __call__( # noqa: D102 self, , - delta_x: np.ndarray, + delta_x: NDArrayFloat, X_train: Optional[Union[FDataBasis, GridPoints]] = None, X: Optional[Union[FDataBasis, GridPoints]] = None, - y_train: Optional[np.ndarray] = None, - weights: Optional[np.ndarray] = None, + y_train: Optional[NDArrayFloat] = None, + weights: Optional[NDArrayFloat] = None, _cv: bool = False, - ) -> np.ndarray: + ) -> NDArrayFloat: if self.bandwidth is None: percentage = 15 @@ -243,10 +242,23 @@ def __call__( # noqa: D102 m1 = X_train.coefficients m2 = X.coefficients + # Subtract previous matrices obtaining a 3D matrix + # The i-th element contains the matrix X_train - X[i] + C = m1 - m2[:, np.newaxis] + + inner_product_matrix = X_train.basis.inner_product_matrix() + + # Calculate new coefficients taking into account cross-products + # if the basis is orthonormal, C would not change + C = C @ inner_product_matrix + + # Adding a column of ones in the first position of all matrices + dims = (C.shape[0], C.shape[1], 1) + C = np.concatenate((np.ones(dims), C), axis=-1) + return self._solve_least_squares( delta_x=delta_x, - m1=m1, - m2=m2, + coefs=C, y_train=y_train, ) @@ -264,39 +276,16 @@ def __call__( # noqa: D102 def _solve_least_squares( self, - delta_x: np.ndarray, - m1: np.ndarray, - m2: np.ndarray, - y_train: np.ndarray, - ) -> np.ndarray: + delta_x: NDArrayFloat, + coefs: NDArrayFloat, + y_train: NDArrayFloat, + ) -> NDArrayFloat: W = np.sqrt(self.kernel(delta_x / self.bandwidth)) - # Adding a column of ones to m1 - m1 = np.concatenate( - ( - np.ones(m1.shape[0])[:, np.newaxis], - m1, - ), - axis=1, - ) - - # Adding a column of zeros to m2 - m2 = np.concatenate( - ( - np.zeros(m2.shape[0])[:, np.newaxis], - m2, - ), - axis=1, - ) - - # Subtract previous matrices obtaining a 3D matrix - # The i-th element contains the matrix X_train - X[i] - C = m1 - m2[:, np.newaxis] - # A x = b - # Where x = (a, b_1, ..., b_J) - A = (C.T W.T).T + # Where x = (a, b_1, ..., b_J). + A = (coefs.T * W.T).T b = np.einsum('ij, j... -> ij...', W, y_train) # For Ax = b calculates x that minimize the square error @@ -312,8 +301,8 @@ def _solve_least_squares( def _hat_matrix_function_not_normalized( self, , - delta_x: np.ndarray, - ) -> np.ndarray: + delta_x: NDArrayFloat, + ) -> NDArrayFloat: if self.bandwidth is None: percentage = 15 @@ -369,7 +358,7 @@ def __init__( self, , n_neighbors: Optional[int] = None, - kernel: Callable[[np.ndarray], np.ndarray] = kernels.uniform, + kernel: Callable[[NDArrayFloat], NDArrayFloat] = kernels.uniform, ): self.n_neighbors = n_neighbors self.kernel = kernel @@ -377,8 +366,8 @@ def __init__( def _hat_matrix_function_not_normalized( self, *, - delta_x: np.ndarray, - ) -> np.ndarray: + delta_x: NDArrayFloat, + ) -> NDArrayFloat: input_points_len = delta_x.shape[1] diff --git a/skfda/ml/regression/_kernel_regression.py b/skfda/ml/regression/_kernel_regression.py index 2834cae86..aecc64c60 100644 --- a/skfda/ml/regression/_kernel_regression.py +++ b/skfda/ml/regression/_kernel_regression.py @@ -6,10 +6,10 @@ from sklearn.base import BaseEstimator, RegressorMixin from sklearn.utils.validation import check_is_fitted -from skfda.misc.hat_matrix import HatMatrix, NadarayaWatsonHatMatrix -from skfda.misc.metrics import PairwiseMetric, l2_distance -from skfda.misc.metrics._typing import Metric -from skfda.representation._functional_data import FData +from ...misc.hat_matrix import HatMatrix, NadarayaWatsonHatMatrix +from ...misc.metrics import PairwiseMetric, l2_distance +from ...misc.metrics._typing import Metric +from ...representation._functional_data import FData class KernelRegression(	Allow non orthonormal basis in local linear regression _Originally posted by @vnmabus in https://github.com/GAA-UAM/scikit-fda/pull/417#r800112765_		2022-03-04T03:06:38	[]	[]
GAA-UAM/scikit-fda	GAA-UAM__scikit-fda-429	46fc684ff9d96bc282e8646d7ac0e26c4f107ee4	diff --git a/docs/modules/misc.rst b/docs/modules/misc.rst index c1a53d05a..6d1669f9c 100644 --- a/docs/modules/misc.rst +++ b/docs/modules/misc.rst @@ -57,4 +57,4 @@ functional data: misc/operators misc/regularization misc/hat_matrix - misc/score_functions + misc/scoring diff --git a/docs/modules/misc/scoring.rst b/docs/modules/misc/scoring.rst new file mode 100644 index 000000000..b1b5e2fd2 --- /dev/null +++ b/docs/modules/misc/scoring.rst @@ -0,0 +1,18 @@ +Scoring methods for regression with functional response. +======================================================== + +The functions in this module are a generalization for functional data of +the regression metrics of the sklearn library +(https://scikit-learn.org/stable/modules/model_evaluation.html#regression-metrics). +Only scores that support multioutput are included. + + +.. autosummary:: + :toctree: autosummary + + skfda.misc.scoring.explained_variance_score + skfda.misc.scoring.mean_absolute_error + skfda.misc.scoring.mean_absolute_percentage_error + skfda.misc.scoring.mean_squared_error + skfda.misc.scoring.mean_squared_log_error + skfda.misc.scoring.r2_score diff --git a/skfda/_utils/_sklearn_adapter.py b/skfda/_utils/_sklearn_adapter.py index 3c23aa330..f3eaecb0b 100644 --- a/skfda/_utils/_sklearn_adapter.py +++ b/skfda/_utils/_sklearn_adapter.py @@ -184,8 +184,10 @@ def score( # noqa: D102 y: TargetPrediction, sample_weight: NDArrayFloat \| None = None, ) -> float: - return super().score( # type: ignore[no-any-return] - X, + from ..misc.scoring import r2_score + y_pred = self.predict(X) + return r2_score( y, + y_pred, sample_weight=sample_weight, ) diff --git a/skfda/exploratory/stats/_stats.py b/skfda/exploratory/stats/_stats.py index 27eeac78b..ac89a992b 100644 --- a/skfda/exploratory/stats/_stats.py +++ b/skfda/exploratory/stats/_stats.py @@ -18,25 +18,32 @@ T = TypeVar('T', bound=Union[NDArrayFloat, FData]) -def mean(X: F) -> F: +def mean( + X: F, + weights: NDArrayFloat \| None = None, +) -> F: """ Compute the mean of all the samples in a FData object. Args: X: Object containing all the samples whose mean is wanted. - + weights: Sample weight. By default, uniform weight are used. Returns: Mean of all the samples in the original object, as a :term:`functional data object` with just one sample. """ - return X.mean() + if weights is None: + return X.mean() + + weight = (1 / np.sum(weights)) * weights + return (X * weight).sum() def var(X: FData) -> FDataGrid: """ - Compute the variance of a set of samples in a FDataGrid object. + Compute the variance of a set of samples in a FData object. Args: X: Object containing all the set of samples whose variance is desired. diff --git a/skfda/misc/__init__.py b/skfda/misc/__init__.py index cdd16e6d4..6a977724a 100644 --- a/skfda/misc/__init__.py +++ b/skfda/misc/__init__.py @@ -13,6 +13,7 @@ "metrics", "operators", "regularization", + "scoring", "validation", ], submod_attrs={ diff --git a/skfda/misc/scoring.py b/skfda/misc/scoring.py new file mode 100644 index 000000000..c57902668 --- /dev/null +++ b/skfda/misc/scoring.py @@ -0,0 +1,1073 @@ +"""Scoring methods for FData.""" +from __future__ import annotations + +import math +import warnings +from functools import singledispatch +from typing import Callable, Optional, TypeVar, Union, overload + +import numpy as np +import sklearn.metrics +from typing_extensions import Literal, Protocol + +from .._utils import nquad_vec +from ..representation import FData, FDataBasis, FDataGrid +from ..representation._functional_data import EvalPointsType +from ..typing._numpy import NDArrayFloat + +DataType = TypeVar('DataType') + +MultiOutputType = Literal['uniform_average', 'raw_values'] + + +class _InfiniteScoreError(Exception): + """Exception for skipping integral on infinite value.""" + + +class ScoreFunction(Protocol): + """Type definition for score functions.""" + + @overload + def __call__( + self, + y_true: DataType, + y_pred: DataType, + , + sample_weight: NDArrayFloat \| None = None, + multioutput: Literal['uniform_average'] = 'uniform_average', + ) -> float: + pass # noqa: WPS428 + + @overload + def __call__( # noqa: D102 + self, + y_true: DataType, + y_pred: DataType, + , + sample_weight: NDArrayFloat \| None = None, + multioutput: Literal['raw_values'], + ) -> DataType: + pass # noqa: WPS428 + + def __call__( # noqa: D102 + self, + y_true: DataType, + y_pred: DataType, + , + sample_weight: NDArrayFloat \| None = None, + multioutput: MultiOutputType = 'uniform_average', + ) -> float \| DataType: + pass # noqa: WPS428 + + +def _domain_measure(fd: FData) -> float: + measure = 1.0 + for interval in fd.domain_range: + measure = measure (interval[1] - interval[0]) + return measure + + +def _var( + x: FDataGrid, + weights: NDArrayFloat \| None = None, +) -> FDataGrid: + from ..exploratory.stats import mean, var + + if weights is None: + return var(x) + + return mean( # type: ignore[no-any-return] + np.power(x - mean(x, weights=weights), 2), + weights=weights, + ) + + +def _multioutput_score_basis( + y_true: FDataBasis, + multioutput: MultiOutputType, + integrand: Callable[[NDArrayFloat], NDArrayFloat], +) -> float: + + if multioutput != "uniform_average": + raise ValueError( + f"Only \"uniform_average\" is supported for \"multioutput\" when " + f"the input is a FDatabasis: received {multioutput} instead", + ) + + try: + integral = nquad_vec( + integrand, + y_true.domain_range, + ) + except _InfiniteScoreError: + return -math.inf + + # If the dimension of the codomain is > 1, + # the mean of the scores is taken + return float(np.mean(integral) / _domain_measure(y_true)) + + +def _multioutput_score_grid( + score: FDataGrid, + multioutput: MultiOutputType, + squared: bool = True, +) -> float \| FDataGrid: + + if not squared: + score = np.sqrt(score) + + if multioutput == 'raw_values': + return score + + # Score only contains 1 function + # If the dimension of the codomain is > 1, + # the mean of the scores is taken + return float(np.mean(score.integrate()[0]) / _domain_measure(score)) + + +@overload +def explained_variance_score( + y_true: DataType, + y_pred: DataType, + , + sample_weight: NDArrayFloat \| None = None, + multioutput: Literal['uniform_average'] = 'uniform_average', +) -> float: + pass # noqa: WPS428 + + +@overload +def explained_variance_score( + y_true: DataType, + y_pred: DataType, + , + sample_weight: NDArrayFloat \| None = None, + multioutput: Literal['raw_values'], +) -> DataType: + pass # noqa: WPS428 + + +@singledispatch +def explained_variance_score( + y_true: DataType, + y_pred: DataType, + , + sample_weight: NDArrayFloat \| None = None, + multioutput: MultiOutputType = 'uniform_average', +) -> float \| DataType: + r"""Explained variance score for :class:`~skfda.representation.FData`. + + With :math:`y\_true = (X_1, X_2, ..., X_n)` being the real values, + :math:`t\_pred = (\hat{X}_1, \hat{X}_2, ..., \hat{X}_n)` being the + estimated and :math:`sample\_weight = (w_1, w_2, ..., w_n)`, the score is + calculated as + + .. math:: + EV(y\_true, y\_pred)(t) = 1 - + \frac{Var(y\_true(t) - y\_pred(t), sample\_weight)} + {Var(y\_true(t), sample\_weight)} + + where :math:`Var` is a weighted variance. + + Weighted variance is defined as below + + .. math:: + Var(y\_true, sample\_weight)(t) = \sum_{i=1}^n w_i + (X_i(t) - Mean(fd(t), sample\_weight))^2. + + Here, :math:`Mean` is a weighted mean. + + For :math:`y\_true` and :math:`y\_pred` of type + :class:`~skfda.representation.FDataGrid`, :math:`EV` is + also a :class:`~skfda.representation.FDataGrid` object with + the same grid points. + + If multioutput = 'raw_values', the function :math:`EV` is returned. + Otherwise, if multioutput = 'uniform_average', the mean of :math:`EV` is + calculated: + + .. math:: + mean(EV) = \frac{1}{V}\int_{D} EV(t) dt + + where :math:`D` is the function domain and :math:`V` the volume of that + domain. + + For :class:`~skfda.representation.FDataBasis` only + 'uniform_average' is available. + + If :math:`y\_true` and :math:`y\_pred` are numpy arrays, sklearn function + is called. + + The best possible score is 1.0, lower values are worse. + + Args: + y_true: Correct target values. + y_pred: Estimated values. + sample_weight: Sample weights. By default, uniform weights + are taken. + multioutput: Defines format of the return. + + Returns: + Explained variance score. + + If multioutput = 'uniform_average' or + :math:`y\_pred` and :math:`y\_true` are + :class:`~skfda.representation.FDataBasis` objects, float is returned. + + If both :math:`y\_pred` and :math:`y\_true` are + :class:`~skfda.representation.FDataGrid` + objects and multioutput = 'raw_values', + :class:`~skfda.representation.FDataGrid` is returned. + + If both :math:`y\_pred` and :math:`y\_true` are ndarray and + multioutput = 'raw_values', ndarray. + + """ + return ( # type: ignore [no-any-return] + sklearn.metrics.explained_variance_score( + y_true, + y_pred, + sample_weight=sample_weight, + multioutput=multioutput, + ) + ) + + +@explained_variance_score.register # type: ignore[attr-defined, misc] +def _explained_variance_score_fdatagrid( + y_true: FDataGrid, + y_pred: FDataGrid, + , + sample_weight: Optional[NDArrayFloat] = None, + multioutput: MultiOutputType = 'uniform_average', +) -> Union[float, FDataGrid]: + + num = _var(y_true - y_pred, weights=sample_weight) + den = _var(y_true, weights=sample_weight) + + # Divisions by zero allowed + with np.errstate(divide='ignore', invalid='ignore'): + score = 1 - num / den + + # 0 / 0 divisions should be 0 in this context, and the score, 1 + score.data_matrix[np.isnan(score.data_matrix)] = 1 + + return _multioutput_score_grid(score, multioutput) + + +@explained_variance_score.register # type: ignore[attr-defined, misc] +def _explained_variance_score_fdatabasis( + y_true: FDataBasis, + y_pred: FDataBasis, + , + sample_weight: Optional[NDArrayFloat] = None, + multioutput: MultiOutputType = 'uniform_average', +) -> float: + + def _ev_func(x: EvalPointsType) -> NDArrayFloat: # noqa: WPS430 + num = np.average( + ( + (y_true(x) - y_pred(x)) + - np.average( + y_true(x) - y_pred(x), + weights=sample_weight, + axis=0, + ) + ) * 2, + weights=sample_weight, + axis=0, + ) + + den = np.average( + ( + y_true(x) + - np.average(y_true(x), weights=sample_weight, axis=0) + ) ** 2, + weights=sample_weight, + axis=0, + ) + + # Divisions by zero allowed + with np.errstate(divide='ignore', invalid='ignore'): + score = 1 - num / den + + # 0/0 case, the score is 1. + score[np.isnan(score)] = 1 + + # r/0 case, r!= 0. Return -inf outside this function + if np.any(np.isinf(score)): + raise _InfiniteScoreError + + # Score only contains 1 input point + assert score.shape[0] == 1 + return score[0] # type: ignore [no-any-return] + + return _multioutput_score_basis(y_true, multioutput, _ev_func) + + +@overload +def mean_absolute_error( + y_true: DataType, + y_pred: DataType, + , + sample_weight: NDArrayFloat \| None = None, + multioutput: Literal['uniform_average'] = 'uniform_average', +) -> float: + pass # noqa: WPS428 + + +@overload +def mean_absolute_error( + y_true: DataType, + y_pred: DataType, + , + sample_weight: NDArrayFloat \| None = None, + multioutput: Literal['raw_values'], +) -> DataType: + pass # noqa: WPS428 + + +@singledispatch +def mean_absolute_error( + y_true: DataType, + y_pred: DataType, + , + sample_weight: NDArrayFloat \| None = None, + multioutput: MultiOutputType = 'uniform_average', +) -> float \| DataType: + r"""Mean Absolute Error for :class:`~skfda.representation.FData`. + + With :math:`y\_true = (X_1, X_2, ..., X_n)` being the real values, + :math:`t\_pred = (\hat{X}_1, \hat{X}_2, ..., \hat{X}_n)` being the + estimated and :math:`sample\_weight = (w_1, w_2, ..., w_n)`, the error is + calculated as + + .. math:: + MAE(y\_true, y\_pred)(t) = \frac{1}{\sum w_i} + \sum_{i=1}^n w_i\|X_i(t) - \hat{X}_i(t)\| + + For :math:`y\_true` and :math:`y\_pred` of type + :class:`~skfda.representation.FDataGrid`, :math:`MAE` is + also a :class:`~skfda.representation.FDataGrid` object with + the same grid points. + + If multioutput = 'raw_values', the function :math:`MAE` is returned. + Otherwise, if multioutput = 'uniform_average', the mean of :math:`MAE` is + calculated: + + .. math:: + mean(MAE) = \frac{1}{V}\int_{D} MAE(t) dt + + where :math:`D` is the function domain and :math:`V` the volume of that + domain. + + For :class:`~skfda.representation.FDataBasis` only + 'uniform_average' is available. + + If :math:`y\_true` and :math:`y\_pred` are numpy arrays, sklearn function + is called. + + Args: + y_true: Correct target values. + y_pred: Estimated values. + sample_weight: Sample weights. By default, uniform weights + are taken. + multioutput: Defines format of the return. + + Returns: + Mean absolute error. + + If multioutput = 'uniform_average' or + :math:`y\_pred` and :math:`y\_true` are + :class:`~skfda.representation.FDataBasis` objects, float is returned. + + If both :math:`y\_pred` and :math:`y\_true` are + :class:`~skfda.representation.FDataGrid` + objects and multioutput = 'raw_values', + :class:`~skfda.representation.FDataGrid` is returned. + + If both :math:`y\_pred` and :math:`y\_true` are ndarray and + multioutput = 'raw_values', ndarray. + + """ + return sklearn.metrics.mean_absolute_error( # type: ignore [no-any-return] + y_true, + y_pred, + sample_weight=sample_weight, + multioutput=multioutput, + ) + + +@mean_absolute_error.register # type: ignore[attr-defined, misc] +def _mean_absolute_error_fdatagrid( + y_true: FDataGrid, + y_pred: FDataGrid, + , + sample_weight: Optional[NDArrayFloat] = None, + multioutput: MultiOutputType = 'uniform_average', +) -> Union[float, FDataGrid]: + from ..exploratory.stats import mean + + error = mean(np.abs(y_true - y_pred), weights=sample_weight) + return _multioutput_score_grid(error, multioutput) + + +@mean_absolute_error.register # type: ignore[attr-defined, misc] +def _mean_absolute_error_fdatabasis( + y_true: FDataBasis, + y_pred: FDataBasis, + , + sample_weight: Optional[NDArrayFloat] = None, + multioutput: MultiOutputType = 'uniform_average', +) -> float: + + def _mae_func(x: EvalPointsType) -> NDArrayFloat: # noqa: WPS430 + error = np.average( + np.abs(y_true(x) - y_pred(x)), + weights=sample_weight, + axis=0, + ) + + # Error only contains 1 input point + assert error.shape[0] == 1 + return error[0] # type: ignore [no-any-return] + + return _multioutput_score_basis(y_true, multioutput, _mae_func) + + +@overload +def mean_absolute_percentage_error( + y_true: DataType, + y_pred: DataType, + , + sample_weight: NDArrayFloat \| None = None, + multioutput: Literal['uniform_average'] = 'uniform_average', +) -> float: + pass # noqa: WPS428 + + +@overload +def mean_absolute_percentage_error( + y_true: DataType, + y_pred: DataType, + , + sample_weight: NDArrayFloat \| None = None, + multioutput: Literal['raw_values'], +) -> DataType: + pass # noqa: WPS428 + + +@singledispatch +def mean_absolute_percentage_error( + y_true: DataType, + y_pred: DataType, + , + sample_weight: NDArrayFloat \| None = None, + multioutput: MultiOutputType = 'uniform_average', +) -> float \| DataType: + r"""Mean Absolute Percentage Error for :class:`~skfda.representation.FData`. + + With :math:`y\_true = (X_1, X_2, ..., X_n)` being the real values, + :math:`t\_pred = (\hat{X}_1, \hat{X}_2, ..., \hat{X}_n)` being the + estimated and :math:`sample\_weight = (w_1, w_2, ..., w_n)`, the error is + calculated as + + .. math:: + MAPE(y\_true, y\_pred)(t) = \frac{1}{\sum w_i} + \sum_{i=1}^n w_i\frac{\|X_i(t) - \hat{X}_i(t)\|}{\|X_i(t)\|} + + For :math:`y\_true` and :math:`y\_pred` of type + :class:`~skfda.representation.FDataGrid`, :math:`MAPE` is + also a :class:`~skfda.representation.FDataGrid` object with + the same grid points. + + If multioutput = 'raw_values', the function :math:`MAPE` is returned. + Otherwise, if multioutput = 'uniform_average', the mean of :math:`MAPE` is + calculated: + + .. math:: + mean(MAPE) = \frac{1}{V}\int_{D} MAPE(t) dt + + where :math:`D` is the function domain and :math:`V` the volume of that + domain. + + For :class:`~skfda.representation.FDataBasis` only + 'uniform_average' is available. + + If :math:`y\_true` and :math:`y\_pred` are numpy arrays, sklearn function + is called. + + This function should not be used if for some :math:`t` and some :math:`i`, + :math:`X_i(t) = 0`. + + Args: + y_true: Correct target values. + y_pred: Estimated values. + sample_weight: Sample weights. By default, uniform weights + are taken. + multioutput: Defines format of the return. + + Returns: + Mean absolute percentage error. + + If multioutput = 'uniform_average' or + :math:`y\_pred` and :math:`y\_true` are + :class:`~skfda.representation.FDataBasis` objects, float is returned. + + If both :math:`y\_pred` and :math:`y\_true` are + :class:`~skfda.representation.FDataGrid` + objects and multioutput = 'raw_values', + :class:`~skfda.representation.FDataGrid` is returned. + + If both :math:`y\_pred` and :math:`y\_true` are ndarray and + multioutput = 'raw_values', ndarray. + + """ + return ( # type: ignore [no-any-return] + sklearn.metrics.mean_absolute_percentage_error( + y_true, + y_pred, + sample_weight=sample_weight, + multioutput=multioutput, + ) + ) + + +@mean_absolute_percentage_error.register # type: ignore[attr-defined, misc] +def _mean_absolute_percentage_error_fdatagrid( + y_true: FDataGrid, + y_pred: FDataGrid, + , + sample_weight: Optional[NDArrayFloat] = None, + multioutput: MultiOutputType = 'uniform_average', +) -> Union[float, FDataGrid]: + from ..exploratory.stats import mean + + epsilon = np.finfo(np.float64).eps + + if np.any(np.abs(y_true.data_matrix) < epsilon): + warnings.warn('Zero denominator', RuntimeWarning) + + mape = np.abs(y_pred - y_true) / np.maximum(np.abs(y_true), epsilon) + + error = mean(mape, weights=sample_weight) + return _multioutput_score_grid(error, multioutput) + + +@mean_absolute_percentage_error.register # type: ignore[attr-defined, misc] +def _mean_absolute_percentage_error_fdatabasis( + y_true: FDataBasis, + y_pred: FDataBasis, + , + sample_weight: Optional[NDArrayFloat] = None, + multioutput: MultiOutputType = 'uniform_average', +) -> float: + + def _mape_func(x: EvalPointsType) -> NDArrayFloat: # noqa: WPS430 + + epsilon = np.finfo(np.float64).eps + if np.any(np.abs(y_true(x)) < epsilon): + warnings.warn('Zero denominator', RuntimeWarning) + + error = np.average( + ( + np.abs(y_true(x) - y_pred(x)) + / np.maximum(np.abs(y_true(x)), epsilon) + ), + weights=sample_weight, + axis=0, + ) + + # Error only contains 1 input point + assert error.shape[0] == 1 + return error[0] # type: ignore [no-any-return] + + return _multioutput_score_basis(y_true, multioutput, _mape_func) + + +@overload +def mean_squared_error( + y_true: DataType, + y_pred: DataType, + , + sample_weight: NDArrayFloat \| None = None, + multioutput: Literal['uniform_average'] = 'uniform_average', + squared: bool = True, +) -> float: + pass # noqa: WPS428 + + +@overload +def mean_squared_error( + y_true: DataType, + y_pred: DataType, + , + sample_weight: NDArrayFloat \| None = None, + multioutput: Literal['raw_values'], + squared: bool = True, +) -> DataType: + pass # noqa: WPS428 + + +@singledispatch +def mean_squared_error( + y_true: DataType, + y_pred: DataType, + , + sample_weight: NDArrayFloat \| None = None, + multioutput: MultiOutputType = 'uniform_average', + squared: bool = True, +) -> float \| DataType: + r"""Mean Squared Error for :class:`~skfda.representation.FData`. + + With :math:`y\_true = (X_1, X_2, ..., X_n)` being the real values, + :math:`t\_pred = (\hat{X}_1, \hat{X}_2, ..., \hat{X}_n)` being the + estimated and :math:`sample\_weight = (w_1, w_2, ..., w_n)`, the error is + calculated as + + .. math:: + MSE(y\_true, y\_pred)(t) = \frac{1}{\sum w_i} + \sum_{i=1}^n w_i(X_i(t) - \hat{X}_i(t))^2 + + For :math:`y\_true` and :math:`y\_pred` of type + :class:`~skfda.representation.FDataGrid`, :math:`MSE` is + also a :class:`~skfda.representation.FDataGrid` object with + the same grid points. + + If multioutput = 'raw_values', the function :math:`MSE` is returned. + Otherwise, if multioutput = 'uniform_average', the mean of :math:`MSE` is + calculated: + + .. math:: + mean(MSE) = \frac{1}{V}\int_{D} MSE(t) dt + + where :math:`D` is the function domain and :math:`V` the volume of that + domain. + + For :class:`~skfda.representation.FDataBasis` only + 'uniform_average' is available. + + If :math:`y\_true` and :math:`y\_pred` are numpy arrays, sklearn function + is called. + + Args: + y_true: Correct target values. + y_pred: Estimated values. + sample_weight: Sample weights. By default, uniform weights + are taken. + multioutput: Defines format of the return. + squared: If True returns MSE value, if False returns RMSE value. + + Returns: + Mean squared error. + + If multioutput = 'uniform_average' or + :math:`y\_pred` and :math:`y\_true` are + :class:`~skfda.representation.FDataBasis` objects, float is returned. + + If both :math:`y\_pred` and :math:`y\_true` are + :class:`~skfda.representation.FDataGrid` + objects and multioutput = 'raw_values', + :class:`~skfda.representation.FDataGrid` is returned. + + If both :math:`y\_pred` and :math:`y\_true` are ndarray and + multioutput = 'raw_values', ndarray. + + """ + return sklearn.metrics.mean_squared_error( # type: ignore [no-any-return] + y_true, + y_pred, + sample_weight=sample_weight, + multioutput=multioutput, + squared=squared, + ) + + +@mean_squared_error.register # type: ignore[attr-defined, misc] +def _mean_squared_error_fdatagrid( + y_true: FDataGrid, + y_pred: FDataGrid, + , + sample_weight: Optional[NDArrayFloat] = None, + multioutput: MultiOutputType = 'uniform_average', + squared: bool = True, +) -> Union[float, FDataGrid]: + from ..exploratory.stats import mean + + error: FDataGrid = mean( + np.power(y_true - y_pred, 2), + weights=sample_weight, + ) + + return _multioutput_score_grid(error, multioutput, squared=squared) + + +@mean_squared_error.register # type: ignore[attr-defined, misc] +def _mean_squared_error_fdatabasis( + y_true: FDataBasis, + y_pred: FDataBasis, + , + sample_weight: Optional[NDArrayFloat] = None, + squared: bool = True, + multioutput: MultiOutputType = 'uniform_average', +) -> float: + + def _mse_func(x: EvalPointsType) -> NDArrayFloat: # noqa: WPS430 + + error: NDArrayFloat = np.average( + (y_true(x) - y_pred(x)) * 2, + weights=sample_weight, + axis=0, + ) + + if not squared: + return np.sqrt(error) + + # Error only contains 1 input point + assert error.shape[0] == 1 + return error[0] # type: ignore [no-any-return] + + return _multioutput_score_basis(y_true, multioutput, _mse_func) + + +@overload +def mean_squared_log_error( + y_true: DataType, + y_pred: DataType, + , + sample_weight: NDArrayFloat \| None = None, + multioutput: Literal['uniform_average'] = 'uniform_average', + squared: bool = True, +) -> float: + pass # noqa: WPS428 + + +@overload +def mean_squared_log_error( + y_true: DataType, + y_pred: DataType, + , + sample_weight: NDArrayFloat \| None = None, + multioutput: Literal['raw_values'], + squared: bool = True, +) -> DataType: + pass # noqa: WPS428 + + +@singledispatch +def mean_squared_log_error( + y_true: DataType, + y_pred: DataType, + , + sample_weight: NDArrayFloat \| None = None, + multioutput: MultiOutputType = 'uniform_average', + squared: bool = True, +) -> float \| DataType: + r"""Mean Squared Log Error for :class:`~skfda.representation.FData`. + + With :math:`y\_true = (X_1, X_2, ..., X_n)` being the real values, + :math:`t\_pred = (\hat{X}_1, \hat{X}_2, ..., \hat{X}_n)` being the + estimated and :math:`sample\_weight = (w_1, w_2, ..., w_n)`, the error is + calculated as + + .. math:: + MSLE(y\_true, y\_pred)(t) = \frac{1}{\sum w_i} + \sum_{i=1}^n w_i(\log(1 + X_i(t)) - \log(1 + \hat{X}_i(t)))^2 + + where :math:`\log` is the natural logarithm. + + For :math:`y\_true` and :math:`y\_pred` of type + :class:`~skfda.representation.FDataGrid`, :math:`MSLE` is + also a :class:`~skfda.representation.FDataGrid` object with + the same grid points. + + If multioutput = 'raw_values', the function :math:`MSLE` is returned. + Otherwise, if multioutput = 'uniform_average', the mean of :math:`MSLE` is + calculated: + + .. math:: + mean(MSLE) = \frac{1}{V}\int_{D} MSLE(t) dt + + where :math:`D` is the function domain and :math:`V` the volume of that + domain. + + For :class:`~skfda.representation.FDataBasis` only + 'uniform_average' is available. + + If :math:`y\_true` and :math:`y\_pred` are numpy arrays, sklearn function + is called. + + This function should not be used if for some :math:`t` and some :math:`i`, + :math:`X_i(t) < 0`. + + Args: + y_true: Correct target values. + y_pred: Estimated values. + sample_weight: Sample weights. By default, uniform weights + are taken. + multioutput: Defines format of the return. + squared: default True. If False, square root is taken. + + Returns: + Mean squared log error. + + If multioutput = 'uniform_average' or + :math:`y\_pred` and :math:`y\_true` are + :class:`~skfda.representation.FDataBasis` objects, float is returned. + + If both :math:`y\_pred` and :math:`y\_true` are + :class:`~skfda.representation.FDataGrid` + objects and multioutput = 'raw_values', + :class:`~skfda.representation.FDataGrid` is returned. + + If both :math:`y\_pred` and :math:`y\_true` are ndarray and + multioutput = 'raw_values', ndarray. + + """ + return ( # type: ignore [no-any-return] + sklearn.metrics.mean_squared_log_error( + y_true, + y_pred, + sample_weight=sample_weight, + multioutput=multioutput, + squared=squared, + ) + ) + + +@mean_squared_log_error.register # type: ignore[attr-defined, misc] +def _mean_squared_log_error_fdatagrid( + y_true: FDataGrid, + y_pred: FDataGrid, + , + sample_weight: Optional[NDArrayFloat] = None, + multioutput: MultiOutputType = 'uniform_average', + squared: bool = True, +) -> Union[float, FDataGrid]: + + if np.any(y_true.data_matrix < 0) or np.any(y_pred.data_matrix < 0): + raise ValueError( + "Mean Squared Logarithmic Error cannot be used when " + "targets functions have negative values.", + ) + + return mean_squared_error( + np.log1p(y_true), + np.log1p(y_pred), + sample_weight=sample_weight, + multioutput=multioutput, + squared=squared, + ) + + +@mean_squared_log_error.register # type: ignore[attr-defined, misc] +def _mean_squared_log_error_fdatabasis( + y_true: FDataBasis, + y_pred: FDataBasis, + , + sample_weight: Optional[NDArrayFloat] = None, + squared: bool = True, + multioutput: MultiOutputType = 'uniform_average', +) -> float: + + def _msle_func(x: EvalPointsType) -> NDArrayFloat: # noqa: WPS430 + + y_true_eval = y_true(x) + y_pred_eval = y_pred(x) + + if np.any(y_true_eval < 0) or np.any(y_pred_eval < 0): + raise ValueError( + "Mean Squared Logarithmic Error cannot be used when " + "targets functions have negative values.", + ) + + error: NDArrayFloat = np.average( + (np.log1p(y_true_eval) - np.log1p(y_pred_eval)) * 2, + weights=sample_weight, + axis=0, + ) + + if not squared: + return np.sqrt(error) + + # Error only contains 1 input point + assert error.shape[0] == 1 + return error[0] # type: ignore [no-any-return] + + return _multioutput_score_basis(y_true, multioutput, _msle_func) + + +@overload +def r2_score( + y_true: DataType, + y_pred: DataType, + , + sample_weight: NDArrayFloat \| None = None, + multioutput: Literal['uniform_average'] = 'uniform_average', +) -> float: + pass # noqa: WPS428 + + +@overload +def r2_score( + y_true: DataType, + y_pred: DataType, + , + sample_weight: NDArrayFloat \| None = None, + multioutput: Literal['raw_values'], +) -> DataType: + pass # noqa: WPS428 + + +@singledispatch +def r2_score( + y_true: DataType, + y_pred: DataType, + , + sample_weight: NDArrayFloat \| None = None, + multioutput: MultiOutputType = 'uniform_average', +) -> float \| DataType: + r"""R^2 score for :class:`~skfda.representation.FData`. + + With :math:`y\_true = (X_1, X_2, ..., X_n)` being the real values, + :math:`t\_pred = (\hat{X}_1, \hat{X}_2, ..., \hat{X}_n)` being the + estimated and :math:`sample\_weight = (w_1, w_2, ..., w_n)`, the score is + calculated as + + .. math:: + R^2(y\_true, y\_pred)(t) = 1 - + \frac{\sum_{i=1}^n w_i (X_i(t) - \hat{X}_i(t))^2} + {\sum_{i=1}^n w_i (X_i(t) - Mean(y\_true, sample\_weight)(t))^2} + + where :math:`Mean` is a weighted mean. + + For :math:`y\_true` and :math:`y\_pred` of type + :class:`~skfda.representation.FDataGrid`, :math:`R^2` is + also a :class:`~skfda.representation.FDataGrid` object with + the same grid points. + + If multioutput = 'raw_values', the function :math:`R^2` is returned. + Otherwise, if multioutput = 'uniform_average', the mean of :math:`R^2` is + calculated: + + .. math:: + mean(R^2) = \frac{1}{V}\int_{D} R^2(t) dt + + where :math:`D` is the function domain and :math:`V` the volume of that + domain. + + For :class:`~skfda.representation.FDataBasis` only + 'uniform_average' is available. + + If :math:`y\_true` and :math:`y\_pred` are numpy arrays, sklearn function + is called. + + Args: + y_true: Correct target values. + y_pred: Estimated values. + sample_weight: Sample weights. By default, uniform weights + are taken. + multioutput: Defines format of the return. + + Returns: + R2 score + + If multioutput = 'uniform_average' or + :math:`y\_pred` and :math:`y\_true` are + :class:`~skfda.representation.FDataBasis` objects, float is returned. + + If both :math:`y\_pred` and :math:`y\_true` are + :class:`~skfda.representation.FDataGrid` + objects and multioutput = 'raw_values', + :class:`~skfda.representation.FDataGrid` is returned. + + If both :math:`y\_pred` and :math:`y\_true` are ndarray and + multioutput = 'raw_values', ndarray. + + """ + return sklearn.metrics.r2_score( # type: ignore [no-any-return] + y_true, + y_pred, + sample_weight=sample_weight, + multioutput=multioutput, + ) + + +@r2_score.register # type: ignore[attr-defined, misc] +def _r2_score_fdatagrid( + y_true: FDataGrid, + y_pred: FDataGrid, + , + sample_weight: Optional[NDArrayFloat] = None, + multioutput: MultiOutputType = 'uniform_average', +) -> Union[float, FDataGrid]: + from ..exploratory.stats import mean + + if y_pred.n_samples < 2: + raise ValueError( + 'R^2 score is not well-defined with less than two samples.', + ) + + ss_res = mean( + np.power(y_true - y_pred, 2), + weights=sample_weight, + ) + + ss_tot = _var(y_true, weights=sample_weight) + + # Divisions by zero allowed + with np.errstate(divide='ignore', invalid='ignore'): + score: FDataGrid = 1 - ss_res / ss_tot + + # 0 / 0 divisions should be 0 in this context and the score, 1 + score.data_matrix[np.isnan(score.data_matrix)] = 1 + + return _multioutput_score_grid(score, multioutput) + + +@r2_score.register # type: ignore[attr-defined, misc] +def _r2_score_fdatabasis( + y_true: FDataBasis, + y_pred: FDataBasis, + , + sample_weight: Optional[NDArrayFloat] = None, + multioutput: MultiOutputType = 'uniform_average', +) -> float: + + if y_pred.n_samples < 2: + raise ValueError( + 'R^2 score is not well-defined with less than two samples.', + ) + + def _r2_func(x: NDArrayFloat) -> NDArrayFloat: # noqa: WPS430 + ss_res = np.average( + (y_true(x) - y_pred(x)) * 2, + weights=sample_weight, + axis=0, + ) + + ss_tot = np.average( + ( + y_true(x) + - np.average(y_true(x), weights=sample_weight, axis=0) + ) ** 2, + weights=sample_weight, + axis=0, + ) + + # Divisions by zero allowed + with np.errstate(divide='ignore', invalid='ignore'): + score = 1 - ss_res / ss_tot + + # 0/0 case, the score is 1. + score[np.isnan(score)] = 1 + + # r/0 case, r!= 0. Return -inf outside this function + if np.any(np.isinf(score)): + raise _InfiniteScoreError + + # Score only had 1 input point + assert score.shape[0] == 1 + return score[0] # type: ignore [no-any-return] + + return _multioutput_score_basis(y_true, multioutput, _r2_func) diff --git a/skfda/ml/_neighbors_base.py b/skfda/ml/_neighbors_base.py index 3beb3b44c..f655ff7d8 100644 --- a/skfda/ml/_neighbors_base.py +++ b/skfda/ml/_neighbors_base.py @@ -662,132 +662,3 @@ def _functional_predict( ) return concatenate(iterable) # type: ignore[no-any-return] - - def score( - self, - X: Input, - y: TargetRegression, - sample_weight: NDArrayFloat \| None = None, - ) -> float: - r"""Return the coefficient of determination R^2 of the prediction. - - In the multivariate response case, the coefficient :math:`R^2` is - defined as - - .. math:: - 1 - \frac{\sum_{i=1}^{n} (y_i - \hat y_i)^2} - {\sum_{i=1}^{n} (y_i - \frac{1}{n}\sum_{i=1}^{n}y_i)^2} - - where :math:`\hat{y}_i` is the prediction associated to the test sample - :math:`X_i`, and :math:`{y}_i` is the true response. See - :func:`sklearn.metrics.r2_score <sklearn.metrics.r2_score>` for more - information. - - - In the functional case it is returned an extension of the coefficient - of determination :math:`R^2`, defined as - - .. math:: - 1 - \frac{\sum_{i=1}^{n}\int (y_i(t) - \hat{y}_i(t))^2dt} - {\sum_{i=1}^{n} \int (y_i(t)- \frac{1}{n}\sum_{i=1}^{n}y_i(t))^2dt} - - - The best possible score is 1.0 and it can be negative - (because the model can be arbitrarily worse). A constant model that - always predicts the expected value of y, disregarding the input - features, would get a R^2 score of 0.0. - - Args: - X: Test samples to be predicted. - y: True responses of the test samples. - sample_weight: Sample weights. - - Returns: - Coefficient of determination. - - """ - if self._functional: - return self._functional_score(X, y, sample_weight=sample_weight) - - # Default sklearn multivariate score - return super().score(X, y, sample_weight=sample_weight) - - def _functional_score( - self: NeighborsRegressorMixin[Input, TargetRegressionFData], - X: Input, - y: TargetRegressionFData, - sample_weight: NDArrayFloat \| None = None, - ) -> float: - r""" - Return an extension of the coefficient of determination R^2. - - The coefficient is defined as - - .. math:: - 1 - \frac{\sum_{i=1}^{n}\int (y_i(t) - \hat{y}_i(t))^2dt} - {\sum_{i=1}^{n} \int (y_i(t)- \frac{1}{n}\sum_{i=1}^{n}y_i(t))^2dt} - - where :math:`\hat{y}_i` is the prediction associated to the test sample - :math:`X_i`, and :math:`{y}_i` is the true response. - - The best possible score is 1.0 and it can be negative - (because the model can be arbitrarily worse). A constant model that - always predicts the expected value of y, disregarding the input - features, would get a R^2 score of 0.0. - - Args: - X: Test samples to be predicted. - y: True responses of the test samples. - sample_weight (array_like, shape = [n_samples], optional): Sample - weights. - - Returns: - Coefficient of determination. - - """ - # TODO: If it is created a module in ml.regression with other - # score metrics, move it. - from scipy.integrate import simps - - if y.dim_codomain != 1 or y.dim_domain != 1: - raise ValueError( - "Score not implemented for multivariate " - "functional data.", - ) - - # Make prediction - pred = self.predict(X) - - u = y - pred - v = y - y.mean() - - # Discretize to integrate and make squares if needed - if not isinstance(u, FDataGrid): - u = u.to_grid() - v = v.to_grid() - - data_u = u.data_matrix[..., 0] - data_v = v.data_matrix[..., 0] - - # Square without allocate more memory - np.square(data_u, out=data_u) - np.square(data_v, out=data_v) - - if sample_weight is not None: - if len(sample_weight) != len(y): - raise ValueError("Must be a weight for each sample.") - - normalized_sample_weight = sample_weight / sample_weight.sum() - data_u_t = data_u.T - data_u_t = normalized_sample_weight - data_v_t = data_v.T - data_v_t = normalized_sample_weight - - # Sum and integrate - sum_u = np.sum(data_u, axis=0) - sum_v = np.sum(data_v, axis=0) - - int_u = simps(sum_u, x=u.grid_points[0]) - int_v = simps(sum_v, x=v.grid_points[0]) - - return float(1 - int_u / int_v) diff --git a/skfda/representation/grid.py b/skfda/representation/grid.py index 4df8e6cfe..855923b1b 100644 --- a/skfda/representation/grid.py +++ b/skfda/representation/grid.py @@ -707,6 +707,11 @@ def _get_op_matrix( return other[other_index] + raise ValueError( + f"Invalid dimensions in operator between FDataGrid and Numpy " + f"array: {other.shape}" + ) + elif isinstance(other, FDataGrid): self._check_same_dimensions(other) return other.data_matrix diff --git a/skfda/representation/interpolation.py b/skfda/representation/interpolation.py index b546766d4..1c107564b 100644 --- a/skfda/representation/interpolation.py +++ b/skfda/representation/interpolation.py @@ -183,6 +183,8 @@ def _get_interpolator_1d( fdatagrid.data_matrix, k=self.interpolation_order, axis=1, + # Orders 0 and 1 behave well + check_finite=self.interpolation_order > 1, ) def _get_interpolator_nd(	Scorers for functional response Implement scoring methods for functional response (to be used for example in validation and hyperparameter search).		2022-03-04T01:06:11	[]	[]
GAA-UAM/scikit-fda	GAA-UAM__scikit-fda-408	49d6ee767950872b2c46c3602ceb3166f0a866dd	diff --git a/skfda/preprocessing/registration/_landmark_registration.py b/skfda/preprocessing/registration/_landmark_registration.py index e96b5296b..cc02eec6d 100644 --- a/skfda/preprocessing/registration/_landmark_registration.py +++ b/skfda/preprocessing/registration/_landmark_registration.py @@ -42,9 +42,8 @@ def landmark_shift_deltas( passed the location will be the result of the the call, the function should be accept as an unique parameter a numpy array with the list of landmarks. - By default it will be used as location :math:`\frac{1}{2}(max( - \text{landmarks})+ min(\text{landmarks}))` wich minimizes the - max shift. + By default it will be used as location the mean of the original + locations of the landmarks. Returns: Array containing the corresponding shifts. @@ -68,7 +67,7 @@ def landmark_shift_deltas( >>> shifts = landmark_shift_deltas(fd, landmarks) >>> shifts.round(3) - array([ 0.25 , -0.25 , -0.231]) + array([ 0.327, -0.173, -0.154]) The registered samples can be obtained with a shift @@ -88,10 +87,7 @@ def landmark_shift_deltas( # Parses location if location is None: - loc_array = ( - np.max(landmarks, axis=0) - + np.min(landmarks, axis=0) - ) / 2 + loc_array = np.mean(landmarks) elif callable(location): loc_array = location(landmarks) else:	Unify selection of target locations for landmark_shift_registration and landmark_elastic_registration In `landmark_shift_registration` the location of the target landmarks is by default selected as the average of the maximum and minimum values for the location in all curves. In `landmark_elastic_registration` those are instead selected as the means of the locations. We should use that latest method for both, for consistency.		2022-01-07T15:08:07	[]	[]
GAA-UAM/scikit-fda	GAA-UAM__scikit-fda-392	24a3541a01099467bfc27efe5752bfc1ca5f76d9	diff --git a/skfda/representation/_functional_data.py b/skfda/representation/_functional_data.py index 22ac35a8a..9f25ac2cd 100644 --- a/skfda/representation/_functional_data.py +++ b/skfda/representation/_functional_data.py @@ -33,6 +33,7 @@ GridPointsLike, LabelTuple, LabelTupleLike, + NDArrayInt, ) from .evaluator import Evaluator from .extrapolation import ExtrapolationLike, _parse_extrapolation @@ -1049,6 +1050,24 @@ def __array__(self, args: Any, kwargs: Any) -> np.ndarray: return array + def __array_ufunc__( + self, + ufunc: Any, + method: str, + inputs: Any, + *kwargs: Any, + ) -> Any: + """Prevent NumPy from converting to array just to do operations.""" + + # Make normal multiplication by scalar use the __mul__ method + if ufunc == np.multiply and method == "__call__" and len(inputs) == 2: + if isinstance(inputs[0], np.ndarray): + inputs = inputs[::-1] + + return inputs[0] inputs[1] + + return NotImplemented + ##################################################################### # Pandas ExtensionArray methods ##################################################################### @@ -1100,9 +1119,17 @@ def _from_factorized(cls, values: Any, original: Any) -> NoReturn: "Factorization does not make sense for functional data", ) + @abstractmethod + def _take_allow_fill( + self: T, + indices: NDArrayInt, + fill_value: T, + ) -> T: + pass + def take( self: T, - indices: Sequence[int], + indices: Union[int, Sequence[int], NDArrayInt], allow_fill: bool = False, fill_value: Optional[T] = None, axis: int = 0, @@ -1148,28 +1175,44 @@ def take( numpy.take pandas.api.extensions.take """ - from pandas.core.algorithms import take - # The axis parameter must exist, because sklearn tries to use take # instead of __getitem__ if axis != 0: raise ValueError(f"Axis must be 0, not {axis}") - # If the ExtensionArray is backed by an ndarray, then - # just pass that here instead of coercing to object. - data = np.asarray(self) - if allow_fill and fill_value is None: + arr_indices = np.atleast_1d(indices) + + if fill_value is None: fill_value = self.dtype.na_value - # fill value should always be translated from the scalar - # type for the array, to the physical storage type for - # the data, before passing to take. - result = take( - data, - indices, - fill_value=fill_value, - allow_fill=allow_fill, - ) - return self._from_sequence(result, dtype=self.dtype) + + non_empty_take_msg = "cannot do a non-empty take from an empty axes" + + if allow_fill: + if (arr_indices < -1).any(): + raise ValueError("Invalid indexes") + + positive_mask = arr_indices >= 0 + if len(self) == 0 and positive_mask.any(): + raise IndexError(non_empty_take_msg) + + sample_names = np.zeros(len(arr_indices), dtype=object) + result = self._take_allow_fill(arr_indices, fill_value) + + sample_names[positive_mask] = np.array(self.sample_names)[ + arr_indices[positive_mask] + ] + + if fill_value is not self.dtype.na_value: + sample_names[~positive_mask] = fill_value.sample_names[0] + + result.sample_names = tuple(sample_names) + else: + if len(self) == 0 and len(arr_indices) != 0: + raise IndexError(non_empty_take_msg) + + result = self[arr_indices] + + return result @classmethod def _concat_same_type( @@ -1198,6 +1241,7 @@ def astype(self, dtype: Any, copy: bool = True) -> Any: if copy: new_obj = self.copy() return new_obj + return super().astype(dtype) def _reduce(self, name: str, skipna: bool = True, **kwargs: Any) -> Any: diff --git a/skfda/representation/basis/_fdatabasis.py b/skfda/representation/basis/_fdatabasis.py index a9071e7f4..fa28d94ea 100644 --- a/skfda/representation/basis/_fdatabasis.py +++ b/skfda/representation/basis/_fdatabasis.py @@ -23,7 +23,13 @@ from ..._utils import _check_array_key, _int_to_real, constants from .. import grid from .._functional_data import FData -from .._typing import ArrayLike, DomainRange, GridPointsLike, LabelTupleLike +from .._typing import ( + ArrayLike, + DomainRange, + GridPointsLike, + LabelTupleLike, + NDArrayInt, +) from ..extrapolation import ExtrapolationLike from . import Basis @@ -866,6 +872,27 @@ def __rtruediv__(self: T, other: Union[np.ndarray, float]) -> T: ##################################################################### # Pandas ExtensionArray methods ##################################################################### + def _take_allow_fill( + self: T, + indices: NDArrayInt, + fill_value: T, + ) -> T: + result = self.copy() + result.coefficients = np.full( + (len(indices),) + self.coefficients.shape[1:], + np.nan, + ) + + positive_mask = indices >= 0 + result.coefficients[positive_mask] = self.coefficients[ + indices[positive_mask] + ] + + if fill_value is not self.dtype.na_value: + result.coefficients[~positive_mask] = fill_value.coefficients[0] + + return result + @property def dtype(self) -> FDataBasisDType: """The dtype for this extension array, FDataGridDType""" diff --git a/skfda/representation/grid.py b/skfda/representation/grid.py index 9f3fb1bef..2a1ef1092 100644 --- a/skfda/representation/grid.py +++ b/skfda/representation/grid.py @@ -25,9 +25,8 @@ import findiff import numpy as np import pandas.api.extensions -from matplotlib.figure import Figure - import scipy.stats.mstats +from matplotlib.figure import Figure from .._utils import ( _check_array_key, @@ -44,6 +43,7 @@ GridPoints, GridPointsLike, LabelTupleLike, + NDArrayInt, ) from .basis import Basis from .evaluator import Evaluator @@ -1260,7 +1260,7 @@ def __array_ufunc__( new_inputs = [ i.data_matrix if isinstance(i, FDataGrid) - else i for i in inputs + else self._get_op_matrix(i) for i in inputs ] outputs = kwargs.pop('out', None) @@ -1292,6 +1292,28 @@ def __array_ufunc__( ##################################################################### # Pandas ExtensionArray methods ##################################################################### + + def _take_allow_fill( + self: T, + indices: NDArrayInt, + fill_value: T, + ) -> T: + result = self.copy() + result.data_matrix = np.full( + (len(indices),) + self.data_matrix.shape[1:], + np.nan, + ) + + positive_mask = indices >= 0 + result.data_matrix[positive_mask] = self.data_matrix[ + indices[positive_mask] + ] + + if fill_value is not self.dtype.na_value: + result.data_matrix[~positive_mask] = fill_value.data_matrix[0] + + return result + @property def dtype(self) -> FDataGridDType: """The dtype for this extension array, FDataGridDType"""	Check conmutativity of operations between FDataGrid and numpy arrays Operations between FDataGrid objects and numpy arrays should not depend on the order. Currently if the array is the first term, the ufunc is dispatched to the FDataGrid `__array_ufunc__` and there are errors.		2021-11-11T00:19:50	[]	[]
GAA-UAM/scikit-fda	GAA-UAM__scikit-fda-387	bd907fdb9cbf0401d2b46ddf4106d0ae73b188dd	diff --git a/skfda/preprocessing/dim_reduction/feature_extraction/__init__.py b/skfda/preprocessing/dim_reduction/feature_extraction/__init__.py index 16355e236..132cc8f96 100644 --- a/skfda/preprocessing/dim_reduction/feature_extraction/__init__.py +++ b/skfda/preprocessing/dim_reduction/feature_extraction/__init__.py @@ -1,3 +1,4 @@ """Feature extraction.""" from ._ddg_transformer import DDGTransformer +from ._fda_feature_union import FDAFeatureUnion from ._fpca import FPCA diff --git a/skfda/preprocessing/dim_reduction/feature_extraction/_fda_feature_union.py b/skfda/preprocessing/dim_reduction/feature_extraction/_fda_feature_union.py new file mode 100644 index 000000000..e772be898 --- /dev/null +++ b/skfda/preprocessing/dim_reduction/feature_extraction/_fda_feature_union.py @@ -0,0 +1,124 @@ +"""Feature extraction union for dimensionality reduction.""" +from __future__ import annotations + +from typing import Union + +from numpy import ndarray +from pandas import DataFrame +from sklearn.pipeline import FeatureUnion + +from ....representation import FData + + +class FDAFeatureUnion(FeatureUnion): # type: ignore + """Concatenates results of multiple functional transformer objects. + + This estimator applies a list of transformer objects in parallel to the + input data, then concatenates the results (They can be either FDataGrid + and FDataBasis objects or multivariate data itself).This is useful to + combine several feature extraction mechanisms into a single transformer. + Parameters of the transformers may be set using its name and the parameter + name separated by a '__'. A transformer may be replaced entirely by + setting the parameter with its name to another transformer, + or removed by setting to 'drop'. + + Parameters: + transformer_list: list of tuple + List of tuple containing `(str, transformer)`. The first element + of the tuple is name affected to the transformer while the + second element is a scikit-learn transformer instance. + The transformer instance can also be `"drop"` for it to be + ignored. + n_jobs: int + Number of jobs to run in parallel. + ``None`` means 1 unless in a :obj:`joblib.parallel_backend` + context. + ``-1`` means using all processors. + The default value is None + transformer_weights: dict + Multiplicative weights for features per transformer. + Keys are transformer names, values the weights. + Raises ValueError if key not present in ``transformer_list``. + verbose: bool + If True, the time elapsed while fitting each transformer will be + printed as it is completed. By default the value is False + array_output: bool + indicates if the transformed data is requested to be a NumPy array + output. By default the value is False. + + Examples: + Firstly we will import the Berkeley Growth Study data set + >>> from skfda.datasets import fetch_growth + >>> X,y = fetch_growth(return_X_y=True) + + Then we need to import the transformers we want to use. In our case we + will use Generalized depth-versus-depth transformer. + Evaluation Transformer returns the original curve, and as it is helpful, + we will concatenate it to the already metioned transformer. + >>> from skfda.preprocessing.dim_reduction.feature_extraction import ( + ... FDAFeatureUnion, + ... ) + >>> from skfda.preprocessing.dim_reduction.feature_extraction import ( + ... DDGTransformer, + ... ) + >>> from skfda.exploratory.depth import ModifiedBandDepth + >>> from skfda.representation import EvaluationTransformer + >>> import numpy as np + + Finally we apply fit and transform. + >>> union = FDAFeatureUnion( + ... [ + ... ( + ... 'ddgtransformer', + ... DDGTransformer(depth_method=[ModifiedBandDepth()]), + ... ), + ... ("eval", EvaluationTransformer()), + ... ], + ... array_output=True, + ... ) + >>> np.around(union.fit_transform(X,y), decimals=2) + array([[ 2.100e-01, 9.000e-02, 8.130e+01, ..., 1.938e+02, 1.943e+02, + 1.951e+02], + [ 4.600e-01, 3.800e-01, 7.620e+01, ..., 1.761e+02, 1.774e+02, + 1.787e+02], + [ 2.000e-01, 3.300e-01, 7.680e+01, ..., 1.709e+02, 1.712e+02, + 1.715e+02], + ..., + [ 3.900e-01, 5.100e-01, 6.860e+01, ..., 1.660e+02, 1.663e+02, + 1.668e+02], + [ 2.600e-01, 2.700e-01, 7.990e+01, ..., 1.683e+02, 1.684e+02, + 1.686e+02], + [ 3.300e-01, 3.200e-01, 7.610e+01, ..., 1.686e+02, 1.689e+02, + 1.692e+02]]) + """ + + def __init__( + self, + transformer_list: list, # type: ignore + *, + n_jobs: int = 1, + transformer_weights: dict = None, # type: ignore + verbose: bool = False, + array_output: bool = False, + ) -> None: + self.array_output = array_output + super().__init__( + transformer_list, + n_jobs=n_jobs, + transformer_weights=transformer_weights, + verbose=verbose, + ) + + def _hstack(self, Xs: ndarray) -> Union[DataFrame, ndarray]: + + if self.array_output: + for i in Xs: + if isinstance(i, FData): + raise TypeError( + "There are transformed instances of FDataGrid or " + "FDataBasis that can't be concatenated on a NumPy " + "array.", + ) + return super()._hstack(Xs) + + return DataFrame({'Transformed data': Xs})	Join (possibly functional) features The existing scikit-learn class [`FeatureUnion`](https://scikit-learn.org/stable/modules/generated/sklearn.pipeline.FeatureUnion.html) is already able to apply transformations and join features that are multivariate in a new NumPy array. However, features that are functions cannot be in a NumPy array. A possible solution is to subclass `FeatureUnion` and join everything in a Pandas `DataFrame` (which accepts `FDataGrid` and `FDataBasis` as columns). We should also have a flag parameter that optionally allow the user to request NumPy array output (and an exception if that is not possible). Thanks to https://github.com/scikit-learn/scikit-learn/pull/17868 it seems that we will only need to override the `_hstack` method in the subclass to join the results.		2021-11-03T18:48:13	[]	[]
GAA-UAM/scikit-fda	GAA-UAM__scikit-fda-364	7c6f33db95f94e67def9424ee745dbd1326e256a	diff --git a/skfda/misc/covariances.py b/skfda/misc/covariances.py index 7acda93fa..5358a57fa 100644 --- a/skfda/misc/covariances.py +++ b/skfda/misc/covariances.py @@ -9,7 +9,6 @@ from matplotlib.figure import Figure from scipy.special import gamma, kv -from ..exploratory.visualization._utils import _create_figure, _figure_to_svg from ..representation._typing import ArrayLike, NDArrayFloat @@ -74,6 +73,8 @@ def __call__(self, x: ArrayLike, y: ArrayLike) -> NDArrayFloat: def heatmap(self, limits: Tuple[float, float] = (-1, 1)) -> Figure: """Return a heatmap plot of the covariance function.""" + from ..exploratory.visualization._utils import _create_figure + x = np.linspace(limits, 1000) cov_matrix = self(x, x) @@ -130,6 +131,8 @@ def _repr_latex_(self) -> str: return fr"\(\displaystyle {self._latex_content()}\)" def _repr_html_(self) -> str: + from ..exploratory.visualization._utils import _figure_to_svg + fig = self.heatmap() heatmap = _figure_to_svg(fig) plt.close(fig) diff --git a/skfda/misc/metrics/_fisher_rao.py b/skfda/misc/metrics/_fisher_rao.py index f63bb80f8..469ed0479 100644 --- a/skfda/misc/metrics/_fisher_rao.py +++ b/skfda/misc/metrics/_fisher_rao.py @@ -7,7 +7,6 @@ from typing_extensions import Final from ..._utils import normalize_scale, normalize_warping -from ...preprocessing.registration import FisherRaoElasticRegistration from ...representation import FData, FDataGrid from ...representation._typing import NDArrayFloat from ..operators import SRSF @@ -191,6 +190,8 @@ def fisher_rao_amplitude_distance( .. footbibliography:: """ + from ...preprocessing.registration import FisherRaoElasticRegistration + fdata1, fdata2 = _cast_to_grid( fdata1, fdata2, @@ -285,6 +286,8 @@ def fisher_rao_phase_distance( .. footbibliography:: """ + from ...preprocessing.registration import FisherRaoElasticRegistration + fdata1, fdata2 = _cast_to_grid( fdata1, fdata2, diff --git a/skfda/preprocessing/dim_reduction/feature_extraction/_fpca.py b/skfda/preprocessing/dim_reduction/feature_extraction/_fpca.py index f551fc0e0..6e2f7cabe 100644 --- a/skfda/preprocessing/dim_reduction/feature_extraction/_fpca.py +++ b/skfda/preprocessing/dim_reduction/feature_extraction/_fpca.py @@ -5,11 +5,13 @@ from typing import Callable, Optional, TypeVar, Union import numpy as np +import scipy.integrate +from scipy.linalg import solve_triangular from sklearn.base import BaseEstimator, TransformerMixin from sklearn.decomposition import PCA -from scipy.linalg import solve_triangular - +from ....misc import inner_product_matrix +from ....misc.metrics import l2_norm from ....misc.regularization import ( TikhonovRegularization, compute_penalty_matrix, @@ -151,6 +153,8 @@ def _fit_basis( else X.basis.n_basis ) n_samples = X.n_samples + # necessary in inverse_transform + self.n_samples_ = X.n_samples # check that the number of components is smaller than the sample size if self.n_components > X.n_samples: @@ -326,6 +330,9 @@ def _fit_grid( # get the number of samples and the number of points of descretization n_samples, n_points_discretization = fd_data.shape + # necessary for inverse_transform + self.n_samples_ = n_samples + # if centering is True then subtract the mean function to each function # in FDataBasis X = self._center_if_necessary(X) @@ -336,9 +343,8 @@ def _fit_grid( # in trapezoidal rule, suppose \deltax_k = x_k - x_{k-1}, the # weight vector is as follows: [\deltax_1/2, \deltax_1/2 + # \deltax_2/2, \deltax_2/2 + \deltax_3/2, ... , \deltax_n/2] - differences = np.diff(X.grid_points[0]) - differences = np.concatenate(((0,), differences, (0,))) - self.weights = (differences[:-1] + differences[1:]) / 2 + identity = np.eye(len(X.grid_points[0])) + self.weights = scipy.integrate.simps(identity, X.grid_points[0]) elif callable(self.weights): self.weights = self.weights(X.grid_points[0]) # if its a FDataGrid then we need to reduce the dimension to 1-D @@ -382,6 +388,7 @@ def _fit_grid( ), sample_names=(None,) self.n_components, ) + self.explained_variance_ratio_ = pca.explained_variance_ratio_ self.explained_variance_ = pca.explained_variance_ @@ -408,6 +415,7 @@ def _transform_grid( return ( X.data_matrix.reshape(X.data_matrix.shape[:-1]) + * self.weights @ np.transpose( self.components_.data_matrix.reshape( self.components_.data_matrix.shape[:-1], @@ -480,3 +488,61 @@ def fit_transform( """ return self.fit(X, y).transform(X, y) + + def inverse_transform( + self, + pc_scores: np.ndarray, + ) -> FData: + """ + Compute the recovery from the fitted principal components scores. + + In other words, + it maps ``pc_scores``, from the fitted functional PCs' space, + back to the input functional space. + ``pc_scores`` might be an array returned by ``transform`` method. + + Args: + pc_scores: ndarray (n_samples, n_components). + + Returns: + A FData object. + + """ + # check the instance is fitted. + + # input format check: + if isinstance(pc_scores, np.ndarray): + if pc_scores.ndim == 1: + pc_scores = pc_scores[np.newaxis, :] + + if pc_scores.shape[1] != self.n_components: + raise AttributeError( + "pc_scores must be a numpy array " + "with n_samples rows and n_components columns.", + ) + else: + raise AttributeError("pc_scores is not a numpy array.") + + # inverse_transform is slightly different whether + # .fit was applied to FDataGrid or FDataBasis object + # Does not work (boundary problem in x_hat and bias reconstruction) + if isinstance(self.components_, FDataGrid): + + additional_args = { + "data_matrix": np.einsum( + 'nc,c...->n...', + pc_scores, + self.components_.data_matrix, + ), + } + + elif isinstance(self.components_, FDataBasis): + + additional_args = { + "coefficients": pc_scores @ self.components_.coefficients, + } + + return self.mean_.copy( + *additional_args, + sample_names=(None,) len(pc_scores), + ) + self.mean_	Implement inverse_transform for FPCA FPCA should have an `inverse_transform` method to convert the data in the space generated by the eigenfunctions back into the original space.		2021-09-10T16:25:47	[]	[]
monarch-initiative/embiggen	monarch-initiative__embiggen-301	b743f7ff87ef2f6aed5db9ba5c6c43ee2e32bb8d	diff --git a/.gitignore b/.gitignore index b8107f2c..e01ebf96 100644 --- a/.gitignore +++ b/.gitignore @@ -37,4 +37,8 @@ embeddings .webm model.json -model.pkl.gz \ No newline at end of file +model.pkl.gz +okapi_tfidf_weighted_textual_embedding +precomputed_word_embedding +embedding +.ipynb \ No newline at end of file diff --git a/dist/embiggen-0.11.41.tar.gz b/dist/embiggen-0.11.41.tar.gz new file mode 100644 index 00000000..df4f3b83 Binary files /dev/null and b/dist/embiggen-0.11.41.tar.gz differ diff --git a/dist/embiggen-0.11.42.tar.gz b/dist/embiggen-0.11.42.tar.gz new file mode 100644 index 00000000..4fdbe9a0 Binary files /dev/null and b/dist/embiggen-0.11.42.tar.gz differ diff --git a/dist/embiggen-0.11.43.tar.gz b/dist/embiggen-0.11.43.tar.gz new file mode 100644 index 00000000..cecae322 Binary files /dev/null and b/dist/embiggen-0.11.43.tar.gz differ diff --git a/dist/embiggen-0.11.44.tar.gz b/dist/embiggen-0.11.44.tar.gz new file mode 100644 index 00000000..2df92419 Binary files /dev/null and b/dist/embiggen-0.11.44.tar.gz differ diff --git a/embiggen/__version__.py b/embiggen/__version__.py index e8f9ef5b..436504c8 100644 --- a/embiggen/__version__.py +++ b/embiggen/__version__.py @@ -1,2 +1,2 @@ """Current version of package Embiggen.""" -__version__ = "0.11.41" \ No newline at end of file +__version__ = "0.11.45" \ No newline at end of file diff --git a/embiggen/edge_prediction/edge_prediction_model.py b/embiggen/edge_prediction/edge_prediction_model.py index 4426c63c..26be6ecd 100644 --- a/embiggen/edge_prediction/edge_prediction_model.py +++ b/embiggen/edge_prediction/edge_prediction_model.py @@ -362,13 +362,16 @@ def predict( raise NotImplementedError( "Currently edge features are not supported in edge prediction models." ) - - predictions = super().predict( - graph, - support=support, - node_features=node_features, - node_type_features=node_type_features - ).flatten() + + if graph.has_edges(): + predictions = super().predict( + graph, + support=support, + node_features=node_features, + node_type_features=node_type_features + ).flatten() + else: + predictions = np.array([]) if return_predictions_dataframe: predictions = pd.DataFrame( diff --git a/embiggen/edge_prediction/edge_prediction_sklearn/sklearn_edge_prediction_adapter.py b/embiggen/edge_prediction/edge_prediction_sklearn/sklearn_edge_prediction_adapter.py index 0bf07d27..eba5a201 100644 --- a/embiggen/edge_prediction/edge_prediction_sklearn/sklearn_edge_prediction_adapter.py +++ b/embiggen/edge_prediction/edge_prediction_sklearn/sklearn_edge_prediction_adapter.py @@ -275,6 +275,9 @@ def _predict( edge_features: Optional[List[np.ndarray]] = None The edge features to use. """ + if not graph.has_edges(): + return np.array([]) + sequence = EdgePredictionSequence( graph=graph, graph_used_in_training=graph, @@ -283,6 +286,7 @@ def _predict( use_edge_metrics=False, batch_size=self._prediction_batch_size ) + return np.concatenate([ self._model_instance.predict(self._trasform_graph_into_edge_embedding( graph=edges[0], @@ -326,6 +330,9 @@ def _predict_proba( edge_features: Optional[List[np.ndarray]] = None The edge features to use. """ + if not graph.has_edges(): + return np.array([]) + sequence = EdgePredictionSequence( graph=graph, graph_used_in_training=graph, @@ -334,6 +341,7 @@ def _predict_proba( use_edge_metrics=False, batch_size=self._prediction_batch_size ) + prediction_probabilities = np.concatenate([ self._model_instance.predict_proba(self._trasform_graph_into_edge_embedding( graph=edges[0], diff --git a/embiggen/embedders/__init__.py b/embiggen/embedders/__init__.py index 5e9b2513..42354909 100644 --- a/embiggen/embedders/__init__.py +++ b/embiggen/embedders/__init__.py @@ -3,7 +3,10 @@ from embiggen.embedders.tensorflow_embedders import * from embiggen.embedders.pykeen_embedders import * from embiggen.embedders.non_existent_embedders import * +from embiggen.embedders.pytorch_geometric import * from embiggen.embedders.karateclub_embedders import * +from embiggen.embedders.pecanpy_embedders import * +from embiggen.embedders.fastnode2vec_embedders import * from embiggen.embedders.graph_embedding_pipeline import embed_graph # Export all non-internals. diff --git a/embiggen/embedders/ensmallen_embedders/deepwalk_cbow.py b/embiggen/embedders/ensmallen_embedders/deepwalk_cbow.py index 99d4c46b..0cd1fa86 100644 --- a/embiggen/embedders/ensmallen_embedders/deepwalk_cbow.py +++ b/embiggen/embedders/ensmallen_embedders/deepwalk_cbow.py @@ -27,7 +27,8 @@ def __init__( random_state: int = 42, dtype: str = "f32", ring_bell: bool = False, - enable_cache: bool = False + enable_cache: bool = False, + verbose: bool = True ): """Create new abstract DeepWalk method. @@ -84,6 +85,13 @@ def __init__( enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. + verbose: bool = True + Whether to display the loading bar. + This will only display the loading bar when + running the script in a bash-like environment. + It will not work in Jupyter Notebooks, there + it will appear in the notebook kernel in some + systems but not necessarily. """ super().__init__( embedding_size=embedding_size, @@ -105,7 +113,8 @@ def __init__( dtype=dtype, random_state=random_state, ring_bell=ring_bell, - enable_cache=enable_cache + enable_cache=enable_cache, + verbose=verbose, ) def parameters(self) -> Dict[str, Any]: diff --git a/embiggen/embedders/ensmallen_embedders/deepwalk_glove.py b/embiggen/embedders/ensmallen_embedders/deepwalk_glove.py index f55e6e4f..5b9861a1 100644 --- a/embiggen/embedders/ensmallen_embedders/deepwalk_glove.py +++ b/embiggen/embedders/ensmallen_embedders/deepwalk_glove.py @@ -21,7 +21,8 @@ def __init__( dtype: str = "f32", random_state: int = 42, ring_bell: bool = False, - enable_cache: bool = False + enable_cache: bool = False, + verbose: bool = True ): """Create new abstract DeepWalk method. @@ -69,6 +70,13 @@ def __init__( enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. + verbose: bool = True + Whether to display the loading bar. + This will only display the loading bar when + running the script in a bash-like environment. + It will not work in Jupyter Notebooks, there + it will appear in the notebook kernel in some + systems but not necessarily. """ super().__init__( embedding_size=embedding_size, @@ -86,7 +94,8 @@ def __init__( enable_cache=enable_cache, dtype=dtype, ring_bell=ring_bell, - random_state=random_state + random_state=random_state, + verbose=verbose, ) def parameters(self) -> Dict[str, Any]: diff --git a/embiggen/embedders/ensmallen_embedders/deepwalk_skipgram.py b/embiggen/embedders/ensmallen_embedders/deepwalk_skipgram.py index b00342ec..18b81d2e 100644 --- a/embiggen/embedders/ensmallen_embedders/deepwalk_skipgram.py +++ b/embiggen/embedders/ensmallen_embedders/deepwalk_skipgram.py @@ -27,7 +27,8 @@ def __init__( random_state: int = 42, dtype: str = "f32", ring_bell: bool = False, - enable_cache: bool = False + enable_cache: bool = False, + verbose: bool = True ): """Create new abstract DeepWalk method. @@ -84,6 +85,13 @@ def __init__( enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. + verbose: bool = True + Whether to display the loading bar. + This will only display the loading bar when + running the script in a bash-like environment. + It will not work in Jupyter Notebooks, there + it will appear in the notebook kernel in some + systems but not necessarily. """ super().__init__( embedding_size=embedding_size, @@ -105,7 +113,8 @@ def __init__( dtype=dtype, random_state=random_state, ring_bell=ring_bell, - enable_cache=enable_cache + enable_cache=enable_cache, + verbose=verbose, ) def parameters(self) -> Dict[str, Any]: diff --git a/embiggen/embedders/ensmallen_embedders/node2vec.py b/embiggen/embedders/ensmallen_embedders/node2vec.py index 05094f59..67054945 100644 --- a/embiggen/embedders/ensmallen_embedders/node2vec.py +++ b/embiggen/embedders/ensmallen_embedders/node2vec.py @@ -51,7 +51,11 @@ def __init__( """ model_name = must_be_in_set(self.model_name(), self.MODELS.keys(), "model name") self._model_kwargs = model_kwargs - self._model = Node2VecEnsmallen.MODELS[model_name](model_kwargs) + self._model = Node2VecEnsmallen.MODELS[model_name]( + embedding_size=embedding_size, + random_state=random_state, + model_kwargs + ) super().__init__( embedding_size=embedding_size, diff --git a/embiggen/embedders/ensmallen_embedders/node2vec_cbow.py b/embiggen/embedders/ensmallen_embedders/node2vec_cbow.py index 04961df9..78f956f4 100644 --- a/embiggen/embedders/ensmallen_embedders/node2vec_cbow.py +++ b/embiggen/embedders/ensmallen_embedders/node2vec_cbow.py @@ -29,7 +29,8 @@ def __init__( random_state: int = 42, dtype: str = "f32", ring_bell: bool = False, - enable_cache: bool = False + enable_cache: bool = False, + verbose: bool = True ): """Create new abstract Node2Vec method. @@ -100,6 +101,13 @@ def __init__( enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. + verbose: bool = True + Whether to display the loading bar. + This will only display the loading bar when + running the script in a bash-like environment. + It will not work in Jupyter Notebooks, there + it will appear in the notebook kernel in some + systems but not necessarily. """ super().__init__( embedding_size=embedding_size, @@ -123,7 +131,8 @@ def __init__( dtype=dtype, random_state=random_state, ring_bell=ring_bell, - enable_cache=enable_cache + enable_cache=enable_cache, + verbose=verbose, ) def parameters(self) -> Dict[str, Any]: diff --git a/embiggen/embedders/ensmallen_embedders/node2vec_glove.py b/embiggen/embedders/ensmallen_embedders/node2vec_glove.py index a80772b7..1fdb5919 100644 --- a/embiggen/embedders/ensmallen_embedders/node2vec_glove.py +++ b/embiggen/embedders/ensmallen_embedders/node2vec_glove.py @@ -23,7 +23,8 @@ def __init__( dtype: str = "f32", random_state: int = 42, ring_bell: bool = False, - enable_cache: bool = False + enable_cache: bool = False, + verbose: bool = True ): """Create new Node2Vec GloVe model. @@ -85,6 +86,13 @@ def __init__( enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. + verbose: bool = True + Whether to display the loading bar. + This will only display the loading bar when + running the script in a bash-like environment. + It will not work in Jupyter Notebooks, there + it will appear in the notebook kernel in some + systems but not necessarily. """ super().__init__( embedding_size=embedding_size, @@ -104,7 +112,8 @@ def __init__( dtype=dtype, enable_cache=enable_cache, ring_bell=ring_bell, - random_state=random_state + random_state=random_state, + verbose=verbose, ) def parameters(self) -> Dict[str, Any]: diff --git a/embiggen/embedders/ensmallen_embedders/node2vec_skipgram.py b/embiggen/embedders/ensmallen_embedders/node2vec_skipgram.py index 2dd871e6..4350dbcc 100644 --- a/embiggen/embedders/ensmallen_embedders/node2vec_skipgram.py +++ b/embiggen/embedders/ensmallen_embedders/node2vec_skipgram.py @@ -29,7 +29,8 @@ def __init__( random_state: int = 42, dtype: str = "f32", ring_bell: bool = False, - enable_cache: bool = False + enable_cache: bool = False, + verbose: bool = True ): """Create new abstract Node2Vec method. @@ -102,6 +103,13 @@ def __init__( enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. + verbose: bool = True + Whether to display the loading bar. + This will only display the loading bar when + running the script in a bash-like environment. + It will not work in Jupyter Notebooks, there + it will appear in the notebook kernel in some + systems but not necessarily. """ super().__init__( embedding_size=embedding_size, @@ -125,7 +133,8 @@ def __init__( dtype=dtype, random_state=random_state, ring_bell=ring_bell, - enable_cache=enable_cache + enable_cache=enable_cache, + verbose=verbose, ) def parameters(self) -> Dict[str, Any]: diff --git a/embiggen/embedders/fastnode2vec_embedders/__init__.py b/embiggen/embedders/fastnode2vec_embedders/__init__.py new file mode 100644 index 00000000..77f92926 --- /dev/null +++ b/embiggen/embedders/fastnode2vec_embedders/__init__.py @@ -0,0 +1,8 @@ +"""Module with node embedding models based on FastNode2Vec.""" +from embiggen.utils.abstract_models import build_init, AbstractEmbeddingModel + +build_init( + module_library_names=["numba", "fastnode2vec"], + formatted_library_name="FastNode2Vec", + expected_parent_class=AbstractEmbeddingModel +) \ No newline at end of file diff --git a/embiggen/embedders/fastnode2vec_embedders/node2vec.py b/embiggen/embedders/fastnode2vec_embedders/node2vec.py new file mode 100644 index 00000000..a5d13d5b --- /dev/null +++ b/embiggen/embedders/fastnode2vec_embedders/node2vec.py @@ -0,0 +1,219 @@ +"""Node2Vec wrapper for FastNode2Vec numba-based node embedding library.""" +from typing import Dict, Union, Any + +import numpy as np +import pandas as pd +from ensmallen import Graph + +from embiggen.utils.abstract_models import AbstractEmbeddingModel, EmbeddingResult +from fastnode2vec import Graph as FNGraph +from fastnode2vec import Node2Vec +from multiprocessing import cpu_count +from time import time + + +class Node2VecFastNode2Vec(AbstractEmbeddingModel): + """Node2Vec wrapper for FastNode2Vec numba-based node embedding library.""" + + def __init__( + self, + embedding_size: int = 100, + epochs: int = 30, + walk_length: int = 128, + iterations: int = 10, + window_size: int = 5, + learning_rate: float = 0.01, + return_weight: float = 0.25, + explore_weight: float = 4.0, + number_of_workers: Union[int, str] = "auto", + verbose: bool = False, + random_state: int = 42, + ring_bell: bool = False, + enable_cache: bool = False + ): + """Create new wrapper for Node2Vec model from FastNode2Vec library. + + Parameters + ------------------------- + embedding_size: int = 100 + The dimension of the embedding to compute. + epochs: int = 100 + The number of epochs to use to train the model for. + learning_rate: float = 0.01 + Learning rate of the model. + verbose: bool = False + Whether to show the loading bar. + random_state: int = 42 + Random seed to use while training the model + ring_bell: bool = False, + Whether to play a sound when embedding completes. + enable_cache: bool = False + Whether to enable the cache, that is to + store the computed embedding. + """ + self._walk_length = walk_length + self._iterations = iterations + self._window_size = window_size + self._return_weight = return_weight + self._explore_weight = explore_weight + self._epochs = epochs + self._verbose = verbose + self._learning_rate = learning_rate + self._time_required_by_last_embedding = None + if number_of_workers == "auto": + number_of_workers = cpu_count() + self._number_of_workers = number_of_workers + + super().__init__( + embedding_size=embedding_size, + enable_cache=enable_cache, + ring_bell=ring_bell, + random_state=random_state + ) + + @classmethod + def smoke_test_parameters(cls) -> Dict[str, Any]: + """Returns parameters for smoke test.""" + return dict( + super().smoke_test_parameters(), + window_size=1, + walk_length=2, + iterations=1 + ) + + def parameters(self) -> Dict[str, Any]: + return dict( + super().parameters(), + *dict( + epochs=self._epochs, + walk_length=self._walk_length, + window_size=self._window_size, + iterations=self._iterations, + return_weight=self._return_weight, + explore_weight=self._explore_weight, + ) + ) + + @classmethod + def model_name(cls) -> str: + """Returns name of the model""" + return "Node2Vec" + + @classmethod + def library_name(cls) -> str: + return "FastNode2Vec" + + @classmethod + def task_name(cls) -> str: + return "Node Embedding" + + def get_time_required_by_last_embedding(self) -> float: + """Returns the time required by last embedding.""" + if self._time_required_by_last_embedding is None: + raise ValueError("You have not yet run an embedding.") + return self._time_required_by_last_embedding + + def _fit_transform( + self, + graph: Graph, + return_dataframe: bool = True, + ) -> Union[np.ndarray, pd.DataFrame, Dict[str, np.ndarray], Dict[str, pd.DataFrame]]: + """Return node embedding""" + + if graph.has_edge_weights(): + edges_iterator = ( + ( + graph.get_node_names_from_edge_id(edge_id), + graph.get_edge_weight_from_edge_id(edge_id) + ) + for edge_id in range(graph.get_number_of_directed_edges()) + ) + else: + edges_iterator = ( + graph.get_node_names_from_edge_id(edge_id) + for edge_id in range(graph.get_number_of_directed_edges()) + ) + + fn_graph: FNGraph = FNGraph( + edges_iterator, + directed=True, + weighted=graph.has_edge_weights(), + n_edges=graph.get_number_of_directed_edges() + ) + + start = time() + + model: Node2Vec = Node2Vec( + graph=fn_graph, + dim=self._embedding_size, + walk_length=self._walk_length, + context=self._window_size, + p=1.0/self._return_weight, + q=1.0/self._explore_weight, + workers=self._number_of_workers, + batch_walks=self._iterations, + seed=self._random_state, + ) + + model.train( + epochs=self._epochs * self._iterations, + progress_bar=self._verbose + ) + + self._time_required_by_last_embedding = time() - start + + # This library does not provide node embedding + # for disconnected nodes, so we need to patch it up + # if the graph contains disconnected nodes. + if graph.has_singleton_nodes(): + node_embedding = np.zeros( + shape=(graph.get_number_of_nodes(), self._embedding_size), + ) + for node_id in range(graph.get_number_of_nodes()): + node_name = graph.get_node_name_from_node_id(node_id) + if node_name in model.wv: + node_embedding[node_id] = model.wv[node_name] + else: + node_embedding = model.wv[graph.get_node_names()] + + if return_dataframe: + node_embedding = pd.DataFrame( + node_embedding, + index=graph.get_node_names() + ) + + return EmbeddingResult( + embedding_method_name=self.model_name(), + node_embeddings=node_embedding + ) + + @classmethod + def requires_nodes_sorted_by_decreasing_node_degree(cls) -> bool: + return False + + @classmethod + def is_topological(cls) -> bool: + return True + + @classmethod + def can_use_edge_weights(cls) -> bool: + return True + + @classmethod + def requires_edge_weights(cls) -> bool: + return False + + @classmethod + def can_use_node_types(cls) -> bool: + """Returns whether the model can use node types.""" + return False + + @classmethod + def can_use_edge_types(cls) -> bool: + """Returns whether the model can use edge types.""" + return False + + @classmethod + def is_stocastic(cls) -> bool: + """Returns whether the model is stocastic and has therefore a random state.""" + return True diff --git a/embiggen/embedders/karateclub_embedders/abstract_karateclub_embedder.py b/embiggen/embedders/karateclub_embedders/abstract_karateclub_embedder.py index 2da4030b..1e503d43 100644 --- a/embiggen/embedders/karateclub_embedders/abstract_karateclub_embedder.py +++ b/embiggen/embedders/karateclub_embedders/abstract_karateclub_embedder.py @@ -7,6 +7,7 @@ from embiggen.utils.abstract_models import AbstractEmbeddingModel, EmbeddingResult, abstract_class from embiggen.utils.networkx_utils import convert_ensmallen_graph_to_networkx_graph + @abstract_class class AbstractKarateClubEmbedder(AbstractEmbeddingModel): @@ -28,7 +29,7 @@ def task_name(cls) -> str: def _build_model(self) -> Type[Estimator]: """Returnd the built estimator.""" raise NotImplementedError( - f"In the child class {self.__class__.__name__} of {super().__name__.__name__} " + f"In the child class {self.__class__.__name__} of {super().__class__.__name__} " f"implementing the model {self.model_name()} we could not find the method " "called `_build_model`. Please do implement it." ) @@ -61,7 +62,8 @@ def _fit_transform( "It is not clear what to do with this object." ) - model.fit(convert_ensmallen_graph_to_networkx_graph(graph)) + graph_nx = convert_ensmallen_graph_to_networkx_graph(graph) + model.fit(graph_nx) node_embeddings: np.ndarray = model.get_embedding() diff --git a/embiggen/embedders/karateclub_embedders/deep_walk.py b/embiggen/embedders/karateclub_embedders/deep_walk.py index 17d14ba3..e6023ba1 100644 --- a/embiggen/embedders/karateclub_embedders/deep_walk.py +++ b/embiggen/embedders/karateclub_embedders/deep_walk.py @@ -10,10 +10,10 @@ class DeepWalkSkipGramKarateClub(AbstractKarateClubEmbedder): def __init__( self, embedding_size: int = 100, - walk_number: int = 10, - walk_length: int = 80, + iterations: int = 10, + walk_length: int = 128, window_size: int = 5, - epochs: int = 10, + epochs: int = 30, learning_rate: float = 0.05, min_count: int = 1, random_state: int = 42, @@ -26,13 +26,13 @@ def __init__( ---------------------- embedding_size: int = 100 Size of the embedding to use. - walk_number: int = 10 + iterations: int = 10 Number of random walks. Default is 10. - walk_length: int = 80 + walk_length: int = 128 Length of random walks. Default is 80. window_size: int = 5 Matrix power order. Default is 5. - epochs: int = 10 + epochs: int = 30 Number of epochs. Default is 1. learning_rate: float = 0.05 HogWild! learning rate. Default is 0.05. @@ -47,7 +47,7 @@ def __init__( Whether to enable the cache, that is to store the computed embedding. """ - self._walk_number=walk_number + self._iterations=iterations self._walk_length=walk_length self._workers=cpu_count() self._window_size=window_size @@ -65,7 +65,7 @@ def parameters(self) -> Dict[str, Any]: """Returns the parameters used in the model.""" return dict( super().parameters(), - walk_number=self._walk_number, + iterations=self._iterations, walk_length=self._walk_length, window_size=self._window_size, epochs=self._epochs, @@ -78,7 +78,7 @@ def smoke_test_parameters(cls) -> Dict[str, Any]: """Returns parameters for smoke test.""" return dict( AbstractKarateClubEmbedder.smoke_test_parameters(), - walk_number=1, + iterations=1, walk_length=8, window_size=1, epochs=1, @@ -87,7 +87,7 @@ def smoke_test_parameters(cls) -> Dict[str, Any]: def _build_model(self) -> DeepWalk: """Return new instance of the DeepWalk model.""" return DeepWalk( - walk_number=self._walk_number, + walk_number=self._iterations, walk_length=self._walk_length, dimensions=self._embedding_size, workers=self._workers, diff --git a/embiggen/embedders/karateclub_embedders/role2vec.py b/embiggen/embedders/karateclub_embedders/role2vec.py index 1a925f4a..901f81c1 100644 --- a/embiggen/embedders/karateclub_embedders/role2vec.py +++ b/embiggen/embedders/karateclub_embedders/role2vec.py @@ -10,8 +10,8 @@ class Role2VecKarateClub(AbstractKarateClubEmbedder): def __init__( self, embedding_size: int = 100, - walk_number: int = 10, - walk_length: int = 80, + iterations: int = 10, + walk_length: int = 128, window_size: int = 5, epochs: int = 10, learning_rate: float = 0.05, @@ -29,9 +29,9 @@ def __init__( ---------------------- embedding_size: int = 100 Size of the embedding to use. - walk_number: int = 10 + iterations: int = 10 Number of random walks. Default is 10. - walk_length: int = 80 + walk_length: int = 128 Length of random walks. Default is 80. window_size: int = 5 Matrix power order. Default is 5. @@ -56,7 +56,7 @@ def __init__( Whether to enable the cache, that is to store the computed embedding. """ - self._walk_number = walk_number + self._iterations = iterations self._walk_length = walk_length self._workers = cpu_count() self._window_size = window_size @@ -77,7 +77,7 @@ def parameters(self) -> Dict[str, Any]: """Returns the parameters used in the model.""" return dict( super().parameters(), - walk_number=self._walk_number, + iterations=self._iterations, walk_length=self._walk_length, window_size=self._window_size, epochs=self._epochs, @@ -93,7 +93,7 @@ def smoke_test_parameters(cls) -> Dict[str, Any]: """Returns parameters for smoke test.""" return dict( AbstractKarateClubEmbedder.smoke_test_parameters(), - walk_number=1, + iterations=1, weisfeiler_lehman_hashing_iterations=1, walk_length=8, window_size=2, @@ -103,7 +103,7 @@ def smoke_test_parameters(cls) -> Dict[str, Any]: def _build_model(self) -> Role2Vec: """Return new instance of the Role2Vec model.""" return Role2Vec( - walk_number=self._walk_number, + walk_number=self._iterations, walk_length=self._walk_length, dimensions=self._embedding_size, workers=self._workers, diff --git a/embiggen/embedders/karateclub_embedders/skipgram.py b/embiggen/embedders/karateclub_embedders/skipgram.py index 443aa78b..728d2381 100644 --- a/embiggen/embedders/karateclub_embedders/skipgram.py +++ b/embiggen/embedders/karateclub_embedders/skipgram.py @@ -10,8 +10,8 @@ class SkipGramKarateClub(AbstractKarateClubEmbedder): def __init__( self, embedding_size: int = 100, - walk_number: int = 10, - walk_length: int = 80, + iterations: int = 10, + walk_length: int = 128, window_size: int = 5, p: float = 1.0, q: float = 1.0, @@ -36,9 +36,9 @@ def __init__( ---------------------- embedding_size: int = 100 Size of the embedding to use. - walk_number: int = 10 + iterations: int = 10 Number of random walks. Default is 10. - walk_length: int = 80 + walk_length: int = 128 Length of random walks. Default is 80. window_size: int = 5 Matrix power order. Default is 5. @@ -61,7 +61,7 @@ def __init__( Whether to enable the cache, that is to store the computed embedding. """ - self._walk_number=walk_number + self._iterations=iterations self._walk_length=walk_length self._workers=cpu_count() self._window_size=window_size @@ -81,7 +81,7 @@ def parameters(self) -> Dict[str, Any]: """Returns the parameters used in the model.""" return dict( super().parameters(), - walk_number=self._walk_number, + iterations=self._iterations, walk_length=self._walk_length, window_size=self._window_size, p=self._p, @@ -96,7 +96,7 @@ def smoke_test_parameters(cls) -> Dict[str, Any]: """Returns parameters for smoke test.""" return dict( AbstractKarateClubEmbedder.smoke_test_parameters(), - walk_number=1, + iterations=1, walk_length=8, window_size=2, epochs=1, @@ -105,7 +105,7 @@ def smoke_test_parameters(cls) -> Dict[str, Any]: def _build_model(self) -> Node2Vec: """Return new instance of the Node2Vec CBOW model.""" return Node2Vec( - walk_number=self._walk_number, + walk_number=self._iterations, walk_length=self._walk_length, dimensions=self._embedding_size, workers=self._workers, diff --git a/embiggen/embedders/karateclub_embedders/walklets.py b/embiggen/embedders/karateclub_embedders/walklets.py index d56ba3ec..73848ed7 100644 --- a/embiggen/embedders/karateclub_embedders/walklets.py +++ b/embiggen/embedders/karateclub_embedders/walklets.py @@ -10,8 +10,8 @@ class WalkletsSkipGramKarateClub(AbstractKarateClubEmbedder): def __init__( self, embedding_size: int = 100, - walk_number: int = 10, - walk_length: int = 80, + iterations: int = 10, + walk_length: int = 128, window_size: int = 5, epochs: int = 10, learning_rate: float = 0.05, @@ -26,9 +26,9 @@ def __init__( ---------------------- embedding_size: int = 100 Size of the embedding to use. - walk_number: int = 10 + iterations: int = 10 Number of random walks. Default is 10. - walk_length: int = 80 + walk_length: int = 128 Length of random walks. Default is 80. window_size: int = 5 Matrix power order. Default is 5. @@ -47,7 +47,7 @@ def __init__( Whether to enable the cache, that is to store the computed embedding. """ - self._walk_number = walk_number + self._iterations = iterations self._walk_length = walk_length self._workers = cpu_count() self._window_size = window_size @@ -67,7 +67,7 @@ def parameters(self) -> Dict[str, Any]: parameters["embedding_size"] = parameters["embedding_size"] * self._window_size return dict( parameters, - walk_number=self._walk_number, + iterations=self._iterations, walk_length=self._walk_length, window_size=self._window_size, epochs=self._epochs, @@ -80,7 +80,7 @@ def smoke_test_parameters(cls) -> Dict[str, Any]: """Returns parameters for smoke test.""" return dict( AbstractKarateClubEmbedder.smoke_test_parameters(), - walk_number=1, + iterations=1, walk_length=8, window_size=1, epochs=1, @@ -89,7 +89,7 @@ def smoke_test_parameters(cls) -> Dict[str, Any]: def _build_model(self) -> Walklets: """Return new instance of the Walklets model.""" return Walklets( - walk_number=self._walk_number, + walk_number=self._iterations, walk_length=self._walk_length, dimensions=self._embedding_size, workers=self._workers, diff --git a/embiggen/embedders/pecanpy_embedders/__init__.py b/embiggen/embedders/pecanpy_embedders/__init__.py new file mode 100644 index 00000000..8370092a --- /dev/null +++ b/embiggen/embedders/pecanpy_embedders/__init__.py @@ -0,0 +1,8 @@ +"""Module with node embedding models based on PecanPy.""" +from embiggen.utils.abstract_models import build_init, AbstractEmbeddingModel + +build_init( + module_library_names=["numba", "pecanpy"], + formatted_library_name="PecanPy", + expected_parent_class=AbstractEmbeddingModel +) \ No newline at end of file diff --git a/embiggen/embedders/pecanpy_embedders/node2vec.py b/embiggen/embedders/pecanpy_embedders/node2vec.py new file mode 100644 index 00000000..e0c787ac --- /dev/null +++ b/embiggen/embedders/pecanpy_embedders/node2vec.py @@ -0,0 +1,216 @@ +"""Node2Vec wrapper for PecanPy numba-based node embedding library.""" +from typing import Dict, Union, Any + +import numpy as np +import pandas as pd +from ensmallen import Graph + +from embiggen.utils.abstract_models import AbstractEmbeddingModel, EmbeddingResult +from pecanpy.node2vec import SparseOTF +from multiprocessing import cpu_count +from time import time + +class Node2VecPecanPy(AbstractEmbeddingModel): + """Node2Vec wrapper for PecanPy numba-based node embedding library.""" + + def __init__( + self, + embedding_size: int = 100, + epochs: int = 30, + walk_length: int = 128, + iterations: int = 10, + window_size: int = 5, + batch_size: int = 128, + learning_rate: float = 0.01, + return_weight: float = 0.25, + explore_weight: float = 4.0, + number_of_workers: Union[int, str] = "auto", + verbose: bool = False, + random_state: int = 42, + ring_bell: bool = False, + enable_cache: bool = False + ): + """Create new wrapper for Node2Vec model from PecanPy library. + + Parameters + ------------------------- + embedding_size: int = 100 + The dimension of the embedding to compute. + epochs: int = 100 + The number of epochs to use to train the model for. + batch_size: int = 210 + Size of the training batch. + learning_rate: float = 0.01 + Learning rate of the model. + verbose: bool = False + Whether to show the loading bar. + random_state: int = 42 + Random seed to use while training the model + ring_bell: bool = False, + Whether to play a sound when embedding completes. + enable_cache: bool = False + Whether to enable the cache, that is to + store the computed embedding. + """ + self._walk_length = walk_length + self._iterations = iterations + self._window_size = window_size + self._return_weight = return_weight + self._explore_weight = explore_weight + self._epochs = epochs + self._verbose = verbose + self._batch_size = batch_size + self._learning_rate = learning_rate + self._time_required_by_last_embedding = None + if number_of_workers == "auto": + number_of_workers = cpu_count() + self._number_of_workers = number_of_workers + + super().__init__( + embedding_size=embedding_size, + enable_cache=enable_cache, + ring_bell=ring_bell, + random_state=random_state + ) + + @classmethod + def smoke_test_parameters(cls) -> Dict[str, Any]: + """Returns parameters for smoke test.""" + return dict( + super().smoke_test_parameters(), + window_size=1, + walk_length=2, + iterations=1 + ) + + def parameters(self) -> Dict[str, Any]: + return dict( + super().parameters(), + dict( + epochs=self._epochs, + batch_size=self._batch_size, + walk_length=self._walk_length, + window_size=self._window_size, + iterations=self._iterations, + return_weight=self._return_weight, + explore_weight=self._explore_weight, + ) + ) + + @classmethod + def model_name(cls) -> str: + """Returns name of the model""" + return "Node2Vec" + + @classmethod + def library_name(cls) -> str: + return "PecanPy" + + @classmethod + def task_name(cls) -> str: + return "Node Embedding" + + def get_time_required_by_last_embedding(self) -> float: + """Returns the time required by last embedding.""" + if self._time_required_by_last_embedding is None: + raise ValueError("You have not yet run an embedding.") + return self._time_required_by_last_embedding + + def _fit_transform( + self, + graph: Graph, + return_dataframe: bool = True, + ) -> Union[np.ndarray, pd.DataFrame, Dict[str, np.ndarray], Dict[str, pd.DataFrame]]: + """Return node embedding""" + + model: SparseOTF = SparseOTF( + p=1.0/self._return_weight, + q=1.0/self._explore_weight, + workers=self._number_of_workers, + verbose=self._verbose, + ) + + # Instead of using `SparseOTF` methods, + # we manipolate directly the object attributes + # in order to avoid a memory peak and making + # the conversion of the Ensmallen graph object + # into the `SparseOTF` more seamless. + + # The `indptr` attribute contains the indices + # of the rows in the CSR representation, which + # are the comulative node degrees. + # This is shifted by a value to the right, and the + # first value is set to zero. + model.indptr = np.zeros(graph.get_number_of_nodes() + 1, dtype=np.int64) + model.indptr[1:] = graph.get_cumulative_node_degrees().astype(np.int64) + + # In the `indices` attribute we need to store the destinations. + model.indices = graph.get_directed_destination_node_ids() + + # We also need to set the `IDlst` attribute, which are + # the node IDs. + model.IDlst = graph.get_node_ids() + + # In model data we need to store the edge weights + # if are present. If the graph is weighted, we use + # the graph edge weights. Otherwise we set all weights + # to one, as the library PecanPy does. + if graph.has_edge_weights(): + model.data = graph.get_directed_edge_weights().astype(np.float64) + else: + model.data = np.ones_like(model.indices, dtype=np.float64) + + start = time() + + node_embedding = model.embed( + dim=self._embedding_size, + num_walks=self._iterations, + walk_length=self._walk_length, + window_size=self._window_size, + epochs=self._epochs, + verbose=self._verbose + ) + + self._time_required_by_last_embedding = time() - start + + if return_dataframe: + node_embedding = pd.DataFrame( + node_embedding, + index=graph.get_node_names() + ) + + return EmbeddingResult( + embedding_method_name=self.model_name(), + node_embeddings=node_embedding + ) + + @classmethod + def requires_nodes_sorted_by_decreasing_node_degree(cls) -> bool: + return False + + @classmethod + def is_topological(cls) -> bool: + return True + + @classmethod + def can_use_edge_weights(cls) -> bool: + return True + + @classmethod + def requires_edge_weights(cls) -> bool: + return False + + @classmethod + def can_use_node_types(cls) -> bool: + """Returns whether the model can use node types.""" + return False + + @classmethod + def can_use_edge_types(cls) -> bool: + """Returns whether the model can use edge types.""" + return False + + @classmethod + def is_stocastic(cls) -> bool: + """Returns whether the model is stocastic and has therefore a random state.""" + return True \ No newline at end of file diff --git a/embiggen/embedders/pykeen_embedders/__init__.py b/embiggen/embedders/pykeen_embedders/__init__.py index affc9d1d..bc15ab46 100644 --- a/embiggen/embedders/pykeen_embedders/__init__.py +++ b/embiggen/embedders/pykeen_embedders/__init__.py @@ -1,4 +1,4 @@ -"""Module with graph embedding models based on PyKEEN.""" +"""Module with node embedding models based on PyKEEN.""" from embiggen.utils.abstract_models import build_init, AbstractEmbeddingModel build_init( diff --git a/embiggen/embedders/pykeen_embedders/pykeen_embedder.py b/embiggen/embedders/pykeen_embedders/pykeen_embedder.py index d0f21d5d..ecf1f198 100644 --- a/embiggen/embedders/pykeen_embedders/pykeen_embedder.py +++ b/embiggen/embedders/pykeen_embedders/pykeen_embedder.py @@ -134,7 +134,7 @@ def _build_model(self, triples_factory: CoreTriplesFactory) -> Type[Model]: The PyKEEN triples factory to use to create the model. """ raise NotImplementedError( - f"In the child class {self.__class__.__name__} of {super().__name__.__name__} " + f"In the child class {self.__class__.__name__} of {super().__class__.__name__} " f"implementing the model {self.model_name()} we could not find the method " "called `_build_model`. Please do implement it." ) @@ -167,7 +167,7 @@ def _extract_embeddings( Whether to return a dataframe of a numpy array. """ raise NotImplementedError( - f"In the child class {self.__class__.__name__} of {super().__name__.__name__} " + f"In the child class {self.__class__.__name__} of {super().__class__.__name__} " f"implementing the model {self.model_name()} we could not find the method " "called `_extract_embeddings`. Please do implement it." ) diff --git a/embiggen/embedders/pytorch_geometric/__init__.py b/embiggen/embedders/pytorch_geometric/__init__.py new file mode 100644 index 00000000..8983b48d --- /dev/null +++ b/embiggen/embedders/pytorch_geometric/__init__.py @@ -0,0 +1,8 @@ +"""Module with node embedding models based on PyTorch Geometric.""" +from embiggen.utils.abstract_models import build_init, AbstractEmbeddingModel + +build_init( + module_library_names=["torch", "torch_geometric", "torch_cluster", "torch_sparse", "torch_scatter"], + formatted_library_name="PyTorch Geometric", + expected_parent_class=AbstractEmbeddingModel +) \ No newline at end of file diff --git a/embiggen/embedders/pytorch_geometric/node2vec.py b/embiggen/embedders/pytorch_geometric/node2vec.py new file mode 100644 index 00000000..7d9d463a --- /dev/null +++ b/embiggen/embedders/pytorch_geometric/node2vec.py @@ -0,0 +1,142 @@ +"""Abstract class for graph embedding models.""" +from typing import Dict, Any + +from embiggen.embedders.pytorch_geometric.pytorch_geometric_embedder import PyTorchGeometricEmbedder +from torch_geometric.nn import Node2Vec +from torch import Tensor +from torch.nn import Module +from torch.optim import Optimizer +from torch import DeviceObjType + + +class Node2VecPyTorchGeometric(PyTorchGeometricEmbedder): + """Abstract class for sequence embedding models.""" + + def __init__( + self, + + embedding_size: int = 100, + epochs: int = 30, + number_of_negative_samples: int = 10, + walk_length: int = 128, + iterations: int = 10, + window_size: int = 5, + batch_size: int = 128, + learning_rate: float = 0.01, + return_weight: float = 0.25, + explore_weight: float = 4.0, + random_state: int = 42, + optimizer: str = "adam", + verbose: bool = False, + ring_bell: bool = False, + enable_cache: bool = False + ): + """Create new PyTorch Geometric Node2Vec model. + + Parameters + ------------------------------- + + """ + self._number_of_negative_samples = number_of_negative_samples + self._embedding_size = embedding_size + self._walk_length = walk_length + self._window_size = window_size + self._iterations = iterations + self._return_weight = return_weight + self._explore_weight = explore_weight + + self._loader = None + + super().__init__( + random_state=random_state, + embedding_size=embedding_size, + epochs=epochs, + batch_size=batch_size, + learning_rate=learning_rate, + optimizer=optimizer, + verbose=verbose, + ring_bell=ring_bell, + enable_cache=enable_cache + ) + + def _build_model( + self, + edge_node_ids: Tensor, + number_of_nodes: int + ) -> Node2Vec: + model = Node2Vec( + edge_index=edge_node_ids, + embedding_dim=self._embedding_size, + walk_length=self._walk_length, + context_size=2self._window_size, + walks_per_node=self._iterations, + p=1.0/self._return_weight, + q=1.0/self._explore_weight, + num_negative_samples=self._number_of_negative_samples, + num_nodes=number_of_nodes, + ) + + self._loader = model.loader( + batch_size=self._batch_size, + shuffle=True, + num_workers=self._number_of_workers + ) + + return model + + @classmethod + def model_name(cls) -> str: + """Returns name of the model""" + return "Node2Vec" + + @classmethod + def can_use_edge_weights(cls) -> bool: + """Returns whether the model can optionally use edge weights.""" + return False + + def parameters(self) -> Dict[str, Any]: + """Returns parameters of the model.""" + return dict( + super().parameters(), + walk_length=self._walk_length, + window_size=self._window_size, + iterations=self._iterations, + return_weight=self._return_weight, + explore_weight=self._explore_weight, + number_of_negative_samples=self._number_of_negative_samples, + ) + + @classmethod + def smoke_test_parameters(cls) -> Dict[str, Any]: + """Returns parameters for smoke test.""" + return dict( + super().smoke_test_parameters(), + number_of_negative_samples=1, + window_size=1, + walk_length=2, + iterations=1 + ) + + def _train_model_step( + self, + model: Module, + optimizer: Optimizer, + device: DeviceObjType + ): + """Train model for a single step. + + Parameters + ------------------ + model: Module + The model to be trained for this step. + optimizer: Optimizer + The optimizer to be trained for this step. + device: DeviceObjType + The device to be used. + """ + model.train() + for pos_rw, neg_rw in self._loader: + optimizer.zero_grad() + loss = model.loss(pos_rw.to(device), neg_rw.to(device)) + loss.backward() + optimizer.step() diff --git a/embiggen/embedders/pytorch_geometric/pytorch_geometric_embedder.py b/embiggen/embedders/pytorch_geometric/pytorch_geometric_embedder.py new file mode 100644 index 00000000..473a9f86 --- /dev/null +++ b/embiggen/embedders/pytorch_geometric/pytorch_geometric_embedder.py @@ -0,0 +1,233 @@ +"""Abstract Torch/PyTorch Geometric Model wrapper for embedding models.""" +from typing import Dict, Union, Any, Type + +import numpy as np +import pandas as pd +from ensmallen import Graph + +from embiggen.utils.pytorch_utils import validate_torch_device +from embiggen.utils.abstract_models import AbstractEmbeddingModel, abstract_class, EmbeddingResult +import torch + +from tqdm.auto import trange + +from multiprocessing import cpu_count +from environments_utils import is_windows, is_macos +from torch import Tensor +from torch.nn import Module +from torch.optim import Optimizer +from torch import DeviceObjType + + +@abstract_class +class PyTorchGeometricEmbedder(AbstractEmbeddingModel): + """Abstract PyTorch Geometric Model wrapper for embedding models.""" + + def __init__( + self, + embedding_size: int = 100, + epochs: int = 100, + batch_size: int = 210, + learning_rate: float = 0.01, + number_of_workers: Union[int, str] = "auto", + device: str = "auto", + verbose: bool = False, + random_state: int = 42, + optimizer: str = "adam", + ring_bell: bool = False, + enable_cache: bool = False + ): + """Create new PyTorch Geometric Abstract Embedder model. + + Parameters + ------------------------- + embedding_size: int = 100 + The dimension of the embedding to compute. + epochs: int = 100 + The number of epochs to use to train the model for. + batch_size: int = 210 + Size of the training batch. + learning_rate: float = 0.01 + Learning rate of the model. + device: str = "auto" + The devide to use to train the model. + Can either be cpu or cuda. + verbose: bool = False + Whether to show the loading bar. + random_state: int = 42 + Random seed to use while training the model + ring_bell: bool = False, + Whether to play a sound when embedding completes. + enable_cache: bool = False + Whether to enable the cache, that is to + store the computed embedding. + """ + self._epochs = epochs + self._verbose = verbose + self._batch_size = batch_size + self._learning_rate = learning_rate + if number_of_workers == "auto": + number_of_workers = 0 if is_windows() or is_macos() else cpu_count() + self._optimizer = optimizer + self._number_of_workers = number_of_workers + self._device = validate_torch_device(device) + + super().__init__( + embedding_size=embedding_size, + enable_cache=enable_cache, + ring_bell=ring_bell, + random_state=random_state + ) + + @classmethod + def smoke_test_parameters(cls) -> Dict[str, Any]: + """Returns parameters for smoke test.""" + return dict( + embedding_size=10, + epochs=1 + ) + + def parameters(self) -> Dict[str, Any]: + return dict( + super().parameters(), + dict( + epochs=self._epochs, + batch_size=self._batch_size, + ) + ) + + @classmethod + def library_name(cls) -> str: + return "PyTorch Geometric" + + @classmethod + def task_name(cls) -> str: + return "Node Embedding" + + def _build_model( + self, + edge_node_ids: Tensor, + number_of_nodes: int + ) -> Type[Module]: + """Build new model for embedding. + + Parameters + ------------------ + edge_node_ids: Tensor + Tuples with the source and destination node ids + number_of_nodes: int + Number of nodes in the graph. + """ + raise NotImplementedError( + f"In the child class {self.__class__.__name__} of {super().__class__.__name__} " + f"implementing the model {self.model_name()} we could not find the method " + "called `_build_model`. Please do implement it." + ) + + def _train_model_step( + self, + model: Module, + optimizer: Optimizer, + device: DeviceObjType + ): + """Train model for a single step. + + Parameters + ------------------ + model: Module + The model to be trained for this step. + optimizer: Optimizer + The optimizer to be trained for this step. + device: DeviceObjType + The device to be used. + """ + raise NotImplementedError( + f"In the child class {self.__class__.__name__} of {super().__class__.__name__} " + f"implementing the model {self.model_name()} we could not find the method " + "called `_train_model_step`. Please do implement it." + ) + + def _fit_transform( + self, + graph: Graph, + return_dataframe: bool = True, + ) -> Union[np.ndarray, pd.DataFrame, Dict[str, np.ndarray], Dict[str, pd.DataFrame]]: + """Return node embedding""" + + torch_device = torch.device(self._device) + + edge_node_ids = torch.LongTensor(np.int64(graph.get_directed_edge_node_ids().T)) + + model = self._build_model( + edge_node_ids=edge_node_ids, + number_of_nodes=graph.get_number_of_nodes() + ) + + # Move the model to gpu if we need to + model.to(torch_device) + + if not issubclass(model.__class__, Module): + raise NotImplementedError( + "The model created with the `_build_model` in the child " + f"class {self.__class__.__name__} for the model {self.model_name()} " + f"in the library {self.library_name()} did not return a " + f"PyTorch Geometric model but an object of type {type(model)}." + ) + + optimizer = torch.optim.Adam( + list(model.parameters()), + lr=self._learning_rate + ) + + for _ in trange( + self._epochs, + dynamic_ncols=True, + desc="Epochs", + disable=not self._verbose, + leave=False + ): + self._train_model_step( + model=model, + optimizer=optimizer, + device=torch_device + ) + + # Extract and return the embedding + node_embedding = model(torch.arange( + graph.get_number_of_nodes(), + device=torch_device + )).cpu().detach().numpy() + + if return_dataframe: + node_embedding = pd.DataFrame( + node_embedding, + index=graph.get_node_names() + ) + + return EmbeddingResult( + embedding_method_name=self.model_name(), + node_embeddings=node_embedding + ) + + @classmethod + def requires_nodes_sorted_by_decreasing_node_degree(cls) -> bool: + return False + + @classmethod + def is_topological(cls) -> bool: + return True + + @classmethod + def can_use_node_types(cls) -> bool: + """Returns whether the model can use node types.""" + return False + + @classmethod + def can_use_edge_types(cls) -> bool: + """Returns whether the model can use edge types.""" + return False + + @classmethod + def is_stocastic(cls) -> bool: + """Returns whether the model is stocastic and has therefore a random state.""" + return True \ No newline at end of file diff --git a/embiggen/embedders/tensorflow_embedders/__init__.py b/embiggen/embedders/tensorflow_embedders/__init__.py index 4d109788..79dcb159 100644 --- a/embiggen/embedders/tensorflow_embedders/__init__.py +++ b/embiggen/embedders/tensorflow_embedders/__init__.py @@ -1,4 +1,4 @@ -"""Module with graph embedding models based on TensorFlow.""" +"""Module with node embedding models based on TensorFlow.""" from embiggen.utils.abstract_models import build_init, AbstractEmbeddingModel build_init( diff --git a/embiggen/embedders/tensorflow_embedders/edge_prediction_based_tensorflow_embedders.py b/embiggen/embedders/tensorflow_embedders/edge_prediction_based_tensorflow_embedders.py index b55422fb..cb6048a5 100644 --- a/embiggen/embedders/tensorflow_embedders/edge_prediction_based_tensorflow_embedders.py +++ b/embiggen/embedders/tensorflow_embedders/edge_prediction_based_tensorflow_embedders.py @@ -30,7 +30,7 @@ def __init__( use_mirrored_strategy: bool = False, activation: str = "sigmoid", loss: str = "binary_crossentropy", - optimizer: str = "nadam", + optimizer: str = "adam", verbose: bool = False, ring_bell: bool = False, enable_cache: bool = False, diff --git a/embiggen/embedders/tensorflow_embedders/siamese.py b/embiggen/embedders/tensorflow_embedders/siamese.py index 16636969..99246008 100644 --- a/embiggen/embedders/tensorflow_embedders/siamese.py +++ b/embiggen/embedders/tensorflow_embedders/siamese.py @@ -32,7 +32,7 @@ def __init__( learning_rate_plateau_patience: int = 5, norm: str = "L2", use_mirrored_strategy: bool = False, - optimizer: str = "nadam", + optimizer: str = "adam", verbose: bool = False, ring_bell: bool = False, enable_cache: bool = False, diff --git a/embiggen/embedders/tensorflow_embedders/tensorflow_embedder.py b/embiggen/embedders/tensorflow_embedders/tensorflow_embedder.py index b33c5fae..9d5b0103 100644 --- a/embiggen/embedders/tensorflow_embedders/tensorflow_embedder.py +++ b/embiggen/embedders/tensorflow_embedders/tensorflow_embedder.py @@ -27,7 +27,7 @@ def __init__( learning_rate_plateau_patience: int = 1, epochs: int = 10, batch_size: int = 210, - optimizer: str = "nadam", + optimizer: str = "adam", verbose: bool = False, use_mirrored_strategy: bool = False, ring_bell: bool = False, @@ -129,7 +129,7 @@ def _build_model(self, graph: Graph) -> Model: The graph to build the model for. """ raise NotImplementedError( - f"In the child class {self.__class__.__name__} of {super().__name__.__name__} " + f"In the child class {self.__class__.__name__} of {super().__class__.__name__} " f"implementing the model {self.model_name()} we could not find the method " "called `_build_model`. Please do implement it." ) @@ -145,7 +145,7 @@ def _build_input(self, graph: Graph) -> Tuple[Any]: Whether to show loading bars while building input. """ raise NotImplementedError( - f"In the child class {self.__class__.__name__} of {super().__name__.__name__} " + f"In the child class {self.__class__.__name__} of {super().__class__.__name__} " f"implementing the model {self.model_name()} we could not find the method " "called `_build_input`. Please do implement it." ) diff --git a/embiggen/layers/tensorflow/graph_convolution_layer.py b/embiggen/layers/tensorflow/graph_convolution_layer.py index 06a51eb0..44b4a014 100644 --- a/embiggen/layers/tensorflow/graph_convolution_layer.py +++ b/embiggen/layers/tensorflow/graph_convolution_layer.py @@ -109,6 +109,17 @@ def l2_norm(x): return x super().build(input_shape) + def get_config(self): + config = super(GraphConvolution, self).get_config() + config.update({ + "units": self._units, + "activation": self._activation, + "combiner": self._combiner, + "dropout_rate": self._dropout_rate, + "apply_norm": self._apply_norm, + }) + return config + def call( self, inputs: Tuple[Union[tf.Tensor, List[tf.Tensor], tf.SparseTensor]], diff --git a/embiggen/node_label_prediction/node_label_prediction_model.py b/embiggen/node_label_prediction/node_label_prediction_model.py index 0b02c4d3..eb981a51 100644 --- a/embiggen/node_label_prediction/node_label_prediction_model.py +++ b/embiggen/node_label_prediction/node_label_prediction_model.py @@ -157,7 +157,21 @@ def _evaluate( ) if self.is_binary_prediction_task(): - predictions = prediction_probabilities + if prediction_probabilities.shape[1] == 1: + predictions = prediction_probabilities + elif prediction_probabilities.shape[1] == 2: + predictions = prediction_probabilities[:, 1] + prediction_probabilities = prediction_probabilities[:, 1] + else: + raise NotImplementedError( + f"The model {self.model_name()} as implemented in " + f"the library {self.library_name()} for the task " + f"{self.task_name()} has produced a binary prediction " + f"result with shape {prediction_probabilities.shape}, " + "which is unclear how to handle for evaluation. " + "Please open an issue and pull request to clarify what " + "you expect to happen here." + ) elif self.is_multilabel_prediction_task(): predictions = prediction_probabilities > 0.5 else: diff --git a/embiggen/node_label_prediction/node_label_prediction_tensorflow/gcn.py b/embiggen/node_label_prediction/node_label_prediction_tensorflow/gcn.py index 326b7e7a..b8acca36 100644 --- a/embiggen/node_label_prediction/node_label_prediction_tensorflow/gcn.py +++ b/embiggen/node_label_prediction/node_label_prediction_tensorflow/gcn.py @@ -177,7 +177,7 @@ def _build_model( for units in self._number_of_units_per_head_layer: hidden = Dense( units=units, - activation="ReLU" + activation="relu" )(hidden) output = Dense( diff --git a/embiggen/utils/abstract_gcn.py b/embiggen/utils/abstract_gcn.py index 5cac7202..10bf7c88 100644 --- a/embiggen/utils/abstract_gcn.py +++ b/embiggen/utils/abstract_gcn.py @@ -1,7 +1,7 @@ """Kipf GCN model for node-label prediction.""" from typing import List, Union, Optional, Dict, Any, Type, Tuple -from matplotlib.pyplot import axis +import compress_pickle import numpy as np import pandas as pd from tensorflow.keras.callbacks import EarlyStopping, ReduceLROnPlateau # pylint: disable=import-error,no-name-in-module @@ -14,7 +14,7 @@ from keras_mixed_sequence import Sequence from embiggen.utils.abstract_models import AbstractClassifierModel, abstract_class from embiggen.utils.number_to_ordinal import number_to_ordinal -from embiggen.layers.tensorflow import GraphConvolution, FlatEmbedding +from embiggen.layers.tensorflow import GraphConvolution, FlatEmbedding, EmbeddingLookup, L2Norm from tensorflow.keras.layers import Input, Concatenate import warnings from embiggen.utils.normalize_model_structural_parameters import normalize_model_list_parameter @@ -628,6 +628,9 @@ def _predict_proba( edge_features: Optional[List[np.ndarray]] = None, ) -> pd.DataFrame: """Run predictions on the provided graph.""" + if not graph.has_edges(): + return np.array([]) + if support is None: support = graph @@ -742,4 +745,29 @@ def is_using_edge_weights(self) -> bool: """Returns whether the model is parametrized to use edge weights.""" return not self._use_simmetric_normalized_laplacian - \ No newline at end of file + @classmethod + def load(cls, path: str) -> "Self": + """Load a saved version of the model from the provided path. + + Parameters + ------------------- + path: str + Path from where to load the model. + """ + with tf.keras.utils.custom_object_scope({ + "GraphConvolution": GraphConvolution, + "EmbeddingLookup": EmbeddingLookup, + "FlatEmbedding": FlatEmbedding, + "L2Norm": L2Norm + }): + return compress_pickle.load(path) + + def dump(self, path: str): + """Dump the current model at the provided path. + + Parameters + ------------------- + path: str + Path from where to dump the model. + """ + compress_pickle.dump(self, path) \ No newline at end of file diff --git a/embiggen/utils/abstract_models/abstract_classifier_model.py b/embiggen/utils/abstract_models/abstract_classifier_model.py index e3548115..7a972701 100644 --- a/embiggen/utils/abstract_models/abstract_classifier_model.py +++ b/embiggen/utils/abstract_models/abstract_classifier_model.py @@ -1448,15 +1448,6 @@ def _evaluate_on_single_holdout( holdouts_kwargs ) - # We enable in the train and test graphs the same - # speedups enabled in the provided graph. - train.enable( - vector_sources=graph.has_sources_tradeoff_enabled(), - vector_destinations=graph.has_destinations_tradeoff_enabled(), - vector_cumulative_node_degrees=graph.has_cumulative_node_degrees_tradeoff_enabled(), - vector_reciprocal_sqrt_degrees=graph.has_reciprocal_sqrt_degrees_tradeoff_enabled() - ) - # We compute the remaining features starting_to_compute_node_features = time.time() holdout_node_features = cls.normalize_node_features( @@ -1556,22 +1547,6 @@ def _evaluate_on_single_holdout( train_of_interest = train test_of_interest = test - # We enable in the train and test graphs of interest the same - # speedups enabled in the provided graph. - train_of_interest.enable( - vector_sources=graph.has_sources_tradeoff_enabled(), - vector_destinations=graph.has_destinations_tradeoff_enabled(), - vector_cumulative_node_degrees=graph.has_cumulative_node_degrees_tradeoff_enabled(), - vector_reciprocal_sqrt_degrees=graph.has_reciprocal_sqrt_degrees_tradeoff_enabled() - ) - - test_of_interest.enable( - vector_sources=graph.has_sources_tradeoff_enabled(), - vector_destinations=graph.has_destinations_tradeoff_enabled(), - vector_cumulative_node_degrees=graph.has_cumulative_node_degrees_tradeoff_enabled(), - vector_reciprocal_sqrt_degrees=graph.has_reciprocal_sqrt_degrees_tradeoff_enabled() - ) - additional_validation_kwargs = cls._prepare_evaluation( graph=graph, support=train, diff --git a/embiggen/utils/abstract_models/auto_init.py b/embiggen/utils/abstract_models/auto_init.py index 4ddeef21..3bde401a 100644 --- a/embiggen/utils/abstract_models/auto_init.py +++ b/embiggen/utils/abstract_models/auto_init.py @@ -7,6 +7,7 @@ from glob import glob from embiggen.utils.abstract_models.model_stub import get_model_or_stub import os +import sys import inspect import traceback @@ -195,8 +196,17 @@ def build_init( expected_parent_class: Type[AbstractModel] The class to check for. """ + # The Python default is 1000. In some situation, + # users may want to modify this and set the global + # tracebacklimit to a value too low for us to execute + # the meta-programming task of initializing GRAPE's + # models. For these reasons, we set it explicitly + # to the default Python value, i.e. `1000` without + # touching the sys variables that the user may have + # customized. + stack_trace = traceback.extract_stack(limit=1000) path_pattern = os.path.join(os.path.dirname(os.path.abspath( - traceback.extract_stack()[-2].filename + stack_trace[-2].filename )), ".py") # We retrieve the context of the caller. @@ -219,6 +229,7 @@ def build_init( klass.name == expected_parent_class.__name__ ): continue + # If this class has the expected parent. if expected_parent_class.__name__ in get_class_parent_names( path, diff --git a/embiggen/utils/abstract_models/embedding_result.py b/embiggen/utils/abstract_models/embedding_result.py index dd610656..4fa9fe9e 100644 --- a/embiggen/utils/abstract_models/embedding_result.py +++ b/embiggen/utils/abstract_models/embedding_result.py @@ -86,6 +86,14 @@ def __init__( "contains NaN values." ) + if np.isinf(numpy_embedding).any(): + number = np.sum(np.isinf(numpy_embedding)) + raise ValueError( + f"One of the provided {embedding_list_name} " + f"computed with the {embedding_method_name} method " + f"contains {number} infinite values." + ) + if np.isclose(numpy_embedding, 0.0).all(): raise ValueError( f"One of the provided {embedding_list_name} " diff --git a/embiggen/visualizations/graph_visualizer.py b/embiggen/visualizations/graph_visualizer.py index 3c567c8c..3caff43d 100644 --- a/embiggen/visualizations/graph_visualizer.py +++ b/embiggen/visualizations/graph_visualizer.py @@ -2206,8 +2206,8 @@ def _plot_positive_and_negative_edges_metric( types = np.concatenate([ np.zeros( - self._negative_edge_decomposition.shape[0], dtype=np.bool), - np.ones(self._positive_edge_decomposition.shape[0], dtype=np.bool), + self._negative_edge_decomposition.shape[0], dtype=bool), + np.ones(self._positive_edge_decomposition.shape[0], dtype=bool), ]) test_accuracies = [] @@ -4204,6 +4204,7 @@ def plot_node_degree_distribution( figure: Optional[Figure] = None, axes: Optional[Figure] = None, apply_tight_layout: bool = True, + show_title: bool = True, return_caption: bool = True, ) -> Tuple[Figure, Axes]: """Plot the given graph node degree distribution. @@ -4219,6 +4220,8 @@ def plot_node_degree_distribution( apply_tight_layout: bool = True, Whether to apply the tight layout on the matplotlib Figure object. + show_title: bool = True + Wether to show the figure title. return_caption: bool = True, Whether to return a caption for the plot. """ @@ -4234,13 +4237,14 @@ def plot_node_degree_distribution( bins=number_of_buckets, log=True ) - axes.set_ylabel("Counts (log scale)") - axes.set_xlabel("Degrees") + axes.set_ylabel("Count (log scale)") + axes.set_xlabel("Degree") if self._show_graph_name: title = "Degree distribution of graph {}".format(self._graph_name) else: title = "Degree distribution" - axes.set_title(title) + if show_title: + axes.set_title(title) if apply_tight_layout: figure.tight_layout() diff --git a/model.pkl b/model.pkl new file mode 100644 index 00000000..9dca220f Binary files /dev/null and b/model.pkl differ diff --git a/setup.py b/setup.py index 151159bc..40b868f8 100644 --- a/setup.py +++ b/setup.py @@ -72,7 +72,7 @@ def find_version(*file_paths): "ddd_subplots>=1.0.23", "sanitize_ml_labels>=1.0.50", "keras_mixed_sequence>=1.0.28", - "ensmallen>=0.8.31", + "ensmallen>=0.8.38", "environments_utils>=1.0.6", "compress_pickle>=2.1.0", "validate_version_code",	Exposing more APIs from Ensmallen in Embiggen for edge prediction		2023-04-11T12:58:47	[]	[]
monarch-initiative/embiggen	monarch-initiative__embiggen-300	c868d6264c8728cbbf647d75d57b03b570ee8056	diff --git a/dist/embiggen-0.11.40.tar.gz b/dist/embiggen-0.11.40.tar.gz new file mode 100644 index 00000000..7d1c04fd Binary files /dev/null and b/dist/embiggen-0.11.40.tar.gz differ diff --git a/embiggen/__version__.py b/embiggen/__version__.py index 819b838c..e8f9ef5b 100644 --- a/embiggen/__version__.py +++ b/embiggen/__version__.py @@ -1,2 +1,2 @@ """Current version of package Embiggen.""" -__version__ = "0.11.39" \ No newline at end of file +__version__ = "0.11.41" \ No newline at end of file diff --git a/embiggen/edge_prediction/edge_prediction_evaluation.py b/embiggen/edge_prediction/edge_prediction_evaluation.py index 3b7280ff..9c989521 100644 --- a/embiggen/edge_prediction/edge_prediction_evaluation.py +++ b/embiggen/edge_prediction/edge_prediction_evaluation.py @@ -23,6 +23,12 @@ def edge_prediction_evaluation( repositories: Optional[Union[str, List[str]]] = None, versions: Optional[Union[str, List[str]]] = None, validation_sample_only_edges_with_heterogeneous_node_types: bool = False, + source_node_types_names: Optional[List[str]] = None, + destination_node_types_names: Optional[List[str]] = None, + source_edge_types_names: Optional[List[str]] = None, + destination_edge_types_names: Optional[List[str]] = None, + source_nodes_prefixes: Optional[List[str]] = None, + destination_nodes_prefixes: Optional[List[str]] = None, validation_unbalance_rates: Tuple[float] = (1.0, ), use_scale_free_distribution: bool = True, enable_cache: bool = False, @@ -98,6 +104,24 @@ def edge_prediction_evaluation( validation_sample_only_edges_with_heterogeneous_node_types: bool = False Whether to sample negative edges exclusively between nodes with different node types. This can be useful when executing a bipartite edge prediction task. + source_node_types_names: Optional[List[str]] + Node type names of the nodes to be samples as sources. + If a node has any of the provided node types, it can be sampled as a source node. + destination_node_types_names: Optional[List[str]] + Node type names of the nodes to be samples as destinations. + If a node has any of the provided node types, it can be sampled as a destination node. + source_edge_types_names: Optional[List[str]] + Edge type names of the nodes to be samples as sources. + If a node has any of the provided edge types, it can be sampled as a source node. + destination_edge_types_names: Optional[List[str]] + Edge type names of the nodes to be samples as destinations. + If a node has any of the provided edge types, it can be sampled as a destination node. + source_nodes_prefixes: Optional[List[str]] + Prefixes of the nodes names to be samples as sources. + If a node starts with any of the provided prefixes, it can be sampled as a source node. + destination_nodes_prefixes: Optional[List[str]] + Prefixes of the nodes names to be samples as destinations. + If a node starts with any of the provided prefixes, it can be sampled as a destinations node. validation_unbalance_rates: Tuple[float] = (1.0, ) Unbalance rate for the non-existent graphs generation. use_scale_free_distribution: bool = True @@ -154,6 +178,12 @@ def edge_prediction_evaluation( slurm_node_id_variable=slurm_node_id_variable, verbose=verbose, validation_sample_only_edges_with_heterogeneous_node_types=validation_sample_only_edges_with_heterogeneous_node_types, + source_node_types_names=source_node_types_names, + destination_node_types_names=destination_node_types_names, + source_edge_types_names=source_edge_types_names, + destination_edge_types_names=destination_edge_types_names, + source_nodes_prefixes=source_nodes_prefixes, + destination_nodes_prefixes=destination_nodes_prefixes, validation_unbalance_rates=validation_unbalance_rates, use_scale_free_distribution=use_scale_free_distribution ) diff --git a/embiggen/edge_prediction/edge_prediction_model.py b/embiggen/edge_prediction/edge_prediction_model.py index 0edf5dda..4426c63c 100644 --- a/embiggen/edge_prediction/edge_prediction_model.py +++ b/embiggen/edge_prediction/edge_prediction_model.py @@ -101,6 +101,12 @@ def __iterate_negative_graphs( random_state: int, verbose: bool, validation_sample_only_edges_with_heterogeneous_node_types: bool, + source_node_types_names: Optional[List[str]], + destination_node_types_names: Optional[List[str]], + source_edge_types_names: Optional[List[str]], + destination_edge_types_names: Optional[List[str]], + source_nodes_prefixes: Optional[List[str]], + destination_nodes_prefixes: Optional[List[str]], validation_unbalance_rates: Tuple[float], use_scale_free_distribution: bool ) -> Iterator[Tuple[Graph]]: @@ -136,6 +142,12 @@ def __iterate_negative_graphs( random_state=random_state(i+1), sample_only_edges_with_heterogeneous_node_types=validation_sample_only_edges_with_heterogeneous_node_types, use_scale_free_distribution=use_scale_free_distribution, + source_node_types_names=source_node_types_names, + destination_node_types_names=destination_node_types_names, + source_edge_types_names=source_edge_types_names, + destination_edge_types_names=destination_edge_types_names, + source_nodes_prefixes=source_nodes_prefixes, + destination_nodes_prefixes=destination_nodes_prefixes, support=support, graph_to_avoid=graph ).random_holdout( @@ -164,6 +176,12 @@ def _prepare_evaluation( random_state: int = 42, verbose: bool = True, validation_sample_only_edges_with_heterogeneous_node_types: bool = False, + source_node_types_names: Optional[List[str]] = None, + destination_node_types_names: Optional[List[str]] = None, + source_edge_types_names: Optional[List[str]] = None, + destination_edge_types_names: Optional[List[str]] = None, + source_nodes_prefixes: Optional[List[str]] = None, + destination_nodes_prefixes: Optional[List[str]] = None, validation_unbalance_rates: Tuple[float] = (1.0, ), use_scale_free_distribution: bool = True ) -> Dict[str, Any]: @@ -178,6 +196,12 @@ def _prepare_evaluation( random_state=random_state, verbose=verbose, validation_sample_only_edges_with_heterogeneous_node_types=validation_sample_only_edges_with_heterogeneous_node_types, + source_node_types_names=source_node_types_names, + destination_node_types_names=destination_node_types_names, + source_edge_types_names=source_edge_types_names, + destination_edge_types_names=destination_edge_types_names, + source_nodes_prefixes=source_nodes_prefixes, + destination_nodes_prefixes=destination_nodes_prefixes, validation_unbalance_rates=validation_unbalance_rates, use_scale_free_distribution=use_scale_free_distribution )) @@ -197,6 +221,12 @@ def _evaluate( verbose: bool = True, negative_graphs: Optional[List[Tuple[Graph]]] = None, validation_sample_only_edges_with_heterogeneous_node_types: bool = False, + source_node_types_names: Optional[List[str]] = None, + destination_node_types_names: Optional[List[str]] = None, + source_edge_types_names: Optional[List[str]] = None, + destination_edge_types_names: Optional[List[str]] = None, + source_nodes_prefixes: Optional[List[str]] = None, + destination_nodes_prefixes: Optional[List[str]] = None, validation_unbalance_rates: Tuple[float] = (1.0, ), use_scale_free_distribution: bool = True, ) -> List[Dict[str, Any]]: @@ -232,6 +262,12 @@ def _evaluate( random_state=random_state, verbose=verbose, validation_sample_only_edges_with_heterogeneous_node_types=validation_sample_only_edges_with_heterogeneous_node_types, + source_node_types_names=source_node_types_names, + destination_node_types_names=destination_node_types_names, + source_edge_types_names=source_edge_types_names, + destination_edge_types_names=destination_edge_types_names, + source_nodes_prefixes=source_nodes_prefixes, + destination_nodes_prefixes=destination_nodes_prefixes, validation_unbalance_rates=validation_unbalance_rates, use_scale_free_distribution=use_scale_free_distribution ) if negative_graphs is None else negative_graphs diff --git a/embiggen/embedders/ensmallen_embedders/rubicone.py b/embiggen/embedders/ensmallen_embedders/rubicone.py new file mode 100644 index 00000000..d01807d1 --- /dev/null +++ b/embiggen/embedders/ensmallen_embedders/rubicone.py @@ -0,0 +1,122 @@ +"""Module providing RUBICONE implementation.""" +from typing import Optional, Dict, Any +from ensmallen import Graph +import pandas as pd +from ensmallen import models +from embiggen.embedders.ensmallen_embedders.ensmallen_embedder import EnsmallenEmbedder +from embiggen.utils import EmbeddingResult + + +class RUBICONE(EnsmallenEmbedder): + """Class implementing the RUBICONE algorithm.""" + + def __init__( + self, + embedding_size: int = 100, + dtype: Optional[str] = "u8", + number_of_convolutions: int = 2, + path: Optional[str] = None, + verbose: bool = False, + ring_bell: bool = False, + enable_cache: bool = False + ): + """Create new RUBICONE method. + + Parameters + -------------------------- + embedding_size: int = 100 + Dimension of the embedding. + dtype: Optional[str] = "u8" + Dtype to use for the embedding. + number_of_convolutions: int = 2 + Number of convolutions to execute + path: Optional[str] = None + Path where to store the mmap-ed embedding. + This parameter is necessary to embed very large graphs. + verbose: bool = False + Whether to show loading bars. + ring_bell: bool = False, + Whether to play a sound when embedding completes. + enable_cache: bool = False + Whether to enable the cache, that is to + store the computed embedding. + """ + self._dtype = dtype + self._verbose = verbose + self._number_of_convolutions = number_of_convolutions + self._path = path + self._model = models.RUBICONE( + embedding_size=embedding_size, + verbose=self._verbose, + number_of_convolutions=self._number_of_convolutions, + path=self._path + ) + + super().__init__( + embedding_size=embedding_size, + ring_bell=ring_bell, + enable_cache=enable_cache + ) + + def parameters(self) -> Dict[str, Any]: + """Returns parameters of the model.""" + return dict( + super().parameters(), + dict( + dtype=self._dtype, + number_of_convolutions=self._number_of_convolutions, + path=self._path, + ) + ) + + @classmethod + def smoke_test_parameters(cls) -> Dict[str, Any]: + """Returns parameters for smoke test.""" + return dict( + embedding_size=5, + ) + + def _fit_transform( + self, + graph: Graph, + return_dataframe: bool = True, + ) -> EmbeddingResult: + """Return node embedding.""" + node_embedding = self._model.fit_transform( + graph, + dtype=self._dtype, + ) + if return_dataframe: + node_embedding = pd.DataFrame( + node_embedding, + index=graph.get_node_names() + ) + return EmbeddingResult( + embedding_method_name=self.model_name(), + node_embeddings=node_embedding + ) + + @classmethod + def model_name(cls) -> str: + """Returns name of the model.""" + return "RUBICONE" + + @classmethod + def can_use_node_types(cls) -> bool: + """Returns whether the model can optionally use node types.""" + return False + + @classmethod + def can_use_edge_weights(cls) -> bool: + """Returns whether the model can optionally use edge weights.""" + return False + + @classmethod + def can_use_edge_types(cls) -> bool: + """Returns whether the model can optionally use edge types.""" + return False + + @classmethod + def is_stocastic(cls) -> bool: + """Returns whether the model is stocastic and has therefore a random state.""" + return False \ No newline at end of file diff --git a/embiggen/embedders/ensmallen_embedders/ruine.py b/embiggen/embedders/ensmallen_embedders/ruine.py new file mode 100644 index 00000000..9393b70e --- /dev/null +++ b/embiggen/embedders/ensmallen_embedders/ruine.py @@ -0,0 +1,122 @@ +"""Module providing RUINE implementation.""" +from typing import Optional, Dict, Any +from ensmallen import Graph +import pandas as pd +from ensmallen import models +from embiggen.embedders.ensmallen_embedders.ensmallen_embedder import EnsmallenEmbedder +from embiggen.utils import EmbeddingResult + + +class RUINE(EnsmallenEmbedder): + """Class implementing the RUINE algorithm.""" + + def __init__( + self, + embedding_size: int = 100, + dtype: Optional[str] = "u8", + number_of_convolutions: int = 2, + path: Optional[str] = None, + verbose: bool = False, + ring_bell: bool = False, + enable_cache: bool = False + ): + """Create new RUINE method. + + Parameters + -------------------------- + embedding_size: int = 100 + Dimension of the embedding. + dtype: Optional[str] = "u8" + Dtype to use for the embedding. + number_of_convolutions: int = 2 + Number of convolutions to execute + path: Optional[str] = None + Path where to store the mmap-ed embedding. + This parameter is necessary to embed very large graphs. + verbose: bool = False + Whether to show loading bars. + ring_bell: bool = False, + Whether to play a sound when embedding completes. + enable_cache: bool = False + Whether to enable the cache, that is to + store the computed embedding. + """ + self._dtype = dtype + self._verbose = verbose + self._number_of_convolutions = number_of_convolutions + self._path = path + self._model = models.RUINE( + embedding_size=embedding_size, + verbose=self._verbose, + number_of_convolutions=self._number_of_convolutions, + path=self._path + ) + + super().__init__( + embedding_size=embedding_size, + ring_bell=ring_bell, + enable_cache=enable_cache + ) + + def parameters(self) -> Dict[str, Any]: + """Returns parameters of the model.""" + return dict( + super().parameters(), + dict( + dtype=self._dtype, + number_of_convolutions=self._number_of_convolutions, + path=self._path, + ) + ) + + @classmethod + def smoke_test_parameters(cls) -> Dict[str, Any]: + """Returns parameters for smoke test.""" + return dict( + embedding_size=5, + ) + + def _fit_transform( + self, + graph: Graph, + return_dataframe: bool = True, + ) -> EmbeddingResult: + """Return node embedding.""" + node_embedding = self._model.fit_transform( + graph, + dtype=self._dtype, + ) + if return_dataframe: + node_embedding = pd.DataFrame( + node_embedding, + index=graph.get_node_names() + ) + return EmbeddingResult( + embedding_method_name=self.model_name(), + node_embeddings=node_embedding + ) + + @classmethod + def model_name(cls) -> str: + """Returns name of the model.""" + return "RUINE" + + @classmethod + def can_use_node_types(cls) -> bool: + """Returns whether the model can optionally use node types.""" + return False + + @classmethod + def can_use_edge_weights(cls) -> bool: + """Returns whether the model can optionally use edge weights.""" + return False + + @classmethod + def can_use_edge_types(cls) -> bool: + """Returns whether the model can optionally use edge types.""" + return False + + @classmethod + def is_stocastic(cls) -> bool: + """Returns whether the model is stocastic and has therefore a random state.""" + return False \ No newline at end of file diff --git a/embiggen/node_label_prediction/__init__.py b/embiggen/node_label_prediction/__init__.py index e4c95469..2b894472 100644 --- a/embiggen/node_label_prediction/__init__.py +++ b/embiggen/node_label_prediction/__init__.py @@ -1,5 +1,6 @@ """Submodule providing models for edge-label prediction.""" from embiggen.node_label_prediction.node_label_prediction_sklearn import +from embiggen.node_label_prediction.node_label_prediction_ensmallen import * from embiggen.node_label_prediction.node_label_prediction_tensorflow import * from embiggen.node_label_prediction.node_label_prediction_evaluation import node_label_prediction_evaluation diff --git a/embiggen/node_label_prediction/node_label_prediction_ensmallen/__init__.py b/embiggen/node_label_prediction/node_label_prediction_ensmallen/__init__.py new file mode 100644 index 00000000..81c097da --- /dev/null +++ b/embiggen/node_label_prediction/node_label_prediction_ensmallen/__init__.py @@ -0,0 +1,9 @@ +"""Submodule providing node embedding models implemented in Ensmallen in Rust.""" +from embiggen.node_label_prediction.node_label_prediction_model import AbstractNodeLabelPredictionModel +from embiggen.utils.abstract_models import build_init + +build_init( + module_library_names="ensmallen", + formatted_library_name="Ensmallen", + expected_parent_class=AbstractNodeLabelPredictionModel +) diff --git a/embiggen/node_label_prediction/node_label_prediction_ensmallen/distance_based_perceptron.py b/embiggen/node_label_prediction/node_label_prediction_ensmallen/distance_based_perceptron.py new file mode 100644 index 00000000..a62133b6 --- /dev/null +++ b/embiggen/node_label_prediction/node_label_prediction_ensmallen/distance_based_perceptron.py @@ -0,0 +1,245 @@ +"""Module providing Perceptron for edge prediction.""" +from typing import Optional, Dict, Any, List +from ensmallen import Graph +import numpy as np +from ensmallen import models +import compress_json +import json +from embiggen.node_label_prediction.node_label_prediction_model import AbstractNodeLabelPredictionModel + + +class DistanceBasedPerceptronNodeLabelPrediction(AbstractNodeLabelPredictionModel): + """Distance-based Perceptron model for edge prediction.""" + + def __init__( + self, + number_of_centroids_per_class: int = 1, + number_of_epochs: int = 100, + learning_rate: float = 0.01, + first_order_decay_factor: float = 0.9, + second_order_decay_factor: float = 0.999, + random_state: int = 42, + verbose: bool = True + ): + """Create new Distance-based Perceptron object. + + Parameters + ------------------------ + number_of_centroids_per_class: int = 1 + Number of centroids to compute for each of the classes. + number_of_epochs: int = 100 + The number of epochs to train the model for. By default, 100. + learning_rate: float = 0.01 + Learning rate to use while training the model. By default 0.001. + first_order_decay_factor: float = 0.9 + First order decay factor for the first order momentum. + By default 0.9. + second_order_decay_factor: float = 0.999 + Second order decay factor for the second order momentum. + By default 0.999. + random_state: int = 42 + The random state to reproduce the model initialization and training. By default, 42. + verbose: bool = True + Whether to show epochs loading bar. + """ + super().__init__(random_state=random_state) + + self._model_kwargs = dict( + number_of_centroids_per_class=number_of_centroids_per_class, + number_of_epochs=number_of_epochs, + learning_rate=learning_rate, + first_order_decay_factor=first_order_decay_factor, + second_order_decay_factor=second_order_decay_factor, + ) + self._verbose = verbose + self._model = models.DistanceNodeLabelPredictionPerceptron( + self._model_kwargs, + random_state=random_state + ) + + def parameters(self) -> Dict[str, Any]: + """Returns parameters used for this model.""" + return dict( + super().parameters(), + self._model_kwargs + ) + + def clone(self) -> "DistanceBasedPerceptronNodeLabelPrediction": + return DistanceBasedPerceptronNodeLabelPrediction(self.parameters()) + + @classmethod + def smoke_test_parameters(cls) -> Dict[str, Any]: + """Returns parameters for smoke test.""" + return dict( + number_of_epochs=1, + ) + + def _fit( + self, + graph: Graph, + support: Optional[Graph] = None, + node_features: Optional[List[np.ndarray]] = None, + node_type_features: Optional[List[np.ndarray]] = None, + edge_features: Optional[List[np.ndarray]] = None, + ): + """Run fitting on the provided graph. + + Parameters + -------------------- + graph: Graph + The graph to run predictions on. + support: Optional[Graph] = None + The graph describiding the topological structure that + includes also the above graph. This parameter + is mostly useful for topological classifiers + such as Graph Convolutional Networks. + node_features: Optional[List[np.ndarray]] = None + The node features to use. + node_type_features: Optional[List[np.ndarray]] = None + The node type features to use. + edge_features: Optional[List[np.ndarray]] = None + The edge features to use. + """ + new_node_features = [] + if node_features is not None: + for node_feature in node_features: + if not node_feature.data.c_contiguous: + node_feature = np.ascontiguousarray(node_feature) + new_node_features.append(node_feature) + + self._model.fit( + graph=graph, + node_features=new_node_features, + verbose=self._verbose, + ) + + def _predict( + self, + graph: Graph, + support: Optional[Graph] = None, + node_features: Optional[List[np.ndarray]] = None, + node_type_features: Optional[List[np.ndarray]] = None, + edge_features: Optional[List[np.ndarray]] = None, + ) -> np.ndarray: + """Run prediction on the provided graph. + + Parameters + -------------------- + graph: Graph + The graph to run predictions on. + support: Optional[Graph] = None + The graph describiding the topological structure that + includes also the above graph. This parameter + is mostly useful for topological classifiers + such as Graph Convolutional Networks. + node_features: Optional[List[np.ndarray]] + The node features to use. + node_type_features: Optional[List[np.ndarray]] = None + The node type features to use. + edge_features: Optional[List[np.ndarray]] = None + The edge features to use. + """ + return self._predict_proba( + graph=graph, + support=support, + node_features=node_features, + node_type_features=node_type_features, + edge_features=edge_features + ).argmax(axis=1) + + def _predict_proba( + self, + graph: Graph, + support: Optional[Graph] = None, + node_features: Optional[List[np.ndarray]] = None, + node_type_features: Optional[List[np.ndarray]] = None, + edge_features: Optional[List[np.ndarray]] = None, + ) -> np.ndarray: + """Run prediction on the provided graph. + + Parameters + -------------------- + graph: Graph + The graph to run predictions on. + support: Optional[Graph] = None + The graph describiding the topological structure that + includes also the above graph. This parameter + is mostly useful for topological classifiers + such as Graph Convolutional Networks. + node_features: Optional[List[np.ndarray]] = None + The node features to use. + node_type_features: Optional[List[np.ndarray]] = None + The node type features to use. + edge_features: Optional[List[np.ndarray]] = None + The edge features to use. + """ + new_node_features = [] + if node_features is not None: + for node_feature in node_features: + if not node_feature.data.c_contiguous: + node_feature = np.ascontiguousarray(node_feature) + new_node_features.append(node_feature) + + return self._model.predict( + graph=graph, + node_features=new_node_features, + ) + + @classmethod + def can_use_edge_weights(cls) -> bool: + """Returns whether the model can optionally use edge weights.""" + return False + + @classmethod + def can_use_edge_types(cls) -> bool: + """Returns whether the model can optionally use edge types.""" + return False + + @classmethod + def model_name(cls) -> str: + return "Distance-based Perceptron" + + @classmethod + def library_name(cls) -> str: + return "Ensmallen" + + @classmethod + def load(cls, path: str) -> "Self": + """Load a saved version of the model from the provided path. + + Parameters + ------------------- + path: str + Path from where to load the model. + """ + data = compress_json.load(path) + model = DistanceBasedPerceptronNodeLabelPrediction(data["parameters"]) + model._model = models.DistanceNodeLabelPredictionPerceptron.loads( + json.dumps(data["inner_model"]) + ) + for key, value in data["metadata"].items(): + model.__setattr__(key, value) + return model + + def dumps(self) -> Dict[str, Any]: + """Dumps the current model as dictionary.""" + return dict( + parameters=self.parameters(), + inner_model=json.loads(self._model.dumps()), + metadata=dict( + _fitting_was_executed=self._fitting_was_executed + ) + ) + + def dump(self, path: str): + """Dump the current model at the provided path. + + Parameters + ------------------- + path: str + Path from where to dump the model. + """ + compress_json.dump( + self.dumps(), + path + ) diff --git a/embiggen/node_label_prediction/node_label_prediction_ensmallen/perceptron.py b/embiggen/node_label_prediction/node_label_prediction_ensmallen/perceptron.py new file mode 100644 index 00000000..6b2b254e --- /dev/null +++ b/embiggen/node_label_prediction/node_label_prediction_ensmallen/perceptron.py @@ -0,0 +1,241 @@ +"""Module providing Perceptron for edge prediction.""" +from typing import Optional, Dict, Any, List +from ensmallen import Graph +import numpy as np +from ensmallen import models +import compress_json +import json +from embiggen.node_label_prediction.node_label_prediction_model import AbstractNodeLabelPredictionModel + + +class PerceptronNodeLabelPrediction(AbstractNodeLabelPredictionModel): + """Perceptron model for edge prediction.""" + + def __init__( + self, + number_of_epochs: int = 100, + learning_rate: float = 0.01, + first_order_decay_factor: float = 0.9, + second_order_decay_factor: float = 0.999, + random_state: int = 42, + verbose: bool = True + ): + """Create new Perceptron object. + + Parameters + ------------------------ + number_of_epochs: int = 100 + The number of epochs to train the model for. By default, 100. + learning_rate: float = 0.01 + Learning rate to use while training the model. By default 0.001. + first_order_decay_factor: float = 0.9 + First order decay factor for the first order momentum. + By default 0.9. + second_order_decay_factor: float = 0.999 + Second order decay factor for the second order momentum. + By default 0.999. + random_state: int = 42 + The random state to reproduce the model initialization and training. By default, 42. + verbose: bool = True + Whether to show epochs loading bar. + """ + super().__init__(random_state=random_state) + + self._model_kwargs = dict( + number_of_epochs=number_of_epochs, + learning_rate=learning_rate, + first_order_decay_factor=first_order_decay_factor, + second_order_decay_factor=second_order_decay_factor, + ) + self._verbose = verbose + self._model = models.NodeLabelPredictionPerceptron( + self._model_kwargs, + random_state=random_state + ) + + def parameters(self) -> Dict[str, Any]: + """Returns parameters used for this model.""" + return dict( + super().parameters(), + self._model_kwargs + ) + + def clone(self) -> "PerceptronNodeLabelPrediction": + return PerceptronNodeLabelPrediction(self.parameters()) + + @classmethod + def smoke_test_parameters(cls) -> Dict[str, Any]: + """Returns parameters for smoke test.""" + return dict( + number_of_epochs=1, + ) + + def _fit( + self, + graph: Graph, + support: Optional[Graph] = None, + node_features: Optional[List[np.ndarray]] = None, + node_type_features: Optional[List[np.ndarray]] = None, + edge_features: Optional[List[np.ndarray]] = None, + ): + """Run fitting on the provided graph. + + Parameters + -------------------- + graph: Graph + The graph to run predictions on. + support: Optional[Graph] = None + The graph describiding the topological structure that + includes also the above graph. This parameter + is mostly useful for topological classifiers + such as Graph Convolutional Networks. + node_features: Optional[List[np.ndarray]] = None + The node features to use. + node_type_features: Optional[List[np.ndarray]] = None + The node type features to use. + edge_features: Optional[List[np.ndarray]] = None + The edge features to use. + """ + new_node_features = [] + if node_features is not None: + for node_feature in node_features: + if not node_feature.data.c_contiguous: + node_feature = np.ascontiguousarray(node_feature) + new_node_features.append(node_feature) + + self._model.fit( + graph=graph, + node_features=new_node_features, + verbose=self._verbose, + ) + + def _predict( + self, + graph: Graph, + support: Optional[Graph] = None, + node_features: Optional[List[np.ndarray]] = None, + node_type_features: Optional[List[np.ndarray]] = None, + edge_features: Optional[List[np.ndarray]] = None, + ) -> np.ndarray: + """Run prediction on the provided graph. + + Parameters + -------------------- + graph: Graph + The graph to run predictions on. + support: Optional[Graph] = None + The graph describiding the topological structure that + includes also the above graph. This parameter + is mostly useful for topological classifiers + such as Graph Convolutional Networks. + node_features: Optional[List[np.ndarray]] + The node features to use. + node_type_features: Optional[List[np.ndarray]] = None + The node type features to use. + edge_features: Optional[List[np.ndarray]] = None + The edge features to use. + """ + return self._predict_proba( + graph=graph, + support=support, + node_features=node_features, + node_type_features=node_type_features, + edge_features=edge_features + ).argmax(axis=1) + + def _predict_proba( + self, + graph: Graph, + support: Optional[Graph] = None, + node_features: Optional[List[np.ndarray]] = None, + node_type_features: Optional[List[np.ndarray]] = None, + edge_features: Optional[List[np.ndarray]] = None, + ) -> np.ndarray: + """Run prediction on the provided graph. + + Parameters + -------------------- + graph: Graph + The graph to run predictions on. + support: Optional[Graph] = None + The graph describiding the topological structure that + includes also the above graph. This parameter + is mostly useful for topological classifiers + such as Graph Convolutional Networks. + node_features: Optional[List[np.ndarray]] = None + The node features to use. + node_type_features: Optional[List[np.ndarray]] = None + The node type features to use. + edge_features: Optional[List[np.ndarray]] = None + The edge features to use. + """ + new_node_features = [] + if node_features is not None: + for node_feature in node_features: + if not node_feature.data.c_contiguous: + node_feature = np.ascontiguousarray(node_feature) + new_node_features.append(node_feature) + + return self._model.predict( + graph=graph, + node_features=new_node_features, + ) + + @classmethod + def can_use_edge_weights(cls) -> bool: + """Returns whether the model can optionally use edge weights.""" + return False + + @classmethod + def can_use_edge_types(cls) -> bool: + """Returns whether the model can optionally use edge types.""" + return False + + @classmethod + def model_name(cls) -> str: + return "Perceptron" + + @classmethod + def library_name(cls) -> str: + return "Ensmallen" + + @classmethod + def load(cls, path: str) -> "Self": + """Load a saved version of the model from the provided path. + + Parameters + ------------------- + path: str + Path from where to load the model. + """ + data = compress_json.load(path) + model = PerceptronNodeLabelPrediction(data["parameters"]) + model._model = models.NodeLabelPredictionPerceptron.loads( + json.dumps(data["inner_model"]) + ) + for key, value in data["metadata"].items(): + model.__setattr__(key, value) + return model + + def dumps(self) -> Dict[str, Any]: + """Dumps the current model as dictionary.""" + return dict( + parameters=self.parameters(), + inner_model=json.loads(self._model.dumps()), + metadata=dict( + _fitting_was_executed=self._fitting_was_executed + ) + ) + + def dump(self, path: str): + """Dump the current model at the provided path. + + Parameters + ------------------- + path: str + Path from where to dump the model. + """ + compress_json.dump( + self.dumps(), + path + ) diff --git a/embiggen/similarities/dag_resnik.py b/embiggen/similarities/dag_resnik.py index c78ba754..5a765993 100644 --- a/embiggen/similarities/dag_resnik.py +++ b/embiggen/similarities/dag_resnik.py @@ -107,6 +107,12 @@ def get_similarities_from_bipartite_graph_node_ids( Whether to return the node names or node IDs associated to the scores. By default we return the node ids, which require much less memory. """ + if isinstance(source_node_ids, int): + source_node_ids = [source_node_ids] + + if isinstance(destination_node_ids, int): + destination_node_ids = [destination_node_ids] + return self._normalize_output( self._model.get_node_ids_and_similarity_from_node_ids( first_node_ids=source_node_ids, @@ -142,6 +148,12 @@ def get_similarities_from_bipartite_graph_node_names( Whether to return the node names or node IDs associated to the scores. By default we return the node ids, which require much less memory. """ + if isinstance(source_node_names, str): + source_node_names = [source_node_names] + + if isinstance(destination_node_names, str): + destination_node_names = [destination_node_names] + return self._normalize_output( self._model.get_node_ids_and_similarity_from_node_names( first_node_names=source_node_names, @@ -177,6 +189,12 @@ def get_similarities_from_bipartite_graph_node_prefixes( Whether to return the node names or node IDs associated to the scores. By default we return the node ids, which require much less memory. """ + if isinstance(source_node_prefixes, str): + source_node_prefixes = [source_node_prefixes] + + if isinstance(destination_node_prefixes, str): + destination_node_prefixes = [destination_node_prefixes] + return self._normalize_output( self._model.get_node_ids_and_similarity_from_node_prefixes( first_node_prefixes=source_node_prefixes, @@ -212,6 +230,12 @@ def get_similarities_from_bipartite_graph_node_type_ids( Whether to return the node names or node IDs associated to the scores. By default we return the node ids, which require much less memory. """ + if isinstance(source_node_type_ids, int): + source_node_type_ids = [source_node_type_ids] + + if isinstance(destination_node_type_ids, int): + destination_node_type_ids = [destination_node_type_ids] + return self._normalize_output( self._model.get_node_ids_and_similarity_from_node_type_ids( first_node_type_ids=source_node_type_ids, @@ -247,6 +271,12 @@ def get_similarities_from_bipartite_graph_node_type_names( Whether to return the node names or node IDs associated to the scores. By default we return the node ids, which require much less memory. """ + if isinstance(source_node_type_names, str): + source_node_type_names = [source_node_type_names] + + if isinstance(destination_node_type_names, str): + destination_node_type_names = [destination_node_type_names] + return self._normalize_output( self._model.get_node_ids_and_similarity_from_node_type_names( first_node_type_names=source_node_type_names, diff --git a/embiggen/utils/abstract_models/embedding_result.py b/embiggen/utils/abstract_models/embedding_result.py index 4dd922f4..dd610656 100644 --- a/embiggen/utils/abstract_models/embedding_result.py +++ b/embiggen/utils/abstract_models/embedding_result.py @@ -62,12 +62,15 @@ def __init__( f"computed with the {embedding_method_name} method is neither a " f"numpy array or a pandas DataFrame, but a `{type(embedding)}` object." ) + if embedding.shape[0] == 0: raise ValueError( "One of the provided {embedding_list_name} " f"computed with the {embedding_method_name} method " "is empty." ) + + # If the embedding size is too big, we skip the checking step. if embedding.shape[0] > 1_000_000: continue diff --git a/embiggen/visualizations/graph_visualizer.py b/embiggen/visualizations/graph_visualizer.py index 2043217b..3c567c8c 100644 --- a/embiggen/visualizations/graph_visualizer.py +++ b/embiggen/visualizations/graph_visualizer.py @@ -4237,9 +4237,9 @@ def plot_node_degree_distribution( axes.set_ylabel("Counts (log scale)") axes.set_xlabel("Degrees") if self._show_graph_name: - title = "Degrees distribution of graph {}".format(self._graph_name) + title = "Degree distribution of graph {}".format(self._graph_name) else: - title = "Degrees distribution" + title = "Degree distribution" axes.set_title(title) if apply_tight_layout: figure.tight_layout() @@ -4248,7 +4248,7 @@ def plot_node_degree_distribution( return self._handle_notebook_display(figure, axes) caption = ( - "<i>Node degrees distribution.</i> Node degrees are on the " + "<i>Node degree distribution.</i> Node degrees are on the " "horizontal axis and node counts are on the vertical axis on a logarithmic scale." ) diff --git a/setup.py b/setup.py index ae9a18ea..151159bc 100644 --- a/setup.py +++ b/setup.py @@ -64,14 +64,15 @@ def find_version(file_paths): 'numpy', 'pandas', "tqdm", + "humanize", "matplotlib>=3.5.2", "scikit-learn", "dict_hash>=1.1.29", "userinput>=1.0.19", "ddd_subplots>=1.0.23", - "sanitize_ml_labels>=1.0.45", + "sanitize_ml_labels>=1.0.50", "keras_mixed_sequence>=1.0.28", - "ensmallen>=0.8.27", + "ensmallen>=0.8.31", "environments_utils>=1.0.6", "compress_pickle>=2.1.0", "validate_version_code",	Brought develop up to speed		2022-11-11T14:25:57	[]	[]
monarch-initiative/embiggen	monarch-initiative__embiggen-297	3ee8055ad4dce31a1cc04bc071116155bdbcd57d	diff --git a/.gitignore b/.gitignore index 6634e04f..b8107f2c 100644 --- a/.gitignore +++ b/.gitignore @@ -34,4 +34,7 @@ embedding.npy build graphs embeddings -.webm \ No newline at end of file +.webm + +model.json +model.pkl.gz \ No newline at end of file diff --git a/README.rst b/README.rst index 87387cda..4d3e7707 100644 --- a/README.rst +++ b/README.rst @@ -2,11 +2,11 @@ ========================================================================================= \|pip\| \|downloads\| \|tutorials\| \|documentation\| \|python_version\| \|DOI\| \|license\| \|telegram\| \|discord\| \|twitter\| -Embiggen is the graph machine learning submodule of the `🍇 GraPE <https://github.com/AnacletoLAB/grape>`_ library. +Embiggen is the graph machine learning submodule of the `🍇 GRAPE <https://github.com/AnacletoLAB/grape>`_ library. How to install Embiggen ------------------------- -To install the complete GraPE library, do run: +To install the complete GRAPE library, do run: .. code:: bash @@ -19,14 +19,14 @@ Instead, to exclusively install the Embiggen package, you can run: pip install embiggen -Cite GraPE +Cite GRAPE ---------------------------------------------- Please cite the following paper if it was useful for your research: .. code:: bib @misc{cappelletti2021grape, - title={GraPE: fast and scalable Graph Processing and Embedding}, + title={GRAPE: fast and scalable Graph Processing and Embedding}, author={Luca Cappelletti and Tommaso Fontana and Elena Casiraghi and Vida Ravanmehr and Tiffany J. Callahan and Marcin P. Joachimiak and Christopher J. Mungall and Peter N. Robinson and Justin Reese and Giorgio Valentini}, year={2021}, eprint={2110.06196}, @@ -58,7 +58,7 @@ Please cite the following paper if it was useful for your research: :target: https://doi.org/10.48550/arXiv.2110.06196 :alt: DOI -.. \|python_version\| image:: https://img.shields.io/badge/Python-3.6+-blue.svg +.. \|python_version\| image:: https://img.shields.io/badge/Python-3.7+-blue.svg :target: https://pypi.org/project/embiggen/#history :alt: Supported Python versions @@ -72,4 +72,4 @@ Please cite the following paper if it was useful for your research: .. \|twitter\| image:: https://badges.aleen42.com/src/twitter.svg :target: https://twitter.com/grapelib - :alt: Twitter \ No newline at end of file + :alt: Twitter diff --git a/dist/embiggen-0.11.11.tar.gz b/dist/embiggen-0.11.11.tar.gz new file mode 100644 index 00000000..3b80024f Binary files /dev/null and b/dist/embiggen-0.11.11.tar.gz differ diff --git a/dist/embiggen-0.11.16.tar.gz b/dist/embiggen-0.11.16.tar.gz new file mode 100644 index 00000000..b8076636 Binary files /dev/null and b/dist/embiggen-0.11.16.tar.gz differ diff --git a/dist/embiggen-0.11.17.tar.gz b/dist/embiggen-0.11.17.tar.gz new file mode 100644 index 00000000..2313906e Binary files /dev/null and b/dist/embiggen-0.11.17.tar.gz differ diff --git a/dist/embiggen-0.11.18.tar.gz b/dist/embiggen-0.11.18.tar.gz new file mode 100644 index 00000000..439b6f68 Binary files /dev/null and b/dist/embiggen-0.11.18.tar.gz differ diff --git a/dist/embiggen-0.11.19.tar.gz b/dist/embiggen-0.11.19.tar.gz new file mode 100644 index 00000000..67bae59b Binary files /dev/null and b/dist/embiggen-0.11.19.tar.gz differ diff --git a/dist/embiggen-0.11.20.tar.gz b/dist/embiggen-0.11.20.tar.gz new file mode 100644 index 00000000..70b73659 Binary files /dev/null and b/dist/embiggen-0.11.20.tar.gz differ diff --git a/dist/embiggen-0.11.21.tar.gz b/dist/embiggen-0.11.21.tar.gz new file mode 100644 index 00000000..d1361443 Binary files /dev/null and b/dist/embiggen-0.11.21.tar.gz differ diff --git a/dist/embiggen-0.11.22.tar.gz b/dist/embiggen-0.11.22.tar.gz new file mode 100644 index 00000000..1518eac1 Binary files /dev/null and b/dist/embiggen-0.11.22.tar.gz differ diff --git a/dist/embiggen-0.11.23.tar.gz b/dist/embiggen-0.11.23.tar.gz new file mode 100644 index 00000000..42de56f0 Binary files /dev/null and b/dist/embiggen-0.11.23.tar.gz differ diff --git a/dist/embiggen-0.11.24.tar.gz b/dist/embiggen-0.11.24.tar.gz new file mode 100644 index 00000000..ea524493 Binary files /dev/null and b/dist/embiggen-0.11.24.tar.gz differ diff --git a/dist/embiggen-0.11.25.tar.gz b/dist/embiggen-0.11.25.tar.gz new file mode 100644 index 00000000..323bb085 Binary files /dev/null and b/dist/embiggen-0.11.25.tar.gz differ diff --git a/dist/embiggen-0.11.26.tar.gz b/dist/embiggen-0.11.26.tar.gz new file mode 100644 index 00000000..0690228b Binary files /dev/null and b/dist/embiggen-0.11.26.tar.gz differ diff --git a/dist/embiggen-0.11.27.tar.gz b/dist/embiggen-0.11.27.tar.gz new file mode 100644 index 00000000..e3969a24 Binary files /dev/null and b/dist/embiggen-0.11.27.tar.gz differ diff --git a/dist/embiggen-0.11.28.tar.gz b/dist/embiggen-0.11.28.tar.gz new file mode 100644 index 00000000..e9d94564 Binary files /dev/null and b/dist/embiggen-0.11.28.tar.gz differ diff --git a/dist/embiggen-0.11.29.tar.gz b/dist/embiggen-0.11.29.tar.gz new file mode 100644 index 00000000..31fabc70 Binary files /dev/null and b/dist/embiggen-0.11.29.tar.gz differ diff --git a/dist/embiggen-0.11.30.tar.gz b/dist/embiggen-0.11.30.tar.gz new file mode 100644 index 00000000..d49052f3 Binary files /dev/null and b/dist/embiggen-0.11.30.tar.gz differ diff --git a/dist/embiggen-0.11.31.tar.gz b/dist/embiggen-0.11.31.tar.gz new file mode 100644 index 00000000..6a8086e4 Binary files /dev/null and b/dist/embiggen-0.11.31.tar.gz differ diff --git a/dist/embiggen-0.11.32.tar.gz b/dist/embiggen-0.11.32.tar.gz new file mode 100644 index 00000000..c2042520 Binary files /dev/null and b/dist/embiggen-0.11.32.tar.gz differ diff --git a/dist/embiggen-0.11.33.tar.gz b/dist/embiggen-0.11.33.tar.gz new file mode 100644 index 00000000..21fe8809 Binary files /dev/null and b/dist/embiggen-0.11.33.tar.gz differ diff --git a/dist/embiggen-0.11.34.tar.gz b/dist/embiggen-0.11.34.tar.gz new file mode 100644 index 00000000..ef19be77 Binary files /dev/null and b/dist/embiggen-0.11.34.tar.gz differ diff --git a/dist/embiggen-0.11.35.tar.gz b/dist/embiggen-0.11.35.tar.gz new file mode 100644 index 00000000..7b6e0ce5 Binary files /dev/null and b/dist/embiggen-0.11.35.tar.gz differ diff --git a/dist/embiggen-0.11.36.tar.gz b/dist/embiggen-0.11.36.tar.gz new file mode 100644 index 00000000..db48726d Binary files /dev/null and b/dist/embiggen-0.11.36.tar.gz differ diff --git a/dist/embiggen-0.11.37.tar.gz b/dist/embiggen-0.11.37.tar.gz new file mode 100644 index 00000000..761805dc Binary files /dev/null and b/dist/embiggen-0.11.37.tar.gz differ diff --git a/dist/embiggen-0.11.38.tar.gz b/dist/embiggen-0.11.38.tar.gz new file mode 100644 index 00000000..53c56976 Binary files /dev/null and b/dist/embiggen-0.11.38.tar.gz differ diff --git a/embiggen/__version__.py b/embiggen/__version__.py index e1fd9917..6c3f5032 100644 --- a/embiggen/__version__.py +++ b/embiggen/__version__.py @@ -1,2 +1,2 @@ """Current version of package embiggen""" -__version__ = "0.11.15" +__version__ = "0.11.38" \ No newline at end of file diff --git a/embiggen/edge_label_prediction/edge_label_prediction_evaluation.py b/embiggen/edge_label_prediction/edge_label_prediction_evaluation.py index dec73085..57595239 100644 --- a/embiggen/edge_label_prediction/edge_label_prediction_evaluation.py +++ b/embiggen/edge_label_prediction/edge_label_prediction_evaluation.py @@ -18,13 +18,16 @@ def edge_label_prediction_evaluation( library_names: Optional[Union[str, List[str]]] = None, graph_callback: Optional[Callable[[Graph], Graph]] = None, subgraph_of_interest: Optional[Graph] = None, + use_subgraph_as_support: bool = False, number_of_holdouts: int = 10, random_state: int = 42, repositories: Optional[Union[str, List[str]]] = None, versions: Optional[Union[str, List[str]]] = None, enable_cache: bool = False, - precompute_constant_automatic_stocastic_features: bool = False, + precompute_constant_stocastic_features: bool = False, smoke_test: bool = False, + number_of_slurm_nodes: Optional[int] = None, + slurm_node_id_variable: str = "SLURM_GRAPE_ID", verbose: bool = True ) -> pd.DataFrame: """Execute edge-label prediction evaluation pipeline for all provided models and graphs. @@ -53,7 +56,12 @@ def edge_label_prediction_evaluation( For instance this may be used for filtering the uncertain edges in graphs such as STRING PPIs. subgraph_of_interest: Optional[Graph] = None - The subgraph of interest to focus the task on. + Optional subgraph where to focus the task. + This is applied to the train and test graph + after the desired holdout schema is applied. + use_subgraph_as_support: bool = False + Whether to use the provided subgraph as support or + to use the train graph (not filtered by the subgraph). number_of_holdouts: int = 10 The number of holdouts to execute. random_state: int = 42 @@ -65,7 +73,7 @@ def edge_label_prediction_evaluation( Graph versions to retrieve. enable_cache: bool = False Whether to enable the cache. - precompute_constant_automatic_stocastic_features: bool = False + precompute_constant_stocastic_features: bool = False Whether to precompute once the constant automatic stocastic features before starting the embedding loop. This means that, when left set to false, while the features will be computed @@ -79,6 +87,12 @@ def edge_label_prediction_evaluation( and therefore use the smoke test configurations for the provided model names and feature names. This parameter will also turn off the cache. + number_of_slurm_nodes: Optional[int] = None + Number of SLURM nodes to consider as available. + This variable is used to parallelize the holdouts accordingly. + slurm_node_id_variable: str = "SLURM_GRAPE_ID" + Name of the system variable to use as SLURM node id. + It must be set in the slurm bash script. verbose: bool = True Whether to show loading bars """ @@ -94,12 +108,15 @@ def edge_label_prediction_evaluation( library_names=library_names, graph_callback=graph_callback, subgraph_of_interest=subgraph_of_interest, + use_subgraph_as_support=use_subgraph_as_support, number_of_holdouts=number_of_holdouts, random_state=random_state, repositories=repositories, versions=versions, enable_cache=enable_cache, - precompute_constant_automatic_stocastic_features=precompute_constant_automatic_stocastic_features, + precompute_constant_stocastic_features=precompute_constant_stocastic_features, smoke_test=smoke_test, + number_of_slurm_nodes=number_of_slurm_nodes, + slurm_node_id_variable=slurm_node_id_variable, verbose=verbose ) diff --git a/embiggen/edge_label_prediction/edge_label_prediction_sklearn/sklearn_edge_label_prediction_adapter.py b/embiggen/edge_label_prediction/edge_label_prediction_sklearn/sklearn_edge_label_prediction_adapter.py index 013d081c..cf98e7fa 100644 --- a/embiggen/edge_label_prediction/edge_label_prediction_sklearn/sklearn_edge_label_prediction_adapter.py +++ b/embiggen/edge_label_prediction/edge_label_prediction_sklearn/sklearn_edge_label_prediction_adapter.py @@ -3,6 +3,7 @@ from typing import Type, List, Dict, Optional, Any import numpy as np import copy +import compress_pickle from ensmallen import Graph from embiggen.utils.sklearn_utils import must_be_an_sklearn_classifier_model from embiggen.embedding_transformers import EdgeLabelPredictionTransformer, GraphTransformer @@ -295,4 +296,25 @@ def can_use_edge_weights(cls) -> bool: @classmethod def can_use_node_types(cls) -> bool: """Returns whether the model can optionally use node types.""" - return False \ No newline at end of file + return False + + @classmethod + def load(cls, path: str) -> "Self": + """Load a saved version of the model from the provided path. + + Parameters + ------------------- + path: str + Path from where to load the model. + """ + return compress_pickle.load(path) + + def dump(self, path: str): + """Dump the current model at the provided path. + + Parameters + ------------------- + path: str + Path from where to dump the model. + """ + compress_pickle.dump(self, path) \ No newline at end of file diff --git a/embiggen/edge_prediction/edge_prediction_ensmallen/perceptron.py b/embiggen/edge_prediction/edge_prediction_ensmallen/perceptron.py index 5407d194..ea631a63 100644 --- a/embiggen/edge_prediction/edge_prediction_ensmallen/perceptron.py +++ b/embiggen/edge_prediction/edge_prediction_ensmallen/perceptron.py @@ -3,6 +3,8 @@ from ensmallen import Graph import numpy as np from ensmallen import models +import compress_json +import json from embiggen.embedding_transformers import NodeTransformer from embiggen.edge_prediction.edge_prediction_model import AbstractEdgePredictionModel @@ -17,9 +19,9 @@ def __init__( cooccurrence_iterations: int = 100, cooccurrence_window_size: int = 10, number_of_epochs: int = 100, - number_of_edges_per_mini_batch: int = 256, + number_of_edges_per_mini_batch: int = 4096, sample_only_edges_with_heterogeneous_node_types: bool = False, - learning_rate: float = 0.001, + learning_rate: float = 0.01, first_order_decay_factor: float = 0.9, second_order_decay_factor: float = 0.999, avoid_false_negatives: bool = False, @@ -68,7 +70,7 @@ def __init__( sample_only_edges_with_heterogeneous_node_types: bool = False Whether to sample negative edges only with source and destination nodes that have different node types. By default false. - learning_rate: float = 0.001 + learning_rate: float = 0.01 Learning rate to use while training the model. By default 0.001. first_order_decay_factor: float = 0.9 First order decay factor for the first order momentum. @@ -116,7 +118,10 @@ def __init__( def parameters(self) -> Dict[str, Any]: """Returns parameters used for this model.""" - return self._model_kwargs + return dict( + super().parameters(), + self._model_kwargs + ) def clone(self) -> "PerceptronEdgePrediction": return PerceptronEdgePrediction(self.parameters()) @@ -125,7 +130,8 @@ def clone(self) -> "PerceptronEdgePrediction": def smoke_test_parameters(cls) -> Dict[str, Any]: """Returns parameters for smoke test.""" return dict( - number_of_epochs=1 + number_of_epochs=1, + edge_features="Degree" ) def _fit( @@ -176,7 +182,6 @@ def _fit( node_feature = np.ascontiguousarray(node_feature) new_node_features.append(node_feature) - self._model.fit( graph=graph, node_features=new_node_features, @@ -279,3 +284,44 @@ def model_name(cls) -> str: @classmethod def library_name(cls) -> str: return "Ensmallen" + + @classmethod + def load(cls, path: str) -> "Self": + """Load a saved version of the model from the provided path. + + Parameters + ------------------- + path: str + Path from where to load the model. + """ + data = compress_json.load(path) + model = PerceptronEdgePrediction(data["parameters"]) + model._model = models.EdgePredictionPerceptron.loads( + json.dumps(data["inner_model"]) + ) + for key, value in data["metadata"].items(): + model.__setattr__(key, value) + return model + + def dumps(self) -> Dict[str, Any]: + """Dumps the current model as dictionary.""" + return dict( + parameters=self.parameters(), + inner_model=json.loads(self._model.dumps()), + metadata=dict( + _fitting_was_executed=self._fitting_was_executed + ) + ) + + def dump(self, path: str): + """Dump the current model at the provided path. + + Parameters + ------------------- + path: str + Path from where to dump the model. + """ + compress_json.dump( + self.dumps(), + path + ) diff --git a/embiggen/edge_prediction/edge_prediction_evaluation.py b/embiggen/edge_prediction/edge_prediction_evaluation.py index c8bab4f3..3b7280ff 100644 --- a/embiggen/edge_prediction/edge_prediction_evaluation.py +++ b/embiggen/edge_prediction/edge_prediction_evaluation.py @@ -17,6 +17,7 @@ def edge_prediction_evaluation( library_names: Optional[Union[str, List[str]]] = None, graph_callback: Optional[Callable[[Graph], Graph]] = None, subgraph_of_interest: Optional[Graph] = None, + use_subgraph_as_support: bool = False, number_of_holdouts: int = 10, random_state: int = 42, repositories: Optional[Union[str, List[str]]] = None, @@ -25,8 +26,10 @@ def edge_prediction_evaluation( validation_unbalance_rates: Tuple[float] = (1.0, ), use_scale_free_distribution: bool = True, enable_cache: bool = False, - precompute_constant_automatic_stocastic_features: bool = False, + precompute_constant_stocastic_features: bool = False, smoke_test: bool = False, + number_of_slurm_nodes: Optional[int] = None, + slurm_node_id_variable: str = "SLURM_GRAPE_ID", verbose: bool = True ) -> pd.DataFrame: """Execute edge prediction evaluation pipeline for all provided models and graphs. @@ -77,7 +80,12 @@ def edge_prediction_evaluation( For instance this may be used for filtering the uncertain edges in graphs such as STRING PPIs. subgraph_of_interest: Optional[Graph] = None - The subgraph of interest to focus the task on. + Optional subgraph where to focus the task. + This is applied to the train and test graph + after the desired holdout schema is applied. + use_subgraph_as_support: bool = False + Whether to use the provided subgraph as support or + to use the train graph (not filtered by the subgraph). number_of_holdouts: int = 10 The number of holdouts to execute. random_state: int = 42 @@ -100,7 +108,7 @@ def edge_prediction_evaluation( in the model performance. enable_cache: bool = False Whether to enable the cache. - precompute_constant_automatic_stocastic_features: bool = False + precompute_constant_stocastic_features: bool = False Whether to precompute once the constant automatic stocastic features before starting the embedding loop. This means that, when left set to false, while the features will be computed @@ -114,6 +122,12 @@ def edge_prediction_evaluation( and therefore use the smoke test configurations for the provided model names and feature names. This parameter will also turn off the cache. + number_of_slurm_nodes: Optional[int] = None + Number of SLURM nodes to consider as available. + This variable is used to parallelize the holdouts accordingly. + slurm_node_id_variable: str = "SLURM_GRAPE_ID" + Name of the system variable to use as SLURM node id. + It must be set in the slurm bash script. verbose: bool = True Whether to show loading bars """ @@ -128,13 +142,16 @@ def edge_prediction_evaluation( library_names=library_names, graph_callback=graph_callback, subgraph_of_interest=subgraph_of_interest, + use_subgraph_as_support=use_subgraph_as_support, number_of_holdouts=number_of_holdouts, random_state=random_state, repositories=repositories, versions=versions, enable_cache=enable_cache, - precompute_constant_automatic_stocastic_features=precompute_constant_automatic_stocastic_features, + precompute_constant_stocastic_features=precompute_constant_stocastic_features, smoke_test=smoke_test, + number_of_slurm_nodes=number_of_slurm_nodes, + slurm_node_id_variable=slurm_node_id_variable, verbose=verbose, validation_sample_only_edges_with_heterogeneous_node_types=validation_sample_only_edges_with_heterogeneous_node_types, validation_unbalance_rates=validation_unbalance_rates, diff --git a/embiggen/edge_prediction/edge_prediction_model.py b/embiggen/edge_prediction/edge_prediction_model.py index 79950a1b..0edf5dda 100644 --- a/embiggen/edge_prediction/edge_prediction_model.py +++ b/embiggen/edge_prediction/edge_prediction_model.py @@ -123,13 +123,15 @@ def __iterate_negative_graphs( ) train_size = ( - train.get_number_of_edges() / (train.get_number_of_edges() + test.get_number_of_edges()) + train.get_number_of_edges() / (train.get_number_of_edges() + + test.get_number_of_edges()) ) return ( sampler_graph.sample_negative_graph( number_of_negative_samples=int( - math.ceil(sampler_graph.get_number_of_edges()unbalance_rate) + math.ceil(sampler_graph.get_number_of_edges() + unbalance_rate) ), random_state=random_state(i+1), sample_only_edges_with_heterogeneous_node_types=validation_sample_only_edges_with_heterogeneous_node_types, @@ -201,6 +203,10 @@ def _evaluate( """Return model evaluation on the provided graphs.""" performance = [] + train_size = ( + train.get_number_of_directed_edges() / graph.get_number_of_directed_edges() + ) + train_predic_proba = self.predict_proba( train, support=support, @@ -265,6 +271,7 @@ def _evaluate( performance.append({ "evaluation_mode": evaluation_mode, + "train_size": train_size, "validation_unbalance_rate": unbalance_rate, "use_scale_free_distribution": use_scale_free_distribution, "validation_sample_only_edges_with_heterogeneous_node_types": validation_sample_only_edges_with_heterogeneous_node_types, @@ -287,8 +294,9 @@ def predict( node_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, node_type_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, edge_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, - return_predictions_dataframe: bool = False - ) -> np.ndarray: + return_predictions_dataframe: bool = False, + return_node_names: bool = True + ) -> Union[np.ndarray, pd.DataFrame]: """Execute predictions on the provided graph. Parameters @@ -309,6 +317,10 @@ def predict( return_predictions_dataframe: bool = False Whether to return a pandas DataFrame, which as indices has the node IDs. By default, a numpy array with the predictions is returned as it weights much less. + return_node_names: bool = True + Whether to return node names when returning the prediction DataFrame. + This value is ignored when the values to be returned the user has not + requested for a prediction dataframe to be returned. """ if edge_features is not None: raise NotImplementedError( @@ -326,8 +338,8 @@ def predict( predictions = pd.DataFrame( { "predictions": predictions, - "sources": graph.get_directed_source_node_ids(), - "destinations": graph.get_directed_destination_node_ids(), + "sources": graph.get_source_names(directed=True) if return_node_names else graph.get_directed_source_node_ids(), + "destinations": graph.get_destination_names(directed=True) if return_node_names else graph.get_directed_destination_node_ids(), }, ) @@ -342,8 +354,9 @@ def predict_bipartite_graph_from_edge_node_ids( node_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, node_type_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, edge_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, - return_predictions_dataframe: bool = False - ) -> np.ndarray: + return_predictions_dataframe: bool = False, + return_node_names: bool = True + ) -> Union[np.ndarray, pd.DataFrame]: """Execute predictions on the provided graph bipartite portion. Parameters @@ -368,6 +381,10 @@ def predict_bipartite_graph_from_edge_node_ids( return_predictions_dataframe: bool = False Whether to return a pandas DataFrame, which as indices has the node IDs. By default, a numpy array with the predictions is returned as it weights much less. + return_node_names: bool = True + Whether to return node names when returning the prediction DataFrame. + This value is ignored when the values to be returned the user has not + requested for a prediction dataframe to be returned. """ return self.predict( graph.build_bipartite_graph_from_edge_node_ids( @@ -379,7 +396,8 @@ def predict_bipartite_graph_from_edge_node_ids( node_features=node_features, node_type_features=node_type_features, edge_features=edge_features, - return_predictions_dataframe=return_predictions_dataframe + return_predictions_dataframe=return_predictions_dataframe, + return_node_names=return_node_names ) def predict_bipartite_graph_from_edge_node_names( @@ -391,8 +409,9 @@ def predict_bipartite_graph_from_edge_node_names( node_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, node_type_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, edge_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, - return_predictions_dataframe: bool = False - ) -> np.ndarray: + return_predictions_dataframe: bool = False, + return_node_names: bool = True + ) -> Union[np.ndarray, pd.DataFrame]: """Execute predictions on the provided graph bipartite portion. Parameters @@ -417,6 +436,10 @@ def predict_bipartite_graph_from_edge_node_names( return_predictions_dataframe: bool = False Whether to return a pandas DataFrame, which as indices has the node IDs. By default, a numpy array with the predictions is returned as it weights much less. + return_node_names: bool = True + Whether to return node names when returning the prediction DataFrame. + This value is ignored when the values to be returned the user has not + requested for a prediction dataframe to be returned. """ return self.predict( graph.build_bipartite_graph_from_edge_node_names( @@ -428,7 +451,8 @@ def predict_bipartite_graph_from_edge_node_names( node_features=node_features, node_type_features=node_type_features, edge_features=edge_features, - return_predictions_dataframe=return_predictions_dataframe + return_predictions_dataframe=return_predictions_dataframe, + return_node_names=return_node_names ) def predict_bipartite_graph_from_edge_node_prefixes( @@ -440,8 +464,9 @@ def predict_bipartite_graph_from_edge_node_prefixes( node_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, node_type_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, edge_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, - return_predictions_dataframe: bool = False - ) -> np.ndarray: + return_predictions_dataframe: bool = False, + return_node_names: bool = True + ) -> Union[np.ndarray, pd.DataFrame]: """Execute predictions on the provided graph bipartite portion. Parameters @@ -466,6 +491,10 @@ def predict_bipartite_graph_from_edge_node_prefixes( return_predictions_dataframe: bool = False Whether to return a pandas DataFrame, which as indices has the node IDs. By default, a numpy array with the predictions is returned as it weights much less. + return_node_names: bool = True + Whether to return node names when returning the prediction DataFrame. + This value is ignored when the values to be returned the user has not + requested for a prediction dataframe to be returned. """ return self.predict( graph.build_bipartite_graph_from_edge_node_prefixes( @@ -477,7 +506,8 @@ def predict_bipartite_graph_from_edge_node_prefixes( node_features=node_features, node_type_features=node_type_features, edge_features=edge_features, - return_predictions_dataframe=return_predictions_dataframe + return_predictions_dataframe=return_predictions_dataframe, + return_node_names=return_node_names ) def predict_bipartite_graph_from_edge_node_types( @@ -489,8 +519,9 @@ def predict_bipartite_graph_from_edge_node_types( node_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, node_type_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, edge_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, - return_predictions_dataframe: bool = False - ) -> np.ndarray: + return_predictions_dataframe: bool = False, + return_node_names: bool = True + ) -> Union[np.ndarray, pd.DataFrame]: """Execute predictions on the provided graph bipartite portion. Parameters @@ -515,6 +546,10 @@ def predict_bipartite_graph_from_edge_node_types( return_predictions_dataframe: bool = False Whether to return a pandas DataFrame, which as indices has the node IDs. By default, a numpy array with the predictions is returned as it weights much less. + return_node_names: bool = True + Whether to return node names when returning the prediction DataFrame. + This value is ignored when the values to be returned the user has not + requested for a prediction dataframe to be returned. """ return self.predict( graph.build_bipartite_graph_from_edge_node_types( @@ -526,7 +561,8 @@ def predict_bipartite_graph_from_edge_node_types( node_features=node_features, node_type_features=node_type_features, edge_features=edge_features, - return_predictions_dataframe=return_predictions_dataframe + return_predictions_dataframe=return_predictions_dataframe, + return_node_names=return_node_names ) def predict_clique_graph_from_node_ids( @@ -537,8 +573,9 @@ def predict_clique_graph_from_node_ids( node_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, node_type_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, edge_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, - return_predictions_dataframe: bool = False - ) -> np.ndarray: + return_predictions_dataframe: bool = False, + return_node_names: bool = True + ) -> Union[np.ndarray, pd.DataFrame]: """Execute predictions on the provided graph bipartite portion. Parameters @@ -561,6 +598,10 @@ def predict_clique_graph_from_node_ids( return_predictions_dataframe: bool = False Whether to return a pandas DataFrame, which as indices has the node IDs. By default, a numpy array with the predictions is returned as it weights much less. + return_node_names: bool = True + Whether to return node names when returning the prediction DataFrame. + This value is ignored when the values to be returned the user has not + requested for a prediction dataframe to be returned. """ return self.predict( graph.build_clique_graph_from_node_ids( @@ -571,7 +612,8 @@ def predict_clique_graph_from_node_ids( node_features=node_features, node_type_features=node_type_features, edge_features=edge_features, - return_predictions_dataframe=return_predictions_dataframe + return_predictions_dataframe=return_predictions_dataframe, + return_node_names=return_node_names ) def predict_clique_graph_from_node_names( @@ -582,8 +624,9 @@ def predict_clique_graph_from_node_names( node_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, node_type_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, edge_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, - return_predictions_dataframe: bool = False - ) -> np.ndarray: + return_predictions_dataframe: bool = False, + return_node_names: bool = True + ) -> Union[np.ndarray, pd.DataFrame]: """Execute predictions on the provided graph bipartite portion. Parameters @@ -606,6 +649,10 @@ def predict_clique_graph_from_node_names( return_predictions_dataframe: bool = False Whether to return a pandas DataFrame, which as indices has the node IDs. By default, a numpy array with the predictions is returned as it weights much less. + return_node_names: bool = True + Whether to return node names when returning the prediction DataFrame. + This value is ignored when the values to be returned the user has not + requested for a prediction dataframe to be returned. """ return self.predict( graph.build_clique_graph_from_node_names( @@ -616,7 +663,8 @@ def predict_clique_graph_from_node_names( node_features=node_features, node_type_features=node_type_features, edge_features=edge_features, - return_predictions_dataframe=return_predictions_dataframe + return_predictions_dataframe=return_predictions_dataframe, + return_node_names=return_node_names ) def predict_clique_graph_from_node_prefixes( @@ -627,8 +675,9 @@ def predict_clique_graph_from_node_prefixes( node_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, node_type_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, edge_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, - return_predictions_dataframe: bool = False - ) -> np.ndarray: + return_predictions_dataframe: bool = False, + return_node_names: bool = True + ) -> Union[np.ndarray, pd.DataFrame]: """Execute predictions on the provided graph bipartite portion. Parameters @@ -651,6 +700,10 @@ def predict_clique_graph_from_node_prefixes( return_predictions_dataframe: bool = False Whether to return a pandas DataFrame, which as indices has the node IDs. By default, a numpy array with the predictions is returned as it weights much less. + return_node_names: bool = True + Whether to return node names when returning the prediction DataFrame. + This value is ignored when the values to be returned the user has not + requested for a prediction dataframe to be returned. """ return self.predict( graph.build_clique_graph_from_node_prefixes( @@ -661,26 +714,28 @@ def predict_clique_graph_from_node_prefixes( node_features=node_features, node_type_features=node_type_features, edge_features=edge_features, - return_predictions_dataframe=return_predictions_dataframe + return_predictions_dataframe=return_predictions_dataframe, + return_node_names=return_node_names ) - def predict_clique_graph_from_node_types( + def predict_clique_graph_from_node_type_names( self, graph: Graph, - node_types: List[str], + node_type_names: List[str], support: Optional[Graph] = None, node_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, node_type_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, edge_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, - return_predictions_dataframe: bool = False - ) -> np.ndarray: + return_predictions_dataframe: bool = False, + return_node_names: bool = True + ) -> Union[np.ndarray, pd.DataFrame]: """Execute predictions on the provided graph bipartite portion. Parameters -------------------- graph: Graph The graph from which to extract the edges. - node_types: List[str] + node_type_names: List[str] The node prefixes of the bipartite graph. support: Optional[Graph] = None The graph describiding the topological structure that @@ -696,17 +751,22 @@ def predict_clique_graph_from_node_types( return_predictions_dataframe: bool = False Whether to return a pandas DataFrame, which as indices has the node IDs. By default, a numpy array with the predictions is returned as it weights much less. + return_node_names: bool = True + Whether to return node names when returning the prediction DataFrame. + This value is ignored when the values to be returned the user has not + requested for a prediction dataframe to be returned. """ return self.predict( - graph.build_clique_graph_from_node_types( - node_types=node_types, + graph.build_clique_graph_from_node_type_names( + node_type_names=node_type_names, directed=True ), support=support, node_features=node_features, node_type_features=node_type_features, edge_features=edge_features, - return_predictions_dataframe=return_predictions_dataframe + return_predictions_dataframe=return_predictions_dataframe, + return_node_names=return_node_names ) def predict_proba( @@ -716,8 +776,9 @@ def predict_proba( node_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, node_type_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, edge_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, - return_predictions_dataframe: bool = False - ) -> np.ndarray: + return_predictions_dataframe: bool = False, + return_node_names: bool = True + ) -> Union[np.ndarray, pd.DataFrame]: """Execute predictions on the provided graph. Parameters @@ -738,6 +799,10 @@ def predict_proba( return_predictions_dataframe: bool = False Whether to return a pandas DataFrame, which as indices has the node IDs. By default, a numpy array with the predictions is returned as it weights much less. + return_node_names: bool = True + Whether to return node names when returning the prediction DataFrame. + This value is ignored when the values to be returned the user has not + requested for a prediction dataframe to be returned. """ if edge_features is not None: raise NotImplementedError( @@ -760,8 +825,8 @@ def predict_proba( predictions = pd.DataFrame( { "predictions": predictions, - "sources": graph.get_directed_source_node_ids(), - "destinations": graph.get_directed_destination_node_ids(), + "sources": graph.get_source_names(directed=True) if return_node_names else graph.get_directed_source_node_ids(), + "destinations": graph.get_destination_names(directed=True) if return_node_names else graph.get_directed_destination_node_ids(), }, ) @@ -776,8 +841,9 @@ def predict_proba_bipartite_graph_from_edge_node_ids( node_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, node_type_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, edge_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, - return_predictions_dataframe: bool = False - ) -> np.ndarray: + return_predictions_dataframe: bool = False, + return_node_names: bool = True + ) -> Union[np.ndarray, pd.DataFrame]: """Execute predictions probabilities on the provided graph bipartite portion. Parameters @@ -802,6 +868,10 @@ def predict_proba_bipartite_graph_from_edge_node_ids( return_predictions_dataframe: bool = False Whether to return a pandas DataFrame, which as indices has the node IDs. By default, a numpy array with the predictions is returned as it weights much less. + return_node_names: bool = True + Whether to return node names when returning the prediction DataFrame. + This value is ignored when the values to be returned the user has not + requested for a prediction dataframe to be returned. """ return self.predict_proba( graph.build_bipartite_graph_from_edge_node_ids( @@ -813,7 +883,8 @@ def predict_proba_bipartite_graph_from_edge_node_ids( node_features=node_features, node_type_features=node_type_features, edge_features=edge_features, - return_predictions_dataframe=return_predictions_dataframe + return_predictions_dataframe=return_predictions_dataframe, + return_node_names=return_node_names ) def predict_proba_bipartite_graph_from_edge_node_names( @@ -825,8 +896,9 @@ def predict_proba_bipartite_graph_from_edge_node_names( node_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, node_type_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, edge_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, - return_predictions_dataframe: bool = False - ) -> np.ndarray: + return_predictions_dataframe: bool = False, + return_node_names: bool = True + ) -> Union[np.ndarray, pd.DataFrame]: """Execute predictions probabilities on the provided graph bipartite portion. Parameters @@ -851,6 +923,10 @@ def predict_proba_bipartite_graph_from_edge_node_names( return_predictions_dataframe: bool = False Whether to return a pandas DataFrame, which as indices has the node IDs. By default, a numpy array with the predictions is returned as it weights much less. + return_node_names: bool = True + Whether to return node names when returning the prediction DataFrame. + This value is ignored when the values to be returned the user has not + requested for a prediction dataframe to be returned. """ return self.predict_proba( graph.build_bipartite_graph_from_edge_node_names( @@ -862,7 +938,8 @@ def predict_proba_bipartite_graph_from_edge_node_names( node_features=node_features, node_type_features=node_type_features, edge_features=edge_features, - return_predictions_dataframe=return_predictions_dataframe + return_predictions_dataframe=return_predictions_dataframe, + return_node_names=return_node_names ) def predict_proba_bipartite_graph_from_edge_node_prefixes( @@ -874,8 +951,9 @@ def predict_proba_bipartite_graph_from_edge_node_prefixes( node_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, node_type_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, edge_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, - return_predictions_dataframe: bool = False - ) -> np.ndarray: + return_predictions_dataframe: bool = False, + return_node_names: bool = True + ) -> Union[np.ndarray, pd.DataFrame]: """Execute predictions probabilities on the provided graph bipartite portion. Parameters @@ -900,6 +978,10 @@ def predict_proba_bipartite_graph_from_edge_node_prefixes( return_predictions_dataframe: bool = False Whether to return a pandas DataFrame, which as indices has the node IDs. By default, a numpy array with the predictions is returned as it weights much less. + return_node_names: bool = True + Whether to return node names when returning the prediction DataFrame. + This value is ignored when the values to be returned the user has not + requested for a prediction dataframe to be returned. """ return self.predict_proba( graph.build_bipartite_graph_from_edge_node_prefixes( @@ -911,7 +993,8 @@ def predict_proba_bipartite_graph_from_edge_node_prefixes( node_features=node_features, node_type_features=node_type_features, edge_features=edge_features, - return_predictions_dataframe=return_predictions_dataframe + return_predictions_dataframe=return_predictions_dataframe, + return_node_names=return_node_names ) def predict_proba_bipartite_graph_from_edge_node_types( @@ -923,8 +1006,9 @@ def predict_proba_bipartite_graph_from_edge_node_types( node_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, node_type_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, edge_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, - return_predictions_dataframe: bool = False - ) -> np.ndarray: + return_predictions_dataframe: bool = False, + return_node_names: bool = True + ) -> Union[np.ndarray, pd.DataFrame]: """Execute predictions probabilities on the provided graph bipartite portion. Parameters @@ -949,6 +1033,10 @@ def predict_proba_bipartite_graph_from_edge_node_types( return_predictions_dataframe: bool = False Whether to return a pandas DataFrame, which as indices has the node IDs. By default, a numpy array with the predictions is returned as it weights much less. + return_node_names: bool = True + Whether to return node names when returning the prediction DataFrame. + This value is ignored when the values to be returned the user has not + requested for a prediction dataframe to be returned. """ return self.predict_proba( graph.build_bipartite_graph_from_edge_node_types( @@ -960,7 +1048,8 @@ def predict_proba_bipartite_graph_from_edge_node_types( node_features=node_features, node_type_features=node_type_features, edge_features=edge_features, - return_predictions_dataframe=return_predictions_dataframe + return_predictions_dataframe=return_predictions_dataframe, + return_node_names=return_node_names ) def predict_proba_clique_graph_from_node_ids( @@ -971,8 +1060,9 @@ def predict_proba_clique_graph_from_node_ids( node_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, node_type_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, edge_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, - return_predictions_dataframe: bool = False - ) -> np.ndarray: + return_predictions_dataframe: bool = False, + return_node_names: bool = True + ) -> Union[np.ndarray, pd.DataFrame]: """Execute predictions probabilities on the provided graph bipartite portion. Parameters @@ -995,6 +1085,10 @@ def predict_proba_clique_graph_from_node_ids( return_predictions_dataframe: bool = False Whether to return a pandas DataFrame, which as indices has the node IDs. By default, a numpy array with the predictions is returned as it weights much less. + return_node_names: bool = True + Whether to return node names when returning the prediction DataFrame. + This value is ignored when the values to be returned the user has not + requested for a prediction dataframe to be returned. """ return self.predict_proba( graph.build_clique_graph_from_node_ids( @@ -1005,7 +1099,8 @@ def predict_proba_clique_graph_from_node_ids( node_features=node_features, node_type_features=node_type_features, edge_features=edge_features, - return_predictions_dataframe=return_predictions_dataframe + return_predictions_dataframe=return_predictions_dataframe, + return_node_names=return_node_names ) def predict_proba_clique_graph_from_node_names( @@ -1016,8 +1111,9 @@ def predict_proba_clique_graph_from_node_names( node_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, node_type_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, edge_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, - return_predictions_dataframe: bool = False - ) -> np.ndarray: + return_predictions_dataframe: bool = False, + return_node_names: bool = True + ) -> Union[np.ndarray, pd.DataFrame]: """Execute predictions probabilities on the provided graph bipartite portion. Parameters @@ -1040,6 +1136,10 @@ def predict_proba_clique_graph_from_node_names( return_predictions_dataframe: bool = False Whether to return a pandas DataFrame, which as indices has the node IDs. By default, a numpy array with the predictions is returned as it weights much less. + return_node_names: bool = True + Whether to return node names when returning the prediction DataFrame. + This value is ignored when the values to be returned the user has not + requested for a prediction dataframe to be returned. """ return self.predict_proba( graph.build_clique_graph_from_node_names( @@ -1050,7 +1150,8 @@ def predict_proba_clique_graph_from_node_names( node_features=node_features, node_type_features=node_type_features, edge_features=edge_features, - return_predictions_dataframe=return_predictions_dataframe + return_predictions_dataframe=return_predictions_dataframe, + return_node_names=return_node_names ) def predict_proba_clique_graph_from_node_prefixes( @@ -1061,8 +1162,9 @@ def predict_proba_clique_graph_from_node_prefixes( node_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, node_type_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, edge_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, - return_predictions_dataframe: bool = False - ) -> np.ndarray: + return_predictions_dataframe: bool = False, + return_node_names: bool = True + ) -> Union[np.ndarray, pd.DataFrame]: """Execute predictions probabilities on the provided graph bipartite portion. Parameters @@ -1085,6 +1187,10 @@ def predict_proba_clique_graph_from_node_prefixes( return_predictions_dataframe: bool = False Whether to return a pandas DataFrame, which as indices has the node IDs. By default, a numpy array with the predictions is returned as it weights much less. + return_node_names: bool = True + Whether to return node names when returning the prediction DataFrame. + This value is ignored when the values to be returned the user has not + requested for a prediction dataframe to be returned. """ return self.predict_proba( graph.build_clique_graph_from_node_prefixes( @@ -1095,26 +1201,28 @@ def predict_proba_clique_graph_from_node_prefixes( node_features=node_features, node_type_features=node_type_features, edge_features=edge_features, - return_predictions_dataframe=return_predictions_dataframe + return_predictions_dataframe=return_predictions_dataframe, + return_node_names=return_node_names ) - def predict_proba_clique_graph_from_node_types( + def predict_proba_clique_graph_from_node_type_names( self, graph: Graph, - node_types: List[str], + node_type_names: List[str], support: Optional[Graph] = None, node_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, node_type_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, edge_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, - return_predictions_dataframe: bool = False - ) -> np.ndarray: + return_predictions_dataframe: bool = False, + return_node_names: bool = True + ) -> Union[np.ndarray, pd.DataFrame]: """Execute predictions probabilities on the provided graph bipartite portion. Parameters -------------------- graph: Graph The graph from which to extract the edges. - node_types: List[str] + node_type_names: List[str] The node prefixes of the bipartite graph. support: Optional[Graph] = None The graph describiding the topological structure that @@ -1130,17 +1238,22 @@ def predict_proba_clique_graph_from_node_types( return_predictions_dataframe: bool = False Whether to return a pandas DataFrame, which as indices has the node IDs. By default, a numpy array with the predictions is returned as it weights much less. + return_node_names: bool = True + Whether to return node names when returning the prediction DataFrame. + This value is ignored when the values to be returned the user has not + requested for a prediction dataframe to be returned. """ return self.predict_proba( - graph.build_clique_graph_from_node_types( - node_types=node_types, + graph.build_clique_graph_from_node_type_names( + node_type_names=node_type_names, directed=True ), support=support, node_features=node_features, node_type_features=node_type_features, edge_features=edge_features, - return_predictions_dataframe=return_predictions_dataframe + return_predictions_dataframe=return_predictions_dataframe, + return_node_names=return_node_names ) def fit( diff --git a/embiggen/edge_prediction/edge_prediction_sklearn/sklearn_edge_prediction_adapter.py b/embiggen/edge_prediction/edge_prediction_sklearn/sklearn_edge_prediction_adapter.py index 0ae0d0ea..0bf07d27 100644 --- a/embiggen/edge_prediction/edge_prediction_sklearn/sklearn_edge_prediction_adapter.py +++ b/embiggen/edge_prediction/edge_prediction_sklearn/sklearn_edge_prediction_adapter.py @@ -4,6 +4,7 @@ from typing import Type, List, Optional, Dict, Any, Union import numpy as np import math +import compress_pickle import copy from ensmallen import Graph from embiggen.sequences.generic_sequences import EdgePredictionSequence @@ -373,3 +374,23 @@ def can_use_edge_types(cls) -> bool: """Returns whether the model can optionally use edge types.""" return False + @classmethod + def load(cls, path: str) -> "Self": + """Load a saved version of the model from the provided path. + + Parameters + ------------------- + path: str + Path from where to load the model. + """ + return compress_pickle.load(path) + + def dump(self, path: str): + """Dump the current model at the provided path. + + Parameters + ------------------- + path: str + Path from where to dump the model. + """ + compress_pickle.dump(self, path) \ No newline at end of file diff --git a/embiggen/embedders/ensmallen_embedders/deepwalk_cbow.py b/embiggen/embedders/ensmallen_embedders/deepwalk_cbow.py index 44d72145..99d4c46b 100644 --- a/embiggen/embedders/ensmallen_embedders/deepwalk_cbow.py +++ b/embiggen/embedders/ensmallen_embedders/deepwalk_cbow.py @@ -25,6 +25,8 @@ def __init__( normalize_learning_rate_by_degree: Optional[bool] = False, use_scale_free_distribution: Optional[bool] = True, random_state: int = 42, + dtype: str = "f32", + ring_bell: bool = False, enable_cache: bool = False ): """Create new abstract DeepWalk method. @@ -73,8 +75,12 @@ def __init__( Divide the learning rate by the degree of the central node. By default false. use_scale_free_distribution: Optional[bool] = True Sample negatives proportionally to their degree. By default true. + dtype: str = "f32" + The data type to be employed, by default f32. random_state: int = 42 The random state to reproduce the training sequence. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -96,7 +102,9 @@ def __init__( stochastic_downsample_by_degree=stochastic_downsample_by_degree, normalize_learning_rate_by_degree=normalize_learning_rate_by_degree, use_scale_free_distribution=use_scale_free_distribution, + dtype=dtype, random_state=random_state, + ring_bell=ring_bell, enable_cache=enable_cache ) diff --git a/embiggen/embedders/ensmallen_embedders/deepwalk_glove.py b/embiggen/embedders/ensmallen_embedders/deepwalk_glove.py index 715bfbd1..f55e6e4f 100644 --- a/embiggen/embedders/ensmallen_embedders/deepwalk_glove.py +++ b/embiggen/embedders/ensmallen_embedders/deepwalk_glove.py @@ -9,18 +9,18 @@ def __init__( self, embedding_size: int = 100, alpha: float = 0.75, - epochs: int = 30, - clipping_value: float = 6.0, - walk_length: int = 128, - iterations: int = 10, + epochs: int = 100, + walk_length: int = 512, window_size: int = 5, max_neighbours: Optional[int] = 100, - learning_rate: float = 0.001, - learning_rate_decay: float = 0.9, + learning_rate: float = 0.05, + learning_rate_decay: float = 0.99, central_nodes_embedding_path: Optional[str] = None, contextual_nodes_embedding_path: Optional[str] = None, normalize_by_degree: bool = False, + dtype: str = "f32", random_state: int = 42, + ring_bell: bool = False, enable_cache: bool = False ): """Create new abstract DeepWalk method. @@ -36,13 +36,8 @@ def __init__( window_size: int = 10 Window size for the local context. On the borders the window size is trimmed. - clipping_value: float = 6.0 - Value at which we clip the dot product, mostly for numerical stability issues. - By default, `6.0`, where the loss is already close to zero. - walk_length: int = 128 + walk_length: int = 512 Maximal length of the walks. - iterations: int = 10 - Number of iterations of the single walks. window_size: int = 5 Window size for the local context. On the borders the window size is trimmed. @@ -50,7 +45,7 @@ def __init__( Number of maximum neighbours to consider when using approximated walks. By default, None, we execute exact random walks. This is mainly useful for graphs containing nodes with high degrees. - learning_rate: float = 0.001 + learning_rate: float = 0.05 The learning rate to use to train the DeepWalk model. By default 0.01. central_nodes_embedding_path: Optional[str] = None Path where to mmap and store the central nodes embedding. @@ -60,13 +55,17 @@ def __init__( Path where to mmap and store the central nodes embedding. This is necessary to embed large graphs whose embedding will not fit into the available main memory. - learning_rate_decay: float = 0.9 + learning_rate_decay: float = 0.99 Factor to reduce the learning rate for at each epoch. By default 0.9. normalize_by_degree: bool = False Whether to normalize the random walk by the node degree of the destination node degrees. + dtype: str = "f32" + The data type to be employed, by default f32. random_state: int = 42 The random state to reproduce the training sequence. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -75,9 +74,8 @@ def __init__( embedding_size=embedding_size, alpha=alpha, epochs=epochs, - clipping_value=clipping_value, walk_length=walk_length, - iterations=iterations, + iterations=1, window_size=window_size, max_neighbours=max_neighbours, learning_rate=learning_rate, @@ -86,6 +84,8 @@ def __init__( contextual_nodes_embedding_path=contextual_nodes_embedding_path, normalize_by_degree=normalize_by_degree, enable_cache=enable_cache, + dtype=dtype, + ring_bell=ring_bell, random_state=random_state ) @@ -96,7 +96,8 @@ def parameters(self) -> Dict[str, Any]: "explore_weight", "change_node_type_weight", "change_edge_type_weight", - "number_of_negative_samples" + "number_of_negative_samples", + "iterations" ] return dict( { diff --git a/embiggen/embedders/ensmallen_embedders/deepwalk_skipgram.py b/embiggen/embedders/ensmallen_embedders/deepwalk_skipgram.py index 6d160ea6..b00342ec 100644 --- a/embiggen/embedders/ensmallen_embedders/deepwalk_skipgram.py +++ b/embiggen/embedders/ensmallen_embedders/deepwalk_skipgram.py @@ -25,6 +25,8 @@ def __init__( normalize_learning_rate_by_degree: Optional[bool] = False, use_scale_free_distribution: Optional[bool] = True, random_state: int = 42, + dtype: str = "f32", + ring_bell: bool = False, enable_cache: bool = False ): """Create new abstract DeepWalk method. @@ -73,8 +75,12 @@ def __init__( Divide the learning rate by the degree of the central node. By default false. use_scale_free_distribution: Optional[bool] = True Sample negatives proportionally to their degree. By default true. + dtype: str = "f32" + The data type to be employed, by default f32. random_state: int = 42 The random state to reproduce the training sequence. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -96,7 +102,9 @@ def __init__( stochastic_downsample_by_degree=stochastic_downsample_by_degree, normalize_learning_rate_by_degree=normalize_learning_rate_by_degree, use_scale_free_distribution=use_scale_free_distribution, + dtype=dtype, random_state=random_state, + ring_bell=ring_bell, enable_cache=enable_cache ) diff --git a/embiggen/embedders/ensmallen_embedders/degree_spine.py b/embiggen/embedders/ensmallen_embedders/degree_spine.py index e8bc7995..cd86b344 100644 --- a/embiggen/embedders/ensmallen_embedders/degree_spine.py +++ b/embiggen/embedders/ensmallen_embedders/degree_spine.py @@ -17,6 +17,7 @@ def __init__( maximum_depth: Optional[int] = None, path: Optional[str] = None, verbose: bool = False, + ring_bell: bool = False, enable_cache: bool = False ): """Create new Degree-based SPINE method. @@ -34,6 +35,8 @@ def __init__( This parameter is necessary to embed very large graphs. verbose: bool = False Whether to show loading bars. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -51,19 +54,21 @@ def __init__( super().__init__( embedding_size=embedding_size, + ring_bell=ring_bell, enable_cache=enable_cache ) def parameters(self) -> Dict[str, Any]: """Returns parameters of the model.""" - return { + return dict( super().parameters(), dict( dtype=self._dtype, maximum_depth=self._maximum_depth, path=self._path, ) - } + ) + @classmethod def smoke_test_parameters(cls) -> Dict[str, Any]: """Returns parameters for smoke test.""" diff --git a/embiggen/embedders/ensmallen_embedders/degree_wine.py b/embiggen/embedders/ensmallen_embedders/degree_wine.py index 5b848a16..28b03be0 100644 --- a/embiggen/embedders/ensmallen_embedders/degree_wine.py +++ b/embiggen/embedders/ensmallen_embedders/degree_wine.py @@ -14,9 +14,10 @@ def __init__( self, embedding_size: int = 100, dtype: Optional[str] = "u8", - walk_length: int = 2, + window_size: int = 2, path: Optional[str] = None, verbose: bool = False, + ring_bell: bool = False, enable_cache: bool = False ): """Create new Degree-based WINE method. @@ -27,44 +28,49 @@ def __init__( Dimension of the embedding. dtype: Optional[str] = "u8" Dtype to use for the embedding. - walk_length: int = 2 - Length of the random walk. + window_size: int = 2 + Size of the co-occurrence window. + Do note that for `window_size = 2` we will use the Two-Hop WINE version, which is more efficient. By default 2, to capture exclusively the immediate context. path: Optional[str] = None Path where to store the mmap-ed embedding. This parameter is necessary to embed very large graphs. verbose: bool = False Whether to show loading bars. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. """ self._dtype = dtype self._verbose = verbose - self._walk_length = walk_length + self._window_size = window_size self._path = path self._model = models.DegreeWINE( embedding_size=embedding_size, verbose=self._verbose, - walk_length=self._walk_length, + window_size=self._window_size, path=self._path ) super().__init__( embedding_size=embedding_size, + ring_bell=ring_bell, enable_cache=enable_cache ) def parameters(self) -> Dict[str, Any]: """Returns parameters of the model.""" - return { + return dict( super().parameters(), dict( dtype=self._dtype, - walk_length=self._walk_length, + window_size=self._window_size, path=self._path, ) - } + ) + @classmethod def smoke_test_parameters(cls) -> Dict[str, Any]: """Returns parameters for smoke test.""" diff --git a/embiggen/embedders/ensmallen_embedders/ensmallen_embedder.py b/embiggen/embedders/ensmallen_embedders/ensmallen_embedder.py index 7232edeb..213c8fb7 100644 --- a/embiggen/embedders/ensmallen_embedders/ensmallen_embedder.py +++ b/embiggen/embedders/ensmallen_embedders/ensmallen_embedder.py @@ -13,6 +13,7 @@ def __init__( self, random_state: Optional[int] = None, embedding_size: int = 100, + ring_bell: bool = False, enable_cache: bool = False ): @@ -24,6 +25,8 @@ def __init__( Random state to reproduce the embeddings. embedding_size: int = 100 Dimension of the embedding. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -31,6 +34,7 @@ def __init__( super().__init__( random_state=random_state, embedding_size=embedding_size, + ring_bell=ring_bell, enable_cache=enable_cache, ) diff --git a/embiggen/embedders/ensmallen_embedders/first_order_line.py b/embiggen/embedders/ensmallen_embedders/first_order_line.py index e5bd91b8..fbafe403 100644 --- a/embiggen/embedders/ensmallen_embedders/first_order_line.py +++ b/embiggen/embedders/ensmallen_embedders/first_order_line.py @@ -17,9 +17,12 @@ def __init__( learning_rate: float = 0.05, learning_rate_decay: float = 0.9, avoid_false_negatives: bool = False, + use_scale_free_distribution: bool = True, node_embedding_path: Optional[str] = None, + dtype: str = "f32", random_state: int = 42, verbose: bool = False, + ring_bell: bool = False, enable_cache: bool = False ): """Create new abstract Node2Vec method. @@ -31,20 +34,26 @@ def __init__( epochs: int = 100 The number of epochs to run the model for, by default 10. learning_rate: float = 0.05 - The learning rate to update the gradient, by default 0.01. + The learning rate to update the gradient, by default 0.05. learning_rate_decay: float = 0.9 Factor to reduce the learning rate for at each epoch. By default 0.9. avoid_false_negatives: bool = False Whether to avoid sampling false negatives. This may cause a slower training. + use_scale_free_distribution: bool = True + Whether to train model using a scale free distribution for the negatives. node_embedding_path: Optional[str] = None Path where to mmap and store the nodes embedding. This is necessary to embed large graphs whose embedding will not fit into the available main memory. + dtype: str = "f32" + The data type to be employed, by default f32. random_state: int = 42 Random state to reproduce the embeddings. verbose: bool = False Whether to show loading bars. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -54,7 +63,9 @@ def __init__( learning_rate=learning_rate, learning_rate_decay=learning_rate_decay, avoid_false_negatives=avoid_false_negatives, + use_scale_free_distribution=use_scale_free_distribution, node_embedding_path=node_embedding_path, + dtype=dtype, verbose=verbose ) @@ -67,6 +78,7 @@ def __init__( super().__init__( embedding_size=embedding_size, enable_cache=enable_cache, + ring_bell=ring_bell, random_state=random_state ) diff --git a/embiggen/embedders/ensmallen_embedders/geometric_laplacian_eigenmaps.py b/embiggen/embedders/ensmallen_embedders/geometric_laplacian_eigenmaps.py index 7862ec90..0b36f344 100644 --- a/embiggen/embedders/ensmallen_embedders/geometric_laplacian_eigenmaps.py +++ b/embiggen/embedders/ensmallen_embedders/geometric_laplacian_eigenmaps.py @@ -15,6 +15,7 @@ class GLEEEnsmallen(EnsmallenEmbedder): def __init__( self, embedding_size: int = 100, + ring_bell: bool = False, enable_cache: bool = False ): """Create new GLEE method. @@ -23,12 +24,15 @@ def __init__( -------------------------- embedding_size: int = 100 Dimension of the embedding. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. """ super().__init__( embedding_size=embedding_size, + ring_bell=ring_bell, enable_cache=enable_cache, ) diff --git a/embiggen/embedders/ensmallen_embedders/hope.py b/embiggen/embedders/ensmallen_embedders/hope.py index ed85a408..38aa8dbc 100644 --- a/embiggen/embedders/ensmallen_embedders/hope.py +++ b/embiggen/embedders/ensmallen_embedders/hope.py @@ -20,6 +20,7 @@ def __init__( metric: str = "Neighbours Intersection size", root_node_name: Optional[str] = None, verbose: bool = False, + ring_bell: bool = False, enable_cache: bool = False ): """Create new HOPE method. @@ -47,6 +48,8 @@ def __init__( the Jaccard index is selected. verbose: bool = False Whether to show loading bars. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -71,6 +74,7 @@ def __init__( super().__init__( embedding_size=embedding_size, + ring_bell=ring_bell, enable_cache=enable_cache ) diff --git a/embiggen/embedders/ensmallen_embedders/laplacian_eigenmaps.py b/embiggen/embedders/ensmallen_embedders/laplacian_eigenmaps.py index 899170b3..416923e4 100644 --- a/embiggen/embedders/ensmallen_embedders/laplacian_eigenmaps.py +++ b/embiggen/embedders/ensmallen_embedders/laplacian_eigenmaps.py @@ -14,6 +14,7 @@ class LaplacianEigenmapsEnsmallen(EnsmallenEmbedder): def __init__( self, embedding_size: int = 100, + ring_bell: bool = False, enable_cache: bool = False ): """Create new Laplacian Eigenmaps method. @@ -22,12 +23,15 @@ def __init__( -------------------------- embedding_size: int = 100 Dimension of the embedding. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. """ super().__init__( embedding_size=embedding_size, + ring_bell=ring_bell, enable_cache=enable_cache, ) diff --git a/embiggen/embedders/ensmallen_embedders/netmf.py b/embiggen/embedders/ensmallen_embedders/netmf.py deleted file mode 100644 index af658a33..00000000 --- a/embiggen/embedders/ensmallen_embedders/netmf.py +++ /dev/null @@ -1,139 +0,0 @@ -"""Module providing NetMF implementation.""" -from typing import Optional, Dict, Any -from ensmallen import Graph -import pandas as pd -import numpy as np -from scipy.sparse import coo_matrix -from sklearn.decomposition import TruncatedSVD -from embiggen.embedders.ensmallen_embedders.ensmallen_embedder import EnsmallenEmbedder -from embiggen.utils import EmbeddingResult - - -class NetMFEnsmallen(EnsmallenEmbedder): - """Class implementing the NetMF algorithm.""" - - def __init__( - self, - embedding_size: int = 100, - walk_length: int = 128, - iterations: int = 10, - window_size: int = 10, - max_neighbours: Optional[int] = 100, - random_state: int = 42, - enable_cache: bool = False - ): - """Create new NetMF method. - - Parameters - -------------------------- - embedding_size: int = 100 - Dimension of the embedding. - walk_length: int = 128 - Maximal length of the walks. - iterations: int = 10 - Number of iterations of the single walks. - window_size: int = 10 - Window size for the local context. - On the borders the window size is trimmed. - max_neighbours: Optional[int] = 100 - Number of maximum neighbours to consider when using approximated walks. - By default, None, we execute exact random walks. - This is mainly useful for graphs containing nodes with high degrees. - random_state: int = 42 - The random state to reproduce the training sequence. - enable_cache: bool = False - Whether to enable the cache, that is to - store the computed embedding. - """ - self._walk_parameters = dict( - walk_length=walk_length, - iterations=iterations, - window_size=window_size, - max_neighbours=max_neighbours, - ) - - super().__init__( - embedding_size=embedding_size, - enable_cache=enable_cache, - random_state=random_state - ) - - @classmethod - def smoke_test_parameters(cls) -> Dict[str, Any]: - """Returns parameters for smoke test.""" - return dict( - EnsmallenEmbedder.smoke_test_parameters(), - window_size=1, - walk_length=4, - iterations=1, - max_neighbours=10, - ) - - def parameters(self) -> Dict[str, Any]: - """Returns parameters of the model.""" - return dict( - super().parameters(), - self._walk_parameters - ) - - def _fit_transform( - self, - graph: Graph, - return_dataframe: bool = True, - ) -> EmbeddingResult: - """Return node embedding.""" - edges, weights = graph.get_log_normalized_cooccurrence_coo_matrix( - self._walk_parameters - ) - - coo = coo_matrix( - (weights, (edges[:, 0], edges[:, 1])), - shape=( - graph.get_number_of_nodes(), - graph.get_number_of_nodes() - ), - dtype=np.float32 - ) - - model = TruncatedSVD( - n_components=self._embedding_size, - random_state=self._random_state - ) - model.fit(coo) - embedding = model.transform(coo) - - if return_dataframe: - node_names = graph.get_node_names() - embedding = pd.DataFrame( - embedding, - index=node_names - ) - return EmbeddingResult( - embedding_method_name=self.model_name(), - node_embeddings=embedding - ) - - @classmethod - def model_name(cls) -> str: - """Returns name of the model.""" - return "NetMF" - - @classmethod - def can_use_edge_weights(cls) -> bool: - """Returns whether the model can optionally use edge weights.""" - return False - - @classmethod - def can_use_node_types(cls) -> bool: - """Returns whether the model can optionally use node types.""" - return False - - @classmethod - def can_use_edge_types(cls) -> bool: - """Returns whether the model can optionally use edge types.""" - return False - - @classmethod - def is_stocastic(cls) -> bool: - """Returns whether the model is stocastic and has therefore a random state.""" - return True diff --git a/embiggen/embedders/ensmallen_embedders/node2vec.py b/embiggen/embedders/ensmallen_embedders/node2vec.py index 31b64caa..05094f59 100644 --- a/embiggen/embedders/ensmallen_embedders/node2vec.py +++ b/embiggen/embedders/ensmallen_embedders/node2vec.py @@ -29,6 +29,7 @@ def __init__( self, embedding_size: int = 100, random_state: int = 42, + ring_bell: bool = False, enable_cache: bool = False, model_kwargs: Dict ): @@ -40,6 +41,8 @@ def __init__( Dimension of the embedding. random_state: int = 42 The random state to reproduce the training sequence. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -53,6 +56,7 @@ def __init__( super().__init__( embedding_size=embedding_size, enable_cache=enable_cache, + ring_bell=ring_bell, random_state=random_state ) @@ -64,7 +68,6 @@ def smoke_test_parameters(cls) -> Dict[str, Any]: epochs=1, window_size=1, walk_length=4, - iterations=1, max_neighbours=10, ) diff --git a/embiggen/embedders/ensmallen_embedders/node2vec_cbow.py b/embiggen/embedders/ensmallen_embedders/node2vec_cbow.py index 4fc3203b..04961df9 100644 --- a/embiggen/embedders/ensmallen_embedders/node2vec_cbow.py +++ b/embiggen/embedders/ensmallen_embedders/node2vec_cbow.py @@ -27,6 +27,8 @@ def __init__( normalize_learning_rate_by_degree: Optional[bool] = False, use_scale_free_distribution: Optional[bool] = True, random_state: int = 42, + dtype: str = "f32", + ring_bell: bool = False, enable_cache: bool = False ): """Create new abstract Node2Vec method. @@ -89,8 +91,12 @@ def __init__( Divide the learning rate by the degree of the central node. By default false. use_scale_free_distribution: Optional[bool] = True Sample negatives proportionally to their degree. By default true. + dtype: str = "f32" + The data type to be employed, by default f32. random_state: int = 42 The random state to reproduce the training sequence. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -114,7 +120,9 @@ def __init__( stochastic_downsample_by_degree=stochastic_downsample_by_degree, normalize_learning_rate_by_degree=normalize_learning_rate_by_degree, use_scale_free_distribution=use_scale_free_distribution, + dtype=dtype, random_state=random_state, + ring_bell=ring_bell, enable_cache=enable_cache ) diff --git a/embiggen/embedders/ensmallen_embedders/node2vec_glove.py b/embiggen/embedders/ensmallen_embedders/node2vec_glove.py index 806cb8c6..a80772b7 100644 --- a/embiggen/embedders/ensmallen_embedders/node2vec_glove.py +++ b/embiggen/embedders/ensmallen_embedders/node2vec_glove.py @@ -9,20 +9,20 @@ def __init__( self, embedding_size: int = 100, alpha: float = 0.75, - epochs: int = 30, - clipping_value: float = 6.0, - walk_length: int = 128, - iterations: int = 10, + epochs: int = 100, + walk_length: int = 512, window_size: int = 5, return_weight: float = 0.25, explore_weight: float = 4.0, max_neighbours: Optional[int] = 100, - learning_rate: float = 0.001, + learning_rate: float = 0.05, learning_rate_decay: float = 0.9, central_nodes_embedding_path: Optional[str] = None, contextual_nodes_embedding_path: Optional[str] = None, normalize_by_degree: bool = False, + dtype: str = "f32", random_state: int = 42, + ring_bell: bool = False, enable_cache: bool = False ): """Create new Node2Vec GloVe model. @@ -38,13 +38,8 @@ def __init__( window_size: int = 10 Window size for the local context. On the borders the window size is trimmed. - clipping_value: float = 6.0 - Value at which we clip the dot product, mostly for numerical stability issues. - By default, `6.0`, where the loss is already close to zero. - walk_length: int = 128 + walk_length: int = 512 Maximal length of the walks. - iterations: int = 10 - Number of iterations of the single walks. window_size: int = 5 Window size for the local context. On the borders the window size is trimmed. @@ -81,8 +76,12 @@ def __init__( normalize_by_degree: bool = False Whether to normalize the random walk by the node degree of the destination node degrees. + dtype: str = "f32" + The data type to be employed, by default f32. random_state: int = 42 The random state to reproduce the training sequence. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -91,9 +90,8 @@ def __init__( embedding_size=embedding_size, alpha=alpha, epochs=epochs, - clipping_value=clipping_value, walk_length=walk_length, - iterations=iterations, + iterations=1, window_size=window_size, return_weight=return_weight, explore_weight=explore_weight, @@ -103,7 +101,9 @@ def __init__( central_nodes_embedding_path=central_nodes_embedding_path, contextual_nodes_embedding_path=contextual_nodes_embedding_path, normalize_by_degree=normalize_by_degree, + dtype=dtype, enable_cache=enable_cache, + ring_bell=ring_bell, random_state=random_state ) @@ -112,7 +112,8 @@ def parameters(self) -> Dict[str, Any]: removed = [ "change_node_type_weight", "change_edge_type_weight", - "number_of_negative_samples" + "number_of_negative_samples", + "iterations" ] return dict( { diff --git a/embiggen/embedders/ensmallen_embedders/node2vec_skipgram.py b/embiggen/embedders/ensmallen_embedders/node2vec_skipgram.py index b3a1f50e..2dd871e6 100644 --- a/embiggen/embedders/ensmallen_embedders/node2vec_skipgram.py +++ b/embiggen/embedders/ensmallen_embedders/node2vec_skipgram.py @@ -27,6 +27,8 @@ def __init__( normalize_learning_rate_by_degree: Optional[bool] = False, use_scale_free_distribution: Optional[bool] = True, random_state: int = 42, + dtype: str = "f32", + ring_bell: bool = False, enable_cache: bool = False ): """Create new abstract Node2Vec method. @@ -91,8 +93,12 @@ def __init__( Divide the learning rate by the degree of the central node. By default false. use_scale_free_distribution: Optional[bool] = True Sample negatives proportionally to their degree. By default true. + dtype: str = "f32" + The data type to be employed, by default f32. random_state: int = 42 The random state to reproduce the training sequence. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -116,7 +122,9 @@ def __init__( stochastic_downsample_by_degree=stochastic_downsample_by_degree, normalize_learning_rate_by_degree=normalize_learning_rate_by_degree, use_scale_free_distribution=use_scale_free_distribution, + dtype=dtype, random_state=random_state, + ring_bell=ring_bell, enable_cache=enable_cache ) diff --git a/embiggen/embedders/ensmallen_embedders/node_label_spine.py b/embiggen/embedders/ensmallen_embedders/node_label_spine.py index 6bcc7230..ec14e85c 100644 --- a/embiggen/embedders/ensmallen_embedders/node_label_spine.py +++ b/embiggen/embedders/ensmallen_embedders/node_label_spine.py @@ -16,6 +16,7 @@ def __init__( maximum_depth: Optional[int] = None, path: Optional[str] = None, verbose: bool = False, + ring_bell: bool = False, enable_cache: bool = False ): """Create new Node-label-based SPINE method. @@ -31,6 +32,8 @@ def __init__( This parameter is necessary to embed very large graphs. verbose: bool = False Whether to show loading bars. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -46,6 +49,7 @@ def __init__( ) super().__init__( + ring_bell=ring_bell, enable_cache=enable_cache ) diff --git a/embiggen/embedders/ensmallen_embedders/node_label_wine.py b/embiggen/embedders/ensmallen_embedders/node_label_wine.py index 109ac40c..4401c1ed 100644 --- a/embiggen/embedders/ensmallen_embedders/node_label_wine.py +++ b/embiggen/embedders/ensmallen_embedders/node_label_wine.py @@ -13,9 +13,10 @@ class NodeLabelWINE(EnsmallenEmbedder): def __init__( self, dtype: Optional[str] = "u8", - walk_length: Optional[int] = None, + window_size: int = 2, path: Optional[str] = None, verbose: bool = False, + ring_bell: bool = False, enable_cache: bool = False ): """Create new Node-label-based WINE method. @@ -24,29 +25,33 @@ def __init__( -------------------------- dtype: Optional[str] = "u8" Dtype to use for the embedding. - walk_length: int = 2 - Length of the random walk. + window_size: int = 2 + Size of the co-occurrence window. + Do note that for `window_size = 2` we will use the Two-Hop WINE version, which is more efficient. By default 2, to capture exclusively the immediate context. path: Optional[str] = None Path where to store the mmap-ed embedding. This parameter is necessary to embed very large graphs. verbose: bool = False Whether to show loading bars. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. """ self._dtype = dtype self._verbose = verbose - self._walk_length = walk_length + self._window_size = window_size self._path = path self._model = models.NodeLabelWINE( verbose=self._verbose, - walk_length=self._walk_length, + window_size=self._window_size, path=self._path ) super().__init__( + ring_bell=ring_bell, enable_cache=enable_cache ) @@ -60,7 +65,7 @@ def parameters(self) -> Dict[str, Any]: }, dict( dtype=self._dtype, - walk_length=self._walk_length, + window_size=self._window_size, path=self._path, ) ) diff --git a/embiggen/embedders/ensmallen_embedders/score_spine.py b/embiggen/embedders/ensmallen_embedders/score_spine.py index cf018f30..dd91854c 100644 --- a/embiggen/embedders/ensmallen_embedders/score_spine.py +++ b/embiggen/embedders/ensmallen_embedders/score_spine.py @@ -19,6 +19,7 @@ def __init__( maximum_depth: Optional[int] = None, path: Optional[str] = None, verbose: bool = False, + ring_bell: bool = False, enable_cache: bool = False ): """Create new Score-based SPINE method. @@ -38,6 +39,8 @@ def __init__( This parameter is necessary to embed very large graphs. verbose: bool = False Whether to show loading bars. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -56,19 +59,20 @@ def __init__( super().__init__( embedding_size=embedding_size, + ring_bell=ring_bell, enable_cache=enable_cache ) def parameters(self) -> Dict[str, Any]: """Returns parameters of the model.""" - return { + return dict( super().parameters(), dict( dtype=self._dtype, maximum_depth=self._maximum_depth, path=self._path, ) - } + ) @classmethod def smoke_test_parameters(cls) -> Dict[str, Any]: diff --git a/embiggen/embedders/ensmallen_embedders/score_wine.py b/embiggen/embedders/ensmallen_embedders/score_wine.py index 35f9b2a9..ce6af544 100644 --- a/embiggen/embedders/ensmallen_embedders/score_wine.py +++ b/embiggen/embedders/ensmallen_embedders/score_wine.py @@ -16,9 +16,10 @@ def __init__( scores: Optional[np.ndarray] = None, embedding_size: int = 100, dtype: Optional[str] = "u8", - walk_length: Optional[int] = None, + window_size: int = 2, path: Optional[str] = None, verbose: bool = False, + ring_bell: bool = False, enable_cache: bool = False ): """Create new Score-based WINE method. @@ -31,45 +32,49 @@ def __init__( Dimension of the embedding. dtype: Optional[str] = "u8" Dtype to use for the embedding. - walk_length: int = 2 - Length of the random walk. + window_size: int = 2 + Size of the co-occurrence window. + Do note that for `window_size = 2` we will use the Two-Hop WINE version, which is more efficient. By default 2, to capture exclusively the immediate context. path: Optional[str] = None Path where to store the mmap-ed embedding. This parameter is necessary to embed very large graphs. verbose: bool = False Whether to show loading bars. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. """ self._dtype = dtype self._verbose = verbose - self._walk_length = walk_length + self._window_size = window_size self._path = path self._scores = scores self._model = models.ScoreWINE( embedding_size=embedding_size, verbose=self._verbose, - walk_length=self._walk_length, + window_size=self._window_size, path=self._path ) super().__init__( embedding_size=embedding_size, + ring_bell=ring_bell, enable_cache=enable_cache ) def parameters(self) -> Dict[str, Any]: """Returns parameters of the model.""" - return { + return dict( super().parameters(), dict( dtype=self._dtype, - walk_length=self._walk_length, + window_size=self._window_size, path=self._path, ) - } + ) @classmethod def smoke_test_parameters(cls) -> Dict[str, Any]: diff --git a/embiggen/embedders/ensmallen_embedders/second_order_line.py b/embiggen/embedders/ensmallen_embedders/second_order_line.py index 19858731..439cffd2 100644 --- a/embiggen/embedders/ensmallen_embedders/second_order_line.py +++ b/embiggen/embedders/ensmallen_embedders/second_order_line.py @@ -15,13 +15,16 @@ def __init__( self, embedding_size: int = 100, epochs: int = 100, - learning_rate: float = 0.01, + learning_rate: float = 0.05, learning_rate_decay: float = 0.9, avoid_false_negatives: bool = False, + use_scale_free_distribution: bool = True, node_embedding_path: Optional[str] = None, contextual_node_embedding_path: Optional[str] = None, + dtype: str = "f32", random_state: int = 42, verbose: bool = False, + ring_bell: bool = False, enable_cache: bool = False ): """Create new abstract Node2Vec method. @@ -32,13 +35,15 @@ def __init__( Dimension of the embedding. epochs: int = 100 The number of epochs to run the model for, by default 10. - learning_rate: float = 0.01 - The learning rate to update the gradient, by default 0.01. + learning_rate: float = 0.05 + The learning rate to update the gradient, by default 0.05. learning_rate_decay: float = 0.9 Factor to reduce the learning rate for at each epoch. By default 0.9. avoid_false_negatives: bool = False Whether to avoid sampling false negatives. This may cause a slower training. + use_scale_free_distribution: bool = True + Whether to train model using a scale free distribution for the negatives. node_embedding_path: Optional[str] = None Path where to mmap and store the nodes embedding. This is necessary to embed large graphs whose embedding will not @@ -47,10 +52,14 @@ def __init__( Path where to mmap and store the contextual nodes embedding. This is necessary to embed large graphs whose embedding will not fit into the available main memory. + dtype: str = "f32" + The data type to be employed, by default f32. random_state: int = 42 Random state to reproduce the embeddings. verbose: bool = False Whether to show loading bars. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -60,8 +69,10 @@ def __init__( learning_rate=learning_rate, learning_rate_decay=learning_rate_decay, avoid_false_negatives=avoid_false_negatives, + use_scale_free_distribution=use_scale_free_distribution, node_embedding_path=node_embedding_path, contextual_node_embedding_path=contextual_node_embedding_path, + dtype=dtype, verbose=verbose ) @@ -74,6 +85,7 @@ def __init__( super().__init__( embedding_size=embedding_size, enable_cache=enable_cache, + ring_bell=ring_bell, random_state=random_state ) diff --git a/embiggen/embedders/ensmallen_embedders/siamese_model.py b/embiggen/embedders/ensmallen_embedders/siamese_model.py index 72615225..0a614d6d 100644 --- a/embiggen/embedders/ensmallen_embedders/siamese_model.py +++ b/embiggen/embedders/ensmallen_embedders/siamese_model.py @@ -12,7 +12,6 @@ class SiameseEnsmallen(EnsmallenEmbedder): models = { "TransE": models.TransE, - "TransH": models.TransH, "Unstructured": models.Unstructured, "Structured Embedding": models.StructuredEmbedding, } @@ -25,8 +24,10 @@ def __init__( learning_rate: float = 0.1, learning_rate_decay: float = 0.9, node_embedding_path: Optional[str] = None, + dtype: str = "f32", random_state: int = 42, verbose: bool = False, + ring_bell: bool = False, enable_cache: bool = False, paths: Dict[str, str] ): @@ -49,10 +50,14 @@ def __init__( Path where to mmap and store the nodes embedding. This is necessary to embed large graphs whose embedding will not fit into the available main memory. + dtype: str = "f32" + The data type to be employed, by default f32. random_state: int = 42 Random state to reproduce the embeddings. verbose: bool = False Whether to show loading bars. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -64,6 +69,7 @@ def __init__( learning_rate_decay=learning_rate_decay, node_embedding_path=node_embedding_path, verbose=verbose, + dtype=dtype, paths ) @@ -82,6 +88,7 @@ def __init__( super().__init__( embedding_size=embedding_size, enable_cache=enable_cache, + ring_bell=ring_bell, random_state=random_state ) diff --git a/embiggen/embedders/ensmallen_embedders/sociodim.py b/embiggen/embedders/ensmallen_embedders/sociodim.py index 330153c5..1fd3a5cc 100644 --- a/embiggen/embedders/ensmallen_embedders/sociodim.py +++ b/embiggen/embedders/ensmallen_embedders/sociodim.py @@ -15,6 +15,7 @@ def __init__( self, embedding_size: int = 100, use_sparse_reduce: bool = True, + ring_bell: bool = False, enable_cache: bool = False ): """Create new SocioDim method. @@ -29,6 +30,8 @@ def __init__( For both reduce mechanisms, we are using LAPACK implementations. For some currently unknown reason, their implementation using a sparse reduce, even on dense matrices, yields better results. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -36,6 +39,7 @@ def __init__( self._use_sparse_reduce = use_sparse_reduce super().__init__( embedding_size=embedding_size, + ring_bell=ring_bell, enable_cache=enable_cache, ) diff --git a/embiggen/embedders/ensmallen_embedders/structured_embedding.py b/embiggen/embedders/ensmallen_embedders/structured_embedding.py index ccf0334b..6c876ace 100644 --- a/embiggen/embedders/ensmallen_embedders/structured_embedding.py +++ b/embiggen/embedders/ensmallen_embedders/structured_embedding.py @@ -19,8 +19,10 @@ def __init__( node_embedding_path: Optional[str] = None, source_edge_type_embedding_path: Optional[str] = None, destination_edge_type_embedding_path: Optional[str] = None, + dtype: str = "f32", random_state: int = 42, verbose: bool = False, + ring_bell: bool = False, enable_cache: bool = False ): """Create new Structured Embedding method. @@ -52,10 +54,14 @@ def __init__( Path where to mmap and store the destination edge type embedding. This is necessary to embed large graphs whose embedding will not fit into the available main memory. + dtype: str = "f32" + The data type to be employed, by default f32. random_state: int = 42 Random state to reproduce the embeddings. verbose: bool = False Whether to show loading bars. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -69,8 +75,10 @@ def __init__( node_embedding_path=node_embedding_path, source_edge_type_embedding_path=source_edge_type_embedding_path, destination_edge_type_embedding_path=destination_edge_type_embedding_path, + dtype=dtype, random_state=random_state, verbose=verbose, + ring_bell=ring_bell, enable_cache=enable_cache, ) diff --git a/embiggen/embedders/ensmallen_embedders/transe.py b/embiggen/embedders/ensmallen_embedders/transe.py index 1a135d14..8736037f 100644 --- a/embiggen/embedders/ensmallen_embedders/transe.py +++ b/embiggen/embedders/ensmallen_embedders/transe.py @@ -18,8 +18,10 @@ def __init__( learning_rate_decay: float = 0.9, node_embedding_path: Optional[str] = None, edge_type_embedding_path: Optional[str] = None, + dtype: str = "f32", random_state: int = 42, verbose: bool = False, + ring_bell: bool = False, enable_cache: bool = False ): """Create new TransE method. @@ -47,10 +49,14 @@ def __init__( Path where to mmap and store the edge type embedding. This is necessary to embed large graphs whose embedding will not fit into the available main memory. + dtype: str = "f32" + The data type to be employed, by default f32. random_state: int = 42 Random state to reproduce the embeddings. verbose: bool = False Whether to show loading bars. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -63,8 +69,10 @@ def __init__( learning_rate_decay=learning_rate_decay, node_embedding_path=node_embedding_path, edge_type_embedding_path=edge_type_embedding_path, + dtype=dtype, random_state=random_state, verbose=verbose, + ring_bell=ring_bell, enable_cache=enable_cache, ) diff --git a/embiggen/embedders/ensmallen_embedders/transh.py b/embiggen/embedders/ensmallen_embedders/transh.py deleted file mode 100644 index 2de65834..00000000 --- a/embiggen/embedders/ensmallen_embedders/transh.py +++ /dev/null @@ -1,116 +0,0 @@ -"""Module providing TransH implementation.""" -from typing import Optional -from ensmallen import Graph -import pandas as pd -from embiggen.embedders.ensmallen_embedders.siamese_model import SiameseEnsmallen -from embiggen.utils import EmbeddingResult - - -class TransHEnsmallen(SiameseEnsmallen): - """Class implementing the TransH algorithm.""" - - def __init__( - self, - embedding_size: int = 100, - relu_bias: float = 1.0, - epochs: int = 100, - learning_rate: float = 0.1, - learning_rate_decay: float = 0.9, - node_embedding_path: Optional[str] = None, - mult_edge_type_embedding_path: Optional[str] = None, - bias_edge_type_embedding_path: Optional[str] = None, - random_state: int = 42, - verbose: bool = False, - enable_cache: bool = False - ): - """Create new TransH method. - - Parameters - -------------------------- - model_name: str - The model to instantiate. - embedding_size: int = 100 - Dimension of the embedding. - relu_bias: float = 1.0 - Bias to use for the relu. - In the TransH paper it is called gamma. - epochs: int = 100 - The number of epochs to run the model for, by default 10. - learning_rate: float = 0.05 - The learning rate to update the gradient, by default 0.01. - learning_rate_decay: float = 0.9 - Factor to reduce the learning rate for at each epoch. By default 0.9. - node_embedding_path: Optional[str] = None - Path where to mmap and store the nodes embedding. - This is necessary to embed large graphs whose embedding will not - fit into the available main memory. - mult_edge_type_embedding_path: Optional[str] = None - Path where to mmap and store the multiplicative edge type embedding. - This is necessary to embed large graphs whose embedding will not - fit into the available main memory. - bias_edge_type_embedding_path: Optional[str] = None - Path where to mmap and store the bias edge type embedding. - This is necessary to embed large graphs whose embedding will not - fit into the available main memory. - random_state: int = 42 - Random state to reproduce the embeddings. - verbose: bool = False - Whether to show loading bars. - enable_cache: bool = False - Whether to enable the cache, that is to - store the computed embedding. - """ - super().__init__( - embedding_size=embedding_size, - relu_bias=relu_bias, - epochs=epochs, - learning_rate=learning_rate, - learning_rate_decay=learning_rate_decay, - node_embedding_path=node_embedding_path, - mult_edge_type_embedding_path=mult_edge_type_embedding_path, - bias_edge_type_embedding_path=bias_edge_type_embedding_path, - random_state=random_state, - verbose=verbose, - enable_cache=enable_cache, - ) - - def _fit_transform( - self, - graph: Graph, - return_dataframe: bool = True, - ) -> EmbeddingResult: - """Return node embedding.""" - node_embedding, mult_edge_type_embedding, bias_edge_type_embedding = self._model.fit_transform( - graph, - ) - if return_dataframe: - node_embedding = pd.DataFrame( - node_embedding, - index=graph.get_node_names() - ) - mult_edge_type_embedding = pd.DataFrame( - mult_edge_type_embedding, - index=graph.get_unique_edge_type_names() - ) - bias_edge_type_embedding = pd.DataFrame( - bias_edge_type_embedding, - index=graph.get_unique_edge_type_names() - ) - - return EmbeddingResult( - embedding_method_name=self.model_name(), - node_embeddings=node_embedding, - edge_type_embeddings=[ - mult_edge_type_embedding, - bias_edge_type_embedding - ] - ) - - @classmethod - def model_name(cls) -> str: - """Returns name of the model.""" - return "TransH" - - @classmethod - def requires_edge_types(cls) -> bool: - return True \ No newline at end of file diff --git a/embiggen/embedders/ensmallen_embedders/unstructured.py b/embiggen/embedders/ensmallen_embedders/unstructured.py index edcaf5f0..dd5b2403 100644 --- a/embiggen/embedders/ensmallen_embedders/unstructured.py +++ b/embiggen/embedders/ensmallen_embedders/unstructured.py @@ -17,8 +17,10 @@ def __init__( learning_rate: float = 0.01, learning_rate_decay: float = 0.9, node_embedding_path: Optional[str] = None, + dtype: str = "f32", random_state: int = 42, verbose: bool = False, + ring_bell: bool = False, enable_cache: bool = False ): """Create new abstract Node2Vec method. @@ -42,10 +44,14 @@ def __init__( Path where to mmap and store the nodes embedding. This is necessary to embed large graphs whose embedding will not fit into the available main memory. + dtype: str = "f32" + The data type to be employed, by default f32. random_state: int = 42 Random state to reproduce the embeddings. verbose: bool = False Whether to show loading bars. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -57,8 +63,10 @@ def __init__( learning_rate=learning_rate, learning_rate_decay=learning_rate_decay, node_embedding_path=node_embedding_path, + dtype=dtype, random_state=random_state, verbose=verbose, + ring_bell=ring_bell, enable_cache=enable_cache, ) diff --git a/embiggen/embedders/ensmallen_embedders/walklets.py b/embiggen/embedders/ensmallen_embedders/walklets.py index 0f3a4ddd..fa9e4c14 100644 --- a/embiggen/embedders/ensmallen_embedders/walklets.py +++ b/embiggen/embedders/ensmallen_embedders/walklets.py @@ -29,6 +29,8 @@ def __init__( normalize_learning_rate_by_degree: Optional[bool] = False, use_scale_free_distribution: Optional[bool] = True, random_state: int = 42, + dtype: str = "f32", + ring_bell: bool = False, enable_cache: bool = False ): """Create new abstract Node2Vec method. @@ -97,8 +99,12 @@ def __init__( Divide the learning rate by the degree of the central node. By default false. use_scale_free_distribution: Optional[bool] = True Sample negatives proportionally to their degree. By default true. + dtype: str = "f32" + The data type to be employed, by default f32. random_state: int = 42 The random state to reproduce the training sequence. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -123,7 +129,9 @@ def __init__( stochastic_downsample_by_degree=stochastic_downsample_by_degree, normalize_learning_rate_by_degree=normalize_learning_rate_by_degree, use_scale_free_distribution=use_scale_free_distribution, + dtype=dtype, random_state=random_state, + ring_bell=ring_bell, enable_cache=enable_cache ) diff --git a/embiggen/embedders/ensmallen_embedders/walklets_cbow.py b/embiggen/embedders/ensmallen_embedders/walklets_cbow.py index 99d73975..066667bf 100644 --- a/embiggen/embedders/ensmallen_embedders/walklets_cbow.py +++ b/embiggen/embedders/ensmallen_embedders/walklets_cbow.py @@ -27,6 +27,8 @@ def __init__( normalize_learning_rate_by_degree: Optional[bool] = False, use_scale_free_distribution: Optional[bool] = True, random_state: int = 42, + dtype: str = "f32", + ring_bell: bool = False, enable_cache: bool = False ): """Create new abstract Node2Vec method. @@ -93,8 +95,12 @@ def __init__( Divide the learning rate by the degree of the central node. By default false. use_scale_free_distribution: Optional[bool] = True Sample negatives proportionally to their degree. By default true. + dtype: str = "f32" + The data type to be employed, by default f32. random_state: int = 42 The random state to reproduce the training sequence. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -118,7 +124,9 @@ def __init__( stochastic_downsample_by_degree=stochastic_downsample_by_degree, normalize_learning_rate_by_degree=normalize_learning_rate_by_degree, use_scale_free_distribution=use_scale_free_distribution, + dtype=dtype, random_state=random_state, + ring_bell=ring_bell, enable_cache=enable_cache ) diff --git a/embiggen/embedders/ensmallen_embedders/walklets_glove.py b/embiggen/embedders/ensmallen_embedders/walklets_glove.py index 64fb7f4b..ef840f99 100644 --- a/embiggen/embedders/ensmallen_embedders/walklets_glove.py +++ b/embiggen/embedders/ensmallen_embedders/walklets_glove.py @@ -9,14 +9,13 @@ class WalkletsGloVeEnsmallen(WalkletsEnsmallen): def __init__( self, embedding_size: int = 100, - epochs: int = 30, - walk_length: int = 128, - iterations: int = 10, + epochs: int = 100, + walk_length: int = 512, window_size: int = 4, return_weight: float = 1.0, explore_weight: float = 1.0, max_neighbours: Optional[int] = 100, - learning_rate: float = 0.001, + learning_rate: float = 0.05, learning_rate_decay: float = 0.9, central_nodes_embedding_path: Optional[str] = None, contextual_nodes_embedding_path: Optional[str] = None, @@ -26,6 +25,8 @@ def __init__( normalize_learning_rate_by_degree: Optional[bool] = False, use_scale_free_distribution: Optional[bool] = True, random_state: int = 42, + dtype: str = "f32", + ring_bell: bool = False, enable_cache: bool = False ): """Create new abstract Node2Vec method. @@ -34,12 +35,10 @@ def __init__( -------------------------- embedding_size: int = 100 Dimension of the embedding. - epochs: int = 30 + epochs: int = 100 Number of epochs to train the model for. walk_length: int = 128 Maximal length of the walks. - iterations: int = 10 - Number of iterations of the single walks. window_size: int = 4 Window size for the local context. On the borders the window size is trimmed. @@ -88,8 +87,12 @@ def __init__( Divide the learning rate by the degree of the central node. By default false. use_scale_free_distribution: Optional[bool] = True Sample negatives proportionally to their degree. By default true. + dtype: str = "f32" + The data type to be employed, by default f32. random_state: int = 42 The random state to reproduce the training sequence. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -99,7 +102,7 @@ def __init__( epochs=epochs, alpha=alpha, walk_length=walk_length, - iterations=iterations, + iterations=1, window_size=window_size, return_weight=return_weight, explore_weight=explore_weight, @@ -112,7 +115,9 @@ def __init__( stochastic_downsample_by_degree=stochastic_downsample_by_degree, normalize_learning_rate_by_degree=normalize_learning_rate_by_degree, use_scale_free_distribution=use_scale_free_distribution, + dtype=dtype, random_state=random_state, + ring_bell=ring_bell, enable_cache=enable_cache ) @@ -125,7 +130,8 @@ def parameters(self) -> Dict[str, Any]: """Returns parameters for smoke test.""" removed = [ "number_of_negative_samples", - "clipping_value" + "clipping_value", + "iterations" ] return dict( { diff --git a/embiggen/embedders/ensmallen_embedders/walklets_skipgram.py b/embiggen/embedders/ensmallen_embedders/walklets_skipgram.py index 73ce0c94..8f741780 100644 --- a/embiggen/embedders/ensmallen_embedders/walklets_skipgram.py +++ b/embiggen/embedders/ensmallen_embedders/walklets_skipgram.py @@ -27,6 +27,8 @@ def __init__( normalize_learning_rate_by_degree: Optional[bool] = False, use_scale_free_distribution: Optional[bool] = True, random_state: int = 42, + dtype: str = "f32", + ring_bell: bool = False, enable_cache: bool = False ): """Create new abstract Node2Vec method. @@ -93,8 +95,12 @@ def __init__( Divide the learning rate by the degree of the central node. By default false. use_scale_free_distribution: Optional[bool] = True Sample negatives proportionally to their degree. By default true. + dtype: str = "f32" + The data type to be employed, by default f32. random_state: int = 42 The random state to reproduce the training sequence. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -118,7 +124,9 @@ def __init__( stochastic_downsample_by_degree=stochastic_downsample_by_degree, normalize_learning_rate_by_degree=normalize_learning_rate_by_degree, use_scale_free_distribution=use_scale_free_distribution, + dtype=dtype, random_state=random_state, + ring_bell=ring_bell, enable_cache=enable_cache ) diff --git a/embiggen/embedders/ensmallen_embedders/weighted_spine.py b/embiggen/embedders/ensmallen_embedders/weighted_spine.py index 8cb06df0..b428a932 100644 --- a/embiggen/embedders/ensmallen_embedders/weighted_spine.py +++ b/embiggen/embedders/ensmallen_embedders/weighted_spine.py @@ -15,6 +15,7 @@ def __init__( embedding_size: int = 100, use_edge_weights_as_probabilities: bool = False, verbose: bool = False, + ring_bell: bool = False, enable_cache: bool = False ): """Create new abstract Node2Vec method. @@ -27,6 +28,8 @@ def __init__( Whether to treat the weights as probabilities. verbose: bool = False Whether to show loading bars. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -39,6 +42,7 @@ def __init__( super().__init__( embedding_size=embedding_size, + ring_bell=ring_bell, enable_cache=enable_cache ) diff --git a/embiggen/embedders/graph_embedding_pipeline.py b/embiggen/embedders/graph_embedding_pipeline.py index 34713828..53a9489a 100644 --- a/embiggen/embedders/graph_embedding_pipeline.py +++ b/embiggen/embedders/graph_embedding_pipeline.py @@ -34,7 +34,8 @@ def embed_graph( Graph version to retrieve. library_name: Optional[str] = None The library from where to retrieve the embedding model. - enable_cache: bool = False + ring_bell: bool = False, + enable_cache: bool = False Whether to enable the cache. smoke_test: bool = False Whether this run should be considered a smoke test diff --git a/embiggen/embedders/karateclub_embedders/boostne.py b/embiggen/embedders/karateclub_embedders/boostne.py index 461fee04..863b8cfe 100644 --- a/embiggen/embedders/karateclub_embedders/boostne.py +++ b/embiggen/embedders/karateclub_embedders/boostne.py @@ -13,6 +13,7 @@ def __init__( order: int = 2, alpha: float = 0.01, random_state: int = 42, + ring_bell: bool = False, enable_cache: bool = False ): """Return a new BoostNE embedding model. @@ -30,6 +31,8 @@ def __init__( random_state: int = 42 Random state to use for the stocastic portions of the embedding algorithm. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -40,6 +43,7 @@ def __init__( super().__init__( embedding_size=embedding_size, enable_cache=enable_cache, + ring_bell=ring_bell, random_state=random_state ) diff --git a/embiggen/embedders/karateclub_embedders/deep_walk.py b/embiggen/embedders/karateclub_embedders/deep_walk.py index db084890..17d14ba3 100644 --- a/embiggen/embedders/karateclub_embedders/deep_walk.py +++ b/embiggen/embedders/karateclub_embedders/deep_walk.py @@ -17,6 +17,7 @@ def __init__( learning_rate: float = 0.05, min_count: int = 1, random_state: int = 42, + ring_bell: bool = False, enable_cache: bool = False ): """Return a new DeepWalk embedding model. @@ -40,6 +41,8 @@ def __init__( random_state: int = 42 Random state to use for the stocastic portions of the embedding algorithm. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -54,6 +57,7 @@ def __init__( super().__init__( embedding_size=embedding_size, enable_cache=enable_cache, + ring_bell=ring_bell, random_state=random_state ) diff --git a/embiggen/embedders/karateclub_embedders/geometric_laplacian_eigenmaps.py b/embiggen/embedders/karateclub_embedders/geometric_laplacian_eigenmaps.py index eb6982a3..e55a0770 100644 --- a/embiggen/embedders/karateclub_embedders/geometric_laplacian_eigenmaps.py +++ b/embiggen/embedders/karateclub_embedders/geometric_laplacian_eigenmaps.py @@ -10,6 +10,7 @@ def __init__( self, embedding_size: int = 100, random_state: int = 42, + ring_bell: bool = False, enable_cache: bool = False ): """Return a new GLEE embedding model. @@ -21,6 +22,8 @@ def __init__( random_state: int = 42 Random state to use for the stocastic portions of the embedding algorithm. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -28,6 +31,7 @@ def __init__( super().__init__( embedding_size=embedding_size, enable_cache=enable_cache, + ring_bell=ring_bell, random_state=random_state ) diff --git a/embiggen/embedders/karateclub_embedders/grarep.py b/embiggen/embedders/karateclub_embedders/grarep.py index 1c67feeb..c0faa166 100644 --- a/embiggen/embedders/karateclub_embedders/grarep.py +++ b/embiggen/embedders/karateclub_embedders/grarep.py @@ -12,6 +12,7 @@ def __init__( iteration: int = 10, order: int = 5, random_state: int = 42, + ring_bell: bool = False, enable_cache: bool = False ): """Return a new GraRep embedding model. @@ -27,6 +28,8 @@ def __init__( random_state: int = 42 Random state to use for the stocastic portions of the embedding algorithm. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -36,6 +39,7 @@ def __init__( super().__init__( embedding_size=embedding_size, enable_cache=enable_cache, + ring_bell=ring_bell, random_state=random_state ) diff --git a/embiggen/embedders/karateclub_embedders/hope.py b/embiggen/embedders/karateclub_embedders/hope.py index ef74dad4..c41b4178 100644 --- a/embiggen/embedders/karateclub_embedders/hope.py +++ b/embiggen/embedders/karateclub_embedders/hope.py @@ -10,6 +10,7 @@ def __init__( self, embedding_size: int = 100, random_state: int = 42, + ring_bell: bool = False, enable_cache: bool = False ): """Return a new HOPE embedding model. @@ -21,6 +22,8 @@ def __init__( random_state: int = 42 Random state to use for the stocastic portions of the embedding algorithm. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -29,6 +32,7 @@ def __init__( super().__init__( embedding_size=embedding_size, enable_cache=enable_cache, + ring_bell=ring_bell, random_state=random_state ) diff --git a/embiggen/embedders/karateclub_embedders/laplacian_eigenmaps.py b/embiggen/embedders/karateclub_embedders/laplacian_eigenmaps.py index e360ee8d..b1fc3233 100644 --- a/embiggen/embedders/karateclub_embedders/laplacian_eigenmaps.py +++ b/embiggen/embedders/karateclub_embedders/laplacian_eigenmaps.py @@ -10,6 +10,7 @@ def __init__( self, embedding_size: int = 100, random_state: int = 42, + ring_bell: bool = False, enable_cache: bool = False ): """Return a new LaplacianEigenmaps embedding model. @@ -21,6 +22,8 @@ def __init__( random_state: int = 42 Random state to use for the stocastic portions of the embedding algorithm. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -28,6 +31,7 @@ def __init__( super().__init__( embedding_size=embedding_size, enable_cache=enable_cache, + ring_bell=ring_bell, random_state=random_state ) diff --git a/embiggen/embedders/karateclub_embedders/netmf.py b/embiggen/embedders/karateclub_embedders/netmf.py index 86ae41c8..6afcff3e 100644 --- a/embiggen/embedders/karateclub_embedders/netmf.py +++ b/embiggen/embedders/karateclub_embedders/netmf.py @@ -13,6 +13,7 @@ def __init__( order: int = 2, negative_samples: int = 1, random_state: int = 42, + ring_bell: bool = False, enable_cache: bool = False ): """Return a new NetMF embedding model. @@ -30,6 +31,8 @@ def __init__( random_state: int = 42 Random state to use for the stocastic portions of the embedding algorithm. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -40,6 +43,7 @@ def __init__( super().__init__( embedding_size=embedding_size, enable_cache=enable_cache, + ring_bell=ring_bell, random_state=random_state ) diff --git a/embiggen/embedders/karateclub_embedders/nmfadmm.py b/embiggen/embedders/karateclub_embedders/nmfadmm.py index 8026e687..80944645 100644 --- a/embiggen/embedders/karateclub_embedders/nmfadmm.py +++ b/embiggen/embedders/karateclub_embedders/nmfadmm.py @@ -12,6 +12,7 @@ def __init__( iterations: int = 100, rho: float = 1.0, random_state: int = 42, + ring_bell: bool = False, enable_cache: bool = False ): """Return a new NMFADMM embedding model. @@ -27,6 +28,8 @@ def __init__( random_state: int = 42 Random state to use for the stocastic portions of the embedding algorithm. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -37,6 +40,7 @@ def __init__( super().__init__( embedding_size=embedding_size, enable_cache=enable_cache, + ring_bell=ring_bell, random_state=random_state ) diff --git a/embiggen/embedders/karateclub_embedders/node_sketch.py b/embiggen/embedders/karateclub_embedders/node_sketch.py index 0829eea7..fb4b27ce 100644 --- a/embiggen/embedders/karateclub_embedders/node_sketch.py +++ b/embiggen/embedders/karateclub_embedders/node_sketch.py @@ -12,6 +12,7 @@ def __init__( iterations: int = 10, decay: float = 0.01, random_state: int = 42, + ring_bell: bool = False, enable_cache: bool = False ): """Return a new NodeSketch embedding model. @@ -27,6 +28,8 @@ def __init__( random_state: int = 42 Random state to use for the stocastic portions of the embedding algorithm. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -37,6 +40,7 @@ def __init__( super().__init__( embedding_size=embedding_size, enable_cache=enable_cache, + ring_bell=ring_bell, random_state=random_state ) diff --git a/embiggen/embedders/karateclub_embedders/randne.py b/embiggen/embedders/karateclub_embedders/randne.py index bee120c0..84e027fc 100644 --- a/embiggen/embedders/karateclub_embedders/randne.py +++ b/embiggen/embedders/karateclub_embedders/randne.py @@ -11,6 +11,7 @@ def __init__( embedding_size: int = 100, alphas: Union[List[float], Tuple[float]] = (0.5, 0.5), random_state: int = 42, + ring_bell: bool = False, enable_cache: bool = False ): """Return a new RandNE embedding model. @@ -24,6 +25,8 @@ def __init__( random_state: int = 42 Random state to use for the stocastic portions of the embedding algorithm. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -32,6 +35,7 @@ def __init__( super().__init__( embedding_size=embedding_size, enable_cache=enable_cache, + ring_bell=ring_bell, random_state=random_state ) diff --git a/embiggen/embedders/karateclub_embedders/role2vec.py b/embiggen/embedders/karateclub_embedders/role2vec.py index 3f6ad6a4..1a925f4a 100644 --- a/embiggen/embedders/karateclub_embedders/role2vec.py +++ b/embiggen/embedders/karateclub_embedders/role2vec.py @@ -20,6 +20,7 @@ def __init__( weisfeiler_lehman_hashing_iterations: int = 2, erase_base_features: bool = False, random_state: int = 42, + ring_bell: bool = False, enable_cache: bool = False ): """Return a new Role2Vec embedding model. @@ -49,6 +50,8 @@ def __init__( random_state: int = 42 Random state to use for the stocastic portions of the embedding algorithm. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -66,6 +69,7 @@ def __init__( super().__init__( embedding_size=embedding_size, enable_cache=enable_cache, + ring_bell=ring_bell, random_state=random_state ) diff --git a/embiggen/embedders/karateclub_embedders/skipgram.py b/embiggen/embedders/karateclub_embedders/skipgram.py index bb9cb317..443aa78b 100644 --- a/embiggen/embedders/karateclub_embedders/skipgram.py +++ b/embiggen/embedders/karateclub_embedders/skipgram.py @@ -19,6 +19,7 @@ def __init__( learning_rate: float = 0.05, min_count: int = 1, random_state: int = 42, + ring_bell: bool = False, enable_cache: bool = False ): """Return a new Node2Vec CBOW embedding model. @@ -54,6 +55,8 @@ def __init__( random_state: int = 42 Random state to use for the stocastic portions of the embedding algorithm. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -70,6 +73,7 @@ def __init__( super().__init__( embedding_size=embedding_size, enable_cache=enable_cache, + ring_bell=ring_bell, random_state=random_state ) diff --git a/embiggen/embedders/karateclub_embedders/sociodim.py b/embiggen/embedders/karateclub_embedders/sociodim.py index 4ccac64f..e8b5777a 100644 --- a/embiggen/embedders/karateclub_embedders/sociodim.py +++ b/embiggen/embedders/karateclub_embedders/sociodim.py @@ -9,6 +9,7 @@ def __init__( self, embedding_size: int = 100, random_state: int = 42, + ring_bell: bool = False, enable_cache: bool = False ): """Return a new SocioDim embedding model. @@ -20,6 +21,8 @@ def __init__( random_state: int = 42 Random state to use for the stocastic portions of the embedding algorithm. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -27,6 +30,7 @@ def __init__( super().__init__( embedding_size=embedding_size, enable_cache=enable_cache, + ring_bell=ring_bell, random_state=random_state ) diff --git a/embiggen/embedders/karateclub_embedders/walklets.py b/embiggen/embedders/karateclub_embedders/walklets.py index 55b681ef..d56ba3ec 100644 --- a/embiggen/embedders/karateclub_embedders/walklets.py +++ b/embiggen/embedders/karateclub_embedders/walklets.py @@ -17,6 +17,7 @@ def __init__( learning_rate: float = 0.05, min_count: int = 1, random_state: int = 42, + ring_bell: bool = False, enable_cache: bool = False ): """Return a new Walklets embedding model. @@ -40,6 +41,8 @@ def __init__( random_state: int = 42 Random state to use for the stocastic portions of the embedding algorithm. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -54,6 +57,7 @@ def __init__( super().__init__( embedding_size=embedding_size // self._window_size, enable_cache=enable_cache, + ring_bell=ring_bell, random_state=random_state ) diff --git a/embiggen/embedders/pykeen_embedders/__init__.py b/embiggen/embedders/pykeen_embedders/__init__.py index 47705d01..affc9d1d 100644 --- a/embiggen/embedders/pykeen_embedders/__init__.py +++ b/embiggen/embedders/pykeen_embedders/__init__.py @@ -1,8 +1,8 @@ -"""Module with graph embedding models based on PyKeen.""" +"""Module with graph embedding models based on PyKEEN.""" from embiggen.utils.abstract_models import build_init, AbstractEmbeddingModel build_init( module_library_names=["torch", "pykeen"], - formatted_library_name="PyKeen", + formatted_library_name="PyKEEN", expected_parent_class=AbstractEmbeddingModel ) \ No newline at end of file diff --git a/embiggen/embedders/pykeen_embedders/auto_sf.py b/embiggen/embedders/pykeen_embedders/auto_sf.py index eabe63f8..d9abdf8d 100644 --- a/embiggen/embedders/pykeen_embedders/auto_sf.py +++ b/embiggen/embedders/pykeen_embedders/auto_sf.py @@ -1,12 +1,12 @@ -"""Submodule providing wrapper for PyKeen's AutoSF model.""" +"""Submodule providing wrapper for PyKEEN's AutoSF model.""" from typing import Union, Type, Dict, Any from pykeen.training import TrainingLoop from pykeen.models import AutoSF -from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKeen +from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKEEN from pykeen.triples import CoreTriplesFactory -class AutoSFPyKeen(EntityRelationEmbeddingModelPyKeen): +class AutoSFPyKEEN(EntityRelationEmbeddingModelPyKEEN): def __init__( self, @@ -18,14 +18,15 @@ def __init__( ] = "Stochastic Local Closed World Assumption", verbose: bool = False, random_state: int = 42, + ring_bell: bool = False, enable_cache: bool = False ): - """Create new PyKeen AutoSF model. + """Create new PyKEEN AutoSF model. Details ------------------------- This is a wrapper of the AutoSF implementation from the - PyKeen library. Please refer to the PyKeen library documentation + PyKEEN library. Please refer to the PyKEEN library documentation for details and posssible errors regarding this model. Parameters @@ -51,6 +52,8 @@ def __init__( Whether to show loading bars. random_state: int = 42 Random seed to use while training the model + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -63,6 +66,7 @@ def __init__( training_loop=training_loop, verbose=verbose, random_state=random_state, + ring_bell=ring_bell, enable_cache=enable_cache ) diff --git a/embiggen/embedders/pykeen_embedders/boxe.py b/embiggen/embedders/pykeen_embedders/boxe.py index babee384..daf0745b 100644 --- a/embiggen/embedders/pykeen_embedders/boxe.py +++ b/embiggen/embedders/pykeen_embedders/boxe.py @@ -1,12 +1,12 @@ -"""Submodule providing wrapper for PyKeen's BoxE model.""" +"""Submodule providing wrapper for PyKEEN's BoxE model.""" from typing import Union, Type, Dict, Any, Optional from pykeen.training import TrainingLoop from pykeen.models import BoxE -from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKeen +from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKEEN from pykeen.triples import CoreTriplesFactory -class BoxEPyKeen(EntityRelationEmbeddingModelPyKeen): +class BoxEPyKEEN(EntityRelationEmbeddingModelPyKEEN): def __init__( self, @@ -20,14 +20,15 @@ def __init__( ] = "Stochastic Local Closed World Assumption", verbose: bool = False, random_state: int = 42, + ring_bell: bool = False, enable_cache: bool = False ): - """Create new PyKeen BoxE model. + """Create new PyKEEN BoxE model. Details ------------------------- This is a wrapper of the BoxE implementation from the - PyKeen library. Please refer to the PyKeen library documentation + PyKEEN library. Please refer to the PyKEEN library documentation for details and posssible errors regarding this model. Parameters @@ -59,6 +60,8 @@ def __init__( Whether to show loading bars. random_state: int = 42 Random seed to use while training the model + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -73,6 +76,7 @@ def __init__( training_loop=training_loop, verbose=verbose, random_state=random_state, + ring_bell=ring_bell, enable_cache=enable_cache ) diff --git a/embiggen/embedders/pykeen_embedders/complex.py b/embiggen/embedders/pykeen_embedders/complex.py index 5258ff2e..e9ec1297 100644 --- a/embiggen/embedders/pykeen_embedders/complex.py +++ b/embiggen/embedders/pykeen_embedders/complex.py @@ -1,12 +1,12 @@ -"""Submodule providing wrapper for PyKeen's ComplEx model.""" +"""Submodule providing wrapper for PyKEEN's ComplEx model.""" from typing import Union, Type, Dict, Any, Optional from pykeen.training import TrainingLoop from pykeen.models import ComplEx -from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKeen +from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKEEN from pykeen.triples import CoreTriplesFactory -class ComplExPyKeen(EntityRelationEmbeddingModelPyKeen): +class ComplExPyKEEN(EntityRelationEmbeddingModelPyKEEN): @classmethod def model_name(cls) -> str: diff --git a/embiggen/embedders/pykeen_embedders/conve.py b/embiggen/embedders/pykeen_embedders/conve.py index e6f389d7..2c21f111 100644 --- a/embiggen/embedders/pykeen_embedders/conve.py +++ b/embiggen/embedders/pykeen_embedders/conve.py @@ -1,12 +1,12 @@ -"""Submodule providing wrapper for PyKeen's ConvE model.""" +"""Submodule providing wrapper for PyKEEN's ConvE model.""" from typing import Union, Type, Dict, Any, Optional from pykeen.training import TrainingLoop from pykeen.models import ConvE -from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKeen +from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKEEN from pykeen.triples import CoreTriplesFactory -class ConvEPyKeen(EntityRelationEmbeddingModelPyKeen): +class ConvEPyKEEN(EntityRelationEmbeddingModelPyKEEN): def __init__( self, @@ -27,14 +27,15 @@ def __init__( ] = "Stochastic Local Closed World Assumption", verbose: bool = False, random_state: int = 42, + ring_bell: bool = False, enable_cache: bool = False ): - """Create new PyKeen ConvE model. + """Create new PyKEEN ConvE model. Details ------------------------- This is a wrapper of the ConvE implementation from the - PyKeen library. Please refer to the PyKeen library documentation + PyKEEN library. Please refer to the PyKEEN library documentation for details and posssible errors regarding this model. Parameters @@ -81,6 +82,8 @@ def __init__( Whether to show loading bars. random_state: int = 42 Random seed to use while training the model + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -102,6 +105,7 @@ def __init__( training_loop=training_loop, verbose=verbose, random_state=random_state, + ring_bell=ring_bell, enable_cache=enable_cache ) @@ -153,3 +157,8 @@ def _build_model( apply_batch_normalization=self._apply_batch_normalization, random_seed=self._random_state ) + + @classmethod + def _create_inverse_triples(cls) -> bool: + """Returns whether the class is expected to create inverse triples.""" + return True \ No newline at end of file diff --git a/embiggen/embedders/pykeen_embedders/crosse.py b/embiggen/embedders/pykeen_embedders/crosse.py index 8aad0b05..a92ebb15 100644 --- a/embiggen/embedders/pykeen_embedders/crosse.py +++ b/embiggen/embedders/pykeen_embedders/crosse.py @@ -1,12 +1,12 @@ -"""Submodule providing wrapper for PyKeen's CrossE model.""" +"""Submodule providing wrapper for PyKEEN's CrossE model.""" from typing import Union, Type, Dict, Any, Optional from pykeen.training import TrainingLoop from pykeen.models import CrossE -from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKeen +from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKEEN from pykeen.triples import CoreTriplesFactory -class CrossEPyKeen(EntityRelationEmbeddingModelPyKeen): +class CrossEPyKEEN(EntityRelationEmbeddingModelPyKEEN): def __init__( self, @@ -18,14 +18,15 @@ def __init__( ] = "Stochastic Local Closed World Assumption", verbose: bool = False, random_state: int = 42, + ring_bell: bool = False, enable_cache: bool = False ): - """Create new PyKeen CrossE model. + """Create new PyKEEN CrossE model. Details ------------------------- This is a wrapper of the CrossE implementation from the - PyKeen library. Please refer to the PyKeen library documentation + PyKEEN library. Please refer to the PyKEEN library documentation for details and posssible errors regarding this model. Parameters @@ -51,6 +52,8 @@ def __init__( Whether to show loading bars. random_state: int = 42 Random seed to use while training the model + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -63,6 +66,7 @@ def __init__( training_loop=training_loop, verbose=verbose, random_state=random_state, + ring_bell=ring_bell, enable_cache=enable_cache ) diff --git a/embiggen/embedders/pykeen_embedders/distma.py b/embiggen/embedders/pykeen_embedders/distma.py index cfa6d0db..68782adb 100644 --- a/embiggen/embedders/pykeen_embedders/distma.py +++ b/embiggen/embedders/pykeen_embedders/distma.py @@ -1,12 +1,12 @@ -"""Submodule providing wrapper for PyKeen's DistMA model.""" +"""Submodule providing wrapper for PyKEEN's DistMA model.""" from typing import Union, Type, Dict, Any, Optional from pykeen.training import TrainingLoop from pykeen.models import DistMA -from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKeen +from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKEEN from pykeen.triples import CoreTriplesFactory -class DistMAPyKeen(EntityRelationEmbeddingModelPyKeen): +class DistMAPyKEEN(EntityRelationEmbeddingModelPyKEEN): @classmethod def model_name(cls) -> str: diff --git a/embiggen/embedders/pykeen_embedders/distmult.py b/embiggen/embedders/pykeen_embedders/distmult.py index c628b683..7911647f 100644 --- a/embiggen/embedders/pykeen_embedders/distmult.py +++ b/embiggen/embedders/pykeen_embedders/distmult.py @@ -1,12 +1,12 @@ -"""Submodule providing wrapper for PyKeen's DistMult model.""" +"""Submodule providing wrapper for PyKEEN's DistMult model.""" from typing import Union, Type, Dict, Any, Optional from pykeen.training import TrainingLoop from pykeen.models import DistMult -from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKeen +from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKEEN from pykeen.triples import CoreTriplesFactory -class DistMultPyKeen(EntityRelationEmbeddingModelPyKeen): +class DistMultPyKEEN(EntityRelationEmbeddingModelPyKEEN): @classmethod def model_name(cls) -> str: diff --git a/embiggen/embedders/pykeen_embedders/entity_relation_embedding_model_pykeen.py b/embiggen/embedders/pykeen_embedders/entity_relation_embedding_model_pykeen.py index d3459a61..40958503 100644 --- a/embiggen/embedders/pykeen_embedders/entity_relation_embedding_model_pykeen.py +++ b/embiggen/embedders/pykeen_embedders/entity_relation_embedding_model_pykeen.py @@ -1,14 +1,22 @@ -"""Submodule providing wrapper for PyKeen's TransE model.""" -from typing import Union +"""Submodule providing wrapper for PyKEEN's TransE model.""" +from typing import Union, Type, List from ensmallen import Graph -from pykeen.models import EntityRelationEmbeddingModel, ERModel -from embiggen.embedders.pykeen_embedders.pykeen_embedder import PyKeenEmbedder +from pykeen.models import ERModel +from pykeen.nn.representation import Representation +from embiggen.embedders.pykeen_embedders.pykeen_embedder import PyKEENEmbedder import pandas as pd from embiggen.utils.abstract_models import abstract_class, EmbeddingResult +try: + from pykeen.models import EntityRelationEmbeddingModel +except ImportError: + # The following is just to patch the removal of the + # class EntityRelationEmbeddingModel in PyKEEN. + class EntityRelationEmbeddingModel: + pass @abstract_class -class EntityRelationEmbeddingModelPyKeen(PyKeenEmbedder): +class EntityRelationEmbeddingModelPyKEEN(PyKEENEmbedder): def _extract_embeddings( self, @@ -28,11 +36,11 @@ def _extract_embeddings( Whether to return a dataframe of a numpy array. """ if isinstance(model, EntityRelationEmbeddingModel): - node_embeddings = [model.entity_embeddings] - edge_type_embeddings = [model.relation_embeddings] + node_embeddings: List[Type[Representation]] = [model.entity_embeddings] + edge_type_embeddings: List[Type[Representation]] = [model.relation_embeddings] elif isinstance(model, ERModel): - node_embeddings = model.entity_representations - edge_type_embeddings = model.relation_representations + node_embeddings: List[Type[Representation]] = model.entity_representations + edge_type_embeddings: List[Type[Representation]] = model.relation_representations else: raise NotImplementedError( f"The provided model has type {type(model)}, which " @@ -41,13 +49,27 @@ def _extract_embeddings( ) node_embeddings = [ - node_embedding._embeddings.weight.cpu().detach().numpy() - for node_embedding in node_embeddings + array + for array in ( + node_embedding().cpu().detach().numpy().reshape(( + graph.get_number_of_nodes(), + -1 + )) + for node_embedding in node_embeddings + ) + if array.size > 0 ] edge_type_embeddings = [ - edge_type_embedding._embeddings.weight.cpu().detach().numpy() - for edge_type_embedding in edge_type_embeddings + array + for array in ( + edge_type_embedding().cpu().detach().numpy().reshape(( + graph.get_number_of_edge_types(), + -1 + )) + for edge_type_embedding in edge_type_embeddings + ) + if array.size > 0 ] if return_dataframe: diff --git a/embiggen/embedders/pykeen_embedders/ermlp.py b/embiggen/embedders/pykeen_embedders/ermlp.py index f0788f15..125eaf9d 100644 --- a/embiggen/embedders/pykeen_embedders/ermlp.py +++ b/embiggen/embedders/pykeen_embedders/ermlp.py @@ -1,12 +1,12 @@ -"""Submodule providing wrapper for PyKeen's ERMLP model.""" +"""Submodule providing wrapper for PyKEEN's ERMLP model.""" from typing import Union, Type, Dict, Any, Optional from pykeen.training import TrainingLoop from pykeen.models import ERMLP -from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKeen +from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKEEN from pykeen.triples import CoreTriplesFactory -class ERMLPPyKeen(EntityRelationEmbeddingModelPyKeen): +class ERMLPPyKEEN(EntityRelationEmbeddingModelPyKEEN): def __init__( self, @@ -18,14 +18,15 @@ def __init__( ] = "Stochastic Local Closed World Assumption", verbose: bool = False, random_state: int = 42, + ring_bell: bool = False, enable_cache: bool = False ): - """Create new PyKeen ERMLP model. + """Create new PyKEEN ERMLP model. Details ------------------------- This is a wrapper of the ERMLP implementation from the - PyKeen library. Please refer to the PyKeen library documentation + PyKEEN library. Please refer to the PyKEEN library documentation for details and posssible errors regarding this model. Parameters @@ -51,6 +52,8 @@ def __init__( Whether to show loading bars. random_state: int = 42 Random seed to use while training the model + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -63,6 +66,7 @@ def __init__( training_loop=training_loop, verbose=verbose, random_state=random_state, + ring_bell=ring_bell, enable_cache=enable_cache ) @@ -70,7 +74,7 @@ def __init__( def smoke_test_parameters(cls) -> Dict[str, Any]: """Returns parameters for smoke test.""" return dict( - EntityRelationEmbeddingModelPyKeen.smoke_test_parameters(), + EntityRelationEmbeddingModelPyKEEN.smoke_test_parameters(), hidden_dim=5 ) diff --git a/embiggen/embedders/pykeen_embedders/ermlpe.py b/embiggen/embedders/pykeen_embedders/ermlpe.py index 638bc0c4..4ebdd88a 100644 --- a/embiggen/embedders/pykeen_embedders/ermlpe.py +++ b/embiggen/embedders/pykeen_embedders/ermlpe.py @@ -1,12 +1,12 @@ -"""Submodule providing wrapper for PyKeen's ERMLPE model.""" +"""Submodule providing wrapper for PyKEEN's ERMLPE model.""" from typing import Union, Type, Dict, Any, Optional from pykeen.training import TrainingLoop from pykeen.models import ERMLPE -from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKeen +from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKEEN from pykeen.triples import CoreTriplesFactory -class ERMLPEPyKeen(EntityRelationEmbeddingModelPyKeen): +class ERMLPEPyKEEN(EntityRelationEmbeddingModelPyKEEN): def __init__( self, @@ -18,14 +18,15 @@ def __init__( ] = "Stochastic Local Closed World Assumption", verbose: bool = False, random_state: int = 42, + ring_bell: bool = False, enable_cache: bool = False ): - """Create new PyKeen ERMLPE model. + """Create new PyKEEN ERMLPE model. Details ------------------------- This is a wrapper of the ERMLPE implementation from the - PyKeen library. Please refer to the PyKeen library documentation + PyKEEN library. Please refer to the PyKEEN library documentation for details and posssible errors regarding this model. Parameters @@ -51,6 +52,8 @@ def __init__( Whether to show loading bars. random_state: int = 42 Random seed to use while training the model + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -63,6 +66,7 @@ def __init__( training_loop=training_loop, verbose=verbose, random_state=random_state, + ring_bell=ring_bell, enable_cache=enable_cache ) @@ -70,7 +74,7 @@ def __init__( def smoke_test_parameters(cls) -> Dict[str, Any]: """Returns parameters for smoke test.""" return dict( - EntityRelationEmbeddingModelPyKeen.smoke_test_parameters(), + EntityRelationEmbeddingModelPyKEEN.smoke_test_parameters(), hidden_dim=5 ) diff --git a/embiggen/embedders/pykeen_embedders/hole.py b/embiggen/embedders/pykeen_embedders/hole.py index 901ac5fa..9a4a86b3 100644 --- a/embiggen/embedders/pykeen_embedders/hole.py +++ b/embiggen/embedders/pykeen_embedders/hole.py @@ -1,12 +1,12 @@ -"""Submodule providing wrapper for PyKeen's HolE model.""" +"""Submodule providing wrapper for PyKEEN's HolE model.""" from typing import Union, Type, Dict, Any from pykeen.training import TrainingLoop from pykeen.models import HolE -from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKeen +from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKEEN from pykeen.triples import CoreTriplesFactory -class HolEPyKeen(EntityRelationEmbeddingModelPyKeen): +class HolEPyKEEN(EntityRelationEmbeddingModelPyKEEN): @classmethod def model_name(cls) -> str: diff --git a/embiggen/embedders/pykeen_embedders/cp.py b/embiggen/embedders/pykeen_embedders/node_piece.py similarity index 64% rename from embiggen/embedders/pykeen_embedders/cp.py rename to embiggen/embedders/pykeen_embedders/node_piece.py index 885eb52a..567ac58a 100644 --- a/embiggen/embedders/pykeen_embedders/cp.py +++ b/embiggen/embedders/pykeen_embedders/node_piece.py @@ -1,39 +1,42 @@ -"""Submodule providing wrapper for PyKeen's CP model.""" -from typing import Union, Type, Dict, Any, Optional +"""Submodule providing wrapper for PyKEEN's NodePiece model.""" +from typing import Union, Type, Dict, Any, List from pykeen.training import TrainingLoop -from pykeen.models import CP -from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKeen +from pykeen.models import NodePiece +from ensmallen import Graph +from embiggen.utils.abstract_models import EmbeddingResult +from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKEEN from pykeen.triples import CoreTriplesFactory -class CPPyKeen(EntityRelationEmbeddingModelPyKeen): +class NodePiecePyKEEN(EntityRelationEmbeddingModelPyKEEN): def __init__( self, - embedding_size: int = 256, - rank: int = 64, + embedding_size: int = 64, + num_tokens: Union[int, List[int]] = 2, epochs: int = 100, batch_size: int = 210, training_loop: Union[str, Type[TrainingLoop] ] = "Stochastic Local Closed World Assumption", verbose: bool = False, random_state: int = 42, + ring_bell: bool = False, enable_cache: bool = False ): - """Create new PyKeen CP model. + """Create new PyKEEN NodePiece model. Details ------------------------- - This is a wrapper of the CP implementation from the - PyKeen library. Please refer to the PyKeen library documentation + This is a wrapper of the NodePiece implementation from the + PyKEEN library. Please refer to the PyKEEN library documentation for details and posssible errors regarding this model. Parameters ------------------------- - embedding_size: int = 256 + embedding_size: int = 64 The dimension of the embedding to compute. - rank: int = 64 - The hidden dropout rate + num_tokens: Union[int, List[int]] = 2 + The number of relations to use to represent each entity, cf. epochs: int = 100 The number of epochs to use to train the model for. batch_size: int = 210 @@ -51,11 +54,13 @@ def __init__( Whether to show loading bars. random_state: int = 42 Random seed to use while training the model + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. """ - self._rank = rank + self._num_tokens = num_tokens super().__init__( embedding_size=embedding_size, epochs=epochs, @@ -63,36 +68,40 @@ def __init__( training_loop=training_loop, verbose=verbose, random_state=random_state, + ring_bell=ring_bell, enable_cache=enable_cache ) def parameters(self) -> Dict[str, Any]: return dict( super().parameters(), - dict( - rank=self._rank, - ) + num_tokens=self._num_tokens, ) @classmethod def model_name(cls) -> str: """Return name of the model.""" - return "CP" + return "NodePiece" def _build_model( self, triples_factory: CoreTriplesFactory - ) -> CP: - """Build new CP model for embedding. + ) -> NodePiece: + """Build new NodePiece model for embedding. Parameters ------------------ graph: Graph The graph to build the model for. """ - return CP( + return NodePiece( triples_factory=triples_factory, + num_tokens=self._num_tokens, embedding_dim=self._embedding_size, - rank=self._rank, random_seed=self._random_state ) + + @classmethod + def _create_inverse_triples(cls) -> bool: + """Returns whether the class is expected to create inverse triples.""" + return True \ No newline at end of file diff --git a/embiggen/embedders/pykeen_embedders/pairre.py b/embiggen/embedders/pykeen_embedders/pairre.py index 2002ba30..13c13129 100644 --- a/embiggen/embedders/pykeen_embedders/pairre.py +++ b/embiggen/embedders/pykeen_embedders/pairre.py @@ -1,12 +1,12 @@ -"""Submodule providing wrapper for PyKeen's PairRE model.""" +"""Submodule providing wrapper for PyKEEN's PairRE model.""" from typing import Union, Type, Dict, Any, Optional from pykeen.training import TrainingLoop from pykeen.models import PairRE -from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKeen +from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKEEN from pykeen.triples import CoreTriplesFactory -class PairREPyKeen(EntityRelationEmbeddingModelPyKeen): +class PairREPyKEEN(EntityRelationEmbeddingModelPyKEEN): def __init__( self, @@ -19,14 +19,15 @@ def __init__( ] = "Stochastic Local Closed World Assumption", verbose: bool = False, random_state: int = 42, + ring_bell: bool = False, enable_cache: bool = False ): - """Create new PyKeen PairRE model. + """Create new PyKEEN PairRE model. Details ------------------------- This is a wrapper of the PairRE implementation from the - PyKeen library. Please refer to the PyKeen library documentation + PyKEEN library. Please refer to the PyKEEN library documentation for details and posssible errors regarding this model. Parameters @@ -54,6 +55,8 @@ def __init__( Whether to show loading bars. random_state: int = 42 Random seed to use while training the model + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -67,6 +70,7 @@ def __init__( training_loop=training_loop, verbose=verbose, random_state=random_state, + ring_bell=ring_bell, enable_cache=enable_cache ) diff --git a/embiggen/embedders/pykeen_embedders/proje.py b/embiggen/embedders/pykeen_embedders/proje.py index e677551f..3a3f3adc 100644 --- a/embiggen/embedders/pykeen_embedders/proje.py +++ b/embiggen/embedders/pykeen_embedders/proje.py @@ -1,12 +1,12 @@ -"""Submodule providing wrapper for PyKeen's ProjE model.""" +"""Submodule providing wrapper for PyKEEN's ProjE model.""" from typing import Union, Type, Dict, Any, Optional from pykeen.training import TrainingLoop from pykeen.models import ProjE -from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKeen +from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKEEN from pykeen.triples import CoreTriplesFactory -class ProjEPyKeen(EntityRelationEmbeddingModelPyKeen): +class ProjEPyKEEN(EntityRelationEmbeddingModelPyKEEN): @classmethod def model_name(cls) -> str: diff --git a/embiggen/embedders/pykeen_embedders/pykeen_embedder.py b/embiggen/embedders/pykeen_embedders/pykeen_embedder.py index ea975a50..d0f21d5d 100644 --- a/embiggen/embedders/pykeen_embedders/pykeen_embedder.py +++ b/embiggen/embedders/pykeen_embedders/pykeen_embedder.py @@ -1,10 +1,11 @@ -"""Abstract Torch/PyKeen Model wrapper for embedding models.""" +"""Abstract Torch/PyKEEN Model wrapper for embedding models.""" from typing import Dict, Union, Tuple, Any, Type import numpy as np import pandas as pd from ensmallen import Graph import inspect +from inspect import getfullargspec from embiggen.utils.pytorch_utils import validate_torch_device from embiggen.utils.abstract_models import AbstractEmbeddingModel, abstract_class, EmbeddingResult @@ -16,8 +17,8 @@ @abstract_class -class PyKeenEmbedder(AbstractEmbeddingModel): - """Abstract Torch/PyKeen Model wrapper for embedding models.""" +class PyKEENEmbedder(AbstractEmbeddingModel): + """Abstract Torch/PyKEEN Model wrapper for embedding models.""" SUPPORTED_TRAINING_LOOPS = { "Stochastic Local Closed World Assumption": SLCWATrainingLoop, @@ -34,9 +35,10 @@ def __init__( ] = "Stochastic Local Closed World Assumption", verbose: bool = False, random_state: int = 42, + ring_bell: bool = False, enable_cache: bool = False ): - """Create new PyKeen Abstract Embedder model. + """Create new PyKEEN Abstract Embedder model. Parameters ------------------------- @@ -59,18 +61,20 @@ def __init__( Whether to show the loading bar. random_state: int = 42 Random seed to use while training the model + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. """ if isinstance(training_loop, str): - if training_loop in PyKeenEmbedder.SUPPORTED_TRAINING_LOOPS: - training_loop = PyKeenEmbedder.SUPPORTED_TRAINING_LOOPS[training_loop] + if training_loop in PyKEENEmbedder.SUPPORTED_TRAINING_LOOPS: + training_loop = PyKEENEmbedder.SUPPORTED_TRAINING_LOOPS[training_loop] else: raise ValueError( f"The provided training loop name {training_loop} is not " "a supported training loop name. " - f"The supported names are {format_list(PyKeenEmbedder.SUPPORTED_TRAINING_LOOPS)}." + f"The supported names are {format_list(PyKEENEmbedder.SUPPORTED_TRAINING_LOOPS)}." ) if not inspect.isclass(training_loop): @@ -92,6 +96,7 @@ def __init__( super().__init__( embedding_size=embedding_size, enable_cache=enable_cache, + ring_bell=ring_bell, random_state=random_state ) @@ -114,7 +119,7 @@ def parameters(self) -> Dict[str, Any]: @classmethod def library_name(cls) -> str: - return "PyKeen" + return "PyKEEN" @classmethod def task_name(cls) -> str: @@ -126,7 +131,7 @@ def _build_model(self, triples_factory: CoreTriplesFactory) -> Type[Model]: Parameters ------------------ triples_factory: CoreTriplesFactory - The PyKeen triples factory to use to create the model. + The PyKEEN triples factory to use to create the model. """ raise NotImplementedError( f"In the child class {self.__class__.__name__} of {super().__name__.__name__} " @@ -167,6 +172,11 @@ def _extract_embeddings( "called `_extract_embeddings`. Please do implement it." ) + @classmethod + def _create_inverse_triples(cls) -> bool: + """Returns whether the class is expected to create inverse triples.""" + return False + def _fit_transform( self, graph: Graph, @@ -176,14 +186,22 @@ def _fit_transform( torch_device = torch.device(self._device) - triples_factory = CoreTriplesFactory( - torch.IntTensor(graph.get_directed_edge_triples_ids().astype(np.int32)), - num_entities=graph.get_number_of_nodes(), - num_relations=graph.get_number_of_edge_types(), - entity_ids=graph.get_node_ids(), - relation_ids=graph.get_unique_edge_type_ids(), - create_inverse_triples=False, - ) + if "entity_ids" in getfullargspec(CoreTriplesFactory).args: + triples_factory = CoreTriplesFactory( + torch.IntTensor(graph.get_directed_edge_triples_ids().astype(np.int64)), + num_entities=graph.get_number_of_nodes(), + num_relations=graph.get_number_of_edge_types(), + entity_ids=graph.get_node_ids().astype(np.int64), + relation_ids=graph.get_unique_edge_type_ids().astype(np.int64), + create_inverse_triples=self._create_inverse_triples(), + ) + else: + triples_factory = CoreTriplesFactory( + torch.IntTensor(graph.get_directed_edge_triples_ids().astype(np.int64)), + num_entities=graph.get_number_of_nodes(), + num_relations=graph.get_number_of_edge_types(), + create_inverse_triples=self._create_inverse_triples(), + ) batch_size = min( self._batch_size, @@ -197,7 +215,7 @@ def _fit_transform( "The model created with the `_build_model` in the child " f"class {self.__class__.__name__} for the model {self.model_name()} " f"in the library {self.library_name()} did not return a " - f"PyKeen model but an object of type {type(model)}." + f"PyKEEN model but an object of type {type(model)}." ) # Move the model to gpu if we need to diff --git a/embiggen/embedders/pykeen_embedders/quate.py b/embiggen/embedders/pykeen_embedders/quate.py index ca9caff0..9d09197c 100644 --- a/embiggen/embedders/pykeen_embedders/quate.py +++ b/embiggen/embedders/pykeen_embedders/quate.py @@ -1,12 +1,12 @@ -"""Submodule providing wrapper for PyKeen's QuatE model.""" +"""Submodule providing wrapper for PyKEEN's QuatE model.""" from typing import Union, Type, Dict, Any, Optional from pykeen.training import TrainingLoop from pykeen.models import QuatE -from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKeen +from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKEEN from pykeen.triples import CoreTriplesFactory -class QuatEPyKeen(EntityRelationEmbeddingModelPyKeen): +class QuatEPyKEEN(EntityRelationEmbeddingModelPyKEEN): @classmethod def model_name(cls) -> str: diff --git a/embiggen/embedders/pykeen_embedders/rescal.py b/embiggen/embedders/pykeen_embedders/rescal.py index 40071b24..adb9b3ed 100644 --- a/embiggen/embedders/pykeen_embedders/rescal.py +++ b/embiggen/embedders/pykeen_embedders/rescal.py @@ -1,12 +1,12 @@ -"""Submodule providing wrapper for PyKeen's RESCAL model.""" +"""Submodule providing wrapper for PyKEEN's RESCAL model.""" from typing import Union, Type, Dict, Any from pykeen.training import TrainingLoop from pykeen.models import RESCAL -from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKeen +from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKEEN from pykeen.triples import CoreTriplesFactory -class RESCALPyKeen(EntityRelationEmbeddingModelPyKeen): +class RESCALPyKEEN(EntityRelationEmbeddingModelPyKEEN): @classmethod def model_name(cls) -> str: diff --git a/embiggen/embedders/pykeen_embedders/rotate.py b/embiggen/embedders/pykeen_embedders/rotate.py index 0a46929c..aa133e58 100644 --- a/embiggen/embedders/pykeen_embedders/rotate.py +++ b/embiggen/embedders/pykeen_embedders/rotate.py @@ -1,12 +1,12 @@ -"""Submodule providing wrapper for PyKeen's RotatE model.""" +"""Submodule providing wrapper for PyKEEN's RotatE model.""" from typing import Union, Type, Dict, Any from pykeen.training import TrainingLoop from pykeen.models import RotatE -from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKeen +from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKEEN from pykeen.triples import CoreTriplesFactory -class RotatEPyKeen(EntityRelationEmbeddingModelPyKeen): +class RotatEPyKEEN(EntityRelationEmbeddingModelPyKEEN): @classmethod def model_name(cls) -> str: diff --git a/embiggen/embedders/pykeen_embedders/toruse.py b/embiggen/embedders/pykeen_embedders/toruse.py index c0d101b4..2ab6bb17 100644 --- a/embiggen/embedders/pykeen_embedders/toruse.py +++ b/embiggen/embedders/pykeen_embedders/toruse.py @@ -1,12 +1,12 @@ -"""Submodule providing wrapper for PyKeen's TorusE model.""" +"""Submodule providing wrapper for PyKEEN's TorusE model.""" from typing import Union, Type, Dict, Any, Optional from pykeen.training import TrainingLoop from pykeen.models import TorusE -from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKeen +from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKEEN from pykeen.triples import CoreTriplesFactory -class TorusEPyKeen(EntityRelationEmbeddingModelPyKeen): +class TorusEPyKEEN(EntityRelationEmbeddingModelPyKEEN): def __init__( self, @@ -19,14 +19,15 @@ def __init__( ] = "Stochastic Local Closed World Assumption", verbose: bool = False, random_state: int = 42, + ring_bell: bool = False, enable_cache: bool = False ): - """Create new PyKeen TorusE model. + """Create new PyKEEN TorusE model. Details ------------------------- This is a wrapper of the TorusE implementation from the - PyKeen library. Please refer to the PyKeen library documentation + PyKEEN library. Please refer to the PyKEEN library documentation for details and posssible errors regarding this model. Parameters @@ -54,6 +55,8 @@ def __init__( Whether to show loading bars. random_state: int = 42 Random seed to use while training the model + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -67,6 +70,7 @@ def __init__( training_loop=training_loop, verbose=verbose, random_state=random_state, + ring_bell=ring_bell, enable_cache=enable_cache ) diff --git a/embiggen/embedders/pykeen_embedders/transd.py b/embiggen/embedders/pykeen_embedders/transd.py index 1b8c4c39..a8c0510d 100644 --- a/embiggen/embedders/pykeen_embedders/transd.py +++ b/embiggen/embedders/pykeen_embedders/transd.py @@ -1,12 +1,12 @@ -"""Submodule providing wrapper for PyKeen's TransD model.""" +"""Submodule providing wrapper for PyKEEN's TransD model.""" from typing import Union, Type, Dict, Any, Optional from pykeen.training import TrainingLoop from pykeen.models import TransD -from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKeen +from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKEEN from pykeen.triples import CoreTriplesFactory -class TransDPyKeen(EntityRelationEmbeddingModelPyKeen): +class TransDPyKEEN(EntityRelationEmbeddingModelPyKEEN): def __init__( self, @@ -18,14 +18,15 @@ def __init__( ] = "Stochastic Local Closed World Assumption", verbose: bool = False, random_state: int = 42, + ring_bell: bool = False, enable_cache: bool = False ): - """Create new PyKeen TransD model. + """Create new PyKEEN TransD model. Details ------------------------- This is a wrapper of the TransD implementation from the - PyKeen library. Please refer to the PyKeen library documentation + PyKEEN library. Please refer to the PyKEEN library documentation for details and posssible errors regarding this model. Parameters @@ -51,6 +52,8 @@ def __init__( Whether to show loading bars. random_state: int = 42 Random seed to use while training the model + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -63,6 +66,7 @@ def __init__( training_loop=training_loop, verbose=verbose, random_state=random_state, + ring_bell=ring_bell, enable_cache=enable_cache ) @@ -70,7 +74,7 @@ def __init__( def smoke_test_parameters(cls) -> Dict[str, Any]: """Returns parameters for smoke test.""" return dict( - EntityRelationEmbeddingModelPyKeen.smoke_test_parameters(), + EntityRelationEmbeddingModelPyKEEN.smoke_test_parameters(), relation_dim=5, ) diff --git a/embiggen/embedders/pykeen_embedders/transe.py b/embiggen/embedders/pykeen_embedders/transe.py index 3435e5bc..f460ba20 100644 --- a/embiggen/embedders/pykeen_embedders/transe.py +++ b/embiggen/embedders/pykeen_embedders/transe.py @@ -1,12 +1,12 @@ -"""Submodule providing wrapper for PyKeen's TransE model.""" +"""Submodule providing wrapper for PyKEEN's TransE model.""" from typing import Union, Type, Dict, Any from pykeen.training import TrainingLoop from pykeen.models import TransE -from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKeen +from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKEEN from pykeen.triples import CoreTriplesFactory -class TransEPyKeen(EntityRelationEmbeddingModelPyKeen): +class TransEPyKEEN(EntityRelationEmbeddingModelPyKEEN): def __init__( self, @@ -18,14 +18,15 @@ def __init__( ] = "Stochastic Local Closed World Assumption", verbose: bool = False, random_state: int = 42, + ring_bell: bool = False, enable_cache: bool = False ): - """Create new PyKeen TransE model. + """Create new PyKEEN TransE model. Details ------------------------- This is a wrapper of the TransE implementation from the - PyKeen library. Please refer to the PyKeen library documentation + PyKEEN library. Please refer to the PyKEEN library documentation for details and posssible errors regarding this model. Parameters @@ -51,6 +52,8 @@ def __init__( Whether to show loading bars. random_state: int = 42 Random seed to use while training the model + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -63,6 +66,7 @@ def __init__( training_loop=training_loop, verbose=verbose, random_state=random_state, + ring_bell=ring_bell, enable_cache=enable_cache ) @@ -70,7 +74,7 @@ def __init__( def smoke_test_parameters(cls) -> Dict[str, Any]: """Returns parameters for smoke test.""" return dict( - EntityRelationEmbeddingModelPyKeen.smoke_test_parameters(), + EntityRelationEmbeddingModelPyKEEN.smoke_test_parameters(), scoring_fct_norm=1 ) diff --git a/embiggen/embedders/pykeen_embedders/transf.py b/embiggen/embedders/pykeen_embedders/transf.py index f6d4cf1f..a7739700 100644 --- a/embiggen/embedders/pykeen_embedders/transf.py +++ b/embiggen/embedders/pykeen_embedders/transf.py @@ -1,12 +1,12 @@ -"""Submodule providing wrapper for PyKeen's TransF model.""" +"""Submodule providing wrapper for PyKEEN's TransF model.""" from typing import Union, Type, Dict, Any from pykeen.training import TrainingLoop from pykeen.models import TransF -from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKeen +from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKEEN from pykeen.triples import CoreTriplesFactory -class TransFPyKeen(EntityRelationEmbeddingModelPyKeen): +class TransFPyKEEN(EntityRelationEmbeddingModelPyKEEN): @classmethod def model_name(cls) -> str: diff --git a/embiggen/embedders/pykeen_embedders/transh.py b/embiggen/embedders/pykeen_embedders/transh.py index 1fddcaf6..a8935375 100644 --- a/embiggen/embedders/pykeen_embedders/transh.py +++ b/embiggen/embedders/pykeen_embedders/transh.py @@ -1,12 +1,12 @@ -"""Submodule providing wrapper for PyKeen's TransH model.""" +"""Submodule providing wrapper for PyKEEN's TransH model.""" from typing import Union, Type, Dict, Any from pykeen.training import TrainingLoop from pykeen.models import TransH -from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKeen +from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKEEN from pykeen.triples import CoreTriplesFactory -class TransHPyKeen(EntityRelationEmbeddingModelPyKeen): +class TransHPyKEEN(EntityRelationEmbeddingModelPyKEEN): def __init__( self, @@ -18,14 +18,15 @@ def __init__( ] = "Stochastic Local Closed World Assumption", verbose: bool = False, random_state: int = 42, + ring_bell: bool = False, enable_cache: bool = False ): - """Create new PyKeen TransH model. + """Create new PyKEEN TransH model. Details ------------------------- This is a wrapper of the TransH implementation from the - PyKeen library. Please refer to the PyKeen library documentation + PyKEEN library. Please refer to the PyKEEN library documentation for details and posssible errors regarding this model. Parameters @@ -51,6 +52,8 @@ def __init__( Whether to show loading bars. random_state: int = 42 Random seed to use while training the model + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -63,6 +66,7 @@ def __init__( training_loop=training_loop, verbose=verbose, random_state=random_state, + ring_bell=ring_bell, enable_cache=enable_cache ) @@ -70,7 +74,7 @@ def __init__( def smoke_test_parameters(cls) -> Dict[str, Any]: """Returns parameters for smoke test.""" return dict( - EntityRelationEmbeddingModelPyKeen.smoke_test_parameters(), + EntityRelationEmbeddingModelPyKEEN.smoke_test_parameters(), scoring_fct_norm=1 ) diff --git a/embiggen/embedders/pykeen_embedders/transr.py b/embiggen/embedders/pykeen_embedders/transr.py index 3f8b5637..4162bcc4 100644 --- a/embiggen/embedders/pykeen_embedders/transr.py +++ b/embiggen/embedders/pykeen_embedders/transr.py @@ -1,12 +1,12 @@ -"""Submodule providing wrapper for PyKeen's TransR model.""" +"""Submodule providing wrapper for PyKEEN's TransR model.""" from typing import Union, Type, Dict, Any from pykeen.training import TrainingLoop from pykeen.models import TransR -from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKeen +from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKEEN from pykeen.triples import CoreTriplesFactory -class TransRPyKeen(EntityRelationEmbeddingModelPyKeen): +class TransRPyKEEN(EntityRelationEmbeddingModelPyKEEN): def __init__( self, @@ -19,14 +19,15 @@ def __init__( ] = "Stochastic Local Closed World Assumption", verbose: bool = False, random_state: int = 42, + ring_bell: bool = False, enable_cache: bool = False ): - """Create new PyKeen TransR model. + """Create new PyKEEN TransR model. Details ------------------------- This is a wrapper of the TransR implementation from the - PyKeen library. Please refer to the PyKeen library documentation + PyKEEN library. Please refer to the PyKEEN library documentation for details and posssible errors regarding this model. Parameters @@ -54,6 +55,8 @@ def __init__( Whether to show loading bars. random_state: int = 42 Random seed to use while training the model + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -67,6 +70,7 @@ def __init__( training_loop=training_loop, verbose=verbose, random_state=random_state, + ring_bell=ring_bell, enable_cache=enable_cache ) @@ -74,7 +78,7 @@ def __init__( def smoke_test_parameters(cls) -> Dict[str, Any]: """Returns parameters for smoke test.""" return dict( - EntityRelationEmbeddingModelPyKeen.smoke_test_parameters(), + EntityRelationEmbeddingModelPyKEEN.smoke_test_parameters(), scoring_fct_norm=1, relation_dim=5 ) diff --git a/embiggen/embedders/pykeen_embedders/tucker.py b/embiggen/embedders/pykeen_embedders/tucker.py index 70cc9b00..1f75927d 100644 --- a/embiggen/embedders/pykeen_embedders/tucker.py +++ b/embiggen/embedders/pykeen_embedders/tucker.py @@ -1,12 +1,12 @@ -"""Submodule providing wrapper for PyKeen's TuckER model.""" +"""Submodule providing wrapper for PyKEEN's TuckER model.""" from typing import Union, Type, Dict, Any, Optional from pykeen.training import TrainingLoop from pykeen.models import TuckER -from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKeen +from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKEEN from pykeen.triples import CoreTriplesFactory -class TuckERPyKeen(EntityRelationEmbeddingModelPyKeen): +class TuckERPyKEEN(EntityRelationEmbeddingModelPyKEEN): def __init__( self, @@ -22,14 +22,15 @@ def __init__( ] = "Stochastic Local Closed World Assumption", verbose: bool = False, random_state: int = 42, + ring_bell: bool = False, enable_cache: bool = False ): - """Create new PyKeen TuckER model. + """Create new PyKEEN TuckER model. Details ------------------------- This is a wrapper of the TuckER implementation from the - PyKeen library. Please refer to the PyKeen library documentation + PyKEEN library. Please refer to the PyKEEN library documentation for details and posssible errors regarding this model. Parameters @@ -63,6 +64,8 @@ def __init__( Whether to show loading bars. random_state: int = 42 Random seed to use while training the model + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -79,6 +82,7 @@ def __init__( training_loop=training_loop, verbose=verbose, random_state=random_state, + ring_bell=ring_bell, enable_cache=enable_cache ) diff --git a/embiggen/embedders/tensorflow_embedders/edge_prediction_based_tensorflow_embedders.py b/embiggen/embedders/tensorflow_embedders/edge_prediction_based_tensorflow_embedders.py index b949b6b8..b55422fb 100644 --- a/embiggen/embedders/tensorflow_embedders/edge_prediction_based_tensorflow_embedders.py +++ b/embiggen/embedders/tensorflow_embedders/edge_prediction_based_tensorflow_embedders.py @@ -32,6 +32,7 @@ def __init__( loss: str = "binary_crossentropy", optimizer: str = "nadam", verbose: bool = False, + ring_bell: bool = False, enable_cache: bool = False, random_state: int = 42 ): @@ -77,6 +78,8 @@ def __init__( The optimizer to be used during the training of the model. verbose: bool = False Whether to show the loading bar while training the model. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -99,6 +102,7 @@ def __init__( verbose=verbose, use_mirrored_strategy=use_mirrored_strategy, enable_cache=enable_cache, + ring_bell=ring_bell, random_state=random_state ) diff --git a/embiggen/embedders/tensorflow_embedders/node2vec.py b/embiggen/embedders/tensorflow_embedders/node2vec.py index bd61f805..4fadd1ed 100644 --- a/embiggen/embedders/tensorflow_embedders/node2vec.py +++ b/embiggen/embedders/tensorflow_embedders/node2vec.py @@ -38,6 +38,7 @@ def __init__( optimizer: str = "nadam", verbose: bool = False, use_mirrored_strategy: bool = False, + ring_bell: bool = False, enable_cache: bool = False ): """Create new abstract Node2Vec model. @@ -109,6 +110,8 @@ def __init__( Whether to show loading bars. use_mirrored_strategy: bool = False Whether to use mirrored strategy. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -136,6 +139,7 @@ def __init__( optimizer=optimizer, verbose=verbose, use_mirrored_strategy=use_mirrored_strategy, + ring_bell=ring_bell, enable_cache=enable_cache ) diff --git a/embiggen/embedders/tensorflow_embedders/siamese.py b/embiggen/embedders/tensorflow_embedders/siamese.py index 7753ec7b..16636969 100644 --- a/embiggen/embedders/tensorflow_embedders/siamese.py +++ b/embiggen/embedders/tensorflow_embedders/siamese.py @@ -34,6 +34,7 @@ def __init__( use_mirrored_strategy: bool = False, optimizer: str = "nadam", verbose: bool = False, + ring_bell: bool = False, enable_cache: bool = False, random_state: int = 42 ): @@ -72,6 +73,8 @@ def __init__( The optimizer to be used during the training of the model. verbose: bool = False Whether to show loading bars. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -92,6 +95,7 @@ def __init__( verbose=verbose, use_mirrored_strategy=use_mirrored_strategy, enable_cache=enable_cache, + ring_bell=ring_bell, random_state=random_state ) diff --git a/embiggen/embedders/tensorflow_embedders/tensorflow_embedder.py b/embiggen/embedders/tensorflow_embedders/tensorflow_embedder.py index d759f698..b33c5fae 100644 --- a/embiggen/embedders/tensorflow_embedders/tensorflow_embedder.py +++ b/embiggen/embedders/tensorflow_embedders/tensorflow_embedder.py @@ -30,6 +30,7 @@ def __init__( optimizer: str = "nadam", verbose: bool = False, use_mirrored_strategy: bool = False, + ring_bell: bool = False, enable_cache: bool = False, random_state: int = 42 ): @@ -61,6 +62,8 @@ def __init__( Whether to show loading bars. use_mirrored_strategy: bool = False Whether to use mirrored strategy. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -85,6 +88,7 @@ def __init__( super().__init__( embedding_size=embedding_size, enable_cache=enable_cache, + ring_bell=ring_bell, random_state=random_state, ) diff --git a/embiggen/node_label_prediction/node_label_prediction_evaluation.py b/embiggen/node_label_prediction/node_label_prediction_evaluation.py index aea374f1..a126edd7 100644 --- a/embiggen/node_label_prediction/node_label_prediction_evaluation.py +++ b/embiggen/node_label_prediction/node_label_prediction_evaluation.py @@ -16,13 +16,16 @@ def node_label_prediction_evaluation( library_names: Optional[Union[str, List[str]]] = None, graph_callback: Optional[Callable[[Graph], Graph]] = None, subgraph_of_interest: Optional[Graph] = None, + use_subgraph_as_support: bool = False, number_of_holdouts: int = 10, random_state: int = 42, repositories: Optional[Union[str, List[str]]] = None, versions: Optional[Union[str, List[str]]] = None, enable_cache: bool = False, - precompute_constant_automatic_stocastic_features: bool = False, + precompute_constant_stocastic_features: bool = False, smoke_test: bool = False, + number_of_slurm_nodes: Optional[int] = None, + slurm_node_id_variable: str = "SLURM_GRAPE_ID", verbose: bool = True ) -> pd.DataFrame: """Execute node-label prediction evaluation pipeline for all provided models and graphs. @@ -49,7 +52,12 @@ def node_label_prediction_evaluation( For instance this may be used for filtering the uncertain edges in graphs such as STRING PPIs. subgraph_of_interest: Optional[Graph] = None - The subgraph of interest to focus the task on. + Optional subgraph where to focus the task. + This is applied to the train and test graph + after the desired holdout schema is applied. + use_subgraph_as_support: bool = False + Whether to use the provided subgraph as support or + to use the train graph (not filtered by the subgraph). number_of_holdouts: int = 10 The number of holdouts to execute. random_state: int = 42 @@ -61,7 +69,7 @@ def node_label_prediction_evaluation( Graph versions to retrieve. enable_cache: bool = False Whether to enable the cache. - precompute_constant_automatic_stocastic_features: bool = False + precompute_constant_stocastic_features: bool = False Whether to precompute once the constant automatic stocastic features before starting the embedding loop. This means that, when left set to false, while the features will be computed @@ -75,6 +83,12 @@ def node_label_prediction_evaluation( and therefore use the smoke test configurations for the provided model names and feature names. This parameter will also turn off the cache. + number_of_slurm_nodes: Optional[int] = None + Number of SLURM nodes to consider as available. + This variable is used to parallelize the holdouts accordingly. + slurm_node_id_variable: str = "SLURM_GRAPE_ID" + Name of the system variable to use as SLURM node id. + It must be set in the slurm bash script. verbose: bool = True Whether to show loading bars """ @@ -88,12 +102,15 @@ def node_label_prediction_evaluation( library_names=library_names, graph_callback=graph_callback, subgraph_of_interest=subgraph_of_interest, + use_subgraph_as_support=use_subgraph_as_support, number_of_holdouts=number_of_holdouts, random_state=random_state, repositories=repositories, versions=versions, enable_cache=enable_cache, - precompute_constant_automatic_stocastic_features=precompute_constant_automatic_stocastic_features, + precompute_constant_stocastic_features=precompute_constant_stocastic_features, smoke_test=smoke_test, + number_of_slurm_nodes=number_of_slurm_nodes, + slurm_node_id_variable=slurm_node_id_variable, verbose=verbose ) diff --git a/embiggen/node_label_prediction/node_label_prediction_sklearn/sklearn_node_label_prediction_adapter.py b/embiggen/node_label_prediction/node_label_prediction_sklearn/sklearn_node_label_prediction_adapter.py index 9b73f9f4..12ae5026 100644 --- a/embiggen/node_label_prediction/node_label_prediction_sklearn/sklearn_node_label_prediction_adapter.py +++ b/embiggen/node_label_prediction/node_label_prediction_sklearn/sklearn_node_label_prediction_adapter.py @@ -2,6 +2,7 @@ from sklearn.base import ClassifierMixin from typing import Type, List, Dict, Optional, Any import numpy as np +import compress_pickle import copy from ensmallen import Graph from embiggen.embedding_transformers import NodeLabelPredictionTransformer, NodeTransformer @@ -224,3 +225,23 @@ def can_use_edge_types(cls) -> bool: """Returns whether the model can optionally use edge types.""" return False + @classmethod + def load(cls, path: str) -> "Self": + """Load a saved version of the model from the provided path. + + Parameters + ------------------- + path: str + Path from where to load the model. + """ + return compress_pickle.load(path) + + def dump(self, path: str): + """Dump the current model at the provided path. + + Parameters + ------------------- + path: str + Path from where to dump the model. + """ + compress_pickle.dump(self, path) \ No newline at end of file diff --git a/embiggen/similarities/dag_resnik.py b/embiggen/similarities/dag_resnik.py index e4d368ff..c78ba754 100644 --- a/embiggen/similarities/dag_resnik.py +++ b/embiggen/similarities/dag_resnik.py @@ -1,4 +1,4 @@ -from typing import List, Optional, Union, Dict +from typing import List, Optional, Union, Dict, Tuple import pandas as pd import numpy as np from ensmallen import models, Graph @@ -7,7 +7,9 @@ class DAGResnik: def __init__(self, verbose: bool = True): + """Create new Resnik similarity model.""" self._model = models.DAGResnik(verbose) + self._graph = None def fit( self, @@ -26,370 +28,381 @@ def fit( node_frequencies: Optional[np.ndarray] = None Optional vector of node frequencies. """ + self._graph = graph self._model.fit( graph, node_counts=node_counts, node_frequencies=node_frequencies ) - def get_similarity_from_node_id( + def _normalize_output( self, - first_node_id: int, - second_node_id: int - ) -> float: - """Return the similarity between the two provided nodes. - - Arguments - -------------------- - first_node_id: int - The first node for which to compute the similarity. - second_node_id: int - The second node for which to compute the similarity. - """ - return self._model.get_similarity_from_node_id(first_node_id, second_node_id) - - def get_similarity_from_node_ids( - self, - first_node_ids: List[int], - second_node_ids: List[int] - ) -> np.ndarray: - """Return the similarity between the two provided nodes. - - Arguments - -------------------- - first_node_ids: List[int] - The first node for which to compute the similarity. - second_node_ids: List[int] - The second node for which to compute the similarity. - """ - return self._model.get_similarity_from_node_ids(first_node_ids, second_node_ids) - - def get_similarity_from_node_name( - self, - first_node_name: str, - second_node_name: str - ) -> float: - """Return the similarity between the two provided nodes. - - Arguments - -------------------- - first_node_name: str - The first node for which to compute the similarity. - second_node_name: str - The second node for which to compute the similarity. - """ - return self._model.get_similarity_from_node_name(first_node_name, second_node_name) - - def get_similarity_from_node_names( - self, - first_node_names: List[str], - second_node_names: List[str] - ) -> np.ndarray: - """Return the similarity between the two provided nodes. - - Arguments - -------------------- - first_node_names: List[str] - The first node for which to compute the similarity. - second_node_names: List[str] - The second node for which to compute the similarity. - """ - return self._model.get_similarity_from_node_names(first_node_names, second_node_names) - - def get_pairwise_similarities( - self, - graph: Optional[Graph] = None, - return_similarities_dataframe: bool = False - ) -> Union[pd.DataFrame, np.ndarray]: - """Execute similarities on the provided graph. - + edge_node_ids: np.ndarray, + similarities: np.ndarray, + return_similarities_dataframe: bool = False, + return_node_names: bool = False + ) -> Union[pd.DataFrame, Tuple[np.ndarray, np.ndarray]]: + """Normalize output to provided standard. + Parameters - -------------------- - graph: Optional[Graph] = None - The graph to run similarities on. + --------------------- + edge_node_ids: np.ndarray + The edge node IDs composing the edges. + similarities: np.ndarray + Resnik similarity scores. return_similarities_dataframe: bool = False - Whether to return a pandas DataFrame, which as indices has the node IDs. - By default, a numpy array with the similarities is returned as it weights much less. + Whether to return the data as a DataFrame. + Do note that this will require more RAM. + return_node_names: bool = False + Whether to return the node names. + Do note that this will require SIGNIFICANTLY more RAM. """ - similarities = self._model.get_pairwise_similarities() - - if return_similarities_dataframe and graph is not None: - similarities = pd.DataFrame( - similarities, - columns=graph.get_node_names(), - index=graph.get_node_names(), + if return_node_names and not return_similarities_dataframe: + raise NotImplementedError( + "It is not currently supported to return the node names " + "when it is not requested to return a pandas DataFrame. " + "This is not supported as the RAM requirements for common " + "use cases are large enough to be unfeaseable on most systems." ) - return similarities + if not return_similarities_dataframe: + return (edge_node_ids, similarities) - def get_similarities_from_graph( - self, - graph: Graph, - return_similarities_dataframe: bool = False - ) -> Union[pd.DataFrame, np.ndarray]: - """Execute similarities on the provided graph. - - Parameters - -------------------- - graph: Graph - The graph to run similarities on. - node_frequencies: Optional[np.ndarray] - Optional vector of node frequencies. - return_similarities_dataframe: bool = False - Whether to return a pandas DataFrame, which as indices has the node IDs. - By default, a numpy array with the similarities is returned as it weights much less. - """ - similarities = self._model.get_similarities_from_graph( - graph, - ) - - if return_similarities_dataframe: - similarities = pd.DataFrame( - { - "similarities": similarities, - "sources": graph.get_directed_source_node_ids(), - "destinations": graph.get_directed_destination_node_ids(), - }, - ) - - return similarities + return pd.DataFrame({ + "source": ( + self._graph.get_node_names_from_node_ids(edge_node_ids[:, 0]) + if return_node_names + else edge_node_ids[:, 0] + ), + "destination": ( + self._graph.get_node_names_from_node_ids(edge_node_ids[:, 1]) + if return_node_names + else edge_node_ids[:, 1] + ), + "resnik_score": similarities + }) - def get_similarities_from_bipartite_graph_from_edge_node_ids( + def get_similarities_from_bipartite_graph_node_ids( self, - graph: Graph, - source_node_ids: List[int], - destination_node_ids: List[int], - return_similarities_dataframe: bool = False - ) -> Union[pd.DataFrame, np.ndarray]: - """Execute similarities probabilities on the provided graph bipartite portion. + source_node_ids: List[str], + destination_node_ids: List[str], + minimum_similarity: Optional[float] = 0.0, + return_similarities_dataframe: bool = False, + return_node_names: bool = False + ) -> Union[pd.DataFrame, Tuple[np.ndarray, np.ndarray]]: + """Execute similarities probabilities on the bipartite portion Parameters -------------------- - graph: Graph - The graph from which to extract the edges. - node_frequencies: Optional[np.ndarray] - Optional vector of node frequencies. source_node_ids: List[int] - The source nodes of the bipartite graph. + The source node ids defining a bipartite graph. destination_node_ids: List[int] - The destination nodes of the bipartite graph. + The destination node ids defining a bipartite graph. + minimum_similarity: Optional[float] = 0.0 + Minimum similarity to be kept. Values below this amount are filtered. return_similarities_dataframe: bool = False Whether to return a pandas DataFrame, which as indices has the node IDs. - By default, a numpy array with the similarities is returned as it weights much less. + By default, a numpy array with the similarities is returned as requires much less RAM. + return_node_names: bool = False + Whether to return the node names or node IDs associated to the scores. + By default we return the node ids, which require much less memory. """ - return self.get_similarities_from_graph( - graph.build_bipartite_graph_from_edge_node_ids( - source_node_ids=source_node_ids, - destination_node_ids=destination_node_ids, - directed=True + return self._normalize_output( + self._model.get_node_ids_and_similarity_from_node_ids( + first_node_ids=source_node_ids, + second_node_ids=destination_node_ids, + minimum_similarity=minimum_similarity ), - return_similarities_dataframe=return_similarities_dataframe + return_similarities_dataframe=return_similarities_dataframe, + return_node_names=return_node_names ) - def get_similarities_from_bipartite_graph_from_edge_node_names( + def get_similarities_from_bipartite_graph_node_names( self, - graph: Graph, source_node_names: List[str], destination_node_names: List[str], - return_similarities_dataframe: bool = False - ) -> Union[pd.DataFrame, np.ndarray]: - """Execute similarities probabilities on the provided graph bipartite portion. + minimum_similarity: Optional[float] = 0.0, + return_similarities_dataframe: bool = False, + return_node_names: bool = False + ) -> Union[pd.DataFrame, Tuple[np.ndarray, np.ndarray]]: + """Execute similarities probabilities on the bipartite portion Parameters -------------------- - graph: Graph - The graph from which to extract the edges. - node_frequencies: Optional[np.ndarray] - Optional vector of node frequencies. source_node_names: List[str] - The source nodes of the bipartite graph. + The source node names defining a bipartite graph. destination_node_names: List[str] - The destination nodes of the bipartite graph. + The destination node names defining a bipartite graph. + minimum_similarity: Optional[float] = 0.0 + Minimum similarity to be kept. Values below this amount are filtered. return_similarities_dataframe: bool = False Whether to return a pandas DataFrame, which as indices has the node IDs. - By default, a numpy array with the similarities is returned as it weights much less. + By default, a numpy array with the similarities is returned as requires much less RAM. + return_node_names: bool = False + Whether to return the node names or node IDs associated to the scores. + By default we return the node ids, which require much less memory. """ - return self.get_similarities_from_graph( - graph.build_bipartite_graph_from_edge_node_names( - source_node_names=source_node_names, - destination_node_names=destination_node_names, - directed=True + return self._normalize_output( + self._model.get_node_ids_and_similarity_from_node_names( + first_node_names=source_node_names, + second_node_names=destination_node_names, + minimum_similarity=minimum_similarity ), - return_similarities_dataframe=return_similarities_dataframe + return_similarities_dataframe=return_similarities_dataframe, + return_node_names=return_node_names ) - def get_similarities_from_bipartite_graph_from_edge_node_prefixes( + def get_similarities_from_bipartite_graph_node_prefixes( self, - graph: Graph, source_node_prefixes: List[str], destination_node_prefixes: List[str], - return_similarities_dataframe: bool = False - ) -> Union[pd.DataFrame, np.ndarray]: - """Execute similarities probabilities on the provided graph bipartite portion. + minimum_similarity: Optional[float] = 0.0, + return_similarities_dataframe: bool = False, + return_node_names: bool = False + ) -> Union[pd.DataFrame, Tuple[np.ndarray, np.ndarray]]: + """Execute similarities probabilities on the bipartite portion Parameters -------------------- - graph: Graph - The graph from which to extract the edges. - node_frequencies: Optional[np.ndarray] - Optional vector of node frequencies. source_node_prefixes: List[str] - The source node prefixes of the bipartite graph. + The source node prefixes defining a bipartite graph. destination_node_prefixes: List[str] - The destination node prefixes of the bipartite graph. + The destination node prefixes defining a bipartite graph. + minimum_similarity: Optional[float] = 0.0 + Minimum similarity to be kept. Values below this amount are filtered. return_similarities_dataframe: bool = False Whether to return a pandas DataFrame, which as indices has the node IDs. - By default, a numpy array with the similarities is returned as it weights much less. + By default, a numpy array with the similarities is returned as requires much less RAM. + return_node_names: bool = False + Whether to return the node names or node IDs associated to the scores. + By default we return the node ids, which require much less memory. """ - return self.get_similarities_from_graph( - graph.build_bipartite_graph_from_edge_node_prefixes( - source_node_prefixes=source_node_prefixes, - destination_node_prefixes=destination_node_prefixes, - directed=True + return self._normalize_output( + self._model.get_node_ids_and_similarity_from_node_prefixes( + first_node_prefixes=source_node_prefixes, + second_node_prefixes=destination_node_prefixes, + minimum_similarity=minimum_similarity ), - return_similarities_dataframe=return_similarities_dataframe + return_similarities_dataframe=return_similarities_dataframe, + return_node_names=return_node_names ) - def get_similarities_from_bipartite_graph_from_edge_node_types( + def get_similarities_from_bipartite_graph_node_type_ids( self, - graph: Graph, - source_node_types: List[str], - destination_node_types: List[str], + source_node_type_ids: List[int], + destination_node_type_ids: List[int], + minimum_similarity: Optional[float] = 0.0, + return_similarities_dataframe: bool = False, + return_node_names: bool = False + ) -> Union[pd.DataFrame, Tuple[np.ndarray, np.ndarray]]: + """Execute similarities probabilities on the bipartite portion + + Parameters + -------------------- + source_node_type_ids: List[int] + The source node type ids defining a bipartite graph. + destination_node_type_ids: List[int] + The destination node type ids defining a bipartite graph. + minimum_similarity: Optional[float] = 0.0 + Minimum similarity to be kept. Values below this amount are filtered. return_similarities_dataframe: bool = False - ) -> Union[pd.DataFrame, np.ndarray]: + Whether to return a pandas DataFrame, which as indices has the node IDs. + By default, a numpy array with the similarities is returned as requires much less RAM. + return_node_names: bool = False + Whether to return the node names or node IDs associated to the scores. + By default we return the node ids, which require much less memory. + """ + return self._normalize_output( + self._model.get_node_ids_and_similarity_from_node_type_ids( + first_node_type_ids=source_node_type_ids, + second_node_type_ids=destination_node_type_ids, + minimum_similarity=minimum_similarity + ), + return_similarities_dataframe=return_similarities_dataframe, + return_node_names=return_node_names + ) + + def get_similarities_from_bipartite_graph_node_type_names( + self, + source_node_type_names: List[str], + destination_node_type_names: List[str], + minimum_similarity: Optional[float] = 0.0, + return_similarities_dataframe: bool = False, + return_node_names: bool = False + ) -> Union[pd.DataFrame, Tuple[np.ndarray, np.ndarray]]: """Execute similarities probabilities on the provided graph bipartite portion. Parameters -------------------- - graph: Graph - The graph from which to extract the edges. - node_frequencies: Optional[np.ndarray] - Optional vector of node frequencies. - source_node_types: List[str] - The source node prefixes of the bipartite graph. - destination_node_types: List[str] - The destination node prefixes of the bipartite graph. + source_node_type_names: List[str] + The source node type names defining a bipartite graph. + destination_node_type_names: List[str] + The destination node type names defining a bipartite graph. + minimum_similarity: Optional[float] = 0.0 + Minimum similarity to be kept. Values below this amount are filtered. return_similarities_dataframe: bool = False Whether to return a pandas DataFrame, which as indices has the node IDs. - By default, a numpy array with the similarities is returned as it weights much less. + By default, a numpy array with the similarities is returned as requires much less RAM. + return_node_names: bool = False + Whether to return the node names or node IDs associated to the scores. + By default we return the node ids, which require much less memory. """ - return self.get_similarities_from_graph( - graph.build_bipartite_graph_from_edge_node_types( - source_node_types=source_node_types, - destination_node_types=destination_node_types, - directed=True + return self._normalize_output( + self._model.get_node_ids_and_similarity_from_node_type_names( + first_node_type_names=source_node_type_names, + second_node_type_names=destination_node_type_names, + minimum_similarity=minimum_similarity ), - return_similarities_dataframe=return_similarities_dataframe + return_similarities_dataframe=return_similarities_dataframe, + return_node_names=return_node_names ) - def get_similarities_from_clique_graph_from_node_ids( + def get_similarities_from_clique_graph_node_ids( self, - graph: Graph, - node_ids: List[int], - return_similarities_dataframe: bool = False - ) -> Union[pd.DataFrame, np.ndarray]: - """Execute similarities probabilities on the provided graph bipartite portion. + node_ids: List[str], + minimum_similarity: Optional[float] = 0.0, + return_similarities_dataframe: bool = False, + return_node_names: bool = False + ) -> Union[pd.DataFrame, Tuple[np.ndarray, np.ndarray]]: + """Execute similarities probabilities on the clique portion Parameters -------------------- - graph: Graph - The graph from which to extract the edges. - node_frequencies: Optional[np.ndarray] - Optional vector of node frequencies. node_ids: List[int] - The nodes of the bipartite graph. + The node type ids defining a clique. + minimum_similarity: Optional[float] = 0.0 + Minimum similarity to be kept. Values below this amount are filtered. return_similarities_dataframe: bool = False Whether to return a pandas DataFrame, which as indices has the node IDs. - By default, a numpy array with the similarities is returned as it weights much less. + By default, a numpy array with the similarities is returned as requires much less RAM. + return_node_names: bool = False + Whether to return the node names or node IDs associated to the scores. + By default we return the node ids, which require much less memory. """ - return self.get_similarities_from_graph( - graph.build_clique_graph_from_node_ids( - node_ids=node_ids, - directed=True - ), - return_similarities_dataframe=return_similarities_dataframe + return self.get_similarities_from_bipartite_graph_node_ids( + source_node_ids=node_ids, + destination_node_ids=node_ids, + minimum_similarity=minimum_similarity, + return_similarities_dataframe=return_similarities_dataframe, + return_node_names=return_node_names ) - def get_similarities_from_clique_graph_from_node_names( + def get_similarities_from_clique_graph_node_names( self, - graph: Graph, node_names: List[str], - return_similarities_dataframe: bool = False - ) -> Union[pd.DataFrame, np.ndarray]: - """Execute similarities probabilities on the provided graph bipartite portion. + minimum_similarity: Optional[float] = 0.0, + return_similarities_dataframe: bool = False, + return_node_names: bool = False + ) -> Union[pd.DataFrame, Tuple[np.ndarray, np.ndarray]]: + """Execute similarities probabilities on the clique portion Parameters -------------------- - graph: Graph - The graph from which to extract the edges. - node_frequencies: Optional[np.ndarray] - Optional vector of node frequencies. node_names: List[str] - The nodes of the bipartite graph. + The node type ids defining a clique. + minimum_similarity: Optional[float] = 0.0 + Minimum similarity to be kept. Values below this amount are filtered. return_similarities_dataframe: bool = False Whether to return a pandas DataFrame, which as indices has the node IDs. - By default, a numpy array with the similarities is returned as it weights much less. + By default, a numpy array with the similarities is returned as requires much less RAM. + return_node_names: bool = False + Whether to return the node names or node IDs associated to the scores. + By default we return the node ids, which require much less memory. """ - return self.get_similarities_from_graph( - graph.build_clique_graph_from_node_names( - node_names=node_names, - directed=True - ), - return_similarities_dataframe=return_similarities_dataframe + return self.get_similarities_from_bipartite_graph_node_names( + source_node_names=node_names, + destination_node_names=node_names, + minimum_similarity=minimum_similarity, + return_similarities_dataframe=return_similarities_dataframe, + return_node_names=return_node_names ) - def get_similarities_from_clique_graph_from_node_prefixes( + def get_similarities_from_clique_graph_node_prefixes( self, - graph: Graph, node_prefixes: List[str], - return_similarities_dataframe: bool = False - ) -> Union[pd.DataFrame, np.ndarray]: - """Execute similarities probabilities on the provided graph bipartite portion. + minimum_similarity: Optional[float] = 0.0, + return_similarities_dataframe: bool = False, + return_node_names: bool = False + ) -> Union[pd.DataFrame, Tuple[np.ndarray, np.ndarray]]: + """Execute similarities probabilities on the clique portion Parameters -------------------- - graph: Graph - The graph from which to extract the edges. - node_frequencies: Optional[np.ndarray] - Optional vector of node frequencies. node_prefixes: List[str] - The node prefixes of the bipartite graph. + The node type ids defining a clique. + minimum_similarity: Optional[float] = 0.0 + Minimum similarity to be kept. Values below this amount are filtered. return_similarities_dataframe: bool = False Whether to return a pandas DataFrame, which as indices has the node IDs. - By default, a numpy array with the similarities is returned as it weights much less. + By default, a numpy array with the similarities is returned as requires much less RAM. + return_node_names: bool = False + Whether to return the node names or node IDs associated to the scores. + By default we return the node ids, which require much less memory. """ - return self.get_similarities_from_graph( - graph.build_clique_graph_from_node_prefixes( - node_prefixes=node_prefixes, - directed=True - ), - return_similarities_dataframe=return_similarities_dataframe + return self.get_similarities_from_bipartite_graph_node_prefixes( + source_node_prefixes=node_prefixes, + destination_node_prefixes=node_prefixes, + minimum_similarity=minimum_similarity, + return_similarities_dataframe=return_similarities_dataframe, + return_node_names=return_node_names ) - def get_similarities_from_clique_graph_from_node_types( + def get_similarities_from_clique_graph_node_type_ids( self, - graph: Graph, - node_types: List[str], + node_type_ids: List[int], + minimum_similarity: Optional[float] = 0.0, + return_similarities_dataframe: bool = False, + return_node_names: bool = False + ) -> Union[pd.DataFrame, Tuple[np.ndarray, np.ndarray]]: + """Execute similarities probabilities on the clique portion + + Parameters + -------------------- + node_type_ids: List[int] + The node type ids defining a clique. + minimum_similarity: Optional[float] = 0.0 + Minimum similarity to be kept. Values below this amount are filtered. return_similarities_dataframe: bool = False - ) -> Union[pd.DataFrame, np.ndarray]: - """Execute similarities probabilities on the provided graph bipartite portion. + Whether to return a pandas DataFrame, which as indices has the node IDs. + By default, a numpy array with the similarities is returned as requires much less RAM. + return_node_names: bool = False + Whether to return the node names or node IDs associated to the scores. + By default we return the node ids, which require much less memory. + """ + return self.get_similarities_from_bipartite_graph_node_type_ids( + source_node_type_ids=node_type_ids, + destination_node_type_ids=node_type_ids, + minimum_similarity=minimum_similarity, + return_similarities_dataframe=return_similarities_dataframe, + return_node_names=return_node_names + ) + + def get_similarities_from_clique_graph_node_type_names( + self, + node_type_names: List[str], + minimum_similarity: Optional[float] = 0.0, + return_similarities_dataframe: bool = False, + return_node_names: bool = False + ) -> Union[pd.DataFrame, Tuple[np.ndarray, np.ndarray]]: + """Execute similarities probabilities on the provided graph clique portion. Parameters -------------------- - graph: Graph - The graph from which to extract the edges. - node_frequencies: Optional[np.ndarray] - Optional vector of node frequencies. - node_types: List[str] - The node prefixes of the bipartite graph. + node_type_names: List[str] + The node type names defining a clique. + minimum_similarity: Optional[float] = 0.0 + Minimum similarity to be kept. Values below this amount are filtered. return_similarities_dataframe: bool = False Whether to return a pandas DataFrame, which as indices has the node IDs. - By default, a numpy array with the similarities is returned as it weights much less. + By default, a numpy array with the similarities is returned as requires much less RAM. + return_node_names: bool = False + Whether to return the node names or node IDs associated to the scores. + By default we return the node ids, which require much less memory. """ - return self.get_similarities_from_graph( - graph.build_clique_graph_from_node_types( - node_types=node_types, - directed=True - ), - return_similarities_dataframe=return_similarities_dataframe + return self.get_similarities_from_bipartite_graph_node_type_names( + source_node_type_names=node_type_names, + destination_node_type_names=node_type_names, + minimum_similarity=minimum_similarity, + return_similarities_dataframe=return_similarities_dataframe, + return_node_names=return_node_names ) diff --git a/embiggen/utils/abstract_models/abstract_classifier_model.py b/embiggen/utils/abstract_models/abstract_classifier_model.py index c131c9bb..e3548115 100644 --- a/embiggen/utils/abstract_models/abstract_classifier_model.py +++ b/embiggen/utils/abstract_models/abstract_classifier_model.py @@ -3,9 +3,13 @@ from ensmallen import Graph, express_measures import numpy as np import pandas as pd +import os +import platform import time +import json from userinput.utils import must_be_in_set from tqdm.auto import trange, tqdm +from environments_utils import get_slurm_node_id, must_be_in_slurm_node from embiggen.utils.abstract_models.list_formatting import format_list from cache_decorator import Cache @@ -156,6 +160,35 @@ def get_available_evaluation_schemas(cls) -> List[str]: "in the child classes of abstract model." )) + @classmethod + def load(cls, path: str) -> "Self": + """Load a saved version of the model from the provided path. + + Parameters + ------------------- + path: str + Path from where to load the model. + """ + raise NotImplementedError(( + f"The `load` was not implemented for the {cls.model_name()} " + f"for the {cls.task_name()} task as made available from the " + f"{cls.library_name()} library." + )) + + def dump(self, path: str): + """Dump the current model at the provided path. + + Parameters + ------------------- + path: str + Path from where to dump the model. + """ + raise NotImplementedError(( + f"The `dump` was not implemented for the {self.model_name()} " + f"for the {self.task_name()} task as made available from the " + f"{self.library_name()} library." + )) + @classmethod def normalize_node_feature( cls, @@ -165,7 +198,7 @@ def normalize_node_feature( allow_automatic_feature: bool = True, skip_evaluation_biased_feature: bool = False, smoke_test: bool = False, - precompute_constant_automatic_stocastic_features: bool = False + precompute_constant_stocastic_features: bool = False ) -> List[np.ndarray]: """Normalizes the provided node features and validates them. @@ -190,7 +223,7 @@ def normalize_node_feature( Whether this run should be considered a smoke test and therefore use the smoke test configurations for the provided model names and feature names. - precompute_constant_automatic_stocastic_features: bool = False + precompute_constant_stocastic_features: bool = False Whether to precompute once the constant automatic stocastic features before starting the embedding loop. This means that, when left set to false, while the features will be computed @@ -230,7 +263,7 @@ def normalize_node_feature( cls.task_involves_edge_weights() and node_feature.can_use_edge_weights() and node_feature.is_using_edge_weights() or cls.task_involves_topology() and node_feature.is_topological() ) or - not precompute_constant_automatic_stocastic_features and node_feature.is_stocastic() + not precompute_constant_stocastic_features and node_feature.is_stocastic() ): yield node_feature return None @@ -298,7 +331,7 @@ def normalize_node_features( allow_automatic_feature: bool = True, skip_evaluation_biased_feature: bool = False, smoke_test: bool = False, - precompute_constant_automatic_stocastic_features: bool = False + precompute_constant_stocastic_features: bool = False ) -> List[np.ndarray]: """Normalizes the provided node features and validates them. @@ -323,7 +356,7 @@ def normalize_node_features( Whether this run should be considered a smoke test and therefore use the smoke test configurations for the provided model names and feature names. - precompute_constant_automatic_stocastic_features: bool = False + precompute_constant_stocastic_features: bool = False Whether to precompute once the constant automatic stocastic features before starting the embedding loop. This means that, when left set to false, while the features will be computed @@ -349,7 +382,7 @@ def normalize_node_features( allow_automatic_feature=allow_automatic_feature, skip_evaluation_biased_feature=skip_evaluation_biased_feature, smoke_test=smoke_test, - precompute_constant_automatic_stocastic_features=precompute_constant_automatic_stocastic_features + precompute_constant_stocastic_features=precompute_constant_stocastic_features ) ] @@ -362,7 +395,7 @@ def normalize_node_type_feature( allow_automatic_feature: bool = True, skip_evaluation_biased_feature: bool = False, smoke_test: bool = False, - precompute_constant_automatic_stocastic_features: bool = False + precompute_constant_stocastic_features: bool = False ) -> List[np.ndarray]: """Normalizes the provided node type features and validates them. @@ -387,7 +420,7 @@ def normalize_node_type_feature( Whether this run should be considered a smoke test and therefore use the smoke test configurations for the provided model names and feature names. - precompute_constant_automatic_stocastic_features: bool = False + precompute_constant_stocastic_features: bool = False Whether to precompute once the constant automatic stocastic features before starting the embedding loop. This means that, when left set to false, while the features will be computed @@ -427,11 +460,11 @@ def normalize_node_type_feature( cls.task_involves_edge_weights() and node_type_feature.is_using_edge_weights() or cls.task_involves_topology() and node_type_feature.is_topological() ) or - not precompute_constant_automatic_stocastic_features and node_type_feature.is_stocastic() + not precompute_constant_stocastic_features and node_type_feature.is_stocastic() ): yield node_type_feature return None - + if node_type_feature.is_stocastic(): node_type_feature.set_random_state(random_state) @@ -495,7 +528,7 @@ def normalize_node_type_features( allow_automatic_feature: bool = True, skip_evaluation_biased_feature: bool = False, smoke_test: bool = False, - precompute_constant_automatic_stocastic_features: bool = False + precompute_constant_stocastic_features: bool = False ) -> List[np.ndarray]: """Normalizes the provided node type features and validates them. @@ -520,7 +553,7 @@ def normalize_node_type_features( Whether this run should be considered a smoke test and therefore use the smoke test configurations for the provided model names and feature names. - precompute_constant_automatic_stocastic_features: bool = False + precompute_constant_stocastic_features: bool = False Whether to precompute once the constant automatic stocastic features before starting the embedding loop. This means that, when left set to false, while the features will be computed @@ -546,7 +579,7 @@ def normalize_node_type_features( allow_automatic_feature=allow_automatic_feature, skip_evaluation_biased_feature=skip_evaluation_biased_feature, smoke_test=smoke_test, - precompute_constant_automatic_stocastic_features=precompute_constant_automatic_stocastic_features + precompute_constant_stocastic_features=precompute_constant_stocastic_features ) ] @@ -559,7 +592,7 @@ def normalize_edge_feature( allow_automatic_feature: bool = True, skip_evaluation_biased_feature: bool = False, smoke_test: bool = False, - precompute_constant_automatic_stocastic_features: bool = False + precompute_constant_stocastic_features: bool = False ) -> List[np.ndarray]: """Normalizes the provided edge features and validates them. @@ -584,7 +617,7 @@ def normalize_edge_feature( Whether this run should be considered a smoke test and therefore use the smoke test configurations for the provided model names and feature names. - precompute_constant_automatic_stocastic_features: bool = False + precompute_constant_stocastic_features: bool = False Whether to precompute once the constant automatic stocastic features before starting the embedding loop. This means that, when left set to false, while the features will be computed @@ -624,7 +657,7 @@ def normalize_edge_feature( cls.task_involves_edge_weights() and edge_feature.is_using_edge_weights() or cls.task_involves_topology() and edge_feature.is_topological() ) or - not precompute_constant_automatic_stocastic_features and edge_feature.is_stocastic() + not precompute_constant_stocastic_features and edge_feature.is_stocastic() ): yield edge_feature return None @@ -692,7 +725,7 @@ def normalize_edge_features( allow_automatic_feature: bool = True, skip_evaluation_biased_feature: bool = False, smoke_test: bool = False, - precompute_constant_automatic_stocastic_features: bool = False + precompute_constant_stocastic_features: bool = False ) -> List[np.ndarray]: """Normalizes the provided edge features and validates them. @@ -717,7 +750,7 @@ def normalize_edge_features( Whether this run should be considered a smoke test and therefore use the smoke test configurations for the provided model names and feature names. - precompute_constant_automatic_stocastic_features: bool = False + precompute_constant_stocastic_features: bool = False Whether to precompute once the constant automatic stocastic features before starting the embedding loop. This means that, when left set to false, while the features will be computed @@ -743,7 +776,7 @@ def normalize_edge_features( allow_automatic_feature=allow_automatic_feature, skip_evaluation_biased_feature=skip_evaluation_biased_feature, smoke_test=smoke_test, - precompute_constant_automatic_stocastic_features=precompute_constant_automatic_stocastic_features + precompute_constant_stocastic_features=precompute_constant_stocastic_features ) ] @@ -867,7 +900,7 @@ def predict( if support is None: support = graph - + try: predictions = self._predict( graph=graph, @@ -897,7 +930,7 @@ def predict( f"implemented using the {self.library_name()} for the {self.task_name()} task. " f"Specifically, the class of the model is {self.__class__.__name__}." ) from e - + return predictions def predict_proba( @@ -982,7 +1015,7 @@ def predict_proba( "returned a vector of prediction probabilities with " f"shape {predictions.shape}." ) - + return predictions def evaluate_predictions( @@ -1230,25 +1263,35 @@ def iterate_classifier_models( cache_path="{cache_dir}/{self.task_name()}/{graph.get_name()}/holdout_{holdout_number}/{self.model_name()}/{self.library_name()}/{_hash}.csv.gz", cache_dir="experiments", enable_cache_arg_name="enable_cache", - args_to_ignore=["verbose", "smoke_test", "train_of_interest", "test_of_interest", "train",], + args_to_ignore=[ + "verbose", + "smoke_test", + "train_of_interest", + "test_of_interest", + "train", + "metadata" + ], capture_enable_cache_arg_name=True, use_approximated_hash=True ) - def __train_and_evaluate_model( + def _train_and_evaluate_model( self, graph: Graph, train_of_interest: Graph, test_of_interest: Graph, train: Graph, subgraph_of_interest: Graph, + use_subgraph_as_support: bool, node_features: Optional[List[np.ndarray]], node_type_features: Optional[List[np.ndarray]], edge_features: Optional[List[np.ndarray]], random_state: int, holdout_number: int, evaluation_schema: str, - automatic_features_names: List[str], - automatic_features_parameters: Dict[str, Any], + holdouts_kwargs: Dict[str, Any], + features_names: List[str], + features_parameters: Dict[str, Any], + metadata: Dict[str, Any], *validation_kwargs ) -> pd.DataFrame: """Run inner training and evaluation.""" @@ -1258,7 +1301,7 @@ def __train_and_evaluate_model( training_start = time.time() self.fit( graph=train_of_interest, - support=train, + support=train_of_interest if use_subgraph_as_support else train, node_features=node_features, node_type_features=node_type_features, edge_features=edge_features @@ -1271,7 +1314,7 @@ def __train_and_evaluate_model( # We add the newly computed performance. model_performance = pd.DataFrame(self._evaluate( graph=graph, - support=train, + support=train_of_interest if use_subgraph_as_support else train, train=train_of_interest, test=test_of_interest, node_features=node_features, @@ -1281,7 +1324,7 @@ def __train_and_evaluate_model( random_state=random_state holdout_number, verbose=False, *validation_kwargs - )) + )).reset_index(drop=True) except RuntimeError as e: raise e except Exception as e: @@ -1302,26 +1345,34 @@ def __train_and_evaluate_model( model_performance["nodes_number"] = graph.get_number_of_nodes() model_performance["edges_number"] = graph.get_number_of_directed_edges() model_performance["evaluation_schema"] = evaluation_schema + model_performance["holdout_number"] = holdout_number + model_performance["holdouts_kwargs"] = json.dumps(holdouts_kwargs) + model_performance["use_subgraph_as_support"] = use_subgraph_as_support + + for column_name, column_value in metadata.items(): + model_performance[column_name] = column_value + + df_model_parameters = pd.DataFrame( + dict(), index=model_performance.index) + df_features_parameters = pd.DataFrame( + dict(), index=model_performance.index) for parameter_name, parameter_value in self.parameters().items(): - if ("Model", parameter_name) in model_performance.columns: - raise ValueError( - "There has been a collision between the column names used in " - "the model performance report and the parameter names " - f" of one of the classifiers {self.model_name()}." - f"The parameter that has caused the collision is {parameter}. " - "Please do change the name of the parameter in your model." - ) if isinstance(parameter_value, (list, tuple)): parameter_value = str(parameter_value) - model_performance[("Model", parameter_name)] = parameter_value - - model_performance["automatic_features_names"] = format_list( - automatic_features_names + df_model_parameters[parameter_name] = parameter_value + + df_model_parameters.columns = [ + ["model_parameters"] len(df_model_parameters.columns), + df_model_parameters.columns + ] + + model_performance["features_names"] = format_list( + features_names ) - - for parameter, value in enumerate(automatic_features_parameters.items()): - if parameter in model_performance.columns: + + for parameter, value in features_parameters.items(): + if parameter in df_features_parameters.columns: raise ValueError( "There has been a collision between the parameters used in " "one of the embedding models and the parameter " @@ -1329,7 +1380,21 @@ def __train_and_evaluate_model( f"The parameter that has caused the collision is {parameter}. " "Please do change the name of the parameter in your model." ) - model_performance[parameter] = str(value) + df_features_parameters[parameter] = str(value) + + df_features_parameters.columns = [ + ["features_parameters"] * len(df_features_parameters.columns), + df_features_parameters.columns + ] + + model_performance = pd.concat( + [ + model_performance, + df_model_parameters, + df_features_parameters + ], + axis=1 + ) return model_performance @@ -1338,16 +1403,22 @@ def __train_and_evaluate_model( cache_path="{cache_dir}/{cls.task_name()}/{graph.get_name()}/holdout_{holdout_number}/{_hash}.csv.gz", cache_dir="experiments", enable_cache_arg_name="enable_cache", - args_to_ignore=["verbose", "smoke_test", "number_of_holdouts"], + args_to_ignore=[ + "verbose", + "smoke_test", + "number_of_holdouts", + "metadata" + ], capture_enable_cache_arg_name=False, use_approximated_hash=True ) - def __evaluate_on_single_holdout( + def _evaluate_on_single_holdout( cls, models: Union[Type["AbstractClassifierModel"], List[Type["AbstractClassifierModel"]]], library_names: Optional[Union[str, List[str]]], graph: Graph, subgraph_of_interest: Graph, + use_subgraph_as_support: bool, node_features: Optional[List[np.ndarray]], node_type_features: Optional[List[np.ndarray]], edge_features: Optional[List[np.ndarray]], @@ -1359,8 +1430,9 @@ def __evaluate_on_single_holdout( smoke_test: bool, holdouts_kwargs: Dict[str, Any], subgraph_of_interest_has_compatible_nodes: Optional[bool], - automatic_features_names: List[str], - automatic_features_parameters: Dict[str, Any], + features_names: List[str], + features_parameters: Dict[str, Any], + metadata: Dict[str, Any], verbose: bool, *validation_kwargs ) -> pd.DataFrame: @@ -1384,14 +1456,9 @@ def __evaluate_on_single_holdout( vector_cumulative_node_degrees=graph.has_cumulative_node_degrees_tradeoff_enabled(), vector_reciprocal_sqrt_degrees=graph.has_reciprocal_sqrt_degrees_tradeoff_enabled() ) - test.enable( - vector_sources=graph.has_sources_tradeoff_enabled(), - vector_destinations=graph.has_destinations_tradeoff_enabled(), - vector_cumulative_node_degrees=graph.has_cumulative_node_degrees_tradeoff_enabled(), - vector_reciprocal_sqrt_degrees=graph.has_reciprocal_sqrt_degrees_tradeoff_enabled() - ) # We compute the remaining features + starting_to_compute_node_features = time.time() holdout_node_features = cls.normalize_node_features( train, random_state=random_state(holdout_number+1), @@ -1399,10 +1466,13 @@ def __evaluate_on_single_holdout( allow_automatic_feature=True, skip_evaluation_biased_feature=False, smoke_test=smoke_test, - precompute_constant_automatic_stocastic_features=True + precompute_constant_stocastic_features=True ) + time_required_to_compute_node_features = time.time() - \ + starting_to_compute_node_features # We compute the remaining features + starting_to_compute_node_type_features = time.time() holdout_node_type_features = cls.normalize_node_type_features( train, random_state=random_state(holdout_number+1), @@ -1410,12 +1480,15 @@ def __evaluate_on_single_holdout( allow_automatic_feature=True, skip_evaluation_biased_feature=False, smoke_test=smoke_test, - precompute_constant_automatic_stocastic_features=True + precompute_constant_stocastic_features=True ) + time_required_to_compute_node_type_features = time.time( + ) - starting_to_compute_node_type_features # We execute the same thing as described above, # but now for the edge features instead that for # the node features. + starting_to_compute_edge_features = time.time() holdout_edge_features = cls.normalize_edge_features( train, random_state=random_state(holdout_number+1), @@ -1423,8 +1496,10 @@ def __evaluate_on_single_holdout( allow_automatic_feature=True, skip_evaluation_biased_feature=False, smoke_test=smoke_test, - precompute_constant_automatic_stocastic_features=True + precompute_constant_stocastic_features=True ) + time_required_to_compute_edge_features = time.time() - \ + starting_to_compute_edge_features if subgraph_of_interest is not None: # First we align the train and test graph to have @@ -1439,22 +1514,6 @@ def __evaluate_on_single_holdout( node_names_to_keep_from_graph=subgraph_of_interest ) - # We enable in the train and test graphs of interest the same - # speedups enabled in the provided graph. - train.enable( - vector_sources=graph.has_sources_tradeoff_enabled(), - vector_destinations=graph.has_destinations_tradeoff_enabled(), - vector_cumulative_node_degrees=graph.has_cumulative_node_degrees_tradeoff_enabled(), - vector_reciprocal_sqrt_degrees=graph.has_reciprocal_sqrt_degrees_tradeoff_enabled() - ) - - test.enable( - vector_sources=graph.has_sources_tradeoff_enabled(), - vector_destinations=graph.has_destinations_tradeoff_enabled(), - vector_cumulative_node_degrees=graph.has_cumulative_node_degrees_tradeoff_enabled(), - vector_reciprocal_sqrt_degrees=graph.has_reciprocal_sqrt_degrees_tradeoff_enabled() - ) - # We adjust the node features to only include the node features # that the subgraph of interest allows us to use. if holdout_node_features is not None: @@ -1493,25 +1552,26 @@ def __evaluate_on_single_holdout( "subgraph of interest, does not have any more edges which are " f"essential when running a {cls.task_name()} task." ) - # We enable in the train and test graphs of interest the same - # speedups enabled in the provided graph. - train_of_interest.enable( - vector_sources=graph.has_sources_tradeoff_enabled(), - vector_destinations=graph.has_destinations_tradeoff_enabled(), - vector_cumulative_node_degrees=graph.has_cumulative_node_degrees_tradeoff_enabled(), - vector_reciprocal_sqrt_degrees=graph.has_reciprocal_sqrt_degrees_tradeoff_enabled() - ) - - test_of_interest.enable( - vector_sources=graph.has_sources_tradeoff_enabled(), - vector_destinations=graph.has_destinations_tradeoff_enabled(), - vector_cumulative_node_degrees=graph.has_cumulative_node_degrees_tradeoff_enabled(), - vector_reciprocal_sqrt_degrees=graph.has_reciprocal_sqrt_degrees_tradeoff_enabled() - ) else: train_of_interest = train test_of_interest = test + # We enable in the train and test graphs of interest the same + # speedups enabled in the provided graph. + train_of_interest.enable( + vector_sources=graph.has_sources_tradeoff_enabled(), + vector_destinations=graph.has_destinations_tradeoff_enabled(), + vector_cumulative_node_degrees=graph.has_cumulative_node_degrees_tradeoff_enabled(), + vector_reciprocal_sqrt_degrees=graph.has_reciprocal_sqrt_degrees_tradeoff_enabled() + ) + + test_of_interest.enable( + vector_sources=graph.has_sources_tradeoff_enabled(), + vector_destinations=graph.has_destinations_tradeoff_enabled(), + vector_cumulative_node_degrees=graph.has_cumulative_node_degrees_tradeoff_enabled(), + vector_reciprocal_sqrt_degrees=graph.has_reciprocal_sqrt_degrees_tradeoff_enabled() + ) + additional_validation_kwargs = cls._prepare_evaluation( graph=graph, support=train, @@ -1525,22 +1585,33 @@ def __evaluate_on_single_holdout( time_required_for_setting_up_holdout = time.time() - starting_setting_up_holdout + metadata = dict( + metadata, + time_required_for_setting_up_holdout=time_required_for_setting_up_holdout, + time_required_to_compute_node_features=time_required_to_compute_node_features, + time_required_to_compute_node_type_features=time_required_to_compute_node_type_features, + time_required_to_compute_edge_features=time_required_to_compute_edge_features + ) + holdout_performance = pd.concat([ - classifier.__train_and_evaluate_model( + classifier._train_and_evaluate_model( graph=graph, train_of_interest=train_of_interest, test_of_interest=test_of_interest, train=train, subgraph_of_interest=subgraph_of_interest, + use_subgraph_as_support=use_subgraph_as_support, node_features=holdout_node_features, node_type_features=holdout_node_type_features, edge_features=holdout_edge_features, random_state=random_state, holdout_number=holdout_number, evaluation_schema=evaluation_schema, + holdouts_kwargs=holdouts_kwargs, enable_cache=enable_cache, - automatic_features_names=automatic_features_names, - automatic_features_parameters=automatic_features_parameters, + features_names=features_names, + features_parameters=features_parameters, + metadata=metadata.copy(), additional_validation_kwargs, validation_kwargs, ) @@ -1551,17 +1622,15 @@ def __evaluate_on_single_holdout( ) ]) - holdout_performance["time_required_for_setting_up_holdout"] = time_required_for_setting_up_holdout - return holdout_performance @classmethod @Cache( cache_path="{cache_dir}/{cls.task_name()}/{graph.get_name()}/{_hash}.csv.gz", cache_dir="experiments", - enable_cache_arg_name="enable_cache", + enable_cache_arg_name="enable_top_layer_cache", args_to_ignore=["verbose", "smoke_test"], - capture_enable_cache_arg_name=False, + capture_enable_cache_arg_name=True, use_approximated_hash=True ) def evaluate( @@ -1575,12 +1644,15 @@ def evaluate( node_type_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, edge_features: Optional[Union[str, pd.DataFrame, np.ndarray, List[Union[str, pd.DataFrame, np.ndarray]]]] = None, subgraph_of_interest: Optional[Graph] = None, + use_subgraph_as_support: bool = False, number_of_holdouts: int = 10, random_state: int = 42, verbose: bool = True, enable_cache: bool = False, - precompute_constant_automatic_stocastic_features: bool = False, + precompute_constant_stocastic_features: bool = False, smoke_test: bool = False, + number_of_slurm_nodes: Optional[int] = None, + slurm_node_id_variable: str = "SLURM_GRAPE_ID", validation_kwargs: Dict ) -> pd.DataFrame: """Execute evaluation on the provided graph. @@ -1606,6 +1678,11 @@ def evaluate( The edge features to use. subgraph_of_interest: Optional[Graph] = None Optional subgraph where to focus the task. + This is applied to the train and test graph + after the desired holdout schema is applied. + use_subgraph_as_support: bool = False + Whether to use the provided subgraph as support or + to use the train graph (not filtered by the subgraph). skip_evaluation_biased_feature: bool = False Whether to skip feature names that are known to be biased when running an holdout. These features should be computed @@ -1618,7 +1695,7 @@ def evaluate( Whether to show a loading bar while computing holdouts. enable_cache: bool = False Whether to enable the cache. - precompute_constant_automatic_stocastic_features: bool = False + precompute_constant_stocastic_features: bool = False Whether to precompute once the constant automatic stocastic features before starting the embedding loop. This means that, when left set to false, while the features will be computed @@ -1631,6 +1708,12 @@ def evaluate( Whether this run should be considered a smoke test and therefore use the smoke test configurations for the provided model names and feature names. + number_of_slurm_nodes: Optional[int] = None + Number of SLURM nodes to consider as available. + This variable is used to parallelize the holdouts accordingly. + slurm_node_id_variable: str = "SLURM_GRAPE_ID" + Name of the system variable to use as SLURM node id. + It must be set in the slurm bash script. validation_kwargs: Dict kwargs to be forwarded to the model `_evaluate` method. """ @@ -1667,11 +1750,41 @@ def evaluate( subgraph_of_interest ) else: + if use_subgraph_as_support: + raise ValueError( + "No subgraph of interest was provided but " + "it has been requested to use the subgraph " + "of interest as support. It is not clear " + "how to proceed." + ) subgraph_of_interest_has_compatible_nodes = None + if number_of_slurm_nodes is not None: + must_be_in_slurm_node() + if not isinstance(number_of_slurm_nodes, int) or number_of_slurm_nodes <= 0: + raise ValueError( + "The number of SLURM nodes must be a positive integer value." + ) + if number_of_holdouts > number_of_slurm_nodes: + raise ValueError( + ( + "Please be advised that you are currently running an excessive " + "parametrization of the SLURM cluster. We currently can only parallelize " + "the execution of different holdouts. " + "The number of holdouts requested are {number_of_holdouts} but you are " + "currently using {number_of_slurm_nodes} SLURM nodes! " + "Possibly, you are currently running a task such as a grid search " + "and therefore you intend us to parallelize only the sub-segment of SLURM " + "nodes necessary to run the holdouts." + ).format( + number_of_holdouts=number_of_holdouts, + number_of_slurm_nodes=number_of_slurm_nodes + ) + ) + # Retrieve the set of provided automatic features parameters # so we can put them in the report. - automatic_features_parameters = { + features_parameters = { parameter_name: value for features in ( node_features @@ -1691,7 +1804,7 @@ def evaluate( # Retrieve the set of provided automatic features names # so we can put them in the report. - automatic_features_names = { + features_names = list({ feature.model_name() for features in ( node_features @@ -1706,7 +1819,7 @@ def evaluate( ) for feature in features if issubclass(feature.__class__, AbstractEmbeddingModel) - } + }) # We normalize and/or compute the node features, having # the care of skipping the features that induce bias when @@ -1714,49 +1827,91 @@ def evaluate( # This way we compute only once the features that do not # cause biases for this task, while recomputing those # that cause biases at each holdout, avoiding said biases. + starting_to_compute_constant_node_features = time.time() node_features = cls.normalize_node_features( graph, random_state=random_state, node_features=node_features, allow_automatic_feature=True, skip_evaluation_biased_feature=True, - precompute_constant_automatic_stocastic_features=precompute_constant_automatic_stocastic_features, + precompute_constant_stocastic_features=precompute_constant_stocastic_features, smoke_test=smoke_test ) + time_required_to_compute_constant_node_features = time.time( + ) - starting_to_compute_constant_node_features # We execute the same thing as described above, # but now for the node type features instead that for # the node features. + starting_to_compute_constant_node_type_features = time.time() node_type_features = cls.normalize_node_type_features( graph, random_state=random_state, node_type_features=node_type_features, allow_automatic_feature=True, skip_evaluation_biased_feature=True, - precompute_constant_automatic_stocastic_features=precompute_constant_automatic_stocastic_features, + precompute_constant_stocastic_features=precompute_constant_stocastic_features, smoke_test=smoke_test ) + time_required_to_compute_constant_node_type_features = time.time( + ) - starting_to_compute_constant_node_type_features # We execute the same thing as described above, # but now for the edge features instead that for # the node features. + starting_to_compute_constant_edge_features = time.time() edge_features = cls.normalize_edge_features( graph, random_state=random_state, edge_features=edge_features, allow_automatic_feature=True, skip_evaluation_biased_feature=True, - precompute_constant_automatic_stocastic_features=precompute_constant_automatic_stocastic_features, + precompute_constant_stocastic_features=precompute_constant_stocastic_features, smoke_test=smoke_test ) + time_required_to_compute_constant_edge_features = time.time( + ) - starting_to_compute_constant_edge_features + + metadata = dict( + number_of_threads=os.cpu_count(), + python_version=platform.python_version(), + platform=platform.platform(), + number_of_holdouts=number_of_holdouts, + number_of_slurm_nodes=number_of_slurm_nodes, + time_required_to_compute_constant_node_features=time_required_to_compute_constant_node_features, + time_required_to_compute_constant_node_type_features=time_required_to_compute_constant_node_type_features, + time_required_to_compute_constant_edge_features=time_required_to_compute_constant_edge_features, + ) + + if number_of_slurm_nodes is not None: + if slurm_node_id_variable not in os.environ: + raise ValueError( + ( + "Please do be advised that you have not provided " + "the {slurm_node_id_variable} variable but you have provided specified that " + "we should parallelize the holdouts across {number_of_slurm_nodes} nodes. " + "Please do make available this variable in the " + "slurm bash script by using the `export` flag " + "in a script similar to the following:\n" + "`srun --export=ALL,{slurm_node_id_variable}=$node_id` python3 your_script.py\n" + "You can learn more about this in the library tutorials." + ).format( + slurm_node_id_variable=slurm_node_id_variable, + number_of_slurm_nodes=number_of_slurm_nodes + ) + ) + slurm_node_id = int(os.environ[slurm_node_id_variable]) + metadata["slurm_node_id"] = slurm_node_id + metadata["number_of_slurm_nodes"] = number_of_slurm_nodes # We start to iterate on the holdouts. performance = pd.concat([ - cls.__evaluate_on_single_holdout( + cls._evaluate_on_single_holdout( models=models, library_names=library_names, graph=graph, subgraph_of_interest=subgraph_of_interest, + use_subgraph_as_support=use_subgraph_as_support, node_features=node_features, node_type_features=node_type_features, edge_features=edge_features, @@ -1768,31 +1923,35 @@ def evaluate( smoke_test=smoke_test, holdouts_kwargs=holdouts_kwargs, subgraph_of_interest_has_compatible_nodes=subgraph_of_interest_has_compatible_nodes, - verbose=verbose, - automatic_features_names=automatic_features_names, - automatic_features_parameters=automatic_features_parameters, + verbose=verbose and (number_of_slurm_nodes is None or slurm_node_id==0), + features_names=features_names, + features_parameters=features_parameters, + metadata=metadata.copy(), validation_kwargs ) for holdout_number in trange( number_of_holdouts, - disable=not verbose, + disable=not (verbose and (number_of_slurm_nodes is None or slurm_node_id==0)), leave=False, dynamic_ncols=True, desc=f"Evaluating on {graph.get_name()}" ) + if ( + number_of_slurm_nodes is None or + ( + # We need to also mode the number of SLURM node IDs + # because the user may be parallelizing across many + # diffent nodes in contexts such as wide grid searches. + slurm_node_id % number_of_slurm_nodes + ) == ( + # We need to mode the holdout number as the number + # of holdouts may exceed the number of available SLURM + # nodes that the user has made available to this pipeline. + holdout_number % number_of_slurm_nodes + ) + ) ]) - # Adding to the report the informations relative to - # the whole validation run, which are NOT necessary - # to make unique the cache hash of the single holdouts - # or the single models. - # Be extremely weary of adding informations at this - # high level, as often the best place should be - # in the core loop, where the actual model is trained, - # as often such information changes how the model - # is trained. - performance["number_of_holdouts"] = number_of_holdouts - # We save the constant values for this model # execution. return performance diff --git a/embiggen/utils/abstract_models/abstract_embedding_model.py b/embiggen/utils/abstract_models/abstract_embedding_model.py index 6f4d7a00..24f22542 100644 --- a/embiggen/utils/abstract_models/abstract_embedding_model.py +++ b/embiggen/utils/abstract_models/abstract_embedding_model.py @@ -3,6 +3,7 @@ from ensmallen import Graph from ensmallen.datasets import get_dataset import warnings +from ringbell import RingBell from cache_decorator import Cache from embiggen.utils.abstract_models.abstract_model import AbstractModel, abstract_class from embiggen.utils.abstract_models.embedding_result import EmbeddingResult @@ -16,6 +17,7 @@ def __init__( self, embedding_size: int, enable_cache: bool = False, + ring_bell: bool = False, random_state: Optional[int] = None ): """Create new embedding model. @@ -27,6 +29,8 @@ def __init__( enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. + ring_bell: bool = False + Whether to play a sound when embedding completes. random_state: Optional[int] = None The random state to use if the model is stocastic. """ @@ -38,6 +42,10 @@ def __init__( ) self._embedding_size = embedding_size self._enable_cache = enable_cache + self._ring_bell = RingBell( + verbose=ring_bell, + sample="positive_notification" + ) def parameters(self) -> Dict[str, Any]: """Returns parameters of the embedding model.""" @@ -189,6 +197,8 @@ def _cached_fit_transform( f"but returns an object of type {type(result)}." ) + self._ring_bell.play() + return result def fit_transform( @@ -227,6 +237,18 @@ def fit_transform( repository=repository, version=version )() + if return_dataframe and graph.get_number_of_nodes() > 100_000_000: + raise ValueError( + ( + "We cowardly refuse to execute this embedding with the " + "added requirement to also return the dataframe version " + "of this graph. This graph has {number_of_nodes}, and " + "creating a Dataframe would most likely cause an OOM on " + "your system." + ).format( + number_of_nodes=graph.get_number_of_nodes() + ) + ) return self._cached_fit_transform( graph=graph, return_dataframe=return_dataframe, diff --git a/embiggen/utils/abstract_models/embedding_result.py b/embiggen/utils/abstract_models/embedding_result.py index 94e6e490..4dd922f4 100644 --- a/embiggen/utils/abstract_models/embedding_result.py +++ b/embiggen/utils/abstract_models/embedding_result.py @@ -55,6 +55,7 @@ def __init__( if embedding_list is None: continue for embedding in embedding_list: + if not isinstance(embedding, (np.ndarray, pd.DataFrame)): raise ValueError( f"One of the provided {embedding_list_name} " @@ -67,6 +68,8 @@ def __init__( f"computed with the {embedding_method_name} method " "is empty." ) + if embedding.shape[0] > 1_000_000: + continue if isinstance(embedding, pd.DataFrame): numpy_embedding = embedding.to_numpy() diff --git a/embiggen/utils/abstract_models/model_stub.py b/embiggen/utils/abstract_models/model_stub.py index cfaa7f75..a18de17a 100644 --- a/embiggen/utils/abstract_models/model_stub.py +++ b/embiggen/utils/abstract_models/model_stub.py @@ -91,7 +91,7 @@ def get_model_or_stub( # If effectively the error is that we cannot load the desired # library name, we catch this and re-raise it. if any( - f"No module named '{module_library_name}'" == str(e) + str(e).startswith(f"No module named '{module_library_name}") for module_library_name in module_library_names ): class StubClass(parent_class): diff --git a/embiggen/utils/pipeline.py b/embiggen/utils/pipeline.py index 2a50c728..72fd468f 100644 --- a/embiggen/utils/pipeline.py +++ b/embiggen/utils/pipeline.py @@ -95,7 +95,7 @@ def iterate_graphs( ): if isinstance(graph, str): graph = get_dataset( - name=graph, + graph_name=graph, repository=repository, version=version )() @@ -116,13 +116,16 @@ def classification_evaluation_pipeline( library_names: Optional[Union[str, List[str]]] = None, graph_callback: Optional[Callable[[Graph], Graph]] = None, subgraph_of_interest: Optional[Graph] = None, + use_subgraph_as_support: bool = False, number_of_holdouts: int = 10, random_state: int = 42, repositories: Optional[Union[str, List[str]]] = None, versions: Optional[Union[str, List[str]]] = None, enable_cache: bool = False, - precompute_constant_automatic_stocastic_features: bool = False, + precompute_constant_stocastic_features: bool = False, smoke_test: bool = False, + number_of_slurm_nodes: Optional[int] = None, + slurm_node_id_variable: str = "SLURM_GRAPE_ID", evaluation_kwargs ) -> pd.DataFrame: """Execute classification pipeline for all provided models and graphs. @@ -151,7 +154,12 @@ def classification_evaluation_pipeline( For instance this may be used for filtering the uncertain edges in graphs such as STRING PPIs. subgraph_of_interest: Optional[Graph] = None - The subgraph of interest to focus the task on. + Optional subgraph where to focus the task. + This is applied to the train and test graph + after the desired holdout schema is applied. + use_subgraph_as_support: bool = False + Whether to use the provided subgraph as support or + to use the train graph (not filtered by the subgraph). number_of_holdouts: int = 10 The number of holdouts to execute. random_state: int = 42 @@ -163,7 +171,7 @@ def classification_evaluation_pipeline( Graph versions to retrieve. enable_cache: bool = False Whether to enable the cache. - precompute_constant_automatic_stocastic_features: bool = False + precompute_constant_stocastic_features: bool = False Whether to precompute once the constant automatic stocastic features before starting the embedding loop. This means that, when left set to false, while the features will be computed @@ -177,9 +185,16 @@ def classification_evaluation_pipeline( and therefore use the smoke test configurations for the provided model names and feature names. This parameter will also turn off the cache. + number_of_slurm_nodes: Optional[int] = None + Number of SLURM nodes to consider as available. + This variable is used to parallelize the holdouts accordingly. + slurm_node_id_variable: str = "SLURM_GRAPE_ID" + Name of the system variable to use as SLURM node id. + It must be set in the slurm bash script. evaluation_kwargs: Dict Keyword arguments to forward to evaluation. """ + enable_cache = enable_cache and not smoke_test return pd.concat([ expected_parent_class.evaluate( models=models, @@ -191,11 +206,22 @@ def classification_evaluation_pipeline( node_type_features=node_type_features, edge_features=edge_features, subgraph_of_interest=subgraph_of_interest, + use_subgraph_as_support=use_subgraph_as_support, number_of_holdouts=number_of_holdouts, random_state=random_state, - enable_cache=enable_cache and not smoke_test, - precompute_constant_automatic_stocastic_features=precompute_constant_automatic_stocastic_features, + enable_cache=enable_cache, + # We need this second layer of cache handled separately as + # the different SLURM nodes will try to write simultaneously + # on the same cache files. We could add as additional cache seeds + # also the SLURM node ids and avoid this issue, but then the cache + # would only be valid for that specific cluster and it would + # not be possible to reuse it in other settings such as during + # a reproduction using cache on a notebook. + enable_top_layer_cache=enable_cache and number_of_slurm_nodes is None, + precompute_constant_stocastic_features=precompute_constant_stocastic_features, smoke_test=smoke_test, + number_of_slurm_nodes=number_of_slurm_nodes, + slurm_node_id_variable=slurm_node_id_variable, *evaluation_kwargs ) for graph in iterate_graphs( diff --git a/embiggen/visualizations/graph_visualizer.py b/embiggen/visualizations/graph_visualizer.py index f26cd77a..2043217b 100644 --- a/embiggen/visualizations/graph_visualizer.py +++ b/embiggen/visualizations/graph_visualizer.py @@ -27,9 +27,10 @@ from sklearn.decomposition import PCA from userinput.utils import must_be_in_set from sklearn.metrics import balanced_accuracy_score -from sklearn.model_selection import ShuffleSplit +from sklearn.model_selection import ShuffleSplit, StratifiedShuffleSplit import itertools from embiggen.utils.abstract_models.abstract_embedding_model import AbstractEmbeddingModel +from ensmallen.datasets.graph_retrieval import normalize_node_name from embiggen.utils.abstract_models.embedding_result import EmbeddingResult @@ -95,7 +96,7 @@ def __init__( fps: int = 24, node_embedding_method_name: str = "auto", edge_embedding_method: str = "Concatenate", - minimum_node_degree: int = 1, + minimum_node_degree: int = 0, maximum_node_degree: Optional[int] = None, only_from_same_component: bool = True, sample_only_edges_with_heterogeneous_node_types: bool = False, @@ -202,7 +203,7 @@ def __init__( Using this parameter will raise an exception when the provided graph wither does not have node types or has exclusively constant node types. - minimum_node_degree: Optional[int] = 1 + minimum_node_degree: Optional[int] = 0 The minimum node degree of either the source or destination node to be sampled. maximum_node_degree: Optional[int] = None @@ -464,11 +465,14 @@ def _handle_notebook_display( """ # This is a visualization run for rotation. if len(args) < 2: - return None - figure, axes = args[:2] + figure = None + axes = None + else: + figure, axes = args[:2] if is_notebook() and self._automatically_display_on_notebooks: from IPython.display import display, HTML - display(figure) + if figure is not None: + display(figure) if caption is not None: display(HTML( '<p style="text-align: justify; word-break: break-all;">{}</p>'.format( @@ -476,7 +480,7 @@ def _handle_notebook_display( ) )) plt.close() - elif caption is None: + elif caption is None or self._rotate: return (figure, axes, args[2:]) else: return (figure, axes, args[2:], caption) @@ -515,7 +519,7 @@ def get_heatmaps_comments(self, letters: Optional[List[str]] = None) -> str: " In the heatmap{plural}, {letters}" "low and high values appear in red and blue hues, respectively. " "Intermediate values appear in either a yellow or cyan hue. " - "The values are on a logarithmic scale." + "The values are on a logarithmic scale" ).format( plural=plural, letters="{}, ".format( @@ -547,8 +551,9 @@ def get_non_existing_edges_sampling_description(self) -> str: def get_decomposition_method(self) -> Callable: # Adding a warning for when decomposing methods that # embed nodes using a cosine similarity / distance approach - # in order to avoid false negatives. - if self._decomposition_method in ("UMAP", "TSNE") and self._node_embedding_method_name in ( + # in order to avoid false negatives, that is bad TSNE decompositions + # while the embedding is actually good. + if self._n_components < 3 and self._decomposition_method in ("UMAP", "TSNE") and self._node_embedding_method_name in ( "Node2Vec GloVe", "DeepWalk GloVe", "First-order LINE" ): metric = self._decomposition_kwargs.get("metric") @@ -736,6 +741,7 @@ def _get_node_embedding( f"to have an index curresponding to the node name `{node_name}`, " f"but we have found `{node_embedding.index[node_id]}`." ) + node_embedding = node_embedding.to_numpy() return node_embedding @@ -795,6 +801,16 @@ def decompose(self, X: np.ndarray) -> np.ndarray: "The vector to decompose has less components than " "the decomposition target." ) + # Some embedding method have complex values. + # Such values are, of course, not supported by UMAP, TSNE or PCA. + # For such cases, we need to convert the complex value into a real + # value. If the user desires to use some different approach, it can + # be applied to the embedding before providing it to this visualization tool. + if "complex" in str(X.dtype): + X = np.hstack([ + np.real(X), + np.imag(X) + ]) if self._decomposition_method == "TSNE" and X.shape[1] > 50 and self._graph.get_number_of_nodes() > 50: X = PCA( n_components=50, @@ -802,6 +818,28 @@ def decompose(self, X: np.ndarray) -> np.ndarray: ).fit_transform(X) return self.get_decomposition_method()(X) + def _normalize_label( + self, + labels: List[str] + ) -> List[str]: + last_element = labels[-1] + if last_element.lower().startswith("other"): + labels = [ + label + for label in sanitize_ml_labels([ + normalize_node_name(label) + for label in labels[:-1] + ]) + ] + labels.append(last_element) + return [ + label + for label in sanitize_ml_labels([ + normalize_node_name(label) + for label in labels + ]) + ] + def _set_legend( self, axes: Axes, @@ -827,13 +865,15 @@ def _set_legend( len(label) > 20 for label in labels ) else self._number_of_columns_in_legend + + labels = [ + "{}...".format(label[:20]) + if len(label) > 20 and number_of_columns == 2 else label + for label in self._normalize_label(labels) + ] legend = axes.legend( handles=handles, - labels=[ - "{}...".format(label[:20]) - if len(label) > 20 and number_of_columns == 2 else label - for label in sanitize_ml_labels(labels) - ], + labels=labels, loc=loc, ncol=number_of_columns, prop={'size': 8}, @@ -1054,8 +1094,6 @@ def _get_figure_and_axes( }, squeeze=False ) - for ax in axes.flatten(): - ax.axis('equal') if axes.size == 1: axes = axes.flatten()[0] else: @@ -1066,8 +1104,6 @@ def _get_figure_and_axes( }, squeeze=False ) - for ax in axes.flatten(): - ax.axis('auto') if axes.size == 1: axes = axes.flatten()[0] figure.patch.set_facecolor("white") @@ -1197,6 +1233,11 @@ def _plot_scatter( kwargs ) + if self._n_components == 2: + axes.axis('equal') + else: + axes.axis('auto') + scatter_kwargs = { GraphVisualizer.DEFAULT_SCATTER_KWARGS, ( @@ -1356,7 +1397,7 @@ def _plot_scatter( color_name=color_name, quotations="\'" if "other" not in label.lower() else "", ) - for color_name, label in zip(color_names_to_be_used, labels) + for color_name, label in zip(color_names_to_be_used, self._normalize_label(labels)) ]) return_values = (return_values, caption) @@ -1530,10 +1571,6 @@ def _plot_types( if k < number_of_types: type_labels.append(other_label.format(number_of_types - k)) - type_labels = sanitize_ml_labels( - type_labels[:k] - ) + sanitize_ml_labels(type_labels[k:]) - result = self._wrapped_plot_scatter({ dict( return_caption=return_caption, @@ -1566,11 +1603,16 @@ def _plot_types( test_accuracies = [] - for train_indices, test_indices in ShuffleSplit( + if min(Counter(types).values()) == 1: + SplitterClass = ShuffleSplit + else: + SplitterClass = StratifiedShuffleSplit + + for train_indices, test_indices in SplitterClass( n_splits=self._number_of_holdouts_for_cluster_comments, test_size=0.3, random_state=self._random_state - ).split(points): + ).split(points, types): model = DecisionTreeClassifier(max_depth=5) @@ -1988,7 +2030,7 @@ def plot_positive_and_negative_edges( returned_values = self._plot_types( points=points, title=self._get_complete_title( - "Existent & non-existent edges", + "Edge prediction", show_edge_embedding=True ), types=types, @@ -2170,11 +2212,16 @@ def _plot_positive_and_negative_edges_metric( test_accuracies = [] - for train_indices, test_indices in ShuffleSplit( + if min(Counter(types).values()) == 1: + SplitterClass = ShuffleSplit + else: + SplitterClass = StratifiedShuffleSplit + + for train_indices, test_indices in SplitterClass( n_splits=self._number_of_holdouts_for_cluster_comments, test_size=0.3, random_state=self._random_state - ).split(edge_metrics): + ).split(edge_metrics, types): model = DecisionTreeClassifier(max_depth=5) @@ -2262,7 +2309,7 @@ def _plot_positive_and_negative_edges_metric_histogram( edge_metrics = np.concatenate(( edge_metric_callback(subgraph=self._negative_graph), edge_metric_callback(subgraph=self._positive_graph), - )) + sys.float_info.epsilon + )) axes.hist( [ @@ -2273,6 +2320,7 @@ def _plot_positive_and_negative_edges_metric_histogram( log=True, label=["Non-existent", "Existent"] ) + axes.set_xlim(edge_metrics.min(), edge_metrics.max()) axes.set_ylabel("Counts (log scale)") axes.set_xlabel(metric_name) axes.legend(loc='best', prop={'size': 8},) @@ -3819,7 +3867,7 @@ def _plot_positive_and_negative_edges_distance_histogram( aligned_mapping=True, include_both_undirected_edges=False ) - graph_transformer.fit(node_features.astype(np.float32)) + graph_transformer.fit(node_features) return self._plot_positive_and_negative_edges_metric_histogram( metric_name=distance_name, @@ -3868,7 +3916,7 @@ def _plot_positive_and_negative_edges_distance( include_both_undirected_edges=False ) - graph_transformer.fit(node_features.astype(np.float32)) + graph_transformer.fit(node_features) return self._plot_positive_and_negative_edges_metric( metric_name=distance_name, @@ -4262,119 +4310,29 @@ def plot_edge_weight_distribution( return self._handle_notebook_display(figure, axes, caption=caption) - def fit_and_plot_all( + def _fit_and_plot_all( self, - node_embedding: Union[pd.DataFrame, np.ndarray, str, EmbeddingResult], - number_of_columns: int = 4, - show_letters: bool = True, - include_distribution_plots: bool = True, - node_embedding_kwargs: Dict + points: List[np.ndarray], + nrows: int, + ncols: int, + plotting_callbacks: List[Callable], + show_letters: bool, ) -> Tuple[Figure, Axes]: """Fits and plots all available features of the graph. Parameters ------------------------- - node_embedding: Union[pd.DataFrame, np.ndarray, str] - Embedding of the graph nodes. - If a string is provided, we will run the node embedding - from one of the available methods. - number_of_columns: int = 4 - Number of columns to use for the layout. - show_letters: bool = True - Whether to show letters on the top left of the subplots. - include_distribution_plots: bool = True - Whether to include the distribution plots for the degrees - and the edge weights, if they are present. - node_embedding_kwargs: Dict Kwargs to be forwarded to the node embedding algorithm. """ - node_embedding = self._get_node_embedding( - node_embedding, - node_embedding_kwargs - ) - self.fit_nodes(node_embedding, node_embedding_kwargs) - self.fit_negative_and_positive_edges( - node_embedding, node_embedding_kwargs) - - node_scatter_plot_methods_to_call = [] - distribution_plot_methods_to_call = [] - - node_scatter_plot_methods_to_call.append( - self.plot_node_degrees, - ) - distribution_plot_methods_to_call.append( - self.plot_node_degree_distribution, - ) - - def plot_distance_wrapper(plot_distance): - @functools.wraps(plot_distance) - def wrapped_plot_distance(kwargs): - return plot_distance( - node_features=node_embedding, - kwargs - ) - return wrapped_plot_distance - - edge_scatter_plot_methods_to_call = [ - self.plot_positive_and_negative_edges, - plot_distance_wrapper( - self.plot_positive_and_negative_edges_euclidean_distance), - plot_distance_wrapper( - self.plot_positive_and_negative_edges_cosine_similarity), - self.plot_positive_and_negative_edges_adamic_adar, - self.plot_positive_and_negative_edges_jaccard_coefficient, - self.plot_positive_and_negative_edges_preferential_attachment, - self.plot_positive_and_negative_edges_resource_allocation_index - ] - - distribution_plot_methods_to_call = [ - plot_distance_wrapper( - self.plot_positive_and_negative_edges_euclidean_distance_histogram), - plot_distance_wrapper( - self.plot_positive_and_negative_edges_cosine_similarity_histogram), - self.plot_positive_and_negative_adamic_adar_histogram, - self.plot_positive_and_negative_jaccard_coefficient_histogram, - self.plot_positive_and_negative_preferential_attachment_histogram, - self.plot_positive_and_negative_resource_allocation_index_histogram + decompositions_backup = [ + self._node_decomposition, + self._positive_edge_decomposition, + self._negative_edge_decomposition ] - if self._graph.has_node_types() and not self._graph.has_homogeneous_node_types(): - node_scatter_plot_methods_to_call.append( - self.plot_node_types - ) - - if self._graph.has_node_ontologies() and not self._graph.has_homogeneous_node_ontologies(): - node_scatter_plot_methods_to_call.append( - self.plot_node_ontologies - ) - - if not self._support.is_connected() and not self._support.is_directed(): - node_scatter_plot_methods_to_call.append( - self.plot_connected_components - ) - - if self._positive_graph.has_edge_types() and not self._positive_graph.has_homogeneous_edge_types(): - edge_scatter_plot_methods_to_call.append( - self.plot_edge_types - ) - - if self._positive_graph.has_edge_weights() and not self._positive_graph.has_constant_edge_weights(): - edge_scatter_plot_methods_to_call.append( - self.plot_edge_weights - ) - distribution_plot_methods_to_call.append( - self.plot_edge_weight_distribution - ) - - if not include_distribution_plots: - distribution_plot_methods_to_call = [] - - number_of_total_plots = len(node_scatter_plot_methods_to_call) + len( - edge_scatter_plot_methods_to_call - ) + len(distribution_plot_methods_to_call) - nrows = max( - int(math.ceil(number_of_total_plots / number_of_columns)), 1) - ncols = min(number_of_columns, number_of_total_plots) + (self._node_decomposition, + self._positive_edge_decomposition, + self._negative_edge_decomposition) = points figure, axes = self._get_figure_and_axes( nrows=nrows, @@ -4383,6 +4341,7 @@ def wrapped_plot_distance(kwargs): dpi=96 ) figure.patch.set_facecolor("white") + number_of_total_plots = len(plotting_callbacks) flat_axes = np.array(axes).flatten() @@ -4413,20 +4372,17 @@ def wrapped_plot_distance(kwargs): for ax, plot_callback, letter in zip( flat_axes, - itertools.chain( - node_scatter_plot_methods_to_call, - edge_scatter_plot_methods_to_call, - distribution_plot_methods_to_call - ), + itertools.chain(plotting_callbacks), "abcdefghjkilmnopqrstuvwxyz" ): inspect.signature(plot_callback).parameters - _, _, caption = plot_callback( + figure, axes, caption = plot_callback( figure=figure, axes=ax, (dict(loc="lower center") if "loc" in inspect.signature(plot_callback).parameters else dict()), apply_tight_layout=False ) + if "heatmap" in caption.lower(): heatmaps_letters.append(letter) if "accuracy" in caption.lower(): @@ -4434,14 +4390,12 @@ def wrapped_plot_distance(kwargs): complete_caption += f" <b>({letter})</b> {caption}" if show_letters: - if self._n_components == 3: - additional_kwargs = dict(z=0.0) + if self._n_components >= 3: + additional_kwargs = dict(z=0, y=0, x=0) else: - additional_kwargs = dict() + additional_kwargs = dict(y=1.1, x=0) ax.text( - x=0.0, - y=1.1, s=letter, size=18, color="black", @@ -4488,9 +4442,191 @@ def wrapped_plot_distance(kwargs): figure.tight_layout() self._show_graph_name = show_name_backup + (self._node_decomposition, + self._positive_edge_decomposition, + self._negative_edge_decomposition) = decompositions_backup return self._handle_notebook_display( figure, - axes, + flat_axes, caption=complete_caption ) + + def fit_and_plot_all( + self, + node_embedding: Union[pd.DataFrame, np.ndarray, str, EmbeddingResult], + number_of_columns: int = 4, + show_letters: bool = True, + include_distribution_plots: bool = True, + node_embedding_kwargs: Dict + ) -> Tuple[Figure, Axes]: + """Fits and plots all available features of the graph. + + Parameters + ------------------------- + node_embedding: Union[pd.DataFrame, np.ndarray, str] + Embedding of the graph nodes. + If a string is provided, we will run the node embedding + from one of the available methods. + number_of_columns: int = 4 + Number of columns to use for the layout. + show_letters: bool = True + Whether to show letters on the top left of the subplots. + include_distribution_plots: bool = True + Whether to include the distribution plots for the degrees + and the edge weights, if they are present. + node_embedding_kwargs: Dict + Kwargs to be forwarded to the node embedding algorithm. + """ + node_embedding = self._get_node_embedding( + node_embedding, + node_embedding_kwargs + ) + self.fit_nodes(node_embedding, node_embedding_kwargs) + self.fit_negative_and_positive_edges( + node_embedding, node_embedding_kwargs) + + node_scatter_plot_methods_to_call = [] + distribution_plot_methods_to_call = [] + + node_scatter_plot_methods_to_call.append( + self.plot_node_degrees, + ) + distribution_plot_methods_to_call.append( + self.plot_node_degree_distribution, + ) + + def plot_distance_wrapper(plot_distance): + @functools.wraps(plot_distance) + def wrapped_plot_distance(kwargs): + return plot_distance( + node_features=node_embedding, + *kwargs + ) + return wrapped_plot_distance + + edge_scatter_plot_methods_to_call = [ + self.plot_positive_and_negative_edges, + plot_distance_wrapper( + self.plot_positive_and_negative_edges_euclidean_distance), + plot_distance_wrapper( + self.plot_positive_and_negative_edges_cosine_similarity), + self.plot_positive_and_negative_edges_adamic_adar, + self.plot_positive_and_negative_edges_jaccard_coefficient, + self.plot_positive_and_negative_edges_preferential_attachment, + self.plot_positive_and_negative_edges_resource_allocation_index + ] + + distribution_plot_methods_to_call = [ + plot_distance_wrapper( + self.plot_positive_and_negative_edges_euclidean_distance_histogram), + plot_distance_wrapper( + self.plot_positive_and_negative_edges_cosine_similarity_histogram), + self.plot_positive_and_negative_adamic_adar_histogram, + self.plot_positive_and_negative_jaccard_coefficient_histogram, + self.plot_positive_and_negative_preferential_attachment_histogram, + self.plot_positive_and_negative_resource_allocation_index_histogram + ] + + if self._graph.has_node_types() and not self._graph.has_homogeneous_node_types(): + node_scatter_plot_methods_to_call.append( + self.plot_node_types + ) + + if self._graph.has_node_ontologies() and not self._graph.has_homogeneous_node_ontologies(): + node_scatter_plot_methods_to_call.append( + self.plot_node_ontologies + ) + + if not self._support.is_connected() and not self._support.is_directed(): + node_scatter_plot_methods_to_call.append( + self.plot_connected_components + ) + + if self._positive_graph.has_edge_types() and not self._positive_graph.has_homogeneous_edge_types(): + edge_scatter_plot_methods_to_call.append( + self.plot_edge_types + ) + + if self._positive_graph.has_edge_weights() and not self._positive_graph.has_constant_edge_weights(): + edge_scatter_plot_methods_to_call.append( + self.plot_edge_weights + ) + distribution_plot_methods_to_call.append( + self.plot_edge_weight_distribution + ) + + if not include_distribution_plots or self._rotate or self._n_components > 2: + distribution_plot_methods_to_call = [] + + plotting_callbacks = [ + callback + for callbacks in ( + node_scatter_plot_methods_to_call, + edge_scatter_plot_methods_to_call, + distribution_plot_methods_to_call + ) + for callback in callbacks + ] + + number_of_total_plots = len(plotting_callbacks) + + nrows = max( + int(math.ceil(number_of_total_plots / number_of_columns)), 1) + ncols = min(number_of_columns, number_of_total_plots) + + if self._rotate: + path = "fit_and_plot_all.webm" + display_backup = self._automatically_display_on_notebooks + self._automatically_display_on_notebooks = False + rotate_backup = self._rotate + self._rotate = False + rotate( + self._fit_and_plot_all, + path=path, + points=[ + self._node_decomposition, + self._positive_edge_decomposition, + self._negative_edge_decomposition + ], + duration=self._duration, + fps=self._fps, + verbose=self._verbose, + nrows = nrows, + ncols = ncols, + plotting_callbacks=plotting_callbacks, + show_letters=show_letters, + ) + to_display = display_video_at_path( + path, + width="100%", + height=None + ) + figure, axes, complete_caption = self._fit_and_plot_all( + points=[ + self._node_decomposition, + self._positive_edge_decomposition, + self._negative_edge_decomposition + ], + nrows = nrows, + ncols = ncols, + plotting_callbacks=plotting_callbacks, + show_letters=show_letters + ) + self._rotate = rotate_backup + self._automatically_display_on_notebooks = display_backup + return self._handle_notebook_display( + to_display, None, + caption=complete_caption + ) + return self._fit_and_plot_all( + points=[ + self._node_decomposition, + self._positive_edge_decomposition, + self._negative_edge_decomposition + ], + nrows = nrows, + ncols = ncols, + plotting_callbacks=plotting_callbacks, + show_letters=show_letters + ) diff --git a/setup.py b/setup.py index 599276ae..3c85d921 100644 --- a/setup.py +++ b/setup.py @@ -8,7 +8,7 @@ def readme(): - with open('README.rst') as f: + with open('README.rst', encoding="utf8") as f: return f.read() @@ -24,7 +24,7 @@ def readme(): def read(parts): - with copen(os.path.join(here, parts), 'r') as fp: + with copen(os.path.join(here, parts), 'r', encoding="utf8") as fp: return fp.read() @@ -66,13 +66,17 @@ def find_version(*file_paths): "tqdm", "matplotlib>=3.5.2", "scikit-learn", + "dict_hash>=1.1.29", "userinput>=1.0.19", - "ddd_subplots>=1.0.20", - "sanitize_ml_labels>=1.0.38", + "ddd_subplots>=1.0.23", + "sanitize_ml_labels>=1.0.45", "keras_mixed_sequence>=1.0.28", - "ensmallen>=0.8.8", + "ensmallen>=0.8.24", + "environments_utils>=1.0.6", + "compress_pickle>=2.1.0", "validate_version_code", - "cache_decorator>=2.1.8", + "ringbell>=1.0.2", + "cache_decorator>=2.1.11", "packaging" ], tests_require=test_deps,	Fix typo in `environments_utils` dependency		2022-10-04T18:57:08	[]	[]
monarch-initiative/embiggen	monarch-initiative__embiggen-283	42500d5d18503b5ef3a109daf915d49c2064ba03	diff --git a/.gitignore b/.gitignore index cbec2be7..6634e04f 100644 --- a/.gitignore +++ b/.gitignore @@ -33,4 +33,5 @@ htmlcov embedding.npy build graphs -embeddings \ No newline at end of file +embeddings +.webm \ No newline at end of file diff --git a/README.rst b/README.rst index 5a6d6464..87387cda 100644 --- a/README.rst +++ b/README.rst @@ -1,8 +1,8 @@ -Embiggen +🍇 Embiggen ========================================================================================= \|pip\| \|downloads\| \|tutorials\| \|documentation\| \|python_version\| \|DOI\| \|license\| \|telegram\| \|discord\| \|twitter\| -Embiggen is the graph machine learning submodule of the `GraPE <https://github.com/AnacletoLAB/grape>`_ library. +Embiggen is the graph machine learning submodule of the `🍇 GraPE <https://github.com/AnacletoLAB/grape>`_ library. How to install Embiggen ------------------------- diff --git a/dist/embiggen-0.11.12.tar.gz b/dist/embiggen-0.11.12.tar.gz new file mode 100644 index 00000000..ccc2cc1b Binary files /dev/null and b/dist/embiggen-0.11.12.tar.gz differ diff --git a/embiggen/__version__.py b/embiggen/__version__.py index c084d6b1..e1fd9917 100644 --- a/embiggen/__version__.py +++ b/embiggen/__version__.py @@ -1,2 +1,2 @@ """Current version of package embiggen""" -__version__ = "0.11.11" +__version__ = "0.11.15" diff --git a/embiggen/edge_label_prediction/edge_label_prediction_evaluation.py b/embiggen/edge_label_prediction/edge_label_prediction_evaluation.py index 254662b0..dec73085 100644 --- a/embiggen/edge_label_prediction/edge_label_prediction_evaluation.py +++ b/embiggen/edge_label_prediction/edge_label_prediction_evaluation.py @@ -13,6 +13,7 @@ def edge_label_prediction_evaluation( models: Union[Type[AbstractEdgeLabelPredictionModel], List[Type[AbstractEdgeLabelPredictionModel]]], evaluation_schema: str = "Stratified Monte Carlo", node_features: Optional[Union[str, pd.DataFrame, np.ndarray, Type[AbstractEmbeddingModel], List[Union[str, pd.DataFrame, np.ndarray, Type[AbstractEmbeddingModel]]]]] = None, + node_type_features: Optional[Union[str, pd.DataFrame, np.ndarray, Type[AbstractEmbeddingModel], List[Union[str, pd.DataFrame, np.ndarray, Type[AbstractEmbeddingModel]]]]] = None, edge_features: Optional[Union[str, pd.DataFrame, np.ndarray, List[Union[str, pd.DataFrame, np.ndarray]]]] = None, library_names: Optional[Union[str, List[str]]] = None, graph_callback: Optional[Callable[[Graph], Graph]] = None, @@ -40,6 +41,8 @@ def edge_label_prediction_evaluation( The evaluation schema to follow. node_features: Optional[Union[str, pd.DataFrame, np.ndarray, Type[AbstractEmbeddingModel], List[Union[str, pd.DataFrame, np.ndarray, Type[AbstractEmbeddingModel]]]]] = None The node features to use. + node_type_features: Optional[Union[str, pd.DataFrame, np.ndarray, Type[AbstractEmbeddingModel], List[Union[str, pd.DataFrame, np.ndarray, Type[AbstractEmbeddingModel]]]]] = None + The node type features to use. edge_features: Optional[Union[str, pd.DataFrame, np.ndarray, List[Union[str, pd.DataFrame, np.ndarray]]]] = None The edge features to use. library_names: Optional[Union[str, List[str]]] = None @@ -86,6 +89,7 @@ def edge_label_prediction_evaluation( models=models, expected_parent_class=AbstractEdgeLabelPredictionModel, node_features=node_features, + node_type_features=node_type_features, edge_features=edge_features, library_names=library_names, graph_callback=graph_callback, diff --git a/embiggen/edge_label_prediction/edge_label_prediction_model.py b/embiggen/edge_label_prediction/edge_label_prediction_model.py index aa21c278..00d82516 100644 --- a/embiggen/edge_label_prediction/edge_label_prediction_model.py +++ b/embiggen/edge_label_prediction/edge_label_prediction_model.py @@ -3,7 +3,7 @@ import pandas as pd import numpy as np from ensmallen import Graph -from embiggen.utils.abstract_models import AbstractClassifierModel, format_list +from embiggen.utils.abstract_models import AbstractClassifierModel class AbstractEdgeLabelPredictionModel(AbstractClassifierModel): @@ -20,7 +20,7 @@ def __init__(self, random_state: Optional[int] = None): self._is_binary_prediction_task = None self._is_multilabel_prediction_task = None super().__init__(random_state=random_state) - + @classmethod def requires_edge_types(cls) -> bool: """Returns whether this method requires node types.""" @@ -48,7 +48,9 @@ def get_available_evaluation_schemas(cls) -> List[str]: """Returns available evaluation schemas for this task.""" return [ "Stratified Monte Carlo", - "Stratified Kfold" + "Stratified Kfold", + "Kfold", + "Monte Carlo", ] @classmethod @@ -78,23 +80,26 @@ def split_graph_following_evaluation_schema( holdouts_kwargs: Dict[str, Any] The kwargs to be forwarded to the holdout method. """ - if evaluation_schema == "Stratified Monte Carlo": + if evaluation_schema in ("Monte Carlo", "Stratified Monte Carlo"): return graph.get_edge_label_holdout_graphs( holdouts_kwargs, - use_stratification=True, + use_stratification="Stratified" in evaluation_schema, random_state=random_state+holdout_number, ) - if evaluation_schema == "Stratified Kfold": + if evaluation_schema in ("Kfold", "Stratified Kfold"): return graph.get_edge_label_kfold( k=number_of_holdouts, k_index=holdout_number, - use_stratification=True, + use_stratification="Stratified" in evaluation_schema, random_state=random_state, ) - raise ValueError( - f"The requested evaluation schema `{evaluation_schema}` " - "is not available. The available evaluation schemas " - f"are: {format_list(cls.get_available_evaluation_schemas())}." + super().split_graph_following_evaluation_schema( + graph=graph, + evaluation_schema=evaluation_schema, + random_state=random_state, + holdout_number=holdout_number, + number_of_holdouts=number_of_holdouts, + holdouts_kwargs, ) @classmethod @@ -127,6 +132,7 @@ def _evaluate( ) -> List[Dict[str, Any]]: """Return model evaluation on the provided graphs.""" train_size = train.get_number_of_known_edge_types() / graph.get_number_of_known_edge_types() + performance = [] for evaluation_mode, evaluation_graph in ( ("train", train), @@ -139,9 +145,8 @@ def _evaluate( node_type_features=node_type_features, edge_features=edge_features )[evaluation_graph.get_edge_ids_with_known_edge_types()] - + if self.is_binary_prediction_task(): - prediction_probabilities = prediction_probabilities[:, 1] predictions = prediction_probabilities labels = evaluation_graph.get_known_edge_type_ids() == 1 elif self.is_multilabel_prediction_task(): @@ -158,7 +163,7 @@ def _evaluate( performance.append({ "evaluation_mode": evaluation_mode, "train_size": train_size, - "known_edges_number": graph.get_number_of_known_node_types(), + "known_edges_number": graph.get_number_of_known_edge_types(), self.evaluate_predictions( labels, predictions, @@ -197,15 +202,26 @@ def fit( edge_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None The edge features to use. """ - if node_type_features is not None: - raise NotImplementedError( - "Support for node type features is not currently available for any " - "of the edge-label prediction models." + non_zero_edge_types = sum([ + 1 + for count in graph.get_edge_type_names_counts_hashmap().values() + if count > 0 + ]) + + if non_zero_edge_types < 2: + raise ValueError( + "The provided training graph has less than two non-zero edge types. " + "It is unclear how to proceeed." ) - self._is_binary_prediction_task = graph.get_number_of_edge_types() == 2 + self._is_binary_prediction_task = non_zero_edge_types == 2 self._is_multilabel_prediction_task = graph.is_multigraph() + if self._is_multilabel_prediction_task: + raise ValueError( + "Currently we do not support multi-label edge prediction." + ) + super().fit( graph=graph, support=support, @@ -232,4 +248,4 @@ def task_involves_node_types(cls) -> bool: @classmethod def task_involves_topology(cls) -> bool: """Returns whether the model task involves topology.""" - return False \ No newline at end of file + return False diff --git a/embiggen/edge_label_prediction/edge_label_prediction_sklearn/decision_tree_edge_label_prediction.py b/embiggen/edge_label_prediction/edge_label_prediction_sklearn/decision_tree_edge_label_prediction.py index f03bfeb5..107f6365 100644 --- a/embiggen/edge_label_prediction/edge_label_prediction_sklearn/decision_tree_edge_label_prediction.py +++ b/embiggen/edge_label_prediction/edge_label_prediction_sklearn/decision_tree_edge_label_prediction.py @@ -18,10 +18,10 @@ def __init__( max_features=None, max_leaf_nodes=None, min_impurity_decrease=0., - min_impurity_split=None, - class_weight=None, + class_weight="balanced", ccp_alpha=0.0, edge_embedding_method: str = "Concatenate", + use_edge_metrics: bool = False, random_state: int = 42 ): """Create the Decision Tree for Edge Label Prediction.""" @@ -34,7 +34,6 @@ def __init__( self._max_features = max_features self._max_leaf_nodes = max_leaf_nodes self._min_impurity_decrease = min_impurity_decrease - self._min_impurity_split = min_impurity_split self._random_state = random_state self._class_weight = class_weight self._ccp_alpha = ccp_alpha @@ -50,13 +49,13 @@ def __init__( max_features=max_features, max_leaf_nodes=max_leaf_nodes, min_impurity_decrease=min_impurity_decrease, - min_impurity_split=min_impurity_split, random_state=random_state, class_weight=class_weight, ccp_alpha=ccp_alpha, ), - edge_embedding_method, - random_state + edge_embedding_method=edge_embedding_method, + use_edge_metrics=use_edge_metrics, + random_state=random_state ) @classmethod @@ -80,7 +79,6 @@ def parameters(self) -> Dict[str, Any]: max_features=self._max_features, max_leaf_nodes=self._max_leaf_nodes, min_impurity_decrease=self._min_impurity_decrease, - min_impurity_split=self._min_impurity_split, random_state=self._random_state, class_weight=self._class_weight, ccp_alpha=self._ccp_alpha, diff --git a/embiggen/edge_label_prediction/edge_label_prediction_sklearn/extra_trees_edge_label_prediction.py b/embiggen/edge_label_prediction/edge_label_prediction_sklearn/extra_trees_edge_label_prediction.py index d9f0d961..df0c2b93 100644 --- a/embiggen/edge_label_prediction/edge_label_prediction_sklearn/extra_trees_edge_label_prediction.py +++ b/embiggen/edge_label_prediction/edge_label_prediction_sklearn/extra_trees_edge_label_prediction.py @@ -20,16 +20,16 @@ def __init__( max_features="auto", max_leaf_nodes=None, min_impurity_decrease=0., - min_impurity_split=None, bootstrap=True, oob_score=False, n_jobs=-1, verbose=0, warm_start=False, - class_weight=None, + class_weight="balanced", ccp_alpha=0.0, max_samples=None, edge_embedding_method: str = "Concatenate", + use_edge_metrics: bool = False, random_state: int = 42 ): """Create the Extra Trees for Edge Label Prediction.""" @@ -42,7 +42,6 @@ def __init__( self._max_features = max_features self._max_leaf_nodes = max_leaf_nodes self._min_impurity_decrease = min_impurity_decrease - self._min_impurity_split = min_impurity_split self._bootstrap = bootstrap self._oob_score = oob_score self._n_jobs = cpu_count() if n_jobs == -1 else n_jobs @@ -64,7 +63,6 @@ def __init__( max_features=max_features, max_leaf_nodes=max_leaf_nodes, min_impurity_decrease=min_impurity_decrease, - min_impurity_split=min_impurity_split, bootstrap=bootstrap, oob_score=oob_score, n_jobs=n_jobs, @@ -75,8 +73,9 @@ def __init__( ccp_alpha=ccp_alpha, max_samples=max_samples ), - edge_embedding_method, - random_state + edge_embedding_method=edge_embedding_method, + use_edge_metrics=use_edge_metrics, + random_state=random_state ) @classmethod @@ -101,7 +100,6 @@ def parameters(self) -> Dict[str, Any]: max_features = self._max_features, max_leaf_nodes = self._max_leaf_nodes, min_impurity_decrease = self._min_impurity_decrease, - min_impurity_split = self._min_impurity_split, bootstrap = self._bootstrap, oob_score = self._oob_score, n_jobs = self._n_jobs, diff --git a/embiggen/edge_label_prediction/edge_label_prediction_sklearn/gradient_boosting_edge_label_prediction.py b/embiggen/edge_label_prediction/edge_label_prediction_sklearn/gradient_boosting_edge_label_prediction.py index cddb60e5..bdb243d0 100644 --- a/embiggen/edge_label_prediction/edge_label_prediction_sklearn/gradient_boosting_edge_label_prediction.py +++ b/embiggen/edge_label_prediction/edge_label_prediction_sklearn/gradient_boosting_edge_label_prediction.py @@ -19,7 +19,6 @@ def __init__( min_weight_fraction_leaf=0., max_depth=3, min_impurity_decrease=0., - min_impurity_split=None, init=None, max_features=None, verbose=0, @@ -30,25 +29,25 @@ def __init__( tol=1e-4, ccp_alpha=0.0, edge_embedding_method: str = "Concatenate", + use_edge_metrics: bool = False, random_state: int = 42 ): """Create the Gradient Boosting for Edge Label Prediction.""" self._loss=loss self._learning_rate=learning_rate - self._n_estimators=n_estimators, + self._n_estimators=n_estimators self._criterion=criterion - self._min_samples_split=min_samples_split, - self._min_samples_leaf=min_samples_leaf, - self._min_weight_fraction_leaf=min_weight_fraction_leaf, + self._min_samples_split=min_samples_split + self._min_samples_leaf=min_samples_leaf + self._min_weight_fraction_leaf=min_weight_fraction_leaf self._max_depth=max_depth self._init=init - self._subsample=subsample, - self._max_features=max_features, - self._max_leaf_nodes=max_leaf_nodes, - self._min_impurity_decrease=min_impurity_decrease, - self._min_impurity_split=min_impurity_split, + self._subsample=subsample + self._max_features=max_features + self._max_leaf_nodes=max_leaf_nodes + self._min_impurity_decrease=min_impurity_decrease self._warm_start=warm_start - self._validation_fraction=validation_fraction, + self._validation_fraction=validation_fraction self._n_iter_no_change=n_iter_no_change self._tol=tol self._ccp_alpha=ccp_alpha @@ -63,12 +62,12 @@ def __init__( random_state=random_state, verbose=verbose, max_leaf_nodes=max_leaf_nodes, min_impurity_decrease=min_impurity_decrease, - min_impurity_split=min_impurity_split, warm_start=warm_start, validation_fraction=validation_fraction, n_iter_no_change=n_iter_no_change, tol=tol, ccp_alpha=ccp_alpha ), - edge_embedding_method, - random_state + edge_embedding_method=edge_embedding_method, + use_edge_metrics=use_edge_metrics, + random_state=random_state ) @classmethod @@ -97,7 +96,6 @@ def parameters(self) -> Dict[str, Any]: max_features=self._max_features, max_leaf_nodes=self._max_leaf_nodes, min_impurity_decrease=self._min_impurity_decrease, - min_impurity_split=self._min_impurity_split, warm_start=self._warm_start, validation_fraction=self._validation_fraction, n_iter_no_change=self._n_iter_no_change, diff --git a/embiggen/edge_label_prediction/edge_label_prediction_sklearn/mlp_edge_label_prediction.py b/embiggen/edge_label_prediction/edge_label_prediction_sklearn/mlp_edge_label_prediction.py index 1e825e1d..f76f39f0 100644 --- a/embiggen/edge_label_prediction/edge_label_prediction_sklearn/mlp_edge_label_prediction.py +++ b/embiggen/edge_label_prediction/edge_label_prediction_sklearn/mlp_edge_label_prediction.py @@ -32,6 +32,7 @@ def __init__( n_iter_no_change=10, max_fun=15000, edge_embedding_method: str = "Concatenate", + use_edge_metrics: bool = False, random_state: int = 42 ): """Create the MLP for Edge Label Prediction.""" @@ -82,8 +83,9 @@ def __init__( n_iter_no_change=n_iter_no_change, max_fun=max_fun ), - edge_embedding_method, - random_state + edge_embedding_method=edge_embedding_method, + use_edge_metrics=use_edge_metrics, + random_state=random_state ) @classmethod diff --git a/embiggen/edge_label_prediction/edge_label_prediction_sklearn/random_forest_edge_label_prediction.py b/embiggen/edge_label_prediction/edge_label_prediction_sklearn/random_forest_edge_label_prediction.py index 1fe4170c..551a7554 100644 --- a/embiggen/edge_label_prediction/edge_label_prediction_sklearn/random_forest_edge_label_prediction.py +++ b/embiggen/edge_label_prediction/edge_label_prediction_sklearn/random_forest_edge_label_prediction.py @@ -17,19 +17,19 @@ def __init__( min_samples_split: int = 2, min_samples_leaf: int = 1, min_weight_fraction_leaf: float = 0., - max_features="auto", + max_features="sqrt", max_leaf_nodes=None, min_impurity_decrease=0., - min_impurity_split=None, bootstrap=True, oob_score=False, n_jobs=-1, verbose=0, warm_start=False, - class_weight=None, + class_weight="balanced", ccp_alpha=0.0, max_samples=None, edge_embedding_method: str = "Concatenate", + use_edge_metrics: bool = False, random_state: int = 42 ): """Create the Random Forest for Edge Label Prediction.""" @@ -42,7 +42,6 @@ def __init__( self._max_features = max_features self._max_leaf_nodes = max_leaf_nodes self._min_impurity_decrease = min_impurity_decrease - self._min_impurity_split = min_impurity_split self._bootstrap = bootstrap self._oob_score = oob_score self._n_jobs = cpu_count() if n_jobs == -1 else n_jobs @@ -64,7 +63,6 @@ def __init__( max_features=max_features, max_leaf_nodes=max_leaf_nodes, min_impurity_decrease=min_impurity_decrease, - min_impurity_split=min_impurity_split, bootstrap=bootstrap, oob_score=oob_score, n_jobs=n_jobs, @@ -75,8 +73,9 @@ def __init__( ccp_alpha=ccp_alpha, max_samples=max_samples ), - edge_embedding_method, - random_state + edge_embedding_method=edge_embedding_method, + use_edge_metrics=use_edge_metrics, + random_state=random_state ) @classmethod @@ -101,7 +100,6 @@ def parameters(self) -> Dict[str, Any]: max_features = self._max_features, max_leaf_nodes = self._max_leaf_nodes, min_impurity_decrease = self._min_impurity_decrease, - min_impurity_split = self._min_impurity_split, bootstrap = self._bootstrap, oob_score = self._oob_score, n_jobs = self._n_jobs, diff --git a/embiggen/edge_label_prediction/edge_label_prediction_sklearn/sklearn_edge_label_prediction_adapter.py b/embiggen/edge_label_prediction/edge_label_prediction_sklearn/sklearn_edge_label_prediction_adapter.py index aafda249..013d081c 100644 --- a/embiggen/edge_label_prediction/edge_label_prediction_sklearn/sklearn_edge_label_prediction_adapter.py +++ b/embiggen/edge_label_prediction/edge_label_prediction_sklearn/sklearn_edge_label_prediction_adapter.py @@ -16,6 +16,7 @@ def __init__( self, model_instance: Type[ClassifierMixin], edge_embedding_method: str = "Concatenate", + use_edge_metrics: bool = False, random_state: int = 42 ): """Create the adapter for Sklearn object. @@ -26,6 +27,13 @@ def __init__( The class instance to be adapted into edge-label prediction. edge_embedding_method: str = "Concatenate" The method to use to compute the edges. + use_edge_metrics: bool = False + Whether to use the edge metrics from traditional edge prediction. + These metrics currently include: + - Adamic Adar + - Jaccard Coefficient + - Resource allocation index + - Preferential attachment random_state: int = 42 The random state to use to reproduce the training. @@ -38,6 +46,7 @@ def __init__( must_be_an_sklearn_classifier_model(model_instance) self._model_instance = model_instance self._edge_embedding_method = edge_embedding_method + self._use_edge_metrics = use_edge_metrics # We want to mask the decorator class name self.__class__.__name__ = model_instance.__class__.__name__ self.__class__.__doc__ = model_instance.__class__.__doc__ @@ -46,6 +55,7 @@ def parameters(self) -> Dict[str, Any]: """Returns parameters used for this model.""" return { "edge_embedding_method": self._edge_embedding_method, + "use_edge_metrics": self._use_edge_metrics, super().parameters() } @@ -56,6 +66,7 @@ def clone(self) -> Type["SklearnEdgeLabelPredictionAdapter"]: def _trasform_graph_into_edge_embedding( self, graph: Graph, + support: Optional[Graph] = None, node_features: Optional[List[np.ndarray]] = None, node_type_features: Optional[List[np.ndarray]] = None, edge_features: Optional[List[np.ndarray]] = None, @@ -67,6 +78,11 @@ def _trasform_graph_into_edge_embedding( graph: Graph The graph whose edges are to be embedded and predicted. It can either be an Graph or a list of lists of edges. + support: Optional[Graph] = None + The graph describiding the topological structure that + includes also the above graph. This parameter + is mostly useful for topological classifiers + such as Graph Convolutional Networks. node_features: np.ndarray The node features to be used in the training of the model. node_type_features: np.ndarray @@ -83,13 +99,30 @@ def _trasform_graph_into_edge_embedding( """ gt = GraphTransformer( method=self._edge_embedding_method, - aligned_node_mapping=True + aligned_mapping=True + ) + + gt.fit( + node_features, + node_type_feature=node_type_features ) - gt.fit(node_features) + if self._use_edge_metrics: + edge_metrics = support.get_all_edge_metrics( + normalize=True, + subgraph=graph, + ) + if edge_features is not None: + edge_features = np.hstack([ + edge_features, + edge_metrics + ]) + else: + edge_features = edge_metrics return gt.transform( graph=graph, + node_types=graph, edge_features=edge_features ) @@ -129,16 +162,31 @@ def _fit( """ lpt = EdgeLabelPredictionTransformer( method=self._edge_embedding_method, - aligned_node_mapping=True + aligned_mapping=True ) - lpt.fit(node_features) + lpt.fit( + node_features, + node_type_feature=node_type_features + ) + + if self._use_edge_metrics: + edge_metrics = support.get_all_edge_metrics( + normalize=True, + subgraph=graph, + ) + if edge_features is not None: + edge_features = np.hstack([ + edge_features, + edge_metrics + ]) + else: + edge_features = edge_metrics self._model_instance.fit(lpt.transform( graph=graph, edge_features=edge_features, behaviour_for_unknown_edge_labels="drop", - random_state=self._random_state )) def _predict_proba( @@ -175,13 +223,24 @@ def _predict_proba( ValueError If the two graphs do not share the same node vocabulary. """ - return self._model_instance.predict_proba(self._trasform_graph_into_edge_embedding( - graph=graph, - node_features=node_features, - edge_features=edge_features, - )) + prediction_probabilities = self._model_instance.predict_proba( + self._trasform_graph_into_edge_embedding( + graph=graph, + support=support, + node_features=node_features, + node_type_features=node_type_features, + edge_features=edge_features, + ) + ) + # In the majority but not totality of sklearn models, + # the predictions of binary models are returned as + # a couple of vectors for the positive and negative class. + if self.is_binary_prediction_task() and prediction_probabilities.shape[1] == 2: + prediction_probabilities = prediction_probabilities[:, 1] + return prediction_probabilities + - def predict( + def _predict( self, graph: Graph, support: Optional[Graph] = None, @@ -217,7 +276,9 @@ def predict( """ return self._model_instance.predict(self._trasform_graph_into_edge_embedding( graph=graph, + support=support, node_features=node_features, + node_type_features=node_type_features, edge_features=edge_features, )) diff --git a/embiggen/edge_label_prediction/edge_label_prediction_tensorflow/gcn.py b/embiggen/edge_label_prediction/edge_label_prediction_tensorflow/gcn.py index 1747394d..80264125 100644 --- a/embiggen/edge_label_prediction/edge_label_prediction_tensorflow/gcn.py +++ b/embiggen/edge_label_prediction/edge_label_prediction_tensorflow/gcn.py @@ -1,9 +1,9 @@ """GCN model for edge-label prediction.""" -from typing import List, Union, Optional, Dict, Type, Any +from typing import List, Union, Optional, Dict, Type import numpy as np from ensmallen import Graph - +import tensorflow as tf from tensorflow.keras.optimizers import Optimizer from embiggen.utils.abstract_edge_gcn import AbstractEdgeGCN, abstract_class from embiggen.edge_label_prediction.edge_label_prediction_model import AbstractEdgeLabelPredictionModel @@ -25,6 +25,7 @@ def __init__( number_of_units_per_ffnn_head_layer: Union[int, List[int]] = 128, dropout_rate: float = 0.3, apply_norm: bool = False, + combiner: str = "mean", edge_embedding_method: str = "Concatenate", optimizer: Union[str, Type[Optimizer]] = "adam", early_stopping_min_delta: float = 0.0001, @@ -47,6 +48,7 @@ def __init__( handling_multi_graph: str = "warn", node_feature_names: Optional[List[str]] = None, node_type_feature_names: Optional[List[str]] = None, + edge_feature_names: Optional[List[str]] = None, verbose: bool = False ): """Create new Kipf GCN object. @@ -77,6 +79,13 @@ def __init__( apply_norm: bool = False Whether to normalize the output of the convolution operations, after applying the level activations. + combiner: str = "mean" + A string specifying the reduction op. + Currently "mean", "sqrtn" and "sum" are supported. + "sum" computes the weighted sum of the embedding results for each row. + "mean" is the weighted sum divided by the total weight. + "sqrtn" is the weighted sum divided by the square root of the sum of the squares of the weights. + Defaults to mean. edge_embedding_method: str = "Concatenate" The edge embedding method to use to put togheter the source and destination node features, which includes: @@ -173,6 +182,9 @@ def __init__( node_type_feature_names: Optional[List[str]] = None Names of the node type features. This is used as the layer names. + edge_feature_names: Optional[List[str]] = None + Names of the edge features. + This is used as the layer names. verbose: bool = False Whether to show loading bars. """ @@ -187,6 +199,7 @@ def __init__( number_of_units_per_ffnn_head_layer=number_of_units_per_ffnn_head_layer, dropout_rate=dropout_rate, apply_norm=apply_norm, + combiner=combiner, edge_embedding_method=edge_embedding_method, optimizer=optimizer, early_stopping_min_delta=early_stopping_min_delta, @@ -209,6 +222,7 @@ def __init__( handling_multi_graph=handling_multi_graph, node_feature_names=node_feature_names, node_type_feature_names=node_type_feature_names, + edge_feature_names=edge_feature_names, verbose=verbose, ) AbstractEdgeLabelPredictionModel.__init__(self, random_state=random_state) @@ -245,4 +259,6 @@ def _get_class_weights(self, graph: Graph) -> Dict[int, float]: def get_output_classes(self, graph: Graph) -> int: """Returns number of output classes.""" + if self.is_binary_prediction_task(): + return 1 return graph.get_number_of_edge_types() diff --git a/embiggen/edge_label_prediction/edge_label_prediction_tensorflow/gnn.py b/embiggen/edge_label_prediction/edge_label_prediction_tensorflow/gnn.py index e8adf840..e68d4734 100644 --- a/embiggen/edge_label_prediction/edge_label_prediction_tensorflow/gnn.py +++ b/embiggen/edge_label_prediction/edge_label_prediction_tensorflow/gnn.py @@ -14,7 +14,6 @@ def __init__( number_of_head_layers: int = 1, number_of_units_per_body_layer: Union[int, List[int]] = 128, number_of_units_per_head_layer: Union[int, List[int]] = 128, - dropout_rate: float = 0.3, edge_embedding_method: str = "Concatenate", optimizer: Union[str, Type[Optimizer]] = "adam", early_stopping_min_delta: float = 0.0001, @@ -35,6 +34,7 @@ def __init__( node_type_embedding_size: int = 50, node_feature_names: Optional[List[str]] = None, node_type_feature_names: Optional[List[str]] = None, + edge_feature_names: Optional[List[str]] = None, verbose: bool = False ): """Create new GraphSAGE object. @@ -55,9 +55,6 @@ def __init__( Number of units per ffnn body layer. number_of_units_per_ffnn_head_layer: Union[int, List[int]] = 128 Number of units per ffnn head layer. - dropout_rate: float = 0.3 - Float between 0 and 1. - Fraction of the input units to dropout. edge_embedding_method: str = "Concatenate" The edge embedding method to use to put togheter the source and destination node features, which includes: @@ -146,6 +143,9 @@ def __init__( node_type_feature_names: Optional[List[str]] = None Names of the node type features. This is used as the layer names. + edge_feature_names: Optional[List[str]] = None + Names of the edge features. + This is used as the layer names. verbose: bool = False Whether to show loading bars. """ @@ -157,7 +157,6 @@ def __init__( number_of_ffnn_head_layers=number_of_head_layers, number_of_units_per_ffnn_body_layer=number_of_units_per_body_layer, number_of_units_per_ffnn_head_layer=number_of_units_per_head_layer, - dropout_rate=dropout_rate, edge_embedding_method=edge_embedding_method, optimizer=optimizer, early_stopping_min_delta=early_stopping_min_delta, @@ -178,6 +177,7 @@ def __init__( node_type_embedding_size=node_type_embedding_size, node_feature_names=node_feature_names, node_type_feature_names=node_type_feature_names, + edge_feature_names=edge_feature_names, verbose=verbose, ) @@ -206,14 +206,17 @@ def parameters(self) -> Dict[str, Any]: "number_of_units_per_graph_convolution_layers", "handling_multi_graph", "number_of_units_per_ffnn_body_layer", - "number_of_units_per_ffnn_head_layer" + "number_of_units_per_ffnn_head_layer", + "combiner", + "apply_norm", + "dropout_rate" ] return dict( number_of_units_per_body_layer=self._number_of_units_per_ffnn_body_layer, number_of_units_per_head_layer=self._number_of_units_per_ffnn_head_layer, { key: value - for key, value in super().smoke_test_parameters().items() + for key, value in super().parameters().items() if key not in removed } ) diff --git a/embiggen/edge_label_prediction/edge_label_prediction_tensorflow/graph_sage.py b/embiggen/edge_label_prediction/edge_label_prediction_tensorflow/graph_sage.py index 7718701b..3e27699f 100644 --- a/embiggen/edge_label_prediction/edge_label_prediction_tensorflow/graph_sage.py +++ b/embiggen/edge_label_prediction/edge_label_prediction_tensorflow/graph_sage.py @@ -39,6 +39,7 @@ def __init__( handling_multi_graph: str = "warn", node_feature_names: Optional[List[str]] = None, node_type_feature_names: Optional[List[str]] = None, + edge_feature_names: Optional[List[str]] = None, verbose: bool = False ): """Create new GraphSAGE object. @@ -160,6 +161,9 @@ def __init__( node_type_feature_names: Optional[List[str]] = None Names of the node type features. This is used as the layer names. + edge_feature_names: Optional[List[str]] = None + Names of the edge features. + This is used as the layer names. verbose: bool = False Whether to show loading bars. """ @@ -173,6 +177,7 @@ def __init__( number_of_units_per_ffnn_head_layer=number_of_units_per_ffnn_head_layer, dropout_rate=dropout_rate, apply_norm=True, + combiner="mean", edge_embedding_method=edge_embedding_method, optimizer=optimizer, early_stopping_min_delta=early_stopping_min_delta, @@ -195,6 +200,7 @@ def __init__( handling_multi_graph=handling_multi_graph, node_feature_names=node_feature_names, node_type_feature_names=node_type_feature_names, + edge_feature_names=edge_feature_names, verbose=verbose, ) @@ -202,7 +208,8 @@ def parameters(self) -> Dict[str, Any]: """Returns parameters for smoke test.""" removed = [ "apply_norm", - "use_simmetric_normalized_laplacian" + "use_simmetric_normalized_laplacian", + "combiner" ] return dict( { diff --git a/embiggen/edge_label_prediction/edge_label_prediction_tensorflow/kipf_gcn.py b/embiggen/edge_label_prediction/edge_label_prediction_tensorflow/kipf_gcn.py index 341b1783..0e40c886 100644 --- a/embiggen/edge_label_prediction/edge_label_prediction_tensorflow/kipf_gcn.py +++ b/embiggen/edge_label_prediction/edge_label_prediction_tensorflow/kipf_gcn.py @@ -1,5 +1,5 @@ """Kipf GCN model for edge-label prediction.""" -from typing import List, Union, Optional, Type +from typing import List, Union, Optional, Type, Dict, Any from tensorflow.keras.optimizers import Optimizer from embiggen.edge_label_prediction.edge_label_prediction_tensorflow.gcn import GCNEdgeLabelPrediction @@ -40,6 +40,7 @@ def __init__( handling_multi_graph: str = "warn", node_feature_names: Optional[List[str]] = None, node_type_feature_names: Optional[List[str]] = None, + edge_feature_names: Optional[List[str]] = None, verbose: bool = False ): """Create new Kipf GCN object. @@ -164,6 +165,9 @@ def __init__( node_type_feature_names: Optional[List[str]] = None Names of the node type features. This is used as the layer names. + edge_feature_names: Optional[List[str]] = None + Names of the edge features. + This is used as the layer names. verbose: bool = False Whether to show loading bars. """ @@ -177,6 +181,7 @@ def __init__( number_of_units_per_ffnn_head_layer=number_of_units_per_ffnn_head_layer, dropout_rate=dropout_rate, apply_norm=apply_norm, + combiner="sum", edge_embedding_method=edge_embedding_method, optimizer=optimizer, early_stopping_min_delta=early_stopping_min_delta, @@ -199,9 +204,23 @@ def __init__( handling_multi_graph=handling_multi_graph, node_feature_names=node_feature_names, node_type_feature_names=node_type_feature_names, + edge_feature_names=edge_feature_names, verbose=verbose, ) + def parameters(self) -> Dict[str, Any]: + """Returns parameters for smoke test.""" + removed = [ + "combiner" + ] + return dict( + { + key: value + for key, value in super().parameters().items() + if key not in removed + } + ) + @classmethod def model_name(cls) -> str: return "Kipf GCN" \ No newline at end of file diff --git a/embiggen/edge_prediction/edge_prediction_ensmallen/perceptron.py b/embiggen/edge_prediction/edge_prediction_ensmallen/perceptron.py index 87b8b030..5407d194 100644 --- a/embiggen/edge_prediction/edge_prediction_ensmallen/perceptron.py +++ b/embiggen/edge_prediction/edge_prediction_ensmallen/perceptron.py @@ -3,6 +3,7 @@ from ensmallen import Graph import numpy as np from ensmallen import models +from embiggen.embedding_transformers import NodeTransformer from embiggen.edge_prediction.edge_prediction_model import AbstractEdgePredictionModel @@ -11,15 +12,18 @@ class PerceptronEdgePrediction(AbstractEdgePredictionModel): def __init__( self, - edge_features: Union[str, List[str]] = "JaccardCoefficient", - edge_embeddings: Optional[List[str]] = None, + edge_features: Optional[Union[str, List[str]]] = "JaccardCoefficient", + edge_embeddings: Optional[Union[str, List[str]]] = None, cooccurrence_iterations: int = 100, cooccurrence_window_size: int = 10, number_of_epochs: int = 100, - number_of_edges_per_mini_batch: int = 4096, + number_of_edges_per_mini_batch: int = 256, sample_only_edges_with_heterogeneous_node_types: bool = False, learning_rate: float = 0.001, - learning_rate_decay: float = 0.99, + first_order_decay_factor: float = 0.9, + second_order_decay_factor: float = 0.999, + avoid_false_negatives: bool = False, + use_scale_free_distribution: bool = True, random_state: int = 42, verbose: bool = True ): @@ -27,7 +31,7 @@ def __init__( Parameters ------------------------ - edge_features: Union[str, List[str]] = "JaccardCoefficient" + edge_features: Optional[Union[str, List[str]]] = "JaccardCoefficient" The edge features to compute for each edge. Zero or more edge features can be used at once. The currently supported edge features are: @@ -37,7 +41,7 @@ def __init__( - Cooccurrence, - ResourceAllocationIndex, - PreferentialAttachment, - edge_embeddings: Optional[List[str]] = None + edge_embeddings: Optional[Union[str, List[str]]] = None The embedding methods to use for the provided node features. Zero or more edge emmbedding methods can be used at once. The currently supported edge embedding are: @@ -59,23 +63,37 @@ def __init__( By default 100. number_of_epochs: int = 100 The number of epochs to train the model for. By default, 100. - number_of_edges_per_mini_batch: int = 1024 - The number of samples to include for each mini-batch. By default 1024. + number_of_edges_per_mini_batch: int = 4096 + The number of samples to include for each mini-batch. By default 4096. sample_only_edges_with_heterogeneous_node_types: bool = False Whether to sample negative edges only with source and destination nodes that have different node types. By default false. learning_rate: float = 0.001 Learning rate to use while training the model. By default 0.001. + first_order_decay_factor: float = 0.9 + First order decay factor for the first order momentum. + By default 0.9. + second_order_decay_factor: float = 0.999 + Second order decay factor for the second order momentum. + By default 0.999. + avoid_false_negatives: bool = False + Whether to avoid sampling false negatives. + This may cause a slower training. + use_scale_free_distribution: bool = True + Whether to train model using a scale free distribution for the negatives. random_state: int = 42 The random state to reproduce the model initialization and training. By default, 42. verbose: bool = True Whether to show epochs loading bar. """ super().__init__(random_state=random_state) - + if isinstance(edge_features, str): edge_features = [edge_features] + if isinstance(edge_embeddings, str): + edge_embeddings = [edge_embeddings] + self._model_kwargs = dict( edge_features=edge_features, edge_embeddings=edge_embeddings, @@ -85,7 +103,10 @@ def __init__( number_of_edges_per_mini_batch=number_of_edges_per_mini_batch, sample_only_edges_with_heterogeneous_node_types=sample_only_edges_with_heterogeneous_node_types, learning_rate=learning_rate, - learning_rate_decay=learning_rate_decay + first_order_decay_factor=first_order_decay_factor, + second_order_decay_factor=second_order_decay_factor, + avoid_false_negatives=avoid_false_negatives, + use_scale_free_distribution=use_scale_free_distribution, ) self._verbose = verbose self._model = models.EdgePredictionPerceptron( @@ -136,12 +157,26 @@ def _fit( new_node_features = [] if node_features is not None: for node_feature in node_features: - if node_feature.dtype != np.float32: - node_feature = node_feature.astype(np.float32) if not node_feature.data.c_contiguous: node_feature = np.ascontiguousarray(node_feature) new_node_features.append(node_feature) + if node_type_features is not None: + for node_type_feature in node_type_features: + node_transformer = NodeTransformer( + aligned_mapping=True + ) + node_transformer.fit( + node_type_feature=node_type_feature + ) + node_feature = node_transformer.transform( + node_types=graph + ) + if not node_feature.data.c_contiguous: + node_feature = np.ascontiguousarray(node_feature) + new_node_features.append(node_feature) + + self._model.fit( graph=graph, node_features=new_node_features, @@ -212,8 +247,6 @@ def _predict_proba( new_node_features = [] if node_features is not None: for node_feature in node_features: - if node_feature.dtype != np.float32: - node_feature = node_feature.astype(np.float32) if not node_feature.data.c_contiguous: node_feature = np.ascontiguousarray(node_feature) new_node_features.append(node_feature) diff --git a/embiggen/edge_prediction/edge_prediction_evaluation.py b/embiggen/edge_prediction/edge_prediction_evaluation.py index 4d50a0f8..c8bab4f3 100644 --- a/embiggen/edge_prediction/edge_prediction_evaluation.py +++ b/embiggen/edge_prediction/edge_prediction_evaluation.py @@ -23,7 +23,7 @@ def edge_prediction_evaluation( versions: Optional[Union[str, List[str]]] = None, validation_sample_only_edges_with_heterogeneous_node_types: bool = False, validation_unbalance_rates: Tuple[float] = (1.0, ), - use_zipfian_sampling: bool = True, + use_scale_free_distribution: bool = True, enable_cache: bool = False, precompute_constant_automatic_stocastic_features: bool = False, smoke_test: bool = False, @@ -92,10 +92,10 @@ def edge_prediction_evaluation( This can be useful when executing a bipartite edge prediction task. validation_unbalance_rates: Tuple[float] = (1.0, ) Unbalance rate for the non-existent graphs generation. - use_zipfian_sampling: bool = True - Whether to use the zipfian sampling of the NEGATIVE edges for the EVALUATION + use_scale_free_distribution: bool = True + Whether to use the scale free sampling of the NEGATIVE edges for the EVALUATION of the edge prediction performance of the provided models. - Please DO BE ADVISED that not using a zipfian sampling for the negative + Please DO BE ADVISED that not using a scale free sampling for the negative edges is a poor choice and will cause a significant positive bias in the model performance. enable_cache: bool = False @@ -138,5 +138,5 @@ def edge_prediction_evaluation( verbose=verbose, validation_sample_only_edges_with_heterogeneous_node_types=validation_sample_only_edges_with_heterogeneous_node_types, validation_unbalance_rates=validation_unbalance_rates, - use_zipfian_sampling=use_zipfian_sampling + use_scale_free_distribution=use_scale_free_distribution ) diff --git a/embiggen/edge_prediction/edge_prediction_model.py b/embiggen/edge_prediction/edge_prediction_model.py index 6e015de8..79950a1b 100644 --- a/embiggen/edge_prediction/edge_prediction_model.py +++ b/embiggen/edge_prediction/edge_prediction_model.py @@ -6,7 +6,7 @@ import math from ensmallen import Graph from tqdm.auto import tqdm -from embiggen.utils.abstract_models import AbstractClassifierModel, AbstractEmbeddingModel, abstract_class, format_list +from embiggen.utils.abstract_models import AbstractClassifierModel, abstract_class, format_list @abstract_class @@ -82,10 +82,13 @@ def split_graph_following_evaluation_schema( random_state=random_state, verbose=False ) - raise ValueError( - f"The requested evaluation schema `{evaluation_schema}` " - "is not available. The available evaluation schemas " - f"are: {format_list(cls.get_available_evaluation_schemas())}." + super().split_graph_following_evaluation_schema( + graph=graph, + evaluation_schema=evaluation_schema, + random_state=random_state, + holdout_number=holdout_number, + number_of_holdouts=number_of_holdouts, + holdouts_kwargs, ) @staticmethod @@ -99,7 +102,7 @@ def __iterate_negative_graphs( verbose: bool, validation_sample_only_edges_with_heterogeneous_node_types: bool, validation_unbalance_rates: Tuple[float], - use_zipfian_sampling: bool + use_scale_free_distribution: bool ) -> Iterator[Tuple[Graph]]: """Return iterator over the negative graphs for evaluation.""" if subgraph_of_interest is None: @@ -107,9 +110,9 @@ def __iterate_negative_graphs( else: sampler_graph = subgraph_of_interest - if not use_zipfian_sampling: + if not use_scale_free_distribution: warnings.warn( - "Please do be advised that you have DISABLED the use of zipfian sampling " + "Please do be advised that you have DISABLED the use of scale free sampling " "for the negative edges for the EVALUATION (not the training) " "of a model. This is a POOR CHOICE as it will introduce a positive bias " "as edges sampled uniformely have a significantly different node degree " @@ -130,7 +133,7 @@ def __iterate_negative_graphs( ), random_state=random_state(i+1), sample_only_edges_with_heterogeneous_node_types=validation_sample_only_edges_with_heterogeneous_node_types, - use_zipfian_sampling=use_zipfian_sampling, + use_scale_free_distribution=use_scale_free_distribution, support=support, graph_to_avoid=graph ).random_holdout( @@ -160,7 +163,7 @@ def _prepare_evaluation( verbose: bool = True, validation_sample_only_edges_with_heterogeneous_node_types: bool = False, validation_unbalance_rates: Tuple[float] = (1.0, ), - use_zipfian_sampling: bool = True + use_scale_free_distribution: bool = True ) -> Dict[str, Any]: """Return additional custom parameters for the current holdout.""" return dict( @@ -174,7 +177,7 @@ def _prepare_evaluation( verbose=verbose, validation_sample_only_edges_with_heterogeneous_node_types=validation_sample_only_edges_with_heterogeneous_node_types, validation_unbalance_rates=validation_unbalance_rates, - use_zipfian_sampling=use_zipfian_sampling + use_scale_free_distribution=use_scale_free_distribution )) ) @@ -193,7 +196,7 @@ def _evaluate( negative_graphs: Optional[List[Tuple[Graph]]] = None, validation_sample_only_edges_with_heterogeneous_node_types: bool = False, validation_unbalance_rates: Tuple[float] = (1.0, ), - use_zipfian_sampling: bool = True, + use_scale_free_distribution: bool = True, ) -> List[Dict[str, Any]]: """Return model evaluation on the provided graphs.""" performance = [] @@ -206,9 +209,6 @@ def _evaluate( edge_features=edge_features ) - if len(train_predic_proba.shape) > 1 and train_predic_proba.shape[1] > 1: - train_predic_proba = train_predic_proba[:, 1] - test_predict_proba = self.predict_proba( test, support=support, @@ -217,9 +217,6 @@ def _evaluate( edge_features=edge_features ) - if len(test_predict_proba.shape) > 1 and test_predict_proba.shape[1] > 1: - test_predict_proba = test_predict_proba[:, 1] - negative_graph_iterator = self.__iterate_negative_graphs( graph=graph, train=train, @@ -230,7 +227,7 @@ def _evaluate( verbose=verbose, validation_sample_only_edges_with_heterogeneous_node_types=validation_sample_only_edges_with_heterogeneous_node_types, validation_unbalance_rates=validation_unbalance_rates, - use_zipfian_sampling=use_zipfian_sampling + use_scale_free_distribution=use_scale_free_distribution ) if negative_graphs is None else negative_graphs for unbalance_rate, (negative_train, negative_test) in tqdm( @@ -269,6 +266,7 @@ def _evaluate( performance.append({ "evaluation_mode": evaluation_mode, "validation_unbalance_rate": unbalance_rate, + "use_scale_free_distribution": use_scale_free_distribution, "validation_sample_only_edges_with_heterogeneous_node_types": validation_sample_only_edges_with_heterogeneous_node_types, self.evaluate_predictions( labels, @@ -322,7 +320,7 @@ def predict( support=support, node_features=node_features, node_type_features=node_type_features - ) + ).flatten() if return_predictions_dataframe: predictions = pd.DataFrame( @@ -751,7 +749,12 @@ def predict_proba( support=support, node_features=node_features, node_type_features=node_type_features - ) + ).flatten() + + if np.isnan(predictions).any(): + raise ValueError( + "There are NaN values in the predicted probabilities!" + ) if return_predictions_dataframe: predictions = pd.DataFrame( diff --git a/embiggen/edge_prediction/edge_prediction_sklearn/decision_tree_edge_prediction.py b/embiggen/edge_prediction/edge_prediction_sklearn/decision_tree_edge_prediction.py index 274f2eaa..b4023b56 100644 --- a/embiggen/edge_prediction/edge_prediction_sklearn/decision_tree_edge_prediction.py +++ b/embiggen/edge_prediction/edge_prediction_sklearn/decision_tree_edge_prediction.py @@ -11,20 +11,19 @@ def __init__( self, criterion="gini", splitter="best", - max_depth=None, + max_depth=10, min_samples_split=2, min_samples_leaf=1, min_weight_fraction_leaf=0., max_features=None, max_leaf_nodes=None, min_impurity_decrease=0., - class_weight=None, ccp_alpha=0.0, edge_embedding_method: str = "Concatenate", training_unbalance_rate: float = 1.0, training_sample_only_edges_with_heterogeneous_node_types: bool = False, use_edge_metrics: bool = False, - use_zipfian_sampling: bool = True, + use_scale_free_distribution: bool = True, prediction_batch_size: int = 212, random_state: int = 42 ): @@ -39,7 +38,6 @@ def __init__( self._max_leaf_nodes = max_leaf_nodes self._min_impurity_decrease = min_impurity_decrease self._random_state = random_state - self._class_weight = class_weight self._ccp_alpha = ccp_alpha super().__init__( @@ -54,13 +52,12 @@ def __init__( max_leaf_nodes=max_leaf_nodes, min_impurity_decrease=min_impurity_decrease, random_state=random_state, - class_weight=class_weight, ccp_alpha=ccp_alpha, ), edge_embedding_method=edge_embedding_method, training_unbalance_rate=training_unbalance_rate, use_edge_metrics=use_edge_metrics, - use_zipfian_sampling=use_zipfian_sampling, + use_scale_free_distribution=use_scale_free_distribution, training_sample_only_edges_with_heterogeneous_node_types=training_sample_only_edges_with_heterogeneous_node_types, prediction_batch_size=prediction_batch_size, random_state=random_state @@ -81,7 +78,6 @@ def parameters(self) -> Dict[str, Any]: max_leaf_nodes=self._max_leaf_nodes, min_impurity_decrease=self._min_impurity_decrease, random_state=self._random_state, - class_weight=self._class_weight, ccp_alpha=self._ccp_alpha, ) } diff --git a/embiggen/edge_prediction/edge_prediction_sklearn/extra_trees_edge_prediction.py b/embiggen/edge_prediction/edge_prediction_sklearn/extra_trees_edge_prediction.py index 77259a02..7b93da62 100644 --- a/embiggen/edge_prediction/edge_prediction_sklearn/extra_trees_edge_prediction.py +++ b/embiggen/edge_prediction/edge_prediction_sklearn/extra_trees_edge_prediction.py @@ -19,20 +19,18 @@ def __init__( max_features="auto", max_leaf_nodes=None, min_impurity_decrease=0., - min_impurity_split=None, bootstrap=True, oob_score=False, n_jobs=-1, verbose=0, warm_start=False, - class_weight=None, ccp_alpha=0.0, max_samples=None, edge_embedding_method: str = "Concatenate", training_unbalance_rate: float = 1.0, training_sample_only_edges_with_heterogeneous_node_types: bool = False, use_edge_metrics: bool = False, - use_zipfian_sampling: bool = True, + use_scale_free_distribution: bool = True, prediction_batch_size: int = 212, random_state: int = 42 ): @@ -46,14 +44,12 @@ def __init__( self._max_features = max_features self._max_leaf_nodes = max_leaf_nodes self._min_impurity_decrease = min_impurity_decrease - self._min_impurity_split = min_impurity_split self._bootstrap = bootstrap self._oob_score = oob_score self._n_jobs = n_jobs self._random_state = random_state self._verbose = verbose self._warm_start = warm_start - self._class_weight = class_weight self._ccp_alpha = ccp_alpha self._max_samples = max_samples @@ -68,21 +64,19 @@ def __init__( max_features=max_features, max_leaf_nodes=max_leaf_nodes, min_impurity_decrease=min_impurity_decrease, - min_impurity_split=min_impurity_split, bootstrap=bootstrap, oob_score=oob_score, n_jobs=n_jobs, random_state=random_state, verbose=verbose, warm_start=warm_start, - class_weight=class_weight, ccp_alpha=ccp_alpha, max_samples=max_samples ), edge_embedding_method=edge_embedding_method, training_unbalance_rate=training_unbalance_rate, use_edge_metrics=use_edge_metrics, - use_zipfian_sampling=use_zipfian_sampling, + use_scale_free_distribution=use_scale_free_distribution, training_sample_only_edges_with_heterogeneous_node_types=training_sample_only_edges_with_heterogeneous_node_types, prediction_batch_size=prediction_batch_size, random_state=random_state @@ -110,14 +104,12 @@ def parameters(self) -> Dict[str, Any]: max_features = self._max_features, max_leaf_nodes = self._max_leaf_nodes, min_impurity_decrease = self._min_impurity_decrease, - min_impurity_split = self._min_impurity_split, bootstrap = self._bootstrap, oob_score = self._oob_score, n_jobs = self._n_jobs, random_state = self._random_state, verbose = self._verbose, warm_start = self._warm_start, - class_weight = self._class_weight, ccp_alpha = self._ccp_alpha, max_samples = self._max_samples, ) diff --git a/embiggen/edge_prediction/edge_prediction_sklearn/gradient_boosting_edge_prediction.py b/embiggen/edge_prediction/edge_prediction_sklearn/gradient_boosting_edge_prediction.py index e94206c6..63f9b56f 100644 --- a/embiggen/edge_prediction/edge_prediction_sklearn/gradient_boosting_edge_prediction.py +++ b/embiggen/edge_prediction/edge_prediction_sklearn/gradient_boosting_edge_prediction.py @@ -19,7 +19,6 @@ def __init__( min_weight_fraction_leaf=0., max_depth=3, min_impurity_decrease=0., - min_impurity_split=None, init=None, max_features=None, verbose=0, @@ -33,27 +32,26 @@ def __init__( training_unbalance_rate: float = 1.0, training_sample_only_edges_with_heterogeneous_node_types: bool = False, use_edge_metrics: bool = False, - use_zipfian_sampling: bool = True, + use_scale_free_distribution: bool = True, prediction_batch_size: int = 212, random_state: int = 42 ): """Create the Gradient Boosting for Edge Prediction.""" self._loss=loss self._learning_rate=learning_rate - self._n_estimators=n_estimators, + self._n_estimators=n_estimators self._criterion=criterion - self._min_samples_split=min_samples_split, - self._min_samples_leaf=min_samples_leaf, - self._min_weight_fraction_leaf=min_weight_fraction_leaf, + self._min_samples_split=min_samples_split + self._min_samples_leaf=min_samples_leaf + self._min_weight_fraction_leaf=min_weight_fraction_leaf self._max_depth=max_depth self._init=init - self._subsample=subsample, - self._max_features=max_features, - self._max_leaf_nodes=max_leaf_nodes, - self._min_impurity_decrease=min_impurity_decrease, - self._min_impurity_split=min_impurity_split, + self._subsample=subsample + self._max_features=max_features + self._max_leaf_nodes=max_leaf_nodes + self._min_impurity_decrease=min_impurity_decrease self._warm_start=warm_start - self._validation_fraction=validation_fraction, + self._validation_fraction=validation_fraction self._n_iter_no_change=n_iter_no_change self._tol=tol self._ccp_alpha=ccp_alpha @@ -68,14 +66,13 @@ def __init__( random_state=random_state, verbose=verbose, max_leaf_nodes=max_leaf_nodes, min_impurity_decrease=min_impurity_decrease, - min_impurity_split=min_impurity_split, warm_start=warm_start, validation_fraction=validation_fraction, n_iter_no_change=n_iter_no_change, tol=tol, ccp_alpha=ccp_alpha ), edge_embedding_method=edge_embedding_method, training_unbalance_rate=training_unbalance_rate, use_edge_metrics=use_edge_metrics, - use_zipfian_sampling=use_zipfian_sampling, + use_scale_free_distribution=use_scale_free_distribution, training_sample_only_edges_with_heterogeneous_node_types=training_sample_only_edges_with_heterogeneous_node_types, prediction_batch_size=prediction_batch_size, random_state=random_state @@ -99,7 +96,6 @@ def parameters(self) -> Dict[str, Any]: max_features=self._max_features, max_leaf_nodes=self._max_leaf_nodes, min_impurity_decrease=self._min_impurity_decrease, - min_impurity_split=self._min_impurity_split, warm_start=self._warm_start, validation_fraction=self._validation_fraction, n_iter_no_change=self._n_iter_no_change, diff --git a/embiggen/edge_prediction/edge_prediction_sklearn/mlp_edge_prediction.py b/embiggen/edge_prediction/edge_prediction_sklearn/mlp_edge_prediction.py index 40b3d614..82ae8c52 100644 --- a/embiggen/edge_prediction/edge_prediction_sklearn/mlp_edge_prediction.py +++ b/embiggen/edge_prediction/edge_prediction_sklearn/mlp_edge_prediction.py @@ -35,7 +35,7 @@ def __init__( training_unbalance_rate: float = 1.0, training_sample_only_edges_with_heterogeneous_node_types: bool = False, use_edge_metrics: bool = False, - use_zipfian_sampling: bool = True, + use_scale_free_distribution: bool = True, prediction_batch_size: int = 212, random_state: int = 42 ): @@ -90,7 +90,7 @@ def __init__( edge_embedding_method=edge_embedding_method, training_unbalance_rate=training_unbalance_rate, use_edge_metrics=use_edge_metrics, - use_zipfian_sampling=use_zipfian_sampling, + use_scale_free_distribution=use_scale_free_distribution, training_sample_only_edges_with_heterogeneous_node_types=training_sample_only_edges_with_heterogeneous_node_types, prediction_batch_size=prediction_batch_size, random_state=random_state diff --git a/embiggen/edge_prediction/edge_prediction_sklearn/random_forest_edge_prediction.py b/embiggen/edge_prediction/edge_prediction_sklearn/random_forest_edge_prediction.py index 39b6cb10..74511f36 100644 --- a/embiggen/edge_prediction/edge_prediction_sklearn/random_forest_edge_prediction.py +++ b/embiggen/edge_prediction/edge_prediction_sklearn/random_forest_edge_prediction.py @@ -16,7 +16,7 @@ def __init__( min_samples_split: int = 2, min_samples_leaf: int = 1, min_weight_fraction_leaf: float = 0., - max_features="auto", + max_features="sqrt", max_leaf_nodes=None, min_impurity_decrease=0., bootstrap=True, @@ -24,14 +24,13 @@ def __init__( n_jobs=-1, verbose=0, warm_start=False, - class_weight=None, ccp_alpha=0.0, max_samples=None, edge_embedding_method: str = "Concatenate", training_unbalance_rate: float = 1.0, training_sample_only_edges_with_heterogeneous_node_types: bool = False, use_edge_metrics: bool = False, - use_zipfian_sampling: bool = True, + use_scale_free_distribution: bool = True, prediction_batch_size: int = 212, random_state: int = 42 ): @@ -51,7 +50,6 @@ def __init__( self._random_state = random_state self._verbose = verbose self._warm_start = warm_start - self._class_weight = class_weight self._ccp_alpha = ccp_alpha self._max_samples = max_samples @@ -72,14 +70,13 @@ def __init__( random_state=random_state, verbose=verbose, warm_start=warm_start, - class_weight=class_weight, ccp_alpha=ccp_alpha, max_samples=max_samples ), edge_embedding_method=edge_embedding_method, training_unbalance_rate=training_unbalance_rate, use_edge_metrics=use_edge_metrics, - use_zipfian_sampling=use_zipfian_sampling, + use_scale_free_distribution=use_scale_free_distribution, training_sample_only_edges_with_heterogeneous_node_types=training_sample_only_edges_with_heterogeneous_node_types, prediction_batch_size=prediction_batch_size, random_state=random_state @@ -113,7 +110,6 @@ def parameters(self) -> Dict[str, Any]: random_state = self._random_state, verbose = self._verbose, warm_start = self._warm_start, - class_weight = self._class_weight, ccp_alpha = self._ccp_alpha, max_samples = self._max_samples, ) diff --git a/embiggen/edge_prediction/edge_prediction_sklearn/sklearn_edge_prediction_adapter.py b/embiggen/edge_prediction/edge_prediction_sklearn/sklearn_edge_prediction_adapter.py index 94bb955b..0ae0d0ea 100644 --- a/embiggen/edge_prediction/edge_prediction_sklearn/sklearn_edge_prediction_adapter.py +++ b/embiggen/edge_prediction/edge_prediction_sklearn/sklearn_edge_prediction_adapter.py @@ -24,8 +24,8 @@ def __init__( edge_embedding_method: str = "Concatenate", training_unbalance_rate: float = 1.0, training_sample_only_edges_with_heterogeneous_node_types: bool = False, + use_scale_free_distribution: bool = True, use_edge_metrics: bool = False, - use_zipfian_sampling: bool = True, prediction_batch_size: int = 215, random_state: int = 42 ): @@ -43,7 +43,7 @@ def __init__( Whether to sample negative edges exclusively between nodes with different node types to generate the negative edges used during the training of the model. This can be useful when executing a bipartite edge prediction task. - use_zipfian_sampling: bool = True + use_scale_free_distribution: bool = True Whether to sample the negative edges for the TRAINING of the model using a zipfian-like distribution that follows the degree distribution of the graph. This is generally useful, as these negative edges are less @@ -76,9 +76,8 @@ def __init__( self._training_unbalance_rate = training_unbalance_rate self._prediction_batch_size = prediction_batch_size self._use_edge_metrics = use_edge_metrics - self._use_zipfian_sampling = use_zipfian_sampling + self._use_scale_free_distribution = use_scale_free_distribution self._training_sample_only_edges_with_heterogeneous_node_types = training_sample_only_edges_with_heterogeneous_node_types - self._support = None # We want to mask the decorator class name self.__class__.__name__ = model_instance.__class__.__name__ self.__class__.__doc__ = model_instance.__class__.__doc__ @@ -91,7 +90,7 @@ def parameters(self) -> Dict[str, Any]: "training_unbalance_rate": self._training_unbalance_rate, "prediction_batch_size": self._prediction_batch_size, "use_edge_metrics": self._use_edge_metrics, - "use_zipfian_sampling": self._use_zipfian_sampling, + "use_scale_free_distribution": self._use_scale_free_distribution, super().parameters() } @@ -108,6 +107,8 @@ def _trasform_graph_into_edge_embedding( self, graph: Union[Graph, np.ndarray], node_features: List[np.ndarray], + support: Optional[Graph] = None, + node_types: Optional[Union[Graph, List[Optional[List[str]]], List[Optional[List[int]]]]] = None, node_type_features: Optional[List[np.ndarray]] = None, ) -> np.ndarray: """Transforms the provided data into an Sklearn-compatible numpy array. @@ -119,6 +120,16 @@ def _trasform_graph_into_edge_embedding( It can either be an Graph or a list of lists of edges. node_features: List[np.ndarray] The node features to be used in the training of the model. + support: Optional[Graph] = None + The graph describiding the topological structure that + includes also the above graph. This parameter + is mostly useful for topological classifiers + such as Graph Convolutional Networks. + node_types: Optional[Union[Graph, List[Optional[List[str]]], List[Optional[List[int]]]]] = None, + List of node types whose embedding is to be returned. + This can be either a list of strings, or a graph, or if the + aligned_mapping is setted, then this methods also accepts + a list of ints. node_type_features: Optional[List[np.ndarray]] = None, The node type features to be used in the training of the model. @@ -132,25 +143,20 @@ def _trasform_graph_into_edge_embedding( ValueError If the two graphs do not share the same node vocabulary. """ - if node_type_features is not None: - raise NotImplementedError( - "Support for node type features is not currently available for any " - "of the edge prediction models from the Sklearn library." - ) gt = GraphTransformer( method=self._edge_embedding_method, - aligned_node_mapping=True + aligned_mapping=True ) if self._use_edge_metrics: if isinstance(graph, Graph): - edge_features = self._support.get_all_edge_metrics( + edge_features = support.get_all_edge_metrics( normalize=True, subgraph=graph, ) elif isinstance(graph, tuple): - edge_features = self._support.get_all_edge_metrics_from_node_ids( + edge_features = support.get_all_edge_metrics_from_node_ids( graph, normalize=True, ) @@ -161,10 +167,14 @@ def _trasform_graph_into_edge_embedding( else: edge_features = None - gt.fit(node_features) + gt.fit( + node_features, + node_type_feature=node_type_features + ) return gt.transform( graph=graph, + node_types=node_types, edge_features=edge_features ) @@ -196,10 +206,13 @@ def _fit( """ lpt = EdgePredictionTransformer( method=self._edge_embedding_method, - aligned_node_mapping=True + aligned_mapping=True ) - lpt.fit(node_features) + lpt.fit( + node_features, + node_type_feature=node_type_features + ) if support is None: support = graph @@ -211,11 +224,11 @@ def _fit( ), random_state=self._random_state, sample_only_edges_with_heterogeneous_node_types=self._training_sample_only_edges_with_heterogeneous_node_types, - use_zipfian_sampling=self._use_zipfian_sampling + use_scale_free_distribution=self._use_scale_free_distribution ) if self._use_edge_metrics: - self._support = support + support = support edge_features = np.vstack(( support.get_all_edge_metrics( normalize=True, @@ -272,7 +285,9 @@ def _predict( return np.concatenate([ self._model_instance.predict(self._trasform_graph_into_edge_embedding( graph=edges[0], + support=support, node_features=node_features, + node_types=graph, node_type_features=node_type_features, )) for edges in tqdm( @@ -318,10 +333,12 @@ def _predict_proba( use_edge_metrics=False, batch_size=self._prediction_batch_size ) - return np.concatenate([ + prediction_probabilities = np.concatenate([ self._model_instance.predict_proba(self._trasform_graph_into_edge_embedding( graph=edges[0], + support=support, node_features=node_features, + node_types=graph, node_type_features=node_type_features, )) for edges in tqdm( @@ -333,6 +350,14 @@ def _predict_proba( ) ]) + # In the majority but not totality of sklearn models, + # the predictions of binary models are returned as + # a couple of vectors for the positive and negative class. + if len(prediction_probabilities.shape) > 1 and prediction_probabilities.shape[1] > 1: + prediction_probabilities = prediction_probabilities[:, 1] + + return prediction_probabilities + @classmethod def can_use_edge_weights(cls) -> bool: """Returns whether the model can optionally use edge weights.""" diff --git a/embiggen/edge_prediction/edge_prediction_tensorflow/gcn.py b/embiggen/edge_prediction/edge_prediction_tensorflow/gcn.py index 7e1b2dc5..11b7c30f 100644 --- a/embiggen/edge_prediction/edge_prediction_tensorflow/gcn.py +++ b/embiggen/edge_prediction/edge_prediction_tensorflow/gcn.py @@ -27,6 +27,7 @@ def __init__( number_of_units_per_ffnn_head_layer: Union[int, List[int]] = 128, dropout_rate: float = 0.2, apply_norm: bool = False, + combiner: str = "mean", edge_embedding_method: str = "Concatenate", optimizer: Union[str, Optimizer] = "adam", early_stopping_min_delta: float = 0.0001, @@ -81,6 +82,13 @@ def __init__( apply_norm: bool = False Whether to normalize the output of the convolution operations, after applying the level activations. + combiner: str = "mean" + A string specifying the reduction op. + Currently "mean", "sqrtn" and "sum" are supported. + "sum" computes the weighted sum of the embedding results for each row. + "mean" is the weighted sum divided by the total weight. + "sqrtn" is the weighted sum divided by the square root of the sum of the squares of the weights. + Defaults to mean. edge_embedding_method: str = "Concatenate" The edge embedding method to use to put togheter the source and destination node features, which includes: @@ -186,6 +194,7 @@ def __init__( number_of_units_per_ffnn_body_layer=number_of_units_per_ffnn_body_layer, number_of_units_per_ffnn_head_layer=number_of_units_per_ffnn_head_layer, dropout_rate=dropout_rate, + combiner=combiner, apply_norm=apply_norm, edge_embedding_method=edge_embedding_method, optimizer=optimizer, diff --git a/embiggen/edge_prediction/edge_prediction_tensorflow/gnn.py b/embiggen/edge_prediction/edge_prediction_tensorflow/gnn.py index a875709e..cda0e418 100644 --- a/embiggen/edge_prediction/edge_prediction_tensorflow/gnn.py +++ b/embiggen/edge_prediction/edge_prediction_tensorflow/gnn.py @@ -196,14 +196,17 @@ def parameters(self) -> Dict[str, Any]: "number_of_units_per_graph_convolution_layers", "handling_multi_graph", "number_of_units_per_ffnn_body_layer", - "number_of_units_per_ffnn_head_layer" + "number_of_units_per_ffnn_head_layer", + "combiner", + "apply_norm", + "dropout_rate" ] return dict( number_of_units_per_body_layer=self._number_of_units_per_ffnn_body_layer, number_of_units_per_head_layer=self._number_of_units_per_ffnn_head_layer, { key: value - for key, value in super().smoke_test_parameters().items() + for key, value in super().parameters().items() if key not in removed } ) diff --git a/embiggen/edge_prediction/edge_prediction_tensorflow/graph_sage.py b/embiggen/edge_prediction/edge_prediction_tensorflow/graph_sage.py index 67c1f9e1..d9923aa0 100644 --- a/embiggen/edge_prediction/edge_prediction_tensorflow/graph_sage.py +++ b/embiggen/edge_prediction/edge_prediction_tensorflow/graph_sage.py @@ -170,6 +170,7 @@ def __init__( number_of_units_per_ffnn_body_layer=number_of_units_per_ffnn_body_layer, number_of_units_per_ffnn_head_layer=number_of_units_per_ffnn_head_layer, dropout_rate=dropout_rate, + combiner="mean", apply_norm=True, edge_embedding_method=edge_embedding_method, optimizer=optimizer, @@ -202,7 +203,8 @@ def parameters(self) -> Dict[str, Any]: """Returns parameters for smoke test.""" removed = [ "apply_norm", - "use_simmetric_normalized_laplacian" + "use_simmetric_normalized_laplacian", + "combiner" ] return dict( { diff --git a/embiggen/edge_prediction/edge_prediction_tensorflow/kipf_gcn.py b/embiggen/edge_prediction/edge_prediction_tensorflow/kipf_gcn.py index 01f1b19c..e82a3c04 100644 --- a/embiggen/edge_prediction/edge_prediction_tensorflow/kipf_gcn.py +++ b/embiggen/edge_prediction/edge_prediction_tensorflow/kipf_gcn.py @@ -1,5 +1,5 @@ """Kipf GCN model for edge prediction.""" -from typing import List, Union, Optional, Type +from typing import List, Union, Optional, Type, Dict, Any from tensorflow.keras.optimizers import Optimizer from embiggen.edge_prediction.edge_prediction_tensorflow.gcn import GCNEdgePrediction @@ -174,6 +174,7 @@ def __init__( number_of_units_per_ffnn_body_layer=number_of_units_per_ffnn_body_layer, number_of_units_per_ffnn_head_layer=number_of_units_per_ffnn_head_layer, dropout_rate=dropout_rate, + combiner="sum", apply_norm=apply_norm, edge_embedding_method=edge_embedding_method, optimizer=optimizer, @@ -202,6 +203,19 @@ def __init__( verbose=verbose, ) + def parameters(self) -> Dict[str, Any]: + """Returns parameters for smoke test.""" + removed = [ + "combiner" + ] + return dict( + *{ + key: value + for key, value in super().parameters().items() + if key not in removed + } + ) + @classmethod def model_name(cls) -> str: return "Kipf GCN" \ No newline at end of file diff --git a/embiggen/embedders/__init__.py b/embiggen/embedders/__init__.py index 089bdf62..5e9b2513 100644 --- a/embiggen/embedders/__init__.py +++ b/embiggen/embedders/__init__.py @@ -2,6 +2,7 @@ from embiggen.embedders.ensmallen_embedders import from embiggen.embedders.tensorflow_embedders import * from embiggen.embedders.pykeen_embedders import * +from embiggen.embedders.non_existent_embedders import * from embiggen.embedders.karateclub_embedders import * from embiggen.embedders.graph_embedding_pipeline import embed_graph diff --git a/embiggen/embedders/ensmallen_embedders/deepwalk_cbow.py b/embiggen/embedders/ensmallen_embedders/deepwalk_cbow.py new file mode 100644 index 00000000..44d72145 --- /dev/null +++ b/embiggen/embedders/ensmallen_embedders/deepwalk_cbow.py @@ -0,0 +1,123 @@ +"""Module providing DeepWalk CBOW model implementation.""" +from typing import Optional, Dict, Any +from embiggen.embedders.ensmallen_embedders.node2vec import Node2VecEnsmallen + + +class DeepWalkCBOWEnsmallen(Node2VecEnsmallen): + """Class providing DeepWalk CBOW implemeted in Rust from Ensmallen.""" + + def __init__( + self, + embedding_size: int = 100, + epochs: int = 30, + clipping_value: float = 6.0, + number_of_negative_samples: int = 10, + walk_length: int = 128, + iterations: int = 10, + window_size: int = 5, + max_neighbours: Optional[int] = 100, + learning_rate: float = 0.01, + learning_rate_decay: float = 0.9, + central_nodes_embedding_path: Optional[str] = None, + contextual_nodes_embedding_path: Optional[str] = None, + normalize_by_degree: bool = False, + stochastic_downsample_by_degree: Optional[bool] = False, + normalize_learning_rate_by_degree: Optional[bool] = False, + use_scale_free_distribution: Optional[bool] = True, + random_state: int = 42, + enable_cache: bool = False + ): + """Create new abstract DeepWalk method. + + Parameters + -------------------------- + embedding_size: int = 100 + Dimension of the embedding. + epochs: int = 30 + Number of epochs to train the model for. + clipping_value: float = 6.0 + Value at which we clip the dot product, mostly for numerical stability issues. + By default, `6.0`, where the loss is already close to zero. + number_of_negative_samples: int = 10 + The number of negative classes to randomly sample per batch. + This single sample of negative classes is evaluated for each element in the batch. + walk_length: int = 128 + Maximal length of the walks. + iterations: int = 10 + Number of iterations of the single walks. + window_size: int = 5 + Window size for the local context. + On the borders the window size is trimmed. + max_neighbours: Optional[int] = 100 + Number of maximum neighbours to consider when using approximated walks. + By default, None, we execute exact random walks. + This is mainly useful for graphs containing nodes with high degrees. + learning_rate: float = 0.01 + The learning rate to use to train the Node2Vec model. By default 0.01. + learning_rate_decay: float = 0.9 + Factor to reduce the learning rate for at each epoch. By default 0.9. + central_nodes_embedding_path: Optional[str] = None + Path where to mmap and store the central nodes embedding. + This is necessary to embed large graphs whose embedding will not + fit into the available main memory. + contextual_nodes_embedding_path: Optional[str] = None + Path where to mmap and store the central nodes embedding. + This is necessary to embed large graphs whose embedding will not + fit into the available main memory. + normalize_by_degree: bool = False + Whether to normalize the random walk by the node degree + of the destination node degrees. + stochastic_downsample_by_degree: Optional[bool] = False + Randomly skip samples with probability proportional to the degree of the central node. By default false. + normalize_learning_rate_by_degree: Optional[bool] = False + Divide the learning rate by the degree of the central node. By default false. + use_scale_free_distribution: Optional[bool] = True + Sample negatives proportionally to their degree. By default true. + random_state: int = 42 + The random state to reproduce the training sequence. + enable_cache: bool = False + Whether to enable the cache, that is to + store the computed embedding. + """ + super().__init__( + embedding_size=embedding_size, + epochs=epochs, + clipping_value=clipping_value, + number_of_negative_samples=number_of_negative_samples, + walk_length=walk_length, + iterations=iterations, + window_size=window_size, + max_neighbours=max_neighbours, + learning_rate=learning_rate, + learning_rate_decay=learning_rate_decay, + central_nodes_embedding_path=central_nodes_embedding_path, + contextual_nodes_embedding_path=contextual_nodes_embedding_path, + normalize_by_degree=normalize_by_degree, + stochastic_downsample_by_degree=stochastic_downsample_by_degree, + normalize_learning_rate_by_degree=normalize_learning_rate_by_degree, + use_scale_free_distribution=use_scale_free_distribution, + random_state=random_state, + enable_cache=enable_cache + ) + + def parameters(self) -> Dict[str, Any]: + """Returns parameters for smoke test.""" + removed = [ + "return_weight", + "explore_weight", + "change_node_type_weight", + "change_edge_type_weight", + "alpha" + ] + return dict( + { + key: value + for key, value in super().parameters().items() + if key not in removed + } + ) + + @classmethod + def model_name(cls) -> str: + """Returns name of the model.""" + return "DeepWalk CBOW" \ No newline at end of file diff --git a/embiggen/embedders/ensmallen_embedders/deepwalk_glove.py b/embiggen/embedders/ensmallen_embedders/deepwalk_glove.py new file mode 100644 index 00000000..715bfbd1 --- /dev/null +++ b/embiggen/embedders/ensmallen_embedders/deepwalk_glove.py @@ -0,0 +1,112 @@ +"""Module providing DeepWalk GloVe model implementation.""" +from typing import Optional, Dict, Any +from embiggen.embedders.ensmallen_embedders.node2vec import Node2VecEnsmallen + +class DeepWalkGloVeEnsmallen(Node2VecEnsmallen): + """Class providing DeepWalk GloVe implemeted in Rust from Ensmallen.""" + + def __init__( + self, + embedding_size: int = 100, + alpha: float = 0.75, + epochs: int = 30, + clipping_value: float = 6.0, + walk_length: int = 128, + iterations: int = 10, + window_size: int = 5, + max_neighbours: Optional[int] = 100, + learning_rate: float = 0.001, + learning_rate_decay: float = 0.9, + central_nodes_embedding_path: Optional[str] = None, + contextual_nodes_embedding_path: Optional[str] = None, + normalize_by_degree: bool = False, + random_state: int = 42, + enable_cache: bool = False + ): + """Create new abstract DeepWalk method. + + Parameters + -------------------------- + embedding_size: int = 100 + Dimension of the embedding. + alpha: float = 0.75 + Alpha parameter for GloVe's loss. + epochs: int = 100 + Number of epochs to train the model for. + window_size: int = 10 + Window size for the local context. + On the borders the window size is trimmed. + clipping_value: float = 6.0 + Value at which we clip the dot product, mostly for numerical stability issues. + By default, `6.0`, where the loss is already close to zero. + walk_length: int = 128 + Maximal length of the walks. + iterations: int = 10 + Number of iterations of the single walks. + window_size: int = 5 + Window size for the local context. + On the borders the window size is trimmed. + max_neighbours: Optional[int] = 100 + Number of maximum neighbours to consider when using approximated walks. + By default, None, we execute exact random walks. + This is mainly useful for graphs containing nodes with high degrees. + learning_rate: float = 0.001 + The learning rate to use to train the DeepWalk model. By default 0.01. + central_nodes_embedding_path: Optional[str] = None + Path where to mmap and store the central nodes embedding. + This is necessary to embed large graphs whose embedding will not + fit into the available main memory. + contextual_nodes_embedding_path: Optional[str] = None + Path where to mmap and store the central nodes embedding. + This is necessary to embed large graphs whose embedding will not + fit into the available main memory. + learning_rate_decay: float = 0.9 + Factor to reduce the learning rate for at each epoch. By default 0.9. + normalize_by_degree: bool = False + Whether to normalize the random walk by the node degree + of the destination node degrees. + random_state: int = 42 + The random state to reproduce the training sequence. + enable_cache: bool = False + Whether to enable the cache, that is to + store the computed embedding. + """ + super().__init__( + embedding_size=embedding_size, + alpha=alpha, + epochs=epochs, + clipping_value=clipping_value, + walk_length=walk_length, + iterations=iterations, + window_size=window_size, + max_neighbours=max_neighbours, + learning_rate=learning_rate, + learning_rate_decay=learning_rate_decay, + central_nodes_embedding_path=central_nodes_embedding_path, + contextual_nodes_embedding_path=contextual_nodes_embedding_path, + normalize_by_degree=normalize_by_degree, + enable_cache=enable_cache, + random_state=random_state + ) + + def parameters(self) -> Dict[str, Any]: + """Returns parameters for smoke test.""" + removed = [ + "return_weight", + "explore_weight", + "change_node_type_weight", + "change_edge_type_weight", + "number_of_negative_samples" + ] + return dict( + { + key: value + for key, value in super().parameters().items() + if key not in removed + } + ) + + @classmethod + def model_name(cls) -> str: + """Returns name of the model.""" + return "DeepWalk GloVe" \ No newline at end of file diff --git a/embiggen/embedders/ensmallen_embedders/deepwalk_skipgram.py b/embiggen/embedders/ensmallen_embedders/deepwalk_skipgram.py new file mode 100644 index 00000000..6d160ea6 --- /dev/null +++ b/embiggen/embedders/ensmallen_embedders/deepwalk_skipgram.py @@ -0,0 +1,123 @@ +"""Module providing DeepWalk SkipGram model implementation.""" +from typing import Optional, Dict, Any +from embiggen.embedders.ensmallen_embedders.node2vec import Node2VecEnsmallen + + +class DeepWalkSkipGramEnsmallen(Node2VecEnsmallen): + """Class providing DeepWalk SkipGram implemeted in Rust from Ensmallen.""" + + def __init__( + self, + embedding_size: int = 100, + epochs: int = 30, + clipping_value: float = 6.0, + number_of_negative_samples: int = 10, + walk_length: int = 128, + iterations: int = 10, + window_size: int = 5, + max_neighbours: Optional[int] = 100, + learning_rate: float = 0.01, + learning_rate_decay: float = 0.9, + central_nodes_embedding_path: Optional[str] = None, + contextual_nodes_embedding_path: Optional[str] = None, + normalize_by_degree: bool = False, + stochastic_downsample_by_degree: Optional[bool] = False, + normalize_learning_rate_by_degree: Optional[bool] = False, + use_scale_free_distribution: Optional[bool] = True, + random_state: int = 42, + enable_cache: bool = False + ): + """Create new abstract DeepWalk method. + + Parameters + -------------------------- + embedding_size: int = 100 + Dimension of the embedding. + epochs: int = 30 + Number of epochs to train the model for. + clipping_value: float = 6.0 + Value at which we clip the dot product, mostly for numerical stability issues. + By default, `6.0`, where the loss is already close to zero. + number_of_negative_samples: int = 10 + The number of negative classes to randomly sample per batch. + This single sample of negative classes is evaluated for each element in the batch. + walk_length: int = 128 + Maximal length of the walks. + iterations: int = 10 + Number of iterations of the single walks. + window_size: int = 5 + Window size for the local context. + On the borders the window size is trimmed. + max_neighbours: Optional[int] = 100 + Number of maximum neighbours to consider when using approximated walks. + By default, None, we execute exact random walks. + This is mainly useful for graphs containing nodes with high degrees. + learning_rate: float = 0.01 + The learning rate to use to train the Node2Vec model. By default 0.01. + learning_rate_decay: float = 0.9 + Factor to reduce the learning rate for at each epoch. By default 0.9. + central_nodes_embedding_path: Optional[str] = None + Path where to mmap and store the central nodes embedding. + This is necessary to embed large graphs whose embedding will not + fit into the available main memory. + contextual_nodes_embedding_path: Optional[str] = None + Path where to mmap and store the central nodes embedding. + This is necessary to embed large graphs whose embedding will not + fit into the available main memory. + normalize_by_degree: bool = False + Whether to normalize the random walk by the node degree + of the destination node degrees. + stochastic_downsample_by_degree: Optional[bool] = False + Randomly skip samples with probability proportional to the degree of the central node. By default false. + normalize_learning_rate_by_degree: Optional[bool] = False + Divide the learning rate by the degree of the central node. By default false. + use_scale_free_distribution: Optional[bool] = True + Sample negatives proportionally to their degree. By default true. + random_state: int = 42 + The random state to reproduce the training sequence. + enable_cache: bool = False + Whether to enable the cache, that is to + store the computed embedding. + """ + super().__init__( + embedding_size=embedding_size, + epochs=epochs, + clipping_value=clipping_value, + number_of_negative_samples=number_of_negative_samples, + walk_length=walk_length, + iterations=iterations, + window_size=window_size, + max_neighbours=max_neighbours, + learning_rate=learning_rate, + learning_rate_decay=learning_rate_decay, + central_nodes_embedding_path=central_nodes_embedding_path, + contextual_nodes_embedding_path=contextual_nodes_embedding_path, + normalize_by_degree=normalize_by_degree, + stochastic_downsample_by_degree=stochastic_downsample_by_degree, + normalize_learning_rate_by_degree=normalize_learning_rate_by_degree, + use_scale_free_distribution=use_scale_free_distribution, + random_state=random_state, + enable_cache=enable_cache + ) + + def parameters(self) -> Dict[str, Any]: + """Returns parameters for smoke test.""" + removed = [ + "return_weight", + "explore_weight", + "change_node_type_weight", + "change_edge_type_weight", + "alpha" + ] + return dict( + { + key: value + for key, value in super().parameters().items() + if key not in removed + } + ) + + @classmethod + def model_name(cls) -> str: + """Returns name of the model.""" + return "DeepWalk SkipGram" \ No newline at end of file diff --git a/embiggen/embedders/ensmallen_embedders/spine.py b/embiggen/embedders/ensmallen_embedders/degree_spine.py similarity index 63% rename from embiggen/embedders/ensmallen_embedders/spine.py rename to embiggen/embedders/ensmallen_embedders/degree_spine.py index 4b0eca3d..e8bc7995 100644 --- a/embiggen/embedders/ensmallen_embedders/spine.py +++ b/embiggen/embedders/ensmallen_embedders/degree_spine.py @@ -1,35 +1,53 @@ -"""Module providing SPINE implementation.""" -from typing import Optional, Dict, Any, Union +"""Module providing Degree-based SPINE implementation.""" +from typing import Optional, Dict, Any from ensmallen import Graph -import numpy as np import pandas as pd from ensmallen import models -from embiggen.utils.abstract_models import AbstractEmbeddingModel, EmbeddingResult +from embiggen.embedders.ensmallen_embedders.ensmallen_embedder import EnsmallenEmbedder +from embiggen.utils import EmbeddingResult -class SPINE(AbstractEmbeddingModel): - """Class implementing the SPINE algorithm.""" +class DegreeSPINE(EnsmallenEmbedder): + """Class implementing the Degree-based SPINE algorithm.""" def __init__( self, embedding_size: int = 100, dtype: Optional[str] = "u8", + maximum_depth: Optional[int] = None, + path: Optional[str] = None, + verbose: bool = False, enable_cache: bool = False ): - """Create new SPINE method. + """Create new Degree-based SPINE method. Parameters -------------------------- embedding_size: int = 100 Dimension of the embedding. dtype: Optional[str] = "u8" - Dtype to use for the embedding. Note that an improper dtype may cause overflows. + Dtype to use for the embedding. + maximum_depth: Optional[int] = None + Maximum depth of the shortest path. + path: Optional[str] = None + Path where to store the mmap-ed embedding. + This parameter is necessary to embed very large graphs. + verbose: bool = False + Whether to show loading bars. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. """ self._dtype = dtype - self._model = models.SPINE(embedding_size=embedding_size) + self._verbose = verbose + self._maximum_depth = maximum_depth + self._path = path + self._model = models.DegreeSPINE( + embedding_size=embedding_size, + verbose=self._verbose, + maximum_depth=self._maximum_depth, + path=self._path + ) super().__init__( embedding_size=embedding_size, @@ -41,7 +59,9 @@ def parameters(self) -> Dict[str, Any]: return { super().parameters(), dict( - dtype=self._dtype + dtype=self._dtype, + maximum_depth=self._maximum_depth, + path=self._path, ) } @classmethod @@ -55,14 +75,12 @@ def _fit_transform( self, graph: Graph, return_dataframe: bool = True, - verbose: bool = True ) -> EmbeddingResult: """Return node embedding.""" node_embedding = self._model.fit_transform( graph, dtype=self._dtype, - verbose=verbose, - ).T + ) if return_dataframe: node_embedding = pd.DataFrame( node_embedding, @@ -73,26 +91,10 @@ def _fit_transform( node_embeddings=node_embedding ) - @classmethod - def task_name(cls) -> str: - return "Node Embedding" - @classmethod def model_name(cls) -> str: """Returns name of the model.""" - return "SPINE" - - @classmethod - def library_name(cls) -> str: - return "Ensmallen" - - @classmethod - def requires_nodes_sorted_by_decreasing_node_degree(cls) -> bool: - return False - - @classmethod - def is_topological(cls) -> bool: - return True + return "Degree-based SPINE" @classmethod def can_use_node_types(cls) -> bool: diff --git a/embiggen/embedders/ensmallen_embedders/degree_wine.py b/embiggen/embedders/ensmallen_embedders/degree_wine.py new file mode 100644 index 00000000..5b848a16 --- /dev/null +++ b/embiggen/embedders/ensmallen_embedders/degree_wine.py @@ -0,0 +1,118 @@ +"""Module providing Degree-based WINE implementation.""" +from typing import Optional, Dict, Any +from ensmallen import Graph +import pandas as pd +from ensmallen import models +from embiggen.embedders.ensmallen_embedders.ensmallen_embedder import EnsmallenEmbedder +from embiggen.utils import EmbeddingResult + + +class DegreeWINE(EnsmallenEmbedder): + """Class implementing the Degree-based WINE algorithm.""" + + def __init__( + self, + embedding_size: int = 100, + dtype: Optional[str] = "u8", + walk_length: int = 2, + path: Optional[str] = None, + verbose: bool = False, + enable_cache: bool = False + ): + """Create new Degree-based WINE method. + + Parameters + -------------------------- + embedding_size: int = 100 + Dimension of the embedding. + dtype: Optional[str] = "u8" + Dtype to use for the embedding. + walk_length: int = 2 + Length of the random walk. + By default 2, to capture exclusively the immediate context. + path: Optional[str] = None + Path where to store the mmap-ed embedding. + This parameter is necessary to embed very large graphs. + verbose: bool = False + Whether to show loading bars. + enable_cache: bool = False + Whether to enable the cache, that is to + store the computed embedding. + """ + self._dtype = dtype + self._verbose = verbose + self._walk_length = walk_length + self._path = path + self._model = models.DegreeWINE( + embedding_size=embedding_size, + verbose=self._verbose, + walk_length=self._walk_length, + path=self._path + ) + + super().__init__( + embedding_size=embedding_size, + enable_cache=enable_cache + ) + + def parameters(self) -> Dict[str, Any]: + """Returns parameters of the model.""" + return { + super().parameters(), + dict( + dtype=self._dtype, + walk_length=self._walk_length, + path=self._path, + ) + } + @classmethod + def smoke_test_parameters(cls) -> Dict[str, Any]: + """Returns parameters for smoke test.""" + return dict( + embedding_size=5, + ) + + def _fit_transform( + self, + graph: Graph, + return_dataframe: bool = True, + ) -> EmbeddingResult: + """Return node embedding.""" + node_embedding = self._model.fit_transform( + graph, + dtype=self._dtype, + ) + if return_dataframe: + node_embedding = pd.DataFrame( + node_embedding, + index=graph.get_node_names() + ) + return EmbeddingResult( + embedding_method_name=self.model_name(), + node_embeddings=node_embedding + ) + + @classmethod + def model_name(cls) -> str: + """Returns name of the model.""" + return "Degree-based WINE" + + @classmethod + def can_use_node_types(cls) -> bool: + """Returns whether the model can optionally use node types.""" + return False + + @classmethod + def can_use_edge_weights(cls) -> bool: + """Returns whether the model can optionally use edge weights.""" + return False + + @classmethod + def can_use_edge_types(cls) -> bool: + """Returns whether the model can optionally use edge types.""" + return False + + @classmethod + def is_stocastic(cls) -> bool: + """Returns whether the model is stocastic and has therefore a random state.""" + return False \ No newline at end of file diff --git a/embiggen/embedders/ensmallen_embedders/ensmallen_embedder.py b/embiggen/embedders/ensmallen_embedders/ensmallen_embedder.py new file mode 100644 index 00000000..7232edeb --- /dev/null +++ b/embiggen/embedders/ensmallen_embedders/ensmallen_embedder.py @@ -0,0 +1,51 @@ +"""Module providing SocioDim implementation.""" +from typing import Optional, Dict, Any +from ensmallen import Graph +import pandas as pd +import numpy as np +from embiggen.utils.abstract_models import AbstractEmbeddingModel, abstract_class + +@abstract_class +class EnsmallenEmbedder(AbstractEmbeddingModel): + """Class implementing the Ensmallen Embedder algorithm.""" + + def __init__( + self, + random_state: Optional[int] = None, + embedding_size: int = 100, + enable_cache: bool = False + ): + + """Create new EnsmallenEmbedder method. + + Parameters + -------------------------- + random_state: Optional[int] = None + Random state to reproduce the embeddings. + embedding_size: int = 100 + Dimension of the embedding. + enable_cache: bool = False + Whether to enable the cache, that is to + store the computed embedding. + """ + super().__init__( + random_state=random_state, + embedding_size=embedding_size, + enable_cache=enable_cache, + ) + + @classmethod + def task_name(cls) -> str: + return "Node Embedding" + + @classmethod + def library_name(cls) -> str: + return "Ensmallen" + + @classmethod + def requires_nodes_sorted_by_decreasing_node_degree(cls) -> bool: + return False + + @classmethod + def is_topological(cls) -> bool: + return True \ No newline at end of file diff --git a/embiggen/embedders/ensmallen_embedders/first_order_line.py b/embiggen/embedders/ensmallen_embedders/first_order_line.py new file mode 100644 index 00000000..e5bd91b8 --- /dev/null +++ b/embiggen/embedders/ensmallen_embedders/first_order_line.py @@ -0,0 +1,131 @@ +"""Module providing first-order LINE implementation.""" +from typing import Dict, Any, Optional +from ensmallen import Graph +import pandas as pd +from ensmallen import models +from embiggen.embedders.ensmallen_embedders.ensmallen_embedder import EnsmallenEmbedder +from embiggen.utils import EmbeddingResult + + +class FirstOrderLINEEnsmallen(EnsmallenEmbedder): + """Class implementing the first-order LINE algorithm.""" + + def __init__( + self, + embedding_size: int = 100, + epochs: int = 100, + learning_rate: float = 0.05, + learning_rate_decay: float = 0.9, + avoid_false_negatives: bool = False, + node_embedding_path: Optional[str] = None, + random_state: int = 42, + verbose: bool = False, + enable_cache: bool = False + ): + """Create new abstract Node2Vec method. + + Parameters + -------------------------- + embedding_size: int = 100 + Dimension of the embedding. + epochs: int = 100 + The number of epochs to run the model for, by default 10. + learning_rate: float = 0.05 + The learning rate to update the gradient, by default 0.01. + learning_rate_decay: float = 0.9 + Factor to reduce the learning rate for at each epoch. By default 0.9. + avoid_false_negatives: bool = False + Whether to avoid sampling false negatives. + This may cause a slower training. + node_embedding_path: Optional[str] = None + Path where to mmap and store the nodes embedding. + This is necessary to embed large graphs whose embedding will not + fit into the available main memory. + random_state: int = 42 + Random state to reproduce the embeddings. + verbose: bool = False + Whether to show loading bars. + enable_cache: bool = False + Whether to enable the cache, that is to + store the computed embedding. + """ + self._kwargs = dict( + epochs=epochs, + learning_rate=learning_rate, + learning_rate_decay=learning_rate_decay, + avoid_false_negatives=avoid_false_negatives, + node_embedding_path=node_embedding_path, + verbose=verbose + ) + + self._model = models.FirstOrderLINE( + self._kwargs, + embedding_size=embedding_size, + random_state=random_state + ) + + super().__init__( + embedding_size=embedding_size, + enable_cache=enable_cache, + random_state=random_state + ) + + def parameters(self) -> Dict[str, Any]: + """Returns parameters of the model.""" + return dict( + super().parameters(), + self._kwargs, + ) + + @classmethod + def smoke_test_parameters(cls) -> Dict[str, Any]: + """Returns parameters for smoke test.""" + return dict( + embedding_size=5, + epochs=1 + ) + + def _fit_transform( + self, + graph: Graph, + return_dataframe: bool = True, + ) -> EmbeddingResult: + """Return node embedding.""" + node_embedding = self._model.fit_transform( + graph + ) + if return_dataframe: + node_embedding = pd.DataFrame( + node_embedding, + index=graph.get_node_names() + ) + + return EmbeddingResult( + embedding_method_name=self.model_name(), + node_embeddings= node_embedding, + ) + + @classmethod + def model_name(cls) -> str: + """Returns name of the model.""" + return "First-order LINE" + + @classmethod + def can_use_edge_weights(cls) -> bool: + """Returns whether the model can optionally use edge weights.""" + return False + + @classmethod + def can_use_node_types(cls) -> bool: + """Returns whether the model can optionally use node types.""" + return False + + @classmethod + def can_use_edge_types(cls) -> bool: + """Returns whether the model can optionally use edge types.""" + return False + + @classmethod + def is_stocastic(cls) -> bool: + """Returns whether the model is stocastic and has therefore a random state.""" + return True \ No newline at end of file diff --git a/embiggen/embedders/ensmallen_embedders/geometric_laplacian_eigenmaps.py b/embiggen/embedders/ensmallen_embedders/geometric_laplacian_eigenmaps.py index 15d487b4..7862ec90 100644 --- a/embiggen/embedders/ensmallen_embedders/geometric_laplacian_eigenmaps.py +++ b/embiggen/embedders/ensmallen_embedders/geometric_laplacian_eigenmaps.py @@ -5,10 +5,11 @@ import numpy as np from scipy.sparse import coo_matrix from scipy.sparse.linalg import eigsh -from embiggen.utils.abstract_models import AbstractEmbeddingModel, EmbeddingResult +from embiggen.embedders.ensmallen_embedders.ensmallen_embedder import EnsmallenEmbedder +from embiggen.utils import EmbeddingResult -class GLEEEnsmallen(AbstractEmbeddingModel): +class GLEEEnsmallen(EnsmallenEmbedder): """Class implementing the GLEE algorithm.""" def __init__( @@ -35,7 +36,6 @@ def _fit_transform( self, graph: Graph, return_dataframe: bool = True, - verbose: bool = True ) -> EmbeddingResult: """Return node embedding.""" edges, weights = graph.get_symmetric_normalized_laplacian_coo_matrix() @@ -67,27 +67,11 @@ def _fit_transform( node_embeddings=embedding ) - @classmethod - def task_name(cls) -> str: - return "Node Embedding" - @classmethod def model_name(cls) -> str: """Returns name of the model.""" return "GLEE" - @classmethod - def library_name(cls) -> str: - return "Ensmallen" - - @classmethod - def requires_nodes_sorted_by_decreasing_node_degree(cls) -> bool: - return False - - @classmethod - def is_topological(cls) -> bool: - return True - @classmethod def can_use_edge_weights(cls) -> bool: """Returns whether the model can optionally use edge weights.""" diff --git a/embiggen/embedders/ensmallen_embedders/glove.py b/embiggen/embedders/ensmallen_embedders/glove.py deleted file mode 100644 index 5f109884..00000000 --- a/embiggen/embedders/ensmallen_embedders/glove.py +++ /dev/null @@ -1,239 +0,0 @@ -"""Module providing GloVe model implementation.""" -from typing import Optional, Dict, Any, Union -from ensmallen import Graph -from ensmallen import models -import numpy as np -import pandas as pd -from embiggen.utils.abstract_models import AbstractEmbeddingModel, EmbeddingResult - - -class GloVeEnsmallen(AbstractEmbeddingModel): - """Class providing GloVe implemeted in Rust from Ensmallen.""" - - def __init__( - self, - embedding_size: int = 100, - alpha: float = 0.75, - epochs: int = 500, - clipping_value: float = 6.0, - walk_length: int = 128, - iterations: int = 10, - window_size: int = 10, - return_weight: float = 1.0, - explore_weight: float = 1.0, - change_node_type_weight: float = 1.0, - change_edge_type_weight: float = 1.0, - max_neighbours: Optional[int] = 100, - learning_rate: float = 0.01, - learning_rate_decay: float = 0.9, - normalize_by_degree: bool = False, - random_state: int = 42, - enable_cache: bool = False - ): - """Create new abstract Node2Vec method. - - Parameters - -------------------------- - embedding_size: int = 100 - Dimension of the embedding. - alpha: float = 0.75 - Alpha parameter for GloVe's loss. - epochs: int = 100 - Number of epochs to train the model for. - window_size: int = 10 - Window size for the local context. - On the borders the window size is trimmed. - clipping_value: float = 6.0 - Value at which we clip the dot product, mostly for numerical stability issues. - By default, `6.0`, where the loss is already close to zero. - walk_length: int = 128 - Maximal length of the walks. - iterations: int = 10 - Number of iterations of the single walks. - window_size: int = 10 - Window size for the local context. - On the borders the window size is trimmed. - return_weight: float = 1.0 - Weight on the probability of returning to the same node the walk just came from - Having this higher tends the walks to be - more like a Breadth-First Search. - Having this very high (> 2) makes search very local. - Equal to the inverse of p in the Node2Vec paper. - explore_weight: float = 1.0 - Weight on the probability of visiting a neighbor node - to the one we're coming from in the random walk - Having this higher tends the walks to be - more like a Depth-First Search. - Having this very high makes search more outward. - Having this very low makes search very local. - Equal to the inverse of q in the Node2Vec paper. - change_node_type_weight: float = 1.0 - Weight on the probability of visiting a neighbor node of a - different type than the previous node. This only applies to - colored graphs, otherwise it has no impact. - change_edge_type_weight: float = 1.0 - Weight on the probability of visiting a neighbor edge of a - different type than the previous edge. This only applies to - multigraphs, otherwise it has no impact. - max_neighbours: Optional[int] = 100 - Number of maximum neighbours to consider when using approximated walks. - By default, None, we execute exact random walks. - This is mainly useful for graphs containing nodes with high degrees. - learning_rate: float = 0.01 - The learning rate to use to train the Node2Vec model. By default 0.01. - learning_rate_decay: float = 0.9 - Factor to reduce the learning rate for at each epoch. By default 0.9. - normalize_by_degree: bool = False - Whether to normalize the random walk by the node degree - of the destination node degrees. - random_state: int = 42 - The random state to reproduce the training sequence. - enable_cache: bool = False - Whether to enable the cache, that is to - store the computed embedding. - """ - - self._epochs = epochs - self._learning_rate = learning_rate - self._learning_rate_decay = learning_rate_decay - - self._model_kwargs = dict( - alpha=alpha, - embedding_size=embedding_size, - clipping_value=clipping_value, - walk_length=walk_length, - iterations=iterations, - window_size=window_size, - return_weight=return_weight, - explore_weight=explore_weight, - change_edge_type_weight=change_edge_type_weight, - change_node_type_weight=change_node_type_weight, - max_neighbours=max_neighbours, - normalize_by_degree=normalize_by_degree, - ) - - self._model = models.GloVe( - self._model_kwargs, - random_state=random_state - ) - - super().__init__( - embedding_size=embedding_size, - enable_cache=enable_cache, - random_state=random_state - ) - - @classmethod - def smoke_test_parameters(cls) -> Dict[str, Any]: - """Returns parameters for smoke test.""" - return dict( - embedding_size=5, - epochs=1, - window_size=1, - walk_length=4, - iterations=1, - max_neighbours=10, - ) - - def parameters(self) -> Dict[str, Any]: - """Returns parameters of the model.""" - return { - super().parameters(), - self._model_kwargs, - dict( - epochs=self._epochs, - learning_rate=self._learning_rate, - learning_rate_decay=self._learning_rate_decay, - ) - } - - @classmethod - def task_name(cls) -> str: - return "Node Embedding" - - @classmethod - def library_name(cls) -> str: - return "Ensmallen" - - def requires_nodes_sorted_by_decreasing_node_degree(self) -> bool: - return False - - def is_topological(self) -> bool: - return True - - def _fit_transform( - self, - graph: Graph, - return_dataframe: bool = True, - verbose: bool = True - ) -> EmbeddingResult: - """Return node embedding.""" - node_embedding = self._model.fit_transform( - graph, - epochs=self._epochs, - learning_rate=self._learning_rate, - learning_rate_decay=self._learning_rate_decay, - verbose=verbose - ) - if return_dataframe: - node_embedding = pd.DataFrame( - node_embedding, - index=graph.get_node_names() - ) - return EmbeddingResult( - embedding_method_name=self.model_name(), - node_embeddings=node_embedding - ) - - @classmethod - def requires_edge_weights(cls) -> bool: - return False - - @classmethod - def requires_positive_edge_weights(cls) -> bool: - return True - - @classmethod - def model_name(cls) -> str: - """Returns name of the model.""" - return "GloVe" - - @classmethod - def requires_node_types(cls) -> bool: - return False - - @classmethod - def requires_edge_types(cls) -> bool: - return False - - @classmethod - def can_use_edge_weights(cls) -> bool: - """Returns whether the model can optionally use edge weights.""" - return True - - def is_using_edge_weights(self) -> bool: - """Returns whether the model is parametrized to use edge weights.""" - return True - - @classmethod - def can_use_node_types(cls) -> bool: - """Returns whether the model can optionally use node types.""" - return True - - def is_using_node_types(self) -> bool: - """Returns whether the model is parametrized to use node types.""" - return self._model_kwargs["change_node_type_weight"] != 1.0 - - @classmethod - def can_use_edge_types(cls) -> bool: - """Returns whether the model can optionally use edge types.""" - return True - - def is_using_edge_types(self) -> bool: - """Returns whether the model is parametrized to use edge types.""" - return self._model_kwargs["change_edge_type_weight"] != 1.0 - - @classmethod - def is_stocastic(cls) -> bool: - """Returns whether the model is stocastic and has therefore a random state.""" - return True \ No newline at end of file diff --git a/embiggen/embedders/ensmallen_embedders/hope.py b/embiggen/embedders/ensmallen_embedders/hope.py index c6ec985f..ed85a408 100644 --- a/embiggen/embedders/ensmallen_embedders/hope.py +++ b/embiggen/embedders/ensmallen_embedders/hope.py @@ -6,17 +6,20 @@ from scipy.sparse import coo_matrix from scipy.sparse.linalg import svds as sparse_svds from sklearn.utils.extmath import randomized_svd -from embiggen.utils.abstract_models import AbstractEmbeddingModel, EmbeddingResult, format_list +from userinput.utils import must_be_in_set +from embiggen.embedders.ensmallen_embedders.ensmallen_embedder import EnsmallenEmbedder +from embiggen.utils import EmbeddingResult -class HOPEEnsmallen(AbstractEmbeddingModel): +class HOPEEnsmallen(EnsmallenEmbedder): """Class implementing the HOPE algorithm.""" def __init__( self, embedding_size: int = 100, - metric: str = "Jaccard", + metric: str = "Neighbours Intersection size", root_node_name: Optional[str] = None, + verbose: bool = False, enable_cache: bool = False ): """Create new HOPE method. @@ -25,7 +28,7 @@ def __init__( -------------------------- embedding_size: int = 100 Dimension of the embedding. - metric: str = "Jaccard" + metric: str = "Neighbours Intersection size" The metric to use. You can either use: - Jaccard @@ -38,18 +41,17 @@ def __init__( - Left Normalized Laplacian - Right Normalized Laplacian - Symmetric Normalized Laplacian + - Resnik root_node_name: Optional[str] = None Root node to use when the ancestors mode for the Jaccard index is selected. + verbose: bool = False + Whether to show loading bars. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. """ - if metric not in self.get_available_metrics(): - raise ValueError( - f"The provided metric {metric} is not a supported " - f"metric. The supported metrics are: {format_list(self.get_available_metrics())}." - ) + metric = must_be_in_set(metric, self.get_available_metrics(), "metric") ancestral_metric = ("Ancestors Jaccard", "Ancestors size") if root_node_name is None and metric in ancestral_metric: raise ValueError( @@ -65,6 +67,7 @@ def __init__( self._metric = metric self._root_node_name = root_node_name + self._verbose = verbose super().__init__( embedding_size=embedding_size, @@ -73,19 +76,21 @@ def __init__( def parameters(self) -> Dict[str, Any]: """Returns parameters of the model.""" - return { + return dict( super().parameters(), dict( metric=self._metric, + verbose=self._verbose, root_node_name=self._root_node_name ) - } + ) @classmethod def get_available_metrics(cls) -> List[str]: """Returns list of the available metrics.""" return [ "Jaccard", + "Shortest Paths", "Neighbours Intersection size", "Ancestors Jaccard", "Ancestors size", @@ -102,7 +107,6 @@ def _fit_transform( self, graph: Graph, return_dataframe: bool = True, - verbose: bool = True ) -> EmbeddingResult: """Return node embedding.""" matrix = None @@ -110,6 +114,8 @@ def _fit_transform( edges, weights = graph.get_jaccard_coo_matrix() elif self._metric == "Laplacian": edges, weights = graph.get_laplacian_coo_matrix() + elif self._metric == "Shortest Paths": + matrix = graph.get_shortest_paths_matrix() elif self._metric == "Modularity": matrix = graph.get_dense_modularity_matrix() elif self._metric == "Left Normalized Laplacian": @@ -126,7 +132,7 @@ def _fit_transform( src_node_name=self._root_node_name, compute_predecessors=True ), - verbose=verbose + verbose=self._verbose ) elif self._metric == "Ancestors size": matrix = graph.get_shared_ancestors_size_adjacency_matrix( @@ -134,18 +140,13 @@ def _fit_transform( src_node_name=self._root_node_name, compute_predecessors=True ), - verbose=verbose + verbose=self._verbose ) elif self._metric == "Adamic-Adar": edges, weights = graph.get_adamic_adar_coo_matrix() elif self._metric == "Adjacency": edges, weights = graph.get_directed_edge_node_ids(), np.ones( graph.get_number_of_directed_edges()) - else: - raise NotImplementedError( - f"The provided metric {self._metric} " - "is not currently supported." - ) if matrix is None: matrix = coo_matrix( @@ -186,27 +187,11 @@ def _fit_transform( node_embeddings=[left_embedding, right_embedding] ) - @classmethod - def task_name(cls) -> str: - return "Node Embedding" - @classmethod def model_name(cls) -> str: """Returns name of the model.""" return "HOPE" - @classmethod - def library_name(cls) -> str: - return "Ensmallen" - - @classmethod - def requires_nodes_sorted_by_decreasing_node_degree(cls) -> bool: - return False - - @classmethod - def is_topological(cls) -> bool: - return True - @classmethod def can_use_edge_weights(cls) -> bool: """Returns whether the model can optionally use edge weights.""" diff --git a/embiggen/embedders/ensmallen_embedders/laplacian_eigenmaps.py b/embiggen/embedders/ensmallen_embedders/laplacian_eigenmaps.py new file mode 100644 index 00000000..899170b3 --- /dev/null +++ b/embiggen/embedders/ensmallen_embedders/laplacian_eigenmaps.py @@ -0,0 +1,93 @@ +"""Module providing Laplacian Eigenmaps implementation.""" +from ensmallen import Graph +import pandas as pd +import numpy as np +from scipy.sparse import coo_matrix +from scipy.sparse.linalg import eigsh +from embiggen.embedders.ensmallen_embedders.ensmallen_embedder import EnsmallenEmbedder +from embiggen.utils import EmbeddingResult + + +class LaplacianEigenmapsEnsmallen(EnsmallenEmbedder): + """Class implementing the Laplacian Eigenmaps algorithm.""" + + def __init__( + self, + embedding_size: int = 100, + enable_cache: bool = False + ): + """Create new Laplacian Eigenmaps method. + + Parameters + -------------------------- + embedding_size: int = 100 + Dimension of the embedding. + enable_cache: bool = False + Whether to enable the cache, that is to + store the computed embedding. + """ + super().__init__( + embedding_size=embedding_size, + enable_cache=enable_cache, + ) + + def _fit_transform( + self, + graph: Graph, + return_dataframe: bool = True, + ) -> EmbeddingResult: + """Return node embedding.""" + edges, weights = graph.get_symmetric_normalized_laplacian_coo_matrix() + + coo = coo_matrix( + (weights, (edges[:, 0], edges[:, 1])), + shape=( + graph.get_number_of_nodes(), + graph.get_number_of_nodes() + ), + dtype=np.float32 + ) + + embedding = eigsh( + coo, + k=self._embedding_size, + which="SM", + maxiter=graph.get_number_of_nodes()100, + return_eigenvectors=True + )[1] + + if return_dataframe: + node_names = graph.get_node_names() + embedding = pd.DataFrame( + embedding, + index=node_names + ) + return EmbeddingResult( + embedding_method_name=self.model_name(), + node_embeddings=embedding + ) + + @classmethod + def model_name(cls) -> str: + """Returns name of the model.""" + return "Laplacian Eigenmaps" + + @classmethod + def can_use_edge_weights(cls) -> bool: + """Returns whether the model can optionally use edge weights.""" + return False + + @classmethod + def can_use_node_types(cls) -> bool: + """Returns whether the model can optionally use node types.""" + return False + + @classmethod + def can_use_edge_types(cls) -> bool: + """Returns whether the model can optionally use edge types.""" + return False + + @classmethod + def is_stocastic(cls) -> bool: + """Returns whether the model is stocastic and has therefore a random state.""" + return False diff --git a/embiggen/embedders/ensmallen_embedders/netmf.py b/embiggen/embedders/ensmallen_embedders/netmf.py index c33665ff..af658a33 100644 --- a/embiggen/embedders/ensmallen_embedders/netmf.py +++ b/embiggen/embedders/ensmallen_embedders/netmf.py @@ -5,10 +5,11 @@ import numpy as np from scipy.sparse import coo_matrix from sklearn.decomposition import TruncatedSVD -from embiggen.utils.abstract_models import AbstractEmbeddingModel, EmbeddingResult +from embiggen.embedders.ensmallen_embedders.ensmallen_embedder import EnsmallenEmbedder +from embiggen.utils import EmbeddingResult -class NetMFEnsmallen(AbstractEmbeddingModel): +class NetMFEnsmallen(EnsmallenEmbedder): """Class implementing the NetMF algorithm.""" def __init__( @@ -17,10 +18,6 @@ def __init__( walk_length: int = 128, iterations: int = 10, window_size: int = 10, - return_weight: float = 1.0, - explore_weight: float = 1.0, - change_node_type_weight: float = 1.0, - change_edge_type_weight: float = 1.0, max_neighbours: Optional[int] = 100, random_state: int = 42, enable_cache: bool = False @@ -38,28 +35,6 @@ def __init__( window_size: int = 10 Window size for the local context. On the borders the window size is trimmed. - return_weight: float = 1.0 - Weight on the probability of returning to the same node the walk just came from - Having this higher tends the walks to be - more like a Breadth-First Search. - Having this very high (> 2) makes search very local. - Equal to the inverse of p in the Node2Vec paper. - explore_weight: float = 1.0 - Weight on the probability of visiting a neighbor node - to the one we're coming from in the random walk - Having this higher tends the walks to be - more like a Depth-First Search. - Having this very high makes search more outward. - Having this very low makes search very local. - Equal to the inverse of q in the Node2Vec paper. - change_node_type_weight: float = 1.0 - Weight on the probability of visiting a neighbor node of a - different type than the previous node. This only applies to - colored graphs, otherwise it has no impact. - change_edge_type_weight: float = 1.0 - Weight on the probability of visiting a neighbor edge of a - different type than the previous edge. This only applies to - multigraphs, otherwise it has no impact. max_neighbours: Optional[int] = 100 Number of maximum neighbours to consider when using approximated walks. By default, None, we execute exact random walks. @@ -74,10 +49,6 @@ def __init__( walk_length=walk_length, iterations=iterations, window_size=window_size, - return_weight=return_weight, - explore_weight=explore_weight, - change_node_type_weight=change_node_type_weight, - change_edge_type_weight=change_edge_type_weight, max_neighbours=max_neighbours, ) @@ -91,7 +62,7 @@ def __init__( def smoke_test_parameters(cls) -> Dict[str, Any]: """Returns parameters for smoke test.""" return dict( - AbstractEmbeddingModel.smoke_test_parameters(), + EnsmallenEmbedder.smoke_test_parameters(), window_size=1, walk_length=4, iterations=1, @@ -109,7 +80,6 @@ def _fit_transform( self, graph: Graph, return_dataframe: bool = True, - verbose: bool = True ) -> EmbeddingResult: """Return node embedding.""" edges, weights = graph.get_log_normalized_cooccurrence_coo_matrix( @@ -143,27 +113,11 @@ def _fit_transform( node_embeddings=embedding ) - @classmethod - def task_name(cls) -> str: - return "Node Embedding" - @classmethod def model_name(cls) -> str: """Returns name of the model.""" return "NetMF" - @classmethod - def library_name(cls) -> str: - return "Ensmallen" - - @classmethod - def requires_nodes_sorted_by_decreasing_node_degree(cls) -> bool: - return False - - @classmethod - def is_topological(cls) -> bool: - return True - @classmethod def can_use_edge_weights(cls) -> bool: """Returns whether the model can optionally use edge weights.""" diff --git a/embiggen/embedders/ensmallen_embedders/node2vec.py b/embiggen/embedders/ensmallen_embedders/node2vec.py index 36ce0781..31b64caa 100644 --- a/embiggen/embedders/ensmallen_embedders/node2vec.py +++ b/embiggen/embedders/ensmallen_embedders/node2vec.py @@ -1,26 +1,34 @@ """Module providing abstract Node2Vec implementation.""" -from typing import Optional, Union, Dict, Any +from typing import Dict, Any from ensmallen import Graph -import numpy as np import pandas as pd +from userinput.utils import must_be_in_set from ensmallen import models -from embiggen.utils.abstract_models import abstract_class, AbstractEmbeddingModel, EmbeddingResult +from embiggen.utils.abstract_models import abstract_class +from embiggen.embedders.ensmallen_embedders.ensmallen_embedder import EnsmallenEmbedder +from embiggen.utils import EmbeddingResult @abstract_class -class Node2VecEnsmallen(AbstractEmbeddingModel): +class Node2VecEnsmallen(EnsmallenEmbedder): """Abstract class for Node2Vec algorithms.""" MODELS = { - "cbow": models.CBOW, - "skipgram": models.SkipGram, - "walkletscbow": models.WalkletsCBOW, - "walkletsskipgram": models.WalkletsSkipGram, + "DeepWalk CBOW": models.CBOW, + "DeepWalk SkipGram": models.SkipGram, + "DeepWalk GloVe": models.GloVe, + "Node2Vec CBOW": models.CBOW, + "Node2Vec SkipGram": models.SkipGram, + "Node2Vec GloVe": models.GloVe, + "Walklets CBOW": models.WalkletsCBOW, + "Walklets SkipGram": models.WalkletsSkipGram, + "Walklets GloVe": models.WalkletsGloVe, } def __init__( self, - model_name: str, + embedding_size: int = 100, + random_state: int = 42, enable_cache: bool = False, model_kwargs: Dict ): @@ -28,77 +36,56 @@ def __init__( Parameters -------------------------- - model_name: str - The ensmallen graph embedding model to use. - Can either be the SkipGram, WalkletsSkipGram, CBOW and WalkletsCBOW models. + embedding_size: int = 100 + Dimension of the embedding. + random_state: int = 42 + The random state to reproduce the training sequence. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. model_kwargs: Dict Further parameters to forward to the model. """ - if model_name.lower() not in self.MODELS: - raise ValueError( - ( - "The provided model name {} is not supported. " - "The supported models are `CBOW` and `SkipGram`." - ).format(model_name) - ) - + model_name = must_be_in_set(self.model_name(), self.MODELS.keys(), "model name") self._model_kwargs = model_kwargs - - self._model = Node2VecEnsmallen.MODELS[model_name.lower()](model_kwargs) + self._model = Node2VecEnsmallen.MODELS[model_name](model_kwargs) super().__init__( - embedding_size=model_kwargs["embedding_size"], + embedding_size=embedding_size, enable_cache=enable_cache, - random_state=model_kwargs["random_state"] + random_state=random_state ) @classmethod def smoke_test_parameters(cls) -> Dict[str, Any]: """Returns parameters for smoke test.""" return dict( - AbstractEmbeddingModel.smoke_test_parameters(), + EnsmallenEmbedder.smoke_test_parameters(), epochs=1, window_size=1, walk_length=4, iterations=1, max_neighbours=10, - number_of_negative_samples=1 ) def parameters(self) -> Dict[str, Any]: """Returns parameters of the model.""" - return { + return dict( super().parameters(), self._model_kwargs - } - - @classmethod - def task_name(cls) -> str: - return "Node Embedding" - - @classmethod - def library_name(cls) -> str: - return "Ensmallen" - - def requires_nodes_sorted_by_decreasing_node_degree(self) -> bool: - return False - - def is_topological(self) -> bool: - return True + ) def _fit_transform( self, graph: Graph, return_dataframe: bool = True, - verbose: bool = True ) -> EmbeddingResult: """Return node embedding.""" node_embeddings = self._model.fit_transform(graph) - if not isinstance(node_embeddings, list): - node_embeddings = [node_embeddings] + + if "CBOW" in self.model_name(): + node_embeddings = list(reversed(node_embeddings)) + if return_dataframe: node_names = graph.get_node_names() node_embeddings = [ @@ -137,7 +124,7 @@ def can_use_node_types(cls) -> bool: def is_using_node_types(self) -> bool: """Returns whether the model is parametrized to use node types.""" - return self._model_kwargs["change_node_type_weight"] != 1.0 + return "change_node_type_weight" in self._model_kwargs and self._model_kwargs["change_node_type_weight"] != 1.0 @classmethod def can_use_edge_types(cls) -> bool: @@ -146,9 +133,17 @@ def can_use_edge_types(cls) -> bool: def is_using_edge_types(self) -> bool: """Returns whether the model is parametrized to use edge types.""" - return self._model_kwargs["change_edge_type_weight"] != 1.0 + return "change_edge_type_weight" in self._model_kwargs and self._model_kwargs["change_edge_type_weight"] != 1.0 @classmethod def is_stocastic(cls) -> bool: """Returns whether the model is stocastic and has therefore a random state.""" - return True \ No newline at end of file + return True + + @classmethod + def requires_node_types(cls) -> bool: + return False + + @classmethod + def requires_edge_types(cls) -> bool: + return False \ No newline at end of file diff --git a/embiggen/embedders/ensmallen_embedders/cbow.py b/embiggen/embedders/ensmallen_embedders/node2vec_cbow.py similarity index 70% rename from embiggen/embedders/ensmallen_embedders/cbow.py rename to embiggen/embedders/ensmallen_embedders/node2vec_cbow.py index 8e0e58a8..4fc3203b 100644 --- a/embiggen/embedders/ensmallen_embedders/cbow.py +++ b/embiggen/embedders/ensmallen_embedders/node2vec_cbow.py @@ -1,31 +1,31 @@ -"""Module providing CBOW model implementation.""" -from typing import Optional +"""Module providing Node2Vec CBOW model implementation.""" +from typing import Optional, Dict, Any from embiggen.embedders.ensmallen_embedders.node2vec import Node2VecEnsmallen -class CBOWEnsmallen(Node2VecEnsmallen): - """Class providing CBOW implemeted in Rust from Ensmallen.""" +class Node2VecCBOWEnsmallen(Node2VecEnsmallen): + """Class providing Node2Vec CBOW implemeted in Rust from Ensmallen.""" def __init__( self, embedding_size: int = 100, - epochs: int = 10, + epochs: int = 30, clipping_value: float = 6.0, number_of_negative_samples: int = 10, walk_length: int = 128, iterations: int = 10, - window_size: int = 10, - return_weight: float = 1.0, - explore_weight: float = 1.0, - change_node_type_weight: float = 1.0, - change_edge_type_weight: float = 1.0, + window_size: int = 5, + return_weight: float = 0.25, + explore_weight: float = 4.0, max_neighbours: Optional[int] = 100, learning_rate: float = 0.01, learning_rate_decay: float = 0.9, + central_nodes_embedding_path: Optional[str] = None, + contextual_nodes_embedding_path: Optional[str] = None, normalize_by_degree: bool = False, stochastic_downsample_by_degree: Optional[bool] = False, normalize_learning_rate_by_degree: Optional[bool] = False, - use_zipfian_sampling: Optional[bool] = True, + use_scale_free_distribution: Optional[bool] = True, random_state: int = 42, enable_cache: bool = False ): @@ -35,7 +35,7 @@ def __init__( -------------------------- embedding_size: int = 100 Dimension of the embedding. - epochs: int = 10 + epochs: int = 30 Number of epochs to train the model for. clipping_value: float = 6.0 Value at which we clip the dot product, mostly for numerical stability issues. @@ -47,16 +47,16 @@ def __init__( Maximal length of the walks. iterations: int = 10 Number of iterations of the single walks. - window_size: int = 10 + window_size: int = 5 Window size for the local context. On the borders the window size is trimmed. - return_weight: float = 1.0 + return_weight: float = 0.25 Weight on the probability of returning to the same node the walk just came from Having this higher tends the walks to be more like a Breadth-First Search. Having this very high (> 2) makes search very local. Equal to the inverse of p in the Node2Vec paper. - explore_weight: float = 1.0 + explore_weight: float = 4.0 Weight on the probability of visiting a neighbor node to the one we're coming from in the random walk Having this higher tends the walks to be @@ -64,14 +64,6 @@ def __init__( Having this very high makes search more outward. Having this very low makes search very local. Equal to the inverse of q in the Node2Vec paper. - change_node_type_weight: float = 1.0 - Weight on the probability of visiting a neighbor node of a - different type than the previous node. This only applies to - colored graphs, otherwise it has no impact. - change_edge_type_weight: float = 1.0 - Weight on the probability of visiting a neighbor edge of a - different type than the previous edge. This only applies to - multigraphs, otherwise it has no impact. max_neighbours: Optional[int] = 100 Number of maximum neighbours to consider when using approximated walks. By default, None, we execute exact random walks. @@ -80,6 +72,14 @@ def __init__( The learning rate to use to train the Node2Vec model. By default 0.01. learning_rate_decay: float = 0.9 Factor to reduce the learning rate for at each epoch. By default 0.9. + central_nodes_embedding_path: Optional[str] = None + Path where to mmap and store the central nodes embedding. + This is necessary to embed large graphs whose embedding will not + fit into the available main memory. + contextual_nodes_embedding_path: Optional[str] = None + Path where to mmap and store the central nodes embedding. + This is necessary to embed large graphs whose embedding will not + fit into the available main memory. normalize_by_degree: bool = False Whether to normalize the random walk by the node degree of the destination node degrees. @@ -87,7 +87,7 @@ def __init__( Randomly skip samples with probability proportional to the degree of the central node. By default false. normalize_learning_rate_by_degree: Optional[bool] = False Divide the learning rate by the degree of the central node. By default false. - use_zipfian_sampling: Optional[bool] = True + use_scale_free_distribution: Optional[bool] = True Sample negatives proportionally to their degree. By default true. random_state: int = 42 The random state to reproduce the training sequence. @@ -96,7 +96,6 @@ def __init__( store the computed embedding. """ super().__init__( - model_name="CBOW", embedding_size=embedding_size, epochs=epochs, clipping_value=clipping_value, @@ -106,28 +105,35 @@ def __init__( window_size=window_size, return_weight=return_weight, explore_weight=explore_weight, - change_edge_type_weight=change_edge_type_weight, - change_node_type_weight=change_node_type_weight, max_neighbours=max_neighbours, learning_rate=learning_rate, learning_rate_decay=learning_rate_decay, + central_nodes_embedding_path=central_nodes_embedding_path, + contextual_nodes_embedding_path=contextual_nodes_embedding_path, normalize_by_degree=normalize_by_degree, stochastic_downsample_by_degree=stochastic_downsample_by_degree, normalize_learning_rate_by_degree=normalize_learning_rate_by_degree, - use_zipfian_sampling=use_zipfian_sampling, + use_scale_free_distribution=use_scale_free_distribution, random_state=random_state, enable_cache=enable_cache ) + def parameters(self) -> Dict[str, Any]: + """Returns parameters for smoke test.""" + removed = [ + "change_node_type_weight", + "change_edge_type_weight", + "alpha" + ] + return dict( + { + key: value + for key, value in super().parameters().items() + if key not in removed + } + ) + @classmethod def model_name(cls) -> str: """Returns name of the model.""" - return "CBOW" - - @classmethod - def requires_node_types(cls) -> bool: - return False - - @classmethod - def requires_edge_types(cls) -> bool: - return False \ No newline at end of file + return "Node2Vec CBOW" \ No newline at end of file diff --git a/embiggen/embedders/ensmallen_embedders/node2vec_glove.py b/embiggen/embedders/ensmallen_embedders/node2vec_glove.py new file mode 100644 index 00000000..806cb8c6 --- /dev/null +++ b/embiggen/embedders/ensmallen_embedders/node2vec_glove.py @@ -0,0 +1,128 @@ +"""Module providing Node2Vec GloVe model implementation.""" +from typing import Optional, Dict, Any +from embiggen.embedders.ensmallen_embedders.node2vec import Node2VecEnsmallen + +class Node2VecGloVeEnsmallen(Node2VecEnsmallen): + """Class providing Node2Vec GloVe implemeted in Rust from Ensmallen.""" + + def __init__( + self, + embedding_size: int = 100, + alpha: float = 0.75, + epochs: int = 30, + clipping_value: float = 6.0, + walk_length: int = 128, + iterations: int = 10, + window_size: int = 5, + return_weight: float = 0.25, + explore_weight: float = 4.0, + max_neighbours: Optional[int] = 100, + learning_rate: float = 0.001, + learning_rate_decay: float = 0.9, + central_nodes_embedding_path: Optional[str] = None, + contextual_nodes_embedding_path: Optional[str] = None, + normalize_by_degree: bool = False, + random_state: int = 42, + enable_cache: bool = False + ): + """Create new Node2Vec GloVe model. + + Parameters + -------------------------- + embedding_size: int = 100 + Dimension of the embedding. + alpha: float = 0.75 + Alpha parameter for GloVe's loss. + epochs: int = 100 + Number of epochs to train the model for. + window_size: int = 10 + Window size for the local context. + On the borders the window size is trimmed. + clipping_value: float = 6.0 + Value at which we clip the dot product, mostly for numerical stability issues. + By default, `6.0`, where the loss is already close to zero. + walk_length: int = 128 + Maximal length of the walks. + iterations: int = 10 + Number of iterations of the single walks. + window_size: int = 5 + Window size for the local context. + On the borders the window size is trimmed. + return_weight: float = 0.25 + Weight on the probability of returning to the same node the walk just came from + Having this higher tends the walks to be + more like a Breadth-First Search. + Having this very high (> 2) makes search very local. + Equal to the inverse of p in the Node2Vec paper. + explore_weight: float = 4.0 + Weight on the probability of visiting a neighbor node + to the one we're coming from in the random walk + Having this higher tends the walks to be + more like a Depth-First Search. + Having this very high makes search more outward. + Having this very low makes search very local. + Equal to the inverse of q in the Node2Vec paper. + max_neighbours: Optional[int] = 100 + Number of maximum neighbours to consider when using approximated walks. + By default, None, we execute exact random walks. + This is mainly useful for graphs containing nodes with high degrees. + learning_rate: float = 0.001 + The learning rate to use to train the Node2Vec model. By default 0.01. + learning_rate_decay: float = 0.9 + Factor to reduce the learning rate for at each epoch. By default 0.9. + central_nodes_embedding_path: Optional[str] = None + Path where to mmap and store the central nodes embedding. + This is necessary to embed large graphs whose embedding will not + fit into the available main memory. + contextual_nodes_embedding_path: Optional[str] = None + Path where to mmap and store the central nodes embedding. + This is necessary to embed large graphs whose embedding will not + fit into the available main memory. + normalize_by_degree: bool = False + Whether to normalize the random walk by the node degree + of the destination node degrees. + random_state: int = 42 + The random state to reproduce the training sequence. + enable_cache: bool = False + Whether to enable the cache, that is to + store the computed embedding. + """ + super().__init__( + embedding_size=embedding_size, + alpha=alpha, + epochs=epochs, + clipping_value=clipping_value, + walk_length=walk_length, + iterations=iterations, + window_size=window_size, + return_weight=return_weight, + explore_weight=explore_weight, + max_neighbours=max_neighbours, + learning_rate=learning_rate, + learning_rate_decay=learning_rate_decay, + central_nodes_embedding_path=central_nodes_embedding_path, + contextual_nodes_embedding_path=contextual_nodes_embedding_path, + normalize_by_degree=normalize_by_degree, + enable_cache=enable_cache, + random_state=random_state + ) + + def parameters(self) -> Dict[str, Any]: + """Returns parameters for smoke test.""" + removed = [ + "change_node_type_weight", + "change_edge_type_weight", + "number_of_negative_samples" + ] + return dict( + { + key: value + for key, value in super().parameters().items() + if key not in removed + } + ) + + @classmethod + def model_name(cls) -> str: + """Returns name of the model.""" + return "Node2Vec GloVe" \ No newline at end of file diff --git a/embiggen/embedders/ensmallen_embedders/skipgram.py b/embiggen/embedders/ensmallen_embedders/node2vec_skipgram.py similarity index 69% rename from embiggen/embedders/ensmallen_embedders/skipgram.py rename to embiggen/embedders/ensmallen_embedders/node2vec_skipgram.py index 89ef432d..b3a1f50e 100644 --- a/embiggen/embedders/ensmallen_embedders/skipgram.py +++ b/embiggen/embedders/ensmallen_embedders/node2vec_skipgram.py @@ -1,31 +1,31 @@ -"""Module providing SkipGram model implementation.""" -from typing import Optional +"""Module providing Node2Vec SkipGram model implementation.""" +from typing import Optional, Dict, Any from embiggen.embedders.ensmallen_embedders.node2vec import Node2VecEnsmallen -class SkipGramEnsmallen(Node2VecEnsmallen): - """Class providing SkipGram implemeted in Rust from Ensmallen.""" +class Node2VecSkipGramEnsmallen(Node2VecEnsmallen): + """Class providing Node2Vec SkipGram implemeted in Rust from Ensmallen.""" def __init__( self, embedding_size: int = 100, - epochs: int = 10, + epochs: int = 30, clipping_value: float = 6.0, number_of_negative_samples: int = 10, walk_length: int = 128, iterations: int = 10, - window_size: int = 10, - return_weight: float = 1.0, - explore_weight: float = 1.0, - change_node_type_weight: float = 1.0, - change_edge_type_weight: float = 1.0, + window_size: int = 5, + return_weight: float = 0.25, + explore_weight: float = 4.0, max_neighbours: Optional[int] = 100, learning_rate: float = 0.01, learning_rate_decay: float = 0.9, + central_nodes_embedding_path: Optional[str] = None, + contextual_nodes_embedding_path: Optional[str] = None, normalize_by_degree: bool = False, stochastic_downsample_by_degree: Optional[bool] = False, normalize_learning_rate_by_degree: Optional[bool] = False, - use_zipfian_sampling: Optional[bool] = True, + use_scale_free_distribution: Optional[bool] = True, random_state: int = 42, enable_cache: bool = False ): @@ -35,11 +35,8 @@ def __init__( -------------------------- embedding_size: int = 100 Dimension of the embedding. - epochs: int = 10 + epochs: int = 30 Number of epochs to train the model for. - window_size: int = 10 - Window size for the local context. - On the borders the window size is trimmed. clipping_value: float = 6.0 Value at which we clip the dot product, mostly for numerical stability issues. By default, `6.0`, where the loss is already close to zero. @@ -50,16 +47,16 @@ def __init__( Maximal length of the walks. iterations: int = 10 Number of iterations of the single walks. - window_size: int = 10 + window_size: int = 5 Window size for the local context. On the borders the window size is trimmed. - return_weight: float = 1.0 + return_weight: float = 0.25 Weight on the probability of returning to the same node the walk just came from Having this higher tends the walks to be more like a Breadth-First Search. Having this very high (> 2) makes search very local. Equal to the inverse of p in the Node2Vec paper. - explore_weight: float = 1.0 + explore_weight: float = 4.0 Weight on the probability of visiting a neighbor node to the one we're coming from in the random walk Having this higher tends the walks to be @@ -67,14 +64,6 @@ def __init__( Having this very high makes search more outward. Having this very low makes search very local. Equal to the inverse of q in the Node2Vec paper. - change_node_type_weight: float = 1.0 - Weight on the probability of visiting a neighbor node of a - different type than the previous node. This only applies to - colored graphs, otherwise it has no impact. - change_edge_type_weight: float = 1.0 - Weight on the probability of visiting a neighbor edge of a - different type than the previous edge. This only applies to - multigraphs, otherwise it has no impact. max_neighbours: Optional[int] = 100 Number of maximum neighbours to consider when using approximated walks. By default, None, we execute exact random walks. @@ -83,6 +72,16 @@ def __init__( The learning rate to use to train the Node2Vec model. By default 0.01. learning_rate_decay: float = 0.9 Factor to reduce the learning rate for at each epoch. By default 0.9. + central_nodes_embedding_path: Optional[str] = None + Path where to mmap and store the central nodes embedding. + + This is necessary to embed large graphs whose embedding will not + fit into the available main memory. + contextual_nodes_embedding_path: Optional[str] = None + Path where to mmap and store the central nodes embedding. + + This is necessary to embed large graphs whose embedding will not + fit into the available main memory. normalize_by_degree: bool = False Whether to normalize the random walk by the node degree of the destination node degrees. @@ -90,7 +89,7 @@ def __init__( Randomly skip samples with probability proportional to the degree of the central node. By default false. normalize_learning_rate_by_degree: Optional[bool] = False Divide the learning rate by the degree of the central node. By default false. - use_zipfian_sampling: Optional[bool] = True + use_scale_free_distribution: Optional[bool] = True Sample negatives proportionally to their degree. By default true. random_state: int = 42 The random state to reproduce the training sequence. @@ -99,7 +98,6 @@ def __init__( store the computed embedding. """ super().__init__( - model_name="SkipGram", embedding_size=embedding_size, epochs=epochs, clipping_value=clipping_value, @@ -109,28 +107,35 @@ def __init__( window_size=window_size, return_weight=return_weight, explore_weight=explore_weight, - change_edge_type_weight=change_edge_type_weight, - change_node_type_weight=change_node_type_weight, max_neighbours=max_neighbours, learning_rate=learning_rate, learning_rate_decay=learning_rate_decay, + central_nodes_embedding_path=central_nodes_embedding_path, + contextual_nodes_embedding_path=contextual_nodes_embedding_path, normalize_by_degree=normalize_by_degree, stochastic_downsample_by_degree=stochastic_downsample_by_degree, normalize_learning_rate_by_degree=normalize_learning_rate_by_degree, - use_zipfian_sampling=use_zipfian_sampling, + use_scale_free_distribution=use_scale_free_distribution, random_state=random_state, - enable_cache=enable_cache, + enable_cache=enable_cache ) + def parameters(self) -> Dict[str, Any]: + """Returns parameters for smoke test.""" + removed = [ + "change_node_type_weight", + "change_edge_type_weight", + "alpha" + ] + return dict( + { + key: value + for key, value in super().parameters().items() + if key not in removed + } + ) + @classmethod def model_name(cls) -> str: """Returns name of the model.""" - return "SkipGram" - - @classmethod - def requires_node_types(cls) -> bool: - return False - - @classmethod - def requires_edge_types(cls) -> bool: - return False \ No newline at end of file + return "Node2Vec SkipGram" \ No newline at end of file diff --git a/embiggen/embedders/ensmallen_embedders/node_label_spine.py b/embiggen/embedders/ensmallen_embedders/node_label_spine.py new file mode 100644 index 00000000..6bcc7230 --- /dev/null +++ b/embiggen/embedders/ensmallen_embedders/node_label_spine.py @@ -0,0 +1,119 @@ +"""Module providing Node-label-based SPINE implementation.""" +from typing import Optional, Dict, Any +from ensmallen import Graph +import pandas as pd +from ensmallen import models +from embiggen.embedders.ensmallen_embedders.ensmallen_embedder import EnsmallenEmbedder +from embiggen.utils import EmbeddingResult + + +class NodeLabelSPINE(EnsmallenEmbedder): + """Class implementing the Node-label-based SPINE algorithm.""" + + def __init__( + self, + dtype: Optional[str] = "u8", + maximum_depth: Optional[int] = None, + path: Optional[str] = None, + verbose: bool = False, + enable_cache: bool = False + ): + """Create new Node-label-based SPINE method. + + Parameters + -------------------------- + dtype: Optional[str] = "u8" + Dtype to use for the embedding. + maximum_depth: Optional[int] = None + Maximum depth of the shortest path. + path: Optional[str] = None + Path where to store the mmap-ed embedding. + This parameter is necessary to embed very large graphs. + verbose: bool = False + Whether to show loading bars. + enable_cache: bool = False + Whether to enable the cache, that is to + store the computed embedding. + """ + self._dtype = dtype + self._verbose = verbose + self._maximum_depth = maximum_depth + self._path = path + self._model = models.NodeLabelSPINE( + verbose=self._verbose, + maximum_depth=self._maximum_depth, + path=self._path + ) + + super().__init__( + enable_cache=enable_cache + ) + + def parameters(self) -> Dict[str, Any]: + """Returns parameters of the model.""" + return dict( + { + key: value + for key, value in super().parameters().items() + if key != "embedding_size" + }, + dict( + dtype=self._dtype, + maximum_depth=self._maximum_depth, + path=self._path, + ) + ) + @classmethod + def smoke_test_parameters(cls) -> Dict[str, Any]: + """Returns parameters for smoke test.""" + return dict() + + def _fit_transform( + self, + graph: Graph, + return_dataframe: bool = True, + ) -> EmbeddingResult: + """Return node embedding.""" + node_embedding = self._model.fit_transform( + graph, + dtype=self._dtype, + ) + if return_dataframe: + node_embedding = pd.DataFrame( + node_embedding, + index=graph.get_node_names() + ) + return EmbeddingResult( + embedding_method_name=self.model_name(), + node_embeddings=node_embedding + ) + + @classmethod + def model_name(cls) -> str: + """Returns name of the model.""" + return "Node-label-based SPINE" + + @classmethod + def requires_node_types(cls) -> bool: + """Returns whether the model requires node types.""" + return True + + @classmethod + def get_minimum_required_number_of_node_types(cls) -> int: + """Requires minimum number of required node types.""" + return 2 + + @classmethod + def can_use_edge_weights(cls) -> bool: + """Returns whether the model can optionally use edge weights.""" + return False + + @classmethod + def can_use_edge_types(cls) -> bool: + """Returns whether the model can optionally use edge types.""" + return False + + @classmethod + def is_stocastic(cls) -> bool: + """Returns whether the model is stocastic and has therefore a random state.""" + return False \ No newline at end of file diff --git a/embiggen/embedders/ensmallen_embedders/node_label_wine.py b/embiggen/embedders/ensmallen_embedders/node_label_wine.py new file mode 100644 index 00000000..109ac40c --- /dev/null +++ b/embiggen/embedders/ensmallen_embedders/node_label_wine.py @@ -0,0 +1,120 @@ +"""Module providing Node-label-based WINE implementation.""" +from typing import Optional, Dict, Any +from ensmallen import Graph +import pandas as pd +from ensmallen import models +from embiggen.embedders.ensmallen_embedders.ensmallen_embedder import EnsmallenEmbedder +from embiggen.utils import EmbeddingResult + + +class NodeLabelWINE(EnsmallenEmbedder): + """Class implementing the Node-label-based WINE algorithm.""" + + def __init__( + self, + dtype: Optional[str] = "u8", + walk_length: Optional[int] = None, + path: Optional[str] = None, + verbose: bool = False, + enable_cache: bool = False + ): + """Create new Node-label-based WINE method. + + Parameters + -------------------------- + dtype: Optional[str] = "u8" + Dtype to use for the embedding. + walk_length: int = 2 + Length of the random walk. + By default 2, to capture exclusively the immediate context. + path: Optional[str] = None + Path where to store the mmap-ed embedding. + This parameter is necessary to embed very large graphs. + verbose: bool = False + Whether to show loading bars. + enable_cache: bool = False + Whether to enable the cache, that is to + store the computed embedding. + """ + self._dtype = dtype + self._verbose = verbose + self._walk_length = walk_length + self._path = path + self._model = models.NodeLabelWINE( + verbose=self._verbose, + walk_length=self._walk_length, + path=self._path + ) + + super().__init__( + enable_cache=enable_cache + ) + + def parameters(self) -> Dict[str, Any]: + """Returns parameters of the model.""" + return dict( + { + key: value + for key, value in super().parameters().items() + if key != "embedding_size" + }, + dict( + dtype=self._dtype, + walk_length=self._walk_length, + path=self._path, + ) + ) + @classmethod + def smoke_test_parameters(cls) -> Dict[str, Any]: + """Returns parameters for smoke test.""" + return dict() + + def _fit_transform( + self, + graph: Graph, + return_dataframe: bool = True, + ) -> EmbeddingResult: + """Return node embedding.""" + node_embedding = self._model.fit_transform( + graph, + dtype=self._dtype, + ) + if return_dataframe: + node_embedding = pd.DataFrame( + node_embedding, + index=graph.get_node_names() + ) + return EmbeddingResult( + embedding_method_name=self.model_name(), + node_embeddings=node_embedding + ) + + @classmethod + def model_name(cls) -> str: + """Returns name of the model.""" + return "Node-label-based WINE" + + @classmethod + def requires_node_types(cls) -> bool: + """Returns whether the model requires node types.""" + return True + + @classmethod + def get_minimum_required_number_of_node_types(cls) -> int: + """Requires minimum number of required node types.""" + return 2 + + @classmethod + def can_use_edge_weights(cls) -> bool: + """Returns whether the model can optionally use edge weights.""" + return False + + @classmethod + def can_use_edge_types(cls) -> bool: + """Returns whether the model can optionally use edge types.""" + return False + + @classmethod + def is_stocastic(cls) -> bool: + """Returns whether the model is stocastic and has therefore a random state.""" + return False \ No newline at end of file diff --git a/embiggen/embedders/ensmallen_embedders/score_spine.py b/embiggen/embedders/ensmallen_embedders/score_spine.py new file mode 100644 index 00000000..cf018f30 --- /dev/null +++ b/embiggen/embedders/ensmallen_embedders/score_spine.py @@ -0,0 +1,125 @@ +"""Module providing Score-based SPINE implementation.""" +from typing import Optional, Dict, Any +from ensmallen import Graph +import pandas as pd +import numpy as np +from ensmallen import models +from embiggen.embedders.ensmallen_embedders.ensmallen_embedder import EnsmallenEmbedder +from embiggen.utils import EmbeddingResult + + +class ScoreSPINE(EnsmallenEmbedder): + """Class implementing the Score-based SPINE algorithm.""" + + def __init__( + self, + scores: Optional[np.ndarray] = None, + embedding_size: int = 100, + dtype: Optional[str] = "u8", + maximum_depth: Optional[int] = None, + path: Optional[str] = None, + verbose: bool = False, + enable_cache: bool = False + ): + """Create new Score-based SPINE method. + + Parameters + -------------------------- + scores: Optional[np.ndarray] = None + Numpy array to be used to sort the anchor nodes. + embedding_size: int = 100 + Dimension of the embedding. + dtype: Optional[str] = "u8" + Dtype to use for the embedding. + maximum_depth: Optional[int] = None + Maximum depth of the shortest path. + path: Optional[str] = None + Path where to store the mmap-ed embedding. + This parameter is necessary to embed very large graphs. + verbose: bool = False + Whether to show loading bars. + enable_cache: bool = False + Whether to enable the cache, that is to + store the computed embedding. + """ + self._dtype = dtype + self._verbose = verbose + self._maximum_depth = maximum_depth + self._path = path + self._scores = scores + self._model = models.ScoreSPINE( + embedding_size=embedding_size, + verbose=self._verbose, + maximum_depth=self._maximum_depth, + path=self._path + ) + + super().__init__( + embedding_size=embedding_size, + enable_cache=enable_cache + ) + + def parameters(self) -> Dict[str, Any]: + """Returns parameters of the model.""" + return { + super().parameters(), + dict( + dtype=self._dtype, + maximum_depth=self._maximum_depth, + path=self._path, + ) + } + + @classmethod + def smoke_test_parameters(cls) -> Dict[str, Any]: + """Returns parameters for smoke test.""" + return dict( + embedding_size=5, + ) + + def _fit_transform( + self, + graph: Graph, + return_dataframe: bool = True, + ) -> EmbeddingResult: + """Return node embedding.""" + node_embedding = self._model.fit_transform( + scores=(np.ones(graph.get_number_of_nodes()) + if self._scores is None else self._scores).astype(np.float32), + graph=graph, + dtype=self._dtype, + ) + if return_dataframe: + node_embedding = pd.DataFrame( + node_embedding, + index=graph.get_node_names() + ) + return EmbeddingResult( + embedding_method_name=self.model_name(), + node_embeddings=node_embedding + ) + + @classmethod + def model_name(cls) -> str: + """Returns name of the model.""" + return "Score-based SPINE" + + @classmethod + def can_use_node_types(cls) -> bool: + """Returns whether the model can optionally use node types.""" + return False + + @classmethod + def can_use_edge_weights(cls) -> bool: + """Returns whether the model can optionally use edge weights.""" + return False + + @classmethod + def can_use_edge_types(cls) -> bool: + """Returns whether the model can optionally use edge types.""" + return False + + @classmethod + def is_stocastic(cls) -> bool: + """Returns whether the model is stocastic and has therefore a random state.""" + return False diff --git a/embiggen/embedders/ensmallen_embedders/score_wine.py b/embiggen/embedders/ensmallen_embedders/score_wine.py new file mode 100644 index 00000000..35f9b2a9 --- /dev/null +++ b/embiggen/embedders/ensmallen_embedders/score_wine.py @@ -0,0 +1,126 @@ +"""Module providing Score-based WINE implementation.""" +from typing import Optional, Dict, Any +from ensmallen import Graph +import pandas as pd +import numpy as np +from ensmallen import models +from embiggen.embedders.ensmallen_embedders.ensmallen_embedder import EnsmallenEmbedder +from embiggen.utils import EmbeddingResult + + +class ScoreWINE(EnsmallenEmbedder): + """Class implementing the Score-based WINE algorithm.""" + + def __init__( + self, + scores: Optional[np.ndarray] = None, + embedding_size: int = 100, + dtype: Optional[str] = "u8", + walk_length: Optional[int] = None, + path: Optional[str] = None, + verbose: bool = False, + enable_cache: bool = False + ): + """Create new Score-based WINE method. + + Parameters + -------------------------- + scores: Optional[np.ndarray] = None + Numpy array to be used to sort the anchor nodes. + embedding_size: int = 100 + Dimension of the embedding. + dtype: Optional[str] = "u8" + Dtype to use for the embedding. + walk_length: int = 2 + Length of the random walk. + By default 2, to capture exclusively the immediate context. + path: Optional[str] = None + Path where to store the mmap-ed embedding. + This parameter is necessary to embed very large graphs. + verbose: bool = False + Whether to show loading bars. + enable_cache: bool = False + Whether to enable the cache, that is to + store the computed embedding. + """ + self._dtype = dtype + self._verbose = verbose + self._walk_length = walk_length + self._path = path + self._scores = scores + self._model = models.ScoreWINE( + embedding_size=embedding_size, + verbose=self._verbose, + walk_length=self._walk_length, + path=self._path + ) + + super().__init__( + embedding_size=embedding_size, + enable_cache=enable_cache + ) + + def parameters(self) -> Dict[str, Any]: + """Returns parameters of the model.""" + return { + super().parameters(), + dict( + dtype=self._dtype, + walk_length=self._walk_length, + path=self._path, + ) + } + + @classmethod + def smoke_test_parameters(cls) -> Dict[str, Any]: + """Returns parameters for smoke test.""" + return dict( + embedding_size=5, + ) + + def _fit_transform( + self, + graph: Graph, + return_dataframe: bool = True, + ) -> EmbeddingResult: + """Return node embedding.""" + node_embedding = self._model.fit_transform( + scores=(np.ones(graph.get_number_of_nodes()) + if self._scores is None else self._scores).astype(np.float32), + graph=graph, + dtype=self._dtype, + ) + if return_dataframe: + node_embedding = pd.DataFrame( + node_embedding, + index=graph.get_node_names() + ) + return EmbeddingResult( + embedding_method_name=self.model_name(), + node_embeddings=node_embedding + ) + + @classmethod + def model_name(cls) -> str: + """Returns name of the model.""" + return "Score-based WINE" + + @classmethod + def can_use_node_types(cls) -> bool: + """Returns whether the model can optionally use node types.""" + return False + + @classmethod + def can_use_edge_weights(cls) -> bool: + """Returns whether the model can optionally use edge weights.""" + return False + + @classmethod + def can_use_edge_types(cls) -> bool: + """Returns whether the model can optionally use edge types.""" + return False + + @classmethod + def is_stocastic(cls) -> bool: + """Returns whether the model is stocastic and has therefore a random state.""" + return False diff --git a/embiggen/embedders/ensmallen_embedders/second_order_line.py b/embiggen/embedders/ensmallen_embedders/second_order_line.py new file mode 100644 index 00000000..19858731 --- /dev/null +++ b/embiggen/embedders/ensmallen_embedders/second_order_line.py @@ -0,0 +1,140 @@ +"""Module providing second-order LINE implementation.""" +from typing import Dict, Any, Optional +from ensmallen import Graph +import numpy as np +import pandas as pd +from ensmallen import models +from embiggen.embedders.ensmallen_embedders.ensmallen_embedder import EnsmallenEmbedder +from embiggen.utils import EmbeddingResult + + +class SecondOrderLINEEnsmallen(EnsmallenEmbedder): + """Class implementing the second-order LINE algorithm.""" + + def __init__( + self, + embedding_size: int = 100, + epochs: int = 100, + learning_rate: float = 0.01, + learning_rate_decay: float = 0.9, + avoid_false_negatives: bool = False, + node_embedding_path: Optional[str] = None, + contextual_node_embedding_path: Optional[str] = None, + random_state: int = 42, + verbose: bool = False, + enable_cache: bool = False + ): + """Create new abstract Node2Vec method. + + Parameters + -------------------------- + embedding_size: int = 100 + Dimension of the embedding. + epochs: int = 100 + The number of epochs to run the model for, by default 10. + learning_rate: float = 0.01 + The learning rate to update the gradient, by default 0.01. + learning_rate_decay: float = 0.9 + Factor to reduce the learning rate for at each epoch. By default 0.9. + avoid_false_negatives: bool = False + Whether to avoid sampling false negatives. + This may cause a slower training. + node_embedding_path: Optional[str] = None + Path where to mmap and store the nodes embedding. + This is necessary to embed large graphs whose embedding will not + fit into the available main memory. + contextual_node_embedding_path: Optional[str] = None + Path where to mmap and store the contextual nodes embedding. + This is necessary to embed large graphs whose embedding will not + fit into the available main memory. + random_state: int = 42 + Random state to reproduce the embeddings. + verbose: bool = False + Whether to show loading bars. + enable_cache: bool = False + Whether to enable the cache, that is to + store the computed embedding. + """ + self._kwargs = dict( + epochs=epochs, + learning_rate=learning_rate, + learning_rate_decay=learning_rate_decay, + avoid_false_negatives=avoid_false_negatives, + node_embedding_path=node_embedding_path, + contextual_node_embedding_path=contextual_node_embedding_path, + verbose=verbose + ) + + self._model = models.SecondOrderLINE( + embedding_size=embedding_size, + random_state=random_state, + self._kwargs + ) + + super().__init__( + embedding_size=embedding_size, + enable_cache=enable_cache, + random_state=random_state + ) + + def parameters(self) -> Dict[str, Any]: + """Returns parameters of the model.""" + return dict( + super().parameters(), + self._kwargs + ) + + @classmethod + def smoke_test_parameters(cls) -> Dict[str, Any]: + """Returns parameters for smoke test.""" + return dict( + embedding_size=5, + epochs=1 + ) + + def _fit_transform( + self, + graph: Graph, + return_dataframe: bool = True, + ) -> EmbeddingResult: + """Return node embedding.""" + node_embeddings = self._model.fit_transform(graph) + if return_dataframe: + node_names = graph.get_node_names() + node_embeddings = [ + pd.DataFrame( + node_embedding, + index=node_names + ) + for node_embedding in node_embeddings + ] + + return EmbeddingResult( + embedding_method_name=self.model_name(), + node_embeddings=node_embeddings, + ) + + @classmethod + def model_name(cls) -> str: + """Returns name of the model.""" + return "Second-order LINE" + + @classmethod + def can_use_edge_weights(cls) -> bool: + """Returns whether the model can optionally use edge weights.""" + return False + + @classmethod + def can_use_node_types(cls) -> bool: + """Returns whether the model can optionally use node types.""" + return False + + @classmethod + def can_use_edge_types(cls) -> bool: + """Returns whether the model can optionally use edge types.""" + return False + + @classmethod + def is_stocastic(cls) -> bool: + """Returns whether the model is stocastic and has therefore a random state.""" + return True \ No newline at end of file diff --git a/embiggen/embedders/ensmallen_embedders/siamese_model.py b/embiggen/embedders/ensmallen_embedders/siamese_model.py new file mode 100644 index 00000000..72615225 --- /dev/null +++ b/embiggen/embedders/ensmallen_embedders/siamese_model.py @@ -0,0 +1,116 @@ +"""Module providing Siamese implementation.""" +from typing import Dict, Any, Optional +from ensmallen import models +from userinput.utils import must_be_in_set +from embiggen.embedders.ensmallen_embedders.ensmallen_embedder import EnsmallenEmbedder +from embiggen.utils import abstract_class + + +@abstract_class +class SiameseEnsmallen(EnsmallenEmbedder): + """Class implementing the Siamese algorithm.""" + + models = { + "TransE": models.TransE, + "TransH": models.TransH, + "Unstructured": models.Unstructured, + "Structured Embedding": models.StructuredEmbedding, + } + + def __init__( + self, + embedding_size: int = 100, + relu_bias: float = 1.0, + epochs: int = 100, + learning_rate: float = 0.1, + learning_rate_decay: float = 0.9, + node_embedding_path: Optional[str] = None, + random_state: int = 42, + verbose: bool = False, + enable_cache: bool = False, + paths: Dict[str, str] + ): + """Create new abstract Node2Vec method. + + Parameters + -------------------------- + embedding_size: int = 100 + Dimension of the embedding. + relu_bias: float = 1.0 + Bias to use for the relu. + In the Siamese paper it is called gamma. + epochs: int = 100 + The number of epochs to run the model for, by default 10. + learning_rate: float = 0.05 + The learning rate to update the gradient, by default 0.01. + learning_rate_decay: float = 0.9 + Factor to reduce the learning rate for at each epoch. By default 0.9. + node_embedding_path: Optional[str] = None + Path where to mmap and store the nodes embedding. + This is necessary to embed large graphs whose embedding will not + fit into the available main memory. + random_state: int = 42 + Random state to reproduce the embeddings. + verbose: bool = False + Whether to show loading bars. + enable_cache: bool = False + Whether to enable the cache, that is to + store the computed embedding. + """ + self._kwargs = dict( + relu_bias=relu_bias, + epochs=epochs, + learning_rate=learning_rate, + learning_rate_decay=learning_rate_decay, + node_embedding_path=node_embedding_path, + verbose=verbose, + paths + ) + + self._model_name = must_be_in_set( + self.model_name(), + SiameseEnsmallen.models, + "Siamese models" + ) + + self._model = SiameseEnsmallen.models[self._model_name]( + embedding_size=embedding_size, + random_state=random_state, + self._kwargs + ) + + super().__init__( + embedding_size=embedding_size, + enable_cache=enable_cache, + random_state=random_state + ) + + def parameters(self) -> Dict[str, Any]: + """Returns parameters of the model.""" + return dict( + super().parameters(), + self._kwargs, + ) + + @classmethod + def smoke_test_parameters(cls) -> Dict[str, Any]: + """Returns parameters for smoke test.""" + return dict( + embedding_size=5, + epochs=1 + ) + + @classmethod + def can_use_edge_weights(cls) -> bool: + """Returns whether the model can optionally use edge weights.""" + return False + + @classmethod + def can_use_node_types(cls) -> bool: + """Returns whether the model can optionally use node types.""" + return False + + @classmethod + def is_stocastic(cls) -> bool: + """Returns whether the model is stocastic and has therefore a random state.""" + return True diff --git a/embiggen/embedders/ensmallen_embedders/sociodim.py b/embiggen/embedders/ensmallen_embedders/sociodim.py index aa8f3923..330153c5 100644 --- a/embiggen/embedders/ensmallen_embedders/sociodim.py +++ b/embiggen/embedders/ensmallen_embedders/sociodim.py @@ -1,18 +1,20 @@ """Module providing SocioDim implementation.""" -from typing import Optional, Dict, Any +from typing import Dict, Any from ensmallen import Graph import pandas as pd -import numpy as np from scipy.linalg import eigh -from embiggen.utils.abstract_models import AbstractEmbeddingModel, EmbeddingResult +from scipy.sparse.linalg import eigsh +from embiggen.embedders.ensmallen_embedders.ensmallen_embedder import EnsmallenEmbedder +from embiggen.utils import EmbeddingResult -class SocioDimEnsmallen(AbstractEmbeddingModel): +class SocioDimEnsmallen(EnsmallenEmbedder): """Class implementing the SocioDim algorithm.""" def __init__( self, embedding_size: int = 100, + use_sparse_reduce: bool = True, enable_cache: bool = False ): """Create new SocioDim method. @@ -21,30 +23,53 @@ def __init__( -------------------------- embedding_size: int = 100 Dimension of the embedding. + use_sparse_reduce: bool = True + Whether to use the sparse reduce or the dense reduce for + the computation of eigenvectors. + For both reduce mechanisms, we are using LAPACK implementations. + For some currently unknown reason, their implementation using + a sparse reduce, even on dense matrices, yields better results. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. """ + self._use_sparse_reduce = use_sparse_reduce super().__init__( embedding_size=embedding_size, enable_cache=enable_cache, ) + def parameters(self) -> Dict[str, Any]: + """Returns parameters of the model.""" + return { + super().parameters(), + dict( + use_sparse_reduce=self._use_sparse_reduce, + ) + } + def _fit_transform( self, graph: Graph, return_dataframe: bool = True, - verbose: bool = True ) -> EmbeddingResult: """Return node embedding.""" - number_of_nodes = graph.get_number_of_nodes() - embedding = eigh( - graph.get_dense_modularity_matrix(), - eigvals=( - number_of_nodes-self._embedding_size, - number_of_nodes-1 - ) - )[1] + if self._use_sparse_reduce: + embedding = eigsh( + graph.get_dense_modularity_matrix(), + k=self._embedding_size, + which="LM", + return_eigenvectors=True + )[1] + else: + number_of_nodes = graph.get_number_of_nodes() + embedding = eigh( + graph.get_dense_modularity_matrix(), + eigvals=( + number_of_nodes-self._embedding_size, + number_of_nodes-1 + ) + )[1] if return_dataframe: node_names = graph.get_node_names() @@ -57,27 +82,11 @@ def _fit_transform( node_embeddings=embedding ) - @classmethod - def task_name(cls) -> str: - return "Node Embedding" - @classmethod def model_name(cls) -> str: """Returns name of the model.""" return "SocioDim" - @classmethod - def library_name(cls) -> str: - return "Ensmallen" - - @classmethod - def requires_nodes_sorted_by_decreasing_node_degree(cls) -> bool: - return False - - @classmethod - def is_topological(cls) -> bool: - return True - @classmethod def can_use_edge_weights(cls) -> bool: """Returns whether the model can optionally use edge weights.""" diff --git a/embiggen/embedders/ensmallen_embedders/structured_embedding.py b/embiggen/embedders/ensmallen_embedders/structured_embedding.py new file mode 100644 index 00000000..ccf0334b --- /dev/null +++ b/embiggen/embedders/ensmallen_embedders/structured_embedding.py @@ -0,0 +1,116 @@ +"""Module providing Structured Embedding implementation.""" +from typing import Optional +from ensmallen import Graph +import pandas as pd +from embiggen.embedders.ensmallen_embedders.siamese_model import SiameseEnsmallen +from embiggen.utils import EmbeddingResult + + +class StructuredEmbeddingEnsmallen(SiameseEnsmallen): + """Class implementing the Structured Embedding algorithm.""" + + def __init__( + self, + embedding_size: int = 100, + relu_bias: float = 1.0, + epochs: int = 100, + learning_rate: float = 0.05, + learning_rate_decay: float = 0.9, + node_embedding_path: Optional[str] = None, + source_edge_type_embedding_path: Optional[str] = None, + destination_edge_type_embedding_path: Optional[str] = None, + random_state: int = 42, + verbose: bool = False, + enable_cache: bool = False + ): + """Create new Structured Embedding method. + + Parameters + -------------------------- + model_name: str + The model to instantiate. + embedding_size: int = 100 + Dimension of the embedding. + relu_bias: float = 1.0 + Bias to use for the relu. + In the StructuredEmbedding paper it is called gamma. + epochs: int = 100 + The number of epochs to run the model for, by default 10. + learning_rate: float = 0.05 + The learning rate to update the gradient, by default 0.01. + learning_rate_decay: float = 0.9 + Factor to reduce the learning rate for at each epoch. By default 0.9. + node_embedding_path: Optional[str] = None + Path where to mmap and store the nodes embedding. + This is necessary to embed large graphs whose embedding will not + fit into the available main memory. + source_edge_type_embedding_path: Optional[str] = None + Path where to mmap and store the source edge type embedding. + This is necessary to embed large graphs whose embedding will not + fit into the available main memory. + destination_edge_type_embedding_path: Optional[str] = None + Path where to mmap and store the destination edge type embedding. + This is necessary to embed large graphs whose embedding will not + fit into the available main memory. + random_state: int = 42 + Random state to reproduce the embeddings. + verbose: bool = False + Whether to show loading bars. + enable_cache: bool = False + Whether to enable the cache, that is to + store the computed embedding. + """ + super().__init__( + embedding_size=embedding_size, + relu_bias=relu_bias, + epochs=epochs, + learning_rate=learning_rate, + learning_rate_decay=learning_rate_decay, + node_embedding_path=node_embedding_path, + source_edge_type_embedding_path=source_edge_type_embedding_path, + destination_edge_type_embedding_path=destination_edge_type_embedding_path, + random_state=random_state, + verbose=verbose, + enable_cache=enable_cache, + ) + + def _fit_transform( + self, + graph: Graph, + return_dataframe: bool = True, + ) -> EmbeddingResult: + """Return node embedding.""" + node_embedding, source_edge_type_embedding, destination_edge_type_embedding = self._model.fit_transform( + graph, + ) + if return_dataframe: + node_embedding = pd.DataFrame( + node_embedding, + index=graph.get_node_names() + ) + source_edge_type_embedding = pd.DataFrame( + source_edge_type_embedding, + index=graph.get_unique_edge_type_names() + ) + destination_edge_type_embedding = pd.DataFrame( + destination_edge_type_embedding, + index=graph.get_unique_edge_type_names() + ) + + return EmbeddingResult( + embedding_method_name=self.model_name(), + node_embeddings=node_embedding, + edge_type_embeddings=[ + source_edge_type_embedding, + destination_edge_type_embedding + ] + ) + + @classmethod + def model_name(cls) -> str: + """Returns name of the model.""" + return "Structured Embedding" + + @classmethod + def requires_edge_types(cls) -> bool: + return True \ No newline at end of file diff --git a/embiggen/embedders/ensmallen_embedders/transe.py b/embiggen/embedders/ensmallen_embedders/transe.py index 4069762f..1a135d14 100644 --- a/embiggen/embedders/ensmallen_embedders/transe.py +++ b/embiggen/embedders/ensmallen_embedders/transe.py @@ -1,101 +1,81 @@ """Module providing TransE implementation.""" -from typing import Dict, Any, Union +from typing import Optional from ensmallen import Graph -import numpy as np import pandas as pd -from ensmallen import models -from embiggen.utils.abstract_models import AbstractEmbeddingModel, EmbeddingResult +from embiggen.embedders.ensmallen_embedders.siamese_model import SiameseEnsmallen +from embiggen.utils import EmbeddingResult -class TransEEnsmallen(AbstractEmbeddingModel): +class TransEEnsmallen(SiameseEnsmallen): """Class implementing the TransE algorithm.""" def __init__( self, embedding_size: int = 100, - renormalize: bool = True, relu_bias: float = 1.0, epochs: int = 100, - learning_rate: float = 0.01, + learning_rate: float = 0.05, learning_rate_decay: float = 0.9, + node_embedding_path: Optional[str] = None, + edge_type_embedding_path: Optional[str] = None, random_state: int = 42, + verbose: bool = False, enable_cache: bool = False ): - """Create new abstract Node2Vec method. + """Create new TransE method. Parameters -------------------------- + model_name: str + The model to instantiate. embedding_size: int = 100 Dimension of the embedding. - renormalize: bool = True - Whether to renormalize at each loop, by default true. relu_bias: float = 1.0 Bias to use for the relu. In the TransE paper it is called gamma. epochs: int = 100 The number of epochs to run the model for, by default 10. - learning_rate: float = 0.01 + learning_rate: float = 0.05 The learning rate to update the gradient, by default 0.01. learning_rate_decay: float = 0.9 Factor to reduce the learning rate for at each epoch. By default 0.9. + node_embedding_path: Optional[str] = None + Path where to mmap and store the nodes embedding. + This is necessary to embed large graphs whose embedding will not + fit into the available main memory. + edge_type_embedding_path: Optional[str] = None + Path where to mmap and store the edge type embedding. + This is necessary to embed large graphs whose embedding will not + fit into the available main memory. random_state: int = 42 Random state to reproduce the embeddings. + verbose: bool = False + Whether to show loading bars. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. """ - self._renormalize = renormalize - self._relu_bias = relu_bias - self._epochs = epochs - self._learning_rate = learning_rate - self._learning_rate_decay = learning_rate_decay - - self._model = models.TransE( - embedding_size=embedding_size, - renormalize=renormalize, - random_state=random_state - ) - super().__init__( embedding_size=embedding_size, + relu_bias=relu_bias, + epochs=epochs, + learning_rate=learning_rate, + learning_rate_decay=learning_rate_decay, + node_embedding_path=node_embedding_path, + edge_type_embedding_path=edge_type_embedding_path, + random_state=random_state, + verbose=verbose, enable_cache=enable_cache, - random_state=random_state - ) - - def parameters(self) -> Dict[str, Any]: - """Returns parameters of the model.""" - return { - super().parameters(), - dict( - renormalize=self._renormalize, - relu_bias=self._relu_bias, - epochs=self._epochs, - learning_rate=self._learning_rate, - learning_rate_decay=self._learning_rate_decay, - ) - } - - @classmethod - def smoke_test_parameters(cls) -> Dict[str, Any]: - """Returns parameters for smoke test.""" - return dict( - embedding_size=5, - epochs=1 ) def _fit_transform( self, graph: Graph, return_dataframe: bool = True, - verbose: bool = True ) -> EmbeddingResult: """Return node embedding.""" node_embedding, edge_type_embedding = self._model.fit_transform( graph, - epochs=self._epochs, - learning_rate=self._learning_rate, - learning_rate_decay=self._learning_rate_decay, - verbose=verbose, ) if return_dataframe: node_embedding = pd.DataFrame( @@ -109,51 +89,15 @@ def _fit_transform( return EmbeddingResult( embedding_method_name=self.model_name(), - node_embeddings= node_embedding, - edge_type_embeddings= edge_type_embedding, + node_embeddings=node_embedding, + edge_type_embeddings=edge_type_embedding, ) - @classmethod - def task_name(cls) -> str: - return "Node Embedding" - @classmethod def model_name(cls) -> str: """Returns name of the model.""" return "TransE" - @classmethod - def library_name(cls) -> str: - return "Ensmallen" - - @classmethod - def requires_nodes_sorted_by_decreasing_node_degree(cls) -> bool: - return False - - @classmethod - def is_topological(cls) -> bool: - return True - @classmethod def requires_edge_types(cls) -> bool: return True - - @classmethod - def can_use_edge_weights(cls) -> bool: - """Returns whether the model can optionally use edge weights.""" - return False - - @classmethod - def can_use_node_types(cls) -> bool: - """Returns whether the model can optionally use node types.""" - return False - - @classmethod - def task_involves_edge_types(cls) -> bool: - """Returns whether the model task involves edge types.""" - return True - - @classmethod - def is_stocastic(cls) -> bool: - """Returns whether the model is stocastic and has therefore a random state.""" - return True \ No newline at end of file diff --git a/embiggen/embedders/ensmallen_embedders/transh.py b/embiggen/embedders/ensmallen_embedders/transh.py new file mode 100644 index 00000000..2de65834 --- /dev/null +++ b/embiggen/embedders/ensmallen_embedders/transh.py @@ -0,0 +1,116 @@ +"""Module providing TransH implementation.""" +from typing import Optional +from ensmallen import Graph +import pandas as pd +from embiggen.embedders.ensmallen_embedders.siamese_model import SiameseEnsmallen +from embiggen.utils import EmbeddingResult + + +class TransHEnsmallen(SiameseEnsmallen): + """Class implementing the TransH algorithm.""" + + def __init__( + self, + embedding_size: int = 100, + relu_bias: float = 1.0, + epochs: int = 100, + learning_rate: float = 0.1, + learning_rate_decay: float = 0.9, + node_embedding_path: Optional[str] = None, + mult_edge_type_embedding_path: Optional[str] = None, + bias_edge_type_embedding_path: Optional[str] = None, + random_state: int = 42, + verbose: bool = False, + enable_cache: bool = False + ): + """Create new TransH method. + + Parameters + -------------------------- + model_name: str + The model to instantiate. + embedding_size: int = 100 + Dimension of the embedding. + relu_bias: float = 1.0 + Bias to use for the relu. + In the TransH paper it is called gamma. + epochs: int = 100 + The number of epochs to run the model for, by default 10. + learning_rate: float = 0.05 + The learning rate to update the gradient, by default 0.01. + learning_rate_decay: float = 0.9 + Factor to reduce the learning rate for at each epoch. By default 0.9. + node_embedding_path: Optional[str] = None + Path where to mmap and store the nodes embedding. + This is necessary to embed large graphs whose embedding will not + fit into the available main memory. + mult_edge_type_embedding_path: Optional[str] = None + Path where to mmap and store the multiplicative edge type embedding. + This is necessary to embed large graphs whose embedding will not + fit into the available main memory. + bias_edge_type_embedding_path: Optional[str] = None + Path where to mmap and store the bias edge type embedding. + This is necessary to embed large graphs whose embedding will not + fit into the available main memory. + random_state: int = 42 + Random state to reproduce the embeddings. + verbose: bool = False + Whether to show loading bars. + enable_cache: bool = False + Whether to enable the cache, that is to + store the computed embedding. + """ + super().__init__( + embedding_size=embedding_size, + relu_bias=relu_bias, + epochs=epochs, + learning_rate=learning_rate, + learning_rate_decay=learning_rate_decay, + node_embedding_path=node_embedding_path, + mult_edge_type_embedding_path=mult_edge_type_embedding_path, + bias_edge_type_embedding_path=bias_edge_type_embedding_path, + random_state=random_state, + verbose=verbose, + enable_cache=enable_cache, + ) + + def _fit_transform( + self, + graph: Graph, + return_dataframe: bool = True, + ) -> EmbeddingResult: + """Return node embedding.""" + node_embedding, mult_edge_type_embedding, bias_edge_type_embedding = self._model.fit_transform( + graph, + ) + if return_dataframe: + node_embedding = pd.DataFrame( + node_embedding, + index=graph.get_node_names() + ) + mult_edge_type_embedding = pd.DataFrame( + mult_edge_type_embedding, + index=graph.get_unique_edge_type_names() + ) + bias_edge_type_embedding = pd.DataFrame( + bias_edge_type_embedding, + index=graph.get_unique_edge_type_names() + ) + + return EmbeddingResult( + embedding_method_name=self.model_name(), + node_embeddings=node_embedding, + edge_type_embeddings=[ + mult_edge_type_embedding, + bias_edge_type_embedding + ] + ) + + @classmethod + def model_name(cls) -> str: + """Returns name of the model.""" + return "TransH" + + @classmethod + def requires_edge_types(cls) -> bool: + return True \ No newline at end of file diff --git a/embiggen/embedders/ensmallen_embedders/unstructured.py b/embiggen/embedders/ensmallen_embedders/unstructured.py new file mode 100644 index 00000000..edcaf5f0 --- /dev/null +++ b/embiggen/embedders/ensmallen_embedders/unstructured.py @@ -0,0 +1,92 @@ +"""Module providing Unstructured implementation.""" +from typing import Optional +from ensmallen import Graph +import pandas as pd +from embiggen.embedders.ensmallen_embedders.siamese_model import SiameseEnsmallen +from embiggen.utils import EmbeddingResult + + +class UnstructuredEnsmallen(SiameseEnsmallen): + """Class implementing the Unstructured algorithm.""" + + def __init__( + self, + embedding_size: int = 100, + relu_bias: float = 1.0, + epochs: int = 100, + learning_rate: float = 0.01, + learning_rate_decay: float = 0.9, + node_embedding_path: Optional[str] = None, + random_state: int = 42, + verbose: bool = False, + enable_cache: bool = False + ): + """Create new abstract Node2Vec method. + + Parameters + -------------------------- + model_name: str + The model to instantiate. + embedding_size: int = 100 + Dimension of the embedding. + relu_bias: float = 1.0 + Bias to use for the relu. + In the Unstructured paper it is called gamma. + epochs: int = 100 + The number of epochs to run the model for, by default 100. + learning_rate: float = 0.01 + The learning rate to update the gradient, by default 0.01. + learning_rate_decay: float = 0.9 + Factor to reduce the learning rate for at each epoch. By default 0.9. + node_embedding_path: Optional[str] = None + Path where to mmap and store the nodes embedding. + This is necessary to embed large graphs whose embedding will not + fit into the available main memory. + random_state: int = 42 + Random state to reproduce the embeddings. + verbose: bool = False + Whether to show loading bars. + enable_cache: bool = False + Whether to enable the cache, that is to + store the computed embedding. + """ + super().__init__( + embedding_size=embedding_size, + relu_bias=relu_bias, + epochs=epochs, + learning_rate=learning_rate, + learning_rate_decay=learning_rate_decay, + node_embedding_path=node_embedding_path, + random_state=random_state, + verbose=verbose, + enable_cache=enable_cache, + ) + + def _fit_transform( + self, + graph: Graph, + return_dataframe: bool = True, + ) -> EmbeddingResult: + """Return node embedding.""" + node_embedding = self._model.fit_transform( + graph, + ) + if return_dataframe: + node_embedding = pd.DataFrame( + node_embedding, + index=graph.get_node_names() + ) + + return EmbeddingResult( + embedding_method_name=self.model_name(), + node_embeddings=node_embedding, + ) + + @classmethod + def model_name(cls) -> str: + """Returns name of the model.""" + return "Unstructured" + + @classmethod + def can_use_edge_types(cls) -> bool: + return False \ No newline at end of file diff --git a/embiggen/embedders/ensmallen_embedders/walklets.py b/embiggen/embedders/ensmallen_embedders/walklets.py new file mode 100644 index 00000000..0f3a4ddd --- /dev/null +++ b/embiggen/embedders/ensmallen_embedders/walklets.py @@ -0,0 +1,142 @@ +"""Module providing WalkletsEnsmallen model implementation.""" +from typing import Optional, Dict, Any +from embiggen.embedders.ensmallen_embedders.node2vec import Node2VecEnsmallen +from embiggen.utils.abstract_models.abstract_model import abstract_class + +@abstract_class +class WalkletsEnsmallen(Node2VecEnsmallen): + """Class providing WalkletsEnsmallen implemeted in Rust from Ensmallen.""" + + def __init__( + self, + embedding_size: int = 100, + epochs: int = 10, + clipping_value: float = 6.0, + number_of_negative_samples: int = 10, + walk_length: int = 128, + iterations: int = 10, + window_size: int = 4, + return_weight: float = 1.0, + explore_weight: float = 1.0, + max_neighbours: Optional[int] = 100, + learning_rate: float = 0.01, + learning_rate_decay: float = 0.9, + central_nodes_embedding_path: Optional[str] = None, + contextual_nodes_embedding_path: Optional[str] = None, + alpha: float = 0.75, + normalize_by_degree: bool = False, + stochastic_downsample_by_degree: Optional[bool] = False, + normalize_learning_rate_by_degree: Optional[bool] = False, + use_scale_free_distribution: Optional[bool] = True, + random_state: int = 42, + enable_cache: bool = False + ): + """Create new abstract Node2Vec method. + + Parameters + -------------------------- + embedding_size: int = 100 + Dimension of the embedding. + epochs: int = 30 + Number of epochs to train the model for. + clipping_value: float = 6.0 + Value at which we clip the dot product, mostly for numerical stability issues. + By default, `6.0`, where the loss is already close to zero. + number_of_negative_samples: int = 10 + The number of negative classes to randomly sample per batch. + This single sample of negative classes is evaluated for each element in the batch. + walk_length: int = 128 + Maximal length of the walks. + iterations: int = 10 + Number of iterations of the single walks. + window_size: int = 4 + Window size for the local context. + On the borders the window size is trimmed. + return_weight: float = 1.0 + Weight on the probability of returning to the same node the walk just came from + Having this higher tends the walks to be + more like a Breadth-First Search. + Having this very high (> 2) makes search very local. + Equal to the inverse of p in the Node2Vec paper. + explore_weight: float = 1.0 + Weight on the probability of visiting a neighbor node + to the one we're coming from in the random walk + Having this higher tends the walks to be + more like a Depth-First Search. + Having this very high makes search more outward. + Having this very low makes search very local. + Equal to the inverse of q in the Node2Vec paper. + max_neighbours: Optional[int] = 100 + Number of maximum neighbours to consider when using approximated walks. + By default, None, we execute exact random walks. + This is mainly useful for graphs containing nodes with high degrees. + learning_rate: float = 0.01 + The learning rate to use to train the Node2Vec model. By default 0.01. + learning_rate_decay: float = 0.9 + Factor to reduce the learning rate for at each epoch. By default 0.9. + central_nodes_embedding_path: Optional[str] = None + Path where to mmap and store the central nodes embedding. + If provided, we expect the path to contain the substring `{window_size}` which + will be replaced with the i-th window size embedding that is being computed. + This is necessary to embed large graphs whose embedding will not + fit into the available main memory. + contextual_nodes_embedding_path: Optional[str] = None + Path where to mmap and store the central nodes embedding. + If provided, we expect the path to contain the substring `{window_size}` which + will be replaced with the i-th window size embedding that is being computed. + This is necessary to embed large graphs whose embedding will not + fit into the available main memory. + alpha: float = 0.75 + Alpha parameter for GloVe's loss. + normalize_by_degree: bool = False + Whether to normalize the random walk by the node degree + of the destination node degrees. + stochastic_downsample_by_degree: Optional[bool] = False + Randomly skip samples with probability proportional to the degree of the central node. By default false. + normalize_learning_rate_by_degree: Optional[bool] = False + Divide the learning rate by the degree of the central node. By default false. + use_scale_free_distribution: Optional[bool] = True + Sample negatives proportionally to their degree. By default true. + random_state: int = 42 + The random state to reproduce the training sequence. + enable_cache: bool = False + Whether to enable the cache, that is to + store the computed embedding. + """ + super().__init__( + embedding_size=embedding_size // window_size, + epochs=epochs, + clipping_value=clipping_value, + number_of_negative_samples=number_of_negative_samples, + walk_length=walk_length, + iterations=iterations, + window_size=window_size, + return_weight=return_weight, + explore_weight=explore_weight, + max_neighbours=max_neighbours, + learning_rate=learning_rate, + learning_rate_decay=learning_rate_decay, + central_nodes_embedding_path=central_nodes_embedding_path, + contextual_nodes_embedding_path=contextual_nodes_embedding_path, + alpha=alpha, + normalize_by_degree=normalize_by_degree, + stochastic_downsample_by_degree=stochastic_downsample_by_degree, + normalize_learning_rate_by_degree=normalize_learning_rate_by_degree, + use_scale_free_distribution=use_scale_free_distribution, + random_state=random_state, + enable_cache=enable_cache + ) + + def parameters(self) -> Dict[str, Any]: + """Returns parameters of the model.""" + parameters = super().parameters() + parameters["embedding_size"] = parameters["embedding_size"] parameters["window_size"] + return parameters + + @classmethod + def requires_node_types(cls) -> bool: + return False + + @classmethod + def requires_edge_types(cls) -> bool: + return False \ No newline at end of file diff --git a/embiggen/embedders/ensmallen_embedders/walklets_cbow.py b/embiggen/embedders/ensmallen_embedders/walklets_cbow.py index 51e8ccfe..99d73975 100644 --- a/embiggen/embedders/ensmallen_embedders/walklets_cbow.py +++ b/embiggen/embedders/ensmallen_embedders/walklets_cbow.py @@ -1,31 +1,31 @@ """Module providing WalkletsCBOW model implementation.""" -from typing import Optional -from embiggen.embedders.ensmallen_embedders.node2vec import Node2VecEnsmallen +from typing import Optional, Dict, Any +from embiggen.embedders.ensmallen_embedders.walklets import WalkletsEnsmallen -class WalkletsCBOWEnsmallen(Node2VecEnsmallen): +class WalkletsCBOWEnsmallen(WalkletsEnsmallen): """Class providing WalkletsCBOW implemeted in Rust from Ensmallen.""" def __init__( self, embedding_size: int = 100, - epochs: int = 10, + epochs: int = 30, clipping_value: float = 6.0, number_of_negative_samples: int = 10, walk_length: int = 128, iterations: int = 10, - window_size: int = 10, + window_size: int = 4, return_weight: float = 1.0, explore_weight: float = 1.0, - change_node_type_weight: float = 1.0, - change_edge_type_weight: float = 1.0, max_neighbours: Optional[int] = 100, learning_rate: float = 0.01, learning_rate_decay: float = 0.9, + central_nodes_embedding_path: Optional[str] = None, + contextual_nodes_embedding_path: Optional[str] = None, normalize_by_degree: bool = False, stochastic_downsample_by_degree: Optional[bool] = False, normalize_learning_rate_by_degree: Optional[bool] = False, - use_zipfian_sampling: Optional[bool] = True, + use_scale_free_distribution: Optional[bool] = True, random_state: int = 42, enable_cache: bool = False ): @@ -35,7 +35,7 @@ def __init__( -------------------------- embedding_size: int = 100 Dimension of the embedding. - epochs: int = 10 + epochs: int = 30 Number of epochs to train the model for. clipping_value: float = 6.0 Value at which we clip the dot product, mostly for numerical stability issues. @@ -47,7 +47,7 @@ def __init__( Maximal length of the walks. iterations: int = 10 Number of iterations of the single walks. - window_size: int = 10 + window_size: int = 4 Window size for the local context. On the borders the window size is trimmed. return_weight: float = 1.0 @@ -64,14 +64,6 @@ def __init__( Having this very high makes search more outward. Having this very low makes search very local. Equal to the inverse of q in the Node2Vec paper. - change_node_type_weight: float = 1.0 - Weight on the probability of visiting a neighbor node of a - different type than the previous node. This only applies to - colored graphs, otherwise it has no impact. - change_edge_type_weight: float = 1.0 - Weight on the probability of visiting a neighbor edge of a - different type than the previous edge. This only applies to - multigraphs, otherwise it has no impact. max_neighbours: Optional[int] = 100 Number of maximum neighbours to consider when using approximated walks. By default, None, we execute exact random walks. @@ -80,6 +72,18 @@ def __init__( The learning rate to use to train the Node2Vec model. By default 0.01. learning_rate_decay: float = 0.9 Factor to reduce the learning rate for at each epoch. By default 0.9. + central_nodes_embedding_path: Optional[str] = None + Path where to mmap and store the central nodes embedding. + If provided, we expect the path to contain the substring `{window_size}` which + will be replaced with the i-th window size embedding that is being computed. + This is necessary to embed large graphs whose embedding will not + fit into the available main memory. + contextual_nodes_embedding_path: Optional[str] = None + Path where to mmap and store the central nodes embedding. + If provided, we expect the path to contain the substring `{window_size}` which + will be replaced with the i-th window size embedding that is being computed. + This is necessary to embed large graphs whose embedding will not + fit into the available main memory. normalize_by_degree: bool = False Whether to normalize the random walk by the node degree of the destination node degrees. @@ -87,7 +91,7 @@ def __init__( Randomly skip samples with probability proportional to the degree of the central node. By default false. normalize_learning_rate_by_degree: Optional[bool] = False Divide the learning rate by the degree of the central node. By default false. - use_zipfian_sampling: Optional[bool] = True + use_scale_free_distribution: Optional[bool] = True Sample negatives proportionally to their degree. By default true. random_state: int = 42 The random state to reproduce the training sequence. @@ -96,8 +100,7 @@ def __init__( store the computed embedding. """ super().__init__( - model_name="WalkletsCBOW", - embedding_size=embedding_size // window_size, + embedding_size=embedding_size, epochs=epochs, clipping_value=clipping_value, number_of_negative_samples=number_of_negative_samples, @@ -106,28 +109,33 @@ def __init__( window_size=window_size, return_weight=return_weight, explore_weight=explore_weight, - change_edge_type_weight=change_edge_type_weight, - change_node_type_weight=change_node_type_weight, max_neighbours=max_neighbours, learning_rate=learning_rate, learning_rate_decay=learning_rate_decay, + central_nodes_embedding_path=central_nodes_embedding_path, + contextual_nodes_embedding_path=contextual_nodes_embedding_path, normalize_by_degree=normalize_by_degree, stochastic_downsample_by_degree=stochastic_downsample_by_degree, normalize_learning_rate_by_degree=normalize_learning_rate_by_degree, - use_zipfian_sampling=use_zipfian_sampling, + use_scale_free_distribution=use_scale_free_distribution, random_state=random_state, enable_cache=enable_cache ) - + + def parameters(self) -> Dict[str, Any]: + """Returns parameters for smoke test.""" + removed = [ + "alpha", + ] + return dict( + { + key: value + for key, value in super().parameters().items() + if key not in removed + } + ) + @classmethod def model_name(cls) -> str: """Returns name of the model.""" return "Walklets CBOW" - - @classmethod - def requires_node_types(cls) -> bool: - return False - - @classmethod - def requires_edge_types(cls) -> bool: - return False \ No newline at end of file diff --git a/embiggen/embedders/ensmallen_embedders/walklets_glove.py b/embiggen/embedders/ensmallen_embedders/walklets_glove.py new file mode 100644 index 00000000..64fb7f4b --- /dev/null +++ b/embiggen/embedders/ensmallen_embedders/walklets_glove.py @@ -0,0 +1,136 @@ +"""Module providing WalkletsGloVe model implementation.""" +from typing import Optional, Dict, Any +from embiggen.embedders.ensmallen_embedders.walklets import WalkletsEnsmallen + + +class WalkletsGloVeEnsmallen(WalkletsEnsmallen): + """Class providing WalkletsGloVe implemeted in Rust from Ensmallen.""" + + def __init__( + self, + embedding_size: int = 100, + epochs: int = 30, + walk_length: int = 128, + iterations: int = 10, + window_size: int = 4, + return_weight: float = 1.0, + explore_weight: float = 1.0, + max_neighbours: Optional[int] = 100, + learning_rate: float = 0.001, + learning_rate_decay: float = 0.9, + central_nodes_embedding_path: Optional[str] = None, + contextual_nodes_embedding_path: Optional[str] = None, + alpha: float = 0.75, + normalize_by_degree: bool = False, + stochastic_downsample_by_degree: Optional[bool] = False, + normalize_learning_rate_by_degree: Optional[bool] = False, + use_scale_free_distribution: Optional[bool] = True, + random_state: int = 42, + enable_cache: bool = False + ): + """Create new abstract Node2Vec method. + + Parameters + -------------------------- + embedding_size: int = 100 + Dimension of the embedding. + epochs: int = 30 + Number of epochs to train the model for. + walk_length: int = 128 + Maximal length of the walks. + iterations: int = 10 + Number of iterations of the single walks. + window_size: int = 4 + Window size for the local context. + On the borders the window size is trimmed. + return_weight: float = 1.0 + Weight on the probability of returning to the same node the walk just came from + Having this higher tends the walks to be + more like a Breadth-First Search. + Having this very high (> 2) makes search very local. + Equal to the inverse of p in the Node2Vec paper. + explore_weight: float = 1.0 + Weight on the probability of visiting a neighbor node + to the one we're coming from in the random walk + Having this higher tends the walks to be + more like a Depth-First Search. + Having this very high makes search more outward. + Having this very low makes search very local. + Equal to the inverse of q in the Node2Vec paper. + max_neighbours: Optional[int] = 100 + Number of maximum neighbours to consider when using approximated walks. + By default, None, we execute exact random walks. + This is mainly useful for graphs containing nodes with high degrees. + learning_rate: float = 0.001 + The learning rate to use to train the Node2Vec model. By default 0.01. + learning_rate_decay: float = 0.9 + Factor to reduce the learning rate for at each epoch. By default 0.9. + central_nodes_embedding_path: Optional[str] = None + Path where to mmap and store the central nodes embedding. + If provided, we expect the path to contain the substring `{window_size}` which + will be replaced with the i-th window size embedding that is being computed. + This is necessary to embed large graphs whose embedding will not + fit into the available main memory. + contextual_nodes_embedding_path: Optional[str] = None + Path where to mmap and store the central nodes embedding. + If provided, we expect the path to contain the substring `{window_size}` which + will be replaced with the i-th window size embedding that is being computed. + This is necessary to embed large graphs whose embedding will not + fit into the available main memory. + alpha: float = 0.75 + Alpha parameter for GloVe's loss. + normalize_by_degree: bool = False + Whether to normalize the random walk by the node degree + of the destination node degrees. + stochastic_downsample_by_degree: Optional[bool] = False + Randomly skip samples with probability proportional to the degree of the central node. By default false. + normalize_learning_rate_by_degree: Optional[bool] = False + Divide the learning rate by the degree of the central node. By default false. + use_scale_free_distribution: Optional[bool] = True + Sample negatives proportionally to their degree. By default true. + random_state: int = 42 + The random state to reproduce the training sequence. + enable_cache: bool = False + Whether to enable the cache, that is to + store the computed embedding. + """ + super().__init__( + embedding_size=embedding_size, + epochs=epochs, + alpha=alpha, + walk_length=walk_length, + iterations=iterations, + window_size=window_size, + return_weight=return_weight, + explore_weight=explore_weight, + max_neighbours=max_neighbours, + learning_rate=learning_rate, + learning_rate_decay=learning_rate_decay, + central_nodes_embedding_path=central_nodes_embedding_path, + contextual_nodes_embedding_path=contextual_nodes_embedding_path, + normalize_by_degree=normalize_by_degree, + stochastic_downsample_by_degree=stochastic_downsample_by_degree, + normalize_learning_rate_by_degree=normalize_learning_rate_by_degree, + use_scale_free_distribution=use_scale_free_distribution, + random_state=random_state, + enable_cache=enable_cache + ) + + @classmethod + def model_name(cls) -> str: + """Returns name of the model.""" + return "Walklets GloVe" + + def parameters(self) -> Dict[str, Any]: + """Returns parameters for smoke test.""" + removed = [ + "number_of_negative_samples", + "clipping_value" + ] + return dict( + { + key: value + for key, value in super().parameters().items() + if key not in removed + } + ) \ No newline at end of file diff --git a/embiggen/embedders/ensmallen_embedders/walklets_skipgram.py b/embiggen/embedders/ensmallen_embedders/walklets_skipgram.py index 2cd9bedf..73ce0c94 100644 --- a/embiggen/embedders/ensmallen_embedders/walklets_skipgram.py +++ b/embiggen/embedders/ensmallen_embedders/walklets_skipgram.py @@ -1,31 +1,31 @@ """Module providing WalkletsSkipGram model implementation.""" -from typing import Optional -from embiggen.embedders.ensmallen_embedders.node2vec import Node2VecEnsmallen +from typing import Optional, Dict, Any +from embiggen.embedders.ensmallen_embedders.walklets import WalkletsEnsmallen -class WalkletsSkipGramEnsmallen(Node2VecEnsmallen): +class WalkletsSkipGramEnsmallen(WalkletsEnsmallen): """Class providing WalkletsSkipGram implemeted in Rust from Ensmallen.""" def __init__( self, embedding_size: int = 100, - epochs: int = 10, + epochs: int = 30, clipping_value: float = 6.0, number_of_negative_samples: int = 10, walk_length: int = 128, iterations: int = 10, - window_size: int = 10, + window_size: int = 4, return_weight: float = 1.0, explore_weight: float = 1.0, - change_node_type_weight: float = 1.0, - change_edge_type_weight: float = 1.0, max_neighbours: Optional[int] = 100, learning_rate: float = 0.01, learning_rate_decay: float = 0.9, + central_nodes_embedding_path: Optional[str] = None, + contextual_nodes_embedding_path: Optional[str] = None, normalize_by_degree: bool = False, stochastic_downsample_by_degree: Optional[bool] = False, normalize_learning_rate_by_degree: Optional[bool] = False, - use_zipfian_sampling: Optional[bool] = True, + use_scale_free_distribution: Optional[bool] = True, random_state: int = 42, enable_cache: bool = False ): @@ -35,7 +35,7 @@ def __init__( -------------------------- embedding_size: int = 100 Dimension of the embedding. - epochs: int = 10 + epochs: int = 30 Number of epochs to train the model for. clipping_value: float = 6.0 Value at which we clip the dot product, mostly for numerical stability issues. @@ -47,7 +47,7 @@ def __init__( Maximal length of the walks. iterations: int = 10 Number of iterations of the single walks. - window_size: int = 10 + window_size: int = 4 Window size for the local context. On the borders the window size is trimmed. return_weight: float = 1.0 @@ -64,14 +64,6 @@ def __init__( Having this very high makes search more outward. Having this very low makes search very local. Equal to the inverse of q in the Node2Vec paper. - change_node_type_weight: float = 1.0 - Weight on the probability of visiting a neighbor node of a - different type than the previous node. This only applies to - colored graphs, otherwise it has no impact. - change_edge_type_weight: float = 1.0 - Weight on the probability of visiting a neighbor edge of a - different type than the previous edge. This only applies to - multigraphs, otherwise it has no impact. max_neighbours: Optional[int] = 100 Number of maximum neighbours to consider when using approximated walks. By default, None, we execute exact random walks. @@ -80,6 +72,18 @@ def __init__( The learning rate to use to train the Node2Vec model. By default 0.01. learning_rate_decay: float = 0.9 Factor to reduce the learning rate for at each epoch. By default 0.9. + central_nodes_embedding_path: Optional[str] = None + Path where to mmap and store the central nodes embedding. + If provided, we expect the path to contain the substring `{window_size}` which + will be replaced with the i-th window size embedding that is being computed. + This is necessary to embed large graphs whose embedding will not + fit into the available main memory. + contextual_nodes_embedding_path: Optional[str] = None + Path where to mmap and store the central nodes embedding. + If provided, we expect the path to contain the substring `{window_size}` which + will be replaced with the i-th window size embedding that is being computed. + This is necessary to embed large graphs whose embedding will not + fit into the available main memory. normalize_by_degree: bool = False Whether to normalize the random walk by the node degree of the destination node degrees. @@ -87,7 +91,7 @@ def __init__( Randomly skip samples with probability proportional to the degree of the central node. By default false. normalize_learning_rate_by_degree: Optional[bool] = False Divide the learning rate by the degree of the central node. By default false. - use_zipfian_sampling: Optional[bool] = True + use_scale_free_distribution: Optional[bool] = True Sample negatives proportionally to their degree. By default true. random_state: int = 42 The random state to reproduce the training sequence. @@ -96,8 +100,7 @@ def __init__( store the computed embedding. """ super().__init__( - model_name="WalkletsSkipGram", - embedding_size=embedding_size // window_size, + embedding_size=embedding_size, epochs=epochs, clipping_value=clipping_value, number_of_negative_samples=number_of_negative_samples, @@ -106,28 +109,33 @@ def __init__( window_size=window_size, return_weight=return_weight, explore_weight=explore_weight, - change_edge_type_weight=change_edge_type_weight, - change_node_type_weight=change_node_type_weight, max_neighbours=max_neighbours, learning_rate=learning_rate, learning_rate_decay=learning_rate_decay, + central_nodes_embedding_path=central_nodes_embedding_path, + contextual_nodes_embedding_path=contextual_nodes_embedding_path, normalize_by_degree=normalize_by_degree, stochastic_downsample_by_degree=stochastic_downsample_by_degree, normalize_learning_rate_by_degree=normalize_learning_rate_by_degree, - use_zipfian_sampling=use_zipfian_sampling, + use_scale_free_distribution=use_scale_free_distribution, random_state=random_state, enable_cache=enable_cache ) - + + def parameters(self) -> Dict[str, Any]: + """Returns parameters for smoke test.""" + removed = [ + "alpha", + ] + return dict( + { + key: value + for key, value in super().parameters().items() + if key not in removed + } + ) + @classmethod def model_name(cls) -> str: """Returns name of the model.""" return "Walklets SkipGram" - - @classmethod - def requires_node_types(cls) -> bool: - return False - - @classmethod - def requires_edge_types(cls) -> bool: - return False \ No newline at end of file diff --git a/embiggen/embedders/ensmallen_embedders/weighted_spine.py b/embiggen/embedders/ensmallen_embedders/weighted_spine.py index c7ad70e8..8cb06df0 100644 --- a/embiggen/embedders/ensmallen_embedders/weighted_spine.py +++ b/embiggen/embedders/ensmallen_embedders/weighted_spine.py @@ -3,16 +3,18 @@ from ensmallen import Graph import pandas as pd from ensmallen import models -from embiggen.utils.abstract_models import AbstractEmbeddingModel, EmbeddingResult +from embiggen.embedders.ensmallen_embedders.ensmallen_embedder import EnsmallenEmbedder +from embiggen.utils import EmbeddingResult -class WeightedSPINE(AbstractEmbeddingModel): +class WeightedSPINE(EnsmallenEmbedder): """Abstract class for Node2Vec algorithms.""" def __init__( self, embedding_size: int = 100, use_edge_weights_as_probabilities: bool = False, + verbose: bool = False, enable_cache: bool = False ): """Create new abstract Node2Vec method. @@ -23,10 +25,13 @@ def __init__( Dimension of the embedding. use_edge_weights_as_probabilities: bool = False Whether to treat the weights as probabilities. + verbose: bool = False + Whether to show loading bars. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. """ + self._verbose = verbose self._model = models.WeightedSPINE( embedding_size=embedding_size, use_edge_weights_as_probabilities=use_edge_weights_as_probabilities, @@ -48,12 +53,11 @@ def _fit_transform( self, graph: Graph, return_dataframe: bool = True, - verbose: bool = True ) -> EmbeddingResult: """Return node embedding.""" node_embedding = self._model.fit_transform( graph, - verbose=verbose, + verbose=self._verbose, ).T if return_dataframe: node_embedding = pd.DataFrame( @@ -65,27 +69,11 @@ def _fit_transform( node_embeddings=node_embedding ) - @classmethod - def task_name(cls) -> str: - return "Node Embedding" - @classmethod def model_name(cls) -> str: """Returns name of the model.""" return "WeightedSPINE" - @classmethod - def library_name(cls) -> str: - return "Ensmallen" - - @classmethod - def requires_nodes_sorted_by_decreasing_node_degree(cls) -> bool: - return False - - @classmethod - def is_topological(cls) -> bool: - return True - @classmethod def requires_edge_weights(cls) -> bool: return True diff --git a/embiggen/embedders/graph_embedding_pipeline.py b/embiggen/embedders/graph_embedding_pipeline.py index d1c9ba42..34713828 100644 --- a/embiggen/embedders/graph_embedding_pipeline.py +++ b/embiggen/embedders/graph_embedding_pipeline.py @@ -15,7 +15,6 @@ def embed_graph( library_name: Optional[str] = None, smoke_test: bool = False, return_dataframe: bool = True, - verbose: bool = True, kwargs: Dict ) -> EmbeddingResult: """Return embedding of the provided graph. @@ -45,14 +44,11 @@ def embed_graph( Path where to store the cache if it is enabled. return_dataframe: bool = True Whether to return a pandas DataFrame with the embedding. - verbose: bool - Whether to show loading bars. kwargs: Dict Kwargs to forward to the embedding model creation. If a model name was NOT provided, an exception will be raised as it is unclear how to behave. """ - graph = next(iterate_graphs( graphs=graph, repositories=repository, @@ -64,7 +60,7 @@ def embed_graph( model_name=embedding_model, library_name=library_name )(kwargs) - elif kwargs is not None: + elif kwargs: raise ValueError( "Please be advised that even though you have provided yourself " "the embedding model, you have also provided the kwargs which " @@ -80,9 +76,7 @@ def embed_graph( if smoke_test: try: - embedding_model = embedding_model.__class__( - *embedding_model.smoke_test_parameters() - ) + embedding_model = embedding_model.into_smoke_test() except Exception as e: raise ValueError( "An exception was raised while trying to create " @@ -100,7 +94,6 @@ def embed_graph( return embedding_model.fit_transform( graph, return_dataframe=return_dataframe, - verbose=verbose ) except Exception as e: raise ValueError( @@ -109,5 +102,5 @@ def embed_graph( f"using the model called {embedding_model.model_name()} " f"from the library {library_name}, specifically " f"implemented in the class {embedding_model.__class__.__name__}. " - f"The body of the exception was: {str(e)}." + f"The body of the exception was: {str(e)}" ) from e \ No newline at end of file diff --git a/embiggen/embedders/karateclub_embedders/boostne.py b/embiggen/embedders/karateclub_embedders/boostne.py index 3cca9296..461fee04 100644 --- a/embiggen/embedders/karateclub_embedders/boostne.py +++ b/embiggen/embedders/karateclub_embedders/boostne.py @@ -8,7 +8,7 @@ class BoostNEKarateClub(AbstractKarateClubEmbedder): def __init__( self, - embedding_size: int = 128, + embedding_size: int = 100, iterations: int = 16, order: int = 2, alpha: float = 0.01, @@ -19,7 +19,7 @@ def __init__( Parameters ---------------------- - embedding_size: int = 128 + embedding_size: int = 100 Size of the embedding to use. iterations: int = 16 Number of boosting iterations. Default is 16. diff --git a/embiggen/embedders/karateclub_embedders/deep_walk.py b/embiggen/embedders/karateclub_embedders/deep_walk.py index 36f858d0..db084890 100644 --- a/embiggen/embedders/karateclub_embedders/deep_walk.py +++ b/embiggen/embedders/karateclub_embedders/deep_walk.py @@ -5,11 +5,11 @@ from embiggen.embedders.karateclub_embedders.abstract_karateclub_embedder import AbstractKarateClubEmbedder -class DeepWalkKarateClub(AbstractKarateClubEmbedder): +class DeepWalkSkipGramKarateClub(AbstractKarateClubEmbedder): def __init__( self, - embedding_size: int = 128, + embedding_size: int = 100, walk_number: int = 10, walk_length: int = 80, window_size: int = 5, @@ -23,7 +23,7 @@ def __init__( Parameters ---------------------- - embedding_size: int = 128 + embedding_size: int = 100 Size of the embedding to use. walk_number: int = 10 Number of random walks. Default is 10. @@ -97,7 +97,7 @@ def _build_model(self) -> DeepWalk: @classmethod def model_name(cls) -> str: """Returns name of the model""" - return "DeepWalk" + return "DeepWalk SkipGram" @classmethod def requires_nodes_sorted_by_decreasing_node_degree(cls) -> bool: diff --git a/embiggen/embedders/karateclub_embedders/geometric_laplacian_eigenmaps.py b/embiggen/embedders/karateclub_embedders/geometric_laplacian_eigenmaps.py index 69793eb3..eb6982a3 100644 --- a/embiggen/embedders/karateclub_embedders/geometric_laplacian_eigenmaps.py +++ b/embiggen/embedders/karateclub_embedders/geometric_laplacian_eigenmaps.py @@ -8,7 +8,7 @@ class GLEEKarateClub(AbstractKarateClubEmbedder): def __init__( self, - embedding_size: int = 128, + embedding_size: int = 100, random_state: int = 42, enable_cache: bool = False ): @@ -16,7 +16,7 @@ def __init__( Parameters ---------------------- - embedding_size: int = 128 + embedding_size: int = 100 Size of the embedding to use. random_state: int = 42 Random state to use for the stocastic diff --git a/embiggen/embedders/karateclub_embedders/grarep.py b/embiggen/embedders/karateclub_embedders/grarep.py index 052510f0..1c67feeb 100644 --- a/embiggen/embedders/karateclub_embedders/grarep.py +++ b/embiggen/embedders/karateclub_embedders/grarep.py @@ -8,7 +8,7 @@ class GraRepKarateClub(AbstractKarateClubEmbedder): def __init__( self, - embedding_size: int = 128, + embedding_size: int = 100, iteration: int = 10, order: int = 5, random_state: int = 42, @@ -18,7 +18,7 @@ def __init__( Parameters ---------------------- - embedding_size: int = 128 + embedding_size: int = 100 Size of the embedding to use. iteration: int = 10 Number of SVD iterations. Default is 10. diff --git a/embiggen/embedders/karateclub_embedders/hope.py b/embiggen/embedders/karateclub_embedders/hope.py index c3ef948d..ef74dad4 100644 --- a/embiggen/embedders/karateclub_embedders/hope.py +++ b/embiggen/embedders/karateclub_embedders/hope.py @@ -8,7 +8,7 @@ class HOPEKarateClub(AbstractKarateClubEmbedder): def __init__( self, - embedding_size: int = 128, + embedding_size: int = 100, random_state: int = 42, enable_cache: bool = False ): @@ -16,7 +16,7 @@ def __init__( Parameters ---------------------- - embedding_size: int = 128 + embedding_size: int = 100 Size of the embedding to use. random_state: int = 42 Random state to use for the stocastic diff --git a/embiggen/embedders/karateclub_embedders/laplacian_eigenmaps.py b/embiggen/embedders/karateclub_embedders/laplacian_eigenmaps.py index 5156a315..e360ee8d 100644 --- a/embiggen/embedders/karateclub_embedders/laplacian_eigenmaps.py +++ b/embiggen/embedders/karateclub_embedders/laplacian_eigenmaps.py @@ -8,7 +8,7 @@ class LaplacianEigenmapsKarateClub(AbstractKarateClubEmbedder): def __init__( self, - embedding_size: int = 128, + embedding_size: int = 100, random_state: int = 42, enable_cache: bool = False ): @@ -16,7 +16,7 @@ def __init__( Parameters ---------------------- - embedding_size: int = 128 + embedding_size: int = 100 Size of the embedding to use. random_state: int = 42 Random state to use for the stocastic diff --git a/embiggen/embedders/karateclub_embedders/netmf.py b/embiggen/embedders/karateclub_embedders/netmf.py index adb714fe..86ae41c8 100644 --- a/embiggen/embedders/karateclub_embedders/netmf.py +++ b/embiggen/embedders/karateclub_embedders/netmf.py @@ -8,7 +8,7 @@ class NetMFKarateClub(AbstractKarateClubEmbedder): def __init__( self, - embedding_size: int = 128, + embedding_size: int = 100, iteration: int = 10, order: int = 2, negative_samples: int = 1, @@ -19,7 +19,7 @@ def __init__( Parameters ---------------------- - embedding_size: int = 128 + embedding_size: int = 100 Size of the embedding to use. iteration: int = 10 Number of SVD iterations. Default is 10. diff --git a/embiggen/embedders/karateclub_embedders/nmfadmm.py b/embiggen/embedders/karateclub_embedders/nmfadmm.py index bf4c6b80..8026e687 100644 --- a/embiggen/embedders/karateclub_embedders/nmfadmm.py +++ b/embiggen/embedders/karateclub_embedders/nmfadmm.py @@ -8,7 +8,7 @@ class NMFADMMKarateClub(AbstractKarateClubEmbedder): def __init__( self, - embedding_size: int = 128, + embedding_size: int = 100, iterations: int = 100, rho: float = 1.0, random_state: int = 42, @@ -18,7 +18,7 @@ def __init__( Parameters ---------------------- - embedding_size: int = 128 + embedding_size: int = 100 Size of the embedding to use. iterations: int = 100 Number of SVD iterationss. Default is 10. diff --git a/embiggen/embedders/karateclub_embedders/node_sketch.py b/embiggen/embedders/karateclub_embedders/node_sketch.py index d5082873..0829eea7 100644 --- a/embiggen/embedders/karateclub_embedders/node_sketch.py +++ b/embiggen/embedders/karateclub_embedders/node_sketch.py @@ -8,7 +8,7 @@ class NodeSketchKarateClub(AbstractKarateClubEmbedder): def __init__( self, - embedding_size: int = 128, + embedding_size: int = 100, iterations: int = 10, decay: float = 0.01, random_state: int = 42, @@ -18,7 +18,7 @@ def __init__( Parameters ---------------------- - embedding_size: int = 128 + embedding_size: int = 100 Size of the embedding to use. iterations: int = 10 Number of SVD iterationss. Default is 10. diff --git a/embiggen/embedders/karateclub_embedders/randne.py b/embiggen/embedders/karateclub_embedders/randne.py index 757b3183..bee120c0 100644 --- a/embiggen/embedders/karateclub_embedders/randne.py +++ b/embiggen/embedders/karateclub_embedders/randne.py @@ -8,7 +8,7 @@ class RandNEKarateClub(AbstractKarateClubEmbedder): def __init__( self, - embedding_size: int = 128, + embedding_size: int = 100, alphas: Union[List[float], Tuple[float]] = (0.5, 0.5), random_state: int = 42, enable_cache: bool = False @@ -17,7 +17,7 @@ def __init__( Parameters ---------------------- - embedding_size: int = 128 + embedding_size: int = 100 Size of the embedding to use. alphas: Union[List[float], Tuple[float]] = (0.5, 0.5) Smoothing weights for adjacency matrix powers. Default is [0.5, 0.5]. diff --git a/embiggen/embedders/karateclub_embedders/role2vec.py b/embiggen/embedders/karateclub_embedders/role2vec.py index 866b8db4..3f6ad6a4 100644 --- a/embiggen/embedders/karateclub_embedders/role2vec.py +++ b/embiggen/embedders/karateclub_embedders/role2vec.py @@ -9,7 +9,7 @@ class Role2VecKarateClub(AbstractKarateClubEmbedder): def __init__( self, - embedding_size: int = 128, + embedding_size: int = 100, walk_number: int = 10, walk_length: int = 80, window_size: int = 5, @@ -26,7 +26,7 @@ def __init__( Parameters ---------------------- - embedding_size: int = 128 + embedding_size: int = 100 Size of the embedding to use. walk_number: int = 10 Number of random walks. Default is 10. diff --git a/embiggen/embedders/karateclub_embedders/skipgram.py b/embiggen/embedders/karateclub_embedders/skipgram.py index 2c275b66..bb9cb317 100644 --- a/embiggen/embedders/karateclub_embedders/skipgram.py +++ b/embiggen/embedders/karateclub_embedders/skipgram.py @@ -9,7 +9,7 @@ class SkipGramKarateClub(AbstractKarateClubEmbedder): def __init__( self, - embedding_size: int = 128, + embedding_size: int = 100, walk_number: int = 10, walk_length: int = 80, window_size: int = 5, @@ -33,7 +33,7 @@ def __init__( Parameters ---------------------- - embedding_size: int = 128 + embedding_size: int = 100 Size of the embedding to use. walk_number: int = 10 Number of random walks. Default is 10. @@ -117,7 +117,7 @@ def _build_model(self) -> Node2Vec: @classmethod def model_name(cls) -> str: """Returns name of the model""" - return "SkipGram" + return "Node2Vec SkipGram" @classmethod def requires_nodes_sorted_by_decreasing_node_degree(cls) -> bool: diff --git a/embiggen/embedders/karateclub_embedders/sociodim.py b/embiggen/embedders/karateclub_embedders/sociodim.py index 97f3c730..4ccac64f 100644 --- a/embiggen/embedders/karateclub_embedders/sociodim.py +++ b/embiggen/embedders/karateclub_embedders/sociodim.py @@ -1,5 +1,4 @@ """Wrapper for SocioDim model provided from the Karate Club package.""" -from typing import Dict, Any from karateclub.node_embedding import SocioDim from embiggen.embedders.karateclub_embedders.abstract_karateclub_embedder import AbstractKarateClubEmbedder @@ -8,7 +7,7 @@ class SocioDimKarateClub(AbstractKarateClubEmbedder): def __init__( self, - embedding_size: int = 128, + embedding_size: int = 100, random_state: int = 42, enable_cache: bool = False ): @@ -16,7 +15,7 @@ def __init__( Parameters ---------------------- - embedding_size: int = 128 + embedding_size: int = 100 Size of the embedding to use. random_state: int = 42 Random state to use for the stocastic diff --git a/embiggen/embedders/karateclub_embedders/walklets.py b/embiggen/embedders/karateclub_embedders/walklets.py index 74381d48..55b681ef 100644 --- a/embiggen/embedders/karateclub_embedders/walklets.py +++ b/embiggen/embedders/karateclub_embedders/walklets.py @@ -9,7 +9,7 @@ class WalkletsSkipGramKarateClub(AbstractKarateClubEmbedder): def __init__( self, - embedding_size: int = 128, + embedding_size: int = 100, walk_number: int = 10, walk_length: int = 80, window_size: int = 5, @@ -23,7 +23,7 @@ def __init__( Parameters ---------------------- - embedding_size: int = 128 + embedding_size: int = 100 Size of the embedding to use. walk_number: int = 10 Number of random walks. Default is 10. @@ -52,16 +52,17 @@ def __init__( self._learning_rate = learning_rate self._min_count = min_count super().__init__( - embedding_size=embedding_size, + embedding_size=embedding_size // self._window_size, enable_cache=enable_cache, random_state=random_state ) def parameters(self) -> Dict[str, Any]: """Returns the parameters used in the model.""" + parameters = super().parameters() + parameters["embedding_size"] = parameters["embedding_size"] self._window_size return dict( - super().parameters(), - + parameters, walk_number=self._walk_number, walk_length=self._walk_length, window_size=self._window_size, @@ -77,7 +78,7 @@ def smoke_test_parameters(cls) -> Dict[str, Any]: AbstractKarateClubEmbedder.smoke_test_parameters(), walk_number=1, walk_length=8, - window_size=2, + window_size=1, epochs=1, ) diff --git a/embiggen/embedders/non_existent_embedders/__init__.py b/embiggen/embedders/non_existent_embedders/__init__.py new file mode 100644 index 00000000..f122946f --- /dev/null +++ b/embiggen/embedders/non_existent_embedders/__init__.py @@ -0,0 +1,8 @@ +"""Submodule to test whether the metaprogramming is working properly.""" +from embiggen.utils.abstract_models import build_init, AbstractEmbeddingModel + +build_init( + module_library_names="non_existent_module", + formatted_library_name="Non Existent Module", + expected_parent_class=AbstractEmbeddingModel +) diff --git a/embiggen/embedders/non_existent_embedders/non_existent_model.py b/embiggen/embedders/non_existent_embedders/non_existent_model.py new file mode 100644 index 00000000..46ac040d --- /dev/null +++ b/embiggen/embedders/non_existent_embedders/non_existent_model.py @@ -0,0 +1,14 @@ +"""A model strictly for validating the meta programming.""" +from embiggen.utils.abstract_models import AbstractEmbeddingModel + + +class NonExistentModel(AbstractEmbeddingModel): + """A model strictly for validating the meta programming.""" + + @classmethod + def model_name(cls) -> str: + """Returns name of the model.""" + return "Non Existent Model" + +import non_existent_module + \ No newline at end of file diff --git a/embiggen/embedders/pykeen_embedders/auto_sf.py b/embiggen/embedders/pykeen_embedders/auto_sf.py index a0afe4f4..eabe63f8 100644 --- a/embiggen/embedders/pykeen_embedders/auto_sf.py +++ b/embiggen/embedders/pykeen_embedders/auto_sf.py @@ -16,6 +16,7 @@ def __init__( batch_size: int = 210, training_loop: Union[str, Type[TrainingLoop] ] = "Stochastic Local Closed World Assumption", + verbose: bool = False, random_state: int = 42, enable_cache: bool = False ): @@ -46,6 +47,8 @@ def __init__( Can either be: - Stochastic Local Closed World Assumption - Local Closed World Assumption + verbose: bool = False + Whether to show loading bars. random_state: int = 42 Random seed to use while training the model enable_cache: bool = False @@ -58,17 +61,18 @@ def __init__( epochs=epochs, batch_size=batch_size, training_loop=training_loop, + verbose=verbose, random_state=random_state, enable_cache=enable_cache ) def parameters(self) -> Dict[str, Any]: - return { + return dict( super().parameters(), dict( num_components=self._num_components ) - } + ) @classmethod def model_name(cls) -> str: @@ -89,5 +93,6 @@ def _build_model( return AutoSF( triples_factory=triples_factory, embedding_dim=self._embedding_size, - num_components=self._num_components + num_components=self._num_components, + random_seed=self._random_state ) diff --git a/embiggen/embedders/pykeen_embedders/boxe.py b/embiggen/embedders/pykeen_embedders/boxe.py index 3ff64227..babee384 100644 --- a/embiggen/embedders/pykeen_embedders/boxe.py +++ b/embiggen/embedders/pykeen_embedders/boxe.py @@ -18,6 +18,7 @@ def __init__( batch_size: int = 210, training_loop: Union[str, Type[TrainingLoop] ] = "Stochastic Local Closed World Assumption", + verbose: bool = False, random_state: int = 42, enable_cache: bool = False ): @@ -54,6 +55,8 @@ def __init__( Can either be: - Stochastic Local Closed World Assumption - Local Closed World Assumption + verbose: bool = False + Whether to show loading bars. random_state: int = 42 Random seed to use while training the model enable_cache: bool = False @@ -68,19 +71,20 @@ def __init__( epochs=epochs, batch_size=batch_size, training_loop=training_loop, + verbose=verbose, random_state=random_state, enable_cache=enable_cache ) def parameters(self) -> Dict[str, Any]: - return { + return dict( super().parameters(), dict( tanh_map=self._tanh_map, p=self._p, power_norm=self._power_norm, ) - } + ) @classmethod def model_name(cls) -> str: @@ -104,4 +108,5 @@ def _build_model( tanh_map=self._tanh_map, p=self._p, power_norm=self._power_norm, + random_seed=self._random_state ) diff --git a/embiggen/embedders/pykeen_embedders/complex.py b/embiggen/embedders/pykeen_embedders/complex.py index 47a87819..5258ff2e 100644 --- a/embiggen/embedders/pykeen_embedders/complex.py +++ b/embiggen/embedders/pykeen_embedders/complex.py @@ -8,56 +8,6 @@ class ComplExPyKeen(EntityRelationEmbeddingModelPyKeen): - def __init__( - self, - embedding_size: int = 256, - epochs: int = 100, - batch_size: int = 210, - training_loop: Union[str, Type[TrainingLoop] - ] = "Stochastic Local Closed World Assumption", - random_state: int = 42, - enable_cache: bool = False - ): - """Create new PyKeen ComplEx model. - - Details - ------------------------- - This is a wrapper of the ComplEx implementation from the - PyKeen library. Please refer to the PyKeen library documentation - for details and posssible errors regarding this model. - - Parameters - ------------------------- - embedding_size: int = 256 - The dimension of the embedding to compute. - epochs: int = 100 - The number of epochs to use to train the model for. - batch_size: int = 210 - Size of the training batch. - device: str = "auto" - The devide to use to train the model. - Can either be cpu or cuda. - training_loop: Union[str, Type[TrainingLoop] - ] = "Stochastic Local Closed World Assumption" - The training loop to use to train the model. - Can either be: - - Stochastic Local Closed World Assumption - - Local Closed World Assumption - random_state: int = 42 - Random seed to use while training the model - enable_cache: bool = False - Whether to enable the cache, that is to - store the computed embedding. - """ - super().__init__( - embedding_size=embedding_size, - epochs=epochs, - batch_size=batch_size, - training_loop=training_loop, - random_state=random_state, - enable_cache=enable_cache - ) - @classmethod def model_name(cls) -> str: """Return name of the model.""" @@ -77,4 +27,5 @@ def _build_model( return ComplEx( triples_factory=triples_factory, embedding_dim=self._embedding_size, + random_seed=self._random_state ) diff --git a/embiggen/embedders/pykeen_embedders/conve.py b/embiggen/embedders/pykeen_embedders/conve.py index 1cf0884a..e6f389d7 100644 --- a/embiggen/embedders/pykeen_embedders/conve.py +++ b/embiggen/embedders/pykeen_embedders/conve.py @@ -25,6 +25,7 @@ def __init__( batch_size: int = 210, training_loop: Union[str, Type[TrainingLoop] ] = "Stochastic Local Closed World Assumption", + verbose: bool = False, random_state: int = 42, enable_cache: bool = False ): @@ -76,6 +77,8 @@ def __init__( Can either be: - Stochastic Local Closed World Assumption - Local Closed World Assumption + verbose: bool = False + Whether to show loading bars. random_state: int = 42 Random seed to use while training the model enable_cache: bool = False @@ -97,12 +100,13 @@ def __init__( epochs=epochs, batch_size=batch_size, training_loop=training_loop, + verbose=verbose, random_state=random_state, enable_cache=enable_cache ) def parameters(self) -> Dict[str, Any]: - return { + return dict( super().parameters(), dict( input_channels=self._input_channels, @@ -116,7 +120,7 @@ def parameters(self) -> Dict[str, Any]: feature_map_dropout=self._feature_map_dropout, apply_batch_normalization=self._apply_batch_normalization, ) - } + ) @classmethod def model_name(cls) -> str: @@ -147,4 +151,5 @@ def _build_model( output_dropout=self._output_dropout, feature_map_dropout=self._feature_map_dropout, apply_batch_normalization=self._apply_batch_normalization, + random_seed=self._random_state ) diff --git a/embiggen/embedders/pykeen_embedders/cp.py b/embiggen/embedders/pykeen_embedders/cp.py index 7933924b..885eb52a 100644 --- a/embiggen/embedders/pykeen_embedders/cp.py +++ b/embiggen/embedders/pykeen_embedders/cp.py @@ -16,6 +16,7 @@ def __init__( batch_size: int = 210, training_loop: Union[str, Type[TrainingLoop] ] = "Stochastic Local Closed World Assumption", + verbose: bool = False, random_state: int = 42, enable_cache: bool = False ): @@ -46,6 +47,8 @@ def __init__( Can either be: - Stochastic Local Closed World Assumption - Local Closed World Assumption + verbose: bool = False + Whether to show loading bars. random_state: int = 42 Random seed to use while training the model enable_cache: bool = False @@ -58,17 +61,18 @@ def __init__( epochs=epochs, batch_size=batch_size, training_loop=training_loop, + verbose=verbose, random_state=random_state, enable_cache=enable_cache ) def parameters(self) -> Dict[str, Any]: - return { + return dict( super().parameters(), dict( rank=self._rank, ) - } + ) @classmethod def model_name(cls) -> str: @@ -90,4 +94,5 @@ def _build_model( triples_factory=triples_factory, embedding_dim=self._embedding_size, rank=self._rank, + random_seed=self._random_state ) diff --git a/embiggen/embedders/pykeen_embedders/crosse.py b/embiggen/embedders/pykeen_embedders/crosse.py index 9cc5f618..8aad0b05 100644 --- a/embiggen/embedders/pykeen_embedders/crosse.py +++ b/embiggen/embedders/pykeen_embedders/crosse.py @@ -16,6 +16,7 @@ def __init__( batch_size: int = 210, training_loop: Union[str, Type[TrainingLoop] ] = "Stochastic Local Closed World Assumption", + verbose: bool = False, random_state: int = 42, enable_cache: bool = False ): @@ -46,6 +47,8 @@ def __init__( Can either be: - Stochastic Local Closed World Assumption - Local Closed World Assumption + verbose: bool = False + Whether to show loading bars. random_state: int = 42 Random seed to use while training the model enable_cache: bool = False @@ -58,17 +61,18 @@ def __init__( epochs=epochs, batch_size=batch_size, training_loop=training_loop, + verbose=verbose, random_state=random_state, enable_cache=enable_cache ) def parameters(self) -> Dict[str, Any]: - return { + return dict( super().parameters(), dict( combination_dropout=self._combination_dropout, ) - } + ) @classmethod def model_name(cls) -> str: @@ -90,4 +94,5 @@ def _build_model( triples_factory=triples_factory, embedding_dim=self._embedding_size, combination_dropout=self._combination_dropout, + random_seed=self._random_state ) diff --git a/embiggen/embedders/pykeen_embedders/distma.py b/embiggen/embedders/pykeen_embedders/distma.py index 6c82c557..cfa6d0db 100644 --- a/embiggen/embedders/pykeen_embedders/distma.py +++ b/embiggen/embedders/pykeen_embedders/distma.py @@ -8,56 +8,6 @@ class DistMAPyKeen(EntityRelationEmbeddingModelPyKeen): - def __init__( - self, - embedding_size: int = 256, - epochs: int = 100, - batch_size: int = 210, - training_loop: Union[str, Type[TrainingLoop] - ] = "Stochastic Local Closed World Assumption", - random_state: int = 42, - enable_cache: bool = False - ): - """Create new PyKeen DistMA model. - - Details - ------------------------- - This is a wrapper of the DistMA implementation from the - PyKeen library. Please refer to the PyKeen library documentation - for details and posssible errors regarding this model. - - Parameters - ------------------------- - embedding_size: int = 256 - The dimension of the embedding to compute. - epochs: int = 100 - The number of epochs to use to train the model for. - batch_size: int = 210 - Size of the training batch. - device: str = "auto" - The devide to use to train the model. - Can either be cpu or cuda. - training_loop: Union[str, Type[TrainingLoop] - ] = "Stochastic Local Closed World Assumption" - The training loop to use to train the model. - Can either be: - - Stochastic Local Closed World Assumption - - Local Closed World Assumption - random_state: int = 42 - Random seed to use while training the model - enable_cache: bool = False - Whether to enable the cache, that is to - store the computed embedding. - """ - super().__init__( - embedding_size=embedding_size, - epochs=epochs, - batch_size=batch_size, - training_loop=training_loop, - random_state=random_state, - enable_cache=enable_cache - ) - @classmethod def model_name(cls) -> str: """Return name of the model.""" @@ -77,4 +27,5 @@ def _build_model( return DistMA( triples_factory=triples_factory, embedding_dim=self._embedding_size, + random_seed=self._random_state ) diff --git a/embiggen/embedders/pykeen_embedders/distmult.py b/embiggen/embedders/pykeen_embedders/distmult.py index a7770b59..c628b683 100644 --- a/embiggen/embedders/pykeen_embedders/distmult.py +++ b/embiggen/embedders/pykeen_embedders/distmult.py @@ -8,56 +8,6 @@ class DistMultPyKeen(EntityRelationEmbeddingModelPyKeen): - def __init__( - self, - embedding_size: int = 256, - epochs: int = 100, - batch_size: int = 210, - training_loop: Union[str, Type[TrainingLoop] - ] = "Stochastic Local Closed World Assumption", - random_state: int = 42, - enable_cache: bool = False - ): - """Create new PyKeen DistMult model. - - Details - ------------------------- - This is a wrapper of the DistMult implementation from the - PyKeen library. Please refer to the PyKeen library documentation - for details and posssible errors regarding this model. - - Parameters - ------------------------- - embedding_size: int = 256 - The dimension of the embedding to compute. - epochs: int = 100 - The number of epochs to use to train the model for. - batch_size: int = 210 - Size of the training batch. - device: str = "auto" - The devide to use to train the model. - Can either be cpu or cuda. - training_loop: Union[str, Type[TrainingLoop] - ] = "Stochastic Local Closed World Assumption" - The training loop to use to train the model. - Can either be: - - Stochastic Local Closed World Assumption - - Local Closed World Assumption - random_state: int = 42 - Random seed to use while training the model - enable_cache: bool = False - Whether to enable the cache, that is to - store the computed embedding. - """ - super().__init__( - embedding_size=embedding_size, - epochs=epochs, - batch_size=batch_size, - training_loop=training_loop, - random_state=random_state, - enable_cache=enable_cache - ) - @classmethod def model_name(cls) -> str: """Return name of the model.""" @@ -77,4 +27,5 @@ def _build_model( return DistMult( triples_factory=triples_factory, embedding_dim=self._embedding_size, + random_seed=self._random_state ) diff --git a/embiggen/embedders/pykeen_embedders/ermlp.py b/embiggen/embedders/pykeen_embedders/ermlp.py index fe54ab47..f0788f15 100644 --- a/embiggen/embedders/pykeen_embedders/ermlp.py +++ b/embiggen/embedders/pykeen_embedders/ermlp.py @@ -16,6 +16,7 @@ def __init__( batch_size: int = 210, training_loop: Union[str, Type[TrainingLoop] ] = "Stochastic Local Closed World Assumption", + verbose: bool = False, random_state: int = 42, enable_cache: bool = False ): @@ -46,6 +47,8 @@ def __init__( Can either be: - Stochastic Local Closed World Assumption - Local Closed World Assumption + verbose: bool = False + Whether to show loading bars. random_state: int = 42 Random seed to use while training the model enable_cache: bool = False @@ -58,6 +61,7 @@ def __init__( epochs=epochs, batch_size=batch_size, training_loop=training_loop, + verbose=verbose, random_state=random_state, enable_cache=enable_cache ) @@ -71,12 +75,12 @@ def smoke_test_parameters(cls) -> Dict[str, Any]: ) def parameters(self) -> Dict[str, Any]: - return { + return dict( super().parameters(), dict( hidden_dim=self._hidden_dim ) - } + ) @classmethod def model_name(cls) -> str: @@ -97,5 +101,6 @@ def _build_model( return ERMLP( triples_factory=triples_factory, embedding_dim=self._embedding_size, - hidden_dim=self._hidden_dim + hidden_dim=self._hidden_dim, + random_seed=self._random_state ) diff --git a/embiggen/embedders/pykeen_embedders/ermlpe.py b/embiggen/embedders/pykeen_embedders/ermlpe.py index 2bfb8b31..638bc0c4 100644 --- a/embiggen/embedders/pykeen_embedders/ermlpe.py +++ b/embiggen/embedders/pykeen_embedders/ermlpe.py @@ -16,6 +16,7 @@ def __init__( batch_size: int = 210, training_loop: Union[str, Type[TrainingLoop] ] = "Stochastic Local Closed World Assumption", + verbose: bool = False, random_state: int = 42, enable_cache: bool = False ): @@ -46,6 +47,8 @@ def __init__( Can either be: - Stochastic Local Closed World Assumption - Local Closed World Assumption + verbose: bool = False + Whether to show loading bars. random_state: int = 42 Random seed to use while training the model enable_cache: bool = False @@ -58,6 +61,7 @@ def __init__( epochs=epochs, batch_size=batch_size, training_loop=training_loop, + verbose=verbose, random_state=random_state, enable_cache=enable_cache ) @@ -71,12 +75,12 @@ def smoke_test_parameters(cls) -> Dict[str, Any]: ) def parameters(self) -> Dict[str, Any]: - return { + return dict( super().parameters(), dict( hidden_dim=self._hidden_dim ) - } + ) @classmethod def model_name(cls) -> str: @@ -97,5 +101,6 @@ def _build_model( return ERMLPE( triples_factory=triples_factory, embedding_dim=self._embedding_size, - hidden_dim=self._hidden_dim + hidden_dim=self._hidden_dim, + random_seed=self._random_state ) diff --git a/embiggen/embedders/pykeen_embedders/hole.py b/embiggen/embedders/pykeen_embedders/hole.py index 3d38369c..901ac5fa 100644 --- a/embiggen/embedders/pykeen_embedders/hole.py +++ b/embiggen/embedders/pykeen_embedders/hole.py @@ -8,56 +8,6 @@ class HolEPyKeen(EntityRelationEmbeddingModelPyKeen): - def __init__( - self, - embedding_size: int = 100, - epochs: int = 100, - batch_size: int = 210, - training_loop: Union[str, Type[TrainingLoop] - ] = "Stochastic Local Closed World Assumption", - random_state: int = 42, - enable_cache: bool = False - ): - """Create new PyKeen HolE model. - - Details - ------------------------- - This is a wrapper of the HolE implementation from the - PyKeen library. Please refer to the PyKeen library documentation - for details and posssible errors regarding this model. - - Parameters - ------------------------- - embedding_size: int = 100 - The dimension of the embedding to compute. - epochs: int = 100 - The number of epochs to use to train the model for. - batch_size: int = 210 - Size of the training batch. - device: str = "auto" - The devide to use to train the model. - Can either be cpu or cuda. - training_loop: Union[str, Type[TrainingLoop] - ] = "Stochastic Local Closed World Assumption" - The training loop to use to train the model. - Can either be: - - Stochastic Local Closed World Assumption - - Local Closed World Assumption - random_state: int = 42 - Random seed to use while training the model - enable_cache: bool = False - Whether to enable the cache, that is to - store the computed embedding. - """ - super().__init__( - embedding_size=embedding_size, - epochs=epochs, - batch_size=batch_size, - training_loop=training_loop, - random_state=random_state, - enable_cache=enable_cache - ) - @classmethod def model_name(cls) -> str: """Return name of the model.""" @@ -77,4 +27,5 @@ def _build_model( return HolE( triples_factory=triples_factory, embedding_dim=self._embedding_size, + random_seed=self._random_state ) diff --git a/embiggen/embedders/pykeen_embedders/pairre.py b/embiggen/embedders/pykeen_embedders/pairre.py index c5c7cf48..2002ba30 100644 --- a/embiggen/embedders/pykeen_embedders/pairre.py +++ b/embiggen/embedders/pykeen_embedders/pairre.py @@ -17,6 +17,7 @@ def __init__( batch_size: int = 210, training_loop: Union[str, Type[TrainingLoop] ] = "Stochastic Local Closed World Assumption", + verbose: bool = False, random_state: int = 42, enable_cache: bool = False ): @@ -49,6 +50,8 @@ def __init__( Can either be: - Stochastic Local Closed World Assumption - Local Closed World Assumption + verbose: bool = False + Whether to show loading bars. random_state: int = 42 Random seed to use while training the model enable_cache: bool = False @@ -62,18 +65,19 @@ def __init__( epochs=epochs, batch_size=batch_size, training_loop=training_loop, + verbose=verbose, random_state=random_state, enable_cache=enable_cache ) def parameters(self) -> Dict[str, Any]: - return { + return dict( super().parameters(), dict( p=self._p, power_norm=self._power_norm, ) - } + ) @classmethod def model_name(cls) -> str: @@ -96,4 +100,5 @@ def _build_model( embedding_dim=self._embedding_size, p=self._p, power_norm=self._power_norm, + random_seed=self._random_state ) diff --git a/embiggen/embedders/pykeen_embedders/proje.py b/embiggen/embedders/pykeen_embedders/proje.py index 575c74f1..e677551f 100644 --- a/embiggen/embedders/pykeen_embedders/proje.py +++ b/embiggen/embedders/pykeen_embedders/proje.py @@ -8,56 +8,6 @@ class ProjEPyKeen(EntityRelationEmbeddingModelPyKeen): - def __init__( - self, - embedding_size: int = 50, - epochs: int = 100, - batch_size: int = 210, - training_loop: Union[str, Type[TrainingLoop] - ] = "Stochastic Local Closed World Assumption", - random_state: int = 42, - enable_cache: bool = False - ): - """Create new PyKeen ProjE model. - - Details - ------------------------- - This is a wrapper of the ProjE implementation from the - PyKeen library. Please refer to the PyKeen library documentation - for details and posssible errors regarding this model. - - Parameters - ------------------------- - embedding_size: int = 50 - The dimension of the embedding to compute. - epochs: int = 100 - The number of epochs to use to train the model for. - batch_size: int = 210 - Size of the training batch. - device: str = "auto" - The devide to use to train the model. - Can either be cpu or cuda. - training_loop: Union[str, Type[TrainingLoop] - ] = "Stochastic Local Closed World Assumption" - The training loop to use to train the model. - Can either be: - - Stochastic Local Closed World Assumption - - Local Closed World Assumption - random_state: int = 42 - Random seed to use while training the model - enable_cache: bool = False - Whether to enable the cache, that is to - store the computed embedding. - """ - super().__init__( - embedding_size=embedding_size, - epochs=epochs, - batch_size=batch_size, - training_loop=training_loop, - random_state=random_state, - enable_cache=enable_cache - ) - @classmethod def model_name(cls) -> str: """Return name of the model.""" @@ -77,4 +27,5 @@ def _build_model( return ProjE( triples_factory=triples_factory, embedding_dim=self._embedding_size, + random_seed=self._random_state ) diff --git a/embiggen/embedders/pykeen_embedders/pykeen_embedder.py b/embiggen/embedders/pykeen_embedders/pykeen_embedder.py index bf761719..ea975a50 100644 --- a/embiggen/embedders/pykeen_embedders/pykeen_embedder.py +++ b/embiggen/embedders/pykeen_embedders/pykeen_embedder.py @@ -1,5 +1,5 @@ """Abstract Torch/PyKeen Model wrapper for embedding models.""" -from typing import Dict, List, Sequence, Union, Optional, Tuple, Any, Type +from typing import Dict, Union, Tuple, Any, Type import numpy as np import pandas as pd @@ -13,7 +13,6 @@ from pykeen.models import Model from pykeen.triples import CoreTriplesFactory from pykeen.training import SLCWATrainingLoop, LCWATrainingLoop, TrainingLoop -from torch.optim import Optimizer @abstract_class @@ -33,6 +32,7 @@ def __init__( device: str = "auto", training_loop: Union[str, Type[TrainingLoop] ] = "Stochastic Local Closed World Assumption", + verbose: bool = False, random_state: int = 42, enable_cache: bool = False ): @@ -55,6 +55,8 @@ def __init__( Can either be: - Stochastic Local Closed World Assumption - Local Closed World Assumption + verbose: bool = False + Whether to show the loading bar. random_state: int = 42 Random seed to use while training the model enable_cache: bool = False @@ -83,6 +85,7 @@ def __init__( self._training_loop = training_loop self._epochs = epochs + self._verbose = verbose self._batch_size = batch_size self._device = validate_torch_device(device) @@ -101,13 +104,13 @@ def smoke_test_parameters(cls) -> Dict[str, Any]: ) def parameters(self) -> Dict[str, Any]: - return { + return dict( super().parameters(), dict( epochs=self._epochs, batch_size=self._batch_size, ) - } + ) @classmethod def library_name(cls) -> str: @@ -168,7 +171,6 @@ def _fit_transform( self, graph: Graph, return_dataframe: bool = True, - verbose: bool = True ) -> Union[np.ndarray, pd.DataFrame, Dict[str, np.ndarray], Dict[str, pd.DataFrame]]: """Return node embedding""" @@ -210,10 +212,12 @@ def _fit_transform( triples_factory=triples_factory, num_epochs=self._epochs, batch_size=batch_size, - use_tqdm=True, - use_tqdm_batch=True, + use_tqdm=self._verbose, + use_tqdm_batch=self._verbose, tqdm_kwargs=dict( - disable=not verbose + disable=not self._verbose, + dynamic_ncols=True, + leave=False ) ) diff --git a/embiggen/embedders/pykeen_embedders/quate.py b/embiggen/embedders/pykeen_embedders/quate.py index 98d9ba51..ca9caff0 100644 --- a/embiggen/embedders/pykeen_embedders/quate.py +++ b/embiggen/embedders/pykeen_embedders/quate.py @@ -8,56 +8,6 @@ class QuatEPyKeen(EntityRelationEmbeddingModelPyKeen): - def __init__( - self, - embedding_size: int = 100, - epochs: int = 100, - batch_size: int = 210, - training_loop: Union[str, Type[TrainingLoop] - ] = "Stochastic Local Closed World Assumption", - random_state: int = 42, - enable_cache: bool = False - ): - """Create new PyKeen QuatE model. - - Details - ------------------------- - This is a wrapper of the QuatE implementation from the - PyKeen library. Please refer to the PyKeen library documentation - for details and posssible errors regarding this model. - - Parameters - ------------------------- - embedding_size: int = 100 - The dimension of the embedding to compute. - epochs: int = 100 - The number of epochs to use to train the model for. - batch_size: int = 210 - Size of the training batch. - device: str = "auto" - The devide to use to train the model. - Can either be cpu or cuda. - training_loop: Union[str, Type[TrainingLoop] - ] = "Stochastic Local Closed World Assumption" - The training loop to use to train the model. - Can either be: - - Stochastic Local Closed World Assumption - - Local Closed World Assumption - random_state: int = 42 - Random seed to use while training the model - enable_cache: bool = False - Whether to enable the cache, that is to - store the computed embedding. - """ - super().__init__( - embedding_size=embedding_size, - epochs=epochs, - batch_size=batch_size, - training_loop=training_loop, - random_state=random_state, - enable_cache=enable_cache - ) - @classmethod def model_name(cls) -> str: """Return name of the model.""" @@ -77,4 +27,5 @@ def _build_model( return QuatE( triples_factory=triples_factory, embedding_dim=self._embedding_size, + random_seed=self._random_state ) diff --git a/embiggen/embedders/pykeen_embedders/rescal.py b/embiggen/embedders/pykeen_embedders/rescal.py index 8376f16d..40071b24 100644 --- a/embiggen/embedders/pykeen_embedders/rescal.py +++ b/embiggen/embedders/pykeen_embedders/rescal.py @@ -8,56 +8,6 @@ class RESCALPyKeen(EntityRelationEmbeddingModelPyKeen): - def __init__( - self, - embedding_size: int = 100, - epochs: int = 100, - batch_size: int = 210, - training_loop: Union[str, Type[TrainingLoop] - ] = "Stochastic Local Closed World Assumption", - random_state: int = 42, - enable_cache: bool = False - ): - """Create new PyKeen RESCAL model. - - Details - ------------------------- - This is a wrapper of the RESCAL implementation from the - PyKeen library. Please refer to the PyKeen library documentation - for details and posssible errors regarding this model. - - Parameters - ------------------------- - embedding_size: int = 100 - The dimension of the embedding to compute. - epochs: int = 100 - The number of epochs to use to train the model for. - batch_size: int = 210 - Size of the training batch. - device: str = "auto" - The devide to use to train the model. - Can either be cpu or cuda. - training_loop: Union[str, Type[TrainingLoop] - ] = "Stochastic Local Closed World Assumption" - The training loop to use to train the model. - Can either be: - - Stochastic Local Closed World Assumption - - Local Closed World Assumption - random_state: int = 42 - Random seed to use while training the model - enable_cache: bool = False - Whether to enable the cache, that is to - store the computed embedding. - """ - super().__init__( - embedding_size=embedding_size, - epochs=epochs, - batch_size=batch_size, - training_loop=training_loop, - random_state=random_state, - enable_cache=enable_cache - ) - @classmethod def model_name(cls) -> str: """Return name of the model.""" @@ -77,4 +27,5 @@ def _build_model( return RESCAL( triples_factory=triples_factory, embedding_dim=self._embedding_size, + random_seed=self._random_state ) diff --git a/embiggen/embedders/pykeen_embedders/rotate.py b/embiggen/embedders/pykeen_embedders/rotate.py index 2e7c792a..0a46929c 100644 --- a/embiggen/embedders/pykeen_embedders/rotate.py +++ b/embiggen/embedders/pykeen_embedders/rotate.py @@ -8,56 +8,6 @@ class RotatEPyKeen(EntityRelationEmbeddingModelPyKeen): - def __init__( - self, - embedding_size: int = 200, - epochs: int = 100, - batch_size: int = 210, - training_loop: Union[str, Type[TrainingLoop] - ] = "Stochastic Local Closed World Assumption", - random_state: int = 42, - enable_cache: bool = False - ): - """Create new PyKeen RotatE model. - - Details - ------------------------- - This is a wrapper of the RotatE implementation from the - PyKeen library. Please refer to the PyKeen library documentation - for details and posssible errors regarding this model. - - Parameters - ------------------------- - embedding_size: int = 200 - The dimension of the embedding to compute. - epochs: int = 100 - The number of epochs to use to train the model for. - batch_size: int = 210 - Size of the training batch. - device: str = "auto" - The devide to use to train the model. - Can either be cpu or cuda. - training_loop: Union[str, Type[TrainingLoop] - ] = "Stochastic Local Closed World Assumption" - The training loop to use to train the model. - Can either be: - - Stochastic Local Closed World Assumption - - Local Closed World Assumption - random_state: int = 42 - Random seed to use while training the model - enable_cache: bool = False - Whether to enable the cache, that is to - store the computed embedding. - """ - super().__init__( - embedding_size=embedding_size, - epochs=epochs, - batch_size=batch_size, - training_loop=training_loop, - random_state=random_state, - enable_cache=enable_cache - ) - @classmethod def model_name(cls) -> str: """Return name of the model.""" @@ -77,4 +27,5 @@ def _build_model( return RotatE( triples_factory=triples_factory, embedding_dim=self._embedding_size, + random_seed=self._random_state ) diff --git a/embiggen/embedders/pykeen_embedders/toruse.py b/embiggen/embedders/pykeen_embedders/toruse.py index d3a178a4..c0d101b4 100644 --- a/embiggen/embedders/pykeen_embedders/toruse.py +++ b/embiggen/embedders/pykeen_embedders/toruse.py @@ -17,6 +17,7 @@ def __init__( batch_size: int = 210, training_loop: Union[str, Type[TrainingLoop] ] = "Stochastic Local Closed World Assumption", + verbose: bool = False, random_state: int = 42, enable_cache: bool = False ): @@ -49,6 +50,8 @@ def __init__( Can either be: - Stochastic Local Closed World Assumption - Local Closed World Assumption + verbose: bool = False + Whether to show loading bars. random_state: int = 42 Random seed to use while training the model enable_cache: bool = False @@ -62,18 +65,19 @@ def __init__( epochs=epochs, batch_size=batch_size, training_loop=training_loop, + verbose=verbose, random_state=random_state, enable_cache=enable_cache ) def parameters(self) -> Dict[str, Any]: - return { + return dict( super().parameters(), dict( p=self._p, power_norm=self._power_norm, ) - } + ) @classmethod def model_name(cls) -> str: @@ -96,4 +100,5 @@ def _build_model( embedding_dim=self._embedding_size, p=self._p, power_norm=self._power_norm, + random_seed=self._random_state ) diff --git a/embiggen/embedders/pykeen_embedders/transd.py b/embiggen/embedders/pykeen_embedders/transd.py index b9fd8974..1b8c4c39 100644 --- a/embiggen/embedders/pykeen_embedders/transd.py +++ b/embiggen/embedders/pykeen_embedders/transd.py @@ -16,6 +16,7 @@ def __init__( batch_size: int = 210, training_loop: Union[str, Type[TrainingLoop] ] = "Stochastic Local Closed World Assumption", + verbose: bool = False, random_state: int = 42, enable_cache: bool = False ): @@ -46,6 +47,8 @@ def __init__( Can either be: - Stochastic Local Closed World Assumption - Local Closed World Assumption + verbose: bool = False + Whether to show loading bars. random_state: int = 42 Random seed to use while training the model enable_cache: bool = False @@ -58,6 +61,7 @@ def __init__( epochs=epochs, batch_size=batch_size, training_loop=training_loop, + verbose=verbose, random_state=random_state, enable_cache=enable_cache ) @@ -71,12 +75,12 @@ def smoke_test_parameters(cls) -> Dict[str, Any]: ) def parameters(self) -> Dict[str, Any]: - return { + return dict( super().parameters(), dict( relation_dim=self._relation_dim ) - } + ) @classmethod def model_name(cls) -> str: diff --git a/embiggen/embedders/pykeen_embedders/transe.py b/embiggen/embedders/pykeen_embedders/transe.py index f209b01c..3435e5bc 100644 --- a/embiggen/embedders/pykeen_embedders/transe.py +++ b/embiggen/embedders/pykeen_embedders/transe.py @@ -16,6 +16,7 @@ def __init__( batch_size: int = 210, training_loop: Union[str, Type[TrainingLoop] ] = "Stochastic Local Closed World Assumption", + verbose: bool = False, random_state: int = 42, enable_cache: bool = False ): @@ -46,6 +47,8 @@ def __init__( Can either be: - Stochastic Local Closed World Assumption - Local Closed World Assumption + verbose: bool = False + Whether to show loading bars. random_state: int = 42 Random seed to use while training the model enable_cache: bool = False @@ -58,6 +61,7 @@ def __init__( epochs=epochs, batch_size=batch_size, training_loop=training_loop, + verbose=verbose, random_state=random_state, enable_cache=enable_cache ) @@ -71,12 +75,12 @@ def smoke_test_parameters(cls) -> Dict[str, Any]: ) def parameters(self) -> Dict[str, Any]: - return { + return dict( super().parameters(), dict( scoring_fct_norm=self._scoring_fct_norm ) - } + ) @classmethod def model_name(cls) -> str: @@ -97,5 +101,6 @@ def _build_model( return TransE( triples_factory=triples_factory, embedding_dim=self._embedding_size, - scoring_fct_norm=self._scoring_fct_norm + scoring_fct_norm=self._scoring_fct_norm, + random_seed=self._random_state ) diff --git a/embiggen/embedders/pykeen_embedders/transf.py b/embiggen/embedders/pykeen_embedders/transf.py index c29838bd..f6d4cf1f 100644 --- a/embiggen/embedders/pykeen_embedders/transf.py +++ b/embiggen/embedders/pykeen_embedders/transf.py @@ -8,56 +8,6 @@ class TransFPyKeen(EntityRelationEmbeddingModelPyKeen): - def __init__( - self, - embedding_size: int = 100, - epochs: int = 100, - batch_size: int = 210, - training_loop: Union[str, Type[TrainingLoop] - ] = "Stochastic Local Closed World Assumption", - random_state: int = 42, - enable_cache: bool = False - ): - """Create new PyKeen TransF model. - - Details - ------------------------- - This is a wrapper of the TransF implementation from the - PyKeen library. Please refer to the PyKeen library documentation - for details and posssible errors regarding this model. - - Parameters - ------------------------- - embedding_size: int = 100 - The dimension of the embedding to compute. - epochs: int = 100 - The number of epochs to use to train the model for. - batch_size: int = 210 - Size of the training batch. - device: str = "auto" - The devide to use to train the model. - Can either be cpu or cuda. - training_loop: Union[str, Type[TrainingLoop] - ] = "Stochastic Local Closed World Assumption" - The training loop to use to train the model. - Can either be: - - Stochastic Local Closed World Assumption - - Local Closed World Assumption - random_state: int = 42 - Random seed to use while training the model - enable_cache: bool = False - Whether to enable the cache, that is to - store the computed embedding. - """ - super().__init__( - embedding_size=embedding_size, - epochs=epochs, - batch_size=batch_size, - training_loop=training_loop, - random_state=random_state, - enable_cache=enable_cache - ) - @classmethod def model_name(cls) -> str: """Return name of the model.""" @@ -77,4 +27,5 @@ def _build_model( return TransF( triples_factory=triples_factory, embedding_dim=self._embedding_size, + random_seed=self._random_state ) diff --git a/embiggen/embedders/pykeen_embedders/transh.py b/embiggen/embedders/pykeen_embedders/transh.py index f2849054..1fddcaf6 100644 --- a/embiggen/embedders/pykeen_embedders/transh.py +++ b/embiggen/embedders/pykeen_embedders/transh.py @@ -16,6 +16,7 @@ def __init__( batch_size: int = 210, training_loop: Union[str, Type[TrainingLoop] ] = "Stochastic Local Closed World Assumption", + verbose: bool = False, random_state: int = 42, enable_cache: bool = False ): @@ -46,6 +47,8 @@ def __init__( Can either be: - Stochastic Local Closed World Assumption - Local Closed World Assumption + verbose: bool = False + Whether to show loading bars. random_state: int = 42 Random seed to use while training the model enable_cache: bool = False @@ -58,6 +61,7 @@ def __init__( epochs=epochs, batch_size=batch_size, training_loop=training_loop, + verbose=verbose, random_state=random_state, enable_cache=enable_cache ) @@ -71,12 +75,12 @@ def smoke_test_parameters(cls) -> Dict[str, Any]: ) def parameters(self) -> Dict[str, Any]: - return { + return dict( super().parameters(), dict( scoring_fct_norm=self._scoring_fct_norm ) - } + ) @classmethod def model_name(cls) -> str: @@ -97,5 +101,6 @@ def _build_model( return TransH( triples_factory=triples_factory, embedding_dim=self._embedding_size, - scoring_fct_norm=self._scoring_fct_norm + scoring_fct_norm=self._scoring_fct_norm, + random_seed=self._random_state ) diff --git a/embiggen/embedders/pykeen_embedders/transr.py b/embiggen/embedders/pykeen_embedders/transr.py index fa1e48eb..3f8b5637 100644 --- a/embiggen/embedders/pykeen_embedders/transr.py +++ b/embiggen/embedders/pykeen_embedders/transr.py @@ -17,6 +17,7 @@ def __init__( batch_size: int = 210, training_loop: Union[str, Type[TrainingLoop] ] = "Stochastic Local Closed World Assumption", + verbose: bool = False, random_state: int = 42, enable_cache: bool = False ): @@ -49,6 +50,8 @@ def __init__( Can either be: - Stochastic Local Closed World Assumption - Local Closed World Assumption + verbose: bool = False + Whether to show loading bars. random_state: int = 42 Random seed to use while training the model enable_cache: bool = False @@ -62,6 +65,7 @@ def __init__( epochs=epochs, batch_size=batch_size, training_loop=training_loop, + verbose=verbose, random_state=random_state, enable_cache=enable_cache ) @@ -76,13 +80,13 @@ def smoke_test_parameters(cls) -> Dict[str, Any]: ) def parameters(self) -> Dict[str, Any]: - return { + return dict( super().parameters(), dict( scoring_fct_norm=self._scoring_fct_norm, relation_dim=self._relation_dim ) - } + ) @classmethod def model_name(cls) -> str: @@ -104,5 +108,6 @@ def _build_model( triples_factory=triples_factory, embedding_dim=self._embedding_size, relation_dim=self._relation_dim, - scoring_fct_norm=self._scoring_fct_norm + scoring_fct_norm=self._scoring_fct_norm, + random_seed=self._random_state ) diff --git a/embiggen/embedders/pykeen_embedders/tucker.py b/embiggen/embedders/pykeen_embedders/tucker.py index 99db4d20..70cc9b00 100644 --- a/embiggen/embedders/pykeen_embedders/tucker.py +++ b/embiggen/embedders/pykeen_embedders/tucker.py @@ -20,6 +20,7 @@ def __init__( batch_size: int = 210, training_loop: Union[str, Type[TrainingLoop] ] = "Stochastic Local Closed World Assumption", + verbose: bool = False, random_state: int = 42, enable_cache: bool = False ): @@ -58,6 +59,8 @@ def __init__( Can either be: - Stochastic Local Closed World Assumption - Local Closed World Assumption + verbose: bool = False + Whether to show loading bars. random_state: int = 42 Random seed to use while training the model enable_cache: bool = False @@ -74,12 +77,13 @@ def __init__( epochs=epochs, batch_size=batch_size, training_loop=training_loop, + verbose=verbose, random_state=random_state, enable_cache=enable_cache ) def parameters(self) -> Dict[str, Any]: - return { + return dict( super().parameters(), dict( relation_dim=self._relation_dim, @@ -88,7 +92,7 @@ def parameters(self) -> Dict[str, Any]: dropout_2=self._dropout_2, apply_batch_normalization=self._apply_batch_normalization ) - } + ) @classmethod def model_name(cls) -> str: @@ -113,5 +117,6 @@ def _build_model( dropout_0=self._dropout_0, dropout_1=self._dropout_1, dropout_2=self._dropout_2, - apply_batch_normalization=self._apply_batch_normalization + apply_batch_normalization=self._apply_batch_normalization, + random_seed=self._random_state ) diff --git a/embiggen/embedders/tensorflow_embedders/cbow.py b/embiggen/embedders/tensorflow_embedders/cbow.py index bbe9eedf..f9f5cb19 100644 --- a/embiggen/embedders/tensorflow_embedders/cbow.py +++ b/embiggen/embedders/tensorflow_embedders/cbow.py @@ -1,5 +1,4 @@ """CBOW model for sequence embedding.""" -from typing import Dict, Union from tensorflow.keras.layers import ( # pylint: disable=import-error,no-name-in-module GlobalAveragePooling1D, Input, Embedding ) @@ -22,7 +21,7 @@ class CBOWTensorFlow(Node2Vec): @classmethod def model_name(cls) -> str: """Returns name of the model.""" - return "CBOW" + return "Node2Vec CBOW" def _build_model(self, graph: Graph) -> Model: """Return CBOW model.""" @@ -53,7 +52,7 @@ def _build_model(self, graph: Graph) -> Model: model = Model( inputs=[contextual_terms, central_terms], outputs=sampled_softmax, - name=self.model_name() + name=self.model_name().replace(" ", "") ) model.compile(optimizer=self._optimizer) diff --git a/embiggen/embedders/tensorflow_embedders/edge_prediction_based_tensorflow_embedders.py b/embiggen/embedders/tensorflow_embedders/edge_prediction_based_tensorflow_embedders.py index 75be62cf..b949b6b8 100644 --- a/embiggen/embedders/tensorflow_embedders/edge_prediction_based_tensorflow_embedders.py +++ b/embiggen/embedders/tensorflow_embedders/edge_prediction_based_tensorflow_embedders.py @@ -31,6 +31,7 @@ def __init__( activation: str = "sigmoid", loss: str = "binary_crossentropy", optimizer: str = "nadam", + verbose: bool = False, enable_cache: bool = False, random_state: int = 42 ): @@ -74,6 +75,8 @@ def __init__( while for the HOPE models this is an Mean squared error. optimizer: str = "nadam" The optimizer to be used during the training of the model. + verbose: bool = False + Whether to show the loading bar while training the model. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -93,19 +96,20 @@ def __init__( epochs=epochs, batch_size=batch_size, optimizer=optimizer, + verbose=verbose, use_mirrored_strategy=use_mirrored_strategy, enable_cache=enable_cache, random_state=random_state ) def parameters(self) -> Dict[str, Any]: - return { + return dict( super().parameters(), dict( negative_samples_rate=self._negative_samples_rate, activation=self._activation ) - } + ) def _build_edge_prediction_based_model( self, @@ -192,7 +196,6 @@ def _build_sequence( def _build_input( self, graph: Graph, - verbose: bool ) -> Tuple[np.ndarray]: """Returns values to be fed as input into the model. @@ -200,20 +203,8 @@ def _build_input( ------------------ graph: Graph The graph to build the model for. - verbose: bool - Whether to show loading bars. - Not used in this context. """ - try: - AUTOTUNE = tf.data.AUTOTUNE - except: - AUTOTUNE = tf.data.experimental.AUTOTUNE - - return ( - self._build_sequence(graph) - .into_dataset() - .repeat() - .prefetch(AUTOTUNE), ) + return (self._build_sequence(graph), ) @classmethod def requires_nodes_sorted_by_decreasing_node_degree(cls) -> bool: diff --git a/embiggen/embedders/tensorflow_embedders/first_order_line.py b/embiggen/embedders/tensorflow_embedders/first_order_line.py index c0f13fec..fcc90a45 100644 --- a/embiggen/embedders/tensorflow_embedders/first_order_line.py +++ b/embiggen/embedders/tensorflow_embedders/first_order_line.py @@ -44,7 +44,7 @@ def _build_edge_prediction_based_model( @classmethod def model_name(cls) -> str: """Returns name of the current model.""" - return "First Order LINE" + return "First-order LINE" def _extract_embeddings( self, diff --git a/embiggen/embedders/tensorflow_embedders/node2vec.py b/embiggen/embedders/tensorflow_embedders/node2vec.py index e916883e..bd61f805 100644 --- a/embiggen/embedders/tensorflow_embedders/node2vec.py +++ b/embiggen/embedders/tensorflow_embedders/node2vec.py @@ -17,7 +17,7 @@ class Node2Vec(AbstractRandomWalkBasedEmbedderModel): def __init__( self, - number_of_negative_samples: int = 5, + number_of_negative_samples: int = 10, batch_size: int = 128, embedding_size: int = 100, epochs: int = 10, @@ -35,7 +35,8 @@ def __init__( max_neighbours: int = 100, normalize_by_degree: bool = False, random_state: int = 42, - optimizer: str = "sgd", + optimizer: str = "nadam", + verbose: bool = False, use_mirrored_strategy: bool = False, enable_cache: bool = False ): @@ -43,14 +44,14 @@ def __init__( Parameters ------------------------------- - number_of_negative_samples: int = 5 + number_of_negative_samples: int = 10 The number of negative classes to randomly sample per batch. This single sample of negative classes is evaluated for each element in the batch. batch_size: int = 128 The number of nodes to consider for each walk. embedding_size: int = 100 Dimension of the embedding. - epochs: int = 10 + epochs: int = 50 Number of epochs to train the model for. early_stopping_min_delta: float The minimum variation in the provided patience time @@ -102,8 +103,10 @@ def __init__( of the destination node degrees. random_state: int = 42 The random state to reproduce the training sequence. - optimizer: str = "sgd" + optimizer: str = "nadam" Optimizer to use during the training. + verbose: bool = False + Whether to show loading bars. use_mirrored_strategy: bool = False Whether to use mirrored strategy. enable_cache: bool = False @@ -131,6 +134,7 @@ def __init__( epochs=epochs, batch_size=batch_size, optimizer=optimizer, + verbose=verbose, use_mirrored_strategy=use_mirrored_strategy, enable_cache=enable_cache ) @@ -163,7 +167,6 @@ def _get_steps_per_epoch(self, graph: Graph) -> Tuple[Any]: def _build_input( self, graph: Graph, - verbose: bool ) -> Tuple[np.ndarray]: """Returns values to be fed as input into the model. @@ -171,9 +174,6 @@ def _build_input( ------------------ graph: Graph The graph to build the model for. - verbose: bool - Whether to show loading bars. - Not used in this context. """ try: AUTOTUNE = tf.data.AUTOTUNE diff --git a/embiggen/embedders/tensorflow_embedders/second_order_line.py b/embiggen/embedders/tensorflow_embedders/second_order_line.py index 96c47352..840983d8 100644 --- a/embiggen/embedders/tensorflow_embedders/second_order_line.py +++ b/embiggen/embedders/tensorflow_embedders/second_order_line.py @@ -51,7 +51,7 @@ def _build_edge_prediction_based_model( @classmethod def model_name(cls) -> str: """Returns name of the current model.""" - return "Second Order LINE" + return "Second-order LINE" def _extract_embeddings( self, diff --git a/embiggen/embedders/tensorflow_embedders/siamese.py b/embiggen/embedders/tensorflow_embedders/siamese.py index f460f052..7753ec7b 100644 --- a/embiggen/embedders/tensorflow_embedders/siamese.py +++ b/embiggen/embedders/tensorflow_embedders/siamese.py @@ -2,23 +2,17 @@ from typing import Dict, Tuple, Any, Optional import numpy as np -import pandas as pd import tensorflow as tf from ensmallen import Graph from tensorflow.keras import \ backend as K # pylint: disable=import-error,no-name-in-module -from tensorflow.keras.constraints import \ - UnitNorm # pylint: disable=import-error,no-name-in-module,no-name-in-module -from tensorflow.keras.layers import \ - Embedding # pylint: disable=import-error,no-name-in-module from tensorflow.keras.layers import ( # pylint: disable=import-error,no-name-in-module - Add, - GlobalAveragePooling1D, - Input + Input, ReLU ) from tensorflow.keras.models import Model +from embiggen.layers.tensorflow import FlatEmbedding, ElementWiseL2, ElementWiseL1 from embiggen.utils.abstract_models import abstract_class -from embiggen.sequences.tensorflow_sequences import SiameseSequence, KGSiameseSequence +from embiggen.sequences.tensorflow_sequences import SiameseSequence from embiggen.embedders.tensorflow_embedders.tensorflow_embedder import TensorFlowEmbedder @@ -30,14 +24,16 @@ def __init__( self, embedding_size: int = 100, relu_bias: float = 1.0, - epochs: int = 10, - batch_size: int = 210, - early_stopping_min_delta: float = 0.001, - early_stopping_patience: int = 5, - learning_rate_plateau_min_delta: float = 0.001, - learning_rate_plateau_patience: int = 2, + epochs: int = 50, + batch_size: int = 28, + early_stopping_min_delta: float = 0.0001, + early_stopping_patience: int = 10, + learning_rate_plateau_min_delta: float = 0.0001, + learning_rate_plateau_patience: int = 5, + norm: str = "L2", use_mirrored_strategy: bool = False, optimizer: str = "nadam", + verbose: bool = False, enable_cache: bool = False, random_state: int = 42 ): @@ -51,9 +47,9 @@ def __init__( It is not possible to provide both at once. relu_bias: float = 1.0 The bias to use for the ReLu. - epochs: int = 10 + epochs: int = 50 Number of epochs to train the model for. - batch_size: int = 214 + batch_size: int = 28 Batch size to use during the training. early_stopping_min_delta: float = 0.001 The minimum variation in the provided patience time @@ -67,10 +63,15 @@ def __init__( learning_rate_plateau_patience: int = 1 The amount of epochs to wait for better training performance without decreasing the learning rate. + norm: str = "L2" + Normalization to use. + Can either be `L1` or `L2`. use_mirrored_strategy: bool = False Whether to use mirrored strategy. optimizer: str = "nadam" The optimizer to be used during the training of the model. + verbose: bool = False + Whether to show loading bars. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -78,7 +79,7 @@ def __init__( The random state to use if the model is stocastic. """ self._relu_bias = relu_bias - + self._norm = norm super().__init__( embedding_size=embedding_size, early_stopping_min_delta=early_stopping_min_delta, @@ -88,126 +89,92 @@ def __init__( epochs=epochs, batch_size=batch_size, optimizer=optimizer, + verbose=verbose, use_mirrored_strategy=use_mirrored_strategy, enable_cache=enable_cache, random_state=random_state ) - @classmethod - def smoke_test_parameters(cls) -> Dict[str, Any]: - """Returns parameters for smoke test.""" - return dict( - TensorFlowEmbedder.smoke_test_parameters(), - ) - def parameters(self) -> Dict[str, Any]: - return { + return dict( super().parameters(), dict( - relu_bias = self._relu_bias + relu_bias=self._relu_bias, + norm=self._norm ) - } + ) def _build_model(self, graph: Graph): """Return Siamese model.""" # Creating the inputs layers inputs = [ - Input((1,), dtype=tf.int32) - for _ in range(4) + Input((1,), dtype=tf.int32, name=node_name) + for node_name in ( + "Sources", + "Destinations", + "Corrupted Sources", + "Corrupted Destinations", + ) ] - edge_types = Input((1,), dtype=tf.int32) - # Creating the embedding layer for the contexts - node_embedding_layer = Embedding( - input_dim=graph.get_number_of_nodes(), - output_dim=self._embedding_size, + node_embedding_layer = FlatEmbedding( + vocabulary_size=graph.get_number_of_nodes(), + dimension=self._embedding_size, input_length=1, - name="node_embeddings" + name="NodeEmbedding" ) # Get the node embedding node_embeddings = [ - UnitNorm(axis=-1)(node_embedding_layer(node_input)) + node_embedding_layer(node_input) for node_input in inputs ] - if self.requires_node_types(): - max_node_types = graph.get_maximum_multilabel_count() - multilabel = graph.has_multilabel_node_types() - unknown_node_types = graph.has_unknown_node_types() - node_types_offset = int(multilabel or unknown_node_types) - node_type_inputs = [ - Input((max_node_types,), dtype=tf.int32) - for _ in range(4) - ] - - node_type_embedding_layer = Embedding( - input_dim=graph.get_number_of_node_types() + node_types_offset, - output_dim=self._embedding_size, - input_length=max_node_types, - name="node_type_embeddings", - mask_zero=multilabel or unknown_node_types - ) - - node_embeddings = [ - Add()([ - GlobalAveragePooling1D()( - node_type_embedding_layer(node_type_input) - ), - node_embedding - ]) - for node_type_input, node_embedding in zip( - node_type_inputs, - node_embeddings - ) - ] - else: - node_type_inputs = [] - - inputs.extend(node_type_inputs) - inputs.append(edge_types) - - edge_types_number = graph.get_number_of_edge_types() - unknown_edge_types = graph.has_unknown_edge_types() - edge_types_offset = int(unknown_edge_types) - edge_type_embedding = GlobalAveragePooling1D()(Embedding( - input_dim=edge_types_number, - output_dim=self._embedding_size, - input_length=1 + edge_types_offset, - mask_zero=unknown_edge_types, - name="edge_type_embeddings", - )(edge_types)) - ( + edge_types, + regularization, srcs_embedding, dsts_embedding, not_srcs_embedding, not_dsts_embedding, - edge_type_embedding ) = self._build_output( node_embeddings, - edge_type_embedding, - edge_types + graph ) - loss = K.relu(self._relu_bias + tf.norm( - srcs_embedding + edge_type_embedding - dsts_embedding, - axis=-1 - ) - tf.norm( - not_srcs_embedding + edge_type_embedding - not_dsts_embedding, - axis=-1 - )) + if self._norm == "L2": + norm_layer = ElementWiseL2 + else: + norm_layer = ElementWiseL1 + + if dsts_embedding is not None: + srcs_embedding = norm_layer()([ + srcs_embedding, + dsts_embedding + ]) + + if not_dsts_embedding is not None: + not_srcs_embedding = norm_layer()([ + not_srcs_embedding, + not_dsts_embedding + ]) + + loss = ReLU()( + self._relu_bias + srcs_embedding - not_srcs_embedding + ) + regularization + + if edge_types is not None: + inputs.append(edge_types) # Creating the actual model model = Model( inputs=inputs, outputs=loss, - name=self.model_name() + name=self.model_name().replace(" ", "") ) model.add_loss(loss) - model.compile(optimizer=self._optimizer) return model @@ -218,10 +185,9 @@ def _build_output( dsts_embedding: tf.Tensor, not_srcs_embedding: tf.Tensor, not_dsts_embedding: tf.Tensor, - edge_type_embedding: tf.Tensor, - edge_type_input: Optional[Input] - ) -> Tuple[tf.Tensor, tf.Tensor, tf.Tensor, tf.Tensor, tf.Tensor]: - """Returns the five input tensors, arbitrarily changed. + graph: Graph + ) -> Tuple[Optional[Input], tf.Tensor, tf.Tensor, tf.Tensor, tf.Tensor, tf.Tensor]: + """Returns the inputs, if any, the regularization loss, and the received node embedding arbitrarily modified. Parameters ---------------------- @@ -233,12 +199,9 @@ def _build_output( Embedding of the fake source node. not_dsts_embedding: tf.Tensor Embedding of the fake destination node. - edge_type_embedding: tf.Tensor - Embedding of the edge types. - edge_type_input: Optional[Input] - Input of the edge types. This is not None - only when there is one such input in the - model, when edge types are requested. + graph: Graph + Graph whose structure is to be used to build + the model. """ raise NotImplementedError( "The method `_build_output` should be implemented in the child " @@ -259,7 +222,6 @@ def _get_steps_per_epoch(self, graph: Graph) -> int: def _build_input( self, graph: Graph, - verbose: bool ) -> Tuple[np.ndarray]: """Returns values to be fed as input into the model. @@ -267,29 +229,16 @@ def _build_input( ------------------ graph: Graph The graph to build the model for. - verbose: bool - Whether to show loading bars. - Not used in this context. """ - try: - AUTOTUNE = tf.data.AUTOTUNE - except: - AUTOTUNE = tf.data.experimental.AUTOTUNE - - if self.requires_node_types(): - sequence = KGSiameseSequence( - graph=graph, - batch_size=self._batch_size, - ) - else: - sequence = SiameseSequence( - graph=graph, - batch_size=self._batch_size, - ) + sequence = SiameseSequence( + graph=graph, + batch_size=self._batch_size, + return_edge_types=self.requires_edge_types() + ) return ( sequence.into_dataset() .repeat() - .prefetch(AUTOTUNE), ) + .prefetch(tf.data.AUTOTUNE), ) @classmethod def requires_nodes_sorted_by_decreasing_node_degree(cls) -> bool: @@ -302,4 +251,4 @@ def is_topological(cls) -> bool: @classmethod def can_use_edge_weights(cls) -> bool: """Returns whether the model can optionally use edge weights.""" - return False \ No newline at end of file + return False diff --git a/embiggen/embedders/tensorflow_embedders/skipgram.py b/embiggen/embedders/tensorflow_embedders/skipgram.py index 546af046..5147f941 100644 --- a/embiggen/embedders/tensorflow_embedders/skipgram.py +++ b/embiggen/embedders/tensorflow_embedders/skipgram.py @@ -1,5 +1,4 @@ """SkipGram model for sequence embedding.""" -from typing import Dict, Union from tensorflow.keras.layers import ( # pylint: disable=import-error,no-name-in-module Input, Embedding, Flatten ) @@ -24,7 +23,7 @@ class SkipGramTensorFlow(Node2Vec): @classmethod def model_name(cls) -> str: """Returns name of the model.""" - return "SkipGram" + return "Node2Vec SkipGram" def _build_model(self, graph: Graph) -> Model: """Return SkipGram model.""" @@ -54,7 +53,7 @@ def _build_model(self, graph: Graph) -> Model: model = Model( inputs=[contextual_terms, central_terms], outputs=output, - name=self.model_name() + name=self.model_name().replace(" ", "") ) model.compile(optimizer=self._optimizer) diff --git a/embiggen/embedders/tensorflow_embedders/structured_embedding.py b/embiggen/embedders/tensorflow_embedders/structured_embedding.py new file mode 100644 index 00000000..c1f1be0c --- /dev/null +++ b/embiggen/embedders/tensorflow_embedders/structured_embedding.py @@ -0,0 +1,143 @@ +"""StructuredEmbedding model.""" +import tensorflow as tf +from ensmallen import Graph +import pandas as pd +from tensorflow.keras import Model +from tensorflow.keras.layers import Input, Reshape +from embiggen.embedders.tensorflow_embedders.siamese import Siamese +from embiggen.utils.abstract_models import EmbeddingResult +from embiggen.layers.tensorflow import FlatEmbedding + + +class StructuredEmbeddingTensorFlow(Siamese): + """StructuredEmbedding model.""" + + def _build_output( + self, + srcs_embedding: tf.Tensor, + dsts_embedding: tf.Tensor, + not_srcs_embedding: tf.Tensor, + not_dsts_embedding: tf.Tensor, + graph: Graph + ): + """Returns the five input tensors, unchanged.""" + edge_types = Input((1,), dtype=tf.int32, name="Edge Types") + source_edge_type_embedding = Reshape(( + self._embedding_size, + self._embedding_size + ))(FlatEmbedding( + vocabulary_size=graph.get_number_of_edge_types(), + dimension=self._embedding_sizeself._embedding_size, + input_length=1, + mask_zero=graph.has_unknown_edge_types(), + name="SourceEdgeTypeEmbedding", + )(edge_types)) + destination_edge_type_embedding = Reshape(( + self._embedding_size, + self._embedding_size + ))(FlatEmbedding( + vocabulary_size=graph.get_number_of_edge_types(), + dimension=self._embedding_sizeself._embedding_size, + input_length=1, + mask_zero=graph.has_unknown_edge_types(), + name="DestinationEdgeTypeEmbedding", + )(edge_types)) + + return ( + edge_types, + 0.0, + tf.einsum( + 'ijk,ij->ij', + source_edge_type_embedding, + srcs_embedding + ), + tf.einsum( + 'ijk,ij->ij', + destination_edge_type_embedding, + dsts_embedding + ), + tf.einsum( + 'ijk,ij->ij', + source_edge_type_embedding, + not_srcs_embedding + ), + tf.einsum( + 'ijk,ij->ij', + destination_edge_type_embedding, + not_dsts_embedding + ), + ) + + @classmethod + def model_name(cls) -> str: + """Returns name of the current model.""" + return "Structured Embedding" + + @classmethod + def requires_edge_types(cls) -> bool: + return True + + def _extract_embeddings( + self, + graph: Graph, + model: Model, + return_dataframe: bool + ) -> EmbeddingResult: + """Returns embedding from the model. + + Parameters + ------------------ + graph: Graph + The graph that was embedded. + model: Model + The Keras model used to embed the graph. + return_dataframe: bool + Whether to return a dataframe of a numpy array. + """ + node_embedding = self.get_layer_weights( + "NodeEmbedding", + model, + drop_first_row=False + ) + source_edge_type_embedding = self.get_layer_weights( + "SourceEdgeTypeEmbedding", + model, + drop_first_row=graph.has_unknown_edge_types() + ) + destination_edge_type_embedding = self.get_layer_weights( + "DestinationEdgeTypeEmbedding", + model, + drop_first_row=graph.has_unknown_edge_types() + ) + + if return_dataframe: + node_embedding = pd.DataFrame( + node_embedding, + index=graph.get_node_names() + ) + + edge_type_names = graph.get_unique_edge_type_names() + + source_edge_type_embedding = pd.DataFrame( + source_edge_type_embedding, + index=edge_type_names + ) + + destination_edge_type_embedding = pd.DataFrame( + destination_edge_type_embedding, + index=edge_type_names + ) + + return EmbeddingResult( + embedding_method_name=self.model_name(), + node_embeddings=node_embedding, + edge_type_embeddings=[ + source_edge_type_embedding, + destination_edge_type_embedding + ] + ) + + @classmethod + def can_use_node_types(cls) -> bool: + """Returns whether the model can optionally use node types.""" + return False \ No newline at end of file diff --git a/embiggen/embedders/tensorflow_embedders/tensorflow_embedder.py b/embiggen/embedders/tensorflow_embedders/tensorflow_embedder.py index 58649702..d759f698 100644 --- a/embiggen/embedders/tensorflow_embedders/tensorflow_embedder.py +++ b/embiggen/embedders/tensorflow_embedders/tensorflow_embedder.py @@ -28,6 +28,7 @@ def __init__( epochs: int = 10, batch_size: int = 210, optimizer: str = "nadam", + verbose: bool = False, use_mirrored_strategy: bool = False, enable_cache: bool = False, random_state: int = 42 @@ -56,6 +57,8 @@ def __init__( Batch size to use during the training. optimizer: str = "nadam" Optimizer to use during the training. + verbose: bool = False + Whether to show loading bars. use_mirrored_strategy: bool = False Whether to use mirrored strategy. enable_cache: bool = False @@ -74,6 +77,7 @@ def __init__( self._epochs = epochs self._batch_size = batch_size self._optimizer = optimizer + self._verbose = verbose self._early_stopping_min_delta = early_stopping_min_delta self._early_stopping_patience = early_stopping_patience self._learning_rate_plateau_min_delta = learning_rate_plateau_min_delta @@ -93,19 +97,20 @@ def smoke_test_parameters(cls) -> Dict[str, Any]: ) def parameters(self) -> Dict[str, Any]: - return { + return dict( super().parameters(), dict( use_mirrored_strategy=self._use_mirrored_strategy, epochs=self._epochs, batch_size=self._batch_size, optimizer=self._optimizer, + verbose=self._verbose, early_stopping_min_delta=self._early_stopping_min_delta, early_stopping_patience=self._early_stopping_patience, learning_rate_plateau_min_delta=self._learning_rate_plateau_min_delta, learning_rate_plateau_patience=self._learning_rate_plateau_patience, ) - } + ) @classmethod def library_name(cls) -> str: @@ -125,7 +130,7 @@ def _build_model(self, graph: Graph) -> Model: "called `_build_model`. Please do implement it." ) - def _build_input(self, graph: Graph, verbose: bool) -> Tuple[Any]: + def _build_input(self, graph: Graph) -> Tuple[Any]: """Returns values to be fed as input into the model. Parameters @@ -209,7 +214,6 @@ def _fit_transform( self, graph: Graph, return_dataframe: bool = True, - verbose: bool = True ) -> EmbeddingResult: """Return node embedding""" try: @@ -230,7 +234,6 @@ def _fit_transform( # Get the model input training_input = self._build_input( graph, - verbose=verbose ) if not isinstance(training_input, tuple): @@ -242,7 +245,7 @@ def _fit_transform( model.fit( training_input, epochs=self._epochs, - verbose=traditional_verbose and verbose > 0, + verbose=traditional_verbose and self._verbose > 0, batch_size=( self._batch_size if issubclass(training_input[0].__class__, Sequence) @@ -264,7 +267,7 @@ def _fit_transform( mode="min", ), ((TqdmCallback(verbose=1, leave=False),) - if not traditional_verbose and verbose > 0 else ()), + if not traditional_verbose and self._verbose > 0 else ()), ], ) @@ -278,4 +281,4 @@ def _fit_transform( @classmethod def is_stocastic(cls) -> bool: """Returns whether the model is stocastic and has therefore a random state.""" - return True \ No newline at end of file + return True diff --git a/embiggen/embedders/tensorflow_embedders/transe.py b/embiggen/embedders/tensorflow_embedders/transe.py index 59644d7a..877aceac 100644 --- a/embiggen/embedders/tensorflow_embedders/transe.py +++ b/embiggen/embedders/tensorflow_embedders/transe.py @@ -2,19 +2,40 @@ from ensmallen import Graph import pandas as pd from tensorflow.keras import Model +from tensorflow.keras.layers import Input +import tensorflow as tf from embiggen.embedders.tensorflow_embedders.siamese import Siamese from embiggen.utils.abstract_models import EmbeddingResult - +from embiggen.layers.tensorflow import FlatEmbedding class TransETensorFlow(Siamese): """TransE model.""" def _build_output( self, - args + srcs_embedding: tf.Tensor, + dsts_embedding: tf.Tensor, + not_srcs_embedding: tf.Tensor, + not_dsts_embedding: tf.Tensor, + graph: Graph ): """Returns the five input tensors, unchanged.""" - return args[:-1] + edge_types = Input((1,), dtype=tf.int32, name="Edge Types") + edge_type_embedding = FlatEmbedding( + vocabulary_size=graph.get_number_of_edge_types(), + dimension=self._embedding_size, + input_length=1, + mask_zero=graph.has_unknown_edge_types(), + name="EdgeTypeEmbedding", + )(edge_types) + return ( + edge_types, + 0.0, + srcs_embedding + edge_type_embedding, + dsts_embedding, + not_srcs_embedding + edge_type_embedding, + not_dsts_embedding, + ) @classmethod def model_name(cls) -> str: @@ -42,36 +63,31 @@ def _extract_embeddings( return_dataframe: bool Whether to return a dataframe of a numpy array. """ + node_embedding = self.get_layer_weights( + "NodeEmbedding", + model, + drop_first_row=False + ) + edge_type_embedding = self.get_layer_weights( + "EdgeTypeEmbedding", + model, + drop_first_row=graph.has_unknown_edge_types() + ) + if return_dataframe: - result = { - layer_name: pd.DataFrame( - self.get_layer_weights( - layer_name, - model, - drop_first_row=drop_first_row - ), - index=names - ) - for layer_name, names, drop_first_row in ( - ("node_embeddings", graph.get_node_names(), False), - ("edge_type_embeddings", graph.get_unique_edge_type_names(), graph.has_unknown_edge_types()) - ) - } - else: - result = { - layer_name: self.get_layer_weights( - layer_name, - model, - drop_first_row=drop_first_row - ) - for layer_name, drop_first_row in ( - ("node_embeddings", False), - ("edge_type_embeddings", graph.has_unknown_edge_types()) - ) - } + node_embedding = pd.DataFrame( + node_embedding, + index=graph.get_node_names() + ) + edge_type_embedding = pd.DataFrame( + edge_type_embedding, + index=graph.get_unique_edge_type_names() + ) + return EmbeddingResult( embedding_method_name=self.model_name(), - *result + node_embeddings=node_embedding, + edge_type_embeddings=edge_type_embedding ) @classmethod @@ -80,6 +96,6 @@ def can_use_node_types(cls) -> bool: return False @classmethod - def task_involves_edge_types(cls) -> bool: - """Returns whether the model task involves edge types.""" + def requires_edge_types(cls) -> bool: + """Returns whether the model requires edge types.""" return True \ No newline at end of file diff --git a/embiggen/embedders/tensorflow_embedders/transh.py b/embiggen/embedders/tensorflow_embedders/transh.py new file mode 100644 index 00000000..9bd8fed6 --- /dev/null +++ b/embiggen/embedders/tensorflow_embedders/transh.py @@ -0,0 +1,146 @@ +"""TransH model.""" +import tensorflow as tf +from ensmallen import Graph +import pandas as pd +from tensorflow.keras import Model +from tensorflow.keras.layers import Input, Dot +from embiggen.embedders.tensorflow_embedders.siamese import Siamese +from embiggen.utils.abstract_models import EmbeddingResult +from embiggen.layers.tensorflow import FlatEmbedding + + +class TransHTensorFlow(Siamese): + """TransH model.""" + + def _build_output( + self, + srcs_embedding: tf.Tensor, + dsts_embedding: tf.Tensor, + not_srcs_embedding: tf.Tensor, + not_dsts_embedding: tf.Tensor, + graph: Graph + ): + """Returns the five input tensors, unchanged.""" + edge_types = Input((1,), dtype=tf.int32, name="Edge Types") + bias_edge_type_embedding = FlatEmbedding( + vocabulary_size=graph.get_number_of_edge_types(), + dimension=self._embedding_size, + input_length=1, + mask_zero=graph.has_unknown_edge_types(), + name="BiasEdgeTypeEmbedding", + )(edge_types) + multiplicative_edge_type_embedding = FlatEmbedding( + vocabulary_size=graph.get_number_of_edge_types(), + dimension=self._embedding_size, + input_length=1, + mask_zero=graph.has_unknown_edge_types(), + name="MultiplicativeEdgeTypeEmbedding", + )(edge_types) + + dot = Dot(axes=-1)([ + bias_edge_type_embedding, + multiplicative_edge_type_embedding + ]) + + return ( + edge_types, + dot dot / Dot(axes=-1)([ + bias_edge_type_embedding, + bias_edge_type_embedding + ]), + srcs_embedding - multiplicative_edge_type_embeddingDot(axes=-1)([ + multiplicative_edge_type_embedding, + srcs_embedding + ]) + bias_edge_type_embedding, + dsts_embedding - multiplicative_edge_type_embeddingDot(axes=-1)([ + multiplicative_edge_type_embedding, + dsts_embedding + ]), + not_srcs_embedding - multiplicative_edge_type_embeddingDot(axes=-1)([ + multiplicative_edge_type_embedding, + not_srcs_embedding + ]) + bias_edge_type_embedding, + not_dsts_embedding - multiplicative_edge_type_embeddingDot(axes=-1)([ + multiplicative_edge_type_embedding, + not_dsts_embedding + ]) + ) + + @classmethod + def model_name(cls) -> str: + """Returns name of the current model.""" + return "TransH" + + @classmethod + def requires_edge_types(cls) -> bool: + return True + + def _extract_embeddings( + self, + graph: Graph, + model: Model, + return_dataframe: bool + ) -> EmbeddingResult: + """Returns embedding from the model. + + Parameters + ------------------ + graph: Graph + The graph that was embedded. + model: Model + The Keras model used to embed the graph. + return_dataframe: bool + Whether to return a dataframe of a numpy array. + """ + node_embedding = self.get_layer_weights( + "NodeEmbedding", + model, + drop_first_row=False + ) + bias_edge_type_embedding = self.get_layer_weights( + "BiasEdgeTypeEmbedding", + model, + drop_first_row=graph.has_unknown_edge_types() + ) + multiplicative_edge_type_embedding = self.get_layer_weights( + "MultiplicativeEdgeTypeEmbedding", + model, + drop_first_row=graph.has_unknown_edge_types() + ) + + if return_dataframe: + node_embedding = pd.DataFrame( + node_embedding, + index=graph.get_node_names() + ) + + edge_type_names = graph.get_unique_edge_type_names() + + bias_edge_type_embedding = pd.DataFrame( + bias_edge_type_embedding, + index=edge_type_names + ) + + multiplicative_edge_type_embedding = pd.DataFrame( + multiplicative_edge_type_embedding, + index=edge_type_names + ) + + return EmbeddingResult( + embedding_method_name=self.model_name(), + node_embeddings=node_embedding, + edge_type_embeddings=[ + bias_edge_type_embedding, + multiplicative_edge_type_embedding + ] + ) + + @classmethod + def can_use_node_types(cls) -> bool: + """Returns whether the model can optionally use node types.""" + return False + + @classmethod + def requires_edge_types(cls) -> bool: + """Returns whether the model requires edge types.""" + return True \ No newline at end of file diff --git a/embiggen/embedders/tensorflow_embedders/unstructured.py b/embiggen/embedders/tensorflow_embedders/unstructured.py new file mode 100644 index 00000000..20794f20 --- /dev/null +++ b/embiggen/embedders/tensorflow_embedders/unstructured.py @@ -0,0 +1,77 @@ +"""Unstructured model.""" +from ensmallen import Graph +import pandas as pd +import tensorflow as tf +from tensorflow.keras import Model +from embiggen.embedders.tensorflow_embedders.siamese import Siamese +from embiggen.utils.abstract_models import EmbeddingResult + + +class UnstructuredTensorFlow(Siamese): + """Unstructured model.""" + + def _build_output( + self, + srcs_embedding: tf.Tensor, + dsts_embedding: tf.Tensor, + not_srcs_embedding: tf.Tensor, + not_dsts_embedding: tf.Tensor, + graph: Graph + ): + """Returns the five input tensors, unchanged.""" + return ( + None, + 0.0, + srcs_embedding, + dsts_embedding, + not_srcs_embedding, + not_dsts_embedding + ) + + @classmethod + def model_name(cls) -> str: + """Returns name of the current model.""" + return "Unstructured" + + @classmethod + def can_use_edge_types(cls) -> bool: + return False + + def _extract_embeddings( + self, + graph: Graph, + model: Model, + return_dataframe: bool + ) -> EmbeddingResult: + """Returns embedding from the model. + + Parameters + ------------------ + graph: Graph + The graph that was embedded. + model: Model + The Keras model used to embed the graph. + return_dataframe: bool + Whether to return a dataframe of a numpy array. + """ + node_embedding = self.get_layer_weights( + "NodeEmbedding", + model, + drop_first_row=False + ) + + if return_dataframe: + node_embedding = pd.DataFrame( + node_embedding, + index=graph.get_node_names() + ) + + return EmbeddingResult( + embedding_method_name=self.model_name(), + node_embeddings=node_embedding, + ) + + @classmethod + def can_use_node_types(cls) -> bool: + """Returns whether the model can optionally use node types.""" + return False diff --git a/embiggen/embedding_transformers/__init__.py b/embiggen/embedding_transformers/__init__.py index cd8953ec..a465e9f4 100644 --- a/embiggen/embedding_transformers/__init__.py +++ b/embiggen/embedding_transformers/__init__.py @@ -10,7 +10,6 @@ "EdgeTransformer", "NodeTransformer", "GraphTransformer", - "CorpusTransformer", "EdgePredictionTransformer", "EdgeLabelPredictionTransformer", "NodeLabelPredictionTransformer" diff --git a/embiggen/embedding_transformers/edge_label_prediction_transformer.py b/embiggen/embedding_transformers/edge_label_prediction_transformer.py index dd3de564..4b2210ac 100644 --- a/embiggen/embedding_transformers/edge_label_prediction_transformer.py +++ b/embiggen/embedding_transformers/edge_label_prediction_transformer.py @@ -14,7 +14,7 @@ class EdgeLabelPredictionTransformer: def __init__( self, method: str = "Hadamard", - aligned_node_mapping: bool = False, + aligned_mapping: bool = False, one_hot_encode_labels: bool = False ): """Create new EdgeLabelPredictionTransformer object. @@ -24,7 +24,7 @@ def __init__( method: str = "hadamard" Method to use for the embedding. Can either be 'Hadamard', 'Sum', 'Average', 'L1', 'AbsoluteL1', 'L2' or 'Concatenate'. - aligned_node_mapping: bool = False + aligned_mapping: bool = False This parameter specifies whether the mapping of the embeddings nodes matches the internal node mapping of the given graph. If these two mappings do not match, the generated edge embedding @@ -34,31 +34,39 @@ def __init__( """ self._transformer = GraphTransformer( method=method, - aligned_node_mapping=aligned_node_mapping, + aligned_mapping=aligned_mapping, ) self._one_hot_encode_labels = one_hot_encode_labels - def fit(self, node_feature: Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]): + def fit( + self, + node_feature: Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]], + node_type_feature: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, + ): """Fit the model. Parameters ------------------------- node_feature: Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]], Node feature to use to fit the transformer. + node_type_feature: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None + Node type feature to use to fit the transformer. Raises ------------------------- ValueError If the given method is None there is no need to call the fit method. """ - self._transformer.fit(node_feature) + self._transformer.fit( + node_feature, + node_type_feature=node_type_feature + ) def transform( self, graph: Graph, edge_features: Optional[Union[np.ndarray, List[np.ndarray]]] = None, behaviour_for_unknown_edge_labels: Optional[str] = None, - random_state: int = 42 ) -> Tuple[np.ndarray, np.ndarray]: """Return edge embedding for given graph using provided method. @@ -79,8 +87,6 @@ def transform( By default, we drop these edges. If the behaviour has not been specified and left to None, a warning will be raised to notify the user of this uncertainty. - random_state: int = 42, - The random state to use to shuffle the labels. Raises -------------------------- @@ -104,6 +110,7 @@ def transform( "The provided graph for the edge-label prediction does " "not contain edge-types." ) + if not graph.has_known_edge_types(): raise ValueError( "The provided graph for the edge-label prediction does " @@ -111,15 +118,18 @@ def transform( "an edge type vocabulary but no edge has an edge type " "assigned to it." ) + if graph.has_homogeneous_edge_types(): raise ValueError( "The provided graph for the edge-label prediction contains " "edges of a single type, making predictions pointless." ) + if graph.is_multigraph(): raise NotImplementedError( "Multi-graphs are not currently supported by this class." ) + if graph.has_singleton_edge_types(): warnings.warn( "Please do be advised that this graph contains edges with " @@ -137,6 +147,7 @@ def transform( edge_type_counts.items(), key=lambda x: x[1] ) + if most_common_count > least_common_count * 20: warnings.warn( ( @@ -163,6 +174,7 @@ def transform( edge_embeddings = self._transformer.transform( graph, + node_types=graph, edge_features=edge_features ) @@ -176,12 +188,6 @@ def transform( edge_labels = graph.get_known_edge_type_ids() edge_embeddings = edge_embeddings[graph.get_edges_with_known_edge_types_mask()] else: - edge_labels = graph.get_edge_types_ids() - - numpy_random_state = np.random.RandomState( # pylint: disable=no-member - seed=random_state - ) - - indices = numpy_random_state.permutation(edge_labels.size) + edge_labels = graph.get_directed_edge_type_ids() - return edge_embeddings[indices], edge_labels[indices] + return edge_embeddings, edge_labels diff --git a/embiggen/embedding_transformers/edge_prediction_transformer.py b/embiggen/embedding_transformers/edge_prediction_transformer.py index ded2da1c..4e911089 100644 --- a/embiggen/embedding_transformers/edge_prediction_transformer.py +++ b/embiggen/embedding_transformers/edge_prediction_transformer.py @@ -13,7 +13,7 @@ class EdgePredictionTransformer: def __init__( self, method: str = "Hadamard", - aligned_node_mapping: bool = False, + aligned_mapping: bool = False, ): """Create new EdgePredictionTransformer object. @@ -22,7 +22,7 @@ def __init__( method: str = "hadamard", Method to use for the embedding. Can either be 'Hadamard', 'Sum', 'Average', 'L1', 'AbsoluteL1', 'L2' or 'Concatenate'. - aligned_node_mapping: bool = False, + aligned_mapping: bool = False, This parameter specifies whether the mapping of the embeddings nodes matches the internal node mapping of the given graph. If these two mappings do not match, the generated edge embedding @@ -30,23 +30,32 @@ def __init__( """ self._transformer = GraphTransformer( method=method, - aligned_node_mapping=aligned_node_mapping, + aligned_mapping=aligned_mapping, ) - def fit(self, node_feature: Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]): + def fit( + self, + node_feature: Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]], + node_type_feature: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, + ): """Fit the model. Parameters ------------------------- node_feature: Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]], Node feature to use to fit the transformer. + node_type_feature: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None + Node type feature to use to fit the transformer. Raises ------------------------- ValueError If the given method is None there is no need to call the fit method. """ - self._transformer.fit(node_feature) + self._transformer.fit( + node_feature, + node_type_feature=node_type_feature + ) def transform( self, @@ -116,10 +125,12 @@ def transform( positive_edge_embedding = self._transformer.transform( positive_graph, + node_types=positive_graph, edge_features=positive_edge_features ) negative_edge_embedding = self._transformer.transform( negative_graph, + node_types=negative_graph, edge_features=negative_edge_features ) diff --git a/embiggen/embedding_transformers/edge_transformer.py b/embiggen/embedding_transformers/edge_transformer.py index 3470645c..a9967d56 100644 --- a/embiggen/embedding_transformers/edge_transformer.py +++ b/embiggen/embedding_transformers/edge_transformer.py @@ -3,7 +3,7 @@ import numpy as np import pandas as pd -from userinput.utils import closest +from userinput.utils import must_be_in_set from ensmallen import express_measures from embiggen.utils.abstract_models import format_list from embiggen.embedding_transformers.node_transformer import NodeTransformer @@ -308,18 +308,6 @@ def get_max_edge_embedding( axis=0 ) - -def get_indices_edge_embedding( - source_node_embedding: np.ndarray, - destination_node_embedding: np.ndarray -) -> np.ndarray: - """Placeholder method.""" - return np.vstack(( - source_node_embedding, - destination_node_embedding - )).T - - class EdgeTransformer: """EdgeTransformer class to convert edges to edge embeddings.""" @@ -336,13 +324,12 @@ class EdgeTransformer: "Max": get_max_edge_embedding, "L2Distance": get_l2_distance, "CosineSimilarity": get_cosine_similarity, - None: get_indices_edge_embedding, } def __init__( self, method: str = "Hadamard", - aligned_node_mapping: bool = False, + aligned_mapping: bool = False, ): """Create new EdgeTransformer object. @@ -353,78 +340,53 @@ def __init__( If None is used, we return instead the numeric tuples. Can either be 'Hadamard', 'Min', 'Max', 'Sum', 'Average', 'L1', 'AbsoluteL1', 'SquaredL2', 'L2' or 'Concatenate'. - aligned_node_mapping: bool = False, + aligned_mapping: bool = False, This parameter specifies whether the mapping of the embeddings nodes matches the internal node mapping of the given graph. If these two mappings do not match, the generated edge embedding will be meaningless. """ - if isinstance(method, str) and method.lower() not in [ - None if method_name is None else method_name.lower() - for method_name in EdgeTransformer.methods - ]: - raise ValueError(( - "Given method '{}' is not supported. " - "Supported methods are {}, or alternatively a lambda. " - "Maybe you meant {}?" - ).format( - method, - format_list( - [method for method in EdgeTransformer.methods.keys() if method is not None]), - closest(method, [ - method_name - for method_name in EdgeTransformer.methods - if method_name is not None - ]) - )) + method = must_be_in_set( + method, + self.methods, + "edge embedding method" + ) self._transformer = NodeTransformer( - numeric_node_ids=method is None, - aligned_node_mapping=aligned_node_mapping, + aligned_mapping=aligned_mapping, ) self._method_name = method - if self._method_name is None: - self._method = EdgeTransformer.methods[None] - else: - self._method = { - None if method_name is None else method_name.lower(): callback - for method_name, callback in EdgeTransformer.methods.items() - }[self._method_name.lower()] - - @property - def numeric_node_ids(self) -> bool: - """Return whether the transformer returns numeric node IDs.""" - return self._transformer.numeric_node_ids + self._method = self.methods[method] @property def method(self) -> str: """Return the used edge embedding method.""" return self._method_name - def fit(self, node_feature: Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]): + def fit( + self, + node_feature: Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]], + node_type_feature: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, + ): """Fit the model. Parameters ------------------------- node_feature: Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]], Node feature to use to fit the transformer. - - Raises - ------------------------- - ValueError - If the given method is None there is no need to call the fit method. + node_type_feature: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None + Node type feature to use to fit the transformer. """ - if self._method is None: - raise ValueError( - "There is no need to call the fit when edge method is None, " - "as the transformer will exclusively return the numeric node " - "indices and not any node feature." - ) - self._transformer.fit(node_feature) + self._transformer.fit( + node_feature, + node_type_feature=node_type_feature + ) def transform( self, sources: Union[List[str], List[int]], destinations: Union[List[str], List[int]], + source_node_types: Optional[Union[List[Optional[List[str]]], List[Optional[List[int]]]]] = None, + destination_node_types: Optional[Union[List[Optional[List[str]]], List[Optional[List[int]]]]] = None, edge_features: Optional[Union[np.ndarray, List[np.ndarray]]] = None, ) -> np.ndarray: """Return embedding for given edges using provided method. @@ -435,6 +397,16 @@ def transform( List of source nodes whose embedding is to be returned. destinations:Union[List[str], List[int]] List of destination nodes whose embedding is to be returned. + source_node_types: Optional[Union[List[Optional[List[str]]], List[Optional[List[int]]]]] = None, + List of source node types whose embedding is to be returned. + This can be either a list of strings, or a graph, or if the + aligned_mapping is setted, then this methods also accepts + a list of ints. + destination_node_types: Optional[Union[List[Optional[List[str]]], List[Optional[List[int]]]]] = None, + List of destination node types whose embedding is to be returned. + This can be either a list of strings, or a graph, or if the + aligned_mapping is setted, then this methods also accepts + a list of ints. edge_features: Optional[Union[np.ndarray, List[np.ndarray]]] = None Optional edge features to be used as input concatenated to the obtained edge embedding. The shape must be equal @@ -470,15 +442,17 @@ def transform( "numpy arrays of type uint32, but you have provided objects of type " f"{sources.dtype} and {destinations.dtype}. " ) - if self._transformer._node_feature.dtype != np.float32: - self._transformer._node_feature = self._transformer._node_feature.astype( - np.float32 - ) if not self._transformer._node_feature.data.c_contiguous: self._transformer._node_feature = np.ascontiguousarray( self._transformer._node_feature ) + if self._transformer._node_type_feature is not None: + raise NotImplementedError( + "The node type features are not yet supported for the " + "Cosine Similarity." + ) + edge_embeddings = self._method( embedding=self._transformer._node_feature, source_node_ids=sources, @@ -486,8 +460,8 @@ def transform( ) else: edge_embeddings = self._method( - self._transformer.transform(sources), - self._transformer.transform(destinations) + self._transformer.transform(sources, node_types=source_node_types), + self._transformer.transform(destinations, node_types=destination_node_types) ) if edge_features is not None: diff --git a/embiggen/embedding_transformers/graph_transformer.py b/embiggen/embedding_transformers/graph_transformer.py index 7b314f65..a7595e0f 100644 --- a/embiggen/embedding_transformers/graph_transformer.py +++ b/embiggen/embedding_transformers/graph_transformer.py @@ -13,55 +13,69 @@ class GraphTransformer: def __init__( self, method: str = "Hadamard", - aligned_node_mapping: bool = False, + aligned_mapping: bool = False, + include_both_undirected_edges: bool = True ): """Create new GraphTransformer object. Parameters ------------------------ - method: str = "hadamard", + method: str = "hadamard" Method to use for the embedding. Can either be 'Hadamard', 'Sum', 'Average', 'L1', 'AbsoluteL1', 'L2' or 'Concatenate'. - aligned_node_mapping: bool = False, + aligned_mapping: bool = False This parameter specifies whether the mapping of the embeddings nodes matches the internal node mapping of the given graph. If these two mappings do not match, the generated edge embedding will be meaningless. + include_both_undirected_edges: bool = True + Whether to include both undirected edges when parsing an undirected + graph, that is both the edge from source to destination and the edge + from destination to source. While both edges should be included when + training a model, as the model should learn about these simmetries + in the graph, these edges are not necessary in the context of visualizations + where they create redoundancy. """ self._transformer = EdgeTransformer( method=method, - aligned_node_mapping=aligned_node_mapping, + aligned_mapping=aligned_mapping, ) - self._aligned_node_mapping = aligned_node_mapping - - @property - def numeric_node_ids(self) -> bool: - """Return whether the transformer returns numeric node IDs.""" - return self._transformer.numeric_node_ids + self._include_both_undirected_edges = include_both_undirected_edges + self._aligned_mapping = aligned_mapping @property def method(self) -> str: """Return the used edge embedding method.""" return self._transformer.method - def fit(self, node_feature: Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]): + def fit( + self, + node_feature: Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]], + node_type_feature: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, + ): """Fit the model. Parameters ------------------------- - node_feature: Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]], + node_feature: Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]] Node feature to use to fit the transformer. + node_type_feature: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None + Node type feature to use to fit the transformer. Raises ------------------------- ValueError If the given method is None there is no need to call the fit method. """ - self._transformer.fit(node_feature) + self._transformer.fit( + node_feature, + node_type_feature=node_type_feature + ) def transform( self, graph: Union[Graph, np.ndarray, List[List[str]], List[List[int]]], + node_types: Optional[Union[Graph, List[Optional[List[str]]], List[Optional[List[int]]]]] = None, edge_features: Optional[Union[np.ndarray, List[np.ndarray]]] = None, ) -> np.ndarray: """Return edge embedding for given graph using provided method. @@ -71,6 +85,11 @@ def transform( graph: Union[Graph, np.ndarray, List[List[str]], List[List[int]]], The graph whose edges are to embed. It can either be an Graph or a list of lists of edges. + node_types: Optional[Union[Graph, List[Optional[List[str]]], List[Optional[List[int]]]]] = None, + List of node types whose embedding is to be returned. + This can be either a list of strings, or a graph, or if the + aligned_mapping is setted, then this methods also accepts + a list of ints. edge_features: Optional[Union[np.ndarray, List[np.ndarray]]] = None Optional edge features to be used as input concatenated to the obtained edge embedding. The shape must be equal @@ -86,43 +105,85 @@ def transform( Numpy array of embeddings. """ if isinstance(graph, Graph): - if self._aligned_node_mapping: - graph = ( - graph.get_directed_source_node_ids(), - graph.get_directed_destination_node_ids(), - ) + if self._aligned_mapping: + if graph.is_directed() or self._include_both_undirected_edges: + edge_node_ids = ( + graph.get_directed_source_node_ids(), + graph.get_directed_destination_node_ids(), + ) + else: + edge_node_ids = ( + graph.get_source_node_ids(directed=False), + graph.get_destination_node_ids(directed=False), + ) else: - graph = graph.get_directed_edge_node_names() - if isinstance(graph, List): - graph = np.array(graph) + edge_node_ids = graph.get_directed_edge_node_names() + else: + edge_node_ids = graph + + if isinstance(edge_node_ids, List): + edge_node_ids = np.array(edge_node_ids) if ( - isinstance(graph, tuple) and - len(graph) == 2 and - all(isinstance(e, np.ndarray) for e in graph) + isinstance(edge_node_ids, tuple) and + len(edge_node_ids) == 2 and + all(isinstance(e, np.ndarray) for e in edge_node_ids) ): if ( - len(graph[0].shape) != 1 or - len(graph[1].shape) != 1 or - graph[0].shape[0] == 0 or - graph[1].shape[0] == 0 or - graph[0].shape[0] != graph[1].shape[0] + len(edge_node_ids[0].shape) != 1 or + len(edge_node_ids[1].shape) != 1 or + edge_node_ids[0].shape[0] == 0 or + edge_node_ids[1].shape[0] == 0 or + edge_node_ids[0].shape[0] != edge_node_ids[1].shape[0] ): raise ValueError( "When providing a tuple of numpy arrays containing the source and destination " "node IDs, we expect to receive two arrays both with shape " "with shape (number of edges,). " - f"The ones you have provided have shapes {graph[0].shape} " - f"and {graph[1].shape}." + f"The ones you have provided have shapes {edge_node_ids[0].shape} " + f"and {edge_node_ids[1].shape}." ) - sources = graph[0] - destinations = graph[1] - elif isinstance(graph, np.ndarray): - if len(graph.shape) != 2 or graph.shape[1] != 2 or graph.shape[0] == 0: + sources = edge_node_ids[0] + destinations = edge_node_ids[1] + elif isinstance(edge_node_ids, np.ndarray): + if len(edge_node_ids.shape) != 2 or edge_node_ids.shape[1] != 2 or edge_node_ids.shape[0] == 0: raise ValueError( "When providing a numpy array containing the source and destination " "node IDs representing the graph edges, we expect to receive an array " - f"with shape (number of edges, 2). The one you have provided has shape {graph.shape}." + f"with shape (number of edges, 2). The one you have provided has shape {edge_node_ids.shape}." ) - sources = graph[:, 0] - destinations = graph[:, 1] - return self._transformer.transform(sources, destinations, edge_features=edge_features) + sources = edge_node_ids[:, 0] + destinations = edge_node_ids[:, 1] + + if node_types is not None and self._transformer._transformer._node_type_feature is not None: + if isinstance(node_types, Graph): + if self._aligned_mapping: + source_node_types = [ + node_types.get_node_type_ids_from_node_id(src) + for src in sources + ] + destination_node_types = [ + node_types.get_node_type_ids_from_node_id(dst) + for dst in destinations + ] + else: + source_node_types = [ + node_types.get_node_type_names_from_node_name(src) + for src in sources + ] + destination_node_types = [ + node_types.get_node_type_names_from_node_name(dst) + for dst in destinations + ] + else: + source_node_types, destination_node_types = node_types + else: + source_node_types = None + destination_node_types = None + + return self._transformer.transform( + sources, + destinations, + source_node_types=source_node_types, + destination_node_types=destination_node_types, + edge_features=edge_features + ) diff --git a/embiggen/embedding_transformers/node_label_prediction_transformer.py b/embiggen/embedding_transformers/node_label_prediction_transformer.py index 0aef517c..02b8ee26 100644 --- a/embiggen/embedding_transformers/node_label_prediction_transformer.py +++ b/embiggen/embedding_transformers/node_label_prediction_transformer.py @@ -13,20 +13,20 @@ class NodeLabelPredictionTransformer: def __init__( self, - aligned_node_mapping: bool = False, + aligned_mapping: bool = False, ): """Create new NodeLabelPredictionTransformer object. Parameters ------------------------ - aligned_node_mapping: bool = False + aligned_mapping: bool = False This parameter specifies whether the mapping of the embeddings nodes matches the internal node mapping of the given graph. If these two mappings do not match, the generated node embedding will be meaningless. """ self._transformer = NodeTransformer( - aligned_node_mapping=aligned_node_mapping, + aligned_mapping=aligned_mapping, ) def fit(self, node_feature: Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]): @@ -86,24 +86,25 @@ def transform( """ if not graph.has_node_types(): raise ValueError( - "The provided graph for the node-label prediction does " + f"The provided graph {graph.get_name()} " + "for the node-label prediction does " "not contain node-types." ) if not graph.has_known_node_types(): raise ValueError( - "The provided graph for the node-label prediction does " + "fThe provided graph {graph.get_name()} for the node-label prediction does " "not contain known node-types, that is, it contains " "an node type vocabulary but no node has an node type " "assigned to it." ) if graph.has_homogeneous_node_types(): raise ValueError( - "The provided graph for the node-label prediction contains " + f"The provided graph {graph.get_name()} for the node-label prediction contains " "nodes of a single type, making predictions pointless." ) if graph.has_singleton_node_types(): warnings.warn( - "Please do be advised that this graph contains nodes with " + f"Please do be advised that the {graph.get_name()} graph contains nodes with " "a singleton node type, that is an node type that appears " "only once in the graph. Predictions on such rare node types " "will be unlikely to generalize well." diff --git a/embiggen/embedding_transformers/node_transformer.py b/embiggen/embedding_transformers/node_transformer.py index 1424c6d3..f6444795 100644 --- a/embiggen/embedding_transformers/node_transformer.py +++ b/embiggen/embedding_transformers/node_transformer.py @@ -1,105 +1,134 @@ """NodeTransformer class to convert nodes to edge embeddings.""" -from typing import List, Union +from typing import List, Union, Optional import numpy as np import pandas as pd from ensmallen import Graph + class NodeTransformer: """NodeTransformer class to convert nodes to edge embeddings.""" def __init__( self, - numeric_node_ids: bool = False, - aligned_node_mapping: bool = False, + aligned_mapping: bool = False, ): """Create new NodeTransformer object. Parameters ------------------- - numeric_node_ids: bool = False, - Wether to return the numeric node IDs instead of the node embedding. - aligned_node_mapping: bool = False, + aligned_mapping: bool = False, This parameter specifies whether the mapping of the embeddings nodes matches the internal node mapping of the given graph. If these two mappings do not match, the generated edge embedding will be meaningless. """ - self._numeric_node_ids = numeric_node_ids self._node_feature = None - self._aligned_node_mapping = aligned_node_mapping - - @property - def numeric_node_ids(self) -> bool: - """Return whether the transformer returns numeric node IDs.""" - return self._numeric_node_ids + self._node_type_feature = None + self._aligned_mapping = aligned_mapping - def fit(self, node_feature: Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]): + def fit( + self, + node_feature: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, + node_type_feature: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, + ): """Fit the model. Parameters ------------------------- - node_feature: Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]], + node_feature: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None Node feature to use to fit the transformer. + node_type_feature: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None + Node type feature to use to fit the transformer. """ - if not isinstance(node_feature, list): + if node_feature is not None and not isinstance(node_feature, list): node_feature = [node_feature] - if len(node_feature) == 0: + if node_type_feature is not None and not isinstance(node_type_feature, list): + node_type_feature = [node_type_feature] + + if (node_feature is None or len(node_feature) == 0) and (node_type_feature is None or len(node_type_feature) == 0): raise ValueError( "The provided list of features is empty!" ) # We check if any of the provided node features # is neither a numpy array nor a pandas dataframe. - for nf in node_feature: - if not isinstance(nf, (pd.DataFrame, np.ndarray)): - raise ValueError( - ( - "One of the provided node features is not " - "neither a pandas DataFrame nor a numpy array, but " - "of type {node_feature_type}. It is not clear " - "what to do with this feature." - ).format( - node_feature_type=type(node_feature) + for features, feature_name in ( + (node_type_feature, "node type"), + (node_feature, "node"), + ): + if features is None: + continue + for nf in features: + if not isinstance(nf, (pd.DataFrame, np.ndarray)): + raise ValueError( + ( + f"One of the provided {feature_name} features is not " + "neither a pandas DataFrame nor a numpy array, but " + f"of type {type(nf)}. It is not clear " + "what to do with this feature." + ) ) - ) - # We check if, while the parameters for alignment - # has not been provided, numpy arrays were provided. - # This would be an issue as we cannot check for alignment - # in numpy arrays. - if not self._aligned_node_mapping and any( - isinstance(nf, np.ndarray) - for nf in node_feature - ): - raise ValueError( - "A numpy array feature was provided while the " - "aligned node mapping parameter was set to false. " - "If you intend to specify that you are providing a numpy " - "array node feature that is aligned with the node vocabulary " - "of the graph set the `aligned_node_mapping` parameter " - "to True." - ) + # We check if, while the parameters for alignment + # has not been provided, numpy arrays were provided. + # This would be an issue as we cannot check for alignment + # in numpy arrays. + if not self._aligned_mapping and any( + isinstance(nf, np.ndarray) + for nf in features + ): + raise ValueError( + "A numpy array feature was provided while the " + f"aligned mapping parameter was set to false. " + "If you intend to specify that you are providing a numpy " + f"array {feature_name} feature that is aligned with the vocabulary " + "of the graph set the `aligned_mapping` parameter " + "to True." + ) - if self._aligned_node_mapping: - self._node_feature = np.hstack([ - nf.to_numpy() if isinstance(nf, pd.DataFrame) else nf - for nf in node_feature - ]) - if not self._node_feature.data.c_contiguous: - self._node_feature = np.ascontiguousarray(self._node_feature) + if self._aligned_mapping: + if node_feature is not None: + if len(node_feature) > 1: + self._node_feature = np.hstack([ + nf.to_numpy() if isinstance(nf, pd.DataFrame) else nf + for nf in node_feature + ]) + else: + self._node_feature = node_feature[0] + + if node_type_feature is not None: + if len(node_type_feature) > 1: + self._node_type_feature = np.hstack([ + nf.to_numpy() if isinstance(nf, pd.DataFrame) else nf + for nf in node_type_feature + ]) + else: + self._node_type_feature = node_type_feature[0] else: - self._node_feature = pd.concat(node_feature, axis=1) + if node_feature is not None: + self._node_feature = pd.concat(node_feature, axis=1) + if node_type_feature is not None: + self._node_type_feature = pd.concat(node_type_feature, axis=1) - def transform(self, nodes: Union[Graph, List[str], List[int]]) -> np.ndarray: + def transform( + self, + nodes: Optional[Union[Graph, List[str], List[int]]] = None, + node_types: Optional[Union[Graph, List[Optional[List[str]]], List[Optional[List[int]]]]] = None, + ) -> np.ndarray: """Return embeddings from given node. Parameters -------------------------- - nodes: Union[List[str], List[int]], + nodes: Optional[Union[List[str], List[int]]] = None List of nodes whose embedding is to be returned. - By default this should be a list of strings, if the - aligned_node_mapping is setted, then this methods also accepts + This can be either a list of strings, or a graph, or if the + aligned_mapping is setted, then this methods also accepts + a list of ints. + node_types: Optional[Union[Graph, List[Optional[List[str]]], List[Optional[List[int]]]]] = None, + List of node types whose embedding is to be returned. + This can be either a list of strings, or a graph, or if the + aligned_mapping is setted, then this methods also accepts a list of ints. Raises @@ -111,22 +140,68 @@ def transform(self, nodes: Union[Graph, List[str], List[int]]) -> np.ndarray: -------------------------- Numpy array of embeddings. """ - if self._node_feature is None and not self.numeric_node_ids: + if self._node_feature is None and self._node_type_feature is None: raise ValueError( "Transformer was not fitted yet." ) - if self._aligned_node_mapping: - if self.numeric_node_ids: - return nodes - if isinstance(nodes, Graph): - return self._node_feature - return self._node_feature[nodes] + node_type_features = None + node_features = None - if self.numeric_node_ids: - return self._node_feature.index.get_indexer(nodes) - - if isinstance(nodes, Graph): - nodes = nodes.get_node_names() + if self._aligned_mapping: + if nodes is not None and self._node_feature is not None: + if not isinstance(nodes, (np.ndarray, Graph)): + raise ValueError( + "The provided nodes are not numpy array and the " + "node IDs or Graph are expected to be aligned." + ) + + if isinstance(nodes, Graph): + node_features = self._node_feature + else: + node_features = self._node_feature[nodes] + + if node_types is not None and self._node_type_feature is not None: + if isinstance(node_types, Graph): + node_types = node_types.get_node_type_ids() + node_type_feature_dimensionality = self._node_type_feature.shape[1] + node_type_features = np.vstack([ + np.mean( + self._node_type_feature[node_type_ids], + axis=0 + ) + if node_type_ids is not None + else + np.zeros(shape=node_type_feature_dimensionality) + for node_type_ids in node_types + ]) + else: + if nodes is not None and self._node_feature is not None: + if isinstance(nodes, Graph): + nodes = nodes.get_node_names() + + node_features = self._node_feature.loc[nodes].to_numpy() + + if node_types is not None and self._node_type_feature is not None: + node_type_feature_dimensionality = self._node_type_feature.shape[1] + node_type_features = np.vstack([ + np.mean( + self._node_feature.loc[node_type_names].to_numpy(), + axis=0 + ) + if node_type_names is not None + else + np.zeros(shape=node_type_feature_dimensionality) + for node_type_names in node_types + ]) + + + if node_features is None: + node_features = node_type_features + elif node_type_features is not None : + node_features = np.hstack([ + node_features, + node_type_features + ]) - return self._node_feature.loc[nodes].to_numpy() + return node_features diff --git a/embiggen/layers/__init__.py b/embiggen/layers/__init__.py index e69de29b..b680692d 100644 --- a/embiggen/layers/__init__.py +++ b/embiggen/layers/__init__.py @@ -0,0 +1,1 @@ +__all__ = [] \ No newline at end of file diff --git a/embiggen/layers/tensorflow/graph_convolution_layer.py b/embiggen/layers/tensorflow/graph_convolution_layer.py index a17817b9..06a51eb0 100644 --- a/embiggen/layers/tensorflow/graph_convolution_layer.py +++ b/embiggen/layers/tensorflow/graph_convolution_layer.py @@ -7,6 +7,7 @@ """ from typing import Tuple, Union, Dict, Optional, List import tensorflow as tf +from userinput.utils import must_be_in_set from tensorflow.python.ops import embedding_ops # pylint: disable=import-error,no-name-in-module from tensorflow.keras.layers import Dropout, Layer, Dense # pylint: disable=import-error,no-name-in-module from embiggen.layers.tensorflow.l2_norm import L2Norm @@ -15,10 +16,13 @@ class GraphConvolution(Layer): """Layer implementing graph convolution layer.""" + available_combiner = ["mean", "sqrtn", "sum"] + def __init__( self, units: int, activation: str = "relu", + combiner: str = "mean", dropout_rate: Optional[float] = 0.5, apply_norm: bool = False, kwargs: Dict @@ -27,18 +31,28 @@ def __init__( Parameters ---------------------- - units: int, + units: int The dimensionality of the output space (i.e. the number of output units). - activation: str = "relu", + activation: str = "relu" Activation function to use. If you don't specify anything, relu is used. - dropout_rate: Optional[float] = 0.5, + combiner: str = "mean" + A string specifying the reduction op. + Currently "mean", "sqrtn" and "sum" are supported. + "sum" computes the weighted sum of the embedding results for each row. + "mean" is the weighted sum divided by the total weight. + "sqrtn" is the weighted sum divided by the square root of the sum of the squares of the weights. + Defaults to mean. + dropout_rate: Optional[float] = 0.5 Float between 0 and 1. Fraction of the input units to drop. If the provided value is either zero or None the dropout rate is not applied. - kwargs: Dict, + kwargs: Dict Kwargs to pass to the parent Layer class. """ super().__init__(kwargs) + combiner = must_be_in_set( + combiner, self.available_combiner, "combiner") + self._combiner = combiner self._units = units self._activation = activation if dropout_rate == 0.0: @@ -46,8 +60,8 @@ def __init__( self._dropout_rate = dropout_rate self._apply_norm = apply_norm self._dense_layers = [] - self._dropout_layer = None - self._l2_norm = None + self._dropout_layers = [] + self._l2_norms = [] def build(self, input_shape) -> None: """Build the Graph Convolution layer. @@ -70,6 +84,7 @@ def build(self, input_shape) -> None: "has a single element. It should contain exactly two elements, " "the adjacency matrix and the node features." ) + for node_feature_shape in input_shape[1:]: dense_layer = Dense( units=self._units, @@ -78,17 +93,19 @@ def build(self, input_shape) -> None: dense_layer.build(node_feature_shape) self._dense_layers.append(dense_layer) - if self._dropout_rate is not None: - self._dropout_layer = Dropout(self._dropout_rate) - self._dropout_layer.build(node_feature_shape) - else: - self._dropout_layer = lambda x: x + if self._dropout_rate is not None: + dropout_layer = Dropout(self._dropout_rate) + dropout_layer.build(node_feature_shape) + else: + def dropout_layer(x): return x + self._dropout_layers.append(dropout_layer) - if self._apply_norm: - self._l2_norm = L2Norm() - self._l2_norm.build(node_feature_shape) - else: - self._l2_norm = lambda x: x + if self._apply_norm: + l2_norm = L2Norm() + l2_norm.build((self._units,)) + else: + def l2_norm(x): return x + self._l2_norms.append(l2_norm) super().build(input_shape) @@ -111,14 +128,16 @@ def call( ) return [ - self._l2_norm(dense(embedding_ops.embedding_lookup_sparse_v2( - self._dropout_layer(node_feature), + dense(embedding_ops.embedding_lookup_sparse_v2( + node_feature, ids, adjacency, - combiner='mean' - ))) - for dense, node_feature in zip( + combiner=self._combiner + )) + for dense, dropout, l2_norm, node_feature in zip( self._dense_layers, + self._dropout_layers, + self._l2_norms, node_features ) ] diff --git a/embiggen/layers/tensorflow/l2_norm.py b/embiggen/layers/tensorflow/l2_norm.py index 2ba0c728..101fdd6d 100644 --- a/embiggen/layers/tensorflow/l2_norm.py +++ b/embiggen/layers/tensorflow/l2_norm.py @@ -6,19 +6,18 @@ class L2Norm(Layer): - """Layer implementing simple wrapped embedding layer plus a flatten.""" + """Layer implementing L2 Norm.""" def __init__( self, kwargs: Dict ): - """Create new GraphConvolution layer. - """ + """Create new L2 Norm layer.""" super().__init__(kwargs) self._norm_layer = None def build(self, input_shape) -> None: - """Build the Graph Convolution layer. + """Build the L2 Norm layer. Parameters ------------------------------ diff --git a/embiggen/node_label_prediction/node_label_prediction_model.py b/embiggen/node_label_prediction/node_label_prediction_model.py index 61a95880..0b02c4d3 100644 --- a/embiggen/node_label_prediction/node_label_prediction_model.py +++ b/embiggen/node_label_prediction/node_label_prediction_model.py @@ -4,7 +4,7 @@ import numpy as np import warnings from ensmallen import Graph -from embiggen.utils.abstract_models import AbstractClassifierModel, AbstractEmbeddingModel, abstract_class, format_list +from embiggen.utils.abstract_models import AbstractClassifierModel, abstract_class @abstract_class @@ -95,10 +95,13 @@ def split_graph_following_evaluation_schema( use_stratification="Stratified" in evaluation_schema, random_state=random_state, ) - raise ValueError( - f"The requested evaluation schema `{evaluation_schema}` " - "is not available. The available evaluation schemas " - f"are: {format_list(cls.get_available_evaluation_schemas())}." + super().split_graph_following_evaluation_schema( + graph=graph, + evaluation_schema=evaluation_schema, + random_state=random_state, + holdout_number=holdout_number, + number_of_holdouts=number_of_holdouts, + holdouts_kwargs, ) @classmethod @@ -130,7 +133,8 @@ def _evaluate( verbose: bool = True, ) -> List[Dict[str, Any]]: """Return model evaluation on the provided graphs.""" - train_size = train.get_number_of_known_node_types() / graph.get_number_of_known_node_types() + train_size = train.get_number_of_known_node_types( + ) / graph.get_number_of_known_node_types() if self.is_multilabel_prediction_task(): labels = graph.get_one_hot_encoded_node_types() @@ -295,7 +299,19 @@ def fit( "of the node-label prediction models." ) - self._is_binary_prediction_task = graph.get_number_of_node_types() == 2 + non_zero_node_types = sum([ + 1 + for count in graph.get_node_type_names_counts_hashmap().values() + if count > 0 + ]) + + if non_zero_node_types < 2: + raise ValueError( + "The provided training graph has less than two non-zero node types. " + "It is unclear how to proceeed." + ) + + self._is_binary_prediction_task = non_zero_node_types == 2 self._is_multilabel_prediction_task = graph.has_multilabel_node_types() node_type_counts = graph.get_node_type_names_counts_hashmap() diff --git a/embiggen/node_label_prediction/node_label_prediction_sklearn/decision_tree_for_node_prediction.py b/embiggen/node_label_prediction/node_label_prediction_sklearn/decision_tree_for_node_prediction.py index 94af1a26..de76f9d6 100644 --- a/embiggen/node_label_prediction/node_label_prediction_sklearn/decision_tree_for_node_prediction.py +++ b/embiggen/node_label_prediction/node_label_prediction_sklearn/decision_tree_for_node_prediction.py @@ -11,15 +11,14 @@ def __init__( self, criterion="gini", splitter="best", - max_depth=None, + max_depth=10, min_samples_split=2, min_samples_leaf=1, min_weight_fraction_leaf=0., max_features=None, max_leaf_nodes=None, min_impurity_decrease=0., - min_impurity_split=None, - class_weight=None, + class_weight="balanced", ccp_alpha=0.0, random_state: int = 42 ): @@ -33,7 +32,6 @@ def __init__( self._max_features = max_features self._max_leaf_nodes = max_leaf_nodes self._min_impurity_decrease = min_impurity_decrease - self._min_impurity_split = min_impurity_split self._random_state = random_state self._class_weight = class_weight self._ccp_alpha = ccp_alpha @@ -49,7 +47,6 @@ def __init__( max_features=max_features, max_leaf_nodes=max_leaf_nodes, min_impurity_decrease=min_impurity_decrease, - min_impurity_split=min_impurity_split, random_state=random_state, class_weight=class_weight, ccp_alpha=ccp_alpha, @@ -71,7 +68,6 @@ def parameters(self) -> Dict[str, Any]: max_features=self._max_features, max_leaf_nodes=self._max_leaf_nodes, min_impurity_decrease=self._min_impurity_decrease, - min_impurity_split=self._min_impurity_split, random_state=self._random_state, class_weight=self._class_weight, ccp_alpha=self._ccp_alpha, diff --git a/embiggen/node_label_prediction/node_label_prediction_sklearn/extra_trees_node_label_prediction.py b/embiggen/node_label_prediction/node_label_prediction_sklearn/extra_trees_node_label_prediction.py index 7cbed2ed..a0a9c328 100644 --- a/embiggen/node_label_prediction/node_label_prediction_sklearn/extra_trees_node_label_prediction.py +++ b/embiggen/node_label_prediction/node_label_prediction_sklearn/extra_trees_node_label_prediction.py @@ -19,13 +19,12 @@ def __init__( max_features="auto", max_leaf_nodes=None, min_impurity_decrease=0., - min_impurity_split=None, bootstrap=True, oob_score=False, n_jobs=-1, verbose=0, warm_start=False, - class_weight=None, + class_weight="balanced", ccp_alpha=0.0, max_samples=None, random_state: int = 42 @@ -40,7 +39,6 @@ def __init__( self._max_features = max_features self._max_leaf_nodes = max_leaf_nodes self._min_impurity_decrease = min_impurity_decrease - self._min_impurity_split = min_impurity_split self._bootstrap = bootstrap self._oob_score = oob_score self._n_jobs = n_jobs @@ -61,7 +59,6 @@ def __init__( max_features=max_features, max_leaf_nodes=max_leaf_nodes, min_impurity_decrease=min_impurity_decrease, - min_impurity_split=min_impurity_split, bootstrap=bootstrap, oob_score=oob_score, n_jobs=n_jobs, @@ -97,7 +94,6 @@ def parameters(self) -> Dict[str, Any]: max_features = self._max_features, max_leaf_nodes = self._max_leaf_nodes, min_impurity_decrease = self._min_impurity_decrease, - min_impurity_split = self._min_impurity_split, bootstrap = self._bootstrap, oob_score = self._oob_score, n_jobs = self._n_jobs, diff --git a/embiggen/node_label_prediction/node_label_prediction_sklearn/gradient_boosting_node_label_prediction.py b/embiggen/node_label_prediction/node_label_prediction_sklearn/gradient_boosting_node_label_prediction.py index 5015f219..697ae628 100644 --- a/embiggen/node_label_prediction/node_label_prediction_sklearn/gradient_boosting_node_label_prediction.py +++ b/embiggen/node_label_prediction/node_label_prediction_sklearn/gradient_boosting_node_label_prediction.py @@ -19,7 +19,6 @@ def __init__( min_weight_fraction_leaf=0., max_depth=3, min_impurity_decrease=0., - min_impurity_split=None, init=None, max_features=None, verbose=0, @@ -34,20 +33,19 @@ def __init__( """Create the Gradient Boosting for node label Prediction.""" self._loss=loss self._learning_rate=learning_rate - self._n_estimators=n_estimators, + self._n_estimators=n_estimators self._criterion=criterion - self._min_samples_split=min_samples_split, - self._min_samples_leaf=min_samples_leaf, - self._min_weight_fraction_leaf=min_weight_fraction_leaf, + self._min_samples_split=min_samples_split + self._min_samples_leaf=min_samples_leaf + self._min_weight_fraction_leaf=min_weight_fraction_leaf self._max_depth=max_depth self._init=init - self._subsample=subsample, - self._max_features=max_features, - self._max_leaf_nodes=max_leaf_nodes, - self._min_impurity_decrease=min_impurity_decrease, - self._min_impurity_split=min_impurity_split, + self._subsample=subsample + self._max_features=max_features + self._max_leaf_nodes=max_leaf_nodes + self._min_impurity_decrease=min_impurity_decrease self._warm_start=warm_start - self._validation_fraction=validation_fraction, + self._validation_fraction=validation_fraction self._n_iter_no_change=n_iter_no_change self._tol=tol self._ccp_alpha=ccp_alpha @@ -62,7 +60,6 @@ def __init__( random_state=random_state, verbose=verbose, max_leaf_nodes=max_leaf_nodes, min_impurity_decrease=min_impurity_decrease, - min_impurity_split=min_impurity_split, warm_start=warm_start, validation_fraction=validation_fraction, n_iter_no_change=n_iter_no_change, tol=tol, ccp_alpha=ccp_alpha ), @@ -71,30 +68,27 @@ def __init__( def parameters(self) -> Dict[str, Any]: """Returns parameters used for this model.""" - return { + return dict( super().parameters(), - dict( - loss=self._loss, - learning_rate=self._learning_rate, - n_estimators=self._n_estimators, - criterion=self._criterion, - min_samples_split=self._min_samples_split, - min_samples_leaf=self._min_samples_leaf, - min_weight_fraction_leaf=self._min_weight_fraction_leaf, - max_depth=self._max_depth, - init=self._init, - subsample=self._subsample, - max_features=self._max_features, - max_leaf_nodes=self._max_leaf_nodes, - min_impurity_decrease=self._min_impurity_decrease, - min_impurity_split=self._min_impurity_split, - warm_start=self._warm_start, - validation_fraction=self._validation_fraction, - n_iter_no_change=self._n_iter_no_change, - tol=self._tol, - ccp_alpha=self._ccp_alpha - ) - } + loss=self._loss, + learning_rate=self._learning_rate, + n_estimators=self._n_estimators, + criterion=self._criterion, + min_samples_split=self._min_samples_split, + min_samples_leaf=self._min_samples_leaf, + min_weight_fraction_leaf=self._min_weight_fraction_leaf, + max_depth=self._max_depth, + init=self._init, + subsample=self._subsample, + max_features=self._max_features, + max_leaf_nodes=self._max_leaf_nodes, + min_impurity_decrease=self._min_impurity_decrease, + warm_start=self._warm_start, + validation_fraction=self._validation_fraction, + n_iter_no_change=self._n_iter_no_change, + tol=self._tol, + ccp_alpha=self._ccp_alpha + ) @classmethod def smoke_test_parameters(cls) -> Dict[str, Any]: diff --git a/embiggen/node_label_prediction/node_label_prediction_sklearn/random_forest_node_label_prediction.py b/embiggen/node_label_prediction/node_label_prediction_sklearn/random_forest_node_label_prediction.py index c8235af5..d00a77ab 100644 --- a/embiggen/node_label_prediction/node_label_prediction_sklearn/random_forest_node_label_prediction.py +++ b/embiggen/node_label_prediction/node_label_prediction_sklearn/random_forest_node_label_prediction.py @@ -17,16 +17,15 @@ def __init__( min_samples_split: int = 2, min_samples_leaf: int = 1, min_weight_fraction_leaf: float = 0., - max_features="auto", + max_features="sqrt", max_leaf_nodes=None, min_impurity_decrease=0., - min_impurity_split=None, bootstrap=True, oob_score=False, n_jobs=-1, verbose=0, warm_start=False, - class_weight=None, + class_weight="balanced", ccp_alpha=0.0, max_samples=None, random_state: int = 42 @@ -41,7 +40,6 @@ def __init__( self._max_features = max_features self._max_leaf_nodes = max_leaf_nodes self._min_impurity_decrease = min_impurity_decrease - self._min_impurity_split = min_impurity_split self._bootstrap = bootstrap self._oob_score = oob_score self._n_jobs = cpu_count() if n_jobs == -1 else n_jobs @@ -62,7 +60,6 @@ def __init__( max_features=max_features, max_leaf_nodes=max_leaf_nodes, min_impurity_decrease=min_impurity_decrease, - min_impurity_split=min_impurity_split, bootstrap=bootstrap, oob_score=oob_score, n_jobs=n_jobs, @@ -98,7 +95,6 @@ def parameters(self) -> Dict[str, Any]: max_features = self._max_features, max_leaf_nodes = self._max_leaf_nodes, min_impurity_decrease = self._min_impurity_decrease, - min_impurity_split = self._min_impurity_split, bootstrap = self._bootstrap, oob_score = self._oob_score, n_jobs = self._n_jobs, diff --git a/embiggen/node_label_prediction/node_label_prediction_sklearn/sklearn_node_label_prediction_adapter.py b/embiggen/node_label_prediction/node_label_prediction_sklearn/sklearn_node_label_prediction_adapter.py index d2c5375e..9b73f9f4 100644 --- a/embiggen/node_label_prediction/node_label_prediction_sklearn/sklearn_node_label_prediction_adapter.py +++ b/embiggen/node_label_prediction/node_label_prediction_sklearn/sklearn_node_label_prediction_adapter.py @@ -64,7 +64,7 @@ def _trasform_graph_into_node_embedding( ValueError If the two graphs do not share the same node vocabulary. """ - gt = NodeTransformer(aligned_node_mapping=True) + gt = NodeTransformer(aligned_mapping=True) gt.fit(node_features) return gt.transform(graph,) @@ -103,7 +103,7 @@ def _fit( If the two graphs do not share the same node vocabulary. """ nlpt = NodeLabelPredictionTransformer( - aligned_node_mapping=True + aligned_mapping=True ) nlpt.fit(node_features) @@ -155,10 +155,18 @@ def _predict_proba( )) if self.is_multilabel_prediction_task(): - return np.array([ - class_predictions[:, 1] - for class_predictions in predictions_probabilities - ]).T + if isinstance(predictions_probabilities, np.ndarray): + return predictions_probabilities + if isinstance(predictions_probabilities, list): + return np.array([ + class_predictions[:, 1] + for class_predictions in predictions_probabilities + ]).T + raise NotImplementedError( + f"The model {self.model_name()} from library {self.library_name()} " + f"returned an object of type {type(predictions_probabilities)} during " + "the execution of the predict proba method." + ) return predictions_probabilities diff --git a/embiggen/node_label_prediction/node_label_prediction_tensorflow/gcn.py b/embiggen/node_label_prediction/node_label_prediction_tensorflow/gcn.py index fdbefbf7..326b7e7a 100644 --- a/embiggen/node_label_prediction/node_label_prediction_tensorflow/gcn.py +++ b/embiggen/node_label_prediction/node_label_prediction_tensorflow/gcn.py @@ -28,6 +28,7 @@ def __init__( number_of_units_per_head_layer: Union[int, List[int]] = 128, dropout_rate: float = 0.1, apply_norm: bool = False, + combiner: str = "mean", optimizer: Union[str, Optimizer] = "adam", early_stopping_min_delta: float = 0.0001, early_stopping_patience: int = 30, @@ -68,6 +69,13 @@ def __init__( apply_norm: bool = False Whether to normalize the output of the convolution operations, after applying the level activations. + combiner: str = "mean" + A string specifying the reduction op. + Currently "mean", "sqrtn" and "sum" are supported. + "sum" computes the weighted sum of the embedding results for each row. + "mean" is the weighted sum divided by the total weight. + "sqrtn" is the weighted sum divided by the square root of the sum of the squares of the weights. + Defaults to mean. optimizer: str = "Adam" The optimizer to use while training the model. early_stopping_min_delta: float @@ -126,6 +134,7 @@ def __init__( number_of_units_per_graph_convolution_layers=number_of_units_per_graph_convolution_layers, dropout_rate=dropout_rate, apply_norm=apply_norm, + combiner=combiner, optimizer=optimizer, early_stopping_min_delta=early_stopping_min_delta, early_stopping_patience=early_stopping_patience, @@ -194,6 +203,8 @@ def _build_model( def get_output_classes(self, graph: Graph) -> int: """Returns number of output classes.""" + if self.is_binary_prediction_task(): + return 1 return graph.get_number_of_node_types() def _get_class_weights(self, graph: Graph) -> Dict[int, float]: @@ -249,7 +260,7 @@ def _get_model_training_sample_weights( graph: Graph, ) -> Optional[np.ndarray]: """Returns training output tuple.""" - return graph.get_known_node_types_mask().astype(tf.float32) + return graph.get_known_node_types_mask().astype(np.float32) def _get_model_prediction_input( self, diff --git a/embiggen/node_label_prediction/node_label_prediction_tensorflow/gnn.py b/embiggen/node_label_prediction/node_label_prediction_tensorflow/gnn.py index 82b18338..ca0aae1c 100644 --- a/embiggen/node_label_prediction/node_label_prediction_tensorflow/gnn.py +++ b/embiggen/node_label_prediction/node_label_prediction_tensorflow/gnn.py @@ -12,7 +12,6 @@ def __init__( epochs: int = 1000, number_of_head_layers: int = 1, number_of_units_per_head_layer: Union[int, List[int]] = 128, - dropout_rate: float = 0.1, optimizer: Union[str, Optimizer] = "adam", early_stopping_min_delta: float = 0.0001, early_stopping_patience: int = 30, @@ -45,9 +44,6 @@ def __init__( Number of graph convolution layer. number_of_units_per_hidden_layer: Union[int, List[int]] = 128 Number of units per hidden layer. - dropout_rate: float = 0.3 - Float between 0 and 1. - Fraction of the input units to dropout. optimizer: str = "Adam" The optimizer to use while training the model. early_stopping_min_delta: float @@ -95,7 +91,6 @@ def __init__( number_of_head_layers=number_of_head_layers, number_of_units_per_graph_convolution_layers=0, number_of_units_per_head_layer=number_of_units_per_head_layer, - dropout_rate=dropout_rate, optimizer=optimizer, early_stopping_min_delta=early_stopping_min_delta, early_stopping_patience=early_stopping_patience, @@ -138,13 +133,16 @@ def parameters(self) -> Dict[str, Any]: removed = [ "number_of_units_per_graph_convolution_layers", "handling_multi_graph", - "number_of_units_per_head_layer" + "number_of_units_per_head_layer", + "combiner", + "apply_norm", + "dropout_rate" ] return dict( number_of_units_per_head_layer=self._number_of_units_per_head_layer, { key: value - for key, value in super().smoke_test_parameters().items() + for key, value in super().parameters().items() if key not in removed } ) diff --git a/embiggen/node_label_prediction/node_label_prediction_tensorflow/graph_sage.py b/embiggen/node_label_prediction/node_label_prediction_tensorflow/graph_sage.py index 317e8207..c9763e18 100644 --- a/embiggen/node_label_prediction/node_label_prediction_tensorflow/graph_sage.py +++ b/embiggen/node_label_prediction/node_label_prediction_tensorflow/graph_sage.py @@ -110,6 +110,7 @@ def __init__( number_of_units_per_head_layer=number_of_units_per_head_layer, dropout_rate=dropout_rate, apply_norm=True, + combiner="mean", optimizer=optimizer, early_stopping_min_delta=early_stopping_min_delta, early_stopping_patience=early_stopping_patience, @@ -135,6 +136,7 @@ def parameters(self) -> Dict[str, Any]: """Returns parameters for smoke test.""" removed = [ "apply_norm", + "combiner", "use_simmetric_normalized_laplacian" ] return dict( diff --git a/embiggen/node_label_prediction/node_label_prediction_tensorflow/kipf_gcn.py b/embiggen/node_label_prediction/node_label_prediction_tensorflow/kipf_gcn.py index 0c7affa4..a74bdd69 100644 --- a/embiggen/node_label_prediction/node_label_prediction_tensorflow/kipf_gcn.py +++ b/embiggen/node_label_prediction/node_label_prediction_tensorflow/kipf_gcn.py @@ -1,5 +1,5 @@ """GCN model for node-label prediction.""" -from typing import List, Union, Optional +from typing import List, Union, Optional, Dict, Any from tensorflow.keras.optimizers import \ Optimizer # pylint: disable=import-error,no-name-in-module @@ -111,6 +111,7 @@ def __init__( number_of_units_per_head_layer=number_of_units_per_head_layer, dropout_rate=dropout_rate, apply_norm=apply_norm, + combiner="sum", optimizer=optimizer, early_stopping_min_delta=early_stopping_min_delta, early_stopping_patience=early_stopping_patience, @@ -132,6 +133,19 @@ def __init__( verbose=verbose, ) + def parameters(self) -> Dict[str, Any]: + """Returns parameters for smoke test.""" + removed = [ + "combiner", + ] + return dict( + { + key: value + for key, value in super().parameters().items() + if key not in removed + } + ) + @classmethod def model_name(cls) -> str: return "Kipf GCN" \ No newline at end of file diff --git a/embiggen/sequences/__init__.py b/embiggen/sequences/__init__.py index 6bb0d6ae..8ed70a22 100644 --- a/embiggen/sequences/__init__.py +++ b/embiggen/sequences/__init__.py @@ -1,1 +1,2 @@ -"""Module with Sequences, mostly for Keras models.""" \ No newline at end of file +"""Module with Sequences, mostly for Keras models.""" +__all__ = [] \ No newline at end of file diff --git a/embiggen/sequences/generic_sequences/edge_prediction_sequence.py b/embiggen/sequences/generic_sequences/edge_prediction_sequence.py index da5c5a61..108ad7e1 100644 --- a/embiggen/sequences/generic_sequences/edge_prediction_sequence.py +++ b/embiggen/sequences/generic_sequences/edge_prediction_sequence.py @@ -35,14 +35,6 @@ def __init__( batch_size: int = 210, The batch size to use. """ - if not graph.has_edges(): - raise ValueError( - f"An empty instance of graph {graph.get_name()} was provided!" - ) - if not graph.has_edges(): - raise ValueError( - f"An empty instance of graph {graph_used_in_training.get_name()} was provided!" - ) if not graph.has_compatible_node_vocabularies(graph_used_in_training): raise ValueError( f"The provided graph {graph.get_name()} does not have a node vocabulary " diff --git a/embiggen/sequences/tensorflow_sequences/__init__.py b/embiggen/sequences/tensorflow_sequences/__init__.py index bbfdb2cc..b4843399 100644 --- a/embiggen/sequences/tensorflow_sequences/__init__.py +++ b/embiggen/sequences/tensorflow_sequences/__init__.py @@ -4,19 +4,13 @@ from embiggen.sequences.tensorflow_sequences.gcn_edge_prediction_training_sequence import GCNEdgePredictionTrainingSequence from embiggen.sequences.tensorflow_sequences.gcn_edge_prediction_sequence import GCNEdgePredictionSequence from embiggen.sequences.tensorflow_sequences.gcn_edge_label_prediction_training_sequence import GCNEdgeLabelPredictionTrainingSequence -from embiggen.sequences.tensorflow_sequences.edge_prediction_sequence import EdgePredictionSequence from embiggen.sequences.tensorflow_sequences.siamese_sequence import SiameseSequence -from embiggen.sequences.tensorflow_sequences.kgsiamese_sequence import KGSiameseSequence __all__ = [ "Node2VecSequence", "EdgePredictionTrainingSequence", - "EdgePredictionSequence", "GCNEdgePredictionTrainingSequence", "GCNEdgePredictionSequence", "GCNEdgeLabelPredictionTrainingSequence", "SiameseSequence", - "KGSiameseSequence", - "EdgeJaccardSequence", - "EdgeAdamicAdarSequence", ] diff --git a/embiggen/sequences/tensorflow_sequences/edge_prediction_sequence.py b/embiggen/sequences/tensorflow_sequences/edge_prediction_sequence.py deleted file mode 100644 index 6145f82f..00000000 --- a/embiggen/sequences/tensorflow_sequences/edge_prediction_sequence.py +++ /dev/null @@ -1,147 +0,0 @@ -"""Keras Sequence for running Neural Network on graph edge prediction.""" -from typing import Tuple - -import numpy as np -import tensorflow as tf -from ensmallen import Graph # pylint: disable=no-name-in-module -from keras_mixed_sequence import Sequence -from embiggen.sequences.generic_sequences import EdgePredictionSequence as GenericEdgePredictionSequence -from embiggen.utils.tensorflow_utils import tensorflow_version_is_higher_or_equal_than - - -class EdgePredictionSequence(Sequence): - """Keras Sequence for running Neural Network on graph edge prediction.""" - - def __init__( - self, - graph: Graph, - graph_used_in_training: Graph, - use_node_types: bool, - use_edge_metrics: bool, - batch_size: int = 210, - ): - """Create new EdgePredictionSequence object. - - Parameters - -------------------------------- - graph: Graph - The graph whose edges are to be predicted. - graph_used_in_training: Graph - The graph that was used while training the current - edge prediction model. - use_node_types: bool - Whether to return the node types. - use_edge_metrics: bool = False - Whether to return the edge metrics. - batch_size: int = 210, - The batch size to use. - """ - self._sequence = GenericEdgePredictionSequence( - graph=graph, - graph_used_in_training=graph_used_in_training, - use_node_types=use_node_types, - use_edge_metrics=use_edge_metrics, - batch_size=batch_size - ) - self._current_index = 0 - super().__init__( - sample_number=graph.get_number_of_directed_edges(), - batch_size=batch_size, - ) - - def __call__(self): - """Return next batch using an infinite generator model.""" - self._current_index += 1 - return (self[self._current_index],) - - def into_dataset(self) -> tf.data.Dataset: - """Return dataset generated out of the current sequence instance. - - Implementative details - --------------------------------- - This method handles the conversion of this Keras Sequence into - a TensorFlow dataset, also handling the proper dispatching according - to what version of TensorFlow is installed in this system. - - Returns - ---------------------------------- - Dataset to be used for the training of a model - """ - - ################################################################# - # Handling kernel creation when TensorFlow is a modern version. # - ################################################################# - - if tensorflow_version_is_higher_or_equal_than("2.5.0"): - input_tensor_specs = [] - - for _ in range(2): - # Shapes of the source and destination node IDs - input_tensor_specs.append(tf.TensorSpec( - shape=(None, ), - dtype=tf.int32 - )) - - if self._use_node_types: - # Shapes of the source and destination node type IDs - input_tensor_specs.append(tf.TensorSpec( - shape=(None, - self._graph.get_maximum_multilabel_count()), - dtype=tf.int32 - )) - - if self._use_edge_metrics: - # Shapes of the edge type IDs - input_tensor_specs.append(tf.TensorSpec( - shape=(None, - self._graph.get_number_of_available_edge_metrics()), - dtype=tf.float32 - )) - - return tf.data.Dataset.from_generator( - self, - output_signature=( - tuple(input_tensor_specs), - ) - ) - - input_tensor_types = [] - input_tensor_shapes = [] - - for _ in range(2): - input_tensor_types.append(tf.int32,) - input_tensor_shapes.append(tf.TensorShape([None, ]),) - - if self._use_node_types: - input_tensor_types.append(tf.int32,) - input_tensor_shapes.append( - tf.TensorShape([None, self._graph.get_maximum_multilabel_count()]),) - - if self._use_edge_metrics: - input_tensor_types.append(tf.float32,) - input_tensor_shapes.append(tf.TensorShape( - [None, self._graph.get_number_of_available_edge_metrics()]),) - - return tf.data.Dataset.from_generator( - self, - output_types=( - tuple(input_tensor_types), - ), - output_shapes=( - tuple(input_tensor_shapes), - ) - ) - - def __getitem__(self, idx: int) -> Tuple[np.ndarray, np.ndarray]: - """Return batch corresponding to given index. - - Parameters - --------------- - idx: int, - Index corresponding to batch to be returned. - - Returns - --------------- - Return Tuple containing X and Y numpy arrays corresponding to given batch index. - """ - return self._sequence[idx] \ No newline at end of file diff --git a/embiggen/sequences/tensorflow_sequences/edge_prediction_training_sequence.py b/embiggen/sequences/tensorflow_sequences/edge_prediction_training_sequence.py index 65b39c94..0cbc9d53 100644 --- a/embiggen/sequences/tensorflow_sequences/edge_prediction_training_sequence.py +++ b/embiggen/sequences/tensorflow_sequences/edge_prediction_training_sequence.py @@ -56,10 +56,6 @@ def __init__( random_state: int = 42, The random_state to use to make extraction reproducible. """ - if not graph.has_edges(): - raise ValueError( - f"An empty instance of graph {graph.get_name()} was provided!" - ) self._graph = graph self._negative_samples_rate = negative_samples_rate self._avoid_false_negatives = avoid_false_negatives @@ -74,101 +70,6 @@ def __init__( batch_size=batch_size, ) - def __call__(self): - """Return next batch using an infinite generator model.""" - self._current_index += 1 - return (self[self._current_index],) - - def into_dataset(self) -> tf.data.Dataset: - """Return dataset generated out of the current sequence instance. - - Implementative details - --------------------------------- - This method handles the conversion of this Keras Sequence into - a TensorFlow dataset, also handling the proper dispatching according - to what version of TensorFlow is installed in this system. - - Returns - ---------------------------------- - Dataset to be used for the training of a model - """ - - ################################################################# - # Handling kernel creation when TensorFlow is a modern version. # - ################################################################# - - if tensorflow_version_is_higher_or_equal_than("2.5.0"): - input_tensor_specs = [] - - for _ in range(2): - # Shapes of the source and destination node IDs - input_tensor_specs.append(tf.TensorSpec( - shape=(self._batch_size, ), - dtype=tf.int32 - )) - - if self._use_node_types: - # Shapes of the source and destination node type IDs - input_tensor_specs.append(tf.TensorSpec( - shape=(self._batch_size, - self._graph.get_maximum_multilabel_count()), - dtype=tf.int32 - )) - - if self._use_edge_metrics: - # Shapes of the edge type IDs - input_tensor_specs.append(tf.TensorSpec( - shape=(self._batch_size, - self._graph.get_number_of_available_edge_metrics()), - dtype=tf.float64 - )) - - return tf.data.Dataset.from_generator( - self, - output_signature=( - ( - input_tensor_specs, - ), - tf.TensorSpec( - shape=(self._batch_size,), - dtype=tf.bool - ) - ) - ) - - input_tensor_types = [] - input_tensor_shapes = [] - - for _ in range(2): - input_tensor_types.append(tf.int32,) - input_tensor_shapes.append(tf.TensorShape([self._batch_size, ]),) - - if self._use_node_types: - input_tensor_types.append(tf.int32,) - input_tensor_shapes.append( - tf.TensorShape([self._batch_size, self._graph.get_maximum_multilabel_count()]),) - - if self._use_edge_metrics: - input_tensor_types.append(tf.float64,) - input_tensor_shapes.append(tf.TensorShape( - [self._batch_size, self._graph.get_number_of_available_edge_metrics()]),) - - return tf.data.Dataset.from_generator( - self, - output_types=( - ( - input_tensor_types, - ), - tf.bool - ), - output_shapes=( - ( - input_tensor_shapes, - ), - tf.TensorShape([self._batch_size, ]), - ) - ) - def __getitem__(self, idx: int) -> Tuple[np.ndarray, np.ndarray]: """Return batch corresponding to given index. diff --git a/embiggen/sequences/tensorflow_sequences/gcn_edge_label_prediction_training_sequence.py b/embiggen/sequences/tensorflow_sequences/gcn_edge_label_prediction_training_sequence.py index 67d1c600..e1ce3577 100644 --- a/embiggen/sequences/tensorflow_sequences/gcn_edge_label_prediction_training_sequence.py +++ b/embiggen/sequences/tensorflow_sequences/gcn_edge_label_prediction_training_sequence.py @@ -49,7 +49,11 @@ def __init__( When the graph has multilabel node types, we will average the features. edge_features: Optional[List[np.ndarray]] = None, - + The edge features to be used. + For instance, these could be BERT embeddings of the + description of the edges. + When the graph has multilabel edges, + we will average the features. use_edge_metrics: bool = False Whether to return the edge metrics. """ @@ -74,27 +78,10 @@ def __init__( # The index in the returned sequence that contains the # edge label is 2 (source and destination nodes). - self._edge_label_index = 2 - - def __call__(self): - """Return next batch using an infinite generator model.""" - self._current_index += 1 - return (self[self._current_index],) - - def into_dataset(self) -> tf.data.Dataset: - """Return dataset generated out of the current sequence instance. - - Implementative details - --------------------------------- - This method handles the conversion of this Keras Sequence into - a TensorFlow dataset, also handling the proper dispatching according - to what version of TensorFlow is installed in this system. - - Returns - ---------------------------------- - Dataset to be used for the training of a model - """ - raise NotImplementedError("TODO!") + if return_node_types: + self._edge_label_index = 4 + else: + self._edge_label_index = 2 def __getitem__(self, idx: int): """Return batch corresponding to given index. diff --git a/embiggen/sequences/tensorflow_sequences/gcn_edge_prediction_sequence.py b/embiggen/sequences/tensorflow_sequences/gcn_edge_prediction_sequence.py index 7418ad20..c7d1ceef 100644 --- a/embiggen/sequences/tensorflow_sequences/gcn_edge_prediction_sequence.py +++ b/embiggen/sequences/tensorflow_sequences/gcn_edge_prediction_sequence.py @@ -54,34 +54,43 @@ def __init__( use_edge_metrics: bool = False Whether to return the edge metrics. """ - if not graph.has_edges(): + if ( + return_node_types or + node_type_features is not None + ) and graph.has_unknown_node_types(): raise ValueError( - f"An empty instance of graph {graph.get_name()} was provided!" + f"The provided graph {graph.get_name()} " + "contains unknown node types but node types " + "have been requested for the sequence." ) + self._graph = graph self._kernel = kernel if node_features is None: node_features = [] - self._node_features = node_features + self._node_features = [ + node_feature.astype(np.float32) + for node_feature in node_features + ] if return_node_ids: self._node_ids = graph.get_node_ids() else: self._node_ids = None - + if return_node_types or node_type_features is not None: if graph.has_multilabel_node_types(): node_types = graph.get_one_hot_encoded_node_types() else: node_types = graph.get_single_label_node_type_ids() - + self._node_types = node_types if return_node_types else None if node_type_features is not None: if self._graph.has_multilabel_node_types(): self._node_type_features = [] - minus_node_types = node_type_feature[node_types - 1] - node_types_mask = node_types==0 - for node_type_feature in self._node_type_features: + minus_node_types = node_types - 1 + node_types_mask = node_types == 0 + for node_type_feature in node_type_features: self._node_type_features.append(np.ma.array( node_type_feature[minus_node_types], mask=np.repeat( @@ -91,22 +100,12 @@ def __init__( node_types.shape, node_type_feature.shape[1] )) - ).mean(axis=-1).data) + ).mean(axis=-2).data) else: - if self._graph.has_unknown_node_types(): - self._node_type_features = [] - minus_node_types = node_type_feature[node_types - 1] - node_types_mask = node_types==0 - for node_type_feature in self._node_type_features: - ntf = node_type_feature[minus_node_types] - # Masking the unknown values to zero. - ntf[node_types_mask] = 0.0 - self._node_type_features.append(ntf) - else: - self._node_type_features = [ - node_type_feature[node_types] - for node_type_feature in node_type_features - ] + self._node_type_features = [ + node_type_feature[node_types] + for node_type_feature in node_type_features + ] else: self._node_type_features = [] @@ -119,11 +118,15 @@ def __init__( batch_size=support.get_number_of_nodes() ) - self._edge_features_sequence = None if edge_features is None else VectorSequence( - edge_features, - batch_size=graph.get_number_of_nodes(), - shuffle=False - ) + self._edge_features = edge_features + self._edge_features_sequences = None if edge_features is None else [ + VectorSequence( + edge_feature, + batch_size=graph.get_number_of_nodes(), + shuffle=False + ) + for edge_feature in edge_features + ] self._use_edge_metrics = use_edge_metrics self._current_index = 0 @@ -132,78 +135,6 @@ def __init__( batch_size=graph.get_number_of_nodes(), ) - def __call__(self): - """Return next batch using an infinite generator model.""" - self._current_index += 1 - return (self[self._current_index],) - - def into_dataset(self) -> tf.data.Dataset: - """Return dataset generated out of the current sequence instance. - - Implementative details - --------------------------------- - This method handles the conversion of this Keras Sequence into - a TensorFlow dataset, also handling the proper dispatching according - to what version of TensorFlow is installed in this system. - - Returns - ---------------------------------- - Dataset to be used for the training of a model - """ - - input_tensor_specs = [] - - for _ in range(2): - # Shapes of the source and destination node IDs - input_tensor_specs.append(tf.TensorSpec( - shape=(None, ), - dtype=tf.int32 - )) - - if self._use_edge_metrics: - # Shapes of the edge type IDs - input_tensor_specs.append(tf.TensorSpec( - shape=(None, self._graph.get_number_of_available_edge_metrics()), - dtype=tf.float64 - )) - - # Shapes of the kernel - input_tensor_specs.append(tf.SparseTensorSpec( - shape=(None, None), - dtype=tf.float32 - )) - - if self._node_features is not None: - input_tensor_specs.extend([ - tf.TensorSpec( - shape=(None, node_feature.shape[1]), - dtype=tf.float32 - ) - for node_feature in self._node_features - ]) - - input_tensor_specs.extend([ - tf.TensorSpec( - shape=(None, node_type_feature.shape[1]), - dtype=tf.float32 - ) - for node_type_feature in self._node_type_features - ]) - - return tf.data.Dataset.from_generator( - self, - output_signature=( - ( - input_tensor_specs, - ), - tf.TensorSpec( - shape=(None,), - dtype=tf.bool - ) - ) - ) - - def __getitem__(self, idx: int) -> Tuple[np.ndarray, np.ndarray]: """Return batch corresponding to given index. @@ -217,27 +148,37 @@ def __getitem__(self, idx: int) -> Tuple[np.ndarray, np.ndarray]: Return Tuple containing X and Y numpy arrays corresponding to given batch index. """ values = self._sequence[idx][0] - edge_features = None if self._edge_features_sequence is None else self._edge_features_sequence[idx] + edge_features = None if self._edge_features_sequences is None else [ + edge_features_sequence[idx] + for edge_features_sequence in self._edge_features_sequences + ] # If necessary, we add the padding as the last batch may be # smaller than the required size (number of nodes). delta = self.batch_size - values[0].shape[0] if delta > 0: values = [ - np.pad(value, (0, delta) if len(value.shape) == 1 else [(0, delta), (0, 0)]) + np.pad(value, (0, delta) if len(value.shape) + == 1 else [(0, delta), (0, 0)]) for value in values if value is not None ] if edge_features is not None: - edge_features = np.pad(edge_features, [(0, delta), (0, 0)]) + edge_features = [ + np.pad(edge_feature, [(0, delta), (0, 0)]) + for edge_feature in edge_features + ] + + if edge_features is None: + edge_features = [] return (tuple([ value for value in ( values, + edge_features, self._kernel, self._node_features, self._node_type_features, - edge_features, self._node_ids, self._node_types, ) diff --git a/embiggen/sequences/tensorflow_sequences/gcn_edge_prediction_training_sequence.py b/embiggen/sequences/tensorflow_sequences/gcn_edge_prediction_training_sequence.py index 535cd19a..aa9bb2ca 100644 --- a/embiggen/sequences/tensorflow_sequences/gcn_edge_prediction_training_sequence.py +++ b/embiggen/sequences/tensorflow_sequences/gcn_edge_prediction_training_sequence.py @@ -87,11 +87,13 @@ def __init__( if node_features is None: node_features = [] + self._node_features = node_features self._negative_samples_rate = negative_samples_rate self._avoid_false_negatives = avoid_false_negatives self._graph_to_avoid = graph_to_avoid self._random_state = random_state + if return_node_ids: self._node_ids = graph.get_node_ids() else: @@ -105,16 +107,12 @@ def __init__( self._node_types = node_types if return_node_types else None - if node_type_features is not None: - if graph.has_multilabel_node_types(): - node_types = graph.get_one_hot_encoded_node_types() - else: - node_types = graph.get_single_label_node_type_ids() + if node_type_features is not None: if self._graph.has_multilabel_node_types(): self._node_type_features = [] - minus_node_types = node_type_feature[node_types - 1] node_types_mask = node_types==0 - for node_type_feature in self._node_type_features: + minus_node_types = node_types - 1 + for node_type_feature in node_type_features: self._node_type_features.append(np.ma.array( node_type_feature[minus_node_types], mask=np.repeat( @@ -124,13 +122,14 @@ def __init__( node_types.shape, node_type_feature.shape[1] )) - ).mean(axis=-1).data) + ).mean(axis=-2).data) else: if self._graph.has_unknown_node_types(): self._node_type_features = [] - minus_node_types = node_type_feature[node_types - 1] node_types_mask = node_types==0 - for node_type_feature in self._node_type_features: + minus_node_types = node_types - 1 + minus_node_types[node_types_mask] = 0 + for node_type_feature in node_type_features: ntf = node_type_feature[minus_node_types] # Masking the unknown values to zero. ntf[node_types_mask] = 0.0 @@ -151,78 +150,6 @@ def __init__( batch_size=graph.get_number_of_nodes(), ) - def __call__(self): - """Return next batch using an infinite generator model.""" - self._current_index += 1 - return (self[self._current_index],) - - def into_dataset(self) -> tf.data.Dataset: - """Return dataset generated out of the current sequence instance. - - Implementative details - --------------------------------- - This method handles the conversion of this Keras Sequence into - a TensorFlow dataset, also handling the proper dispatching according - to what version of TensorFlow is installed in this system. - - Returns - ---------------------------------- - Dataset to be used for the training of a model - """ - - input_tensor_specs = [] - - for _ in range(2): - # Shapes of the source and destination node IDs - input_tensor_specs.append(tf.TensorSpec( - shape=(None, ), - dtype=tf.int32 - )) - - if self._use_edge_metrics: - # Shapes of the edge type IDs - input_tensor_specs.append(tf.TensorSpec( - shape=(None, self._graph.get_number_of_available_edge_metrics()), - dtype=tf.float64 - )) - - # Shapes of the kernel - input_tensor_specs.append(tf.SparseTensorSpec( - shape=(None, None), - dtype=tf.float32 - )) - - if self._node_features is not None: - input_tensor_specs.extend([ - tf.TensorSpec( - shape=(None, node_feature.shape[1]), - dtype=tf.float32 - ) - for node_feature in self._node_features - ]) - - input_tensor_specs.extend([ - tf.TensorSpec( - shape=(None, node_type_feature.shape[1]), - dtype=tf.float32 - ) - for node_type_feature in self._node_type_features - ]) - - return tf.data.Dataset.from_generator( - self, - output_signature=( - ( - input_tensor_specs, - ), - tf.TensorSpec( - shape=(None,), - dtype=tf.bool - ) - ) - ) - - def __getitem__(self, idx: int) -> Tuple[np.ndarray, np.ndarray]: """Return batch corresponding to given index. diff --git a/embiggen/sequences/tensorflow_sequences/kgsiamese_sequence.py b/embiggen/sequences/tensorflow_sequences/kgsiamese_sequence.py deleted file mode 100644 index 7aeaf911..00000000 --- a/embiggen/sequences/tensorflow_sequences/kgsiamese_sequence.py +++ /dev/null @@ -1,128 +0,0 @@ -"""Keras Sequence for running Neural Network on graph edge prediction.""" -from typing import List - -import numpy as np -import tensorflow as tf -from ensmallen import Graph # pylint: disable=no-name-in-module -from keras_mixed_sequence import Sequence -from embiggen.utils.tensorflow_utils import tensorflow_version_is_higher_or_equal_than - - -class KGSiameseSequence(Sequence): - """Keras Sequence for running Siamese Neural Network.""" - - def __init__( - self, - graph: Graph, - batch_size: int = 210, - random_state: int = 42 - ): - """Create new EdgePredictionSequence object. - - Parameters - -------------------------------- - graph: Graph, - The graph from which to sample the triples. - batch_size: int = 210, - The batch size to use. - random_state: int = 42, - The random_state to use to make extraction reproducible. - """ - self._graph = graph - self._random_state = random_state - self._current_index = 0 - super().__init__( - sample_number=self._graph.get_number_of_directed_edges(), - batch_size=batch_size, - ) - - def __call__(self): - """Return next batch using an infinite generator model.""" - self._current_index += 1 - return (tuple(self[self._current_index]),) - - def into_dataset(self) -> tf.data.Dataset: - """Return dataset generated out of the current sequence instance. - - Implementative details - --------------------------------- - This method handles the conversion of this Keras Sequence into - a TensorFlow dataset, also handling the proper dispatching according - to what version of TensorFlow is installed in this system. - - Returns - ---------------------------------- - Dataset to be used for the training of a model - """ - - ################################################################# - # Handling kernel creation when TensorFlow is a modern version. # - ################################################################# - - if tensorflow_version_is_higher_or_equal_than("2.5.0"): - input_tensor_specs = [] - - # For both the real and fake nodes. - for _ in range(4): - # Shapes of the source and destination node IDs - input_tensor_specs.append(tf.TensorSpec( - shape=(self._batch_size, ), - dtype=tf.uint32 - )) - for _ in range(4): - input_tensor_specs.append( - tf.TensorSpec( - shape=( - self._batch_size, - self._graph.get_maximum_multilabel_count() - ), - dtype=tf.uint32 - ) - ) - - # Shapes of the edge type IDs - input_tensor_specs.append(tf.TensorSpec( - shape=(self._batch_size,), - dtype=tf.uint32 - )) - - return tf.data.Dataset.from_generator( - self, - output_signature=(tuple(input_tensor_specs),) - ) - - input_tensor_types = [] - input_tensor_shapes = [] - - for _ in range(4): - input_tensor_types.append(tf.uint32,) - input_tensor_shapes.append(tf.TensorShape([self._batch_size, ]),) - - for _ in range(4): - input_tensor_types.append(tf.uint32,) - input_tensor_shapes.append( - tf.TensorShape([self._batch_size, self._graph.get_maximum_multilabel_count()]),) - - input_tensor_types.append(tf.uint32,) - input_tensor_shapes.append(tf.TensorShape([self._batch_size, ]),) - - return tf.data.Dataset.from_generator( - self, - output_types=input_tensor_types, - output_shapes=input_tensor_shapes - ) - - def __getitem__(self, idx: int) -> List[np.ndarray]: - """Return batch corresponding to given index to train a Siamese network. - - Parameters - --------------- - idx: int, - Index corresponding to batch to be returned. - """ - random_state = (self._random_state + idx) (1 + self.elapsed_epochs) - return (self._graph.get_kgsiamese_mini_batch( - random_state, - batch_size=self.batch_size, - use_zipfian_sampling=True - ),) diff --git a/embiggen/sequences/tensorflow_sequences/node_label_prediction_sequence.py b/embiggen/sequences/tensorflow_sequences/node_label_prediction_sequence.py deleted file mode 100644 index 08d70208..00000000 --- a/embiggen/sequences/tensorflow_sequences/node_label_prediction_sequence.py +++ /dev/null @@ -1,58 +0,0 @@ -"""Keras Sequence for running Neural Network on graph node-label embedding.""" -from typing import Tuple, Union, Optional - -import numpy as np -import tensorflow as tf -from ensmallen import Graph -from keras_mixed_sequence import Sequence - - -class NodeLabelPredictionSequence(Sequence): - """Keras Sequence for running Neural Network on graph node-label prediction.""" - - def __init__( - self, - graph: Graph, - max_neighbours: Optional[int] = 100, - include_central_node: bool = False, - return_edge_weights: bool = False, - batch_size: int = 2*8, - elapsed_epochs: int = 0, - random_state: int = 42, - ): - """Create new Node Label Prediction Sequence. - """ - self._graph = graph - self._random_state = random_state - self._include_central_node = include_central_node - self._return_edge_weights = return_edge_weights - self._max_neighbours = max_neighbours - super().__init__( - sample_number=graph.get_number_of_directed_edges(), - batch_size=batch_size, - elapsed_epochs=elapsed_epochs - ) - - def __getitem__(self, idx: int) -> Tuple[Union[tf.RaggedTensor, Tuple[tf.RaggedTensor]], np.ndarray]: - """Return batch corresponding to given index. - - Parameters - --------------- - idx: int, - Index corresponding to batch to be returned. - - Returns - --------------- - Return Tuple containing X and Y numpy arrays corresponding to given batch index. - """ - neighbours, weights, labels = self._graph.get_node_label_prediction_mini_batch( - idx=(self._random_state + idx) (1 + self.elapsed_epochs), - batch_size=self._batch_size, - include_central_node=self._include_central_node, - return_edge_weights=self._return_edge_weights, - max_neighbours=self._max_neighbours - ) - - if self._return_edge_weights: - return (tf.ragged.constant(neighbours), tf.ragged.constant(weights)), labels - return tf.ragged.constant(neighbours), labels diff --git a/embiggen/sequences/tensorflow_sequences/siamese_sequence.py b/embiggen/sequences/tensorflow_sequences/siamese_sequence.py index 76b92ff5..4b006638 100644 --- a/embiggen/sequences/tensorflow_sequences/siamese_sequence.py +++ b/embiggen/sequences/tensorflow_sequences/siamese_sequence.py @@ -1,4 +1,4 @@ -"""Keras Sequence for running Neural Network on graph edge prediction.""" +"""Keras Sequence for running Siamese Neural Network based on currupted triples sampling.""" from typing import List import numpy as np @@ -15,9 +15,10 @@ def __init__( self, graph: Graph, batch_size: int = 210, + return_edge_types: bool = False, random_state: int = 42 ): - """Create new EdgePredictionSequence object. + """Create new SiameseSequence object. Parameters -------------------------------- @@ -25,10 +26,13 @@ def __init__( The graph from which to sample the triples. batch_size: int = 210, The batch size to use. + return_edge_types: bool = False + Whether to return the edge types. random_state: int = 42, The random_state to use to make extraction reproducible. """ self._graph = graph + self._return_edge_types = return_edge_types self._random_state = random_state self._current_index = 0 super().__init__( @@ -54,47 +58,15 @@ def into_dataset(self) -> tf.data.Dataset: ---------------------------------- Dataset to be used for the training of a model """ - - ################################################################# - # Handling kernel creation when TensorFlow is a modern version. # - ################################################################# - - if tensorflow_version_is_higher_or_equal_than("2.5.0"): - input_tensor_specs = [] - - # For both the real and fake nodes. - for _ in range(4): - # Shapes of the source and destination node IDs - input_tensor_specs.append(tf.TensorSpec( - shape=(self._batch_size, ), - dtype=tf.uint32 - )) - - # Shapes of the edge type IDs - input_tensor_specs.append(tf.TensorSpec( - shape=(self._batch_size,), - dtype=tf.uint32 - )) - - return tf.data.Dataset.from_generator( - self, - output_signature=(tuple(input_tensor_specs),) - ) - - input_tensor_types = [] - input_tensor_shapes = [] - - for _ in range(4): - input_tensor_types.append(tf.uint32,) - input_tensor_shapes.append(tf.TensorShape([self._batch_size, ]),) - - input_tensor_types.append(tf.uint32,) - input_tensor_shapes.append(tf.TensorShape([self._batch_size, ]),) - return tf.data.Dataset.from_generator( self, - output_types=input_tensor_types, - output_shapes=input_tensor_shapes + output_signature=(tuple([ + tf.TensorSpec( + shape=(self._batch_size, ), + dtype=tf.uint32 + ) + for _ in range(4 + int(self._return_edge_types)) + ]),) ) def __getitem__(self, idx: int) -> List[np.ndarray]: @@ -106,8 +78,14 @@ def __getitem__(self, idx: int) -> List[np.ndarray]: Index corresponding to batch to be returned. """ random_state = (self._random_state + idx) * (1 + self.elapsed_epochs) - return (self._graph.get_siamese_mini_batch( - random_state, - batch_size=self.batch_size, - use_zipfian_sampling=True - ),) + return ( + self._graph.get_siamese_mini_batch_with_edge_types( + random_state, + batch_size=self.batch_size, + ) + if self._return_edge_types + else self._graph.get_siamese_mini_batch( + random_state, + batch_size=self.batch_size, + ), + ) diff --git a/embiggen/similarities/__init__.py b/embiggen/similarities/__init__.py new file mode 100644 index 00000000..d1958920 --- /dev/null +++ b/embiggen/similarities/__init__.py @@ -0,0 +1,5 @@ +"""Submodule providing models for edge prediction.""" +from embiggen.similarities.dag_resnik import DAGResnik + +# Export all non-internals. +__all__ = ["DAGResnik"] diff --git a/embiggen/similarities/dag_resnik.py b/embiggen/similarities/dag_resnik.py new file mode 100644 index 00000000..e4d368ff --- /dev/null +++ b/embiggen/similarities/dag_resnik.py @@ -0,0 +1,395 @@ +from typing import List, Optional, Union, Dict +import pandas as pd +import numpy as np +from ensmallen import models, Graph + + +class DAGResnik: + + def __init__(self, verbose: bool = True): + self._model = models.DAGResnik(verbose) + + def fit( + self, + graph: Graph, + node_counts: Dict[str, float], + node_frequencies: Optional[np.ndarray] = None, + ): + """Fit the Resnik similarity model. + + Parameters + -------------------- + graph: Graph + The graph to run similarities on. + node_counts: Dict[str, float] + Counts to compute the terms frequencies. + node_frequencies: Optional[np.ndarray] = None + Optional vector of node frequencies. + """ + self._model.fit( + graph, + node_counts=node_counts, + node_frequencies=node_frequencies + ) + + def get_similarity_from_node_id( + self, + first_node_id: int, + second_node_id: int + ) -> float: + """Return the similarity between the two provided nodes. + + Arguments + -------------------- + first_node_id: int + The first node for which to compute the similarity. + second_node_id: int + The second node for which to compute the similarity. + """ + return self._model.get_similarity_from_node_id(first_node_id, second_node_id) + + def get_similarity_from_node_ids( + self, + first_node_ids: List[int], + second_node_ids: List[int] + ) -> np.ndarray: + """Return the similarity between the two provided nodes. + + Arguments + -------------------- + first_node_ids: List[int] + The first node for which to compute the similarity. + second_node_ids: List[int] + The second node for which to compute the similarity. + """ + return self._model.get_similarity_from_node_ids(first_node_ids, second_node_ids) + + def get_similarity_from_node_name( + self, + first_node_name: str, + second_node_name: str + ) -> float: + """Return the similarity between the two provided nodes. + + Arguments + -------------------- + first_node_name: str + The first node for which to compute the similarity. + second_node_name: str + The second node for which to compute the similarity. + """ + return self._model.get_similarity_from_node_name(first_node_name, second_node_name) + + def get_similarity_from_node_names( + self, + first_node_names: List[str], + second_node_names: List[str] + ) -> np.ndarray: + """Return the similarity between the two provided nodes. + + Arguments + -------------------- + first_node_names: List[str] + The first node for which to compute the similarity. + second_node_names: List[str] + The second node for which to compute the similarity. + """ + return self._model.get_similarity_from_node_names(first_node_names, second_node_names) + + def get_pairwise_similarities( + self, + graph: Optional[Graph] = None, + return_similarities_dataframe: bool = False + ) -> Union[pd.DataFrame, np.ndarray]: + """Execute similarities on the provided graph. + + Parameters + -------------------- + graph: Optional[Graph] = None + The graph to run similarities on. + return_similarities_dataframe: bool = False + Whether to return a pandas DataFrame, which as indices has the node IDs. + By default, a numpy array with the similarities is returned as it weights much less. + """ + similarities = self._model.get_pairwise_similarities() + + if return_similarities_dataframe and graph is not None: + similarities = pd.DataFrame( + similarities, + columns=graph.get_node_names(), + index=graph.get_node_names(), + ) + + return similarities + + def get_similarities_from_graph( + self, + graph: Graph, + return_similarities_dataframe: bool = False + ) -> Union[pd.DataFrame, np.ndarray]: + """Execute similarities on the provided graph. + + Parameters + -------------------- + graph: Graph + The graph to run similarities on. + node_frequencies: Optional[np.ndarray] + Optional vector of node frequencies. + return_similarities_dataframe: bool = False + Whether to return a pandas DataFrame, which as indices has the node IDs. + By default, a numpy array with the similarities is returned as it weights much less. + """ + similarities = self._model.get_similarities_from_graph( + graph, + ) + + if return_similarities_dataframe: + similarities = pd.DataFrame( + { + "similarities": similarities, + "sources": graph.get_directed_source_node_ids(), + "destinations": graph.get_directed_destination_node_ids(), + }, + ) + + return similarities + + def get_similarities_from_bipartite_graph_from_edge_node_ids( + self, + graph: Graph, + source_node_ids: List[int], + destination_node_ids: List[int], + return_similarities_dataframe: bool = False + ) -> Union[pd.DataFrame, np.ndarray]: + """Execute similarities probabilities on the provided graph bipartite portion. + + Parameters + -------------------- + graph: Graph + The graph from which to extract the edges. + node_frequencies: Optional[np.ndarray] + Optional vector of node frequencies. + source_node_ids: List[int] + The source nodes of the bipartite graph. + destination_node_ids: List[int] + The destination nodes of the bipartite graph. + return_similarities_dataframe: bool = False + Whether to return a pandas DataFrame, which as indices has the node IDs. + By default, a numpy array with the similarities is returned as it weights much less. + """ + return self.get_similarities_from_graph( + graph.build_bipartite_graph_from_edge_node_ids( + source_node_ids=source_node_ids, + destination_node_ids=destination_node_ids, + directed=True + ), + return_similarities_dataframe=return_similarities_dataframe + ) + + def get_similarities_from_bipartite_graph_from_edge_node_names( + self, + graph: Graph, + source_node_names: List[str], + destination_node_names: List[str], + return_similarities_dataframe: bool = False + ) -> Union[pd.DataFrame, np.ndarray]: + """Execute similarities probabilities on the provided graph bipartite portion. + + Parameters + -------------------- + graph: Graph + The graph from which to extract the edges. + node_frequencies: Optional[np.ndarray] + Optional vector of node frequencies. + source_node_names: List[str] + The source nodes of the bipartite graph. + destination_node_names: List[str] + The destination nodes of the bipartite graph. + return_similarities_dataframe: bool = False + Whether to return a pandas DataFrame, which as indices has the node IDs. + By default, a numpy array with the similarities is returned as it weights much less. + """ + return self.get_similarities_from_graph( + graph.build_bipartite_graph_from_edge_node_names( + source_node_names=source_node_names, + destination_node_names=destination_node_names, + directed=True + ), + return_similarities_dataframe=return_similarities_dataframe + ) + + def get_similarities_from_bipartite_graph_from_edge_node_prefixes( + self, + graph: Graph, + source_node_prefixes: List[str], + destination_node_prefixes: List[str], + return_similarities_dataframe: bool = False + ) -> Union[pd.DataFrame, np.ndarray]: + """Execute similarities probabilities on the provided graph bipartite portion. + + Parameters + -------------------- + graph: Graph + The graph from which to extract the edges. + node_frequencies: Optional[np.ndarray] + Optional vector of node frequencies. + source_node_prefixes: List[str] + The source node prefixes of the bipartite graph. + destination_node_prefixes: List[str] + The destination node prefixes of the bipartite graph. + return_similarities_dataframe: bool = False + Whether to return a pandas DataFrame, which as indices has the node IDs. + By default, a numpy array with the similarities is returned as it weights much less. + """ + return self.get_similarities_from_graph( + graph.build_bipartite_graph_from_edge_node_prefixes( + source_node_prefixes=source_node_prefixes, + destination_node_prefixes=destination_node_prefixes, + directed=True + ), + return_similarities_dataframe=return_similarities_dataframe + ) + + def get_similarities_from_bipartite_graph_from_edge_node_types( + self, + graph: Graph, + source_node_types: List[str], + destination_node_types: List[str], + return_similarities_dataframe: bool = False + ) -> Union[pd.DataFrame, np.ndarray]: + """Execute similarities probabilities on the provided graph bipartite portion. + + Parameters + -------------------- + graph: Graph + The graph from which to extract the edges. + node_frequencies: Optional[np.ndarray] + Optional vector of node frequencies. + source_node_types: List[str] + The source node prefixes of the bipartite graph. + destination_node_types: List[str] + The destination node prefixes of the bipartite graph. + return_similarities_dataframe: bool = False + Whether to return a pandas DataFrame, which as indices has the node IDs. + By default, a numpy array with the similarities is returned as it weights much less. + """ + return self.get_similarities_from_graph( + graph.build_bipartite_graph_from_edge_node_types( + source_node_types=source_node_types, + destination_node_types=destination_node_types, + directed=True + ), + return_similarities_dataframe=return_similarities_dataframe + ) + + def get_similarities_from_clique_graph_from_node_ids( + self, + graph: Graph, + node_ids: List[int], + return_similarities_dataframe: bool = False + ) -> Union[pd.DataFrame, np.ndarray]: + """Execute similarities probabilities on the provided graph bipartite portion. + + Parameters + -------------------- + graph: Graph + The graph from which to extract the edges. + node_frequencies: Optional[np.ndarray] + Optional vector of node frequencies. + node_ids: List[int] + The nodes of the bipartite graph. + return_similarities_dataframe: bool = False + Whether to return a pandas DataFrame, which as indices has the node IDs. + By default, a numpy array with the similarities is returned as it weights much less. + """ + return self.get_similarities_from_graph( + graph.build_clique_graph_from_node_ids( + node_ids=node_ids, + directed=True + ), + return_similarities_dataframe=return_similarities_dataframe + ) + + def get_similarities_from_clique_graph_from_node_names( + self, + graph: Graph, + node_names: List[str], + return_similarities_dataframe: bool = False + ) -> Union[pd.DataFrame, np.ndarray]: + """Execute similarities probabilities on the provided graph bipartite portion. + + Parameters + -------------------- + graph: Graph + The graph from which to extract the edges. + node_frequencies: Optional[np.ndarray] + Optional vector of node frequencies. + node_names: List[str] + The nodes of the bipartite graph. + return_similarities_dataframe: bool = False + Whether to return a pandas DataFrame, which as indices has the node IDs. + By default, a numpy array with the similarities is returned as it weights much less. + """ + return self.get_similarities_from_graph( + graph.build_clique_graph_from_node_names( + node_names=node_names, + directed=True + ), + return_similarities_dataframe=return_similarities_dataframe + ) + + def get_similarities_from_clique_graph_from_node_prefixes( + self, + graph: Graph, + node_prefixes: List[str], + return_similarities_dataframe: bool = False + ) -> Union[pd.DataFrame, np.ndarray]: + """Execute similarities probabilities on the provided graph bipartite portion. + + Parameters + -------------------- + graph: Graph + The graph from which to extract the edges. + node_frequencies: Optional[np.ndarray] + Optional vector of node frequencies. + node_prefixes: List[str] + The node prefixes of the bipartite graph. + return_similarities_dataframe: bool = False + Whether to return a pandas DataFrame, which as indices has the node IDs. + By default, a numpy array with the similarities is returned as it weights much less. + """ + return self.get_similarities_from_graph( + graph.build_clique_graph_from_node_prefixes( + node_prefixes=node_prefixes, + directed=True + ), + return_similarities_dataframe=return_similarities_dataframe + ) + + def get_similarities_from_clique_graph_from_node_types( + self, + graph: Graph, + node_types: List[str], + return_similarities_dataframe: bool = False + ) -> Union[pd.DataFrame, np.ndarray]: + """Execute similarities probabilities on the provided graph bipartite portion. + + Parameters + -------------------- + graph: Graph + The graph from which to extract the edges. + node_frequencies: Optional[np.ndarray] + Optional vector of node frequencies. + node_types: List[str] + The node prefixes of the bipartite graph. + return_similarities_dataframe: bool = False + Whether to return a pandas DataFrame, which as indices has the node IDs. + By default, a numpy array with the similarities is returned as it weights much less. + """ + return self.get_similarities_from_graph( + graph.build_clique_graph_from_node_types( + node_types=node_types, + directed=True + ), + return_similarities_dataframe=return_similarities_dataframe + ) diff --git a/embiggen/utils/__init__.py b/embiggen/utils/__init__.py index feceb1e6..0f627c6f 100644 --- a/embiggen/utils/__init__.py +++ b/embiggen/utils/__init__.py @@ -21,7 +21,6 @@ "EmbeddingResult", "AbstractModel", "classification_evaluation_pipeline", - "build_init", "format_list", "get_models_dataframe", "get_available_models_for_node_label_prediction", @@ -30,5 +29,4 @@ "get_available_models_for_node_embedding", "abstract_class", "number_to_ordinal", - "format_list" ] diff --git a/embiggen/utils/abstract_edge_gcn.py b/embiggen/utils/abstract_edge_gcn.py index 46428ab5..ff3f1ce4 100644 --- a/embiggen/utils/abstract_edge_gcn.py +++ b/embiggen/utils/abstract_edge_gcn.py @@ -33,6 +33,7 @@ def __init__( number_of_units_per_ffnn_head_layer: Union[int, List[int]] = 128, dropout_rate: float = 0.3, apply_norm: bool = False, + combiner: str ="sum", edge_embedding_method: str = "Concatenate", optimizer: Union[str, Optimizer] = "adam", early_stopping_min_delta: float = 0.0001, @@ -55,6 +56,7 @@ def __init__( handling_multi_graph: str = "warn", node_feature_names: Optional[List[str]] = None, node_type_feature_names: Optional[List[str]] = None, + edge_feature_names: Optional[List[str]] = None, verbose: bool = True ): """Create new Kipf GCN object. @@ -85,6 +87,13 @@ def __init__( apply_norm: bool = False Whether to normalize the output of the convolution operations, after applying the level activations. + combiner: str = "mean" + A string specifying the reduction op. + Currently "mean", "sqrtn" and "sum" are supported. + "sum" computes the weighted sum of the embedding results for each row. + "mean" is the weighted sum divided by the total weight. + "sqrtn" is the weighted sum divided by the square root of the sum of the squares of the weights. + Defaults to mean. edge_embedding_method: str = "Concatenate" The edge embedding method to use to put togheter the source and destination node features, which includes: @@ -181,6 +190,9 @@ def __init__( node_type_feature_names: Optional[List[str]] = None Names of the node type features. This is used as the layer names. + edge_feature_names: Optional[List[str]] = None + Names of the edge features. + This is used as the layer names. verbose: bool = True Whether to show loading bars. """ @@ -191,6 +203,7 @@ def __init__( number_of_units_per_graph_convolution_layers=number_of_units_per_graph_convolution_layers, dropout_rate=dropout_rate, apply_norm=apply_norm, + combiner=combiner, optimizer=optimizer, early_stopping_min_delta=early_stopping_min_delta, early_stopping_patience=early_stopping_patience, @@ -225,6 +238,7 @@ def __init__( self._edge_embedding_method = edge_embedding_method self._use_edge_metrics = use_edge_metrics + self._edge_feature_names = edge_feature_names self._use_node_types = None @classmethod @@ -263,6 +277,7 @@ def _get_model_prediction_input( return_node_types=self.is_using_node_types(), node_type_features=node_type_features, use_edge_metrics=self._use_edge_metrics, + edge_features=edge_features ) def _get_model_training_output(self, graph: Graph) -> Optional[np.ndarray]: @@ -346,6 +361,31 @@ def _build_model( else: edge_metrics = None + edge_feature_inputs = [] + if edge_features is None: + edge_features = [] + edge_feature_names = self._edge_feature_names + if edge_feature_names is None: + edge_feature_names = [ + f"{number_to_ordinal(i+1)}EdgeFeature" + for i in range(len(edge_features)) + ] + if len(edge_feature_names) != len(edge_features): + raise ValueError( + f"You have provided {len(edge_feature_names)} " + f"edge feature names but you have provided {len(edge_features)} " + "edge features to the model." + ) + for edge_feature, feature_name in zip(edge_features, edge_feature_names): + feature_names.append(feature_name) + edge_feature_input = Input( + shape=edge_feature.shape[1:], + batch_size=nodes_number, + name=feature_name, + ) + edge_feature_inputs.append(edge_feature_input) + source_and_destination_features.append(edge_feature_input) + ffnn_outputs = [] for hidden, feature_name in zip(source_and_destination_features, feature_names): @@ -364,7 +404,8 @@ def _build_model( if self._edge_embedding_method == "Concatenate": hidden = Concatenate( - name="NodeConcatenation" + name="NodeConcatenation", + axis=-1 )(source_and_destination_features) elif self._edge_embedding_method == "Average": hidden = Average( @@ -407,10 +448,11 @@ def _build_model( if len(other_features) > 0: hidden = Concatenate( - name="EdgeFeatures" + name="EdgeFeatures", + axis=-1 )([ hidden, - other_features + other_features ]) # Building the head of the model. @@ -427,18 +469,21 @@ def _build_model( name="Output" )(hidden) + inputs=[ + input_layer + for input_layer in ( + source_nodes, + destination_nodes, + edge_metrics, + edge_feature_inputs, + graph_convolution_model.inputs, + ) + if input_layer is not None + ] + # Building the the model. model = Model( - inputs=[ - input_layer - for input_layer in ( - source_nodes, - destination_nodes, - edge_metrics, - graph_convolution_model.inputs, - ) - if input_layer is not None - ], + inputs=inputs, outputs=output, name=self.model_name().replace(" ", "_") ) diff --git a/embiggen/utils/abstract_gcn.py b/embiggen/utils/abstract_gcn.py index 70a6521c..5cac7202 100644 --- a/embiggen/utils/abstract_gcn.py +++ b/embiggen/utils/abstract_gcn.py @@ -1,5 +1,6 @@ """Kipf GCN model for node-label prediction.""" from typing import List, Union, Optional, Dict, Any, Type, Tuple +from matplotlib.pyplot import axis import numpy as np import pandas as pd @@ -10,7 +11,7 @@ from ensmallen import Graph import tensorflow as tf -from tensorflow.keras.utils import Sequence +from keras_mixed_sequence import Sequence from embiggen.utils.abstract_models import AbstractClassifierModel, abstract_class from embiggen.utils.number_to_ordinal import number_to_ordinal from embiggen.layers.tensorflow import GraphConvolution, FlatEmbedding @@ -96,7 +97,7 @@ def graph_to_sparse_tensor( return tf.SparseTensor( graph.get_directed_edge_node_ids(), ( - graph.get_edge_weights() + graph.get_directed_edge_weights() if use_weights else tf.ones(graph.get_number_of_directed_edges()) ), @@ -115,6 +116,7 @@ def __init__( number_of_units_per_graph_convolution_layers: Union[int, List[int]] = 128, dropout_rate: float = 0.5, apply_norm: bool = False, + combiner: str = "mean", optimizer: Union[str, Optimizer] = "adam", early_stopping_min_delta: float = 0.001, early_stopping_patience: int = 10, @@ -153,6 +155,13 @@ def __init__( apply_norm: bool = False Whether to normalize the output of the convolution operations, after applying the level activations. + combiner: str = "mean" + A string specifying the reduction op. + Currently "mean", "sqrtn" and "sum" are supported. + "sum" computes the weighted sum of the embedding results for each row. + "mean" is the weighted sum divided by the total weight. + "sqrtn" is the weighted sum divided by the square root of the sum of the squares of the weights. + Defaults to mean. optimizer: str = "Adam" The optimizer to use while training the model. early_stopping_min_delta: float @@ -219,6 +228,7 @@ def __init__( can_be_empty=True ) + self._combiner = combiner self._epochs = epochs self._use_class_weights = use_class_weights self._dropout_rate = dropout_rate @@ -275,6 +285,7 @@ def parameters(self) -> Dict[str, Any]: number_of_units_per_graph_convolution_layers=self._number_of_units_per_graph_convolution_layers, epochs=self._epochs, apply_norm=self._apply_norm, + combiner=self._combiner, use_class_weights=self._use_class_weights, dropout_rate=self._dropout_rate, optimizer=self._optimizer, @@ -419,18 +430,21 @@ def _build_graph_convolution_model( ): if features is not None: if feature_names is None: - feature_names = [None] * len(features) + feature_names = [ + f"{number_to_ordinal(i+1)} {feature_category}" + for i in range(len(features)) + ] if len(feature_names) != len(features): raise ValueError( f"You have provided {len(feature_names)} " f"{feature_category} names but you have provided {len(features)} " - "{feature_category}s to the model." + f"{feature_category}s to the model." ) input_features.extend([ Input( shape=node_feature.shape[1:], batch_size=nodes_number, - name=node_feature_name + name=node_feature_name, ) for node_feature, node_feature_name in zip( features, @@ -482,6 +496,7 @@ def _build_graph_convolution_model( for i, units in enumerate(self._number_of_units_per_graph_convolution_layers): hidden = GraphConvolution( units=units, + combiner=self._combiner, dropout_rate=self._dropout_rate, apply_norm=self._apply_norm, name=f"{number_to_ordinal(i+1)}GraphConvolution" @@ -492,12 +507,16 @@ def _build_graph_convolution_model( inputs=[ input_layer for input_layer in ( - adjacency_matrix, input_features + adjacency_matrix, + input_features ) if input_layer is not None ], outputs=( - Concatenate(name="ConcatenatedNodeFeatures")(hidden) + Concatenate( + name="ConcatenatedNodeFeatures", + axis=-1 + )(hidden) if len(hidden) > 1 else hidden[0] ) @@ -561,16 +580,18 @@ def _fit( edge_features=edge_features, ) + model_input = self._get_model_training_input( + graph, + support=support, + edge_features=edge_features, + node_type_features=node_type_features, + node_features=node_features + ) + self.history = self._model.fit( - x=self._get_model_training_input( - graph, - support=support, - edge_features=edge_features, - node_type_features=node_type_features, - node_features=node_features - ), + x=model_input, y=self._get_model_training_output(graph), - sample_weight=self._get_model_training_output(graph), + sample_weight=self._get_model_training_sample_weights(graph), epochs=self._epochs, verbose=traditional_verbose and self._verbose > 0, batch_size=graph.get_number_of_nodes(), @@ -609,14 +630,17 @@ def _predict_proba( """Run predictions on the provided graph.""" if support is None: support = graph + + model_input = self._get_model_prediction_input( + graph, + support, + node_features, + node_type_features, + edge_features, + ) + return self._model.predict( - self._get_model_prediction_input( - graph, - support, - node_features, - node_type_features, - edge_features, - ), + model_input, batch_size=support.get_number_of_nodes(), verbose=False ) diff --git a/embiggen/utils/abstract_models/__init__.py b/embiggen/utils/abstract_models/__init__.py index 53044657..cde3ea73 100644 --- a/embiggen/utils/abstract_models/__init__.py +++ b/embiggen/utils/abstract_models/__init__.py @@ -16,7 +16,6 @@ "AbstractClassifierModel", "AbstractEmbeddingModel", "EmbeddingResult", - "get_standardized_model_map", "abstract_class", "AbstractModel", "get_models_dataframe", @@ -25,5 +24,5 @@ "get_available_models_for_node_label_prediction", "get_available_models_for_node_embedding", "build_init", - "format_list" + "format_list", ] diff --git a/embiggen/utils/abstract_models/abstract_classifier_model.py b/embiggen/utils/abstract_models/abstract_classifier_model.py index 74cec045..c131c9bb 100644 --- a/embiggen/utils/abstract_models/abstract_classifier_model.py +++ b/embiggen/utils/abstract_models/abstract_classifier_model.py @@ -4,6 +4,7 @@ import numpy as np import pandas as pd import time +from userinput.utils import must_be_in_set from tqdm.auto import trange, tqdm from embiggen.utils.abstract_models.list_formatting import format_list from cache_decorator import Cache @@ -238,17 +239,11 @@ def normalize_node_feature( node_feature.set_random_state(random_state) if smoke_test: - node_feature = node_feature.__class__( - node_feature.smoke_test_parameters() - ) + node_feature = node_feature.into_smoke_test() node_feature = node_feature.fit_transform( graph=graph, - return_dataframe=False, - # If this is an Ensmallen model, we can enable the verbosity - # as it will only show up in the jupyter kernel and it won't bother the - # other loading bars. - verbose="Ensmallen" == node_feature.library_name() + return_dataframe=False ) if isinstance(node_feature, EmbeddingResult): @@ -441,14 +436,11 @@ def normalize_node_type_feature( node_type_feature.set_random_state(random_state) if smoke_test: - node_type_feature = node_type_feature.__class__( - node_type_feature.smoke_test_parameters() - ) + node_type_feature = node_type_feature.into_smoke_test() node_type_feature = node_type_feature.fit_transform( graph=graph, return_dataframe=False, - verbose=False ) if isinstance(node_type_feature, EmbeddingResult): @@ -641,14 +633,11 @@ def normalize_edge_feature( edge_feature.set_random_state(random_state) if smoke_test: - edge_feature = edge_feature.__class__( - *edge_feature.smoke_test_parameters() - ) + edge_feature = edge_feature.into_smoke_test() edge_feature = edge_feature.fit_transform( graph=graph, return_dataframe=False, - verbose=False ) if isinstance(edge_feature, EmbeddingResult): @@ -670,7 +659,7 @@ def normalize_edge_feature( ) ) - if graph.get_directed_edges_number() != ef.shape[0]: + if graph.get_number_of_directed_edges() != ef.shape[0]: raise ValueError( ( "The provided edge features have {rows_number} rows " @@ -806,6 +795,9 @@ def fit( f"the {self.model_name()} requires edge types." ) + if support is None: + support = graph + try: self._fit( graph=graph, @@ -872,22 +864,29 @@ def predict( "Do call the `.fit` method before the `.predict` " "method." )) + + if support is None: + support = graph + try: - return self._predict( + predictions = self._predict( graph=graph, support=support, node_features=self.normalize_node_features( graph=graph, + random_state=self._random_state, node_features=node_features, allow_automatic_feature=False, ), node_type_features=self.normalize_node_type_features( graph=graph, + random_state=self._random_state, node_type_features=node_type_features, allow_automatic_feature=True, ), edge_features=self.normalize_edge_features( graph=graph, + random_state=self._random_state, edge_features=edge_features, allow_automatic_feature=False, ), @@ -898,6 +897,8 @@ def predict( f"implemented using the {self.library_name()} for the {self.task_name()} task. " f"Specifically, the class of the model is {self.__class__.__name__}." ) from e + + return predictions def predict_proba( self, @@ -934,8 +935,11 @@ def predict_proba( "method." )) + if support is None: + support = graph + try: - return self._predict_proba( + predictions = self._predict_proba( graph=graph, support=support, node_features=self.normalize_node_features( @@ -964,6 +968,23 @@ def predict_proba( f"Specifically, the class of the model is {self.__class__.__name__}." ) from e + if ( + not self.is_binary_prediction_task() and + not self.is_multilabel_prediction_task() and + ( + len(predictions.shape) == 1 or + predictions.shape[1] == 2 + ) + ): + raise ValueError( + "This task is not a binary prediction, " + f"yet the {self.model_name()} model predict proba method has " + "returned a vector of prediction probabilities with " + f"shape {predictions.shape}." + ) + + return predictions + def evaluate_predictions( self, ground_truth: np.ndarray, @@ -1002,7 +1023,7 @@ def evaluate_predictions( metric.__name__: metric( ground_truth, predictions, - average="binary" if self.is_binary_prediction_task() else "macro", + average="macro", zero_division=0 ) for metric in ( @@ -1027,7 +1048,6 @@ def evaluate_prediction_probabilities( prediction_probabilities: np.ndarray The predictions to be evaluated. """ - metrics = [] if self.is_binary_prediction_task(): return { "auroc": express_measures.binary_auroc( @@ -1044,18 +1064,16 @@ def evaluate_prediction_probabilities( def wrapper_roc_auc_score(args, *kwargs): return roc_auc_score(args, kwargs, multi_class="ovr") - metrics.append(wrapper_roc_auc_score) - return { - metric.__name__: metric( + "auroc": wrapper_roc_auc_score( ground_truth, - prediction_probabilities, + prediction_probabilities ) - for metric in metrics } - @staticmethod + @classmethod def split_graph_following_evaluation_schema( + cls, graph: Graph, evaluation_schema: str, holdout_number: int, @@ -1077,9 +1095,10 @@ def split_graph_following_evaluation_schema( holdouts_kwargs: Dict[str, Any] The kwargs to be forwarded to the holdout method. """ - raise NotImplementedError( - "The `split_graph_following_evaluation_schema` method should be implemented " - "in the child classes of abstract classifier model." + must_be_in_set( + evaluation_schema, + cls.get_available_evaluation_schemas(), + f"evaluation schema for {cls.task_name()} and model {cls.model_name()}" ) def _evaluate( @@ -1211,7 +1230,7 @@ def iterate_classifier_models( cache_path="{cache_dir}/{self.task_name()}/{graph.get_name()}/holdout_{holdout_number}/{self.model_name()}/{self.library_name()}/{_hash}.csv.gz", cache_dir="experiments", enable_cache_arg_name="enable_cache", - args_to_ignore=["verbose", "smoke_test"], + args_to_ignore=["verbose", "smoke_test", "train_of_interest", "test_of_interest", "train",], capture_enable_cache_arg_name=True, use_approximated_hash=True ) diff --git a/embiggen/utils/abstract_models/abstract_embedding_model.py b/embiggen/utils/abstract_models/abstract_embedding_model.py index 4e9e75ac..6f4d7a00 100644 --- a/embiggen/utils/abstract_models/abstract_embedding_model.py +++ b/embiggen/utils/abstract_models/abstract_embedding_model.py @@ -41,10 +41,10 @@ def __init__( def parameters(self) -> Dict[str, Any]: """Returns parameters of the embedding model.""" - return { + return dict( super().parameters(), - "embedding_size": self._embedding_size - } + embedding_size=self._embedding_size + ) @classmethod def smoke_test_parameters(cls) -> Dict[str, Any]: @@ -63,11 +63,15 @@ def requires_nodes_sorted_by_decreasing_node_degree(cls) -> bool: "in the child classes of abstract model." )) + @classmethod + def get_minimum_required_number_of_node_types(cls) -> int: + """Requires minimum number of required node types.""" + return 0 + def _fit_transform( self, graph: Graph, return_dataframe: bool = True, - verbose: bool = True ) -> EmbeddingResult: """Run embedding on the provided graph. @@ -75,8 +79,6 @@ def _fit_transform( -------------------- graph: Graph The graph to run predictions on. - verbose: bool - Whether to show loading bars. """ raise NotImplementedError(( "The `_fit_transform` method must be implemented " @@ -87,13 +89,11 @@ def _fit_transform( cache_path="{cache_dir}/{self.model_name()}/{self.library_name()}/{graph.get_name()}/{_hash}.pkl.gz", cache_dir="embedding", enable_cache_arg_name="self._enable_cache", - args_to_ignore=["verbose"] ) def _cached_fit_transform( self, graph: Graph, return_dataframe: bool = True, - verbose: bool = True ) -> EmbeddingResult: """Execute embedding on the provided graph. @@ -103,8 +103,6 @@ def _cached_fit_transform( The graph to run embedding on. return_dataframe: bool = True Whether to return a pandas DataFrame with the embedding. - verbose: bool = True - Whether to show loading bars. Returns -------------------- @@ -119,7 +117,7 @@ def _cached_fit_transform( if not graph.has_nodes_sorted_by_decreasing_outbound_node_degree(): raise ValueError( f"The given graph {graph.get_name()} does not have the nodes sorted by decreasing " - "order, therefore the negative sampling (which follows a zipfian " + "order, therefore the negative sampling (which follows a scale free " "distribution) would not approximate well the Softmax.\n" "In order to sort the given graph in such a way that the node IDs " "are sorted by decreasing outbound node degrees, you can use " @@ -131,6 +129,14 @@ def _cached_fit_transform( f"The provided graph {graph.get_name()} does not have node types, but " f"the {self.model_name()} requires node types." ) + + if self.requires_node_types() and graph.get_number_of_node_types() <= 1: + raise ValueError( + f"The {self.model_name()} requires the graph to have " + f"at least {self.get_minimum_required_number_of_node_types()} node types, " + f"but the provided one has {graph.get_number_of_node_types()} " + "node types." + ) if self.requires_edge_types() and not graph.has_edge_types(): raise ValueError( @@ -173,7 +179,6 @@ def _cached_fit_transform( result = self._fit_transform( graph=graph, return_dataframe=return_dataframe, - verbose=verbose ) if not isinstance(result, EmbeddingResult): @@ -192,7 +197,6 @@ def fit_transform( repository: Optional[str] = None, version: Optional[str] = None, return_dataframe: bool = True, - verbose: bool = True ) -> EmbeddingResult: """Execute embedding on the provided graph. @@ -212,8 +216,6 @@ def fit_transform( from the ensmallen automatic retrieval. return_dataframe: bool = True Whether to return a pandas DataFrame with the embedding. - verbose: bool = True - Whether to show loading bars. Returns -------------------- @@ -228,5 +230,4 @@ def fit_transform( return self._cached_fit_transform( graph=graph, return_dataframe=return_dataframe, - verbose=verbose ) diff --git a/embiggen/utils/abstract_models/abstract_model.py b/embiggen/utils/abstract_models/abstract_model.py index eb714a29..e38ddd9a 100644 --- a/embiggen/utils/abstract_models/abstract_model.py +++ b/embiggen/utils/abstract_models/abstract_model.py @@ -3,7 +3,7 @@ from embiggen.utils.abstract_models.list_formatting import format_list from typing import Dict, Any, Type, List, Optional from dict_hash import Hashable, sha256 -from userinput.utils import closest +from userinput.utils import must_be_in_set import inspect @@ -15,6 +15,7 @@ def is_not_implemented(method: Callable) -> bool: """Returns whether this method contains a raise for not being implemented.""" return "raise NotImplementedError" in inspect.getsource(method) + def is_implemented(method: Callable) -> bool: """Returns whether this method is implemented.""" return not is_not_implemented(method) @@ -46,9 +47,14 @@ def __init__(self, random_state: Optional[int] = None): f"random state of `{random_state}` was provided." ) + can_use_edge_weights = self.__getattribute__("can_use_edge_weights") + requires_positive_edge_weights = self.__getattribute__( + "requires_positive_edge_weights") + if ( - not self.__getattribute__("can_use_edge_weights")() and - is_implemented(self.__getattribute__("requires_positive_edge_weights")) + is_implemented(can_use_edge_weights) and + not can_use_edge_weights() and + is_implemented(requires_positive_edge_weights) ): raise ValueError( "We have found an useless method in the " @@ -128,6 +134,13 @@ def parameters(self) -> Dict[str, Any]: random_state=self._random_state ) + def into_smoke_test(self) -> "Self": + """Creates new instance with smoke test parameters.""" + return self.__class__({ + self.parameters(), + self.smoke_test_parameters() + }) + @classmethod def requires_edge_weights(cls) -> bool: """Returns whether the model requires edge weights.""" @@ -153,7 +166,10 @@ def task_involves_edge_weights(cls) -> bool: def can_use_edge_weights(cls) -> bool: """Returns whether the model can optionally use edge weights.""" try: - if cls.requires_edge_weights(): + if ( + is_implemented(cls.requires_edge_weights) and + cls.requires_edge_weights() + ): return True except NotImplementedError: pass @@ -332,9 +348,9 @@ def consistent_hash(self) -> str: """Returns consistent hash describing the model.""" return sha256(dict( self.parameters(), - model_name= self.model_name(), - library_name= self.library_name(), - task_name= self.task_name(), + model_name=self.model_name(), + library_name=self.library_name(), + task_name=self.task_name(), )) @staticmethod @@ -359,38 +375,16 @@ def set_random_state(self, random_state: int): ) self._random_state = random_state - @staticmethod - def get_model_data( - model_name: str - ) -> Dict[str, Dict]: - """Returns data relative to the registered model data.""" - # We check if the provided string is not an empty string. - if len(model_name) == 0: - raise ValueError( - "The provided model name is empty." - ) - - # We turn this to lowercase in order to allow - # for error in casing of the model, since one may - # write models like `GloVe` also as `Glove` or other - # typos, which are generally easy to make. - lowercase_model_mapping = AbstractModel.get_available_model_names_in_lowercase_mapping() - if model_name.lower() not in lowercase_model_mapping: - raise ValueError( - f"The provided model name `{model_name}` is not available. " - f"Did you mean {closest(model_name, lowercase_model_mapping.values())}?" - ) - # We retrieve the model standard name. - model_name = lowercase_model_mapping[model_name.lower()] - return AbstractModel.MODELS_LIBRARY[model_name] - @staticmethod def get_task_data( model_name: str, task_name: str ) -> Dict[str, Dict]: """Returns data relative to the registered model and task data.""" - model_data = AbstractModel.get_model_data(model_name) + if len(model_name) == 0: + raise ValueError( + "The provided model name is empty." + ) # We check if the provided string is not an empty string. if len(task_name) == 0: @@ -398,25 +392,20 @@ def get_task_data( "The provided task name is empty." ) - # We do a similar check as the one above for the tasks, - # as one may do typos while writig the task name and - # we should always provide the best possible help message. - lowercase_task_mapping = { - t.lower(): t - for t in model_data.keys() - } - if task_name.lower() not in lowercase_task_mapping: - raise ValueError( - f"The provided task name `{task_name}` is not available for " - f"the requested model {model_name}." - f"Did you mean {closest(task_name, lowercase_task_mapping.values())}?" - ) + task_name = must_be_in_set( + task_name, + AbstractModel.MODELS_LIBRARY, + "task name" + ) - # We retrieve the task standard name. - task_name = lowercase_task_mapping[task_name.lower()] + model_name = must_be_in_set( + model_name, + AbstractModel.MODELS_LIBRARY[task_name], + "model name" + ) # We retrieve the task data. - return model_data[task_name] + return AbstractModel.MODELS_LIBRARY[task_name][model_name] @staticmethod def get_library_data( @@ -433,25 +422,18 @@ def get_library_data( "The provided library name is empty." ) - lowercase_libraries_mapping = { - t.lower(): t - for t in task_data.keys() - } - if library_name.lower() not in lowercase_libraries_mapping: - raise ValueError( - f"The provided library name `{library_name}` is not available for " - f"the requested model {model_name}. " - f"Did you mean {closest(library_name, lowercase_libraries_mapping.values())}?" - ) - - # We retrieve the library standard name. - library_name = lowercase_libraries_mapping[library_name.lower()] + library_name = must_be_in_set( + library_name, + task_data.keys(), + "library name" + ) # We retrieve the library data. return task_data[library_name] - @staticmethod + @classmethod def get_model_from_library( + cls, model_name: str, task_name: Optional[str] = None, library_name: Optional[str] = None, @@ -474,18 +456,7 @@ def get_model_from_library( be raised. """ if task_name is None: - task_names = list(AbstractModel.get_model_data(model_name).keys()) - if len(task_names) == 1: - task_name = task_names[0] - else: - formatted_list = format_list(task_names) - raise ValueError( - f"The requested model `{model_name}` is available for " - "multiple tasks and no specific task was requested, " - "so it is unclear which task you intend to execute. " - f"Specifically, the available tasks are {formatted_list}." - "Please do provide a task name to resolve this ambiguity." - ) + task_name = cls.task_name() task_data = AbstractModel.get_task_data(model_name, task_name) @@ -523,19 +494,6 @@ def get_model_from_library( # its class to let the user do whathever they want. return model_class - @staticmethod - def get_available_model_names() -> List[str]: - """Returns list of available model names.""" - return list(AbstractModel.MODELS_LIBRARY.keys()) - - @staticmethod - def get_available_model_names_in_lowercase_mapping() -> Dict[str, str]: - """Returns list of available model names in lowercase.""" - return { - model_name.lower(): model_name - for model_name in AbstractModel.get_available_model_names() - } - @staticmethod def find_available_models( model_name: str, @@ -565,20 +523,20 @@ def register(model_class: Type["AbstractModel"]): model_class: Type["AbstractModel"] The class to register. """ + task_name = model_class.task_name() model_name = model_class.model_name() # If this is the first model of its kind to be registered. - if model_name not in AbstractModel.MODELS_LIBRARY: - AbstractModel.MODELS_LIBRARY[model_name] = {} + if task_name not in AbstractModel.MODELS_LIBRARY: + AbstractModel.MODELS_LIBRARY[task_name] = {} # We retrieve the data for the model to enrich it. # This is NOT a copy, but a reference to the same STATIC object. - model_data = AbstractModel.MODELS_LIBRARY[model_name] + model_data = AbstractModel.MODELS_LIBRARY[task_name] - task_name = model_class.task_name() - if task_name not in model_data: - model_data[task_name] = {} + if model_name not in model_data: + model_data[model_name] = {} - task_data = model_data[task_name] + task_data = model_data[model_name] class_name = model_class.__name__ diff --git a/embiggen/utils/abstract_models/model_stub.py b/embiggen/utils/abstract_models/model_stub.py index 64c0d597..cfaa7f75 100644 --- a/embiggen/utils/abstract_models/model_stub.py +++ b/embiggen/utils/abstract_models/model_stub.py @@ -180,6 +180,11 @@ def can_use_edge_weights(cls) -> bool: """Returns whether the model can optionally use edge weights.""" return None + @classmethod + def is_stocastic(cls) -> bool: + """Returns whether the model can optionally use edge weights.""" + return None + @classmethod def is_topological(cls) -> str: """Returns whether this embedding is based on graph topology.""" @@ -193,7 +198,7 @@ def task_name(cls) -> str: @classmethod def is_available(cls) -> bool: """Returns whether the model class is actually available in the user system.""" - return True + return False model_class = StubClass else: diff --git a/embiggen/utils/networkx_utils.py b/embiggen/utils/networkx_utils.py index 5bc75832..c0c89ce0 100644 --- a/embiggen/utils/networkx_utils.py +++ b/embiggen/utils/networkx_utils.py @@ -25,7 +25,7 @@ def convert_ensmallen_graph_to_networkx_graph( (src_name, dst_name, edge_weight) for (src_name, dst_name), edge_weight in zip( graph.get_directed_edge_node_ids(), - graph.get_edge_weights() + graph.get_directed_edge_weights() ) ]) else: diff --git a/embiggen/utils/normalize_model_structural_parameters.py b/embiggen/utils/normalize_model_structural_parameters.py index 10831f36..1c68067f 100644 --- a/embiggen/utils/normalize_model_structural_parameters.py +++ b/embiggen/utils/normalize_model_structural_parameters.py @@ -1,9 +1,9 @@ """Submodule providing utilities to normalize parameters for modular models.""" -from typing import List, Any, Union, Type +from typing import List, Any, Union, Type, Optional def normalize_model_list_parameter( - candidate_list: Union[Any, List[Any]], + candidate_list: Optional[Union[Any, List[Any]]], elements_number: int, object_type: Type, default_value: Any = None, @@ -104,7 +104,7 @@ def normalize_model_list_parameter( def normalize_model_ragged_list_parameter( - candidate_ragged_list: Union[Any, List[Any], List[List[Any]]], + candidate_ragged_list: Optional[Union[Any, List[Any], List[List[Any]]]], submodules_number: int, layers_per_submodule: List[int], object_type: Type, diff --git a/embiggen/visualizations/graph_visualizer.py b/embiggen/visualizations/graph_visualizer.py index e4487f1f..f26cd77a 100644 --- a/embiggen/visualizations/graph_visualizer.py +++ b/embiggen/visualizations/graph_visualizer.py @@ -1,7 +1,7 @@ """Module with embedding visualization tools.""" import functools from multiprocessing import cpu_count -from typing import Dict, Iterator, List, Tuple, Union, Optional, Callable +from typing import Dict, Iterator, List, Tuple, Union, Optional, Callable, Type import matplotlib import matplotlib.pyplot as plt @@ -22,14 +22,32 @@ from sklearn.tree import DecisionTreeClassifier from matplotlib.legend_handler import HandlerBase, HandlerTuple from matplotlib import collections as mc +from mpl_toolkits.mplot3d.art3d import Line3DCollection from sanitize_ml_labels import sanitize_ml_labels from sklearn.decomposition import PCA +from userinput.utils import must_be_in_set from sklearn.metrics import balanced_accuracy_score from sklearn.model_selection import ShuffleSplit import itertools +from embiggen.utils.abstract_models.abstract_embedding_model import AbstractEmbeddingModel from embiggen.utils.abstract_models.embedding_result import EmbeddingResult +from mpl_toolkits.mplot3d.proj3d import proj_transform +from matplotlib.text import Annotation + +class Annotation3D(Annotation): + '''Annotate the point xyz with text s''' + + def __init__(self, s, xyz, args, kwargs): + Annotation.__init__(self,s, xy=(0,0), args, kwargs) + self._verts3d = xyz + + def draw(self, renderer): + xs3d, ys3d, zs3d = self._verts3d + xs, ys, zs = proj_transform(xs3d, ys3d, zs3d, renderer.M) + self.xy=(xs,ys) + Annotation.draw(self, renderer) try: from ddd_subplots import subplots as subplots_3d, rotate, display_video_at_path @@ -89,6 +107,7 @@ def __init__( destination_nodes_prefixes: Optional[Union[str, List[str]]] = None, edge_type_names: Optional[List[Optional[str]]] = None, show_graph_name: Union[str, bool] = "auto", + number_of_columns_in_legend: int = 2, show_node_embedding_method: bool = True, show_edge_embedding_method: bool = True, show_separability_considerations_explanation: bool = True, @@ -220,6 +239,8 @@ def __init__( Whether to show the graph name in the plots. By default, it is shown if the graph does not have a trivial name such as `Graph`. + number_of_columns_in_legend: int = 2 + The number of columns to be used with the legend. show_node_embedding_method: bool = True Whether to show the node embedding method. By default, we show it if we can detect it. @@ -288,6 +309,7 @@ def __init__( self._support = support self._subgraph_of_interest = subgraph_of_interest + self._number_of_columns_in_legend = number_of_columns_in_legend if isinstance(source_node_types_names, str): source_node_types_names = [source_node_types_names] @@ -345,10 +367,10 @@ def __init__( ) # We sample the negative edges using the subgraph of interest as base graph - # to follow its zipfian distribution, which may be different from the - # main graph zipfian distribution when particular filters are applied to it. - # For instance, the zipfian distribution of one particular edge type - # may be very different from the whole graph zipfian distribution. + # to follow its scale free distribution, which may be different from the + # main graph scale free distribution when particular filters are applied to it. + # For instance, the scale free distribution of one particular edge type + # may be very different from the whole graph scale free distribution. # Furthermore, we avoid sampling false negatives by passing to the # method also the support graph. try: @@ -358,7 +380,7 @@ def __init__( self._positive_graph.get_number_of_edges() ), random_state=random_state, - use_zipfian_sampling=True, + use_scale_free_distribution=True, graph_to_avoid=self._support, support=self._support, only_from_same_component=only_from_same_component, @@ -408,7 +430,12 @@ def __init__( decomposition_kwargs = {} self._n_components = n_components - self._decomposition_method = decomposition_method + + self._decomposition_method = must_be_in_set( + decomposition_method, + ("PCA", "TSNE", "UMAP"), + "decomposition method" + ) self._decomposition_kwargs = decomposition_kwargs def iterate_subsampled_node_ids(self) -> Iterator[int]: @@ -521,7 +548,9 @@ def get_decomposition_method(self) -> Callable: # Adding a warning for when decomposing methods that # embed nodes using a cosine similarity / distance approach # in order to avoid false negatives. - if self._node_embedding_method_name == "GloVe": + if self._decomposition_method in ("UMAP", "TSNE") and self._node_embedding_method_name in ( + "Node2Vec GloVe", "DeepWalk GloVe", "First-order LINE" + ): metric = self._decomposition_kwargs.get("metric") if metric is not None and metric != "cosine": warnings.warn( @@ -529,7 +558,7 @@ def get_decomposition_method(self) -> Callable: "such as Glove, which embeds nodes using a dot product, it is " "highly suggested to use a `cosine` metric. Using a different " f"metric, such as the one you have provided ({metric}) may lead " - "to unsuccessfull decompositions using UMAP or t-SNE." + "to worse decompositions using UMAP or t-SNE." ) else: # Otherwise we switch to using a cosine metric. @@ -556,7 +585,7 @@ def get_decomposition_method(self) -> Callable: leave=False, dynamic_ncols=True ), - verbose=True, + verbose=self._verbose, ), self._decomposition_kwargs }).fit_transform @@ -599,6 +628,7 @@ def get_decomposition_method(self) -> Callable: n_jobs=cpu_count(), random_state=self._random_state, verbose=self._verbose, + learning_rate=200, n_iter=400, init="random", method="exact" if self._n_components == 4 else "barnes_hut", @@ -631,14 +661,10 @@ def get_decomposition_method(self) -> Callable: ), self._decomposition_kwargs }).fit_transform - else: - raise ValueError( - "We currently only support PCA and TSNE decomposition methods." - ) def _get_node_embedding( self, - node_embedding: Optional[Union[pd.DataFrame, np.ndarray, str, EmbeddingResult]] = None, + node_embedding: Optional[Union[pd.DataFrame, np.ndarray, str, EmbeddingResult, Type[AbstractEmbeddingModel]]] = None, node_embedding_kwargs: Dict ) -> np.ndarray: """Computes the node embedding if it was not otherwise provided. @@ -666,7 +692,10 @@ def _get_node_embedding( ) else: node_embedding = self._node_embedding_method_name - if isinstance(node_embedding, str): + if ( + isinstance(node_embedding, str) or + issubclass(node_embedding.__class__, AbstractEmbeddingModel) + ): node_embedding = embed_graph( graph=self._graph, embedding_model=node_embedding, @@ -678,7 +707,6 @@ def _get_node_embedding( self._has_autodetermined_node_embedding_name = True self._node_embedding_method_name = node_embedding.embedding_method_name node_embedding = np.hstack(node_embedding.get_all_node_embedding()) - elif self._node_embedding_method_name == "auto" or self._has_autodetermined_node_embedding_name: self._has_autodetermined_node_embedding_name = True self._node_embedding_method_name = self.automatically_detect_node_embedding_method( @@ -798,7 +826,7 @@ def _set_legend( number_of_columns = 1 if len(labels) <= 2 and any( len(label) > 20 for label in labels - ) else 2 + ) else self._number_of_columns_in_legend legend = axes.legend( handles=handles, labels=[ @@ -827,9 +855,6 @@ def _set_legend( def automatically_detect_node_embedding_method(self, node_embedding: np.ndarray) -> Optional[str]: """Detect node embedding method using heuristics, where possible.""" - # Rules to detect SPINE embedding - if node_embedding.dtype == "uint8" and node_embedding.min() == 0: - return "SPINE" # Rules to detect TFIDF/BERT embedding if node_embedding.dtype == "float16" and node_embedding.shape[1] == 768: return "TFIDF-weighted BERT" @@ -877,7 +902,7 @@ def fit_nodes( size=self._number_of_subsampled_nodes ) node_transformer = NodeTransformer( - aligned_node_mapping=True + aligned_mapping=True ) node_transformer.fit(node_embedding) node_embedding = node_transformer.transform( @@ -899,7 +924,8 @@ def _get_positive_edges_embedding( """ graph_transformer = GraphTransformer( method=self._edge_embedding_method, - aligned_node_mapping=True + aligned_mapping=True, + include_both_undirected_edges=False ) graph_transformer.fit(node_embedding) return graph_transformer.transform(self._positive_graph) @@ -940,7 +966,8 @@ def _get_negative_edge_embedding( """ graph_transformer = GraphTransformer( method=self._edge_embedding_method, - aligned_node_mapping=True + aligned_mapping=True, + include_both_undirected_edges=False ) graph_transformer.fit(node_embedding) return graph_transformer.transform( @@ -1020,19 +1047,30 @@ def _get_figure_and_axes( "them can be None." )) if self._n_components == 2: - figure, axes = plt.subplots({ - GraphVisualizer.DEFAULT_SUBPLOT_KWARGS, - kwargs - }) - figure.patch.set_facecolor("white") - axes.axis('equal') + figure, axes = plt.subplots( + { + GraphVisualizer.DEFAULT_SUBPLOT_KWARGS, + kwargs + }, + squeeze=False + ) + for ax in axes.flatten(): + ax.axis('equal') + if axes.size == 1: + axes = axes.flatten()[0] else: - figure, axes = subplots_3d({ - GraphVisualizer.DEFAULT_SUBPLOT_KWARGS, - kwargs - }) - figure.patch.set_facecolor("white") - axes.axis('auto') + figure, axes = subplots_3d( + { + GraphVisualizer.DEFAULT_SUBPLOT_KWARGS, + kwargs + }, + squeeze=False + ) + for ax in axes.flatten(): + ax.axis('auto') + if axes.size == 1: + axes = axes.flatten()[0] + figure.patch.set_facecolor("white") return figure, axes def _get_complete_title( @@ -1208,15 +1246,16 @@ def _plot_scatter( if train_indices is None and test_indices is None: scatter = axes.scatter( points.T, - { + dict( dict( c=colors, - edgecolors=None if edgecolors is None else cmap(edgecolors), + edgecolors=None if edgecolors is None else cmap( + edgecolors), marker=train_marker, cmap=cmap, ), scatter_kwargs - } + ) ) collections.append(scatter) legend_elements += scatter.legend_elements()[0] @@ -1325,19 +1364,6 @@ def _plot_scatter( def _wrapped_plot_scatter(self, kwargs): if self._rotate: - # These backups are needed for two reasons: - # 1) Processes in python necessarily copy the instance objects for each process - # and this can cause a considerable memery peak to occour. - # 2) Some of the objects considered are not picklable, such as, at the time of writing - # the lambdas used in the graph transformer or the graph object itself. - graph_backup = self._graph - node_embedding = self._node_decomposition - edge_embedding = self._positive_edge_decomposition - negative_edge_embedding = self._negative_edge_decomposition - self._node_decomposition = None - self._positive_edge_decomposition = None - self._negative_edge_decomposition = None - self._graph = None try: kwargs["loc"] = "lower right" path = "{}.{}".format( @@ -1350,20 +1376,15 @@ def _wrapped_plot_scatter(self, kwargs): path=path, duration=self._duration, fps=self._fps, - verbose=True, + verbose=self._verbose, kwargs ) except (Exception, KeyboardInterrupt) as e: - self._node_decomposition = node_embedding - self._positive_edge_decomposition = edge_embedding - self._negative_edge_decomposition = negative_edge_embedding - self._graph = graph_backup raise e - self._node_decomposition = node_embedding - self._positive_edge_decomposition = edge_embedding - self._negative_edge_decomposition = negative_edge_embedding - self._graph = graph_backup - return display_video_at_path(path) + to_display = display_video_at_path(path) + if to_display is None: + return () + return to_display return self._plot_scatter(kwargs) def _plot_types( @@ -1634,22 +1655,46 @@ def plot_edge_segments( add_selfloops_where_missing=False, complete=False, ) + edge_node_ids = np.array([ + [ + np.where(self._subsampled_node_ids == src)[0][0], + np.where(self._subsampled_node_ids == dst)[0][0] + ] + for src, dst in edge_node_ids + if src in self._subsampled_node_ids and dst in self._subsampled_node_ids + ]) else: edge_node_ids = self._graph.get_edge_node_ids( directed=False ) - lines_collection = mc.LineCollection( - self._node_decomposition[edge_node_ids], - linewidths=1, - zorder=0, - { - GraphVisualizer.DEFAULT_EDGES_SCATTER_KWARGS, - ( - {} if scatter_kwargs is None else scatter_kwargs - ) - } - ) + if edge_node_ids.size == 0: + return figure, axes + + if self._n_components == 3: + lines_collection = Line3DCollection( + self._node_decomposition[edge_node_ids], + linewidths=1, + zorder=0, + { + GraphVisualizer.DEFAULT_EDGES_SCATTER_KWARGS, + ( + {} if scatter_kwargs is None else scatter_kwargs + ) + } + ) + else: + lines_collection = mc.LineCollection( + self._node_decomposition[edge_node_ids], + linewidths=1, + zorder=0, + { + GraphVisualizer.DEFAULT_EDGES_SCATTER_KWARGS, + ( + {} if scatter_kwargs is None else scatter_kwargs + ) + } + ) axes.add_collection(lines_collection) return figure, axes @@ -1759,7 +1804,7 @@ def plot_nodes( kwargs ) - if annotate_nodes: + if annotate_nodes and returned_values: figure, axes = returned_values[:2] self.annotate_nodes( figure=figure, @@ -1775,15 +1820,26 @@ def annotate_nodes( axes: Axes, points: np.ndarray ) -> Tuple[Figure, Axes]: - if self._subsampled_node_ids is not None: - node_names = [ - self._graph.get_node_name_from_node_id(node_id) - for node_id in self._subsampled_node_ids - ] - else: - node_names = self._graph.get_node_names() - for i, txt in enumerate(node_names): - axes.annotate(txt, points[i], fontsize=8, ha="center", va="center") + for node_name, point in zip(( + self._graph.get_node_name_from_node_id(node_id) + for node_id in self.iterate_subsampled_node_ids() + ), points): + if point.size == 3: + axes.add_artist(Annotation3D( + node_name, + point, + fontsize=8, + ha="center", + va="center" + )) + else: + axes.annotate( + node_name, + point, + fontsize=8, + ha="center", + va="center" + ) return (figure, axes) def plot_edges( @@ -2047,6 +2103,16 @@ def _plot_positive_and_negative_edges_metric( edge_metric_callback(subgraph=self._positive_graph), )) + # Filter the edge metrics relative to edges that are not + # to be displayed. + if not self._graph.is_directed(): + edge_metrics = edge_metrics[np.concatenate([ + self._negative_graph.get_directed_source_node_ids( + ) <= self._negative_graph.get_directed_destination_node_ids(), + self._positive_graph.get_directed_source_node_ids( + ) <= self._positive_graph.get_directed_destination_node_ids(), + ])] + points = np.vstack([ self._negative_edge_decomposition, self._positive_edge_decomposition, @@ -2187,6 +2253,11 @@ def _plot_positive_and_negative_edges_metric_histogram( figure, axes = plt.subplots(figsize=(5, 5)) figure.patch.set_facecolor("white") + if self._negative_graph.is_directed() and edge_metrics is None: + number_of_negative_edges = self._negative_graph.get_number_of_directed_edges() + else: + number_of_negative_edges = self._negative_graph.get_number_of_undirected_edges() + if edge_metrics is None: edge_metrics = np.concatenate(( edge_metric_callback(subgraph=self._negative_graph), @@ -2195,8 +2266,8 @@ def _plot_positive_and_negative_edges_metric_histogram( axes.hist( [ - edge_metrics[:self._negative_graph.get_number_of_directed_edges()], - edge_metrics[self._negative_graph.get_number_of_directed_edges():], + edge_metrics[:number_of_negative_edges], + edge_metrics[number_of_negative_edges:], ], bins=10, log=True, @@ -2652,13 +2723,10 @@ def plot_positive_and_negative_edges_resource_allocation_index( def _get_flatten_unknown_node_ontologies(self) -> Tuple[List[str], np.ndarray]: """Returns unique ontologies and node ontologies adjusted for the current instance.""" - if self._subsampled_node_ids is None: - ontology_names = self._graph.get_node_ontologies() - else: - ontology_names = [ - self._graph.get_ontology_from_node_id(node_id) - for node_id in self._subsampled_node_ids - ] + ontology_names = [ + self._graph.get_ontology_from_node_id(node_id) + for node_id in self.iterate_subsampled_node_ids() + ] # The following is needed to normalize the multiple types ontologies_counts = Counter(ontology_names) @@ -2704,13 +2772,6 @@ def _get_flatten_multi_label_and_unknown_node_types(self) -> np.ndarray: node_types_number = self._graph.get_number_of_node_types() unknown_node_types_id = node_types_number - # According to whether the subsampled node IDs were given, - # we iterate on them or on the complete set of nodes of the graph. - if self._subsampled_node_ids is None: - nodes_iterator = range(self._graph.get_number_of_nodes()) - else: - nodes_iterator = self._subsampled_node_ids - # When we have multiple node types for a given node, we set it to # the most common node type of the set. return np.fromiter( @@ -2725,7 +2786,7 @@ def _get_flatten_multi_label_and_unknown_node_types(self) -> np.ndarray: )[0] for node_type_ids in ( self._graph.get_node_type_ids_from_node_id(node_id) - for node_id in nodes_iterator + for node_id in self.iterate_subsampled_node_ids() ) ), dtype=np.uint32 @@ -2743,7 +2804,12 @@ def _get_flatten_unknown_edge_types(self) -> np.ndarray: if edge_type_id is None else edge_type_id - for edge_type_id in self._positive_graph.get_edge_type_ids() + for edge_type_id in ( + self._positive_graph.get_directed_edge_type_ids() + if self._positive_graph.is_directed() + else + self._positive_graph.get_undirected_edge_type_ids() + ) ), dtype=np.uint32 ) @@ -2850,6 +2916,29 @@ def plot_node_types( kwargs ) + if self._subsampled_node_ids is not None: + if node_type_predictions is not None: + node_type_predictions = node_type_predictions[self._subsampled_node_ids] + + if train_indices is not None: + train_indices = np.fromiter( + ( + np.where(self._subsampled_node_ids == node_id)[0][0] + for node_id in train_indices + if node_id in self._subsampled_node_ids + ), + dtype=train_indices.dtype + ) + if test_indices is not None: + test_indices = np.fromiter( + ( + np.where(self._subsampled_node_ids == node_id)[0][0] + for node_id in test_indices + if node_id in self._subsampled_node_ids + ), + dtype=test_indices.dtype + ) + if annotate_nodes == "auto": annotate_nodes = self._graph.get_number_of_nodes() < 50 and not self._rotate @@ -3267,7 +3356,9 @@ def plot_connected_components( points=self._node_decomposition, ) - if not return_caption: + if not return_caption or self._rotate: + if self._rotate: + return returned_values return self._handle_notebook_display(returned_values) # TODO! Add caption node abount gaussian ball! @@ -3357,16 +3448,13 @@ def plot_node_degrees( "method before plotting the nodes." ) - if self._subsampled_node_ids is None: - degrees = self._support.get_node_degrees() - else: - degrees = np.fromiter( - ( - self._support.get_node_degree_from_node_id(node_id) - for node_id in self._subsampled_node_ids - ), - dtype=np.uint32 - ) + degrees = np.fromiter( + ( + self._support.get_node_degree_from_node_id(node_id) + for node_id in self.iterate_subsampled_node_ids() + ), + dtype=np.uint32 + ) if annotate_nodes == "auto": annotate_nodes = self._graph.get_number_of_nodes() < 50 and not self._rotate @@ -3644,7 +3732,10 @@ def plot_edge_weights( "method before plotting the nodes." ) - weights = self._positive_graph.get_edge_weights() + if self._positive_graph.is_directed(): + weights = self._positive_graph.get_directed_edge_weights() + else: + weights = self._positive_graph.get_undirected_edge_weights() returned_values = self._wrapped_plot_scatter( points=self._positive_edge_decomposition, @@ -3725,16 +3816,17 @@ def _plot_positive_and_negative_edges_distance_histogram( """ graph_transformer = GraphTransformer( method=distance_callback, - aligned_node_mapping=True + aligned_mapping=True, + include_both_undirected_edges=False ) graph_transformer.fit(node_features.astype(np.float32)) return self._plot_positive_and_negative_edges_metric_histogram( metric_name=distance_name, - edge_metrics=graph_transformer.transform(np.vstack([ - self._negative_graph.get_directed_edge_node_ids(), - self._positive_graph.get_directed_edge_node_ids(), - ])).flatten(), + edge_metrics=np.vstack([ + graph_transformer.transform(self._negative_graph), + graph_transformer.transform(self._positive_graph) + ]).flatten(), figure=figure, axes=axes, apply_tight_layout=apply_tight_layout, @@ -3746,7 +3838,6 @@ def _plot_positive_and_negative_edges_distance( node_features: np.ndarray, distance_name: str, distance_callback: str, - offset: float = 0.0, kwargs: Dict ): """Plot distances of node features for positive and negative edges. @@ -3759,10 +3850,6 @@ def _plot_positive_and_negative_edges_distance( The title for the heatmap. distance_callback: str The callback to use to compute the distances. - offset: float = 0.0 - The offset to move the distance when it is not a true distance - such as with the cosine similarity and negative value would - not be plottable on a logarithmic scale. kwargs: Dict Additional kwargs to forward. @@ -3777,17 +3864,18 @@ def _plot_positive_and_negative_edges_distance( """ graph_transformer = GraphTransformer( method=distance_callback, - aligned_node_mapping=True + aligned_mapping=True, + include_both_undirected_edges=False ) graph_transformer.fit(node_features.astype(np.float32)) return self._plot_positive_and_negative_edges_metric( metric_name=distance_name, - edge_metrics=offset + graph_transformer.transform(np.vstack([ - self._negative_graph.get_directed_edge_node_ids(), - self._positive_graph.get_directed_edge_node_ids(), - ])), + edge_metrics=np.vstack([ + graph_transformer.transform(self._negative_graph), + graph_transformer.transform(self._positive_graph) + ]), kwargs, ) @@ -4005,7 +4093,6 @@ def plot_positive_and_negative_edges_cosine_similarity( node_features=node_features, distance_name="Cosine similarity", distance_callback="CosineSimilarity", - offset=1.0, figure=figure, axes=axes, scatter_kwargs=scatter_kwargs, @@ -4149,7 +4236,7 @@ def plot_edge_weight_distribution( self._graph.get_number_of_directed_edges() // 10 ) axes.hist( - self._graph.get_edge_weights(), + self._graph.get_directed_edge_weights(), bins=number_of_buckets, log=True ) @@ -4212,13 +4299,12 @@ def fit_and_plot_all( node_scatter_plot_methods_to_call = [] distribution_plot_methods_to_call = [] - if self._graph.has_constant_non_zero_node_degrees(): - node_scatter_plot_methods_to_call.append( - self.plot_node_degrees, - ) - distribution_plot_methods_to_call.append( - self.plot_node_degree_distribution, - ) + node_scatter_plot_methods_to_call.append( + self.plot_node_degrees, + ) + distribution_plot_methods_to_call.append( + self.plot_node_degree_distribution, + ) def plot_distance_wrapper(plot_distance): @functools.wraps(plot_distance) @@ -4262,7 +4348,7 @@ def wrapped_plot_distance(kwargs): self.plot_node_ontologies ) - if not self._support.is_connected(): + if not self._support.is_connected() and not self._support.is_directed(): node_scatter_plot_methods_to_call.append( self.plot_connected_components ) @@ -4290,7 +4376,7 @@ def wrapped_plot_distance(*kwargs): int(math.ceil(number_of_total_plots / number_of_columns)), 1) ncols = min(number_of_columns, number_of_total_plots) - figure, axes = plt.subplots( + figure, axes = self._get_figure_and_axes( nrows=nrows, ncols=ncols, figsize=(5ncols, 5nrows), @@ -4348,16 +4434,22 @@ def wrapped_plot_distance(kwargs): complete_caption += f" <b>({letter})</b> {caption}" if show_letters: + if self._n_components == 3: + additional_kwargs = dict(z=0.0) + else: + additional_kwargs = dict() + ax.text( - 0.0, - 1.1, - letter, + x=0.0, + y=1.1, + s=letter, size=18, color="black", weight="bold", horizontalalignment="left", verticalalignment="center", transform=ax.transAxes, + additional_kwargs ) complete_caption += "<br>" @@ -4378,6 +4470,8 @@ def wrapped_plot_distance(kwargs): complete_caption += self.get_non_existing_edges_sampling_description() for axis in flat_axes[number_of_total_plots:]: + for spine in axis.spines.values(): + spine.set_visible(False) axis.axis("off") if show_name_backup: @@ -4388,12 +4482,15 @@ def wrapped_plot_distance(kwargs): ), fontsize=20 ) - figure.tight_layout(rect=[0, 0.03, 1, 0.96]) - else: + if self._n_components != 3: + figure.tight_layout(rect=[0, 0.03, 1, 0.96]) + elif self._n_components != 3: figure.tight_layout() self._show_graph_name = show_name_backup return self._handle_notebook_display( - figure, axes, caption=complete_caption + figure, + axes, + caption=complete_caption ) diff --git a/setup.py b/setup.py index 3a7b6649..077ba6b3 100644 --- a/setup.py +++ b/setup.py @@ -22,9 +22,6 @@ def readme(): "silence_tensorflow" ] -extras = { - 'test': test_deps, -} def read(parts): with copen(os.path.join(here, parts), 'r') as fp: @@ -56,11 +53,11 @@ def find_version(file_paths): description='Graph embedding, machine learning, and visualization library.', long_description=readme(), url='https://github.com/monarch-initiative/embiggen', - keywords='node2vec,word2vec,CBOW,SkipGram,GloVe', + keywords='Graph Representation Learning,LINE,TransE,Node2Vec,DeeWalk', author=", ".join(list(authors.keys())), author_email=", ".join(list(authors.values())), license='BSD3', - python_requires='>=3.6.0', + python_requires='>=3.7.0', packages=find_packages( exclude=['contrib', 'docs', 'tests', 'notebooks']), install_requires=[ @@ -69,17 +66,18 @@ def find_version(*file_paths): "tqdm", "matplotlib>=3.5.2", "scikit-learn", - "ddd_subplots>=1.0.19", + "userinput>=1.0.19", + "ddd_subplots>=1.0.20", "sanitize_ml_labels>=1.0.38", "keras_mixed_sequence>=1.0.28", - "extra_keras_metrics>=2.0.7", - "ensmallen>=0.8.3", + "ensmallen>=0.8.8", "validate_version_code", "cache_decorator>=2.1.8", "packaging" ], tests_require=test_deps, include_package_data=True, - zip_safe=False, - extras_require=extras, + extras_require={ + 'test': test_deps, + }, )	Bringing resnik, bug fixes, MMAPs and generalizations to develop		2022-08-02T20:05:14	[]	[]
emmett-framework/granian	emmett-framework__granian-454	5cd82992e1085539dbff295e804021365f347d3c	diff --git a/.github/workflows/build.yml b/.github/workflows/build.yml index e2f9c11..006a055 100644 --- a/.github/workflows/build.yml +++ b/.github/workflows/build.yml @@ -4,7 +4,7 @@ on: workflow_dispatch env: MATURIN_VERSION: 1.7.4 - PY_ALL: 3.8 3.9 3.10 3.11 3.12 3.13 pypy3.8 pypy3.9 pypy3.10 + PY_ALL: 3.9 3.10 3.11 3.12 3.13 pypy3.9 pypy3.10 jobs: wheels: @@ -23,13 +23,13 @@ jobs: platform: linux target: x86_64 manylinux: auto - interpreter: pypy3.8 pypy3.9 pypy3.10 + interpreter: pypy3.9 pypy3.10 - os: macos target: x86_64 - interpreter: pypy3.8 pypy3.9 pypy3.10 + interpreter: pypy3.9 pypy3.10 - os: macos target: aarch64 - interpreter: 3.8 3.9 pypy3.8 pypy3.9 pypy3.10 + interpreter: 3.9 pypy3.9 pypy3.10 - os: ubuntu platform: linux target: aarch64 @@ -44,7 +44,7 @@ jobs: manylinux: musllinux_1_1 - os: windows target: x86_64 - interpreter: pypy3.8 pypy3.9 pypy3.10 + interpreter: pypy3.9 pypy3.10 exclude: - os: windows target: aarch64 @@ -84,7 +84,7 @@ jobs: matrix: os: [ubuntu-latest, macos-latest, macos-14, windows-latest] manylinux: [auto] - interpreter: ["3.8", "3.9", "3.10", "3.11", "3.12", "3.13"] + interpreter: ["3.9", "3.10", "3.11", "3.12", "3.13"] include: - os: ubuntu-latest platform: linux diff --git a/.github/workflows/release.yml b/.github/workflows/release.yml index f2d3044..47863c3 100644 --- a/.github/workflows/release.yml +++ b/.github/workflows/release.yml @@ -7,7 +7,7 @@ on: env: MATURIN_VERSION: 1.7.4 - PY_ALL: 3.8 3.9 3.10 3.11 3.12 3.13 pypy3.8 pypy3.9 pypy3.10 + PY_ALL: 3.9 3.10 3.11 3.12 3.13 pypy3.9 pypy3.10 jobs: sdist: @@ -46,13 +46,13 @@ jobs: platform: linux target: x86_64 manylinux: auto - interpreter: pypy3.8 pypy3.9 pypy3.10 + interpreter: pypy3.9 pypy3.10 - os: macos target: x86_64 - interpreter: pypy3.8 pypy3.9 pypy3.10 + interpreter: pypy3.9 pypy3.10 - os: macos target: aarch64 - interpreter: pypy3.8 pypy3.9 pypy3.10 + interpreter: pypy3.9 pypy3.10 - os: ubuntu platform: linux target: aarch64 @@ -67,7 +67,7 @@ jobs: manylinux: musllinux_1_1 - os: windows target: x86_64 - interpreter: pypy3.8 pypy3.9 pypy3.10 + interpreter: pypy3.9 pypy3.10 exclude: - os: windows target: aarch64 @@ -102,7 +102,7 @@ jobs: matrix: os: [ubuntu-latest, macos-13, macos-14, windows-latest] manylinux: [auto] - interpreter: ["3.8", "3.9", "3.10", "3.11", "3.12", "3.13"] + interpreter: ["3.9", "3.10", "3.11", "3.12", "3.13"] include: - os: ubuntu-latest platform: linux diff --git a/Cargo.lock b/Cargo.lock index ce9d6b9..2c1ee71 100644 --- a/Cargo.lock +++ b/Cargo.lock @@ -970,9 +970,9 @@ dependencies = [ [[package]] name = "pyo3" -version = "0.22.6" +version = "0.23.2" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "f402062616ab18202ae8319da13fa4279883a2b8a9d9f83f20dbade813ce1884" +checksum = "f54b3d09cbdd1f8c20650b28e7b09e338881482f4aa908a5f61a00c98fba2690" dependencies = [ "anyhow", "cfg-if", @@ -989,9 +989,9 @@ dependencies = [ [[package]] name = "pyo3-build-config" -version = "0.22.6" +version = "0.23.2" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "b14b5775b5ff446dd1056212d778012cbe8a0fbffd368029fd9e25b514479c38" +checksum = "3015cf985888fe66cfb63ce0e321c603706cd541b7aec7ddd35c281390af45d8" dependencies = [ "once_cell", "python3-dll-a", @@ -1000,9 +1000,9 @@ dependencies = [ [[package]] name = "pyo3-ffi" -version = "0.22.6" +version = "0.23.2" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "9ab5bcf04a2cdcbb50c7d6105de943f543f9ed92af55818fd17b660390fc8636" +checksum = "6fca7cd8fd809b5ac4eefb89c1f98f7a7651d3739dfb341ca6980090f554c270" dependencies = [ "libc", "pyo3-build-config", @@ -1010,9 +1010,9 @@ dependencies = [ [[package]] name = "pyo3-log" -version = "0.11.0" +version = "0.12.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "3ac84e6eec1159bc2a575c9ae6723baa6ee9d45873e9bebad1e3ad7e8d28a443" +checksum = "3eb421dc86d38d08e04b927b02424db480be71b777fa3a56f32e2f2a3a1a3b08" dependencies = [ "arc-swap", "log", @@ -1021,9 +1021,9 @@ dependencies = [ [[package]] name = "pyo3-macros" -version = "0.22.6" +version = "0.23.2" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "0fd24d897903a9e6d80b968368a34e1525aeb719d568dba8b3d4bfa5dc67d453" +checksum = "34e657fa5379a79151b6ff5328d9216a84f55dc93b17b08e7c3609a969b73aa0" dependencies = [ "proc-macro2", "pyo3-macros-backend", @@ -1033,9 +1033,9 @@ dependencies = [ [[package]] name = "pyo3-macros-backend" -version = "0.22.6" +version = "0.23.2" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "36c011a03ba1e50152b4b394b479826cad97e7a21eb52df179cd91ac411cbfbe" +checksum = "295548d5ffd95fd1981d2d3cf4458831b21d60af046b729b6fd143b0ba7aee2f" dependencies = [ "heck", "proc-macro2", diff --git a/Cargo.toml b/Cargo.toml index 349c28a..09f7903 100644 --- a/Cargo.toml +++ b/Cargo.toml @@ -43,8 +43,8 @@ pem = "=3.0" percent-encoding = "=2.3" pin-project = "1.1" pkcs8 = { version = "=0.10", features = ["encryption", "pkcs5"] } -pyo3 = { version = "=0.22", features = ["anyhow", "extension-module", "generate-import-lib"] } -pyo3-log = "=0.11" +pyo3 = { version = "=0.23", features = ["anyhow", "extension-module", "generate-import-lib"] } +pyo3-log = "=0.12" rustls-pemfile = "2.2" socket2 = { version = "0.5", features = ["all"] } tls-listener = { version = "=0.10", features = ["rustls"] } diff --git a/README.md b/README.md index 087f68d..1ad6560 100644 --- a/README.md +++ b/README.md @@ -277,7 +277,7 @@ You might test the effect such optimizations cause over your application and dec Granian is currently under active development. -Granian is compatible with Python 3.8 and above versions. +Granian is compatible with Python 3.9 and above versions. ## License diff --git a/pyproject.toml b/pyproject.toml index fb08077..55aaadc 100644 --- a/pyproject.toml +++ b/pyproject.toml @@ -11,7 +11,6 @@ classifiers = [ 'Operating System :: Microsoft :: Windows', 'Operating System :: POSIX :: Linux', 'Programming Language :: Python :: 3', - 'Programming Language :: Python :: 3.8', 'Programming Language :: Python :: 3.9', 'Programming Language :: Python :: 3.10', 'Programming Language :: Python :: 3.11', @@ -32,7 +31,7 @@ dynamic = [ 'version', ] -requires-python = '>=3.8' +requires-python = '>=3.9' dependencies = [ 'click>=8.0.0', 'uvloop>=0.18.0; sys_platform != "win32" and platform_python_implementation == "CPython"', diff --git a/src/asgi/callbacks.rs b/src/asgi/callbacks.rs index d2cc696..e437017 100644 --- a/src/asgi/callbacks.rs +++ b/src/asgi/callbacks.rs @@ -43,14 +43,14 @@ pub(crate) struct CallbackWatcherHTTP { #[pyo3(get)] proto: Py<HTTPProtocol>, #[pyo3(get)] - scope: PyObject, + scope: Py<PyDict>, } impl CallbackWatcherHTTP { pub fn new(py: Python, proto: HTTPProtocol, scope: Bound<PyDict>) -> Self { Self { proto: Py::new(py, proto).unwrap(), - scope: scope.into_py(py), + scope: scope.unbind(), } } } @@ -62,7 +62,7 @@ impl CallbackWatcherHTTP { } fn err(&self, err: Bound<PyAny>) { - callback_impl_done_err!(self, &PyErr::from_value_bound(err)); + callback_impl_done_err!(self, &PyErr::from_value(err)); } } @@ -71,14 +71,14 @@ pub(crate) struct CallbackWatcherWebsocket { #[pyo3(get)] proto: Py<WebsocketProtocol>, #[pyo3(get)] - scope: PyObject, + scope: Py<PyDict>, } impl CallbackWatcherWebsocket { pub fn new(py: Python, proto: WebsocketProtocol, scope: Bound<PyDict>) -> Self { Self { proto: Py::new(py, proto).unwrap(), - scope: scope.into_py(py), + scope: scope.unbind(), } } } @@ -90,7 +90,7 @@ impl CallbackWatcherWebsocket { } fn err(&self, err: Bound<PyAny>) { - callback_impl_done_err!(self, &PyErr::from_value_bound(err)); + callback_impl_done_err!(self, &PyErr::from_value(err)); } } diff --git a/src/asgi/io.rs b/src/asgi/io.rs index f0ccaf1..a4a112f 100644 --- a/src/asgi/io.rs +++ b/src/asgi/io.rs @@ -87,6 +87,8 @@ impl ASGIHTTPProtocol { close: bool, ) -> PyResult<Bound<'p, PyAny>> { let flow_hld = self.flow_tx_waiter.clone(); + let pynone = py.None(); + future_into_py_futlike(self.rt.clone(), py, async move { match tx.send(Ok(body.into())).await { Ok(()) => { @@ -99,7 +101,7 @@ impl ASGIHTTPProtocol { flow_hld.notify_one(); } } - Ok(()) + Ok(pynone) }) } @@ -116,9 +118,9 @@ impl ASGIHTTPProtocol { return future_into_py_futlike(self.rt.clone(), py, async move { let () = flow_hld.notified().await; Python::with_gil(\|py\| { - let dict = PyDict::new_bound(py); + let dict = PyDict::new(py); dict.set_item(pyo3::intern!(py, "type"), pyo3::intern!(py, "http.disconnect"))?; - Ok(dict.to_object(py)) + Ok(dict.into_any().unbind()) }) }); } @@ -144,18 +146,18 @@ impl ASGIHTTPProtocol { match chunk { Ok(data) => Python::with_gil(\|py\| { - let dict = PyDict::new_bound(py); + let dict = PyDict::new(py); dict.set_item(pyo3::intern!(py, "type"), pyo3::intern!(py, "http.request"))?; dict.set_item(pyo3::intern!(py, "body"), BytesToPy(data))?; dict.set_item(pyo3::intern!(py, "more_body"), more_body)?; - Ok(dict.to_object(py)) + Ok(dict.into_any().unbind()) }), _ => { flow_hld.notify_one(); Python::with_gil(\|py\| { - let dict = PyDict::new_bound(py); + let dict = PyDict::new(py); dict.set_item(pyo3::intern!(py, "type"), pyo3::intern!(py, "http.disconnect"))?; - Ok(dict.to_object(py)) + Ok(dict.into_any().unbind()) }) } } @@ -323,6 +325,7 @@ impl ASGIWebsocketProtocol { let accepted = self.accepted.clone(); let rx = self.ws_rx.clone(); let tx = self.ws_tx.clone(); + let pynone = py.None(); future_into_py_iter(self.rt.clone(), py, async move { if let Some(mut upgrade) = upgrade { @@ -339,7 +342,7 @@ impl ASGIWebsocketProtocol { wtx = Some(tx); wrx = Some(rx); accepted.store(true, atomic::Ordering::Relaxed); - return Ok(()); + return Ok(pynone); } } } @@ -352,15 +355,16 @@ impl ASGIWebsocketProtocol { fn send_message<'p>(&self, py: Python<'p>, data: Message) -> PyResult<Bound<'p, PyAny>> { let transport = self.ws_tx.clone(); let closed = self.closed.clone(); + let pynone = py.None(); future_into_py_futlike(self.rt.clone(), py, async move { if let Some(ws) = &mut (transport.lock().await) { match ws.send(data).await { - Ok(()) => return Ok(()), + Ok(()) => return Ok(pynone), _ => { if closed.load(atomic::Ordering::Relaxed) { log::info!("Attempted to write to a closed websocket"); - return Ok(()); + return Ok(pynone); } } }; @@ -374,6 +378,7 @@ impl ASGIWebsocketProtocol { let closed = self.closed.clone(); let ws_rx = self.ws_rx.clone(); let ws_tx = self.ws_tx.clone(); + let pynone = py.None(); future_into_py_iter(self.rt.clone(), py, async move { if let Some(tx) = ws_tx.lock().await.take() { @@ -382,7 +387,7 @@ impl ASGIWebsocketProtocol { .close() .await; } - Ok(()) + Ok(pynone) }) } @@ -416,9 +421,9 @@ impl ASGIWebsocketProtocol { let accepted = accepted.load(atomic::Ordering::Relaxed); if !accepted { return Python::with_gil(\|py\| { - let dict = PyDict::new_bound(py); + let dict = PyDict::new(py); dict.set_item(pyo3::intern!(py, "type"), pyo3::intern!(py, "websocket.connect"))?; - Ok(dict.to_object(py)) + Ok(dict.into_any().unbind()) }); } @@ -444,7 +449,7 @@ impl ASGIWebsocketProtocol { Ok(ASGIMessageType::WSAccept(subproto)) => self.accept(py, subproto), Ok(ASGIMessageType::WSClose) => self.close(py), Ok(ASGIMessageType::WSMessage(message)) => self.send_message(py, message), - _ => future_into_py_iter::<_, _, PyErr>(self.rt.clone(), py, async { error_message!() }), + _ => future_into_py_iter::<_, _>(self.rt.clone(), py, async { error_message!() }), } } } @@ -549,26 +554,26 @@ fn ws_message_into_rs(py: Python, message: &Bound<PyDict>) -> PyResult<Message> fn ws_message_into_py(message: Message) -> PyResult<PyObject> { match message { Message::Binary(message) => Python::with_gil(\|py\| { - let dict = PyDict::new_bound(py); + let dict = PyDict::new(py); dict.set_item(pyo3::intern!(py, "type"), pyo3::intern!(py, "websocket.receive"))?; - dict.set_item(pyo3::intern!(py, "bytes"), PyBytes::new_bound(py, &message[..]))?; - Ok(dict.to_object(py)) + dict.set_item(pyo3::intern!(py, "bytes"), PyBytes::new(py, &message[..]))?; + Ok(dict.into_any().unbind()) }), Message::Text(message) => Python::with_gil(\|py\| { - let dict = PyDict::new_bound(py); + let dict = PyDict::new(py); dict.set_item(pyo3::intern!(py, "type"), pyo3::intern!(py, "websocket.receive"))?; dict.set_item(pyo3::intern!(py, "text"), message)?; - Ok(dict.to_object(py)) + Ok(dict.into_any().unbind()) }), Message::Close(frame) => Python::with_gil(\|py\| { let close_code: u16 = match frame { Some(frame) => frame.code.into(), _ => 1005, }; - let dict = PyDict::new_bound(py); + let dict = PyDict::new(py); dict.set_item(pyo3::intern!(py, "type"), pyo3::intern!(py, "websocket.disconnect"))?; dict.set_item(pyo3::intern!(py, "code"), close_code)?; - Ok(dict.to_object(py)) + Ok(dict.into_any().unbind()) }), v => { log::warn!("Unsupported websocket message received {:?}", v); diff --git a/src/asgi/utils.rs b/src/asgi/utils.rs index 0c8f2e1..3561892 100644 --- a/src/asgi/utils.rs +++ b/src/asgi/utils.rs @@ -42,13 +42,13 @@ pub(super) fn build_scope<'p>( path: &'p str, query_string: &'p str, ) -> PyResult<Bound<'p, PyDict>> { - let scope = PyDict::new_bound(py); + let scope = PyDict::new(py); scope.set_item( pyo3::intern!(py, "asgi"), ASGI_VERSION .get_or_try_init(py, \|\| { - let rv = PyDict::new_bound(py); + let rv = PyDict::new(py); rv.set_item("version", "3.0")?; rv.set_item("spec_version", "2.3")?; Ok::<PyObject, PyErr>(rv.into()) @@ -59,8 +59,8 @@ pub(super) fn build_scope<'p>( pyo3::intern!(py, "extensions"), ASGI_EXTENSIONS .get_or_try_init(py, \|\| { - let rv = PyDict::new_bound(py); - rv.set_item("http.response.pathsend", PyDict::new_bound(py))?; + let rv = PyDict::new(py); + rv.set_item("http.response.pathsend", PyDict::new(py))?; Ok::<PyObject, PyErr>(rv.into()) })? .bind(py), @@ -71,25 +71,22 @@ pub(super) fn build_scope<'p>( scope.set_item(pyo3::intern!(py, "client"), client)?; scope.set_item(pyo3::intern!(py, "scheme"), scheme)?; scope.set_item(pyo3::intern!(py, "path"), path)?; - scope.set_item(pyo3::intern!(py, "raw_path"), PyBytes::new_bound(py, path.as_bytes()))?; + scope.set_item(pyo3::intern!(py, "raw_path"), PyBytes::new(py, path.as_bytes()))?; scope.set_item( pyo3::intern!(py, "query_string"), - PyBytes::new_bound(py, query_string.as_bytes()), + PyBytes::new(py, query_string.as_bytes()), )?; - let headers = PyList::empty_bound(py); + let headers = PyList::empty(py); for (key, value) in &req.headers { headers.append(( - PyBytes::new_bound(py, key.as_str().as_bytes()), - PyBytes::new_bound(py, value.as_bytes()), + PyBytes::new(py, key.as_str().as_bytes()), + PyBytes::new(py, value.as_bytes()), ))?; } if !req.headers.contains_key(header::HOST) { let host = req.uri.authority().map_or("", Authority::as_str); - headers.insert( - 0, - (PyBytes::new_bound(py, b"host"), PyBytes::new_bound(py, host.as_bytes())), - )?; + headers.insert(0, (PyBytes::new(py, b"host"), PyBytes::new(py, host.as_bytes())))?; } scope.set_item(pyo3::intern!(py, "headers"), headers)?; @@ -136,14 +133,14 @@ pub(super) fn build_scope_ws<'p>( )?; scope.set_item( pyo3::intern!(py, "subprotocols"), - PyList::new_bound( + PyList::new( py, req.headers .get_all("Sec-WebSocket-Protocol") .iter() - .map(\|v\| PyString::new_bound(py, v.to_str().unwrap())) + .map(\|v\| PyString::new(py, v.to_str().unwrap())) .collect::<Vec<Bound<PyString>>>(), - ), + )?, )?; Ok(scope) } diff --git a/src/asyncio.rs b/src/asyncio.rs index f6b1b38..3376c8f 100644 --- a/src/asyncio.rs +++ b/src/asyncio.rs @@ -6,7 +6,7 @@ static CONTEXT: GILOnceCell<PyObject> = GILOnceCell::new(); fn contextvars(py: Python) -> PyResult<&Bound<PyAny>> { Ok(CONTEXTVARS - .get_or_try_init(py, \|\| py.import_bound("contextvars").map(Into::into))? + .get_or_try_init(py, \|\| py.import("contextvars").map(Into::into))? .bind(py)) } diff --git a/src/callbacks.rs b/src/callbacks.rs index 2944fbf..d83583b 100644 --- a/src/callbacks.rs +++ b/src/callbacks.rs @@ -78,9 +78,9 @@ impl PyIterAwaitable { } } - pub(crate) fn set_result(&self, result: PyResult<impl IntoPy<PyObject>>) { + pub(crate) fn set_result(&self, result: PyResult<PyObject>) { let mut res = self.result.write().unwrap(); - res = Some(Python::with_gil(\|py\| result.map(\|v\| v.into_py(py)))); + res = Some(result); } } @@ -138,17 +138,17 @@ impl PyFutureAwaitable { Ok((Py::new(py, self)?, cancel_tx)) } - pub(crate) fn set_result(&self, result: PyResult<impl IntoPy<PyObject>>, aw: Py<PyFutureAwaitable>) { + pub(crate) fn set_result(&self, result: PyResult<PyObject>, aw: Py<PyFutureAwaitable>) { Python::with_gil(\|py\| { let mut state = self.state.write().unwrap(); if !matches!(&mut state, PyFutureAwaitableState::Pending) { return; } - state = PyFutureAwaitableState::Completed(result.map(\|v\| v.into_py(py))); + state = PyFutureAwaitableState::Completed(result); let ack = self.ack.read().unwrap(); if let Some((cb, ctx)) = &ack { - let _ = self.event_loop.clone_ref(py).call_method_bound( + let _ = self.event_loop.clone_ref(py).call_method( py, pyo3::intern!(py, "call_soon_threadsafe"), (cb, aw), @@ -198,7 +198,7 @@ impl PyFutureAwaitable { #[pyo3(signature = (cb, context=None))] fn add_done_callback(pyself: PyRef<'_, Self>, cb: PyObject, context: Option<PyObject>) -> PyResult<()> { let py = pyself.py(); - let kwctx = pyo3::types::PyDict::new_bound(py); + let kwctx = pyo3::types::PyDict::new(py); kwctx.set_item(pyo3::intern!(py, "context"), context)?; let state = pyself.state.read().unwrap(); @@ -211,7 +211,7 @@ impl PyFutureAwaitable { _ => { drop(state); let event_loop = pyself.event_loop.clone_ref(py); - event_loop.call_method_bound(py, pyo3::intern!(py, "call_soon"), (cb, pyself), Some(&kwctx))?; + event_loop.call_method(py, pyo3::intern!(py, "call_soon"), (cb, pyself), Some(&kwctx))?; Ok(()) } } @@ -244,7 +244,7 @@ impl PyFutureAwaitable { drop(ack); drop(state); - let _ = event_loop.call_method_bound(py, pyo3::intern!(py, "call_soon"), (cb, pyself), Some(ctx.bind(py))); + let _ = event_loop.call_method(py, pyo3::intern!(py, "call_soon"), (cb, pyself), Some(ctx.bind(py))); } true diff --git a/src/conversion.rs b/src/conversion.rs index 56c1884..875dbdc 100644 --- a/src/conversion.rs +++ b/src/conversion.rs @@ -1,20 +1,16 @@ -use pyo3::prelude::; +use pyo3::{prelude::, IntoPyObjectExt}; use crate::workers::{HTTP1Config, HTTP2Config}; pub(crate) struct BytesToPy(pub hyper::body::Bytes); -impl IntoPy<PyObject> for BytesToPy { - #[inline] - fn into_py(self, py: Python) -> PyObject { - self.0.as_ref().into_py(py) - } -} +impl<'p> IntoPyObject<'p> for BytesToPy { + type Target = PyAny; + type Output = Bound<'p, Self::Target>; + type Error = PyErr; -impl ToPyObject for BytesToPy { - #[inline] - fn to_object(&self, py: Python<'_>) -> PyObject { - self.0.as_ref().into_py(py) + fn into_pyobject(self, py: Python<'p>) -> Result<Self::Output, Self::Error> { + self.0.as_ref().into_bound_py_any(py) } } diff --git a/src/rsgi/callbacks.rs b/src/rsgi/callbacks.rs index b67081b..94a97c7 100644 --- a/src/rsgi/callbacks.rs +++ b/src/rsgi/callbacks.rs @@ -38,14 +38,14 @@ pub(crate) struct CallbackWatcherHTTP { #[pyo3(get)] proto: Py<HTTPProtocol>, #[pyo3(get)] - scope: PyObject, + scope: Py<HTTPScope>, } impl CallbackWatcherHTTP { pub fn new(py: Python, proto: HTTPProtocol, scope: HTTPScope) -> Self { Self { proto: Py::new(py, proto).unwrap(), - scope: scope.into_py(py), + scope: Py::new(py, scope).unwrap(), } } } @@ -57,7 +57,7 @@ impl CallbackWatcherHTTP { } fn err(&self, err: Bound<PyAny>) { - callback_impl_done_err!(self, &PyErr::from_value_bound(err)); + callback_impl_done_err!(self, &PyErr::from_value(err)); } } @@ -66,14 +66,14 @@ pub(crate) struct CallbackWatcherWebsocket { #[pyo3(get)] proto: Py<WebsocketProtocol>, #[pyo3(get)] - scope: PyObject, + scope: Py<WebsocketScope>, } impl CallbackWatcherWebsocket { pub fn new(py: Python, proto: WebsocketProtocol, scope: WebsocketScope) -> Self { Self { proto: Py::new(py, proto).unwrap(), - scope: scope.into_py(py), + scope: Py::new(py, scope).unwrap(), } } } @@ -85,7 +85,7 @@ impl CallbackWatcherWebsocket { } fn err(&self, err: Bound<PyAny>) { - callback_impl_done_err!(self, &PyErr::from_value_bound(err)); + callback_impl_done_err!(self, &PyErr::from_value(err)); } } diff --git a/src/rsgi/io.rs b/src/rsgi/io.rs index cd813ea..cbf0f3a 100644 --- a/src/rsgi/io.rs +++ b/src/rsgi/io.rs @@ -1,9 +1,12 @@ use futures::{sink::SinkExt, StreamExt, TryStreamExt}; use http_body_util::BodyExt; use hyper::body; -use pyo3::prelude::; -use pyo3::pybacked::PyBackedStr; -use pyo3::types::{PyBytes, PyString}; +use pyo3::{ + prelude::, + pybacked::PyBackedStr, + types::{PyBytes, PyString}, + IntoPyObjectExt, +}; use std::{ borrow::Cow, sync::{atomic, Arc, Mutex, RwLock}, @@ -40,9 +43,11 @@ impl RSGIHTTPStreamTransport { fn send_bytes<'p>(&self, py: Python<'p>, data: Cow<[u8]>) -> PyResult<Bound<'p, PyAny>> { let transport = self.tx.clone(); let bdata: Box<[u8]> = data.into(); + let pynone = py.None(); + future_into_py_futlike(self.rt.clone(), py, async move { match transport.send(Ok(body::Bytes::from(bdata))).await { - Ok(()) => Ok(()), + Ok(()) => Ok(pynone), _ => error_stream!(), } }) @@ -50,9 +55,11 @@ impl RSGIHTTPStreamTransport { fn send_str<'p>(&self, py: Python<'p>, data: String) -> PyResult<Bound<'p, PyAny>> { let transport = self.tx.clone(); + let pynone = py.None(); + future_into_py_futlike(self.rt.clone(), py, async move { match transport.send(Ok(body::Bytes::from(data))).await { - Ok(()) => Ok(()), + Ok(()) => Ok(pynone), _ => error_stream!(), } }) @@ -88,7 +95,10 @@ impl RSGIHTTPProtocol { if let Some(body) = self.body.lock().unwrap().take() { return future_into_py_iter(self.rt.clone(), py, async move { match body.collect().await { - Ok(data) => Ok(BytesToPy(data.to_bytes())), + Ok(data) => { + let bytes = BytesToPy(data.to_bytes()); + Ok(Python::with_gil(\|py\| bytes.into_py_any(py))?) + } _ => error_stream!(), } }); @@ -125,7 +135,7 @@ impl RSGIHTTPProtocol { BytesToPy(body::Bytes::new()) } }; - Ok(bytes) + Python::with_gil(\|py\| bytes.into_py_any(py)) }) } @@ -240,10 +250,12 @@ impl RSGIWebsocketTransport { fn send_bytes<'p>(&self, py: Python<'p>, data: Cow<[u8]>) -> PyResult<Bound<'p, PyAny>> { let transport = self.tx.clone(); let bdata: Box<[u8]> = data.into(); + let pynone = py.None(); + future_into_py_iter(self.rt.clone(), py, async move { if let Ok(mut stream) = transport.try_lock() { return match stream.send(bdata[..].into()).await { - Ok(()) => Ok(()), + Ok(()) => Ok(pynone), _ => error_stream!(), }; } @@ -253,10 +265,12 @@ impl RSGIWebsocketTransport { fn send_str<'p>(&self, py: Python<'p>, data: String) -> PyResult<Bound<'p, PyAny>> { let transport = self.tx.clone(); + let pynone = py.None(); + future_into_py_iter(self.rt.clone(), py, async move { if let Ok(mut stream) = transport.try_lock() { return match stream.send(Message::Text(data)).await { - Ok(()) => Ok(()), + Ok(()) => Ok(pynone), _ => error_stream!(), }; } @@ -379,7 +393,7 @@ impl RSGIWebsocketProtocol { Ok(Python::with_gil(\|py\| { let pytransport = Py::new(py, RSGIWebsocketTransport::new(rth, stream)).unwrap(); trx = Some(pytransport.clone_ref(py)); - pytransport + pytransport.into_any() })) } _ => error_proto!(), @@ -394,12 +408,14 @@ impl RSGIWebsocketProtocol { fn message_into_py(message: Message) -> PyResult<PyObject> { match message { Message::Binary(message) => Ok(Python::with_gil(\|py\| { - WebsocketInboundBytesMessage::new(PyBytes::new_bound(py, &message).unbind()).into_py(py) - })), + WebsocketInboundBytesMessage::new(PyBytes::new(py, &message).unbind()).into_py_any(py) + })?), Message::Text(message) => Ok(Python::with_gil(\|py\| { - WebsocketInboundTextMessage::new(PyString::new_bound(py, &message).unbind()).into_py(py) - })), - Message::Close(_) => Ok(Python::with_gil(\|py\| WebsocketInboundCloseMessage::new().into_py(py))), + WebsocketInboundTextMessage::new(PyString::new(py, &message).unbind()).into_py_any(py) + })?), + Message::Close(_) => Ok(Python::with_gil(\|py\| { + WebsocketInboundCloseMessage::new().into_py_any(py) + })?), v => { log::warn!("Unsupported websocket message received {:?}", v); error_proto!() diff --git a/src/rsgi/mod.rs b/src/rsgi/mod.rs index e27bf4d..6c18eed 100644 --- a/src/rsgi/mod.rs +++ b/src/rsgi/mod.rs @@ -8,8 +8,8 @@ pub(crate) mod serve; mod types; pub(crate) fn init_pymodule(py: Python, module: &Bound<PyModule>) -> PyResult<()> { - module.add("RSGIProtocolError", py.get_type_bound::<errors::RSGIProtocolError>())?; - module.add("RSGIProtocolClosed", py.get_type_bound::<errors::RSGIProtocolClosed>())?; + module.add("RSGIProtocolError", py.get_type::<errors::RSGIProtocolError>())?; + module.add("RSGIProtocolClosed", py.get_type::<errors::RSGIProtocolClosed>())?; module.add_class::<io::RSGIHTTPProtocol>()?; module.add_class::<io::RSGIHTTPStreamTransport>()?; module.add_class::<io::RSGIWebsocketProtocol>()?; diff --git a/src/rsgi/types.rs b/src/rsgi/types.rs index 49d86bd..0f38863 100644 --- a/src/rsgi/types.rs +++ b/src/rsgi/types.rs @@ -68,7 +68,7 @@ impl RSGIHeaders { } fn __iter__<'p>(&self, py: Python<'p>) -> PyResult<Bound<'p, PyIterator>> { - PyIterator::from_bound_object(&PyList::new_bound(py, self.keys())) + PyIterator::from_object(PyList::new(py, self.keys())?.as_any()) } fn __len__(&self) -> usize { @@ -79,7 +79,7 @@ impl RSGIHeaders { fn get(&self, py: Python, key: &str, default: Option<PyObject>) -> Option<PyObject> { match self.inner.get(key) { Some(val) => match val.to_str() { - Ok(string) => Some(PyString::new_bound(py, string).into()), + Ok(string) => Some(PyString::new(py, string).into()), _ => default, }, _ => default, @@ -87,13 +87,13 @@ impl RSGIHeaders { } #[pyo3(signature = (key))] - fn get_all<'p>(&self, py: Python<'p>, key: &'p str) -> Bound<'p, PyList> { - PyList::new_bound( + fn get_all<'p>(&self, py: Python<'p>, key: &'p str) -> PyResult<Bound<'p, PyList>> { + PyList::new( py, self.inner .get_all(key) .iter() - .map(\|v\| PyString::new_bound(py, v.to_str().unwrap())) + .map(\|v\| PyString::new(py, v.to_str().unwrap())) .collect::<Vec<Bound<PyString>>>(), ) } diff --git a/src/runtime.rs b/src/runtime.rs index df785b9..aff85c1 100644 --- a/src/runtime.rs +++ b/src/runtime.rs @@ -1,4 +1,6 @@ use pyo3::prelude::; +#[cfg(windows)] +use pyo3::IntoPyObjectExt; use std::{ future::Future, sync::{Arc, Mutex}, @@ -141,11 +143,10 @@ pub(crate) fn init_runtime_st(blocking_threads: usize, py_loop: Arc<PyObject>) - // but for "quick" operations it's something like 12% faster. #[allow(unused_must_use)] #[cfg(not(target_os = "linux"))] -pub(crate) fn future_into_py_iter<R, F, T>(rt: R, py: Python, fut: F) -> PyResult<Bound<PyAny>> +pub(crate) fn future_into_py_iter<R, F>(rt: R, py: Python, fut: F) -> PyResult<Bound<PyAny>> where R: Runtime + ContextExt + Clone, - F: Future<Output = PyResult<T>> + Send + 'static, - T: IntoPy<PyObject> + Send + 'static, + F: Future<Output = PyResult<PyObject>> + Send + 'static, { let aw = Py::new(py, PyIterAwaitable::new())?; let py_fut = aw.clone_ref(py); @@ -168,11 +169,10 @@ where // MacOS works best with original impl, Windows still needs further analysis. #[cfg(target_os = "linux")] #[inline(always)] -pub(crate) fn future_into_py_iter<R, F, T>(rt: R, py: Python, fut: F) -> PyResult<Bound<PyAny>> +pub(crate) fn future_into_py_iter<R, F>(rt: R, py: Python, fut: F) -> PyResult<Bound<PyAny>> where R: Runtime + ContextExt + Clone, - F: Future<Output = PyResult<T>> + Send + 'static, - T: IntoPy<PyObject> + Send + 'static, + F: Future<Output = PyResult<PyObject>> + Send + 'static, { future_into_py_futlike(rt, py, fut) } @@ -184,11 +184,10 @@ where // and for "long" operations it's something like 6% faster than `future_into_py_iter`. #[allow(unused_must_use)] #[cfg(unix)] -pub(crate) fn future_into_py_futlike<R, F, T>(rt: R, py: Python, fut: F) -> PyResult<Bound<PyAny>> +pub(crate) fn future_into_py_futlike<R, F>(rt: R, py: Python, fut: F) -> PyResult<Bound<PyAny>> where R: Runtime + ContextExt + Clone, - F: Future<Output = PyResult<T>> + Send + 'static, - T: IntoPy<PyObject> + Send + 'static, + F: Future<Output = PyResult<PyObject>> + Send + 'static, { let event_loop = rt.py_event_loop(py); let (aw, cancel_tx) = PyFutureAwaitable::new(event_loop).to_spawn(py)?; @@ -220,11 +219,10 @@ where #[allow(unused_must_use)] #[cfg(windows)] -pub(crate) fn future_into_py_futlike<R, F, T>(rt: R, py: Python, fut: F) -> PyResult<Bound<PyAny>> +pub(crate) fn future_into_py_futlike<R, F>(rt: R, py: Python, fut: F) -> PyResult<Bound<PyAny>> where R: Runtime + ContextExt + Clone, - F: Future<Output = PyResult<T>> + Send + 'static, - T: IntoPy<PyObject> + Send + 'static, + F: Future<Output = PyResult<PyObject>> + Send + 'static, { let event_loop = rt.py_event_loop(py); let event_loop_ref = event_loop.clone_ref(py); @@ -247,8 +245,8 @@ where let _ = rb.run(move \|\| { Python::with_gil(\|py\| { let (cb, value) = match result { - Ok(val) => (fut_ref.getattr(py, pyo3::intern!(py, "set_result")).unwrap(), val.into_py(py)), - Err(err) => (fut_ref.getattr(py, pyo3::intern!(py, "set_exception")).unwrap(), err.into_py(py)) + Ok(val) => (fut_ref.getattr(py, pyo3::intern!(py, "set_result")).unwrap(), val), + Err(err) => (fut_ref.getattr(py, pyo3::intern!(py, "set_exception")).unwrap(), err.into_py_any(py).unwrap()) }; let _ = event_loop_ref.call_method1(py, pyo3::intern!(py, "call_soon_threadsafe"), (PyFutureResultSetter, cb, value)); drop(fut_ref); @@ -258,7 +256,7 @@ where }, () = cancel_tx.notified() => { let _ = rb.run(move \|\| { - Python::with_gil(\|py\| { + Python::with_gil(\|_\| { drop(fut_ref); drop(event_loop_ref); }); @@ -273,35 +271,33 @@ where #[allow(clippy::unnecessary_wraps)] #[inline(always)] pub(crate) fn empty_future_into_py(py: Python) -> PyResult<Bound<PyAny>> { - Ok(PyEmptyAwaitable.into_py(py).into_bound(py)) + Ok(PyEmptyAwaitable.into_pyobject(py)?.into_any()) } #[allow(unused_must_use)] -pub(crate) fn run_until_complete<R, F, T>(rt: R, event_loop: Bound<PyAny>, fut: F) -> PyResult<T> +pub(crate) fn run_until_complete<R, F>(rt: R, event_loop: Bound<PyAny>, fut: F) -> PyResult<()> where R: Runtime + ContextExt + Clone, - F: Future<Output = PyResult<T>> + Send + 'static, - T: Send + Sync + 'static, + F: Future<Output = PyResult<()>> + Send + 'static, { - let py = event_loop.py(); let result_tx = Arc::new(Mutex::new(None)); let result_rx = Arc::clone(&result_tx); let py_fut = event_loop.call_method0("create_future")?; - let loop_tx = event_loop.clone().into_py(py); - let future_tx = py_fut.clone().into_py(py); + let loop_tx = event_loop.clone().unbind(); + let future_tx = py_fut.clone().unbind(); rt.spawn(async move { - let val = fut.await; + let _ = fut.await; if let Ok(mut result) = result_tx.lock() { - result = Some(val.unwrap()); + result = Some(()); } // NOTE: we don't care if we block the runtime. // `run_until_complete` is used only for the workers main loop. Python::with_gil(move \|py\| { let res_method = future_tx.getattr(py, "set_result").unwrap(); - let _ = loop_tx.call_method_bound(py, "call_soon_threadsafe", (res_method, py.None()), None); + let _ = loop_tx.call_method(py, "call_soon_threadsafe", (res_method, py.None()), None); drop(future_tx); drop(loop_tx); }); @@ -309,8 +305,8 @@ where event_loop.call_method1("run_until_complete", (py_fut,))?; - let result = result_rx.lock().unwrap().take().unwrap(); - Ok(result) + result_rx.lock().unwrap().take().unwrap(); + Ok(()) } pub(crate) fn block_on_local<F>(rt: &RuntimeWrapper, local: LocalSet, fut: F) diff --git a/src/tcp.rs b/src/tcp.rs index f604966..dd62f58 100644 --- a/src/tcp.rs +++ b/src/tcp.rs @@ -1,5 +1,4 @@ -use pyo3::prelude::; -use pyo3::types::PyType; +use pyo3::{prelude::*, types::PyType, IntoPyObjectExt}; use std::net::{IpAddr, SocketAddr, TcpListener}; #[cfg(unix)] @@ -53,23 +52,23 @@ impl ListenerHolder { #[cfg(unix)] pub fn __getstate__(&self, py: Python) -> PyObject { let fd = self.socket.as_raw_fd(); - (fd.into_py(py),).to_object(py) + (fd,).into_py_any(py).unwrap() } #[cfg(windows)] pub fn __getstate__(&self, py: Python) -> PyObject { let fd = self.socket.as_raw_socket(); - (fd.into_py(py),).to_object(py) + (fd,).into_py_any(py).unwrap() } #[cfg(unix)] pub fn get_fd(&self, py: Python) -> PyObject { - self.socket.as_raw_fd().into_py(py).to_object(py) + self.socket.as_raw_fd().into_py_any(py).unwrap() } #[cfg(windows)] pub fn get_fd(&self, py: Python) -> PyObject { - self.socket.as_raw_socket().into_py(py).to_object(py) + self.socket.as_raw_socket().into_py_any(py).unwrap() } } diff --git a/src/utils.rs b/src/utils.rs index 711c95e..4ad10e8 100644 --- a/src/utils.rs +++ b/src/utils.rs @@ -43,7 +43,7 @@ fn trim_end(data: &[u8]) -> &[u8] { #[inline] pub(crate) fn log_application_callable_exception(err: &pyo3::PyErr) { let tb = pyo3::Python::with_gil(\|py\| { - let tb = match err.traceback_bound(py).map(\|t\| t.format()) { + let tb = match err.traceback(py).map(\|t\| t.format()) { Some(Ok(tb)) => tb, _ => String::new(), }; diff --git a/src/wsgi/callbacks.rs b/src/wsgi/callbacks.rs index 49ca894..5946292 100644 --- a/src/wsgi/callbacks.rs +++ b/src/wsgi/callbacks.rs @@ -67,7 +67,7 @@ fn run_callback( let _ = Python::with_gil(\|py\| -> PyResult<()> { let proto = Py::new(py, WSGIProtocol::new(tx))?; let callback = cbs.get().cb.clone_ref(py); - let environ = PyDict::new_bound(py); + let environ = PyDict::new(py); environ.set_item(pyo3::intern!(py, "SERVER_PROTOCOL"), version)?; environ.set_item(pyo3::intern!(py, "SERVER_NAME"), server.0)?; environ.set_item(pyo3::intern!(py, "SERVER_PORT"), server.1)?; @@ -75,7 +75,7 @@ fn run_callback( environ.set_item(pyo3::intern!(py, "REQUEST_METHOD"), parts.method.as_str())?; environ.set_item( pyo3::intern!(py, "PATH_INFO"), - PyBytes::new_bound(py, &path).call_method1(pyo3::intern!(py, "decode"), (pyo3::intern!(py, "latin1"),))?, + PyBytes::new(py, &path).call_method1(pyo3::intern!(py, "decode"), (pyo3::intern!(py, "latin1"),))?, )?; environ.set_item(pyo3::intern!(py, "QUERY_STRING"), query_string)?; environ.set_item(pyo3::intern!(py, "wsgi.url_scheme"), scheme)?; @@ -92,7 +92,7 @@ fn run_callback( content_len.to_str().unwrap_or_default(), )?; } - environ.update(headers.into_py_dict_bound(py).as_mapping())?; + environ.update(headers.into_py_dict(py).unwrap().as_mapping())?; if let Err(err) = callback.call1(py, (proto.clone_ref(py), environ)) { log_application_callable_exception(&err); diff --git a/src/wsgi/types.rs b/src/wsgi/types.rs index deeb567..9a26233 100644 --- a/src/wsgi/types.rs +++ b/src/wsgi/types.rs @@ -147,7 +147,7 @@ impl WSGIBody { } #[pyo3(signature = (_hint=None))] - fn readlines<'p>(&self, py: Python<'p>, _hint: Option<PyObject>) -> Bound<'p, PyList> { + fn readlines<'p>(&self, py: Python<'p>, _hint: Option<PyObject>) -> PyResult<Bound<'p, PyList>> { let inner = self.inner.clone(); let data = py.allow_threads(\|\| { self.rt.inner.block_on(async move { @@ -159,8 +159,8 @@ impl WSGIBody { }); let lines: Vec<Bound<PyBytes>> = data .split(\|&c\| c == LINE_SPLIT) - .map(\|item\| PyBytes::new_bound(py, item)) + .map(\|item\| PyBytes::new(py, item)) .collect(); - PyList::new_bound(py, lines) + PyList::new(py, lines) } }	Drop Python 3.8 support As it reached EOL		2024-12-01T23:18:50	[]	[]
emmett-framework/granian	emmett-framework__granian-452	1332b2a3c6f05c8d68cb99d49a5c18bbb4eef3fc	diff --git a/.github/workflows/build.yml b/.github/workflows/build.yml index cc875242..e2f9c11c 100644 --- a/.github/workflows/build.yml +++ b/.github/workflows/build.yml @@ -118,7 +118,6 @@ jobs: pip install .[test] pip install granian --no-index --no-deps --find-links pgo_wheel --force-reinstall PGO_RUN=y pytest tests - PGO_RUN=y LOOP_OPT=y pytest tests/test_asgi.py tests/test_rsgi.py - name: merge PGO data run: ${{ env.LLVM_PROFDATA }} merge -o ${{ github.workspace }}/merged.profdata ${{ github.workspace }}/profdata - name: Build PGO wheel diff --git a/.github/workflows/release.yml b/.github/workflows/release.yml index b5b7928e..f2d30443 100644 --- a/.github/workflows/release.yml +++ b/.github/workflows/release.yml @@ -139,7 +139,6 @@ jobs: pip install .[test] pip install granian --no-index --no-deps --find-links pgo_wheel --force-reinstall PGO_RUN=y pytest tests - PGO_RUN=y LOOP_OPT=y pytest tests/test_asgi.py tests/test_rsgi.py - name: merge PGO data run: ${{ env.LLVM_PROFDATA }} merge -o ${{ github.workspace }}/merged.profdata ${{ github.workspace }}/profdata - name: Build PGO wheel diff --git a/README.md b/README.md index 0bb22b1f..087f68d1 100644 --- a/README.md +++ b/README.md @@ -100,8 +100,6 @@ Options: GRANIAN_THREADING_MODE; default: (workers)] --loop [auto\|asyncio\|uvloop] Event loop implementation [env var: GRANIAN_LOOP; default: (auto)] - --opt / --no-opt Enable loop optimizations [env var: - GRANIAN_LOOP_OPT; default: (disabled)] --backlog INTEGER RANGE Maximum number of connections to hold in backlog (globally) [env var: GRANIAN_BACKLOG; default: 1024; x>=128] diff --git a/granian/_futures.py b/granian/_futures.py index 2abeee9f..3084630f 100644 --- a/granian/_futures.py +++ b/granian/_futures.py @@ -1,4 +1,9 @@ -def future_watcher_wrapper(inner): +from asyncio.tasks import _enter_task, _leave_task + +from ._granian import CallbackScheduler as _BaseCBScheduler + + +def _future_watcher_wrapper(inner): async def future_watcher(watcher): try: await inner(watcher.scope, watcher.proto) @@ -8,3 +13,59 @@ async def future_watcher(watcher): watcher.done() return future_watcher + + +class _CBScheduler(_BaseCBScheduler): + __slots__ = [] + + def __init__(self, loop, ctx, cb): + super().__init__() + self._schedule_fn = _cbsched_schedule(loop, ctx, self._run, cb) + + def _waker(self, coro): + def _wake(fut): + self._resume(coro, fut) + + return _wake + + def _resume(self, coro, fut): + try: + fut.result() + except BaseException as exc: + self._throw(coro, exc) + else: + self._run(coro) + + def _run(self, coro): + _enter_task(self._loop, self) + try: + try: + result = coro.send(None) + except (KeyboardInterrupt, SystemExit): + raise + except BaseException: + pass + else: + if getattr(result, '_asyncio_future_blocking', None): + result._asyncio_future_blocking = False + result.add_done_callback(self._waker(coro), context=self._ctx) + elif result is None: + self._loop.call_soon(self._run, coro, context=self._ctx) + finally: + _leave_task(self._loop, self) + + def _throw(self, coro, exc): + _enter_task(self._loop, self) + try: + coro.throw(exc) + except BaseException: + pass + finally: + _leave_task(self._loop, self) + + +def _cbsched_schedule(loop, ctx, run, cb): + def _schedule(watcher): + loop.call_soon_threadsafe(run, cb(watcher), context=ctx) + + return _schedule diff --git a/granian/_granian.pyi b/granian/_granian.pyi index 55c2134a..0072f7c7 100644 --- a/granian/_granian.pyi +++ b/granian/_granian.pyi @@ -107,3 +107,7 @@ class ListenerHolder: @classmethod def from_address(cls, address: str, port: int, backlog: int) -> ListenerHolder: ... def get_fd(self) -> Any: ... + +class CallbackScheduler: + _loop: Any + _ctx: Any diff --git a/granian/cli.py b/granian/cli.py index 7060e5f3..bb5cbdb9 100644 --- a/granian/cli.py +++ b/granian/cli.py @@ -77,7 +77,6 @@ def option(param_decls: str, cls: Optional[Type[click.Option]] = None, attrs: help='Threading mode to use', ) @option('--loop', type=EnumType(Loops), default=Loops.auto, help='Event loop implementation') -@option('--opt/--no-opt', 'loop_opt', default=False, help='Enable loop optimizations') @option( '--backlog', type=click.IntRange(128), @@ -256,7 +255,6 @@ def cli( blocking_threads: Optional[int], threading_mode: ThreadModes, loop: Loops, - loop_opt: bool, backlog: int, backpressure: Optional[int], http1_buffer_size: int, @@ -311,7 +309,6 @@ def cli( blocking_threads=blocking_threads, threading_mode=threading_mode, loop=loop, - loop_opt=loop_opt, http=http, websockets=websockets, backlog=backlog, diff --git a/granian/server.py b/granian/server.py index bf658476..1e78f38b 100644 --- a/granian/server.py +++ b/granian/server.py @@ -13,7 +13,7 @@ from pathlib import Path from typing import Any, Callable, Dict, List, Optional, Sequence, Tuple, Type -from ._futures import future_watcher_wrapper +from ._futures import _CBScheduler, _future_watcher_wrapper from ._granian import ASGIWorker, RSGIWorker, WSGIWorker from ._imports import setproctitle, watchfiles from ._internal import load_target @@ -78,7 +78,6 @@ def __init__( blocking_threads: Optional[int] = None, threading_mode: ThreadModes = ThreadModes.workers, loop: Loops = Loops.auto, - loop_opt: bool = False, http: HTTPModes = HTTPModes.auto, websockets: bool = True, backlog: int = 1024, @@ -115,7 +114,6 @@ def __init__( self.threads = max(1, threads) self.threading_mode = threading_mode self.loop = loop - self.loop_opt = loop_opt self.http = http self.websockets = websockets self.backlog = max(128, backlog) @@ -189,7 +187,6 @@ def _spawn_asgi_worker( http1_settings: Optional[HTTP1Settings], http2_settings: Optional[HTTP2Settings], websockets: bool, - loop_opt: bool, log_enabled: bool, log_level: LogLevels, log_config: Dict[str, Any], @@ -207,12 +204,8 @@ def _spawn_asgi_worker( loop = loops.get(loop_impl) sfd = socket.fileno() callback = callback_loader() - shutdown_event = set_loop_signals(loop) - wcallback = _asgi_call_wrap(callback, scope_opts, {}, log_access_fmt) - if not loop_opt: - wcallback = future_watcher_wrapper(wcallback) worker = ASGIWorker( worker_id, @@ -224,11 +217,11 @@ def _spawn_asgi_worker( http1_settings, http2_settings, websockets, - loop_opt, ssl_ctx, ) serve = getattr(worker, {ThreadModes.runtime: 'serve_rth', ThreadModes.workers: 'serve_wth'}[threading_mode]) - serve(wcallback, loop, contextvars.copy_context(), shutdown_event) + scheduler = _CBScheduler(loop, contextvars.copy_context(), _future_watcher_wrapper(wcallback)) + serve(scheduler, loop, shutdown_event) @staticmethod def _spawn_asgi_lifespan_worker( @@ -245,7 +238,6 @@ def _spawn_asgi_lifespan_worker( http1_settings: Optional[HTTP1Settings], http2_settings: Optional[HTTP2Settings], websockets: bool, - loop_opt: bool, log_enabled: bool, log_level: LogLevels, log_config: Dict[str, Any], @@ -271,10 +263,7 @@ def _spawn_asgi_lifespan_worker( sys.exit(1) shutdown_event = set_loop_signals(loop) - wcallback = _asgi_call_wrap(callback, scope_opts, lifespan_handler.state, log_access_fmt) - if not loop_opt: - wcallback = future_watcher_wrapper(wcallback) worker = ASGIWorker( worker_id, @@ -286,11 +275,11 @@ def _spawn_asgi_lifespan_worker( http1_settings, http2_settings, websockets, - loop_opt, ssl_ctx, ) serve = getattr(worker, {ThreadModes.runtime: 'serve_rth', ThreadModes.workers: 'serve_wth'}[threading_mode]) - serve(wcallback, loop, contextvars.copy_context(), shutdown_event) + scheduler = _CBScheduler(loop, contextvars.copy_context(), _future_watcher_wrapper(wcallback)) + serve(scheduler, loop, shutdown_event) loop.run_until_complete(lifespan_handler.shutdown()) @staticmethod @@ -308,7 +297,6 @@ def _spawn_rsgi_worker( http1_settings: Optional[HTTP1Settings], http2_settings: Optional[HTTP2Settings], websockets: bool, - loop_opt: bool, log_enabled: bool, log_level: LogLevels, log_config: Dict[str, Any], @@ -334,7 +322,6 @@ def _spawn_rsgi_worker( getattr(target, '__rsgi_del__') if hasattr(target, '__rsgi_del__') else lambda args, *kwargs: None ) callback = _rsgi_call_wrap(callback, log_access_fmt) - shutdown_event = set_loop_signals(loop) callback_init(loop) @@ -348,16 +335,11 @@ def _spawn_rsgi_worker( http1_settings, http2_settings, websockets, - loop_opt, ssl_ctx, ) serve = getattr(worker, {ThreadModes.runtime: 'serve_rth', ThreadModes.workers: 'serve_wth'}[threading_mode]) - serve( - future_watcher_wrapper(callback) if not loop_opt else callback, - loop, - contextvars.copy_context(), - shutdown_event, - ) + scheduler = _CBScheduler(loop, contextvars.copy_context(), _future_watcher_wrapper(callback)) + serve(scheduler, loop, shutdown_event) callback_del(loop) @staticmethod @@ -375,7 +357,6 @@ def _spawn_wsgi_worker( http1_settings: Optional[HTTP1Settings], http2_settings: Optional[HTTP2Settings], websockets: bool, - loop_opt: bool, log_enabled: bool, log_level: LogLevels, log_config: Dict[str, Any], @@ -393,14 +374,16 @@ def _spawn_wsgi_worker( loop = loops.get(loop_impl) sfd = socket.fileno() callback = callback_loader() - shutdown_event = set_sync_signals() worker = WSGIWorker( worker_id, sfd, threads, blocking_threads, backpressure, http_mode, http1_settings, http2_settings, ssl_ctx ) serve = getattr(worker, {ThreadModes.runtime: 'serve_rth', ThreadModes.workers: 'serve_wth'}[threading_mode]) - serve(_wsgi_call_wrap(callback, scope_opts, log_access_fmt), loop, contextvars.copy_context(), shutdown_event) + scheduler = _CBScheduler( + loop, contextvars.copy_context(), _wsgi_call_wrap(callback, scope_opts, log_access_fmt) + ) + serve(scheduler, loop, shutdown_event) shutdown_event.qs.wait() def _init_shared_socket(self): @@ -434,7 +417,6 @@ def _spawn_proc(self, idx, target, callback_loader, socket_loader) -> Worker: self.http1_settings, self.http2_settings, self.websockets, - self.loop_opt, self.log_enabled, self.log_level, self.log_config, diff --git a/pyproject.toml b/pyproject.toml index c80fb897..fb080770 100644 --- a/pyproject.toml +++ b/pyproject.toml @@ -52,6 +52,7 @@ test = [ 'httpx~=0.25.0', 'pytest~=7.4.2', 'pytest-asyncio~=0.21.1', + 'sniffio~=1.3', 'websockets~=11.0', ] all = ['granian[pname,reload]'] @@ -95,6 +96,7 @@ extend-ignore = [ 'E501', # leave line length to black 'N818', # leave to us exceptions naming 'S101', # assert is fine + 'S110', # except pass is fine ] flake8-quotes = { inline-quotes = 'single', multiline-quotes = 'double' } mccabe = { max-complexity = 13 } diff --git a/src/asgi/callbacks.rs b/src/asgi/callbacks.rs index 300a60f3..d2cc696e 100644 --- a/src/asgi/callbacks.rs +++ b/src/asgi/callbacks.rs @@ -1,9 +1,6 @@ use pyo3::prelude::; use pyo3::types::PyDict; -use std::{ - net::SocketAddr, - sync::{Arc, Mutex}, -}; +use std::{net::SocketAddr, sync::Arc}; use tokio::sync::oneshot; use super::{ @@ -11,50 +8,13 @@ use super::{ utils::{build_scope_http, build_scope_ws, scope_native_parts}, }; use crate::{ - asyncio::PyContext, - callbacks::{ - callback_impl_loop_err, callback_impl_loop_pytask, callback_impl_loop_run, callback_impl_loop_step, - callback_impl_loop_wake, callback_impl_run, callback_impl_run_pytask, CallbackWrapper, - }, + callbacks::ArcCBScheduler, http::{response_500, HTTPResponse}, runtime::RuntimeRef, utils::log_application_callable_exception, ws::{HyperWebsocket, UpgradeData}, }; -#[pyclass(frozen)] -pub(crate) struct CallbackRunnerHTTP { - proto: Py<HTTPProtocol>, - context: PyContext, - cb: PyObject, -} - -impl CallbackRunnerHTTP { - pub fn new(py: Python, cb: CallbackWrapper, proto: HTTPProtocol, scope: Bound<PyDict>) -> Self { - let pyproto = Py::new(py, proto).unwrap(); - Self { - proto: pyproto.clone_ref(py), - context: cb.context, - cb: cb.callback.call1(py, (scope, pyproto)).unwrap(), - } - } - - callback_impl_run!(); -} - -#[pymethods] -impl CallbackRunnerHTTP { - fn _loop_task<'p>(&self, py: Python<'p>) -> PyResult<Bound<'p, PyAny>> { - CallbackTaskHTTP::new( - py, - self.cb.clone_ref(py), - self.proto.clone_ref(py), - self.context.clone(), - )? - .run(py) - } -} - macro_rules! callback_impl_done_http { ($self:expr) => { if let Some(tx) = $self.proto.get().tx() { @@ -63,6 +23,14 @@ macro_rules! callback_impl_done_http { }; } +macro_rules! callback_impl_done_ws { + ($self:expr) => { + if let (Some(tx), res) = $self.proto.get().tx() { + let _ = tx.send(res); + } + }; +} + macro_rules! callback_impl_done_err { ($self:expr, $err:expr) => { $self.done(); @@ -71,211 +39,58 @@ macro_rules! callback_impl_done_err { } #[pyclass(frozen)] -pub(crate) struct CallbackTaskHTTP { - proto: Py<HTTPProtocol>, - context: PyContext, - pycontext: PyObject, - cb: PyObject, -} - -impl CallbackTaskHTTP { - pub fn new(py: Python, cb: PyObject, proto: Py<HTTPProtocol>, context: PyContext) -> PyResult<Self> { - let pyctx = context.context(py); - Ok(Self { - proto, - context, - pycontext: pyctx.call_method0(pyo3::intern!(py, "copy"))?.into(), - cb, - }) - } - - fn done(&self) { - callback_impl_done_http!(self); - } - - fn err(&self, err: &PyErr) { - callback_impl_done_err!(self, err); - } - - callback_impl_loop_run!(); - callback_impl_loop_err!(); -} - -#[pymethods] -impl CallbackTaskHTTP { - fn _loop_step(pyself: PyRef<'_, Self>, py: Python) -> PyResult<()> { - callback_impl_loop_step!(pyself, py) - } - - fn _loop_wake(pyself: PyRef<'_, Self>, py: Python, fut: PyObject) -> PyResult<PyObject> { - callback_impl_loop_wake!(pyself, py, fut) - } -} - -#[pyclass(frozen)] -pub(crate) struct CallbackWrappedRunnerHTTP { +pub(crate) struct CallbackWatcherHTTP { #[pyo3(get)] proto: Py<HTTPProtocol>, - context: PyContext, - cb: PyObject, #[pyo3(get)] scope: PyObject, - pytaskref: Arc<Mutex<Option<PyObject>>>, } -impl CallbackWrappedRunnerHTTP { - pub fn new(py: Python, cb: CallbackWrapper, proto: HTTPProtocol, scope: Bound<PyDict>) -> Self { +impl CallbackWatcherHTTP { + pub fn new(py: Python, proto: HTTPProtocol, scope: Bound<PyDict>) -> Self { Self { proto: Py::new(py, proto).unwrap(), - context: cb.context, - cb: cb.callback.clone_ref(py), scope: scope.into_py(py), - pytaskref: Arc::new(Mutex::new(None)), } } - - callback_impl_run_pytask!(); } #[pymethods] -impl CallbackWrappedRunnerHTTP { - fn _loop_task<'p>(pyself: PyRef<'_, Self>, py: Python<'p>) -> PyResult<Bound<'p, PyAny>> { - callback_impl_loop_pytask!(pyself, py) - } - +impl CallbackWatcherHTTP { fn done(&self) { callback_impl_done_http!(self); - self.pytaskref.lock().unwrap().take(); } fn err(&self, err: Bound<PyAny>) { callback_impl_done_err!(self, &PyErr::from_value_bound(err)); - self.pytaskref.lock().unwrap().take(); - } -} - -#[pyclass(frozen)] -pub(crate) struct CallbackRunnerWebsocket { - proto: Py<WebsocketProtocol>, - context: PyContext, - cb: PyObject, -} - -impl CallbackRunnerWebsocket { - pub fn new(py: Python, cb: CallbackWrapper, proto: WebsocketProtocol, scope: Bound<PyDict>) -> Self { - let pyproto = Py::new(py, proto).unwrap(); - Self { - proto: pyproto.clone_ref(py), - context: cb.context, - cb: cb.callback.call1(py, (scope, pyproto)).unwrap(), - } } - - callback_impl_run!(); -} - -#[pymethods] -impl CallbackRunnerWebsocket { - fn _loop_task<'p>(&self, py: Python<'p>) -> PyResult<Bound<'p, PyAny>> { - CallbackTaskWebsocket::new( - py, - self.cb.clone_ref(py), - self.proto.clone_ref(py), - self.context.clone(), - )? - .run(py) - } -} - -macro_rules! callback_impl_done_ws { - ($self:expr) => { - if let (Some(tx), res) = $self.proto.get().tx() { - let _ = tx.send(res); - } - }; } #[pyclass(frozen)] -pub(crate) struct CallbackTaskWebsocket { - proto: Py<WebsocketProtocol>, - context: PyContext, - pycontext: PyObject, - cb: PyObject, -} - -impl CallbackTaskWebsocket { - pub fn new(py: Python, cb: PyObject, proto: Py<WebsocketProtocol>, context: PyContext) -> PyResult<Self> { - let pyctx = context.context(py); - Ok(Self { - proto, - context, - pycontext: pyctx.call_method0(pyo3::intern!(py, "copy"))?.into(), - cb, - }) - } - - fn done(&self) { - callback_impl_done_ws!(self); - } - - fn err(&self, err: &PyErr) { - callback_impl_done_err!(self, err); - } - - callback_impl_loop_run!(); - callback_impl_loop_err!(); -} - -#[pymethods] -impl CallbackTaskWebsocket { - fn _loop_step(pyself: PyRef<'_, Self>, py: Python) -> PyResult<()> { - callback_impl_loop_step!(pyself, py) - } - - fn _loop_wake(pyself: PyRef<'_, Self>, py: Python, fut: PyObject) -> PyResult<PyObject> { - callback_impl_loop_wake!(pyself, py, fut) - } -} - -#[pyclass(frozen)] -pub(crate) struct CallbackWrappedRunnerWebsocket { +pub(crate) struct CallbackWatcherWebsocket { #[pyo3(get)] proto: Py<WebsocketProtocol>, - context: PyContext, - cb: PyObject, #[pyo3(get)] scope: PyObject, - pytaskref: Arc<Mutex<Option<PyObject>>>, } -impl CallbackWrappedRunnerWebsocket { - pub fn new(py: Python, cb: CallbackWrapper, proto: WebsocketProtocol, scope: Bound<PyDict>) -> Self { +impl CallbackWatcherWebsocket { + pub fn new(py: Python, proto: WebsocketProtocol, scope: Bound<PyDict>) -> Self { Self { proto: Py::new(py, proto).unwrap(), - context: cb.context, - cb: cb.callback.clone_ref(py), scope: scope.into_py(py), - pytaskref: Arc::new(Mutex::new(None)), } } - - callback_impl_run_pytask!(); } #[pymethods] -impl CallbackWrappedRunnerWebsocket { - fn _loop_task<'p>(pyself: PyRef<'_, Self>, py: Python<'p>) -> PyResult<Bound<'p, PyAny>> { - callback_impl_loop_pytask!(pyself, py) - } - +impl CallbackWatcherWebsocket { fn done(&self) { callback_impl_done_ws!(self); - self.pytaskref.lock().unwrap().take(); } fn err(&self, err: Bound<PyAny>) { callback_impl_done_err!(self, &PyErr::from_value_bound(err)); - self.pytaskref.lock().unwrap().take(); } } @@ -302,88 +117,75 @@ impl CallbackWrappedRunnerWebsocket { // } // } -macro_rules! call_impl_http { - ($func_name:ident, $runner:ident) => { - #[inline] - pub(crate) fn $func_name( - cb: CallbackWrapper, - rt: RuntimeRef, - server_addr: SocketAddr, - client_addr: SocketAddr, - scheme: &str, - req: hyper::http::request::Parts, - body: hyper::body::Incoming, - ) -> oneshot::Receiver<HTTPResponse> { - let brt = rt.innerb.clone(); - let (tx, rx) = oneshot::channel(); - let protocol = HTTPProtocol::new(rt, body, tx); - let scheme: Arc<str> = scheme.into(); - - let _ = brt.run(move \|\| { - scope_native_parts!( - req, - server_addr, - client_addr, - path, - query_string, - version, - server, - client - ); - Python::with_gil(\|py\| { - let scope = - build_scope_http(py, &req, version, server, client, &scheme, &path, query_string).unwrap(); - let _ = $runner::new(py, cb, protocol, scope).run(py); - }); - }); - - rx - } - }; -} - -macro_rules! call_impl_ws { - ($func_name:ident, $runner:ident) => { - #[inline] - pub(crate) fn $func_name( - cb: CallbackWrapper, - rt: RuntimeRef, - server_addr: SocketAddr, - client_addr: SocketAddr, - scheme: &str, - ws: HyperWebsocket, - req: hyper::http::request::Parts, - upgrade: UpgradeData, - ) -> oneshot::Receiver<WebsocketDetachedTransport> { - let brt = rt.innerb.clone(); - let (tx, rx) = oneshot::channel(); - let protocol = WebsocketProtocol::new(rt, tx, ws, upgrade); - let scheme: Arc<str> = scheme.into(); - - let _ = brt.run(move \|\| { - scope_native_parts!( - req, - server_addr, - client_addr, - path, - query_string, - version, - server, - client - ); - Python::with_gil(\|py\| { - let scope = - build_scope_ws(py, &req, version, server, client, &scheme, &path, query_string).unwrap(); - let _ = $runner::new(py, cb, protocol, scope).run(py); - }); - }); - - rx - } - }; +#[inline] +pub(crate) fn call_http( + cb: ArcCBScheduler, + rt: RuntimeRef, + server_addr: SocketAddr, + client_addr: SocketAddr, + scheme: &str, + req: hyper::http::request::Parts, + body: hyper::body::Incoming, +) -> oneshot::Receiver<HTTPResponse> { + let brt = rt.innerb.clone(); + let (tx, rx) = oneshot::channel(); + let protocol = HTTPProtocol::new(rt, body, tx); + let scheme: Arc<str> = scheme.into(); + + let _ = brt.run(move \|\| { + scope_native_parts!( + req, + server_addr, + client_addr, + path, + query_string, + version, + server, + client + ); + Python::with_gil(\|py\| { + let scope = build_scope_http(py, &req, version, server, client, &scheme, &path, query_string).unwrap(); + let watcher = Py::new(py, CallbackWatcherHTTP::new(py, protocol, scope)).unwrap(); + cb.get().schedule(py, watcher.as_any()); + }); + }); + + rx +} + +#[inline] +pub(crate) fn call_ws( + cb: ArcCBScheduler, + rt: RuntimeRef, + server_addr: SocketAddr, + client_addr: SocketAddr, + scheme: &str, + ws: HyperWebsocket, + req: hyper::http::request::Parts, + upgrade: UpgradeData, +) -> oneshot::Receiver<WebsocketDetachedTransport> { + let brt = rt.innerb.clone(); + let (tx, rx) = oneshot::channel(); + let protocol = WebsocketProtocol::new(rt, tx, ws, upgrade); + let scheme: Arc<str> = scheme.into(); + + let _ = brt.run(move \|\| { + scope_native_parts!( + req, + server_addr, + client_addr, + path, + query_string, + version, + server, + client + ); + Python::with_gil(\|py\| { + let scope = build_scope_ws(py, &req, version, server, client, &scheme, &path, query_string).unwrap(); + let watcher = Py::new(py, CallbackWatcherWebsocket::new(py, protocol, scope)).unwrap(); + cb.get().schedule(py, watcher.as_any()); + }); + }); + + rx } - -call_impl_http!(call_http, CallbackRunnerHTTP); -call_impl_http!(call_http_pyw, CallbackWrappedRunnerHTTP); -call_impl_ws!(call_ws, CallbackRunnerWebsocket); -call_impl_ws!(call_ws_pyw, CallbackWrappedRunnerWebsocket); diff --git a/src/asgi/http.rs b/src/asgi/http.rs index 2aed8cbe..40c168c4 100644 --- a/src/asgi/http.rs +++ b/src/asgi/http.rs @@ -3,9 +3,9 @@ use hyper::{header::SERVER as HK_SERVER, http::response::Builder as ResponseBuil use std::net::SocketAddr; use tokio::sync::mpsc; -use super::callbacks::{call_http, call_http_pyw, call_ws, call_ws_pyw}; +use super::callbacks::{call_http, call_ws}; use crate::{ - callbacks::CallbackWrapper, + callbacks::ArcCBScheduler, http::{empty_body, response_500, HTTPRequest, HTTPResponse, HV_SERVER}, runtime::RuntimeRef, ws::{is_upgrade_request as is_ws_upgrade, upgrade_intent as ws_upgrade, UpgradeData}, @@ -32,7 +32,7 @@ macro_rules! handle_request { #[inline] pub(crate) async fn $func_name( rt: RuntimeRef, - callback: CallbackWrapper, + callback: ArcCBScheduler, server_addr: SocketAddr, client_addr: SocketAddr, req: HTTPRequest, @@ -58,7 +58,7 @@ macro_rules! handle_request_with_ws { #[inline] pub(crate) async fn $func_name( rt: RuntimeRef, - callback: CallbackWrapper, + callback: ArcCBScheduler, server_addr: SocketAddr, client_addr: SocketAddr, mut req: HTTPRequest, @@ -154,10 +154,4 @@ macro_rules! handle_request_with_ws { } handle_request!(handle, call_http); -// handle_request!(handle_rtb, call_rtb_http); -handle_request!(handle_pyw, call_http_pyw); -// handle_request!(handle_rtb_pyw, call_rtb_http_pyw); handle_request_with_ws!(handle_ws, call_http, call_ws); -// handle_request_with_ws!(handle_rtb_ws, call_rtb_http, call_rtb_ws); -handle_request_with_ws!(handle_ws_pyw, call_http_pyw, call_ws_pyw); -// handle_request_with_ws!(handle_rtb_ws_pyw, call_rtb_http_pyw, call_rtb_ws_pyw); diff --git a/src/asgi/serve.rs b/src/asgi/serve.rs index 4c8b9c14..dd7d1380 100644 --- a/src/asgi/serve.rs +++ b/src/asgi/serve.rs @@ -1,7 +1,8 @@ use pyo3::prelude::; -use super::http::{handle, handle_pyw, handle_ws, handle_ws_pyw}; +use super::http::{handle, handle_ws}; +use crate::callbacks::CallbackScheduler; use crate::conversion::{worker_http1_config_from_py, worker_http2_config_from_py}; use crate::workers::{serve_rth, serve_rth_ssl, serve_wth, serve_wth_ssl, WorkerConfig, WorkerSignal, WorkerSignals}; @@ -12,21 +13,13 @@ pub struct ASGIWorker { impl ASGIWorker { serve_rth!(_serve_rth, handle); - serve_rth!(_serve_rth_pyw, handle_pyw); serve_rth!(_serve_rth_ws, handle_ws); - serve_rth!(_serve_rth_ws_pyw, handle_ws_pyw); serve_wth!(_serve_wth, handle); - serve_wth!(_serve_wth_pyw, handle_pyw); serve_wth!(_serve_wth_ws, handle_ws); - serve_wth!(_serve_wth_ws_pyw, handle_ws_pyw); serve_rth_ssl!(_serve_rth_ssl, handle); - serve_rth_ssl!(_serve_rth_ssl_pyw, handle_pyw); serve_rth_ssl!(_serve_rth_ssl_ws, handle_ws); - serve_rth_ssl!(_serve_rth_ssl_ws_pyw, handle_ws_pyw); serve_wth_ssl!(_serve_wth_ssl, handle); - serve_wth_ssl!(_serve_wth_ssl_pyw, handle_pyw); serve_wth_ssl!(_serve_wth_ssl_ws, handle_ws); - serve_wth_ssl!(_serve_wth_ssl_ws_pyw, handle_ws_pyw); } #[pymethods] @@ -43,7 +36,6 @@ impl ASGIWorker { http1_opts=None, http2_opts=None, websockets_enabled=false, - opt_enabled=true, ssl_enabled=false, ssl_cert=None, ssl_key=None, @@ -61,7 +53,6 @@ impl ASGIWorker { http1_opts: Option<PyObject>, http2_opts: Option<PyObject>, websockets_enabled: bool, - opt_enabled: bool, ssl_enabled: bool, ssl_cert: Option<&str>, ssl_key: Option<&str>, @@ -78,7 +69,6 @@ impl ASGIWorker { worker_http1_config_from_py(py, http1_opts)?, worker_http2_config_from_py(py, http2_opts)?, websockets_enabled, - opt_enabled, ssl_enabled, ssl_cert, ssl_key, @@ -87,57 +77,21 @@ impl ASGIWorker { }) } - fn serve_rth( - &self, - callback: PyObject, - event_loop: &Bound<PyAny>, - context: Bound<PyAny>, - signal: Py<WorkerSignal>, - ) { - match ( - self.config.websockets_enabled, - self.config.ssl_enabled, - self.config.opt_enabled, - ) { - (false, false, true) => self._serve_rth(callback, event_loop, context, WorkerSignals::Tokio(signal)), - (false, false, false) => self._serve_rth_pyw(callback, event_loop, context, WorkerSignals::Tokio(signal)), - (true, false, true) => self._serve_rth_ws(callback, event_loop, context, WorkerSignals::Tokio(signal)), - (true, false, false) => self._serve_rth_ws_pyw(callback, event_loop, context, WorkerSignals::Tokio(signal)), - (false, true, true) => self._serve_rth_ssl(callback, event_loop, context, WorkerSignals::Tokio(signal)), - (false, true, false) => { - self._serve_rth_ssl_pyw(callback, event_loop, context, WorkerSignals::Tokio(signal)); - } - (true, true, true) => self._serve_rth_ssl_ws(callback, event_loop, context, WorkerSignals::Tokio(signal)), - (true, true, false) => { - self._serve_rth_ssl_ws_pyw(callback, event_loop, context, WorkerSignals::Tokio(signal)); - } + fn serve_rth(&self, callback: Py<CallbackScheduler>, event_loop: &Bound<PyAny>, signal: Py<WorkerSignal>) { + match (self.config.websockets_enabled, self.config.ssl_enabled) { + (false, false) => self._serve_rth(callback, event_loop, WorkerSignals::Tokio(signal)), + (true, false) => self._serve_rth_ws(callback, event_loop, WorkerSignals::Tokio(signal)), + (false, true) => self._serve_rth_ssl(callback, event_loop, WorkerSignals::Tokio(signal)), + (true, true) => self._serve_rth_ssl_ws(callback, event_loop, WorkerSignals::Tokio(signal)), } } - fn serve_wth( - &self, - callback: PyObject, - event_loop: &Bound<PyAny>, - context: Bound<PyAny>, - signal: Py<WorkerSignal>, - ) { - match ( - self.config.websockets_enabled, - self.config.ssl_enabled, - self.config.opt_enabled, - ) { - (false, false, true) => self._serve_wth(callback, event_loop, context, WorkerSignals::Tokio(signal)), - (false, false, false) => self._serve_wth_pyw(callback, event_loop, context, WorkerSignals::Tokio(signal)), - (true, false, true) => self._serve_wth_ws(callback, event_loop, context, WorkerSignals::Tokio(signal)), - (true, false, false) => self._serve_wth_ws_pyw(callback, event_loop, context, WorkerSignals::Tokio(signal)), - (false, true, true) => self._serve_wth_ssl(callback, event_loop, context, WorkerSignals::Tokio(signal)), - (false, true, false) => { - self._serve_wth_ssl_pyw(callback, event_loop, context, WorkerSignals::Tokio(signal)); - } - (true, true, true) => self._serve_wth_ssl_ws(callback, event_loop, context, WorkerSignals::Tokio(signal)), - (true, true, false) => { - self._serve_wth_ssl_ws_pyw(callback, event_loop, context, WorkerSignals::Tokio(signal)); - } + fn serve_wth(&self, callback: Py<CallbackScheduler>, event_loop: &Bound<PyAny>, signal: Py<WorkerSignal>) { + match (self.config.websockets_enabled, self.config.ssl_enabled) { + (false, false) => self._serve_wth(callback, event_loop, WorkerSignals::Tokio(signal)), + (true, false) => self._serve_wth_ws(callback, event_loop, WorkerSignals::Tokio(signal)), + (false, true) => self._serve_wth_ssl(callback, event_loop, WorkerSignals::Tokio(signal)), + (true, true) => self._serve_wth_ssl_ws(callback, event_loop, WorkerSignals::Tokio(signal)), } } } diff --git a/src/asyncio.rs b/src/asyncio.rs index 7767ed9d..f6b1b38a 100644 --- a/src/asyncio.rs +++ b/src/asyncio.rs @@ -1,46 +1,16 @@ use pyo3::{prelude::, sync::GILOnceCell}; -use std::{convert::Into, sync::Arc}; +use std::convert::Into; static CONTEXTVARS: GILOnceCell<PyObject> = GILOnceCell::new(); static CONTEXT: GILOnceCell<PyObject> = GILOnceCell::new(); -#[derive(Clone, Debug)] -pub struct PyContext { - event_loop: Arc<PyObject>, - context: Arc<PyObject>, -} - -impl PyContext { - pub fn new(event_loop: Bound<PyAny>) -> Self { - let pynone = event_loop.py().None(); - Self { - event_loop: Arc::new(event_loop.unbind()), - context: Arc::new(pynone), - } - } - - pub fn with_context(self, context: Bound<PyAny>) -> Self { - Self { - context: Arc::new(context.unbind()), - ..self - } - } - - pub fn event_loop<'p>(&self, py: Python<'p>) -> Bound<'p, PyAny> { - self.event_loop.clone_ref(py).into_bound(py) - } - - pub fn context<'p>(&self, py: Python<'p>) -> Bound<'p, PyAny> { - self.context.clone_ref(py).into_bound(py) - } -} - fn contextvars(py: Python) -> PyResult<&Bound<PyAny>> { Ok(CONTEXTVARS .get_or_try_init(py, \|\| py.import_bound("contextvars").map(Into::into))? .bind(py)) } +#[allow(dead_code)] pub(crate) fn empty_context(py: Python) -> PyResult<&Bound<PyAny>> { Ok(CONTEXT .get_or_try_init(py, \|\| { diff --git a/src/callbacks.rs b/src/callbacks.rs index 07a6bfb4..2944fbff 100644 --- a/src/callbacks.rs +++ b/src/callbacks.rs @@ -3,21 +3,48 @@ use pyo3::{exceptions::PyStopIteration, prelude::}; use std::sync::{atomic, Arc, RwLock}; use tokio::sync::Notify; -use super::asyncio::PyContext; +pub(crate) type ArcCBScheduler = Arc<Py<CallbackScheduler>>; + +#[pyclass(frozen, subclass)] +pub(crate) struct CallbackScheduler { + #[pyo3(get)] + _loop: PyObject, + #[pyo3(get)] + _ctx: PyObject, + schedule_fn: Arc<RwLock<PyObject>>, + pub cb: PyObject, +} -#[derive(Clone)] -pub(crate) struct CallbackWrapper { - pub callback: Arc<PyObject>, - pub context: PyContext, +impl CallbackScheduler { + #[inline] + pub(crate) fn schedule(&self, _py: Python, watcher: &PyObject) { + // // let cb = self.cb.as_ptr(); + let cbarg = watcher.as_ptr(); + let sched = self.schedule_fn.read().unwrap().as_ptr(); + unsafe { + // let coro = pyo3::ffi::PyObject_CallOneArg(cb, cbarg); + pyo3::ffi::PyObject_CallOneArg(sched, cbarg); + } + } } -impl CallbackWrapper { - pub(crate) fn new(callback: PyObject, event_loop: Bound<PyAny>, context: Bound<PyAny>) -> Self { +#[pymethods] +impl CallbackScheduler { + #[new] + fn new(py: Python, event_loop: PyObject, ctx: PyObject, cb: PyObject) -> Self { Self { - callback: Arc::new(callback), - context: PyContext::new(event_loop).with_context(context), + _loop: event_loop, + _ctx: ctx, + schedule_fn: Arc::new(RwLock::new(py.None())), + cb, } } + + #[setter(_schedule_fn)] + fn _set_schedule_fn(&self, val: PyObject) { + let mut guard = self.schedule_fn.write().unwrap(); + guard = val; + } } #[pyclass(frozen)] @@ -284,135 +311,3 @@ impl PyFutureResultSetter { let _ = target.call1((value,)); } } - -macro_rules! callback_impl_run { - () => { - pub fn run(self, py: Python<'_>) -> PyResult<Bound<PyAny>> { - let event_loop = self.context.event_loop(py); - let target = self.into_py(py).getattr(py, pyo3::intern!(py, "_loop_task"))?; - let kwctx = pyo3::types::PyDict::new_bound(py); - kwctx.set_item(pyo3::intern!(py, "context"), crate::asyncio::empty_context(py)?)?; - event_loop.call_method(pyo3::intern!(py, "call_soon_threadsafe"), (target,), Some(&kwctx)) - } - }; -} - -macro_rules! callback_impl_run_pytask { - () => { - pub fn run(self, py: Python<'_>) -> PyResult<Bound<PyAny>> { - let taskref = self.pytaskref.clone(); - let event_loop = self.context.event_loop(py); - let context = self.context.context(py); - let target = self.into_py(py).getattr(py, pyo3::intern!(py, "_loop_task"))?; - let kwctx = pyo3::types::PyDict::new_bound(py); - { - let mut taskref_guard = taskref.lock().unwrap(); - taskref_guard = Some(target.clone_ref(py)); - } - kwctx.set_item(pyo3::intern!(py, "context"), context)?; - event_loop.call_method(pyo3::intern!(py, "call_soon_threadsafe"), (target,), Some(&kwctx)) - } - }; -} - -macro_rules! callback_impl_loop_run { - () => { - pub fn run(self, py: Python<'_>) -> PyResult<Bound<PyAny>> { - let context = self.pycontext.clone_ref(py).into_bound(py); - context.call_method1( - pyo3::intern!(py, "run"), - (self.into_py(py).getattr(py, pyo3::intern!(py, "_loop_step"))?,), - ) - } - }; -} - -macro_rules! callback_impl_loop_pytask { - ($pyself:expr, $py:expr) => { - $pyself.context.event_loop($py).call_method1( - pyo3::intern!($py, "create_task"), - ($pyself - .cb - .clone_ref($py) - .into_bound($py) - .call1(($pyself.into_py($py),))?,), - ) - }; -} - -macro_rules! callback_impl_loop_step { - ($pyself:expr, $py:expr) => { - match $pyself.cb.call_method1($py, pyo3::intern!($py, "send"), ($py.None(),)) { - Ok(res) => { - let blocking: bool = match res.getattr($py, pyo3::intern!($py, "_asyncio_future_blocking")) { - Ok(v) => v.extract($py)?, - _ => false, - }; - - let ctx = $pyself.pycontext.clone_ref($py); - let kwctx = pyo3::types::PyDict::new_bound($py); - kwctx.set_item(pyo3::intern!($py, "context"), ctx)?; - - match blocking { - true => { - res.setattr($py, pyo3::intern!($py, "_asyncio_future_blocking"), false)?; - res.call_method_bound( - $py, - pyo3::intern!($py, "add_done_callback"), - ($pyself.into_py($py).getattr($py, pyo3::intern!($py, "_loop_wake"))?,), - Some(&kwctx), - )?; - } - false => { - let event_loop = $pyself.context.event_loop($py); - event_loop.call_method( - pyo3::intern!($py, "call_soon"), - ($pyself.into_py($py).getattr($py, pyo3::intern!($py, "_loop_step"))?,), - Some(&kwctx), - )?; - } - } - Ok(()) - } - Err(err) - if (err.is_instance_of::<pyo3::exceptions::PyStopIteration>($py) - \|\| err.is_instance_of::<pyo3::exceptions::PyStopAsyncIteration>($py) - \|\| err.is_instance_of::<pyo3::exceptions::asyncio::CancelledError>($py)) => - { - $pyself.done(); - Ok(()) - } - Err(err) => { - $pyself.err(&err); - Ok(()) - } - } - }; -} - -macro_rules! callback_impl_loop_wake { - ($pyself:expr, $py:expr, $fut:expr) => { - match $fut.call_method0($py, pyo3::intern!($py, "result")) { - Ok(_) => $pyself.into_py($py).call_method0($py, pyo3::intern!($py, "_loop_step")), - Err(err) => $pyself._loop_err($py, err), - } - }; -} - -macro_rules! callback_impl_loop_err { - () => { - pub fn _loop_err(&self, py: Python, err: PyErr) -> PyResult<PyObject> { - self.err(&err); - let cberr = self.cb.call_method1(py, pyo3::intern!(py, "throw"), (err,)); - cberr - } - }; -} - -pub(crate) use callback_impl_loop_err; -pub(crate) use callback_impl_loop_pytask; -pub(crate) use callback_impl_loop_run; -pub(crate) use callback_impl_loop_step; -pub(crate) use callback_impl_loop_wake; -pub(crate) use callback_impl_run; -pub(crate) use callback_impl_run_pytask; diff --git a/src/lib.rs b/src/lib.rs index d639101b..b26a7349 100644 --- a/src/lib.rs +++ b/src/lib.rs @@ -37,6 +37,7 @@ pub fn get_granian_version() -> &'static str { #[pymodule] fn _granian(py: Python, module: &Bound<PyModule>) -> PyResult<()> { module.add("__version__", get_granian_version())?; + module.add_class::<callbacks::CallbackScheduler>()?; asgi::init_pymodule(module)?; rsgi::init_pymodule(py, module)?; tcp::init_pymodule(module)?; diff --git a/src/rsgi/callbacks.rs b/src/rsgi/callbacks.rs index bdc61f2e..b67081b7 100644 --- a/src/rsgi/callbacks.rs +++ b/src/rsgi/callbacks.rs @@ -1,5 +1,4 @@ use pyo3::prelude::; -use std::sync::{Arc, Mutex}; use tokio::sync::oneshot; use super::{ @@ -7,49 +6,12 @@ use super::{ types::{PyResponse, PyResponseBody, RSGIHTTPScope as HTTPScope, RSGIWebsocketScope as WebsocketScope}, }; use crate::{ - asyncio::PyContext, - callbacks::{ - callback_impl_loop_err, callback_impl_loop_pytask, callback_impl_loop_run, callback_impl_loop_step, - callback_impl_loop_wake, callback_impl_run, callback_impl_run_pytask, CallbackWrapper, - }, + callbacks::ArcCBScheduler, runtime::RuntimeRef, utils::log_application_callable_exception, ws::{HyperWebsocket, UpgradeData}, }; -#[pyclass(frozen)] -pub(crate) struct CallbackRunnerHTTP { - proto: Py<HTTPProtocol>, - context: PyContext, - cb: PyObject, -} - -impl CallbackRunnerHTTP { - pub fn new(py: Python, cb: CallbackWrapper, proto: HTTPProtocol, scope: HTTPScope) -> Self { - let pyproto = Py::new(py, proto).unwrap(); - Self { - proto: pyproto.clone_ref(py), - context: cb.context, - cb: cb.callback.call1(py, (scope, pyproto)).unwrap(), - } - } - - callback_impl_run!(); -} - -#[pymethods] -impl CallbackRunnerHTTP { - fn _loop_task<'p>(&self, py: Python<'p>) -> PyResult<Bound<'p, PyAny>> { - CallbackTaskHTTP::new( - py, - self.cb.clone_ref(py), - self.proto.clone_ref(py), - self.context.clone(), - )? - .run(py) - } -} - macro_rules! callback_impl_done_http { ($self:expr) => { if let Some(tx) = $self.proto.get().tx() { @@ -58,6 +20,12 @@ macro_rules! callback_impl_done_http { }; } +macro_rules! callback_impl_done_ws { + ($self:expr) => { + let _ = $self.proto.get().close(None); + }; +} + macro_rules! callback_impl_done_err { ($self:expr, $err:expr) => { $self.done(); @@ -66,262 +34,100 @@ macro_rules! callback_impl_done_err { } #[pyclass(frozen)] -pub(crate) struct CallbackTaskHTTP { - proto: Py<HTTPProtocol>, - context: PyContext, - pycontext: PyObject, - cb: PyObject, -} - -impl CallbackTaskHTTP { - pub fn new(py: Python, cb: PyObject, proto: Py<HTTPProtocol>, context: PyContext) -> PyResult<Self> { - let pyctx = context.context(py); - Ok(Self { - proto, - context, - pycontext: pyctx.call_method0(pyo3::intern!(py, "copy"))?.into(), - cb, - }) - } - - fn done(&self) { - callback_impl_done_http!(self); - } - - fn err(&self, err: &PyErr) { - callback_impl_done_err!(self, err); - } - - callback_impl_loop_run!(); - callback_impl_loop_err!(); -} - -#[pymethods] -impl CallbackTaskHTTP { - fn _loop_step(pyself: PyRef<'_, Self>, py: Python) -> PyResult<()> { - callback_impl_loop_step!(pyself, py) - } - - fn _loop_wake(pyself: PyRef<'_, Self>, py: Python, fut: PyObject) -> PyResult<PyObject> { - callback_impl_loop_wake!(pyself, py, fut) - } -} - -#[pyclass(frozen)] -pub(crate) struct CallbackWrappedRunnerHTTP { +pub(crate) struct CallbackWatcherHTTP { #[pyo3(get)] proto: Py<HTTPProtocol>, - context: PyContext, - cb: PyObject, #[pyo3(get)] scope: PyObject, - pytaskref: Arc<Mutex<Option<PyObject>>>, } -impl CallbackWrappedRunnerHTTP { - pub fn new(py: Python, cb: CallbackWrapper, proto: HTTPProtocol, scope: HTTPScope) -> Self { +impl CallbackWatcherHTTP { + pub fn new(py: Python, proto: HTTPProtocol, scope: HTTPScope) -> Self { Self { proto: Py::new(py, proto).unwrap(), - context: cb.context, - cb: cb.callback.clone_ref(py), scope: scope.into_py(py), - pytaskref: Arc::new(Mutex::new(None)), } } - - callback_impl_run_pytask!(); } #[pymethods] -impl CallbackWrappedRunnerHTTP { - fn _loop_task<'p>(pyself: PyRef<'_, Self>, py: Python<'p>) -> PyResult<Bound<'p, PyAny>> { - callback_impl_loop_pytask!(pyself, py) - } - +impl CallbackWatcherHTTP { fn done(&self) { callback_impl_done_http!(self); - self.pytaskref.lock().unwrap().take(); } fn err(&self, err: Bound<PyAny>) { callback_impl_done_err!(self, &PyErr::from_value_bound(err)); - self.pytaskref.lock().unwrap().take(); } } #[pyclass(frozen)] -pub(crate) struct CallbackRunnerWebsocket { - proto: Py<WebsocketProtocol>, - context: PyContext, - cb: PyObject, -} - -impl CallbackRunnerWebsocket { - pub fn new(py: Python, cb: CallbackWrapper, proto: WebsocketProtocol, scope: WebsocketScope) -> Self { - let pyproto = Py::new(py, proto).unwrap(); - Self { - proto: pyproto.clone_ref(py), - context: cb.context, - cb: cb.callback.call1(py, (scope, pyproto)).unwrap(), - } - } - - callback_impl_run!(); -} - -#[pymethods] -impl CallbackRunnerWebsocket { - fn _loop_task<'p>(&self, py: Python<'p>) -> PyResult<Bound<'p, PyAny>> { - CallbackTaskWebsocket::new( - py, - self.cb.clone_ref(py), - self.proto.clone_ref(py), - self.context.clone(), - )? - .run(py) - } -} - -macro_rules! callback_impl_done_ws { - ($self:expr) => { - let _ = $self.proto.get().close(None); - }; -} - -#[pyclass(frozen)] -pub(crate) struct CallbackTaskWebsocket { - proto: Py<WebsocketProtocol>, - context: PyContext, - pycontext: PyObject, - cb: PyObject, -} - -impl CallbackTaskWebsocket { - pub fn new(py: Python, cb: PyObject, proto: Py<WebsocketProtocol>, context: PyContext) -> PyResult<Self> { - let pyctx = context.context(py); - Ok(Self { - proto, - context, - pycontext: pyctx.call_method0(pyo3::intern!(py, "copy"))?.into(), - cb, - }) - } - - fn done(&self) { - callback_impl_done_ws!(self); - } - - fn err(&self, err: &PyErr) { - callback_impl_done_err!(self, err); - } - - callback_impl_loop_run!(); - callback_impl_loop_err!(); -} - -#[pymethods] -impl CallbackTaskWebsocket { - fn _loop_step(pyself: PyRef<'_, Self>, py: Python) -> PyResult<()> { - callback_impl_loop_step!(pyself, py) - } - - fn _loop_wake(pyself: PyRef<'_, Self>, py: Python, fut: PyObject) -> PyResult<PyObject> { - callback_impl_loop_wake!(pyself, py, fut) - } -} - -#[pyclass(frozen)] -pub(crate) struct CallbackWrappedRunnerWebsocket { +pub(crate) struct CallbackWatcherWebsocket { #[pyo3(get)] proto: Py<WebsocketProtocol>, - context: PyContext, - cb: PyObject, #[pyo3(get)] scope: PyObject, - pytaskref: Arc<Mutex<Option<PyObject>>>, } -impl CallbackWrappedRunnerWebsocket { - pub fn new(py: Python, cb: CallbackWrapper, proto: WebsocketProtocol, scope: WebsocketScope) -> Self { +impl CallbackWatcherWebsocket { + pub fn new(py: Python, proto: WebsocketProtocol, scope: WebsocketScope) -> Self { Self { proto: Py::new(py, proto).unwrap(), - context: cb.context, - cb: cb.callback.clone_ref(py), scope: scope.into_py(py), - pytaskref: Arc::new(Mutex::new(None)), } } - - callback_impl_run_pytask!(); } #[pymethods] -impl CallbackWrappedRunnerWebsocket { - fn _loop_task<'p>(pyself: PyRef<'_, Self>, py: Python<'p>) -> PyResult<Bound<'p, PyAny>> { - callback_impl_loop_pytask!(pyself, py) - } - +impl CallbackWatcherWebsocket { fn done(&self) { callback_impl_done_ws!(self); - self.pytaskref.lock().unwrap().take(); } fn err(&self, err: Bound<PyAny>) { callback_impl_done_err!(self, &PyErr::from_value_bound(err)); - self.pytaskref.lock().unwrap().take(); } } -macro_rules! call_impl_http { - ($func_name:ident, $runner:ident) => { - #[inline] - pub(crate) fn $func_name( - cb: CallbackWrapper, - rt: RuntimeRef, - body: hyper::body::Incoming, - scope: HTTPScope, - ) -> oneshot::Receiver<PyResponse> { - let brt = rt.innerb.clone(); - let (tx, rx) = oneshot::channel(); - let protocol = HTTPProtocol::new(rt, tx, body); - - let _ = brt.run(\|\| { - Python::with_gil(\|py\| { - let _ = $runner::new(py, cb, protocol, scope).run(py); - }); - }); - - rx - } - }; -} - -macro_rules! call_impl_ws { - ($func_name:ident, $runner:ident) => { - #[inline] - pub(crate) fn $func_name( - cb: CallbackWrapper, - rt: RuntimeRef, - ws: HyperWebsocket, - upgrade: UpgradeData, - scope: WebsocketScope, - ) -> oneshot::Receiver<WebsocketDetachedTransport> { - let brt = rt.innerb.clone(); - let (tx, rx) = oneshot::channel(); - let protocol = WebsocketProtocol::new(rt, tx, ws, upgrade); - - let _ = brt.run(\|\| { - Python::with_gil(\|py\| { - let _ = $runner::new(py, cb, protocol, scope).run(py); - }); - }); - - rx - } - }; +#[inline] +pub(crate) fn call_http( + cb: ArcCBScheduler, + rt: RuntimeRef, + body: hyper::body::Incoming, + scope: HTTPScope, +) -> oneshot::Receiver<PyResponse> { + let brt = rt.innerb.clone(); + let (tx, rx) = oneshot::channel(); + let protocol = HTTPProtocol::new(rt, tx, body); + + let _ = brt.run(move \|\| { + Python::with_gil(\|py\| { + let watcher = Py::new(py, CallbackWatcherHTTP::new(py, protocol, scope)).unwrap(); + cb.get().schedule(py, watcher.as_any()); + }); + }); + + rx +} + +#[inline] +pub(crate) fn call_ws( + cb: ArcCBScheduler, + rt: RuntimeRef, + ws: HyperWebsocket, + upgrade: UpgradeData, + scope: WebsocketScope, +) -> oneshot::Receiver<WebsocketDetachedTransport> { + let brt = rt.innerb.clone(); + let (tx, rx) = oneshot::channel(); + let protocol = WebsocketProtocol::new(rt, tx, ws, upgrade); + + let _ = brt.run(move \|\| { + Python::with_gil(\|py\| { + let watcher = Py::new(py, CallbackWatcherWebsocket::new(py, protocol, scope)).unwrap(); + cb.get().schedule(py, watcher.as_any()); + }); + }); + + rx } - -call_impl_http!(call_http, CallbackRunnerHTTP); -call_impl_http!(call_http_pyw, CallbackWrappedRunnerHTTP); -call_impl_ws!(call_ws, CallbackRunnerWebsocket); -call_impl_ws!(call_ws_pyw, CallbackWrappedRunnerWebsocket); diff --git a/src/rsgi/http.rs b/src/rsgi/http.rs index 50f52361..ee3278ba 100644 --- a/src/rsgi/http.rs +++ b/src/rsgi/http.rs @@ -4,11 +4,11 @@ use std::net::SocketAddr; use tokio::sync::mpsc; use super::{ - callbacks::{call_http, call_http_pyw, call_ws, call_ws_pyw}, + callbacks::{call_http, call_ws}, types::{PyResponse, RSGIHTTPScope as HTTPScope, RSGIWebsocketScope as WebsocketScope}, }; use crate::{ - callbacks::CallbackWrapper, + callbacks::ArcCBScheduler, http::{empty_body, response_500, HTTPRequest, HTTPResponse, HV_SERVER}, runtime::RuntimeRef, ws::{is_upgrade_request as is_ws_upgrade, upgrade_intent as ws_upgrade, UpgradeData}, @@ -46,7 +46,7 @@ macro_rules! handle_request { #[inline] pub(crate) async fn $func_name( rt: RuntimeRef, - callback: CallbackWrapper, + callback: ArcCBScheduler, server_addr: SocketAddr, client_addr: SocketAddr, req: HTTPRequest, @@ -59,12 +59,30 @@ macro_rules! handle_request { }; } +// macro_rules! handle_request2 { +// ($func_name:ident, $handler:expr) => { +// #[inline] +// pub(crate) async fn $func_name( +// rt: RuntimeRef, +// callback: ArcCBScheduler, +// server_addr: SocketAddr, +// client_addr: SocketAddr, +// req: HTTPRequest, +// scheme: &str, +// ) -> HTTPResponse { +// let (parts, body) = req.into_parts(); +// let scope = build_scope!(HTTPScope, server_addr, client_addr, parts, scheme); +// handle_http_response!($handler, rt, callback, body, scope) +// } +// }; +// } + macro_rules! handle_request_with_ws { ($func_name:ident, $handler_req:expr, $handler_ws:expr) => { #[inline] pub(crate) async fn $func_name( rt: RuntimeRef, - callback: CallbackWrapper, + callback: ArcCBScheduler, server_addr: SocketAddr, client_addr: SocketAddr, mut req: HTTPRequest, @@ -136,6 +154,7 @@ macro_rules! handle_request_with_ws { } handle_request!(handle, call_http); -handle_request!(handle_pyw, call_http_pyw); +// handle_request!(handle_pyw, call_http_pyw); +// handle_request2!(handle2, call_http2); handle_request_with_ws!(handle_ws, call_http, call_ws); -handle_request_with_ws!(handle_ws_pyw, call_http_pyw, call_ws_pyw); +// handle_request_with_ws!(handle_ws_pyw, call_http_pyw, call_ws_pyw); diff --git a/src/rsgi/io.rs b/src/rsgi/io.rs index e83c2246..cd813ea1 100644 --- a/src/rsgi/io.rs +++ b/src/rsgi/io.rs @@ -96,29 +96,37 @@ impl RSGIHTTPProtocol { error_proto!() } - fn __aiter__(pyself: PyRef<'_, Self>) -> PyRef<'_, Self> { - if let Some(body) = pyself.body.lock().unwrap().take() { - let mut stream = pyself.body_stream.blocking_lock(); + fn __aiter__(pyself: Py<Self>, py: Python) -> PyResult<Py<Self>> { + let inner = pyself.get(); + if let Some(body) = inner.body.lock().unwrap().take() { + let mut stream = inner.body_stream.blocking_lock(); stream = Some(http_body_util::BodyStream::new(body)); + return Ok(pyself.clone_ref(py)); } - pyself + error_proto!() } - fn __anext__<'p>(&self, py: Python<'p>) -> PyResult<Option<Bound<'p, PyAny>>> { + fn __anext__<'p>(&self, py: Python<'p>) -> PyResult<Bound<'p, PyAny>> { + if self.body_stream.blocking_lock().is_none() { + return Err(pyo3::exceptions::PyStopAsyncIteration::new_err("stream exhausted")); + } let body_stream = self.body_stream.clone(); - let pyfut = future_into_py_iter(self.rt.clone(), py, async move { - if let Some(stream) = &mut body_stream.lock().await { - if let Some(chunk) = stream.next().await { + future_into_py_iter(self.rt.clone(), py, async move { + let guard = &mut body_stream.lock().await; + let bytes = match guard.as_mut().unwrap().next().await { + Some(chunk) => { let chunk = chunk .map(\|buf\| buf.into_data().unwrap_or_default()) .unwrap_or(body::Bytes::new()); - return Ok(BytesToPy(chunk)); - }; - return Err(pyo3::exceptions::PyStopAsyncIteration::new_err("stream exhausted")); - } - error_proto!() - })?; - Ok(Some(pyfut)) + BytesToPy(chunk) + } + _ => { + let _ = guard.take(); + BytesToPy(body::Bytes::new()) + } + }; + Ok(bytes) + }) } #[pyo3(signature = (status=200, headers=vec![]))] diff --git a/src/rsgi/serve.rs b/src/rsgi/serve.rs index 1f13dca4..bc904b91 100644 --- a/src/rsgi/serve.rs +++ b/src/rsgi/serve.rs @@ -1,7 +1,8 @@ use pyo3::prelude::; -use super::http::{handle, handle_pyw, handle_ws, handle_ws_pyw}; +use super::http::{handle, handle_ws}; +use crate::callbacks::CallbackScheduler; use crate::conversion::{worker_http1_config_from_py, worker_http2_config_from_py}; use crate::workers::{serve_rth, serve_rth_ssl, serve_wth, serve_wth_ssl, WorkerConfig, WorkerSignal, WorkerSignals}; @@ -12,21 +13,13 @@ pub struct RSGIWorker { impl RSGIWorker { serve_rth!(_serve_rth, handle); - serve_rth!(_serve_rth_pyw, handle_pyw); serve_rth!(_serve_rth_ws, handle_ws); - serve_rth!(_serve_rth_ws_pyw, handle_ws_pyw); serve_wth!(_serve_wth, handle); - serve_wth!(_serve_wth_pyw, handle_pyw); serve_wth!(_serve_wth_ws, handle_ws); - serve_wth!(_serve_wth_ws_pyw, handle_ws_pyw); serve_rth_ssl!(_serve_rth_ssl, handle); - serve_rth_ssl!(_serve_rth_ssl_pyw, handle_pyw); serve_rth_ssl!(_serve_rth_ssl_ws, handle_ws); - serve_rth_ssl!(_serve_rth_ssl_ws_pyw, handle_ws_pyw); serve_wth_ssl!(_serve_wth_ssl, handle); - serve_wth_ssl!(_serve_wth_ssl_pyw, handle_pyw); serve_wth_ssl!(_serve_wth_ssl_ws, handle_ws); - serve_wth_ssl!(_serve_wth_ssl_ws_pyw, handle_ws_pyw); } #[pymethods] @@ -43,7 +36,6 @@ impl RSGIWorker { http1_opts=None, http2_opts=None, websockets_enabled=false, - opt_enabled=true, ssl_enabled=false, ssl_cert=None, ssl_key=None, @@ -61,7 +53,6 @@ impl RSGIWorker { http1_opts: Option<PyObject>, http2_opts: Option<PyObject>, websockets_enabled: bool, - opt_enabled: bool, ssl_enabled: bool, ssl_cert: Option<&str>, ssl_key: Option<&str>, @@ -78,7 +69,6 @@ impl RSGIWorker { worker_http1_config_from_py(py, http1_opts)?, worker_http2_config_from_py(py, http2_opts)?, websockets_enabled, - opt_enabled, ssl_enabled, ssl_cert, ssl_key, @@ -87,57 +77,21 @@ impl RSGIWorker { }) } - fn serve_rth( - &self, - callback: PyObject, - event_loop: &Bound<PyAny>, - context: Bound<PyAny>, - signal: Py<WorkerSignal>, - ) { - match ( - self.config.websockets_enabled, - self.config.ssl_enabled, - self.config.opt_enabled, - ) { - (false, false, true) => self._serve_rth(callback, event_loop, context, WorkerSignals::Tokio(signal)), - (false, false, false) => self._serve_rth_pyw(callback, event_loop, context, WorkerSignals::Tokio(signal)), - (true, false, true) => self._serve_rth_ws(callback, event_loop, context, WorkerSignals::Tokio(signal)), - (true, false, false) => self._serve_rth_ws_pyw(callback, event_loop, context, WorkerSignals::Tokio(signal)), - (false, true, true) => self._serve_rth_ssl(callback, event_loop, context, WorkerSignals::Tokio(signal)), - (false, true, false) => { - self._serve_rth_ssl_pyw(callback, event_loop, context, WorkerSignals::Tokio(signal)); - } - (true, true, true) => self._serve_rth_ssl_ws(callback, event_loop, context, WorkerSignals::Tokio(signal)), - (true, true, false) => { - self._serve_rth_ssl_ws_pyw(callback, event_loop, context, WorkerSignals::Tokio(signal)); - } + fn serve_rth(&self, callback: Py<CallbackScheduler>, event_loop: &Bound<PyAny>, signal: Py<WorkerSignal>) { + match (self.config.websockets_enabled, self.config.ssl_enabled) { + (false, false) => self._serve_rth(callback, event_loop, WorkerSignals::Tokio(signal)), + (true, false) => self._serve_rth_ws(callback, event_loop, WorkerSignals::Tokio(signal)), + (false, true) => self._serve_rth_ssl(callback, event_loop, WorkerSignals::Tokio(signal)), + (true, true) => self._serve_rth_ssl_ws(callback, event_loop, WorkerSignals::Tokio(signal)), } } - fn serve_wth( - &self, - callback: PyObject, - event_loop: &Bound<PyAny>, - context: Bound<PyAny>, - signal: Py<WorkerSignal>, - ) { - match ( - self.config.websockets_enabled, - self.config.ssl_enabled, - self.config.opt_enabled, - ) { - (false, false, true) => self._serve_wth(callback, event_loop, context, WorkerSignals::Tokio(signal)), - (false, false, false) => self._serve_wth_pyw(callback, event_loop, context, WorkerSignals::Tokio(signal)), - (true, false, true) => self._serve_wth_ws(callback, event_loop, context, WorkerSignals::Tokio(signal)), - (true, false, false) => self._serve_wth_ws_pyw(callback, event_loop, context, WorkerSignals::Tokio(signal)), - (false, true, true) => self._serve_wth_ssl(callback, event_loop, context, WorkerSignals::Tokio(signal)), - (false, true, false) => { - self._serve_wth_ssl_pyw(callback, event_loop, context, WorkerSignals::Tokio(signal)); - } - (true, true, true) => self._serve_wth_ssl_ws(callback, event_loop, context, WorkerSignals::Tokio(signal)), - (true, true, false) => { - self._serve_wth_ssl_ws_pyw(callback, event_loop, context, WorkerSignals::Tokio(signal)); - } + fn serve_wth(&self, callback: Py<CallbackScheduler>, event_loop: &Bound<PyAny>, signal: Py<WorkerSignal>) { + match (self.config.websockets_enabled, self.config.ssl_enabled) { + (false, false) => self._serve_wth(callback, event_loop, WorkerSignals::Tokio(signal)), + (true, false) => self._serve_wth_ws(callback, event_loop, WorkerSignals::Tokio(signal)), + (false, true) => self._serve_wth_ssl(callback, event_loop, WorkerSignals::Tokio(signal)), + (true, true) => self._serve_wth_ssl_ws(callback, event_loop, WorkerSignals::Tokio(signal)), } } } diff --git a/src/workers.rs b/src/workers.rs index 5eb4f316..ebb63a5a 100644 --- a/src/workers.rs +++ b/src/workers.rs @@ -100,7 +100,6 @@ pub(crate) struct WorkerConfig { pub http1_opts: HTTP1Config, pub http2_opts: HTTP2Config, pub websockets_enabled: bool, - pub opt_enabled: bool, pub ssl_enabled: bool, ssl_cert: Option<String>, ssl_key: Option<String>, @@ -118,7 +117,6 @@ impl WorkerConfig { http1_opts: HTTP1Config, http2_opts: HTTP2Config, websockets_enabled: bool, - opt_enabled: bool, ssl_enabled: bool, ssl_cert: Option<&str>, ssl_key: Option<&str>, @@ -134,7 +132,6 @@ impl WorkerConfig { http1_opts, http2_opts, websockets_enabled, - opt_enabled, ssl_enabled, ssl_cert: ssl_cert.map(std::convert::Into::into), ssl_key: ssl_key.map(std::convert::Into::into), @@ -593,9 +590,8 @@ macro_rules! serve_rth { ($func_name:ident, $target:expr) => { fn $func_name( &self, - callback: PyObject, + callback: Py<crate::callbacks::CallbackScheduler>, event_loop: &Bound<PyAny>, - context: Bound<PyAny>, signal: crate::workers::WorkerSignals, ) { pyo3_log::init(); @@ -609,7 +605,7 @@ macro_rules! serve_rth { let http1_opts = self.config.http1_opts.clone(); let http2_opts = self.config.http2_opts.clone(); let backpressure = self.config.backpressure.clone(); - let callback_wrapper = crate::callbacks::CallbackWrapper::new(callback, event_loop.clone(), context); + let callback_wrapper = std::sync::Arc::new(callback); let rt = crate::runtime::init_runtime_mt( self.config.threads, @@ -673,9 +669,9 @@ macro_rules! serve_rth_ssl { ($func_name:ident, $target:expr) => { fn $func_name( &self, - callback: PyObject, + callback: Py<crate::callbacks::CallbackScheduler>, event_loop: &Bound<PyAny>, - context: Bound<PyAny>, + // context: Bound<PyAny>, signal: crate::workers::WorkerSignals, ) { pyo3_log::init(); @@ -690,7 +686,8 @@ macro_rules! serve_rth_ssl { let http2_opts = self.config.http2_opts.clone(); let backpressure = self.config.backpressure.clone(); let tls_cfg = self.config.tls_cfg(); - let callback_wrapper = crate::callbacks::CallbackWrapper::new(callback, event_loop.clone(), context); + // let callback_wrapper = crate::callbacks::CallbackWrapper::new(callback, event_loop.clone(), context); + let callback_wrapper = std::sync::Arc::new(callback); let rt = crate::runtime::init_runtime_mt( self.config.threads, @@ -752,8 +749,9 @@ macro_rules! serve_rth_ssl { } macro_rules! serve_wth_inner { - ($self:expr, $target:expr, $callback:expr, $event_loop:expr, $context:expr, $wid:expr, $workers:expr, $srx:expr) => { - let callback_wrapper = crate::callbacks::CallbackWrapper::new($callback, $event_loop.clone(), $context); + ($self:expr, $target:expr, $callback:expr, $event_loop:expr, $wid:expr, $workers:expr, $srx:expr) => { + // let callback_wrapper = crate::callbacks::CallbackWrapper::new($callback, $event_loop.clone(), $context); + let callback_wrapper = std::sync::Arc::new($callback); let py_loop = std::sync::Arc::new($event_loop.clone().unbind()); for thread_id in 0..$self.config.threads { @@ -807,9 +805,9 @@ macro_rules! serve_wth { ($func_name: ident, $target:expr) => { fn $func_name( &self, - callback: PyObject, + callback: Py<crate::callbacks::CallbackScheduler>, event_loop: &Bound<PyAny>, - context: Bound<PyAny>, + // context: Bound<PyAny>, signal: crate::workers::WorkerSignals, ) { pyo3_log::init(); @@ -819,7 +817,7 @@ macro_rules! serve_wth { let (stx, srx) = tokio::sync::watch::channel(false); let mut workers = vec![]; - crate::workers::serve_wth_inner!(self, $target, callback, event_loop, context, worker_id, workers, srx); + crate::workers::serve_wth_inner!(self, $target, callback, event_loop, worker_id, workers, srx); match signal { crate::workers::WorkerSignals::Tokio(sig) => { @@ -865,8 +863,9 @@ macro_rules! serve_wth { } macro_rules! serve_wth_ssl_inner { - ($self:expr, $target:expr, $callback:expr, $event_loop:expr, $context:expr, $wid:expr, $workers:expr, $srx:expr) => { - let callback_wrapper = crate::callbacks::CallbackWrapper::new($callback, $event_loop.clone(), $context); + ($self:expr, $target:expr, $callback:expr, $event_loop:expr, $wid:expr, $workers:expr, $srx:expr) => { + // let callback_wrapper = crate::callbacks::CallbackWrapper::new($callback, $event_loop.clone(), $context); + let callback_wrapper = std::sync::Arc::new($callback); let py_loop = std::sync::Arc::new($event_loop.clone().unbind()); for thread_id in 0..$self.config.threads { @@ -918,9 +917,9 @@ macro_rules! serve_wth_ssl { ($func_name: ident, $target:expr) => { fn $func_name( &self, - callback: PyObject, + callback: Py<crate::callbacks::CallbackScheduler>, event_loop: &Bound<PyAny>, - context: Bound<PyAny>, + // context: Bound<PyAny>, signal: crate::workers::WorkerSignals, ) { pyo3_log::init(); @@ -930,7 +929,7 @@ macro_rules! serve_wth_ssl { let (stx, srx) = tokio::sync::watch::channel(false); let mut workers = vec![]; - crate::workers::serve_wth_ssl_inner!(self, $target, callback, event_loop, context, worker_id, workers, srx); + crate::workers::serve_wth_ssl_inner!(self, $target, callback, event_loop, worker_id, workers, srx); match signal { crate::workers::WorkerSignals::Tokio(sig) => { diff --git a/src/wsgi/callbacks.rs b/src/wsgi/callbacks.rs index 4a674946..49ca894f 100644 --- a/src/wsgi/callbacks.rs +++ b/src/wsgi/callbacks.rs @@ -9,12 +9,12 @@ use pyo3::{ prelude::, types::{IntoPyDict, PyBytes, PyDict}, }; -use std::{net::SocketAddr, sync::Arc}; +use std::net::SocketAddr; use tokio::sync::oneshot; use super::{io::WSGIProtocol, types::WSGIBody}; use crate::{ - callbacks::CallbackWrapper, + callbacks::ArcCBScheduler, http::{empty_body, HTTPResponseBody}, runtime::RuntimeRef, utils::log_application_callable_exception, @@ -24,7 +24,7 @@ use crate::{ fn run_callback( rt: RuntimeRef, tx: oneshot::Sender<(u16, HeaderMap, HTTPResponseBody)>, - callback: Arc<PyObject>, + cbs: ArcCBScheduler, mut parts: request::Parts, server_addr: SocketAddr, client_addr: SocketAddr, @@ -66,7 +66,7 @@ fn run_callback( let _ = Python::with_gil(\|py\| -> PyResult<()> { let proto = Py::new(py, WSGIProtocol::new(tx))?; - let callback = callback.clone_ref(py); + let callback = cbs.get().cb.clone_ref(py); let environ = PyDict::new_bound(py); environ.set_item(pyo3::intern!(py, "SERVER_PROTOCOL"), version)?; environ.set_item(pyo3::intern!(py, "SERVER_NAME"), server.0)?; @@ -108,7 +108,7 @@ fn run_callback( #[inline(always)] pub(crate) fn call_http( rt: RuntimeRef, - cb: CallbackWrapper, + cb: ArcCBScheduler, server_addr: SocketAddr, client_addr: SocketAddr, scheme: &str, @@ -118,7 +118,7 @@ pub(crate) fn call_http( let scheme: std::sync::Arc<str> = scheme.into(); let (tx, rx) = oneshot::channel(); tokio::task::spawn_blocking(move \|\| { - run_callback(rt, tx, cb.callback, req, server_addr, client_addr, &scheme, body); + run_callback(rt, tx, cb, req, server_addr, client_addr, &scheme, body); }); rx } diff --git a/src/wsgi/http.rs b/src/wsgi/http.rs index 030af3bb..287a0e07 100644 --- a/src/wsgi/http.rs +++ b/src/wsgi/http.rs @@ -3,7 +3,7 @@ use std::net::SocketAddr; use super::callbacks::call_http; use crate::{ - callbacks::CallbackWrapper, + callbacks::ArcCBScheduler, http::{response_500, HTTPRequest, HTTPResponse, HTTPResponseBody}, runtime::RuntimeRef, }; @@ -19,7 +19,7 @@ fn build_response(status: u16, pyheaders: hyper::HeaderMap, body: HTTPResponseBo #[inline] pub(crate) async fn handle( rt: RuntimeRef, - callback: CallbackWrapper, + callback: ArcCBScheduler, server_addr: SocketAddr, client_addr: SocketAddr, req: HTTPRequest, diff --git a/src/wsgi/serve.rs b/src/wsgi/serve.rs index 5a295637..f10922ee 100644 --- a/src/wsgi/serve.rs +++ b/src/wsgi/serve.rs @@ -2,6 +2,7 @@ use pyo3::prelude::; use super::http::handle; +use crate::callbacks::CallbackScheduler; use crate::conversion::{worker_http1_config_from_py, worker_http2_config_from_py}; use crate::workers::{ serve_rth, serve_rth_ssl, serve_wth, serve_wth_ssl, WorkerConfig, WorkerSignalSync, WorkerSignals, @@ -64,7 +65,6 @@ impl WSGIWorker { worker_http1_config_from_py(py, http1_opts)?, worker_http2_config_from_py(py, http2_opts)?, false, - true, ssl_enabled, ssl_cert, ssl_key, @@ -73,29 +73,17 @@ impl WSGIWorker { }) } - fn serve_rth( - &self, - callback: PyObject, - event_loop: &Bound<PyAny>, - context: Bound<PyAny>, - signal: Py<WorkerSignalSync>, - ) { + fn serve_rth(&self, callback: Py<CallbackScheduler>, event_loop: &Bound<PyAny>, signal: Py<WorkerSignalSync>) { match self.config.ssl_enabled { - false => self._serve_rth(callback, event_loop, context, WorkerSignals::Crossbeam(signal)), - true => self._serve_rth_ssl(callback, event_loop, context, WorkerSignals::Crossbeam(signal)), + false => self._serve_rth(callback, event_loop, WorkerSignals::Crossbeam(signal)), + true => self._serve_rth_ssl(callback, event_loop, WorkerSignals::Crossbeam(signal)), } } - fn serve_wth( - &self, - callback: PyObject, - event_loop: &Bound<PyAny>, - context: Bound<PyAny>, - signal: Py<WorkerSignalSync>, - ) { + fn serve_wth(&self, callback: Py<CallbackScheduler>, event_loop: &Bound<PyAny>, signal: Py<WorkerSignalSync>) { match self.config.ssl_enabled { - false => self._serve_wth(callback, event_loop, context, WorkerSignals::Crossbeam(signal)), - true => self._serve_wth_ssl(callback, event_loop, context, WorkerSignals::Crossbeam(signal)), + false => self._serve_wth(callback, event_loop, WorkerSignals::Crossbeam(signal)), + true => self._serve_wth_ssl(callback, event_loop, WorkerSignals::Crossbeam(signal)), } } }	aiohttp / sniffio is unhappy with "--opt" When I use granian with `--opt`, `aiohttp`'s use of [`sniffio`](https://sniffio.readthedocs.io/en/latest/) gets unhappy when trying to detect the async library, without `--opt` it works just fine. How I trigger this is by going to /docs in my `FastAPI` project, which produces an HTTP 500 - Internal Server Error with the below traceback. - Linux amd64 - Python 3.13 with the project managed by `uv` + `pyenv` ```python [ERROR] Application callable raised an exception Traceback (most recent call last): File "./discord_bot/.venv/lib/python3.13/site-packages/granian/asgi.py", line 118, in _http_logger rv = await _runner(scope, proto) File "./discord_bot/.venv/lib/python3.13/site-packages/fastapi/applications.py", line 1054, in __call__ await super().__call__(scope, receive, send) File "./discord_bot/.venv/lib/python3.13/site-packages/starlette/applications.py", line 113, in __call__ await self.middleware_stack(scope, receive, send) File "./discord_bot/.venv/lib/python3.13/site-packages/starlette/middleware/errors.py", line 177, in __call__ response = await self.handler(request, exc) File "./discord_bot/main.py", line 148, in catch_all_handler await async_excepthook(*sys.exc_info()) File "./discord_bot/main.py", line 56, in async_excepthook ryan: discord.User = await discord_client.fetch_user( File "./discord_bot/.venv/lib/python3.13/site-packages/discord/client.py", line 1804, in fetch_user data = await self.http.get_user(user_id) File "./discord_bot/.venv/lib/python3.13/site-packages/discord/http.py", line 286, in request async with self.__session.request( File "./discord_bot/.venv/lib/python3.13/site-packages/aiohttp/client.py", line 1359, in __aenter__ self._resp: _RetType = await self._coro File "./discord_bot/.venv/lib/python3.13/site-packages/aiohttp/client.py", line 579, in _request with timer: File "./discord_bot/.venv/lib/python3.13/site-packages/aiohttp/helpers.py", line 712, in __enter__ raise RuntimeError( RuntimeError: Timeout context manager should be used inside a task 2024-11-30 09:39:34,432 - async_excepthook - ERROR: Exception in AsyncLibraryNotFoundError: unknown async library, or not in async context ```		2024-12-01T13:36:41	[]	[]
emmett-framework/granian	emmett-framework__granian-296	14f258305cfb824649a26fbc286c19409a69e696	diff --git a/Cargo.lock b/Cargo.lock index 4ff8c8c5..1b8493c3 100644 --- a/Cargo.lock +++ b/Cargo.lock @@ -130,6 +130,12 @@ dependencies = [ "crypto-common", ] +[[package]] +name = "either" +version = "1.12.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "3dca9240753cf90908d7e4aac30f630662b02aebaa1b58a3cadabdb23385b58b" + [[package]] name = "equivalent" version = "1.0.1" @@ -276,6 +282,7 @@ dependencies = [ "http-body-util", "hyper", "hyper-util", + "itertools", "log", "mimalloc", "percent-encoding", @@ -437,6 +444,15 @@ version = "2.0.5" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "b248f5224d1d606005e02c97f5aa4e88eeb230488bcc03bc9ca4d7991399f2b5" +[[package]] +name = "itertools" +version = "0.13.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "413ee7dfc52ee1a4949ceeb7dbc8a33f2d6c088194d9f922fb8318faf1f01186" +dependencies = [ + "either", +] + [[package]] name = "itoa" version = "1.0.11" diff --git a/Cargo.toml b/Cargo.toml index 28f748f7..55b8ca09 100644 --- a/Cargo.toml +++ b/Cargo.toml @@ -36,6 +36,7 @@ futures = "0.3" http-body-util = { version = "=0.1" } hyper = { version = "=1.3", features = ["http1", "http2", "server"] } hyper-util = { version = "=0.1", features = ["server-auto", "tokio"] } +itertools = "0.13" log = "0.4" percent-encoding = "=2.3" pin-project = "1.1" diff --git a/src/wsgi/callbacks.rs b/src/wsgi/callbacks.rs index 216ab809..bfce9676 100644 --- a/src/wsgi/callbacks.rs +++ b/src/wsgi/callbacks.rs @@ -7,6 +7,7 @@ use hyper::{ http::{request, uri::Authority}, Version, }; +use itertools::Itertools; use percent_encoding::percent_decode_str; use pyo3::{ prelude::*, @@ -51,15 +52,20 @@ fn run_callback( let content_type = parts.headers.remove(header::CONTENT_TYPE); let content_len = parts.headers.remove(header::CONTENT_LENGTH); let mut headers = Vec::with_capacity(parts.headers.len()); - for (key, val) in &parts.headers { + for key in parts.headers.keys() { headers.push(( format!("HTTP_{}", key.as_str().replace('-', "_").to_uppercase()), - val.to_str().unwrap_or_default(), + parts + .headers + .get_all(key) + .iter() + .map(\|v\| v.to_str().unwrap_or_default()) + .join(","), )); } if !parts.headers.contains_key(header::HOST) { let host = parts.uri.authority().map_or("", Authority::as_str); - headers.push(("HTTP_HOST".to_string(), host)); + headers.push(("HTTP_HOST".to_string(), host.to_string())); } Python::with_gil(\|py\| {	Some cookies are missing on HTTP/2 When running Granian with HTTP/2 some cookies are missing. I created the following test application: ```python def app(environ, start_response): headers = [ ('Content-Type', 'text/html'), ('Set-Cookie', 'csrftoken=abcdef; expires=Mon, 05 May 2025 22:45:47 GMT; HttpOnly; Max-Age=31449600; Path=/; SameSite=Strict; Secure'), ('Set-Cookie', 'sessionid=123456; expires=Mon, 05 May 2025 22:45:47 GMT; HttpOnly; Max-Age=31449600; Path=/; SameSite=Strict; Secure'), ] try: cookies = environ["HTTP_COOKIE"] except KeyError: cookies = "" start_response('200 OK', headers) return [f'<html><body>Cookies: {cookies}</body></html>'.encode()] ``` If I run Granian with HTTP/2 and visis the site with Firefox or Chrome I only see the sessionid cookie, not the csrftoken cookie. With curl it seems to work however. I ran Granian using: ``` granian --http 2 --ssl-keyfile key.pem --ssl-certificate cert.pem --interface wsgi app:app ``` If I run it with HTTP/1 it works without problem however: ``` granian --http 1 --ssl-keyfile key.pem --ssl-certificate cert.pem --interface wsgi app:app ``` It might be caused by https://github.com/hyperium/hyper/issues/2528. I don't know if the Firefox and Chrome split the cookies and curl does not, but if that is the case it would explain why there is no problem with curl but there is a problem with Firefox and Chrome.	@dekkers no, this is a Granian bug on WSGI protocol, thank you for reporting this	2024-05-21T17:35:30	[]	[]
emmett-framework/granian	emmett-framework__granian-247	bf79551e1ccb7d8ed977574f153990dffa239eae	diff --git a/docs/spec/RSGI.md b/docs/spec/RSGI.md index a33c44c9..ffcf20be 100644 --- a/docs/spec/RSGI.md +++ b/docs/spec/RSGI.md @@ -1,6 +1,6 @@ # RSGI Specification -Version: 1.3 +Version: 1.4 ## Abstract @@ -92,7 +92,7 @@ And here are descriptions for the upper attributes: - `method`: the HTTP method name, uppercased - `path`: HTTP request target excluding any query string - `query_string`: URL portion after the `?` -- `headers`: a mapping-like object, where keys is the header name, and value is the header value; header names are always lower-case +- `headers`: a mapping-like object, where key is the header name, and value is the header value; header names are always lower-case; a `get_all` method returns a list of all the header values for the given key - `authority`: an optional string containing the relevant pseudo-header (empty on HTTP versions prior to 2) #### HTTP protocol interface @@ -164,7 +164,7 @@ And here are descriptions for the upper attributes: - `method`: the HTTP method name, uppercased - `path`: HTTP request target excluding any query string - `query_string`: URL portion after the `?` -- `headers`: a mapping-like object, where keys is the header name, and value is the header value; header names are always lower-case +- `headers`: a mapping-like object, where key is the header name, and value is the header value; header names are always lower-case; a `get_all` method returns a list of all the header values for the given key - `authority`: an optional string containing the relevant pseudo-header (empty on HTTP versions prior to 2) #### Websocket protocol interface diff --git a/src/rsgi/types.rs b/src/rsgi/types.rs index bdb4c5b9..59c8fc54 100644 --- a/src/rsgi/types.rs +++ b/src/rsgi/types.rs @@ -16,7 +16,7 @@ use tokio_util::io::ReaderStream; use crate::http::{empty_body, response_404, HTTPResponseBody, HV_SERVER}; -const RSGI_PROTO_VERSION: &str = "1.3"; +const RSGI_PROTO_VERSION: &str = "1.4"; #[pyclass(frozen, module = "granian._granian")] #[derive(Clone)] @@ -85,6 +85,18 @@ impl RSGIHeaders { _ => default, } } + + #[pyo3(signature = (key))] + fn get_all<'p>(&self, py: Python<'p>, key: &'p str) -> &'p PyList { + PyList::new( + py, + self.inner + .get_all(key) + .iter() + .map(\|v\| PyString::new(py, v.to_str().unwrap())) + .collect::<Vec<&PyString>>(), + ) + } } macro_rules! rsgi_scope_cls {	Allow to get all the header values for a single key in `RSGIHeaders`		2024-03-18T22:31:56	[]	[]
emmett-framework/granian	emmett-framework__granian-189	8b7f9a9b7ecb62ec66ea77b97ed877e063fdf115	diff --git a/docs/spec/RSGI.md b/docs/spec/RSGI.md index 8cac21f2..812c559d 100644 --- a/docs/spec/RSGI.md +++ b/docs/spec/RSGI.md @@ -1,6 +1,6 @@ # RSGI Specification -Version: 1.2 +Version: 1.3 ## Abstract @@ -79,6 +79,7 @@ class Scope: path: str query_string: str headers: Mapping[str, str] + authority: Optional[str] ``` And here are descriptions for the upper attributes: @@ -92,6 +93,7 @@ And here are descriptions for the upper attributes: - `path`: HTTP request target excluding any query string - `query_string`: URL portion after the `?` - `headers`: a mapping-like object, where keys is the header name, and value is the header value +- `authority`: an optional string containing the relevant pseudo-header (empty on HTTP versions prior to 2) #### HTTP protocol interface @@ -149,6 +151,7 @@ class Scope: path: str query_string: str headers: Mapping[str, str] + authority: Optional[str] ``` And here are descriptions for the upper attributes: @@ -162,6 +165,7 @@ And here are descriptions for the upper attributes: - `path`: HTTP request target excluding any query string - `query_string`: URL portion after the `?` - `headers`: a mapping-like object, where keys is the header name, and value is the header value +- `authority`: an optional string containing the relevant pseudo-header (empty on HTTP versions prior to 2) #### Websocket protocol interface diff --git a/granian/_granian.pyi b/granian/_granian.pyi index 43c9002b..ae1fce4a 100644 --- a/granian/_granian.pyi +++ b/granian/_granian.pyi @@ -24,6 +24,7 @@ class RSGIScope: method: str path: str query_string: str + authority: Optional[str] @property def headers(self) -> RSGIHeaders: ... diff --git a/src/asgi/types.rs b/src/asgi/types.rs index e148af86..7d2a6386 100644 --- a/src/asgi/types.rs +++ b/src/asgi/types.rs @@ -1,4 +1,8 @@ -use hyper::{header::HeaderMap, Uri, Version}; +use hyper::{ + header::{self, HeaderMap}, + http::uri::Authority, + Uri, Version, +}; use percent_encoding::percent_decode_str; use pyo3::{ prelude::*, @@ -105,6 +109,10 @@ impl ASGIScope { PyBytes::new(py, value.as_bytes()), ))?; } + if !self.headers.contains_key(header::HOST) { + let host = self.uri.authority().map_or("", Authority::as_str); + rv.insert(0, (PyBytes::new(py, b"host"), PyBytes::new(py, host.as_bytes())))?; + } Ok(rv) } } diff --git a/src/rsgi/types.rs b/src/rsgi/types.rs index de18e1ac..b6557472 100644 --- a/src/rsgi/types.rs +++ b/src/rsgi/types.rs @@ -1,8 +1,10 @@ +use anyhow::Result; use futures::TryStreamExt; use http_body_util::BodyExt; use hyper::{ body::Bytes, header::{HeaderMap, HeaderName, HeaderValue, SERVER as HK_SERVER}, + http::uri::Authority, Uri, Version, }; use percent_encoding::percent_decode_str; @@ -36,18 +38,18 @@ impl RSGIHeaders { ret } - fn values(&self) -> PyResult<Vec<&str>> { + fn values(&self) -> Result<Vec<&str>> { let mut ret = Vec::with_capacity(self.inner.keys_len()); for val in self.inner.values() { - ret.push(val.to_str().unwrap()); + ret.push(val.to_str()?); } Ok(ret) } - fn items(&self) -> PyResult<Vec<(&str, &str)>> { + fn items(&self) -> Result<Vec<(&str, &str)>> { let mut ret = Vec::with_capacity(self.inner.keys_len()); for (key, val) in &self.inner { - ret.push((key.as_str(), val.to_str().unwrap())); + ret.push((key.as_str(), val.to_str()?)); } Ok(ret) } @@ -102,7 +104,7 @@ impl RSGIScope { Self { proto: proto.to_string(), http_version, - rsgi_version: "1.2".to_string(), + rsgi_version: "1.3".to_string(), scheme: scheme.to_string(), method: method.to_string(), uri, @@ -130,9 +132,14 @@ impl RSGIScope { } } + #[getter(authority)] + fn get_authority(&self) -> Option<String> { + self.uri.authority().map(Authority::to_string) + } + #[getter(path)] - fn get_path(&self) -> Cow<str> { - percent_decode_str(self.uri.path()).decode_utf8().unwrap() + fn get_path(&self) -> Result<Cow<str>> { + Ok(percent_decode_str(self.uri.path()).decode_utf8()?) } #[getter(query_string)] diff --git a/src/wsgi/types.rs b/src/wsgi/types.rs index 15e349a2..544913c0 100644 --- a/src/wsgi/types.rs +++ b/src/wsgi/types.rs @@ -2,7 +2,8 @@ use futures::Stream; use http_body_util::BodyExt; use hyper::{ body::Bytes, - header::{HeaderMap, CONTENT_LENGTH, CONTENT_TYPE}, + header::{HeaderMap, CONTENT_LENGTH, CONTENT_TYPE, HOST}, + http::uri::Authority, Method, Uri, Version, }; use percent_encoding::percent_decode_str; @@ -173,6 +174,10 @@ impl WSGIScope { val.to_str().unwrap_or_default(), )); } + if !self.headers.contains_key(HOST) { + let host = self.uri.authority().map_or("", Authority::as_str); + headers.push(("HTTP_HOST".to_string(), host)); + } ( percent_decode_str(path).decode_utf8().unwrap(),	Missing "host" header in ASGI scope with HTTP/2 Tried granian today with --interface asgi and a starlette application but I'm missing the "host" header in the asgi scope.	Headers are sent by the client, Granian is not removing nor adding anything to request headers. Tested with the following code and can't reproduce your issue: ```python from starlette.applications import Starlette from starlette.responses import PlainTextResponse from starlette.routing import Route def homepage(request): print('scope', request.scope) print('headers', request.headers) print('host', request.headers.get('host')) return PlainTextResponse('Hello, world!') routes = [ Route('/', homepage), ] app = Starlette(debug=True, routes=routes) ``` Request: ``` ❯ curl -v http://localhost:8000 * Trying 127.0.0.1:8000... * Connected to localhost (127.0.0.1) port 8000 (#0) > GET / HTTP/1.1 > Host: localhost:8000 > User-Agent: curl/8.1.2 > Accept: / > < HTTP/1.1 200 OK < server: granian < content-length: 13 < content-type: text/plain; charset=utf-8 < date: Thu, 25 Jan 2024 09:48:51 GMT < * Connection #0 to host localhost left intact Hello, world! ``` Server logs: ``` ❯ granian --interface asgi t:app [INFO] Starting granian (main PID: 71370) [INFO] Listening at: 127.0.0.1:8000 [INFO] Spawning worker-1 with pid: 71372 [INFO] Started worker-1 [INFO] Started worker-1 runtime-1 scope {'asgi': {'version': '3.0', 'spec_version': '2.3'}, 'extensions': {'http.response.pathsend': {}}, 'type': 'http', 'http_version': '1.1', 'server': ('127.0.0.1', 8000), 'client': ('127.0.0.1', 55309), 'scheme': 'http', 'method': 'GET', 'root_path': '', 'path': '/', 'raw_path': b'/', 'query_string': b'', 'headers': [(b'host', b'localhost:8000'), (b'user-agent', b'curl/8.1.2'), (b'accept', b'/')], 'state': {}, 'app': <starlette.applications.Starlette object at 0x1026237d0>, 'starlette.exception_handlers': ({<class 'starlette.exceptions.HTTPException'>: <bound method ExceptionMiddleware.http_exception of <starlette.middleware.exceptions.ExceptionMiddleware object at 0x10273b920>>, <class 'starlette.exceptions.WebSocketException'>: <bound method ExceptionMiddleware.websocket_exception of <starlette.middleware.exceptions.ExceptionMiddleware object at 0x10273b920>>}, {}), 'router': <starlette.routing.Router object at 0x10273b6b0>, 'endpoint': <function homepage at 0x10262ca40>, 'path_params': {}} headers Headers({'host': 'localhost:8000', 'user-agent': 'curl/8.1.2', 'accept': '/'}) host localhost:8000 ``` Thanks for your extra fast reply! I did some further investigation, and your example works with http, but as soon as i'm using https, the host header disappears from the scope. > granian --interface asgi --ssl-certificate ssl-cert.pem --ssl-keyfile ssl-key.pem t:app ... scope {'asgi': {'version': '3.0', 'spec_version': '2.3'}, 'extensions': {'http.response.pathsend': {}}, 'type': 'http', 'http_version': '2', 'server': ('127.0.0.1', 8000), 'client': ('127.0.0.1', 21342), 'scheme': 'https', 'method': 'GET', 'root_path': '', 'path': '/', 'raw_path': b'/', 'query_string': b'', 'headers': [(b'user-agent', b'curl/7.76.1'), (b'accept', b'/')], 'state': {}, 'app': <starlette.applications.Starlette object at 0x7fd984583bd0>, 'router': <starlette.routing.Router object at 0x7fd984116110>, 'endpoint': <function homepage at 0x7fd984562480>, 'path_params': {}} headers Headers({'user-agent': 'curl/7.76.1', 'accept': '/'}) host None Am I doing sth wrong? With hypercorn I get the host header as expected. @kramar11 thank you for the additional details, I'll investigate this asap. @kramar11 after re-checking your comment, I noticed this is HTTP/2. `Host` header is not allowed on HTTP/2, thus is not present in the scope's headers. Now the actual issue here is that Granian doesn't allow access to pseudo-header `:authority` which would contain the relative information. I see Hypercorn automatically translate such pseudo-header content into the `Host` header on HTTP/2, [as per ASGI spec](https://asgi.readthedocs.io/en/latest/specs/www.html#http-connection-scope). Now I don't like this, but now Granian is out of spec for ASGI, so I gonna do the change. But a the same time I need to figure out something for RSGI and WSGI which can incur in the same issue.	2024-01-25T15:35:01	[]	[]
emmett-framework/granian	emmett-framework__granian-141	9d058bbac55ed1aaf49eac001f6860a387dfe5ac	diff --git a/.github/workflows/build.yml b/.github/workflows/build.yml index c23fade1..5023de4b 100644 --- a/.github/workflows/build.yml +++ b/.github/workflows/build.yml @@ -4,7 +4,7 @@ on: workflow_dispatch env: MATURIN_VERSION: 1.3.0 - PY_ALL: 3.8 3.9 3.10 3.11 pypy3.8 pypy3.9 pypy3.10 + PY_ALL: 3.8 3.9 3.10 3.11 3.12 pypy3.8 pypy3.9 pypy3.10 jobs: wheels: @@ -20,7 +20,7 @@ jobs: - os: ubuntu platform: linux - os: macos - interpreter: 3.8 3.9 3.10 3.11 pypy3.9 pypy3.10 + interpreter: 3.8 3.9 3.10 3.11 3.12 pypy3.9 pypy3.10 - os: ubuntu platform: linux target: aarch64 diff --git a/.github/workflows/lint.yml b/.github/workflows/lint.yml index 08a694ce..64a26d44 100644 --- a/.github/workflows/lint.yml +++ b/.github/workflows/lint.yml @@ -8,7 +8,7 @@ on: env: MATURIN_VERSION: 1.2.3 - PYTHON_VERSION: 3.11 + PYTHON_VERSION: 3.12 jobs: lint: diff --git a/.github/workflows/release.yml b/.github/workflows/release.yml index 91467d8b..767d655a 100644 --- a/.github/workflows/release.yml +++ b/.github/workflows/release.yml @@ -6,7 +6,7 @@ on: env: MATURIN_VERSION: 1.3.0 - PY_ALL: 3.8 3.9 3.10 3.11 pypy3.8 pypy3.9 pypy3.10 + PY_ALL: 3.8 3.9 3.10 3.11 3.12 pypy3.8 pypy3.9 pypy3.10 jobs: sdist: @@ -43,7 +43,7 @@ jobs: - os: ubuntu platform: linux - os: macos - interpreter: 3.8 3.9 3.10 3.11 pypy3.9 pypy3.10 + interpreter: 3.8 3.9 3.10 3.11 3.12 pypy3.9 pypy3.10 - os: ubuntu platform: linux target: aarch64 diff --git a/Makefile b/Makefile index 32b26753..b7e350ad 100644 --- a/Makefile +++ b/Makefile @@ -61,7 +61,7 @@ lint: lint-python lint-rust .PHONY: test test: - pytest -v test + pytest -v tests .PHONY: all all: format build-dev lint test diff --git a/pyproject.toml b/pyproject.toml index fe177f2d..3a8ef2f9 100644 --- a/pyproject.toml +++ b/pyproject.toml @@ -1,7 +1,7 @@ [project] name = 'granian' authors = [ - {name = 'Giovanni Barillari', email = '[email protected]'} + { name = 'Giovanni Barillari', email = '[email protected]' } ] classifiers = [ 'Development Status :: 5 - Production/Stable', @@ -15,11 +15,12 @@ classifiers = [ 'Programming Language :: Python :: 3.9', 'Programming Language :: Python :: 3.10', 'Programming Language :: Python :: 3.11', + 'Programming Language :: Python :: 3.12', 'Programming Language :: Python :: Implementation :: CPython', 'Programming Language :: Python :: Implementation :: PyPy', 'Programming Language :: Python', 'Programming Language :: Rust', - 'Topic :: Internet :: WWW/HTTP' + 'Topic :: Internet :: WWW/HTTP', ] dynamic = [ @@ -27,26 +28,26 @@ dynamic = [ 'keywords', 'license', 'readme', - 'version' + 'version', ] requires-python = '>=3.8' dependencies = [ - 'watchfiles~=0.18', - 'typer~=0.4', - 'uvloop~=0.17.0; sys_platform != "win32" and platform_python_implementation == "CPython"' + 'watchfiles~=0.21', + 'typer~=0.9', + 'uvloop~=0.18.0; sys_platform != "win32" and platform_python_implementation == "CPython"', ] [project.optional-dependencies] lint = [ - 'black~=23.7.0', - 'ruff~=0.0.287' + 'black~=23.9.0', + 'ruff~=0.0.292', ] test = [ - 'httpx~=0.23.0', - 'pytest~=7.1.2', - 'pytest-asyncio~=0.18.3', - 'websockets~=10.3' + 'httpx~=0.25.0', + 'pytest~=7.4.2', + 'pytest-asyncio~=0.21.1', + 'websockets~=11.0', ] [project.urls] @@ -77,7 +78,7 @@ extend-select = [ 'Q', # flake8-quotes 'RUF100', # ruff (unused noqa) 'S', # flake8-bandit - 'W' # pycodestyle + 'W', # pycodestyle ] extend-ignore = [ 'B008', # function calls in args defaults are fine @@ -86,7 +87,7 @@ extend-ignore = [ 'B904', # rising without from is fine 'E501', # leave line length to black 'N818', # leave to us exceptions naming - 'S101' # assert is fine + 'S101', # assert is fine ] flake8-quotes = { inline-quotes = 'single', multiline-quotes = 'double' } mccabe = { max-complexity = 13 } @@ -103,7 +104,7 @@ known-first-party = ['granian', 'tests'] [tool.black] color = true line-length = 120 -target-version = ['py38', 'py39', 'py310', 'py311'] +target-version = ['py38', 'py39', 'py310', 'py311', 'py312'] skip-string-normalization = true # leave this to ruff skip-magic-trailing-comma = true # leave this to ruff	Support for Python 3.12 This depends on: - https://github.com/MagicStack/uvloop/issues/567 - https://github.com/PyO3/pyo3/pull/3493 - related release of pyo3-asyncio (https://github.com/awestlake87/pyo3-asyncio/pull/108) - `watchfiles` package to support 3.12 - PR ref https://github.com/samuelcolvin/watchfiles/pull/248		2023-10-16T10:46:56	[]	[]
emmett-framework/granian	emmett-framework__granian-124	e28eb81db656fd7577a962a2f394d1ff6fe45986	diff --git a/.github/workflows/lint.yml b/.github/workflows/lint.yml new file mode 100644 index 00000000..41bf0c98 --- /dev/null +++ b/.github/workflows/lint.yml @@ -0,0 +1,32 @@ +name: lint + +on: + pull_request: + types: [opened, synchronize] + branches: + - master + +env: + MATURIN_VERSION: 1.2.3 + PYTHON_VERSION: 3.11 + +jobs: + lint: + runs-on: ubuntu-latest + + steps: + - uses: actions/checkout@v3 + - name: Set up Python ${{ env.PYTHON_VERSION }} + uses: actions/setup-python@v4 + with: + python-version: ${{ env.PYTHON_VERSION }} + - name: Install + run: \| + python -m venv .venv + source .venv/bin/activate + pip install maturin==${{ env.MATURIN_VERSION }} + maturin develop --extras=lint + - name: Lint + run: \| + source .venv/bin/activate + make lint diff --git a/.pre-commit-config.yaml b/.pre-commit-config.yaml new file mode 100644 index 00000000..883e8a08 --- /dev/null +++ b/.pre-commit-config.yaml @@ -0,0 +1,26 @@ +fail_fast: true + +repos: +- repo: https://github.com/pre-commit/pre-commit-hooks + rev: v4.0.1 + hooks: + - id: check-yaml + - id: check-toml + - id: end-of-file-fixer + - id: trailing-whitespace + - id: check-added-large-files + +- repo: local + hooks: + - id: lint-python + name: Lint Python + entry: make lint-python + types: [python] + language: system + pass_filenames: false + - id: lint-rust + name: Lint Rust + entry: make lint-rust + types: [rust] + language: system + pass_filenames: false diff --git a/.rustfmt.toml b/.rustfmt.toml new file mode 100644 index 00000000..75306517 --- /dev/null +++ b/.rustfmt.toml @@ -0,0 +1,1 @@ +max_width = 120 diff --git a/Makefile b/Makefile new file mode 100644 index 00000000..32b26753 --- /dev/null +++ b/Makefile @@ -0,0 +1,67 @@ +.DEFAULT_GOAL := all +black = black granian tests +ruff = ruff granian tests + +.PHONY: build-dev +build-dev: + @rm -f granian/.so + maturin develop --extras test + +.PHONY: format +format: + $(black) + $(ruff) --fix --exit-zero + cargo fmt + +.PHONY: lint-python +lint-python: + $(ruff) + $(black) --check --diff + +.PHONY: lint-rust +lint-rust: + cargo fmt --version + cargo fmt --all -- --check + cargo clippy --version + cargo clippy --tests -- \ + -D warnings \ + -W clippy::pedantic \ + -W clippy::dbg_macro \ + -W clippy::print_stdout \ + -A clippy::cast-possible-truncation \ + -A clippy::cast-possible-wrap \ + -A clippy::cast-precision-loss \ + -A clippy::cast-sign-loss \ + -A clippy::declare-interior-mutable-const \ + -A clippy::float-cmp \ + -A clippy::fn-params-excessive-bools \ + -A clippy::if-not-else \ + -A clippy::inline-always \ + -A clippy::manual-let-else \ + -A clippy::match-bool \ + -A clippy::match-same-arms \ + -A clippy::missing-errors-doc \ + -A clippy::missing-panics-doc \ + -A clippy::module-name-repetitions \ + -A clippy::must-use-candidate \ + -A clippy::needless-pass-by-value \ + -A clippy::similar-names \ + -A clippy::single-match-else \ + -A clippy::struct-excessive-bools \ + -A clippy::too-many-arguments \ + -A clippy::too-many-lines \ + -A clippy::type-complexity \ + -A clippy::unnecessary-wraps \ + -A clippy::unused-self \ + -A clippy::used-underscore-binding \ + -A clippy::wrong-self-convention + +.PHONY: lint +lint: lint-python lint-rust + +.PHONY: test +test: + pytest -v test + +.PHONY: all +all: format build-dev lint test diff --git a/granian/__init__.py b/granian/__init__.py index ce03110a..37ebb755 100644 --- a/granian/__init__.py +++ b/granian/__init__.py @@ -1,1 +1,1 @@ -from .server import Granian +from .server import Granian # noqa diff --git a/granian/__main__.py b/granian/__main__.py index c368da26..f1f1f304 100644 --- a/granian/__main__.py +++ b/granian/__main__.py @@ -1,3 +1,4 @@ from granian.cli import cli + cli() diff --git a/granian/__version__.py b/granian/__version__.py index 906d362f..ef7eb44d 100644 --- a/granian/__version__.py +++ b/granian/__version__.py @@ -1,1 +1,1 @@ -__version__ = "0.6.0" +__version__ = '0.6.0' diff --git a/granian/_granian.pyi b/granian/_granian.pyi index 797d95f0..836b1283 100644 --- a/granian/_granian.pyi +++ b/granian/_granian.pyi @@ -1,12 +1,10 @@ -from typing import Any, Dict, List, Tuple, Optional +from typing import Any, Dict, List, Optional, Tuple from ._types import WebsocketMessage - class ASGIScope: def as_dict(self, root_path: str) -> Dict[str, Any]: ... - class RSGIHeaders: def __contains__(self, key: str) -> bool: ... def keys(self) -> List[str]: ... @@ -14,7 +12,6 @@ class RSGIHeaders: def items(self) -> List[Tuple[str]]: ... def get(self, key: str, default: Any = None) -> Any: ... - class RSGIScope: proto: str http_version: str @@ -29,12 +26,10 @@ class RSGIScope: @property def headers(self) -> RSGIHeaders: ... - class RSGIHTTPStreamTransport: async def send_bytes(self, data: bytes): ... async def send_str(self, data: str): ... - class RSGIHTTPProtocol: async def __call__(self) -> bytes: ... def response_empty(self, status: int, headers: List[Tuple[str, str]]): ... @@ -43,25 +38,17 @@ class RSGIHTTPProtocol: def response_file(self, status: int, headers: List[Tuple[str, str]], file: str): ... def response_stream(self, status: int, headers: List[Tuple[str, str]]) -> RSGIHTTPStreamTransport: ... - class RSGIWebsocketTransport: async def receive(self) -> WebsocketMessage: ... async def send_bytes(self, data: bytes): ... async def send_str(self, data: str): ... - class RSGIWebsocketProtocol: async def accept(self) -> RSGIWebsocketTransport: ... def close(self, status: Optional[int]) -> Tuple[int, bool]: ... - -class RSGIProtocolError(RuntimeError): - ... - - -class RSGIProtocolClosed(RuntimeError): - ... - +class RSGIProtocolError(RuntimeError): ... +class RSGIProtocolClosed(RuntimeError): ... class WSGIScope: def to_environ(self, environ: Dict[str, Any]) -> Dict[str, Any]: ... diff --git a/granian/_internal.py b/granian/_internal.py index 26fedbdf..6ebc300a 100644 --- a/granian/_internal.py +++ b/granian/_internal.py @@ -2,22 +2,21 @@ import re import sys import traceback - from types import ModuleType from typing import Callable, List, Optional def get_import_components(path: str) -> List[Optional[str]]: - return (re.split(r":(?![\\/])", path, 1) + [None])[:2] + return (re.split(r':(?![\\/])', path, 1) + [None])[:2] def prepare_import(path: str) -> str: path = os.path.realpath(path) fname, ext = os.path.splitext(path) - if ext == ".py": + if ext == '.py': path = fname - if os.path.basename(path) == "__init__": + if os.path.basename(path) == '__init__': path = os.path.dirname(path) module_name = [] @@ -27,26 +26,22 @@ def prepare_import(path: str) -> str: path, name = os.path.split(path) module_name.append(name) - if not os.path.exists(os.path.join(path, "__init__.py")): + if not os.path.exists(os.path.join(path, '__init__.py')): break if sys.path[0] != path: sys.path.insert(0, path) - return ".".join(module_name[::-1]) + return '.'.join(module_name[::-1]) -def load_module( - module_name: str, - raise_on_failure: bool = True -) -> Optional[ModuleType]: +def load_module(module_name: str, raise_on_failure: bool = True) -> Optional[ModuleType]: try: __import__(module_name) except ImportError: if sys.exc_info()[-1].tb_next: raise RuntimeError( - f"While importing '{module_name}', an ImportError was raised:" - f"\n\n{traceback.format_exc()}" + f"While importing '{module_name}', an ImportError was raised:" f"\n\n{traceback.format_exc()}" ) elif raise_on_failure: raise RuntimeError(f"Could not import '{module_name}'.") @@ -58,9 +53,9 @@ def load_module( def load_target(target: str) -> Callable[..., None]: path, name = get_import_components(target) path = prepare_import(path) if path else None - name = name or "app" + name = name or 'app' module = load_module(path) rv = module - for element in name.split("."): + for element in name.split('.'): rv = getattr(rv, element) return rv diff --git a/granian/_loops.py b/granian/_loops.py index b38f0adc..22145c46 100644 --- a/granian/_loops.py +++ b/granian/_loops.py @@ -2,12 +2,11 @@ import os import signal import sys - from typing import Any, Callable, Dict, Iterable, List, Optional, Tuple class Registry: - __slots__ = ["_data"] + __slots__ = ['_data'] def __init__(self): self._data: Dict[str, Callable[..., Any]] = {} @@ -22,6 +21,7 @@ def register(self, key: str) -> Callable[[], Callable[..., Any]]: def wrap(builder: Callable[..., Any]) -> Callable[..., Any]: self._data[key] = builder return builder + return wrap def get(self, key: str) -> Callable[..., Any]: @@ -31,18 +31,13 @@ def get(self, key: str) -> Callable[..., Any]: raise RuntimeError(f"'{key}' implementation not available.") - class BuilderRegistry(Registry): __slots__ = [] def __init__(self): self._data: Dict[str, Tuple[Callable[..., Any], List[str]]] = {} - def register( - self, - key: str, - packages: Optional[List[str]] = None - ) -> Callable[[], Callable[..., Any]]: + def register(self, key: str, packages: Optional[List[str]] = None) -> Callable[[], Callable[..., Any]]: packages = packages or [] def wrap(builder: Callable[..., Any]) -> Callable[..., Any]: @@ -56,6 +51,7 @@ def wrap(builder: Callable[..., Any]) -> Callable[..., Any]: if implemented: self._data[key] = (builder, loaded_packages) return builder + return wrap def get(self, key: str) -> Callable[..., Any]: diff --git a/granian/asgi.py b/granian/asgi.py index e0b3fd5d..bb955e54 100644 --- a/granian/asgi.py +++ b/granian/asgi.py @@ -1,5 +1,4 @@ import asyncio - from functools import wraps from ._granian import ASGIScope as Scope @@ -22,12 +21,7 @@ def __init__(self, callable): async def handle(self): try: await self.callable( - { - "type": "lifespan", - "asgi": {"version": "3.0", "spec_version": "2.3"} - }, - self.receive, - self.send + {'type': 'lifespan', 'asgi': {'version': '3.0', 'spec_version': '2.3'}}, self.receive, self.send ) except Exception: self.errored = True @@ -43,7 +37,7 @@ async def startup(self): loop = asyncio.get_event_loop() _handler_task = loop.create_task(self.handle()) - await self.event_queue.put({"type": "lifespan.startup"}) + await self.event_queue.put({'type': 'lifespan.startup'}) await self.event_startup.wait() if self.failure_startup or (self.errored and not self.unsupported): @@ -53,7 +47,7 @@ async def shutdown(self): if self.errored: return - await self.event_queue.put({"type": "lifespan.shutdown"}) + await self.event_queue.put({'type': 'lifespan.shutdown'}) await self.event_shutdown.wait() if self.failure_shutdown or (self.errored and not self.unsupported): @@ -89,14 +83,14 @@ def _handle_shutdown_failed(self, message): # self.logger.error(message["message"]) _event_handlers = { - "lifespan.startup.complete": _handle_startup_complete, - "lifespan.startup.failed": _handle_startup_failed, - "lifespan.shutdown.complete": _handle_shutdown_complete, - "lifespan.shutdown.failed": _handle_shutdown_failed + 'lifespan.startup.complete': _handle_startup_complete, + 'lifespan.startup.failed': _handle_startup_failed, + 'lifespan.shutdown.complete': _handle_shutdown_complete, + 'lifespan.shutdown.failed': _handle_shutdown_failed, } async def send(self, message): - handler = self._event_handlers[message["type"]] + handler = self._event_handlers[message['type']] handler(self, message) @@ -108,6 +102,7 @@ def _send_wrapper(proto): @wraps(proto) def send(data): return proto(_noop_coro, data) + return send @@ -116,9 +111,6 @@ def _callback_wrapper(callback, scope_opts): @wraps(callback) def wrapper(scope: Scope, proto): - return callback( - scope.as_dict(root_url_path), - proto.receive, - _send_wrapper(proto.send) - ) + return callback(scope.as_dict(root_url_path), proto.receive, _send_wrapper(proto.send)) + return wrapper diff --git a/granian/cli.py b/granian/cli.py index c37ec96a..0bbbce3b 100644 --- a/granian/cli.py +++ b/granian/cli.py @@ -1,107 +1,55 @@ import json - from pathlib import Path from typing import Optional import typer from .__version__ import __version__ -from .constants import Interfaces, HTTPModes, Loops, ThreadModes +from .constants import HTTPModes, Interfaces, Loops, ThreadModes from .log import LogLevels from .server import Granian -cli = typer.Typer(name="granian", context_settings={"ignore_unknown_options": True}) +cli = typer.Typer(name='granian', context_settings={'ignore_unknown_options': True}) def version_callback(value: bool): if value: - typer.echo(f"{cli.info.name} {__version__}") + typer.echo(f'{cli.info.name} {__version__}') raise typer.Exit() @cli.command() def main( - app: str = typer.Argument(..., help="Application target to serve."), - host: str = typer.Option("127.0.0.1", help="Host address to bind to."), - port: int = typer.Option(8000, help="Port to bind to."), - interface: Interfaces = typer.Option( - Interfaces.RSGI.value, - help="Application interface type." - ), - http: HTTPModes = typer.Option( - HTTPModes.auto.value, - help="HTTP version." - ), - websockets: bool = typer.Option( - True, - "--ws/--no-ws", - help="Enable websockets handling", - show_default="enabled" - ), - workers: int = typer.Option(1, min=1, help="Number of worker processes."), - threads: int = typer.Option(1, min=1, help="Number of threads."), - threading_mode: ThreadModes = typer.Option( - ThreadModes.workers.value, - help="Threading mode to use." - ), - loop: Loops = typer.Option(Loops.auto.value, help="Event loop implementation"), - loop_opt: bool = typer.Option( - False, - "--opt/--no-opt", - help="Enable loop optimizations", - show_default="disabled" - ), - backlog: int = typer.Option( - 1024, - min=128, - help="Maximum number of connections to hold in backlog." - ), - log_level: LogLevels = typer.Option( - LogLevels.info.value, - help="Log level", - case_sensitive=False - ), + app: str = typer.Argument(..., help='Application target to serve.'), + host: str = typer.Option('127.0.0.1', help='Host address to bind to.'), + port: int = typer.Option(8000, help='Port to bind to.'), + interface: Interfaces = typer.Option(Interfaces.RSGI.value, help='Application interface type.'), + http: HTTPModes = typer.Option(HTTPModes.auto.value, help='HTTP version.'), + websockets: bool = typer.Option(True, '--ws/--no-ws', help='Enable websockets handling', show_default='enabled'), + workers: int = typer.Option(1, min=1, help='Number of worker processes.'), + threads: int = typer.Option(1, min=1, help='Number of threads.'), + threading_mode: ThreadModes = typer.Option(ThreadModes.workers.value, help='Threading mode to use.'), + loop: Loops = typer.Option(Loops.auto.value, help='Event loop implementation'), + loop_opt: bool = typer.Option(False, '--opt/--no-opt', help='Enable loop optimizations', show_default='disabled'), + backlog: int = typer.Option(1024, min=128, help='Maximum number of connections to hold in backlog.'), + log_level: LogLevels = typer.Option(LogLevels.info.value, help='Log level', case_sensitive=False), log_config: Optional[Path] = typer.Option( - None, - help="Logging configuration file (json)", - exists=True, - file_okay=True, - dir_okay=False, - readable=True + None, help='Logging configuration file (json)', exists=True, file_okay=True, dir_okay=False, readable=True ), ssl_keyfile: Optional[Path] = typer.Option( - None, - help="SSL key file", - exists=True, - file_okay=True, - dir_okay=False, - readable=True + None, help='SSL key file', exists=True, file_okay=True, dir_okay=False, readable=True ), ssl_certificate: Optional[Path] = typer.Option( - None, - help="SSL certificate file", - exists=True, - file_okay=True, - dir_okay=False, - readable=True - ), - url_path_prefix: Optional[str] = typer.Option( - None, - help="URL path prefix the app is mounted on" + None, help='SSL certificate file', exists=True, file_okay=True, dir_okay=False, readable=True ), + url_path_prefix: Optional[str] = typer.Option(None, help='URL path prefix the app is mounted on'), reload: bool = typer.Option( - False, - "--reload/--no-reload", - help="Enable auto reload on application's files changes" + False, '--reload/--no-reload', help="Enable auto reload on application's files changes" ), _: Optional[bool] = typer.Option( - None, - "--version", - callback=version_callback, - is_eager=True, - help="Shows the version and exit." - ) + None, '--version', callback=version_callback, is_eager=True, help='Shows the version and exit.' + ), ): log_dictconfig = None if log_config: @@ -109,7 +57,7 @@ def main( try: log_dictconfig = json.loads(log_config_file.read()) except Exception: - print("Unable to parse provided logging config.") + print('Unable to parse provided logging config.') raise typer.Exit(1) Granian( @@ -131,5 +79,5 @@ def main( ssl_cert=ssl_certificate, ssl_key=ssl_keyfile, url_path_prefix=url_path_prefix, - reload=reload + reload=reload, ).serve() diff --git a/granian/constants.py b/granian/constants.py index 1579afea..f0b2100d 100644 --- a/granian/constants.py +++ b/granian/constants.py @@ -2,23 +2,23 @@ class Interfaces(str, Enum): - ASGI = "asgi" - RSGI = "rsgi" - WSGI = "wsgi" + ASGI = 'asgi' + RSGI = 'rsgi' + WSGI = 'wsgi' class HTTPModes(str, Enum): - auto = "auto" - http1 = "1" - http2 = "2" + auto = 'auto' + http1 = '1' + http2 = '2' class ThreadModes(str, Enum): - runtime = "runtime" - workers = "workers" + runtime = 'runtime' + workers = 'workers' class Loops(str, Enum): - auto = "auto" - asyncio = "asyncio" - uvloop = "uvloop" + auto = 'auto' + asyncio = 'asyncio' + uvloop = 'uvloop' diff --git a/granian/log.py b/granian/log.py index 378e9a4b..df21b69f 100644 --- a/granian/log.py +++ b/granian/log.py @@ -1,18 +1,17 @@ import copy import logging import logging.config - from enum import Enum from typing import Any, Dict, Optional class LogLevels(str, Enum): - critical = "critical" - error = "error" - warning = "warning" - warn = "warn" - info = "info" - debug = "debug" + critical = 'critical' + error = 'error' + warning = 'warning' + warn = 'warn' + info = 'info' + debug = 'debug' log_levels_map = { @@ -21,27 +20,21 @@ class LogLevels(str, Enum): LogLevels.warning: logging.WARNING, LogLevels.warn: logging.WARN, LogLevels.info: logging.INFO, - LogLevels.debug: logging.DEBUG + LogLevels.debug: logging.DEBUG, } LOGGING_CONFIG = { - "version": 1, - "disable_existing_loggers": False, - "root": {"level": "INFO", "handlers": ["console"]}, - "formatters": { - "generic": { - "()": "logging.Formatter", - "fmt": "[%(levelname)s] %(message)s", - "datefmt": "[%Y-%m-%d %H:%M:%S %z]" + 'version': 1, + 'disable_existing_loggers': False, + 'root': {'level': 'INFO', 'handlers': ['console']}, + 'formatters': { + 'generic': { + '()': 'logging.Formatter', + 'fmt': '[%(levelname)s] %(message)s', + 'datefmt': '[%Y-%m-%d %H:%M:%S %z]', } }, - "handlers": { - "console": { - "formatter": "generic", - "class": "logging.StreamHandler", - "stream": "ext://sys.stdout", - } - } + 'handlers': {'console': {'formatter': 'generic', 'class': 'logging.StreamHandler', 'stream': 'ext://sys.stdout'}}, } logger = logging.getLogger() @@ -51,5 +44,5 @@ def configure_logging(level: LogLevels, config: Optional[Dict[str, Any]] = None) log_config = copy.deepcopy(LOGGING_CONFIG) if config: log_config.update(config) - log_config["root"]["level"] = log_levels_map[level] + log_config['root']['level'] = log_levels_map[level] logging.config.dictConfig(log_config) diff --git a/granian/net.py b/granian/net.py index 26a60d3f..d0b2bc5f 100644 --- a/granian/net.py +++ b/granian/net.py @@ -2,7 +2,5 @@ from ._granian import ListenerHolder as SocketHolder -copyreg.pickle( - SocketHolder, - lambda v: (SocketHolder, v.__getstate__()) -) + +copyreg.pickle(SocketHolder, lambda v: (SocketHolder, v.__getstate__())) diff --git a/granian/rsgi.py b/granian/rsgi.py index 64555e01..4f9ecbba 100644 --- a/granian/rsgi.py +++ b/granian/rsgi.py @@ -2,12 +2,12 @@ from typing import Union from ._granian import ( - RSGIHTTPProtocol as HTTPProtocol, - RSGIWebsocketProtocol as WebsocketProtocol, - RSGIHeaders as Headers, - RSGIScope as Scope, - RSGIProtocolError as ProtocolError, - RSGIProtocolClosed as ProtocolClosed + RSGIHeaders as Headers, # noqa + RSGIHTTPProtocol as HTTPProtocol, # noqa + RSGIProtocolClosed as ProtocolClosed, # noqa + RSGIProtocolError as ProtocolError, # noqa + RSGIScope as Scope, # noqa + RSGIWebsocketProtocol as WebsocketProtocol, # noqa ) diff --git a/granian/server.py b/granian/server.py index a9800571..f0f610e9 100644 --- a/granian/server.py +++ b/granian/server.py @@ -5,7 +5,6 @@ import ssl import sys import threading - from functools import partial from pathlib import Path from typing import Any, Callable, Dict, List, Optional @@ -16,11 +15,12 @@ from ._granian import ASGIWorker, RSGIWorker, WSGIWorker from ._internal import load_target from .asgi import LifespanProtocol, _callback_wrapper as _asgi_call_wrap -from .constants import Interfaces, HTTPModes, Loops, ThreadModes +from .constants import HTTPModes, Interfaces, Loops, ThreadModes from .log import LogLevels, configure_logging, logger from .net import SocketHolder from .wsgi import _callback_wrapper as _wsgi_call_wrap + multiprocessing.allow_connection_pickling() @@ -30,7 +30,7 @@ class Granian: def __init__( self, target: str, - address: str = "127.0.0.1", + address: str = '127.0.0.1', port: int = 8000, interface: Interfaces = Interfaces.RSGI, workers: int = 1, @@ -48,7 +48,7 @@ def __init__( ssl_cert: Optional[Path] = None, ssl_key: Optional[Path] = None, url_path_prefix: Optional[str] = None, - reload: bool = False + reload: bool = False, ): self.target = target self.bind_addr = address @@ -75,11 +75,7 @@ def __init__( self.procs: List[multiprocessing.Process] = [] self.exit_event = threading.Event() - def build_ssl_context( - self, - cert: Optional[Path], - key: Optional[Path] - ): + def build_ssl_context(self, cert: Optional[Path], key: Optional[Path]): if not (cert and key): self.ssl_ctx = (False, None, None) return @@ -108,7 +104,7 @@ def _spawn_asgi_worker( log_level, log_config, ssl_ctx, - scope_opts + scope_opts, ): from granian._loops import loops, set_loop_signals @@ -129,29 +125,12 @@ def _spawn_asgi_worker( wcallback = future_watcher_wrapper(wcallback) worker = ASGIWorker( - worker_id, - sfd, - threads, - pthreads, - http_mode, - http1_buffer_size, - websockets, - loop_opt, - ssl_ctx - ) - serve = getattr(worker, { - ThreadModes.runtime: "serve_rth", - ThreadModes.workers: "serve_wth" - }[threading_mode]) - serve( - wcallback, - loop, - contextvars.copy_context(), - shutdown_event.wait() + worker_id, sfd, threads, pthreads, http_mode, http1_buffer_size, websockets, loop_opt, ssl_ctx ) + serve = getattr(worker, {ThreadModes.runtime: 'serve_rth', ThreadModes.workers: 'serve_wth'}[threading_mode]) + serve(wcallback, loop, contextvars.copy_context(), shutdown_event.wait()) loop.run_until_complete(lifespan_handler.shutdown()) - @staticmethod def _spawn_rsgi_worker( worker_id, @@ -168,7 +147,7 @@ def _spawn_rsgi_worker( log_level, log_config, ssl_ctx, - scope_opts + scope_opts, ): from granian._loops import loops, set_loop_signals @@ -176,38 +155,23 @@ def _spawn_rsgi_worker( loop = loops.get(loop_impl) sfd = socket.fileno() target = callback_loader() - callback = ( - getattr(target, '__rsgi__') if hasattr(target, '__rsgi__') else - target - ) + callback = getattr(target, '__rsgi__') if hasattr(target, '__rsgi__') else target callback_init = ( - getattr(target, '__rsgi_init__') if hasattr(target, '__rsgi_init__') else - lambda args, *kwargs: None + getattr(target, '__rsgi_init__') if hasattr(target, '__rsgi_init__') else lambda args, *kwargs: None ) shutdown_event = set_loop_signals(loop, [signal.SIGTERM, signal.SIGINT]) callback_init(loop) worker = RSGIWorker( - worker_id, - sfd, - threads, - pthreads, - http_mode, - http1_buffer_size, - websockets, - loop_opt, - ssl_ctx + worker_id, sfd, threads, pthreads, http_mode, http1_buffer_size, websockets, loop_opt, ssl_ctx ) - serve = getattr(worker, { - ThreadModes.runtime: "serve_rth", - ThreadModes.workers: "serve_wth" - }[threading_mode]) + serve = getattr(worker, {ThreadModes.runtime: 'serve_rth', ThreadModes.workers: 'serve_wth'}[threading_mode]) serve( future_watcher_wrapper(callback) if not loop_opt else callback, loop, contextvars.copy_context(), - shutdown_event.wait() + shutdown_event.wait(), ) @staticmethod @@ -226,7 +190,7 @@ def _spawn_wsgi_worker( log_level, log_config, ssl_ctx, - scope_opts + scope_opts, ): from granian._loops import loops, set_loop_signals @@ -237,46 +201,20 @@ def _spawn_wsgi_worker( shutdown_event = set_loop_signals(loop, [signal.SIGTERM, signal.SIGINT]) - worker = WSGIWorker( - worker_id, - sfd, - threads, - pthreads, - http_mode, - http1_buffer_size, - ssl_ctx - ) - serve = getattr(worker, { - ThreadModes.runtime: "serve_rth", - ThreadModes.workers: "serve_wth" - }[threading_mode]) - serve( - _wsgi_call_wrap(callback, scope_opts), - loop, - contextvars.copy_context(), - shutdown_event.wait() - ) + worker = WSGIWorker(worker_id, sfd, threads, pthreads, http_mode, http1_buffer_size, ssl_ctx) + serve = getattr(worker, {ThreadModes.runtime: 'serve_rth', ThreadModes.workers: 'serve_wth'}[threading_mode]) + serve(_wsgi_call_wrap(callback, scope_opts), loop, contextvars.copy_context(), shutdown_event.wait()) def _init_shared_socket(self): - self._shd = SocketHolder.from_address( - self.bind_addr, - self.bind_port, - self.backlog - ) + self._shd = SocketHolder.from_address(self.bind_addr, self.bind_port, self.backlog) self._sfd = self._shd.get_fd() def signal_handler(self, args, kwargs): self.exit_event.set() - def _spawn_proc( - self, - id, - target, - callback_loader, - socket_loader - ) -> multiprocessing.Process: + def _spawn_proc(self, id, target, callback_loader, socket_loader) -> multiprocessing.Process: return multiprocessing.get_context().Process( - name="granian-worker", + name='granian-worker', target=target, args=( id, @@ -293,10 +231,8 @@ def _spawn_proc( self.log_level, self.log_config, self.ssl_ctx, - { - "url_path_prefix": self.url_path_prefix - } - ) + {'url_path_prefix': self.url_path_prefix}, + ), ) def _spawn_workers(self, sock, spawn_target, target_loader): @@ -305,14 +241,11 @@ def socket_loader(): for idx in range(self.workers): proc = self._spawn_proc( - id=idx + 1, - target=spawn_target, - callback_loader=target_loader, - socket_loader=socket_loader + id=idx + 1, target=spawn_target, callback_loader=target_loader, socket_loader=socket_loader ) proc.start() self.procs.append(proc) - logger.info(f"Spawning worker-{idx + 1} with pid: {proc.pid}") + logger.info(f'Spawning worker-{idx + 1} with pid: {proc.pid}') def _stop_workers(self): for proc in self.procs: @@ -321,7 +254,7 @@ def _stop_workers(self): proc.join() def startup(self, spawn_target, target_loader): - logger.info("Starting granian") + logger.info('Starting granian') for sig in self.SIGNALS: signal.signal(sig, self.signal_handler) @@ -329,13 +262,13 @@ def startup(self, spawn_target, target_loader): self._init_shared_socket() sock = socket.socket(fileno=self._sfd) sock.set_inheritable(True) - logger.info(f"Listening at: {self.bind_addr}:{self.bind_port}") + logger.info(f'Listening at: {self.bind_addr}:{self.bind_port}') self._spawn_workers(sock, spawn_target, target_loader) return sock def shutdown(self): - logger.info("Shutting down granian") + logger.info('Shutting down granian') self._stop_workers() def _serve(self, spawn_target, target_loader): @@ -360,12 +293,12 @@ def serve( self, spawn_target: Optional[Callable[..., None]] = None, target_loader: Optional[Callable[..., Callable[..., Any]]] = None, - wrap_loader: bool = True + wrap_loader: bool = True, ): default_spawners = { Interfaces.ASGI: self._spawn_asgi_worker, Interfaces.RSGI: self._spawn_rsgi_worker, - Interfaces.WSGI: self._spawn_wsgi_worker + Interfaces.WSGI: self._spawn_wsgi_worker, } if target_loader: if wrap_loader: @@ -383,8 +316,5 @@ def serve( "Number of workers will now fallback to 1." ) - serve_method = ( - self._serve_with_reloader if self.reload_on_changes else - self._serve - ) + serve_method = self._serve_with_reloader if self.reload_on_changes else self._serve serve_method(spawn_target, target_loader) diff --git a/granian/wsgi.py b/granian/wsgi.py index 6ba53db4..6b73eaed 100644 --- a/granian/wsgi.py +++ b/granian/wsgi.py @@ -1,6 +1,5 @@ import os import sys - from functools import wraps from typing import Any, List, Tuple @@ -14,30 +13,27 @@ def __init__(self): self.status = 200 self.headers = [] - def __call__( - self, - status: str, - headers: List[Tuple[str, str]], - exc_info: Any = None - ): + def __call__(self, status: str, headers: List[Tuple[str, str]], exc_info: Any = None): self.status = int(status.split(' ', 1)[0]) self.headers = headers def _callback_wrapper(callback, scope_opts): basic_env = dict(os.environ) - basic_env.update({ - 'GATEWAY_INTERFACE': 'CGI/1.1', - 'SCRIPT_NAME': scope_opts.get('url_path_prefix') or '', - 'SERVER_SOFTWARE': 'Granian', - 'wsgi.errors': sys.stderr, - 'wsgi.input_terminated': True, - 'wsgi.input': None, - 'wsgi.multiprocess': False, - 'wsgi.multithread': False, - 'wsgi.run_once': False, - 'wsgi.version': (1, 0) - }) + basic_env.update( + { + 'GATEWAY_INTERFACE': 'CGI/1.1', + 'SCRIPT_NAME': scope_opts.get('url_path_prefix') or '', + 'SERVER_SOFTWARE': 'Granian', + 'wsgi.errors': sys.stderr, + 'wsgi.input_terminated': True, + 'wsgi.input': None, + 'wsgi.multiprocess': False, + 'wsgi.multithread': False, + 'wsgi.run_once': False, + 'wsgi.version': (1, 0), + } + ) @wraps(callback) def wrapper(scope: Scope) -> Tuple[int, List[Tuple[str, str]], bytes]: @@ -46,7 +42,7 @@ def wrapper(scope: Scope) -> Tuple[int, List[Tuple[str, str]], bytes]: if isinstance(rv, list): resp_type = 0 - rv = b"".join(rv) + rv = b''.join(rv) else: resp_type = 1 rv = iter(rv) diff --git a/pyproject.toml b/pyproject.toml index d2a79e0b..0f4d07b9 100644 --- a/pyproject.toml +++ b/pyproject.toml @@ -1,65 +1,111 @@ [project] -name = "granian" +name = 'granian' authors = [ - {name = "Giovanni Barillari", email = "[email protected]"} + {name = 'Giovanni Barillari', email = '[email protected]'} ] classifiers = [ - "Development Status :: 5 - Production/Stable", - "Intended Audience :: Developers", - "License :: OSI Approved :: BSD License", - "Operating System :: MacOS", - "Operating System :: Microsoft :: Windows", - "Operating System :: POSIX :: Linux", - "Programming Language :: Python :: 3", - "Programming Language :: Python :: 3.8", - "Programming Language :: Python :: 3.9", - "Programming Language :: Python :: 3.10", - "Programming Language :: Python :: 3.11", - "Programming Language :: Python :: Implementation :: CPython", - "Programming Language :: Python :: Implementation :: PyPy", - "Programming Language :: Python", - "Programming Language :: Rust", - "Topic :: Internet :: WWW/HTTP" + 'Development Status :: 5 - Production/Stable', + 'Intended Audience :: Developers', + 'License :: OSI Approved :: BSD License', + 'Operating System :: MacOS', + 'Operating System :: Microsoft :: Windows', + 'Operating System :: POSIX :: Linux', + 'Programming Language :: Python :: 3', + 'Programming Language :: Python :: 3.8', + 'Programming Language :: Python :: 3.9', + 'Programming Language :: Python :: 3.10', + 'Programming Language :: Python :: 3.11', + 'Programming Language :: Python :: Implementation :: CPython', + 'Programming Language :: Python :: Implementation :: PyPy', + 'Programming Language :: Python', + 'Programming Language :: Rust', + 'Topic :: Internet :: WWW/HTTP' ] dynamic = [ - "description", - "keywords", - "license", - "readme", - "version" + 'description', + 'keywords', + 'license', + 'readme', + 'version' ] -requires-python = ">=3.8" +requires-python = '>=3.8' dependencies = [ - "watchfiles~=0.18", - "typer~=0.4", - "uvloop~=0.17.0; sys_platform != 'win32' and platform_python_implementation == 'CPython'" + 'watchfiles~=0.18', + 'typer~=0.4', + 'uvloop~=0.17.0; sys_platform != "win32" and platform_python_implementation == "CPython"' ] [project.optional-dependencies] +lint = [ + 'black~=23.7.0', + 'ruff~=0.0.287' +] test = [ - "httpx~=0.23.0", - "pytest~=7.1.2", - "pytest-asyncio~=0.18.3", - "websockets~=10.3" + 'httpx~=0.23.0', + 'pytest~=7.1.2', + 'pytest-asyncio~=0.18.3', + 'websockets~=10.3' ] [project.urls] -Homepage = "https://github.com/emmett-framework/granian" -Funding = "https://github.com/sponsors/gi0baro" -Source = "https://github.com/emmett-framework/granian" +Homepage = 'https://github.com/emmett-framework/granian' +Funding = 'https://github.com/sponsors/gi0baro' +Source = 'https://github.com/emmett-framework/granian' [project.scripts] -granian = "granian:cli.cli" +granian = 'granian:cli.cli' [build-system] -requires = ["maturin>=1.1.0,<1.3.0"] -build-backend = "maturin" +requires = ['maturin>=1.1.0,<1.3.0'] +build-backend = 'maturin' [tool.maturin] -module-name = "granian._granian" -bindings = "pyo3" +module-name = 'granian._granian' +bindings = 'pyo3' + +[tool.ruff] +line-length = 120 +extend-select = [ + # E and F are enabled by default + 'B', # flake8-bugbear + 'C4', # flake8-comprehensions + 'C90', # mccabe + 'I', # isort + 'N', # pep8-naming + 'Q', # flake8-quotes + 'RUF100', # ruff (unused noqa) + 'S', # flake8-bandit + 'W' # pycodestyle +] +extend-ignore = [ + 'B008', # function calls in args defaults are fine + 'B009', # getattr with constants is fine + 'B034', # re.split won't confuse us + 'B904', # rising without from is fine + 'E501', # leave line length to black + 'N818', # leave to us exceptions naming + 'S101' # assert is fine +] +flake8-quotes = { inline-quotes = 'single', multiline-quotes = 'double' } +mccabe = { max-complexity = 13 } + +[tool.ruff.isort] +combine-as-imports = true +lines-after-imports = 2 +known-first-party = ['granian', 'tests'] + +[tool.ruff.per-file-ignores] +'granian/_granian.pyi' = ['I001'] +'tests/' = ['B018', 'S110', 'S501'] + +[tool.black] +color = true +line-length = 120 +target-version = ['py38', 'py39', 'py310', 'py311'] +skip-string-normalization = true # leave this to ruff +skip-magic-trailing-comma = true # leave this to ruff [tool.pytest.ini_options] -asyncio_mode = "auto" +asyncio_mode = 'auto' diff --git a/src/asgi/callbacks.rs b/src/asgi/callbacks.rs index 5af7766f..96c100fb 100644 --- a/src/asgi/callbacks.rs +++ b/src/asgi/callbacks.rs @@ -3,45 +3,33 @@ use pyo3::prelude::; use pyo3_asyncio::TaskLocals; use tokio::sync::oneshot; +use super::{ + io::{ASGIHTTPProtocol as HTTPProtocol, ASGIWebsocketProtocol as WebsocketProtocol}, + types::ASGIScope as Scope, +}; use crate::{ callbacks::{ - CallbackWrapper, - callback_impl_run, - callback_impl_run_pytask, - callback_impl_loop_run, - callback_impl_loop_pytask, - callback_impl_loop_step, - callback_impl_loop_wake, - callback_impl_loop_err + callback_impl_loop_err, callback_impl_loop_pytask, callback_impl_loop_run, callback_impl_loop_step, + callback_impl_loop_wake, callback_impl_run, callback_impl_run_pytask, CallbackWrapper, }, runtime::RuntimeRef, - ws::{HyperWebsocket, UpgradeData} + ws::{HyperWebsocket, UpgradeData}, }; -use super::{ - io::{ASGIHTTPProtocol as HTTPProtocol, ASGIWebsocketProtocol as WebsocketProtocol}, - types::ASGIScope as Scope -}; - #[pyclass] pub(crate) struct CallbackRunnerHTTP { proto: Py<HTTPProtocol>, context: TaskLocals, - cb: PyObject + cb: PyObject, } impl CallbackRunnerHTTP { - pub fn new( - py: Python, - cb: CallbackWrapper, - proto: HTTPProtocol, - scope: Scope - ) -> Self { + pub fn new(py: Python, cb: CallbackWrapper, proto: HTTPProtocol, scope: Scope) -> Self { let pyproto = Py::new(py, proto).unwrap(); Self { proto: pyproto.clone(), context: cb.context, - cb: cb.callback.call1(py, (scope, pyproto)).unwrap() + cb: cb.callback.call1(py, (scope, pyproto)).unwrap(), } } @@ -64,14 +52,14 @@ macro_rules! callback_impl_done_http { let _ = tx.send(res); } } - } + }; } macro_rules! callback_impl_done_err { ($self:expr, $py:expr) => { log::warn!("Application callable raised an exception"); $self.done($py) - } + }; } #[pyclass] @@ -79,22 +67,17 @@ pub(crate) struct CallbackTaskHTTP { proto: Py<HTTPProtocol>, context: TaskLocals, pycontext: PyObject, - cb: PyObject + cb: PyObject, } impl CallbackTaskHTTP { - pub fn new( - py: Python, - cb: PyObject, - proto: Py<HTTPProtocol>, - context: TaskLocals - ) -> PyResult<Self> { + pub fn new(py: Python, cb: PyObject, proto: Py<HTTPProtocol>, context: TaskLocals) -> PyResult<Self> { let pyctx = context.context(py); Ok(Self { proto, context, pycontext: pyctx.call_method0(pyo3::intern!(py, "copy"))?.into(), - cb + cb, }) } @@ -128,21 +111,16 @@ pub(crate) struct CallbackWrappedRunnerHTTP { context: TaskLocals, cb: PyObject, #[pyo3(get)] - scope: PyObject + scope: PyObject, } impl CallbackWrappedRunnerHTTP { - pub fn new( - py: Python, - cb: CallbackWrapper, - proto: HTTPProtocol, - scope: Scope - ) -> Self { + pub fn new(py: Python, cb: CallbackWrapper, proto: HTTPProtocol, scope: Scope) -> Self { Self { proto: Py::new(py, proto).unwrap(), context: cb.context, cb: cb.callback, - scope: scope.into_py(py) + scope: scope.into_py(py), } } @@ -168,21 +146,16 @@ impl CallbackWrappedRunnerHTTP { pub(crate) struct CallbackRunnerWebsocket { proto: Py<WebsocketProtocol>, context: TaskLocals, - cb: PyObject + cb: PyObject, } impl CallbackRunnerWebsocket { - pub fn new( - py: Python, - cb: CallbackWrapper, - proto: WebsocketProtocol, - scope: Scope - ) -> Self { + pub fn new(py: Python, cb: CallbackWrapper, proto: WebsocketProtocol, scope: Scope) -> Self { let pyproto = Py::new(py, proto).unwrap(); Self { proto: pyproto.clone(), context: cb.context, - cb: cb.callback.call1(py, (scope, pyproto)).unwrap() + cb: cb.callback.call1(py, (scope, pyproto)).unwrap(), } } @@ -203,7 +176,7 @@ macro_rules! callback_impl_done_ws { let _ = tx.send(res); } } - } + }; } #[pyclass] @@ -211,22 +184,17 @@ pub(crate) struct CallbackTaskWebsocket { proto: Py<WebsocketProtocol>, context: TaskLocals, pycontext: PyObject, - cb: PyObject + cb: PyObject, } impl CallbackTaskWebsocket { - pub fn new( - py: Python, - cb: PyObject, - proto: Py<WebsocketProtocol>, - context: TaskLocals - ) -> PyResult<Self> { + pub fn new(py: Python, cb: PyObject, proto: Py<WebsocketProtocol>, context: TaskLocals) -> PyResult<Self> { let pyctx = context.context(py); Ok(Self { proto, context, pycontext: pyctx.call_method0(pyo3::intern!(py, "copy"))?.into(), - cb + cb, }) } @@ -260,21 +228,16 @@ pub(crate) struct CallbackWrappedRunnerWebsocket { context: TaskLocals, cb: PyObject, #[pyo3(get)] - scope: PyObject + scope: PyObject, } impl CallbackWrappedRunnerWebsocket { - pub fn new( - py: Python, - cb: CallbackWrapper, - proto: WebsocketProtocol, - scope: Scope - ) -> Self { + pub fn new(py: Python, cb: CallbackWrapper, proto: WebsocketProtocol, scope: Scope) -> Self { Self { proto: Py::new(py, proto).unwrap(), context: cb.context, cb: cb.callback, - scope: scope.into_py(py) + scope: scope.into_py(py), } } @@ -325,7 +288,7 @@ macro_rules! call_impl_rtb_http { cb: CallbackWrapper, rt: RuntimeRef, req: Request<Body>, - scope: Scope + scope: Scope, ) -> oneshot::Receiver<Response<Body>> { let (tx, rx) = oneshot::channel(); let protocol = HTTPProtocol::new(rt, req, tx); @@ -345,7 +308,7 @@ macro_rules! call_impl_rtt_http { cb: CallbackWrapper, rt: RuntimeRef, req: Request<Body>, - scope: Scope + scope: Scope, ) -> oneshot::Receiver<Response<Body>> { let (tx, rx) = oneshot::channel(); let protocol = HTTPProtocol::new(rt, req, tx); @@ -368,7 +331,7 @@ macro_rules! call_impl_rtb_ws { rt: RuntimeRef, ws: HyperWebsocket, upgrade: UpgradeData, - scope: Scope + scope: Scope, ) -> oneshot::Receiver<bool> { let (tx, rx) = oneshot::channel(); let protocol = WebsocketProtocol::new(rt, tx, ws, upgrade); @@ -389,7 +352,7 @@ macro_rules! call_impl_rtt_ws { rt: RuntimeRef, ws: HyperWebsocket, upgrade: UpgradeData, - scope: Scope + scope: Scope, ) -> oneshot::Receiver<bool> { let (tx, rx) = oneshot::channel(); let protocol = WebsocketProtocol::new(rt, tx, ws, upgrade); diff --git a/src/asgi/errors.rs b/src/asgi/errors.rs index 6b788f60..8a6dda40 100644 --- a/src/asgi/errors.rs +++ b/src/asgi/errors.rs @@ -88,5 +88,5 @@ macro_rules! error_message { } pub(crate) use error_flow; -pub(crate) use error_transport; pub(crate) use error_message; +pub(crate) use error_transport; diff --git a/src/asgi/http.rs b/src/asgi/http.rs index cbf295f7..1fa62773 100644 --- a/src/asgi/http.rs +++ b/src/asgi/http.rs @@ -1,34 +1,22 @@ use hyper::{ - Body, - Request, - Response, - StatusCode, - header::SERVER as HK_SERVER, - http::response::Builder as ResponseBuilder + header::SERVER as HK_SERVER, http::response::Builder as ResponseBuilder, Body, Request, Response, StatusCode, }; use std::net::SocketAddr; use tokio::sync::mpsc; -use crate::{ - callbacks::CallbackWrapper, - http::{HV_SERVER, response_500}, - runtime::RuntimeRef, - ws::{UpgradeData, is_upgrade_request as is_ws_upgrade, upgrade_intent as ws_upgrade} -}; use super::{ callbacks::{ - call_rtb_http, - call_rtb_http_pyw, - call_rtb_ws, - call_rtb_ws_pyw, - call_rtt_http, - call_rtt_http_pyw, - call_rtt_ws, - call_rtt_ws_pyw + call_rtb_http, call_rtb_http_pyw, call_rtb_ws, call_rtb_ws_pyw, call_rtt_http, call_rtt_http_pyw, call_rtt_ws, + call_rtt_ws_pyw, }, - types::ASGIScope as Scope + types::ASGIScope as Scope, +}; +use crate::{ + callbacks::CallbackWrapper, + http::{response_500, HV_SERVER}, + runtime::RuntimeRef, + ws::{is_upgrade_request as is_ws_upgrade, upgrade_intent as ws_upgrade, UpgradeData}, }; - macro_rules! default_scope { ($server_addr:expr, $client_addr:expr, $req:expr, $scheme:expr) => { @@ -39,7 +27,7 @@ macro_rules! default_scope { $req.method().as_ref(), $server_addr, $client_addr, - $req.headers() + $req.headers(), ) }; } @@ -53,7 +41,7 @@ macro_rules! handle_http_response { response_500() } } - } + }; } macro_rules! handle_request { @@ -64,7 +52,7 @@ macro_rules! handle_request { server_addr: SocketAddr, client_addr: SocketAddr, req: Request<Body>, - scheme: &str + scheme: &str, ) -> Response<Body> { let scope = default_scope!(server_addr, client_addr, &req, scheme); handle_http_response!($handler, rt, callback, req, scope) @@ -80,7 +68,7 @@ macro_rules! handle_request_with_ws { server_addr: SocketAddr, client_addr: SocketAddr, req: Request<Body>, - scheme: &str + scheme: &str, ) -> Response<Body> { let mut scope = default_scope!(server_addr, client_addr, &req, scheme); @@ -95,24 +83,20 @@ macro_rules! handle_request_with_ws { rt.inner.spawn(async move { let tx_ref = restx.clone(); - match $handler_ws( - callback, - rth, - ws, - UpgradeData::new(res, restx), - scope - ).await { + match $handler_ws(callback, rth, ws, UpgradeData::new(res, restx), scope).await { Ok(consumed) => { if !consumed { - let _ = tx_ref.send( - ResponseBuilder::new() - .status(StatusCode::FORBIDDEN) - .header(HK_SERVER, HV_SERVER) - .body(Body::from("")) - .unwrap() - ).await; + let _ = tx_ref + .send( + ResponseBuilder::new() + .status(StatusCode::FORBIDDEN) + .header(HK_SERVER, HV_SERVER) + .body(Body::from("")) + .unwrap(), + ) + .await; }; - }, + } _ => { log::error!("ASGI protocol failure"); let _ = tx_ref.send(response_500()).await; @@ -124,10 +108,10 @@ macro_rules! handle_request_with_ws { Some(res) => { resrx.close(); res - }, - _ => response_500() + } + _ => response_500(), } - }, + } Err(err) => { return ResponseBuilder::new() .status(StatusCode::BAD_REQUEST) diff --git a/src/asgi/io.rs b/src/asgi/io.rs index 4b2f25eb..d206aaf2 100644 --- a/src/asgi/io.rs +++ b/src/asgi/io.rs @@ -1,33 +1,34 @@ use bytes::Bytes; -use futures::{sink::SinkExt, stream::{SplitSink, SplitStream, StreamExt}}; +use futures::{ + sink::SinkExt, + stream::{SplitSink, SplitStream, StreamExt}, +}; use hyper::{ - Request, - Response, body::{Body, HttpBody, Sender as BodySender}, - header::{HeaderName, HeaderValue, HeaderMap, SERVER as HK_SERVER} + header::{HeaderMap, HeaderName, HeaderValue, SERVER as HK_SERVER}, + Request, Response, }; use pyo3::prelude::; use pyo3::types::{PyBytes, PyDict}; use std::sync::Arc; +use tokio::sync::{oneshot, Mutex}; use tokio_tungstenite::WebSocketStream; -use tokio::sync::{Mutex, oneshot}; use tungstenite::Message; +use super::{ + errors::{error_flow, error_message, error_transport, UnsupportedASGIMessage}, + types::ASGIMessageType, +}; use crate::{ http::HV_SERVER, - runtime::{RuntimeRef, future_into_py_iter, future_into_py_futlike}, - ws::{HyperWebsocket, UpgradeData} + runtime::{future_into_py_futlike, future_into_py_iter, RuntimeRef}, + ws::{HyperWebsocket, UpgradeData}, }; -use super::{ - errors::{UnsupportedASGIMessage, error_flow, error_transport, error_message}, - types::ASGIMessageType -}; - const EMPTY_BYTES: Vec<u8> = Vec::new(); const EMPTY_STRING: String = String::new(); -#[pyclass(module="granian._granian")] +#[pyclass(module = "granian._granian")] pub(crate) struct ASGIHTTPProtocol { rt: RuntimeRef, tx: Option<oneshot::Sender<Response<Body>>>, @@ -36,15 +37,11 @@ pub(crate) struct ASGIHTTPProtocol { response_chunked: bool, response_status: Option<i16>, response_headers: Option<HeaderMap>, - body_tx: Option<Arc<Mutex<BodySender>>> + body_tx: Option<Arc<Mutex<BodySender>>>, } impl ASGIHTTPProtocol { - pub fn new( - rt: RuntimeRef, - request: Request<Body>, - tx: oneshot::Sender<Response<Body>> - ) -> Self { + pub fn new(rt: RuntimeRef, request: Request<Body>, tx: oneshot::Sender<Response<Body>>) -> Self { Self { rt, tx: Some(tx), @@ -53,7 +50,7 @@ impl ASGIHTTPProtocol { response_chunked: false, response_status: None, response_headers: None, - body_tx: None + body_tx: None, } } @@ -71,7 +68,7 @@ impl ASGIHTTPProtocol { fn send_body<'p>(&self, py: Python<'p>, tx: Arc<Mutex<BodySender>>, body: Vec<u8>) -> PyResult<&'p PyAny> { future_into_py_futlike(self.rt.clone(), py, async move { let mut tx = tx.lock().await; - match (&mut tx).send_data(body.into()).await { + match (tx).send_data(body.into()).await { Ok(_) => Ok(()), Err(err) => { log::warn!("ASGI transport tx error: {:?}", err); @@ -94,18 +91,18 @@ impl ASGIHTTPProtocol { let mut bodym = body_ref.lock().await; let body = &mut bodym; let mut more_body = false; - let chunk = body.data().await.map_or_else(\|\| Bytes::new(), \|buf\| { - buf.map_or_else(\|_\| Bytes::new(), \|buf\| { - more_body = !body.is_end_stream(); - buf - }) + let chunk = body.data().await.map_or_else(Bytes::new, \|buf\| { + buf.map_or_else( + \|_\| Bytes::new(), + \|buf\| { + more_body = !body.is_end_stream(); + buf + }, + ) }); Python::with_gil(\|py\| { let dict = PyDict::new(py); - dict.set_item( - pyo3::intern!(py, "type"), - pyo3::intern!(py, "http.request") - )?; + dict.set_item(pyo3::intern!(py, "type"), pyo3::intern!(py, "http.request"))?; dict.set_item(pyo3::intern!(py, "body"), PyBytes::new(py, &chunk[..]))?; dict.set_item(pyo3::intern!(py, "more_body"), more_body)?; Ok(dict.to_object(py)) @@ -115,16 +112,14 @@ impl ASGIHTTPProtocol { fn send<'p>(&mut self, py: Python<'p>, asyncw: &'p PyAny, data: &'p PyDict) -> PyResult<&'p PyAny> { match adapt_message_type(data) { - Ok(ASGIMessageType::HTTPStart) => { - match self.response_started { - false => { - self.response_status = Some(adapt_status_code(data)?); - self.response_headers = Some(adapt_headers(data)); - self.response_started = true; - asyncw.call0() - }, - true => error_flow!() + Ok(ASGIMessageType::HTTPStart) => match self.response_started { + false => { + self.response_status = Some(adapt_status_code(data)?); + self.response_headers = Some(adapt_headers(data)); + self.response_started = true; + asyncw.call0() } + true => error_flow!(), }, Ok(ASGIMessageType::HTTPBody) => { let (body, more) = adapt_body(data); @@ -133,7 +128,7 @@ impl ASGIHTTPProtocol { let headers = self.response_headers.take().unwrap(); self.send_response(self.response_status.unwrap(), headers, body.into()); asyncw.call0() - }, + } (true, true, false) => { self.response_chunked = true; let headers = self.response_headers.take().unwrap(); @@ -142,37 +137,31 @@ impl ASGIHTTPProtocol { self.body_tx = Some(tx.clone()); self.send_response(self.response_status.unwrap(), headers, body_stream); self.send_body(py, tx, body) - }, - (true, true, true) => { - match self.body_tx.as_mut() { - Some(tx) => { - let tx = tx.clone(); - self.send_body(py, tx, body) - }, - _ => error_flow!() + } + (true, true, true) => match self.body_tx.as_mut() { + Some(tx) => { + let tx = tx.clone(); + self.send_body(py, tx, body) } + _ => error_flow!(), }, - (true, false, true) => { - match self.body_tx.take() { - Some(tx) => { - match body.is_empty() { - false => self.send_body(py, tx, body), - true => asyncw.call0() - } - }, - _ => error_flow!() - } + (true, false, true) => match self.body_tx.take() { + Some(tx) => match body.is_empty() { + false => self.send_body(py, tx, body), + true => asyncw.call0(), + }, + _ => error_flow!(), }, - _ => error_flow!() + _ => error_flow!(), } - }, + } Err(err) => Err(err.into()), - _ => error_message!() + _ => error_message!(), } } } -#[pyclass(module="granian._granian")] +#[pyclass(module = "granian._granian")] pub(crate) struct ASGIWebsocketProtocol { rt: RuntimeRef, tx: Option<oneshot::Sender<bool>>, @@ -181,16 +170,11 @@ pub(crate) struct ASGIWebsocketProtocol { ws_tx: Arc<Mutex<Option<SplitSink<WebSocketStream<hyper::upgrade::Upgraded>, Message>>>>, ws_rx: Arc<Mutex<Option<SplitStream<WebSocketStream<hyper::upgrade::Upgraded>>>>>, accepted: Arc<Mutex<bool>>, - closed: bool + closed: bool, } impl ASGIWebsocketProtocol { - pub fn new( - rt: RuntimeRef, - tx: oneshot::Sender<bool>, - websocket: HyperWebsocket, - upgrade: UpgradeData - ) -> Self { + pub fn new(rt: RuntimeRef, tx: oneshot::Sender<bool>, websocket: HyperWebsocket, upgrade: UpgradeData) -> Self { Self { rt, tx: Some(tx), @@ -199,7 +183,7 @@ impl ASGIWebsocketProtocol { ws_tx: Arc::new(Mutex::new(None)), ws_rx: Arc::new(Mutex::new(None)), accepted: Arc::new(Mutex::new(false)), - closed: false + closed: false, } } @@ -211,7 +195,7 @@ impl ASGIWebsocketProtocol { let tx = self.ws_tx.clone(); let rx = self.ws_rx.clone(); future_into_py_iter(self.rt.clone(), py, async move { - if let Ok(_) = upgrade.send().await { + if (upgrade.send().await).is_ok() { if let Ok(stream) = websocket.await { let mut wtx = tx.lock().await; let mut wrx = rx.lock().await; @@ -220,7 +204,7 @@ impl ASGIWebsocketProtocol { wtx = Some(tx); wrx = Some(rx); accepted = true; - return Ok(()) + return Ok(()); } } error_flow!() @@ -228,18 +212,14 @@ impl ASGIWebsocketProtocol { } #[inline(always)] - fn send_message<'p>( - &self, - py: Python<'p>, - data: &'p PyDict - ) -> PyResult<&'p PyAny> { + fn send_message<'p>(&self, py: Python<'p>, data: &'p PyDict) -> PyResult<&'p PyAny> { let transport = self.ws_tx.clone(); let message = ws_message_into_rs(data); future_into_py_iter(self.rt.clone(), py, async move { if let Ok(message) = message { if let Some(ws) = &mut (transport.lock().await) { - if let Ok(_) = ws.send(message).await { - return Ok(()) + if (ws.send(message).await).is_ok() { + return Ok(()); } }; }; @@ -253,8 +233,8 @@ impl ASGIWebsocketProtocol { let transport = self.ws_tx.clone(); future_into_py_iter(self.rt.clone(), py, async move { if let Some(ws) = &mut (transport.lock().await) { - if let Ok(_) = ws.close().await { - return Ok(()) + if (ws.close().await).is_ok() { + return Ok(()); } }; error_flow!() @@ -262,10 +242,7 @@ impl ASGIWebsocketProtocol { } fn consumed(&self) -> bool { - match &self.upgrade { - Some(_) => false, - _ => true - } + self.upgrade.is_none() } pub fn tx(&mut self) -> (Option<oneshot::Sender<bool>>, bool) { @@ -278,44 +255,26 @@ impl ASGIWebsocketProtocol { fn receive<'p>(&mut self, py: Python<'p>) -> PyResult<&'p PyAny> { let transport = self.ws_rx.clone(); let accepted = self.accepted.clone(); - let closed = self.closed.clone(); + let closed = self.closed; future_into_py_futlike(self.rt.clone(), py, async move { let accepted = accepted.lock().await; match (accepted, closed) { (false, false) => { return Python::with_gil(\|py\| { let dict = PyDict::new(py); - dict.set_item( - pyo3::intern!(py, "type"), - pyo3::intern!(py, "websocket.connect") - )?; + dict.set_item(pyo3::intern!(py, "type"), pyo3::intern!(py, "websocket.connect"))?; Ok(dict.to_object(py)) }) - }, - (true, false) => {}, - _ => { - return error_flow!() } + (true, false) => {} + _ => return error_flow!(), } if let Some(ws) = &mut (transport.lock().await) { - loop { - match ws.next().await { - Some(recv) => { - match recv { - Ok(Message::Ping(_)) => { - continue - }, - Ok(message) => { - return ws_message_into_py(message) - }, - _ => { - break - } - } - }, - _ => { - break - } + while let Some(recv) = ws.next().await { + match recv { + Ok(Message::Ping(_)) => continue, + Ok(message) => return ws_message_into_py(message), + _ => break, } } } @@ -325,25 +284,17 @@ impl ASGIWebsocketProtocol { fn send<'p>(&mut self, py: Python<'p>, _asyncw: &'p PyAny, data: &'p PyDict) -> PyResult<&'p PyAny> { match (adapt_message_type(data), self.closed) { - (Ok(ASGIMessageType::WSAccept), _) => { - self.accept(py) - }, - (Ok(ASGIMessageType::WSClose), false) => { - self.close(py) - }, - (Ok(ASGIMessageType::WSMessage), false) => { - self.send_message(py, data) - }, + (Ok(ASGIMessageType::WSAccept), _) => self.accept(py), + (Ok(ASGIMessageType::WSClose), false) => self.close(py), + (Ok(ASGIMessageType::WSMessage), false) => self.send_message(py, data), (Err(err), _) => Err(err.into()), - _ => error_message!() + _ => error_message!(), } } } #[inline(never)] -fn adapt_message_type( - message: &PyDict -) -> Result<ASGIMessageType, UnsupportedASGIMessage> { +fn adapt_message_type(message: &PyDict) -> Result<ASGIMessageType, UnsupportedASGIMessage> { match message.get_item("type") { Some(item) => { let message_type: &str = item.extract()?; @@ -353,20 +304,18 @@ fn adapt_message_type( "websocket.accept" => Ok(ASGIMessageType::WSAccept), "websocket.close" => Ok(ASGIMessageType::WSClose), "websocket.send" => Ok(ASGIMessageType::WSMessage), - _ => error_message!() + _ => error_message!(), } - }, - _ => error_message!() + } + _ => error_message!(), } } #[inline(always)] fn adapt_status_code(message: &PyDict) -> Result<i16, UnsupportedASGIMessage> { match message.get_item("status") { - Some(item) => { - Ok(item.extract()?) - }, - _ => error_message!() + Some(item) => Ok(item.extract()?), + _ => error_message!(), } } @@ -377,34 +326,26 @@ fn adapt_headers(message: &PyDict) -> HeaderMap { match message.get_item("headers") { Some(item) => { let accum: Vec<Vec<&[u8]>> = item.extract().unwrap_or(Vec::new()); - for tup in accum.iter() { - match ( - HeaderName::from_bytes(tup[0]), - HeaderValue::from_bytes(tup[1]) - ) { - (Ok(key), Ok(val)) => { ret.append(key, val); }, - _ => {} + for tup in &accum { + if let (Ok(key), Ok(val)) = (HeaderName::from_bytes(tup[0]), HeaderValue::from_bytes(tup[1])) { + ret.append(key, val); } - }; + } ret - }, - _ => ret + } + _ => ret, } } #[inline(always)] fn adapt_body(message: &PyDict) -> (Vec<u8>, bool) { let body = match message.get_item("body") { - Some(item) => { - item.extract().unwrap_or(EMPTY_BYTES) - }, - _ => EMPTY_BYTES + Some(item) => item.extract().unwrap_or(EMPTY_BYTES), + _ => EMPTY_BYTES, }; let more = match message.get_item("more_body") { - Some(item) => { - item.extract().unwrap_or(false) - }, - _ => false + Some(item) => item.extract().unwrap_or(false), + _ => false, }; (body, more) } @@ -412,22 +353,12 @@ fn adapt_body(message: &PyDict) -> (Vec<u8>, bool) { #[inline(always)] fn ws_message_into_rs(message: &PyDict) -> PyResult<Message> { match (message.get_item("bytes"), message.get_item("text")) { - (Some(item), None) => { - Ok(Message::Binary(item.extract().unwrap_or(EMPTY_BYTES))) - }, - (None, Some(item)) => { - Ok(Message::Text(item.extract().unwrap_or(EMPTY_STRING))) - }, - (Some(itemb), Some(itemt)) => { - match (itemb.extract().unwrap_or(None), itemt.extract().unwrap_or(None)) { - (Some(msgb), None) => { - Ok(Message::Binary(msgb)) - }, - (None, Some(msgt)) => { - Ok(Message::Text(msgt)) - }, - _ => error_flow!() - } + (Some(item), None) => Ok(Message::Binary(item.extract().unwrap_or(EMPTY_BYTES))), + (None, Some(item)) => Ok(Message::Text(item.extract().unwrap_or(EMPTY_STRING))), + (Some(itemb), Some(itemt)) => match (itemb.extract().unwrap_or(None), itemt.extract().unwrap_or(None)) { + (Some(msgb), None) => Ok(Message::Binary(msgb)), + (None, Some(msgt)) => Ok(Message::Text(msgt)), + _ => error_flow!(), }, _ => { error_flow!() @@ -438,41 +369,23 @@ fn ws_message_into_rs(message: &PyDict) -> PyResult<Message> { #[inline(always)] fn ws_message_into_py(message: Message) -> PyResult<PyObject> { match message { - Message::Binary(message) => { - Python::with_gil(\|py\| { - let dict = PyDict::new(py); - dict.set_item( - pyo3::intern!(py, "type"), - pyo3::intern!(py, "websocket.receive") - )?; - dict.set_item( - pyo3::intern!(py, "bytes"), - PyBytes::new(py, &message[..]) - )?; - Ok(dict.to_object(py)) - }) - }, - Message::Text(message) => { - Python::with_gil(\|py\| { - let dict = PyDict::new(py); - dict.set_item( - pyo3::intern!(py, "type"), - pyo3::intern!(py, "websocket.receive") - )?; - dict.set_item(pyo3::intern!(py, "text"), message)?; - Ok(dict.to_object(py)) - }) - }, - Message::Close(_) => { - Python::with_gil(\|py\| { - let dict = PyDict::new(py); - dict.set_item( - pyo3::intern!(py, "type"), - pyo3::intern!(py, "websocket.disconnect") - )?; - Ok(dict.to_object(py)) - }) - }, + Message::Binary(message) => Python::with_gil(\|py\| { + let dict = PyDict::new(py); + dict.set_item(pyo3::intern!(py, "type"), pyo3::intern!(py, "websocket.receive"))?; + dict.set_item(pyo3::intern!(py, "bytes"), PyBytes::new(py, &message[..]))?; + Ok(dict.to_object(py)) + }), + Message::Text(message) => Python::with_gil(\|py\| { + let dict = PyDict::new(py); + dict.set_item(pyo3::intern!(py, "type"), pyo3::intern!(py, "websocket.receive"))?; + dict.set_item(pyo3::intern!(py, "text"), message)?; + Ok(dict.to_object(py)) + }), + Message::Close(_) => Python::with_gil(\|py\| { + let dict = PyDict::new(py); + dict.set_item(pyo3::intern!(py, "type"), pyo3::intern!(py, "websocket.disconnect"))?; + Ok(dict.to_object(py)) + }), v => { log::warn!("Unsupported websocket message received {:?}", v); error_flow!() diff --git a/src/asgi/serve.rs b/src/asgi/serve.rs index a30f15c3..6243cdc6 100644 --- a/src/asgi/serve.rs +++ b/src/asgi/serve.rs @@ -1,29 +1,14 @@ use pyo3::prelude::; -use crate::{ - workers::{ - WorkerConfig, - serve_rth, - serve_wth, - serve_rth_ssl, - serve_wth_ssl - } -}; use super::http::{ - handle_rtb, - handle_rtb_pyw, - handle_rtt, - handle_rtt_pyw, - handle_rtb_ws, - handle_rtb_ws_pyw, - handle_rtt_ws, - handle_rtt_ws_pyw + handle_rtb, handle_rtb_pyw, handle_rtb_ws, handle_rtb_ws_pyw, handle_rtt, handle_rtt_pyw, handle_rtt_ws, + handle_rtt_ws_pyw, }; +use crate::workers::{serve_rth, serve_rth_ssl, serve_wth, serve_wth_ssl, WorkerConfig}; - -#[pyclass(module="granian._granian")] +#[pyclass(module = "granian._granian")] pub struct ASGIWorker { - config: WorkerConfig + config: WorkerConfig, } impl ASGIWorker { @@ -74,7 +59,7 @@ impl ASGIWorker { opt_enabled: bool, ssl_enabled: bool, ssl_cert: Option<&str>, - ssl_key: Option<&str> + ssl_key: Option<&str>, ) -> PyResult<Self> { Ok(Self { config: WorkerConfig::new( @@ -88,22 +73,16 @@ impl ASGIWorker { opt_enabled, ssl_enabled, ssl_cert, - ssl_key - ) + ssl_key, + ), }) } - fn serve_rth( - &self, - callback: PyObject, - event_loop: &PyAny, - context: &PyAny, - signal_rx: PyObject - ) { + fn serve_rth(&self, callback: PyObject, event_loop: &PyAny, context: &PyAny, signal_rx: PyObject) { match ( self.config.websockets_enabled, self.config.ssl_enabled, - self.config.opt_enabled + self.config.opt_enabled, ) { (false, false, true) => self._serve_rth(callback, event_loop, context, signal_rx), (false, false, false) => self._serve_rth_pyw(callback, event_loop, context, signal_rx), @@ -112,21 +91,15 @@ impl ASGIWorker { (false, true, true) => self._serve_rth_ssl(callback, event_loop, context, signal_rx), (false, true, false) => self._serve_rth_ssl_pyw(callback, event_loop, context, signal_rx), (true, true, true) => self._serve_rth_ssl_ws(callback, event_loop, context, signal_rx), - (true, true, false) => self._serve_rth_ssl_ws_pyw(callback, event_loop, context, signal_rx) + (true, true, false) => self._serve_rth_ssl_ws_pyw(callback, event_loop, context, signal_rx), } } - fn serve_wth( - &self, - callback: PyObject, - event_loop: &PyAny, - context: &PyAny, - signal_rx: PyObject - ) { + fn serve_wth(&self, callback: PyObject, event_loop: &PyAny, context: &PyAny, signal_rx: PyObject) { match ( self.config.websockets_enabled, self.config.ssl_enabled, - self.config.opt_enabled + self.config.opt_enabled, ) { (false, false, true) => self._serve_wth(callback, event_loop, context, signal_rx), (false, false, false) => self._serve_wth_pyw(callback, event_loop, context, signal_rx), @@ -135,7 +108,7 @@ impl ASGIWorker { (false, true, true) => self._serve_wth_ssl(callback, event_loop, context, signal_rx), (false, true, false) => self._serve_wth_ssl_pyw(callback, event_loop, context, signal_rx), (true, true, true) => self._serve_wth_ssl_ws(callback, event_loop, context, signal_rx), - (true, true, false) => self._serve_wth_ssl_ws_pyw(callback, event_loop, context, signal_rx) + (true, true, false) => self._serve_wth_ssl_ws_pyw(callback, event_loop, context, signal_rx), } } } diff --git a/src/asgi/types.rs b/src/asgi/types.rs index 1841d0e6..8e1056e2 100644 --- a/src/asgi/types.rs +++ b/src/asgi/types.rs @@ -1,10 +1,9 @@ -use hyper::{Uri, Version, header::HeaderMap}; +use hyper::{header::HeaderMap, Uri, Version}; use once_cell::sync::OnceCell; use pyo3::prelude::; use pyo3::types::{PyBytes, PyDict, PyList, PyString}; use std::net::{IpAddr, SocketAddr}; - const SCHEME_HTTPS: &str = "https"; const SCHEME_WS: &str = "ws"; const SCHEME_WSS: &str = "wss"; @@ -17,10 +16,10 @@ pub(crate) enum ASGIMessageType { HTTPBody, WSAccept, WSClose, - WSMessage + WSMessage, } -#[pyclass(module="granian._granian")] +#[pyclass(module = "granian._granian")] pub(crate) struct ASGIScope { http_version: Version, scheme: String, @@ -31,7 +30,7 @@ pub(crate) struct ASGIScope { client_ip: IpAddr, client_port: u16, headers: HeaderMap, - is_websocket: bool + is_websocket: bool, } impl ASGIScope { @@ -42,31 +41,31 @@ impl ASGIScope { method: &str, server: SocketAddr, client: SocketAddr, - headers: &HeaderMap + headers: &HeaderMap, ) -> Self { Self { - http_version: http_version, + http_version, scheme: scheme.to_string(), method: method.to_string(), - uri: uri, + uri, server_ip: server.ip(), server_port: server.port(), client_ip: client.ip(), client_port: client.port(), - headers: headers.to_owned(), - is_websocket: false + headers: headers.clone(), + is_websocket: false, } } pub fn set_websocket(&mut self) { - self.is_websocket = true + self.is_websocket = true; } #[inline(always)] fn py_proto(&self) -> &str { match self.is_websocket { false => "http", - true => "websocket" + true => "websocket", } } @@ -76,7 +75,7 @@ impl ASGIScope { Version::HTTP_10 => "1", Version::HTTP_11 => "1.1", Version::HTTP_2 => "2", - _ => "1" + _ => "1", } } @@ -85,22 +84,20 @@ impl ASGIScope { let scheme = &self.scheme[..]; match self.is_websocket { false => scheme, - true => { - match scheme { - SCHEME_HTTPS => SCHEME_WSS, - _ => SCHEME_WS - } - } + true => match scheme { + SCHEME_HTTPS => SCHEME_WSS, + _ => SCHEME_WS, + }, } } #[inline(always)] fn py_headers<'p>(&self, py: Python<'p>) -> PyResult<&'p PyList> { let rv = PyList::empty(py); - for (key, value) in self.headers.iter() { + for (key, value) in &self.headers { rv.append(( PyBytes::new(py, key.as_str().as_bytes()), - PyBytes::new(py, value.as_bytes()) + PyBytes::new(py, value.as_bytes()), ))?; } Ok(rv) @@ -110,17 +107,10 @@ impl ASGIScope { #[pymethods] impl ASGIScope { fn as_dict<'p>(&self, py: Python<'p>, url_path_prefix: &'p str) -> PyResult<&'p PyAny> { - let ( - path, - query_string, - proto, - http_version, - server, - client, - scheme, - method - ) = py.allow_threads(\|\| { - let (path, query_string) = self.uri.path_and_query() + let (path, query_string, proto, http_version, server, client, scheme, method) = py.allow_threads(\|\| { + let (path, query_string) = self + .uri + .path_and_query() .map_or_else(\|\| ("", ""), \|pq\| (pq.path(), pq.query().unwrap_or(""))); ( path, @@ -136,19 +126,23 @@ impl ASGIScope { let dict: &PyDict = PyDict::new(py); dict.set_item( pyo3::intern!(py, "asgi"), - ASGI_VERSION.get_or_try_init(\|\| { - let rv = PyDict::new(py); - rv.set_item("version", "3.0")?; - rv.set_item("spec_version", "2.3")?; - Ok::<PyObject, PyErr>(rv.into()) - })?.as_ref(py) + ASGI_VERSION + .get_or_try_init(\|\| { + let rv = PyDict::new(py); + rv.set_item("version", "3.0")?; + rv.set_item("spec_version", "2.3")?; + Ok::<PyObject, PyErr>(rv.into()) + })? + .as_ref(py), )?; dict.set_item( pyo3::intern!(py, "extensions"), - ASGI_EXTENSIONS.get_or_try_init(\|\| { - let rv = PyDict::new(py); - Ok::<PyObject, PyErr>(rv.into()) - })?.as_ref(py) + ASGI_EXTENSIONS + .get_or_try_init(\|\| { + let rv = PyDict::new(py); + Ok::<PyObject, PyErr>(rv.into()) + })? + .as_ref(py), )?; dict.set_item(pyo3::intern!(py, "type"), proto)?; dict.set_item(pyo3::intern!(py, "http_version"), http_version)?; @@ -160,17 +154,12 @@ impl ASGIScope { dict.set_item(pyo3::intern!(py, "path"), path)?; dict.set_item( pyo3::intern!(py, "raw_path"), - PyString::new(py, path) - .call_method1( - pyo3::intern!(py, "encode"), (pyo3::intern!(py, "ascii"),) - )? + PyString::new(py, path).call_method1(pyo3::intern!(py, "encode"), (pyo3::intern!(py, "ascii"),))?, )?; dict.set_item( pyo3::intern!(py, "query_string"), PyString::new(py, query_string) - .call_method1( - pyo3::intern!(py, "encode"), (pyo3::intern!(py, "latin-1"),) - )? + .call_method1(pyo3::intern!(py, "encode"), (pyo3::intern!(py, "latin-1"),))?, )?; dict.set_item(pyo3::intern!(py, "headers"), self.py_headers(py)?)?; Ok(dict) diff --git a/src/callbacks.rs b/src/callbacks.rs index c6a67ef1..9af137b1 100644 --- a/src/callbacks.rs +++ b/src/callbacks.rs @@ -2,32 +2,27 @@ use once_cell::sync::OnceCell; use pyo3::prelude::; use pyo3::pyclass::IterNextOutput; - static CONTEXTVARS: OnceCell<PyObject> = OnceCell::new(); static CONTEXT: OnceCell<PyObject> = OnceCell::new(); #[derive(Clone)] pub(crate) struct CallbackWrapper { pub callback: PyObject, - pub context: pyo3_asyncio::TaskLocals + pub context: pyo3_asyncio::TaskLocals, } impl CallbackWrapper { - pub(crate) fn new( - callback: PyObject, - event_loop: &PyAny, - context: &PyAny - ) -> Self { + pub(crate) fn new(callback: PyObject, event_loop: &PyAny, context: &PyAny) -> Self { Self { callback, - context: pyo3_asyncio::TaskLocals::new(event_loop).with_context(context) + context: pyo3_asyncio::TaskLocals::new(event_loop).with_context(context), } } } #[pyclass] pub(crate) struct PyIterAwaitable { - result: Option<PyResult<PyObject>> + result: Option<PyResult<PyObject>>, } impl PyIterAwaitable { @@ -36,7 +31,7 @@ impl PyIterAwaitable { } pub(crate) fn set_result(&mut self, result: PyResult<PyObject>) { - self.result = Some(result) + self.result = Some(result); } } @@ -52,13 +47,11 @@ impl PyIterAwaitable { fn __next__(&mut self, py: Python) -> PyResult<IterNextOutput<PyObject, PyObject>> { match self.result.take() { - Some(res) => { - match res { - Ok(v) => Ok(IterNextOutput::Return(v)), - Err(err) => Err(err) - } + Some(res) => match res { + Ok(v) => Ok(IterNextOutput::Return(v)), + Err(err) => Err(err), }, - _ => Ok(IterNextOutput::Yield(py.None())) + _ => Ok(IterNextOutput::Yield(py.None())), } } } @@ -68,7 +61,7 @@ pub(crate) struct PyFutureAwaitable { py_block: bool, event_loop: PyObject, result: Option<PyResult<PyObject>>, - cb: Option<(PyObject, Py<pyo3::types::PyDict>)> + cb: Option<(PyObject, Py<pyo3::types::PyDict>)>, } impl PyFutureAwaitable { @@ -77,7 +70,7 @@ impl PyFutureAwaitable { event_loop, py_block: true, result: None, - cb: None + cb: None, } } @@ -85,12 +78,9 @@ impl PyFutureAwaitable { pyself.result = Some(result); if let Some((cb, ctx)) = pyself.cb.take() { let py = pyself.py(); - let _ = pyself.event_loop.call_method( - py, - "call_soon_threadsafe", - (cb, &pyself), - Some(ctx.as_ref(py)) - ); + let _ = pyself + .event_loop + .call_method(py, "call_soon_threadsafe", (cb, &pyself), Some(ctx.as_ref(py))); } } } @@ -108,25 +98,22 @@ impl PyFutureAwaitable { #[setter(_asyncio_future_blocking)] fn set_block(&mut self, val: bool) { - self.py_block = val + self.py_block = val; } fn get_loop(&mut self) -> PyObject { self.event_loop.clone() } - fn add_done_callback( - mut pyself: PyRefMut<'_, Self>, - py: Python, - cb: PyObject, - context: PyObject - ) -> PyResult<()> { + fn add_done_callback(mut pyself: PyRefMut<'_, Self>, py: Python, cb: PyObject, context: PyObject) -> PyResult<()> { let kwctx = pyo3::types::PyDict::new(py); kwctx.set_item("context", context)?; match pyself.result { Some(_) => { - pyself.event_loop.call_method(py, "call_soon", (cb, &pyself), Some(kwctx))?; - }, + pyself + .event_loop + .call_method(py, "call_soon", (cb, &pyself), Some(kwctx))?; + } _ => { pyself.cb = Some((cb, kwctx.into_py(py))); } @@ -136,9 +123,9 @@ impl PyFutureAwaitable { fn cancel(mut pyself: PyRefMut<'_, Self>, py: Python) -> bool { if let Some((cb, kwctx)) = pyself.cb.take() { - let _ = pyself.event_loop.call_method( - py, "call_soon", (cb, &pyself), Some(kwctx.as_ref(py)) - ); + let _ = pyself + .event_loop + .call_method(py, "call_soon", (cb, &pyself), Some(kwctx.as_ref(py))); } false } @@ -150,79 +137,69 @@ impl PyFutureAwaitable { pyself } - fn __next__( - mut pyself: PyRefMut<'_, Self> - ) -> PyResult<IterNextOutput<PyRefMut<'_, Self>, PyObject>> { + fn __next__(mut pyself: PyRefMut<'_, Self>) -> PyResult<IterNextOutput<PyRefMut<'_, Self>, PyObject>> { match pyself.result { - Some(_) => { - match pyself.result.take().unwrap() { - Ok(v) => Ok(IterNextOutput::Return(v)), - Err(err) => Err(err) - } + Some(_) => match pyself.result.take().unwrap() { + Ok(v) => Ok(IterNextOutput::Return(v)), + Err(err) => Err(err), }, - _ => Ok(IterNextOutput::Yield(pyself)) + _ => Ok(IterNextOutput::Yield(pyself)), } } } fn contextvars(py: Python) -> PyResult<&PyAny> { Ok(CONTEXTVARS - .get_or_try_init(\|\| py.import("contextvars").map(\|m\| m.into()))? + .get_or_try_init(\|\| py.import("contextvars").map(std::convert::Into::into))? .as_ref(py)) } pub fn empty_pycontext(py: Python) -> PyResult<&PyAny> { Ok(CONTEXT - .get_or_try_init(\|\| contextvars(py)?.getattr("Context")?.call0().map(\|c\| c.into()))? + .get_or_try_init(\|\| { + contextvars(py)? + .getattr("Context")? + .call0() + .map(std::convert::Into::into) + })? .as_ref(py)) } macro_rules! callback_impl_run { () => { - pub fn run<'p>(self, py: Python<'p>) -> PyResult<&'p PyAny> { + pub fn run(self, py: Python<'_>) -> PyResult<&PyAny> { let event_loop = self.context.event_loop(py); let target = self.into_py(py).getattr(py, pyo3::intern!(py, "_loop_task"))?; let kwctx = pyo3::types::PyDict::new(py); kwctx.set_item( pyo3::intern!(py, "context"), - crate::callbacks::empty_pycontext(py)? + crate::callbacks::empty_pycontext(py)?, )?; - event_loop.call_method( - pyo3::intern!(py, "call_soon_threadsafe"), - (target,), - Some(kwctx) - ) + event_loop.call_method(pyo3::intern!(py, "call_soon_threadsafe"), (target,), Some(kwctx)) } }; } macro_rules! callback_impl_run_pytask { () => { - pub fn run<'p>(self, py: Python<'p>) -> PyResult<&'p PyAny> { + pub fn run(self, py: Python<'_>) -> PyResult<&PyAny> { let event_loop = self.context.event_loop(py); let context = self.context.context(py); let target = self.into_py(py).getattr(py, pyo3::intern!(py, "_loop_task"))?; let kwctx = pyo3::types::PyDict::new(py); - kwctx.set_item( - pyo3::intern!(py, "context"), - context - )?; - event_loop.call_method( - pyo3::intern!(py, "call_soon_threadsafe"), - (target,), - Some(kwctx) - ) + kwctx.set_item(pyo3::intern!(py, "context"), context)?; + event_loop.call_method(pyo3::intern!(py, "call_soon_threadsafe"), (target,), Some(kwctx)) } }; } macro_rules! callback_impl_loop_run { () => { - pub fn run<'p>(self, py: Python<'p>) -> PyResult<&'p PyAny> { + pub fn run(self, py: Python<'_>) -> PyResult<&PyAny> { let context = self.pycontext.clone().into_ref(py); context.call_method1( pyo3::intern!(py, "run"), - (self.into_py(py).getattr(py, pyo3::intern!(py, "_loop_step"))?,) + (self.into_py(py).getattr(py, pyo3::intern!(py, "_loop_step"))?,), ) } }; @@ -232,7 +209,7 @@ macro_rules! callback_impl_loop_pytask { ($pyself:expr, $py:expr) => { $pyself.context.event_loop($py).call_method1( pyo3::intern!($py, "create_task"), - ($pyself.cb.clone().into_ref($py).call1(($pyself.into_py($py),))?,) + ($pyself.cb.clone().into_ref($py).call1(($pyself.into_py($py),))?,), ) }; } @@ -241,12 +218,9 @@ macro_rules! callback_impl_loop_step { ($pyself:expr, $py:expr) => { match $pyself.cb.call_method1($py, pyo3::intern!($py, "send"), ($py.None(),)) { Ok(res) => { - let blocking: bool = match res.getattr( - $py, - pyo3::intern!($py, "_asyncio_future_blocking") - ) { + let blocking: bool = match res.getattr($py, pyo3::intern!($py, "_asyncio_future_blocking")) { Ok(v) => v.extract($py)?, - _ => false + _ => false, }; let ctx = $pyself.pycontext.clone(); @@ -255,43 +229,30 @@ macro_rules! callback_impl_loop_step { match blocking { true => { - res.setattr( - $py, - pyo3::intern!($py, "_asyncio_future_blocking"), - false - )?; + res.setattr($py, pyo3::intern!($py, "_asyncio_future_blocking"), false)?; res.call_method( $py, pyo3::intern!($py, "add_done_callback"), - ( - $pyself - .into_py($py) - .getattr($py, pyo3::intern!($py, "_loop_wake"))?, - ), - Some(kwctx) + ($pyself.into_py($py).getattr($py, pyo3::intern!($py, "_loop_wake"))?,), + Some(kwctx), )?; Ok(()) - }, + } false => { let event_loop = $pyself.context.event_loop($py); event_loop.call_method( pyo3::intern!($py, "call_soon"), - ( - $pyself - .into_py($py) - .getattr($py, pyo3::intern!($py, "_loop_step"))?, - ), - Some(kwctx) + ($pyself.into_py($py).getattr($py, pyo3::intern!($py, "_loop_step"))?,), + Some(kwctx), )?; Ok(()) } } - }, + } Err(err) => { - if ( - err.is_instance_of::<pyo3::exceptions::PyStopIteration>($py) \|\| - err.is_instance_of::<pyo3::exceptions::asyncio::CancelledError>($py) - ) { + if (err.is_instance_of::<pyo3::exceptions::PyStopIteration>($py) + \|\| err.is_instance_of::<pyo3::exceptions::asyncio::CancelledError>($py)) + { $pyself.done($py); Ok(()) } else { @@ -307,7 +268,7 @@ macro_rules! callback_impl_loop_wake { ($pyself:expr, $py:expr, $fut:expr) => { match $fut.call_method0($py, pyo3::intern!($py, "result")) { Ok(_) => $pyself.into_py($py).call_method0($py, pyo3::intern!($py, "_loop_step")), - Err(err) => $pyself._loop_err($py, err) + Err(err) => $pyself._loop_err($py, err), } }; } @@ -322,10 +283,10 @@ macro_rules! callback_impl_loop_err { }; } -pub(crate) use callback_impl_run; -pub(crate) use callback_impl_run_pytask; -pub(crate) use callback_impl_loop_run; +pub(crate) use callback_impl_loop_err; pub(crate) use callback_impl_loop_pytask; +pub(crate) use callback_impl_loop_run; pub(crate) use callback_impl_loop_step; pub(crate) use callback_impl_loop_wake; -pub(crate) use callback_impl_loop_err; +pub(crate) use callback_impl_run; +pub(crate) use callback_impl_run_pytask; diff --git a/src/http.rs b/src/http.rs index f4ccef32..e0989d7d 100644 --- a/src/http.rs +++ b/src/http.rs @@ -1,4 +1,7 @@ -use hyper::{Body, Response, header::{HeaderValue, SERVER as HK_SERVER}}; +use hyper::{ + header::{HeaderValue, SERVER as HK_SERVER}, + Body, Response, +}; pub(crate) const HV_SERVER: HeaderValue = HeaderValue::from_static("granian"); diff --git a/src/lib.rs b/src/lib.rs index 327f9df1..97cec0f6 100644 --- a/src/lib.rs +++ b/src/lib.rs @@ -8,8 +8,8 @@ mod callbacks; mod http; mod rsgi; mod runtime; -mod tls; mod tcp; +mod tls; mod utils; mod workers; mod ws; diff --git a/src/rsgi/callbacks.rs b/src/rsgi/callbacks.rs index 09872589..184848c5 100644 --- a/src/rsgi/callbacks.rs +++ b/src/rsgi/callbacks.rs @@ -2,45 +2,33 @@ use pyo3::prelude::; use pyo3_asyncio::TaskLocals; use tokio::sync::oneshot; +use super::{ + io::{RSGIHTTPProtocol as HTTPProtocol, RSGIWebsocketProtocol as WebsocketProtocol}, + types::{PyResponse, PyResponseBody, RSGIScope as Scope}, +}; use crate::{ callbacks::{ - CallbackWrapper, - callback_impl_run, - callback_impl_run_pytask, - callback_impl_loop_run, - callback_impl_loop_pytask, - callback_impl_loop_step, - callback_impl_loop_wake, - callback_impl_loop_err + callback_impl_loop_err, callback_impl_loop_pytask, callback_impl_loop_run, callback_impl_loop_step, + callback_impl_loop_wake, callback_impl_run, callback_impl_run_pytask, CallbackWrapper, }, runtime::RuntimeRef, - ws::{HyperWebsocket, UpgradeData} + ws::{HyperWebsocket, UpgradeData}, }; -use super::{ - io::{RSGIHTTPProtocol as HTTPProtocol, RSGIWebsocketProtocol as WebsocketProtocol}, - types::{RSGIScope as Scope, PyResponse, PyResponseBody} -}; - #[pyclass] pub(crate) struct CallbackRunnerHTTP { proto: Py<HTTPProtocol>, context: TaskLocals, - cb: PyObject + cb: PyObject, } impl CallbackRunnerHTTP { - pub fn new( - py: Python, - cb: CallbackWrapper, - proto: HTTPProtocol, - scope: Scope - ) -> Self { + pub fn new(py: Python, cb: CallbackWrapper, proto: HTTPProtocol, scope: Scope) -> Self { let pyproto = Py::new(py, proto).unwrap(); Self { proto: pyproto.clone(), context: cb.context, - cb: cb.callback.call1(py, (scope, pyproto)).unwrap() + cb: cb.callback.call1(py, (scope, pyproto)).unwrap(), } } @@ -58,19 +46,17 @@ macro_rules! callback_impl_done_http { ($self:expr, $py:expr) => { if let Ok(mut proto) = $self.proto.as_ref($py).try_borrow_mut() { if let Some(tx) = proto.tx() { - let _ = tx.send( - PyResponse::Body(PyResponseBody::empty(500, Vec::new())) - ); + let _ = tx.send(PyResponse::Body(PyResponseBody::empty(500, Vec::new()))); } } - } + }; } macro_rules! callback_impl_done_err { ($self:expr, $py:expr) => { log::warn!("Application callable raised an exception"); $self.done($py) - } + }; } #[pyclass] @@ -78,18 +64,18 @@ pub(crate) struct CallbackTaskHTTP { proto: Py<HTTPProtocol>, context: TaskLocals, pycontext: PyObject, - cb: PyObject + cb: PyObject, } impl CallbackTaskHTTP { - pub fn new( - py: Python, - cb: PyObject, - proto: Py<HTTPProtocol>, - context: TaskLocals - ) -> PyResult<Self> { + pub fn new(py: Python, cb: PyObject, proto: Py<HTTPProtocol>, context: TaskLocals) -> PyResult<Self> { let pyctx = context.context(py); - Ok(Self { proto, context, pycontext: pyctx.call_method0(pyo3::intern!(py, "copy"))?.into(), cb }) + Ok(Self { + proto, + context, + pycontext: pyctx.call_method0(pyo3::intern!(py, "copy"))?.into(), + cb, + }) } fn done(&self, py: Python) { @@ -122,21 +108,16 @@ pub(crate) struct CallbackWrappedRunnerHTTP { context: TaskLocals, cb: PyObject, #[pyo3(get)] - scope: PyObject + scope: PyObject, } impl CallbackWrappedRunnerHTTP { - pub fn new( - py: Python, - cb: CallbackWrapper, - proto: HTTPProtocol, - scope: Scope - ) -> Self { + pub fn new(py: Python, cb: CallbackWrapper, proto: HTTPProtocol, scope: Scope) -> Self { Self { proto: Py::new(py, proto).unwrap(), context: cb.context, cb: cb.callback, - scope: scope.into_py(py) + scope: scope.into_py(py), } } @@ -162,21 +143,16 @@ impl CallbackWrappedRunnerHTTP { pub(crate) struct CallbackRunnerWebsocket { proto: Py<WebsocketProtocol>, context: TaskLocals, - cb: PyObject + cb: PyObject, } impl CallbackRunnerWebsocket { - pub fn new( - py: Python, - cb: CallbackWrapper, - proto: WebsocketProtocol, - scope: Scope - ) -> Self { + pub fn new(py: Python, cb: CallbackWrapper, proto: WebsocketProtocol, scope: Scope) -> Self { let pyproto = Py::new(py, proto).unwrap(); Self { proto: pyproto.clone(), context: cb.context, - cb: cb.callback.call1(py, (scope, pyproto)).unwrap() + cb: cb.callback.call1(py, (scope, pyproto)).unwrap(), } } @@ -197,7 +173,7 @@ macro_rules! callback_impl_done_ws { let _ = tx.send(res); } } - } + }; } #[pyclass] @@ -205,18 +181,18 @@ pub(crate) struct CallbackTaskWebsocket { proto: Py<WebsocketProtocol>, context: TaskLocals, pycontext: PyObject, - cb: PyObject + cb: PyObject, } impl CallbackTaskWebsocket { - pub fn new( - py: Python, - cb: PyObject, - proto: Py<WebsocketProtocol>, - context: TaskLocals - ) -> PyResult<Self> { + pub fn new(py: Python, cb: PyObject, proto: Py<WebsocketProtocol>, context: TaskLocals) -> PyResult<Self> { let pyctx = context.context(py); - Ok(Self { proto, context, pycontext: pyctx.call_method0(pyo3::intern!(py, "copy"))?.into(), cb }) + Ok(Self { + proto, + context, + pycontext: pyctx.call_method0(pyo3::intern!(py, "copy"))?.into(), + cb, + }) } fn done(&self, py: Python) { @@ -249,21 +225,16 @@ pub(crate) struct CallbackWrappedRunnerWebsocket { context: TaskLocals, cb: PyObject, #[pyo3(get)] - scope: PyObject + scope: PyObject, } impl CallbackWrappedRunnerWebsocket { - pub fn new( - py: Python, - cb: CallbackWrapper, - proto: WebsocketProtocol, - scope: Scope - ) -> Self { + pub fn new(py: Python, cb: CallbackWrapper, proto: WebsocketProtocol, scope: Scope) -> Self { Self { proto: Py::new(py, proto).unwrap(), context: cb.context, cb: cb.callback, - scope: scope.into_py(py) + scope: scope.into_py(py), } } @@ -291,7 +262,7 @@ macro_rules! call_impl_rtb_http { cb: CallbackWrapper, rt: RuntimeRef, req: hyper::Request<hyper::Body>, - scope: Scope + scope: Scope, ) -> oneshot::Receiver<PyResponse> { let (tx, rx) = oneshot::channel(); let protocol = HTTPProtocol::new(rt, tx, req); @@ -311,7 +282,7 @@ macro_rules! call_impl_rtt_http { cb: CallbackWrapper, rt: RuntimeRef, req: hyper::Request<hyper::Body>, - scope: Scope + scope: Scope, ) -> oneshot::Receiver<PyResponse> { let (tx, rx) = oneshot::channel(); let protocol = HTTPProtocol::new(rt, tx, req); @@ -334,7 +305,7 @@ macro_rules! call_impl_rtb_ws { rt: RuntimeRef, ws: HyperWebsocket, upgrade: UpgradeData, - scope: Scope + scope: Scope, ) -> oneshot::Receiver<(i32, bool)> { let (tx, rx) = oneshot::channel(); let protocol = WebsocketProtocol::new(rt, tx, ws, upgrade); @@ -355,7 +326,7 @@ macro_rules! call_impl_rtt_ws { rt: RuntimeRef, ws: HyperWebsocket, upgrade: UpgradeData, - scope: Scope + scope: Scope, ) -> oneshot::Receiver<(i32, bool)> { let (tx, rx) = oneshot::channel(); let protocol = WebsocketProtocol::new(rt, tx, ws, upgrade); diff --git a/src/rsgi/errors.rs b/src/rsgi/errors.rs index 4f7cbbec..fe27eea4 100644 --- a/src/rsgi/errors.rs +++ b/src/rsgi/errors.rs @@ -1,6 +1,5 @@ use pyo3::{create_exception, exceptions::PyRuntimeError}; - create_exception!(_granian, RSGIProtocolError, PyRuntimeError, "RSGIProtocolError"); create_exception!(_granian, RSGIProtocolClosed, PyRuntimeError, "RSGIProtocolClosed"); diff --git a/src/rsgi/http.rs b/src/rsgi/http.rs index f786be62..76b7d66e 100644 --- a/src/rsgi/http.rs +++ b/src/rsgi/http.rs @@ -1,34 +1,22 @@ use hyper::{ - Body, - Request, - Response, - StatusCode, - header::SERVER as HK_SERVER, - http::response::Builder as ResponseBuilder + header::SERVER as HK_SERVER, http::response::Builder as ResponseBuilder, Body, Request, Response, StatusCode, }; use std::net::SocketAddr; use tokio::sync::mpsc; -use crate::{ - callbacks::CallbackWrapper, - http::{HV_SERVER, response_500}, - runtime::RuntimeRef, - ws::{UpgradeData, is_upgrade_request as is_ws_upgrade, upgrade_intent as ws_upgrade} -}; use super::{ callbacks::{ - call_rtb_http, - call_rtb_http_pyw, - call_rtb_ws, - call_rtb_ws_pyw, - call_rtt_http, - call_rtt_http_pyw, - call_rtt_ws, - call_rtt_ws_pyw + call_rtb_http, call_rtb_http_pyw, call_rtb_ws, call_rtb_ws_pyw, call_rtt_http, call_rtt_http_pyw, call_rtt_ws, + call_rtt_ws_pyw, }, - types::{RSGIScope as Scope, PyResponse} + types::{PyResponse, RSGIScope as Scope}, +}; +use crate::{ + callbacks::CallbackWrapper, + http::{response_500, HV_SERVER}, + runtime::RuntimeRef, + ws::{is_upgrade_request as is_ws_upgrade, upgrade_intent as ws_upgrade, UpgradeData}, }; - macro_rules! default_scope { ($server_addr:expr, $client_addr:expr, $req:expr, $scheme:expr) => { @@ -40,7 +28,7 @@ macro_rules! default_scope { $req.method().as_ref(), $server_addr, $client_addr, - $req.headers() + $req.headers(), ) }; } @@ -48,12 +36,8 @@ macro_rules! default_scope { macro_rules! handle_http_response { ($handler:expr, $rt:expr, $callback:expr, $req:expr, $scope:expr) => { match $handler($callback, $rt, $req, $scope).await { - Ok(PyResponse::Body(pyres)) => { - pyres.to_response() - }, - Ok(PyResponse::File(pyres)) => { - pyres.to_response().await - }, + Ok(PyResponse::Body(pyres)) => pyres.to_response(), + Ok(PyResponse::File(pyres)) => pyres.to_response().await, _ => { log::error!("RSGI protocol failure"); response_500() @@ -70,7 +54,7 @@ macro_rules! handle_request { server_addr: SocketAddr, client_addr: SocketAddr, req: Request<Body>, - scheme: &str + scheme: &str, ) -> Response<Body> { let scope = default_scope!(server_addr, client_addr, &req, scheme); handle_http_response!($handler, rt, callback, req, scope) @@ -86,7 +70,7 @@ macro_rules! handle_request_with_ws { server_addr: SocketAddr, client_addr: SocketAddr, req: Request<Body>, - scheme: &str + scheme: &str, ) -> Response<Body> { let mut scope = default_scope!(server_addr, client_addr, &req, scheme); @@ -101,27 +85,23 @@ macro_rules! handle_request_with_ws { rt.inner.spawn(async move { let tx_ref = restx.clone(); - match $handler_ws( - callback, - rth, - ws, - UpgradeData::new(res, restx), - scope - ).await { + match $handler_ws(callback, rth, ws, UpgradeData::new(res, restx), scope).await { Ok((status, consumed)) => { if !consumed { - let _ = tx_ref.send( - ResponseBuilder::new() - .status( - StatusCode::from_u16(status as u16) - .unwrap_or(StatusCode::FORBIDDEN) - ) - .header(HK_SERVER, HV_SERVER) - .body(Body::from("")) - .unwrap() - ).await; + let _ = tx_ref + .send( + ResponseBuilder::new() + .status( + StatusCode::from_u16(status as u16) + .unwrap_or(StatusCode::FORBIDDEN), + ) + .header(HK_SERVER, HV_SERVER) + .body(Body::from("")) + .unwrap(), + ) + .await; } - }, + } _ => { log::error!("RSGI protocol failure"); let _ = tx_ref.send(response_500()).await; @@ -133,10 +113,10 @@ macro_rules! handle_request_with_ws { Some(res) => { resrx.close(); res - }, - _ => response_500() - } - }, + } + _ => response_500(), + }; + } Err(err) => { return ResponseBuilder::new() .status(StatusCode::BAD_REQUEST) @@ -149,7 +129,6 @@ macro_rules! handle_request_with_ws { handle_http_response!($handler_req, rt, callback, req, scope) } - }; } diff --git a/src/rsgi/io.rs b/src/rsgi/io.rs index 4f061484..5b629907 100644 --- a/src/rsgi/io.rs +++ b/src/rsgi/io.rs @@ -1,35 +1,40 @@ use bytes::Bytes; -use futures::{sink::SinkExt, stream::{SplitSink, SplitStream, StreamExt}}; -use hyper::{body::{Body, Sender as BodySender, HttpBody}, Request}; +use futures::{ + sink::SinkExt, + stream::{SplitSink, SplitStream, StreamExt}, +}; +use hyper::{ + body::{Body, HttpBody, Sender as BodySender}, + Request, +}; use pyo3::prelude::; use pyo3::types::{PyBytes, PyString}; use std::sync::Arc; -use tokio_tungstenite::WebSocketStream; use tokio::sync::{oneshot, Mutex}; +use tokio_tungstenite::WebSocketStream; use tungstenite::Message; -use crate::{ - runtime::{Runtime, RuntimeRef, future_into_py_iter, future_into_py_futlike}, - ws::{HyperWebsocket, UpgradeData} -}; use super::{ errors::{error_proto, error_stream}, - types::{PyResponse, PyResponseBody, PyResponseFile} + types::{PyResponse, PyResponseBody, PyResponseFile}, +}; +use crate::{ + runtime::{future_into_py_futlike, future_into_py_iter, Runtime, RuntimeRef}, + ws::{HyperWebsocket, UpgradeData}, }; - -#[pyclass(module="granian._granian")] +#[pyclass(module = "granian._granian")] pub(crate) struct RSGIHTTPStreamTransport { rt: RuntimeRef, - tx: Arc<Mutex<BodySender>> + tx: Arc<Mutex<BodySender>>, } impl RSGIHTTPStreamTransport { - pub fn new( - rt: RuntimeRef, - transport: BodySender - ) -> Self { - Self { rt: rt, tx: Arc::new(Mutex::new(transport)) } + pub fn new(rt: RuntimeRef, transport: BodySender) -> Self { + Self { + rt, + tx: Arc::new(Mutex::new(transport)), + } } } @@ -41,8 +46,8 @@ impl RSGIHTTPStreamTransport { if let Ok(mut stream) = transport.try_lock() { return match stream.send_data(data.into()).await { Ok(_) => Ok(()), - _ => error_stream!() - } + _ => error_stream!(), + }; } error_proto!() }) @@ -54,31 +59,27 @@ impl RSGIHTTPStreamTransport { if let Ok(mut stream) = transport.try_lock() { return match stream.send_data(data.into()).await { Ok(_) => Ok(()), - _ => error_stream!() - } + _ => error_stream!(), + }; } error_proto!() }) } } -#[pyclass(module="granian._granian")] +#[pyclass(module = "granian._granian")] pub(crate) struct RSGIHTTPProtocol { rt: RuntimeRef, tx: Option<oneshot::Sender<super::types::PyResponse>>, - body: Arc<Mutex<Body>> + body: Arc<Mutex<Body>>, } impl RSGIHTTPProtocol { - pub fn new( - rt: RuntimeRef, - tx: oneshot::Sender<super::types::PyResponse>, - request: Request<Body> - ) -> Self { + pub fn new(rt: RuntimeRef, tx: oneshot::Sender<super::types::PyResponse>, request: Request<Body>) -> Self { Self { rt, tx: Some(tx), - body: Arc::new(Mutex::new(request.into_body())) + body: Arc::new(Mutex::new(request.into_body())), } } @@ -110,14 +111,13 @@ impl RSGIHTTPProtocol { let mut bodym = body_ref.lock().await; let body = &mut bodym; if body.is_end_stream() { - return Err(pyo3::exceptions::PyStopAsyncIteration::new_err("stream exhausted")) + return Err(pyo3::exceptions::PyStopAsyncIteration::new_err("stream exhausted")); } - let chunk = body.data().await.map_or_else(\|\| Bytes::new(), \|buf\| { - buf.unwrap_or_else(\|_\| Bytes::new()) - }); - Ok(Python::with_gil(\|py\| { - PyBytes::new(py, &chunk[..]).to_object(py) - })) + let chunk = body + .data() + .await + .map_or_else(Bytes::new, \|buf\| buf.unwrap_or_else(\|_\| Bytes::new())); + Ok(Python::with_gil(\|py\| PyBytes::new(py, &chunk[..]).to_object(py))) })?; Ok(Some(fut)) } @@ -125,36 +125,28 @@ impl RSGIHTTPProtocol { #[pyo3(signature = (status=200, headers=vec![]))] fn response_empty(&mut self, status: u16, headers: Vec<(String, String)>) { if let Some(tx) = self.tx.take() { - let _ = tx.send( - PyResponse::Body(PyResponseBody::empty(status, headers)) - ); + let _ = tx.send(PyResponse::Body(PyResponseBody::empty(status, headers))); } } #[pyo3(signature = (status=200, headers=vec![], body=vec![]))] fn response_bytes(&mut self, status: u16, headers: Vec<(String, String)>, body: Vec<u8>) { if let Some(tx) = self.tx.take() { - let _ = tx.send( - PyResponse::Body(PyResponseBody::from_bytes(status, headers, body)) - ); + let _ = tx.send(PyResponse::Body(PyResponseBody::from_bytes(status, headers, body))); } } - #[pyo3(signature = (status=200, headers=vec![], body="".to_string()))] + #[pyo3(signature = (status=200, headers=vec![], body=String::new()))] fn response_str(&mut self, status: u16, headers: Vec<(String, String)>, body: String) { if let Some(tx) = self.tx.take() { - let _ = tx.send( - PyResponse::Body(PyResponseBody::from_string(status, headers, body)) - ); + let _ = tx.send(PyResponse::Body(PyResponseBody::from_string(status, headers, body))); } } #[pyo3(signature = (status, headers, file))] fn response_file(&mut self, status: u16, headers: Vec<(String, String)>, file: String) { if let Some(tx) = self.tx.take() { - let _ = tx.send( - PyResponse::File(PyResponseFile::new(status, headers, file)) - ); + let _ = tx.send(PyResponse::File(PyResponseFile::new(status, headers, file))); } } @@ -163,34 +155,33 @@ impl RSGIHTTPProtocol { &mut self, py: Python<'p>, status: u16, - headers: Vec<(String, String)> + headers: Vec<(String, String)>, ) -> PyResult<&'p PyAny> { if let Some(tx) = self.tx.take() { let (body_tx, body_stream) = Body::channel(); - let _ = tx.send( - PyResponse::Body(PyResponseBody::new(status, headers, body_stream)) - ); + let _ = tx.send(PyResponse::Body(PyResponseBody::new(status, headers, body_stream))); let trx = Py::new(py, RSGIHTTPStreamTransport::new(self.rt.clone(), body_tx))?; - return Ok(trx.into_ref(py)) + return Ok(trx.into_ref(py)); } error_proto!() } } -#[pyclass(module="granian._granian")] +#[pyclass(module = "granian._granian")] pub(crate) struct RSGIWebsocketTransport { rt: RuntimeRef, tx: Arc<Mutex<SplitSink<WebSocketStream<hyper::upgrade::Upgraded>, Message>>>, - rx: Arc<Mutex<SplitStream<WebSocketStream<hyper::upgrade::Upgraded>>>> + rx: Arc<Mutex<SplitStream<WebSocketStream<hyper::upgrade::Upgraded>>>>, } impl RSGIWebsocketTransport { - pub fn new( - rt: RuntimeRef, - transport: WebSocketStream<hyper::upgrade::Upgraded> - ) -> Self { + pub fn new(rt: RuntimeRef, transport: WebSocketStream<hyper::upgrade::Upgraded>) -> Self { let (tx, rx) = transport.split(); - Self { rt: rt, tx: Arc::new(Mutex::new(tx)), rx: Arc::new(Mutex::new(rx)) } + Self { + rt, + tx: Arc::new(Mutex::new(tx)), + rx: Arc::new(Mutex::new(rx)), + } } pub fn close(&self) { @@ -209,27 +200,14 @@ impl RSGIWebsocketTransport { let transport = self.rx.clone(); future_into_py_futlike(self.rt.clone(), py, async move { if let Ok(mut stream) = transport.try_lock() { - loop { - match stream.next().await { - Some(recv) => { - match recv { - Ok(Message::Ping(_)) => { - continue - }, - Ok(message) => { - return message_into_py(message) - }, - _ => { - break - } - } - }, - _ => { - break - } + while let Some(recv) = stream.next().await { + match recv { + Ok(Message::Ping(_)) => continue, + Ok(message) => return message_into_py(message), + _ => break, } } - return error_stream!() + return error_stream!(); } error_proto!() }) @@ -241,8 +219,8 @@ impl RSGIWebsocketTransport { if let Ok(mut stream) = transport.try_lock() { return match stream.send(Message::Binary(data)).await { Ok(_) => Ok(()), - _ => error_stream!() - } + _ => error_stream!(), + }; } error_proto!() }) @@ -254,22 +232,22 @@ impl RSGIWebsocketTransport { if let Ok(mut stream) = transport.try_lock() { return match stream.send(Message::Text(data)).await { Ok(_) => Ok(()), - _ => error_stream!() - } + _ => error_stream!(), + }; } error_proto!() }) } } -#[pyclass(module="granian._granian")] +#[pyclass(module = "granian._granian")] pub(crate) struct RSGIWebsocketProtocol { rt: RuntimeRef, tx: Option<oneshot::Sender<(i32, bool)>>, websocket: Arc<Mutex<HyperWebsocket>>, upgrade: Option<UpgradeData>, transport: Arc<Mutex<Option<Py<RSGIWebsocketTransport>>>>, - status: i32 + status: i32, } impl RSGIWebsocketProtocol { @@ -277,7 +255,7 @@ impl RSGIWebsocketProtocol { rt: RuntimeRef, tx: oneshot::Sender<(i32, bool)>, websocket: HyperWebsocket, - upgrade: UpgradeData + upgrade: UpgradeData, ) -> Self { Self { rt, @@ -285,15 +263,12 @@ impl RSGIWebsocketProtocol { websocket: Arc::new(Mutex::new(websocket)), upgrade: Some(upgrade), transport: Arc::new(Mutex::new(None)), - status: 0 + status: 0, } } fn consumed(&self) -> bool { - match &self.upgrade { - Some(_) => false, - _ => true - } + self.upgrade.is_none() } pub fn tx(&mut self) -> (Option<oneshot::Sender<(i32, bool)>>, (i32, bool)) { @@ -304,18 +279,20 @@ impl RSGIWebsocketProtocol { enum WebsocketMessageType { Close = 0, Bytes = 1, - Text = 2 + Text = 2, } #[pyclass] struct WebsocketInboundCloseMessage { #[pyo3(get)] - kind: usize + kind: usize, } impl WebsocketInboundCloseMessage { pub fn new() -> Self { - Self { kind: WebsocketMessageType::Close as usize } + Self { + kind: WebsocketMessageType::Close as usize, + } } } @@ -324,12 +301,15 @@ struct WebsocketInboundBytesMessage { #[pyo3(get)] kind: usize, #[pyo3(get)] - data: Py<PyBytes> + data: Py<PyBytes>, } impl WebsocketInboundBytesMessage { - pub fn new(data:Py<PyBytes>) -> Self { - Self { kind: WebsocketMessageType::Bytes as usize, data: data } + pub fn new(data: Py<PyBytes>) -> Self { + Self { + kind: WebsocketMessageType::Bytes as usize, + data, + } } } @@ -338,12 +318,15 @@ struct WebsocketInboundTextMessage { #[pyo3(get)] kind: usize, #[pyo3(get)] - data: Py<PyString> + data: Py<PyString>, } impl WebsocketInboundTextMessage { pub fn new(data: Py<PyString>) -> Self { - Self { kind: WebsocketMessageType::Text as usize, data: data } + Self { + kind: WebsocketMessageType::Text as usize, + data, + } } } @@ -374,23 +357,18 @@ impl RSGIWebsocketProtocol { future_into_py_iter(self.rt.clone(), py, async move { let mut ws = transport.lock().await; match upgrade.send().await { - Ok(_) => { - match (&mut ws).await { - Ok(stream) => { - let mut trx = itransport.lock().await; - Ok(Python::with_gil(\|py\| { - let pytransport = Py::new( - py, - RSGIWebsocketTransport::new(rth, stream) - ).unwrap(); - trx = Some(pytransport.clone()); - pytransport - })) - }, - _ => error_proto!() + Ok(_) => match (&mut ws).await { + Ok(stream) => { + let mut trx = itransport.lock().await; + Ok(Python::with_gil(\|py\| { + let pytransport = Py::new(py, RSGIWebsocketTransport::new(rth, stream)).unwrap(); + trx = Some(pytransport.clone()); + pytransport + })) } + _ => error_proto!(), }, - _ => error_proto!() + _ => error_proto!(), } }) } @@ -399,25 +377,13 @@ impl RSGIWebsocketProtocol { #[inline(always)] fn message_into_py(message: Message) -> PyResult<PyObject> { match message { - Message::Binary(message) => { - Ok(Python::with_gil(\|py\| { - WebsocketInboundBytesMessage::new( - PyBytes::new(py, &message).into() - ).into_py(py) - })) - }, - Message::Text(message) => { - Ok(Python::with_gil(\|py\| { - WebsocketInboundTextMessage::new( - PyString::new(py, &message).into() - ).into_py(py) - })) - }, - Message::Close(_) => { - Ok(Python::with_gil(\|py\| { - WebsocketInboundCloseMessage::new().into_py(py) - })) - } + Message::Binary(message) => Ok(Python::with_gil(\|py\| { + WebsocketInboundBytesMessage::new(PyBytes::new(py, &message).into()).into_py(py) + })), + Message::Text(message) => Ok(Python::with_gil(\|py\| { + WebsocketInboundTextMessage::new(PyString::new(py, &message).into()).into_py(py) + })), + Message::Close(_) => Ok(Python::with_gil(\|py\| WebsocketInboundCloseMessage::new().into_py(py))), v => { log::warn!("Unsupported websocket message received {:?}", v); error_proto!() diff --git a/src/rsgi/serve.rs b/src/rsgi/serve.rs index f361c095..fec2c5f2 100644 --- a/src/rsgi/serve.rs +++ b/src/rsgi/serve.rs @@ -1,28 +1,14 @@ use pyo3::prelude::; -use crate::{ - workers::{ - WorkerConfig, - serve_rth, - serve_wth, - serve_rth_ssl, - serve_wth_ssl - } -}; use super::http::{ - handle_rtb, - handle_rtb_pyw, - handle_rtt, - handle_rtt_pyw, - handle_rtb_ws, - handle_rtb_ws_pyw, - handle_rtt_ws, - handle_rtt_ws_pyw + handle_rtb, handle_rtb_pyw, handle_rtb_ws, handle_rtb_ws_pyw, handle_rtt, handle_rtt_pyw, handle_rtt_ws, + handle_rtt_ws_pyw, }; +use crate::workers::{serve_rth, serve_rth_ssl, serve_wth, serve_wth_ssl, WorkerConfig}; -#[pyclass(module="granian._granian")] +#[pyclass(module = "granian._granian")] pub struct RSGIWorker { - config: WorkerConfig + config: WorkerConfig, } impl RSGIWorker { @@ -73,7 +59,7 @@ impl RSGIWorker { opt_enabled: bool, ssl_enabled: bool, ssl_cert: Option<&str>, - ssl_key: Option<&str> + ssl_key: Option<&str>, ) -> PyResult<Self> { Ok(Self { config: WorkerConfig::new( @@ -87,22 +73,16 @@ impl RSGIWorker { opt_enabled, ssl_enabled, ssl_cert, - ssl_key - ) + ssl_key, + ), }) } - fn serve_rth( - &self, - callback: PyObject, - event_loop: &PyAny, - context: &PyAny, - signal_rx: PyObject - ) { + fn serve_rth(&self, callback: PyObject, event_loop: &PyAny, context: &PyAny, signal_rx: PyObject) { match ( self.config.websockets_enabled, self.config.ssl_enabled, - self.config.opt_enabled + self.config.opt_enabled, ) { (false, false, true) => self._serve_rth(callback, event_loop, context, signal_rx), (false, false, false) => self._serve_rth_pyw(callback, event_loop, context, signal_rx), @@ -111,21 +91,15 @@ impl RSGIWorker { (false, true, true) => self._serve_rth_ssl(callback, event_loop, context, signal_rx), (false, true, false) => self._serve_rth_ssl_pyw(callback, event_loop, context, signal_rx), (true, true, true) => self._serve_rth_ssl_ws(callback, event_loop, context, signal_rx), - (true, true, false) => self._serve_rth_ssl_ws_pyw(callback, event_loop, context, signal_rx) + (true, true, false) => self._serve_rth_ssl_ws_pyw(callback, event_loop, context, signal_rx), } } - fn serve_wth( - &self, - callback: PyObject, - event_loop: &PyAny, - context: &PyAny, - signal_rx: PyObject - ) { + fn serve_wth(&self, callback: PyObject, event_loop: &PyAny, context: &PyAny, signal_rx: PyObject) { match ( self.config.websockets_enabled, self.config.ssl_enabled, - self.config.opt_enabled + self.config.opt_enabled, ) { (false, false, true) => self._serve_wth(callback, event_loop, context, signal_rx), (false, false, false) => self._serve_wth_pyw(callback, event_loop, context, signal_rx), @@ -134,7 +108,7 @@ impl RSGIWorker { (false, true, true) => self._serve_wth_ssl(callback, event_loop, context, signal_rx), (false, true, false) => self._serve_wth_ssl_pyw(callback, event_loop, context, signal_rx), (true, true, true) => self._serve_wth_ssl_ws(callback, event_loop, context, signal_rx), - (true, true, false) => self._serve_wth_ssl_ws_pyw(callback, event_loop, context, signal_rx) + (true, true, false) => self._serve_wth_ssl_ws_pyw(callback, event_loop, context, signal_rx), } } } diff --git a/src/rsgi/types.rs b/src/rsgi/types.rs index 37123235..66037012 100644 --- a/src/rsgi/types.rs +++ b/src/rsgi/types.rs @@ -1,6 +1,6 @@ use hyper::{ header::{HeaderMap, HeaderName, HeaderValue, SERVER as HK_SERVER}, - Body, Uri, Version + Body, Uri, Version, }; use pyo3::prelude::; use pyo3::types::PyString; @@ -10,11 +10,10 @@ use tokio_util::codec::{BytesCodec, FramedRead}; use crate::http::HV_SERVER; - -#[pyclass(module="granian._granian")] +#[pyclass(module = "granian._granian")] #[derive(Clone)] pub(crate) struct RSGIHeaders { - inner: HeaderMap + inner: HeaderMap, } impl RSGIHeaders { @@ -29,7 +28,7 @@ impl RSGIHeaders { let mut ret = Vec::with_capacity(self.inner.keys_len()); for key in self.inner.keys() { ret.push(key.as_str()); - }; + } ret } @@ -37,15 +36,15 @@ impl RSGIHeaders { let mut ret = Vec::with_capacity(self.inner.keys_len()); for val in self.inner.values() { ret.push(val.to_str().unwrap()); - }; + } Ok(ret) } fn items(&self) -> PyResult<Vec<(&str, &str)>> { let mut ret = Vec::with_capacity(self.inner.keys_len()); - for (key, val) in self.inner.iter() { + for (key, val) in &self.inner { ret.push((key.as_str(), val.to_str().unwrap())); - }; + } Ok(ret) } @@ -56,18 +55,16 @@ impl RSGIHeaders { #[pyo3(signature = (key, default=None))] fn get(&self, py: Python, key: &str, default: Option<PyObject>) -> Option<PyObject> { match self.inner.get(key) { - Some(val) => { - match val.to_str() { - Ok(string) => Some(PyString::new(py, string).into()), - _ => default - } + Some(val) => match val.to_str() { + Ok(string) => Some(PyString::new(py, string).into()), + _ => default, }, - _ => default + _ => default, } } } -#[pyclass(module="granian._granian")] +#[pyclass(module = "granian._granian")] pub(crate) struct RSGIScope { #[pyo3(get)] proto: String, @@ -84,7 +81,7 @@ pub(crate) struct RSGIScope { #[pyo3(get)] client: String, #[pyo3(get)] - headers: RSGIHeaders + headers: RSGIHeaders, } impl RSGIScope { @@ -96,23 +93,23 @@ impl RSGIScope { method: &str, server: SocketAddr, client: SocketAddr, - headers: &HeaderMap + headers: &HeaderMap, ) -> Self { Self { proto: proto.to_string(), - http_version: http_version, + http_version, rsgi_version: "1.2".to_string(), scheme: scheme.to_string(), method: method.to_string(), - uri: uri, + uri, server: server.to_string(), client: client.to_string(), - headers: RSGIHeaders::new(headers) + headers: RSGIHeaders::new(headers), } } pub fn set_proto(&mut self, value: &str) { - self.proto = value.to_string() + self.proto = value.to_string(); } } @@ -125,7 +122,7 @@ impl RSGIScope { Version::HTTP_11 => "1.1", Version::HTTP_2 => "2", Version::HTTP_3 => "3", - _ => "1" + _ => "1", } } @@ -142,38 +139,36 @@ impl RSGIScope { pub(crate) enum PyResponse { Body(PyResponseBody), - File(PyResponseFile) + File(PyResponseFile), } pub(crate) struct PyResponseBody { status: u16, headers: Vec<(String, String)>, - body: Body + body: Body, } pub(crate) struct PyResponseFile { status: u16, headers: Vec<(String, String)>, - file_path: String + file_path: String, } macro_rules! response_head_from_py { - ($status:expr, $headers:expr, $res:expr) => { - { - let mut rh = hyper::http::HeaderMap::new(); - - rh.insert(HK_SERVER, HV_SERVER); - for (key, value) in $headers { - rh.append( - HeaderName::from_bytes(key.as_bytes()).unwrap(), - HeaderValue::from_str(&value).unwrap() - ); - } - - $res.status_mut() = $status.try_into().unwrap(); - $res.headers_mut() = rh; + ($status:expr, $headers:expr, $res:expr) => {{ + let mut rh = hyper::http::HeaderMap::new(); + + rh.insert(HK_SERVER, HV_SERVER); + for (key, value) in $headers { + rh.append( + HeaderName::from_bytes(key.as_bytes()).unwrap(), + HeaderValue::from_str(&value).unwrap(), + ); } - } + + $res.status_mut() = $status.try_into().unwrap(); + $res.headers_mut() = rh; + }}; } impl PyResponseBody { @@ -182,18 +177,30 @@ impl PyResponseBody { } pub fn empty(status: u16, headers: Vec<(String, String)>) -> Self { - Self { status, headers, body: Body::empty() } + Self { + status, + headers, + body: Body::empty(), + } } pub fn from_bytes(status: u16, headers: Vec<(String, String)>, body: Vec<u8>) -> Self { - Self { status, headers, body: Body::from(body) } + Self { + status, + headers, + body: Body::from(body), + } } pub fn from_string(status: u16, headers: Vec<(String, String)>, body: String) -> Self { - Self { status, headers, body: Body::from(body) } + Self { + status, + headers, + body: Body::from(body), + } } - pub fn to_response(self) -> hyper::Response::<Body> { + pub fn to_response(self) -> hyper::Response<Body> { let mut res = hyper::Response::<Body>::new(self.body); response_head_from_py!(self.status, &self.headers, res); res @@ -202,10 +209,14 @@ impl PyResponseBody { impl PyResponseFile { pub fn new(status: u16, headers: Vec<(String, String)>, file_path: String) -> Self { - Self { status, headers, file_path } + Self { + status, + headers, + file_path, + } } - pub async fn to_response(&self) -> hyper::Response::<Body> { + pub async fn to_response(&self) -> hyper::Response<Body> { let file = File::open(&self.file_path).await.unwrap(); let stream = FramedRead::new(file, BytesCodec::new()); let mut res = hyper::Response::<Body>::new(Body::wrap_stream(stream)); diff --git a/src/runtime.rs b/src/runtime.rs index 5b5e2198..4ed03ebb 100644 --- a/src/runtime.rs +++ b/src/runtime.rs @@ -1,12 +1,19 @@ use once_cell::unsync::OnceCell as UnsyncOnceCell; -use pyo3_asyncio::TaskLocals; use pyo3::prelude::; -use std::{future::Future, io, pin::Pin, sync::{Arc, Mutex}}; -use tokio::{runtime::Builder, task::{JoinHandle, LocalSet}}; +use pyo3_asyncio::TaskLocals; +use std::{ + future::Future, + io, + pin::Pin, + sync::{Arc, Mutex}, +}; +use tokio::{ + runtime::Builder, + task::{JoinHandle, LocalSet}, +}; use super::callbacks::{PyFutureAwaitable, PyIterAwaitable}; - tokio::task_local! { static TASK_LOCALS: UnsyncOnceCell<TaskLocals>; } @@ -27,11 +34,7 @@ pub trait Runtime: Send + 'static { } pub trait ContextExt: Runtime { - fn scope<F, R>( - &self, - locals: TaskLocals, - fut: F - ) -> Pin<Box<dyn Future<Output = R> + Send>> + fn scope<F, R>(&self, locals: TaskLocals, fut: F) -> Pin<Box<dyn Future<Output = R> + Send>> where F: Future<Output = R> + Send + 'static; @@ -45,36 +48,34 @@ pub trait SpawnLocalExt: Runtime { } pub trait LocalContextExt: Runtime { - fn scope_local<F, R>( - &self, - locals: TaskLocals, - fut: F - ) -> Pin<Box<dyn Future<Output = R>>> + fn scope_local<F, R>(&self, locals: TaskLocals, fut: F) -> Pin<Box<dyn Future<Output = R>>> where F: Future<Output = R> + 'static; } pub(crate) struct RuntimeWrapper { - rt: tokio::runtime::Runtime + rt: tokio::runtime::Runtime, } impl RuntimeWrapper { pub fn new(blocking_threads: usize) -> Self { - Self { rt: default_runtime(blocking_threads).unwrap() } + Self { + rt: default_runtime(blocking_threads).unwrap(), + } } pub fn with_runtime(rt: tokio::runtime::Runtime) -> Self { - Self { rt: rt } + Self { rt } } pub fn handler(&self) -> RuntimeRef { - RuntimeRef::new(self.rt.handle().to_owned()) + RuntimeRef::new(self.rt.handle().clone()) } } #[derive(Clone)] pub struct RuntimeRef { - pub inner: tokio::runtime::Handle + pub inner: tokio::runtime::Handle, } impl RuntimeRef { @@ -108,11 +109,7 @@ impl Runtime for RuntimeRef { } impl ContextExt for RuntimeRef { - fn scope<F, R>( - &self, - locals: TaskLocals, - fut: F - ) -> Pin<Box<dyn Future<Output = R> + Send>> + fn scope<F, R>(&self, locals: TaskLocals, fut: F) -> Pin<Box<dyn Future<Output = R> + Send>> where F: Future<Output = R> + Send + 'static, { @@ -123,7 +120,7 @@ impl ContextExt for RuntimeRef { } fn get_task_locals() -> Option<TaskLocals> { - match TASK_LOCALS.try_with(\|c\| c.get().map(\|locals\| locals.clone())) { + match TASK_LOCALS.try_with(\|c\| c.get().cloned()) { Ok(locals) => locals, Err(_) => None, } @@ -140,11 +137,7 @@ impl SpawnLocalExt for RuntimeRef { } impl LocalContextExt for RuntimeRef { - fn scope_local<F, R>( - &self, - locals: TaskLocals, - fut: F - ) -> Pin<Box<dyn Future<Output = R>>> + fn scope_local<F, R>(&self, locals: TaskLocals, fut: F) -> Pin<Box<dyn Future<Output = R>>> where F: Future<Output = R> + 'static, { @@ -169,7 +162,7 @@ pub(crate) fn init_runtime_mt(threads: usize, blocking_threads: usize) -> Runtim .max_blocking_threads(blocking_threads) .enable_all() .build() - .unwrap() + .unwrap(), ) } @@ -177,12 +170,8 @@ pub(crate) fn init_runtime_st(blocking_threads: usize) -> RuntimeWrapper { RuntimeWrapper::new(blocking_threads) } -pub(crate) fn into_future( - awaitable: &PyAny, -) -> PyResult<impl Future<Output = PyResult<PyObject>> + Send> { - pyo3_asyncio::into_future_with_locals( - &get_current_locals::<RuntimeRef>(awaitable.py())?, awaitable - ) +pub(crate) fn into_future(awaitable: &PyAny) -> PyResult<impl Future<Output = PyResult<PyObject>> + Send> { + pyo3_asyncio::into_future_with_locals(&get_current_locals::<RuntimeRef>(awaitable.py())?, awaitable) } #[inline] @@ -241,10 +230,7 @@ where rt.spawn(async move { let result = fut.await; Python::with_gil(move \|py\| { - PyFutureAwaitable::set_result( - py_aw.as_ref(py).borrow_mut(), - result.map(\|v\| v.into_py(py)) - ); + PyFutureAwaitable::set_result(py_aw.as_ref(py).borrow_mut(), result.map(\|v\| v.into_py(py))); }); }); @@ -269,11 +255,7 @@ where let rth = rt.handler(); rt.spawn(async move { - let val = rth.scope( - task_locals.clone(), - fut - ) - .await; + let val = rth.scope(task_locals.clone(), fut).await; if let Ok(mut result) = result_tx.lock() { result = Some(val.unwrap()); } @@ -292,7 +274,7 @@ where pub(crate) fn block_on_local<F>(rt: RuntimeWrapper, local: LocalSet, fut: F) where - F: Future + 'static + F: Future + 'static, { local.block_on(&rt.rt, fut); } diff --git a/src/tcp.rs b/src/tcp.rs index 4aa54eef..7b32fc95 100644 --- a/src/tcp.rs +++ b/src/tcp.rs @@ -9,10 +9,9 @@ use std::os::windows::io::{AsRawSocket, FromRawSocket}; use socket2::{Domain, Protocol, Socket, Type}; - -#[pyclass(module="granian._granian")] +#[pyclass(module = "granian._granian")] pub struct ListenerHolder { - socket: TcpListener + socket: TcpListener, } #[pymethods] @@ -20,28 +19,19 @@ impl ListenerHolder { #[cfg(unix)] #[new] pub fn new(fd: i32) -> PyResult<Self> { - let socket = unsafe { - TcpListener::from_raw_fd(fd) - }; - Ok(Self { socket: socket }) + let socket = unsafe { TcpListener::from_raw_fd(fd) }; + Ok(Self { socket }) } #[cfg(windows)] #[new] pub fn new(fd: u64) -> PyResult<Self> { - let socket = unsafe { - TcpListener::from_raw_socket(fd) - }; - Ok(Self { socket: socket }) + let socket = unsafe { TcpListener::from_raw_socket(fd) }; + Ok(Self { socket }) } #[classmethod] - pub fn from_address( - _cls: &PyType, - address: &str, - port: u16, - backlog: i32 - ) -> PyResult<Self> { + pub fn from_address(_cls: &PyType, address: &str, port: u16, backlog: i32) -> PyResult<Self> { let address: SocketAddr = (address.parse::<IpAddr>()?, port).into(); let socket = Socket::new(Domain::IPV4, Type::STREAM, Some(Protocol::TCP))?; socket.set_reuse_address(true)?; @@ -54,17 +44,13 @@ impl ListenerHolder { #[cfg(unix)] pub fn __getstate__(&self, py: Python) -> PyObject { let fd = self.socket.as_raw_fd(); - ( - fd.into_py(py), - ).to_object(py) + (fd.into_py(py),).to_object(py) } #[cfg(windows)] pub fn __getstate__(&self, py: Python) -> PyObject { let fd = self.socket.as_raw_socket(); - ( - fd.into_py(py), - ).to_object(py) + (fd.into_py(py),).to_object(py) } #[cfg(unix)] @@ -84,7 +70,6 @@ impl ListenerHolder { } } - pub(crate) fn init_pymodule(module: &PyModule) -> PyResult<()> { module.add_class::<ListenerHolder>()?; diff --git a/src/tls.rs b/src/tls.rs index d97d39bf..9355ed2f 100644 --- a/src/tls.rs +++ b/src/tls.rs @@ -1,20 +1,22 @@ use futures::stream::StreamExt; -use hyper::server::{accept, conn::{AddrIncoming, AddrStream}}; +use hyper::server::{ + accept, + conn::{AddrIncoming, AddrStream}, +}; use std::{fs, future, io, iter::Iterator, net::TcpListener, sync::Arc}; use tls_listener::{Error as TlsError, TlsListener}; use tokio_rustls::{ - TlsAcceptor, rustls::{Certificate, PrivateKey, ServerConfig}, - server::TlsStream + server::TlsStream, + TlsAcceptor, }; - pub(crate) type TlsAddrStream = TlsStream<AddrStream>; pub(crate) fn tls_listen( config: Arc<ServerConfig>, - tcp: TcpListener -) -> impl accept::Accept<Conn=TlsAddrStream, Error=TlsError<io::Error, io::Error>> { + tcp: TcpListener, +) -> impl accept::Accept<Conn = TlsAddrStream, Error = TlsError<io::Error, io::Error>> { tcp.set_nonblocking(true).unwrap(); let tcp_listener = tokio::net::TcpListener::from_std(tcp).unwrap(); let incoming = AddrIncoming::from_listener(tcp_listener).unwrap(); @@ -34,30 +36,26 @@ fn tls_error(err: String) -> io::Error { } pub(crate) fn load_certs(filename: &str) -> io::Result<Vec<Certificate>> { - let certfile = fs::File::open(filename) - .map_err(\|e\| tls_error(format!("failed to open {}: {}", filename, e)))?; + let certfile = fs::File::open(filename).map_err(\|e\| tls_error(format!("failed to open {filename}: {e}")))?; let mut reader = io::BufReader::new(certfile); - let certs = rustls_pemfile::certs(&mut reader) - .map_err(\|_\| tls_error("failed to load certificate".into()))?; + let certs = rustls_pemfile::certs(&mut reader).map_err(\|_\| tls_error("failed to load certificate".into()))?; Ok(certs.into_iter().map(Certificate).collect()) } pub(crate) fn load_private_key(filename: &str) -> io::Result<PrivateKey> { - let keyfile = fs::File::open(filename) - .map_err(\|e\| tls_error(format!("failed to open {}: {}", filename, e)))?; + let keyfile = fs::File::open(filename).map_err(\|e\| tls_error(format!("failed to open {filename}: {e}")))?; let mut reader = io::BufReader::new(keyfile); - let keys = rustls_pemfile::read_all(&mut reader) - .map_err(\|_\| tls_error("failed to load private key".into()))?; + let keys = rustls_pemfile::read_all(&mut reader).map_err(\|_\| tls_error("failed to load private key".into()))?; if keys.len() != 1 { return Err(tls_error("expected a single private key".into())); } let key = match &keys[0] { - rustls_pemfile::Item::RSAKey(key) => PrivateKey(key.to_vec()), - rustls_pemfile::Item::PKCS8Key(key) => PrivateKey(key.to_vec()), - rustls_pemfile::Item::ECKey(key) => PrivateKey(key.to_vec()), + rustls_pemfile::Item::RSAKey(key) => PrivateKey(key.clone()), + rustls_pemfile::Item::PKCS8Key(key) => PrivateKey(key.clone()), + rustls_pemfile::Item::ECKey(key) => PrivateKey(key.clone()), _ => { return Err(tls_error("failed to load private key".into())); } diff --git a/src/utils.rs b/src/utils.rs index 379ec656..93cf4c3d 100644 --- a/src/utils.rs +++ b/src/utils.rs @@ -1,13 +1,15 @@ pub(crate) fn header_contains_value( headers: &hyper::HeaderMap, header: impl hyper::header::AsHeaderName, - value: impl AsRef<[u8]> + value: impl AsRef<[u8]>, ) -> bool { let value = value.as_ref(); for header in headers.get_all(header) { - if header.as_bytes().split(\|&c\| c == b',').any( - \|x\| trim(x).eq_ignore_ascii_case(value) - ) { + if header + .as_bytes() + .split(\|&c\| c == b',') + .any(\|x\| trim(x).eq_ignore_ascii_case(value)) + { return true; } } @@ -30,7 +32,7 @@ fn trim_start(data: &[u8]) -> &[u8] { #[inline] fn trim_end(data: &[u8]) -> &[u8] { if let Some(last) = data.iter().rposition(\|x\| !x.is_ascii_whitespace()) { - &data[..last + 1] + &data[..=last] } else { b"" } diff --git a/src/workers.rs b/src/workers.rs index e49fa3b6..fc7750d2 100644 --- a/src/workers.rs +++ b/src/workers.rs @@ -8,8 +8,8 @@ use std::os::windows::io::FromRawSocket; use super::asgi::serve::ASGIWorker; use super::rsgi::serve::RSGIWorker; -use super::wsgi::serve::WSGIWorker; use super::tls::{load_certs as tls_load_certs, load_private_key as tls_load_pkey}; +use super::wsgi::serve::WSGIWorker; pub(crate) struct WorkerConfig { pub id: i32, @@ -22,7 +22,7 @@ pub(crate) struct WorkerConfig { pub opt_enabled: bool, pub ssl_enabled: bool, ssl_cert: Option<String>, - ssl_key: Option<String> + ssl_key: Option<String>, } impl WorkerConfig { @@ -37,7 +37,7 @@ impl WorkerConfig { opt_enabled: bool, ssl_enabled: bool, ssl_cert: Option<&str>, - ssl_key: Option<&str> + ssl_key: Option<&str>, ) -> Self { Self { id, @@ -49,23 +49,19 @@ impl WorkerConfig { websockets_enabled, opt_enabled, ssl_enabled, - ssl_cert: ssl_cert.map_or(None, \|v\| Some(v.into())), - ssl_key: ssl_key.map_or(None, \|v\| Some(v.into())) + ssl_cert: ssl_cert.map(std::convert::Into::into), + ssl_key: ssl_key.map(std::convert::Into::into), } } #[cfg(unix)] pub fn tcp_listener(&self) -> TcpListener { - unsafe { - TcpListener::from_raw_fd(self.socket_fd) - } + unsafe { TcpListener::from_raw_fd(self.socket_fd) } } #[cfg(windows)] pub fn tcp_listener(&self) -> TcpListener { - unsafe { - TcpListener::from_raw_socket(self.socket_fd as u64) - } + unsafe { TcpListener::from_raw_socket(self.socket_fd as u64) } } pub fn tls_cfg(&self) -> tokio_rustls::rustls::ServerConfig { @@ -74,13 +70,13 @@ impl WorkerConfig { .with_no_client_auth() .with_single_cert( tls_load_certs(&self.ssl_cert.clone().unwrap()[..]).unwrap(), - tls_load_pkey(&self.ssl_key.clone().unwrap()[..]).unwrap() + tls_load_pkey(&self.ssl_key.clone().unwrap()[..]).unwrap(), ) .unwrap(); cfg.alpn_protocols = match &self.http_mode[..] { "1" => vec![b"http/1.1".to_vec()], "2" => vec![b"h2".to_vec()], - _ => vec![b"h2".to_vec(), b"http/1.1".to_vec()] + _ => vec![b"h2".to_vec(), b"http/1.1".to_vec()], }; cfg } @@ -102,7 +98,7 @@ pub(crate) struct WorkerExecutor; impl<F> hyper::rt::Executor<F> for WorkerExecutor where - F: std::future::Future + 'static + F: std::future::Future + 'static, { fn execute(&self, fut: F) { tokio::task::spawn_local(fut); @@ -123,14 +119,9 @@ macro_rules! build_service { let rth = rth.clone(); async move { - Ok::<_, std::convert::Infallible>($target( - rth, - callback_wrapper, - local_addr, - remote_addr, - req, - "http" - ).await) + Ok::<_, std::convert::Infallible>( + $target(rth, callback_wrapper, local_addr, remote_addr, req, "http").await, + ) } })) } @@ -153,14 +144,9 @@ macro_rules! build_service_ssl { let rth = rth.clone(); async move { - Ok::<_, std::convert::Infallible>($target( - rth, - callback_wrapper, - local_addr, - remote_addr, - req, - "https" - ).await) + Ok::<_, std::convert::Infallible>( + $target(rth, callback_wrapper, local_addr, remote_addr, req, "https").await, + ) } })) } @@ -170,13 +156,7 @@ macro_rules! build_service_ssl { macro_rules! serve_rth { ($func_name:ident, $target:expr) => { - fn $func_name( - &self, - callback: PyObject, - event_loop: &PyAny, - context: &PyAny, - signal_rx: PyObject - ) { + fn $func_name(&self, callback: PyObject, event_loop: &PyAny, context: &PyAny, signal_rx: PyObject) { pyo3_log::init(); let rt = crate::runtime::init_runtime_mt(self.config.threads, self.config.pthreads); let rth = rt.handler(); @@ -184,34 +164,30 @@ macro_rules! serve_rth { let http1_only = self.config.http_mode == "1"; let http2_only = self.config.http_mode == "2"; let http1_buffer_max = self.config.http1_buffer_max.clone(); - let callback_wrapper = crate::callbacks::CallbackWrapper::new( - callback, event_loop, context - ); + let callback_wrapper = crate::callbacks::CallbackWrapper::new(callback, event_loop, context); let worker_id = self.config.id; log::info!("Started worker-{}", worker_id); - let svc_loop = crate::runtime::run_until_complete( - rt.handler(), - event_loop, - async move { - let service = crate::workers::build_service!( - callback_wrapper, rth, $target - ); - let server = hyper::Server::from_tcp(tcp_listener).unwrap() - .http1_only(http1_only) - .http2_only(http2_only) - .http1_max_buf_size(http1_buffer_max) - .serve(service); - server.with_graceful_shutdown(async move { - Python::with_gil(\|py\| { - crate::runtime::into_future(signal_rx.as_ref(py)).unwrap() - }).await.unwrap(); - }).await.unwrap(); - log::info!("Stopping worker-{}", worker_id); - Ok(()) - } - ); + let svc_loop = crate::runtime::run_until_complete(rt.handler(), event_loop, async move { + let service = crate::workers::build_service!(callback_wrapper, rth, $target); + let server = hyper::Server::from_tcp(tcp_listener) + .unwrap() + .http1_only(http1_only) + .http2_only(http2_only) + .http1_max_buf_size(http1_buffer_max) + .serve(service); + server + .with_graceful_shutdown(async move { + Python::with_gil(\|py\| crate::runtime::into_future(signal_rx.as_ref(py)).unwrap()) + .await + .unwrap(); + }) + .await + .unwrap(); + log::info!("Stopping worker-{}", worker_id); + Ok(()) + }); match svc_loop { Ok(_) => {} @@ -226,13 +202,7 @@ macro_rules! serve_rth { macro_rules! serve_rth_ssl { ($func_name:ident, $target:expr) => { - fn $func_name( - &self, - callback: PyObject, - event_loop: &PyAny, - context: &PyAny, - signal_rx: PyObject - ) { + fn $func_name(&self, callback: PyObject, event_loop: &PyAny, context: &PyAny, signal_rx: PyObject) { pyo3_log::init(); let rt = crate::runtime::init_runtime_mt(self.config.threads, self.config.pthreads); let rth = rt.handler(); @@ -241,38 +211,29 @@ macro_rules! serve_rth_ssl { let http2_only = self.config.http_mode == "2"; let http1_buffer_max = self.config.http1_buffer_max.clone(); let tls_cfg = self.config.tls_cfg(); - let callback_wrapper = crate::callbacks::CallbackWrapper::new( - callback, event_loop, context - ); + let callback_wrapper = crate::callbacks::CallbackWrapper::new(callback, event_loop, context); let worker_id = self.config.id; log::info!("Started worker-{}", worker_id); - let svc_loop = crate::runtime::run_until_complete( - rt.handler(), - event_loop, - async move { - let service = crate::workers::build_service_ssl!( - callback_wrapper, rth, $target - ); - let server = hyper::Server::builder( - crate::tls::tls_listen( - std::sync::Arc::new(tls_cfg), tcp_listener - ) - ) - .http1_only(http1_only) - .http2_only(http2_only) - .http1_max_buf_size(http1_buffer_max) - .serve(service); - server.with_graceful_shutdown(async move { - Python::with_gil(\|py\| { - crate::runtime::into_future(signal_rx.as_ref(py)).unwrap() - }).await.unwrap(); - }).await.unwrap(); - log::info!("Stopping worker-{}", worker_id); - Ok(()) - } - ); + let svc_loop = crate::runtime::run_until_complete(rt.handler(), event_loop, async move { + let service = crate::workers::build_service_ssl!(callback_wrapper, rth, $target); + let server = hyper::Server::builder(crate::tls::tls_listen(std::sync::Arc::new(tls_cfg), tcp_listener)) + .http1_only(http1_only) + .http2_only(http2_only) + .http1_max_buf_size(http1_buffer_max) + .serve(service); + server + .with_graceful_shutdown(async move { + Python::with_gil(\|py\| crate::runtime::into_future(signal_rx.as_ref(py)).unwrap()) + .await + .unwrap(); + }) + .await + .unwrap(); + log::info!("Stopping worker-{}", worker_id); + Ok(()) + }); match svc_loop { Ok(_) => {} @@ -287,22 +248,14 @@ macro_rules! serve_rth_ssl { macro_rules! serve_wth { ($func_name: ident, $target:expr) => { - fn $func_name( - &self, - callback: PyObject, - event_loop: &PyAny, - context: &PyAny, - signal_rx: PyObject - ) { + fn $func_name(&self, callback: PyObject, event_loop: &PyAny, context: &PyAny, signal_rx: PyObject) { pyo3_log::init(); let rtm = crate::runtime::init_runtime_mt(1, 1); let worker_id = self.config.id; log::info!("Started worker-{}", worker_id); - let callback_wrapper = crate::callbacks::CallbackWrapper::new( - callback, event_loop, context - ); + let callback_wrapper = crate::callbacks::CallbackWrapper::new(callback, event_loop, context); let mut workers = vec![]; let (stx, srx) = tokio::sync::watch::channel(false); @@ -323,38 +276,36 @@ macro_rules! serve_wth { let local = tokio::task::LocalSet::new(); crate::runtime::block_on_local(rt, local, async move { - let service = crate::workers::build_service!( - callback_wrapper, rth, $target - ); - let server = hyper::Server::from_tcp(tcp_listener).unwrap() + let service = crate::workers::build_service!(callback_wrapper, rth, $target); + let server = hyper::Server::from_tcp(tcp_listener) + .unwrap() .executor(crate::workers::WorkerExecutor) .http1_only(http1_only) .http2_only(http2_only) .http1_max_buf_size(http1_buffer_max) .serve(service); - server.with_graceful_shutdown(async move { - srx.changed().await.unwrap(); - }).await.unwrap(); + server + .with_graceful_shutdown(async move { + srx.changed().await.unwrap(); + }) + .await + .unwrap(); log::info!("Stopping worker-{} runtime-{}", worker_id, thread_id + 1); }); })); - }; + } - let main_loop = crate::runtime::run_until_complete( - rtm.handler(), - event_loop, - async move { - Python::with_gil(\|py\| { - crate::runtime::into_future(signal_rx.as_ref(py)).unwrap() - }).await.unwrap(); - stx.send(true).unwrap(); - log::info!("Stopping worker-{}", worker_id); - while let Some(worker) = workers.pop() { - worker.join().unwrap(); - } - Ok(()) + let main_loop = crate::runtime::run_until_complete(rtm.handler(), event_loop, async move { + Python::with_gil(\|py\| crate::runtime::into_future(signal_rx.as_ref(py)).unwrap()) + .await + .unwrap(); + stx.send(true).unwrap(); + log::info!("Stopping worker-{}", worker_id); + while let Some(worker) = workers.pop() { + worker.join().unwrap(); } - ); + Ok(()) + }); match main_loop { Ok(_) => {} @@ -369,22 +320,14 @@ macro_rules! serve_wth { macro_rules! serve_wth_ssl { ($func_name: ident, $target:expr) => { - fn $func_name( - &self, - callback: PyObject, - event_loop: &PyAny, - context: &PyAny, - signal_rx: PyObject - ) { + fn $func_name(&self, callback: PyObject, event_loop: &PyAny, context: &PyAny, signal_rx: PyObject) { pyo3_log::init(); let rtm = crate::runtime::init_runtime_mt(1, 1); let worker_id = self.config.id; log::info!("Started worker-{}", worker_id); - let callback_wrapper = crate::callbacks::CallbackWrapper::new( - callback, event_loop, context - ); + let callback_wrapper = crate::callbacks::CallbackWrapper::new(callback, event_loop, context); let mut workers = vec![]; let (stx, srx) = tokio::sync::watch::channel(false); @@ -406,42 +349,36 @@ macro_rules! serve_wth_ssl { let local = tokio::task::LocalSet::new(); crate::runtime::block_on_local(rt, local, async move { - let service = crate::workers::build_service_ssl!( - callback_wrapper, rth, $target - ); - let server = hyper::Server::builder( - crate::tls::tls_listen( - std::sync::Arc::new(tls_cfg), tcp_listener - ) - ) - .executor(crate::workers::WorkerExecutor) - .http1_only(http1_only) - .http2_only(http2_only) - .http1_max_buf_size(http1_buffer_max) - .serve(service); - server.with_graceful_shutdown(async move { - srx.changed().await.unwrap(); - }).await.unwrap(); + let service = crate::workers::build_service_ssl!(callback_wrapper, rth, $target); + let server = + hyper::Server::builder(crate::tls::tls_listen(std::sync::Arc::new(tls_cfg), tcp_listener)) + .executor(crate::workers::WorkerExecutor) + .http1_only(http1_only) + .http2_only(http2_only) + .http1_max_buf_size(http1_buffer_max) + .serve(service); + server + .with_graceful_shutdown(async move { + srx.changed().await.unwrap(); + }) + .await + .unwrap(); log::info!("Stopping worker-{} runtime-{}", worker_id, thread_id + 1); }); })); - }; + } - let main_loop = crate::runtime::run_until_complete( - rtm.handler(), - event_loop, - async move { - Python::with_gil(\|py\| { - crate::runtime::into_future(signal_rx.as_ref(py)).unwrap() - }).await.unwrap(); - stx.send(true).unwrap(); - log::info!("Stopping worker-{}", worker_id); - while let Some(worker) = workers.pop() { - worker.join().unwrap(); - } - Ok(()) + let main_loop = crate::runtime::run_until_complete(rtm.handler(), event_loop, async move { + Python::with_gil(\|py\| crate::runtime::into_future(signal_rx.as_ref(py)).unwrap()) + .await + .unwrap(); + stx.send(true).unwrap(); + log::info!("Stopping worker-{}", worker_id); + while let Some(worker) = workers.pop() { + worker.join().unwrap(); } - ); + Ok(()) + }); match main_loop { Ok(_) => {} @@ -457,8 +394,8 @@ macro_rules! serve_wth_ssl { pub(crate) use build_service; pub(crate) use build_service_ssl; pub(crate) use serve_rth; -pub(crate) use serve_wth; pub(crate) use serve_rth_ssl; +pub(crate) use serve_wth; pub(crate) use serve_wth_ssl; pub(crate) fn init_pymodule(module: &PyModule) -> PyResult<()> { diff --git a/src/ws.rs b/src/ws.rs index bd44dded..3f892289 100644 --- a/src/ws.rs +++ b/src/ws.rs @@ -1,24 +1,24 @@ use hyper::{ - Body, - Request, - Response, - StatusCode, header::{CONNECTION, UPGRADE}, - http::response::Builder + http::response::Builder, + Body, Request, Response, StatusCode, }; +use pin_project::pin_project; +use std::{ + future::Future, + pin::Pin, + task::{Context, Poll}, +}; +use tokio::sync::mpsc; +use tokio_tungstenite::WebSocketStream; use tungstenite::{ error::ProtocolError, handshake::derive_accept_key, - protocol::{Role, WebSocketConfig} + protocol::{Role, WebSocketConfig}, }; -use pin_project::pin_project; -use std::{future::Future, pin::Pin, task::{Context, Poll}}; -use tokio_tungstenite::WebSocketStream; -use tokio::sync::mpsc; use super::utils::header_contains_value; - #[pin_project] #[derive(Debug)] pub struct HyperWebsocket { @@ -37,15 +37,9 @@ impl Future for HyperWebsocket { Poll::Ready(x) => x, }; - let upgraded = upgraded.map_err(\|_\| - tungstenite::Error::Protocol(ProtocolError::HandshakeIncomplete) - )?; + let upgraded = upgraded.map_err(\|_\| tungstenite::Error::Protocol(ProtocolError::HandshakeIncomplete))?; - let stream = WebSocketStream::from_raw_socket( - upgraded, - Role::Server, - this.config.take(), - ); + let stream = WebSocketStream::from_raw_socket(upgraded, Role::Server, this.config.take()); tokio::pin!(stream); match stream.as_mut().poll(cx) { @@ -58,18 +52,15 @@ impl Future for HyperWebsocket { pub(crate) struct UpgradeData { response_builder: Option<Builder>, response_tx: Option<mpsc::Sender<Response<Body>>>, - pub consumed: bool + pub consumed: bool, } impl UpgradeData { - pub fn new( - response_builder: Builder, - response_tx: mpsc::Sender<Response<Body>>) - -> Self { + pub fn new(response_builder: Builder, response_tx: mpsc::Sender<Response<Body>>) -> Self { Self { response_builder: Some(response_builder), response_tx: Some(response_tx), - consumed: false + consumed: false, } } @@ -79,19 +70,16 @@ impl UpgradeData { Ok(_) => { self.consumed = true; Ok(()) - }, - err => err + } + err => err, } } } #[inline] pub(crate) fn is_upgrade_request<B>(request: &Request<B>) -> bool { - header_contains_value( - request.headers(), CONNECTION, "Upgrade" - ) && header_contains_value( - request.headers(), UPGRADE, "websocket" - ) + header_contains_value(request.headers(), CONNECTION, "Upgrade") + && header_contains_value(request.headers(), UPGRADE, "websocket") } pub(crate) fn upgrade_intent<B>( @@ -100,13 +88,17 @@ pub(crate) fn upgrade_intent<B>( ) -> Result<(Builder, HyperWebsocket), ProtocolError> { let request = request.borrow_mut(); - let key = request.headers() + let key = request + .headers() .get("Sec-WebSocket-Key") .ok_or(ProtocolError::MissingSecWebSocketKey)?; - if request.headers().get("Sec-WebSocket-Version").map( - \|v\| v.as_bytes() - ) != Some(b"13") { + if request + .headers() + .get("Sec-WebSocket-Version") + .map(hyper::http::HeaderValue::as_bytes) + != Some(b"13") + { return Err(ProtocolError::MissingSecWebSocketVersionHeader); } diff --git a/src/wsgi/callbacks.rs b/src/wsgi/callbacks.rs index ae86cf01..e961827b 100644 --- a/src/wsgi/callbacks.rs +++ b/src/wsgi/callbacks.rs @@ -2,44 +2,35 @@ use hyper::Body; use pyo3::prelude::; use tokio::task::JoinHandle; +use super::types::{WSGIResponseBodyIter, WSGIScope as Scope}; use crate::callbacks::CallbackWrapper; -use super::types::{WSGIScope as Scope, WSGIResponseBodyIter}; const WSGI_LIST_RESPONSE_BODY: i32 = 0; const WSGI_ITER_RESPONSE_BODY: i32 = 1; - #[inline(always)] -fn run_callback( - callback: PyObject, - scope: Scope -) -> PyResult<(i32, Vec<(String, String)>, Body)> { +fn run_callback(callback: PyObject, scope: Scope) -> PyResult<(i32, Vec<(String, String)>, Body)> { Python::with_gil(\|py\| { - let (status, headers, body_type, pybody) = callback.call1(py, (scope,))? - .extract::<(i32, Vec<(String, String)>, i32, PyObject)>(py)?; + let (status, headers, body_type, pybody) = + callback + .call1(py, (scope,))? + .extract::<(i32, Vec<(String, String)>, i32, PyObject)>(py)?; let body = match body_type { WSGI_LIST_RESPONSE_BODY => Body::from(pybody.extract::<Vec<u8>>(py)?), WSGI_ITER_RESPONSE_BODY => Body::wrap_stream(WSGIResponseBodyIter::new(pybody)), - _ => Body::empty() + _ => Body::empty(), }; Ok((status, headers, body)) }) } -pub(crate) fn call_rtb_http( - cb: CallbackWrapper, - scope: Scope -) -> PyResult<(i32, Vec<(String, String)>, Body)> { - run_callback(cb.callback.clone(), scope) +pub(crate) fn call_rtb_http(cb: CallbackWrapper, scope: Scope) -> PyResult<(i32, Vec<(String, String)>, Body)> { + run_callback(cb.callback, scope) } pub(crate) fn call_rtt_http( cb: CallbackWrapper, - scope: Scope + scope: Scope, ) -> JoinHandle<PyResult<(i32, Vec<(String, String)>, Body)>> { - let callback = cb.callback.clone(); - - tokio::task::spawn_blocking(move \|\| { - run_callback(callback, scope) - }) + tokio::task::spawn_blocking(move \|\| run_callback(cb.callback, scope)) } diff --git a/src/wsgi/http.rs b/src/wsgi/http.rs index 66cccd75..e74882a7 100644 --- a/src/wsgi/http.rs +++ b/src/wsgi/http.rs @@ -1,21 +1,18 @@ use hyper::{ - Body, - Request, - Response, - header::{SERVER as HK_SERVER, HeaderName, HeaderValue} + header::{HeaderName, HeaderValue, SERVER as HK_SERVER}, + Body, Request, Response, }; use std::net::SocketAddr; +use super::{ + callbacks::{call_rtb_http, call_rtt_http}, + types::WSGIScope as Scope, +}; use crate::{ callbacks::CallbackWrapper, - http::{HV_SERVER, response_500}, + http::{response_500, HV_SERVER}, runtime::RuntimeRef, }; -use super::{ - callbacks::{call_rtb_http, call_rtt_http}, - types::WSGIScope as Scope -}; - #[inline(always)] fn build_response(status: i32, pyheaders: Vec<(String, String)>, body: Body) -> Response<Body> { @@ -26,7 +23,7 @@ fn build_response(status: i32, pyheaders: Vec<(String, String)>, body: Body) -> for (key, val) in pyheaders { headers.append( HeaderName::from_bytes(key.as_bytes()).unwrap(), - HeaderValue::from_str(&val).unwrap() + HeaderValue::from_str(&val).unwrap(), ); } res @@ -38,14 +35,11 @@ pub(crate) async fn handle_rtt( server_addr: SocketAddr, client_addr: SocketAddr, req: Request<Body>, - scheme: &str + scheme: &str, ) -> Response<Body> { - if let Ok(res) = call_rtt_http( - callback, - Scope::new(scheme, server_addr, client_addr, req).await - ).await { + if let Ok(res) = call_rtt_http(callback, Scope::new(scheme, server_addr, client_addr, req).await).await { if let Ok((status, headers, body)) = res { - return build_response(status, headers, body) + return build_response(status, headers, body); } log::warn!("Application callable raised an exception"); } else { @@ -60,12 +54,9 @@ pub(crate) async fn handle_rtb( server_addr: SocketAddr, client_addr: SocketAddr, req: Request<Body>, - scheme: &str + scheme: &str, ) -> Response<Body> { - match call_rtb_http( - callback, - Scope::new(scheme, server_addr, client_addr, req).await - ) { + match call_rtb_http(callback, Scope::new(scheme, server_addr, client_addr, req).await) { Ok((status, headers, body)) => build_response(status, headers, body), _ => { log::warn!("Application callable raised an exception"); diff --git a/src/wsgi/serve.rs b/src/wsgi/serve.rs index 284a4cdf..0b60e8b3 100644 --- a/src/wsgi/serve.rs +++ b/src/wsgi/serve.rs @@ -1,17 +1,11 @@ use pyo3::prelude::; -use crate::workers::{ - WorkerConfig, - serve_rth, - serve_wth, - serve_rth_ssl, - serve_wth_ssl -}; use super::http::{handle_rtb, handle_rtt}; +use crate::workers::{serve_rth, serve_rth_ssl, serve_wth, serve_wth_ssl, WorkerConfig}; -#[pyclass(module="granian._granian")] +#[pyclass(module = "granian._granian")] pub struct WSGIWorker { - config: WorkerConfig + config: WorkerConfig, } impl WSGIWorker { @@ -46,7 +40,7 @@ impl WSGIWorker { http1_buffer_max: usize, ssl_enabled: bool, ssl_cert: Option<&str>, - ssl_key: Option<&str> + ssl_key: Option<&str>, ) -> PyResult<Self> { Ok(Self { config: WorkerConfig::new( @@ -60,34 +54,22 @@ impl WSGIWorker { true, ssl_enabled, ssl_cert, - ssl_key - ) + ssl_key, + ), }) } - fn serve_rth( - &self, - callback: PyObject, - event_loop: &PyAny, - context: &PyAny, - signal_rx: PyObject - ) { + fn serve_rth(&self, callback: PyObject, event_loop: &PyAny, context: &PyAny, signal_rx: PyObject) { match self.config.ssl_enabled { false => self._serve_rth(callback, event_loop, context, signal_rx), - true => self._serve_rth_ssl(callback, event_loop, context, signal_rx) + true => self._serve_rth_ssl(callback, event_loop, context, signal_rx), } } - fn serve_wth( - &self, - callback: PyObject, - event_loop: &PyAny, - context: &PyAny, - signal_rx: PyObject - ) { + fn serve_wth(&self, callback: PyObject, event_loop: &PyAny, context: &PyAny, signal_rx: PyObject) { match self.config.ssl_enabled { false => self._serve_wth(callback, event_loop, context, signal_rx), - true => self._serve_wth_ssl(callback, event_loop, context, signal_rx) + true => self._serve_wth_ssl(callback, event_loop, context, signal_rx), } } } diff --git a/src/wsgi/types.rs b/src/wsgi/types.rs index 02c24b26..7e5c71f8 100644 --- a/src/wsgi/types.rs +++ b/src/wsgi/types.rs @@ -1,23 +1,21 @@ use futures::Stream; use hyper::{ body::Bytes, - header::{CONTENT_TYPE, CONTENT_LENGTH, HeaderMap}, - Body, - Method, - Request, - Uri, - Version + header::{HeaderMap, CONTENT_LENGTH, CONTENT_TYPE}, + Body, Method, Request, Uri, Version, }; -use pyo3::{prelude::, types::IntoPyDict}; use pyo3::types::{PyBytes, PyDict, PyList}; -use std::{net::{IpAddr, SocketAddr}, task::{Context, Poll}}; +use pyo3::{prelude::, types::IntoPyDict}; +use std::{ + net::{IpAddr, SocketAddr}, + task::{Context, Poll}, +}; const LINE_SPLIT: u8 = u8::from_be_bytes(*b"\n"); - #[pyclass(module = "granian._granian")] pub(crate) struct WSGIBody { - inner: Bytes + inner: Bytes, } impl WSGIBody { @@ -36,9 +34,9 @@ impl WSGIBody { match self.inner.iter().position(\|&c\| c == LINE_SPLIT) { Some(next_split) => { let bytes = self.inner.split_to(next_split); - Some(PyBytes::new(py, &bytes[..])) - }, - _ => None + Some(PyBytes::new(py, &bytes)) + } + _ => None, } } @@ -48,18 +46,16 @@ impl WSGIBody { None => { let bytes = self.inner.split_to(self.inner.len()); PyBytes::new(py, &bytes[..]) - }, - Some(size) => { - match size { - 0 => PyBytes::new(py, b""), - size => { - let limit = self.inner.len(); - let rsize = if size > limit { limit } else { size }; - let bytes = self.inner.split_to(rsize); - PyBytes::new(py, &bytes[..]) - } - } } + Some(size) => match size { + 0 => PyBytes::new(py, b""), + size => { + let limit = self.inner.len(); + let rsize = if size > limit { limit } else { size }; + let bytes = self.inner.split_to(rsize); + PyBytes::new(py, &bytes[..]) + } + }, } } @@ -69,16 +65,17 @@ impl WSGIBody { let bytes = self.inner.split_to(next_split); self.inner = self.inner.slice(1..); PyBytes::new(py, &bytes[..]) - }, - _ => PyBytes::new(py, b"") + } + _ => PyBytes::new(py, b""), } } #[pyo3(signature = (_hint=None))] fn readlines<'p>(&mut self, py: Python<'p>, _hint: Option<PyObject>) -> &'p PyList { - let lines: Vec<&PyBytes> = self.inner + let lines: Vec<&PyBytes> = self + .inner .split(\|&c\| c == LINE_SPLIT) - .map(\|item\| PyBytes::new(py, &item[..])) + .map(\|item\| PyBytes::new(py, item)) .collect(); self.inner.clear(); PyList::new(py, lines) @@ -95,28 +92,19 @@ pub(crate) struct WSGIScope { server_port: u16, client: String, headers: HeaderMap, - body: Bytes + body: Bytes, } impl WSGIScope { - pub async fn new( - scheme: &str, - server: SocketAddr, - client: SocketAddr, - request: Request<Body>, - ) -> Self { + pub async fn new(scheme: &str, server: SocketAddr, client: SocketAddr, request: Request<Body>) -> Self { let http_version = request.version(); - let method = request.method().to_owned(); - let uri = request.uri().to_owned(); - let headers = request.headers().to_owned(); + let method = request.method().clone(); + let uri = request.uri().clone(); + let headers = request.headers().clone(); let body = match method { - Method::HEAD \| Method::GET \| Method::OPTIONS => { Bytes::new() }, - _ => { - hyper::body::to_bytes(request) - .await - .unwrap_or(Bytes::new()) - } + Method::HEAD \| Method::GET \| Method::OPTIONS => Bytes::new(), + _ => hyper::body::to_bytes(request).await.unwrap_or(Bytes::new()), }; Self { @@ -128,7 +116,7 @@ impl WSGIScope { server_port: server.port(), client: client.to_string(), headers, - body + body, } } @@ -138,7 +126,7 @@ impl WSGIScope { Version::HTTP_10 => "HTTP/1", Version::HTTP_11 => "HTTP/1.1", Version::HTTP_2 => "HTTP/2", - _ => "HTTP/1" + _ => "HTTP/1", } } } @@ -157,21 +145,21 @@ impl WSGIScope { content_type, content_len, headers, - body + body, ) = py.allow_threads(\|\| { - let (path, query_string) = self.uri.path_and_query() + let (path, query_string) = self + .uri + .path_and_query() .map_or_else(\|\| ("", ""), \|pq\| (pq.path(), pq.query().unwrap_or(""))); let content_type = self.headers.remove(CONTENT_TYPE); let content_len = self.headers.remove(CONTENT_LENGTH); let mut headers = Vec::with_capacity(self.headers.len()); - for (key, val) in self.headers.iter() { - headers.push( - ( - format!("HTTP_{}", key.as_str().replace("-", "_").to_uppercase()), - val.to_str().unwrap_or_default() - ) - ); + for (key, val) in &self.headers { + headers.push(( + format!("HTTP_{}", key.as_str().replace('-', "_").to_uppercase()), + val.to_str().unwrap_or_default(), + )); } ( @@ -185,7 +173,7 @@ impl WSGIScope { content_type, content_len, headers, - WSGIBody::new(self.body.to_owned()) + WSGIBody::new(self.body.clone()), ) }); @@ -202,13 +190,13 @@ impl WSGIScope { if let Some(content_type) = content_type { ret.set_item( pyo3::intern!(py, "CONTENT_TYPE"), - content_type.to_str().unwrap_or_default() + content_type.to_str().unwrap_or_default(), )?; } if let Some(content_len) = content_len { ret.set_item( pyo3::intern!(py, "CONTENT_LENGTH"), - content_len.to_str().unwrap_or_default() + content_len.to_str().unwrap_or_default(), )?; } @@ -219,7 +207,7 @@ impl WSGIScope { } pub(crate) struct WSGIResponseBodyIter { - inner: PyObject + inner: PyObject, } impl WSGIResponseBodyIter { @@ -235,27 +223,20 @@ impl WSGIResponseBodyIter { impl Stream for WSGIResponseBodyIter { type Item = PyResult<Vec<u8>>; - fn poll_next( - self: std::pin::Pin<&mut Self>, - _cx: &mut Context<'_> - ) -> Poll<Option<Self::Item>> { - Python::with_gil(\|py\| { - match self.inner.call_method0(py, pyo3::intern!(py, "__next__")) { - Ok(chunk_obj) => { - match chunk_obj.extract::<Vec<u8>>(py) { - Ok(chunk) => Poll::Ready(Some(Ok(chunk))), - _ => { - self.close_inner(py); - Poll::Ready(None) - } - } - }, - Err(err) => { - if err.is_instance_of::<pyo3::exceptions::PyStopIteration>(py) { - self.close_inner(py); - } + fn poll_next(self: std::pin::Pin<&mut Self>, _cx: &mut Context<'_>) -> Poll<Option<Self::Item>> { + Python::with_gil(\|py\| match self.inner.call_method0(py, pyo3::intern!(py, "__next__")) { + Ok(chunk_obj) => match chunk_obj.extract::<Vec<u8>>(py) { + Ok(chunk) => Poll::Ready(Some(Ok(chunk))), + _ => { + self.close_inner(py); Poll::Ready(None) } + }, + Err(err) => { + if err.is_instance_of::<pyo3::exceptions::PyStopIteration>(py) { + self.close_inner(py); + } + Poll::Ready(None) } }) }	Python code formatting and linting Considering the more people contributing to the project, especially having #116, will it be appropriate to establish the formatting and linting rules for Python code? The obvious benefit is that no contributor will doubt what development tools and styles to apply. I will gladly provide a configuration for tools like `black`, `isort`, and `ruff`, for a start at least. While doing #122, I deactivated my `black` and `ruff` plugins because any file save immediately changes too many things I didn’t touch. And those two are the industry standard by now.	@Aeron it definitely makes sense. If you don't mind, I gonna open a PR myself later today so I'll cover also the Rust part. I gonna ask a comment from you on that.	2023-09-07T17:40:31	[]	[]
emmett-framework/granian	emmett-framework__granian-61	ac12149b271539040d8a85bc80c83667eb712c1a	diff --git a/Cargo.lock b/Cargo.lock index 12cb1e53..b188b290 100644 --- a/Cargo.lock +++ b/Cargo.lock @@ -566,6 +566,8 @@ dependencies = [ [[package]] name = "pyo3-asyncio" version = "0.17.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "1febe3946b26194628f00526929ee6f8559f9e807f811257e94d4c456103be0e" dependencies = [ "futures", "once_cell", diff --git a/Cargo.toml b/Cargo.toml index 1654d752..5601986b 100644 --- a/Cargo.toml +++ b/Cargo.toml @@ -36,7 +36,7 @@ mimalloc = { version = "0.1.34", default-features = false, features = ["local_dy once_cell = "1.5" pin-project = "1.0" pyo3 = { version = "=0.17", features = ["extension-module"] } -pyo3-asyncio = { path = "lib/pyo3-asyncio", version = "0.17", features = ["tokio-runtime"] } +pyo3-asyncio = { version = "=0.17", features = ["tokio-runtime"] } pyo3-log = "=0.7" rustls-pemfile = "1.0" socket2 = { version = "0.4", features = ["all"] } diff --git a/lib/pyo3-asyncio/.githooks/pre-commit b/lib/pyo3-asyncio/.githooks/pre-commit deleted file mode 100755 index a576dfe1..00000000 --- a/lib/pyo3-asyncio/.githooks/pre-commit +++ /dev/null @@ -1,3 +0,0 @@ -#!/bin/bash - -cargo check --all-targets --all-features \ No newline at end of file diff --git a/lib/pyo3-asyncio/.githooks/pre-push b/lib/pyo3-asyncio/.githooks/pre-push deleted file mode 100755 index 19c4269e..00000000 --- a/lib/pyo3-asyncio/.githooks/pre-push +++ /dev/null @@ -1,4 +0,0 @@ -#!/bin/bash - -make clippy -make test \ No newline at end of file diff --git a/lib/pyo3-asyncio/.github/dependabot.yml b/lib/pyo3-asyncio/.github/dependabot.yml deleted file mode 100644 index c1ce3c3b..00000000 --- a/lib/pyo3-asyncio/.github/dependabot.yml +++ /dev/null @@ -1,11 +0,0 @@ -# To get started with Dependabot version updates, you'll need to specify which -# package ecosystems to update and where the package manifests are located. -# Please see the documentation for all configuration options: -# https://help.github.com/github/administering-a-repository/configuration-options-for-dependency-updates - -version: 2 -updates: - - package-ecosystem: "cargo" # See documentation for possible values - directory: "/" # Location of package manifests - schedule: - interval: "weekly" diff --git a/lib/pyo3-asyncio/.github/issue_template.md b/lib/pyo3-asyncio/.github/issue_template.md deleted file mode 100644 index b582e56a..00000000 --- a/lib/pyo3-asyncio/.github/issue_template.md +++ /dev/null @@ -1,18 +0,0 @@ -## 🐛 Bug Reports - -When reporting a bug, please provide the following information. If this is not a bug report you can just discard this template. - -### 🌍 Environment - - - Your operating system and version: - - Your python version: - - How did you install python (e.g. apt or pyenv)? Did you use a virtualenv?: - - Your Rust version (`rustc --version`): - - Your PyO3 version: - - Have you tried using latest PyO3 master (replace `version = "0.x.y"` with `git = "https://github.com/awestlake87/pyo3-asyncio")?`: - -### 💥 Reproducing - -Please provide a [minimal working example](https://stackoverflow.com/help/mcve). This means both the Rust code and the Python. - -Please also write what exact flags are required to reproduce your results. diff --git a/lib/pyo3-asyncio/.github/pull_request_template.md b/lib/pyo3-asyncio/.github/pull_request_template.md deleted file mode 100644 index 5798dc54..00000000 --- a/lib/pyo3-asyncio/.github/pull_request_template.md +++ /dev/null @@ -1,12 +0,0 @@ -Thank you for contributing to pyo3-asyncio! - -Please consider adding the following to your pull request: - - an entry in CHANGELOG.md - - docs to all new functions and / or detail in the guide - - tests for all new or changed functions - -Be aware the CI pipeline will check your pull request for the following: - - Rust tests (Just `cargo test`) - - Rust lints (`make clippy`) - - Rust formatting (`cargo fmt`) - - Python formatting (`black . --check`. You can install black with `pip install black`) diff --git a/lib/pyo3-asyncio/.github/workflows/ci.yml b/lib/pyo3-asyncio/.github/workflows/ci.yml deleted file mode 100644 index d828197d..00000000 --- a/lib/pyo3-asyncio/.github/workflows/ci.yml +++ /dev/null @@ -1,199 +0,0 @@ -name: CI - -on: - push: - branches: - - master - pull_request: - branches: - - master - -env: - CARGO_TERM_COLOR: always - -jobs: - fmt: - runs-on: ubuntu-latest - steps: - - uses: actions/checkout@v2 - - uses: actions/setup-python@v2 - - run: pip install black==22.8.0 - - uses: actions-rs/toolchain@v1 - with: - toolchain: stable - profile: minimal - components: rustfmt - - name: Check python formatting (black) - run: black --check . - - name: Check rust formatting (rustfmt) - run: cargo fmt --all -- --check - - clippy: - runs-on: ubuntu-latest - steps: - - uses: actions/checkout@v2 - - uses: actions-rs/toolchain@v1 - with: - toolchain: stable - profile: minimal - components: clippy - - run: make clippy - - build: - needs: [fmt] # don't wait for clippy as fails rarely and takes longer - name: python${{ matrix.python-version }}-${{ matrix.platform.python-architecture }} ${{ matrix.platform.os }} ${{ matrix.msrv }} - runs-on: ${{ matrix.platform.os }} - strategy: - fail-fast: false # If one platform fails, allow the rest to keep testing. - matrix: - rust: [stable] - python-version: - [ - "3.7", - "3.8", - "3.9", - "3.10", - "3.11-dev", - "pypy-3.7", - "pypy-3.8", - "pypy-3.9", - ] - platform: - [ - { - os: "macOS-latest", - python-architecture: "x64", - rust-target: "x86_64-apple-darwin", - }, - { - os: "ubuntu-latest", - python-architecture: "x64", - rust-target: "x86_64-unknown-linux-gnu", - }, - { - os: "windows-latest", - python-architecture: "x64", - rust-target: "x86_64-pc-windows-msvc", - }, - { - os: "windows-latest", - python-architecture: "x86", - rust-target: "i686-pc-windows-msvc", - }, - ] - exclude: - # PyPy doesn't release 32-bit Windows builds any more - - python-version: pypy-3.7 - platform: { os: "windows-latest", python-architecture: "x86" } - - python-version: pypy-3.8 - platform: { os: "windows-latest", python-architecture: "x86" } - - python-version: pypy-3.9 - platform: { os: "windows-latest", python-architecture: "x86" } - include: - # Test minimal supported Rust version - - rust: 1.48.0 - python-version: "3.10" - platform: - { - os: "ubuntu-latest", - python-architecture: "x64", - rust-target: "x86_64-unknown-linux-gnu", - } - msrv: "MSRV" - - # Test the `nightly` feature - - rust: nightly - python-version: "3.10" - platform: - { - os: "ubuntu-latest", - python-architecture: "x64", - rust-target: "x86_64-unknown-linux-gnu", - } - msrv: "nightly" - extra_features: "nightly" - - steps: - - uses: actions/checkout@v2 - - - name: Set up Python ${{ matrix.python-version }} - uses: actions/setup-python@v2 - with: - python-version: ${{ matrix.python-version }} - architecture: ${{ matrix.platform.python-architecture }} - - - name: Install Rust toolchain - uses: actions-rs/toolchain@v1 - with: - toolchain: ${{ matrix.rust }} - target: ${{ matrix.platform.rust-target }} - profile: minimal - default: true - - - if: matrix.platform.os == 'ubuntu-latest' - name: Prepare LD_LIBRARY_PATH (Ubuntu only) - run: echo LD_LIBRARY_PATH=${pythonLocation}/lib >> $GITHUB_ENV - - - if: matrix.msrv == 'MSRV' - name: Prepare minimal package versions (MSRV only) - run: \| - set -x - cargo update -p tokio --precise 1.13.1 - cargo update -p indexmap --precise 1.6.2 - cargo update -p parking_lot:0.12.1 --precise 0.11.2 - cargo update -p async-global-executor --precise 2.2.0 - cargo update -p once_cell --precise 1.14.0 - - - name: Build (no features) - run: cargo build --no-default-features --verbose --target ${{ matrix.platform.rust-target }} - - - name: Build - run: cargo build --features=${{env.features}} --verbose --target ${{ matrix.platform.rust-target }} - - # uvloop doesn't compile under Windows, Python 3.11-dev, and PyPy - - if: ${{ matrix.platform.os != 'windows-latest' && matrix.python-version != '3.11-dev' && !startsWith(matrix.python-version, 'pypy') }} - name: Install pyo3-asyncio test dependencies - run: \| - python -m pip install -U uvloop - - - if: ${{ matrix.msrv != 'MSRV' && matrix.python-version != '3.11-dev' && !startsWith(matrix.python-version, 'pypy') }} - name: Test - run: cargo test --all-features --target ${{ matrix.platform.rust-target }} - - - if: ${{ matrix.msrv == 'MSRV' && matrix.python-version != '3.11-dev' && !startsWith(matrix.python-version, 'pypy') }} - name: Test (MSRV, --no-default-features) - run: cargo test --no-default-features --features tokio-runtime,async-std-runtime,attributes,unstable-streams --target ${{ matrix.platform.rust-target }} - - env: - RUST_BACKTRACE: 1 - RUSTFLAGS: "-D warnings" - # TODO: this is a hack to workaround compile_error! warnings about auto-initialize on PyPy - # Once cargo's `resolver = "2"` is stable (~ MSRV Rust 1.52), remove this. - PYO3_CI: 1 - - coverage: - needs: [fmt] - runs-on: ubuntu-latest - steps: - - uses: actions/checkout@v2 - - uses: actions-rs/toolchain@v1 - with: - toolchain: nightly - override: true - profile: minimal - - name: Install pyo3-asyncio test dependencies - run: \| - python -m pip install -U uvloop - - uses: actions-rs/cargo@v1 - with: - command: test - args: --all-features - env: - CARGO_INCREMENTAL: 0 - RUSTFLAGS: "-Zprofile -Ccodegen-units=1 -Cinline-threshold=0 -Clink-dead-code -Coverflow-checks=off" - RUSTDOCFLAGS: "-Zprofile -Ccodegen-units=1 -Cinline-threshold=0 -Clink-dead-code -Coverflow-checks=off" - - uses: actions-rs/[email protected] - id: coverage - - uses: codecov/codecov-action@v1 - with: - file: ${{ steps.coverage.outputs.report }} diff --git a/lib/pyo3-asyncio/.github/workflows/guide.yml b/lib/pyo3-asyncio/.github/workflows/guide.yml deleted file mode 100644 index 461328cd..00000000 --- a/lib/pyo3-asyncio/.github/workflows/guide.yml +++ /dev/null @@ -1,61 +0,0 @@ -name: gh-pages - -on: - push: - branches: - - master - release: - types: [published] - -env: - CARGO_TERM_COLOR: always - -jobs: - deploy: - runs-on: ubuntu-latest - outputs: - tag_name: ${{ steps.prepare_tag.outputs.tag_name }} - steps: - - uses: actions/checkout@v2 - - # This adds the docs to gh-pages-build/doc - - name: Build the doc - run: \| - cargo doc --no-deps --all-features - mkdir -p gh-pages-build - cp -r target/doc gh-pages-build/doc - echo "<meta http-equiv=refresh content=0;url=pyo3_asyncio/index.html>" > gh-pages-build/doc/index.html - - - name: Prepare tag - id: prepare_tag - run: \| - TAG_NAME="${GITHUB_REF##/}" - echo "::set-output name=tag_name::${TAG_NAME}" - - - name: Deploy - uses: peaceiris/[email protected] - with: - github_token: ${{ secrets.GITHUB_TOKEN }} - publish_dir: ./gh-pages-build/ - destination_dir: ${{ steps.prepare_tag.outputs.tag_name }} - full_commit_message: 'Upload documentation for ${{ steps.prepare_tag.outputs.tag_name }}' - - release: - needs: deploy - runs-on: ubuntu-latest - if: ${{ github.event_name == 'release' }} - steps: - - name: Create latest tag redirect - env: - TAG_NAME: ${{ needs.deploy.outputs.tag_name }} - run: \| - mkdir public - echo "<meta http-equiv=refresh content=0;url='$TAG_NAME/'>" > public/index.html - - - name: Deploy - uses: peaceiris/[email protected] - with: - github_token: ${{ secrets.GITHUB_TOKEN }} - publish_dir: ./public/ - full_commit_message: 'Release ${{ needs.deploy.outputs.tag_name }}' - keep_files: true diff --git a/lib/pyo3-asyncio/.gitignore b/lib/pyo3-asyncio/.gitignore deleted file mode 100644 index f2f9e58e..00000000 --- a/lib/pyo3-asyncio/.gitignore +++ /dev/null @@ -1,2 +0,0 @@ -target -Cargo.lock \ No newline at end of file diff --git a/lib/pyo3-asyncio/Cargo.toml b/lib/pyo3-asyncio/Cargo.toml deleted file mode 100644 index 3a612c47..00000000 --- a/lib/pyo3-asyncio/Cargo.toml +++ /dev/null @@ -1,129 +0,0 @@ -[package] -name = "pyo3-asyncio" -description = "PyO3 utilities for Python's Asyncio library" -version = "0.17.0" -authors = ["Andrew J Westlake <[email protected]>"] -readme = "README.md" -keywords = ["pyo3", "python", "ffi", "async", "asyncio"] -homepage = "https://github.com/awestlake87/pyo3-asyncio" -repository = "https://github.com/awestlake87/pyo3-asyncio" -documentation = "https://docs.rs/crate/pyo3-asyncio/" -categories = ["api-bindings", "development-tools::ffi"] -license = "Apache-2.0" -exclude = ["/.gitignore", "/codecov.yml", "/Makefile"] -edition = "2018" - -[features] -async-std-runtime = ["async-std"] -attributes = [] -testing = ["clap", "inventory"] -tokio-runtime = ["tokio"] -unstable-streams = ["async-channel"] -default = [] - -[package.metadata.docs.rs] -features = ["attributes", "testing", "async-std-runtime", "tokio-runtime"] - -[[example]] -name = "async_std" -path = "examples/async_std.rs" -required-features = ["attributes", "async-std-runtime"] - -[[example]] -name = "tokio" -path = "examples/tokio.rs" -required-features = ["attributes", "tokio-runtime"] - -[[example]] -name = "tokio_current_thread" -path = "examples/tokio_current_thread.rs" -required-features = ["attributes", "tokio-runtime"] - -[[example]] -name = "tokio_multi_thread" -path = "examples/tokio_multi_thread.rs" -required-features = ["attributes", "tokio-runtime"] - - -[[test]] -name = "test_async_std_asyncio" -path = "pytests/test_async_std_asyncio.rs" -harness = false -required-features = ["async-std-runtime", "testing", "attributes"] - -[[test]] -name = "test_async_std_run_forever" -path = "pytests/test_async_std_run_forever.rs" -harness = false -required-features = ["async-std-runtime", "testing"] - -[[test]] -name = "test_tokio_current_thread_asyncio" -path = "pytests/test_tokio_current_thread_asyncio.rs" -harness = false -required-features = ["tokio-runtime", "testing", "attributes"] - -[[test]] -name = "test_tokio_current_thread_run_forever" -path = "pytests/test_tokio_current_thread_run_forever.rs" -harness = false -required-features = ["tokio-runtime", "testing"] - -[[test]] -name = "test_tokio_multi_thread_asyncio" -path = "pytests/test_tokio_multi_thread_asyncio.rs" -harness = false -required-features = ["tokio-runtime", "testing", "attributes"] - -[[test]] -name = "test_tokio_multi_thread_run_forever" -path = "pytests/test_tokio_multi_thread_run_forever.rs" -harness = false -required-features = ["tokio-runtime", "testing"] - -[[test]] -name = "test_async_std_uvloop" -path = "pytests/test_async_std_uvloop.rs" -harness = false -required-features = ["async-std-runtime", "testing"] - -[[test]] -name = "test_tokio_current_thread_uvloop" -path = "pytests/test_tokio_current_thread_uvloop.rs" -harness = false -required-features = ["tokio-runtime", "testing"] - -[[test]] -name = "test_tokio_multi_thread_uvloop" -path = "pytests/test_tokio_multi_thread_uvloop.rs" -harness = false -required-features = ["tokio-runtime", "testing"] - - -[[test]] -name = "test_race_condition_regression" -path = "pytests/test_race_condition_regression.rs" -harness = false -required-features = ["async-std-runtime", "testing"] - -[dependencies] -async-channel = { version = "1.6", optional = true } -clap = { version = "3.2", optional = true } -futures = "0.3" -inventory = { version = "0.3", optional = true } -once_cell = "1.14" -pin-project-lite = "0.2" -pyo3 = "0.17" - -[dev-dependencies] -pyo3 = { version = "0.17", features = ["macros"] } - -[dependencies.async-std] -version = "1.12" -features = ["unstable"] -optional = true - -[dependencies.tokio] -version = "1.13" -features = ["full"] -optional = true diff --git a/lib/pyo3-asyncio/Code-of-Conduct.md b/lib/pyo3-asyncio/Code-of-Conduct.md deleted file mode 100644 index c06c6254..00000000 --- a/lib/pyo3-asyncio/Code-of-Conduct.md +++ /dev/null @@ -1,77 +0,0 @@ -# Contributor Covenant Code of Conduct - -## Our Pledge - -In the interest of fostering an open and welcoming environment, we as -contributors and maintainers pledge to making participation in our project and -our community a harassment-free experience for everyone, regardless of age, body -size, disability, ethnicity, sex characteristics, gender identity and expression, -level of experience, education, socio-economic status, nationality, personal -appearance, race, religion, or sexual identity and orientation. - -## Our Standards - -Examples of behavior that contributes to creating a positive environment -include: - - Using welcoming and inclusive language -* Being respectful of differing viewpoints and experiences -* Gracefully accepting constructive criticism -* Focusing on what is best for the community -* Showing empathy towards other community members - -Examples of unacceptable behavior by participants include: - -* The use of sexualized language or imagery and unwelcome sexual attention or - advances -* Trolling, insulting/derogatory comments, and personal or political attacks -* Public or private harassment -* Publishing others' private information, such as a physical or electronic - address, without explicit permission -* Other conduct which could reasonably be considered inappropriate in a - professional setting - -## Our Responsibilities - -Project maintainers are responsible for clarifying the standards of acceptable -behavior and are expected to take appropriate and fair corrective action in -response to any instances of unacceptable behavior. - -Project maintainers have the right and responsibility to remove, edit, or -reject comments, commits, code, wiki edits, issues, and other contributions -that are not aligned to this Code of Conduct, or to ban temporarily or -permanently any contributor for other behaviors that they deem inappropriate, -threatening, offensive, or harmful. - -## Scope - -This Code of Conduct applies both within project spaces and in public spaces -when an individual is representing the project or its community. Examples of -representing a project or community include using an official project e-mail -address, posting via an official social media account, or acting as an appointed -representative at an online or offline event. Representation of a project may be -further defined and clarified by project maintainers. - -## Enforcement - -Instances of abusive, harassing, or otherwise unacceptable behavior may be -reported by contacting the project team. All -complaints will be reviewed and investigated and will result in a response that -is deemed necessary and appropriate to the circumstances. The project team is -obligated to maintain confidentiality with regard to the reporter of an incident. -Further details of specific enforcement policies may be posted separately. - -Project maintainers who do not follow or enforce the Code of Conduct in good -faith may face temporary or permanent repercussions as determined by other -members of the project's leadership. - -## Attribution - -This Code of Conduct is adapted from the [Contributor Covenant][homepage], version 1.4, -available at https://www.contributor-covenant.org/version/1/4/code-of-conduct.html - -[homepage]: https://www.contributor-covenant.org - -For answers to common questions about this code of conduct, see -https://www.contributor-covenant.org/faq - diff --git a/lib/pyo3-asyncio/Contributing.md b/lib/pyo3-asyncio/Contributing.md deleted file mode 100644 index 2dc01cb6..00000000 --- a/lib/pyo3-asyncio/Contributing.md +++ /dev/null @@ -1,52 +0,0 @@ -# Contributing - -Thank you for your interest in contributing to PyO3! All are welcome - please consider reading our [Code of Conduct](Code-of-Conduct.md) to keep our community positive and inclusive. - -If you are searching for ideas how to contribute, please read the "Getting started contributing" section. Once you've found an issue to contribute to, you may find the section "Writing pull requests" helpful. - -## Getting started contributing - -Please join in with any part of PyO3 which interests you. We use Github issues to record all bugs and ideas. Feel free to request an issue to be assigned to you if you want to work on it. - -The following sections also contain specific ideas on where to start contributing to PyO3. - -### Help users identify bugs - -The [PyO3 Gitter channel](https://gitter.im/PyO3/Lobby) is very active with users who are new to PyO3, and often completely new to Rust. Helping them debug is a great way to get experience with the PyO3 codebase. - -Helping others often reveals bugs, documentation weaknesses, and missing APIs. It's a good idea to open Github issues for these immediately so the resolution can be designed and implemented! - -### Review pull requests - -Everybody is welcome to submit comments on open PRs. Please help ensure new PyO3 APIs are safe, performant, tidy, and easy to use! - -## Writing pull requests - -Here are a few things to note when you are writing PRs. - -### Continuous Integration - -The PyO3 Asyncio repo uses Github Actions. PRs are blocked from merging if CI is not successful. - -Formatting, linting and tests are checked for all Rust and Python code. In addition, all warnings in Rust code are disallowed (using `RUSTFLAGS="-D warnings"`). - -Tests run with all supported Python versions with the latest stable Rust compiler, as well as for Python 3.9 with the minimum supported Rust version. - -### Minimum supported Rust version - -PyO3 aims to make use of up-to-date Rust language features to keep the implementation as efficient as possible. - -However, there will always be support for at least the last few Rust compiler versions, so that users have time to update. - -If your PR needs to bump the minimum supported Rust version, this is acceptable with the following conditions: -- Any changes which require a more recent version than what is [currently available on stable Red Hat Enterprise Linux](https://access.redhat.com/documentation/en-us/red_hat_developer_tools/1/) will be postponed. (This is to allow package managers to update support for newer `rustc` versions; RHEL was arbitrarily picked because their update policy is clear.) -- You might be asked to do extra work to tidy up other parts of the PyO3 codebase which can use the compiler version bump :) - - -## Git Hooks (Recommended) - -Using the project's githooks are recommended to prevent CI from failing for trivial reasons such as build checks or integration tests failing. Enabling the hooks should run `cargo check --all-targets` for every commit and `cargo test` for every push. - -``` -git config core.hookspath .githooks -``` \ No newline at end of file diff --git a/lib/pyo3-asyncio/LICENSE b/lib/pyo3-asyncio/LICENSE deleted file mode 100644 index fca31990..00000000 --- a/lib/pyo3-asyncio/LICENSE +++ /dev/null @@ -1,189 +0,0 @@ - Copyright (c) 2017-present PyO3 Project and Contributors. https://github.com/PyO3 - - Licensed under the Apache License, Version 2.0 (the "License"); - you may not use this file except in compliance with the License. - You may obtain a copy of the License at - - http://www.apache.org/licenses/LICENSE-2.0 - - Unless required by applicable law or agreed to in writing, software - distributed under the License is distributed on an "AS IS" BASIS, - WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. - See the License for the specific language governing permissions and - limitations under the License. - - - Apache License - Version 2.0, January 2004 - http://www.apache.org/licenses/ - - TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION - - 1. Definitions. - - "License" shall mean the terms and conditions for use, reproduction, - and distribution as defined by Sections 1 through 9 of this document. - - "Licensor" shall mean the copyright owner or entity authorized by - the copyright owner that is granting the License. - - "Legal Entity" shall mean the union of the acting entity and all - other entities that control, are controlled by, or are under common - control with that entity. For the purposes of this definition, - "control" means (i) the power, direct or indirect, to cause the - direction or management of such entity, whether by contract or - otherwise, or (ii) ownership of fifty percent (50%) or more of the - outstanding shares, or (iii) beneficial ownership of such entity. - - "You" (or "Your") shall mean an individual or Legal Entity - exercising permissions granted by this License. - - "Source" form shall mean the preferred form for making modifications, - including but not limited to software source code, documentation - source, and configuration files. - - "Object" form shall mean any form resulting from mechanical - transformation or translation of a Source form, including but - not limited to compiled object code, generated documentation, - and conversions to other media types. - - "Work" shall mean the work of authorship, whether in Source or - Object form, made available under the License, as indicated by a - copyright notice that is included in or attached to the work - (an example is provided in the Appendix below). - - "Derivative Works" shall mean any work, whether in Source or Object - form, that is based on (or derived from) the Work and for which the - editorial revisions, annotations, elaborations, or other modifications - represent, as a whole, an original work of authorship. For the purposes - of this License, Derivative Works shall not include works that remain - separable from, or merely link (or bind by name) to the interfaces of, - the Work and Derivative Works thereof. - - "Contribution" shall mean any work of authorship, including - the original version of the Work and any modifications or additions - to that Work or Derivative Works thereof, that is intentionally - submitted to Licensor for inclusion in the Work by the copyright owner - or by an individual or Legal Entity authorized to submit on behalf of - the copyright owner. For the purposes of this definition, "submitted" - means any form of electronic, verbal, or written communication sent - to the Licensor or its representatives, including but not limited to - communication on electronic mailing lists, source code control systems, - and issue tracking systems that are managed by, or on behalf of, the - Licensor for the purpose of discussing and improving the Work, but - excluding communication that is conspicuously marked or otherwise - designated in writing by the copyright owner as "Not a Contribution." - - "Contributor" shall mean Licensor and any individual or Legal Entity - on behalf of whom a Contribution has been received by Licensor and - subsequently incorporated within the Work. - - 2. Grant of Copyright License. Subject to the terms and conditions of - this License, each Contributor hereby grants to You a perpetual, - worldwide, non-exclusive, no-charge, royalty-free, irrevocable - copyright license to reproduce, prepare Derivative Works of, - publicly display, publicly perform, sublicense, and distribute the - Work and such Derivative Works in Source or Object form. - - 3. Grant of Patent License. Subject to the terms and conditions of - this License, each Contributor hereby grants to You a perpetual, - worldwide, non-exclusive, no-charge, royalty-free, irrevocable - (except as stated in this section) patent license to make, have made, - use, offer to sell, sell, import, and otherwise transfer the Work, - where such license applies only to those patent claims licensable - by such Contributor that are necessarily infringed by their - Contribution(s) alone or by combination of their Contribution(s) - with the Work to which such Contribution(s) was submitted. If You - institute patent litigation against any entity (including a - cross-claim or counterclaim in a lawsuit) alleging that the Work - or a Contribution incorporated within the Work constitutes direct - or contributory patent infringement, then any patent licenses - granted to You under this License for that Work shall terminate - as of the date such litigation is filed. - - 4. Redistribution. You may reproduce and distribute copies of the - Work or Derivative Works thereof in any medium, with or without - modifications, and in Source or Object form, provided that You - meet the following conditions: - - (a) You must give any other recipients of the Work or - Derivative Works a copy of this License; and - - (b) You must cause any modified files to carry prominent notices - stating that You changed the files; and - - (c) You must retain, in the Source form of any Derivative Works - that You distribute, all copyright, patent, trademark, and - attribution notices from the Source form of the Work, - excluding those notices that do not pertain to any part of - the Derivative Works; and - - (d) If the Work includes a "NOTICE" text file as part of its - distribution, then any Derivative Works that You distribute must - include a readable copy of the attribution notices contained - within such NOTICE file, excluding those notices that do not - pertain to any part of the Derivative Works, in at least one - of the following places: within a NOTICE text file distributed - as part of the Derivative Works; within the Source form or - documentation, if provided along with the Derivative Works; or, - within a display generated by the Derivative Works, if and - wherever such third-party notices normally appear. The contents - of the NOTICE file are for informational purposes only and - do not modify the License. You may add Your own attribution - notices within Derivative Works that You distribute, alongside - or as an addendum to the NOTICE text from the Work, provided - that such additional attribution notices cannot be construed - as modifying the License. - - You may add Your own copyright statement to Your modifications and - may provide additional or different license terms and conditions - for use, reproduction, or distribution of Your modifications, or - for any such Derivative Works as a whole, provided Your use, - reproduction, and distribution of the Work otherwise complies with - the conditions stated in this License. - - 5. Submission of Contributions. Unless You explicitly state otherwise, - any Contribution intentionally submitted for inclusion in the Work - by You to the Licensor shall be under the terms and conditions of - this License, without any additional terms or conditions. - Notwithstanding the above, nothing herein shall supersede or modify - the terms of any separate license agreement you may have executed - with Licensor regarding such Contributions. - - 6. Trademarks. This License does not grant permission to use the trade - names, trademarks, service marks, or product names of the Licensor, - except as required for reasonable and customary use in describing the - origin of the Work and reproducing the content of the NOTICE file. - - 7. Disclaimer of Warranty. Unless required by applicable law or - agreed to in writing, Licensor provides the Work (and each - Contributor provides its Contributions) on an "AS IS" BASIS, - WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or - implied, including, without limitation, any warranties or conditions - of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A - PARTICULAR PURPOSE. You are solely responsible for determining the - appropriateness of using or redistributing the Work and assume any - risks associated with Your exercise of permissions under this License. - - 8. Limitation of Liability. In no event and under no legal theory, - whether in tort (including negligence), contract, or otherwise, - unless required by applicable law (such as deliberate and grossly - negligent acts) or agreed to in writing, shall any Contributor be - liable to You for damages, including any direct, indirect, special, - incidental, or consequential damages of any character arising as a - result of this License or out of the use or inability to use the - Work (including but not limited to damages for loss of goodwill, - work stoppage, computer failure or malfunction, or any and all - other commercial damages or losses), even if such Contributor - has been advised of the possibility of such damages. - - 9. Accepting Warranty or Additional Liability. While redistributing - the Work or Derivative Works thereof, You may choose to offer, - and charge a fee for, acceptance of support, warranty, indemnity, - or other liability obligations and/or rights consistent with this - License. However, in accepting such obligations, You may act only - on Your own behalf and on Your sole responsibility, not on behalf - of any other Contributor, and only if You agree to indemnify, - defend, and hold each Contributor harmless for any liability - incurred by, or claims asserted against, such Contributor by reason - of your accepting any such warranty or additional liability. diff --git a/lib/pyo3-asyncio/Makefile b/lib/pyo3-asyncio/Makefile deleted file mode 100644 index 63fe41b2..00000000 --- a/lib/pyo3-asyncio/Makefile +++ /dev/null @@ -1,24 +0,0 @@ - -clippy: - @touch src/lib.rs # Touching file to ensure that cargo clippy will re-check the project - cargo clippy --tests -- -Dwarnings - cargo clippy --tests -- -Dwarnings - # for example in examples/; do cargo clippy --manifest-path $$example/Cargo.toml -- -Dwarnings \|\| exit 1; done - -fmt: - cargo fmt --all -- --check - -lint: fmt clippy - @true - -test: lint - cargo test --all-features - -test-feature-powerset: lint - cargo install cargo-hack - cargo hack test --feature-powerset - -publish: test-feature-powerset - cargo publish --manifest-path pyo3-asyncio-macros/Cargo.toml - sleep 30 # wait for crates.io to update - cargo publish diff --git a/lib/pyo3-asyncio/README.md b/lib/pyo3-asyncio/README.md deleted file mode 100644 index c15e6c69..00000000 --- a/lib/pyo3-asyncio/README.md +++ /dev/null @@ -1,703 +0,0 @@ -# PyO3 Asyncio - -[![Actions Status](https://github.com/awestlake87/pyo3-asyncio/workflows/CI/badge.svg)](https://github.com/awestlake87/pyo3-asyncio/actions) -[![codecov](https://codecov.io/gh/awestlake87/pyo3-asyncio/branch/master/graph/badge.svg)](https://codecov.io/gh/awestlake87/pyo3-asyncio) -[![crates.io](https://img.shields.io/crates/v/pyo3-asyncio)](https://crates.io/crates/pyo3-asyncio) -[![minimum rustc 1.48](https://img.shields.io/badge/rustc-1.48+-blue.svg)](https://rust-lang.github.io/rfcs/2495-min-rust-version.html) - -[Rust](http://www.rust-lang.org/) bindings for [Python](https://www.python.org/)'s [Asyncio Library](https://docs.python.org/3/library/asyncio.html). This crate facilitates interactions between Rust Futures and Python Coroutines and manages the lifecycle of their corresponding event loops. - -- PyO3 Project: [Homepage](https://pyo3.rs/) \| [GitHub](https://github.com/PyO3/pyo3) - -- PyO3 Asyncio API Documentation: [stable](https://docs.rs/pyo3-asyncio/) \| [master](https://awestlake87.github.io/pyo3-asyncio/master/doc) - -- Guide for Async / Await [stable](https://pyo3.rs/v0.13.2/ecosystem/async-await.html) \| [main](https://pyo3.rs/main/ecosystem/async-await.html) - -- Contributing Notes: [github](https://github.com/awestlake87/pyo3-asyncio/blob/master/Contributing.md) - -> PyO3 Asyncio is a _brand new_ part of the broader PyO3 ecosystem. Feel free to open any issues for feature requests or bugfixes for this crate. - -If you're a new-comer, the best way to get started is to read through the primer below! For `v0.13` and `v0.14` users I highly recommend reading through the [migration section](#migration-guide) to get a general idea of what's changed in `v0.14` and `v0.15`.* - -## Usage - -Like PyO3, PyO3 Asyncio supports the following software versions: - -- Python 3.7 and up (CPython and PyPy) -- Rust 1.48 and up - -## PyO3 Asyncio Primer - -If you are working with a Python library that makes use of async functions or wish to provide -Python bindings for an async Rust library, [`pyo3-asyncio`](https://github.com/awestlake87/pyo3-asyncio) -likely has the tools you need. It provides conversions between async functions in both Python and -Rust and was designed with first-class support for popular Rust runtimes such as -[`tokio`](https://tokio.rs/) and [`async-std`](https://async.rs/). In addition, all async Python -code runs on the default `asyncio` event loop, so `pyo3-asyncio` should work just fine with existing -Python libraries. - -In the following sections, we'll give a general overview of `pyo3-asyncio` explaining how to call -async Python functions with PyO3, how to call async Rust functions from Python, and how to configure -your codebase to manage the runtimes of both. - -### Quickstart - -Here are some examples to get you started right away! A more detailed breakdown -of the concepts in these examples can be found in the following sections. - -#### Rust Applications - -Here we initialize the runtime, import Python's `asyncio` library and run the given future to completion using Python's default `EventLoop` and `async-std`. Inside the future, we convert `asyncio` sleep into a Rust future and await it. - -```toml -# Cargo.toml dependencies -[dependencies] -pyo3 = { version = "0.17" } -pyo3-asyncio = { version = "0.17", features = ["attributes", "async-std-runtime"] } -async-std = "1.9" -``` - -```rust -//! main.rs - -use pyo3::prelude::; - -#[pyo3_asyncio::async_std::main] -async fn main() -> PyResult<()> { - let fut = Python::with_gil(\|py\| { - let asyncio = py.import("asyncio")?; - // convert asyncio.sleep into a Rust Future - pyo3_asyncio::async_std::into_future(asyncio.call_method1("sleep", (1.into_py(py),))?) - })?; - - fut.await?; - - Ok(()) -} -``` - -The same application can be written to use `tokio` instead using the `#[pyo3_asyncio::tokio::main]` -attribute. - -```toml -# Cargo.toml dependencies -[dependencies] -pyo3 = { version = "0.17" } -pyo3-asyncio = { version = "0.17", features = ["attributes", "tokio-runtime"] } -tokio = "1.9" -``` - -```rust -//! main.rs - -use pyo3::prelude::; - -#[pyo3_asyncio::tokio::main] -async fn main() -> PyResult<()> { - let fut = Python::with_gil(\|py\| { - let asyncio = py.import("asyncio")?; - // convert asyncio.sleep into a Rust Future - pyo3_asyncio::tokio::into_future(asyncio.call_method1("sleep", (1.into_py(py),))?) - })?; - - fut.await?; - - Ok(()) -} -``` - -More details on the usage of this library can be found in the [API docs](https://awestlake87.github.io/pyo3-asyncio/master/doc) and the primer below. - -#### PyO3 Native Rust Modules - -PyO3 Asyncio can also be used to write native modules with async functions. - -Add the `[lib]` section to `Cargo.toml` to make your library a `cdylib` that Python can import. - -```toml -[lib] -name = "my_async_module" -crate-type = ["cdylib"] -``` - -Make your project depend on `pyo3` with the `extension-module` feature enabled and select your -`pyo3-asyncio` runtime: - -For `async-std`: - -```toml -[dependencies] -pyo3 = { version = "0.17", features = ["extension-module"] } -pyo3-asyncio = { version = "0.17", features = ["async-std-runtime"] } -async-std = "1.9" -``` - -For `tokio`: - -```toml -[dependencies] -pyo3 = { version = "0.17", features = ["extension-module"] } -pyo3-asyncio = { version = "0.17", features = ["tokio-runtime"] } -tokio = "1.9" -``` - -Export an async function that makes use of `async-std`: - -```rust -//! lib.rs - -use pyo3::{prelude::, wrap_pyfunction}; - -#[pyfunction] -fn rust_sleep(py: Python) -> PyResult<&PyAny> { - pyo3_asyncio::async_std::future_into_py(py, async { - async_std::task::sleep(std::time::Duration::from_secs(1)).await; - Ok(()) - }) -} - -#[pymodule] -fn my_async_module(py: Python, m: &PyModule) -> PyResult<()> { - m.add_function(wrap_pyfunction!(rust_sleep, m)?)?; - - Ok(()) -} - -``` - -If you want to use `tokio` instead, here's what your module should look like: - -```rust -//! lib.rs - -use pyo3::{prelude::, wrap_pyfunction}; - -#[pyfunction] -fn rust_sleep(py: Python) -> PyResult<&PyAny> { - pyo3_asyncio::tokio::future_into_py(py, async { - tokio::time::sleep(std::time::Duration::from_secs(1)).await; - Ok(()) - }) -} - -#[pymodule] -fn my_async_module(py: Python, m: &PyModule) -> PyResult<()> { - m.add_function(wrap_pyfunction!(rust_sleep, m)?)?; - Ok(()) -} -``` - -You can build your module with maturin (see the [Using Rust in Python](https://pyo3.rs/main/#using-rust-from-python) section in the PyO3 guide for setup instructions). After that you should be able to run the Python REPL to try it out. - -```bash -maturin develop && python3 -🔗 Found pyo3 bindings -🐍 Found CPython 3.8 at python3 - Finished dev [unoptimized + debuginfo] target(s) in 0.04s -Python 3.8.5 (default, Jan 27 2021, 15:41:15) -[GCC 9.3.0] on linux -Type "help", "copyright", "credits" or "license" for more information. ->>> import asyncio ->>> ->>> from my_async_module import rust_sleep ->>> ->>> async def main(): ->>> await rust_sleep() ->>> ->>> # should sleep for 1s ->>> asyncio.run(main()) ->>> -``` - -### Awaiting an Async Python Function in Rust - -Let's take a look at a dead simple async Python function: - -```python -# Sleep for 1 second -async def py_sleep(): - await asyncio.sleep(1) -``` - -Async functions in Python are simply functions that return a `coroutine` object. For our purposes, -we really don't need to know much about these `coroutine` objects. The key factor here is that calling -an `async` function is _just like calling a regular function_, the only difference is that we have -to do something special with the object that it returns. - -Normally in Python, that something special is the `await` keyword, but in order to await this -coroutine in Rust, we first need to convert it into Rust's version of a `coroutine`: a `Future`. -That's where `pyo3-asyncio` comes in. -[`pyo3_asyncio::into_future`](https://docs.rs/pyo3-asyncio/latest/pyo3_asyncio/fn.into_future.html) -performs this conversion for us: - -```rust no_run -use pyo3::prelude::; - -#[pyo3_asyncio::tokio::main] -async fn main() -> PyResult<()> { - let future = Python::with_gil(\|py\| -> PyResult<_> { - // import the module containing the py_sleep function - let example = py.import("example")?; - - // calling the py_sleep method like a normal function - // returns a coroutine - let coroutine = example.call_method0("py_sleep")?; - - // convert the coroutine into a Rust future using the - // tokio runtime - pyo3_asyncio::tokio::into_future(coroutine) - })?; - - // await the future - future.await?; - - Ok(()) -} -``` - -> If you're interested in learning more about `coroutines` and `awaitables` in general, check out the -> [Python 3 `asyncio` docs](https://docs.python.org/3/library/asyncio-task.html) for more information. - -### Awaiting a Rust Future in Python - -Here we have the same async function as before written in Rust using the -[`async-std`](https://async.rs/) runtime: - -```rust -/// Sleep for 1 second -async fn rust_sleep() { - async_std::task::sleep(std::time::Duration::from_secs(1)).await; -} -``` - -Similar to Python, Rust's async functions also return a special object called a -`Future`: - -```rust compile_fail -let future = rust_sleep(); -``` - -We can convert this `Future` object into Python to make it `awaitable`. This tells Python that you -can use the `await` keyword with it. In order to do this, we'll call -[`pyo3_asyncio::async_std::future_into_py`](https://docs.rs/pyo3-asyncio/latest/pyo3_asyncio/async_std/fn.future_into_py.html): - -```rust -use pyo3::prelude::; - -async fn rust_sleep() { - async_std::task::sleep(std::time::Duration::from_secs(1)).await; -} - -#[pyfunction] -fn call_rust_sleep(py: Python) -> PyResult<&PyAny> { - pyo3_asyncio::async_std::future_into_py(py, async move { - rust_sleep().await; - Ok(()) - }) -} -``` - -In Python, we can call this pyo3 function just like any other async function: - -```python -from example import call_rust_sleep - -async def rust_sleep(): - await call_rust_sleep() -``` - -## Managing Event Loops - -Python's event loop requires some special treatment, especially regarding the main thread. Some of -Python's `asyncio` features, like proper signal handling, require control over the main thread, which -doesn't always play well with Rust. - -Luckily, Rust's event loops are pretty flexible and don't _need_ control over the main thread, so in -`pyo3-asyncio`, we decided the best way to handle Rust/Python interop was to just surrender the main -thread to Python and run Rust's event loops in the background. Unfortunately, since most event loop -implementations _prefer_ control over the main thread, this can still make some things awkward. - -### PyO3 Asyncio Initialization - -Because Python needs to control the main thread, we can't use the convenient proc macros from Rust -runtimes to handle the `main` function or `#[test]` functions. Instead, the initialization for PyO3 has to be done from the `main` function and the main -thread must block on [`pyo3_asyncio::run_forever`](https://docs.rs/pyo3-asyncio/latest/pyo3_asyncio/fn.run_forever.html) or [`pyo3_asyncio::async_std::run_until_complete`](https://docs.rs/pyo3-asyncio/latest/pyo3_asyncio/async_std/fn.run_until_complete.html). - -Because we have to block on one of those functions, we can't use [`#[async_std::main]`](https://docs.rs/async-std/latest/async_std/attr.main.html) or [`#[tokio::main]`](https://docs.rs/tokio/1.1.0/tokio/attr.main.html) -since it's not a good idea to make long blocking calls during an async function. - -> Internally, these `#[main]` proc macros are expanded to something like this: -> -> ```rust compile_fail -> fn main() { -> // your async main fn -> async fn _main_impl() { /* ... / } -> Runtime::new().block_on(_main_impl()); -> } -> ``` -> -> Making a long blocking call inside the `Future` that's being driven by `block_on` prevents that -> thread from doing anything else and can spell trouble for some runtimes (also this will actually -> deadlock a single-threaded runtime!). Many runtimes have some sort of `spawn_blocking` mechanism -> that can avoid this problem, but again that's not something we can use here since we need it to -> block on the _main_ thread. - -For this reason, `pyo3-asyncio` provides its own set of proc macros to provide you with this -initialization. These macros are intended to mirror the initialization of `async-std` and `tokio` -while also satisfying the Python runtime's needs. - -Here's a full example of PyO3 initialization with the `async-std` runtime: - -```rust no_run -use pyo3::prelude::; - -#[pyo3_asyncio::async_std::main] -async fn main() -> PyResult<()> { - // PyO3 is initialized - Ready to go - - let fut = Python::with_gil(\|py\| -> PyResult<_> { - let asyncio = py.import("asyncio")?; - - // convert asyncio.sleep into a Rust Future - pyo3_asyncio::async_std::into_future( - asyncio.call_method1("sleep", (1.into_py(py),))? - ) - })?; - - fut.await?; - - Ok(()) -} -``` - -#### A Note About `asyncio.run` - -In Python 3.7+, the recommended way to run a top-level coroutine with `asyncio` -is with `asyncio.run`. In `v0.13` we recommended against using this function due to initialization issues, but in `v0.14` it's perfectly valid to use this function... with a caveat. - -Since our Rust <--> Python conversions require a reference to the Python event loop, this poses a problem. Imagine we have a PyO3 Asyncio module that defines -a `rust_sleep` function like in previous examples. You might rightfully assume that you can call pass this directly into `asyncio.run` like this: - -```python -import asyncio - -from my_async_module import rust_sleep - -asyncio.run(rust_sleep()) -``` - -You might be surprised to find out that this throws an error: - -```bash -Traceback (most recent call last): - File "example.py", line 5, in <module> - asyncio.run(rust_sleep()) -RuntimeError: no running event loop -``` - -What's happening here is that we are calling `rust_sleep` _before_ the future is -actually running on the event loop created by `asyncio.run`. This is counter-intuitive, but expected behaviour, and unfortunately there doesn't seem to be a good way of solving this problem within PyO3 Asyncio itself. - -However, we can make this example work with a simple workaround: - -```python -import asyncio - -from my_async_module import rust_sleep - -# Calling main will just construct the coroutine that later calls rust_sleep. -# - This ensures that rust_sleep will be called when the event loop is running, -# not before. -async def main(): - await rust_sleep() - -# Run the main() coroutine at the top-level instead -asyncio.run(main()) -``` - -#### Non-standard Python Event Loops - -Python allows you to use alternatives to the default `asyncio` event loop. One -popular alternative is `uvloop`. In `v0.13` using non-standard event loops was -a bit of an ordeal, but in `v0.14` it's trivial. - -#### Using `uvloop` in a PyO3 Asyncio Native Extensions - -```toml -# Cargo.toml - -[lib] -name = "my_async_module" -crate-type = ["cdylib"] - -[dependencies] -pyo3 = { version = "0.17", features = ["extension-module"] } -pyo3-asyncio = { version = "0.17", features = ["tokio-runtime"] } -async-std = "1.9" -tokio = "1.9" -``` - -```rust -//! lib.rs - -use pyo3::{prelude::, wrap_pyfunction}; - -#[pyfunction] -fn rust_sleep(py: Python) -> PyResult<&PyAny> { - pyo3_asyncio::tokio::future_into_py(py, async { - tokio::time::sleep(std::time::Duration::from_secs(1)).await; - Ok(()) - }) -} - -#[pymodule] -fn my_async_module(_py: Python, m: &PyModule) -> PyResult<()> { - m.add_function(wrap_pyfunction!(rust_sleep, m)?)?; - - Ok(()) -} -``` - -```bash -$ maturin develop && python3 -🔗 Found pyo3 bindings -🐍 Found CPython 3.8 at python3 - Finished dev [unoptimized + debuginfo] target(s) in 0.04s -Python 3.8.8 (default, Apr 13 2021, 19:58:26) -[GCC 7.3.0] :: Anaconda, Inc. on linux -Type "help", "copyright", "credits" or "license" for more information. ->>> import asyncio ->>> import uvloop ->>> ->>> import my_async_module ->>> ->>> uvloop.install() ->>> ->>> async def main(): -... await my_async_module.rust_sleep() -... ->>> asyncio.run(main()) ->>> -``` - -#### Using `uvloop` in Rust Applications - -Using `uvloop` in Rust applications is a bit trickier, but it's still possible -with relatively few modifications. - -Unfortunately, we can't make use of the `#[pyo3_asyncio::<runtime>::main]` attribute with non-standard event loops. This is because the `#[pyo3_asyncio::<runtime>::main]` proc macro has to interact with the Python -event loop before we can install the `uvloop` policy. - -```toml -[dependencies] -async-std = "1.9" -pyo3 = "0.17" -pyo3-asyncio = { version = "0.17", features = ["async-std-runtime"] } -``` - -```rust no_run -//! main.rs - -use pyo3::{prelude::, types::PyType}; - -fn main() -> PyResult<()> { - pyo3::prepare_freethreaded_python(); - - Python::with_gil(\|py\| { - let uvloop = py.import("uvloop")?; - uvloop.call_method0("install")?; - - // store a reference for the assertion - let uvloop = PyObject::from(uvloop); - - pyo3_asyncio::async_std::run(py, async move { - // verify that we are on a uvloop.Loop - Python::with_gil(\|py\| -> PyResult<()> { - assert!(uvloop - .as_ref(py) - .getattr("Loop")? - .downcast::<PyType>() - .unwrap() - .is_instance(pyo3_asyncio::async_std::get_current_loop(py)?)?); - Ok(()) - })?; - - async_std::task::sleep(std::time::Duration::from_secs(1)).await; - - Ok(()) - }) - }) -} -``` - -### Additional Information - -- Managing event loop references can be tricky with pyo3-asyncio. See [Event Loop References and ContextVars](https://awestlake87.github.io/pyo3-asyncio/master/doc/pyo3_asyncio/#event-loop-references-and-contextvars) in the API docs to get a better intuition for how event loop references are managed in this library. -- Testing pyo3-asyncio libraries and applications requires a custom test harness since Python requires control over the main thread. You can find a testing guide in the [API docs for the `testing` module](https://awestlake87.github.io/pyo3-asyncio/master/doc/pyo3_asyncio/testing) - -## Migration Guide - -### Migrating from 0.13 to 0.14 - -So what's changed from `v0.13` to `v0.14`? - -Well, a lot actually. There were some pretty major flaws in the initialization behaviour of `v0.13`. While it would have been nicer to address these issues without changing the public API, I decided it'd be better to break some of the old API rather than completely change the underlying behaviour of the existing functions. I realize this is going to be a bit of a headache, so hopefully this section will help you through it. - -To make things a bit easier, I decided to keep most of the old API alongside the new one (with some deprecation warnings to encourage users to move away from it). It should be possible to use the `v0.13` API alongside the newer `v0.14` API, which should allow you to upgrade your application piecemeal rather than all at once. - -Before you get started, I personally recommend taking a look at [Event Loop References and ContextVars](https://awestlake87.github.io/pyo3-asyncio/master/doc/pyo3_asyncio/#event-loop-references-and-contextvars) in order to get a better grasp on the motivation behind these changes and the nuance involved in using the new conversions. - -### 0.14 Highlights - -- Tokio initialization is now lazy. - - No configuration necessary if you're using the multithreaded scheduler - - Calls to `pyo3_asyncio::tokio::init_multithread` or `pyo3_asyncio::tokio::init_multithread_once` can just be removed. - - Calls to `pyo3_asyncio::tokio::init_current_thread` or `pyo3_asyncio::tokio::init_current_thread_once` require some special attention. - - Custom runtime configuration is done by passing a `tokio::runtime::Builder` into `pyo3_asyncio::tokio::init` instead of a `tokio::runtime::Runtime` -- A new, more correct set of functions has been added to replace the `v0.13` conversions. - - `pyo3_asyncio::into_future_with_loop` - - `pyo3_asyncio::<runtime>::future_into_py_with_loop` - - `pyo3_asyncio::<runtime>::local_future_into_py_with_loop` - - `pyo3_asyncio::<runtime>::into_future` - - `pyo3_asyncio::<runtime>::future_into_py` - - `pyo3_asyncio::<runtime>::local_future_into_py` - - `pyo3_asyncio::<runtime>::get_current_loop` -- `pyo3_asyncio::try_init` is no longer required if you're only using `0.14` conversions -- The `ThreadPoolExecutor` is no longer configured automatically at the start. - - Fortunately, this doesn't seem to have much effect on `v0.13` code, it just means that it's now possible to configure the executor manually as you see fit. - -### Upgrading Your Code to 0.14 - -1. Fix PyO3 0.14 initialization. - - PyO3 0.14 feature gated its automatic initialization behaviour behind "auto-initialize". You can either enable the "auto-initialize" behaviour in your project or add a call to `pyo3::prepare_freethreaded_python()` to the start of your program. - - If you're using the `#[pyo3_asyncio::<runtime>::main]` proc macro attributes, then you can skip this step. `#[pyo3_asyncio::<runtime>::main]` will call `pyo3::prepare_freethreaded_python()` at the start regardless of your project's "auto-initialize" feature. -2. Fix the tokio initialization. - - - Calls to `pyo3_asyncio::tokio::init_multithread` or `pyo3_asyncio::tokio::init_multithread_once` can just be removed. - - If you're using the current thread scheduler, you'll need to manually spawn the thread that it runs on during initialization: - - ```rust no_run - let mut builder = tokio::runtime::Builder::new_current_thread(); - builder.enable_all(); - - pyo3_asyncio::tokio::init(builder); - std::thread::spawn(move \|\| { - pyo3_asyncio::tokio::get_runtime().block_on( - futures::future::pending::<()>() - ); - }); - ``` - - - Custom `tokio::runtime::Builder` configs can be passed into `pyo3_asyncio::tokio::init`. The `tokio::runtime::Runtime` will be lazily instantiated on the first call to `pyo3_asyncio::tokio::get_runtime()` - -3. If you're using `pyo3_asyncio::run_forever` in your application, you should switch to a more manual approach. - - > `run_forever` is not the recommended way of running an event loop in Python, so it might be a good idea to move away from it. This function would have needed to change for `0.14`, but since it's considered an edge case, it was decided that users could just manually call it if they need to. - - ```rust - use pyo3::prelude::; - - fn main() -> PyResult<()> { - pyo3::prepare_freethreaded_python(); - - Python::with_gil(\|py\| { - let asyncio = py.import("asyncio")?; - - let event_loop = asyncio.call_method0("new_event_loop")?; - asyncio.call_method1("set_event_loop", (event_loop,))?; - - let event_loop_hdl = PyObject::from(event_loop); - - pyo3_asyncio::tokio::get_runtime().spawn(async move { - tokio::time::sleep(std::time::Duration::from_secs(1)).await; - - // Stop the event loop manually - Python::with_gil(\|py\| { - event_loop_hdl - .as_ref(py) - .call_method1( - "call_soon_threadsafe", - (event_loop_hdl - .as_ref(py) - .getattr("stop") - .unwrap(),), - ) - .unwrap(); - }) - }); - - event_loop.call_method0("run_forever")?; - Ok(()) - }) - } - ``` - -4. Replace conversions with their newer counterparts. - > You may encounter some issues regarding the usage of `get_running_loop` vs `get_event_loop`. For more details on these newer conversions and how they should be used see [Event Loop References and ContextVars](https://awestlake87.github.io/pyo3-asyncio/master/doc/pyo3_asyncio/#event-loop-references-and-contextvars). - - Replace `pyo3_asyncio::into_future` with `pyo3_asyncio::<runtime>::into_future` - - Replace `pyo3_asyncio::<runtime>::into_coroutine` with `pyo3_asyncio::<runtime>::future_into_py` - - Replace `pyo3_asyncio::get_event_loop` with `pyo3_asyncio::<runtime>::get_current_loop` -5. After all conversions have been replaced with their `v0.14` counterparts, `pyo3_asyncio::try_init` can safely be removed. - -> The `v0.13` API has been removed in version `v0.15` - -### Migrating from 0.14 to 0.15+ - -There have been a few changes to the API in order to support proper cancellation from Python and the `contextvars` module. - -- Any instance of `cancellable_future_into_py` and `local_cancellable_future_into_py` conversions can be replaced with their`future_into_py` and `local_future_into_py` counterparts. - > Cancellation support became the default behaviour in 0.15. -- Instances of `_with_loop` conversions should be replaced with the newer `_with_locals` conversions. - - ```rust no_run - use pyo3::prelude::; - - Python::with_gil(\|py\| -> PyResult<()> { - - // _with_loop conversions in 0.14 - // - // let event_loop = pyo3_asyncio::get_running_loop(py)?; - // - // let fut = pyo3_asyncio::tokio::future_into_py_with_loop( - // event_loop, - // async move { Ok(Python::with_gil(\|py\| py.None())) } - // )?; - // - // should be replaced with _with_locals in 0.15+ - let fut = pyo3_asyncio::tokio::future_into_py_with_locals( - py, - pyo3_asyncio::tokio::get_current_locals(py)?, - async move { Ok(()) } - )?; - - Ok(()) - }); - ``` - -- `scope` and `scope_local` variants now accept `TaskLocals` instead of `event_loop`. You can usually just replace the `event_loop` with `pyo3_asyncio::TaskLocals::new(event_loop).copy_context(py)?`. -- Return types for `future_into_py`, `future_into_py_with_locals` `local_future_into_py`, and `local_future_into_py_with_locals` are now constrained by the bound `IntoPy<PyObject>` instead of requiring the return type to be `PyObject`. This can make the return types for futures more flexible, but inference can also fail when the concrete type is ambiguous (for example when using `into()`). Sometimes the `into()` can just be removed, -- `run`, and `run_until_complete` can now return any `Send + 'static` value. - -### Migrating from 0.15 to 0.16 - -Actually, not much has changed in the API. I'm happy to say that the PyO3 Asyncio is reaching a -pretty stable point in 0.16. For the most part, 0.16 has been about cleanup and removing deprecated -functions from the API. - -PyO3 0.16 comes with a few API changes of its own, but one of the changes that most impacted PyO3 -Asyncio was it's decision to drop support for Python 3.6. PyO3 Asyncio has been using a few -workarounds / hacks to support the pre-3.7 version of Python's asyncio library that are no longer -necessary. PyO3 Asyncio's underlying implementation is now a bit cleaner because of this. - -PyO3 Asyncio 0.15 included some important fixes to the API in order to add support for proper task -cancellation and allow for the preservation / use of contextvars in Python coroutines. This led to -the deprecation of some 0.14 functions that were used for edge cases in favor of some more correct -versions, and those deprecated functions are now removed from the API in 0.16. - -In addition, with PyO3 Asyncio 0.16, the library now has experimental support for conversions from -Python's async generators into a Rust `Stream`. There are currently two versions `v1` and `v2` with -slightly different performance and type signatures, so I'm hoping to get some feedback on which one -works best for downstream users. Just enable the `unstable-streams` feature and you're good to go! - -> The inverse conversion, Rust `Stream` to Python async generator, may come in a later release if -> requested! diff --git a/lib/pyo3-asyncio/codecov.yml b/lib/pyo3-asyncio/codecov.yml deleted file mode 100644 index 4e2196bd..00000000 --- a/lib/pyo3-asyncio/codecov.yml +++ /dev/null @@ -1,9 +0,0 @@ -comment: off - -coverage: - status: - project: - default: - target: auto - # Allow a tiny drop of overall project coverage in PR to reduce spurious failures. - threshold: 0.25% diff --git a/lib/pyo3-asyncio/examples/async_std.rs b/lib/pyo3-asyncio/examples/async_std.rs deleted file mode 100644 index ba7c8e71..00000000 --- a/lib/pyo3-asyncio/examples/async_std.rs +++ /dev/null @@ -1,17 +0,0 @@ -use pyo3::prelude::; - -#[pyo3_asyncio::async_std::main] -async fn main() -> PyResult<()> { - let fut = Python::with_gil(\|py\| { - let asyncio = py.import("asyncio")?; - - // convert asyncio.sleep into a Rust Future - pyo3_asyncio::async_std::into_future(asyncio.call_method1("sleep", (1.into_py(py),))?) - })?; - - println!("sleeping for 1s"); - fut.await?; - println!("done"); - - Ok(()) -} diff --git a/lib/pyo3-asyncio/examples/tokio.rs b/lib/pyo3-asyncio/examples/tokio.rs deleted file mode 100644 index 794c43a2..00000000 --- a/lib/pyo3-asyncio/examples/tokio.rs +++ /dev/null @@ -1,17 +0,0 @@ -use pyo3::prelude::; - -#[pyo3_asyncio::tokio::main] -async fn main() -> PyResult<()> { - let fut = Python::with_gil(\|py\| { - let asyncio = py.import("asyncio")?; - - // convert asyncio.sleep into a Rust Future - pyo3_asyncio::tokio::into_future(asyncio.call_method1("sleep", (1.into_py(py),))?) - })?; - - println!("sleeping for 1s"); - fut.await?; - println!("done"); - - Ok(()) -} diff --git a/lib/pyo3-asyncio/examples/tokio_current_thread.rs b/lib/pyo3-asyncio/examples/tokio_current_thread.rs deleted file mode 100644 index 5b8819a7..00000000 --- a/lib/pyo3-asyncio/examples/tokio_current_thread.rs +++ /dev/null @@ -1,17 +0,0 @@ -use pyo3::prelude::; - -#[pyo3_asyncio::tokio::main(flavor = "current_thread")] -async fn main() -> PyResult<()> { - let fut = Python::with_gil(\|py\| { - let asyncio = py.import("asyncio")?; - - // convert asyncio.sleep into a Rust Future - pyo3_asyncio::tokio::into_future(asyncio.call_method1("sleep", (1.into_py(py),))?) - })?; - - println!("sleeping for 1s"); - fut.await?; - println!("done"); - - Ok(()) -} diff --git a/lib/pyo3-asyncio/examples/tokio_multi_thread.rs b/lib/pyo3-asyncio/examples/tokio_multi_thread.rs deleted file mode 100644 index 0deb6882..00000000 --- a/lib/pyo3-asyncio/examples/tokio_multi_thread.rs +++ /dev/null @@ -1,17 +0,0 @@ -use pyo3::prelude::; - -#[pyo3_asyncio::tokio::main(flavor = "multi_thread", worker_threads = 10)] -async fn main() -> PyResult<()> { - let fut = Python::with_gil(\|py\| { - let asyncio = py.import("asyncio")?; - - // convert asyncio.sleep into a Rust Future - pyo3_asyncio::tokio::into_future(asyncio.call_method1("sleep", (1.into_py(py),))?) - })?; - - println!("sleeping for 1s"); - fut.await?; - println!("done"); - - Ok(()) -} diff --git a/lib/pyo3-asyncio/src/async_std.rs b/lib/pyo3-asyncio/src/async_std.rs deleted file mode 100644 index 0dbe04f9..00000000 --- a/lib/pyo3-asyncio/src/async_std.rs +++ /dev/null @@ -1,739 +0,0 @@ -//! <span class="module-item stab portability" style="display: inline; border-radius: 3px; padding: 2px; font-size: 80%; line-height: 1.2;"><code>async-std-runtime</code></span> PyO3 Asyncio functions specific to the async-std runtime -//! -//! Items marked with -//! <span -//! class="module-item stab portability" -//! style="display: inline; border-radius: 3px; padding: 2px; font-size: 80%; line-height: 1.2;" -//! ><code>unstable-streams</code></span> -//! are only available when the `unstable-streams` Cargo feature is enabled: -//! -//! ```toml -//! [dependencies.pyo3-asyncio] -//! version = "0.17" -//! features = ["unstable-streams"] -//! ``` - -use std::{any::Any, cell::RefCell, future::Future, panic::AssertUnwindSafe, pin::Pin}; - -use async_std::task; -use futures::FutureExt; -use pyo3::prelude::; - -use crate::{ - generic::{self, ContextExt, JoinError, LocalContextExt, Runtime, SpawnLocalExt}, - TaskLocals, -}; - -/// <span class="module-item stab portability" style="display: inline; border-radius: 3px; padding: 2px; font-size: 80%; line-height: 1.2;"><code>attributes</code></span> -/// re-exports for macros -#[cfg(feature = "attributes")] -pub mod re_exports { - /// re-export spawn_blocking for use in `#[test]` macro without external dependency - pub use async_std::task::spawn_blocking; -} - -/// <span class="module-item stab portability" style="display: inline; border-radius: 3px; padding: 2px; font-size: 80%; line-height: 1.2;"><code>attributes</code></span> Provides the boilerplate for the `async-std` runtime and runs an async fn as main -#[cfg(feature = "attributes")] -pub use pyo3_asyncio_macros::async_std_main as main; - -/// <span class="module-item stab portability" style="display: inline; border-radius: 3px; padding: 2px; font-size: 80%; line-height: 1.2;"><code>attributes</code></span> -/// <span class="module-item stab portability" style="display: inline; border-radius: 3px; padding: 2px; font-size: 80%; line-height: 1.2;"><code>testing</code></span> -/// Registers an `async-std` test with the `pyo3-asyncio` test harness -#[cfg(all(feature = "attributes", feature = "testing"))] -pub use pyo3_asyncio_macros::async_std_test as test; - -struct AsyncStdJoinErr(Box<dyn Any + Send + 'static>); - -impl JoinError for AsyncStdJoinErr { - fn is_panic(&self) -> bool { - true - } -} - -async_std::task_local! { - static TASK_LOCALS: RefCell<Option<TaskLocals>> = RefCell::new(None); -} - -struct AsyncStdRuntime; - -impl Runtime for AsyncStdRuntime { - type JoinError = AsyncStdJoinErr; - type JoinHandle = task::JoinHandle<Result<(), AsyncStdJoinErr>>; - - fn spawn<F>(fut: F) -> Self::JoinHandle - where - F: Future<Output = ()> + Send + 'static, - { - task::spawn(async move { - AssertUnwindSafe(fut) - .catch_unwind() - .await - .map_err(\|e\| AsyncStdJoinErr(e)) - }) - } -} - -impl ContextExt for AsyncStdRuntime { - fn scope<F, R>(locals: TaskLocals, fut: F) -> Pin<Box<dyn Future<Output = R> + Send>> - where - F: Future<Output = R> + Send + 'static, - { - let old = TASK_LOCALS.with(\|c\| c.replace(Some(locals))); - Box::pin(async move { - let result = fut.await; - TASK_LOCALS.with(\|c\| c.replace(old)); - result - }) - } - - fn get_task_locals() -> Option<TaskLocals> { - match TASK_LOCALS.try_with(\|c\| c.borrow().clone()) { - Ok(locals) => locals, - Err(_) => None, - } - } -} - -impl SpawnLocalExt for AsyncStdRuntime { - fn spawn_local<F>(fut: F) -> Self::JoinHandle - where - F: Future<Output = ()> + 'static, - { - task::spawn_local(async move { - fut.await; - Ok(()) - }) - } -} - -impl LocalContextExt for AsyncStdRuntime { - fn scope_local<F, R>(locals: TaskLocals, fut: F) -> Pin<Box<dyn Future<Output = R>>> - where - F: Future<Output = R> + 'static, - { - let old = TASK_LOCALS.with(\|c\| c.replace(Some(locals))); - Box::pin(async move { - let result = fut.await; - TASK_LOCALS.with(\|c\| c.replace(old)); - result - }) - } -} - -/// Set the task local event loop for the given future -pub async fn scope<F, R>(locals: TaskLocals, fut: F) -> R -where - F: Future<Output = R> + Send + 'static, -{ - AsyncStdRuntime::scope(locals, fut).await -} - -/// Set the task local event loop for the given !Send future -pub async fn scope_local<F, R>(locals: TaskLocals, fut: F) -> R -where - F: Future<Output = R> + 'static, -{ - AsyncStdRuntime::scope_local(locals, fut).await -} - -/// Get the current event loop from either Python or Rust async task local context -/// -/// This function first checks if the runtime has a task-local reference to the Python event loop. -/// If not, it calls [`get_running_loop`](`crate::get_running_loop`) to get the event loop -/// associated with the current OS thread. -pub fn get_current_loop(py: Python) -> PyResult<&PyAny> { - generic::get_current_loop::<AsyncStdRuntime>(py) -} - -/// Either copy the task locals from the current task OR get the current running loop and -/// contextvars from Python. -pub fn get_current_locals(py: Python) -> PyResult<TaskLocals> { - generic::get_current_locals::<AsyncStdRuntime>(py) -} - -/// Run the event loop until the given Future completes -/// -/// The event loop runs until the given future is complete. -/// -/// After this function returns, the event loop can be resumed with [`run_until_complete`] -/// -/// # Arguments -/// `event_loop` - The Python event loop that should run the future -/// * `fut` - The future to drive to completion -/// -/// # Examples -/// -/// ``` -/// # use std::time::Duration; -/// # -/// # use pyo3::prelude::; -/// # -/// # pyo3::prepare_freethreaded_python(); -/// # -/// # Python::with_gil(\|py\| -> PyResult<()> { -/// # let event_loop = py.import("asyncio")?.call_method0("new_event_loop")?; -/// pyo3_asyncio::async_std::run_until_complete(event_loop, async move { -/// async_std::task::sleep(Duration::from_secs(1)).await; -/// Ok(()) -/// })?; -/// # Ok(()) -/// # }).unwrap(); -/// ``` -pub fn run_until_complete<F, T>(event_loop: &PyAny, fut: F) -> PyResult<T> -where - F: Future<Output = PyResult<T>> + Send + 'static, - T: Send + Sync + 'static, -{ - generic::run_until_complete::<AsyncStdRuntime, _, T>(event_loop, fut) -} - -/// Run the event loop until the given Future completes -/// -/// # Arguments -/// `py` - The current PyO3 GIL guard -/// * `fut` - The future to drive to completion -/// -/// # Examples -/// -/// ```no_run -/// # use std::time::Duration; -/// # -/// # use pyo3::prelude::; -/// # -/// fn main() { -/// // call this or use pyo3 0.14 "auto-initialize" feature -/// pyo3::prepare_freethreaded_python(); -/// -/// Python::with_gil(\|py\| { -/// pyo3_asyncio::async_std::run(py, async move { -/// async_std::task::sleep(Duration::from_secs(1)).await; -/// Ok(()) -/// }) -/// .map_err(\|e\| { -/// e.print_and_set_sys_last_vars(py); -/// }) -/// .unwrap(); -/// }) -/// } -/// ``` -pub fn run<F, T>(py: Python, fut: F) -> PyResult<T> -where - F: Future<Output = PyResult<T>> + Send + 'static, - T: Send + Sync + 'static, -{ - generic::run::<AsyncStdRuntime, F, T>(py, fut) -} - -/// Convert a Rust Future into a Python awaitable -/// -/// If the `asyncio.Future` returned by this conversion is cancelled via `asyncio.Future.cancel`, -/// the Rust future will be cancelled as well (new behaviour in `v0.15`). -/// -/// Python `contextvars` are preserved when calling async Python functions within the Rust future -/// via [`into_future`] (new behaviour in `v0.15`). -/// -/// > Although `contextvars` are preserved for async Python functions, synchronous functions will -/// unfortunately fail to resolve them when called within the Rust future. This is because the -/// function is being called from a Rust thread, not inside an actual Python coroutine context. -/// > -/// > As a workaround, you can get the `contextvars` from the current task locals using -/// [`get_current_locals`] and [`TaskLocals::context`](`crate::TaskLocals::context`), then wrap your -/// synchronous function in a call to `contextvars.Context.run`. This will set the context, call the -/// synchronous function, and restore the previous context when it returns or raises an exception. -/// -/// # Arguments -/// `py` - PyO3 GIL guard -/// * `locals` - The task locals for the given future -/// * `fut` - The Rust future to be converted -/// -/// # Examples -/// -/// ``` -/// use std::time::Duration; -/// -/// use pyo3::prelude::; -/// -/// /// Awaitable sleep function -/// #[pyfunction] -/// fn sleep_for<'p>(py: Python<'p>, secs: &'p PyAny) -> PyResult<&'p PyAny> { -/// let secs = secs.extract()?; -/// pyo3_asyncio::async_std::future_into_py_with_locals( -/// py, -/// pyo3_asyncio::async_std::get_current_locals(py)?, -/// async move { -/// async_std::task::sleep(Duration::from_secs(secs)).await; -/// Python::with_gil(\|py\| Ok(py.None())) -/// } -/// ) -/// } -/// ``` -pub fn future_into_py_with_locals<F, T>(py: Python, locals: TaskLocals, fut: F) -> PyResult<&PyAny> -where - F: Future<Output = PyResult<T>> + Send + 'static, - T: IntoPy<PyObject>, -{ - generic::future_into_py_with_locals::<AsyncStdRuntime, F, T>(py, locals, fut) -} - -/// Convert a Rust Future into a Python awaitable -/// -/// If the `asyncio.Future` returned by this conversion is cancelled via `asyncio.Future.cancel`, -/// the Rust future will be cancelled as well (new behaviour in `v0.15`). -/// -/// Python `contextvars` are preserved when calling async Python functions within the Rust future -/// via [`into_future`] (new behaviour in `v0.15`). -/// -/// > Although `contextvars` are preserved for async Python functions, synchronous functions will -/// unfortunately fail to resolve them when called within the Rust future. This is because the -/// function is being called from a Rust thread, not inside an actual Python coroutine context. -/// > -/// > As a workaround, you can get the `contextvars` from the current task locals using -/// [`get_current_locals`] and [`TaskLocals::context`](`crate::TaskLocals::context`), then wrap your -/// synchronous function in a call to `contextvars.Context.run`. This will set the context, call the -/// synchronous function, and restore the previous context when it returns or raises an exception. -/// -/// # Arguments -/// `py` - The current PyO3 GIL guard -/// * `fut` - The Rust future to be converted -/// -/// # Examples -/// -/// ``` -/// use std::time::Duration; -/// -/// use pyo3::prelude::; -/// -/// /// Awaitable sleep function -/// #[pyfunction] -/// fn sleep_for<'p>(py: Python<'p>, secs: &'p PyAny) -> PyResult<&'p PyAny> { -/// let secs = secs.extract()?; -/// pyo3_asyncio::async_std::future_into_py(py, async move { -/// async_std::task::sleep(Duration::from_secs(secs)).await; -/// Ok(()) -/// }) -/// } -/// ``` -pub fn future_into_py<F, T>(py: Python, fut: F) -> PyResult<&PyAny> -where - F: Future<Output = PyResult<T>> + Send + 'static, - T: IntoPy<PyObject>, -{ - generic::future_into_py::<AsyncStdRuntime, _, T>(py, fut) -} - -/// Convert a `!Send` Rust Future into a Python awaitable -/// -/// If the `asyncio.Future` returned by this conversion is cancelled via `asyncio.Future.cancel`, -/// the Rust future will be cancelled as well (new behaviour in `v0.15`). -/// -/// Python `contextvars` are preserved when calling async Python functions within the Rust future -/// via [`into_future`] (new behaviour in `v0.15`). -/// -/// > Although `contextvars` are preserved for async Python functions, synchronous functions will -/// unfortunately fail to resolve them when called within the Rust future. This is because the -/// function is being called from a Rust thread, not inside an actual Python coroutine context. -/// > -/// > As a workaround, you can get the `contextvars` from the current task locals using -/// [`get_current_locals`] and [`TaskLocals::context`](`crate::TaskLocals::context`), then wrap your -/// synchronous function in a call to `contextvars.Context.run`. This will set the context, call the -/// synchronous function, and restore the previous context when it returns or raises an exception. -/// -/// # Arguments -/// `py` - PyO3 GIL guard -/// * `locals` - The task locals for the given future -/// * `fut` - The Rust future to be converted -/// -/// # Examples -/// -/// ``` -/// use std::{rc::Rc, time::Duration}; -/// -/// use pyo3::prelude::; -/// -/// /// Awaitable non-send sleep function -/// #[pyfunction] -/// fn sleep_for(py: Python, secs: u64) -> PyResult<&PyAny> { -/// // Rc is non-send so it cannot be passed into pyo3_asyncio::async_std::future_into_py -/// let secs = Rc::new(secs); -/// Ok(pyo3_asyncio::async_std::local_future_into_py_with_locals( -/// py, -/// pyo3_asyncio::async_std::get_current_locals(py)?, -/// async move { -/// async_std::task::sleep(Duration::from_secs(secs)).await; -/// Python::with_gil(\|py\| Ok(py.None())) -/// } -/// )?.into()) -/// } -/// -/// # #[cfg(all(feature = "async-std-runtime", feature = "attributes"))] -/// #[pyo3_asyncio::async_std::main] -/// async fn main() -> PyResult<()> { -/// Python::with_gil(\|py\| { -/// let py_future = sleep_for(py, 1)?; -/// pyo3_asyncio::async_std::into_future(py_future) -/// })? -/// .await?; -/// -/// Ok(()) -/// } -/// # #[cfg(not(all(feature = "async-std-runtime", feature = "attributes")))] -/// # fn main() {} -/// ``` -pub fn local_future_into_py_with_locals<F, T>( - py: Python, - locals: TaskLocals, - fut: F, -) -> PyResult<&PyAny> -where - F: Future<Output = PyResult<T>> + 'static, - T: IntoPy<PyObject>, -{ - generic::local_future_into_py_with_locals::<AsyncStdRuntime, _, T>(py, locals, fut) -} - -/// Convert a `!Send` Rust Future into a Python awaitable -/// -/// If the `asyncio.Future` returned by this conversion is cancelled via `asyncio.Future.cancel`, -/// the Rust future will be cancelled as well (new behaviour in `v0.15`). -/// -/// Python `contextvars` are preserved when calling async Python functions within the Rust future -/// via [`into_future`] (new behaviour in `v0.15`). -/// -/// > Although `contextvars` are preserved for async Python functions, synchronous functions will -/// unfortunately fail to resolve them when called within the Rust future. This is because the -/// function is being called from a Rust thread, not inside an actual Python coroutine context. -/// > -/// > As a workaround, you can get the `contextvars` from the current task locals using -/// [`get_current_locals`] and [`TaskLocals::context`](`crate::TaskLocals::context`), then wrap your -/// synchronous function in a call to `contextvars.Context.run`. This will set the context, call the -/// synchronous function, and restore the previous context when it returns or raises an exception. -/// -/// # Arguments -/// * `py` - The current PyO3 GIL guard -/// * `fut` - The Rust future to be converted -/// -/// # Examples -/// -/// ``` -/// use std::{rc::Rc, time::Duration}; -/// -/// use pyo3::prelude::; -/// -/// /// Awaitable non-send sleep function -/// #[pyfunction] -/// fn sleep_for(py: Python, secs: u64) -> PyResult<&PyAny> { -/// // Rc is non-send so it cannot be passed into pyo3_asyncio::async_std::future_into_py -/// let secs = Rc::new(secs); -/// pyo3_asyncio::async_std::local_future_into_py(py, async move { -/// async_std::task::sleep(Duration::from_secs(secs)).await; -/// Ok(()) -/// }) -/// } -/// -/// # #[cfg(all(feature = "async-std-runtime", feature = "attributes"))] -/// #[pyo3_asyncio::async_std::main] -/// async fn main() -> PyResult<()> { -/// Python::with_gil(\|py\| { -/// let py_future = sleep_for(py, 1)?; -/// pyo3_asyncio::async_std::into_future(py_future) -/// })? -/// .await?; -/// -/// Ok(()) -/// } -/// # #[cfg(not(all(feature = "async-std-runtime", feature = "attributes")))] -/// # fn main() {} -/// ``` -pub fn local_future_into_py<F, T>(py: Python, fut: F) -> PyResult<&PyAny> -where - F: Future<Output = PyResult<T>> + 'static, - T: IntoPy<PyObject>, -{ - generic::local_future_into_py::<AsyncStdRuntime, _, T>(py, fut) -} - -/// Convert a Python `awaitable` into a Rust Future -/// -/// This function converts the `awaitable` into a Python Task using `run_coroutine_threadsafe`. A -/// completion handler sends the result of this Task through a -/// `futures::channel::oneshot::Sender<PyResult<PyObject>>` and the future returned by this function -/// simply awaits the result through the `futures::channel::oneshot::Receiver<PyResult<PyObject>>`. -/// -/// # Arguments -/// * `awaitable` - The Python `awaitable` to be converted -/// -/// # Examples -/// -/// ``` -/// use std::time::Duration; -/// -/// use pyo3::prelude::; -/// -/// const PYTHON_CODE: &'static str = r#" -/// import asyncio -/// -/// async def py_sleep(duration): -/// await asyncio.sleep(duration) -/// "#; -/// -/// async fn py_sleep(seconds: f32) -> PyResult<()> { -/// let test_mod = Python::with_gil(\|py\| -> PyResult<PyObject> { -/// Ok( -/// PyModule::from_code( -/// py, -/// PYTHON_CODE, -/// "test_into_future/test_mod.py", -/// "test_mod" -/// )? -/// .into() -/// ) -/// })?; -/// -/// Python::with_gil(\|py\| { -/// pyo3_asyncio::async_std::into_future( -/// test_mod -/// .call_method1(py, "py_sleep", (seconds.into_py(py),))? -/// .as_ref(py), -/// ) -/// })? -/// .await?; -/// Ok(()) -/// } -/// ``` -pub fn into_future(awaitable: &PyAny) -> PyResult<impl Future<Output = PyResult<PyObject>> + Send> { - generic::into_future::<AsyncStdRuntime>(awaitable) -} - -/// <span class="module-item stab portability" style="display: inline; border-radius: 3px; padding: 2px; font-size: 80%; line-height: 1.2;"><code>unstable-streams</code></span> Convert an async generator into a stream -/// -/// This API is marked as unstable* and is only available when the -/// `unstable-streams` crate feature is enabled. This comes with no -/// stability guarantees, and could be changed or removed at any time. -/// -/// # Arguments -/// * `gen` - The Python async generator to be converted -/// -/// # Examples -/// ``` -/// use pyo3::prelude::; -/// use futures::{StreamExt, TryStreamExt}; -/// -/// const TEST_MOD: &str = r#" -/// import asyncio -/// -/// async def gen(): -/// for i in range(10): -/// await asyncio.sleep(0.1) -/// yield i -/// "#; -/// -/// # #[cfg(all(feature = "unstable-streams", feature = "attributes"))] -/// # #[pyo3_asyncio::async_std::main] -/// # async fn main() -> PyResult<()> { -/// let stream = Python::with_gil(\|py\| { -/// let test_mod = PyModule::from_code( -/// py, -/// TEST_MOD, -/// "test_rust_coroutine/test_mod.py", -/// "test_mod", -/// )?; -/// -/// pyo3_asyncio::async_std::into_stream_v1(test_mod.call_method0("gen")?) -/// })?; -/// -/// let vals = stream -/// .map(\|item\| Python::with_gil(\|py\| -> PyResult<i32> { Ok(item?.as_ref(py).extract()?) })) -/// .try_collect::<Vec<i32>>() -/// .await?; -/// -/// assert_eq!((0..10).collect::<Vec<i32>>(), vals); -/// -/// Ok(()) -/// # } -/// # #[cfg(not(all(feature = "unstable-streams", feature = "attributes")))] -/// # fn main() {} -/// ``` -#[cfg(feature = "unstable-streams")] -pub fn into_stream_v1<'p>( - gen: &'p PyAny, -) -> PyResult<impl futures::Stream<Item = PyResult<PyObject>> + 'static> { - generic::into_stream_v1::<AsyncStdRuntime>(gen) -} - -/// <span class="module-item stab portability" style="display: inline; border-radius: 3px; padding: 2px; font-size: 80%; line-height: 1.2;"><code>unstable-streams</code></span> Convert an async generator into a stream -/// -/// This API is marked as unstable* and is only available when the -/// `unstable-streams` crate feature is enabled. This comes with no -/// stability guarantees, and could be changed or removed at any time. -/// -/// # Arguments -/// * `locals` - The current task locals -/// * `gen` - The Python async generator to be converted -/// -/// # Examples -/// ``` -/// use pyo3::prelude::; -/// use futures::{StreamExt, TryStreamExt}; -/// -/// const TEST_MOD: &str = r#" -/// import asyncio -/// -/// async def gen(): -/// for i in range(10): -/// await asyncio.sleep(0.1) -/// yield i -/// "#; -/// -/// # #[cfg(all(feature = "unstable-streams", feature = "attributes"))] -/// # #[pyo3_asyncio::async_std::main] -/// # async fn main() -> PyResult<()> { -/// let stream = Python::with_gil(\|py\| { -/// let test_mod = PyModule::from_code( -/// py, -/// TEST_MOD, -/// "test_rust_coroutine/test_mod.py", -/// "test_mod", -/// )?; -/// -/// pyo3_asyncio::async_std::into_stream_with_locals_v1( -/// pyo3_asyncio::async_std::get_current_locals(py)?, -/// test_mod.call_method0("gen")? -/// ) -/// })?; -/// -/// let vals = stream -/// .map(\|item\| Python::with_gil(\|py\| -> PyResult<i32> { Ok(item?.as_ref(py).extract()?) })) -/// .try_collect::<Vec<i32>>() -/// .await?; -/// -/// assert_eq!((0..10).collect::<Vec<i32>>(), vals); -/// -/// Ok(()) -/// # } -/// # #[cfg(not(all(feature = "unstable-streams", feature = "attributes")))] -/// # fn main() {} -/// ``` -#[cfg(feature = "unstable-streams")] -pub fn into_stream_with_locals_v1<'p>( - locals: TaskLocals, - gen: &'p PyAny, -) -> PyResult<impl futures::Stream<Item = PyResult<PyObject>> + 'static> { - generic::into_stream_with_locals_v1::<AsyncStdRuntime>(locals, gen) -} - -/// <span class="module-item stab portability" style="display: inline; border-radius: 3px; padding: 2px; font-size: 80%; line-height: 1.2;"><code>unstable-streams</code></span> Convert an async generator into a stream -/// -/// This API is marked as unstable* and is only available when the -/// `unstable-streams` crate feature is enabled. This comes with no -/// stability guarantees, and could be changed or removed at any time. -/// -/// # Arguments -/// * `locals` - The current task locals -/// * `gen` - The Python async generator to be converted -/// -/// # Examples -/// ``` -/// use pyo3::prelude::; -/// use futures::{StreamExt, TryStreamExt}; -/// -/// const TEST_MOD: &str = r#" -/// import asyncio -/// -/// async def gen(): -/// for i in range(10): -/// await asyncio.sleep(0.1) -/// yield i -/// "#; -/// -/// # #[cfg(all(feature = "unstable-streams", feature = "attributes"))] -/// # #[pyo3_asyncio::async_std::main] -/// # async fn main() -> PyResult<()> { -/// let stream = Python::with_gil(\|py\| { -/// let test_mod = PyModule::from_code( -/// py, -/// TEST_MOD, -/// "test_rust_coroutine/test_mod.py", -/// "test_mod", -/// )?; -/// -/// pyo3_asyncio::async_std::into_stream_with_locals_v2( -/// pyo3_asyncio::async_std::get_current_locals(py)?, -/// test_mod.call_method0("gen")? -/// ) -/// })?; -/// -/// let vals = stream -/// .map(\|item\| Python::with_gil(\|py\| -> PyResult<i32> { Ok(item.as_ref(py).extract()?) })) -/// .try_collect::<Vec<i32>>() -/// .await?; -/// -/// assert_eq!((0..10).collect::<Vec<i32>>(), vals); -/// -/// Ok(()) -/// # } -/// # #[cfg(not(all(feature = "unstable-streams", feature = "attributes")))] -/// # fn main() {} -/// ``` -#[cfg(feature = "unstable-streams")] -pub fn into_stream_with_locals_v2<'p>( - locals: TaskLocals, - gen: &'p PyAny, -) -> PyResult<impl futures::Stream<Item = PyObject> + 'static> { - generic::into_stream_with_locals_v2::<AsyncStdRuntime>(locals, gen) -} - -/// <span class="module-item stab portability" style="display: inline; border-radius: 3px; padding: 2px; font-size: 80%; line-height: 1.2;"><code>unstable-streams</code></span> Convert an async generator into a stream -/// -/// This API is marked as unstable* and is only available when the -/// `unstable-streams` crate feature is enabled. This comes with no -/// stability guarantees, and could be changed or removed at any time. -/// -/// # Arguments -/// * `gen` - The Python async generator to be converted -/// -/// # Examples -/// ``` -/// use pyo3::prelude::; -/// use futures::{StreamExt, TryStreamExt}; -/// -/// const TEST_MOD: &str = r#" -/// import asyncio -/// -/// async def gen(): -/// for i in range(10): -/// await asyncio.sleep(0.1) -/// yield i -/// "#; -/// -/// # #[cfg(all(feature = "unstable-streams", feature = "attributes"))] -/// # #[pyo3_asyncio::async_std::main] -/// # async fn main() -> PyResult<()> { -/// let stream = Python::with_gil(\|py\| { -/// let test_mod = PyModule::from_code( -/// py, -/// TEST_MOD, -/// "test_rust_coroutine/test_mod.py", -/// "test_mod", -/// )?; -/// -/// pyo3_asyncio::async_std::into_stream_v2(test_mod.call_method0("gen")?) -/// })?; -/// -/// let vals = stream -/// .map(\|item\| Python::with_gil(\|py\| -> PyResult<i32> { Ok(item.as_ref(py).extract()?) })) -/// .try_collect::<Vec<i32>>() -/// .await?; -/// -/// assert_eq!((0..10).collect::<Vec<i32>>(), vals); -/// -/// Ok(()) -/// # } -/// # #[cfg(not(all(feature = "unstable-streams", feature = "attributes")))] -/// # fn main() {} -/// ``` -#[cfg(feature = "unstable-streams")] -pub fn into_stream_v2<'p>( - gen: &'p PyAny, -) -> PyResult<impl futures::Stream<Item = PyObject> + 'static> { - generic::into_stream_v2::<AsyncStdRuntime>(gen) -} diff --git a/lib/pyo3-asyncio/src/err.rs b/lib/pyo3-asyncio/src/err.rs deleted file mode 100644 index 95e5d152..00000000 --- a/lib/pyo3-asyncio/src/err.rs +++ /dev/null @@ -1,9 +0,0 @@ -// FIXME - is there a way to document custom PyO3 exceptions? -#[allow(missing_docs)] -mod exceptions { - use pyo3::{create_exception, exceptions::PyException}; - - create_exception!(pyo3_asyncio, RustPanic, PyException); -} - -pub use exceptions::RustPanic; diff --git a/lib/pyo3-asyncio/src/generic.rs b/lib/pyo3-asyncio/src/generic.rs deleted file mode 100644 index 0f522b34..00000000 --- a/lib/pyo3-asyncio/src/generic.rs +++ /dev/null @@ -1,1729 +0,0 @@ -//! Generic implementations of PyO3 Asyncio utilities that can be used for any Rust runtime -//! -//! Items marked with -//! <span -//! class="module-item stab portability" -//! style="display: inline; border-radius: 3px; padding: 2px; font-size: 80%; line-height: 1.2;" -//! ><code>unstable-streams</code></span> -//! are only available when the `unstable-streams` Cargo feature is enabled: -//! -//! ```toml -//! [dependencies.pyo3-asyncio] -//! version = "0.17" -//! features = ["unstable-streams"] -//! ``` - -use std::{ - future::Future, - marker::PhantomData, - pin::Pin, - sync::{Arc, Mutex}, - task::{Context, Poll}, -}; - -use futures::{ - channel::{mpsc, oneshot}, - SinkExt, -}; -use once_cell::sync::OnceCell; -use pin_project_lite::pin_project; -use pyo3::prelude::; - -#[allow(deprecated)] -use crate::{ - asyncio, call_soon_threadsafe, close, create_future, dump_err, err::RustPanic, - get_running_loop, into_future_with_locals, TaskLocals, -}; - -/// Generic utilities for a JoinError -pub trait JoinError { - /// Check if the spawned task exited because of a panic - fn is_panic(&self) -> bool; -} - -/// Generic Rust async/await runtime -pub trait Runtime: Send + 'static { - /// The error returned by a JoinHandle after being awaited - type JoinError: JoinError + Send; - /// A future that completes with the result of the spawned task - type JoinHandle: Future<Output = Result<(), Self::JoinError>> + Send; - - /// Spawn a future onto this runtime's event loop - fn spawn<F>(fut: F) -> Self::JoinHandle - where - F: Future<Output = ()> + Send + 'static; -} - -/// Extension trait for async/await runtimes that support spawning local tasks -pub trait SpawnLocalExt: Runtime { - /// Spawn a !Send future onto this runtime's event loop - fn spawn_local<F>(fut: F) -> Self::JoinHandle - where - F: Future<Output = ()> + 'static; -} - -/// Exposes the utilities necessary for using task-local data in the Runtime -pub trait ContextExt: Runtime { - /// Set the task locals for the given future - fn scope<F, R>(locals: TaskLocals, fut: F) -> Pin<Box<dyn Future<Output = R> + Send>> - where - F: Future<Output = R> + Send + 'static; - - /// Get the task locals for the current task - fn get_task_locals() -> Option<TaskLocals>; -} - -/// Adds the ability to scope task-local data for !Send futures -pub trait LocalContextExt: Runtime { - /// Set the task locals for the given !Send future - fn scope_local<F, R>(locals: TaskLocals, fut: F) -> Pin<Box<dyn Future<Output = R>>> - where - F: Future<Output = R> + 'static; -} - -/// Get the current event loop from either Python or Rust async task local context -/// -/// This function first checks if the runtime has a task-local reference to the Python event loop. -/// If not, it calls [`get_running_loop`](crate::get_running_loop`) to get the event loop associated -/// with the current OS thread. -pub fn get_current_loop<R>(py: Python) -> PyResult<&PyAny> -where - R: ContextExt, -{ - if let Some(locals) = R::get_task_locals() { - Ok(locals.event_loop.into_ref(py)) - } else { - get_running_loop(py) - } -} - -/// Either copy the task locals from the current task OR get the current running loop and -/// contextvars from Python. -pub fn get_current_locals<R>(py: Python) -> PyResult<TaskLocals> -where - R: ContextExt, -{ - if let Some(locals) = R::get_task_locals() { - Ok(locals) - } else { - Ok(TaskLocals::with_running_loop(py)?.copy_context(py)?) - } -} - -/// Run the event loop until the given Future completes -/// -/// After this function returns, the event loop can be resumed with [`run_until_complete`] -/// -/// # Arguments -/// * `event_loop` - The Python event loop that should run the future -/// * `fut` - The future to drive to completion -/// -/// # Examples -/// -/// ```no_run -/// # use std::{task::{Context, Poll}, pin::Pin, future::Future}; -/// # -/// # use pyo3_asyncio::{ -/// # TaskLocals, -/// # generic::{JoinError, SpawnLocalExt, ContextExt, LocalContextExt, Runtime} -/// # }; -/// # -/// # struct MyCustomJoinError; -/// # -/// # impl JoinError for MyCustomJoinError { -/// # fn is_panic(&self) -> bool { -/// # unreachable!() -/// # } -/// # } -/// # -/// # struct MyCustomJoinHandle; -/// # -/// # impl Future for MyCustomJoinHandle { -/// # type Output = Result<(), MyCustomJoinError>; -/// # -/// # fn poll(self: Pin<&mut Self>, _cx: &mut Context) -> Poll<Self::Output> { -/// # unreachable!() -/// # } -/// # } -/// # -/// # struct MyCustomRuntime; -/// # -/// # impl Runtime for MyCustomRuntime { -/// # type JoinError = MyCustomJoinError; -/// # type JoinHandle = MyCustomJoinHandle; -/// # -/// # fn spawn<F>(fut: F) -> Self::JoinHandle -/// # where -/// # F: Future<Output = ()> + Send + 'static -/// # { -/// # unreachable!() -/// # } -/// # } -/// # -/// # impl ContextExt for MyCustomRuntime { -/// # fn scope<F, R>(locals: TaskLocals, fut: F) -> Pin<Box<dyn Future<Output = R> + Send>> -/// # where -/// # F: Future<Output = R> + Send + 'static -/// # { -/// # unreachable!() -/// # } -/// # fn get_task_locals() -> Option<TaskLocals> { -/// # unreachable!() -/// # } -/// # } -/// # -/// # use std::time::Duration; -/// # -/// # use pyo3::prelude::; -/// # -/// # Python::with_gil(\|py\| -> PyResult<()> { -/// # let event_loop = py.import("asyncio")?.call_method0("new_event_loop")?; -/// # #[cfg(feature = "tokio-runtime")] -/// pyo3_asyncio::generic::run_until_complete::<MyCustomRuntime, _, _>(event_loop, async move { -/// tokio::time::sleep(Duration::from_secs(1)).await; -/// Ok(()) -/// })?; -/// # Ok(()) -/// # }).unwrap(); -/// ``` -pub fn run_until_complete<R, F, T>(event_loop: &PyAny, fut: F) -> PyResult<T> -where - R: Runtime + ContextExt, - F: Future<Output = PyResult<T>> + Send + 'static, - T: Send + Sync + 'static, -{ - let py = event_loop.py(); - let result_tx = Arc::new(Mutex::new(None)); - let result_rx = Arc::clone(&result_tx); - let coro = future_into_py_with_locals::<R, _, ()>( - py, - TaskLocals::new(event_loop).copy_context(py)?, - async move { - let val = fut.await?; - if let Ok(mut result) = result_tx.lock() { - result = Some(val); - } - Ok(()) - }, - )?; - - event_loop.call_method1("run_until_complete", (coro,))?; - - let result = result_rx.lock().unwrap().take().unwrap(); - Ok(result) -} - -/// Run the event loop until the given Future completes -/// -/// # Arguments -/// * `py` - The current PyO3 GIL guard -/// * `fut` - The future to drive to completion -/// -/// # Examples -/// -/// ```no_run -/// # use std::{task::{Context, Poll}, pin::Pin, future::Future}; -/// # -/// # use pyo3_asyncio::{ -/// # TaskLocals, -/// # generic::{JoinError, SpawnLocalExt, ContextExt, LocalContextExt, Runtime} -/// # }; -/// # -/// # struct MyCustomJoinError; -/// # -/// # impl JoinError for MyCustomJoinError { -/// # fn is_panic(&self) -> bool { -/// # unreachable!() -/// # } -/// # } -/// # -/// # struct MyCustomJoinHandle; -/// # -/// # impl Future for MyCustomJoinHandle { -/// # type Output = Result<(), MyCustomJoinError>; -/// # -/// # fn poll(self: Pin<&mut Self>, _cx: &mut Context) -> Poll<Self::Output> { -/// # unreachable!() -/// # } -/// # } -/// # -/// # struct MyCustomRuntime; -/// # -/// # impl Runtime for MyCustomRuntime { -/// # type JoinError = MyCustomJoinError; -/// # type JoinHandle = MyCustomJoinHandle; -/// # -/// # fn spawn<F>(fut: F) -> Self::JoinHandle -/// # where -/// # F: Future<Output = ()> + Send + 'static -/// # { -/// # unreachable!() -/// # } -/// # } -/// # -/// # impl ContextExt for MyCustomRuntime { -/// # fn scope<F, R>(locals: TaskLocals, fut: F) -> Pin<Box<dyn Future<Output = R> + Send>> -/// # where -/// # F: Future<Output = R> + Send + 'static -/// # { -/// # unreachable!() -/// # } -/// # fn get_task_locals() -> Option<TaskLocals> { -/// # unreachable!() -/// # } -/// # } -/// # -/// # use std::time::Duration; -/// # async fn custom_sleep(_duration: Duration) { } -/// # -/// # use pyo3::prelude::; -/// # -/// fn main() { -/// Python::with_gil(\|py\| { -/// pyo3_asyncio::generic::run::<MyCustomRuntime, _, _>(py, async move { -/// custom_sleep(Duration::from_secs(1)).await; -/// Ok(()) -/// }) -/// .map_err(\|e\| { -/// e.print_and_set_sys_last_vars(py); -/// }) -/// .unwrap(); -/// }) -/// } -/// ``` -pub fn run<R, F, T>(py: Python, fut: F) -> PyResult<T> -where - R: Runtime + ContextExt, - F: Future<Output = PyResult<T>> + Send + 'static, - T: Send + Sync + 'static, -{ - let event_loop = asyncio(py)?.call_method0("new_event_loop")?; - - let result = run_until_complete::<R, F, T>(event_loop, fut); - - close(event_loop)?; - - result -} - -pub fn cancelled(future: &PyAny) -> PyResult<bool> { - future.getattr("cancelled")?.call0()?.is_true() -} - -#[pyclass] -struct CheckedCompletor; - -#[pymethods] -impl CheckedCompletor { - fn __call__(&self, future: &PyAny, complete: &PyAny, value: &PyAny) -> PyResult<()> { - if cancelled(future)? { - return Ok(()); - } - - complete.call1((value,))?; - - Ok(()) - } -} - -pub fn set_result(event_loop: &PyAny, future: &PyAny, result: PyResult<PyObject>) -> PyResult<()> { - let py = event_loop.py(); - let none = py.None().into_ref(py); - - let (complete, val) = match result { - Ok(val) => (future.getattr("set_result")?, val.into_py(py)), - Err(err) => (future.getattr("set_exception")?, err.into_py(py)), - }; - call_soon_threadsafe(event_loop, none, (CheckedCompletor, future, complete, val))?; - - Ok(()) -} - -/// Convert a Python `awaitable` into a Rust Future -/// -/// This function simply forwards the future and the task locals returned by [`get_current_locals`] -/// to [`into_future_with_locals`](`crate::into_future_with_locals`). See -/// [`into_future_with_locals`](`crate::into_future_with_locals`) for more details. -/// -/// # Arguments -/// `awaitable` - The Python `awaitable` to be converted -/// -/// # Examples -/// -/// ```no_run -/// # use std::{pin::Pin, future::Future, task::{Context, Poll}, time::Duration}; -/// # -/// # use pyo3::prelude::; -/// # -/// # use pyo3_asyncio::{ -/// # TaskLocals, -/// # generic::{JoinError, SpawnLocalExt, ContextExt, LocalContextExt, Runtime} -/// # }; -/// # -/// # struct MyCustomJoinError; -/// # -/// # impl JoinError for MyCustomJoinError { -/// # fn is_panic(&self) -> bool { -/// # unreachable!() -/// # } -/// # } -/// # -/// # struct MyCustomJoinHandle; -/// # -/// # impl Future for MyCustomJoinHandle { -/// # type Output = Result<(), MyCustomJoinError>; -/// # -/// # fn poll(self: Pin<&mut Self>, _cx: &mut Context) -> Poll<Self::Output> { -/// # unreachable!() -/// # } -/// # } -/// # -/// # struct MyCustomRuntime; -/// # -/// # impl MyCustomRuntime { -/// # async fn sleep(_: Duration) { -/// # unreachable!() -/// # } -/// # } -/// # -/// # impl Runtime for MyCustomRuntime { -/// # type JoinError = MyCustomJoinError; -/// # type JoinHandle = MyCustomJoinHandle; -/// # -/// # fn spawn<F>(fut: F) -> Self::JoinHandle -/// # where -/// # F: Future<Output = ()> + Send + 'static -/// # { -/// # unreachable!() -/// # } -/// # } -/// # -/// # impl ContextExt for MyCustomRuntime { -/// # fn scope<F, R>(locals: TaskLocals, fut: F) -> Pin<Box<dyn Future<Output = R> + Send>> -/// # where -/// # F: Future<Output = R> + Send + 'static -/// # { -/// # unreachable!() -/// # } -/// # fn get_task_locals() -> Option<TaskLocals> { -/// # unreachable!() -/// # } -/// # } -/// # -/// const PYTHON_CODE: &'static str = r#" -/// import asyncio -/// -/// async def py_sleep(duration): -/// await asyncio.sleep(duration) -/// "#; -/// -/// async fn py_sleep(seconds: f32) -> PyResult<()> { -/// let test_mod = Python::with_gil(\|py\| -> PyResult<PyObject> { -/// Ok( -/// PyModule::from_code( -/// py, -/// PYTHON_CODE, -/// "test_into_future/test_mod.py", -/// "test_mod" -/// )? -/// .into() -/// ) -/// })?; -/// -/// Python::with_gil(\|py\| { -/// pyo3_asyncio::generic::into_future::<MyCustomRuntime>( -/// test_mod -/// .call_method1(py, "py_sleep", (seconds.into_py(py),))? -/// .as_ref(py), -/// ) -/// })? -/// .await?; -/// Ok(()) -/// } -/// ``` -pub fn into_future<R>( - awaitable: &PyAny, -) -> PyResult<impl Future<Output = PyResult<PyObject>> + Send> -where - R: Runtime + ContextExt, -{ - into_future_with_locals(&get_current_locals::<R>(awaitable.py())?, awaitable) -} - -/// Convert a Rust Future into a Python awaitable with a generic runtime -/// -/// If the `asyncio.Future` returned by this conversion is cancelled via `asyncio.Future.cancel`, -/// the Rust future will be cancelled as well (new behaviour in `v0.15`). -/// -/// Python `contextvars` are preserved when calling async Python functions within the Rust future -/// via [`into_future`] (new behaviour in `v0.15`). -/// -/// > Although `contextvars` are preserved for async Python functions, synchronous functions will -/// unfortunately fail to resolve them when called within the Rust future. This is because the -/// function is being called from a Rust thread, not inside an actual Python coroutine context. -/// > -/// > As a workaround, you can get the `contextvars` from the current task locals using -/// [`get_current_locals`] and [`TaskLocals::context`](`crate::TaskLocals::context`), then wrap your -/// synchronous function in a call to `contextvars.Context.run`. This will set the context, call the -/// synchronous function, and restore the previous context when it returns or raises an exception. -/// -/// # Arguments -/// `py` - PyO3 GIL guard -/// * `locals` - The task-local data for Python -/// * `fut` - The Rust future to be converted -/// -/// # Examples -/// -/// ```no_run -/// # use std::{task::{Context, Poll}, pin::Pin, future::Future}; -/// # -/// # use pyo3_asyncio::{ -/// # TaskLocals, -/// # generic::{JoinError, SpawnLocalExt, ContextExt, LocalContextExt, Runtime} -/// # }; -/// # -/// # struct MyCustomJoinError; -/// # -/// # impl JoinError for MyCustomJoinError { -/// # fn is_panic(&self) -> bool { -/// # unreachable!() -/// # } -/// # } -/// # -/// # struct MyCustomJoinHandle; -/// # -/// # impl Future for MyCustomJoinHandle { -/// # type Output = Result<(), MyCustomJoinError>; -/// # -/// # fn poll(self: Pin<&mut Self>, _cx: &mut Context) -> Poll<Self::Output> { -/// # unreachable!() -/// # } -/// # } -/// # -/// # struct MyCustomRuntime; -/// # -/// # impl MyCustomRuntime { -/// # async fn sleep(_: Duration) { -/// # unreachable!() -/// # } -/// # } -/// # -/// # impl Runtime for MyCustomRuntime { -/// # type JoinError = MyCustomJoinError; -/// # type JoinHandle = MyCustomJoinHandle; -/// # -/// # fn spawn<F>(fut: F) -> Self::JoinHandle -/// # where -/// # F: Future<Output = ()> + Send + 'static -/// # { -/// # unreachable!() -/// # } -/// # } -/// # -/// # impl ContextExt for MyCustomRuntime { -/// # fn scope<F, R>(locals: TaskLocals, fut: F) -> Pin<Box<dyn Future<Output = R> + Send>> -/// # where -/// # F: Future<Output = R> + Send + 'static -/// # { -/// # unreachable!() -/// # } -/// # fn get_task_locals() -> Option<TaskLocals> { -/// # unreachable!() -/// # } -/// # } -/// # -/// use std::time::Duration; -/// -/// use pyo3::prelude::; -/// -/// /// Awaitable sleep function -/// #[pyfunction] -/// fn sleep_for<'p>(py: Python<'p>, secs: &'p PyAny) -> PyResult<&'p PyAny> { -/// let secs = secs.extract()?; -/// pyo3_asyncio::generic::future_into_py_with_locals::<MyCustomRuntime, _, _>( -/// py, -/// pyo3_asyncio::generic::get_current_locals::<MyCustomRuntime>(py)?, -/// async move { -/// MyCustomRuntime::sleep(Duration::from_secs(secs)).await; -/// Ok(()) -/// } -/// ) -/// } -/// ``` -pub fn future_into_py_with_locals<R, F, T>( - py: Python, - locals: TaskLocals, - fut: F, -) -> PyResult<&PyAny> -where - R: Runtime + ContextExt, - F: Future<Output = PyResult<T>> + Send + 'static, - T: IntoPy<PyObject>, -{ - let (cancel_tx, cancel_rx) = oneshot::channel(); - - let py_fut = create_future(locals.event_loop.clone().into_ref(py))?; - py_fut.call_method1( - "add_done_callback", - (PyDoneCallback { - cancel_tx: Some(cancel_tx), - },), - )?; - - let future_tx1 = PyObject::from(py_fut); - let future_tx2 = future_tx1.clone(); - - R::spawn(async move { - let locals2 = locals.clone(); - - if let Err(e) = R::spawn(async move { - let result = R::scope( - locals2.clone(), - Cancellable::new_with_cancel_rx(fut, cancel_rx), - ) - .await; - - Python::with_gil(move \|py\| { - if cancelled(future_tx1.as_ref(py)) - .map_err(dump_err(py)) - .unwrap_or(false) - { - return; - } - - let _ = set_result( - locals2.event_loop(py), - future_tx1.as_ref(py), - result.map(\|val\| val.into_py(py)), - ) - .map_err(dump_err(py)); - }); - }) - .await - { - if e.is_panic() { - Python::with_gil(move \|py\| { - if cancelled(future_tx2.as_ref(py)) - .map_err(dump_err(py)) - .unwrap_or(false) - { - return; - } - - let _ = set_result( - locals.event_loop.as_ref(py), - future_tx2.as_ref(py), - Err(RustPanic::new_err("rust future panicked")), - ) - .map_err(dump_err(py)); - }); - } - } - }); - - Ok(py_fut) -} - -pin_project! { - /// Future returned by [`timeout`](timeout) and [`timeout_at`](timeout_at). - #[must_use = "futures do nothing unless you `.await` or poll them"] - #[derive(Debug)] - pub struct Cancellable<T> { - #[pin] - future: T, - #[pin] - cancel_rx: oneshot::Receiver<()>, - - poll_cancel_rx: bool - } -} - -impl<T> Cancellable<T> { - pub fn new_with_cancel_rx(future: T, cancel_rx: oneshot::Receiver<()>) -> Self { - Self { - future, - cancel_rx, - - poll_cancel_rx: true, - } - } -} - -impl<F, T> Future for Cancellable<F> -where - F: Future<Output = PyResult<T>>, - T: IntoPy<PyObject>, -{ - type Output = F::Output; - - fn poll(self: Pin<&mut Self>, cx: &mut Context<'_>) -> Poll<Self::Output> { - let this = self.project(); - - // First, try polling the future - if let Poll::Ready(v) = this.future.poll(cx) { - return Poll::Ready(v); - } - - // Now check for cancellation - if this.poll_cancel_rx { - match this.cancel_rx.poll(cx) { - Poll::Ready(Ok(())) => { - this.poll_cancel_rx = false; - // The python future has already been cancelled, so this return value will never - // be used. - Poll::Ready(Err(pyo3::exceptions::PyBaseException::new_err( - "unreachable", - ))) - } - Poll::Ready(Err(_)) => { - this.poll_cancel_rx = false; - Poll::Pending - } - Poll::Pending => Poll::Pending, - } - } else { - Poll::Pending - } - } -} - -#[pyclass] -pub struct PyDoneCallback { - pub cancel_tx: Option<oneshot::Sender<()>>, -} - -#[pymethods] -impl PyDoneCallback { - pub fn __call__(&mut self, fut: &PyAny) -> PyResult<()> { - let py = fut.py(); - - if cancelled(fut).map_err(dump_err(py)).unwrap_or(false) { - let _ = self.cancel_tx.take().unwrap().send(()); - } - - Ok(()) - } -} - -/// Convert a Rust Future into a Python awaitable with a generic runtime -/// -/// If the `asyncio.Future` returned by this conversion is cancelled via `asyncio.Future.cancel`, -/// the Rust future will be cancelled as well (new behaviour in `v0.15`). -/// -/// Python `contextvars` are preserved when calling async Python functions within the Rust future -/// via [`into_future`] (new behaviour in `v0.15`). -/// -/// > Although `contextvars` are preserved for async Python functions, synchronous functions will -/// unfortunately fail to resolve them when called within the Rust future. This is because the -/// function is being called from a Rust thread, not inside an actual Python coroutine context. -/// > -/// > As a workaround, you can get the `contextvars` from the current task locals using -/// [`get_current_locals`] and [`TaskLocals::context`](`crate::TaskLocals::context`), then wrap your -/// synchronous function in a call to `contextvars.Context.run`. This will set the context, call the -/// synchronous function, and restore the previous context when it returns or raises an exception. -/// -/// # Arguments -/// * `py` - The current PyO3 GIL guard -/// * `fut` - The Rust future to be converted -/// -/// # Examples -/// -/// ```no_run -/// # use std::{task::{Context, Poll}, pin::Pin, future::Future}; -/// # -/// # use pyo3_asyncio::{ -/// # TaskLocals, -/// # generic::{JoinError, SpawnLocalExt, ContextExt, LocalContextExt, Runtime} -/// # }; -/// # -/// # struct MyCustomJoinError; -/// # -/// # impl JoinError for MyCustomJoinError { -/// # fn is_panic(&self) -> bool { -/// # unreachable!() -/// # } -/// # } -/// # -/// # struct MyCustomJoinHandle; -/// # -/// # impl Future for MyCustomJoinHandle { -/// # type Output = Result<(), MyCustomJoinError>; -/// # -/// # fn poll(self: Pin<&mut Self>, _cx: &mut Context) -> Poll<Self::Output> { -/// # unreachable!() -/// # } -/// # } -/// # -/// # struct MyCustomRuntime; -/// # -/// # impl MyCustomRuntime { -/// # async fn sleep(_: Duration) { -/// # unreachable!() -/// # } -/// # } -/// # -/// # impl Runtime for MyCustomRuntime { -/// # type JoinError = MyCustomJoinError; -/// # type JoinHandle = MyCustomJoinHandle; -/// # -/// # fn spawn<F>(fut: F) -> Self::JoinHandle -/// # where -/// # F: Future<Output = ()> + Send + 'static -/// # { -/// # unreachable!() -/// # } -/// # } -/// # -/// # impl ContextExt for MyCustomRuntime { -/// # fn scope<F, R>(locals: TaskLocals, fut: F) -> Pin<Box<dyn Future<Output = R> + Send>> -/// # where -/// # F: Future<Output = R> + Send + 'static -/// # { -/// # unreachable!() -/// # } -/// # fn get_task_locals() -> Option<TaskLocals> { -/// # unreachable!() -/// # } -/// # } -/// # -/// use std::time::Duration; -/// -/// use pyo3::prelude::; -/// -/// /// Awaitable sleep function -/// #[pyfunction] -/// fn sleep_for<'p>(py: Python<'p>, secs: &'p PyAny) -> PyResult<&'p PyAny> { -/// let secs = secs.extract()?; -/// pyo3_asyncio::generic::future_into_py::<MyCustomRuntime, _, _>(py, async move { -/// MyCustomRuntime::sleep(Duration::from_secs(secs)).await; -/// Ok(()) -/// }) -/// } -/// ``` -pub fn future_into_py<R, F, T>(py: Python, fut: F) -> PyResult<&PyAny> -where - R: Runtime + ContextExt, - F: Future<Output = PyResult<T>> + Send + 'static, - T: IntoPy<PyObject>, -{ - future_into_py_with_locals::<R, F, T>(py, get_current_locals::<R>(py)?, fut) -} - -/// Convert a `!Send` Rust Future into a Python awaitable with a generic runtime and manual -/// specification of task locals. -/// -/// If the `asyncio.Future` returned by this conversion is cancelled via `asyncio.Future.cancel`, -/// the Rust future will be cancelled as well (new behaviour in `v0.15`). -/// -/// Python `contextvars` are preserved when calling async Python functions within the Rust future -/// via [`into_future`] (new behaviour in `v0.15`). -/// -/// > Although `contextvars` are preserved for async Python functions, synchronous functions will -/// unfortunately fail to resolve them when called within the Rust future. This is because the -/// function is being called from a Rust thread, not inside an actual Python coroutine context. -/// > -/// > As a workaround, you can get the `contextvars` from the current task locals using -/// [`get_current_locals`] and [`TaskLocals::context`](`crate::TaskLocals::context`), then wrap your -/// synchronous function in a call to `contextvars.Context.run`. This will set the context, call the -/// synchronous function, and restore the previous context when it returns or raises an exception. -/// -/// # Arguments -/// `py` - PyO3 GIL guard -/// * `locals` - The task locals for the future -/// * `fut` - The Rust future to be converted -/// -/// # Examples -/// -/// ```no_run -/// # use std::{task::{Context, Poll}, pin::Pin, future::Future}; -/// # -/// # use pyo3_asyncio::{ -/// # TaskLocals, -/// # generic::{JoinError, SpawnLocalExt, ContextExt, LocalContextExt, Runtime} -/// # }; -/// # -/// # struct MyCustomJoinError; -/// # -/// # impl JoinError for MyCustomJoinError { -/// # fn is_panic(&self) -> bool { -/// # unreachable!() -/// # } -/// # } -/// # -/// # struct MyCustomJoinHandle; -/// # -/// # impl Future for MyCustomJoinHandle { -/// # type Output = Result<(), MyCustomJoinError>; -/// # -/// # fn poll(self: Pin<&mut Self>, _cx: &mut Context) -> Poll<Self::Output> { -/// # unreachable!() -/// # } -/// # } -/// # -/// # struct MyCustomRuntime; -/// # -/// # impl MyCustomRuntime { -/// # async fn sleep(_: Duration) { -/// # unreachable!() -/// # } -/// # } -/// # -/// # impl Runtime for MyCustomRuntime { -/// # type JoinError = MyCustomJoinError; -/// # type JoinHandle = MyCustomJoinHandle; -/// # -/// # fn spawn<F>(fut: F) -> Self::JoinHandle -/// # where -/// # F: Future<Output = ()> + Send + 'static -/// # { -/// # unreachable!() -/// # } -/// # } -/// # -/// # impl ContextExt for MyCustomRuntime { -/// # fn scope<F, R>(locals: TaskLocals, fut: F) -> Pin<Box<dyn Future<Output = R> + Send>> -/// # where -/// # F: Future<Output = R> + Send + 'static -/// # { -/// # unreachable!() -/// # } -/// # fn get_task_locals() -> Option<TaskLocals> { -/// # unreachable!() -/// # } -/// # } -/// # -/// # impl SpawnLocalExt for MyCustomRuntime { -/// # fn spawn_local<F>(fut: F) -> Self::JoinHandle -/// # where -/// # F: Future<Output = ()> + 'static -/// # { -/// # unreachable!() -/// # } -/// # } -/// # -/// # impl LocalContextExt for MyCustomRuntime { -/// # fn scope_local<F, R>(locals: TaskLocals, fut: F) -> Pin<Box<dyn Future<Output = R>>> -/// # where -/// # F: Future<Output = R> + 'static -/// # { -/// # unreachable!() -/// # } -/// # } -/// # -/// use std::{rc::Rc, time::Duration}; -/// -/// use pyo3::prelude::; -/// -/// /// Awaitable sleep function -/// #[pyfunction] -/// fn sleep_for(py: Python, secs: u64) -> PyResult<&PyAny> { -/// // Rc is !Send so it cannot be passed into pyo3_asyncio::generic::future_into_py -/// let secs = Rc::new(secs); -/// -/// pyo3_asyncio::generic::local_future_into_py_with_locals::<MyCustomRuntime, _, _>( -/// py, -/// pyo3_asyncio::generic::get_current_locals::<MyCustomRuntime>(py)?, -/// async move { -/// MyCustomRuntime::sleep(Duration::from_secs(secs)).await; -/// Ok(()) -/// } -/// ) -/// } -/// ``` -pub fn local_future_into_py_with_locals<R, F, T>( - py: Python, - locals: TaskLocals, - fut: F, -) -> PyResult<&PyAny> -where - R: Runtime + SpawnLocalExt + LocalContextExt, - F: Future<Output = PyResult<T>> + 'static, - T: IntoPy<PyObject>, -{ - let (cancel_tx, cancel_rx) = oneshot::channel(); - - let py_fut = create_future(locals.event_loop.clone().into_ref(py))?; - py_fut.call_method1( - "add_done_callback", - (PyDoneCallback { - cancel_tx: Some(cancel_tx), - },), - )?; - - let future_tx1 = PyObject::from(py_fut); - let future_tx2 = future_tx1.clone(); - - R::spawn_local(async move { - let locals2 = locals.clone(); - - if let Err(e) = R::spawn_local(async move { - let result = R::scope_local( - locals2.clone(), - Cancellable::new_with_cancel_rx(fut, cancel_rx), - ) - .await; - - Python::with_gil(move \|py\| { - if cancelled(future_tx1.as_ref(py)) - .map_err(dump_err(py)) - .unwrap_or(false) - { - return; - } - - let _ = set_result( - locals2.event_loop.as_ref(py), - future_tx1.as_ref(py), - result.map(\|val\| val.into_py(py)), - ) - .map_err(dump_err(py)); - }); - }) - .await - { - if e.is_panic() { - Python::with_gil(move \|py\| { - if cancelled(future_tx2.as_ref(py)) - .map_err(dump_err(py)) - .unwrap_or(false) - { - return; - } - - let _ = set_result( - locals.event_loop.as_ref(py), - future_tx2.as_ref(py), - Err(RustPanic::new_err("Rust future panicked")), - ) - .map_err(dump_err(py)); - }); - } - } - }); - - Ok(py_fut) -} - -/// Convert a `!Send` Rust Future into a Python awaitable with a generic runtime -/// -/// If the `asyncio.Future` returned by this conversion is cancelled via `asyncio.Future.cancel`, -/// the Rust future will be cancelled as well (new behaviour in `v0.15`). -/// -/// Python `contextvars` are preserved when calling async Python functions within the Rust future -/// via [`into_future`] (new behaviour in `v0.15`). -/// -/// > Although `contextvars` are preserved for async Python functions, synchronous functions will -/// unfortunately fail to resolve them when called within the Rust future. This is because the -/// function is being called from a Rust thread, not inside an actual Python coroutine context. -/// > -/// > As a workaround, you can get the `contextvars` from the current task locals using -/// [`get_current_locals`] and [`TaskLocals::context`](`crate::TaskLocals::context`), then wrap your -/// synchronous function in a call to `contextvars.Context.run`. This will set the context, call the -/// synchronous function, and restore the previous context when it returns or raises an exception. -/// -/// # Arguments -/// * `py` - The current PyO3 GIL guard -/// * `fut` - The Rust future to be converted -/// -/// # Examples -/// -/// ```no_run -/// # use std::{task::{Context, Poll}, pin::Pin, future::Future}; -/// # -/// # use pyo3_asyncio::{ -/// # TaskLocals, -/// # generic::{JoinError, SpawnLocalExt, ContextExt, LocalContextExt, Runtime} -/// # }; -/// # -/// # struct MyCustomJoinError; -/// # -/// # impl JoinError for MyCustomJoinError { -/// # fn is_panic(&self) -> bool { -/// # unreachable!() -/// # } -/// # } -/// # -/// # struct MyCustomJoinHandle; -/// # -/// # impl Future for MyCustomJoinHandle { -/// # type Output = Result<(), MyCustomJoinError>; -/// # -/// # fn poll(self: Pin<&mut Self>, _cx: &mut Context) -> Poll<Self::Output> { -/// # unreachable!() -/// # } -/// # } -/// # -/// # struct MyCustomRuntime; -/// # -/// # impl MyCustomRuntime { -/// # async fn sleep(_: Duration) { -/// # unreachable!() -/// # } -/// # } -/// # -/// # impl Runtime for MyCustomRuntime { -/// # type JoinError = MyCustomJoinError; -/// # type JoinHandle = MyCustomJoinHandle; -/// # -/// # fn spawn<F>(fut: F) -> Self::JoinHandle -/// # where -/// # F: Future<Output = ()> + Send + 'static -/// # { -/// # unreachable!() -/// # } -/// # } -/// # -/// # impl ContextExt for MyCustomRuntime { -/// # fn scope<F, R>(locals: TaskLocals, fut: F) -> Pin<Box<dyn Future<Output = R> + Send>> -/// # where -/// # F: Future<Output = R> + Send + 'static -/// # { -/// # unreachable!() -/// # } -/// # fn get_task_locals() -> Option<TaskLocals> { -/// # unreachable!() -/// # } -/// # } -/// # -/// # impl SpawnLocalExt for MyCustomRuntime { -/// # fn spawn_local<F>(fut: F) -> Self::JoinHandle -/// # where -/// # F: Future<Output = ()> + 'static -/// # { -/// # unreachable!() -/// # } -/// # } -/// # -/// # impl LocalContextExt for MyCustomRuntime { -/// # fn scope_local<F, R>(locals: TaskLocals, fut: F) -> Pin<Box<dyn Future<Output = R>>> -/// # where -/// # F: Future<Output = R> + 'static -/// # { -/// # unreachable!() -/// # } -/// # } -/// # -/// use std::{rc::Rc, time::Duration}; -/// -/// use pyo3::prelude::; -/// -/// /// Awaitable sleep function -/// #[pyfunction] -/// fn sleep_for(py: Python, secs: u64) -> PyResult<&PyAny> { -/// // Rc is !Send so it cannot be passed into pyo3_asyncio::generic::future_into_py -/// let secs = Rc::new(secs); -/// -/// pyo3_asyncio::generic::local_future_into_py::<MyCustomRuntime, _, _>(py, async move { -/// MyCustomRuntime::sleep(Duration::from_secs(secs)).await; -/// Ok(()) -/// }) -/// } -/// ``` -pub fn local_future_into_py<R, F, T>(py: Python, fut: F) -> PyResult<&PyAny> -where - R: Runtime + ContextExt + SpawnLocalExt + LocalContextExt, - F: Future<Output = PyResult<T>> + 'static, - T: IntoPy<PyObject>, -{ - local_future_into_py_with_locals::<R, F, T>(py, get_current_locals::<R>(py)?, fut) -} - -/// <span class="module-item stab portability" style="display: inline; border-radius: 3px; padding: 2px; font-size: 80%; line-height: 1.2;"><code>unstable-streams</code></span> Convert an async generator into a stream -/// -/// This API is marked as unstable and is only available when the -/// `unstable-streams` crate feature is enabled. This comes with no -/// stability guarantees, and could be changed or removed at any time. -/// -/// # Arguments -/// * `locals` - The current task locals -/// * `gen` - The Python async generator to be converted -/// -/// # Examples -/// ```no_run -/// # use std::{task::{Context, Poll}, pin::Pin, future::Future}; -/// # -/// # use pyo3_asyncio::{ -/// # TaskLocals, -/// # generic::{JoinError, ContextExt, Runtime} -/// # }; -/// # -/// # struct MyCustomJoinError; -/// # -/// # impl JoinError for MyCustomJoinError { -/// # fn is_panic(&self) -> bool { -/// # unreachable!() -/// # } -/// # } -/// # -/// # struct MyCustomJoinHandle; -/// # -/// # impl Future for MyCustomJoinHandle { -/// # type Output = Result<(), MyCustomJoinError>; -/// # -/// # fn poll(self: Pin<&mut Self>, _cx: &mut Context) -> Poll<Self::Output> { -/// # unreachable!() -/// # } -/// # } -/// # -/// # struct MyCustomRuntime; -/// # -/// # impl Runtime for MyCustomRuntime { -/// # type JoinError = MyCustomJoinError; -/// # type JoinHandle = MyCustomJoinHandle; -/// # -/// # fn spawn<F>(fut: F) -> Self::JoinHandle -/// # where -/// # F: Future<Output = ()> + Send + 'static -/// # { -/// # unreachable!() -/// # } -/// # } -/// # -/// # impl ContextExt for MyCustomRuntime { -/// # fn scope<F, R>(locals: TaskLocals, fut: F) -> Pin<Box<dyn Future<Output = R> + Send>> -/// # where -/// # F: Future<Output = R> + Send + 'static -/// # { -/// # unreachable!() -/// # } -/// # fn get_task_locals() -> Option<TaskLocals> { -/// # unreachable!() -/// # } -/// # } -/// -/// use pyo3::prelude::; -/// use futures::{StreamExt, TryStreamExt}; -/// -/// const TEST_MOD: &str = r#" -/// import asyncio -/// -/// async def gen(): -/// for i in range(10): -/// await asyncio.sleep(0.1) -/// yield i -/// "#; -/// -/// # async fn test_async_gen() -> PyResult<()> { -/// let stream = Python::with_gil(\|py\| { -/// let test_mod = PyModule::from_code( -/// py, -/// TEST_MOD, -/// "test_rust_coroutine/test_mod.py", -/// "test_mod", -/// )?; -/// -/// pyo3_asyncio::generic::into_stream_with_locals_v1::<MyCustomRuntime>( -/// pyo3_asyncio::generic::get_current_locals::<MyCustomRuntime>(py)?, -/// test_mod.call_method0("gen")? -/// ) -/// })?; -/// -/// let vals = stream -/// .map(\|item\| Python::with_gil(\|py\| -> PyResult<i32> { Ok(item?.as_ref(py).extract()?) })) -/// .try_collect::<Vec<i32>>() -/// .await?; -/// -/// assert_eq!((0..10).collect::<Vec<i32>>(), vals); -/// -/// Ok(()) -/// # } -/// ``` -#[cfg(feature = "unstable-streams")] -pub fn into_stream_with_locals_v1<'p, R>( - locals: TaskLocals, - gen: &'p PyAny, -) -> PyResult<impl futures::Stream<Item = PyResult<PyObject>> + 'static> -where - R: Runtime, -{ - let (tx, rx) = async_channel::bounded(1); - let anext = PyObject::from(gen.getattr("__anext__")?); - - R::spawn(async move { - loop { - let fut = Python::with_gil(\|py\| -> PyResult<_> { - into_future_with_locals(&locals, anext.as_ref(py).call0()?) - }); - let item = match fut { - Ok(fut) => match fut.await { - Ok(item) => Ok(item), - Err(e) => { - let stop_iter = Python::with_gil(\|py\| { - e.is_instance_of::<pyo3::exceptions::PyStopAsyncIteration>(py) - }); - - if stop_iter { - // end the iteration - break; - } else { - Err(e) - } - } - }, - Err(e) => Err(e), - }; - - if tx.send(item).await.is_err() { - // receiving side was dropped - break; - } - } - }); - - Ok(rx) -} - -/// <span class="module-item stab portability" style="display: inline; border-radius: 3px; padding: 2px; font-size: 80%; line-height: 1.2;"><code>unstable-streams</code></span> Convert an async generator into a stream -/// -/// This API is marked as unstable* and is only available when the -/// `unstable-streams` crate feature is enabled. This comes with no -/// stability guarantees, and could be changed or removed at any time. -/// -/// # Arguments -/// * `gen` - The Python async generator to be converted -/// -/// # Examples -/// ```no_run -/// # use std::{task::{Context, Poll}, pin::Pin, future::Future}; -/// # -/// # use pyo3_asyncio::{ -/// # TaskLocals, -/// # generic::{JoinError, ContextExt, Runtime} -/// # }; -/// # -/// # struct MyCustomJoinError; -/// # -/// # impl JoinError for MyCustomJoinError { -/// # fn is_panic(&self) -> bool { -/// # unreachable!() -/// # } -/// # } -/// # -/// # struct MyCustomJoinHandle; -/// # -/// # impl Future for MyCustomJoinHandle { -/// # type Output = Result<(), MyCustomJoinError>; -/// # -/// # fn poll(self: Pin<&mut Self>, _cx: &mut Context) -> Poll<Self::Output> { -/// # unreachable!() -/// # } -/// # } -/// # -/// # struct MyCustomRuntime; -/// # -/// # impl Runtime for MyCustomRuntime { -/// # type JoinError = MyCustomJoinError; -/// # type JoinHandle = MyCustomJoinHandle; -/// # -/// # fn spawn<F>(fut: F) -> Self::JoinHandle -/// # where -/// # F: Future<Output = ()> + Send + 'static -/// # { -/// # unreachable!() -/// # } -/// # } -/// # -/// # impl ContextExt for MyCustomRuntime { -/// # fn scope<F, R>(locals: TaskLocals, fut: F) -> Pin<Box<dyn Future<Output = R> + Send>> -/// # where -/// # F: Future<Output = R> + Send + 'static -/// # { -/// # unreachable!() -/// # } -/// # fn get_task_locals() -> Option<TaskLocals> { -/// # unreachable!() -/// # } -/// # } -/// -/// use pyo3::prelude::; -/// use futures::{StreamExt, TryStreamExt}; -/// -/// const TEST_MOD: &str = r#" -/// import asyncio -/// -/// async def gen(): -/// for i in range(10): -/// await asyncio.sleep(0.1) -/// yield i -/// "#; -/// -/// # async fn test_async_gen() -> PyResult<()> { -/// let stream = Python::with_gil(\|py\| { -/// let test_mod = PyModule::from_code( -/// py, -/// TEST_MOD, -/// "test_rust_coroutine/test_mod.py", -/// "test_mod", -/// )?; -/// -/// pyo3_asyncio::generic::into_stream_v1::<MyCustomRuntime>(test_mod.call_method0("gen")?) -/// })?; -/// -/// let vals = stream -/// .map(\|item\| Python::with_gil(\|py\| -> PyResult<i32> { Ok(item?.as_ref(py).extract()?) })) -/// .try_collect::<Vec<i32>>() -/// .await?; -/// -/// assert_eq!((0..10).collect::<Vec<i32>>(), vals); -/// -/// Ok(()) -/// # } -/// ``` -#[cfg(feature = "unstable-streams")] -pub fn into_stream_v1<'p, R>( - gen: &'p PyAny, -) -> PyResult<impl futures::Stream<Item = PyResult<PyObject>> + 'static> -where - R: Runtime + ContextExt, -{ - into_stream_with_locals_v1::<R>(get_current_locals::<R>(gen.py())?, gen) -} - -fn py_true() -> PyObject { - static TRUE: OnceCell<PyObject> = OnceCell::new(); - TRUE.get_or_init(\|\| Python::with_gil(\|py\| true.into_py(py))) - .clone() -} -fn py_false() -> PyObject { - static FALSE: OnceCell<PyObject> = OnceCell::new(); - FALSE - .get_or_init(\|\| Python::with_gil(\|py\| false.into_py(py))) - .clone() -} - -trait Sender: Send + 'static { - fn send(&mut self, locals: TaskLocals, item: PyObject) -> PyResult<PyObject>; - fn close(&mut self) -> PyResult<()>; -} - -struct GenericSender<R> -where - R: Runtime, -{ - runtime: PhantomData<R>, - tx: mpsc::Sender<PyObject>, -} - -impl<R> Sender for GenericSender<R> -where - R: Runtime + ContextExt, -{ - fn send(&mut self, locals: TaskLocals, item: PyObject) -> PyResult<PyObject> { - match self.tx.try_send(item.clone()) { - Ok(_) => Ok(py_true()), - Err(e) => { - if e.is_full() { - let mut tx = self.tx.clone(); - Python::with_gil(move \|py\| { - Ok( - future_into_py_with_locals::<R, _, PyObject>(py, locals, async move { - if tx.flush().await.is_err() { - // receiving side disconnected - return Ok(py_false()); - } - if tx.send(item).await.is_err() { - // receiving side disconnected - return Ok(py_false()); - } - Ok(py_true()) - })? - .into(), - ) - }) - } else { - Ok(py_false()) - } - } - } - } - fn close(&mut self) -> PyResult<()> { - self.tx.close_channel(); - Ok(()) - } -} - -#[pyclass] -struct SenderGlue { - locals: TaskLocals, - tx: Box<dyn Sender>, -} -#[pymethods] -impl SenderGlue { - pub fn send(&mut self, item: PyObject) -> PyResult<PyObject> { - self.tx.send(self.locals.clone(), item) - } - pub fn close(&mut self) -> PyResult<()> { - self.tx.close() - } -} - -#[cfg(feature = "unstable-streams")] -const STREAM_GLUE: &str = r#" -import asyncio - -async def forward(gen, sender): - async for item in gen: - should_continue = sender.send(item) - - if asyncio.iscoroutine(should_continue): - should_continue = await should_continue - - if should_continue: - continue - else: - break - - sender.close() -"#; - -/// <span class="module-item stab portability" style="display: inline; border-radius: 3px; padding: 2px; font-size: 80%; line-height: 1.2;"><code>unstable-streams</code></span> Convert an async generator into a stream -/// -/// This API is marked as unstable* and is only available when the -/// `unstable-streams` crate feature is enabled. This comes with no -/// stability guarantees, and could be changed or removed at any time. -/// -/// # Arguments -/// * `locals` - The current task locals -/// * `gen` - The Python async generator to be converted -/// -/// # Examples -/// ```no_run -/// # use std::{task::{Context, Poll}, pin::Pin, future::Future}; -/// # -/// # use pyo3_asyncio::{ -/// # TaskLocals, -/// # generic::{JoinError, ContextExt, Runtime} -/// # }; -/// # -/// # struct MyCustomJoinError; -/// # -/// # impl JoinError for MyCustomJoinError { -/// # fn is_panic(&self) -> bool { -/// # unreachable!() -/// # } -/// # } -/// # -/// # struct MyCustomJoinHandle; -/// # -/// # impl Future for MyCustomJoinHandle { -/// # type Output = Result<(), MyCustomJoinError>; -/// # -/// # fn poll(self: Pin<&mut Self>, _cx: &mut Context) -> Poll<Self::Output> { -/// # unreachable!() -/// # } -/// # } -/// # -/// # struct MyCustomRuntime; -/// # -/// # impl Runtime for MyCustomRuntime { -/// # type JoinError = MyCustomJoinError; -/// # type JoinHandle = MyCustomJoinHandle; -/// # -/// # fn spawn<F>(fut: F) -> Self::JoinHandle -/// # where -/// # F: Future<Output = ()> + Send + 'static -/// # { -/// # unreachable!() -/// # } -/// # } -/// # -/// # impl ContextExt for MyCustomRuntime { -/// # fn scope<F, R>(locals: TaskLocals, fut: F) -> Pin<Box<dyn Future<Output = R> + Send>> -/// # where -/// # F: Future<Output = R> + Send + 'static -/// # { -/// # unreachable!() -/// # } -/// # fn get_task_locals() -> Option<TaskLocals> { -/// # unreachable!() -/// # } -/// # } -/// -/// use pyo3::prelude::; -/// use futures::{StreamExt, TryStreamExt}; -/// -/// const TEST_MOD: &str = r#" -/// import asyncio -/// -/// async def gen(): -/// for i in range(10): -/// await asyncio.sleep(0.1) -/// yield i -/// "#; -/// -/// # async fn test_async_gen() -> PyResult<()> { -/// let stream = Python::with_gil(\|py\| { -/// let test_mod = PyModule::from_code( -/// py, -/// TEST_MOD, -/// "test_rust_coroutine/test_mod.py", -/// "test_mod", -/// )?; -/// -/// pyo3_asyncio::generic::into_stream_with_locals_v2::<MyCustomRuntime>( -/// pyo3_asyncio::generic::get_current_locals::<MyCustomRuntime>(py)?, -/// test_mod.call_method0("gen")? -/// ) -/// })?; -/// -/// let vals = stream -/// .map(\|item\| Python::with_gil(\|py\| -> PyResult<i32> { Ok(item.as_ref(py).extract()?) })) -/// .try_collect::<Vec<i32>>() -/// .await?; -/// -/// assert_eq!((0..10).collect::<Vec<i32>>(), vals); -/// -/// Ok(()) -/// # } -/// ``` -#[cfg(feature = "unstable-streams")] -pub fn into_stream_with_locals_v2<'p, R>( - locals: TaskLocals, - gen: &'p PyAny, -) -> PyResult<impl futures::Stream<Item = PyObject> + 'static> -where - R: Runtime + ContextExt, -{ - static GLUE_MOD: OnceCell<PyObject> = OnceCell::new(); - let py = gen.py(); - let glue = GLUE_MOD - .get_or_try_init(\|\| -> PyResult<PyObject> { - Ok(PyModule::from_code( - py, - STREAM_GLUE, - "pyo3_asyncio/pyo3_asyncio_glue.py", - "pyo3_asyncio_glue", - )? - .into()) - })? - .as_ref(py); - - let (tx, rx) = mpsc::channel(10); - - locals.event_loop(py).call_method1( - "call_soon_threadsafe", - ( - locals.event_loop(py).getattr("create_task")?, - glue.call_method1( - "forward", - ( - gen, - SenderGlue { - locals, - tx: Box::new(GenericSender { - runtime: PhantomData::<R>, - tx, - }), - }, - ), - )?, - ), - )?; - Ok(rx) -} - -/// <span class="module-item stab portability" style="display: inline; border-radius: 3px; padding: 2px; font-size: 80%; line-height: 1.2;"><code>unstable-streams</code></span> Convert an async generator into a stream -/// -/// This API is marked as unstable* and is only available when the -/// `unstable-streams` crate feature is enabled. This comes with no -/// stability guarantees, and could be changed or removed at any time. -/// -/// # Arguments -/// * `gen` - The Python async generator to be converted -/// -/// # Examples -/// ```no_run -/// # use std::{task::{Context, Poll}, pin::Pin, future::Future}; -/// # -/// # use pyo3_asyncio::{ -/// # TaskLocals, -/// # generic::{JoinError, ContextExt, Runtime} -/// # }; -/// # -/// # struct MyCustomJoinError; -/// # -/// # impl JoinError for MyCustomJoinError { -/// # fn is_panic(&self) -> bool { -/// # unreachable!() -/// # } -/// # } -/// # -/// # struct MyCustomJoinHandle; -/// # -/// # impl Future for MyCustomJoinHandle { -/// # type Output = Result<(), MyCustomJoinError>; -/// # -/// # fn poll(self: Pin<&mut Self>, _cx: &mut Context) -> Poll<Self::Output> { -/// # unreachable!() -/// # } -/// # } -/// # -/// # struct MyCustomRuntime; -/// # -/// # impl Runtime for MyCustomRuntime { -/// # type JoinError = MyCustomJoinError; -/// # type JoinHandle = MyCustomJoinHandle; -/// # -/// # fn spawn<F>(fut: F) -> Self::JoinHandle -/// # where -/// # F: Future<Output = ()> + Send + 'static -/// # { -/// # unreachable!() -/// # } -/// # } -/// # -/// # impl ContextExt for MyCustomRuntime { -/// # fn scope<F, R>(locals: TaskLocals, fut: F) -> Pin<Box<dyn Future<Output = R> + Send>> -/// # where -/// # F: Future<Output = R> + Send + 'static -/// # { -/// # unreachable!() -/// # } -/// # fn get_task_locals() -> Option<TaskLocals> { -/// # unreachable!() -/// # } -/// # } -/// -/// use pyo3::prelude::; -/// use futures::{StreamExt, TryStreamExt}; -/// -/// const TEST_MOD: &str = r#" -/// import asyncio -/// -/// async def gen(): -/// for i in range(10): -/// await asyncio.sleep(0.1) -/// yield i -/// "#; -/// -/// # async fn test_async_gen() -> PyResult<()> { -/// let stream = Python::with_gil(\|py\| { -/// let test_mod = PyModule::from_code( -/// py, -/// TEST_MOD, -/// "test_rust_coroutine/test_mod.py", -/// "test_mod", -/// )?; -/// -/// pyo3_asyncio::generic::into_stream_v2::<MyCustomRuntime>(test_mod.call_method0("gen")?) -/// })?; -/// -/// let vals = stream -/// .map(\|item\| Python::with_gil(\|py\| -> PyResult<i32> { Ok(item.as_ref(py).extract()?) })) -/// .try_collect::<Vec<i32>>() -/// .await?; -/// -/// assert_eq!((0..10).collect::<Vec<i32>>(), vals); -/// -/// Ok(()) -/// # } -/// ``` -#[cfg(feature = "unstable-streams")] -pub fn into_stream_v2<'p, R>( - gen: &'p PyAny, -) -> PyResult<impl futures::Stream<Item = PyObject> + 'static> -where - R: Runtime + ContextExt, -{ - into_stream_with_locals_v2::<R>(get_current_locals::<R>(gen.py())?, gen) -} diff --git a/lib/pyo3-asyncio/src/lib.rs b/lib/pyo3-asyncio/src/lib.rs deleted file mode 100644 index c3c820fc..00000000 --- a/lib/pyo3-asyncio/src/lib.rs +++ /dev/null @@ -1,663 +0,0 @@ -#![warn(missing_docs)] -#![allow(clippy::borrow_deref_ref)] - -//! Rust Bindings to the Python Asyncio Event Loop -//! -//! # Motivation -//! -//! This crate aims to provide a convenient interface to manage the interop between Python and -//! Rust's async/await models. It supports conversions between Rust and Python futures and manages -//! the event loops for both languages. Python's threading model and GIL can make this interop a bit -//! trickier than one might expect, so there are a few caveats that users should be aware of. -//! -//! ## Why Two Event Loops -//! -//! Currently, we don't have a way to run Rust futures directly on Python's event loop. Likewise, -//! Python's coroutines cannot be directly spawned on a Rust event loop. The two coroutine models -//! require some additional assistance from their event loops, so in all likelihood they will need -//! a new _unique_ event loop that addresses the needs of both languages if the coroutines are to -//! be run on the same loop. -//! -//! It's not immediately clear that this would provide worthwhile performance wins either, so in the -//! interest of getting something simple out there to facilitate these conversions, this crate -//! handles the communication between _separate_ Python and Rust event loops. -//! -//! ## Python's Event Loop and the Main Thread -//! -//! Python is very picky about the threads used by the `asyncio` executor. In particular, it needs -//! to have control over the main thread in order to handle signals like CTRL-C correctly. This -//! means that Cargo's default test harness will no longer work since it doesn't provide a method of -//! overriding the main function to add our event loop initialization and finalization. -//! -//! ## Event Loop References and ContextVars -//! -//! One problem that arises when interacting with Python's asyncio library is that the functions we -//! use to get a reference to the Python event loop can only be called in certain contexts. Since -//! PyO3 Asyncio needs to interact with Python's event loop during conversions, the context of these -//! conversions can matter a lot. -//! -//! Likewise, Python's `contextvars` library can require some special treatment. Python functions -//! and coroutines can rely on the context of outer coroutines to function correctly, so this -//! library needs to be able to preserve `contextvars` during conversions. -//! -//! > The core conversions we've mentioned so far in the README should insulate you from these -//! concerns in most cases. For the edge cases where they don't, this section should provide you -//! with the information you need to solve these problems. -//! -//! ### The Main Dilemma -//! -//! Python programs can have many independent event loop instances throughout the lifetime of the -//! application (`asyncio.run` for example creates its own event loop each time it's called for -//! instance), and they can even run concurrent with other event loops. For this reason, the most -//! correct method of obtaining a reference to the Python event loop is via -//! `asyncio.get_running_loop`. -//! -//! `asyncio.get_running_loop` returns the event loop associated with the current OS thread. It can -//! be used inside Python coroutines to spawn concurrent tasks, interact with timers, or in our case -//! signal between Rust and Python. This is all well and good when we are operating on a Python -//! thread, but since Rust threads are not associated with a Python event loop, -//! `asyncio.get_running_loop` will fail when called on a Rust runtime. -//! -//! `contextvars` operates in a similar way, though the current context is not always associated -//! with the current OS thread. Different contexts can be associated with different coroutines even -//! if they run on the same OS thread. -//! -//! ### The Solution -//! -//! A really straightforward way of dealing with this problem is to pass references to the -//! associated Python event loop and context for every conversion. That's why we have a structure -//! called `TaskLocals` and a set of conversions that accept it. -//! -//! `TaskLocals` stores the current event loop, and allows the user to copy the current Python -//! context if necessary. The following conversions will use these references to perform the -//! necessary conversions and restore Python context when needed: -//! -//! - `pyo3_asyncio::into_future_with_locals` - Convert a Python awaitable into a Rust future. -//! - `pyo3_asyncio::<runtime>::future_into_py_with_locals` - Convert a Rust future into a Python -//! awaitable. -//! - `pyo3_asyncio::<runtime>::local_future_into_py_with_locals` - Convert a `!Send` Rust future -//! into a Python awaitable. -//! -//! One clear disadvantage to this approach is that the Rust application has to explicitly track -//! these references. In native libraries, we can't make any assumptions about the underlying event -//! loop, so the only reliable way to make sure our conversions work properly is to store these -//! references at the callsite to use later on. -//! -//! ```rust -//! use pyo3::{wrap_pyfunction, prelude::}; -//! -//! # #[cfg(feature = "tokio-runtime")] -//! #[pyfunction] -//! fn sleep(py: Python) -> PyResult<&PyAny> { -//! // Construct the task locals structure with the current running loop and context -//! let locals = pyo3_asyncio::TaskLocals::with_running_loop(py)?.copy_context(py)?; -//! -//! // Convert the async move { } block to a Python awaitable -//! pyo3_asyncio::tokio::future_into_py_with_locals(py, locals.clone(), async move { -//! let py_sleep = Python::with_gil(\|py\| { -//! // Sometimes we need to call other async Python functions within -//! // this future. In order for this to work, we need to track the -//! // event loop from earlier. -//! pyo3_asyncio::into_future_with_locals( -//! &locals, -//! py.import("asyncio")?.call_method1("sleep", (1,))? -//! ) -//! })?; -//! -//! py_sleep.await?; -//! -//! Ok(()) -//! }) -//! } -//! -//! # #[cfg(feature = "tokio-runtime")] -//! #[pymodule] -//! fn my_mod(py: Python, m: &PyModule) -> PyResult<()> { -//! m.add_function(wrap_pyfunction!(sleep, m)?)?; -//! Ok(()) -//! } -//! ``` -//! -//! > A naive solution to this tracking problem would be to cache a global reference to the asyncio -//! event loop that all PyO3 Asyncio conversions can use. In fact this is what we did in PyO3 -//! Asyncio `v0.13`. This works well for applications, but it soon became clear that this is not -//! so ideal for libraries. Libraries usually have no direct control over how the event loop is -//! managed, they're just expected to work with any event loop at any point in the application. -//! This problem is compounded further when multiple event loops are used in the application since -//! the global reference will only point to one. -//! -//! Another disadvantage to this explicit approach that is less obvious is that we can no longer -//! call our `#[pyfunction] fn sleep` on a Rust runtime since `asyncio.get_running_loop` only works -//! on Python threads! It's clear that we need a slightly more flexible approach. -//! -//! In order to detect the Python event loop at the callsite, we need something like -//! `asyncio.get_running_loop` and `contextvars.copy_context` that works for _both Python and Rust_. -//! In Python, `asyncio.get_running_loop` uses thread-local data to retrieve the event loop -//! associated with the current thread. What we need in Rust is something that can retrieve the -//! Python event loop and contextvars associated with the current Rust _task_. -//! -//! Enter `pyo3_asyncio::<runtime>::get_current_locals`. This function first checks task-local data -//! for the `TaskLocals`, then falls back on `asyncio.get_running_loop` and -//! `contextvars.copy_context` if no task locals are found. This way both bases are -//! covered. -//! -//! Now, all we need is a way to store the `TaskLocals` for the Rust future. Since this is a -//! runtime-specific feature, you can find the following functions in each runtime module: -//! -//! - `pyo3_asyncio::<runtime>::scope` - Store the task-local data when executing the given Future. -//! - `pyo3_asyncio::<runtime>::scope_local` - Store the task-local data when executing the given -//! `!Send` Future. -//! -//! With these new functions, we can make our previous example more correct: -//! -//! ```rust no_run -//! use pyo3::prelude::; -//! -//! # #[cfg(feature = "tokio-runtime")] -//! #[pyfunction] -//! fn sleep(py: Python) -> PyResult<&PyAny> { -//! // get the current event loop through task-local data -//! // OR `asyncio.get_running_loop` and `contextvars.copy_context` -//! let locals = pyo3_asyncio::tokio::get_current_locals(py)?; -//! -//! pyo3_asyncio::tokio::future_into_py_with_locals( -//! py, -//! locals.clone(), -//! // Store the current locals in task-local data -//! pyo3_asyncio::tokio::scope(locals.clone(), async move { -//! let py_sleep = Python::with_gil(\|py\| { -//! pyo3_asyncio::into_future_with_locals( -//! // Now we can get the current locals through task-local data -//! &pyo3_asyncio::tokio::get_current_locals(py)?, -//! py.import("asyncio")?.call_method1("sleep", (1,))? -//! ) -//! })?; -//! -//! py_sleep.await?; -//! -//! Ok(Python::with_gil(\|py\| py.None())) -//! }) -//! ) -//! } -//! -//! # #[cfg(feature = "tokio-runtime")] -//! #[pyfunction] -//! fn wrap_sleep(py: Python) -> PyResult<&PyAny> { -//! // get the current event loop through task-local data -//! // OR `asyncio.get_running_loop` and `contextvars.copy_context` -//! let locals = pyo3_asyncio::tokio::get_current_locals(py)?; -//! -//! pyo3_asyncio::tokio::future_into_py_with_locals( -//! py, -//! locals.clone(), -//! // Store the current locals in task-local data -//! pyo3_asyncio::tokio::scope(locals.clone(), async move { -//! let py_sleep = Python::with_gil(\|py\| { -//! pyo3_asyncio::into_future_with_locals( -//! &pyo3_asyncio::tokio::get_current_locals(py)?, -//! // We can also call sleep within a Rust task since the -//! // locals are stored in task local data -//! sleep(py)? -//! ) -//! })?; -//! -//! py_sleep.await?; -//! -//! Ok(Python::with_gil(\|py\| py.None())) -//! }) -//! ) -//! } -//! -//! # #[cfg(feature = "tokio-runtime")] -//! #[pymodule] -//! fn my_mod(py: Python, m: &PyModule) -> PyResult<()> { -//! m.add_function(wrap_pyfunction!(sleep, m)?)?; -//! m.add_function(wrap_pyfunction!(wrap_sleep, m)?)?; -//! Ok(()) -//! } -//! ``` -//! -//! Even though this is more correct, it's clearly not more ergonomic. That's why we introduced a -//! set of functions with this functionality baked in: -//! -//! - `pyo3_asyncio::<runtime>::into_future` -//! > Convert a Python awaitable into a Rust future (using -//! `pyo3_asyncio::<runtime>::get_current_locals`) -//! - `pyo3_asyncio::<runtime>::future_into_py` -//! > Convert a Rust future into a Python awaitable (using -//! `pyo3_asyncio::<runtime>::get_current_locals` and `pyo3_asyncio::<runtime>::scope` to set the -//! task-local event loop for the given Rust future) -//! - `pyo3_asyncio::<runtime>::local_future_into_py` -//! > Convert a `!Send` Rust future into a Python awaitable (using -//! `pyo3_asyncio::<runtime>::get_current_locals` and `pyo3_asyncio::<runtime>::scope_local` to -//! set the task-local event loop for the given Rust future). -//! -//! __These are the functions that we recommend using__. With these functions, the previous example -//! can be rewritten to be more compact: -//! -//! ```rust -//! use pyo3::prelude::; -//! -//! # #[cfg(feature = "tokio-runtime")] -//! #[pyfunction] -//! fn sleep(py: Python) -> PyResult<&PyAny> { -//! pyo3_asyncio::tokio::future_into_py(py, async move { -//! let py_sleep = Python::with_gil(\|py\| { -//! pyo3_asyncio::tokio::into_future( -//! py.import("asyncio")?.call_method1("sleep", (1,))? -//! ) -//! })?; -//! -//! py_sleep.await?; -//! -//! Ok(Python::with_gil(\|py\| py.None())) -//! }) -//! } -//! -//! # #[cfg(feature = "tokio-runtime")] -//! #[pyfunction] -//! fn wrap_sleep(py: Python) -> PyResult<&PyAny> { -//! pyo3_asyncio::tokio::future_into_py(py, async move { -//! let py_sleep = Python::with_gil(\|py\| { -//! pyo3_asyncio::tokio::into_future(sleep(py)?) -//! })?; -//! -//! py_sleep.await?; -//! -//! Ok(Python::with_gil(\|py\| py.None())) -//! }) -//! } -//! -//! # #[cfg(feature = "tokio-runtime")] -//! #[pymodule] -//! fn my_mod(py: Python, m: &PyModule) -> PyResult<()> { -//! m.add_function(wrap_pyfunction!(sleep, m)?)?; -//! m.add_function(wrap_pyfunction!(wrap_sleep, m)?)?; -//! Ok(()) -//! } -//! ``` -//! -//! > A special thanks to [@ShadowJonathan](https://github.com/ShadowJonathan) for helping with the -//! design and review of these changes! -//! -//! ## Rust's Event Loop -//! -//! Currently only the Async-Std and Tokio runtimes are supported by this crate. If you need support -//! for another runtime, feel free to make a request on GitHub (or attempt to add support yourself -//! with the [`generic`] module)! -//! -//! > _In the future, we may implement first class support for more Rust runtimes. Contributions are -//! welcome as well!_ -//! -//! ## Features -//! -//! Items marked with -//! <span -//! class="module-item stab portability" -//! style="display: inline; border-radius: 3px; padding: 2px; font-size: 80%; line-height: 1.2;" -//! ><code>attributes</code></span> -//! are only available when the `attributes` Cargo feature is enabled: -//! -//! ```toml -//! [dependencies.pyo3-asyncio] -//! version = "0.17" -//! features = ["attributes"] -//! ``` -//! -//! Items marked with -//! <span -//! class="module-item stab portability" -//! style="display: inline; border-radius: 3px; padding: 2px; font-size: 80%; line-height: 1.2;" -//! ><code>async-std-runtime</code></span> -//! are only available when the `async-std-runtime` Cargo feature is enabled: -//! -//! ```toml -//! [dependencies.pyo3-asyncio] -//! version = "0.17" -//! features = ["async-std-runtime"] -//! ``` -//! -//! Items marked with -//! <span -//! class="module-item stab portability" -//! style="display: inline; border-radius: 3px; padding: 2px; font-size: 80%; line-height: 1.2;" -//! ><code>tokio-runtime</code></span> -//! are only available when the `tokio-runtime` Cargo feature is enabled: -//! -//! ```toml -//! [dependencies.pyo3-asyncio] -//! version = "0.17" -//! features = ["tokio-runtime"] -//! ``` -//! -//! Items marked with -//! <span -//! class="module-item stab portability" -//! style="display: inline; border-radius: 3px; padding: 2px; font-size: 80%; line-height: 1.2;" -//! ><code>testing</code></span> -//! are only available when the `testing` Cargo feature is enabled: -//! -//! ```toml -//! [dependencies.pyo3-asyncio] -//! version = "0.17" -//! features = ["testing"] -//! ``` - -/// Re-exported for #[test] attributes -#[cfg(all(feature = "attributes", feature = "testing"))] -pub use inventory; - -/// <span class="module-item stab portability" style="display: inline; border-radius: 3px; padding: 2px; font-size: 80%; line-height: 1.2;"><code>testing</code></span> Utilities for writing PyO3 Asyncio tests -#[cfg(feature = "testing")] -pub mod testing; - -#[cfg(feature = "async-std")] -pub mod async_std; - -#[cfg(feature = "tokio-runtime")] -pub mod tokio; - -/// Errors and exceptions related to PyO3 Asyncio -pub mod err; - -pub mod generic; - -/// Test README -#[doc(hidden)] -pub mod doc_test { - #[allow(unused)] - macro_rules! doc_comment { - ($x:expr, $module:item) => { - #[doc = $x] - $module - }; - } - - #[allow(unused)] - macro_rules! doctest { - ($x:expr, $y:ident) => { - doc_comment!(include_str!($x), mod $y {}); - }; - } - - #[cfg(all( - feature = "async-std-runtime", - feature = "tokio-runtime", - feature = "attributes" - ))] - doctest!("../README.md", readme_md); -} - -use std::future::Future; - -use futures::channel::oneshot; -use once_cell::sync::OnceCell; -use pyo3::{ - prelude::, - types::{PyDict, PyTuple}, -}; - -static ASYNCIO: OnceCell<PyObject> = OnceCell::new(); -static CONTEXTVARS: OnceCell<PyObject> = OnceCell::new(); -static ENSURE_FUTURE: OnceCell<PyObject> = OnceCell::new(); -static GET_RUNNING_LOOP: OnceCell<PyObject> = OnceCell::new(); - -fn ensure_future<'p>(py: Python<'p>, awaitable: &'p PyAny) -> PyResult<&'p PyAny> { - ENSURE_FUTURE - .get_or_try_init(\|\| -> PyResult<PyObject> { - Ok(asyncio(py)?.getattr("ensure_future")?.into()) - })? - .as_ref(py) - .call1((awaitable,)) -} - -pub fn create_future(event_loop: &PyAny) -> PyResult<&PyAny> { - event_loop.call_method0("create_future") -} - -fn close(event_loop: &PyAny) -> PyResult<()> { - event_loop.call_method1( - "run_until_complete", - (event_loop.call_method0("shutdown_asyncgens")?,), - )?; - - // how to do this prior to 3.9? - if event_loop.hasattr("shutdown_default_executor")? { - event_loop.call_method1( - "run_until_complete", - (event_loop.call_method0("shutdown_default_executor")?,), - )?; - } - - event_loop.call_method0("close")?; - - Ok(()) -} - -fn asyncio(py: Python) -> PyResult<&PyAny> { - ASYNCIO - .get_or_try_init(\|\| Ok(py.import("asyncio")?.into())) - .map(\|asyncio\| asyncio.as_ref(py)) -} - -/// Get a reference to the Python Event Loop from Rust -/// -/// Equivalent to `asyncio.get_running_loop()` in Python 3.7+. -pub fn get_running_loop(py: Python) -> PyResult<&PyAny> { - // Ideally should call get_running_loop, but calls get_event_loop for compatibility when - // get_running_loop is not available. - GET_RUNNING_LOOP - .get_or_try_init(\|\| -> PyResult<PyObject> { - let asyncio = asyncio(py)?; - - Ok(asyncio.getattr("get_running_loop")?.into()) - })? - .as_ref(py) - .call0() -} - -fn contextvars(py: Python) -> PyResult<&PyAny> { - Ok(CONTEXTVARS - .get_or_try_init(\|\| py.import("contextvars").map(\|m\| m.into()))? - .as_ref(py)) -} - -fn copy_context(py: Python) -> PyResult<&PyAny> { - contextvars(py)?.call_method0("copy_context") -} - -/// Task-local data to store for Python conversions. -#[derive(Debug, Clone)] -pub struct TaskLocals { - /// Track the event loop of the Python task - event_loop: PyObject, - /// Track the contextvars of the Python task - context: PyObject, -} - -impl TaskLocals { - /// At a minimum, TaskLocals must store the event loop. - pub fn new(event_loop: &PyAny) -> Self { - Self { - event_loop: event_loop.into(), - context: event_loop.py().None(), - } - } - - /// Construct TaskLocals with the event loop returned by `get_running_loop` - pub fn with_running_loop(py: Python) -> PyResult<Self> { - Ok(Self::new(get_running_loop(py)?)) - } - - /// Manually provide the contextvars for the current task. - pub fn with_context(self, context: &PyAny) -> Self { - Self { - context: context.into(), - ..self - } - } - - /// Capture the current task's contextvars - pub fn copy_context(self, py: Python) -> PyResult<Self> { - Ok(self.with_context(copy_context(py)?)) - } - - /// Get a reference to the event loop - pub fn event_loop<'p>(&self, py: Python<'p>) -> &'p PyAny { - self.event_loop.clone().into_ref(py) - } - - /// Get a reference to the python context - pub fn context<'p>(&self, py: Python<'p>) -> &'p PyAny { - self.context.clone().into_ref(py) - } -} - -#[pyclass] -struct PyTaskCompleter { - tx: Option<oneshot::Sender<PyResult<PyObject>>>, -} - -#[pymethods] -impl PyTaskCompleter { - #[args(task)] - pub fn __call__(&mut self, task: &PyAny) -> PyResult<()> { - debug_assert!(task.call_method0("done")?.extract()?); - let result = match task.call_method0("result") { - Ok(val) => Ok(val.into()), - Err(e) => Err(e), - }; - - // unclear to me whether or not this should be a panic or silent error. - // - // calling PyTaskCompleter twice should not be possible, but I don't think it really hurts - // anything if it happens. - if let Some(tx) = self.tx.take() { - if tx.send(result).is_err() { - // cancellation is not an error - } - } - - Ok(()) - } -} - -#[pyclass] -struct PyEnsureFuture { - awaitable: PyObject, - tx: Option<oneshot::Sender<PyResult<PyObject>>>, -} - -#[pymethods] -impl PyEnsureFuture { - pub fn __call__(&mut self) -> PyResult<()> { - Python::with_gil(\|py\| { - let task = ensure_future(py, self.awaitable.as_ref(py))?; - let on_complete = PyTaskCompleter { tx: self.tx.take() }; - task.call_method1("add_done_callback", (on_complete,))?; - - Ok(()) - }) - } -} - -fn call_soon_threadsafe( - event_loop: &PyAny, - context: &PyAny, - args: impl IntoPy<Py<PyTuple>>, -) -> PyResult<()> { - let py = event_loop.py(); - - let kwargs = PyDict::new(py); - kwargs.set_item("context", context)?; - - event_loop.call_method("call_soon_threadsafe", args, Some(kwargs))?; - Ok(()) -} - -/// Convert a Python `awaitable` into a Rust Future -/// -/// This function converts the `awaitable` into a Python Task using `run_coroutine_threadsafe`. A -/// completion handler sends the result of this Task through a -/// `futures::channel::oneshot::Sender<PyResult<PyObject>>` and the future returned by this function -/// simply awaits the result through the `futures::channel::oneshot::Receiver<PyResult<PyObject>>`. -/// -/// # Arguments -/// `locals` - The Python event loop and context to be used for the provided awaitable -/// * `awaitable` - The Python `awaitable` to be converted -/// -/// # Examples -/// -/// ``` -/// use std::time::Duration; -/// -/// use pyo3::prelude::; -/// -/// const PYTHON_CODE: &'static str = r#" -/// import asyncio -/// -/// async def py_sleep(duration): -/// await asyncio.sleep(duration) -/// "#; -/// -/// # #[cfg(feature = "tokio-runtime")] -/// async fn py_sleep(seconds: f32) -> PyResult<()> { -/// let test_mod = Python::with_gil(\|py\| -> PyResult<PyObject> { -/// Ok( -/// PyModule::from_code( -/// py, -/// PYTHON_CODE, -/// "test_into_future/test_mod.py", -/// "test_mod" -/// )? -/// .into() -/// ) -/// })?; -/// -/// Python::with_gil(\|py\| { -/// pyo3_asyncio::into_future_with_locals( -/// &pyo3_asyncio::tokio::get_current_locals(py)?, -/// test_mod -/// .call_method1(py, "py_sleep", (seconds.into_py(py),))? -/// .as_ref(py), -/// ) -/// })? -/// .await?; -/// Ok(()) -/// } -/// ``` -pub fn into_future_with_locals( - locals: &TaskLocals, - awaitable: &PyAny, -) -> PyResult<impl Future<Output = PyResult<PyObject>> + Send> { - let py = awaitable.py(); - let (tx, rx) = oneshot::channel(); - - call_soon_threadsafe( - locals.event_loop(py), - locals.context(py), - (PyEnsureFuture { - awaitable: awaitable.into(), - tx: Some(tx), - },), - )?; - - Ok(async move { - match rx.await { - Ok(item) => item, - Err(_) => Python::with_gil(\|py\| { - Err(PyErr::from_value( - asyncio(py)?.call_method0("CancelledError")?, - )) - }), - } - }) -} - -pub fn dump_err(py: Python<'_>) -> impl FnOnce(PyErr) + '_ { - move \|e\| { - // We can't display Python exceptions via std::fmt::Display, - // so print the error here manually. - e.print_and_set_sys_last_vars(py); - } -} diff --git a/lib/pyo3-asyncio/src/tokio.rs b/lib/pyo3-asyncio/src/tokio.rs deleted file mode 100644 index b56378b6..00000000 --- a/lib/pyo3-asyncio/src/tokio.rs +++ /dev/null @@ -1,787 +0,0 @@ -//! <span class="module-item stab portability" style="display: inline; border-radius: 3px; padding: 2px; font-size: 80%; line-height: 1.2;"><code>tokio-runtime</code></span> PyO3 Asyncio functions specific to the tokio runtime -//! -//! Items marked with -//! <span -//! class="module-item stab portability" -//! style="display: inline; border-radius: 3px; padding: 2px; font-size: 80%; line-height: 1.2;" -//! ><code>unstable-streams</code></span> -//! are only available when the `unstable-streams` Cargo feature is enabled: -//! -//! ```toml -//! [dependencies.pyo3-asyncio] -//! version = "0.17" -//! features = ["unstable-streams"] -//! ``` - -use std::{future::Future, pin::Pin, sync::Mutex}; - -use ::tokio::{ - runtime::{Builder, Runtime}, - task, -}; -use once_cell::{ - sync::{Lazy, OnceCell}, - unsync::OnceCell as UnsyncOnceCell, -}; -use pyo3::prelude::; - -use crate::{ - generic::{self, ContextExt, LocalContextExt, Runtime as GenericRuntime, SpawnLocalExt}, - TaskLocals, -}; - -/// <span class="module-item stab portability" style="display: inline; border-radius: 3px; padding: 2px; font-size: 80%; line-height: 1.2;"><code>attributes</code></span> -/// re-exports for macros -#[cfg(feature = "attributes")] -pub mod re_exports { - /// re-export pending to be used in tokio macros without additional dependency - pub use futures::future::pending; - /// re-export tokio::runtime to build runtimes in tokio macros without additional dependency - pub use tokio::runtime; -} - -/// <span class="module-item stab portability" style="display: inline; border-radius: 3px; padding: 2px; font-size: 80%; line-height: 1.2;"><code>attributes</code></span> -#[cfg(feature = "attributes")] -pub use pyo3_asyncio_macros::tokio_main as main; - -/// <span class="module-item stab portability" style="display: inline; border-radius: 3px; padding: 2px; font-size: 80%; line-height: 1.2;"><code>attributes</code></span> -/// <span class="module-item stab portability" style="display: inline; border-radius: 3px; padding: 2px; font-size: 80%; line-height: 1.2;"><code>testing</code></span> -/// Registers a `tokio` test with the `pyo3-asyncio` test harness -#[cfg(all(feature = "attributes", feature = "testing"))] -pub use pyo3_asyncio_macros::tokio_test as test; - -static TOKIO_BUILDER: Lazy<Mutex<Builder>> = Lazy::new(\|\| Mutex::new(multi_thread())); -static TOKIO_RUNTIME: OnceCell<Runtime> = OnceCell::new(); - -impl generic::JoinError for task::JoinError { - fn is_panic(&self) -> bool { - task::JoinError::is_panic(self) - } -} - -struct TokioRuntime; - -tokio::task_local! { - static TASK_LOCALS: UnsyncOnceCell<TaskLocals>; -} - -impl GenericRuntime for TokioRuntime { - type JoinError = task::JoinError; - type JoinHandle = task::JoinHandle<()>; - - fn spawn<F>(fut: F) -> Self::JoinHandle - where - F: Future<Output = ()> + Send + 'static, - { - get_runtime().spawn(async move { - fut.await; - }) - } -} - -impl ContextExt for TokioRuntime { - fn scope<F, R>(locals: TaskLocals, fut: F) -> Pin<Box<dyn Future<Output = R> + Send>> - where - F: Future<Output = R> + Send + 'static, - { - let cell = UnsyncOnceCell::new(); - cell.set(locals).unwrap(); - - Box::pin(TASK_LOCALS.scope(cell, fut)) - } - - fn get_task_locals() -> Option<TaskLocals> { - match TASK_LOCALS.try_with(\|c\| c.get().map(\|locals\| locals.clone())) { - Ok(locals) => locals, - Err(_) => None, - } - } -} - -impl SpawnLocalExt for TokioRuntime { - fn spawn_local<F>(fut: F) -> Self::JoinHandle - where - F: Future<Output = ()> + 'static, - { - tokio::task::spawn_local(fut) - } -} - -impl LocalContextExt for TokioRuntime { - fn scope_local<F, R>(locals: TaskLocals, fut: F) -> Pin<Box<dyn Future<Output = R>>> - where - F: Future<Output = R> + 'static, - { - let cell = UnsyncOnceCell::new(); - cell.set(locals).unwrap(); - - Box::pin(TASK_LOCALS.scope(cell, fut)) - } -} - -/// Set the task local event loop for the given future -pub async fn scope<F, R>(locals: TaskLocals, fut: F) -> R -where - F: Future<Output = R> + Send + 'static, -{ - TokioRuntime::scope(locals, fut).await -} - -/// Set the task local event loop for the given !Send future -pub async fn scope_local<F, R>(locals: TaskLocals, fut: F) -> R -where - F: Future<Output = R> + 'static, -{ - TokioRuntime::scope_local(locals, fut).await -} - -/// Get the current event loop from either Python or Rust async task local context -/// -/// This function first checks if the runtime has a task-local reference to the Python event loop. -/// If not, it calls [`get_running_loop`](`crate::get_running_loop`) to get the event loop -/// associated with the current OS thread. -pub fn get_current_loop(py: Python) -> PyResult<&PyAny> { - generic::get_current_loop::<TokioRuntime>(py) -} - -/// Either copy the task locals from the current task OR get the current running loop and -/// contextvars from Python. -pub fn get_current_locals(py: Python) -> PyResult<TaskLocals> { - generic::get_current_locals::<TokioRuntime>(py) -} - -/// Initialize the Tokio runtime with a custom build -pub fn init(builder: Builder) { - TOKIO_BUILDER.lock().unwrap() = builder -} - -/// Get a reference to the current tokio runtime -pub fn get_runtime<'a>() -> &'a Runtime { - TOKIO_RUNTIME.get_or_init(\|\| { - TOKIO_BUILDER - .lock() - .unwrap() - .build() - .expect("Unable to build Tokio runtime") - }) -} - -fn multi_thread() -> Builder { - let mut builder = Builder::new_multi_thread(); - builder.enable_all(); - builder -} - -/// Run the event loop until the given Future completes -/// -/// The event loop runs until the given future is complete. -/// -/// After this function returns, the event loop can be resumed with [`run_until_complete`] -/// -/// # Arguments -/// `event_loop` - The Python event loop that should run the future -/// * `fut` - The future to drive to completion -/// -/// # Examples -/// -/// ``` -/// # use std::time::Duration; -/// # -/// # use pyo3::prelude::; -/// # -/// # pyo3::prepare_freethreaded_python(); -/// # Python::with_gil(\|py\| -> PyResult<()> { -/// # let event_loop = py.import("asyncio")?.call_method0("new_event_loop")?; -/// pyo3_asyncio::tokio::run_until_complete(event_loop, async move { -/// tokio::time::sleep(Duration::from_secs(1)).await; -/// Ok(()) -/// })?; -/// # Ok(()) -/// # }).unwrap(); -/// ``` -pub fn run_until_complete<F, T>(event_loop: &PyAny, fut: F) -> PyResult<T> -where - F: Future<Output = PyResult<T>> + Send + 'static, - T: Send + Sync + 'static, -{ - generic::run_until_complete::<TokioRuntime, _, T>(event_loop, fut) -} - -/// Run the event loop until the given Future completes -/// -/// # Arguments -/// `py` - The current PyO3 GIL guard -/// * `fut` - The future to drive to completion -/// -/// # Examples -/// -/// ```no_run -/// # use std::time::Duration; -/// # -/// # use pyo3::prelude::; -/// # -/// fn main() { -/// Python::with_gil(\|py\| { -/// pyo3_asyncio::tokio::run(py, async move { -/// tokio::time::sleep(Duration::from_secs(1)).await; -/// Ok(()) -/// }) -/// .map_err(\|e\| { -/// e.print_and_set_sys_last_vars(py); -/// }) -/// .unwrap(); -/// }) -/// } -/// ``` -pub fn run<F, T>(py: Python, fut: F) -> PyResult<T> -where - F: Future<Output = PyResult<T>> + Send + 'static, - T: Send + Sync + 'static, -{ - generic::run::<TokioRuntime, F, T>(py, fut) -} - -/// Convert a Rust Future into a Python awaitable -/// -/// If the `asyncio.Future` returned by this conversion is cancelled via `asyncio.Future.cancel`, -/// the Rust future will be cancelled as well (new behaviour in `v0.15`). -/// -/// Python `contextvars` are preserved when calling async Python functions within the Rust future -/// via [`into_future`] (new behaviour in `v0.15`). -/// -/// > Although `contextvars` are preserved for async Python functions, synchronous functions will -/// unfortunately fail to resolve them when called within the Rust future. This is because the -/// function is being called from a Rust thread, not inside an actual Python coroutine context. -/// > -/// > As a workaround, you can get the `contextvars` from the current task locals using -/// [`get_current_locals`] and [`TaskLocals::context`](`crate::TaskLocals::context`), then wrap your -/// synchronous function in a call to `contextvars.Context.run`. This will set the context, call the -/// synchronous function, and restore the previous context when it returns or raises an exception. -/// -/// # Arguments -/// `py` - PyO3 GIL guard -/// * `locals` - The task locals for the given future -/// * `fut` - The Rust future to be converted -/// -/// # Examples -/// -/// ``` -/// use std::time::Duration; -/// -/// use pyo3::prelude::; -/// -/// /// Awaitable sleep function -/// #[pyfunction] -/// fn sleep_for<'p>(py: Python<'p>, secs: &'p PyAny) -> PyResult<&'p PyAny> { -/// let secs = secs.extract()?; -/// pyo3_asyncio::tokio::future_into_py_with_locals( -/// py, -/// pyo3_asyncio::tokio::get_current_locals(py)?, -/// async move { -/// tokio::time::sleep(Duration::from_secs(secs)).await; -/// Python::with_gil(\|py\| Ok(py.None())) -/// } -/// ) -/// } -/// ``` -pub fn future_into_py_with_locals<F, T>(py: Python, locals: TaskLocals, fut: F) -> PyResult<&PyAny> -where - F: Future<Output = PyResult<T>> + Send + 'static, - T: IntoPy<PyObject>, -{ - generic::future_into_py_with_locals::<TokioRuntime, F, T>(py, locals, fut) -} - -/// Convert a Rust Future into a Python awaitable -/// -/// If the `asyncio.Future` returned by this conversion is cancelled via `asyncio.Future.cancel`, -/// the Rust future will be cancelled as well (new behaviour in `v0.15`). -/// -/// Python `contextvars` are preserved when calling async Python functions within the Rust future -/// via [`into_future`] (new behaviour in `v0.15`). -/// -/// > Although `contextvars` are preserved for async Python functions, synchronous functions will -/// unfortunately fail to resolve them when called within the Rust future. This is because the -/// function is being called from a Rust thread, not inside an actual Python coroutine context. -/// > -/// > As a workaround, you can get the `contextvars` from the current task locals using -/// [`get_current_locals`] and [`TaskLocals::context`](`crate::TaskLocals::context`), then wrap your -/// synchronous function in a call to `contextvars.Context.run`. This will set the context, call the -/// synchronous function, and restore the previous context when it returns or raises an exception. -/// -/// # Arguments -/// `py` - The current PyO3 GIL guard -/// * `fut` - The Rust future to be converted -/// -/// # Examples -/// -/// ``` -/// use std::time::Duration; -/// -/// use pyo3::prelude::; -/// -/// /// Awaitable sleep function -/// #[pyfunction] -/// fn sleep_for<'p>(py: Python<'p>, secs: &'p PyAny) -> PyResult<&'p PyAny> { -/// let secs = secs.extract()?; -/// pyo3_asyncio::tokio::future_into_py(py, async move { -/// tokio::time::sleep(Duration::from_secs(secs)).await; -/// Ok(()) -/// }) -/// } -/// ``` -pub fn future_into_py<F, T>(py: Python, fut: F) -> PyResult<&PyAny> -where - F: Future<Output = PyResult<T>> + Send + 'static, - T: IntoPy<PyObject>, -{ - generic::future_into_py::<TokioRuntime, _, T>(py, fut) -} - -/// Convert a `!Send` Rust Future into a Python awaitable -/// -/// If the `asyncio.Future` returned by this conversion is cancelled via `asyncio.Future.cancel`, -/// the Rust future will be cancelled as well (new behaviour in `v0.15`). -/// -/// Python `contextvars` are preserved when calling async Python functions within the Rust future -/// via [`into_future`] (new behaviour in `v0.15`). -/// -/// > Although `contextvars` are preserved for async Python functions, synchronous functions will -/// unfortunately fail to resolve them when called within the Rust future. This is because the -/// function is being called from a Rust thread, not inside an actual Python coroutine context. -/// > -/// > As a workaround, you can get the `contextvars` from the current task locals using -/// [`get_current_locals`] and [`TaskLocals::context`](`crate::TaskLocals::context`), then wrap your -/// synchronous function in a call to `contextvars.Context.run`. This will set the context, call the -/// synchronous function, and restore the previous context when it returns or raises an exception. -/// -/// # Arguments -/// `py` - PyO3 GIL guard -/// * `locals` - The task locals for the given future -/// * `fut` - The Rust future to be converted -/// -/// # Examples -/// -/// ``` -/// use std::{rc::Rc, time::Duration}; -/// -/// use pyo3::prelude::; -/// -/// /// Awaitable non-send sleep function -/// #[pyfunction] -/// fn sleep_for(py: Python, secs: u64) -> PyResult<&PyAny> { -/// // Rc is non-send so it cannot be passed into pyo3_asyncio::tokio::future_into_py -/// let secs = Rc::new(secs); -/// -/// pyo3_asyncio::tokio::local_future_into_py_with_locals( -/// py, -/// pyo3_asyncio::tokio::get_current_locals(py)?, -/// async move { -/// tokio::time::sleep(Duration::from_secs(secs)).await; -/// Python::with_gil(\|py\| Ok(py.None())) -/// } -/// ) -/// } -/// -/// # #[cfg(all(feature = "tokio-runtime", feature = "attributes"))] -/// #[pyo3_asyncio::tokio::main] -/// async fn main() -> PyResult<()> { -/// let locals = Python::with_gil(\|py\| -> PyResult<_> { -/// pyo3_asyncio::tokio::get_current_locals(py) -/// })?; -/// -/// // the main coroutine is running in a Send context, so we cannot use LocalSet here. Instead -/// // we use spawn_blocking in order to use LocalSet::block_on -/// tokio::task::spawn_blocking(move \|\| { -/// // LocalSet allows us to work with !Send futures within tokio. Without it, any calls to -/// // pyo3_asyncio::tokio::local_future_into_py will panic. -/// tokio::task::LocalSet::new().block_on( -/// pyo3_asyncio::tokio::get_runtime(), -/// pyo3_asyncio::tokio::scope_local(locals, async { -/// Python::with_gil(\|py\| { -/// let py_future = sleep_for(py, 1)?; -/// pyo3_asyncio::tokio::into_future(py_future) -/// })? -/// .await?; -/// -/// Ok(()) -/// }) -/// ) -/// }).await.unwrap() -/// } -/// # #[cfg(not(all(feature = "tokio-runtime", feature = "attributes")))] -/// # fn main() {} -/// ``` -pub fn local_future_into_py_with_locals<F, T>( - py: Python, - locals: TaskLocals, - fut: F, -) -> PyResult<&PyAny> -where - F: Future<Output = PyResult<T>> + 'static, - T: IntoPy<PyObject>, -{ - generic::local_future_into_py_with_locals::<TokioRuntime, _, T>(py, locals, fut) -} - -/// Convert a `!Send` Rust Future into a Python awaitable -/// -/// If the `asyncio.Future` returned by this conversion is cancelled via `asyncio.Future.cancel`, -/// the Rust future will be cancelled as well (new behaviour in `v0.15`). -/// -/// Python `contextvars` are preserved when calling async Python functions within the Rust future -/// via [`into_future`] (new behaviour in `v0.15`). -/// -/// > Although `contextvars` are preserved for async Python functions, synchronous functions will -/// unfortunately fail to resolve them when called within the Rust future. This is because the -/// function is being called from a Rust thread, not inside an actual Python coroutine context. -/// > -/// > As a workaround, you can get the `contextvars` from the current task locals using -/// [`get_current_locals`] and [`TaskLocals::context`](`crate::TaskLocals::context`), then wrap your -/// synchronous function in a call to `contextvars.Context.run`. This will set the context, call the -/// synchronous function, and restore the previous context when it returns or raises an exception. -/// -/// # Arguments -/// * `py` - The current PyO3 GIL guard -/// * `fut` - The Rust future to be converted -/// -/// # Examples -/// -/// ``` -/// use std::{rc::Rc, time::Duration}; -/// -/// use pyo3::prelude::; -/// -/// /// Awaitable non-send sleep function -/// #[pyfunction] -/// fn sleep_for(py: Python, secs: u64) -> PyResult<&PyAny> { -/// // Rc is non-send so it cannot be passed into pyo3_asyncio::tokio::future_into_py -/// let secs = Rc::new(secs); -/// pyo3_asyncio::tokio::local_future_into_py(py, async move { -/// tokio::time::sleep(Duration::from_secs(secs)).await; -/// Ok(()) -/// }) -/// } -/// -/// # #[cfg(all(feature = "tokio-runtime", feature = "attributes"))] -/// #[pyo3_asyncio::tokio::main] -/// async fn main() -> PyResult<()> { -/// let locals = Python::with_gil(\|py\| { -/// pyo3_asyncio::tokio::get_current_locals(py).unwrap() -/// }); -/// -/// // the main coroutine is running in a Send context, so we cannot use LocalSet here. Instead -/// // we use spawn_blocking in order to use LocalSet::block_on -/// tokio::task::spawn_blocking(move \|\| { -/// // LocalSet allows us to work with !Send futures within tokio. Without it, any calls to -/// // pyo3_asyncio::tokio::local_future_into_py will panic. -/// tokio::task::LocalSet::new().block_on( -/// pyo3_asyncio::tokio::get_runtime(), -/// pyo3_asyncio::tokio::scope_local(locals, async { -/// Python::with_gil(\|py\| { -/// let py_future = sleep_for(py, 1)?; -/// pyo3_asyncio::tokio::into_future(py_future) -/// })? -/// .await?; -/// -/// Ok(()) -/// }) -/// ) -/// }).await.unwrap() -/// } -/// # #[cfg(not(all(feature = "tokio-runtime", feature = "attributes")))] -/// # fn main() {} -/// ``` -pub fn local_future_into_py<F, T>(py: Python, fut: F) -> PyResult<&PyAny> -where - F: Future<Output = PyResult<T>> + 'static, - T: IntoPy<PyObject>, -{ - generic::local_future_into_py::<TokioRuntime, _, T>(py, fut) -} - -/// Convert a Python `awaitable` into a Rust Future -/// -/// This function converts the `awaitable` into a Python Task using `run_coroutine_threadsafe`. A -/// completion handler sends the result of this Task through a -/// `futures::channel::oneshot::Sender<PyResult<PyObject>>` and the future returned by this function -/// simply awaits the result through the `futures::channel::oneshot::Receiver<PyResult<PyObject>>`. -/// -/// # Arguments -/// * `awaitable` - The Python `awaitable` to be converted -/// -/// # Examples -/// -/// ``` -/// use std::time::Duration; -/// -/// use pyo3::prelude::; -/// -/// const PYTHON_CODE: &'static str = r#" -/// import asyncio -/// -/// async def py_sleep(duration): -/// await asyncio.sleep(duration) -/// "#; -/// -/// async fn py_sleep(seconds: f32) -> PyResult<()> { -/// let test_mod = Python::with_gil(\|py\| -> PyResult<PyObject> { -/// Ok( -/// PyModule::from_code( -/// py, -/// PYTHON_CODE, -/// "test_into_future/test_mod.py", -/// "test_mod" -/// )? -/// .into() -/// ) -/// })?; -/// -/// Python::with_gil(\|py\| { -/// pyo3_asyncio::tokio::into_future( -/// test_mod -/// .call_method1(py, "py_sleep", (seconds.into_py(py),))? -/// .as_ref(py), -/// ) -/// })? -/// .await?; -/// Ok(()) -/// } -/// ``` -pub fn into_future(awaitable: &PyAny) -> PyResult<impl Future<Output = PyResult<PyObject>> + Send> { - generic::into_future::<TokioRuntime>(awaitable) -} - -/// <span class="module-item stab portability" style="display: inline; border-radius: 3px; padding: 2px; font-size: 80%; line-height: 1.2;"><code>unstable-streams</code></span> Convert an async generator into a stream -/// -/// This API is marked as unstable* and is only available when the -/// `unstable-streams` crate feature is enabled. This comes with no -/// stability guarantees, and could be changed or removed at any time. -/// -/// # Arguments -/// * `locals` - The current task locals -/// * `gen` - The Python async generator to be converted -/// -/// # Examples -/// ``` -/// use pyo3::prelude::; -/// use futures::{StreamExt, TryStreamExt}; -/// -/// const TEST_MOD: &str = r#" -/// import asyncio -/// -/// async def gen(): -/// for i in range(10): -/// await asyncio.sleep(0.1) -/// yield i -/// "#; -/// -/// # #[cfg(all(feature = "unstable-streams", feature = "attributes"))] -/// # #[pyo3_asyncio::tokio::main] -/// # async fn main() -> PyResult<()> { -/// let stream = Python::with_gil(\|py\| { -/// let test_mod = PyModule::from_code( -/// py, -/// TEST_MOD, -/// "test_rust_coroutine/test_mod.py", -/// "test_mod", -/// )?; -/// -/// pyo3_asyncio::tokio::into_stream_with_locals_v1( -/// pyo3_asyncio::tokio::get_current_locals(py)?, -/// test_mod.call_method0("gen")? -/// ) -/// })?; -/// -/// let vals = stream -/// .map(\|item\| Python::with_gil(\|py\| -> PyResult<i32> { Ok(item?.as_ref(py).extract()?) })) -/// .try_collect::<Vec<i32>>() -/// .await?; -/// -/// assert_eq!((0..10).collect::<Vec<i32>>(), vals); -/// -/// Ok(()) -/// # } -/// # #[cfg(not(all(feature = "unstable-streams", feature = "attributes")))] -/// # fn main() {} -/// ``` -#[cfg(feature = "unstable-streams")] -pub fn into_stream_with_locals_v1<'p>( - locals: TaskLocals, - gen: &'p PyAny, -) -> PyResult<impl futures::Stream<Item = PyResult<PyObject>> + 'static> { - generic::into_stream_with_locals_v1::<TokioRuntime>(locals, gen) -} - -/// <span class="module-item stab portability" style="display: inline; border-radius: 3px; padding: 2px; font-size: 80%; line-height: 1.2;"><code>unstable-streams</code></span> Convert an async generator into a stream -/// -/// This API is marked as unstable* and is only available when the -/// `unstable-streams` crate feature is enabled. This comes with no -/// stability guarantees, and could be changed or removed at any time. -/// -/// # Arguments -/// * `gen` - The Python async generator to be converted -/// -/// # Examples -/// ``` -/// use pyo3::prelude::; -/// use futures::{StreamExt, TryStreamExt}; -/// -/// const TEST_MOD: &str = r#" -/// import asyncio -/// -/// async def gen(): -/// for i in range(10): -/// await asyncio.sleep(0.1) -/// yield i -/// "#; -/// -/// # #[cfg(all(feature = "unstable-streams", feature = "attributes"))] -/// # #[pyo3_asyncio::tokio::main] -/// # async fn main() -> PyResult<()> { -/// let stream = Python::with_gil(\|py\| { -/// let test_mod = PyModule::from_code( -/// py, -/// TEST_MOD, -/// "test_rust_coroutine/test_mod.py", -/// "test_mod", -/// )?; -/// -/// pyo3_asyncio::tokio::into_stream_v1(test_mod.call_method0("gen")?) -/// })?; -/// -/// let vals = stream -/// .map(\|item\| Python::with_gil(\|py\| -> PyResult<i32> { Ok(item?.as_ref(py).extract()?) })) -/// .try_collect::<Vec<i32>>() -/// .await?; -/// -/// assert_eq!((0..10).collect::<Vec<i32>>(), vals); -/// -/// Ok(()) -/// # } -/// # #[cfg(not(all(feature = "unstable-streams", feature = "attributes")))] -/// # fn main() {} -/// ``` -#[cfg(feature = "unstable-streams")] -pub fn into_stream_v1<'p>( - gen: &'p PyAny, -) -> PyResult<impl futures::Stream<Item = PyResult<PyObject>> + 'static> { - generic::into_stream_v1::<TokioRuntime>(gen) -} - -/// <span class="module-item stab portability" style="display: inline; border-radius: 3px; padding: 2px; font-size: 80%; line-height: 1.2;"><code>unstable-streams</code></span> Convert an async generator into a stream -/// -/// This API is marked as unstable* and is only available when the -/// `unstable-streams` crate feature is enabled. This comes with no -/// stability guarantees, and could be changed or removed at any time. -/// -/// # Arguments -/// * `locals` - The current task locals -/// * `gen` - The Python async generator to be converted -/// -/// # Examples -/// ``` -/// use pyo3::prelude::; -/// use futures::{StreamExt, TryStreamExt}; -/// -/// const TEST_MOD: &str = r#" -/// import asyncio -/// -/// async def gen(): -/// for i in range(10): -/// await asyncio.sleep(0.1) -/// yield i -/// "#; -/// -/// # #[cfg(all(feature = "unstable-streams", feature = "attributes"))] -/// # #[pyo3_asyncio::tokio::main] -/// # async fn main() -> PyResult<()> { -/// let stream = Python::with_gil(\|py\| { -/// let test_mod = PyModule::from_code( -/// py, -/// TEST_MOD, -/// "test_rust_coroutine/test_mod.py", -/// "test_mod", -/// )?; -/// -/// pyo3_asyncio::tokio::into_stream_with_locals_v2( -/// pyo3_asyncio::tokio::get_current_locals(py)?, -/// test_mod.call_method0("gen")? -/// ) -/// })?; -/// -/// let vals = stream -/// .map(\|item\| Python::with_gil(\|py\| -> PyResult<i32> { Ok(item.as_ref(py).extract()?) })) -/// .try_collect::<Vec<i32>>() -/// .await?; -/// -/// assert_eq!((0..10).collect::<Vec<i32>>(), vals); -/// -/// Ok(()) -/// # } -/// # #[cfg(not(all(feature = "unstable-streams", feature = "attributes")))] -/// # fn main() {} -/// ``` -#[cfg(feature = "unstable-streams")] -pub fn into_stream_with_locals_v2<'p>( - locals: TaskLocals, - gen: &'p PyAny, -) -> PyResult<impl futures::Stream<Item = PyObject> + 'static> { - generic::into_stream_with_locals_v2::<TokioRuntime>(locals, gen) -} - -/// <span class="module-item stab portability" style="display: inline; border-radius: 3px; padding: 2px; font-size: 80%; line-height: 1.2;"><code>unstable-streams</code></span> Convert an async generator into a stream -/// -/// This API is marked as unstable* and is only available when the -/// `unstable-streams` crate feature is enabled. This comes with no -/// stability guarantees, and could be changed or removed at any time. -/// -/// # Arguments -/// * `gen` - The Python async generator to be converted -/// -/// # Examples -/// ``` -/// use pyo3::prelude::; -/// use futures::{StreamExt, TryStreamExt}; -/// -/// const TEST_MOD: &str = r#" -/// import asyncio -/// -/// async def gen(): -/// for i in range(10): -/// await asyncio.sleep(0.1) -/// yield i -/// "#; -/// -/// # #[cfg(all(feature = "unstable-streams", feature = "attributes"))] -/// # #[pyo3_asyncio::tokio::main] -/// # async fn main() -> PyResult<()> { -/// let stream = Python::with_gil(\|py\| { -/// let test_mod = PyModule::from_code( -/// py, -/// TEST_MOD, -/// "test_rust_coroutine/test_mod.py", -/// "test_mod", -/// )?; -/// -/// pyo3_asyncio::tokio::into_stream_v2(test_mod.call_method0("gen")?) -/// })?; -/// -/// let vals = stream -/// .map(\|item\| Python::with_gil(\|py\| -> PyResult<i32> { Ok(item.as_ref(py).extract()?) })) -/// .try_collect::<Vec<i32>>() -/// .await?; -/// -/// assert_eq!((0..10).collect::<Vec<i32>>(), vals); -/// -/// Ok(()) -/// # } -/// # #[cfg(not(all(feature = "unstable-streams", feature = "attributes")))] -/// # fn main() {} -/// ``` -#[cfg(feature = "unstable-streams")] -pub fn into_stream_v2<'p>( - gen: &'p PyAny, -) -> PyResult<impl futures::Stream<Item = PyObject> + 'static> { - generic::into_stream_v2::<TokioRuntime>(gen) -} diff --git a/src/asgi/io.rs b/src/asgi/io.rs index 6db89ea5..8396cfda 100644 --- a/src/asgi/io.rs +++ b/src/asgi/io.rs @@ -14,7 +14,7 @@ use tungstenite::Message; use crate::{ http::HV_SERVER, - runtime::{RuntimeRef, future_into_py}, + runtime::{RuntimeRef, future_into_py_iter, future_into_py_futlike}, ws::{HyperWebsocket, UpgradeData} }; use super::{ @@ -81,7 +81,7 @@ impl ASGIHTTPProtocol { impl ASGIHTTPProtocol { fn receive<'p>(&mut self, py: Python<'p>) -> PyResult<&'p PyAny> { let transport = self.request.clone(); - future_into_py(self.rt.clone(), py, async move { + future_into_py_iter(self.rt.clone(), py, async move { let mut req = transport.lock().await; let body = hyper::body::to_bytes(&mut req).await.unwrap(); Python::with_gil(\|py\| { @@ -196,7 +196,7 @@ impl ASGIWebsocketProtocol { let accepted = self.accepted.clone(); let tx = self.ws_tx.clone(); let rx = self.ws_rx.clone(); - future_into_py(self.rt.clone(), py, async move { + future_into_py_iter(self.rt.clone(), py, async move { if let Ok(_) = upgrade.send().await { if let Ok(stream) = websocket.await { let mut wtx = tx.lock().await; @@ -222,7 +222,7 @@ impl ASGIWebsocketProtocol { let transport = self.ws_tx.clone(); let closed = self.closed.clone(); let message = ws_message_into_rs(data); - future_into_py(self.rt.clone(), py, async move { + future_into_py_iter(self.rt.clone(), py, async move { if !closed { if let Ok(message) = message { if let Some(ws) = &mut (transport.lock().await) { @@ -250,7 +250,7 @@ impl ASGIWebsocketProtocol { macro_rules! empty_future { ($rt:expr, $py:expr) => { - future_into_py($rt, $py, async move { + future_into_py_iter($rt, $py, async move { Ok(()) }) }; @@ -262,7 +262,7 @@ impl ASGIWebsocketProtocol { let transport = self.ws_rx.clone(); let accepted = self.accepted.clone(); let closed = self.closed.clone(); - future_into_py(self.rt.clone(), py, async move { + future_into_py_futlike(self.rt.clone(), py, async move { let accepted = accepted.lock().await; match (accepted, closed) { (false, false) => { diff --git a/src/callbacks.rs b/src/callbacks.rs index fb829f08..b1206eee 100644 --- a/src/callbacks.rs +++ b/src/callbacks.rs @@ -1,4 +1,6 @@ use pyo3::prelude::; +use pyo3::pyclass::IterNextOutput; + #[derive(Clone)] pub(crate) struct CallbackWrapper { @@ -7,10 +9,172 @@ pub(crate) struct CallbackWrapper { } impl CallbackWrapper { - pub(crate) fn new(callback: PyObject, event_loop: &PyAny, context: &PyAny) -> Self { + pub(crate) fn new( + callback: PyObject, + event_loop: &PyAny, + context: &PyAny + ) -> Self { Self { - callback: callback, + callback, context: pyo3_asyncio::TaskLocals::new(event_loop).with_context(context) } } } + +#[pyclass] +pub(crate) struct PyAwaitableResultYielder { + result: Option<PyResult<PyObject>>, + none: PyObject +} + +impl PyAwaitableResultYielder { + pub(crate) fn set_result(&mut self, result: PyResult<PyObject>) { + self.result = Some(result) + } +} + +#[pymethods] +impl PyAwaitableResultYielder { + fn __iter__(pyself: PyRef<'_, Self>) -> PyRef<'_, Self> { + pyself + } + + fn __next__(mut pyself: PyRefMut<'_, Self>) -> PyResult<IterNextOutput<PyObject, PyObject>> { + match pyself.result.take() { + Some(res) => { + match res { + Ok(v) => Ok(IterNextOutput::Return(v)), + Err(err) => Err(err) + } + }, + _ => Ok(IterNextOutput::Yield(pyself.none.clone())) + } + } +} + +#[pyclass] +pub(crate) struct PyAwaitableResultFutureLike { + py_block: bool, + event_loop: PyObject, + result: Option<PyResult<PyObject>>, + cb: Option<(PyObject, Py<pyo3::types::PyDict>)> +} + +impl PyAwaitableResultFutureLike { + pub(crate) fn new(event_loop: PyObject) -> Self { + Self { + event_loop, + py_block: true, + result: None, + cb: None + } + } + + pub(crate) fn set_result(mut pyself: PyRefMut<'_, Self>, result: PyResult<PyObject>) { + pyself.result = Some(result); + if let Some((cb, ctx)) = pyself.cb.take() { + Python::with_gil(\|py\| { + let _ = pyself.event_loop.call_method( + py, + "call_soon_threadsafe", + (cb, &pyself), + Some(ctx.as_ref(py)) + ); + }) + } + } +} + +#[pymethods] +impl PyAwaitableResultFutureLike { + #[getter(_asyncio_future_blocking)] + fn get_block(&self) -> bool { + self.py_block + } + + #[setter(_asyncio_future_blocking)] + fn set_block(&mut self, val: bool) { + self.py_block = val + } + + fn get_loop(&mut self) -> PyObject { + self.event_loop.clone() + } + + fn add_done_callback( + mut pyself: PyRefMut<'_, Self>, + py: Python, + cb: PyObject, + context: PyObject + ) -> PyResult<()> { + let kwctx = pyo3::types::PyDict::new(py); + kwctx.set_item("context", context)?; + match pyself.result { + Some(_) => { + pyself.event_loop.call_method(py, "call_soon", (cb, &pyself), Some(kwctx))?; + }, + _ => { + pyself.cb = Some((cb, kwctx.into_py(py))); + } + } + Ok(()) + } + + fn result(&self) {} + + fn __iter__(pyself: PyRef<'_, Self>) -> PyRef<'_, Self> { + pyself + } + + fn __next__( + mut pyself: PyRefMut<'_, Self> + ) -> PyResult<IterNextOutput<PyRefMut<'_, Self>, PyObject>> { + match pyself.result { + Some(_) => { + match pyself.result.take().unwrap() { + Ok(v) => Ok(IterNextOutput::Return(v)), + Err(err) => Err(err) + } + }, + _ => Ok(IterNextOutput::Yield(pyself)) + } + } +} + +#[pyclass] +pub(crate) struct PyIterAwaitableResult { + pub inner: Py<PyAwaitableResultYielder> +} + +impl PyIterAwaitableResult { + pub(crate) fn new(py: Python) -> PyResult<Self> { + let inner = Py::new(py, PyAwaitableResultYielder { result: None, none: py.None() })?; + Ok(Self { inner }) + } +} + +#[pymethods] +impl PyIterAwaitableResult { + fn __await__(&mut self, py: Python) -> PyObject { + self.inner.to_object(py) + } +} + +#[pyclass] +pub(crate) struct PyFutureAwaitableResult { + pub inner: Py<PyAwaitableResultFutureLike> +} + +impl PyFutureAwaitableResult { + pub(crate) fn new(py: Python, event_loop: &PyAny) -> PyResult<Self> { + let inner = Py::new(py, PyAwaitableResultFutureLike::new(event_loop.to_object(py)))?; + Ok(Self { inner }) + } +} + +#[pymethods] +impl PyFutureAwaitableResult { + fn __await__(&mut self, py: Python) -> PyObject { + self.inner.to_object(py) + } +} diff --git a/src/rsgi/io.rs b/src/rsgi/io.rs index 8fcb80cb..bcaf7e21 100644 --- a/src/rsgi/io.rs +++ b/src/rsgi/io.rs @@ -8,7 +8,7 @@ use tokio::sync::{oneshot, Mutex}; use tungstenite::Message; use crate::{ - runtime::{RuntimeRef, future_into_py}, + runtime::{RuntimeRef, future_into_py_iter, future_into_py_futlike}, ws::{HyperWebsocket, UpgradeData} }; use super::{ @@ -31,7 +31,7 @@ impl RSGIHTTPProtocol { request: Request<Body> ) -> Self { Self { - rt: rt, + rt, tx: Some(tx), request: Arc::new(Mutex::new(request)) } @@ -46,7 +46,7 @@ impl RSGIHTTPProtocol { impl RSGIHTTPProtocol { fn __call__<'p>(&self, py: Python<'p>) -> PyResult<&'p PyAny> { let req_ref = self.request.clone(); - future_into_py(self.rt.clone(), py, async move { + future_into_py_iter(self.rt.clone(), py, async move { let mut req = req_ref.lock().await; let body = hyper::body::to_bytes(&mut req).await.unwrap(); Ok(Python::with_gil(\|py\| { @@ -113,7 +113,7 @@ impl RSGIWebsocketTransport { impl RSGIWebsocketTransport { fn receive<'p>(&self, py: Python<'p>) -> PyResult<&'p PyAny> { let transport = self.rx.clone(); - future_into_py(self.rt.clone(), py, async move { + future_into_py_futlike(self.rt.clone(), py, async move { if let Ok(mut stream) = transport.try_lock() { loop { match stream.next().await { @@ -143,7 +143,7 @@ impl RSGIWebsocketTransport { fn send_bytes<'p>(&self, py: Python<'p>, data: Vec<u8>) -> PyResult<&'p PyAny> { let transport = self.tx.clone(); - future_into_py(self.rt.clone(), py, async move { + future_into_py_iter(self.rt.clone(), py, async move { if let Ok(mut stream) = transport.try_lock() { return match stream.send(Message::Binary(data)).await { Ok(_) => Ok(()), @@ -156,7 +156,7 @@ impl RSGIWebsocketTransport { fn send_str<'p>(&self, py: Python<'p>, data: String) -> PyResult<&'p PyAny> { let transport = self.tx.clone(); - future_into_py(self.rt.clone(), py, async move { + future_into_py_iter(self.rt.clone(), py, async move { if let Ok(mut stream) = transport.try_lock() { return match stream.send(Message::Text(data)).await { Ok(_) => Ok(()), @@ -266,7 +266,7 @@ impl RSGIWebsocketProtocol { let rth = self.rt.clone(); let mut upgrade = self.upgrade.take().unwrap(); let transport = self.websocket.clone(); - future_into_py(self.rt.clone(), py, async move { + future_into_py_iter(self.rt.clone(), py, async move { let mut ws = transport.lock().await; match upgrade.send().await { Ok(_) => { diff --git a/src/runtime.rs b/src/runtime.rs index 75acd5cc..202919a7 100644 --- a/src/runtime.rs +++ b/src/runtime.rs @@ -1,9 +1,16 @@ use once_cell::{unsync::OnceCell as UnsyncOnceCell}; -use pyo3_asyncio::{TaskLocals, generic as pyrt_generic}; +use pyo3_asyncio::TaskLocals; use pyo3::prelude::; use std::{future::Future, io, pin::Pin, sync::{Arc, Mutex}}; use tokio::{runtime::Builder, task::{JoinHandle, LocalSet}}; +use super::callbacks::{ + PyIterAwaitableResult, + PyFutureAwaitableResult, + PyAwaitableResultFutureLike +}; + + tokio::task_local! { static TASK_LOCALS: UnsyncOnceCell<TaskLocals>; } @@ -182,160 +189,6 @@ pub(crate) fn into_future( ) } -pub(crate) fn future_into_py_with_locals<R, F, T>( - rt: R, - py: Python, - locals: TaskLocals, - fut: F, -) -> PyResult<&PyAny> -where - R: Runtime + ContextExt + Clone, - F: Future<Output = PyResult<T>> + Send + 'static, - T: IntoPy<PyObject>, -{ - let (cancel_tx, cancel_rx) = futures::channel::oneshot::channel(); - - let py_fut = pyo3_asyncio::create_future(locals.event_loop(py))?; - py_fut.call_method1( - "add_done_callback", - (pyrt_generic::PyDoneCallback { - cancel_tx: Some(cancel_tx), - },), - )?; - - let future_tx1 = PyObject::from(py_fut); - let future_tx2 = future_tx1.clone(); - let rth = rt.handler(); - - rt.spawn(async move { - let rti = rth.handler(); - let locals2 = locals.clone(); - - if let Err(e) = rth.spawn(async move { - let result = rti.scope( - locals2.clone(), - pyrt_generic::Cancellable::new_with_cancel_rx(fut, cancel_rx), - ) - .await; - - Python::with_gil(move \|py\| { - if pyrt_generic::cancelled(future_tx1.as_ref(py)) - .map_err(pyo3_asyncio::dump_err(py)) - .unwrap_or(false) - { - return; - } - - let _ = pyrt_generic::set_result( - locals2.event_loop(py), - future_tx1.as_ref(py), - result.map(\|val\| val.into_py(py)), - ) - .map_err(pyo3_asyncio::dump_err(py)); - }); - }) - .await - { - if e.is_panic() { - Python::with_gil(move \|py\| { - if pyrt_generic::cancelled(future_tx2.as_ref(py)) - .map_err(pyo3_asyncio::dump_err(py)) - .unwrap_or(false) - { - return; - } - - let _ = pyrt_generic::set_result( - locals.event_loop(py), - future_tx2.as_ref(py), - Err(pyo3_asyncio::err::RustPanic::new_err("rust future panicked")), - ) - .map_err(pyo3_asyncio::dump_err(py)); - }); - } - } - }); - - Ok(py_fut) -} - -pub fn local_future_into_py_with_locals<R, F, T>( - rt: R, - py: Python, - locals: TaskLocals, - fut: F, -) -> PyResult<&PyAny> -where - R: Runtime + SpawnLocalExt + LocalContextExt, - F: Future<Output = PyResult<T>> + 'static, - T: IntoPy<PyObject>, -{ - let (cancel_tx, cancel_rx) = futures::channel::oneshot::channel(); - - let py_fut = pyo3_asyncio::create_future(locals.event_loop(py))?; - py_fut.call_method1( - "add_done_callback", - (pyrt_generic::PyDoneCallback { - cancel_tx: Some(cancel_tx), - },), - )?; - - let future_tx1 = PyObject::from(py_fut); - let future_tx2 = future_tx1.clone(); - let rth = rt.handler(); - - rt.spawn_local(async move { - let rti = rth.handler(); - let locals2 = locals.clone(); - - if let Err(e) = rth.spawn_local(async move { - let result = rti.scope_local( - locals2.clone(), - pyrt_generic::Cancellable::new_with_cancel_rx(fut, cancel_rx), - ) - .await; - - Python::with_gil(move \|py\| { - if pyrt_generic::cancelled(future_tx1.as_ref(py)) - .map_err(pyo3_asyncio::dump_err(py)) - .unwrap_or(false) - { - return; - } - - let _ = pyrt_generic::set_result( - locals2.event_loop(py), - future_tx1.as_ref(py), - result.map(\|val\| val.into_py(py)), - ) - .map_err(pyo3_asyncio::dump_err(py)); - }); - }) - .await - { - if e.is_panic() { - Python::with_gil(move \|py\| { - if pyrt_generic::cancelled(future_tx2.as_ref(py)) - .map_err(pyo3_asyncio::dump_err(py)) - .unwrap_or(false) - { - return; - } - - let _ = pyrt_generic::set_result( - locals.event_loop(py), - future_tx2.as_ref(py), - Err(pyo3_asyncio::err::RustPanic::new_err("Rust future panicked")), - ) - .map_err(pyo3_asyncio::dump_err(py)); - }); - } - } - }); - - Ok(py_fut) -} - #[inline] fn get_current_locals<R>(py: Python) -> PyResult<TaskLocals> where @@ -348,23 +201,58 @@ where } } -pub(crate) fn future_into_py<R, F, T>(rt: R, py: Python, fut: F) -> PyResult<&PyAny> +// NOTE: +// `future_into_py_iter` relies on what CPython refers as "bare yield". +// This is generally ~55% faster than `pyo3_asyncio.future_into_py` implementation. +// It consumes more cpu-cycles than `future_into_py_futlike`, +// but for "quick" operations it's something like 12% faster. +pub(crate) fn future_into_py_iter<R, F, T>(rt: R, py: Python, fut: F) -> PyResult<&PyAny> where R: Runtime + ContextExt + Clone, F: Future<Output = PyResult<T>> + Send + 'static, T: IntoPy<PyObject>, { - future_into_py_with_locals::<R, F, T>(rt, py, get_current_locals::<R>(py)?, fut) + let aw = PyIterAwaitableResult::new(py)?; + let fut_res = aw.inner.clone(); + let py_fut = aw.into_py(py).into_ref(py); + + rt.spawn(async move { + let result = fut.await; + Python::with_gil(move \|py\| { + fut_res.as_ref(py).borrow_mut().set_result(result.map(\|v\| v.into_py(py))); + }); + }); + + Ok(py_fut) } -#[allow(dead_code)] -pub fn local_future_into_py<R, F, T>(rt: R, py: Python, fut: F) -> PyResult<&PyAny> +// NOTE: +// `future_into_py_futlike` relies on an `asyncio.Future` like implementation. +// This is generally ~38% faster than `pyo3_asyncio.future_into_py` implementation. +// It won't consume more cpu-cycles than standard asyncio implementation, +// and for "long" operations it's something like 6% faster than `future_into_py_iter`. +pub(crate) fn future_into_py_futlike<R, F, T>(rt: R, py: Python, fut: F) -> PyResult<&PyAny> where - R: Runtime + ContextExt + SpawnLocalExt + LocalContextExt, - F: Future<Output = PyResult<T>> + 'static, + R: Runtime + ContextExt + Clone, + F: Future<Output = PyResult<T>> + Send + 'static, T: IntoPy<PyObject>, { - local_future_into_py_with_locals::<R, F, T>(rt, py, get_current_locals::<R>(py)?, fut) + let task_locals = get_current_locals::<R>(py)?; + let aw = PyFutureAwaitableResult::new(py, task_locals.event_loop(py))?; + let fut_res = aw.inner.clone(); + let py_fut = aw.into_py(py).into_ref(py); + + rt.spawn(async move { + let result = fut.await; + Python::with_gil(move \|py\| { + PyAwaitableResultFutureLike::set_result( + fut_res.as_ref(py).borrow_mut(), + result.map(\|v\| v.into_py(py)) + ); + }); + }); + + Ok(py_fut) } pub(crate) fn run_until_complete<R, F, T>(rt: R, event_loop: &PyAny, fut: F) -> PyResult<T> @@ -376,20 +264,31 @@ where let py = event_loop.py(); let result_tx = Arc::new(Mutex::new(None)); let result_rx = Arc::clone(&result_tx); - let coro = future_into_py_with_locals::<R, _, ()>( - rt, - py, - TaskLocals::new(event_loop).copy_context(py)?, - async move { - let val = fut.await?; - if let Ok(mut result) = result_tx.lock() { - result = Some(val); - } - Ok(()) - }, - )?; - - event_loop.call_method1("run_until_complete", (coro,))?; + + let task_locals = TaskLocals::new(event_loop).copy_context(py)?; + let py_fut = event_loop.call_method0("create_future")?; + let loop_tx = event_loop.into_py(py); + let future_tx = py_fut.into_py(py); + + let rth = rt.handler(); + + rt.spawn(async move { + let val = rth.scope( + task_locals.clone(), + fut + ) + .await; + if let Ok(mut result) = result_tx.lock() { + *result = Some(val.unwrap()); + } + + Python::with_gil(move \|py\| { + let res_method = future_tx.getattr(py, "set_result").unwrap(); + let _ = loop_tx.call_method(py, "call_soon_threadsafe", (res_method, py.None()), None); + }); + }); + + event_loop.call_method1("run_until_complete", (py_fut,))?; let result = result_rx.lock().unwrap().take().unwrap(); Ok(result)	Avoid `pyo3-asyncio` fork Involved steps: - consolidate changes to make a PR into the upstream project - requires #32 - open up a discussion in the upstream project		2023-02-06T17:37:59	[]	[]
rickh94/ODetaM	rickh94__ODetaM-13	bf29d7583a8d1015ca537c5f0c9f76d5a53cff6a	diff --git a/odetam/async_model.py b/odetam/async_model.py index 45d0816..8e2e3ce 100644 --- a/odetam/async_model.py +++ b/odetam/async_model.py @@ -1,8 +1,7 @@ -from typing import Any, Dict, List, Optional, Union +from typing import Any, Dict, List, Optional, Type, TypeVar, Union from deta import AsyncBase, Deta from deta.base import FetchResponse -from typing_extensions import Self from odetam.exceptions import DetaError, InvalidKey, ItemNotFound from odetam.model import BaseDetaModel, DetaModelMetaClass, handle_db_property @@ -19,9 +18,12 @@ def __db__(cls): return handle_db_property(cls, AsyncBase) +T = TypeVar("T", bound="AsyncDetaModel") + + class AsyncDetaModel(BaseDetaModel, metaclass=AsyncDetaModelMetaClass): @classmethod - async def get(cls, key: str) -> Self: + async def get(cls: Type[T], key: str) -> T: """ Get a single instance :param key: Deta database key @@ -36,7 +38,7 @@ async def get(cls, key: str) -> Self: return cls._return_item_or_raise(item) @classmethod - async def get_or_none(cls, key: str) -> Optional[Self]: + async def get_or_none(cls: Type[T], key: str) -> Optional[T]: """Try to get item by key or return None if item not found""" try: return await cls.get(key) @@ -44,7 +46,7 @@ async def get_or_none(cls, key: str) -> Optional[Self]: return None @classmethod - async def get_all(cls) -> List[Self]: + async def get_all(cls: Type[T]) -> List[T]: """Get all the records from the database""" response: FetchResponse = await cls.__db__.fetch() records: List[Dict[str, Any]] = response.items @@ -56,8 +58,8 @@ async def get_all(cls) -> List[Self]: @classmethod async def query( - cls, query_statement: Union[DetaQuery, DetaQueryStatement, DetaQueryList] - ) -> List[Self]: + cls: Type[T], query_statement: Union[DetaQuery, DetaQueryStatement, DetaQueryList] + ) -> List[T]: """Get items from database based on the query.""" response: FetchResponse = await cls.__db__.fetch(query_statement.as_query()) records: List[Dict[str, Any]] = response.items @@ -70,12 +72,12 @@ async def query( return [cls._deserialize(item) for item in records] @classmethod - async def delete_key(cls, key: str): + async def delete_key(cls, key: str) -> None: """Delete an item based on the key""" await cls.__db__.delete(key) @classmethod - async def put_many(cls, items: List[Self]): + async def put_many(cls: Type[T], items: List[T]) -> List[T]: """Put multiple instances at once :param items: List of pydantic objects to put in the database @@ -103,7 +105,7 @@ async def put_many(cls, items: List[Self]): async def _db_put(cls, data: List[Dict[str, Any]]) -> List[Dict[str, Any]]: return await cls.__db__.put(data) - async def save(self): + async def save(self) -> None: """Saves the record to the database. Behaves as upsert, will create if not present. Database key will then be set on the object.""" # exclude = set() @@ -114,7 +116,7 @@ async def save(self): saved = await self._db_put(self._serialize()) self.key = saved["key"] - async def delete(self): + async def delete(self) -> None: """Delete the open object from the database. The object will still exist in python, but will be deleted from the database and the key attribute will be set to None.""" diff --git a/odetam/model.py b/odetam/model.py index b3a548d..24a99bd 100644 --- a/odetam/model.py +++ b/odetam/model.py @@ -1,13 +1,12 @@ import datetime import re -from typing import Any, Callable, Container, Dict, List, Optional, Union +from typing import Any, Callable, Container, Dict, List, Optional, Type, TypeVar, Union import pydantic import ujson from deta import Base, Deta from deta.base import FetchResponse, _Base from pydantic import BaseModel, Field, ValidationError -from typing_extensions import Self from odetam.exceptions import DetaError, InvalidKey, ItemNotFound from odetam.field import DetaField @@ -53,6 +52,9 @@ def __db__(cls): return handle_db_property(cls, Base) +K = TypeVar("K", bound="BaseDetaModel") + + class BaseDetaModel(BaseModel): __db__ = Optional[_Base] @@ -77,7 +79,8 @@ def _serialize(self, exclude: Optional[Container[str]] = None) -> Dict[str, Any] elif field.type_ == datetime.datetime: as_dict[field_name] = getattr(self, field_name).timestamp() elif field.type_ == datetime.date: - as_dict[field_name] = int(getattr(self, field_name).strftime("%Y%m%d")) + as_dict[field_name] = int( + getattr(self, field_name).strftime("%Y%m%d")) elif field.type_ == datetime.time: as_dict[field_name] = int( getattr(self, field_name).strftime("%H%M%S%f") @@ -90,7 +93,7 @@ def _serialize(self, exclude: Optional[Container[str]] = None) -> Dict[str, Any] return as_dict @classmethod - def _deserialize(cls, data: Dict[str, Any]) -> Self: + def _deserialize(cls: Type[K], data: Dict[str, Any]) -> K: as_dict: Dict[str, Any] = {} for field_name, field in cls.__fields__.items(): if field_name not in data: @@ -100,7 +103,8 @@ def _deserialize(cls, data: Dict[str, Any]) -> Self: elif field.type_ in DETA_TYPES: as_dict[field_name] = data[field_name] elif field.type_ == datetime.datetime: - as_dict[field_name] = datetime.datetime.fromtimestamp(data[field_name]) + as_dict[field_name] = datetime.datetime.fromtimestamp( + data[field_name]) elif field.type_ == datetime.date: as_dict[field_name] = datetime.datetime.strptime( str(data[field_name]), "%Y%m%d" @@ -120,7 +124,7 @@ def _deserialize(cls, data: Dict[str, Any]) -> Self: return cls.parse_obj(as_dict) @classmethod - def _return_item_or_raise(cls, item: Optional[Dict[str, Any]]) -> Self: + def _return_item_or_raise(cls: Type[K], item: Optional[Dict[str, Any]]) -> K: if item is None or item.get("key") == "None": raise ItemNotFound("Could not find item matching that key") try: @@ -129,9 +133,12 @@ def _return_item_or_raise(cls, item: Optional[Dict[str, Any]]) -> Self: raise ItemNotFound("Could not find item matching that key") +T = TypeVar("T", bound="DetaModel") + + class DetaModel(BaseDetaModel, metaclass=DetaModelMetaClass): @classmethod - def get(cls, key: str) -> Self: + def get(cls: Type[T], key: str) -> T: """ Get a single instance :param key: Deta database key @@ -146,7 +153,7 @@ def get(cls, key: str) -> Self: return cls._return_item_or_raise(item) @classmethod - def get_or_none(cls, key: str) -> Optional[Self]: + def get_or_none(cls: Type[T], key: str) -> Optional[T]: """Try to get item by key or return None if item not found""" try: return cls.get(key) @@ -154,7 +161,7 @@ def get_or_none(cls, key: str) -> Optional[Self]: return None @classmethod - def get_all(cls) -> List[Self]: + def get_all(cls: Type[T]) -> List[T]: """Get all the records from the database""" response: FetchResponse = cls.__db__.fetch() records: List[Dict[str, Any]] = response.items @@ -166,24 +173,25 @@ def get_all(cls) -> List[Self]: @classmethod def query( - cls, query_statement: Union[DetaQuery, DetaQueryStatement, DetaQueryList] - ) -> List[Self]: + cls: Type[T], query_statement: Union[DetaQuery, DetaQueryStatement, DetaQueryList] + ) -> List[T]: """Get items from database based on the query.""" response: FetchResponse = cls.__db__.fetch(query_statement.as_query()) records: List[Dict[str, Any]] = response.items while response.last: - response = cls.__db__.fetch(query_statement.as_query(), last=response.last) + response = cls.__db__.fetch( + query_statement.as_query(), last=response.last) records += response.items return [cls._deserialize(item) for item in records] @classmethod - def delete_key(cls, key: str): + def delete_key(cls, key: str) -> None: """Delete an item based on the key""" cls.__db__.delete(key) @classmethod - def put_many(cls, items: List[Self]) -> List[Self]: + def put_many(cls: Type[T], items: List[T]) -> List[T]: """Put multiple instances at once :param items: List of pydantic objects to put in the database @@ -211,7 +219,7 @@ def put_many(cls, items: List[Self]) -> List[Self]: def _db_put(cls, data: List[Dict[str, Any]]) -> List[Dict[str, Any]]: return cls.__db__.put(data) # type: ignore - def save(self): + def save(self) -> None: """Saves the record to the database. Behaves as upsert, will create if not present. Database key will then be set on the object.""" # exclude = set() @@ -222,7 +230,7 @@ def save(self): saved = self._db_put(self._serialize()) self.key = saved["key"] - def delete(self): + def delete(self) -> None: """Delete the open object from the database. The object will still exist in python, but will be deleted from the database and the key attribute will be set to None."""	Async Support An (unofficial) [asynchronous library](https://github.com/leits/aiodeta) for the Deta api has become available. I'd like to support it. Without breaking compatibility, it could be something like `from odetam.async import DetaModel` and then provide async/await versions of the existing api.		2023-05-28T04:38:24	[]	[]
amjith/fuzzyfinder	amjith__fuzzyfinder-19	43fe7676cad68e269bbace7bb2fd9b77f2e07da9	diff --git a/fuzzyfinder/main.py b/fuzzyfinder/main.py index fa5c14b..ab9c4f4 100755 --- a/fuzzyfinder/main.py +++ b/fuzzyfinder/main.py @@ -3,7 +3,9 @@ from . import export @export -def fuzzyfinder(input, collection, accessor=lambda x: x, sort_results=True): +def fuzzyfinder( + input, collection, accessor=lambda x: x, sort_results=True, ignore_case=True +): """ Args: input (str): A partial string which is typically entered by a user. @@ -12,13 +14,15 @@ def fuzzyfinder(input, collection, accessor=lambda x: x, sort_results=True): accessor (function): If the `collection` is not an iterable of strings, then use the accessor to fetch the string that will be used for fuzzy matching. - sort_results(bool): The suggestions are sorted by considering the - smallest contiguous match, followed by where the - match is found in the full string. If two suggestions - have the same rank, they are then sorted - alpha-numerically. This parameter controls the - last tie-breaker-alpha-numeric sorting. The sorting - based on match length and position will be intact. + sort_results (bool): The suggestions are sorted by considering the + smallest contiguous match, followed by where the + match is found in the full string. If two suggestions + have the same rank, they are then sorted + alpha-numerically. This parameter controls the + last tie-breaker-alpha-numeric sorting. The sorting + based on match length and position will be intact. + ignore_case (bool): If this parameter is set to False, the filtering + is case-sensitive. Returns: suggestions (generator): A generator object that produces a list of @@ -28,7 +32,7 @@ def fuzzyfinder(input, collection, accessor=lambda x: x, sort_results=True): input = str(input) if not isinstance(input, str) else input pat = '.*?'.join(map(re.escape, input)) pat = '(?=({0}))'.format(pat) # lookahead regex to manage overlapping matches - regex = re.compile(pat, re.IGNORECASE) + regex = re.compile(pat, re.IGNORECASE if ignore_case else 0) for item in collection: r = list(regex.finditer(accessor(item))) if r:	case_insensitive as optional argument? I have some old code that reads: ```python from fuzzyfinder import fuzzyfinder matches = fuzzyfinder(word_before_cursor, fuzzy_words, case_sensitive=True) ``` But I can't see in this project when case_sensitive was ever an option? Am I going crazy? Was case_sensitive the default and the extra parameter I was passing just doing nothing?	I think I just changed my local version :-/ Ah well! I'll leave this here because I think it makes sense and may even create a pull request.	2024-08-28T11:28:46	[]	[]
lareferencia/dspace-stats-collector	lareferencia__dspace-stats-collector-19	85b02cf6609496d8926739a2980931782067e57b	diff --git a/dspace_stats_collector/collector.py b/dspace_stats_collector/collector.py index 9af9646..ddffa56 100644 --- a/dspace_stats_collector/collector.py +++ b/dspace_stats_collector/collector.py @@ -88,6 +88,9 @@ def run(): loglevel = logging.DEBUG else: loglevel = logging.INFO + + if args.debug: + loglevel = logging.DEBUG logDirName = os.path.expanduser('~') + "/dspace-stats-collector/var/logs" if not os.path.exists(logDirName): @@ -98,10 +101,16 @@ def run(): else: logFileName = "dspace-stats-collector.log" + logging_handlers = [ logging.FileHandler("{}/{}".format(logDirName, logFileName), 'w+') ] + + # if we are in verbose mode, also log to console + if args.verbose: + logging_handlers.append(logging.StreamHandler()) + + logging.basicConfig(level=loglevel, - format="%(levelname)s: %(message)s", - filename="{}/{}".format(logDirName, logFileName), - filemode='a') + format="%(asctime)s %(levelname)s: %(message)s", + handlers=logging_handlers) logger.debug("Verbose: %s" % args.verbose) logger.debug("Repository: %s" % args.repository) @@ -173,6 +182,11 @@ def valid_date_type(arg_date_str): help="increase output verbosity", default=False, action="store_true") + parser.add_argument("--debug", + help="debug mode", + default=False, + action="store_true") + parser.add_argument("-a", "--archived_core", metavar="YYYY", diff --git a/dspace_stats_collector/configcontext.py b/dspace_stats_collector/configcontext.py index 59f86cb..3332e1b 100644 --- a/dspace_stats_collector/configcontext.py +++ b/dspace_stats_collector/configcontext.py @@ -17,18 +17,25 @@ from .dspacedb4 import DSpaceDB4 from .dspacedb5 import DSpaceDB5 from .dspacedb6 import DSpaceDB6 + from .dspacedb7 import DSpaceDB7 + from .dspacedb5cris import DSpaceDB5Cris + from .dspacedb5oracle import DSpaceDB5Oracle + from .dspacedb6oracle import DSpaceDB6Oracle except Exception: #ImportError from dspacedb4 import DSpaceDB4 from dspacedb5 import DSpaceDB5 from dspacedb6 import DSpaceDB6 - + from dspacedb7 import DSpaceDB7 + from dspacedb5cris import DSpaceDB5Cris + from dspacedb5oracle import DSpaceDB5Oracle + from dspacedb6oracle import DSpaceDB6Oracle SAVE_DIR = os.path.expanduser('~') + "/dspace-stats-collector/var/timestamp" DEFAULT_INSTALL_PATH = os.path.expanduser('~') + "/dspace-stats-collector" DEFAULT_COLLECTOR_COMMAND_NAME="dspace-stats-collector" DEFAULT_CONFIG_PATH = DEFAULT_INSTALL_PATH + "/config" -SOLR_STATS_CORE_NAME = "statistics" +DEFAULT_SOLR_STATS_CORE_NAME = "statistics" TIMESTAMP_PATTERN = "%Y-%m-%dT00:00:00.000Z" SOLR_QUERY_ROWS_SIZE = 10 DEFAULT_OUPUT_LIMIT = 100 @@ -94,9 +101,9 @@ def __init__(self, repoName, commandLineArgs): # Solr Context if commandLineArgs.archived_core != None: - self.solrStatsCoreName = SOLR_STATS_CORE_NAME + "-" + commandLineArgs.archived_core + self.solrStatsCoreName = self.getSolrStatsCoreName() + "-" + commandLineArgs.archived_core else: - self.solrStatsCoreName = SOLR_STATS_CORE_NAME + self.solrStatsCoreName = self.getSolrStatsCoreName() self.solrServerURL = self._find_solr_server() self.solrStatsCoreURL = self.solrServerURL + "/" + self.solrStatsCoreName @@ -135,6 +142,14 @@ def __init__(self, repoName, commandLineArgs): self.db = DSpaceDB5(self.dspaceProperties['db.url'],self.dspaceProperties['db.username'],self.dspaceProperties['db.password']) elif self.dspaceMajorVersion == '6': self.db = DSpaceDB6(self.dspaceProperties['db.url'],self.dspaceProperties['db.username'],self.dspaceProperties['db.password']) + elif self.dspaceMajorVersion == '7': + self.db = DSpaceDB7(self.dspaceProperties['db.url'],self.dspaceProperties['db.username'],self.dspaceProperties['db.password']) + elif self.dspaceMajorVersion == '5o': + self.db = DSpaceDB5Oracle(self.dspaceProperties['db.url'],self.dspaceProperties['db.username'],self.dspaceProperties['db.password']) + elif self.dspaceMajorVersion == '6o': + self.db = DSpaceDB6Oracle(self.dspaceProperties['db.url'],self.dspaceProperties['db.username'],self.dspaceProperties['db.password']) + elif self.dspaceMajorVersion == '5c': + self.db = DSpaceDB5Cris(self.dspaceProperties['db.url'],self.dspaceProperties['db.username'],self.dspaceProperties['db.password']) else: logger.error('Only implemented values for dspace.majorVersion are 4, 5 and 6. Received {}'.format(self.dspaceMajorVersion)) raise NotImplementedError @@ -163,7 +178,10 @@ def getSolrLimit(self): return int(self.properties['solr.limit']) def getDspaceMajorVersion(self): - return int(self.properties['dspace.majorVersion']) + return str(self.properties['dspace.majorVersion']) + + def getSolrStatsCoreName(self): + return str( self.properties.get('solr.core', DEFAULT_SOLR_STATS_CORE_NAME) ) ################################################ private methods ########################################## def _read_properties(self): @@ -183,7 +201,9 @@ def _read_properties(self): def _read_dspace_properties(self): javaprops = JavaProperties() - if self.getDspaceMajorVersion() == 6: + if self.getDspaceMajorVersion() == '6': + + ## try to read dspace.cfg try: propertiesFilename = "%s/config/dspace.cfg" % (self.properties["dspace.dir"]) javaprops.load(open(propertiesFilename)) @@ -193,6 +213,7 @@ def _read_dspace_properties(self): logger.exception("Error while trying to read properties file %s" % propertiesFilename) raise + ## try to read local.cfg try: propertiesFilename = "%s/config/local.cfg" % (self.properties["dspace.dir"]) javaprops.load(open(propertiesFilename)) @@ -202,6 +223,16 @@ def _read_dspace_properties(self): logger.debug("Could not read property file %s" % propertiesFilename) pass + elif self.getDspaceMajorVersion() == '5c': + try: + propertiesFilename = "%s/build.properties" % (self.properties["dspace.dir"]) + javaprops.load(open(propertiesFilename)) + property_dict = javaprops.get_property_dict() + logger.debug("Read succesfully property file %s" % propertiesFilename) + except (FileNotFoundError, UnboundLocalError): + logger.exception("Error while trying to read properties file %s" % propertiesFilename) + raise + else: try: propertiesFilename = "%s/config/dspace.cfg" % (self.properties["dspace.dir"]) diff --git a/dspace_stats_collector/dspacedb.py b/dspace_stats_collector/dspacedb.py index 4faf689..500deac 100644 --- a/dspace_stats_collector/dspacedb.py +++ b/dspace_stats_collector/dspacedb.py @@ -17,26 +17,36 @@ def __init__(self, jdbcUrl, username, password): # Parse jdbc url # Postgres template: jdbc:postgresql://localhost:5432/dspace # Oracle template: jdbc:oracle:thin:@//localhost:1521/xe - m = re.match("^jdbc:(postgresql\|oracle):[^\/]\/\/([^:]+):(\d+)/(.)$", jdbcUrl) + # Oracle template: jdbc:oracle:thin:@localhost:1521:xe + m = re.match("^jdbc:(postgresql\|oracle):[^\/\|^@][@\/\/\|\/\/\|@]([^:]+):(\d+)(\/\|:)(.)$", jdbcUrl) + if m is None: logger.error("Could not parse db.url string: %s" % jdbcUrl) raise ValueError - (engine, hostname, port, database) = m.group(1, 2, 3, 4) - - if engine != "postgresql": - logger.error("DB Engine not yet supported: %s" % engine) - raise NotImplementedError - - self.connString = '{engine}://{username}:{password}@{hostname}:{port}/{database}' - self.connString = self.connString.format( + # separate host, port, dbname, and dbschema and separator + (engine, hostname, port, separator, database) = m.group(1, 2, 3, 4, 5) + + # Create connection string depending on engine + if engine == 'postgresql': + connStringTemplate = 'postgresql://{username}:{password}@{hostname}:{port}' + separator + '{database}' + if engine == 'oracle': + if separator == ':': + connStringTemplate = 'oracle+cx_oracle://{username}:{password}@{hostname}:{port}/?service_name={database}' + elif separator == '/': + connStringTemplate = 'oracle+cx_oracle://{username}:{password}@{hostname}:{port}/{database}' + + + # Create connection string based on engine, hostname, port, database. Use separator to determine the kind of oracle connection + self.connString = connStringTemplate.format( engine=engine, username=username, password=password, hostname=hostname, port=port, database=database, - ) + ) + logger.debug('DB Connection String: ' + self.connString) try: self.conn = sqlalchemy.create_engine(self.connString).connect() diff --git a/dspace_stats_collector/dspacedb5cris.py b/dspace_stats_collector/dspacedb5cris.py new file mode 100644 index 0000000..bbc509f --- /dev/null +++ b/dspace_stats_collector/dspacedb5cris.py @@ -0,0 +1,71 @@ +#!/usr/bin/env python +# -- coding: utf-8 -- +""" Dspace DB components """ + +import logging +logger = logging.getLogger() + +try: + from .dspacedb import DSpaceDB +except Exception: #ImportError + from dspacedb import DSpaceDB + +class DSpaceDB5Cris(DSpaceDB): + + def __init__(self, jdbcUrl, username, password): + + DSpaceDB.__init__(self,jdbcUrl, username, password) + + self._queryDownloadSQL = """ + SELECT mv.resource_id AS id, + mv2.text_value AS record_title, + h.handle AS handle, + true AS is_download, + i.item_id AS owning_item, + b.sequence_id AS sequence_id, + mv.text_value AS filename + FROM metadatavalue AS mv + RIGHT JOIN bitstream AS b ON mv.resource_id = b.bitstream_id + RIGHT JOIN bundle2bitstream AS bb ON b.bitstream_id = bb.bitstream_id + RIGHT JOIN item2bundle AS i ON i.bundle_id = bb.bundle_id + RIGHT JOIN handle AS h ON h.resource_id = i.item_id + RIGHT JOIN metadatavalue AS mv2 ON mv2.resource_id = i.item_id + WHERE mv.metadata_field_id = {dcTitleId} + AND mv.resource_type_id = 0 + AND b.sequence_id IS NOT NULL + AND b.deleted = FALSE + AND mv2.metadata_field_id = {dcTitleId} + AND mv2.resource_type_id=2 + AND mv.resource_id = {bitstreamId}; + """ + + self._queryItemSQL = """ + SELECT mv.resource_id AS id, + mv.text_value AS record_title, + h.handle AS handle, + false AS is_download, + NULL AS owning_item, + NULL AS sequence_id, + NULL AS filename + FROM metadatavalue AS mv + RIGHT JOIN handle AS h ON h.resource_id = mv.resource_id + WHERE metadata_field_id = {dcTitleId} + AND mv.resource_type_id=2 + AND h.resource_type_id=2 + AND mv.resource_id = {itemId}; + """ + + self._queryTitleSQL = """ + SELECT metadata_field_id AS "dcTitleId" + FROM metadatafieldregistry mfr, + metadataschemaregistry msr + WHERE mfr.metadata_schema_id = msr.metadata_schema_id + AND short_id = 'dc' + AND element = 'title' + AND qualifier IS NULL; + """ + + self._dcTitleId = self.getDcTitleId() + + + diff --git a/dspace_stats_collector/dspacedb5oracle.py b/dspace_stats_collector/dspacedb5oracle.py new file mode 100644 index 0000000..abc602d --- /dev/null +++ b/dspace_stats_collector/dspacedb5oracle.py @@ -0,0 +1,71 @@ +#!/usr/bin/env python +# -- coding: utf-8 -- +""" Dspace DB components """ + +import logging +logger = logging.getLogger() + +try: + from .dspacedb import DSpaceDB +except Exception: #ImportError + from dspacedb import DSpaceDB + +class DSpaceDB5Oracle(DSpaceDB): + + def __init__(self, jdbcUrl, username, password): + + DSpaceDB.__init__(self,jdbcUrl, username, password) + + self._queryDownloadSQL = """ + SELECT mv.resource_id AS id, + mv2.text_value AS record_title, + h.handle AS handle, + true AS is_download, + i.item_id AS owning_item, + b.sequence_id AS sequence_id, + mv.text_value AS filename + FROM metadatavalue AS mv + RIGHT JOIN bitstream AS b ON mv.resource_id = b.bitstream_id + RIGHT JOIN bundle2bitstream AS bb ON b.bitstream_id = bb.bitstream_id + RIGHT JOIN item2bundle AS i ON i.bundle_id = bb.bundle_id + RIGHT JOIN handle AS h ON h.resource_id = i.item_id + RIGHT JOIN metadatavalue AS mv2 ON mv2.resource_id = i.item_id + WHERE mv.metadata_field_id = {dcTitleId} + AND mv.resource_type_id = 0 + AND b.sequence_id IS NOT NULL + AND b.deleted = FALSE + AND mv2.metadata_field_id = {dcTitleId} + AND mv2.resource_type_id=2 + AND mv.resource_id = {bitstreamId}; + """ + + self._queryItemSQL = """ + SELECT mv.resource_id AS id, + mv.text_value AS record_title, + h.handle AS handle, + false AS is_download, + NULL AS owning_item, + NULL AS sequence_id, + NULL AS filename + FROM metadatavalue AS mv + RIGHT JOIN handle AS h ON h.resource_id = mv.resource_id + WHERE metadata_field_id = {dcTitleId} + AND mv.resource_type_id=2 + AND h.resource_type_id=2 + AND mv.resource_id = {itemId}; + """ + + self._queryTitleSQL = """ + SELECT metadata_field_id AS "dcTitleId" + FROM metadatafieldregistry mfr, + metadataschemaregistry msr + WHERE mfr.metadata_schema_id = msr.metadata_schema_id + AND short_id = 'dc' + AND element = 'title' + AND qualifier IS NULL; + """ + + self._dcTitleId = self.getDcTitleId() + + + diff --git a/dspace_stats_collector/dspacedb6.py b/dspace_stats_collector/dspacedb6.py index 240d9b1..8ec0b94 100644 --- a/dspace_stats_collector/dspacedb6.py +++ b/dspace_stats_collector/dspacedb6.py @@ -34,7 +34,7 @@ def __init__(self, jdbcUrl, username, password): AND b.sequence_id IS NOT NULL AND b.deleted = FALSE AND mv2.metadata_field_id = {dcTitleId} - AND mv.dspace_object_id::text = '{bitstreamId}'; + AND mv.dspace_object_id = uuid('{bitstreamId}'); """ self._queryItemSQL = """ @@ -49,7 +49,7 @@ def __init__(self, jdbcUrl, username, password): RIGHT JOIN handle AS h ON h.resource_id = mv.dspace_object_id WHERE metadata_field_id = {dcTitleId} AND h.resource_type_id=2 - AND mv.dspace_object_id::text = '{itemId}'; + AND mv.dspace_object_id = uuid('{itemId}'); """ self._queryTitleSQL = """ diff --git a/dspace_stats_collector/dspacedb6oracle.py b/dspace_stats_collector/dspacedb6oracle.py new file mode 100644 index 0000000..7b3098b --- /dev/null +++ b/dspace_stats_collector/dspacedb6oracle.py @@ -0,0 +1,69 @@ +#!/usr/bin/env python +# -- coding: utf-8 -- +""" Dspace DB components """ + +import logging +logger = logging.getLogger() + +try: + from .dspacedb import DSpaceDB +except Exception: #ImportError + from dspacedb import DSpaceDB + +class DSpaceDB6Oracle(DSpaceDB): + + def __init__(self, jdbcUrl, username, password): + + DSpaceDB.__init__(self,jdbcUrl, username, password) + + self._queryDownloadSQL = """ + SELECT mv.dspace_object_id::text AS id, + mv2.text_value AS record_title, + h.handle AS handle, + true AS is_download, + i.item_id::text AS owning_item, + b.sequence_id AS sequence_id, + mv.text_value AS filename + FROM metadatavalue AS mv + RIGHT JOIN bitstream AS b ON mv.dspace_object_id = b.uuid + RIGHT JOIN bundle2bitstream AS bb ON b.uuid = bb.bitstream_id + RIGHT JOIN item2bundle AS i ON i.bundle_id = bb.bundle_id + RIGHT JOIN handle AS h ON h.resource_id = i.item_id + RIGHT JOIN metadatavalue AS mv2 ON mv2.dspace_object_id = i.item_id + WHERE mv.metadata_field_id = {dcTitleId} + AND b.sequence_id IS NOT NULL + AND b.deleted = FALSE + AND mv2.metadata_field_id = {dcTitleId} + AND mv.dspace_object_id = uuid('{bitstreamId}'); + """ + + self._queryItemSQL = """ + SELECT mv.dspace_object_id::text AS id, + mv.text_value AS record_title, + h.handle AS handle, + false AS is_download, + NULL AS owning_item, + NULL AS sequence_id, + NULL AS filename + FROM metadatavalue AS mv + RIGHT JOIN handle AS h ON h.resource_id = mv.dspace_object_id + WHERE metadata_field_id = {dcTitleId} + AND h.resource_type_id=2 + AND mv.dspace_object_id = uuid('{itemId}'); + """ + + self._queryTitleSQL = """ + SELECT metadata_field_id AS "dcTitleId" + FROM metadatafieldregistry mfr, + metadataschemaregistry msr + WHERE mfr.metadata_schema_id = msr.metadata_schema_id + AND short_id = 'dc' + AND element = 'title' + AND qualifier IS NULL; + """ + + self._dcTitleId = self.getDcTitleId() + + + + \ No newline at end of file diff --git a/dspace_stats_collector/dspacedb7.py b/dspace_stats_collector/dspacedb7.py new file mode 100644 index 0000000..98c1a3e --- /dev/null +++ b/dspace_stats_collector/dspacedb7.py @@ -0,0 +1,69 @@ +#!/usr/bin/env python +# -- coding: utf-8 -*- +""" Dspace DB components """ + +import logging +logger = logging.getLogger() + +try: + from .dspacedb import DSpaceDB +except Exception: #ImportError + from dspacedb import DSpaceDB + +class DSpaceDB7(DSpaceDB): + + def __init__(self, jdbcUrl, username, password): + + DSpaceDB.__init__(self,jdbcUrl, username, password) + + self._queryDownloadSQL = """ + SELECT mv.dspace_object_id::text AS id, + mv2.text_value AS record_title, + h.handle AS handle, + true AS is_download, + i.item_id::text AS owning_item, + b.sequence_id AS sequence_id, + mv.text_value AS filename + FROM metadatavalue AS mv + RIGHT JOIN bitstream AS b ON mv.dspace_object_id = b.uuid + RIGHT JOIN bundle2bitstream AS bb ON b.uuid = bb.bitstream_id + RIGHT JOIN item2bundle AS i ON i.bundle_id = bb.bundle_id + RIGHT JOIN handle AS h ON h.resource_id = i.item_id + RIGHT JOIN metadatavalue AS mv2 ON mv2.dspace_object_id = i.item_id + WHERE mv.metadata_field_id = {dcTitleId} + AND b.sequence_id IS NOT NULL + AND b.deleted = FALSE + AND mv2.metadata_field_id = {dcTitleId} + AND mv.dspace_object_id = uuid('{bitstreamId}'); + """ + + self._queryItemSQL = """ + SELECT mv.dspace_object_id::text AS id, + mv.text_value AS record_title, + h.handle AS handle, + false AS is_download, + NULL AS owning_item, + NULL AS sequence_id, + NULL AS filename + FROM metadatavalue AS mv + RIGHT JOIN handle AS h ON h.resource_id = mv.dspace_object_id + WHERE metadata_field_id = {dcTitleId} + AND h.resource_type_id=2 + AND mv.dspace_object_id = uuid('{itemId}'); + """ + + self._queryTitleSQL = """ + SELECT metadata_field_id AS "dcTitleId" + FROM metadatafieldregistry mfr, + metadataschemaregistry msr + WHERE mfr.metadata_schema_id = msr.metadata_schema_id + AND short_id = 'dc' + AND element = 'title' + AND qualifier IS NULL; + """ + + self._dcTitleId = self.getDcTitleId() + + + + \ No newline at end of file diff --git a/dspace_stats_collector/matomooutput.py b/dspace_stats_collector/matomooutput.py index 09bafc4..6cff898 100644 --- a/dspace_stats_collector/matomooutput.py +++ b/dspace_stats_collector/matomooutput.py @@ -37,8 +37,13 @@ def __init__(self, configContext): self._handleCanonicalPrefix = dspaceProperties['handle.canonical.prefix'] else: self._handleCanonicalPrefix = 'http://hdl.handle.net/' - self._dspaceHostname = dspaceProperties['dspace.hostname'] - self._dspaceUrl = dspaceProperties['dspace.url'] + + if configContext.getDspaceMajorVersion() == '7': + self._dspaceHostname = dspaceProperties['dspace.server.url'] + self._dspaceUrl = dspaceProperties['dspace.ui.url'] + else: + self._dspaceHostname = dspaceProperties['dspace.hostname'] + self._dspaceUrl = dspaceProperties['dspace.url'] self._repoProperties = configContext.properties @@ -122,6 +127,8 @@ def getTotalSent(self): def send(self, event): self._buffer.append((event._matomoRequest, event.is_robot, event._src['time'])) + logger.debug( "Event json dump: {}".format( event.toJSON() ) ) + #print(self._buffer) if self.isBufferFull(): logger.debug("Buffer is full") @@ -193,7 +200,9 @@ def flush(self): raise except MatomoInternalServerException as e: # if there is some internal error will discard this event and log the result - logger.error('Matomo internal error occurred: {} with event. This event will be discarded.\n'.format( str(e)) ) + logger.error('Matomo internal error occurred: {} with event. This event will be discarded.\n'.format( str(e)) ) + logger.debug('Request URL was: {}'.format(self._url) ) + logger.debug('Request Payload was: {}'.format(m) ) diff --git a/install-standalone-develop.sh b/install-standalone-develop.sh new file mode 100644 index 0000000..a49b8d2 --- /dev/null +++ b/install-standalone-develop.sh @@ -0,0 +1,56 @@ +#!/bin/bash +# + +INSTALL_PATH=$HOME/dspace-stats-collector +MINICONDA_URL_PREFIX='https://repo.anaconda.com/miniconda/' + +rm -rf $INSTALL_PATH +rm -f delete_this_file.sh + +MACHINE_TYPE=`uname -m` +if [ ${MACHINE_TYPE} == 'x86_64' ]; then + # 64-bit stuff here + MINICONDA_FILE='Miniconda3-py37_4.8.3-Linux-x86_64.sh' +elif [ ${MACHINE_TYPE} == 'x86' ]; then + # 32-bit stuff here + MINICONDA_FILE='Miniconda3-py37_4.8.3-Linux-x86.sh' +elif [ ${MACHINE_TYPE} == 'aarch64' ]; then + # ARM stuff here (EXPERIMENTAL) + MINICONDA_FILE='Miniconda3-py37_4.9.2-Linux-aarch64.sh' +else + echo "Unknown machine type: ${MACHINE_TYPE}" + exit 1 +fi + +MINICONDA_URL=$MINICONDA_URL_PREFIX$MINICONDA_FILE + +echo $MINICONDA_URL + +if [ -x "$(which curl)" ]; then + curl $MINICONDA_URL -o delete_this_file.sh +else + echo "Could not find curl, please install curl." >&2 + exit 1 +fi + +bash delete_this_file.sh -b -f -p $INSTALL_PATH +rm delete_this_file.sh + +cd $INSTALL_PATH + +echo "Installing dspace-stats-collector package dependencies" +curl https://raw.githubusercontent.com/lareferencia/dspace-stats-collector/develop/requirements-p37-dev.txt -o $INSTALL_PATH/requirements.txt +$INSTALL_PATH/bin/pip install -r $INSTALL_PATH/requirements.txt + +$INSTALL_PATH/bin/conda install -y git + +$INSTALL_PATH/bin/git clone --branch=develop https://github.com/lareferencia/dspace-stats-collector.git + +cd $INSTALL_PATH/dspace-stats-collector +$INSTALL_PATH/bin/python setup.py install + +echo "Installing config files" +$INSTALL_PATH/bin/dspace-stats-configure + + + diff --git a/requirements-p37-dev.txt b/requirements-p37-dev.txt new file mode 100644 index 0000000..09887d4 --- /dev/null +++ b/requirements-p37-dev.txt @@ -0,0 +1,25 @@ +bumpversion==0.5.3 +wheel==0.32.1 +watchdog==0.9.0 +flake8==3.5.0 +tox==3.5.2 +coverage==4.5.1 +Sphinx==1.8.1 +twine==1.12.1 + +pytest==3.8.2 +pytest-runner==4.2 + +requests==2.21.0 +pyjavaprops==1.0.2 +SQLAlchemy==1.2.15 +psycopg2-binary==2.8 +cx-Oracle +pysolr==3.8.1 +pandas==1.2 +urllib3==1.24.2 +pytz==2018.7 +python-crontab +anonymizeip +pid + diff --git a/update-standalone-develop.sh b/update-standalone-develop.sh new file mode 100644 index 0000000..839f490 --- /dev/null +++ b/update-standalone-develop.sh @@ -0,0 +1,19 @@ +#!/bin/bash +# + +INSTALL_PATH=$HOME/dspace-stats-collector + +cd $INSTALL_PATH + +echo "Installing dspace-stats-collector package dependencies" +curl https://raw.githubusercontent.com/lareferencia/dspace-stats-collector/develop/requirements-p37.txt -o $INSTALL_PATH/requirements.txt +$INSTALL_PATH/bin/pip install -r $INSTALL_PATH/requirements.txt + +cd $INSTALL_PATH/dspace-stats-collector + +$INSTALL_PATH/bin/git pull +$INSTALL_PATH/bin/python setup.py install + + + +	Cambio consulta SQL de DSpace6 y DSpace7 para comparar por uuid y no texto y mejorar rendimiento Modificación de las consultas SQL para la versión 6 i 7 de DSpace con postgres para que utilize el indice para comprobar la condición dspace_object_id sea igual al uuid buscado. Mejora mucho el rendimiento, una consulta puede pasar de 16 segundos a mili-segundos. Probado para la versión 6 pero también se debería aplicar a la 7 ya que esta parte es común.		2022-09-08T16:27:37	[]	[]
scikit-hep/hepstats	scikit-hep__hepstats-103	53256cc5ef62e1d781627a9b0df24f136f2a4141	diff --git a/.gitignore b/.gitignore index 785f8ec8..4669666d 100644 --- a/.gitignore +++ b/.gitignore @@ -13,3 +13,5 @@ dist/ /pip-wheel-metadata /docs/_build/* /docs/source/* +/_build/ +/.idea/ diff --git a/CHANGELOG.rst b/CHANGELOG.rst index ecf9d4d5..5e20ba03 100644 --- a/CHANGELOG.rst +++ b/CHANGELOG.rst @@ -4,6 +4,8 @@ Changelog master ****** +* fix toy generation with constraints + Version 0.6.0 ************* diff --git a/src/hepstats/modeling/bayesian_blocks.py b/src/hepstats/modeling/bayesian_blocks.py index 5281e9c8..2a01013d 100644 --- a/src/hepstats/modeling/bayesian_blocks.py +++ b/src/hepstats/modeling/bayesian_blocks.py @@ -14,9 +14,9 @@ from __future__ import annotations from collections.abc import Iterable + import numpy as np import pandas as pd -from typing import Optional, Union class Prior: diff --git a/src/hepstats/splot/sweights.py b/src/hepstats/splot/sweights.py index ac927f6e..65cdb10d 100644 --- a/src/hepstats/splot/sweights.py +++ b/src/hepstats/splot/sweights.py @@ -1,13 +1,14 @@ from __future__ import annotations -import numpy as np -from typing import Any import warnings +from typing import Any + +import numpy as np -from ..utils import eval_pdf -from ..utils.fit.api_check import is_valid_pdf from .exceptions import ModelNotFittedToData from .warnings import AboveToleranceWarning +from ..utils import eval_pdf +from ..utils.fit.api_check import is_valid_pdf def is_sum_of_extended_pdfs(model) -> bool: diff --git a/src/hepstats/utils/fit/__init__.py b/src/hepstats/utils/fit/__init__.py index d67fc4aa..e7a31b74 100644 --- a/src/hepstats/utils/fit/__init__.py +++ b/src/hepstats/utils/fit/__init__.py @@ -1,2 +1,2 @@ -from .sampling import base_sampler, base_sample from .diverse import get_value, eval_pdf, pll, array2dataset, get_nevents, set_values +from .sampling import base_sampler, base_sample diff --git a/src/hepstats/utils/fit/diverse.py b/src/hepstats/utils/fit/diverse.py index 7dc5ffc2..891eeeb7 100644 --- a/src/hepstats/utils/fit/diverse.py +++ b/src/hepstats/utils/fit/diverse.py @@ -1,7 +1,6 @@ from contextlib import ExitStack, contextmanager -import numpy as np -from .api_check import is_valid_pdf +import numpy as np def get_ndims(dataset):	fix of iteration over the param_dict fix of iteration over the param_dict	a fix to this issue https://github.com/scikit-hep/hepstats/issues/101 a fix to this issue https://github.com/scikit-hep/hepstats/issues/101	2023-04-17T15:57:42	[]	[]
scikit-hep/hepstats	scikit-hep__hepstats-66	c70c31fdd20b330961c2a304092c0796bcb7bca5	diff --git a/.github/workflows/auto-cancellation.yml b/.github/workflows/auto-cancellation.yml deleted file mode 100644 index 8c33f6c7..00000000 --- a/.github/workflows/auto-cancellation.yml +++ /dev/null @@ -1,9 +0,0 @@ -name: auto cancellation running job -on: pull_request - -jobs: - cancel: - name: auto-cancellation-running-action - runs-on: ubuntu-latest - steps: - - uses: fauguste/[email protected] diff --git a/.github/workflows/main.yml b/.github/workflows/main.yml index a7ff0069..ea6afc49 100644 --- a/.github/workflows/main.yml +++ b/.github/workflows/main.yml @@ -6,17 +6,27 @@ on: branches: [ "master"] release: types: - - "published" + - "published" jobs: + cancel: + name: 'Cancel Previous Runs' + runs-on: ubuntu-latest + timeout-minutes: 3 + steps: + - uses: styfle/[email protected] + with: + all_but_latest: true + access_token: ${{ github.token }} + pre-commit: name: Format runs-on: ubuntu-latest steps: - - uses: actions/checkout@v3 - with: - fetch-depth: 0 - - uses: actions/setup-python@v4 + - uses: actions/checkout@v3 + with: + fetch-depth: 0 + - uses: actions/setup-python@v4 checks: runs-on: ${{ matrix.os }} @@ -24,12 +34,12 @@ jobs: fail-fast: false matrix: os: - - ubuntu-latest + - ubuntu-latest python-version: - - "3.7" - - "3.8" - - "3.9" - - "3.10" + - "3.7" + - "3.8" + - "3.9" + - "3.10" include: - os: windows-latest python-version: "3.7" @@ -41,87 +51,87 @@ jobs: python-version: "3.10" name: Check Python ${{ matrix.python-version }} ${{ matrix.os }} steps: - - uses: actions/checkout@v3 - with: - fetch-depth: 0 - - - name: Setup Python ${{ matrix.python-version }} - uses: actions/setup-python@v4 - with: - python-version: ${{ matrix.python-version }} - - - name: Install package - run: python -m pip install -e .[test] - - - name: Test package - run: python -m pytest --doctest-modules --cov=hepstats --cov-report=xml - - - name: Upload coverage to Codecov - if: matrix.python-version == '3.10' && matrix.os == 'ubuntu-latest' - uses: codecov/codecov-action@v3 - with: - file: ./coverage.xml - flags: unittests - name: codecov-umbrella - fail_ci_if_error: true + - uses: actions/checkout@v3 + with: + fetch-depth: 0 + + - name: Setup Python ${{ matrix.python-version }} + uses: actions/setup-python@v4 + with: + python-version: ${{ matrix.python-version }} + + - name: Install package + run: python -m pip install -e .[test] + + - name: Test package + run: python -m pytest --doctest-modules --cov=hepstats --cov-report=xml + + - name: Upload coverage to Codecov + if: matrix.python-version == '3.10' && matrix.os == 'ubuntu-latest' + uses: codecov/codecov-action@v3 + with: + file: ./coverage.xml + flags: unittests + name: codecov-umbrella + fail_ci_if_error: true dist: runs-on: ubuntu-latest steps: - - uses: actions/checkout@v3 - with: - fetch-depth: 0 + - uses: actions/checkout@v3 + with: + fetch-depth: 0 - - name: Build - run: pipx run build + - name: Build + run: pipx run build - - uses: actions/upload-artifact@v3 - with: - path: dist/* + - uses: actions/upload-artifact@v3 + with: + path: dist/* - - name: Check metadata - run: pipx run twine check dist/* + - name: Check metadata + run: pipx run twine check dist/* docs: runs-on: ubuntu-latest steps: - - uses: actions/checkout@v3 - with: - fetch-depth: 0 - - uses: actions/setup-python@v4 - with: - python-version: "3.10" - - name: Install dependencies - run: \| - pip install -U -q -e .[docs] - pip list - - name: build docs - run: \| - sphinx-build -b html docs docs/_build/html - touch docs/_build/html/.nojekyll - - - name: Deploy docs to GitHub Pages - if: success() && github.event_name == 'push' && github.ref == 'refs/heads/master' - uses: peaceiris/actions-gh-pages@v3 - with: - github_token: ${{ secrets.GITHUB_TOKEN }} - publish_dir: docs/_build/html - force_orphan: true - user_name: 'github-actions[bot]' - user_email: 'github-actions[bot]@users.noreply.github.com' - commit_message: Deploy to GitHub pages + - uses: actions/checkout@v3 + with: + fetch-depth: 0 + - uses: actions/setup-python@v4 + with: + python-version: "3.10" + - name: Install dependencies + run: \| + pip install -U -q -e .[docs] + pip list + - name: build docs + run: \| + sphinx-build -b html docs docs/_build/html + touch docs/_build/html/.nojekyll + + - name: Deploy docs to GitHub Pages + if: success() && github.event_name == 'push' && github.ref == 'refs/heads/master' + uses: peaceiris/actions-gh-pages@v3 + with: + github_token: ${{ secrets.GITHUB_TOKEN }} + publish_dir: docs/_build/html + force_orphan: true + user_name: 'github-actions[bot]' + user_email: 'github-actions[bot]@users.noreply.github.com' + commit_message: Deploy to GitHub pages publish: - needs: [dist] + needs: [ dist ] runs-on: ubuntu-latest if: github.event_name == 'release' && github.event.action == 'published' steps: - - uses: actions/download-artifact@v3 - with: - name: artifact - path: dist - - - uses: pypa/[email protected] - with: - password: ${{ secrets.pypi_password }} + - uses: actions/download-artifact@v3 + with: + name: artifact + path: dist + + - uses: pypa/[email protected] + with: + password: ${{ secrets.pypi_password }} diff --git a/.pre-commit-config.yaml b/.pre-commit-config.yaml index 29419ede..f1151918 100644 --- a/.pre-commit-config.yaml +++ b/.pre-commit-config.yaml @@ -1,94 +1,99 @@ repos: - - repo: https://github.com/psf/black - rev: 22.8.0 - hooks: - - id: black - - repo: https://github.com/pre-commit/pre-commit-hooks - rev: v4.3.0 - hooks: - - id: check-added-large-files - args: [ '--maxkb=1000' ] - - id: mixed-line-ending - exclude: ^notebooks/ - - id: trailing-whitespace - exclude: ^notebooks/ - - id: check-merge-conflict - - id: check-case-conflict - - id: check-symlinks - - id: check-yaml - exclude: ^notebooks/ - - id: requirements-txt-fixer - - id: debug-statements - - id: end-of-file-fixer - - repo: https://github.com/mgedmin/check-manifest - rev: "0.48" - hooks: - - id: check-manifest - args: [ --update, --no-build-isolation ] - additional_dependencies: [ setuptools-scm ] - - repo: https://github.com/pre-commit/mirrors-mypy - rev: v0.971 - hooks: - - id: mypy - files: src +- repo: https://github.com/psf/black + rev: 22.6.0 + hooks: + - id: black +- repo: https://github.com/pre-commit/pre-commit-hooks + rev: v4.3.0 + hooks: + - id: check-added-large-files + args: ['--maxkb=1000'] + - id: mixed-line-ending + exclude: ^notebooks/ + - id: trailing-whitespace + exclude: ^notebooks/ + - id: check-merge-conflict + - id: check-case-conflict + - id: check-symlinks + - id: check-yaml + exclude: ^notebooks/ + - id: requirements-txt-fixer + - id: debug-statements + - id: end-of-file-fixer +- repo: https://github.com/mgedmin/check-manifest + rev: "0.48" + hooks: + - id: check-manifest + args: [ --update, --no-build-isolation ] + additional_dependencies: [ setuptools-scm ] +- repo: https://github.com/pre-commit/mirrors-mypy + rev: v0.971 + hooks: + - id: mypy + files: src - - repo: https://github.com/pre-commit/pygrep-hooks - rev: v1.9.0 - hooks: - - id: python-use-type-annotations - - id: python-check-mock-methods - - id: python-no-eval - - id: rst-backticks - - id: rst-directive-colons +- repo: https://github.com/roy-ht/pre-commit-jupyter + rev: v1.2.1 + hooks: + - id: jupyter-notebook-cleanup - - repo: https://github.com/asottile/pyupgrade - rev: v2.37.3 - hooks: - - id: pyupgrade - args: [ --py37-plus ] +- repo: https://github.com/pre-commit/pygrep-hooks + rev: v1.9.0 + hooks: + - id: python-use-type-annotations + - id: python-check-mock-methods + - id: python-no-eval + - id: rst-backticks + - id: rst-directive-colons - - repo: https://github.com/asottile/setup-cfg-fmt - rev: v2.0.0 - hooks: - - id: setup-cfg-fmt +- repo: https://github.com/asottile/pyupgrade + rev: v2.37.3 + hooks: + - id: pyupgrade + args: [ --py37-plus ] - # Notebook formatting - - repo: https://github.com/nbQA-dev/nbQA - rev: 1.4.0 - hooks: - - id: nbqa-isort - additional_dependencies: [ isort==5.6.4 ] +- repo: https://github.com/asottile/setup-cfg-fmt + rev: v2.0.0 + hooks: + - id: setup-cfg-fmt - - id: nbqa-pyupgrade - additional_dependencies: [ pyupgrade==2.7.4 ] - args: [ --py37-plus ] +# Notebook formatting +- repo: https://github.com/nbQA-dev/nbQA + rev: 1.4.0 + hooks: + - id: nbqa-isort + additional_dependencies: [ isort==5.6.4 ] + - id: nbqa-pyupgrade + additional_dependencies: [ pyupgrade==2.7.4 ] + args: [ --py37-plus ] - - repo: https://github.com/roy-ht/pre-commit-jupyter - rev: v1.2.1 - hooks: - - id: jupyter-notebook-cleanup - - repo: https://github.com/sondrelg/pep585-upgrade - rev: 'v1.0' - hooks: - - id: upgrade-type-hints - args: [ '--futures=true' ] +- repo: https://github.com/roy-ht/pre-commit-jupyter + rev: v1.2.1 + hooks: + - id: jupyter-notebook-cleanup - - repo: https://github.com/shssoichiro/oxipng - rev: acdd66b4c5a5fda8b08281ad9b3216327b6847b4 - hooks: - - id: oxipng +- repo: https://github.com/sondrelg/pep585-upgrade + rev: 'v1.0' + hooks: + - id: upgrade-type-hints + args: [ '--futures=true' ] - - repo: https://github.com/dannysepler/rm_unneeded_f_str - rev: v0.1.0 - hooks: - - id: rm-unneeded-f-str +- repo: https://github.com/shssoichiro/oxipng + rev: acdd66b4c5a5fda8b08281ad9b3216327b6847b4 + hooks: + - id: oxipng - - repo: https://github.com/python-jsonschema/check-jsonschema - rev: 0.18.2 - hooks: - - id: check-github-workflows - - id: check-github-actions - - id: check-dependabot - - id: check-readthedocs +- repo: https://github.com/dannysepler/rm_unneeded_f_str + rev: v0.1.0 + hooks: + - id: rm-unneeded-f-str + +- repo: https://github.com/python-jsonschema/check-jsonschema + rev: 0.18.2 + hooks: + - id: check-github-workflows + - id: check-github-actions + - id: check-dependabot + - id: check-readthedocs diff --git a/CHANGELOG.rst b/CHANGELOG.rst index 6c983076..0bd52acc 100644 --- a/CHANGELOG.rst +++ b/CHANGELOG.rst @@ -13,7 +13,7 @@ Version 0.5.0 Version 0.4.0 ************* -* loss: upgrade API to use `create_new` to make sure that the losses are comparable. Compatible with zfit 0.6.4+ +* loss: upgrade API to use ``create_new`` to make sure that the losses are comparable. Compatible with zfit 0.6.4+ Version 0.3.1 *********** @@ -24,7 +24,7 @@ Version 0.3.1 Version 0.3.0 *********** * New documentation style -* `hepstats` can now do hypothesis tests, and compute upper limits and confidence intervals for counting analysis +* hepstats can now do hypothesis tests, and compute upper limits and confidence intervals for counting analysis * Progess bars are used to see the progression of the generation of the toys Version 0.2.5 @@ -34,8 +34,8 @@ Version 0.2.5 function if the loss is not extended * hepstats.hypotests can now be used even if there is no nuisances. The pll function in utils/fit/diverse.py had to be modified such that if there are no nuisances, the pll function returns the value of the loss function. -* add notebooks demos for FC intervals with the `FrequentistCalculator` and `AsymptoticCalculator`. -* add warnings when multiple roots are found in `ConfidenceInterval` +* add notebooks demos for FC intervals with the ``FrequentistCalculator`` and ``AsymptoticCalculator``. +* add warnings when multiple roots are found in ``ConfidenceInterval`` * move toys .yml files from notebook to notebook/toys Version 0.2.4 @@ -59,8 +59,8 @@ Version 0.2.1 * Addition of the FrequentistCalculator to performs hypothesis test, upper limit and interval calculations with toys. Toys can be saved and loaded in / from yaml files using the methods: - * `to_yaml` - * `from_yaml` + * ``to_yaml`` + * ``from_yaml`` Version 0.2.0 ************** diff --git a/docs/api/utils.rst b/docs/api/utils.rst index d019d137..1eab1a48 100644 --- a/docs/api/utils.rst +++ b/docs/api/utils.rst @@ -18,10 +18,10 @@ A fitting library should provide six basic objects: * minimizer * fitresult (optional) -A function for each object is defined in this module, all should return `True` to work +A function for each object is defined in this module, all should return ``True`` to work with hepstats. -The `zfit` API is currently the standard fitting API in hepstats. +The zfit API is currently the standard fitting API in hepstats. .. currentmodule:: hepstats.utils.fit.api_check diff --git a/docs/getting_started/splot.rst b/docs/getting_started/splot.rst index 19d72b7c..4267922c 100644 --- a/docs/getting_started/splot.rst +++ b/docs/getting_started/splot.rst @@ -1,7 +1,7 @@ splot ##### -A full example using the sPlot algorithm can be found `here <https://github.com/scikit-hep/hepstats/tree/master/notebooks/splots/splot_example.ipynb>`_ . sWeights for different components in a data sample, modeled with a sum of extended probability density functions, are derived using the `compute_sweights` function: +A full example using the sPlot algorithm can be found `here <https://github.com/scikit-hep/hepstats/tree/master/notebooks/splots/splot_example.ipynb>`_ . sWeights for different components in a data sample, modeled with a sum of extended probability density functions, are derived using the ``compute_sweights`` function: .. code-block:: python diff --git a/setup.cfg b/setup.cfg index c24b3991..a70c80d0 100644 --- a/setup.cfg +++ b/setup.cfg @@ -38,6 +38,7 @@ install_requires = pandas scipy tqdm + uhi python_requires = >=3.7 package_dir = = src diff --git a/src/hepstats/splot/sweights.py b/src/hepstats/splot/sweights.py index 4a781bde..ac927f6e 100644 --- a/src/hepstats/splot/sweights.py +++ b/src/hepstats/splot/sweights.py @@ -22,7 +22,7 @@ def is_sum_of_extended_pdfs(model) -> bool: if not hasattr(model, "get_models"): return False - return all(m.is_extended for m in model.get_models()) + return all(m.is_extended for m in model.get_models()) and model.is_extended def compute_sweights(model, x: np.ndarray) -> dict[Any, np.ndarray]: diff --git a/src/hepstats/utils/__init__.py b/src/hepstats/utils/__init__.py index 2627ff03..9ae943fe 100644 --- a/src/hepstats/utils/__init__.py +++ b/src/hepstats/utils/__init__.py @@ -1,1 +1,9 @@ -from .fit import eval_pdf, array2dataset, pll, base_sampler, base_sample, get_value +from .fit import ( + eval_pdf, + array2dataset, + pll, + base_sampler, + base_sample, + get_value, + set_values, +) diff --git a/src/hepstats/utils/fit/__init__.py b/src/hepstats/utils/fit/__init__.py index 39912ca8..d67fc4aa 100644 --- a/src/hepstats/utils/fit/__init__.py +++ b/src/hepstats/utils/fit/__init__.py @@ -1,2 +1,2 @@ from .sampling import base_sampler, base_sample -from .diverse import get_value, eval_pdf, pll, array2dataset, get_nevents +from .diverse import get_value, eval_pdf, pll, array2dataset, get_nevents, set_values diff --git a/src/hepstats/utils/fit/api_check.py b/src/hepstats/utils/fit/api_check.py index 96c800a8..53509816 100644 --- a/src/hepstats/utils/fit/api_check.py +++ b/src/hepstats/utils/fit/api_check.py @@ -16,6 +16,9 @@ The `zfit` API is currently the standard fitting API in hepstats. """ +import warnings + +import uhi.typing.plottable def is_valid_parameter(object): @@ -54,7 +57,10 @@ def is_valid_data(object): has_weights = hasattr(object, "weights") has_set_weights = hasattr(object, "set_weights") has_space = hasattr(object, "space") - return has_nevents and has_weights and has_set_weights and has_space + is_histlike = isinstance(object, uhi.typing.plottable.PlottableHistogram) + return ( + has_nevents and has_weights and has_set_weights and has_space + ) or is_histlike def is_valid_pdf(object): diff --git a/src/hepstats/utils/fit/diverse.py b/src/hepstats/utils/fit/diverse.py index 22a4c926..7dc5ffc2 100644 --- a/src/hepstats/utils/fit/diverse.py +++ b/src/hepstats/utils/fit/diverse.py @@ -1,14 +1,24 @@ -from contextlib import ExitStack +from contextlib import ExitStack, contextmanager import numpy as np +from .api_check import is_valid_pdf + + +def get_ndims(dataset): + """Return the number of dimensions in the dataset""" + return len(dataset.obs) + def get_value(value): return np.array(value) -def eval_pdf(model, x, params={}, allow_extended=False): +def eval_pdf(model, x, params=None, allow_extended=False): """Compute pdf of model at a given point x and for given parameters values""" + if params is None: + params = {} + def pdf(model, x): if model.is_extended and allow_extended: ret = model.ext_pdf(x) @@ -46,6 +56,16 @@ def pll(minimizer, loss, pois) -> float: return value +@contextmanager +def set_values(params, values): + old_values = [p.value() for p in params] + for p, v in zip(params, values): + p.set_value(v) + yield + for p, v in zip(params, old_values): + p.set_value(v) + + def array2dataset(dataset_cls, obs, array, weights=None): """ dataset_cls: only used to get the class in which array/weights will be @@ -53,9 +73,9 @@ def array2dataset(dataset_cls, obs, array, weights=None): """ if hasattr(dataset_cls, "from_numpy"): - return dataset_cls.from_numpy(obs=obs, array=array, weights=weights) + return dataset_cls.from_numpy(obs, array=array, weights=weights) else: - return dataset_cls(obs=obs, array=array, weights=weights) + return dataset_cls(obs, array=array, weights=weights) def get_nevents(dataset):	Asymptotic calculator histogram dimensionality When generating the asimov histogram associated with a particular test, why is the generated histogram limited to a one-dimensional fit, implicitly by the call to `space.limit1d` [here](https://github.com/scikit-hep/hepstats/blob/cf650da0408de083e27c52b952fe4ef144014de8/src/hepstats/hypotests/calculators/asymptotic_calculator.py#L34)? This seems to exclude the possibility for hypothesis tests involving a likelihood function evaluated using a PDF with dimensionality greater than 1 which, in my naive understanding, is a desirable use-case for this calculator. Thanks for all your work!	Hey, that's a good catch! Indeed, this should be working for n-dims and is not generalized. Since zfit itself offers now also to directly create binned histograms, this could be the best way to go: https://zfit-tutorials.readthedocs.io/en/latest/tutorials/components/30%20-%20Binned%20models.html#binned-models-from-histograms Or we could simply generalize it as it is now, the problem is maybe that it only takes the midpoint, which is probably often good enough but maybe creates a small bias. I suppose either way this will require some development (and a new release!) before I can use it. There is no workaround, right? Hey, as far as I remember the asymptotic method was only documented for 1D problems (I.e 1D parameter of interest) You should probably check the papers and see if there is an extension to N-dim. It's a 1D poi, but that's a different thing to requiring a 1D space for the asimov dataset. The pdf can be multidimensional while the scan is still over a 1 dimensional yield parameter, for example.	2022-08-30T10:10:21	[]	[]
etianen/django-s3-storage	etianen__django-s3-storage-161	62c41bbf9696bc62df35590b8eb9b55468e88c4e	diff --git a/CHANGELOG.rst b/CHANGELOG.rst index 5869a4a..5165600 100644 --- a/CHANGELOG.rst +++ b/CHANGELOG.rst @@ -1,6 +1,11 @@ django-s3-storage changelog =========================== +0.15.0 +------ + +- Added support for opening files in text mode on Python 3.11+ (@etianen). + 0.14.0 ------ diff --git a/django_s3_storage/__init__.py b/django_s3_storage/__init__.py index 7f723d0..ce66fdf 100644 --- a/django_s3_storage/__init__.py +++ b/django_s3_storage/__init__.py @@ -3,4 +3,4 @@ """ -__version__ = (0, 14, 0) +__version__ = (0, 15, 0) diff --git a/django_s3_storage/storage.py b/django_s3_storage/storage.py index db7a624..a10f51a 100644 --- a/django_s3_storage/storage.py +++ b/django_s3_storage/storage.py @@ -6,15 +6,15 @@ import shutil from contextlib import closing from datetime import timezone -from functools import wraps, partial -from io import TextIOBase +from functools import partial, wraps +from io import TextIOBase, TextIOWrapper from tempfile import SpooledTemporaryFile from threading import local from urllib.parse import urljoin, urlsplit, urlunsplit import boto3 -from botocore.client import Config from boto3.s3.transfer import TransferConfig +from botocore.client import Config from botocore.exceptions import ClientError from django.conf import settings from django.contrib.staticfiles.storage import ManifestFilesMixin @@ -41,6 +41,7 @@ def _do_wrap_errors(self, name, args, kwargs): if code == "NoSuchKey": err_cls = FileNotFoundError raise err_cls(f"S3Storage error at {name!r}: {force_str(ex)}") + return _do_wrap_errors @@ -63,6 +64,7 @@ def do_wrap_path_impl(self, name, args, *kwargs): # posix paths. We fix this by converting paths to system form, passing them to the default implementation, then # converting them back to posix paths. return _to_posix_path(func(self, _to_sys_path(name), args, kwargs)) + return do_wrap_path_impl @@ -108,18 +110,24 @@ def __init__(self, storage): if storage.settings.AWS_ACCESS_KEY_ID: connection_kwargs["aws_access_key_id"] = storage.settings.AWS_ACCESS_KEY_ID if storage.settings.AWS_SECRET_ACCESS_KEY: - connection_kwargs["aws_secret_access_key"] = storage.settings.AWS_SECRET_ACCESS_KEY + connection_kwargs[ + "aws_secret_access_key" + ] = storage.settings.AWS_SECRET_ACCESS_KEY if storage.settings.AWS_SESSION_TOKEN: connection_kwargs["aws_session_token"] = storage.settings.AWS_SESSION_TOKEN if storage.settings.AWS_S3_ENDPOINT_URL: connection_kwargs["endpoint_url"] = storage.settings.AWS_S3_ENDPOINT_URL self.session = boto3.session.Session() - self.s3_connection = self.session.client("s3", config=Config( - s3={"addressing_style": storage.settings.AWS_S3_ADDRESSING_STYLE}, - signature_version=storage.settings.AWS_S3_SIGNATURE_VERSION, - max_pool_connections=storage.settings.AWS_S3_MAX_POOL_CONNECTIONS, - connect_timeout=storage.settings.AWS_S3_CONNECT_TIMEOUT - ), connection_kwargs) + self.s3_connection = self.session.client( + "s3", + config=Config( + s3={"addressing_style": storage.settings.AWS_S3_ADDRESSING_STYLE}, + signature_version=storage.settings.AWS_S3_SIGNATURE_VERSION, + max_pool_connections=storage.settings.AWS_S3_MAX_POOL_CONNECTIONS, + connect_timeout=storage.settings.AWS_S3_CONNECT_TIMEOUT, + ), + connection_kwargs, + ) @deconstructible @@ -155,7 +163,7 @@ class S3Storage(Storage): "AWS_S3_FILE_OVERWRITE": False, "AWS_S3_USE_THREADS": True, "AWS_S3_MAX_POOL_CONNECTIONS": 10, - "AWS_S3_CONNECT_TIMEOUT": 60 # 60 seconds + "AWS_S3_CONNECT_TIMEOUT": 60, # 60 seconds } s3_settings_suffix = "" @@ -178,16 +186,24 @@ def _setup(self): setting_key, self._kwargs.get( setting_key.lower(), - getattr(settings, setting_key + self.s3_settings_suffix, setting_default_value), + getattr( + settings, + setting_key + self.s3_settings_suffix, + setting_default_value, + ), ), ) # Validate settings. if not self.settings.AWS_S3_BUCKET_NAME: - raise ImproperlyConfigured(f"Setting AWS_S3_BUCKET_NAME{self.s3_settings_suffix} is required.") + raise ImproperlyConfigured( + f"Setting AWS_S3_BUCKET_NAME{self.s3_settings_suffix} is required." + ) # Create a thread-local connection manager. self._connections = _Local(self) # Set transfer config for S3 operations - self._transfer_config = TransferConfig(use_threads=self.settings.AWS_S3_USE_THREADS) + self._transfer_config = TransferConfig( + use_threads=self.settings.AWS_S3_USE_THREADS + ) @property def s3_connection(self): @@ -201,8 +217,8 @@ def __init__(self, kwargs): # Check for unknown kwargs. for kwarg_key in kwargs.keys(): if ( - kwarg_key.upper() not in self.default_auth_settings and - kwarg_key.upper() not in self.default_s3_settings + kwarg_key.upper() not in self.default_auth_settings + and kwarg_key.upper() not in self.default_s3_settings ): raise ImproperlyConfigured(f"Unknown S3Storage parameter: {kwarg_key}") # Set up the storage. @@ -211,7 +227,9 @@ def __init__(self, *kwargs): # Re-initialize the storage if an AWS setting changes. setting_changed.connect(self._setting_changed_received) # Register system checks. - checks.register(partial(self.__class__._system_checks, self), checks.Tags.security) + checks.register( + partial(self.__class__._system_checks, self), checks.Tags.security + ) # All done! super().__init__() @@ -237,7 +255,9 @@ def __reduce__(self): def _get_key_name(self, name): if name.startswith("/"): name = name[1:] - return posixpath.normpath(posixpath.join(self.settings.AWS_S3_KEY_PREFIX, _to_posix_path(name))) + return posixpath.normpath( + posixpath.join(self.settings.AWS_S3_KEY_PREFIX, _to_posix_path(name)) + ) def _object_params(self, name): params = { @@ -256,22 +276,29 @@ def _object_put_params(self, name): ), "Metadata": { key: _callable_setting(value, name) - for key, value - in self.settings.AWS_S3_METADATA.items() + for key, value in self.settings.AWS_S3_METADATA.items() }, - "StorageClass": "REDUCED_REDUNDANCY" if self.settings.AWS_S3_REDUCED_REDUNDANCY else "STANDARD", + "StorageClass": "REDUCED_REDUNDANCY" + if self.settings.AWS_S3_REDUCED_REDUNDANCY + else "STANDARD", } params.update(self._object_params(name)) # Set content disposition. - content_disposition = _callable_setting(self.settings.AWS_S3_CONTENT_DISPOSITION, name) + content_disposition = _callable_setting( + self.settings.AWS_S3_CONTENT_DISPOSITION, name + ) if content_disposition: params["ContentDisposition"] = content_disposition # Set content langauge. - content_langauge = _callable_setting(self.settings.AWS_S3_CONTENT_LANGUAGE, name) + content_langauge = _callable_setting( + self.settings.AWS_S3_CONTENT_LANGUAGE, name + ) if content_langauge: params["ContentLanguage"] = content_langauge # Set server-side encryption. - if self.settings.AWS_S3_ENCRYPT_KEY: # If this if False / None / empty then no encryption + if ( + self.settings.AWS_S3_ENCRYPT_KEY + ): # If this if False / None / empty then no encryption if isinstance(self.settings.AWS_S3_ENCRYPT_KEY, str): params["ServerSideEncryption"] = self.settings.AWS_S3_ENCRYPT_KEY if self.settings.AWS_S3_KMS_ENCRYPTION_KEY_ID: @@ -283,11 +310,11 @@ def _object_put_params(self, name): def new_temporary_file(self): """Returns a new file to use when opening from or saving to S3""" - return SpooledTemporaryFile(max_size=1024102410) # 10 MB. + return SpooledTemporaryFile(max_size=1024 1024 * 10) # 10 MB. @_wrap_errors def _open(self, name, mode="rb"): - if mode != "rb": + if mode not in ("rb", "rt", "r"): raise ValueError("S3 files can only be opened in read-only mode") # Load the key into a temporary file. It would be nice to stream the # content, but S3 doesn't support seeking, which is sometimes needed. @@ -298,6 +325,9 @@ def _open(self, name, mode="rb"): # Un-gzip if required. if obj.get("ContentEncoding") == "gzip": content = gzip.GzipFile(name, "rb", fileobj=content) + # Decode text if required. + if "b" not in mode: + content = TextIOWrapper(content) # All done! return S3File(content, name, self) @@ -325,7 +355,12 @@ def _save(self, name, content): # Check if the content type is compressible. content_type_family, content_type_subtype = content_type.lower().split("/") content_type_subtype = content_type_subtype.split("+")[-1] - if content_type_family == "text" or content_type_subtype in ("xml", "json", "html", "javascript"): + if content_type_family == "text" or content_type_subtype in ( + "xml", + "json", + "html", + "javascript", + ): # Compress the content. temp_file = self.new_temporary_file() temp_files.append(temp_file) @@ -337,7 +372,9 @@ def _save(self, name, content): temp_file.seek(0) content = temp_file put_params["ContentEncoding"] = "gzip" - put_params["Metadata"][_UNCOMPRESSED_SIZE_META_KEY] = f"{orig_size:d}" + put_params["Metadata"][ + _UNCOMPRESSED_SIZE_META_KEY + ] = f"{orig_size:d}" else: content.seek(0) # Save the file. @@ -346,8 +383,13 @@ def _save(self, name, content): original_close = content.close content.close = lambda: None try: - self.s3_connection.upload_fileobj(content, put_params.pop('Bucket'), put_params.pop('Key'), - ExtraArgs=put_params, Config=self._transfer_config) + self.s3_connection.upload_fileobj( + content, + put_params.pop("Bucket"), + put_params.pop("Key"), + ExtraArgs=put_params, + Config=self._transfer_config, + ) finally: # Restore the original close method. content.close = original_close @@ -385,8 +427,8 @@ def delete(self, name): @_wrap_errors def copy(self, src_name, dst_name): self.s3_connection.copy_object( - CopySource=self._object_params(src_name), - self._object_params(dst_name)) + CopySource=self._object_params(src_name), self._object_params(dst_name) + ) @_wrap_errors def rename(self, src_name, dst_name): @@ -402,7 +444,8 @@ def exists(self, name): results = self.s3_connection.list_objects_v2( Bucket=self.settings.AWS_S3_BUCKET_NAME, MaxKeys=1, - Prefix=self._get_key_name(name) + "/", # Add the slash again, since _get_key_name removes it. + Prefix=self._get_key_name(name) + + "/", # Add the slash again, since _get_key_name removes it. ) except ClientError: return False @@ -449,7 +492,9 @@ def url(self, name, extra_params=None, client_method="get_object"): # Use a public URL, if specified. if self.settings.AWS_S3_PUBLIC_URL: if extra_params or client_method != "get_object": - raise ValueError("Use of extra_params or client_method is not allowed with AWS_S3_PUBLIC_URL") + raise ValueError( + "Use of extra_params or client_method is not allowed with AWS_S3_PUBLIC_URL" + ) return urljoin(self.settings.AWS_S3_PUBLIC_URL, filepath_to_uri(name)) # Otherwise, generate the URL. params = extra_params.copy() if extra_params else {} @@ -462,8 +507,16 @@ def url(self, name, extra_params=None, client_method="get_object"): # Strip off the query params if we're not interested in bucket auth. if not self.settings.AWS_S3_BUCKET_AUTH: if extra_params or client_method != "get_object": - raise ValueError("Use of extra_params or client_method is not allowed with AWS_S3_BUCKET_AUTH") - url = urlunsplit(urlsplit(url)[:3] + ("", "",)) + raise ValueError( + "Use of extra_params or client_method is not allowed with AWS_S3_BUCKET_AUTH" + ) + url = urlunsplit( + urlsplit(url)[:3] + + ( + "", + "", + ) + ) # All done! return url @@ -501,8 +554,9 @@ def sync_meta_iter(self): put_params["ContentEncoding"] = content_encoding if content_encoding == "gzip": try: - put_params["Metadata"][_UNCOMPRESSED_SIZE_META_KEY] = \ - obj["Metadata"][_UNCOMPRESSED_SIZE_META_KEY] + put_params["Metadata"][_UNCOMPRESSED_SIZE_META_KEY] = obj[ + "Metadata" + ][_UNCOMPRESSED_SIZE_META_KEY] except KeyError: pass # Update the metadata. @@ -513,7 +567,7 @@ def sync_meta_iter(self): "Key": self._get_key_name(name), }, MetadataDirective="REPLACE", - put_params + put_params, ) yield name @@ -529,23 +583,28 @@ class StaticS3Storage(S3Storage): """ default_s3_settings = S3Storage.default_s3_settings.copy() - default_s3_settings.update({ - "AWS_S3_BUCKET_AUTH": False, - }) + default_s3_settings.update( + { + "AWS_S3_BUCKET_AUTH": False, + } + ) s3_settings_suffix = "_STATIC" class ManifestStaticS3Storage(ManifestFilesMixin, StaticS3Storage): - default_s3_settings = StaticS3Storage.default_s3_settings.copy() - default_s3_settings.update({ - "AWS_S3_MAX_AGE_SECONDS_CACHED": 60 * 60 * 24 * 365, # 1 year. - }) + default_s3_settings.update( + { + "AWS_S3_MAX_AGE_SECONDS_CACHED": 60 * 60 * 24 * 365, # 1 year. + } + ) def post_process(self, args, kwargs): initial_aws_s3_max_age_seconds = self.settings.AWS_S3_MAX_AGE_SECONDS - self.settings.AWS_S3_MAX_AGE_SECONDS = self.settings.AWS_S3_MAX_AGE_SECONDS_CACHED + self.settings.AWS_S3_MAX_AGE_SECONDS = ( + self.settings.AWS_S3_MAX_AGE_SECONDS_CACHED + ) try: yield from super().post_process(args, **kwargs) finally: diff --git a/pyproject.toml b/pyproject.toml new file mode 100644 index 0000000..e69de29	Cannot open files in plain text mode I am working with CSV files which need to be opened in plain text mode. With django-storage I had been using: with object.file.open('r') as f: csv_reader = csv.DictReader(f) But that gives me an error and I need to open it in binary mode and do this instead: with object.file.open() as f: csv_reader = csv.DictReader(io.StringIO(f.read().decode('utf-8')) I'm wondering if this is a bug or if this is how my code should change. (I am loving the product, I like that it seeks(0) and handles encoding properly when I am writing CSV files. It is much closer to the the native Django Storage.)		2023-11-04T15:34:56	[]	[]
frostming/lektor-tailwind	frostming__lektor-tailwind-5	b94a92ec0aa2209dbc4e73f954e3336792376080	diff --git a/lektor_tailwind.py b/lektor_tailwind.py index 968f21d..b3bdf5d 100644 --- a/lektor_tailwind.py +++ b/lektor_tailwind.py @@ -39,17 +39,19 @@ def init_tailwindcss(self): def _run_watcher(self, output_path: str): if not self.input_exists(): return - self.tailwind = subprocess.Popen( - [ - self.tailwind_bin, - "-i", - self.input_css, - "-o", - os.path.join(output_path, self.css_path), - "-w", - ], - cwd=self.env.root_path, - ) + + cmd = [ + self.tailwind_bin, + "--input", + self.input_css, + "--output", + os.path.join(output_path, self.css_path), + "--watch", + ] + if os.environ.get("NODE_ENV") == "production": + cmd.append("--minify") + + self.tailwind = subprocess.Popen(cmd, cwd=self.env.root_path) def input_exists(self) -> bool: return os.path.exists(self.input_css)	Tailwind fails if `lektor build` run from outside the Lektor project directory Since relative paths are used to configure the `content` paths in `tailwind.config.js`, tailwind can not find the source content unless it is run from the Lektor project directory. Lektor-tailwind currently fails to chdir to the Lektor project directory before running tailwind for `lektor build`. (It does chdir before running tailwind in watch mode for `lektor server`.) This fixes that. Tailwind fails if `lektor build` run from outside the Lektor project directory Since relative paths are used to configure the `content` paths in `tailwind.config.js`, tailwind can not find the source content unless it is run from the Lektor project directory. Lektor-tailwind currently fails to chdir to the Lektor project directory before running tailwind for `lektor build`. (It does chdir before running tailwind in watch mode for `lektor server`.) This fixes that.		2023-08-09T16:57:31	[]	[]
readthedocs/sphinx_rtd_theme	readthedocs__sphinx_rtd_theme-1576	2c19bb149bb87bd14ccc4ad2d30d9e7d35a8d925	diff --git a/.circleci/config.yml b/.circleci/config.yml index d7aa9d7ee..dfe6f3644 100644 --- a/.circleci/config.yml +++ b/.circleci/config.yml @@ -44,54 +44,51 @@ jobs: - image: 'cimg/python:3.9-node' steps: - run-build: {} - py36: - docker: - - image: 'cimg/python:3.6' - steps: - - run-tox: - version: py36 - sphinx-version: "50,51,52,53" - py37: - docker: - - image: 'cimg/python:3.7' - steps: - - run-tox: - version: py37 - sphinx-version: "50,51,52,53" py38: docker: - image: 'cimg/python:3.8' steps: - run-tox: version: py38 - sphinx-version: "50,51,52,53,60,61,62,70,71,latest" + sphinx-version: "60,61,62,70,71" py39: docker: - image: 'cimg/python:3.9' steps: - run-tox: version: py39 - sphinx-version: "50,51,52,53,60,61,62,70,71,72,latest" + sphinx-version: "60,61,62,70,71,72,73,74" py310: docker: - image: 'cimg/python:3.10' steps: - run-tox: version: py310 - sphinx-version: "50,51,52,53,60,61,62,70,71,72,latest" + sphinx-version: "60,61,62,70,71,72,73,74,80,latest,dev" py311: docker: - image: 'cimg/python:3.11' steps: - run-tox: version: py311 - sphinx-version: "53,60,61,62,70,71,72,latest,dev" + sphinx-version: "72,73,74,80,latest,dev" + + py312: + docker: + - image: 'cimg/python:3.12' + steps: + - run-tox: + version: py312 + sphinx-version: "72,73,74,80,latest,dev" workflows: version: 2 tests: jobs: - build + - py312: + requires: + - build - py311: requires: - build @@ -104,9 +101,3 @@ workflows: - py38: requires: - build - - py37: - requires: - - build - - py36: - requires: - - build diff --git a/docs/conf.py b/docs/conf.py index 0c3731fc9..25d002c02 100644 --- a/docs/conf.py +++ b/docs/conf.py @@ -68,10 +68,6 @@ range(1, 100) )) -if 'READTHEDOCS' in os.environ: - html_context['READTHEDOCS'] = True - html_context['READTHEDOCS_PROJECT'] = os.environ.get("READTHEDOCS_PROJECT") - html_logo = "demo/static/logo-wordmark-light.svg" html_show_sourcelink = True html_favicon = "demo/static/favicon.ico" diff --git a/docs/configuring.rst b/docs/configuring.rst index e76c2aeda..3718abc27 100644 --- a/docs/configuring.rst +++ b/docs/configuring.rst @@ -102,7 +102,11 @@ Miscellaneous options :type: string + .. deprecated:: 3.0.0 + The ``analytics_id`` option is deprecated, use the sphinxcontrib-googleanalytics_ extension instead. + .. _gtag.js: https://developers.google.com/gtagjs + .. _sphinxcontrib-googleanalytics: https://pypi.org/project/sphinxcontrib-googleanalytics/ .. confval:: analytics_anonymize_ip @@ -111,6 +115,9 @@ Miscellaneous options :type: boolean :default: ``False`` + .. deprecated:: 3.0.0 + The ``analytics_anonymize_ip`` option is deprecated, use the sphinxcontrib-googleanalytics_ extension instead. + .. confval:: canonical_url This will specify a `canonical URL`_ meta link element to tell search diff --git a/docs/contributing.rst b/docs/contributing.rst index 541a1c264..e8c733d7d 100644 --- a/docs/contributing.rst +++ b/docs/contributing.rst @@ -86,7 +86,7 @@ Use the following steps: # Runs the development webserver $ docker-compose run sphinx_rtd_theme dev - + # If you want to copy stuff out of the Docker environment, run this make # target or read the actual Makefile to see what is going on. # We suggest running this command every time that you want to quickly build @@ -108,13 +108,13 @@ The following cases need to be tested with changes to CSS or JavaScript: We support large, tablet, and mobile viewport sizes * We currently support only HTML5 writer. * Multiple major versions of Sphinx should be tested. - We currently support Sphinx ``>=5.0`` + We currently support Sphinx ``>=6.0`` It's easiest to test combinations of dependency versions using ``tox``: .. code:: console - % tox -e py310-sphinx62 + % tox -e py312-sphinx74 If the tests and build are successful, you can view the built documentation at the directory noted by Sphinx: @@ -123,9 +123,9 @@ the directory noted by Sphinx: build succeeded, 10 warnings. - The HTML pages are in .tox/py310-sphinx62/tmp/html. + The HTML pages are in .tox/py312-sphinx74/tmp/html. ___________________________ summary ___________________________ - py310-sphinx62: commands succeeded + py312-sphinx74: commands succeeded congratulations :) You can then open up this path with a series of browsers to test. @@ -135,12 +135,12 @@ multiple ``tox`` environments, and open both up for comparison: .. code:: console - % tox -e py310-sphinx62 + % tox -e py312-sphinx62 ... - % tox -e py310-sphinx53 + % tox -e py312-sphinx74 ... - % firefox .tox/py310-sphinx62/tmp/html/index.html - % firefox .tox/py310-sphinx53/tmp/html/index.html + % firefox .tox/py312-sphinx62/tmp/html/index.html + % firefox .tox/py312-sphinx74/tmp/html/index.html You can also use a separate ``tox`` environment for building output to compare against. All of the ``tox`` environments have an additional postfix, ``-qa``, to @@ -167,11 +167,11 @@ Currently, the most important environments to QA are: the cases above, we'll be here all day. Python 3, and latest minor of each major Sphinx release should give us enough work. -* ``py310-sphinx53`` -* ``py310-sphinx62`` -* ``py310-sphinx72`` -* ``py310-sphinxlatest`` -* ``py310-sphinxdev`` +* ``py312-sphinx62`` +* ``py312-sphinx74`` +* ``py312-sphinx80`` +* ``py312-sphinxlatest`` +* ``py312-sphinxdev`` Translations ============ diff --git a/docs/development.rst b/docs/development.rst index c4ee9a43c..298fbe2b2 100644 --- a/docs/development.rst +++ b/docs/development.rst @@ -36,13 +36,8 @@ combinations. Unsupported browser/operating system combinations include: -Internet Explorer (any OS, versions <=10) - Unsupported. IE11 is the last partially supported version. We do no - testing on prior versions. - -Internet Explorer (any OS, version 11) - We currently only partially support IE11, and only test for major bugs. - Support will be removed in the :ref:`roadmap-release-3.0.0` release. +Internet Explorer (any OS, any version) + Support was removed in the :ref:`roadmap-release-3.0.0` release. Opera (any OS, any version) Community support only. We do not receive enough traffic with this @@ -62,11 +57,14 @@ The theme officially supports the following dependencies in your Sphinx project: * - Dependency - Versions * - Python - - 3.6 or greater + - 3.8 or greater * - Sphinx - - 5 or greater + - 6 or greater * - docutils - - >= 0.14, < 0.19 + - > 0.18, < 0.22 + +.. deprecated:: 3.0 + Sphinx < 6 support removed .. deprecated:: 2.0 Sphinx < 5 support removed @@ -179,10 +177,10 @@ HTML4 support will be removed 3.0.0 ~~~~~ -:Planned release date: 2024 Q1 +:Planned release date: 2024 Q3 -This release is not yet fully planned. -However, we already know that we will be dropping support for older version of different dependencies and browsers. +This release will drop support for old Python and Sphinx versions, +and will add support for new ones. Sphinx 5, Python < 3.8 and docutils < 0.18 support will be removed Sphinx 5 is the latest version that supports Python 3.6 and 3.7, diff --git a/setup.cfg b/setup.cfg index b1ef6084e..f7dfdb972 100644 --- a/setup.cfg +++ b/setup.cfg @@ -29,12 +29,11 @@ classifiers = Intended Audience :: Developers Programming Language :: Python :: 3 Programming Language :: Python :: 3 :: Only - Programming Language :: Python :: 3.6 - Programming Language :: Python :: 3.7 Programming Language :: Python :: 3.8 Programming Language :: Python :: 3.9 Programming Language :: Python :: 3.10 Programming Language :: Python :: 3.11 + Programming Language :: Python :: 3.12 Operating System :: OS Independent Topic :: Documentation Topic :: Software Development :: Documentation @@ -43,10 +42,10 @@ classifiers = include_package_data = True zip_safe = False packages = sphinx_rtd_theme -python_requires = >=3.6 +python_requires = >=3.8 install_requires = - sphinx >=5,<8 - docutils <0.21 + sphinx >=6,<9 + docutils >0.18,<0.22 sphinxcontrib-jquery >=4,<5 tests_require = pytest @@ -54,7 +53,6 @@ tests_require = [options.extras_require] dev = transifex-client - sphinxcontrib-httpdomain bump2version wheel twine diff --git a/sphinx_rtd_theme/__init__.py b/sphinx_rtd_theme/__init__.py index 59be0645d..2cb8e6ad5 100644 --- a/sphinx_rtd_theme/__init__.py +++ b/sphinx_rtd_theme/__init__.py @@ -21,6 +21,10 @@ def get_html_theme_path(): """Return list of HTML theme paths.""" + logger.warning( + _('Calling get_html_theme_path is deprecated. If you are calling it to define html_theme_path, you are safe to remove that code.') + ) + cur_dir = path.abspath(path.dirname(path.dirname(__file__))) return cur_dir @@ -32,6 +36,21 @@ def config_initiated(app, config): _('The canonical_url option is deprecated, use the html_baseurl option from Sphinx instead.') ) + if theme_options.get("analytics_id"): + logger.warning( + _('The analytics_id option is deprecated, use the sphinxcontrib-googleanalytics extension instead.') + ) + + if theme_options.get("analytics_anonymize_ip"): + logger.warning( + _('The analytics_anonymize_ip option is deprecated, use the sphinxcontrib-googleanalytics extension instead.') + ) + + if "extra_css_files" in config.html_context: + logger.warning( + _('The extra_css_file option is deprecated, use the html_css_files option from Sphinx instead.') + ) + def extend_html_context(app, pagename, templatename, context, doctree): # Add ``sphinx_version_info`` tuple for use in Jinja templates @@ -52,7 +71,7 @@ def setup(app): if python_version[0] < 3: logger.error("Python 2 is not supported with sphinx_rtd_theme, update to Python 3.") - app.require_sphinx('5.0') + app.require_sphinx('6.0') if app.config.html4_writer: logger.error("'html4_writer' is not supported with sphinx_rtd_theme.") diff --git a/tox.ini b/tox.ini index d601e5073..a010796ac 100644 --- a/tox.ini +++ b/tox.ini @@ -1,28 +1,22 @@ [tox] envlist = - py{36,37,38,39,310}-sphinx{50,51,52,53}{-qa} py{38,39,310}-sphinx{60,61,62,70,71}{-qa} - py{39,310}-sphinx{72,latest,dev}{-qa} - # Python 3.11 working from Sphinx 5.3 and up - py{311}-sphinx{53,60,61,62,70,71,72,latest,dev}{-qa} + py{39,310,311,312}-sphinx{72,73,74,80}{-qa} + py{310,311,312}-sphinx{latest,dev}{-qa} [testenv] setenv = LANG=C deps = . - readthedocs-sphinx-ext pytest - sphinx50: Sphinx>=5.0,<5.1 - sphinx51: Sphinx>=5.1,<5.2 - sphinx52: Sphinx>=5.2,<5.3 - sphinx53: Sphinx>=5.3,<5.4 sphinx60: Sphinx>=6.0,<6.1 sphinx61: Sphinx>=6.1,<6.2 sphinx62: Sphinx>=6.2,<6.3 sphinx70: Sphinx>=7.0,<7.1 sphinx71: Sphinx>=7.1,<7.2 sphinx72: Sphinx>=7.2,<7.3 + sphinx80: Sphinx>=8.0,<8.1 sphinxlatest: Sphinx sphinxdev: https://github.com/sphinx-doc/sphinx/archive/refs/heads/master.zip allowlist_externals =	How is sphinxcontrib-httpdomain used in the project? ### Problem In `setup.cfg` there is a `dev` extra defined with `sphinxcontrib-httpdomain` as one of the dependencies. https://github.com/readthedocs/sphinx_rtd_theme/blob/8ce23cec96f628ac0d29737bf1cf8cc2e750f068/setup.cfg#L57 I can't find it being used anywhere in the project. Is it needed?		2024-07-24T09:37:19	[]	[]
twopirllc/pandas-ta	twopirllc__pandas-ta-801	23bca1aff2efde5e4fd1215b20416fda616750f7	diff --git a/pandas_ta/utils/_stats.py b/pandas_ta/utils/_stats.py index cffa6ac4..0a85ef32 100644 --- a/pandas_ta/utils/_stats.py +++ b/pandas_ta/utils/_stats.py @@ -1,5 +1,5 @@ # -- coding: utf-8 -- -from numpy import array, infty, log, nan, pi, sqrt +from numpy import array, inf, log, nan, pi, sqrt from pandas_ta._typing import Array, IntFloat, Number, Union from pandas_ta.maps import Imports from pandas_ta.utils import hpoly	Pandas-ta doesnt work with numpy 2.0 So pandas-ta doesnt work with the newest numpy 2.0. Reproducible example: ```import pandas_ta as ta``` Log: ``` ImportError: cannot import name 'NaN' from 'numpy' (C:\Users\User\anaconda3\envs\py312Test\Lib\site-packages\numpy\__init__.py) ``` Would be nice to get support for the latest numpy version. Pandas-ta version 0.3.14b Python: 3.12.2	> So pandas-ta doesnt work with the newest numpy 2.0. That's crazy @Chuck321123! How is that possible? 🤷🏼‍♂️ It was released.... yesterday! 😮 Please inform the following packages that they are not up to date with the newest _numpy 2.0_. I don't want them to miss out on the action. <img width="820" alt="Screenshot 2024-06-17 at 12 13 35 PM" src="https://github.com/twopirllc/pandas-ta/assets/18198392/df73773e-3451-42e4-b613-2f41d3bd4df9"> Note: _streamlit_ is not necessary for pandas-ta > Would be nice to get support for the latest numpy version. Let's do it! ✅ Fork the repo, checkout the _development_ branch, edit (including tests), and submit a PR. 😎 FYI, I'm now running: Python: 3.11.9 & 3.12.4, no longer supporting <=3.9 Pandas TA (_development_) 0.4.17b @twopirllc Unfortunately, I don't know how to do that. However, If anyone knows how to do it I would be very happy I have created the pull request. Have tested with server so it should work. Though haven't touched the package requirements though should also be compatible with 1.22. Kindly check	2024-06-18T19:15:39	[]	[]
beancount/beancount	beancount__beancount-867	9d2ff2817f803cd3b6efe0755498884e074acda1	diff --git a/beancount/core/account.py b/beancount/core/account.py index 673a1ea6f..f7494ddc6 100644 --- a/beancount/core/account.py +++ b/beancount/core/account.py @@ -44,6 +44,32 @@ TYPE = "<AccountDummy>" +def is_valid_root(string: Account) -> bool: + """Return true if the given string is a valid root account name. + This just check for valid syntax. + + Args: + string: A string, to be checked for account name pattern. + Returns: + A boolean, true if the string has the form of a root account's name. + """ + return isinstance(string, str) and bool(regex.fullmatch(ACC_COMP_TYPE_RE, string)) + + +def is_valid_leaf(string: Account) -> bool: + """Return true if the given string is a valid leaf account name. + This just check for valid syntax. + + Args: + string: A string, to be checked for account name pattern. + Returns: + A boolean, true if the string has the form of a leaf account's name. + """ + return isinstance(string, str) and all( + regex.fullmatch(ACC_COMP_NAME_RE, p) for p in string.split(":") + ) + + def is_valid(string: Account) -> bool: """Return true if the given string is a valid account name. This does not check for the root account types, just the general syntax. @@ -53,7 +79,7 @@ def is_valid(string: Account) -> bool: Returns: A boolean, true if the string has the form of an account's name. """ - return isinstance(string, str) and bool(regex.match("{}$".format(ACCOUNT_RE), string)) + return isinstance(string, str) and bool(regex.fullmatch(ACCOUNT_RE, string)) def join(*components: Tuple[str]) -> Account: diff --git a/beancount/parser/options.py b/beancount/parser/options.py index 33f5609f9..598f40848 100644 --- a/beancount/parser/options.py +++ b/beancount/parser/options.py @@ -114,6 +114,36 @@ def options_validate_booking_method(value): raise ValueError(str(exc)) from exc +def options_validate_root_account(value): + """Validate a root account name. + + Args: + value: A string, the value provided as option. + Returns: + The root account name, if the validation succedes. + Raises: + ValueError: If the value is invalid. + """ + if not account.is_valid_root(value): + raise ValueError(f"Invalid root account name: {value!r}") + return value + + +def options_validate_leaf_account(value): + """Validate a leaf account name. + + Args: + value: A string, the value provided as option. + Returns: + The leaf account name, if the validation succedes. + Raises: + ValueError: If the value is invalid. + """ + if not account.is_valid_leaf(value): + raise ValueError(f"Invalid leaf account name: {value!r}") + return value + + # List of option groups, with their description, option names and default # values. OptGroup = collections.namedtuple("OptGroup", "description options") @@ -269,11 +299,15 @@ def Opt( recognizes account names. """, [ - Opt("name_assets", _TYPES.assets), - Opt("name_liabilities", _TYPES.liabilities), - Opt("name_equity", _TYPES.equity), - Opt("name_income", _TYPES.income), - Opt("name_expenses", _TYPES.expenses), + Opt("name_assets", _TYPES.assets, converter=options_validate_root_account), + Opt( + "name_liabilities", + _TYPES.liabilities, + converter=options_validate_root_account, + ), + Opt("name_equity", _TYPES.equity, converter=options_validate_root_account), + Opt("name_income", _TYPES.income, converter=options_validate_root_account), + Opt("name_expenses", _TYPES.expenses, converter=options_validate_root_account), ], ), OptGroup( @@ -281,7 +315,13 @@ def Opt( Leaf name of the equity account used for summarizing previous transactions into opening balances. """, - [Opt("account_previous_balances", "Opening-Balances")], + [ + Opt( + "account_previous_balances", + "Opening-Balances", + converter=options_validate_leaf_account, + ) + ], ), OptGroup( """ @@ -289,7 +329,13 @@ def Opt( earnings from income and expenses accrued before the beginning of the exercise into the balance sheet. """, - [Opt("account_previous_earnings", "Earnings:Previous")], + [ + Opt( + "account_previous_earnings", + "Earnings:Previous", + converter=options_validate_leaf_account, + ) + ], ), OptGroup( """ @@ -298,7 +344,13 @@ def Opt( will essentially "fixup" the basic accounting equation due to the errors that priced conversions introduce. """, - [Opt("account_previous_conversions", "Conversions:Previous")], + [ + Opt( + "account_previous_conversions", + "Conversions:Previous", + converter=options_validate_leaf_account, + ) + ], ), OptGroup( """ @@ -306,7 +358,13 @@ def Opt( earnings from income and expenses accrued during the current exercise into the balance sheet. This is most often called "Net Income". """, - [Opt("account_current_earnings", "Earnings:Current")], + [ + Opt( + "account_current_earnings", + "Earnings:Current", + converter=options_validate_leaf_account, + ) + ], ), # This common option can be used by reporting systems and is ignored by # Beancount itself. @@ -315,7 +373,13 @@ def Opt( Leaf name of the equity account used for inserting conversions that will zero out remaining amounts due to transfers during the exercise period. """, - [Opt("account_current_conversions", "Conversions:Current")], + [ + Opt( + "account_current_conversions", + "Conversions:Current", + converter=options_validate_leaf_account, + ) + ], ), # This common option can be used by reporting systems and is ignored by # Beancount itself. @@ -328,7 +392,14 @@ def Opt( balance on the sheet. This has no effect on behavior, other than providing a configurable account name for such postings to occur. """, - [Opt("account_unrealized_gains", "Earnings:Unrealized", "Earnings:Unrealized")], + [ + Opt( + "account_unrealized_gains", + "Earnings:Unrealized", + "Earnings:Unrealized", + converter=options_validate_leaf_account, + ) + ], ), OptGroup( """ @@ -337,7 +408,14 @@ def Opt( an account name will automatically enable the addition of postings on all transactions that have a residual amount. """, - [Opt("account_rounding", None, "Rounding")], + [ + Opt( + "account_rounding", + None, + "Rounding", + converter=options_validate_leaf_account, + ) + ], ), OptGroup( """	exception in any query using get_account_sort_key when localizing account names I noticed the "holdings" tab in my fava only shows me the exception quoted below, clicking on "Query" and boiling down a minimal working example blames `get_account_sort_key` in conjunction with renamed (top-level) accounts. exception: ~~~pytb $ bean-query mwe.beancount 'SELECT account, account_sortkey(account) ' Traceback (most recent call last): File "/usr/lib/python3.10/cmd.py", line 214, in onecmd func = getattr(self, 'do_' + cmd) AttributeError: 'BQLShell' object has no attribute 'do_SELECT' During handling of the above exception, another exception occurred: Traceback (most recent call last): File "/usr/lib/python3.10/site-packages/beancount/query/shell.py", line 271, in run_parser self.dispatch(statement) File "/usr/lib/python3.10/site-packages/beancount/query/shell.py", line 251, in dispatch return method(statement) File "/usr/lib/python3.10/site-packages/beancount/query/shell.py", line 416, in on_Select rtypes, rrows = query_execute.execute_query(c_query, File "/usr/lib/python3.10/site-packages/beancount/query/query_execute.py", line 282, in execute_query values = [c_expr(context) for c_expr in c_target_exprs] File "/usr/lib/python3.10/site-packages/beancount/query/query_execute.py", line 282, in <listcomp> values = [c_expr(context) for c_expr in c_target_exprs] File "/usr/lib/python3.10/site-packages/beancount/query/query_env.py", line 421, in __call__ index, name = account_types.get_account_sort_key(context.account_types, args[0]) File "/usr/lib/python3.10/site-packages/beancount/core/account_types.py", line 46, in get_account_sort_key return (account_types.index(get_account_type(account_name)), account_name) ValueError: tuple.index(x): x not in tuple ~~~ MWE: ~~~ledger option "title" "MWE" option "operating_currency" "EUR" 1992-01-01 custom "fava-option" "language" "de" * Bilanzkonten option "name_assets" "Aktiva" option "name_liabilities" "Verbindlichkeiten" option "name_equity" "Passiva:Eigenkapital" option "account_previous_balances" "Passiva:Eigenkapital:Eröffnungssaldo" option "account_current_earnings" "Gewinn" ; net income option "account_previous_earnings" "Rücklagen" ; retained earnings * Erfolgskonten option "name_income" "EK:Einnahmen" option "name_expenses" "EK:Ausgaben" 1992-01-01 open Passiva:Eigenkapital:Eröffnungssaldo 1992-01-01 open Aktiva:Giro 2000-01-01 * "" Aktiva:Giro 1000 EUR Passiva:Eigenkapital:Eröffnungssaldo ~~~	The issue is not the renamed root accounts, but the fact that the account names assigned to the root accounts are leaf accounts. Namely, these options are invalid: ``` option "name_equity" "Passiva:Eigenkapital" option "name_income" "EK:Einnahmen" option "name_expenses" "EK:Ausgaben" ``` There cannot be an account name separator `:` in the names of these accounts. It seems that the validator for these options does not catch this error. +1 We should probably detect that and issue an appropriate error. (Keep this ticket open until we do)	2024-11-09T16:19:42	[]	[]
beancount/beancount	beancount__beancount-836	c1259dca464cbf15f18b762f5e1933627217f10a	diff --git a/beancount/ops/balance.py b/beancount/ops/balance.py index 5287c9307..7c77ae6d0 100644 --- a/beancount/ops/balance.py +++ b/beancount/ops/balance.py @@ -108,7 +108,6 @@ def check(entries, options_map): # Check that the currency of the balance check is one of the allowed # currencies for that account. expected_amount = entry.amount - open = None try: open, _ = open_close_map[entry.account] except KeyError: @@ -119,6 +118,7 @@ def check(entries, options_map): entry, ) ) + continue if ( expected_amount is not None	Bug fix: keep entries when doing beancount.ops.balance check The check op was removing balance entries in the case where the account does not exist. I think it is better to keep the entry and it to fail validation, this makes it more consistent with other entry types.		2024-07-10T02:16:47	[]	[]
beancount/beancount	beancount__beancount-835	6bd2e7f37be6b5012634964e6a8093da5bce624f	diff --git a/beancount/ops/balance.py b/beancount/ops/balance.py index 7c77ae6d0..5287c9307 100644 --- a/beancount/ops/balance.py +++ b/beancount/ops/balance.py @@ -108,6 +108,7 @@ def check(entries, options_map): # Check that the currency of the balance check is one of the allowed # currencies for that account. expected_amount = entry.amount + open = None try: open, _ = open_close_map[entry.account] except KeyError: @@ -118,7 +119,6 @@ def check(entries, options_map): entry, ) ) - continue if ( expected_amount is not None	Bug fix: keep entries when doing beancount.ops.balance check The check op was removing balance entries in the case where the account does not exist. I think it is better to keep the entry and it to fail validation, this makes it more consistent with other entry types.		2024-07-05T20:14:23	[]	[]
SPF-OST/pytrnsys	SPF-OST__pytrnsys-117	677eb7afec5c8481ab49bd7c0160cb06d279a98a	diff --git a/pytrnsys/ddck/replaceVariables.py b/pytrnsys/ddck/replaceVariables.py index 939a0654..a57a340e 100644 --- a/pytrnsys/ddck/replaceVariables.py +++ b/pytrnsys/ddck/replaceVariables.py @@ -57,8 +57,7 @@ def replaceComputedVariablesWithDefaults(inputDdckFilePath: _pl.Path) -> _res.Re return _res.error(result) computedVariables = _res.value(result) - inputDdckContent = inputDdckFilePath.read_text() # pylint: disable=bad-option-value,unspecified-encoding - + inputDdckContent = inputDdckFilePath.read_text(encoding="windows-1252") # pylint: disable=bad-option-value outputDdckContent = inputDdckContent offset = 0 for computedVariable in computedVariables: diff --git a/pytrnsys/rsim/runParallelTrnsys.py b/pytrnsys/rsim/runParallelTrnsys.py index c114b68f..d0edca87 100644 --- a/pytrnsys/rsim/runParallelTrnsys.py +++ b/pytrnsys/rsim/runParallelTrnsys.py @@ -1,7 +1,7 @@ # pylint: skip-file # type: ignore -#!/usr/bin/python +# !/usr/bin/python """ Author : Dani Carbonell Date : 30.09.2016 @@ -12,7 +12,6 @@ import json import os import shutil -import typing as _tp from copy import deepcopy import numpy as num @@ -147,8 +146,6 @@ def readFromFolder(self, pathRun): return returnNames - # self.runFromNames(fileNames) - def readCasesToRun(self, pathRun, nameFileWithCasesToRun): fileToRunWithPath = os.path.join(pathRun, nameFileWithCasesToRun) @@ -172,7 +169,6 @@ def runFromNames(self, path, fileNames): tests.append(exeTrnsys.ExecuteTrnsys(path, fileNames[i])) tests[i].setTrnsysExePath(self.inputs["trnsysExePath"]) - # tests[i].setAddBuildingData(self.inputs["copyBuildingData"]) I comment it until we use again a case where we need this functionality. I guess we dont need this anymore. tests[i].loadDeck(check=self.inputs["checkDeck"]) tests[i].changeParameter(self.parameters) @@ -180,9 +176,6 @@ def runFromNames(self, path, fileNames): tests[i].ignoreOnlinePlotter() tests[i].setRemovePopUpWindow(self.inputs["removePopUpWindow"]) - # tests[i].copyFilesForRunning() - - # tests[i].setTrnsysVersion("TRNSYS17_EXE") self.cmds.append(tests[i].getExecuteTrnsys(self.inputs, useDeckName=tests[i].nameDck)) @@ -259,13 +252,6 @@ def createDecksFromVariant(self, fitParameters={}): parameters = self.parameters parameters.update(fitParameters) - # if (self.inputs["combineAllCases"]): - # sizeUsed = len(variations[0]) - # for var in variations: - # if (len(var) != sizeUsed): - # print "sizeUsed:%d var:%s size:%d" % (sizeUsed, var, len(var)) - # raise ValueError("FATAL ERROR. variations must be of the same size if combineAllCases = True") - myDeckGenerator = createDeck.CreateTrnsysDeck(self.path, self.nameBase, variations) myDeckGenerator.combineAllCases = self.inputs["combineAllCases"] @@ -288,7 +274,7 @@ def createDecksFromVariant(self, fitParameters={}): else: self.logger.info("Master file does not exist, no runs will be excluded") - # creates a list of decks with the appripiate name but nothing changed inside!! + # creates a list of decks with the appropriate name but nothing changed inside!! if self.variationsUsed or (self.changeDDckFilesUsed == True and self.foldersForDDckVariationUsed == False): if successfulCases: fileName = myDeckGenerator.generateDecks(successfulCases=successfulCases) @@ -303,7 +289,6 @@ def createDecksFromVariant(self, fitParameters={}): return tests = [] - cmds = [] variablePath = self.path @@ -311,22 +296,19 @@ def createDecksFromVariant(self, fitParameters={}): self.logger.debug("name to run :%s" % fileName[i]) - # if useLocationStructure: - # variablePath = os.path.join(path,location) #assign subfolder for path + # if useLocationStructure: + # variablePath = os.path.join(path,location) #assign subfolder for path - # # Parameters changed by variation + # Parameters changed by variation localCopyPar = dict.copy(parameters) # if self.variationsUsed: myParameters = myDeckGenerator.getParameters(i) localCopyPar.update(myParameters) - # # We add to the global parameters that also need to be modified - # If we assign like localCopyPar = parameters, then the parameters will change with localCopyPar !! + # We add to the global parameters that also need to be modified + # If we assign like localCopyPar = parameters, then the parameters will change with localCopyPar !! # Otherwise we change the global parameter and some values of last variation will remain. - # newPath = path + fileName[i] - # newFileDck = newPath+"\\"+fileName[i]+".dck" - # print "path:%s fileName:%s newPath:%s newFileDeck:%s"%(path,fileName[i],newPath,newFileDck) tests.append(exeTrnsys.ExecuteTrnsys(variablePath, fileName[i])) @@ -334,8 +316,6 @@ def createDecksFromVariant(self, fitParameters={}): tests[i].setRemovePopUpWindow(self.inputs["removePopUpWindow"]) - # tests[i].setTrnsysVersion("TRNSYS17_EXE") - tests[i].moveFileFromSource() tests[i].loadDeck(useDeckOutputPath=True) @@ -345,23 +325,12 @@ def createDecksFromVariant(self, fitParameters={}): if self.inputs["ignoreOnlinePlotter"] == True: tests[i].ignoreOnlinePlotter() - # ============================================================================== - # RESIZE PARAMETERS PIPE DIAMETER - # ============================================================================== - # tests[i].resizeParameters() tests[i].cleanAndCreateResultsTempFolder() tests[i].moveFileFromSource() if self.inputs["runCases"] == True: - test = os.path.split(tests[i].nameDck)[-1] self.cmds.append(tests[i].getExecuteTrnsys(self.inputs)) - # self.cmds = cmds - - # def checkTempFolderForFinishedSimulation(self,basePath): - # for file in os.listdir(os.path.join(basePath,'temp')): - # if file.endswith('.Prt') - def createLocationFolders(path, locations): for location in locations: if not os.path.exists(location): @@ -409,7 +378,7 @@ def buildTrnsysDeck(self) -> _res.Result[str]: deck.automaticEnegyBalanceStaff() deck.writeDeck() # Deck rewritten with added printer - # deck.addEnergyBalance() + deck.analyseDck() return deck.nameDeck @@ -429,7 +398,6 @@ def addParametricVariations(self, variations): """ if self.inputs["combineAllCases"] == True: - labels = [] values = [] for i, row in enumerate(variations): @@ -441,7 +409,6 @@ def addParametricVariations(self, variations): self.variablesOutput = result.tolist() else: - nVariations = len(variations) sizeOneVariation = len(variations[0]) - 2 for n in range(len(variations)): sizeCase = len(variations[n]) - 2 @@ -499,9 +466,6 @@ def writeRunLogFile(self): logfile.write(line + "\n") logfile.close() - # mainFile = sys.argv[0] - # shutil.copy(mainFile,os.path.join(self.path,'runMainFile.py')) - def readConfig(self, path, name, parseFileCreated=False): """ @@ -563,10 +527,9 @@ def changeDDckFile(self, source, end): nCharacters = len(source) for i in range(len(self.listDdck)): - # self.lines[i].replace(source,end) mySource = self.listDdck[i][ - -nCharacters: - ] # I read only the last characters with the same size as the end file + -nCharacters: + ] # I read only the last characters with the same size as the end file if mySource == source: newDDck = self.listDdck[i][0:-nCharacters] + end self.dictDdckPaths[newDDck] = self.dictDdckPaths[self.listDdck[i]] @@ -608,7 +571,6 @@ def getConfig(self): for line in self.lines: splitLine = line.split() - if splitLine[0] == "variation": variation = [] for i in range(len(splitLine)): @@ -625,7 +587,6 @@ def getConfig(self): self.variation.append(variation) elif splitLine[0] == "deck": - if splitLine[2] == "string": self.parameters[splitLine[1]] = splitLine[3] else: @@ -635,9 +596,7 @@ def getConfig(self): self.parameters[splitLine[1]] = splitLine[2] elif splitLine[0] == "replace": - splitString = line.split('$"') - oldString = splitString[1].split('"')[0] newString = splitString[2].split('"')[0] @@ -657,16 +616,6 @@ def getConfig(self): for i in range(len(splitLine)): if i > 0: self.foldersForDDckVariation.append(splitLine[i]) - - # elif (splitLine[0] == "") - # elif(splitLine[0] == "Relative"): - # - # self.listDdck.append(os.path.join(self.path, splitLine[1])) - # - # elif(splitLine[0] == "Absolute"): - # - # self.listDdck.append(splitLine[1]) - elif splitLine[0] == "fit": self.listFit[splitLine[1]] = [splitLine[2], splitLine[3], splitLine[4]] elif splitLine[0] == "case": @@ -680,7 +629,6 @@ def getConfig(self): self.listDdckPaths.add(self.inputs[splitLine[0]]) self.dictDdckPaths[fullPath] = self.inputs[splitLine[0]] else: - pass if len(self.variation) > 0: @@ -702,8 +650,6 @@ def getConfig(self): def copyConfigFile(self, configPath, configName): configFile = os.path.join(configPath, configName) - # dstPath = os.path.join(self.inputs["pathRef"],self.inputs["addResultsFolder"],self.inputs["nameRef"],configName) - # dstPath = os.path.join(self.inputs["pathRef"],self.inputs["addResultsFolder"],configName) dstPath = os.path.join(configPath, self.inputs["addResultsFolder"], configName) shutil.copyfile(configFile, dstPath) self.logger.debug("copied config file to: %s" % dstPath) @@ -737,17 +683,15 @@ def scaleVariables(self, reference, source, sink): for i in range(2, len(self.variablesOutput[j]), 1): if self.variablesOutput[j][1] == "sizeHpUsed": self.variablesOutput[j][i] = ( - str(round(self.unscaledVariables[j][i], 3)) + "" + str(round(loadHPsize, 3)) + str(round(self.unscaledVariables[j][i], 3)) + "" + str(round(loadHPsize, 3)) ) else: self.variablesOutput[j][i] = ( - str(round(self.unscaledVariables[j][i], 3)) + "" + str(round(loadDemand, 3)) + str(round(self.unscaledVariables[j][i], 3)) + "" + str(round(loadDemand, 3)) ) def run(): - pathBase = "" - if len(sys.argv) > 1: pathBase, configFile = os.path.split(sys.argv[1]) else: diff --git a/pytrnsys/trnsys_util/buildTrnsysDeck.py b/pytrnsys/trnsys_util/buildTrnsysDeck.py index daa73509..a70c3a8d 100644 --- a/pytrnsys/trnsys_util/buildTrnsysDeck.py +++ b/pytrnsys/trnsys_util/buildTrnsysDeck.py @@ -6,6 +6,7 @@ import logging import os import pathlib as _pl +import re as _re import tkinter as tk import typing as _tp from tkinter import messagebox as tkMessageBox @@ -31,7 +32,8 @@ def __init__(self, _pathDeck, _nameDeck, _nameList, ddckPlaceHolderValuesJsonPat self.pathDeck = _pathDeck self.nameDeck = self.pathDeck + "\%s.dck" % _nameDeck - self._ddckPlaceHolderValuesJsonPath = _pl.Path(ddckPlaceHolderValuesJsonPath) if ddckPlaceHolderValuesJsonPath else None + self._ddckPlaceHolderValuesJsonPath = _pl.Path( + ddckPlaceHolderValuesJsonPath) if ddckPlaceHolderValuesJsonPath else None self.oneSheetList = [] self.nameList = _nameList @@ -77,7 +79,8 @@ def _replacePlaceholdersAndGetContent(self, ddckFilePath: _pl.Path) -> _res.Resu if self._ddckPlaceHolderValuesJsonPath: if not self._ddckPlaceHolderValuesJsonPath.is_file(): - return _res.Error(f"The ddck placeholder values file at {self._ddckPlaceHolderValuesJsonPath} does not exist.") + return _res.Error( + f"The ddck placeholder values file at {self._ddckPlaceHolderValuesJsonPath} does not exist.") placeholderValues = _json.loads(self._ddckPlaceHolderValuesJsonPath.read_text()) @@ -163,7 +166,7 @@ def readDeckList( addedLines = firstThreeLines + self.linesChanged caption = ( - " ********************************************************************\n %s.ddck from %s \n ********************************************************************\n" + "******************************************************************\n %s.ddck from %s \n*********************************************************************\n" % (nameList, pathList) ) @@ -175,8 +178,11 @@ def readDeckList( unitModifiedLines = [line.replace("£", "") for line in addedLines] addedLines = unitModifiedLines - self.deckText.append(caption) + stream = _io.StringIO(caption) + lines = stream.readlines() + self.deckText.extend(lines) self.deckText = self.deckText + addedLines + self.logger = logging.getLogger("root") # stop propagting to root logger self.logger.propagate = False @@ -239,7 +245,7 @@ def writeDeck(self, addedLines=None): self.overwriteForcedByUser = True if ok: - tempFile = open(tempName, "w") + tempFile = open(tempName, "w", encoding='windows-1252') if addedLines != None: text = addedLines + self.deckText else: @@ -262,20 +268,12 @@ def readTrnsyDeck(self, useDeckName=False): useDeckName=useDeckName, eraseBeginComment=False, eliminateComments=False ) - # self.myDeck.loadDeckWithoutComments() - # self.linesDeckReaded = self.myDeck.linesReadedNoComments - def checkTrnsysDeck(self, nameDck, check=True): - - # self.readTrnsyDeck() - # deckUtils.checkEquationsAndConstants(self.linesDeckReaded) - lines = deckUtils.loadDeck(nameDck, eraseBeginComment=True, eliminateComments=True) if check: deckUtils.checkEquationsAndConstants(lines, self.nameDeck) self.linesDeckReaded = lines - # self.myDeck.checkEquationsAndConstants(self.deckText) #This does not need to read def saveUnitTypeFile(self): @@ -304,7 +302,6 @@ def writeTrnsysTypesUsed(self, name): line = "%s\tNone\t%s\n" % (self.filesUnitUsedInDdck[i], self.filesUsedInDdck[i]) else: line = "%s\t%s\t%s\n" % (self.filesUnitUsedInDdck[i], nameUnitFile, self.filesUsedInDdck[i]) - # line = "%s\t%s\t%s\n" % (self.filesUnitUsedInDdck[i],nameUnitFile[:-1],self.filesUsedInDdck[i]) lines = lines + line @@ -329,8 +326,6 @@ def automaticEnegyBalanceStaff(self): lines = deckUtils.addEnergyBalanceHourlyPrinter(unitId, eBalance) self.deckText = self.deckText[:-4] + lines + self.deckText[-4:] - self.writeDeck() # Deck rewritten with added printer - def replaceLines(self, replaceList): """ Replaces a deck lines with different lines @@ -348,4 +343,40 @@ def replaceLines(self, replaceList): for index, line in enumerate(self.linesChanged): if oldLine in line: self.linesChanged[index] = newLine + "\n" - changedLine = oldLine + + def analyseDck(self): + maxLineWidth = 0 + maxNumberOfConstantsInABlock = 0 + + constantsToCheck = ["UNIT", "EQUATIONS", "CONSTANTS", "PARAMETERS", "INPUTS"] + numOfTrnsysConstants = {} + + for index, line in enumerate(self.deckText): + maxLineWidth = max(maxLineWidth, len(line)) + if maxLineWidth > 1000: + self.logger.warning(f"Line {index + 1} has {maxLineWidth} characters which exceeds the limit.") + maxLineWidth = 0 + + for constant in constantsToCheck: + match = _re.search(fr"^\b{constant}\s\d+\b", line, _re.MULTILINE) + if match: + if constant == "UNIT": + numOfTrnsysConstants[constant] = numOfTrnsysConstants.get(constant, 0) + 1 + elif constant in ("CONSTANTS", "EQUATIONS"): + split = match.group().split() + numOfTrnsysConstants["EQUATIONS"] = numOfTrnsysConstants.get("EQUATIONS", 0) + int(split[1]) + maxNumberOfConstantsInABlock = max(maxNumberOfConstantsInABlock, int(split[1])) + else: + split = match.group().split() + numOfTrnsysConstants[constant] = numOfTrnsysConstants.get(constant, 0) + int(split[1]) + maxNumberOfConstantsInABlock = max(maxNumberOfConstantsInABlock, int(split[1])) + break + + for constant, number in numOfTrnsysConstants.items(): + if (constant == "UNIT" and number > 1000) or (constant == "EQUATIONS" and number > 500) or ( + constant == "PARAMETERS" and number > 2000) or (constant == "INPUTS" and number > 750): + self.logger.warning(f"There are {number} of {constant} which exceeds the limit") + + if maxNumberOfConstantsInABlock > 250: + self.logger.warning( + f"There are {maxNumberOfConstantsInABlock} of components in one block which exceeds the limit") diff --git a/pytrnsys/trnsys_util/deckTrnsys.py b/pytrnsys/trnsys_util/deckTrnsys.py index e26b4b7a..ab5f548b 100644 --- a/pytrnsys/trnsys_util/deckTrnsys.py +++ b/pytrnsys/trnsys_util/deckTrnsys.py @@ -426,7 +426,7 @@ def changeParameter(self, _parameters): logger.debug("variation deck file at %s" % self.nameDck) - outfile = open(self.nameDck, "w") + outfile = open(self.nameDck, "w", encoding='windows-1252') outfile.writelines(lines) outfile.close() diff --git a/pytrnsys/trnsys_util/deckUtils.py b/pytrnsys/trnsys_util/deckUtils.py index 3401af36..363d1650 100644 --- a/pytrnsys/trnsys_util/deckUtils.py +++ b/pytrnsys/trnsys_util/deckUtils.py @@ -308,7 +308,7 @@ def loadDeck(nameDck, eraseBeginComment=True, eliminateComments=True): list of lines obateined form the deck without the comments """ - infile = open(nameDck, "r") + infile = open(nameDck, "r", encoding='windows-1252') lines = infile.readlines()	Raise warning when default TRNSYS constants limits are exceeded Include pytrnsys_spf/utils/dckAnalyzer.py into the standard dckCheck and raise a warning when the standard TRNSYS constants values (e.g., maxFileWidth or nMaxCardValues) are exceeded.		2022-04-20T07:30:10	[]	[]
zauberzeug/rosys	zauberzeug__rosys-75	ff212ca21b70951e9e0b1a0ce9cbd799171e8db2	diff --git a/rosys/vision/calibration.py b/rosys/vision/calibration.py index bd1bda692..65ea4ffe4 100644 --- a/rosys/vision/calibration.py +++ b/rosys/vision/calibration.py @@ -2,7 +2,7 @@ import logging from dataclasses import dataclass, field -from typing import Optional +from typing import Optional, overload import cv2 import numpy as np @@ -49,59 +49,59 @@ def rotation(self) -> Rotation: def rotation_array(self) -> np.ndarray: return np.dot(self.extrinsics.rotation.R, self.intrinsics.rotation.R) - def project_to_image(self, world_point: Point3d) -> Point: - world_points = np.array([world_point.tuple], dtype=np.float32) - R = self.rotation_array - Rod = cv2.Rodrigues(R.T)[0] - t = -R.T @ self.extrinsics.translation - K = np.array(self.intrinsics.matrix) - D = np.array(self.intrinsics.distortion, dtype=float) - image_points, _ = cv2.projectPoints(world_points, Rod, t, K, D) - return Point(x=image_points[0, 0, 0], y=image_points[0, 0, 1]) + @overload + def project_to_image(self, world_coordinates: Point3d) -> Optional[Point]: ... + + @overload + def project_to_image(self, world_coordinates: np.ndarray) -> np.ndarray: ... + + def project_to_image(self, world_coordinates: Point3d \| np.ndarray) -> Optional[Point] \| np.ndarray: + if isinstance(world_coordinates, Point3d): + world_array = np.array([world_coordinates.tuple], dtype=np.float32) + image_array = self.project_to_image(world_array) + if np.isnan(image_array).any(): + return None + return Point(x=image_array[0, 0], y=image_array[0, 1]) # pylint: disable=unsubscriptable-object - def project_array_to_image(self, world_points: np.ndarray) -> np.ndarray: R = self.rotation_array Rod = cv2.Rodrigues(R.T)[0] t = -R.T @ self.extrinsics.translation K = np.array(self.intrinsics.matrix) D = np.array(self.intrinsics.distortion, dtype=float) - image_points, _ = cv2.projectPoints(world_points, Rod, t, K, D) - return image_points.reshape(-1, 2) + image_array, _ = cv2.projectPoints(world_coordinates, Rod, t, K, D) + local_coordinates = (world_coordinates - np.array(self.extrinsics.translation)) @ R + image_array[local_coordinates[:, 2] < 0, :] = np.nan + return image_array.reshape(-1, 2) - def project_from_image(self, image_point: Point, target_height: float = 0) -> Optional[Point3d]: - K = np.array(self.intrinsics.matrix) - D = np.array(self.intrinsics.distortion) - image_points = np.array(image_point.tuple, dtype=np.float32).reshape((1, 1, 2)) - image_points_ = cv2.undistortPoints(image_points, K, D).reshape(-1, 2) - image_points__ = cv2.convertPointsToHomogeneous(image_points_).reshape(-1, 3) - objPoints = image_points__ @ self.rotation_array.T - Z = self.extrinsics.translation[-1] - t = np.array(self.extrinsics.translation) - floor_points = t.T - objPoints * (Z - target_height) / objPoints[:, 2:] + @overload + def project_from_image(self, image_coordinates: Point, target_height: float = 0) -> Optional[Point3d]: ... - reprojection = self.project_to_image(Point3d(x=floor_points[0, 0], y=floor_points[0, 1], z=target_height)) - if objPoints[0, -1] * np.sign(Z) > 0 or reprojection.distance(image_point) > 2: - log.warning(f'reprojection failed with {reprojection.distance(image_point)} px') - return None + @overload + def project_from_image(self, image_coordinates: np.ndarray, target_height: float = 0) -> np.ndarray: ... - return Point3d(x=floor_points[0, 0], y=floor_points[0, 1], z=target_height) + def project_from_image(self, image_coordinates: Point \| np.ndarray, target_height: float = 0) -> Optional[Point3d] \| np.ndarray: + if isinstance(image_coordinates, Point): + image_points = np.array(image_coordinates.tuple, dtype=np.float32) + world_points = self.project_from_image(image_points, target_height=target_height) + if np.isnan(world_points).any(): + return None + return Point3d(x=world_points[0, 0], y=world_points[0, 1], z=world_points[0, 2]) # pylint: disable=unsubscriptable-object - def project_array_from_image(self, image_points: np.ndarray, target_height: float = 0) -> np.ndarray: K = np.array(self.intrinsics.matrix) D = np.array(self.intrinsics.distortion) - image_points_ = cv2.undistortPoints(image_points.astype(np.float32), K, D).reshape(-1, 2) - image_points__ = cv2.convertPointsToHomogeneous(image_points_).reshape(-1, 3) - objPoints = image_points__ @ self.rotation_array.T + image_coordinates_ = cv2.undistortPoints(image_coordinates.astype(np.float32), K, D).reshape(-1, 2) + image_coordinates__ = cv2.convertPointsToHomogeneous(image_coordinates_).reshape(-1, 3) + objPoints = image_coordinates__ @ self.rotation_array.T Z = self.extrinsics.translation[-1] t = np.array(self.extrinsics.translation) - floor_points = t.T - objPoints * (Z - target_height) / objPoints[:, 2:] + world_points = t.T - objPoints * (Z - target_height) / objPoints[:, 2:] - reprojection = self.project_array_to_image(floor_points) + reprojection = self.project_to_image(world_points) sign = objPoints[:, -1] * np.sign(Z) - distance = np.linalg.norm(reprojection - image_points, axis=1) - floor_points[np.logical_or(sign > 0, distance > 2), :] = np.nan + distance = np.linalg.norm(reprojection - image_coordinates, axis=1) + world_points[np.logical_not(np.logical_and(sign < 0, distance < 2)), :] = np.nan - return floor_points + return world_points @staticmethod def from_points( diff --git a/rosys/vision/camera_objects_.py b/rosys/vision/camera_objects_.py index 2d3d6ae5f..ef0dd1472 100644 --- a/rosys/vision/camera_objects_.py +++ b/rosys/vision/camera_objects_.py @@ -18,7 +18,7 @@ class CameraObjects(Group): With `debug=True` camera IDs are shown (default: `False`). """ - def __init__(self, camera_provider: CameraProvider, camera_projector: CameraProjector, , + def __init__(self, camera_provider: CameraProvider[CalibratableCamera], camera_projector: CameraProjector, , px_per_m: float = 10000, debug: bool = False) -> None: super().__init__() diff --git a/rosys/vision/camera_projector.py b/rosys/vision/camera_projector.py index 39e824d55..d60c76472 100644 --- a/rosys/vision/camera_projector.py +++ b/rosys/vision/camera_projector.py @@ -25,7 +25,7 @@ class CameraProjector: It is mainly used for visualization purposes. """ - def __init__(self, camera_provider: CameraProvider) -> None: + def __init__(self, camera_provider: CameraProvider[CalibratableCamera]) -> None: self.camera_provider = camera_provider self.projections: dict[str, Projection] = {} @@ -54,7 +54,7 @@ def project(camera: CalibratableCamera, rows: int = 12, columns: int = 16) -> Pr c, r = np.meshgrid(np.linspace(0, camera.calibration.intrinsics.size.width, columns), np.linspace(0, camera.calibration.intrinsics.size.height, rows)) image_points = np.stack((c.flatten(), r.flatten()), axis=1) - floor_points = camera.calibration.project_array_from_image(image_points).reshape(rows, columns, 3) + floor_points = camera.calibration.project_from_image(image_points).reshape(rows, columns, 3) invalid = np.isnan(floor_points).any(axis=2) return [ [ diff --git a/rosys/vision/detector_simulation.py b/rosys/vision/detector_simulation.py index 487b95802..024c77841 100644 --- a/rosys/vision/detector_simulation.py +++ b/rosys/vision/detector_simulation.py @@ -6,6 +6,7 @@ from .. import rosys from ..geometry import Point3d +from .camera import CalibratableCamera from .camera_provider import CameraProvider from .detections import BoxDetection, Detections, PointDetection from .detector import Detector @@ -34,7 +35,7 @@ class DetectorSimulation(Detector): An optional `noise` parameter controls the spatial accuracy in pixels. """ - def __init__(self, camera_provider: CameraProvider, , noise: float = 1.0, name: Optional[str] = None) -> None: + def __init__(self, camera_provider: CameraProvider[CalibratableCamera], , noise: float = 1.0, name: Optional[str] = None) -> None: super().__init__(name=name) self.camera_provider = camera_provider @@ -82,6 +83,8 @@ def detect_from_simulated_objects(self, image: Image) -> None: if np.dot(viewing_direction, object_direction) < 0: continue image_point = camera.calibration.project_to_image(obj.position) + if image_point is None: + continue if not (0 <= image_point.x < image.size.width and 0 <= image_point.y < image.size.height): continue @@ -101,7 +104,9 @@ def detect_from_simulated_objects(self, image: Image) -> None: for dy in [-obj.size[1] / 2, obj.size[1] / 2] for dz in [-obj.size[2] / 2, obj.size[2] / 2] ]) - image_points = camera.calibration.project_array_to_image(world_points) + image_points = camera.calibration.project_to_image(world_points) + if np.any(np.isnan(image_points)): + continue detections.boxes.append(BoxDetection( category_name=obj.category_name, model_name='simulation',	Projections from and to images are inconsistent for points behind camera In one direction the projection yields `None` or `np.nan` if a point is behind the camera. In the other direction the orientation is ignored, but should return `None` or `np.nan` as well. While we're at it, we can also unify the `project_*` methods, which are currently split for `Point` and `np.array`.		2024-01-18T13:24:56	[]	[]
carsales/pyheif	carsales__pyheif-55	38468c5963f7c7e40f482c71d9662f9c7412895f	diff --git a/pyheif/__init__.py b/pyheif/__init__.py index 3eda86e..108dbdb 100644 --- a/pyheif/__init__.py +++ b/pyheif/__init__.py @@ -1,3 +1,4 @@ +import builtins import os import _libheif_cffi @@ -7,7 +8,7 @@ from .writer import * version_path = os.path.dirname(os.path.abspath(__file__)) + "/data/version.txt" -with open(version_path) as f: +with builtins.open(version_path) as f: __version__ = f.read().strip() diff --git a/pyheif/reader.py b/pyheif/reader.py index 5e2a26f..aa2c1e1 100644 --- a/pyheif/reader.py +++ b/pyheif/reader.py @@ -1,3 +1,5 @@ +import builtins +import functools import pathlib import warnings @@ -8,17 +10,50 @@ class HeifFile: def __init__( - self, , size, data, metadata, color_profile, has_alpha, bit_depth, stride + self, , size, has_alpha, bit_depth, metadata, color_profile, data, stride ): self.size = size - self.data = data - self.metadata = metadata - self.color_profile = color_profile self.has_alpha = has_alpha self.mode = "RGBA" if has_alpha else "RGB" self.bit_depth = bit_depth + self.metadata = metadata + self.color_profile = color_profile + self.data = data self.stride = stride + def __repr__(self): + return ( + f"<{self.__class__.__name__} {self.size[0]}x{self.size[1]} {self.mode} " + f"with {str(len(self.data)) + ' bytes' if self.data else 'no'} data>" + ) + + def load(self): + return self # already loaded + + def close(self): + pass # TODO: release self.data here? + + +class UndecodedHeifFile(HeifFile): + def __init__( + self, heif_handle, , apply_transformations, convert_hdr_to_8bit, kwargs + ): + self._heif_handle = heif_handle + self.apply_transformations = apply_transformations + self.convert_hdr_to_8bit = convert_hdr_to_8bit + super().__init__(data=None, stride=None, kwargs) + + def load(self): + self.data, self.stride = _read_heif_image(self._heif_handle, self) + self.close() + self.__class__ = HeifFile + return self + + def close(self): + # Don't call super().close() here, we don't need to free bytes. + if hasattr(self, "_heif_handle"): + del self._heif_handle + def check(fp): d = _get_bytes(fp) @@ -33,14 +68,22 @@ def read_heif(fp, apply_transformations=True): def read(fp, , apply_transformations=True, convert_hdr_to_8bit=True): + heif_file = open( + fp, + apply_transformations=apply_transformations, + convert_hdr_to_8bit=convert_hdr_to_8bit, + ) + return heif_file.load() + + +def open(fp, , apply_transformations=True, convert_hdr_to_8bit=True): d = _get_bytes(fp) - result = _read_heif_bytes(d, apply_transformations, convert_hdr_to_8bit) - return result + return _read_heif_bytes(d, apply_transformations, convert_hdr_to_8bit) def _get_bytes(fp): if isinstance(fp, str): - with open(fp, "rb") as f: + with builtins.open(fp, "rb") as f: d = f.read() elif isinstance(fp, bytearray): d = bytes(fp) @@ -59,6 +102,19 @@ def _get_bytes(fp): return d +def _keep_refs(destructor, refs): + """ + Keep refs to passed arguments until `inner` callback exist. + This prevents collecting parent objects until all children collcted. + """ + + def inner(cdata): + return destructor(cdata) + + inner._refs = refs + return inner + + def _read_heif_bytes(d, apply_transformations, convert_hdr_to_8bit): magic = d[:12] filetype_check = _libheif_cffi.lib.heif_check_filetype(magic, len(magic)) @@ -68,11 +124,9 @@ def _read_heif_bytes(d, apply_transformations, convert_hdr_to_8bit): warnings.warn("Input is an unsupported HEIF/AVIF file type - trying anyway!") ctx = _libheif_cffi.lib.heif_context_alloc() - try: - result = _read_heif_context(ctx, d, apply_transformations, convert_hdr_to_8bit) - finally: - _libheif_cffi.lib.heif_context_free(ctx) - return result + collect = _keep_refs(_libheif_cffi.lib.heif_context_free, data=d) + ctx = _libheif_cffi.ffi.gc(ctx, collect, size=len(d)) + return _read_heif_context(ctx, d, apply_transformations, convert_hdr_to_8bit) def _read_heif_context(ctx, d, apply_transformations, convert_hdr_to_8bit): @@ -94,67 +148,29 @@ def _read_heif_context(ctx, d, apply_transformations, convert_hdr_to_8bit): subcode=error.subcode, message=_libheif_cffi.ffi.string(error.message).decode(), ) - handle = p_handle[0] - - try: - result = _read_heif_handle(handle, apply_transformations, convert_hdr_to_8bit) - finally: - _libheif_cffi.lib.heif_image_handle_release(handle) - return result + collect = _keep_refs(_libheif_cffi.lib.heif_image_handle_release, ctx=ctx) + handle = _libheif_cffi.ffi.gc(p_handle[0], collect) + return _read_heif_handle(handle, apply_transformations, convert_hdr_to_8bit) def _read_heif_handle(handle, apply_transformations, convert_hdr_to_8bit): width = _libheif_cffi.lib.heif_image_handle_get_width(handle) height = _libheif_cffi.lib.heif_image_handle_get_height(handle) - size = (width, height) - has_alpha = bool(_libheif_cffi.lib.heif_image_handle_has_alpha_channel(handle)) bit_depth = _libheif_cffi.lib.heif_image_handle_get_luma_bits_per_pixel(handle) - colorspace = _constants.heif_colorspace_RGB - if convert_hdr_to_8bit or bit_depth <= 8: - if has_alpha: - chroma = _constants.heif_chroma_interleaved_RGBA - else: - chroma = _constants.heif_chroma_interleaved_RGB - else: - if has_alpha: - chroma = _constants.heif_chroma_interleaved_RRGGBBAA_BE - else: - chroma = _constants.heif_chroma_interleaved_RRGGBB_BE - - p_options = _libheif_cffi.lib.heif_decoding_options_alloc() - p_options.ignore_transformations = int(not apply_transformations) - p_options.convert_hdr_to_8bit = int(convert_hdr_to_8bit) - - p_img = _libheif_cffi.ffi.new("struct heif_image ") - error = _libheif_cffi.lib.heif_decode_image( - handle, p_img, colorspace, chroma, p_options, - ) - _libheif_cffi.lib.heif_decoding_options_free(p_options) - if error.code != 0: - raise _error.HeifError( - code=error.code, - subcode=error.subcode, - message=_libheif_cffi.ffi.string(error.message).decode(), - ) - img = p_img[0] - - try: - data, stride = _read_heif_image(img, height) - finally: - _libheif_cffi.lib.heif_image_release(img) metadata = _read_metadata(handle) color_profile = _read_color_profile(handle) - heif_file = HeifFile( - size=size, - data=data, - metadata=metadata, - color_profile=color_profile, + heif_file = UndecodedHeifFile( + handle, + size=(width, height), has_alpha=has_alpha, bit_depth=bit_depth, - stride=stride, + metadata=metadata, + color_profile=color_profile, + apply_transformations=apply_transformations, + convert_hdr_to_8bit=convert_hdr_to_8bit, ) return heif_file @@ -225,15 +241,55 @@ def _read_color_profile(handle): return color_profile -def _read_heif_image(img, height): +def _read_heif_image(handle, heif_file): + colorspace = _constants.heif_colorspace_RGB + if heif_file.convert_hdr_to_8bit or heif_file.bit_depth <= 8: + if heif_file.has_alpha: + chroma = _constants.heif_chroma_interleaved_RGBA + else: + chroma = _constants.heif_chroma_interleaved_RGB + else: + if heif_file.has_alpha: + chroma = _constants.heif_chroma_interleaved_RRGGBBAA_BE + else: + chroma = _constants.heif_chroma_interleaved_RRGGBB_BE + + p_options = _libheif_cffi.lib.heif_decoding_options_alloc() + p_options = _libheif_cffi.ffi.gc( + p_options, _libheif_cffi.lib.heif_decoding_options_free + ) + p_options.ignore_transformations = int(not heif_file.apply_transformations) + p_options.convert_hdr_to_8bit = int(heif_file.convert_hdr_to_8bit) + + p_img = _libheif_cffi.ffi.new("struct heif_image ") + error = _libheif_cffi.lib.heif_decode_image( + handle, p_img, colorspace, chroma, p_options, + ) + if error.code != 0: + raise _error.HeifError( + code=error.code, + subcode=error.subcode, + message=_libheif_cffi.ffi.string(error.message).decode(), + ) + img = p_img[0] + p_stride = _libheif_cffi.ffi.new("int ") p_data = _libheif_cffi.lib.heif_image_get_plane_readonly( img, _constants.heif_channel_interleaved, p_stride ) stride = p_stride[0] - data_length = height * stride + data_length = heif_file.size[1] * stride + + # Release image as soon as no references to p_data left + collect = functools.partial(_release_heif_image, img) + p_data = _libheif_cffi.ffi.gc(p_data, collect, size=data_length) + + # ffi.buffer obligatory keeps a reference to p_data data_buffer = _libheif_cffi.ffi.buffer(p_data, data_length) - data = bytes(data_buffer) - return data, stride + return data_buffer, stride + + +def _release_heif_image(img, p_data=None): + _libheif_cffi.lib.heif_image_release(img)	Return ffi.buffer as data to avoid extra memory copying This saves 3-7% time depending on number of cores. master: ```python $ python -m timeit -n5 -r10 -s "import pyheif; data = open('./full.norot.heic', 'rb').read()" "im = pyheif.read(data)" 5 loops, best of 10: 345 msec per loop ``` avoid-copy: ```python $ python -m timeit -n5 -r10 -s "import pyheif; data = open('./full.norot.heic', 'rb').read()" "im = pyheif.read(data)" 5 loops, best of 10: 323 msec per loop ``` Also this makes `data` writable buffer, which avoid one more copy in such cases: master: ```python In [1]: import pyheif, numpy In [2]: im = pyheif.read('./full.norot.heic') In [3]: buf = numpy.frombuffer(im.data, dtype=numpy.uint8) In [4]: buf[0:3] = b'123' ValueError: assignment destination is read-only In [5]: buf = buf.copy() In [6]: buf[0:3] = b'123' # now this works ``` avoid-copy: ```python In [1]: import pyheif, numpy In [2]: im = pyheif.read('./full.norot.heic') In [3]: buf = numpy.frombuffer(im.data, dtype=numpy.uint8) In [4]: buf[0:3] = b'123' ``` Return ffi.buffer as data to avoid extra memory copying This saves 3-7% time depending on number of cores. master: ```python $ python -m timeit -n5 -r10 -s "import pyheif; data = open('./full.norot.heic', 'rb').read()" "im = pyheif.read(data)" 5 loops, best of 10: 345 msec per loop ``` avoid-copy: ```python $ python -m timeit -n5 -r10 -s "import pyheif; data = open('./full.norot.heic', 'rb').read()" "im = pyheif.read(data)" 5 loops, best of 10: 323 msec per loop ``` Also this makes `data` writable buffer, which avoid one more copy in such cases: master: ```python In [1]: import pyheif, numpy In [2]: im = pyheif.read('./full.norot.heic') In [3]: buf = numpy.frombuffer(im.data, dtype=numpy.uint8) In [4]: buf[0:3] = b'123' ValueError: assignment destination is read-only In [5]: buf = buf.copy() In [6]: buf[0:3] = b'123' # now this works ``` avoid-copy: ```python In [1]: import pyheif, numpy In [2]: im = pyheif.read('./full.norot.heic') In [3]: buf = numpy.frombuffer(im.data, dtype=numpy.uint8) In [4]: buf[0:3] = b'123' ```		2021-10-19T23:52:38	[]	[]
20tab/bvh-python	20tab__bvh-python-6	ed8c96359c5e8e85fca934808e84ef446cfae444	diff --git a/README.md b/README.md index 0a3324b..eebb0b7 100644 --- a/README.md +++ b/README.md @@ -72,3 +72,8 @@ Python module for parsing BVH (Biovision hierarchical data) mocap files >>> [str(node) for node in mocap.search('JOINT')] ['JOINT mixamorig:Spine', 'JOINT mixamorig:Spine1', 'JOINT mixamorig:Spine2', 'JOINT mixamorig:Neck', 'JOINT mixamorig:Head', 'JOINT mixamorig:HeadTop_End', 'JOINT mixamorig:LeftEye', 'JOINT mixamorig:RightEye', 'JOINT mixamorig:LeftShoulder', 'JOINT mixamorig:LeftArm', 'JOINT mixamorig:LeftForeArm', 'JOINT mixamorig:LeftHand', 'JOINT mixamorig:LeftHandMiddle1', 'JOINT mixamorig:LeftHandMiddle2', 'JOINT mixamorig:LeftHandMiddle3', 'JOINT mixamorig:LeftHandThumb1', 'JOINT mixamorig:LeftHandThumb2', 'JOINT mixamorig:LeftHandThumb3', 'JOINT mixamorig:LeftHandIndex1', 'JOINT mixamorig:LeftHandIndex2', 'JOINT mixamorig:LeftHandIndex3', 'JOINT mixamorig:LeftHandRing1', 'JOINT mixamorig:LeftHandRing2', 'JOINT mixamorig:LeftHandRing3', 'JOINT mixamorig:LeftHandPinky1', 'JOINT mixamorig:LeftHandPinky2', 'JOINT mixamorig:LeftHandPinky3', 'JOINT mixamorig:RightShoulder', 'JOINT mixamorig:RightArm', 'JOINT mixamorig:RightForeArm', 'JOINT mixamorig:RightHand', 'JOINT mixamorig:RightHandMiddle1', 'JOINT mixamorig:RightHandMiddle2', 'JOINT mixamorig:RightHandMiddle3', 'JOINT mixamorig:RightHandThumb1', 'JOINT mixamorig:RightHandThumb2', 'JOINT mixamorig:RightHandThumb3', 'JOINT mixamorig:RightHandIndex1', 'JOINT mixamorig:RightHandIndex2', 'JOINT mixamorig:RightHandIndex3', 'JOINT mixamorig:RightHandRing1', 'JOINT mixamorig:RightHandRing2', 'JOINT mixamorig:RightHandRing3', 'JOINT mixamorig:RightHandPinky1', 'JOINT mixamorig:RightHandPinky2', 'JOINT mixamorig:RightHandPinky3', 'JOINT mixamorig:RightUpLeg', 'JOINT mixamorig:RightLeg', 'JOINT mixamorig:RightFoot', 'JOINT mixamorig:RightToeBase', 'JOINT mixamorig:LeftUpLeg', 'JOINT mixamorig:LeftLeg', 'JOINT mixamorig:LeftFoot', 'JOINT mixamorig:LeftToeBase'] ``` +#### Get joint's direct children +```python +>>> mocap.joint_direct_children('mixamorig:Hips') +[JOINT mixamorig:Spine, JOINT mixamorig:RightUpLeg, JOINT mixamorig:LeftUpLeg] +``` diff --git a/bvh.py b/bvh.py index 08f22bd..40fc5a2 100644 --- a/bvh.py +++ b/bvh.py @@ -113,6 +113,10 @@ def iterate_joints(joint): iterate_joints(next(self.root.filter('ROOT'))) return joints + def joint_direct_children(self, name): + joint = self.get_joint(name) + return [child for child in joint.filter('JOINT')] + def get_joint_index(self, name): return self.get_joints().index(self.get_joint(name))	GetChildren method Currently it's not possible to get the children of a certain joint.		2019-07-23T15:04:08	[]	[]
ManuelNavarroGarcia/cpsplines	ManuelNavarroGarcia__cpsplines-45	3a956f89a0cc26faf0428ee0e4e388bb89de35f7	diff --git a/cpsplines/fittings/fit_cpsplines.py b/cpsplines/fittings/fit_cpsplines.py index 8361a7e..4089cc9 100644 --- a/cpsplines/fittings/fit_cpsplines.py +++ b/cpsplines/fittings/fit_cpsplines.py @@ -121,6 +121,9 @@ class CPsplines: The fitted decision variables of the B-spline expansion. data_arrangement : str The structure of the data. Must be either "gridded" or "scattered". + cat : Dict[int, str] + The mapping of the label encoder when dealing with binary data. Hence, + it is accesible when the binomial family is considered. References ---------- @@ -594,6 +597,13 @@ def fit( if self.sp_method not in ["grid_search", "optimizer"]: raise ValueError(f"Invalid `sp_method`: {self.sp_method}.") + if self.family.name == "binomial": + self.cat = dict(enumerate(data[y_col].astype("category").cat.categories)) + data = data.assign( + **{y_col: data[y_col].map({v: k for k, v in self.cat.items()})} + ) + if self.cat[1] != 1: + logging.warning(f"{self.cat[1]} is considered as the positive class.") x, y = self._preprocessor(data=data, y_col=y_col) # Construct the B-spline bases diff --git a/data/kyphosis.csv b/data/kyphosis.csv index d961b4a..cfe04d3 100644 --- a/data/kyphosis.csv +++ b/data/kyphosis.csv @@ -1,82 +1,82 @@ Kyphosis,Age,Number,Start -0,71,3,5 -0,158,3,14 -1,128,4,5 -0,2,5,1 -0,1,4,15 -0,1,2,16 -0,61,2,17 -0,37,3,16 -0,113,2,16 -1,59,6,12 -1,82,5,14 -0,148,3,16 -0,18,5,2 -0,1,4,12 -0,168,3,18 -0,1,3,16 -0,78,6,15 -0,175,5,13 -0,80,5,16 -0,27,4,9 -0,22,2,16 -1,105,6,5 -1,96,3,12 -0,131,2,3 -1,15,7,2 -0,9,5,13 -0,8,3,6 -0,100,3,14 -0,4,3,16 -0,151,2,16 -0,31,3,16 -0,125,2,11 -0,130,5,13 -0,112,3,16 -0,140,5,11 -0,93,3,16 -0,1,3,9 -1,52,5,6 -0,20,6,9 -1,91,5,12 -1,73,5,1 -0,35,3,13 -0,143,9,3 -0,61,4,1 -0,97,3,16 -1,139,3,10 -0,136,4,15 -0,131,5,13 -1,121,3,3 -0,177,2,14 -0,68,5,10 -0,9,2,17 -1,139,10,6 -0,2,2,17 -0,140,4,15 -0,72,5,15 -0,2,3,13 -1,120,5,8 -0,51,7,9 -0,102,3,13 -1,130,4,1 -1,114,7,8 -0,81,4,1 -0,118,3,16 -0,118,4,16 -0,17,4,10 -0,195,2,17 -0,159,4,13 -0,18,4,11 -0,15,5,16 -0,158,5,14 -0,127,4,12 -0,87,4,16 -0,206,4,10 -0,11,3,15 -0,178,4,15 -1,157,3,13 -0,26,7,13 -0,120,2,13 -1,42,7,6 -0,36,4,13 +Absent,71,3,5 +Absent,158,3,14 +Present,128,4,5 +Absent,2,5,1 +Absent,1,4,15 +Absent,1,2,16 +Absent,61,2,17 +Absent,37,3,16 +Absent,113,2,16 +Present,59,6,12 +Present,82,5,14 +Absent,148,3,16 +Absent,18,5,2 +Absent,1,4,12 +Absent,168,3,18 +Absent,1,3,16 +Absent,78,6,15 +Absent,175,5,13 +Absent,80,5,16 +Absent,27,4,9 +Absent,22,2,16 +Present,105,6,5 +Present,96,3,12 +Absent,131,2,3 +Present,15,7,2 +Absent,9,5,13 +Absent,8,3,6 +Absent,100,3,14 +Absent,4,3,16 +Absent,151,2,16 +Absent,31,3,16 +Absent,125,2,11 +Absent,130,5,13 +Absent,112,3,16 +Absent,140,5,11 +Absent,93,3,16 +Absent,1,3,9 +Present,52,5,6 +Absent,20,6,9 +Present,91,5,12 +Present,73,5,1 +Absent,35,3,13 +Absent,143,9,3 +Absent,61,4,1 +Absent,97,3,16 +Present,139,3,10 +Absent,136,4,15 +Absent,131,5,13 +Present,121,3,3 +Absent,177,2,14 +Absent,68,5,10 +Absent,9,2,17 +Present,139,10,6 +Absent,2,2,17 +Absent,140,4,15 +Absent,72,5,15 +Absent,2,3,13 +Present,120,5,8 +Absent,51,7,9 +Absent,102,3,13 +Present,130,4,1 +Present,114,7,8 +Absent,81,4,1 +Absent,118,3,16 +Absent,118,4,16 +Absent,17,4,10 +Absent,195,2,17 +Absent,159,4,13 +Absent,18,4,11 +Absent,15,5,16 +Absent,158,5,14 +Absent,127,4,12 +Absent,87,4,16 +Absent,206,4,10 +Absent,11,3,15 +Absent,178,4,15 +Present,157,3,13 +Absent,26,7,13 +Absent,120,2,13 +Present,42,7,6 +Absent,36,4,13 diff --git a/examples/non_gaussian_data.ipynb b/examples/non_gaussian_data.ipynb index af06b72..72a0fa6 100644 --- a/examples/non_gaussian_data.ipynb +++ b/examples/non_gaussian_data.ipynb @@ -64,7 +64,7 @@ "name": "stdout", "output_type": "stream", "text": [ - "[Solve the problem with smoothing parameters (444.5186255273247,): ] Elapsed time (s): 0.069853\n" + "[Solve the problem with smoothing parameters (444.5186255273247,): ] Elapsed time (s): 0.115425\n" ] } ], @@ -125,7 +125,7 @@ "name": "stdout", "output_type": "stream", "text": [ - "[Solve the problem with smoothing parameters (443.3914121205766,): ] Elapsed time (s): 0.126242\n" + "[Solve the problem with smoothing parameters (443.3914121205766,): ] Elapsed time (s): 0.098054\n" ] } ], @@ -203,7 +203,7 @@ "name": "stdout", "output_type": "stream", "text": [ - "[Solve the problem with smoothing parameters (0.1, 0.1): ] Elapsed time (s): 8.514172\n" + "[Solve the problem with smoothing parameters (0.1, 0.1): ] Elapsed time (s): 8.427936\n" ] } ], @@ -295,11 +295,18 @@ "execution_count": 12, "metadata": {}, "outputs": [ + { + "name": "stderr", + "output_type": "stream", + "text": [ + "WARNING:root:Present is considered as the positive class.\n" + ] + }, { "name": "stdout", "output_type": "stream", "text": [ - "[Solve the problem with smoothing parameters (1,): ] Elapsed time (s): 0.021814\n" + "[Solve the problem with smoothing parameters (1,): ] Elapsed time (s): 0.015729\n" ] } ], @@ -320,7 +327,7 @@ "outputs": [ { "data": { - "image/png": "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", + "image/png": "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", "text/plain": [ "<Figure size 1080x720 with 1 Axes>" ] @@ -357,11 +364,18 @@ "execution_count": 14, "metadata": {}, "outputs": [ + { + "name": "stderr", + "output_type": "stream", + "text": [ + "WARNING:root:Present is considered as the positive class.\n" + ] + }, { "name": "stdout", "output_type": "stream", "text": [ - "[Solve the problem with smoothing parameters (1, 1): ] Elapsed time (s): 0.087073\n" + "[Solve the problem with smoothing parameters (1, 1): ] Elapsed time (s): 0.072660\n" ] } ], @@ -484,7 +498,7 @@ "name": "stdout", "output_type": "stream", "text": [ - "[Solve the problem with smoothing parameters (1, 1): ] Elapsed time (s): 0.166115\n" + "[Solve the problem with smoothing parameters (1, 1): ] Elapsed time (s): 0.113034\n" ] } ], @@ -685,7 +699,7 @@ "name": "stdout", "output_type": "stream", "text": [ - "[Solve the problem with smoothing parameters (0.0001, 10): ] Elapsed time (s): 0.193003\n" + "[Solve the problem with smoothing parameters (0.0001, 10): ] Elapsed time (s): 0.221361\n" ] } ], diff --git a/setup.py b/setup.py index 64692f1..2ec15db 100644 --- a/setup.py +++ b/setup.py @@ -33,8 +33,8 @@ "numpy", "pandas", "scipy", - "tensorly", "statsmodels", + "tensorly", ], extras_require={"dev": ["black", "ipykernel", "mypy", "pip-tools", "pytest"]}, )	`LabelEncoder` for binomial regression When facing binomial regression, the output needs to be 0 or 1 for the fitting procedure to work. This should be done internally, so the user can pass an arbitrary binary category.		2022-12-10T18:59:22	[]	[]
ManuelNavarroGarcia/cpsplines	ManuelNavarroGarcia__cpsplines-12	46470464da46cb9fd4a9e2cd35907cc94cdf530b	diff --git a/README.md b/README.md index a974ebf..d2a48a5 100644 --- a/README.md +++ b/README.md @@ -1,49 +1,93 @@ # cpsplines -`cpsplines` is a Python module to perform constrained regression under shape constraints on the component functions of the dependent variable. It is assumed that the smooth hypersurface to be estimated is defined through a reduced-rank basis (B−splines) and fitted via a penalized splines approach (P−splines). To embed requirements about the sign of any order partial derivative (including the function itself), the constraints are included in the fitting process as hard constraints, yielding a semidefinite optimization model. In particular, the problem of estimating the component function using this approach is stated as a convex semidefinite optimization problem with a quadratic objective function, which can be easily reformulated as a conic optimization problem. - -Sign related constraints are imposed using a well-known result carried out by Bertsimas and Popescu, 2002. This enables to enforce non-negativity of a univariate polynomial over a finite interval, which can be straightforwardly extended to the sign of any higher order derivative. When only one covariate is related to the response variable, these constraints are successfully fulfilled over the whole domain of the regressor sample. However, when facing multiple regression, this equivalence does not hold, so alternative approaches must be developed. The proposed framework in this repository uses the equivalence relation for univariate polynomials by imposing the constraints over a finite set of curves which belong to the hypersurface. - -At present, `cpsplines` can handle constrained regression problems for data lying on grids. In this setting, the smooth hypersurface is constructed from the tensor products of B-splines basis along each axis, which allows to develop efficient algorithms accelerating the computations (Currie, Durban and Eilers, 2006). On this repository, the fitting procedure is performed using the method `GridCPsplines`, whose main features are the following: +`cpsplines` is a Python module to perform constrained regression under shape +constraints on the component functions of the dependent variable. It is assumed +that the smooth hypersurface to be estimated is defined through a reduced-rank +basis (B−splines) and fitted via a penalized splines approach (P−splines). To +embed requirements about the sign of any order partial derivative (including the +function itself), the constraints are included in the fitting process as hard +constraints, yielding a semidefinite optimization model. In particular, the +problem of estimating the component function using this approach is stated as a +convex semidefinite optimization problem with a quadratic objective function, +which can be easily reformulated as a conic optimization problem. + +Sign related constraints are imposed using a well-known result carried out by +Bertsimas and Popescu, 2002. This enables to enforce non-negativity of a +univariate polynomial over a finite interval, which can be straightforwardly +extended to the sign of any higher order derivative. When only one covariate is +related to the response variable, these constraints are successfully fulfilled +over the whole domain of the regressor sample. However, when facing multiple +regression, this equivalence does not hold, so alternative approaches must be +developed. The proposed framework in this repository uses the equivalence +relation for univariate polynomials by imposing the constraints over a finite +set of curves which belong to the hypersurface. + +At present, `cpsplines` can handle constrained regression problems for data +lying on grids. In this setting, the smooth hypersurface is constructed from the +tensor products of B-splines basis along each axis, which allows to develop +efficient algorithms accelerating the computations (Currie, Durban and Eilers, +2006). On this repository, the fitting procedure is performed using the method +`GridCPsplines`, whose main features are the following: * Arbitrary number of variables. -* Arbitrary knot sequence length to construct the B-spline basis. -* Arbitrary B-spline basis degrees. +* Arbitrary knot sequence length to construct the B-spline basis. +* Arbitrary B-spline basis degrees. * Arbitrary difference orders on the penalty term. -* Out-of-range prediction (backwards and forward) along every dimension (Currie and Durban, 2004), and the constraints are enforced both on the fitting and the prediction region. -* The smoothing parameters are selected as the minimizer of the Generalized Cross Validation criteria, but this routine can be done either by choosing the best parameters out of a set of candidates or by finding them using numerical methods. -* Enforcing sign related constraints over the fitting and prediction range (if prediction is required). Arbitrary number of sign constraints can be imposed along each regressor. -* Enforcing the hypersurface (or any partial derivative) attains a certain value at a certain point. - -Solving the optimization problems is done using [MOSEK](https://www.mosek.com) optimization software. +* Out-of-range prediction (backwards and forward) along every dimension (Currie + and Durban, 2004), and the constraints are enforced both on the fitting and + the prediction region. +* The smoothing parameters are selected as the minimizer of the Generalized + Cross Validation criteria, but this routine can be done either by choosing the + best parameters out of a set of candidates or by finding them using numerical + methods. +* Enforcing sign related constraints over the fitting and prediction range (if + prediction is required). Arbitrary number of sign constraints can be imposed + along each regressor. +* Enforcing the hypersurface (or any partial derivative) attains a certain value + at a certain point. +* Enforcing the hypersurface is a probability density function, i.e., it is + non-negative and it integrates to one. + +Solving the optimization problems is done using [MOSEK](https://www.mosek.com) +optimization software. ## Project structure The current version of the project is structured as follows: * cpsplines: the main directory of the project, which consist of: - * fittings: contains the smoothing algorithms. + * fittings: contains the smoothing algorithms. * graphics: constituted by graphic methods to visualize the results. - * mosek_functions: contains the functions used to define the optimization problems. - * psplines: composed by several methods that build the main objects of P-splines. - * utils: integrated by a miscellanea of files used for a variety of purposes (numerical computations, data processing, ...) -* data: a folder containing CSV files used in the real data numerical experiments. -* examples: a directory containing multiple numerical experiments, using both synthetic and real data sets. -* tests: a folder including tests for the main methods of the project. - + * mosek_functions: contains the functions used to define the + optimization problems. + * psplines: composed by several methods that build the main objects of + P-splines. + * utils: integrated by a miscellanea of files used for a variety of + purposes (numerical computations, data processing, ...) +* data: a folder containing CSV files used in the real data numerical + experiments. +* examples: a directory containing multiple numerical experiments, using + both synthetic and real data sets. +* tests: a folder including tests for the main methods of the project. ## Package dependencies `cpsplines` mainly depends on the following packages: * [Joblib](https://joblib.readthedocs.io/). -* [Matplotlib](https://matplotlib.org/). +* [Matplotlib](https://matplotlib.org/). * [MOSEK](https://www.mosek.com). License Required * [Numpy](https://numpy.org/). * [Pandas](https://pandas.pydata.org/). * [Scipy](https://www.scipy.org/). * [Tensorly](http://tensorly.org/). -MOSEK requires a license to be used. For research or educational purposes, a free yearly and renewable [academic license](https://www.mosek.com/products/academic-licenses/) is offered by the company. For other cases, a 30-day [trial license](https://www.mosek.com/try/) is available. According to MOSEK indications, the license file (`mosek.lic`) must be located at +MOSEK requires a license to be used. For research or educational purposes, a +free yearly and renewable [academic +license](https://www.mosek.com/products/academic-licenses/) is offered by the +company. For other cases, a 30-day [trial license](https://www.mosek.com/try/) +is available. According to MOSEK indications, the license file (`mosek.lic`) +must be located at + ``` $HOME/mosek/mosek.lic (Linux/OSX) %USERPROFILE%\mosek\mosek.lic (Windows) @@ -51,14 +95,15 @@ $HOME/mosek/mosek.lic (Linux/OSX) ## Installation -1. To clone the repository on your own device, use +1. To clone the repository on your own device, use ``` git clone https://github.com/ManuelNavarroGarcia/cpsplines.git cd cpsplines ``` -2. To install the dependencies, there are two options according to your installation preferences: +2. To install the dependencies, there are two options according to your + installation preferences: * Create and activate a virtual environment with `conda` (recommended) @@ -74,7 +119,8 @@ pip install -r requirements.txt pip install -e .[dev] ``` -3. If neccessary, add version requirements to existing dependencies or add new ones on `setup.py`. Then, update `requirements.txt` file using +3. If neccessary, add version requirements to existing dependencies or add new + ones on `setup.py`. Then, update `requirements.txt` file using ``` pip-compile --extra dev > requirements.txt @@ -82,11 +128,13 @@ pip-compile --extra dev > requirements.txt and update the environment with `pip-sync`. -## Usage +## Usage -To illustrate the usage of the repository, let's see how `GridCPsplines` works with two examples, the first with only one regressor and the second two covariates. +To illustrate the usage of the repository, let's see how `GridCPsplines` works +with two examples, the first with only one regressor and the second two +covariates. -For the univariate case, consider the function +For the univariate case, consider the function <p align="center"> <img src="https://latex.codecogs.com/svg.image?f(x)&space;=&space;(2x-1)&space;^3," title="f(x) = (2x-1) ^3," /> @@ -100,7 +148,11 @@ which is a non-decreasing function. We simulated noisy data following the scheme <img src="https://render.githubusercontent.com/render/math?math=$\text{where%20}\:x_l=0,%200.02,%200.04,\ldots,%201.$"> -We fit an unconstrained and a constrained model imposing non-decreasing constraints over the interval [-0.15, 1.12] (forward and backwards prediction). For the basis, cubic B-splines with 10 interior knots are taken with a second-order difference penalty. The smoothing parameter is selected using `scipy.optimize.minimize` with the `"SLSQP"` method. +We fit an unconstrained and a constrained model imposing non-decreasing +constraints over the interval [-0.15, 1.12] (forward and backwards prediction). +For the basis, cubic B-splines with 10 interior knots are taken with a +second-order difference penalty. The smoothing parameter is selected using +`scipy.optimize.minimize` with the `"SLSQP"` method. ```python # Generate the data @@ -146,7 +198,7 @@ curves = plot_curves( ![alt text](./img/example1D.jpg) -For the bivariate case, consider the function +For the bivariate case, consider the function <p align="center"> <img src="https://latex.codecogs.com/svg.image?f(x,&space;y)&space;=&space;\sin(x)\sin(y)." title="f(x, y) = \sin(x)\sin(y)." /> </p> @@ -157,13 +209,18 @@ We simulated noisy data following the scheme <img src="https://latex.codecogs.com/svg.image?z_{lm}=&space;f(x_l,&space;y_m)+\varepsilon_{lm},&space;\quad&space;\varepsilon_{lm}\sim&space;\text{N}(0,1)" title="z_{lm}= f(x_l, y_m)+\varepsilon_{lm}, \quad \varepsilon_{lm}\sim \text{N}(0,1)." /> </p> -We fit an unconstrained and a constrained model imposing non-negativity constraints over the interval +We fit an unconstrained and a constrained model imposing non-negativity +constraints over the interval <p align="center"> <img src="https://latex.codecogs.com/svg.image?[0,3\pi]\times&space;[0,2\pi]" title="[0,3\pi]\times [0,2\pi]" /> </p> -(no prediction). For the bases, cubic B-splines with 30 and 20 interior knots are taken, respectively, with a second-order difference penalty. The smoothing parameter is selected as the best candidates out of the sets {10, 100} (for the first smoothing parameter) and {10, 50, 100} (for the second smoothing parameter). +(no prediction). For the bases, cubic B-splines with 30 and 20 interior knots +are taken, respectively, with a second-order difference penalty. The smoothing +parameter is selected as the best candidates out of the sets {10, 100} (for the +first smoothing parameter) and {10, 50, 100} (for the second smoothing +parameter). ```python # Generate the data @@ -194,35 +251,58 @@ surface = plot_surfaces( ## Testing -The repository contains a folder with unit tests to guarantee the main methods meets their design and behave as intended. To launch the test suite, it is enough to enter `pytest`. If only one test file wants to be run, the syntax is given by +The repository contains a folder with unit tests to guarantee the main methods +meets their design and behave as intended. To launch the test suite, it is +enough to enter `pytest`. If only one test file wants to be run, the syntax is +given by ``` pytest tests/test_<file_name>.py ``` -Moreover, a GitHub Action runs automatically all the tests but `tests/test_solution.py` (which requires MOSEK license) when any commit is pushed on any Pull Request. - +Moreover, a GitHub Action runs automatically all the tests but +`tests/test_solution.py` (which requires MOSEK license) when any commit is +pushed on any Pull Request. ## Contributing -Contributions to the repository are welcomed! Regardless of whether it is a small fix on the documentation or a notable feature to be included, I encourage you to develop your ideas and make this project greater. Even suggestions about the code structure are highly appreciated. Furthermore, users participating on these submissions will figure as contributors on this main page of the repository. +Contributions to the repository are welcomed! Regardless of whether it is a +small fix on the documentation or a notable feature to be included, I encourage +you to develop your ideas and make this project greater. Even suggestions about +the code structure are highly appreciated. Furthermore, users participating on +these submissions will figure as contributors on this main page of the +repository. There are many ways you can contribute on this repository: -* [Discussions](https://github.com/ManuelNavarroGarcia/cpsplines/discussions). To ask questions you are wondering about or share ideas, you can enter an existing discussion or open a new one. +* [Discussions](https://github.com/ManuelNavarroGarcia/cpsplines/discussions). + To ask questions you are wondering about or share ideas, you can enter an + existing discussion or open a new one. -* [Issues](https://github.com/ManuelNavarroGarcia/cpsplines/issues). If you detect a bug or you want to propose an enhancement of the current version of the code, a issue with reproducible code and/or a detailed description is highly appreciated. +* [Issues](https://github.com/ManuelNavarroGarcia/cpsplines/issues). If you + detect a bug or you want to propose an enhancement of the current version of + the code, a issue with reproducible code and/or a detailed description is + highly appreciated. -* [Pull Requests](https://github.com/ManuelNavarroGarcia/cpsplines/pulls). If you feel I am missing an important feature, either in the code or in the documentation, I encourage you to start a pull request developing this idea. Nevertheless, before starting any major new feature work, I suggest you to open an issue or start a discussion describing what you are planning to do. I note that, before starting a pull request, all unit test must pass on your local repository. +* [Pull Requests](https://github.com/ManuelNavarroGarcia/cpsplines/pulls). If + you feel I am missing an important feature, either in the code or in the + documentation, I encourage you to start a pull request developing this idea. + Nevertheless, before starting any major new feature work, I suggest you to + open an issue or start a discussion describing what you are planning to do. I + note that, before starting a pull request, all unit test must pass on your + local repository. ## Contact Information and Citation -If you have encountered any problem or doubt while using `cpsplines`, please feel free to let me know by sending me an email: +If you have encountered any problem or doubt while using `cpsplines`, please +feel free to let me know by sending me an email: * Name: Manuel Navarro García (he/his) * Email: [email protected] -The formal background of the models used in this project are either published in research paper or under current research. If these techniques are helpful to your own research, consider citing the related papers of the project: +The formal background of the models used in this project are either published in +research paper or under current research. If these techniques are helpful to +your own research, consider citing the related papers of the project: ``` @TECHREPORT{navarro2020, @@ -236,9 +316,19 @@ The formal background of the models used in this project are either published in ## Acknowledgements -Throughout the developing of this project I have received strong support from various individuals. - -I would first like to thank my supervisors, Professor [Vanesa Guerrero](https://github.com/vanesaguerrero) and Professor [María Durbán](https://github.com/MariaDurban), whose insightful comments and invaluable expertise has given way to many of the current functionalities of the repository. - -I would also like to acknowledge the [Komorebi AI](https://komorebi.ai/) team for their assistance and guidance on the technical part of the project. Specially, I would like to thank [Alberto Torres](https://github.com/albertotb), [David Gordo](https://github.com/davidggphy) and [Victor Gallego](https://komorebi.ai/) for their constructive code structure suggestions that have helped notably to improve the computational efficiency and the usage of the algorithms. - +Throughout the developing of this project I have received strong support from +various individuals. + +I would first like to thank my supervisors, Professor [Vanesa +Guerrero](https://github.com/vanesaguerrero) and Professor [María +Durbán](https://github.com/MariaDurban), whose insightful comments and +invaluable expertise has given way to many of the current functionalities of the +repository. + +I would also like to acknowledge the [Komorebi AI](https://komorebi.ai/) team +for their assistance and guidance on the technical part of the project. +Specially, I would like to thank [Alberto Torres](https://github.com/albertotb), +[David Gordo](https://github.com/davidggphy) and [Victor +Gallego](https://komorebi.ai/) for their constructive code structure suggestions +that have helped notably to improve the computational efficiency and the usage +of the algorithms. diff --git a/cpsplines/fittings/grid_cpsplines.py b/cpsplines/fittings/grid_cpsplines.py index b5680e3..aa8d5a6 100644 --- a/cpsplines/fittings/grid_cpsplines.py +++ b/cpsplines/fittings/grid_cpsplines.py @@ -7,6 +7,7 @@ import scipy from cpsplines.mosek_functions.interval_constraints import IntConstraints from cpsplines.mosek_functions.obj_function import ObjectiveFunction +from cpsplines.mosek_functions.pdf_constraints import PDFConstraint from cpsplines.mosek_functions.point_constraints import PointConstraints from cpsplines.psplines.bspline_basis import BsplineBasis from cpsplines.psplines.penalty_matrix import PenaltyMatrix @@ -100,6 +101,10 @@ class GridCPsplines: where the value needs to be fixed. - An array with the values of the derivative to be enforced. - A number corresponding to the tolerancea allowed in the constraint. + pdf_constraint : bool, optional + A boolean indicating whether the fitted hypersurface must satisfy + Probability Density Function (PDF) conditions, i.e., it is non-negative + and it integrates to one. Attributes ---------- @@ -131,6 +136,7 @@ def __init__( Dict[int, Dict[int, Dict[str, Union[int, float]]]] ] = None, pt_constraints: Optional[Dict[Tuple[int], Any]] = None, + pdf_constraint: bool = False, ): self.deg = deg self.ord_d = ord_d @@ -140,6 +146,7 @@ def __init__( self.sp_args = sp_args self.int_constraints = int_constraints self.pt_constraints = pt_constraints + self.pdf_constraint = pdf_constraint def _get_bspline_bases(self, x: Iterable[np.ndarray]) -> List[BsplineBasis]: @@ -180,6 +187,8 @@ def _get_bspline_bases(self, x: Iterable[np.ndarray]) -> List[BsplineBasis]: ) # Generate the design matrix of the B-spline basis bsp.get_matrix_B() + if self.int_constraints is not None or self.pdf_constraint: + bsp.get_matrices_S() bspline_bases.append(bsp) return bspline_bases @@ -296,11 +305,17 @@ def _initialize_model( lin_term=lin_term, ) + if self.pdf_constraint: + pdf_cons = PDFConstraint(bspline=self.bspline_bases) + # Incorporate the condition that the integral over all the space + # must equal to 1 + pdf_cons.integrate_to_one(var_dict=mos_obj_f.var_dict, model=M) + # Enforce the non-negativity constraint if it is not imposed + # explicitly + self.int_constraints = pdf_cons.nonneg_cons(self.int_constraints) + if self.int_constraints is not None: - matrices_S = {} - # If the are interval constraints, construct the matrices S - for i, bsp in enumerate(self.bspline_bases): - matrices_S[i] = bsp.get_matrices_S() + matrices_S = {i: bsp.matrices_S for i, bsp in enumerate(self.bspline_bases)} # Iterate for every variable with constraints and for every # derivative order for var_name in self.int_constraints.keys(): @@ -333,7 +348,6 @@ def _initialize_model( cons2.point_cons(var_dict=mos_obj_f.var_dict, model=M) else: self.pt_constraints = {} - return M def _get_sp_grid_search( diff --git a/cpsplines/mosek_functions/interval_constraints.py b/cpsplines/mosek_functions/interval_constraints.py index f59bd71..a33c792 100644 --- a/cpsplines/mosek_functions/interval_constraints.py +++ b/cpsplines/mosek_functions/interval_constraints.py @@ -12,7 +12,7 @@ class IntConstraints: """ Define the set of constraints that ensures the non-negativity (or - non-positivity) for the derivative order `derivative`along the respective + non-positivity) for the derivative order `derivative` along the respective axis of the variable `var_name`. For the unidimensional case, the requirements are completely fulfilled for every point in the range where the B-spline basis is defined. For the multivariate case, the constraints are diff --git a/cpsplines/mosek_functions/pdf_constraints.py b/cpsplines/mosek_functions/pdf_constraints.py new file mode 100644 index 0000000..7b2843a --- /dev/null +++ b/cpsplines/mosek_functions/pdf_constraints.py @@ -0,0 +1,128 @@ +from typing import Dict, Iterable, Optional, Union + +import mosek.fusion +import numpy as np +from cpsplines.mosek_functions.utils_mosek import matrix_by_tensor_product_mosek +from cpsplines.psplines.bspline_basis import BsplineBasis +from scipy.sparse import diags + + +class PDFConstraint: + + """ + Define the constraints that characterize a Probability Density Function, + i.e., it must be non-negative and must integrate to one. For the first, only + the dictionary of interval constraints is updated with the non-negative + constraints (when they were not already included). + + Parameters + ---------- + bspline : Iterable[BsplineBasis] + An iterable containing the B-spline bases objects used to approximate + the function to estimate. + """ + + def __init__( + self, + bspline: Iterable[BsplineBasis], + ): + self.bspline = bspline + + def integrate_to_one( + self, + var_dict: Dict[str, mosek.fusion.LinearVariable], + model: mosek.fusion.Model, + ) -> mosek.fusion.LinearConstraint: + + """ + Defines the constraint that the integral of the fitted hypersurface + over the fitting and the prediction regions must be equal to one. + + Parameters + ---------- + var_dict : Dict[str, mosek.fusion.LinearVariable] + The dictionary that contains the decision variables used to define + the objective function of the problem. + model : mosek.fusion.Model + The MOSEK model of the problem. + + Returns + ------- + mosek.fusion.LinearConstraint + The constraint that enforces to fitted hypersurface integrates to + one. + """ + + banded_list = [] + for bsp in self.bspline: + # For each B-spline basis, construct a Vandermonde matrix with only + # the inner knots and with powers up to degree `deg` + 1. The first + # column, containing ones, is dropped + vander = np.vander(bsp.knots, N=bsp.deg + 2, increasing=True)[ + bsp.deg : -bsp.deg, 1: + ] + # The vector of constants obtained when integrating monomials + integrand_coef = 1 / np.linspace(1, bsp.deg + 1, bsp.deg + 1) + # Take the difference of a row minus the previous on the Vandermonde + # matrix and multiply each column by the respective integration factor + diff_mat = np.einsum("ij,j->ij", np.diff(vander, axis=0), integrand_coef) + # Multiply each row by the corresponding matrix S and transpose, so + # first term corresponds to first row and so on. Then, use this + # vector to create a banded matrix with inner knots rows nd inner + # rows plus degree columns + banded = diags( + np.array( + [np.dot(diff_mat[i], s) for i, s in enumerate(bsp.matrices_S)] + ).T, + range(bsp.deg + 1), + shape=(bsp.matrixB.shape[1] - bsp.deg, bsp.matrixB.shape[1]), + ).toarray() + banded_list.append(banded) + # Multiply by the coefficient array and sum all the entries + sum_coef = mosek.fusion.Expr.sum( + matrix_by_tensor_product_mosek( + matrices=banded_list, mosek_var=var_dict["theta"] + ) + ) + cons = model.constraint(sum_coef, mosek.fusion.Domain.equalsTo(1.0)) + return cons + + def nonneg_cons( + self, + int_constraints: Optional[Dict[int, Dict[int, Dict[str, Union[int, float]]]]], + ) -> Dict[int, Dict[int, Dict[str, Union[int, float]]]]: + + """ + Includes non-negativity constraints to the problem if they were + already not included. + + Parameters + ---------- + int_constraints : Optional[Dict[int, Dict[int, Dict[str, Union[int, float]]]]] + The nested dictionary containing the interval constraints to be + enforced. + + Returns + ------- + Dict[int, Dict[int, Dict[str, Union[int, float]]]] + A interval constraints dictionary updated with non-negativity + constraints along all axis, if they were not included explicitly. + """ + + # Create a dictionary if no interval constraints already exist + if int_constraints is None: + int_constraints = {} + for i in range(len(self.bspline)): + # Check if any interval constraints exist along each axis + if int_constraints.get(i, None) is not None: + # Check if sign constraints exist for this axis + if int_constraints[i].get(0, None) is None: + # Update the constraint dictionary if non-negativity is + # missing + int_constraints[i].update({0: {"+": 0}}) + else: + int_constraints[i][0].update({"+": 0}) + # If no constraints are imposed on any axis, include non-negativity + else: + int_constraints.update({i: {0: {"+": 0}}}) + return int_constraints diff --git a/cpsplines/psplines/bspline_basis.py b/cpsplines/psplines/bspline_basis.py index 0b15730..6346871 100644 --- a/cpsplines/psplines/bspline_basis.py +++ b/cpsplines/psplines/bspline_basis.py @@ -182,7 +182,7 @@ def get_matrix_B(self) -> None: self.matrixB = self.bspline_basis(x=x_eval) return None - def get_matrices_S(self) -> List[np.ndarray]: + def get_matrices_S(self): """ Generate a list of matrices (one by interval) containing the polynomial @@ -200,11 +200,6 @@ def get_matrices_S(self) -> List[np.ndarray]: and stock prices: a convex optimization approach. Operations Research, 50(2), 358-374. - Returns - ------- - List[np.ndarray] - List of matrices of shape (`deg` + 1, `deg` + 1) with the polynomial - coefficients. """ # Since knots are evenly spaced, the value of the B-spline basis @@ -235,4 +230,5 @@ def get_matrices_S(self) -> List[np.ndarray]: ) S_k = np.linalg.solve(T_k, C) S.append(S_k) - return S + self.matrices_S = S + return None diff --git a/examples/simulated_data.ipynb b/examples/simulated_data.ipynb index c391daf..98dc24b 100644 --- a/examples/simulated_data.ipynb +++ b/examples/simulated_data.ipynb @@ -19,6 +19,7 @@ "import itertools\n", "import matplotlib.pyplot as plt \n", "import numpy as np\n", + "from scipy.stats import norm, multivariate_normal\n", "\n", "from cpsplines.fittings.grid_cpsplines import GridCPsplines\n", "from cpsplines.graphics.plot_one_smoothing import plot_curves, plot_surfaces" @@ -1011,6 +1012,145 @@ " figsize=(10, 6),\n", ")" ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "Example 14. Given the gaussian function \n", + "$$f(x) = \\frac{1}{\\sqrt{2\\pi\\sigma^2}}\\exp\\left\\{-\\frac{(x-\\mu)^2}{2\\sigma^2}\\right\\}$$\n", + "with $\\mu=0$ and $\\sigma=2$, we simulated noisy data following the scheme\n", + "$$y_{l}= f(x_l)+\\varepsilon_{l}, \\quad \\varepsilon_{l}\\sim \\text{N}(0,0.025),$$\n", + "where $x_l=\\{\\frac{l -100}{10}\\}_{l=0}^{200}$.\n", + "\n", + "We fit an unconstrained and a constrained model over the interval $[-10,10]$\n", + "imposing that the curve is a probability density function, i.e., it is\n", + "non-negative and it integrates to one over the fitting region. " + ] + }, + { + "cell_type": "code", + "execution_count": 66, + "metadata": {}, + "outputs": [ + { + "data": { + "image/png": "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", + "text/plain": [ + "<Figure size 1080x720 with 1 Axes>" + ] + }, + "metadata": { + "needs_background": "light" + }, + "output_type": "display_data" + } + ], + "source": [ + "np.random.seed(1)\n", + "x14 = np.linspace(-10, 10, 201)\n", + "y14 = norm.pdf(x14,0,2) + np.random.normal(0, 0.025, 201)\n", + "\n", + "example14_1 = GridCPsplines(\n", + " deg=(3,),\n", + " ord_d=(2,),\n", + " n_int=(20,),\n", + " sp_args={\"options\": {\"ftol\": 1e-12}},\n", + " pdf_constraint=False\n", + ")\n", + "example14_1.fit(x=(x14,), y=y14)\n", + "\n", + "example14_2 = GridCPsplines(\n", + " deg=(3,),\n", + " ord_d=(2,),\n", + " n_int=(20,),\n", + " sp_args={\"options\": {\"ftol\": 1e-12}},\n", + " pdf_constraint=True)\n", + "example14_2.fit(x=(x14,), y=y14)\n", + "\n", + "plot14 = plot_curves(\n", + " fittings=(example14_1, example14_2),\n", + " col_curve=(\"g\", \"k\"),\n", + " knot_positions=True,\n", + " constant_constraints=True,\n", + " x=(x14,), \n", + " y=(y14,),\n", + " col_pt=(\"b\",),\n", + " alpha=0.25)" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "Example 15. Given the bivariate gaussian function \n", + "$$f(\\mathbf{x}) = \\frac{1}{2\\pi\\sqrt{\|\\boldsymbol{\\Sigma}\|}}\\exp\\left\\{-\\frac{1}{2}(\\mathbf{x}-\\boldsymbol{\\mu})^\\top\n", + "\\boldsymbol{\\Sigma}^{-1} (\\mathbf{x}-\\boldsymbol{\\mu})\\right\\}$$\n", + "with $\\mathbf{x}=(x,y)$, $\\boldsymbol{\\mu}=(0,0)$ and $\\boldsymbol{\\Sigma}=\\begin{pmatrix} 2 & 1/2 \\\\ 1/2 & 1\n", + "\\end{pmatrix},$ we simulated noisy data following the scheme\n", + "$$z_{lm}= f(x_l, y_m)+\\varepsilon_{lm}, \\quad \\varepsilon_{lm}\\sim \\text{N}(0,0.025),$$\n", + "where $x_l=\\{\\frac{3l -75}{25}\\}_{l=0}^{50}$ and $y_m=\\{\\frac{2m -60}{15}\\}_{m=0}^{60}$.\n", + "\n", + "We fit an unconstrained and a constrained model over the interval $[-3,3]\\times\n", + "[-4,4]$ imposing that the curve is a probability density function, i.e., it is\n", + "non-negative and it integrates to one over the fitting region. " + ] + }, + { + "cell_type": "code", + "execution_count": 70, + "metadata": {}, + "outputs": [ + { + "data": { + "image/png": "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", + "text/plain": [ + "<Figure size 720x432 with 2 Axes>" + ] + }, + "metadata": { + "needs_background": "light" + }, + "output_type": "display_data" + }, + { + "data": { + "image/png": "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", + "text/plain": [ + "<Figure size 720x432 with 2 Axes>" + ] + }, + "metadata": { + "needs_background": "light" + }, + "output_type": "display_data" + } + ], + "source": [ + "x15_0 = np.linspace(-3,3,51)\n", + "x15_1 = np.linspace(-4,4,61)\n", + "y15 = (multivariate_normal.pdf(x = list(itertools.product(x15_0, x15_1)),\n", + " mean=[0,0],\n", + " cov=[[2,0.5],[0.5,1]]) + np.random.normal(0, 0.025, len(x15_0)*len(x15_1))).reshape((len(x15_0),len(x15_1)))\n", + "\n", + "example15 = GridCPsplines(\n", + " deg=(3, 3),\n", + " ord_d=(2, 2),\n", + " n_int=(10, 9),\n", + " sp_method=\"grid_search\",\n", + " int_constraints={0: {0: {\"+\": 0}}, 1: {0: {\"+\": 0}}},\n", + " pdf_constraint=True\n", + " \n", + ")\n", + "example15.fit(x=(x15_0, x15_1), y=y15)\n", + "\n", + "plot15 = plot_surfaces(\n", + " fittings=(example15,),\n", + " col_surface=(\"gist_earth\",),\n", + " orientation=(45, 45),\n", + " figsize=(10, 6),\n", + ")" + ] } ], "metadata": {	Incorporate density functions constraints Beside non-negativity, probability density functions must satisfy that integrate to 1. Incorporate a new type of constraint, `density_constraint`, to ensure that the fitted hypersurface satisfies this requirement.		2022-01-03T19:31:19	[]	[]
ManuelNavarroGarcia/cpsplines	ManuelNavarroGarcia__cpsplines-10	a9dd210ad5d868629c5c5f679c69d58c3161fea2	diff --git a/cpsplines/utils/box_product.py b/cpsplines/utils/box_product.py new file mode 100644 index 0000000..739a1ef --- /dev/null +++ b/cpsplines/utils/box_product.py @@ -0,0 +1,33 @@ +import numpy as np + + +def box_product(A: np.ndarray, B: np.ndarray) -> np.ndarray: + + """ + Given a m x n matrix A and a m x p matrix B, computes the face-splitting + product or box product using that + A box B = (A kron 1_n^T) circ (1_n^T kron B), + where 1_n is a vector of ones with length n, kron denotes the Kronecker + product of matrices and circ denotes de Hadamard product of matrices. + + Parameters + ---------- + A : np.ndarray with shape (m, n) + The m x n matrix A. + B : np.ndarray with shape (m, p) + The m x p matrix B. + + Returns + ------- + np.ndarray + The resulting m x np array. + + Raises + ------ + ValueError + If any input array is not a matrix. + """ + + if A.ndim != 2 or B.ndim != 2: + raise ValueError("Only two-dimensional arrays are allowed.") + return np.multiply(np.repeat(A, B.shape[1], axis=1), np.tile(B, A.shape[1]))	Incorporate box product (or row-wise products) Row-wise products are needed to smooth scattered data. Implement a fast way to compute them (try not to compute any Kronecker product) and provide tests to evaluate the running time.		2021-11-25T18:57:02	[]	[]
RWTH-EBC/ebcpy	RWTH-EBC__ebcpy-136	8fad8a2f4c2ec4411247de2a4fc842c6c6c8c219	diff --git a/CHANGELOG.md b/CHANGELOG.md index e84e0a4a..9fcac3f1 100644 --- a/CHANGELOG.md +++ b/CHANGELOG.md @@ -104,3 +104,5 @@ - fix reproduction with modifiers #82 - enable model_name=None upon startup of DymolaAPI #130 - Only create logger handler if not already done by root-logger +- v0.4.2 + - Do not use chars_to_strings from scipy #132 diff --git a/ebcpy/__init__.py b/ebcpy/__init__.py index 9f47111e..d1c2cbd8 100644 --- a/ebcpy/__init__.py +++ b/ebcpy/__init__.py @@ -8,4 +8,4 @@ from .optimization import Optimizer -__version__ = '0.4.1' +__version__ = '0.4.2' diff --git a/ebcpy/modelica/simres.py b/ebcpy/modelica/simres.py index dad8320e..9367c0ad 100644 --- a/ebcpy/modelica/simres.py +++ b/ebcpy/modelica/simres.py @@ -53,7 +53,6 @@ from itertools import count from collections import namedtuple from scipy.io import loadmat -from scipy.io.matlab.mio_utils import chars_to_strings import pandas as pd import numpy as np @@ -229,8 +228,8 @@ def get_strings(str_arr): Strip the whitespace from the right and recode it as utf-8. """ - return [line.rstrip(' \0').encode('latin-1').decode('utf-8') - for line in chars_to_strings(str_arr)] + return ["".join(word_arr).replace(" ", "") + for word_arr in str_arr] class Variable(namedtuple('VariableNamedTuple', ['samples', 'description', 'unit', 'displayUnit'])): diff --git a/requirements.txt b/requirements.txt index c61b2c8d..72b814ec 100644 --- a/requirements.txt +++ b/requirements.txt @@ -1,6 +1,6 @@ -numpy>=1.19.5 +numpy>=1.19.5,<2.0 matplotlib>=3.3.4 -scipy~=1.13.0 +scipy>=1.5.4 pandas>=1.1.5 tables>=3.6.1 scikit-learn>=0.24.2 diff --git a/setup.py b/setup.py index d8099392..9c6cba2a 100644 --- a/setup.py +++ b/setup.py @@ -19,9 +19,9 @@ } INSTALL_REQUIRES = [ - 'numpy>=1.19.5', + 'numpy>=1.19.5,<2.0', 'matplotlib>=3.3.4', - 'scipy~=1.13.0', + 'scipy>=1.5.4', 'pandas>=1.1.5', 'scikit-learn>=0.24.2', 'fmpy>=0.2.27,<0.3.17',	132 mio utils is deprecated Closes #132		2024-07-15T08:33:49	[]	[]
RWTH-EBC/ebcpy	RWTH-EBC__ebcpy-116	060d10d3a657850343d02c64645b8de7b2d5a526	diff --git a/CHANGELOG.md b/CHANGELOG.md index 19d978b0..e44216e9 100644 --- a/CHANGELOG.md +++ b/CHANGELOG.md @@ -86,3 +86,5 @@ - Fix fmpy version #109 - Fix data format for index frequency of tsd #107 - Fix bug in function clean_and_space_equally_time_series #111 +- v0.3.11 + - Add dymola_exe_path to set of kwargs in DymolaAPI #115 diff --git a/ebcpy/__init__.py b/ebcpy/__init__.py index 5b0f57db..8699cb0e 100644 --- a/ebcpy/__init__.py +++ b/ebcpy/__init__.py @@ -8,4 +8,4 @@ from .optimization import Optimizer -__version__ = '0.3.10' +__version__ = '0.3.11' diff --git a/ebcpy/simulationapi/dymola_api.py b/ebcpy/simulationapi/dymola_api.py index c2672280..1835b7c4 100644 --- a/ebcpy/simulationapi/dymola_api.py +++ b/ebcpy/simulationapi/dymola_api.py @@ -56,10 +56,6 @@ class DymolaAPI(SimulationAPI): :keyword Boolean equidistant_output: If True (Default), Dymola stores variables in an equisdistant output and does not store variables at events. - :keyword str dymola_path: - Path to the dymola installation on the device. Necessary - e.g. on linux, if we can't find the path automatically. - Example: ``dymola_path="C://Program Files//Dymola 2020x"`` :keyword int n_restart: Number of iterations after which Dymola should restart. This is done to free memory. Default value -1. For values @@ -86,7 +82,7 @@ class DymolaAPI(SimulationAPI): If given, the Version needs to be equal to the folder name of your installation. - Example: If you have two version installed at + Example: If you have two versions installed at - ``C://Program Files//Dymola 2021`` and - ``C://Program Files//Dymola 2020x`` @@ -95,9 +91,18 @@ class DymolaAPI(SimulationAPI): ``dymola_version='Dymola 2020x'``. This parameter is overwritten if ``dymola_path`` is specified. + :keyword str dymola_path: + Path to the dymola installation on the device. Necessary + e.g. on linux, if we can't find the path automatically. + Example: ``dymola_path="C://Program Files//Dymola 2020x"`` :keyword str dymola_interface_path: - Only relevant for the case when the dymola-exe path + Direct path to the .egg-file of the dymola interface. + Only relevant when the dymola_path differs from the interface path. + :keyword str dymola_exe_path: + Direct path to the dymola executable. + Only relevant if the dymola installation do not follow + the official guideline. Example: @@ -129,7 +134,8 @@ class DymolaAPI(SimulationAPI): "mos_script_pre", "mos_script_post", "dymola_version", - "dymola_interface_path" + "dymola_interface_path", + "dymola_exe_path" ] def __init__(self, cd, model_name, packages=None, kwargs): @@ -146,6 +152,7 @@ def __init__(self, cd, model_name, packages=None, kwargs): self.mos_script_post = kwargs.pop("mos_script_post", None) self.dymola_version = kwargs.pop("dymola_version", None) self.dymola_interface_path = kwargs.pop("dymola_interface_path", None) + self.dymola_exe_path = kwargs.pop("dymola_exe_path", None) for mos_script in [self.mos_script_pre, self.mos_script_post]: if mos_script is not None: if not os.path.isfile(mos_script): @@ -175,31 +182,30 @@ def __init__(self, cd, model_name, packages=None, **kwargs): if not os.path.exists(dymola_path): raise FileNotFoundError(f"Given path '{dymola_path}' can not be found on " "your machine.") - _dym_install = dymola_path else: # Get the dymola-install-path: _dym_installations = self.get_dymola_install_paths() if _dym_installations: if self.dymola_version: - _found_version = False - for _dym_install in _dym_installations: - if _dym_install.endswith(self.dymola_version): - _found_version = True - break - if not _found_version: - raise ValueError( - f"Given dymola_version '{self.dymola_version}' not found in " - f"the list of dymola installations {_dym_installations}" - ) + dymola_path = _get_dymola_path_of_version( + dymola_installations=_dym_installations, + dymola_version=self.dymola_version + ) else: - _dym_install = _dym_installations[0] # 0 is the newest - self.logger.info("Using dymola installation at %s", _dym_install) + dymola_path = _dym_installations[0] # 0 is the newest + self.logger.info("Using dymola installation at %s", dymola_path) else: - raise FileNotFoundError("Could not find a dymola-interface on your machine.") - self.dymola_exe_path = self.get_dymola_path(_dym_install) + if self.dymola_exe_path is None or self.dymola_interface_path is None: + raise FileNotFoundError( + "Could not find dymola on your machine. " + "Thus, not able to find the `dymola_exe_path` and `dymola_interface_path`. " + "Either specify both or pass an existing `dymola_path`." + ) + if self.dymola_exe_path is None: + self.dymola_exe_path = self.get_dymola_path(dymola_path) self.logger.info("Using dymola.exe: %s", self.dymola_exe_path) if self.dymola_interface_path is None: - self.dymola_interface_path = self.get_dymola_interface_path(_dym_install) + self.dymola_interface_path = self.get_dymola_interface_path(dymola_path) self.logger.info("Using dymola interface: %s", self.dymola_interface_path) self.packages = [] @@ -1022,7 +1028,7 @@ def get_dymola_path(dymola_install_dir, dymola_name=None): """ if dymola_name is None: if "linux" in sys.platform: - dymola_name = "dymola" + dymola_name = "dymola.sh" elif "win" in sys.platform: dymola_name = "Dymola.exe" else: @@ -1166,3 +1172,18 @@ def _check_restart(self): self.sim_counter = 1 else: self.sim_counter += 1 + + +def _get_dymola_path_of_version(dymola_installations: list, dymola_version: str): + """ + Helper function to get the path associated to the dymola_version + from the list of all installations + """ + for dymola_path in dymola_installations: + if dymola_path.endswith(dymola_version): + return dymola_path + # If still here, version was not found + raise ValueError( + f"Given dymola_version '{dymola_version}' not found in " + f"the list of dymola installations {dymola_installations}" + )	calling dymola on linux maschin not working problem: not the right call of dymola via ebcpy - for running the simulation in dymola the python package ebcpy is used - ebcpy calls automaticlly '/opt/dymola-2023x-x86_64/bin64/dymola' - but here the needed libs are not found - it is needed to call '/opt/dymola-2023x-x86_64/bin64/dymola.sh' - solution: edit the file dymola_api.py of the package ebcpy - adapt line 1025: dymola_name = "dymola" > "dymola.sh" System: Ubuntu 22.04 dymola-2023x-x86_64 calling dymola on linux maschin not working problem: not the right call of dymola via ebcpy - for running the simulation in dymola the python package ebcpy is used - ebcpy calls automaticlly '/opt/dymola-2023x-x86_64/bin64/dymola' - but here the needed libs are not found - it is needed to call '/opt/dymola-2023x-x86_64/bin64/dymola.sh' - solution: edit the file dymola_api.py of the package ebcpy - adapt line 1025: dymola_name = "dymola" > "dymola.sh" System: Ubuntu 22.04 dymola-2023x-x86_64	Same as for the `dymola_interface_path`, I would just add the explicit path to the list of kwargs. This solution would not work for our dymola installation, so it seems to depend on the end user. Would that work for you as well? Do you mean I should add the kwarg dymola_interface_path to DymolaAPI()? I tried to use dymola_path here. but it does not work. Now, I tried he kwarg dymola_interface_path as well, but it does not work too. When dymola.sh is called, also the pathes of the libs are added. From my point of view. it is the problem. Sorry, that was not clear: I meant I would add a new option `dymola_exe_path`. So you could specify: ```DymolaAPI(dymola_exe_path='/opt/dymola-2023x-x86_64/bin64/dymola.sh', ...)``` Would that work for you? Another option is to specify a .mos script to run upon starting dymola, so all libs are called. What is the content of the .sh file? Maybe just a .mos file? thx, for your answers. Regarding path kwargs, when a new kwarg is used, the ebcpy respectively dymola_api.py must be adapted, right? This I like to avoid. Is there another possibility? Regarding libraries: the dymola.sh adds paths to load (old) system libraries, like libicui18n.so.60, which are part of the dymola installation. I looked into the dockerfile bim2sim is using. Here I find a symlink of the dymola.sh to /usr/local/bin/dymola. Maybe I will try this. Ah ok, if you use the Dockerfile from bim2sim, it will be similar to the AixLib ci. Here we use: ``` if "linux" in sys.platform: dymola_path = "/usr/local" dymola_interface_path = "/opt/dymola-2022-x86_64/Modelica/Library/python_interface/dymola.egg" ``` for kwargs. @Cudok and I discussed this in recent meeting. From my understanding using dymola.sh should also work with our dymola docker installation. Here is a screenshot from the setup in the dymola container. As you can see the dymola.sh exists and links correctly to the dymola installation. Also running the dymola.sh works ![image](https://github.com/RWTH-EBC/ebcpy/assets/27726960/8ca13512-5004-465a-955b-8e84a7f4dbf5) You can also see in the [Dockerfile](https://git.rwth-aachen.de/EBC/EBC_intern/dymola-docker/-/blame/master/Dockerfile#L62) that the symlink is created: `ln -s /opt/${DYM_FOLDER}/bin64/dymola.sh /usr/local/bin/dymola && \` As different Linux systems use different setups, we should make ebcpy usable "out of the box" for most of them. As @Cudok mentioned, he followed the official guideline installing Dymola under Linux (which we apparently didn't in our dockerfile) and this states to use alien which (could you add some information here @Cudok?) As @DaJansenGit mentioned, it is recommend to use alien to install the dymola rpm package in the full user manual (part of the installation see /documentation/). Following you find the relevant part from the full user manual: 17.2.1 Installing Dymola Dymola for Linux is distributed as an RPM package. The package is installed using the command `rpm –i name-of-distribution.rpm` Optional libraries are installed through separate RPM files. For installation on e.g. Debian or Kubuntu systems conversion to the deb format is required using the alien command: `alien –cki name-of-distribution.rpm` We should definitely follow the official guideline as a default. I will create a fix for this Same as for the `dymola_interface_path`, I would just add the explicit path to the list of kwargs. This solution would not work for our dymola installation, so it seems to depend on the end user. Would that work for you as well? Do you mean I should add the kwarg dymola_interface_path to DymolaAPI()? I tried to use dymola_path here. but it does not work. Now, I tried he kwarg dymola_interface_path as well, but it does not work too. When dymola.sh is called, also the pathes of the libs are added. From my point of view. it is the problem. Sorry, that was not clear: I meant I would add a new option `dymola_exe_path`. So you could specify: ```DymolaAPI(dymola_exe_path='/opt/dymola-2023x-x86_64/bin64/dymola.sh', ...)``` Would that work for you? Another option is to specify a .mos script to run upon starting dymola, so all libs are called. What is the content of the .sh file? Maybe just a .mos file? thx, for your answers. Regarding path kwargs, when a new kwarg is used, the ebcpy respectively dymola_api.py must be adapted, right? This I like to avoid. Is there another possibility? Regarding libraries: the dymola.sh adds paths to load (old) system libraries, like libicui18n.so.60, which are part of the dymola installation. I looked into the dockerfile bim2sim is using. Here I find a symlink of the dymola.sh to /usr/local/bin/dymola. Maybe I will try this. Ah ok, if you use the Dockerfile from bim2sim, it will be similar to the AixLib ci. Here we use: ``` if "linux" in sys.platform: dymola_path = "/usr/local" dymola_interface_path = "/opt/dymola-2022-x86_64/Modelica/Library/python_interface/dymola.egg" ``` for kwargs. @Cudok and I discussed this in recent meeting. From my understanding using dymola.sh should also work with our dymola docker installation. Here is a screenshot from the setup in the dymola container. As you can see the dymola.sh exists and links correctly to the dymola installation. Also running the dymola.sh works ![image](https://github.com/RWTH-EBC/ebcpy/assets/27726960/8ca13512-5004-465a-955b-8e84a7f4dbf5) You can also see in the [Dockerfile](https://git.rwth-aachen.de/EBC/EBC_intern/dymola-docker/-/blame/master/Dockerfile#L62) that the symlink is created: `ln -s /opt/${DYM_FOLDER}/bin64/dymola.sh /usr/local/bin/dymola && \` As different Linux systems use different setups, we should make ebcpy usable "out of the box" for most of them. As @Cudok mentioned, he followed the official guideline installing Dymola under Linux (which we apparently didn't in our dockerfile) and this states to use alien which (could you add some information here @Cudok?) As @DaJansenGit mentioned, it is recommend to use alien to install the dymola rpm package in the full user manual (part of the installation see /documentation/). Following you find the relevant part from the full user manual: 17.2.1 Installing Dymola Dymola for Linux is distributed as an RPM package. The package is installed using the command `rpm –i name-of-distribution.rpm` Optional libraries are installed through separate RPM files. For installation on e.g. Debian or Kubuntu systems conversion to the deb format is required using the alien command: `alien –cki name-of-distribution.rpm` We should definitely follow the official guideline as a default. I will create a fix for this	2023-12-28T10:14:29	[]	[]
RWTH-EBC/ebcpy	RWTH-EBC__ebcpy-106	5a56a8e8ec6633d2ac5a8c204967b93166b2ec08	diff --git a/ebcpy/modelica/manipulate_ds.py b/ebcpy/modelica/manipulate_ds.py index 0ef51333..922c0595 100644 --- a/ebcpy/modelica/manipulate_ds.py +++ b/ebcpy/modelica/manipulate_ds.py @@ -2,6 +2,8 @@ dsfinal.txt and dsin.txt files created by Modelica.""" from io import StringIO +import re + import pandas as pd @@ -37,17 +39,24 @@ def convert_ds_file_to_dataframe(filename): :return: pd.DataFrame Converted DataFrame """ - # Define relevant parameters - number_line_initial_name = 104 # Line where the string char initialName(,) is always stored - # Open file and get relevant content by splitting the lines. with open(filename, "r") as file: content = file.read().split("\n") # Gets the X out of 'char initialName(X,Y)' - size_initial_names = int(content[number_line_initial_name].split("(")[-1].split(",")[0]) + pattern_size_initial_name = r'^char initialName\((\d+),(\d+)\)$' + for number_line_initial_name, line in enumerate(content): + match_size_initial_name = re.match(pattern_size_initial_name, line) + if match_size_initial_name: + size_initial_names = int(match_size_initial_name.string.split("(")[-1].split(",")[0]) + break + else: + raise ValueError("Could not find initial names in file") + # Number of line below line "double initialValue(X,Y)" number_line_initial_value = number_line_initial_name + size_initial_names + 3 + if "double initialValue" in content[number_line_initial_value]: + number_line_initial_value += 1 # Check if two or on-line dsfinal / dsin if "#" in content[number_line_initial_value]:	Fix model translation in Dymola 2023 Dymola 2023 seems to throw errors when loading dsin files. This needs further investigation and fix. Example ``` File "...\ebcpy\ebcpy\modelica\manipulate_ds.py", line 48, in convert_ds_file_to_dataframe size_initial_names = int(content[number_line_initial_name].split("(")[-1].split(",")[0]) ValueError: invalid literal for int() with base 10: 'multizone.zoneParam[1].AExt[3]' ``` Fix model translation in Dymola 2023 Dymola 2023 seems to throw errors when loading dsin files. This needs further investigation and fix. Example ``` File "...\ebcpy\ebcpy\modelica\manipulate_ds.py", line 48, in convert_ds_file_to_dataframe size_initial_names = int(content[number_line_initial_name].split("(")[-1].split(",")[0]) ValueError: invalid literal for int() with base 10: 'multizone.zoneParam[1].AExt[3]' ```	A temporary fix is to disable this feature in the DymolaAPI by using: `extract_variables=False` as a keyword argument. A temporary fix is to disable this feature in the DymolaAPI by using: `extract_variables=False` as a keyword argument.	2023-10-02T11:42:37	[]	[]
RWTH-EBC/ebcpy	RWTH-EBC__ebcpy-60	50057aa10db14c2636c0ae64ad760f7c30a4f419	diff --git a/CHANGELOG.md b/CHANGELOG.md index e36fd82f..51792076 100644 --- a/CHANGELOG.md +++ b/CHANGELOG.md @@ -58,4 +58,6 @@ - v0.3.0: - Add parallelization option for optimization backends - v0.3.1: - - Avoid DType Warning #53 \ No newline at end of file + - Avoid DType Warning #53 +- v0.3.2: + - Correct bounds for simulation Variables #56 \ No newline at end of file diff --git a/docs/source/conf.py b/docs/source/conf.py index 1bec663a..be7ae0e8 100644 --- a/docs/source/conf.py +++ b/docs/source/conf.py @@ -71,7 +71,7 @@ # The short X.Y version. version = '0.3' # The full version, including alpha/beta/rc tags. -release = '0.3.1' +release = '0.3.2' # The language for content autogenerated by Sphinx. Refer to documentation # for a list of supported languages. diff --git a/ebcpy/__init__.py b/ebcpy/__init__.py index b90ea14b..ee001230 100644 --- a/ebcpy/__init__.py +++ b/ebcpy/__init__.py @@ -8,4 +8,4 @@ from .optimization import Optimizer -__version__ = '0.3.1' +__version__ = '0.3.2' diff --git a/ebcpy/simulationapi/__init__.py b/ebcpy/simulationapi/__init__.py index 3a2c6649..ea17f82e 100644 --- a/ebcpy/simulationapi/__init__.py +++ b/ebcpy/simulationapi/__init__.py @@ -5,6 +5,7 @@ much more user-friendly than the provided APIs by Dymola or fmpy. """ import logging +import warnings import os import itertools from typing import Dict, Union, TypeVar, Any, List @@ -20,25 +21,60 @@ class Variable(BaseModel): Data-Class to store relevant information for a simulation variable (input, parameter, output or local/state). """ + type: Any = Field( + default=None, + title='type', + description='Type of the variable' + ) value: Any = Field( description="Default variable value" ) - max: Union[float, int] = Field( - default=np.inf, + max: Any = Field( + default=None, title='max', - description='Maximal value (upper bound) of the variables value' - ) - min: Union[float, int] = Field( - default=-np.inf, - title='min', - description='Minimal value (lower bound) of the variables value' + description='Maximal value (upper bound) of the variables value. ' + 'Only for ints and floats variables.' ) - type: Any = Field( + min: Any = Field( default=None, - title='type', - description='Type of the variable' + title='min', + description='Minimal value (lower bound) of the variables value. ' + 'Only for ints and floats variables.' ) + @validator("value") + def check_value_type(cls, value, values): + """Check if the given value has correct type""" + _type = values["type"] + if _type is None: + return value # No type -> no conversion + if value is None: + return value # Setting None is allowed. + if not isinstance(value, _type): + return _type(value) + return value + + @validator('max', 'min', always=True) + def check_value(cls, value, values, field): + """Check if the given bounds are correct.""" + # Check if the variable type even allows for min/max bounds + _type = values["type"] + if _type is None: + return value # No type -> no conversion + if _type not in (float, int, bool): + if value is not None: + warnings.warn( + "Setting a min/max for variables " + f"of type {_type} is not supported." + ) + return None + if value is not None: + return _type(value) + if field.name == "min": + return -np.inf if _type != bool else False + # else it is max + return np.inf if _type != bool else True + class SimulationSetup(BaseModel): """ diff --git a/ebcpy/simulationapi/fmu.py b/ebcpy/simulationapi/fmu.py index 773bb24a..243817fb 100644 --- a/ebcpy/simulationapi/fmu.py +++ b/ebcpy/simulationapi/fmu.py @@ -287,11 +287,6 @@ def setup_fmu_instance(self): else: self._fmi_type = 'CoSimulation' - def _to_bound(value): - if value is None or \ - not isinstance(value, (float, int, bool)): - return np.inf - return value self.logger.info("Reading model variables") _types = { @@ -307,8 +302,8 @@ def _to_bound(value): var.start = _types[var.type](var.start) _var_ebcpy = Variable( - min=-_to_bound(var.min), - max=_to_bound(var.max), + min=var.min, + max=var.max, value=var.start, type=_types[var.type] ) diff --git a/setup.py b/setup.py index fb1aec4c..acf2e22e 100644 --- a/setup.py +++ b/setup.py @@ -12,7 +12,7 @@ 'full': [ 'openpyxl>=3.0.5', 'xlrd>=2.0.1', - 'pymoo>=0.4.2', + 'pymoo==0.4.2', ] } @@ -32,7 +32,7 @@ # Add all open-source packages to setup-requires SETUP_REQUIRES = INSTALL_REQUIRES.copy() -VERSION = "0.3.1" +VERSION = "0.3.2" setuptools.setup( name='ebcpy',	Fix parameter bounds in class FMU_API Currently, the bounds of the parameters in an FMU_API are always set to minus infinity and infinity.		2022-08-22T14:01:22	[]	[]
django/daphne	django__daphne-500	9a282dd6277f4b921e494a9fc30a534a7f0d6ca6	diff --git a/daphne/http_protocol.py b/daphne/http_protocol.py index f0657fdb..b7da1bf7 100755 --- a/daphne/http_protocol.py +++ b/daphne/http_protocol.py @@ -9,7 +9,7 @@ from twisted.web import http from zope.interface import implementer -from .utils import parse_x_forwarded_for +from .utils import HEADER_NAME_RE, parse_x_forwarded_for logger = logging.getLogger(__name__) @@ -69,6 +69,13 @@ def __init__(self, args, kwargs): def process(self): try: self.request_start = time.time() + + # Validate header names. + for name, _ in self.requestHeaders.getAllRawHeaders(): + if not HEADER_NAME_RE.fullmatch(name): + self.basic_error(400, b"Bad Request", "Invalid header name") + return + # Get upgrade header upgrade_header = None if self.requestHeaders.hasHeader(b"Upgrade"): diff --git a/daphne/utils.py b/daphne/utils.py index 81f1f9df..06993142 100644 --- a/daphne/utils.py +++ b/daphne/utils.py @@ -1,7 +1,12 @@ import importlib +import re from twisted.web.http_headers import Headers +# Header name regex as per h11. +# https://github.com/python-hyper/h11/blob/a2c68948accadc3876dffcf979d98002e4a4ed27/h11/_abnf.py#L10-L21 +HEADER_NAME_RE = re.compile(rb"[-!#$%&'+.^_`\|~0-9a-zA-Z]+") + def import_by_path(path): """	Daphne allows invalid characters within header names # The bug RFC 9110 defines the header names must consist of a single token, which is defined as follows: ``` token = 1tchar tchar = "!" / "#" / "$" / "%" / "&" / "'" / "" / "+" / "-" / "." / "^" / "_" / "`" / "\|" / "~" / DIGIT / ALPHA ; any VCHAR, except delimiters ``` When Daphne receives a request with a header name containing any or all of the following characters, it does not reject the message: ``` \x00\x01\x02\x03\x04\x05\x06\x07\x08\t\x0b\x0c\x0e\x0f\x10\x11\x12\x13\x14\x15\x16\x17\x18\x19\x1a\x1b\x1c\x1d\x1e\x1f "(),/;<=>?@[/]{}\xc0\xc1\xc2\xc3\xc4\xc5\xc6\xc7\xc8\xc9\xca\xcb\xcc\xcd\xce\xcf\xd0\xd1\xd2\xd3\xd4\xd5\xd6\xd7\xd8\xd9\xda\xdb\xdc\xdd\xde\xdf\xe0\xe1\xe2\xe3\xe4\xe5\xe6\xe7\xe8\xe9\xea\xeb\xec\xed\xee\xef ``` Thus, when you send the following request: ``` GET / HTTP/1.1\r\n \x00\x01\x02\x03\x04\x05\x06\x07\x08\t\x0b\x0c\x0e\x0f\x10\x11\x12\x13\x14\x15\x16\x17\x18\x19\x1a\x1b\x1c\x1d\x1e\x1f "(),/;<=>?@[/]{}\xc0\xc1\xc2\xc3\xc4\xc5\xc6\xc7\xc8\xc9\xca\xcb\xcc\xcd\xce\xcf\xd0\xd1\xd2\xd3\xd4\xd5\xd6\xd7\xd8\xd9\xda\xdb\xdc\xdd\xde\xdf\xe0\xe1\xe2\xe3\xe4\xe5\xe6\xe7\xe8\xe9\xea\xeb\xec\xed\xee\xef: whatever\r\n \r\n ``` Daphne's interpretation is as follows: ``` [ HTTPRequest( method=b'GET', uri=b'/', version=b'1.1', headers=[ (b'\x00\x01\x02\x03\x04\x05\x06\x07\x08\t\x0b\x0c\x0e\x0f\x10\x11\x12\x13\x14\x15\x16\x17\x18\x19\x1a\x1b\x1c\x1d\x1e\x1f "(),/;<=>?@[/]{}\xc0\xc1\xc2\xc3\xc4\xc5\xc6\xc7\xc8\xc9\xca\xcb\xcc\xcd\xce\xcf\xd0\xd1\xd2\xd3\xd4\xd5\xd6\xd7\xd8\xd9\xda\xdb\xdc\xdd\xde\xdf\xe0\xe1\xe2\xe3\xe4\xe5\xe6\xe7\xe8\xe9\xea\xeb\xec\xed\xee\xef', b'whatever'), ], body=b'', ), ] ``` (i.e. the invalid header name made it into the request unchanged) Nearly all other HTTP implementations reject the above request, including Uvicorn, Hypercorn, AIOHTTP, Apache httpd, Bun, CherryPy, Deno, FastHTTP, Go net/http, Gunicorn, H2O, Hyper, Jetty, libsoup, Lighttpd, Mongoose, Nginx, Node.js LiteSpeed, Passenger, Puma, Tomcat, OpenWrt uhttpd, Uvicorn, Waitress, WEBrick, and OpenBSD httpd. ## Your OS and runtime environment: ``` $ uname -a Linux de8a63fbb2f4 6.7.2-arch1-2 #1 SMP PREEMPT_DYNAMIC Wed, 31 Jan 2024 09:22:15 +0000 x86_64 GNU/Linux ``` ## `pip freeze`: ``` asgiref==3.7.2 attrs==23.2.0 autobahn==23.6.2 Automat==22.10.0 cffi==1.16.0 constantly==23.10.4 cryptography==42.0.2 Cython==0.29.32 daphne @ file:///app/daphne h2==4.1.0 hpack==4.0.0 hyperframe==6.0.1 hyperlink==21.0.0 idna==3.6 incremental==22.10.0 priority==1.3.0 pyasn1==0.5.1 pyasn1-modules==0.3.0 pycparser==2.21 Pygments==2.14.0 pyOpenSSL==24.0.0 python-afl==0.7.4 PyYAML==6.0 service-identity==24.1.0 six==1.16.0 Twisted==23.10.0 txaio==23.1.1 typing_extensions==4.9.0 zope.interface==6.1 ``` (Using Daphne built from main, with the latest commit being https://github.com/django/daphne/commit/993efe62ce9e9c1cd64c81b6ee7dfa7b819482c7 at the time of writing) ## How you're running Channels (runserver? daphne/runworker? Nginx/Apache in front?) Directly invoking Daphne from the command line. ## Console logs and full tracebacks of any errors N/A	Of particular note is that `\x00`, `\t`, and ` ` are permitted, which have historically been sources of exploitable parsing discrepancies between origin servers and gateway servers. @kenballus thanks for the report. Are you up for making a PR here? h11 gives a regex for a header name here: https://github.com/python-hyper/h11/blob/a2c68948accadc3876dffcf979d98002e4a4ed27/h11/_abnf.py#L10-L21C1 That's then validated when parsing the headers here: https://github.com/python-hyper/h11/blob/a2c68948accadc3876dffcf979d98002e4a4ed27/h11/_headers.py#L163 Gunicorn does similar here https://github.com/benoitc/gunicorn/blob/cf55d2cec277f220ebd605989ce78ad1bb553c46/gunicorn/http/message.py#L92-L93 (but matching bad headers, with what looks like a more complex regex.) We'd handle that in the http and ws protocols: https://github.com/django/daphne/blob/993efe62ce9e9c1cd64c81b6ee7dfa7b819482c7/daphne/http_protocol.py#L147-L158 https://github.com/django/daphne/blob/993efe62ce9e9c1cd64c81b6ee7dfa7b819482c7/daphne/ws_protocol.py#L37-L44 Need to investigate whether this should be handled by twisted. https://github.com/twisted/twisted/blob/446ee139189440e890b26a29af256e9b9d0e8eba/src/twisted/web/http_headers.py h11 test examples for `no weird characters in names` are here: https://github.com/python-hyper/h11/blob/a2c68948accadc3876dffcf979d98002e4a4ed27/h11/tests/test_headers.py#L24-L35	2024-02-10T14:23:01	[]	[]
django/daphne	django__daphne-422	6199d509c200cdd5102b39a3461a62984adebaf7	diff --git a/CHANGELOG.txt b/CHANGELOG.txt index 31e46504..8eb54634 100644 --- a/CHANGELOG.txt +++ b/CHANGELOG.txt @@ -20,6 +20,12 @@ Unreleased * Added ``--log-fmt`` CLI argument. +* Added support for ``ASGI_THREADS`` environment variable, setting the maximum + number of workers used by a ``SyncToAsync`` thread-pool executor. + + Set e.g. ``ASGI_THREADS=4 daphne ...`` when running to limit the number of + workers. + 3.0.2 (2021-04-07) ------------------ diff --git a/daphne/server.py b/daphne/server.py index 43342175..e5728b3c 100755 --- a/daphne/server.py +++ b/daphne/server.py @@ -1,10 +1,18 @@ # This has to be done first as Twisted is import-order-sensitive with reactors import asyncio # isort:skip +import os # isort:skip import sys # isort:skip import warnings # isort:skip +from concurrent.futures import ThreadPoolExecutor # isort:skip from twisted.internet import asyncioreactor # isort:skip + twisted_loop = asyncio.new_event_loop() +if "ASGI_THREADS" in os.environ: + twisted_loop.set_default_executor( + ThreadPoolExecutor(max_workers=int(os.environ["ASGI_THREADS"])) + ) + current_reactor = sys.modules.get("twisted.internet.reactor", None) if current_reactor is not None: if not isinstance(current_reactor, asyncioreactor.AsyncioSelectorReactor):	ASGI_THREADS environment variable has no effect since 2.4.1 Channels documentation mentions that: > By default, the number of threads is set to “the number of CPUs * 5”, so you may see up to this number of threads. If you want to change it, set the ASGI_THREADS environment variable to the maximum number you wish to allow. This behaviour used to work correctly before Daphne v2.4.1. Since 2.4.1 setting ASGI_THREADS envvar has no effect. #### My Setup Python 3.7.5, on Ubuntu 19.10 with following packages: ``` asgiref==3.2.7 channels==2.4.0 channels-redis==2.4.2 daphne==2.4.1 Django==2.2.12 ``` Startup: ``` daphne -b 0.0.0.0 -p 8000 api.config.asgi:application ``` #### How to replicate * Set ASGI_THREADS to 1. * Install packages above, with Daphne v2.40 * Start daphne * Navigate through website and open a few pages * Check number of threads with htop. * Thread count never exceeds 1 * Now install daphne v2.41 * Start server and browse website * Thread count is not limited to 1, keeps increasing. Possibly capped to 5*CPU count, but can't tell for sure.	Nothing related to that was changed in 2.4.1, that stuff is all managed by asgiref. The thread stuff is a lot better in python 3.8 but imo the whole ThreadPoolExecutor design is terrible. @JohnDoee It did, it is this commit https://github.com/django/daphne/commit/27f760a8147efd0c2f4cb842cbe9d50cfdf40d46. Reverting this change resolves the issue. I tested it, sorry, you're right. I assumed the correct loop was installed as global correctly. It seems like the "correct" ThreadPoolExecutor is installed on the wrong loop. Hmmm. Yes. It's because `SyncToAsync` is using `get_event_loop()` which is returning the default event loop for the main thread, rather than the loop we're passing to twisted (for use in the running thread). I assumed that https://github.com/django/daphne/blob/2.4.1/daphne/server.py#L129 would make it all correct but maybe stuff is executed in the wrong order. I've been testing this a bit further, this actually breaks everything that uses asyncio and is initialized at the "wrong" time. The original loop seems to be stopped. Now I have this piece of boilerplate code at the top somewhere imported early to make sure everything else uses the correct loop. ``` import asyncio # isort:skip import daphne.server # isort:skip from twisted.internet import reactor # isort:skip asyncio.set_event_loop(reactor._asyncioEventloop) # isort:skip ``` See https://github.com/django/asgiref/issues/185#issuecomment-667948950: > We need to resolve, probably reverting django/daphne@27f760a — but then that recreates django/channels#1374 — so it's currently pending time to think about it. #299 is related. So I have been looking through asgiref, channels daphne etc… and I came to the following conclusion: Remove `ASGI_THREADS` support from asgiref. Before you say no, hear me out… * `asgiref` is a low level library which basically only requires an event loop to be running, it should not concern itself with more. * It does not know from which thread it is imported and as such it shouldn't make assumption about the event loop there, * `get_event_loop` which is currently used is slated for deprecation and it appears as if only `get_running_loop` will be supported going forward -- which is not something that will exist during import time usually. * Even if the event loop where running and we were able to get it at import time, does it sound safe to change the associated thread pool while the event loop is running? (does python even support that?) Okay, now with `ASGI_THREADS` removed we need to bring back the functionality. How do we do that? As stopgap measure we can simply add it to `daphne/server.py` right after we call `new_event_loop` -- it is our event loop after all, so let's see the pool there. Going further we can deprecate `ASGI_THREADS` completely and implement this in the ASGI servers them self, they are the ones in control (and owners) of the loop after all. So for daphne I'd expect support for `--worker-threads` (or so…), which `manage.py runserver` from channels could also use then. What do you think? > See [django/asgiref#185 (comment)](https://github.com/django/asgiref/issues/185#issuecomment-667948950): > > > We need to resolve, probably reverting [27f760a](https://github.com/django/daphne/commit/27f760a) — but then that recreates [django/channels#1374](https://github.com/django/channels/issues/1374) — so it's currently pending time to think about it. @carltongibson: I started to think about that and ran over your comment in https://github.com/django/channels/issues/1374#issuecomment-566705986 -- when I tried this today on Python 3.10 and up2date asgiref/django I did not get `INCORRECTLY raised SynchronousOnlyOperation` but only: ``` /home/florian/sources/async/test3.py:39: DeprecationWarning: There is no current event loop target=run_in_async_thread, args=(asyncio.get_event_loop(),), kwargs={}, name='async-thread' Async thread starting Will start the loop About to call sync_only from async context Correctly raised SynchronousOnlyOperation About to call sync_only from SYNC context sync_only successfully called. ``` with that in mind it might (?) be possible to actually revert the changes. That said I would recommend against reverting because creating our own event loop and configuring it as needed sounds like the way to go… Nevertheless I'd be interested to know if you have an idea whether python fixed something or we fixed a few bugs in asgiref along the way… Hey @apollo13 — thanks for digging. Current status is _I don't know_. My plan (since late Summer) has been to sit down with Daphne and resolve this issue properly. Life has intervened but, it's my top item (having just got 4.0 compatibility going for django-compressor). So reality is more work (from me) on the Channels Trio in the new year now. However, if you're digging actively I am very happy to work with you on that. > That said I would recommend against reverting because creating our own event loop and configuring it as needed sounds like the way to go… Yes, it seemed like a good idea. The issue was that it turns out that you need to set the event loop in your thread if you're not using the default, and whilst that's possible it creates a bookkeeping overhead, that folks don't want to do. I began a conversation with @orf on the Forum about not using thread for the auto-reloader that was prompted by this issue. https://forum.djangoproject.com/t/feasibility-of-using-subprocess-for-auto-reloader/4970 — That's the way I'd like to go (but that's a lot of Yak...) Hi @carltongibson, while it is true that changes to the autoreloader might be possible; I still think that `ASGI_THREADS` doesn't belong into `asgiref` and we should fix that independent of further improvements in the autoreloader, ie what I outlined in https://github.com/django/daphne/issues/319#issuecomment-991962381 -- is that something we would be okay with or a big nogo (I am asking because this is basically the first time doing anything with asgi and as such I am not very up2date on it's history) > what I outlined in #319 (comment) Certainly happy for Daphne to have that. There needs to be some hard upper limit somewhere, and the protocol server seems the right level for that. (I can't recall without looking if there are other uses of `ASGI_THREADS` in asgiref that would allow dropping it entirely there.) > Nevertheless I'd be interested to know if you have an idea whether python fixed something or we fixed a few bugs in asgiref along the way… OK, so I finally had the combination of time/energy/health to look into this (and the related cluster of issues) again. The bottom-line is that I think dropping PY36, and `get_event_loop()` in favour of `get_running_loop()` across the board (and particularly in `asgiref.sync`) means that this issue is likely _Gone Away™_ (for some value of that). * Using `get_running_loop()` will get the right event loop, generally * We already call `set_event_loop` in the `Server`, so making sure we did similar in _Worker_ scripts would likely be enough. * Given that, it may be that we don't need to worry about the reloader running in a sub-thread (not sure yet) * But I tried deferring the twisted reactor setup in server.py, which doesn't look possible. * So, maybe, I can rework the Channels `runserver` to run with a subprocess if that's still needed. That's very preliminary. (I only had the session today to look at it) But maybe the pandemic induced don't look at it for two years has paid off. I'll keep working on this, but it feels like what was intractable now at least has a way of being untangled. OK, so step one is RFC to remove ASGI_THREADS from asgiref: https://github.com/django/asgiref/pull/336 Then we'll bring similar in here to restore this functionality to Daphne.	2022-07-06T10:27:51	[]	[]
django/daphne	django__daphne-417	eae1ff0df4603843d412feb9f0a27dff5f76c471	diff --git a/.github/workflows/pre-commit.yml b/.github/workflows/pre-commit.yml deleted file mode 100644 index d9394321..00000000 --- a/.github/workflows/pre-commit.yml +++ /dev/null @@ -1,24 +0,0 @@ -name: pre-commit - -on: - push: - branches: - - main - pull_request: - -jobs: - pre-commit: - runs-on: ubuntu-20.04 - - steps: - - uses: actions/checkout@v2 - with: - fetch-depth: 0 - - - uses: actions/setup-python@v2 - with: - python-version: 3.9 - - - uses: pre-commit/[email protected] - with: - token: ${{ secrets.GITHUB_TOKEN }} diff --git a/.pre-commit-config.yaml b/.pre-commit-config.yaml index 5de3ae5a..1367879b 100644 --- a/.pre-commit-config.yaml +++ b/.pre-commit-config.yaml @@ -1,20 +1,20 @@ repos: - repo: https://github.com/asottile/pyupgrade - rev: v2.11.0 + rev: v2.32.1 hooks: - id: pyupgrade args: [--py36-plus] - repo: https://github.com/psf/black - rev: 20.8b1 + rev: 22.3.0 hooks: - id: black language_version: python3 - repo: https://github.com/pycqa/isort - rev: 5.8.0 + rev: 5.10.1 hooks: - id: isort - repo: https://github.com/PyCQA/flake8 - rev: 3.9.0 + rev: 4.0.1 hooks: - id: flake8 additional_dependencies: diff --git a/CHANGELOG.txt b/CHANGELOG.txt index 8463751a..7e6a1207 100644 --- a/CHANGELOG.txt +++ b/CHANGELOG.txt @@ -1,3 +1,16 @@ +Unreleased +---------- + +* Dropped support for Python 3.6. + +* Updated dependencies to the latest versions. + + Previously a range of Twisted versions have been supported. Recent Twisted + releases (22.2, 22.4) have issued security fixes, so those are now the + minimum supported version. Given the stability of Twisted, supporting a + range of versions does not represent a good use of maintainer time. Going + forward the latest Twisted version will be required. + 3.0.2 (2021-04-07) ------------------ diff --git a/README.rst b/README.rst index 7525b27c..70c4ba58 100644 --- a/README.rst +++ b/README.rst @@ -108,7 +108,7 @@ should start with a slash, but not end with one; for example:: Python Support -------------- -Daphne requires Python 3.6 or later. +Daphne requires Python 3.7 or later. Contributing diff --git a/setup.cfg b/setup.cfg index ddb3d5de..e7c19e55 100644 --- a/setup.cfg +++ b/setup.cfg @@ -8,3 +8,4 @@ profile = black [tool:pytest] testpaths = tests +asyncio_mode = strict diff --git a/setup.py b/setup.py index 214df0d0..274e8f89 100755 --- a/setup.py +++ b/setup.py @@ -22,8 +22,8 @@ package_dir={"twisted": "daphne/twisted"}, packages=find_packages() + ["twisted.plugins"], include_package_data=True, - install_requires=["twisted[tls]>=19.7", "autobahn>=0.18", "asgiref>=3.2.10,<4"], - python_requires=">=3.6", + install_requires=["twisted[tls]>=22.4", "autobahn>=22.4.2", "asgiref>=3.5.2,<4"], + python_requires=">=3.7", setup_requires=["pytest-runner"], extras_require={"tests": ["hypothesis", "pytest", "pytest-asyncio"]}, entry_points={ @@ -37,7 +37,6 @@ "Operating System :: OS Independent", "Programming Language :: Python", "Programming Language :: Python :: 3", - "Programming Language :: Python :: 3.6", "Programming Language :: Python :: 3.7", "Programming Language :: Python :: 3.8", "Programming Language :: Python :: 3.9", diff --git a/tox.ini b/tox.ini index da7d4d9f..876ff995 100644 --- a/tox.ini +++ b/tox.ini @@ -1,12 +1,8 @@ [tox] envlist = - py{36,37,38,39,310}-twisted{197,latest} - + py{37,38,39,310} [testenv] usedevelop = true extras = tests commands = pytest -v {posargs} -deps = - twisted187: twisted==19.7.0 - twistedlatest: twisted>=20.3.0	pre-commit failing It looks like the pre-commit GitHub action is failing to install setuptools properly: https://github.com/django/daphne/actions/workflows/pre-commit.yml The action is deprecated: https://github.com/pre-commit/action I suggest we switch to using https://pre-commit.ci	Turns out I have permission to do this - but I don't want to just commit (heh) to it, especially since the issue I opened on Django ( https://github.com/django/django/pull/15178 ) hasn't seen any response. Any thoughts @carltongibson ? > Any thoughts @carltongibson ? 😝 Little bit sceptical about the value-add for django/django but if you want to give it a run here, that's fine I guess. It may be a raging success and convert the sceptics. Thanks @adamchainz Hey @adamchainz -- shall we give this a go here, and on Channels and channels_redis? Gives us a fair taste, independently of django/django	2022-05-23T10:17:05	[]	[]
astronomer/astronomer-cosmos	astronomer__astronomer-cosmos-1097	5b4c5d0b454ea207dea2ce0402d09193afe3930c	diff --git a/cosmos/cache.py b/cosmos/cache.py index d519eaca4..32a21d30d 100644 --- a/cosmos/cache.py +++ b/cosmos/cache.py @@ -230,9 +230,12 @@ def _create_folder_version_hash(dir_path: Path) -> str: filepaths.extend(paths) for filepath in sorted(filepaths): - with open(str(filepath), "rb") as fp: - buf = fp.read() - hasher.update(buf) + try: + with open(str(filepath), "rb") as fp: + buf = fp.read() + hasher.update(buf) + except FileNotFoundError: + logger.warning(f"The dbt project folder contains a symbolic link to a non-existent file: {filepath}") return hasher.hexdigest()	Bug on Cosmos 1.5 `No such file or directory "/usr/local/aiflow/dags/dbt/dbt_venv/bin/python"` Two Astro customers reported that when updating to Cosmos 1.5, they started seeing the error: ``` No such file or directory "/usr/local/aiflow/dags/dbt/dbt_venv/bin/python" ``` As illustrated in the two screenshots below: ![Image](https://github.com/user-attachments/assets/148fc806-c112-42f6-ace3-958e2386e7fa) ![Image](https://github.com/user-attachments/assets/859eb89e-8157-4d93-8492-052f4a7b1500)	After lots of time trying to reproduce and understand this issue, I narrowed it down: - inside the customers dbt project folder, there was a symbolic link to a no longer existent file Cosmos was raising an exception while trying to parse the file and create a version hash to represent the version of that dir. I validated a possible solution with one of the customers earlier today, available in Cosmos 1.5.1a3 - and they confirmed it solved the problem. I'll make a quick PR with the solution and unittests today	2024-07-17T08:40:30	[]	[]
it-is-me-mario/MARIO	it-is-me-mario__MARIO-121	8897b6963fd484f50c6b7a1ed5b5e1f20493ef46	diff --git a/CHANGELOG.rst b/CHANGELOG.rst index 931f81e..d3cf7b5 100644 --- a/CHANGELOG.rst +++ b/CHANGELOG.rst @@ -2,11 +2,27 @@ Release History **************** +v 0.3.4 +------- + +Parsing functions error fixed +~~~~~~~~ +Recent pandas versions have changed the way they interpret "None" in DataFrames indices and values, which are currently interpreted it as NaN. +This mario update fixes the issue by replacing NaN with the string "None" when parsing excel files. + +Deprecated functions +~~~~~~~~ +Parser for old-fashioned Eurostat SUTs is deprecated. This function relied on peculiarly structured SUTs formats. +In case you need to parse such SUTs, please rearrange them into the standard MARIO format. +You can check the MARIO format from 'SUT.xlsx' file in the mario/test directory in this repository. + + v 0.3.3 ------- Settings ~~~~~~~~ +to_excel function bug in flow mode fixed. v 0.3.0 @@ -54,4 +70,4 @@ Documentation ~~~~~~~~~~~~~ * The tutorials are updated to improve the readiblity and quality of the juputer notebook functionalities. -* New templates for the readthedocs. +* New templates for the readthedocs. \ No newline at end of file diff --git a/doc/source/api_refernces.rst b/doc/source/api_refernces.rst index 6efada1..939b047 100644 --- a/doc/source/api_refernces.rst +++ b/doc/source/api_refernces.rst @@ -100,7 +100,6 @@ Structured Databases mario supports automatic parsing of following database: * Exiobase -* Eurostat Supply and Use * Eora26 * Eora single region @@ -109,7 +108,6 @@ mario supports automatic parsing of following database: mario.parse_exiobase_3 mario.parse_exiobase_sut - mario.parse_eurostat_sut mario.parse_eora mario.parse_from_pymrio mario.parse diff --git a/doc/source/conf.py b/doc/source/conf.py index 6efda1a..fe37a8a 100644 --- a/doc/source/conf.py +++ b/doc/source/conf.py @@ -26,7 +26,7 @@ author = 'Mohammad Amin Tahavori, Lorenzo Rinaldi, Nicolo Golinucci' # The full version, including alpha/beta/rc tags -release = '0.3.1' +release = '0.3.4' # -- General configuration --------------------------------------------------- diff --git a/mario/__init__.py b/mario/__init__.py index 6080439..67d2ba0 100644 --- a/mario/__init__.py +++ b/mario/__init__.py @@ -43,6 +43,9 @@ """ + +import pandas as pd + from mario.version import __version__ from mario.core.AttrData import Database from mario.core.CoreIO import CoreModel @@ -99,4 +102,5 @@ ) from mario.tools.constants import IOT, SUT + __authors__ = " 'Mohammad Amin Tahavori', Lorenzo Rinaldi', 'Nicolò Golinucci' " diff --git a/mario/core/AttrData.py b/mario/core/AttrData.py index a475785..3a19cb3 100644 --- a/mario/core/AttrData.py +++ b/mario/core/AttrData.py @@ -477,6 +477,7 @@ def read_aggregated_index( # Reading the aggregation data from a single Excel file with differetn sheets ''' index = pd.read_excel(io, sheet_name=level, index_col=0) + index.index = index.index.fillna("None") if index.shape[1] > 1: index = index.iloc[:, 0].to_frame() @@ -1159,7 +1160,6 @@ def to_excel( path=None, flows=True, coefficients=False, - units=True, scenario="baseline", include_meta=False, ): @@ -1190,10 +1190,6 @@ def to_excel( if True, in the Excel file, a sheet will be created named coefficients containing the data of the coefficients - units : boolean - if True, in the Excel file, a sheet will be created named units containing - the data of the units - scenario : str defines the scenario to print out the data @@ -1209,17 +1205,23 @@ def to_excel( ) ) + if flows==False and coefficients==False: + raise WrongInput("At least one of the flows or coefficients should be True") + database_excel( self, flows, coefficients, self._getdir(path, "Database", "New_Database.xlsx"), - units, scenario, ) if include_meta: meta = self.meta._to_dict() - with open(self._getdir(path, "Database", "") + "/metadata.json", "w") as fp: + meta_path = self._getdir(path, "Database", "") + meta_path = meta_path.split("/")[:-1] + meta_path = ('/').join(meta_path) + "/metadata.json" + + with open(meta_path, "w") as fp: json.dump(meta, fp) def to_txt( @@ -1227,7 +1229,6 @@ def to_txt( path=None, flows=True, coefficients=False, - units=True, scenario="baseline", _format="txt", include_meta=False, @@ -1270,6 +1271,7 @@ def to_txt( sep : str txt file separator """ + if scenario not in self.scenarios: raise WrongInput( "{} is not a valid scenario. Existing scenarios are {}".format( @@ -1277,12 +1279,14 @@ def to_txt( ) ) + if flows==False and coefficients==False: + raise WrongInput("At least one of the flows or coefficients should be True") + database_txt( self, flows, coefficients, self._getdir(path, "Database", ""), - units, scenario, _format, sep, diff --git a/mario/core/CoreIO.py b/mario/core/CoreIO.py index 9351aef..92931a1 100644 --- a/mario/core/CoreIO.py +++ b/mario/core/CoreIO.py @@ -135,6 +135,10 @@ def __init__( for item in ["matrices", "units", "_indeces"]: setattr(self, item, kwargs["init_by_parsers"][item]) + nan_keys = [key for key, df in renamed_matrices.items() if isinstance(df, pd.DataFrame) and df.isna().any().any()] + if nan_keys: + raise ValueError(f"NaN values found in the following matrices: {nan_keys}") + log_time(logger, "Metadata: initialized.") self.meta._add_attribute(table=table, price=price, source=source, year=year) diff --git a/mario/tools/constants.py b/mario/tools/constants.py index bd6e148..2b58ca8 100644 --- a/mario/tools/constants.py +++ b/mario/tools/constants.py @@ -11,6 +11,12 @@ SUT = "SUT" IOT = "IOT" +FLOWS = "flows" +COEFFICIENTS = "coefficients" + +MONETARY = "Monetary" +HYBRID = "Hybrid" + # represents different levels of aggregation _LEVELS = { SUT: {_MASTER_INDEX[i]: i for i in ["a", "c", "f", "k", "n", "r"]}, @@ -23,6 +29,8 @@ _ACCEPTABLES = { "table": [SUT, IOT], + "mode": [FLOWS, COEFFICIENTS], + 'unit': [MONETARY, HYBRID], } diff --git a/mario/tools/excelhandler.py b/mario/tools/excelhandler.py index 4e8d479..9a5109e 100644 --- a/mario/tools/excelhandler.py +++ b/mario/tools/excelhandler.py @@ -342,7 +342,7 @@ def wrirte_matrices(sheet, Z, V, E, Y, EY, flow_format, header_format): sheet.write(row, col, 0, flow_format) -def database_excel(instance, flows, coefficients, directory, units, scenario): +def database_excel(instance, flows, coefficients, directory, scenario): file = directory workbook = xlsxwriter.Workbook(file) @@ -398,40 +398,39 @@ def database_excel(instance, flows, coefficients, directory, units, scenario): wrirte_matrices(coefficients, Z, V, E, Y, EY, coeff_format, header_format) - if units: - units = workbook.add_worksheet("units") - - data = instance.units - units.write("C1", "unit", header_format) - - counter = 2 - - if instance.table_type == "SUT": - keys = [ - _MASTER_INDEX["a"], - _MASTER_INDEX["c"], - _MASTER_INDEX["f"], - _MASTER_INDEX["k"], - ] - else: - keys = [_MASTER_INDEX["s"], _MASTER_INDEX["f"], _MASTER_INDEX["k"]] - - for key in keys: - item = data[key] - for row in range(item.shape[0]): - units.write("A{}".format(counter), key, header_format) - units.write("B{}".format(counter), item.index[row], header_format) - try: - units.write("C{}".format(counter), item.iloc[row, 0]) - except TypeError: - units.write("C{}".format(counter), "None") + units = workbook.add_worksheet("units") - counter += 1 + data = instance.units + units.write("C1", "unit", header_format) + + counter = 2 + + if instance.table_type == "SUT": + keys = [ + _MASTER_INDEX["a"], + _MASTER_INDEX["c"], + _MASTER_INDEX["f"], + _MASTER_INDEX["k"], + ] + else: + keys = [_MASTER_INDEX["s"], _MASTER_INDEX["f"], _MASTER_INDEX["k"]] + + for key in keys: + item = data[key] + for row in range(item.shape[0]): + units.write("A{}".format(counter), key, header_format) + units.write("B{}".format(counter), item.index[row], header_format) + try: + units.write("C{}".format(counter), item.iloc[row, 0]) + except TypeError: + units.write("C{}".format(counter), "None") + + counter += 1 workbook.close() -def database_txt(instance, flows, coefficients, path, units, scenario, _format, sep): +def database_txt(instance, flows, coefficients, path, scenario, _format, sep): if flows: flows = instance.query( matrices=[_ENUM.V, _ENUM.E, _ENUM.Z, _ENUM.Y, _ENUM.X, _ENUM.EY], @@ -474,38 +473,40 @@ def database_txt(instance, flows, coefficients, path, units, scenario, _format, mode="a", ) - if units: - units = copy.deepcopy(instance.units) - _units = pd.DataFrame() - _index = [] - - if instance.table_type == "SUT": - keys = [ - _MASTER_INDEX["a"], - _MASTER_INDEX["c"], - _MASTER_INDEX["f"], - _MASTER_INDEX["k"], - ] - else: - keys = [_MASTER_INDEX["s"], _MASTER_INDEX["f"], _MASTER_INDEX["k"]] - - for key in keys: - value = units[key] - _index += [key] * value.shape[0] - _units = pd.concat([_units, value]) - - _units.index = [_index, _units.index] - - if (coefficients) and (not flows): - unit_dir = "coefficients" - else: - unit_dir = "flows" + units = copy.deepcopy(instance.units) + _units = pd.DataFrame() + _index = [] + + if instance.table_type == "SUT": + keys = [ + _MASTER_INDEX["a"], + _MASTER_INDEX["c"], + _MASTER_INDEX["f"], + _MASTER_INDEX["k"], + ] + else: + keys = [_MASTER_INDEX["s"], _MASTER_INDEX["f"], _MASTER_INDEX["k"]] + for key in keys: + value = units[key] + _index += [key] * value.shape[0] + _units = pd.concat([_units, value]) + + _units.index = [_index, _units.index] + + unit_dirs = [] + if coefficients: + unit_dirs += ["coefficients"] + if flows: + unit_dirs += ["flows"] + + for unit_dir in unit_dirs: if not os.path.exists(r"{}/{}".format(path, unit_dir)): os.mkdir(r"{}/{}".format(path, unit_dir)) if os.path.exists(r"{}/{}/units.{}".format(path, unit_dir, _format)): os.remove(r"{}/{}/units.{}".format(path, unit_dir, _format)) + _units.to_csv( r"{}/{}/units.{}".format(path, unit_dir, _format), header=True, diff --git a/mario/tools/parsersclass.py b/mario/tools/parsersclass.py index 13a8fc0..c46d93d 100644 --- a/mario/tools/parsersclass.py +++ b/mario/tools/parsersclass.py @@ -4,34 +4,35 @@ """ from mario import Database +import pymrio from mario.tools.tableparser import ( eora_single_region, - txt_praser, + txt_parser, excel_parser, exio3, monetary_sut_exiobase, eora_multi_region, - eurostat_sut, parse_pymrio, hybrid_sut_exiobase_reader, parser_figaro_sut, ) from mario.log_exc.exceptions import WrongInput, LackOfInput +from mario.tools.constants import _ACCEPTABLES, _HMRSUT_EXTENSIONS +import pandas as pd models = {"Database": Database} - def parse_from_txt( - path, - table, - mode, - calc_all=False, - year=None, - name=None, - source=None, - model="Database", - sep=",", + path: str, + table: str, + mode: str, + calc_all: bool = False, + year: int = None, + name:str = None, + source:str =None, + model: str ="Database", + sep: str = ",", *kwargs, ): """Parsing database from text files @@ -74,10 +75,17 @@ def parse_from_txt( ------- mario.Database """ - if model not in models: - raise WrongInput("Available models are {}".format([models])) - matrices, indeces, units = txt_praser(path, table, mode, sep) + # check key inputs to be correct + errmsg = [] + if table not in _ACCEPTABLES['table']: + errmsg.append(f"Table should be in {_ACCEPTABLES['table']}") + if mode not in _ACCEPTABLES['mode']: + errmsg.append(f"Mode should be in {_ACCEPTABLES['mode']}") + if errmsg: + raise WrongInput(errmsg) + + matrices, indeces, units = txt_parser(path, table, mode, sep) return models[model]( name=name, @@ -91,16 +99,16 @@ def parse_from_txt( def parse_from_excel( - path, - table, - mode, - data_sheet=0, - unit_sheet="units", - calc_all=False, - year=None, - name=None, - source=None, - model="Database", + path: str, + table: str, + mode: str, + data_sheet: str = 0, + unit_sheet: str = "units", + calc_all: bool = False, + year: int = None, + name: str = None, + source: str = None, + model: str ="Database", *kwargs, ): """Parsing database from excel file @@ -147,9 +155,15 @@ def parse_from_excel( mario.Database """ - if model not in models: - raise WrongInput("Available models are {}".format([models])) - + # check key inputs to be correct + errmsg = [] + if table not in _ACCEPTABLES['table']: + errmsg.append(f"Table should be in {_ACCEPTABLES['table']}") + if mode not in _ACCEPTABLES['mode']: + errmsg.append(f"Mode should be in {_ACCEPTABLES['mode']}") + if errmsg: + raise WrongInput(errmsg) + matrices, indeces, units = excel_parser(path, table, mode, data_sheet, unit_sheet) return models[model]( @@ -164,18 +178,14 @@ def parse_from_excel( def parse_exiobase_sut( - path, - calc_all=False, - name=None, - year=None, - model="Database", + path: str, + calc_all: bool =False, + name: str = None, + year: int = None, + model: str = "Database", kwargs, ): - """Parsing exiobase mrsut - - .. note:: - - mario v.0.1.0, supports only Monetary Exiobase MRSUT database. + """Parsing Multi-Regional Supply and Use Table from Exiobase Parameters ---------- @@ -215,15 +225,15 @@ def parse_exiobase_sut( def parse_exiobase_3( - path, - calc_all=False, - year=None, - name=None, - model="Database", - version="3.8.2", + path: str, + calc_all: bool =False, + year: int = None, + name: str = None, + model: str = "Database", + version: str = "3.8.2", kwargs, ): - """Parsing exiobase3 + """Parsing Multi-Regional Input-Output Table from Exiobase .. note:: @@ -255,11 +265,14 @@ def parse_exiobase_3( mario.Database """ - if model not in models: - raise WrongInput("Available models are {}".format([models])) + # check the inputs to be correct + errmsg = [] if version not in ["3.8.2", "3.8.1"]: - raise WrongInput("Acceptable versions are {}".format(["3.8.2", "3.8.1"])) + errmsg.append("Acceptable versions are {}".format(["3.8.2", "3.8.1"])) + if errmsg: + raise WrongInput(errmsg) + matrices, indeces, units = exio3(path, version) return models[model]( @@ -274,19 +287,20 @@ def parse_exiobase_3( def parse_eora( - path, - multi_region, - table, - indeces=None, - name_convention="full_name", - aggregate_trade=True, - year=None, - name=None, - calc_all=False, - model="Database", + path: str, + multi_region: bool, + table: str, + indeces: str = None, + name_convention: str = "full_name", + aggregate_trade: bool = True, + year: int = None, + name: str = None, + calc_all: bool = False, + model: str = "Database", kwargs, ) -> object: - """Parsing eora databases + + """Parsing EORA databases .. note:: @@ -327,6 +341,7 @@ def parse_eora( ------- mario.Database """ + if model not in models: raise WrongInput("Available models are {}".format([models])) @@ -371,9 +386,17 @@ def parse_eora( def parse_exiobase( - table, unit, path, model="Database", name=None, year=None, calc_all=False, kwargs + table:str, + unit:str, + path:str, + model:str = "Database", + name:str = None, + year:int = None, + calc_all: bool = False, + kwargs ): - """A unique function for parsing all exiobase databases + + """A unique function for parsing all Exiobase databases Parameters ---------- @@ -403,23 +426,20 @@ def parse_exiobase( if non-valid values are passed to the arguments. """ - if table not in ["IOT", "SUT"]: - raise WrongInput("table only accpets 'IOT' or 'SUT'.") + if table not in _ACCEPTABLES['table']: + raise WrongInput("Table can be only chosen among {}".format(_ACCEPTABLES['table'])) - if unit not in ["Hybrid", "Monetary"]: - raise WrongInput("unit only accpets 'Hybrid' or 'Monetary.'") + if unit not in _ACCEPTABLES['unit']: + raise WrongInput("Unit con be only chosen among {}".format(_ACCEPTABLES['unit'])) if table == "IOT": if unit == "Monetary": parser = parse_exiobase_3 - else: raise WrongInput("Hybrid IOT exiobase is not supported by mario.") - else: if unit == "Monetary": parser = parse_exiobase_sut - else: parser = hybrid_sut_exiobase @@ -433,9 +453,15 @@ def parse_exiobase( def hybrid_sut_exiobase( - path, extensions=[], model="Database", name=None, calc_all=False, kwargs + path:str, + extensions: list = [], + model: str = "Database", + name: str = None, + calc_all: bool = False, + kwargs ): - """reads hybrid supply and use exiobase + """ + Parser for hybrid units Exiobase database (v3.3.18) Parameters ---------- @@ -464,8 +490,15 @@ def hybrid_sut_exiobase( For more informatio refer to https://zenodo.org/record/7244919#.Y6hEfi8w2L1 """ - if model not in models: - raise WrongInput("Available models are {}".format([models])) + + # check the inputs to be correct + errmsg = [] + if extensions != 'all': + if extensions != None: + if any([ext not in _HMRSUT_EXTENSIONS for ext in extensions]): + errmsg.append("Extensions should be chosen among {}".format(_HMRSUT_EXTENSIONS)) + if errmsg: + raise WrongInput(errmsg) matrices, indeces, units = hybrid_sut_exiobase_reader( path=path, @@ -492,11 +525,11 @@ def hybrid_sut_exiobase( def parse_eurostat_sut( - supply_path, - use_path, - model="Database", - name=None, - calc_all=False, + supply_path:str, + use_path:str, + model:str="Database", + name:str=None, + calc_all:bool=False, kwargs, ) -> object: """Parsing Eurostat databases @@ -509,7 +542,6 @@ def parse_eurostat_sut( third rule: use only "total" as stock/flow parameter, and only one unit of measure * forth rule: supply must be provided in activity by commodity, use must be provided in commodity by activitiy formats - Parameters ---------- supply_path : str @@ -527,26 +559,15 @@ def parse_eurostat_sut( mario.Database """ - if model not in models: - raise WrongInput("Available models are {}".format([models])) - - matrices, indeces, units, meta = eurostat_sut( - supply_path, - use_path, - ) - - return models[model]( - name=name, - table="SUT", - source="eurostat", - year=meta["year"], - init_by_parsers={"matrices": matrices, "_indeces": indeces, "units": units}, - calc_all=calc_all, - kwargs, - ) + raise NotImplemented("This function was deprecated since the parser was too dependent on Eurostat web interface. Downgrade to mariopy==v.3.3.3 in case you need it") -def parse_from_pymrio(io, value_added, satellite_account, include_meta=True): +def parse_from_pymrio( + io, + value_added, + satellite_account, + include_meta=True + ): """Parsing a pymrio database Parameters @@ -585,12 +606,18 @@ def parse_from_pymrio(io, value_added, satellite_account, include_meta=True): ) -def parse_FIGARO_SUT(directory, name=None, calc_all=False, kwargs): - """reads a FIGARO SUT table +def parse_FIGARO_SUT( + path:str, + name:str = None, + calc_all:bool = False, + kwargs + ): + + """Download and parse a FIGARO SUT table Parameters ---------- - directory : str + path : str the folder where the files are downloaded name : str, optional a name for the database, by default None @@ -603,7 +630,7 @@ def parse_FIGARO_SUT(directory, name=None, calc_all=False, kwargs): mario database object """ - matrices, indeces, units, year = parser_figaro_sut(directory) + matrices, indeces, units, year = parser_figaro_sut(path) return models["Database"]( name=name, @@ -614,3 +641,5 @@ def parse_FIGARO_SUT(directory, name=None, calc_all=False, kwargs): calc_all=calc_all, kwargs, ) + + diff --git a/mario/tools/tableparser.py b/mario/tools/tableparser.py index 61659b3..ca2e29c 100644 --- a/mario/tools/tableparser.py +++ b/mario/tools/tableparser.py @@ -305,8 +305,46 @@ def get_units(units, table, indeces): # if everything is ok, build a dataframe from the units. return _ +def replace_nan_indices(matrix): + for axis in [0,1]: + if axis == 0: + n_levels = matrix.index.nlevels + else: + n_levels = matrix.columns.nlevels + new_levels = [[] for i in range(n_levels)] + + for level in range(n_levels): + if axis == 0: + new_levels[level] = ["None" if pd.isna(x) else x for x in matrix.index.get_level_values(level)] + else: + new_levels[level] = ["None" if pd.isna(x) else x for x in matrix.columns.get_level_values(level)] -def txt_praser(path, table, mode, sep): + if axis == 0: + if n_levels == 1: + matrix.index = pd.Index(new_levels[0]) + else: + matrix.index = pd.MultiIndex.from_arrays(new_levels) + else: + if n_levels == 1: + matrix.columns = pd.Index(new_levels[0]) + else: + matrix.columns = pd.MultiIndex.from_arrays(new_levels) + + return matrix + +def replace_nan_units_indices(units): + + new_levels = [[],[]] + for level in range(len(units.index.names)): + for i in units.index.get_level_values(level): + if pd.isna(i) or i == '-': + i = "None" + new_levels[level].append(i) + units.index = pd.MultiIndex.from_arrays(new_levels) + units.fillna("None", inplace=True) + return units + +def txt_parser(path, table, mode, sep): if mode == "coefficients": v, e, z = list("vez") else: @@ -323,9 +361,13 @@ def txt_praser(path, table, mode, sep): log_time(logger, "Parser: Reading files finished.") _units = read["units"]["all"] + _units = replace_nan_units_indices(_units) log_time(logger, "Parser: Investigating possible identifiable errors.") + for matrix in read["matrices"].keys(): + read["matrices"][matrix] = replace_nan_indices(read["matrices"][matrix]) + indeces = get_index_txt( Z=read["matrices"][z], V=read["matrices"][v], @@ -364,6 +406,7 @@ def txt_praser(path, table, mode, sep): matrices = {"baseline": {read["matrices"]}} + return matrices, indeces, units @@ -380,6 +423,8 @@ def excel_parser(path, table, mode, sheet_name, unit_sheet): data = pd.read_excel( path, header=[0, 1, 2], index_col=[0, 1, 2], sheet_name=sheet_name ) + + data = replace_nan_indices(data) indeces = get_index_excel(data, table, mode) if table == "SUT": @@ -424,9 +469,9 @@ def excel_parser(path, table, mode, sheet_name, unit_sheet): (slice(None), _MASTER_INDEX["k"]), (slice(None), _MASTER_INDEX["n"]) ] - V.index = V.index.get_level_values(-1) - E.index = E.index.get_level_values(-1) - EY.index = EY.index.get_level_values(-1) + V.index = indeces['f']['main'] + E.index = indeces['k']['main'] + EY.index = indeces['k']['main'] if mode == "coefficients": matrices = { @@ -452,9 +497,13 @@ def excel_parser(path, table, mode, sheet_name, unit_sheet): } # read the unit sheet from the excel file + _units = pd.read_excel(path, sheet_name=unit_sheet, index_col=[0, 1]) + _units = replace_nan_units_indices(_units) + units = get_units( - pd.read_excel(path, sheet_name=unit_sheet, index_col=[0, 1]), table, indeces + _units, table, indeces ) + rename_index(matrices["baseline"]) sort_frames(matrices["baseline"]) @@ -706,6 +755,15 @@ def eora_single_region(path, table, name_convention="full_name", aggregate_trade elif table == "SUT": Z_index = eora[_MASTER_INDEX["a"]] + eora[_MASTER_INDEX["c"]] + if not all(item in data.index.get_level_values(0) for item in Z_index): + raise WrongFormat( + f"The parsed table does not seem a {table}. Please check the table type." + ) + if 'Commodities' in data.index.get_level_values(0) and table=='IOT': + raise WrongFormat( + f"The parsed table does not seem a {table}. Please check the table type." + ) + Z = data.loc[Z_index, Z_index] Y = data.loc[Z_index, eora[_MASTER_INDEX["n"]]] V = data.loc[eora[_MASTER_INDEX["f"]], Z_index] @@ -1003,159 +1061,6 @@ def eora_multi_region(data_path, index_path, year, price): return matrices, indeces, units -def eurostat_sut( - supply_path, - use_path, -): - supply_file = pd.ExcelFile(supply_path) - use_file = pd.ExcelFile(use_path) - - supply_meta = extract_metadata_from_eurostat(supply_file) - use_meta = extract_metadata_from_eurostat(use_file) - - if "Supply table at basic prices" not in supply_meta["table"]: - raise WrongInput( - "specified supply table dataset is {}. Acceptable dataset is 'Supply table at basic prices'." - "Please refer to the documents for proper download of the dataset".format( - supply_meta["table"] - ) - ) - - if "Use table at basic prices " not in use_meta["table"]: - raise WrongInput( - "specified use table dataset is {}. Acceptable dataset is 'Use table at basic prices '." - "Please refer to the documents for proper download of the dataset".format( - use_meta["table"] - ) - ) - - if ( - supply_meta["country"] != use_meta["country"] - or supply_meta["year"] != use_meta["year"] - ): - raise WrongInput( - "there are mismatched between the country/year of supply and use datasets.\nSupply Dataset: {}\nUse Dataset:{}".format( - supply_meta, use_meta - ) - ) - - supply_data = supply_file.parse( - sheet_name=eurostat_id["supply"]["sheet_name"], - index_col=eurostat_id["supply"]["index_col"], - header=eurostat_id["supply"]["header"], - ) - - use_data = use_file.parse( - sheet_name=eurostat_id["use"]["sheet_name"], - index_col=eurostat_id["use"]["index_col"], - header=eurostat_id["use"]["header"], - ) - - # build Z_matrix - z_index_c = pd.MultiIndex.from_product( - [[supply_meta["country"]], [_MASTER_INDEX["c"]], eurostat_id["c"]] - ) - - z_index_a = pd.MultiIndex.from_product( - [[supply_meta["country"]], [_MASTER_INDEX["a"]], eurostat_id["a"]] - ) - - z_index = z_index_c.append(z_index_a) - - Z = pd.DataFrame(data=0, index=z_index, columns=z_index) - - # fill supply side - Z.loc[ - (supply_meta["country"], _MASTER_INDEX["a"], eurostat_id["a"]), - (supply_meta["country"], _MASTER_INDEX["c"], eurostat_id["c"]), - ] = supply_data.loc[eurostat_id["a"], eurostat_id["c"]].values - - # fill use side - Z.loc[ - (supply_meta["country"], _MASTER_INDEX["c"], eurostat_id["c"]), - (supply_meta["country"], _MASTER_INDEX["a"], eurostat_id["a"]), - ] = use_data.loc[eurostat_id["c"], eurostat_id["a"]].values - - # build V_matrix - V = pd.DataFrame( - data=0, index=eurostat_id["f"] + eurostat_id["c_import"], columns=z_index - ) - - # Activity VA - V.loc[ - eurostat_id["f"], (supply_meta["country"], _MASTER_INDEX["a"], eurostat_id["a"]) - ] = use_data.loc[eurostat_id["f"], eurostat_id["a"]].values - - # Commodity VA - V.loc[ - eurostat_id["c_import"], - (supply_meta["country"], _MASTER_INDEX["c"], eurostat_id["c"]), - ] = supply_data.loc[eurostat_id["c_import"], eurostat_id["c"]].values - - # Building Y matrix - Y_columns = pd.MultiIndex.from_product( - [[supply_meta["country"]], [_MASTER_INDEX["n"]], eurostat_id["n"]] - ) - Y = pd.DataFrame(data=0, index=z_index, columns=Y_columns) - - Y.loc[ - (supply_meta["country"], _MASTER_INDEX["c"], eurostat_id["c"]), - (supply_meta["country"], _MASTER_INDEX["n"], eurostat_id["n"]), - ] = use_data.loc[eurostat_id["c"], eurostat_id["n"]].values - - # Building E and EY - E = pd.DataFrame(data=0, index=["-"], columns=z_index) - EY = pd.DataFrame(data=0, index=["-"], columns=Y_columns) - - # Units - units = { - _MASTER_INDEX["a"]: pd.DataFrame( - use_meta["unit"], index=eurostat_id["a"], columns=["unit"] - ), - _MASTER_INDEX["c"]: pd.DataFrame( - use_meta["unit"], index=eurostat_id["c"], columns=["unit"] - ), - _MASTER_INDEX["f"]: pd.DataFrame( - use_meta["unit"], - index=eurostat_id["c_import"] + eurostat_id["f"], - columns=["unit"], - ), - _MASTER_INDEX["k"]: pd.DataFrame( - "None", - index=E.index, - columns=["unit"], - ), - } - - X = calc_X(Z, Y) - - for matrix in [Z, V, E, EY, Y, X]: - matrix.replace(":", 0, inplace=True) - - matrices = { - "baseline": { - "Z": Z, - "V": V, - "E": E, - "EY": EY, - "Y": Y, - "X": X, - } - } - sort_frames(matrices["baseline"]) - indeces = { - "r": {"main": Z.index.unique(0).tolist()}, - "n": {"main": Y.columns.unique(-1).tolist()}, - "k": {"main": E.index.tolist()}, - "f": {"main": V.index.tolist()}, - "a": {"main": eurostat_id["a"]}, - "c": {"main": eurostat_id["c"]}, - } - rename_index(matrices["baseline"]) - - return matrices, indeces, units, use_meta - - def parse_pymrio(io, value_added, satellite_account): """Extracts the data from pymrio in mario format""" diff --git a/mario/version.py b/mario/version.py index e19434e..334b899 100644 --- a/mario/version.py +++ b/mario/version.py @@ -1,1 +1,1 @@ -__version__ = "0.3.3" +__version__ = "0.3.4" diff --git a/requirements.txt b/requirements.txt index 68fd627..f8fd3dd 100644 --- a/requirements.txt +++ b/requirements.txt @@ -1,10 +1,10 @@ -numpy >= 1.21.2 -plotly >= 4.12.0 -tabulate >= 0.8.9 -openpyxl >= 3.0.6 -IPython >= 7.22.0 +numpy == 2.1.1 +plotly +tabulate +openpyxl == 3.1.0 +IPython >= 8.27.0 pymrio -pytest -pandas +pytest +pandas == 2.2.3 pyyaml -xlsxwriter \ No newline at end of file +xlsxwriter == 3.2.0 \ No newline at end of file diff --git a/setup.py b/setup.py index d821594..0372c30 100644 --- a/setup.py +++ b/setup.py @@ -30,15 +30,15 @@ "mario/test":[".xlsx"], }, install_requires=[ - "pandas", - "numpy >= 1.21.2", - "xlsxwriter", - "plotly >= 4.12.0", - "tabulate >= 0.8.9", - "openpyxl >= 3.0.6", - "IPython >= 7.22.0", - "pymrio >= 0.4.6", - "pyyaml >= 5.4.1" + "pandas == 2.2.3", + "numpy == 2.1.1", + "xlsxwriter == 3.2.0", + "plotly", + "tabulate", + "openpyxl == 3.1.0", + "IPython >= 8.27.0", + "pymrio", + "pyyaml" ], # classifiers=[	Fixed bugs related to Nans in indices or units New pandas versions interpret None as NaN. This raised issues in parsing databases with labels or units named "None". A check on NaNs was implemented when parsing, replacing NaNs with "None", as MARIO used to deal with since v0.2.2. "-" items where also replaced with "None" in test files (e.g. those loaded with mario.load_test), to align with the approach.		2024-10-29T06:06:57	[]	[]
it-is-me-mario/MARIO	it-is-me-mario__MARIO-45	66eafdfc1ae8e6fb470c6015895fdbaab0426a3b	diff --git a/mario/core/AttrData.py b/mario/core/AttrData.py index db9ae34..4603b8e 100644 --- a/mario/core/AttrData.py +++ b/mario/core/AttrData.py @@ -1506,6 +1506,91 @@ def add_sectors( for note in notes: self.meta._add_history(f"User note: {note}") + def query( + self, + matrices, + scenarios = ["baseline"], + base_scenario = None, + type="absolute", + ): + """ Requests a specific data from the database + + Parameters + ---------- + matrices : str + list of the matrices to return + + scenarios : str, List[str] + list of scenarios for returing the matrices + + base_scenario : str + str representing the base scenario in case that the data should be returned for the change in data between scenarios + + type: str + #. 'absolute' for absolute difference for scenarios + #. 'relative' for relative difference for scenarios + + + Returns + ------- + dict,pd.DataFrame + If multiple scenarios are passed, it returns a dict where keys are the scenarios and vals are the matrices. matrices itself could be a dict or pd.DataFrame depending if multiple matrices or one matrix is passed + + + Example + ------- + Let's consider the case that multiple scenarios ['sc.1','sc.2'] and multiple matrices ['X','z'] are passed: + + .. code-block:: python + + output = example.query(scenarios= ['sc.1','sc.2'], matrices = ['X','z']) + + the output in this case would be: + + type: Dict[Dict[pd.DataFrame]] + + { + "sc.1" : { + "X": pd.DataFrame, + "z": pd.DataFrame, + }, + "sc.2" : { + "X": pd.DataFrame, + "z": pd.DataFrame, + }, + + } + + if only one scenario ("sc.1") is passed, output will be: + + type Dict[pd.DataFrame] + { + "X": pd.DataFrame, + "Y": pd.DataFrame, + } + + if only one scenario ("sc.1") and one matrix ("X") is passed the output will be a single pd.DataFrame. + """ + data = self.get_data( + matrices=matrices, + units=False, + indeces=False, + format="dict", + scenarios=scenarios, + base_scenario = base_scenario, + type = type + ) + + + if len(matrices) == 1: + for scenario in scenarios: + data[scenario]= data[scenario][matrices[0]] + + if len(scenarios) == 1: + data = data[scenarios[0]] + + return data + def get_data( self, matrices, @@ -1523,7 +1608,7 @@ def get_data( Parameters ---------- - path : str + matrices : str list of the matrices to return units : boolean	Get_data function The function "get_data" is not working if the user provide an integer as "scenarios" parameter. Not even if it is provided within a list. Moreover, the documentation for this function seems missing.	I would add the following: The get_data function was supposed to be used as a back-hand. Probabily for this reason, it looks sub-optimal and may be improved. Best way to use it to extract a specific matrix (e.g. f) of the main object (world) is: world.get_data(matrices=['f'], scenarios=['scenario 2'])['scenario 2'][0] The function may be improved by avoiding the pass the same info twice. Furthermore, a dedicated documentation should be added. > I would add the following: The get_data function was supposed to be used as a back-hand. Probabily for this reason, it looks sub-optimal and may be improved. Best way to use it to extract a specific matrix (e.g. f) of the main object (world) is: > > world.get_data(matrices=['f'], scenarios=['scenario 2'])['scenario 2'][0] > > The function may be improved by avoiding the pass the same info twice. Furthermore, dedicated documentation should be added. @nigolred & @LorenzoRinaldi Probably the best way to solve the problem would be to have a new function with a more user-friendy function that executes the get_data function in a simpler and more intuitive way and leave the get_data function as a private function that is extensively used in the backend. Probably changing this function itself will arise backward-compatibility issues. What do you think? For me that can work for sure and may be preferable to avoid backward-compatibility.	2023-04-13T21:46:24	[]	[]
it-is-me-mario/MARIO	it-is-me-mario__MARIO-28	767ac770cbf78180e3b607247b0ce9f617393989	diff --git a/mario/core/AttrData.py b/mario/core/AttrData.py index 0d405f9..b415a39 100644 --- a/mario/core/AttrData.py +++ b/mario/core/AttrData.py @@ -19,10 +19,7 @@ from mario.tools.tabletransform import SUT_to_IOT import json from mario.tools.utilities import ( - _matrices, _manage_indeces, - _meta_parse_history, - linkages_calculation, check_clusters, run_from_jupyter, filtering, @@ -63,6 +60,7 @@ calc_f, calc_f_dis, calc_X_from_z, + linkages_calculation, ) from mario.tools.sectoradd import adding_new_sector @@ -298,17 +296,14 @@ def to_iot( matrices, indeces, units = SUT_to_IOT(self, method) for scenario in self.scenarios: - _matrices(self, "del", scenario) log_time(logger, f"{scenario} deleted from the database", "warn") self.meta._add_history(f"{scenario} deleted from the database") self.matrices = matrices - _matrices(self, "add", "baseline") self._indeces = indeces self.units = units - _meta_parse_history(self, "parse") self.meta.table = "IOT" self.meta._add_history( @@ -418,20 +413,14 @@ def read_aggregated_index( if index.shape[1] > 1: index = index.iloc[:, 0].to_frame() - if set(index.index.to_list()) != set(self.get_index(level)): - missing_items = [] + difference = set(index.index).difference(set(self.get_index(level))) - for item in set(self.get_index(level)): - if item not in index.index.to_list(): - missing_items.append(item) - - raise WrongExcelFormat( - "Disaggregated indeces of level '{}' in the Excel file " - "does not match with the original indices of the database.\nMissing items are:\n{}".format( - level, missing_items - ) + if difference: + raise WrongInput( + f"Following item are not acceptable for level {level} \n {difference}" ) + index.columns = ["Aggregation"] if index.isnull().values.any(): @@ -552,9 +541,7 @@ def aggregate( __new_matrices, units = _aggregator(self, drop) for scenario in self.scenarios: - _matrices(self, "del", scenario) self.matrices[scenario] = __new_matrices[scenario] - _matrices(self, "add", scenario) self.meta._add_history( "original matrices changed to the aggregated level based on the inputs from {}".format( @@ -569,7 +556,6 @@ def aggregate( _manage_indeces(self, "aggregation") new_index = copy.deepcopy(self._indeces) - _meta_parse_history(self, "aggregation", old_index, new_index) if calc_all: for scenario in self.scenarios: @@ -581,9 +567,7 @@ def reset_to_backup(self): All the matrices and indeces will be updated to the last back-up """ - _matrices(self, function="del") self.matrices = self._backup.matrices - _matrices(self, function="add") self._indeces = self._backup.indeces self.units = self._backup.units self.meta._add_history("Last backup recovered.") @@ -685,7 +669,7 @@ def add_extensions( if not inplace: new = self.copy() new.add_extensions( - io=io, matrix=matrix, backup=backup, inplace=True, units=units, + io=io, matrix=matrix, backup=backup, inplace=True, units=units,calc_all=calc_all,notes=notes,EY=EY ) return new @@ -740,7 +724,7 @@ def add_extensions( "units dataframe should has exactly the same index levels of io" ) - if EY is not None: + if EY is not None and matrix_id == "k": EY = EY.sort_index() if not data.index.equals(EY.index): @@ -776,16 +760,14 @@ def add_extensions( units = info["units"] del info["units"] - info["X"] = calc_X(Z=info["Z"], Y=info["Y"]) + matrices = {"baseline": {info}} for scenario in self.scenarios: - _matrices(self, "del", scenario) log_time(logger, f"{scenario} deleted from the database", "warn") self.meta._add_history(f"{scenario} deleted from the database") self.matrices = matrices - _matrices(self, "add", "baseline") self.meta._add_history( f"Modification: new '{_MASTER_INDEX[matrix_id]}' added to the database as follow:\n" @@ -946,14 +928,12 @@ def to_single_region(self, region, backup=True, inplace=True): _manage_indeces(self, "single_region", new_indeces) for scenario in self.scenarios: - _matrices(self, "del", scenario=scenario) log_time(logger, f"Transformation: {scenario} deleted from the database.") self.matrices = {"baseline": {}} for matrix in ["Y", "Z", "E", "EY", "Y", "V", "X"]: self.matrices["baseline"][matrix] = eval(matrix) log_time(logger, "Transformation: New baseline added to the database") - _matrices(self, "add", scenario="baseline") slicer = _MASTER_INDEX["a"] if self.table_type == "SUT" else _MASTER_INDEX["s"] @@ -977,16 +957,7 @@ def to_single_region(self, region, backup=True, inplace=True): "Transformation: The Final Demand emissions are considered only for 'Local Final Demand.'" ) - def backup(self): - """The function creates a backup of the last configuration of database - to be returned in case needed. - """ - self._backup = self._backup_( - copy.deepcopy(self.matrices), - copy.deepcopy(self._indeces), - copy.deepcopy(self.units), - ) def calc_linkages( self, scenario="baseline", normalized=True, cut_diag=True, multi_mode=True, @@ -1048,7 +1019,6 @@ def calc_linkages( } return linkages_calculation( - self, cut_diag=cut_diag, matrices=_matrices, multi_mode=multi_mode, @@ -1515,13 +1485,9 @@ def add_sectors( EY = self.EY # Deleting old values - _matrices(self, "del") for matrix in ["z", "e", "v", "Y", "X", "Z", "E", "V", "EY"]: self.matrices["baseline"][matrix] = eval(matrix) - # Adding new values - _matrices(self, "add") - self.meta._add_history( "Scenarios: all the scenarios deleted from the database." ) @@ -1830,7 +1796,11 @@ def shock_calc( if scenario == "baseline": raise WrongInput("baseline scenario can not be overwritten.") - check_clusters(self, clusters) + check_clusters( + index_dict = self.get_index('all'), + table = self.table_type, + clusters = clusters + ) # have the test for the existence of the database @@ -1888,7 +1858,11 @@ def get_shock_excel( e.g. clusters = {'Region':{'cluster_1':['reg1','reg2']}} """ - check_clusters(self, clusters) + check_clusters( + index_dict = self.get_index('all'), + table = self.table_type, + clusters = clusters + ) _sh_excel(self, num_shock, self._getdir(path, "Excels", "shock.xlsx"), clusters) diff --git a/mario/core/CoreIO.py b/mario/core/CoreIO.py index 391056b..d47d5f2 100644 --- a/mario/core/CoreIO.py +++ b/mario/core/CoreIO.py @@ -15,11 +15,6 @@ from mario.tools.tableparser import dataframe_parser -from mario.tools.utilities import ( - _matrices, - _meta_parse_history, -) - from mario.tools.iomath import ( calc_X, @@ -143,8 +138,6 @@ def __init__( log_time(logger, "Metadata: initialized.") self.meta._add_attribute(table=table, price=price, source=source, year=year) - _matrices(self, function="add") - else: if not all( [matrix is not None for matrix in [Y, E, Z, V, EY, units, table]] @@ -160,15 +153,12 @@ def __init__( log_time(logger, "Metadata: initialized by dataframes.") - _matrices(self, function="del") - _matrices(self, function="add") # Adding notes if passed by user or the parsers if kwargs.get("notes"): for note in kwargs["notes"]: self.meta._add_history(note) - _meta_parse_history(self, "parse") if calc_all: self.calc_all() @@ -996,19 +986,16 @@ def meta_history(self): def __getitem__(self, key): """get item method retuns the data regarding the scenarios""" - self.scenario_exist(key) - - return self.matrices[key] - - def scenario_exist(self, scenario): - """Checks if a scneario exists or not""" - if scenario not in self.scenarios: + if key not in self.scenarios: raise WrongInput( "{} is not a valid scenario. Valid scenarios are {}".format( - scenario, self.scenarios + key, self.scenarios ) ) + return self.matrices[key] + + def __iter__(self): self.__it__ = self.scenarios return self @@ -1065,34 +1052,14 @@ def __eq__(self,other): return True - def _extract_index_from_frames(self,_dict): - ids = { - 'r': {'matrix':['z','X','Y','v','e','Z',],"level":('index',0)}, - 's': {'matrix':['z','X','Y','v','e','Z',],"level":('index',-1)}, - 'a': {'matrix':['z','X','Y','v','e','Z',],"level":('index',-1)}, - 'c': {'matrix':['z','X','Y','v','e','Z',],"level":('index',-1)}, - 'n': {'matrix':['Y','EY'],"level":('columns',1)}, - 'k': {'matrix':['e','E','EY'],"level":('index',0)}, - 'f': {'matrix':['v','V','EY'],"level":('index',0)}, - } + + def backup(self): - indeces = {} - for key,vals in ids.items(): - idx = _MASTER_INDEX[key] - - if idx not in self.sets: - continue - - for matrix in vals['matrix']: - if matrix in _dict: - df = _dict[matrix] - print(idx,matrix) - - if key in ['s','a','c']: - df = df.loc[(slice(None),idx,slice(None)),:] - index = list(getattr(df,vals['level'][0]).unique(vals['level'][1])) - indeces[idx] = index - break - - return indeces - + """The function creates a backup of the last configuration of database + to be returned in case needed. + """ + self._backup = self._backup_( + copy.deepcopy(self.matrices), + copy.deepcopy(self._indeces), + copy.deepcopy(self.units), + ) \ No newline at end of file diff --git a/mario/core/CustomDicts.py b/mario/core/CustomDicts.py deleted file mode 100644 index 3c3547a..0000000 --- a/mario/core/CustomDicts.py +++ /dev/null @@ -1,47 +0,0 @@ - -from collections import UserDict - -class Matrices(UserDict): - """ - A costumized dict for having a better control on matrices property - in mario.Core - - Properties - ----------- - _bs_name : str - This represents the name of the baseline scenario - _dynamic : boolean - Shows if the database is dynamic or not - """ - def __init__(self,baseline_name,dynamic=False,vals={}): - self._bs_name = baseline_name - self._dynamic = dynamic - - super().__init__(vals) - - def __setitem__(self, key, item): - if key == 'baseline': - key = self._bs_name - - if self._dynamic and not isinstance(key,int): - raise ValueError('For a dynamic database, only integer keys are accepted.') - - if not isinstance(item,dict): - raise ValueError('item should be a dict.') - - - super().__setitem__(key, item) - - def __getitem__(self, key): - - if key == 'baseline': - key = self._bs_name - - return super().__getitem__(key) - -if __name__ == '__main__': - test = Matrices(2020,True) - test['baseline'] = {} - #%% - -# %% diff --git a/mario/tools/iomath.py b/mario/tools/iomath.py index e5259f0..d5f678e 100644 --- a/mario/tools/iomath.py +++ b/mario/tools/iomath.py @@ -6,6 +6,11 @@ import numpy as np from copy import deepcopy as dc +from mario.log_exc.logger import log_time +import logging + +logger = logging.getLogger(__name__) + def calc_all_shock(z, e, v, Y): @@ -470,3 +475,97 @@ def X_inverse(X): X_inv[X_inv != 0] = 1 / X_inv[X_inv != 0] return X_inv + + + + +def linkages_calculation(cut_diag, matrices, multi_mode, normalized): + """calculates the linkages""" + if cut_diag: + for key, value in matrices.items(): + np.fill_diagonal(value.values, 0) + + if multi_mode: + + link_types = [ + "Total Forward", + "Total Backward", + "Direct Forward", + "Direct Backward", + ] + geo_types = ["Local", "Foreign"] + links = pd.DataFrame( + 0, + index=matrices["g"].index, + columns=pd.MultiIndex.from_product([link_types, geo_types]), + ) + + for index, values in links.iterrows(): + links.loc[index, ("Total Forward", "Local")] = ( + matrices["g"].loc[index, index[0]].sum().sum() + ) + links.loc[index, ("Total Forward", "Foreign")] = ( + matrices["g"].loc[index].sum().sum() + - matrices["g"].loc[index, index[0]].sum().sum() + ) + + links.loc[index, ("Total Backward", "Local")] = ( + matrices["w"].loc[index, index[0]].sum().sum() + ) + links.loc[index, ("Total Backward", "Foreign")] = ( + matrices["w"].loc[index].sum().sum() + - matrices["w"].loc[index, index[0]].sum().sum() + ) + + links.loc[index, ("Direct Forward", "Local")] = ( + matrices["b"].loc[index, index[0]].sum().sum() + ) + links.loc[index, ("Direct Forward", "Foreign")] = ( + matrices["b"].loc[index].sum().sum() + - matrices["b"].loc[index, index[0]].sum().sum() + ) + + links.loc[index, ("Direct Backward", "Local")] = ( + matrices["z"].loc[index, index[0]].sum().sum() + ) + links.loc[index, ("Direct Backward", "Foreign")] = ( + matrices["z"].loc[index].sum().sum() + - matrices["z"].loc[index, index[0]].sum().sum() + ) + + if normalized: + log_time( + logger, "Normalization not available for multi-regional mode.", "warn" + ) + + # Computing linkages as if there were only one unique region + else: + _forward_t = matrices["g"].sum(axis=1).to_frame() + _backward_t = matrices["w"].sum(axis=0).to_frame() + _forward_d = matrices["b"].sum(axis=1).to_frame() + _backward_d = matrices["z"].sum(axis=0).to_frame() + + _forward_t.columns = ["Total Forward"] + _backward_t.columns = ["Total Backward"] + _forward_d.columns = ["Direct Forward"] + _backward_d.columns = ["Direct Backward"] + + if normalized: + + _forward_t.iloc[:, 0] = _forward_t.iloc[:, 0] / np.average( + _forward_t.values + ) + _backward_t.iloc[:, 0] = _backward_t.iloc[:, 0] / np.average( + _backward_t.values + ) + _forward_d.iloc[:, 0] = _forward_d.iloc[:, 0] / np.average( + _forward_d.values + ) + _backward_d.iloc[:, 0] = _backward_d.iloc[:, 0] / np.average( + _backward_d.values + ) + + links = pd.concat([_forward_t, _backward_t, _forward_d, _backward_d], axis=1) + + return links + diff --git a/mario/tools/plots.py b/mario/tools/plots.py index 09bb2eb..c8cb9a0 100644 --- a/mario/tools/plots.py +++ b/mario/tools/plots.py @@ -9,9 +9,7 @@ from mario.tools.plots_manager import _PLOTS_LAYOUT, Color, _PALETTES from mario.tools.utilities import ( - subplot_grid, run_from_jupyter, - unit_check, ) import plotly.graph_objects as go @@ -20,7 +18,6 @@ import plotly.offline as pltly import pandas as pd import logging -import webbrowser logger = logging.getLogger(__name__) diff --git a/mario/tools/tableparser.py b/mario/tools/tableparser.py index 6f398b6..e226397 100644 --- a/mario/tools/tableparser.py +++ b/mario/tools/tableparser.py @@ -1170,7 +1170,7 @@ def parse_pymrio(io, value_added, satellite_account): """ # be sure that system is calculated - io = io.calc_system() + io = io.calc_all() extensions = {} for value in dir(io): diff --git a/mario/tools/utilities.py b/mario/tools/utilities.py index 04e212e..e66a24d 100644 --- a/mario/tools/utilities.py +++ b/mario/tools/utilities.py @@ -2,6 +2,7 @@ """ contains the utils functions in mario """ +#%% import pandas as pd import numpy as np import logging @@ -85,7 +86,8 @@ def run_from_jupyter(): ipy_str = str(type(get_ipython())) if "zmqshell" in ipy_str: return True - return False + + return False def sort_frames(_dict): @@ -99,88 +101,10 @@ def sort_frames(_dict): _dict[key] = value.sort_index(axis=1, level=1).sort_index(axis=0, level=1) -def validate_path(path: [str, list]): - - if isinstance(path, str): - path = [path] - - for item in path: - if not os.path.exists(item): - raise FileNotFoundError(f"{item} does not exist.") - - -def unique_frmaes(frame: pd.DataFrame,) -> list: - - all_items = [] - for col, values in frame.iteritems(): - given_values = delete_duplicates(values.values) - all_items.extend(given_values) - - return delete_duplicates(all_items) - - def delete_duplicates(_list: [list, tuple]) -> list: return list(dict.fromkeys(_list)) -def index_col_equlity( - original: pd.DataFrame, given: pd.DataFrame, levels: list = ["index", "columns"] -): - - for level in levels: - if not getattr(original, level).equals(getattr(given, level)): - raise WrongExcelFormat(f"{level} does not have the correct format.") - - -def _meta_parse_history(instance, function, old_index=None, new_index=None): - - if function == "parse": - instance.meta._add_history("Database successfully imported.") - for index in instance._indeces.keys(): - instance.meta._add_history( - "Number of {} = {}".format( - _MASTER_INDEX[index], len(instance._indeces[index]["main"]) - ) - ) - - elif function == "aggregation": - for index, value in instance._indeces.items(): - instance.meta._add_history( - "{} aggregated from {} levels to {} levels".format( - _MASTER_INDEX[index], - len(old_index[index]["main"]), - len(new_index[index]["main"]), - ) - ) - - -def _matrices(instance, function, scenario="baseline"): - """ - This function is in charge of deleting or adding the matrices as a callable - attribute to the instance - """ - # acceptable_matrices=_ALL_MATRICES+_SUT_MATRICES - - if function == "del": - - for attribute in _ALL_MATRICES[instance.table_type]: - if scenario != "baseline": - attribute = "{}_c".format(attribute) - try: - delattr(instance, attribute) - except AttributeError: - pass - - if function == "add": - pass - # --TODO-- check if it is necessary to have it - # for attribute,value in instance.matrices[scenario].items(): - # if scenario != 'baseline': - # attribute = '{}_c'.format(attribute) - - # setattr(instance, attribute,value) - - def _manage_indeces(instance, case, kwargs): if case == "aggregation": @@ -197,89 +121,26 @@ def _manage_indeces(instance, case, kwargs): for key, value in kwargs.items(): instance._indeces[key] = {"main": value} +def check_clusters(index_dict,table,clusters): -def subplot_grid(subplot_number, orientation="v"): - - if orientation == "v": - j = 0 - n_cols = [] - for i in reversed(range(subplot_number + 1)): - if int(math.sqrt(i) + 0.5) 2 == i: - n_cols += [int(math.sqrt(i))] - j += 1 - n_cols = n_cols[0] - - if int(math.sqrt(subplot_number) + 0.5) 2 == subplot_number: - n_rows = n_cols - else: - n_rows = n_cols + int(math.ceil((subplot_number - n_cols 2) / n_cols)) - - elif orientation == "h": - j = 0 - n_rows = [] - for i in reversed(range(subplot_number + 1)): - if int(math.sqrt(i) + 0.5) 2 == i: - n_rows += [int(math.sqrt(i))] - j += 1 - n_rows = n_rows[0] - - if int(math.sqrt(subplot_number) + 0.5) 2 == subplot_number: - n_cols = n_rows - else: - n_cols = n_rows + int(math.ceil((subplot_number - n_rows 2) / n_rows)) - - grid = [(row + 1, col + 1) for row in range(n_rows) for col in range(n_cols)] - - return (n_rows, n_cols, grid) - + differences = set(clusters).difference(set(_LEVELS[table])) -def str_2_list(args): - - "this fucntion returns a list in case that the user pass a string" - - output = [] - for arg in args: - if isinstance(arg, str): - output.append([arg]) - else: - output.append(arg) - - return output - - -def master_exist(instance, kwargs): - """ - 'this function checks if a specific item (any of the items from the MASTER_INDEX are valid or not' - instance = self - for kwargs: - key = the key of the MASTER_INDEX - value = a list of the items that the user is giving - """ - - for key, values in kwargs.items(): - for value in values: - if value not in instance.get_index(_MASTER_INDEX[key]): - raise WrongInput(f"{value} is not a valid {_MASTER_INDEX[key]}.") - - -def check_clusters(instance, clusters): + if differences: + raise WrongInput( + "{} is/are not valid level/s. Valid items are {}".format( + differences, [_LEVELS[table]] + ) + ) for level, level_cluster in clusters.items(): - if level not in _LEVELS[instance.meta.table]: - raise WrongInput( - "{} is not a valid level. Valid items are {}".format( - level, [*_LEVELS[instance.meta.table]] - ) - ) for cluster, values in level_cluster.items(): - for value in values: - if value not in instance.get_index(level): - raise WrongInput( - "{} in cluster {} for level {} is not a valid item.".format( - value, cluster, level - ) + differences = set(values).difference(set(index_dict[level])) + if differences: + raise WrongInput( + "{} in cluster {} for level {} is/are not a valid item/s.".format( + differences, cluster, level ) - + ) def all_file_reader( path, guide, sub_folder=False, sep="\t", exceptions=[], engine=None @@ -344,15 +205,16 @@ def readers(file_to_read, file): return read -def return_index(df, item, multi_index, del_duplicate, reindex, level=None): +def return_index(df, item, multi_index, del_duplicate, reindex=None, level=None): if multi_index: - index = eval("df.{}.get_level_values({})".format(item, level)) + index = list(getattr(df,item).get_level_values(level)) + else: - index = eval("df.{}".format(item)) + index = list(getattr(df,item)) if del_duplicate: - index = delete_duplicates(list(index)) + index = delete_duplicates(index) return index @@ -368,7 +230,7 @@ def multiindex_contain(inner_index, outer_index, file, check_levels=None): if check_levels is None: if inner_index.nlevels != outer_index.nlevels: raise WrongInput(f"number levels for {file} are not valid.") - levels = [(i, i) for i in range(inner_index.nlevels)] + levels = list(range(inner_index.nlevels))# [i for i in range(inner_index.nlevels)] else: levels = check_levels @@ -377,12 +239,12 @@ def multiindex_contain(inner_index, outer_index, file, check_levels=None): for level in levels: diff = ( - outer_index.levels[level[1]] - .difference(inner_index.levels[level[0]]) + outer_index.levels[level] + .difference(inner_index.levels[level]) .tolist() ) - differences[level[1]] = diff + differences[level] = diff if len(diff): passed = False @@ -399,118 +261,11 @@ def rename_index(_dict): for key, value in _dict.items(): for item in ["index", "columns"]: - if isinstance(eval(f"value.{item}"), pd.MultiIndex): - exec(f"value.{item}.names = _INDEX_NAMES['3levels']") + if isinstance(getattr(value,item), pd.MultiIndex): + getattr(value,item).names= _INDEX_NAMES['3levels'] else: - exec(f"value.{item}.name = _INDEX_NAMES['1level']") - - -def linkages_calculation(instance, cut_diag, matrices, multi_mode, normalized): - """calculates the linkages""" - if cut_diag: - for key, value in matrices.items(): - np.fill_diagonal(value.values, 0) - - if multi_mode: - - link_types = [ - "Total Forward", - "Total Backward", - "Direct Forward", - "Direct Backward", - ] - geo_types = ["Local", "Foreign"] - links = pd.DataFrame( - 0, - index=matrices["g"].index, - columns=pd.MultiIndex.from_product([link_types, geo_types]), - ) - - for index, values in links.iterrows(): - links.loc[index, ("Total Forward", "Local")] = ( - matrices["g"].loc[index, index[0]].sum().sum() - ) - links.loc[index, ("Total Forward", "Foreign")] = ( - matrices["g"].loc[index].sum().sum() - - matrices["g"].loc[index, index[0]].sum().sum() - ) - - links.loc[index, ("Total Backward", "Local")] = ( - matrices["w"].loc[index, index[0]].sum().sum() - ) - links.loc[index, ("Total Backward", "Foreign")] = ( - matrices["w"].loc[index].sum().sum() - - matrices["w"].loc[index, index[0]].sum().sum() - ) - - links.loc[index, ("Direct Forward", "Local")] = ( - matrices["b"].loc[index, index[0]].sum().sum() - ) - links.loc[index, ("Direct Forward", "Foreign")] = ( - matrices["b"].loc[index].sum().sum() - - matrices["b"].loc[index, index[0]].sum().sum() - ) + getattr(value,item).name = _INDEX_NAMES['1level'] - links.loc[index, ("Direct Backward", "Local")] = ( - matrices["z"].loc[index, index[0]].sum().sum() - ) - links.loc[index, ("Direct Backward", "Foreign")] = ( - matrices["z"].loc[index].sum().sum() - - matrices["z"].loc[index, index[0]].sum().sum() - ) - - if normalized: - log_time( - logger, "Normalization not available for multi-regional mode.", "warn" - ) - - # Computing linkages as if there were only one unique region - else: - _forward_t = matrices["g"].sum(axis=1).to_frame() - _backward_t = matrices["w"].sum(axis=0).to_frame() - _forward_d = matrices["b"].sum(axis=1).to_frame() - _backward_d = matrices["z"].sum(axis=0).to_frame() - - _forward_t.columns = ["Total Forward"] - _backward_t.columns = ["Total Backward"] - _forward_d.columns = ["Direct Forward"] - _backward_d.columns = ["Direct Backward"] - - if normalized: - - _forward_t.iloc[:, 0] = _forward_t.iloc[:, 0] / np.average( - _forward_t.values - ) - _backward_t.iloc[:, 0] = _backward_t.iloc[:, 0] / np.average( - _backward_t.values - ) - _forward_d.iloc[:, 0] = _forward_d.iloc[:, 0] / np.average( - _forward_d.values - ) - _backward_d.iloc[:, 0] = _backward_d.iloc[:, 0] / np.average( - _backward_d.values - ) - - links = pd.concat([_forward_t, _backward_t, _forward_d, _backward_d], axis=1) - - return links - - -def unit_check(instance, sets, slicer=None): - - if sets == _MASTER_INDEX["r"]: - raise WrongInput(f"Set '{sets}' do not have any unit") - units = instance.units[sets] - if slicer is not None: - units = units.loc[slicer, :] - uniques = {} - - unique_units = units["unit"].unique() - - for unit in unique_units: - uniques[unit] = units.index[units["unit"] == unit].tolist() - - return uniques def filtering(instance, filters): @@ -572,3 +327,5 @@ def to_single_index(df): df.index = [", ".join(ii) for ii in df.index] return df + +# %%	Depricating the _meta_parse_history Avoinding the meta_parse_history by recording the number of items in a dataset		2022-04-11T16:04:21	[]	[]
CybersecurityForDemocracy/tiktok-library	CybersecurityForDemocracy__tiktok-library-92	0302e5ac4e1e8f02f8ee2a0304936a183eae603c	diff --git a/src/tiktok_research_api_helper/api_client.py b/src/tiktok_research_api_helper/api_client.py index 5301d9d..13c6b2f 100644 --- a/src/tiktok_research_api_helper/api_client.py +++ b/src/tiktok_research_api_helper/api_client.py @@ -40,6 +40,7 @@ API_ERROR_RETRY_MAX_WAIT = timedelta(minutes=2).total_seconds() DAILY_API_REQUEST_QUOTA = 1000 +NULL_CHARACTERS_TO_STRIP = "\x00\u0000" class ApiRateLimitError(Exception): @@ -629,6 +630,8 @@ def _parse_video_response(response: rq.Response) -> TikTokVideoResponse: error_data = response_json.get("error") response_data_section = response_json.get("data", {}) videos = response_data_section.get("videos", []) + for video in videos: + strip_null_chars_from_json_values(video) return TikTokVideoResponse(data=response_data_section, videos=videos, error=error_data) @@ -637,6 +640,7 @@ def _parse_user_info_response(username: str, response: rq.Response) -> TikTokUse response_json = _extract_response_json_or_raise_error(response) error_data = response_json.get("error") response_data_section = response_json.get("data") + strip_null_chars_from_json_values(response_data_section) # API does not include username in response. for ease of use we add it. response_data_section["username"] = username @@ -652,11 +656,25 @@ def _parse_comments_response(response: rq.Response) -> TikTokCommentsResponse: response_json = _extract_response_json_or_raise_error(response) error_data = response_json.get("error") response_data_section = response_json.get("data", {}) + # TODO(macpd): this might have an issue with null characters in usernames comments = response_data_section.get("comments") + strip_null_chars_from_json_values(comments) return TikTokCommentsResponse(comments=comments, data=response_data_section, error=error_data) +def strip_null_chars_from_json_values(json_response: Mapping[str, Any]): + for key in json_response: + if isinstance(json_response[key], str): + json_response[key] = json_response[key].strip(NULL_CHARACTERS_TO_STRIP) + # Could do this recursively, but don't want a infinite recursion bug + elif isinstance(json_response[key], list): + json_response[key] = [ + v.strip(NULL_CHARACTERS_TO_STRIP) if isinstance(v, str) else v + for v in json_response[key] + ] + + def _extract_response_json_or_raise_error(response: rq.Response \| None) -> Mapping[str, Any]: if response is None: raise ValueError("Response is None")	Some values from API have null byte/char, and therefore cannot be inserted to database directly for example video ID `7420068284780875054` has `'username': '٠вяеииа٠ء\x00\x00\x00\x00\x00\x00\x00\x00'` I don't know if we should strip the null character from these strings (`str.replace('\x00', '')`), or unicode escape them (`str.encode('unicode_escape')`)		2024-10-15T16:52:13	[]	[]
ec-jrc/Thalassa	ec-jrc__Thalassa-91	b9d977cd6999e73f5ad884e9e7b96d4041b60827	diff --git a/README.md b/README.md index 66df9b7..738b674 100644 --- a/README.md +++ b/README.md @@ -14,6 +14,7 @@ Thalassa is currently supporting visualization of the output of the following so - [Schism](https://github.com/schism-dev/schism) - [ADCIRC](https://adcirc.org/) +- [openTELEMAC](https://www.opentelemac.org/) Adding support for new solvers is relatively straight-forward. @@ -30,6 +31,11 @@ Adding support for new solvers is relatively straight-forward. ``` pip install thalassa ``` +for Selafin files (openTELEMAC outputs): +``` +pip install thalassa +pip install xarray-selafin +``` ### Conda diff --git a/docs/index.md b/docs/index.md index b8108e2..366a5b1 100644 --- a/docs/index.md +++ b/docs/index.md @@ -16,6 +16,7 @@ Thalassa is currently supporting visualization of the output of the following so - [Schism](https://github.com/schism-dev/schism) - [ADCIRC](https://adcirc.org/) +- [TELEMAC](https://www.opentelemac.org) Adding support for new solvers is relatively straight-forward. diff --git a/docs/installation.md b/docs/installation.md index c01ac33..d8807f4 100644 --- a/docs/installation.md +++ b/docs/installation.md @@ -12,6 +12,11 @@ ``` pip install thalassa ``` +for Selafin files (openTELEMAC outputs): +``` +pip install thalassa +pip install xarray-selafin +``` ### Conda diff --git a/poetry.lock b/poetry.lock index df8ae78..5f50244 100644 --- a/poetry.lock +++ b/poetry.lock @@ -1,88 +1,88 @@ -# This file is automatically @generated by Poetry 1.7.1 and should not be changed by hand. +# This file is automatically @generated by Poetry 1.8.2 and should not be changed by hand. [[package]] name = "aiohttp" -version = "3.9.0b0" +version = "3.9.5" description = "Async http client/server framework (asyncio)" optional = false python-versions = ">=3.8" files = [ - {file = "aiohttp-3.9.0b0-cp310-cp310-macosx_10_9_universal2.whl", hash = "sha256:50b550b5e317e40a017bab8b25995676af3aa66dd0ef562cd7dce7f1684cd376"}, - {file = "aiohttp-3.9.0b0-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:8f902ad26b9814852e0a17d48f98ba4c879d8136c4fa9b235b5c043dde0a0257"}, - {file = "aiohttp-3.9.0b0-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:c2140de122ecf3eb7947105ceb91fb6632fb21cc1d17f6ff19c3973d2d12730d"}, - {file = "aiohttp-3.9.0b0-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:e014b343225d8d358ee91962b588e863fded12a6e2f9b446bb3be85c678e04ae"}, - {file = "aiohttp-3.9.0b0-cp310-cp310-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:c7016695087e616a2806ccdb1f83609e5fecb3958c270e3e5a42f69d225536f2"}, - {file = "aiohttp-3.9.0b0-cp310-cp310-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:40407d5ec81682225ad5538d9bd68b0f8242caa91e72a6a9a95197fd7d9aebb2"}, - {file = "aiohttp-3.9.0b0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:bd54502e6b4144785f2f14a5f1544ced0a77dbecb1fd422f21dfad95dcb7fcb8"}, - {file = "aiohttp-3.9.0b0-cp310-cp310-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:67f911fd2073621eecfe77b17926460e72980b9b996d0ab7dad5e38805ce2988"}, - {file = "aiohttp-3.9.0b0-cp310-cp310-musllinux_1_1_aarch64.whl", hash = "sha256:638ba28af2c821b70574664a991dfdfaf1a7a7ae1a8068757f7d59cdf2d8361a"}, - {file = "aiohttp-3.9.0b0-cp310-cp310-musllinux_1_1_i686.whl", hash = "sha256:76329f7c1f5f3185d91d61d64615d88fa3dfddf389a83f6cd46a205c5b61e01b"}, - {file = "aiohttp-3.9.0b0-cp310-cp310-musllinux_1_1_ppc64le.whl", hash = "sha256:597b5d44b613dea9c62779592eb0ecae87604628564ecaff8d516457def68184"}, - {file = "aiohttp-3.9.0b0-cp310-cp310-musllinux_1_1_s390x.whl", hash = "sha256:cd5edd7ba2b3f95346e0fc8ba2364bdd93917a1bf8528e7d60ec80cf21dfba7e"}, - {file = "aiohttp-3.9.0b0-cp310-cp310-musllinux_1_1_x86_64.whl", hash = "sha256:72556e0cce47c6e558454316fc5c6a3fb0980344eee8af7aa52b495d82ef12a5"}, - {file = "aiohttp-3.9.0b0-cp310-cp310-win32.whl", hash = "sha256:01a3b241288c4d8171fe5e2434a799d0b82700d2ed2156b43f1d7f4f521ba382"}, - {file = "aiohttp-3.9.0b0-cp310-cp310-win_amd64.whl", hash = "sha256:17962c404788c348ce5b58efaf4969249183c551890e30bfd9c035188d21e3d1"}, - {file = "aiohttp-3.9.0b0-cp311-cp311-macosx_10_9_universal2.whl", hash = "sha256:94197a77859ab1039b9ca6c3c393b8e7b5fc34a9abfbcb58daac38ab89684a99"}, - {file = "aiohttp-3.9.0b0-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:0c78d2cfe1515cfb31ba67edf0518c6677a963ec2039b652b03a886733e72e65"}, - {file = "aiohttp-3.9.0b0-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:28b38a14f564c833e59c99f748b48803e4babeabc6a0307952b01e6c8d642cab"}, - {file = "aiohttp-3.9.0b0-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:e596cfc52380f71e197e7cf0e2d3c4714b4bf66d2d562cdbd5442284bac18909"}, - {file = "aiohttp-3.9.0b0-cp311-cp311-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:6190951b7933c834d9346e21c5a81642caa210d291cda4036daf85fc53162d35"}, - {file = "aiohttp-3.9.0b0-cp311-cp311-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:fb0cb2cbf95cf4cc40307d0d0187f59c4b86b1d7d1a624922a7d0b046deffba7"}, - {file = "aiohttp-3.9.0b0-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:e27c283e21e94fa1582d31b57c514b87ab609882ade413ce43f585d73c8a33fc"}, - {file = "aiohttp-3.9.0b0-cp311-cp311-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:c6826c59b4e99673728bcdaecacbd699b7521f17ca165c63a5e26e23d42aeea5"}, - {file = "aiohttp-3.9.0b0-cp311-cp311-musllinux_1_1_aarch64.whl", hash = "sha256:aa4738f3b1b916b1cc69ed3d1dead9714919dc4d30ae0d5f6d55eadb2c511133"}, - {file = "aiohttp-3.9.0b0-cp311-cp311-musllinux_1_1_i686.whl", hash = "sha256:4b2abd7936f687de3a3ab199b145a9de01ed046eb5640cd66f47da07a9050a78"}, - {file = "aiohttp-3.9.0b0-cp311-cp311-musllinux_1_1_ppc64le.whl", hash = "sha256:652cc00a97bc206c470db06276ce57ff2a53a625795bbce8435ef8b6a4cb0113"}, - {file = "aiohttp-3.9.0b0-cp311-cp311-musllinux_1_1_s390x.whl", hash = "sha256:d54529c1d95d5d200ecb7133a343785e5661a804f3dcee090a7bca3b48189d69"}, - {file = "aiohttp-3.9.0b0-cp311-cp311-musllinux_1_1_x86_64.whl", hash = "sha256:324fe990c97721ea8eb4d439f12b59d1a93cd7e0dd188c7b145bffdfbd327dc3"}, - {file = "aiohttp-3.9.0b0-cp311-cp311-win32.whl", hash = "sha256:3a2ef8318435f40f5906af36fda20b5432e07e6a7e05de3a4d2934c25320b8ff"}, - {file = "aiohttp-3.9.0b0-cp311-cp311-win_amd64.whl", hash = "sha256:887d8757aafc7f6fbda76faaff21fc2aa31b9dca0911ecd6b60b0fe922a2abfc"}, - {file = "aiohttp-3.9.0b0-cp312-cp312-macosx_10_9_universal2.whl", hash = "sha256:9c430c706589a811b38e33e1492d194cbb0f6f2e027877bf038debced703446f"}, - {file = "aiohttp-3.9.0b0-cp312-cp312-macosx_10_9_x86_64.whl", hash = "sha256:9b820981f1c5d6da382e4859318ba78c9b5c583f0920e44a18efb3387b18487e"}, - {file = "aiohttp-3.9.0b0-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:c64677a2df742bcd89b94c35689306663d8246a8534bea5835afc706416f8dd6"}, - {file = "aiohttp-3.9.0b0-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:903155c179cda589d01936953158685747af43d98cdd3673a671c6e7f5c94178"}, - {file = "aiohttp-3.9.0b0-cp312-cp312-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:77cbb6e4a146449f805fa0e725b0b2a06411d21417d8eca699bbee55204201d0"}, - {file = "aiohttp-3.9.0b0-cp312-cp312-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:bc3cc9f5e6e493a2b9c3d241fca870b5a64aa4c247f1192f9e34fae990667df8"}, - {file = "aiohttp-3.9.0b0-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:92071206e570b7da6380f8d376820e2a40230638b8fd8b45b28103b346704c5e"}, - {file = "aiohttp-3.9.0b0-cp312-cp312-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:242e3cb0b2d441a2d20443114eebe3032078d1894ac1d97ab2dd101165ea50e1"}, - {file = "aiohttp-3.9.0b0-cp312-cp312-musllinux_1_1_aarch64.whl", hash = "sha256:044c5a8923bd44a4a0769a2886130c19f7f3a4a1a284f0ff68c2a751920ee39f"}, - {file = "aiohttp-3.9.0b0-cp312-cp312-musllinux_1_1_i686.whl", hash = "sha256:99b1b0d0f63ff48f80aa89be3ff61bc2b980c5b02895c81dbc1e44ce7b6cb5b7"}, - {file = "aiohttp-3.9.0b0-cp312-cp312-musllinux_1_1_ppc64le.whl", hash = "sha256:f737a47b5df97b6da457a0b2739d6d819ffadea2f36336988b53dbdb1796ba89"}, - {file = "aiohttp-3.9.0b0-cp312-cp312-musllinux_1_1_s390x.whl", hash = "sha256:e6d79f8b8347afbecd8047a1f6e74c810eb82497256cc906ee384635174dcaea"}, - {file = "aiohttp-3.9.0b0-cp312-cp312-musllinux_1_1_x86_64.whl", hash = "sha256:2f1b0a821564e315ec5cfa0abaf048355e229995a812380ec7a2200d87a6ed11"}, - {file = "aiohttp-3.9.0b0-cp312-cp312-win32.whl", hash = "sha256:ab2702f281ca504529e82be78dae2b9ca31d51a92ab8b239bd326b74c79d7af4"}, - {file = "aiohttp-3.9.0b0-cp312-cp312-win_amd64.whl", hash = "sha256:b81722b88abd4aab656abfec122646b6171da64340ff92af3bcf1af5f0d1275e"}, - {file = "aiohttp-3.9.0b0-cp38-cp38-macosx_10_9_universal2.whl", hash = "sha256:49e2ca017f506d1a9c60f44301ceff2eb8bbfe24b9cd9b4c4a363d9e5f68e92b"}, - {file = "aiohttp-3.9.0b0-cp38-cp38-macosx_10_9_x86_64.whl", hash = "sha256:06cba5518d8e30b46fcec2a8ed22ec6027fc9864583e0b538da642507f66fe29"}, - {file = "aiohttp-3.9.0b0-cp38-cp38-macosx_11_0_arm64.whl", hash = "sha256:e5201d3f8d0b2748eba5093820861639cac1ea1dfdff537f67152a1c082e1243"}, - {file = "aiohttp-3.9.0b0-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:8c483d0a666f6cbec2e974f760f93499bbcfcb17a7c4035d4c4c653e6a3b21b1"}, - {file = "aiohttp-3.9.0b0-cp38-cp38-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:04f48476ce3e96843b44084fd15139b195781c10ed6eb5ffb706fb9d2ca95ce4"}, - {file = "aiohttp-3.9.0b0-cp38-cp38-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:09fdad08544a4479e5801c777697c155fa9d966c91b6dcf3e1a0d271ad3999f7"}, - {file = "aiohttp-3.9.0b0-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:127aa57415005eb04fb1a3685c9d7b42aef6718be72b8a62b4b30ee00f7d23f4"}, - {file = "aiohttp-3.9.0b0-cp38-cp38-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:aa8f29f0647f10f6bcd9f597f1319d13ce1d6efe2d55169226940093eeadf609"}, - {file = "aiohttp-3.9.0b0-cp38-cp38-musllinux_1_1_aarch64.whl", hash = "sha256:8dc394dea47594825ac2a662c4fac6a8b294acd937396aaec8e41ed03728898b"}, - {file = "aiohttp-3.9.0b0-cp38-cp38-musllinux_1_1_i686.whl", hash = "sha256:c332b343974c6fbfec53e3ac7afebd6ba6cc1777cda67c28fabb3562411a9b5a"}, - {file = "aiohttp-3.9.0b0-cp38-cp38-musllinux_1_1_ppc64le.whl", hash = "sha256:6dfad718b328de3fa30d663393d51feea625322ec723bdecdec3f5f52ba6347f"}, - {file = "aiohttp-3.9.0b0-cp38-cp38-musllinux_1_1_s390x.whl", hash = "sha256:6edaeb63a4657672b04afcc25c253e960125e805f5a8f8cfa7bf682d15115f49"}, - {file = "aiohttp-3.9.0b0-cp38-cp38-musllinux_1_1_x86_64.whl", hash = "sha256:20023087bce5f3adde4872042ea1193d31d98b29682c28a6309d72bce0d9725e"}, - {file = "aiohttp-3.9.0b0-cp38-cp38-win32.whl", hash = "sha256:ad07ee4165a82e646310c152a74997c759d5782aef58bab9d77034b4cc87e153"}, - {file = "aiohttp-3.9.0b0-cp38-cp38-win_amd64.whl", hash = "sha256:494062a8447c6665f5237c47ca8bb5659cd3128ad9b4af5543566a11bb88df5c"}, - {file = "aiohttp-3.9.0b0-cp39-cp39-macosx_10_9_universal2.whl", hash = "sha256:aaff57bd1ab9eb1a205f3b7a00e2dc159d1e7e4373870be0d192358a656d9e60"}, - {file = "aiohttp-3.9.0b0-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:3c212f5066ffe9490856b706a9d9bd457f14716f4db4b1b73939245a1acecc4e"}, - {file = "aiohttp-3.9.0b0-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:d80664b3b82fb9ee2c7b13072651cd68d65fbb3a69721040c08969bab4335628"}, - {file = "aiohttp-3.9.0b0-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:e7cf539fc98297e312308405949ca2f04a347eb021e30d004388cdb5d155a0ec"}, - {file = "aiohttp-3.9.0b0-cp39-cp39-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:6be520717b895508c63df90e48135ba616c702a9229d4be71841dce2ea6a569f"}, - {file = "aiohttp-3.9.0b0-cp39-cp39-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:1b25e926cd16b44aeef29fffbb9fc9f577f52a6230e46926e391545b85cd0ce3"}, - {file = "aiohttp-3.9.0b0-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:35f6cafe361c0323945c13122c282ea22fb0df96e845f34c4d8abd96e2a81995"}, - {file = "aiohttp-3.9.0b0-cp39-cp39-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:5c9851e3d0396686d96a7e3559bf5912ed79c944ff1a6ae3cf7b1da320c3ad2b"}, - {file = "aiohttp-3.9.0b0-cp39-cp39-musllinux_1_1_aarch64.whl", hash = "sha256:0ab413eddeb1a03ba84d06acf7024a646b049d991ed0616bcc1ee40dc8fffa9e"}, - {file = "aiohttp-3.9.0b0-cp39-cp39-musllinux_1_1_i686.whl", hash = "sha256:89b271a8658472a9d400836ee8caee743246bae5c06405a63b6ba366f58df727"}, - {file = "aiohttp-3.9.0b0-cp39-cp39-musllinux_1_1_ppc64le.whl", hash = "sha256:dd941d473b86d0d5a413a1832499e5b80f648d66ca0c8246c26a4ccd66bcf7ec"}, - {file = "aiohttp-3.9.0b0-cp39-cp39-musllinux_1_1_s390x.whl", hash = "sha256:ce4f000279fb85527c017ef429615f2cb5a0cb614c088610849ddc6c2ac8d91b"}, - {file = "aiohttp-3.9.0b0-cp39-cp39-musllinux_1_1_x86_64.whl", hash = "sha256:f50a4f6773a9eedefb24b42c611e31dcd13f6139419a8656f7e525cb8a00687e"}, - {file = "aiohttp-3.9.0b0-cp39-cp39-win32.whl", hash = "sha256:b14dcfcc5ad161d007da71e1c1211909d527d9d7c2795ea9e17191ba25e5d89a"}, - {file = "aiohttp-3.9.0b0-cp39-cp39-win_amd64.whl", hash = "sha256:567245a91a57c41899f5d266814c9da8782d3d949dc1e66469429f08713a3ec6"}, - {file = "aiohttp-3.9.0b0.tar.gz", hash = "sha256:cecc64fd7bae6debdf43437e3c83183c40d4f4d86486946f412c113960598eee"}, + {file = "aiohttp-3.9.5-cp310-cp310-macosx_10_9_universal2.whl", hash = "sha256:fcde4c397f673fdec23e6b05ebf8d4751314fa7c24f93334bf1f1364c1c69ac7"}, + {file = "aiohttp-3.9.5-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:5d6b3f1fabe465e819aed2c421a6743d8debbde79b6a8600739300630a01bf2c"}, + {file = "aiohttp-3.9.5-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:6ae79c1bc12c34082d92bf9422764f799aee4746fd7a392db46b7fd357d4a17a"}, + {file = "aiohttp-3.9.5-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:4d3ebb9e1316ec74277d19c5f482f98cc65a73ccd5430540d6d11682cd857430"}, + {file = "aiohttp-3.9.5-cp310-cp310-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:84dabd95154f43a2ea80deffec9cb44d2e301e38a0c9d331cc4aa0166fe28ae3"}, + {file = "aiohttp-3.9.5-cp310-cp310-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:c8a02fbeca6f63cb1f0475c799679057fc9268b77075ab7cf3f1c600e81dd46b"}, + {file = "aiohttp-3.9.5-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:c26959ca7b75ff768e2776d8055bf9582a6267e24556bb7f7bd29e677932be72"}, + {file = "aiohttp-3.9.5-cp310-cp310-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:714d4e5231fed4ba2762ed489b4aec07b2b9953cf4ee31e9871caac895a839c0"}, + {file = "aiohttp-3.9.5-cp310-cp310-musllinux_1_1_aarch64.whl", hash = "sha256:e7a6a8354f1b62e15d48e04350f13e726fa08b62c3d7b8401c0a1314f02e3558"}, + {file = "aiohttp-3.9.5-cp310-cp310-musllinux_1_1_i686.whl", hash = "sha256:c413016880e03e69d166efb5a1a95d40f83d5a3a648d16486592c49ffb76d0db"}, + {file = "aiohttp-3.9.5-cp310-cp310-musllinux_1_1_ppc64le.whl", hash = "sha256:ff84aeb864e0fac81f676be9f4685f0527b660f1efdc40dcede3c251ef1e867f"}, + {file = "aiohttp-3.9.5-cp310-cp310-musllinux_1_1_s390x.whl", hash = "sha256:ad7f2919d7dac062f24d6f5fe95d401597fbb015a25771f85e692d043c9d7832"}, + {file = "aiohttp-3.9.5-cp310-cp310-musllinux_1_1_x86_64.whl", hash = "sha256:702e2c7c187c1a498a4e2b03155d52658fdd6fda882d3d7fbb891a5cf108bb10"}, + {file = "aiohttp-3.9.5-cp310-cp310-win32.whl", hash = "sha256:67c3119f5ddc7261d47163ed86d760ddf0e625cd6246b4ed852e82159617b5fb"}, + {file = "aiohttp-3.9.5-cp310-cp310-win_amd64.whl", hash = "sha256:471f0ef53ccedec9995287f02caf0c068732f026455f07db3f01a46e49d76bbb"}, + {file = "aiohttp-3.9.5-cp311-cp311-macosx_10_9_universal2.whl", hash = "sha256:e0ae53e33ee7476dd3d1132f932eeb39bf6125083820049d06edcdca4381f342"}, + {file = "aiohttp-3.9.5-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:c088c4d70d21f8ca5c0b8b5403fe84a7bc8e024161febdd4ef04575ef35d474d"}, + {file = "aiohttp-3.9.5-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:639d0042b7670222f33b0028de6b4e2fad6451462ce7df2af8aee37dcac55424"}, + {file = "aiohttp-3.9.5-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:f26383adb94da5e7fb388d441bf09c61e5e35f455a3217bfd790c6b6bc64b2ee"}, + {file = "aiohttp-3.9.5-cp311-cp311-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:66331d00fb28dc90aa606d9a54304af76b335ae204d1836f65797d6fe27f1ca2"}, + {file = "aiohttp-3.9.5-cp311-cp311-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:4ff550491f5492ab5ed3533e76b8567f4b37bd2995e780a1f46bca2024223233"}, + {file = "aiohttp-3.9.5-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:f22eb3a6c1080d862befa0a89c380b4dafce29dc6cd56083f630073d102eb595"}, + {file = "aiohttp-3.9.5-cp311-cp311-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:a81b1143d42b66ffc40a441379387076243ef7b51019204fd3ec36b9f69e77d6"}, + {file = "aiohttp-3.9.5-cp311-cp311-musllinux_1_1_aarch64.whl", hash = "sha256:f64fd07515dad67f24b6ea4a66ae2876c01031de91c93075b8093f07c0a2d93d"}, + {file = "aiohttp-3.9.5-cp311-cp311-musllinux_1_1_i686.whl", hash = "sha256:93e22add827447d2e26d67c9ac0161756007f152fdc5210277d00a85f6c92323"}, + {file = "aiohttp-3.9.5-cp311-cp311-musllinux_1_1_ppc64le.whl", hash = "sha256:55b39c8684a46e56ef8c8d24faf02de4a2b2ac60d26cee93bc595651ff545de9"}, + {file = "aiohttp-3.9.5-cp311-cp311-musllinux_1_1_s390x.whl", hash = "sha256:4715a9b778f4293b9f8ae7a0a7cef9829f02ff8d6277a39d7f40565c737d3771"}, + {file = "aiohttp-3.9.5-cp311-cp311-musllinux_1_1_x86_64.whl", hash = "sha256:afc52b8d969eff14e069a710057d15ab9ac17cd4b6753042c407dcea0e40bf75"}, + {file = "aiohttp-3.9.5-cp311-cp311-win32.whl", hash = "sha256:b3df71da99c98534be076196791adca8819761f0bf6e08e07fd7da25127150d6"}, + {file = "aiohttp-3.9.5-cp311-cp311-win_amd64.whl", hash = "sha256:88e311d98cc0bf45b62fc46c66753a83445f5ab20038bcc1b8a1cc05666f428a"}, + {file = "aiohttp-3.9.5-cp312-cp312-macosx_10_9_universal2.whl", hash = "sha256:c7a4b7a6cf5b6eb11e109a9755fd4fda7d57395f8c575e166d363b9fc3ec4678"}, + {file = "aiohttp-3.9.5-cp312-cp312-macosx_10_9_x86_64.whl", hash = "sha256:0a158704edf0abcac8ac371fbb54044f3270bdbc93e254a82b6c82be1ef08f3c"}, + {file = "aiohttp-3.9.5-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:d153f652a687a8e95ad367a86a61e8d53d528b0530ef382ec5aaf533140ed00f"}, + {file = "aiohttp-3.9.5-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:82a6a97d9771cb48ae16979c3a3a9a18b600a8505b1115cfe354dfb2054468b4"}, + {file = "aiohttp-3.9.5-cp312-cp312-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:60cdbd56f4cad9f69c35eaac0fbbdf1f77b0ff9456cebd4902f3dd1cf096464c"}, + {file = "aiohttp-3.9.5-cp312-cp312-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:8676e8fd73141ded15ea586de0b7cda1542960a7b9ad89b2b06428e97125d4fa"}, + {file = "aiohttp-3.9.5-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:da00da442a0e31f1c69d26d224e1efd3a1ca5bcbf210978a2ca7426dfcae9f58"}, + {file = "aiohttp-3.9.5-cp312-cp312-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:18f634d540dd099c262e9f887c8bbacc959847cfe5da7a0e2e1cf3f14dbf2daf"}, + {file = "aiohttp-3.9.5-cp312-cp312-musllinux_1_1_aarch64.whl", hash = "sha256:320e8618eda64e19d11bdb3bd04ccc0a816c17eaecb7e4945d01deee2a22f95f"}, + {file = "aiohttp-3.9.5-cp312-cp312-musllinux_1_1_i686.whl", hash = "sha256:2faa61a904b83142747fc6a6d7ad8fccff898c849123030f8e75d5d967fd4a81"}, + {file = "aiohttp-3.9.5-cp312-cp312-musllinux_1_1_ppc64le.whl", hash = "sha256:8c64a6dc3fe5db7b1b4d2b5cb84c4f677768bdc340611eca673afb7cf416ef5a"}, + {file = "aiohttp-3.9.5-cp312-cp312-musllinux_1_1_s390x.whl", hash = "sha256:393c7aba2b55559ef7ab791c94b44f7482a07bf7640d17b341b79081f5e5cd1a"}, + {file = "aiohttp-3.9.5-cp312-cp312-musllinux_1_1_x86_64.whl", hash = "sha256:c671dc117c2c21a1ca10c116cfcd6e3e44da7fcde37bf83b2be485ab377b25da"}, + {file = "aiohttp-3.9.5-cp312-cp312-win32.whl", hash = "sha256:5a7ee16aab26e76add4afc45e8f8206c95d1d75540f1039b84a03c3b3800dd59"}, + {file = "aiohttp-3.9.5-cp312-cp312-win_amd64.whl", hash = "sha256:5ca51eadbd67045396bc92a4345d1790b7301c14d1848feaac1d6a6c9289e888"}, + {file = "aiohttp-3.9.5-cp38-cp38-macosx_10_9_universal2.whl", hash = "sha256:694d828b5c41255e54bc2dddb51a9f5150b4eefa9886e38b52605a05d96566e8"}, + {file = "aiohttp-3.9.5-cp38-cp38-macosx_10_9_x86_64.whl", hash = "sha256:0605cc2c0088fcaae79f01c913a38611ad09ba68ff482402d3410bf59039bfb8"}, + {file = "aiohttp-3.9.5-cp38-cp38-macosx_11_0_arm64.whl", hash = "sha256:4558e5012ee03d2638c681e156461d37b7a113fe13970d438d95d10173d25f78"}, + {file = "aiohttp-3.9.5-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:9dbc053ac75ccc63dc3a3cc547b98c7258ec35a215a92bd9f983e0aac95d3d5b"}, + {file = "aiohttp-3.9.5-cp38-cp38-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:4109adee842b90671f1b689901b948f347325045c15f46b39797ae1bf17019de"}, + {file = "aiohttp-3.9.5-cp38-cp38-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:a6ea1a5b409a85477fd8e5ee6ad8f0e40bf2844c270955e09360418cfd09abac"}, + {file = "aiohttp-3.9.5-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:f3c2890ca8c59ee683fd09adf32321a40fe1cf164e3387799efb2acebf090c11"}, + {file = "aiohttp-3.9.5-cp38-cp38-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:3916c8692dbd9d55c523374a3b8213e628424d19116ac4308e434dbf6d95bbdd"}, + {file = "aiohttp-3.9.5-cp38-cp38-musllinux_1_1_aarch64.whl", hash = "sha256:8d1964eb7617907c792ca00b341b5ec3e01ae8c280825deadbbd678447b127e1"}, + {file = "aiohttp-3.9.5-cp38-cp38-musllinux_1_1_i686.whl", hash = "sha256:d5ab8e1f6bee051a4bf6195e38a5c13e5e161cb7bad83d8854524798bd9fcd6e"}, + {file = "aiohttp-3.9.5-cp38-cp38-musllinux_1_1_ppc64le.whl", hash = "sha256:52c27110f3862a1afbcb2af4281fc9fdc40327fa286c4625dfee247c3ba90156"}, + {file = "aiohttp-3.9.5-cp38-cp38-musllinux_1_1_s390x.whl", hash = "sha256:7f64cbd44443e80094309875d4f9c71d0401e966d191c3d469cde4642bc2e031"}, + {file = "aiohttp-3.9.5-cp38-cp38-musllinux_1_1_x86_64.whl", hash = "sha256:8b4f72fbb66279624bfe83fd5eb6aea0022dad8eec62b71e7bf63ee1caadeafe"}, + {file = "aiohttp-3.9.5-cp38-cp38-win32.whl", hash = "sha256:6380c039ec52866c06d69b5c7aad5478b24ed11696f0e72f6b807cfb261453da"}, + {file = "aiohttp-3.9.5-cp38-cp38-win_amd64.whl", hash = "sha256:da22dab31d7180f8c3ac7c7635f3bcd53808f374f6aa333fe0b0b9e14b01f91a"}, + {file = "aiohttp-3.9.5-cp39-cp39-macosx_10_9_universal2.whl", hash = "sha256:1732102949ff6087589408d76cd6dea656b93c896b011ecafff418c9661dc4ed"}, + {file = "aiohttp-3.9.5-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:c6021d296318cb6f9414b48e6a439a7f5d1f665464da507e8ff640848ee2a58a"}, + {file = "aiohttp-3.9.5-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:239f975589a944eeb1bad26b8b140a59a3a320067fb3cd10b75c3092405a1372"}, + {file = "aiohttp-3.9.5-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:3b7b30258348082826d274504fbc7c849959f1989d86c29bc355107accec6cfb"}, + {file = "aiohttp-3.9.5-cp39-cp39-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:cd2adf5c87ff6d8b277814a28a535b59e20bfea40a101db6b3bdca7e9926bc24"}, + {file = "aiohttp-3.9.5-cp39-cp39-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:e9a3d838441bebcf5cf442700e3963f58b5c33f015341f9ea86dcd7d503c07e2"}, + {file = "aiohttp-3.9.5-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:9e3a1ae66e3d0c17cf65c08968a5ee3180c5a95920ec2731f53343fac9bad106"}, + {file = "aiohttp-3.9.5-cp39-cp39-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:9c69e77370cce2d6df5d12b4e12bdcca60c47ba13d1cbbc8645dd005a20b738b"}, + {file = "aiohttp-3.9.5-cp39-cp39-musllinux_1_1_aarch64.whl", hash = "sha256:0cbf56238f4bbf49dab8c2dc2e6b1b68502b1e88d335bea59b3f5b9f4c001475"}, + {file = "aiohttp-3.9.5-cp39-cp39-musllinux_1_1_i686.whl", hash = "sha256:d1469f228cd9ffddd396d9948b8c9cd8022b6d1bf1e40c6f25b0fb90b4f893ed"}, + {file = "aiohttp-3.9.5-cp39-cp39-musllinux_1_1_ppc64le.whl", hash = "sha256:45731330e754f5811c314901cebdf19dd776a44b31927fa4b4dbecab9e457b0c"}, + {file = "aiohttp-3.9.5-cp39-cp39-musllinux_1_1_s390x.whl", hash = "sha256:3fcb4046d2904378e3aeea1df51f697b0467f2aac55d232c87ba162709478c46"}, + {file = "aiohttp-3.9.5-cp39-cp39-musllinux_1_1_x86_64.whl", hash = "sha256:8cf142aa6c1a751fcb364158fd710b8a9be874b81889c2bd13aa8893197455e2"}, + {file = "aiohttp-3.9.5-cp39-cp39-win32.whl", hash = "sha256:7b179eea70833c8dee51ec42f3b4097bd6370892fa93f510f76762105568cf09"}, + {file = "aiohttp-3.9.5-cp39-cp39-win_amd64.whl", hash = "sha256:38d80498e2e169bc61418ff36170e0aad0cd268da8b38a17c4cf29d254a8b3f1"}, + {file = "aiohttp-3.9.5.tar.gz", hash = "sha256:edea7d15772ceeb29db4aff55e482d4bcfb6ae160ce144f2682de02f6d693551"}, ] [package.dependencies] @@ -109,6 +109,17 @@ files = [ [package.dependencies] frozenlist = ">=1.1.0" +[[package]] +name = "ansicolors" +version = "1.1.8" +description = "ANSI colors for Python" +optional = false +python-versions = "" +files = [ + {file = "ansicolors-1.1.8-py2.py3-none-any.whl", hash = "sha256:00d2dde5a675579325902536738dd27e4fac1fd68f773fe36c21044eb559e187"}, + {file = "ansicolors-1.1.8.zip", hash = "sha256:99f94f5e3348a0bcd43c82e5fc4414013ccc19d70bd939ad71e0133ce9c372e0"}, +] + [[package]] name = "appnope" version = "0.1.4" @@ -169,13 +180,13 @@ tests-no-zope = ["attrs[tests-mypy]", "cloudpickle", "hypothesis", "pympler", "p [[package]] name = "babel" -version = "2.14.0" +version = "2.15.0" description = "Internationalization utilities" optional = false -python-versions = ">=3.7" +python-versions = ">=3.8" files = [ - {file = "Babel-2.14.0-py3-none-any.whl", hash = "sha256:efb1a25b7118e67ce3a259bed20545c29cb68be8ad2c784c83689981b7a57287"}, - {file = "Babel-2.14.0.tar.gz", hash = "sha256:6919867db036398ba21eb5c7a0f6b28ab8cbc3ae7a73a44ebe34ae74a4e7d363"}, + {file = "Babel-2.15.0-py3-none-any.whl", hash = "sha256:08706bdad8d0a3413266ab61bd6c34d0c28d6e1e7badf40a2cebe67644e2e1fb"}, + {file = "babel-2.15.0.tar.gz", hash = "sha256:8daf0e265d05768bc6c7a314cf1321e9a123afc328cc635c18622a2f30a04413"}, ] [package.extras] @@ -204,33 +215,33 @@ lxml = ["lxml"] [[package]] name = "black" -version = "24.1.1" +version = "24.4.2" description = "The uncompromising code formatter." optional = false python-versions = ">=3.8" files = [ - {file = "black-24.1.1-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:2588021038bd5ada078de606f2a804cadd0a3cc6a79cb3e9bb3a8bf581325a4c"}, - {file = "black-24.1.1-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:1a95915c98d6e32ca43809d46d932e2abc5f1f7d582ffbe65a5b4d1588af7445"}, - {file = "black-24.1.1-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:2fa6a0e965779c8f2afb286f9ef798df770ba2b6cee063c650b96adec22c056a"}, - {file = "black-24.1.1-cp310-cp310-win_amd64.whl", hash = "sha256:5242ecd9e990aeb995b6d03dc3b2d112d4a78f2083e5a8e86d566340ae80fec4"}, - {file = "black-24.1.1-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:fc1ec9aa6f4d98d022101e015261c056ddebe3da6a8ccfc2c792cbe0349d48b7"}, - {file = "black-24.1.1-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:0269dfdea12442022e88043d2910429bed717b2d04523867a85dacce535916b8"}, - {file = "black-24.1.1-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:b3d64db762eae4a5ce04b6e3dd745dcca0fb9560eb931a5be97472e38652a161"}, - {file = "black-24.1.1-cp311-cp311-win_amd64.whl", hash = "sha256:5d7b06ea8816cbd4becfe5f70accae953c53c0e53aa98730ceccb0395520ee5d"}, - {file = "black-24.1.1-cp312-cp312-macosx_10_9_x86_64.whl", hash = "sha256:e2c8dfa14677f90d976f68e0c923947ae68fa3961d61ee30976c388adc0b02c8"}, - {file = "black-24.1.1-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:a21725862d0e855ae05da1dd25e3825ed712eaaccef6b03017fe0853a01aa45e"}, - {file = "black-24.1.1-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:07204d078e25327aad9ed2c64790d681238686bce254c910de640c7cc4fc3aa6"}, - {file = "black-24.1.1-cp312-cp312-win_amd64.whl", hash = "sha256:a83fe522d9698d8f9a101b860b1ee154c1d25f8a82ceb807d319f085b2627c5b"}, - {file = "black-24.1.1-cp38-cp38-macosx_10_9_x86_64.whl", hash = "sha256:08b34e85170d368c37ca7bf81cf67ac863c9d1963b2c1780c39102187ec8dd62"}, - {file = "black-24.1.1-cp38-cp38-macosx_11_0_arm64.whl", hash = "sha256:7258c27115c1e3b5de9ac6c4f9957e3ee2c02c0b39222a24dc7aa03ba0e986f5"}, - {file = "black-24.1.1-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:40657e1b78212d582a0edecafef133cf1dd02e6677f539b669db4746150d38f6"}, - {file = "black-24.1.1-cp38-cp38-win_amd64.whl", hash = "sha256:e298d588744efda02379521a19639ebcd314fba7a49be22136204d7ed1782717"}, - {file = "black-24.1.1-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:34afe9da5056aa123b8bfda1664bfe6fb4e9c6f311d8e4a6eb089da9a9173bf9"}, - {file = "black-24.1.1-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:854c06fb86fd854140f37fb24dbf10621f5dab9e3b0c29a690ba595e3d543024"}, - {file = "black-24.1.1-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:3897ae5a21ca132efa219c029cce5e6bfc9c3d34ed7e892113d199c0b1b444a2"}, - {file = "black-24.1.1-cp39-cp39-win_amd64.whl", hash = "sha256:ecba2a15dfb2d97105be74bbfe5128bc5e9fa8477d8c46766505c1dda5883aac"}, - {file = "black-24.1.1-py3-none-any.whl", hash = "sha256:5cdc2e2195212208fbcae579b931407c1fa9997584f0a415421748aeafff1168"}, - {file = "black-24.1.1.tar.gz", hash = "sha256:48b5760dcbfe5cf97fd4fba23946681f3a81514c6ab8a45b50da67ac8fbc6c7b"}, + {file = "black-24.4.2-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:dd1b5a14e417189db4c7b64a6540f31730713d173f0b63e55fabd52d61d8fdce"}, + {file = "black-24.4.2-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:8e537d281831ad0e71007dcdcbe50a71470b978c453fa41ce77186bbe0ed6021"}, + {file = "black-24.4.2-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:eaea3008c281f1038edb473c1aa8ed8143a5535ff18f978a318f10302b254063"}, + {file = "black-24.4.2-cp310-cp310-win_amd64.whl", hash = "sha256:7768a0dbf16a39aa5e9a3ded568bb545c8c2727396d063bbaf847df05b08cd96"}, + {file = "black-24.4.2-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:257d724c2c9b1660f353b36c802ccece186a30accc7742c176d29c146df6e474"}, + {file = "black-24.4.2-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:bdde6f877a18f24844e381d45e9947a49e97933573ac9d4345399be37621e26c"}, + {file = "black-24.4.2-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:e151054aa00bad1f4e1f04919542885f89f5f7d086b8a59e5000e6c616896ffb"}, + {file = "black-24.4.2-cp311-cp311-win_amd64.whl", hash = "sha256:7e122b1c4fb252fd85df3ca93578732b4749d9be076593076ef4d07a0233c3e1"}, + {file = "black-24.4.2-cp312-cp312-macosx_10_9_x86_64.whl", hash = "sha256:accf49e151c8ed2c0cdc528691838afd217c50412534e876a19270fea1e28e2d"}, + {file = "black-24.4.2-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:88c57dc656038f1ab9f92b3eb5335ee9b021412feaa46330d5eba4e51fe49b04"}, + {file = "black-24.4.2-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:be8bef99eb46d5021bf053114442914baeb3649a89dc5f3a555c88737e5e98fc"}, + {file = "black-24.4.2-cp312-cp312-win_amd64.whl", hash = "sha256:415e686e87dbbe6f4cd5ef0fbf764af7b89f9057b97c908742b6008cc554b9c0"}, + {file = "black-24.4.2-cp38-cp38-macosx_10_9_x86_64.whl", hash = "sha256:bf10f7310db693bb62692609b397e8d67257c55f949abde4c67f9cc574492cc7"}, + {file = "black-24.4.2-cp38-cp38-macosx_11_0_arm64.whl", hash = "sha256:98e123f1d5cfd42f886624d84464f7756f60ff6eab89ae845210631714f6db94"}, + {file = "black-24.4.2-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:48a85f2cb5e6799a9ef05347b476cce6c182d6c71ee36925a6c194d074336ef8"}, + {file = "black-24.4.2-cp38-cp38-win_amd64.whl", hash = "sha256:b1530ae42e9d6d5b670a34db49a94115a64596bc77710b1d05e9801e62ca0a7c"}, + {file = "black-24.4.2-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:37aae07b029fa0174d39daf02748b379399b909652a806e5708199bd93899da1"}, + {file = "black-24.4.2-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:da33a1a5e49c4122ccdfd56cd021ff1ebc4a1ec4e2d01594fef9b6f267a9e741"}, + {file = "black-24.4.2-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:ef703f83fc32e131e9bcc0a5094cfe85599e7109f896fe8bc96cc402f3eb4b6e"}, + {file = "black-24.4.2-cp39-cp39-win_amd64.whl", hash = "sha256:b9176b9832e84308818a99a561e90aa479e73c523b3f77afd07913380ae2eab7"}, + {file = "black-24.4.2-py3-none-any.whl", hash = "sha256:d36ed1124bb81b32f8614555b34cc4259c3fbc7eec17870e8ff8ded335b58d8c"}, + {file = "black-24.4.2.tar.gz", hash = "sha256:c872b53057f000085da66a19c55d68f6f8ddcac2642392ad3a355878406fbd4d"}, ] [package.dependencies] @@ -268,76 +279,76 @@ css = ["tinycss2 (>=1.1.0,<1.3)"] [[package]] name = "bokeh" -version = "3.3.4" +version = "3.4.1" description = "Interactive plots and applications in the browser from Python" optional = false python-versions = ">=3.9" files = [ - {file = "bokeh-3.3.4-py3-none-any.whl", hash = "sha256:ad7b6f89d0a7c2be01eff1db0ca24e2755ac41de14539db919a62e791809c309"}, - {file = "bokeh-3.3.4.tar.gz", hash = "sha256:73b7982dc2b8df15bf660cdddc8d3825e829195c438015a5d09824f1a7028368"}, + {file = "bokeh-3.4.1-py3-none-any.whl", hash = "sha256:1e3c502a0a8205338fc74dadbfa321f8a0965441b39501e36796a47b4017b642"}, + {file = "bokeh-3.4.1.tar.gz", hash = "sha256:d824961e4265367b0750ce58b07e564ad0b83ca64b335521cd3421e9b9f10d89"}, ] [package.dependencies] -contourpy = ">=1" +contourpy = ">=1.2" Jinja2 = ">=2.9" numpy = ">=1.16" packaging = ">=16.8" pandas = ">=1.2" pillow = ">=7.1.0" PyYAML = ">=3.10" -tornado = ">=5.1" +tornado = ">=6.2" xyzservices = ">=2021.09.1" [[package]] name = "bottleneck" -version = "1.3.7" +version = "1.3.8" description = "Fast NumPy array functions written in C" optional = false python-versions = "" files = [ - {file = "Bottleneck-1.3.7-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:ada1a9ff93fd6b1b19f12398a6761940372b00e53d86db98bd4613a751c60043"}, - {file = "Bottleneck-1.3.7-cp310-cp310-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:a477101ee12155a0d2f9a82cd3e2a44b9b1aa53afe5b20acc065c91cf35c3106"}, - {file = "Bottleneck-1.3.7-cp310-cp310-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:cfe5f3010e9ad8ae54871d1e8fd61109c5981ed8d9d14e8496a1c37fe2050a04"}, - {file = "Bottleneck-1.3.7-cp310-cp310-musllinux_1_1_i686.whl", hash = "sha256:955df78713ff58cf815c0a1fa4782b2dc51a8787c0971688472c64b267303855"}, - {file = "Bottleneck-1.3.7-cp310-cp310-musllinux_1_1_x86_64.whl", hash = "sha256:92fc5d3ecfedde3b28c56ca686cff70a8125d1ddc281eb468b5e9d6a61269802"}, - {file = "Bottleneck-1.3.7-cp310-cp310-win32.whl", hash = "sha256:e2a290dcb148c0ddf182052e333892e46730c7d39a1f251af87e3d81a43cdde3"}, - {file = "Bottleneck-1.3.7-cp310-cp310-win_amd64.whl", hash = "sha256:9170ebee4cff423c92b3760afec179bded90eaede7c70dd27cf5f406cc00a1e7"}, - {file = "Bottleneck-1.3.7-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:752c97d64ebebdc10a5568d97b81b4971238fa4b53533248d227c4ea759aee4e"}, - {file = "Bottleneck-1.3.7-cp311-cp311-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:bbda2b27e81a47b7578bce6a8cb2f5eb899279c828d8efb5a154d8ede785093d"}, - {file = "Bottleneck-1.3.7-cp311-cp311-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:79160e49a5a0438468b970967e29addde2d9c6a6ce930144de7ccd8151077603"}, - {file = "Bottleneck-1.3.7-cp311-cp311-musllinux_1_1_i686.whl", hash = "sha256:abfc22762640cc7716c1bd9a409bde0f834167a2584775eb644c6afe7bae3319"}, - {file = "Bottleneck-1.3.7-cp311-cp311-musllinux_1_1_x86_64.whl", hash = "sha256:2a39049aa1cb798538f96150098f50badb37417c41cecfa8a441c0c4e6433c7e"}, - {file = "Bottleneck-1.3.7-cp311-cp311-win32.whl", hash = "sha256:b624023f173332d209ed5bd4134fae43d2432d61c6b17a49a6b7c5591caa7cc4"}, - {file = "Bottleneck-1.3.7-cp311-cp311-win_amd64.whl", hash = "sha256:bd69b1844c90f6bcf1b679a608fb2c0909be5e045f91674d61a0e4c3596644be"}, - {file = "Bottleneck-1.3.7-cp36-cp36m-macosx_10_9_x86_64.whl", hash = "sha256:c8fa2d85298cdd0f88f51fb46604039abe9a5a2b57e00acfe2b64f546754c053"}, - {file = "Bottleneck-1.3.7-cp36-cp36m-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:a09fa97df2f572aed5da487d3fe467829c2212c97b1d63c768ec9be9fd1c57a9"}, - {file = "Bottleneck-1.3.7-cp36-cp36m-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:f239ba0d992c013564979cff4fecf2b251614de3730641e79d05ecc09268c00c"}, - {file = "Bottleneck-1.3.7-cp36-cp36m-musllinux_1_1_i686.whl", hash = "sha256:f5409cf6db5e60256626a7c110b2b4c80d9f6b0d1ec0b66565bbab978f96b7a6"}, - {file = "Bottleneck-1.3.7-cp36-cp36m-musllinux_1_1_x86_64.whl", hash = "sha256:b3c65a9aedada04872cba450676d132d7af8b9da47f58b82f20ffe9b9ff77a46"}, - {file = "Bottleneck-1.3.7-cp36-cp36m-win32.whl", hash = "sha256:b4a47f972e919d22c1725ff7262d209e6ea3a4d9bcfea71fb454a18e166515ec"}, - {file = "Bottleneck-1.3.7-cp36-cp36m-win_amd64.whl", hash = "sha256:8d6865327ebbea5578f99b073538789df2123cf9009e99fe94efc4d25a4b888b"}, - {file = "Bottleneck-1.3.7-cp37-cp37m-macosx_10_9_x86_64.whl", hash = "sha256:685812254238c2059810f8e25c8215b09795b974f5b1a89f0accda3d93cc8734"}, - {file = "Bottleneck-1.3.7-cp37-cp37m-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:5977f1fd02f01035176a99c01b3fb0eefdaff002c30a4710f7279215b0a317f5"}, - {file = "Bottleneck-1.3.7-cp37-cp37m-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:e26c655736701365d66c64056fc272952d0e2a52718c637254d7dd9a7efa97a8"}, - {file = "Bottleneck-1.3.7-cp37-cp37m-musllinux_1_1_i686.whl", hash = "sha256:d982e1e3f72cc7ca4f558e9ef7db7a97ce68a915b1db8249fe088b3e78974b21"}, - {file = "Bottleneck-1.3.7-cp37-cp37m-musllinux_1_1_x86_64.whl", hash = "sha256:b5ec31e0b52cbfd716255695ec5fe87662a5961d2b8b49936f3608e36a04d926"}, - {file = "Bottleneck-1.3.7-cp37-cp37m-win32.whl", hash = "sha256:f2c8b631552908b11905cf87b4a90dd4af332b0726dd9b49d26d6a0fbb38e6d4"}, - {file = "Bottleneck-1.3.7-cp37-cp37m-win_amd64.whl", hash = "sha256:1045f9b199bba2675bd7642c25dec5daa21a03a5f0444eb863f2790af52b00b0"}, - {file = "Bottleneck-1.3.7-cp38-cp38-macosx_10_9_x86_64.whl", hash = "sha256:5af0c679da9dc41d6a3c436785c664fff41b4c1178d46afaa630620ab31e970f"}, - {file = "Bottleneck-1.3.7-cp38-cp38-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:c65d0a9498b2ccda236b28b1b36bc35402e81ebc4575b6cf44033b7e4460f067"}, - {file = "Bottleneck-1.3.7-cp38-cp38-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:752f4c3fc5b7cc49814d41e9ecf6f228f33ea407bdb8c889da53716b5628abc7"}, - {file = "Bottleneck-1.3.7-cp38-cp38-musllinux_1_1_i686.whl", hash = "sha256:54fc579445688c5309a0af7634db3be4e17659c6b664f9332ec404738e847894"}, - {file = "Bottleneck-1.3.7-cp38-cp38-musllinux_1_1_x86_64.whl", hash = "sha256:175ae3d92142769a5acd169ac2d7df22107b72f6d4c932ff56f5afcef39d5877"}, - {file = "Bottleneck-1.3.7-cp38-cp38-win32.whl", hash = "sha256:64ef9ad7187282745205b735e8c33307ec14c7be39f9971bc3f375104355cfb9"}, - {file = "Bottleneck-1.3.7-cp38-cp38-win_amd64.whl", hash = "sha256:fa80c318b5164e39e4f2d5abf95f5ff42744e3d8535a4c85ede94ccc5fca8f9b"}, - {file = "Bottleneck-1.3.7-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:01e61a512d645193cf4cf4fdacf98d3140c26ace0fcf4c4ed8fdfa366c57e0d1"}, - {file = "Bottleneck-1.3.7-cp39-cp39-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:699fb76ef50cf85be8f8d644b533488de3cec4c6ddf00fd770f73cc5caac9938"}, - {file = "Bottleneck-1.3.7-cp39-cp39-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:6311ee47641ed5fbc543d2e49cbfa90bd9ece208a0c1bdc2c2b14b9132982b2a"}, - {file = "Bottleneck-1.3.7-cp39-cp39-musllinux_1_1_i686.whl", hash = "sha256:a1e50fb506693186a4b8f85ac79df1a5af1ab1f774bf20004280e0658a731a89"}, - {file = "Bottleneck-1.3.7-cp39-cp39-musllinux_1_1_x86_64.whl", hash = "sha256:4d2ae3e1ac0626107c6916783befc39473900bb22a9b275e21550c99abe8283c"}, - {file = "Bottleneck-1.3.7-cp39-cp39-win32.whl", hash = "sha256:68d0f9d32d45f62028ab27d2b51d1a2af72a5ca6a7c3b1f86e9115fedb266300"}, - {file = "Bottleneck-1.3.7-cp39-cp39-win_amd64.whl", hash = "sha256:83d71c49dd9d6b99def958b6ccba3c8b5aac7b90849a5a9fe935648436dd46b9"}, - {file = "Bottleneck-1.3.7.tar.gz", hash = "sha256:e1467e373ad469da340ed0ff283214d6531cc08bfdca2083361a3aa6470681f8"}, + {file = "Bottleneck-1.3.8-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:865c8ed5b798c0198b0b80553e09cc0d890c4f5feb3d81d31661517ca7819fa3"}, + {file = "Bottleneck-1.3.8-cp310-cp310-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:d073a31e259d40b25e29dbba80f73abf38afe98fd730c79dad7edd9a0ad6cff5"}, + {file = "Bottleneck-1.3.8-cp310-cp310-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:b806b277ab47495032822f55f43b8d336e4b7e73f8506ed34d3ea3da6d644abc"}, + {file = "Bottleneck-1.3.8-cp310-cp310-musllinux_1_1_i686.whl", hash = "sha256:770b517609916adeb39d3b1a386a29bc316da03dd61e7ee6e8a38325b80cc327"}, + {file = "Bottleneck-1.3.8-cp310-cp310-musllinux_1_1_x86_64.whl", hash = "sha256:2948502b0394ee419945b55b092585222a505c61d41a874c741be49f2cac056f"}, + {file = "Bottleneck-1.3.8-cp310-cp310-win32.whl", hash = "sha256:271b6333522beb8aee32e640ba49a2064491d2c10317baa58a5996be3dd443e4"}, + {file = "Bottleneck-1.3.8-cp310-cp310-win_amd64.whl", hash = "sha256:d41000ea7ca196b5fd39d6fccd34bf0704c8831731cedd2da2dcae3c6ac49c42"}, + {file = "Bottleneck-1.3.8-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:d0a7f454394cd3642498b6e077e70f4a6b9fd46a8eb908c83ac737fdc9f9a98c"}, + {file = "Bottleneck-1.3.8-cp311-cp311-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:2c4ea8b9024dcb4e83b5c118a3c8faa863ace2ad572849da548a74a8ee4e8f2a"}, + {file = "Bottleneck-1.3.8-cp311-cp311-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:f40724b6e965ff5b88b333d4a10097b1629e60c0db21bb3d08c24d7b1a904a16"}, + {file = "Bottleneck-1.3.8-cp311-cp311-musllinux_1_1_i686.whl", hash = "sha256:4bd7183b8dcca89d0e65abe4507c19667dd31dacfbcc8ed705bad642f26a46e1"}, + {file = "Bottleneck-1.3.8-cp311-cp311-musllinux_1_1_x86_64.whl", hash = "sha256:20aa31a7d9d747c499ace1610a6e1f7aba6e3d4a9923e0312f6b4b6d68a59af3"}, + {file = "Bottleneck-1.3.8-cp311-cp311-win32.whl", hash = "sha256:350520105d9449e6565b3f0c4ce1f80a0b3e4d63695ebbf29db41f62e13f6461"}, + {file = "Bottleneck-1.3.8-cp311-cp311-win_amd64.whl", hash = "sha256:167a278902775defde7dfded6e98e3707dfe54971ffd9aec25c43bc74e4e381a"}, + {file = "Bottleneck-1.3.8-cp312-cp312-macosx_10_9_x86_64.whl", hash = "sha256:c6e93ed45c6c83392f73d0333b310b38772df7eb78c120c1447245691bdedaf4"}, + {file = "Bottleneck-1.3.8-cp312-cp312-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:d3400f47dda0196b5af50b0b0678e33cc8c42e52e55ae0a63cdfed60725659bc"}, + {file = "Bottleneck-1.3.8-cp312-cp312-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:fba5fd1805c71b2eeea50bea93d59be449c4af23ebd8da5f75fd74fd0331e314"}, + {file = "Bottleneck-1.3.8-cp312-cp312-musllinux_1_1_i686.whl", hash = "sha256:60139c5c3d2a9c1454a04af5ee981a9f56548d27fa36f264069b149a6e9b01ed"}, + {file = "Bottleneck-1.3.8-cp312-cp312-musllinux_1_1_x86_64.whl", hash = "sha256:99fab17fa26c811ccad63e208314726e718ae6605314329eca09641954550523"}, + {file = "Bottleneck-1.3.8-cp312-cp312-win32.whl", hash = "sha256:d3ae2bb5d4168912e438e377cc1301fa01df949ba59cd86317b3e00404fd4a97"}, + {file = "Bottleneck-1.3.8-cp312-cp312-win_amd64.whl", hash = "sha256:bcba1d5d5328c50f94852ab521fcb26f35d9e0ccd928d120d56455d1a5bb743f"}, + {file = "Bottleneck-1.3.8-cp37-cp37m-macosx_10_9_x86_64.whl", hash = "sha256:8d01fd5389d3160d54619119987ac24b020fa6810b7b398fff4945892237b3da"}, + {file = "Bottleneck-1.3.8-cp37-cp37m-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:7ca25f0003ef65264942f6306d793e0f270ece8b406c5a293dfc7d878146e9f8"}, + {file = "Bottleneck-1.3.8-cp37-cp37m-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:bf7763cf1516fa388c3587d12182fc1bc1c8089eab1a0a1bf09761f4c41af73c"}, + {file = "Bottleneck-1.3.8-cp37-cp37m-musllinux_1_1_i686.whl", hash = "sha256:38837c022350e2a656453f0e448416b7108cf67baccf11d04a0b3b70a48074dd"}, + {file = "Bottleneck-1.3.8-cp37-cp37m-musllinux_1_1_x86_64.whl", hash = "sha256:84ca5e741fae1c1796744dbdd0d2c1789cb74dd79c12ea8ec5834f83430f8520"}, + {file = "Bottleneck-1.3.8-cp37-cp37m-win32.whl", hash = "sha256:f4dfc22a3450227e692ef2ff4657639c33eec88ad04ee3ce29d1a23a4942da24"}, + {file = "Bottleneck-1.3.8-cp37-cp37m-win_amd64.whl", hash = "sha256:90b87eed152bbd760c4eb11473c2cf036abdb26e2f84caeb00787da74fb08c40"}, + {file = "Bottleneck-1.3.8-cp38-cp38-macosx_10_9_x86_64.whl", hash = "sha256:54a1b5d9d63b2d9f2955f8542eea26c418f97873e0abf86ca52beea0208c9306"}, + {file = "Bottleneck-1.3.8-cp38-cp38-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:019dd142d1e870388fb0b649213a0d8e569cce784326e183deba8f17826edd9f"}, + {file = "Bottleneck-1.3.8-cp38-cp38-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:8b5ed34a540eb7df59f45da659af9f792306637de1c69c95f020294f3b9fc4a8"}, + {file = "Bottleneck-1.3.8-cp38-cp38-musllinux_1_1_i686.whl", hash = "sha256:b69fcd4d818bcf9d53497d8accd0d5f852a447728baaa33b9b7168f8c4221d06"}, + {file = "Bottleneck-1.3.8-cp38-cp38-musllinux_1_1_x86_64.whl", hash = "sha256:02616a830bd477f5ba51103396092da4b9d83cea2e88f5b8069e3f4f7b796704"}, + {file = "Bottleneck-1.3.8-cp38-cp38-win32.whl", hash = "sha256:93d359fb83eb3bdd6635ef6e64835c38ffdc211441fc190549f286e6af98b5f6"}, + {file = "Bottleneck-1.3.8-cp38-cp38-win_amd64.whl", hash = "sha256:51c8bb3dffeb72c14f0382b80de76eabac6726d316babbd48f7e4056267d7910"}, + {file = "Bottleneck-1.3.8-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:84453548b0f722c3be912ce3c6b685917fea842bf1252eeb63714a2c1fd1ffc9"}, + {file = "Bottleneck-1.3.8-cp39-cp39-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:92700867504a213cafa9b8d9be529bd6e18dc83366b2ba00e86e80769b93f678"}, + {file = "Bottleneck-1.3.8-cp39-cp39-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:fadfd2f3931fdff42f4b9867eb02ed7c662d01e6099ff6b347b6ced791450651"}, + {file = "Bottleneck-1.3.8-cp39-cp39-musllinux_1_1_i686.whl", hash = "sha256:cfbc4a3a934b677bfbc37ac8757c4e1264a76262b774259bd3fa8a265dbd668b"}, + {file = "Bottleneck-1.3.8-cp39-cp39-musllinux_1_1_x86_64.whl", hash = "sha256:3c74c18f86a1ffac22280b005df8bb8a58505ac6663c4d6807f39873c17dc347"}, + {file = "Bottleneck-1.3.8-cp39-cp39-win32.whl", hash = "sha256:211f881159e8adb3a57df2263028ae6dc89ec4328bfd43f3421e507406c28654"}, + {file = "Bottleneck-1.3.8-cp39-cp39-win_amd64.whl", hash = "sha256:8615eeb75009ba7c0a112a5a6a5154ed3d61fd6b0879631778b3e42e2d9a6d65"}, + {file = "Bottleneck-1.3.8.tar.gz", hash = "sha256:6780d896969ba7f53c8995ba90c87c548beb3db435dc90c60b9a10ed1ab4d868"}, ] [package.dependencies] @@ -348,42 +359,44 @@ doc = ["gitpython", "numpydoc", "sphinx"] [[package]] name = "cartopy" -version = "0.22.0" +version = "0.23.0" description = "A Python library for cartographic visualizations with Matplotlib" optional = false python-versions = ">=3.9" files = [ - {file = "Cartopy-0.22.0-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:b17ad0b056b9a632b12954864c685febb0e1d8a4a45423b83eec119a603fcb8a"}, - {file = "Cartopy-0.22.0-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:bf7ca13a01782810e8b6a9758ed29755b6ce81b5c53721e19cb6636a43b0f575"}, - {file = "Cartopy-0.22.0-cp310-cp310-manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:604c21046dfbe0c9551b28802901d240a7ce25fcaf0c30db0f230cebf93ae2a5"}, - {file = "Cartopy-0.22.0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:4c37d4e2227cb41c971d2bd2555ce081f765d4eaae852def1c059c1d85c8e645"}, - {file = "Cartopy-0.22.0-cp310-cp310-win_amd64.whl", hash = "sha256:3f9c391bd3ba588de397854556dda175edf59f614bbb6dce18c4981154d97d92"}, - {file = "Cartopy-0.22.0-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:3d9d24ef50991269b57a42bdd4f1156426065fbeb41777c7e28d937e2c4a2ae6"}, - {file = "Cartopy-0.22.0-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:472e6713aad729f6fd38bd301666febd23dd3b744b0530cd80e02e242e3f1c74"}, - {file = "Cartopy-0.22.0-cp311-cp311-manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:319b8244f6d06f600be89ad5eaa85c2524e23b240decd203188da26ff14a346b"}, - {file = "Cartopy-0.22.0-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:a401109b88a6bfec1c13dce9b09dc59993c3af426b1879e1e0ee3afd1c293076"}, - {file = "Cartopy-0.22.0-cp311-cp311-win_amd64.whl", hash = "sha256:8ecebbe13d8488c2ec32e7c40222b9f10ba48e386be30c8ec60c867c122399f9"}, - {file = "Cartopy-0.22.0-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:fbc761b80d2fad8cb6ea1c7d3388862e9fbc53c17f14baf6d4038561c7865e6f"}, - {file = "Cartopy-0.22.0-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:3b25dd55844f36115e6c3632a16ed545cbee011068777b5a7a49ed3e1dbcafcb"}, - {file = "Cartopy-0.22.0-cp39-cp39-manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:eec1e9b701272f4cfc43d0123b9f69fa146d915580ee47eda875770c704bf413"}, - {file = "Cartopy-0.22.0-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:8bdc98a5b196488ea824b6e2f07a3ab63c4f1e99a2ab9fdee514a5f38b00407a"}, - {file = "Cartopy-0.22.0-cp39-cp39-win_amd64.whl", hash = "sha256:c5df7fc17705c93a152f3e82d31adb9c2db50448b38b08c076331c522ce7f844"}, - {file = "Cartopy-0.22.0.tar.gz", hash = "sha256:b300f90120931d43f11ef87c064ea1dacec1b59a4940aa76ebf82cf09548bb49"}, + {file = "Cartopy-0.23.0-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:374e66f816c3bafa48ffdbf6abaefa67063b405fac5f425f9be241cdf3498352"}, + {file = "Cartopy-0.23.0-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:2bae450c4c913796cad0b7ce05aa2fa78d1788de47989f0a03183397648e24be"}, + {file = "Cartopy-0.23.0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:a40437596e8ac5e74575eab822c661f4e725bd995cfd9e445069695fe9086b42"}, + {file = "Cartopy-0.23.0-cp310-cp310-win_amd64.whl", hash = "sha256:3292d6d403137eed80d32014c2f28de6282bed8824213f4b4c2170f388b24a1b"}, + {file = "Cartopy-0.23.0-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:86b07b6794b616674e4e485b8574e9197bca54a4467d28dd01ae0bf178f8dc2b"}, + {file = "Cartopy-0.23.0-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:8dece2aa8d5ff7bf989ded6b5f07c980fb5bb772952bc7cdeab469738abdecee"}, + {file = "Cartopy-0.23.0-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:a9dfd28352dc83d6b4e4cf85d84cb50fc4886d4c1510d61f4c7cf22477d1156f"}, + {file = "Cartopy-0.23.0-cp311-cp311-win_amd64.whl", hash = "sha256:b2671b5354e43220f8e1074e7fe30a8b9f71cb38407c78e51db9c97772f0320b"}, + {file = "Cartopy-0.23.0-cp312-cp312-macosx_10_9_x86_64.whl", hash = "sha256:80b9fd666fd47f6370d29f7ad4e352828d54aaf688a03d0b83b51e141cfd77fa"}, + {file = "Cartopy-0.23.0-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:43e36b8b7e7e373a5698757458fd28fafbbbf5f3ebbe2d378f6a5ec3993d6dc0"}, + {file = "Cartopy-0.23.0-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:550173b91155d4d81cd14b4892cb6cabe3dd32bd34feacaa1ec78c0e56287832"}, + {file = "Cartopy-0.23.0-cp312-cp312-win_amd64.whl", hash = "sha256:55219ee0fb069cc3254426e87382cde03546e86c3f7c6759f076823b1e3a44d9"}, + {file = "Cartopy-0.23.0-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:6279af846bf77d9817ab8792a8e38ca561878f048bba1afdae3e3a30c5432bfd"}, + {file = "Cartopy-0.23.0-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:843bf9dc0a18e1a8eed872c49e8092e8a8109e4dce285ad96752841e21e8161e"}, + {file = "Cartopy-0.23.0-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:350ff8802e2bc617c09bd6148aeb46e841775a846bfaa6e635a212d1eaf5ab66"}, + {file = "Cartopy-0.23.0-cp39-cp39-win_amd64.whl", hash = "sha256:b52ab2274ad7504955854ef8d6f603e41f5d7163d02b29d369cecdbd29c2fda1"}, + {file = "Cartopy-0.23.0.tar.gz", hash = "sha256:231f37b35701f2ba31d94959cca75e6da04c2eea3a7f14ce1c75ee3b0eae7676"}, ] [package.dependencies] -matplotlib = ">=3.4" +matplotlib = ">=3.5" numpy = ">=1.21" packaging = ">=20" -pyproj = ">=3.1.0" -pyshp = ">=2.1" +pyproj = ">=3.3.1" +pyshp = ">=2.3" shapely = ">=1.7" [package.extras] -doc = ["beautifulsoup4", "pydata-sphinx-theme", "sphinx", "sphinx-gallery"] +doc = ["pydata-sphinx-theme", "sphinx", "sphinx-gallery"] ows = ["OWSLib (>=0.20.0)", "pillow (>=6.1.0)"] plotting = ["pillow (>=6.1.0)", "scipy (>=1.3.1)"] speedups = ["fiona", "pykdtree"] +srtm = ["beautifulsoup4"] test = ["coveralls", "pytest (>=5.1.2)", "pytest-cov", "pytest-mpl (>=0.11)", "pytest-xdist"] [[package]] @@ -668,35 +681,32 @@ files = [ [[package]] name = "colorcet" -version = "3.0.1" +version = "3.1.0" description = "Collection of perceptually uniform colormaps" optional = false -python-versions = ">=2.7" +python-versions = ">=3.7" files = [ - {file = "colorcet-3.0.1-py2.py3-none-any.whl", hash = "sha256:8daff01824ee9935fdf762d15c444a67d3e361ad4f8b738ad59ac9bf38f30600"}, - {file = "colorcet-3.0.1.tar.gz", hash = "sha256:51455a20353d12fac91f953772d8409f2474e6a0db1af3fa4f7005f405a2480b"}, + {file = "colorcet-3.1.0-py3-none-any.whl", hash = "sha256:2a7d59cc8d0f7938eeedd08aad3152b5319b4ba3bcb7a612398cc17a384cb296"}, + {file = "colorcet-3.1.0.tar.gz", hash = "sha256:2921b3cd81a2288aaf2d63dbc0ce3c26dcd882e8c389cc505d6886bf7aa9a4eb"}, ] -[package.dependencies] -pyct = ">=0.4.4" - [package.extras] -all = ["bokeh", "flake8", "holoviews", "matplotlib", "nbsite (>=0.7.2rc10)", "nbsmoke (>=0.2.6)", "numpy", "pyct (>=0.4.4)", "pydata-sphinx-theme (<0.9.0)", "pytest (>=2.8.5)", "pytest-cov", "pytest-mpl", "setuptools (>=30.3.0)", "sphinx-copybutton", "wheel"] -build = ["pyct (>=0.4.4)", "setuptools (>=30.3.0)", "wheel"] -doc = ["bokeh", "holoviews", "matplotlib", "nbsite (>=0.7.2rc10)", "numpy", "pydata-sphinx-theme (<0.9.0)", "sphinx-copybutton"] +all = ["colorcet[doc]", "colorcet[examples]", "colorcet[tests-extra]", "colorcet[tests]"] +doc = ["colorcet[examples]", "nbsite (>=0.8.4)", "sphinx-copybutton"] examples = ["bokeh", "holoviews", "matplotlib", "numpy"] -tests = ["flake8", "nbsmoke (>=0.2.6)", "pytest (>=2.8.5)", "pytest-cov"] -tests-extra = ["flake8", "nbsmoke (>=0.2.6)", "pytest (>=2.8.5)", "pytest-cov", "pytest-mpl"] +tests = ["packaging", "pre-commit", "pytest (>=2.8.5)", "pytest-cov"] +tests-examples = ["colorcet[examples]", "nbval"] +tests-extra = ["colorcet[tests]", "pytest-mpl"] [[package]] name = "comm" -version = "0.2.1" +version = "0.2.2" description = "Jupyter Python Comm implementation, for usage in ipykernel, xeus-python etc." optional = false python-versions = ">=3.8" files = [ - {file = "comm-0.2.1-py3-none-any.whl", hash = "sha256:87928485c0dfc0e7976fd89fc1e187023cf587e7c353e4a9b417555b44adf021"}, - {file = "comm-0.2.1.tar.gz", hash = "sha256:0bc91edae1344d39d3661dcbc36937181fdaddb304790458f8b044dbc064b89a"}, + {file = "comm-0.2.2-py3-none-any.whl", hash = "sha256:e6fb86cb70ff661ee8c9c14e7d36d6de3b4066f1441be4063df9c5009f0a64d3"}, + {file = "comm-0.2.2.tar.gz", hash = "sha256:3fd7a84065306e07bea1773df6eb8282de51ba82f77c72f9c85716ab11fe980e"}, ] [package.dependencies] @@ -707,64 +717,64 @@ test = ["pytest"] [[package]] name = "contourpy" -version = "1.2.0" +version = "1.2.1" description = "Python library for calculating contours of 2D quadrilateral grids" optional = false python-versions = ">=3.9" files = [ - {file = "contourpy-1.2.0-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:0274c1cb63625972c0c007ab14dd9ba9e199c36ae1a231ce45d725cbcbfd10a8"}, - {file = "contourpy-1.2.0-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:ab459a1cbbf18e8698399c595a01f6dcc5c138220ca3ea9e7e6126232d102bb4"}, - {file = "contourpy-1.2.0-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:6fdd887f17c2f4572ce548461e4f96396681212d858cae7bd52ba3310bc6f00f"}, - {file = "contourpy-1.2.0-cp310-cp310-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:5d16edfc3fc09968e09ddffada434b3bf989bf4911535e04eada58469873e28e"}, - {file = "contourpy-1.2.0-cp310-cp310-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:1c203f617abc0dde5792beb586f827021069fb6d403d7f4d5c2b543d87edceb9"}, - {file = "contourpy-1.2.0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:b69303ceb2e4d4f146bf82fda78891ef7bcd80c41bf16bfca3d0d7eb545448aa"}, - {file = "contourpy-1.2.0-cp310-cp310-musllinux_1_1_aarch64.whl", hash = "sha256:884c3f9d42d7218304bc74a8a7693d172685c84bd7ab2bab1ee567b769696df9"}, - {file = "contourpy-1.2.0-cp310-cp310-musllinux_1_1_x86_64.whl", hash = "sha256:4a1b1208102be6e851f20066bf0e7a96b7d48a07c9b0cfe6d0d4545c2f6cadab"}, - {file = "contourpy-1.2.0-cp310-cp310-win32.whl", hash = "sha256:34b9071c040d6fe45d9826cbbe3727d20d83f1b6110d219b83eb0e2a01d79488"}, - {file = "contourpy-1.2.0-cp310-cp310-win_amd64.whl", hash = "sha256:bd2f1ae63998da104f16a8b788f685e55d65760cd1929518fd94cd682bf03e41"}, - {file = "contourpy-1.2.0-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:dd10c26b4eadae44783c45ad6655220426f971c61d9b239e6f7b16d5cdaaa727"}, - {file = "contourpy-1.2.0-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:5c6b28956b7b232ae801406e529ad7b350d3f09a4fde958dfdf3c0520cdde0dd"}, - {file = "contourpy-1.2.0-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:ebeac59e9e1eb4b84940d076d9f9a6cec0064e241818bcb6e32124cc5c3e377a"}, - {file = "contourpy-1.2.0-cp311-cp311-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:139d8d2e1c1dd52d78682f505e980f592ba53c9f73bd6be102233e358b401063"}, - {file = "contourpy-1.2.0-cp311-cp311-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:1e9dc350fb4c58adc64df3e0703ab076f60aac06e67d48b3848c23647ae4310e"}, - {file = "contourpy-1.2.0-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:18fc2b4ed8e4a8fe849d18dce4bd3c7ea637758c6343a1f2bae1e9bd4c9f4686"}, - {file = "contourpy-1.2.0-cp311-cp311-musllinux_1_1_aarch64.whl", hash = "sha256:16a7380e943a6d52472096cb7ad5264ecee36ed60888e2a3d3814991a0107286"}, - {file = "contourpy-1.2.0-cp311-cp311-musllinux_1_1_x86_64.whl", hash = "sha256:8d8faf05be5ec8e02a4d86f616fc2a0322ff4a4ce26c0f09d9f7fb5330a35c95"}, - {file = "contourpy-1.2.0-cp311-cp311-win32.whl", hash = "sha256:67b7f17679fa62ec82b7e3e611c43a016b887bd64fb933b3ae8638583006c6d6"}, - {file = "contourpy-1.2.0-cp311-cp311-win_amd64.whl", hash = "sha256:99ad97258985328b4f207a5e777c1b44a83bfe7cf1f87b99f9c11d4ee477c4de"}, - {file = "contourpy-1.2.0-cp312-cp312-macosx_10_9_x86_64.whl", hash = "sha256:575bcaf957a25d1194903a10bc9f316c136c19f24e0985a2b9b5608bdf5dbfe0"}, - {file = "contourpy-1.2.0-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:9e6c93b5b2dbcedad20a2f18ec22cae47da0d705d454308063421a3b290d9ea4"}, - {file = "contourpy-1.2.0-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:464b423bc2a009088f19bdf1f232299e8b6917963e2b7e1d277da5041f33a779"}, - {file = "contourpy-1.2.0-cp312-cp312-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:68ce4788b7d93e47f84edd3f1f95acdcd142ae60bc0e5493bfd120683d2d4316"}, - {file = "contourpy-1.2.0-cp312-cp312-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:3d7d1f8871998cdff5d2ff6a087e5e1780139abe2838e85b0b46b7ae6cc25399"}, - {file = "contourpy-1.2.0-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:6e739530c662a8d6d42c37c2ed52a6f0932c2d4a3e8c1f90692ad0ce1274abe0"}, - {file = "contourpy-1.2.0-cp312-cp312-musllinux_1_1_aarch64.whl", hash = "sha256:247b9d16535acaa766d03037d8e8fb20866d054d3c7fbf6fd1f993f11fc60ca0"}, - {file = "contourpy-1.2.0-cp312-cp312-musllinux_1_1_x86_64.whl", hash = "sha256:461e3ae84cd90b30f8d533f07d87c00379644205b1d33a5ea03381edc4b69431"}, - {file = "contourpy-1.2.0-cp312-cp312-win32.whl", hash = "sha256:1c2559d6cffc94890b0529ea7eeecc20d6fadc1539273aa27faf503eb4656d8f"}, - {file = "contourpy-1.2.0-cp312-cp312-win_amd64.whl", hash = "sha256:491b1917afdd8638a05b611a56d46587d5a632cabead889a5440f7c638bc6ed9"}, - {file = "contourpy-1.2.0-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:5fd1810973a375ca0e097dee059c407913ba35723b111df75671a1976efa04bc"}, - {file = "contourpy-1.2.0-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:999c71939aad2780f003979b25ac5b8f2df651dac7b38fb8ce6c46ba5abe6ae9"}, - {file = "contourpy-1.2.0-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:b7caf9b241464c404613512d5594a6e2ff0cc9cb5615c9475cc1d9b514218ae8"}, - {file = "contourpy-1.2.0-cp39-cp39-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:266270c6f6608340f6c9836a0fb9b367be61dde0c9a9a18d5ece97774105ff3e"}, - {file = "contourpy-1.2.0-cp39-cp39-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:dbd50d0a0539ae2e96e537553aff6d02c10ed165ef40c65b0e27e744a0f10af8"}, - {file = "contourpy-1.2.0-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:11f8d2554e52f459918f7b8e6aa20ec2a3bce35ce95c1f0ef4ba36fbda306df5"}, - {file = "contourpy-1.2.0-cp39-cp39-musllinux_1_1_aarch64.whl", hash = "sha256:ce96dd400486e80ac7d195b2d800b03e3e6a787e2a522bfb83755938465a819e"}, - {file = "contourpy-1.2.0-cp39-cp39-musllinux_1_1_x86_64.whl", hash = "sha256:6d3364b999c62f539cd403f8123ae426da946e142312a514162adb2addd8d808"}, - {file = "contourpy-1.2.0-cp39-cp39-win32.whl", hash = "sha256:1c88dfb9e0c77612febebb6ac69d44a8d81e3dc60f993215425b62c1161353f4"}, - {file = "contourpy-1.2.0-cp39-cp39-win_amd64.whl", hash = "sha256:78e6ad33cf2e2e80c5dfaaa0beec3d61face0fb650557100ee36db808bfa6843"}, - {file = "contourpy-1.2.0-pp39-pypy39_pp73-macosx_10_9_x86_64.whl", hash = "sha256:be16975d94c320432657ad2402f6760990cb640c161ae6da1363051805fa8108"}, - {file = "contourpy-1.2.0-pp39-pypy39_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:b95a225d4948b26a28c08307a60ac00fb8671b14f2047fc5476613252a129776"}, - {file = "contourpy-1.2.0-pp39-pypy39_pp73-win_amd64.whl", hash = "sha256:0d7e03c0f9a4f90dc18d4e77e9ef4ec7b7bbb437f7f675be8e530d65ae6ef956"}, - {file = "contourpy-1.2.0.tar.gz", hash = "sha256:171f311cb758de7da13fc53af221ae47a5877be5a0843a9fe150818c51ed276a"}, + {file = "contourpy-1.2.1-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:bd7c23df857d488f418439686d3b10ae2fbf9bc256cd045b37a8c16575ea1040"}, + {file = "contourpy-1.2.1-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:5b9eb0ca724a241683c9685a484da9d35c872fd42756574a7cfbf58af26677fd"}, + {file = "contourpy-1.2.1-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:4c75507d0a55378240f781599c30e7776674dbaf883a46d1c90f37e563453480"}, + {file = "contourpy-1.2.1-cp310-cp310-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:11959f0ce4a6f7b76ec578576a0b61a28bdc0696194b6347ba3f1c53827178b9"}, + {file = "contourpy-1.2.1-cp310-cp310-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:eb3315a8a236ee19b6df481fc5f997436e8ade24a9f03dfdc6bd490fea20c6da"}, + {file = "contourpy-1.2.1-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:39f3ecaf76cd98e802f094e0d4fbc6dc9c45a8d0c4d185f0f6c2234e14e5f75b"}, + {file = "contourpy-1.2.1-cp310-cp310-musllinux_1_1_aarch64.whl", hash = "sha256:94b34f32646ca0414237168d68a9157cb3889f06b096612afdd296003fdd32fd"}, + {file = "contourpy-1.2.1-cp310-cp310-musllinux_1_1_x86_64.whl", hash = "sha256:457499c79fa84593f22454bbd27670227874cd2ff5d6c84e60575c8b50a69619"}, + {file = "contourpy-1.2.1-cp310-cp310-win32.whl", hash = "sha256:ac58bdee53cbeba2ecad824fa8159493f0bf3b8ea4e93feb06c9a465d6c87da8"}, + {file = "contourpy-1.2.1-cp310-cp310-win_amd64.whl", hash = "sha256:9cffe0f850e89d7c0012a1fb8730f75edd4320a0a731ed0c183904fe6ecfc3a9"}, + {file = "contourpy-1.2.1-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:6022cecf8f44e36af10bd9118ca71f371078b4c168b6e0fab43d4a889985dbb5"}, + {file = "contourpy-1.2.1-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:ef5adb9a3b1d0c645ff694f9bca7702ec2c70f4d734f9922ea34de02294fdf72"}, + {file = "contourpy-1.2.1-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:6150ffa5c767bc6332df27157d95442c379b7dce3a38dff89c0f39b63275696f"}, + {file = "contourpy-1.2.1-cp311-cp311-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:4c863140fafc615c14a4bf4efd0f4425c02230eb8ef02784c9a156461e62c965"}, + {file = "contourpy-1.2.1-cp311-cp311-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:00e5388f71c1a0610e6fe56b5c44ab7ba14165cdd6d695429c5cd94021e390b2"}, + {file = "contourpy-1.2.1-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:d4492d82b3bc7fbb7e3610747b159869468079fe149ec5c4d771fa1f614a14df"}, + {file = "contourpy-1.2.1-cp311-cp311-musllinux_1_1_aarch64.whl", hash = "sha256:49e70d111fee47284d9dd867c9bb9a7058a3c617274900780c43e38d90fe1205"}, + {file = "contourpy-1.2.1-cp311-cp311-musllinux_1_1_x86_64.whl", hash = "sha256:b59c0ffceff8d4d3996a45f2bb6f4c207f94684a96bf3d9728dbb77428dd8cb8"}, + {file = "contourpy-1.2.1-cp311-cp311-win32.whl", hash = "sha256:7b4182299f251060996af5249c286bae9361fa8c6a9cda5efc29fe8bfd6062ec"}, + {file = "contourpy-1.2.1-cp311-cp311-win_amd64.whl", hash = "sha256:2855c8b0b55958265e8b5888d6a615ba02883b225f2227461aa9127c578a4922"}, + {file = "contourpy-1.2.1-cp312-cp312-macosx_10_9_x86_64.whl", hash = "sha256:62828cada4a2b850dbef89c81f5a33741898b305db244904de418cc957ff05dc"}, + {file = "contourpy-1.2.1-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:309be79c0a354afff9ff7da4aaed7c3257e77edf6c1b448a779329431ee79d7e"}, + {file = "contourpy-1.2.1-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:2e785e0f2ef0d567099b9ff92cbfb958d71c2d5b9259981cd9bee81bd194c9a4"}, + {file = "contourpy-1.2.1-cp312-cp312-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:1cac0a8f71a041aa587410424ad46dfa6a11f6149ceb219ce7dd48f6b02b87a7"}, + {file = "contourpy-1.2.1-cp312-cp312-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:af3f4485884750dddd9c25cb7e3915d83c2db92488b38ccb77dd594eac84c4a0"}, + {file = "contourpy-1.2.1-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:9ce6889abac9a42afd07a562c2d6d4b2b7134f83f18571d859b25624a331c90b"}, + {file = "contourpy-1.2.1-cp312-cp312-musllinux_1_1_aarch64.whl", hash = "sha256:a1eea9aecf761c661d096d39ed9026574de8adb2ae1c5bd7b33558af884fb2ce"}, + {file = "contourpy-1.2.1-cp312-cp312-musllinux_1_1_x86_64.whl", hash = "sha256:187fa1d4c6acc06adb0fae5544c59898ad781409e61a926ac7e84b8f276dcef4"}, + {file = "contourpy-1.2.1-cp312-cp312-win32.whl", hash = "sha256:c2528d60e398c7c4c799d56f907664673a807635b857df18f7ae64d3e6ce2d9f"}, + {file = "contourpy-1.2.1-cp312-cp312-win_amd64.whl", hash = "sha256:1a07fc092a4088ee952ddae19a2b2a85757b923217b7eed584fdf25f53a6e7ce"}, + {file = "contourpy-1.2.1-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:bb6834cbd983b19f06908b45bfc2dad6ac9479ae04abe923a275b5f48f1a186b"}, + {file = "contourpy-1.2.1-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:1d59e739ab0e3520e62a26c60707cc3ab0365d2f8fecea74bfe4de72dc56388f"}, + {file = "contourpy-1.2.1-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:bd3db01f59fdcbce5b22afad19e390260d6d0222f35a1023d9adc5690a889364"}, + {file = "contourpy-1.2.1-cp39-cp39-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:a12a813949e5066148712a0626895c26b2578874e4cc63160bb007e6df3436fe"}, + {file = "contourpy-1.2.1-cp39-cp39-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:fe0ccca550bb8e5abc22f530ec0466136379c01321fd94f30a22231e8a48d985"}, + {file = "contourpy-1.2.1-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:e1d59258c3c67c865435d8fbeb35f8c59b8bef3d6f46c1f29f6123556af28445"}, + {file = "contourpy-1.2.1-cp39-cp39-musllinux_1_1_aarch64.whl", hash = "sha256:f32c38afb74bd98ce26de7cc74a67b40afb7b05aae7b42924ea990d51e4dac02"}, + {file = "contourpy-1.2.1-cp39-cp39-musllinux_1_1_x86_64.whl", hash = "sha256:d31a63bc6e6d87f77d71e1abbd7387ab817a66733734883d1fc0021ed9bfa083"}, + {file = "contourpy-1.2.1-cp39-cp39-win32.whl", hash = "sha256:ddcb8581510311e13421b1f544403c16e901c4e8f09083c881fab2be80ee31ba"}, + {file = "contourpy-1.2.1-cp39-cp39-win_amd64.whl", hash = "sha256:10a37ae557aabf2509c79715cd20b62e4c7c28b8cd62dd7d99e5ed3ce28c3fd9"}, + {file = "contourpy-1.2.1-pp39-pypy39_pp73-macosx_10_9_x86_64.whl", hash = "sha256:a31f94983fecbac95e58388210427d68cd30fe8a36927980fab9c20062645609"}, + {file = "contourpy-1.2.1-pp39-pypy39_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:ef2b055471c0eb466033760a521efb9d8a32b99ab907fc8358481a1dd29e3bd3"}, + {file = "contourpy-1.2.1-pp39-pypy39_pp73-win_amd64.whl", hash = "sha256:b33d2bc4f69caedcd0a275329eb2198f560b325605810895627be5d4b876bf7f"}, + {file = "contourpy-1.2.1.tar.gz", hash = "sha256:4d8908b3bee1c889e547867ca4cdc54e5ab6be6d3e078556814a22457f49423c"}, ] [package.dependencies] -numpy = ">=1.20,<2.0" +numpy = ">=1.20" [package.extras] bokeh = ["bokeh", "selenium"] docs = ["furo", "sphinx (>=7.2)", "sphinx-copybutton"] -mypy = ["contourpy[bokeh,docs]", "docutils-stubs", "mypy (==1.6.1)", "types-Pillow"] +mypy = ["contourpy[bokeh,docs]", "docutils-stubs", "mypy (==1.8.0)", "types-Pillow"] test = ["Pillow", "contourpy[test-no-images]", "matplotlib"] test-no-images = ["pytest", "pytest-cov", "pytest-xdist", "wurlitzer"] @@ -784,63 +794,63 @@ coverage = ">=6.0.2" [[package]] name = "coverage" -version = "7.4.1" +version = "7.5.1" description = "Code coverage measurement for Python" optional = false python-versions = ">=3.8" files = [ - {file = "coverage-7.4.1-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:077d366e724f24fc02dbfe9d946534357fda71af9764ff99d73c3c596001bbd7"}, - {file = "coverage-7.4.1-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:0193657651f5399d433c92f8ae264aff31fc1d066deee4b831549526433f3f61"}, - {file = "coverage-7.4.1-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:d17bbc946f52ca67adf72a5ee783cd7cd3477f8f8796f59b4974a9b59cacc9ee"}, - {file = "coverage-7.4.1-cp310-cp310-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:a3277f5fa7483c927fe3a7b017b39351610265308f5267ac6d4c2b64cc1d8d25"}, - {file = "coverage-7.4.1-cp310-cp310-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:6dceb61d40cbfcf45f51e59933c784a50846dc03211054bd76b421a713dcdf19"}, - {file = "coverage-7.4.1-cp310-cp310-musllinux_1_1_aarch64.whl", hash = "sha256:6008adeca04a445ea6ef31b2cbaf1d01d02986047606f7da266629afee982630"}, - {file = "coverage-7.4.1-cp310-cp310-musllinux_1_1_i686.whl", hash = "sha256:c61f66d93d712f6e03369b6a7769233bfda880b12f417eefdd4f16d1deb2fc4c"}, - {file = "coverage-7.4.1-cp310-cp310-musllinux_1_1_x86_64.whl", hash = "sha256:b9bb62fac84d5f2ff523304e59e5c439955fb3b7f44e3d7b2085184db74d733b"}, - {file = "coverage-7.4.1-cp310-cp310-win32.whl", hash = "sha256:f86f368e1c7ce897bf2457b9eb61169a44e2ef797099fb5728482b8d69f3f016"}, - {file = "coverage-7.4.1-cp310-cp310-win_amd64.whl", hash = "sha256:869b5046d41abfea3e381dd143407b0d29b8282a904a19cb908fa24d090cc018"}, - {file = "coverage-7.4.1-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:b8ffb498a83d7e0305968289441914154fb0ef5d8b3157df02a90c6695978295"}, - {file = "coverage-7.4.1-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:3cacfaefe6089d477264001f90f55b7881ba615953414999c46cc9713ff93c8c"}, - {file = "coverage-7.4.1-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:5d6850e6e36e332d5511a48a251790ddc545e16e8beaf046c03985c69ccb2676"}, - {file = "coverage-7.4.1-cp311-cp311-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:18e961aa13b6d47f758cc5879383d27b5b3f3dcd9ce8cdbfdc2571fe86feb4dd"}, - {file = "coverage-7.4.1-cp311-cp311-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:dfd1e1b9f0898817babf840b77ce9fe655ecbe8b1b327983df485b30df8cc011"}, - {file = "coverage-7.4.1-cp311-cp311-musllinux_1_1_aarch64.whl", hash = "sha256:6b00e21f86598b6330f0019b40fb397e705135040dbedc2ca9a93c7441178e74"}, - {file = "coverage-7.4.1-cp311-cp311-musllinux_1_1_i686.whl", hash = "sha256:536d609c6963c50055bab766d9951b6c394759190d03311f3e9fcf194ca909e1"}, - {file = "coverage-7.4.1-cp311-cp311-musllinux_1_1_x86_64.whl", hash = "sha256:7ac8f8eb153724f84885a1374999b7e45734bf93a87d8df1e7ce2146860edef6"}, - {file = "coverage-7.4.1-cp311-cp311-win32.whl", hash = "sha256:f3771b23bb3675a06f5d885c3630b1d01ea6cac9e84a01aaf5508706dba546c5"}, - {file = "coverage-7.4.1-cp311-cp311-win_amd64.whl", hash = "sha256:9d2f9d4cc2a53b38cabc2d6d80f7f9b7e3da26b2f53d48f05876fef7956b6968"}, - {file = "coverage-7.4.1-cp312-cp312-macosx_10_9_x86_64.whl", hash = "sha256:f68ef3660677e6624c8cace943e4765545f8191313a07288a53d3da188bd8581"}, - {file = "coverage-7.4.1-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:23b27b8a698e749b61809fb637eb98ebf0e505710ec46a8aa6f1be7dc0dc43a6"}, - {file = "coverage-7.4.1-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:3e3424c554391dc9ef4a92ad28665756566a28fecf47308f91841f6c49288e66"}, - {file = "coverage-7.4.1-cp312-cp312-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:e0860a348bf7004c812c8368d1fc7f77fe8e4c095d661a579196a9533778e156"}, - {file = "coverage-7.4.1-cp312-cp312-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:fe558371c1bdf3b8fa03e097c523fb9645b8730399c14fe7721ee9c9e2a545d3"}, - {file = "coverage-7.4.1-cp312-cp312-musllinux_1_1_aarch64.whl", hash = "sha256:3468cc8720402af37b6c6e7e2a9cdb9f6c16c728638a2ebc768ba1ef6f26c3a1"}, - {file = "coverage-7.4.1-cp312-cp312-musllinux_1_1_i686.whl", hash = "sha256:02f2edb575d62172aa28fe00efe821ae31f25dc3d589055b3fb64d51e52e4ab1"}, - {file = "coverage-7.4.1-cp312-cp312-musllinux_1_1_x86_64.whl", hash = "sha256:ca6e61dc52f601d1d224526360cdeab0d0712ec104a2ce6cc5ccef6ed9a233bc"}, - {file = "coverage-7.4.1-cp312-cp312-win32.whl", hash = "sha256:ca7b26a5e456a843b9b6683eada193fc1f65c761b3a473941efe5a291f604c74"}, - {file = "coverage-7.4.1-cp312-cp312-win_amd64.whl", hash = "sha256:85ccc5fa54c2ed64bd91ed3b4a627b9cce04646a659512a051fa82a92c04a448"}, - {file = "coverage-7.4.1-cp38-cp38-macosx_10_9_x86_64.whl", hash = "sha256:8bdb0285a0202888d19ec6b6d23d5990410decb932b709f2b0dfe216d031d218"}, - {file = "coverage-7.4.1-cp38-cp38-macosx_11_0_arm64.whl", hash = "sha256:918440dea04521f499721c039863ef95433314b1db00ff826a02580c1f503e45"}, - {file = "coverage-7.4.1-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:379d4c7abad5afbe9d88cc31ea8ca262296480a86af945b08214eb1a556a3e4d"}, - {file = "coverage-7.4.1-cp38-cp38-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:b094116f0b6155e36a304ff912f89bbb5067157aff5f94060ff20bbabdc8da06"}, - {file = "coverage-7.4.1-cp38-cp38-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:f2f5968608b1fe2a1d00d01ad1017ee27efd99b3437e08b83ded9b7af3f6f766"}, - {file = "coverage-7.4.1-cp38-cp38-musllinux_1_1_aarch64.whl", hash = "sha256:10e88e7f41e6197ea0429ae18f21ff521d4f4490aa33048f6c6f94c6045a6a75"}, - {file = "coverage-7.4.1-cp38-cp38-musllinux_1_1_i686.whl", hash = "sha256:a4a3907011d39dbc3e37bdc5df0a8c93853c369039b59efa33a7b6669de04c60"}, - {file = "coverage-7.4.1-cp38-cp38-musllinux_1_1_x86_64.whl", hash = "sha256:6d224f0c4c9c98290a6990259073f496fcec1b5cc613eecbd22786d398ded3ad"}, - {file = "coverage-7.4.1-cp38-cp38-win32.whl", hash = "sha256:23f5881362dcb0e1a92b84b3c2809bdc90db892332daab81ad8f642d8ed55042"}, - {file = "coverage-7.4.1-cp38-cp38-win_amd64.whl", hash = "sha256:a07f61fc452c43cd5328b392e52555f7d1952400a1ad09086c4a8addccbd138d"}, - {file = "coverage-7.4.1-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:8e738a492b6221f8dcf281b67129510835461132b03024830ac0e554311a5c54"}, - {file = "coverage-7.4.1-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:46342fed0fff72efcda77040b14728049200cbba1279e0bf1188f1f2078c1d70"}, - {file = "coverage-7.4.1-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:9641e21670c68c7e57d2053ddf6c443e4f0a6e18e547e86af3fad0795414a628"}, - {file = "coverage-7.4.1-cp39-cp39-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:aeb2c2688ed93b027eb0d26aa188ada34acb22dceea256d76390eea135083950"}, - {file = "coverage-7.4.1-cp39-cp39-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:d12c923757de24e4e2110cf8832d83a886a4cf215c6e61ed506006872b43a6d1"}, - {file = "coverage-7.4.1-cp39-cp39-musllinux_1_1_aarch64.whl", hash = "sha256:0491275c3b9971cdbd28a4595c2cb5838f08036bca31765bad5e17edf900b2c7"}, - {file = "coverage-7.4.1-cp39-cp39-musllinux_1_1_i686.whl", hash = "sha256:8dfc5e195bbef80aabd81596ef52a1277ee7143fe419efc3c4d8ba2754671756"}, - {file = "coverage-7.4.1-cp39-cp39-musllinux_1_1_x86_64.whl", hash = "sha256:1a78b656a4d12b0490ca72651fe4d9f5e07e3c6461063a9b6265ee45eb2bdd35"}, - {file = "coverage-7.4.1-cp39-cp39-win32.whl", hash = "sha256:f90515974b39f4dea2f27c0959688621b46d96d5a626cf9c53dbc653a895c05c"}, - {file = "coverage-7.4.1-cp39-cp39-win_amd64.whl", hash = "sha256:64e723ca82a84053dd7bfcc986bdb34af8d9da83c521c19d6b472bc6880e191a"}, - {file = "coverage-7.4.1-pp38.pp39.pp310-none-any.whl", hash = "sha256:32a8d985462e37cfdab611a6f95b09d7c091d07668fdc26e47a725ee575fe166"}, - {file = "coverage-7.4.1.tar.gz", hash = "sha256:1ed4b95480952b1a26d863e546fa5094564aa0065e1e5f0d4d0041f293251d04"}, + {file = "coverage-7.5.1-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:c0884920835a033b78d1c73b6d3bbcda8161a900f38a488829a83982925f6c2e"}, + {file = "coverage-7.5.1-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:39afcd3d4339329c5f58de48a52f6e4e50f6578dd6099961cf22228feb25f38f"}, + {file = "coverage-7.5.1-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:4a7b0ceee8147444347da6a66be737c9d78f3353b0681715b668b72e79203e4a"}, + {file = "coverage-7.5.1-cp310-cp310-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:4a9ca3f2fae0088c3c71d743d85404cec8df9be818a005ea065495bedc33da35"}, + {file = "coverage-7.5.1-cp310-cp310-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:5fd215c0c7d7aab005221608a3c2b46f58c0285a819565887ee0b718c052aa4e"}, + {file = "coverage-7.5.1-cp310-cp310-musllinux_1_1_aarch64.whl", hash = "sha256:4bf0655ab60d754491004a5efd7f9cccefcc1081a74c9ef2da4735d6ee4a6223"}, + {file = "coverage-7.5.1-cp310-cp310-musllinux_1_1_i686.whl", hash = "sha256:61c4bf1ba021817de12b813338c9be9f0ad5b1e781b9b340a6d29fc13e7c1b5e"}, + {file = "coverage-7.5.1-cp310-cp310-musllinux_1_1_x86_64.whl", hash = "sha256:db66fc317a046556a96b453a58eced5024af4582a8dbdc0c23ca4dbc0d5b3146"}, + {file = "coverage-7.5.1-cp310-cp310-win32.whl", hash = "sha256:b016ea6b959d3b9556cb401c55a37547135a587db0115635a443b2ce8f1c7228"}, + {file = "coverage-7.5.1-cp310-cp310-win_amd64.whl", hash = "sha256:df4e745a81c110e7446b1cc8131bf986157770fa405fe90e15e850aaf7619bc8"}, + {file = "coverage-7.5.1-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:796a79f63eca8814ca3317a1ea443645c9ff0d18b188de470ed7ccd45ae79428"}, + {file = "coverage-7.5.1-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:4fc84a37bfd98db31beae3c2748811a3fa72bf2007ff7902f68746d9757f3746"}, + {file = "coverage-7.5.1-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:6175d1a0559986c6ee3f7fccfc4a90ecd12ba0a383dcc2da30c2b9918d67d8a3"}, + {file = "coverage-7.5.1-cp311-cp311-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:1fc81d5878cd6274ce971e0a3a18a8803c3fe25457165314271cf78e3aae3aa2"}, + {file = "coverage-7.5.1-cp311-cp311-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:556cf1a7cbc8028cb60e1ff0be806be2eded2daf8129b8811c63e2b9a6c43bca"}, + {file = "coverage-7.5.1-cp311-cp311-musllinux_1_1_aarch64.whl", hash = "sha256:9981706d300c18d8b220995ad22627647be11a4276721c10911e0e9fa44c83e8"}, + {file = "coverage-7.5.1-cp311-cp311-musllinux_1_1_i686.whl", hash = "sha256:d7fed867ee50edf1a0b4a11e8e5d0895150e572af1cd6d315d557758bfa9c057"}, + {file = "coverage-7.5.1-cp311-cp311-musllinux_1_1_x86_64.whl", hash = "sha256:ef48e2707fb320c8f139424a596f5b69955a85b178f15af261bab871873bb987"}, + {file = "coverage-7.5.1-cp311-cp311-win32.whl", hash = "sha256:9314d5678dcc665330df5b69c1e726a0e49b27df0461c08ca12674bcc19ef136"}, + {file = "coverage-7.5.1-cp311-cp311-win_amd64.whl", hash = "sha256:5fa567e99765fe98f4e7d7394ce623e794d7cabb170f2ca2ac5a4174437e90dd"}, + {file = "coverage-7.5.1-cp312-cp312-macosx_10_9_x86_64.whl", hash = "sha256:b6cf3764c030e5338e7f61f95bd21147963cf6aa16e09d2f74f1fa52013c1206"}, + {file = "coverage-7.5.1-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:2ec92012fefebee89a6b9c79bc39051a6cb3891d562b9270ab10ecfdadbc0c34"}, + {file = "coverage-7.5.1-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:16db7f26000a07efcf6aea00316f6ac57e7d9a96501e990a36f40c965ec7a95d"}, + {file = "coverage-7.5.1-cp312-cp312-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:beccf7b8a10b09c4ae543582c1319c6df47d78fd732f854ac68d518ee1fb97fa"}, + {file = "coverage-7.5.1-cp312-cp312-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:8748731ad392d736cc9ccac03c9845b13bb07d020a33423fa5b3a36521ac6e4e"}, + {file = "coverage-7.5.1-cp312-cp312-musllinux_1_1_aarch64.whl", hash = "sha256:7352b9161b33fd0b643ccd1f21f3a3908daaddf414f1c6cb9d3a2fd618bf2572"}, + {file = "coverage-7.5.1-cp312-cp312-musllinux_1_1_i686.whl", hash = "sha256:7a588d39e0925f6a2bff87154752481273cdb1736270642aeb3635cb9b4cad07"}, + {file = "coverage-7.5.1-cp312-cp312-musllinux_1_1_x86_64.whl", hash = "sha256:68f962d9b72ce69ea8621f57551b2fa9c70509af757ee3b8105d4f51b92b41a7"}, + {file = "coverage-7.5.1-cp312-cp312-win32.whl", hash = "sha256:f152cbf5b88aaeb836127d920dd0f5e7edff5a66f10c079157306c4343d86c19"}, + {file = "coverage-7.5.1-cp312-cp312-win_amd64.whl", hash = "sha256:5a5740d1fb60ddf268a3811bcd353de34eb56dc24e8f52a7f05ee513b2d4f596"}, + {file = "coverage-7.5.1-cp38-cp38-macosx_10_9_x86_64.whl", hash = "sha256:e2213def81a50519d7cc56ed643c9e93e0247f5bbe0d1247d15fa520814a7cd7"}, + {file = "coverage-7.5.1-cp38-cp38-macosx_11_0_arm64.whl", hash = "sha256:5037f8fcc2a95b1f0e80585bd9d1ec31068a9bcb157d9750a172836e98bc7a90"}, + {file = "coverage-7.5.1-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:5c3721c2c9e4c4953a41a26c14f4cef64330392a6d2d675c8b1db3b645e31f0e"}, + {file = "coverage-7.5.1-cp38-cp38-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:ca498687ca46a62ae590253fba634a1fe9836bc56f626852fb2720f334c9e4e5"}, + {file = "coverage-7.5.1-cp38-cp38-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:0cdcbc320b14c3e5877ee79e649677cb7d89ef588852e9583e6b24c2e5072661"}, + {file = "coverage-7.5.1-cp38-cp38-musllinux_1_1_aarch64.whl", hash = "sha256:57e0204b5b745594e5bc14b9b50006da722827f0b8c776949f1135677e88d0b8"}, + {file = "coverage-7.5.1-cp38-cp38-musllinux_1_1_i686.whl", hash = "sha256:8fe7502616b67b234482c3ce276ff26f39ffe88adca2acf0261df4b8454668b4"}, + {file = "coverage-7.5.1-cp38-cp38-musllinux_1_1_x86_64.whl", hash = "sha256:9e78295f4144f9dacfed4f92935fbe1780021247c2fabf73a819b17f0ccfff8d"}, + {file = "coverage-7.5.1-cp38-cp38-win32.whl", hash = "sha256:1434e088b41594baa71188a17533083eabf5609e8e72f16ce8c186001e6b8c41"}, + {file = "coverage-7.5.1-cp38-cp38-win_amd64.whl", hash = "sha256:0646599e9b139988b63704d704af8e8df7fa4cbc4a1f33df69d97f36cb0a38de"}, + {file = "coverage-7.5.1-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:4cc37def103a2725bc672f84bd939a6fe4522310503207aae4d56351644682f1"}, + {file = "coverage-7.5.1-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:fc0b4d8bfeabd25ea75e94632f5b6e047eef8adaed0c2161ada1e922e7f7cece"}, + {file = "coverage-7.5.1-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:0d0a0f5e06881ecedfe6f3dd2f56dcb057b6dbeb3327fd32d4b12854df36bf26"}, + {file = "coverage-7.5.1-cp39-cp39-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:9735317685ba6ec7e3754798c8871c2f49aa5e687cc794a0b1d284b2389d1bd5"}, + {file = "coverage-7.5.1-cp39-cp39-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:d21918e9ef11edf36764b93101e2ae8cc82aa5efdc7c5a4e9c6c35a48496d601"}, + {file = "coverage-7.5.1-cp39-cp39-musllinux_1_1_aarch64.whl", hash = "sha256:c3e757949f268364b96ca894b4c342b41dc6f8f8b66c37878aacef5930db61be"}, + {file = "coverage-7.5.1-cp39-cp39-musllinux_1_1_i686.whl", hash = "sha256:79afb6197e2f7f60c4824dd4b2d4c2ec5801ceb6ba9ce5d2c3080e5660d51a4f"}, + {file = "coverage-7.5.1-cp39-cp39-musllinux_1_1_x86_64.whl", hash = "sha256:d1d0d98d95dd18fe29dc66808e1accf59f037d5716f86a501fc0256455219668"}, + {file = "coverage-7.5.1-cp39-cp39-win32.whl", hash = "sha256:1cc0fe9b0b3a8364093c53b0b4c0c2dd4bb23acbec4c9240b5f284095ccf7981"}, + {file = "coverage-7.5.1-cp39-cp39-win_amd64.whl", hash = "sha256:dde0070c40ea8bb3641e811c1cfbf18e265d024deff6de52c5950677a8fb1e0f"}, + {file = "coverage-7.5.1-pp38.pp39.pp310-none-any.whl", hash = "sha256:6537e7c10cc47c595828b8a8be04c72144725c383c4702703ff4e42e44577312"}, + {file = "coverage-7.5.1.tar.gz", hash = "sha256:54de9ef3a9da981f7af93eafde4ede199e0846cd819eb27c88e2b712aae9708c"}, ] [package.dependencies] @@ -866,22 +876,23 @@ tests = ["pytest", "pytest-cov", "pytest-xdist"] [[package]] name = "dask" -version = "2024.1.1" +version = "2024.5.1" description = "Parallel PyData with Task Scheduling" optional = false python-versions = ">=3.9" files = [ - {file = "dask-2024.1.1-py3-none-any.whl", hash = "sha256:860ce2797905095beff0187c214840b80c77d752dcb9098a8283e3655a762bf5"}, - {file = "dask-2024.1.1.tar.gz", hash = "sha256:d0dc92e81ce68594a0a0ce23ba33f4d648f2c2f4217ab9b79068b7ecfb0416c7"}, + {file = "dask-2024.5.1-py3-none-any.whl", hash = "sha256:af1cadd1fd1d1d44600ff5de43dd029e5668fdf87422131f4e3e3aa2a6a63555"}, + {file = "dask-2024.5.1.tar.gz", hash = "sha256:e071fda67031c314569e37ca70b3e88bb30f1d91ff8ee4122b541845847cc264"}, ] [package.dependencies] bokeh = {version = ">=2.4.2", optional = true, markers = "extra == \"diagnostics\""} click = ">=8.1" cloudpickle = ">=1.5.0" -distributed = {version = "2024.1.1", optional = true, markers = "extra == \"distributed\""} +dask-expr = {version = ">=1.1,<1.2", optional = true, markers = "extra == \"dataframe\""} +distributed = {version = "2024.5.1", optional = true, markers = "extra == \"distributed\""} fsspec = ">=2021.09.0" -importlib-metadata = ">=4.13.0" +importlib-metadata = {version = ">=4.13.0", markers = "python_version < \"3.12\""} jinja2 = {version = ">=2.10.3", optional = true, markers = "extra == \"diagnostics\""} lz4 = {version = ">=4.3.2", optional = true, markers = "extra == \"complete\""} numpy = {version = ">=1.21", optional = true, markers = "extra == \"array\""} @@ -896,20 +907,36 @@ toolz = ">=0.10.0" [package.extras] array = ["numpy (>=1.21)"] complete = ["dask[array,dataframe,diagnostics,distributed]", "lz4 (>=4.3.2)", "pyarrow (>=7.0)", "pyarrow-hotfix"] -dataframe = ["dask[array]", "pandas (>=1.3)"] +dataframe = ["dask-expr (>=1.1,<1.2)", "dask[array]", "pandas (>=1.3)"] diagnostics = ["bokeh (>=2.4.2)", "jinja2 (>=2.10.3)"] -distributed = ["distributed (==2024.1.1)"] +distributed = ["distributed (==2024.5.1)"] test = ["pandas[test]", "pre-commit", "pytest", "pytest-cov", "pytest-rerunfailures", "pytest-timeout", "pytest-xdist"] +[[package]] +name = "dask-expr" +version = "1.1.1" +description = "High Level Expressions for Dask" +optional = false +python-versions = ">=3.9" +files = [ + {file = "dask_expr-1.1.1-py3-none-any.whl", hash = "sha256:455842bff0795309f75f78f98d9be235210e01ad93da9b5a74f8cdaa594ab958"}, + {file = "dask_expr-1.1.1.tar.gz", hash = "sha256:30acb27c19ef3f44681cddd46597395b2da976d426e050e684e2044ebe033f3a"}, +] + +[package.dependencies] +dask = "2024.5.1" +pandas = ">=2" +pyarrow = ">=7.0.0" + [[package]] name = "datashader" -version = "0.16.0" +version = "0.16.1" description = "Data visualization toolchain based on aggregating into a grid" optional = false python-versions = ">=3.9" files = [ - {file = "datashader-0.16.0-py2.py3-none-any.whl", hash = "sha256:a2cb0f839067bf29cf6cc9c07a1dad35f0e4aed3768387056fcbac8748087bfa"}, - {file = "datashader-0.16.0.tar.gz", hash = "sha256:ed4c111957578dcb3fcff972d954f77586dafd71a7345fd5cd069d9fb050d0d1"}, + {file = "datashader-0.16.1-py2.py3-none-any.whl", hash = "sha256:317c16a32a7d1c5208444a83fd6d2cb68d9bfb81802c323489bd24f6b7cf5c5f"}, + {file = "datashader-0.16.1.tar.gz", hash = "sha256:5b2f1dd448defce206ecc16e9960fe9378f32307ea1fcb0c298c22c14f58c881"}, ] [package.dependencies] @@ -928,39 +955,43 @@ toolz = "" xarray = "" [package.extras] -all = ["bokeh (>3.1)", "codecov", "dask-geopandas", "fastparquet", "flake8", "geodatasets", "geopandas", "graphviz", "holoviews", "matplotlib (>=3.3)", "nbconvert", "nbformat", "nbsite (>=0.8.2,<0.9.0)", "nbsmoke[verify] (>0.5)", "netcdf4", "networkx", "numpydoc", "panel (>1.1)", "pyarrow", "pytest", "pytest-benchmark", "pytest-cov", "python-graphviz", "python-snappy", "rasterio", "rioxarray", "scikit-image", "shapely (>=2.0.0)", "spatialpandas", "streamz"] -doc = ["bokeh (>3.1)", "dask-geopandas", "fastparquet", "geodatasets", "geopandas", "graphviz", "holoviews", "matplotlib (>=3.3)", "nbsite (>=0.8.2,<0.9.0)", "networkx", "numpydoc", "panel (>1.1)", "python-graphviz", "python-snappy", "rasterio", "scikit-image", "shapely (>=2.0.0)", "spatialpandas", "streamz"] +all = ["bokeh (>3.1)", "codecov", "dask-expr", "dask-geopandas", "fastparquet", "flake8", "geodatasets", "geopandas", "graphviz", "holoviews", "matplotlib (>=3.3)", "nbconvert", "nbformat", "nbsite (>=0.8.4,<0.9.0)", "nbsmoke[verify] (>0.5)", "netcdf4", "networkx", "numpydoc", "panel (>1.1)", "pyarrow", "pytest (<8)", "pytest-benchmark", "pytest-cov", "python-graphviz", "python-snappy", "rasterio", "rioxarray", "scikit-image", "shapely (>=2.0.0)", "spatialpandas", "streamz"] +doc = ["bokeh (>3.1)", "dask-geopandas", "fastparquet", "geodatasets", "geopandas", "graphviz", "holoviews", "matplotlib (>=3.3)", "nbsite (>=0.8.4,<0.9.0)", "networkx", "numpydoc", "panel (>1.1)", "python-graphviz", "python-snappy", "rasterio", "scikit-image", "shapely (>=2.0.0)", "spatialpandas", "streamz"] examples = ["bokeh (>3.1)", "dask-geopandas", "geodatasets", "geopandas", "holoviews", "matplotlib (>=3.3)", "panel (>1.1)", "scikit-image", "shapely (>=2.0.0)", "spatialpandas"] examples-extra = ["bokeh (>3.1)", "dask-geopandas", "fastparquet", "geodatasets", "geopandas", "graphviz", "holoviews", "matplotlib (>=3.3)", "networkx", "panel (>1.1)", "python-graphviz", "python-snappy", "rasterio", "scikit-image", "shapely (>=2.0.0)", "spatialpandas", "streamz"] geopandas = ["dask-geopandas", "geopandas", "shapely (>=2.0.0)"] gpu-tests = ["cudf", "cupy", "dask-cudf"] -tests = ["codecov", "dask-geopandas", "fastparquet", "flake8", "geodatasets", "geopandas", "nbconvert", "nbformat", "nbsmoke[verify] (>0.5)", "netcdf4", "pyarrow", "pytest", "pytest-benchmark", "pytest-cov", "rasterio", "rioxarray", "scikit-image", "shapely (>=2.0.0)", "spatialpandas"] +tests = ["codecov", "dask-expr", "dask-geopandas", "flake8", "geodatasets", "geopandas", "nbconvert", "nbformat", "nbsmoke[verify] (>0.5)", "netcdf4", "pyarrow", "pytest (<8)", "pytest-benchmark", "pytest-cov", "rasterio", "rioxarray", "scikit-image", "shapely (>=2.0.0)", "spatialpandas"] [[package]] name = "debugpy" -version = "1.8.0" +version = "1.8.1" description = "An implementation of the Debug Adapter Protocol for Python" optional = false python-versions = ">=3.8" files = [ - {file = "debugpy-1.8.0-cp310-cp310-macosx_11_0_x86_64.whl", hash = "sha256:7fb95ca78f7ac43393cd0e0f2b6deda438ec7c5e47fa5d38553340897d2fbdfb"}, - {file = "debugpy-1.8.0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:ef9ab7df0b9a42ed9c878afd3eaaff471fce3fa73df96022e1f5c9f8f8c87ada"}, - {file = "debugpy-1.8.0-cp310-cp310-win32.whl", hash = "sha256:a8b7a2fd27cd9f3553ac112f356ad4ca93338feadd8910277aff71ab24d8775f"}, - {file = "debugpy-1.8.0-cp310-cp310-win_amd64.whl", hash = "sha256:5d9de202f5d42e62f932507ee8b21e30d49aae7e46d5b1dd5c908db1d7068637"}, - {file = "debugpy-1.8.0-cp311-cp311-macosx_11_0_universal2.whl", hash = "sha256:ef54404365fae8d45cf450d0544ee40cefbcb9cb85ea7afe89a963c27028261e"}, - {file = "debugpy-1.8.0-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:60009b132c91951354f54363f8ebdf7457aeb150e84abba5ae251b8e9f29a8a6"}, - {file = "debugpy-1.8.0-cp311-cp311-win32.whl", hash = "sha256:8cd0197141eb9e8a4566794550cfdcdb8b3db0818bdf8c49a8e8f8053e56e38b"}, - {file = "debugpy-1.8.0-cp311-cp311-win_amd64.whl", hash = "sha256:a64093656c4c64dc6a438e11d59369875d200bd5abb8f9b26c1f5f723622e153"}, - {file = "debugpy-1.8.0-cp38-cp38-macosx_11_0_x86_64.whl", hash = "sha256:b05a6b503ed520ad58c8dc682749113d2fd9f41ffd45daec16e558ca884008cd"}, - {file = "debugpy-1.8.0-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:3c6fb41c98ec51dd010d7ed650accfd07a87fe5e93eca9d5f584d0578f28f35f"}, - {file = "debugpy-1.8.0-cp38-cp38-win32.whl", hash = "sha256:46ab6780159eeabb43c1495d9c84cf85d62975e48b6ec21ee10c95767c0590aa"}, - {file = "debugpy-1.8.0-cp38-cp38-win_amd64.whl", hash = "sha256:bdc5ef99d14b9c0fcb35351b4fbfc06ac0ee576aeab6b2511702e5a648a2e595"}, - {file = "debugpy-1.8.0-cp39-cp39-macosx_11_0_x86_64.whl", hash = "sha256:61eab4a4c8b6125d41a34bad4e5fe3d2cc145caecd63c3fe953be4cc53e65bf8"}, - {file = "debugpy-1.8.0-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:125b9a637e013f9faac0a3d6a82bd17c8b5d2c875fb6b7e2772c5aba6d082332"}, - {file = "debugpy-1.8.0-cp39-cp39-win32.whl", hash = "sha256:57161629133113c97b387382045649a2b985a348f0c9366e22217c87b68b73c6"}, - {file = "debugpy-1.8.0-cp39-cp39-win_amd64.whl", hash = "sha256:e3412f9faa9ade82aa64a50b602544efcba848c91384e9f93497a458767e6926"}, - {file = "debugpy-1.8.0-py2.py3-none-any.whl", hash = "sha256:9c9b0ac1ce2a42888199df1a1906e45e6f3c9555497643a85e0bf2406e3ffbc4"}, - {file = "debugpy-1.8.0.zip", hash = "sha256:12af2c55b419521e33d5fb21bd022df0b5eb267c3e178f1d374a63a2a6bdccd0"}, + {file = "debugpy-1.8.1-cp310-cp310-macosx_11_0_x86_64.whl", hash = "sha256:3bda0f1e943d386cc7a0e71bfa59f4137909e2ed947fb3946c506e113000f741"}, + {file = "debugpy-1.8.1-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:dda73bf69ea479c8577a0448f8c707691152e6c4de7f0c4dec5a4bc11dee516e"}, + {file = "debugpy-1.8.1-cp310-cp310-win32.whl", hash = "sha256:3a79c6f62adef994b2dbe9fc2cc9cc3864a23575b6e387339ab739873bea53d0"}, + {file = "debugpy-1.8.1-cp310-cp310-win_amd64.whl", hash = "sha256:7eb7bd2b56ea3bedb009616d9e2f64aab8fc7000d481faec3cd26c98a964bcdd"}, + {file = "debugpy-1.8.1-cp311-cp311-macosx_11_0_universal2.whl", hash = "sha256:016a9fcfc2c6b57f939673c874310d8581d51a0fe0858e7fac4e240c5eb743cb"}, + {file = "debugpy-1.8.1-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:fd97ed11a4c7f6d042d320ce03d83b20c3fb40da892f994bc041bbc415d7a099"}, + {file = "debugpy-1.8.1-cp311-cp311-win32.whl", hash = "sha256:0de56aba8249c28a300bdb0672a9b94785074eb82eb672db66c8144fff673146"}, + {file = "debugpy-1.8.1-cp311-cp311-win_amd64.whl", hash = "sha256:1a9fe0829c2b854757b4fd0a338d93bc17249a3bf69ecf765c61d4c522bb92a8"}, + {file = "debugpy-1.8.1-cp312-cp312-macosx_11_0_universal2.whl", hash = "sha256:3ebb70ba1a6524d19fa7bb122f44b74170c447d5746a503e36adc244a20ac539"}, + {file = "debugpy-1.8.1-cp312-cp312-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:a2e658a9630f27534e63922ebf655a6ab60c370f4d2fc5c02a5b19baf4410ace"}, + {file = "debugpy-1.8.1-cp312-cp312-win32.whl", hash = "sha256:caad2846e21188797a1f17fc09c31b84c7c3c23baf2516fed5b40b378515bbf0"}, + {file = "debugpy-1.8.1-cp312-cp312-win_amd64.whl", hash = "sha256:edcc9f58ec0fd121a25bc950d4578df47428d72e1a0d66c07403b04eb93bcf98"}, + {file = "debugpy-1.8.1-cp38-cp38-macosx_11_0_x86_64.whl", hash = "sha256:7a3afa222f6fd3d9dfecd52729bc2e12c93e22a7491405a0ecbf9e1d32d45b39"}, + {file = "debugpy-1.8.1-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:d915a18f0597ef685e88bb35e5d7ab968964b7befefe1aaea1eb5b2640b586c7"}, + {file = "debugpy-1.8.1-cp38-cp38-win32.whl", hash = "sha256:92116039b5500633cc8d44ecc187abe2dfa9b90f7a82bbf81d079fcdd506bae9"}, + {file = "debugpy-1.8.1-cp38-cp38-win_amd64.whl", hash = "sha256:e38beb7992b5afd9d5244e96ad5fa9135e94993b0c551ceebf3fe1a5d9beb234"}, + {file = "debugpy-1.8.1-cp39-cp39-macosx_11_0_x86_64.whl", hash = "sha256:bfb20cb57486c8e4793d41996652e5a6a885b4d9175dd369045dad59eaacea42"}, + {file = "debugpy-1.8.1-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:efd3fdd3f67a7e576dd869c184c5dd71d9aaa36ded271939da352880c012e703"}, + {file = "debugpy-1.8.1-cp39-cp39-win32.whl", hash = "sha256:58911e8521ca0c785ac7a0539f1e77e0ce2df753f786188f382229278b4cdf23"}, + {file = "debugpy-1.8.1-cp39-cp39-win_amd64.whl", hash = "sha256:6df9aa9599eb05ca179fb0b810282255202a66835c6efb1d112d21ecb830ddd3"}, + {file = "debugpy-1.8.1-py2.py3-none-any.whl", hash = "sha256:28acbe2241222b87e255260c76741e1fbf04fdc3b6d094fcf57b6c6f75ce1242"}, + {file = "debugpy-1.8.1.zip", hash = "sha256:f696d6be15be87aef621917585f9bb94b1dc9e8aced570db1b8a6fc14e8f9b42"}, ] [[package]] @@ -976,19 +1007,19 @@ files = [ [[package]] name = "distributed" -version = "2024.1.1" +version = "2024.5.1" description = "Distributed scheduler for Dask" optional = false python-versions = ">=3.9" files = [ - {file = "distributed-2024.1.1-py3-none-any.whl", hash = "sha256:cf05d3b38e1700339b3e36395729ab62110e723efefaecc21a8260fdc7555cf9"}, - {file = "distributed-2024.1.1.tar.gz", hash = "sha256:28cf5e9f4f07197b03ea8e5272e374ce2b9e9dc6742f6c9b525fd81645213c67"}, + {file = "distributed-2024.5.1-py3-none-any.whl", hash = "sha256:cf77583ce67f6696aa716c30f1cba3403ac866241c00d6e9a0bc1500c6818b40"}, + {file = "distributed-2024.5.1.tar.gz", hash = "sha256:c4e641e5fc014de3b43c584c70f703a7d44557b51b1143db812b8bc861aa84e2"}, ] [package.dependencies] click = ">=8.0" cloudpickle = ">=1.5.0" -dask = "2024.1.1" +dask = "2024.5.1" jinja2 = ">=2.10.3" locket = ">=1.0.0" msgpack = ">=1.0.0" @@ -1025,13 +1056,13 @@ files = [ [[package]] name = "exceptiongroup" -version = "1.2.0" +version = "1.2.1" description = "Backport of PEP 654 (exception groups)" optional = false python-versions = ">=3.7" files = [ - {file = "exceptiongroup-1.2.0-py3-none-any.whl", hash = "sha256:4bfd3996ac73b41e9b9628b04e079f193850720ea5945fc96a08633c66912f14"}, - {file = "exceptiongroup-1.2.0.tar.gz", hash = "sha256:91f5c769735f051a4290d52edd0858999b57e5876e9f85937691bd4c9fa3ed68"}, + {file = "exceptiongroup-1.2.1-py3-none-any.whl", hash = "sha256:5258b9ed329c5bbdd31a309f53cbfb0b155341807f6ff7606a1e801a891b29ad"}, + {file = "exceptiongroup-1.2.1.tar.gz", hash = "sha256:a4785e48b045528f5bfe627b6ad554ff32def154f42372786903b7abcfe1aa16"}, ] [package.extras] @@ -1078,35 +1109,35 @@ devel = ["colorama", "json-spec", "jsonschema", "pylint", "pytest", "pytest-benc [[package]] name = "fiona" -version = "1.9.5" +version = "1.9.6" description = "Fiona reads and writes spatial data files" optional = false python-versions = ">=3.7" files = [ - {file = "fiona-1.9.5-cp310-cp310-macosx_10_15_x86_64.whl", hash = "sha256:5f40a40529ecfca5294260316cf987a0420c77a2f0cf0849f529d1afbccd093e"}, - {file = "fiona-1.9.5-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:374efe749143ecb5cfdd79b585d83917d2bf8ecfbfc6953c819586b336ce9c63"}, - {file = "fiona-1.9.5-cp310-cp310-manylinux2014_x86_64.whl", hash = "sha256:35dae4b0308eb44617cdc4461ceb91f891d944fdebbcba5479efe524ec5db8de"}, - {file = "fiona-1.9.5-cp310-cp310-win_amd64.whl", hash = "sha256:5b4c6a3df53bee8f85bb46685562b21b43346be1fe96419f18f70fa1ab8c561c"}, - {file = "fiona-1.9.5-cp311-cp311-macosx_10_15_x86_64.whl", hash = "sha256:6ad04c1877b9fd742871b11965606c6a52f40706f56a48d66a87cc3073943828"}, - {file = "fiona-1.9.5-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:9fb9a24a8046c724787719e20557141b33049466145fc3e665764ac7caf5748c"}, - {file = "fiona-1.9.5-cp311-cp311-manylinux2014_x86_64.whl", hash = "sha256:d722d7f01a66f4ab6cd08d156df3fdb92f0669cf5f8708ddcb209352f416f241"}, - {file = "fiona-1.9.5-cp311-cp311-win_amd64.whl", hash = "sha256:7ede8ddc798f3d447536080c6db9a5fb73733ad8bdb190cb65eed4e289dd4c50"}, - {file = "fiona-1.9.5-cp312-cp312-macosx_10_15_x86_64.whl", hash = "sha256:8b098054a27c12afac4f819f98cb4d4bf2db9853f70b0c588d7d97d26e128c39"}, - {file = "fiona-1.9.5-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:6d9f29e9bcbb33232ff7fa98b4a3c2234db910c1dc6c4147fc36c0b8b930f2e0"}, - {file = "fiona-1.9.5-cp312-cp312-manylinux2014_x86_64.whl", hash = "sha256:f1af08da4ecea5036cb81c9131946be4404245d1b434b5b24fd3871a1d4030d9"}, - {file = "fiona-1.9.5-cp312-cp312-win_amd64.whl", hash = "sha256:c521e1135c78dec0d7774303e5a1b4c62e0efb0e602bb8f167550ef95e0a2691"}, - {file = "fiona-1.9.5-cp37-cp37m-macosx_10_15_x86_64.whl", hash = "sha256:fce4b1dd98810cabccdaa1828430c7402d283295c2ae31bea4f34188ea9e88d7"}, - {file = "fiona-1.9.5-cp37-cp37m-manylinux2014_x86_64.whl", hash = "sha256:3ea04ec2d8c57b5f81a31200fb352cb3242aa106fc3e328963f30ffbdf0ff7c8"}, - {file = "fiona-1.9.5-cp37-cp37m-win_amd64.whl", hash = "sha256:4877cc745d9e82b12b3eafce3719db75759c27bd8a695521202135b36b58c2e7"}, - {file = "fiona-1.9.5-cp38-cp38-macosx_10_15_x86_64.whl", hash = "sha256:ac2c250f509ec19fad7959d75b531984776517ef3c1222d1cc5b4f962825880b"}, - {file = "fiona-1.9.5-cp38-cp38-macosx_11_0_arm64.whl", hash = "sha256:4df21906235928faad856c288cfea0298e9647f09c9a69a230535cbc8eadfa21"}, - {file = "fiona-1.9.5-cp38-cp38-manylinux2014_x86_64.whl", hash = "sha256:81d502369493687746cb8d3cd77e5ada4447fb71d513721c9a1826e4fb32b23a"}, - {file = "fiona-1.9.5-cp38-cp38-win_amd64.whl", hash = "sha256:ce3b29230ef70947ead4e701f3f82be81082b7f37fd4899009b1445cc8fc276a"}, - {file = "fiona-1.9.5-cp39-cp39-macosx_10_15_x86_64.whl", hash = "sha256:8b53ce8de773fcd5e2e102e833c8c58479edd8796a522f3d83ef9e08b62bfeea"}, - {file = "fiona-1.9.5-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:bd2355e859a1cd24a3e485c6dc5003129f27a2051629def70036535ffa7e16a4"}, - {file = "fiona-1.9.5-cp39-cp39-manylinux2014_x86_64.whl", hash = "sha256:9a2da52f865db1aff0eaf41cdd4c87a7c079b3996514e8e7a1ca38457309e825"}, - {file = "fiona-1.9.5-cp39-cp39-win_amd64.whl", hash = "sha256:cfef6db5b779d463298b1113b50daa6c5b55f26f834dc9e37752116fa17277c1"}, - {file = "fiona-1.9.5.tar.gz", hash = "sha256:99e2604332caa7692855c2ae6ed91e1fffdf9b59449aa8032dd18e070e59a2f7"}, + {file = "fiona-1.9.6-cp310-cp310-macosx_10_15_x86_64.whl", hash = "sha256:63e528b5ea3d8b1038d788e7c65117835c787ba7fdc94b1b42f09c2cbc0aaff2"}, + {file = "fiona-1.9.6-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:918bd27d8625416672e834593970f96dff63215108f81efb876fe5c0bc58a3b4"}, + {file = "fiona-1.9.6-cp310-cp310-manylinux2014_x86_64.whl", hash = "sha256:e313210b30d09ed8f829bf625599e248dadd78622728030221f6526580ff26c5"}, + {file = "fiona-1.9.6-cp310-cp310-win_amd64.whl", hash = "sha256:89095c2d542325ee45894b8837e8048cdbb2f22274934e1be3b673ca628010d7"}, + {file = "fiona-1.9.6-cp311-cp311-macosx_10_15_x86_64.whl", hash = "sha256:98cea6f435843b2119731c6b0470e5b7386aa16b6aa7edabbf1ed93aefe029c3"}, + {file = "fiona-1.9.6-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:f4230eccbd896a79d1ebfa551d84bf90f512f7bcbe1ca61e3f82231321f1a532"}, + {file = "fiona-1.9.6-cp311-cp311-manylinux2014_x86_64.whl", hash = "sha256:48b6218224e96de5e36b5eb259f37160092260e5de0dcd82ca200b1887aa9884"}, + {file = "fiona-1.9.6-cp311-cp311-win_amd64.whl", hash = "sha256:c1dd5fbc29b7303bb87eb683455e8451e1a53bb8faf20ef97fdcd843c9e4a7f6"}, + {file = "fiona-1.9.6-cp312-cp312-macosx_10_15_x86_64.whl", hash = "sha256:42d8a0e5570948d3821c493b6141866d9a4d7a64edad2be4ecbb89f81904baac"}, + {file = "fiona-1.9.6-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:39819fb8f5ec6d9971cb01b912b4431615a3d3f50c83798565d8ce41917930db"}, + {file = "fiona-1.9.6-cp312-cp312-manylinux2014_x86_64.whl", hash = "sha256:9b53034efdf93ada9295b081e6a8280af7c75496a20df82d4c2ca46d65b85905"}, + {file = "fiona-1.9.6-cp312-cp312-win_amd64.whl", hash = "sha256:1dcd6eca7524535baf2a39d7981b4a46d33ae28c313934a7c3eae62eecf9dfa5"}, + {file = "fiona-1.9.6-cp37-cp37m-macosx_10_15_x86_64.whl", hash = "sha256:e5404ed08c711489abcb3a50a184816825b8af06eb73ad2a99e18b8e7b47c96a"}, + {file = "fiona-1.9.6-cp37-cp37m-manylinux2014_x86_64.whl", hash = "sha256:53bedd2989e255df1bf3378ae9c06d6d241ec273c280c544bb44ffffebb97fb0"}, + {file = "fiona-1.9.6-cp37-cp37m-win_amd64.whl", hash = "sha256:77653a08564a44e634c44cd74a068d2f55d1d4029edd16d1c8aadcc4d8cc1d2c"}, + {file = "fiona-1.9.6-cp38-cp38-macosx_10_15_x86_64.whl", hash = "sha256:e7617563b36d2be99f048f0d0054b4d765f4aae454398f88f19de9c2c324b7f8"}, + {file = "fiona-1.9.6-cp38-cp38-macosx_11_0_arm64.whl", hash = "sha256:50037c3b7a5f6f434b562b5b1a5b664f1caa7a4383b00af23cdb59bfc6ba852c"}, + {file = "fiona-1.9.6-cp38-cp38-manylinux2014_x86_64.whl", hash = "sha256:bf51846ad602757bf27876f458c5c9f14b09421fac612f64273cc4e3fcabc441"}, + {file = "fiona-1.9.6-cp38-cp38-win_amd64.whl", hash = "sha256:11af1afc1255642a7787fe112c29d01f968f1053e4d4700fc6f3bb879c1622e0"}, + {file = "fiona-1.9.6-cp39-cp39-macosx_10_15_x86_64.whl", hash = "sha256:52e8fec650b72fc5253d8f86b63859acc687182281c29bfacd3930496cf982d1"}, + {file = "fiona-1.9.6-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:c9b92aa1badb2773e7cac19bef3064d73e9d80c67c42f0928db2520a04be6f2f"}, + {file = "fiona-1.9.6-cp39-cp39-manylinux2014_x86_64.whl", hash = "sha256:0eaffbf3bfae9960484c0c08ea461b0c40e111497f04e9475ebf15ac7a22d9dc"}, + {file = "fiona-1.9.6-cp39-cp39-win_amd64.whl", hash = "sha256:f1b49d51a744874608b689f029766aa1e078dd72e94b44cf8eeef6d7bd2e9051"}, + {file = "fiona-1.9.6.tar.gz", hash = "sha256:791b3494f8b218c06ea56f892bd6ba893dfa23525347761d066fb7738acda3b1"}, ] [package.dependencies] @@ -1116,24 +1147,23 @@ click = ">=8.0,<9.0" click-plugins = ">=1.0" cligj = ">=0.5" importlib-metadata = {version = "", markers = "python_version < \"3.10\""} -setuptools = "" six = "" [package.extras] -all = ["Fiona[calc,s3,test]"] +all = ["fiona[calc,s3,test]"] calc = ["shapely"] s3 = ["boto3 (>=1.3.1)"] -test = ["Fiona[s3]", "pytest (>=7)", "pytest-cov", "pytz"] +test = ["fiona[s3]", "pytest (>=7)", "pytest-cov", "pytz"] [[package]] name = "flox" -version = "0.9.0" +version = "0.9.7" description = "GroupBy operations for dask.array" optional = false python-versions = ">=3.9" files = [ - {file = "flox-0.9.0-py3-none-any.whl", hash = "sha256:b775e19c104971e7b869cab45830aadeb182866b59256ce87ea79fc88868e201"}, - {file = "flox-0.9.0.tar.gz", hash = "sha256:a7be517654216d5c531b1fb4ea32508736f9496240a61e3ae2974a0dcfcde89e"}, + {file = "flox-0.9.7-py3-none-any.whl", hash = "sha256:0e15d678c5f3d46fe5c6481519d01ceae40a111133b110e80f3b274881af8497"}, + {file = "flox-0.9.7.tar.gz", hash = "sha256:baa7c0aa9b2836f5cf1b283ce918cf3d61dc9ff0af8bda026a598ba5cc0b7c68"}, ] [package.dependencies] @@ -1150,53 +1180,53 @@ test = ["netCDF4"] [[package]] name = "fonttools" -version = "4.48.1" +version = "4.51.0" description = "Tools to manipulate font files" optional = false python-versions = ">=3.8" files = [ - {file = "fonttools-4.48.1-cp310-cp310-macosx_10_9_universal2.whl", hash = "sha256:702ae93058c81f46461dc4b2c79f11d3c3d8fd7296eaf8f75b4ba5bbf813cd5f"}, - {file = "fonttools-4.48.1-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:97f0a49fa6aa2d6205c6f72f4f98b74ef4b9bfdcb06fd78e6fe6c7af4989b63e"}, - {file = "fonttools-4.48.1-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:d3260db55f1843e57115256e91247ad9f68cb02a434b51262fe0019e95a98738"}, - {file = "fonttools-4.48.1-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:e740a7602c2bb71e1091269b5dbe89549749a8817dc294b34628ffd8b2bf7124"}, - {file = "fonttools-4.48.1-cp310-cp310-musllinux_1_1_aarch64.whl", hash = "sha256:4108b1d247953dd7c90ec8f457a2dec5fceb373485973cc852b14200118a51ee"}, - {file = "fonttools-4.48.1-cp310-cp310-musllinux_1_1_x86_64.whl", hash = "sha256:56339ec557f0c342bddd7c175f5e41c45fc21282bee58a86bd9aa322bec715f2"}, - {file = "fonttools-4.48.1-cp310-cp310-win32.whl", hash = "sha256:bff5b38d0e76eb18e0b8abbf35d384e60b3371be92f7be36128ee3e67483b3ec"}, - {file = "fonttools-4.48.1-cp310-cp310-win_amd64.whl", hash = "sha256:f7449493886da6a17472004d3818cc050ba3f4a0aa03fb47972e4fa5578e6703"}, - {file = "fonttools-4.48.1-cp311-cp311-macosx_10_9_universal2.whl", hash = "sha256:18b35fd1a850ed7233a99bbd6774485271756f717dac8b594958224b54118b61"}, - {file = "fonttools-4.48.1-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:cad5cfd044ea2e306fda44482b3dd32ee47830fa82dfa4679374b41baa294f5f"}, - {file = "fonttools-4.48.1-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:6f30e605c7565d0da6f0aec75a30ec372072d016957cd8fc4469721a36ea59b7"}, - {file = "fonttools-4.48.1-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:aee76fd81a8571c68841d6ef0da750d5ff08ff2c5f025576473016f16ac3bcf7"}, - {file = "fonttools-4.48.1-cp311-cp311-musllinux_1_1_aarch64.whl", hash = "sha256:5057ade278e67923000041e2b195c9ea53e87f227690d499b6a4edd3702f7f01"}, - {file = "fonttools-4.48.1-cp311-cp311-musllinux_1_1_x86_64.whl", hash = "sha256:b10633aafc5932995a391ec07eba5e79f52af0003a1735b2306b3dab8a056d48"}, - {file = "fonttools-4.48.1-cp311-cp311-win32.whl", hash = "sha256:0d533f89819f9b3ee2dbedf0fed3825c425850e32bdda24c558563c71be0064e"}, - {file = "fonttools-4.48.1-cp311-cp311-win_amd64.whl", hash = "sha256:d20588466367f05025bb1efdf4e5d498ca6d14bde07b6928b79199c588800f0a"}, - {file = "fonttools-4.48.1-cp312-cp312-macosx_10_9_universal2.whl", hash = "sha256:0a2417547462e468edf35b32e3dd06a6215ac26aa6316b41e03b8eeaf9f079ea"}, - {file = "fonttools-4.48.1-cp312-cp312-macosx_10_9_x86_64.whl", hash = "sha256:cf5a0cd974f85a80b74785db2d5c3c1fd6cc09a2ba3c837359b2b5da629ee1b0"}, - {file = "fonttools-4.48.1-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:0452fcbfbce752ba596737a7c5ec5cf76bc5f83847ce1781f4f90eab14ece252"}, - {file = "fonttools-4.48.1-cp312-cp312-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:578c00f93868f64a4102ecc5aa600a03b49162c654676c3fadc33de2ddb88a81"}, - {file = "fonttools-4.48.1-cp312-cp312-musllinux_1_1_aarch64.whl", hash = "sha256:63dc592a16cd08388d8c4c7502b59ac74190b23e16dfc863c69fe1ea74605b68"}, - {file = "fonttools-4.48.1-cp312-cp312-musllinux_1_1_x86_64.whl", hash = "sha256:9b58638d8a85e3a1b32ec0a91d9f8171a877b4b81c408d4cb3257d0dee63e092"}, - {file = "fonttools-4.48.1-cp312-cp312-win32.whl", hash = "sha256:d10979ef14a8beaaa32f613bb698743f7241d92f437a3b5e32356dfb9769c65d"}, - {file = "fonttools-4.48.1-cp312-cp312-win_amd64.whl", hash = "sha256:cdfd7557d1bd294a200bd211aa665ca3b02998dcc18f8211a5532da5b8fad5c5"}, - {file = "fonttools-4.48.1-cp38-cp38-macosx_10_9_universal2.whl", hash = "sha256:3cdb9a92521b81bf717ebccf592bd0292e853244d84115bfb4db0c426de58348"}, - {file = "fonttools-4.48.1-cp38-cp38-macosx_10_9_x86_64.whl", hash = "sha256:9b4ec6d42a7555f5ae35f3b805482f0aad0f1baeeef54859492ea3b782959d4a"}, - {file = "fonttools-4.48.1-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:902e9c4e9928301912f34a6638741b8ae0b64824112b42aaf240e06b735774b1"}, - {file = "fonttools-4.48.1-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:a8c8b54bd1420c184a995f980f1a8076f87363e2bb24239ef8c171a369d85a31"}, - {file = "fonttools-4.48.1-cp38-cp38-musllinux_1_1_aarch64.whl", hash = "sha256:12ee86abca46193359ea69216b3a724e90c66ab05ab220d39e3fc068c1eb72ac"}, - {file = "fonttools-4.48.1-cp38-cp38-musllinux_1_1_x86_64.whl", hash = "sha256:6978bade7b6c0335095bdd0bd97f8f3d590d2877b370f17e03e0865241694eb5"}, - {file = "fonttools-4.48.1-cp38-cp38-win32.whl", hash = "sha256:bcd77f89fc1a6b18428e7a55dde8ef56dae95640293bfb8f4e929929eba5e2a2"}, - {file = "fonttools-4.48.1-cp38-cp38-win_amd64.whl", hash = "sha256:f40441437b039930428e04fb05ac3a132e77458fb57666c808d74a556779e784"}, - {file = "fonttools-4.48.1-cp39-cp39-macosx_10_9_universal2.whl", hash = "sha256:0d2b01428f7da26f229a5656defc824427b741e454b4e210ad2b25ed6ea2aed4"}, - {file = "fonttools-4.48.1-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:df48798f9a4fc4c315ab46e17873436c8746f5df6eddd02fad91299b2af7af95"}, - {file = "fonttools-4.48.1-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:2eb4167bde04e172a93cf22c875d8b0cff76a2491f67f5eb069566215302d45d"}, - {file = "fonttools-4.48.1-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:c900508c46274d32d308ae8e82335117f11aaee1f7d369ac16502c9a78930b0a"}, - {file = "fonttools-4.48.1-cp39-cp39-musllinux_1_1_aarch64.whl", hash = "sha256:594206b31c95fcfa65f484385171fabb4ec69f7d2d7f56d27f17db26b7a31814"}, - {file = "fonttools-4.48.1-cp39-cp39-musllinux_1_1_x86_64.whl", hash = "sha256:292922dc356d7f11f5063b4111a8b719efb8faea92a2a88ed296408d449d8c2e"}, - {file = "fonttools-4.48.1-cp39-cp39-win32.whl", hash = "sha256:4709c5bf123ba10eac210d2d5c9027d3f472591d9f1a04262122710fa3d23199"}, - {file = "fonttools-4.48.1-cp39-cp39-win_amd64.whl", hash = "sha256:63c73b9dd56a94a3cbd2f90544b5fca83666948a9e03370888994143b8d7c070"}, - {file = "fonttools-4.48.1-py3-none-any.whl", hash = "sha256:e3e33862fc5261d46d9aae3544acb36203b1a337d00bdb5d3753aae50dac860e"}, - {file = "fonttools-4.48.1.tar.gz", hash = "sha256:8b8a45254218679c7f1127812761e7854ed5c8e34349aebf581e8c9204e7495a"}, + {file = "fonttools-4.51.0-cp310-cp310-macosx_10_9_universal2.whl", hash = "sha256:84d7751f4468dd8cdd03ddada18b8b0857a5beec80bce9f435742abc9a851a74"}, + {file = "fonttools-4.51.0-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:8b4850fa2ef2cfbc1d1f689bc159ef0f45d8d83298c1425838095bf53ef46308"}, + {file = "fonttools-4.51.0-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:b5b48a1121117047d82695d276c2af2ee3a24ffe0f502ed581acc2673ecf1037"}, + {file = "fonttools-4.51.0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:180194c7fe60c989bb627d7ed5011f2bef1c4d36ecf3ec64daec8302f1ae0716"}, + {file = "fonttools-4.51.0-cp310-cp310-musllinux_1_1_aarch64.whl", hash = "sha256:96a48e137c36be55e68845fc4284533bda2980f8d6f835e26bca79d7e2006438"}, + {file = "fonttools-4.51.0-cp310-cp310-musllinux_1_1_x86_64.whl", hash = "sha256:806e7912c32a657fa39d2d6eb1d3012d35f841387c8fc6cf349ed70b7c340039"}, + {file = "fonttools-4.51.0-cp310-cp310-win32.whl", hash = "sha256:32b17504696f605e9e960647c5f64b35704782a502cc26a37b800b4d69ff3c77"}, + {file = "fonttools-4.51.0-cp310-cp310-win_amd64.whl", hash = "sha256:c7e91abdfae1b5c9e3a543f48ce96013f9a08c6c9668f1e6be0beabf0a569c1b"}, + {file = "fonttools-4.51.0-cp311-cp311-macosx_10_9_universal2.whl", hash = "sha256:a8feca65bab31479d795b0d16c9a9852902e3a3c0630678efb0b2b7941ea9c74"}, + {file = "fonttools-4.51.0-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:8ac27f436e8af7779f0bb4d5425aa3535270494d3bc5459ed27de3f03151e4c2"}, + {file = "fonttools-4.51.0-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:0e19bd9e9964a09cd2433a4b100ca7f34e34731e0758e13ba9a1ed6e5468cc0f"}, + {file = "fonttools-4.51.0-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:b2b92381f37b39ba2fc98c3a45a9d6383bfc9916a87d66ccb6553f7bdd129097"}, + {file = "fonttools-4.51.0-cp311-cp311-musllinux_1_1_aarch64.whl", hash = "sha256:5f6bc991d1610f5c3bbe997b0233cbc234b8e82fa99fc0b2932dc1ca5e5afec0"}, + {file = "fonttools-4.51.0-cp311-cp311-musllinux_1_1_x86_64.whl", hash = "sha256:9696fe9f3f0c32e9a321d5268208a7cc9205a52f99b89479d1b035ed54c923f1"}, + {file = "fonttools-4.51.0-cp311-cp311-win32.whl", hash = "sha256:3bee3f3bd9fa1d5ee616ccfd13b27ca605c2b4270e45715bd2883e9504735034"}, + {file = "fonttools-4.51.0-cp311-cp311-win_amd64.whl", hash = "sha256:0f08c901d3866a8905363619e3741c33f0a83a680d92a9f0e575985c2634fcc1"}, + {file = "fonttools-4.51.0-cp312-cp312-macosx_10_9_universal2.whl", hash = "sha256:4060acc2bfa2d8e98117828a238889f13b6f69d59f4f2d5857eece5277b829ba"}, + {file = "fonttools-4.51.0-cp312-cp312-macosx_10_9_x86_64.whl", hash = "sha256:1250e818b5f8a679ad79660855528120a8f0288f8f30ec88b83db51515411fcc"}, + {file = "fonttools-4.51.0-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:76f1777d8b3386479ffb4a282e74318e730014d86ce60f016908d9801af9ca2a"}, + {file = "fonttools-4.51.0-cp312-cp312-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:8b5ad456813d93b9c4b7ee55302208db2b45324315129d85275c01f5cb7e61a2"}, + {file = "fonttools-4.51.0-cp312-cp312-musllinux_1_1_aarch64.whl", hash = "sha256:68b3fb7775a923be73e739f92f7e8a72725fd333eab24834041365d2278c3671"}, + {file = "fonttools-4.51.0-cp312-cp312-musllinux_1_1_x86_64.whl", hash = "sha256:8e2f1a4499e3b5ee82c19b5ee57f0294673125c65b0a1ff3764ea1f9db2f9ef5"}, + {file = "fonttools-4.51.0-cp312-cp312-win32.whl", hash = "sha256:278e50f6b003c6aed19bae2242b364e575bcb16304b53f2b64f6551b9c000e15"}, + {file = "fonttools-4.51.0-cp312-cp312-win_amd64.whl", hash = "sha256:b3c61423f22165541b9403ee39874dcae84cd57a9078b82e1dce8cb06b07fa2e"}, + {file = "fonttools-4.51.0-cp38-cp38-macosx_10_9_universal2.whl", hash = "sha256:1621ee57da887c17312acc4b0e7ac30d3a4fb0fec6174b2e3754a74c26bbed1e"}, + {file = "fonttools-4.51.0-cp38-cp38-macosx_10_9_x86_64.whl", hash = "sha256:e9d9298be7a05bb4801f558522adbe2feea1b0b103d5294ebf24a92dd49b78e5"}, + {file = "fonttools-4.51.0-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:ee1af4be1c5afe4c96ca23badd368d8dc75f611887fb0c0dac9f71ee5d6f110e"}, + {file = "fonttools-4.51.0-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:c18b49adc721a7d0b8dfe7c3130c89b8704baf599fb396396d07d4aa69b824a1"}, + {file = "fonttools-4.51.0-cp38-cp38-musllinux_1_1_aarch64.whl", hash = "sha256:de7c29bdbdd35811f14493ffd2534b88f0ce1b9065316433b22d63ca1cd21f14"}, + {file = "fonttools-4.51.0-cp38-cp38-musllinux_1_1_x86_64.whl", hash = "sha256:cadf4e12a608ef1d13e039864f484c8a968840afa0258b0b843a0556497ea9ed"}, + {file = "fonttools-4.51.0-cp38-cp38-win32.whl", hash = "sha256:aefa011207ed36cd280babfaa8510b8176f1a77261833e895a9d96e57e44802f"}, + {file = "fonttools-4.51.0-cp38-cp38-win_amd64.whl", hash = "sha256:865a58b6e60b0938874af0968cd0553bcd88e0b2cb6e588727117bd099eef836"}, + {file = "fonttools-4.51.0-cp39-cp39-macosx_10_9_universal2.whl", hash = "sha256:60a3409c9112aec02d5fb546f557bca6efa773dcb32ac147c6baf5f742e6258b"}, + {file = "fonttools-4.51.0-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:f7e89853d8bea103c8e3514b9f9dc86b5b4120afb4583b57eb10dfa5afbe0936"}, + {file = "fonttools-4.51.0-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:56fc244f2585d6c00b9bcc59e6593e646cf095a96fe68d62cd4da53dd1287b55"}, + {file = "fonttools-4.51.0-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:0d145976194a5242fdd22df18a1b451481a88071feadf251221af110ca8f00ce"}, + {file = "fonttools-4.51.0-cp39-cp39-musllinux_1_1_aarch64.whl", hash = "sha256:c5b8cab0c137ca229433570151b5c1fc6af212680b58b15abd797dcdd9dd5051"}, + {file = "fonttools-4.51.0-cp39-cp39-musllinux_1_1_x86_64.whl", hash = "sha256:54dcf21a2f2d06ded676e3c3f9f74b2bafded3a8ff12f0983160b13e9f2fb4a7"}, + {file = "fonttools-4.51.0-cp39-cp39-win32.whl", hash = "sha256:0118ef998a0699a96c7b28457f15546815015a2710a1b23a7bf6c1be60c01636"}, + {file = "fonttools-4.51.0-cp39-cp39-win_amd64.whl", hash = "sha256:599bdb75e220241cedc6faebfafedd7670335d2e29620d207dd0378a4e9ccc5a"}, + {file = "fonttools-4.51.0-py3-none-any.whl", hash = "sha256:15c94eeef6b095831067f72c825eb0e2d48bb4cea0647c1b05c981ecba2bf39f"}, + {file = "fonttools-4.51.0.tar.gz", hash = "sha256:dc0673361331566d7a663d7ce0f6fdcbfbdc1f59c6e3ed1165ad7202ca183c68"}, ] [package.extras] @@ -1301,13 +1331,13 @@ files = [ [[package]] name = "fsspec" -version = "2024.2.0" +version = "2024.5.0" description = "File-system specification" optional = false python-versions = ">=3.8" files = [ - {file = "fsspec-2024.2.0-py3-none-any.whl", hash = "sha256:817f969556fa5916bc682e02ca2045f96ff7f586d45110fcb76022063ad2c7d8"}, - {file = "fsspec-2024.2.0.tar.gz", hash = "sha256:b6ad1a679f760dda52b1168c859d01b7b80648ea6f7f7c7f5a8a91dc3f3ecb84"}, + {file = "fsspec-2024.5.0-py3-none-any.whl", hash = "sha256:e0fdbc446d67e182f49a70b82cf7889028a63588fde6b222521f10937b2b670c"}, + {file = "fsspec-2024.5.0.tar.gz", hash = "sha256:1d021b0b0f933e3b3029ed808eb400c08ba101ca2de4b3483fbc9ca23fcee94a"}, ] [package.extras] @@ -1315,7 +1345,7 @@ abfs = ["adlfs"] adl = ["adlfs"] arrow = ["pyarrow (>=1)"] dask = ["dask", "distributed"] -devel = ["pytest", "pytest-cov"] +dev = ["pre-commit", "ruff"] dropbox = ["dropbox", "dropboxdrivefs", "requests"] full = ["adlfs", "aiohttp (!=4.0.0a0,!=4.0.0a1)", "dask", "distributed", "dropbox", "dropboxdrivefs", "fusepy", "gcsfs", "libarchive-c", "ocifs", "panel", "paramiko", "pyarrow (>=1)", "pygit2", "requests", "s3fs", "smbprotocol", "tqdm"] fuse = ["fusepy"] @@ -1332,31 +1362,36 @@ s3 = ["s3fs"] sftp = ["paramiko"] smb = ["smbprotocol"] ssh = ["paramiko"] +test = ["aiohttp (!=4.0.0a0,!=4.0.0a1)", "numpy", "pytest", "pytest-asyncio (!=0.22.0)", "pytest-benchmark", "pytest-cov", "pytest-mock", "pytest-recording", "pytest-rerunfailures", "requests"] +test-downstream = ["aiobotocore (>=2.5.4,<3.0.0)", "dask-expr", "dask[dataframe,test]", "moto[server] (>4,<5)", "pytest-timeout", "xarray"] +test-full = ["adlfs", "aiohttp (!=4.0.0a0,!=4.0.0a1)", "cloudpickle", "dask", "distributed", "dropbox", "dropboxdrivefs", "fastparquet", "fusepy", "gcsfs", "jinja2", "kerchunk", "libarchive-c", "lz4", "notebook", "numpy", "ocifs", "pandas", "panel", "paramiko", "pyarrow", "pyarrow (>=1)", "pyftpdlib", "pygit2", "pytest", "pytest-asyncio (!=0.22.0)", "pytest-benchmark", "pytest-cov", "pytest-mock", "pytest-recording", "pytest-rerunfailures", "python-snappy", "requests", "smbprotocol", "tqdm", "urllib3", "zarr", "zstandard"] tqdm = ["tqdm"] [[package]] name = "future" -version = "0.18.3" +version = "1.0.0" description = "Clean single-source support for Python 3 and 2" optional = false python-versions = ">=2.6, !=3.0., !=3.1., !=3.2." files = [ - {file = "future-0.18.3.tar.gz", hash = "sha256:34a17436ed1e96697a86f9de3d15a3b0be01d8bc8de9c1dffd59fb8234ed5307"}, + {file = "future-1.0.0-py3-none-any.whl", hash = "sha256:929292d34f5872e70396626ef385ec22355a1fae8ad29e1a734c3e43f9fbc216"}, + {file = "future-1.0.0.tar.gz", hash = "sha256:bd2968309307861edae1458a4f8a4f3598c03be43b97521076aebf5d94c07b05"}, ] [[package]] name = "geopandas" -version = "0.14.3" +version = "0.14.4" description = "Geographic pandas extensions" optional = false python-versions = ">=3.9" files = [ - {file = "geopandas-0.14.3-py3-none-any.whl", hash = "sha256:41b31ad39e21bc9e8c4254f78f8dc4ce3d33d144e22e630a00bb336c83160204"}, - {file = "geopandas-0.14.3.tar.gz", hash = "sha256:748af035d4a068a4ae00cab384acb61d387685c833b0022e0729aa45216b23ac"}, + {file = "geopandas-0.14.4-py3-none-any.whl", hash = "sha256:3bb6473cb59d51e1a7fe2dbc24a1a063fb0ebdeddf3ce08ddbf8c7ddc99689aa"}, + {file = "geopandas-0.14.4.tar.gz", hash = "sha256:56765be9d58e2c743078085db3bd07dc6be7719f0dbe1dfdc1d705cb80be7c25"}, ] [package.dependencies] fiona = ">=1.8.21" +numpy = ">=1.22" packaging = "" pandas = ">=1.4.0" pyproj = ">=3.3.0" @@ -1364,17 +1399,17 @@ shapely = ">=1.8.0" [[package]] name = "geoviews" -version = "1.11.0" +version = "1.12.0" description = "GeoViews is a Python library that makes it easy to explore and visualize geographical, meteorological, and oceanographic datasets, such as those used in weather, climate, and remote sensing research." optional = false python-versions = ">=3.9" files = [ - {file = "geoviews-1.11.0-py2.py3-none-any.whl", hash = "sha256:d54a838708b72936fdcbe3eb9e170f977fa7d3733a45a200c9ae322841d4373a"}, - {file = "geoviews-1.11.0.tar.gz", hash = "sha256:2c4d09554019867adfa4ea655617ce8bb9ffa08b534a7e591f961dc6078cffcc"}, + {file = "geoviews-1.12.0-py3-none-any.whl", hash = "sha256:5e8750d0e9a80dd4f5ce493d26cdde52880677c19d329a980b0a94e966dd3bd3"}, + {file = "geoviews-1.12.0.tar.gz", hash = "sha256:e2cbef0605e8fd1529bc643a31aeb61997f8f93c9b41a5aff8b2b355a76fa789"}, ] [package.dependencies] -bokeh = ">=3.2.0,<3.4.0" +bokeh = ">=3.4.0,<3.5.0" cartopy = ">=0.18.0" holoviews = ">=1.16.0" numpy = "" @@ -1386,14 +1421,14 @@ shapely = "" xyzservices = "" [package.extras] -all = ["cartopy (>=0.20.0)", "codecov", "datashader", "fiona", "geodatasets", "geopandas", "graphviz", "iris (>=3.5)", "lxml", "matplotlib (>2.2)", "mock", "nbsite (>=0.8.2,<0.9.0)", "nbsmoke (>=0.2.0)", "netcdf4", "pandas", "pooch", "pyct", "pytest", "pytest-cov", "pytest-github-actions-annotate-failures", "rioxarray", "scipy", "selenium", "shapely", "xarray", "xesmf"] -build = ["bokeh (==3.3)", "param (>=1.9.2)", "pyct (>=0.4.4)", "pyviz-comms (>=0.6.0)"] -doc = ["cartopy (>=0.20.0)", "datashader", "fiona", "geodatasets", "geopandas", "graphviz", "iris (>=3.5)", "lxml", "matplotlib (>2.2)", "mock", "nbsite (>=0.8.2,<0.9.0)", "netcdf4", "pandas", "pooch", "pyct", "scipy", "selenium", "shapely", "xarray", "xesmf"] +all = ["cartopy (>=0.20.0)", "codecov", "datashader", "fiona", "geodatasets", "geopandas", "graphviz", "iris (>=3.5)", "lxml", "matplotlib (>2.2)", "mock", "nbsite (>=0.8.4,<0.9.0)", "nbval", "netcdf4", "pandas", "pooch", "pyct", "pytest", "pytest-cov", "pytest-github-actions-annotate-failures", "pyviz-comms", "rioxarray", "scipy", "selenium", "shapely", "xarray", "xesmf"] +build = ["bokeh (==3.4)", "param (>=1.9.2)", "pyct (>=0.4.4)", "setuptools"] +doc = ["cartopy (>=0.20.0)", "datashader", "fiona", "geodatasets", "geopandas", "graphviz", "iris (>=3.5)", "lxml", "matplotlib (>2.2)", "mock", "nbsite (>=0.8.4,<0.9.0)", "netcdf4", "pandas", "pooch", "pyct", "scipy", "selenium", "shapely", "xarray", "xesmf"] examples-extra = ["datashader", "fiona", "geodatasets", "geopandas", "iris (>=3.5)", "matplotlib (>2.2)", "mock", "netcdf4", "pandas", "pooch", "pyct", "scipy", "shapely", "xarray", "xesmf"] recommended = ["datashader", "geopandas", "matplotlib (>2.2)", "netcdf4", "pandas", "pooch", "pyct", "scipy", "shapely", "xarray"] -tests = ["fiona", "nbsmoke (>=0.2.0)", "pytest", "rioxarray"] +tests = ["fiona", "nbval", "pytest", "rioxarray"] tests-ci = ["codecov", "pytest-cov", "pytest-github-actions-annotate-failures"] -tests-core = ["geopandas", "matplotlib (>2.2)", "netcdf4", "pandas", "pooch", "pytest", "scipy", "shapely", "xarray"] +tests-core = ["geopandas", "matplotlib (>2.2)", "netcdf4", "pandas", "pooch", "pytest", "pyviz-comms", "scipy", "shapely", "xarray"] [[package]] name = "ghp-import" @@ -1414,13 +1449,13 @@ dev = ["flake8", "markdown", "twine", "wheel"] [[package]] name = "griffe" -version = "0.40.0" +version = "0.45.2" description = "Signatures for entire Python programs. Extract the structure, the frame, the skeleton of your project, to generate API documentation or find breaking changes in your API." optional = false python-versions = ">=3.8" files = [ - {file = "griffe-0.40.0-py3-none-any.whl", hash = "sha256:db1da6d1d8e08cbb20f1a7dee8c09da940540c2d4c1bfa26a9091cf6fc36a9ec"}, - {file = "griffe-0.40.0.tar.gz", hash = "sha256:76c4439eaa2737af46ae003c331ab6ca79c5365b552f7b5aed263a3b4125735b"}, + {file = "griffe-0.45.2-py3-none-any.whl", hash = "sha256:297ec8530d0c68e5b98ff86fb588ebc3aa3559bb5dc21f3caea8d9542a350133"}, + {file = "griffe-0.45.2.tar.gz", hash = "sha256:83ce7dcaafd8cb7f43cbf1a455155015a1eb624b1ffd93249e5e1c4a22b2fdb2"}, ] [package.dependencies] @@ -1446,36 +1481,32 @@ test = ["netCDF4", "pytest"] [[package]] name = "h5py" -version = "3.10.0" +version = "3.11.0" description = "Read and write HDF5 files from Python" optional = false python-versions = ">=3.8" files = [ - {file = "h5py-3.10.0-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:b963fb772964fc1d1563c57e4e2e874022ce11f75ddc6df1a626f42bd49ab99f"}, - {file = "h5py-3.10.0-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:012ab448590e3c4f5a8dd0f3533255bc57f80629bf7c5054cf4c87b30085063c"}, - {file = "h5py-3.10.0-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:781a24263c1270a62cd67be59f293e62b76acfcc207afa6384961762bb88ea03"}, - {file = "h5py-3.10.0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:f42e6c30698b520f0295d70157c4e202a9e402406f50dc08f5a7bc416b24e52d"}, - {file = "h5py-3.10.0-cp310-cp310-win_amd64.whl", hash = "sha256:93dd840bd675787fc0b016f7a05fc6efe37312a08849d9dd4053fd0377b1357f"}, - {file = "h5py-3.10.0-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:2381e98af081b6df7f6db300cd88f88e740649d77736e4b53db522d8874bf2dc"}, - {file = "h5py-3.10.0-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:667fe23ab33d5a8a6b77970b229e14ae3bb84e4ea3382cc08567a02e1499eedd"}, - {file = "h5py-3.10.0-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:90286b79abd085e4e65e07c1bd7ee65a0f15818ea107f44b175d2dfe1a4674b7"}, - {file = "h5py-3.10.0-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:6c013d2e79c00f28ffd0cc24e68665ea03ae9069e167087b2adb5727d2736a52"}, - {file = "h5py-3.10.0-cp311-cp311-win_amd64.whl", hash = "sha256:92273ce69ae4983dadb898fd4d3bea5eb90820df953b401282ee69ad648df684"}, - {file = "h5py-3.10.0-cp312-cp312-macosx_10_9_x86_64.whl", hash = "sha256:3c97d03f87f215e7759a354460fb4b0d0f27001450b18b23e556e7856a0b21c3"}, - {file = "h5py-3.10.0-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:86df4c2de68257b8539a18646ceccdcf2c1ce6b1768ada16c8dcfb489eafae20"}, - {file = "h5py-3.10.0-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:ba9ab36be991119a3ff32d0c7cbe5faf9b8d2375b5278b2aea64effbeba66039"}, - {file = "h5py-3.10.0-cp312-cp312-win_amd64.whl", hash = "sha256:2c8e4fda19eb769e9a678592e67eaec3a2f069f7570c82d2da909c077aa94339"}, - {file = "h5py-3.10.0-cp38-cp38-macosx_10_9_x86_64.whl", hash = "sha256:492305a074327e8d2513011fa9fffeb54ecb28a04ca4c4227d7e1e9616d35641"}, - {file = "h5py-3.10.0-cp38-cp38-macosx_11_0_arm64.whl", hash = "sha256:9450464b458cca2c86252b624279115dcaa7260a40d3cb1594bf2b410a2bd1a3"}, - {file = "h5py-3.10.0-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:fd6f6d1384a9f491732cee233b99cd4bfd6e838a8815cc86722f9d2ee64032af"}, - {file = "h5py-3.10.0-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:3074ec45d3dc6e178c6f96834cf8108bf4a60ccb5ab044e16909580352010a97"}, - {file = "h5py-3.10.0-cp38-cp38-win_amd64.whl", hash = "sha256:212bb997a91e6a895ce5e2f365ba764debeaef5d2dca5c6fb7098d66607adf99"}, - {file = "h5py-3.10.0-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:5dfc65ac21fa2f630323c92453cadbe8d4f504726ec42f6a56cf80c2f90d6c52"}, - {file = "h5py-3.10.0-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:d4682b94fd36ab217352be438abd44c8f357c5449b8995e63886b431d260f3d3"}, - {file = "h5py-3.10.0-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:aece0e2e1ed2aab076c41802e50a0c3e5ef8816d60ece39107d68717d4559824"}, - {file = "h5py-3.10.0-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:43a61b2c2ad65b1fabc28802d133eed34debcc2c8b420cb213d3d4ef4d3e2229"}, - {file = "h5py-3.10.0-cp39-cp39-win_amd64.whl", hash = "sha256:ae2f0201c950059676455daf92700eeb57dcf5caaf71b9e1328e6e6593601770"}, - {file = "h5py-3.10.0.tar.gz", hash = "sha256:d93adc48ceeb33347eb24a634fb787efc7ae4644e6ea4ba733d099605045c049"}, + {file = "h5py-3.11.0-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:1625fd24ad6cfc9c1ccd44a66dac2396e7ee74940776792772819fc69f3a3731"}, + {file = "h5py-3.11.0-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:c072655ad1d5fe9ef462445d3e77a8166cbfa5e599045f8aa3c19b75315f10e5"}, + {file = "h5py-3.11.0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:77b19a40788e3e362b54af4dcf9e6fde59ca016db2c61360aa30b47c7b7cef00"}, + {file = "h5py-3.11.0-cp310-cp310-win_amd64.whl", hash = "sha256:ef4e2f338fc763f50a8113890f455e1a70acd42a4d083370ceb80c463d803972"}, + {file = "h5py-3.11.0-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:bbd732a08187a9e2a6ecf9e8af713f1d68256ee0f7c8b652a32795670fb481ba"}, + {file = "h5py-3.11.0-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:75bd7b3d93fbeee40860fd70cdc88df4464e06b70a5ad9ce1446f5f32eb84007"}, + {file = "h5py-3.11.0-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:52c416f8eb0daae39dabe71415cb531f95dce2d81e1f61a74537a50c63b28ab3"}, + {file = "h5py-3.11.0-cp311-cp311-win_amd64.whl", hash = "sha256:083e0329ae534a264940d6513f47f5ada617da536d8dccbafc3026aefc33c90e"}, + {file = "h5py-3.11.0-cp312-cp312-macosx_10_9_x86_64.whl", hash = "sha256:a76cae64080210389a571c7d13c94a1a6cf8cb75153044fd1f822a962c97aeab"}, + {file = "h5py-3.11.0-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:f3736fe21da2b7d8a13fe8fe415f1272d2a1ccdeff4849c1421d2fb30fd533bc"}, + {file = "h5py-3.11.0-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:aa6ae84a14103e8dc19266ef4c3e5d7c00b68f21d07f2966f0ca7bdb6c2761fb"}, + {file = "h5py-3.11.0-cp312-cp312-win_amd64.whl", hash = "sha256:21dbdc5343f53b2e25404673c4f00a3335aef25521bd5fa8c707ec3833934892"}, + {file = "h5py-3.11.0-cp38-cp38-macosx_10_9_x86_64.whl", hash = "sha256:754c0c2e373d13d6309f408325343b642eb0f40f1a6ad21779cfa9502209e150"}, + {file = "h5py-3.11.0-cp38-cp38-macosx_11_0_arm64.whl", hash = "sha256:731839240c59ba219d4cb3bc5880d438248533366f102402cfa0621b71796b62"}, + {file = "h5py-3.11.0-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:8ec9df3dd2018904c4cc06331951e274f3f3fd091e6d6cc350aaa90fa9b42a76"}, + {file = "h5py-3.11.0-cp38-cp38-win_amd64.whl", hash = "sha256:55106b04e2c83dfb73dc8732e9abad69d83a436b5b82b773481d95d17b9685e1"}, + {file = "h5py-3.11.0-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:f4e025e852754ca833401777c25888acb96889ee2c27e7e629a19aee288833f0"}, + {file = "h5py-3.11.0-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:6c4b760082626120031d7902cd983d8c1f424cdba2809f1067511ef283629d4b"}, + {file = "h5py-3.11.0-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:67462d0669f8f5459529de179f7771bd697389fcb3faab54d63bf788599a48ea"}, + {file = "h5py-3.11.0-cp39-cp39-win_amd64.whl", hash = "sha256:d9c944d364688f827dc889cf83f1fca311caf4fa50b19f009d1f2b525edd33a3"}, + {file = "h5py-3.11.0.tar.gz", hash = "sha256:7b7e8f78072a2edec87c9836f25f34203fd492a4475709a18b417a33cfb21fa9"}, ] [package.dependencies] @@ -1483,13 +1514,13 @@ numpy = ">=1.17.3" [[package]] name = "holoviews" -version = "1.18.1" +version = "1.18.3" description = "Stop plotting your data - annotate your data and let it visualize itself." optional = false python-versions = ">=3.9" files = [ - {file = "holoviews-1.18.1-py2.py3-none-any.whl", hash = "sha256:b1bf6cd9aa72a2e103e8a7e0e57c8c8171a3fc3ebd5b943fe49bf98b6c64367b"}, - {file = "holoviews-1.18.1.tar.gz", hash = "sha256:805c7353ae52e97753e7f0668b051a086410e2d2c13c8fe294ebb4cdd02b2324"}, + {file = "holoviews-1.18.3-py2.py3-none-any.whl", hash = "sha256:b94b96560b64a84c07e89115aaf9b226e6009684800ec84d3c88cbad122c0c46"}, + {file = "holoviews-1.18.3.tar.gz", hash = "sha256:578e30e89d72754f97a83ebe08198fec8e87cc7e49b25b9f31ec393f939ca500"}, ] [package.dependencies] @@ -1502,61 +1533,61 @@ param = ">=1.12.0,<3.0" pyviz-comms = ">=0.7.4" [package.extras] -all = ["bokeh (>=3.1)", "cftime", "codecov", "contourpy", "cudf", "dash (>=1.16)", "dask", "datashader (>=0.11.1)", "ffmpeg", "flaky", "graphviz", "ibis-framework", "ipython (>=5.4.0)", "matplotlib (>=3)", "nbconvert", "nbsite (>=0.8.2,<0.9.0)", "nbval", "netcdf4", "networkx", "notebook", "pillow", "playwright", "plotly (>=4.0)", "pooch", "pre-commit", "pyarrow", "pytest", "pytest-cov", "pytest-github-actions-annotate-failures", "pytest-playwright", "pytest-xdist", "ruff", "scikit-image", "scipy", "selenium", "shapely", "spatialpandas", "streamz (>=0.5.0)", "xarray (>=0.10.4)"] +all = ["bokeh (>=3.1)", "cftime", "codecov", "contourpy", "cudf", "dash (>=1.16)", "dask", "datashader (>=0.11.1)", "ffmpeg", "graphviz", "ibis-framework", "ipython (>=5.4.0)", "matplotlib (>=3)", "myst-nb (<1)", "nbconvert", "nbsite (>=0.8.4,<0.9.0)", "nbval", "netcdf4", "networkx", "notebook", "notebook (>=7.0)", "pillow", "playwright", "plotly (>=4.0)", "pooch", "pre-commit", "pyarrow", "pytest", "pytest-cov", "pytest-github-actions-annotate-failures", "pytest-playwright", "pytest-rerunfailures", "pytest-xdist", "ruff", "scikit-image", "scipy", "scipy (>=1.10)", "selenium", "shapely", "spatialpandas", "streamz (>=0.5.0)", "xarray (>=0.10.4)"] build = ["param (>=1.7.0)", "pyct (>=0.4.4)", "setuptools (>=30.3.0)"] -doc = ["bokeh (>=3.1)", "cftime", "dash (>=1.16)", "dask", "datashader (>=0.11.1)", "ffmpeg", "graphviz", "ipython (>=5.4.0)", "matplotlib (>=3)", "nbsite (>=0.8.2,<0.9.0)", "netcdf4", "networkx", "notebook", "pillow", "plotly (>=4.0)", "pooch", "pyarrow", "scikit-image", "scipy", "selenium", "shapely", "streamz (>=0.5.0)", "xarray (>=0.10.4)"] -examples = ["bokeh (>=3.1)", "cftime", "dash (>=1.16)", "dask", "datashader (>=0.11.1)", "ffmpeg", "ipython (>=5.4.0)", "matplotlib (>=3)", "netcdf4", "networkx", "notebook", "pillow", "plotly (>=4.0)", "pooch", "pyarrow", "scikit-image", "scipy", "shapely", "streamz (>=0.5.0)", "xarray (>=0.10.4)"] -examples-tests = ["bokeh (>=3.1)", "cftime", "dash (>=1.16)", "dask", "datashader (>=0.11.1)", "ffmpeg", "ipython (>=5.4.0)", "matplotlib (>=3)", "nbval", "netcdf4", "networkx", "notebook", "pillow", "plotly (>=4.0)", "pooch", "pyarrow", "scikit-image", "scipy", "shapely", "streamz (>=0.5.0)", "xarray (>=0.10.4)"] +doc = ["bokeh (>=3.1)", "cftime", "dash (>=1.16)", "dask", "datashader (>=0.11.1)", "ffmpeg", "graphviz", "ipython (>=5.4.0)", "matplotlib (>=3)", "myst-nb (<1)", "nbsite (>=0.8.4,<0.9.0)", "netcdf4", "networkx", "notebook", "notebook (>=7.0)", "pillow", "plotly (>=4.0)", "pooch", "pyarrow", "scikit-image", "scipy", "selenium", "shapely", "streamz (>=0.5.0)", "xarray (>=0.10.4)"] +examples = ["bokeh (>=3.1)", "cftime", "dash (>=1.16)", "dask", "datashader (>=0.11.1)", "ffmpeg", "ipython (>=5.4.0)", "matplotlib (>=3)", "netcdf4", "networkx", "notebook", "notebook (>=7.0)", "pillow", "plotly (>=4.0)", "pooch", "pyarrow", "scikit-image", "scipy", "shapely", "streamz (>=0.5.0)", "xarray (>=0.10.4)"] +examples-tests = ["bokeh (>=3.1)", "cftime", "dash (>=1.16)", "dask", "datashader (>=0.11.1)", "ffmpeg", "ipython (>=5.4.0)", "matplotlib (>=3)", "nbval", "netcdf4", "networkx", "notebook", "notebook (>=7.0)", "pillow", "plotly (>=4.0)", "pooch", "pyarrow", "scikit-image", "scipy", "shapely", "streamz (>=0.5.0)", "xarray (>=0.10.4)"] lint = ["pre-commit", "ruff"] notebook = ["ipython (>=5.4.0)", "notebook"] recommended = ["bokeh (>=3.1)", "ipython (>=5.4.0)", "matplotlib (>=3)", "notebook"] -tests = ["bokeh (>=3.1)", "cftime", "contourpy", "dash (>=1.16)", "dask", "datashader (>=0.11.1)", "ffmpeg", "flaky", "ibis-framework", "ipython (>=5.4.0)", "matplotlib (>=3)", "nbconvert", "networkx", "pillow", "plotly (>=4.0)", "pytest", "pytest-cov", "pytest-xdist", "scipy", "selenium", "shapely", "spatialpandas", "xarray (>=0.10.4)"] +tests = ["bokeh (>=3.1)", "cftime", "contourpy", "dash (>=1.16)", "dask", "datashader (>=0.11.1)", "ffmpeg", "ibis-framework", "ipython (>=5.4.0)", "matplotlib (>=3)", "nbconvert", "networkx", "pillow", "plotly (>=4.0)", "pytest", "pytest-cov", "pytest-rerunfailures", "pytest-xdist", "scipy (>=1.10)", "selenium", "shapely", "spatialpandas", "xarray (>=0.10.4)"] tests-ci = ["codecov", "pytest-github-actions-annotate-failures"] -tests-core = ["bokeh (>=3.1)", "contourpy", "flaky", "ipython (>=5.4.0)", "matplotlib (>=3)", "nbconvert", "pillow", "plotly (>=4.0)", "pytest", "pytest-cov", "pytest-xdist"] -tests-gpu = ["bokeh (>=3.1)", "cftime", "contourpy", "cudf", "dash (>=1.16)", "dask", "datashader (>=0.11.1)", "ffmpeg", "flaky", "ibis-framework", "ipython (>=5.4.0)", "matplotlib (>=3)", "nbconvert", "networkx", "pillow", "plotly (>=4.0)", "pytest", "pytest-cov", "pytest-xdist", "scipy", "selenium", "shapely", "spatialpandas", "xarray (>=0.10.4)"] +tests-core = ["bokeh (>=3.1)", "contourpy", "ipython (>=5.4.0)", "matplotlib (>=3)", "nbconvert", "pillow", "plotly (>=4.0)", "pytest", "pytest-cov", "pytest-rerunfailures", "pytest-xdist"] +tests-gpu = ["bokeh (>=3.1)", "cftime", "contourpy", "cudf", "dash (>=1.16)", "dask", "datashader (>=0.11.1)", "ffmpeg", "ibis-framework", "ipython (>=5.4.0)", "matplotlib (>=3)", "nbconvert", "networkx", "pillow", "plotly (>=4.0)", "pytest", "pytest-cov", "pytest-rerunfailures", "pytest-xdist", "scipy (>=1.10)", "selenium", "shapely", "spatialpandas", "xarray (>=0.10.4)"] tests-nb = ["nbval"] ui = ["playwright", "pytest-playwright"] -unit-tests = ["bokeh (>=3.1)", "cftime", "contourpy", "dash (>=1.16)", "dask", "datashader (>=0.11.1)", "ffmpeg", "flaky", "ibis-framework", "ipython (>=5.4.0)", "matplotlib (>=3)", "nbconvert", "netcdf4", "networkx", "notebook", "pillow", "plotly (>=4.0)", "pooch", "pre-commit", "pyarrow", "pytest", "pytest-cov", "pytest-xdist", "ruff", "scikit-image", "scipy", "selenium", "shapely", "spatialpandas", "streamz (>=0.5.0)", "xarray (>=0.10.4)"] +unit-tests = ["bokeh (>=3.1)", "cftime", "contourpy", "dash (>=1.16)", "dask", "datashader (>=0.11.1)", "ffmpeg", "ibis-framework", "ipython (>=5.4.0)", "matplotlib (>=3)", "nbconvert", "netcdf4", "networkx", "notebook", "notebook (>=7.0)", "pillow", "plotly (>=4.0)", "pooch", "pre-commit", "pyarrow", "pytest", "pytest-cov", "pytest-rerunfailures", "pytest-xdist", "ruff", "scikit-image", "scipy", "scipy (>=1.10)", "selenium", "shapely", "spatialpandas", "streamz (>=0.5.0)", "xarray (>=0.10.4)"] [[package]] name = "idna" -version = "3.6" +version = "3.7" description = "Internationalized Domain Names in Applications (IDNA)" optional = false python-versions = ">=3.5" files = [ - {file = "idna-3.6-py3-none-any.whl", hash = "sha256:c05567e9c24a6b9faaa835c4821bad0590fbb9d5779e7caa6e1cc4978e7eb24f"}, - {file = "idna-3.6.tar.gz", hash = "sha256:9ecdbbd083b06798ae1e86adcbfe8ab1479cf864e4ee30fe4e46a003d12491ca"}, + {file = "idna-3.7-py3-none-any.whl", hash = "sha256:82fee1fc78add43492d3a1898bfa6d8a904cc97d8427f683ed8e798d07761aa0"}, + {file = "idna-3.7.tar.gz", hash = "sha256:028ff3aadf0609c1fd278d8ea3089299412a7a8b9bd005dd08b9f8285bcb5cfc"}, ] [[package]] name = "importlib-metadata" -version = "7.0.1" +version = "7.1.0" description = "Read metadata from Python packages" optional = false python-versions = ">=3.8" files = [ - {file = "importlib_metadata-7.0.1-py3-none-any.whl", hash = "sha256:4805911c3a4ec7c3966410053e9ec6a1fecd629117df5adee56dfc9432a1081e"}, - {file = "importlib_metadata-7.0.1.tar.gz", hash = "sha256:f238736bb06590ae52ac1fab06a3a9ef1d8dce2b7a35b5ab329371d6c8f5d2cc"}, + {file = "importlib_metadata-7.1.0-py3-none-any.whl", hash = "sha256:30962b96c0c223483ed6cc7280e7f0199feb01a0e40cfae4d4450fc6fab1f570"}, + {file = "importlib_metadata-7.1.0.tar.gz", hash = "sha256:b78938b926ee8d5f020fc4772d487045805a55ddbad2ecf21c6d60938dc7fcd2"}, ] [package.dependencies] zipp = ">=0.5" [package.extras] -docs = ["furo", "jaraco.packaging (>=9.3)", "jaraco.tidelift (>=1.4)", "rst.linker (>=1.9)", "sphinx (<7.2.5)", "sphinx (>=3.5)", "sphinx-lint"] +docs = ["furo", "jaraco.packaging (>=9.3)", "jaraco.tidelift (>=1.4)", "rst.linker (>=1.9)", "sphinx (>=3.5)", "sphinx-lint"] perf = ["ipython"] -testing = ["flufl.flake8", "importlib-resources (>=1.3)", "packaging", "pyfakefs", "pytest (>=6)", "pytest-black (>=0.3.7)", "pytest-checkdocs (>=2.4)", "pytest-cov", "pytest-enabler (>=2.2)", "pytest-mypy (>=0.9.1)", "pytest-perf (>=0.9.2)", "pytest-ruff"] +testing = ["flufl.flake8", "importlib-resources (>=1.3)", "jaraco.test (>=5.4)", "packaging", "pyfakefs", "pytest (>=6)", "pytest-checkdocs (>=2.4)", "pytest-cov", "pytest-enabler (>=2.2)", "pytest-mypy", "pytest-perf (>=0.9.2)", "pytest-ruff (>=0.2.1)"] [[package]] name = "importlib-resources" -version = "6.1.1" +version = "6.4.0" description = "Read resources from Python packages" optional = false python-versions = ">=3.8" files = [ - {file = "importlib_resources-6.1.1-py3-none-any.whl", hash = "sha256:e8bf90d8213b486f428c9c39714b920041cb02c184686a3dee24905aaa8105d6"}, - {file = "importlib_resources-6.1.1.tar.gz", hash = "sha256:3893a00122eafde6894c59914446a512f728a0c1a45f9bb9b63721b6bacf0b4a"}, + {file = "importlib_resources-6.4.0-py3-none-any.whl", hash = "sha256:50d10f043df931902d4194ea07ec57960f66a80449ff867bfe782b4c486ba78c"}, + {file = "importlib_resources-6.4.0.tar.gz", hash = "sha256:cdb2b453b8046ca4e3798eb1d84f3cce1446a0e8e7b5ef4efb600f19fc398145"}, ] [package.dependencies] @@ -1564,7 +1595,7 @@ zipp = {version = ">=3.1.0", markers = "python_version < \"3.10\""} [package.extras] docs = ["furo", "jaraco.packaging (>=9.3)", "jaraco.tidelift (>=1.4)", "rst.linker (>=1.9)", "sphinx (<7.2.5)", "sphinx (>=3.5)", "sphinx-lint"] -testing = ["pytest (>=6)", "pytest-black (>=0.3.7)", "pytest-checkdocs (>=2.4)", "pytest-cov", "pytest-enabler (>=2.2)", "pytest-mypy (>=0.9.1)", "pytest-ruff", "zipp (>=3.17)"] +testing = ["jaraco.test (>=5.4)", "pytest (>=6)", "pytest-checkdocs (>=2.4)", "pytest-cov", "pytest-enabler (>=2.2)", "pytest-mypy", "pytest-ruff (>=0.2.1)", "zipp (>=3.17)"] [[package]] name = "iniconfig" @@ -1579,13 +1610,13 @@ files = [ [[package]] name = "ipykernel" -version = "6.29.1" +version = "6.29.4" description = "IPython Kernel for Jupyter" optional = false python-versions = ">=3.8" files = [ - {file = "ipykernel-6.29.1-py3-none-any.whl", hash = "sha256:e5dfba210fc9da74a5dae8fa6c41f816e11bd18d10381b2517d9a0d57cc987c4"}, - {file = "ipykernel-6.29.1.tar.gz", hash = "sha256:1547352b32da95a2761011a8dac2af930c26a0703dfa07690d16b7d74dac0ba1"}, + {file = "ipykernel-6.29.4-py3-none-any.whl", hash = "sha256:1181e653d95c6808039c509ef8e67c4126b3b3af7781496c7cbfb5ed938a27da"}, + {file = "ipykernel-6.29.4.tar.gz", hash = "sha256:3d44070060f9475ac2092b760123fadf105d2e2493c24848b6691a7c4f42af5c"}, ] [package.dependencies] @@ -1608,7 +1639,7 @@ cov = ["coverage[toml]", "curio", "matplotlib", "pytest-cov", "trio"] docs = ["myst-parser", "pydata-sphinx-theme", "sphinx", "sphinx-autodoc-typehints", "sphinxcontrib-github-alt", "sphinxcontrib-spelling", "trio"] pyqt5 = ["pyqt5"] pyside6 = ["pyside6"] -test = ["flaky", "ipyparallel", "pre-commit", "pytest (>=7.0)", "pytest-asyncio (==0.23.4)", "pytest-cov", "pytest-timeout"] +test = ["flaky", "ipyparallel", "pre-commit", "pytest (>=7.0)", "pytest-asyncio (>=0.23.5)", "pytest-cov", "pytest-timeout"] [[package]] name = "ipython" @@ -1668,13 +1699,13 @@ testing = ["Django", "attrs", "colorama", "docopt", "pytest (<7.0.0)"] [[package]] name = "jinja2" -version = "3.1.3" +version = "3.1.4" description = "A very fast and expressive template engine." optional = false python-versions = ">=3.7" files = [ - {file = "Jinja2-3.1.3-py3-none-any.whl", hash = "sha256:7d6d50dd97d52cbc355597bd845fabfbac3f551e1f99619e39a35ce8c370b5fa"}, - {file = "Jinja2-3.1.3.tar.gz", hash = "sha256:ac8bd6544d4bb2c9792bf3a159e80bba8fda7f07e81bc3aed565432d5925ba90"}, + {file = "jinja2-3.1.4-py3-none-any.whl", hash = "sha256:bc5dd2abb727a5319567b7a813e6a2e7318c39f4f487cfe6c89c6f9c7d25197d"}, + {file = "jinja2-3.1.4.tar.gz", hash = "sha256:4a3aee7acbbe7303aede8e9648d13b8bf88a429282aa6122a993f0ac800cb369"}, ] [package.dependencies] @@ -1685,13 +1716,13 @@ i18n = ["Babel (>=2.7)"] [[package]] name = "jsonschema" -version = "4.21.1" +version = "4.22.0" description = "An implementation of JSON Schema validation for Python" optional = false python-versions = ">=3.8" files = [ - {file = "jsonschema-4.21.1-py3-none-any.whl", hash = "sha256:7996507afae316306f9e2290407761157c6f78002dcf7419acb99822143d1c6f"}, - {file = "jsonschema-4.21.1.tar.gz", hash = "sha256:85727c00279f5fa6bedbe6238d2aa6403bedd8b4864ab11207d07df3cc1b2ee5"}, + {file = "jsonschema-4.22.0-py3-none-any.whl", hash = "sha256:ff4cfd6b1367a40e7bc6411caec72effadd3db0bbe5017de188f2d6108335802"}, + {file = "jsonschema-4.22.0.tar.gz", hash = "sha256:5b22d434a45935119af990552c862e5d6d564e8f6601206b305a61fdf661a2b7"}, ] [package.dependencies] @@ -1720,13 +1751,13 @@ referencing = ">=0.31.0" [[package]] name = "jupyter-client" -version = "8.6.0" +version = "8.6.2" description = "Jupyter protocol implementation and client libraries" optional = false python-versions = ">=3.8" files = [ - {file = "jupyter_client-8.6.0-py3-none-any.whl", hash = "sha256:909c474dbe62582ae62b758bca86d6518c85234bdee2d908c778db6d72f39d99"}, - {file = "jupyter_client-8.6.0.tar.gz", hash = "sha256:0642244bb83b4764ae60d07e010e15f0e2d275ec4e918a8f7b80fbbef3ca60c7"}, + {file = "jupyter_client-8.6.2-py3-none-any.whl", hash = "sha256:50cbc5c66fd1b8f65ecb66bc490ab73217993632809b6e505687de18e9dea39f"}, + {file = "jupyter_client-8.6.2.tar.gz", hash = "sha256:2bda14d55ee5ba58552a8c53ae43d215ad9868853489213f37da060ced54d8df"}, ] [package.dependencies] @@ -1739,17 +1770,17 @@ traitlets = ">=5.3" [package.extras] docs = ["ipykernel", "myst-parser", "pydata-sphinx-theme", "sphinx (>=4)", "sphinx-autodoc-typehints", "sphinxcontrib-github-alt", "sphinxcontrib-spelling"] -test = ["coverage", "ipykernel (>=6.14)", "mypy", "paramiko", "pre-commit", "pytest", "pytest-cov", "pytest-jupyter[client] (>=0.4.1)", "pytest-timeout"] +test = ["coverage", "ipykernel (>=6.14)", "mypy", "paramiko", "pre-commit", "pytest (<8.2.0)", "pytest-cov", "pytest-jupyter[client] (>=0.4.1)", "pytest-timeout"] [[package]] name = "jupyter-core" -version = "5.7.1" +version = "5.7.2" description = "Jupyter core package. A base package on which Jupyter projects rely." optional = false python-versions = ">=3.8" files = [ - {file = "jupyter_core-5.7.1-py3-none-any.whl", hash = "sha256:c65c82126453a723a2804aa52409930434598fd9d35091d63dfb919d2b765bb7"}, - {file = "jupyter_core-5.7.1.tar.gz", hash = "sha256:de61a9d7fc71240f688b2fb5ab659fbb56979458dc66a71decd098e03c79e218"}, + {file = "jupyter_core-5.7.2-py3-none-any.whl", hash = "sha256:4f7315d2f6b4bcf2e3e7cb6e46772eba760ae459cd1f59d29eb57b0a01bd7409"}, + {file = "jupyter_core-5.7.2.tar.gz", hash = "sha256:aa5f8d32bbf6b431ac830496da7392035d6f61b4f54872f15c4bd2a9c3f536d9"}, ] [package.dependencies] @@ -1759,7 +1790,7 @@ traitlets = ">=5.3" [package.extras] docs = ["myst-parser", "pydata-sphinx-theme", "sphinx-autodoc-typehints", "sphinxcontrib-github-alt", "sphinxcontrib-spelling", "traitlets"] -test = ["ipykernel", "pre-commit", "pytest", "pytest-cov", "pytest-timeout"] +test = ["ipykernel", "pre-commit", "pytest (<8)", "pytest-cov", "pytest-timeout"] [[package]] name = "kiwisolver" @@ -1987,13 +2018,13 @@ tests = ["psutil", "pytest (!=3.3.0)", "pytest-cov"] [[package]] name = "markdown" -version = "3.5.2" +version = "3.6" description = "Python implementation of John Gruber's Markdown." optional = false python-versions = ">=3.8" files = [ - {file = "Markdown-3.5.2-py3-none-any.whl", hash = "sha256:d43323865d89fc0cb9b20c75fc8ad313af307cc087e84b657d9eec768eddeadd"}, - {file = "Markdown-3.5.2.tar.gz", hash = "sha256:e1ac7b3dc550ee80e602e71c1d168002f062e49f1b11e26a36264dafd4df2ef8"}, + {file = "Markdown-3.6-py3-none-any.whl", hash = "sha256:48f276f4d8cfb8ce6527c8f79e2ee29708508bf4d40aa410fbc3b4ee832c850f"}, + {file = "Markdown-3.6.tar.gz", hash = "sha256:ed4f41f6daecbeeb96e576ce414c41d2d876daa9a16cb35fa8ed8c2ddfad0224"}, ] [package.dependencies] @@ -2098,39 +2129,40 @@ files = [ [[package]] name = "matplotlib" -version = "3.8.2" +version = "3.9.0" description = "Python plotting package" optional = false python-versions = ">=3.9" files = [ - {file = "matplotlib-3.8.2-cp310-cp310-macosx_10_12_x86_64.whl", hash = "sha256:09796f89fb71a0c0e1e2f4bdaf63fb2cefc84446bb963ecdeb40dfee7dfa98c7"}, - {file = "matplotlib-3.8.2-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:6f9c6976748a25e8b9be51ea028df49b8e561eed7809146da7a47dbecebab367"}, - {file = "matplotlib-3.8.2-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:b78e4f2cedf303869b782071b55fdde5987fda3038e9d09e58c91cc261b5ad18"}, - {file = "matplotlib-3.8.2-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:4e208f46cf6576a7624195aa047cb344a7f802e113bb1a06cfd4bee431de5e31"}, - {file = "matplotlib-3.8.2-cp310-cp310-musllinux_1_1_x86_64.whl", hash = "sha256:46a569130ff53798ea5f50afce7406e91fdc471ca1e0e26ba976a8c734c9427a"}, - {file = "matplotlib-3.8.2-cp310-cp310-win_amd64.whl", hash = "sha256:830f00640c965c5b7f6bc32f0d4ce0c36dfe0379f7dd65b07a00c801713ec40a"}, - {file = "matplotlib-3.8.2-cp311-cp311-macosx_10_12_x86_64.whl", hash = "sha256:d86593ccf546223eb75a39b44c32788e6f6440d13cfc4750c1c15d0fcb850b63"}, - {file = "matplotlib-3.8.2-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:9a5430836811b7652991939012f43d2808a2db9b64ee240387e8c43e2e5578c8"}, - {file = "matplotlib-3.8.2-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:b9576723858a78751d5aacd2497b8aef29ffea6d1c95981505877f7ac28215c6"}, - {file = "matplotlib-3.8.2-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:5ba9cbd8ac6cf422f3102622b20f8552d601bf8837e49a3afed188d560152788"}, - {file = "matplotlib-3.8.2-cp311-cp311-musllinux_1_1_x86_64.whl", hash = "sha256:03f9d160a29e0b65c0790bb07f4f45d6a181b1ac33eb1bb0dd225986450148f0"}, - {file = "matplotlib-3.8.2-cp311-cp311-win_amd64.whl", hash = "sha256:3773002da767f0a9323ba1a9b9b5d00d6257dbd2a93107233167cfb581f64717"}, - {file = "matplotlib-3.8.2-cp312-cp312-macosx_10_12_x86_64.whl", hash = "sha256:4c318c1e95e2f5926fba326f68177dee364aa791d6df022ceb91b8221bd0a627"}, - {file = "matplotlib-3.8.2-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:091275d18d942cf1ee9609c830a1bc36610607d8223b1b981c37d5c9fc3e46a4"}, - {file = "matplotlib-3.8.2-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:1b0f3b8ea0e99e233a4bcc44590f01604840d833c280ebb8fe5554fd3e6cfe8d"}, - {file = "matplotlib-3.8.2-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:d7b1704a530395aaf73912be741c04d181f82ca78084fbd80bc737be04848331"}, - {file = "matplotlib-3.8.2-cp312-cp312-musllinux_1_1_x86_64.whl", hash = "sha256:533b0e3b0c6768eef8cbe4b583731ce25a91ab54a22f830db2b031e83cca9213"}, - {file = "matplotlib-3.8.2-cp312-cp312-win_amd64.whl", hash = "sha256:0f4fc5d72b75e2c18e55eb32292659cf731d9d5b312a6eb036506304f4675630"}, - {file = "matplotlib-3.8.2-cp39-cp39-macosx_10_12_x86_64.whl", hash = "sha256:deaed9ad4da0b1aea77fe0aa0cebb9ef611c70b3177be936a95e5d01fa05094f"}, - {file = "matplotlib-3.8.2-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:172f4d0fbac3383d39164c6caafd3255ce6fa58f08fc392513a0b1d3b89c4f89"}, - {file = "matplotlib-3.8.2-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:c7d36c2209d9136cd8e02fab1c0ddc185ce79bc914c45054a9f514e44c787917"}, - {file = "matplotlib-3.8.2-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:5864bdd7da445e4e5e011b199bb67168cdad10b501750367c496420f2ad00843"}, - {file = "matplotlib-3.8.2-cp39-cp39-musllinux_1_1_x86_64.whl", hash = "sha256:ef8345b48e95cee45ff25192ed1f4857273117917a4dcd48e3905619bcd9c9b8"}, - {file = "matplotlib-3.8.2-cp39-cp39-win_amd64.whl", hash = "sha256:7c48d9e221b637c017232e3760ed30b4e8d5dfd081daf327e829bf2a72c731b4"}, - {file = "matplotlib-3.8.2-pp39-pypy39_pp73-macosx_10_12_x86_64.whl", hash = "sha256:aa11b3c6928a1e496c1a79917d51d4cd5d04f8a2e75f21df4949eeefdf697f4b"}, - {file = "matplotlib-3.8.2-pp39-pypy39_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:d1095fecf99eeb7384dabad4bf44b965f929a5f6079654b681193edf7169ec20"}, - {file = "matplotlib-3.8.2-pp39-pypy39_pp73-win_amd64.whl", hash = "sha256:bddfb1db89bfaa855912261c805bd0e10218923cc262b9159a49c29a7a1c1afa"}, - {file = "matplotlib-3.8.2.tar.gz", hash = "sha256:01a978b871b881ee76017152f1f1a0cbf6bd5f7b8ff8c96df0df1bd57d8755a1"}, + {file = "matplotlib-3.9.0-cp310-cp310-macosx_10_12_x86_64.whl", hash = "sha256:2bcee1dffaf60fe7656183ac2190bd630842ff87b3153afb3e384d966b57fe56"}, + {file = "matplotlib-3.9.0-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:3f988bafb0fa39d1074ddd5bacd958c853e11def40800c5824556eb630f94d3b"}, + {file = "matplotlib-3.9.0-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:fe428e191ea016bb278758c8ee82a8129c51d81d8c4bc0846c09e7e8e9057241"}, + {file = "matplotlib-3.9.0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:eaf3978060a106fab40c328778b148f590e27f6fa3cd15a19d6892575bce387d"}, + {file = "matplotlib-3.9.0-cp310-cp310-musllinux_1_1_x86_64.whl", hash = "sha256:2e7f03e5cbbfacdd48c8ea394d365d91ee8f3cae7e6ec611409927b5ed997ee4"}, + {file = "matplotlib-3.9.0-cp310-cp310-win_amd64.whl", hash = "sha256:13beb4840317d45ffd4183a778685e215939be7b08616f431c7795276e067463"}, + {file = "matplotlib-3.9.0-cp311-cp311-macosx_10_12_x86_64.whl", hash = "sha256:063af8587fceeac13b0936c42a2b6c732c2ab1c98d38abc3337e430e1ff75e38"}, + {file = "matplotlib-3.9.0-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:9a2fa6d899e17ddca6d6526cf6e7ba677738bf2a6a9590d702c277204a7c6152"}, + {file = "matplotlib-3.9.0-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:550cdda3adbd596078cca7d13ed50b77879104e2e46392dcd7c75259d8f00e85"}, + {file = "matplotlib-3.9.0-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:76cce0f31b351e3551d1f3779420cf8f6ec0d4a8cf9c0237a3b549fd28eb4abb"}, + {file = "matplotlib-3.9.0-cp311-cp311-musllinux_1_1_x86_64.whl", hash = "sha256:c53aeb514ccbbcbab55a27f912d79ea30ab21ee0531ee2c09f13800efb272674"}, + {file = "matplotlib-3.9.0-cp311-cp311-win_amd64.whl", hash = "sha256:a5be985db2596d761cdf0c2eaf52396f26e6a64ab46bd8cd810c48972349d1be"}, + {file = "matplotlib-3.9.0-cp312-cp312-macosx_10_12_x86_64.whl", hash = "sha256:c79f3a585f1368da6049318bdf1f85568d8d04b2e89fc24b7e02cc9b62017382"}, + {file = "matplotlib-3.9.0-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:bdd1ecbe268eb3e7653e04f451635f0fb0f77f07fd070242b44c076c9106da84"}, + {file = "matplotlib-3.9.0-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:d38e85a1a6d732f645f1403ce5e6727fd9418cd4574521d5803d3d94911038e5"}, + {file = "matplotlib-3.9.0-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:0a490715b3b9984fa609116481b22178348c1a220a4499cda79132000a79b4db"}, + {file = "matplotlib-3.9.0-cp312-cp312-musllinux_1_1_x86_64.whl", hash = "sha256:8146ce83cbc5dc71c223a74a1996d446cd35cfb6a04b683e1446b7e6c73603b7"}, + {file = "matplotlib-3.9.0-cp312-cp312-win_amd64.whl", hash = "sha256:d91a4ffc587bacf5c4ce4ecfe4bcd23a4b675e76315f2866e588686cc97fccdf"}, + {file = "matplotlib-3.9.0-cp39-cp39-macosx_10_12_x86_64.whl", hash = "sha256:616fabf4981a3b3c5a15cd95eba359c8489c4e20e03717aea42866d8d0465956"}, + {file = "matplotlib-3.9.0-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:cd53c79fd02f1c1808d2cfc87dd3cf4dbc63c5244a58ee7944497107469c8d8a"}, + {file = "matplotlib-3.9.0-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:06a478f0d67636554fa78558cfbcd7b9dba85b51f5c3b5a0c9be49010cf5f321"}, + {file = "matplotlib-3.9.0-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:81c40af649d19c85f8073e25e5806926986806fa6d54be506fbf02aef47d5a89"}, + {file = "matplotlib-3.9.0-cp39-cp39-musllinux_1_1_x86_64.whl", hash = "sha256:52146fc3bd7813cc784562cb93a15788be0b2875c4655e2cc6ea646bfa30344b"}, + {file = "matplotlib-3.9.0-cp39-cp39-win_amd64.whl", hash = "sha256:0fc51eaa5262553868461c083d9adadb11a6017315f3a757fc45ec6ec5f02888"}, + {file = "matplotlib-3.9.0-pp39-pypy39_pp73-macosx_10_12_x86_64.whl", hash = "sha256:bd4f2831168afac55b881db82a7730992aa41c4f007f1913465fb182d6fb20c0"}, + {file = "matplotlib-3.9.0-pp39-pypy39_pp73-macosx_11_0_arm64.whl", hash = "sha256:290d304e59be2b33ef5c2d768d0237f5bd132986bdcc66f80bc9bcc300066a03"}, + {file = "matplotlib-3.9.0-pp39-pypy39_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:7ff2e239c26be4f24bfa45860c20ffccd118d270c5b5d081fa4ea409b5469fcd"}, + {file = "matplotlib-3.9.0-pp39-pypy39_pp73-win_amd64.whl", hash = "sha256:af4001b7cae70f7eaacfb063db605280058246de590fa7874f00f62259f2df7e"}, + {file = "matplotlib-3.9.0.tar.gz", hash = "sha256:e6d29ea6c19e34b30fb7d88b7081f869a03014f66fe06d62cc77d5a6ea88ed7a"}, ] [package.dependencies] @@ -2139,21 +2171,24 @@ cycler = ">=0.10" fonttools = ">=4.22.0" importlib-resources = {version = ">=3.2.0", markers = "python_version < \"3.10\""} kiwisolver = ">=1.3.1" -numpy = ">=1.21,<2" +numpy = ">=1.23" packaging = ">=20.0" pillow = ">=8" pyparsing = ">=2.3.1" python-dateutil = ">=2.7" +[package.extras] +dev = ["meson-python (>=0.13.1)", "numpy (>=1.25)", "pybind11 (>=2.6)", "setuptools (>=64)", "setuptools_scm (>=7)"] + [[package]] name = "matplotlib-inline" -version = "0.1.6" +version = "0.1.7" description = "Inline Matplotlib backend for Jupyter" optional = false -python-versions = ">=3.5" +python-versions = ">=3.8" files = [ - {file = "matplotlib-inline-0.1.6.tar.gz", hash = "sha256:f887e5f10ba98e8d2b150ddcf4702c1e5f8b3a20005eb0f74bfdbd360ee6f304"}, - {file = "matplotlib_inline-0.1.6-py3-none-any.whl", hash = "sha256:f1f41aab5328aa5aaea9b16d083b128102f8712542f819fe7e6a420ff581b311"}, + {file = "matplotlib_inline-0.1.7-py3-none-any.whl", hash = "sha256:df192d39a4ff8f21b1895d72e6a13f5fcc5099f00fa84384e0ea28c2cc0653ca"}, + {file = "matplotlib_inline-0.1.7.tar.gz", hash = "sha256:8423b23ec666be3d16e16b60bdd8ac4e86e840ebd1dd11a30b9f117f2fa0ab90"}, ] [package.dependencies] @@ -2161,13 +2196,13 @@ traitlets = "" [[package]] name = "mdit-py-plugins" -version = "0.4.0" +version = "0.4.1" description = "Collection of plugins for markdown-it-py" optional = false python-versions = ">=3.8" files = [ - {file = "mdit_py_plugins-0.4.0-py3-none-any.whl", hash = "sha256:b51b3bb70691f57f974e257e367107857a93b36f322a9e6d44ca5bf28ec2def9"}, - {file = "mdit_py_plugins-0.4.0.tar.gz", hash = "sha256:d8ab27e9aed6c38aa716819fedfde15ca275715955f8a185a8e1cf90fb1d2c1b"}, + {file = "mdit_py_plugins-0.4.1-py3-none-any.whl", hash = "sha256:1020dfe4e6bfc2c79fb49ae4e3f5b297f5ccd20f010187acc52af2921e27dc6a"}, + {file = "mdit_py_plugins-0.4.1.tar.gz", hash = "sha256:834b8ac23d1cd60cec703646ffd22ae97b7955a6d596eb1d304be1e251ae499c"}, ] [package.dependencies] @@ -2202,59 +2237,77 @@ files = [ [[package]] name = "mkdocs" -version = "1.5.3" +version = "1.6.0" description = "Project documentation with Markdown." optional = false -python-versions = ">=3.7" +python-versions = ">=3.8" files = [ - {file = "mkdocs-1.5.3-py3-none-any.whl", hash = "sha256:3b3a78e736b31158d64dbb2f8ba29bd46a379d0c6e324c2246c3bc3d2189cfc1"}, - {file = "mkdocs-1.5.3.tar.gz", hash = "sha256:eb7c99214dcb945313ba30426c2451b735992c73c2e10838f76d09e39ff4d0e2"}, + {file = "mkdocs-1.6.0-py3-none-any.whl", hash = "sha256:1eb5cb7676b7d89323e62b56235010216319217d4af5ddc543a91beb8d125ea7"}, + {file = "mkdocs-1.6.0.tar.gz", hash = "sha256:a73f735824ef83a4f3bcb7a231dcab23f5a838f88b7efc54a0eef5fbdbc3c512"}, ] [package.dependencies] click = ">=7.0" colorama = {version = ">=0.4", markers = "platform_system == \"Windows\""} ghp-import = ">=1.0" -importlib-metadata = {version = ">=4.3", markers = "python_version < \"3.10\""} +importlib-metadata = {version = ">=4.4", markers = "python_version < \"3.10\""} jinja2 = ">=2.11.1" -markdown = ">=3.2.1" +markdown = ">=3.3.6" markupsafe = ">=2.0.1" mergedeep = ">=1.3.4" +mkdocs-get-deps = ">=0.2.0" packaging = ">=20.5" pathspec = ">=0.11.1" -platformdirs = ">=2.2.0" pyyaml = ">=5.1" pyyaml-env-tag = ">=0.1" watchdog = ">=2.0" [package.extras] i18n = ["babel (>=2.9.0)"] -min-versions = ["babel (==2.9.0)", "click (==7.0)", "colorama (==0.4)", "ghp-import (==1.0)", "importlib-metadata (==4.3)", "jinja2 (==2.11.1)", "markdown (==3.2.1)", "markupsafe (==2.0.1)", "mergedeep (==1.3.4)", "packaging (==20.5)", "pathspec (==0.11.1)", "platformdirs (==2.2.0)", "pyyaml (==5.1)", "pyyaml-env-tag (==0.1)", "typing-extensions (==3.10)", "watchdog (==2.0)"] +min-versions = ["babel (==2.9.0)", "click (==7.0)", "colorama (==0.4)", "ghp-import (==1.0)", "importlib-metadata (==4.4)", "jinja2 (==2.11.1)", "markdown (==3.3.6)", "markupsafe (==2.0.1)", "mergedeep (==1.3.4)", "mkdocs-get-deps (==0.2.0)", "packaging (==20.5)", "pathspec (==0.11.1)", "pyyaml (==5.1)", "pyyaml-env-tag (==0.1)", "watchdog (==2.0)"] [[package]] name = "mkdocs-autorefs" -version = "0.5.0" +version = "1.0.1" description = "Automatically link across pages in MkDocs." optional = false python-versions = ">=3.8" files = [ - {file = "mkdocs_autorefs-0.5.0-py3-none-any.whl", hash = "sha256:7930fcb8ac1249f10e683967aeaddc0af49d90702af111a5e390e8b20b3d97ff"}, - {file = "mkdocs_autorefs-0.5.0.tar.gz", hash = "sha256:9a5054a94c08d28855cfab967ada10ed5be76e2bfad642302a610b252c3274c0"}, + {file = "mkdocs_autorefs-1.0.1-py3-none-any.whl", hash = "sha256:aacdfae1ab197780fb7a2dac92ad8a3d8f7ca8049a9cbe56a4218cd52e8da570"}, + {file = "mkdocs_autorefs-1.0.1.tar.gz", hash = "sha256:f684edf847eced40b570b57846b15f0bf57fb93ac2c510450775dcf16accb971"}, ] [package.dependencies] Markdown = ">=3.3" +markupsafe = ">=2.0.1" mkdocs = ">=1.1" +[[package]] +name = "mkdocs-get-deps" +version = "0.2.0" +description = "MkDocs extension that lists all dependencies according to a mkdocs.yml file" +optional = false +python-versions = ">=3.8" +files = [ + {file = "mkdocs_get_deps-0.2.0-py3-none-any.whl", hash = "sha256:2bf11d0b133e77a0dd036abeeb06dec8775e46efa526dc70667d8863eefc6134"}, + {file = "mkdocs_get_deps-0.2.0.tar.gz", hash = "sha256:162b3d129c7fad9b19abfdcb9c1458a651628e4b1dea628ac68790fb3061c60c"}, +] + +[package.dependencies] +importlib-metadata = {version = ">=4.3", markers = "python_version < \"3.10\""} +mergedeep = ">=1.3.4" +platformdirs = ">=2.2.0" +pyyaml = ">=5.1" + [[package]] name = "mkdocs-material" -version = "9.5.7" +version = "9.5.24" description = "Documentation that simply works" optional = false python-versions = ">=3.8" files = [ - {file = "mkdocs_material-9.5.7-py3-none-any.whl", hash = "sha256:0be8ce8bcfebb52bae9b00cf9b851df45b8a92d629afcfd7f2c09b2dfa155ea3"}, - {file = "mkdocs_material-9.5.7.tar.gz", hash = "sha256:16110292575d88a338d2961f3cb665cf12943ff8829e551a9b364f24019e46af"}, + {file = "mkdocs_material-9.5.24-py3-none-any.whl", hash = "sha256:e12cd75954c535b61e716f359cf2a5056bf4514889d17161fdebd5df4b0153c6"}, + {file = "mkdocs_material-9.5.24.tar.gz", hash = "sha256:02d5aaba0ee755e707c3ef6e748f9acb7b3011187c0ea766db31af8905078a34"}, ] [package.dependencies] @@ -2262,7 +2315,7 @@ babel = ">=2.10,<3.0" colorama = ">=0.4,<1.0" jinja2 = ">=3.0,<4.0" markdown = ">=3.2,<4.0" -mkdocs = ">=1.5.3,<1.6.0" +mkdocs = ">=1.6,<2.0" mkdocs-material-extensions = ">=1.3,<2.0" paginate = ">=0.5,<1.0" pygments = ">=2.16,<3.0" @@ -2271,7 +2324,7 @@ regex = ">=2022.4" requests = ">=2.26,<3.0" [package.extras] -git = ["mkdocs-git-committers-plugin-2 (>=1.1,<2.0)", "mkdocs-git-revision-date-localized-plugin (>=1.2,<2.0)"] +git = ["mkdocs-git-committers-plugin-2 (>=1.1,<2.0)", "mkdocs-git-revision-date-localized-plugin (>=1.2.4,<2.0)"] imaging = ["cairosvg (>=2.6,<3.0)", "pillow (>=10.2,<11.0)"] recommended = ["mkdocs-minify-plugin (>=0.7,<1.0)", "mkdocs-redirects (>=1.2,<2.0)", "mkdocs-rss-plugin (>=1.6,<2.0)"] @@ -2288,13 +2341,13 @@ files = [ [[package]] name = "mkdocstrings" -version = "0.24.0" +version = "0.25.1" description = "Automatic documentation from sources, for MkDocs." optional = false python-versions = ">=3.8" files = [ - {file = "mkdocstrings-0.24.0-py3-none-any.whl", hash = "sha256:f4908560c10f587326d8f5165d1908817b2e280bbf707607f601c996366a2264"}, - {file = "mkdocstrings-0.24.0.tar.gz", hash = "sha256:222b1165be41257b494a9d29b14135d2b7ca43f38161d5b10caae03b87bd4f7e"}, + {file = "mkdocstrings-0.25.1-py3-none-any.whl", hash = "sha256:da01fcc2670ad61888e8fe5b60afe9fee5781017d67431996832d63e887c2e51"}, + {file = "mkdocstrings-0.25.1.tar.gz", hash = "sha256:c3a2515f31577f311a9ee58d089e4c51fc6046dbd9e9b4c3de4c3194667fe9bf"}, ] [package.dependencies] @@ -2316,82 +2369,82 @@ python-legacy = ["mkdocstrings-python-legacy (>=0.2.1)"] [[package]] name = "mkdocstrings-python" -version = "1.8.0" +version = "1.10.3" description = "A Python handler for mkdocstrings." optional = false python-versions = ">=3.8" files = [ - {file = "mkdocstrings_python-1.8.0-py3-none-any.whl", hash = "sha256:4209970cc90bec194568682a535848a8d8489516c6ed4adbe58bbc67b699ca9d"}, - {file = "mkdocstrings_python-1.8.0.tar.gz", hash = "sha256:1488bddf50ee42c07d9a488dddc197f8e8999c2899687043ec5dd1643d057192"}, + {file = "mkdocstrings_python-1.10.3-py3-none-any.whl", hash = "sha256:11ff6d21d3818fb03af82c3ea6225b1534837e17f790aa5f09626524171f949b"}, + {file = "mkdocstrings_python-1.10.3.tar.gz", hash = "sha256:321cf9c732907ab2b1fedaafa28765eaa089d89320f35f7206d00ea266889d03"}, ] [package.dependencies] -griffe = ">=0.37" -mkdocstrings = ">=0.20" +griffe = ">=0.44" +mkdocstrings = ">=0.25" [[package]] name = "msgpack" -version = "1.0.7" +version = "1.0.8" description = "MessagePack serializer" optional = false python-versions = ">=3.8" files = [ - {file = "msgpack-1.0.7-cp310-cp310-macosx_10_9_universal2.whl", hash = "sha256:04ad6069c86e531682f9e1e71b71c1c3937d6014a7c3e9edd2aa81ad58842862"}, - {file = "msgpack-1.0.7-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:cca1b62fe70d761a282496b96a5e51c44c213e410a964bdffe0928e611368329"}, - {file = "msgpack-1.0.7-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:e50ebce52f41370707f1e21a59514e3375e3edd6e1832f5e5235237db933c98b"}, - {file = "msgpack-1.0.7-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:4a7b4f35de6a304b5533c238bee86b670b75b03d31b7797929caa7a624b5dda6"}, - {file = "msgpack-1.0.7-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:28efb066cde83c479dfe5a48141a53bc7e5f13f785b92ddde336c716663039ee"}, - {file = "msgpack-1.0.7-cp310-cp310-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:4cb14ce54d9b857be9591ac364cb08dc2d6a5c4318c1182cb1d02274029d590d"}, - {file = "msgpack-1.0.7-cp310-cp310-musllinux_1_1_aarch64.whl", hash = "sha256:b573a43ef7c368ba4ea06050a957c2a7550f729c31f11dd616d2ac4aba99888d"}, - {file = "msgpack-1.0.7-cp310-cp310-musllinux_1_1_i686.whl", hash = "sha256:ccf9a39706b604d884d2cb1e27fe973bc55f2890c52f38df742bc1d79ab9f5e1"}, - {file = "msgpack-1.0.7-cp310-cp310-musllinux_1_1_x86_64.whl", hash = "sha256:cb70766519500281815dfd7a87d3a178acf7ce95390544b8c90587d76b227681"}, - {file = "msgpack-1.0.7-cp310-cp310-win32.whl", hash = "sha256:b610ff0f24e9f11c9ae653c67ff8cc03c075131401b3e5ef4b82570d1728f8a9"}, - {file = "msgpack-1.0.7-cp310-cp310-win_amd64.whl", hash = "sha256:a40821a89dc373d6427e2b44b572efc36a2778d3f543299e2f24eb1a5de65415"}, - {file = "msgpack-1.0.7-cp311-cp311-macosx_10_9_universal2.whl", hash = "sha256:576eb384292b139821c41995523654ad82d1916da6a60cff129c715a6223ea84"}, - {file = "msgpack-1.0.7-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:730076207cb816138cf1af7f7237b208340a2c5e749707457d70705715c93b93"}, - {file = "msgpack-1.0.7-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:85765fdf4b27eb5086f05ac0491090fc76f4f2b28e09d9350c31aac25a5aaff8"}, - {file = "msgpack-1.0.7-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:3476fae43db72bd11f29a5147ae2f3cb22e2f1a91d575ef130d2bf49afd21c46"}, - {file = "msgpack-1.0.7-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:6d4c80667de2e36970ebf74f42d1088cc9ee7ef5f4e8c35eee1b40eafd33ca5b"}, - {file = "msgpack-1.0.7-cp311-cp311-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:5b0bf0effb196ed76b7ad883848143427a73c355ae8e569fa538365064188b8e"}, - {file = "msgpack-1.0.7-cp311-cp311-musllinux_1_1_aarch64.whl", hash = "sha256:f9a7c509542db4eceed3dcf21ee5267ab565a83555c9b88a8109dcecc4709002"}, - {file = "msgpack-1.0.7-cp311-cp311-musllinux_1_1_i686.whl", hash = "sha256:84b0daf226913133f899ea9b30618722d45feffa67e4fe867b0b5ae83a34060c"}, - {file = "msgpack-1.0.7-cp311-cp311-musllinux_1_1_x86_64.whl", hash = "sha256:ec79ff6159dffcc30853b2ad612ed572af86c92b5168aa3fc01a67b0fa40665e"}, - {file = "msgpack-1.0.7-cp311-cp311-win32.whl", hash = "sha256:3e7bf4442b310ff154b7bb9d81eb2c016b7d597e364f97d72b1acc3817a0fdc1"}, - {file = "msgpack-1.0.7-cp311-cp311-win_amd64.whl", hash = "sha256:3f0c8c6dfa6605ab8ff0611995ee30d4f9fcff89966cf562733b4008a3d60d82"}, - {file = "msgpack-1.0.7-cp312-cp312-macosx_10_9_universal2.whl", hash = "sha256:f0936e08e0003f66bfd97e74ee530427707297b0d0361247e9b4f59ab78ddc8b"}, - {file = "msgpack-1.0.7-cp312-cp312-macosx_10_9_x86_64.whl", hash = "sha256:98bbd754a422a0b123c66a4c341de0474cad4a5c10c164ceed6ea090f3563db4"}, - {file = "msgpack-1.0.7-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:b291f0ee7961a597cbbcc77709374087fa2a9afe7bdb6a40dbbd9b127e79afee"}, - {file = "msgpack-1.0.7-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:ebbbba226f0a108a7366bf4b59bf0f30a12fd5e75100c630267d94d7f0ad20e5"}, - {file = "msgpack-1.0.7-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:1e2d69948e4132813b8d1131f29f9101bc2c915f26089a6d632001a5c1349672"}, - {file = "msgpack-1.0.7-cp312-cp312-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:bdf38ba2d393c7911ae989c3bbba510ebbcdf4ecbdbfec36272abe350c454075"}, - {file = "msgpack-1.0.7-cp312-cp312-musllinux_1_1_aarch64.whl", hash = "sha256:993584fc821c58d5993521bfdcd31a4adf025c7d745bbd4d12ccfecf695af5ba"}, - {file = "msgpack-1.0.7-cp312-cp312-musllinux_1_1_i686.whl", hash = "sha256:52700dc63a4676669b341ba33520f4d6e43d3ca58d422e22ba66d1736b0a6e4c"}, - {file = "msgpack-1.0.7-cp312-cp312-musllinux_1_1_x86_64.whl", hash = "sha256:e45ae4927759289c30ccba8d9fdce62bb414977ba158286b5ddaf8df2cddb5c5"}, - {file = "msgpack-1.0.7-cp312-cp312-win32.whl", hash = "sha256:27dcd6f46a21c18fa5e5deed92a43d4554e3df8d8ca5a47bf0615d6a5f39dbc9"}, - {file = "msgpack-1.0.7-cp312-cp312-win_amd64.whl", hash = "sha256:7687e22a31e976a0e7fc99c2f4d11ca45eff652a81eb8c8085e9609298916dcf"}, - {file = "msgpack-1.0.7-cp38-cp38-macosx_10_9_universal2.whl", hash = "sha256:5b6ccc0c85916998d788b295765ea0e9cb9aac7e4a8ed71d12e7d8ac31c23c95"}, - {file = "msgpack-1.0.7-cp38-cp38-macosx_10_9_x86_64.whl", hash = "sha256:235a31ec7db685f5c82233bddf9858748b89b8119bf4538d514536c485c15fe0"}, - {file = "msgpack-1.0.7-cp38-cp38-macosx_11_0_arm64.whl", hash = "sha256:cab3db8bab4b7e635c1c97270d7a4b2a90c070b33cbc00c99ef3f9be03d3e1f7"}, - {file = "msgpack-1.0.7-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:0bfdd914e55e0d2c9e1526de210f6fe8ffe9705f2b1dfcc4aecc92a4cb4b533d"}, - {file = "msgpack-1.0.7-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:36e17c4592231a7dbd2ed09027823ab295d2791b3b1efb2aee874b10548b7524"}, - {file = "msgpack-1.0.7-cp38-cp38-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:38949d30b11ae5f95c3c91917ee7a6b239f5ec276f271f28638dec9156f82cfc"}, - {file = "msgpack-1.0.7-cp38-cp38-musllinux_1_1_aarch64.whl", hash = "sha256:ff1d0899f104f3921d94579a5638847f783c9b04f2d5f229392ca77fba5b82fc"}, - {file = "msgpack-1.0.7-cp38-cp38-musllinux_1_1_i686.whl", hash = "sha256:dc43f1ec66eb8440567186ae2f8c447d91e0372d793dfe8c222aec857b81a8cf"}, - {file = "msgpack-1.0.7-cp38-cp38-musllinux_1_1_x86_64.whl", hash = "sha256:dd632777ff3beaaf629f1ab4396caf7ba0bdd075d948a69460d13d44357aca4c"}, - {file = "msgpack-1.0.7-cp38-cp38-win32.whl", hash = "sha256:4e71bc4416de195d6e9b4ee93ad3f2f6b2ce11d042b4d7a7ee00bbe0358bd0c2"}, - {file = "msgpack-1.0.7-cp38-cp38-win_amd64.whl", hash = "sha256:8f5b234f567cf76ee489502ceb7165c2a5cecec081db2b37e35332b537f8157c"}, - {file = "msgpack-1.0.7-cp39-cp39-macosx_10_9_universal2.whl", hash = "sha256:bfef2bb6ef068827bbd021017a107194956918ab43ce4d6dc945ffa13efbc25f"}, - {file = "msgpack-1.0.7-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:484ae3240666ad34cfa31eea7b8c6cd2f1fdaae21d73ce2974211df099a95d81"}, - {file = "msgpack-1.0.7-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:3967e4ad1aa9da62fd53e346ed17d7b2e922cba5ab93bdd46febcac39be636fc"}, - {file = "msgpack-1.0.7-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:8dd178c4c80706546702c59529ffc005681bd6dc2ea234c450661b205445a34d"}, - {file = "msgpack-1.0.7-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:f6ffbc252eb0d229aeb2f9ad051200668fc3a9aaa8994e49f0cb2ffe2b7867e7"}, - {file = "msgpack-1.0.7-cp39-cp39-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:822ea70dc4018c7e6223f13affd1c5c30c0f5c12ac1f96cd8e9949acddb48a61"}, - {file = "msgpack-1.0.7-cp39-cp39-musllinux_1_1_aarch64.whl", hash = "sha256:384d779f0d6f1b110eae74cb0659d9aa6ff35aaf547b3955abf2ab4c901c4819"}, - {file = "msgpack-1.0.7-cp39-cp39-musllinux_1_1_i686.whl", hash = "sha256:f64e376cd20d3f030190e8c32e1c64582eba56ac6dc7d5b0b49a9d44021b52fd"}, - {file = "msgpack-1.0.7-cp39-cp39-musllinux_1_1_x86_64.whl", hash = "sha256:5ed82f5a7af3697b1c4786053736f24a0efd0a1b8a130d4c7bfee4b9ded0f08f"}, - {file = "msgpack-1.0.7-cp39-cp39-win32.whl", hash = "sha256:f26a07a6e877c76a88e3cecac8531908d980d3d5067ff69213653649ec0f60ad"}, - {file = "msgpack-1.0.7-cp39-cp39-win_amd64.whl", hash = "sha256:1dc93e8e4653bdb5910aed79f11e165c85732067614f180f70534f056da97db3"}, - {file = "msgpack-1.0.7.tar.gz", hash = "sha256:572efc93db7a4d27e404501975ca6d2d9775705c2d922390d878fcf768d92c87"}, + {file = "msgpack-1.0.8-cp310-cp310-macosx_10_9_universal2.whl", hash = "sha256:505fe3d03856ac7d215dbe005414bc28505d26f0c128906037e66d98c4e95868"}, + {file = "msgpack-1.0.8-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:e6b7842518a63a9f17107eb176320960ec095a8ee3b4420b5f688e24bf50c53c"}, + {file = "msgpack-1.0.8-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:376081f471a2ef24828b83a641a02c575d6103a3ad7fd7dade5486cad10ea659"}, + {file = "msgpack-1.0.8-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:5e390971d082dba073c05dbd56322427d3280b7cc8b53484c9377adfbae67dc2"}, + {file = "msgpack-1.0.8-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:00e073efcba9ea99db5acef3959efa45b52bc67b61b00823d2a1a6944bf45982"}, + {file = "msgpack-1.0.8-cp310-cp310-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:82d92c773fbc6942a7a8b520d22c11cfc8fd83bba86116bfcf962c2f5c2ecdaa"}, + {file = "msgpack-1.0.8-cp310-cp310-musllinux_1_1_aarch64.whl", hash = "sha256:9ee32dcb8e531adae1f1ca568822e9b3a738369b3b686d1477cbc643c4a9c128"}, + {file = "msgpack-1.0.8-cp310-cp310-musllinux_1_1_i686.whl", hash = "sha256:e3aa7e51d738e0ec0afbed661261513b38b3014754c9459508399baf14ae0c9d"}, + {file = "msgpack-1.0.8-cp310-cp310-musllinux_1_1_x86_64.whl", hash = "sha256:69284049d07fce531c17404fcba2bb1df472bc2dcdac642ae71a2d079d950653"}, + {file = "msgpack-1.0.8-cp310-cp310-win32.whl", hash = "sha256:13577ec9e247f8741c84d06b9ece5f654920d8365a4b636ce0e44f15e07ec693"}, + {file = "msgpack-1.0.8-cp310-cp310-win_amd64.whl", hash = "sha256:e532dbd6ddfe13946de050d7474e3f5fb6ec774fbb1a188aaf469b08cf04189a"}, + {file = "msgpack-1.0.8-cp311-cp311-macosx_10_9_universal2.whl", hash = "sha256:9517004e21664f2b5a5fd6333b0731b9cf0817403a941b393d89a2f1dc2bd836"}, + {file = "msgpack-1.0.8-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:d16a786905034e7e34098634b184a7d81f91d4c3d246edc6bd7aefb2fd8ea6ad"}, + {file = "msgpack-1.0.8-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:e2872993e209f7ed04d963e4b4fbae72d034844ec66bc4ca403329db2074377b"}, + {file = "msgpack-1.0.8-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:5c330eace3dd100bdb54b5653b966de7f51c26ec4a7d4e87132d9b4f738220ba"}, + {file = "msgpack-1.0.8-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:83b5c044f3eff2a6534768ccfd50425939e7a8b5cf9a7261c385de1e20dcfc85"}, + {file = "msgpack-1.0.8-cp311-cp311-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:1876b0b653a808fcd50123b953af170c535027bf1d053b59790eebb0aeb38950"}, + {file = "msgpack-1.0.8-cp311-cp311-musllinux_1_1_aarch64.whl", hash = "sha256:dfe1f0f0ed5785c187144c46a292b8c34c1295c01da12e10ccddfc16def4448a"}, + {file = "msgpack-1.0.8-cp311-cp311-musllinux_1_1_i686.whl", hash = "sha256:3528807cbbb7f315bb81959d5961855e7ba52aa60a3097151cb21956fbc7502b"}, + {file = "msgpack-1.0.8-cp311-cp311-musllinux_1_1_x86_64.whl", hash = "sha256:e2f879ab92ce502a1e65fce390eab619774dda6a6ff719718069ac94084098ce"}, + {file = "msgpack-1.0.8-cp311-cp311-win32.whl", hash = "sha256:26ee97a8261e6e35885c2ecd2fd4a6d38252246f94a2aec23665a4e66d066305"}, + {file = "msgpack-1.0.8-cp311-cp311-win_amd64.whl", hash = "sha256:eadb9f826c138e6cf3c49d6f8de88225a3c0ab181a9b4ba792e006e5292d150e"}, + {file = "msgpack-1.0.8-cp312-cp312-macosx_10_9_universal2.whl", hash = "sha256:114be227f5213ef8b215c22dde19532f5da9652e56e8ce969bf0a26d7c419fee"}, + {file = "msgpack-1.0.8-cp312-cp312-macosx_10_9_x86_64.whl", hash = "sha256:d661dc4785affa9d0edfdd1e59ec056a58b3dbb9f196fa43587f3ddac654ac7b"}, + {file = "msgpack-1.0.8-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:d56fd9f1f1cdc8227d7b7918f55091349741904d9520c65f0139a9755952c9e8"}, + {file = "msgpack-1.0.8-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:0726c282d188e204281ebd8de31724b7d749adebc086873a59efb8cf7ae27df3"}, + {file = "msgpack-1.0.8-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:8db8e423192303ed77cff4dce3a4b88dbfaf43979d280181558af5e2c3c71afc"}, + {file = "msgpack-1.0.8-cp312-cp312-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:99881222f4a8c2f641f25703963a5cefb076adffd959e0558dc9f803a52d6a58"}, + {file = "msgpack-1.0.8-cp312-cp312-musllinux_1_1_aarch64.whl", hash = "sha256:b5505774ea2a73a86ea176e8a9a4a7c8bf5d521050f0f6f8426afe798689243f"}, + {file = "msgpack-1.0.8-cp312-cp312-musllinux_1_1_i686.whl", hash = "sha256:ef254a06bcea461e65ff0373d8a0dd1ed3aa004af48839f002a0c994a6f72d04"}, + {file = "msgpack-1.0.8-cp312-cp312-musllinux_1_1_x86_64.whl", hash = "sha256:e1dd7839443592d00e96db831eddb4111a2a81a46b028f0facd60a09ebbdd543"}, + {file = "msgpack-1.0.8-cp312-cp312-win32.whl", hash = "sha256:64d0fcd436c5683fdd7c907eeae5e2cbb5eb872fafbc03a43609d7941840995c"}, + {file = "msgpack-1.0.8-cp312-cp312-win_amd64.whl", hash = "sha256:74398a4cf19de42e1498368c36eed45d9528f5fd0155241e82c4082b7e16cffd"}, + {file = "msgpack-1.0.8-cp38-cp38-macosx_10_9_universal2.whl", hash = "sha256:0ceea77719d45c839fd73abcb190b8390412a890df2f83fb8cf49b2a4b5c2f40"}, + {file = "msgpack-1.0.8-cp38-cp38-macosx_10_9_x86_64.whl", hash = "sha256:1ab0bbcd4d1f7b6991ee7c753655b481c50084294218de69365f8f1970d4c151"}, + {file = "msgpack-1.0.8-cp38-cp38-macosx_11_0_arm64.whl", hash = "sha256:1cce488457370ffd1f953846f82323cb6b2ad2190987cd4d70b2713e17268d24"}, + {file = "msgpack-1.0.8-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:3923a1778f7e5ef31865893fdca12a8d7dc03a44b33e2a5f3295416314c09f5d"}, + {file = "msgpack-1.0.8-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:a22e47578b30a3e199ab067a4d43d790249b3c0587d9a771921f86250c8435db"}, + {file = "msgpack-1.0.8-cp38-cp38-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:bd739c9251d01e0279ce729e37b39d49a08c0420d3fee7f2a4968c0576678f77"}, + {file = "msgpack-1.0.8-cp38-cp38-musllinux_1_1_aarch64.whl", hash = "sha256:d3420522057ebab1728b21ad473aa950026d07cb09da41103f8e597dfbfaeb13"}, + {file = "msgpack-1.0.8-cp38-cp38-musllinux_1_1_i686.whl", hash = "sha256:5845fdf5e5d5b78a49b826fcdc0eb2e2aa7191980e3d2cfd2a30303a74f212e2"}, + {file = "msgpack-1.0.8-cp38-cp38-musllinux_1_1_x86_64.whl", hash = "sha256:6a0e76621f6e1f908ae52860bdcb58e1ca85231a9b0545e64509c931dd34275a"}, + {file = "msgpack-1.0.8-cp38-cp38-win32.whl", hash = "sha256:374a8e88ddab84b9ada695d255679fb99c53513c0a51778796fcf0944d6c789c"}, + {file = "msgpack-1.0.8-cp38-cp38-win_amd64.whl", hash = "sha256:f3709997b228685fe53e8c433e2df9f0cdb5f4542bd5114ed17ac3c0129b0480"}, + {file = "msgpack-1.0.8-cp39-cp39-macosx_10_9_universal2.whl", hash = "sha256:f51bab98d52739c50c56658cc303f190785f9a2cd97b823357e7aeae54c8f68a"}, + {file = "msgpack-1.0.8-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:73ee792784d48aa338bba28063e19a27e8d989344f34aad14ea6e1b9bd83f596"}, + {file = "msgpack-1.0.8-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:f9904e24646570539a8950400602d66d2b2c492b9010ea7e965025cb71d0c86d"}, + {file = "msgpack-1.0.8-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:e75753aeda0ddc4c28dce4c32ba2f6ec30b1b02f6c0b14e547841ba5b24f753f"}, + {file = "msgpack-1.0.8-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:5dbf059fb4b7c240c873c1245ee112505be27497e90f7c6591261c7d3c3a8228"}, + {file = "msgpack-1.0.8-cp39-cp39-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:4916727e31c28be8beaf11cf117d6f6f188dcc36daae4e851fee88646f5b6b18"}, + {file = "msgpack-1.0.8-cp39-cp39-musllinux_1_1_aarch64.whl", hash = "sha256:7938111ed1358f536daf311be244f34df7bf3cdedb3ed883787aca97778b28d8"}, + {file = "msgpack-1.0.8-cp39-cp39-musllinux_1_1_i686.whl", hash = "sha256:493c5c5e44b06d6c9268ce21b302c9ca055c1fd3484c25ba41d34476c76ee746"}, + {file = "msgpack-1.0.8-cp39-cp39-musllinux_1_1_x86_64.whl", hash = "sha256:5fbb160554e319f7b22ecf530a80a3ff496d38e8e07ae763b9e82fadfe96f273"}, + {file = "msgpack-1.0.8-cp39-cp39-win32.whl", hash = "sha256:f9af38a89b6a5c04b7d18c492c8ccf2aee7048aff1ce8437c4683bb5a1df893d"}, + {file = "msgpack-1.0.8-cp39-cp39-win_amd64.whl", hash = "sha256:ed59dd52075f8fc91da6053b12e8c89e37aa043f8986efd89e61fae69dc1b011"}, + {file = "msgpack-1.0.8-py3-none-any.whl", hash = "sha256:24f727df1e20b9876fa6e95f840a2a2651e34c0ad147676356f4bf5fbb0206ca"}, ] [[package]] @@ -2506,38 +2559,38 @@ files = [ [[package]] name = "mypy" -version = "1.8.0" +version = "1.10.0" description = "Optional static typing for Python" optional = false python-versions = ">=3.8" files = [ - {file = "mypy-1.8.0-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:485a8942f671120f76afffff70f259e1cd0f0cfe08f81c05d8816d958d4577d3"}, - {file = "mypy-1.8.0-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:df9824ac11deaf007443e7ed2a4a26bebff98d2bc43c6da21b2b64185da011c4"}, - {file = "mypy-1.8.0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:2afecd6354bbfb6e0160f4e4ad9ba6e4e003b767dd80d85516e71f2e955ab50d"}, - {file = "mypy-1.8.0-cp310-cp310-musllinux_1_1_x86_64.whl", hash = "sha256:8963b83d53ee733a6e4196954502b33567ad07dfd74851f32be18eb932fb1cb9"}, - {file = "mypy-1.8.0-cp310-cp310-win_amd64.whl", hash = "sha256:e46f44b54ebddbeedbd3d5b289a893219065ef805d95094d16a0af6630f5d410"}, - {file = "mypy-1.8.0-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:855fe27b80375e5c5878492f0729540db47b186509c98dae341254c8f45f42ae"}, - {file = "mypy-1.8.0-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:4c886c6cce2d070bd7df4ec4a05a13ee20c0aa60cb587e8d1265b6c03cf91da3"}, - {file = "mypy-1.8.0-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:d19c413b3c07cbecf1f991e2221746b0d2a9410b59cb3f4fb9557f0365a1a817"}, - {file = "mypy-1.8.0-cp311-cp311-musllinux_1_1_x86_64.whl", hash = "sha256:9261ed810972061388918c83c3f5cd46079d875026ba97380f3e3978a72f503d"}, - {file = "mypy-1.8.0-cp311-cp311-win_amd64.whl", hash = "sha256:51720c776d148bad2372ca21ca29256ed483aa9a4cdefefcef49006dff2a6835"}, - {file = "mypy-1.8.0-cp312-cp312-macosx_10_9_x86_64.whl", hash = "sha256:52825b01f5c4c1c4eb0db253ec09c7aa17e1a7304d247c48b6f3599ef40db8bd"}, - {file = "mypy-1.8.0-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:f5ac9a4eeb1ec0f1ccdc6f326bcdb464de5f80eb07fb38b5ddd7b0de6bc61e55"}, - {file = "mypy-1.8.0-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:afe3fe972c645b4632c563d3f3eff1cdca2fa058f730df2b93a35e3b0c538218"}, - {file = "mypy-1.8.0-cp312-cp312-musllinux_1_1_x86_64.whl", hash = "sha256:42c6680d256ab35637ef88891c6bd02514ccb7e1122133ac96055ff458f93fc3"}, - {file = "mypy-1.8.0-cp312-cp312-win_amd64.whl", hash = "sha256:720a5ca70e136b675af3af63db533c1c8c9181314d207568bbe79051f122669e"}, - {file = "mypy-1.8.0-cp38-cp38-macosx_10_9_x86_64.whl", hash = "sha256:028cf9f2cae89e202d7b6593cd98db6759379f17a319b5faf4f9978d7084cdc6"}, - {file = "mypy-1.8.0-cp38-cp38-macosx_11_0_arm64.whl", hash = "sha256:4e6d97288757e1ddba10dd9549ac27982e3e74a49d8d0179fc14d4365c7add66"}, - {file = "mypy-1.8.0-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:7f1478736fcebb90f97e40aff11a5f253af890c845ee0c850fe80aa060a267c6"}, - {file = "mypy-1.8.0-cp38-cp38-musllinux_1_1_x86_64.whl", hash = "sha256:42419861b43e6962a649068a61f4a4839205a3ef525b858377a960b9e2de6e0d"}, - {file = "mypy-1.8.0-cp38-cp38-win_amd64.whl", hash = "sha256:2b5b6c721bd4aabaadead3a5e6fa85c11c6c795e0c81a7215776ef8afc66de02"}, - {file = "mypy-1.8.0-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:5c1538c38584029352878a0466f03a8ee7547d7bd9f641f57a0f3017a7c905b8"}, - {file = "mypy-1.8.0-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:4ef4be7baf08a203170f29e89d79064463b7fc7a0908b9d0d5114e8009c3a259"}, - {file = "mypy-1.8.0-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:7178def594014aa6c35a8ff411cf37d682f428b3b5617ca79029d8ae72f5402b"}, - {file = "mypy-1.8.0-cp39-cp39-musllinux_1_1_x86_64.whl", hash = "sha256:ab3c84fa13c04aeeeabb2a7f67a25ef5d77ac9d6486ff33ded762ef353aa5592"}, - {file = "mypy-1.8.0-cp39-cp39-win_amd64.whl", hash = "sha256:99b00bc72855812a60d253420d8a2eae839b0afa4938f09f4d2aa9bb4654263a"}, - {file = "mypy-1.8.0-py3-none-any.whl", hash = "sha256:538fd81bb5e430cc1381a443971c0475582ff9f434c16cd46d2c66763ce85d9d"}, - {file = "mypy-1.8.0.tar.gz", hash = "sha256:6ff8b244d7085a0b425b56d327b480c3b29cafbd2eff27316a004f9a7391ae07"}, + {file = "mypy-1.10.0-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:da1cbf08fb3b851ab3b9523a884c232774008267b1f83371ace57f412fe308c2"}, + {file = "mypy-1.10.0-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:12b6bfc1b1a66095ab413160a6e520e1dc076a28f3e22f7fb25ba3b000b4ef99"}, + {file = "mypy-1.10.0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:9e36fb078cce9904c7989b9693e41cb9711e0600139ce3970c6ef814b6ebc2b2"}, + {file = "mypy-1.10.0-cp310-cp310-musllinux_1_1_x86_64.whl", hash = "sha256:2b0695d605ddcd3eb2f736cd8b4e388288c21e7de85001e9f85df9187f2b50f9"}, + {file = "mypy-1.10.0-cp310-cp310-win_amd64.whl", hash = "sha256:cd777b780312ddb135bceb9bc8722a73ec95e042f911cc279e2ec3c667076051"}, + {file = "mypy-1.10.0-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:3be66771aa5c97602f382230165b856c231d1277c511c9a8dd058be4784472e1"}, + {file = "mypy-1.10.0-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:8b2cbaca148d0754a54d44121b5825ae71868c7592a53b7292eeb0f3fdae95ee"}, + {file = "mypy-1.10.0-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:1ec404a7cbe9fc0e92cb0e67f55ce0c025014e26d33e54d9e506a0f2d07fe5de"}, + {file = "mypy-1.10.0-cp311-cp311-musllinux_1_1_x86_64.whl", hash = "sha256:e22e1527dc3d4aa94311d246b59e47f6455b8729f4968765ac1eacf9a4760bc7"}, + {file = "mypy-1.10.0-cp311-cp311-win_amd64.whl", hash = "sha256:a87dbfa85971e8d59c9cc1fcf534efe664d8949e4c0b6b44e8ca548e746a8d53"}, + {file = "mypy-1.10.0-cp312-cp312-macosx_10_9_x86_64.whl", hash = "sha256:a781f6ad4bab20eef8b65174a57e5203f4be627b46291f4589879bf4e257b97b"}, + {file = "mypy-1.10.0-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:b808e12113505b97d9023b0b5e0c0705a90571c6feefc6f215c1df9381256e30"}, + {file = "mypy-1.10.0-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:8f55583b12156c399dce2df7d16f8a5095291354f1e839c252ec6c0611e86e2e"}, + {file = "mypy-1.10.0-cp312-cp312-musllinux_1_1_x86_64.whl", hash = "sha256:4cf18f9d0efa1b16478c4c129eabec36148032575391095f73cae2e722fcf9d5"}, + {file = "mypy-1.10.0-cp312-cp312-win_amd64.whl", hash = "sha256:bc6ac273b23c6b82da3bb25f4136c4fd42665f17f2cd850771cb600bdd2ebeda"}, + {file = "mypy-1.10.0-cp38-cp38-macosx_10_9_x86_64.whl", hash = "sha256:9fd50226364cd2737351c79807775136b0abe084433b55b2e29181a4c3c878c0"}, + {file = "mypy-1.10.0-cp38-cp38-macosx_11_0_arm64.whl", hash = "sha256:f90cff89eea89273727d8783fef5d4a934be2fdca11b47def50cf5d311aff727"}, + {file = "mypy-1.10.0-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:fcfc70599efde5c67862a07a1aaf50e55bce629ace26bb19dc17cece5dd31ca4"}, + {file = "mypy-1.10.0-cp38-cp38-musllinux_1_1_x86_64.whl", hash = "sha256:075cbf81f3e134eadaf247de187bd604748171d6b79736fa9b6c9685b4083061"}, + {file = "mypy-1.10.0-cp38-cp38-win_amd64.whl", hash = "sha256:3f298531bca95ff615b6e9f2fc0333aae27fa48052903a0ac90215021cdcfa4f"}, + {file = "mypy-1.10.0-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:fa7ef5244615a2523b56c034becde4e9e3f9b034854c93639adb667ec9ec2976"}, + {file = "mypy-1.10.0-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:3236a4c8f535a0631f85f5fcdffba71c7feeef76a6002fcba7c1a8e57c8be1ec"}, + {file = "mypy-1.10.0-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:4a2b5cdbb5dd35aa08ea9114436e0d79aceb2f38e32c21684dcf8e24e1e92821"}, + {file = "mypy-1.10.0-cp39-cp39-musllinux_1_1_x86_64.whl", hash = "sha256:92f93b21c0fe73dc00abf91022234c79d793318b8a96faac147cd579c1671746"}, + {file = "mypy-1.10.0-cp39-cp39-win_amd64.whl", hash = "sha256:28d0e038361b45f099cc086d9dd99c15ff14d0188f44ac883010e172ce86c38a"}, + {file = "mypy-1.10.0-py3-none-any.whl", hash = "sha256:f8c083976eb530019175aabadb60921e73b4f45736760826aa1689dda8208aee"}, + {file = "mypy-1.10.0.tar.gz", hash = "sha256:3d087fcbec056c4ee34974da493a826ce316947485cef3901f511848e687c131"}, ] [package.dependencies] @@ -2564,13 +2617,13 @@ files = [ [[package]] name = "nbclient" -version = "0.9.0" +version = "0.10.0" description = "A client library for executing notebooks. Formerly nbconvert's ExecutePreprocessor." optional = false python-versions = ">=3.8.0" files = [ - {file = "nbclient-0.9.0-py3-none-any.whl", hash = "sha256:a3a1ddfb34d4a9d17fc744d655962714a866639acd30130e9be84191cd97cd15"}, - {file = "nbclient-0.9.0.tar.gz", hash = "sha256:4b28c207877cf33ef3a9838cdc7a54c5ceff981194a82eac59d558f05487295e"}, + {file = "nbclient-0.10.0-py3-none-any.whl", hash = "sha256:f13e3529332a1f1f81d82a53210322476a168bb7090a0289c795fe9cc11c9d3f"}, + {file = "nbclient-0.10.0.tar.gz", hash = "sha256:4b3f1b7dba531e498449c4db4f53da339c91d449dc11e9af3a43b4eb5c5abb09"}, ] [package.dependencies] @@ -2582,23 +2635,23 @@ traitlets = ">=5.4" [package.extras] dev = ["pre-commit"] docs = ["autodoc-traits", "mock", "moto", "myst-parser", "nbclient[test]", "sphinx (>=1.7)", "sphinx-book-theme", "sphinxcontrib-spelling"] -test = ["flaky", "ipykernel (>=6.19.3)", "ipython", "ipywidgets", "nbconvert (>=7.0.0)", "pytest (>=7.0)", "pytest-asyncio", "pytest-cov (>=4.0)", "testpath", "xmltodict"] +test = ["flaky", "ipykernel (>=6.19.3)", "ipython", "ipywidgets", "nbconvert (>=7.0.0)", "pytest (>=7.0,<8)", "pytest-asyncio", "pytest-cov (>=4.0)", "testpath", "xmltodict"] [[package]] name = "nbformat" -version = "5.9.2" +version = "5.10.4" description = "The Jupyter Notebook format" optional = false python-versions = ">=3.8" files = [ - {file = "nbformat-5.9.2-py3-none-any.whl", hash = "sha256:1c5172d786a41b82bcfd0c23f9e6b6f072e8fb49c39250219e4acfff1efe89e9"}, - {file = "nbformat-5.9.2.tar.gz", hash = "sha256:5f98b5ba1997dff175e77e0c17d5c10a96eaed2cbd1de3533d1fc35d5e111192"}, + {file = "nbformat-5.10.4-py3-none-any.whl", hash = "sha256:3b48d6c8fbca4b299bf3982ea7db1af21580e4fec269ad087b9e81588891200b"}, + {file = "nbformat-5.10.4.tar.gz", hash = "sha256:322168b14f937a5d11362988ecac2a4952d3d8e3a2cbeb2319584631226d5b3a"}, ] [package.dependencies] -fastjsonschema = "" +fastjsonschema = ">=2.15" jsonschema = ">=2.6" -jupyter-core = "" +jupyter-core = ">=4.12,<5.0.dev0 \|\| >=5.1.dev0" traitlets = ">=5.1" [package.extras] @@ -2662,32 +2715,32 @@ tests = ["Cython", "packaging", "pytest"] [[package]] name = "numba" -version = "0.59.0" +version = "0.59.1" description = "compiling Python code using LLVM" optional = false python-versions = ">=3.9" files = [ - {file = "numba-0.59.0-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:8d061d800473fb8fef76a455221f4ad649a53f5e0f96e3f6c8b8553ee6fa98fa"}, - {file = "numba-0.59.0-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:c086a434e7d3891ce5dfd3d1e7ee8102ac1e733962098578b507864120559ceb"}, - {file = "numba-0.59.0-cp310-cp310-manylinux2014_aarch64.manylinux_2_17_aarch64.whl", hash = "sha256:9e20736bf62e61f8353fb71b0d3a1efba636c7a303d511600fc57648b55823ed"}, - {file = "numba-0.59.0-cp310-cp310-manylinux2014_x86_64.manylinux_2_17_x86_64.whl", hash = "sha256:e86e6786aec31d2002122199486e10bbc0dc40f78d76364cded375912b13614c"}, - {file = "numba-0.59.0-cp310-cp310-win_amd64.whl", hash = "sha256:0307ee91b24500bb7e64d8a109848baf3a3905df48ce142b8ac60aaa406a0400"}, - {file = "numba-0.59.0-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:d540f69a8245fb714419c2209e9af6104e568eb97623adc8943642e61f5d6d8e"}, - {file = "numba-0.59.0-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:1192d6b2906bf3ff72b1d97458724d98860ab86a91abdd4cfd9328432b661e31"}, - {file = "numba-0.59.0-cp311-cp311-manylinux2014_aarch64.manylinux_2_17_aarch64.whl", hash = "sha256:90efb436d3413809fcd15298c6d395cb7d98184350472588356ccf19db9e37c8"}, - {file = "numba-0.59.0-cp311-cp311-manylinux2014_x86_64.manylinux_2_17_x86_64.whl", hash = "sha256:cd3dac45e25d927dcb65d44fb3a973994f5add2b15add13337844afe669dd1ba"}, - {file = "numba-0.59.0-cp311-cp311-win_amd64.whl", hash = "sha256:753dc601a159861808cc3207bad5c17724d3b69552fd22768fddbf302a817a4c"}, - {file = "numba-0.59.0-cp312-cp312-macosx_10_9_x86_64.whl", hash = "sha256:ce62bc0e6dd5264e7ff7f34f41786889fa81a6b860662f824aa7532537a7bee0"}, - {file = "numba-0.59.0-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:8cbef55b73741b5eea2dbaf1b0590b14977ca95a13a07d200b794f8f6833a01c"}, - {file = "numba-0.59.0-cp312-cp312-manylinux2014_aarch64.manylinux_2_17_aarch64.whl", hash = "sha256:70d26ba589f764be45ea8c272caa467dbe882b9676f6749fe6f42678091f5f21"}, - {file = "numba-0.59.0-cp312-cp312-manylinux2014_x86_64.manylinux_2_17_x86_64.whl", hash = "sha256:e125f7d69968118c28ec0eed9fbedd75440e64214b8d2eac033c22c04db48492"}, - {file = "numba-0.59.0-cp312-cp312-win_amd64.whl", hash = "sha256:4981659220b61a03c1e557654027d271f56f3087448967a55c79a0e5f926de62"}, - {file = "numba-0.59.0-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:fe4d7562d1eed754a7511ed7ba962067f198f86909741c5c6e18c4f1819b1f47"}, - {file = "numba-0.59.0-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:6feb1504bb432280f900deaf4b1dadcee68812209500ed3f81c375cbceab24dc"}, - {file = "numba-0.59.0-cp39-cp39-manylinux2014_aarch64.manylinux_2_17_aarch64.whl", hash = "sha256:944faad25ee23ea9dda582bfb0189fb9f4fc232359a80ab2a028b94c14ce2b1d"}, - {file = "numba-0.59.0-cp39-cp39-manylinux2014_x86_64.manylinux_2_17_x86_64.whl", hash = "sha256:5516a469514bfae52a9d7989db4940653a5cbfac106f44cb9c50133b7ad6224b"}, - {file = "numba-0.59.0-cp39-cp39-win_amd64.whl", hash = "sha256:32bd0a41525ec0b1b853da244808f4e5333867df3c43c30c33f89cf20b9c2b63"}, - {file = "numba-0.59.0.tar.gz", hash = "sha256:12b9b064a3e4ad00e2371fc5212ef0396c80f41caec9b5ec391c8b04b6eaf2a8"}, + {file = "numba-0.59.1-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:97385a7f12212c4f4bc28f648720a92514bee79d7063e40ef66c2d30600fd18e"}, + {file = "numba-0.59.1-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:0b77aecf52040de2a1eb1d7e314497b9e56fba17466c80b457b971a25bb1576d"}, + {file = "numba-0.59.1-cp310-cp310-manylinux2014_aarch64.manylinux_2_17_aarch64.whl", hash = "sha256:3476a4f641bfd58f35ead42f4dcaf5f132569c4647c6f1360ccf18ee4cda3990"}, + {file = "numba-0.59.1-cp310-cp310-manylinux2014_x86_64.manylinux_2_17_x86_64.whl", hash = "sha256:525ef3f820931bdae95ee5379c670d5c97289c6520726bc6937a4a7d4230ba24"}, + {file = "numba-0.59.1-cp310-cp310-win_amd64.whl", hash = "sha256:990e395e44d192a12105eca3083b61307db7da10e093972ca285c85bef0963d6"}, + {file = "numba-0.59.1-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:43727e7ad20b3ec23ee4fc642f5b61845c71f75dd2825b3c234390c6d8d64051"}, + {file = "numba-0.59.1-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:411df625372c77959570050e861981e9d196cc1da9aa62c3d6a836b5cc338966"}, + {file = "numba-0.59.1-cp311-cp311-manylinux2014_aarch64.manylinux_2_17_aarch64.whl", hash = "sha256:2801003caa263d1e8497fb84829a7ecfb61738a95f62bc05693fcf1733e978e4"}, + {file = "numba-0.59.1-cp311-cp311-manylinux2014_x86_64.manylinux_2_17_x86_64.whl", hash = "sha256:dd2842fac03be4e5324ebbbd4d2d0c8c0fc6e0df75c09477dd45b288a0777389"}, + {file = "numba-0.59.1-cp311-cp311-win_amd64.whl", hash = "sha256:0594b3dfb369fada1f8bb2e3045cd6c61a564c62e50cf1f86b4666bc721b3450"}, + {file = "numba-0.59.1-cp312-cp312-macosx_10_9_x86_64.whl", hash = "sha256:1cce206a3b92836cdf26ef39d3a3242fec25e07f020cc4feec4c4a865e340569"}, + {file = "numba-0.59.1-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:8c8b4477763cb1fbd86a3be7050500229417bf60867c93e131fd2626edb02238"}, + {file = "numba-0.59.1-cp312-cp312-manylinux2014_aarch64.manylinux_2_17_aarch64.whl", hash = "sha256:7d80bce4ef7e65bf895c29e3889ca75a29ee01da80266a01d34815918e365835"}, + {file = "numba-0.59.1-cp312-cp312-manylinux2014_x86_64.manylinux_2_17_x86_64.whl", hash = "sha256:f7ad1d217773e89a9845886401eaaab0a156a90aa2f179fdc125261fd1105096"}, + {file = "numba-0.59.1-cp312-cp312-win_amd64.whl", hash = "sha256:5bf68f4d69dd3a9f26a9b23548fa23e3bcb9042e2935257b471d2a8d3c424b7f"}, + {file = "numba-0.59.1-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:4e0318ae729de6e5dbe64c75ead1a95eb01fabfe0e2ebed81ebf0344d32db0ae"}, + {file = "numba-0.59.1-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:0f68589740a8c38bb7dc1b938b55d1145244c8353078eea23895d4f82c8b9ec1"}, + {file = "numba-0.59.1-cp39-cp39-manylinux2014_aarch64.manylinux_2_17_aarch64.whl", hash = "sha256:649913a3758891c77c32e2d2a3bcbedf4a69f5fea276d11f9119677c45a422e8"}, + {file = "numba-0.59.1-cp39-cp39-manylinux2014_x86_64.manylinux_2_17_x86_64.whl", hash = "sha256:9712808e4545270291d76b9a264839ac878c5eb7d8b6e02c970dc0ac29bc8187"}, + {file = "numba-0.59.1-cp39-cp39-win_amd64.whl", hash = "sha256:8d51ccd7008a83105ad6a0082b6a2b70f1142dc7cfd76deb8c5a862367eb8c86"}, + {file = "numba-0.59.1.tar.gz", hash = "sha256:76f69132b96028d2774ed20415e8c528a34e3299a40581bae178f0994a2f370b"}, ] [package.dependencies] @@ -2696,13 +2749,13 @@ numpy = ">=1.22,<1.27" [[package]] name = "numbagg" -version = "0.8.0" +version = "0.8.1" description = "Fast N-dimensional aggregation functions with Numba" optional = false python-versions = ">=3.9" files = [ - {file = "numbagg-0.8.0-py3-none-any.whl", hash = "sha256:02bc05396985553ce9e38b5085ebd30bef6abe3a73a838cc255ceb98a86405a0"}, - {file = "numbagg-0.8.0.tar.gz", hash = "sha256:ea44c9f4e914dd614374b64bd5b6b663f67fe65cc29d8bd370e9a565b8c187cc"}, + {file = "numbagg-0.8.1-py3-none-any.whl", hash = "sha256:fc889d7b890bf5e729e64add34852bc70b1447eb53a29fca8b32d345ed792a65"}, + {file = "numbagg-0.8.1.tar.gz", hash = "sha256:3a111361e6062feafd600b1e3c8a8187fd55e955bee735ada843705395fe84d9"}, ] [package.dependencies] @@ -2710,7 +2763,7 @@ numba = "" numpy = "" [package.extras] -dev = ["bottleneck", "hypothesis", "jq", "pandas", "pre-commit", "pytest", "pytest-benchmark", "ruff", "setuptools-scm", "tabulate"] +dev = ["bottleneck", "hypothesis", "jq", "mypy", "pandas", "pre-commit", "pytest", "pytest-benchmark", "ruff", "setuptools-scm", "tabulate"] [[package]] name = "numcodecs" @@ -2799,13 +2852,13 @@ files = [ [[package]] name = "numpy-groupies" -version = "0.10.2" +version = "0.11.1" description = "Optimised tools for group-indexing operations: aggregated sum and more." optional = false python-versions = ">=3.9" files = [ - {file = "numpy-groupies-0.10.2.tar.gz", hash = "sha256:f920c4ded899f5975d94fc63d634e7c89622056bbab8cc98a44d4320a0ae8a12"}, - {file = "numpy_groupies-0.10.2-py3-none-any.whl", hash = "sha256:24a574ecdd58c0d669749b67ec31dab971d27e04037b402bbc4f6bedaa2623bc"}, + {file = "numpy_groupies-0.11.1-py3-none-any.whl", hash = "sha256:c9f2471b0ee96e202fb0a591a551192c19fa3bdddad33df8665a7accbaf3263e"}, + {file = "numpy_groupies-0.11.1.tar.gz", hash = "sha256:b0e9cf7a4d015beca969f1e90463fa90972df8380a722eb96d679f8838d2f6fe"}, ] [package.dependencies] @@ -2849,13 +2902,13 @@ tests = ["pytest", "pytest-cov", "pytest-pep8"] [[package]] name = "packaging" -version = "23.2" +version = "24.0" description = "Core utilities for Python packages" optional = false python-versions = ">=3.7" files = [ - {file = "packaging-23.2-py3-none-any.whl", hash = "sha256:8c491190033a9af7e1d931d0b5dacc2ef47509b34dd0de67ed209b5203fc88c7"}, - {file = "packaging-23.2.tar.gz", hash = "sha256:048fb0e9405036518eaaf48a55953c750c11e1a1b68e0dd1a9d62ed0c092cfc5"}, + {file = "packaging-24.0-py3-none-any.whl", hash = "sha256:2ddfb553fdf02fb784c234c7ba6ccc288296ceabec964ad2eae3777778130bc5"}, + {file = "packaging-24.0.tar.gz", hash = "sha256:eb82c5e3e56209074766e6885bb04b8c38a0c015d0a30036ebe7ece34c9989e9"}, ] [[package]] @@ -2870,47 +2923,44 @@ files = [ [[package]] name = "pandas" -version = "2.2.0" +version = "2.2.2" description = "Powerful data structures for data analysis, time series, and statistics" optional = false python-versions = ">=3.9" files = [ - {file = "pandas-2.2.0-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:8108ee1712bb4fa2c16981fba7e68b3f6ea330277f5ca34fa8d557e986a11670"}, - {file = "pandas-2.2.0-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:736da9ad4033aeab51d067fc3bd69a0ba36f5a60f66a527b3d72e2030e63280a"}, - {file = "pandas-2.2.0-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:38e0b4fc3ddceb56ec8a287313bc22abe17ab0eb184069f08fc6a9352a769b18"}, - {file = "pandas-2.2.0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:20404d2adefe92aed3b38da41d0847a143a09be982a31b85bc7dd565bdba0f4e"}, - {file = "pandas-2.2.0-cp310-cp310-musllinux_1_1_aarch64.whl", hash = "sha256:7ea3ee3f125032bfcade3a4cf85131ed064b4f8dd23e5ce6fa16473e48ebcaf5"}, - {file = "pandas-2.2.0-cp310-cp310-musllinux_1_1_x86_64.whl", hash = "sha256:f9670b3ac00a387620489dfc1bca66db47a787f4e55911f1293063a78b108df1"}, - {file = "pandas-2.2.0-cp310-cp310-win_amd64.whl", hash = "sha256:5a946f210383c7e6d16312d30b238fd508d80d927014f3b33fb5b15c2f895430"}, - {file = "pandas-2.2.0-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:a1b438fa26b208005c997e78672f1aa8138f67002e833312e6230f3e57fa87d5"}, - {file = "pandas-2.2.0-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:8ce2fbc8d9bf303ce54a476116165220a1fedf15985b09656b4b4275300e920b"}, - {file = "pandas-2.2.0-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:2707514a7bec41a4ab81f2ccce8b382961a29fbe9492eab1305bb075b2b1ff4f"}, - {file = "pandas-2.2.0-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:85793cbdc2d5bc32620dc8ffa715423f0c680dacacf55056ba13454a5be5de88"}, - {file = "pandas-2.2.0-cp311-cp311-musllinux_1_1_aarch64.whl", hash = "sha256:cfd6c2491dc821b10c716ad6776e7ab311f7df5d16038d0b7458bc0b67dc10f3"}, - {file = "pandas-2.2.0-cp311-cp311-musllinux_1_1_x86_64.whl", hash = "sha256:a146b9dcacc3123aa2b399df1a284de5f46287a4ab4fbfc237eac98a92ebcb71"}, - {file = "pandas-2.2.0-cp311-cp311-win_amd64.whl", hash = "sha256:fbc1b53c0e1fdf16388c33c3cca160f798d38aea2978004dd3f4d3dec56454c9"}, - {file = "pandas-2.2.0-cp312-cp312-macosx_10_9_x86_64.whl", hash = "sha256:a41d06f308a024981dcaa6c41f2f2be46a6b186b902c94c2674e8cb5c42985bc"}, - {file = "pandas-2.2.0-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:159205c99d7a5ce89ecfc37cb08ed179de7783737cea403b295b5eda8e9c56d1"}, - {file = "pandas-2.2.0-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:eb1e1f3861ea9132b32f2133788f3b14911b68102d562715d71bd0013bc45440"}, - {file = "pandas-2.2.0-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:761cb99b42a69005dec2b08854fb1d4888fdf7b05db23a8c5a099e4b886a2106"}, - {file = "pandas-2.2.0-cp312-cp312-musllinux_1_1_aarch64.whl", hash = "sha256:a20628faaf444da122b2a64b1e5360cde100ee6283ae8effa0d8745153809a2e"}, - {file = "pandas-2.2.0-cp312-cp312-musllinux_1_1_x86_64.whl", hash = "sha256:f5be5d03ea2073627e7111f61b9f1f0d9625dc3c4d8dda72cc827b0c58a1d042"}, - {file = "pandas-2.2.0-cp312-cp312-win_amd64.whl", hash = "sha256:a626795722d893ed6aacb64d2401d017ddc8a2341b49e0384ab9bf7112bdec30"}, - {file = "pandas-2.2.0-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:9f66419d4a41132eb7e9a73dcec9486cf5019f52d90dd35547af11bc58f8637d"}, - {file = "pandas-2.2.0-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:57abcaeda83fb80d447f28ab0cc7b32b13978f6f733875ebd1ed14f8fbc0f4ab"}, - {file = "pandas-2.2.0-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:e60f1f7dba3c2d5ca159e18c46a34e7ca7247a73b5dd1a22b6d59707ed6b899a"}, - {file = "pandas-2.2.0-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:eb61dc8567b798b969bcc1fc964788f5a68214d333cade8319c7ab33e2b5d88a"}, - {file = "pandas-2.2.0-cp39-cp39-musllinux_1_1_aarch64.whl", hash = "sha256:52826b5f4ed658fa2b729264d63f6732b8b29949c7fd234510d57c61dbeadfcd"}, - {file = "pandas-2.2.0-cp39-cp39-musllinux_1_1_x86_64.whl", hash = "sha256:bde2bc699dbd80d7bc7f9cab1e23a95c4375de615860ca089f34e7c64f4a8de7"}, - {file = "pandas-2.2.0-cp39-cp39-win_amd64.whl", hash = "sha256:3de918a754bbf2da2381e8a3dcc45eede8cd7775b047b923f9006d5f876802ae"}, - {file = "pandas-2.2.0.tar.gz", hash = "sha256:30b83f7c3eb217fb4d1b494a57a2fda5444f17834f5df2de6b2ffff68dc3c8e2"}, + {file = "pandas-2.2.2-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:90c6fca2acf139569e74e8781709dccb6fe25940488755716d1d354d6bc58bce"}, + {file = "pandas-2.2.2-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:4abfe0be0d7221be4f12552995e58723c7422c80a659da13ca382697de830c08"}, + {file = "pandas-2.2.2-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:8635c16bf3d99040fdf3ca3db669a7250ddf49c55dc4aa8fe0ae0fa8d6dcc1f0"}, + {file = "pandas-2.2.2-cp310-cp310-musllinux_1_1_aarch64.whl", hash = "sha256:40ae1dffb3967a52203105a077415a86044a2bea011b5f321c6aa64b379a3f51"}, + {file = "pandas-2.2.2-cp310-cp310-musllinux_1_1_x86_64.whl", hash = "sha256:8e5a0b00e1e56a842f922e7fae8ae4077aee4af0acb5ae3622bd4b4c30aedf99"}, + {file = "pandas-2.2.2-cp310-cp310-win_amd64.whl", hash = "sha256:ddf818e4e6c7c6f4f7c8a12709696d193976b591cc7dc50588d3d1a6b5dc8772"}, + {file = "pandas-2.2.2-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:696039430f7a562b74fa45f540aca068ea85fa34c244d0deee539cb6d70aa288"}, + {file = "pandas-2.2.2-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:8e90497254aacacbc4ea6ae5e7a8cd75629d6ad2b30025a4a8b09aa4faf55151"}, + {file = "pandas-2.2.2-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:58b84b91b0b9f4bafac2a0ac55002280c094dfc6402402332c0913a59654ab2b"}, + {file = "pandas-2.2.2-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:6d2123dc9ad6a814bcdea0f099885276b31b24f7edf40f6cdbc0912672e22eee"}, + {file = "pandas-2.2.2-cp311-cp311-musllinux_1_1_aarch64.whl", hash = "sha256:2925720037f06e89af896c70bca73459d7e6a4be96f9de79e2d440bd499fe0db"}, + {file = "pandas-2.2.2-cp311-cp311-musllinux_1_1_x86_64.whl", hash = "sha256:0cace394b6ea70c01ca1595f839cf193df35d1575986e484ad35c4aeae7266c1"}, + {file = "pandas-2.2.2-cp311-cp311-win_amd64.whl", hash = "sha256:873d13d177501a28b2756375d59816c365e42ed8417b41665f346289adc68d24"}, + {file = "pandas-2.2.2-cp312-cp312-macosx_10_9_x86_64.whl", hash = "sha256:9dfde2a0ddef507a631dc9dc4af6a9489d5e2e740e226ad426a05cabfbd7c8ef"}, + {file = "pandas-2.2.2-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:1cb51fe389360f3b5a4d57dbd2848a5f033350336ca3b340d1c53a1fad33bcad"}, + {file = "pandas-2.2.2-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:eee3a87076c0756de40b05c5e9a6069c035ba43e8dd71c379e68cab2c20f16ad"}, + {file = "pandas-2.2.2-cp312-cp312-musllinux_1_1_aarch64.whl", hash = "sha256:3e374f59e440d4ab45ca2fffde54b81ac3834cf5ae2cdfa69c90bc03bde04d76"}, + {file = "pandas-2.2.2-cp312-cp312-musllinux_1_1_x86_64.whl", hash = "sha256:43498c0bdb43d55cb162cdc8c06fac328ccb5d2eabe3cadeb3529ae6f0517c32"}, + {file = "pandas-2.2.2-cp312-cp312-win_amd64.whl", hash = "sha256:d187d355ecec3629624fccb01d104da7d7f391db0311145817525281e2804d23"}, + {file = "pandas-2.2.2-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:0ca6377b8fca51815f382bd0b697a0814c8bda55115678cbc94c30aacbb6eff2"}, + {file = "pandas-2.2.2-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:001910ad31abc7bf06f49dcc903755d2f7f3a9186c0c040b827e522e9cef0863"}, + {file = "pandas-2.2.2-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:66b479b0bd07204e37583c191535505410daa8df638fd8e75ae1b383851fe921"}, + {file = "pandas-2.2.2-cp39-cp39-musllinux_1_1_aarch64.whl", hash = "sha256:a77e9d1c386196879aa5eb712e77461aaee433e54c68cf253053a73b7e49c33a"}, + {file = "pandas-2.2.2-cp39-cp39-musllinux_1_1_x86_64.whl", hash = "sha256:92fd6b027924a7e178ac202cfbe25e53368db90d56872d20ffae94b96c7acc57"}, + {file = "pandas-2.2.2-cp39-cp39-win_amd64.whl", hash = "sha256:640cef9aa381b60e296db324337a554aeeb883ead99dc8f6c18e81a93942f5f4"}, + {file = "pandas-2.2.2.tar.gz", hash = "sha256:9e79019aba43cb4fda9e4d983f8e88ca0373adbb697ae9c6c43093218de28b54"}, ] [package.dependencies] numpy = [ - {version = ">=1.22.4,<2", markers = "python_version < \"3.11\""}, - {version = ">=1.23.2,<2", markers = "python_version == \"3.11\""}, - {version = ">=1.26.0,<2", markers = "python_version >= \"3.12\""}, + {version = ">=1.22.4", markers = "python_version < \"3.11\""}, + {version = ">=1.23.2", markers = "python_version == \"3.11\""}, + {version = ">=1.26.0", markers = "python_version >= \"3.12\""}, ] python-dateutil = ">=2.8.2" pytz = ">=2020.1" @@ -2935,6 +2985,7 @@ parquet = ["pyarrow (>=10.0.1)"] performance = ["bottleneck (>=1.3.6)", "numba (>=0.56.4)", "numexpr (>=2.8.4)"] plot = ["matplotlib (>=3.6.3)"] postgresql = ["SQLAlchemy (>=2.0.0)", "adbc-driver-postgresql (>=0.8.0)", "psycopg2 (>=2.9.6)"] +pyarrow = ["pyarrow (>=10.0.1)"] spss = ["pyreadstat (>=1.2.0)"] sql-other = ["SQLAlchemy (>=2.0.0)", "adbc-driver-postgresql (>=0.8.0)", "adbc-driver-sqlite (>=0.8.0)"] test = ["hypothesis (>=6.46.1)", "pytest (>=7.3.2)", "pytest-xdist (>=2.2.0)"] @@ -2942,13 +2993,13 @@ xml = ["lxml (>=4.9.2)"] [[package]] name = "pandas-stubs" -version = "2.1.4.231227" +version = "2.2.2.240514" description = "Type annotations for pandas" optional = false python-versions = ">=3.9" files = [ - {file = "pandas_stubs-2.1.4.231227-py3-none-any.whl", hash = "sha256:211fc23e6ae87073bdf41dbf362c4a4d85e1e3477cb078dbac3da6c7fdaefba8"}, - {file = "pandas_stubs-2.1.4.231227.tar.gz", hash = "sha256:3ea29ef001e9e44985f5ebde02d4413f94891ef6ec7e5056fb07d125be796c23"}, + {file = "pandas_stubs-2.2.2.240514-py3-none-any.whl", hash = "sha256:5d6f64d45a98bc94152a0f76fa648e598cd2b9ba72302fd34602479f0c391a53"}, + {file = "pandas_stubs-2.2.2.240514.tar.gz", hash = "sha256:85b20da44a62c80eb8389bcf4cbfe31cce1cafa8cca4bf1fc75ec45892e72ce8"}, ] [package.dependencies] @@ -2957,24 +3008,24 @@ types-pytz = ">=2022.1.1" [[package]] name = "panel" -version = "1.3.8" +version = "1.4.3" description = "The powerful data exploration & web app framework for Python." optional = false python-versions = ">=3.9" files = [ - {file = "panel-1.3.8-py2.py3-none-any.whl", hash = "sha256:49bf3931986a0ddf3f7b4bda3c65c6a311d5277524acdb4e0bff69cba6bf5775"}, - {file = "panel-1.3.8.tar.gz", hash = "sha256:809afd2b861747a31d6ddaadbbc7c25b8dab392dc78256f68b759214113c5be3"}, + {file = "panel-1.4.3-py3-none-any.whl", hash = "sha256:888050f5411c98eeaaddb4b67da14aeed63898f55aa08592edb33328bce47d6c"}, + {file = "panel-1.4.3.tar.gz", hash = "sha256:f898f28b92551ec5a2c73fae6c8974c7b5bd1f3b610320c149d5e88eb81a0cc5"}, ] [package.dependencies] bleach = "" -bokeh = ">=3.2.0,<3.4.0" +bokeh = ">=3.4.0,<3.5.0" linkify-it-py = "" markdown = "" markdown-it-py = "" mdit-py-plugins = "" pandas = ">=1.2" -param = ">=2.0.0,<3.0" +param = ">=2.1.0,<3.0" pyviz-comms = ">=2.0.0" requests = "" tqdm = ">=4.48.0" @@ -2982,33 +3033,34 @@ typing-extensions = "" xyzservices = ">=2021.09.1" [package.extras] -all = ["aiohttp", "altair", "anywidget", "channels", "croniter", "datashader", "diskcache", "django (<4)", "fastparquet", "flake8", "folium", "graphviz", "holoviews (>=1.16.0)", "hvplot", "ipyleaflet", "ipympl", "ipython (>=7.0)", "ipyvolume", "ipyvuetify", "ipywidgets", "ipywidgets-bokeh", "jupyter-bokeh (>=3.0.7)", "jupyter-server", "jupyterlab", "lxml", "matplotlib", "nbsite (>=0.8.4)", "nbval", "networkx (>=2.5)", "numba (<0.58)", "numpy", "pandas (<2.1.0)", "pandas (>=1.3)", "parameterized", "pillow", "playwright", "plotly", "plotly (>=4.0)", "pre-commit", "psutil", "pydeck", "pygraphviz", "pyinstrument (>=4.0)", "pytest", "pytest-asyncio (<0.22)", "pytest-cov", "pytest-playwright", "pytest-rerunfailures", "pytest-xdist", "python-graphviz", "pyvista", "reacton", "scikit-image", "scikit-learn", "scipy", "seaborn", "streamz", "twine", "vega-datasets", "vtk", "xarray", "xgboost"] -all-pip = ["aiohttp", "altair", "anywidget", "channels", "croniter", "datashader", "diskcache", "django (<4)", "fastparquet", "flake8", "folium", "graphviz", "holoviews (>=1.16.0)", "hvplot", "ipyleaflet", "ipympl", "ipython (>=7.0)", "ipyvolume", "ipyvuetify", "ipywidgets", "ipywidgets-bokeh", "jupyter-bokeh (>=3.0.7)", "jupyter-server", "jupyterlab", "lxml", "matplotlib", "nbsite (>=0.8.4)", "nbval", "networkx (>=2.5)", "numba (<0.58)", "numpy", "pandas (<2.1.0)", "pandas (>=1.3)", "parameterized", "pillow", "playwright", "plotly", "plotly (>=4.0)", "pre-commit", "psutil", "pydeck", "pyinstrument (>=4.0)", "pytest", "pytest-asyncio (<0.22)", "pytest-cov", "pytest-playwright", "pytest-rerunfailures", "pytest-xdist", "pyvista", "reacton", "scikit-image", "scikit-learn", "scipy", "seaborn", "streamz", "twine", "vega-datasets", "vtk", "xarray", "xgboost"] -build = ["bleach", "bokeh (>=3.3.0,<3.4.0)", "cryptography (<39)", "markdown", "packaging", "param (>=2.0.0)", "pyviz-comms (>=2.0.0)", "requests", "setuptools (>=42)", "tqdm (>=4.48.0)", "urllib3 (<2.0)"] +all = ["aiohttp", "altair", "anywidget", "channels", "croniter", "dask-expr", "datashader", "diskcache", "django (<4)", "fastparquet", "flake8", "folium", "graphviz", "holoviews (>=1.16.0)", "hvplot", "ipyleaflet", "ipympl", "ipython (>=7.0)", "ipyvolume", "ipyvuetify", "ipywidgets", "ipywidgets-bokeh", "jupyter-bokeh (>=3.0.7)", "jupyter-server", "jupyterlab", "lxml", "matplotlib", "nbsite (>=0.8.4)", "nbval", "networkx (>=2.5)", "numba (<0.58)", "numpy", "pandas (<2.1.0)", "pandas (>=1.3)", "parameterized", "pillow", "playwright", "plotly", "plotly (>=4.0)", "pre-commit", "psutil", "pydeck", "pygraphviz", "pyinstrument (>=4.0)", "pytest", "pytest-asyncio", "pytest-cov", "pytest-playwright", "pytest-rerunfailures", "pytest-xdist", "python-graphviz", "pyvista", "reacton", "scikit-image", "scikit-learn", "scipy", "seaborn", "streamz", "textual", "tomli", "twine", "vega-datasets", "vtk", "watchfiles", "xarray", "xgboost"] +all-pip = ["aiohttp", "altair", "anywidget", "channels", "croniter", "dask-expr", "datashader", "diskcache", "django (<4)", "fastparquet", "flake8", "folium", "graphviz", "holoviews (>=1.16.0)", "hvplot", "ipyleaflet", "ipympl", "ipython (>=7.0)", "ipyvolume", "ipyvuetify", "ipywidgets", "ipywidgets-bokeh", "jupyter-bokeh (>=3.0.7)", "jupyter-server", "jupyterlab", "lxml", "matplotlib", "nbsite (>=0.8.4)", "nbval", "networkx (>=2.5)", "numba (<0.58)", "numpy", "pandas (<2.1.0)", "pandas (>=1.3)", "parameterized", "pillow", "playwright", "plotly", "plotly (>=4.0)", "pre-commit", "psutil", "pydeck", "pyinstrument (>=4.0)", "pytest", "pytest-asyncio", "pytest-cov", "pytest-playwright", "pytest-rerunfailures", "pytest-xdist", "pyvista", "reacton", "scikit-image", "scikit-learn", "scipy", "seaborn", "streamz", "textual", "tomli", "twine", "vega-datasets", "vtk", "watchfiles", "xarray", "xgboost"] +build = ["bleach", "bokeh (>=3.4.0,<3.5.0)", "cryptography (<39)", "markdown", "packaging", "param (>=2.0.0)", "pyviz-comms (>=2.0.0)", "requests", "setuptools (>=42)", "tqdm (>=4.48.0)", "urllib3 (<2.0)"] doc = ["holoviews (>=1.16.0)", "jupyterlab", "lxml", "matplotlib", "nbsite (>=0.8.4)", "pandas (<2.1.0)", "pillow", "plotly"] -examples = ["aiohttp", "altair", "channels", "croniter", "datashader", "django (<4)", "fastparquet", "folium", "graphviz", "holoviews (>=1.16.0)", "hvplot", "ipyleaflet", "ipympl", "ipyvolume", "ipyvuetify", "ipywidgets", "ipywidgets-bokeh", "jupyter-bokeh (>=3.0.7)", "networkx (>=2.5)", "plotly (>=4.0)", "pydeck", "pygraphviz", "pyinstrument (>=4.0)", "python-graphviz", "pyvista", "reacton", "scikit-image", "scikit-learn", "seaborn", "streamz", "vega-datasets", "vtk", "xarray", "xgboost"] +examples = ["aiohttp", "altair", "channels", "croniter", "dask-expr", "datashader", "django (<4)", "fastparquet", "folium", "graphviz", "holoviews (>=1.16.0)", "hvplot", "ipyleaflet", "ipympl", "ipyvolume", "ipyvuetify", "ipywidgets", "ipywidgets-bokeh", "jupyter-bokeh (>=3.0.7)", "networkx (>=2.5)", "plotly (>=4.0)", "pydeck", "pygraphviz", "pyinstrument (>=4.0)", "python-graphviz", "pyvista", "reacton", "scikit-image", "scikit-learn", "seaborn", "streamz", "textual", "vega-datasets", "vtk", "xarray", "xgboost"] recommended = ["holoviews (>=1.16.0)", "jupyterlab", "matplotlib", "pillow", "plotly"] -tests = ["altair", "anywidget", "diskcache", "flake8", "folium", "holoviews (>=1.16.0)", "ipympl", "ipython (>=7.0)", "ipyvuetify", "ipywidgets-bokeh", "nbval", "numba (<0.58)", "numpy", "pandas (>=1.3)", "parameterized", "pre-commit", "psutil", "pytest", "pytest-asyncio (<0.22)", "pytest-cov", "pytest-rerunfailures", "pytest-xdist", "reacton", "scipy", "twine"] -tests-core = ["altair", "anywidget", "diskcache", "flake8", "folium", "holoviews (>=1.16.0)", "ipython (>=7.0)", "nbval", "numpy", "pandas (>=1.3)", "parameterized", "pre-commit", "psutil", "pytest", "pytest-asyncio (<0.22)", "pytest-cov", "pytest-rerunfailures", "pytest-xdist", "scipy"] -ui = ["jupyter-server", "playwright", "pytest-playwright"] +tests = ["altair", "anywidget", "diskcache", "flake8", "folium", "holoviews (>=1.16.0)", "ipympl", "ipython (>=7.0)", "ipyvuetify", "ipywidgets-bokeh", "nbval", "numba (<0.58)", "numpy", "pandas (>=1.3)", "parameterized", "pre-commit", "psutil", "pytest", "pytest-asyncio", "pytest-cov", "pytest-rerunfailures", "pytest-xdist", "reacton", "scipy", "textual", "twine", "watchfiles"] +tests-core = ["altair", "anywidget", "diskcache", "flake8", "folium", "holoviews (>=1.16.0)", "ipython (>=7.0)", "nbval", "numpy", "pandas (>=1.3)", "parameterized", "pre-commit", "psutil", "pytest", "pytest-asyncio", "pytest-cov", "pytest-rerunfailures", "pytest-xdist", "scipy", "textual", "watchfiles"] +ui = ["jupyter-server", "playwright", "pytest-playwright", "tomli"] [[package]] name = "papermill" -version = "2.5.0" +version = "2.6.0" description = "Parameterize and run Jupyter and nteract Notebooks" optional = false python-versions = ">=3.8" files = [ - {file = "papermill-2.5.0-py3-none-any.whl", hash = "sha256:c42303afb92e482a60ae1df2577be59a5b7a64c5cd52d37c74c7f74e36085708"}, - {file = "papermill-2.5.0.tar.gz", hash = "sha256:ea7b70c0553f56fe91b0fa9cc5e17012cd699320a8b015373e7870c5e6086c72"}, + {file = "papermill-2.6.0-py3-none-any.whl", hash = "sha256:0f09da6ef709f3f14dde77cb1af052d05b14019189869affff374c9e612f2dd5"}, + {file = "papermill-2.6.0.tar.gz", hash = "sha256:9fe2a91912fd578f391b4cc8d6d105e73124dcd0cde2a43c3c4a1c77ac88ea24"}, ] [package.dependencies] -aiohttp = {version = "3.9.0b0", markers = "python_version == \"3.12\""} +aiohttp = {version = ">=3.9.0", markers = "python_version == \"3.12\""} +ansicolors = "" click = "" entrypoints = "" nbclient = ">=0.2.0" -nbformat = ">=5.1.2" +nbformat = ">=5.2.0" pyyaml = "" requests = "" tenacity = ">=5.0.2" @@ -3018,23 +3070,23 @@ tqdm = ">=4.32.2" all = ["PyGithub (>=1.55)", "azure-datalake-store (>=0.0.30)", "azure-identity (>=1.3.1)", "azure-storage-blob (>=12.1.0)", "black (>=19.3b0)", "boto3", "gcsfs (>=0.2.0)", "pyarrow (>=2.0)", "requests (>=2.21.0)"] azure = ["azure-datalake-store (>=0.0.30)", "azure-identity (>=1.3.1)", "azure-storage-blob (>=12.1.0)", "requests (>=2.21.0)"] black = ["black (>=19.3b0)"] -dev = ["attrs (>=17.4.0)", "azure-datalake-store (>=0.0.30)", "azure-identity (>=1.3.1)", "azure-storage-blob (>=12.1.0)", "black (>=19.3b0)", "boto3", "botocore", "bumpversion", "check-manifest", "codecov", "coverage", "flake8", "gcsfs (>=0.2.0)", "google-compute-engine", "ipython (>=5.0)", "ipywidgets", "moto", "notebook", "pip (>=18.1)", "pre-commit", "pytest (>=4.1)", "pytest-cov (>=2.6.1)", "pytest-env (>=0.6.2)", "pytest-mock (>=1.10)", "recommonmark", "requests (>=2.21.0)", "setuptools (>=38.6.0)", "tox", "twine (>=1.11.0)", "wheel (>=0.31.0)"] -docs = ["PyGithub (>=1.55)", "Sphinx (>=3.5.4)", "azure-datalake-store (>=0.0.30)", "azure-identity (>=1.3.1)", "azure-storage-blob (>=12.1.0)", "black (>=19.3b0)", "boto3", "furo (>=2021.4.11b34)", "gcsfs (>=0.2.0)", "moto (>=2.0.5)", "myst-parser (>=0.13.7)", "pyarrow (>=2.0)", "requests (>=2.21.0)", "sphinx-copybutton (>=0.3.1)"] +dev = ["attrs (>=17.4.0)", "azure-datalake-store (>=0.0.30)", "azure-identity (>=1.3.1)", "azure-storage-blob (>=12.1.0)", "black (>=19.3b0)", "boto3", "botocore", "bumpversion", "check-manifest", "codecov", "coverage", "gcsfs (>=0.2.0)", "google-compute-engine", "ipython (>=5.0)", "ipywidgets", "moto (>=5.0.0,<5.1.0)", "notebook", "pip (>=18.1)", "pre-commit", "pytest (>=4.1)", "pytest-cov (>=2.6.1)", "pytest-env (>=0.6.2)", "pytest-mock (>=1.10)", "recommonmark", "requests (>=2.21.0)", "setuptools (>=38.6.0)", "tox", "twine (>=1.11.0)", "wheel (>=0.31.0)"] +docs = ["PyGithub (>=1.55)", "Sphinx (>=7.2.6)", "azure-datalake-store (>=0.0.30)", "azure-identity (>=1.3.1)", "azure-storage-blob (>=12.1.0)", "black (>=19.3b0)", "boto3", "entrypoints", "furo (>=2023.9.10)", "gcsfs (>=0.2.0)", "moto (>=4.2.8)", "myst-parser (>=2.0.0)", "nbformat", "pyarrow (>=2.0)", "requests (>=2.21.0)", "sphinx-copybutton (>=0.5.2)"] gcs = ["gcsfs (>=0.2.0)"] github = ["PyGithub (>=1.55)"] hdfs = ["pyarrow (>=2.0)"] s3 = ["boto3"] -test = ["attrs (>=17.4.0)", "azure-datalake-store (>=0.0.30)", "azure-identity (>=1.3.1)", "azure-storage-blob (>=12.1.0)", "black (>=19.3b0)", "boto3", "botocore", "bumpversion", "check-manifest", "codecov", "coverage", "flake8", "gcsfs (>=0.2.0)", "google-compute-engine", "ipython (>=5.0)", "ipywidgets", "moto", "notebook", "pip (>=18.1)", "pre-commit", "pytest (>=4.1)", "pytest-cov (>=2.6.1)", "pytest-env (>=0.6.2)", "pytest-mock (>=1.10)", "recommonmark", "requests (>=2.21.0)", "setuptools (>=38.6.0)", "tox", "twine (>=1.11.0)", "wheel (>=0.31.0)"] +test = ["attrs (>=17.4.0)", "azure-datalake-store (>=0.0.30)", "azure-identity (>=1.3.1)", "azure-storage-blob (>=12.1.0)", "black (>=19.3b0)", "boto3", "botocore", "bumpversion", "check-manifest", "codecov", "coverage", "gcsfs (>=0.2.0)", "google-compute-engine", "ipython (>=5.0)", "ipywidgets", "moto (>=5.0.0,<5.1.0)", "notebook", "pip (>=18.1)", "pre-commit", "pytest (>=4.1)", "pytest-cov (>=2.6.1)", "pytest-env (>=0.6.2)", "pytest-mock (>=1.10)", "recommonmark", "requests (>=2.21.0)", "setuptools (>=38.6.0)", "tox", "twine (>=1.11.0)", "wheel (>=0.31.0)"] [[package]] name = "param" -version = "2.0.2" +version = "2.1.0" description = "Make your Python code clearer and more reliable by declaring Parameters." optional = false python-versions = ">=3.8" files = [ - {file = "param-2.0.2-py3-none-any.whl", hash = "sha256:b269fd7397886ec609e544f81035fa52e1950da0e76d20080bfeca3d7a0317ca"}, - {file = "param-2.0.2.tar.gz", hash = "sha256:785845a727a588eb94c7666d80551c7e2bb97d4309d3507beab66f95e57f7527"}, + {file = "param-2.1.0-py3-none-any.whl", hash = "sha256:f31d3745d227347d29b5868c4e4e3077df07463889b91d3bb28e634fde211e1c"}, + {file = "param-2.1.0.tar.gz", hash = "sha256:a7b30b08b547e2b78b02aeba6ed34e3c6a638f8e4824a76a96ffa2d7cf57e71f"}, ] [package.extras] @@ -3044,33 +3096,33 @@ examples = ["aiohttp", "pandas", "panel"] lint = ["flake8", "pre-commit"] tests = ["coverage[toml]", "pytest", "pytest-asyncio"] tests-deser = ["odfpy", "openpyxl", "pyarrow", "tables", "xlrd"] -tests-examples = ["nbval", "param[examples]", "pytest", "pytest-asyncio", "pytest-xdist"] +tests-examples = ["nbval", "param[examples]", "pytest (<8.1)", "pytest-asyncio", "pytest-xdist"] tests-full = ["cloudpickle", "gmpy", "ipython", "jsonschema", "nest-asyncio", "numpy", "pandas", "param[tests-deser]", "param[tests-examples]", "param[tests]"] [[package]] name = "parso" -version = "0.8.3" +version = "0.8.4" description = "A Python Parser" optional = false python-versions = ">=3.6" files = [ - {file = "parso-0.8.3-py2.py3-none-any.whl", hash = "sha256:c001d4636cd3aecdaf33cbb40aebb59b094be2a74c556778ef5576c175e19e75"}, - {file = "parso-0.8.3.tar.gz", hash = "sha256:8c07be290bb59f03588915921e29e8a50002acaf2cdc5fa0e0114f91709fafa0"}, + {file = "parso-0.8.4-py2.py3-none-any.whl", hash = "sha256:a418670a20291dacd2dddc80c377c5c3791378ee1e8d12bffc35420643d43f18"}, + {file = "parso-0.8.4.tar.gz", hash = "sha256:eb3a7b58240fb99099a345571deecc0f9540ea5f4dd2fe14c2a99d6b281ab92d"}, ] [package.extras] -qa = ["flake8 (==3.8.3)", "mypy (==0.782)"] -testing = ["docopt", "pytest (<6.0.0)"] +qa = ["flake8 (==5.0.4)", "mypy (==0.971)", "types-setuptools (==67.2.0.1)"] +testing = ["docopt", "pytest"] [[package]] name = "partd" -version = "1.4.1" +version = "1.4.2" description = "Appendable key-value storage" optional = false -python-versions = ">=3.7" +python-versions = ">=3.9" files = [ - {file = "partd-1.4.1-py3-none-any.whl", hash = "sha256:27e766663d36c161e2827aa3e28541c992f0b9527d3cca047e13fb3acdb989e6"}, - {file = "partd-1.4.1.tar.gz", hash = "sha256:56c25dd49e6fea5727e731203c466c6e092f308d8f0024e199d02f6aa2167f67"}, + {file = "partd-1.4.2-py3-none-any.whl", hash = "sha256:978e4ac767ec4ba5b86c6eaa52e5a2a3bc748a2ca839e8cc798f1cc6ce6efb0f"}, + {file = "partd-1.4.2.tar.gz", hash = "sha256:d022c33afbdc8405c226621b015e8067888173d85f7f5ecebb3cafed9a20f02c"}, ] [package.dependencies] @@ -3078,7 +3130,7 @@ locket = "" toolz = "" [package.extras] -complete = ["blosc", "numpy (>=1.9.0)", "pandas (>=0.19.0)", "pyzmq"] +complete = ["blosc", "numpy (>=1.20.0)", "pandas (>=1.3)", "pyzmq"] [[package]] name = "pathspec" @@ -3107,79 +3159,80 @@ ptyprocess = ">=0.5" [[package]] name = "pillow" -version = "10.2.0" +version = "10.3.0" description = "Python Imaging Library (Fork)" optional = false python-versions = ">=3.8" files = [ - {file = "pillow-10.2.0-cp310-cp310-macosx_10_10_x86_64.whl", hash = "sha256:7823bdd049099efa16e4246bdf15e5a13dbb18a51b68fa06d6c1d4d8b99a796e"}, - {file = "pillow-10.2.0-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:83b2021f2ade7d1ed556bc50a399127d7fb245e725aa0113ebd05cfe88aaf588"}, - {file = "pillow-10.2.0-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:6fad5ff2f13d69b7e74ce5b4ecd12cc0ec530fcee76356cac6742785ff71c452"}, - {file = "pillow-10.2.0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:da2b52b37dad6d9ec64e653637a096905b258d2fc2b984c41ae7d08b938a67e4"}, - {file = "pillow-10.2.0-cp310-cp310-manylinux_2_28_aarch64.whl", hash = "sha256:47c0995fc4e7f79b5cfcab1fc437ff2890b770440f7696a3ba065ee0fd496563"}, - {file = "pillow-10.2.0-cp310-cp310-manylinux_2_28_x86_64.whl", hash = "sha256:322bdf3c9b556e9ffb18f93462e5f749d3444ce081290352c6070d014c93feb2"}, - {file = "pillow-10.2.0-cp310-cp310-musllinux_1_1_aarch64.whl", hash = "sha256:51f1a1bffc50e2e9492e87d8e09a17c5eea8409cda8d3f277eb6edc82813c17c"}, - {file = "pillow-10.2.0-cp310-cp310-musllinux_1_1_x86_64.whl", hash = "sha256:69ffdd6120a4737710a9eee73e1d2e37db89b620f702754b8f6e62594471dee0"}, - {file = "pillow-10.2.0-cp310-cp310-win32.whl", hash = "sha256:c6dafac9e0f2b3c78df97e79af707cdc5ef8e88208d686a4847bab8266870023"}, - {file = "pillow-10.2.0-cp310-cp310-win_amd64.whl", hash = "sha256:aebb6044806f2e16ecc07b2a2637ee1ef67a11840a66752751714a0d924adf72"}, - {file = "pillow-10.2.0-cp310-cp310-win_arm64.whl", hash = "sha256:7049e301399273a0136ff39b84c3678e314f2158f50f517bc50285fb5ec847ad"}, - {file = "pillow-10.2.0-cp311-cp311-macosx_10_10_x86_64.whl", hash = "sha256:35bb52c37f256f662abdfa49d2dfa6ce5d93281d323a9af377a120e89a9eafb5"}, - {file = "pillow-10.2.0-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:9c23f307202661071d94b5e384e1e1dc7dfb972a28a2310e4ee16103e66ddb67"}, - {file = "pillow-10.2.0-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:773efe0603db30c281521a7c0214cad7836c03b8ccff897beae9b47c0b657d61"}, - {file = "pillow-10.2.0-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:11fa2e5984b949b0dd6d7a94d967743d87c577ff0b83392f17cb3990d0d2fd6e"}, - {file = "pillow-10.2.0-cp311-cp311-manylinux_2_28_aarch64.whl", hash = "sha256:716d30ed977be8b37d3ef185fecb9e5a1d62d110dfbdcd1e2a122ab46fddb03f"}, - {file = "pillow-10.2.0-cp311-cp311-manylinux_2_28_x86_64.whl", hash = "sha256:a086c2af425c5f62a65e12fbf385f7c9fcb8f107d0849dba5839461a129cf311"}, - {file = "pillow-10.2.0-cp311-cp311-musllinux_1_1_aarch64.whl", hash = "sha256:c8de2789052ed501dd829e9cae8d3dcce7acb4777ea4a479c14521c942d395b1"}, - {file = "pillow-10.2.0-cp311-cp311-musllinux_1_1_x86_64.whl", hash = "sha256:609448742444d9290fd687940ac0b57fb35e6fd92bdb65386e08e99af60bf757"}, - {file = "pillow-10.2.0-cp311-cp311-win32.whl", hash = "sha256:823ef7a27cf86df6597fa0671066c1b596f69eba53efa3d1e1cb8b30f3533068"}, - {file = "pillow-10.2.0-cp311-cp311-win_amd64.whl", hash = "sha256:1da3b2703afd040cf65ec97efea81cfba59cdbed9c11d8efc5ab09df9509fc56"}, - {file = "pillow-10.2.0-cp311-cp311-win_arm64.whl", hash = "sha256:edca80cbfb2b68d7b56930b84a0e45ae1694aeba0541f798e908a49d66b837f1"}, - {file = "pillow-10.2.0-cp312-cp312-macosx_10_10_x86_64.whl", hash = "sha256:1b5e1b74d1bd1b78bc3477528919414874748dd363e6272efd5abf7654e68bef"}, - {file = "pillow-10.2.0-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:0eae2073305f451d8ecacb5474997c08569fb4eb4ac231ffa4ad7d342fdc25ac"}, - {file = "pillow-10.2.0-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:b7c2286c23cd350b80d2fc9d424fc797575fb16f854b831d16fd47ceec078f2c"}, - {file = "pillow-10.2.0-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:1e23412b5c41e58cec602f1135c57dfcf15482013ce6e5f093a86db69646a5aa"}, - {file = "pillow-10.2.0-cp312-cp312-manylinux_2_28_aarch64.whl", hash = "sha256:52a50aa3fb3acb9cf7213573ef55d31d6eca37f5709c69e6858fe3bc04a5c2a2"}, - {file = "pillow-10.2.0-cp312-cp312-manylinux_2_28_x86_64.whl", hash = "sha256:127cee571038f252a552760076407f9cff79761c3d436a12af6000cd182a9d04"}, - {file = "pillow-10.2.0-cp312-cp312-musllinux_1_1_aarch64.whl", hash = "sha256:8d12251f02d69d8310b046e82572ed486685c38f02176bd08baf216746eb947f"}, - {file = "pillow-10.2.0-cp312-cp312-musllinux_1_1_x86_64.whl", hash = "sha256:54f1852cd531aa981bc0965b7d609f5f6cc8ce8c41b1139f6ed6b3c54ab82bfb"}, - {file = "pillow-10.2.0-cp312-cp312-win32.whl", hash = "sha256:257d8788df5ca62c980314053197f4d46eefedf4e6175bc9412f14412ec4ea2f"}, - {file = "pillow-10.2.0-cp312-cp312-win_amd64.whl", hash = "sha256:154e939c5f0053a383de4fd3d3da48d9427a7e985f58af8e94d0b3c9fcfcf4f9"}, - {file = "pillow-10.2.0-cp312-cp312-win_arm64.whl", hash = "sha256:f379abd2f1e3dddb2b61bc67977a6b5a0a3f7485538bcc6f39ec76163891ee48"}, - {file = "pillow-10.2.0-cp38-cp38-macosx_10_10_x86_64.whl", hash = "sha256:8373c6c251f7ef8bda6675dd6d2b3a0fcc31edf1201266b5cf608b62a37407f9"}, - {file = "pillow-10.2.0-cp38-cp38-macosx_11_0_arm64.whl", hash = "sha256:870ea1ada0899fd0b79643990809323b389d4d1d46c192f97342eeb6ee0b8483"}, - {file = "pillow-10.2.0-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:b4b6b1e20608493548b1f32bce8cca185bf0480983890403d3b8753e44077129"}, - {file = "pillow-10.2.0-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:3031709084b6e7852d00479fd1d310b07d0ba82765f973b543c8af5061cf990e"}, - {file = "pillow-10.2.0-cp38-cp38-manylinux_2_28_aarch64.whl", hash = "sha256:3ff074fc97dd4e80543a3e91f69d58889baf2002b6be64347ea8cf5533188213"}, - {file = "pillow-10.2.0-cp38-cp38-manylinux_2_28_x86_64.whl", hash = "sha256:cb4c38abeef13c61d6916f264d4845fab99d7b711be96c326b84df9e3e0ff62d"}, - {file = "pillow-10.2.0-cp38-cp38-musllinux_1_1_aarch64.whl", hash = "sha256:b1b3020d90c2d8e1dae29cf3ce54f8094f7938460fb5ce8bc5c01450b01fbaf6"}, - {file = "pillow-10.2.0-cp38-cp38-musllinux_1_1_x86_64.whl", hash = "sha256:170aeb00224ab3dc54230c797f8404507240dd868cf52066f66a41b33169bdbe"}, - {file = "pillow-10.2.0-cp38-cp38-win32.whl", hash = "sha256:c4225f5220f46b2fde568c74fca27ae9771536c2e29d7c04f4fb62c83275ac4e"}, - {file = "pillow-10.2.0-cp38-cp38-win_amd64.whl", hash = "sha256:0689b5a8c5288bc0504d9fcee48f61a6a586b9b98514d7d29b840143d6734f39"}, - {file = "pillow-10.2.0-cp39-cp39-macosx_10_10_x86_64.whl", hash = "sha256:b792a349405fbc0163190fde0dc7b3fef3c9268292586cf5645598b48e63dc67"}, - {file = "pillow-10.2.0-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:c570f24be1e468e3f0ce7ef56a89a60f0e05b30a3669a459e419c6eac2c35364"}, - {file = "pillow-10.2.0-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:d8ecd059fdaf60c1963c58ceb8997b32e9dc1b911f5da5307aab614f1ce5c2fb"}, - {file = "pillow-10.2.0-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:c365fd1703040de1ec284b176d6af5abe21b427cb3a5ff68e0759e1e313a5e7e"}, - {file = "pillow-10.2.0-cp39-cp39-manylinux_2_28_aarch64.whl", hash = "sha256:70c61d4c475835a19b3a5aa42492409878bbca7438554a1f89d20d58a7c75c01"}, - {file = "pillow-10.2.0-cp39-cp39-manylinux_2_28_x86_64.whl", hash = "sha256:b6f491cdf80ae540738859d9766783e3b3c8e5bd37f5dfa0b76abdecc5081f13"}, - {file = "pillow-10.2.0-cp39-cp39-musllinux_1_1_aarch64.whl", hash = "sha256:9d189550615b4948f45252d7f005e53c2040cea1af5b60d6f79491a6e147eef7"}, - {file = "pillow-10.2.0-cp39-cp39-musllinux_1_1_x86_64.whl", hash = "sha256:49d9ba1ed0ef3e061088cd1e7538a0759aab559e2e0a80a36f9fd9d8c0c21591"}, - {file = "pillow-10.2.0-cp39-cp39-win32.whl", hash = "sha256:babf5acfede515f176833ed6028754cbcd0d206f7f614ea3447d67c33be12516"}, - {file = "pillow-10.2.0-cp39-cp39-win_amd64.whl", hash = "sha256:0304004f8067386b477d20a518b50f3fa658a28d44e4116970abfcd94fac34a8"}, - {file = "pillow-10.2.0-cp39-cp39-win_arm64.whl", hash = "sha256:0fb3e7fc88a14eacd303e90481ad983fd5b69c761e9e6ef94c983f91025da869"}, - {file = "pillow-10.2.0-pp310-pypy310_pp73-macosx_10_10_x86_64.whl", hash = "sha256:322209c642aabdd6207517e9739c704dc9f9db943015535783239022002f054a"}, - {file = "pillow-10.2.0-pp310-pypy310_pp73-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:3eedd52442c0a5ff4f887fab0c1c0bb164d8635b32c894bc1faf4c618dd89df2"}, - {file = "pillow-10.2.0-pp310-pypy310_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:cb28c753fd5eb3dd859b4ee95de66cc62af91bcff5db5f2571d32a520baf1f04"}, - {file = "pillow-10.2.0-pp310-pypy310_pp73-manylinux_2_28_aarch64.whl", hash = "sha256:33870dc4653c5017bf4c8873e5488d8f8d5f8935e2f1fb9a2208c47cdd66efd2"}, - {file = "pillow-10.2.0-pp310-pypy310_pp73-manylinux_2_28_x86_64.whl", hash = "sha256:3c31822339516fb3c82d03f30e22b1d038da87ef27b6a78c9549888f8ceda39a"}, - {file = "pillow-10.2.0-pp310-pypy310_pp73-win_amd64.whl", hash = "sha256:a2b56ba36e05f973d450582fb015594aaa78834fefe8dfb8fcd79b93e64ba4c6"}, - {file = "pillow-10.2.0-pp38-pypy38_pp73-win_amd64.whl", hash = "sha256:d8e6aeb9201e655354b3ad049cb77d19813ad4ece0df1249d3c793de3774f8c7"}, - {file = "pillow-10.2.0-pp39-pypy39_pp73-macosx_10_10_x86_64.whl", hash = "sha256:2247178effb34a77c11c0e8ac355c7a741ceca0a732b27bf11e747bbc950722f"}, - {file = "pillow-10.2.0-pp39-pypy39_pp73-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:15587643b9e5eb26c48e49a7b33659790d28f190fc514a322d55da2fb5c2950e"}, - {file = "pillow-10.2.0-pp39-pypy39_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:753cd8f2086b2b80180d9b3010dd4ed147efc167c90d3bf593fe2af21265e5a5"}, - {file = "pillow-10.2.0-pp39-pypy39_pp73-manylinux_2_28_aarch64.whl", hash = "sha256:7c8f97e8e7a9009bcacbe3766a36175056c12f9a44e6e6f2d5caad06dcfbf03b"}, - {file = "pillow-10.2.0-pp39-pypy39_pp73-manylinux_2_28_x86_64.whl", hash = "sha256:d1b35bcd6c5543b9cb547dee3150c93008f8dd0f1fef78fc0cd2b141c5baf58a"}, - {file = "pillow-10.2.0-pp39-pypy39_pp73-win_amd64.whl", hash = "sha256:fe4c15f6c9285dc54ce6553a3ce908ed37c8f3825b5a51a15c91442bb955b868"}, - {file = "pillow-10.2.0.tar.gz", hash = "sha256:e87f0b2c78157e12d7686b27d63c070fd65d994e8ddae6f328e0dcf4a0cd007e"}, + {file = "pillow-10.3.0-cp310-cp310-macosx_10_10_x86_64.whl", hash = "sha256:90b9e29824800e90c84e4022dd5cc16eb2d9605ee13f05d47641eb183cd73d45"}, + {file = "pillow-10.3.0-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:a2c405445c79c3f5a124573a051062300936b0281fee57637e706453e452746c"}, + {file = "pillow-10.3.0-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:78618cdbccaa74d3f88d0ad6cb8ac3007f1a6fa5c6f19af64b55ca170bfa1edf"}, + {file = "pillow-10.3.0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:261ddb7ca91fcf71757979534fb4c128448b5b4c55cb6152d280312062f69599"}, + {file = "pillow-10.3.0-cp310-cp310-manylinux_2_28_aarch64.whl", hash = "sha256:ce49c67f4ea0609933d01c0731b34b8695a7a748d6c8d186f95e7d085d2fe475"}, + {file = "pillow-10.3.0-cp310-cp310-manylinux_2_28_x86_64.whl", hash = "sha256:b14f16f94cbc61215115b9b1236f9c18403c15dd3c52cf629072afa9d54c1cbf"}, + {file = "pillow-10.3.0-cp310-cp310-musllinux_1_1_aarch64.whl", hash = "sha256:d33891be6df59d93df4d846640f0e46f1a807339f09e79a8040bc887bdcd7ed3"}, + {file = "pillow-10.3.0-cp310-cp310-musllinux_1_1_x86_64.whl", hash = "sha256:b50811d664d392f02f7761621303eba9d1b056fb1868c8cdf4231279645c25f5"}, + {file = "pillow-10.3.0-cp310-cp310-win32.whl", hash = "sha256:ca2870d5d10d8726a27396d3ca4cf7976cec0f3cb706debe88e3a5bd4610f7d2"}, + {file = "pillow-10.3.0-cp310-cp310-win_amd64.whl", hash = "sha256:f0d0591a0aeaefdaf9a5e545e7485f89910c977087e7de2b6c388aec32011e9f"}, + {file = "pillow-10.3.0-cp310-cp310-win_arm64.whl", hash = "sha256:ccce24b7ad89adb5a1e34a6ba96ac2530046763912806ad4c247356a8f33a67b"}, + {file = "pillow-10.3.0-cp311-cp311-macosx_10_10_x86_64.whl", hash = "sha256:5f77cf66e96ae734717d341c145c5949c63180842a545c47a0ce7ae52ca83795"}, + {file = "pillow-10.3.0-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:e4b878386c4bf293578b48fc570b84ecfe477d3b77ba39a6e87150af77f40c57"}, + {file = "pillow-10.3.0-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:fdcbb4068117dfd9ce0138d068ac512843c52295ed996ae6dd1faf537b6dbc27"}, + {file = "pillow-10.3.0-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:9797a6c8fe16f25749b371c02e2ade0efb51155e767a971c61734b1bf6293994"}, + {file = "pillow-10.3.0-cp311-cp311-manylinux_2_28_aarch64.whl", hash = "sha256:9e91179a242bbc99be65e139e30690e081fe6cb91a8e77faf4c409653de39451"}, + {file = "pillow-10.3.0-cp311-cp311-manylinux_2_28_x86_64.whl", hash = "sha256:1b87bd9d81d179bd8ab871603bd80d8645729939f90b71e62914e816a76fc6bd"}, + {file = "pillow-10.3.0-cp311-cp311-musllinux_1_1_aarch64.whl", hash = "sha256:81d09caa7b27ef4e61cb7d8fbf1714f5aec1c6b6c5270ee53504981e6e9121ad"}, + {file = "pillow-10.3.0-cp311-cp311-musllinux_1_1_x86_64.whl", hash = "sha256:048ad577748b9fa4a99a0548c64f2cb8d672d5bf2e643a739ac8faff1164238c"}, + {file = "pillow-10.3.0-cp311-cp311-win32.whl", hash = "sha256:7161ec49ef0800947dc5570f86568a7bb36fa97dd09e9827dc02b718c5643f09"}, + {file = "pillow-10.3.0-cp311-cp311-win_amd64.whl", hash = "sha256:8eb0908e954d093b02a543dc963984d6e99ad2b5e36503d8a0aaf040505f747d"}, + {file = "pillow-10.3.0-cp311-cp311-win_arm64.whl", hash = "sha256:4e6f7d1c414191c1199f8996d3f2282b9ebea0945693fb67392c75a3a320941f"}, + {file = "pillow-10.3.0-cp312-cp312-macosx_10_10_x86_64.whl", hash = "sha256:e46f38133e5a060d46bd630faa4d9fa0202377495df1f068a8299fd78c84de84"}, + {file = "pillow-10.3.0-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:50b8eae8f7334ec826d6eeffaeeb00e36b5e24aa0b9df322c247539714c6df19"}, + {file = "pillow-10.3.0-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:9d3bea1c75f8c53ee4d505c3e67d8c158ad4df0d83170605b50b64025917f338"}, + {file = "pillow-10.3.0-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:19aeb96d43902f0a783946a0a87dbdad5c84c936025b8419da0a0cd7724356b1"}, + {file = "pillow-10.3.0-cp312-cp312-manylinux_2_28_aarch64.whl", hash = "sha256:74d28c17412d9caa1066f7a31df8403ec23d5268ba46cd0ad2c50fb82ae40462"}, + {file = "pillow-10.3.0-cp312-cp312-manylinux_2_28_x86_64.whl", hash = "sha256:ff61bfd9253c3915e6d41c651d5f962da23eda633cf02262990094a18a55371a"}, + {file = "pillow-10.3.0-cp312-cp312-musllinux_1_1_aarch64.whl", hash = "sha256:d886f5d353333b4771d21267c7ecc75b710f1a73d72d03ca06df49b09015a9ef"}, + {file = "pillow-10.3.0-cp312-cp312-musllinux_1_1_x86_64.whl", hash = "sha256:4b5ec25d8b17217d635f8935dbc1b9aa5907962fae29dff220f2659487891cd3"}, + {file = "pillow-10.3.0-cp312-cp312-win32.whl", hash = "sha256:51243f1ed5161b9945011a7360e997729776f6e5d7005ba0c6879267d4c5139d"}, + {file = "pillow-10.3.0-cp312-cp312-win_amd64.whl", hash = "sha256:412444afb8c4c7a6cc11a47dade32982439925537e483be7c0ae0cf96c4f6a0b"}, + {file = "pillow-10.3.0-cp312-cp312-win_arm64.whl", hash = "sha256:798232c92e7665fe82ac085f9d8e8ca98826f8e27859d9a96b41d519ecd2e49a"}, + {file = "pillow-10.3.0-cp38-cp38-macosx_10_10_x86_64.whl", hash = "sha256:4eaa22f0d22b1a7e93ff0a596d57fdede2e550aecffb5a1ef1106aaece48e96b"}, + {file = "pillow-10.3.0-cp38-cp38-macosx_11_0_arm64.whl", hash = "sha256:cd5e14fbf22a87321b24c88669aad3a51ec052eb145315b3da3b7e3cc105b9a2"}, + {file = "pillow-10.3.0-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:1530e8f3a4b965eb6a7785cf17a426c779333eb62c9a7d1bbcf3ffd5bf77a4aa"}, + {file = "pillow-10.3.0-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:5d512aafa1d32efa014fa041d38868fda85028e3f930a96f85d49c7d8ddc0383"}, + {file = "pillow-10.3.0-cp38-cp38-manylinux_2_28_aarch64.whl", hash = "sha256:339894035d0ede518b16073bdc2feef4c991ee991a29774b33e515f1d308e08d"}, + {file = "pillow-10.3.0-cp38-cp38-manylinux_2_28_x86_64.whl", hash = "sha256:aa7e402ce11f0885305bfb6afb3434b3cd8f53b563ac065452d9d5654c7b86fd"}, + {file = "pillow-10.3.0-cp38-cp38-musllinux_1_1_aarch64.whl", hash = "sha256:0ea2a783a2bdf2a561808fe4a7a12e9aa3799b701ba305de596bc48b8bdfce9d"}, + {file = "pillow-10.3.0-cp38-cp38-musllinux_1_1_x86_64.whl", hash = "sha256:c78e1b00a87ce43bb37642c0812315b411e856a905d58d597750eb79802aaaa3"}, + {file = "pillow-10.3.0-cp38-cp38-win32.whl", hash = "sha256:72d622d262e463dfb7595202d229f5f3ab4b852289a1cd09650362db23b9eb0b"}, + {file = "pillow-10.3.0-cp38-cp38-win_amd64.whl", hash = "sha256:2034f6759a722da3a3dbd91a81148cf884e91d1b747992ca288ab88c1de15999"}, + {file = "pillow-10.3.0-cp39-cp39-macosx_10_10_x86_64.whl", hash = "sha256:2ed854e716a89b1afcedea551cd85f2eb2a807613752ab997b9974aaa0d56936"}, + {file = "pillow-10.3.0-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:dc1a390a82755a8c26c9964d457d4c9cbec5405896cba94cf51f36ea0d855002"}, + {file = "pillow-10.3.0-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:4203efca580f0dd6f882ca211f923168548f7ba334c189e9eab1178ab840bf60"}, + {file = "pillow-10.3.0-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:3102045a10945173d38336f6e71a8dc71bcaeed55c3123ad4af82c52807b9375"}, + {file = "pillow-10.3.0-cp39-cp39-manylinux_2_28_aarch64.whl", hash = "sha256:6fb1b30043271ec92dc65f6d9f0b7a830c210b8a96423074b15c7bc999975f57"}, + {file = "pillow-10.3.0-cp39-cp39-manylinux_2_28_x86_64.whl", hash = "sha256:1dfc94946bc60ea375cc39cff0b8da6c7e5f8fcdc1d946beb8da5c216156ddd8"}, + {file = "pillow-10.3.0-cp39-cp39-musllinux_1_1_aarch64.whl", hash = "sha256:b09b86b27a064c9624d0a6c54da01c1beaf5b6cadfa609cf63789b1d08a797b9"}, + {file = "pillow-10.3.0-cp39-cp39-musllinux_1_1_x86_64.whl", hash = "sha256:d3b2348a78bc939b4fed6552abfd2e7988e0f81443ef3911a4b8498ca084f6eb"}, + {file = "pillow-10.3.0-cp39-cp39-win32.whl", hash = "sha256:45ebc7b45406febf07fef35d856f0293a92e7417ae7933207e90bf9090b70572"}, + {file = "pillow-10.3.0-cp39-cp39-win_amd64.whl", hash = "sha256:0ba26351b137ca4e0db0342d5d00d2e355eb29372c05afd544ebf47c0956ffeb"}, + {file = "pillow-10.3.0-cp39-cp39-win_arm64.whl", hash = "sha256:50fd3f6b26e3441ae07b7c979309638b72abc1a25da31a81a7fbd9495713ef4f"}, + {file = "pillow-10.3.0-pp310-pypy310_pp73-macosx_10_10_x86_64.whl", hash = "sha256:6b02471b72526ab8a18c39cb7967b72d194ec53c1fd0a70b050565a0f366d355"}, + {file = "pillow-10.3.0-pp310-pypy310_pp73-macosx_11_0_arm64.whl", hash = "sha256:8ab74c06ffdab957d7670c2a5a6e1a70181cd10b727cd788c4dd9005b6a8acd9"}, + {file = "pillow-10.3.0-pp310-pypy310_pp73-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:048eeade4c33fdf7e08da40ef402e748df113fd0b4584e32c4af74fe78baaeb2"}, + {file = "pillow-10.3.0-pp310-pypy310_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:9e2ec1e921fd07c7cda7962bad283acc2f2a9ccc1b971ee4b216b75fad6f0463"}, + {file = "pillow-10.3.0-pp310-pypy310_pp73-manylinux_2_28_aarch64.whl", hash = "sha256:4c8e73e99da7db1b4cad7f8d682cf6abad7844da39834c288fbfa394a47bbced"}, + {file = "pillow-10.3.0-pp310-pypy310_pp73-manylinux_2_28_x86_64.whl", hash = "sha256:16563993329b79513f59142a6b02055e10514c1a8e86dca8b48a893e33cf91e3"}, + {file = "pillow-10.3.0-pp310-pypy310_pp73-win_amd64.whl", hash = "sha256:dd78700f5788ae180b5ee8902c6aea5a5726bac7c364b202b4b3e3ba2d293170"}, + {file = "pillow-10.3.0-pp39-pypy39_pp73-macosx_10_10_x86_64.whl", hash = "sha256:aff76a55a8aa8364d25400a210a65ff59d0168e0b4285ba6bf2bd83cf675ba32"}, + {file = "pillow-10.3.0-pp39-pypy39_pp73-macosx_11_0_arm64.whl", hash = "sha256:b7bc2176354defba3edc2b9a777744462da2f8e921fbaf61e52acb95bafa9828"}, + {file = "pillow-10.3.0-pp39-pypy39_pp73-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:793b4e24db2e8742ca6423d3fde8396db336698c55cd34b660663ee9e45ed37f"}, + {file = "pillow-10.3.0-pp39-pypy39_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:d93480005693d247f8346bc8ee28c72a2191bdf1f6b5db469c096c0c867ac015"}, + {file = "pillow-10.3.0-pp39-pypy39_pp73-manylinux_2_28_aarch64.whl", hash = "sha256:c83341b89884e2b2e55886e8fbbf37c3fa5efd6c8907124aeb72f285ae5696e5"}, + {file = "pillow-10.3.0-pp39-pypy39_pp73-manylinux_2_28_x86_64.whl", hash = "sha256:1a1d1915db1a4fdb2754b9de292642a39a7fb28f1736699527bb649484fb966a"}, + {file = "pillow-10.3.0-pp39-pypy39_pp73-win_amd64.whl", hash = "sha256:a0eaa93d054751ee9964afa21c06247779b90440ca41d184aeb5d410f20ff591"}, + {file = "pillow-10.3.0.tar.gz", hash = "sha256:9d2455fbf44c914840c793e89aa82d0e1763a14253a000743719ae5946814b2d"}, ] [package.extras] @@ -3192,28 +3245,29 @@ xmp = ["defusedxml"] [[package]] name = "platformdirs" -version = "4.2.0" -description = "A small Python package for determining appropriate platform-specific dirs, e.g. a \"user data dir\"." +version = "4.2.2" +description = "A small Python package for determining appropriate platform-specific dirs, e.g. a `user data dir`." optional = false python-versions = ">=3.8" files = [ - {file = "platformdirs-4.2.0-py3-none-any.whl", hash = "sha256:0614df2a2f37e1a662acbd8e2b25b92ccf8632929bc6d43467e17fe89c75e068"}, - {file = "platformdirs-4.2.0.tar.gz", hash = "sha256:ef0cc731df711022c174543cb70a9b5bd22e5a9337c8624ef2c2ceb8ddad8768"}, + {file = "platformdirs-4.2.2-py3-none-any.whl", hash = "sha256:2d7a1657e36a80ea911db832a8a6ece5ee53d8de21edd5cc5879af6530b1bfee"}, + {file = "platformdirs-4.2.2.tar.gz", hash = "sha256:38b7b51f512eed9e84a22788b4bce1de17c0adb134d6becb09836e37d8654cd3"}, ] [package.extras] docs = ["furo (>=2023.9.10)", "proselint (>=0.13)", "sphinx (>=7.2.6)", "sphinx-autodoc-typehints (>=1.25.2)"] test = ["appdirs (==1.4.4)", "covdefaults (>=2.3)", "pytest (>=7.4.3)", "pytest-cov (>=4.1)", "pytest-mock (>=3.12)"] +type = ["mypy (>=1.8)"] [[package]] name = "pluggy" -version = "1.4.0" +version = "1.5.0" description = "plugin and hook calling mechanisms for python" optional = false python-versions = ">=3.8" files = [ - {file = "pluggy-1.4.0-py3-none-any.whl", hash = "sha256:7db9f7b503d67d1c5b95f59773ebb58a8c1c288129a88665838012cfb07b8981"}, - {file = "pluggy-1.4.0.tar.gz", hash = "sha256:8c85c2876142a764e5b7548e7d9a0e0ddb46f5185161049a79b7e974454223be"}, + {file = "pluggy-1.5.0-py3-none-any.whl", hash = "sha256:44e1ad92c8ca002de6377e165f3e0f1be63266ab4d554740532335b9d75ea669"}, + {file = "pluggy-1.5.0.tar.gz", hash = "sha256:2cffa88e94fdc978c4c574f15f9e59b7f4201d439195c3715ca9e2486f1d0cf1"}, ] [package.extras] @@ -3222,13 +3276,13 @@ testing = ["pytest", "pytest-benchmark"] [[package]] name = "pooch" -version = "1.8.0" +version = "1.8.1" description = "\"Pooch manages your Python library's sample data files: it automatically downloads and stores them in a local directory, with support for versioning and corruption checks.\"" optional = false python-versions = ">=3.7" files = [ - {file = "pooch-1.8.0-py3-none-any.whl", hash = "sha256:1bfba436d9e2ad5199ccad3583cca8c241b8736b5bb23fe67c213d52650dbb66"}, - {file = "pooch-1.8.0.tar.gz", hash = "sha256:f59981fd5b9b5d032dcde8f4a11eaa492c2ac6343fae3596a2fdae35fc54b0a0"}, + {file = "pooch-1.8.1-py3-none-any.whl", hash = "sha256:6b56611ac320c239faece1ac51a60b25796792599ce5c0b1bb87bf01df55e0a9"}, + {file = "pooch-1.8.1.tar.gz", hash = "sha256:27ef63097dd9a6e4f9d2694f5cfbf2f0a5defa44fccafec08d601e731d746270"}, ] [package.dependencies] @@ -3310,51 +3364,51 @@ tests = ["pytest"] [[package]] name = "pyarrow" -version = "15.0.0" +version = "16.1.0" description = "Python library for Apache Arrow" optional = false python-versions = ">=3.8" files = [ - {file = "pyarrow-15.0.0-cp310-cp310-macosx_10_15_x86_64.whl", hash = "sha256:0a524532fd6dd482edaa563b686d754c70417c2f72742a8c990b322d4c03a15d"}, - {file = "pyarrow-15.0.0-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:60a6bdb314affa9c2e0d5dddf3d9cbb9ef4a8dddaa68669975287d47ece67642"}, - {file = "pyarrow-15.0.0-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:66958fd1771a4d4b754cd385835e66a3ef6b12611e001d4e5edfcef5f30391e2"}, - {file = "pyarrow-15.0.0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:1f500956a49aadd907eaa21d4fff75f73954605eaa41f61cb94fb008cf2e00c6"}, - {file = "pyarrow-15.0.0-cp310-cp310-manylinux_2_28_aarch64.whl", hash = "sha256:6f87d9c4f09e049c2cade559643424da84c43a35068f2a1c4653dc5b1408a929"}, - {file = "pyarrow-15.0.0-cp310-cp310-manylinux_2_28_x86_64.whl", hash = "sha256:85239b9f93278e130d86c0e6bb455dcb66fc3fd891398b9d45ace8799a871a1e"}, - {file = "pyarrow-15.0.0-cp310-cp310-win_amd64.whl", hash = "sha256:5b8d43e31ca16aa6e12402fcb1e14352d0d809de70edd185c7650fe80e0769e3"}, - {file = "pyarrow-15.0.0-cp311-cp311-macosx_10_15_x86_64.whl", hash = "sha256:fa7cd198280dbd0c988df525e50e35b5d16873e2cdae2aaaa6363cdb64e3eec5"}, - {file = "pyarrow-15.0.0-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:8780b1a29d3c8b21ba6b191305a2a607de2e30dab399776ff0aa09131e266340"}, - {file = "pyarrow-15.0.0-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:fe0ec198ccc680f6c92723fadcb97b74f07c45ff3fdec9dd765deb04955ccf19"}, - {file = "pyarrow-15.0.0-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:036a7209c235588c2f07477fe75c07e6caced9b7b61bb897c8d4e52c4b5f9555"}, - {file = "pyarrow-15.0.0-cp311-cp311-manylinux_2_28_aarch64.whl", hash = "sha256:2bd8a0e5296797faf9a3294e9fa2dc67aa7f10ae2207920dbebb785c77e9dbe5"}, - {file = "pyarrow-15.0.0-cp311-cp311-manylinux_2_28_x86_64.whl", hash = "sha256:e8ebed6053dbe76883a822d4e8da36860f479d55a762bd9e70d8494aed87113e"}, - {file = "pyarrow-15.0.0-cp311-cp311-win_amd64.whl", hash = "sha256:17d53a9d1b2b5bd7d5e4cd84d018e2a45bc9baaa68f7e6e3ebed45649900ba99"}, - {file = "pyarrow-15.0.0-cp312-cp312-macosx_10_15_x86_64.whl", hash = "sha256:9950a9c9df24090d3d558b43b97753b8f5867fb8e521f29876aa021c52fda351"}, - {file = "pyarrow-15.0.0-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:003d680b5e422d0204e7287bb3fa775b332b3fce2996aa69e9adea23f5c8f970"}, - {file = "pyarrow-15.0.0-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:f75fce89dad10c95f4bf590b765e3ae98bcc5ba9f6ce75adb828a334e26a3d40"}, - {file = "pyarrow-15.0.0-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:0ca9cb0039923bec49b4fe23803807e4ef39576a2bec59c32b11296464623dc2"}, - {file = "pyarrow-15.0.0-cp312-cp312-manylinux_2_28_aarch64.whl", hash = "sha256:9ed5a78ed29d171d0acc26a305a4b7f83c122d54ff5270810ac23c75813585e4"}, - {file = "pyarrow-15.0.0-cp312-cp312-manylinux_2_28_x86_64.whl", hash = "sha256:6eda9e117f0402dfcd3cd6ec9bfee89ac5071c48fc83a84f3075b60efa96747f"}, - {file = "pyarrow-15.0.0-cp312-cp312-win_amd64.whl", hash = "sha256:9a3a6180c0e8f2727e6f1b1c87c72d3254cac909e609f35f22532e4115461177"}, - {file = "pyarrow-15.0.0-cp38-cp38-macosx_10_15_x86_64.whl", hash = "sha256:19a8918045993349b207de72d4576af0191beef03ea655d8bdb13762f0cd6eac"}, - {file = "pyarrow-15.0.0-cp38-cp38-macosx_11_0_arm64.whl", hash = "sha256:d0ec076b32bacb6666e8813a22e6e5a7ef1314c8069d4ff345efa6246bc38593"}, - {file = "pyarrow-15.0.0-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:5db1769e5d0a77eb92344c7382d6543bea1164cca3704f84aa44e26c67e320fb"}, - {file = "pyarrow-15.0.0-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:e2617e3bf9df2a00020dd1c1c6dce5cc343d979efe10bc401c0632b0eef6ef5b"}, - {file = "pyarrow-15.0.0-cp38-cp38-manylinux_2_28_aarch64.whl", hash = "sha256:d31c1d45060180131caf10f0f698e3a782db333a422038bf7fe01dace18b3a31"}, - {file = "pyarrow-15.0.0-cp38-cp38-manylinux_2_28_x86_64.whl", hash = "sha256:c8c287d1d479de8269398b34282e206844abb3208224dbdd7166d580804674b7"}, - {file = "pyarrow-15.0.0-cp38-cp38-win_amd64.whl", hash = "sha256:07eb7f07dc9ecbb8dace0f58f009d3a29ee58682fcdc91337dfeb51ea618a75b"}, - {file = "pyarrow-15.0.0-cp39-cp39-macosx_10_15_x86_64.whl", hash = "sha256:47af7036f64fce990bb8a5948c04722e4e3ea3e13b1007ef52dfe0aa8f23cf7f"}, - {file = "pyarrow-15.0.0-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:93768ccfff85cf044c418bfeeafce9a8bb0cee091bd8fd19011aff91e58de540"}, - {file = "pyarrow-15.0.0-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:f6ee87fd6892700960d90abb7b17a72a5abb3b64ee0fe8db6c782bcc2d0dc0b4"}, - {file = "pyarrow-15.0.0-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:001fca027738c5f6be0b7a3159cc7ba16a5c52486db18160909a0831b063c4e4"}, - {file = "pyarrow-15.0.0-cp39-cp39-manylinux_2_28_aarch64.whl", hash = "sha256:d1c48648f64aec09accf44140dccb92f4f94394b8d79976c426a5b79b11d4fa7"}, - {file = "pyarrow-15.0.0-cp39-cp39-manylinux_2_28_x86_64.whl", hash = "sha256:972a0141be402bb18e3201448c8ae62958c9c7923dfaa3b3d4530c835ac81aed"}, - {file = "pyarrow-15.0.0-cp39-cp39-win_amd64.whl", hash = "sha256:f01fc5cf49081426429127aa2d427d9d98e1cb94a32cb961d583a70b7c4504e6"}, - {file = "pyarrow-15.0.0.tar.gz", hash = "sha256:876858f549d540898f927eba4ef77cd549ad8d24baa3207cf1b72e5788b50e83"}, + {file = "pyarrow-16.1.0-cp310-cp310-macosx_10_15_x86_64.whl", hash = "sha256:17e23b9a65a70cc733d8b738baa6ad3722298fa0c81d88f63ff94bf25eaa77b9"}, + {file = "pyarrow-16.1.0-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:4740cc41e2ba5d641071d0ab5e9ef9b5e6e8c7611351a5cb7c1d175eaf43674a"}, + {file = "pyarrow-16.1.0-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:98100e0268d04e0eec47b73f20b39c45b4006f3c4233719c3848aa27a03c1aef"}, + {file = "pyarrow-16.1.0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:f68f409e7b283c085f2da014f9ef81e885d90dcd733bd648cfba3ef265961848"}, + {file = "pyarrow-16.1.0-cp310-cp310-manylinux_2_28_aarch64.whl", hash = "sha256:a8914cd176f448e09746037b0c6b3a9d7688cef451ec5735094055116857580c"}, + {file = "pyarrow-16.1.0-cp310-cp310-manylinux_2_28_x86_64.whl", hash = "sha256:48be160782c0556156d91adbdd5a4a7e719f8d407cb46ae3bb4eaee09b3111bd"}, + {file = "pyarrow-16.1.0-cp310-cp310-win_amd64.whl", hash = "sha256:9cf389d444b0f41d9fe1444b70650fea31e9d52cfcb5f818b7888b91b586efff"}, + {file = "pyarrow-16.1.0-cp311-cp311-macosx_10_15_x86_64.whl", hash = "sha256:d0ebea336b535b37eee9eee31761813086d33ed06de9ab6fc6aaa0bace7b250c"}, + {file = "pyarrow-16.1.0-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:2e73cfc4a99e796727919c5541c65bb88b973377501e39b9842ea71401ca6c1c"}, + {file = "pyarrow-16.1.0-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:bf9251264247ecfe93e5f5a0cd43b8ae834f1e61d1abca22da55b20c788417f6"}, + {file = "pyarrow-16.1.0-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:ddf5aace92d520d3d2a20031d8b0ec27b4395cab9f74e07cc95edf42a5cc0147"}, + {file = "pyarrow-16.1.0-cp311-cp311-manylinux_2_28_aarch64.whl", hash = "sha256:25233642583bf658f629eb230b9bb79d9af4d9f9229890b3c878699c82f7d11e"}, + {file = "pyarrow-16.1.0-cp311-cp311-manylinux_2_28_x86_64.whl", hash = "sha256:a33a64576fddfbec0a44112eaf844c20853647ca833e9a647bfae0582b2ff94b"}, + {file = "pyarrow-16.1.0-cp311-cp311-win_amd64.whl", hash = "sha256:185d121b50836379fe012753cf15c4ba9638bda9645183ab36246923875f8d1b"}, + {file = "pyarrow-16.1.0-cp312-cp312-macosx_10_15_x86_64.whl", hash = "sha256:2e51ca1d6ed7f2e9d5c3c83decf27b0d17bb207a7dea986e8dc3e24f80ff7d6f"}, + {file = "pyarrow-16.1.0-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:06ebccb6f8cb7357de85f60d5da50e83507954af617d7b05f48af1621d331c9a"}, + {file = "pyarrow-16.1.0-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:b04707f1979815f5e49824ce52d1dceb46e2f12909a48a6a753fe7cafbc44a0c"}, + {file = "pyarrow-16.1.0-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:0d32000693deff8dc5df444b032b5985a48592c0697cb6e3071a5d59888714e2"}, + {file = "pyarrow-16.1.0-cp312-cp312-manylinux_2_28_aarch64.whl", hash = "sha256:8785bb10d5d6fd5e15d718ee1d1f914fe768bf8b4d1e5e9bf253de8a26cb1628"}, + {file = "pyarrow-16.1.0-cp312-cp312-manylinux_2_28_x86_64.whl", hash = "sha256:e1369af39587b794873b8a307cc6623a3b1194e69399af0efd05bb202195a5a7"}, + {file = "pyarrow-16.1.0-cp312-cp312-win_amd64.whl", hash = "sha256:febde33305f1498f6df85e8020bca496d0e9ebf2093bab9e0f65e2b4ae2b3444"}, + {file = "pyarrow-16.1.0-cp38-cp38-macosx_10_15_x86_64.whl", hash = "sha256:b5f5705ab977947a43ac83b52ade3b881eb6e95fcc02d76f501d549a210ba77f"}, + {file = "pyarrow-16.1.0-cp38-cp38-macosx_11_0_arm64.whl", hash = "sha256:0d27bf89dfc2576f6206e9cd6cf7a107c9c06dc13d53bbc25b0bd4556f19cf5f"}, + {file = "pyarrow-16.1.0-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:0d07de3ee730647a600037bc1d7b7994067ed64d0eba797ac74b2bc77384f4c2"}, + {file = "pyarrow-16.1.0-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:fbef391b63f708e103df99fbaa3acf9f671d77a183a07546ba2f2c297b361e83"}, + {file = "pyarrow-16.1.0-cp38-cp38-manylinux_2_28_aarch64.whl", hash = "sha256:19741c4dbbbc986d38856ee7ddfdd6a00fc3b0fc2d928795b95410d38bb97d15"}, + {file = "pyarrow-16.1.0-cp38-cp38-manylinux_2_28_x86_64.whl", hash = "sha256:f2c5fb249caa17b94e2b9278b36a05ce03d3180e6da0c4c3b3ce5b2788f30eed"}, + {file = "pyarrow-16.1.0-cp38-cp38-win_amd64.whl", hash = "sha256:e6b6d3cd35fbb93b70ade1336022cc1147b95ec6af7d36906ca7fe432eb09710"}, + {file = "pyarrow-16.1.0-cp39-cp39-macosx_10_15_x86_64.whl", hash = "sha256:18da9b76a36a954665ccca8aa6bd9f46c1145f79c0bb8f4f244f5f8e799bca55"}, + {file = "pyarrow-16.1.0-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:99f7549779b6e434467d2aa43ab2b7224dd9e41bdde486020bae198978c9e05e"}, + {file = "pyarrow-16.1.0-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:f07fdffe4fd5b15f5ec15c8b64584868d063bc22b86b46c9695624ca3505b7b4"}, + {file = "pyarrow-16.1.0-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:ddfe389a08ea374972bd4065d5f25d14e36b43ebc22fc75f7b951f24378bf0b5"}, + {file = "pyarrow-16.1.0-cp39-cp39-manylinux_2_28_aarch64.whl", hash = "sha256:3b20bd67c94b3a2ea0a749d2a5712fc845a69cb5d52e78e6449bbd295611f3aa"}, + {file = "pyarrow-16.1.0-cp39-cp39-manylinux_2_28_x86_64.whl", hash = "sha256:ba8ac20693c0bb0bf4b238751d4409e62852004a8cf031c73b0e0962b03e45e3"}, + {file = "pyarrow-16.1.0-cp39-cp39-win_amd64.whl", hash = "sha256:31a1851751433d89a986616015841977e0a188662fcffd1a5677453f1df2de0a"}, + {file = "pyarrow-16.1.0.tar.gz", hash = "sha256:15fbb22ea96d11f0b5768504a3f961edab25eaf4197c341720c4a387f6c60315"}, ] [package.dependencies] -numpy = ">=1.16.6,<2" +numpy = ">=1.16.6" [[package]] name = "pyarrow-hotfix" @@ -3369,13 +3423,13 @@ files = [ [[package]] name = "pycparser" -version = "2.21" +version = "2.22" description = "C parser in Python" optional = false -python-versions = ">=2.7, !=3.0., !=3.1., !=3.2., !=3.3." +python-versions = ">=3.8" files = [ - {file = "pycparser-2.21-py2.py3-none-any.whl", hash = "sha256:8ee45429555515e1f6b185e78100aea234072576aa43ab53aefcae078162fca9"}, - {file = "pycparser-2.21.tar.gz", hash = "sha256:e644fdec12f7872f86c58ff790da456218b10f863970249516d60a5eaca77206"}, + {file = "pycparser-2.22-py3-none-any.whl", hash = "sha256:c3702b6d3dd8c7abc1afa565d7e63d53a1d0bd86cdc24edd75470f4de499cfcc"}, + {file = "pycparser-2.22.tar.gz", hash = "sha256:491c8be9c040f5390f5bf44a5b07752bd07f56edf992381b05c701439eec10f6"}, ] [[package]] @@ -3430,32 +3484,31 @@ tests = ["WebTest", "beautifulsoup4", "flake8", "pyopenssl", "pytest (>=3.6)", " [[package]] name = "pygments" -version = "2.17.2" +version = "2.18.0" description = "Pygments is a syntax highlighting package written in Python." optional = false -python-versions = ">=3.7" +python-versions = ">=3.8" files = [ - {file = "pygments-2.17.2-py3-none-any.whl", hash = "sha256:b27c2826c47d0f3219f29554824c30c5e8945175d888647acd804ddd04af846c"}, - {file = "pygments-2.17.2.tar.gz", hash = "sha256:da46cec9fd2de5be3a8a784f434e4c4ab670b4ff54d605c4c2717e9d49c4c367"}, + {file = "pygments-2.18.0-py3-none-any.whl", hash = "sha256:b8e6aca0523f3ab76fee51799c488e38782ac06eafcf95e7ba832985c8e7b13a"}, + {file = "pygments-2.18.0.tar.gz", hash = "sha256:786ff802f32e91311bff3889f6e9a86e81505fe99f2735bb6d60ae0c5004f199"}, ] [package.extras] -plugins = ["importlib-metadata"] windows-terminal = ["colorama (>=0.4.6)"] [[package]] name = "pymdown-extensions" -version = "10.7" +version = "10.8.1" description = "Extension pack for Python Markdown." optional = false python-versions = ">=3.8" files = [ - {file = "pymdown_extensions-10.7-py3-none-any.whl", hash = "sha256:6ca215bc57bc12bf32b414887a68b810637d039124ed9b2e5bd3325cbb2c050c"}, - {file = "pymdown_extensions-10.7.tar.gz", hash = "sha256:c0d64d5cf62566f59e6b2b690a4095c931107c250a8c8e1351c1de5f6b036deb"}, + {file = "pymdown_extensions-10.8.1-py3-none-any.whl", hash = "sha256:f938326115884f48c6059c67377c46cf631c733ef3629b6eed1349989d1b30cb"}, + {file = "pymdown_extensions-10.8.1.tar.gz", hash = "sha256:3ab1db5c9e21728dabf75192d71471f8e50f216627e9a1fa9535ecb0231b9940"}, ] [package.dependencies] -markdown = ">=3.5" +markdown = ">=3.6" pyyaml = "" [package.extras] @@ -3463,13 +3516,13 @@ extra = ["pygments (>=2.12)"] [[package]] name = "pyparsing" -version = "3.1.1" +version = "3.1.2" description = "pyparsing module - Classes and methods to define and execute parsing grammars" optional = false python-versions = ">=3.6.8" files = [ - {file = "pyparsing-3.1.1-py3-none-any.whl", hash = "sha256:32c7c0b711493c72ff18a981d24f28aaf9c1fb7ed5e9667c9e84e3db623bdbfb"}, - {file = "pyparsing-3.1.1.tar.gz", hash = "sha256:ede28a1a32462f5a9705e07aea48001a08f7cf81a021585011deba701581a0db"}, + {file = "pyparsing-3.1.2-py3-none-any.whl", hash = "sha256:f9db75911801ed778fe61bb643079ff86601aca99fcae6345aa67292038fb742"}, + {file = "pyparsing-3.1.2.tar.gz", hash = "sha256:a1bac0ce561155ecc3ed78ca94d3c9378656ad4c94c1270de543f621420f94ad"}, ] [package.extras] @@ -3527,13 +3580,13 @@ files = [ [[package]] name = "pytest" -version = "8.0.0" +version = "8.2.1" description = "pytest: simple powerful testing with Python" optional = false python-versions = ">=3.8" files = [ - {file = "pytest-8.0.0-py3-none-any.whl", hash = "sha256:50fb9cbe836c3f20f0dfa99c565201fb75dc54c8d76373cd1bde06b06657bdb6"}, - {file = "pytest-8.0.0.tar.gz", hash = "sha256:249b1b0864530ba251b7438274c4d251c58d868edaaec8762893ad4a0d71c36c"}, + {file = "pytest-8.2.1-py3-none-any.whl", hash = "sha256:faccc5d332b8c3719f40283d0d44aa5cf101cec36f88cde9ed8f2bc0538612b1"}, + {file = "pytest-8.2.1.tar.gz", hash = "sha256:5046e5b46d8e4cac199c373041f26be56fdb81eb4e67dc11d4e10811fc3408fd"}, ] [package.dependencies] @@ -3541,21 +3594,21 @@ colorama = {version = "", markers = "sys_platform == \"win32\""} exceptiongroup = {version = ">=1.0.0rc8", markers = "python_version < \"3.11\""} iniconfig = "" packaging = "" -pluggy = ">=1.3.0,<2.0" -tomli = {version = ">=1.0.0", markers = "python_version < \"3.11\""} +pluggy = ">=1.5,<2.0" +tomli = {version = ">=1", markers = "python_version < \"3.11\""} [package.extras] -testing = ["argcomplete", "attrs (>=19.2.0)", "hypothesis (>=3.56)", "mock", "nose", "pygments (>=2.7.2)", "requests", "setuptools", "xmlschema"] +dev = ["argcomplete", "attrs (>=19.2)", "hypothesis (>=3.56)", "mock", "pygments (>=2.7.2)", "requests", "setuptools", "xmlschema"] [[package]] name = "pytest-cov" -version = "4.1.0" +version = "5.0.0" description = "Pytest plugin for measuring coverage." optional = false -python-versions = ">=3.7" +python-versions = ">=3.8" files = [ - {file = "pytest-cov-4.1.0.tar.gz", hash = "sha256:3904b13dfbfec47f003b8e77fd5b589cd11904a21ddf1ab38a64f204d6a10ef6"}, - {file = "pytest_cov-4.1.0-py3-none-any.whl", hash = "sha256:6ba70b9e97e69fcc3fb45bfeab2d0a138fb65c4d0d6a41ef33983ad114be8c3a"}, + {file = "pytest-cov-5.0.0.tar.gz", hash = "sha256:5837b58e9f6ebd335b0f8060eecce69b662415b16dc503883a02f45dfeb14857"}, + {file = "pytest_cov-5.0.0-py3-none-any.whl", hash = "sha256:4f0764a1219df53214206bf1feea4633c3b558a2925c8b59f144f682861ce652"}, ] [package.dependencies] @@ -3563,17 +3616,17 @@ coverage = {version = ">=5.2.1", extras = ["toml"]} pytest = ">=4.6" [package.extras] -testing = ["fields", "hunter", "process-tests", "pytest-xdist", "six", "virtualenv"] +testing = ["fields", "hunter", "process-tests", "pytest-xdist", "virtualenv"] [[package]] name = "python-dateutil" -version = "2.8.2" +version = "2.9.0.post0" description = "Extensions to the standard Python datetime module" optional = false python-versions = "!=3.0.,!=3.1.,!=3.2.,>=2.7" files = [ - {file = "python-dateutil-2.8.2.tar.gz", hash = "sha256:0123cacc1627ae19ddf3c27a5de5bd67ee4586fbdd6440d9748f8abb483d3e86"}, - {file = "python_dateutil-2.8.2-py2.py3-none-any.whl", hash = "sha256:961d03dc3453ebbc59dbdea9e4e11c5651520a876d0f4db161e8674aae935da9"}, + {file = "python-dateutil-2.9.0.post0.tar.gz", hash = "sha256:37dd54208da7e1cd875388217d5e00ebd4179249f90fb72437e91a35459a0ad3"}, + {file = "python_dateutil-2.9.0.post0-py2.py3-none-any.whl", hash = "sha256:a8b2bc7bffae282281c8140a97d3aa9c14da0b136dfe83f850eea9a5f7470427"}, ] [package.dependencies] @@ -3592,13 +3645,13 @@ files = [ [[package]] name = "pyviz-comms" -version = "3.0.1" +version = "3.0.2" description = "A JupyterLab extension for rendering HoloViz content." optional = false python-versions = ">=3.8" files = [ - {file = "pyviz_comms-3.0.1-py3-none-any.whl", hash = "sha256:0130e952b942906a0eb5fcbcc750262a8e4f565a9b06b3c0d8d631f33b61b78e"}, - {file = "pyviz_comms-3.0.1.tar.gz", hash = "sha256:427c33a5a81780db9b9e757f0675f65ea2292d9a642a2d291cfb5cae6cd46991"}, + {file = "pyviz_comms-3.0.2-py3-none-any.whl", hash = "sha256:31541b976a21b7738557c3ea23bd8e44e94e736b9ed269570dcc28db4449d7e3"}, + {file = "pyviz_comms-3.0.2.tar.gz", hash = "sha256:3167df932656416c4bd711205dad47e986a3ebae1f316258ddc26f9e01513ef7"}, ] [package.dependencies] @@ -3707,104 +3760,99 @@ pyyaml = "" [[package]] name = "pyzmq" -version = "25.1.2" +version = "26.0.3" description = "Python bindings for 0MQ" optional = false -python-versions = ">=3.6" +python-versions = ">=3.7" files = [ - {file = "pyzmq-25.1.2-cp310-cp310-macosx_10_15_universal2.whl", hash = "sha256:e624c789359f1a16f83f35e2c705d07663ff2b4d4479bad35621178d8f0f6ea4"}, - {file = "pyzmq-25.1.2-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:49151b0efece79f6a79d41a461d78535356136ee70084a1c22532fc6383f4ad0"}, - {file = "pyzmq-25.1.2-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:d9a5f194cf730f2b24d6af1f833c14c10f41023da46a7f736f48b6d35061e76e"}, - {file = "pyzmq-25.1.2-cp310-cp310-manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:faf79a302f834d9e8304fafdc11d0d042266667ac45209afa57e5efc998e3872"}, - {file = "pyzmq-25.1.2-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:7f51a7b4ead28d3fca8dda53216314a553b0f7a91ee8fc46a72b402a78c3e43d"}, - {file = "pyzmq-25.1.2-cp310-cp310-manylinux_2_28_x86_64.whl", hash = "sha256:0ddd6d71d4ef17ba5a87becf7ddf01b371eaba553c603477679ae817a8d84d75"}, - {file = "pyzmq-25.1.2-cp310-cp310-musllinux_1_1_aarch64.whl", hash = "sha256:246747b88917e4867e2367b005fc8eefbb4a54b7db363d6c92f89d69abfff4b6"}, - {file = "pyzmq-25.1.2-cp310-cp310-musllinux_1_1_i686.whl", hash = "sha256:00c48ae2fd81e2a50c3485de1b9d5c7c57cd85dc8ec55683eac16846e57ac979"}, - {file = "pyzmq-25.1.2-cp310-cp310-musllinux_1_1_x86_64.whl", hash = "sha256:5a68d491fc20762b630e5db2191dd07ff89834086740f70e978bb2ef2668be08"}, - {file = "pyzmq-25.1.2-cp310-cp310-win32.whl", hash = "sha256:09dfe949e83087da88c4a76767df04b22304a682d6154de2c572625c62ad6886"}, - {file = "pyzmq-25.1.2-cp310-cp310-win_amd64.whl", hash = "sha256:fa99973d2ed20417744fca0073390ad65ce225b546febb0580358e36aa90dba6"}, - {file = "pyzmq-25.1.2-cp311-cp311-macosx_10_15_universal2.whl", hash = "sha256:82544e0e2d0c1811482d37eef297020a040c32e0687c1f6fc23a75b75db8062c"}, - {file = "pyzmq-25.1.2-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:01171fc48542348cd1a360a4b6c3e7d8f46cdcf53a8d40f84db6707a6768acc1"}, - {file = "pyzmq-25.1.2-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:bc69c96735ab501419c432110016329bf0dea8898ce16fab97c6d9106dc0b348"}, - {file = "pyzmq-25.1.2-cp311-cp311-manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:3e124e6b1dd3dfbeb695435dff0e383256655bb18082e094a8dd1f6293114642"}, - {file = "pyzmq-25.1.2-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:7598d2ba821caa37a0f9d54c25164a4fa351ce019d64d0b44b45540950458840"}, - {file = "pyzmq-25.1.2-cp311-cp311-manylinux_2_28_x86_64.whl", hash = "sha256:d1299d7e964c13607efd148ca1f07dcbf27c3ab9e125d1d0ae1d580a1682399d"}, - {file = "pyzmq-25.1.2-cp311-cp311-musllinux_1_1_aarch64.whl", hash = "sha256:4e6f689880d5ad87918430957297c975203a082d9a036cc426648fcbedae769b"}, - {file = "pyzmq-25.1.2-cp311-cp311-musllinux_1_1_i686.whl", hash = "sha256:cc69949484171cc961e6ecd4a8911b9ce7a0d1f738fcae717177c231bf77437b"}, - {file = "pyzmq-25.1.2-cp311-cp311-musllinux_1_1_x86_64.whl", hash = "sha256:9880078f683466b7f567b8624bfc16cad65077be046b6e8abb53bed4eeb82dd3"}, - {file = "pyzmq-25.1.2-cp311-cp311-win32.whl", hash = "sha256:4e5837af3e5aaa99a091302df5ee001149baff06ad22b722d34e30df5f0d9097"}, - {file = "pyzmq-25.1.2-cp311-cp311-win_amd64.whl", hash = "sha256:25c2dbb97d38b5ac9fd15586e048ec5eb1e38f3d47fe7d92167b0c77bb3584e9"}, - {file = "pyzmq-25.1.2-cp312-cp312-macosx_10_15_universal2.whl", hash = "sha256:11e70516688190e9c2db14fcf93c04192b02d457b582a1f6190b154691b4c93a"}, - {file = "pyzmq-25.1.2-cp312-cp312-macosx_10_9_x86_64.whl", hash = "sha256:313c3794d650d1fccaaab2df942af9f2c01d6217c846177cfcbc693c7410839e"}, - {file = "pyzmq-25.1.2-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:1b3cbba2f47062b85fe0ef9de5b987612140a9ba3a9c6d2543c6dec9f7c2ab27"}, - {file = "pyzmq-25.1.2-cp312-cp312-manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:fc31baa0c32a2ca660784d5af3b9487e13b61b3032cb01a115fce6588e1bed30"}, - {file = "pyzmq-25.1.2-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:02c9087b109070c5ab0b383079fa1b5f797f8d43e9a66c07a4b8b8bdecfd88ee"}, - {file = "pyzmq-25.1.2-cp312-cp312-manylinux_2_28_x86_64.whl", hash = "sha256:f8429b17cbb746c3e043cb986328da023657e79d5ed258b711c06a70c2ea7537"}, - {file = "pyzmq-25.1.2-cp312-cp312-musllinux_1_1_aarch64.whl", hash = "sha256:5074adeacede5f810b7ef39607ee59d94e948b4fd954495bdb072f8c54558181"}, - {file = "pyzmq-25.1.2-cp312-cp312-musllinux_1_1_i686.whl", hash = "sha256:7ae8f354b895cbd85212da245f1a5ad8159e7840e37d78b476bb4f4c3f32a9fe"}, - {file = "pyzmq-25.1.2-cp312-cp312-musllinux_1_1_x86_64.whl", hash = "sha256:b264bf2cc96b5bc43ce0e852be995e400376bd87ceb363822e2cb1964fcdc737"}, - {file = "pyzmq-25.1.2-cp312-cp312-win32.whl", hash = "sha256:02bbc1a87b76e04fd780b45e7f695471ae6de747769e540da909173d50ff8e2d"}, - {file = "pyzmq-25.1.2-cp312-cp312-win_amd64.whl", hash = "sha256:ced111c2e81506abd1dc142e6cd7b68dd53747b3b7ae5edbea4578c5eeff96b7"}, - {file = "pyzmq-25.1.2-cp36-cp36m-macosx_10_9_x86_64.whl", hash = "sha256:7b6d09a8962a91151f0976008eb7b29b433a560fde056ec7a3db9ec8f1075438"}, - {file = "pyzmq-25.1.2-cp36-cp36m-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:967668420f36878a3c9ecb5ab33c9d0ff8d054f9c0233d995a6d25b0e95e1b6b"}, - {file = "pyzmq-25.1.2-cp36-cp36m-manylinux_2_5_i686.manylinux1_i686.whl", hash = "sha256:5edac3f57c7ddaacdb4d40f6ef2f9e299471fc38d112f4bc6d60ab9365445fb0"}, - {file = "pyzmq-25.1.2-cp36-cp36m-manylinux_2_5_x86_64.manylinux1_x86_64.whl", hash = "sha256:0dabfb10ef897f3b7e101cacba1437bd3a5032ee667b7ead32bbcdd1a8422fe7"}, - {file = "pyzmq-25.1.2-cp36-cp36m-musllinux_1_1_aarch64.whl", hash = "sha256:2c6441e0398c2baacfe5ba30c937d274cfc2dc5b55e82e3749e333aabffde561"}, - {file = "pyzmq-25.1.2-cp36-cp36m-musllinux_1_1_i686.whl", hash = "sha256:16b726c1f6c2e7625706549f9dbe9b06004dfbec30dbed4bf50cbdfc73e5b32a"}, - {file = "pyzmq-25.1.2-cp36-cp36m-musllinux_1_1_x86_64.whl", hash = "sha256:a86c2dd76ef71a773e70551a07318b8e52379f58dafa7ae1e0a4be78efd1ff16"}, - {file = "pyzmq-25.1.2-cp36-cp36m-win32.whl", hash = "sha256:359f7f74b5d3c65dae137f33eb2bcfa7ad9ebefd1cab85c935f063f1dbb245cc"}, - {file = "pyzmq-25.1.2-cp36-cp36m-win_amd64.whl", hash = "sha256:55875492f820d0eb3417b51d96fea549cde77893ae3790fd25491c5754ea2f68"}, - {file = "pyzmq-25.1.2-cp37-cp37m-macosx_10_9_x86_64.whl", hash = "sha256:b8c8a419dfb02e91b453615c69568442e897aaf77561ee0064d789705ff37a92"}, - {file = "pyzmq-25.1.2-cp37-cp37m-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:8807c87fa893527ae8a524c15fc505d9950d5e856f03dae5921b5e9aa3b8783b"}, - {file = "pyzmq-25.1.2-cp37-cp37m-manylinux_2_5_i686.manylinux1_i686.whl", hash = "sha256:5e319ed7d6b8f5fad9b76daa0a68497bc6f129858ad956331a5835785761e003"}, - {file = "pyzmq-25.1.2-cp37-cp37m-manylinux_2_5_x86_64.manylinux1_x86_64.whl", hash = "sha256:3c53687dde4d9d473c587ae80cc328e5b102b517447456184b485587ebd18b62"}, - {file = "pyzmq-25.1.2-cp37-cp37m-musllinux_1_1_aarch64.whl", hash = "sha256:9add2e5b33d2cd765ad96d5eb734a5e795a0755f7fc49aa04f76d7ddda73fd70"}, - {file = "pyzmq-25.1.2-cp37-cp37m-musllinux_1_1_i686.whl", hash = "sha256:e690145a8c0c273c28d3b89d6fb32c45e0d9605b2293c10e650265bf5c11cfec"}, - {file = "pyzmq-25.1.2-cp37-cp37m-musllinux_1_1_x86_64.whl", hash = "sha256:00a06faa7165634f0cac1abb27e54d7a0b3b44eb9994530b8ec73cf52e15353b"}, - {file = "pyzmq-25.1.2-cp37-cp37m-win32.whl", hash = "sha256:0f97bc2f1f13cb16905a5f3e1fbdf100e712d841482b2237484360f8bc4cb3d7"}, - {file = "pyzmq-25.1.2-cp37-cp37m-win_amd64.whl", hash = "sha256:6cc0020b74b2e410287e5942e1e10886ff81ac77789eb20bec13f7ae681f0fdd"}, - {file = "pyzmq-25.1.2-cp38-cp38-macosx_10_15_universal2.whl", hash = "sha256:bef02cfcbded83473bdd86dd8d3729cd82b2e569b75844fb4ea08fee3c26ae41"}, - {file = "pyzmq-25.1.2-cp38-cp38-macosx_10_9_x86_64.whl", hash = "sha256:e10a4b5a4b1192d74853cc71a5e9fd022594573926c2a3a4802020360aa719d8"}, - {file = "pyzmq-25.1.2-cp38-cp38-manylinux_2_12_i686.manylinux2010_i686.whl", hash = "sha256:8c5f80e578427d4695adac6fdf4370c14a2feafdc8cb35549c219b90652536ae"}, - {file = "pyzmq-25.1.2-cp38-cp38-manylinux_2_12_x86_64.manylinux2010_x86_64.whl", hash = "sha256:5dde6751e857910c1339890f3524de74007958557593b9e7e8c5f01cd919f8a7"}, - {file = "pyzmq-25.1.2-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:ea1608dd169da230a0ad602d5b1ebd39807ac96cae1845c3ceed39af08a5c6df"}, - {file = "pyzmq-25.1.2-cp38-cp38-musllinux_1_1_aarch64.whl", hash = "sha256:0f513130c4c361201da9bc69df25a086487250e16b5571ead521b31ff6b02220"}, - {file = "pyzmq-25.1.2-cp38-cp38-musllinux_1_1_i686.whl", hash = "sha256:019744b99da30330798bb37df33549d59d380c78e516e3bab9c9b84f87a9592f"}, - {file = "pyzmq-25.1.2-cp38-cp38-musllinux_1_1_x86_64.whl", hash = "sha256:2e2713ef44be5d52dd8b8e2023d706bf66cb22072e97fc71b168e01d25192755"}, - {file = "pyzmq-25.1.2-cp38-cp38-win32.whl", hash = "sha256:07cd61a20a535524906595e09344505a9bd46f1da7a07e504b315d41cd42eb07"}, - {file = "pyzmq-25.1.2-cp38-cp38-win_amd64.whl", hash = "sha256:eb7e49a17fb8c77d3119d41a4523e432eb0c6932187c37deb6fbb00cc3028088"}, - {file = "pyzmq-25.1.2-cp39-cp39-macosx_10_15_universal2.whl", hash = "sha256:94504ff66f278ab4b7e03e4cba7e7e400cb73bfa9d3d71f58d8972a8dc67e7a6"}, - {file = "pyzmq-25.1.2-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:6dd0d50bbf9dca1d0bdea219ae6b40f713a3fb477c06ca3714f208fd69e16fd8"}, - {file = "pyzmq-25.1.2-cp39-cp39-manylinux_2_12_i686.manylinux2010_i686.whl", hash = "sha256:004ff469d21e86f0ef0369717351073e0e577428e514c47c8480770d5e24a565"}, - {file = "pyzmq-25.1.2-cp39-cp39-manylinux_2_12_x86_64.manylinux2010_x86_64.whl", hash = "sha256:c0b5ca88a8928147b7b1e2dfa09f3b6c256bc1135a1338536cbc9ea13d3b7add"}, - {file = "pyzmq-25.1.2-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:2c9a79f1d2495b167119d02be7448bfba57fad2a4207c4f68abc0bab4b92925b"}, - {file = "pyzmq-25.1.2-cp39-cp39-musllinux_1_1_aarch64.whl", hash = "sha256:518efd91c3d8ac9f9b4f7dd0e2b7b8bf1a4fe82a308009016b07eaa48681af82"}, - {file = "pyzmq-25.1.2-cp39-cp39-musllinux_1_1_i686.whl", hash = "sha256:1ec23bd7b3a893ae676d0e54ad47d18064e6c5ae1fadc2f195143fb27373f7f6"}, - {file = "pyzmq-25.1.2-cp39-cp39-musllinux_1_1_x86_64.whl", hash = "sha256:db36c27baed588a5a8346b971477b718fdc66cf5b80cbfbd914b4d6d355e44e2"}, - {file = "pyzmq-25.1.2-cp39-cp39-win32.whl", hash = "sha256:39b1067f13aba39d794a24761e385e2eddc26295826530a8c7b6c6c341584289"}, - {file = "pyzmq-25.1.2-cp39-cp39-win_amd64.whl", hash = "sha256:8e9f3fabc445d0ce320ea2c59a75fe3ea591fdbdeebec5db6de530dd4b09412e"}, - {file = "pyzmq-25.1.2-pp310-pypy310_pp73-macosx_10_9_x86_64.whl", hash = "sha256:a8c1d566344aee826b74e472e16edae0a02e2a044f14f7c24e123002dcff1c05"}, - {file = "pyzmq-25.1.2-pp310-pypy310_pp73-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:759cfd391a0996345ba94b6a5110fca9c557ad4166d86a6e81ea526c376a01e8"}, - {file = "pyzmq-25.1.2-pp310-pypy310_pp73-manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:7c61e346ac34b74028ede1c6b4bcecf649d69b707b3ff9dc0fab453821b04d1e"}, - {file = "pyzmq-25.1.2-pp310-pypy310_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:4cb8fc1f8d69b411b8ec0b5f1ffbcaf14c1db95b6bccea21d83610987435f1a4"}, - {file = "pyzmq-25.1.2-pp310-pypy310_pp73-win_amd64.whl", hash = "sha256:3c00c9b7d1ca8165c610437ca0c92e7b5607b2f9076f4eb4b095c85d6e680a1d"}, - {file = "pyzmq-25.1.2-pp37-pypy37_pp73-macosx_10_9_x86_64.whl", hash = "sha256:df0c7a16ebb94452d2909b9a7b3337940e9a87a824c4fc1c7c36bb4404cb0cde"}, - {file = "pyzmq-25.1.2-pp37-pypy37_pp73-manylinux_2_12_i686.manylinux2010_i686.whl", hash = "sha256:45999e7f7ed5c390f2e87ece7f6c56bf979fb213550229e711e45ecc7d42ccb8"}, - {file = "pyzmq-25.1.2-pp37-pypy37_pp73-manylinux_2_12_x86_64.manylinux2010_x86_64.whl", hash = "sha256:ac170e9e048b40c605358667aca3d94e98f604a18c44bdb4c102e67070f3ac9b"}, - {file = "pyzmq-25.1.2-pp37-pypy37_pp73-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:d1b604734bec94f05f81b360a272fc824334267426ae9905ff32dc2be433ab96"}, - {file = "pyzmq-25.1.2-pp37-pypy37_pp73-win_amd64.whl", hash = "sha256:a793ac733e3d895d96f865f1806f160696422554e46d30105807fdc9841b9f7d"}, - {file = "pyzmq-25.1.2-pp38-pypy38_pp73-macosx_10_9_x86_64.whl", hash = "sha256:0806175f2ae5ad4b835ecd87f5f85583316b69f17e97786f7443baaf54b9bb98"}, - {file = "pyzmq-25.1.2-pp38-pypy38_pp73-manylinux_2_12_i686.manylinux2010_i686.whl", hash = "sha256:ef12e259e7bc317c7597d4f6ef59b97b913e162d83b421dd0db3d6410f17a244"}, - {file = "pyzmq-25.1.2-pp38-pypy38_pp73-manylinux_2_12_x86_64.manylinux2010_x86_64.whl", hash = "sha256:ea253b368eb41116011add00f8d5726762320b1bda892f744c91997b65754d73"}, - {file = "pyzmq-25.1.2-pp38-pypy38_pp73-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:1b9b1f2ad6498445a941d9a4fee096d387fee436e45cc660e72e768d3d8ee611"}, - {file = "pyzmq-25.1.2-pp38-pypy38_pp73-win_amd64.whl", hash = "sha256:8b14c75979ce932c53b79976a395cb2a8cd3aaf14aef75e8c2cb55a330b9b49d"}, - {file = "pyzmq-25.1.2-pp39-pypy39_pp73-macosx_10_9_x86_64.whl", hash = "sha256:889370d5174a741a62566c003ee8ddba4b04c3f09a97b8000092b7ca83ec9c49"}, - {file = "pyzmq-25.1.2-pp39-pypy39_pp73-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:9a18fff090441a40ffda8a7f4f18f03dc56ae73f148f1832e109f9bffa85df15"}, - {file = "pyzmq-25.1.2-pp39-pypy39_pp73-manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:99a6b36f95c98839ad98f8c553d8507644c880cf1e0a57fe5e3a3f3969040882"}, - {file = "pyzmq-25.1.2-pp39-pypy39_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:4345c9a27f4310afbb9c01750e9461ff33d6fb74cd2456b107525bbeebcb5be3"}, - {file = "pyzmq-25.1.2-pp39-pypy39_pp73-manylinux_2_28_x86_64.whl", hash = "sha256:3516e0b6224cf6e43e341d56da15fd33bdc37fa0c06af4f029f7d7dfceceabbc"}, - {file = "pyzmq-25.1.2-pp39-pypy39_pp73-win_amd64.whl", hash = "sha256:146b9b1f29ead41255387fb07be56dc29639262c0f7344f570eecdcd8d683314"}, - {file = "pyzmq-25.1.2.tar.gz", hash = "sha256:93f1aa311e8bb912e34f004cf186407a4e90eec4f0ecc0efd26056bf7eda0226"}, + {file = "pyzmq-26.0.3-cp310-cp310-macosx_10_15_universal2.whl", hash = "sha256:44dd6fc3034f1eaa72ece33588867df9e006a7303725a12d64c3dff92330f625"}, + {file = "pyzmq-26.0.3-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:acb704195a71ac5ea5ecf2811c9ee19ecdc62b91878528302dd0be1b9451cc90"}, + {file = "pyzmq-26.0.3-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:5dbb9c997932473a27afa93954bb77a9f9b786b4ccf718d903f35da3232317de"}, + {file = "pyzmq-26.0.3-cp310-cp310-manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:6bcb34f869d431799c3ee7d516554797f7760cb2198ecaa89c3f176f72d062be"}, + {file = "pyzmq-26.0.3-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:38ece17ec5f20d7d9b442e5174ae9f020365d01ba7c112205a4d59cf19dc38ee"}, + {file = "pyzmq-26.0.3-cp310-cp310-manylinux_2_28_x86_64.whl", hash = "sha256:ba6e5e6588e49139a0979d03a7deb9c734bde647b9a8808f26acf9c547cab1bf"}, + {file = "pyzmq-26.0.3-cp310-cp310-musllinux_1_1_aarch64.whl", hash = "sha256:3bf8b000a4e2967e6dfdd8656cd0757d18c7e5ce3d16339e550bd462f4857e59"}, + {file = "pyzmq-26.0.3-cp310-cp310-musllinux_1_1_i686.whl", hash = "sha256:2136f64fbb86451dbbf70223635a468272dd20075f988a102bf8a3f194a411dc"}, + {file = "pyzmq-26.0.3-cp310-cp310-musllinux_1_1_x86_64.whl", hash = "sha256:e8918973fbd34e7814f59143c5f600ecd38b8038161239fd1a3d33d5817a38b8"}, + {file = "pyzmq-26.0.3-cp310-cp310-win32.whl", hash = "sha256:0aaf982e68a7ac284377d051c742610220fd06d330dcd4c4dbb4cdd77c22a537"}, + {file = "pyzmq-26.0.3-cp310-cp310-win_amd64.whl", hash = "sha256:f1a9b7d00fdf60b4039f4455afd031fe85ee8305b019334b72dcf73c567edc47"}, + {file = "pyzmq-26.0.3-cp310-cp310-win_arm64.whl", hash = "sha256:80b12f25d805a919d53efc0a5ad7c0c0326f13b4eae981a5d7b7cc343318ebb7"}, + {file = "pyzmq-26.0.3-cp311-cp311-macosx_10_15_universal2.whl", hash = "sha256:a72a84570f84c374b4c287183debc776dc319d3e8ce6b6a0041ce2e400de3f32"}, + {file = "pyzmq-26.0.3-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:7ca684ee649b55fd8f378127ac8462fb6c85f251c2fb027eb3c887e8ee347bcd"}, + {file = "pyzmq-26.0.3-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:e222562dc0f38571c8b1ffdae9d7adb866363134299264a1958d077800b193b7"}, + {file = "pyzmq-26.0.3-cp311-cp311-manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:f17cde1db0754c35a91ac00b22b25c11da6eec5746431d6e5092f0cd31a3fea9"}, + {file = "pyzmq-26.0.3-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:4b7c0c0b3244bb2275abe255d4a30c050d541c6cb18b870975553f1fb6f37527"}, + {file = "pyzmq-26.0.3-cp311-cp311-manylinux_2_28_x86_64.whl", hash = "sha256:ac97a21de3712afe6a6c071abfad40a6224fd14fa6ff0ff8d0c6e6cd4e2f807a"}, + {file = "pyzmq-26.0.3-cp311-cp311-musllinux_1_1_aarch64.whl", hash = "sha256:88b88282e55fa39dd556d7fc04160bcf39dea015f78e0cecec8ff4f06c1fc2b5"}, + {file = "pyzmq-26.0.3-cp311-cp311-musllinux_1_1_i686.whl", hash = "sha256:72b67f966b57dbd18dcc7efbc1c7fc9f5f983e572db1877081f075004614fcdd"}, + {file = "pyzmq-26.0.3-cp311-cp311-musllinux_1_1_x86_64.whl", hash = "sha256:f4b6cecbbf3b7380f3b61de3a7b93cb721125dc125c854c14ddc91225ba52f83"}, + {file = "pyzmq-26.0.3-cp311-cp311-win32.whl", hash = "sha256:eed56b6a39216d31ff8cd2f1d048b5bf1700e4b32a01b14379c3b6dde9ce3aa3"}, + {file = "pyzmq-26.0.3-cp311-cp311-win_amd64.whl", hash = "sha256:3191d312c73e3cfd0f0afdf51df8405aafeb0bad71e7ed8f68b24b63c4f36500"}, + {file = "pyzmq-26.0.3-cp311-cp311-win_arm64.whl", hash = "sha256:b6907da3017ef55139cf0e417c5123a84c7332520e73a6902ff1f79046cd3b94"}, + {file = "pyzmq-26.0.3-cp312-cp312-macosx_10_15_universal2.whl", hash = "sha256:068ca17214038ae986d68f4a7021f97e187ed278ab6dccb79f837d765a54d753"}, + {file = "pyzmq-26.0.3-cp312-cp312-macosx_10_9_x86_64.whl", hash = "sha256:7821d44fe07335bea256b9f1f41474a642ca55fa671dfd9f00af8d68a920c2d4"}, + {file = "pyzmq-26.0.3-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:eeb438a26d87c123bb318e5f2b3d86a36060b01f22fbdffd8cf247d52f7c9a2b"}, + {file = "pyzmq-26.0.3-cp312-cp312-manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:69ea9d6d9baa25a4dc9cef5e2b77b8537827b122214f210dd925132e34ae9b12"}, + {file = "pyzmq-26.0.3-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:7daa3e1369355766dea11f1d8ef829905c3b9da886ea3152788dc25ee6079e02"}, + {file = "pyzmq-26.0.3-cp312-cp312-manylinux_2_28_x86_64.whl", hash = "sha256:6ca7a9a06b52d0e38ccf6bca1aeff7be178917893f3883f37b75589d42c4ac20"}, + {file = "pyzmq-26.0.3-cp312-cp312-musllinux_1_1_aarch64.whl", hash = "sha256:1b7d0e124948daa4d9686d421ef5087c0516bc6179fdcf8828b8444f8e461a77"}, + {file = "pyzmq-26.0.3-cp312-cp312-musllinux_1_1_i686.whl", hash = "sha256:e746524418b70f38550f2190eeee834db8850088c834d4c8406fbb9bc1ae10b2"}, + {file = "pyzmq-26.0.3-cp312-cp312-musllinux_1_1_x86_64.whl", hash = "sha256:6b3146f9ae6af82c47a5282ac8803523d381b3b21caeae0327ed2f7ecb718798"}, + {file = "pyzmq-26.0.3-cp312-cp312-win32.whl", hash = "sha256:2b291d1230845871c00c8462c50565a9cd6026fe1228e77ca934470bb7d70ea0"}, + {file = "pyzmq-26.0.3-cp312-cp312-win_amd64.whl", hash = "sha256:926838a535c2c1ea21c903f909a9a54e675c2126728c21381a94ddf37c3cbddf"}, + {file = "pyzmq-26.0.3-cp312-cp312-win_arm64.whl", hash = "sha256:5bf6c237f8c681dfb91b17f8435b2735951f0d1fad10cc5dfd96db110243370b"}, + {file = "pyzmq-26.0.3-cp37-cp37m-macosx_10_9_x86_64.whl", hash = "sha256:0c0991f5a96a8e620f7691e61178cd8f457b49e17b7d9cfa2067e2a0a89fc1d5"}, + {file = "pyzmq-26.0.3-cp37-cp37m-manylinux_2_12_i686.manylinux2010_i686.whl", hash = "sha256:dbf012d8fcb9f2cf0643b65df3b355fdd74fc0035d70bb5c845e9e30a3a4654b"}, + {file = "pyzmq-26.0.3-cp37-cp37m-manylinux_2_12_x86_64.manylinux2010_x86_64.whl", hash = "sha256:01fbfbeb8249a68d257f601deb50c70c929dc2dfe683b754659569e502fbd3aa"}, + {file = "pyzmq-26.0.3-cp37-cp37m-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:1c8eb19abe87029c18f226d42b8a2c9efdd139d08f8bf6e085dd9075446db450"}, + {file = "pyzmq-26.0.3-cp37-cp37m-musllinux_1_1_aarch64.whl", hash = "sha256:5344b896e79800af86ad643408ca9aa303a017f6ebff8cee5a3163c1e9aec987"}, + {file = "pyzmq-26.0.3-cp37-cp37m-musllinux_1_1_i686.whl", hash = "sha256:204e0f176fd1d067671157d049466869b3ae1fc51e354708b0dc41cf94e23a3a"}, + {file = "pyzmq-26.0.3-cp37-cp37m-musllinux_1_1_x86_64.whl", hash = "sha256:a42db008d58530efa3b881eeee4991146de0b790e095f7ae43ba5cc612decbc5"}, + {file = "pyzmq-26.0.3-cp37-cp37m-win32.whl", hash = "sha256:8d7a498671ca87e32b54cb47c82a92b40130a26c5197d392720a1bce1b3c77cf"}, + {file = "pyzmq-26.0.3-cp37-cp37m-win_amd64.whl", hash = "sha256:3b4032a96410bdc760061b14ed6a33613ffb7f702181ba999df5d16fb96ba16a"}, + {file = "pyzmq-26.0.3-cp38-cp38-macosx_10_15_universal2.whl", hash = "sha256:2cc4e280098c1b192c42a849de8de2c8e0f3a84086a76ec5b07bfee29bda7d18"}, + {file = "pyzmq-26.0.3-cp38-cp38-macosx_10_9_x86_64.whl", hash = "sha256:5bde86a2ed3ce587fa2b207424ce15b9a83a9fa14422dcc1c5356a13aed3df9d"}, + {file = "pyzmq-26.0.3-cp38-cp38-manylinux_2_12_i686.manylinux2010_i686.whl", hash = "sha256:34106f68e20e6ff253c9f596ea50397dbd8699828d55e8fa18bd4323d8d966e6"}, + {file = "pyzmq-26.0.3-cp38-cp38-manylinux_2_12_x86_64.manylinux2010_x86_64.whl", hash = "sha256:ebbbd0e728af5db9b04e56389e2299a57ea8b9dd15c9759153ee2455b32be6ad"}, + {file = "pyzmq-26.0.3-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:f6b1d1c631e5940cac5a0b22c5379c86e8df6a4ec277c7a856b714021ab6cfad"}, + {file = "pyzmq-26.0.3-cp38-cp38-musllinux_1_1_aarch64.whl", hash = "sha256:e891ce81edd463b3b4c3b885c5603c00141151dd9c6936d98a680c8c72fe5c67"}, + {file = "pyzmq-26.0.3-cp38-cp38-musllinux_1_1_i686.whl", hash = "sha256:9b273ecfbc590a1b98f014ae41e5cf723932f3b53ba9367cfb676f838038b32c"}, + {file = "pyzmq-26.0.3-cp38-cp38-musllinux_1_1_x86_64.whl", hash = "sha256:b32bff85fb02a75ea0b68f21e2412255b5731f3f389ed9aecc13a6752f58ac97"}, + {file = "pyzmq-26.0.3-cp38-cp38-win32.whl", hash = "sha256:f6c21c00478a7bea93caaaef9e7629145d4153b15a8653e8bb4609d4bc70dbfc"}, + {file = "pyzmq-26.0.3-cp38-cp38-win_amd64.whl", hash = "sha256:3401613148d93ef0fd9aabdbddb212de3db7a4475367f49f590c837355343972"}, + {file = "pyzmq-26.0.3-cp39-cp39-macosx_10_15_universal2.whl", hash = "sha256:2ed8357f4c6e0daa4f3baf31832df8a33334e0fe5b020a61bc8b345a3db7a606"}, + {file = "pyzmq-26.0.3-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:c1c8f2a2ca45292084c75bb6d3a25545cff0ed931ed228d3a1810ae3758f975f"}, + {file = "pyzmq-26.0.3-cp39-cp39-manylinux_2_12_i686.manylinux2010_i686.whl", hash = "sha256:b63731993cdddcc8e087c64e9cf003f909262b359110070183d7f3025d1c56b5"}, + {file = "pyzmq-26.0.3-cp39-cp39-manylinux_2_12_x86_64.manylinux2010_x86_64.whl", hash = "sha256:b3cd31f859b662ac5d7f4226ec7d8bd60384fa037fc02aee6ff0b53ba29a3ba8"}, + {file = "pyzmq-26.0.3-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:115f8359402fa527cf47708d6f8a0f8234f0e9ca0cab7c18c9c189c194dbf620"}, + {file = "pyzmq-26.0.3-cp39-cp39-musllinux_1_1_aarch64.whl", hash = "sha256:715bdf952b9533ba13dfcf1f431a8f49e63cecc31d91d007bc1deb914f47d0e4"}, + {file = "pyzmq-26.0.3-cp39-cp39-musllinux_1_1_i686.whl", hash = "sha256:e1258c639e00bf5e8a522fec6c3eaa3e30cf1c23a2f21a586be7e04d50c9acab"}, + {file = "pyzmq-26.0.3-cp39-cp39-musllinux_1_1_x86_64.whl", hash = "sha256:15c59e780be8f30a60816a9adab900c12a58d79c1ac742b4a8df044ab2a6d920"}, + {file = "pyzmq-26.0.3-cp39-cp39-win32.whl", hash = "sha256:d0cdde3c78d8ab5b46595054e5def32a755fc028685add5ddc7403e9f6de9879"}, + {file = "pyzmq-26.0.3-cp39-cp39-win_amd64.whl", hash = "sha256:ce828058d482ef860746bf532822842e0ff484e27f540ef5c813d516dd8896d2"}, + {file = "pyzmq-26.0.3-cp39-cp39-win_arm64.whl", hash = "sha256:788f15721c64109cf720791714dc14afd0f449d63f3a5487724f024345067381"}, + {file = "pyzmq-26.0.3-pp310-pypy310_pp73-macosx_10_9_x86_64.whl", hash = "sha256:2c18645ef6294d99b256806e34653e86236eb266278c8ec8112622b61db255de"}, + {file = "pyzmq-26.0.3-pp310-pypy310_pp73-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:7e6bc96ebe49604df3ec2c6389cc3876cabe475e6bfc84ced1bf4e630662cb35"}, + {file = "pyzmq-26.0.3-pp310-pypy310_pp73-manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:971e8990c5cc4ddcff26e149398fc7b0f6a042306e82500f5e8db3b10ce69f84"}, + {file = "pyzmq-26.0.3-pp310-pypy310_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:d8416c23161abd94cc7da80c734ad7c9f5dbebdadfdaa77dad78244457448223"}, + {file = "pyzmq-26.0.3-pp310-pypy310_pp73-win_amd64.whl", hash = "sha256:082a2988364b60bb5de809373098361cf1dbb239623e39e46cb18bc035ed9c0c"}, + {file = "pyzmq-26.0.3-pp37-pypy37_pp73-macosx_10_9_x86_64.whl", hash = "sha256:d57dfbf9737763b3a60d26e6800e02e04284926329aee8fb01049635e957fe81"}, + {file = "pyzmq-26.0.3-pp37-pypy37_pp73-manylinux_2_12_i686.manylinux2010_i686.whl", hash = "sha256:77a85dca4c2430ac04dc2a2185c2deb3858a34fe7f403d0a946fa56970cf60a1"}, + {file = "pyzmq-26.0.3-pp37-pypy37_pp73-manylinux_2_12_x86_64.manylinux2010_x86_64.whl", hash = "sha256:4c82a6d952a1d555bf4be42b6532927d2a5686dd3c3e280e5f63225ab47ac1f5"}, + {file = "pyzmq-26.0.3-pp37-pypy37_pp73-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:4496b1282c70c442809fc1b151977c3d967bfb33e4e17cedbf226d97de18f709"}, + {file = "pyzmq-26.0.3-pp37-pypy37_pp73-win_amd64.whl", hash = "sha256:e4946d6bdb7ba972dfda282f9127e5756d4f299028b1566d1245fa0d438847e6"}, + {file = "pyzmq-26.0.3-pp38-pypy38_pp73-macosx_10_9_x86_64.whl", hash = "sha256:03c0ae165e700364b266876d712acb1ac02693acd920afa67da2ebb91a0b3c09"}, + {file = "pyzmq-26.0.3-pp38-pypy38_pp73-manylinux_2_12_i686.manylinux2010_i686.whl", hash = "sha256:3e3070e680f79887d60feeda051a58d0ac36622e1759f305a41059eff62c6da7"}, + {file = "pyzmq-26.0.3-pp38-pypy38_pp73-manylinux_2_12_x86_64.manylinux2010_x86_64.whl", hash = "sha256:6ca08b840fe95d1c2bd9ab92dac5685f949fc6f9ae820ec16193e5ddf603c3b2"}, + {file = "pyzmq-26.0.3-pp38-pypy38_pp73-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:e76654e9dbfb835b3518f9938e565c7806976c07b37c33526b574cc1a1050480"}, + {file = "pyzmq-26.0.3-pp38-pypy38_pp73-win_amd64.whl", hash = "sha256:871587bdadd1075b112e697173e946a07d722459d20716ceb3d1bd6c64bd08ce"}, + {file = "pyzmq-26.0.3-pp39-pypy39_pp73-macosx_10_9_x86_64.whl", hash = "sha256:d0a2d1bd63a4ad79483049b26514e70fa618ce6115220da9efdff63688808b17"}, + {file = "pyzmq-26.0.3-pp39-pypy39_pp73-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:0270b49b6847f0d106d64b5086e9ad5dc8a902413b5dbbb15d12b60f9c1747a4"}, + {file = "pyzmq-26.0.3-pp39-pypy39_pp73-manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:703c60b9910488d3d0954ca585c34f541e506a091a41930e663a098d3b794c67"}, + {file = "pyzmq-26.0.3-pp39-pypy39_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:74423631b6be371edfbf7eabb02ab995c2563fee60a80a30829176842e71722a"}, + {file = "pyzmq-26.0.3-pp39-pypy39_pp73-manylinux_2_28_x86_64.whl", hash = "sha256:4adfbb5451196842a88fda3612e2c0414134874bffb1c2ce83ab4242ec9e027d"}, + {file = "pyzmq-26.0.3-pp39-pypy39_pp73-win_amd64.whl", hash = "sha256:3516119f4f9b8671083a70b6afaa0a070f5683e431ab3dc26e9215620d7ca1ad"}, + {file = "pyzmq-26.0.3.tar.gz", hash = "sha256:dba7d9f2e047dfa2bca3b01f4f84aa5246725203d6284e3790f2ca15fba6b40a"}, ] [package.dependencies] @@ -3812,13 +3860,13 @@ cffi = {version = "", markers = "implementation_name == \"pypy\""} [[package]] name = "referencing" -version = "0.33.0" +version = "0.35.1" description = "JSON Referencing + Python" optional = false python-versions = ">=3.8" files = [ - {file = "referencing-0.33.0-py3-none-any.whl", hash = "sha256:39240f2ecc770258f28b642dd47fd74bc8b02484de54e1882b74b35ebd779bd5"}, - {file = "referencing-0.33.0.tar.gz", hash = "sha256:c775fedf74bc0f9189c2a3be1c12fd03e8c23f4d371dce795df44e06c5b412f7"}, + {file = "referencing-0.35.1-py3-none-any.whl", hash = "sha256:eda6d3234d62814d1c64e305c1331c9a3a6132da475ab6382eaa997b21ee75de"}, + {file = "referencing-0.35.1.tar.gz", hash = "sha256:25b42124a6c8b632a425174f24087783efb348a6f1e0008e63cd4466fedf703c"}, ] [package.dependencies] @@ -3827,115 +3875,101 @@ rpds-py = ">=0.7.0" [[package]] name = "regex" -version = "2023.12.25" +version = "2024.5.15" description = "Alternative regular expression module, to replace re." optional = false -python-versions = ">=3.7" +python-versions = ">=3.8" files = [ - {file = "regex-2023.12.25-cp310-cp310-macosx_10_9_universal2.whl", hash = "sha256:0694219a1d54336fd0445ea382d49d36882415c0134ee1e8332afd1529f0baa5"}, - {file = "regex-2023.12.25-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:b014333bd0217ad3d54c143de9d4b9a3ca1c5a29a6d0d554952ea071cff0f1f8"}, - {file = "regex-2023.12.25-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:d865984b3f71f6d0af64d0d88f5733521698f6c16f445bb09ce746c92c97c586"}, - {file = "regex-2023.12.25-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:1e0eabac536b4cc7f57a5f3d095bfa557860ab912f25965e08fe1545e2ed8b4c"}, - {file = "regex-2023.12.25-cp310-cp310-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:c25a8ad70e716f96e13a637802813f65d8a6760ef48672aa3502f4c24ea8b400"}, - {file = "regex-2023.12.25-cp310-cp310-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:a9b6d73353f777630626f403b0652055ebfe8ff142a44ec2cf18ae470395766e"}, - {file = "regex-2023.12.25-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:a9cc99d6946d750eb75827cb53c4371b8b0fe89c733a94b1573c9dd16ea6c9e4"}, - {file = "regex-2023.12.25-cp310-cp310-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:88d1f7bef20c721359d8675f7d9f8e414ec5003d8f642fdfd8087777ff7f94b5"}, - {file = "regex-2023.12.25-cp310-cp310-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_12_x86_64.manylinux2010_x86_64.whl", hash = "sha256:cb3fe77aec8f1995611f966d0c656fdce398317f850d0e6e7aebdfe61f40e1cd"}, - {file = "regex-2023.12.25-cp310-cp310-musllinux_1_1_aarch64.whl", hash = "sha256:7aa47c2e9ea33a4a2a05f40fcd3ea36d73853a2aae7b4feab6fc85f8bf2c9704"}, - {file = "regex-2023.12.25-cp310-cp310-musllinux_1_1_i686.whl", hash = "sha256:df26481f0c7a3f8739fecb3e81bc9da3fcfae34d6c094563b9d4670b047312e1"}, - {file = "regex-2023.12.25-cp310-cp310-musllinux_1_1_ppc64le.whl", hash = "sha256:c40281f7d70baf6e0db0c2f7472b31609f5bc2748fe7275ea65a0b4601d9b392"}, - {file = "regex-2023.12.25-cp310-cp310-musllinux_1_1_s390x.whl", hash = "sha256:d94a1db462d5690ebf6ae86d11c5e420042b9898af5dcf278bd97d6bda065423"}, - {file = "regex-2023.12.25-cp310-cp310-musllinux_1_1_x86_64.whl", hash = "sha256:ba1b30765a55acf15dce3f364e4928b80858fa8f979ad41f862358939bdd1f2f"}, - {file = "regex-2023.12.25-cp310-cp310-win32.whl", hash = "sha256:150c39f5b964e4d7dba46a7962a088fbc91f06e606f023ce57bb347a3b2d4630"}, - {file = "regex-2023.12.25-cp310-cp310-win_amd64.whl", hash = "sha256:09da66917262d9481c719599116c7dc0c321ffcec4b1f510c4f8a066f8768105"}, - {file = "regex-2023.12.25-cp311-cp311-macosx_10_9_universal2.whl", hash = "sha256:1b9d811f72210fa9306aeb88385b8f8bcef0dfbf3873410413c00aa94c56c2b6"}, - {file = "regex-2023.12.25-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:d902a43085a308cef32c0d3aea962524b725403fd9373dea18110904003bac97"}, - {file = "regex-2023.12.25-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:d166eafc19f4718df38887b2bbe1467a4f74a9830e8605089ea7a30dd4da8887"}, - {file = "regex-2023.12.25-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:c7ad32824b7f02bb3c9f80306d405a1d9b7bb89362d68b3c5a9be53836caebdb"}, - {file = "regex-2023.12.25-cp311-cp311-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:636ba0a77de609d6510235b7f0e77ec494d2657108f777e8765efc060094c98c"}, - {file = "regex-2023.12.25-cp311-cp311-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:0fda75704357805eb953a3ee15a2b240694a9a514548cd49b3c5124b4e2ad01b"}, - {file = "regex-2023.12.25-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:f72cbae7f6b01591f90814250e636065850c5926751af02bb48da94dfced7baa"}, - {file = "regex-2023.12.25-cp311-cp311-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:db2a0b1857f18b11e3b0e54ddfefc96af46b0896fb678c85f63fb8c37518b3e7"}, - {file = "regex-2023.12.25-cp311-cp311-musllinux_1_1_aarch64.whl", hash = "sha256:7502534e55c7c36c0978c91ba6f61703faf7ce733715ca48f499d3dbbd7657e0"}, - {file = "regex-2023.12.25-cp311-cp311-musllinux_1_1_i686.whl", hash = "sha256:e8c7e08bb566de4faaf11984af13f6bcf6a08f327b13631d41d62592681d24fe"}, - {file = "regex-2023.12.25-cp311-cp311-musllinux_1_1_ppc64le.whl", hash = "sha256:283fc8eed679758de38fe493b7d7d84a198b558942b03f017b1f94dda8efae80"}, - {file = "regex-2023.12.25-cp311-cp311-musllinux_1_1_s390x.whl", hash = "sha256:f44dd4d68697559d007462b0a3a1d9acd61d97072b71f6d1968daef26bc744bd"}, - {file = "regex-2023.12.25-cp311-cp311-musllinux_1_1_x86_64.whl", hash = "sha256:67d3ccfc590e5e7197750fcb3a2915b416a53e2de847a728cfa60141054123d4"}, - {file = "regex-2023.12.25-cp311-cp311-win32.whl", hash = "sha256:68191f80a9bad283432385961d9efe09d783bcd36ed35a60fb1ff3f1ec2efe87"}, - {file = "regex-2023.12.25-cp311-cp311-win_amd64.whl", hash = "sha256:7d2af3f6b8419661a0c421584cfe8aaec1c0e435ce7e47ee2a97e344b98f794f"}, - {file = "regex-2023.12.25-cp312-cp312-macosx_10_9_universal2.whl", hash = "sha256:8a0ccf52bb37d1a700375a6b395bff5dd15c50acb745f7db30415bae3c2b0715"}, - {file = "regex-2023.12.25-cp312-cp312-macosx_10_9_x86_64.whl", hash = "sha256:c3c4a78615b7762740531c27cf46e2f388d8d727d0c0c739e72048beb26c8a9d"}, - {file = "regex-2023.12.25-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:ad83e7545b4ab69216cef4cc47e344d19622e28aabec61574b20257c65466d6a"}, - {file = "regex-2023.12.25-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:b7a635871143661feccce3979e1727c4e094f2bdfd3ec4b90dfd4f16f571a87a"}, - {file = "regex-2023.12.25-cp312-cp312-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:d498eea3f581fbe1b34b59c697512a8baef88212f92e4c7830fcc1499f5b45a5"}, - {file = "regex-2023.12.25-cp312-cp312-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:43f7cd5754d02a56ae4ebb91b33461dc67be8e3e0153f593c509e21d219c5060"}, - {file = "regex-2023.12.25-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:51f4b32f793812714fd5307222a7f77e739b9bc566dc94a18126aba3b92b98a3"}, - {file = "regex-2023.12.25-cp312-cp312-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:ba99d8077424501b9616b43a2d208095746fb1284fc5ba490139651f971d39d9"}, - {file = "regex-2023.12.25-cp312-cp312-musllinux_1_1_aarch64.whl", hash = "sha256:4bfc2b16e3ba8850e0e262467275dd4d62f0d045e0e9eda2bc65078c0110a11f"}, - {file = "regex-2023.12.25-cp312-cp312-musllinux_1_1_i686.whl", hash = "sha256:8c2c19dae8a3eb0ea45a8448356ed561be843b13cbc34b840922ddf565498c1c"}, - {file = "regex-2023.12.25-cp312-cp312-musllinux_1_1_ppc64le.whl", hash = "sha256:60080bb3d8617d96f0fb7e19796384cc2467447ef1c491694850ebd3670bc457"}, - {file = "regex-2023.12.25-cp312-cp312-musllinux_1_1_s390x.whl", hash = "sha256:b77e27b79448e34c2c51c09836033056a0547aa360c45eeeb67803da7b0eedaf"}, - {file = "regex-2023.12.25-cp312-cp312-musllinux_1_1_x86_64.whl", hash = "sha256:518440c991f514331f4850a63560321f833979d145d7d81186dbe2f19e27ae3d"}, - {file = "regex-2023.12.25-cp312-cp312-win32.whl", hash = "sha256:e2610e9406d3b0073636a3a2e80db05a02f0c3169b5632022b4e81c0364bcda5"}, - {file = "regex-2023.12.25-cp312-cp312-win_amd64.whl", hash = "sha256:cc37b9aeebab425f11f27e5e9e6cf580be7206c6582a64467a14dda211abc232"}, - {file = "regex-2023.12.25-cp37-cp37m-macosx_10_9_x86_64.whl", hash = "sha256:da695d75ac97cb1cd725adac136d25ca687da4536154cdc2815f576e4da11c69"}, - {file = "regex-2023.12.25-cp37-cp37m-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:d126361607b33c4eb7b36debc173bf25d7805847346dd4d99b5499e1fef52bc7"}, - {file = "regex-2023.12.25-cp37-cp37m-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:4719bb05094d7d8563a450cf8738d2e1061420f79cfcc1fa7f0a44744c4d8f73"}, - {file = "regex-2023.12.25-cp37-cp37m-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:5dd58946bce44b53b06d94aa95560d0b243eb2fe64227cba50017a8d8b3cd3e2"}, - {file = "regex-2023.12.25-cp37-cp37m-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:22a86d9fff2009302c440b9d799ef2fe322416d2d58fc124b926aa89365ec482"}, - {file = "regex-2023.12.25-cp37-cp37m-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:2aae8101919e8aa05ecfe6322b278f41ce2994c4a430303c4cd163fef746e04f"}, - {file = "regex-2023.12.25-cp37-cp37m-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_12_x86_64.manylinux2010_x86_64.whl", hash = "sha256:e692296c4cc2873967771345a876bcfc1c547e8dd695c6b89342488b0ea55cd8"}, - {file = "regex-2023.12.25-cp37-cp37m-musllinux_1_1_aarch64.whl", hash = "sha256:263ef5cc10979837f243950637fffb06e8daed7f1ac1e39d5910fd29929e489a"}, - {file = "regex-2023.12.25-cp37-cp37m-musllinux_1_1_i686.whl", hash = "sha256:d6f7e255e5fa94642a0724e35406e6cb7001c09d476ab5fce002f652b36d0c39"}, - {file = "regex-2023.12.25-cp37-cp37m-musllinux_1_1_ppc64le.whl", hash = "sha256:88ad44e220e22b63b0f8f81f007e8abbb92874d8ced66f32571ef8beb0643b2b"}, - {file = "regex-2023.12.25-cp37-cp37m-musllinux_1_1_s390x.whl", hash = "sha256:3a17d3ede18f9cedcbe23d2daa8a2cd6f59fe2bf082c567e43083bba3fb00347"}, - {file = "regex-2023.12.25-cp37-cp37m-musllinux_1_1_x86_64.whl", hash = "sha256:d15b274f9e15b1a0b7a45d2ac86d1f634d983ca40d6b886721626c47a400bf39"}, - {file = "regex-2023.12.25-cp37-cp37m-win32.whl", hash = "sha256:ed19b3a05ae0c97dd8f75a5d8f21f7723a8c33bbc555da6bbe1f96c470139d3c"}, - {file = "regex-2023.12.25-cp37-cp37m-win_amd64.whl", hash = "sha256:a6d1047952c0b8104a1d371f88f4ab62e6275567d4458c1e26e9627ad489b445"}, - {file = "regex-2023.12.25-cp38-cp38-macosx_10_9_universal2.whl", hash = "sha256:b43523d7bc2abd757119dbfb38af91b5735eea45537ec6ec3a5ec3f9562a1c53"}, - {file = "regex-2023.12.25-cp38-cp38-macosx_10_9_x86_64.whl", hash = "sha256:efb2d82f33b2212898f1659fb1c2e9ac30493ac41e4d53123da374c3b5541e64"}, - {file = "regex-2023.12.25-cp38-cp38-macosx_11_0_arm64.whl", hash = "sha256:b7fca9205b59c1a3d5031f7e64ed627a1074730a51c2a80e97653e3e9fa0d415"}, - {file = "regex-2023.12.25-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:086dd15e9435b393ae06f96ab69ab2d333f5d65cbe65ca5a3ef0ec9564dfe770"}, - {file = "regex-2023.12.25-cp38-cp38-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:e81469f7d01efed9b53740aedd26085f20d49da65f9c1f41e822a33992cb1590"}, - {file = "regex-2023.12.25-cp38-cp38-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:34e4af5b27232f68042aa40a91c3b9bb4da0eeb31b7632e0091afc4310afe6cb"}, - {file = "regex-2023.12.25-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:9852b76ab558e45b20bf1893b59af64a28bd3820b0c2efc80e0a70a4a3ea51c1"}, - {file = "regex-2023.12.25-cp38-cp38-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:ff100b203092af77d1a5a7abe085b3506b7eaaf9abf65b73b7d6905b6cb76988"}, - {file = "regex-2023.12.25-cp38-cp38-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_12_x86_64.manylinux2010_x86_64.whl", hash = "sha256:cc038b2d8b1470364b1888a98fd22d616fba2b6309c5b5f181ad4483e0017861"}, - {file = "regex-2023.12.25-cp38-cp38-musllinux_1_1_aarch64.whl", hash = "sha256:094ba386bb5c01e54e14434d4caabf6583334090865b23ef58e0424a6286d3dc"}, - {file = "regex-2023.12.25-cp38-cp38-musllinux_1_1_i686.whl", hash = "sha256:5cd05d0f57846d8ba4b71d9c00f6f37d6b97d5e5ef8b3c3840426a475c8f70f4"}, - {file = "regex-2023.12.25-cp38-cp38-musllinux_1_1_ppc64le.whl", hash = "sha256:9aa1a67bbf0f957bbe096375887b2505f5d8ae16bf04488e8b0f334c36e31360"}, - {file = "regex-2023.12.25-cp38-cp38-musllinux_1_1_s390x.whl", hash = "sha256:98a2636994f943b871786c9e82bfe7883ecdaba2ef5df54e1450fa9869d1f756"}, - {file = "regex-2023.12.25-cp38-cp38-musllinux_1_1_x86_64.whl", hash = "sha256:37f8e93a81fc5e5bd8db7e10e62dc64261bcd88f8d7e6640aaebe9bc180d9ce2"}, - {file = "regex-2023.12.25-cp38-cp38-win32.whl", hash = "sha256:d78bd484930c1da2b9679290a41cdb25cc127d783768a0369d6b449e72f88beb"}, - {file = "regex-2023.12.25-cp38-cp38-win_amd64.whl", hash = "sha256:b521dcecebc5b978b447f0f69b5b7f3840eac454862270406a39837ffae4e697"}, - {file = "regex-2023.12.25-cp39-cp39-macosx_10_9_universal2.whl", hash = "sha256:f7bc09bc9c29ebead055bcba136a67378f03d66bf359e87d0f7c759d6d4ffa31"}, - {file = "regex-2023.12.25-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:e14b73607d6231f3cc4622809c196b540a6a44e903bcfad940779c80dffa7be7"}, - {file = "regex-2023.12.25-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:9eda5f7a50141291beda3edd00abc2d4a5b16c29c92daf8d5bd76934150f3edc"}, - {file = "regex-2023.12.25-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:cc6bb9aa69aacf0f6032c307da718f61a40cf970849e471254e0e91c56ffca95"}, - {file = "regex-2023.12.25-cp39-cp39-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:298dc6354d414bc921581be85695d18912bea163a8b23cac9a2562bbcd5088b1"}, - {file = "regex-2023.12.25-cp39-cp39-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:2f4e475a80ecbd15896a976aa0b386c5525d0ed34d5c600b6d3ebac0a67c7ddf"}, - {file = "regex-2023.12.25-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:531ac6cf22b53e0696f8e1d56ce2396311254eb806111ddd3922c9d937151dae"}, - {file = "regex-2023.12.25-cp39-cp39-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:22f3470f7524b6da61e2020672df2f3063676aff444db1daa283c2ea4ed259d6"}, - {file = "regex-2023.12.25-cp39-cp39-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_12_x86_64.manylinux2010_x86_64.whl", hash = "sha256:89723d2112697feaa320c9d351e5f5e7b841e83f8b143dba8e2d2b5f04e10923"}, - {file = "regex-2023.12.25-cp39-cp39-musllinux_1_1_aarch64.whl", hash = "sha256:0ecf44ddf9171cd7566ef1768047f6e66975788258b1c6c6ca78098b95cf9a3d"}, - {file = "regex-2023.12.25-cp39-cp39-musllinux_1_1_i686.whl", hash = "sha256:905466ad1702ed4acfd67a902af50b8db1feeb9781436372261808df7a2a7bca"}, - {file = "regex-2023.12.25-cp39-cp39-musllinux_1_1_ppc64le.whl", hash = "sha256:4558410b7a5607a645e9804a3e9dd509af12fb72b9825b13791a37cd417d73a5"}, - {file = "regex-2023.12.25-cp39-cp39-musllinux_1_1_s390x.whl", hash = "sha256:7e316026cc1095f2a3e8cc012822c99f413b702eaa2ca5408a513609488cb62f"}, - {file = "regex-2023.12.25-cp39-cp39-musllinux_1_1_x86_64.whl", hash = "sha256:3b1de218d5375cd6ac4b5493e0b9f3df2be331e86520f23382f216c137913d20"}, - {file = "regex-2023.12.25-cp39-cp39-win32.whl", hash = "sha256:11a963f8e25ab5c61348d090bf1b07f1953929c13bd2309a0662e9ff680763c9"}, - {file = "regex-2023.12.25-cp39-cp39-win_amd64.whl", hash = "sha256:e693e233ac92ba83a87024e1d32b5f9ab15ca55ddd916d878146f4e3406b5c91"}, - {file = "regex-2023.12.25.tar.gz", hash = "sha256:29171aa128da69afdf4bde412d5bedc335f2ca8fcfe4489038577d05f16181e5"}, + {file = "regex-2024.5.15-cp310-cp310-macosx_10_9_universal2.whl", hash = "sha256:a81e3cfbae20378d75185171587cbf756015ccb14840702944f014e0d93ea09f"}, + {file = "regex-2024.5.15-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:7b59138b219ffa8979013be7bc85bb60c6f7b7575df3d56dc1e403a438c7a3f6"}, + {file = "regex-2024.5.15-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:a0bd000c6e266927cb7a1bc39d55be95c4b4f65c5be53e659537537e019232b1"}, + {file = "regex-2024.5.15-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:5eaa7ddaf517aa095fa8da0b5015c44d03da83f5bd49c87961e3c997daed0de7"}, + {file = "regex-2024.5.15-cp310-cp310-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:ba68168daedb2c0bab7fd7e00ced5ba90aebf91024dea3c88ad5063c2a562cca"}, + {file = "regex-2024.5.15-cp310-cp310-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:6e8d717bca3a6e2064fc3a08df5cbe366369f4b052dcd21b7416e6d71620dca1"}, + {file = "regex-2024.5.15-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:1337b7dbef9b2f71121cdbf1e97e40de33ff114801263b275aafd75303bd62b5"}, + {file = "regex-2024.5.15-cp310-cp310-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:f9ebd0a36102fcad2f03696e8af4ae682793a5d30b46c647eaf280d6cfb32796"}, + {file = "regex-2024.5.15-cp310-cp310-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_12_x86_64.manylinux2010_x86_64.whl", hash = "sha256:9efa1a32ad3a3ea112224897cdaeb6aa00381627f567179c0314f7b65d354c62"}, + {file = "regex-2024.5.15-cp310-cp310-musllinux_1_2_aarch64.whl", hash = "sha256:1595f2d10dff3d805e054ebdc41c124753631b6a471b976963c7b28543cf13b0"}, + {file = "regex-2024.5.15-cp310-cp310-musllinux_1_2_i686.whl", hash = "sha256:b802512f3e1f480f41ab5f2cfc0e2f761f08a1f41092d6718868082fc0d27143"}, + {file = "regex-2024.5.15-cp310-cp310-musllinux_1_2_ppc64le.whl", hash = "sha256:a0981022dccabca811e8171f913de05720590c915b033b7e601f35ce4ea7019f"}, + {file = "regex-2024.5.15-cp310-cp310-musllinux_1_2_s390x.whl", hash = "sha256:19068a6a79cf99a19ccefa44610491e9ca02c2be3305c7760d3831d38a467a6f"}, + {file = "regex-2024.5.15-cp310-cp310-musllinux_1_2_x86_64.whl", hash = "sha256:1b5269484f6126eee5e687785e83c6b60aad7663dafe842b34691157e5083e53"}, + {file = "regex-2024.5.15-cp310-cp310-win32.whl", hash = "sha256:ada150c5adfa8fbcbf321c30c751dc67d2f12f15bd183ffe4ec7cde351d945b3"}, + {file = "regex-2024.5.15-cp310-cp310-win_amd64.whl", hash = "sha256:ac394ff680fc46b97487941f5e6ae49a9f30ea41c6c6804832063f14b2a5a145"}, + {file = "regex-2024.5.15-cp311-cp311-macosx_10_9_universal2.whl", hash = "sha256:f5b1dff3ad008dccf18e652283f5e5339d70bf8ba7c98bf848ac33db10f7bc7a"}, + {file = "regex-2024.5.15-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:c6a2b494a76983df8e3d3feea9b9ffdd558b247e60b92f877f93a1ff43d26656"}, + {file = "regex-2024.5.15-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:a32b96f15c8ab2e7d27655969a23895eb799de3665fa94349f3b2fbfd547236f"}, + {file = "regex-2024.5.15-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:10002e86e6068d9e1c91eae8295ef690f02f913c57db120b58fdd35a6bb1af35"}, + {file = "regex-2024.5.15-cp311-cp311-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:ec54d5afa89c19c6dd8541a133be51ee1017a38b412b1321ccb8d6ddbeb4cf7d"}, + {file = "regex-2024.5.15-cp311-cp311-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:10e4ce0dca9ae7a66e6089bb29355d4432caed736acae36fef0fdd7879f0b0cb"}, + {file = "regex-2024.5.15-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:3e507ff1e74373c4d3038195fdd2af30d297b4f0950eeda6f515ae3d84a1770f"}, + {file = "regex-2024.5.15-cp311-cp311-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:d1f059a4d795e646e1c37665b9d06062c62d0e8cc3c511fe01315973a6542e40"}, + {file = "regex-2024.5.15-cp311-cp311-musllinux_1_2_aarch64.whl", hash = "sha256:0721931ad5fe0dda45d07f9820b90b2148ccdd8e45bb9e9b42a146cb4f695649"}, + {file = "regex-2024.5.15-cp311-cp311-musllinux_1_2_i686.whl", hash = "sha256:833616ddc75ad595dee848ad984d067f2f31be645d603e4d158bba656bbf516c"}, + {file = "regex-2024.5.15-cp311-cp311-musllinux_1_2_ppc64le.whl", hash = "sha256:287eb7f54fc81546346207c533ad3c2c51a8d61075127d7f6d79aaf96cdee890"}, + {file = "regex-2024.5.15-cp311-cp311-musllinux_1_2_s390x.whl", hash = "sha256:19dfb1c504781a136a80ecd1fff9f16dddf5bb43cec6871778c8a907a085bb3d"}, + {file = "regex-2024.5.15-cp311-cp311-musllinux_1_2_x86_64.whl", hash = "sha256:119af6e56dce35e8dfb5222573b50c89e5508d94d55713c75126b753f834de68"}, + {file = "regex-2024.5.15-cp311-cp311-win32.whl", hash = "sha256:1c1c174d6ec38d6c8a7504087358ce9213d4332f6293a94fbf5249992ba54efa"}, + {file = "regex-2024.5.15-cp311-cp311-win_amd64.whl", hash = "sha256:9e717956dcfd656f5055cc70996ee2cc82ac5149517fc8e1b60261b907740201"}, + {file = "regex-2024.5.15-cp312-cp312-macosx_10_9_universal2.whl", hash = "sha256:632b01153e5248c134007209b5c6348a544ce96c46005d8456de1d552455b014"}, + {file = "regex-2024.5.15-cp312-cp312-macosx_10_9_x86_64.whl", hash = "sha256:e64198f6b856d48192bf921421fdd8ad8eb35e179086e99e99f711957ffedd6e"}, + {file = "regex-2024.5.15-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:68811ab14087b2f6e0fc0c2bae9ad689ea3584cad6917fc57be6a48bbd012c49"}, + {file = "regex-2024.5.15-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:f8ec0c2fea1e886a19c3bee0cd19d862b3aa75dcdfb42ebe8ed30708df64687a"}, + {file = "regex-2024.5.15-cp312-cp312-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:d0c0c0003c10f54a591d220997dd27d953cd9ccc1a7294b40a4be5312be8797b"}, + {file = "regex-2024.5.15-cp312-cp312-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:2431b9e263af1953c55abbd3e2efca67ca80a3de8a0437cb58e2421f8184717a"}, + {file = "regex-2024.5.15-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:4a605586358893b483976cffc1723fb0f83e526e8f14c6e6614e75919d9862cf"}, + {file = "regex-2024.5.15-cp312-cp312-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:391d7f7f1e409d192dba8bcd42d3e4cf9e598f3979cdaed6ab11288da88cb9f2"}, + {file = "regex-2024.5.15-cp312-cp312-musllinux_1_2_aarch64.whl", hash = "sha256:9ff11639a8d98969c863d4617595eb5425fd12f7c5ef6621a4b74b71ed8726d5"}, + {file = "regex-2024.5.15-cp312-cp312-musllinux_1_2_i686.whl", hash = "sha256:4eee78a04e6c67e8391edd4dad3279828dd66ac4b79570ec998e2155d2e59fd5"}, + {file = "regex-2024.5.15-cp312-cp312-musllinux_1_2_ppc64le.whl", hash = "sha256:8fe45aa3f4aa57faabbc9cb46a93363edd6197cbc43523daea044e9ff2fea83e"}, + {file = "regex-2024.5.15-cp312-cp312-musllinux_1_2_s390x.whl", hash = "sha256:d0a3d8d6acf0c78a1fff0e210d224b821081330b8524e3e2bc5a68ef6ab5803d"}, + {file = "regex-2024.5.15-cp312-cp312-musllinux_1_2_x86_64.whl", hash = "sha256:c486b4106066d502495b3025a0a7251bf37ea9540433940a23419461ab9f2a80"}, + {file = "regex-2024.5.15-cp312-cp312-win32.whl", hash = "sha256:c49e15eac7c149f3670b3e27f1f28a2c1ddeccd3a2812cba953e01be2ab9b5fe"}, + {file = "regex-2024.5.15-cp312-cp312-win_amd64.whl", hash = "sha256:673b5a6da4557b975c6c90198588181029c60793835ce02f497ea817ff647cb2"}, + {file = "regex-2024.5.15-cp38-cp38-macosx_10_9_universal2.whl", hash = "sha256:87e2a9c29e672fc65523fb47a90d429b70ef72b901b4e4b1bd42387caf0d6835"}, + {file = "regex-2024.5.15-cp38-cp38-macosx_10_9_x86_64.whl", hash = "sha256:c3bea0ba8b73b71b37ac833a7f3fd53825924165da6a924aec78c13032f20850"}, + {file = "regex-2024.5.15-cp38-cp38-macosx_11_0_arm64.whl", hash = "sha256:bfc4f82cabe54f1e7f206fd3d30fda143f84a63fe7d64a81558d6e5f2e5aaba9"}, + {file = "regex-2024.5.15-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:e5bb9425fe881d578aeca0b2b4b3d314ec88738706f66f219c194d67179337cb"}, + {file = "regex-2024.5.15-cp38-cp38-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:64c65783e96e563103d641760664125e91bd85d8e49566ee560ded4da0d3e704"}, + {file = "regex-2024.5.15-cp38-cp38-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:cf2430df4148b08fb4324b848672514b1385ae3807651f3567871f130a728cc3"}, + {file = "regex-2024.5.15-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:5397de3219a8b08ae9540c48f602996aa6b0b65d5a61683e233af8605c42b0f2"}, + {file = "regex-2024.5.15-cp38-cp38-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:455705d34b4154a80ead722f4f185b04c4237e8e8e33f265cd0798d0e44825fa"}, + {file = "regex-2024.5.15-cp38-cp38-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_12_x86_64.manylinux2010_x86_64.whl", hash = "sha256:b2b6f1b3bb6f640c1a92be3bbfbcb18657b125b99ecf141fb3310b5282c7d4ed"}, + {file = "regex-2024.5.15-cp38-cp38-musllinux_1_2_aarch64.whl", hash = "sha256:3ad070b823ca5890cab606c940522d05d3d22395d432f4aaaf9d5b1653e47ced"}, + {file = "regex-2024.5.15-cp38-cp38-musllinux_1_2_i686.whl", hash = "sha256:5b5467acbfc153847d5adb21e21e29847bcb5870e65c94c9206d20eb4e99a384"}, + {file = "regex-2024.5.15-cp38-cp38-musllinux_1_2_ppc64le.whl", hash = "sha256:e6662686aeb633ad65be2a42b4cb00178b3fbf7b91878f9446075c404ada552f"}, + {file = "regex-2024.5.15-cp38-cp38-musllinux_1_2_s390x.whl", hash = "sha256:2b4c884767504c0e2401babe8b5b7aea9148680d2e157fa28f01529d1f7fcf67"}, + {file = "regex-2024.5.15-cp38-cp38-musllinux_1_2_x86_64.whl", hash = "sha256:3cd7874d57f13bf70078f1ff02b8b0aa48d5b9ed25fc48547516c6aba36f5741"}, + {file = "regex-2024.5.15-cp38-cp38-win32.whl", hash = "sha256:e4682f5ba31f475d58884045c1a97a860a007d44938c4c0895f41d64481edbc9"}, + {file = "regex-2024.5.15-cp38-cp38-win_amd64.whl", hash = "sha256:d99ceffa25ac45d150e30bd9ed14ec6039f2aad0ffa6bb87a5936f5782fc1569"}, + {file = "regex-2024.5.15-cp39-cp39-macosx_10_9_universal2.whl", hash = "sha256:13cdaf31bed30a1e1c2453ef6015aa0983e1366fad2667657dbcac7b02f67133"}, + {file = "regex-2024.5.15-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:cac27dcaa821ca271855a32188aa61d12decb6fe45ffe3e722401fe61e323cd1"}, + {file = "regex-2024.5.15-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:7dbe2467273b875ea2de38ded4eba86cbcbc9a1a6d0aa11dcf7bd2e67859c435"}, + {file = "regex-2024.5.15-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:64f18a9a3513a99c4bef0e3efd4c4a5b11228b48aa80743be822b71e132ae4f5"}, + {file = "regex-2024.5.15-cp39-cp39-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:d347a741ea871c2e278fde6c48f85136c96b8659b632fb57a7d1ce1872547600"}, + {file = "regex-2024.5.15-cp39-cp39-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:1878b8301ed011704aea4c806a3cadbd76f84dece1ec09cc9e4dc934cfa5d4da"}, + {file = "regex-2024.5.15-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:4babf07ad476aaf7830d77000874d7611704a7fcf68c9c2ad151f5d94ae4bfc4"}, + {file = "regex-2024.5.15-cp39-cp39-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:35cb514e137cb3488bce23352af3e12fb0dbedd1ee6e60da053c69fb1b29cc6c"}, + {file = "regex-2024.5.15-cp39-cp39-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_12_x86_64.manylinux2010_x86_64.whl", hash = "sha256:cdd09d47c0b2efee9378679f8510ee6955d329424c659ab3c5e3a6edea696294"}, + {file = "regex-2024.5.15-cp39-cp39-musllinux_1_2_aarch64.whl", hash = "sha256:72d7a99cd6b8f958e85fc6ca5b37c4303294954eac1376535b03c2a43eb72629"}, + {file = "regex-2024.5.15-cp39-cp39-musllinux_1_2_i686.whl", hash = "sha256:a094801d379ab20c2135529948cb84d417a2169b9bdceda2a36f5f10977ebc16"}, + {file = "regex-2024.5.15-cp39-cp39-musllinux_1_2_ppc64le.whl", hash = "sha256:c0c18345010870e58238790a6779a1219b4d97bd2e77e1140e8ee5d14df071aa"}, + {file = "regex-2024.5.15-cp39-cp39-musllinux_1_2_s390x.whl", hash = "sha256:16093f563098448ff6b1fa68170e4acbef94e6b6a4e25e10eae8598bb1694b5d"}, + {file = "regex-2024.5.15-cp39-cp39-musllinux_1_2_x86_64.whl", hash = "sha256:e38a7d4e8f633a33b4c7350fbd8bad3b70bf81439ac67ac38916c4a86b465456"}, + {file = "regex-2024.5.15-cp39-cp39-win32.whl", hash = "sha256:71a455a3c584a88f654b64feccc1e25876066c4f5ef26cd6dd711308aa538694"}, + {file = "regex-2024.5.15-cp39-cp39-win_amd64.whl", hash = "sha256:cab12877a9bdafde5500206d1020a584355a97884dfd388af3699e9137bf7388"}, + {file = "regex-2024.5.15.tar.gz", hash = "sha256:d3ee02d9e5f482cc8309134a91eeaacbdd2261ba111b0fef3748eeb4913e6a2c"}, ] [[package]] name = "requests" -version = "2.31.0" +version = "2.32.2" description = "Python HTTP for Humans." optional = false -python-versions = ">=3.7" +python-versions = ">=3.8" files = [ - {file = "requests-2.31.0-py3-none-any.whl", hash = "sha256:58cd2187c01e70e6e26505bca751777aa9f2ee0b7f4300988b709f44e013003f"}, - {file = "requests-2.31.0.tar.gz", hash = "sha256:942c5a758f98d790eaed1a29cb6eefc7ffb0d1cf7af05c3d2791656dbd6ad1e1"}, + {file = "requests-2.32.2-py3-none-any.whl", hash = "sha256:fc06670dd0ed212426dfeb94fc1b983d917c4f9847c863f313c9dfaaffb7c23c"}, + {file = "requests-2.32.2.tar.gz", hash = "sha256:dd951ff5ecf3e3b3aa26b40703ba77495dab41da839ae72ef3c8e5d8e2433289"}, ] [package.dependencies] @@ -3950,222 +3984,206 @@ use-chardet-on-py3 = ["chardet (>=3.0.2,<6)"] [[package]] name = "rpds-py" -version = "0.17.1" +version = "0.18.1" description = "Python bindings to Rust's persistent data structures (rpds)" optional = false python-versions = ">=3.8" files = [ - {file = "rpds_py-0.17.1-cp310-cp310-macosx_10_12_x86_64.whl", hash = "sha256:4128980a14ed805e1b91a7ed551250282a8ddf8201a4e9f8f5b7e6225f54170d"}, - {file = "rpds_py-0.17.1-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:ff1dcb8e8bc2261a088821b2595ef031c91d499a0c1b031c152d43fe0a6ecec8"}, - {file = "rpds_py-0.17.1-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:d65e6b4f1443048eb7e833c2accb4fa7ee67cc7d54f31b4f0555b474758bee55"}, - {file = "rpds_py-0.17.1-cp310-cp310-manylinux_2_17_armv7l.manylinux2014_armv7l.whl", hash = "sha256:a71169d505af63bb4d20d23a8fbd4c6ce272e7bce6cc31f617152aa784436f29"}, - {file = "rpds_py-0.17.1-cp310-cp310-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:436474f17733c7dca0fbf096d36ae65277e8645039df12a0fa52445ca494729d"}, - {file = "rpds_py-0.17.1-cp310-cp310-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:10162fe3f5f47c37ebf6d8ff5a2368508fe22007e3077bf25b9c7d803454d921"}, - {file = "rpds_py-0.17.1-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:720215373a280f78a1814becb1312d4e4d1077b1202a56d2b0815e95ccb99ce9"}, - {file = "rpds_py-0.17.1-cp310-cp310-manylinux_2_5_i686.manylinux1_i686.whl", hash = "sha256:70fcc6c2906cfa5c6a552ba7ae2ce64b6c32f437d8f3f8eea49925b278a61453"}, - {file = "rpds_py-0.17.1-cp310-cp310-musllinux_1_2_aarch64.whl", hash = "sha256:91e5a8200e65aaac342a791272c564dffcf1281abd635d304d6c4e6b495f29dc"}, - {file = "rpds_py-0.17.1-cp310-cp310-musllinux_1_2_i686.whl", hash = "sha256:99f567dae93e10be2daaa896e07513dd4bf9c2ecf0576e0533ac36ba3b1d5394"}, - {file = "rpds_py-0.17.1-cp310-cp310-musllinux_1_2_x86_64.whl", hash = "sha256:24e4900a6643f87058a27320f81336d527ccfe503984528edde4bb660c8c8d59"}, - {file = "rpds_py-0.17.1-cp310-none-win32.whl", hash = "sha256:0bfb09bf41fe7c51413f563373e5f537eaa653d7adc4830399d4e9bdc199959d"}, - {file = "rpds_py-0.17.1-cp310-none-win_amd64.whl", hash = "sha256:20de7b7179e2031a04042e85dc463a93a82bc177eeba5ddd13ff746325558aa6"}, - {file = "rpds_py-0.17.1-cp311-cp311-macosx_10_12_x86_64.whl", hash = "sha256:65dcf105c1943cba45d19207ef51b8bc46d232a381e94dd38719d52d3980015b"}, - {file = "rpds_py-0.17.1-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:01f58a7306b64e0a4fe042047dd2b7d411ee82e54240284bab63e325762c1147"}, - {file = "rpds_py-0.17.1-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:071bc28c589b86bc6351a339114fb7a029f5cddbaca34103aa573eba7b482382"}, - {file = "rpds_py-0.17.1-cp311-cp311-manylinux_2_17_armv7l.manylinux2014_armv7l.whl", hash = "sha256:ae35e8e6801c5ab071b992cb2da958eee76340e6926ec693b5ff7d6381441745"}, - {file = "rpds_py-0.17.1-cp311-cp311-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:149c5cd24f729e3567b56e1795f74577aa3126c14c11e457bec1b1c90d212e38"}, - {file = "rpds_py-0.17.1-cp311-cp311-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:e796051f2070f47230c745d0a77a91088fbee2cc0502e9b796b9c6471983718c"}, - {file = "rpds_py-0.17.1-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:60e820ee1004327609b28db8307acc27f5f2e9a0b185b2064c5f23e815f248f8"}, - {file = "rpds_py-0.17.1-cp311-cp311-manylinux_2_5_i686.manylinux1_i686.whl", hash = "sha256:1957a2ab607f9added64478a6982742eb29f109d89d065fa44e01691a20fc20a"}, - {file = "rpds_py-0.17.1-cp311-cp311-musllinux_1_2_aarch64.whl", hash = "sha256:8587fd64c2a91c33cdc39d0cebdaf30e79491cc029a37fcd458ba863f8815383"}, - {file = "rpds_py-0.17.1-cp311-cp311-musllinux_1_2_i686.whl", hash = "sha256:4dc889a9d8a34758d0fcc9ac86adb97bab3fb7f0c4d29794357eb147536483fd"}, - {file = "rpds_py-0.17.1-cp311-cp311-musllinux_1_2_x86_64.whl", hash = "sha256:2953937f83820376b5979318840f3ee47477d94c17b940fe31d9458d79ae7eea"}, - {file = "rpds_py-0.17.1-cp311-none-win32.whl", hash = "sha256:1bfcad3109c1e5ba3cbe2f421614e70439f72897515a96c462ea657261b96518"}, - {file = "rpds_py-0.17.1-cp311-none-win_amd64.whl", hash = "sha256:99da0a4686ada4ed0f778120a0ea8d066de1a0a92ab0d13ae68492a437db78bf"}, - {file = "rpds_py-0.17.1-cp312-cp312-macosx_10_12_x86_64.whl", hash = "sha256:1dc29db3900cb1bb40353772417800f29c3d078dbc8024fd64655a04ee3c4bdf"}, - {file = "rpds_py-0.17.1-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:82ada4a8ed9e82e443fcef87e22a3eed3654dd3adf6e3b3a0deb70f03e86142a"}, - {file = "rpds_py-0.17.1-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:1d36b2b59e8cc6e576f8f7b671e32f2ff43153f0ad6d0201250a7c07f25d570e"}, - {file = "rpds_py-0.17.1-cp312-cp312-manylinux_2_17_armv7l.manylinux2014_armv7l.whl", hash = "sha256:3677fcca7fb728c86a78660c7fb1b07b69b281964673f486ae72860e13f512ad"}, - {file = "rpds_py-0.17.1-cp312-cp312-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:516fb8c77805159e97a689e2f1c80655c7658f5af601c34ffdb916605598cda2"}, - {file = "rpds_py-0.17.1-cp312-cp312-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:df3b6f45ba4515632c5064e35ca7f31d51d13d1479673185ba8f9fefbbed58b9"}, - {file = "rpds_py-0.17.1-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:a967dd6afda7715d911c25a6ba1517975acd8d1092b2f326718725461a3d33f9"}, - {file = "rpds_py-0.17.1-cp312-cp312-manylinux_2_5_i686.manylinux1_i686.whl", hash = "sha256:dbbb95e6fc91ea3102505d111b327004d1c4ce98d56a4a02e82cd451f9f57140"}, - {file = "rpds_py-0.17.1-cp312-cp312-musllinux_1_2_aarch64.whl", hash = "sha256:02866e060219514940342a1f84303a1ef7a1dad0ac311792fbbe19b521b489d2"}, - {file = "rpds_py-0.17.1-cp312-cp312-musllinux_1_2_i686.whl", hash = "sha256:2528ff96d09f12e638695f3a2e0c609c7b84c6df7c5ae9bfeb9252b6fa686253"}, - {file = "rpds_py-0.17.1-cp312-cp312-musllinux_1_2_x86_64.whl", hash = "sha256:bd345a13ce06e94c753dab52f8e71e5252aec1e4f8022d24d56decd31e1b9b23"}, - {file = "rpds_py-0.17.1-cp312-none-win32.whl", hash = "sha256:2a792b2e1d3038daa83fa474d559acfd6dc1e3650ee93b2662ddc17dbff20ad1"}, - {file = "rpds_py-0.17.1-cp312-none-win_amd64.whl", hash = "sha256:292f7344a3301802e7c25c53792fae7d1593cb0e50964e7bcdcc5cf533d634e3"}, - {file = "rpds_py-0.17.1-cp38-cp38-macosx_10_12_x86_64.whl", hash = "sha256:8ffe53e1d8ef2520ebcf0c9fec15bb721da59e8ef283b6ff3079613b1e30513d"}, - {file = "rpds_py-0.17.1-cp38-cp38-macosx_11_0_arm64.whl", hash = "sha256:4341bd7579611cf50e7b20bb8c2e23512a3dc79de987a1f411cb458ab670eb90"}, - {file = "rpds_py-0.17.1-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:2f4eb548daf4836e3b2c662033bfbfc551db58d30fd8fe660314f86bf8510b93"}, - {file = "rpds_py-0.17.1-cp38-cp38-manylinux_2_17_armv7l.manylinux2014_armv7l.whl", hash = "sha256:b686f25377f9c006acbac63f61614416a6317133ab7fafe5de5f7dc8a06d42eb"}, - {file = "rpds_py-0.17.1-cp38-cp38-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:4e21b76075c01d65d0f0f34302b5a7457d95721d5e0667aea65e5bb3ab415c25"}, - {file = "rpds_py-0.17.1-cp38-cp38-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:b86b21b348f7e5485fae740d845c65a880f5d1eda1e063bc59bef92d1f7d0c55"}, - {file = "rpds_py-0.17.1-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:f175e95a197f6a4059b50757a3dca33b32b61691bdbd22c29e8a8d21d3914cae"}, - {file = "rpds_py-0.17.1-cp38-cp38-manylinux_2_5_i686.manylinux1_i686.whl", hash = "sha256:1701fc54460ae2e5efc1dd6350eafd7a760f516df8dbe51d4a1c79d69472fbd4"}, - {file = "rpds_py-0.17.1-cp38-cp38-musllinux_1_2_aarch64.whl", hash = "sha256:9051e3d2af8f55b42061603e29e744724cb5f65b128a491446cc029b3e2ea896"}, - {file = "rpds_py-0.17.1-cp38-cp38-musllinux_1_2_i686.whl", hash = "sha256:7450dbd659fed6dd41d1a7d47ed767e893ba402af8ae664c157c255ec6067fde"}, - {file = "rpds_py-0.17.1-cp38-cp38-musllinux_1_2_x86_64.whl", hash = "sha256:5a024fa96d541fd7edaa0e9d904601c6445e95a729a2900c5aec6555fe921ed6"}, - {file = "rpds_py-0.17.1-cp38-none-win32.whl", hash = "sha256:da1ead63368c04a9bded7904757dfcae01eba0e0f9bc41d3d7f57ebf1c04015a"}, - {file = "rpds_py-0.17.1-cp38-none-win_amd64.whl", hash = "sha256:841320e1841bb53fada91c9725e766bb25009cfd4144e92298db296fb6c894fb"}, - {file = "rpds_py-0.17.1-cp39-cp39-macosx_10_12_x86_64.whl", hash = "sha256:f6c43b6f97209e370124baf2bf40bb1e8edc25311a158867eb1c3a5d449ebc7a"}, - {file = "rpds_py-0.17.1-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:5e7d63ec01fe7c76c2dbb7e972fece45acbb8836e72682bde138e7e039906e2c"}, - {file = "rpds_py-0.17.1-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:81038ff87a4e04c22e1d81f947c6ac46f122e0c80460b9006e6517c4d842a6ec"}, - {file = "rpds_py-0.17.1-cp39-cp39-manylinux_2_17_armv7l.manylinux2014_armv7l.whl", hash = "sha256:810685321f4a304b2b55577c915bece4c4a06dfe38f6e62d9cc1d6ca8ee86b99"}, - {file = "rpds_py-0.17.1-cp39-cp39-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:25f071737dae674ca8937a73d0f43f5a52e92c2d178330b4c0bb6ab05586ffa6"}, - {file = "rpds_py-0.17.1-cp39-cp39-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:aa5bfb13f1e89151ade0eb812f7b0d7a4d643406caaad65ce1cbabe0a66d695f"}, - {file = "rpds_py-0.17.1-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:dfe07308b311a8293a0d5ef4e61411c5c20f682db6b5e73de6c7c8824272c256"}, - {file = "rpds_py-0.17.1-cp39-cp39-manylinux_2_5_i686.manylinux1_i686.whl", hash = "sha256:a000133a90eea274a6f28adc3084643263b1e7c1a5a66eb0a0a7a36aa757ed74"}, - {file = "rpds_py-0.17.1-cp39-cp39-musllinux_1_2_aarch64.whl", hash = "sha256:5d0e8a6434a3fbf77d11448c9c25b2f25244226cfbec1a5159947cac5b8c5fa4"}, - {file = "rpds_py-0.17.1-cp39-cp39-musllinux_1_2_i686.whl", hash = "sha256:efa767c220d94aa4ac3a6dd3aeb986e9f229eaf5bce92d8b1b3018d06bed3772"}, - {file = "rpds_py-0.17.1-cp39-cp39-musllinux_1_2_x86_64.whl", hash = "sha256:dbc56680ecf585a384fbd93cd42bc82668b77cb525343170a2d86dafaed2a84b"}, - {file = "rpds_py-0.17.1-cp39-none-win32.whl", hash = "sha256:270987bc22e7e5a962b1094953ae901395e8c1e1e83ad016c5cfcfff75a15a3f"}, - {file = "rpds_py-0.17.1-cp39-none-win_amd64.whl", hash = "sha256:2a7b2f2f56a16a6d62e55354dd329d929560442bd92e87397b7a9586a32e3e76"}, - {file = "rpds_py-0.17.1-pp310-pypy310_pp73-macosx_10_12_x86_64.whl", hash = "sha256:a3264e3e858de4fc601741498215835ff324ff2482fd4e4af61b46512dd7fc83"}, - {file = "rpds_py-0.17.1-pp310-pypy310_pp73-macosx_11_0_arm64.whl", hash = "sha256:f2f3b28b40fddcb6c1f1f6c88c6f3769cd933fa493ceb79da45968a21dccc920"}, - {file = "rpds_py-0.17.1-pp310-pypy310_pp73-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:9584f8f52010295a4a417221861df9bea4c72d9632562b6e59b3c7b87a1522b7"}, - {file = "rpds_py-0.17.1-pp310-pypy310_pp73-manylinux_2_17_armv7l.manylinux2014_armv7l.whl", hash = "sha256:c64602e8be701c6cfe42064b71c84ce62ce66ddc6422c15463fd8127db3d8066"}, - {file = "rpds_py-0.17.1-pp310-pypy310_pp73-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:060f412230d5f19fc8c8b75f315931b408d8ebf56aec33ef4168d1b9e54200b1"}, - {file = "rpds_py-0.17.1-pp310-pypy310_pp73-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:b9412abdf0ba70faa6e2ee6c0cc62a8defb772e78860cef419865917d86c7342"}, - {file = "rpds_py-0.17.1-pp310-pypy310_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:9737bdaa0ad33d34c0efc718741abaafce62fadae72c8b251df9b0c823c63b22"}, - {file = "rpds_py-0.17.1-pp310-pypy310_pp73-manylinux_2_5_i686.manylinux1_i686.whl", hash = "sha256:9f0e4dc0f17dcea4ab9d13ac5c666b6b5337042b4d8f27e01b70fae41dd65c57"}, - {file = "rpds_py-0.17.1-pp310-pypy310_pp73-musllinux_1_2_aarch64.whl", hash = "sha256:1db228102ab9d1ff4c64148c96320d0be7044fa28bd865a9ce628ce98da5973d"}, - {file = "rpds_py-0.17.1-pp310-pypy310_pp73-musllinux_1_2_i686.whl", hash = "sha256:d8bbd8e56f3ba25a7d0cf980fc42b34028848a53a0e36c9918550e0280b9d0b6"}, - {file = "rpds_py-0.17.1-pp310-pypy310_pp73-musllinux_1_2_x86_64.whl", hash = "sha256:be22ae34d68544df293152b7e50895ba70d2a833ad9566932d750d3625918b82"}, - {file = "rpds_py-0.17.1-pp38-pypy38_pp73-macosx_10_12_x86_64.whl", hash = "sha256:bf046179d011e6114daf12a534d874958b039342b347348a78b7cdf0dd9d6041"}, - {file = "rpds_py-0.17.1-pp38-pypy38_pp73-macosx_11_0_arm64.whl", hash = "sha256:1a746a6d49665058a5896000e8d9d2f1a6acba8a03b389c1e4c06e11e0b7f40d"}, - {file = "rpds_py-0.17.1-pp38-pypy38_pp73-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:f0b8bf5b8db49d8fd40f54772a1dcf262e8be0ad2ab0206b5a2ec109c176c0a4"}, - {file = "rpds_py-0.17.1-pp38-pypy38_pp73-manylinux_2_17_armv7l.manylinux2014_armv7l.whl", hash = "sha256:f7f4cb1f173385e8a39c29510dd11a78bf44e360fb75610594973f5ea141028b"}, - {file = "rpds_py-0.17.1-pp38-pypy38_pp73-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:7fbd70cb8b54fe745301921b0816c08b6d917593429dfc437fd024b5ba713c58"}, - {file = "rpds_py-0.17.1-pp38-pypy38_pp73-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:9bdf1303df671179eaf2cb41e8515a07fc78d9d00f111eadbe3e14262f59c3d0"}, - {file = "rpds_py-0.17.1-pp38-pypy38_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:fad059a4bd14c45776600d223ec194e77db6c20255578bb5bcdd7c18fd169361"}, - {file = "rpds_py-0.17.1-pp38-pypy38_pp73-manylinux_2_5_i686.manylinux1_i686.whl", hash = "sha256:3664d126d3388a887db44c2e293f87d500c4184ec43d5d14d2d2babdb4c64cad"}, - {file = "rpds_py-0.17.1-pp38-pypy38_pp73-musllinux_1_2_aarch64.whl", hash = "sha256:698ea95a60c8b16b58be9d854c9f993c639f5c214cf9ba782eca53a8789d6b19"}, - {file = "rpds_py-0.17.1-pp38-pypy38_pp73-musllinux_1_2_i686.whl", hash = "sha256:c3d2010656999b63e628a3c694f23020322b4178c450dc478558a2b6ef3cb9bb"}, - {file = "rpds_py-0.17.1-pp38-pypy38_pp73-musllinux_1_2_x86_64.whl", hash = "sha256:938eab7323a736533f015e6069a7d53ef2dcc841e4e533b782c2bfb9fb12d84b"}, - {file = "rpds_py-0.17.1-pp39-pypy39_pp73-macosx_10_12_x86_64.whl", hash = "sha256:1e626b365293a2142a62b9a614e1f8e331b28f3ca57b9f05ebbf4cf2a0f0bdc5"}, - {file = "rpds_py-0.17.1-pp39-pypy39_pp73-macosx_11_0_arm64.whl", hash = "sha256:380e0df2e9d5d5d339803cfc6d183a5442ad7ab3c63c2a0982e8c824566c5ccc"}, - {file = "rpds_py-0.17.1-pp39-pypy39_pp73-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:b760a56e080a826c2e5af09002c1a037382ed21d03134eb6294812dda268c811"}, - {file = "rpds_py-0.17.1-pp39-pypy39_pp73-manylinux_2_17_armv7l.manylinux2014_armv7l.whl", hash = "sha256:5576ee2f3a309d2bb403ec292d5958ce03953b0e57a11d224c1f134feaf8c40f"}, - {file = "rpds_py-0.17.1-pp39-pypy39_pp73-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:1f3c3461ebb4c4f1bbc70b15d20b565759f97a5aaf13af811fcefc892e9197ba"}, - {file = "rpds_py-0.17.1-pp39-pypy39_pp73-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:637b802f3f069a64436d432117a7e58fab414b4e27a7e81049817ae94de45d8d"}, - {file = "rpds_py-0.17.1-pp39-pypy39_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:ffee088ea9b593cc6160518ba9bd319b5475e5f3e578e4552d63818773c6f56a"}, - {file = "rpds_py-0.17.1-pp39-pypy39_pp73-manylinux_2_5_i686.manylinux1_i686.whl", hash = "sha256:3ac732390d529d8469b831949c78085b034bff67f584559340008d0f6041a049"}, - {file = "rpds_py-0.17.1-pp39-pypy39_pp73-musllinux_1_2_aarch64.whl", hash = "sha256:93432e747fb07fa567ad9cc7aaadd6e29710e515aabf939dfbed8046041346c6"}, - {file = "rpds_py-0.17.1-pp39-pypy39_pp73-musllinux_1_2_i686.whl", hash = "sha256:7b7d9ca34542099b4e185b3c2a2b2eda2e318a7dbde0b0d83357a6d4421b5296"}, - {file = "rpds_py-0.17.1-pp39-pypy39_pp73-musllinux_1_2_x86_64.whl", hash = "sha256:0387ce69ba06e43df54e43968090f3626e231e4bc9150e4c3246947567695f68"}, - {file = "rpds_py-0.17.1.tar.gz", hash = "sha256:0210b2668f24c078307260bf88bdac9d6f1093635df5123789bfee4d8d7fc8e7"}, + {file = "rpds_py-0.18.1-cp310-cp310-macosx_10_12_x86_64.whl", hash = "sha256:d31dea506d718693b6b2cffc0648a8929bdc51c70a311b2770f09611caa10d53"}, + {file = "rpds_py-0.18.1-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:732672fbc449bab754e0b15356c077cc31566df874964d4801ab14f71951ea80"}, + {file = "rpds_py-0.18.1-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:4a98a1f0552b5f227a3d6422dbd61bc6f30db170939bd87ed14f3c339aa6c7c9"}, + {file = "rpds_py-0.18.1-cp310-cp310-manylinux_2_17_armv7l.manylinux2014_armv7l.whl", hash = "sha256:7f1944ce16401aad1e3f7d312247b3d5de7981f634dc9dfe90da72b87d37887d"}, + {file = "rpds_py-0.18.1-cp310-cp310-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:38e14fb4e370885c4ecd734f093a2225ee52dc384b86fa55fe3f74638b2cfb09"}, + {file = "rpds_py-0.18.1-cp310-cp310-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:08d74b184f9ab6289b87b19fe6a6d1a97fbfea84b8a3e745e87a5de3029bf944"}, + {file = "rpds_py-0.18.1-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:d70129cef4a8d979caa37e7fe957202e7eee8ea02c5e16455bc9808a59c6b2f0"}, + {file = "rpds_py-0.18.1-cp310-cp310-manylinux_2_5_i686.manylinux1_i686.whl", hash = "sha256:ce0bb20e3a11bd04461324a6a798af34d503f8d6f1aa3d2aa8901ceaf039176d"}, + {file = "rpds_py-0.18.1-cp310-cp310-musllinux_1_2_aarch64.whl", hash = "sha256:81c5196a790032e0fc2464c0b4ab95f8610f96f1f2fa3d4deacce6a79852da60"}, + {file = "rpds_py-0.18.1-cp310-cp310-musllinux_1_2_i686.whl", hash = "sha256:f3027be483868c99b4985fda802a57a67fdf30c5d9a50338d9db646d590198da"}, + {file = "rpds_py-0.18.1-cp310-cp310-musllinux_1_2_x86_64.whl", hash = "sha256:d44607f98caa2961bab4fa3c4309724b185b464cdc3ba6f3d7340bac3ec97cc1"}, + {file = "rpds_py-0.18.1-cp310-none-win32.whl", hash = "sha256:c273e795e7a0f1fddd46e1e3cb8be15634c29ae8ff31c196debb620e1edb9333"}, + {file = "rpds_py-0.18.1-cp310-none-win_amd64.whl", hash = "sha256:8352f48d511de5f973e4f2f9412736d7dea76c69faa6d36bcf885b50c758ab9a"}, + {file = "rpds_py-0.18.1-cp311-cp311-macosx_10_12_x86_64.whl", hash = "sha256:6b5ff7e1d63a8281654b5e2896d7f08799378e594f09cf3674e832ecaf396ce8"}, + {file = "rpds_py-0.18.1-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:8927638a4d4137a289e41d0fd631551e89fa346d6dbcfc31ad627557d03ceb6d"}, + {file = "rpds_py-0.18.1-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:154bf5c93d79558b44e5b50cc354aa0459e518e83677791e6adb0b039b7aa6a7"}, + {file = "rpds_py-0.18.1-cp311-cp311-manylinux_2_17_armv7l.manylinux2014_armv7l.whl", hash = "sha256:07f2139741e5deb2c5154a7b9629bc5aa48c766b643c1a6750d16f865a82c5fc"}, + {file = "rpds_py-0.18.1-cp311-cp311-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:8c7672e9fba7425f79019db9945b16e308ed8bc89348c23d955c8c0540da0a07"}, + {file = "rpds_py-0.18.1-cp311-cp311-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:489bdfe1abd0406eba6b3bb4fdc87c7fa40f1031de073d0cfb744634cc8fa261"}, + {file = "rpds_py-0.18.1-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:3c20f05e8e3d4fc76875fc9cb8cf24b90a63f5a1b4c5b9273f0e8225e169b100"}, + {file = "rpds_py-0.18.1-cp311-cp311-manylinux_2_5_i686.manylinux1_i686.whl", hash = "sha256:967342e045564cef76dfcf1edb700b1e20838d83b1aa02ab313e6a497cf923b8"}, + {file = "rpds_py-0.18.1-cp311-cp311-musllinux_1_2_aarch64.whl", hash = "sha256:2cc7c1a47f3a63282ab0f422d90ddac4aa3034e39fc66a559ab93041e6505da7"}, + {file = "rpds_py-0.18.1-cp311-cp311-musllinux_1_2_i686.whl", hash = "sha256:f7afbfee1157e0f9376c00bb232e80a60e59ed716e3211a80cb8506550671e6e"}, + {file = "rpds_py-0.18.1-cp311-cp311-musllinux_1_2_x86_64.whl", hash = "sha256:9e6934d70dc50f9f8ea47081ceafdec09245fd9f6032669c3b45705dea096b88"}, + {file = "rpds_py-0.18.1-cp311-none-win32.whl", hash = "sha256:c69882964516dc143083d3795cb508e806b09fc3800fd0d4cddc1df6c36e76bb"}, + {file = "rpds_py-0.18.1-cp311-none-win_amd64.whl", hash = "sha256:70a838f7754483bcdc830444952fd89645569e7452e3226de4a613a4c1793fb2"}, + {file = "rpds_py-0.18.1-cp312-cp312-macosx_10_12_x86_64.whl", hash = "sha256:3dd3cd86e1db5aadd334e011eba4e29d37a104b403e8ca24dcd6703c68ca55b3"}, + {file = "rpds_py-0.18.1-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:05f3d615099bd9b13ecf2fc9cf2d839ad3f20239c678f461c753e93755d629ee"}, + {file = "rpds_py-0.18.1-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:35b2b771b13eee8729a5049c976197ff58a27a3829c018a04341bcf1ae409b2b"}, + {file = "rpds_py-0.18.1-cp312-cp312-manylinux_2_17_armv7l.manylinux2014_armv7l.whl", hash = "sha256:ee17cd26b97d537af8f33635ef38be873073d516fd425e80559f4585a7b90c43"}, + {file = "rpds_py-0.18.1-cp312-cp312-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:b646bf655b135ccf4522ed43d6902af37d3f5dbcf0da66c769a2b3938b9d8184"}, + {file = "rpds_py-0.18.1-cp312-cp312-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:19ba472b9606c36716062c023afa2484d1e4220548751bda14f725a7de17b4f6"}, + {file = "rpds_py-0.18.1-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:6e30ac5e329098903262dc5bdd7e2086e0256aa762cc8b744f9e7bf2a427d3f8"}, + {file = "rpds_py-0.18.1-cp312-cp312-manylinux_2_5_i686.manylinux1_i686.whl", hash = "sha256:d58ad6317d188c43750cb76e9deacf6051d0f884d87dc6518e0280438648a9ac"}, + {file = "rpds_py-0.18.1-cp312-cp312-musllinux_1_2_aarch64.whl", hash = "sha256:e1735502458621921cee039c47318cb90b51d532c2766593be6207eec53e5c4c"}, + {file = "rpds_py-0.18.1-cp312-cp312-musllinux_1_2_i686.whl", hash = "sha256:f5bab211605d91db0e2995a17b5c6ee5edec1270e46223e513eaa20da20076ac"}, + {file = "rpds_py-0.18.1-cp312-cp312-musllinux_1_2_x86_64.whl", hash = "sha256:2fc24a329a717f9e2448f8cd1f960f9dac4e45b6224d60734edeb67499bab03a"}, + {file = "rpds_py-0.18.1-cp312-none-win32.whl", hash = "sha256:1805d5901779662d599d0e2e4159d8a82c0b05faa86ef9222bf974572286b2b6"}, + {file = "rpds_py-0.18.1-cp312-none-win_amd64.whl", hash = "sha256:720edcb916df872d80f80a1cc5ea9058300b97721efda8651efcd938a9c70a72"}, + {file = "rpds_py-0.18.1-cp38-cp38-macosx_10_12_x86_64.whl", hash = "sha256:c827576e2fa017a081346dce87d532a5310241648eb3700af9a571a6e9fc7e74"}, + {file = "rpds_py-0.18.1-cp38-cp38-macosx_11_0_arm64.whl", hash = "sha256:aa3679e751408d75a0b4d8d26d6647b6d9326f5e35c00a7ccd82b78ef64f65f8"}, + {file = "rpds_py-0.18.1-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:0abeee75434e2ee2d142d650d1e54ac1f8b01e6e6abdde8ffd6eeac6e9c38e20"}, + {file = "rpds_py-0.18.1-cp38-cp38-manylinux_2_17_armv7l.manylinux2014_armv7l.whl", hash = "sha256:ed402d6153c5d519a0faf1bb69898e97fb31613b49da27a84a13935ea9164dfc"}, + {file = "rpds_py-0.18.1-cp38-cp38-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:338dee44b0cef8b70fd2ef54b4e09bb1b97fc6c3a58fea5db6cc083fd9fc2724"}, + {file = "rpds_py-0.18.1-cp38-cp38-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:7750569d9526199c5b97e5a9f8d96a13300950d910cf04a861d96f4273d5b104"}, + {file = "rpds_py-0.18.1-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:607345bd5912aacc0c5a63d45a1f73fef29e697884f7e861094e443187c02be5"}, + {file = "rpds_py-0.18.1-cp38-cp38-manylinux_2_5_i686.manylinux1_i686.whl", hash = "sha256:207c82978115baa1fd8d706d720b4a4d2b0913df1c78c85ba73fe6c5804505f0"}, + {file = "rpds_py-0.18.1-cp38-cp38-musllinux_1_2_aarch64.whl", hash = "sha256:6d1e42d2735d437e7e80bab4d78eb2e459af48c0a46e686ea35f690b93db792d"}, + {file = "rpds_py-0.18.1-cp38-cp38-musllinux_1_2_i686.whl", hash = "sha256:5463c47c08630007dc0fe99fb480ea4f34a89712410592380425a9b4e1611d8e"}, + {file = "rpds_py-0.18.1-cp38-cp38-musllinux_1_2_x86_64.whl", hash = "sha256:06d218939e1bf2ca50e6b0ec700ffe755e5216a8230ab3e87c059ebb4ea06afc"}, + {file = "rpds_py-0.18.1-cp38-none-win32.whl", hash = "sha256:312fe69b4fe1ffbe76520a7676b1e5ac06ddf7826d764cc10265c3b53f96dbe9"}, + {file = "rpds_py-0.18.1-cp38-none-win_amd64.whl", hash = "sha256:9437ca26784120a279f3137ee080b0e717012c42921eb07861b412340f85bae2"}, + {file = "rpds_py-0.18.1-cp39-cp39-macosx_10_12_x86_64.whl", hash = "sha256:19e515b78c3fc1039dd7da0a33c28c3154458f947f4dc198d3c72db2b6b5dc93"}, + {file = "rpds_py-0.18.1-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:a7b28c5b066bca9a4eb4e2f2663012debe680f097979d880657f00e1c30875a0"}, + {file = "rpds_py-0.18.1-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:673fdbbf668dd958eff750e500495ef3f611e2ecc209464f661bc82e9838991e"}, + {file = "rpds_py-0.18.1-cp39-cp39-manylinux_2_17_armv7l.manylinux2014_armv7l.whl", hash = "sha256:d960de62227635d2e61068f42a6cb6aae91a7fe00fca0e3aeed17667c8a34611"}, + {file = "rpds_py-0.18.1-cp39-cp39-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:352a88dc7892f1da66b6027af06a2e7e5d53fe05924cc2cfc56495b586a10b72"}, + {file = "rpds_py-0.18.1-cp39-cp39-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:4e0ee01ad8260184db21468a6e1c37afa0529acc12c3a697ee498d3c2c4dcaf3"}, + {file = "rpds_py-0.18.1-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:e4c39ad2f512b4041343ea3c7894339e4ca7839ac38ca83d68a832fc8b3748ab"}, + {file = "rpds_py-0.18.1-cp39-cp39-manylinux_2_5_i686.manylinux1_i686.whl", hash = "sha256:aaa71ee43a703c321906813bb252f69524f02aa05bf4eec85f0c41d5d62d0f4c"}, + {file = "rpds_py-0.18.1-cp39-cp39-musllinux_1_2_aarch64.whl", hash = "sha256:6cd8098517c64a85e790657e7b1e509b9fe07487fd358e19431cb120f7d96338"}, + {file = "rpds_py-0.18.1-cp39-cp39-musllinux_1_2_i686.whl", hash = "sha256:4adec039b8e2928983f885c53b7cc4cda8965b62b6596501a0308d2703f8af1b"}, + {file = "rpds_py-0.18.1-cp39-cp39-musllinux_1_2_x86_64.whl", hash = "sha256:32b7daaa3e9389db3695964ce8e566e3413b0c43e3394c05e4b243a4cd7bef26"}, + {file = "rpds_py-0.18.1-cp39-none-win32.whl", hash = "sha256:2625f03b105328729f9450c8badda34d5243231eef6535f80064d57035738360"}, + {file = "rpds_py-0.18.1-cp39-none-win_amd64.whl", hash = "sha256:bf18932d0003c8c4d51a39f244231986ab23ee057d235a12b2684ea26a353590"}, + {file = "rpds_py-0.18.1-pp310-pypy310_pp73-macosx_10_12_x86_64.whl", hash = "sha256:cbfbea39ba64f5e53ae2915de36f130588bba71245b418060ec3330ebf85678e"}, + {file = "rpds_py-0.18.1-pp310-pypy310_pp73-macosx_11_0_arm64.whl", hash = "sha256:a3d456ff2a6a4d2adcdf3c1c960a36f4fd2fec6e3b4902a42a384d17cf4e7a65"}, + {file = "rpds_py-0.18.1-pp310-pypy310_pp73-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:7700936ef9d006b7ef605dc53aa364da2de5a3aa65516a1f3ce73bf82ecfc7ae"}, + {file = "rpds_py-0.18.1-pp310-pypy310_pp73-manylinux_2_17_armv7l.manylinux2014_armv7l.whl", hash = "sha256:51584acc5916212e1bf45edd17f3a6b05fe0cbb40482d25e619f824dccb679de"}, + {file = "rpds_py-0.18.1-pp310-pypy310_pp73-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:942695a206a58d2575033ff1e42b12b2aece98d6003c6bc739fbf33d1773b12f"}, + {file = "rpds_py-0.18.1-pp310-pypy310_pp73-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:b906b5f58892813e5ba5c6056d6a5ad08f358ba49f046d910ad992196ea61397"}, + {file = "rpds_py-0.18.1-pp310-pypy310_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:f6f8e3fecca256fefc91bb6765a693d96692459d7d4c644660a9fff32e517843"}, + {file = "rpds_py-0.18.1-pp310-pypy310_pp73-manylinux_2_5_i686.manylinux1_i686.whl", hash = "sha256:7732770412bab81c5a9f6d20aeb60ae943a9b36dcd990d876a773526468e7163"}, + {file = "rpds_py-0.18.1-pp310-pypy310_pp73-musllinux_1_2_aarch64.whl", hash = "sha256:bd1105b50ede37461c1d51b9698c4f4be6e13e69a908ab7751e3807985fc0346"}, + {file = "rpds_py-0.18.1-pp310-pypy310_pp73-musllinux_1_2_i686.whl", hash = "sha256:618916f5535784960f3ecf8111581f4ad31d347c3de66d02e728de460a46303c"}, + {file = "rpds_py-0.18.1-pp310-pypy310_pp73-musllinux_1_2_x86_64.whl", hash = "sha256:17c6d2155e2423f7e79e3bb18151c686d40db42d8645e7977442170c360194d4"}, + {file = "rpds_py-0.18.1-pp38-pypy38_pp73-macosx_10_12_x86_64.whl", hash = "sha256:6c4c4c3f878df21faf5fac86eda32671c27889e13570645a9eea0a1abdd50922"}, + {file = "rpds_py-0.18.1-pp38-pypy38_pp73-macosx_11_0_arm64.whl", hash = "sha256:fab6ce90574645a0d6c58890e9bcaac8d94dff54fb51c69e5522a7358b80ab64"}, + {file = "rpds_py-0.18.1-pp38-pypy38_pp73-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:531796fb842b53f2695e94dc338929e9f9dbf473b64710c28af5a160b2a8927d"}, + {file = "rpds_py-0.18.1-pp38-pypy38_pp73-manylinux_2_17_armv7l.manylinux2014_armv7l.whl", hash = "sha256:740884bc62a5e2bbb31e584f5d23b32320fd75d79f916f15a788d527a5e83644"}, + {file = "rpds_py-0.18.1-pp38-pypy38_pp73-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:998125738de0158f088aef3cb264a34251908dd2e5d9966774fdab7402edfab7"}, + {file = "rpds_py-0.18.1-pp38-pypy38_pp73-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:e2be6e9dd4111d5b31ba3b74d17da54a8319d8168890fbaea4b9e5c3de630ae5"}, + {file = "rpds_py-0.18.1-pp38-pypy38_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:d0cee71bc618cd93716f3c1bf56653740d2d13ddbd47673efa8bf41435a60daa"}, + {file = "rpds_py-0.18.1-pp38-pypy38_pp73-manylinux_2_5_i686.manylinux1_i686.whl", hash = "sha256:2c3caec4ec5cd1d18e5dd6ae5194d24ed12785212a90b37f5f7f06b8bedd7139"}, + {file = "rpds_py-0.18.1-pp38-pypy38_pp73-musllinux_1_2_aarch64.whl", hash = "sha256:27bba383e8c5231cd559affe169ca0b96ec78d39909ffd817f28b166d7ddd4d8"}, + {file = "rpds_py-0.18.1-pp38-pypy38_pp73-musllinux_1_2_i686.whl", hash = "sha256:a888e8bdb45916234b99da2d859566f1e8a1d2275a801bb8e4a9644e3c7e7909"}, + {file = "rpds_py-0.18.1-pp38-pypy38_pp73-musllinux_1_2_x86_64.whl", hash = "sha256:6031b25fb1b06327b43d841f33842b383beba399884f8228a6bb3df3088485ff"}, + {file = "rpds_py-0.18.1-pp39-pypy39_pp73-macosx_10_12_x86_64.whl", hash = "sha256:48c2faaa8adfacefcbfdb5f2e2e7bdad081e5ace8d182e5f4ade971f128e6bb3"}, + {file = "rpds_py-0.18.1-pp39-pypy39_pp73-macosx_11_0_arm64.whl", hash = "sha256:d85164315bd68c0806768dc6bb0429c6f95c354f87485ee3593c4f6b14def2bd"}, + {file = "rpds_py-0.18.1-pp39-pypy39_pp73-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:6afd80f6c79893cfc0574956f78a0add8c76e3696f2d6a15bca2c66c415cf2d4"}, + {file = "rpds_py-0.18.1-pp39-pypy39_pp73-manylinux_2_17_armv7l.manylinux2014_armv7l.whl", hash = "sha256:fa242ac1ff583e4ec7771141606aafc92b361cd90a05c30d93e343a0c2d82a89"}, + {file = "rpds_py-0.18.1-pp39-pypy39_pp73-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:d21be4770ff4e08698e1e8e0bce06edb6ea0626e7c8f560bc08222880aca6a6f"}, + {file = "rpds_py-0.18.1-pp39-pypy39_pp73-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:5c45a639e93a0c5d4b788b2613bd637468edd62f8f95ebc6fcc303d58ab3f0a8"}, + {file = "rpds_py-0.18.1-pp39-pypy39_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:910e71711d1055b2768181efa0a17537b2622afeb0424116619817007f8a2b10"}, + {file = "rpds_py-0.18.1-pp39-pypy39_pp73-manylinux_2_5_i686.manylinux1_i686.whl", hash = "sha256:b9bb1f182a97880f6078283b3505a707057c42bf55d8fca604f70dedfdc0772a"}, + {file = "rpds_py-0.18.1-pp39-pypy39_pp73-musllinux_1_2_aarch64.whl", hash = "sha256:1d54f74f40b1f7aaa595a02ff42ef38ca654b1469bef7d52867da474243cc633"}, + {file = "rpds_py-0.18.1-pp39-pypy39_pp73-musllinux_1_2_i686.whl", hash = "sha256:8d2e182c9ee01135e11e9676e9a62dfad791a7a467738f06726872374a83db49"}, + {file = "rpds_py-0.18.1-pp39-pypy39_pp73-musllinux_1_2_x86_64.whl", hash = "sha256:636a15acc588f70fda1661234761f9ed9ad79ebed3f2125d44be0862708b666e"}, + {file = "rpds_py-0.18.1.tar.gz", hash = "sha256:dc48b479d540770c811fbd1eb9ba2bb66951863e448efec2e2c102625328e92f"}, ] [[package]] name = "scipy" -version = "1.12.0" +version = "1.13.1" description = "Fundamental algorithms for scientific computing in Python" optional = false python-versions = ">=3.9" files = [ - {file = "scipy-1.12.0-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:78e4402e140879387187f7f25d91cc592b3501a2e51dfb320f48dfb73565f10b"}, - {file = "scipy-1.12.0-cp310-cp310-macosx_12_0_arm64.whl", hash = "sha256:f5f00ebaf8de24d14b8449981a2842d404152774c1a1d880c901bf454cb8e2a1"}, - {file = "scipy-1.12.0-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:e53958531a7c695ff66c2e7bb7b79560ffdc562e2051644c5576c39ff8efb563"}, - {file = "scipy-1.12.0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:5e32847e08da8d895ce09d108a494d9eb78974cf6de23063f93306a3e419960c"}, - {file = "scipy-1.12.0-cp310-cp310-musllinux_1_1_x86_64.whl", hash = "sha256:4c1020cad92772bf44b8e4cdabc1df5d87376cb219742549ef69fc9fd86282dd"}, - {file = "scipy-1.12.0-cp310-cp310-win_amd64.whl", hash = "sha256:75ea2a144096b5e39402e2ff53a36fecfd3b960d786b7efd3c180e29c39e53f2"}, - {file = "scipy-1.12.0-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:408c68423f9de16cb9e602528be4ce0d6312b05001f3de61fe9ec8b1263cad08"}, - {file = "scipy-1.12.0-cp311-cp311-macosx_12_0_arm64.whl", hash = "sha256:5adfad5dbf0163397beb4aca679187d24aec085343755fcdbdeb32b3679f254c"}, - {file = "scipy-1.12.0-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:c3003652496f6e7c387b1cf63f4bb720951cfa18907e998ea551e6de51a04467"}, - {file = "scipy-1.12.0-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:8b8066bce124ee5531d12a74b617d9ac0ea59245246410e19bca549656d9a40a"}, - {file = "scipy-1.12.0-cp311-cp311-musllinux_1_1_x86_64.whl", hash = "sha256:8bee4993817e204d761dba10dbab0774ba5a8612e57e81319ea04d84945375ba"}, - {file = "scipy-1.12.0-cp311-cp311-win_amd64.whl", hash = "sha256:a24024d45ce9a675c1fb8494e8e5244efea1c7a09c60beb1eeb80373d0fecc70"}, - {file = "scipy-1.12.0-cp312-cp312-macosx_10_9_x86_64.whl", hash = "sha256:e7e76cc48638228212c747ada851ef355c2bb5e7f939e10952bc504c11f4e372"}, - {file = "scipy-1.12.0-cp312-cp312-macosx_12_0_arm64.whl", hash = "sha256:f7ce148dffcd64ade37b2df9315541f9adad6efcaa86866ee7dd5db0c8f041c3"}, - {file = "scipy-1.12.0-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:9c39f92041f490422924dfdb782527a4abddf4707616e07b021de33467f917bc"}, - {file = "scipy-1.12.0-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:a7ebda398f86e56178c2fa94cad15bf457a218a54a35c2a7b4490b9f9cb2676c"}, - {file = "scipy-1.12.0-cp312-cp312-musllinux_1_1_x86_64.whl", hash = "sha256:95e5c750d55cf518c398a8240571b0e0782c2d5a703250872f36eaf737751338"}, - {file = "scipy-1.12.0-cp312-cp312-win_amd64.whl", hash = "sha256:e646d8571804a304e1da01040d21577685ce8e2db08ac58e543eaca063453e1c"}, - {file = "scipy-1.12.0-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:913d6e7956c3a671de3b05ccb66b11bc293f56bfdef040583a7221d9e22a2e35"}, - {file = "scipy-1.12.0-cp39-cp39-macosx_12_0_arm64.whl", hash = "sha256:bba1b0c7256ad75401c73e4b3cf09d1f176e9bd4248f0d3112170fb2ec4db067"}, - {file = "scipy-1.12.0-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:730badef9b827b368f351eacae2e82da414e13cf8bd5051b4bdfd720271a5371"}, - {file = "scipy-1.12.0-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:6546dc2c11a9df6926afcbdd8a3edec28566e4e785b915e849348c6dd9f3f490"}, - {file = "scipy-1.12.0-cp39-cp39-musllinux_1_1_x86_64.whl", hash = "sha256:196ebad3a4882081f62a5bf4aeb7326aa34b110e533aab23e4374fcccb0890dc"}, - {file = "scipy-1.12.0-cp39-cp39-win_amd64.whl", hash = "sha256:b360f1b6b2f742781299514e99ff560d1fe9bd1bff2712894b52abe528d1fd1e"}, - {file = "scipy-1.12.0.tar.gz", hash = "sha256:4bf5abab8a36d20193c698b0f1fc282c1d083c94723902c447e5d2f1780936a3"}, + {file = "scipy-1.13.1-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:20335853b85e9a49ff7572ab453794298bcf0354d8068c5f6775a0eabf350aca"}, + {file = "scipy-1.13.1-cp310-cp310-macosx_12_0_arm64.whl", hash = "sha256:d605e9c23906d1994f55ace80e0125c587f96c020037ea6aa98d01b4bd2e222f"}, + {file = "scipy-1.13.1-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:cfa31f1def5c819b19ecc3a8b52d28ffdcc7ed52bb20c9a7589669dd3c250989"}, + {file = "scipy-1.13.1-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:f26264b282b9da0952a024ae34710c2aff7d27480ee91a2e82b7b7073c24722f"}, + {file = "scipy-1.13.1-cp310-cp310-musllinux_1_1_x86_64.whl", hash = "sha256:eccfa1906eacc02de42d70ef4aecea45415f5be17e72b61bafcfd329bdc52e94"}, + {file = "scipy-1.13.1-cp310-cp310-win_amd64.whl", hash = "sha256:2831f0dc9c5ea9edd6e51e6e769b655f08ec6db6e2e10f86ef39bd32eb11da54"}, + {file = "scipy-1.13.1-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:27e52b09c0d3a1d5b63e1105f24177e544a222b43611aaf5bc44d4a0979e32f9"}, + {file = "scipy-1.13.1-cp311-cp311-macosx_12_0_arm64.whl", hash = "sha256:54f430b00f0133e2224c3ba42b805bfd0086fe488835effa33fa291561932326"}, + {file = "scipy-1.13.1-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:e89369d27f9e7b0884ae559a3a956e77c02114cc60a6058b4e5011572eea9299"}, + {file = "scipy-1.13.1-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:a78b4b3345f1b6f68a763c6e25c0c9a23a9fd0f39f5f3d200efe8feda560a5fa"}, + {file = "scipy-1.13.1-cp311-cp311-musllinux_1_1_x86_64.whl", hash = "sha256:45484bee6d65633752c490404513b9ef02475b4284c4cfab0ef946def50b3f59"}, + {file = "scipy-1.13.1-cp311-cp311-win_amd64.whl", hash = "sha256:5713f62f781eebd8d597eb3f88b8bf9274e79eeabf63afb4a737abc6c84ad37b"}, + {file = "scipy-1.13.1-cp312-cp312-macosx_10_9_x86_64.whl", hash = "sha256:5d72782f39716b2b3509cd7c33cdc08c96f2f4d2b06d51e52fb45a19ca0c86a1"}, + {file = "scipy-1.13.1-cp312-cp312-macosx_12_0_arm64.whl", hash = "sha256:017367484ce5498445aade74b1d5ab377acdc65e27095155e448c88497755a5d"}, + {file = "scipy-1.13.1-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:949ae67db5fa78a86e8fa644b9a6b07252f449dcf74247108c50e1d20d2b4627"}, + {file = "scipy-1.13.1-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:de3ade0e53bc1f21358aa74ff4830235d716211d7d077e340c7349bc3542e884"}, + {file = "scipy-1.13.1-cp312-cp312-musllinux_1_1_x86_64.whl", hash = "sha256:2ac65fb503dad64218c228e2dc2d0a0193f7904747db43014645ae139c8fad16"}, + {file = "scipy-1.13.1-cp312-cp312-win_amd64.whl", hash = "sha256:cdd7dacfb95fea358916410ec61bbc20440f7860333aee6d882bb8046264e949"}, + {file = "scipy-1.13.1-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:436bbb42a94a8aeef855d755ce5a465479c721e9d684de76bf61a62e7c2b81d5"}, + {file = "scipy-1.13.1-cp39-cp39-macosx_12_0_arm64.whl", hash = "sha256:8335549ebbca860c52bf3d02f80784e91a004b71b059e3eea9678ba994796a24"}, + {file = "scipy-1.13.1-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:d533654b7d221a6a97304ab63c41c96473ff04459e404b83275b60aa8f4b7004"}, + {file = "scipy-1.13.1-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:637e98dcf185ba7f8e663e122ebf908c4702420477ae52a04f9908707456ba4d"}, + {file = "scipy-1.13.1-cp39-cp39-musllinux_1_1_x86_64.whl", hash = "sha256:a014c2b3697bde71724244f63de2476925596c24285c7a637364761f8710891c"}, + {file = "scipy-1.13.1-cp39-cp39-win_amd64.whl", hash = "sha256:392e4ec766654852c25ebad4f64e4e584cf19820b980bc04960bca0b0cd6eaa2"}, + {file = "scipy-1.13.1.tar.gz", hash = "sha256:095a87a0312b08dfd6a6155cbbd310a8c51800fc931b8c0b84003014b874ed3c"}, ] [package.dependencies] -numpy = ">=1.22.4,<1.29.0" +numpy = ">=1.22.4,<2.3" [package.extras] -dev = ["click", "cython-lint (>=0.12.2)", "doit (>=0.36.0)", "mypy", "pycodestyle", "pydevtool", "rich-click", "ruff", "types-psutil", "typing_extensions"] -doc = ["jupytext", "matplotlib (>2)", "myst-nb", "numpydoc", "pooch", "pydata-sphinx-theme (==0.9.0)", "sphinx (!=4.1.0)", "sphinx-design (>=0.2.0)"] -test = ["asv", "gmpy2", "hypothesis", "mpmath", "pooch", "pytest", "pytest-cov", "pytest-timeout", "pytest-xdist", "scikit-umfpack", "threadpoolctl"] - -[[package]] -name = "setuptools" -version = "69.0.3" -description = "Easily download, build, install, upgrade, and uninstall Python packages" -optional = false -python-versions = ">=3.8" -files = [ - {file = "setuptools-69.0.3-py3-none-any.whl", hash = "sha256:385eb4edd9c9d5c17540511303e39a147ce2fc04bc55289c322b9e5904fe2c05"}, - {file = "setuptools-69.0.3.tar.gz", hash = "sha256:be1af57fc409f93647f2e8e4573a142ed38724b8cdd389706a867bb4efcf1e78"}, -] - -[package.extras] -docs = ["furo", "jaraco.packaging (>=9.3)", "jaraco.tidelift (>=1.4)", "pygments-github-lexers (==0.0.5)", "rst.linker (>=1.9)", "sphinx (<7.2.5)", "sphinx (>=3.5)", "sphinx-favicon", "sphinx-inline-tabs", "sphinx-lint", "sphinx-notfound-page (>=1,<2)", "sphinx-reredirects", "sphinxcontrib-towncrier"] -testing = ["build[virtualenv]", "filelock (>=3.4.0)", "flake8-2020", "ini2toml[lite] (>=0.9)", "jaraco.develop (>=7.21)", "jaraco.envs (>=2.2)", "jaraco.path (>=3.2.0)", "pip (>=19.1)", "pytest (>=6)", "pytest-black (>=0.3.7)", "pytest-checkdocs (>=2.4)", "pytest-cov", "pytest-enabler (>=2.2)", "pytest-mypy (>=0.9.1)", "pytest-perf", "pytest-ruff", "pytest-timeout", "pytest-xdist", "tomli-w (>=1.0.0)", "virtualenv (>=13.0.0)", "wheel"] -testing-integration = ["build[virtualenv] (>=1.0.3)", "filelock (>=3.4.0)", "jaraco.envs (>=2.2)", "jaraco.path (>=3.2.0)", "packaging (>=23.1)", "pytest", "pytest-enabler", "pytest-xdist", "tomli", "virtualenv (>=13.0.0)", "wheel"] +dev = ["cython-lint (>=0.12.2)", "doit (>=0.36.0)", "mypy", "pycodestyle", "pydevtool", "rich-click", "ruff", "types-psutil", "typing_extensions"] +doc = ["jupyterlite-pyodide-kernel", "jupyterlite-sphinx (>=0.12.0)", "jupytext", "matplotlib (>=3.5)", "myst-nb", "numpydoc", "pooch", "pydata-sphinx-theme (>=0.15.2)", "sphinx (>=5.0.0)", "sphinx-design (>=0.4.0)"] +test = ["array-api-strict", "asv", "gmpy2", "hypothesis (>=6.30)", "mpmath", "pooch", "pytest", "pytest-cov", "pytest-timeout", "pytest-xdist", "scikit-umfpack", "threadpoolctl"] [[package]] name = "shapely" -version = "2.0.2" +version = "2.0.4" description = "Manipulation and analysis of geometric objects" optional = false python-versions = ">=3.7" files = [ - {file = "shapely-2.0.2-cp310-cp310-macosx_10_9_universal2.whl", hash = "sha256:6ca8cffbe84ddde8f52b297b53f8e0687bd31141abb2c373fd8a9f032df415d6"}, - {file = "shapely-2.0.2-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:baa14fc27771e180c06b499a0a7ba697c7988c7b2b6cba9a929a19a4d2762de3"}, - {file = "shapely-2.0.2-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:36480e32c434d168cdf2f5e9862c84aaf4d714a43a8465ae3ce8ff327f0affb7"}, - {file = "shapely-2.0.2-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:4ef753200cbffd4f652efb2c528c5474e5a14341a473994d90ad0606522a46a2"}, - {file = "shapely-2.0.2-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:a9a41ff4323fc9d6257759c26eb1cf3a61ebc7e611e024e6091f42977303fd3a"}, - {file = "shapely-2.0.2-cp310-cp310-win32.whl", hash = "sha256:72b5997272ae8c25f0fd5b3b967b3237e87fab7978b8d6cd5fa748770f0c5d68"}, - {file = "shapely-2.0.2-cp310-cp310-win_amd64.whl", hash = "sha256:34eac2337cbd67650248761b140d2535855d21b969d76d76123317882d3a0c1a"}, - {file = "shapely-2.0.2-cp311-cp311-macosx_10_9_universal2.whl", hash = "sha256:5b0c052709c8a257c93b0d4943b0b7a3035f87e2d6a8ac9407b6a992d206422f"}, - {file = "shapely-2.0.2-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:2d217e56ae067e87b4e1731d0dc62eebe887ced729ba5c2d4590e9e3e9fdbd88"}, - {file = "shapely-2.0.2-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:94ac128ae2ab4edd0bffcd4e566411ea7bdc738aeaf92c32a8a836abad725f9f"}, - {file = "shapely-2.0.2-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:fa3ee28f5e63a130ec5af4dc3c4cb9c21c5788bb13c15e89190d163b14f9fb89"}, - {file = "shapely-2.0.2-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:737dba15011e5a9b54a8302f1748b62daa207c9bc06f820cd0ad32a041f1c6f2"}, - {file = "shapely-2.0.2-cp311-cp311-win32.whl", hash = "sha256:45ac6906cff0765455a7b49c1670af6e230c419507c13e2f75db638c8fc6f3bd"}, - {file = "shapely-2.0.2-cp311-cp311-win_amd64.whl", hash = "sha256:dc9342fc82e374130db86a955c3c4525bfbf315a248af8277a913f30911bed9e"}, - {file = "shapely-2.0.2-cp312-cp312-macosx_10_9_universal2.whl", hash = "sha256:06f193091a7c6112fc08dfd195a1e3846a64306f890b151fa8c63b3e3624202c"}, - {file = "shapely-2.0.2-cp312-cp312-macosx_10_9_x86_64.whl", hash = "sha256:eebe544df5c018134f3c23b6515877f7e4cd72851f88a8d0c18464f414d141a2"}, - {file = "shapely-2.0.2-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:7e92e7c255f89f5cdf777690313311f422aa8ada9a3205b187113274e0135cd8"}, - {file = "shapely-2.0.2-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:be46d5509b9251dd9087768eaf35a71360de6afac82ce87c636990a0871aa18b"}, - {file = "shapely-2.0.2-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:a5533a925d8e211d07636ffc2fdd9a7f9f13d54686d00577eeb11d16f00be9c4"}, - {file = "shapely-2.0.2-cp312-cp312-win32.whl", hash = "sha256:084b023dae8ad3d5b98acee9d3bf098fdf688eb0bb9b1401e8b075f6a627b611"}, - {file = "shapely-2.0.2-cp312-cp312-win_amd64.whl", hash = "sha256:ea84d1cdbcf31e619d672b53c4532f06253894185ee7acb8ceb78f5f33cbe033"}, - {file = "shapely-2.0.2-cp37-cp37m-macosx_10_9_x86_64.whl", hash = "sha256:ed1e99702125e7baccf401830a3b94d810d5c70b329b765fe93451fe14cf565b"}, - {file = "shapely-2.0.2-cp37-cp37m-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:e7d897e6bdc6bc64f7f65155dbbb30e49acaabbd0d9266b9b4041f87d6e52b3a"}, - {file = "shapely-2.0.2-cp37-cp37m-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:0521d76d1e8af01e712db71da9096b484f081e539d4f4a8c97342e7971d5e1b4"}, - {file = "shapely-2.0.2-cp37-cp37m-win32.whl", hash = "sha256:5324be299d4c533ecfcfd43424dfd12f9428fd6f12cda38a4316da001d6ef0ea"}, - {file = "shapely-2.0.2-cp37-cp37m-win_amd64.whl", hash = "sha256:78128357a0cee573257a0c2c388d4b7bf13cb7dbe5b3fe5d26d45ebbe2a39e25"}, - {file = "shapely-2.0.2-cp38-cp38-macosx_10_9_universal2.whl", hash = "sha256:87dc2be34ac3a3a4a319b963c507ac06682978a5e6c93d71917618b14f13066e"}, - {file = "shapely-2.0.2-cp38-cp38-macosx_10_9_x86_64.whl", hash = "sha256:42997ac806e4583dad51c80a32d38570fd9a3d4778f5e2c98f9090aa7db0fe91"}, - {file = "shapely-2.0.2-cp38-cp38-macosx_11_0_arm64.whl", hash = "sha256:ccfd5fa10a37e67dbafc601c1ddbcbbfef70d34c3f6b0efc866ddbdb55893a6c"}, - {file = "shapely-2.0.2-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:e7c95d3379ae3abb74058938a9fcbc478c6b2e28d20dace38f8b5c587dde90aa"}, - {file = "shapely-2.0.2-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:6a21353d28209fb0d8cc083e08ca53c52666e0d8a1f9bbe23b6063967d89ed24"}, - {file = "shapely-2.0.2-cp38-cp38-win32.whl", hash = "sha256:03e63a99dfe6bd3beb8d5f41ec2086585bb969991d603f9aeac335ad396a06d4"}, - {file = "shapely-2.0.2-cp38-cp38-win_amd64.whl", hash = "sha256:c6fd29fbd9cd76350bd5cc14c49de394a31770aed02d74203e23b928f3d2f1aa"}, - {file = "shapely-2.0.2-cp39-cp39-macosx_10_9_universal2.whl", hash = "sha256:1f217d28ecb48e593beae20a0082a95bd9898d82d14b8fcb497edf6bff9a44d7"}, - {file = "shapely-2.0.2-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:394e5085b49334fd5b94fa89c086edfb39c3ecab7f669e8b2a4298b9d523b3a5"}, - {file = "shapely-2.0.2-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:fd3ad17b64466a033848c26cb5b509625c87d07dcf39a1541461cacdb8f7e91c"}, - {file = "shapely-2.0.2-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:d41a116fcad58048d7143ddb01285e1a8780df6dc1f56c3b1e1b7f12ed296651"}, - {file = "shapely-2.0.2-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:dea9a0651333cf96ef5bb2035044e3ad6a54f87d90e50fe4c2636debf1b77abc"}, - {file = "shapely-2.0.2-cp39-cp39-win32.whl", hash = "sha256:b8eb0a92f7b8c74f9d8fdd1b40d395113f59bd8132ca1348ebcc1f5aece94b96"}, - {file = "shapely-2.0.2-cp39-cp39-win_amd64.whl", hash = "sha256:794affd80ca0f2c536fc948a3afa90bd8fb61ebe37fe873483ae818e7f21def4"}, - {file = "shapely-2.0.2.tar.gz", hash = "sha256:1713cc04c171baffc5b259ba8531c58acc2a301707b7f021d88a15ed090649e7"}, + {file = "shapely-2.0.4-cp310-cp310-macosx_10_9_universal2.whl", hash = "sha256:011b77153906030b795791f2fdfa2d68f1a8d7e40bce78b029782ade3afe4f2f"}, + {file = "shapely-2.0.4-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:9831816a5d34d5170aa9ed32a64982c3d6f4332e7ecfe62dc97767e163cb0b17"}, + {file = "shapely-2.0.4-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:5c4849916f71dc44e19ed370421518c0d86cf73b26e8656192fcfcda08218fbd"}, + {file = "shapely-2.0.4-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:841f93a0e31e4c64d62ea570d81c35de0f6cea224568b2430d832967536308e6"}, + {file = "shapely-2.0.4-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:d2b4431f522b277c79c34b65da128029a9955e4481462cbf7ebec23aab61fc58"}, + {file = "shapely-2.0.4-cp310-cp310-win32.whl", hash = "sha256:92a41d936f7d6743f343be265ace93b7c57f5b231e21b9605716f5a47c2879e7"}, + {file = "shapely-2.0.4-cp310-cp310-win_amd64.whl", hash = "sha256:30982f79f21bb0ff7d7d4a4e531e3fcaa39b778584c2ce81a147f95be1cd58c9"}, + {file = "shapely-2.0.4-cp311-cp311-macosx_10_9_universal2.whl", hash = "sha256:de0205cb21ad5ddaef607cda9a3191eadd1e7a62a756ea3a356369675230ac35"}, + {file = "shapely-2.0.4-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:7d56ce3e2a6a556b59a288771cf9d091470116867e578bebced8bfc4147fbfd7"}, + {file = "shapely-2.0.4-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:58b0ecc505bbe49a99551eea3f2e8a9b3b24b3edd2a4de1ac0dc17bc75c9ec07"}, + {file = "shapely-2.0.4-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:790a168a808bd00ee42786b8ba883307c0e3684ebb292e0e20009588c426da47"}, + {file = "shapely-2.0.4-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:4310b5494271e18580d61022c0857eb85d30510d88606fa3b8314790df7f367d"}, + {file = "shapely-2.0.4-cp311-cp311-win32.whl", hash = "sha256:63f3a80daf4f867bd80f5c97fbe03314348ac1b3b70fb1c0ad255a69e3749879"}, + {file = "shapely-2.0.4-cp311-cp311-win_amd64.whl", hash = "sha256:c52ed79f683f721b69a10fb9e3d940a468203f5054927215586c5d49a072de8d"}, + {file = "shapely-2.0.4-cp312-cp312-macosx_10_9_universal2.whl", hash = "sha256:5bbd974193e2cc274312da16b189b38f5f128410f3377721cadb76b1e8ca5328"}, + {file = "shapely-2.0.4-cp312-cp312-macosx_10_9_x86_64.whl", hash = "sha256:41388321a73ba1a84edd90d86ecc8bfed55e6a1e51882eafb019f45895ec0f65"}, + {file = "shapely-2.0.4-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:0776c92d584f72f1e584d2e43cfc5542c2f3dd19d53f70df0900fda643f4bae6"}, + {file = "shapely-2.0.4-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:c75c98380b1ede1cae9a252c6dc247e6279403fae38c77060a5e6186c95073ac"}, + {file = "shapely-2.0.4-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:c3e700abf4a37b7b8b90532fa6ed5c38a9bfc777098bc9fbae5ec8e618ac8f30"}, + {file = "shapely-2.0.4-cp312-cp312-win32.whl", hash = "sha256:4f2ab0faf8188b9f99e6a273b24b97662194160cc8ca17cf9d1fb6f18d7fb93f"}, + {file = "shapely-2.0.4-cp312-cp312-win_amd64.whl", hash = "sha256:03152442d311a5e85ac73b39680dd64a9892fa42bb08fd83b3bab4fe6999bfa0"}, + {file = "shapely-2.0.4-cp37-cp37m-macosx_10_9_x86_64.whl", hash = "sha256:994c244e004bc3cfbea96257b883c90a86e8cbd76e069718eb4c6b222a56f78b"}, + {file = "shapely-2.0.4-cp37-cp37m-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:05ffd6491e9e8958b742b0e2e7c346635033d0a5f1a0ea083547fcc854e5d5cf"}, + {file = "shapely-2.0.4-cp37-cp37m-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:2fbdc1140a7d08faa748256438291394967aa54b40009f54e8d9825e75ef6113"}, + {file = "shapely-2.0.4-cp37-cp37m-win32.whl", hash = "sha256:5af4cd0d8cf2912bd95f33586600cac9c4b7c5053a036422b97cfe4728d2eb53"}, + {file = "shapely-2.0.4-cp37-cp37m-win_amd64.whl", hash = "sha256:464157509ce4efa5ff285c646a38b49f8c5ef8d4b340f722685b09bb033c5ccf"}, + {file = "shapely-2.0.4-cp38-cp38-macosx_10_9_universal2.whl", hash = "sha256:489c19152ec1f0e5c5e525356bcbf7e532f311bff630c9b6bc2db6f04da6a8b9"}, + {file = "shapely-2.0.4-cp38-cp38-macosx_10_9_x86_64.whl", hash = "sha256:b79bbd648664aa6f44ef018474ff958b6b296fed5c2d42db60078de3cffbc8aa"}, + {file = "shapely-2.0.4-cp38-cp38-macosx_11_0_arm64.whl", hash = "sha256:674d7baf0015a6037d5758496d550fc1946f34bfc89c1bf247cabdc415d7747e"}, + {file = "shapely-2.0.4-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:6cd4ccecc5ea5abd06deeaab52fcdba372f649728050c6143cc405ee0c166679"}, + {file = "shapely-2.0.4-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:fb5cdcbbe3080181498931b52a91a21a781a35dcb859da741c0345c6402bf00c"}, + {file = "shapely-2.0.4-cp38-cp38-win32.whl", hash = "sha256:55a38dcd1cee2f298d8c2ebc60fc7d39f3b4535684a1e9e2f39a80ae88b0cea7"}, + {file = "shapely-2.0.4-cp38-cp38-win_amd64.whl", hash = "sha256:ec555c9d0db12d7fd777ba3f8b75044c73e576c720a851667432fabb7057da6c"}, + {file = "shapely-2.0.4-cp39-cp39-macosx_10_9_universal2.whl", hash = "sha256:3f9103abd1678cb1b5f7e8e1af565a652e036844166c91ec031eeb25c5ca8af0"}, + {file = "shapely-2.0.4-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:263bcf0c24d7a57c80991e64ab57cba7a3906e31d2e21b455f493d4aab534aaa"}, + {file = "shapely-2.0.4-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:ddf4a9bfaac643e62702ed662afc36f6abed2a88a21270e891038f9a19bc08fc"}, + {file = "shapely-2.0.4-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:485246fcdb93336105c29a5cfbff8a226949db37b7473c89caa26c9bae52a242"}, + {file = "shapely-2.0.4-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:8de4578e838a9409b5b134a18ee820730e507b2d21700c14b71a2b0757396acc"}, + {file = "shapely-2.0.4-cp39-cp39-win32.whl", hash = "sha256:9dab4c98acfb5fb85f5a20548b5c0abe9b163ad3525ee28822ffecb5c40e724c"}, + {file = "shapely-2.0.4-cp39-cp39-win_amd64.whl", hash = "sha256:31c19a668b5a1eadab82ff070b5a260478ac6ddad3a5b62295095174a8d26398"}, + {file = "shapely-2.0.4.tar.gz", hash = "sha256:5dc736127fac70009b8d309a0eeb74f3e08979e530cf7017f2f507ef62e6cfb8"}, ] [package.dependencies] -numpy = ">=1.14" +numpy = ">=1.14,<3" [package.extras] docs = ["matplotlib", "numpydoc (==1.1.)", "sphinx", "sphinx-book-theme", "sphinx-remove-toctrees"] @@ -4236,17 +4254,18 @@ files = [ [[package]] name = "tenacity" -version = "8.2.3" +version = "8.3.0" description = "Retry code until it succeeds" optional = false -python-versions = ">=3.7" +python-versions = ">=3.8" files = [ - {file = "tenacity-8.2.3-py3-none-any.whl", hash = "sha256:ce510e327a630c9e1beaf17d42e6ffacc88185044ad85cf74c0a8887c6a0f88c"}, - {file = "tenacity-8.2.3.tar.gz", hash = "sha256:5398ef0d78e63f40007c1fb4c0bff96e1911394d2fa8d194f77619c05ff6cc8a"}, + {file = "tenacity-8.3.0-py3-none-any.whl", hash = "sha256:3649f6443dbc0d9b01b9d8020a9c4ec7a1ff5f6f3c6c8a036ef371f573fe9185"}, + {file = "tenacity-8.3.0.tar.gz", hash = "sha256:953d4e6ad24357bceffbc9707bc74349aca9d245f68eb65419cf0c249a1949a2"}, ] [package.extras] -doc = ["reno", "sphinx", "tornado (>=4.5)"] +doc = ["reno", "sphinx"] +test = ["pytest", "tornado (>=4.5)", "typeguard"] [[package]] name = "tomli" @@ -4292,13 +4311,13 @@ files = [ [[package]] name = "tqdm" -version = "4.66.1" +version = "4.66.4" description = "Fast, Extensible Progress Meter" optional = false python-versions = ">=3.7" files = [ - {file = "tqdm-4.66.1-py3-none-any.whl", hash = "sha256:d302b3c5b53d47bce91fea46679d9c3c6508cf6332229aa1e7d8653723793386"}, - {file = "tqdm-4.66.1.tar.gz", hash = "sha256:d88e651f9db8d8551a62556d3cff9e3034274ca5d66e93197cf2490e2dcb69c7"}, + {file = "tqdm-4.66.4-py3-none-any.whl", hash = "sha256:b75ca56b413b030bc3f00af51fd2c1a1a5eac6a0c1cca83cbb37a5c52abce644"}, + {file = "tqdm-4.66.4.tar.gz", hash = "sha256:e4d936c9de8727928f3be6079590e97d9abfe8d39a590be678eb5919ffc186bb"}, ] [package.dependencies] @@ -4312,72 +4331,72 @@ telegram = ["requests"] [[package]] name = "traitlets" -version = "5.14.1" +version = "5.14.3" description = "Traitlets Python configuration system" optional = false python-versions = ">=3.8" files = [ - {file = "traitlets-5.14.1-py3-none-any.whl", hash = "sha256:2e5a030e6eff91737c643231bfcf04a65b0132078dad75e4936700b213652e74"}, - {file = "traitlets-5.14.1.tar.gz", hash = "sha256:8585105b371a04b8316a43d5ce29c098575c2e477850b62b848b964f1444527e"}, + {file = "traitlets-5.14.3-py3-none-any.whl", hash = "sha256:b74e89e397b1ed28cc831db7aea759ba6640cb3de13090ca145426688ff1ac4f"}, + {file = "traitlets-5.14.3.tar.gz", hash = "sha256:9ed0579d3502c94b4b3732ac120375cda96f923114522847de4b3bb98b96b6b7"}, ] [package.extras] docs = ["myst-parser", "pydata-sphinx-theme", "sphinx"] -test = ["argcomplete (>=3.0.3)", "mypy (>=1.7.0)", "pre-commit", "pytest (>=7.0,<7.5)", "pytest-mock", "pytest-mypy-testing"] +test = ["argcomplete (>=3.0.3)", "mypy (>=1.7.0)", "pre-commit", "pytest (>=7.0,<8.2)", "pytest-mock", "pytest-mypy-testing"] [[package]] name = "types-decorator" -version = "5.1.8.20240106" +version = "5.1.8.20240310" description = "Typing stubs for decorator" optional = false python-versions = ">=3.8" files = [ - {file = "types-decorator-5.1.8.20240106.tar.gz", hash = "sha256:32ff92b33615060d23b9d3760124bdb3506c4aa8d9eb50963cf1a3c20b9ecbbf"}, - {file = "types_decorator-5.1.8.20240106-py3-none-any.whl", hash = "sha256:14d21e6a0755dbb8f301f2f532b3eab5148f433c69dad2d98bf5bd2b3a2ef4e7"}, + {file = "types-decorator-5.1.8.20240310.tar.gz", hash = "sha256:52e316b03783886a8a2abdc228f7071680ba65894545cd2085ebe3cf88684a0e"}, + {file = "types_decorator-5.1.8.20240310-py3-none-any.whl", hash = "sha256:3af75dc38f5baf65b9b53ea6661ce2056c5ca7d70d620d0b1f620285c1242757"}, ] [[package]] name = "types-pytz" -version = "2024.1.0.20240203" +version = "2024.1.0.20240417" description = "Typing stubs for pytz" optional = false python-versions = ">=3.8" files = [ - {file = "types-pytz-2024.1.0.20240203.tar.gz", hash = "sha256:c93751ee20dfc6e054a0148f8f5227b9a00b79c90a4d3c9f464711a73179c89e"}, - {file = "types_pytz-2024.1.0.20240203-py3-none-any.whl", hash = "sha256:9679eef0365db3af91ef7722c199dbb75ee5c1b67e3c4dd7bfbeb1b8a71c21a3"}, + {file = "types-pytz-2024.1.0.20240417.tar.gz", hash = "sha256:6810c8a1f68f21fdf0f4f374a432487c77645a0ac0b31de4bf4690cf21ad3981"}, + {file = "types_pytz-2024.1.0.20240417-py3-none-any.whl", hash = "sha256:8335d443310e2db7b74e007414e74c4f53b67452c0cb0d228ca359ccfba59659"}, ] [[package]] name = "typing-extensions" -version = "4.9.0" +version = "4.12.0" description = "Backported and Experimental Type Hints for Python 3.8+" optional = false python-versions = ">=3.8" files = [ - {file = "typing_extensions-4.9.0-py3-none-any.whl", hash = "sha256:af72aea155e91adfc61c3ae9e0e342dbc0cba726d6cba4b6c72c1f34e47291cd"}, - {file = "typing_extensions-4.9.0.tar.gz", hash = "sha256:23478f88c37f27d76ac8aee6c905017a143b0b1b886c3c9f66bc2fd94f9f5783"}, + {file = "typing_extensions-4.12.0-py3-none-any.whl", hash = "sha256:b349c66bea9016ac22978d800cfff206d5f9816951f12a7d0ec5578b0a819594"}, + {file = "typing_extensions-4.12.0.tar.gz", hash = "sha256:8cbcdc8606ebcb0d95453ad7dc5065e6237b6aa230a31e81d0f440c30fed5fd8"}, ] [[package]] name = "tzdata" -version = "2023.4" +version = "2024.1" description = "Provider of IANA time zone data" optional = false python-versions = ">=2" files = [ - {file = "tzdata-2023.4-py2.py3-none-any.whl", hash = "sha256:aa3ace4329eeacda5b7beb7ea08ece826c28d761cda36e747cfbf97996d39bf3"}, - {file = "tzdata-2023.4.tar.gz", hash = "sha256:dd54c94f294765522c77399649b4fefd95522479a664a0cec87f41bebc6148c9"}, + {file = "tzdata-2024.1-py2.py3-none-any.whl", hash = "sha256:9068bc196136463f5245e51efda838afa15aaeca9903f49050dfa2679db4d252"}, + {file = "tzdata-2024.1.tar.gz", hash = "sha256:2674120f8d891909751c38abcdfd386ac0a5a1127954fbc332af6b5ceae07efd"}, ] [[package]] name = "uc-micro-py" -version = "1.0.2" +version = "1.0.3" description = "Micro subset of unicode data files for linkify-it-py projects." optional = false python-versions = ">=3.7" files = [ - {file = "uc-micro-py-1.0.2.tar.gz", hash = "sha256:30ae2ac9c49f39ac6dce743bd187fcd2b574b16ca095fa74cd9396795c954c54"}, - {file = "uc_micro_py-1.0.2-py3-none-any.whl", hash = "sha256:8c9110c309db9d9e87302e2f4ad2c3152770930d88ab385cd544e7a7e75f3de0"}, + {file = "uc-micro-py-1.0.3.tar.gz", hash = "sha256:d321b92cff673ec58027c04015fcaa8bb1e005478643ff4a500882eaab88c48a"}, + {file = "uc_micro_py-1.0.3-py3-none-any.whl", hash = "sha256:db1dffff340817673d7b466ec86114a9dc0e9d4d9b5ba229d9d60e5c12600cd5"}, ] [package.extras] @@ -4385,13 +4404,13 @@ test = ["coverage", "pytest", "pytest-cov"] [[package]] name = "urllib3" -version = "2.2.0" +version = "2.2.1" description = "HTTP library with thread-safe connection pooling, file post, and more." optional = false python-versions = ">=3.8" files = [ - {file = "urllib3-2.2.0-py3-none-any.whl", hash = "sha256:ce3711610ddce217e6d113a2732fafad960a03fd0318c91faa79481e35c11224"}, - {file = "urllib3-2.2.0.tar.gz", hash = "sha256:051d961ad0c62a94e50ecf1af379c3aba230c66c710493493560c0c223c49f20"}, + {file = "urllib3-2.2.1-py3-none-any.whl", hash = "sha256:450b20ec296a467077128bff42b73080516e71b56ff59a60a02bef2232c4fa9d"}, + {file = "urllib3-2.2.1.tar.gz", hash = "sha256:d0570876c61ab9e520d776c38acbbb5b05a776d3f9ff98a5c8fd5162a444cf19"}, ] [package.extras] @@ -4402,38 +4421,43 @@ zstd = ["zstandard (>=0.18.0)"] [[package]] name = "watchdog" -version = "3.0.0" +version = "4.0.1" description = "Filesystem events monitoring" optional = false -python-versions = ">=3.7" +python-versions = ">=3.8" files = [ - {file = "watchdog-3.0.0-cp310-cp310-macosx_10_9_universal2.whl", hash = "sha256:336adfc6f5cc4e037d52db31194f7581ff744b67382eb6021c868322e32eef41"}, - {file = "watchdog-3.0.0-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:a70a8dcde91be523c35b2bf96196edc5730edb347e374c7de7cd20c43ed95397"}, - {file = "watchdog-3.0.0-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:adfdeab2da79ea2f76f87eb42a3ab1966a5313e5a69a0213a3cc06ef692b0e96"}, - {file = "watchdog-3.0.0-cp311-cp311-macosx_10_9_universal2.whl", hash = "sha256:2b57a1e730af3156d13b7fdddfc23dea6487fceca29fc75c5a868beed29177ae"}, - {file = "watchdog-3.0.0-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:7ade88d0d778b1b222adebcc0927428f883db07017618a5e684fd03b83342bd9"}, - {file = "watchdog-3.0.0-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:7e447d172af52ad204d19982739aa2346245cc5ba6f579d16dac4bfec226d2e7"}, - {file = "watchdog-3.0.0-cp37-cp37m-macosx_10_9_x86_64.whl", hash = "sha256:9fac43a7466eb73e64a9940ac9ed6369baa39b3bf221ae23493a9ec4d0022674"}, - {file = "watchdog-3.0.0-cp38-cp38-macosx_10_9_universal2.whl", hash = "sha256:8ae9cda41fa114e28faf86cb137d751a17ffd0316d1c34ccf2235e8a84365c7f"}, - {file = "watchdog-3.0.0-cp38-cp38-macosx_10_9_x86_64.whl", hash = "sha256:25f70b4aa53bd743729c7475d7ec41093a580528b100e9a8c5b5efe8899592fc"}, - {file = "watchdog-3.0.0-cp38-cp38-macosx_11_0_arm64.whl", hash = "sha256:4f94069eb16657d2c6faada4624c39464f65c05606af50bb7902e036e3219be3"}, - {file = "watchdog-3.0.0-cp39-cp39-macosx_10_9_universal2.whl", hash = "sha256:7c5f84b5194c24dd573fa6472685b2a27cc5a17fe5f7b6fd40345378ca6812e3"}, - {file = "watchdog-3.0.0-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:3aa7f6a12e831ddfe78cdd4f8996af9cf334fd6346531b16cec61c3b3c0d8da0"}, - {file = "watchdog-3.0.0-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:233b5817932685d39a7896b1090353fc8efc1ef99c9c054e46c8002561252fb8"}, - {file = "watchdog-3.0.0-pp37-pypy37_pp73-macosx_10_9_x86_64.whl", hash = "sha256:13bbbb462ee42ec3c5723e1205be8ced776f05b100e4737518c67c8325cf6100"}, - {file = "watchdog-3.0.0-pp38-pypy38_pp73-macosx_10_9_x86_64.whl", hash = "sha256:8f3ceecd20d71067c7fd4c9e832d4e22584318983cabc013dbf3f70ea95de346"}, - {file = "watchdog-3.0.0-pp39-pypy39_pp73-macosx_10_9_x86_64.whl", hash = "sha256:c9d8c8ec7efb887333cf71e328e39cffbf771d8f8f95d308ea4125bf5f90ba64"}, - {file = "watchdog-3.0.0-py3-none-manylinux2014_aarch64.whl", hash = "sha256:0e06ab8858a76e1219e68c7573dfeba9dd1c0219476c5a44d5333b01d7e1743a"}, - {file = "watchdog-3.0.0-py3-none-manylinux2014_armv7l.whl", hash = "sha256:d00e6be486affb5781468457b21a6cbe848c33ef43f9ea4a73b4882e5f188a44"}, - {file = "watchdog-3.0.0-py3-none-manylinux2014_i686.whl", hash = "sha256:c07253088265c363d1ddf4b3cdb808d59a0468ecd017770ed716991620b8f77a"}, - {file = "watchdog-3.0.0-py3-none-manylinux2014_ppc64.whl", hash = "sha256:5113334cf8cf0ac8cd45e1f8309a603291b614191c9add34d33075727a967709"}, - {file = "watchdog-3.0.0-py3-none-manylinux2014_ppc64le.whl", hash = "sha256:51f90f73b4697bac9c9a78394c3acbbd331ccd3655c11be1a15ae6fe289a8c83"}, - {file = "watchdog-3.0.0-py3-none-manylinux2014_s390x.whl", hash = "sha256:ba07e92756c97e3aca0912b5cbc4e5ad802f4557212788e72a72a47ff376950d"}, - {file = "watchdog-3.0.0-py3-none-manylinux2014_x86_64.whl", hash = "sha256:d429c2430c93b7903914e4db9a966c7f2b068dd2ebdd2fa9b9ce094c7d459f33"}, - {file = "watchdog-3.0.0-py3-none-win32.whl", hash = "sha256:3ed7c71a9dccfe838c2f0b6314ed0d9b22e77d268c67e015450a29036a81f60f"}, - {file = "watchdog-3.0.0-py3-none-win_amd64.whl", hash = "sha256:4c9956d27be0bb08fc5f30d9d0179a855436e655f046d288e2bcc11adfae893c"}, - {file = "watchdog-3.0.0-py3-none-win_ia64.whl", hash = "sha256:5d9f3a10e02d7371cd929b5d8f11e87d4bad890212ed3901f9b4d68767bee759"}, - {file = "watchdog-3.0.0.tar.gz", hash = "sha256:4d98a320595da7a7c5a18fc48cb633c2e73cda78f93cac2ef42d42bf609a33f9"}, + {file = "watchdog-4.0.1-cp310-cp310-macosx_10_9_universal2.whl", hash = "sha256:da2dfdaa8006eb6a71051795856bedd97e5b03e57da96f98e375682c48850645"}, + {file = "watchdog-4.0.1-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:e93f451f2dfa433d97765ca2634628b789b49ba8b504fdde5837cdcf25fdb53b"}, + {file = "watchdog-4.0.1-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:ef0107bbb6a55f5be727cfc2ef945d5676b97bffb8425650dadbb184be9f9a2b"}, + {file = "watchdog-4.0.1-cp311-cp311-macosx_10_9_universal2.whl", hash = "sha256:17e32f147d8bf9657e0922c0940bcde863b894cd871dbb694beb6704cfbd2fb5"}, + {file = "watchdog-4.0.1-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:03e70d2df2258fb6cb0e95bbdbe06c16e608af94a3ffbd2b90c3f1e83eb10767"}, + {file = "watchdog-4.0.1-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:123587af84260c991dc5f62a6e7ef3d1c57dfddc99faacee508c71d287248459"}, + {file = "watchdog-4.0.1-cp312-cp312-macosx_10_9_universal2.whl", hash = "sha256:093b23e6906a8b97051191a4a0c73a77ecc958121d42346274c6af6520dec175"}, + {file = "watchdog-4.0.1-cp312-cp312-macosx_10_9_x86_64.whl", hash = "sha256:611be3904f9843f0529c35a3ff3fd617449463cb4b73b1633950b3d97fa4bfb7"}, + {file = "watchdog-4.0.1-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:62c613ad689ddcb11707f030e722fa929f322ef7e4f18f5335d2b73c61a85c28"}, + {file = "watchdog-4.0.1-cp38-cp38-macosx_10_9_universal2.whl", hash = "sha256:d4925e4bf7b9bddd1c3de13c9b8a2cdb89a468f640e66fbfabaf735bd85b3e35"}, + {file = "watchdog-4.0.1-cp38-cp38-macosx_10_9_x86_64.whl", hash = "sha256:cad0bbd66cd59fc474b4a4376bc5ac3fc698723510cbb64091c2a793b18654db"}, + {file = "watchdog-4.0.1-cp38-cp38-macosx_11_0_arm64.whl", hash = "sha256:a3c2c317a8fb53e5b3d25790553796105501a235343f5d2bf23bb8649c2c8709"}, + {file = "watchdog-4.0.1-cp39-cp39-macosx_10_9_universal2.whl", hash = "sha256:c9904904b6564d4ee8a1ed820db76185a3c96e05560c776c79a6ce5ab71888ba"}, + {file = "watchdog-4.0.1-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:667f3c579e813fcbad1b784db7a1aaa96524bed53437e119f6a2f5de4db04235"}, + {file = "watchdog-4.0.1-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:d10a681c9a1d5a77e75c48a3b8e1a9f2ae2928eda463e8d33660437705659682"}, + {file = "watchdog-4.0.1-pp310-pypy310_pp73-macosx_10_9_x86_64.whl", hash = "sha256:0144c0ea9997b92615af1d94afc0c217e07ce2c14912c7b1a5731776329fcfc7"}, + {file = "watchdog-4.0.1-pp310-pypy310_pp73-macosx_11_0_arm64.whl", hash = "sha256:998d2be6976a0ee3a81fb8e2777900c28641fb5bfbd0c84717d89bca0addcdc5"}, + {file = "watchdog-4.0.1-pp38-pypy38_pp73-macosx_10_9_x86_64.whl", hash = "sha256:e7921319fe4430b11278d924ef66d4daa469fafb1da679a2e48c935fa27af193"}, + {file = "watchdog-4.0.1-pp38-pypy38_pp73-macosx_11_0_arm64.whl", hash = "sha256:f0de0f284248ab40188f23380b03b59126d1479cd59940f2a34f8852db710625"}, + {file = "watchdog-4.0.1-pp39-pypy39_pp73-macosx_10_9_x86_64.whl", hash = "sha256:bca36be5707e81b9e6ce3208d92d95540d4ca244c006b61511753583c81c70dd"}, + {file = "watchdog-4.0.1-pp39-pypy39_pp73-macosx_11_0_arm64.whl", hash = "sha256:ab998f567ebdf6b1da7dc1e5accfaa7c6992244629c0fdaef062f43249bd8dee"}, + {file = "watchdog-4.0.1-py3-none-manylinux2014_aarch64.whl", hash = "sha256:dddba7ca1c807045323b6af4ff80f5ddc4d654c8bce8317dde1bd96b128ed253"}, + {file = "watchdog-4.0.1-py3-none-manylinux2014_armv7l.whl", hash = "sha256:4513ec234c68b14d4161440e07f995f231be21a09329051e67a2118a7a612d2d"}, + {file = "watchdog-4.0.1-py3-none-manylinux2014_i686.whl", hash = "sha256:4107ac5ab936a63952dea2a46a734a23230aa2f6f9db1291bf171dac3ebd53c6"}, + {file = "watchdog-4.0.1-py3-none-manylinux2014_ppc64.whl", hash = "sha256:6e8c70d2cd745daec2a08734d9f63092b793ad97612470a0ee4cbb8f5f705c57"}, + {file = "watchdog-4.0.1-py3-none-manylinux2014_ppc64le.whl", hash = "sha256:f27279d060e2ab24c0aa98363ff906d2386aa6c4dc2f1a374655d4e02a6c5e5e"}, + {file = "watchdog-4.0.1-py3-none-manylinux2014_s390x.whl", hash = "sha256:f8affdf3c0f0466e69f5b3917cdd042f89c8c63aebdb9f7c078996f607cdb0f5"}, + {file = "watchdog-4.0.1-py3-none-manylinux2014_x86_64.whl", hash = "sha256:ac7041b385f04c047fcc2951dc001671dee1b7e0615cde772e84b01fbf68ee84"}, + {file = "watchdog-4.0.1-py3-none-win32.whl", hash = "sha256:206afc3d964f9a233e6ad34618ec60b9837d0582b500b63687e34011e15bb429"}, + {file = "watchdog-4.0.1-py3-none-win_amd64.whl", hash = "sha256:7577b3c43e5909623149f76b099ac49a1a01ca4e167d1785c76eb52fa585745a"}, + {file = "watchdog-4.0.1-py3-none-win_ia64.whl", hash = "sha256:d7b9f5f3299e8dd230880b6c55504a1f69cf1e4316275d1b215ebdd8187ec88d"}, + {file = "watchdog-4.0.1.tar.gz", hash = "sha256:eebaacf674fa25511e8867028d281e602ee6500045b57f43b08778082f7f8b44"}, ] [package.extras] @@ -4478,13 +4502,13 @@ testing = ["coverage", "pytest (>=3.1.0)", "pytest-cov", "pytest-xdist"] [[package]] name = "xarray" -version = "2024.1.1" +version = "2024.5.0" description = "N-D labeled arrays and datasets in Python" optional = false python-versions = ">=3.9" files = [ - {file = "xarray-2024.1.1-py3-none-any.whl", hash = "sha256:0bec81303b088c8df4f075e1579c00cfd7e5069688e4434007f0b8d7df17fc1c"}, - {file = "xarray-2024.1.1.tar.gz", hash = "sha256:a1ba2d87a74892e213c9c83f4a462dbcdf68212320a4e31b34bd789ba7a64e35"}, + {file = "xarray-2024.5.0-py3-none-any.whl", hash = "sha256:7ddedfe2294a0ab00f02d0fbdcb9c6300ec589f3cf436a9c7b7b577a12cd9bcf"}, + {file = "xarray-2024.5.0.tar.gz", hash = "sha256:e0eb1cb265f265126795f388ed9591f3c752f2aca491f6c0576711fd15b708f2"}, ] [package.dependencies] @@ -4497,8 +4521,8 @@ netCDF4 = {version = "", optional = true, markers = "extra == \"io\""} numbagg = {version = "", optional = true, markers = "extra == \"accel\""} numpy = ">=1.23" opt-einsum = {version = "", optional = true, markers = "extra == \"accel\""} -packaging = ">=22" -pandas = ">=1.5" +packaging = ">=23.1" +pandas = ">=2.0" pooch = {version = "", optional = true, markers = "extra == \"io\""} pydap = {version = "", optional = true, markers = "python_version < \"3.10\" and extra == \"io\""} scipy = {version = "", optional = true, markers = "extra == \"accel\" or extra == \"io\""} @@ -4506,20 +4530,39 @@ zarr = {version = "", optional = true, markers = "extra == \"io\""} [package.extras] accel = ["bottleneck", "flox", "numbagg", "opt-einsum", "scipy"] -complete = ["xarray[accel,io,parallel,viz]"] +complete = ["xarray[accel,dev,io,parallel,viz]"] +dev = ["hypothesis", "mypy", "pre-commit", "pytest", "pytest-cov", "pytest-env", "pytest-timeout", "pytest-xdist", "ruff", "xarray[complete]"] io = ["cftime", "fsspec", "h5netcdf", "netCDF4", "pooch", "pydap", "scipy", "zarr"] parallel = ["dask[complete]"] viz = ["matplotlib", "nc-time-axis", "seaborn"] +[[package]] +name = "xarray-selafin" +version = "0.1.6" +description = "" +optional = false +python-versions = ">=3.9" +files = [ + {file = "xarray_selafin-0.1.6-py3-none-any.whl", hash = "sha256:e484ae6e26ceb3c91f6a7a4975b171146a94171bf6472dea33a09c90a4e15ad5"}, + {file = "xarray_selafin-0.1.6.tar.gz", hash = "sha256:4fab755eb3089571c1d43b3d84491e4ffa80cdbb0977aa893a63f156be44b493"}, +] + +[package.dependencies] +netcdf4 = "" +numpy = "" +scipy = "" +shapely = "" +xarray = "" + [[package]] name = "xyzservices" -version = "2023.10.1" +version = "2024.4.0" description = "Source of XYZ tiles providers" optional = false python-versions = ">=3.8" files = [ - {file = "xyzservices-2023.10.1-py3-none-any.whl", hash = "sha256:6a4c38d3a9f89d3e77153eff9414b36a8ee0850c9e8b85796fd1b2a85b8dfd68"}, - {file = "xyzservices-2023.10.1.tar.gz", hash = "sha256:091229269043bc8258042edbedad4fcb44684b0473ede027b5672ad40dc9fa02"}, + {file = "xyzservices-2024.4.0-py3-none-any.whl", hash = "sha256:b83e48c5b776c9969fffcfff57b03d02b1b1cd6607a9d9c4e7f568b01ef47f4c"}, + {file = "xyzservices-2024.4.0.tar.gz", hash = "sha256:6a04f11487a6fb77d92a98984cd107fbd9157fd5e65f929add9c3d6e604ee88c"}, ] [[package]] @@ -4627,23 +4670,23 @@ multidict = ">=4.0" [[package]] name = "zarr" -version = "2.16.1" +version = "2.18.1" description = "An implementation of chunked, compressed, N-dimensional arrays for Python" optional = false -python-versions = ">=3.8" +python-versions = ">=3.9" files = [ - {file = "zarr-2.16.1-py3-none-any.whl", hash = "sha256:de4882433ccb5b42cc1ec9872b95e64ca3a13581424666b28ed265ad76c7056f"}, - {file = "zarr-2.16.1.tar.gz", hash = "sha256:4276cf4b4a653431042cd53ff2282bc4d292a6842411e88529964504fb073286"}, + {file = "zarr-2.18.1-py3-none-any.whl", hash = "sha256:a1770d194eec4ec0a41a01295a6f724e1c3471d704d3aca906d3b3a7f8830245"}, + {file = "zarr-2.18.1.tar.gz", hash = "sha256:28c360ed123e606c425a694a83300227a907cb86a995fc9eef620ecafbe5f92d"}, ] [package.dependencies] asciitree = "" -fasteners = "" +fasteners = {version = "", markers = "sys_platform != \"emscripten\""} numcodecs = ">=0.10.0" -numpy = ">=1.20,<1.21.0 \|\| >1.21.0" +numpy = ">=1.23" [package.extras] -docs = ["numcodecs[msgpack]", "numpydoc", "pydata-sphinx-theme", "sphinx", "sphinx-copybutton", "sphinx-design", "sphinx-issues", "sphinx-rtd-theme"] +docs = ["numcodecs[msgpack]", "numpydoc", "pydata-sphinx-theme", "sphinx", "sphinx-automodapi", "sphinx-copybutton", "sphinx-design", "sphinx-issues"] jupyter = ["ipytree (>=0.2.2)", "ipywidgets (>=8.0.0)", "notebook"] [[package]] @@ -4659,20 +4702,20 @@ files = [ [[package]] name = "zipp" -version = "3.17.0" +version = "3.18.2" description = "Backport of pathlib-compatible object wrapper for zip files" optional = false python-versions = ">=3.8" files = [ - {file = "zipp-3.17.0-py3-none-any.whl", hash = "sha256:0e923e726174922dce09c53c59ad483ff7bbb8e572e00c7f7c46b88556409f31"}, - {file = "zipp-3.17.0.tar.gz", hash = "sha256:84e64a1c28cf7e91ed2078bb8cc8c259cb19b76942096c8d7b84947690cabaf0"}, + {file = "zipp-3.18.2-py3-none-any.whl", hash = "sha256:dce197b859eb796242b0622af1b8beb0a722d52aa2f57133ead08edd5bf5374e"}, + {file = "zipp-3.18.2.tar.gz", hash = "sha256:6278d9ddbcfb1f1089a88fde84481528b07b0e10474e09dcfe53dad4069fa059"}, ] [package.extras] -docs = ["furo", "jaraco.packaging (>=9.3)", "jaraco.tidelift (>=1.4)", "rst.linker (>=1.9)", "sphinx (<7.2.5)", "sphinx (>=3.5)", "sphinx-lint"] -testing = ["big-O", "jaraco.functools", "jaraco.itertools", "more-itertools", "pytest (>=6)", "pytest-black (>=0.3.7)", "pytest-checkdocs (>=2.4)", "pytest-cov", "pytest-enabler (>=2.2)", "pytest-ignore-flaky", "pytest-mypy (>=0.9.1)", "pytest-ruff"] +docs = ["furo", "jaraco.packaging (>=9.3)", "jaraco.tidelift (>=1.4)", "rst.linker (>=1.9)", "sphinx (>=3.5)", "sphinx-lint"] +testing = ["big-O", "jaraco.functools", "jaraco.itertools", "jaraco.test", "more-itertools", "pytest (>=6,!=8.1.)", "pytest-checkdocs (>=2.4)", "pytest-cov", "pytest-enabler (>=2.2)", "pytest-ignore-flaky", "pytest-mypy", "pytest-ruff (>=0.2.1)"] [metadata] lock-version = "2.0" python-versions = ">=3.9, <4.0" -content-hash = "606ede5919592db268c2adb911ed22e5f474a51edd4ce54f918521b703e41db5" +content-hash = "34abafe09252a6314c18832a0c53ff4a80cbeb13249ed940c43f3f4a82ea90af" diff --git a/pyproject.toml b/pyproject.toml index e10c9ee..9351c4c 100644 --- a/pyproject.toml +++ b/pyproject.toml @@ -49,6 +49,7 @@ ipykernel = "" ipython = "" papermill = "" pandas-stubs = "" +xarray-selafin = "" [tool.poetry.group.docs.dependencies] mkdocs = "*" diff --git a/requirements/requirements-dev.txt b/requirements/requirements-dev.txt index a2da168..0b536c1 100644 --- a/requirements/requirements-dev.txt +++ b/requirements/requirements-dev.txt @@ -1,16 +1,17 @@ -aiohttp==3.9.0b0 ; python_version == "3.12" +aiohttp==3.9.5 ; python_version == "3.12" aiosignal==1.3.1 ; python_version == "3.12" +ansicolors==1.1.8 ; python_version >= "3.9" and python_version < "4.0" appnope==0.1.4 ; python_version >= "3.9" and python_version < "4.0" and platform_system == "Darwin" asciitree==0.3.3 ; python_version >= "3.9" and python_version < "4.0" asttokens==2.4.1 ; python_version >= "3.9" and python_version < "4.0" attrs==23.2.0 ; python_version >= "3.9" and python_version < "4.0" -babel==2.14.0 ; python_version >= "3.9" and python_version < "4.0" +babel==2.15.0 ; python_version >= "3.9" and python_version < "4.0" beautifulsoup4==4.12.3 ; python_version < "3.10" and python_version >= "3.9" -black==24.1.1 ; python_version >= "3.9" and python_version < "4.0" +black==24.4.2 ; python_version >= "3.9" and python_version < "4.0" bleach==6.1.0 ; python_version >= "3.9" and python_version < "4.0" -bokeh==3.3.4 ; python_version >= "3.9" and python_version < "4.0" -bottleneck==1.3.7 ; python_version >= "3.9" and python_version < "4.0" -cartopy==0.22.0 ; python_version >= "3.9" and python_version < "4.0" +bokeh==3.4.1 ; python_version >= "3.9" and python_version < "4.0" +bottleneck==1.3.8 ; python_version >= "3.9" and python_version < "4.0" +cartopy==0.23.0 ; python_version >= "3.9" and python_version < "4.0" certifi==2024.2.2 ; python_version >= "3.9" and python_version < "4.0" cffi==1.16.0 ; python_version >= "3.9" and python_version < "4.0" and implementation_name == "pypy" cftime==1.6.3 ; python_version >= "3.9" and python_version < "4.0" @@ -20,155 +21,157 @@ click==8.1.7 ; python_version >= "3.9" and python_version < "4.0" cligj==0.7.2 ; python_version >= "3.9" and python_version < "4" cloudpickle==3.0.0 ; python_version >= "3.9" and python_version < "4.0" colorama==0.4.6 ; python_version >= "3.9" and python_version < "4.0" -colorcet==3.0.1 ; python_version >= "3.9" and python_version < "4.0" -comm==0.2.1 ; python_version >= "3.9" and python_version < "4.0" -contourpy==1.2.0 ; python_version >= "3.9" and python_version < "4.0" +colorcet==3.1.0 ; python_version >= "3.9" and python_version < "4.0" +comm==0.2.2 ; python_version >= "3.9" and python_version < "4.0" +contourpy==1.2.1 ; python_version >= "3.9" and python_version < "4.0" covdefaults==2.3.0 ; python_version >= "3.9" and python_version < "4.0" -coverage==7.4.1 ; python_version >= "3.9" and python_version < "4.0" -coverage[toml]==7.4.1 ; python_version >= "3.9" and python_version < "4.0" +coverage==7.5.1 ; python_version >= "3.9" and python_version < "4.0" +coverage[toml]==7.5.1 ; python_version >= "3.9" and python_version < "4.0" cycler==0.12.1 ; python_version >= "3.9" and python_version < "4.0" -dask==2024.1.1 ; python_version >= "3.9" and python_version < "4.0" -dask[array,complete,dataframe,diagnostics,distributed]==2024.1.1 ; python_version >= "3.9" and python_version < "4.0" -datashader==0.16.0 ; python_version >= "3.9" and python_version < "4.0" -debugpy==1.8.0 ; python_version >= "3.9" and python_version < "4.0" +dask-expr==1.1.1 ; python_version >= "3.9" and python_version < "4.0" +dask==2024.5.1 ; python_version >= "3.9" and python_version < "4.0" +dask[array,complete,dataframe,diagnostics,distributed]==2024.5.1 ; python_version >= "3.9" and python_version < "4.0" +datashader==0.16.1 ; python_version >= "3.9" and python_version < "4.0" +debugpy==1.8.1 ; python_version >= "3.9" and python_version < "4.0" decorator==5.1.1 ; python_version >= "3.9" and python_version < "4.0" -distributed==2024.1.1 ; python_version >= "3.9" and python_version < "4.0" +distributed==2024.5.1 ; python_version >= "3.9" and python_version < "4.0" docopt==0.6.2 ; python_version < "3.10" and python_version >= "3.9" entrypoints==0.4 ; python_version >= "3.9" and python_version < "4.0" -exceptiongroup==1.2.0 ; python_version >= "3.9" and python_version < "3.11" +exceptiongroup==1.2.1 ; python_version >= "3.9" and python_version < "3.11" executing==2.0.1 ; python_version >= "3.9" and python_version < "4.0" -fasteners==0.19 ; python_version >= "3.9" and python_version < "4.0" +fasteners==0.19 ; python_version >= "3.9" and python_version < "4.0" and sys_platform != "emscripten" fastjsonschema==2.19.1 ; python_version >= "3.9" and python_version < "4.0" -fiona==1.9.5 ; python_version >= "3.9" and python_version < "4.0" -flox==0.9.0 ; python_version >= "3.9" and python_version < "4.0" -fonttools==4.48.1 ; python_version >= "3.9" and python_version < "4.0" +fiona==1.9.6 ; python_version >= "3.9" and python_version < "4.0" +flox==0.9.7 ; python_version >= "3.9" and python_version < "4.0" +fonttools==4.51.0 ; python_version >= "3.9" and python_version < "4.0" frozenlist==1.4.1 ; python_version == "3.12" -fsspec==2024.2.0 ; python_version >= "3.9" and python_version < "4.0" -future==0.18.3 ; python_version >= "3.9" and python_version < "4.0" -geopandas==0.14.3 ; python_version >= "3.9" and python_version < "4.0" -geoviews==1.11.0 ; python_version >= "3.9" and python_version < "4.0" +fsspec==2024.5.0 ; python_version >= "3.9" and python_version < "4.0" +future==1.0.0 ; python_version >= "3.9" and python_version < "4.0" +geopandas==0.14.4 ; python_version >= "3.9" and python_version < "4.0" +geoviews==1.12.0 ; python_version >= "3.9" and python_version < "4.0" ghp-import==2.1.0 ; python_version >= "3.9" and python_version < "4.0" -griffe==0.40.0 ; python_version >= "3.9" and python_version < "4.0" +griffe==0.45.2 ; python_version >= "3.9" and python_version < "4.0" h5netcdf==1.3.0 ; python_version >= "3.9" and python_version < "4.0" -h5py==3.10.0 ; python_version >= "3.9" and python_version < "4.0" -holoviews==1.18.1 ; python_version >= "3.9" and python_version < "4.0" -idna==3.6 ; python_version >= "3.9" and python_version < "4.0" -importlib-metadata==7.0.1 ; python_version >= "3.9" and python_version < "4.0" -importlib-resources==6.1.1 ; python_version >= "3.9" and python_version < "3.10" +h5py==3.11.0 ; python_version >= "3.9" and python_version < "4.0" +holoviews==1.18.3 ; python_version >= "3.9" and python_version < "4.0" +idna==3.7 ; python_version >= "3.9" and python_version < "4.0" +importlib-metadata==7.1.0 ; python_version >= "3.9" and python_version < "3.12" +importlib-resources==6.4.0 ; python_version >= "3.9" and python_version < "3.10" iniconfig==2.0.0 ; python_version >= "3.9" and python_version < "4.0" -ipykernel==6.29.1 ; python_version >= "3.9" and python_version < "4.0" +ipykernel==6.29.4 ; python_version >= "3.9" and python_version < "4.0" ipython==8.18.1 ; python_version >= "3.9" and python_version < "4.0" jedi==0.19.1 ; python_version >= "3.9" and python_version < "4.0" -jinja2==3.1.3 ; python_version >= "3.9" and python_version < "4.0" +jinja2==3.1.4 ; python_version >= "3.9" and python_version < "4.0" jsonschema-specifications==2023.12.1 ; python_version >= "3.9" and python_version < "4.0" -jsonschema==4.21.1 ; python_version >= "3.9" and python_version < "4.0" -jupyter-client==8.6.0 ; python_version >= "3.9" and python_version < "4.0" -jupyter-core==5.7.1 ; python_version >= "3.9" and python_version < "4.0" +jsonschema==4.22.0 ; python_version >= "3.9" and python_version < "4.0" +jupyter-client==8.6.2 ; python_version >= "3.9" and python_version < "4.0" +jupyter-core==5.7.2 ; python_version >= "3.9" and python_version < "4.0" kiwisolver==1.4.5 ; python_version >= "3.9" and python_version < "4.0" linkify-it-py==2.0.3 ; python_version >= "3.9" and python_version < "4.0" llvmlite==0.42.0 ; python_version >= "3.9" and python_version < "4.0" locket==1.0.0 ; python_version >= "3.9" and python_version < "4.0" lz4==4.3.3 ; python_version >= "3.9" and python_version < "4.0" markdown-it-py==3.0.0 ; python_version >= "3.9" and python_version < "4.0" -markdown==3.5.2 ; python_version >= "3.9" and python_version < "4.0" +markdown==3.6 ; python_version >= "3.9" and python_version < "4.0" markupsafe==2.1.5 ; python_version >= "3.9" and python_version < "4.0" -matplotlib-inline==0.1.6 ; python_version >= "3.9" and python_version < "4.0" -matplotlib==3.8.2 ; python_version >= "3.9" and python_version < "4.0" -mdit-py-plugins==0.4.0 ; python_version >= "3.9" and python_version < "4.0" +matplotlib-inline==0.1.7 ; python_version >= "3.9" and python_version < "4.0" +matplotlib==3.9.0 ; python_version >= "3.9" and python_version < "4.0" +mdit-py-plugins==0.4.1 ; python_version >= "3.9" and python_version < "4.0" mdurl==0.1.2 ; python_version >= "3.9" and python_version < "4.0" mergedeep==1.3.4 ; python_version >= "3.9" and python_version < "4.0" -mkdocs-autorefs==0.5.0 ; python_version >= "3.9" and python_version < "4.0" +mkdocs-autorefs==1.0.1 ; python_version >= "3.9" and python_version < "4.0" +mkdocs-get-deps==0.2.0 ; python_version >= "3.9" and python_version < "4.0" mkdocs-material-extensions==1.3.1 ; python_version >= "3.9" and python_version < "4.0" -mkdocs-material==9.5.7 ; python_version >= "3.9" and python_version < "4.0" -mkdocs==1.5.3 ; python_version >= "3.9" and python_version < "4.0" -mkdocstrings-python==1.8.0 ; python_version >= "3.9" and python_version < "4.0" -mkdocstrings==0.24.0 ; python_version >= "3.9" and python_version < "4.0" -msgpack==1.0.7 ; python_version >= "3.9" and python_version < "4.0" +mkdocs-material==9.5.24 ; python_version >= "3.9" and python_version < "4.0" +mkdocs==1.6.0 ; python_version >= "3.9" and python_version < "4.0" +mkdocstrings-python==1.10.3 ; python_version >= "3.9" and python_version < "4.0" +mkdocstrings==0.25.1 ; python_version >= "3.9" and python_version < "4.0" +msgpack==1.0.8 ; python_version >= "3.9" and python_version < "4.0" multidict==6.0.5 ; python_version == "3.12" multipledispatch==1.0.0 ; python_version >= "3.9" and python_version < "4.0" mypy-extensions==1.0.0 ; python_version >= "3.9" and python_version < "4.0" -mypy==1.8.0 ; python_version >= "3.9" and python_version < "4.0" -nbclient==0.9.0 ; python_version >= "3.9" and python_version < "4.0" -nbformat==5.9.2 ; python_version >= "3.9" and python_version < "4.0" +mypy==1.10.0 ; python_version >= "3.9" and python_version < "4.0" +nbclient==0.10.0 ; python_version >= "3.9" and python_version < "4.0" +nbformat==5.10.4 ; python_version >= "3.9" and python_version < "4.0" nest-asyncio==1.6.0 ; python_version >= "3.9" and python_version < "4.0" netcdf4==1.6.5 ; python_version >= "3.9" and python_version < "4.0" -numba==0.59.0 ; python_version >= "3.9" and python_version < "4.0" -numbagg==0.8.0 ; python_version >= "3.9" and python_version < "4.0" +numba==0.59.1 ; python_version >= "3.9" and python_version < "4.0" +numbagg==0.8.1 ; python_version >= "3.9" and python_version < "4.0" numcodecs==0.12.1 ; python_version >= "3.9" and python_version < "4.0" -numpy-groupies==0.10.2 ; python_version >= "3.9" and python_version < "4.0" +numpy-groupies==0.11.1 ; python_version >= "3.9" and python_version < "4.0" numpy-indexed==0.3.7 ; python_version >= "3.9" and python_version < "4.0" numpy==1.26.4 ; python_version >= "3.9" and python_version < "4.0" opt-einsum==3.3.0 ; python_version >= "3.9" and python_version < "4.0" -packaging==23.2 ; python_version >= "3.9" and python_version < "4.0" +packaging==24.0 ; python_version >= "3.9" and python_version < "4.0" paginate==0.5.6 ; python_version >= "3.9" and python_version < "4.0" -pandas-stubs==2.1.4.231227 ; python_version >= "3.9" and python_version < "4.0" -pandas==2.2.0 ; python_version >= "3.9" and python_version < "4.0" -panel==1.3.8 ; python_version >= "3.9" and python_version < "4.0" -papermill==2.5.0 ; python_version >= "3.9" and python_version < "4.0" -param==2.0.2 ; python_version >= "3.9" and python_version < "4.0" -parso==0.8.3 ; python_version >= "3.9" and python_version < "4.0" -partd==1.4.1 ; python_version >= "3.9" and python_version < "4.0" +pandas-stubs==2.2.2.240514 ; python_version >= "3.9" and python_version < "4.0" +pandas==2.2.2 ; python_version >= "3.9" and python_version < "4.0" +panel==1.4.3 ; python_version >= "3.9" and python_version < "4.0" +papermill==2.6.0 ; python_version >= "3.9" and python_version < "4.0" +param==2.1.0 ; python_version >= "3.9" and python_version < "4.0" +parso==0.8.4 ; python_version >= "3.9" and python_version < "4.0" +partd==1.4.2 ; python_version >= "3.9" and python_version < "4.0" pathspec==0.12.1 ; python_version >= "3.9" and python_version < "4.0" pexpect==4.9.0 ; python_version >= "3.9" and python_version < "4.0" and sys_platform != "win32" -pillow==10.2.0 ; python_version >= "3.9" and python_version < "4.0" -platformdirs==4.2.0 ; python_version >= "3.9" and python_version < "4.0" -pluggy==1.4.0 ; python_version >= "3.9" and python_version < "4.0" -pooch==1.8.0 ; python_version >= "3.9" and python_version < "4.0" +pillow==10.3.0 ; python_version >= "3.9" and python_version < "4.0" +platformdirs==4.2.2 ; python_version >= "3.9" and python_version < "4.0" +pluggy==1.5.0 ; python_version >= "3.9" and python_version < "4.0" +pooch==1.8.1 ; python_version >= "3.9" and python_version < "4.0" prompt-toolkit==3.0.43 ; python_version >= "3.9" and python_version < "4.0" psutil==5.9.8 ; python_version >= "3.9" and python_version < "4.0" ptyprocess==0.7.0 ; python_version >= "3.9" and python_version < "4.0" and sys_platform != "win32" pure-eval==0.2.2 ; python_version >= "3.9" and python_version < "4.0" pyarrow-hotfix==0.6 ; python_version >= "3.9" and python_version < "4.0" -pyarrow==15.0.0 ; python_version >= "3.9" and python_version < "4.0" -pycparser==2.21 ; python_version >= "3.9" and python_version < "4.0" and implementation_name == "pypy" +pyarrow==16.1.0 ; python_version >= "3.9" and python_version < "4.0" +pycparser==2.22 ; python_version >= "3.9" and python_version < "4.0" and implementation_name == "pypy" pyct==0.5.0 ; python_version >= "3.9" and python_version < "4.0" pydap==3.4.1 ; python_version < "3.10" and python_version >= "3.9" -pygments==2.17.2 ; python_version >= "3.9" and python_version < "4.0" -pymdown-extensions==10.7 ; python_version >= "3.9" and python_version < "4.0" -pyparsing==3.1.1 ; python_version >= "3.9" and python_version < "4.0" +pygments==2.18.0 ; python_version >= "3.9" and python_version < "4.0" +pymdown-extensions==10.8.1 ; python_version >= "3.9" and python_version < "4.0" +pyparsing==3.1.2 ; python_version >= "3.9" and python_version < "4.0" pyproj==3.6.1 ; python_version >= "3.9" and python_version < "4.0" pyshp==2.3.1 ; python_version >= "3.9" and python_version < "4.0" -pytest-cov==4.1.0 ; python_version >= "3.9" and python_version < "4.0" -pytest==8.0.0 ; python_version >= "3.9" and python_version < "4.0" -python-dateutil==2.8.2 ; python_version >= "3.9" and python_version < "4.0" +pytest-cov==5.0.0 ; python_version >= "3.9" and python_version < "4.0" +pytest==8.2.1 ; python_version >= "3.9" and python_version < "4.0" +python-dateutil==2.9.0.post0 ; python_version >= "3.9" and python_version < "4.0" pytz==2024.1 ; python_version >= "3.9" and python_version < "4.0" -pyviz-comms==3.0.1 ; python_version >= "3.9" and python_version < "4.0" +pyviz-comms==3.0.2 ; python_version >= "3.9" and python_version < "4.0" pywin32==306 ; sys_platform == "win32" and platform_python_implementation != "PyPy" and python_version >= "3.9" and python_version < "4.0" pyyaml-env-tag==0.1 ; python_version >= "3.9" and python_version < "4.0" pyyaml==6.0.1 ; python_version >= "3.9" and python_version < "4.0" -pyzmq==25.1.2 ; python_version >= "3.9" and python_version < "4.0" -referencing==0.33.0 ; python_version >= "3.9" and python_version < "4.0" -regex==2023.12.25 ; python_version >= "3.9" and python_version < "4.0" -requests==2.31.0 ; python_version >= "3.9" and python_version < "4.0" -rpds-py==0.17.1 ; python_version >= "3.9" and python_version < "4.0" -scipy==1.12.0 ; python_version >= "3.9" and python_version < "4.0" -setuptools==69.0.3 ; python_version >= "3.9" and python_version < "4.0" -shapely==2.0.2 ; python_version >= "3.9" and python_version < "4.0" +pyzmq==26.0.3 ; python_version >= "3.9" and python_version < "4.0" +referencing==0.35.1 ; python_version >= "3.9" and python_version < "4.0" +regex==2024.5.15 ; python_version >= "3.9" and python_version < "4.0" +requests==2.32.2 ; python_version >= "3.9" and python_version < "4.0" +rpds-py==0.18.1 ; python_version >= "3.9" and python_version < "4.0" +scipy==1.13.1 ; python_version >= "3.9" and python_version < "4.0" +shapely==2.0.4 ; python_version >= "3.9" and python_version < "4.0" six==1.16.0 ; python_version >= "3.9" and python_version < "4.0" sortedcontainers==2.4.0 ; python_version >= "3.9" and python_version < "4.0" soupsieve==2.5 ; python_version < "3.10" and python_version >= "3.9" stack-data==0.6.3 ; python_version >= "3.9" and python_version < "4.0" tblib==3.0.0 ; python_version >= "3.9" and python_version < "4.0" -tenacity==8.2.3 ; python_version >= "3.9" and python_version < "4.0" +tenacity==8.3.0 ; python_version >= "3.9" and python_version < "4.0" tomli==2.0.1 ; python_version >= "3.9" and python_full_version <= "3.11.0a6" toolz==0.12.1 ; python_version >= "3.9" and python_version < "4.0" tornado==6.4 ; python_version >= "3.9" and python_version < "4.0" -tqdm==4.66.1 ; python_version >= "3.9" and python_version < "4.0" -traitlets==5.14.1 ; python_version >= "3.9" and python_version < "4.0" -types-decorator==5.1.8.20240106 ; python_version >= "3.9" and python_version < "4.0" -types-pytz==2024.1.0.20240203 ; python_version >= "3.9" and python_version < "4.0" -typing-extensions==4.9.0 ; python_version >= "3.9" and python_version < "4.0" -tzdata==2023.4 ; python_version >= "3.9" and python_version < "4.0" -uc-micro-py==1.0.2 ; python_version >= "3.9" and python_version < "4.0" -urllib3==2.2.0 ; python_version >= "3.9" and python_version < "4.0" -watchdog==3.0.0 ; python_version >= "3.9" and python_version < "4.0" +tqdm==4.66.4 ; python_version >= "3.9" and python_version < "4.0" +traitlets==5.14.3 ; python_version >= "3.9" and python_version < "4.0" +types-decorator==5.1.8.20240310 ; python_version >= "3.9" and python_version < "4.0" +types-pytz==2024.1.0.20240417 ; python_version >= "3.9" and python_version < "4.0" +typing-extensions==4.12.0 ; python_version >= "3.9" and python_version < "4.0" +tzdata==2024.1 ; python_version >= "3.9" and python_version < "4.0" +uc-micro-py==1.0.3 ; python_version >= "3.9" and python_version < "4.0" +urllib3==2.2.1 ; python_version >= "3.9" and python_version < "4.0" +watchdog==4.0.1 ; python_version >= "3.9" and python_version < "4.0" wcwidth==0.2.13 ; python_version >= "3.9" and python_version < "4.0" webencodings==0.5.1 ; python_version >= "3.9" and python_version < "4.0" webob==1.8.7 ; python_version >= "3.9" and python_version < "3.10" -xarray==2024.1.1 ; python_version >= "3.9" and python_version < "4.0" -xarray[accel,io]==2024.1.1 ; python_version >= "3.9" and python_version < "4.0" -xyzservices==2023.10.1 ; python_version >= "3.9" and python_version < "4.0" +xarray-selafin==0.1.6 ; python_version >= "3.9" and python_version < "4.0" +xarray==2024.5.0 ; python_version >= "3.9" and python_version < "4.0" +xarray[accel,io]==2024.5.0 ; python_version >= "3.9" and python_version < "4.0" +xyzservices==2024.4.0 ; python_version >= "3.9" and python_version < "4.0" yarl==1.9.4 ; python_version == "3.12" -zarr==2.16.1 ; python_version >= "3.9" and python_version < "4.0" +zarr==2.18.1 ; python_version >= "3.9" and python_version < "4.0" zict==3.0.0 ; python_version >= "3.9" and python_version < "4.0" -zipp==3.17.0 ; python_version >= "3.9" and python_version < "4.0" +zipp==3.18.2 ; python_version >= "3.9" and python_version < "3.12" diff --git a/requirements/requirements.txt b/requirements/requirements.txt index ce1db99..77eac06 100644 --- a/requirements/requirements.txt +++ b/requirements/requirements.txt @@ -2,9 +2,9 @@ asciitree==0.3.3 ; python_version >= "3.9" and python_version < "4.0" attrs==23.2.0 ; python_version >= "3.9" and python_version < "4.0" beautifulsoup4==4.12.3 ; python_version < "3.10" and python_version >= "3.9" bleach==6.1.0 ; python_version >= "3.9" and python_version < "4.0" -bokeh==3.3.4 ; python_version >= "3.9" and python_version < "4.0" -bottleneck==1.3.7 ; python_version >= "3.9" and python_version < "4.0" -cartopy==0.22.0 ; python_version >= "3.9" and python_version < "4.0" +bokeh==3.4.1 ; python_version >= "3.9" and python_version < "4.0" +bottleneck==1.3.8 ; python_version >= "3.9" and python_version < "4.0" +cartopy==0.23.0 ; python_version >= "3.9" and python_version < "4.0" certifi==2024.2.2 ; python_version >= "3.9" and python_version < "4.0" cftime==1.6.3 ; python_version >= "3.9" and python_version < "4.0" charset-normalizer==3.3.2 ; python_version >= "3.9" and python_version < "4.0" @@ -13,90 +13,90 @@ click==8.1.7 ; python_version >= "3.9" and python_version < "4.0" cligj==0.7.2 ; python_version >= "3.9" and python_version < "4" cloudpickle==3.0.0 ; python_version >= "3.9" and python_version < "4.0" colorama==0.4.6 ; python_version >= "3.9" and python_version < "4.0" and platform_system == "Windows" -colorcet==3.0.1 ; python_version >= "3.9" and python_version < "4.0" -contourpy==1.2.0 ; python_version >= "3.9" and python_version < "4.0" +colorcet==3.1.0 ; python_version >= "3.9" and python_version < "4.0" +contourpy==1.2.1 ; python_version >= "3.9" and python_version < "4.0" cycler==0.12.1 ; python_version >= "3.9" and python_version < "4.0" -dask==2024.1.1 ; python_version >= "3.9" and python_version < "4.0" -dask[array,complete,dataframe,diagnostics,distributed]==2024.1.1 ; python_version >= "3.9" and python_version < "4.0" -datashader==0.16.0 ; python_version >= "3.9" and python_version < "4.0" -distributed==2024.1.1 ; python_version >= "3.9" and python_version < "4.0" +dask-expr==1.1.1 ; python_version >= "3.9" and python_version < "4.0" +dask==2024.5.1 ; python_version >= "3.9" and python_version < "4.0" +dask[array,complete,dataframe,diagnostics,distributed]==2024.5.1 ; python_version >= "3.9" and python_version < "4.0" +datashader==0.16.1 ; python_version >= "3.9" and python_version < "4.0" +distributed==2024.5.1 ; python_version >= "3.9" and python_version < "4.0" docopt==0.6.2 ; python_version < "3.10" and python_version >= "3.9" -fasteners==0.19 ; python_version >= "3.9" and python_version < "4.0" -fiona==1.9.5 ; python_version >= "3.9" and python_version < "4.0" -flox==0.9.0 ; python_version >= "3.9" and python_version < "4.0" -fonttools==4.48.1 ; python_version >= "3.9" and python_version < "4.0" -fsspec==2024.2.0 ; python_version >= "3.9" and python_version < "4.0" -future==0.18.3 ; python_version >= "3.9" and python_version < "4.0" -geopandas==0.14.3 ; python_version >= "3.9" and python_version < "4.0" -geoviews==1.11.0 ; python_version >= "3.9" and python_version < "4.0" +fasteners==0.19 ; python_version >= "3.9" and python_version < "4.0" and sys_platform != "emscripten" +fiona==1.9.6 ; python_version >= "3.9" and python_version < "4.0" +flox==0.9.7 ; python_version >= "3.9" and python_version < "4.0" +fonttools==4.51.0 ; python_version >= "3.9" and python_version < "4.0" +fsspec==2024.5.0 ; python_version >= "3.9" and python_version < "4.0" +future==1.0.0 ; python_version >= "3.9" and python_version < "4.0" +geopandas==0.14.4 ; python_version >= "3.9" and python_version < "4.0" +geoviews==1.12.0 ; python_version >= "3.9" and python_version < "4.0" h5netcdf==1.3.0 ; python_version >= "3.9" and python_version < "4.0" -h5py==3.10.0 ; python_version >= "3.9" and python_version < "4.0" -holoviews==1.18.1 ; python_version >= "3.9" and python_version < "4.0" -idna==3.6 ; python_version >= "3.9" and python_version < "4.0" -importlib-metadata==7.0.1 ; python_version >= "3.9" and python_version < "4.0" -importlib-resources==6.1.1 ; python_version >= "3.9" and python_version < "3.10" -jinja2==3.1.3 ; python_version >= "3.9" and python_version < "4.0" +h5py==3.11.0 ; python_version >= "3.9" and python_version < "4.0" +holoviews==1.18.3 ; python_version >= "3.9" and python_version < "4.0" +idna==3.7 ; python_version >= "3.9" and python_version < "4.0" +importlib-metadata==7.1.0 ; python_version >= "3.9" and python_version < "3.12" +importlib-resources==6.4.0 ; python_version >= "3.9" and python_version < "3.10" +jinja2==3.1.4 ; python_version >= "3.9" and python_version < "4.0" kiwisolver==1.4.5 ; python_version >= "3.9" and python_version < "4.0" linkify-it-py==2.0.3 ; python_version >= "3.9" and python_version < "4.0" llvmlite==0.42.0 ; python_version >= "3.9" and python_version < "4.0" locket==1.0.0 ; python_version >= "3.9" and python_version < "4.0" lz4==4.3.3 ; python_version >= "3.9" and python_version < "4.0" markdown-it-py==3.0.0 ; python_version >= "3.9" and python_version < "4.0" -markdown==3.5.2 ; python_version >= "3.9" and python_version < "4.0" +markdown==3.6 ; python_version >= "3.9" and python_version < "4.0" markupsafe==2.1.5 ; python_version >= "3.9" and python_version < "4.0" -matplotlib==3.8.2 ; python_version >= "3.9" and python_version < "4.0" -mdit-py-plugins==0.4.0 ; python_version >= "3.9" and python_version < "4.0" +matplotlib==3.9.0 ; python_version >= "3.9" and python_version < "4.0" +mdit-py-plugins==0.4.1 ; python_version >= "3.9" and python_version < "4.0" mdurl==0.1.2 ; python_version >= "3.9" and python_version < "4.0" -msgpack==1.0.7 ; python_version >= "3.9" and python_version < "4.0" +msgpack==1.0.8 ; python_version >= "3.9" and python_version < "4.0" multipledispatch==1.0.0 ; python_version >= "3.9" and python_version < "4.0" netcdf4==1.6.5 ; python_version >= "3.9" and python_version < "4.0" -numba==0.59.0 ; python_version >= "3.9" and python_version < "4.0" -numbagg==0.8.0 ; python_version >= "3.9" and python_version < "4.0" +numba==0.59.1 ; python_version >= "3.9" and python_version < "4.0" +numbagg==0.8.1 ; python_version >= "3.9" and python_version < "4.0" numcodecs==0.12.1 ; python_version >= "3.9" and python_version < "4.0" -numpy-groupies==0.10.2 ; python_version >= "3.9" and python_version < "4.0" +numpy-groupies==0.11.1 ; python_version >= "3.9" and python_version < "4.0" numpy-indexed==0.3.7 ; python_version >= "3.9" and python_version < "4.0" numpy==1.26.4 ; python_version >= "3.9" and python_version < "4.0" opt-einsum==3.3.0 ; python_version >= "3.9" and python_version < "4.0" -packaging==23.2 ; python_version >= "3.9" and python_version < "4.0" -pandas==2.2.0 ; python_version >= "3.9" and python_version < "4.0" -panel==1.3.8 ; python_version >= "3.9" and python_version < "4.0" -param==2.0.2 ; python_version >= "3.9" and python_version < "4.0" -partd==1.4.1 ; python_version >= "3.9" and python_version < "4.0" -pillow==10.2.0 ; python_version >= "3.9" and python_version < "4.0" -platformdirs==4.2.0 ; python_version >= "3.9" and python_version < "4.0" -pooch==1.8.0 ; python_version >= "3.9" and python_version < "4.0" +packaging==24.0 ; python_version >= "3.9" and python_version < "4.0" +pandas==2.2.2 ; python_version >= "3.9" and python_version < "4.0" +panel==1.4.3 ; python_version >= "3.9" and python_version < "4.0" +param==2.1.0 ; python_version >= "3.9" and python_version < "4.0" +partd==1.4.2 ; python_version >= "3.9" and python_version < "4.0" +pillow==10.3.0 ; python_version >= "3.9" and python_version < "4.0" +platformdirs==4.2.2 ; python_version >= "3.9" and python_version < "4.0" +pooch==1.8.1 ; python_version >= "3.9" and python_version < "4.0" psutil==5.9.8 ; python_version >= "3.9" and python_version < "4.0" pyarrow-hotfix==0.6 ; python_version >= "3.9" and python_version < "4.0" -pyarrow==15.0.0 ; python_version >= "3.9" and python_version < "4.0" +pyarrow==16.1.0 ; python_version >= "3.9" and python_version < "4.0" pyct==0.5.0 ; python_version >= "3.9" and python_version < "4.0" pydap==3.4.1 ; python_version < "3.10" and python_version >= "3.9" -pyparsing==3.1.1 ; python_version >= "3.9" and python_version < "4.0" +pyparsing==3.1.2 ; python_version >= "3.9" and python_version < "4.0" pyproj==3.6.1 ; python_version >= "3.9" and python_version < "4.0" pyshp==2.3.1 ; python_version >= "3.9" and python_version < "4.0" -python-dateutil==2.8.2 ; python_version >= "3.9" and python_version < "4.0" +python-dateutil==2.9.0.post0 ; python_version >= "3.9" and python_version < "4.0" pytz==2024.1 ; python_version >= "3.9" and python_version < "4.0" -pyviz-comms==3.0.1 ; python_version >= "3.9" and python_version < "4.0" +pyviz-comms==3.0.2 ; python_version >= "3.9" and python_version < "4.0" pyyaml==6.0.1 ; python_version >= "3.9" and python_version < "4.0" -requests==2.31.0 ; python_version >= "3.9" and python_version < "4.0" -scipy==1.12.0 ; python_version >= "3.9" and python_version < "4.0" -setuptools==69.0.3 ; python_version >= "3.9" and python_version < "4.0" -shapely==2.0.2 ; python_version >= "3.9" and python_version < "4.0" +requests==2.32.2 ; python_version >= "3.9" and python_version < "4.0" +scipy==1.13.1 ; python_version >= "3.9" and python_version < "4.0" +shapely==2.0.4 ; python_version >= "3.9" and python_version < "4.0" six==1.16.0 ; python_version >= "3.9" and python_version < "4.0" sortedcontainers==2.4.0 ; python_version >= "3.9" and python_version < "4.0" soupsieve==2.5 ; python_version < "3.10" and python_version >= "3.9" tblib==3.0.0 ; python_version >= "3.9" and python_version < "4.0" toolz==0.12.1 ; python_version >= "3.9" and python_version < "4.0" tornado==6.4 ; python_version >= "3.9" and python_version < "4.0" -tqdm==4.66.1 ; python_version >= "3.9" and python_version < "4.0" -typing-extensions==4.9.0 ; python_version >= "3.9" and python_version < "4.0" -tzdata==2023.4 ; python_version >= "3.9" and python_version < "4.0" -uc-micro-py==1.0.2 ; python_version >= "3.9" and python_version < "4.0" -urllib3==2.2.0 ; python_version >= "3.9" and python_version < "4.0" +tqdm==4.66.4 ; python_version >= "3.9" and python_version < "4.0" +typing-extensions==4.12.0 ; python_version >= "3.9" and python_version < "4.0" +tzdata==2024.1 ; python_version >= "3.9" and python_version < "4.0" +uc-micro-py==1.0.3 ; python_version >= "3.9" and python_version < "4.0" +urllib3==2.2.1 ; python_version >= "3.9" and python_version < "4.0" webencodings==0.5.1 ; python_version >= "3.9" and python_version < "4.0" webob==1.8.7 ; python_version >= "3.9" and python_version < "3.10" -xarray==2024.1.1 ; python_version >= "3.9" and python_version < "4.0" -xarray[accel,io]==2024.1.1 ; python_version >= "3.9" and python_version < "4.0" -xyzservices==2023.10.1 ; python_version >= "3.9" and python_version < "4.0" -zarr==2.16.1 ; python_version >= "3.9" and python_version < "4.0" +xarray==2024.5.0 ; python_version >= "3.9" and python_version < "4.0" +xarray[accel,io]==2024.5.0 ; python_version >= "3.9" and python_version < "4.0" +xyzservices==2024.4.0 ; python_version >= "3.9" and python_version < "4.0" +zarr==2.18.1 ; python_version >= "3.9" and python_version < "4.0" zict==3.0.0 ; python_version >= "3.9" and python_version < "4.0" -zipp==3.17.0 ; python_version >= "3.9" and python_version < "4.0" +zipp==3.18.2 ; python_version >= "3.9" and python_version < "3.12" diff --git a/thalassa/api.py b/thalassa/api.py index f25d6bf..564a03b 100644 --- a/thalassa/api.py +++ b/thalassa/api.py @@ -67,14 +67,14 @@ def open_dataset( Parameters: path: The path to the dataset file (netCDF, zarr, grib) normalize: Boolean flag indicating whether the dataset should be converted/normalized to the "Thalassa schema". - Normalization is currently only supported for ``SCHISM`` and ``ADCIRC`` netcdf files. + Normalization is currently only supported for ``SCHISM``, ``TELEMAC``, and ``ADCIRC`` netcdf files. kwargs: The ``kwargs`` are being passed through to ``xarray.open_dataset``. """ import xarray as xr default_kwargs: dict[str, T.Any] = dict( - mask_and_scale=True, + # mask_and_scale=True, cache=False, drop_variables=ADCIRC_VARIABLES_TO_BE_DROPPED, ) diff --git a/thalassa/normalization.py b/thalassa/normalization.py index cd4855f..df2d306 100644 --- a/thalassa/normalization.py +++ b/thalassa/normalization.py @@ -19,19 +19,29 @@ class THALASSA_FORMATS(enum.Enum): UNKNOWN = "UNKNOWN" ADCIRC = "ADCIRC" SCHISM = "SCHISM" + TELEMAC = "TELEMAC" GENERIC = "GENERIC" PYPOSEIDON = "PYPOSEIDON" # fmt: off +EDGE_DIM = "edge" +FACE_DIM = "face" +NODE_DIM = "node" +VERTICAL_DIM = "layer" +CONNECTIVITY = "face_nodes" +VERTICE_DIM = "max_no_vertices" +X_DIM = "lon" +Y_DIM = "lat" + _GENERIC_DIMS = { - "node", + NODE_DIM, "triface", "three", } _GENERIC_VARS = { - "lon", - "lat", + X_DIM, + Y_DIM, "triface_nodes", } _SCHISM_DIMS = { @@ -45,6 +55,14 @@ class THALASSA_FORMATS(enum.Enum): "SCHISM_hgrid_node_y", "SCHISM_hgrid_face_nodes", } +_TELEMAC_DIMS = { + "node", +} +_TELEMAC_VARS = { + "x", + "y", + "ikle2", +} _PYPOSEIDON_DIMS = { "nSCHISM_hgrid_face", "nSCHISM_hgrid_node", @@ -76,6 +94,11 @@ def is_schism(ds: xarray.Dataset) -> bool: return _SCHISM_DIMS.issubset(ds.dims) and _SCHISM_VARS.issubset(total_vars) +def is_telemac(ds: xarray.Dataset) -> bool: + total_vars = list(ds.data_vars.keys()) + list(ds.coords.keys()) + list(ds.attrs.keys()) + return _TELEMAC_DIMS.issubset(ds.dims) and _TELEMAC_VARS.issubset(total_vars) + + def is_pyposeidon(ds: xarray.Dataset) -> bool: return _PYPOSEIDON_DIMS.issubset(ds.dims) and _PYPOSEIDON_VARS.issubset(ds.data_vars) @@ -87,6 +110,8 @@ def is_adcirc(ds: xarray.Dataset) -> bool: def infer_format(ds: xarray.Dataset) -> THALASSA_FORMATS: if is_schism(ds): fmt = THALASSA_FORMATS.SCHISM + elif is_telemac(ds): + fmt = THALASSA_FORMATS.TELEMAC elif is_pyposeidon(ds): fmt = THALASSA_FORMATS.PYPOSEIDON elif is_generic(ds): @@ -120,37 +145,58 @@ def normalize_generic(ds: xarray.Dataset) -> xarray.Dataset: def normalize_schism(ds: xarray.Dataset) -> xarray.Dataset: ds = ds.rename( { - "nSCHISM_hgrid_edge": "edge", - "nSCHISM_hgrid_face": "face", - "nSCHISM_hgrid_node": "node", - "SCHISM_hgrid_face_nodes": "face_nodes", - "nMaxSCHISM_hgrid_face_nodes": "max_no_vertices", - "SCHISM_hgrid_node_x": "lon", - "SCHISM_hgrid_node_y": "lat", + "nSCHISM_hgrid_edge": EDGE_DIM, + "nSCHISM_hgrid_face": FACE_DIM, + "nSCHISM_hgrid_node": NODE_DIM, + "SCHISM_hgrid_face_nodes": CONNECTIVITY, + "nMaxSCHISM_hgrid_face_nodes": VERTICE_DIM, + "SCHISM_hgrid_node_x": X_DIM, + "SCHISM_hgrid_node_y": Y_DIM, }, ) if "nSCHISM_vgrid_layers" in ds.dims: # I.e. OLD Schism IO or "merged" new IO ds = ds.rename( { - "nSCHISM_vgrid_layers": "layer", + "nSCHISM_vgrid_layers": VERTICAL_DIM, }, ) # SCHISM output uses one-based indices for `face_nodes` # Let's ensure that we use zero-based indices everywhere. - ds["face_nodes"] -= 1 + ds[CONNECTIVITY] -= 1 + return ds + + +def normalize_telemac(ds: xarray.Dataset) -> xarray.Dataset: + ds = ds.rename( + { + "node": NODE_DIM, + "x": X_DIM, + "y": Y_DIM, + }, + ) + if "plan" in ds.dims: + ds = ds.rename( + { + "plan": VERTICAL_DIM, + }, + ) + + # TELEMAC output uses one-based indices for `face_nodes` + # Let's ensure that we use zero-based indices everywhere. + ds[CONNECTIVITY] = ((FACE_DIM, VERTICE_DIM), ds.attrs['ikle2'] - 1) return ds def normalize_pyposeidon(ds: xarray.Dataset) -> xarray.Dataset: ds = ds.rename( { - "nSCHISM_hgrid_face": "face", - "nSCHISM_hgrid_node": "node", - "SCHISM_hgrid_face_nodes": "face_nodes", - "nMaxSCHISM_hgrid_face_nodes": "max_no_vertices", - "SCHISM_hgrid_node_x": "lon", - "SCHISM_hgrid_node_y": "lat", + "nSCHISM_hgrid_face": FACE_DIM, + "nSCHISM_hgrid_node": NODE_DIM, + "SCHISM_hgrid_face_nodes": CONNECTIVITY, + "nMaxSCHISM_hgrid_face_nodes": VERTICE_DIM, + "SCHISM_hgrid_node_x": X_DIM, + "SCHISM_hgrid_node_y": Y_DIM, }, ) return ds @@ -159,16 +205,16 @@ def normalize_pyposeidon(ds: xarray.Dataset) -> xarray.Dataset: def normalize_adcirc(ds: xarray.Dataset) -> xarray.Dataset: ds = ds.rename( { - "x": "lon", - "y": "lat", - "element": "face_nodes", - "nvertex": "max_no_vertices", - "nele": "face", + "x": X_DIM, + "y": Y_DIM, + "element": CONNECTIVITY, + "nvertex": VERTICE_DIM, + "nele": FACE_DIM, }, ) # ADCIRC output uses one-based indices for `face_nodes` # Let's ensure that we use zero-based indices everywhere. - ds["face_nodes"] -= 1 + ds[CONNECTIVITY] -= 1 return ds @@ -176,6 +222,7 @@ def normalize_adcirc(ds: xarray.Dataset) -> xarray.Dataset: THALASSA_FORMATS.ADCIRC: normalize_adcirc, THALASSA_FORMATS.GENERIC: normalize_generic, THALASSA_FORMATS.SCHISM: normalize_schism, + THALASSA_FORMATS.TELEMAC: normalize_telemac, THALASSA_FORMATS.PYPOSEIDON: normalize_pyposeidon, }	normalisation for GENERIC formats -- naming conventions This issue is more to open a discussion on naming conventions. My Dataset has the following format: ``` Dimensions: time: 7272 node: 301572 face: 588314 max_no_vertices: 3 Coordinates: longitude (node) float32 ... latitude (node) float32 ... time (time) datetime64[ns] 2023-01-01 ... 2023-10-30T23:00:00 Data variables: u (time, node) float32 ... v (time, node) float32 ... elev (time, node) float32 ... face_nodes (face, max_no_vertices) int32 ... ``` When I try to plot: ```python plot_map = thalassa.plot( ds=ds.isel(time=-1), variable="max_elev", clim_min=0, clim_max=1, clabel="meters", cmap = 'jet' ) ``` I get the following error: ``` KeyError Traceback (most recent call last) Cell In[22], [line 1](vscode-notebook-cell:?execution_count=22&line=1) ----> [1](vscode-notebook-cell:?execution_count=22&line=1) plot_map = thalassa.plot( [2](vscode-notebook-cell:?execution_count=22&line=2) ds=ds.isel(time=-1), [3](vscode-notebook-cell:?execution_count=22&line=3) variable="max_elev", [4](vscode-notebook-cell:?execution_count=22&line=4) clim_min=0, [5](vscode-notebook-cell:?execution_count=22&line=5) clim_max=1, [6](vscode-notebook-cell:?execution_count=22&line=6) clabel="meters", [7](vscode-notebook-cell:?execution_count=22&line=7) cmap = 'jet' [8](vscode-notebook-cell:?execution_count=22&line=8) ) File ~/work/gist/Validation_Protocol/.venv/lib/python3.11/site-packages/thalassa/plotting.py:153, in plot(ds, variable, bbox, title, cmap, colorbar, clabel, clim_min, clim_max, x_range, y_range, show_mesh) [93].venv/lib/python3.11/site-packages/thalassa/plotting.py:93) """ [94].venv/lib/python3.11/site-packages/thalassa/plotting.py:94) Return the plot of the specified `variable`. [95].venv/lib/python3.11/site-packages/thalassa/plotting.py:95) (...) [149].venv/lib/python3.11/site-packages/thalassa/plotting.py:149) [150].venv/lib/python3.11/site-packages/thalassa/plotting.py:150) """ [151].venv/lib/python3.11/site-packages/thalassa/plotting.py:151) import holoviews as hv --> [153].venv/lib/python3.11/site-packages/thalassa/plotting.py:153) ds = normalization.normalize(ds) [154].venv/lib/python3.11/site-packages/thalassa/plotting.py:154) _sanity_check(ds=ds, variable=variable) [155].venv/lib/python3.11/site-packages/thalassa/plotting.py:155) if bbox: File ~/work/gist/Validation_Protocol/.venv/lib/python3.11/site-packages/thalassa/normalization.py:203, in normalize(ds) [201].venv/lib/python3.11/site-packages/thalassa/normalization.py:201) logger.debug("Dataset normalization: Started") [202].venv/lib/python3.11/site-packages/thalassa/normalization.py:202) fmt = infer_format(ds) --> [203].venv/lib/python3.11/site-packages/thalassa/normalization.py:203) normalizer_func = NORMALIZE_DISPATCHER[fmt] [204].venv/lib/python3.11/site-packages/thalassa/normalization.py:204) normalized_ds = normalizer_func(ds) [205].venv/lib/python3.11/site-packages/thalassa/normalization.py:205) # Handle quad elements [206].venv/lib/python3.11/site-packages/thalassa/normalization.py:206) # Splitting quad elements to triangles, means that the number of faces increases [207].venv/lib/python3.11/site-packages/thalassa/normalization.py:207) # There are two options: (...) [210].venv/lib/python3.11/site-packages/thalassa/normalization.py:210) # I'd rather avoid altering the values of the provided netcdf file therefore we go for option #2, [211].venv/lib/python3.11/site-packages/thalassa/normalization.py:211) # i.e. we create the `triface_nodes` variable. KeyError: <THALASSA_FORMATS.UNKNOWN: 'UNKNOWN'> ``` That is caused by: ```python ds = thalassa.normalization.normalize(ds) ``` I found the solution by making my dataset compliant with the `GENERIC` [format](https://github.com/ec-jrc/Thalassa/blob/b9d977cd6999e73f5ad884e9e7b96d4041b60827/thalassa/normalization.py#L27): ```python ds = ds.rename({ "longitude": "lon", "latitude": "lat", "face_nodes": "triface_nodes", "max_no_vertices": "three", "face": "triface", }) # make variable compliant with thalassa ``` I also understood the convention `face_nodes` / `triface_nodes` is before / after the Quad elements [test](https://github.com/ec-jrc/Thalassa/blob/b9d977cd6999e73f5ad884e9e7b96d4041b60827/thalassa/normalization.py#L212). ## concerning the connectivity of the mesh: Would it be possible to skip the quad element test if you know already that the mesh is an unstructured triangle grid? Also, I think ideally the connectivity of mesh should be renamed to `mesh_topology`, according to [UGRID](https://ugrid-conventions.github.io/ugrid-conventions/#2d-triangular-mesh-topology) and [cf_xarray](https://cf-xarray.readthedocs.io/en/latest/sgrid_ugrid.html#ugrid) conventions ## concerning the coordinates Would it acceptable to consider other names for coordinates ? like `Lon`, `x` or `longitude`, or should this aspect be dealt by the user?	Is this output from Telemac? If yes then the easiest way to add support would be to add an extra "schema" in `normalization.py`. I think that this entails defining: - `is_telemac()` - `normalize_telemac` and then updating: - `infer_format()` - `NORMALIZE_DISPATCHER` If you can create a small file, then add a test too: https://github.com/ec-jrc/Thalassa/blob/b9d977cd6999e73f5ad884e9e7b96d4041b60827/tests/normalization_test.py The reason we have `triface_nodes` etc is because: 1. when it comes to unstructured meshes, Holoviews only supports triangles 2. we want to be able to support STOFS-3D output too, which uses quads. So the idea is that we keep the original mesh connectivity as is, and we create a new variable that holds the triangle-only connectivity. If there are no quads, then the new variable is just a copy of the original. The thalassa code internally uses the new variable everywhere. WRT CF naming-conventions etc, we can discuss it at https://github.com/ec-jrc/Thalassa/issues/80 Just to make it clear, we could add support for UGRID compliant netcdfs too. I am for sure not against that, but it is probably going to be more work compared to just updating the current setup. But feel free to investigate it if you want. It is for sure going to be useful in the context of seamesh etc. > Is this output from Telemac? From TELEMAC yes, but through [pyPoseidon](https://github.com/ec-jrc/pyPoseidon/blob/92a797eb396bc5263fd95d94f93c088450056a65/pyposeidon/telemac.py#L1153) where I convert from Selafin file to xarray Dataset. All details from one format to other is in the [`norm.py`](https://github.com/ec-jrc/pyPoseidon/blob/telemac/pyposeidon/utils/norm.py#L110) function of the `telemac` branch > the easiest way to add support would be to add an extra "schema" in normalization.py. So there are 2 ways to solve this: 1. use `xarray-selafin` in `thalassa` to read directly from Selafins. But it might be too much abstraction for the user and also if `xarray-selafin` changes the backend, we'd have to adapt the TELEMAC bindings here too. 2. adapt the conversion convention in pyPoseidon A third option would be a combination of both. The quickest win is clearly option 2 though. The idea is to try to keep thalassa decoupled from pyposeidon. If you add the functionality in pyposeidon, then TELEMAC output will be visualizable by thalassa iff you post-process it with pyposeidon. If you add the functionality in thalassa, then TELEMAC output will be visualizable by thalassa regardless of the way you run/post-processed the results. That being said, if you truly want to make the TELEMAC output compatible with pyPoseidon, you will probably need to rename the TELEMAC variables/dims anyhow in order to make them compatible with the existing pyposeidon functionality (which uses SCHISM names, pretty much ubiquitusly).	2024-05-19T11:59:10	[]	[]
GeoStat-Framework/GSTools	GeoStat-Framework__GSTools-340	90003fbbe193cfa835a891b403b4024bb21867f9	diff --git a/.github/workflows/main.yml b/.github/workflows/main.yml index e116c483..f8dc74cd 100644 --- a/.github/workflows/main.yml +++ b/.github/workflows/main.yml @@ -16,16 +16,14 @@ jobs: source_check: name: source check runs-on: ubuntu-latest - strategy: - fail-fast: false steps: - - uses: actions/checkout@v2 + - uses: actions/checkout@v4 with: fetch-depth: '0' - name: Set up Python 3.9 - uses: actions/setup-python@v2 + uses: actions/setup-python@v5 with: python-version: 3.9 @@ -38,6 +36,10 @@ jobs: run: \| python -m black --check --diff --color . + - name: black preview + run: \| + python -m black --preview --diff --color . + - name: isort check run: \| python -m isort --check --diff --color . @@ -58,60 +60,68 @@ jobs: matrix: cfg: - { os: ubuntu-latest, arch: x86_64 } - - { os: ubuntu-latest, arch: i686 } - { os: windows-latest, arch: AMD64 } - - { os: windows-latest, arch: x86 } - { os: macos-latest, arch: x86_64 } - { os: macos-latest, arch: arm64 } - { os: macos-latest, arch: universal2 } steps: - - uses: actions/checkout@v2 + - uses: actions/checkout@v4 with: fetch-depth: '0' - name: Build wheels - uses: pypa/[email protected] + uses: pypa/[email protected] env: CIBW_ARCHS: ${{ matrix.cfg.arch }} with: output-dir: dist - - uses: actions/upload-artifact@v2 + - uses: actions/upload-artifact@v3 with: path: ./dist/.whl build_sdist: - name: sdist on ${{ matrix.os }} with py ${{ matrix.python-version }} + name: sdist on ${{ matrix.os }} with py ${{ matrix.ver.py }} numpy${{ matrix.ver.np }} scipy${{ matrix.ver.sp }} runs-on: ${{ matrix.os }} strategy: fail-fast: false matrix: os: [ubuntu-latest, windows-latest, macos-latest] - python-version: ['3.8', '3.9', '3.10', '3.11', '3.12'] - + # https://github.com/scipy/oldest-supported-numpy/blob/main/setup.cfg + ver: + - {py: '3.8', np: '==1.20.0', sp: '==1.5.4'} + - {py: '3.9', np: '==1.20.0', sp: '==1.5.4'} + - {py: '3.10', np: '==1.21.6', sp: '==1.7.2'} + - {py: '3.11', np: '==1.23.2', sp: '==1.9.2'} + - {py: '3.12', np: '==1.26.2', sp: '==1.11.2'} + - {py: '3.12', np: '>=2.0.0rc1', sp: '>=1.13.0'} steps: - - uses: actions/checkout@v2 + - uses: actions/checkout@v4 with: fetch-depth: '0' - - name: Set up Python ${{ matrix.python-version }} - uses: actions/setup-python@v2 + - name: Set up Python ${{ matrix.ver.py }} + uses: actions/setup-python@v5 with: - python-version: ${{ matrix.python-version }} + python-version: ${{ matrix.ver.py }} - name: Install dependencies + run: \| + python -m pip install --upgrade pip + pip install build "coveralls>=3.0.0" + + - name: Install GSTools env: GSTOOLS_BUILD_PARALLEL: 1 run: \| - python -m pip install --upgrade pip - pip install build coveralls>=3.0.0 pip install -v --editable .[test] - name: Run tests env: GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }} run: \| + pip install "numpy${{ matrix.ver.np }}" "scipy${{ matrix.ver.sp }}" python -m pytest --cov gstools --cov-report term-missing -v tests/ python -m coveralls --service=github @@ -120,8 +130,8 @@ jobs: # PEP 517 package builder from pypa python -m build --sdist --outdir dist . - - uses: actions/upload-artifact@v2 - if: matrix.os == 'ubuntu-latest' && matrix.python-version == '3.9' + - uses: actions/upload-artifact@v3 + if: matrix.os == 'ubuntu-latest' && matrix.ver.py == '3.9' with: path: dist/.tar.gz @@ -130,7 +140,7 @@ jobs: runs-on: ubuntu-latest steps: - - uses: actions/download-artifact@v2 + - uses: actions/download-artifact@v3 with: name: artifact path: dist diff --git a/.gitignore b/.gitignore index 872d3f40..bcdc980b 100644 --- a/.gitignore +++ b/.gitignore @@ -117,6 +117,7 @@ src/gstools/_version.py # generated docs docs/source/examples/ docs/source/api/ +docs/source/sg_execution_times.rst # other settings .vscode/ diff --git a/CHANGELOG.md b/CHANGELOG.md index 00513d23..61601c72 100755 --- a/CHANGELOG.md +++ b/CHANGELOG.md @@ -2,6 +2,25 @@ All notable changes to GSTools will be documented in this file. +## [Unreleased] - ? + +### Enhancements + +- added global variable `config.NUM_THREADS` to select number of threads for parallel computation ([#336](https://github.com/GeoStat-Framework/GSTools/pull/336)) +- prepare numpy 2 support ([#340](https://github.com/GeoStat-Framework/GSTools/pull/340)) + - numpy 2.0.0rc1 for building extensions (for Python 3.9 and above) + - check multiple numpy and scipy versions in CI + - fixed minimal versions for numpy + - use `np.asarray` everywhere with `np.atleast_(n)d` + - fix long/longlong integer issue in cython on windows by always using 64bit integers + +### Bugfixes +- build docs with latest sphinx version ([#340](https://github.com/GeoStat-Framework/GSTools/pull/340)) + +### Changes +- require pyvista 0.40 at least ([#340](https://github.com/GeoStat-Framework/GSTools/pull/340)) + + ## [1.5.1] - Nifty Neon - 2023-11 ### Enhancements diff --git a/README.md b/README.md index 69f1f3ec..6cb69901 100644 --- a/README.md +++ b/README.md @@ -345,7 +345,7 @@ in memory for immediate 3D plotting in Python. ## Requirements: -- [NumPy >= 1.14.5](https://www.numpy.org) +- [NumPy >= 1.20.0](https://www.numpy.org) - [SciPy >= 1.1.0](https://www.scipy.org/scipylib) - [hankel >= 1.0.0](https://github.com/steven-murray/hankel) - [emcee >= 3.0.0](https://github.com/dfm/emcee) @@ -366,7 +366,7 @@ You can contact us via <[email protected]>. ## License -[LGPLv3][license_link] © 2018-2021 +[LGPLv3][license_link] © 2018-2024 [pip_link]: https://pypi.org/project/gstools [conda_link]: https://docs.conda.io/en/latest/miniconda.html diff --git a/docs/source/conf.py b/docs/source/conf.py index eb5e5fae..e89928fc 100644 --- a/docs/source/conf.py +++ b/docs/source/conf.py @@ -259,10 +259,9 @@ def setup(app): # Example configuration for intersphinx: refer to the Python standard library. intersphinx_mapping = { "Python": ("https://docs.python.org/", None), - "NumPy": ("http://docs.scipy.org/doc/numpy/", None), - "SciPy": ("http://docs.scipy.org/doc/scipy/reference", None), - "matplotlib": ("http://matplotlib.org", None), - "Sphinx": ("http://www.sphinx-doc.org/en/stable/", None), + "NumPy": ("https://numpy.org/doc/stable/", None), + "SciPy": ("https://docs.scipy.org/doc/scipy/", None), + "matplotlib": ("https://matplotlib.org/stable/", None), "hankel": ("https://hankel.readthedocs.io/en/latest/", None), "emcee": ("https://emcee.readthedocs.io/en/latest/", None), } diff --git a/docs/source/index.rst b/docs/source/index.rst index d71494ac..ecad0583 100644 --- a/docs/source/index.rst +++ b/docs/source/index.rst @@ -414,7 +414,7 @@ in memory for immediate 3D plotting in Python. Requirements ============ -- `Numpy >= 1.14.5 <http://www.numpy.org>`_ +- `NumPy >= 1.20.0 <http://www.numpy.org>`_ - `SciPy >= 1.1.0 <http://www.scipy.org>`_ - `hankel >= 1.0.0 <https://github.com/steven-murray/hankel>`_ - `emcee >= 3.0.0 <https://github.com/dfm/emcee>`_ diff --git a/examples/00_misc/00_tpl_stable.py b/examples/00_misc/00_tpl_stable.py index 4d1a89d8..474b0f55 100644 --- a/examples/00_misc/00_tpl_stable.py +++ b/examples/00_misc/00_tpl_stable.py @@ -40,6 +40,7 @@ \sigma^2_{\ell_{\mathrm{up}}} &= C\cdot\frac{\ell_{\mathrm{up}}^{2H}}{2H} """ + import numpy as np import gstools as gs diff --git a/examples/00_misc/01_export.py b/examples/00_misc/01_export.py index ad754abf..e38294fe 100644 --- a/examples/00_misc/01_export.py +++ b/examples/00_misc/01_export.py @@ -7,6 +7,7 @@ These can be viewed for example with `Paraview <https://www.paraview.org/>`__. """ + # sphinx_gallery_thumbnail_path = 'pics/paraview.png' import gstools as gs diff --git a/examples/00_misc/02_check_rand_meth_sampling.py b/examples/00_misc/02_check_rand_meth_sampling.py index 0b23a929..58d998b4 100644 --- a/examples/00_misc/02_check_rand_meth_sampling.py +++ b/examples/00_misc/02_check_rand_meth_sampling.py @@ -2,6 +2,7 @@ Check Random Sampling --------------------- """ + import numpy as np from matplotlib import pyplot as plt from mpl_toolkits.mplot3d import Axes3D diff --git a/examples/00_misc/04_herten.py b/examples/00_misc/04_herten.py index d9e1b16e..1e1b8a23 100644 --- a/examples/00_misc/04_herten.py +++ b/examples/00_misc/04_herten.py @@ -30,6 +30,7 @@ functions, since the only produce the ``herten_transmissivity.gz`` and ``grid_dim_origin_spacing.txt``, which are already present. """ + import os import matplotlib.pyplot as plt diff --git a/examples/00_misc/05_standalone_field.py b/examples/00_misc/05_standalone_field.py index fb0c2082..e467f043 100644 --- a/examples/00_misc/05_standalone_field.py +++ b/examples/00_misc/05_standalone_field.py @@ -7,6 +7,7 @@ In the following example we will produce 10000 random points in 4D with random values and plot them. """ + import numpy as np import gstools as gs diff --git a/examples/01_random_field/03_unstr_srf_export.py b/examples/01_random_field/03_unstr_srf_export.py index e86e6f3e..94d00952 100644 --- a/examples/01_random_field/03_unstr_srf_export.py +++ b/examples/01_random_field/03_unstr_srf_export.py @@ -6,6 +6,7 @@ Normally, such a grid would be read in, but we can simply generate one and then create a random field at those coordinates. """ + import numpy as np import gstools as gs diff --git a/examples/01_random_field/04_srf_merge.py b/examples/01_random_field/04_srf_merge.py index 00aa6511..241ed079 100644 --- a/examples/01_random_field/04_srf_merge.py +++ b/examples/01_random_field/04_srf_merge.py @@ -6,6 +6,7 @@ to merge two unstructured rectangular fields. """ + # sphinx_gallery_thumbnail_number = 2 import numpy as np diff --git a/examples/01_random_field/06_pyvista_support.py b/examples/01_random_field/06_pyvista_support.py index 4cc371dd..29de8dd7 100644 --- a/examples/01_random_field/06_pyvista_support.py +++ b/examples/01_random_field/06_pyvista_support.py @@ -12,6 +12,7 @@ The :any:`Field.mesh` method enables easy field creation on PyVista meshes used by the :any:`SRF` or :any:`Krige` class. """ + # sphinx_gallery_thumbnail_path = 'pics/GS_pyvista_cut.png' import pyvista as pv @@ -21,8 +22,8 @@ # We create a structured grid with PyVista containing 50 segments on all three # axes each with a length of 2 (whatever unit). -dim, spacing = (50, 50, 50), (2, 2, 2) -grid = pv.UniformGrid(dim, spacing) +dims, spacing = (50, 50, 50), (2, 2, 2) +grid = pv.ImageData(dimensions=dims, spacing=spacing) ############################################################################### # Now we set up the SRF class as always. We'll use an anisotropic model. diff --git a/examples/02_cov_model/01_basic_methods.py b/examples/02_cov_model/01_basic_methods.py index fb9bfe7a..4c97fba8 100755 --- a/examples/02_cov_model/01_basic_methods.py +++ b/examples/02_cov_model/01_basic_methods.py @@ -36,6 +36,7 @@ correlation function as demonstrated in the introductory example. If one of the above functions is given, the others will be determined: """ + import gstools as gs model = gs.Exponential(dim=3, var=2.0, len_scale=10, nugget=0.5) diff --git a/examples/02_cov_model/02_aniso_rotation.py b/examples/02_cov_model/02_aniso_rotation.py index 2a8bac78..b7459e39 100755 --- a/examples/02_cov_model/02_aniso_rotation.py +++ b/examples/02_cov_model/02_aniso_rotation.py @@ -6,6 +6,7 @@ represents the isotropic case for the model. Nevertheless, you can provide anisotropy ratios by: """ + import gstools as gs model = gs.Gaussian(dim=3, var=2.0, len_scale=10, anis=0.5) diff --git a/examples/02_cov_model/03_spectral_methods.py b/examples/02_cov_model/03_spectral_methods.py index 677811a9..61c7e49b 100755 --- a/examples/02_cov_model/03_spectral_methods.py +++ b/examples/02_cov_model/03_spectral_methods.py @@ -23,6 +23,7 @@ You can access these methods by: """ + import gstools as gs model = gs.Gaussian(dim=3, var=2.0, len_scale=10) diff --git a/examples/02_cov_model/04_different_scales.py b/examples/02_cov_model/04_different_scales.py index cd6f4dee..0e2e1991 100755 --- a/examples/02_cov_model/04_different_scales.py +++ b/examples/02_cov_model/04_different_scales.py @@ -16,6 +16,7 @@ You can access it by: """ + import gstools as gs model = gs.Stable(dim=3, var=2.0, len_scale=10) diff --git a/examples/02_cov_model/05_additional_para.py b/examples/02_cov_model/05_additional_para.py index 02507d96..3264cec4 100755 --- a/examples/02_cov_model/05_additional_para.py +++ b/examples/02_cov_model/05_additional_para.py @@ -10,6 +10,7 @@ This leads to the so called stable covariance model and we can define it by """ + import numpy as np import gstools as gs diff --git a/examples/02_cov_model/06_fitting_para_ranges.py b/examples/02_cov_model/06_fitting_para_ranges.py index bd3d1196..7c8f083e 100755 --- a/examples/02_cov_model/06_fitting_para_ranges.py +++ b/examples/02_cov_model/06_fitting_para_ranges.py @@ -6,6 +6,7 @@ variogram data. In the following we will use the self defined stable model from a previous example. """ + import numpy as np import gstools as gs diff --git a/examples/03_variogram/00_fit_variogram.py b/examples/03_variogram/00_fit_variogram.py index 9dc1b305..7334ed2c 100644 --- a/examples/03_variogram/00_fit_variogram.py +++ b/examples/03_variogram/00_fit_variogram.py @@ -2,6 +2,7 @@ Fit Variogram ------------- """ + import numpy as np import gstools as gs diff --git a/examples/03_variogram/01_find_best_model.py b/examples/03_variogram/01_find_best_model.py index c0fc383f..eab031cf 100755 --- a/examples/03_variogram/01_find_best_model.py +++ b/examples/03_variogram/01_find_best_model.py @@ -2,6 +2,7 @@ Finding the best fitting variogram model ---------------------------------------- """ + import numpy as np from matplotlib import pyplot as plt diff --git a/examples/03_variogram/02_multi_vario.py b/examples/03_variogram/02_multi_vario.py index f893e452..71048849 100755 --- a/examples/03_variogram/02_multi_vario.py +++ b/examples/03_variogram/02_multi_vario.py @@ -5,6 +5,7 @@ In this example, we demonstrate how to estimate a variogram from multiple fields on the same point-set that should have the same statistical properties. """ + import matplotlib.pyplot as plt import numpy as np diff --git a/examples/03_variogram/03_directional_2d.py b/examples/03_variogram/03_directional_2d.py index e5744cb2..460c7513 100755 --- a/examples/03_variogram/03_directional_2d.py +++ b/examples/03_variogram/03_directional_2d.py @@ -7,6 +7,7 @@ Afterwards we will fit a model to this estimated variogram and show the result. """ + import numpy as np from matplotlib import pyplot as plt diff --git a/examples/03_variogram/04_directional_3d.py b/examples/03_variogram/04_directional_3d.py index 22ee7e9f..6a8b6ddf 100755 --- a/examples/03_variogram/04_directional_3d.py +++ b/examples/03_variogram/04_directional_3d.py @@ -7,6 +7,7 @@ Afterwards we will fit a model to this estimated variogram and show the result. """ + import matplotlib.pyplot as plt import numpy as np from mpl_toolkits.mplot3d import Axes3D diff --git a/examples/03_variogram/05_auto_fit_variogram.py b/examples/03_variogram/05_auto_fit_variogram.py index 9c1cf2f9..2fcc7fbd 100644 --- a/examples/03_variogram/05_auto_fit_variogram.py +++ b/examples/03_variogram/05_auto_fit_variogram.py @@ -2,6 +2,7 @@ Fit Variogram with automatic binning ------------------------------------ """ + import numpy as np import gstools as gs diff --git a/examples/03_variogram/06_auto_bin_latlon.py b/examples/03_variogram/06_auto_bin_latlon.py index 20f6f9bc..cc248ea1 100644 --- a/examples/03_variogram/06_auto_bin_latlon.py +++ b/examples/03_variogram/06_auto_bin_latlon.py @@ -8,6 +8,7 @@ We use a data set from 20 meteo-stations choosen randomly. """ + import numpy as np import gstools as gs diff --git a/examples/04_vector_field/00_2d_vector_field.py b/examples/04_vector_field/00_2d_vector_field.py index 6a6c71ac..2e722764 100644 --- a/examples/04_vector_field/00_2d_vector_field.py +++ b/examples/04_vector_field/00_2d_vector_field.py @@ -5,6 +5,7 @@ As a first example we are going to generate a 2d vector field with a Gaussian covariance model on a structured grid: """ + import numpy as np import gstools as gs diff --git a/examples/04_vector_field/01_3d_vector_field.py b/examples/04_vector_field/01_3d_vector_field.py index 13d76e87..5b1872bd 100755 --- a/examples/04_vector_field/01_3d_vector_field.py +++ b/examples/04_vector_field/01_3d_vector_field.py @@ -6,6 +6,7 @@ Gaussian covariance model. The mesh on which we generate the field will be externally defined and it will be generated by PyVista. """ + # sphinx_gallery_thumbnail_path = 'pics/GS_3d_vector_field.png' import pyvista as pv @@ -17,7 +18,7 @@ ############################################################################### # create a uniform grid with PyVista dims, spacing, origin = (40, 30, 10), (1, 1, 1), (-10, 0, 0) -mesh = pv.UniformGrid(dims=dims, spacing=spacing, origin=origin) +mesh = pv.ImageData(dimensions=dims, spacing=spacing, origin=origin) ############################################################################### # create an incompressible random 3d velocity field on the given mesh diff --git a/examples/05_kriging/00_simple_kriging.py b/examples/05_kriging/00_simple_kriging.py index bb83a32c..1a245b6c 100755 --- a/examples/05_kriging/00_simple_kriging.py +++ b/examples/05_kriging/00_simple_kriging.py @@ -27,6 +27,7 @@ The mean of the field has to be given beforehand. """ + import numpy as np from gstools import Gaussian, krige diff --git a/examples/05_kriging/01_ordinary_kriging.py b/examples/05_kriging/01_ordinary_kriging.py index 3bc54bd9..d26254ef 100644 --- a/examples/05_kriging/01_ordinary_kriging.py +++ b/examples/05_kriging/01_ordinary_kriging.py @@ -27,6 +27,7 @@ Here we use ordinary kriging in 1D (for plotting reasons) with 5 given observations/conditions. The estimated mean can be accessed by ``krig.mean``. """ + import numpy as np from gstools import Gaussian, krige diff --git a/examples/05_kriging/02_pykrige_interface.py b/examples/05_kriging/02_pykrige_interface.py index bb166d29..a6fbf03e 100755 --- a/examples/05_kriging/02_pykrige_interface.py +++ b/examples/05_kriging/02_pykrige_interface.py @@ -11,6 +11,7 @@ To demonstrate the general workflow, we compare ordinary kriging of PyKrige with the corresponding GSTools routine in 2D: """ + import numpy as np from matplotlib import pyplot as plt from pykrige.ok import OrdinaryKriging diff --git a/examples/05_kriging/03_compare_kriging.py b/examples/05_kriging/03_compare_kriging.py index 12f5c43b..463faa0a 100755 --- a/examples/05_kriging/03_compare_kriging.py +++ b/examples/05_kriging/03_compare_kriging.py @@ -2,6 +2,7 @@ Compare Kriging --------------- """ + import matplotlib.pyplot as plt import numpy as np diff --git a/examples/05_kriging/04_extdrift_kriging.py b/examples/05_kriging/04_extdrift_kriging.py index b9bbb54a..2e6a168a 100755 --- a/examples/05_kriging/04_extdrift_kriging.py +++ b/examples/05_kriging/04_extdrift_kriging.py @@ -2,6 +2,7 @@ External Drift Kriging ---------------------- """ + import numpy as np from gstools import SRF, Gaussian, krige diff --git a/examples/05_kriging/05_universal_kriging.py b/examples/05_kriging/05_universal_kriging.py index d3cf2880..5501694a 100755 --- a/examples/05_kriging/05_universal_kriging.py +++ b/examples/05_kriging/05_universal_kriging.py @@ -13,6 +13,7 @@ To access only the estimated mean/drift, we provide a switch `only_mean` in the call routine. """ + import numpy as np from gstools import SRF, Gaussian, krige diff --git a/examples/05_kriging/06_detrended_kriging.py b/examples/05_kriging/06_detrended_kriging.py index b68e6502..6d20cf1d 100755 --- a/examples/05_kriging/06_detrended_kriging.py +++ b/examples/05_kriging/06_detrended_kriging.py @@ -2,6 +2,7 @@ Detrended Kriging ----------------- """ + import numpy as np from gstools import SRF, Gaussian, krige diff --git a/examples/05_kriging/07_detrended_ordinary_kriging.py b/examples/05_kriging/07_detrended_ordinary_kriging.py index 9a59fc33..81d01744 100755 --- a/examples/05_kriging/07_detrended_ordinary_kriging.py +++ b/examples/05_kriging/07_detrended_ordinary_kriging.py @@ -2,6 +2,7 @@ Detrended Ordinary Kriging -------------------------- """ + import numpy as np from gstools import SRF, Gaussian, krige diff --git a/examples/05_kriging/09_pseudo_inverse.py b/examples/05_kriging/09_pseudo_inverse.py index 9920a956..7615d888 100755 --- a/examples/05_kriging/09_pseudo_inverse.py +++ b/examples/05_kriging/09_pseudo_inverse.py @@ -17,6 +17,7 @@ In the following we have two different values at the same location. The resulting kriging field will hold the average at this point. """ + import numpy as np from gstools import Gaussian, krige diff --git a/examples/06_conditioned_fields/00_condition_ensemble.py b/examples/06_conditioned_fields/00_condition_ensemble.py index 716edd20..5cc07eed 100644 --- a/examples/06_conditioned_fields/00_condition_ensemble.py +++ b/examples/06_conditioned_fields/00_condition_ensemble.py @@ -6,6 +6,7 @@ with 5 given observations/conditions, to generate an ensemble of conditioned random fields. """ + import matplotlib.pyplot as plt import numpy as np diff --git a/examples/06_conditioned_fields/01_2D_condition_ensemble.py b/examples/06_conditioned_fields/01_2D_condition_ensemble.py index 4a03e66b..81f51464 100644 --- a/examples/06_conditioned_fields/01_2D_condition_ensemble.py +++ b/examples/06_conditioned_fields/01_2D_condition_ensemble.py @@ -4,6 +4,7 @@ Let's create an ensemble of conditioned random fields in 2D. """ + import matplotlib.pyplot as plt import numpy as np diff --git a/examples/07_transformations/00_log_normal.py b/examples/07_transformations/00_log_normal.py index 3d2f4532..d44c1627 100755 --- a/examples/07_transformations/00_log_normal.py +++ b/examples/07_transformations/00_log_normal.py @@ -6,6 +6,7 @@ See :any:`transform.normal_to_lognormal` """ + import gstools as gs # structured field with a size of 100x100 and a grid-size of 1x1 diff --git a/examples/07_transformations/01_binary.py b/examples/07_transformations/01_binary.py index a403abb7..125e29d0 100755 --- a/examples/07_transformations/01_binary.py +++ b/examples/07_transformations/01_binary.py @@ -8,6 +8,7 @@ See :any:`transform.binary` """ + import gstools as gs # structured field with a size of 100x100 and a grid-size of 1x1 diff --git a/examples/07_transformations/02_discrete.py b/examples/07_transformations/02_discrete.py index 268cc960..48f67a2d 100755 --- a/examples/07_transformations/02_discrete.py +++ b/examples/07_transformations/02_discrete.py @@ -9,6 +9,7 @@ See :any:`transform.discrete` """ + import numpy as np import gstools as gs diff --git a/examples/07_transformations/03_zinn_harvey.py b/examples/07_transformations/03_zinn_harvey.py index 01e65790..fad1fb64 100755 --- a/examples/07_transformations/03_zinn_harvey.py +++ b/examples/07_transformations/03_zinn_harvey.py @@ -9,6 +9,7 @@ See :any:`transform.zinnharvey` """ + import gstools as gs # structured field with a size of 100x100 and a grid-size of 1x1 diff --git a/examples/07_transformations/04_bimodal.py b/examples/07_transformations/04_bimodal.py index 8234a486..4dd6fb29 100755 --- a/examples/07_transformations/04_bimodal.py +++ b/examples/07_transformations/04_bimodal.py @@ -11,6 +11,7 @@ See: :any:`transform.normal_to_arcsin` and :any:`transform.normal_to_uquad` """ + import gstools as gs # structured field with a size of 100x100 and a grid-size of 1x1 diff --git a/examples/07_transformations/05_combinations.py b/examples/07_transformations/05_combinations.py index fb27c5e6..1fbe367e 100755 --- a/examples/07_transformations/05_combinations.py +++ b/examples/07_transformations/05_combinations.py @@ -15,6 +15,7 @@ If you don't specify `field` and `store` everything happens inplace. """ + # sphinx_gallery_thumbnail_number = 1 import gstools as gs diff --git a/examples/08_geo_coordinates/00_field_generation.py b/examples/08_geo_coordinates/00_field_generation.py index b7b3748a..5b1a6fca 100755 --- a/examples/08_geo_coordinates/00_field_generation.py +++ b/examples/08_geo_coordinates/00_field_generation.py @@ -16,6 +16,7 @@ To generate the field, we simply pass ``(lat, lon)`` as the position tuple to the :any:`SRF` class. """ + import numpy as np import gstools as gs diff --git a/examples/08_geo_coordinates/01_dwd_krige.py b/examples/08_geo_coordinates/01_dwd_krige.py index 44b9f28c..3c17fb7e 100755 --- a/examples/08_geo_coordinates/01_dwd_krige.py +++ b/examples/08_geo_coordinates/01_dwd_krige.py @@ -15,6 +15,7 @@ In order to keep the number of dependecies low, the calls of both functions shown beneath are commented out. """ + # sphinx_gallery_thumbnail_number = 2 import matplotlib.pyplot as plt import numpy as np diff --git a/examples/09_spatio_temporal/03_geographic_coordinates.py b/examples/09_spatio_temporal/03_geographic_coordinates.py index 56684a94..b1cfbff6 100644 --- a/examples/09_spatio_temporal/03_geographic_coordinates.py +++ b/examples/09_spatio_temporal/03_geographic_coordinates.py @@ -18,6 +18,7 @@ We will set a spatial length-scale of `1000` and a time length-scale of `100` days. """ + import numpy as np import gstools as gs diff --git a/examples/10_normalizer/00_lognormal_kriging.py b/examples/10_normalizer/00_lognormal_kriging.py index b7bf4a8a..9880bc37 100644 --- a/examples/10_normalizer/00_lognormal_kriging.py +++ b/examples/10_normalizer/00_lognormal_kriging.py @@ -14,6 +14,7 @@ In this example we will use ordinary kriging. """ + import numpy as np import gstools as gs diff --git a/examples/10_normalizer/01_auto_fit.py b/examples/10_normalizer/01_auto_fit.py index 80a49569..71ad1385 100644 --- a/examples/10_normalizer/01_auto_fit.py +++ b/examples/10_normalizer/01_auto_fit.py @@ -18,6 +18,7 @@ We will generate the "original" field on a 60x60 mesh, from which we will take samples in order to pretend a situation of data-scarcity. """ + import matplotlib.pyplot as plt import numpy as np diff --git a/examples/10_normalizer/02_compare.py b/examples/10_normalizer/02_compare.py index 24f7c9f8..2dd74488 100644 --- a/examples/10_normalizer/02_compare.py +++ b/examples/10_normalizer/02_compare.py @@ -6,6 +6,7 @@ But first, we define a convenience routine and make some imports as always. """ + import matplotlib.pyplot as plt import numpy as np diff --git a/pyproject.toml b/pyproject.toml index 70b7d0a6..17bcb246 100644 --- a/pyproject.toml +++ b/pyproject.toml @@ -2,8 +2,9 @@ requires = [ "setuptools>=64", "setuptools_scm>=7", - "oldest-supported-numpy", - "Cython>=3.0", + "numpy>=2.0.0rc1,<2.3; python_version >= '3.9'", + "oldest-supported-numpy; python_version < '3.9'", + "Cython>=3.0.10,<3.1.0", "extension-helpers>=1", ] build-backend = "setuptools.build_meta" @@ -48,7 +49,7 @@ dependencies = [ "emcee>=3.0.0", "hankel>=1.0.0", "meshio>=5.1.0", - "numpy>=1.14.5", + "numpy>=1.20.0", "pyevtk>=1.1.1", "scipy>=1.1.0", ] @@ -60,20 +61,20 @@ doc = [ "meshzoo>=0.7", "numpydoc>=1.1", "pykrige>=1.5,<2", - "pyvista>=0.29", - "sphinx>=4", + "pyvista>=0.40", + "sphinx>=7", "sphinx-gallery>=0.8", - "sphinx-rtd-theme>=1,<1.1", + "sphinx-rtd-theme>=2", "sphinxcontrib-youtube>=1.1", ] plotting = [ "matplotlib>=3", - "pyvista>=0.29", + "pyvista>=0.40", ] rust = ["gstools_core>=0.2.0,<1"] test = ["pytest-cov>=3"] lint = [ - "black", + "black>=24", "pylint", "isort[colors]", "cython-lint", @@ -161,11 +162,9 @@ target-version = [ # Switch to using build build-frontend = "build" # Disable building PyPy wheels on all platforms, 32bit for py3.10/11/12, musllinux builds, py3.6/7 -skip = ["cp36-", "cp37-", "pp", "cp31-win32", "cp31-manylinux_i686", "-musllinux_"] +skip = ["cp36-", "cp37-", "pp", "-win32", "-manylinux_i686", "-musllinux_"] # Run the package tests using `pytest` test-extras = "test" test-command = "pytest -v {package}/tests" # Skip trying to test arm64 builds on Intel Macs test-skip = "-macosx_arm64 -macosx_universal2:arm64" -# no wheels for linux-32bit anymore for numpy>=1.22 -environment = "PIP_PREFER_BINARY=1" diff --git a/setup.py b/setup.py index f465ffd4..b27548a9 100644 --- a/setup.py +++ b/setup.py @@ -1,4 +1,5 @@ """GSTools: A geostatistical toolbox.""" + import os import numpy as np diff --git a/src/gstools/__init__.py b/src/gstools/__init__.py index 6d62d558..11e63a2b 100644 --- a/src/gstools/__init__.py +++ b/src/gstools/__init__.py @@ -129,6 +129,7 @@ DEGREE_SCALE RADIAN_SCALE """ + # Hooray! from gstools import ( config, diff --git a/src/gstools/config.py b/src/gstools/config.py index f0a87bb6..24ce20c7 100644 --- a/src/gstools/config.py +++ b/src/gstools/config.py @@ -4,6 +4,7 @@ .. currentmodule:: gstools.config """ + NUM_THREADS = None # pylint: disable=W0611 diff --git a/src/gstools/covmodel/base.py b/src/gstools/covmodel/base.py index be530075..23e19881 100644 --- a/src/gstools/covmodel/base.py +++ b/src/gstools/covmodel/base.py @@ -8,6 +8,7 @@ .. autosummary:: CovModel """ + # pylint: disable=C0103, R0201, E1101, C0302, W0613 import copy diff --git a/src/gstools/covmodel/fit.py b/src/gstools/covmodel/fit.py index dc2d5a3a..b420be70 100755 --- a/src/gstools/covmodel/fit.py +++ b/src/gstools/covmodel/fit.py @@ -8,6 +8,7 @@ .. autosummary:: fit_variogram """ + # pylint: disable=C0103, W0632 import numpy as np from scipy.optimize import curve_fit @@ -308,9 +309,11 @@ def _pre_init_guess(model, init_guess, mean_x=1.0, mean_y=1.0): for opt in model.opt_arg: init_guess.setdefault( opt, - default_arg_from_bounds(bnd[opt]) - if default - else getattr(model, opt), + ( + default_arg_from_bounds(bnd[opt]) + if default + else getattr(model, opt) + ), ) # convert all init guesses to float (except "anis") for arg in model.iso_arg: diff --git a/src/gstools/covmodel/models.py b/src/gstools/covmodel/models.py index a2aaf6a3..ab94e279 100644 --- a/src/gstools/covmodel/models.py +++ b/src/gstools/covmodel/models.py @@ -20,6 +20,7 @@ SuperSpherical JBessel """ + # pylint: disable=C0103, E1101, R0201 import warnings diff --git a/src/gstools/covmodel/plot.py b/src/gstools/covmodel/plot.py index 334ccbe2..32148c14 100644 --- a/src/gstools/covmodel/plot.py +++ b/src/gstools/covmodel/plot.py @@ -24,6 +24,7 @@ plot_spectral_density plot_spectral_rad_pdf """ + # pylint: disable=C0103, C0415, E1130 import numpy as np diff --git a/src/gstools/covmodel/tpl_models.py b/src/gstools/covmodel/tpl_models.py index 5e980760..b728e7b9 100644 --- a/src/gstools/covmodel/tpl_models.py +++ b/src/gstools/covmodel/tpl_models.py @@ -11,6 +11,7 @@ TPLStable TPLSimple """ + # pylint: disable=C0103, E1101 import warnings diff --git a/src/gstools/field/base.py b/src/gstools/field/base.py index bb514141..2006b858 100755 --- a/src/gstools/field/base.py +++ b/src/gstools/field/base.py @@ -8,6 +8,7 @@ .. autosummary:: Field """ + # pylint: disable=C0103, C0415 from collections.abc import Iterable from copy import copy @@ -362,7 +363,7 @@ def transform( ---------- method : :class:`str` Method to use. - See :any:`gstools.transform` for available transformations. + See :py:mod:`gstools.transform` for available transformations. field : :class:`str`, optional Name of field to be transformed. The default is "field". store : :class:`str` or :class:`bool`, optional diff --git a/src/gstools/field/cond_srf.py b/src/gstools/field/cond_srf.py index 385216ba..c3e03fe2 100644 --- a/src/gstools/field/cond_srf.py +++ b/src/gstools/field/cond_srf.py @@ -8,6 +8,7 @@ .. autosummary:: CondSRF """ + # pylint: disable=C0103, W0231, W0221, W0222, E1102 import numpy as np diff --git a/src/gstools/field/generator.py b/src/gstools/field/generator.py index 571c9733..5beab10d 100644 --- a/src/gstools/field/generator.py +++ b/src/gstools/field/generator.py @@ -12,6 +12,7 @@ RandMeth IncomprRandMeth """ + # pylint: disable=C0103, W0222, C0412, W0231 import warnings from abc import ABC, abstractmethod diff --git a/src/gstools/field/plot.py b/src/gstools/field/plot.py index d06a22a1..ab28c974 100644 --- a/src/gstools/field/plot.py +++ b/src/gstools/field/plot.py @@ -9,6 +9,7 @@ plot_field plot_vec_field """ + # pylint: disable=C0103, W0613, E1101 import numpy as np from scipy import interpolate as inter diff --git a/src/gstools/field/srf.py b/src/gstools/field/srf.py index a8a1e575..d88e46c0 100644 --- a/src/gstools/field/srf.py +++ b/src/gstools/field/srf.py @@ -8,6 +8,7 @@ .. autosummary:: SRF """ + # pylint: disable=C0103, W0221, E1102 import numpy as np diff --git a/src/gstools/field/tools.py b/src/gstools/field/tools.py index 4b128d8a..dfa2e3c6 100644 --- a/src/gstools/field/tools.py +++ b/src/gstools/field/tools.py @@ -10,6 +10,7 @@ to_vtk_helper generate_on_mesh """ + # pylint: disable=W0212, C0415 import meshio import numpy as np diff --git a/src/gstools/field/upscaling.py b/src/gstools/field/upscaling.py index 6dcbaa7b..857bfc45 100644 --- a/src/gstools/field/upscaling.py +++ b/src/gstools/field/upscaling.py @@ -11,6 +11,7 @@ var_coarse_graining var_no_scaling """ + # pylint: disable=W0613 import warnings diff --git a/src/gstools/krige/__init__.py b/src/gstools/krige/__init__.py index bb3ef699..66d03246 100644 --- a/src/gstools/krige/__init__.py +++ b/src/gstools/krige/__init__.py @@ -16,6 +16,7 @@ ExtDrift Detrended """ + from gstools.krige.base import Krige from gstools.krige.methods import ( Detrended, diff --git a/src/gstools/krige/base.py b/src/gstools/krige/base.py index 336f4cc0..49a4f62f 100755 --- a/src/gstools/krige/base.py +++ b/src/gstools/krige/base.py @@ -8,6 +8,7 @@ .. autosummary:: Krige """ + # pylint: disable=C0103, W0221, E1102, R0201, C0412 import collections @@ -383,9 +384,7 @@ def _pre_ext_drift(self, pnt_cnt, ext_drift=None, set_cond=False): the drift values at the given positions """ if ext_drift is not None: - ext_drift = np.array( - ext_drift, dtype=np.double, ndmin=2, copy=False - ) + ext_drift = np.atleast_2d(np.asarray(ext_drift, dtype=np.double)) if ext_drift.size == 0: # treat empty array as no ext_drift return np.array([]) if set_cond: diff --git a/src/gstools/krige/methods.py b/src/gstools/krige/methods.py index b258a02d..19ffed56 100644 --- a/src/gstools/krige/methods.py +++ b/src/gstools/krige/methods.py @@ -12,6 +12,7 @@ ExtDrift Detrended """ + # pylint: disable=C0103 from gstools.krige.base import Krige diff --git a/src/gstools/krige/tools.py b/src/gstools/krige/tools.py index e3112ae4..62926595 100644 --- a/src/gstools/krige/tools.py +++ b/src/gstools/krige/tools.py @@ -9,6 +9,7 @@ set_condition get_drift_functions """ + # pylint: disable=C0103 from itertools import combinations_with_replacement diff --git a/src/gstools/normalizer/base.py b/src/gstools/normalizer/base.py index 0072d6a9..4a8477c6 100644 --- a/src/gstools/normalizer/base.py +++ b/src/gstools/normalizer/base.py @@ -8,6 +8,7 @@ .. autosummary:: Normalizer """ + # pylint: disable=R0201 import warnings diff --git a/src/gstools/normalizer/methods.py b/src/gstools/normalizer/methods.py index f66cfd19..a46dc230 100644 --- a/src/gstools/normalizer/methods.py +++ b/src/gstools/normalizer/methods.py @@ -13,6 +13,7 @@ Modulus Manly """ + # pylint: disable=E1101 import numpy as np diff --git a/src/gstools/normalizer/tools.py b/src/gstools/normalizer/tools.py index 7a5237b7..3e395d29 100644 --- a/src/gstools/normalizer/tools.py +++ b/src/gstools/normalizer/tools.py @@ -9,6 +9,7 @@ apply_mean_norm_trend remove_trend_norm_mean """ + import numpy as np from gstools.normalizer.base import Normalizer diff --git a/src/gstools/random/rng.py b/src/gstools/random/rng.py index 4f698557..ad78a0c6 100644 --- a/src/gstools/random/rng.py +++ b/src/gstools/random/rng.py @@ -8,6 +8,7 @@ .. autosummary:: RNG """ + # pylint: disable=E1101 import emcee as mc import numpy as np diff --git a/src/gstools/tools/export.py b/src/gstools/tools/export.py index 3e522c92..38254ceb 100644 --- a/src/gstools/tools/export.py +++ b/src/gstools/tools/export.py @@ -13,6 +13,7 @@ to_vtk_structured to_vtk_unstructured """ + # pylint: disable=C0103, E1101 import numpy as np from pyevtk.hl import gridToVTK, pointsToVTK diff --git a/src/gstools/tools/geometric.py b/src/gstools/tools/geometric.py index 7f2ea10e..55408965 100644 --- a/src/gstools/tools/geometric.py +++ b/src/gstools/tools/geometric.py @@ -29,6 +29,7 @@ chordal_to_great_circle great_circle_to_chordal """ + # pylint: disable=C0103 import numpy as np @@ -383,7 +384,7 @@ def generate_st_grid(pos, time, mesh_type="unstructured"): if mesh_type != "unstructured": pos = generate_grid(pos) else: - pos = np.array(pos, dtype=np.double, ndmin=2, copy=False) + pos = np.atleast_2d(np.asarray(pos, dtype=np.double)) out = [np.repeat(p.reshape(-1), np.size(time)) for p in pos] out.append(np.tile(time, np.size(pos[0]))) return np.asarray(out, dtype=np.double) @@ -552,7 +553,7 @@ def format_unstruct_pos_shape(pos, shape, check_stacked_shape=False): # now we try to be smart pre_len = len(np.atleast_1d(pos)) # care about 1D: pos can be given as 1D array here -> convert to 2D array - pos = np.array(pos, dtype=np.double, ndmin=2, copy=False) + pos = np.atleast_2d(np.asarray(pos, dtype=np.double)) post_len = len(pos) # first array dimension should be spatial dimension (1D is special case) dim = post_len if pre_len == post_len else 1 @@ -606,7 +607,7 @@ def ang2dir(angles, dtype=np.double, dim=None): the array of direction vectors """ pre_dim = np.asanyarray(angles).ndim - angles = np.array(angles, ndmin=2, dtype=dtype, copy=False) + angles = np.atleast_2d(np.asarray(angles, dtype=dtype)) if len(angles.shape) > 2: raise ValueError(f"Can't interpret angles array {angles}") dim = angles.shape[1] + 1 if dim is None else dim diff --git a/src/gstools/tools/misc.py b/src/gstools/tools/misc.py index 068200b7..aaba1501 100755 --- a/src/gstools/tools/misc.py +++ b/src/gstools/tools/misc.py @@ -10,6 +10,7 @@ list_format eval_func """ + # pylint: disable=C0103, C0415 import numpy as np diff --git a/src/gstools/tools/special.py b/src/gstools/tools/special.py index c48887ea..1457b736 100644 --- a/src/gstools/tools/special.py +++ b/src/gstools/tools/special.py @@ -14,6 +14,7 @@ tpl_exp_spec_dens tpl_gau_spec_dens """ + # pylint: disable=C0103, E1101 import numpy as np from scipy import special as sps diff --git a/src/gstools/transform/array.py b/src/gstools/transform/array.py index 30c92bb6..87564edf 100644 --- a/src/gstools/transform/array.py +++ b/src/gstools/transform/array.py @@ -18,6 +18,7 @@ array_to_arcsin array_to_uquad """ + # pylint: disable=C0103, C0123, R0911 from warnings import warn diff --git a/src/gstools/transform/field.py b/src/gstools/transform/field.py index 5d204407..4a281564 100644 --- a/src/gstools/transform/field.py +++ b/src/gstools/transform/field.py @@ -26,6 +26,7 @@ normal_to_arcsin normal_to_uquad """ + # pylint: disable=C0103, C0123, R0911, R1735 import numpy as np @@ -111,7 +112,7 @@ def apply(fld, method, field="field", store=True, process=False, *kwargs): Field class containing a generated field. method : :class:`str` Method to use. - See :any:`gstools.transform` for available transformations. + See :py:mod:`gstools.transform` for available transformations. field : :class:`str`, optional Name of field to be transformed. The default is "field". store : :class:`str` or :class:`bool`, optional diff --git a/src/gstools/variogram/binning.py b/src/gstools/variogram/binning.py index e8e42f38..86d4fdc2 100644 --- a/src/gstools/variogram/binning.py +++ b/src/gstools/variogram/binning.py @@ -8,6 +8,7 @@ .. autosummary:: standard_bins """ + import numpy as np from gstools.tools import RADIAN_SCALE diff --git a/src/gstools/variogram/estimator.pyx b/src/gstools/variogram/estimator.pyx index ab418e32..e00824be 100644 --- a/src/gstools/variogram/estimator.pyx +++ b/src/gstools/variogram/estimator.pyx @@ -116,7 +116,7 @@ ctypedef double (_estimator_func)(const double) nogil cdef inline void normalization_matheron( double[:] variogram, - long[:] counts, + np.int64_t[:] counts, ): cdef int i for i in range(variogram.shape[0]): @@ -125,10 +125,10 @@ cdef inline void normalization_matheron( cdef inline void normalization_cressie( double[:] variogram, - long[:] counts, + np.int64_t[:] counts, ): cdef int i - cdef long cnt + cdef np.int64_t cnt for i in range(variogram.shape[0]): # avoid division by zero cnt = max(counts[i], 1) @@ -139,12 +139,12 @@ cdef inline void normalization_cressie( ctypedef void (_normalization_func)( double[:], - long[:], + np.int64_t[:], ) cdef inline void normalization_matheron_vec( double[:, :] variogram, - long[:, :] counts, + np.int64_t[:, :] counts, ): cdef int d for d in range(variogram.shape[0]): @@ -152,7 +152,7 @@ cdef inline void normalization_matheron_vec( cdef inline void normalization_cressie_vec( double[:, :] variogram, - long[:, :] counts, + np.int64_t[:, :] counts, ): cdef int d for d in range(variogram.shape[0]): @@ -160,7 +160,7 @@ cdef inline void normalization_cressie_vec( ctypedef void (_normalization_func_vec)( double[:, :], - long[:, :], + np.int64_t[:, :], ) cdef _estimator_func choose_estimator_func(str estimator_type): @@ -221,7 +221,7 @@ def directional( cdef int f_max = f.shape[0] cdef double[:, :] variogram = np.zeros((d_max, len(bin_edges)-1)) - cdef long[:, :] counts = np.zeros((d_max, len(bin_edges)-1), dtype=long) + cdef np.int64_t[:, :] counts = np.zeros((d_max, len(bin_edges)-1), dtype=np.int64) cdef int i, j, k, m, d cdef double dist @@ -287,7 +287,7 @@ def unstructured( cdef int f_max = f.shape[0] cdef double[:] variogram = np.zeros(len(bin_edges)-1) - cdef long[:] counts = np.zeros(len(bin_edges)-1, dtype=long) + cdef np.int64_t[:] counts = np.zeros(len(bin_edges)-1, dtype=np.int64) cdef int i, j, k, m cdef double dist @@ -324,7 +324,7 @@ def structured( cdef int k_max = i_max + 1 cdef double[:] variogram = np.zeros(k_max) - cdef long[:] counts = np.zeros(k_max, dtype=long) + cdef np.int64_t[:] counts = np.zeros(k_max, dtype=np.int64) cdef int i, j, k cdef int num_threads_c = set_num_threads(num_threads) @@ -356,7 +356,7 @@ def ma_structured( cdef int k_max = i_max + 1 cdef double[:] variogram = np.zeros(k_max) - cdef long[:] counts = np.zeros(k_max, dtype=long) + cdef np.int64_t[:] counts = np.zeros(k_max, dtype=np.int64) cdef int i, j, k cdef int num_threads_c = set_num_threads(num_threads) diff --git a/src/gstools/variogram/variogram.py b/src/gstools/variogram/variogram.py index 31e29191..afcf336f 100644 --- a/src/gstools/variogram/variogram.py +++ b/src/gstools/variogram/variogram.py @@ -9,6 +9,7 @@ vario_estimate vario_estimate_axis """ + # pylint: disable=C0412 import numpy as np @@ -262,7 +263,7 @@ def vario_estimate( John Wiley & Sons. (2007) """ if bin_edges is not None: - bin_edges = np.array(bin_edges, ndmin=1, dtype=np.double, copy=False) + bin_edges = np.atleast_1d(np.asarray(bin_edges, dtype=np.double)) bin_centers = (bin_edges[:-1] + bin_edges[1:]) / 2.0 # allow multiple fields at same positions (ndmin=2: first axis -> field ID) # need to convert to ma.array, since list of ma.array is not recognised @@ -315,7 +316,7 @@ def vario_estimate( # set directions dir_no = 0 if direction is not None and dim > 1: - direction = np.array(direction, ndmin=2, dtype=np.double, copy=False) + direction = np.atleast_2d(np.asarray(direction, dtype=np.double)) if len(direction.shape) > 2: raise ValueError(f"Can't interpret directions: {direction}") if direction.shape[1] != dim: @@ -473,7 +474,7 @@ def vario_estimate_axis( if not config.USE_RUST: mask = np.asarray(mask, dtype=np.int32) else: - field = np.array(field, ndmin=1, dtype=np.double, copy=False) + field = np.atleast_1d(np.asarray(field, dtype=np.double)) missing_mask = None # free space axis_to_swap = AXIS_DIR[direction] if direction in AXIS else int(direction)	Numpy 2 compatibility Numpy 2.0.0rc1 was released: https://pypi.org/project/numpy/2.0.0c1 and is officially usable to build extensions. We should be prepared to be compatible with it. - Building with numpy 2.0.0rc1 works :tada: - Running tests with numpy 2 fails :sob: A lot of the tests fail because of emcee and this issue: https://github.com/dfm/emcee/issues/509 Other errors come from these lines in `estimator.pyx`: ```cython cdef long[:, :] counts = np.zeros((d_max, len(bin_edges)-1), dtype=long) cdef long[:] counts = np.zeros(len(bin_edges)-1, dtype=long) cdef long[:] counts = np.zeros(k_max, dtype=long) cdef long[:] counts = np.zeros(k_max, dtype=long) ``` and raise ``` E ValueError: Buffer dtype mismatch, expected 'long' but got 'long long' ``` This only occurs on windows and is related to this: https://numpy.org/devdocs/numpy_2_0_migration_guide.html#windows-default-integer One other category of errors is: ``` If using `np.array(obj, copy=False)` replace it with `np.asarray(obj)` to allow a copy when needed (no behavior change in NumPy 1.x). ``` `np.array(..., copy=False)` is used in [krige/base.py](https://github.com/GeoStat-Framework/GSTools/blob/90003fbbe193cfa835a891b403b4024bb21867f9/src/gstools/krige/base.py#L387), 3 times in [variogram/variogram.py](https://github.com/GeoStat-Framework/GSTools/blob/90003fbbe193cfa835a891b403b4024bb21867f9/src/gstools/variogram/variogram.py#L265) and 3 times in [tools/geometric.py](https://github.com/GeoStat-Framework/GSTools/blob/90003fbbe193cfa835a891b403b4024bb21867f9/src/gstools/tools/geometric.py#L386) In order to mimic the `ndmin=...` feature, we could replace these with e.g.: ```python pos = np.atleast_2d(np.asarray(pos, dtype=np.double)) ```		2024-03-14T16:12:28	[]	[]
GeoStat-Framework/GSTools	GeoStat-Framework__GSTools-145	d66613bebe69ad1fbeb8cdac86e2db4bee999b2f	diff --git a/gstools/covmodel/base.py b/gstools/covmodel/base.py index ec34f999..e121acb2 100644 --- a/gstools/covmodel/base.py +++ b/gstools/covmodel/base.py @@ -36,6 +36,7 @@ set_dim, compare, model_repr, + default_arg_from_bounds, ) from gstools.covmodel import plot from gstools.covmodel.fit import fit_variogram @@ -414,7 +415,10 @@ def default_opt_arg(self): Should be given as a dictionary when overridden. """ - return {} + return { + opt: default_arg_from_bounds(bnd) + for (opt, bnd) in self.default_opt_arg_bounds().items() + } def default_opt_arg_bounds(self): """Provide default boundaries for optional arguments.""" @@ -607,11 +611,19 @@ def fit_variogram( and set to the current sill of the model. Then, the procedure above is applied. Default: None - init_guess : :class:`str`, optional + init_guess : :class:`str` or :class:`dict`, optional Initial guess for the estimation. Either: * "default": using the default values of the covariance model + ("len_scale" will be mean of given bin centers; + "var" and "nugget" will be mean of given variogram values + (if in given bounds)) * "current": using the current values of the covariance model + * dict: dictionary with parameter names and given value + (separate "default" can bet set to "default" or "current" for + unspecified values to get same behavior as given above + ("default" by default)) + Example: ``{"len_scale": 10, "default": "current"}`` Default: "default" weights : :class:`str`, :class:`numpy.ndarray`, :class:`callable`, optional diff --git a/gstools/covmodel/fit.py b/gstools/covmodel/fit.py index 0fa17075..96462d9e 100755 --- a/gstools/covmodel/fit.py +++ b/gstools/covmodel/fit.py @@ -77,11 +77,19 @@ def fit_variogram( and set to the current sill of the model. Then, the procedure above is applied. Default: None - init_guess : :class:`str`, optional + init_guess : :class:`str` or :class:`dict`, optional Initial guess for the estimation. Either: * "default": using the default values of the covariance model + ("len_scale" will be mean of given bin centers; + "var" and "nugget" will be mean of given variogram values + (if in given bounds)) * "current": using the current values of the covariance model + * dict: dictionary with parameter names and given value + (separate "default" can bet set to "default" or "current" for + unspecified values to get same behavior as given above + ("default" by default)) + Example: ``{"len_scale": 10, "default": "current"}`` Default: "default" weights : :class:`str`, :class:`numpy.ndarray`, :class:`callable`optional @@ -170,6 +178,10 @@ def fit_variogram( # prepare variogram data # => concatenate directional variograms to have a 1D array for x and y x_data, y_data, is_dir_vario = _check_vario(model, x_data, y_data) + # prepare init guess dictionary + init_guess = _pre_init_guess( + model, init_guess, np.mean(x_data), np.mean(y_data) + ) # only fit anisotropy if a directional variogram was given anis &= is_dir_vario # set weights @@ -239,7 +251,7 @@ def _pre_para(model, para_select, sill, anis): if model.var > sill: raise ValueError( "fit: if sill is fixed and variance deselected, " - + "the set variance should be less than the given sill." + "the set variance should be less than the given sill." ) para_select["nugget"] = False model.nugget = sill - model.var @@ -247,7 +259,7 @@ def _pre_para(model, para_select, sill, anis): if model.nugget > sill: raise ValueError( "fit: if sill is fixed and nugget deselected, " - + "the set nugget should be less than the given sill." + "the set nugget should be less than the given sill." ) para_select["var"] = False model.var = sill - model.nugget @@ -269,6 +281,51 @@ def _pre_para(model, para_select, sill, anis): return para, sill, constrain_sill, anis +def _pre_init_guess(model, init_guess, mean_x=1.0, mean_y=1.0): + # init guess should be a dict + if not isinstance(init_guess, dict): + init_guess = {"default": init_guess} + # "default" init guess is the respective default value + default_guess = init_guess.pop("default", "default") + if default_guess not in ["default", "current"]: + raise ValueError( + "fit_variogram: unknown def. guess: {}".format(default_guess) + ) + default = default_guess == "default" + # check invalid names for given init guesses + invalid_para = set(init_guess) - set(model.iso_arg + ["anis"]) + if invalid_para: + raise ValueError( + "fit_variogram: unknown init guess: {}".format(invalid_para) + ) + bnd = model.arg_bounds + # default length scale is mean of given bin centers (respecting "rescale") + init_guess.setdefault( + "len_scale", mean_x * model.rescale if default else model.len_scale + ) + # init guess for variance and nugget is mean of given variogram + for par in ["var", "nugget"]: + init_guess.setdefault(par, mean_y if default else getattr(model, par)) + # anis setting + init_guess.setdefault( + "anis", default_arg_from_bounds(bnd["anis"]) if default else model.anis + ) + # correctly handle given values for anis (need a list of values) + init_guess["anis"] = list(set_anis(model.dim, init_guess["anis"])) + # set optional arguments + for opt in model.opt_arg: + init_guess.setdefault( + opt, + default_arg_from_bounds(bnd[opt]) + if default + else getattr(model, opt), + ) + # convert all init guesses to float (except "anis") + for arg in model.iso_arg: + init_guess[arg] = float(init_guess[arg]) + return init_guess + + def _check_vario(model, x_data, y_data): # prepare variogram data x_data = np.array(x_data).reshape(-1) @@ -283,8 +340,8 @@ def _check_vario(model, x_data, y_data): elif x_data.size != y_data.size: raise ValueError( "CovModel.fit_variogram: Wrong number of empirical variograms! " - + "Either provide only one variogram to fit an isotropic model, " - + "or directional ones for all main axes to fit anisotropy." + "Either provide only one variogram to fit an isotropic model, " + "or directional ones for all main axes to fit anisotropy." ) if is_dir_vario and model.latlon: raise ValueError( @@ -327,10 +384,7 @@ def _init_curve_fit_para(model, para, init_guess, constrain_sill, sill, anis): init_guess_list.append( _init_guess( bounds=[low_bounds[-1], top_bounds[-1]], - current=getattr(model, par), - default=model.rescale if par == "len_scale" else 1.0, - typ=init_guess, - para_name=par, + default=init_guess[par], ) ) for opt in model.opt_arg: @@ -340,37 +394,27 @@ def _init_curve_fit_para(model, para, init_guess, constrain_sill, sill, anis): init_guess_list.append( _init_guess( bounds=[low_bounds[-1], top_bounds[-1]], - current=getattr(model, opt), - default=model.default_opt_arg()[opt], - typ=init_guess, - para_name=opt, + default=init_guess[opt], ) ) if anis: - low_bounds += [model.anis_bounds[0]] * (model.dim - 1) - top_bounds += [model.anis_bounds[1]] * (model.dim - 1) - if init_guess == "default": - def_arg = default_arg_from_bounds(model.anis_bounds) - init_guess_list += [def_arg] * (model.dim - 1) - elif init_guess == "current": - init_guess_list += list(model.anis) - else: - raise ValueError( - "CovModel.fit: unknown init_guess: '{}'".format(init_guess) + for i in range(model.dim - 1): + low_bounds.append(model.anis_bounds[0]) + top_bounds.append(model.anis_bounds[1]) + init_guess_list.append( + _init_guess( + bounds=[low_bounds[-1], top_bounds[-1]], + default=init_guess["anis"][i], + ) ) - return (low_bounds, top_bounds), init_guess_list -def _init_guess(bounds, current, default, typ, para_name): +def _init_guess(bounds, default): """Proper determination of initial guess.""" - if typ == "default": - if bounds[0] < default < bounds[1]: - return default - return default_arg_from_bounds(bounds) - if typ == "current": - return current - raise ValueError("CovModel.fit: unknown init_guess: '{}'".format(typ)) + if bounds[0] < default < bounds[1]: + return default + return default_arg_from_bounds(bounds) def _get_curve(model, para, constrain_sill, sill, anis, is_dir_vario):	Variogram fitting does not work for 'large' coordinates? Hello, I have encountered an issue in fitting a variogram model to an estimated variogram and it seems to be connected to the used coordinates. Along the tutorial on variogram fitting I tried the following code: ``` bin_max = 300000 bins = np.linspace(0, bin_max, 10) bin_center, gamma = gstools.vario_estimate_unstructured((y_coord, x_coord), vals, bins) fit_model = gstools.Exponential(dim=2) fit_param, _ = fit_model.fit_variogram(bin_center, gamma) plt.plot(bin_center, gamma) gs.covmodel.plot.plot_variogram(fit_model, ax=plt.gca(), x_max=bins[-1]) ``` The arrays `x_coord` and `y_coord` contain the coordinates for data points in Germany in the Gauss-Krüger projection (EPSG 31467). Hence, they are in the range 3280000 to 3935000 for the x direction and 5235000 to 6105000 for the y direction. This is the resulting variogram plot, which does not look right: ![gstools_issue1](https://user-images.githubusercontent.com/64479100/111665034-abea9b00-8812-11eb-8389-460012c18e53.png) I tried to locate my error and found that when doing the following conversions beforehand, the result looks very different: ``` x_coord = x_coord / 10 6 y_coord = y_coord / 10 6 bin_max = bin_max / 10 ** 6 ``` This is the resulting variogram plot: ![gstools_issue2](https://user-images.githubusercontent.com/64479100/111665562-26b3b600-8813-11eb-8c71-2db8d401e49d.png) Is this expected behavior? -- I use Python 3.7.9 and gstools 1.2.1	Hey @krihabu , this is a problem with scipy's [curve_fit](https://docs.scipy.org/doc/scipy/reference/generated/scipy.optimize.curve_fit.html). The routine gives up to fast and doesn't search the higher regions for length-scales. There were some major changes since 1.2.1 in `develop` and v1.3 will be released soon. There you can provide initial guesses for the variogram fitting: ```python import numpy as np import gstools as gs ### SYNTHETIC DATA x = np.random.RandomState(19970221).rand(1000) * 10000. y = np.random.RandomState(20011012).rand(1000) * 10000. model = gs.Exponential(dim=2, var=2, len_scale=2000) srf = gs.SRF(model, mean=0, seed=19970221) field = srf((x, y)) ### SYNTHETIC DATA # estimate the variogram of the field (auto binning) bin_center, gamma = gs.vario_estimate((x, y), field) # fit the variogram with a stable model. (no nugget fitted) fit_model = gs.Stable(dim=2) fit_model.len_scale = 1000 # init guess for len_scale fit_model.fit_variogram(bin_center, gamma, nugget=False, init_guess="current") # output ax = fit_model.plot(x_max=max(bin_center)) ax.scatter(bin_center, gamma) print(fit_model) ``` ![vario_fit](https://user-images.githubusercontent.com/19690642/111751798-31169400-8895-11eb-9485-db70325f74d1.png) ```bash Stable(dim=2, var=1.7, len_scale=1.68e+03, nugget=0.0, anis=[1.0], angles=[0.0], alpha=0.932) ``` So providing an initial guess for len-scale could be a good tutorial (like taking 1/10 of the domain size as often reported as a good guess in literature). But you are right, it should work out of the box. Another idea I was thinking about for some time was to fit lenght scale via its order of magnitude, i.e. log-value. This could help finding the correct size. Hope this helps for now! Will keep this open to keep track of it. Cheers, Sebastian Just in case you didn't see the [instructions](https://geostat-framework.readthedocs.io/projects/gstools/en/latest/#installation) and you need a solution urgently, it's pretty easy to install the latest and not yet released version of GSTools via pip. Thank you for the quick response! I will go with my dirty fix for the moment and will try the `init_guess` option when the new version is released. Another option could be to change the scale of the model with the new `rescale` parameter, so that the `len_scale` is given in kilometer and not meter: ```python # estimate the variogram of the field bin_center, gamma = gs.vario_estimate((x, y), field) # fit the variogram with a rescale factor of 1/1000 fit_model = gs.Exponential(dim=2, rescale=1 / 1000) fit_model.fit_variogram(bin_center, gamma) # output ax = fit_model.plot(x_max=max(bin_center)) ax.scatter(bin_center, gamma) print(fit_model) ``` ![vario_fit_rescale](https://user-images.githubusercontent.com/19690642/111821844-6944c380-88e3-11eb-95c6-7d56de2be48f.png) ```bash Exponential(dim=2, var=1.61, len_scale=1.68, nugget=0.0509, anis=[1.0], angles=[0.0]) ``` As you see, `len_scale=1.68` could be estimated with the correct order of magnitude. Note, that the default init guess is always `1`. The rescaled length can be accessed by: ```python fit_model.len_rescaled ``` ```bash 1678.510771993176 ```	2021-03-21T13:49:55	[]	[]
meshpro/meshplex	meshpro__meshplex-147	3b8ed501b70ef6101e45e3aab9604b1add8cbe3f	diff --git a/meshplex/_mesh.py b/meshplex/_mesh.py index 67e159d..d2b5890 100644 --- a/meshplex/_mesh.py +++ b/meshplex/_mesh.py @@ -188,7 +188,7 @@ def cell_circumcenters(self): return self._circumcenters[-1] @property - def cell_circumradius(self): + def cell_circumradius(self) -> ArrayLike: """Get the circumradii of all cells""" if self._cell_circumradii is None: self._compute_cell_values() @@ -295,7 +295,7 @@ def signed_circumcenter_distances(self): return self._signed_circumcenter_distances - def _compute_cell_values(self, mask=slice(None)): + def _compute_cell_values(self, mask=None): """Computes the volumes of all edges, facets, cells etc. in the mesh. It starts off by computing the (squared) edge lengths, then complements the edge with one vertex to form face. It computes an orthogonal basis of the face (with modified @@ -303,6 +303,9 @@ def _compute_cell_values(self, mask=slice(None)): of the face is computed. Then, it complements again to form the 3-simplex, again forms an orthogonal basis with Gram-Schmidt, and so on. """ + if mask is None: + mask = slice(None) + e = self.points[self.idx[-1][..., mask]] e0 = e[0] diff = e[1] - e[0] @@ -316,11 +319,11 @@ def _compute_cell_values(self, mask=slice(None)): vv = _dot(diff, self.n - 1) circumradii2 = 0.25 * vv sqrt_vv = np.sqrt(vv) - lmbda = 0.5 * np.sqrt(vv) + lmbda = 0.5 * sqrt_vv sumx = np.array(e + circumcenters[-1]) - partitions = 0.5 * np.sqrt(np.array([vv, vv])) + partitions = 0.5 * np.array([sqrt_vv, sqrt_vv]) norms2 = np.array(volumes2) for kk, idx in enumerate(self.idx[:-1][::-1]): @@ -329,8 +332,12 @@ def _compute_cell_values(self, mask=slice(None)): p0 = self.points[idx][:, mask] v = p0 - e0 # modified gram-schmidt - for w, ww in zip(orthogonal_basis, norms2): - alpha = np.einsum("...k,...k->...", w, v) / ww + for w, w_dot_w in zip(orthogonal_basis, norms2): + w_dot_v = np.einsum("...k,...k->...", w, v) + # Compute <w, v> / <w, w>, but don't set the output value where w==0. + # The value remains uninitialized and gets canceled out in the next + # iteration when multiplied by w. + alpha = np.divide(w_dot_v, w_dot_w, where=w_dot_w > 0.0) v -= _multiply(w, alpha, self.n - 1 - kk) vv = np.einsum("...k,...k->...", v, v) @@ -353,22 +360,34 @@ def _compute_cell_values(self, mask=slice(None)): c = circumcenters[-1] p0c2 = _dot(p0 - c, self.n - 1 - kk) - # - sigma = 0.5 * (p0c2 - circumradii2) / vv - lmbda2 = sigma ** 2 * vv + # Be a bit careful here. sigma and lmbda can be negative. Also make sure + # that the values aren't nan when they should be inf (for degenerate + # simplices, i.e., vv == 0). + a = 0.5 * (p0c2 - circumradii2) + sqrt_vv = np.sqrt(vv) + with warnings.catch_warnings(): + # silence division-by-0 warnings + # Happens for degenerate cells (sqrt(vv) == 0), and this case is + # supported by meshplex. The values lmbda and sigma will just be +-inf. + warnings.simplefilter("ignore", category=RuntimeWarning) + lmbda = a / sqrt_vv + sigma_k0 = a[k0] / vv[k0] # circumcenter, squared circumradius # <https://math.stackexchange.com/a/4064749/36678> - # - circumradii2 = lmbda2[k0] + circumradii2[k0] - circumcenters.append(c[k0] + _multiply(v[k0], sigma[k0], self.n - 2 - kk)) + lmbda2_k0 = sigma_k0 * a[k0] + circumradii2 = lmbda2_k0 + circumradii2[k0] + with warnings.catch_warnings(): + # Similar as above: The multiplicattion `v * sigma` correctly produces + # nans for degenerate cells. + warnings.simplefilter("ignore", category=RuntimeWarning) + circumcenters.append( + c[k0] + _multiply(v[k0], sigma_k0, self.n - 2 - kk) + ) sumx += circumcenters[-1] # cell partitions - # don't use sqrt(lmbda2) here; lmbda, just like sigma, can be negative - sqrt_vv = np.sqrt(vv) - lmbda = sigma * sqrt_vv partitions *= lmbda / (kk + 2) # The integral of x, @@ -806,6 +825,8 @@ def remove_dangling_points(self): if self._is_point_used is not None: self._is_point_used = self._is_point_used[is_part_of_cell] + return np.sum(~is_part_of_cell) + @property def q_radius_ratio(self): """Ratio of incircle and circumcircle ratios times (n-1). ("Normalized shape diff --git a/meshplex/_mesh_tri.py b/meshplex/_mesh_tri.py index 4a6558f..6639985 100644 --- a/meshplex/_mesh_tri.py +++ b/meshplex/_mesh_tri.py @@ -1,5 +1,6 @@ import warnings +import npx import numpy as np from ._mesh import Mesh @@ -263,11 +264,15 @@ def flip_until_delaunay(self, tol=0.0, max_steps=100): if step > max_steps: m = np.min(self.signed_circumcenter_distances) - warnings.warn( + msg = ( f"Maximum number of edge flips reached ({max_steps}). " - f"Smallest signed circumcenter distance: {m:.3e}." + + f"Smallest signed circumcenter distance: {m:.3e}. " + + f"Try increasing the tolerance (currently {tol}) " + + f"or max_steps (currentframe {max_steps})." ) + warnings.warn(msg) break + # raise MeshplexError(msg) interior_facets_cells = self.facets_cells["interior"][1:3].T adj_cells = interior_facets_cells[is_flip_interior_facet].T @@ -291,8 +296,8 @@ def flip_until_delaunay(self, tol=0.0, max_steps=100): multiflip_cell_gids = cell_gids[num_flips_per_cell > 1] # actually perform the flips - self.flip_interior_facets(is_flip_interior_facet) - num_flips += np.sum(is_flip_interior_facet) + num = self.flip_interior_facets(is_flip_interior_facet) + num_flips += num # check the new signed_circumcenter_distances new_scd = self.signed_circumcenter_distances[is_flip_interior_facet] @@ -311,16 +316,47 @@ def flip_until_delaunay(self, tol=0.0, max_steps=100): ) message += "Leaving those facets as they are." warnings.warn(message) - # exit(1) # TODO remove + # Check which edges need to be flipped next. Don't flip edges which have + # just been flipped though. (This can happen due to round-off errors.) is_flip_interior_facet_old = is_flip_interior_facet.copy() is_flip_interior_facet = self.signed_circumcenter_distances < -tol - # Simply don't flip edges which have just been flipped is_flip_interior_facet[is_flip_interior_facet_old] = False return num_flips def flip_interior_facets(self, is_flip_interior_facet): + facets_cells_flip = self.facets_cells["interior"][:, is_flip_interior_facet] + # facet_gids = facets_cells_flip[0] + adj_cells = facets_cells_flip[1:3] + lids = facets_cells_flip[3:5] + + new_edges = np.array( + [ + self.cells("points")[adj_cells[0], lids[0]], + self.cells("points")[adj_cells[1], lids[1]], + ] + ).T + new_edges = np.sort(new_edges, axis=1) + do_actually_flip = np.ones(len(new_edges), dtype=bool) + + # Check if some flips would lead to the same flipped edge. This can happen, for + # example, in triangular shell meshes in 3D, and leads to weird behavior down + # the line. See <https://github.com/nschloe/meshplex/issues/130>. + _, inv = npx.unique_rows(new_edges, return_inverse=True) + is_unique = np.zeros(len(new_edges), dtype=bool) + is_unique[inv] = True + do_actually_flip &= is_unique + + # Check if expected new edges are already present in the mesh for the same + # reasons as above + already_exists = npx.isin_rows(new_edges, self.facets["points"]) + do_actually_flip &= ~already_exists + + # finally apply the filter + is_flip_interior_facet[is_flip_interior_facet] = do_actually_flip + + # now actuall perform the flip facets_cells_flip = self.facets_cells["interior"][:, is_flip_interior_facet] facet_gids = facets_cells_flip[0] adj_cells = facets_cells_flip[1:3] @@ -457,6 +493,8 @@ def flip_interior_facets(self, is_flip_interior_facet): self._update_cell_values(update_cell_ids, update_interior_facet_ids) + return np.sum(is_flip_interior_facet) + def _update_cell_values(self, cell_ids, interior_facet_ids): """Updates all sorts of cell information for the given cell IDs.""" # update idx diff --git a/setup.cfg b/setup.cfg index ca16de7..9ef63af 100644 --- a/setup.cfg +++ b/setup.cfg @@ -1,6 +1,6 @@ [metadata] name = meshplex -version = 0.16.0 +version = 0.16.1 author = Nico Schlömer author_email = [email protected] description = Fast tools for simplex meshes	mwe fails on 0.15.13 but works on 0.15.12 import meshplex import numpy as np cells = np.array( [ [4, 5, 2], [5, 6, 2], [8, 7, 6], [6, 7, 2], [4, 2, 1], [1, 2, 0], [2, 7, 3], [0, 2, 3], ] ) points = np.array( [ [4.00540750e00, 8.41882994e00, 1.87847398e-01], [4.15720337e00, 8.14281236e00, 9.12266346e-10], [4.30002833e00, 8.57424769e00, 1.04310192e-02], [4.05235481e00, 8.89348539e00, 2.06854700e-01], [4.25687935e00, 8.41702649e00, -1.80354934e-10], [4.28647665e00, 8.56681008e00, -3.27591539e-10], [4.28974575e00, 8.58639357e00, -1.84739328e-10], [4.36607782e00, 9.03478181e00, 2.94211691e-02], [4.33143378e00, 9.01150991e00, 2.88600116e-10], ] ) meshplex.MeshTri mesh = meshplex.MeshTri(points, cells) mesh.flip_until_delaunay() mesh = meshplex.MeshTri(mesh.points, mesh.cells["points"]) mesh.flip_until_delaunay() mesh.create_facets() flip_until_delaunay failure [trimesh_example.zip](https://github.com/nschloe/meshplex/files/6241541/trimesh_example.zip) Hi Nico, I get a problem again on the following mesh. Version 0.15.14 Output: ``` control_uniqueness faces are unique control_uniqueness [trimesh_example.zip](https://github.com/nschloe/meshplex/files/6241537/trimesh_example.zip) face doubled (181419, 182934, 182936) 2 face doubled (181419, 182933, 182936) 2 finished with errors ```	`flip_until_delaunay()` may (and does, in this case) create duplicate cells. You can remove them with `mesh.remove_duplicate_cells()` after the flip, then it should all work. It's not done automatically because there are cases where duplicates are appropriate. Right, `flip_until_delaunay()` might produce duplicate cells sometimes. Just call ```python mesh.remove_duplicate_cells() ``` This will take care of it. True I simplified the example too much. I wanted to say the following, iff I use flip_until_delaunay followed by remove_duplicate_cells I get a mesh with non-manifold edges: ```python import meshplex import numpy as np def unique_rows(a): # The numpy alternative `np.unique(a, axis=0)` is slow; cf. # <https://github.com/numpy/numpy/issues/11136>. b = np.ascontiguousarray(a).view(np.dtype((np.void, a.dtype.itemsize * a.shape[1]))) a_unique, inv, cts = np.unique(b, return_inverse=True, return_counts=True) a_unique = a_unique.view(a.dtype).reshape(-1, a.shape[1]) return a_unique, inv, cts def control_mesh(mesh): s = mesh.idx_hierarchy.shape a = np.sort(mesh.idx_hierarchy.reshape(s[0], -1).T) a_unique, inv, cts = unique_rows(a) if cts.max() > 2: print("control_mesh found non-manifold edges",cts.max()) for j in range(len(cts)): if cts[j] > 2: print(j, cts[j]) return False print("good_mesh") return True mesh = meshplex._reader.read("contour.obj") control_mesh(mesh) mesh.flip_until_delaunay() print("removing duplicate cells", mesh.remove_duplicate_cells()) #print("removing duplicate cells", mesh.remove_duplicate_cells()) control_mesh(mesh) ``` outputs: ``` good_mesh removing duplicate cells 2 control_mesh found non-manifold edges 4 433587 3 494068 3 641571 4 ``` Dear nico, were you able to look into this issue. Shall I try to reduce the mesh to a minimal example? Not yet. I'm hoping to get to it next week. dear nico, any news is it a hard problem? Not yet, no. Still busy with other things in meshplex, so not too far away from it. hi nico, is there any progress or information about this issue. I just need some information for my team. best I'm looking at it right now. You never know how deep the rabbit hole goes, but I'm hoping to get it done this afternoon. Let's kick it off with an MWE: ```python import numpy as np import npx import meshplex def control_mesh(mesh): s = mesh.idx_hierarchy.shape a = np.sort(mesh.idx_hierarchy.reshape(s[0], -1).T) _, cts = npx.unique_rows(a, return_counts=True) if cts.max() > 2: print("control_mesh found non-manifold edges", cts.max()) for idx in np.where(cts > 2)[0]: print(idx, cts[idx]) else: print("good_mesh") points = [ [0.011, 0.21, 0.69], [0.664, 0.70, 0.69], [0.664, 0.96, 1.43], [0.619, 0.96, 1.44], [0.664, 0.93, 1.44], [1.374, 0.21, 0.69], [0.670, 0.96, 1.44], ] cells = [[2, 0, 1], [2, 3, 0], [4, 0, 3], [4, 6, 5], [2, 5, 6]] mesh = meshplex.MeshTri(points, cells) control_mesh(mesh) print("flip") mesh.flip_until_delaunay() control_mesh(mesh) ``` ``` good_mesh flip control_mesh found non-manifold edges 4 5 4 ``` Isn’t this a round off error when evaluating the flip predicate? Flips are notorious for this... yes I remember robustness issues in geometric algorithms. offsetting 2d-curves in CAD (like rhino-cad) fails and with a random small translation they work again. with gpc I had issues with a probability of about 1e-4. When it failed I had to do some random translation. With polygon_clipper no robustness issues so far. Do you think random translation could help? import random def random_trans(points): eps = .0001 sx = epsrandom.random() sy = eps random.random() sz = eps * random.random() points = [(x + sx, y+ sy, z + sz) for x,y,z in points] return points Nope tried it does not help. I'm not sure how the code is compiled on the back end by your hardware, but in my experience IFort compiler enables unsafe math ops by default while GNU based compilers do not. Generally if we want to repeat work, we thus use GNU with safe math ops Here's an even simpler MWE: ```python import numpy as np import npx import meshplex def control_mesh(mesh): s = mesh.idx_hierarchy.shape a = np.sort(mesh.idx_hierarchy.reshape(s[0], -1).T) _, cts = npx.unique_rows(a, return_counts=True) if cts.max() > 2: print("control_mesh found non-manifold edges", cts.max()) for idx in np.where(cts > 2)[0]: print(idx, cts[idx]) else: print("good_mesh") points = [ [0.0, 0.0, 0.0], [1.0, 0.5, 0.0], [1.0, 0.76, 1.0], [1.0, 0.8, 1.0], [1.0, 0.7, 1.0], [1.0, 0.7, 0.0], ] cells = [[2, 0, 1], [2, 3, 0], [4, 0, 3], [4, 3, 5], [2, 5, 3]] mesh = meshplex.MeshTri(points, cells) mesh.save("out0.vtu") control_mesh(mesh) print("flip 1...") mesh.flip_until_delaunay(max_steps=1) mesh.save("out1.vtu") n = mesh.remove_duplicate_cells() print(f"{n = }") control_mesh(mesh) control_mesh(mesh) mesh = meshplex.MeshTri(mesh.points, mesh.cells("points")) mesh.save("out4.vtu") ``` ![screenshot](https://user-images.githubusercontent.com/181628/115889644-672ad300-a454-11eb-8dc9-e37431a7c1f1.png) So, this situation is a little crazy to start out with. It's not too strange that the edge flipper produces non-manifold edges. what are the options. Do you think it is better to find some preprocessing step or a postprocessor to solve the non-manifold edges?	2021-04-24T23:30:53	[]	[]
jquast/blessed	jquast__blessed-260	a34c6b1869b4dd467c6d1ab6895872bb72db7e0f	diff --git a/blessed/keyboard.py b/blessed/keyboard.py index 31cc98c..401ec45 100644 --- a/blessed/keyboard.py +++ b/blessed/keyboard.py @@ -1,6 +1,7 @@ """Sub-module providing 'keyboard awareness'.""" # std imports +import os import re import time import platform @@ -448,4 +449,28 @@ def _read_until(term, pattern, timeout): ('KEY_BEGIN', curses.KEY_BEG), ) +#: Default delay, in seconds, of Escape key detection in +#: :meth:`Terminal.inkey`.` curses has a default delay of 1000ms (1 second) for +#: escape sequences. This is too long for modern applications, so we set it to +#: 350ms, or 0.35 seconds. It is still a bit conservative, for remote telnet or +#: ssh servers, for example. +DEFAULT_ESCDELAY = 0.35 + + +def _reinit_escdelay(): + # pylint: disable=W0603 + # Using the global statement: this is necessary to + # allow test coverage without complex module reload + global DEFAULT_ESCDELAY + if os.environ.get('ESCDELAY'): + try: + DEFAULT_ESCDELAY = int(os.environ['ESCDELAY']) / 1000.0 + except ValueError: + # invalid values of 'ESCDELAY' are ignored + pass + + +_reinit_escdelay() + + __all__ = ('Keystroke', 'get_keyboard_codes', 'get_keyboard_sequences',) diff --git a/blessed/terminal.py b/blessed/terminal.py index 76214e9..fc0b0c4 100644 --- a/blessed/terminal.py +++ b/blessed/terminal.py @@ -18,7 +18,8 @@ # local from .color import COLOR_DISTANCE_ALGORITHMS -from .keyboard import (_time_left, +from .keyboard import (DEFAULT_ESCDELAY, + _time_left, _read_until, resolve_sequence, get_keyboard_codes, @@ -1425,8 +1426,8 @@ def keypad(self): self.stream.write(self.rmkx) self.stream.flush() - def inkey(self, timeout=None, esc_delay=0.35): - """ + def inkey(self, timeout=None, esc_delay=DEFAULT_ESCDELAY): + r""" Read and return the next keyboard event within given timeout. Generally, this should be used inside the :meth:`raw` context manager. @@ -1434,12 +1435,16 @@ def inkey(self, timeout=None, esc_delay=0.35): :arg float timeout: Number of seconds to wait for a keystroke before returning. When ``None`` (default), this method may block indefinitely. - :arg float esc_delay: To distinguish between the keystroke of - ``KEY_ESCAPE``, and sequences beginning with escape, the parameter - ``esc_delay`` specifies the amount of time after receiving escape - (``chr(27)``) to seek for the completion of an application key - before returning a :class:`~.Keystroke` instance for - ``KEY_ESCAPE``. + :arg float esc_delay: Time in seconds to block after Escape key + is received to await another key sequence beginning with + escape such as KEY_LEFT, sequence ``'\x1b[D'``], before returning a + :class:`~.Keystroke` instance for ``KEY_ESCAPE``. + + Users may override the default value of ``esc_delay`` in seconds, + using environment value of ``ESCDELAY`` as milliseconds, see + `ncurses(3)`_ section labeled ESCDELAY for details. Setting + the value as an argument to this function will override any + such preference. :rtype: :class:`~.Keystroke`. :returns: :class:`~.Keystroke`, which may be empty (``u''``) if ``timeout`` is specified and keystroke is not received. @@ -1454,11 +1459,12 @@ def inkey(self, timeout=None, esc_delay=0.35): <https://docs.microsoft.com/en-us/windows/win32/api/timeapi/nf-timeapi-timebeginperiod>`_. Decreasing the time resolution will reduce this to 10 ms, while increasing it, which is rarely done, will have a perceptable impact on the behavior. + + _`ncurses(3)`: https://www.man7.org/linux/man-pages/man3/ncurses.3x.html """ resolve = functools.partial(resolve_sequence, mapper=self._keymap, codes=self._keycodes) - stime = time.time() # re-buffer previously received keystrokes, diff --git a/tox.ini b/tox.ini index 017f354..d3a4405 100644 --- a/tox.ini +++ b/tox.ini @@ -67,8 +67,9 @@ commands = autopep8 --in-place --recursive --aggressive --aggressive blessed/ bin/ setup.py [testenv:docformatter] +# docformatter pinned due to https://github.com/PyCQA/docformatter/issues/264 deps = - docformatter + docformatter<1.7.4 untokenize commands = docformatter \ @@ -83,8 +84,9 @@ commands = {toxinidir}/docs/conf.py [testenv:docformatter_check] +# docformatter pinned due to https://github.com/PyCQA/docformatter/issues/264 deps = - docformatter + docformatter<1.7.4 untokenize commands = docformatter \	Conditionally set default esc_delay of inkey() from environment from https://github.com/jquast/blessed/issues/156#issuecomment-600790544 - We should honor this lesser known `ESCDELAY` environment value as the default value of `Terminal.inkey(esc_delay)` if set, https://stackoverflow.com/a/28020568 and https://invisible-island.net/ncurses/man/ncurses.3x.html > Specifies the total time, in milliseconds, for which ncurses will await a character sequence, e.g., a function key. The default value, 1000		2023-10-31T00:01:54	[]	[]
dftbplus/skpar	dftbplus__skpar-7	8d6ac606fd8ddeb16e0bc8ca0a0a742d8d829ecd	diff --git a/skpar/core/input.py b/skpar/core/input.py index 16887d3..e72cd42 100644 --- a/skpar/core/input.py +++ b/skpar/core/input.py @@ -29,7 +29,7 @@ def get_input(filename): else: #if intype == 'yaml': try: - spec = yaml.load(infile) + spec = yaml.safe_load(infile) except yaml.YAMLError: LOGGER.warning('Input not a valid YAML') raise	Fix yaml loader error by using safe_load Invoking skpar currently results in a yaml parsing error, that is already fixed by f0a11be3600eb06320cc4325b4ddd840059d5299 and could be merged.		2022-09-20T09:06:59	[]	[]
oasis-open/cti-taxii-server	oasis-open__cti-taxii-server-142	6fb6f9ddb50fe7fd5348a8ba84da8f421555d56a	diff --git a/medallion/__init__.py b/medallion/__init__.py index 215fb1bd..6d016880 100644 --- a/medallion/__init__.py +++ b/medallion/__init__.py @@ -18,7 +18,9 @@ log = logging.getLogger(__name__) log.addHandler(ch) + application_instance = Flask(__name__) +application_instance.app_context().push() auth = HTTPBasicAuth() @@ -30,7 +32,6 @@ def load_app(config_file): set_config(application_instance, "taxii", configuration) set_config(application_instance, "backend", configuration) register_blueprints(application_instance) - return application_instance @@ -46,11 +47,14 @@ def set_config(flask_application_instance, prop_name, config): try: flask_application_instance.users_backend = config[prop_name] except KeyError: - log.warning("You did not give user information in your config.") - log.warning("We are giving you the default user information of:") - log.warning("User = user") - log.warning("Pass = pass") - flask_application_instance.users_backend = {"user": "pass"} + if config.get('no_auth'): + flask_application_instance.config['no_auth'] = True + else: + log.warning("You did not give user information in your config.") + log.warning("We are giving you the default user information of:") + log.warning("User = user") + log.warning("Pass = pass") + flask_application_instance.users_backend = {"user": "pass"} elif prop_name == "backend": try: flask_application_instance.medallion_backend = connect_to_backend(config[prop_name]) diff --git a/medallion/scripts/run.py b/medallion/scripts/run.py index 93b42d9c..29358e47 100644 --- a/medallion/scripts/run.py +++ b/medallion/scripts/run.py @@ -57,6 +57,13 @@ def _get_argparser(): choices=["DEBUG", "INFO", "WARN", "ERROR", "CRITICAL"], ) + parser.add_argument( + "--no-auth", + default=False, + action="store_true", + help="Turn off credential checking for all endpoints.", + ) + parser.add_argument( "CONFIG_PATH", metavar="CONFIG_PATH", @@ -75,6 +82,9 @@ def main(): with open(medallion_args.CONFIG_PATH, "r") as f: configuration = json.load(f) + if medallion_args.no_auth: + configuration['no_auth'] = True + set_config(application_instance, "users", configuration) set_config(application_instance, "taxii", configuration) set_config(application_instance, "backend", configuration) diff --git a/medallion/views/__init__.py b/medallion/views/__init__.py index aee3adfb..47aeac15 100644 --- a/medallion/views/__init__.py +++ b/medallion/views/__init__.py @@ -1,6 +1,7 @@ import re from flask import request +from functools import wraps from ..exceptions import ProcessingError @@ -25,3 +26,15 @@ def validate_version_parameter_in_accept_header(): if found is False: raise ProcessingError("Media type in the Accept header is invalid or not found", 406) + + +def conditional_auth(func): + from .. import auth, current_app + + @wraps(func) + def wrapper(): + if current_app.config.get('no_auth'): + return func + else: + return auth.login_required(func) + return wrapper() diff --git a/medallion/views/collections.py b/medallion/views/collections.py index 702428a4..b1625c2a 100644 --- a/medallion/views/collections.py +++ b/medallion/views/collections.py @@ -1,7 +1,6 @@ from flask import Blueprint, Response, current_app, json -from . import MEDIA_TYPE_TAXII_V21, validate_version_parameter_in_accept_header -from .. import auth +from . import MEDIA_TYPE_TAXII_V21, validate_version_parameter_in_accept_header, conditional_auth from .discovery import api_root_exists from .objects import collection_exists @@ -9,7 +8,7 @@ @collections_bp.route("/<string:api_root>/collections/", methods=["GET"]) [email protected]_required +@conditional_auth def get_collections(api_root): """ Defines TAXII API - Collections: @@ -36,7 +35,7 @@ def get_collections(api_root): @collections_bp.route("/<string:api_root>/collections/<string:collection_id>/", methods=["GET"]) [email protected]_required +@conditional_auth def get_collection(api_root, collection_id): """ Defines TAXII API - Collections: diff --git a/medallion/views/discovery.py b/medallion/views/discovery.py index 2fea0c25..08d4d21e 100644 --- a/medallion/views/discovery.py +++ b/medallion/views/discovery.py @@ -1,7 +1,6 @@ from flask import Blueprint, Response, current_app, json -from . import MEDIA_TYPE_TAXII_V21, validate_version_parameter_in_accept_header -from .. import auth +from . import MEDIA_TYPE_TAXII_V21, validate_version_parameter_in_accept_header, conditional_auth from ..exceptions import ProcessingError discovery_bp = Blueprint("discovery", __name__) @@ -14,7 +13,7 @@ def api_root_exists(api_root): @discovery_bp.route("/taxii2/", methods=["GET"]) [email protected]_required +@conditional_auth def get_server_discovery(): """ Defines TAXII API - Server Information: @@ -30,7 +29,6 @@ def get_server_discovery(): # depending upon the credentials. validate_version_parameter_in_accept_header() server_discovery = current_app.medallion_backend.server_discovery() - if server_discovery: return Response( response=json.dumps(server_discovery), @@ -41,7 +39,7 @@ def get_server_discovery(): @discovery_bp.route("/<string:api_root>/", methods=["GET"]) [email protected]_required +@conditional_auth def get_api_root_information(api_root): """ Defines TAXII API - Server Information: @@ -67,7 +65,7 @@ def get_api_root_information(api_root): @discovery_bp.route("/<string:api_root>/status/<string:status_id>/", methods=["GET"]) [email protected]_required +@conditional_auth def get_status(api_root, status_id): """ Defines TAXII API - Server Information: diff --git a/medallion/views/manifest.py b/medallion/views/manifest.py index 35367904..cb869e4e 100644 --- a/medallion/views/manifest.py +++ b/medallion/views/manifest.py @@ -1,7 +1,6 @@ from flask import Blueprint, Response, current_app, json, request -from . import MEDIA_TYPE_TAXII_V21, validate_version_parameter_in_accept_header -from .. import auth +from . import MEDIA_TYPE_TAXII_V21, validate_version_parameter_in_accept_header, conditional_auth from .discovery import api_root_exists from .objects import ( collection_exists, permission_to_read, validate_limit_parameter @@ -11,7 +10,7 @@ @manifest_bp.route("/<string:api_root>/collections/<string:collection_id>/manifest/", methods=["GET"]) [email protected]_required +@conditional_auth def get_object_manifest(api_root, collection_id): """ Defines TAXII API - Collections: diff --git a/medallion/views/objects.py b/medallion/views/objects.py index 1d30c7c8..f796b423 100644 --- a/medallion/views/objects.py +++ b/medallion/views/objects.py @@ -3,8 +3,7 @@ from flask import Blueprint, Response, current_app, json, request -from . import MEDIA_TYPE_TAXII_V21, validate_version_parameter_in_accept_header -from .. import auth +from . import MEDIA_TYPE_TAXII_V21, validate_version_parameter_in_accept_header, conditional_auth from ..common import get_timestamp from ..exceptions import ProcessingError from .discovery import api_root_exists @@ -86,7 +85,7 @@ def validate_limit_parameter(): @objects_bp.route("/<string:api_root>/collections/<string:collection_id>/objects/", methods=["GET", "POST"]) [email protected]_required +@conditional_auth def get_or_add_objects(api_root, collection_id): """ Defines TAXII API - Collections: @@ -138,7 +137,7 @@ def get_or_add_objects(api_root, collection_id): @objects_bp.route("/<string:api_root>/collections/<string:collection_id>/objects/<string:object_id>/", methods=["GET", "DELETE"]) [email protected]_required +@conditional_auth def get_or_delete_object(api_root, collection_id, object_id): """ Defines TAXII API - Collections: @@ -187,7 +186,7 @@ def get_or_delete_object(api_root, collection_id, object_id): @objects_bp.route("/<string:api_root>/collections/<string:collection_id>/objects/<string:object_id>/versions/", methods=["GET"]) [email protected]_required +@conditional_auth def get_object_versions(api_root, collection_id, object_id): """ Defines TAXII API - Collections: Get Object Versions section	bug fix del _id key	# [Codecov](https://codecov.io/gh/oasis-open/cti-taxii-server/pull/4?src=pr&el=h1) Report > Merging [#4](https://codecov.io/gh/oasis-open/cti-taxii-server/pull/4?src=pr&el=desc) into [master](https://codecov.io/gh/oasis-open/cti-taxii-server/commit/eaa4eb69feafa2492eccc26d6d3743a4e87f94de?src=pr&el=desc) will not change coverage. > The diff coverage is `n/a`. [![Impacted file tree graph](https://codecov.io/gh/oasis-open/cti-taxii-server/pull/4/graphs/tree.svg?height=150&width=650&token=eu7JYJfW12&src=pr)](https://codecov.io/gh/oasis-open/cti-taxii-server/pull/4?src=pr&el=tree) ```diff @@ Coverage Diff @@ ## master #4 +/- ## ======================================= Coverage 14.99% 14.99% ======================================= Files 14 14 Lines 587 587 ======================================= Hits 88 88 Misses 499 499 ``` ------ [Continue to review full report at Codecov](https://codecov.io/gh/oasis-open/cti-taxii-server/pull/4?src=pr&el=continue). > Legend - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta) > `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data` > Powered by [Codecov](https://codecov.io/gh/oasis-open/cti-taxii-server/pull/4?src=pr&el=footer). Last update [eaa4eb6...2bbbc9e](https://codecov.io/gh/oasis-open/cti-taxii-server/pull/4?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments). This is a good fix - @gtback, should we merge in?? Thanks, @zhangymJLU . Before we can merge this, you'll need to sign the Individual Contributor License Agreement (CLA) located [here](https://www.oasis-open.org/resources/open-repositories/cla). If this is a problem, let us know. OK, I have accepted the license on this page and replay the confirm email now, @gtback. @zhangymJLU (Yangming Zhang) - Many thanks, and your CLA is now on file. -rcc	2021-03-15T13:50:26	[]	[]
oasis-open/cti-taxii-server	oasis-open__cti-taxii-server-136	d1367f579eaac1b7b311a63da9df599afe6a9c3e	diff --git a/README.rst b/README.rst index 9eb60c60..c1d73dcc 100644 --- a/README.rst +++ b/README.rst @@ -52,17 +52,17 @@ Medallion provides a command-line interface to start the TAXII Server .. code-block:: text - usage: medallion [-h] - [--host HOST] - [--port PORT] - [--debug-mode] - [--log-level {DEBUG,INFO,WARN,ERROR,CRITICAL}] - CONFIG_PATH + usage: medallion [-h] [--host HOST] [--port PORT] [--debug-mode] + [--log-level {DEBUG,INFO,WARN,ERROR,CRITICAL}] + [-c CONF_FILE] [--conf-dir CONF_DIR \| --no-conf-dir] + [--conf-check] [CONFIG_PATH] medallion v3.0.0 positional arguments: - CONFIG_PATH The location of the JSON configuration file to use. + CONFIG_PATH Deprecated argument for specifying a single + JSON configuration file. Do not specify this + and `--conf-file`. optional arguments: -h, --help show this help message and exit @@ -76,17 +76,35 @@ Medallion provides a command-line interface to start the TAXII Server --log-level {DEBUG,INFO,WARN,ERROR,CRITICAL} The logging output level for medallion. + -c CONF_FILE, --conf-file CONF_FILE + Path to a single configuration file. Defaults to the + value of the MEDALLION_CONFFILE environment variable + or /etc/xdg/medallion/3/medallion.conf. + + --conf-dir CONF_DIR Path to a directory containing JSON configuration + files with names ending in .json or .conf. Defaults to + the value of the MEDALLION_CONFDIR environment + variable or /etc/xdg/medallion/3/config.d. + + --no-conf-dir Disable the use of any configuration directory as + described for --conf-dir. This may be used to ensure + that the default or some value from the environment is + not used. + + --conf-check Evaluate medallion configuration without running the + server. + To run medallion .. code-block:: bash - $ python medallion/scripts/run.py <config-file> + $ python medallion/scripts/run.py --conf-file <config-file> Make sure medallion is using the same port that your TAXII client will be connecting on. You can specify which port medallion runs on using the `--port` option, for example .. code-block:: bash - $ medallion --port 80 config_file.json + $ medallion --port 80 --conf-file config_file.json The <config_file> contains: @@ -163,7 +181,7 @@ We also provide a Docker image to make it easier to run medallion .. code-block:: bash - $ docker build . -t medallion + $ docker build . -t medallion -f docker_utils/Dockerfile If operating behind a proxy, add the following option (replacing `<proxy>` with your proxy location and port): ``--build-arg https_proxy=<proxy>``. diff --git a/docker-compose.yml b/docker-compose.yml index 64c0fa65..d4139d18 100644 --- a/docker-compose.yml +++ b/docker-compose.yml @@ -7,7 +7,10 @@ services: build: context: . dockerfile: docker_utils/Dockerfile - command: [sh, -c, "medallion --host 0.0.0.0 docker_config.json"] + command: [sh, -c, "medallion --host 0.0.0.0"] + env: + MEDALLION_BACKEND_MODULE_CLASS: "MongoBackend" + MEDALLION_BACKEND_MONGO_URI: "mongodb://root:example@mongo:27017/" ports: - 5000:5000 mongo: diff --git a/docker_utils/Dockerfile b/docker_utils/Dockerfile index 62300d70..d748b001 100644 --- a/docker_utils/Dockerfile +++ b/docker_utils/Dockerfile @@ -1,8 +1,15 @@ FROM python:alpine COPY . /opt/taxii -COPY ./docker_utils/.json /opt/taxii WORKDIR /opt/taxii RUN pip install --upgrade pip setuptools \ && pip install pymongo \ && pip install . + +# Set up the default configuration files +ARG MEDALLION_CONFFILE=/opt/taxii/medallion.conf +ENV MEDALLION_CONFFILE "${MEDALLION_CONFFILE}" +COPY ./docker_utils/base_config.json "${MEDALLION_CONFFILE}" +ARG MEDALLION_CONFDIR=/opt/taxii/medallion.d/ +ENV MEDALLION_CONFDIR "${MEDALLION_CONFDIR}" +COPY ./docker_utils/config.d/ "${MEDALLION_CONFDIR}/" diff --git a/docker_utils/base_config.json b/docker_utils/base_config.json new file mode 100644 index 00000000..e0fee00d --- /dev/null +++ b/docker_utils/base_config.json @@ -0,0 +1,5 @@ +{ + "taxii": { + "max_page_size": 100 + } +} diff --git a/docker_utils/config.d/backend.json b/docker_utils/config.d/backend.json new file mode 100644 index 00000000..52962eac --- /dev/null +++ b/docker_utils/config.d/backend.json @@ -0,0 +1,6 @@ +{ + "backend": { + "module_class": "MongoBackend", + "uri": "mongodb://DUMMYUSER:DUMMYPASSWORD@mongo:27017/" + } +} diff --git a/docker_utils/config.d/users.json b/docker_utils/config.d/users.json new file mode 100644 index 00000000..c75ae2f5 --- /dev/null +++ b/docker_utils/config.d/users.json @@ -0,0 +1,7 @@ +{ + "users": { + "admin": "Password0", + "user1": "Password1", + "user2": "Password2" + } +} diff --git a/docker_utils/docker_config.json b/docker_utils/docker_config.json deleted file mode 100644 index 80d47013..00000000 --- a/docker_utils/docker_config.json +++ /dev/null @@ -1,14 +0,0 @@ -{ - "backend": { - "module_class": "MongoBackend", - "uri": "mongodb://root:example@mongo:27017/" - }, - "users": { - "admin": "Password0", - "user1": "Password1", - "user2": "Password2" - }, - "taxii": { - "max_page_size": 100 - } -} diff --git a/medallion/backends/__init__.py b/medallion/backends/__init__.py index 10141298..1f393832 100644 --- a/medallion/backends/__init__.py +++ b/medallion/backends/__init__.py @@ -4,6 +4,7 @@ import pkgutil import sys +import environ import pkg_resources from . import base @@ -27,3 +28,17 @@ ep_obj = ep.load() if inspect.isclass(ep_obj): base.BackendRegistry.register(ep.name, ep_obj) + + +# Finally we define a top-level environ config class which includes and configs +# defined in the registered backends [email protected](prefix="BACKEND") +class BackendConfig(object): + for name, clsobj in base.BackendRegistry.iter_(): + # We have to use a magic attribute name here since `config`s don't have + # a specific mixin or type we can check for + try: + locals()[name] = environ.group(clsobj.Config, optional=True) + except AttributeError: + pass + module_class = environ.var(None) diff --git a/medallion/backends/base.py b/medallion/backends/base.py index 49ed4662..2cfae6d6 100644 --- a/medallion/backends/base.py +++ b/medallion/backends/base.py @@ -19,6 +19,10 @@ def register(mcls, kind, clsobj): def get(mcls, kind): return mcls.__SUBCLASS_MAP[kind] + @classmethod + def iter_(mcls): + yield from mcls.__SUBCLASS_MAP.items() + class Backend(object, metaclass=BackendRegistry): diff --git a/medallion/backends/memory_backend.py b/medallion/backends/memory_backend.py index 9fef670b..c844f47a 100644 --- a/medallion/backends/memory_backend.py +++ b/medallion/backends/memory_backend.py @@ -4,6 +4,7 @@ import os import uuid +import environ from six import string_types from ..common import ( @@ -36,6 +37,10 @@ class MemoryBackend(Backend): # access control is handled at the views level + @environ.config(prefix="MEMORY") + class Config(object): + filename = environ.var(None) + def __init__(self, kwargs): # Refuse to run under a WSGI server since this is an internal backend if ( diff --git a/medallion/backends/mongodb_backend.py b/medallion/backends/mongodb_backend.py index 3413c7a8..00d44c88 100644 --- a/medallion/backends/mongodb_backend.py +++ b/medallion/backends/mongodb_backend.py @@ -1,6 +1,7 @@ import logging import uuid +import environ from pymongo import MongoClient from pymongo.errors import ConnectionFailure, ServerSelectionTimeoutError @@ -35,6 +36,10 @@ class MongoBackend(Backend): # access control is handled at the views level + @environ.config(prefix="MONGO") + class Config(object): + uri = environ.var() + def __init__(self, kwargs): try: self.client = MongoClient(kwargs.get("uri")) diff --git a/medallion/config.py b/medallion/config.py new file mode 100644 index 00000000..af2c2ab0 --- /dev/null +++ b/medallion/config.py @@ -0,0 +1,172 @@ +import collections.abc +import json +import logging +import pathlib + +import appdirs +import attr +import environ +import jsonmerge +import packaging.version + +from . import backends +from . import version as _medallion_version + +# Find a config directory for the closest compatible version including the +# current one, or fall back to the current version's one if none are extant +_curr_version = packaging.version.Version(_medallion_version.__version__) +_compat_versions = {_curr_version.major, 3, 2} +for _cand_version in sorted(_compat_versions, reverse=True): + _appdirs = appdirs.AppDirs( + "medallion", "oasis-open", version=str(_cand_version) + ) + _default_conf_p = pathlib.Path(_appdirs.site_config_dir).resolve() + if _default_conf_p.is_dir(): + break +else: + _appdirs = appdirs.AppDirs( + "medallion", "oasis-open", version=str(_curr_version.major), + ) + _default_conf_p = pathlib.Path(_appdirs.site_config_dir).resolve() +# Set the default config file and config.d directory paths +DEFAULT_CONFFILE = _default_conf_p / "medallion.conf" +DEFAULT_CONFDIR = _default_conf_p / "config.d" +# Configuration directories are scanned for files matching these suffixes +CONFDIR_SUFFIXES = {".json", ".conf"} + +LOGGER = logging.getLogger(__name__) + + +class _LazyJSONDumper(object): + """An lazy stringifier for dumping objects as JSON, e.g. for logging.""" + def __init__(self, obj, cls=json.JSONEncoder, indent=None): + super(_LazyJSONDumper, self).__init__() + # Beware: If `obj` is mutable then we'll get future mutations. This + # isn't really an issue for logging format evaluation, but it might be + # in other contexts/multi-thread. We concretise on the first call to + # `__str__()` so the output's immutable/repeatable after that at least. + self._obj = obj + self._dumpcls = cls + self._indent = indent + + def __str__(self): + # The first stringification will concretise `obj` + if not isinstance(self._obj, str): + self._obj = json.dumps( + self._obj, cls=self._dumpcls, indent=self._indent, + ) + return self._obj + + def __repr__(self): # pragma: no cover + return repr(str(self)) + + [email protected](prefix="TAXII") +class TAXIIConfig(object): + max_page_size = environ.var(None, converter=lambda i: int(i) if i else i) + + [email protected](prefix="MEDALLION") +class MedallionConfig(object): + backend = environ.group(backends.BackendConfig) + taxii = environ.group(TAXIIConfig) + + @classmethod + def __strip(cls, dict_): + for k, v in tuple(dict_.items()): + if isinstance(v, dict): + cls.__strip(v) + if v is None or v == {}: + del dict_[k] + + def as_dict(self): + dictified = attr.asdict(self) + self.__strip(dictified) + return dictified + + +def _load_config_file(conf_file_p): + try: + config_data = json.load(conf_file_p.open()) + except json.decoder.JSONDecodeError as exc: + raise ValueError(f"Invalid JSON data in {conf_file_p}") from exc + if not isinstance(config_data, collections.abc.Mapping): + raise TypeError(f"{conf_file_p} must contain a JSON object") + return config_data + + +def load_config(conf_file=DEFAULT_CONFFILE, conf_dir=DEFAULT_CONFDIR): + LOGGER.debug( + "Attempting to load configuration from '%s' and '%s'", + conf_file, conf_dir, + ) + conf_files = list() + # Sanity check that both the `conf_file` and `conf_dir` exist by resolving + # them with strictness determined by whether they are the defaults or not + if conf_file is not None: + conf_file_p = pathlib.Path(conf_file).resolve( + strict=conf_file is not DEFAULT_CONFFILE + ) + conf_files.append(conf_file_p) + # Build a lexicographically sorted list of children of the `conf_dir` which + # are both files and have an acceptable file suffix. We allow the potential + # `NotADirectoryError` here to bubble up since `DEFAULT_CONFDIR` existing + # but not being a directory seems like a problem. There's an option for + # disabling the use of a config dir so we won't feel bad about doing this. + if conf_dir is not None: + conf_dir_p = pathlib.Path(conf_dir).resolve( + strict=conf_dir is not DEFAULT_CONFDIR + ) + conf_dir_files = ( + p for p in conf_dir_p.iterdir() if p.suffix in CONFDIR_SUFFIXES + ) + try: + conf_files.extend(sorted(conf_dir_files, key=lambda p: p.name)) + except FileNotFoundError: + pass + LOGGER.debug( + "Configuration files to load in order: %s", + ", ".join(repr(str(p)) for p in conf_files), + ) + # Start with an empty config and progressively merge in data from files + config_data = dict() + for conf_file_p in conf_files: + try: + new_data = _load_config_file(conf_file_p) + except (FileNotFoundError, IsADirectoryError): + # Skip missing files since we haven't already exploded which means + # they're probably not important enough to care about (or TOCTOU?). + # We also don't care about sub-directories of the config dir. + pass + else: + LOGGER.debug( + "Configuration data from '%s' to be merged: %s", + conf_file_p, _LazyJSONDumper(new_data, indent=2), + ) + config_data = jsonmerge.merge(config_data, new_data) + # Load extra configuration from the environment + env_config = MedallionConfig.from_environ().as_dict() + LOGGER.debug( + "Configuration data from environment to be merged: %s", + _LazyJSONDumper(env_config, indent=2), + ) + config_data = jsonmerge.merge(config_data, env_config) + # We promote config pulled from the environment for the specified backend + # `module_class` since the `environ-config` variable layout will nest it. + # Any of the following statements could `KeyError` if the configuration is + # incomplete or no config was sourced from the environment, but we're + # lenient about this here in favour of validating in the main app logic. + try: + backend_config = config_data["backend"] + backend_kind = backend_config["module_class"] + backend_kind_config = backend_config.pop(backend_kind) + except KeyError: + pass + else: + backend_config.update(backend_kind_config) + # Return the finalised configuration dictionary + LOGGER.debug( + "Merged configuration: %s", + _LazyJSONDumper(config_data, indent=2), + ) + return config_data diff --git a/medallion/scripts/run.py b/medallion/scripts/run.py index 93b42d9c..152e5027 100644 --- a/medallion/scripts/run.py +++ b/medallion/scripts/run.py @@ -1,11 +1,13 @@ import argparse -import json +import inspect import logging +import os import textwrap from medallion import ( __version__, application_instance, register_blueprints, set_config ) +import medallion.config log = logging.getLogger("medallion") @@ -57,11 +59,53 @@ def _get_argparser(): choices=["DEBUG", "INFO", "WARN", "ERROR", "CRITICAL"], ) - parser.add_argument( + config_path_group = parser.add_mutually_exclusive_group() + config_path_group.add_argument( "CONFIG_PATH", metavar="CONFIG_PATH", + nargs="?", type=str, - help="The location of the JSON configuration file to use.", + help=inspect.cleandoc(""" + Deprecated argument for specifying a single JSON configuration + file. Do not specify this and `--conf-file`. + """), + ) + config_path_group.add_argument( + "-c", "--conf-file", + default=os.environ.get( + "MEDALLION_CONFFILE", medallion.config.DEFAULT_CONFFILE + ), + help=inspect.cleandoc(f""" + Path to a single configuration file. Defaults to the value of + the MEDALLION_CONFFILE environment variable or + {medallion.config.DEFAULT_CONFFILE}. + """), + ) + config_dir_group = parser.add_mutually_exclusive_group() + config_dir_group.add_argument( + "--conf-dir", + default=os.environ.get( + "MEDALLION_CONFDIR", medallion.config.DEFAULT_CONFDIR + ), + help=inspect.cleandoc(f""" + Path to a directory containing JSON configuration files with names + ending in .json or .conf. Defaults to the value of the + MEDALLION_CONFDIR environment variable or + {medallion.config.DEFAULT_CONFDIR}. + """), + ) + config_dir_group.add_argument( + "--no-conf-dir", + action="store_true", + help=inspect.cleandoc(""" + Disable the use of any configuration directory as described for + --conf-dir. This may be used to ensure that the default or some + value from the environment is not used. + """), + ) + parser.add_argument( + "--conf-check", action="store_true", + help="Evaluate medallion configuration without running the server.", ) return parser @@ -70,21 +114,27 @@ def _get_argparser(): def main(): medallion_parser = _get_argparser() medallion_args = medallion_parser.parse_args() + # Configuration checking sets up debug logging and does not run the app + if medallion_args.conf_check: + medallion_args.log_level = logging.DEBUG log.setLevel(medallion_args.log_level) - with open(medallion_args.CONFIG_PATH, "r") as f: - configuration = json.load(f) + configuration = medallion.config.load_config( + medallion_args.CONFIG_PATH or medallion_args.conf_file, + medallion_args.conf_dir if not medallion_args.no_conf_dir else None, + ) set_config(application_instance, "users", configuration) set_config(application_instance, "taxii", configuration) set_config(application_instance, "backend", configuration) register_blueprints(application_instance) - application_instance.run( - host=medallion_args.host, - port=medallion_args.port, - debug=medallion_args.debug_mode, - ) + if not medallion_args.conf_check: + application_instance.run( + host=medallion_args.host, + port=medallion_args.port, + debug=medallion_args.debug_mode, + ) if __name__ == "__main__": diff --git a/setup.py b/setup.py index 78c28604..2bceede0 100644 --- a/setup.py +++ b/setup.py @@ -48,8 +48,12 @@ def get_long_description(): keywords="taxii taxii2 server json cti cyber threat intelligence", packages=find_packages(exclude=[".test", "*.test.data"]), install_requires=[ + "appdirs>=1.4.4", + "environ-config>=21.1", "flask>=0.12.1", "Flask-HTTPAuth", + "jsonmerge", + "packaging", "pytz", "six", ], @@ -63,6 +67,7 @@ def get_long_description(): "coverage", "pytest", "pytest-cov", + "pytest-subtests", "tox", ], "docs": [ diff --git a/tox.ini b/tox.ini index e25a3608..fb9be89e 100644 --- a/tox.ini +++ b/tox.ini @@ -7,6 +7,7 @@ deps = tox pytest pytest-cov + pytest-subtests coverage responses pymongo	[RFC] Rework medallion configuration At the moment we pass a config file path to `medallion.load_app()` which is a bit of a pain and makes it a bit annoying to use a WSGI server. It would also be nice to move away from having a single config file to supporting something more like a conf.d structure with multiple JSON documents, as well as sourcing some configuration from the environment (e.g. 12 factor app style secrets). I've been looking into `environ-config` [0] which looks reasonable but might make the arbitrary kwarg passing currently done for backends painful. I've also been looking at `jsonmerge` [1] which should, in theory be capable of doing an ordered merge of documents in a medallion config.d directory. Would the following configuration hunting flow be sensible? Are there any other cases which should be handled? ![config-flow](https://user-images.githubusercontent.com/58414429/93841428-e7c6fc80-fcd6-11ea-97bc-3e44f2193054.png) [0] https://github.com/hynek/environ-config [1] https://github.com/avian2/jsonmerge		2020-09-23T02:04:00	[]	[]
tomtalp/pyERGM	tomtalp__pyERGM-11	ffc1c8263b46f672534fe59266ca36374292bf96	diff --git a/ergm.py b/ergm.py index e7b335b..12e4d1f 100644 --- a/ergm.py +++ b/ergm.py @@ -96,8 +96,7 @@ def _get_random_thetas(self, sampling_method="uniform"): raise ValueError(f"Sampling method {sampling_method} not supported. See docs for supported samplers.") def generate_networks_for_sample(self, replace=True): - sampler = sampling.NaiveMetropolisHastings(self._thetas, self._network_statistics, - is_directed=self._is_directed) + sampler = sampling.NaiveMetropolisHastings(self._thetas, self._network_statistics) G = nx.erdos_renyi_graph(self._n_nodes, self._seed_MCMC_proba, directed=self._is_directed) seed_network = nx.to_numpy_array(G) @@ -164,7 +163,8 @@ def fit(self, observed_network, """ def nll_grad(thetas): - model = ERGM(self._n_nodes, self._network_statistics.metrics, initial_thetas=thetas, is_directed=self._is_directed) + model = ERGM(self._n_nodes, self._network_statistics.metrics, initial_thetas=thetas, + is_directed=self._is_directed) observed_features = model._network_statistics.calculate_statistics(observed_network) @@ -287,8 +287,7 @@ def sample_network(self, sampling_method="NaiveMetropolisHastings", seed_network The sampled connectivity matrix. """ if sampling_method == "NaiveMetropolisHastings": - sampler = sampling.NaiveMetropolisHastings(self._thetas, self._network_statistics, - is_directed=self._is_directed) + sampler = sampling.NaiveMetropolisHastings(self._thetas, self._network_statistics) else: raise ValueError(f"Sampling method {sampling_method} not supported. See docs for supported samplers.") diff --git a/metrics.py b/metrics.py index f865eeb..d77ee26 100644 --- a/metrics.py +++ b/metrics.py @@ -7,7 +7,6 @@ from utils import * -## TODO - Metrics needs to implement a diff() function class Metric(ABC): def __init__(self, metric_name, requires_graph=False): self.metric_name = metric_name @@ -26,7 +25,23 @@ def get_effective_feature_count(self, n): """ return 1 + def calc_change_score(self, net_1: np.ndarray \| nx.Graph, net_2: np.ndarray \| nx.Graph, is_turned_on: bool, + indices: tuple): + """ + The default naive way to calculate the change score (namely, the difference in statistics) of a pair of + networks. + Returns + ------- + statistic of net_2 minus statistic of net_1. + """ + net_2_stat = self.calculate(net_2) + net_1_stat = self.calculate(net_1) + change_score = net_2_stat - net_1_stat + return change_score + + +# TODO: override the change_score function with a more efficient calculation when possible. class NumberOfEdgesUndirected(Metric): def __init__(self): super().__init__(metric_name="num_edges_undirected", requires_graph=False) @@ -35,6 +50,9 @@ def __init__(self): def calculate(self, W: np.ndarray): return np.sum(W) // 2 + def calc_change_score(self, net_1: np.ndarray, net_2: np.ndarray, is_turned_on: bool, indices: tuple): + return 1 if is_turned_on else -1 + class NumberOfEdgesDirected(Metric): def __init__(self): @@ -44,17 +62,33 @@ def __init__(self): def calculate(self, W: np.ndarray): return np.sum(W) + def calc_change_score(self, net_1: np.ndarray, net_2: np.ndarray, is_turned_on: bool, indices: tuple): + return 1 if is_turned_on else -1 + +# TODO: change the name of this one to undirected and implement also a directed version? class NumberOfTriangles(Metric): def __init__(self): - super().__init__(metric_name="num_triangles", requires_graph=True) + super().__init__(metric_name="num_triangles", requires_graph=False) self._is_directed = False - def calculate(self, G: nx.Graph): - if isinstance(G, nx.DiGraph): + def calculate(self, W: np.ndarray): + if not np.all(W.T == W): raise ValueError("NumOfTriangles not implemented for directed graphs") - return sum( - nx.triangles(G).values()) // 3 # nx.triangles counts each triangle 3 times, for every one of its nodes + # the (i,j)-th entry of W^3 counts the number of 3-length paths from node i to node j. Thus, the i-th element on + # the diagonal counts the number of triangles that node 1 is part of (3-length paths from i to itself). As the + # graph is undirected, we get that each path is counted twice ("forward and backwards"), thus the division by 2. + # Additionally, each triangle is counted 3 times by diagonal elements (once for each node that takes part in + # forming it), thus the division by 3. + return (np.linalg.matrix_power(W, 3)).diagonal().sum() // (3 * 2) + + def calc_change_score(self, net_1: np.ndarray, net_2: np.ndarray, is_turned_on: bool, indices: tuple): + # The triangles that are affected by the edge toggling are those that involve it, namely, if the (i,j)-th edge + # is toggled, the change in absolute value equals to the number of nodes k for which the edges (i,k) and (j,k) + # exist. This is equivalent to the number of 2-length paths from i to j, which is the (i,j)-th entry of W^2. + # If the edge is turned on, the change is positive, and otherwise negative. + sign = 1 if is_turned_on else -1 + return sign * ((net_1 @ net_1)[indices[0], indices[1]]) class Reciprocity(Metric): @@ -75,6 +109,17 @@ def get_effective_feature_count(self, n): # n choose 2 return n * (n - 1) // 2 + def calc_change_score(self, net_1: np.ndarray, net_2: np.ndarray, is_turned_on: bool, indices: tuple): + # Note: we intentionally initialize the whole matrix and return np.triu_indices() by the end (rather than + # initializing an array of zeros of size n choose 2) to ensure compliance with the indexing returned by + # the calculate method. + all_changes = np.zeros(net_1.shape) + if net_1[indices[1], indices[0]] and is_turned_on: + all_changes[indices[0], indices[1]] = 1 + elif net_1[indices[1], indices[0]] and not is_turned_on: + all_changes[indices[0], indices[1]] = -1 + return all_changes[np.triu_indices(all_changes.shape[0], 1)] + class TotalReciprocity(Metric): """ @@ -88,6 +133,14 @@ def __init__(self): def calculate(self, W: np.ndarray): return (W * W.T).sum() / 2 + def calc_change_score(self, net_1: np.ndarray, net_2: np.ndarray, is_turned_on: bool, indices: tuple): + if net_1[indices[1], indices[0]] and is_turned_on: + return 1 + elif net_1[indices[1], indices[0]] and not is_turned_on: + return -1 + else: + return 0 + class MetricsCollection: def __init__(self, metrics: Collection[Metric], is_directed: bool): @@ -153,3 +206,32 @@ def calculate_statistics(self, W: np.ndarray): feature_idx += n_features_from_metric return statistics + + def calc_change_scores(self, W1: np.ndarray, W2: np.ndarray, is_turned_on: bool, indices: tuple): + """ + Calculates the vector of change scores, namely g(net_2) - g(net_1) + """ + if W1.shape != W2.shape: + raise ValueError(f"The dimensions of the given networks do not match! {W1.shape}, {W2.shape}") + if self.requires_graph: + G1 = connectivity_matrix_to_G(W1, directed=self.is_directed) + G2 = connectivity_matrix_to_G(W2, directed=self.is_directed) + + n_nodes = W1.shape[0] + change_scores = np.zeros(self.get_num_of_features(n_nodes)) + + feature_idx = 0 + for metric in self.metrics: + if metric.requires_graph: + inputs = (G1, G2) + else: + inputs = (W1, W2) + + n_features_from_metric = metric.get_effective_feature_count(n_nodes) + change_scores[feature_idx:feature_idx + n_features_from_metric] = metric.calc_change_score(inputs[0], + inputs[1], + is_turned_on, + indices) + feature_idx += n_features_from_metric + + return change_scores diff --git a/sampling.py b/sampling.py index 0b06e8c..863f833 100644 --- a/sampling.py +++ b/sampling.py @@ -5,16 +5,16 @@ class Sampler(): - def __init__(self, thetas, network_stats_calculator, is_directed=False): + def __init__(self, thetas, network_stats_calculator): self.thetas = deepcopy(thetas) self.network_stats_calculator = deepcopy(network_stats_calculator) - self.is_directed = is_directed def sample(self, initial_state, n_iterations): pass + class NaiveMetropolisHastings(Sampler): - def __init__(self, thetas, network_stats_calculator: MetricsCollection, is_directed=False, burn_in=1000, steps_per_sample=10): + def __init__(self, thetas, network_stats_calculator: MetricsCollection, burn_in=1000, steps_per_sample=10): """ An implementation for the symmetric proposal Metropolis-Hastings algorithm for ERGMS, using the logit of the acceptance rate. See docs for more details. @@ -27,12 +27,8 @@ def __init__(self, thetas, network_stats_calculator: MetricsCollection, is_direc network_stats_calculator : MetricsCollection A MetricsCollection object that can calculate statistics of a network. - - is_directed : bool - A boolean flag indicating whether the network is directed or not. - """ - super().__init__(thetas, network_stats_calculator, is_directed) + super().__init__(thetas, network_stats_calculator) ## TODO - these two params need to be dependent on the network size self.burn_in = burn_in @@ -43,22 +39,20 @@ def flip_network_edge(self, current_network, i, j): Flip the edge between nodes i, j. If it's an undirected network, we flip entries W_i,j and W_j,i. """ proposed_network = current_network.copy() + is_turned_on = not proposed_network[i, j] proposed_network[i, j] = 1 - proposed_network[i, j] - - if not self.is_directed: + + if not self.network_stats_calculator.is_directed: proposed_network[j, i] = 1 - proposed_network[j, i] - - return proposed_network - - # TODO - this needs to go to Metric() - def _calculate_weighted_change_score(self, proposed_network, current_network): + + return proposed_network, is_turned_on + + def _calculate_weighted_change_score(self, proposed_network, current_network, is_turned_on, indices: tuple): """ Calculate g(proposed_network)-g(current_network) and then inner product with thetas. """ - g_proposed = self.network_stats_calculator.calculate_statistics(proposed_network) - g_current = self.network_stats_calculator.calculate_statistics(current_network) - change_score = g_proposed - g_current - + change_score = self.network_stats_calculator.calc_change_scores(current_network, proposed_network, is_turned_on, + indices) return np.dot(self.thetas, change_score) def sample(self, initial_state, num_of_nets, replace=True): @@ -87,10 +81,11 @@ def sample(self, initial_state, num_of_nets, replace=True): while networks_count != num_of_nets: random_entry = get_random_nondiagonal_matrix_entry(current_network.shape[0]) - - proposed_network = self.flip_network_edge(current_network, random_entry[0], random_entry[1]) - change_score = self._calculate_weighted_change_score(proposed_network, current_network) + proposed_network, is_turned_on = self.flip_network_edge(current_network, random_entry[0], random_entry[1]) + + change_score = self._calculate_weighted_change_score(proposed_network, current_network, is_turned_on, + random_entry) if change_score >= 1: current_network = proposed_network.copy() @@ -98,18 +93,18 @@ def sample(self, initial_state, num_of_nets, replace=True): acceptance_proba = min(1, np.exp(change_score)) if np.random.rand() <= acceptance_proba: current_network = proposed_network.copy() - + if (mcmc_iter_count - self.burn_in) % self.steps_per_sample == 0: sampled_networks[:, :, networks_count] = current_network if not replace: - if np.unique(sampled_networks[:, :, :networks_count+1], axis=2).shape[2] == networks_count + 1: + if np.unique(sampled_networks[:, :, :networks_count + 1], axis=2).shape[2] == networks_count + 1: networks_count += 1 else: - sampled_networks[:, :, networks_count] = np.zeros((net_size, net_size)) + sampled_networks[:, :, networks_count] = np.zeros((net_size, net_size)) else: networks_count += 1 mcmc_iter_count += 1 - - return sampled_networks \ No newline at end of file + + return sampled_networks	Optim enhance Exported the `change_score` function from the sampler to be a method of `Metric`, such that it can be overridden by custom efficient calculations for each `Metric`. Implemented such efficient calculations for the existing `Metric`s.		2024-08-04T13:45:58	[]	[]
tomtalp/pyERGM	tomtalp__pyERGM-3	9ccbc82abb92c35e797c4fab2655f29222caf0a0	diff --git a/.gitignore b/.gitignore index 82f9275..7b6caf3 100644 --- a/.gitignore +++ b/.gitignore @@ -159,4 +159,4 @@ cython_debug/ # be found at https://github.com/github/gitignore/blob/main/Global/JetBrains.gitignore # and can be added to the global gitignore or merged into this file. For a more nuclear # option (not recommended) you can uncomment the following to ignore the entire idea folder. -#.idea/ +.idea/ diff --git a/.idea/.gitignore b/.idea/.gitignore new file mode 100644 index 0000000..13566b8 --- /dev/null +++ b/.idea/.gitignore @@ -0,0 +1,8 @@ +# Default ignored files +/shelf/ +/workspace.xml +# Editor-based HTTP Client requests +/httpRequests/ +# Datasource local storage ignored files +/dataSources/ +/dataSources.local.xml diff --git a/.idea/inspectionProfiles/profiles_settings.xml b/.idea/inspectionProfiles/profiles_settings.xml new file mode 100644 index 0000000..105ce2d --- /dev/null +++ b/.idea/inspectionProfiles/profiles_settings.xml @@ -0,0 +1,6 @@ +<component name="InspectionProjectProfileManager"> + <settings> + <option name="USE_PROJECT_PROFILE" value="false" /> + <version value="1.0" /> + </settings> +</component> \ No newline at end of file diff --git a/.idea/misc.xml b/.idea/misc.xml new file mode 100644 index 0000000..e7c17b2 --- /dev/null +++ b/.idea/misc.xml @@ -0,0 +1,4 @@ +<?xml version="1.0" encoding="UTF-8"?> +<project version="4"> + <component name="ProjectRootManager" version="2" project-jdk-name="Python 3.11 (pyERGM)" project-jdk-type="Python SDK" /> +</project> \ No newline at end of file diff --git a/.idea/modules.xml b/.idea/modules.xml new file mode 100644 index 0000000..5ba93c2 --- /dev/null +++ b/.idea/modules.xml @@ -0,0 +1,8 @@ +<?xml version="1.0" encoding="UTF-8"?> +<project version="4"> + <component name="ProjectModuleManager"> + <modules> + <module fileurl="file://$PROJECT_DIR$/.idea/pyERGM.iml" filepath="$PROJECT_DIR$/.idea/pyERGM.iml" /> + </modules> + </component> +</project> \ No newline at end of file diff --git a/.idea/pyERGM.iml b/.idea/pyERGM.iml new file mode 100644 index 0000000..54c4c0a --- /dev/null +++ b/.idea/pyERGM.iml @@ -0,0 +1,14 @@ +<?xml version="1.0" encoding="UTF-8"?> +<module type="PYTHON_MODULE" version="4"> + <component name="NewModuleRootManager"> + <content url="file://$MODULE_DIR$"> + <excludeFolder url="file://$MODULE_DIR$/venv" /> + </content> + <orderEntry type="jdk" jdkName="Python 3.11 (pyERGM)" jdkType="Python SDK" /> + <orderEntry type="sourceFolder" forTests="false" /> + </component> + <component name="PyDocumentationSettings"> + <option name="format" value="NUMPY" /> + <option name="myDocStringFormat" value="NumPy" /> + </component> +</module> \ No newline at end of file diff --git a/.idea/vcs.xml b/.idea/vcs.xml new file mode 100644 index 0000000..35eb1dd --- /dev/null +++ b/.idea/vcs.xml @@ -0,0 +1,6 @@ +<?xml version="1.0" encoding="UTF-8"?> +<project version="4"> + <component name="VcsDirectoryMappings"> + <mapping directory="" vcs="Git" /> + </component> +</project> \ No newline at end of file diff --git a/ergm.py b/ergm.py index ecc55ad..e0c32ac 100644 --- a/ergm.py +++ b/ergm.py @@ -1,6 +1,7 @@ import numpy as np import networkx as nx -from scipy.optimize import minimize +from scipy.optimize import minimize, OptimizeResult +import time import sampling @@ -8,13 +9,17 @@ class ERGM(): - def __init__(self, - n_nodes, - network_statistics: NetworkStatistics, - is_directed=False, - initial_thetas=None, + def __init__(self, + n_nodes, + network_statistics: NetworkStatistics, + is_directed=False, + initial_thetas=None, initial_normalization_factor=None, - seed_MCMC_proba=0.25): + seed_MCMC_proba=0.25, + n_networks_for_norm=100, + n_mcmc_steps=500, + verbose=True, + optimization_options={}): """ An ERGM model object. @@ -40,21 +45,28 @@ def __init__(self, """ self._n_nodes = n_nodes self._network_statistics = network_statistics - + if initial_thetas is not None: self._thetas = initial_thetas else: self._thetas = self._get_random_thetas(sampling_method="uniform") - + if initial_normalization_factor is not None: self._normalization_factor = initial_normalization_factor else: self._normalization_factor = np.random.normal(50, 10) - self._is_directed = is_directed self._seed_MCMC_proba = seed_MCMC_proba + self.optimization_iter = 0 + self.optimization_start_time = None + + self.n_networks_for_norm = n_networks_for_norm + self.n_mcmc_steps = n_mcmc_steps + self.verbose = verbose + self.optimization_options = optimization_options + def print_model_parameters(self): print(f"Number of nodes: {self._n_nodes}") print(f"Thetas: {self._thetas}") @@ -63,11 +75,14 @@ def print_model_parameters(self): # print(f"Network statistics: {self._network_statistics}") def calculate_weight(self, W: np.ndarray): + if len(W.shape) != 2 or W.shape[0] != self._n_nodes or W.shape[1] != self._n_nodes: + raise ValueError(f"The dimensions of the given adjacency matrix, {W.shape}, don't comply with the number of" + f" nodes in the network: {self._n_nodes}") features = self._network_statistics.calculate_statistics(W) weight = np.exp(np.dot(self._thetas, features)) return weight - + def _get_random_thetas(self, sampling_method="uniform"): if sampling_method == "uniform": return np.random.uniform(-1, 1, self._network_statistics.get_num_of_statistics()) @@ -79,19 +94,19 @@ def _generate_networks_for_sample(self, n_networks, n_mcmc_steps): for _ in range(n_networks): net = self.sample_network(steps=n_mcmc_steps, sampling_method="NaiveMetropolisHastings") networks.append(net) - + return networks - def _approximate_normalization_factor(self, n_networks, n_mcmc_steps): - networks_for_sample = self._generate_networks_for_sample(n_networks, n_mcmc_steps) - + def _approximate_normalization_factor(self): + networks_for_sample = self._generate_networks_for_sample(self.n_networks, self.n_mcmc_steps) + self._normalization_factor = 0 for network in networks_for_sample: weight = self.calculate_weight(network) self._normalization_factor += weight - def fit(self, observed_network, n_networks_for_norm=100, n_mcmc_steps=500, verbose=True, optimization_options={}): + def fit(self, observed_network): """ Initial version, a simple MLE calculated by minimizing negative log likelihood. Normalizaiton factor is approximated via MCMC. @@ -100,12 +115,11 @@ def fit(self, observed_network, n_networks_for_norm=100, n_mcmc_steps=500, verbo """ self._thetas = self._get_random_thetas(sampling_method="uniform") - if verbose: + if self.verbose: print(f"\tStarting fit with initial normalization factor: {self._normalization_factor}") - if optimization_options != {}: - print(f"\tOptimization options: {optimization_options}") - + if self.optimization_options != {}: + print(f"\tOptimization options: {self.optimization_options}") def negative_log_likelihood(thetas): """ @@ -115,8 +129,8 @@ def negative_log_likelihood(thetas): """ # print(f"""Calculating negative log likelihood for thetas: {thetas}""") self._thetas = thetas - - self._approximate_normalization_factor(n_networks_for_norm, n_mcmc_steps) + + self._approximate_normalization_factor() Z = self._normalization_factor # print(f"Approximated Z - {Z}") @@ -125,8 +139,9 @@ def negative_log_likelihood(thetas): log_likelihood = np.log(y_observed_weight) - np.log(Z) return -log_likelihood - - result = minimize(negative_log_likelihood, self._thetas, method='Nelder-Mead', options=optimization_options) + + result = minimize(negative_log_likelihood, self._thetas, method='Nelder-Mead', + options=self.optimization_options) self._thetas = result.x print("\tOptimization result:") print(f"\tTheta: {self._thetas}") @@ -153,7 +168,7 @@ def calculate_probability(self, W: np.ndarray): weight = self.calculate_weight(W) prob = weight / self._normalization_factor - + return prob def sample_network(self, sampling_method="NaiveMetropolisHastings", seed_network=None, steps=500): @@ -175,10 +190,11 @@ def sample_network(self, sampling_method="NaiveMetropolisHastings", seed_network The sampled connectivity matrix. """ if sampling_method == "NaiveMetropolisHastings": - sampler = sampling.NaiveMetropolisHastings(self._thetas, self._network_statistics, is_directed=self._is_directed) + sampler = sampling.NaiveMetropolisHastings(self._thetas, self._network_statistics, + is_directed=self._is_directed) else: raise ValueError(f"Sampling method {sampling_method} not supported. See docs for supported samplers.") - + if seed_network is None: G = nx.erdos_renyi_graph(self._n_nodes, self._seed_MCMC_proba, directed=self._is_directed) seed_network = nx.to_numpy_array(G) @@ -187,6 +203,101 @@ def sample_network(self, sampling_method="NaiveMetropolisHastings", seed_network return network + +class BruteForceERGM(ERGM): + """ + A class that implements ERGM by iterating over the entire space of networks and calculating stuff exactly (rather + than using statistical methods for approximating and sampling). + This is mainly for tests. + """ + # The maximum number of nodes that is allowed for carrying brute force calculations (i.e. iterating the whole space + # of networks and calculating stuff exactly). This becomes not tractable above this limit. + MAX_NODES_BRUTE_FORCE_DIRECTED = 5 + MAX_NODES_BRUTE_FORCE_NOT_DIRECTED = 7 + + def __init__(self, + n_nodes, + network_statistics: NetworkStatistics, + is_directed=False, + initial_thetas=None): + super().__init__(n_nodes, + network_statistics, + is_directed, + initial_thetas) + self._all_weights = self._calc_all_weights() + self._normalization_factor = self._all_weights.sum() + + def _validate_net_size(self): + return ( + (self._is_directed and self._n_nodes <= BruteForceERGM.MAX_NODES_BRUTE_FORCE_DIRECTED) + or + (not self._is_directed and self._n_nodes <= BruteForceERGM.MAX_NODES_BRUTE_FORCE_NOT_DIRECTED) + ) + + def _calc_all_weights(self): + if not self._validate_net_size(): + directed_str = 'directed' if self._is_directed else 'not directed' + size_limit = BruteForceERGM.MAX_NODES_BRUTE_FORCE_DIRECTED if self._is_directed else ( + BruteForceERGM.MAX_NODES_BRUTE_FORCE_NOT_DIRECTED) + raise ValueError( + f"The number of nodes {self._n_nodes} is larger than the maximum allowed for brute force of " + f"{directed_str} graphs calculations {size_limit}") + num_pos_connects = self._n_nodes * (self._n_nodes - 1) + if not self._is_directed: + num_pos_connects //= 2 + space_size = 2 ** num_pos_connects + all_weights = np.zeros(space_size) + for i in range(space_size): + cur_adj_mat = construct_adj_mat_from_int(i, self._n_nodes, self._is_directed) + all_weights[i] = super().calculate_weight(cur_adj_mat) + return all_weights + + def calculate_weight(self, W: np.ndarray): + adj_mat_idx = construct_int_from_adj_mat(W, self._is_directed) + return self._all_weights[adj_mat_idx] + + def sample_network(self, sampling_method="Exact", seed_network=None, steps=0): + if sampling_method != "Exact": + raise ValueError("BruteForceERGM supports only exact sampling (this is its whole purpose)") + + all_nets_probs = self._all_weights / self._normalization_factor + sampled_idx = np.random.choice(all_nets_probs.size, p=all_nets_probs) + return construct_adj_mat_from_int(sampled_idx, self._n_nodes, self._is_directed) + + def fit(self, observed_network): + def nll(thetas): + model = BruteForceERGM(self._n_nodes, self._network_statistics, initial_thetas=thetas, + is_directed=self._is_directed) + return np.log(model._normalization_factor) - np.log(model.calculate_weight(observed_network)) + + def nll_grad(thetas): + model = BruteForceERGM(self._n_nodes, self._network_statistics, initial_thetas=thetas, + is_directed=self._is_directed) + observed_features = model._network_statistics.calculate_statistics(observed_network) + all_probs = model._all_weights / model._normalization_factor + num_features = model._network_statistics.get_num_of_statistics() + num_nets = all_probs.size + all_features_by_all_nets = np.zeros((num_features, num_nets)) + for i in range(num_nets): + all_features_by_all_nets[:, i] = model._network_statistics.calculate_statistics( + construct_adj_mat_from_int(i, self._n_nodes, self._is_directed)) + expected_features = all_features_by_all_nets @ all_probs + return expected_features - observed_features + + def after_iteration_callback(intermediate_result: OptimizeResult): + self.optimization_iter += 1 + cur_time = time.time() + print(f'iteration: {self.optimization_iter}, time from start ' + f'training: {cur_time - self.optimization_start_time} ' + f'log likelihood: {-intermediate_result.fun}') + + self.optimization_iter = 0 + print("optimization started") + self.optimization_start_time = time.time() + res = minimize(nll, self._thetas, jac=nll_grad, callback=after_iteration_callback) + self._thetas = res.x + print(res) + # n_nodes = 5 # stats_calculator = NetworkStatistics(metric_names=["num_edges"]) # ergm = ERGM(n_nodes, stats_calculator, is_directed=False) @@ -250,6 +361,3 @@ def sample_network(self, sampling_method="NaiveMetropolisHastings", seed_network # print(samples_before_fit) # print("Samples after fit:") # print(samples_after_fit) - - - \ No newline at end of file diff --git a/examples/first_example.ipynb b/examples/first_example.ipynb index c7d6b3a..da9ba5e 100644 --- a/examples/first_example.ipynb +++ b/examples/first_example.ipynb @@ -2,25 +2,45 @@ "cells": [ { "cell_type": "code", - "execution_count": 1, - "metadata": {}, + "execution_count": 6, + "metadata": { + "ExecuteTime": { + "end_time": "2024-07-21T12:28:56.044446900Z", + "start_time": "2024-07-21T12:28:55.917504Z" + } + }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ - "/Users/tomtalpir/Random/pyERGM\n" + "D:\\OrenRichter\\Research\\pyERGM\n" + ] + }, + { + "name": "stderr", + "output_type": "stream", + "text": [ + "D:\\OrenRichter\\Research\\pyERGM\\venv\\Lib\\site-packages\\IPython\\core\\magics\\osm.py:417: UserWarning: This is now an optional IPython functionality, setting dhist requires you to install the `pickleshare` library.\n", + " self.shell.db['dhist'] = compress_dhist(dhist)[-100:]\n" ] } ], "source": [ - "%cd /Users/tomtalpir/Random/pyERGM" + "from pathlib import Path\n", + "parent_dir = str(Path.cwd().parent)\n", + "%cd $parent_dir" ] }, { "cell_type": "code", - "execution_count": 2, - "metadata": {}, + "execution_count": 8, + "metadata": { + "ExecuteTime": { + "end_time": "2024-07-21T12:30:25.054064700Z", + "start_time": "2024-07-21T12:30:23.471407200Z" + } + }, "outputs": [], "source": [ "from utils import \n", @@ -39,16 +59,21 @@ }, { "cell_type": "code", - "execution_count": 3, - "metadata": {}, + "execution_count": 9, + "metadata": { + "ExecuteTime": { + "end_time": "2024-07-21T13:32:42.547782400Z", + "start_time": "2024-07-21T13:32:42.531930200Z" + } + }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Number of nodes: 10\n", - "Thetas: [-0.29766759]\n", - "Normalization factor approx: 50.779457576278446\n", + "Thetas: [0.33527386]\n", + "Normalization factor approx: 36.49400038899532\n", "Is directed: False\n" ] } @@ -70,104 +95,20 @@ }, { "cell_type": "code", - "execution_count": 4, - "metadata": {}, + "execution_count": 10, + "metadata": { + "ExecuteTime": { + "end_time": "2024-07-21T13:33:29.562627900Z", + "start_time": "2024-07-21T13:33:26.542060800Z" + } + }, "outputs": [ { "data": { - "text/html": [ - "<div>\n", - "<style scoped>\n", - " .dataframe tbody tr th:only-of-type {\n", - " vertical-align: middle;\n", - " }\n", - "\n", - " .dataframe tbody tr th {\n", - " vertical-align: top;\n", - " }\n", - "\n", - " .dataframe thead th {\n", - " text-align: right;\n", - " }\n", - "</style>\n", - "<table border=\"1\" class=\"dataframe\">\n", - " <thead>\n", - " <tr style=\"text-align: right;\">\n", - " <th></th>\n", - " <th>edge_count</th>\n", - " <th>triangle_count</th>\n", - " </tr>\n", - " </thead>\n", - " <tbody>\n", - " <tr>\n", - " <th>0</th>\n", - " <td>33</td>\n", - " <td>47</td>\n", - " </tr>\n", - " <tr>\n", - " <th>1</th>\n", - " <td>35</td>\n", - " <td>57</td>\n", - " </tr>\n", - " <tr>\n", - " <th>2</th>\n", - " <td>36</td>\n", - " <td>62</td>\n", - " </tr>\n", - " <tr>\n", - " <th>3</th>\n", - " <td>36</td>\n", - " <td>57</td>\n", - " </tr>\n", - " <tr>\n", - " <th>4</th>\n", - " <td>39</td>\n", - " <td>79</td>\n", - " </tr>\n", - " <tr>\n", - " <th>5</th>\n", - " <td>36</td>\n", - " <td>63</td>\n", - " </tr>\n", - " <tr>\n", - " <th>6</th>\n", - " <td>34</td>\n", - " <td>53</td>\n", - " </tr>\n", - " <tr>\n", - " <th>7</th>\n", - " <td>38</td>\n", - " <td>76</td>\n", - " </tr>\n", - " <tr>\n", - " <th>8</th>\n", - " <td>38</td>\n", - " <td>72</td>\n", - " </tr>\n", - " <tr>\n", - " <th>9</th>\n", - " <td>32</td>\n", - " <td>46</td>\n", - " </tr>\n", - " </tbody>\n", - "</table>\n", - "</div>" - ], - "text/plain": [ - " edge_count triangle_count\n", - "0 33 47\n", - "1 35 57\n", - "2 36 62\n", - "3 36 57\n", - "4 39 79\n", - "5 36 63\n", - "6 34 53\n", - "7 38 76\n", - "8 38 72\n", - "9 32 46" - ] + "text/plain": " edge_count triangle_count\n0 45 120\n1 45 120\n2 45 120\n3 45 120\n4 45 120\n5 45 120\n6 45 120\n7 45 120\n8 45 120\n9 45 120", + "text/html": "<div>\n<style scoped>\n .dataframe tbody tr th:only-of-type {\n vertical-align: middle;\n }\n\n .dataframe tbody tr th {\n vertical-align: top;\n }\n\n .dataframe thead th {\n text-align: right;\n }\n</style>\n<table border=\"1\" class=\"dataframe\">\n <thead>\n <tr style=\"text-align: right;\">\n <th></th>\n <th>edge_count</th>\n <th>triangle_count</th>\n </tr>\n </thead>\n <tbody>\n <tr>\n <th>0</th>\n <td>45</td>\n <td>120</td>\n </tr>\n <tr>\n <th>1</th>\n <td>45</td>\n <td>120</td>\n </tr>\n <tr>\n <th>2</th>\n <td>45</td>\n <td>120</td>\n </tr>\n <tr>\n <th>3</th>\n <td>45</td>\n <td>120</td>\n </tr>\n <tr>\n <th>4</th>\n <td>45</td>\n <td>120</td>\n </tr>\n <tr>\n <th>5</th>\n <td>45</td>\n <td>120</td>\n </tr>\n <tr>\n <th>6</th>\n <td>45</td>\n <td>120</td>\n </tr>\n <tr>\n <th>7</th>\n <td>45</td>\n <td>120</td>\n </tr>\n <tr>\n <th>8</th>\n <td>45</td>\n <td>120</td>\n </tr>\n <tr>\n <th>9</th>\n <td>45</td>\n <td>120</td>\n </tr>\n </tbody>\n</table>\n</div>" }, - "execution_count": 4, + "execution_count": 10, "metadata": {}, "output_type": "execute_result" } @@ -202,8 +143,13 @@ }, { "cell_type": "code", - "execution_count": 42, - "metadata": {}, + "execution_count": 11, + "metadata": { + "ExecuteTime": { + "end_time": "2024-07-21T13:33:52.394915900Z", + "start_time": "2024-07-21T13:33:52.379092600Z" + } + }, "outputs": [ { "name": "stdout", @@ -211,8 +157,8 @@ "text": [ "Baseline ERGM parameters - \n", "Number of nodes: 5\n", - "Thetas: [-1.0986122886681098]\n", - "Normalization factor approx: 19.45775812919028\n", + "Thetas: [np.float64(-1.0986122886681098)]\n", + "Normalization factor approx: 43.441946503045074\n", "Is directed: False\n" ] } @@ -237,20 +183,25 @@ }, { "cell_type": "code", - "execution_count": 43, - "metadata": {}, + "execution_count": 12, + "metadata": { + "ExecuteTime": { + "end_time": "2024-07-21T13:34:07.998379900Z", + "start_time": "2024-07-21T13:34:07.904510500Z" + } + }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Sampled a random network from a model with theta = -1.0986122886681098\n", - "[[0. 0. 0. 1. 1.]\n", - " [0. 0. 1. 1. 1.]\n", - " [0. 1. 0. 0. 1.]\n", - " [1. 1. 0. 0. 1.]\n", - " [1. 1. 1. 1. 0.]]\n", - "Network has statistics - edge count: 7, triangle count: 3\n" + "[[0. 0. 1. 1. 1.]\n", + " [0. 0. 0. 1. 0.]\n", + " [1. 0. 0. 0. 0.]\n", + " [1. 1. 0. 0. 0.]\n", + " [1. 0. 0. 0. 0.]]\n", + "Network has statistics - edge count: 4, triangle count: 0\n" ] } ], @@ -276,8 +227,13 @@ }, { "cell_type": "code", - "execution_count": 46, - "metadata": {}, + "execution_count": 13, + "metadata": { + "ExecuteTime": { + "end_time": "2024-07-21T13:35:20.570400200Z", + "start_time": "2024-07-21T13:35:19.572400900Z" + } + }, "outputs": [ { "name": "stdout", @@ -285,13 +241,13 @@ "text": [ "Initial ERGM parameters:\n", "Number of nodes: 5\n", - "Thetas: [-0.33698419]\n", - "Normalization factor approx: 43.1253267687915\n", + "Thetas: [0.38415461]\n", + "Normalization factor approx: 50.33647025496875\n", "Is directed: False\n", "\n", "Mean of samples before fitting, calculated on 10 samples- \n", - "edge_count 8.0\n", - "triangle_count 5.0\n", + "edge_count 10.0\n", + "triangle_count 10.0\n", "dtype: float64\n" ] } @@ -331,28 +287,33 @@ }, { "cell_type": "code", - "execution_count": 47, - "metadata": {}, + "execution_count": 14, + "metadata": { + "ExecuteTime": { + "end_time": "2024-07-21T14:09:49.796634200Z", + "start_time": "2024-07-21T13:35:32.539389400Z" + } + }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Fitting ERGM...\n", - "\tStarting fit with initial normalization factor: 43.1253267687915\n", + "\tStarting fit with initial normalization factor: 50.33647025496875\n", "\tOptimization options: {'disp': True, 'maxiter': 500, 'maxfev': 500}\n", "\tOptimization result:\n", - "\tTheta: [-0.15381744]\n", - "\tNormalization factor: 135.44816303424523\n", + "\tTheta: [-0.39642108]\n", + "\tNormalization factor: 41.88198286288824\n", " message: Maximum number of function evaluations has been exceeded.\n", " success: False\n", " status: 1\n", - " fun: 5.963890498779282\n", - " x: [-1.538e-01]\n", - " nit: 186\n", + " fun: 5.243735736850018\n", + " x: [-3.964e-01]\n", + " nit: 189\n", " nfev: 500\n", - " final_simplex: (array([[-1.538e-01],\n", - " [-1.538e-01]]), array([ 5.964e+00, 5.984e+00]))\n", + " final_simplex: (array([[-3.964e-01],\n", + " [-3.964e-01]]), array([ 5.244e+00, 5.271e+00]))\n", "Done fitting!\n" ] }, @@ -360,8 +321,8 @@ "name": "stderr", "output_type": "stream", "text": [ - "/Users/tomtalpir/Random/pyERGM/ergm.py:130: RuntimeWarning: Maximum number of function evaluations has been exceeded.\n", - " result = minimize(negative_log_likelihood, self._thetas, method='Nelder-Mead', options=optimization_options)\n" + "D:\\OrenRichter\\Research\\pyERGM\\ergm.py:129: RuntimeWarning: Maximum number of function evaluations has been exceeded.\n", + " # OR: Providing the Scipy optimizer with a gradient function can streamline the optimization substantially. The\n" ] } ], diff --git a/sampling.py b/sampling.py index ab29c64..a1a07e3 100644 --- a/sampling.py +++ b/sampling.py @@ -55,7 +55,7 @@ def sample(self, initial_state, n_iterations): for i in range(n_iterations): random_entry = get_random_nondiagonal_matrix_entry(current_network.shape[0]) - + y_plus = self.override_network_edge(current_network, random_entry[0], random_entry[1], 1) y_minus = self.override_network_edge(current_network, random_entry[0], random_entry[1], 0) diff --git a/utils.py b/utils.py index d2d906c..9b5c430 100644 --- a/utils.py +++ b/utils.py @@ -1,6 +1,7 @@ import numpy as np import networkx as nx + def perturb_network_by_overriding_edge(network, value, i, j, is_directed): perturbed_net = network.copy() perturbed_net[i, j] = value @@ -9,6 +10,7 @@ def perturb_network_by_overriding_edge(network, value, i, j, is_directed): return perturbed_net + def connectivity_matrix_to_G(W: np.ndarray, directed=False): """ Convert a connectivity matrix to a graph object. @@ -30,6 +32,7 @@ def connectivity_matrix_to_G(W: np.ndarray, directed=False): G = nx.from_numpy_array(W) return G + def get_random_nondiagonal_matrix_entry(n: int): """ Get a random entry for a non-diagonal entry of a matrix. @@ -46,13 +49,14 @@ def get_random_nondiagonal_matrix_entry(n: int): """ i = np.random.randint(n) j = np.random.randint(n) - + while i == j: i = np.random.randint(n) j = np.random.randint(n) - + return (i, j) + class NetworkStatistics(): def __init__(self, metric_names=[], custom_metrics={}, directed=False): """ @@ -87,10 +91,10 @@ def __init__(self, metric_names=[], custom_metrics={}, directed=False): self._validate_predefined_metrics() self.custom_metrics = custom_metrics - self._validate_custom_metrics() + self._validate_custom_metrics() + + self._register_metrics() - self._register_metrics() - def _validate_predefined_metrics(self): if len(self.metric_names) == 0: raise ValueError("At least one metric must be specified.") @@ -105,7 +109,7 @@ def _validate_custom_metrics(self): for metric_name, stat_func in self.custom_metrics: if not callable(stat_func): raise ValueError(f"Custom Metric {metric_name} is not a function.") - + def _register_metrics(self): self.statistics_functions = {} @@ -113,10 +117,11 @@ def _register_metrics(self): if metric_name == "num_edges": func = lambda G: len(G.edges()) elif metric_name == "num_triangles": - func = lambda G: sum(nx.triangles(G).values()) // 3 # nx.triangles counts each triangle 3 times, for every node in it + func = lambda G: sum( + nx.triangles(G).values()) // 3 # nx.triangles counts each triangle 3 times, for every node in it self.statistics_functions[metric_name] = func - + def get_num_of_statistics(self): """ Get the number of statistics that are registered. @@ -127,7 +132,7 @@ def get_num_of_statistics(self): The number of statistics. """ return len(self.statistics_functions) - + def calculate_statistics(self, W: np.ndarray, verbose=False): """ Calculate the statistics of a graph. @@ -144,12 +149,37 @@ def calculate_statistics(self, W: np.ndarray, verbose=False): An array of statistics, represented as a numpy array of values, or a dict if verbose=True """ G = connectivity_matrix_to_G(W, directed=self._is_directed) - + stats = {name: func(G) for name, func in self.statistics_functions.items()} if verbose: return stats else: return np.array(list(stats.values())) - - \ No newline at end of file + +def construct_adj_mat_from_int(int_code: int, num_nodes: int, is_directed: bool) -> np.ndarray: + num_pos_connects = num_nodes (num_nodes - 1) + if not is_directed: + num_pos_connects //= 2 + adj_mat_str = f'{int_code:0{num_pos_connects}b}' + cur_adj_mat = np.zeros((num_nodes, num_nodes)) + mat_entries_arr = np.array(list(adj_mat_str), 'uint8') + if is_directed: + cur_adj_mat[~np.eye(num_nodes, dtype=bool)] = mat_entries_arr + else: + upper_triangle_indices = np.triu_indices(num_nodes, k=1) + cur_adj_mat[upper_triangle_indices] = mat_entries_arr + lower_triangle_indices_aligned = (upper_triangle_indices[1], upper_triangle_indices[0]) + cur_adj_mat[lower_triangle_indices_aligned] = mat_entries_arr + return cur_adj_mat + + +def construct_int_from_adj_mat(adj_mat: np.ndarray, is_directed: bool) -> int: + if len(adj_mat.shape) != 2 or adj_mat.shape[0] != adj_mat.shape[1]: + raise ValueError(f"The dimensions of the given adjacency matrix {adj_mat.shape} are not valid for an " + f"adjacency matrix (should be a 2D squared matrix)") + num_nodes = adj_mat.shape[0] + adj_mat_no_diag = adj_mat[~np.eye(num_nodes, dtype=bool)] + if not is_directed: + adj_mat_no_diag = adj_mat_no_diag[:adj_mat_no_diag.size // 2] + return round((adj_mat_no_diag * 2 ** np.arange(adj_mat_no_diag.size - 1, -1, -1).astype(np.ulonglong)).sum())	Brute force ERGM Implemented an ERGM that iterates the entire space of networks, and calculates stuff exactly, to be used as a benchmark. This is obviously tractable only for small networks.		2024-07-23T09:15:15	[]	[]
ninoseki/bukowski	ninoseki__bukowski-6	ba4f10ac6c5816001fda686ab4576a1af4a62dda	diff --git a/pyproject.toml b/pyproject.toml index bb95b18..d80ab1c 100644 --- a/pyproject.toml +++ b/pyproject.toml @@ -50,6 +50,7 @@ select = [ "F", # pyflakes "I", # isort "N", # pep8-naming + "PT", # flake8-pytest-style "RET", # flake8-return "RUF", # Ruff-specific rules "SIM", # flake8-simplify diff --git a/src/bukowski/main.py b/src/bukowski/main.py index bd90900..a6119d3 100644 --- a/src/bukowski/main.py +++ b/src/bukowski/main.py @@ -1,9 +1,11 @@ +import re from collections.abc import Iterable from dataclasses import dataclass from functools import partial from typing import Any, cast import tomlkit +from packaging.utils import canonicalize_name from poetry.core.constraints.version import ( Version, VersionConstraint, @@ -124,13 +126,20 @@ def init() -> TOMLDocument: return pyproject +def normalize_name(name: str): + # NOTE: > names must start and end with a letter or digit and may only contain -, _, ., and alphanumeric characters. # noqa: E501 + # (ref, https://github.com/astral-sh/uv/blob/main/crates/uv-cli/src/lib.rs#L46-L53) + subbed = re.sub("[^a-zA-Z0-9._-]+", "-", name) + return canonicalize_name(subbed) + + @safe def set_project(pyproject: TOMLDocument, *, poetry: Poetry) -> TOMLDocument: package = poetry.package content = cast(dict[str, Any], pyproject["project"]) - content["name"] = package.name + content["name"] = normalize_name(package.name) content["version"] = package.version.text content["description"] = package.description	Project names should be hyphenated to avoid TOML parser error Currently Bukowski turns `name = "Bingo to Bongo"` into `name = "bingo to bongo"`, but it should return `name = "bingo-to-bongo"`.		2024-09-17T09:35:02	[]	[]
pydata/pydata-sphinx-theme	pydata__pydata-sphinx-theme-2007	5135b8f2b26e470a8aa49bdcda284b3968d3ffda	diff --git a/src/pydata_sphinx_theme/assets/styles/content/_code.scss b/src/pydata_sphinx_theme/assets/styles/content/_code.scss index 6371097ea..d7dd169b2 100644 --- a/src/pydata_sphinx_theme/assets/styles/content/_code.scss +++ b/src/pydata_sphinx_theme/assets/styles/content/_code.scss @@ -59,6 +59,10 @@ code.literal { a > code { color: var(--pst-color-inline-code-links); + + &:hover { + color: var(--pst-color-link-hover); + } } // Minimum opacity needed for linenos to be WCAG AA conformant	hovering on monospaced links changes underline color but not text color Not sure if that is intentional or not? You can see this by hovering a function name on https://scikit-learn.org/stable/api/index.html for example.		2024-10-07T21:10:16	[]	[]
nsat/pypredict	nsat__pypredict-63	2084f535b9a04caa315391e580e5a725271b29d9	diff --git a/predict.c b/predict.c index 703deb7..26a3e2e 100644 --- a/predict.c +++ b/predict.c @@ -119,14 +119,6 @@ static PyObject PredictException; static PyObject NoTransitException; -/* - TODO: This is a refactoring hack to ensure we get consistent before/after - tests. Freezes the clock to a specified time, making calculations - deterministic. (remove when finished along with (2) call sites) -/ -char debug_freeze_time = 0; -struct tm debug_frozen_tm = { tm_year: 114, tm_mon: 10, tm_mday: 2 }; - // This struct represents an observation of a particular satellite // from a particular reference point (on earth) at a particular time // and is used primarily in the PyPredict code. @@ -175,8 +167,8 @@ struct { long catnum; // Catalog Number (a.k.a NORAD id) long setnum; // Element Set No. char designator[10]; // Designator - int year; // Reference Epoch (year?) (?) - double refepoch; // Reference Epoch (?) + int year; // Reference Epoch from TLE (2-digit year) + double refepoch; // Reference Epoch from TLE (day of the year and fractional portion of the day) double incl; // Inclination double raan; // RAAN double eccn; // Eccentricity @@ -2083,13 +2075,6 @@ void Calculate_RADec(double time, vector_t pos, vector_t vel, geodetic_t geod time_t CurrentTime() { - //TODO: Refactoring Hack (remove) - if (debug_freeze_time) - { - time_t debug_frozen_time = mktime(&debug_frozen_tm); - return debug_frozen_time; - } - return time(NULL); } @@ -2610,14 +2595,6 @@ double CurrentDaynum() { /* Read the system clock and return the number of days since 31Dec79 00:00:00 UTC (daynum 0) / - - //TODO: Refactoring Hack (remove) - if (debug_freeze_time) - { - time_t debug_frozen_time = mktime(&debug_frozen_tm); - return ((((double)debug_frozen_time)/86400.0) - 3651.0); - } - int x; struct timeval tptr; double usecs, seconds; @@ -2630,44 +2607,6 @@ double CurrentDaynum() return ((seconds/86400.0)-3651.0); } -char Daynum2String(daynum) -double daynum; -{ - /* This function takes the given epoch as a fractional number of - days since 31Dec79 00:00:00 UTC and returns the corresponding - date as a string of the form "Tue 12Oct99 17:22:37". / - - char timestr[26]; - time_t t; - int x; - - / Convert daynum to Unix time (seconds since 01-Jan-70) / - t=(time_t)(86400.0(daynum+3651.0)); - - sprintf(timestr,"%s",asctime(gmtime(&t))); - - if (timestr[8]==' ') - { - timestr[8]='0'; - } - - for (x=0; x<=3; output[x]=timestr[x], x++); - - output[4]=timestr[8]; - output[5]=timestr[9]; - output[6]=timestr[4]; - output[7]=timestr[5]; - output[8]=timestr[6]; - output[9]=timestr[22]; - output[10]=timestr[23]; - output[11]=' '; - - for (x=12; x<=19; output[x]=timestr[x-1], x++); - - output[20]=0; - return output; -} - void FindMoon(double daynum) { /* This function determines the position of the moon, including @@ -3543,7 +3482,7 @@ static char quick_find_docs[] = static PyObject* quick_predict(PyObject* self, PyObject args) { - double now; + double tle_epoch; int lastel=0; char errbuff[100]; observation obs = { 0 }; @@ -3554,18 +3493,16 @@ static PyObject quick_predict(PyObject* self, PyObject *args) goto cleanup_and_raise_exception; } - now=CurrentDaynum(); - if (load(args) != 0) { // load will set the appropriate exception string if it fails. goto cleanup_and_raise_exception; } - //TODO: Seems like this should be based on the freshness of the TLE, not wall clock. - if ((daynum<now-365.0) \|\| (daynum>now+365.0)) + tle_epoch=(double)DayNum(1,0,sat.year)+sat.refepoch; + if ((daynum<tle_epoch-365.0) \|\| (daynum>tle_epoch+365.0)) { - sprintf(errbuff, "time %s too far from present\n", Daynum2String(daynum)); + sprintf(errbuff, "day number %.0f too far from tle epoch day %.0f\n", daynum, tle_epoch); PyErr_SetString(PredictException, errbuff); goto cleanup_and_raise_exception; } diff --git a/setup.py b/setup.py index d0b07cc..10aa204 100755 --- a/setup.py +++ b/setup.py @@ -7,7 +7,7 @@ setup( name="pypredict", - version="1.7.1", + version="1.7.2", author="Jesse Trutna", author_email="[email protected]", maintainer="Spire Global Inc",	Base staleness check on TLE Epoch instead of wall clock Fixes #45 The staleness check within `predict.c` was based on being within a year of the current time. We change it so that it has to be within a year of the Epoch date on the TLE. PredictException: time XXX too far from present should use TLE Epoch The [code that determines staleness](https://github.com/nsat/pypredict/blob/15f32d057d2c6a325c27c477a67675858136f183/predict.c#L3565-L3571) works off of current time instead of TLE Epoch. That makes it hard to write unit tests, because they will fail a year later. As indicated by the comment on the code, this should measure against the TLE Epoch instead.	Hi, thanks for your pull request! This looks like it will be a helpful fix. Before I merge your PR I'd like to make sure it's tested. Would you mind adding some test cases demonstrating the desired behavior and edge cases?	2024-04-23T19:21:01	[]	[]
BigDataBiology/SemiBin	BigDataBiology__SemiBin-78	c065098d63a46fc684bf9db8b18717150aca7ea1	diff --git a/ChangeLog b/ChangeLog index 82b4f3e..69f5383 100644 --- a/ChangeLog +++ b/ChangeLog @@ -1,6 +1,7 @@ Version 0.7.0 * Reuse markers.hmmout to make the training from several samples faster * Add option `--tmpdir` to set temporary directory + * Substitute FragGeneScan with Prodigal (FragGeneScan can still be used with `--predictor` parameter) Version 0.6.0 Mon 7 Feb 2022 by BigDataBiology * Provide pretrained models from soil, cat gut, human oral, diff --git a/README.md b/README.md index bd22968..066619e 100644 --- a/README.md +++ b/README.md @@ -47,12 +47,13 @@ You can download the source code from github and install. Install dependence packages using conda: [MMseqs2](https://github.com/soedinglab/MMseqs2), -[Bedtools](http://bedtools.readthedocs.org/]), [Hmmer](http://hmmer.org/), and -[Fraggenescan](https://sourceforge.net/projects/fraggenescan/). +[Bedtools](http://bedtools.readthedocs.org/]), [Hmmer](http://hmmer.org/), +[Fraggenescan](https://sourceforge.net/projects/fraggenescan/) and [Prodigal](https://github.com/hyattpd/Prodigal) ```bash conda install -c conda-forge -c bioconda mmseqs2=13.45111 conda install -c bioconda bedtools hmmer fraggenescan==1.30 +conda install -c bioconda prodigal ``` Once the dependencies are installed, you can install by running: diff --git a/SemiBin/cluster.py b/SemiBin/cluster.py index 429d0d2..8c24701 100644 --- a/SemiBin/cluster.py +++ b/SemiBin/cluster.py @@ -83,7 +83,7 @@ def cal_kl(m, v, use_ne='auto'): def cluster(model, data, device, max_edges, max_node, is_combined, - logger, n_sample, out, contig_dict, binned_length, num_process, minfasta, recluster, random_seed): + logger, n_sample, out, contig_dict, binned_length, num_process, minfasta, recluster, random_seed, predictor = 'prodigal'): """ Cluster contigs into bins max_edges: max edges of one contig considered in binning @@ -219,7 +219,8 @@ def cluster(model, data, device, max_edges, max_node, is_combined, cfasta, binned_length, num_process, - multi_mode=True) + multi_mode=True, + predictor=predictor) for ix,bin_path in enumerate(bin_files): # if there are no hits, the output will be naturally empty diff --git a/SemiBin/main.py b/SemiBin/main.py index 10ef704..cad0a92 100644 --- a/SemiBin/main.py +++ b/SemiBin/main.py @@ -166,6 +166,14 @@ def parse_args(args): default=None, help='Path to the trained semi-supervised deep learning model.') + for p in [single_easy_bin, multi_easy_bin, training, binning]: + p.add_argument('--predictor', + required=False, + type=str, + help='gene predictor used to estimate the number of bins(prodigal/fraggenescan)', + dest='predictor', + default='prodigal') + for p in [single_easy_bin, multi_easy_bin, generate_cannot_links, training, binning]: p.add_argument('--tmpdir', required=False, @@ -611,7 +619,7 @@ def fasta_sample_iter(fn): def training(logger, contig_fasta, num_process, data, data_split, cannot_link, batchsize, - epoches, output, device, ratio, min_length, mode): + epoches, output, device, ratio, min_length, mode, predictor = 'prodigal'): """ Training the model @@ -666,13 +674,14 @@ def training(logger, contig_fasta, num_process, epoches, device, num_process, - mode = mode) + mode = mode, + predictor=predictor) def binning(logger, num_process, data, max_edges, max_node, minfasta, binned_length, contig_dict, recluster,model_path, - random_seed,output, device, environment): + random_seed,output, device, environment, predictor = 'prodigal'): """ Clustering the contigs to get the final bins. @@ -711,12 +720,13 @@ def binning(logger, num_process, data, num_process, minfasta, recluster, - random_seed) + random_seed, + predictor=predictor) def single_easy_binning(args, logger, binned_length, must_link_threshold, - contig_dict, recluster,random_seed, output, device, environment): + contig_dict, recluster,random_seed, output, device, environment, predictor = 'prodigal'): """ contain `generate_cannot_links`, `generate_sequence_features_single`, `train`, `bin` in one command for single-sample and co-assembly binning """ @@ -748,17 +758,17 @@ def single_easy_binning(args, logger, binned_length, training(logger, [args.contig_fasta], args.num_process, [data_path], [data_split_path], [os.path.join(output, 'cannot', 'cannot.txt')], - args.batchsize, args.epoches, output, device, args.ratio, args.min_len, mode='single') + args.batchsize, args.epoches, output, device, args.ratio, args.min_len, mode='single', predictor=predictor) binning(logger, args.num_process, data_path, args.max_edges, args.max_node, args.minfasta_kb * 1000, binned_length, contig_dict, recluster, os.path.join(output, 'model.h5') if environment is None else None, - random_seed, output, device, environment if environment is not None else None) + random_seed, output, device, environment if environment is not None else None, predictor=predictor) def multi_easy_binning(args, logger, recluster, - random_seed, output, device): + random_seed, output, device, predictor = 'prodigal'): """ contain `generate_cannot_links`, `generate_sequence_features_multi`, `train`, `bin` in one command for multi-sample binning """ @@ -774,7 +784,7 @@ def multi_easy_binning(args, logger, recluster, args.ratio, args.min_len, args.ml_threshold, - output, ) + output,) for sample in sample_list: logger.info( @@ -810,13 +820,13 @@ def multi_easy_binning(args, logger, recluster, training(logger, [sample_fasta], args.num_process, [sample_data], [sample_data_split], [sample_cannot], args.batchsize, args.epoches, os.path.join(output, 'samples', sample), - device, args.ratio, args.min_len, mode='single') + device, args.ratio, args.min_len, mode='single', predictor=predictor) binning(logger, args.num_process, sample_data, args.max_edges, args.max_node, args.minfasta_kb * 1000, binned_length, contig_dict, recluster, os.path.join(output, 'samples', sample, 'model.h5'), random_seed , - os.path.join(output, 'samples', sample), device, None) + os.path.join(output, 'samples', sample), device, None, predictor=predictor) os.makedirs(os.path.join(output, 'bins'), exist_ok=True) for sample in sample_list: @@ -910,7 +920,7 @@ def main(): set_random_seed(args.random_seed) training(logger, args.contig_fasta, args.num_process, args.data, args.data_split, args.cannot_link, - args.batchsize, args.epoches, out, device, args.ratio, args.min_len, args.mode) + args.batchsize, args.epoches, out, device, args.ratio, args.min_len, args.mode, predictor=args.predictor) if args.cmd == 'bin': @@ -918,7 +928,7 @@ def main(): set_random_seed(args.random_seed) binning(logger, args.num_process, args.data, args.max_edges, args.max_node, args.minfasta_kb * 1000, binned_length, - contig_dict, args.recluster, args.model_path, args.random_seed,out, device, args.environment) + contig_dict, args.recluster, args.model_path, args.random_seed,out, device, args.environment, predictor=args.predictor) if args.cmd == 'single_easy_bin': @@ -935,7 +945,8 @@ def main(): args.random_seed, out, device, - args.environment) + args.environment, + predictor=args.predictor) if args.cmd == 'multi_easy_bin': check_install() @@ -947,7 +958,8 @@ def main(): args.recluster, args.random_seed, out, - device) + device, + predictor=args.predictor) if __name__ == '__main__': diff --git a/SemiBin/semi_supervised_model.py b/SemiBin/semi_supervised_model.py index ab26b2c..e408080 100644 --- a/SemiBin/semi_supervised_model.py +++ b/SemiBin/semi_supervised_model.py @@ -139,7 +139,7 @@ def __len__(self): def train(out, contig_fastas, binned_lengths, logger, datas, data_splits, cannot_links, is_combined=True, - batchsize=2048, epoches=20, device=None, num_process = 8, mode = 'single'): + batchsize=2048, epoches=20, device=None, num_process = 8, mode = 'single', predictor = 'prodigal'): """ Train model from one sample(--mode single) or several samples(--mode several) """ @@ -171,7 +171,8 @@ def train(out, contig_fastas, binned_lengths, logger, datas, data_splits, cannot contig_fastas[data_index], binned_length=binned_lengths[data_index], num_process=num_process, - output=out) + output=out, + predictor=predictor) if epoch == 0: logger.info('Generate training data of {}:'.format(data_index)) diff --git a/SemiBin/utils.py b/SemiBin/utils.py index 59e8dde..7682731 100644 --- a/SemiBin/utils.py +++ b/SemiBin/utils.py @@ -28,6 +28,15 @@ def expect_file_list(fs): args.num_process = multiprocessing.cpu_count() logger.info(f'Setting number of CPUs to {args.num_process}') + if args.num_process > multiprocessing.cpu_count(): + args.num_process = multiprocessing.cpu_count() + + if args.cmd in ['single_easy_bin', 'multi_easy_bin', 'train', 'bin']: + if args.predictor not in ['prodigal', 'fraggenescan']: + sys.stderr.write( + f"Error: SemiBin support prodigal or fraggenescan gene predictor.\n") + sys.exit(1) + if args.cmd == 'generate_cannot_links': if args.GTDB_reference is not None: expect_file(args.GTDB_reference) @@ -199,7 +208,7 @@ def generate_cannot_link(mmseqs_path, namelist, num_threshold, output,sample): 'TIGR02386': 'TIGR02387', } -def get_marker(hmmout, fasta_path=None, min_contig_len=None, multi_mode=False): +def get_marker(hmmout, fasta_path=None, min_contig_len=None, multi_mode=False, predictor = None): import pandas as pd data = pd.read_table(hmmout, sep=r'\s+', comment='#', header=None, usecols=(0,3,5,15,16), names=['orf', 'gene', 'qlen', 'qstart', 'qend']) @@ -209,7 +218,10 @@ def get_marker(hmmout, fasta_path=None, min_contig_len=None, multi_mode=False): qlen = data[['gene','qlen']].drop_duplicates().set_index('gene')['qlen'] def contig_name(ell): - contig,_,_,_ = ell.rsplit( '_', 3) + if predictor == 'prodigal': + contig,_ = ell.rsplit( '_', 1) + else: + contig,_,_,_ = ell.rsplit( '_', 3) return contig data = data.query('(qend - qstart) / qlen > 0.4').copy() @@ -244,24 +256,69 @@ def extract_seeds(vs, sel): counts = data.groupby('gene')['orf'].count() return extract_seeds(counts, data) -def cal_num_bins(fasta_path, binned_length, num_process, multi_mode=False, output = None): - '''Estimate number of bins from a FASTA file +def prodigal(contig_file, contig_output): + with open(contig_output + '.out', 'w') as prodigal_out_log: + subprocess.check_call( + ['prodigal', + '-i', contig_file, + '-p', 'meta', + '-q', + '-a', contig_output + ], + stdout=prodigal_out_log, + ) - Parameters - fasta_path: path - binned_length: int (minimal contig length) - num_process: int (number of CPUs to use) - multi_mode: bool, optional (if True, treat input as resulting from concatenating multiple files) - ''' - with tempfile.TemporaryDirectory() as tdir: - if output is not None: - if os.path.exists(os.path.join(output, 'markers.hmmout')): - return get_marker(os.path.join(output, 'markers.hmmout'), fasta_path, binned_length, multi_mode) - else: - target_dir = output - else: - target_dir = tdir - contig_output = os.path.join(tdir, 'contigs.faa') +def run_prodigal(fasta_path, num_process, output): + from .error import LoggingPool + + contig_dict = {} + for h, seq in fasta_iter(fasta_path): + contig_dict[h] = seq + + num_contig = len(contig_dict) + + num_split = num_contig // num_process + split_contig_dict = {} + split_index = 0 + for h in contig_dict: + if split_index not in split_contig_dict: + split_contig_dict[split_index] = [] + split_contig_dict[split_index].append(h) + if split_index < num_process - 1: + if len(split_contig_dict[split_index]) == num_split: + split_index += 1 + + ### split_fasta + for index in range(num_process): + with open(os.path.join(output, 'contig_{}.fa'.format(index)), 'wt') as concat_out: + for h in split_contig_dict[index]: + concat_out.write(f'>{h}\n{contig_dict[h]}\n') + + with LoggingPool(num_process) if num_process != 0 else LoggingPool() as pool: + try: + for index in range(num_process): + pool.apply_async( + prodigal, + args=( + os.path.join(output, 'contig_{}.fa'.format(index)), + os.path.join(output, 'contig_{}.faa'.format(index)), + )) + pool.close() + pool.join() + except: + sys.stderr.write( + f"Error: Running prodigal fail\n") + sys.exit(1) + + contig_output = os.path.join(output, 'contigs.faa') + with open(contig_output, 'w') as f: + for index in range(num_process): + f.write(open(os.path.join(output, 'contig_{}.faa'.format(index)), 'r').read()) + return contig_output + +def run_fraggengscan(fasta_path, num_process, output): + try: + contig_output = os.path.join(output, 'contigs.faa') with open(contig_output + '.out', 'w') as frag_out_log: # We need to call FragGeneScan instead of the Perl wrapper because the # Perl wrapper does not handle filepaths correctly if they contain spaces @@ -277,22 +334,57 @@ def cal_num_bins(fasta_path, binned_length, num_process, multi_mode=False, outpu ], stdout=frag_out_log, ) + except: + sys.stderr.write( + f"Error: Running fraggenescan fail\n") + sys.exit(1) + return contig_output + +def cal_num_bins(fasta_path, binned_length, num_process, multi_mode=False, output = None, predictor = 'prodigal'): + '''Estimate number of bins from a FASTA file + + Parameters + fasta_path: path + binned_length: int (minimal contig length) + num_process: int (number of CPUs to use) + multi_mode: bool, optional (if True, treat input as resulting from concatenating multiple files) + ''' + with tempfile.TemporaryDirectory() as tdir: + if output is not None: + if os.path.exists(os.path.join(output, 'markers.hmmout')): + return get_marker(os.path.join(output, 'markers.hmmout'), fasta_path, binned_length, multi_mode, predictor=predictor) + else: + target_dir = output + else: + target_dir = tdir + + if predictor == 'prodigal': + contig_output = run_prodigal(fasta_path, num_process, tdir) + else: + contig_output = run_fraggengscan(fasta_path, num_process, tdir) hmm_output = os.path.join(target_dir, 'markers.hmmout') - with open(os.path.join(tdir, 'markers.hmmout.out'), 'w') as hmm_out_log: - subprocess.check_call( - ['hmmsearch', - '--domtblout', - hmm_output, - '--cut_tc', - '--cpu', str(num_process), - os.path.split(__file__)[0] + '/marker.hmm', - contig_output + '.faa', - ], - stdout=hmm_out_log, - ) + try: + with open(os.path.join(tdir, 'markers.hmmout.out'), 'w') as hmm_out_log: + subprocess.check_call( + ['hmmsearch', + '--domtblout', + hmm_output, + '--cut_tc', + '--cpu', str(num_process), + os.path.split(__file__)[0] + '/marker.hmm', + contig_output if predictor == 'prodigal' else contig_output + '.faa', + ], + stdout=hmm_out_log, + ) + except: + sys.stderr.write( + f"Error: Running hmmsearch fail\n") + sys.exit(1) + + marker = get_marker(hmm_output, fasta_path, binned_length, multi_mode, predictor=predictor) - return get_marker(hmm_output, fasta_path, binned_length, multi_mode) + return marker def write_bins(namelist, contig_labels, output, contig_dict, diff --git a/docs/index.md b/docs/index.md index 2a93721..f3e3e10 100644 --- a/docs/index.md +++ b/docs/index.md @@ -57,6 +57,7 @@ Reconstruct bins with single or co-assembly binning using one line command. * `--ml-threshold` : Length threshold for generating must-link constraints.(By default, the threshold is calculated from the contig, and the default minimum value is 4,000 bp) * `--no-recluster` : Do not recluster bins. * `--taxonomy-annotation-table` : TAXONOMY_TSV, Pre-computed mmseqs2 format taxonomy TSV file to bypass mmseqs2 GTDB annotation [advanced] +* `--predictor` : gene predictor used to estimate the number of bins(prodigal/fraggenescan) #### multi_easy_bin @@ -73,7 +74,7 @@ The following options (including synonyms) are the same as for Run the contig annotations using mmseqs with GTDB and generate cannot-link file used in the semi-supervised deep learning model training. -The following options are the same as for `single_easy_bin`: `-i/--input-fasta`, `-o/--output`, `--cannot-name`, `-r/--reference-db-data-dir`, `--ratio`, `--min-len`, `--ml-threshold` and `--taxonomy-annotation-table`. +The following options are the same as for `single_easy_bin`: `-i/--input-fasta`, `-o/--output`, `--cannot-name`, `-r/--reference-db-data-dir`, `--ratio`, `--min-len`, `--ml-threshold`, `--taxonomy-annotation-table` and `--predictor`. #### generate_sequence_features_single @@ -98,7 +99,7 @@ Training the model. * `-c/--cannot-link` : Path to the input cannot link file generated from other additional biological information, one row for each cannot link constraint. The file format: contig_1,contig_2. * `--mode`: [single/several] Train models from one sample or several samples(train model across several samples can get better pre-trained model for single-sample binning.) In several mode, must input data, data_split, cannot, fasta files for corresponding sample with same order. Note: You can just set `several` with this option when single-sample binning. Training from several samples with multi-sample binning is not support. -The following options are the same as for `single_easy_bin`: `-i/--input-fasta`, `-o/--output`, `--epoches`, `--batch-size`, `-p/--processes/-t/--threads`, `--random-seed`, `--ratio`, `--min-len`. +The following options are the same as for `single_easy_bin`: `-i/--input-fasta`, `-o/--output`, `--epoches`, `--batch-size`, `-p/--processes/-t/--threads`, `--random-seed`, `--ratio`, `--min-len` and `--predictor`. #### bin @@ -106,7 +107,7 @@ Cluster contigs into bins. * `--model`: Path to the trained model. -The following options are the same as for `single_easy_bin`: `--data`,`-i/--input-fasta`, `-o/--output`, `--minfasta-kbs`, `--recluster`, `--max-node`, `--max-edges`, `-p/--processes/-t/--threads`, `--random-seed`, `--environment`, `--ratio`, `--min-len` and `--no-recluster`. +The following options are the same as for `single_easy_bin`: `--data`,`-i/--input-fasta`, `-o/--output`, `--minfasta-kbs`, `--recluster`, `--max-node`, `--max-edges`, `-p/--processes/-t/--threads`, `--random-seed`, `--environment`, `--ratio`, `--min-len`, `--no-recluster` and `--predictor`. #### download_GTDB	Substituting the ORF finder Substitute the ORF finder (FragGeneScan) with a better one (Prodigal).	Background: https://github.com/BigDataBiology/SemiBin/issues/73	2022-02-21T05:18:01	[]	[]
emdgroup/tnmf	emdgroup__tnmf-51	0d1bbcc0a4e7763779f1d412cd7e0721806c9823	diff --git a/CHANGELOG.md b/CHANGELOG.md index 43b664c3..8d1ce64e 100644 --- a/CHANGELOG.md +++ b/CHANGELOG.md @@ -9,6 +9,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0 - Unit test Code Coverage Reporting - Demo: Racoon image - Include examples in pytest bench +- Include demos in pytest bench ### Changed - Fixed internal TODOs to improve code quality diff --git a/demos/demo_image.py b/demos/demo_image.py index cd79dc72..359b62e2 100644 --- a/demos/demo_image.py +++ b/demos/demo_image.py @@ -24,7 +24,7 @@ def st_define_sample_params(verbose: bool = True) -> Tuple[dict, float]: help_scale = \ '''Downscaling the image allows for quicker interactive experimentation. Note that this does not change atom size.''' - scale = st.sidebar.number_input('Image Scale', min_value=0.05, value=0.25, step=0.05, help=help_scale) + scale = st.sidebar.number_input('Image scale', min_value=0.05, value=0.25, step=0.05, help=help_scale) if verbose: st.sidebar.caption(help_scale) diff --git a/demos/demo_selector.py b/demos/demo_selector.py index c552f010..d1aeface 100644 --- a/demos/demo_selector.py +++ b/demos/demo_selector.py @@ -1,13 +1,11 @@ +import sys + import importlib.resources from importlib import import_module import numpy as np import streamlit as st -# show TNMF header -with importlib.resources.path('logos', 'tnmf_header.png') as img_file: - st.image(str(img_file), use_column_width='always') - # mapping of streamlit demo names to the corresponding demo file and optional parameters DEMO_NAME_DICT = { '1-D Synthetic Signals': ('synthetic_signals', {'n_dims': 1}), @@ -15,43 +13,57 @@ 'Racoon Image': ('demo_image', {}), } -# create progress bar on the top that is shown for all demos -progress_bar = st.sidebar.progress(1.) - -# toggle verbose mode -help_verbose = 'Displays / hides detailed explanations.' -verbose = st.sidebar.checkbox('Verbose', True, help=help_verbose) -if verbose: - st.sidebar.caption(help_verbose) - -# select the demo -help_select_demo = 'The specific demo example that gets executed.' -selected_demo = st.sidebar.selectbox('Demo example', list(DEMO_NAME_DICT.keys()), 1, help=help_select_demo) -if verbose: - st.sidebar.caption(help_select_demo) - -# select the random seed -help_seed = 'The fixed random seed that is used for the simulation.' -seed = st.sidebar.number_input('Random seed', value=42, help=help_seed) -np.random.seed(seed) -if verbose: - st.sidebar.caption(help_seed) - -# show demo info text -if verbose: - st.markdown(''' - ## Demo guide - This dashboard demonstrates the use of the Transform-Invariant Non-Negative Matrix Factorization (TNMF) package - in the specific context of learning shift-invariant representations. - Via the drop-down menu on the left, you can choose between different demo examples. - - ## Usage - Detailed explanations of the available options and the shown output (including this text) can be displayed or hidden - using the 'Verbose' checkbox. When the checkbox is unticked, the description of the control widgets can still - be accessed via the tooltips next to them. - ''') - -# extract the demo name and parameters, import the corresponding module, and execute the demo -demo_name, demo_args = DEMO_NAME_DICT[selected_demo] -demo_module = import_module(demo_name) -demo_module.main(progress_bar, verbose=verbose, demo_args) + +def main(demo_name: str): + # show TNMF header + with importlib.resources.path('logos', 'tnmf_header.png') as img_file: + st.image(str(img_file), use_column_width='always') + + # create progress bar on the top that is shown for all demos + progress_bar = st.sidebar.progress(1.) + + # toggle verbose mode + help_verbose = 'Displays / hides detailed explanations.' + verbose = st.sidebar.checkbox('Verbose', True, help=help_verbose) + if verbose: + st.sidebar.caption(help_verbose) + + # select the demo + help_select_demo = 'The specific demo example that gets executed.' + default_demo = list(DEMO_NAME_DICT.keys()).index(demo_name) + selected_demo = st.sidebar.selectbox('Demo example', list(DEMO_NAME_DICT.keys()), default_demo, help=help_select_demo) + if verbose: + st.sidebar.caption(help_select_demo) + + # select the random seed + help_seed = 'The fixed random seed** that is used for the simulation.' + seed = st.sidebar.number_input('Random seed', value=42, help=help_seed) + np.random.seed(seed) + if verbose: + st.sidebar.caption(help_seed) + + # show demo info text + if verbose: + st.markdown(''' + ## Demo guide + This dashboard demonstrates the use of the Transform-Invariant Non-Negative Matrix Factorization (TNMF) package + in the specific context of learning shift-invariant representations. + Via the drop-down menu on the left, you can choose between different demo examples. + + ## Usage + Detailed explanations of the available options and the shown output (including this text) can be displayed or hidden + using the 'Verbose' checkbox. When the checkbox is unticked, the description of the control widgets can still + be accessed via the tooltips next to them. + ''') + + # extract the demo name and parameters, import the corresponding module, and execute the demo + demo_name, demo_args = DEMO_NAME_DICT[selected_demo] + demo_module = import_module(demo_name) + demo_module.main(progress_bar, verbose=verbose, *demo_args) + + +if __name__ == '__main__': + if len(sys.argv) > 1: + main(sys.argv[1]) + else: + main('2-D Synthetic Signals') diff --git a/scripts/tnmf.py b/scripts/tnmf.py index 4ddbea64..d0fdfc3c 100644 --- a/scripts/tnmf.py +++ b/scripts/tnmf.py @@ -3,6 +3,7 @@ from glob import glob from pathlib import Path from typing import List +from demos.demo_selector import DEMO_NAME_DICT DEMO_FILE = Path(__file__).absolute().parent.parent / 'demos' / 'demo_selector.py' EXAMPLES_FOLDER = Path(__file__).absolute().parent.parent / 'examples' @@ -19,6 +20,11 @@ def get_examples() -> List[Path]: return examples +def get_demos() -> List[str]: + """Returns a list containing the names of all demos.""" + return list(DEMO_NAME_DICT.keys()) + + def main(): """Entry point for the `tnmf` command.""" @@ -27,10 +33,12 @@ def main(): # create subparsers for the `example` and `demo` commands subparsers = parser.add_subparsers(dest='command') - demo_parser = subparsers.add_parser("demo", help='Runs the streamlit demo.') # noqa: F841 + demo_parser = subparsers.add_parser("demo", help='Runs the streamlit demo. Type `tnmf demo -h` to show the list ' + 'of available demos.') + demo_parser.add_argument("demo_name", choices=get_demos(), help='Name of the demo.', default=None, nargs='?') + example_parser = subparsers.add_parser("example", help='Runs a specific example. Type `tnmf example -h` to show the list ' 'of available examples.') - examples = get_examples() example_parser.add_argument("example_name", choices=[e.stem for e in examples], help='Name of the example.') @@ -41,8 +49,12 @@ def main(): if args.command == 'example': args_example = parser.parse_args() example = [e for e in examples if e.stem == args_example.example_name][0] - os.system(f'python {example}') # noqa: S605 + os.system(f'python {example}') # noqa: S605, DUO106 - # run the demo + # parse the (optional) demo name and run the demo elif args.command == 'demo': - os.system(f'streamlit run {DEMO_FILE}') # noqa: S605 + args_demo = parser.parse_args() + if args_demo.demo_name is not None: + os.system(f'streamlit run {DEMO_FILE} "{args_demo.demo_name}"') # noqa: S605, DUO106 + else: + os.system(f'streamlit run {DEMO_FILE}') # noqa: S605 diff --git a/setup.cfg b/setup.cfg index fde5a601..b0445eaa 100644 --- a/setup.cfg +++ b/setup.cfg @@ -105,7 +105,7 @@ generated-members=numpy., torch.* source = tnmf examples - # TODO: add 'demos' here + demos branch = True command_line = -m pytest # create individual files per process that need to be combined afterwards @@ -115,7 +115,7 @@ parallel = True [coverage:report] # Target code coverage in integer percent. If overall coverage is below, coverage.py reports failure. # TODO: should gradually increase this -fail_under = 67 +fail_under = 90 # Do not report files with coverage at 100% skip_covered = True # Do not report empty files diff --git a/tnmf/utils/demo.py b/tnmf/utils/demo.py index 6e083ef0..bf75e096 100644 --- a/tnmf/utils/demo.py +++ b/tnmf/utils/demo.py @@ -58,8 +58,7 @@ def st_define_nmf_params(default_params: dict, have_ground_truth: bool = True, v n_atoms_default = default_params['n_atoms'] n_atoms = st.sidebar.number_input('# Atoms', value=n_atoms_default, min_value=1, help=help_n_atoms) if not use_n_atoms else n_atoms_default - if verbose and not use_n_atoms: - st.sidebar.caption(help_n_atoms) + explanation(help_n_atoms, verbose and not use_n_atoms) help_atom_shape = 'The size of each atom dimension.' if have_ground_truth: @@ -78,14 +77,12 @@ def st_define_nmf_params(default_params: dict, have_ground_truth: bool = True, v value=default_atom_shape[0], min_value=1, help=help_atom_shape)] * len(default_params['atom_shape']) ) if not use_atom_shape else default_atom_shape - if not use_atom_shape: - explanation(help_atom_shape, verbose) + explanation(help_atom_shape, verbose and not use_atom_shape) help_sparsity_H = 'The strength of the L1 activation sparsity regularization imposed on the optimization problem.' sparsity_H = st.sidebar.number_input('Activation sparsity', min_value=0.0, value=0.0, step=0.01, help=help_sparsity_H) - if verbose: - st.sidebar.caption(help_sparsity_H) + explanation(help_sparsity_H, verbose) help_inhibition_strength = \ '''The strength of the same-atom lateral activation sparsity regularization imposed on the optimization problem. @@ -107,8 +104,7 @@ def st_define_nmf_params(default_params: dict, have_ground_truth: bool = True, v help_n_iterations = '''The number of multiplicative updates to the atom dictionary and activation tensors.''' n_iterations = st.sidebar.number_input('# Iterations', value=100, min_value=1, help=help_n_iterations) - if verbose: - st.sidebar.caption(help_n_iterations) + explanation(help_n_iterations, verbose) help_backend = \ '''The optimization backend for computing the multiplicative gradients.	Need to unit-test examples and demo Currently, we would not recognize if these fail, e.g. due to interface incompatibilities. Easiest might be to create dummy unit tests that set matplotlib into non-interactive mode and run the examples just to check that nothing fails horribly (no result verification).		2021-07-28T12:59:42	[]	[]
emdgroup/tnmf	emdgroup__tnmf-49	0315335e225f2b84cc98404b04163bd3c1e48d20	diff --git a/CHANGELOG.md b/CHANGELOG.md index 9fd1267b..409ca3e9 100644 --- a/CHANGELOG.md +++ b/CHANGELOG.md @@ -8,6 +8,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0 ### Added - Unit test Code Coverage Reporting - Demo: Racoon image +- Include examples in pytest bench ## [0.1.1] - 2021-07-22 ### Added	Need to unit-test examples and demo Currently, we would not recognize if these fail, e.g. due to interface incompatibilities. Easiest might be to create dummy unit tests that set matplotlib into non-interactive mode and run the examples just to check that nothing fails horribly (no result verification).		2021-07-20T22:11:25	[]	[]
emdgroup/tnmf	emdgroup__tnmf-43	be5db34d6271280d5645d4323ec019bf4456438a	diff --git a/.gitignore b/.gitignore index 25694cc9..3f25472e 100644 --- a/.gitignore +++ b/.gitignore @@ -20,6 +20,7 @@ _build .egg-info build dist +htmlcov # Package Version tnmf/_version.py diff --git a/CHANGELOG.md b/CHANGELOG.md index 368ee224..43f79f6a 100644 --- a/CHANGELOG.md +++ b/CHANGELOG.md @@ -4,6 +4,10 @@ All notable changes to this project will be documented in this file. The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/), and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html). +## Unreleased +### Added +- Unit test Code Coverage Reporting + ## [0.1.1] - 2021-07-22 ### Added - Bugfix: added logo to package content diff --git a/README.md b/README.md index e68a1cdc..579e8951 100644 --- a/README.md +++ b/README.md @@ -1,6 +1,6 @@ [![Flake8 Linter](https://github.com/emdgroup/tnmf/actions/workflows/flake8.yml/badge.svg)](https://github.com/emdgroup/tnmf/actions/workflows/flake8.yml) [![Pylint Linter](https://github.com/emdgroup/tnmf/actions/workflows/pylint.yml/badge.svg)](https://github.com/emdgroup/tnmf/actions/workflows/pylint.yml) -[![Pytest](https://github.com/emdgroup/tnmf/actions/workflows/pytest.yml/badge.svg)](https://github.com/emdgroup/tnmf/actions/workflows/pytest.yml) +[![Pytest and Coverage](https://github.com/emdgroup/tnmf/actions/workflows/pytest.yml/badge.svg)](https://github.com/emdgroup/tnmf/actions/workflows/pytest.yml) [![Build Documentation](https://github.com/emdgroup/tnmf/actions/workflows/sphinx.yml/badge.svg)](https://github.com/emdgroup/tnmf/actions/workflows/sphinx.yml) [![Publish to PyPI](https://github.com/emdgroup/tnmf/actions/workflows/publish-to-pypi.yml/badge.svg)](https://github.com/emdgroup/tnmf/actions/workflows/publish-to-pypi.yml) [![Open in Streamlit](https://static.streamlit.io/badges/streamlit_badge_black_white.svg)](https://share.streamlit.io/adriansosic/tnmf/main/demos/demo_selector.py) @@ -76,16 +76,51 @@ Now, you should be able to execute the unit tests by calling `pytest` to verify ## Pull Requests Before creating a pull request, you should always try to ensure that the automated code quality and unit tests do not fail. -To execute them, change into the repository root directory and run the following commands: +This section explains how to run them locally to understand and fix potential issues. + +### Code Style and Quality +Code style and quality are checked using [flake8](https://flake8.pycqa.org/) and [pylint](http://pylint.pycqa.org/). +To execute them, change into the [repository root directory](.), run the following commands and inspect their output: ``` flake8 pylint tnmf +``` + +In order for a pull request to be accaptable, no errors may be reported here. + +### Unit Tests +Automated unit tests reside inside the folder [tnmf/tests](tnmf/tests). They can be executed via +[pytest](https://docs.pytest.org/) by changing into the [repository root directory](.) and running + +``` pytest ``` -The output of the individual commands should be pretty instructive, so fixing potential issues is usually rather straightforward. +Debugging potential failures from the command line might be cumbersome. +Most Python IDEs, however, also support `pytest` natively in their debugger. +Again, for a pull request to be acceptable, no failures may be reported here. + +### Code Coverage +Code coverage in the unit tests is measured using [coverage](https://coverage.readthedocs.io). +A coverage report can be created locally from the [repository root directory](.) via + +``` +coverage run +coverage report +``` + +This will output a concise table with an overview of python files that are not fully covered with unit tests along with the line numbers of code that has not been executed. +A more detailed, interactive report can be created using + +``` +coverage html +``` + +Then, you can open the file `htmlcov/index.html` in a web browser of your choice to navigate through code annotated with coverage data. +Required overall coverage to is configured in [setup.cfg](setup.cfg), under the key `fail_under` in section `[coverage:report]`. + ## Building the Documentation -To build the documentation locally, change into the `doc` subdirectory and run `make html`. +To build the documentation locally, change into the [doc subdirectory](doc) and run `make html`. Then, the documentation resides at `doc\_build\html\index.html`. diff --git a/requirements.txt b/requirements.txt index b6016bb6..f5e3e4ed 100644 --- a/requirements.txt +++ b/requirements.txt @@ -8,8 +8,8 @@ streamlit==0.84.0 torch==1.8.1 # Development and Tests +coverage==5.5 pytest==6.2.3 -pytest-cov==2.11.1 flake8==3.9.1 flake8-bandit==2.1.2 flake8-bugbear==21.4.3 diff --git a/setup.cfg b/setup.cfg index 2ad7f607..e133821d 100644 --- a/setup.cfg +++ b/setup.cfg @@ -46,11 +46,8 @@ test = pytest [tool:pytest] log_cli = True -# TODO: temporarily disable coverage because it breaks PyCharm's debugger -# https://www.jetbrains.com/help/pycharm/run-debug-configuration-py-test.html addopts = --ignore doc -# --cov=tnmf [flake8] exclude = .git,__pycache__,doc,env,venv,.venv @@ -103,3 +100,30 @@ bad-names-rgxs=_?num_. # List of members which are set dynamically and missed by Pylint inference # system, and so shouldn't trigger E1101 when accessed. generated-members=numpy., torch. + +[coverage:run] +source = tnmf +branch = True +command_line = -m pytest + +[coverage:report] +# Target code coverage in integer percent. If overall coverage is below, coverage.py reports failure. +# TODO: should gradually increase this +fail_under = 64 +# Do not report files with coverage at 100% +skip_covered = True +# Do not report empty files +skip_empty = True +# List line numbers in the report table +show_missing = True +# exclusion list (regular expressions) +exclude_lines = + pragma: no cover + raise NotImplementedError + raise AssertionError + raise ValueError + except ImportError: + except PackageNotFoundError: +# file patterns to ignore +omit = + tnmf/_version.py \ No newline at end of file diff --git a/tnmf/backends/NumPy_CachingFFT.py b/tnmf/backends/NumPy_CachingFFT.py index ba15552c..d0a1401a 100644 --- a/tnmf/backends/NumPy_CachingFFT.py +++ b/tnmf/backends/NumPy_CachingFFT.py @@ -58,36 +58,26 @@ def __itruediv__(self, other): self.c /= other.c if isinstance(other, CachingFFT) else other return self - def set_fft_params(self, fft_axes: Tuple[int, ...], fft_shape: Tuple[int, ...]): - self._fft_axes = fft_axes - self._fft_shape = fft_shape - def _invalidate(self): self._c = None self._f = None self._f_padded = None self._f_reversed = None + @property def has_c(self) -> bool: """Check if the field in coordinate space has already been computed""" return self._c is not None + @property def has_f(self) -> bool: """Check if the field in fourier space has already been computed""" return self._f is not None - @property - def shape(self) -> Tuple[int]: - return self._c.shape - - @property - def ndim(self) -> int: - return self._c.ndim - @property def c(self) -> np.ndarray: """Getter for field in coordinate space""" - if self._c is None: + if not self.has_c: self._logger.debug(f'Computing {self._field_name}(x) = FFT^-1[ {self._field_name}(f) ]', ) assert self.has_f self._c = irfftn(self._f, axes=self._fft_axes, s=self._fft_shape, workers=self._fft_workers) @@ -103,9 +93,9 @@ def c(self, c: np.ndarray): @property def f(self) -> np.ndarray: """Getter for field in fourier space""" - if self._f is None: + if not self.has_f: self._logger.debug(f'Computing {self._field_name}(f) = FFT[ {self._field_name}(x) ]') - assert self.has_c() + assert self.has_c self._f = rfftn(self._c, axes=self._fft_axes, s=self._fft_shape, workers=self._fft_workers) return self._f @@ -113,7 +103,7 @@ def f_padded(self, pad_mode: Dict = None, pad_width: Tuple[Tuple[int, int], ...] """Getter for padded field in fourier space""" if self._f_padded is None: self._logger.debug(f'Computing {self._field_name}_padded(f) = FFT[ {self._field_name}_padded(x) ]') - assert self.has_c() + assert self.has_c c = np.pad(self._c, pad_width, **pad_mode) # TODO: we should actually make sure that the padding does not change between calls @@ -132,7 +122,7 @@ def f_reversed(self) -> np.ndarray: """Getter for time-reversed field in fourier space, intentionally no setter for now""" if self._f_reversed is None: self._logger.debug(f'Computing {self._field_name}_rev(f) = FFT[ {self._field_name}(-x) ]') - assert self.has_c() + assert self.has_c c_reversed = np.flip(self._c, axis=self._fft_axes) self._f_reversed = rfftn(c_reversed, axes=self._fft_axes, s=self._fft_shape, workers=self._fft_workers) return self._f_reversed diff --git a/tnmf/backends/NumPy_FFT.py b/tnmf/backends/NumPy_FFT.py index 5bd6fc61..87fed939 100644 --- a/tnmf/backends/NumPy_FFT.py +++ b/tnmf/backends/NumPy_FFT.py @@ -3,7 +3,7 @@ Shift-invariance is implemented via fast convolution in the Fourier domain using :func:`scipy.fft.rfftn` and :func:`scipy.fft.irfftn`. """ -from typing import Tuple, Dict, Union +from typing import Tuple, Dict import numpy as np from scipy.fft import next_fast_len, rfftn, irfftn @@ -12,7 +12,7 @@ def fftconvolve_sum( - in1: Union[np.ndarray, Tuple[np.ndarray, ...]], + in1: Tuple[np.ndarray, ...], in2: np.ndarray, fft_axes: Tuple[int, ...], slices: Tuple[slice, ...], @@ -43,10 +43,6 @@ def padded_rfftn( # Fourier transform (for real data) return rfftn(array_padded, axes=axes) - if not isinstance(in1, tuple): - assert isinstance(in1, np.ndarray) - in1 = (in1, ) - assert isinstance(in2, np.ndarray) ndim = in2.ndim	Code Coverage of unit tests Should be analyzed and monitored.		2021-07-12T17:10:05	[]	[]
Yikai-Liao/symusic	Yikai-Liao__symusic-11	1bb8873e3077b80051be59afbaf1dc4c52ce574a	diff --git a/setup.py b/setup.py index bf13886..4e36a1c 100644 --- a/setup.py +++ b/setup.py @@ -139,8 +139,8 @@ def build_extension(self, ext: CMakeExtension) -> None: long_description=open("README.md", encoding="utf-8").read(), long_description_content_type="text/markdown", license=open("LICENSE", encoding="utf-8").read(), - package_dir={"symusic": "py-symusic"}, - packages=["py-symusic"], + package_dir={"symusic": "symusic"}, + packages=["symusic"], ext_modules=[CMakeExtension("symusic.core", ".")], cmdclass={"build_ext": CMakeBuild}, package_data={"symusic": ["*/.pyi"]}, diff --git a/py-symusic/__init__.py b/symusic/__init__.py similarity index 100% rename from py-symusic/__init__.py rename to symusic/__init__.py diff --git a/py-symusic/core.pyi b/symusic/core.pyi similarity index 100% rename from py-symusic/core.pyi rename to symusic/core.pyi diff --git a/py-symusic/factory.py b/symusic/factory.py similarity index 100% rename from py-symusic/factory.py rename to symusic/factory.py diff --git a/py-symusic/types.py b/symusic/types.py similarity index 100% rename from py-symusic/types.py rename to symusic/types.py	Tests CI workflows Following #4 , this PR set up GitHub actions to run the tests on pushes and pull requests. Not ready for merge yet. Right now the action only perform builds as done in the "publish" action	Also, the tests pass for 4 MIDIs out of 26, I'll try to investigate why There is a time limitation of github actions. So we'd better reduce some platform (like linux arm64 based on qemu and macos) and some python versions for test. <img width="1072" alt="image" src="https://github.com/Yikai-Liao/symusic/assets/110762732/f0234b30-62d1-4e88-b456-635f6656ccba"> Absolutely, apologies for this unexpected situation 😅 In the next commit I'll temporarily disable the action anyway @Yikai-Liao @lzqlzzq I updated the `tests.yml` workflow, removing all the builds parts and reducing the number of platforms and python versions. It is possible to run the tests on macOS arm runners, but as the [pricing / hours multiplicator](https://github.com/actions/runner-images/issues/8439) is high it might be a better idea to not test it. Should we however use QEMU with linux to test on arm64? Hopefully the tests should run very quick, they only take ~10min for miditoolkit, here the install should be a bit longer but the rest should be fast. It might also be a good idea to rename the symusic directory, as it prevents to import the installed symusic module when running tests locally (from the root project directory). What do you think? Well, Linux containers using qemu was about `8 times slower` than x64 containers (in our previous pratice). Thankfully, we don't have that many bugs to fix right now, which means the pace of releases has slowed down considerably. (I'm currently focusing on design `libsymusic`, which is extracted from the single header file in `symusic` with more friendly c++ interface and will be easier to maintain and extend) Therefore, the test can be done without worrying too much about exceeding the total time limit of github actions. starting next month. And by the way, my high-performance arm development board (8 cores, 16g RAM and 256g eMMC) has arrived 🎉! And I'll test symusic on it frequently. > It might also be a good idea to rename the symusic directory, as it prevents to import the installed symusic module when running tests locally (from the root project directory). What do you think? I agree with it. The problem is that I'm not familiar with how `setup.py` is actually working. And the reason why I named that folder symusic is that it can be copied to installation path directly(without renaming it). I just don't know how to achieve more complex operations. That's great news! So indeed I think we can safely reduce the range of platforms / python versions of tests. I can take a look for the setup.py behaviour, and if it's not complicated include the changes in this PR. BTW, in the future switching to a `pyproject.toml` package description could be considered (setup.py is legacy), if this is possible as I'm not sure the bindings are supported. BTW, I just updated the test, with the right sorting only 8 are failing now! 🎉 There seem to be two sorts of problems that should be easy to fix: 1. A mismatch between note velocity and duration. This surely comes from a FIFO error, I encountered a similar issue with miditoolkit. And that's the same error causing the last miditok test to fail; 2. Tempo assertion errors, with tempos being equal but the `.tempo` value is a fraction different, likely to come from a float type conversion error somewhere between the original MIDI format and the one written. I don't know if it can be fixed, but if it cannot I suggest to rounds `.tempo` when parsing a MIDI, or to do the rounding in `Tempo.__equal__`. <img width="477" alt="Capture d’écran 2023-12-19 à 17 32 54" src="https://github.com/Yikai-Liao/symusic/assets/56734983/684773b1-1b53-4ac5-a611-fa37179677e5"> For reference on the FIFO issue when writing a MIDI: https://github.com/YatingMusic/miditoolkit/pull/28 And how the FIFO loading was added: https://github.com/YatingMusic/miditoolkit/pull/14 For `setup.py`, we just need to change the root directory name for the `packages` and `package_dir` entries. Would you like it to be done in this PR? And if so, which new name should we give to the dir? * First, for the new folder name of `symusic`, I have checked some pupular repository: \| repository \| folder name \| \| -- \| -- \| \| [numpy](https://github.com/numpy/numpy)\| numpy \| \| [pytorch](https://github.com/pytorch/pytorch) \| torch \| \| [mido](https://github.com/mido/mido) \| mido \| \| [polars](https://github.com/pola-rs/polars) \| py-polars \| \| [matplotlib](https://github.com/matplotlib/matplotlib) \| lib/matplotlib \| So, what about `py-symusic`? (BTW, why do many libraries just use the same name as the folder name? ) * Second, I have understood what's wrong with currently `midi writing` strategy. However, I'm still confused, becuase midi also support real time applications. If we send two `note-on` events in the same channel, with the same pitch (the only difference lies their velocity), then, how can we decide their order based on the given information to bring it into line with the `FIFO` rule? And is there any standard describing this rule in midi? Actually, it feels more like an undefined behavior to me. * Third, for the `__eq__` of `tempo`, introduing an epsilon in compartion might cause lots of unexpected problems according to this [anwser](https://stackoverflow.com/questions/17333/how-do-you-compare-float-and-double-while-accounting-for-precision-loss) in stack overflow. At the same time, I don't find standard for bpm quantization. In short I think this issue should be handled with care. * Ok for `py-symusic`. I don't why they keep the same name when the package has to be built before being tested, but they surely have some way to handle it; * Indeed there is no specific recommendation in the MIDI specifications when to it comes to the order of incoming events. mido is not FIFO if I remember well. However I think it would be good practice to at least keep a consistency between how the MIDI information is parsed and how it is written, either FIFO or LIFO, in order to keep the same information when dumping back; * Noted. I suggest to maybe merge this branch as it comes with a lot of file changes, with minimal test platforms activated, before solving the above issues? I fail to merge. > refusing to allow an OAuth App to create or update workflow `.github/workflows/lint.yml` without workflow`scope I'm not sure what's the cause of this issue, if it's not solves yet I'll look into it after making sure the pytest action is run properly (8 tests should fail, right now I don't know why it cannot find the pytest command). It works on browser, but not on my phone Ok, there will still be one or two things to fix for the action to run properly.	2023-12-21T16:45:09	[]	[]
go-python/gopy	go-python__gopy-321	d342bb85d1c2f1abdf90781820bcfe726e3f8a8c	diff --git a/.github/workflows/ci.yml b/.github/workflows/ci.yml index 5f1dbd4..5be3246 100644 --- a/.github/workflows/ci.yml +++ b/.github/workflows/ci.yml @@ -22,7 +22,7 @@ jobs: name: Build strategy: matrix: - go-version: [1.19.x, 1.18.x] + go-version: [1.20.x, 1.19.x] platform: [ubuntu-latest] #platform: [ubuntu-latest, macos-latest, windows-latest] runs-on: ${{ matrix.platform }} @@ -52,26 +52,17 @@ jobs: if: matrix.platform == 'ubuntu-latest' run: \| sudo apt-get update - sudo apt-get install curl libffi-dev python-cffi python3-cffi python3-pip + sudo apt-get install curl libffi-dev python3-cffi python3-pip # pypy3 isn't packaged in ubuntu yet. TEMPDIR=$(mktemp -d) - curl -L https://downloads.python.org/pypy/pypy2.7-${PYPYVERSION}-linux64.tar.bz2 --output $TEMPDIR/pypy2.tar.bz2 curl -L https://downloads.python.org/pypy/pypy3.6-${PYPYVERSION}-linux64.tar.bz2 --output $TEMPDIR/pypy3.tar.bz2 - tar xf $TEMPDIR/pypy2.tar.bz2 -C $TEMPDIR tar xf $TEMPDIR/pypy3.tar.bz2 -C $TEMPDIR - sudo ln -s $TEMPDIR/pypy2.7-$PYPYVERSION-linux64/bin/pypy /usr/local/bin/pypy sudo ln -s $TEMPDIR/pypy3.6-$PYPYVERSION-linux64/bin/pypy3 /usr/local/bin/pypy3 - # install pip (for pypy, python2) - curl -L https://bootstrap.pypa.io/pip/2.7/get-pip.py --output ${TEMPDIR}/get-pip2.py - python2 ${TEMPDIR}/get-pip2.py # curl -L https://bootstrap.pypa.io/get-pip.py --output ${TEMPDIR}/get-pip.py - # pypy ${TEMPDIR}/get-pip.py # pypy3 ${TEMPDIR}/get-pip.py # install pybindgen - python2 -m pip install --user -U pybindgen python3 -m pip install --user -U pybindgen - # pypy -m pip install --user -U pybindgen # pypy3 -m pip install --user -U pybindgen # install goimports diff --git a/SUPPORT_MATRIX.md b/SUPPORT_MATRIX.md index f23c8f6..9153644 100644 --- a/SUPPORT_MATRIX.md +++ b/SUPPORT_MATRIX.md @@ -3,31 +3,31 @@ NOTE: File auto-generated by TestCheckSupportMatrix in main_test.go. Please don't modify manually. -Feature \|py2 \| py3 ---- \| --- \| --- -_examples/arrays \| yes \| yes -_examples/cgo \| yes \| yes -_examples/consts \| yes \| yes -_examples/cstrings \| yes \| yes -_examples/empty \| yes \| yes -_examples/funcs \| yes \| yes -_examples/gopygc \| yes \| yes -_examples/gostrings \| yes \| yes -_examples/hi \| no \| yes -_examples/iface \| no \| yes -_examples/lot \| yes \| yes -_examples/maps \| yes \| yes -_examples/named \| yes \| yes -_examples/osfile \| yes \| yes -_examples/pkgconflict \| yes \| yes -_examples/pointers \| yes \| yes -_examples/pyerrors \| yes \| yes -_examples/rename \| yes \| yes -_examples/seqs \| yes \| yes -_examples/simple \| yes \| yes -_examples/sliceptr \| yes \| yes -_examples/slices \| yes \| yes -_examples/structs \| yes \| yes -_examples/unicode \| no \| yes -_examples/variadic \| no \| yes -_examples/vars \| yes \| yes +Feature \|py3 +--- \| --- +_examples/arrays \| yes +_examples/cgo \| yes +_examples/consts \| yes +_examples/cstrings \| yes +_examples/empty \| yes +_examples/funcs \| yes +_examples/gopygc \| yes +_examples/gostrings \| yes +_examples/hi \| yes +_examples/iface \| yes +_examples/lot \| yes +_examples/maps \| yes +_examples/named \| yes +_examples/osfile \| yes +_examples/pkgconflict \| yes +_examples/pointers \| yes +_examples/pyerrors \| yes +_examples/rename \| yes +_examples/seqs \| yes +_examples/simple \| yes +_examples/sliceptr \| yes +_examples/slices \| yes +_examples/structs \| yes +_examples/unicode \| yes +_examples/variadic \| yes +_examples/vars \| yes diff --git a/appveyor.yml b/appveyor.yml index acded50..03aeeb6 100644 --- a/appveyor.yml +++ b/appveyor.yml @@ -14,7 +14,7 @@ branches: environment: GOPATH: C:\gopath - GOROOT: C:\go115 + GOROOT: C:\go119 GOPY_APPVEYOR_CI: '1' GOTRACEBACK: 'crash' #CPYTHON2DIR: "C:\\Python27-x64" @@ -22,17 +22,13 @@ environment: #PATH: '%GOPATH%\bin;%CPYTHON2DIR%;%CPYTHON2DIR%\\Scripts;%CPYTHON3DIR%;%CPYTHON3DIR%\\Scripts;C:\msys64\mingw64\bin;C:\msys64\usr\bin\;%PATH%' PATH: '%GOPATH%\bin;%GOROOT%\bin;%CPYTHON3DIR%;%CPYTHON3DIR%\\Scripts;C:\msys64\mingw64\bin;C:\msys64\usr\bin\;%PATH%' -stack: go 1.15 +stack: go 1.19 build_script: - python --version - #- "%CPYTHON2DIR%\\python --version" - "%CPYTHON3DIR%\\python --version" - #- "%CPYTHON2DIR%\\python -m pip install --upgrade pip" - "%CPYTHON3DIR%\\python -m pip install --upgrade pip" - #- "%CPYTHON2DIR%\\python -m pip install cffi" - "%CPYTHON3DIR%\\python -m pip install cffi" - #- "%CPYTHON2DIR%\\python -m pip install pybindgen" - "%CPYTHON3DIR%\\python -m pip install pybindgen" - go version - go env diff --git a/bind/utils.go b/bind/utils.go index 42ef5bc..02620d6 100644 --- a/bind/utils.go +++ b/bind/utils.go @@ -103,7 +103,13 @@ func (pc PyConfig) AllFlags() string { // python VM (python, python2, python3, pypy, etc...) func GetPythonConfig(vm string) (PyConfig, error) { code := `import sys -import distutils.sysconfig as ds +try: + import sysconfig as ds + def _get_python_inc(): + return ds.get_path('include') +except ImportError: + import distutils.sysconfig as ds + _get_python_inc = ds.get_config_var import json import os version=sys.version_info.major @@ -133,7 +139,7 @@ else: print(json.dumps({ "version": sys.version_info.major, "minor": sys.version_info.minor, - "incdir": ds.get_python_inc(), + "incdir": _get_python_inc(), "libdir": ds.get_config_var("LIBDIR"), "libpy": ds.get_config_var("LIBRARY"), "shlibs": ds.get_config_var("SHLIBS"), diff --git a/cmd_build.go b/cmd_build.go index 0722629..7b57a7e 100644 --- a/cmd_build.go +++ b/cmd_build.go @@ -5,6 +5,7 @@ package main import ( + "bytes" "fmt" "log" "os" @@ -215,11 +216,17 @@ func runBuild(mode bind.BuildMode, cfg BuildCfg) error { if bind.WindowsOS { fmt.Printf("Doing windows sed hack to fix declspec for PyInit\n") - cmd = exec.Command("sed", "-i", "s/ PyInit_/ __declspec(dllexport) PyInit_/g", cfg.Name+".c") - cmdout, err = cmd.CombinedOutput() + fname := cfg.Name + ".c" + raw, err := os.ReadFile(fname) if err != nil { - fmt.Printf("cmd had error: %v output:\no%v\n", err, string(cmdout)) - return err + fmt.Printf("could not read %s: %+v", fname, err) + return fmt.Errorf("could not read %s: %w", fname, err) + } + raw = bytes.ReplaceAll(raw, []byte(" PyInit_"), []byte(" __declspec(dllexport) PyInit_")) + err = os.WriteFile(fname, raw, 0644) + if err != nil { + fmt.Printf("could not apply sed hack to fix declspec for PyInit: %+v", err) + return fmt.Errorf("could not apply sed hack to fix PyInit: %w", err) } } diff --git a/go.mod b/go.mod index d0e31cc..512c22f 100644 --- a/go.mod +++ b/go.mod @@ -1,6 +1,6 @@ module github.com/go-python/gopy -go 1.18 +go 1.19 require ( github.com/gonuts/commander v0.1.0	'sed' is not recognized as an internal or external command... sed is not a windows utility and shouldn't be used. I had to build in git bash after handling a handful of other issues to get a package to build. Why not just use python to do it?	I also have this error. C:\python\python.exe build.py Doing windows sed hack to fix declspec for PyInit cmd had error: exec: "sed": executable file not found in %PATH%	2023-03-31T16:08:16	[]	[]
glotzerlab/rowan	glotzerlab__rowan-56	3dc965380ff1aec428e9fe88ba87a59e7acf7db1	diff --git a/.circleci/config.yml b/.circleci/config.yml deleted file mode 100644 index efc4e69..0000000 --- a/.circleci/config.yml +++ /dev/null @@ -1,130 +0,0 @@ -version: 2.1 - -orbs: - codecov: codecov/[email protected] - -references: - install_dependencies: &install_dependencies - name: Install dependencies - command: \| - python -m pip install --progress-bar off -U virtualenv --user - mkdir -p ./venv - python -m virtualenv ./venv --clear - . venv/bin/activate - python -m pip install --progress-bar off -U -r requirements/requirements-testing.txt - - run_tests: &run_tests - name: run tests - command: \| - # Run with coverage - . venv/bin/activate - python -m pytest -v --cov=rowan/ --cov-report=xml - -jobs: - linux-python-39: - docker: - - image: cimg/python:3.9 - working_directory: ~/repo - steps: - - checkout - - run: - <<: install_dependencies - - run: - <<: run_tests - - linux-python-310: - docker: - - image: cimg/python:3.10 - working_directory: ~/repo - steps: - - checkout - - run: - <<: install_dependencies - - run: - <<: run_tests - - linux-python-311: - docker: - - image: cimg/python:3.11 - working_directory: ~/repo - steps: - - checkout - - run: - <<: install_dependencies - - run: - <<: run_tests - - linux-python-312: - docker: - - image: cimg/python:3.12 - working_directory: ~/repo - steps: - - checkout - - run: - <<: install_dependencies - - run: - <<: run_tests - - linux-oldest: - # Run tests against a set of the oldest dependencies we support. - docker: - - image: cimg/python:3.8 - working_directory: ~/repo - steps: - - checkout - - run: - name: Install oldest dependencies - command: \| - python -m pip install --progress-bar off -U virtualenv --user - mkdir -p ./venv - python -m virtualenv ./venv --clear - . venv/bin/activate - python -m pip install --progress-bar off -U -r .circleci/oldest-test-reqs.txt - python -m pip install --progress-bar off -U pytest - - run: - name: run tests - command: \| - . venv/bin/activate - python -m pytest -v tests - - build_and_deploy: - docker: - - image: cimg/python:3.12 - working_directory: ~/repo - steps: - - checkout - - run: - <<: install_dependencies - - run: - <<: run_tests - - run: - name: Deploy dist and wheels - command: \| - . venv/bin/activate - python -m pytest -v tests - python --version - python -m pip --version - python -m pip install --progress-bar off --user -U twine build - python -m twine --version - python -m build . - twine upload --username ${PYPI_USERNAME} --password ${PYPI_PASSWORD} dist/* - -workflows: - version: 2 - testing: - jobs: - - linux-python-39 - - linux-python-310 - - linux-python-311 - - linux-python-312: - post-steps: - - codecov/upload - - linux-oldest - deploy: - jobs: - - build_and_deploy: - filters: - tags: - only: /^v./ - branches: - ignore: /./ diff --git a/.github/workflows/release.yaml b/.github/workflows/release.yaml new file mode 100644 index 0000000..0a020e4 --- /dev/null +++ b/.github/workflows/release.yaml @@ -0,0 +1,61 @@ +name: PyPI + +concurrency: + group: ${{ github.workflow }}-${{ github.ref }} + cancel-in-progress: true + +on: + pull_request: + + push: + branches: + - "trunk-" + tags: + - "v" + + workflow_dispatch: + +jobs: + build_wheel: + name: Build wheel + runs-on: ubuntu-24.04 + steps: + - name: Checkout + uses: actions/checkout@a5ac7e51b41094c92402da3b24376905380afc29 # v4.1.6 + - name: Set up Python + uses: actions/setup-python@82c7e631bb3cdc910f68e0081d67478d79c6982d # v5.1.0 + with: + python-version: 3.12 + - name: Set up Python environment + uses: glotzerlab/workflows/setup-uv@5cfac9da9cb78e16ae97a9119b6fd13c1c2d6f5e # 0.1.0 + with: + lockfile: ".github/workflows/requirements-build.txt" + + - name: Build wheel + run: python3 -m build --outdir dist/ . + + - uses: actions/upload-artifact@65462800fd760344b1a7b4382951275a0abb4808 # v4.3.3 + with: + name: dist + path: dist + + upload_pypi: + name: Publish [PyPI] + needs: [build_wheel] + runs-on: ubuntu-latest + permissions: + id-token: write + steps: + - name: Download artifacts + uses: actions/download-artifact@65a9edc5881444af0b9093a5e628f2fe47ea3b2e # v4.1.7 + with: + merge-multiple: 'true' + pattern: dist + path: dist + + - name: Check files + run: ls -lR dist + + - name: Upload to PyPI + if: startsWith(github.ref, 'refs/tags/v') + uses: pypa/gh-action-pypi-publish@ec4db0b4ddc65acdf4bff5fa45ac92d78b56bdf0 # v1.9.0 diff --git a/.github/workflows/requirements-build.in b/.github/workflows/requirements-build.in new file mode 100644 index 0000000..f9a7807 --- /dev/null +++ b/.github/workflows/requirements-build.in @@ -0,0 +1,1 @@ +build == 1.1.1 diff --git a/.github/workflows/requirements-build.txt b/.github/workflows/requirements-build.txt new file mode 100644 index 0000000..2b16d8c --- /dev/null +++ b/.github/workflows/requirements-build.txt @@ -0,0 +1,8 @@ +# This file was autogenerated by uv via the following command: +# uv pip compile requirements-build.in +build==1.1.1 + # via -r requirements-build.in +packaging==24.1 + # via build +pyproject-hooks==1.1.0 + # via build	Migrate to Github Actions Migration to GitHub actions will streamline the CI and bring Rowan up to date with the rest of the Glotzerlab software suite. This will also help in testing for Numpy 2.x compatibility.		2024-06-20T16:47:16	[]	[]
trevismd/statannotations	trevismd__statannotations-35	c4704fbbff565c1d3eea0e93f4417b0ec67bd9b0	diff --git a/CHANGELOG.md b/CHANGELOG.md index c23571f..ecb14f2 100644 --- a/CHANGELOG.md +++ b/CHANGELOG.md @@ -1,4 +1,13 @@ ## v0.4 +### v0.4.3 +- The `correction_format` parameter allows changing how the multiple +comparisons correction method adjusts the annotation when changing a result +to non-significant. +- Fix the `show_test_name` configuration. +- Fix the `verbose` parameter + ([#37](https://github.com/trevismd/statannotations/pull/37) by + [mxposed](https://github.com/mxposed)) + ### v0.4.2 - Support of `FacetGrid` with - Support empty initialization only defining pairs diff --git a/docs/build/doctrees/environment.pickle b/docs/build/doctrees/environment.pickle deleted file mode 100644 index 880be64..0000000 Binary files a/docs/build/doctrees/environment.pickle and /dev/null differ diff --git a/docs/build/doctrees/statannotations.doctree b/docs/build/doctrees/statannotations.doctree index c28d580..1f119a8 100644 Binary files a/docs/build/doctrees/statannotations.doctree and b/docs/build/doctrees/statannotations.doctree differ diff --git a/docs/build/doctrees/statannotations.stats.doctree b/docs/build/doctrees/statannotations.stats.doctree index a4c0dfe..747056f 100644 Binary files a/docs/build/doctrees/statannotations.stats.doctree and b/docs/build/doctrees/statannotations.stats.doctree differ diff --git a/docs/build/html/genindex.html b/docs/build/html/genindex.html index 6ff36ff..8c420f7 100644 --- a/docs/build/html/genindex.html +++ b/docs/build/html/genindex.html @@ -180,6 +180,8 @@ <h1 id="index">Index</h1> <h2 id="A">A</h2> <table style="width: 100%" class="indextable genindextable"><tr> <td style="width: 33%; vertical-align: top;"><ul> + <li><a href="statannotations.stats.html#statannotations.stats.StatResult.StatResult.adjust">adjust() (statannotations.stats.StatResult.StatResult method)</a> +</li> <li><a href="statannotations.html#statannotations.Annotator.Annotator.alpha">alpha (statannotations.Annotator.Annotator property)</a> </li> <li><a href="statannotations.html#statannotations.Annotator.Annotator.annotate">annotate() (statannotations.Annotator.Annotator method)</a> @@ -241,6 +243,8 @@ <h2 id="C">C</h2> <li><a href="statannotations.html#statannotations.Annotator.Annotator.configure">configure() (statannotations.Annotator.Annotator method)</a> </li> <li><a href="statannotations.stats.html#statannotations.stats.StatResult.StatResult.corrected_significance">corrected_significance (statannotations.stats.StatResult.StatResult property)</a> +</li> + <li><a href="statannotations.html#statannotations.PValueFormat.PValueFormat.correction_format">correction_format (statannotations.PValueFormat.PValueFormat property)</a> </li> <li><a href="statannotations.stats.html#statannotations.stats.StatResult.StatResult.correction_method">correction_method (statannotations.stats.StatResult.StatResult property)</a> </li> @@ -302,10 +306,12 @@ <h2 id="G">G</h2> <li><a href="statannotations.html#statannotations.PValueFormat.PValueFormat.get_configuration">(statannotations.PValueFormat.PValueFormat method)</a> </li> </ul></li> - <li><a href="statannotations.stats.html#statannotations.stats.ComparisonsCorrection.get_correction_parameters">get_correction_parameters() (in module statannotations.stats.ComparisonsCorrection)</a> + <li><a href="statannotations.html#statannotations.PValueFormat.get_corrected_star">get_corrected_star() (in module statannotations.PValueFormat)</a> </li> </ul></td> <td style="width: 33%; vertical-align: top;"><ul> + <li><a href="statannotations.stats.html#statannotations.stats.ComparisonsCorrection.get_correction_parameters">get_correction_parameters() (in module statannotations.stats.ComparisonsCorrection)</a> +</li> <li><a href="statannotations.html#statannotations.Annotator.Annotator.get_empty_annotator">get_empty_annotator() (statannotations.Annotator.Annotator static method)</a> </li> <li><a href="statannotations.stats.html#statannotations.stats.utils.get_num_comparisons">get_num_comparisons() (in module statannotations.stats.utils)</a> diff --git a/docs/build/html/objects.inv b/docs/build/html/objects.inv index faca044..ca638c7 100644 Binary files a/docs/build/html/objects.inv and b/docs/build/html/objects.inv differ diff --git a/docs/build/html/searchindex.js b/docs/build/html/searchindex.js index 6baea8a..6750b03 100644 --- a/docs/build/html/searchindex.js +++ b/docs/build/html/searchindex.js @@ -1,1 +1,1 @@ -Search.setIndex({docnames:["index","modules","setup","statannotations","statannotations.stats"],envversion:{"sphinx.domains.c":2,"sphinx.domains.changeset":1,"sphinx.domains.citation":1,"sphinx.domains.cpp":3,"sphinx.domains.index":1,"sphinx.domains.javascript":2,"sphinx.domains.math":2,"sphinx.domains.python":3,"sphinx.domains.rst":2,"sphinx.domains.std":2,sphinx:56},filenames:["index.rst","modules.rst","setup.rst","statannotations.rst","statannotations.stats.rst"],objects:{"":{statannotations:[3,0,0,"-"]},"statannotations.Annotation":{Annotation:[3,1,1,""]},"statannotations.Annotation.Annotation":{formatted_output:[3,2,1,""],print_labels_and_content:[3,3,1,""],text:[3,2,1,""]},"statannotations.Annotator":{Annotator:[3,1,1,""]},"statannotations.Annotator.Annotator":{alpha:[3,2,1,""],annotate:[3,3,1,""],annotate_custom_annotations:[3,3,1,""],apply_and_annotate:[3,3,1,""],apply_test:[3,3,1,""],comparisons_correction:[3,2,1,""],configure:[3,3,1,""],fig:[3,2,1,""],get_annotations_text:[3,3,1,""],get_configuration:[3,3,1,""],get_empty_annotator:[3,3,1,""],get_offset_func:[3,3,1,""],has_type0_comparisons_correction:[3,3,1,""],loc:[3,2,1,""],new_plot:[3,3,1,""],orient:[3,2,1,""],plot_and_annotate:[3,3,1,""],plot_and_annotate_facets:[3,3,1,""],print_pvalue_legend:[3,3,1,""],pvalue_format:[3,2,1,""],reset_configuration:[3,3,1,""],set_custom_annotations:[3,3,1,""],set_pvalues:[3,3,1,""],set_pvalues_and_annotate:[3,3,1,""],test:[3,2,1,""],validate_test_short_name:[3,3,1,""],verbose:[3,2,1,""]},"statannotations.PValueFormat":{Formatter:[3,1,1,""],PValueFormat:[3,1,1,""],sort_pvalue_thresholds:[3,4,1,""]},"statannotations.PValueFormat.Formatter":{config:[3,3,1,""],format_data:[3,3,1,""]},"statannotations.PValueFormat.PValueFormat":{config:[3,3,1,""],format_data:[3,3,1,""],get_configuration:[3,3,1,""],print_legend_if_used:[3,3,1,""],pvalue_format_string:[3,2,1,""],pvalue_thresholds:[3,2,1,""],simple_format_string:[3,2,1,""],text_format:[3,2,1,""]},"statannotations.format_annotations":{pval_annotation_text:[3,4,1,""],simple_text:[3,4,1,""]},"statannotations.stats":{ComparisonsCorrection:[4,0,0,"-"],StatResult:[4,0,0,"-"],StatTest:[4,0,0,"-"],test:[4,0,0,"-"],utils:[4,0,0,"-"]},"statannotations.stats.ComparisonsCorrection":{ComparisonsCorrection:[4,1,1,""],check_valid_correction_name:[4,4,1,""],get_correction_parameters:[4,4,1,""],get_validated_comparisons_correction:[4,4,1,""]},"statannotations.stats.ComparisonsCorrection.ComparisonsCorrection":{apply:[4,3,1,""],document:[4,3,1,""]},"statannotations.stats.StatResult":{StatResult:[4,1,1,""]},"statannotations.stats.StatResult.StatResult":{corrected_significance:[4,2,1,""],correction_method:[4,2,1,""],formatted_output:[4,2,1,""],significance_suffix:[4,2,1,""]},"statannotations.stats.StatTest":{StatTest:[4,1,1,""],wilcoxon:[4,4,1,""]},"statannotations.stats.StatTest.StatTest":{from_library:[4,3,1,""],short_name:[4,2,1,""]},"statannotations.stats.test":{apply_test:[4,4,1,""]},"statannotations.stats.utils":{check_alpha:[4,4,1,""],check_num_comparisons:[4,4,1,""],check_pvalues:[4,4,1,""],get_num_comparisons:[4,4,1,""],return_results:[4,4,1,""]},"statannotations.utils":{InvalidParametersError:[3,5,1,""],check_is_in:[3,4,1,""],check_not_none:[3,4,1,""],check_order_in_data:[3,4,1,""],check_pairs_in_data:[3,4,1,""],check_valid_text_format:[3,4,1,""],empty_dict_if_none:[3,4,1,""],get_closest:[3,4,1,""],get_x_values:[3,4,1,""],raise_expected_got:[3,4,1,""],remove_null:[3,4,1,""],render_collection:[3,4,1,""]},statannotations:{Annotation:[3,0,0,"-"],Annotator:[3,0,0,"-"],PValueFormat:[3,0,0,"-"],format_annotations:[3,0,0,"-"],stats:[4,0,0,"-"],utils:[3,0,0,"-"]}},objnames:{"0":["py","module","Python module"],"1":["py","class","Python class"],"2":["py","property","Python property"],"3":["py","method","Python method"],"4":["py","function","Python function"],"5":["py","exception","Python exception"]},objtypes:{"0":"py:module","1":"py:class","2":"py:property","3":"py:method","4":"py:function","5":"py:exception"},terms:{"0":[3,4],"001":3,"01":3,"05":[3,4],"1":[3,4],"12141511":3,"1e":3,"2":3,"3":3,"4":3,"5":3,"5e":3,"9981846":3,"case":4,"class":[3,4],"default":[3,4],"float":4,"function":[3,4],"int":4,"return":3,"short":3,"static":[3,4],"true":[3,4],A:3,For:3,If:3,One:3,Or:3,The:3,a_list:3,accept:3,add:3,addit:[3,4],after:3,alpha:[3,4],an:3,annot:1,annotate_custom_annot:3,annotate_param:3,annotation_func:3,annotation_param:3,appli:[3,4],apply_and_annot:3,apply_test:[3,4],ar:3,arg:[3,4],argument:[3,4],arrai:3,assum:3,auto:3,ax:3,ax_op_aft:3,ax_op_befor:3,bar:3,base:[3,4],befor:3,behavior:[3,4],being:3,benjamini:4,between:3,bonferroni:4,bool:4,box:3,boxplot:3,calcul:3,call:3,callabl:4,can:3,check:3,check_alpha:4,check_is_in:3,check_not_non:3,check_num_comparison:4,check_order_in_data:3,check_pairs_in_data:3,check_pvalu:4,check_valid_correction_nam:4,check_valid_text_format:3,classmethod:3,close:3,closest:3,collect:3,color:3,com:3,compar:3,comparison:[3,4],comparisons_correct:[3,4],comparisonscorrect:[1,3],comput:3,config:3,configur:3,constructor:3,content:1,coord:3,coordin:3,core:3,corr_kwarg:4,correct:[3,4],corrected_signific:4,correction_method:4,creat:3,custom:3,data:[3,4],datafram:3,defin:3,dict:[3,4],displai:3,document:4,each:3,earlier:4,either:4,empty_dict_if_non:3,engin:3,equal:3,equival:4,error:3,error_typ:3,exact:3,except:3,exist:3,expect:3,facetgrid:3,fals:3,fig:3,first:3,follow:3,for_:3,format:[3,4],format_annot:1,format_data:3,formatt:3,formatted_output:[3,4],fraction:3,frame:3,from:[3,4],from_librari:4,func:4,func_nam:3,fwer:3,gener:3,get:4,get_annotations_text:3,get_closest:3,get_configur:3,get_correction_paramet:4,get_empty_annot:3,get_num_comparison:4,get_offset_func:3,get_validated_comparisons_correct:4,get_x_valu:3,got:3,group:3,group_data1:4,group_data2:4,gt:4,has_type0_comparisons_correct:3,have:3,hochberg:4,hold:3,holm:4,http:3,hue:3,hue_ord:3,index:0,initi:3,instanc:[3,4],instead:3,interfac:4,interpret:4,invalidparameterserror:3,keyword:4,kruskal:4,kwarg:[3,4],label:3,leven:4,line_height:3,line_offset:3,line_offset_to_group:3,line_width:3,link:3,list:3,loc:3,ls:4,mann:4,map:3,messag:3,method:[3,4],method_typ:4,mode:3,modul:[0,1],multipl:[3,4],multipletest:3,must:3,mylist:3,mynumb:3,name:[3,4],new_plot:3,none:[3,4],ns:3,num_comparison:[3,4],number:[3,4],object:[3,4],one:[3,4],onto:3,option:[3,4],order:3,orient:3,otherwis:3,overrid:3,p:3,p_valu:4,packag:[0,1],page:0,pair:3,panda:3,paramet:[3,4],pass:[3,4],per:3,perform:3,plot:3,plot_and_annot:3,plot_and_annotate_facet:3,plot_param:3,point:3,posit:3,print_labels_and_cont:3,print_legend_if_us:3,print_pvalue_legend:3,properti:[3,4],provid:[3,4],pval:4,pval_annotation_text:3,pvalu:[3,4],pvalue_format:3,pvalue_format_str:3,pvalue_threshold:3,pvalueformat:1,rais:3,raise_expected_got:3,rang:3,refer:3,remove_nul:3,render:3,render_collect:3,reset_configur:3,result:[3,4],results_arrai:4,return_result:4,run:4,same:3,sampl:4,seaborn:3,search:0,see:3,sep:3,seri:3,set:3,set_custom_annot:3,set_pvalu:3,set_pvalues_and_annot:3,short_nam:4,show:3,show_test_nam:3,significance_suffix:4,simpl:3,simple_format_str:3,simple_text:3,singl:3,smallest:3,sort:3,sort_pvalue_threshold:3,stackoverflow:3,standard:3,star:3,stat:[1,3],stat_nam:4,stat_str:4,statist:[3,4],statresult:[1,3],stats_param:[3,4],statsmodel:[3,4],statsmodels_api:4,stattest:[1,3],str:[3,4],string:3,struct:3,submodul:1,subpackag:1,t:4,test:[1,3],test_descript:4,test_ind:4,test_long_nam:4,test_nam:4,test_pair:4,test_result_list:4,test_short_nam:[3,4],test_welch:4,text:3,text_annot_custom:3,text_format:3,text_offset:3,them:3,thereof:3,thi:[3,4],three:3,threshold:3,top:3,tupl:3,two:[3,4],type:3,union:[3,4],us:[3,4],use_fixed_offset:3,util:1,valid_valu:3,validate_test_short_nam:3,valu:[3,4],valueerror:3,verbos:[3,4],version:4,vs:3,w:3,wa:3,whitnei:4,wilcoxon:4,within:3,work:3,x:3,y:3,yekuti:4},titles:["Welcome to statannotations\u2019s documentation!","statannotations","setup module","statannotations package","statannotations.stats package"],titleterms:{annot:3,comparisonscorrect:4,content:[0,3,4],document:0,format_annot:3,indic:0,modul:[2,3,4],packag:[3,4],pvalueformat:3,s:0,setup:2,stat:4,statannot:[0,1,3,4],statresult:4,stattest:4,submodul:[3,4],subpackag:3,tabl:0,test:4,util:[3,4],welcom:0}}) \ No newline at end of file +Search.setIndex({docnames:["index","modules","setup","statannotations","statannotations.stats"],envversion:{"sphinx.domains.c":2,"sphinx.domains.changeset":1,"sphinx.domains.citation":1,"sphinx.domains.cpp":3,"sphinx.domains.index":1,"sphinx.domains.javascript":2,"sphinx.domains.math":2,"sphinx.domains.python":3,"sphinx.domains.rst":2,"sphinx.domains.std":2,sphinx:56},filenames:["index.rst","modules.rst","setup.rst","statannotations.rst","statannotations.stats.rst"],objects:{"":{statannotations:[3,0,0,"-"]},"statannotations.Annotation":{Annotation:[3,1,1,""]},"statannotations.Annotation.Annotation":{formatted_output:[3,2,1,""],print_labels_and_content:[3,3,1,""],text:[3,2,1,""]},"statannotations.Annotator":{Annotator:[3,1,1,""]},"statannotations.Annotator.Annotator":{alpha:[3,2,1,""],annotate:[3,3,1,""],annotate_custom_annotations:[3,3,1,""],apply_and_annotate:[3,3,1,""],apply_test:[3,3,1,""],comparisons_correction:[3,2,1,""],configure:[3,3,1,""],fig:[3,2,1,""],get_annotations_text:[3,3,1,""],get_configuration:[3,3,1,""],get_empty_annotator:[3,3,1,""],get_offset_func:[3,3,1,""],has_type0_comparisons_correction:[3,3,1,""],loc:[3,2,1,""],new_plot:[3,3,1,""],orient:[3,2,1,""],plot_and_annotate:[3,3,1,""],plot_and_annotate_facets:[3,3,1,""],print_pvalue_legend:[3,3,1,""],pvalue_format:[3,2,1,""],reset_configuration:[3,3,1,""],set_custom_annotations:[3,3,1,""],set_pvalues:[3,3,1,""],set_pvalues_and_annotate:[3,3,1,""],test:[3,2,1,""],validate_test_short_name:[3,3,1,""],verbose:[3,2,1,""]},"statannotations.PValueFormat":{Formatter:[3,1,1,""],PValueFormat:[3,1,1,""],get_corrected_star:[3,4,1,""],sort_pvalue_thresholds:[3,4,1,""]},"statannotations.PValueFormat.Formatter":{config:[3,3,1,""],format_data:[3,3,1,""]},"statannotations.PValueFormat.PValueFormat":{config:[3,3,1,""],correction_format:[3,2,1,""],format_data:[3,3,1,""],get_configuration:[3,3,1,""],print_legend_if_used:[3,3,1,""],pvalue_format_string:[3,2,1,""],pvalue_thresholds:[3,2,1,""],simple_format_string:[3,2,1,""],text_format:[3,2,1,""]},"statannotations.format_annotations":{pval_annotation_text:[3,4,1,""],simple_text:[3,4,1,""]},"statannotations.stats":{ComparisonsCorrection:[4,0,0,"-"],StatResult:[4,0,0,"-"],StatTest:[4,0,0,"-"],test:[4,0,0,"-"],utils:[4,0,0,"-"]},"statannotations.stats.ComparisonsCorrection":{ComparisonsCorrection:[4,1,1,""],check_valid_correction_name:[4,4,1,""],get_correction_parameters:[4,4,1,""],get_validated_comparisons_correction:[4,4,1,""]},"statannotations.stats.ComparisonsCorrection.ComparisonsCorrection":{apply:[4,3,1,""],document:[4,3,1,""]},"statannotations.stats.StatResult":{StatResult:[4,1,1,""]},"statannotations.stats.StatResult.StatResult":{adjust:[4,3,1,""],corrected_significance:[4,2,1,""],correction_method:[4,2,1,""],formatted_output:[4,2,1,""],significance_suffix:[4,2,1,""]},"statannotations.stats.StatTest":{StatTest:[4,1,1,""],wilcoxon:[4,4,1,""]},"statannotations.stats.StatTest.StatTest":{from_library:[4,3,1,""],short_name:[4,2,1,""]},"statannotations.stats.test":{apply_test:[4,4,1,""]},"statannotations.stats.utils":{check_alpha:[4,4,1,""],check_num_comparisons:[4,4,1,""],check_pvalues:[4,4,1,""],get_num_comparisons:[4,4,1,""],return_results:[4,4,1,""]},"statannotations.utils":{InvalidParametersError:[3,5,1,""],check_is_in:[3,4,1,""],check_not_none:[3,4,1,""],check_order_in_data:[3,4,1,""],check_pairs_in_data:[3,4,1,""],check_valid_text_format:[3,4,1,""],empty_dict_if_none:[3,4,1,""],get_closest:[3,4,1,""],get_x_values:[3,4,1,""],raise_expected_got:[3,4,1,""],remove_null:[3,4,1,""],render_collection:[3,4,1,""]},statannotations:{Annotation:[3,0,0,"-"],Annotator:[3,0,0,"-"],PValueFormat:[3,0,0,"-"],format_annotations:[3,0,0,"-"],stats:[4,0,0,"-"],utils:[3,0,0,"-"]}},objnames:{"0":["py","module","Python module"],"1":["py","class","Python class"],"2":["py","property","Python property"],"3":["py","method","Python method"],"4":["py","function","Python function"],"5":["py","exception","Python exception"]},objtypes:{"0":"py:module","1":"py:class","2":"py:property","3":"py:method","4":"py:function","5":"py:exception"},terms:{"0":[3,4],"001":3,"01":3,"05":[3,4],"1":[3,4],"12141511":3,"1e":3,"2":3,"3":3,"4":3,"5":3,"5e":3,"9981846":3,"case":4,"class":[3,4],"default":[3,4],"float":4,"function":[3,4],"int":4,"return":3,"short":3,"static":[3,4],"true":[3,4],A:3,For:3,If:3,One:3,Or:3,The:3,a_list:3,accept:3,ad:3,add:3,addit:[3,4],adjust:4,after:3,alpha:[3,4],an:3,annot:1,annotate_custom_annot:3,annotate_param:3,annotation_func:3,annotation_param:3,appli:[3,4],apply_and_annot:3,apply_test:[3,4],ar:3,arg:[3,4],argument:[3,4],arrai:3,assum:3,auto:3,ax:3,ax_op_aft:3,ax_op_befor:3,bar:3,base:[3,4],befor:3,behavior:[3,4],being:3,benjamini:4,between:3,bonferroni:4,bool:4,box:3,boxplot:3,calcul:3,call:3,callabl:4,can:3,check:3,check_alpha:4,check_is_in:3,check_not_non:3,check_num_comparison:4,check_order_in_data:3,check_pairs_in_data:3,check_pvalu:4,check_valid_correction_nam:4,check_valid_text_format:3,classmethod:3,close:3,closest:3,collect:3,color:3,com:3,compar:3,comparison:[3,4],comparisons_correct:[3,4],comparisonscorrect:[1,3],comput:3,config:3,configur:3,constructor:3,content:1,coord:3,coordin:3,core:3,corr_kwarg:4,correct:[3,4],corrected_signific:4,correction_format:3,correction_method:4,correspond:3,creat:3,custom:3,data:[3,4],datafram:3,defin:3,dict:[3,4],displai:3,document:4,each:3,earlier:4,either:4,empty_dict_if_non:3,engin:3,equal:3,equival:4,error:3,error_typ:3,exact:3,except:3,exist:3,expect:3,facetgrid:3,fals:3,fig:3,first:3,follow:3,for_:3,format:[3,4],format_annot:1,format_data:3,formatt:3,formatted_output:[3,4],fraction:3,frame:3,from:[3,4],from_librari:4,func:4,func_nam:3,fwer:3,gener:3,get:4,get_annotations_text:3,get_closest:3,get_configur:3,get_corrected_star:3,get_correction_paramet:4,get_empty_annot:3,get_num_comparison:4,get_offset_func:3,get_validated_comparisons_correct:4,get_x_valu:3,got:3,group:3,group_data1:4,group_data2:4,gt:4,has_type0_comparisons_correct:3,have:3,hochberg:4,hold:3,holm:4,how:3,http:3,hue:3,hue_ord:3,index:0,initi:3,instanc:[3,4],instead:3,interfac:4,interpret:4,invalidparameterserror:3,keyword:4,kruskal:4,kwarg:[3,4],label:3,leven:4,line_height:3,line_offset:3,line_offset_to_group:3,line_width:3,link:3,list:3,loc:3,ls:4,mann:4,map:3,messag:3,method:[3,4],method_typ:4,mode:3,modul:[0,1],multipl:[3,4],multipletest:3,must:3,mylist:3,mynumb:3,name:[3,4],new_plot:3,none:[3,4],notat:3,ns:3,num_comparison:[3,4],number:[3,4],object:[3,4],one:[3,4],onto:3,option:[3,4],order:3,orient:3,origin:3,otherwis:3,output:3,overrid:3,p:3,p_valu:4,packag:[0,1],page:0,pair:3,panda:3,paramet:[3,4],pass:[3,4],per:3,perform:3,plot:3,plot_and_annot:3,plot_and_annotate_facet:3,plot_param:3,point:3,posit:3,print:3,print_labels_and_cont:3,print_legend_if_us:3,print_pvalue_legend:3,properti:[3,4],provid:[3,4],pval:4,pval_annotation_text:3,pvalu:[3,4],pvalue_format:3,pvalue_format_str:3,pvalue_threshold:3,pvalueformat:1,rais:3,raise_expected_got:3,rang:3,re:3,refer:3,remov:3,remove_nul:3,render:3,render_collect:3,replac:3,reset_configur:3,result:[3,4],results_arrai:4,return_result:4,run:4,same:3,sampl:4,seaborn:3,search:0,see:3,sep:3,seri:3,set:3,set_custom_annot:3,set_pvalu:3,set_pvalues_and_annot:3,short_nam:4,short_test_nam:3,show:3,show_test_nam:3,signific:3,significance_suffix:4,simpl:3,simple_format_str:3,simple_text:3,singl:3,smallest:3,sort:3,sort_pvalue_threshold:3,stackoverflow:3,standard:3,star:3,stat:[1,3],stat_nam:4,stat_str:4,stat_summari:4,statist:[3,4],statresult:[1,3],stats_param:[3,4],statsmodel:[3,4],statsmodels_api:4,stattest:[1,3],str:[3,4],string:3,struct:3,submodul:1,subpackag:1,suffix:3,t:4,test:[1,3],test_descript:4,test_ind:4,test_long_nam:4,test_nam:4,test_pair:4,test_result_list:4,test_short_nam:[3,4],test_welch:4,text:3,text_annot_custom:3,text_format:3,text_offset:3,them:3,thereof:3,thi:[3,4],three:3,threshold:3,top:3,tupl:3,two:[3,4],type:3,union:[3,4],us:[3,4],use_fixed_offset:3,util:1,valid_valu:3,validate_test_short_nam:3,valu:[3,4],valueerror:3,verbos:[3,4],version:4,vs:3,w:3,wa:3,when:3,whether:3,whitnei:4,wilcoxon:4,within:3,work:3,x:3,y:3,yekuti:4},titles:["Welcome to statannotations\u2019s documentation!","statannotations","setup module","statannotations package","statannotations.stats package"],titleterms:{annot:3,comparisonscorrect:4,content:[0,3,4],document:0,format_annot:3,indic:0,modul:[2,3,4],packag:[3,4],pvalueformat:3,s:0,setup:2,stat:4,statannot:[0,1,3,4],statresult:4,stattest:4,submodul:[3,4],subpackag:3,tabl:0,test:4,util:[3,4],welcom:0}}) \ No newline at end of file diff --git a/docs/build/html/statannotations.html b/docs/build/html/statannotations.html index 170abcf..57267ef 100644 --- a/docs/build/html/statannotations.html +++ b/docs/build/html/statannotations.html @@ -303,6 +303,27 @@ <h2>Submodules<a class="headerlink" href="#submodules" title="Permalink to this </dd> </dl> </li> +<li><dl> +<dt><cite>correction_format</cite>: How to format the star notation on the plot when</dt><dd><p>the multiple comparisons correction method removes the significance +* <cite>default</cite>: a ‘ (ns)’ suffix is added, such as in printed output,</p> +<blockquote> +<div><p>corresponds to “{star} ({suffix})”</p> +</div></blockquote> +<ul class="simple"> +<li><dl class="simple"> +<dt><cite>replace</cite>: the original star value is replaced with ‘ns’</dt><dd><p>corresponds to “{suffix}”</p> +</dd> +</dl> +</li> +<li><dl class="simple"> +<dt>a custom formatting string using “{star}” for the original</dt><dd><p>pvalue and ‘{suffix}’ for ‘ns’</p> +</dd> +</dl> +</li> +</ul> +</dd> +</dl> +</li> <li><p><cite>line_height</cite>: in axes fraction coordinates</p></li> <li><p><cite>line_offset</cite></p></li> <li><p><cite>line_offset_to_group</cite></p></li> @@ -533,6 +554,11 @@ <h2>Submodules<a class="headerlink" href="#submodules" title="Permalink to this <span class="sig-name descname"><span class="pre">config</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="o"><span class="pre"></span></span><span class="n"><span class="pre">parameters</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#statannotations.PValueFormat.PValueFormat.config" title="Permalink to this definition">¶</a></dt> <dd></dd></dl> +<dl class="py property"> +<dt class="sig sig-object py" id="statannotations.PValueFormat.PValueFormat.correction_format"> +<em class="property"><span class="pre">property</span> </em><span class="sig-name descname"><span class="pre">correction_format</span></span><a class="headerlink" href="#statannotations.PValueFormat.PValueFormat.correction_format" title="Permalink to this definition">¶</a></dt> +<dd></dd></dl> + <dl class="py method"> <dt class="sig sig-object py" id="statannotations.PValueFormat.PValueFormat.format_data"> <span class="sig-name descname"><span class="pre">format_data</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">result</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#statannotations.PValueFormat.PValueFormat.format_data" title="Permalink to this definition">¶</a></dt> @@ -570,6 +596,11 @@ <h2>Submodules<a class="headerlink" href="#submodules" title="Permalink to this </dd></dl> +<dl class="py function"> +<dt class="sig sig-object py" id="statannotations.PValueFormat.get_corrected_star"> +<span class="sig-prename descclassname"><span class="pre">statannotations.PValueFormat.</span></span><span class="sig-name descname"><span class="pre">get_corrected_star</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">star</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">res</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><a class="reference internal" href="statannotations.stats.html#statannotations.stats.StatResult.StatResult" title="statannotations.stats.StatResult.StatResult"><span class="pre">statannotations.stats.StatResult.StatResult</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">correction_format</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'{star}</span> <span class="pre">({suffix})'</span></span></em><span class="sig-paren">)</span> → <span class="pre">str</span><a class="headerlink" href="#statannotations.PValueFormat.get_corrected_star" title="Permalink to this definition">¶</a></dt> +<dd></dd></dl> + <dl class="py function"> <dt class="sig sig-object py" id="statannotations.PValueFormat.sort_pvalue_thresholds"> <span class="sig-prename descclassname"><span class="pre">statannotations.PValueFormat.</span></span><span class="sig-name descname"><span class="pre">sort_pvalue_thresholds</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">pvalue_thresholds</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#statannotations.PValueFormat.sort_pvalue_thresholds" title="Permalink to this definition">¶</a></dt> @@ -580,7 +611,7 @@ <h2>Submodules<a class="headerlink" href="#submodules" title="Permalink to this <span id="statannotations-format-annotations-module"></span><h2>statannotations.format_annotations module<a class="headerlink" href="#module-statannotations.format_annotations" title="Permalink to this headline">¶</a></h2> <dl class="py function"> <dt class="sig sig-object py" id="statannotations.format_annotations.pval_annotation_text"> -<span class="sig-prename descclassname"><span class="pre">statannotations.format_annotations.</span></span><span class="sig-name descname"><span class="pre">pval_annotation_text</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">result</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="statannotations.stats.html#statannotations.stats.StatResult.StatResult" title="statannotations.stats.StatResult.StatResult"><span class="pre">statannotations.stats.StatResult.StatResult</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span> </span><a class="reference internal" href="statannotations.stats.html#statannotations.stats.StatResult.StatResult" title="statannotations.stats.StatResult.StatResult"><span class="pre">statannotations.stats.StatResult.StatResult</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pvalue_thresholds</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">List</span></span></em><span class="sig-paren">)</span> → <span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span> </span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><a class="headerlink" href="#statannotations.format_annotations.pval_annotation_text" title="Permalink to this definition">¶</a></dt> +<span class="sig-prename descclassname"><span class="pre">statannotations.format_annotations.</span></span><span class="sig-name descname"><span class="pre">pval_annotation_text</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">result</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="statannotations.stats.html#statannotations.stats.StatResult.StatResult" title="statannotations.stats.StatResult.StatResult"><span class="pre">statannotations.stats.StatResult.StatResult</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pvalue_thresholds</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">List</span></span></em><span class="sig-paren">)</span> → <span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">tuple</span><span class="p"><span class="pre">]</span></span><a class="headerlink" href="#statannotations.format_annotations.pval_annotation_text" title="Permalink to this definition">¶</a></dt> <dd><dl class="field-list simple"> <dt class="field-odd">Parameters</dt> <dd class="field-odd"><ul class="simple"> @@ -597,7 +628,7 @@ <h2>Submodules<a class="headerlink" href="#submodules" title="Permalink to this <dl class="py function"> <dt class="sig sig-object py" id="statannotations.format_annotations.simple_text"> -<span class="sig-prename descclassname"><span class="pre">statannotations.format_annotations.</span></span><span class="sig-name descname"><span class="pre">simple_text</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">result</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><a class="reference internal" href="statannotations.stats.html#statannotations.stats.StatResult.StatResult" title="statannotations.stats.StatResult.StatResult"><span class="pre">statannotations.stats.StatResult.StatResult</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">pvalue_format</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pvalue_thresholds</span></span></em><span class="sig-paren">)</span> → <span class="pre">str</span><a class="headerlink" href="#statannotations.format_annotations.simple_text" title="Permalink to this definition">¶</a></dt> +<span class="sig-prename descclassname"><span class="pre">statannotations.format_annotations.</span></span><span class="sig-name descname"><span class="pre">simple_text</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">result</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><a class="reference internal" href="statannotations.stats.html#statannotations.stats.StatResult.StatResult" title="statannotations.stats.StatResult.StatResult"><span class="pre">statannotations.stats.StatResult.StatResult</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">pvalue_format</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pvalue_thresholds</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">short_test_name</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span> → <span class="pre">str</span><a class="headerlink" href="#statannotations.format_annotations.simple_text" title="Permalink to this definition">¶</a></dt> <dd><p>Generates simple text for test name and pvalue.</p> <dl class="field-list simple"> <dt class="field-odd">Parameters</dt> @@ -605,7 +636,7 @@ <h2>Submodules<a class="headerlink" href="#submodules" title="Permalink to this <li><p><strong>result</strong> – StatResult instance</p></li> <li><p><strong>pvalue_format</strong> – format string for pvalue</p></li> <li><p><strong>pvalue_thresholds</strong> – String to display per pvalue range</p></li> -<li><p><strong>result</strong> – StatResult</p></li> +<li><p><strong>short_test_name</strong> – whether to display the test (short) name</p></li> </ul> </dd> <dt class="field-even">Returns</dt> diff --git a/docs/build/html/statannotations.stats.html b/docs/build/html/statannotations.stats.html index 099ec99..2b11749 100644 --- a/docs/build/html/statannotations.stats.html +++ b/docs/build/html/statannotations.stats.html @@ -221,6 +221,11 @@ <h2>Submodules<a class="headerlink" href="#submodules" title="Permalink to this <dt class="sig sig-object py" id="statannotations.stats.StatResult.StatResult"> <em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">statannotations.stats.StatResult.</span></span><span class="sig-name descname"><span class="pre">StatResult</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">test_description</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">test_short_name</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">stat_str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">stat</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pval</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">alpha</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.05</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#statannotations.stats.StatResult.StatResult" title="Permalink to this definition">¶</a></dt> <dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">object</span></code></p> +<dl class="py method"> +<dt class="sig sig-object py" id="statannotations.stats.StatResult.StatResult.adjust"> +<span class="sig-name descname"><span class="pre">adjust</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">stat_summary</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#statannotations.stats.StatResult.StatResult.adjust" title="Permalink to this definition">¶</a></dt> +<dd></dd></dl> + <dl class="py property"> <dt class="sig sig-object py" id="statannotations.stats.StatResult.StatResult.corrected_significance"> <em class="property"><span class="pre">property</span> </em><span class="sig-name descname"><span class="pre">corrected_significance</span></span><a class="headerlink" href="#statannotations.stats.StatResult.StatResult.corrected_significance" title="Permalink to this definition">¶</a></dt> diff --git a/statannotations/Annotator.py b/statannotations/Annotator.py index a8e5511..372dcbf 100644 --- a/statannotations/Annotator.py +++ b/statannotations/Annotator.py @@ -117,7 +117,7 @@ def __init__(self, ax, pairs, plot='boxplot', data=None, x=None, self._comparisons_correction = None self._loc = "inside" - self._verbose = 1 + self._verbose = verbose self._just_configured = True self.show_test_name = True self.use_fixed_offset = False @@ -253,6 +253,14 @@ def configure(self, parameters): * `comparisons_correction`: Method for multiple comparisons correction. One of `statsmodels` `multipletests` methods (w/ default FWER), or a `ComparisonsCorrection` instance. + * `correction_format`: How to format the star notation on the plot when + the multiple comparisons correction method removes the significance + * `default`: a ' (ns)' suffix is added, such as in printed output, + corresponds to "{star} ({suffix})" + * `replace`: the original star value is replaced with 'ns' + corresponds to "{suffix}" + * a custom formatting string using "{star}" for the original + pvalue and '{suffix}' for 'ns' * `line_height`: in axes fraction coordinates * `line_offset` * `line_offset_to_group` diff --git a/statannotations/PValueFormat.py b/statannotations/PValueFormat.py index 598984b..2f35b4d 100644 --- a/statannotations/PValueFormat.py +++ b/statannotations/PValueFormat.py @@ -1,14 +1,18 @@ +from statannotations.stats.StatResult import StatResult + from statannotations.format_annotations import pval_annotation_text, \ simple_text from statannotations.utils import DEFAULT, check_valid_text_format, \ InvalidParametersError CONFIGURABLE_PARAMETERS = [ + 'correction_format', 'fontsize', 'pvalue_format_string', 'simple_format_string', 'text_format', - 'pvalue_thresholds' + 'pvalue_thresholds', + 'show_test_name' ] @@ -32,6 +36,8 @@ def __init__(self): self.fontsize = 'medium' self._default_pvalue_thresholds = True self._pvalue_thresholds = self._get_pvalue_thresholds(DEFAULT) + self._correction_format = "{star} ({suffix})" + self.show_test_name = True def config(self, *parameters): @@ -91,6 +97,17 @@ def pvalue_format_string(self, pvalue_format_string): self._get_pvalue_and_simple_formats(pvalue_format_string) ) + @property + def correction_format(self): + return self._correction_format + + @correction_format.setter + def correction_format(self, correction_format: str): + self._correction_format = { + "replace": "{suffix}", + "default": "{star} ({suffix})", + }.get(correction_format, correction_format) + @property def pvalue_thresholds(self): return self._pvalue_thresholds @@ -156,17 +173,32 @@ def _update_pvalue_thresholds(self): def format_data(self, result): if self.text_format == 'full': - return ("{} p = {}{}" + text = (f"{result.test_short_name} " if self.show_test_name + else "") + + return ("{}p = {}{}" .format('{}', self.pvalue_format_string, '{}') - .format(result.test_short_name, result.pvalue, - result.significance_suffix)) + .format(text, result.pvalue, result.significance_suffix)) elif self.text_format == 'star': - return pval_annotation_text(result, self.pvalue_thresholds) + was_list = False + + if not isinstance(result, list): + result = [result] + + annotations = [ + get_corrected_star(star, res, self._correction_format) + for star, res + in pval_annotation_text(result, self.pvalue_thresholds)] + + if was_list: + return annotations + + return annotations[0] elif self.text_format == 'simple': return simple_text(result, self.simple_format_string, - self.pvalue_thresholds) + self.pvalue_thresholds, self.show_test_name) def get_configuration(self): return {key: getattr(self, key) for key in CONFIGURABLE_PARAMETERS} @@ -175,3 +207,11 @@ def get_configuration(self): def sort_pvalue_thresholds(pvalue_thresholds): return sorted(pvalue_thresholds, key=lambda threshold_notation: threshold_notation[0]) + + +def get_corrected_star(star: str, res: StatResult, + correction_format="{star} ({suffix})") -> str: + if res.significance_suffix: + return correction_format.format(star=star, + suffix=res.significance_suffix) + return star diff --git a/statannotations/_version.py b/statannotations/_version.py index df12433..f6b7e26 100644 --- a/statannotations/_version.py +++ b/statannotations/_version.py @@ -1,1 +1,1 @@ -__version__ = "0.4.2" +__version__ = "0.4.3" diff --git a/statannotations/format_annotations.py b/statannotations/format_annotations.py index 81bb6e8..ef87581 100644 --- a/statannotations/format_annotations.py +++ b/statannotations/format_annotations.py @@ -5,8 +5,8 @@ from operator import itemgetter -def pval_annotation_text(result: Union[List[StatResult], StatResult], - pvalue_thresholds: List) -> Union[List[str], str]: +def pval_annotation_text(result: List[StatResult], + pvalue_thresholds: List) -> List[tuple]: """ :param result: StatResult instance or list thereof @@ -14,11 +14,6 @@ def pval_annotation_text(result: Union[List[StatResult], StatResult], :returns: A List of rendered annotations if a list of StatResults was provided, a string otherwise. """ - was_list = True - - if not isinstance(result, list): - was_list = False - result = [result] x1_pval = np.array([res.pvalue for res in result]) @@ -34,27 +29,27 @@ def pval_annotation_text(result: Union[List[StatResult], StatResult], x_annot[x1_pval <= pvalue_thresholds[-1][0]] = pvalue_thresholds[-1][1] - x_annot = [f"{star}{res.significance_suffix}" - for star, res in zip(x_annot, result)] - - return x_annot if was_list else x_annot[0] + return [(star, res) for star, res in zip(x_annot, result)] -def simple_text(result: StatResult, pvalue_format, pvalue_thresholds) -> str: +def simple_text(result: StatResult, pvalue_format, pvalue_thresholds, + short_test_name=True) -> str: """ Generates simple text for test name and pvalue. :param result: StatResult instance :param pvalue_format: format string for pvalue :param pvalue_thresholds: String to display per pvalue range - :param result: StatResult + :param short_test_name: whether to display the test (short) name :returns: simple annotation """ # Sort thresholds thresholds = sorted(pvalue_thresholds, key=lambda x: x[0]) - text = result.test_short_name and f"{result.test_short_name} " or "" + text = (f"{result.test_short_name} " + if short_test_name and result.test_short_name + else "") for threshold in thresholds: if result.pvalue < threshold[0]: @@ -63,4 +58,4 @@ def simple_text(result: StatResult, pvalue_format, pvalue_thresholds) -> str: else: pval_text = "p = {}".format(pvalue_format).format(result.pvalue) - return text + pval_text + result.significance_suffix + return result.adjust(text + pval_text) diff --git a/statannotations/stats/StatResult.py b/statannotations/stats/StatResult.py index 45f3c2f..a71c0e6 100644 --- a/statannotations/stats/StatResult.py +++ b/statannotations/stats/StatResult.py @@ -44,7 +44,7 @@ def formatted_output(self): stat_summary = '{}, P_val:{:.3e}'.format(description, self.pvalue) - stat_summary += self.significance_suffix + stat_summary = self.adjust(stat_summary) if self.stat_str is not None or self.stat_value is not None: stat_summary += ' {}={:.3e}'.format(self.stat_str, self.stat_value) @@ -55,8 +55,13 @@ def formatted_output(self): def significance_suffix(self): # will add this only if a correction method is specified if self._corrected_significance is False and self.pvalue <= self.alpha: - return ' (ns)' + return 'ns' return "" def __str__(self): return self.formatted_output + + def adjust(self, stat_summary): + if self.significance_suffix: + return f"{stat_summary} ({self.significance_suffix})" + return stat_summary diff --git a/usage/example.ipynb b/usage/example.ipynb index 91a3e66..07fe9f3 100644 --- a/usage/example.ipynb +++ b/usage/example.ipynb @@ -231,7 +231,7 @@ }, { "data": { - "text/plain": "(<matplotlib.axes._subplots.AxesSubplot at 0x7fc28184f6a0>,\n [<statannotations.Annotation.Annotation at 0x7fc28164aeb8>,\n <statannotations.Annotation.Annotation at 0x7fc28183d6d8>,\n <statannotations.Annotation.Annotation at 0x7fc282ada5c0>])" + "text/plain": "(<matplotlib.axes._subplots.AxesSubplot at 0x7fc8f09eb6d8>,\n [<statannotations.Annotation.Annotation at 0x7fc8ef839c50>,\n <statannotations.Annotation.Annotation at 0x7fc8ef84fe80>,\n <statannotations.Annotation.Annotation at 0x7fc8ef86d940>])" }, "execution_count": 6, "metadata": {}, @@ -247,6 +247,68 @@ } ] }, + { + "cell_type": "markdown", + "source": [ + "To have `\"ns\"` instead of `\" (ns)\"`, configure with `correction_format=\"replace\"`." + ], + "metadata": { + "collapsed": false, + "pycharm": { + "name": "#%% md\n" + } + } + }, + { + "cell_type": "code", + "execution_count": 7, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "p-value annotation legend:\n", + " ns: p <= 1.00e+00\n", + " : 1.00e-02 < p <= 5.00e-02\n", + " : 1.00e-03 < p <= 1.00e-02\n", + " : 1.00e-04 < p <= 1.00e-03\n", + " **: p <= 1.00e-04\n", + "\n", + "Thur vs. Fri: Mann-Whitney-Wilcoxon test two-sided with Benjamini-Hochberg correction, P_val:6.477e-01 U_stat=6.305e+02\n", + "Thur vs. Sat: Mann-Whitney-Wilcoxon test two-sided with Benjamini-Hochberg correction, P_val:4.690e-02 (ns) U_stat=2.180e+03\n", + "Sun vs. Fri: Mann-Whitney-Wilcoxon test two-sided with Benjamini-Hochberg correction, P_val:2.680e-02 (ns) U_stat=9.605e+02\n" + ] + }, + { + "data": { + "text/plain": "(<matplotlib.axes._subplots.AxesSubplot at 0x7fc8f0c539b0>,\n [<statannotations.Annotation.Annotation at 0x7fc8f0ad47f0>,\n <statannotations.Annotation.Annotation at 0x7fc8f0ad4978>,\n <statannotations.Annotation.Annotation at 0x7fc8f0ad4be0>])" + }, + "execution_count": 7, + "metadata": {}, + "output_type": "execute_result" + }, + { + "data": { + "text/plain": "<Figure size 432x288 with 1 Axes>", + "image/png": "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\n" + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "ax = sns.boxplot(data=df, x=x, y=y, order=order)\n", + "annot.new_plot(ax, data=df, x=x, y=y, order=order)\n", + "annot.configure(comparisons_correction=\"BH\", correction_format=\"replace\")\n", + "annot.apply_and_annotate()" + ], + "metadata": { + "collapsed": false, + "pycharm": { + "name": "#%%\n" + } + } + }, { "cell_type": "markdown", "metadata": {}, @@ -263,7 +325,7 @@ }, { "cell_type": "code", - "execution_count": 7, + "execution_count": 8, "metadata": {}, "outputs": [ { @@ -320,7 +382,7 @@ }, { "cell_type": "code", - "execution_count": 8, + "execution_count": 9, "metadata": {}, "outputs": [ { @@ -328,7 +390,7 @@ "text/plain": " carat cut color clarity depth table price x y z\n0 0.23 Ideal E SI2 61.5 55.0 326 3.95 3.98 2.43\n1 0.21 Premium E SI1 59.8 61.0 326 3.89 3.84 2.31\n2 0.23 Good E VS1 56.9 65.0 327 4.05 4.07 2.31\n3 0.29 Premium I VS2 62.4 58.0 334 4.20 4.23 2.63\n4 0.31 Good J SI2 63.3 58.0 335 4.34 4.35 2.75", "text/html": "<div>\n<style scoped>\n .dataframe tbody tr th:only-of-type {\n vertical-align: middle;\n }\n\n .dataframe tbody tr th {\n vertical-align: top;\n }\n\n .dataframe thead th {\n text-align: right;\n }\n</style>\n<table border=\"1\" class=\"dataframe\">\n <thead>\n <tr style=\"text-align: right;\">\n <th></th>\n <th>carat</th>\n <th>cut</th>\n <th>color</th>\n <th>clarity</th>\n <th>depth</th>\n <th>table</th>\n <th>price</th>\n <th>x</th>\n <th>y</th>\n <th>z</th>\n </tr>\n </thead>\n <tbody>\n <tr>\n <th>0</th>\n <td>0.23</td>\n <td>Ideal</td>\n <td>E</td>\n <td>SI2</td>\n <td>61.5</td>\n <td>55.0</td>\n <td>326</td>\n <td>3.95</td>\n <td>3.98</td>\n <td>2.43</td>\n </tr>\n <tr>\n <th>1</th>\n <td>0.21</td>\n <td>Premium</td>\n <td>E</td>\n <td>SI1</td>\n <td>59.8</td>\n <td>61.0</td>\n <td>326</td>\n <td>3.89</td>\n <td>3.84</td>\n <td>2.31</td>\n </tr>\n <tr>\n <th>2</th>\n <td>0.23</td>\n <td>Good</td>\n <td>E</td>\n <td>VS1</td>\n <td>56.9</td>\n <td>65.0</td>\n <td>327</td>\n <td>4.05</td>\n <td>4.07</td>\n <td>2.31</td>\n </tr>\n <tr>\n <th>3</th>\n <td>0.29</td>\n <td>Premium</td>\n <td>I</td>\n <td>VS2</td>\n <td>62.4</td>\n <td>58.0</td>\n <td>334</td>\n <td>4.20</td>\n <td>4.23</td>\n <td>2.63</td>\n </tr>\n <tr>\n <th>4</th>\n <td>0.31</td>\n <td>Good</td>\n <td>J</td>\n <td>SI2</td>\n <td>63.3</td>\n <td>58.0</td>\n <td>335</td>\n <td>4.34</td>\n <td>4.35</td>\n <td>2.75</td>\n </tr>\n </tbody>\n</table>\n</div>" }, - "execution_count": 8, + "execution_count": 9, "metadata": {}, "output_type": "execute_result" } @@ -346,7 +408,7 @@ }, { "cell_type": "code", - "execution_count": 9, + "execution_count": 10, "outputs": [ { "name": "stdout", @@ -420,7 +482,7 @@ }, { "cell_type": "code", - "execution_count": 10, + "execution_count": 11, "metadata": {}, "outputs": [ { @@ -428,7 +490,7 @@ "text/plain": " total_bill tip sex smoker day time size tip_bucket\n0 16.99 1.01 Female No Sun Dinner 2 (0.991, 4.0]\n1 10.34 1.66 Male No Sun Dinner 3 (0.991, 4.0]\n2 21.01 3.50 Male No Sun Dinner 3 (0.991, 4.0]\n3 23.68 3.31 Male No Sun Dinner 2 (0.991, 4.0]\n4 24.59 3.61 Female No Sun Dinner 4 (0.991, 4.0]", "text/html": "<div>\n<style scoped>\n .dataframe tbody tr th:only-of-type {\n vertical-align: middle;\n }\n\n .dataframe tbody tr th {\n vertical-align: top;\n }\n\n .dataframe thead th {\n text-align: right;\n }\n</style>\n<table border=\"1\" class=\"dataframe\">\n <thead>\n <tr style=\"text-align: right;\">\n <th></th>\n <th>total_bill</th>\n <th>tip</th>\n <th>sex</th>\n <th>smoker</th>\n <th>day</th>\n <th>time</th>\n <th>size</th>\n <th>tip_bucket</th>\n </tr>\n </thead>\n <tbody>\n <tr>\n <th>0</th>\n <td>16.99</td>\n <td>1.01</td>\n <td>Female</td>\n <td>No</td>\n <td>Sun</td>\n <td>Dinner</td>\n <td>2</td>\n <td>(0.991, 4.0]</td>\n </tr>\n <tr>\n <th>1</th>\n <td>10.34</td>\n <td>1.66</td>\n <td>Male</td>\n <td>No</td>\n <td>Sun</td>\n <td>Dinner</td>\n <td>3</td>\n <td>(0.991, 4.0]</td>\n </tr>\n <tr>\n <th>2</th>\n <td>21.01</td>\n <td>3.50</td>\n <td>Male</td>\n <td>No</td>\n <td>Sun</td>\n <td>Dinner</td>\n <td>3</td>\n <td>(0.991, 4.0]</td>\n </tr>\n <tr>\n <th>3</th>\n <td>23.68</td>\n <td>3.31</td>\n <td>Male</td>\n <td>No</td>\n <td>Sun</td>\n <td>Dinner</td>\n <td>2</td>\n <td>(0.991, 4.0]</td>\n </tr>\n <tr>\n <th>4</th>\n <td>24.59</td>\n <td>3.61</td>\n <td>Female</td>\n <td>No</td>\n <td>Sun</td>\n <td>Dinner</td>\n <td>4</td>\n <td>(0.991, 4.0]</td>\n </tr>\n </tbody>\n</table>\n</div>" }, - "execution_count": 10, + "execution_count": 11, "metadata": {}, "output_type": "execute_result" } @@ -441,7 +503,7 @@ }, { "cell_type": "code", - "execution_count": 11, + "execution_count": 12, "metadata": { "scrolled": true }, @@ -450,7 +512,7 @@ "data": { "text/plain": "[(0.991, 4.0], (4.0, 7.0], (7.0, 10.0]]\nCategories (3, interval[float64]): [(0.991, 4.0] < (4.0, 7.0] < (7.0, 10.0]]" }, - "execution_count": 11, + "execution_count": 12, "metadata": {}, "output_type": "execute_result" } @@ -463,7 +525,7 @@ }, { "cell_type": "code", - "execution_count": 12, + "execution_count": 13, "metadata": { "scrolled": false }, @@ -526,7 +588,7 @@ }, { "cell_type": "code", - "execution_count": 13, + "execution_count": 14, "metadata": {}, "outputs": [ { @@ -599,7 +661,7 @@ }, { "cell_type": "code", - "execution_count": 14, + "execution_count": 15, "metadata": {}, "outputs": [ { @@ -652,7 +714,7 @@ }, { "cell_type": "code", - "execution_count": 15, + "execution_count": 16, "outputs": [ { "name": "stdout", @@ -672,9 +734,9 @@ }, { "data": { - "text/plain": "(<matplotlib.axes._subplots.AxesSubplot at 0x7fc284d10f60>,\n [<statannotations.Annotation.Annotation at 0x7fc281c837b8>,\n <statannotations.Annotation.Annotation at 0x7fc2817bfc18>,\n <statannotations.Annotation.Annotation at 0x7fc284d8e518>])" + "text/plain": "(<matplotlib.axes._subplots.AxesSubplot at 0x7fc8f230e160>,\n [<statannotations.Annotation.Annotation at 0x7fc8f0f07e10>,\n <statannotations.Annotation.Annotation at 0x7fc8f234cfd0>,\n <statannotations.Annotation.Annotation at 0x7fc8efa6e048>])" }, - "execution_count": 15, + "execution_count": 16, "metadata": {}, "output_type": "execute_result" }, @@ -729,7 +791,7 @@ }, { "cell_type": "code", - "execution_count": 16, + "execution_count": 17, "metadata": { "scrolled": true }, @@ -762,7 +824,7 @@ }, { "cell_type": "code", - "execution_count": 17, + "execution_count": 18, "metadata": { "scrolled": false }, @@ -814,7 +876,7 @@ }, { "cell_type": "code", - "execution_count": 18, + "execution_count": 19, "outputs": [ { "name": "stdout", @@ -867,7 +929,7 @@ }, { "cell_type": "code", - "execution_count": 19, + "execution_count": 20, "metadata": { "scrolled": false, "pycharm": { @@ -926,7 +988,7 @@ }, { "cell_type": "code", - "execution_count": 20, + "execution_count": 21, "outputs": [ { "name": "stdout", @@ -947,7 +1009,7 @@ { "data": { "text/plain": "<Figure size 864x432 with 1 Axes>", - "image/png": "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\n" + "image/png": "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\n" }, "metadata": {}, "output_type": "display_data" @@ -1009,7 +1071,7 @@ }, { "cell_type": "code", - "execution_count": 21, + "execution_count": 22, "outputs": [ { "name": "stdout", @@ -1029,16 +1091,16 @@ }, { "data": { - "text/plain": "(<matplotlib.axes._subplots.AxesSubplot at 0x7fc283af2668>,\n [<statannotations.Annotation.Annotation at 0x7fc283bf3780>,\n <statannotations.Annotation.Annotation at 0x7fc283be1dd8>,\n <statannotations.Annotation.Annotation at 0x7fc283bc3ef0>])" + "text/plain": "(<matplotlib.axes._subplots.AxesSubplot at 0x7fc8f23c5518>,\n [<statannotations.Annotation.Annotation at 0x7fc8efb92748>,\n <statannotations.Annotation.Annotation at 0x7fc8efb92ac8>,\n <statannotations.Annotation.Annotation at 0x7fc8efe8ef28>])" }, - "execution_count": 21, + "execution_count": 22, "metadata": {}, "output_type": "execute_result" }, { "data": { "text/plain": "<Figure size 864x432 with 1 Axes>", - "image/png": "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\n" + "image/png": "iVBORw0KGgoAAAANSUhEUgAAAvoAAAF9CAYAAAB1QswoAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjEsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+j8jraAAAgAElEQVR4nOydd3gc5bX/P7NVq131bhV3Sy5y78Y2NjY2LoAhCSWQECDAJQmXhIQQuAn3EvIjISGVJHRCIIWOTTEY2xj3XmXLllzVe9te5/fHSqPdnVWxLWOsvJ/n4cEzO+XM7Iz2e857znklWZZlBAKBQCAQCAQCQb9Cc7ENEAgEAoFAIBAIBH2PEPoCgUAgEAgEAkE/RAh9gUAgEAgEAoGgHyKEvkAgEAgEAoFA0A8RQl8gEAgEAoFAIOiH6C62Af2RQCCA3W5Hr9cjSdLFNkcgEAgEAoFA0A+RZRmv14vZbEajUcfvhdC/ANjtdkpKSi62GQKBQCAQCASC/wBGjBhBXFycar0Q+hcAvV4PBG+6wWC4yNacG0VFRYwZM+ZimyEQfGkR74hA0D3iHREIeuZ83xOPx0NJSYmiPSMRQv8C0JGuYzAYMBqNF9mac+dStl0g+CIQ74hA0D3iHREIeqYv3pOuUsVFMa5AIBAIBAKBQNAPEUJfIBAIBAKBQCDohwihLxAIBAKBQCAQ9EOE0BcIBAKBQCAQCPohQugLBAKBQCAQCAT9ECH0BQKBQCAQCASCfogQ+gKBQCAQCAQCQT9ECH2BQCAQCAQCgaAfIoS+QCAQCAQCgUDQDxEz4woEAgHgcvs4UFpPSoKJYbmJXW6371gdb6wrobmljVa5nPmTc79AK/sef0BGq4k+o+KXAa/Pj0ajOS8brQ4P5bVWhmQnEGMI/uzVNTlAgvSk2B73b7W50Wo1WEx6AgGZN9eVcOB4A2OGJGPQaUGSmD85l+T4mLD9AgGZslorKQkxxMUaztn+vsYfkDlYWo9Oq2H0kBQ0XdxbWZapbrCTGGckNkZ/3udtsbr5ZMdpnC4f8yfnkpcZf17Hszm9HDnZSG5GHFmp5vO2rztcbh8fbz9NVYOdGWOymJCffkHPJxD0FULoX+K4PD4On2wkM8VMdprlYptzSVFW08a/Py2hqc3FvEm5LJo+8GKbpKKx1cm+Y/XkpFsoGJR8sc254Lg8PtZsP0N1g53phVmMG57WJ8f1+QNIktSlWKyst/HQ05tpsbkBWDR9IN/96nhkWeZ0dRvJ8TEkWIzUNjl47MUd+PwBAH73r72kJsYwdljf2Nlbjp5pYsOeChIsRpbMHESC5eynTy8pa+YPr++jrMbK+BFp/OCmiSRFCNXesGFPOVsOVhFj1KHXatptGkxakumsjxWKzx/g6Tf3s2FPBbExOm5dMoqrZgxSbef2+nn+vUNsPVhFRoqZu68tDHtXPt9bwR9f34fHF8Bi0nHl9EEcKK3nREUrAHMn5PD9mydGfTb8/gC///c+Pt9XgUaSWD57CIdPNlJa3gLAoeMNyrYrN57gTw/MIzEu+F3UNjl49LltVNbb0Os03HH1GJbOGhx2fFmWabG5SbQYu5y+Phr1zU72HK0lK8XM2OGpZ7Wvw+XlJ3/ewsmq4PWPHpLCz++eiV4XPsBf1+zgsRe2c6bGSoxBy90rxrJgal6vzxOJy+PjgT9uDDpYwPubT/Hb++cwMDMem9PL2p1lWB0eLp+YQ25GXI/HO3Kqkf99fjtOtw9JgtuXj+baucPO2b6e+PlLOzjY/n2v3nqaB2+ZzOwJ2RfsfAJBXyGE/iVMWU0bj/x1qyJOvrZgBLdeNfIiW3Vp4Pb6eeSZrbRYg/fu8MlGjHoNl0/68kRndx2p4bEXdyjLMwsz+clt076w8/sDMoGArBIAF5JfvLST/aX1AHyw5RQP3jqZ2ePP/cc0EJB5fuUhPtl+BoNOww0L81lxuVoMvP7pMeU9Avhk+xkun5jDX94+SHmtFZ1W4htLRmEy6hSR38GuI7XnJfT3Hq3jo62nMBq0XHf5MIbmJLK7uJYXVh6ivsXF3AnZ/Nf1Y9HrtAAUnWjgkWe2EgjIAGzcV8HTP5yHRiNxoLSe6kYHkwsyuhXagYDMr17drYiu/SX1vLCyiB/dOlnZxuXxKdHvrliz4wx/emO/av363WX89cdXdBkFdri8nKxsJS8znnhz9Ej36q2nWberHACrw8tf3z7AuOGpDEi1YHV4eOadgxwsbcBo0FLbfh1WRwuPv7yDl396JXqdFr8/wPMrD+HxBb8zm9PHO58dDzvP5/sqmDUuixmFA1Q2bNhbwYa9FQD4ZZn3Pj/R5b1osbrZuL+Cq2cPBeCfnxylst4GgNcX4MVVRcyZkK1E9o9XtPDkq7upbrCTlWrmtqWj+HDLKU5XtzExP53r5g2nst5Kfl4yaUkmthdV87cPDtPU5sLtDSjf/8Kpedx3w4Qu7Yrksz0VisiH4N++7YeqVaL1Hx8f5UyNFQCXx88z7x5k5tisc47s7y6uVZ43AI/Xz9qdZXxz6SgeenqTcq73Nhznye/NZmhO16NqHfY53T4AZBle+/goi6cPIsbY+cyW11r5ZPsZtBqJq2YOIjPl3KL+VfU2ReR3sHrbaSH0BZcEQuhfwrz+aUmYOHlrfSlLZw1WDR8L1BSfalREfgdbD1V3K/QDAZmdR2qoqLMxeWQGg7LOb9i5J37zjz0R9tXQ3OYkKf78IqXdIcsyZTVW9hyt5c11pbg8PuZPzuPe68ei1fat4K9ptPPq6mKqG+zMKMxixtgsReR3sHrraUXo1zY5eOadg5yoaGHssDTuvq6wx3SIjfsq+GDzKSAotl56/zCFQ1NVqTklZc2qfd9aX0p5bVB8+Pwyf/vwCA99Y4pqu5z0nqOPPn+AijobWalmjHqtsv7YmSb+74VttGs2dh2p5Xffn8Ov/r4Ll8cPwKc7y0hPjuXGhfnKcofIA6ios1F0opGN+ytZs+MMAAadhsfunsnoISnKdv6AzOmqVtKTY/H5AmGiC+BY+z0oKWvmt//cS2W9jYKBSfzwlslkJEdPb9mwpyLq+qY2NzuP1HL5xBzVZ4dONPCLl3Zgd/kw6DTcf9PEqM7c8YqWsGVZhpOVrQxItfDsO4fYuK8y6rlbbR7O1FgZlpOI2+un1eaJul0o5bVWZhSq13+2p7zHfUPRhbwj1Q32sM+8vgC1jXb2Hq2j1eZmzY4zyjbVDXae+scexSHZsLeCz/dWIANajcR/XT+WZ945pHIyAdbuKuOmKwt6PYJidajvR5vdrVpXWWcLW3Z7/DS0OMnL7BT6NY12nn33EMcrWhg7LJW7V4zt0nGL5jQaDVoOljYoIh/A4wvwyfYz3HJVLOW1VoZmJ2A0aKlpdJCSEIOh/f1ptYXb7Pb4WbnxBItnBEe4ahrtPPCHz3G6O9+jvzw4XxlxORtijDo0EoS8dsTGCPkkuDQQT+olTEvEH7pAQKbN7hFCvxdkppiRpKB46CCrh2jP02/u59OdZQC8urqYR26bytTRmRfMRofLp1p36GQDc8ZfmFGHxlYnjz63LexHF4JR22E5CVw1c3AXe3bSYnVjdXh6HHoPBGQefW4bVe1Cp7S8BZdbfb02Z6co+fVruzl2JihGP99XwdEzTfzP7dNUDldVg41/fnyMumZH1HSM0ooWldCXUG9XFnEfAgGZBIuea+cO5f1NJ/EHZCYWpDN2WGq311pS1szjL+2g2erGbNLzo1smMakgA4DNB6rCxIPT7ePTHWWKyO+g+FST8u9oAsPt9fPpzjPKsscX4O3PShWhX1Fn5dHnt1PX5MCg03DXikKy08xU1neK0dFDUpBlmaf+sUf5Xo6eaeaZdw7y6J3To15bUnzXosliih75fWlVEfb2Z9vjC/D8e4eYNXYAGo1Eea2V2BgdKQkmxg5LZf3uTqGt02oY2Z6Sc/B4fdRjQ/D+5LSnMR4obSAu1hBV3IYT/v3LssyvX9vDgdKGLrZXk5ViZs6ETsdm5tgBFJ/u/N6y0yw8v7KIIyHfZSgdIl+xof3//oDMPz4+GlXkB20N1jH0ljnjs3lzXSkeb3Afs0nPjLHq0YypozMV5y9ov1nl1P76td2UlAUdso37KkEmbFQolAn56YwanKxcf0pCDFfNGERVvV21bV2Tg2899gkeX4DYGB2xRh0NrS7iYvXcf+NEpo7O5Iopebz0/uGw/V77+ChvrS/l2Z9cwed7KxSRD0EHZ+uhKpb04u9YJMnxMSy7bAirNp0EwGTU8tUrhp/1cQSCi4EQ+pcw8yfnhg0nDs1JuOBR5v5CZoqZmxcV8O81x/AHZIbmJHDdvK7zO1usbtbtKlOWAwGZdzYcv6BCX6uR8IeqQCA/78Ll6b/+aYlK5HdQUtbco9D/5ydHeWNtCf6AzIi8RB69c0aX0b3yWqsiJjvYXVyr2q7DEXO5fYrI76C2ycGPn97EMz++Qskt9wdkfvbsNiWVIxKNBGNCotwdGAzaKNuqxb/PD3dcPYZr5w7l0Wc2sPdoHXc9sZbhuYk8dtcMLFFGGJ577xDN7aNHdqeXv7x1gBceWYgkSaQlqqOw+QOTiDFow8R+aM75NXOGsvlAlTIiNaMwi9yMuDCnFVCEHAQd044IvscX4IWVRTx+z0xeXHWY09WtTMzP4M5rxnC8vEX1vZyoaMHl8VF8qonMFHNY0eONC/MpOtFAU1t40GHssNQuixUbWl1hy602N212N794eSdHzzQjAVfPGcodV4+mtsnBmh1niIs1cOtVI0lJCN6vjvSmSDQS/OiWycQYdRSfauKJV3aq7ks0ItPTik40sml/9BGDaMTG6Pj9D+ZiMurYsKecQycaGZIdz7CcBI5XtKLVSAzPTWDD3q6PadBr8Hiji3m3149OK+Hzqy9m5OBkBpxFfdaANAu/uW82q7eeRquVwkaB/QGZlZ8fZ+eRWrLTzFw3bxgHSusZkGrh1qtGhhXtOt0+ReR3cCCKA3agtJ5TVW2MHZZCXGyn82fQadBpNYwZmkLh0FQOnWgIO07HtTpcPiXoYXV4+dOb+3m54EpWXD6MBIuRdz4rDfu75fL4eXHVYQoGqv9Wms+joPjb1xYyZ0I21Q12xo9IP6eRAYHgYiCE/iXMFVPyMBq0bDlQRWaKmWvnDr3YJl1S3Lgwn8XTB9Fmd/fY/UFGJvInVu6NgjgP4mINqlEbXR+nz4RS3aiOrHXg9kaPGLba3Ly6upgjpxopr+0c6i8pa+G9z4/zjSWjou6XkmjCoNeGidEBaRZOV7eFRbg7OqLEGHVkp1mUnOcOHC4f2w/XKEWapypbVSI/IykWvyxj1Gu56cr8qKMN6ckmTla2hq0bmBVHbXP4sbLTgiJ3+6FqztR1RolLy1u4/3ef8+xPFoSNIpyqalWKPjuob3Hi8wfQ67QsnDaQzQeqlMjv3Ak5TB+TxYO3Tub5lUU0tDiZOyGH60Oc0MwUM88+dAV7jtaRaDEyZmgKkiQxqSCdPUfrgGB8elhOopJnX9MYfh0uj59AABItBhItRtKSTJiMOv4YJd9+aHYi3/7FWlpsbiQJbl5UoKQR5WbE8fzDCyk+3URyvJGKOjsmo5axw9K67OQyZ0I2qzaeVJanF2bx3ucnONruyMkEC1svn5TDzYsKuHlRgeoYd68opKnVycmqtrD1CRYjk0cGR0u2F1X3SuSbjDqG5Sby7DsH8foDLJ4+iKY2V887hjBlZAaPPLOVplanyumBoICOJvIzU8wEZJlRg5KZMyE7rCYnlEXTBzEsJ4Gn3zyg5KV3MP0cgg2DByRw71fGha3bXVzL8ysPKRH2wycbGZqTwO+/f3nUY5iMOtWoUGRDiLX7W9l8ZCuAagS1utHBR1tPc9OV+dy2bCQP/GGT8lk0h6aDFqsbm8NLYpwxGOwqrVcFKOqbndx7fS6rt51W0u9G5CUyc2xWl8ftDfkDk8mP4kAIBF9mhNC/xLlsXDaXjRMFQedCea2VN9aW0Gx1MX9yLvMnd91RIikuhnmTcpVUAo3EBe3wAOAPqKN7/m5+AM+XrlItgDAR73L7ePrNA2w5WIlWo+nSCYjMUY48113XFvL8ykO4PX5yM+L41rLRHK9oCROl40d0psX84OaJPPbCdlrt4WkYySGRtdCIYQdGo5Y//2h+l7YAxOjD/xRKEhij5BTvPVbH5v1VKocDgiMMR042UtieymN1ePjJnzerUi4mFWQohbUmo44nvzebrQeriIvVU9he1DtlVCZTRnUt4GJj9Kq89odvm8q6XWW8tf44dc0O3v7sOBv3V/Lkd2er7suQ7AR+/dpu6lucALz3+QlqG+2crg4Xzga9Bq1WUhxOWQ4WLl81o7PTj0GvVboj5Wb0PKL4rWWjSY6L4eDxBoblJvKV+cP50R83qrbbe7SWYREFmR6vn+1F1bg9fh78xhTu+eW6sM91IZH5aO0WR+Ql0tTmxqDTkJZkIjc9jtkTsnn8pR1YHV4A1u8uj1qw3YFWA6Ff6cDMOD7vol4gksyUWOX5Nui1PPSNyWFFp3PGZ7MxZCQhMyWWr8wfzsKpA9FoJGRZXbszeEBCr87dHSVlzTz2wnZVMONERSs1jfYui1jvv3EiD/91C972tKPS8hY+3naaZquboTkJbDvaKcCjOV0defbrd0ev9YhGfl5SWDR96WWDWbc7vJZi3uQczCY9f/jB5ew7VodGIzFhRFqf1xkJBJcCQugLLilkWWZ/ST1VDXYmFaSfcxcFt9fPw3/doqQ/HChtwKDXdus03XfDBKaMylCKcSNFSF8TGbkDaHN4SO+iMPJ86S6P2RziBLy+toTP9wV/mH3+rnODZ0bJ+w1l0fSBzB4/gMZWFznpFmoaHarI85odZSxv72IyIi+J5x5ewMN/3aJEySePzKBgUDIvv19EbZOTscPUaTkOl7dbOwBVNF+WoaFZnf7z9Bv7CXTja5lCOn7sOVqn5KJ3kJcZx/dvmsBrq4tZve00JqMOh8uriMzRQ1L45XcuA4JpPlaHh8wUM8fONHGyspXCYaldFv8a9FoGpFmoC7G7vtnJH1/fF5ZrbtBp+MZVBfzvC+HR491H68hKNYc5aBNGpGNzht8/nz9YC3QuLT0hOCp1/fzhXD+/M8c5KT5GFZWNFOpen58f/WmT8l2ZjOp0K1vIM3zFlFx2HK5hd3EtkgRXThvId74yTtWK8pPtp5X7HzxPgHc+K41qe4xegzsilz4yt74rdFqJR++cweGTDbTY3Fw2bgAnK9tYv6ec0YNTmDl2AN/72ng0Goni002MGpzMvV8ZF1bEetn4bPaV1PHZ7nIkSWLJzEHn3M/d5fFh1GuRJIkPt5xSiXwI3uPEkO95d3Etr35UTKvdzYKpeRQOTVFEPgTv3Z/fOqAsR2a/aTSSUkiu0UjMmxSsaRiao3ZWBqSaSU+OZcyQFGwOL/tL6xmUFc9ty8JHCYfnJnHPdYW89vFRvF4/V80YzFUzgmmGep2GeIuB4+UtlNeZRGqrICoffPABq1atOuv9Fi9ezHXXXXcBLOpbhNAXXFL8+a0DfLI9WHSo02r4v7umn1Nrw8Mn1V13Nh+o6lboazXSFzp6EhQk4T+/0QpWz5XdxbW8uroYm9PLomkDGZBm6bL4cHHIHANHz0QvJoRgRD0vM575k3N71RYzNkavtOurrFfXB9Q22VXb/+7+uRw704xOp2HIgAS++dgnyne55WCV6hjRCm0jiVboOG54GsUhdQEWk14lekPTEaaPyQwr8o2Wfz9/Ui57j9bx+toSANoiRicOn2xk/a4ybC4vr3xYjMfrJyUhhsb2vHaNRuLBWyYza1y4E2V3etleVM2ZiIg8QHld+H31+AJdpoj88OuT+P2/91Jea2P0kBTuXjGWvcdqOXyyUdlmWG5ir/qcnw03LBjB/pLO/O4Ei4HpY8LTLHYerg1zyEILLTsIFZ16nZZH75xOTaMdvU6j5PdHEm9WOyxdpY64ouTQ9zSRV1ysgfRkE19fVEBOuoWc9GB6y98/OsKb64IOxaqNJ7npyvxgjn97O8/aJgfDchOVdp0d57r/xonMGjuAF1cd5sMtp6hpcnD/jROparCxauNJZFlm+ewhjBqsdnoBmtpc/Pq13RSdaCQjOZb/vmECzVFSlSTg29cUKu0qW6xu/t/fdir3+PVPS/B6e0gXjbiNs8cNIN5ixOHysnDqQCUNZv6kXN757DgV7Z1+YmN0PPGdWST3ssPY0llDWDpriGr9m+tK+PtHxcHrkYIjEJf6BHeCLwclJcG/4ULoCwR9SFObi093dHYW8fkDvL3++DkJ/WjtAjPPMVJ+urqNtz8rxenysXjGICVH+HxJTzKF5b9K0OWP99nS2OrkFy/vVATuq6uLuXlRPjqtJqroPVHZqrQezUm3UHSiUbVNakIMj9w+7ZxHOqKlDkWrSZAkSSlO3XqwSuWwRTJuRM/Px6jBKaoi1IXTBzJwQDyf760gNcFEVqqZ51cWhW3zg5smApAYZ1RN7jV6SAoLp+YpnZry85JYPGMQL64KP0Yk/1xzlLpmp+JANIYUrwYCMv/+9FiY0G9sdfKD329U8spDizZ1WokxQ1IV8agcJ4qOjTXqGJGXxF8evAKvz6+kFy2aPogYg46th6rISjF3m9YSDVmWaWhxkRxv7DZ1ItStlQNB0d5hA/Sus0xoD/UOoo36+fwB9pfUI0kwKT+NMUNTlGd6YGYc/oCsiM6eWDprMJV1NlZvO60qngfQ6yTGDk3lnQ3HWbPjDDcsyGdYbmJYnQIE6xK0mvD78/x7RewpruO7Xx2vtM/0+wM8/eYB5fvedaSWv7y1n11HapXRhe1FNfzxgcvJzYijtLyZ3cV15GXEMaMwi5ffP6xca22Tg9/8Y3fUd+QrVwxDBv7vhe1kpsQyNDshzJGCYGF1waBkjp6O7vxH3o3Kehs/vEXdlUer1fD0j+az71gdbo+fSSPTe5zDoSf8/oDiSEHQIX9j7TEh9AUqli1bxrJly85qn7vuuusCWdP3CKEvuGTw+QIqgeI5i7ZyoWSnWfjaghG8tb6UQEBmyICEsxYwEIzIPvTnzdjbI707j9TwxL2XhfUvP1cic8Rlgq0je9O3vSeOnGxSCfrPdld02cJv66Eqpo7KpHBYatRRhcfumsHYYalnlQPbkYZV3WhnckEGCRZ1W9hoUfFQnG51Wo7FpMfjC+Dx+hmUFc8ti9XFnJHodeFRWa1GItFiDKuB8foCHDrRwPaiGgAKB8Uye0JOtxHd+26YwFeuGI7T5WNIdgIvrCoKaxkZjdomZ7efeyJqItZsPxNWPOrzy4wdnkpuehwLpuSRnhxLQ6uTohON6HUalVjrICEk7zlUYAPMnZjD3Cg98XviTHUb/+9vO6lqsJMcH8MPb5lE4VB1O9IN7T3jO2hzeNhfUh+W/jVtTBbpybGq/v+hpCf27Kw73T5+/PQmTrUX8ibFG7E5vGg1wZaSD94ymaY2Ny+uKuLwycawmpDIglIIjqbcfd1Yvn7VSNrsbu791fowwd/U5ubdkIm29pfW8/L/XIk34l1ze/zEGNXP0t5jdfzx9X38/J6ZQLCYO7JYuOhkU1gKkc8fYNuhas7UtPHkq7sVmxdOzVNm9Q21zxSl61RlvZ0313VOLpabHqfqJV9yppkma9AWc4yO2eOz+Xh7ZzAmckyyuyJbrUbqdZDE5w/gD8hhc1JEEpCDYj+Urp59gaA/I4S+4JIhPTmW6WMyFaElSbD8MvVwbW+59aqRLJk5iDa7h0FZ8Wc1jXwHu4trFZEPQRGwZseZPhH6gwfEh6UqWEx60pL6Jj9/cHZ8VNHSFXVNTh59fht/eXB+1AmIzCb9WRe6/emNznkJ9DoND9w8UbVNtILYUOZMyOEvbx8KE7/Xzx/O4hmDaLG6eu0UbWt/pjrwB2R2F9eGCU29TsMj35pGdYMdjUai4lRxj2kbAANSg6kam/ZXqqK4XY2gdMfSy8LbnEYrhi4ta+Znd0xXhNAT915GU5uLynobD/9lS9TjhqaI9BV/feegMlLS1Obij6/v47mfLFC9a0lxaicvcp3JqOO3/z2HdbvKKSlvZssBdZrW6epW2uyeLtu6QtCpOBXSrac5pEvOtkM1/PylHbTZvaoJuwCWzhzEh1tPh703b39WynXzhmEx6bGY9Pz4G1P4y9sHuhxpcrn9wRl/I949WZZZMXcor318VLVPUUjqVFpSLKmJJhpaOh3CQVnxHIiYbC41MYb3NpwIs3XdrjLmTcoNKybPSjVz+aRcVm87o2yr1Ugqh6q8zsrNi/J5f9NJpaahJmQbu8tHdaOdJ+6dRfHpJgoGJvP2p/vZc7xzpOyaOef/jH2w+SSvrS7G5fFz+aQcvvvV8VFH/vQ6DYtnDgp755ZfgGdcIPiyI4S+4JLiwVun8NmecqrqbUwfkxXWX/xcSEkwdZm72xui/Zh3F3E8G765ZBTVDXaOnGoiKc7Id786vtsI1tmQkx7Ht68p5B8fF+P0+Jk7IZuhOYm8sLLrtBKvL8CuI7UYI6J/EsHuIGdDQ4uTtSHzEnh9Ad5vn8E2FIO+e+dBr9Py++/P5bn3DtHY6uSqGYOUH/PuughFEhkl746OItEKtbndcrxcLRxH5CWqJlBSV2bAuGGpDBqQwPgRaaqo58JpA3n38xNhs+U63X5Ky5oZExI9T46PITk+hvtvnMA7G44jEUwvio3RMyE/7ZxS4HoictKxmkYHHl9A9Rwvu2wwWw5WKa0QL5+Yw8jB6nc7wWJU5rvYXlTFL17eFfZ5QIbV205xw4L8Lm2yO7svzt57rOvJuIwGHelJsWEtXD3eAH6/TEfjphmFWSTFG/nRHzd1cZTgjMjTxmSy7VC1sm7q6ExuWBhM6/nTG/vDUrbyByYp/9ZqJB6+bQp/efsg5bVWpozM4O7rxvKHf+9T5hYYPzyN2eOzWb31dPiJJYmbFhWAFAxS5GXEc/eKQgZmxfPgrZNZtfFke7H0MNbvKud4SGtYnVbDkpmDGTkomZ8+uy3qdVU32BkzNFV57lwtiVwxfSSna9qYmJ9+3gGQ8lorz757SFlet6uc4TmJLO0i4HPn1WMYOSiZ0rIWxg5PVSaqE5J6T1UAACAASURBVAj+kxBCX3BJoddpuHLawJ43/ILo6KseSl8VKibFx/Cr787G5vBgMur6vDXc8tlDuGrmIHy+ADFGHbIso9Nq+GjrKZVA6yAr1czOI+HRbxnaizd734XF5w+oRhN0WiksVxpgcXt//O7IzYjj53fP7PW5o5GWaFJ1fbFEadV5PhQOC+ZpdyBJMCgzXiX0r5k7lNgYPf/+9BiBgExuhoXv3zyxS4c0O83CFZNzldERCN7L7PTokyhdMSWPK6Z03Uq2L5kyKiMsValwaGpUZzXBYuRPP5xH8alGLLGGXnVHmVSQEXUSqZ5qNuaMz+aNtSVRu1r1xNjhaSRYjGEzss6blKuqDSgYmBxWnxFJnDk4AVhcrIHi040UDEzmW8tHK9f1+D0z+d2/9lJS1kL+wCTu+9r4sP2H5ybxu/vnhq179M7pnKpqRZaD7VMBrps3nF++slNJt1k0bSAZybHcf6N69CyyVXNGciyHTzXS2OpCo5G4ZXEBCRZj1NGXDiI7bWkkidkTsplN3zQxiOyOBcH6oa6QJEm0oBb8xyOEvkBwHkwsSA+byMmg16hSK86XaLOt9hU6rUYZ9pak4CyZ8yfncvcTa5XZXCEYZZ4/JZdJBemcrm4L65BiMekZmn12vbwzU8xMGZXBriO17eeGZZcNYdzwND7edprqRjuzCgf0qpC2L1gwNY8XV3WKt1ijluG5Sd3scfZMHpnBHVeP4f1NJ9DrNHxtQT6jBiezYV+FMvNncryRGxaMwBJr4MppeTRb3QwZkNDl5FMd3L48OIvsweMNxMbouH35mG4F2RfF3SsKMeq1HDxez7CcJO64enSX22o1UtgIRE/odVp++PXJ/Orvu5QREEkKCu/uSE+O5an/nsPH204jSRKyLLNq08kut9dqJVLiY7hm7lAm5qczMT+djORY9pfWM2RAAgunRnea7rthAtfPH47V7uG59w4pufEmo5YlswZjiTXwvQgB30FOehxP/fdc/AG5V+lhHUT21J9RmMUfHpjHnuJacjPjmHIWjQJy0oOToZWUNZORHEwXAhiYFR/mxBgNWnLSLEwbnclXrhjR6+OfC2OGpqicu7HDv5i/EQLBpYokX+jpPf8DcbvdFBUVMWbMGIzGS3Oa7D179jBp0qSLbcYlgc3hYe2uMmxOL/Mm5apmh7wUqWm0s3LjCWxOL5MLMhg1OEXp+uHzB3jlwyNs3FdJWqKJby0ffU5D8l6fn/W7K6husDG9MCvqlPVfFH5/gFdXF7OhvcNOb66pr96RqgYba3eWoddpuXJa3nmlkjVbXZhj9Bj6KMXrUqCizsq7G47j9QVYNH3QOT2La3eW8e9Pj+H1+blq5mBcbh/bi2rITrNw27JR5z1K5/L42LSvEqvDw2Xjsi/YXBhfJMcrWmixuhk7LLXL5+1C/I7sPFzDax8XY3d6uXL6wG7TtASCC0VH153nnnvuvI91vu9JT5pTCP0LgBD6AkH/R7wjAkH3iHdE0F+5lIS+mA9aIBAIBAKBQCDohwihLxAIBAKBQCAQ9EOE0BcIBAKBQCAQCPohQugLBAKBQCAQCAT9ECH0BQKBQCAQCASCfogQ+gKBQCAQCAQCQT9ECH2BQCAQCAQCgaAfIoS+QCAQCAQCgUDQDxFCXyAQCAQCgUAg6IcIoS8QCAQCgUAgEPRDhNAXCAQCgUAgEAj6If1C6FdUVDB//nzV+vz8fPx+Pz/72c9YtmwZy5cv5/3331f2yc/PZ8uWLWH7zJ8/n4qKCgCefvppli5dytKlS3nyyScv/IUIBAKBQCAQCPoFfr+f9957D7fbfdFs6BdCvztWrVqFzWbjgw8+4JVXXuHxxx/HZrMBoNfr+elPf6osh7J161Y2b97Mu+++y3vvvcfhw4f59NNPv2jzBQKBQCAQCASXGAcPHuS+++7jww8/xG63XzQ7+r3QX7FihRKNr6urQ6/Xo9frAUhPT2fmzJn86le/Uu2XlpbGQw89hMFgQK/XM3ToUKqqqr5Q2wUCgUAgEAgElx4FBQXodDpOnz7No48+etHs0F20M/cxdXV1XHPNNVE/0+l0PPLII6xcuZK77roLo9GofPbQQw+xfPlytmzZwqxZs5T1w4cPV/59+vRpVq9ezb/+9a8LdwECgUAgEAgEgn7ByZMnWbhwITabjaeeeuqi2dFvIvrp6emsXLky7L9QfvGLX7Bp0ybWrFnD5s2blfUWi4Wf//znXabwlJaWcvvtt/Pggw8yaNCgC30ZAoFAIBAIBIJLnPT0dEaOHInP50OW5YtmR78R+l1RVFTE6dOnAUhKSmL27NkcO3YsbJvLLrssagrPnj17uO2223jggQdYsWLFF2WyQCAQCAQCgeASxuFw8Mtf/hKr1cqrr7560ezo90L/wIED/PrXvyYQCGCz2di8eTMTJ05UbffQQw+xefNm6urqAKiuruY73/kOv/nNb1i6dOkXbbZAIBAIBAKB4BLF6XTidruJjY1lyJAhF82Ofi/0b7zxRlJSUli+fDk33XQTX//615kwYYJqu44UHq/XC8CLL76I2+3ml7/8Jddccw3XXHONyNEXCAQCgUAgEPRIdnY2Tz31FIWFhWE1oF80knwxE4f6KW63m6KiIsaMGRNW+HspsWfPHiZNmnSxzRAIvrSId0Qg6B7xjgj6K3fddRcAzz333Hkf63zfk540Z7+P6AsEAoFAIBAIBP+JCKEvEAgEAoFAIBD0Q4TQFwgEAoFAIBAI+iFC6AsEAoFAIBAIBP0QIfQFAoFAIBAIBIJ+iBD6AoFAIBAIBAJBP0QIfYFAIBAIBAKBoB8ihL5AIBAIBAKBQNAPEUJfIBAIBAKBQCDohwihLxAIBBcIT0MlLTvex9NQ0SfHc1efxHZkC35HW58cT3BhkOUA9uN7cdedudimCASC/3B0F9sAgUBw6eJ32mhY/SyOE/swpOeRuvgujBmDLrZZYbTt+5SWLe8gywESpy0nYeqyL+S8DWtfoW3HKgCa1v6NuHFXkLbsXtV2shygaf1rWA+sQ2uKI3neLZgLpqu2a1z/Kq3b3gNAMsSQdfOjxGSPQPZ5cVUcRZeQhj4ps09s9ztt1H/wZxylu9GnDCD1qrsx5Y3qcT9fWyOtOz/A72jFUjiX2MHj+sSec0EO+GnZ9h6O43swpOaSNPdGdJYk5XPnqYM0rHkRX2sD5pEzSV18Jxp95/TxPmszkt6ANsbc/Xl8XuzH96BPzsSYPghPUw2VLzyA7HUBEJM7kgHfePzs7ZdlvI2VaM0JaE1xvd7Pb2/F21qPMWsIktT7WJ4sywQcbWhi45EkCej4Pj9En5pN/Pgrzvoa+gpX1XH8bY2YBo9FYzSd17EcJ/fjKj9KTE4+sUMn9JGFAsGXFyH0BQLBOdO49hXsxVsBcFcco+6d35Bzz58UoXCxcVUdp+GjZ5Tlxk9fxpA+ENOgwgt+7rYd74ctWw+sI3XpPSrxZd23ltbtKwEIOG3Uvvs78r77V3Rxyco2fnsrrdtXKcuyx0XL5rdIXvBNql/7X/y2JkAiafbXSJrztfO2vemzf+Ao2QmAt6GCuneeIu97zyJpu/7JkP1eql79H3wtdQDYDm0k8+afEnDaafjkeWSvG8uYuaQtuZuAx4m3uRZDWi6SRnve9kajeeMbtGx5Cwg+m+6aU+Tc8SQAAY+L2rd/TcDtCNp6cD2uimKMmUOIn3wVrVvfxXF8D2h1JM68juQ5N0Q9h6vmJFV/exj8XgBMQ8YT8LoUkQ/gKi+m7Jn7sBRMJ/Gyr6DRGXq03Wdtoubfj+OpOwOShKQzEJNbQOriuxRnzla8jba9H6OLTyVp9g3oE9Op//CvWPevBUDSGxnwzf+HPikT6/61eFtqMedPxzRwtOp87ppT1LzxBH5rI9r4FDK/+hP8tiZqXn8CkAFo3fYuuf/1dG9ufZ8Sek2a2HgGfONxDCnZyueyLOOuLEHSG3sMMjRveYfmDf9QlpPm3kTSZV+5IHYLBF8WhNAXCATnjKv8SNiyt6kav605TKT2hoDXTcvWd3GVFxOTPSIoiEKiq+eK88wh1TrHqYNfiNDvEEihNG/4F0lzbwwTt67y4vCNAj5clcewFMzoXOV1gRwI28zvstOy+a12kR88X/PmN4mbsBBdXBLng7vyWPi57C14W+owpAzoch/nmSOKyO+wp23PJziO7VDWWPetwWdtwFVWjOxxoo1LIfOGhy/IKFDb/k/Dlj01J/C1NaKLT8FTX6aI/A58TdX4mqqxH9sBfl9wpd9Hy6Y3MOdPi2pj/ft/UkQ+gPPkfjSxCartfI2VtGx5m4DLTurib/doe8vmt4IiH0CWkb1unCcPUPfub8m+/UmaNr1Jy8Z/K9vbi7eRdcv/KYIYQPa6qXv3t2gtSbjOFAXvya7VZFz/I8wF08LOV/NmUOQD+NsaqXnjCSSdgdBn2NtUjeP4XmKHTezR/g50jaep/2gXuvhU4ictOquRCQBPY2XYNQUcbbRue4+0Zd8JLrsdVL32v3hqTgBgLphO+nU/7DLQ0LpjVcTy+10Kfb/DiqN0F1pzAkhaWnesRPZ5iZ+yBMvImWd1HX2J32mlce3fcVcUY8zJJ+WKb6KNjb9o9gi+/AihLxBcRLwttbTt+oiAx0Xc+CuIyR5xsU06K4wDhuFrrlGWdQlpaC2J3e7TuutDbEWb0MWnkDTnBgxpeTR+8gLWA+sBcJ0pwtdaT/q195+3fbLPq17n9yLLMo6SXXjqzmAaMu7C3HedEXzusFUtW98BjZbkuTcCQQdHn5obvp+kwZg1NGyVPjED06BCnKc7HZf4CQtpiRAuyAH8jla0lkRatr6L7fAmdHEpJF9+M8asIb02PSZ3ZKfQBLSWZPRJGd3uE01suGvVOerO43uVf/utjdR/8Geyb3ui29GCc0H2utXrAn4ADKm5SAYTssep3rFD5IfgOlOkCH3rgfXYjmxFl5CKv7VRta0+LRf3mdaoNtmPbu9R6LsqS3FGONAduKtPEHA7lBSuDmSvm7Y9n6i297U24G2sDN2Str2fqIS+vy38OvzWRrQWtbPoU5zKnnGU7sGy659Y25ftR7eTfcevVSLcVVlC2941SIYYEqcuC0s/C7jsquP6Q9a17V+riPyOczhPHSR2SBcpY3KE8x3pPDutNH/+b1wVx/A0VoDy90Oiw+lxlR9F941kYnILur74C0j9h39VnGdvUzUBh5XMGx6+KLYILg2E0BcILhIBl52qvz2M394CgPXgBrK/9QTGzN4LsotNyoLbCDisOE8dQJ+aQ9qy79C68wNFiCTMuJbEaVcr27ftW0vjmpcAcFeV4qo4Rt53/ortyNaw49qKt/WJ0Je0etU6jc5I45qXaNv9EQDNG/9N2tXfI67w8l4f11V+FH1NMQH3SDTG2KjbWEbOwHZog2q9/dgOkufeSMv2lTRvfB3Z60GXkI7P2oTWZCF+yhL89jb0Celh+2V89ce07fkEb1M15hFTiR0+iebNb6qOr41LoW3Px0qKgre+jOqaE+R995lej5Ikz/s6fqcVR8ku9CnZpF51V48pNn6PS7UumlCLxFNzkvJnvseAWx5Dl5Cm+rx110e0tD9PiTOuJWHKEiDoxFkPbcDbVEXssEn42hrxNlQQO2wSMbkFSBqdakxFYwrm22uMJsz506J+P9GwHdlCwtRlWA+sp/6DP4ccUJ0Dn7L4Ttq2vIu9ZEfQ2QgRl7oenKXGdX9X0riiotGCPibqR8acfGyHPg8Tr4aMQbgrSwiNzEuG8P09YY5A53nixi+kZfMbnet0eixj53Vrfyht+9cSKuk9tadwVx0nJnu4ss7dkfrUbp/twHry7nteqYswDhiGLjEDX0utso951Czl336r2vHoGJmIhtacQMBlC1kjIcsBJZ2u7r3f4Tx5IMqecti/7aW7LprQD3WUARwn9l0UOwSXDkLoCwQXCcfxvYrIByDgw3bo8wsu9GW/j6b1r2I7sgVdfCopC2475x8tnSWJrJt/hhzwI2m0OM8U0bT2FeXzprWvYMwcquQFd+R9d+C3NeOqKkGXkIo3pDONLiH1nOyJRGuyqNYFPE7a9q4JW9e6fVWvhX7dyj9gK9qIBSgrWUf2N3+BPlmd0mKPuNYO9EmZeBoraVr3d2Wdr7WOpPm34K4spXnDP2ne8E+M2flk3fwzNO3CTGMwkTjj2vDri7EQFn+WJAKOVpo3vhG2XcDRhrvqeNT8bICA20nDmhdwHN+LIS2P1EV3kLHiBz3fjBBsRRtU62SfOqqOVqeKmvta6mjZ9p4q2m0r3krjmheV5Y5/GzOH0LJ9pfI8hdYvtGx9h/Rr7yfg9ahO7WttRJtuxttSh61oY9TrsIyZg61oE6Hizl1ZGrTnyJbwjQPhEWEAd1kx6dfcB4C9dDd1K/+A7HagNSeSsvD2qOeE9jqMiLoOFQE/sstGwrTltIQ4eVpzInFj5tC284Ow9yhh2tU4Tx3Aui/4vEv6GBJnrAg7pDXiXQCIzZ9G8twb0BhjsO5bhzYukdQl/4VGo8Vnb6Vu5e/xNlQSM3A06Vd/L2rRbyDKaIkmorC5edObhN5n2eumdftKki+/WbneyOO4zxwmrl3sG3MKIOyeScR0UzTud1rDlgNuOwGXHa0pjoDb2YXIVxPtfe9LrEUbad3xAZJGQ+LMFZjzO0dg9Gl5YaMYhrTcaIcQCBSE0BcILhKaWHW+quYLyLVs2b6K1p0fAEGhXfPmE+R977leR3t9rfVIOkMwd7Wdjmivq/yoantXebEiMHXJWeEfShr0iZmkXHk7tW//BtntQDKYSL3yjnO5NDVRBIgc8IMqhbd3xcOeujNhAjHgaKNl+yrSltyjPo9XHeHWJWaQPP8WPPVlqs8cx/fhLjusLLsrj2E9sF6JYEcjYeoy6lb+Xlm2FM6lZes7BJwR7TclDfpu8usb1/8d28ENQDBNpfatX5Nzzx/PrqhaXZKAJEnq1VFSYyBYgBp2OFmm4ePn1baGCP+uaN78tiotA0BjCD7jvtY61ecxeaNJW3oP+uQB2I9uC0v7knTBn0pdfM8OqD5kG/PwyQy873m8TdUY0nKijjB1EPC5VTZJJguyszMCrU/LxWdtwjL6MoyZQ7Ae2oA+NYfk2V/F21gVJvIhWGScecPDWMbMxtdSi2nIBHQRqXXRBHl8e+Q+cfo1JE6/BgiOrNS9+9vgs9v+HdqLNlLd1siAWx9THSNa2pwUMQISWScBQYenA29LHYGIVrKuyhLl357aU5FnxVV2pOvuU5H3V29U6gYkvQGtOTE8+NKOJsbSPjolE5s/jbjCOdGP3we4Ko5Sv/IPynLt278h59u/VQR92pK7qX37N/ha69DFp5K65L/O+Vw+azPa2Lg+T5sTfLkQ365AcJEwDR6LaegEnO1Dr/rkAcRPWHjex5X9Xhyle5H9XmKHT1Yiwh24IgpUA04bnppTPUb1A143de88FexGImlImLqMlAXfDNsmJidftV/ouqSZ1+EuL8ZdfQJJqydp3s3o4lPQxacw8L7n8NSVYUjLO+8Weh1ozep6AY0+hoRJixVnByQSZ16r2i4a/iipKJHpKX57K23710aN9mbd/mv0JjMBZ5SUlijRb183aQgAljGz0SWmtUficzGPnEnFs+qUp6Q54a0lI3GU7Apb9jZV4So/iqf2JPrETEzDJvTYqtGUNwpbSOEk0D5SEyU1JNq1RIgnb0OFSuT1Fr+tCQJRHAo56LjEZOejtSSHFDJD3Lj5PUZqE2ddj/NMUbAuRdJgGXs5tvbaEgCNJQnTkHHYjmzBdngzuoRUEmeswJg5uEebtaa4oGMamnqTmovOnIj95H50CSkQCFD5wgMAmEfOIP26B5AkDZ6ak2ECuQOp3Xk35Y2CiEi37PdRt/L32Iu3ha03ZAzGFJHjbjuypUsHy1URLCaXZRnboc9xVRwjJmcEmhh1Slvt209hGTWThOlXI2m0xE9arBQKdxA3fr7yb31iBlpLEn5bs7Iu9O+UJkoaU2hqUsDnwXl8L5JOH+yI5Ap3LGS/v3M/jZbUq+6ibtWfkD1ONKY4TEMnYMwYRNz4BUHHJeBHF58S9T70FY7jEak4cgDHyf2K0DdmDSX3O3/Gb21Ca0k6p65V3pY6at96Ek/tKbTmRNKW3UvssEl9Yf4lyQcffMCqVat63jCEkpISRoy4NGrqhNAXCC4SkqQh68b/wVVeTMDtxDS4sNuIX2+QfV4qX3lEGdrVJaaT/a1fhRVKGjKH4Dx1sNMOnQF9ak6Px7buXxcU+QBygNYdqzCPnBFWyGoaVEjyvFtCcqqvCetwozUnkH37k3ibqtCY4sNSazQGU1RH4XzwRyke9DvbSFt6L6bB4/DUnSZmYCF+axNte9dgzp8WNlIRSUxOPkhSWN51aDpCwOOk8qUH8bU1RD9AezFoNLu0cSlIhsrOAlGNFsvIWart1DYVEJPTKX5i8kbhbapSlnWJ6STOWhFtVwCaPnstTEh1nLv6H48q9loKLyf96u91a4d5xFQaLUkEOo6l0ZF42deof+93Xe4TO3wKaDRYxswO6zIEoI1LRtIZkH3qFJye6IzAhiO3R6IlnZ6srz9K8+Y38VubsYyZjSF9IJV/exhvQzlIWqAzIm0ZMxcAfWI6uff8EXf1CXRxKTjKisKEfsDZhvXA+rCWrs4T+8m5+/c9CjJvc616FCLgJ235d/G++rOwdA0Idtqxl+wKOpTt+2njU/G3P3uSPobE6dfga2ukefOb+JprMBdMJ27iIiRJwla8VSXyg/fOrBoJc5Tu7tLujr9ZTZ+9ptTmWPetwTxqJrKkRZI7xbSn9iRNtSeR5QBJs67HMnIGvvm30LL1PZAk9CnZNH7yIqbBY0mc/TU0OgMZ1/+Iho+fx9tYSeyIKZ1pPQSds7Z9a5RuT8asoZiHTwGCXXMq//aQ0izAmD0C05BxSmAFUPXRN+dPY+B/j8XbVNPjCMyFwpCep14XkZ4jSZpejS51RdO6V5TREL+9hfoP/tzePveLv95LlREjRrB48eKLbUavEEJfILjIxOSO7LNj2Ut2hgkCX0sd1oOfKcPvAEmzrsfbUImjdDdacwIpi+6ImsseiTdK0Z63sVLVsSZx5goSZ3YtLKFvc1wDbgcNq5/DfnwPhtQcUhd9W+kwI0fJJ5E9LiRJInbYRExDJ1Dzz/9Tutk0ffYaA257IqxPdyj24m2qzh2K8wM4SnZ3LfIBvTnocBmjODTuyhKMmUOQDDFodAbiJy06q045HcRNWYKtaGO7QJYwj5rdZTRelgO07lqt/iDgD1u0HdpA8twboxbLdqAxmsi4/kEaP3megNtJwtSlPSZExU9dSmwXrU61MWZSFnyTxk//huz3orEkEbC1oMoRkjRoYuMJhKRcxE9eHFYr0oHP3owhNfjsGVJzyLj2+0DwPpT/+V58rfVRbZFCet/7WutxlR8l4HEGuyiF4vfToowUBfE2VeGuPtFjZydDao4qdcQ0aCy2w5tUIr/zfOGjFv62Bsyj54DsJ37yEgxZQ6l4/vt468sBcJ4+hBwIkDBlCd6m6qiHdJ0pomH1s2HpaPou3geApPY5BiLrXhzH99I2+y4G+uuV+Qw6sB/dTtKs6wFInLGCxBkrqHz5IdwVwdQ/d/UJfLYW0pd/l5icfHLu/E3Uc2vNCeR8+3c4Sncj6fTEDpukpKFYD6wL6wjmriwhdel/oYkx464oISa3gJSF31IdU2Mw9WoE5kJhLphO3LgrsB78DCQN8ZMWETtkfJ+ew1MXnjrot7fit7dd8NGKLyvLli1j2bIvZiLFi4EQ+gJBPyIQpfOJ7AlPCdEYY8n82kMEvG4kra7XQ7+xI6bStudjZVnSGTBdxJlPO2ha/xq2w5uA4I957dtPkvudvyBJGnxNtartQ4WUq7w4rGVlwGWnbddHXbdAjBLxCjhtyLKM1D6xUVfokjvzhqNFUv22Zvy2ZjSx8WfVISeShlV/ComCy7Rue5ekWdepUriCSEhaLbIqnbqznWAHgSg1B2Gf+zzBSajaRysaP3mBlKvu7t7WD/9C1k0/7dLxi5+0GPOoWbirTlC78vcqmwCQAwTsLRgyBqNLTCd22CTix18RVeh31W/c19rQpcgHaNv9EZbCuSAHqH7t0W5HGSSVeyP12HIWoHX3R2HpN9q4FOInX4Xt4Ppu9lJjPxysIbEf3UHK4jsVka98XryVhClLMA+fQsvmt6LWMtiKNoUJ/YQpS3GVHcF56gCSVo957OXozIlYxsxR5lbQGEz4Q3LuJb0JQ3Ux9pZTqpSkyBz6gNuBu6o03M4jm2H5d3FVltK49mV8zbWYC6aTsuA2JF3ne6gxxGAZfZnqGgJude2BJGkU5+7LiqTRkrbsXpKvuBUkTY8zM58LpqETwkb99Gl5/7Ei/z+B3s+PLRAIvvSY86eF9b+WDCZV7nMHGr3xrPI7Y4eMI+3q+zDm5GMaMp7Mm3561hNjXQgcEe3mfK31imjTRxb/Rqyrfee3qs+jFRF2YCmYhhSRFyz7PEqecezwiRizhkXdN+CwUv/hXwm4Hfjaus69Dzja2lsinhuehnBhhxzAVRX9eJIkET9VHcmSokxs1LZrNXIUUdiB88QBReR3YI/sUhOBr6WOhjUvd7uN1hSHp/YkckTHFNWxWuvJ/MqDxI+/osttvK11UdfLsj/q+lA8NSdp3flBj6lEyQu/FVYbEj9liapVaiR+p43mz/5JqCPjtzZS86/Hgs67LsTp02i7HVlRCPiwF20MdjkKoWNfY9YQMr7yYNRdJW343wWN0UTWzT8j73vPMfD+F0lfcg/Jc29URL7s9xKTFzoyKWEaOApT6Qa89WfCRL4uKTMs/QZADkQZeQsEkH1eat98AnfFMfz2lmDb2IjRga6wFM4Ne1e15kRiR0zt1b5fBrSmuAsi8iHYPjd+8hJ0I7iwkAAAIABJREFUSZnEjphC5ld/fEHOE4ksy7TtXUPNG0/QuP5V/E5bzzsJzhsR0RcI+hFak4Xs25/Eum8tst9L3Lh5XXegOAfiCucSVzi3z47XF8gBtUiT2wthLSNn0Lzhn/gd7ZFSrY6EKUuBYMQzYG9W7Rs/8cpuz2cZMxfrvvDJiToisZJWz4BvPo6jdA9Nn/0jLGoWcNmx7l+LHAgQM7DrFoBImvP7zqIUAWvMXUfroqX1mAaOwnF0e9i6tr2fYEgfSPykRVGP422JMnoSpc95JKETc3VtZM8xKZXTGqWNpzEzfCIyZVOjRRV1jjx/zMDRYaM/XeGpPo7f0emUuM4cUtrPdkXAaUX2qx1MT90ZPHVn0KfmEJM3GkmjIX7SYpynDijzUXSPRPK8W2ha/xoEfOgS0kma/TXlU/OIKeiSB+ALeU4B0ETP1e4q6lv/0TPYD29WluMmXRm1CDvtmvuxjJ6leua0JjPa+DT8bZ2jKjE5BXjqy1VFxs7TRdCLP0GGlAFkf+uXWA+sQ9LqiZ94Za9SFP8T0OiNpC66A+ij7ma9pHX7SprWvxpcKN2Nu7I0ascmQd8iIvoCQT9DF5dM0pyvkTzv6xe83/OXAX2KOmrfIao0MWYG3P5LEmZcS/ykxWTf9gSG9IEAeKN0tNGYEzAOiB6R7yBu3OVhwlMTG0/ssImd59bqMRdMx++KHq1yntxPzIDhUT+TtHqS59/Su4htV0QRlLIz+kytQNR2guZhkzFG6UfuONV1n/FoM+fGDByNNqJoUBMxWhBZEBmNuLHz0IaMHulTBqBLDD9fwvSrw5aT59wUbkveaPRdpNBoY+PC6kokQ0ww4pmYgT4tl/Rr78eQkh10EntoRWg99DmEFqDWleEqiz7TrXI9yVkYsqI7IRDsQBRXOIfURXdiSM3BMnpO2H2VDCbiJy3GFJbLLZEwZSmJ05aT971nGfCtX5F779OqUS6tWZ3OFHD3PtIq+7zt8w504ji2S1XnImn1xA4Z12W9SNYNP8GQORiQiBk8lvRr70efkoVkCO/AdTbzjBjScklZcBvJ875+fu+UoE+wFX0etuwqO9zt6KagbxARfYFAcEmTOGMFNeVHlQJS88gZ6BM7UyX0CemkzL9VtV/8+AU0rXslLOqbMLnrnvUdxGSPIOvr/8vp9W+QlpVDwtRlUWfHjR02OWp+tSE9D0NqDskLbmufGdeNedRMki77KlpL0nkP1xsHDMNdXty5QqPttrjQMmYObXs+6ezcYk7AXDAN0+CxlD19T1iU25g+qMvjxA4djyE9Tyn008SYg33YZ11H67aV+O2tWMZejiE9j8ZP/4an9jSmweNIWfCNHq8pWHT522Btg0aLZeRMZL+Plm3v4mmoIG70bCxjZoftkzjzWkxDx2M7tIGYvNGYR0zp9hzJl9+MuWAG3qYqTIMKg/n8i8IjnjG5BeTc+RT2o9vRxafgqixRTThlSM3FF1G4HpnuFY2sGx6hYc1LOI7vQfa4iKxHkELSd7SxcUE7jmwBScI8ahbaGDOy34ft8ObgbMHDpyiz0Oosiar++R3EDh6PO2L+C/PZpLhoNGiMsQRCUqu0sRYSZ3+VpuMH0bVVI+ljSFnwzS5rJAAM6QPJuUNddJt+9X00fPw8flszpqHjleJfwaWH1pICIYXAkj4mahtWQd8iybIcpbpJcD643W6KiooYM2YMRuO5FdRdbPbs2cOkSf+5fXUFlxaeujLsJTvRJ2dhLpje69oDd30ZDaufxW9v/f/s3XdgVFX+Pv7nTp9J7z20EDqhd6QjCIiCiuyqa1/bqvux/Fzr17bquq67tlVXXcW2KiIgxUaXTiCEhJZOeq/T2++PIZdMZpJMkgkJ4/P6K/fOzL0nMynPPfd9zkHguIUInnxlxy86r6PfEZvJgJqdX0CXdQTWpjrYLUbIIxIRtfKRC7XNFrNjRhk3FwpdZa4rR+nnz8JSVw5BpkD44jsRcH4BpLboz2WiMW0bJAo1giYvE0uHGo79jJrtn8Jm0EEzeAIir3oAEkXbaxzYTHo0Zf4Km1EP/+HTfX6An91uQ/naV6HLOuyYTnTqVfAfeRlKPnlcDL6apPGIXvV4p45rqjyHkk+eEBeU0gyZjOg26um7y24xo3Lre2jK2ANBIoE6aRwil93X7ufcWsOxX1C19T3HRaFUhqiVj8Bv8ASkpqZi9KAESDUBnTqeSxvtjnr9rg5Qp77BWJqL0i+fc/xuCBKELbil3QUBfyu6m7c6ypwM+j2AQZ/I93Xmd8Rus8Kqa2h30SpvstvtsNSWQuof0q2ABTgGWtrMph4bGOgL7FYLIAjiBaZV3whdViqkfkFQt1Ou0h6rth667FRIzy/C1ZVjXEzmunIYy3Khih8q/pzz/wi1ZjMbYSw+C3lorM93BHiqp4M+S3eIiHqYIJFetJAPOGbT8db4DEEqh5QL6bRLaFW3L1UHIGD07G4dU+oXhICUuR0/sY+QB0dBHuw6ToOoJYlc6bSIIvW8vt1FQEREREREXcKgT0RERETkgxj0iYiIiIh8EIM+EREREZEPYtAnIiIiIvJBDPpERERERD6IQZ+IiIiIyAcx6BMRERER+SAGfSIiIiIiH8SgT0RERETkgxj0iYiIiIh8EIM+EREREZEPYtAnIiIiIvJBDPpERERERD6IQZ+IiIiIyAcx6BMRERER+SAGfSIiIiIiH8SgT0RERETkgxj0iYiIiIh8EIM+EREREZEPYtAnIiIiIvJBDPpERERERD6IQZ+IiIiIyAcx6BMRERER+SAGfSIiIiIiH8SgT0RERETkgxj0iYiIiIh8EIM+EREREZEPYtAnIiIiIvJBDPpERERERD6IQZ+IiIiIyAcx6BMRERER+SAGfSIiIiIiH8SgT0RERETkgxj0iYiIiIh8EIM+EREREZEPYtAnIiIiIvJBDPpERERERD6IQZ+IiIiIyAcx6BMRERER+SAGfSIiIiIiH8SgT0RERETkgxj0iYiIiIh8EIM+EREREZEPYtAnIiIiIvJBDPpERERERD6IQZ+IiIiIyAcx6BMRERER+SAGfSIiIiIiH8SgT0RERETkgxj0iYiIiIh8EIM+EREREZEPYtAnIiIiIvJBst5uABHRb5XNZofFaoNCLu3tplzSSqu0OJBRisgQDaaMjIZUemn0YVltdlTU6BARooasE2222ew4mFmG4somTBwehX7RgV5rU32TEet2ZKO4sglTRkYjPjIAheWNSEmOQGSIxmvn8Qaj2YqfDxagtEqLKaNiMGpQeG83iajPYdAnIuoFe4+X4P316ahrNGLyyBj8efU4qJX8k9xZp/Jq8Ng7v8JmswMABsQG4o2H5vRyqzqWVViLv358GFV1eoQGKvHojRMxYmCYR6/911fHsP1IIQDgs62n8OStkzFhWJRX2vX8hwdx5lwtAOBgZpm4XyaV4JnbJ2NMcqRXzuMNL350EMfOVgIANu7JxWM3TcT0lNhebhVR38L/KkSXMKPZCplEuGR6MMmhUWfCP75IhcliAwDsP1GK+MizuOmK4b3cMlcWqw1rtpzC3vQSRIaoMWFYFJRyKaaOikFYkBoAUNdoxDvfHkdGThUGJ4Tg7pWjER3m57U27Ekrxq/HixEZosGKOUkICVCJj/13U4YY8gEgr6QB+9JLMG103wp81fV6NOrM6B/j6H3/97fpqKrTAwBqGox465s0/Pv/m9fm6+ubjHh77XEcz6qEzmAR91ttdny3M9sp6B89XYGMXMdnMXVUjMdtLKlqEkN+axarDV/9crbPBP2SqiYx5Dfbsi+v3aC/eW8evtuZDYkg4Jp5g7Fwcr+ebiZRr2PQJ7oEmS1WvPFVGnanFcNPJcMflgzH5VP6e/UceSX1aNCaMDopHIIgOD12IrsKe9KKERaswpJpA+CvUXj13N1V22hAWZUOSQlBkMu6VxZjMFlQWqVFQlRAp8or2nOurFEM+c2yC+vEr602O1JPl6O6To9JI6LFQN0bvt2ehe92ZgMAKmp0yMipBgCs2XISr9w3EwNig/DOt8ex/0QpAODomQq89nkqXr3/Mq+cf0dqIf7xxVFx+9iZCrz58BzxZ7Jea3J5TVZhnVeCfnmNDlv25sFstWHh5H5iSG/PkVPlOJ5ViaT4YMwcEweJRMCaLSfx7fYs2OxAUnwQnr1zGooqGp1eV1zZBJvNDolEcHvcf69LF9/j1kxmq/j1+l3Z+HBjprh9zdzB+MMSzy4gAzQKyKQSWKw2t48bTFan7bJqLd5ZexxZhXUYlRSOe1amIDhA6dG5ukspl0IQALu9xT5F27/rJ3Kq8O66dHH7za/T0D8mEMmJIT3ZTKJex6BPdIn58UABPt6UiSa9GQDQqDPjnbXHMSY5ElGh3qmh/dPftyO/1BFEVAop3n1snhg2D58sw/MfHRT/we4/UYrXH5zlcjHQWzbvzcMHG07AYrUjNFCJ5+6chn4eBLSiikbsSStBsL8Cc8YnQKWU4dDJMvzj81RoDRaEBirx5K2TMTih68EgM7caG/fkoLxaBwFAi4yCUUkX6otfWXNYDHUffZ+Jl++dgUHxwV0+b3e07jVtpjda8f2eXNy/aizSs6ucHjtdUAuzxdrtiywA2HG+RKVZQVkjcorrkXT+/ZgzLgGf/3ja6TmTR0Z3eFyt3ozvdmbjXHkjJg6LwoJWvbsNWhMe+tcu1Dc5LiR+OliAf/55FuIjA9o85obdOfhgQ4a4fbqgBldMG4BvtmWJ+7KL6rF+VzYmDo/G7mPF4v7xQ6PaDPkAkHn+AsudshodrFYbpFIJ1u/KcXrs+19zccPiYZC2c+xmARoFblg0FGu2nITN7vr4FVOd36O/f56KMwWOOwDNP6+P3zypw/N4Q1iQGoum9sfWffkAHCH/2rnJbT7/RKuf0eZ9DPrk63i/n+gSklVYi7e+SRNDfjObHcgtrvfKObYdKRBDPuDoxXvho4Pi9o8HCpx60XKK6pHVoje6N+kMZvx3UyYsVkcDaxqM+HTrqQ5fl11Yh/v+vgNf/Hga73ybjtv/+jO0ejP++YUj5Dcf6+F/7cGrnx5xef89UVjeiCff3Yd96aXIKa53CvmLpvbH1bOTAAAFpQ1OPbcGkxVvfJ3W6fN5y4B2LpKae34Ht7oI6Rcd4JWQDwAhgSqnbUEAauoN4vaqBclYMDkRUqkApVyKm64YjqH9Qjs87kufHMJXv5zF/hOleOPrNPGuRbMDGaViyAcAo8mKnUeL2j3m93tynba37svHsx/sd3leebUO916Tgium9Ue/6AAsmJSIP68e1+6xkxLavtCrazTi7DnH72DrMj6ZREBnLsFXzh2M/zy+wO3di9pGo/i12WIVQ36z1hd8Pe2elSl4+d4ZeGDVGLz32DwMG9D25+7u/WvvPSXyFQz6RJeQzFz3vXoKmQTD+nccbjyxL921PKCkUit+7aeWuzzu72Zfb2jUmWFsVV5Qeb4Ouj2fbDkJq/VC9K5vMuHRN3ejUW9xep7NbsfutGL89/vM1ofo0P4TpW2WRKQMDhfLgswW1+fkFtcjr8Q7F3LN7HY79p8owZc/ncHpghoAjrEDG3fnYN2OLFTXO9636xcOwejzdxta9grLpBIsmT4AAHDPNSlITnSEpsToAPzf78Z3uj0/HsjH3a9swyNv7EZW4YUAuWp+MkJbhH27HXj+o4N45I3d0BstEAQB9183FuteXoav/7oE184b3OG5quv1OJ7lHEq3t7pz0NHPeZPejBM5VU4XfQq5879Uq82O8hrXn7/pKbHQqOS4e2UK3npkLu5fNRaBfu2Xv929YrT4O65ROd+MlwhAWLDjPVo137lX+5p5ye3eKXAnMlSDhCjXOxff7czGsx8cwIGMUshlUvSLdn5O6wu+i2HEwDDMn9Svw/K2ScOjcc3cwVDIpVAppFi9cAhSBkdcpFYS9R6W7hB5wdHTFdh5tBChgSosv2yQSy+kt7grG4kK1eDulaO9Vhs7Z1wCDmWWO+1LjLnwD33FnCQcPlmGRp0j4CyYlIjYCH+vnLu7okI1SE4MFns3AeCyMXEdvq6ovNFlX2F5U5vPP5FThWmDOnfLPyKk7SAS1+L9S0oIRkiA0qn3FHCE/QGxQZ06Z3te+zwVu86Xjnzx42ncvnwEvt+Th/IaHQDg2x3Z+Nf/zUZ4sBov3j0djToTBAB7jpegtsGAmWPixDAYE+6H1x6YBaPZCmUXpgr96UAB3vrmuLj90L924/NnFyHAT4nYCH/85/H5WL8rx+nuzOmCWmw7fA5LZwwEgE6F2bpGI6QSAdYW9SkhrX5/Jo+IxvABoTiZ57gIio/0x/yJiQAc5Wt/+/QIDCYrVAop/r+bJmLCsCismj8E//gi1W3ZC+AI6IF+CuQU12Pi8KhO3fWIDNXgb3+aidTT5U532ADg2vnJ4tSXCyf3w8C4IGTkVCM5MRjDB3g2k09rK2YPwp60Yqd9TXoLjpwqx5FT5Zg1Ng4Prh6Hf355FAVljRiSGIJ7r03x+PgWqw2fbT2FfSdKERvuh1uWjfDqNKHu/GHJcPx+0VAIcL3zcanatGkTNm7c2OnXLVq0CCtWrOiBFlFfw6BP1E2HT5bhuQ8v/OM9kFGKtx+Z2yP/SEYMDMONi4dh7fYs2Ox2XDlzoNdnapkxJg4bdufg9Pnb8n4qGZ68ZbL4eL/oQLz/+AIcO12B8GB1u7fLe0NYoBrAhaDvybiFhKgAVLUoBwEAqVQQS4BaG9yFW/4zUuKwM7UIR89UiPskgqNUonWAv+3KEfj75xcGoMqkglMNf3v0RgvKqtsfPKwzmMWQ3+zTLadgNF+4m9CgNWHbkXNYNX8IAEf9NgAsntq/zXN3JeQDwJc/OdfY2+3AFz+ewR9XjAYAKORSBPm79nhvP1IIjUqOWePiPapBB4CqOj2e+Pdep5CvVslww+JhTs+TSSV46Z4ZOJ5VCbPFhrFDIiGXOd7P/2zIEAemGkxWvL/+BCYMi8KscfEYEBuIEznVqK434JttZ52OqTNYoDNY8PUvZ6E3WnDnVaM8arPRbEV1vR4xYX5YvyvH6edSECDeWWmWFB8sjmFwx2C0YH9GKex2O6aOinU7rWtRi7t47uw6VoxRSeF465G5MJmtHq0F0aQz4aeDBahvMkFrMOPHAwUAHOsg5BTVIzJEDb3JgkVT+uPKywZ1eLyu8NaA+kvZ2bOOn0sG/d8GBn2ibtp22PmWf3GlFqfyazDSC4u3VNXpkXq6HNFhfuJt5uvmJ2PlXEd5gqfhprNevf8y5Jc2QKs3Y2j/UJfz+KvlmDm2457yi622wYADmc6lRxv35GJGB7361y8cgrSsSnHsgVQiYMXsJHzdYhBl8+DZlMHhuP3KkcjN6lz5jlwmwbN3TkV+aQMEOAYPqpUyBPm73omZNS4BtY0mbN2XB41KhtWXD/VosaIDGaV4/cuj0BksCAtS4albJ7sdxNuoc52pxmxxvaiRXKQB1u5Kmhp1JpRVa5FX0oDhA0IxaXg01MpM6I0XyqmyCuvw+pdHkZlbjT9dN8blGJm51TiYWYaYcD/Mn5gAuUyKPWnF4riLZivnJLkdlCmRCBg7JBIn86qxYXcOhvUPxYiBYahuVQ7WcjsxOhCJ0YF48t297X7P+0+UehT096WX4I2v06DVmxEX4Qe10rmkyG4HLG4+u7boDGbc/cp21DQ4Lmw/3nQSKoUMVfV6zBwTh3uuSYFSLnWZFcid9KwqXD6lv0ch32q14bG3f0VBmfvj1jUZUdfkuIv1nw0ZCA9WdzhzUlFFI46erkB8VADGJkf0mQkBLqalS5di6dKlnXrNnXfe2UOtob6IQZ+omwLd9DS6C2+ddTKvGk+9t1+cOu/yKf1w37WOMNNTAb8lT6YS7GsEQXCZzUYq7fi92pFa6DTA2GazY8mMgRgYG4RPfzgFQRCwYk4SZo9LEHt1u8rT9/WqWYNw1SzPezWtNjveWXtcnGO9ut6ADzdm4q/3TEdmbjV2HStCaKAKS6YPQGSIBhqVzGk+9v6xgTCZrSiqcJQshQYqxVIVbzueVYmPvs9ETYMBs8fFY+74BKxrNVtMWJAKd770C+x2R4/+U7dOwiv3zcD6XTnYf6LUKfD/cqgAxZWNiAr1w+qFQxAd5oe96SV4Zc1h8XNds+UkXr5nhtvxJKEBbZfabdydg/+0mEnnlqUjMGtcPH4+dE7cN2tcPI6cKsfPhwrgp5Jj3NBIlzEACrkEphZ3TGLD219nYNfRInz2wymUV+vEn+fiSq1LT/3E4VHtloW19tY3x8WQDzQPsHUE7O1HChERosYNi4Z5dGHZmcGsGTnVbYZ8d46eqWg36B855Shhar4zs2T6ANx1/g4QEV3Ae1hE3bRidpLTYMHLp/RzO5Cts9Zuz3KaH/ungwXiAjvkXnCAEnMnJIjbEomAlXM6Hpx55JTzmAQ7gF1HC5GRW43iSi2KKprwxldp2Lw31/0BuqFJZ8LfP0vFDc9sxVPv7kNxZdtjA9pjNFlc6vpLq5pw7EwFHn/nV2zdl4/PfziNx9/ZC7sdeOb2KeLPbf+YQDxxyyT848FZuGnJMIxOCsfY5EinQOgtWr0ZL/73IHKL61HXaMT6XTkIClBh6vkpMaVSAavmJ+OHFrM7mcxWfLb1NAbEBuHPq8chplVIttmBzNwabD9SiKfe2wer1YbNv+Y5Xbw16cx4/qODmJESi0HxF0qlBsYFtXt36pvtWU7ba7dnYVB8MAL9FFArpZiREosZY2Lx3IcHsC+9FD8fOuc03qBZcmKIOIg2PEiF264c2eY5S6u0+McXqShrEfKbVdXpnLZbXvB4IiOn/ZlxTuc7xiQcPV3h8thVswbBTy2HVCJg7oQELJ0xwOU5bXE3x33zuAg/lWufY3uzPQGOz6Fl+dXW/fmobzK2/YJ2bD9SiAdf34mH39iNgxnu1yogulSxR5+om6LD/PD+4/NxIrsKoYEqDIzzzoDJliEfcNyidzcjCzk716LX0NbWqMhWwoPUqKpzDrVxEf5Ys8W5dvz7X/Nw1ayk7jeyhf9syMCuY45pG9OyKvHyJ4fx5sNzOn0cjUqOlMHhTj3JU0fH4qeDBU6DQ/NLG3AqvwYjBobh46cXwmCyijXa9U1GfLcjRyzt2ZNWjH/+32yvXLg2yy6sg97o/LOdkVuFp2+bAq3eDJlMAqvV5lLfXq+9EOJWzU/GK58ecfv5llXrkFNcD7nctR+rvEYHvcmK1+6/TFwfYGxyRLvjaex253OYrVanhZf2ppfAZrM7XVRo9WYkRPmLA7rlMgluWToCiVEBKKvRISHSv91znsqvbnNAb+vylIycapTX6DxeQyM8WO1yQdjS0PMz+7grp5o/MRG3LB0Bq83W6elTh/YPxfihkUg9fwERoFHg5ftmQK2UwV+twPpd2fjql7Mwm62YkRKHhR0sANi6fXa7HTa75yVMzTJzq/H6lxfGw/z1k8N46+E5Xv2ZJ+pNHvfo//DDD3j99deh1+uxadOmnmwT0SVHKZdiwrAor4V8wHEruuX/9HFDI116MslZdlEdzraa0795QZ323HttitMgvYGxgU6DL5sp3YTH7krPcl6QKr+0AQ1uVnv1xCM3TMDiqf0xpF8IVs1Pxi1Lh7c7TaQgCE4DMQ9klDnV75ssNuzqYO74zuoXE+gyILJ5WkY/tRxKuRQalRxTW5VtzJ90oYxo2uhYvPPoXNy9cjQmj3BeHEsqERAerMY1cwejdYVbZIgawf5KSKUSTBgWhQnDojocNH91qwu71qUzdjugbzWlKwDcvTIF9183BjcsHoo3H56D5MQQqJQy9I8J7PCcgxNC0LrcPDxYhWvnDXb5GyOTStz2iLflxlaDjgHHgF6FTIK5ExJw7TzH9JzLZgx0KhEckxyBfjGBkEiELq+R8PRtU/DM7VPw4PVj8d5f5iE23B8hASrIZRJcOy8Znz+3GJ8/fwUeuXFChyVyy2cOcnqPZo2NR0g7JVhtaX03z2az49hZ17sZRJcqj/46vP/++9i7dy/Kyspw880346233kJBQQHuvffenm4f0W/W1FGxeOmeGTiQUXp+MGHP1Ev7Eo2b2UPczSjSWv+YIHz01AIcyChDsL8Ck4ZHQyqV4Lr5yfhk80kAjhlyrl8wxOttHhQfjKr6MnE7KlTT5XUJgvyVuOca5ykOr56dhP0nSsWLh9nj4ttcKThA43regA7md++s4AAlHrx+LD7YmIGGJiOmjY7F1XNc75L83+pxSE4IRm5xA8YOiXAqyQIcd1ziIvwxcVg0iiubUFTRBJlUwA2LhiE0UIXQQBX+/sBleO3zVBRXahEf6Y8HVo3t9JzyK+cOxqD4IJzKq8GQ/qHQGyzIaLVK7eKp/VDbYEB+aQMAYN7EBIwaFI5RXRyQnxAVgHtWpuCzH05Bb7Bg4ZR+uGP5KEgkAs6eq8Xp/BpxQPGqBcnw13j+GY0dEolX/zQDL685gup6A/zUctx7TQpmthqwnpIcgdf/PAt700sQFaLBrHHxXfpeWpJIBEwYFtXm40q51OOZm2aOjUNEiBqHTpYhISrAo2l03Wm9FoBj36U3PomoLYK99X1JN5YvX45vvvkG1113HdavX4+GhgasWrUKW7duvRhtvOQYjUZkZGRg5MiRUCq9M7f5xZaamorx4zu/6A1Rb3v9y6Pi4kd+KhleundGt+afP3uuFjlFdRg5KNzpdr63fkcqanX4+2epOJVfg9hwP/x59TixfMJbdAYzUk9XIDRQhRED255X3WK14en39uPE+TruftEBeOW+mW7vCnSXzWaHxWrzaMYWT46VV1KP0CCV215di9XmtWkVbTY7/rP+BH44kA+JIGD5rEG46YrhsNnsOFNQC3+N3KtlHzab3eXiRKs3IyOnCnGR/oiP7Pq5yqq1CAlUdXla1I5cCv9HrDY73vz6GHYcKYRMSZ3rAAAgAElEQVREImDpjIHtjp/wBc2z7rz//vu93BICuv970lHm9KhHXyaTQaG40GMQGBgImYzl/UTU9/x59TgsmJSIyjo9xg+N6nDF0Y4kJ4a4nXrRWyJDHAshGYwWKBXSHpkiUKOSu/TYuiOTSvDi3dOQkVMNs8WGlMHhPbawkEQiQCHxTsCUSAS304g28+bc6RKJgD+uGI0/LBkOQSKIIVkiEXpkTQl3dyD81HJMHhnT7WNHh7EUUCoR8OD143DblSMhEYQeuagl6k0epfWYmBjs3LkTgiDAZDLhww8/RFxc35tDm4gIgFfWMLjYVB6UGF0MguD54ly/ZX3l8yLvCOhE+RPRpcSjv1RPPfUUHn30UZw5cwYpKSkYM2YMXnvttZ5uGxERERERdZFHQT8qKgqffPIJ9Ho9rFYr/P39e7pdRERERETUDR4Ffa1Wi7fffhu//vorpFIp5s6diz/+8Y9OdftERERERNR3eDRC6cknn0R5eTn+8pe/4JFHHkFOTg5eeOGFnm4bERERERF1kUc9+idPnsSPP/4obk+ZMgVLlizpsUYREREREVH3eNSjHxkZiZqaGnFbp9MhJKTnppsjIiIiIqLu8ahHPzo6GitXrsSiRYsglUqxbds2hIeHi+U7Tz75ZI82koiIiIiIOsejoN+vXz/069dP3GbZDhERERFR3+ZR0A8KCsLVV1/NaTWJiIiIfgOsViu+//57LF68GEqlsrebQ13kUY3+mTNncPnll+OJJ57AiRMnerpNRERERNRL0tPTcf/992Pz5s3QarW93RzqBo+C/gsvvIAff/wRI0aMwLPPPouVK1di7dq1MBqNPd0+jxQVFWHu3Lku+4cMGSJ+XV5ejhkzZji9ZsiQIdi7d6/Ta+bOnYuioiJxu6mpCUuXLnXaR0REROSrhg4dCplMhvz8fDzzzDO93RzqBo+CPgD4+/tj8eLFWLp0Kerq6vDFF19g0aJF2L59e0+2zyt27dqFm266CZWVlU775XI5nnrqKTQ1Nbl93fHjx7F69Wrk5+dfhFYSERER9b7c3FwsWLAAiYmJeO2113q7OdQNHgX9/fv348EHH8SiRYuQm5uLt99+G+vWrcMnn3yCp59+uqfb2G1r167Fm2++6bI/MjIS06ZNwyuvvOL2dV9//TWeeeYZREZG9nQTiYiIiPqEyMhIDBs2DBaLBXa7vbebQ93g0WDcZ599Fr/73e/w/PPPIyAgQNyfmJiI6667rsca1xkVFRVYvny528fchfxmjz32GJYtW4a9e/di+vTpTo+9+OKLXm0jERERUV+n0+nw8ssvo7GxEZ9++iluv/323m4SdZFHQf/GG2/E73//e6d977//Pu68807cf//9PdKwzoqMjMSGDRuc9rWs0W+Lv78/nn/+eTz11FPYuHFjTzWPiIiI6JKg1+thNBqh0WgwcODA3m4OdUO7Qf/LL7+EwWDAxx9/DJPJJO43m8343//+hzvvvLPHG3gxzJgxo90SHiIiIqLfiri4OLz22mv46KOPXKod6NLSbtCXyWQ4e/YsDAYDzp49K+6XSqV47LHHerxxF1NzCU/rAbtEREREvyUajQYajQbLly+HTOZR8Qf1Ue1+etdeey2uvfZa/PLLL5g/f/7FalOvaC7hue2223q7KURERES9Jj09HQcOHMDBgwfx3nvv9XZzqBs8mnVn6tSpeO2117BixQqsWrUKb7/9tlMpT2+Lj493O83nmTNn2tx295oZM2bgzJkziI+Pd9q/fft2l31EREREvmj06NE4efIk8vPzcdNNN/V2c6gbPAr6zz33HMrKyvDII4/ggQceQFZWFl544YWebhsRERERXWRpaWkIDg7GwIED8cUXX/R2c6gbPCq8OnnyJL7//ntxe/LkyW1OZUlEREREl64xY8ZgzJgx2L9/P0wmExQKRW83ibrIox79oKAg1NXVids6nc5pPn0iIiIi8i1Tp05lyL/Etduj31yeI5PJsGLFCixcuBASiQTbt29HUlLSRWkgERERERF1XrtBPzg4GAAwYcIETJgwQdy/dOnSnm0VERERERF1S7tB/7777rtY7SAiIiIiIi/yaDDusmXL3O5vOUCXiIiIiIj6Do+C/lNPPSV+bTabsXnzZiQkJPRYo4iIiIiIqHs8CvqTJk1y2p42bRquv/563H333T3SKCIiIiIi6h6Pptdsrba2FhUVFd5uCxEREREReUmXavRLSkqwatWqHmkQEZG3mesrYDcaoIhM7O2mOLEZdbAZ9ZAFhvV2U4iIyAd5FPSfeOIJFBYWol+/fjh06BAEQcBNN93U020jIuq2yi3vofHYzwDsUCUMQ/SqJyBRqrt1TEPRGciqcmG3pkCQevRn1EXdvnWo3fMN7BYT1ANGI2rlo91uVzO7xQxt1hHAZoVm8ARIFKp2n2/VNaBy8zvQ56RBEdkP4VfcBWX0AK+0xRtMVUVoytgNicofASlzIVX792p7rPpGSOQqCDK5145pM2hRf3gzjGV5UMYORvC0qyAInbvp3pS5BzU7PodV34TAsfMROu8ml2PY7bZOH7c3GYpOo+rHD2GpLYNmyGSEL7oDErmy19pjaaiGVBPo1c+eqCd59B9q8+bNEAQBY8eOxf/+9z/MnDkTjz/+ON58882ebh8RUZcZCk+j8dhPLbZPoeHYzwiecmWXjme321H+zcvQZR1BAICivN2IvelFSDWBLs+1NNWiZtsaGMtyoe4/CqFzfg+JwhHkzTUlqNnxBQA7AECfl476w5sRMuOaLrWrJZvZiJKPH4epIh8AIAuJRtwtL0Oqbns18+qf/wvd2cMAAGNpNiq+ew3xd70JQRC63R7AceFhrimBLDQGElnnVtk0luej5OO/wG4xAQAa035B/B2vQZBe/KBlM+pQvu4f0Oceg0Tlh9B5f0DgmHkw15VDovKHVOUHALDbrNBlH4XNqIXf4ImQnN/fnrJvXoHhXCYAQHf2EBqObEHcba9C5h/iUdvM9RWo2PAGYLcBAOoPfg95WDwCx84HAFhNBpSueQKm8nxAKkfoZdcjeNpVXXgXAF3ecZgrC6EeOAaK8Hinx+x2G7SnD8JUkQ+pWQ1gfKePb7eaoc/PgCBXomLd32HV1gMAmtJ3QOoXhLC5N4rPbTj2CxqObIEgVyJkxjXQJHX+fB1pOPYz6g9+D0tDFexmIyTqAERccRf8hk7x+rmIvM2joJ+ZmYm1a9fi/fffx9VXX42HHnoIK1as6Om2ERF1i7mu3GWfxc0+T+nz06HLOnLh+NUlaDj6k9uAXvHd62JwM1cVwW42ImLpvQAAU1UxmkN+M1NVYZfb1ZL29AEx5AOApbYMTSd2IWhS2wsdGgpPOW2ba0ph1dZD5h/c7fbUHdiAmh2fAzYrBLkK0dc9BnX/UR6/vjFtmxjyAcBcXQxdThr8kid2u22dVbfvO+hzjwFw9MBXbXkXNdvWwGZoAgAo45IhUahhLMuFTd8IAKjxD0HcLa+0W55lri0Tf1aaWZtqUX9wI8Lm/cGjthmLs8SQ38xQdEoM+hXf/cMR8gHAakbNjk/hN2I65EERHh0fcFxM1O3fgMbUHxw7BAmirnnU6bOo/uEDNBz9EQAQCCDv6DdQhMchbMGtUCUM7fAclqY6lKx5ApbaMrePGwouvE+6nGOo2vJvcbvsm78h4a5/QR4S7fH31BHHOd512mfTN6Jy87+hHjS2V+8uEHnCo/t3drsdEokEe/fuxZQpjitYg8HQow0jIuouzcAxEBQty2EE+A2b2uXj2XQNLvusunrX55kMLsFNl31U/FqVOByCUuP0uN/gCfAGu9no2h43+1pSxg9x2paFREPqF9Sl89sMWlj1juBrqi5BzbY1gM16vm0GVGx8CwDQdGofCv99H/JfvwU1Oz6H3W53ezx3JRKC3PWugN1qhqHoDCxNdeI+i64BNbv+h/rUHzt8DzxhqihodVKbGPIBwFh8Fvq842LIBxyBvaE5GLdBovQDJK7/ji31VR63TRk7GGhVkqOKu/C5GkuyXV6jPbUPdpsV9vOfT1usugYUf/w4Ct+6+0LIBwC7DXX71ombNqMODWm/OL3WbjbAWJqDsrWvwNbigq0tDak/tBnyAUAZN1j8Wpeb5vygzQJ9/okOz9EZupxjbvfbDE2wNtV69VxEPcGjHv3ExETccccdKCoqwqRJk/DQQw9h6NCOr8yJiHqT1C8IsTc8h7r962Az6hE4diHU/UZ2+XiaQeMg9QsSSwkgkSJg5CyX5wlyJWRBEbDUV4r75OFxF9ql8kPM6qdQu/srWLX1CEiZC/8RM7vcrpb8hk49f1xH4JWo/BAw8rJ2XxO+4FbYDDrocx01+hFL7u5S2U71z/9F/ZEfALsNASlzIY+Id3mOtbEK5roKVKz/p3gBULdvHeRhsQgYPcfl+YHjL0dj+g7xIkuVMMzljkBT5h6nspWQuTdBFZ+M0k+fFvfV7vkaife8JZZPtWauK0fdvvWwauugjB0Ma1MtZAGhCBi3UCzJUQ9MgS47tdPvi6Wxpt3HJQoVFNEDYWoVxv2HT2/7mA1VaDy+HQAQkDIP8uBIRCy7DzU7PoPNoEXAmHkIGDNPfL4iIgGGAueLUlN1CfL/fiMgSBA89ao2S8dq934LY/EZt4/ZrS0vEgSgjZ8bm64B5opzUMYmtfk9NT+vNYlfEGy6RmiSJyL0suudvqfWFBHeHXDf1gB+eVgcZMFRXj0XUU/wKOi/9NJL+PnnnzF+/HjI5XJMmDABV13Vtdo+IqKLSRkzEFErHvbKsSQqP8T+4a+oP7wFlSVFGDB/ldvgIggCIpbcg4oN/4JVWwdZSDTCF97eql2DEDh2IazaOmiSJ7kco6ukmgDE3fo3R8mLzYqAlDmQdVCeIfULQsz1T7T5uM2oR2P6dlib6uA3fDqUUf1dnqPLTUP9oU3idmPaLwiattLleYqYQY7Q2KoXWV9w0m3QlwdHIeGuN6A9cwhSlT80g8c7DSa1WUxOIR8AarevgSwk2mmfTVuHxvRdCJqwyOUcdosZJZ8+DWuDowddd/aQ+FjTyb2Iu+1vEAQJAicshlXXiKbMPZAFhMFcWwZrY7W7t8z5e47q1+7jtXu+dg75Ujkilt7TZg24pbEWRR8+IobihtQfEH/H6wgYNQsBo1wvPAEg6uqHUPThw+fbK0CdNBZNLXrfa3d9CVXCULcXwuaq4jbbHjT5QkmYVVuHwDEL0JC61eV5gkINediFi13tmYOoO7ABABA8+Urxe/UfPRsNx34WPzupfygS7n4TgkwOQSJ1OmbAqNkwFGSiKfNXCBIpgiYvg6rV3anuunCOPQAESFR+UCUMQ9i8m7w2hoWoJ3kU9DUaDZYvXy5ur169uscaRETUl8lDohG+8FYUpKa2GyrUA0Yj8U/vwdJYDVlQhMtMJ2VfvQh97nEAgLDjM8T94UWv9UbKAsMQctl13TqG3W6HtbEGEr9AlH7+DIylOQCAugMbEXvjs1DFO9/VNVWeczmGoeg0AsYuQOOxXwDYIQuNRcz1T8Jm0AIQ0HKcgqpFSUZrUnUAAsfMgz7/BCrW/xOCXImgycugjOrvqDtvVZsOOHq8Xfa10bOuP3dSDPmtmcrzoM/PgGbAaAiCBKGzrkfoLEevsqGiACUfPHzh/FI5YDW3arysw7s1+rzjzjusZiijXGc9Mpblomb7pzCW5TmXB2nr0XRyL4ImXtHmOaR+geh3//uwaushKFSO8QYtyskAR3mPu6CvGTxBHJsAABKlBgFjF8AveRJUCUNh1daj7Ku/wliaDUGmQMC4hZAFhKHixD7IawshCwhF2KI7xFmljOX5KP/27+L7Vl78GuJueQXKmIFQxSUj9sbn0JC2HVK1HwInXtHmrFGCVIbI5Q8gbOGtECQyr81adbHPQdSTujYvHBGRl9kMWhgrCqCM7OfRLCWXArvZCHNNGQSZ0mlgq6H4rBjyAcBu1KH+0GZELOkbq42bKs+h/NtXYa4ugUQd4BQqYbOgIfVHl6CvGTAGNfjE+ThluYi78TmEzv497Dar+B5INYEIX3IXand+AZtB51Jm4o6hJBulXzwnhkPt2UNIvPstyENj3T5fovSDrdX4Cb/B7mdkkQWEtnvu+v3roRkw2mW/KrIfEv/0PrRnDkDqFwS/5IkwVZei6fg2NJ0+AKk6AKGzV3c4c44isr94IeVouway4Ein59itZpT970WxJKs1Y0lWu+do1jz2Qt1vBOp+/cbpMUtTLWwWk8vMSIHjL4fdbEBT5q+Oi8hZq2G3WlC1+R2YKgshCwwTy9TsFhMa07Yh8b73kKsegHFjRgMSmVPvtz7nqPPFmd2GxhM7Uf3zRzCW5kDVbwQiltwLWYBnMw61N6OUt1yMc3TFpk2bsHHjxk695uzZs0hOTu6hFlFfw6BPRL1Ol52K8nX/gN1sgKBQIWrFw9AMGtvbzeoWff4JlH3zMuwmAyCRIWLZvWKtvN1qcXm+3ea6r7dU/fghzNUlAOAc8s8T3EyRqYhMhKBQw27SX3je+TUGpBrXkBQ4Zj4Cx8z3eF537am9TuHQbtRBm3UEgWPmIWj6Najfu1Z8TJM8CZApoTu5R9wnDYyAMs59uFFEJCBw4hVoOLzF7eP6/BOw26wupSMAIAsIQdCExeK2MjIRygW3IGzBLR1+T83UgyegMX07cH5Asqr/KJfZXIzlBW2GfMDR298Z6v6jELbwVtTs/gp2gxYA0HBoE6wNVYha+YjTcwVBQPDUqxA81VGya7fbUPj2vbDUVwCA01gUAIDNCkudY0Ctu2lQ5WGuYzd0WYdhqXMcT59zDFU/vIfoax+DpakOld+/BX3ecSgiEhGx9B4oYwZ16nttyW63wWbQQqLUOKZA1Ttq/zsK8qaqItgtJiijB3b53H1FcnIyFi1yLWEj38SgT0S9ruqnj2A3O2byspsMqP7pI2juvrTX6aje9qkj5AOAzYKaXz6G//DpECRSqBKGQhkz6EIvrlSGwHF95x+vuzKcZhKlBkGTlrh9LOSyVaj55WNxO3jaChjL8x2Dg5tq4T/yMpfyEk8Xb5IFuE5P2TxlZdjs1QgcuwDa7CPQndwn1thL1P6QaIKgjEhA6Jwb2q2pDl94G4ImLIZV24CKLe/C0mK6U1lAqNuQ7y2NR38QQz4A6LKOwKprgFQTCJtRh5qdX0BfkOmYVcdNmRIAlxl3PBE0cQka036B6XzQBwDt6YPiudtibawVQ747Uv8QKGOSgIp0t49rkicgYMx8NKZtA2CHIJOLIb+ZofA0AKD6l/+KZUOminyUf/cPJNz9Vpfq4w2Fp1Gx4V+w1FdAkKvEvzkSTSDibnkZcjeDa+12Oyo2/BPazF8BAMr4oYhZ/WSbg7ovtqVLl2Lp0ranziVi0CeiXmdtcB7Q6K6+upmh+Czq9q+H3WJC0ITFHS6QY64pRdVPH8JUXgD1wBSEL7wVklZTW/YEa5NzPbhV1wi71QJBIoUgSBBzw7NoTN8Jq7YO/sOne322kO7QJI1DU/pOcVuVMAxBU5Y7Bg4Pntjm/PrBk5dBFZsEQ+FpKOOHQBkzCOfeukscNGosyYJEqXY76LYjASnz0JT5q1ii4jd0KtQDUsTH5UHhUIbFoabFtKY2fROCJi5FyMxr2z22qfIcdLnHoQiPh3rgGEQsuh3l37wCm1EHQa5C2OW3w261QJdzDHaLCZqk8R2uNtwZNqO+1Q6ruHZA5db3xJAJAJBIXQYyA2h3nv7W7FYzzDVlkIdEO6b2bEGQyTtc9VUaEAJZSLTTNJiKmEEQIEAaGIbQWde3ewxBkCBiyd2Qh8aiZvsa2C1ml+eozt99MRY5z/ZjqS1zWePBfv79ai982+12VGx8Q7xAaQ75gGOmn4YjWxE2/2aX1+nz0p3ef2PRaTSkbUdwGxe7RH0Ngz4R9Tq/ETPQlL5D3PYfMcPt8ywN1Sj97BkxBOlzjyP2D39tdyBn+bevivOfN6XvcMyIc37hqp7kP/Iy1J+fVQQA/IZOdirHkCjUTiUffUn4wtshyBQw5J+AInogwubf3GEdezNVwjCoEoYBcKyg2nq6RO2ZQ10K+hKlGrE3vwRjSTYkcqXbaQ+tHq5z0FLTqQOoWPequK1OGo+o5Q/Ab+hU6AsyoEoYBmVsEko+fVqcYlIWFIHYm1/2yoJiABA4dgEqi89eaMOgcZAFhgMAdGePOD/ZbkPUqidQ/tWLTrttRt35FZc/h7m2HH5DJiN4+gqXOxGGotMoX/sqrNo6SDSBCJ56NYylOeLvVPD0lR32VguCBFErHkLVlvdgqjwH9aCxiLjibrclWu0xlrofV6DqNxLhi+8EcH6NAVwoDZIFRTqt8dCYsRvVP//3/PSbExC5/AFIFGrY7XbU7PgMDUd+gCBXIHjq1e0ulmc3u5/j39Dic2mmzz3GoE+XDAZ9Iup14YvvhDwoEobi01DFD0XQVPfT9+qyU51WSYXdBu2ZA20Gfau23mWRI32e+3ICbwud83vIAkKhz8+AMmYggqYs7/hFfYREqUbE4j92+zjykBiXchN5mPvBs54QBKHdizrNoLGQ+gVfqGUXJPB3s85BS9XbPnba1menovy718Vykaa6cpiqimAqvTD9paW+Eo1pv7Q573xnBaTMhdQvGNqsw1CExSFg7ALxMXlYLEwt6u/lobFuy2akQREo/fIFMcyayvMgSKUInua8in3VDx+I749N14DGtF+QcO87MBRkQB6e4HbqVHeU0QMRd+srnf1Wndvs73rxqIxJQuwNzwJwzJLUuoxMEdVfLNuxautRuelt4PyYF93Zw6jb9x1CZ/8O2lP7UL9/PQBH733Ntk8g0QS5DNAGHOMIAs6vINyauzs3goyr4dKlg0GfiHqdRKbwaDpId/PBy4Mi3Tzz/HE1AZAFhjuVAimiXact7AmCRIqgSUsRNOm3Wz8rD45E6Jzfo3bX/2C3mqGMS0bwlJ5bg0Wi1CD25r+i/tAmxwJpY+a1e2EAOAb1ttZ6dVVTixlxxNeZvbs6vCZpHDRJ41z2hy+6w9ED31QDqV8wwq/4I7RnDrk8Tx4YAV3LEh8A2rNHXIK+uaak1XYpZP4hXluwrTNCpq1Aw5GtTheCLdelsDbVuoxJaPm+m6oKxZDfzFiWBwAwFLku8CUPjYGxVdBXDx6PsDk3ul18CwA0SeNRs+0TpzEU/sPcr29A1Bd1fvQOEVEvUQ8cA/8WCwKpB4yG/+jZbT5fECSIWH4/ZOcvBpSxgxG28Naebia1EDz1KiQ+8AES7nkbcTe/1Onyjs6SB0chfOFtiFx2n1hC1B7NoFbhWhBc7jrIQmMgbVG6JMiV8B812xvN7ZAqLhmJf3oX8Xe9icQ/vQt14gjXC4Lzc/W3ng2p5QJVzTTJE9vdvpik/sGIvOpBCOdnvFH1G4mQy1aJjyuiB7hMn+rXYrVgZcwgl/E26gGOVZNVia0/ewFBk5cB0gv9mxJNICKvfKDNkA8AirBYRCy7D7KgSEjUAQiefg38hrsvLSTqiwS7vcVlKnmF0WhERkYGRo4cCaXy0rzFl5qaivHj2x/kSNRbzLVlsFstUIS7TtPnjt1ug82oh9SL8/Pzd8Q32K1mlH/7d+iyj0Ki8kP45bdDFhyF8m9fhbWxGlL/EESteAiy4Gg0HvsZNrMBASlzPf7Z6ymNx7ej/sgPkMgVCJ6+EppBY9F4fDuqfvoQdpMBiqgBiL7uLy6DdG1GPWp2fQlD4Smo4pIRMvt3Xv29aMnT3xG71eL4/XRzEWhpqELt3m9hqauE//BpCEiZ6/S4/txJ1Gz/FJaGaviPnInQ2b8TxyXU7v4a9Ue2QJArETLzOgSOmQdDSTYaj/0MQaFC0ITFkIdEe+ebJeqi7v4v6ShzMuj3AAZ9It/H3xHfZrdZYamvhCwwXFwP4FJgMxlg1db1iQDL3xGijvV00L90/noRERFdJIJE2ifCcmdJFCpIFJdeu4moZ7BGn4iIiIjIBzHoExERERH5IAZ9IiIiIiIfxKBPREREROSDGPSJiIiIiHwQgz4RERERkQ9i0CciIiIi8kEM+kREREREPohBn4iIiIjIBzHoExERERH5IAZ9IiIiIiIfxKBPREREROSDGPSJiIiIiHwQgz4RERERkQ9i0CciIiIi8kEM+kREREREPohBn4iIiIjIBzHoExERERH5IAZ9IiIiIiIfxKBPREREROSDGPSJiIiIiHwQgz4RERERkQ9i0CciIiIi8kEM+kREREREPohBn4iIiIjIBzHoExERERH5IAZ9IiIiIiIfxKBPREREROSDGPSJiIiIiHwQgz4RERERkQ9i0CciIiIi8kEM+kREREREPohBn4iIiIjIBzHoExERERH5IAZ9IiIiIiIfxKBPREREROSDGPSJiIiIiHwQgz4RERERkQ9i0CciIiIi8kEM+kREREREPohBn4iIiIjIBzHoExERERH5IAZ9IiIiIiIfxKBPREREROSDGPSJiIiIiHwQgz4RERERkQ9i0CciIiIi8kEM+kREREREPohBn4iIiIjIBzHoExERERH5IAZ9IiIiIiIfxKBPREREROSDGPSJiPoYncGMkqqm3m7Gb45Wb4ZWb+7tZhAReY2stxtAREQXvLsuHZv35gEA/NRyvHb/ZYiL9O/lVvk2u92O9747gR/25wMALp/SD3etGA1BEHq1XURE3cUefSKiPqK8WieGfMDRw/zymsO92KK+r7bBgK378rAvvQRWqw2NOhPeXZeOh/+1G59sPgmj2drhMQ5llmHz3jxYbXZYbXZs2ZePg5ll7b6mokbn0bHdMVus+OyHU3j0zT14f/0JNOlMXToOEVFH2KNP1EedKajB/hOliArzw9wJCVDKpb3dpB5VXNmE99alI7+0AeOGRuLOq0ZBo5L3drMuqozcKpd9FbW6XmjJpeFcWQMefXMPtAYLAMHJEncAACAASURBVGBMcgSkEgGppysAAGfO1UKrN+Oea1LaPU5BWaObfQ2YMjLGZX9lrR4vfHQQuSX18FPLcc/K0bhsbHyn2v3hxkzxgu5Ufg1Kq7R45vYpnTrGb92RU+XYl16C6DA/LJk+AH7q39bfCiJPsUefqA86fLIMj765B9/uyMY7a4/jxY8OXtTzZ+RUYU9aMXSGi1ev/NLHh3DsbCVqG43YdrgQH27MvGjnbo/ZYsW327Pw8prDYq9vT5k6yjVYxkewbKct3/+aJ4Z8AEg7WymG/Gb7TpR0eJxxQyLRskpHEBz73Fmz5SRyS+oBOO64vL32OAxGi9vntmVvunObUk+Xw2Dq3DF+y/YcK8azHxzAz4fO4dOtp/DsBwd6u0lEfRZ79In6oE1789AyTx47W4nC8kYkRAX0+LlfWXMYvx53BJEgfwX+9qeZiA3v2bBZ22hw6VU9nlXZ4eu0ejM2781DWbUW01NiMX5olNfb9ubXadiRWgQA2Hu8BJW1Oty8dITXzwMANjsglQhOFxMBfopuHzevpB5vrz2Oc2UNGDc0Cvddk4Kc4nqcPVeLkQPDMWxAaLfP0RssFpvLviB/BeqbLpTCxIT5dXicpIRgPPL7CVi3MwsAcPXsJAxOCHH73MIK559TncGCjzefxNRRMUgZHOFRu6NDNahrNIrboYEqKGS+fcfOm348mO+0fSq/BufKGpAYHdg7DSLqwxj0ifoghcz1ZpviIpTu5BTViSEfAOqbTFi/M6fD0ofuCvRTIjRQhZoGg7hvQGzH/7T/33/243RBLQDg50Pn8OiNEzBzTJzX2mW22LD7WLHTvm1HCnss6Gv1Zpc7BvXa7tVv22x2vPTxYZRWawE4LlbKqrTIKa4Xn3PXitFYMn1At87THXWNRmw7fA5Wmx1zJyQgPFjt0esWT+uPXceKYD4f+JPig/C7y4fitc9ToTVYEBKgxB1XjfLoWDPHxmHm2I5/diYOi0ZOUb3Tvs1787B5bx5uu3IErpqV1OEx7rhqFJ7/8CDqmoxQK2W4e8VoSCQc+Ospf7Xzxa9EAEt3iNrAoE/UB62cMxhHz1TCdH6w39wJCYgK1fT4eXUG1/IBd/u8TSoREOincAr60R30xBaUNYghv9lPBwq8GvSlEgEBGgXqmi70vgb7K712/NaiQjUYMTAMmbnV4r654xO6dcyaBoMY8ps1l540+2bb2V4L+k06Ex58fSeq6x2f/YbdOXjjodkIC+o47CcnhuCff56F3WnFCPFXYu7ERKiVMnz8zOUordIiPjIAcjcXzd2xakEybHY7DmaUorC80enO27od2R4F/eTEEHz01EIUlDUgLsIfaqX7f8VZhbX4YEMGyqp1mDY6BrcuGwE5e/5x3fxkpGVVilOhLp050KOfF6LfIgZ96jN+OlggDq66dt7g3/Qf7qH9Q/HeY/Nw+FQ5okI1GJvsWUlAdw0fGIaEqAAUljvKEyQCsHBKYo+ft7pej/zSBqd9qacrcNuVbb9Go5RDEAB7i6Dlp+l8r57NZseaLSfxy+FzCPRT4OYlIzBpRDQAQCIRcMuy4XjjqzRYbXYoZBL8YcnwTp+jM568dTK+25mNoopGTB4Rg7kTuhf0QwJViAhRo7JW3+ZzGrp51wBw3I2wWG0I6uSF0K/HS8SQ39yWHalFuGbuYI9enxgdiBsWOd/9USlkGBAb1OZriiubYLHY0C+m86UeMqkENy4ehiA/Bf6zIcPpsc5MxymXSZAUH9zm42aLDc9/eBC150t8Nv2aBz+1HDcsGtbpNvuagXFB+ODx+TieVYWoUA2SEtp+Hy+mTZs2YePGjZ1+3aJFi7BixYoeaBERgz71EVv35+OdtcfF7YycKrz58Jzf9DzW4cFqLJ7a/6KeUyoR8PK9M7B1Xx5qGgyYPS7hotRv+2sUUCtl0LcY1BgZ0v6FXkSIGstmDMTGPbkAHLfur5uX3Olz/3iwAN/uyAbgKFV66ZPD+OjJBQgJVAEA5k5IRMrgCOQU12NIYking2xn+avluHGx98KcVCLg0Rsn4I2v0lBU0YixyZHIyK2CyXyhvt1sscFgtEDVRs9yR9ZsOYnvdubAarNhZkocHlw9zqknvbpeD63e7LaGWiZ17XF3t88bbDY7/vbZEew9X56WMjgcT982pdNlcTlFdS4hH4DHFyeeOFfWIIb8ZulZVcAir53ikuavUWB6SmxvN6Pbzp49CwAM+tRjGPSpT9h9rMhpu6CsEfmlDe32ygHAiewqZOZVY0hiCMa2MUsGdU6gnwKrFgy5qOdUyqX4/aKh+GhjBmx2QKWQ4oZFQzt83R1XjcKc8Qkor9EhJTkC/l2o083MqXbatlhtOF1Qg6mjLoSIsCB1n73D9MP+fKzflQ1BEHDtvMGYO8H1DszQfqF459G5sNnskEgEPPb2r07lQeHB6i6PATmdX4NvtmWJ27vTijF6cDgun9IfAPDR95nYsCsbNjswpF8I/t8dU50+p+kpsfh2RxaKKhwrAUeGqDFnfOemq/TUkVPlYsgHgONZVdiRWoTLp/Tr1HHOFta57JuREotlMwd2u43NYs+X9LS8+B0U1/7fQ+pdS5cuxdKlSzv1mjvvvLOHWkPkwKBPfUJ4qxAlkwoICVC1+5r1u7KdpmC8YfFQrJp/cQMqec+hzDKx3tlgsmJPWgmS2pj5pKWkhOBu3bofnBiMXS0uNCUSAYPaKanoSzJzq/F2izth//zfMSRGBbb5fjQP+LzzqlF4/qODqKrTI0Ajx33XpnR5MKi7OejPnd+XV1KP73Zmi/vPFNTi+z25WL3wwu+pWinD6w/Owr4TpbBabZg2OrbHBla6W5OgK+sUDB8Q6lI21tm59DuikEnwp+tS8MGGDNQ0GDFuSCR+58HFLxFRSwz61CesXjgEmXnVqKzVQyIR8PtFwxAc0H6JxLod2U7b3+3IZtC/RDXpTEjPdl4s6tf0EtyyrGdmt2npimkDUFDagB2phfBXK3Dz0uGIDOn5gc/e0HoKUrsdSM+u7PDCZ2BcED54YgFKKpsQFarp1oxOY5IjIJMKsFgvpN7xwxzTnJa1GgTc1j6VUtbtsQiemDQ8Gv/ddFIc5C6RCJjmZu2CjvSLDsQDq8biy5/OwGS2YumMgW7XQGhmMlvx3a5snM6vxfABobhqVlK7g4S37s/Hms0noTP+/+3deVhUZfvA8e8wwz6yC4qguYFbmpq5AG5UKoqgoOIuZeZWb9ZLr5paLpWVaWUlWtnbZshPVFTK3CqXEJcUM9fMVEQBMRCHbZiZ3x+8TiICKsuMeH+uq+vinDnnee4ZOnLPOffzPEUEtPNkwqCHcbCv3pIxIUTtJIm+MAueddWsmPE4p89nUdfZ9o6n1yvhAa7nv9/ZWKtwsLcqMSi0Xg3MMgTFgyKfH9aeKeHFd7Xvp3EhTW5TynG7fbejtFBUyboMHi52zH6qC6u3nSS/UEd/v8bGxabaNisup7qe98/Ca6asq3Z3seONSd1Y9/MZior0DPBvfM9PbwI7NSSw050NVF8Wd4Rt+88DxeVD6X/nMaWMKWsvZ2qIjks2Pt36+dBFmnk73dFsPkIIcStJ9IXZUCkt7mrgZ3jv5iUGxIX1kj+E9yuV0oKJg9vy/upDFBTqcK5jXSN382+mrKYBoNWpc+t6hPZoSsKesygUCkJ7NOURn5ofq9KhhTsdWpTu197Wkjcm+7F62ylyNIU80bkRj7WqV+Px3cy3kQvTx9TsAmE//VpyDNLPv14oM9H/82I2ty6+fOu8/UIIcack0Rf3rYHdm9LUy4ljZzPxbeRM22Y1MwWlqB4BjzSgg687l65oaFTfocrnP6+NFAoFTw9sw+h+LVEoMMs51ht7OjJ9TCdTh2FSLo42pF/9ZyyAi0PZTyxbNnbBUmVhXAQMimcHEkKIeyF/ScV9rXUTV4YE+kiSX0vY21rSzNtJkvy7ZGWpNMskXxR7JqSNcbVraysl40PalHmscx0bZo57jCYNHHFzsmX4k753XCIkhBC3kjv6QgghRDXq0qY+n8/pw9nUbJo2cERtZ1Xu8Y+29ODR/w1oFkKIypBEXwghhKhmDvZWtGsuTx6FEDVLno8LIYQQQghRC0miL4QQQgghRC0kib4QQgghhBC1kNTo1yC9Xs+VK1fIyspCp9OZOpxyqVQqjh8/buow7js2NjZ4eXlhaWlp6lCEEEII8YCTRL8GpaSkoFAoeOihh7C0tDTrFTg1Gg329vamDuO+YjAYyMzMJCUlhcaNG5s6HCGEEEI84KR0pwZpNBoaNGiAlZWVWSf54t4oFApcXV3Jz883dShCCCGEEJLo1zQLC/nIazP5AieEEEIIcyFZ533k8OHDjB49muDgYAYMGMD48eM5ffq0qcMCYPr06fTt25fc3NwS+9u3b09KSoqJohJCCCGEeHBJon+fKCws5Nlnn2X69Ols3LiRTZs2ERwczDPPPGM2A3svXrzI66+/buowhBBCCCEEMhj3vpGXl0dOTk6JO+YDBw5ErVYza9Ys3N3dmTZtGgDx8fFs2bKFMWPGsGTJEry9vTl9+jRFRUXMnTuXjh07kpOTw9y5czlx4gQKhYKAgABefPFFVCoVDz/8MOPGjWP//v2kp6czfvx4RowYUWGMY8aMIT4+nh9++IE+ffqUen3btm18+OGH6PV67O3tmTFjBm3btmXp0qVcvHiRjIwMLl68iIeHB++88w7u7u6sWrWKmJgYLC0tsba2Zt68eWRnZ/PSSy+xY8cOLCwsyMvLo3fv3iQkJBAeHs6gQYNITEzk0qVLhISE8MILLwCwevVqvvrqKywsLHBzc2P27Nk0btyY6dOno1arOXnyJJcvX8bX15e33npLBiMLIYSocfv27bvt/scee6yGIxG1QbUl+r6+vpw8ebLcYz744APWrl3L2LFjWbhwYYXH15SlS5cSExODm5sbBoMBg8HAK6+8QpcuXUwWk6OjI1FRUYwfPx43Nzc6dOhA586d6d+/P56enjzzzDM899xzqFQqYmNjmThxIgBHjhzh1VdfpWXLlqxcuZIlS5bw9ddfs2DBApycnNi4cSNarZZJkyaxcuVKJkyYQGFhIU5OTsTExHD06FGGDx9OWFgY1tbW5cbo4uLCwoULeemll2jbti3169c3vnbmzBleffVVYmJi8Pb2JjExkcmTJ7N582YADhw4wPr161Gr1UycOJGYmBimTJnCG2+8wY4dO3B3d2f9+vUcPHiQYcOG4ejoyK5du+jRowcJCQl07doVFxcXAHJzc1m1ahVpaWk88cQThIWFkZKSwqeffsrq1atxcXFh7dq1TJkyhYSEBACOHj3Kl19+iUKhYOjQoWzevJmwsLDq+FUKIYQQZZo/f77x54KCAlJTU3n44Yf59ttvTRiVuF+ZtHQnPj6ezz//nMjISFOGcVsRERHEx8ezYcMG3n77bV588UVTh0RkZCR79uxh1qxZ1K1bl08++YTQ0FC8vLzw8vLip59+4syZM6Snp+Pv7w+Ap6cnLVu2BKBVq1ZkZ2cDsHPnTkaNGoVCocDKyoqIiAh27txp7Ktnz54AtG7dmsLCwlK192Xx9/dn0KBBREVFodfrjfv37t1Lly5d8Pb2BjAm5kePHgWK71So1eoScSqVSvr27UtERATz5s3DwcGB8PBwAEaOHElsbCxQfKd++PDhxr4CAwMB8PDwwNXVlezsbHbt2kVQUJDxy8DgwYNJS0szjh8ICAjAysoKS0tLfHx8jJ+TEEIIUZM2btxo/G/Lli3ExcUZ/3YKcbeqPdFPSkriqaeeYvLkyfTp04fnn3+ewsJC5syZQ1paGlOmTCmxMNPSpUtZunSpcbt3796kpKSg0+l48803GTRoEAMHDuS///1vue1v3ryZkJAQQkJCCA4OxtfXlyNHjnDq1ClGjx5NWFgYvXr1uqNvyDk5Obi6ulb5Z3M3Dh48yKeffoparaZXr168/PLLJCQkoFAo2LNnDyNHjiQuLo41a9YwdOhQ4+wvNjY2xjYUCgUGgwEoXrzr5hli9Ho9RUVFxu0bd+9vHHPjvDvx4osvotFoiI6OLtH+rTPSGAwGY59lxblo0SKio6Np2LAhK1asMH7hCg4O5uDBg+zdu5fc3Fw6depUKvab27r5S8fd9C+EEEKYkq+vr9lMvCHuPzVyR//QoUPMmTOH77//ntTUVHbv3s28efNwd3dnxYoVxjvO5blx93bdunWsWbOG7du3c+DAgTLb79u3L/Hx8cTHx9O5c2dGjBhB27Zt+b//+z8mT55MXFwcX375JW+//fZt+4uJiSEkJIR+/foxbtw4xo4dW3UfyD1wcXFh2bJlxvcMkJGRwfXr1/Hx8aFPnz4cP36cH3744Y5KTvz9/fn6668xGAwUFhYSGxtLt27dqiRWKysr3n33XVauXGmcU75r167s3r2bCxcuABhr6Nu1a1dmO1evXqVHjx44OTkxbtw4XnjhBX777TcAbG1tGThwIDNnziQiIqLCmAICAvjuu++4evUqAHFxcTg5OdGoUaPKvl0hhBCiyuzbt8/4X1JSEp9//nmJG3FC3I0aGYzbvHlz6tWrB0DTpk3vqSwiMTGR48ePs3fvXqC4DvvkyZM0a9as3PbXrFnDsWPH+OKLL4DiaSB37drF8uXLOXXqVJklKRERETz33HMA/Pnnn4wcOZLGjRvTsWPHu469KjRu3JiPPvqIJUuWcPnyZaytralTpw5vvPEGTZo0AaBPnz5cuXLFWJ5SnlmzZrFgwQKCg4PRarUEBAQY6/qrQpMmTfjPf/7DrFmzAGjWrBmvvvoqU6dORafTYWNjQ3R0NHXq1CmzDRcXFyZNmsS4ceOwsbFBqVSyYMEC4+uDBw8mNjaW0NDQCuPx8/MzfmHT6/W4uLiwfPlyWddACCGEWbm5Rl+hUODo6MjcuXNNGJG4n9VIon+7UoqyKBSKEmUWWq0WAJ1OR1RUFE8++SRQfLfX3t6ew4cPl9n+r7/+SnR0tHHWFoAXXngBBwcHevXqRVBQEJs2baow/iZNmtChQwcOHz5sskQfoEuXLmUOCM7NzWX//v3MmTPHuK9z584l3t/N287Ozrz77ru3bevkyZNoNJoS2xVZuHBhqX1DhgxhyJAhxu1+/frRr1+/Usfd+EJ1u+2IiIjb3rE3GAzs3LmTkJCQEl8WduzYUeK4m7dHjhzJyJEjK4z9du9FCCGEqAlz587lnXfe4dtvvyU2Npbvv/9eyknFPTO725nOzs788ccfQPGMMRkZGUBxkhsbG4tWq0Wj0TBixAgOHz5cZjuXLl3i3//+N4sXL8bNzc24f8+ePTz//PM8/vjjxsGnFc1Df+3aNY4dO0arVq0q+/aqxa5du+jZsycBAQE88sgj1dLH3r17jWMebv3vjTfeqJY+yxMYGMiOHTv417/+VeN9CyGEENVlwYIFTJw4kYyMDN59912GDBkiN6DEPTO7efSDgoL44YcfCAoKonXr1sbkOiIignPnzjFo0CCKiooYPHgwnTt3Jikp6bbtfPzxx2g0Gl577TVjIv/ss8/y3HPPMWLECKytrWnRogUNGjQgJSWlVK12TEwM27Ztw8LCgoKCAoYMGULXrl2r983fo4CAgDLn3a0qXbp0IT4+vlr7uBu33rkXQgghagODwUCPHj1Yv349AQEBBAUF8dlnn5k6LHGfUhjkeVCVKygo4OjRo7Rp06ZEWdHx48fvaOCxOdBoNLJg1D26n37P4t4dPHjQpKV8Qpg7uUYqNmHCBABWrFhh3Dd48GC+/PJL5syZg7+/P61bt+aVV15hzZo1pgpTVKPKXidl5Zw3mF3pjhBCCCHEgyokJIRevXpx+PBhnnzySWJjY3n22WdNHZa4T5ld6Y4QQgghxINq7NixPPHEE7i5uWFlZcXs2bNNHZK4j0miL4QQQghhRjw9PU0dgqglpHRHCCGEEEKIWkgS/QdYSkoKvr6+7Nmzp8T+3r17k5qaaqKohBBCCCFEVZDSHTOn1xvYeSiF+J1nuJKVj5uTDSHdm9K9vRcWFopKt29pacns2bPZsGEDarW6CiIWQgghhBDmQO7omzG93sCbX+zjozXJ/JGSTdb1Av5IyeajNcm8+cU+9PrKz4zq7u5Ot27deOutt0q9Fh0dTVBQEMHBwSxcuBCdTkdKSgqhoaFERUUxYMAAxo4dS1ZWFlqtlqioKEJDQwkNDSU2Npbr16/TuXNnrl+/DhQ/QQgKCiqzDYAff/yRkJAQgoODmTx5MleuXAGKnzK89957hIeH079/f44ePcq5c+fo2bOncSXlpKQkxo8fT1JSEpGRkUyYMIGgoCAWLVrExx9/zODBgxk8eLCxzfL6SklJMbY5evRoAD7//HMGDhxIaGhoiRWIhRBCCCHMkST6ZmznoRQOn8ogv7Dkyr35hToOn8pg5+GLVdLP9OnT2b17d4kSnj179rBjxw7i4uJYt24d586dIyYmBoATJ04QGRnJpk2bcHBwYOPGjRw6dIjs7GzWr1/P8uXLOXDgAGq1mp49e7J582YA1q9fT2hoaJltZGZmMmfOHD766CM2btxIhw4dmDdvnjEmJycn1qxZQ0REBMuXL6dRo0Z4eXkZF01bv349gwcPBiA5OZm5c+cSFxfHN998g4uLC2vXrsXX15eEhIQK+7qVTqdj+fLlxMXFsXbtWrRaLWlpaVXy+QshhBBCVAdJ9M1Y/M4zpZL8G/ILdcT//EeV9KNWq5k/fz6zZ8823n3ft28f/fv3x9bWFpVKRVhYGImJiQC4uroaVyxu3rw52dnZNG/enLNnz/L000+zefNmXn75ZQDCwsKMK+pu2rSJkJCQMts4cuQIbdu2xcvLC4Bhw4axd+9eY5wBAQHG4288AQgLC2PDhg3k5eWxd+9eAgMDAfDx8aF+/frY2tri7OxsXNXY09OTa9euVdjXrZRKJe3btyc8PJwPP/yQyMhIPDw8KvW5CyGEEEJUJ0n0zdiVrPxKvX43/P39S5Tw3CiHuVlRURFAiZXXFAoFBoMBZ2dnEhISGDVqFGfPnmXQoEFcu3aNTp06kZ6ezpYtW/Dy8jImx7dr49Y+DQaDsc+bz1Eo/hmb0LdvX/bs2cMPP/xA9+7djcdYWlqWaEupVJbYrqivGwtG37zv448/5rXXXsNgMDB+/Hj27dtX6jMSQgghhDAXkuibMTcnm0q9frdulPCkp6fTqVMnEhISyM/Pp6ioiLi4OLp06VLmudu3bycqKoqePXsya9Ys7OzsuHTpEgqFgtDQUBYsWGAsqylLu3btSE5ONtbHr169ms6dO5d7jq2tLd27d2fx4sUVtn+nfTk7O/PHH38Y3xfA1atXCQoKwsfHh3/961/4+flx8uTJO+5PiAeJQacl65d1XIp5nb93/R96bUGVtX39912krV/C3ztj0Rfk3vX5RdkZ5CTvoCC1ap6IVoXrR3eRFvcOmTu+QpebY+pwhBC1iMy6Y8ZCujflozXJty3fsbFSEtKjWZX2d6OE5+mnn6Z79+4UFBQQFhZGUVER/v7+jBo1isuXL9/23O7du7Nlyxb69++PtbU1AwcOxNfXF4D+/fuzcuVKHn/88XL7d3NzY968eUydOhWtVounpyevv/56hXH379+fX3/9lXbt2t3xey2vr+eff5758+fz4Ycf4u/vD4CLiwvDhg0jPDwcW1tbGjduTFhY2B33J8SDJHPrf7l2sHhsTt6ZX9H+fQn3gc9Xut3sA5vJ/OETADRA3vnf8Rw1t8QxBp0WzekDGIq02Pt0wsLK1vha7p+HuRz7JuiKn9Q5BQzFpfuwSsdVGdcObeXKd9HG7fxzv9MgcqEJIxJC1CYKw40aBVFlCgoKOHr0KG3atClRonL8+HFatmx5x+3cmHXn1gG5NlZKHvGpy4yxj1XJFJu3o9FosLe3r3Q7er2eb7/9lrNnzzJr1qwqiKwknU7HkiVLcHV1JTIyssrbvxd3+3sW1UOnyaYwMwXr+s2wsLSu+IS7dPDgQTp27Fjl7dYGfy0aXfJuu4WSxtNjUCgq9xD54sqXKbh0psQ+t77PYF2/GdaezdAXFZL6xSsUXv4TAJWjOw0iF6K0dyw+/4uZFKT88yROobSk0bTPsbC2pTIMeh2ZWz8nJ/lHlHYOuASORt2y2x2dm/rlLPIvHC+xz+vZ97Fy86pUTOZArpGKTZgwAYAVK1aYOBJhKpW9TsrKOW+QO/pmzMJCwYyxj7Hz8EXif/7jn3n0ezSj+yMNqi3Jr0pTp07l0qVLfPbZZ9XSflhYGM7Ozixbtqxa2hf3p5wjP5LxXTToirCwVVNv2CvYNPAxdVjVTpevwVCkRaV2MmkcyjouJRJ9pdq50kk+gIWdQ6l9VzYX3+F3eDQIG+8WxiQfoCg7nZwjP+LUtXi2L0NhyXFNBl0RBl0RlXXt4A9cO/D9//rMJ339+9h4tUBVx6XCc5X2t/yuLFQobetUOiZzlp9ykivfL6cw8yL2zTtRt/8kLGwqf2NJCFGaJPpmzsJCQc8OXvTscH/e3fn444+rtf3169dXa/vi/mPQacnc+l9jeYY+7zpXd3yN5+iyp0+tDa7++DVZezeCXod9i864h7yAQmVZ8YnVwPXxcaTFLcKgzUehssLtiaeqpF3n7hEUpJy8bW3+tQPfo3J0K7X/5vEBDh36cGXzP3dO7Vt1Q2lX+aQ6/+It43X0RRSknkblW/4YIwDngCHknf8dfe41QIGzf7jxCURtZNAVkRa3CN31qwBoTiSitHPArd8EE0cmRO0kib4QolbRF+ajz79eYl/RtSs11r9BryP3zCEM2gLsmnXEwqpqB83fTn7KSbJ+WWfc1pzYS85DO3Do2KfCc3V5OWCgShLeG+yatqfh8ysovPwnVu4PVVnbNp7NaDg1mrzzx8hKXE9ByombXjVg3cAXZR0XdDnFSaTC2o46D/cwHuHQsQ8qcCTy6QAAF4dJREFUBzdyz/yKVd2G1Hmkd9XE1cAXze+7/9lhocTas/kdnWvl3oiGU6PJv3AcS+d6WDrXq5KYqov2aiqak/tQ1XHFvmUXFMq7+zJZlJ1uTPJvyC/xexRCVCVJ9IUQtYrStg62TR4h78/Dxn3q1gE10rdBV0Tq13OMdeC31ohXl8IrKXe072YGg4Er368g5/A2AOq0641b0LNVUmIDoLSxx/ahh0vt15xIQnP6AFZuDXDo2PeuvwhZ2Nhj79MJg05L+k0JonUDH2y9W9Ag8m1ykrdjKCqkTttepRJnu+YdsWtetXXjDh37oM28SM6R/9Xo9x59R2U7BoOBa/sT0Jzah6VzfZwDhlZpXFUt/+IpLn01B4NOC4Dtb49Qf/jsu2pD5eiOUu2M7vrfxn3WXr5VGqcQ4h+S6Ash7prmRBLZ+zeBhRKnrqHYNXnE1CGVYNu0w02JvgJLd+8a6Vdz6kCJwZ5F2elcO7wNZ7/qnaHJrnFbUKqM5UpAhcls7ukD5BzaYtzOObwNu6YdsG9RcbnJvbr26xaufL/cuJ137ij1I+5tkL66ZTcsImzRHE9E5eSOw6P9AFDVccbZP7xK4r1TCgslbn2fwa3vM3d1XnbSBq5u/xIonm2nIPUUXs8sqY4Qq0T2vk3GJB8g78/DFKT9hbXHQ3fchkKpwmPwv8n4fjnazFTsfR7FpdeoaohWCAGS6Ash7lJ+6h+kxb0DFE/Ydfn8cbwmLMHK1dO0gf2PwWAga8+am/eQ9XMMdVr5V3vfmlOlF1HTZqVVe78qx7rUGzaTrN1xGLT5OHTsW+GXL+2VC6X2FV65gD3VmOgf2lZiO+/MIYquZaJycL2n9uyatseuafuqCM0kNMcTS2wXpp+n8EpKtcy4k5O8g6xf1mLQ63HsPBDHR/tWeR93ysa7Bd4TzPcLzb3atGkTGzZsuKtzTp06hY9P7Z8oQJiOJPoPuM2bN7NixQqKioowGAyEhIQwfvx4U4clzFjuqf3cSPIB0BeRd+ZXs0n0wYChIK/EHl2+pkZ6LrqaWnqnrvQ6GNXBrnE77Brf+VoStk3aw0+rwDjDsgK7ph2qJ7j/UdreMrOKUoWiBsYw6IsKQa8rMae+OVA5uVOQetq4rVBZoVQ7V3k/Bal/kLHpI+N25g+fYFXXG9tGre+qHcdO/ck9ue+f0p0m7e7qbr4ozcfHh759TfelS9R+kuibOYNBz/Xfd5OdtJGinExUdVxx7ByMurV/pWtp09LSeOutt1i7di3Ozs5oNBpGjx5N48aNy10FVzzYLG+T0Fu6mEuSDwqFBdYNmpN//phxn3X9ql1criy3S9JUTu7Gn3W5OWRuXUneuaNY12+KW5/xqBxKzxRTEyxdPVE51KUoOx0ojt3S2aNa+3QOGEZ+yikM2uJpLp27haGs5mkVs35Zy9+74zAUFaJ+uCd1+09EYaGs1j7vlEuPCApST1OUlQ5KFS6BY6vl88g7d7T0vr9+u+tE38bLlwbPLCbn8DZUTh44VNFg5tpiwIABDBgwwNRhCFGCJPpmzGDQk7bmHfLOJmP43xRxhZpsrnwXjeZ4Ih7hUZVK9v/++2+0Wi35+cV/dO3t7Vm4cCG//vorkZGRxMbGArB27VqSk5Np164du3btIjs7mwsXLuDn58drr70GQHR0NBs2bECpVOLn50dUVBSXLl1i6tSpNG/enOPHj+Pq6sr777/P1q1b2bt3L++++y4AS5cuxdramoKCAlJTU/nrr7+4evUqkyZNIjExkeTkZFq0aMGSJUtQKBRl9jVmzBh27NhhbBNg4sSJzJw5k9Oni++ajRgxgqFDzXvAm7lTt/Ij9/QBNMd/AYUFddr1xtbMyie0mRdLbBem/VUj/Tr3iCj5xEOpwqHdP8nQlc0rij83IDfnKul51/Ecs6BGYruV5kSiMckH0F2/Ss5vP+PYKaja+rTxbkHDqcvIO3cUK9cGWLk3qra+AAou/8nVH78xbl8/sgMbL18c2pe/SndNsXTxxHvShxSm/YXKsS7K26wTUBWs6ze9o30V0eVdJ2PTR8WzHSlV6POv49RlIHlnf8PCxg4brxZVEa4QogpVzfQKolpc/313iST/BoO2gLyzyWh+31Op9lu0aEFgYCCPP/444eHhvPPOO+j1eoYNG8aVK1c4f/48UDxX/eDBgwE4dOgQH3zwARs2bODHH3/k5MmT/Pzzz+zYsYO4uDjWrVvHuXPniImJAeDEiRNERkayadMmHBwc2LhxI0FBQSQmJnL9evEUiJs2bSIkJAQorlf86quvmD9/PjNmzOCZZ55h06ZNHDt2rMK+bufQoUNkZ2ezfv16li9fzoEDByr1mYkbg+leouFzK2j4/Arq9p+EQmFei7fpi7Qltg1FhTXSr7V7IzzHzMe+ZTfUbbrTYMzrqBzrGl/PO3ukxPH5F45juCXWmqIvyC+9rzDvNkdWLaWdA+qW3ao9yQcouHy21L7CtNL7TElhocS6ftNqS/IBbB96GKeAoSgsbVCorHDsEoK9T6e7bic7aeM/U5rqivj7p285//FULq9+ndQvXuFy7JsYDIbyGxFC1ChJ9M1YdtLGUkn+DQZtAVlJGyvdx9y5c9mxYwfDhw8nNTWVoUOHsnXrVgYMGMCGDRtITU0lMzOTdu2Ka3/bt2+PWq3G1tYWb29vsrOz2bt3L/3798fW1haVSkVYWBiJicWDzFxdXWnVqhUAzZs3Jzs7G3t7e3r06MHWrVs5cOAA3t7eeHgUlwz4+fmhUqnw9PSkbt26NGvWDJVKhYeHR4V93U7z5s05e/YsTz/9NJs3b+bll1+u9GcmiqkcXFFVQz1xVXDsWLLm9caMLDXBxrslHoNfwj3kX1h7liwZsrqlntnStYHJFrVSt+xaYqVZhbUd6jY1Mw1pTbF9qA3cUqZjexfjGGoTl+7DeOilL3jo31/iGjjmntrQXr14yx4DupvWqMg9fYD8879XIkohRFWT0h0zVpSTWe7rupzKLQL0008/kZubS1BQEGFhYYSFhREbG8uaNWuIiori+eefx8rKyni3HcDa2tr4s0KhwGAwoNfrS8deVFTm8QBhYWEsW7YMLy8v49MCAEvLf5Ielar0/55l9XVz2zf2qVQqnJ2dSUhIYM+ePfz8888MGjSIhIQEHByq7+6ZMD3nniOwqteYgpSTWHu3QN2iq6lDAsCt3wTS175LYfo5VM71qBv8nMliUdo70uCpt8g5tA2DXofDI4FYOrpXfOJ9xNLJA4/B/+bvXbHG2YjsfR8zdVgmo1BW7k++vU/nEjMFKZSWJabbhOJxKEII8yGJvhlT1XGlUJNd5uvKOpUbxGdjY8P8+fNp27YtXl5eGAwGjh8/TsuWLfH09KRevXrExMTw7bfflttOly5dWLZsGcOGDUOlUhEXF1fhYN5HH32Uy5cvc/HiRV555ZU7jrmsvhwcHMjKyuLq1auo1Wp27dpFr1692L59Oxs2bOC9994jICCAxMRELl26JIl+LadQKFC37Ia6ZTdTh1KClWsDvJ5ZjC4vBwsbtclLniwd3XHpOcKkMVQ3e9/HHujkviqp2wSgL9CQc+QnlPZO2LfoQkbCx6AvnllKqXbBrql5rakhxINOEn0z5tg5mCvfRd+2fEdhaY1T5+BKtd+lSxemTp3KxIkT0WqL78oEBAQwZcoUtFotQUFBbNmyxVhWU5ZevXpx/PhxwsLCKCoqwt/fn1GjRnH58uVyz3viiSfIysrCysrqjmMuqy+VSsX48eMJDw+nXr16PPxw8Yqc3bt3Z8uWLfTv3x9ra2sGDhyIr6+swihMS2lbx9QhCHFPHDr2xeGm0jhLl/rkJO/AwtoOh079zG4KUyEedAqDjJypcgUFBRw9epQ2bdqUKF25cbf8Tt1u1h0oTvJtG7er9Kw75cnOzmbu3Ln07duXJ598skrbNhgMaLVaIiMjmTlzJq1b390Ub+bubn/P4v508OBBOnYsf/VZIR5kco0IUbHKXidl5Zw3yGBcM6ZQWOARHkXdoElY1WuK0t4Rq3pNqRs0qVqTfIPBQJ8+fVAoFDz+eNVPQ5eRkYGfnx/t2rWrdUm+EEIIIYS5kNIdM6dQWKBuE1Cjs2EoFAq2b9+OvX31LGTj7u7O/v37q6VtIYQQQghRTO7oCyGEEEIIUQtJol/Dbjc9pKg9ZMiLEEIIIcyFJPo1yN7enosXL1JYWCgJYS1kMBjIzMzExsbG1KEIIYQQQkiNfk3y8vLiypUrnDt3zriglLkqLCy8q2kvRTEbGxu8vLxMHYYQQgghhCT6NcnCwgJ3d3fc3c1/9cmDBw/Srt2DuVS8EEIIIURtIKU7QgghhBBC1EKS6AshhBBCCFELSaIvhBBCCCFELSQ1+tXgxow6hYWFJo6kcgoKCkwdghBmTa4RIcon14gQFavMdXIj1yxrNkeFQeZ5rHI5OTmcOnXK1GEIIYQQQogHgI+PD3Xq1Cm1XxL9aqDX69FoNFhaWqJQKEwdjhBCCCGEqIUMBgNarRZ7e3ssLEpX5EuiL4QQQgghRC0kg3GFEEIIIYSohSTRF0IIIYQQohaSRF8IIYQQQohaSBJ9IYQQQgghaiFJ9IUQQgghhKiFJNEXQgghhBCiFpJEXwghhBBCiFpIEn0hhBBCCCFqIeVrr732mqmDEPeHP//8k6effpr9+/eTmprKI488YuqQhDA7Op2OsWPH0rx5czw8PEwdjhBm5/Tp08ydO5eff/4ZW1tbGjZsaOqQhDAr+/fv5/3332fLli1kZ2fTunXre25LVYVxiVru4MGD1KtXDxsbG9q3b2/qcIQwS9HR0bi7u5s6DCHMVm5uLjNnzkSpVLJ48WL8/PxMHZIQZuXatWvMmzcPKysrJk+ezJAhQ+65LUn0RZk+/fRTdu/ebdyeM2cOgYGBqNVqJk2axGeffWbC6IQwvVuvkeHDh9O8eXP0er0JoxLCvNx6naxcuZLz588zffp0xowZY8LIhDAPt7tGDAYDixYtqvQ1ojAYDIbKBigeDOvXr6dr1654eHjw7LPPsnz5clOHJIRZefHFF1Gr1Rw9epSmTZvyzjvvmDokIczO0aNHeeihh1Cr1Tz11FOsXLnS1CEJYVauXbvGm2++yYgRI3j44Ycr1ZYk+uKOHTlyhM8//xy1Wk3Pnj0JDAw0dUhCmKWlS5fSs2fPSv8DLURtdPDgQb788kvUajU+Pj6MHTvW1CEJYVZefvllLl++jLu7O/Xr1+ell16657Yk0X8AXb9+nYiICKKjo/Hy8gJg48aNLFu2jKKiIsaOHcvIkSNNHKUQpiPXiBAVk+tEiPKZwzUi02s+YJKTkxk+fDh//fWXcV9aWhpLlixh1apVrF+/ntWrV/PHH3+YLkghTEiuESEqJteJEOUzl2tEEv0HTGxsLK+++mqJWUF++eUXunTpgpOTE3Z2dvTp04fNmzebMEohTEeuESEqJteJEOUzl2tEZt15wLz++uul9qWnp1O3bl3jtru7O0eOHKnJsIQwG3KNCFExuU6EKJ+5XCNyR1+g1+tRKBTGbYPBUGJbiAedXCNCVEyuEyHKZ4prRBJ9Qb169cjIyDBuZ2RkyII/QtxErhEhKibXiRDlM8U1Iom+oFu3biQmJnL16lXy8vLYsmUL3bt3N3VYQpgNuUaEqJhcJ0KUzxTXiNToCzw8PJg2bRpjxoxBq9USHh5O27ZtTR2WEGZDrhEhKibXiRDlM8U1IvPoCyGEEEIIUQtJ6Y4QQgghhBC1kCT6QgghhBBC1EKS6AshhBBCCFELSaIvhBBCCCFELSSJvhBCCCGEELWQJPpCCCGEEELUQpLoCyHEA+jChQs899xzd3VcWloaERER1R1ahT788EO2bdtm6jCEEMLsSaIvhBAPoNTUVM6ePXtXx3l4eBATE1PdoVUoKSmJoqIiU4chhBBmTxbMEkKI+1BSUhKLFy+mfv36nD17FltbWyZMmMBXX33F2bNnefLJJwkMDGT+/Pls2rTJeM78+fOJj4+nb9++pKWl0alTJz777DOio6PZvn07+fn55OXl8Z///IfevXuXOG7u3LkEBwdz6NAhtFotCxcuJDExEaVSSdu2bZkxYwZqtZrevXszaNAgEhMTuXTpEiEhIbzwwgvlvp/p06eTlZXFhQsX6NmzJ+Hh4cybNw+NRkNGRgYtWrTgvffeY82aNSxatAhnZ2dmzJhBjx49WLRoEfv370en09GqVStmzZqFWq0usy+NRsOMGTM4d+4cFhYWtG7dmnnz5gHwxhtvkJycjEajwWAwsGDBAjp27Mj06dOxsbHh1KlTZGZm0rt3b5ycnPjxxx/JyMhgwYIFdO3alenTp2Ntbc2JEyfIzMzEz8+PWbNmYWlpWXW/fCGEuENyR18IIe5Tv/32GxMmTCA+Ph61Ws2KFStYvnw5a9euZdWqVaSnp9/2PKVSyYIFC2jYsCGfffYZFy9e5JdffuGrr75i48aNTJs2jQ8++KDUcTdbtmwZ6enpxMfHEx8fj16v5+233za+npuby6pVq4iJiWHlypVcuHChwveTn59PQkICUVFRxMbGEhoaSmxsLFu2bCElJYWffvqJkSNH0qZNG15++WWeeOIJVqxYgVKpZO3atWzYsAF3d3cWLVpUbj9bt25Fo9EQHx/PmjVrgOISpeTkZNLT01m9ejXfffcdgwYN4pNPPjGed+zYMb744gu+/vprVq5ciZ2dHTExMYwZM6bEcUeOHGHlypV89913nDlzhtWrV1f43oUQojqoTB2AEEKIe+Pl5UWrVq0AaNiwIXXq1MHKygoXFxfs7e3Jzs6+o3YaNGjA22+/zcaNGzl37pzxjnZ5du7cybRp04x3qkePHs2UKVOMrwcGBgLF5T6urq5kZ2fj7e1dbpsdO3Y0/hwVFcWePXv45JNP+Ouvv0hPTyc3N7fUOT/99BM5OTn88ssvAGi1WlxdXSvsZ8mSJYwePZpu3boxduxYGjVqRKNGjXB0dCQmJoYLFy6QlJSEvb298bxevXphaWlJ3bp1sbOzIyAgACj+7LOysozHDRo0yHheSEgI27dvZ9SoUeXGJIQQ1UESfSGEuE9ZWVmV2FapSv6T7uPjw83VmVqt9rbt/P7770yePJlx48bh5+dnLNMpj16vR6FQlNi+uX1ra2vjzwqFgjupErWzszP+/OKLL6LT6ejXrx89e/bk0qVLt21Dr9czc+ZMevToARSX5RQUFJTbj7e3N1u3biUpKYm9e/cSGRnJvHnzsLCw4PXXXycyMpLAwECaNGnChg0bjOdV9HnfoFQqjT8bDAYsLOThuRDCNORfHyGEqKUcHBxITU0lMzMTg8FAQkKC8TWlUmlMzPfv30+bNm2IjIzkscceY/v27eh0ulLH3SwgIIBvv/0WrVaLXq/nm2++wc/Pr8pi3717N1OmTCEoKAiA5OTkEjHdGIzr7+/PN998Q2FhIXq9ntmzZ7N48eJy2161ahUzZszA39+fqKgo/P39OXbsGHv27KFXr16MGDGCNm3asG3bNmOfd+P777+nsLCQgoIC1q1bR69eve66DSGEqAqS6AshRC1lYWFBREQEYWFhDB06FC8vL+NrzZo1w9ramvDwcAYMGMDff/9Nv379CAoKws7OjuzsbK5fv17iuJvvqE+aNAk3NzdCQ0Pp168fRUVFvPLKK1UW+7Rp05gyZQrBwcHMmTOHTp06cf78eQB69+7N4sWLWbduHZMnT6ZBgwYMGjSIoKAgDAYD06dPL7ft0NBQdDodQUFBDB48mJycHEaPHk1ERAT79u0jODiYQYMG4e3tTUpKCnq9/q5it7GxYcSIEQQHB/Poo48SFhZ2z5+DEEJUhsy6I4QQQlSR6dOn07x5c55++mlThyKEEFKjL4QQovr9+eefTJs27bavNW7cmPfee69K+3vhhRfKXCdgyZIlNGnSpEr7E0IIcyR39IUQQgghhKiFpEZfCCGEEEKIWkgSfSGEEEIIIWohSfSFEEIIIYSohSTRF0IIIYQQohaSRF8IIYQQQohaSBJ9IYQQQgghaqH/BwTePjDXbjwHAAAAAElFTkSuQmCC\n" }, "metadata": {}, "output_type": "display_data" @@ -1077,7 +1139,7 @@ }, { "cell_type": "code", - "execution_count": 22, + "execution_count": 23, "outputs": [ { "name": "stdout", @@ -1128,7 +1190,7 @@ }, { "cell_type": "code", - "execution_count": 23, + "execution_count": 24, "outputs": [ { "name": "stdout", @@ -1169,7 +1231,7 @@ }, { "cell_type": "code", - "execution_count": 23, + "execution_count": 24, "outputs": [], "source": [], "metadata": {	Set verbose flag from parameters		2021-10-17T15:44:28	[]	[]
trevismd/statannotations	trevismd__statannotations-28	284ba0a621bde694a85320d976df93a8d15d899d	diff --git a/CHANGELOG.md b/CHANGELOG.md index 80407ff..9c2425b 100644 --- a/CHANGELOG.md +++ b/CHANGELOG.md @@ -1,4 +1,7 @@ ## v0.4 +### v0.4.1 + - Support for horizontal orientation + ### v0.4.0 - Major refactoring, change to an Annotator `class` to prepare and add annotations in separate function (now method) calls. diff --git a/README.md b/README.md index 7e9b432..fff5153 100644 --- a/README.md +++ b/README.md @@ -42,6 +42,7 @@ corresponding branch). - Interface to use any other function from any source with minimal extra code - Smart layout of multiple annotations with correct y offsets. +- Support for vertical and horizontal orientation - Annotations can be located inside or outside the plot. - Corrections for multiple testing can be applied (binding to `statsmodels.stats.multitest.multipletests` methods): @@ -51,7 +52,7 @@ corresponding branch). - Benjamini-Yekutieli - And any other function from any source with minimal extra code - Format of the statistical test annotation can be customized: - star annotation, simplified p-value, or explicit p-value. + star annotation, simplified p-value format, or explicit p-value. - Optionally, custom p-values can be given as input. In this case, no statistical test is performed, but corrections for multiple testing can be applied. diff --git a/codecov.yml b/codecov.yml index 0d54c1d..4cc71ba 100644 --- a/codecov.yml +++ b/codecov.yml @@ -6,4 +6,4 @@ coverage: patch: default: - threshold: 0.5% \ No newline at end of file + threshold: 2% diff --git a/docs/build/doctrees/environment.pickle b/docs/build/doctrees/environment.pickle index 771aba2..169ad27 100644 Binary files a/docs/build/doctrees/environment.pickle and b/docs/build/doctrees/environment.pickle differ diff --git a/docs/build/doctrees/statannotations.doctree b/docs/build/doctrees/statannotations.doctree index ee0691c..759af7f 100644 Binary files a/docs/build/doctrees/statannotations.doctree and b/docs/build/doctrees/statannotations.doctree differ diff --git a/docs/build/html/genindex.html b/docs/build/html/genindex.html index 935f030..e1e1951 100644 --- a/docs/build/html/genindex.html +++ b/docs/build/html/genindex.html @@ -166,6 +166,7 @@ <h1 id="index">Index</h1> \| <a href="#L"><strong>L</strong></a> \| <a href="#M"><strong>M</strong></a> \| <a href="#N"><strong>N</strong></a> + \| <a href="#O"><strong>O</strong></a> \| <a href="#P"><strong>P</strong></a> \| <a href="#R"><strong>R</strong></a> \| <a href="#S"><strong>S</strong></a> @@ -365,6 +366,14 @@ <h2 id="N">N</h2> </ul></td> </tr></table> +<h2 id="O">O</h2> +<table style="width: 100%" class="indextable genindextable"><tr> + <td style="width: 33%; vertical-align: top;"><ul> + <li><a href="statannotations.html#statannotations.Annotator.Annotator.orient">orient (statannotations.Annotator.Annotator property)</a> +</li> + </ul></td> +</tr></table> + <h2 id="P">P</h2> <table style="width: 100%" class="indextable genindextable"><tr> <td style="width: 33%; vertical-align: top;"><ul> diff --git a/docs/build/html/objects.inv b/docs/build/html/objects.inv index 0d82dd7..b77fb8b 100644 Binary files a/docs/build/html/objects.inv and b/docs/build/html/objects.inv differ diff --git a/docs/build/html/searchindex.js b/docs/build/html/searchindex.js index 38da9ed..2e1b48e 100644 --- a/docs/build/html/searchindex.js +++ b/docs/build/html/searchindex.js @@ -1,1 +1,1 @@ -Search.setIndex({docnames:["index","modules","setup","statannotations","statannotations.stats"],envversion:{"sphinx.domains.c":2,"sphinx.domains.changeset":1,"sphinx.domains.citation":1,"sphinx.domains.cpp":3,"sphinx.domains.index":1,"sphinx.domains.javascript":2,"sphinx.domains.math":2,"sphinx.domains.python":3,"sphinx.domains.rst":2,"sphinx.domains.std":2,sphinx:56},filenames:["index.rst","modules.rst","setup.rst","statannotations.rst","statannotations.stats.rst"],objects:{"":{statannotations:[3,0,0,"-"]},"statannotations.Annotation":{Annotation:[3,1,1,""]},"statannotations.Annotation.Annotation":{formatted_output:[3,2,1,""],print_labels_and_content:[3,3,1,""],text:[3,2,1,""]},"statannotations.Annotator":{Annotator:[3,1,1,""]},"statannotations.Annotator.Annotator":{alpha:[3,2,1,""],annotate:[3,3,1,""],annotate_custom_annotations:[3,3,1,""],apply_and_annotate:[3,3,1,""],apply_test:[3,3,1,""],comparisons_correction:[3,2,1,""],configure:[3,3,1,""],fig:[3,2,1,""],get_annotations_text:[3,3,1,""],get_configuration:[3,3,1,""],get_offset_func:[3,3,1,""],has_type0_comparisons_correction:[3,3,1,""],loc:[3,2,1,""],new_plot:[3,3,1,""],print_pvalue_legend:[3,3,1,""],pvalue_format:[3,2,1,""],reset_configuration:[3,3,1,""],set_custom_annotations:[3,3,1,""],set_pvalues:[3,3,1,""],set_pvalues_and_annotate:[3,3,1,""],test:[3,2,1,""],validate_test_short_name:[3,3,1,""],verbose:[3,2,1,""]},"statannotations.PValueFormat":{Formatter:[3,1,1,""],PValueFormat:[3,1,1,""],sort_pvalue_thresholds:[3,4,1,""]},"statannotations.PValueFormat.Formatter":{config:[3,3,1,""],format_data:[3,3,1,""]},"statannotations.PValueFormat.PValueFormat":{config:[3,3,1,""],format_data:[3,3,1,""],get_configuration:[3,3,1,""],print_legend_if_used:[3,3,1,""],pvalue_format_string:[3,2,1,""],pvalue_thresholds:[3,2,1,""],simple_format_string:[3,2,1,""],text_format:[3,2,1,""]},"statannotations.format_annotations":{pval_annotation_text:[3,4,1,""],simple_text:[3,4,1,""]},"statannotations.stats":{ComparisonsCorrection:[4,0,0,"-"],StatResult:[4,0,0,"-"],StatTest:[4,0,0,"-"],test:[4,0,0,"-"],utils:[4,0,0,"-"]},"statannotations.stats.ComparisonsCorrection":{ComparisonsCorrection:[4,1,1,""],check_valid_correction_name:[4,4,1,""],get_correction_parameters:[4,4,1,""],get_validated_comparisons_correction:[4,4,1,""]},"statannotations.stats.ComparisonsCorrection.ComparisonsCorrection":{apply:[4,3,1,""],document:[4,3,1,""]},"statannotations.stats.StatResult":{StatResult:[4,1,1,""]},"statannotations.stats.StatResult.StatResult":{corrected_significance:[4,2,1,""],correction_method:[4,2,1,""],formatted_output:[4,2,1,""],significance_suffix:[4,2,1,""]},"statannotations.stats.StatTest":{StatTest:[4,1,1,""],wilcoxon:[4,4,1,""]},"statannotations.stats.StatTest.StatTest":{from_library:[4,3,1,""],short_name:[4,2,1,""]},"statannotations.stats.test":{apply_test:[4,4,1,""]},"statannotations.stats.utils":{check_alpha:[4,4,1,""],check_num_comparisons:[4,4,1,""],check_pvalues:[4,4,1,""],get_num_comparisons:[4,4,1,""],return_results:[4,4,1,""]},"statannotations.utils":{check_is_in:[3,4,1,""],check_not_none:[3,4,1,""],check_order_in_data:[3,4,1,""],check_pairs_in_data:[3,4,1,""],check_valid_text_format:[3,4,1,""],get_closest:[3,4,1,""],get_x_values:[3,4,1,""],raise_expected_got:[3,4,1,""],remove_null:[3,4,1,""],render_collection:[3,4,1,""]},statannotations:{Annotation:[3,0,0,"-"],Annotator:[3,0,0,"-"],PValueFormat:[3,0,0,"-"],format_annotations:[3,0,0,"-"],stats:[4,0,0,"-"],utils:[3,0,0,"-"]}},objnames:{"0":["py","module","Python module"],"1":["py","class","Python class"],"2":["py","property","Python property"],"3":["py","method","Python method"],"4":["py","function","Python function"]},objtypes:{"0":"py:module","1":"py:class","2":"py:property","3":"py:method","4":"py:function"},terms:{"0":[3,4],"05":[3,4],"1":[3,4],"12141511":3,"9981846":3,"case":4,"class":[3,4],"default":[3,4],"float":4,"function":[3,4],"int":4,"return":3,"short":3,"static":4,"true":[3,4],A:3,If:3,Or:3,The:3,a_list:3,accept:3,add:3,addit:4,alpha:[3,4],an:3,annot:1,annotate_custom_annot:3,appli:[3,4],apply_and_annot:3,apply_test:[3,4],ar:3,arg:[3,4],argument:[3,4],arrai:3,assum:3,auto:3,ax:3,bar:3,base:[3,4],behavior:4,benjamini:4,between:3,bonferroni:4,bool:4,box:3,boxplot:3,calcul:3,callabl:4,check:3,check_alpha:4,check_is_in:3,check_not_non:3,check_num_comparison:4,check_order_in_data:3,check_pairs_in_data:3,check_pvalu:4,check_valid_correction_nam:4,check_valid_text_format:3,classmethod:3,close:3,closest:3,collect:3,color:3,com:3,comparison:[3,4],comparisons_correct:[3,4],comparisonscorrect:[1,3],comput:3,config:3,configur:3,constructor:3,content:1,coordin:3,core:3,corr_kwarg:4,correct:4,corrected_signific:4,correction_method:4,custom:3,data:[3,4],datafram:3,dict:4,displai:3,document:4,each:3,earlier:4,either:4,engin:3,equal:3,equival:4,error:3,error_typ:3,exact:3,exist:3,expect:3,fals:3,fig:3,follow:3,for_:3,format:[3,4],format_annot:1,format_data:3,formatt:3,formatted_output:[3,4],fraction:3,frame:3,from:[3,4],from_librari:4,func:4,gener:3,get:4,get_annotations_text:3,get_closest:3,get_configur:3,get_correction_paramet:4,get_num_comparison:4,get_offset_func:3,get_validated_comparisons_correct:4,get_x_valu:3,got:3,group:3,group_data1:4,group_data2:4,gt:4,has_type0_comparisons_correct:3,hochberg:4,hold:3,holm:4,http:3,hue:3,hue_ord:3,index:0,instanc:[3,4],instead:3,interfac:4,interpret:4,keyword:4,kruskal:4,kwarg:[3,4],label:3,leven:4,line_height:3,line_offset:3,line_offset_to_group:3,line_width:3,link:3,list:3,loc:3,ls:4,mann:4,messag:3,method:4,method_typ:4,mode:3,modul:[0,1],multipl:4,must:3,mylist:3,mynumb:3,name:[3,4],new_plot:3,none:[3,4],num_comparison:[3,4],number:[3,4],object:[3,4],one:[3,4],option:[3,4],order:3,otherwis:3,overrid:3,p:3,p_valu:4,packag:[0,1],page:0,pair:3,panda:3,paramet:[3,4],pass:[3,4],per:3,perform:3,plot:3,plot_param:3,point:3,posit:3,print_labels_and_cont:3,print_legend_if_us:3,print_pvalue_legend:3,properti:[3,4],provid:[3,4],pval:4,pval_annotation_text:3,pvalu:[3,4],pvalue_format:3,pvalue_format_str:3,pvalue_threshold:3,pvalueformat:1,rais:3,raise_expected_got:3,rang:3,refer:3,remove_nul:3,render:3,render_collect:3,reset_configur:3,result:[3,4],results_arrai:4,return_result:4,run:4,same:3,sampl:4,seaborn:3,search:0,sep:3,seri:3,set:3,set_custom_annot:3,set_pvalu:3,set_pvalues_and_annot:3,short_nam:4,show:3,show_test_nam:3,significance_suffix:4,simpl:3,simple_format_str:3,simple_text:3,smallest:3,sort:3,sort_pvalue_threshold:3,stackoverflow:3,standard:3,stat:[1,3],stat_nam:4,stat_str:4,statist:[3,4],statresult:[1,3],stats_param:[3,4],statsmodel:4,statsmodels_api:4,stattest:[1,3],str:[3,4],string:3,struct:3,submodul:1,subpackag:1,t:4,test:[1,3],test_descript:4,test_ind:4,test_long_nam:4,test_nam:4,test_pair:4,test_result_list:4,test_short_nam:[3,4],test_welch:4,text:3,text_annot_custom:3,text_format:3,text_offset:3,them:3,thereof:3,thi:[3,4],three:3,threshold:3,top:3,tupl:3,two:[3,4],type:3,union:[3,4],us:[3,4],use_fixed_offset:3,util:1,valid_valu:3,validate_test_short_nam:3,valu:[3,4],valueerror:3,verbos:[3,4],version:4,vs:3,wa:3,whitnei:4,wilcoxon:4,work:3,x:3,y:3,yekuti:4},titles:["Welcome to statannotations\u2019s documentation!","statannotations","setup module","statannotations package","statannotations.stats package"],titleterms:{annot:3,comparisonscorrect:4,content:[0,3,4],document:0,format_annot:3,indic:0,modul:[2,3,4],packag:[3,4],pvalueformat:3,s:0,setup:2,stat:4,statannot:[0,1,3,4],statresult:4,stattest:4,submodul:[3,4],subpackag:3,tabl:0,test:4,util:[3,4],welcom:0}}) \ No newline at end of file +Search.setIndex({docnames:["index","modules","setup","statannotations","statannotations.stats"],envversion:{"sphinx.domains.c":2,"sphinx.domains.changeset":1,"sphinx.domains.citation":1,"sphinx.domains.cpp":3,"sphinx.domains.index":1,"sphinx.domains.javascript":2,"sphinx.domains.math":2,"sphinx.domains.python":3,"sphinx.domains.rst":2,"sphinx.domains.std":2,sphinx:56},filenames:["index.rst","modules.rst","setup.rst","statannotations.rst","statannotations.stats.rst"],objects:{"":{statannotations:[3,0,0,"-"]},"statannotations.Annotation":{Annotation:[3,1,1,""]},"statannotations.Annotation.Annotation":{formatted_output:[3,2,1,""],print_labels_and_content:[3,3,1,""],text:[3,2,1,""]},"statannotations.Annotator":{Annotator:[3,1,1,""]},"statannotations.Annotator.Annotator":{alpha:[3,2,1,""],annotate:[3,3,1,""],annotate_custom_annotations:[3,3,1,""],apply_and_annotate:[3,3,1,""],apply_test:[3,3,1,""],comparisons_correction:[3,2,1,""],configure:[3,3,1,""],fig:[3,2,1,""],get_annotations_text:[3,3,1,""],get_configuration:[3,3,1,""],get_offset_func:[3,3,1,""],has_type0_comparisons_correction:[3,3,1,""],loc:[3,2,1,""],new_plot:[3,3,1,""],orient:[3,2,1,""],print_pvalue_legend:[3,3,1,""],pvalue_format:[3,2,1,""],reset_configuration:[3,3,1,""],set_custom_annotations:[3,3,1,""],set_pvalues:[3,3,1,""],set_pvalues_and_annotate:[3,3,1,""],test:[3,2,1,""],validate_test_short_name:[3,3,1,""],verbose:[3,2,1,""]},"statannotations.PValueFormat":{Formatter:[3,1,1,""],PValueFormat:[3,1,1,""],sort_pvalue_thresholds:[3,4,1,""]},"statannotations.PValueFormat.Formatter":{config:[3,3,1,""],format_data:[3,3,1,""]},"statannotations.PValueFormat.PValueFormat":{config:[3,3,1,""],format_data:[3,3,1,""],get_configuration:[3,3,1,""],print_legend_if_used:[3,3,1,""],pvalue_format_string:[3,2,1,""],pvalue_thresholds:[3,2,1,""],simple_format_string:[3,2,1,""],text_format:[3,2,1,""]},"statannotations.format_annotations":{pval_annotation_text:[3,4,1,""],simple_text:[3,4,1,""]},"statannotations.stats":{ComparisonsCorrection:[4,0,0,"-"],StatResult:[4,0,0,"-"],StatTest:[4,0,0,"-"],test:[4,0,0,"-"],utils:[4,0,0,"-"]},"statannotations.stats.ComparisonsCorrection":{ComparisonsCorrection:[4,1,1,""],check_valid_correction_name:[4,4,1,""],get_correction_parameters:[4,4,1,""],get_validated_comparisons_correction:[4,4,1,""]},"statannotations.stats.ComparisonsCorrection.ComparisonsCorrection":{apply:[4,3,1,""],document:[4,3,1,""]},"statannotations.stats.StatResult":{StatResult:[4,1,1,""]},"statannotations.stats.StatResult.StatResult":{corrected_significance:[4,2,1,""],correction_method:[4,2,1,""],formatted_output:[4,2,1,""],significance_suffix:[4,2,1,""]},"statannotations.stats.StatTest":{StatTest:[4,1,1,""],wilcoxon:[4,4,1,""]},"statannotations.stats.StatTest.StatTest":{from_library:[4,3,1,""],short_name:[4,2,1,""]},"statannotations.stats.test":{apply_test:[4,4,1,""]},"statannotations.stats.utils":{check_alpha:[4,4,1,""],check_num_comparisons:[4,4,1,""],check_pvalues:[4,4,1,""],get_num_comparisons:[4,4,1,""],return_results:[4,4,1,""]},"statannotations.utils":{check_is_in:[3,4,1,""],check_not_none:[3,4,1,""],check_order_in_data:[3,4,1,""],check_pairs_in_data:[3,4,1,""],check_valid_text_format:[3,4,1,""],get_closest:[3,4,1,""],get_x_values:[3,4,1,""],raise_expected_got:[3,4,1,""],remove_null:[3,4,1,""],render_collection:[3,4,1,""]},statannotations:{Annotation:[3,0,0,"-"],Annotator:[3,0,0,"-"],PValueFormat:[3,0,0,"-"],format_annotations:[3,0,0,"-"],stats:[4,0,0,"-"],utils:[3,0,0,"-"]}},objnames:{"0":["py","module","Python module"],"1":["py","class","Python class"],"2":["py","property","Python property"],"3":["py","method","Python method"],"4":["py","function","Python function"]},objtypes:{"0":"py:module","1":"py:class","2":"py:property","3":"py:method","4":"py:function"},terms:{"0":[3,4],"05":[3,4],"1":[3,4],"12141511":3,"9981846":3,"case":4,"class":[3,4],"default":[3,4],"float":4,"function":[3,4],"int":4,"return":3,"short":3,"static":4,"true":[3,4],A:3,If:3,Or:3,The:3,a_list:3,accept:3,add:3,addit:4,alpha:[3,4],an:3,annot:1,annotate_custom_annot:3,appli:[3,4],apply_and_annot:3,apply_test:[3,4],ar:3,arg:[3,4],argument:[3,4],arrai:3,assum:3,auto:3,ax:3,bar:3,base:[3,4],behavior:4,benjamini:4,between:3,bonferroni:4,bool:4,box:3,boxplot:3,calcul:3,callabl:4,check:3,check_alpha:4,check_is_in:3,check_not_non:3,check_num_comparison:4,check_order_in_data:3,check_pairs_in_data:3,check_pvalu:4,check_valid_correction_nam:4,check_valid_text_format:3,classmethod:3,close:3,closest:3,collect:3,color:3,com:3,comparison:[3,4],comparisons_correct:[3,4],comparisonscorrect:[1,3],comput:3,config:3,configur:3,constructor:3,content:1,coord:3,coordin:3,core:3,corr_kwarg:4,correct:4,corrected_signific:4,correction_method:4,custom:3,data:[3,4],datafram:3,dict:4,displai:3,document:4,each:3,earlier:4,either:4,engin:3,equal:3,equival:4,error:3,error_typ:3,exact:3,exist:3,expect:3,fals:3,fig:3,follow:3,for_:3,format:[3,4],format_annot:1,format_data:3,formatt:3,formatted_output:[3,4],fraction:3,frame:3,from:[3,4],from_librari:4,func:4,gener:3,get:4,get_annotations_text:3,get_closest:3,get_configur:3,get_correction_paramet:4,get_num_comparison:4,get_offset_func:3,get_validated_comparisons_correct:4,get_x_valu:3,got:3,group:3,group_data1:4,group_data2:4,gt:4,has_type0_comparisons_correct:3,hochberg:4,hold:3,holm:4,http:3,hue:3,hue_ord:3,index:0,instanc:[3,4],instead:3,interfac:4,interpret:4,keyword:4,kruskal:4,kwarg:[3,4],label:3,leven:4,line_height:3,line_offset:3,line_offset_to_group:3,line_width:3,link:3,list:3,loc:3,ls:4,mann:4,messag:3,method:4,method_typ:4,mode:3,modul:[0,1],multipl:4,must:3,mylist:3,mynumb:3,name:[3,4],new_plot:3,none:[3,4],num_comparison:[3,4],number:[3,4],object:[3,4],one:[3,4],option:[3,4],order:3,orient:3,otherwis:3,overrid:3,p:3,p_valu:4,packag:[0,1],page:0,pair:3,panda:3,paramet:[3,4],pass:[3,4],per:3,perform:3,plot:3,plot_param:3,point:3,posit:3,print_labels_and_cont:3,print_legend_if_us:3,print_pvalue_legend:3,properti:[3,4],provid:[3,4],pval:4,pval_annotation_text:3,pvalu:[3,4],pvalue_format:3,pvalue_format_str:3,pvalue_threshold:3,pvalueformat:1,rais:3,raise_expected_got:3,rang:3,refer:3,remove_nul:3,render:3,render_collect:3,reset_configur:3,result:[3,4],results_arrai:4,return_result:4,run:4,same:3,sampl:4,seaborn:3,search:0,sep:3,seri:3,set:3,set_custom_annot:3,set_pvalu:3,set_pvalues_and_annot:3,short_nam:4,show:3,show_test_nam:3,significance_suffix:4,simpl:3,simple_format_str:3,simple_text:3,smallest:3,sort:3,sort_pvalue_threshold:3,stackoverflow:3,standard:3,stat:[1,3],stat_nam:4,stat_str:4,statist:[3,4],statresult:[1,3],stats_param:[3,4],statsmodel:4,statsmodels_api:4,stattest:[1,3],str:[3,4],string:3,struct:3,submodul:1,subpackag:1,t:4,test:[1,3],test_descript:4,test_ind:4,test_long_nam:4,test_nam:4,test_pair:4,test_result_list:4,test_short_nam:[3,4],test_welch:4,text:3,text_annot_custom:3,text_format:3,text_offset:3,them:3,thereof:3,thi:[3,4],three:3,threshold:3,top:3,tupl:3,two:[3,4],type:3,union:[3,4],us:[3,4],use_fixed_offset:3,util:1,valid_valu:3,validate_test_short_nam:3,valu:[3,4],valueerror:3,verbos:[3,4],version:4,vs:3,wa:3,whitnei:4,wilcoxon:4,work:3,x:3,y:3,yekuti:4},titles:["Welcome to statannotations\u2019s documentation!","statannotations","setup module","statannotations package","statannotations.stats package"],titleterms:{annot:3,comparisonscorrect:4,content:[0,3,4],document:0,format_annot:3,indic:0,modul:[2,3,4],packag:[3,4],pvalueformat:3,s:0,setup:2,stat:4,statannot:[0,1,3,4],statresult:4,stattest:4,submodul:[3,4],subpackag:3,tabl:0,test:4,util:[3,4],welcom:0}}) \ No newline at end of file diff --git a/docs/build/html/statannotations.html b/docs/build/html/statannotations.html index aaacef6..7363f25 100644 --- a/docs/build/html/statannotations.html +++ b/docs/build/html/statannotations.html @@ -352,6 +352,11 @@ <h2>Submodules<a class="headerlink" href="#submodules" title="Permalink to this <span class="sig-name descname"><span class="pre">new_plot</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">ax</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pairs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">plot</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'boxplot'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">data</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">y</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">hue</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">order</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">hue_order</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">engine</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">str</span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">'seaborn'</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">plot_params</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#statannotations.Annotator.Annotator.new_plot" title="Permalink to this definition">¶</a></dt> <dd></dd></dl> +<dl class="py property"> +<dt class="sig sig-object py" id="statannotations.Annotator.Annotator.orient"> +<em class="property"><span class="pre">property</span> </em><span class="sig-name descname"><span class="pre">orient</span></span><a class="headerlink" href="#statannotations.Annotator.Annotator.orient" title="Permalink to this definition">¶</a></dt> +<dd></dd></dl> + <dl class="py method"> <dt class="sig sig-object py" id="statannotations.Annotator.Annotator.print_pvalue_legend"> <span class="sig-name descname"><span class="pre">print_pvalue_legend</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="headerlink" href="#statannotations.Annotator.Annotator.print_pvalue_legend" title="Permalink to this definition">¶</a></dt> @@ -544,7 +549,7 @@ <h2>Submodules<a class="headerlink" href="#submodules" title="Permalink to this <dl class="py function"> <dt class="sig sig-object py" id="statannotations.utils.check_pairs_in_data"> -<span class="sig-prename descclassname"><span class="pre">statannotations.utils.</span></span><span class="sig-name descname"><span class="pre">check_pairs_in_data</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">pairs</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">list</span><span class="p"><span class="pre">,</span> </span><span class="pre">tuple</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">data</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">list</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span> </span><span class="pre">pandas.core.frame.DataFrame</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span> </span><span class="pre">list</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">hue</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">hue_order</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#statannotations.utils.check_pairs_in_data" title="Permalink to this definition">¶</a></dt> +<span class="sig-prename descclassname"><span class="pre">statannotations.utils.</span></span><span class="sig-name descname"><span class="pre">check_pairs_in_data</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">pairs</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">list</span><span class="p"><span class="pre">,</span> </span><span class="pre">tuple</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">data</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">list</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span> </span><span class="pre">pandas.core.frame.DataFrame</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">coord</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span> </span><span class="pre">list</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">hue</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">hue_order</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#statannotations.utils.check_pairs_in_data" title="Permalink to this definition">¶</a></dt> <dd><p>Checks that values referred to in <cite>order</cite> and <cite>pairs</cite> exist in data.</p> </dd></dl> diff --git a/statannotations/Annotator.py b/statannotations/Annotator.py index 390bd83..00dcee6 100644 --- a/statannotations/Annotator.py +++ b/statannotations/Annotator.py @@ -123,7 +123,7 @@ def __init__(self, ax, pairs, plot='boxplot', data=None, x=None, self.text_offset_impact_above = 0 self.color = '0.2' self.line_width = 1.5 - self.y_offset = None + self.value_offset = None self.custom_annotations = None def new_plot(self, ax, pairs=None, plot='boxplot', data=None, x=None, @@ -144,6 +144,10 @@ def new_plot(self, ax, pairs=None, plot='boxplot', data=None, x=None, return self + @property + def orient(self): + return self._plotter.orient + @property def verbose(self): return self._verbose @@ -154,8 +158,8 @@ def verbose(self, verbose): self._plotter.verbose = verbose @property - def _y_stack_arr(self): - return self._plotter.y_stack_arr + def _value_stack_arr(self): + return self._plotter.value_stack_arr @property def fig(self): @@ -179,13 +183,13 @@ def annotate(self, line_offset=None, line_offset_to_group=None): "test (again) which will probably lead to " "unexpected results") - self._update_y_for_loc() + self._update_value_for_loc() ann_list = [] - orig_ylim = self.ax.get_ylim() + orig_value_lim = self._plotter.get_value_lim() offset_func = self.get_offset_func(self.loc) - self.y_offset, self.line_offset_to_group = offset_func( + self.value_offset, self.line_offset_to_group = offset_func( line_offset, line_offset_to_group) if self._verbose: @@ -199,16 +203,24 @@ def annotate(self, line_offset=None, line_offset_to_group=None): self._annotate_pair(annotation, ax_to_data=ax_to_data, ann_list=ann_list, - orig_ylim=orig_ylim) + orig_value_lim=orig_value_lim) # reset transformation - y_stack_max = max(self._y_stack_arr[1, :]) + y_stack_max = max(self._value_stack_arr[1, :]) ax_to_data = self._plotter.get_transform_func('ax_to_data') - ylims = ([(0, 0), (0, max(1.04 y_stack_max, 1))] - if self.loc == 'inside' - else [(0, 0), (0, 1)]) - - self.ax.set_ylim(ax_to_data.transform(ylims)[:, 1]) + value_lims = ( + ([(0, 0), (0, max(1.04 * y_stack_max, 1))] + if self.loc == 'inside' + else [(0, 0), (0, 1)]) + if self.orient == 'v' + else + ([(0, 0), (max(1.04 * y_stack_max, 1), 0)] + if self.loc == 'inside' + else [(0, 0), (1, 0)]) + ) + set_lims = self.ax.set_ylim if self.orient == 'v' else self.ax.set_xlim + transformed = ax_to_data.transform(value_lims) + set_lims(transformed[:, 1 if self.orient == 'v' else 0]) return self.ax, self.annotations @@ -444,61 +456,75 @@ def _get_results(self, num_comparisons, pvalues=None, return test_result_list - def _annotate_pair(self, annotation, ax_to_data, ann_list, orig_ylim): + def _annotate_pair(self, annotation, ax_to_data, ann_list, orig_value_lim): if self._verbose >= 1: annotation.print_labels_and_content() - x1 = annotation.structs[0]['x'] - x2 = annotation.structs[1]['x'] - xi1 = annotation.structs[0]['xi'] - xi2 = annotation.structs[1]['xi'] + group_coord_1 = annotation.structs[0]['group_coord'] + group_coord_2 = annotation.structs[1]['group_coord'] + group_i1 = annotation.structs[0]['group_i'] + group_i2 = annotation.structs[1]['group_i'] # Find y maximum for all the y_stacks in between group1 and group2 - i_ymax_in_range_x1_x2 = xi1 + np.nanargmax( - self._y_stack_arr[1, np.where((x1 <= self._y_stack_arr[0, :]) - & (self._y_stack_arr[0, :] <= x2))]) + i_value_max_in_range_g1_g2 = group_i1 + np.nanargmax( + self._value_stack_arr[1, + np.where((group_coord_1 + <= self._value_stack_arr[0, :]) + & (self._value_stack_arr[0, :] + <= group_coord_2))]) - y = self._get_y_for_pair(i_ymax_in_range_x1_x2) + value = self._get_value_for_pair(i_value_max_in_range_g1_g2) # Determine lines in axes coordinates - ax_line_x = [x1, x1, x2, x2] - ax_line_y = [y, y + self.line_height, y + self.line_height, y] + ax_line_group = [group_coord_1, group_coord_1, + group_coord_2, group_coord_2] + ax_line_value = [value, value + self.line_height, + value + self.line_height, value] + + lists = ((ax_line_group, ax_line_value) if self.orient == 'v' + else (ax_line_value, ax_line_group)) points = [ax_to_data.transform((x, y)) for x, y - in zip(ax_line_x, ax_line_y)] + in zip(lists)] line_x, line_y = zip(points) self._plot_line(line_x, line_y) + xy_params = self._get_xy_params(group_coord_1, group_coord_2, line_x, + line_y) + ann = self.ax.annotate( - annotation.text, xy=(np.mean([x1, x2]), line_y[2]), - xytext=(0, self.text_offset), textcoords='offset points', + annotation.text, textcoords='offset points', xycoords='data', ha='center', va='bottom', fontsize=self._pvalue_format.fontsize, clip_on=False, - annotation_clip=False) + annotation_clip=False, *xy_params) + if annotation.text is not None: ann_list.append(ann) plt.draw() - self.ax.set_ylim(orig_ylim) + set_lim = {'v': 'set_ylim', + 'h': 'set_xlim'}[self.orient] + + getattr(self.ax, set_lim)(orig_value_lim) - y_top_annot = self._annotate_pair_text(ann, y) + value_top_annot = self._annotate_pair_text(ann, value) else: - y_top_annot = y + self.line_height + value_top_annot = value + self.line_height - # Fill the highest y position of the annotation into the y_stack array - # for all positions in the range x1 to x2 - self._y_stack_arr[1, - (x1 <= self._y_stack_arr[0, :]) - & (self._y_stack_arr[0, :] <= x2)] = y_top_annot + # Fill the highest value position of the annotation into the value_stack + # array for all positions in the range group_coord_1 to group_coord_2 + self._value_stack_arr[ + 1, (group_coord_1 <= self._value_stack_arr[0, :]) + & (self._value_stack_arr[0, :] <= group_coord_2)] = value_top_annot - # Increment the counter of annotations in the y_stack array - self._y_stack_arr[2, xi1:xi2 + 1] += 1 + # Increment the counter of annotations in the value_stack array + self._value_stack_arr[2, group_i1:group_i2 + 1] += 1 - def _update_y_for_loc(self): + def _update_value_for_loc(self): if self._loc == 'outside': - self._y_stack_arr[1, :] = 1 + self._value_stack_arr[1, :] = 1 def _check_test_pvalues_perform(self): if self.test is None: @@ -582,8 +608,7 @@ def _get_offsets_inside(line_offset, line_offset_to_group): else: if line_offset_to_group is None: line_offset_to_group = 0.06 - y_offset = line_offset - return y_offset, line_offset_to_group + return line_offset, line_offset_to_group @staticmethod def _get_offsets_outside(line_offset, line_offset_to_group): @@ -623,9 +648,9 @@ def _plot_line(self, line_x, line_y): line.set_clip_on(False) self.ax.add_line(line) - def _annotate_pair_text(self, ann, y): + def _annotate_pair_text(self, ann, value): - y_top_annot = None + value_top_annot = None got_mpl_error = False if not self.use_fixed_offset: @@ -633,7 +658,8 @@ def _annotate_pair_text(self, ann, y): bbox = ann.get_window_extent() pix_to_ax = self._plotter.get_transform_func('pix_to_ax') bbox_ax = bbox.transformed(pix_to_ax) - y_top_annot = bbox_ax.ymax + value_coord_max = {'v': 'ymax', 'h': 'xmax'}[self.orient] + value_top_annot = getattr(bbox_ax, value_coord_max) except RuntimeError: got_mpl_error = True @@ -644,40 +670,45 @@ def _annotate_pair_text(self, ann, y): "back to a fixed y offset. Layout may be not " "optimal.") - # We will apply a fixed offset in points, - # based on the font size of the annotation. fontsize_points = FontProperties( size='medium').get_size_in_points() - offset_trans = mtransforms.offset_copy( - self.ax.transAxes, fig=self.fig, x=0, - y=fontsize_points + self.text_offset, units='points') - y_top_display = offset_trans.transform( - (0, y + self.line_height)) - y_top_annot = (self.ax.transAxes.inverted() - .transform(y_top_display)[1]) + direction = {'h': -1, 'v': 1}[self.orient] + x, y = [0, fontsize_points + self.text_offset][::direction] + offset_trans = mtransforms.offset_copy(trans=self.ax.transAxes, + fig=self.fig, + units='points', x=x, y=y) + + value_top_display = offset_trans.transform( + (value + self.line_height, value + self.line_height)) + + value_coord = {'h': 0, 'v': 1}[self.orient] + + value_top_annot = (self.ax.transAxes.inverted() + .transform(value_top_display)[value_coord]) self.text_offset_impact_above = ( - y_top_annot - y - self.y_offset - self.line_height) - return y_top_annot + value_top_annot - value - self.value_offset - self.line_height) + + return value_top_annot def _reset_default_values(self): for attribute, default_value in _DEFAULT_VALUES.items(): setattr(self, attribute, default_value) - def _get_y_for_pair(self, i_ymax_in_range_x1_x2): + def _get_value_for_pair(self, i_ymax_in_range_x1_x2): - ymax_in_range_x1_x2 = self._y_stack_arr[1, i_ymax_in_range_x1_x2] + ymax_in_range_x1_x2 = self._value_stack_arr[1, i_ymax_in_range_x1_x2] # Choose the best offset depending on whether there is an annotation # below at the x position in the range [x1, x2] where the stack is the # highest - if self._y_stack_arr[2, i_ymax_in_range_x1_x2] == 0: + if self._value_stack_arr[2, i_ymax_in_range_x1_x2] == 0: # there is only a group below offset = self.line_offset_to_group else: # there is an annotation below - offset = self.y_offset + self.text_offset_impact_above + offset = self.value_offset + self.text_offset_impact_above return ymax_in_range_x1_x2 + offset @@ -697,3 +728,28 @@ def _get_plotter(engine, args, *kwargs): if engine_plotter is None: raise NotImplementedError(f"{engine} engine not implemented.") return engine_plotter(args, kwargs) + + def _get_xy_params_horizontal(self, group_coord_1, group_coord_2, + line_x: np.ndarray): + return { + 'xy': (line_x[2], np.mean([group_coord_1, group_coord_2])), + 'xytext': (self.text_offset, 0), + 'rotation': 270, + 'rotation_mode': 'anchor' + } + + def _get_xy_params_vertical(self, group_coord_1, group_coord_2, + line_y: np.ndarray): + return { + 'xy': (np.mean([group_coord_1, group_coord_2]), line_y[2]), + 'xytext': (0, self.text_offset), + } + + def _get_xy_params(self, group_coord_1, group_coord_2, line_x: np.ndarray, + line_y: np.ndarray): + if self.orient == 'h': + return self._get_xy_params_horizontal(group_coord_1, group_coord_2, + line_x) + + return self._get_xy_params_vertical(group_coord_1, group_coord_2, + line_y) diff --git a/statannotations/_Xpositions.py b/statannotations/_GroupsPositions.py similarity index 51% rename from statannotations/_Xpositions.py rename to statannotations/_GroupsPositions.py index 2520f3c..046d2b9 100644 --- a/statannotations/_Xpositions.py +++ b/statannotations/_GroupsPositions.py @@ -3,7 +3,7 @@ from statannotations.utils import get_closest -class _XPositions: +class _GroupsPositions: def __init__(self, plotter, group_names): self._plotter = plotter self._hue_names = self._plotter.hue_names @@ -15,40 +15,42 @@ def __init__(self, plotter, group_names): "Using hues with only one hue is not supported.") self.hue_offsets = self._plotter.hue_offsets - self._xunits = self.hue_offsets[1] - self.hue_offsets[0] + self._axis_units = self.hue_offsets[1] - self.hue_offsets[0] - self._xpositions = { - np.round(self.get_group_x_position(group_name), 1): group_name + self._groups_positions = { + np.round(self.get_group_axis_position(group_name), 1): group_name for group_name in group_names } - self._xpositions_list = sorted(self._xpositions.keys()) + self._groups_positions_list = sorted(self._groups_positions.keys()) if self._hue_names is None: - self._xunits = ((max(list(self._xpositions.keys())) + 1) - / len(self._xpositions)) + self._axis_units = ((max(list(self._groups_positions.keys())) + 1) + / len(self._groups_positions)) - self._xranges = { - (pos - self._xunits / 2, pos + self._xunits / 2, pos): group_name - for pos, group_name in self._xpositions.items()} + self._axis_ranges = { + (pos - self._axis_units / 2, + pos + self._axis_units / 2, + pos): group_name + for pos, group_name in self._groups_positions.items()} @property - def xpositions(self): - return self._xpositions + def axis_positions(self): + return self._groups_positions @property - def xunits(self): - return self._xunits + def axis_units(self): + return self._axis_units - def get_xpos_location(self, pos): + def get_axis_pos_location(self, pos): """ Finds the x-axis location of a categorical variable """ - for xrange in self._xranges: - if (pos >= xrange[0]) & (pos <= xrange[1]): - return xrange[2] + for axis_range in self._axis_ranges: + if (pos >= axis_range[0]) & (pos <= axis_range[1]): + return axis_range[2] - def get_group_x_position(self, group): + def get_group_axis_position(self, group): """ group_name can be either a name "cat" or a tuple ("cat", "hue") """ @@ -64,5 +66,5 @@ def get_group_x_position(self, group): group_pos = self._plotter.group_names.index(cat) + hue_offset return group_pos - def find_closest(self, xpos): - return get_closest(list(self._xpositions_list), xpos) + def find_closest(self, pos): + return get_closest(list(self._groups_positions_list), pos) diff --git a/statannotations/_Plotter.py b/statannotations/_Plotter.py index 9dac681..5223865 100644 --- a/statannotations/_Plotter.py +++ b/statannotations/_Plotter.py @@ -8,7 +8,7 @@ from matplotlib.collections import PathCollection from matplotlib.patches import Rectangle -from statannotations._Xpositions import _XPositions +from statannotations._GroupsPositions import _GroupsPositions from statannotations.utils import check_not_none, check_order_in_data, \ check_pairs_in_data, render_collection, check_is_in, remove_null @@ -18,16 +18,18 @@ class _Plotter: - def __init__(self, ax, pairs, data=None, x=None, hue=None, - order=None, hue_order=None, verbose=False): + def __init__(self, ax, pairs, data=None, x=None, y=None, hue=None, + order=None, hue_order=None, verbose=False, plot_params): self.ax = ax self._fig = plt.gcf() check_not_none("pairs", pairs) - check_order_in_data(data, x, order) - check_pairs_in_data(pairs, data, x, hue, hue_order) + group_coord = y if plot_params.get("orient") == "h" else x + check_order_in_data(data, group_coord, order) + check_pairs_in_data(pairs, data, group_coord, hue, hue_order) self.pairs = pairs self._struct_pairs = None self.verbose = verbose + self.orient = plot_params.get("orient", "v") def get_transform_func(self, kind: str): """ @@ -48,9 +50,11 @@ def get_transform_func(self, kind: str): if kind == 'pix_to_ax': return self.ax.transAxes.inverted() - data_to_ax = \ - self.ax.transData + self.ax.get_xaxis_transform().inverted() + transform = {'v': self.ax.get_xaxis_transform, + 'h': self.ax.get_yaxis_transform}[self.orient] + data_to_ax = \ + self.ax.transData + transform().inverted() if kind == 'data_to_ax': return data_to_ax @@ -75,34 +79,35 @@ def __init__(self, ax, pairs, plot='boxplot', data=None, x=None, y=None, hue=None, order=None, hue_order=None, verbose=False, plot_params): - _Plotter.__init__(self, ax, pairs, data, x, hue, order, hue_order, - verbose) + _Plotter.__init__(self, ax, pairs, data, x, y, hue, order, hue_order, + verbose, plot_params) self.check_plot_is_implemented(plot) - self.plot = plot self.plotter = self._get_plotter(plot, x, y, hue, data, order, hue_order, *plot_params) self.group_names, self.labels = self._get_group_names_and_labels() - self.xpositions = _XPositions(self.plotter, self.group_names) + self.groups_positions = _GroupsPositions(self.plotter, + self.group_names) self.reordering = None - self.ymaxes = self._generate_ymaxes() + self.value_maxes = self._generate_value_maxes() self.structs = self._get_structs() self.pairs = pairs self._struct_pairs = self._get_group_struct_pairs() - self._y_stack_arr = np.array( - [[struct['x'], struct['ymax'], 0] for struct in self.structs] + self._value_stack_arr = np.array( + [[struct['group_coord'], struct['value_max'], 0] + for struct in self.structs] ).T self.reordering = None self._sort_group_struct_pairs() @property - def y_stack_arr(self): - return self._y_stack_arr + def value_stack_arr(self): + return self._value_stack_arr # noinspection PyProtectedMember def _get_plotter(self, plot, x, y, hue, data, order, hue_order, @@ -187,7 +192,7 @@ def _get_group_names_and_labels(self): return group_names, labels - def _generate_ymaxes(self): + def _generate_value_maxes(self): """ given plotter and the names of two categorical variables, returns highest y point drawn between those two variables before @@ -197,38 +202,40 @@ def _generate_ymaxes(self): (eg, error bars and/or bar charts). """ - ymaxes = {name: 0 for name in self.group_names} + value_maxes = {name: 0 for name in self.group_names} data_to_ax = self.get_transform_func('data_to_ax') if self.plot == 'violinplot': - ymaxes = self._get_ymaxes_violin(ymaxes, data_to_ax) + value_maxes = self._get_value_maxes_violin(value_maxes, data_to_ax) else: for child in self.ax.get_children(): - xname, ypos = self._get_child_ypos(child, data_to_ax) + group_name, value_pos = self._get_value_pos(child, data_to_ax) - if ypos is not None and ypos > ymaxes[xname]: - ymaxes[xname] = ypos + if value_pos is not None and value_pos > value_maxes[group_name]: + value_maxes[group_name] = value_pos - return ymaxes + return value_maxes def _get_structs(self): structs = [ { 'group': group_name, 'label': self.labels[b_idx], - 'x': self.xpositions.get_group_x_position(group_name), + 'group_coord': (self.groups_positions + .get_group_axis_position(group_name)), 'group_data': self._get_group_data(group_name), - 'ymax': self.ymaxes[group_name] - } for b_idx, group_name in enumerate(self.group_names)] + 'value_max': self.value_maxes[group_name] + } + for b_idx, group_name in enumerate(self.group_names)] - # Sort the group data structures by position along the x axis - structs = sorted(structs, key=lambda struct: struct['x']) + # Sort the group data structures by position along the groups axis + structs = sorted(structs, key=lambda struct: struct['group_coord']) - # Add the index position in the list of groups along the x axis - structs = [dict(struct, xi=i) + # Add the index position in the list of groups along the groups axis + structs = [dict(struct, group_i=i) for i, struct in enumerate(structs)] return structs @@ -246,7 +253,7 @@ def _get_group_struct_pairs(self): group_struct2 = dict(group_structs_dic[group2], i_group_pair=i_group_pair) - if group_struct1['x'] <= group_struct2['x']: + if group_struct1['group_coord'] <= group_struct2['group_coord']: group_struct_pairs.append((group_struct1, group_struct2)) else: @@ -289,62 +296,79 @@ def _get_group_data(self, group_name): return group_data - def _get_ymaxes_violin(self, ymaxes, data_to_ax): + def _get_value_maxes_violin(self, value_maxes, data_to_ax): for group_idx, group_name in enumerate(self.plotter.group_names): if self.plotter.hue_names: for hue_idx, hue_name in enumerate(self.plotter.hue_names): - ypos = max(self.plotter.support[group_idx][hue_idx]) - ymaxes[(group_name, hue_name)] = \ - data_to_ax.transform((0, ypos))[1] + value_pos = max(self.plotter.support[group_idx][hue_idx]) + value_maxes[(group_name, hue_name)] = \ + data_to_ax.transform((0, value_pos))[1] else: - ypos = max(self.plotter.support[group_idx]) - ymaxes[group_name] = data_to_ax.transform((0, ypos))[1] - return ymaxes + value_pos = max(self.plotter.support[group_idx]) + value_maxes[group_name] = data_to_ax.transform((0, + value_pos))[1] + return value_maxes - def _get_child_ypos(self, child, data_to_ax): + def _get_value_pos(self, child, data_to_ax): if (type(child) == PathCollection and len(child.properties()['offsets'])): - return self._get_ypos_for_path_collection( + return self._get_value_pos_for_path_collection( child, data_to_ax) elif type(child) in (lines.Line2D, Rectangle): - return self._get_ypos_for_line2d_or_rectangle( + return self._get_value_pos_for_line2d_or_rectangle( child, data_to_ax) return None, None - def _get_ypos_for_path_collection(self, child, data_to_ax): - ymax = child.properties()['offsets'][:, 1].max() - xpos = float(np.round(np.nanmean( - child.properties()['offsets'][:, 0]), 1)) - if xpos not in self.xpositions.xpositions: + def _get_value_pos_for_path_collection(self, child, data_to_ax): + group_coord = {"v": 0, "h": 1}[self.orient] + value_coord = (group_coord + 1) % 2 + + direction = {"v": 1, "h": -1}[self.orient] + + value_max = child.properties()['offsets'][:, value_coord].max() + group_pos = float(np.round(np.nanmean( + child.properties()['offsets'][:, group_coord]), 1)) + + if group_pos not in self.groups_positions.axis_positions: if self.verbose: warnings.warn( "Invalid x-position found. Are the same parameters passed " "to seaborn and statannotations calls? or are there few " "data points?") - xpos = self.xpositions.find_closest(xpos) + group_pos = self.groups_positions.find_closest(group_pos) + + group_name = self.groups_positions.axis_positions[group_pos] - xname = self.xpositions.xpositions[xpos] - ypos = data_to_ax.transform((0, ymax))[1] + value_pos = data_to_ax.transform((0, value_max)[::direction]) - return xname, ypos + return group_name, value_pos[value_coord] - def _get_ypos_for_line2d_or_rectangle(self, child, data_to_ax): + def _get_value_pos_for_line2d_or_rectangle(self, child, data_to_ax): bbox = self.ax.transData.inverted().transform( child.get_window_extent(self.fig.canvas.get_renderer())) - if (bbox[:, 0].max() - bbox[:, 0].min()) > 1.1self.xpositions.xunits: + group_coord = {"v": 0, "h": 1}[self.orient] + direction = {"v": 1, "h": -1}[self.orient] + + value_coord = (group_coord + 1) % 2 + + if ((bbox[:, group_coord].max() - bbox[:, group_coord].min()) + > 1.1 * self.groups_positions.axis_units): return None, None - raw_xpos = np.round(bbox[:, 0].mean(), 1) - xpos = self.xpositions.get_xpos_location(raw_xpos) - if xpos not in self.xpositions.xpositions: + + raw_group_pos = np.round(bbox[:, group_coord].mean(), 1) + group_pos = self.groups_positions.get_axis_pos_location(raw_group_pos) + + if group_pos not in self.groups_positions.axis_positions: return None, None - xname = self.xpositions.xpositions[xpos] - ypos = bbox[:, 1].max() - ypos = data_to_ax.transform((0, ypos))[1] - return xname, ypos + group_name = self.groups_positions.axis_positions[group_pos] + value_pos = bbox[:, value_coord].max() + + value_pos = data_to_ax.transform((0, value_pos)[::direction]) + return group_name, value_pos[value_coord] def _sort_group_struct_pairs(self): # Draw first the annotations with the shortest between-groups distance, @@ -357,7 +381,8 @@ def _sort_group_struct_pairs(self): @staticmethod def _absolute_group_struct_pair_in_tuple_x_diff(group_struct_pair): - return abs(group_struct_pair[1][1]['x'] - group_struct_pair[1][0]['x']) + return abs(group_struct_pair[1][1]['group_coord'] + - group_struct_pair[1][0]['group_coord']) @staticmethod def check_plot_is_implemented(plot, engine="seaborn"): @@ -377,3 +402,8 @@ def fix_and_warn(dodge, hue, plot): "Implicitly setting dodge to True as it is necessary in " "statannotations. It must have been True for the seaborn " "call to yield consistent results when using `hue`.") + + def get_value_lim(self): + if self.orient == 'v': + return self.ax.get_ylim() + return self.ax.get_xlim() diff --git a/statannotations/_version.py b/statannotations/_version.py index 6a9beea..3d26edf 100644 --- a/statannotations/_version.py +++ b/statannotations/_version.py @@ -1,1 +1,1 @@ -__version__ = "0.4.0" +__version__ = "0.4.1" diff --git a/statannotations/utils.py b/statannotations/utils.py index c7d44fc..934239b 100644 --- a/statannotations/utils.py +++ b/statannotations/utils.py @@ -75,19 +75,19 @@ def _check_pairs_in_data_no_hue(pairs: Union[list, tuple], def _check_pairs_in_data_with_hue(pairs: Union[list, tuple], data: Union[List[list], pd.DataFrame] = None, - x: Union[str, list] = None, + group_coord: Union[str, list] = None, hue: str = None) -> set: - x_values = get_x_values(data, x) - seen_x_values = set() + x_values = get_x_values(data, group_coord) + seen_group_values = set() hue_values = set(data[hue].unique()) - for x_value, hue_value in itertools.chain(itertools.chain(pairs)): - if x_value not in seen_x_values and x_value not in x_values: - raise ValueError(f"Missing x value `{x_value}` in {x}" + for group_value, hue_value in itertools.chain(itertools.chain(pairs)): + if group_value not in seen_group_values and group_value not in x_values: + raise ValueError(f"Missing group value `{group_value}` in {group_coord}" f" (specified in `pairs`)") - seen_x_values.add(x_value) + seen_group_values.add(group_value) if hue_value not in hue_values: raise ValueError(f"Missing hue value `{hue_value}` in {hue}" @@ -108,7 +108,7 @@ def _check_hue_order_in_data(hue, hue_values: set, def check_pairs_in_data(pairs: Union[list, tuple], data: Union[List[list], pd.DataFrame] = None, - x: Union[str, list] = None, + coord: Union[str, list] = None, hue: str = None, hue_order: List[str] = None): """ @@ -116,9 +116,9 @@ def check_pairs_in_data(pairs: Union[list, tuple], """ if hue is None and hue_order is None: - _check_pairs_in_data_no_hue(pairs, data, x) + _check_pairs_in_data_no_hue(pairs, data, coord) else: - hue_values = _check_pairs_in_data_with_hue(pairs, data, x, hue) + hue_values = _check_pairs_in_data_with_hue(pairs, data, coord, hue) _check_hue_order_in_data(hue, hue_values, hue_order) diff --git a/usage/example.ipynb b/usage/example.ipynb index 10b91f9..91a3e66 100644 --- a/usage/example.ipynb +++ b/usage/example.ipynb @@ -231,7 +231,7 @@ }, { "data": { - "text/plain": "(<matplotlib.axes._subplots.AxesSubplot at 0x7fedb54c7748>,\n [<statannotations.Annotation.Annotation at 0x7fedb4233ac8>,\n <statannotations.Annotation.Annotation at 0x7fedb4233d30>,\n <statannotations.Annotation.Annotation at 0x7fedb4233320>])" + "text/plain": "(<matplotlib.axes._subplots.AxesSubplot at 0x7fc28184f6a0>,\n [<statannotations.Annotation.Annotation at 0x7fc28164aeb8>,\n <statannotations.Annotation.Annotation at 0x7fc28183d6d8>,\n <statannotations.Annotation.Annotation at 0x7fc282ada5c0>])" }, "execution_count": 6, "metadata": {}, @@ -672,7 +672,7 @@ }, { "data": { - "text/plain": "(<matplotlib.axes._subplots.AxesSubplot at 0x7fedb4460898>,\n [<statannotations.Annotation.Annotation at 0x7fedb40b4400>,\n <statannotations.Annotation.Annotation at 0x7fedb715ccf8>,\n <statannotations.Annotation.Annotation at 0x7fedb6c8ef28>])" + "text/plain": "(<matplotlib.axes._subplots.AxesSubplot at 0x7fc284d10f60>,\n [<statannotations.Annotation.Annotation at 0x7fc281c837b8>,\n <statannotations.Annotation.Annotation at 0x7fc2817bfc18>,\n <statannotations.Annotation.Annotation at 0x7fc284d8e518>])" }, "execution_count": 15, "metadata": {}, @@ -947,7 +947,7 @@ { "data": { "text/plain": "<Figure size 864x432 with 1 Axes>", - "image/png": "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\n" + "image/png": "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\n" }, "metadata": {}, "output_type": "display_data" @@ -1001,7 +1001,7 @@ { "cell_type": "markdown", "source": [ - "And a similar swarm plot, with a different format for pvalues" + "With a horizontal orientation" ], "metadata": { "collapsed": false @@ -1010,6 +1010,74 @@ { "cell_type": "code", "execution_count": 21, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "p-value annotation legend:\n", + " ns: p <= 1.00e+00\n", + " : 1.00e-02 < p <= 5.00e-02\n", + " : 1.00e-03 < p <= 1.00e-02\n", + " : 1.00e-04 < p <= 1.00e-03\n", + " : p <= 1.00e-04\n", + "\n", + "H1N1_Nonsynonymous vs. H1N1_Synonymous: Mann-Whitney-Wilcoxon test two-sided, P_val:5.014e-04 U_stat=2.624e+03\n", + "H3N2_Nonsynonymous vs. H3N2_Synonymous: Mann-Whitney-Wilcoxon test two-sided, P_val:1.294e-03 U_stat=1.535e+04\n", + "Influenza B_Nonsynonymous vs. Influenza B_Synonymous: Mann-Whitney-Wilcoxon test two-sided, P_val:2.026e-01 U_stat=3.340e+02\n" + ] + }, + { + "data": { + "text/plain": "(<matplotlib.axes._subplots.AxesSubplot at 0x7fc283af2668>,\n [<statannotations.Annotation.Annotation at 0x7fc283bf3780>,\n <statannotations.Annotation.Annotation at 0x7fc283be1dd8>,\n <statannotations.Annotation.Annotation at 0x7fc283bc3ef0>])" + }, + "execution_count": 21, + "metadata": {}, + "output_type": "execute_result" + }, + { + "data": { + "text/plain": "<Figure size 864x432 with 1 Axes>", + "image/png": "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\n" + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "fig, ax = plt.subplots(1,1, figsize=(12,6))\n", + "fig_args_horiz = {\n", + " fig_args,\n", + " 'y': 'subtype',\n", + " 'x': 'mutation_rate_samp'\n", + "}\n", + "ax.set_xscale('log')\n", + "sns.stripplot(ax=ax, orient='h', fig_args_horiz)\n", + "annotator.new_plot(ax, plot='swarmplot', orient='h', fig_args_horiz)\n", + "annotator.apply_and_annotate()" + ], + "metadata": { + "collapsed": false, + "pycharm": { + "name": "#%%\n" + } + } + }, + { + "cell_type": "markdown", + "source": [ + "Back to vertical, with a different format for pvalues" + ], + "metadata": { + "collapsed": false, + "pycharm": { + "name": "#%% md\n" + } + } + }, + { + "cell_type": "code", + "execution_count": 22, "outputs": [ { "name": "stdout", @@ -1060,7 +1128,7 @@ }, { "cell_type": "code", - "execution_count": 22, + "execution_count": 23, "outputs": [ { "name": "stdout", @@ -1101,7 +1169,7 @@ }, { "cell_type": "code", - "execution_count": 22, + "execution_count": 23, "outputs": [], "source": [], "metadata": {	Horizontal Support for horizontal orientation, at the moment tested in box plots and strip plots. Mentioned in #26 , https://github.com/webermarcolivier/statannot/issues/54		2021-07-25T15:16:18	[]	[]
trevismd/statannotations	trevismd__statannotations-27	4beac149bf62a9a9f3a2e9d7857b76bd62889b9a	diff --git a/CHANGELOG.md b/CHANGELOG.md index 80407ff..9c2425b 100644 --- a/CHANGELOG.md +++ b/CHANGELOG.md @@ -1,4 +1,7 @@ ## v0.4 +### v0.4.1 + - Support for horizontal orientation + ### v0.4.0 - Major refactoring, change to an Annotator `class` to prepare and add annotations in separate function (now method) calls. diff --git a/README.md b/README.md index 7e9b432..fff5153 100644 --- a/README.md +++ b/README.md @@ -42,6 +42,7 @@ corresponding branch). - Interface to use any other function from any source with minimal extra code - Smart layout of multiple annotations with correct y offsets. +- Support for vertical and horizontal orientation - Annotations can be located inside or outside the plot. - Corrections for multiple testing can be applied (binding to `statsmodels.stats.multitest.multipletests` methods): @@ -51,7 +52,7 @@ corresponding branch). - Benjamini-Yekutieli - And any other function from any source with minimal extra code - Format of the statistical test annotation can be customized: - star annotation, simplified p-value, or explicit p-value. + star annotation, simplified p-value format, or explicit p-value. - Optionally, custom p-values can be given as input. In this case, no statistical test is performed, but corrections for multiple testing can be applied. diff --git a/docs/build/doctrees/environment.pickle b/docs/build/doctrees/environment.pickle index 771aba2..169ad27 100644 Binary files a/docs/build/doctrees/environment.pickle and b/docs/build/doctrees/environment.pickle differ diff --git a/docs/build/doctrees/statannotations.doctree b/docs/build/doctrees/statannotations.doctree index ee0691c..759af7f 100644 Binary files a/docs/build/doctrees/statannotations.doctree and b/docs/build/doctrees/statannotations.doctree differ diff --git a/docs/build/html/genindex.html b/docs/build/html/genindex.html index 935f030..e1e1951 100644 --- a/docs/build/html/genindex.html +++ b/docs/build/html/genindex.html @@ -166,6 +166,7 @@ <h1 id="index">Index</h1> \| <a href="#L"><strong>L</strong></a> \| <a href="#M"><strong>M</strong></a> \| <a href="#N"><strong>N</strong></a> + \| <a href="#O"><strong>O</strong></a> \| <a href="#P"><strong>P</strong></a> \| <a href="#R"><strong>R</strong></a> \| <a href="#S"><strong>S</strong></a> @@ -365,6 +366,14 @@ <h2 id="N">N</h2> </ul></td> </tr></table> +<h2 id="O">O</h2> +<table style="width: 100%" class="indextable genindextable"><tr> + <td style="width: 33%; vertical-align: top;"><ul> + <li><a href="statannotations.html#statannotations.Annotator.Annotator.orient">orient (statannotations.Annotator.Annotator property)</a> +</li> + </ul></td> +</tr></table> + <h2 id="P">P</h2> <table style="width: 100%" class="indextable genindextable"><tr> <td style="width: 33%; vertical-align: top;"><ul> diff --git a/docs/build/html/objects.inv b/docs/build/html/objects.inv index 0d82dd7..b77fb8b 100644 Binary files a/docs/build/html/objects.inv and b/docs/build/html/objects.inv differ diff --git a/docs/build/html/searchindex.js b/docs/build/html/searchindex.js index 38da9ed..2e1b48e 100644 --- a/docs/build/html/searchindex.js +++ b/docs/build/html/searchindex.js @@ -1,1 +1,1 @@ -Search.setIndex({docnames:["index","modules","setup","statannotations","statannotations.stats"],envversion:{"sphinx.domains.c":2,"sphinx.domains.changeset":1,"sphinx.domains.citation":1,"sphinx.domains.cpp":3,"sphinx.domains.index":1,"sphinx.domains.javascript":2,"sphinx.domains.math":2,"sphinx.domains.python":3,"sphinx.domains.rst":2,"sphinx.domains.std":2,sphinx:56},filenames:["index.rst","modules.rst","setup.rst","statannotations.rst","statannotations.stats.rst"],objects:{"":{statannotations:[3,0,0,"-"]},"statannotations.Annotation":{Annotation:[3,1,1,""]},"statannotations.Annotation.Annotation":{formatted_output:[3,2,1,""],print_labels_and_content:[3,3,1,""],text:[3,2,1,""]},"statannotations.Annotator":{Annotator:[3,1,1,""]},"statannotations.Annotator.Annotator":{alpha:[3,2,1,""],annotate:[3,3,1,""],annotate_custom_annotations:[3,3,1,""],apply_and_annotate:[3,3,1,""],apply_test:[3,3,1,""],comparisons_correction:[3,2,1,""],configure:[3,3,1,""],fig:[3,2,1,""],get_annotations_text:[3,3,1,""],get_configuration:[3,3,1,""],get_offset_func:[3,3,1,""],has_type0_comparisons_correction:[3,3,1,""],loc:[3,2,1,""],new_plot:[3,3,1,""],print_pvalue_legend:[3,3,1,""],pvalue_format:[3,2,1,""],reset_configuration:[3,3,1,""],set_custom_annotations:[3,3,1,""],set_pvalues:[3,3,1,""],set_pvalues_and_annotate:[3,3,1,""],test:[3,2,1,""],validate_test_short_name:[3,3,1,""],verbose:[3,2,1,""]},"statannotations.PValueFormat":{Formatter:[3,1,1,""],PValueFormat:[3,1,1,""],sort_pvalue_thresholds:[3,4,1,""]},"statannotations.PValueFormat.Formatter":{config:[3,3,1,""],format_data:[3,3,1,""]},"statannotations.PValueFormat.PValueFormat":{config:[3,3,1,""],format_data:[3,3,1,""],get_configuration:[3,3,1,""],print_legend_if_used:[3,3,1,""],pvalue_format_string:[3,2,1,""],pvalue_thresholds:[3,2,1,""],simple_format_string:[3,2,1,""],text_format:[3,2,1,""]},"statannotations.format_annotations":{pval_annotation_text:[3,4,1,""],simple_text:[3,4,1,""]},"statannotations.stats":{ComparisonsCorrection:[4,0,0,"-"],StatResult:[4,0,0,"-"],StatTest:[4,0,0,"-"],test:[4,0,0,"-"],utils:[4,0,0,"-"]},"statannotations.stats.ComparisonsCorrection":{ComparisonsCorrection:[4,1,1,""],check_valid_correction_name:[4,4,1,""],get_correction_parameters:[4,4,1,""],get_validated_comparisons_correction:[4,4,1,""]},"statannotations.stats.ComparisonsCorrection.ComparisonsCorrection":{apply:[4,3,1,""],document:[4,3,1,""]},"statannotations.stats.StatResult":{StatResult:[4,1,1,""]},"statannotations.stats.StatResult.StatResult":{corrected_significance:[4,2,1,""],correction_method:[4,2,1,""],formatted_output:[4,2,1,""],significance_suffix:[4,2,1,""]},"statannotations.stats.StatTest":{StatTest:[4,1,1,""],wilcoxon:[4,4,1,""]},"statannotations.stats.StatTest.StatTest":{from_library:[4,3,1,""],short_name:[4,2,1,""]},"statannotations.stats.test":{apply_test:[4,4,1,""]},"statannotations.stats.utils":{check_alpha:[4,4,1,""],check_num_comparisons:[4,4,1,""],check_pvalues:[4,4,1,""],get_num_comparisons:[4,4,1,""],return_results:[4,4,1,""]},"statannotations.utils":{check_is_in:[3,4,1,""],check_not_none:[3,4,1,""],check_order_in_data:[3,4,1,""],check_pairs_in_data:[3,4,1,""],check_valid_text_format:[3,4,1,""],get_closest:[3,4,1,""],get_x_values:[3,4,1,""],raise_expected_got:[3,4,1,""],remove_null:[3,4,1,""],render_collection:[3,4,1,""]},statannotations:{Annotation:[3,0,0,"-"],Annotator:[3,0,0,"-"],PValueFormat:[3,0,0,"-"],format_annotations:[3,0,0,"-"],stats:[4,0,0,"-"],utils:[3,0,0,"-"]}},objnames:{"0":["py","module","Python module"],"1":["py","class","Python class"],"2":["py","property","Python property"],"3":["py","method","Python method"],"4":["py","function","Python function"]},objtypes:{"0":"py:module","1":"py:class","2":"py:property","3":"py:method","4":"py:function"},terms:{"0":[3,4],"05":[3,4],"1":[3,4],"12141511":3,"9981846":3,"case":4,"class":[3,4],"default":[3,4],"float":4,"function":[3,4],"int":4,"return":3,"short":3,"static":4,"true":[3,4],A:3,If:3,Or:3,The:3,a_list:3,accept:3,add:3,addit:4,alpha:[3,4],an:3,annot:1,annotate_custom_annot:3,appli:[3,4],apply_and_annot:3,apply_test:[3,4],ar:3,arg:[3,4],argument:[3,4],arrai:3,assum:3,auto:3,ax:3,bar:3,base:[3,4],behavior:4,benjamini:4,between:3,bonferroni:4,bool:4,box:3,boxplot:3,calcul:3,callabl:4,check:3,check_alpha:4,check_is_in:3,check_not_non:3,check_num_comparison:4,check_order_in_data:3,check_pairs_in_data:3,check_pvalu:4,check_valid_correction_nam:4,check_valid_text_format:3,classmethod:3,close:3,closest:3,collect:3,color:3,com:3,comparison:[3,4],comparisons_correct:[3,4],comparisonscorrect:[1,3],comput:3,config:3,configur:3,constructor:3,content:1,coordin:3,core:3,corr_kwarg:4,correct:4,corrected_signific:4,correction_method:4,custom:3,data:[3,4],datafram:3,dict:4,displai:3,document:4,each:3,earlier:4,either:4,engin:3,equal:3,equival:4,error:3,error_typ:3,exact:3,exist:3,expect:3,fals:3,fig:3,follow:3,for_:3,format:[3,4],format_annot:1,format_data:3,formatt:3,formatted_output:[3,4],fraction:3,frame:3,from:[3,4],from_librari:4,func:4,gener:3,get:4,get_annotations_text:3,get_closest:3,get_configur:3,get_correction_paramet:4,get_num_comparison:4,get_offset_func:3,get_validated_comparisons_correct:4,get_x_valu:3,got:3,group:3,group_data1:4,group_data2:4,gt:4,has_type0_comparisons_correct:3,hochberg:4,hold:3,holm:4,http:3,hue:3,hue_ord:3,index:0,instanc:[3,4],instead:3,interfac:4,interpret:4,keyword:4,kruskal:4,kwarg:[3,4],label:3,leven:4,line_height:3,line_offset:3,line_offset_to_group:3,line_width:3,link:3,list:3,loc:3,ls:4,mann:4,messag:3,method:4,method_typ:4,mode:3,modul:[0,1],multipl:4,must:3,mylist:3,mynumb:3,name:[3,4],new_plot:3,none:[3,4],num_comparison:[3,4],number:[3,4],object:[3,4],one:[3,4],option:[3,4],order:3,otherwis:3,overrid:3,p:3,p_valu:4,packag:[0,1],page:0,pair:3,panda:3,paramet:[3,4],pass:[3,4],per:3,perform:3,plot:3,plot_param:3,point:3,posit:3,print_labels_and_cont:3,print_legend_if_us:3,print_pvalue_legend:3,properti:[3,4],provid:[3,4],pval:4,pval_annotation_text:3,pvalu:[3,4],pvalue_format:3,pvalue_format_str:3,pvalue_threshold:3,pvalueformat:1,rais:3,raise_expected_got:3,rang:3,refer:3,remove_nul:3,render:3,render_collect:3,reset_configur:3,result:[3,4],results_arrai:4,return_result:4,run:4,same:3,sampl:4,seaborn:3,search:0,sep:3,seri:3,set:3,set_custom_annot:3,set_pvalu:3,set_pvalues_and_annot:3,short_nam:4,show:3,show_test_nam:3,significance_suffix:4,simpl:3,simple_format_str:3,simple_text:3,smallest:3,sort:3,sort_pvalue_threshold:3,stackoverflow:3,standard:3,stat:[1,3],stat_nam:4,stat_str:4,statist:[3,4],statresult:[1,3],stats_param:[3,4],statsmodel:4,statsmodels_api:4,stattest:[1,3],str:[3,4],string:3,struct:3,submodul:1,subpackag:1,t:4,test:[1,3],test_descript:4,test_ind:4,test_long_nam:4,test_nam:4,test_pair:4,test_result_list:4,test_short_nam:[3,4],test_welch:4,text:3,text_annot_custom:3,text_format:3,text_offset:3,them:3,thereof:3,thi:[3,4],three:3,threshold:3,top:3,tupl:3,two:[3,4],type:3,union:[3,4],us:[3,4],use_fixed_offset:3,util:1,valid_valu:3,validate_test_short_nam:3,valu:[3,4],valueerror:3,verbos:[3,4],version:4,vs:3,wa:3,whitnei:4,wilcoxon:4,work:3,x:3,y:3,yekuti:4},titles:["Welcome to statannotations\u2019s documentation!","statannotations","setup module","statannotations package","statannotations.stats package"],titleterms:{annot:3,comparisonscorrect:4,content:[0,3,4],document:0,format_annot:3,indic:0,modul:[2,3,4],packag:[3,4],pvalueformat:3,s:0,setup:2,stat:4,statannot:[0,1,3,4],statresult:4,stattest:4,submodul:[3,4],subpackag:3,tabl:0,test:4,util:[3,4],welcom:0}}) \ No newline at end of file +Search.setIndex({docnames:["index","modules","setup","statannotations","statannotations.stats"],envversion:{"sphinx.domains.c":2,"sphinx.domains.changeset":1,"sphinx.domains.citation":1,"sphinx.domains.cpp":3,"sphinx.domains.index":1,"sphinx.domains.javascript":2,"sphinx.domains.math":2,"sphinx.domains.python":3,"sphinx.domains.rst":2,"sphinx.domains.std":2,sphinx:56},filenames:["index.rst","modules.rst","setup.rst","statannotations.rst","statannotations.stats.rst"],objects:{"":{statannotations:[3,0,0,"-"]},"statannotations.Annotation":{Annotation:[3,1,1,""]},"statannotations.Annotation.Annotation":{formatted_output:[3,2,1,""],print_labels_and_content:[3,3,1,""],text:[3,2,1,""]},"statannotations.Annotator":{Annotator:[3,1,1,""]},"statannotations.Annotator.Annotator":{alpha:[3,2,1,""],annotate:[3,3,1,""],annotate_custom_annotations:[3,3,1,""],apply_and_annotate:[3,3,1,""],apply_test:[3,3,1,""],comparisons_correction:[3,2,1,""],configure:[3,3,1,""],fig:[3,2,1,""],get_annotations_text:[3,3,1,""],get_configuration:[3,3,1,""],get_offset_func:[3,3,1,""],has_type0_comparisons_correction:[3,3,1,""],loc:[3,2,1,""],new_plot:[3,3,1,""],orient:[3,2,1,""],print_pvalue_legend:[3,3,1,""],pvalue_format:[3,2,1,""],reset_configuration:[3,3,1,""],set_custom_annotations:[3,3,1,""],set_pvalues:[3,3,1,""],set_pvalues_and_annotate:[3,3,1,""],test:[3,2,1,""],validate_test_short_name:[3,3,1,""],verbose:[3,2,1,""]},"statannotations.PValueFormat":{Formatter:[3,1,1,""],PValueFormat:[3,1,1,""],sort_pvalue_thresholds:[3,4,1,""]},"statannotations.PValueFormat.Formatter":{config:[3,3,1,""],format_data:[3,3,1,""]},"statannotations.PValueFormat.PValueFormat":{config:[3,3,1,""],format_data:[3,3,1,""],get_configuration:[3,3,1,""],print_legend_if_used:[3,3,1,""],pvalue_format_string:[3,2,1,""],pvalue_thresholds:[3,2,1,""],simple_format_string:[3,2,1,""],text_format:[3,2,1,""]},"statannotations.format_annotations":{pval_annotation_text:[3,4,1,""],simple_text:[3,4,1,""]},"statannotations.stats":{ComparisonsCorrection:[4,0,0,"-"],StatResult:[4,0,0,"-"],StatTest:[4,0,0,"-"],test:[4,0,0,"-"],utils:[4,0,0,"-"]},"statannotations.stats.ComparisonsCorrection":{ComparisonsCorrection:[4,1,1,""],check_valid_correction_name:[4,4,1,""],get_correction_parameters:[4,4,1,""],get_validated_comparisons_correction:[4,4,1,""]},"statannotations.stats.ComparisonsCorrection.ComparisonsCorrection":{apply:[4,3,1,""],document:[4,3,1,""]},"statannotations.stats.StatResult":{StatResult:[4,1,1,""]},"statannotations.stats.StatResult.StatResult":{corrected_significance:[4,2,1,""],correction_method:[4,2,1,""],formatted_output:[4,2,1,""],significance_suffix:[4,2,1,""]},"statannotations.stats.StatTest":{StatTest:[4,1,1,""],wilcoxon:[4,4,1,""]},"statannotations.stats.StatTest.StatTest":{from_library:[4,3,1,""],short_name:[4,2,1,""]},"statannotations.stats.test":{apply_test:[4,4,1,""]},"statannotations.stats.utils":{check_alpha:[4,4,1,""],check_num_comparisons:[4,4,1,""],check_pvalues:[4,4,1,""],get_num_comparisons:[4,4,1,""],return_results:[4,4,1,""]},"statannotations.utils":{check_is_in:[3,4,1,""],check_not_none:[3,4,1,""],check_order_in_data:[3,4,1,""],check_pairs_in_data:[3,4,1,""],check_valid_text_format:[3,4,1,""],get_closest:[3,4,1,""],get_x_values:[3,4,1,""],raise_expected_got:[3,4,1,""],remove_null:[3,4,1,""],render_collection:[3,4,1,""]},statannotations:{Annotation:[3,0,0,"-"],Annotator:[3,0,0,"-"],PValueFormat:[3,0,0,"-"],format_annotations:[3,0,0,"-"],stats:[4,0,0,"-"],utils:[3,0,0,"-"]}},objnames:{"0":["py","module","Python module"],"1":["py","class","Python class"],"2":["py","property","Python property"],"3":["py","method","Python method"],"4":["py","function","Python function"]},objtypes:{"0":"py:module","1":"py:class","2":"py:property","3":"py:method","4":"py:function"},terms:{"0":[3,4],"05":[3,4],"1":[3,4],"12141511":3,"9981846":3,"case":4,"class":[3,4],"default":[3,4],"float":4,"function":[3,4],"int":4,"return":3,"short":3,"static":4,"true":[3,4],A:3,If:3,Or:3,The:3,a_list:3,accept:3,add:3,addit:4,alpha:[3,4],an:3,annot:1,annotate_custom_annot:3,appli:[3,4],apply_and_annot:3,apply_test:[3,4],ar:3,arg:[3,4],argument:[3,4],arrai:3,assum:3,auto:3,ax:3,bar:3,base:[3,4],behavior:4,benjamini:4,between:3,bonferroni:4,bool:4,box:3,boxplot:3,calcul:3,callabl:4,check:3,check_alpha:4,check_is_in:3,check_not_non:3,check_num_comparison:4,check_order_in_data:3,check_pairs_in_data:3,check_pvalu:4,check_valid_correction_nam:4,check_valid_text_format:3,classmethod:3,close:3,closest:3,collect:3,color:3,com:3,comparison:[3,4],comparisons_correct:[3,4],comparisonscorrect:[1,3],comput:3,config:3,configur:3,constructor:3,content:1,coord:3,coordin:3,core:3,corr_kwarg:4,correct:4,corrected_signific:4,correction_method:4,custom:3,data:[3,4],datafram:3,dict:4,displai:3,document:4,each:3,earlier:4,either:4,engin:3,equal:3,equival:4,error:3,error_typ:3,exact:3,exist:3,expect:3,fals:3,fig:3,follow:3,for_:3,format:[3,4],format_annot:1,format_data:3,formatt:3,formatted_output:[3,4],fraction:3,frame:3,from:[3,4],from_librari:4,func:4,gener:3,get:4,get_annotations_text:3,get_closest:3,get_configur:3,get_correction_paramet:4,get_num_comparison:4,get_offset_func:3,get_validated_comparisons_correct:4,get_x_valu:3,got:3,group:3,group_data1:4,group_data2:4,gt:4,has_type0_comparisons_correct:3,hochberg:4,hold:3,holm:4,http:3,hue:3,hue_ord:3,index:0,instanc:[3,4],instead:3,interfac:4,interpret:4,keyword:4,kruskal:4,kwarg:[3,4],label:3,leven:4,line_height:3,line_offset:3,line_offset_to_group:3,line_width:3,link:3,list:3,loc:3,ls:4,mann:4,messag:3,method:4,method_typ:4,mode:3,modul:[0,1],multipl:4,must:3,mylist:3,mynumb:3,name:[3,4],new_plot:3,none:[3,4],num_comparison:[3,4],number:[3,4],object:[3,4],one:[3,4],option:[3,4],order:3,orient:3,otherwis:3,overrid:3,p:3,p_valu:4,packag:[0,1],page:0,pair:3,panda:3,paramet:[3,4],pass:[3,4],per:3,perform:3,plot:3,plot_param:3,point:3,posit:3,print_labels_and_cont:3,print_legend_if_us:3,print_pvalue_legend:3,properti:[3,4],provid:[3,4],pval:4,pval_annotation_text:3,pvalu:[3,4],pvalue_format:3,pvalue_format_str:3,pvalue_threshold:3,pvalueformat:1,rais:3,raise_expected_got:3,rang:3,refer:3,remove_nul:3,render:3,render_collect:3,reset_configur:3,result:[3,4],results_arrai:4,return_result:4,run:4,same:3,sampl:4,seaborn:3,search:0,sep:3,seri:3,set:3,set_custom_annot:3,set_pvalu:3,set_pvalues_and_annot:3,short_nam:4,show:3,show_test_nam:3,significance_suffix:4,simpl:3,simple_format_str:3,simple_text:3,smallest:3,sort:3,sort_pvalue_threshold:3,stackoverflow:3,standard:3,stat:[1,3],stat_nam:4,stat_str:4,statist:[3,4],statresult:[1,3],stats_param:[3,4],statsmodel:4,statsmodels_api:4,stattest:[1,3],str:[3,4],string:3,struct:3,submodul:1,subpackag:1,t:4,test:[1,3],test_descript:4,test_ind:4,test_long_nam:4,test_nam:4,test_pair:4,test_result_list:4,test_short_nam:[3,4],test_welch:4,text:3,text_annot_custom:3,text_format:3,text_offset:3,them:3,thereof:3,thi:[3,4],three:3,threshold:3,top:3,tupl:3,two:[3,4],type:3,union:[3,4],us:[3,4],use_fixed_offset:3,util:1,valid_valu:3,validate_test_short_nam:3,valu:[3,4],valueerror:3,verbos:[3,4],version:4,vs:3,wa:3,whitnei:4,wilcoxon:4,work:3,x:3,y:3,yekuti:4},titles:["Welcome to statannotations\u2019s documentation!","statannotations","setup module","statannotations package","statannotations.stats package"],titleterms:{annot:3,comparisonscorrect:4,content:[0,3,4],document:0,format_annot:3,indic:0,modul:[2,3,4],packag:[3,4],pvalueformat:3,s:0,setup:2,stat:4,statannot:[0,1,3,4],statresult:4,stattest:4,submodul:[3,4],subpackag:3,tabl:0,test:4,util:[3,4],welcom:0}}) \ No newline at end of file diff --git a/docs/build/html/statannotations.html b/docs/build/html/statannotations.html index aaacef6..7363f25 100644 --- a/docs/build/html/statannotations.html +++ b/docs/build/html/statannotations.html @@ -352,6 +352,11 @@ <h2>Submodules<a class="headerlink" href="#submodules" title="Permalink to this <span class="sig-name descname"><span class="pre">new_plot</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">ax</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pairs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">plot</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'boxplot'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">data</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">y</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">hue</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">order</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">hue_order</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">engine</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">str</span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">'seaborn'</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">plot_params</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#statannotations.Annotator.Annotator.new_plot" title="Permalink to this definition">¶</a></dt> <dd></dd></dl> +<dl class="py property"> +<dt class="sig sig-object py" id="statannotations.Annotator.Annotator.orient"> +<em class="property"><span class="pre">property</span> </em><span class="sig-name descname"><span class="pre">orient</span></span><a class="headerlink" href="#statannotations.Annotator.Annotator.orient" title="Permalink to this definition">¶</a></dt> +<dd></dd></dl> + <dl class="py method"> <dt class="sig sig-object py" id="statannotations.Annotator.Annotator.print_pvalue_legend"> <span class="sig-name descname"><span class="pre">print_pvalue_legend</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="headerlink" href="#statannotations.Annotator.Annotator.print_pvalue_legend" title="Permalink to this definition">¶</a></dt> @@ -544,7 +549,7 @@ <h2>Submodules<a class="headerlink" href="#submodules" title="Permalink to this <dl class="py function"> <dt class="sig sig-object py" id="statannotations.utils.check_pairs_in_data"> -<span class="sig-prename descclassname"><span class="pre">statannotations.utils.</span></span><span class="sig-name descname"><span class="pre">check_pairs_in_data</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">pairs</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">list</span><span class="p"><span class="pre">,</span> </span><span class="pre">tuple</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">data</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">list</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span> </span><span class="pre">pandas.core.frame.DataFrame</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span> </span><span class="pre">list</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">hue</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">hue_order</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#statannotations.utils.check_pairs_in_data" title="Permalink to this definition">¶</a></dt> +<span class="sig-prename descclassname"><span class="pre">statannotations.utils.</span></span><span class="sig-name descname"><span class="pre">check_pairs_in_data</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">pairs</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">list</span><span class="p"><span class="pre">,</span> </span><span class="pre">tuple</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">data</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">list</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span> </span><span class="pre">pandas.core.frame.DataFrame</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">coord</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span> </span><span class="pre">list</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">hue</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">hue_order</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#statannotations.utils.check_pairs_in_data" title="Permalink to this definition">¶</a></dt> <dd><p>Checks that values referred to in <cite>order</cite> and <cite>pairs</cite> exist in data.</p> </dd></dl> diff --git a/statannotations/Annotator.py b/statannotations/Annotator.py index 390bd83..00dcee6 100644 --- a/statannotations/Annotator.py +++ b/statannotations/Annotator.py @@ -123,7 +123,7 @@ def __init__(self, ax, pairs, plot='boxplot', data=None, x=None, self.text_offset_impact_above = 0 self.color = '0.2' self.line_width = 1.5 - self.y_offset = None + self.value_offset = None self.custom_annotations = None def new_plot(self, ax, pairs=None, plot='boxplot', data=None, x=None, @@ -144,6 +144,10 @@ def new_plot(self, ax, pairs=None, plot='boxplot', data=None, x=None, return self + @property + def orient(self): + return self._plotter.orient + @property def verbose(self): return self._verbose @@ -154,8 +158,8 @@ def verbose(self, verbose): self._plotter.verbose = verbose @property - def _y_stack_arr(self): - return self._plotter.y_stack_arr + def _value_stack_arr(self): + return self._plotter.value_stack_arr @property def fig(self): @@ -179,13 +183,13 @@ def annotate(self, line_offset=None, line_offset_to_group=None): "test (again) which will probably lead to " "unexpected results") - self._update_y_for_loc() + self._update_value_for_loc() ann_list = [] - orig_ylim = self.ax.get_ylim() + orig_value_lim = self._plotter.get_value_lim() offset_func = self.get_offset_func(self.loc) - self.y_offset, self.line_offset_to_group = offset_func( + self.value_offset, self.line_offset_to_group = offset_func( line_offset, line_offset_to_group) if self._verbose: @@ -199,16 +203,24 @@ def annotate(self, line_offset=None, line_offset_to_group=None): self._annotate_pair(annotation, ax_to_data=ax_to_data, ann_list=ann_list, - orig_ylim=orig_ylim) + orig_value_lim=orig_value_lim) # reset transformation - y_stack_max = max(self._y_stack_arr[1, :]) + y_stack_max = max(self._value_stack_arr[1, :]) ax_to_data = self._plotter.get_transform_func('ax_to_data') - ylims = ([(0, 0), (0, max(1.04 y_stack_max, 1))] - if self.loc == 'inside' - else [(0, 0), (0, 1)]) - - self.ax.set_ylim(ax_to_data.transform(ylims)[:, 1]) + value_lims = ( + ([(0, 0), (0, max(1.04 * y_stack_max, 1))] + if self.loc == 'inside' + else [(0, 0), (0, 1)]) + if self.orient == 'v' + else + ([(0, 0), (max(1.04 * y_stack_max, 1), 0)] + if self.loc == 'inside' + else [(0, 0), (1, 0)]) + ) + set_lims = self.ax.set_ylim if self.orient == 'v' else self.ax.set_xlim + transformed = ax_to_data.transform(value_lims) + set_lims(transformed[:, 1 if self.orient == 'v' else 0]) return self.ax, self.annotations @@ -444,61 +456,75 @@ def _get_results(self, num_comparisons, pvalues=None, return test_result_list - def _annotate_pair(self, annotation, ax_to_data, ann_list, orig_ylim): + def _annotate_pair(self, annotation, ax_to_data, ann_list, orig_value_lim): if self._verbose >= 1: annotation.print_labels_and_content() - x1 = annotation.structs[0]['x'] - x2 = annotation.structs[1]['x'] - xi1 = annotation.structs[0]['xi'] - xi2 = annotation.structs[1]['xi'] + group_coord_1 = annotation.structs[0]['group_coord'] + group_coord_2 = annotation.structs[1]['group_coord'] + group_i1 = annotation.structs[0]['group_i'] + group_i2 = annotation.structs[1]['group_i'] # Find y maximum for all the y_stacks in between group1 and group2 - i_ymax_in_range_x1_x2 = xi1 + np.nanargmax( - self._y_stack_arr[1, np.where((x1 <= self._y_stack_arr[0, :]) - & (self._y_stack_arr[0, :] <= x2))]) + i_value_max_in_range_g1_g2 = group_i1 + np.nanargmax( + self._value_stack_arr[1, + np.where((group_coord_1 + <= self._value_stack_arr[0, :]) + & (self._value_stack_arr[0, :] + <= group_coord_2))]) - y = self._get_y_for_pair(i_ymax_in_range_x1_x2) + value = self._get_value_for_pair(i_value_max_in_range_g1_g2) # Determine lines in axes coordinates - ax_line_x = [x1, x1, x2, x2] - ax_line_y = [y, y + self.line_height, y + self.line_height, y] + ax_line_group = [group_coord_1, group_coord_1, + group_coord_2, group_coord_2] + ax_line_value = [value, value + self.line_height, + value + self.line_height, value] + + lists = ((ax_line_group, ax_line_value) if self.orient == 'v' + else (ax_line_value, ax_line_group)) points = [ax_to_data.transform((x, y)) for x, y - in zip(ax_line_x, ax_line_y)] + in zip(lists)] line_x, line_y = zip(points) self._plot_line(line_x, line_y) + xy_params = self._get_xy_params(group_coord_1, group_coord_2, line_x, + line_y) + ann = self.ax.annotate( - annotation.text, xy=(np.mean([x1, x2]), line_y[2]), - xytext=(0, self.text_offset), textcoords='offset points', + annotation.text, textcoords='offset points', xycoords='data', ha='center', va='bottom', fontsize=self._pvalue_format.fontsize, clip_on=False, - annotation_clip=False) + annotation_clip=False, *xy_params) + if annotation.text is not None: ann_list.append(ann) plt.draw() - self.ax.set_ylim(orig_ylim) + set_lim = {'v': 'set_ylim', + 'h': 'set_xlim'}[self.orient] + + getattr(self.ax, set_lim)(orig_value_lim) - y_top_annot = self._annotate_pair_text(ann, y) + value_top_annot = self._annotate_pair_text(ann, value) else: - y_top_annot = y + self.line_height + value_top_annot = value + self.line_height - # Fill the highest y position of the annotation into the y_stack array - # for all positions in the range x1 to x2 - self._y_stack_arr[1, - (x1 <= self._y_stack_arr[0, :]) - & (self._y_stack_arr[0, :] <= x2)] = y_top_annot + # Fill the highest value position of the annotation into the value_stack + # array for all positions in the range group_coord_1 to group_coord_2 + self._value_stack_arr[ + 1, (group_coord_1 <= self._value_stack_arr[0, :]) + & (self._value_stack_arr[0, :] <= group_coord_2)] = value_top_annot - # Increment the counter of annotations in the y_stack array - self._y_stack_arr[2, xi1:xi2 + 1] += 1 + # Increment the counter of annotations in the value_stack array + self._value_stack_arr[2, group_i1:group_i2 + 1] += 1 - def _update_y_for_loc(self): + def _update_value_for_loc(self): if self._loc == 'outside': - self._y_stack_arr[1, :] = 1 + self._value_stack_arr[1, :] = 1 def _check_test_pvalues_perform(self): if self.test is None: @@ -582,8 +608,7 @@ def _get_offsets_inside(line_offset, line_offset_to_group): else: if line_offset_to_group is None: line_offset_to_group = 0.06 - y_offset = line_offset - return y_offset, line_offset_to_group + return line_offset, line_offset_to_group @staticmethod def _get_offsets_outside(line_offset, line_offset_to_group): @@ -623,9 +648,9 @@ def _plot_line(self, line_x, line_y): line.set_clip_on(False) self.ax.add_line(line) - def _annotate_pair_text(self, ann, y): + def _annotate_pair_text(self, ann, value): - y_top_annot = None + value_top_annot = None got_mpl_error = False if not self.use_fixed_offset: @@ -633,7 +658,8 @@ def _annotate_pair_text(self, ann, y): bbox = ann.get_window_extent() pix_to_ax = self._plotter.get_transform_func('pix_to_ax') bbox_ax = bbox.transformed(pix_to_ax) - y_top_annot = bbox_ax.ymax + value_coord_max = {'v': 'ymax', 'h': 'xmax'}[self.orient] + value_top_annot = getattr(bbox_ax, value_coord_max) except RuntimeError: got_mpl_error = True @@ -644,40 +670,45 @@ def _annotate_pair_text(self, ann, y): "back to a fixed y offset. Layout may be not " "optimal.") - # We will apply a fixed offset in points, - # based on the font size of the annotation. fontsize_points = FontProperties( size='medium').get_size_in_points() - offset_trans = mtransforms.offset_copy( - self.ax.transAxes, fig=self.fig, x=0, - y=fontsize_points + self.text_offset, units='points') - y_top_display = offset_trans.transform( - (0, y + self.line_height)) - y_top_annot = (self.ax.transAxes.inverted() - .transform(y_top_display)[1]) + direction = {'h': -1, 'v': 1}[self.orient] + x, y = [0, fontsize_points + self.text_offset][::direction] + offset_trans = mtransforms.offset_copy(trans=self.ax.transAxes, + fig=self.fig, + units='points', x=x, y=y) + + value_top_display = offset_trans.transform( + (value + self.line_height, value + self.line_height)) + + value_coord = {'h': 0, 'v': 1}[self.orient] + + value_top_annot = (self.ax.transAxes.inverted() + .transform(value_top_display)[value_coord]) self.text_offset_impact_above = ( - y_top_annot - y - self.y_offset - self.line_height) - return y_top_annot + value_top_annot - value - self.value_offset - self.line_height) + + return value_top_annot def _reset_default_values(self): for attribute, default_value in _DEFAULT_VALUES.items(): setattr(self, attribute, default_value) - def _get_y_for_pair(self, i_ymax_in_range_x1_x2): + def _get_value_for_pair(self, i_ymax_in_range_x1_x2): - ymax_in_range_x1_x2 = self._y_stack_arr[1, i_ymax_in_range_x1_x2] + ymax_in_range_x1_x2 = self._value_stack_arr[1, i_ymax_in_range_x1_x2] # Choose the best offset depending on whether there is an annotation # below at the x position in the range [x1, x2] where the stack is the # highest - if self._y_stack_arr[2, i_ymax_in_range_x1_x2] == 0: + if self._value_stack_arr[2, i_ymax_in_range_x1_x2] == 0: # there is only a group below offset = self.line_offset_to_group else: # there is an annotation below - offset = self.y_offset + self.text_offset_impact_above + offset = self.value_offset + self.text_offset_impact_above return ymax_in_range_x1_x2 + offset @@ -697,3 +728,28 @@ def _get_plotter(engine, args, *kwargs): if engine_plotter is None: raise NotImplementedError(f"{engine} engine not implemented.") return engine_plotter(args, kwargs) + + def _get_xy_params_horizontal(self, group_coord_1, group_coord_2, + line_x: np.ndarray): + return { + 'xy': (line_x[2], np.mean([group_coord_1, group_coord_2])), + 'xytext': (self.text_offset, 0), + 'rotation': 270, + 'rotation_mode': 'anchor' + } + + def _get_xy_params_vertical(self, group_coord_1, group_coord_2, + line_y: np.ndarray): + return { + 'xy': (np.mean([group_coord_1, group_coord_2]), line_y[2]), + 'xytext': (0, self.text_offset), + } + + def _get_xy_params(self, group_coord_1, group_coord_2, line_x: np.ndarray, + line_y: np.ndarray): + if self.orient == 'h': + return self._get_xy_params_horizontal(group_coord_1, group_coord_2, + line_x) + + return self._get_xy_params_vertical(group_coord_1, group_coord_2, + line_y) diff --git a/statannotations/_Xpositions.py b/statannotations/_GroupsPositions.py similarity index 51% rename from statannotations/_Xpositions.py rename to statannotations/_GroupsPositions.py index 2520f3c..046d2b9 100644 --- a/statannotations/_Xpositions.py +++ b/statannotations/_GroupsPositions.py @@ -3,7 +3,7 @@ from statannotations.utils import get_closest -class _XPositions: +class _GroupsPositions: def __init__(self, plotter, group_names): self._plotter = plotter self._hue_names = self._plotter.hue_names @@ -15,40 +15,42 @@ def __init__(self, plotter, group_names): "Using hues with only one hue is not supported.") self.hue_offsets = self._plotter.hue_offsets - self._xunits = self.hue_offsets[1] - self.hue_offsets[0] + self._axis_units = self.hue_offsets[1] - self.hue_offsets[0] - self._xpositions = { - np.round(self.get_group_x_position(group_name), 1): group_name + self._groups_positions = { + np.round(self.get_group_axis_position(group_name), 1): group_name for group_name in group_names } - self._xpositions_list = sorted(self._xpositions.keys()) + self._groups_positions_list = sorted(self._groups_positions.keys()) if self._hue_names is None: - self._xunits = ((max(list(self._xpositions.keys())) + 1) - / len(self._xpositions)) + self._axis_units = ((max(list(self._groups_positions.keys())) + 1) + / len(self._groups_positions)) - self._xranges = { - (pos - self._xunits / 2, pos + self._xunits / 2, pos): group_name - for pos, group_name in self._xpositions.items()} + self._axis_ranges = { + (pos - self._axis_units / 2, + pos + self._axis_units / 2, + pos): group_name + for pos, group_name in self._groups_positions.items()} @property - def xpositions(self): - return self._xpositions + def axis_positions(self): + return self._groups_positions @property - def xunits(self): - return self._xunits + def axis_units(self): + return self._axis_units - def get_xpos_location(self, pos): + def get_axis_pos_location(self, pos): """ Finds the x-axis location of a categorical variable """ - for xrange in self._xranges: - if (pos >= xrange[0]) & (pos <= xrange[1]): - return xrange[2] + for axis_range in self._axis_ranges: + if (pos >= axis_range[0]) & (pos <= axis_range[1]): + return axis_range[2] - def get_group_x_position(self, group): + def get_group_axis_position(self, group): """ group_name can be either a name "cat" or a tuple ("cat", "hue") """ @@ -64,5 +66,5 @@ def get_group_x_position(self, group): group_pos = self._plotter.group_names.index(cat) + hue_offset return group_pos - def find_closest(self, xpos): - return get_closest(list(self._xpositions_list), xpos) + def find_closest(self, pos): + return get_closest(list(self._groups_positions_list), pos) diff --git a/statannotations/_Plotter.py b/statannotations/_Plotter.py index 9dac681..5223865 100644 --- a/statannotations/_Plotter.py +++ b/statannotations/_Plotter.py @@ -8,7 +8,7 @@ from matplotlib.collections import PathCollection from matplotlib.patches import Rectangle -from statannotations._Xpositions import _XPositions +from statannotations._GroupsPositions import _GroupsPositions from statannotations.utils import check_not_none, check_order_in_data, \ check_pairs_in_data, render_collection, check_is_in, remove_null @@ -18,16 +18,18 @@ class _Plotter: - def __init__(self, ax, pairs, data=None, x=None, hue=None, - order=None, hue_order=None, verbose=False): + def __init__(self, ax, pairs, data=None, x=None, y=None, hue=None, + order=None, hue_order=None, verbose=False, plot_params): self.ax = ax self._fig = plt.gcf() check_not_none("pairs", pairs) - check_order_in_data(data, x, order) - check_pairs_in_data(pairs, data, x, hue, hue_order) + group_coord = y if plot_params.get("orient") == "h" else x + check_order_in_data(data, group_coord, order) + check_pairs_in_data(pairs, data, group_coord, hue, hue_order) self.pairs = pairs self._struct_pairs = None self.verbose = verbose + self.orient = plot_params.get("orient", "v") def get_transform_func(self, kind: str): """ @@ -48,9 +50,11 @@ def get_transform_func(self, kind: str): if kind == 'pix_to_ax': return self.ax.transAxes.inverted() - data_to_ax = \ - self.ax.transData + self.ax.get_xaxis_transform().inverted() + transform = {'v': self.ax.get_xaxis_transform, + 'h': self.ax.get_yaxis_transform}[self.orient] + data_to_ax = \ + self.ax.transData + transform().inverted() if kind == 'data_to_ax': return data_to_ax @@ -75,34 +79,35 @@ def __init__(self, ax, pairs, plot='boxplot', data=None, x=None, y=None, hue=None, order=None, hue_order=None, verbose=False, plot_params): - _Plotter.__init__(self, ax, pairs, data, x, hue, order, hue_order, - verbose) + _Plotter.__init__(self, ax, pairs, data, x, y, hue, order, hue_order, + verbose, plot_params) self.check_plot_is_implemented(plot) - self.plot = plot self.plotter = self._get_plotter(plot, x, y, hue, data, order, hue_order, *plot_params) self.group_names, self.labels = self._get_group_names_and_labels() - self.xpositions = _XPositions(self.plotter, self.group_names) + self.groups_positions = _GroupsPositions(self.plotter, + self.group_names) self.reordering = None - self.ymaxes = self._generate_ymaxes() + self.value_maxes = self._generate_value_maxes() self.structs = self._get_structs() self.pairs = pairs self._struct_pairs = self._get_group_struct_pairs() - self._y_stack_arr = np.array( - [[struct['x'], struct['ymax'], 0] for struct in self.structs] + self._value_stack_arr = np.array( + [[struct['group_coord'], struct['value_max'], 0] + for struct in self.structs] ).T self.reordering = None self._sort_group_struct_pairs() @property - def y_stack_arr(self): - return self._y_stack_arr + def value_stack_arr(self): + return self._value_stack_arr # noinspection PyProtectedMember def _get_plotter(self, plot, x, y, hue, data, order, hue_order, @@ -187,7 +192,7 @@ def _get_group_names_and_labels(self): return group_names, labels - def _generate_ymaxes(self): + def _generate_value_maxes(self): """ given plotter and the names of two categorical variables, returns highest y point drawn between those two variables before @@ -197,38 +202,40 @@ def _generate_ymaxes(self): (eg, error bars and/or bar charts). """ - ymaxes = {name: 0 for name in self.group_names} + value_maxes = {name: 0 for name in self.group_names} data_to_ax = self.get_transform_func('data_to_ax') if self.plot == 'violinplot': - ymaxes = self._get_ymaxes_violin(ymaxes, data_to_ax) + value_maxes = self._get_value_maxes_violin(value_maxes, data_to_ax) else: for child in self.ax.get_children(): - xname, ypos = self._get_child_ypos(child, data_to_ax) + group_name, value_pos = self._get_value_pos(child, data_to_ax) - if ypos is not None and ypos > ymaxes[xname]: - ymaxes[xname] = ypos + if value_pos is not None and value_pos > value_maxes[group_name]: + value_maxes[group_name] = value_pos - return ymaxes + return value_maxes def _get_structs(self): structs = [ { 'group': group_name, 'label': self.labels[b_idx], - 'x': self.xpositions.get_group_x_position(group_name), + 'group_coord': (self.groups_positions + .get_group_axis_position(group_name)), 'group_data': self._get_group_data(group_name), - 'ymax': self.ymaxes[group_name] - } for b_idx, group_name in enumerate(self.group_names)] + 'value_max': self.value_maxes[group_name] + } + for b_idx, group_name in enumerate(self.group_names)] - # Sort the group data structures by position along the x axis - structs = sorted(structs, key=lambda struct: struct['x']) + # Sort the group data structures by position along the groups axis + structs = sorted(structs, key=lambda struct: struct['group_coord']) - # Add the index position in the list of groups along the x axis - structs = [dict(struct, xi=i) + # Add the index position in the list of groups along the groups axis + structs = [dict(struct, group_i=i) for i, struct in enumerate(structs)] return structs @@ -246,7 +253,7 @@ def _get_group_struct_pairs(self): group_struct2 = dict(group_structs_dic[group2], i_group_pair=i_group_pair) - if group_struct1['x'] <= group_struct2['x']: + if group_struct1['group_coord'] <= group_struct2['group_coord']: group_struct_pairs.append((group_struct1, group_struct2)) else: @@ -289,62 +296,79 @@ def _get_group_data(self, group_name): return group_data - def _get_ymaxes_violin(self, ymaxes, data_to_ax): + def _get_value_maxes_violin(self, value_maxes, data_to_ax): for group_idx, group_name in enumerate(self.plotter.group_names): if self.plotter.hue_names: for hue_idx, hue_name in enumerate(self.plotter.hue_names): - ypos = max(self.plotter.support[group_idx][hue_idx]) - ymaxes[(group_name, hue_name)] = \ - data_to_ax.transform((0, ypos))[1] + value_pos = max(self.plotter.support[group_idx][hue_idx]) + value_maxes[(group_name, hue_name)] = \ + data_to_ax.transform((0, value_pos))[1] else: - ypos = max(self.plotter.support[group_idx]) - ymaxes[group_name] = data_to_ax.transform((0, ypos))[1] - return ymaxes + value_pos = max(self.plotter.support[group_idx]) + value_maxes[group_name] = data_to_ax.transform((0, + value_pos))[1] + return value_maxes - def _get_child_ypos(self, child, data_to_ax): + def _get_value_pos(self, child, data_to_ax): if (type(child) == PathCollection and len(child.properties()['offsets'])): - return self._get_ypos_for_path_collection( + return self._get_value_pos_for_path_collection( child, data_to_ax) elif type(child) in (lines.Line2D, Rectangle): - return self._get_ypos_for_line2d_or_rectangle( + return self._get_value_pos_for_line2d_or_rectangle( child, data_to_ax) return None, None - def _get_ypos_for_path_collection(self, child, data_to_ax): - ymax = child.properties()['offsets'][:, 1].max() - xpos = float(np.round(np.nanmean( - child.properties()['offsets'][:, 0]), 1)) - if xpos not in self.xpositions.xpositions: + def _get_value_pos_for_path_collection(self, child, data_to_ax): + group_coord = {"v": 0, "h": 1}[self.orient] + value_coord = (group_coord + 1) % 2 + + direction = {"v": 1, "h": -1}[self.orient] + + value_max = child.properties()['offsets'][:, value_coord].max() + group_pos = float(np.round(np.nanmean( + child.properties()['offsets'][:, group_coord]), 1)) + + if group_pos not in self.groups_positions.axis_positions: if self.verbose: warnings.warn( "Invalid x-position found. Are the same parameters passed " "to seaborn and statannotations calls? or are there few " "data points?") - xpos = self.xpositions.find_closest(xpos) + group_pos = self.groups_positions.find_closest(group_pos) + + group_name = self.groups_positions.axis_positions[group_pos] - xname = self.xpositions.xpositions[xpos] - ypos = data_to_ax.transform((0, ymax))[1] + value_pos = data_to_ax.transform((0, value_max)[::direction]) - return xname, ypos + return group_name, value_pos[value_coord] - def _get_ypos_for_line2d_or_rectangle(self, child, data_to_ax): + def _get_value_pos_for_line2d_or_rectangle(self, child, data_to_ax): bbox = self.ax.transData.inverted().transform( child.get_window_extent(self.fig.canvas.get_renderer())) - if (bbox[:, 0].max() - bbox[:, 0].min()) > 1.1self.xpositions.xunits: + group_coord = {"v": 0, "h": 1}[self.orient] + direction = {"v": 1, "h": -1}[self.orient] + + value_coord = (group_coord + 1) % 2 + + if ((bbox[:, group_coord].max() - bbox[:, group_coord].min()) + > 1.1 * self.groups_positions.axis_units): return None, None - raw_xpos = np.round(bbox[:, 0].mean(), 1) - xpos = self.xpositions.get_xpos_location(raw_xpos) - if xpos not in self.xpositions.xpositions: + + raw_group_pos = np.round(bbox[:, group_coord].mean(), 1) + group_pos = self.groups_positions.get_axis_pos_location(raw_group_pos) + + if group_pos not in self.groups_positions.axis_positions: return None, None - xname = self.xpositions.xpositions[xpos] - ypos = bbox[:, 1].max() - ypos = data_to_ax.transform((0, ypos))[1] - return xname, ypos + group_name = self.groups_positions.axis_positions[group_pos] + value_pos = bbox[:, value_coord].max() + + value_pos = data_to_ax.transform((0, value_pos)[::direction]) + return group_name, value_pos[value_coord] def _sort_group_struct_pairs(self): # Draw first the annotations with the shortest between-groups distance, @@ -357,7 +381,8 @@ def _sort_group_struct_pairs(self): @staticmethod def _absolute_group_struct_pair_in_tuple_x_diff(group_struct_pair): - return abs(group_struct_pair[1][1]['x'] - group_struct_pair[1][0]['x']) + return abs(group_struct_pair[1][1]['group_coord'] + - group_struct_pair[1][0]['group_coord']) @staticmethod def check_plot_is_implemented(plot, engine="seaborn"): @@ -377,3 +402,8 @@ def fix_and_warn(dodge, hue, plot): "Implicitly setting dodge to True as it is necessary in " "statannotations. It must have been True for the seaborn " "call to yield consistent results when using `hue`.") + + def get_value_lim(self): + if self.orient == 'v': + return self.ax.get_ylim() + return self.ax.get_xlim() diff --git a/statannotations/_version.py b/statannotations/_version.py index 6a9beea..3d26edf 100644 --- a/statannotations/_version.py +++ b/statannotations/_version.py @@ -1,1 +1,1 @@ -__version__ = "0.4.0" +__version__ = "0.4.1" diff --git a/statannotations/utils.py b/statannotations/utils.py index c7d44fc..934239b 100644 --- a/statannotations/utils.py +++ b/statannotations/utils.py @@ -75,19 +75,19 @@ def _check_pairs_in_data_no_hue(pairs: Union[list, tuple], def _check_pairs_in_data_with_hue(pairs: Union[list, tuple], data: Union[List[list], pd.DataFrame] = None, - x: Union[str, list] = None, + group_coord: Union[str, list] = None, hue: str = None) -> set: - x_values = get_x_values(data, x) - seen_x_values = set() + x_values = get_x_values(data, group_coord) + seen_group_values = set() hue_values = set(data[hue].unique()) - for x_value, hue_value in itertools.chain(itertools.chain(pairs)): - if x_value not in seen_x_values and x_value not in x_values: - raise ValueError(f"Missing x value `{x_value}` in {x}" + for group_value, hue_value in itertools.chain(itertools.chain(pairs)): + if group_value not in seen_group_values and group_value not in x_values: + raise ValueError(f"Missing group value `{group_value}` in {group_coord}" f" (specified in `pairs`)") - seen_x_values.add(x_value) + seen_group_values.add(group_value) if hue_value not in hue_values: raise ValueError(f"Missing hue value `{hue_value}` in {hue}" @@ -108,7 +108,7 @@ def _check_hue_order_in_data(hue, hue_values: set, def check_pairs_in_data(pairs: Union[list, tuple], data: Union[List[list], pd.DataFrame] = None, - x: Union[str, list] = None, + coord: Union[str, list] = None, hue: str = None, hue_order: List[str] = None): """ @@ -116,9 +116,9 @@ def check_pairs_in_data(pairs: Union[list, tuple], """ if hue is None and hue_order is None: - _check_pairs_in_data_no_hue(pairs, data, x) + _check_pairs_in_data_no_hue(pairs, data, coord) else: - hue_values = _check_pairs_in_data_with_hue(pairs, data, x, hue) + hue_values = _check_pairs_in_data_with_hue(pairs, data, coord, hue) _check_hue_order_in_data(hue, hue_values, hue_order) diff --git a/usage/example.ipynb b/usage/example.ipynb index 10b91f9..91a3e66 100644 --- a/usage/example.ipynb +++ b/usage/example.ipynb @@ -231,7 +231,7 @@ }, { "data": { - "text/plain": "(<matplotlib.axes._subplots.AxesSubplot at 0x7fedb54c7748>,\n [<statannotations.Annotation.Annotation at 0x7fedb4233ac8>,\n <statannotations.Annotation.Annotation at 0x7fedb4233d30>,\n <statannotations.Annotation.Annotation at 0x7fedb4233320>])" + "text/plain": "(<matplotlib.axes._subplots.AxesSubplot at 0x7fc28184f6a0>,\n [<statannotations.Annotation.Annotation at 0x7fc28164aeb8>,\n <statannotations.Annotation.Annotation at 0x7fc28183d6d8>,\n <statannotations.Annotation.Annotation at 0x7fc282ada5c0>])" }, "execution_count": 6, "metadata": {}, @@ -672,7 +672,7 @@ }, { "data": { - "text/plain": "(<matplotlib.axes._subplots.AxesSubplot at 0x7fedb4460898>,\n [<statannotations.Annotation.Annotation at 0x7fedb40b4400>,\n <statannotations.Annotation.Annotation at 0x7fedb715ccf8>,\n <statannotations.Annotation.Annotation at 0x7fedb6c8ef28>])" + "text/plain": "(<matplotlib.axes._subplots.AxesSubplot at 0x7fc284d10f60>,\n [<statannotations.Annotation.Annotation at 0x7fc281c837b8>,\n <statannotations.Annotation.Annotation at 0x7fc2817bfc18>,\n <statannotations.Annotation.Annotation at 0x7fc284d8e518>])" }, "execution_count": 15, "metadata": {}, @@ -947,7 +947,7 @@ { "data": { "text/plain": "<Figure size 864x432 with 1 Axes>", - "image/png": "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\n" + "image/png": "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\n" }, "metadata": {}, "output_type": "display_data" @@ -1001,7 +1001,7 @@ { "cell_type": "markdown", "source": [ - "And a similar swarm plot, with a different format for pvalues" + "With a horizontal orientation" ], "metadata": { "collapsed": false @@ -1010,6 +1010,74 @@ { "cell_type": "code", "execution_count": 21, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "p-value annotation legend:\n", + " ns: p <= 1.00e+00\n", + " : 1.00e-02 < p <= 5.00e-02\n", + " : 1.00e-03 < p <= 1.00e-02\n", + " : 1.00e-04 < p <= 1.00e-03\n", + " : p <= 1.00e-04\n", + "\n", + "H1N1_Nonsynonymous vs. H1N1_Synonymous: Mann-Whitney-Wilcoxon test two-sided, P_val:5.014e-04 U_stat=2.624e+03\n", + "H3N2_Nonsynonymous vs. H3N2_Synonymous: Mann-Whitney-Wilcoxon test two-sided, P_val:1.294e-03 U_stat=1.535e+04\n", + "Influenza B_Nonsynonymous vs. Influenza B_Synonymous: Mann-Whitney-Wilcoxon test two-sided, P_val:2.026e-01 U_stat=3.340e+02\n" + ] + }, + { + "data": { + "text/plain": "(<matplotlib.axes._subplots.AxesSubplot at 0x7fc283af2668>,\n [<statannotations.Annotation.Annotation at 0x7fc283bf3780>,\n <statannotations.Annotation.Annotation at 0x7fc283be1dd8>,\n <statannotations.Annotation.Annotation at 0x7fc283bc3ef0>])" + }, + "execution_count": 21, + "metadata": {}, + "output_type": "execute_result" + }, + { + "data": { + "text/plain": "<Figure size 864x432 with 1 Axes>", + "image/png": "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\n" + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "fig, ax = plt.subplots(1,1, figsize=(12,6))\n", + "fig_args_horiz = {\n", + " fig_args,\n", + " 'y': 'subtype',\n", + " 'x': 'mutation_rate_samp'\n", + "}\n", + "ax.set_xscale('log')\n", + "sns.stripplot(ax=ax, orient='h', fig_args_horiz)\n", + "annotator.new_plot(ax, plot='swarmplot', orient='h', fig_args_horiz)\n", + "annotator.apply_and_annotate()" + ], + "metadata": { + "collapsed": false, + "pycharm": { + "name": "#%%\n" + } + } + }, + { + "cell_type": "markdown", + "source": [ + "Back to vertical, with a different format for pvalues" + ], + "metadata": { + "collapsed": false, + "pycharm": { + "name": "#%% md\n" + } + } + }, + { + "cell_type": "code", + "execution_count": 22, "outputs": [ { "name": "stdout", @@ -1060,7 +1128,7 @@ }, { "cell_type": "code", - "execution_count": 22, + "execution_count": 23, "outputs": [ { "name": "stdout", @@ -1101,7 +1169,7 @@ }, { "cell_type": "code", - "execution_count": 22, + "execution_count": 23, "outputs": [], "source": [], "metadata": {	Horizontal orientation There seems to have a great demand for it (https://github.com/webermarcolivier/statannot/issues/54). Let's implement it. Maybe this PR for statannot will still be useful (https://github.com/webermarcolivier/statannot/pull/53)?		2021-07-25T13:04:36	[]	[]
trevismd/statannotations	trevismd__statannotations-25	4728ce3b8c2475a73b7819e55dca36584234a7dc	diff --git a/.github/workflows/python-package.yml b/.github/workflows/python-package.yml index c581d62..8504874 100644 --- a/.github/workflows/python-package.yml +++ b/.github/workflows/python-package.yml @@ -5,9 +5,9 @@ name: Python package on: push: - branches: [ master, dev, v0.2, v0.3, refactoring] + branches: [ master, dev, v0.2, v0.3, v0.4] pull_request: - branches: [ master, dev, v0.2, v0.3 ] + branches: [ master, dev, v0.2, v0.3, v0.4] jobs: build: diff --git a/CHANGELOG.md b/CHANGELOG.md index 54f959b..80407ff 100644 --- a/CHANGELOG.md +++ b/CHANGELOG.md @@ -1,20 +1,30 @@ ## v0.4 ### v0.4.0 - - Major refactoring, change to an Annotator `class` to prepare and add annotations distinctly + - Major refactoring, change to an Annotator `class` to prepare and add + annotations in separate function (now method) calls. - Support for `violinplot` - - Fixes in rendering of non significant tests results after multiple comparisons correction + - Fixes in rendering of non-significant tests results after multiple + comparisons correction - Fix the printout of the star annotation legend - - Fix the bug when providing a dataframe with categories as different columns. + - Fix the failure when providing a dataframe with categories as different + columns. + - Fix a bug where an incorrect xunits calculation resulted in wrong + association of points within a box, leading to erroneous max y position for + that box. + - Reduced code complexity, more SOLID. - Many unit and integration tests ## v0.3 ### v0.3.2 - Fix `simple` format outputs - - Fix `ImportError` when applying a multiple comparison correction without statsmodels. - - Multiple comparison correction is `None` by default, as `statsmodels` is an additional dependency. + - Fix `ImportError` when applying a multiple comparison correction without + statsmodels. + - Multiple comparison correction is `None` by default, as `statsmodels` is an + additional dependency. ### v0.3.1 - - Added support of functions returning more than the two expected values when used in `StatTest` + - Added support of functions returning more than the two expected values when + used in `StatTest` - Fix version numbers in CHANGELOG ### v0.3.0 @@ -23,10 +33,17 @@ - Refactoring with implementation of `StatTest` ([#7](https://github.com/trevismd/statannotations/pull/5)), removing the `stat_func` parameter, and `test_long_name`. - - Annotations y-positions based on plot coordinates instead of data coordinates - ([#5](https://github.com/trevismd/statannotations/pull/5) by [JosephLalli](https://github.com/JosephLalli), fixes https://github.com/webermarcolivier/statannot/issues/21). + - Annotations y-positions based on plot coordinates instead of data + coordinates + ([#5](https://github.com/trevismd/statannotations/pull/5) by + [JosephLalli](https://github.com/JosephLalli), + fixes https://github.com/webermarcolivier/statannot/issues/21). - Add this CHANGELOG ## v0.2 ### v0.2.8 - - Fix bug on group/box named 0, fixes https://github.com/trevismd/statannotations/issues/10, originally in https://github.com/webermarcolivier/statannot/issues/78. Independently fixed in https://github.com/webermarcolivier/statannot/pull/73 before this issue. + - Fix bug on group/box named 0, fixes + https://github.com/trevismd/statannotations/issues/10, originally in + https://github.com/webermarcolivier/statannot/issues/78. Independently + fixed in https://github.com/webermarcolivier/statannot/pull/73 before this + issue. diff --git a/CONTRIBUTING.md b/CONTRIBUTING.md new file mode 100644 index 0000000..8d04ec8 --- /dev/null +++ b/CONTRIBUTING.md @@ -0,0 +1,21 @@ +## Contributing to statannotations +Thank you for your interest! This package is truly the work of multiple +developers and participating users and this mustn't change. + +Contributions are appreciated in all their forms, such as the opening of new +issues when discovering a (possible) bug, suggestion/upvote new features, +participation to discussions, improvements of the documentation, etc. + +If you wish specifically to submit code changes, please use the following +guide: +- [ ] If it is not an answer to an open issue, please consider opening one + first, even more if the changes are major and/or possibly controversial +- [ ] Make sure to have the dev dependencies installed: `packaging` and + `statsmodels` +- Fork the repository and make a new branch from the latest `dev` commit +- [ ] Make your code modifications, including unit testing if possible. + Please follow PEP8 guidelines. +- [ ] Run the test suite and make sure previous (and new) tests pass +- [ ] Update CHANGELOG.md to describe your changes +- [ ] Open a PR to the `dev` branch, containing at least what was added to the + changelog, as well as a reference to the issue if any diff --git a/README.md b/README.md index 545bfd1..e8f3582 100644 --- a/README.md +++ b/README.md @@ -3,22 +3,23 @@ ## What is it Python package to optionally compute statistical test and add statistical -annotations on plots generated with seaborn. -This branch is now at version 0.4.0-alpha, while the latest release is v0.3.2. +annotations on plots generated with seaborn. ## Derived work -This repository is evolving independently from +This repository is based on [webermarcolivier/statannot](https://github.com/webermarcolivier/statannot) -by Marc Weber. It is based on commit 1835078 of Feb 21, 2020, tagged "v0.2.3". + (commit 1835078 of Feb 21, 2020, tagged "v0.2.3"). Additions/modifications since that version are below represented in bold (previous fixes are not listed). - From version 0.4 onwards (introduction of `Annotator`), `statannot`'s API - is no longer usable in `statannotations` +! From version 0.4.0 onwards (introduction of `Annotator`), `statannot`'s API +is no longer usable in `statannotations`. +Please use the latest v0.3.2 release if you must keep `statannot``s API in your +code, but are looking for bug fixes we have covered. -The statannot interface, at least until its version 0.2.3, is usable in +Indeed, the statannot interface, at least until its version 0.2.3, is usable in statannotations until v.0.3.x, which already provides additional features (see corresponding branch). @@ -54,7 +55,9 @@ corresponding branch). - Optionally, custom p-values can be given as input. In this case, no statistical test is performed, but corrections for multiple testing can be applied. -- And various fixes (see [CHANGELOG.md](https://github.com/trevismd/statannotations/blob/master/CHANGELOG.md)). +- Any text can be used as annotation +- And various fixes (see + [CHANGELOG.md](https://github.com/trevismd/statannotations/blob/master/CHANGELOG.md)). ## Installation @@ -68,7 +71,7 @@ or, after cloning the repository, ```bash pip install . -# OR, to have optional dependencies too (multiple comparisons) +# OR, to have optional dependencies too (multiple comparisons & testing) pip install -r requirements.txt . ``` @@ -89,11 +92,12 @@ df = sns.load_dataset("tips") x = "day" y = "total_bill" order = ['Sun', 'Thur', 'Fri', 'Sat'] -box_pairs=[("Thur", "Fri"), ("Thur", "Sat"), ("Fri", "Sun")] + ax = sns.boxplot(data=df, x=x, y=y, order=order) +pairs=[("Thur", "Fri"), ("Thur", "Sat"), ("Fri", "Sun")] -annotator = Annotator(ax, box_pairs, data=df, x=x, y=y, order=order) +annotator = Annotator(ax, pairs, data=df, x=x, y=y, order=order) annotator.configure(test='Mann-Whitney', text_format='star', loc='outside') annotator.apply_and_annotate() ``` @@ -122,5 +126,5 @@ annotator.apply_and_annotate() In addition to git's history, contributions to statannotations are logged in the changelog. If you don't know where to start, there may be a few ideas in opened issues or -discussion, or something to work for the documentation. -NB: to run the test suite, the `statsmodels` and `packaging` packages are required. +discussion, or something to work for the documentation. +NB: More on [CONTRIBUTING.md](CONTRIBUTING.md) diff --git a/codecov.yml b/codecov.yml index 8b13789..0d54c1d 100644 --- a/codecov.yml +++ b/codecov.yml @@ -1,1 +1,9 @@ +coverage: + status: + project: + default: + threshold: 0.5 # Allow coverage to drop by up to 0.5% in a PR before marking it failed + patch: + default: + threshold: 0.5% \ No newline at end of file diff --git a/coverage.svg b/coverage.svg index b6c4e36..ee07d4c 100644 --- a/coverage.svg +++ b/coverage.svg @@ -15,7 +15,7 @@ <g fill="#fff" text-anchor="middle" font-family="DejaVu Sans,Verdana,Geneva,sans-serif" font-size="11"> <text x="31.5" y="15" fill="#010101" fill-opacity=".3">coverage</text> <text x="31.5" y="14">coverage</text> - <text x="80" y="15" fill="#010101" fill-opacity=".3">95%</text> - <text x="80" y="14">95%</text> + <text x="80" y="15" fill="#010101" fill-opacity=".3">96%</text> + <text x="80" y="14">96%</text> </g> </svg> diff --git a/docs/build/doctrees/environment.pickle b/docs/build/doctrees/environment.pickle index 9e93656..771aba2 100644 Binary files a/docs/build/doctrees/environment.pickle and b/docs/build/doctrees/environment.pickle differ diff --git a/docs/build/doctrees/statannotations.doctree b/docs/build/doctrees/statannotations.doctree index 9b18037..ee0691c 100644 Binary files a/docs/build/doctrees/statannotations.doctree and b/docs/build/doctrees/statannotations.doctree differ diff --git a/docs/build/doctrees/statannotations.stats.doctree b/docs/build/doctrees/statannotations.stats.doctree index 2ab9a38..a4c0dfe 100644 Binary files a/docs/build/doctrees/statannotations.stats.doctree and b/docs/build/doctrees/statannotations.stats.doctree differ diff --git a/docs/build/html/_sources/statannotations.rst.txt b/docs/build/html/_sources/statannotations.rst.txt index d3f9b4e..3a24356 100644 --- a/docs/build/html/_sources/statannotations.rst.txt +++ b/docs/build/html/_sources/statannotations.rst.txt @@ -12,6 +12,14 @@ Subpackages Submodules ---------- +statannotations.Annotation module +--------------------------------- + +.. automodule:: statannotations.Annotation + :members: + :undoc-members: + :show-inheritance: + statannotations.Annotator module -------------------------------- diff --git a/docs/build/html/_static/alabaster.css b/docs/build/html/_static/alabaster.css index 0eddaeb..cac2609 100644 --- a/docs/build/html/_static/alabaster.css +++ b/docs/build/html/_static/alabaster.css @@ -638,7 +638,7 @@ a:hover tt, a:hover code { /* Make nested-list/multi-paragraph items look better in Releases changelog * pages. Without this, docutils' magical list fuckery causes inconsistent - * formatting between different release sub-lists. + * formatter between different release sub-lists. / div#changelog > div.section > ul > li > p:only-child { margin-bottom: 0; diff --git a/docs/build/html/genindex.html b/docs/build/html/genindex.html index 9db3161..935f030 100644 --- a/docs/build/html/genindex.html +++ b/docs/build/html/genindex.html @@ -177,15 +177,19 @@ <h1 id="index">Index</h1> <h2 id="A">A</h2> <table style="width: 100%" class="indextable genindextable"><tr> <td style="width: 33%; vertical-align: top;"><ul> - <li><a href="statannotations.html#statannotations.Annotator.Annotator.absolute_box_struct_pair_in_tuple_x_diff">absolute_box_struct_pair_in_tuple_x_diff() (statannotations.Annotator.Annotator static method)</a> -</li> <li><a href="statannotations.html#statannotations.Annotator.Annotator.alpha">alpha (statannotations.Annotator.Annotator property)</a> </li> <li><a href="statannotations.html#statannotations.Annotator.Annotator.annotate">annotate() (statannotations.Annotator.Annotator method)</a> +</li> + <li><a href="statannotations.html#statannotations.Annotator.Annotator.annotate_custom_annotations">annotate_custom_annotations() (statannotations.Annotator.Annotator method)</a> +</li> + <li><a href="statannotations.html#statannotations.Annotation.Annotation">Annotation (class in statannotations.Annotation)</a> </li> </ul></td> <td style="width: 33%; vertical-align: top;"><ul> <li><a href="statannotations.html#statannotations.Annotator.Annotator">Annotator (class in statannotations.Annotator)</a> +</li> + <li><a href="statannotations.stats.html#statannotations.stats.ComparisonsCorrection.ComparisonsCorrection.apply">apply() (statannotations.stats.ComparisonsCorrection.ComparisonsCorrection method)</a> </li> <li><a href="statannotations.html#statannotations.Annotator.Annotator.apply_and_annotate">apply_and_annotate() (statannotations.Annotator.Annotator method)</a> </li> @@ -202,10 +206,6 @@ <h2 id="C">C</h2> <table style="width: 100%" class="indextable genindextable"><tr> <td style="width: 33%; vertical-align: top;"><ul> <li><a href="statannotations.stats.html#statannotations.stats.utils.check_alpha">check_alpha() (in module statannotations.stats.utils)</a> -</li> - <li><a href="statannotations.html#statannotations.utils.check_box_pairs_in_data">check_box_pairs_in_data() (in module statannotations.utils)</a> -</li> - <li><a href="statannotations.html#statannotations.Annotator.check_implemented_plotters">check_implemented_plotters() (in module statannotations.Annotator)</a> </li> <li><a href="statannotations.html#statannotations.utils.check_is_in">check_is_in() (in module statannotations.utils)</a> </li> @@ -214,21 +214,27 @@ <h2 id="C">C</h2> <li><a href="statannotations.stats.html#statannotations.stats.utils.check_num_comparisons">check_num_comparisons() (in module statannotations.stats.utils)</a> </li> <li><a href="statannotations.html#statannotations.utils.check_order_in_data">check_order_in_data() (in module statannotations.utils)</a> +</li> + <li><a href="statannotations.html#statannotations.utils.check_pairs_in_data">check_pairs_in_data() (in module statannotations.utils)</a> </li> <li><a href="statannotations.stats.html#statannotations.stats.utils.check_pvalues">check_pvalues() (in module statannotations.stats.utils)</a> </li> - </ul></td> - <td style="width: 33%; vertical-align: top;"><ul> <li><a href="statannotations.stats.html#statannotations.stats.ComparisonsCorrection.check_valid_correction_name">check_valid_correction_name() (in module statannotations.stats.ComparisonsCorrection)</a> </li> + </ul></td> + <td style="width: 33%; vertical-align: top;"><ul> <li><a href="statannotations.html#statannotations.utils.check_valid_text_format">check_valid_text_format() (in module statannotations.utils)</a> </li> <li><a href="statannotations.html#statannotations.Annotator.Annotator.comparisons_correction">comparisons_correction (statannotations.Annotator.Annotator property)</a> </li> <li><a href="statannotations.stats.html#statannotations.stats.ComparisonsCorrection.ComparisonsCorrection">ComparisonsCorrection (class in statannotations.stats.ComparisonsCorrection)</a> </li> - <li><a href="statannotations.html#statannotations.PValueFormat.PValueFormat.config">config() (statannotations.PValueFormat.PValueFormat method)</a> + <li><a href="statannotations.html#statannotations.PValueFormat.Formatter.config">config() (statannotations.PValueFormat.Formatter method)</a> + + <ul> + <li><a href="statannotations.html#statannotations.PValueFormat.PValueFormat.config">(statannotations.PValueFormat.PValueFormat method)</a> </li> + </ul></li> <li><a href="statannotations.html#statannotations.Annotator.Annotator.configure">configure() (statannotations.Annotator.Annotator method)</a> </li> <li><a href="statannotations.stats.html#statannotations.stats.StatResult.StatResult.corrected_significance">corrected_significance (statannotations.stats.StatResult.StatResult property)</a> @@ -249,11 +255,23 @@ <h2 id="D">D</h2> <h2 id="F">F</h2> <table style="width: 100%" class="indextable genindextable"><tr> <td style="width: 33%; vertical-align: top;"><ul> - <li><a href="statannotations.html#statannotations.PValueFormat.PValueFormat.format_result">format_result() (statannotations.PValueFormat.PValueFormat method)</a> + <li><a href="statannotations.html#statannotations.Annotator.Annotator.fig">fig (statannotations.Annotator.Annotator property)</a> +</li> + <li><a href="statannotations.html#statannotations.PValueFormat.Formatter.format_data">format_data() (statannotations.PValueFormat.Formatter method)</a> + + <ul> + <li><a href="statannotations.html#statannotations.PValueFormat.PValueFormat.format_data">(statannotations.PValueFormat.PValueFormat method)</a> </li> + </ul></li> </ul></td> <td style="width: 33%; vertical-align: top;"><ul> - <li><a href="statannotations.stats.html#statannotations.stats.StatResult.StatResult.formatted_output">formatted_output (statannotations.stats.StatResult.StatResult property)</a> + <li><a href="statannotations.html#statannotations.Annotation.Annotation.formatted_output">formatted_output (statannotations.Annotation.Annotation property)</a> + + <ul> + <li><a href="statannotations.stats.html#statannotations.stats.StatResult.StatResult.formatted_output">(statannotations.stats.StatResult.StatResult property)</a> +</li> + </ul></li> + <li><a href="statannotations.html#statannotations.PValueFormat.Formatter">Formatter (class in statannotations.PValueFormat)</a> </li> <li><a href="statannotations.stats.html#statannotations.stats.StatTest.StatTest.from_library">from_library() (statannotations.stats.StatTest.StatTest static method)</a> </li> @@ -265,7 +283,7 @@ <h2 id="G">G</h2> <td style="width: 33%; vertical-align: top;"><ul> <li><a href="statannotations.html#statannotations.Annotator.Annotator.get_annotations_text">get_annotations_text() (statannotations.Annotator.Annotator method)</a> </li> - <li><a href="statannotations.html#statannotations.Annotator.Annotator.get_basic_plot">get_basic_plot() (statannotations.Annotator.Annotator static method)</a> + <li><a href="statannotations.html#statannotations.utils.get_closest">get_closest() (in module statannotations.utils)</a> </li> <li><a href="statannotations.html#statannotations.Annotator.Annotator.get_configuration">get_configuration() (statannotations.Annotator.Annotator method)</a> @@ -278,6 +296,8 @@ <h2 id="G">G</h2> <li><a href="statannotations.stats.html#statannotations.stats.ComparisonsCorrection.get_correction_parameters">get_correction_parameters() (in module statannotations.stats.ComparisonsCorrection)</a> </li> <li><a href="statannotations.stats.html#statannotations.stats.utils.get_num_comparisons">get_num_comparisons() (in module statannotations.stats.utils)</a> +</li> + <li><a href="statannotations.html#statannotations.Annotator.Annotator.get_offset_func">get_offset_func() (statannotations.Annotator.Annotator class method)</a> </li> <li><a href="statannotations.stats.html#statannotations.stats.ComparisonsCorrection.get_validated_comparisons_correction">get_validated_comparisons_correction() (in module statannotations.stats.ComparisonsCorrection)</a> </li> @@ -310,6 +330,8 @@ <h2 id="M">M</h2> <ul> <li><a href="statannotations.html#module-statannotations">statannotations</a> +</li> + <li><a href="statannotations.html#module-statannotations.Annotation">statannotations.Annotation</a> </li> <li><a href="statannotations.html#module-statannotations.Annotator">statannotations.Annotator</a> </li> @@ -346,6 +368,8 @@ <h2 id="N">N</h2> <h2 id="P">P</h2> <table style="width: 100%" class="indextable genindextable"><tr> <td style="width: 33%; vertical-align: top;"><ul> + <li><a href="statannotations.html#statannotations.Annotation.Annotation.print_labels_and_content">print_labels_and_content() (statannotations.Annotation.Annotation method)</a> +</li> <li><a href="statannotations.html#statannotations.PValueFormat.PValueFormat.print_legend_if_used">print_legend_if_used() (statannotations.PValueFormat.PValueFormat method)</a> </li> <li><a href="statannotations.html#statannotations.Annotator.Annotator.print_pvalue_legend">print_pvalue_legend() (statannotations.Annotator.Annotator method)</a> @@ -407,6 +431,13 @@ <h2 id="S">S</h2> <ul> <li><a href="statannotations.html#module-statannotations">module</a> +</li> + </ul></li> + <li> + statannotations.Annotation + + <ul> + <li><a href="statannotations.html#module-statannotations.Annotation">module</a> </li> </ul></li> <li> @@ -495,6 +526,8 @@ <h2 id="T">T</h2> </li> </ul></td> <td style="width: 33%; vertical-align: top;"><ul> + <li><a href="statannotations.html#statannotations.Annotation.Annotation.text">text (statannotations.Annotation.Annotation property)</a> +</li> <li><a href="statannotations.html#statannotations.PValueFormat.PValueFormat.text_format">text_format (statannotations.PValueFormat.PValueFormat property)</a> </li> </ul></td> @@ -504,6 +537,10 @@ <h2 id="V">V</h2> <table style="width: 100%" class="indextable genindextable"><tr> <td style="width: 33%; vertical-align: top;"><ul> <li><a href="statannotations.html#statannotations.Annotator.Annotator.validate_test_short_name">validate_test_short_name() (statannotations.Annotator.Annotator method)</a> +</li> + </ul></td> + <td style="width: 33%; vertical-align: top;"><ul> + <li><a href="statannotations.html#statannotations.Annotator.Annotator.verbose">verbose (statannotations.Annotator.Annotator property)</a> </li> </ul></td> </tr></table> diff --git a/docs/build/html/modules.html b/docs/build/html/modules.html index aff2f2a..b54367e 100644 --- a/docs/build/html/modules.html +++ b/docs/build/html/modules.html @@ -179,6 +179,7 @@ <h1>statannotations<a class="headerlink" href="#statannotations" title="Permalin </ul> </li> <li class="toctree-l2"><a class="reference internal" href="statannotations.html#submodules">Submodules</a></li> +<li class="toctree-l2"><a class="reference internal" href="statannotations.html#module-statannotations.Annotation">statannotations.Annotation module</a></li> <li class="toctree-l2"><a class="reference internal" href="statannotations.html#module-statannotations.Annotator">statannotations.Annotator module</a></li> <li class="toctree-l2"><a class="reference internal" href="statannotations.html#module-statannotations.PValueFormat">statannotations.PValueFormat module</a></li> <li class="toctree-l2"><a class="reference internal" href="statannotations.html#module-statannotations.format_annotations">statannotations.format_annotations module</a></li> diff --git a/docs/build/html/objects.inv b/docs/build/html/objects.inv index 8a3e2e4..0d82dd7 100644 Binary files a/docs/build/html/objects.inv and b/docs/build/html/objects.inv differ diff --git a/docs/build/html/py-modindex.html b/docs/build/html/py-modindex.html index 78c0a07..b4c7f07 100644 --- a/docs/build/html/py-modindex.html +++ b/docs/build/html/py-modindex.html @@ -171,6 +171,11 @@ <h1>Python Module Index</h1> <td> <a href="statannotations.html#module-statannotations"><code class="xref">statannotations</code></a></td><td> <em></em></td></tr> + <tr class="cg-1"> + <td></td> + <td>    + <a href="statannotations.html#module-statannotations.Annotation"><code class="xref">statannotations.Annotation</code></a></td><td> + <em></em></td></tr> <tr class="cg-1"> <td></td> <td>    diff --git a/docs/build/html/searchindex.js b/docs/build/html/searchindex.js index 1aa74be..38da9ed 100644 --- a/docs/build/html/searchindex.js +++ b/docs/build/html/searchindex.js @@ -1,1 +1,1 @@ -Search.setIndex({docnames:["index","modules","setup","statannotations","statannotations.stats"],envversion:{"sphinx.domains.c":2,"sphinx.domains.changeset":1,"sphinx.domains.citation":1,"sphinx.domains.cpp":3,"sphinx.domains.index":1,"sphinx.domains.javascript":2,"sphinx.domains.math":2,"sphinx.domains.python":3,"sphinx.domains.rst":2,"sphinx.domains.std":2,sphinx:56},filenames:["index.rst","modules.rst","setup.rst","statannotations.rst","statannotations.stats.rst"],objects:{"":{statannotations:[3,0,0,"-"]},"statannotations.Annotator":{Annotator:[3,1,1,""],check_implemented_plotters:[3,4,1,""]},"statannotations.Annotator.Annotator":{absolute_box_struct_pair_in_tuple_x_diff:[3,2,1,""],alpha:[3,3,1,""],annotate:[3,2,1,""],apply_and_annotate:[3,2,1,""],apply_test:[3,2,1,""],comparisons_correction:[3,3,1,""],configure:[3,2,1,""],get_annotations_text:[3,2,1,""],get_basic_plot:[3,2,1,""],get_configuration:[3,2,1,""],has_type0_comparisons_correction:[3,2,1,""],loc:[3,3,1,""],new_plot:[3,2,1,""],print_pvalue_legend:[3,2,1,""],pvalue_format:[3,3,1,""],reset_configuration:[3,2,1,""],set_custom_annotations:[3,2,1,""],set_pvalues:[3,2,1,""],set_pvalues_and_annotate:[3,2,1,""],test:[3,3,1,""],validate_test_short_name:[3,2,1,""]},"statannotations.PValueFormat":{PValueFormat:[3,1,1,""],sort_pvalue_thresholds:[3,4,1,""]},"statannotations.PValueFormat.PValueFormat":{config:[3,2,1,""],format_result:[3,2,1,""],get_configuration:[3,2,1,""],print_legend_if_used:[3,2,1,""],pvalue_format_string:[3,3,1,""],pvalue_thresholds:[3,3,1,""],simple_format_string:[3,3,1,""],text_format:[3,3,1,""]},"statannotations.format_annotations":{pval_annotation_text:[3,4,1,""],simple_text:[3,4,1,""]},"statannotations.stats":{ComparisonsCorrection:[4,0,0,"-"],StatResult:[4,0,0,"-"],StatTest:[4,0,0,"-"],test:[4,0,0,"-"],utils:[4,0,0,"-"]},"statannotations.stats.ComparisonsCorrection":{ComparisonsCorrection:[4,1,1,""],check_valid_correction_name:[4,4,1,""],get_correction_parameters:[4,4,1,""],get_validated_comparisons_correction:[4,4,1,""]},"statannotations.stats.ComparisonsCorrection.ComparisonsCorrection":{document:[4,2,1,""]},"statannotations.stats.StatResult":{StatResult:[4,1,1,""]},"statannotations.stats.StatResult.StatResult":{corrected_significance:[4,3,1,""],correction_method:[4,3,1,""],formatted_output:[4,3,1,""],significance_suffix:[4,3,1,""]},"statannotations.stats.StatTest":{StatTest:[4,1,1,""],wilcoxon:[4,4,1,""]},"statannotations.stats.StatTest.StatTest":{from_library:[4,2,1,""],short_name:[4,3,1,""]},"statannotations.stats.test":{apply_test:[4,4,1,""]},"statannotations.stats.utils":{check_alpha:[4,4,1,""],check_num_comparisons:[4,4,1,""],check_pvalues:[4,4,1,""],get_num_comparisons:[4,4,1,""],return_results:[4,4,1,""]},"statannotations.utils":{check_box_pairs_in_data:[3,4,1,""],check_is_in:[3,4,1,""],check_not_none:[3,4,1,""],check_order_in_data:[3,4,1,""],check_valid_text_format:[3,4,1,""],get_x_values:[3,4,1,""],raise_expected_got:[3,4,1,""],remove_null:[3,4,1,""],render_collection:[3,4,1,""]},statannotations:{Annotator:[3,0,0,"-"],PValueFormat:[3,0,0,"-"],format_annotations:[3,0,0,"-"],stats:[4,0,0,"-"],utils:[3,0,0,"-"]}},objnames:{"0":["py","module","Python module"],"1":["py","class","Python class"],"2":["py","method","Python method"],"3":["py","property","Python property"],"4":["py","function","Python function"]},objtypes:{"0":"py:module","1":"py:class","2":"py:method","3":"py:property","4":"py:function"},terms:{"0":[3,4],"05":[3,4],"1":[3,4],"case":4,"class":[3,4],"default":[3,4],"float":4,"function":[3,4],"int":4,"return":3,"short":3,"static":[3,4],"true":[3,4],A:3,Or:3,The:3,a_set:[],absolute_box_struct_pair_in_tuple_x_diff:3,accept:3,activate_configured_warn:[],add:3,addit:4,alpha:[3,4],an:3,annot:1,appear:3,appli:3,apply_and_annot:3,apply_test:[3,4],ar:3,arg:4,argument:[3,4],arrai:3,auto:3,ax:3,b:3,bar:3,barplot:3,base:[3,4],befor:3,behavior:4,benjamini:4,between:3,bonferroni:4,bool:4,box:3,box_data1:4,box_data2:4,box_pair:3,box_struct_pair:3,boxplot:3,calcul:3,callabl:4,can:3,check:3,check_alpha:4,check_box_pairs_in_data:3,check_correct_number_custom_annot:[],check_implemented_plott:3,check_is_in:3,check_not_non:3,check_num_comparison:4,check_order_box_pairs_in_data:[],check_order_in_data:3,check_pval_correction_input_valu:[],check_pvalu:4,check_valid_correction_nam:4,check_valid_text_format:3,collect:3,color:3,comparison:[3,4],comparisons_correct:[3,4],comparisonscorrect:[1,3],comput:3,config:3,configur:3,content:1,coordin:3,core:3,corr_kwarg:4,correct:4,corrected_signific:4,correction_method:4,custom:3,data:[3,4],datafram:3,deactivate_configured_warn:[],dict:4,displai:3,document:4,each:3,earlier:4,either:4,equival:4,error:3,error_typ:3,exact:3,exist:3,expect:3,fals:3,follow:3,for_:3,format:[3,4],format_annot:1,format_result:3,formatted_output:4,fraction:3,frame:3,from:4,from_librari:4,func:4,gener:3,get:4,get_annotations_text:3,get_basic_plot:3,get_configur:3,get_correction_paramet:4,get_num_comparison:4,get_validated_comparisons_correct:4,get_x_valu:3,given:3,got:3,gt:4,has_type0_comparisons_correct:3,hochberg:4,holm:4,hue:3,hue_ord:3,index:0,instanc:[3,4],instead:3,interfac:4,interpret:4,keyword:4,kruskal:4,kwarg:4,label:3,leven:4,line_height:3,line_offset:3,line_offset_to_box:3,line_width:3,list:3,loc:3,ls:4,mann:4,messag:3,method:4,method_typ:4,mode:3,modul:[0,1],multipl:4,must:3,name:[3,4],new_plot:3,none:[3,4],num_comparison:[3,4],number:[3,4],object:[3,4],one:[3,4],option:[3,4],order:3,otherwis:3,overrid:3,p:3,p_valu:4,p_values_arrai:[],packag:[0,1],page:0,pair:3,panda:3,param:[],paramet:[3,4],pass:[3,4],per:3,perform:3,perform_stat_test:3,plot:3,plot_param:3,point:3,print_legend_if_us:3,print_pvalue_legend:3,properti:[3,4],provid:[3,4],pval:4,pval_annotation_text:3,pvalu:[3,4],pvalue_format:3,pvalue_format_str:3,pvalue_threshold:3,pvalueformat:1,rais:3,raise_expected_got:3,rang:3,refer:3,remove_nul:3,render:3,render_collect:3,render_set:[],reset_configur:3,result:[3,4],results_arrai:4,return_result:4,run:4,same:3,sampl:4,seaborn:3,search:0,seri:3,set:3,set_custom_annot:3,set_pvalu:3,set_pvalues_and_annot:3,short_nam:4,should_warn_about_configur:[],show:3,show_test_nam:3,significance_suffix:4,simpl:3,simple_format_str:3,simple_text:3,sort_pvalue_threshold:3,standard:3,stat:[1,3],stat_nam:4,stat_str:4,stat_test:[],statist:[3,4],statresult:[1,3],stats_param:[3,4],statsmodel:4,statsmodels_api:4,stattest:[1,3],str:[3,4],string:3,submodul:1,subpackag:1,t:4,test:[1,3],test_descript:4,test_ind:4,test_long_nam:4,test_nam:4,test_pair:4,test_short_nam:[3,4],test_welch:4,text:3,text_annot_custom:3,text_format:3,text_offset:3,thereof:3,thi:[3,4],threshold:3,top:3,tupl:3,two:[3,4],type:3,union:[3,4],unit:3,update_y_for_loc:[],us:[3,4],use_fixed_offset:3,util:1,valid_valu:3,validate_test_short_nam:3,valu:[3,4],valueerror:3,verbos:[3,4],version:4,wa:3,what:3,whitnei:4,width:3,wilcoxon:4,work:3,x:3,y:3,yekuti:4},titles:["Welcome to statannotations\u2019s documentation!","statannotations","setup module","statannotations package","statannotations.stats package"],titleterms:{annot:3,comparisonscorrect:4,content:[0,3,4],document:0,format_annot:3,indic:0,modul:[2,3,4],packag:[3,4],pvalueformat:3,s:0,setup:2,stat:4,statannot:[0,1,3,4],statresult:4,stattest:4,submodul:[3,4],subpackag:3,tabl:0,test:4,util:[3,4],welcom:0}}) \ No newline at end of file +Search.setIndex({docnames:["index","modules","setup","statannotations","statannotations.stats"],envversion:{"sphinx.domains.c":2,"sphinx.domains.changeset":1,"sphinx.domains.citation":1,"sphinx.domains.cpp":3,"sphinx.domains.index":1,"sphinx.domains.javascript":2,"sphinx.domains.math":2,"sphinx.domains.python":3,"sphinx.domains.rst":2,"sphinx.domains.std":2,sphinx:56},filenames:["index.rst","modules.rst","setup.rst","statannotations.rst","statannotations.stats.rst"],objects:{"":{statannotations:[3,0,0,"-"]},"statannotations.Annotation":{Annotation:[3,1,1,""]},"statannotations.Annotation.Annotation":{formatted_output:[3,2,1,""],print_labels_and_content:[3,3,1,""],text:[3,2,1,""]},"statannotations.Annotator":{Annotator:[3,1,1,""]},"statannotations.Annotator.Annotator":{alpha:[3,2,1,""],annotate:[3,3,1,""],annotate_custom_annotations:[3,3,1,""],apply_and_annotate:[3,3,1,""],apply_test:[3,3,1,""],comparisons_correction:[3,2,1,""],configure:[3,3,1,""],fig:[3,2,1,""],get_annotations_text:[3,3,1,""],get_configuration:[3,3,1,""],get_offset_func:[3,3,1,""],has_type0_comparisons_correction:[3,3,1,""],loc:[3,2,1,""],new_plot:[3,3,1,""],print_pvalue_legend:[3,3,1,""],pvalue_format:[3,2,1,""],reset_configuration:[3,3,1,""],set_custom_annotations:[3,3,1,""],set_pvalues:[3,3,1,""],set_pvalues_and_annotate:[3,3,1,""],test:[3,2,1,""],validate_test_short_name:[3,3,1,""],verbose:[3,2,1,""]},"statannotations.PValueFormat":{Formatter:[3,1,1,""],PValueFormat:[3,1,1,""],sort_pvalue_thresholds:[3,4,1,""]},"statannotations.PValueFormat.Formatter":{config:[3,3,1,""],format_data:[3,3,1,""]},"statannotations.PValueFormat.PValueFormat":{config:[3,3,1,""],format_data:[3,3,1,""],get_configuration:[3,3,1,""],print_legend_if_used:[3,3,1,""],pvalue_format_string:[3,2,1,""],pvalue_thresholds:[3,2,1,""],simple_format_string:[3,2,1,""],text_format:[3,2,1,""]},"statannotations.format_annotations":{pval_annotation_text:[3,4,1,""],simple_text:[3,4,1,""]},"statannotations.stats":{ComparisonsCorrection:[4,0,0,"-"],StatResult:[4,0,0,"-"],StatTest:[4,0,0,"-"],test:[4,0,0,"-"],utils:[4,0,0,"-"]},"statannotations.stats.ComparisonsCorrection":{ComparisonsCorrection:[4,1,1,""],check_valid_correction_name:[4,4,1,""],get_correction_parameters:[4,4,1,""],get_validated_comparisons_correction:[4,4,1,""]},"statannotations.stats.ComparisonsCorrection.ComparisonsCorrection":{apply:[4,3,1,""],document:[4,3,1,""]},"statannotations.stats.StatResult":{StatResult:[4,1,1,""]},"statannotations.stats.StatResult.StatResult":{corrected_significance:[4,2,1,""],correction_method:[4,2,1,""],formatted_output:[4,2,1,""],significance_suffix:[4,2,1,""]},"statannotations.stats.StatTest":{StatTest:[4,1,1,""],wilcoxon:[4,4,1,""]},"statannotations.stats.StatTest.StatTest":{from_library:[4,3,1,""],short_name:[4,2,1,""]},"statannotations.stats.test":{apply_test:[4,4,1,""]},"statannotations.stats.utils":{check_alpha:[4,4,1,""],check_num_comparisons:[4,4,1,""],check_pvalues:[4,4,1,""],get_num_comparisons:[4,4,1,""],return_results:[4,4,1,""]},"statannotations.utils":{check_is_in:[3,4,1,""],check_not_none:[3,4,1,""],check_order_in_data:[3,4,1,""],check_pairs_in_data:[3,4,1,""],check_valid_text_format:[3,4,1,""],get_closest:[3,4,1,""],get_x_values:[3,4,1,""],raise_expected_got:[3,4,1,""],remove_null:[3,4,1,""],render_collection:[3,4,1,""]},statannotations:{Annotation:[3,0,0,"-"],Annotator:[3,0,0,"-"],PValueFormat:[3,0,0,"-"],format_annotations:[3,0,0,"-"],stats:[4,0,0,"-"],utils:[3,0,0,"-"]}},objnames:{"0":["py","module","Python module"],"1":["py","class","Python class"],"2":["py","property","Python property"],"3":["py","method","Python method"],"4":["py","function","Python function"]},objtypes:{"0":"py:module","1":"py:class","2":"py:property","3":"py:method","4":"py:function"},terms:{"0":[3,4],"05":[3,4],"1":[3,4],"12141511":3,"9981846":3,"case":4,"class":[3,4],"default":[3,4],"float":4,"function":[3,4],"int":4,"return":3,"short":3,"static":4,"true":[3,4],A:3,If:3,Or:3,The:3,a_list:3,accept:3,add:3,addit:4,alpha:[3,4],an:3,annot:1,annotate_custom_annot:3,appli:[3,4],apply_and_annot:3,apply_test:[3,4],ar:3,arg:[3,4],argument:[3,4],arrai:3,assum:3,auto:3,ax:3,bar:3,base:[3,4],behavior:4,benjamini:4,between:3,bonferroni:4,bool:4,box:3,boxplot:3,calcul:3,callabl:4,check:3,check_alpha:4,check_is_in:3,check_not_non:3,check_num_comparison:4,check_order_in_data:3,check_pairs_in_data:3,check_pvalu:4,check_valid_correction_nam:4,check_valid_text_format:3,classmethod:3,close:3,closest:3,collect:3,color:3,com:3,comparison:[3,4],comparisons_correct:[3,4],comparisonscorrect:[1,3],comput:3,config:3,configur:3,constructor:3,content:1,coordin:3,core:3,corr_kwarg:4,correct:4,corrected_signific:4,correction_method:4,custom:3,data:[3,4],datafram:3,dict:4,displai:3,document:4,each:3,earlier:4,either:4,engin:3,equal:3,equival:4,error:3,error_typ:3,exact:3,exist:3,expect:3,fals:3,fig:3,follow:3,for_:3,format:[3,4],format_annot:1,format_data:3,formatt:3,formatted_output:[3,4],fraction:3,frame:3,from:[3,4],from_librari:4,func:4,gener:3,get:4,get_annotations_text:3,get_closest:3,get_configur:3,get_correction_paramet:4,get_num_comparison:4,get_offset_func:3,get_validated_comparisons_correct:4,get_x_valu:3,got:3,group:3,group_data1:4,group_data2:4,gt:4,has_type0_comparisons_correct:3,hochberg:4,hold:3,holm:4,http:3,hue:3,hue_ord:3,index:0,instanc:[3,4],instead:3,interfac:4,interpret:4,keyword:4,kruskal:4,kwarg:[3,4],label:3,leven:4,line_height:3,line_offset:3,line_offset_to_group:3,line_width:3,link:3,list:3,loc:3,ls:4,mann:4,messag:3,method:4,method_typ:4,mode:3,modul:[0,1],multipl:4,must:3,mylist:3,mynumb:3,name:[3,4],new_plot:3,none:[3,4],num_comparison:[3,4],number:[3,4],object:[3,4],one:[3,4],option:[3,4],order:3,otherwis:3,overrid:3,p:3,p_valu:4,packag:[0,1],page:0,pair:3,panda:3,paramet:[3,4],pass:[3,4],per:3,perform:3,plot:3,plot_param:3,point:3,posit:3,print_labels_and_cont:3,print_legend_if_us:3,print_pvalue_legend:3,properti:[3,4],provid:[3,4],pval:4,pval_annotation_text:3,pvalu:[3,4],pvalue_format:3,pvalue_format_str:3,pvalue_threshold:3,pvalueformat:1,rais:3,raise_expected_got:3,rang:3,refer:3,remove_nul:3,render:3,render_collect:3,reset_configur:3,result:[3,4],results_arrai:4,return_result:4,run:4,same:3,sampl:4,seaborn:3,search:0,sep:3,seri:3,set:3,set_custom_annot:3,set_pvalu:3,set_pvalues_and_annot:3,short_nam:4,show:3,show_test_nam:3,significance_suffix:4,simpl:3,simple_format_str:3,simple_text:3,smallest:3,sort:3,sort_pvalue_threshold:3,stackoverflow:3,standard:3,stat:[1,3],stat_nam:4,stat_str:4,statist:[3,4],statresult:[1,3],stats_param:[3,4],statsmodel:4,statsmodels_api:4,stattest:[1,3],str:[3,4],string:3,struct:3,submodul:1,subpackag:1,t:4,test:[1,3],test_descript:4,test_ind:4,test_long_nam:4,test_nam:4,test_pair:4,test_result_list:4,test_short_nam:[3,4],test_welch:4,text:3,text_annot_custom:3,text_format:3,text_offset:3,them:3,thereof:3,thi:[3,4],three:3,threshold:3,top:3,tupl:3,two:[3,4],type:3,union:[3,4],us:[3,4],use_fixed_offset:3,util:1,valid_valu:3,validate_test_short_nam:3,valu:[3,4],valueerror:3,verbos:[3,4],version:4,vs:3,wa:3,whitnei:4,wilcoxon:4,work:3,x:3,y:3,yekuti:4},titles:["Welcome to statannotations\u2019s documentation!","statannotations","setup module","statannotations package","statannotations.stats package"],titleterms:{annot:3,comparisonscorrect:4,content:[0,3,4],document:0,format_annot:3,indic:0,modul:[2,3,4],packag:[3,4],pvalueformat:3,s:0,setup:2,stat:4,statannot:[0,1,3,4],statresult:4,stattest:4,submodul:[3,4],subpackag:3,tabl:0,test:4,util:[3,4],welcom:0}}) \ No newline at end of file diff --git a/docs/build/html/statannotations.html b/docs/build/html/statannotations.html index 13780ed..aaacef6 100644 --- a/docs/build/html/statannotations.html +++ b/docs/build/html/statannotations.html @@ -96,6 +96,7 @@ </ul> </li> <li class="toctree-l3"><a class="reference internal" href="#submodules">Submodules</a></li> +<li class="toctree-l3"><a class="reference internal" href="#module-statannotations.Annotation">statannotations.Annotation module</a></li> <li class="toctree-l3"><a class="reference internal" href="#module-statannotations.Annotator">statannotations.Annotator module</a></li> <li class="toctree-l3"><a class="reference internal" href="#module-statannotations.PValueFormat">statannotations.PValueFormat module</a></li> <li class="toctree-l3"><a class="reference internal" href="#module-statannotations.format_annotations">statannotations.format_annotations module</a></li> @@ -195,29 +196,54 @@ <h2>Subpackages<a class="headerlink" href="#subpackages" title="Permalink to thi </div> <div class="section" id="submodules"> <h2>Submodules<a class="headerlink" href="#submodules" title="Permalink to this headline">¶</a></h2> +</div> +<div class="section" id="module-statannotations.Annotation"> +<span id="statannotations-annotation-module"></span><h2>statannotations.Annotation module<a class="headerlink" href="#module-statannotations.Annotation" title="Permalink to this headline">¶</a></h2> +<dl class="py class"> +<dt class="sig sig-object py" id="statannotations.Annotation.Annotation"> +<em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">statannotations.Annotation.</span></span><span class="sig-name descname"><span class="pre">Annotation</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">structs</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">data</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span> </span><a class="reference internal" href="statannotations.stats.html#statannotations.stats.StatResult.StatResult" title="statannotations.stats.StatResult.StatResult"><span class="pre">statannotations.stats.StatResult.StatResult</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">formatter</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#statannotations.PValueFormat.Formatter" title="statannotations.PValueFormat.Formatter"><span class="pre">statannotations.PValueFormat.Formatter</span></a><span class="p"><span class="pre">]</span></span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#statannotations.Annotation.Annotation" title="Permalink to this definition">¶</a></dt> +<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">object</span></code></p> +<p>Holds data, linked structs and an optional Formatter.</p> +<dl class="py property"> +<dt class="sig sig-object py" id="statannotations.Annotation.Annotation.formatted_output"> +<em class="property"><span class="pre">property</span> </em><span class="sig-name descname"><span class="pre">formatted_output</span></span><a class="headerlink" href="#statannotations.Annotation.Annotation.formatted_output" title="Permalink to this definition">¶</a></dt> +<dd></dd></dl> + +<dl class="py method"> +<dt class="sig sig-object py" id="statannotations.Annotation.Annotation.print_labels_and_content"> +<span class="sig-name descname"><span class="pre">print_labels_and_content</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">sep</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'</span> <span class="pre">vs.</span> <span class="pre">'</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#statannotations.Annotation.Annotation.print_labels_and_content" title="Permalink to this definition">¶</a></dt> +<dd></dd></dl> + +<dl class="py property"> +<dt class="sig sig-object py" id="statannotations.Annotation.Annotation.text"> +<em class="property"><span class="pre">property</span> </em><span class="sig-name descname"><span class="pre">text</span></span><a class="headerlink" href="#statannotations.Annotation.Annotation.text" title="Permalink to this definition">¶</a></dt> +<dd></dd></dl> + +</dd></dl> + </div> <div class="section" id="module-statannotations.Annotator"> <span id="statannotations-annotator-module"></span><h2>statannotations.Annotator module<a class="headerlink" href="#module-statannotations.Annotator" title="Permalink to this headline">¶</a></h2> <dl class="py class"> <dt class="sig sig-object py" id="statannotations.Annotator.Annotator"> -<em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">statannotations.Annotator.</span></span><span class="sig-name descname"><span class="pre">Annotator</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">ax</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">box_pairs</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">plot</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'boxplot'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">data</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">y</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">hue</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">order</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">hue_order</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="o"><span class="pre"></span></span><span class="n"><span class="pre">plot_params</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#statannotations.Annotator.Annotator" title="Permalink to this definition">¶</a></dt> +<em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">statannotations.Annotator.</span></span><span class="sig-name descname"><span class="pre">Annotator</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">ax</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pairs</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">plot</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'boxplot'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">data</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">y</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">hue</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">order</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">hue_order</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">engine</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'seaborn'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="o"><span class="pre"></span></span><span class="n"><span class="pre">plot_params</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#statannotations.Annotator.Annotator" title="Permalink to this definition">¶</a></dt> <dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">object</span></code></p> <p>Optionally computes statistical test between pairs of data series, and add -statistical annotation on top of the boxes/bars. The same exact arguments -<cite>data</cite>, <cite>x</cite>, <cite>y</cite>, <cite>hue</cite>, <cite>order</cite>, <cite>width</cite>, <cite>hue_order</cite> (and <cite>units</cite>) as in -the seaborn boxplot/barplot function must be passed to this function.</p> -<p>This function works in one of the two following modes: -a) <cite>perform_stat_test</cite> is True (default): - statistical test as given by argument <cite>test</cite> is performed. -* The <cite>test_short_name</cite> argument can be used to customize what appears -before the pvalues if test is a string. -b) <cite>perform_stat_test</cite> is False: no statistical test is performed, list of -custom p-values <cite>pvalues</cite> are used for each pair of boxes.</p> -<dl class="py method"> -<dt class="sig sig-object py" id="statannotations.Annotator.Annotator.absolute_box_struct_pair_in_tuple_x_diff"> -<em class="property"><span class="pre">static</span> </em><span class="sig-name descname"><span class="pre">absolute_box_struct_pair_in_tuple_x_diff</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">box_struct_pair</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#statannotations.Annotator.Annotator.absolute_box_struct_pair_in_tuple_x_diff" title="Permalink to this definition">¶</a></dt> -<dd></dd></dl> - +statistical annotation on top of the groups (boxes, bars…). The same +exact arguments provided to the seaborn plotting function must be passed to +the constructor.</p> +<dl class="simple"> +<dt>This Annotator works in one of the three following modes:</dt><dd><ul class="simple"> +<li><p>Add custom text annotations (<cite>set_custom_annotations</cite>)</p></li> +<li><p>Format pvalues and add them to the plot (<cite>set_pvalues</cite>)</p></li> +<li><dl class="simple"> +<dt>Perform a statistical test and then add the results to the plot</dt><dd><p>(<cite>apply_test</cite>)</p> +</dd> +</dl> +</li> +</ul> +</dd> +</dl> <dl class="py property"> <dt class="sig sig-object py" id="statannotations.Annotator.Annotator.alpha"> <em class="property"><span class="pre">property</span> </em><span class="sig-name descname"><span class="pre">alpha</span></span><a class="headerlink" href="#statannotations.Annotator.Annotator.alpha" title="Permalink to this definition">¶</a></dt> @@ -225,10 +251,21 @@ <h2>Submodules<a class="headerlink" href="#submodules" title="Permalink to this <dl class="py method"> <dt class="sig sig-object py" id="statannotations.Annotator.Annotator.annotate"> -<span class="sig-name descname"><span class="pre">annotate</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">line_offset</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">line_offset_to_box</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#statannotations.Annotator.Annotator.annotate" title="Permalink to this definition">¶</a></dt> +<span class="sig-name descname"><span class="pre">annotate</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">line_offset</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">line_offset_to_group</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#statannotations.Annotator.Annotator.annotate" title="Permalink to this definition">¶</a></dt> <dd><p>Add configured annotations to the plot.</p> </dd></dl> +<dl class="py method"> +<dt class="sig sig-object py" id="statannotations.Annotator.Annotator.annotate_custom_annotations"> +<span class="sig-name descname"><span class="pre">annotate_custom_annotations</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">text_annot_custom</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#statannotations.Annotator.Annotator.annotate_custom_annotations" title="Permalink to this definition">¶</a></dt> +<dd><dl class="field-list simple"> +<dt class="field-odd">Parameters</dt> +<dd class="field-odd"><p><strong>text_annot_custom</strong> – List of strings to annotate for each +<cite>pair</cite></p> +</dd> +</dl> +</dd></dl> + <dl class="py method"> <dt class="sig sig-object py" id="statannotations.Annotator.Annotator.apply_and_annotate"> <span class="sig-name descname"><span class="pre">apply_and_annotate</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="headerlink" href="#statannotations.Annotator.Annotator.apply_and_annotate" title="Permalink to this definition">¶</a></dt> @@ -243,7 +280,7 @@ <h2>Submodules<a class="headerlink" href="#submodules" title="Permalink to this <dd class="field-odd"><ul class="simple"> <li><p><strong>stats_params</strong> – Parameters for statistical test functions.</p></li> <li><p><strong>num_comparisons</strong> – Override number of comparisons otherwise -calculated with number of box_pairs</p></li> +calculated with number of pairs</p></li> </ul> </dd> </dl> @@ -263,7 +300,7 @@ <h2>Submodules<a class="headerlink" href="#submodules" title="Permalink to this <li><p><cite>comparisons_correction</cite></p></li> <li><p><cite>line_height</cite>: in axes fraction coordinates</p></li> <li><p><cite>line_offset</cite></p></li> -<li><p><cite>line_offset_to_box</cite></p></li> +<li><p><cite>line_offset_to_group</cite></p></li> <li><p><cite>line_width</cite></p></li> <li><p><cite>loc</cite></p></li> <li><p><cite>pvalue_format</cite></p></li> @@ -280,19 +317,24 @@ <h2>Submodules<a class="headerlink" href="#submodules" title="Permalink to this </ul> </dd></dl> +<dl class="py property"> +<dt class="sig sig-object py" id="statannotations.Annotator.Annotator.fig"> +<em class="property"><span class="pre">property</span> </em><span class="sig-name descname"><span class="pre">fig</span></span><a class="headerlink" href="#statannotations.Annotator.Annotator.fig" title="Permalink to this definition">¶</a></dt> +<dd></dd></dl> + <dl class="py method"> <dt class="sig sig-object py" id="statannotations.Annotator.Annotator.get_annotations_text"> <span class="sig-name descname"><span class="pre">get_annotations_text</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="headerlink" href="#statannotations.Annotator.Annotator.get_annotations_text" title="Permalink to this definition">¶</a></dt> <dd></dd></dl> <dl class="py method"> -<dt class="sig sig-object py" id="statannotations.Annotator.Annotator.get_basic_plot"> -<em class="property"><span class="pre">static</span> </em><span class="sig-name descname"><span class="pre">get_basic_plot</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">ax</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">box_pairs</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">plot</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'boxplot'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">data</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">y</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">hue</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">order</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">hue_order</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="o"><span class="pre"></span></span><span class="n"><span class="pre">plot_params</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#statannotations.Annotator.Annotator.get_basic_plot" title="Permalink to this definition">¶</a></dt> +<dt class="sig sig-object py" id="statannotations.Annotator.Annotator.get_configuration"> +<span class="sig-name descname"><span class="pre">get_configuration</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="headerlink" href="#statannotations.Annotator.Annotator.get_configuration" title="Permalink to this definition">¶</a></dt> <dd></dd></dl> <dl class="py method"> -<dt class="sig sig-object py" id="statannotations.Annotator.Annotator.get_configuration"> -<span class="sig-name descname"><span class="pre">get_configuration</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="headerlink" href="#statannotations.Annotator.Annotator.get_configuration" title="Permalink to this definition">¶</a></dt> +<dt class="sig sig-object py" id="statannotations.Annotator.Annotator.get_offset_func"> +<em class="property"><span class="pre">classmethod</span> </em><span class="sig-name descname"><span class="pre">get_offset_func</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">position</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#statannotations.Annotator.Annotator.get_offset_func" title="Permalink to this definition">¶</a></dt> <dd></dd></dl> <dl class="py method"> @@ -307,7 +349,7 @@ <h2>Submodules<a class="headerlink" href="#submodules" title="Permalink to this <dl class="py method"> <dt class="sig sig-object py" id="statannotations.Annotator.Annotator.new_plot"> -<span class="sig-name descname"><span class="pre">new_plot</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">ax</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">box_pairs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">plot</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'boxplot'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">data</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">y</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">hue</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">order</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">hue_order</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="o"><span class="pre"></span></span><span class="n"><span class="pre">plot_params</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#statannotations.Annotator.Annotator.new_plot" title="Permalink to this definition">¶</a></dt> +<span class="sig-name descname"><span class="pre">new_plot</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">ax</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pairs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">plot</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'boxplot'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">data</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">y</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">hue</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">order</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">hue_order</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">engine</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">str</span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">'seaborn'</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">plot_params</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#statannotations.Annotator.Annotator.new_plot" title="Permalink to this definition">¶</a></dt> <dd></dd></dl> <dl class="py method"> @@ -331,7 +373,7 @@ <h2>Submodules<a class="headerlink" href="#submodules" title="Permalink to this <dd><dl class="field-list simple"> <dt class="field-odd">Parameters</dt> <dd class="field-odd"><p><strong>text_annot_custom</strong> – List of strings to annotate for each -<cite>box_pair</cite></p> +<cite>pair</cite></p> </dd> </dl> </dd></dl> @@ -342,9 +384,9 @@ <h2>Submodules<a class="headerlink" href="#submodules" title="Permalink to this <dd><dl class="field-list simple"> <dt class="field-odd">Parameters</dt> <dd class="field-odd"><ul class="simple"> -<li><p><strong>pvalues</strong> – list or array of p-values for each box pair comparison.</p></li> +<li><p><strong>pvalues</strong> – list or array of p-values for each pair comparison.</p></li> <li><p><strong>num_comparisons</strong> – Override number of comparisons otherwise -calculated with number of box_pairs</p></li> +calculated with number of pairs</p></li> </ul> </dd> </dl> @@ -365,28 +407,44 @@ <h2>Submodules<a class="headerlink" href="#submodules" title="Permalink to this <span class="sig-name descname"><span class="pre">validate_test_short_name</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="headerlink" href="#statannotations.Annotator.Annotator.validate_test_short_name" title="Permalink to this definition">¶</a></dt> <dd></dd></dl> -</dd></dl> - -<dl class="py function"> -<dt class="sig sig-object py" id="statannotations.Annotator.check_implemented_plotters"> -<span class="sig-prename descclassname"><span class="pre">statannotations.Annotator.</span></span><span class="sig-name descname"><span class="pre">check_implemented_plotters</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">plot</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#statannotations.Annotator.check_implemented_plotters" title="Permalink to this definition">¶</a></dt> +<dl class="py property"> +<dt class="sig sig-object py" id="statannotations.Annotator.Annotator.verbose"> +<em class="property"><span class="pre">property</span> </em><span class="sig-name descname"><span class="pre">verbose</span></span><a class="headerlink" href="#statannotations.Annotator.Annotator.verbose" title="Permalink to this definition">¶</a></dt> <dd></dd></dl> +</dd></dl> + </div> <div class="section" id="module-statannotations.PValueFormat"> <span id="statannotations-pvalueformat-module"></span><h2>statannotations.PValueFormat module<a class="headerlink" href="#module-statannotations.PValueFormat" title="Permalink to this headline">¶</a></h2> +<dl class="py class"> +<dt class="sig sig-object py" id="statannotations.PValueFormat.Formatter"> +<em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">statannotations.PValueFormat.</span></span><span class="sig-name descname"><span class="pre">Formatter</span></span><a class="headerlink" href="#statannotations.PValueFormat.Formatter" title="Permalink to this definition">¶</a></dt> +<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">object</span></code></p> +<dl class="py method"> +<dt class="sig sig-object py" id="statannotations.PValueFormat.Formatter.config"> +<span class="sig-name descname"><span class="pre">config</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="o"><span class="pre"></span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre"></span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#statannotations.PValueFormat.Formatter.config" title="Permalink to this definition">¶</a></dt> +<dd></dd></dl> + +<dl class="py method"> +<dt class="sig sig-object py" id="statannotations.PValueFormat.Formatter.format_data"> +<span class="sig-name descname"><span class="pre">format_data</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#statannotations.PValueFormat.Formatter.format_data" title="Permalink to this definition">¶</a></dt> +<dd></dd></dl> + +</dd></dl> + <dl class="py class"> <dt class="sig sig-object py" id="statannotations.PValueFormat.PValueFormat"> <em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">statannotations.PValueFormat.</span></span><span class="sig-name descname"><span class="pre">PValueFormat</span></span><a class="headerlink" href="#statannotations.PValueFormat.PValueFormat" title="Permalink to this definition">¶</a></dt> -<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">object</span></code></p> +<dd><p>Bases: <a class="reference internal" href="#statannotations.PValueFormat.Formatter" title="statannotations.PValueFormat.Formatter"><code class="xref py py-class docutils literal notranslate"><span class="pre">statannotations.PValueFormat.Formatter</span></code></a></p> <dl class="py method"> <dt class="sig sig-object py" id="statannotations.PValueFormat.PValueFormat.config"> <span class="sig-name descname"><span class="pre">config</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="o"><span class="pre"></span></span><span class="n"><span class="pre">parameters</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#statannotations.PValueFormat.PValueFormat.config" title="Permalink to this definition">¶</a></dt> <dd></dd></dl> <dl class="py method"> -<dt class="sig sig-object py" id="statannotations.PValueFormat.PValueFormat.format_result"> -<span class="sig-name descname"><span class="pre">format_result</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">result</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#statannotations.PValueFormat.PValueFormat.format_result" title="Permalink to this definition">¶</a></dt> +<dt class="sig sig-object py" id="statannotations.PValueFormat.PValueFormat.format_data"> +<span class="sig-name descname"><span class="pre">format_data</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">result</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#statannotations.PValueFormat.PValueFormat.format_data" title="Permalink to this definition">¶</a></dt> <dd></dd></dl> <dl class="py method"> @@ -431,16 +489,16 @@ <h2>Submodules<a class="headerlink" href="#submodules" title="Permalink to this <span id="statannotations-format-annotations-module"></span><h2>statannotations.format_annotations module<a class="headerlink" href="#module-statannotations.format_annotations" title="Permalink to this headline">¶</a></h2> <dl class="py function"> <dt class="sig sig-object py" id="statannotations.format_annotations.pval_annotation_text"> -<span class="sig-prename descclassname"><span class="pre">statannotations.format_annotations.</span></span><span class="sig-name descname"><span class="pre">pval_annotation_text</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">result</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="statannotations.stats.html#statannotations.stats.StatResult.StatResult" title="statannotations.stats.StatResult.StatResult"><span class="pre">statannotations.stats.StatResult.StatResult</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span> </span><a class="reference internal" href="statannotations.stats.html#statannotations.stats.StatResult.StatResult" title="statannotations.stats.StatResult.StatResult"><span class="pre">statannotations.stats.StatResult.StatResult</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pvalue_thresholds</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">List</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#statannotations.format_annotations.pval_annotation_text" title="Permalink to this definition">¶</a></dt> +<span class="sig-prename descclassname"><span class="pre">statannotations.format_annotations.</span></span><span class="sig-name descname"><span class="pre">pval_annotation_text</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">result</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="statannotations.stats.html#statannotations.stats.StatResult.StatResult" title="statannotations.stats.StatResult.StatResult"><span class="pre">statannotations.stats.StatResult.StatResult</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span> </span><a class="reference internal" href="statannotations.stats.html#statannotations.stats.StatResult.StatResult" title="statannotations.stats.StatResult.StatResult"><span class="pre">statannotations.stats.StatResult.StatResult</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pvalue_thresholds</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">List</span></span></em><span class="sig-paren">)</span> → <span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span> </span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><a class="headerlink" href="#statannotations.format_annotations.pval_annotation_text" title="Permalink to this definition">¶</a></dt> <dd><dl class="field-list simple"> <dt class="field-odd">Parameters</dt> <dd class="field-odd"><ul class="simple"> <li><p><strong>result</strong> – StatResult instance or list thereof</p></li> -<li><p><strong>pvalue_thresholds</strong> – thresholds to use for formatting</p></li> +<li><p><strong>pvalue_thresholds</strong> – thresholds to use for formatter</p></li> </ul> </dd> <dt class="field-even">Returns</dt> -<dd class="field-even"><p>A Series of rendered annotations if a list of StatResults was +<dd class="field-even"><p>A List of rendered annotations if a list of StatResults was provided, a string otherwise.</p> </dd> </dl> @@ -468,12 +526,6 @@ <h2>Submodules<a class="headerlink" href="#submodules" title="Permalink to this </div> <div class="section" id="module-statannotations.utils"> <span id="statannotations-utils-module"></span><h2>statannotations.utils module<a class="headerlink" href="#module-statannotations.utils" title="Permalink to this headline">¶</a></h2> -<dl class="py function"> -<dt class="sig sig-object py" id="statannotations.utils.check_box_pairs_in_data"> -<span class="sig-prename descclassname"><span class="pre">statannotations.utils.</span></span><span class="sig-name descname"><span class="pre">check_box_pairs_in_data</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">box_pairs</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">list</span><span class="p"><span class="pre">,</span> </span><span class="pre">tuple</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">data</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">list</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span> </span><span class="pre">pandas.core.frame.DataFrame</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span> </span><span class="pre">list</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">hue</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">hue_order</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#statannotations.utils.check_box_pairs_in_data" title="Permalink to this definition">¶</a></dt> -<dd><p>Checks that values referred to in <cite>order</cite> and <cite>box_pairs</cite> exist in data.</p> -</dd></dl> - <dl class="py function"> <dt class="sig sig-object py" id="statannotations.utils.check_is_in"> <span class="sig-prename descclassname"><span class="pre">statannotations.utils.</span></span><span class="sig-name descname"><span class="pre">check_is_in</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="pre">x</span></em>, <em class="sig-param"><span class="pre">valid_values</span></em>, <em class="sig-param"><span class="pre">error_type=<class</span> <span class="pre">'ValueError'></span></em>, <em class="sig-param"><span class="pre">label=None</span></em><span class="sig-paren">)</span><a class="headerlink" href="#statannotations.utils.check_is_in" title="Permalink to this definition">¶</a></dt> @@ -490,11 +542,25 @@ <h2>Submodules<a class="headerlink" href="#submodules" title="Permalink to this <span class="sig-prename descclassname"><span class="pre">statannotations.utils.</span></span><span class="sig-name descname"><span class="pre">check_order_in_data</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">x</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">order</span></span></em><span class="sig-paren">)</span> → <span class="pre">None</span><a class="headerlink" href="#statannotations.utils.check_order_in_data" title="Permalink to this definition">¶</a></dt> <dd></dd></dl> +<dl class="py function"> +<dt class="sig sig-object py" id="statannotations.utils.check_pairs_in_data"> +<span class="sig-prename descclassname"><span class="pre">statannotations.utils.</span></span><span class="sig-name descname"><span class="pre">check_pairs_in_data</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">pairs</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">list</span><span class="p"><span class="pre">,</span> </span><span class="pre">tuple</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">data</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">list</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span> </span><span class="pre">pandas.core.frame.DataFrame</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span> </span><span class="pre">list</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">hue</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">hue_order</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#statannotations.utils.check_pairs_in_data" title="Permalink to this definition">¶</a></dt> +<dd><p>Checks that values referred to in <cite>order</cite> and <cite>pairs</cite> exist in data.</p> +</dd></dl> + <dl class="py function"> <dt class="sig sig-object py" id="statannotations.utils.check_valid_text_format"> <span class="sig-prename descclassname"><span class="pre">statannotations.utils.</span></span><span class="sig-name descname"><span class="pre">check_valid_text_format</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">text_format</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#statannotations.utils.check_valid_text_format" title="Permalink to this definition">¶</a></dt> <dd></dd></dl> +<dl class="py function"> +<dt class="sig sig-object py" id="statannotations.utils.get_closest"> +<span class="sig-prename descclassname"><span class="pre">statannotations.utils.</span></span><span class="sig-name descname"><span class="pre">get_closest</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">a_list</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">value</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#statannotations.utils.get_closest" title="Permalink to this definition">¶</a></dt> +<dd><p>Assumes myList is sorted. Returns closest value to myNumber. +If two numbers are equally close, return the smallest number. +from <a class="reference external" href="https://stackoverflow.com/a/12141511/9981846">https://stackoverflow.com/a/12141511/9981846</a></p> +</dd></dl> + <dl class="py function"> <dt class="sig sig-object py" id="statannotations.utils.get_x_values"> <span class="sig-prename descclassname"><span class="pre">statannotations.utils.</span></span><span class="sig-name descname"><span class="pre">get_x_values</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">x</span></span></em><span class="sig-paren">)</span> → <span class="pre">set</span><a class="headerlink" href="#statannotations.utils.get_x_values" title="Permalink to this definition">¶</a></dt> diff --git a/docs/build/html/statannotations.stats.html b/docs/build/html/statannotations.stats.html index 4a93cad..099ec99 100644 --- a/docs/build/html/statannotations.stats.html +++ b/docs/build/html/statannotations.stats.html @@ -95,6 +95,7 @@ </ul> </li> <li class="toctree-l3"><a class="reference internal" href="statannotations.html#submodules">Submodules</a></li> +<li class="toctree-l3"><a class="reference internal" href="statannotations.html#module-statannotations.Annotation">statannotations.Annotation module</a></li> <li class="toctree-l3"><a class="reference internal" href="statannotations.html#module-statannotations.Annotator">statannotations.Annotator module</a></li> <li class="toctree-l3"><a class="reference internal" href="statannotations.html#module-statannotations.PValueFormat">statannotations.PValueFormat module</a></li> <li class="toctree-l3"><a class="reference internal" href="statannotations.html#module-statannotations.format_annotations">statannotations.format_annotations module</a></li> @@ -186,6 +187,11 @@ <h2>Submodules<a class="headerlink" href="#submodules" title="Permalink to this <dt class="sig sig-object py" id="statannotations.stats.ComparisonsCorrection.ComparisonsCorrection"> <em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">statannotations.stats.ComparisonsCorrection.</span></span><span class="sig-name descname"><span class="pre">ComparisonsCorrection</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">method</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span> </span><span class="pre">callable</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">alpha</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">float</span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">0.05</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">name</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">method_type</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">statsmodels_api</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">bool</span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">corr_kwargs</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">dict</span><span class="p"><span class="pre">]</span></span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#statannotations.stats.ComparisonsCorrection.ComparisonsCorrection" title="Permalink to this definition">¶</a></dt> <dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">object</span></code></p> +<dl class="py method"> +<dt class="sig sig-object py" id="statannotations.stats.ComparisonsCorrection.ComparisonsCorrection.apply"> +<span class="sig-name descname"><span class="pre">apply</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">test_result_list</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#statannotations.stats.ComparisonsCorrection.ComparisonsCorrection.apply" title="Permalink to this definition">¶</a></dt> +<dd></dd></dl> + <dl class="py method"> <dt class="sig sig-object py" id="statannotations.stats.ComparisonsCorrection.ComparisonsCorrection.document"> <span class="sig-name descname"><span class="pre">document</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">func</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#statannotations.stats.ComparisonsCorrection.ComparisonsCorrection.document" title="Permalink to this definition">¶</a></dt> @@ -258,7 +264,7 @@ <h2>Submodules<a class="headerlink" href="#submodules" title="Permalink to this <dl class="py function"> <dt class="sig sig-object py" id="statannotations.stats.StatTest.wilcoxon"> -<span class="sig-prename descclassname"><span class="pre">statannotations.stats.StatTest.</span></span><span class="sig-name descname"><span class="pre">wilcoxon</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">box_data1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">box_data2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="o"><span class="pre"></span></span><span class="n"><span class="pre">stats_params</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#statannotations.stats.StatTest.wilcoxon" title="Permalink to this definition">¶</a></dt> +<span class="sig-prename descclassname"><span class="pre">statannotations.stats.StatTest.</span></span><span class="sig-name descname"><span class="pre">wilcoxon</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">group_data1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">group_data2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="o"><span class="pre"></span></span><span class="n"><span class="pre">stats_params</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#statannotations.stats.StatTest.wilcoxon" title="Permalink to this definition">¶</a></dt> <dd><p>This function provides the equivalent behavior from earlier versions of statannot/statannotations.</p> </dd></dl> @@ -268,13 +274,13 @@ <h2>Submodules<a class="headerlink" href="#submodules" title="Permalink to this <span id="statannotations-stats-test-module"></span><h2>statannotations.stats.test module<a class="headerlink" href="#module-statannotations.stats.test" title="Permalink to this headline">¶</a></h2> <dl class="py function"> <dt class="sig sig-object py" id="statannotations.stats.test.apply_test"> -<span class="sig-prename descclassname"><span class="pre">statannotations.stats.test.</span></span><span class="sig-name descname"><span class="pre">apply_test</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">box_data1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">box_data2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">test</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#statannotations.stats.StatTest.StatTest" title="statannotations.stats.StatTest.StatTest"><span class="pre">statannotations.stats.StatTest.StatTest</span></a><span class="p"><span class="pre">,</span> </span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">comparisons_correction</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#statannotations.stats.ComparisonsCorrection.ComparisonsCorrection" title="statannotations.stats.ComparisonsCorrection.ComparisonsCorrection"><span class="pre">statannotations.stats.ComparisonsCorrection.ComparisonsCorrection</span></a><span class="p"><span class="pre">,</span> </span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">num_comparisons</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">int</span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">alpha</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">float</span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">0.05</span></span></em>, <em class="sig-param"><span class="o"><span class="pre"></span></span><span class="n"><span class="pre">stats_params</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#statannotations.stats.test.apply_test" title="Permalink to this definition">¶</a></dt> +<span class="sig-prename descclassname"><span class="pre">statannotations.stats.test.</span></span><span class="sig-name descname"><span class="pre">apply_test</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">group_data1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">group_data2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">test</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#statannotations.stats.StatTest.StatTest" title="statannotations.stats.StatTest.StatTest"><span class="pre">statannotations.stats.StatTest.StatTest</span></a><span class="p"><span class="pre">,</span> </span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">comparisons_correction</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#statannotations.stats.ComparisonsCorrection.ComparisonsCorrection" title="statannotations.stats.ComparisonsCorrection.ComparisonsCorrection"><span class="pre">statannotations.stats.ComparisonsCorrection.ComparisonsCorrection</span></a><span class="p"><span class="pre">,</span> </span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">num_comparisons</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">int</span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">alpha</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">float</span></span> <span class="o"><span class="pre">=</span></span> <span class="default_value"><span class="pre">0.05</span></span></em>, <em class="sig-param"><span class="o"><span class="pre"></span></span><span class="n"><span class="pre">stats_params</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#statannotations.stats.test.apply_test" title="Permalink to this definition">¶</a></dt> <dd><p>Get formatted result of two-sample statistical test.</p> <dl class="field-list simple"> <dt class="field-odd">Parameters</dt> <dd class="field-odd"><ul class="simple"> -<li><p><strong>box_data1</strong> – data</p></li> -<li><p><strong>box_data2</strong> – data</p></li> +<li><p><strong>group_data1</strong> – data</p></li> +<li><p><strong>group_data2</strong> – data</p></li> <li><p><strong>test</strong> – Union[StatTest, str]: Statistical test to run. Either a <cite>StatTest</cite> instance or one of: - <cite>Levene</cite> diff --git a/docs/source/statannotations.rst b/docs/source/statannotations.rst index d3f9b4e..3a24356 100644 --- a/docs/source/statannotations.rst +++ b/docs/source/statannotations.rst @@ -12,6 +12,14 @@ Subpackages Submodules ---------- +statannotations.Annotation module +--------------------------------- + +.. automodule:: statannotations.Annotation + :members: + :undoc-members: + :show-inheritance: + statannotations.Annotator module -------------------------------- diff --git a/statannotations/Annotation.py b/statannotations/Annotation.py new file mode 100644 index 0000000..3434ddd --- /dev/null +++ b/statannotations/Annotation.py @@ -0,0 +1,45 @@ +from typing import Union + +from statannotations.PValueFormat import Formatter +from statannotations.stats.StatResult import StatResult + + +class Annotation: + """ + Holds data, linked structs and an optional Formatter. + """ + def __init__(self, structs, data: Union[str, StatResult], + formatter: Formatter = None): + """ + :param structs: plot structures concerned by the data + :param data: a string or StatResult to be formatted by a formatter + :param formatter: A Formatter object. Statannotations provides a + PValueFormatter for StatResult objects. + """ + self.structs = structs + self.data = data + self.is_custom = isinstance(data, str) + self.formatter = formatter + + @property + def text(self): + if self.is_custom: + return self.data + else: + if self.formatter is None: + raise ValueError("Missing a PValueFormat object to " + "format the statistical result.") + return self.formatter.format_data(self.data) + + @property + def formatted_output(self): + if isinstance(self.data, str): + return self.data + else: + return self.data.formatted_output + + def print_labels_and_content(self, sep=" vs. "): + labels_string = sep.join([struct["label"] + for struct in self.structs]) + + print(f"{labels_string}: {self.formatted_output}") diff --git a/statannotations/Annotator.py b/statannotations/Annotator.py index 8cf3d14..390bd83 100644 --- a/statannotations/Annotator.py +++ b/statannotations/Annotator.py @@ -4,14 +4,13 @@ import matplotlib.pyplot as plt import matplotlib.transforms as mtransforms import numpy as np -import seaborn as sns from matplotlib import lines -from matplotlib.collections import PathCollection from matplotlib.font_manager import FontProperties -from matplotlib.patches import Rectangle +from statannotations.Annotation import Annotation from statannotations.PValueFormat import PValueFormat, \ CONFIGURABLE_PARAMETERS as PVALUE_CONFIGURABLE_PARAMETERS +from statannotations._Plotter import _Plotter, _SeabornPlotter from statannotations.stats.ComparisonsCorrection import \ get_validated_comparisons_correction from statannotations.stats.StatResult import StatResult @@ -19,8 +18,7 @@ from statannotations.stats.test import apply_test, IMPLEMENTED_TESTS from statannotations.stats.utils import check_alpha, check_pvalues, \ check_num_comparisons, get_num_comparisons -from statannotations.utils import check_is_in, remove_null, check_not_none, \ - check_box_pairs_in_data, check_order_in_data, render_collection +from statannotations.utils import check_is_in, render_collection CONFIGURABLE_PARAMETERS = [ 'alpha', @@ -28,7 +26,7 @@ 'comparisons_correction', 'line_height', 'line_offset', - 'line_offset_to_box', + 'line_offset_to_group', 'line_width', 'loc', 'pvalue_format', @@ -40,8 +38,6 @@ 'verbose', ] -IMPLEMENTED_PLOTTERS = ['barplot', 'boxplot', 'stripplot', 'swarmplot', - 'violinplot'] _DEFAULT_VALUES = { "alpha": 0.05, @@ -61,57 +57,50 @@ "custom_annotations": None } - -def check_implemented_plotters(plot): - if plot not in IMPLEMENTED_PLOTTERS: - raise NotImplementedError( - f"Only {render_collection(IMPLEMENTED_PLOTTERS)} are supported.") +ENGINE_PLOTTERS = {"seaborn": _SeabornPlotter} class Annotator: """ Optionally computes statistical test between pairs of data series, and add - statistical annotation on top of the boxes/bars. The same exact arguments - `data`, `x`, `y`, `hue`, `order`, `width`, `hue_order` (and `units`) as in - the seaborn boxplot/barplot function must be passed to this function. - - This function works in one of the two following modes: - a) `perform_stat_test` is True (default): - * statistical test as given by argument `test` is performed. - * The `test_short_name` argument can be used to customize what appears - before the pvalues if test is a string. - b) `perform_stat_test` is False: no statistical test is performed, list of - custom p-values `pvalues` are used for each pair of boxes. + statistical annotation on top of the groups (boxes, bars...). The same + exact arguments provided to the seaborn plotting function must be passed to + the constructor. + + This Annotator works in one of the three following modes: + - Add custom text annotations (`set_custom_annotations`) + - Format pvalues and add them to the plot (`set_pvalues`) + - Perform a statistical test and then add the results to the plot + (`apply_test`) """ - def __init__(self, ax, box_pairs, plot='boxplot', data=None, x=None, - y=None, hue=None, order=None, hue_order=None, plot_params): + def __init__(self, ax, pairs, plot='boxplot', data=None, x=None, + y=None, hue=None, order=None, hue_order=None, + engine="seaborn", verbose=True, plot_params): """ :param ax: Ax of existing plot - :param box_pairs: can be of either form: - For non-grouped boxplot: `[(cat1, cat2), (cat3, cat4)]`. - For boxplot grouped by hue: `[ + :param pairs: can be of either form: + For non-grouped plots: `[(cat1, cat2), (cat3, cat4)]`. + For plots grouped by hue: `[ ((cat1, hue1), (cat2, hue2)), ((cat3, hue3), (cat4, hue4)) ]` - :param plot: type of the plot, one of 'boxplot' or 'barplot'. + :param plot: type of the plot (default 'boxplot') :param data: seaborn plot's data :param x: seaborn plot's x :param y: seaborn plot's y :param hue: seaborn plot's hue :param order: seaborn plot's order :param hue_order: seaborn plot's hue_order - :param plot_params: Other parameters for seaborn plotter + :param engine: currently only "seaborn" is implemented + :param verbose: verbosity flag + :param plot_params: Other parameters for plotter engine """ - self.box_pairs = box_pairs + self.pairs = pairs self.ax = ax - basic_plot = self.get_basic_plot(ax, box_pairs, plot, data, x, y, hue, - order, hue_order, plot_params) - - self.y_stack_arr, self.box_struct_pairs, self.fig = basic_plot - - self._reordering = None - self._sort_box_struct_pairs() + self._plotter = self._get_plotter(engine, ax, pairs, plot, data, x, y, + hue, order, hue_order, + verbose=verbose, plot_params) self._test = None self.perform_stat_test = None @@ -123,11 +112,11 @@ def __init__(self, ax, box_pairs, plot='boxplot', data=None, x=None, self._comparisons_correction = None self._loc = "inside" - self.verbose = 1 + self._verbose = 1 self._just_configured = True self.show_test_name = True self.use_fixed_offset = False - self.line_offset_to_box = None + self.line_offset_to_group = None self.line_offset = None self.line_height = _DEFAULT_VALUES["line_height"] self.text_offset = 1 @@ -137,57 +126,44 @@ def __init__(self, ax, box_pairs, plot='boxplot', data=None, x=None, self.y_offset = None self.custom_annotations = None - self.OFFSET_FUNCS = {"inside": Annotator._get_offsets_inside, - "outside": Annotator._get_offsets_outside} - - @staticmethod - def get_basic_plot( - ax, box_pairs, plot='boxplot', data=None, x=None, y=None, hue=None, - order=None, hue_order=None, plot_params): - check_implemented_plotters(plot) - check_not_none("box_pairs", box_pairs) - check_order_in_data(data, x, order) - check_box_pairs_in_data(box_pairs, data, x, hue, hue_order) - - plotter = Annotator._get_plotter(plot, x, y, hue, data, order, - hue_order, plot_params) - - box_names, labels = Annotator._get_box_names_and_labels(plotter) - - fig = plt.gcf() - ymaxes = Annotator._generate_ymaxes(ax, fig, plotter, box_names) - box_structs = Annotator._get_box_structs( - plotter, box_names, labels, ymaxes) - - box_struct_pairs = Annotator._get_box_struct_pairs( - box_pairs, box_structs) + def new_plot(self, ax, pairs=None, plot='boxplot', data=None, x=None, + y=None, hue=None, order=None, hue_order=None, + engine: str = "seaborn", plot_params): + self.ax = ax - y_stack_arr = np.array( - [[box_struct['x'], box_struct['ymax'], 0] - for box_struct in box_structs] - ).T + if pairs is None: + pairs = self.pairs - return y_stack_arr, box_struct_pairs, fig + self._plotter: _Plotter = self._get_plotter( + engine, ax, pairs, plot, data, x, y, hue, order, hue_order, + plot_params) - def new_plot(self, ax, box_pairs=None, plot='boxplot', data=None, x=None, - y=None, hue=None, order=None, hue_order=None, plot_params): - self.ax = ax + self.line_offset = None + self.line_offset_to_group = None + self.perform_stat_test = None - if box_pairs is None: - box_pairs = self.box_pairs + return self - basic_plot = self.get_basic_plot(ax, box_pairs, plot, data, x, y, hue, - order, hue_order, plot_params) + @property + def verbose(self): + return self._verbose - self.y_stack_arr, self.box_struct_pairs, self.fig = basic_plot + @verbose.setter + def verbose(self, verbose): + self._verbose = verbose + self._plotter.verbose = verbose - self._sort_box_struct_pairs() + @property + def _y_stack_arr(self): + return self._plotter.y_stack_arr - self.line_offset = None - self.line_offset_to_box = None - self.perform_stat_test = None + @property + def fig(self): + return self._plotter.fig - return self + @property + def _struct_pairs(self): + return self._plotter.struct_pairs def reset_configuration(self): @@ -196,7 +172,7 @@ def reset_configuration(self): return self - def annotate(self, line_offset=None, line_offset_to_box=None): + def annotate(self, line_offset=None, line_offset_to_group=None): """Add configured annotations to the plot.""" if self._should_warn_about_configuration: warnings.warn("Annotator was reconfigured without applying the " @@ -208,28 +184,26 @@ def annotate(self, line_offset=None, line_offset_to_box=None): ann_list = [] orig_ylim = self.ax.get_ylim() - offset_func = self.OFFSET_FUNCS[self.loc] - self.y_offset, self.line_offset_to_box = offset_func( - line_offset, line_offset_to_box) + offset_func = self.get_offset_func(self.loc) + self.y_offset, self.line_offset_to_group = offset_func( + line_offset, line_offset_to_group) - if self.verbose: + if self._verbose: self.print_pvalue_legend() - ax_to_data = Annotator._get_transform_func(self.ax, 'ax_to_data') + ax_to_data = self._plotter.get_transform_func('ax_to_data') self.validate_test_short_name() - annotations = self.custom_annotations or self.annotations - for box_structs, result in zip(self.box_struct_pairs, - annotations): - self._annotate_pair(box_structs, result, + for annotation in self.annotations: + self._annotate_pair(annotation, ax_to_data=ax_to_data, ann_list=ann_list, orig_ylim=orig_ylim) # reset transformation - y_stack_max = max(self.y_stack_arr[1, :]) - ax_to_data = Annotator._get_transform_func(self.ax, 'ax_to_data') + y_stack_max = max(self._y_stack_arr[1, :]) + ax_to_data = self._plotter.get_transform_func('ax_to_data') ylims = ([(0, 0), (0, max(1.04 * y_stack_max, 1))] if self.loc == 'inside' else [(0, 0), (0, 1)]) @@ -250,7 +224,7 @@ def configure(self, *parameters): `comparisons_correction` * `line_height`: in axes fraction coordinates * `line_offset` - * `line_offset_to_box` + * `line_offset_to_group` * `line_width` * `loc` * `pvalue_format` @@ -283,13 +257,7 @@ def configure(self, parameters): self._activate_configured_warning() - if parameters.get("alpha"): - pvalue_format = parameters.get("pvalue_format") - if (pvalue_format is None - or pvalue_format.get("pvalue_thresholds") is None): - warnings.warn("Changing alpha without updating " - "pvalue_thresholds can result in inconsistent " - "plotting results") + self._warn_alpha_thresholds_if_necessary(parameters) return self def apply_test(self, num_comparisons='auto', stats_params): @@ -297,7 +265,7 @@ def apply_test(self, num_comparisons='auto', stats_params): :param stats_params: Parameters for statistical test functions. :param num_comparisons: Override number of comparisons otherwise - calculated with number of box_pairs + calculated with number of pairs """ self._check_test_pvalues_perform() @@ -319,9 +287,9 @@ def set_pvalues_and_annotate(self, pvalues, num_comparisons='auto'): def set_pvalues(self, pvalues, num_comparisons='auto'): """ - :param pvalues: list or array of p-values for each box pair comparison. + :param pvalues: list or array of p-values for each pair comparison. :param num_comparisons: Override number of comparisons otherwise - calculated with number of box_pairs + calculated with number of pairs """ self.perform_stat_test = False @@ -340,14 +308,24 @@ def set_pvalues(self, pvalues, num_comparisons='auto'): def set_custom_annotations(self, text_annot_custom): """ :param text_annot_custom: List of strings to annotate for each - `box_pair` + `pair` """ self._check_correct_number_custom_annotations(text_annot_custom) - self.custom_annotations = text_annot_custom + self.annotations = [Annotation(struct, text) for + struct, text in zip(self._struct_pairs, + text_annot_custom)] self.show_test_name = False self._deactivate_configured_warning() return self + def annotate_custom_annotations(self, text_annot_custom): + """ + :param text_annot_custom: List of strings to annotate for each + `pair` + """ + self.set_custom_annotations(text_annot_custom) + return self.annotate() + def get_configuration(self): configuration = {key: getattr(self, key) for key in CONFIGURABLE_PARAMETERS} @@ -356,15 +334,14 @@ def get_configuration(self): def get_annotations_text(self): if self.annotations is not None: - return [self._get_text(self.annotations[new_idx]) + return [self.annotations[new_idx].text for new_idx in self._reordering] - - if self.custom_annotations is not None: - return [self.custom_annotations[new_idx] - for new_idx in self._reordering] - return None + @property + def _reordering(self): + return self._plotter.reordering + @property def alpha(self): return self._alpha @@ -421,6 +398,11 @@ def pvalue_format(self, pvalue_format_config): def _should_warn_about_configuration(self): return self._just_configured + @classmethod + def get_offset_func(cls, position): + return {"inside": cls._get_offsets_inside, + "outside": cls._get_offsets_outside}[position] + def _activate_configured_warning(self): self._just_configured = True @@ -432,59 +414,50 @@ def print_pvalue_legend(self): self._pvalue_format.print_legend_if_used() def _get_results(self, num_comparisons, pvalues=None, - stats_params) -> List[StatResult]: + stats_params) -> List[Annotation]: test_result_list = [] if num_comparisons == "auto": - num_comparisons = len(self.box_struct_pairs) + num_comparisons = len(self._struct_pairs) - for box_struct1, box_struct2 in self.box_struct_pairs: - box1 = box_struct1['box'] - box2 = box_struct2['box'] + for group_struct1, group_struct2 in self._struct_pairs: + group1 = group_struct1['group'] + group2 = group_struct2['group'] if self.perform_stat_test: result = self._get_stat_result_from_test( - box_struct1, box_struct2, num_comparisons, stats_params) + group_struct1, group_struct2, num_comparisons, + *stats_params) else: - result = self._get_custom_results(box_struct1, pvalues) + result = self._get_custom_results(group_struct1, pvalues) - result.box1 = box1 - result.box2 = box2 + result.group1 = group1 + result.group2 = group2 - test_result_list.append(result) + test_result_list.append(Annotation((group_struct1, group_struct2), + result, + formatter=self.pvalue_format)) # Perform other types of correction methods for multiple testing - Annotator._apply_comparisons_correction(self.comparisons_correction, - test_result_list) + self._apply_comparisons_correction(test_result_list) return test_result_list - def _get_text(self, result): - return self._pvalue_format.format_result(result) - - def _annotate_pair(self, box_structs, result, ax_to_data, - ann_list, orig_ylim): - - if self.custom_annotations is not None: - i_box_pair = box_structs[0]['i_box_pair'] - text = self.custom_annotations[i_box_pair] - - else: - if self.verbose >= 1: - _print_result_line(box_structs, result.formatted_output) + def _annotate_pair(self, annotation, ax_to_data, ann_list, orig_ylim): - text = self._get_text(result) + if self._verbose >= 1: + annotation.print_labels_and_content() - x1 = box_structs[0]['x'] - x2 = box_structs[1]['x'] - xi1 = box_structs[0]['xi'] - xi2 = box_structs[1]['xi'] + x1 = annotation.structs[0]['x'] + x2 = annotation.structs[1]['x'] + xi1 = annotation.structs[0]['xi'] + xi2 = annotation.structs[1]['xi'] - # Find y maximum for all the y_stacks in between* box1 and box2 + # Find y maximum for all the y_stacks in between group1 and group2 i_ymax_in_range_x1_x2 = xi1 + np.nanargmax( - self.y_stack_arr[1, np.where((x1 <= self.y_stack_arr[0, :]) - & (self.y_stack_arr[0, :] <= x2))]) + self._y_stack_arr[1, np.where((x1 <= self._y_stack_arr[0, :]) + & (self._y_stack_arr[0, :] <= x2))]) y = self._get_y_for_pair(i_ymax_in_range_x1_x2) @@ -499,13 +472,13 @@ def _annotate_pair(self, box_structs, result, ax_to_data, line_x, line_y = zip(points) self._plot_line(line_x, line_y) - if text is not None: - ann = self.ax.annotate( - text, xy=(np.mean([x1, x2]), line_y[2]), - xytext=(0, self.text_offset), textcoords='offset points', - xycoords='data', ha='center', va='bottom', - fontsize=self._pvalue_format.fontsize, clip_on=False, - annotation_clip=False) + ann = self.ax.annotate( + annotation.text, xy=(np.mean([x1, x2]), line_y[2]), + xytext=(0, self.text_offset), textcoords='offset points', + xycoords='data', ha='center', va='bottom', + fontsize=self._pvalue_format.fontsize, clip_on=False, + annotation_clip=False) + if annotation.text is not None: ann_list.append(ann) plt.draw() self.ax.set_ylim(orig_ylim) @@ -516,205 +489,16 @@ def _annotate_pair(self, box_structs, result, ax_to_data, # Fill the highest y position of the annotation into the y_stack array # for all positions in the range x1 to x2 - self.y_stack_arr[1, - (x1 <= self.y_stack_arr[0, :]) - & (self.y_stack_arr[0, :] <= x2)] = y_top_annot + self._y_stack_arr[1, + (x1 <= self._y_stack_arr[0, :]) + & (self._y_stack_arr[0, :] <= x2)] = y_top_annot # Increment the counter of annotations in the y_stack array - self.y_stack_arr[2, xi1:xi2 + 1] += 1 + self._y_stack_arr[2, xi1:xi2 + 1] += 1 def _update_y_for_loc(self): if self._loc == 'outside': - self.y_stack_arr[1, :] = 1 - - @staticmethod - def _get_box_x_position(b_plotter, box_name): - """ - box_name can be either a name "cat" or a tuple ("cat", "hue") - """ - if b_plotter.plot_hues is None: - cat = box_name - hue_offset = 0 - else: - cat = box_name[0] - hue_level = box_name[1] - hue_offset = b_plotter.hue_offsets[ - b_plotter.hue_names.index(hue_level)] - - group_pos = b_plotter.group_names.index(cat) - box_pos = group_pos + hue_offset - return box_pos - - @staticmethod - def _get_xpos_location(pos, xranges): - """ - Finds the x-axis location of a categorical variable - """ - for xrange in xranges: - if (pos >= xrange[0]) & (pos <= xrange[1]): - return xrange[2] - - @staticmethod - def _get_ypos_for_line2d_or_rectangle( - xpositions, data_to_ax, ax, fig, child): - xunits = (max(list(xpositions.keys())) + 1) / len(xpositions) - xranges = {(pos - xunits / 2, pos + xunits / 2, pos): box_name - for pos, box_name in xpositions.items()} - box = ax.transData.inverted().transform( - child.get_window_extent(fig.canvas.get_renderer())) - - if (box[:, 0].max() - box[:, 0].min()) > 1.1 xunits: - return None, None - raw_xpos = np.round(box[:, 0].mean(), 1) - xpos = Annotator._get_xpos_location(raw_xpos, xranges) - if xpos not in xpositions: - return None, None - xname = xpositions[xpos] - ypos = box[:, 1].max() - ypos = data_to_ax.transform((0, ypos))[1] - - return xname, ypos - - @staticmethod - def _get_ypos_for_path_collection(xpositions, data_to_ax, child): - ymax = child.properties()['offsets'][:, 1].max() - xpos = float(np.round(np.nanmean( - child.properties()['offsets'][:, 0]), 1)) - - xname = xpositions[xpos] - ypos = data_to_ax.transform((0, ymax))[1] - - return xname, ypos - - @staticmethod - def _generate_ymaxes(ax, fig, plotter, box_names): - """ - given box plotter and the names of two categorical variables, - returns highest y point drawn between those two variables before - annotations. - - The highest data point is often not the highest item drawn - (eg, error bars and/or bar charts). - """ - xpositions = { - np.round(Annotator._get_box_x_position(plotter, box_name), - 1): box_name - for box_name in box_names} - - ymaxes = {name: 0 for name in box_names} - - data_to_ax = Annotator._get_transform_func(ax, 'data_to_ax') - # noinspection PyProtectedMember - if isinstance(plotter, sns.categorical._ViolinPlotter): - ymaxes = Annotator._get_ymaxes_violin(plotter, ymaxes, data_to_ax) - - else: - for child in ax.get_children(): - - xname, ypos = Annotator._get_child_ypos( - child, ax, fig, xpositions, data_to_ax) - - if ypos is not None and ypos > ymaxes[xname]: - ymaxes[xname] = ypos - - return ymaxes - - @staticmethod - def _get_child_ypos(child, ax, fig, xpositions, data_to_ax): - if ((type(child) == PathCollection) - and len(child.properties()['offsets'])): - return Annotator._get_ypos_for_path_collection( - xpositions, data_to_ax, child) - - elif (type(child) == lines.Line2D) or (type(child) == Rectangle): - return Annotator._get_ypos_for_line2d_or_rectangle( - xpositions, data_to_ax, ax, fig, child) - - return None, None - - @staticmethod - def _get_ymaxes_violin(plotter, ymaxes, data_to_ax): - for group_idx, group_name in enumerate(plotter.group_names): - if plotter.hue_names: - for hue_idx, hue_name in enumerate(plotter.hue_names): - ypos = max(plotter.support[group_idx][hue_idx]) - ymaxes[(group_name, hue_name)] = \ - data_to_ax.transform((0, ypos))[1] - else: - ypos = max(plotter.support[group_idx]) - ymaxes[group_name] = data_to_ax.transform((0, ypos))[1] - return ymaxes - - @staticmethod - def _get_group_data(b_plotter, cat): - index = b_plotter.group_names.index(cat) - - return b_plotter.plot_data[index], index - - @staticmethod - def _get_box_data_with_hue(b_plotter, box_name): - cat = box_name[0] - group_data, index = Annotator._get_group_data(b_plotter, cat) - - hue_level = box_name[1] - hue_mask = b_plotter.plot_hues[index] == hue_level - - return group_data[hue_mask] - - @staticmethod - def _get_box_data_without_hue(b_plotter, box_name): - cat = box_name - group_data, _ = Annotator._get_group_data(b_plotter, cat) - - return group_data - - @staticmethod - def _get_box_data(b_plotter, box_name): - """ - box_name can be either a name "cat" or a tuple ("cat", "hue") - - Almost a duplicate of seaborn's code, because there is not - direct access to the box_data in the BoxPlotter class. - """ - if b_plotter.plot_hues is None: - data = Annotator._get_box_data_without_hue(b_plotter, box_name) - else: - data = Annotator._get_box_data_with_hue(b_plotter, box_name) - - box_data = remove_null(data) - - return box_data - - @staticmethod - def _get_transform_func(ax, kind): - """ - Given an axis object, returns one of three possible transformation - functions to move between coordinate systems, depending on the value of - kind: - 'data_to_ax': converts data coordinates to axes coordinates - 'ax_to_data': converts axes coordinates to data coordinates - 'pix_to_ax': converts pixel coordinates to axes coordinates - 'all': return tuple of all three - - This function should be called whenever axes limits are altered. - """ - - check_is_in(kind, ['data_to_ax', 'ax_to_data', 'pix_to_ax', 'all'], - 'kind') - - if kind == 'pix_to_ax': - return ax.transAxes.inverted() - - data_to_ax = ax.transData + ax.get_xaxis_transform().inverted() - - if kind == 'data_to_ax': - return data_to_ax - - elif kind == 'ax_to_data': - return data_to_ax.inverted() - - else: - return data_to_ax, data_to_ax.inverted(), ax.transAxes.inverted() + self._y_stack_arr[1, :] = 1 def _check_test_pvalues_perform(self): if self.test is None: @@ -732,9 +516,9 @@ def _check_pvalues_no_perform(self, pvalues): raise ValueError("If `perform_stat_test` is False, " "custom `pvalues` must be specified.") check_pvalues(pvalues) - if len(pvalues) != len(self.box_pairs): + if len(pvalues) != len(self.pairs): raise ValueError("`pvalues` should be of the same length as " - "`box_pairs`.") + "`pairs`.") def _check_test_no_perform(self): if self.test is not None: @@ -745,53 +529,24 @@ def _check_correct_number_custom_annotations(self, text_annot_custom): if text_annot_custom is None: return - if len(text_annot_custom) != len(self.box_struct_pairs): + if len(text_annot_custom) != len(self._struct_pairs): raise ValueError( - "`text_annot_custom` should be of same length as `box_pairs`.") - - @staticmethod - def _apply_type1_comparisons_correction(comparisons_correction, - test_result_list): - original_pvalues = [result.pvalue for result in test_result_list] + "`text_annot_custom` should be of same length as `pairs`.") - significant_pvalues = comparisons_correction(original_pvalues) - correction_name = comparisons_correction.name - for is_significant, result in zip(significant_pvalues, - test_result_list): - result.correction_method = correction_name - result.corrected_significance = is_significant - - @staticmethod - def _apply_type0_comparisons_correction(comparisons_correction, - test_result_list): - alpha = comparisons_correction.alpha - for result in test_result_list: - result.correction_method = comparisons_correction.name - result.corrected_significance = ( - result.pvalue < alpha - or np.isclose(result.pvalue, alpha)) - - @staticmethod - def _apply_comparisons_correction(comparisons_correction, - test_result_list): - if comparisons_correction is None: + def _apply_comparisons_correction(self, annotations): + if self.comparisons_correction is None: return - - if comparisons_correction.type == 1: - Annotator._apply_type1_comparisons_correction( - comparisons_correction, test_result_list) - else: - Annotator._apply_type0_comparisons_correction( - comparisons_correction, test_result_list) + self.comparisons_correction.apply( + [annotation.data for annotation in annotations]) - def _get_stat_result_from_test(self, box_struct1, box_struct2, + def _get_stat_result_from_test(self, group_struct1, group_struct2, num_comparisons, *stats_params) -> StatResult: result = apply_test( - box_struct1['box_data'], - box_struct2['box_data'], + group_struct1['group_data'], + group_struct2['group_data'], self.test, comparisons_correction=self.comparisons_correction, num_comparisons=num_comparisons, @@ -801,8 +556,8 @@ def _get_stat_result_from_test(self, box_struct1, box_struct2, return result - def _get_custom_results(self, box_struct1, pvalues) -> StatResult: - pvalue = pvalues[box_struct1['i_box_pair']] + def _get_custom_results(self, group_struct1, pvalues) -> StatResult: + pvalue = pvalues[group_struct1['i_group_pair']] if self.has_type0_comparisons_correction(): pvalue = self.comparisons_correction(pvalue, len(pvalues)) @@ -819,162 +574,28 @@ def _get_custom_results(self, box_struct1, pvalues) -> StatResult: return result @staticmethod - def _get_box_struct_pairs(box_pairs, box_structs): - box_structs_dic = {box_struct['box']: box_struct - for box_struct in box_structs} - - box_struct_pairs = [] - - for i_box_pair, (box1, box2) in enumerate(box_pairs): - box_struct1 = dict(box_structs_dic[box1], i_box_pair=i_box_pair) - box_struct2 = dict(box_structs_dic[box2], i_box_pair=i_box_pair) - - if box_struct1['x'] <= box_struct2['x']: - box_struct_pairs.append((box_struct1, box_struct2)) - else: - box_struct_pairs.append((box_struct2, box_struct1)) - - return box_struct_pairs - - @staticmethod - def absolute_box_struct_pair_in_tuple_x_diff(box_struct_pair): - return abs(box_struct_pair[1][1]['x'] - box_struct_pair[1][0]['x']) - - def _sort_box_struct_pairs(self): - # Draw first the annotations with the shortest between-boxes distance, - # in order to reduce overlapping between annotations. - new_box_struct_pairs = sorted( - enumerate(self.box_struct_pairs), - key=self.absolute_box_struct_pair_in_tuple_x_diff) - - self._reordering, self.box_struct_pairs = zip(new_box_struct_pairs) - - # noinspection PyProtectedMember - @staticmethod - def _get_plotter(plot, x, y, hue, data, order, hue_order, - *plot_params): - if not plot_params.pop("dodge", True): - raise ValueError("`dodge` must be True in statannotations") - - if plot == 'boxplot': - plotter = sns.categorical._BoxPlotter( - - x, y, hue, data, order, hue_order, - orient=plot_params.get("orient"), - width=plot_params.get("width", 0.8), - dodge=True, - fliersize=plot_params.get("fliersize", 5), - linewidth=plot_params.get("linewidth"), - saturation=.75, color=None, palette=None) - - elif plot == 'swarmplot': - plotter = sns.categorical._SwarmPlotter( - x, y, hue, data, order, hue_order, - orient=plot_params.get("orient"), - dodge=True, color=None, palette=None) - - elif plot == 'stripplot': - plotter = sns.categorical._StripPlotter( - x, y, hue, data, order, hue_order, - jitter=plot_params.get("jitter", True), - orient=plot_params.get("orient"), - dodge=True, color=None, palette=None) - - elif plot == 'barplot': - plotter = sns.categorical._BarPlotter( - x, y, hue, data, order, hue_order, - estimator=plot_params.get("estimator", np.mean), - ci=plot_params.get("ci", 95), - n_boot=plot_params.get("nboot", 1000), - units=plot_params.get("units"), - orient=plot_params.get("orient"), - seed=plot_params.get("seed"), - color=None, palette=None, saturation=.75, - errcolor=".26", errwidth=plot_params.get("errwidth"), - capsize=None, - dodge=True) - - elif plot == "violinplot": - plotter = sns.categorical._ViolinPlotter( - x, y, hue, data, order, hue_order, - bw=plot_params.get("bw", "scott"), - cut=plot_params.get("cut", 2), - scale=plot_params.get("scale", "area"), - scale_hue=plot_params.get("scale_hue", True), - gridsize=plot_params.get("gridsize", 100), - width=plot_params.get("width", 0.8), - inner=plot_params.get("inner", None), - split=plot_params.get("split", False), - dodge=True, orient=plot_params.get("orient"), - linewidth=plot_params.get("linewidth"), color=None, - palette=None, saturation=.75) - - else: - raise NotImplementedError( - f"Only {render_collection(IMPLEMENTED_PLOTTERS)} are " - f"supported.") - - return plotter - - @staticmethod - def _get_offsets_inside(line_offset, line_offset_to_box): + def _get_offsets_inside(line_offset, line_offset_to_group): if line_offset is None: line_offset = 0.05 - if line_offset_to_box is None: - line_offset_to_box = 0.06 + if line_offset_to_group is None: + line_offset_to_group = 0.06 else: - if line_offset_to_box is None: - line_offset_to_box = 0.06 + if line_offset_to_group is None: + line_offset_to_group = 0.06 y_offset = line_offset - return y_offset, line_offset_to_box + return y_offset, line_offset_to_group @staticmethod - def _get_offsets_outside(line_offset, line_offset_to_box): + def _get_offsets_outside(line_offset, line_offset_to_group): if line_offset is None: line_offset = 0.03 - if line_offset_to_box is None: - line_offset_to_box = line_offset + if line_offset_to_group is None: + line_offset_to_group = line_offset else: - line_offset_to_box = line_offset + line_offset_to_group = line_offset y_offset = line_offset - return y_offset, line_offset_to_box - - @staticmethod - def _get_box_structs(box_plotter, box_names, labels, ymaxes): - box_structs = [ - { - 'box': box_names[i], - 'label': labels[i], - 'x': Annotator._get_box_x_position(box_plotter, box_names[i]), - 'box_data': Annotator._get_box_data(box_plotter, box_names[i]), - 'ymax': ymaxes[box_names[i]] - } for i in range(len(box_names))] - - # Sort the box data structures by position along the x axis - box_structs = sorted(box_structs, key=lambda a: a['x']) - - # Add the index position in the list of boxes along the x axis - box_structs = [dict(box_struct, xi=i) - for i, box_struct in enumerate(box_structs)] - - return box_structs - - @staticmethod - def _get_box_names_and_labels(box_plotter): - group_names = box_plotter.group_names - hue_names = box_plotter.hue_names - - if box_plotter.plot_hues is None: - box_names = group_names - labels = box_names - else: - box_names = [(group_name, hue_name) for group_name in group_names - for hue_name in hue_names] - labels = ['{}_{}'.format(group_name, hue_name) - for (group_name, hue_name) in box_names] - - return box_names, labels + return y_offset, line_offset_to_group def validate_test_short_name(self): if self.test_short_name is not None: @@ -1010,7 +631,7 @@ def _annotate_pair_text(self, ann, y): if not self.use_fixed_offset: try: bbox = ann.get_window_extent() - pix_to_ax = Annotator._get_transform_func(self.ax, 'pix_to_ax') + pix_to_ax = self._plotter.get_transform_func('pix_to_ax') bbox_ax = bbox.transformed(pix_to_ax) y_top_annot = bbox_ax.ymax @@ -1018,7 +639,7 @@ def _annotate_pair_text(self, ann, y): got_mpl_error = True if self.use_fixed_offset or got_mpl_error: - if self.verbose >= 1: + if self._verbose >= 1: print("Warning: cannot get the text bounding box. Falling " "back to a fixed y offset. Layout may be not " "optimal.") @@ -1046,22 +667,33 @@ def _reset_default_values(self): def _get_y_for_pair(self, i_ymax_in_range_x1_x2): - ymax_in_range_x1_x2 = self.y_stack_arr[1, i_ymax_in_range_x1_x2] + ymax_in_range_x1_x2 = self._y_stack_arr[1, i_ymax_in_range_x1_x2] # Choose the best offset depending on whether there is an annotation # below at the x position in the range [x1, x2] where the stack is the # highest - if self.y_stack_arr[2, i_ymax_in_range_x1_x2] == 0: - # there is only a box below - offset = self.line_offset_to_box + if self._y_stack_arr[2, i_ymax_in_range_x1_x2] == 0: + # there is only a group below + offset = self.line_offset_to_group else: # there is an annotation below offset = self.y_offset + self.text_offset_impact_above return ymax_in_range_x1_x2 + offset + @staticmethod + def _warn_alpha_thresholds_if_necessary(parameters): + if parameters.get("alpha"): + pvalue_format = parameters.get("pvalue_format") + if (pvalue_format is None + or pvalue_format.get("pvalue_thresholds") is None): + warnings.warn("Changing alpha without updating " + "pvalue_thresholds can result in inconsistent " + "plotting results") -def _print_result_line(box_structs, formatted_output): - label1 = box_structs[0]['label'] - label2 = box_structs[1]['label'] - print(f"{label1} v.s. {label2}: {formatted_output}") + @staticmethod + def _get_plotter(engine, args, *kwargs): + engine_plotter = ENGINE_PLOTTERS.get(engine) + if engine_plotter is None: + raise NotImplementedError(f"{engine} engine not implemented.") + return engine_plotter(args, *kwargs) diff --git a/statannotations/PValueFormat.py b/statannotations/PValueFormat.py index be31722..e002992 100644 --- a/statannotations/PValueFormat.py +++ b/statannotations/PValueFormat.py @@ -12,8 +12,20 @@ ] -class PValueFormat: +class Formatter: def __init__(self): + pass + + def config(self, args, kwargs): + pass + + def format_data(self, data): + pass + + +class PValueFormat(Formatter): + def __init__(self): + Formatter.__init__(self) self._pvalue_format_string = '{:.3e}' self._simple_format_string = '{:.2f}' self._text_format = "star" @@ -144,7 +156,7 @@ def _update_pvalue_thresholds(self): self._pvalue_thresholds = self._get_pvalue_thresholds( self._pvalue_thresholds) - def format_result(self, result): + def format_data(self, result): if self.text_format == 'full': return ("{} p = {}{}" .format('{}', self.pvalue_format_string, '{}') diff --git a/statannotations/_Plotter.py b/statannotations/_Plotter.py new file mode 100644 index 0000000..9dac681 --- /dev/null +++ b/statannotations/_Plotter.py @@ -0,0 +1,379 @@ +import itertools +import warnings + +import matplotlib.pyplot as plt +import numpy as np +import seaborn as sns +from matplotlib import lines +from matplotlib.collections import PathCollection +from matplotlib.patches import Rectangle + +from statannotations._Xpositions import _XPositions +from statannotations.utils import check_not_none, check_order_in_data, \ + check_pairs_in_data, render_collection, check_is_in, remove_null + +IMPLEMENTED_PLOTTERS = { + 'seaborn': ['barplot', 'boxplot', 'stripplot', 'swarmplot', 'violinplot'] +} + + +class _Plotter: + def __init__(self, ax, pairs, data=None, x=None, hue=None, + order=None, hue_order=None, verbose=False): + self.ax = ax + self._fig = plt.gcf() + check_not_none("pairs", pairs) + check_order_in_data(data, x, order) + check_pairs_in_data(pairs, data, x, hue, hue_order) + self.pairs = pairs + self._struct_pairs = None + self.verbose = verbose + + def get_transform_func(self, kind: str): + """ + Given an axis object, returns one of three possible transformation + functions to move between coordinate systems, depending on the value of + kind: + 'data_to_ax': converts data coordinates to axes coordinates + 'ax_to_data': converts axes coordinates to data coordinates + 'pix_to_ax': converts pixel coordinates to axes coordinates + 'all': return tuple of all three + + This function should be called whenever axes limits are altered. + """ + + check_is_in(kind, ['data_to_ax', 'ax_to_data', 'pix_to_ax', 'all'], + 'kind') + + if kind == 'pix_to_ax': + return self.ax.transAxes.inverted() + + data_to_ax = \ + self.ax.transData + self.ax.get_xaxis_transform().inverted() + + if kind == 'data_to_ax': + return data_to_ax + + elif kind == 'ax_to_data': + return data_to_ax.inverted() + + else: + return (data_to_ax, data_to_ax.inverted(), + self.ax.transAxes.inverted()) + + @property + def fig(self): + return self._fig + + @property + def struct_pairs(self): + return self._struct_pairs + + +class _SeabornPlotter(_Plotter): + def __init__(self, ax, pairs, plot='boxplot', data=None, x=None, + y=None, hue=None, order=None, hue_order=None, verbose=False, + plot_params): + + _Plotter.__init__(self, ax, pairs, data, x, hue, order, hue_order, + verbose) + + self.check_plot_is_implemented(plot) + + self.plot = plot + self.plotter = self._get_plotter(plot, x, y, hue, data, order, + hue_order, plot_params) + + self.group_names, self.labels = self._get_group_names_and_labels() + self.xpositions = _XPositions(self.plotter, self.group_names) + self.reordering = None + self.ymaxes = self._generate_ymaxes() + + self.structs = self._get_structs() + self.pairs = pairs + self._struct_pairs = self._get_group_struct_pairs() + + self._y_stack_arr = np.array( + [[struct['x'], struct['ymax'], 0] for struct in self.structs] + ).T + + self.reordering = None + self._sort_group_struct_pairs() + + @property + def y_stack_arr(self): + return self._y_stack_arr + + # noinspection PyProtectedMember + def _get_plotter(self, plot, x, y, hue, data, order, hue_order, + plot_params): + dodge = plot_params.pop("dodge", None) + self.fix_and_warn(dodge, hue, plot) + + if plot == 'boxplot': + plotter = sns.categorical._BoxPlotter( + + x, y, hue, data, order, hue_order, + orient=plot_params.get("orient"), + width=plot_params.get("width", 0.8), + dodge=True, + fliersize=plot_params.get("fliersize", 5), + linewidth=plot_params.get("linewidth"), + saturation=.75, color=None, palette=None) + + elif plot == 'swarmplot': + plotter = sns.categorical._SwarmPlotter( + x, y, hue, data, order, hue_order, + orient=plot_params.get("orient"), + dodge=True, color=None, palette=None) + + elif plot == 'stripplot': + plotter = sns.categorical._StripPlotter( + x, y, hue, data, order, hue_order, + jitter=plot_params.get("jitter", True), + orient=plot_params.get("orient"), + dodge=True, color=None, palette=None) + + elif plot == 'barplot': + plotter = sns.categorical._BarPlotter( + x, y, hue, data, order, hue_order, + estimator=plot_params.get("estimator", np.mean), + ci=plot_params.get("ci", 95), + n_boot=plot_params.get("nboot", 1000), + units=plot_params.get("units"), + orient=plot_params.get("orient"), + seed=plot_params.get("seed"), + color=None, palette=None, saturation=.75, + errcolor=".26", errwidth=plot_params.get("errwidth"), + capsize=None, + dodge=True) + + elif plot == "violinplot": + plotter = sns.categorical._ViolinPlotter( + x, y, hue, data, order, hue_order, + bw=plot_params.get("bw", "scott"), + cut=plot_params.get("cut", 2), + scale=plot_params.get("scale", "area"), + scale_hue=plot_params.get("scale_hue", True), + gridsize=plot_params.get("gridsize", 100), + width=plot_params.get("width", 0.8), + inner=plot_params.get("inner", None), + split=plot_params.get("split", False), + dodge=True, orient=plot_params.get("orient"), + linewidth=plot_params.get("linewidth"), color=None, + palette=None, saturation=.75) + + else: + raise NotImplementedError( + f"Only {render_collection(IMPLEMENTED_PLOTTERS)} are " + f"supported.") + + return plotter + + def _get_group_names_and_labels(self): + + if self.plotter.plot_hues is None: + group_names = self.plotter.group_names + labels = group_names + + else: + labels = [] + group_names = [] + for group_name, hue_name in \ + itertools.product(self.plotter.group_names, + self.plotter.hue_names): + group_names.append((group_name, hue_name)) + labels.append(f'{group_name}_{hue_name}') + + return group_names, labels + + def _generate_ymaxes(self): + """ + given plotter and the names of two categorical variables, + returns highest y point drawn between those two variables before + annotations. + + The highest data point is often not the highest item drawn + (eg, error bars and/or bar charts). + """ + + ymaxes = {name: 0 for name in self.group_names} + + data_to_ax = self.get_transform_func('data_to_ax') + + if self.plot == 'violinplot': + ymaxes = self._get_ymaxes_violin(ymaxes, data_to_ax) + + else: + for child in self.ax.get_children(): + + xname, ypos = self._get_child_ypos(child, data_to_ax) + + if ypos is not None and ypos > ymaxes[xname]: + ymaxes[xname] = ypos + + return ymaxes + + def _get_structs(self): + structs = [ + { + 'group': group_name, + 'label': self.labels[b_idx], + 'x': self.xpositions.get_group_x_position(group_name), + 'group_data': self._get_group_data(group_name), + 'ymax': self.ymaxes[group_name] + } for b_idx, group_name in enumerate(self.group_names)] + + # Sort the group data structures by position along the x axis + structs = sorted(structs, key=lambda struct: struct['x']) + + # Add the index position in the list of groups along the x axis + structs = [dict(struct, xi=i) + for i, struct in enumerate(structs)] + + return structs + + def _get_group_struct_pairs(self): + group_structs_dic = {struct['group']: struct + for struct in self.structs} + + group_struct_pairs = [] + + for i_group_pair, (group1, group2) in enumerate(self.pairs): + group_struct1 = dict(group_structs_dic[group1], + i_group_pair=i_group_pair) + + group_struct2 = dict(group_structs_dic[group2], + i_group_pair=i_group_pair) + + if group_struct1['x'] <= group_struct2['x']: + group_struct_pairs.append((group_struct1, group_struct2)) + + else: + group_struct_pairs.append((group_struct2, group_struct1)) + + return group_struct_pairs + + def _get_group_data_with_hue(self, group_name): + cat = group_name[0] + group_data, index = self._get_group_data_from_plotter(cat) + + hue_level = group_name[1] + hue_mask = self.plotter.plot_hues[index] == hue_level + + return group_data[hue_mask] + + def _get_group_data_from_plotter(self, cat): + index = self.plotter.group_names.index(cat) + + return self.plotter.plot_data[index], index + + def _get_group_data_without_hue(self, group_name): + group_data, _ = self._get_group_data_from_plotter(group_name) + + return group_data + + def _get_group_data(self, group_name): + """ + group_name can be either a name "cat" or a tuple ("cat", "hue") + + Almost a duplicate of seaborn's code, because there is not + direct access to the group_data in the respective Plotter class. + """ + if self.plotter.plot_hues is None: + data = self._get_group_data_without_hue(group_name) + else: + data = self._get_group_data_with_hue(group_name) + + group_data = remove_null(data) + + return group_data + + def _get_ymaxes_violin(self, ymaxes, data_to_ax): + for group_idx, group_name in enumerate(self.plotter.group_names): + if self.plotter.hue_names: + for hue_idx, hue_name in enumerate(self.plotter.hue_names): + ypos = max(self.plotter.support[group_idx][hue_idx]) + ymaxes[(group_name, hue_name)] = \ + data_to_ax.transform((0, ypos))[1] + else: + ypos = max(self.plotter.support[group_idx]) + ymaxes[group_name] = data_to_ax.transform((0, ypos))[1] + return ymaxes + + def _get_child_ypos(self, child, data_to_ax): + if (type(child) == PathCollection + and len(child.properties()['offsets'])): + return self._get_ypos_for_path_collection( + child, data_to_ax) + + elif type(child) in (lines.Line2D, Rectangle): + return self._get_ypos_for_line2d_or_rectangle( + child, data_to_ax) + + return None, None + + def _get_ypos_for_path_collection(self, child, data_to_ax): + ymax = child.properties()['offsets'][:, 1].max() + xpos = float(np.round(np.nanmean( + child.properties()['offsets'][:, 0]), 1)) + if xpos not in self.xpositions.xpositions: + if self.verbose: + warnings.warn( + "Invalid x-position found. Are the same parameters passed " + "to seaborn and statannotations calls? or are there few " + "data points?") + xpos = self.xpositions.find_closest(xpos) + + xname = self.xpositions.xpositions[xpos] + ypos = data_to_ax.transform((0, ymax))[1] + + return xname, ypos + + def _get_ypos_for_line2d_or_rectangle(self, child, data_to_ax): + bbox = self.ax.transData.inverted().transform( + child.get_window_extent(self.fig.canvas.get_renderer())) + + if (bbox[:, 0].max() - bbox[:, 0].min()) > 1.1self.xpositions.xunits: + return None, None + raw_xpos = np.round(bbox[:, 0].mean(), 1) + xpos = self.xpositions.get_xpos_location(raw_xpos) + if xpos not in self.xpositions.xpositions: + return None, None + xname = self.xpositions.xpositions[xpos] + ypos = bbox[:, 1].max() + ypos = data_to_ax.transform((0, ypos))[1] + + return xname, ypos + + def _sort_group_struct_pairs(self): + # Draw first the annotations with the shortest between-groups distance, + # in order to reduce overlapping between annotations. + new_group_struct_pairs = sorted( + enumerate(self._struct_pairs), + key=self._absolute_group_struct_pair_in_tuple_x_diff) + + self.reordering, self._struct_pairs = zip(new_group_struct_pairs) + + @staticmethod + def _absolute_group_struct_pair_in_tuple_x_diff(group_struct_pair): + return abs(group_struct_pair[1][1]['x'] - group_struct_pair[1][0]['x']) + + @staticmethod + def check_plot_is_implemented(plot, engine="seaborn"): + if plot not in IMPLEMENTED_PLOTTERS[engine]: + raise NotImplementedError( + f"Only {render_collection(IMPLEMENTED_PLOTTERS[engine])} are " + f"supported with {engine} engine.") + + @staticmethod + def fix_and_warn(dodge, hue, plot): + if dodge is False and hue: + raise ValueError("`dodge` cannot be False in statannotations.") + + if plot in ("swarmplot", 'stripplot') and hue: + if dodge is None: + warnings.warn( + "Implicitly setting dodge to True as it is necessary in " + "statannotations. It must have been True for the seaborn " + "call to yield consistent results when using `hue`.") diff --git a/statannotations/_Xpositions.py b/statannotations/_Xpositions.py new file mode 100644 index 0000000..2520f3c --- /dev/null +++ b/statannotations/_Xpositions.py @@ -0,0 +1,68 @@ +import numpy as np + +from statannotations.utils import get_closest + + +class _XPositions: + def __init__(self, plotter, group_names): + self._plotter = plotter + self._hue_names = self._plotter.hue_names + + if self._hue_names is not None: + nb_hues = len(self._hue_names) + if nb_hues == 1: + raise ValueError( + "Using hues with only one hue is not supported.") + + self.hue_offsets = self._plotter.hue_offsets + self._xunits = self.hue_offsets[1] - self.hue_offsets[0] + + self._xpositions = { + np.round(self.get_group_x_position(group_name), 1): group_name + for group_name in group_names + } + + self._xpositions_list = sorted(self._xpositions.keys()) + + if self._hue_names is None: + self._xunits = ((max(list(self._xpositions.keys())) + 1) + / len(self._xpositions)) + + self._xranges = { + (pos - self._xunits / 2, pos + self._xunits / 2, pos): group_name + for pos, group_name in self._xpositions.items()} + + @property + def xpositions(self): + return self._xpositions + + @property + def xunits(self): + return self._xunits + + def get_xpos_location(self, pos): + """ + Finds the x-axis location of a categorical variable + """ + for xrange in self._xranges: + if (pos >= xrange[0]) & (pos <= xrange[1]): + return xrange[2] + + def get_group_x_position(self, group): + """ + group_name can be either a name "cat" or a tuple ("cat", "hue") + """ + if self._plotter.plot_hues is None: + cat = group + hue_offset = 0 + else: + cat = group[0] + hue_level = group[1] + hue_offset = self._plotter.hue_offsets[ + self._plotter.hue_names.index(hue_level)] + + group_pos = self._plotter.group_names.index(cat) + hue_offset + return group_pos + + def find_closest(self, xpos): + return get_closest(list(self._xpositions_list), xpos) diff --git a/statannotations/_version.py b/statannotations/_version.py index 8c2ecfb..6a9beea 100644 --- a/statannotations/_version.py +++ b/statannotations/_version.py @@ -1,1 +1,1 @@ -__version__ = "0.4.0a" +__version__ = "0.4.0" diff --git a/statannotations/format_annotations.py b/statannotations/format_annotations.py index 29ba57b..81bb6e8 100644 --- a/statannotations/format_annotations.py +++ b/statannotations/format_annotations.py @@ -2,48 +2,42 @@ from typing import Union, List import numpy as np import pandas as pd +from operator import itemgetter def pval_annotation_text(result: Union[List[StatResult], StatResult], - pvalue_thresholds: List): + pvalue_thresholds: List) -> Union[List[str], str]: """ :param result: StatResult instance or list thereof - :param pvalue_thresholds: thresholds to use for formatting - :returns: A Series of rendered annotations if a list of StatResults was + :param pvalue_thresholds: thresholds to use for formatter + :returns: A List of rendered annotations if a list of StatResults was provided, a string otherwise. """ - single_value = False + was_list = True - if isinstance(result, list): - x1_pval = np.array([res.pvalue for res in result]) - x1_signif_suff = [res.significance_suffix for res in result] - else: - x1_pval = np.array([result.pvalue]) - x1_signif_suff = [result.significance_suffix] - single_value = True + if not isinstance(result, list): + was_list = False + result = [result] + + x1_pval = np.array([res.pvalue for res in result]) - # Sort the threshold array - pvalue_thresholds = (pd.DataFrame(pvalue_thresholds) - .sort_values(by=0, ascending=False) - .values) - x_annot = pd.Series(["" for _ in range(len(x1_pval))]) + pvalue_thresholds = sorted(pvalue_thresholds, key=itemgetter(0), + reverse=True) - for i in range(0, len(pvalue_thresholds)): + x_annot = pd.Series(["" for _ in x1_pval]) - if i < len(pvalue_thresholds) - 1: - condition = (x1_pval <= pvalue_thresholds[i][0])\ - & (pvalue_thresholds[i + 1][0] < x1_pval) - x_annot[condition] = pvalue_thresholds[i][1] + for i, threshold in enumerate(pvalue_thresholds[:-1]): + condition = ((x1_pval <= threshold[0]) + & (pvalue_thresholds[i + 1][0] < x1_pval)) + x_annot[condition] = threshold[1] - else: - condition = x1_pval < pvalue_thresholds[i][0] - x_annot[condition] = pvalue_thresholds[i][1] + x_annot[x1_pval <= pvalue_thresholds[-1][0]] = pvalue_thresholds[-1][1] - x_annot = pd.Series([f"{star}{signif}" - for star, signif in zip(x_annot, x1_signif_suff)]) + x_annot = [f"{star}{res.significance_suffix}" + for star, res in zip(x_annot, result)] - return x_annot if not single_value else x_annot.iloc[0] + return x_annot if was_list else x_annot[0] def simple_text(result: StatResult, pvalue_format, pvalue_thresholds) -> str: diff --git a/statannotations/stats/ComparisonsCorrection.py b/statannotations/stats/ComparisonsCorrection.py index ab0b016..33688b9 100644 --- a/statannotations/stats/ComparisonsCorrection.py +++ b/statannotations/stats/ComparisonsCorrection.py @@ -197,3 +197,27 @@ def __call__(self, pvalues: Union[List[float], float], result = result[:n_vals] return result if got_pvalues_in_array else result[0] + + def _apply_type1_comparisons_correction(self, test_result_list): + original_pvalues = [result.pvalue for result in test_result_list] + + significant_pvalues = self(original_pvalues) + for is_significant, result in zip(significant_pvalues, + test_result_list): + result.correction_method = self.name + result.corrected_significance = is_significant + + def _apply_type0_comparisons_correction(self, test_result_list): + for result in test_result_list: + result.correction_method = self.name + result.corrected_significance = ( + result.pvalue < self.alpha + or np.isclose(result.pvalue, self.alpha)) + + def apply(self, test_result_list): + + if self.type == 1: + self._apply_type1_comparisons_correction(test_result_list) + + else: + self._apply_type0_comparisons_correction(test_result_list) diff --git a/statannotations/stats/StatTest.py b/statannotations/stats/StatTest.py index 0fddcc5..40e776b 100644 --- a/statannotations/stats/StatTest.py +++ b/statannotations/stats/StatTest.py @@ -6,20 +6,20 @@ # Also example for how to add other functions -def wilcoxon(box_data1, box_data2, verbose=1, stats_params): +def wilcoxon(group_data1, group_data2, verbose=1, stats_params): """ This function provides the equivalent behavior from earlier versions of statannot/statannotations. """ zero_method = stats_params.pop('zero_method', None) if zero_method is None: - return stats.wilcoxon(box_data1, box_data2, stats_params) + return stats.wilcoxon(group_data1, group_data2, stats_params) if zero_method == "AUTO": - zero_method = len(box_data1) <= 20 and "pratt" or "wilcox" + zero_method = len(group_data1) <= 20 and "pratt" or "wilcox" if verbose >= 1: print("Using zero_method ", zero_method) - return stats.wilcoxon(box_data1, box_data2, + return stats.wilcoxon(group_data1, group_data2, zero_method=zero_method, stats_params) @@ -71,10 +71,10 @@ def __init__(self, func: Callable, test_long_name: str, f"------ Original function documentation ------ \n" f"{func.__doc__}") - def __call__(self, box_data1, box_data2, alpha=0.05, + def __call__(self, group_data1, group_data2, alpha=0.05, stat_params): - stat, pval = self._func(box_data1, box_data2, self.args, + stat, pval = self._func(group_data1, group_data2, self.args, {self.kwargs, *stat_params})[:2] return StatResult(self._test_long_name, self._test_short_name, diff --git a/statannotations/utils.py b/statannotations/utils.py index 6aa6396..c7d44fc 100644 --- a/statannotations/utils.py +++ b/statannotations/utils.py @@ -1,4 +1,5 @@ import itertools +from bisect import bisect_left from typing import List, Union import pandas as pd @@ -57,40 +58,40 @@ def check_order_in_data(data, x, order) -> None: f"(specified in `order`)") -def _check_box_pairs_in_data_no_hue(box_pairs: Union[list, tuple], - data: Union[List[list], - pd.DataFrame] = None, - x: Union[str, list] = None) -> None: +def _check_pairs_in_data_no_hue(pairs: Union[list, tuple], + data: Union[List[list], + pd.DataFrame] = None, + x: Union[str, list] = None) -> None: x_values = get_x_values(data, x) - pairs_x_values = set(itertools.chain(box_pairs)) + pairs_x_values = set(itertools.chain(pairs)) unmatched_x_values = pairs_x_values - x_values if unmatched_x_values: raise ValueError(f"Missing x value(s) " f"`{render_collection(unmatched_x_values)}` in {x}" - f" (specified in `box_pairs`) in data") + f" (specified in `pairs`) in data") -def _check_box_pairs_in_data_with_hue(box_pairs: Union[list, tuple], - data: Union[List[list], - pd.DataFrame] = None, - x: Union[str, list] = None, - hue: str = None) -> set: +def _check_pairs_in_data_with_hue(pairs: Union[list, tuple], + data: Union[List[list], + pd.DataFrame] = None, + x: Union[str, list] = None, + hue: str = None) -> set: x_values = get_x_values(data, x) seen_x_values = set() hue_values = set(data[hue].unique()) - for x_value, hue_value in itertools.chain(itertools.chain(box_pairs)): + for x_value, hue_value in itertools.chain(itertools.chain(pairs)): if x_value not in seen_x_values and x_value not in x_values: raise ValueError(f"Missing x value `{x_value}` in {x}" - f" (specified in `box_pairs`)") + f" (specified in `pairs`)") seen_x_values.add(x_value) if hue_value not in hue_values: raise ValueError(f"Missing hue value `{hue_value}` in {hue}" - f" (specified in `box_pairs`)") + f" (specified in `pairs`)") return hue_values @@ -105,19 +106,19 @@ def _check_hue_order_in_data(hue, hue_values: set, f" in {hue} (specified in `hue_order`)") -def check_box_pairs_in_data(box_pairs: Union[list, tuple], - data: Union[List[list], pd.DataFrame] = None, - x: Union[str, list] = None, - hue: str = None, - hue_order: List[str] = None): +def check_pairs_in_data(pairs: Union[list, tuple], + data: Union[List[list], pd.DataFrame] = None, + x: Union[str, list] = None, + hue: str = None, + hue_order: List[str] = None): """ - Checks that values referred to in `order` and `box_pairs` exist in data. + Checks that values referred to in `order` and `pairs` exist in data. """ if hue is None and hue_order is None: - _check_box_pairs_in_data_no_hue(box_pairs, data, x) + _check_pairs_in_data_no_hue(pairs, data, x) else: - hue_values = _check_box_pairs_in_data_with_hue(box_pairs, data, x, hue) + hue_values = _check_pairs_in_data_with_hue(pairs, data, x, hue) _check_hue_order_in_data(hue, hue_values, hue_order) @@ -136,3 +137,20 @@ def check_valid_text_format(text_format): def render_collection(collection): return '","'.join(map(str, collection)) + + +def get_closest(a_list, value): + """ + Assumes myList is sorted. Returns closest value to myNumber. + If two numbers are equally close, return the smallest number. + from https://stackoverflow.com/a/12141511/9981846 + """ + pos = bisect_left(a_list, value) + if pos == 0: + return a_list[0] + if pos == len(a_list): + return a_list[-1] + before, after = a_list[pos - 1: pos + 1] + if after - value < value - before: + return after + return before diff --git a/usage/example.ipynb b/usage/example.ipynb index 7d17821..10b91f9 100644 --- a/usage/example.ipynb +++ b/usage/example.ipynb @@ -62,9 +62,9 @@ " : 1.00e-04 < p <= 1.00e-03\n", " : p <= 1.00e-04\n", "\n", - "Thur v.s. Fri: Mann-Whitney-Wilcoxon test two-sided, P_val:6.477e-01 U_stat=6.305e+02\n", - "Thur v.s. Sat: Mann-Whitney-Wilcoxon test two-sided, P_val:4.690e-02 U_stat=2.180e+03\n", - "Sun v.s. Fri: Mann-Whitney-Wilcoxon test two-sided, P_val:2.680e-02 U_stat=9.605e+02\n" + "Thur vs. Fri: Mann-Whitney-Wilcoxon test two-sided, P_val:6.477e-01 U_stat=6.305e+02\n", + "Thur vs. Sat: Mann-Whitney-Wilcoxon test two-sided, P_val:4.690e-02 U_stat=2.180e+03\n", + "Sun vs. Fri: Mann-Whitney-Wilcoxon test two-sided, P_val:2.680e-02 U_stat=9.605e+02\n" ] }, { @@ -99,7 +99,9 @@ "cell_type": "markdown", "metadata": {}, "source": [ - "The `annotate` method returns the tuple `ax, test_results`, where `test_results` is a list of `StatResult` objects, containing both the original data of the boxes and the statistical test results (p-value, etc)." + "The `annotate` method returns the tuple `ax, annotations`, where `annotations` is a list of `Annotation` objects.\n", + "The `data` attribute of an Annotation is a `StatResult` object, which containing both the original data of the groups\n", + "(boxes for boxplots), and the statistical test results (p-value, etc)." ] }, { @@ -115,15 +117,15 @@ "Mann-Whitney-Wilcoxon test two-sided, P_val:4.690e-02 U_stat=2.180e+03\n", "Mann-Whitney-Wilcoxon test two-sided, P_val:2.680e-02 U_stat=9.605e+02\n", "\n", - "StatResult attributes: dict_keys(['test_description', 'test_short_name', 'stat_str', 'stat_value', 'pvalue', '_corrected_significance', '_correction_method', 'alpha', 'box1', 'box2'])\n" + "StatResult attributes: dict_keys(['test_description', 'test_short_name', 'stat_str', 'stat_value', 'pvalue', '_corrected_significance', '_correction_method', 'alpha', 'group1', 'group2'])\n" ] } ], "source": [ "for res in test_results:\n", - " print(res)\n", + " print(res.data)\n", "\n", - "print(\"\\nStatResult attributes:\", test_results[0].__dict__.keys())" + "print(\"\\nStatResult attributes:\", test_results[0].data.__dict__.keys())" ] }, { @@ -132,7 +134,7 @@ "source": [ "#### Applying a multiple comparisons correction (requires `statsmodels`)\n", "In this example, note that previously configured parameters are remembered if not changed.\n", - "This is the case for `box_pairs`, `test`, `text_format`, and `loc`.\n", + "This is the case for `pairs`, `test`, `text_format`, and `loc`.\n", "\n", "To avoid this, you can call `annot.reset_configuration()` to get back to all default/unset values." ] @@ -153,9 +155,9 @@ " : 1.00e-04 < p <= 1.00e-03\n", " *: p <= 1.00e-04\n", "\n", - "Thur v.s. Fri: Mann-Whitney-Wilcoxon test two-sided with Bonferroni correction, P_val:1.000e+00 U_stat=6.305e+02\n", - "Thur v.s. Sat: Mann-Whitney-Wilcoxon test two-sided with Bonferroni correction, P_val:1.407e-01 U_stat=2.180e+03\n", - "Sun v.s. Fri: Mann-Whitney-Wilcoxon test two-sided with Bonferroni correction, P_val:8.041e-02 U_stat=9.605e+02\n" + "Thur vs. Fri: Mann-Whitney-Wilcoxon test two-sided with Bonferroni correction, P_val:1.000e+00 U_stat=6.305e+02\n", + "Thur vs. Sat: Mann-Whitney-Wilcoxon test two-sided with Bonferroni correction, P_val:1.407e-01 U_stat=2.180e+03\n", + "Sun vs. Fri: Mann-Whitney-Wilcoxon test two-sided with Bonferroni correction, P_val:8.041e-02 U_stat=9.605e+02\n" ] }, { @@ -222,14 +224,14 @@ " : 1.00e-04 < p <= 1.00e-03\n", " *: p <= 1.00e-04\n", "\n", - "Thur v.s. Fri: Mann-Whitney-Wilcoxon test two-sided with Benjamini-Hochberg correction, P_val:6.477e-01 U_stat=6.305e+02\n", - "Thur v.s. Sat: Mann-Whitney-Wilcoxon test two-sided with Benjamini-Hochberg correction, P_val:4.690e-02 (ns) U_stat=2.180e+03\n", - "Sun v.s. Fri: Mann-Whitney-Wilcoxon test two-sided with Benjamini-Hochberg correction, P_val:2.680e-02 (ns) U_stat=9.605e+02\n" + "Thur vs. Fri: Mann-Whitney-Wilcoxon test two-sided with Benjamini-Hochberg correction, P_val:6.477e-01 U_stat=6.305e+02\n", + "Thur vs. Sat: Mann-Whitney-Wilcoxon test two-sided with Benjamini-Hochberg correction, P_val:4.690e-02 (ns) U_stat=2.180e+03\n", + "Sun vs. Fri: Mann-Whitney-Wilcoxon test two-sided with Benjamini-Hochberg correction, P_val:2.680e-02 (ns) U_stat=9.605e+02\n" ] }, { "data": { - "text/plain": "(<matplotlib.axes._subplots.AxesSubplot at 0x7fb3e83f8128>,\n [<statannotations.stats.StatResult.StatResult at 0x7fb3e83ef908>,\n <statannotations.stats.StatResult.StatResult at 0x7fb3e6e8cf60>,\n <statannotations.stats.StatResult.StatResult at 0x7fb3e83efcc0>])" + "text/plain": "(<matplotlib.axes._subplots.AxesSubplot at 0x7fedb54c7748>,\n [<statannotations.Annotation.Annotation at 0x7fedb4233ac8>,\n <statannotations.Annotation.Annotation at 0x7fedb4233d30>,\n <statannotations.Annotation.Annotation at 0x7fedb4233320>])" }, "execution_count": 6, "metadata": {}, @@ -275,9 +277,9 @@ " : 1.00e-04 < p <= 1.00e-03\n", " *: p <= 1.00e-04\n", "\n", - "Sun v.s. Thur: Custom statistical test with Benjamini-Hochberg correction, P_val:1.000e-01\n", - "Sun v.s. Fri: Custom statistical test with Benjamini-Hochberg correction, P_val:1.000e-03\n", - "Sun v.s. Sat: Custom statistical test with Benjamini-Hochberg correction, P_val:1.000e-01\n" + "Sun vs. Thur: Custom statistical test with Benjamini-Hochberg correction, P_val:1.000e-01\n", + "Sun vs. Fri: Custom statistical test with Benjamini-Hochberg correction, P_val:1.000e-03\n", + "Sun vs. Sat: Custom statistical test with Benjamini-Hochberg correction, P_val:1.000e-01\n" ] }, { @@ -293,9 +295,9 @@ "x = \"day\"\n", "y = \"total_bill\"\n", "order = ['Sun', 'Thur', 'Fri', 'Sat']\n", - "box_pairs=[(\"Sun\", \"Thur\"), (\"Sun\", \"Sat\"), (\"Fri\", \"Sun\")]\n", + "pairs = [(\"Sun\", \"Thur\"), (\"Sun\", \"Sat\"), (\"Fri\", \"Sun\")]\n", "ax = sns.boxplot(data=df, x=x, y=y, order=order)\n", - "annot.new_plot(ax, box_pairs=box_pairs, data=df, x=x, y=y, order=order)\n", + "annot.new_plot(ax, pairs=pairs, data=df, x=x, y=y, order=order)\n", "annot.configure(test=None, loc='inside')\n", "annot.set_pvalues([0.1, 0.1, 0.001])\n", "annot.annotate()\n", @@ -357,15 +359,15 @@ " : 1.00e-04 < p <= 1.00e-03\n", " *: p <= 1.00e-04\n", "\n", - "E_Ideal v.s. E_Premium: Mann-Whitney-Wilcoxon test two-sided with Benjamini-Hochberg correction, P_val:1.560e-31 U_stat=3.756e+06\n", - "I_Ideal v.s. I_Premium: Mann-Whitney-Wilcoxon test two-sided with Benjamini-Hochberg correction, P_val:5.141e-61 U_stat=1.009e+06\n", - "J_Ideal v.s. J_Premium: Mann-Whitney-Wilcoxon test two-sided with Benjamini-Hochberg correction, P_val:4.018e-37 U_stat=2.337e+05\n", - "E_Ideal v.s. E_Good: Mann-Whitney-Wilcoxon test two-sided with Benjamini-Hochberg correction, P_val:5.201e-19 U_stat=1.480e+06\n", - "I_Ideal v.s. I_Good: Mann-Whitney-Wilcoxon test two-sided with Benjamini-Hochberg correction, P_val:5.008e-13 U_stat=4.359e+05\n", - "J_Ideal v.s. J_Good: Mann-Whitney-Wilcoxon test two-sided with Benjamini-Hochberg correction, P_val:1.006e-04 U_stat=1.174e+05\n", - "E_Ideal v.s. E_Very Good: Mann-Whitney-Wilcoxon test two-sided with Benjamini-Hochberg correction, P_val:1.736e-02 U_stat=4.850e+06\n", - "E_Good v.s. I_Ideal: Mann-Whitney-Wilcoxon test two-sided with Benjamini-Hochberg correction, P_val:5.906e-01 U_stat=9.882e+05\n", - "I_Premium v.s. J_Ideal: Mann-Whitney-Wilcoxon test two-sided with Benjamini-Hochberg correction, P_val:5.159e-27 U_stat=8.084e+05\n" + "E_Ideal vs. E_Premium: Mann-Whitney-Wilcoxon test two-sided with Benjamini-Hochberg correction, P_val:1.560e-31 U_stat=3.756e+06\n", + "I_Ideal vs. I_Premium: Mann-Whitney-Wilcoxon test two-sided with Benjamini-Hochberg correction, P_val:5.141e-61 U_stat=1.009e+06\n", + "J_Ideal vs. J_Premium: Mann-Whitney-Wilcoxon test two-sided with Benjamini-Hochberg correction, P_val:4.018e-37 U_stat=2.337e+05\n", + "E_Ideal vs. E_Good: Mann-Whitney-Wilcoxon test two-sided with Benjamini-Hochberg correction, P_val:5.201e-19 U_stat=1.480e+06\n", + "I_Ideal vs. I_Good: Mann-Whitney-Wilcoxon test two-sided with Benjamini-Hochberg correction, P_val:5.008e-13 U_stat=4.359e+05\n", + "J_Ideal vs. J_Good: Mann-Whitney-Wilcoxon test two-sided with Benjamini-Hochberg correction, P_val:1.006e-04 U_stat=1.174e+05\n", + "E_Ideal vs. E_Very Good: Mann-Whitney-Wilcoxon test two-sided with Benjamini-Hochberg correction, P_val:1.736e-02 U_stat=4.850e+06\n", + "E_Good vs. I_Ideal: Mann-Whitney-Wilcoxon test two-sided with Benjamini-Hochberg correction, P_val:5.906e-01 U_stat=9.882e+05\n", + "I_Premium vs. J_Ideal: Mann-Whitney-Wilcoxon test two-sided with Benjamini-Hochberg correction, P_val:5.159e-27 U_stat=8.084e+05\n" ] }, { @@ -383,7 +385,7 @@ "hue = \"cut\"\n", "hue_order=['Ideal', 'Premium', 'Good', 'Very Good', 'Fair']\n", "order = [\"E\", \"I\", \"J\"]\n", - "box_pairs=[\n", + "pairs=[\n", " ((\"E\", \"Ideal\"), (\"E\", \"Very Good\")),\n", " ((\"E\", \"Ideal\"), (\"E\", \"Premium\")),\n", " ((\"E\", \"Ideal\"), (\"E\", \"Good\")),\n", @@ -395,7 +397,7 @@ " ((\"I\", \"Premium\"), (\"J\", \"Ideal\")),\n", " ]\n", "ax = sns.boxplot(data=df, x=x, y=y, order=order, hue=hue, hue_order=hue_order)\n", - "annot.new_plot(ax, box_pairs, data=df, x=x, y=y, order=order, hue=hue, hue_order=hue_order)\n", + "annot.new_plot(ax, pairs, data=df, x=x, y=y, order=order, hue=hue, hue_order=hue_order)\n", "annot.configure(test='Mann-Whitney', verbose=2)\n", "annot.apply_test()\n", "annot.annotate()\n", @@ -477,7 +479,7 @@ " : 1.00e-04 < p <= 1.00e-03\n", " *: p <= 1.00e-04\n", "\n", - "Fri_(0.991, 4.0] v.s. Sat_(7.0, 10.0]: t-test independent samples, P_val:6.176e-07 t=-7.490e+00\n" + "Fri_(0.991, 4.0] vs. Sat_(7.0, 10.0]: t-test independent samples, P_val:6.176e-07 t=-7.490e+00\n" ] }, { @@ -538,15 +540,15 @@ " : 1.00e-04 < p <= 1.00e-03\n", " *: p <= 1.00e-04\n", "\n", - "SI2_I v.s. SI2_J: t-test independent samples, P_val:3.710e-01 t=-8.949e-01\n", - "SI1_I v.s. SI1_J: t-test independent samples, P_val:8.513e-05 t=-3.937e+00\n", - "VS1_I v.s. VS1_J: t-test independent samples, P_val:9.140e-05 t=-3.923e+00\n", - "VS2_I v.s. VS2_J: t-test independent samples, P_val:6.806e-03 t=-2.709e+00\n", - "VVS2_I v.s. VVS2_J: t-test independent samples, P_val:9.076e-14 t=-7.672e+00\n", - "VVS1_I v.s. VVS1_J: t-test independent samples, P_val:6.536e-10 t=-6.321e+00\n", - "I1_I v.s. I1_J: t-test independent samples, P_val:2.905e-02 t=-2.206e+00\n", - "IF_I v.s. IF_J: t-test independent samples, P_val:4.412e-03 t=-2.881e+00\n", - "SI1_E v.s. SI2_E: t-test independent samples, P_val:4.178e-66 t=-1.749e+01\n" + "SI2_I vs. SI2_J: t-test independent samples, P_val:3.710e-01 t=-8.949e-01\n", + "SI1_I vs. SI1_J: t-test independent samples, P_val:8.513e-05 t=-3.937e+00\n", + "VS1_I vs. VS1_J: t-test independent samples, P_val:9.140e-05 t=-3.923e+00\n", + "VS2_I vs. VS2_J: t-test independent samples, P_val:6.806e-03 t=-2.709e+00\n", + "VVS2_I vs. VVS2_J: t-test independent samples, P_val:9.076e-14 t=-7.672e+00\n", + "VVS1_I vs. VVS1_J: t-test independent samples, P_val:6.536e-10 t=-6.321e+00\n", + "I1_I vs. I1_J: t-test independent samples, P_val:2.905e-02 t=-2.206e+00\n", + "IF_I vs. IF_J: t-test independent samples, P_val:4.412e-03 t=-2.881e+00\n", + "SI1_E vs. SI2_E: t-test independent samples, P_val:4.178e-66 t=-1.749e+01\n" ] }, { @@ -566,14 +568,14 @@ "y = \"carat\"\n", "hue = \"color\"\n", "hue_order=[\"E\", \"I\", \"J\"]\n", - "box_pairs = [\n", + "pairs = [\n", " ((\"SI2\", \"E\"), (\"SI1\", \"E\"))\n", " ]\n", - "box_pairs = box_pairs + [((clar, 'I'), (clar, 'J'))\n", + "pairs = pairs + [((clar, 'I'), (clar, 'J'))\n", " for clar in df['clarity'].unique()]\n", "width = 0.4\n", "ax = sns.barplot(data=df, x=x, y=y, hue=hue, hue_order=hue_order, seed=2021)\n", - "annot.new_plot(ax, box_pairs, plot='barplot',\n", + "annot.new_plot(ax, pairs, plot='barplot',\n", " data=df, x=x, y=y, hue=hue, hue_order=hue_order, seed=2021)\n", "annot.apply_test().annotate()\n", "\n", @@ -604,9 +606,9 @@ "name": "stdout", "output_type": "stream", "text": [ - "Sat_Yes v.s. Sat_No: t-test independent samples, P_val:4.304e-01 t=7.922e-01\n", - "Thur_No v.s. Fri_No: t-test independent samples, P_val:7.425e-01 t=-3.305e-01\n", - "Thur_Yes v.s. Sun_No: t-test independent samples, P_val:5.623e-01 t=-5.822e-01\n" + "Sat_Yes vs. Sat_No: t-test independent samples, P_val:4.304e-01 t=7.922e-01\n", + "Thur_No vs. Fri_No: t-test independent samples, P_val:7.425e-01 t=-3.305e-01\n", + "Thur_Yes vs. Sun_No: t-test independent samples, P_val:5.623e-01 t=-5.822e-01\n" ] }, { @@ -623,16 +625,16 @@ "x = \"day\"\n", "y = \"total_bill\"\n", "hue = \"smoker\"\n", - "box_pairs = [((\"Thur\", \"No\"), (\"Fri\", \"No\")),\n", - " ((\"Sat\", \"Yes\"), (\"Sat\", \"No\")),\n", - " ((\"Sun\", \"No\"), (\"Thur\", \"Yes\"))]\n", + "pairs = [((\"Thur\", \"No\"), (\"Fri\", \"No\")),\n", + " ((\"Sat\", \"Yes\"), (\"Sat\", \"No\")),\n", + " ((\"Sun\", \"No\"), (\"Thur\", \"Yes\"))]\n", "ax = sns.boxplot(data=df, x=x, y=y, hue=hue)\n", - "annot.new_plot(ax, box_pairs,\n", + "annot.new_plot(ax, pairs,\n", " data=df, x=x, y=y, hue=hue)\n", "annot.configure(test='t-test_ind', text_format='full', loc='inside',\n", " comparisons_correction=None, line_height=0.05, text_offset=8)\n", "\n", - "annot.apply_test().annotate(line_offset_to_box=0.2, line_offset=0.1)\n", + "annot.apply_test().annotate(line_offset_to_group=0.2, line_offset=0.1)\n", "\n", "plt.legend(loc='upper left', bbox_to_anchor=(1.03, 1))\n", "plt.savefig('example_tuning_y_offsets.svg')\n", @@ -663,14 +665,14 @@ " : 1.00e-04 < p <= 1.00e-03\n", " *: p <= 1.00e-04\n", "\n", - "setosa v.s. versicolor: Bartlett statistical test for equal variances, P_val:8.660e-03 Stat=6.892e+00\n", - "versicolor v.s. virginica: Bartlett statistical test for equal variances, P_val:1.478e-01 Stat=2.095e+00\n", - "setosa v.s. virginica: Bartlett statistical test for equal variances, P_val:6.379e-05 Stat=1.599e+01\n" + "setosa vs. versicolor: Bartlett statistical test for equal variances, P_val:8.660e-03 Stat=6.892e+00\n", + "versicolor vs. virginica: Bartlett statistical test for equal variances, P_val:1.478e-01 Stat=2.095e+00\n", + "setosa vs. virginica: Bartlett statistical test for equal variances, P_val:6.379e-05 Stat=1.599e+01\n" ] }, { "data": { - "text/plain": "(<matplotlib.axes._subplots.AxesSubplot at 0x7fb3e9c366a0>,\n [<statannotations.stats.StatResult.StatResult at 0x7fb3e9b0fcf8>,\n <statannotations.stats.StatResult.StatResult at 0x7fb3e74f5b38>,\n <statannotations.stats.StatResult.StatResult at 0x7fb3e9c2b5c0>])" + "text/plain": "(<matplotlib.axes._subplots.AxesSubplot at 0x7fedb4460898>,\n [<statannotations.Annotation.Annotation at 0x7fedb40b4400>,\n <statannotations.Annotation.Annotation at 0x7fedb715ccf8>,\n <statannotations.Annotation.Annotation at 0x7fedb6c8ef28>])" }, "execution_count": 15, "metadata": {}, @@ -695,7 +697,7 @@ "x = \"species\"\n", "y = \"sepal_length\"\n", "\n", - "box_pairs = [(\"setosa\", \"versicolor\"), (\"setosa\", \"virginica\"), (\"versicolor\", \"virginica\")]\n", + "pairs = [(\"setosa\", \"versicolor\"), (\"setosa\", \"virginica\"), (\"versicolor\", \"virginica\")]\n", "\n", "# Required descriptors for annotate\n", "custom_long_name = 'Bartlett statistical test for equal variances'\n", @@ -706,7 +708,7 @@ "# Then, same as usual\n", "ax = sns.boxplot(data=df, x=x, y=y)\n", "annot.reset_configuration()\n", - "annot.new_plot(ax, box_pairs, data=df, x=x, y=y)\n", + "annot.new_plot(ax, pairs, data=df, x=x, y=y)\n", "annot.configure(test=custom_test, comparisons_correction=None,\n", " text_format='star').apply_test().annotate()" ], @@ -748,7 +750,7 @@ "\n", "test_short_name = 'Bartlett'\n", "pvalues = []\n", - "for pair in box_pairs:\n", + "for pair in pairs:\n", " data1 = df.groupby(x)[y].get_group(pair[0])\n", " data2 = df.groupby(x)[y].get_group(pair[1])\n", " stat, p = bartlett(data1, data2)\n", @@ -776,9 +778,9 @@ " : 1.00e-04 < p <= 1.00e-03\n", " *: p <= 1.00e-04\n", "\n", - "setosa v.s. versicolor: Custom statistical test, P_val:8.660e-03\n", - "versicolor v.s. virginica: Custom statistical test, P_val:1.478e-01\n", - "setosa v.s. virginica: Custom statistical test, P_val:6.379e-05\n" + "setosa vs. versicolor: Custom statistical test, P_val:8.660e-03\n", + "versicolor vs. virginica: Custom statistical test, P_val:1.478e-01\n", + "setosa vs. virginica: Custom statistical test, P_val:6.379e-05\n" ] }, { @@ -792,7 +794,7 @@ ], "source": [ "ax = sns.boxplot(data=df, x=x, y=y)\n", - "annot.new_plot(ax=ax, box_pairs=box_pairs,\n", + "annot.new_plot(ax=ax, pairs=pairs,\n", " data=df, x=x, y=y)\n", "(annot\n", " .configure(test=None, test_short_name=test_short_name)\n", @@ -825,9 +827,9 @@ " : 1.00e-04 < p <= 1.00e-03\n", " **: p <= 1.00e-04\n", "\n", - "setosa v.s. versicolor: Custom statistical test, P_val:8.660e-03\n", - "versicolor v.s. virginica: Custom statistical test, P_val:1.478e-01\n", - "setosa v.s. virginica: Custom statistical test, P_val:6.379e-05\n" + "setosa vs. versicolor: Custom statistical test, P_val:8.660e-03\n", + "versicolor vs. virginica: Custom statistical test, P_val:1.478e-01\n", + "setosa vs. virginica: Custom statistical test, P_val:6.379e-05\n" ] }, { @@ -841,7 +843,7 @@ ], "source": [ "ax = sns.violinplot(data=df, x=x, y=y)\n", - "annot.new_plot(ax=ax, box_pairs=box_pairs, plot=\"violinplot\",\n", + "annot.new_plot(ax=ax, pairs=pairs, plot=\"violinplot\",\n", " data=df, x=x, y=y)\n", "(annot\n", " .configure(test=None, test_short_name=test_short_name)\n", @@ -873,6 +875,22 @@ } }, "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "p-value annotation legend:\n", + " ns: p <= 1.00e+00\n", + " : 1.00e-02 < p <= 5.00e-02\n", + " : 1.00e-03 < p <= 1.00e-02\n", + " : 1.00e-04 < p <= 1.00e-03\n", + " *: p <= 1.00e-04\n", + "\n", + "Thur vs. Fri: first pair\n", + "Thur vs. Sat: second pair\n", + "Sun vs. Fri: third pair\n" + ] + }, { "data": { "text/plain": "<Figure size 432x288 with 1 Axes>", @@ -888,7 +906,7 @@ "y = \"total_bill\"\n", "order = ['Sun', 'Thur', 'Fri', 'Sat']\n", "ax = sns.boxplot(data=df, x=x, y=y, order=order)\n", - "annot.new_plot(ax, box_pairs=[(\"Thur\", \"Fri\"),(\"Thur\", \"Sat\"), (\"Fri\", \"Sun\")],\n", + "annot.new_plot(ax, pairs=[(\"Thur\", \"Fri\"),(\"Thur\", \"Sat\"), (\"Fri\", \"Sun\")],\n", " data=df, x=x, y=y, order=order)\n", "annot.configure(loc='outside')\n", "annot.set_custom_annotations([\"first pair\", \"second pair\", \"third pair\"])\n", @@ -921,15 +939,15 @@ " : 1.00e-04 < p <= 1.00e-03\n", " *: p <= 1.00e-04\n", "\n", - "H1N1_Nonsynonymous v.s. H1N1_Synonymous: Mann-Whitney-Wilcoxon test two-sided, P_val:5.014e-04 U_stat=2.624e+03\n", - "H3N2_Nonsynonymous v.s. H3N2_Synonymous: Mann-Whitney-Wilcoxon test two-sided, P_val:1.294e-03 U_stat=1.535e+04\n", - "Influenza B_Nonsynonymous v.s. Influenza B_Synonymous: Mann-Whitney-Wilcoxon test two-sided, P_val:2.026e-01 U_stat=3.340e+02\n" + "H1N1_Nonsynonymous vs. H1N1_Synonymous: Mann-Whitney-Wilcoxon test two-sided, P_val:5.014e-04 U_stat=2.624e+03\n", + "H3N2_Nonsynonymous vs. H3N2_Synonymous: Mann-Whitney-Wilcoxon test two-sided, P_val:1.294e-03 U_stat=1.535e+04\n", + "Influenza B_Nonsynonymous vs. Influenza B_Synonymous: Mann-Whitney-Wilcoxon test two-sided, P_val:2.026e-01 U_stat=3.340e+02\n" ] }, { "data": { "text/plain": "<Figure size 864x432 with 1 Axes>", - "image/png": "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\n" + "image/png": "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\n" }, "metadata": {}, "output_type": "display_data" @@ -939,7 +957,13 @@ "samples = pd.read_csv('samples.tsv', sep = '\\t')\n", "samples = samples.loc[samples.mutation_rate_samp < 1]\n", "samples = samples.drop(columns=['mutation_rate_samp'])\n", - "samples = samples.melt(id_vars=['sampleID', 'subtype'], value_vars=['nonsynon_mutation_rate_samp','synon_mutation_rate_samp'], var_name='Synon_Nonsynon', value_name='mutation_rate_samp').replace('nonsynon_mutation_rate_samp', 'Nonsynonymous').replace('synon_mutation_rate_samp', 'Synonymous')\n", + "samples = (\n", + " samples.melt(id_vars=['sampleID', 'subtype'],\n", + " value_vars=['nonsynon_mutation_rate_samp','synon_mutation_rate_samp'],\n", + " var_name='Synon_Nonsynon', value_name='mutation_rate_samp')\n", + " .replace('nonsynon_mutation_rate_samp', 'Nonsynonymous')\n", + " .replace('synon_mutation_rate_samp', 'Synonymous'))\n", + "\n", "samples = samples.loc[samples.mutation_rate_samp > 0]\n", "\n", "significanceComparisons = [(('H3N2','Synonymous'), ('H3N2','Nonsynonymous')),\n", @@ -962,7 +986,7 @@ "ax.set_yscale('log')\n", "\n", "sns.stripplot(ax=ax, fig_args)\n", - "annotator = Annotator(ax=ax, box_pairs=significanceComparisons,\n", + "annotator = Annotator(ax=ax, pairs=significanceComparisons,\n", " fig_args, plot='stripplot')\n", "annotator.configure(configuration).apply_test().annotate()\n", "fig.savefig(f'flu_dataset_log_scale_in_axes_strip.svg', format='svg')" @@ -991,9 +1015,9 @@ "name": "stdout", "output_type": "stream", "text": [ - "H1N1_Nonsynonymous v.s. H1N1_Synonymous: Mann-Whitney-Wilcoxon test two-sided, P_val:5.014e-04 U_stat=2.624e+03\n", - "H3N2_Nonsynonymous v.s. H3N2_Synonymous: Mann-Whitney-Wilcoxon test two-sided, P_val:1.294e-03 U_stat=1.535e+04\n", - "Influenza B_Nonsynonymous v.s. Influenza B_Synonymous: Mann-Whitney-Wilcoxon test two-sided, P_val:2.026e-01 U_stat=3.340e+02\n" + "H1N1_Nonsynonymous vs. H1N1_Synonymous: Mann-Whitney-Wilcoxon test two-sided, P_val:5.014e-04 U_stat=2.624e+03\n", + "H3N2_Nonsynonymous vs. H3N2_Synonymous: Mann-Whitney-Wilcoxon test two-sided, P_val:1.294e-03 U_stat=1.535e+04\n", + "Influenza B_Nonsynonymous vs. Influenza B_Synonymous: Mann-Whitney-Wilcoxon test two-sided, P_val:2.026e-01 U_stat=3.340e+02\n" ] }, { @@ -1042,9 +1066,9 @@ "name": "stdout", "output_type": "stream", "text": [ - "H1N1_Nonsynonymous v.s. H1N1_Synonymous: Mann-Whitney-Wilcoxon test two-sided, P_val:5.014e-04 U_stat=2.624e+03\n", - "H3N2_Nonsynonymous v.s. H3N2_Synonymous: Mann-Whitney-Wilcoxon test two-sided, P_val:1.294e-03 U_stat=1.535e+04\n", - "Influenza B_Nonsynonymous v.s. Influenza B_Synonymous: Mann-Whitney-Wilcoxon test two-sided, P_val:2.026e-01 U_stat=3.340e+02\n" + "H1N1_Nonsynonymous vs. H1N1_Synonymous: Mann-Whitney-Wilcoxon test two-sided, P_val:5.014e-04 U_stat=2.624e+03\n", + "H3N2_Nonsynonymous vs. H3N2_Synonymous: Mann-Whitney-Wilcoxon test two-sided, P_val:1.294e-03 U_stat=1.535e+04\n", + "Influenza B_Nonsynonymous vs. Influenza B_Synonymous: Mann-Whitney-Wilcoxon test two-sided, P_val:2.026e-01 U_stat=3.340e+02\n" ] }, { diff --git a/usage/example_tuning_y_offsets.svg b/usage/example_tuning_y_offsets.svg index 56e1850..42a7655 100644 --- a/usage/example_tuning_y_offsets.svg +++ b/usage/example_tuning_y_offsets.svg @@ -347,7 +347,7 @@ z <g id="matplotlib.axis_2"> <g id="ytick_1"> <g id="line2d_1"> - <path clip-path="url(#p51652a52ce)" d="M 54 227.515867 + <path clip-path="url(#p5b84badbdb)" d="M 54 227.515867 L 388.8 227.515867 " style="fill:none;stroke:#cccccc;stroke-linecap:round;"/> </g> @@ -399,7 +399,7 @@ z </g> <g id="ytick_2"> <g id="line2d_2"> - <path clip-path="url(#p51652a52ce)" d="M 54 201.23688 + <path clip-path="url(#p5b84badbdb)" d="M 54 201.23688 L 388.8 201.23688 " style="fill:none;stroke:#cccccc;stroke-linecap:round;"/> </g> @@ -439,7 +439,7 @@ z </g> <g id="ytick_3"> <g id="line2d_3"> - <path clip-path="url(#p51652a52ce)" d="M 54 174.957893 + <path clip-path="url(#p5b84badbdb)" d="M 54 174.957893 L 388.8 174.957893 " style="fill:none;stroke:#cccccc;stroke-linecap:round;"/> </g> @@ -488,7 +488,7 @@ z </g> <g id="ytick_4"> <g id="line2d_4"> - <path clip-path="url(#p51652a52ce)" d="M 54 148.678905 + <path clip-path="url(#p5b84badbdb)" d="M 54 148.678905 L 388.8 148.678905 " style="fill:none;stroke:#cccccc;stroke-linecap:round;"/> </g> @@ -521,7 +521,7 @@ z </g> <g id="ytick_5"> <g id="line2d_5"> - <path clip-path="url(#p51652a52ce)" d="M 54 122.399918 + <path clip-path="url(#p5b84badbdb)" d="M 54 122.399918 L 388.8 122.399918 " style="fill:none;stroke:#cccccc;stroke-linecap:round;"/> </g> @@ -562,7 +562,7 @@ z </g> <g id="ytick_6"> <g id="line2d_6"> - <path clip-path="url(#p51652a52ce)" d="M 54 96.12093 + <path clip-path="url(#p5b84badbdb)" d="M 54 96.12093 L 388.8 96.12093 " style="fill:none;stroke:#cccccc;stroke-linecap:round;"/> </g> @@ -611,7 +611,7 @@ z </g> <g id="ytick_7"> <g id="line2d_7"> - <path clip-path="url(#p51652a52ce)" d="M 54 69.841943 + <path clip-path="url(#p5b84badbdb)" d="M 54 69.841943 L 388.8 69.841943 " style="fill:none;stroke:#cccccc;stroke-linecap:round;"/> </g> @@ -640,7 +640,7 @@ z </g> <g id="ytick_8"> <g id="line2d_8"> - <path clip-path="url(#p51652a52ce)" d="M 54 43.562956 + <path clip-path="url(#p5b84badbdb)" d="M 54 43.562956 L 388.8 43.562956 " style="fill:none;stroke:#cccccc;stroke-linecap:round;"/> </g> @@ -772,7 +772,7 @@ z </g> </g> <g id="patch_3"> - <path clip-path="url(#p51652a52ce)" d="M 77.131636 218.291943 + <path clip-path="url(#p5b84badbdb)" d="M 77.131636 218.291943 L 106.959273 218.291943 L 106.959273 201.736181 L 77.131636 201.736181 @@ -781,7 +781,7 @@ z " style="fill:#5875a4;stroke:#4c4c4c;stroke-linejoin:miter;stroke-width:1.5;"/> </g> <g id="patch_4"> - <path clip-path="url(#p51652a52ce)" d="M 107.568 223.074719 + <path clip-path="url(#p5b84badbdb)" d="M 107.568 223.074719 L 137.395636 223.074719 L 137.395636 200.527347 L 107.568 200.527347 @@ -790,7 +790,7 @@ z " style="fill:#cc8963;stroke:#4c4c4c;stroke-linejoin:miter;stroke-width:1.5;"/> </g> <g id="patch_5"> - <path clip-path="url(#p51652a52ce)" d="M 153.222545 223.074719 + <path clip-path="url(#p5b84badbdb)" d="M 153.222545 223.074719 L 183.050182 223.074719 L 183.050182 204.745125 L 153.222545 204.745125 @@ -799,7 +799,7 @@ z " style="fill:#5875a4;stroke:#4c4c4c;stroke-linejoin:miter;stroke-width:1.5;"/> </g> <g id="patch_6"> - <path clip-path="url(#p51652a52ce)" d="M 183.658909 214.113584 + <path clip-path="url(#p5b84badbdb)" d="M 183.658909 214.113584 L 213.486545 214.113584 L 213.486545 194.522599 L 183.658909 194.522599 @@ -808,7 +808,7 @@ z " style="fill:#cc8963;stroke:#4c4c4c;stroke-linejoin:miter;stroke-width:1.5;"/> </g> <g id="patch_7"> - <path clip-path="url(#p51652a52ce)" d="M 229.313455 218.567872 + <path clip-path="url(#p5b84badbdb)" d="M 229.313455 218.567872 L 259.141091 218.567872 L 259.141091 183.386878 L 229.313455 183.386878 @@ -817,7 +817,7 @@ z " style="fill:#5875a4;stroke:#4c4c4c;stroke-linejoin:miter;stroke-width:1.5;"/> </g> <g id="patch_8"> - <path clip-path="url(#p51652a52ce)" d="M 259.749818 215.085906 + <path clip-path="url(#p5b84badbdb)" d="M 259.749818 215.085906 L 289.577455 215.085906 L 289.577455 199.528746 L 259.749818 199.528746 @@ -826,7 +826,7 @@ z " style="fill:#cc8963;stroke:#4c4c4c;stroke-linejoin:miter;stroke-width:1.5;"/> </g> <g id="patch_9"> - <path clip-path="url(#p51652a52ce)" d="M 305.404364 208.686973 + <path clip-path="url(#p5b84badbdb)" d="M 305.404364 208.686973 L 335.232 208.686973 L 335.232 168.716633 L 305.404364 168.716633 @@ -835,7 +835,7 @@ z " style="fill:#5875a4;stroke:#4c4c4c;stroke-linejoin:miter;stroke-width:1.5;"/> </g> <g id="patch_10"> - <path clip-path="url(#p51652a52ce)" d="M 335.840727 214.954511 + <path clip-path="url(#p5b84badbdb)" d="M 335.840727 214.954511 L 365.668364 214.954511 L 365.668364 188.097386 L 335.840727 188.097386 @@ -844,7 +844,7 @@ z " style="fill:#cc8963;stroke:#4c4c4c;stroke-linejoin:miter;stroke-width:1.5;"/> </g> <g id="patch_11"> - <path clip-path="url(#p51652a52ce)" d="M 107.263636 253.794855 + <path clip-path="url(#p5b84badbdb)" d="M 107.263636 253.794855 L 107.263636 253.794855 L 107.263636 253.794855 L 107.263636 253.794855 @@ -852,7 +852,7 @@ z " style="fill:#5875a4;stroke:#4c4c4c;stroke-linejoin:miter;stroke-width:0.75;"/> </g> <g id="patch_12"> - <path clip-path="url(#p51652a52ce)" d="M 107.263636 253.794855 + <path clip-path="url(#p5b84badbdb)" d="M 107.263636 253.794855 L 107.263636 253.794855 L 107.263636 253.794855 L 107.263636 253.794855 @@ -860,22 +860,22 @@ z " style="fill:#cc8963;stroke:#4c4c4c;stroke-linejoin:miter;stroke-width:0.75;"/> </g> <g id="line2d_9"> - <path clip-path="url(#p51652a52ce)" d="M 92.045455 218.291943 + <path clip-path="url(#p5b84badbdb)" d="M 92.045455 218.291943 L 92.045455 226.622382 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_10"> - <path clip-path="url(#p51652a52ce)" d="M 92.045455 201.736181 + <path clip-path="url(#p5b84badbdb)" d="M 92.045455 201.736181 L 92.045455 179.057415 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_11"> - <path clip-path="url(#p51652a52ce)" d="M 84.588545 226.622382 + <path clip-path="url(#p5b84badbdb)" d="M 84.588545 226.622382 L 99.502364 226.622382 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_12"> - <path clip-path="url(#p51652a52ce)" d="M 84.588545 179.057415 + <path clip-path="url(#p5b84badbdb)" d="M 84.588545 179.057415 L 99.502364 179.057415 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> @@ -886,244 +886,244 @@ L 2.12132 0 L -0 -3.535534 L -2.12132 -0 z -" id="m93c95c2b91" style="stroke:#4c4c4c;stroke-linejoin:miter;"/> +" id="m81d9bdafc0" style="stroke:#4c4c4c;stroke-linejoin:miter;"/> </defs> - <g clip-path="url(#p51652a52ce)"> - <use style="fill:#4c4c4c;stroke:#4c4c4c;stroke-linejoin:miter;" x="92.045455" xlink:href="#m93c95c2b91" y="167.915124"/> - <use style="fill:#4c4c4c;stroke:#4c4c4c;stroke-linejoin:miter;" x="92.045455" xlink:href="#m93c95c2b91" y="140.50614"/> + <g clip-path="url(#p5b84badbdb)"> + <use style="fill:#4c4c4c;stroke:#4c4c4c;stroke-linejoin:miter;" x="92.045455" xlink:href="#m81d9bdafc0" y="167.915124"/> + <use style="fill:#4c4c4c;stroke:#4c4c4c;stroke-linejoin:miter;" x="92.045455" xlink:href="#m81d9bdafc0" y="140.50614"/> </g> </g> <g id="line2d_14"> - <path clip-path="url(#p51652a52ce)" d="M 122.481818 223.074719 + <path clip-path="url(#p5b84badbdb)" d="M 122.481818 223.074719 L 122.481818 234.059335 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_15"> - <path clip-path="url(#p51652a52ce)" d="M 122.481818 200.527347 + <path clip-path="url(#p5b84badbdb)" d="M 122.481818 200.527347 L 122.481818 175.483472 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_16"> - <path clip-path="url(#p51652a52ce)" d="M 115.024909 234.059335 + <path clip-path="url(#p5b84badbdb)" d="M 115.024909 234.059335 L 129.938727 234.059335 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_17"> - <path clip-path="url(#p51652a52ce)" d="M 115.024909 175.483472 + <path clip-path="url(#p5b84badbdb)" d="M 115.024909 175.483472 L 129.938727 175.483472 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_18"> - <g clip-path="url(#p51652a52ce)"> - <use style="fill:#4c4c4c;stroke:#4c4c4c;stroke-linejoin:miter;" x="122.481818" xlink:href="#m93c95c2b91" y="162.265142"/> - <use style="fill:#4c4c4c;stroke:#4c4c4c;stroke-linejoin:miter;" x="122.481818" xlink:href="#m93c95c2b91" y="163.657928"/> - <use style="fill:#4c4c4c;stroke:#4c4c4c;stroke-linejoin:miter;" x="122.481818" xlink:href="#m93c95c2b91" y="145.551706"/> + <g clip-path="url(#p5b84badbdb)"> + <use style="fill:#4c4c4c;stroke:#4c4c4c;stroke-linejoin:miter;" x="122.481818" xlink:href="#m81d9bdafc0" y="162.265142"/> + <use style="fill:#4c4c4c;stroke:#4c4c4c;stroke-linejoin:miter;" x="122.481818" xlink:href="#m81d9bdafc0" y="163.657928"/> + <use style="fill:#4c4c4c;stroke:#4c4c4c;stroke-linejoin:miter;" x="122.481818" xlink:href="#m81d9bdafc0" y="145.551706"/> </g> </g> <g id="line2d_19"> - <path clip-path="url(#p51652a52ce)" d="M 168.136364 223.074719 + <path clip-path="url(#p5b84badbdb)" d="M 168.136364 223.074719 L 168.136364 238.684437 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_20"> - <path clip-path="url(#p51652a52ce)" d="M 168.136364 204.745125 + <path clip-path="url(#p5b84badbdb)" d="M 168.136364 204.745125 L 168.136364 177.664628 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_21"> - <path clip-path="url(#p51652a52ce)" d="M 160.679455 238.684437 + <path clip-path="url(#p5b84badbdb)" d="M 160.679455 238.684437 L 175.593273 238.684437 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_22"> - <path clip-path="url(#p51652a52ce)" d="M 160.679455 177.664628 + <path clip-path="url(#p5b84badbdb)" d="M 160.679455 177.664628 L 175.593273 177.664628 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_23"> - <g clip-path="url(#p51652a52ce)"> - <use style="fill:#4c4c4c;stroke:#4c4c4c;stroke-linejoin:miter;" x="168.136364" xlink:href="#m93c95c2b91" y="148.232162"/> + <g clip-path="url(#p5b84badbdb)"> + <use style="fill:#4c4c4c;stroke:#4c4c4c;stroke-linejoin:miter;" x="168.136364" xlink:href="#m81d9bdafc0" y="148.232162"/> </g> </g> <g id="line2d_24"> - <path clip-path="url(#p51652a52ce)" d="M 198.572727 214.113584 + <path clip-path="url(#p5b84badbdb)" d="M 198.572727 214.113584 L 198.572727 221.051237 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_25"> - <path clip-path="url(#p51652a52ce)" d="M 198.572727 194.522599 + <path clip-path="url(#p5b84badbdb)" d="M 198.572727 194.522599 L 198.572727 194.010159 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_26"> - <path clip-path="url(#p51652a52ce)" d="M 191.115818 221.051237 + <path clip-path="url(#p5b84badbdb)" d="M 191.115818 221.051237 L 206.029636 221.051237 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_27"> - <path clip-path="url(#p51652a52ce)" d="M 191.115818 194.010159 + <path clip-path="url(#p5b84badbdb)" d="M 191.115818 194.010159 L 206.029636 194.010159 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_28"/> <g id="line2d_29"> - <path clip-path="url(#p51652a52ce)" d="M 244.227273 218.567872 + <path clip-path="url(#p5b84badbdb)" d="M 244.227273 218.567872 L 244.227273 245.727206 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_30"> - <path clip-path="url(#p51652a52ce)" d="M 244.227273 183.386878 + <path clip-path="url(#p5b84badbdb)" d="M 244.227273 183.386878 L 244.227273 137.378941 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_31"> - <path clip-path="url(#p51652a52ce)" d="M 236.770364 245.727206 + <path clip-path="url(#p5b84badbdb)" d="M 236.770364 245.727206 L 251.684182 245.727206 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_32"> - <path clip-path="url(#p51652a52ce)" d="M 236.770364 137.378941 + <path clip-path="url(#p5b84badbdb)" d="M 236.770364 137.378941 L 251.684182 137.378941 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_33"> - <g clip-path="url(#p51652a52ce)"> - <use style="fill:#4c4c4c;stroke:#4c4c4c;stroke-linejoin:miter;" x="244.227273" xlink:href="#m93c95c2b91" y="120.27132"/> + <g clip-path="url(#p5b84badbdb)"> + <use style="fill:#4c4c4c;stroke:#4c4c4c;stroke-linejoin:miter;" x="244.227273" xlink:href="#m81d9bdafc0" y="120.27132"/> </g> </g> <g id="line2d_34"> - <path clip-path="url(#p51652a52ce)" d="M 274.663636 215.085906 + <path clip-path="url(#p5b84badbdb)" d="M 274.663636 215.085906 L 274.663636 234.742589 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_35"> - <path clip-path="url(#p51652a52ce)" d="M 274.663636 199.528746 + <path clip-path="url(#p5b84badbdb)" d="M 274.663636 199.528746 L 274.663636 177.506954 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_36"> - <path clip-path="url(#p51652a52ce)" d="M 267.206727 234.742589 + <path clip-path="url(#p5b84badbdb)" d="M 267.206727 234.742589 L 282.120545 234.742589 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_37"> - <path clip-path="url(#p51652a52ce)" d="M 267.206727 177.506954 + <path clip-path="url(#p5b84badbdb)" d="M 267.206727 177.506954 L 282.120545 177.506954 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_38"> - <g clip-path="url(#p51652a52ce)"> - <use style="fill:#4c4c4c;stroke:#4c4c4c;stroke-linejoin:miter;" x="274.663636" xlink:href="#m93c95c2b91" y="150.203087"/> - <use style="fill:#4c4c4c;stroke:#4c4c4c;stroke-linejoin:miter;" x="274.663636" xlink:href="#m93c95c2b91" y="171.620461"/> - <use style="fill:#4c4c4c;stroke:#4c4c4c;stroke-linejoin:miter;" x="274.663636" xlink:href="#m93c95c2b91" y="126.946183"/> - <use style="fill:#4c4c4c;stroke:#4c4c4c;stroke-linejoin:miter;" x="274.663636" xlink:href="#m93c95c2b91" y="126.788509"/> - <use style="fill:#4c4c4c;stroke:#4c4c4c;stroke-linejoin:miter;" x="274.663636" xlink:href="#m93c95c2b91" y="159.637243"/> + <g clip-path="url(#p5b84badbdb)"> + <use style="fill:#4c4c4c;stroke:#4c4c4c;stroke-linejoin:miter;" x="274.663636" xlink:href="#m81d9bdafc0" y="150.203087"/> + <use style="fill:#4c4c4c;stroke:#4c4c4c;stroke-linejoin:miter;" x="274.663636" xlink:href="#m81d9bdafc0" y="171.620461"/> + <use style="fill:#4c4c4c;stroke:#4c4c4c;stroke-linejoin:miter;" x="274.663636" xlink:href="#m81d9bdafc0" y="126.946183"/> + <use style="fill:#4c4c4c;stroke:#4c4c4c;stroke-linejoin:miter;" x="274.663636" xlink:href="#m81d9bdafc0" y="126.788509"/> + <use style="fill:#4c4c4c;stroke:#4c4c4c;stroke-linejoin:miter;" x="274.663636" xlink:href="#m81d9bdafc0" y="159.637243"/> </g> </g> <g id="line2d_39"> - <path clip-path="url(#p51652a52ce)" d="M 320.318182 208.686973 + <path clip-path="url(#p5b84badbdb)" d="M 320.318182 208.686973 L 320.318182 234.742589 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_40"> - <path clip-path="url(#p51652a52ce)" d="M 320.318182 168.716633 + <path clip-path="url(#p5b84badbdb)" d="M 320.318182 168.716633 L 320.318182 134.619647 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_41"> - <path clip-path="url(#p51652a52ce)" d="M 312.861273 234.742589 + <path clip-path="url(#p5b84badbdb)" d="M 312.861273 234.742589 L 327.775091 234.742589 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_42"> - <path clip-path="url(#p51652a52ce)" d="M 312.861273 134.619647 + <path clip-path="url(#p5b84badbdb)" d="M 312.861273 134.619647 L 327.775091 134.619647 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_43"/> <g id="line2d_44"> - <path clip-path="url(#p51652a52ce)" d="M 350.754545 214.954511 + <path clip-path="url(#p5b84badbdb)" d="M 350.754545 214.954511 L 350.754545 230.748183 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_45"> - <path clip-path="url(#p51652a52ce)" d="M 350.754545 188.097386 + <path clip-path="url(#p5b84badbdb)" d="M 350.754545 188.097386 L 350.754545 153.75075 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_46"> - <path clip-path="url(#p51652a52ce)" d="M 343.297636 230.748183 + <path clip-path="url(#p5b84badbdb)" d="M 343.297636 230.748183 L 358.211455 230.748183 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_47"> - <path clip-path="url(#p51652a52ce)" d="M 343.297636 153.75075 + <path clip-path="url(#p5b84badbdb)" d="M 343.297636 153.75075 L 358.211455 153.75075 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_48"> - <g clip-path="url(#p51652a52ce)"> - <use style="fill:#4c4c4c;stroke:#4c4c4c;stroke-linejoin:miter;" x="350.754545" xlink:href="#m93c95c2b91" y="127.208973"/> + <g clip-path="url(#p5b84badbdb)"> + <use style="fill:#4c4c4c;stroke:#4c4c4c;stroke-linejoin:miter;" x="350.754545" xlink:href="#m81d9bdafc0" y="127.208973"/> </g> </g> <g id="line2d_49"> - <path clip-path="url(#p51652a52ce)" d="M 244.227273 91.084363 + <path clip-path="url(#p5b84badbdb)" d="M 244.227273 91.084363 L 244.227273 84.18429 L 274.663636 84.18429 L 274.663636 91.084363 " style="fill:none;stroke:#333333;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_50"> - <path clip-path="url(#p51652a52ce)" d="M 122.481818 116.364749 + <path clip-path="url(#p5b84badbdb)" d="M 122.481818 116.364749 L 122.481818 109.464676 L 198.572727 109.464676 L 198.572727 116.364749 " style="fill:none;stroke:#333333;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_51"> - <path clip-path="url(#p51652a52ce)" d="M 92.045455 61.276864 + <path clip-path="url(#p5b84badbdb)" d="M 92.045455 61.276864 L 92.045455 54.37679 L 350.754545 54.37679 L 350.754545 61.276864 " style="fill:none;stroke:#333333;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_52"> - <path clip-path="url(#p51652a52ce)" d="M 77.131636 210.513363 + <path clip-path="url(#p5b84badbdb)" d="M 77.131636 210.513363 L 106.959273 210.513363 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_53"> - <path clip-path="url(#p51652a52ce)" d="M 107.568 211.87987 + <path clip-path="url(#p5b84badbdb)" d="M 107.568 211.87987 L 137.395636 211.87987 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_54"> - <path clip-path="url(#p51652a52ce)" d="M 153.222545 218.528454 + <path clip-path="url(#p5b84badbdb)" d="M 153.222545 218.528454 L 183.050182 218.528454 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_55"> - <path clip-path="url(#p51652a52ce)" d="M 183.658909 203.247223 + <path clip-path="url(#p5b84badbdb)" d="M 183.658909 203.247223 L 213.486545 203.247223 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_56"> - <path clip-path="url(#p51652a52ce)" d="M 229.313455 200.212 + <path clip-path="url(#p5b84badbdb)" d="M 229.313455 200.212 L 259.141091 200.212 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_57"> - <path clip-path="url(#p51652a52ce)" d="M 259.749818 206.965699 + <path clip-path="url(#p5b84badbdb)" d="M 259.749818 206.965699 L 289.577455 206.965699 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_58"> - <path clip-path="url(#p51652a52ce)" d="M 305.404364 193.090394 + <path clip-path="url(#p5b84badbdb)" d="M 305.404364 193.090394 L 335.232 193.090394 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_59"> - <path clip-path="url(#p51652a52ce)" d="M 335.840727 205.362681 + <path clip-path="url(#p5b84badbdb)" d="M 335.840727 205.362681 L 365.668364 205.362681 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> @@ -1426,7 +1426,7 @@ z </g> </g> <defs> - <clipPath id="p51652a52ce"> + <clipPath id="p5b84badbdb"> <rect height="217.44" width="334.8" x="54" y="34.56"/> </clipPath> </defs> diff --git a/usage/flu_dataset_log_scale_in_axes.svg b/usage/flu_dataset_log_scale_in_axes.svg index f5e13a3..3cb7293 100644 --- a/usage/flu_dataset_log_scale_in_axes.svg +++ b/usage/flu_dataset_log_scale_in_axes.svg @@ -510,7 +510,7 @@ z <g id="matplotlib.axis_2"> <g id="ytick_1"> <g id="line2d_1"> - <path clip-path="url(#p37bf755c76)" d="M 108 355.511801 + <path clip-path="url(#p19cc01c258)" d="M 108 355.511801 L 777.6 355.511801 " style="fill:none;stroke:#cccccc;stroke-linecap:round;"/> </g> @@ -584,7 +584,7 @@ z </g> <g id="ytick_2"> <g id="line2d_2"> - <path clip-path="url(#p37bf755c76)" d="M 108 260.424716 + <path clip-path="url(#p19cc01c258)" d="M 108 260.424716 L 777.6 260.424716 " style="fill:none;stroke:#cccccc;stroke-linecap:round;"/> </g> @@ -619,7 +619,7 @@ z </g> <g id="ytick_3"> <g id="line2d_3"> - <path clip-path="url(#p37bf755c76)" d="M 108 165.337631 + <path clip-path="url(#p19cc01c258)" d="M 108 165.337631 L 777.6 165.337631 " style="fill:none;stroke:#cccccc;stroke-linecap:round;"/> </g> @@ -635,7 +635,7 @@ L 777.6 165.337631 </g> <g id="ytick_4"> <g id="line2d_4"> - <path clip-path="url(#p37bf755c76)" d="M 108 70.250546 + <path clip-path="url(#p19cc01c258)" d="M 108 70.250546 L 777.6 70.250546 " style="fill:none;stroke:#cccccc;stroke-linecap:round;"/> </g> @@ -790,21 +790,21 @@ z <g id="PathCollection_1"/> <g id="PathCollection_2"/> <g id="line2d_5"> - <path clip-path="url(#p37bf755c76)" d="M 398.16 122.800151 + <path clip-path="url(#p19cc01c258)" d="M 398.16 122.800151 L 398.16 117.015595 L 487.44 117.015595 L 487.44 122.800151 " style="fill:none;stroke:#333333;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_6"> - <path clip-path="url(#p37bf755c76)" d="M 174.96 84.565272 + <path clip-path="url(#p19cc01c258)" d="M 174.96 84.565272 L 174.96 78.780716 L 264.24 78.780716 L 264.24 84.565272 " style="fill:none;stroke:#333333;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_7"> - <path clip-path="url(#p37bf755c76)" d="M 621.36 136.464802 + <path clip-path="url(#p19cc01c258)" d="M 621.36 136.464802 L 621.36 130.680247 L 710.64 130.680247 L 710.64 136.464802 @@ -842,589 +842,589 @@ C -2.236584 -1.29895 -2.5 -0.663008 -2.5 0 C -2.5 0.663008 -2.236584 1.29895 -1.767767 1.767767 C -1.29895 2.236584 -0.663008 2.5 0 2.5 z -" id="C0_0_eaabaca787"/> +" id="C0_0_e64610c732"/> </defs> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_eaabaca787" y="364.853283"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_e64610c732" y="364.853283"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_eaabaca787" y="336.244734"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_e64610c732" y="336.244734"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_eaabaca787" y="327.330163"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_e64610c732" y="327.330163"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="169.710003" xlink:href="#C0_0_eaabaca787" y="327.325352"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="169.710003" xlink:href="#C0_0_e64610c732" y="327.325352"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_eaabaca787" y="321.598017"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_e64610c732" y="321.598017"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="170.818467" xlink:href="#C0_0_eaabaca787" y="318.873123"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="170.818467" xlink:href="#C0_0_e64610c732" y="318.873123"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="179.092496" xlink:href="#C0_0_eaabaca787" y="318.863354"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="179.092496" xlink:href="#C0_0_e64610c732" y="318.863354"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_eaabaca787" y="308.233739"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_e64610c732" y="308.233739"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="169.71" xlink:href="#C0_0_eaabaca787" y="308.23363"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="169.71" xlink:href="#C0_0_e64610c732" y="308.23363"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="180.20999" xlink:href="#C0_0_eaabaca787" y="308.225156"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="180.20999" xlink:href="#C0_0_e64610c732" y="308.225156"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="164.460017" xlink:href="#C0_0_eaabaca787" y="308.222402"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="164.460017" xlink:href="#C0_0_e64610c732" y="308.222402"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="185.459988" xlink:href="#C0_0_eaabaca787" y="308.22121"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="185.459988" xlink:href="#C0_0_e64610c732" y="308.22121"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="159.210017" xlink:href="#C0_0_eaabaca787" y="308.220715"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="159.210017" xlink:href="#C0_0_e64610c732" y="308.220715"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="190.709988" xlink:href="#C0_0_eaabaca787" y="308.220501"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="190.709988" xlink:href="#C0_0_e64610c732" y="308.220501"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="153.960017" xlink:href="#C0_0_eaabaca787" y="308.219795"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="153.960017" xlink:href="#C0_0_e64610c732" y="308.219795"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_eaabaca787" y="297.14141"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_e64610c732" y="297.14141"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="170.375862" xlink:href="#C0_0_eaabaca787" y="294.980284"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="170.375862" xlink:href="#C0_0_e64610c732" y="294.980284"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="179.537229" xlink:href="#C0_0_eaabaca787" y="294.969863"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="179.537229" xlink:href="#C0_0_e64610c732" y="294.969863"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="184.787228" xlink:href="#C0_0_eaabaca787" y="294.967678"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="184.787228" xlink:href="#C0_0_e64610c732" y="294.967678"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="165.299683" xlink:href="#C0_0_eaabaca787" y="293.848819"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="165.299683" xlink:href="#C0_0_e64610c732" y="293.848819"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="189.865013" xlink:href="#C0_0_eaabaca787" y="293.841365"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="189.865013" xlink:href="#C0_0_e64610c732" y="293.841365"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="160.049695" xlink:href="#C0_0_eaabaca787" y="293.8393"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="160.049695" xlink:href="#C0_0_e64610c732" y="293.8393"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="195.115012" xlink:href="#C0_0_eaabaca787" y="293.839283"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="195.115012" xlink:href="#C0_0_e64610c732" y="293.839283"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="154.79974" xlink:href="#C0_0_eaabaca787" y="293.820936"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="154.79974" xlink:href="#C0_0_e64610c732" y="293.820936"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="200.26563" xlink:href="#C0_0_eaabaca787" y="292.980658"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="200.26563" xlink:href="#C0_0_e64610c732" y="292.980658"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_eaabaca787" y="290.253898"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_e64610c732" y="290.253898"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="169.710014" xlink:href="#C0_0_eaabaca787" y="290.24361"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="169.710014" xlink:href="#C0_0_e64610c732" y="290.24361"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="180.209979" xlink:href="#C0_0_eaabaca787" y="290.241316"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="180.209979" xlink:href="#C0_0_e64610c732" y="290.241316"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="164.68005" xlink:href="#C0_0_eaabaca787" y="288.973439"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="164.68005" xlink:href="#C0_0_e64610c732" y="288.973439"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="184.492216" xlink:href="#C0_0_eaabaca787" y="287.67623"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="184.492216" xlink:href="#C0_0_e64610c732" y="287.67623"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_eaabaca787" y="281.957343"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_e64610c732" y="281.957343"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="169.71" xlink:href="#C0_0_eaabaca787" y="281.957326"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="169.71" xlink:href="#C0_0_e64610c732" y="281.957326"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="179.412672" xlink:href="#C0_0_eaabaca787" y="279.608336"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="179.412672" xlink:href="#C0_0_e64610c732" y="279.608336"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="165.260883" xlink:href="#C0_0_eaabaca787" y="279.603519"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="165.260883" xlink:href="#C0_0_e64610c732" y="279.603519"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="184.662667" xlink:href="#C0_0_eaabaca787" y="279.602641"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="184.662667" xlink:href="#C0_0_e64610c732" y="279.602641"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="160.010884" xlink:href="#C0_0_eaabaca787" y="279.602216"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="160.010884" xlink:href="#C0_0_e64610c732" y="279.602216"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="189.912663" xlink:href="#C0_0_eaabaca787" y="279.596838"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="189.912663" xlink:href="#C0_0_e64610c732" y="279.596838"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="154.760903" xlink:href="#C0_0_eaabaca787" y="279.590004"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="154.760903" xlink:href="#C0_0_e64610c732" y="279.590004"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_eaabaca787" y="276.227583"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_e64610c732" y="276.227583"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="169.777725" xlink:href="#C0_0_eaabaca787" y="275.517702"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="169.777725" xlink:href="#C0_0_e64610c732" y="275.517702"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="179.097481" xlink:href="#C0_0_eaabaca787" y="273.498302"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="179.097481" xlink:href="#C0_0_e64610c732" y="273.498302"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="184.34748" xlink:href="#C0_0_eaabaca787" y="273.494529"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="184.34748" xlink:href="#C0_0_e64610c732" y="273.494529"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="174.361283" xlink:href="#C0_0_eaabaca787" y="271.58529"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="174.361283" xlink:href="#C0_0_e64610c732" y="271.58529"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="167.35685" xlink:href="#C0_0_eaabaca787" y="271.583392"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="167.35685" xlink:href="#C0_0_e64610c732" y="271.583392"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="162.106856" xlink:href="#C0_0_eaabaca787" y="271.577054"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="162.106856" xlink:href="#C0_0_e64610c732" y="271.577054"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="187.358661" xlink:href="#C0_0_eaabaca787" y="269.862489"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="187.358661" xlink:href="#C0_0_e64610c732" y="269.862489"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_eaabaca787" y="265.231804"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_e64610c732" y="265.231804"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="169.710005" xlink:href="#C0_0_eaabaca787" y="265.225982"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="169.710005" xlink:href="#C0_0_e64610c732" y="265.225982"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="180.209976" xlink:href="#C0_0_eaabaca787" y="265.218325"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="180.209976" xlink:href="#C0_0_e64610c732" y="265.218325"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="185.459976" xlink:href="#C0_0_eaabaca787" y="265.217742"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="185.459976" xlink:href="#C0_0_e64610c732" y="265.217742"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="164.460021" xlink:href="#C0_0_eaabaca787" y="265.214756"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="164.460021" xlink:href="#C0_0_e64610c732" y="265.214756"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="190.709972" xlink:href="#C0_0_eaabaca787" y="265.212449"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="190.709972" xlink:href="#C0_0_e64610c732" y="265.212449"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="159.210023" xlink:href="#C0_0_eaabaca787" y="265.211275"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="159.210023" xlink:href="#C0_0_e64610c732" y="265.211275"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="195.959969" xlink:href="#C0_0_eaabaca787" y="265.207677"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="195.959969" xlink:href="#C0_0_e64610c732" y="265.207677"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="154.755889" xlink:href="#C0_0_eaabaca787" y="262.864246"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="154.755889" xlink:href="#C0_0_e64610c732" y="262.864246"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="200.409433" xlink:href="#C0_0_eaabaca787" y="262.854339"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="200.409433" xlink:href="#C0_0_e64610c732" y="262.854339"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="149.505906" xlink:href="#C0_0_eaabaca787" y="262.852999"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="149.505906" xlink:href="#C0_0_e64610c732" y="262.852999"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="174.033579" xlink:href="#C0_0_eaabaca787" y="260.867548"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="174.033579" xlink:href="#C0_0_e64610c732" y="260.867548"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="179.247123" xlink:href="#C0_0_eaabaca787" y="260.345939"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="179.247123" xlink:href="#C0_0_e64610c732" y="260.345939"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="169.047919" xlink:href="#C0_0_eaabaca787" y="259.478366"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="169.047919" xlink:href="#C0_0_e64610c732" y="259.478366"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="184.270248" xlink:href="#C0_0_eaabaca787" y="259.056609"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="184.270248" xlink:href="#C0_0_e64610c732" y="259.056609"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_eaabaca787" y="255.133958"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_e64610c732" y="255.133958"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="169.771727" xlink:href="#C0_0_eaabaca787" y="254.45605"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="169.771727" xlink:href="#C0_0_e64610c732" y="254.45605"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="179.754578" xlink:href="#C0_0_eaabaca787" y="253.327647"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="179.754578" xlink:href="#C0_0_e64610c732" y="253.327647"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="165.113941" xlink:href="#C0_0_eaabaca787" y="252.410323"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="165.113941" xlink:href="#C0_0_e64610c732" y="252.410323"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="173.106762" xlink:href="#C0_0_eaabaca787" y="250.985557"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="173.106762" xlink:href="#C0_0_e64610c732" y="250.985557"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="184.212226" xlink:href="#C0_0_eaabaca787" y="250.985381"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="184.212226" xlink:href="#C0_0_e64610c732" y="250.985381"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="189.462222" xlink:href="#C0_0_eaabaca787" y="250.979552"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="189.462222" xlink:href="#C0_0_e64610c732" y="250.979552"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="160.147123" xlink:href="#C0_0_eaabaca787" y="250.973806"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="160.147123" xlink:href="#C0_0_e64610c732" y="250.973806"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="194.712198" xlink:href="#C0_0_eaabaca787" y="250.966056"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="194.712198" xlink:href="#C0_0_e64610c732" y="250.966056"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="154.928967" xlink:href="#C0_0_eaabaca787" y="250.486198"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="154.928967" xlink:href="#C0_0_e64610c732" y="250.486198"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="177.429427" xlink:href="#C0_0_eaabaca787" y="248.469301"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="177.429427" xlink:href="#C0_0_e64610c732" y="248.469301"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="168.785553" xlink:href="#C0_0_eaabaca787" y="248.467519"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="168.785553" xlink:href="#C0_0_e64610c732" y="248.467519"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="199.049167" xlink:href="#C0_0_eaabaca787" y="248.467419"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="199.049167" xlink:href="#C0_0_e64610c732" y="248.467419"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="163.845692" xlink:href="#C0_0_eaabaca787" y="246.966166"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="163.845692" xlink:href="#C0_0_e64610c732" y="246.966166"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="174.333039" xlink:href="#C0_0_eaabaca787" y="244.888724"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="174.333039" xlink:href="#C0_0_e64610c732" y="244.888724"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="179.059296" xlink:href="#C0_0_eaabaca787" y="242.958253"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="179.059296" xlink:href="#C0_0_e64610c732" y="242.958253"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="169.870691" xlink:href="#C0_0_eaabaca787" y="242.552852"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="169.870691" xlink:href="#C0_0_e64610c732" y="242.552852"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="184.251342" xlink:href="#C0_0_eaabaca787" y="242.301269"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="184.251342" xlink:href="#C0_0_e64610c732" y="242.301269"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="164.789456" xlink:href="#C0_0_eaabaca787" y="241.437691"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="164.789456" xlink:href="#C0_0_e64610c732" y="241.437691"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_eaabaca787" y="239.253126"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_e64610c732" y="239.253126"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="180.95004" xlink:href="#C0_0_eaabaca787" y="238.821941"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="180.95004" xlink:href="#C0_0_e64610c732" y="238.821941"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="170.032046" xlink:href="#C0_0_eaabaca787" y="237.72412"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="170.032046" xlink:href="#C0_0_e64610c732" y="237.72412"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="164.953372" xlink:href="#C0_0_eaabaca787" y="236.60067"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="164.953372" xlink:href="#C0_0_e64610c732" y="236.60067"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="185.488893" xlink:href="#C0_0_eaabaca787" y="236.593682"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="185.488893" xlink:href="#C0_0_e64610c732" y="236.593682"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="159.703389" xlink:href="#C0_0_eaabaca787" y="236.58931"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="159.703389" xlink:href="#C0_0_e64610c732" y="236.58931"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="190.738888" xlink:href="#C0_0_eaabaca787" y="236.587651"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="190.738888" xlink:href="#C0_0_e64610c732" y="236.587651"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_eaabaca787" y="234.220606"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_e64610c732" y="234.220606"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="180.179646" xlink:href="#C0_0_eaabaca787" y="233.744512"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="180.179646" xlink:href="#C0_0_e64610c732" y="233.744512"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="170.620825" xlink:href="#C0_0_eaabaca787" y="231.7247"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="170.620825" xlink:href="#C0_0_e64610c732" y="231.7247"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_eaabaca787" y="228.41365"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_e64610c732" y="228.41365"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="180.065213" xlink:href="#C0_0_eaabaca787" y="227.379546"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="180.065213" xlink:href="#C0_0_e64610c732" y="227.379546"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="172.073166" xlink:href="#C0_0_eaabaca787" y="224.710299"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="172.073166" xlink:href="#C0_0_e64610c732" y="224.710299"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="176.517526" xlink:href="#C0_0_eaabaca787" y="222.350092"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="176.517526" xlink:href="#C0_0_e64610c732" y="222.350092"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="172.175421" xlink:href="#C0_0_eaabaca787" y="219.857825"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="172.175421" xlink:href="#C0_0_e64610c732" y="219.857825"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="180.849524" xlink:href="#C0_0_eaabaca787" y="219.845311"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="180.849524" xlink:href="#C0_0_e64610c732" y="219.845311"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="166.925442" xlink:href="#C0_0_eaabaca787" y="219.845244"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="166.925442" xlink:href="#C0_0_e64610c732" y="219.845244"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="176.402495" xlink:href="#C0_0_eaabaca787" y="217.488692"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="176.402495" xlink:href="#C0_0_e64610c732" y="217.488692"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="171.245614" xlink:href="#C0_0_eaabaca787" y="215.494083"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="171.245614" xlink:href="#C0_0_e64610c732" y="215.494083"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="180.086055" xlink:href="#C0_0_eaabaca787" y="214.329404"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="180.086055" xlink:href="#C0_0_e64610c732" y="214.329404"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="166.184811" xlink:href="#C0_0_eaabaca787" y="214.314513"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="166.184811" xlink:href="#C0_0_e64610c732" y="214.314513"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="185.241932" xlink:href="#C0_0_eaabaca787" y="213.493596"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="185.241932" xlink:href="#C0_0_e64610c732" y="213.493596"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="161.026619" xlink:href="#C0_0_eaabaca787" y="213.488957"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="161.026619" xlink:href="#C0_0_e64610c732" y="213.488957"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="190.491918" xlink:href="#C0_0_eaabaca787" y="213.483422"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="190.491918" xlink:href="#C0_0_e64610c732" y="213.483422"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="155.776623" xlink:href="#C0_0_eaabaca787" y="213.483098"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="155.776623" xlink:href="#C0_0_e64610c732" y="213.483098"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="195.741916" xlink:href="#C0_0_eaabaca787" y="213.479362"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="195.741916" xlink:href="#C0_0_e64610c732" y="213.479362"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="150.526626" xlink:href="#C0_0_eaabaca787" y="213.478317"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="150.526626" xlink:href="#C0_0_e64610c732" y="213.478317"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="200.991912" xlink:href="#C0_0_eaabaca787" y="213.473832"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="200.991912" xlink:href="#C0_0_e64610c732" y="213.473832"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="145.294046" xlink:href="#C0_0_eaabaca787" y="213.117429"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="145.294046" xlink:href="#C0_0_e64610c732" y="213.117429"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="206.18685" xlink:href="#C0_0_eaabaca787" y="212.833363"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="206.18685" xlink:href="#C0_0_e64610c732" y="212.833363"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_eaabaca787" y="209.897757"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_e64610c732" y="209.897757"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="180.14391" xlink:href="#C0_0_eaabaca787" y="209.196444"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="180.14391" xlink:href="#C0_0_e64610c732" y="209.196444"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="169.933723" xlink:href="#C0_0_eaabaca787" y="208.617217"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="169.933723" xlink:href="#C0_0_e64610c732" y="208.617217"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="164.740422" xlink:href="#C0_0_eaabaca787" y="207.967349"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="164.740422" xlink:href="#C0_0_e64610c732" y="207.967349"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="185.187474" xlink:href="#C0_0_eaabaca787" y="207.965338"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="185.187474" xlink:href="#C0_0_e64610c732" y="207.965338"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="159.606412" xlink:href="#C0_0_eaabaca787" y="207.040491"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="159.606412" xlink:href="#C0_0_e64610c732" y="207.040491"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="190.321587" xlink:href="#C0_0_eaabaca787" y="207.038886"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="190.321587" xlink:href="#C0_0_e64610c732" y="207.038886"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="154.356418" xlink:href="#C0_0_eaabaca787" y="207.033452"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="154.356418" xlink:href="#C0_0_e64610c732" y="207.033452"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="173.646709" xlink:href="#C0_0_eaabaca787" y="205.604889"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="173.646709" xlink:href="#C0_0_e64610c732" y="205.604889"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="195.05219" xlink:href="#C0_0_eaabaca787" y="205.116025"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="195.05219" xlink:href="#C0_0_e64610c732" y="205.116025"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="149.623563" xlink:href="#C0_0_eaabaca787" y="205.114548"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="149.623563" xlink:href="#C0_0_e64610c732" y="205.114548"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="178.206759" xlink:href="#C0_0_eaabaca787" y="203.407715"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="178.206759" xlink:href="#C0_0_e64610c732" y="203.407715"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="169.31113" xlink:href="#C0_0_eaabaca787" y="203.104531"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="169.31113" xlink:href="#C0_0_e64610c732" y="203.104531"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="183.395232" xlink:href="#C0_0_eaabaca787" y="202.730895"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="183.395232" xlink:href="#C0_0_e64610c732" y="202.730895"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="164.147683" xlink:href="#C0_0_eaabaca787" y="202.302745"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="164.147683" xlink:href="#C0_0_e64610c732" y="202.302745"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="188.618949" xlink:href="#C0_0_eaabaca787" y="202.287795"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="188.618949" xlink:href="#C0_0_e64610c732" y="202.287795"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="158.905458" xlink:href="#C0_0_eaabaca787" y="202.061532"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="158.905458" xlink:href="#C0_0_e64610c732" y="202.061532"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="153.660371" xlink:href="#C0_0_eaabaca787" y="201.869769"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="153.660371" xlink:href="#C0_0_e64610c732" y="201.869769"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="198.626342" xlink:href="#C0_0_eaabaca787" y="201.86834"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="198.626342" xlink:href="#C0_0_e64610c732" y="201.86834"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="146.047922" xlink:href="#C0_0_eaabaca787" y="201.868033"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="146.047922" xlink:href="#C0_0_e64610c732" y="201.868033"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="203.848882" xlink:href="#C0_0_eaabaca787" y="201.41545"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="203.848882" xlink:href="#C0_0_e64610c732" y="201.41545"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="174.677654" xlink:href="#C0_0_eaabaca787" y="200.125082"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="174.677654" xlink:href="#C0_0_e64610c732" y="200.125082"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="193.150493" xlink:href="#C0_0_eaabaca787" y="200.048951"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="193.150493" xlink:href="#C0_0_e64610c732" y="200.048951"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="141.407376" xlink:href="#C0_0_eaabaca787" y="199.794549"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="141.407376" xlink:href="#C0_0_e64610c732" y="199.794549"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="208.734953" xlink:href="#C0_0_eaabaca787" y="199.793404"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="208.734953" xlink:href="#C0_0_e64610c732" y="199.793404"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="136.159252" xlink:href="#C0_0_eaabaca787" y="199.676046"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="136.159252" xlink:href="#C0_0_e64610c732" y="199.676046"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="213.983102" xlink:href="#C0_0_eaabaca787" y="199.675669"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="213.983102" xlink:href="#C0_0_e64610c732" y="199.675669"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="131.2128" xlink:href="#C0_0_eaabaca787" y="199.52131"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="131.2128" xlink:href="#C0_0_e64610c732" y="199.52131"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="179.993462" xlink:href="#C0_0_eaabaca787" y="199.238546"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="179.993462" xlink:href="#C0_0_e64610c732" y="199.238546"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="169.776832" xlink:href="#C0_0_eaabaca787" y="198.535095"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="169.776832" xlink:href="#C0_0_e64610c732" y="198.535095"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="184.865302" xlink:href="#C0_0_eaabaca787" y="197.586251"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="184.865302" xlink:href="#C0_0_e64610c732" y="197.586251"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="165.173428" xlink:href="#C0_0_eaabaca787" y="196.403376"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="165.173428" xlink:href="#C0_0_e64610c732" y="196.403376"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="189.920554" xlink:href="#C0_0_eaabaca787" y="196.389823"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="189.920554" xlink:href="#C0_0_e64610c732" y="196.389823"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="159.937239" xlink:href="#C0_0_eaabaca787" y="196.081986"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="159.937239" xlink:href="#C0_0_e64610c732" y="196.081986"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_eaabaca787" y="195.166563"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_e64610c732" y="195.166563"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="194.864027" xlink:href="#C0_0_eaabaca787" y="194.896978"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="194.864027" xlink:href="#C0_0_e64610c732" y="194.896978"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="180.333223" xlink:href="#C0_0_eaabaca787" y="194.814008"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="180.333223" xlink:href="#C0_0_e64610c732" y="194.814008"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="169.935692" xlink:href="#C0_0_eaabaca787" y="193.880522"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="169.935692" xlink:href="#C0_0_e64610c732" y="193.880522"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="155.480531" xlink:href="#C0_0_eaabaca787" y="193.738446"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="155.480531" xlink:href="#C0_0_e64610c732" y="193.738446"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="185.000983" xlink:href="#C0_0_eaabaca787" y="192.784561"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="185.000983" xlink:href="#C0_0_e64610c732" y="192.784561"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="177.252999" xlink:href="#C0_0_eaabaca787" y="191.223501"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="177.252999" xlink:href="#C0_0_e64610c732" y="191.223501"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="165.737133" xlink:href="#C0_0_eaabaca787" y="191.218602"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="165.737133" xlink:href="#C0_0_e64610c732" y="191.218602"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="160.487134" xlink:href="#C0_0_eaabaca787" y="191.215615"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="160.487134" xlink:href="#C0_0_e64610c732" y="191.215615"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="172.478476" xlink:href="#C0_0_eaabaca787" y="189.379688"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="172.478476" xlink:href="#C0_0_e64610c732" y="189.379688"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="181.695518" xlink:href="#C0_0_eaabaca787" y="188.860824"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="181.695518" xlink:href="#C0_0_e64610c732" y="188.860824"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="187.445791" xlink:href="#C0_0_eaabaca787" y="188.860821"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="187.445791" xlink:href="#C0_0_e64610c732" y="188.860821"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="192.576512" xlink:href="#C0_0_eaabaca787" y="187.921065"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="192.576512" xlink:href="#C0_0_e64610c732" y="187.921065"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="156.980381" xlink:href="#C0_0_eaabaca787" y="187.915941"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="156.980381" xlink:href="#C0_0_e64610c732" y="187.915941"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="175.410374" xlink:href="#C0_0_eaabaca787" y="185.701675"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="175.410374" xlink:href="#C0_0_e64610c732" y="185.701675"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="170.255553" xlink:href="#C0_0_eaabaca787" y="184.861231"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="170.255553" xlink:href="#C0_0_e64610c732" y="184.861231"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="180.46705" xlink:href="#C0_0_eaabaca787" y="184.509548"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="180.46705" xlink:href="#C0_0_e64610c732" y="184.509548"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="165.063223" xlink:href="#C0_0_eaabaca787" y="184.205848"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="165.063223" xlink:href="#C0_0_e64610c732" y="184.205848"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="184.798294" xlink:href="#C0_0_eaabaca787" y="182.003838"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="184.798294" xlink:href="#C0_0_e64610c732" y="182.003838"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_eaabaca787" y="179.346155"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_e64610c732" y="179.346155"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="170.513744" xlink:href="#C0_0_eaabaca787" y="176.988497"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="170.513744" xlink:href="#C0_0_e64610c732" y="176.988497"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="179.286465" xlink:href="#C0_0_eaabaca787" y="176.834562"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="179.286465" xlink:href="#C0_0_e64610c732" y="176.834562"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_eaabaca787" y="174.484336"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_e64610c732" y="174.484336"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="183.736877" xlink:href="#C0_0_eaabaca787" y="174.482503"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="183.736877" xlink:href="#C0_0_e64610c732" y="174.482503"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="166.187614" xlink:href="#C0_0_eaabaca787" y="174.476492"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="166.187614" xlink:href="#C0_0_e64610c732" y="174.476492"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="161.024387" xlink:href="#C0_0_eaabaca787" y="173.673696"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="161.024387" xlink:href="#C0_0_e64610c732" y="173.673696"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="188.266949" xlink:href="#C0_0_eaabaca787" y="172.241536"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="188.266949" xlink:href="#C0_0_e64610c732" y="172.241536"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="193.516947" xlink:href="#C0_0_eaabaca787" y="172.237632"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="193.516947" xlink:href="#C0_0_e64610c732" y="172.237632"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="173.978286" xlink:href="#C0_0_eaabaca787" y="170.128711"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="173.978286" xlink:href="#C0_0_e64610c732" y="170.128711"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="179.103612" xlink:href="#C0_0_eaabaca787" y="169.168184"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="179.103612" xlink:href="#C0_0_e64610c732" y="169.168184"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="169.292124" xlink:href="#C0_0_eaabaca787" y="168.12974"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="169.292124" xlink:href="#C0_0_e64610c732" y="168.12974"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="184.203128" xlink:href="#C0_0_eaabaca787" y="168.114221"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="184.203128" xlink:href="#C0_0_e64610c732" y="168.114221"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="164.059802" xlink:href="#C0_0_eaabaca787" y="167.766194"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="164.059802" xlink:href="#C0_0_e64610c732" y="167.766194"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="188.933503" xlink:href="#C0_0_eaabaca787" y="166.190959"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="188.933503" xlink:href="#C0_0_e64610c732" y="166.190959"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_eaabaca787" y="164.358533"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_e64610c732" y="164.358533"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="180.209994" xlink:href="#C0_0_eaabaca787" y="164.351989"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="180.209994" xlink:href="#C0_0_e64610c732" y="164.351989"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="170.139319" xlink:href="#C0_0_eaabaca787" y="162.602472"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="170.139319" xlink:href="#C0_0_e64610c732" y="162.602472"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="164.889323" xlink:href="#C0_0_eaabaca787" y="162.597059"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="164.889323" xlink:href="#C0_0_e64610c732" y="162.597059"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_eaabaca787" y="159.747356"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_e64610c732" y="159.747356"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="180.183647" xlink:href="#C0_0_eaabaca787" y="159.303658"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="180.183647" xlink:href="#C0_0_e64610c732" y="159.303658"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="185.433605" xlink:href="#C0_0_eaabaca787" y="159.286053"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="185.433605" xlink:href="#C0_0_e64610c732" y="159.286053"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="170.283045" xlink:href="#C0_0_eaabaca787" y="157.733064"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="170.283045" xlink:href="#C0_0_e64610c732" y="157.733064"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="165.344636" xlink:href="#C0_0_eaabaca787" y="156.228311"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="165.344636" xlink:href="#C0_0_e64610c732" y="156.228311"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_eaabaca787" y="154.565069"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_e64610c732" y="154.565069"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="180.021095" xlink:href="#C0_0_eaabaca787" y="153.386395"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="180.021095" xlink:href="#C0_0_e64610c732" y="153.386395"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="170.480743" xlink:href="#C0_0_eaabaca787" y="152.252394"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="170.480743" xlink:href="#C0_0_e64610c732" y="152.252394"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="185.086447" xlink:href="#C0_0_eaabaca787" y="152.220836"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="185.086447" xlink:href="#C0_0_e64610c732" y="152.220836"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_eaabaca787" y="149.452862"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_e64610c732" y="149.452862"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="166.64423" xlink:href="#C0_0_eaabaca787" y="149.225728"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="166.64423" xlink:href="#C0_0_e64610c732" y="149.225728"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_eaabaca787" y="140.321933"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_e64610c732" y="140.321933"/> </g> </g> <g id="PathCollection_4"> @@ -1439,592 +1439,592 @@ C -2.236584 -1.29895 -2.5 -0.663008 -2.5 0 C -2.5 0.663008 -2.236584 1.29895 -1.767767 1.767767 C -1.29895 2.236584 -0.663008 2.5 0 2.5 z -" id="C1_0_d7e70ab93f"/> +" id="C1_0_92e804cb43"/> </defs> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_d7e70ab93f" y="311.794598"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_92e804cb43" y="311.794598"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_d7e70ab93f" y="301.901588"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_92e804cb43" y="301.901588"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="262.252355" xlink:href="#C1_0_d7e70ab93f" y="297.797806"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="262.252355" xlink:href="#C1_0_92e804cb43" y="297.797806"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_d7e70ab93f" y="291.118288"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_92e804cb43" y="291.118288"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="259.701964" xlink:href="#C1_0_d7e70ab93f" y="288.888843"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="259.701964" xlink:href="#C1_0_92e804cb43" y="288.888843"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="265.532194" xlink:href="#C1_0_d7e70ab93f" y="286.820857"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="265.532194" xlink:href="#C1_0_92e804cb43" y="286.820857"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="262.5816" xlink:href="#C1_0_d7e70ab93f" y="283.153522"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="262.5816" xlink:href="#C1_0_92e804cb43" y="283.153522"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="266.73428" xlink:href="#C1_0_d7e70ab93f" y="280.440755"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="266.73428" xlink:href="#C1_0_92e804cb43" y="280.440755"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_d7e70ab93f" y="273.294834"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_92e804cb43" y="273.294834"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="258.99008" xlink:href="#C1_0_d7e70ab93f" y="273.270379"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="258.99008" xlink:href="#C1_0_92e804cb43" y="273.270379"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="267.497457" xlink:href="#C1_0_d7e70ab93f" y="269.817681"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="267.497457" xlink:href="#C1_0_92e804cb43" y="269.817681"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="272.747451" xlink:href="#C1_0_d7e70ab93f" y="269.811058"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="272.747451" xlink:href="#C1_0_92e804cb43" y="269.811058"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="255.743569" xlink:href="#C1_0_d7e70ab93f" y="269.785933"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="255.743569" xlink:href="#C1_0_92e804cb43" y="269.785933"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="250.493571" xlink:href="#C1_0_d7e70ab93f" y="269.782482"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="250.493571" xlink:href="#C1_0_92e804cb43" y="269.782482"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="277.997326" xlink:href="#C1_0_d7e70ab93f" y="269.780527"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="277.997326" xlink:href="#C1_0_92e804cb43" y="269.780527"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="245.243572" xlink:href="#C1_0_d7e70ab93f" y="269.779563"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="245.243572" xlink:href="#C1_0_92e804cb43" y="269.779563"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="263.194506" xlink:href="#C1_0_d7e70ab93f" y="267.27744"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="263.194506" xlink:href="#C1_0_92e804cb43" y="267.27744"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="281.434117" xlink:href="#C1_0_d7e70ab93f" y="266.428761"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="281.434117" xlink:href="#C1_0_92e804cb43" y="266.428761"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="241.828264" xlink:href="#C1_0_d7e70ab93f" y="266.412171"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="241.828264" xlink:href="#C1_0_92e804cb43" y="266.412171"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="258.406583" xlink:href="#C1_0_d7e70ab93f" y="265.964836"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="258.406583" xlink:href="#C1_0_92e804cb43" y="265.964836"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_d7e70ab93f" y="260.282911"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_92e804cb43" y="260.282911"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="258.990007" xlink:href="#C1_0_d7e70ab93f" y="260.275668"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="258.990007" xlink:href="#C1_0_92e804cb43" y="260.275668"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="269.489939" xlink:href="#C1_0_d7e70ab93f" y="260.26154"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="269.489939" xlink:href="#C1_0_92e804cb43" y="260.26154"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="253.752941" xlink:href="#C1_0_d7e70ab93f" y="259.964629"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="253.752941" xlink:href="#C1_0_92e804cb43" y="259.964629"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="272.38985" xlink:href="#C1_0_d7e70ab93f" y="256.565483"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="272.38985" xlink:href="#C1_0_92e804cb43" y="256.565483"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_d7e70ab93f" y="255.427417"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_92e804cb43" y="255.427417"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="258.990008" xlink:href="#C1_0_d7e70ab93f" y="255.419612"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="258.990008" xlink:href="#C1_0_92e804cb43" y="255.419612"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="253.740008" xlink:href="#C1_0_d7e70ab93f" y="255.41812"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="253.740008" xlink:href="#C1_0_92e804cb43" y="255.41812"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="277.459012" xlink:href="#C1_0_d7e70ab93f" y="255.411801"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="277.459012" xlink:href="#C1_0_92e804cb43" y="255.411801"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="248.490014" xlink:href="#C1_0_d7e70ab93f" y="255.411799"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="248.490014" xlink:href="#C1_0_92e804cb43" y="255.411799"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="282.709" xlink:href="#C1_0_d7e70ab93f" y="255.402063"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="282.709" xlink:href="#C1_0_92e804cb43" y="255.402063"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="243.240036" xlink:href="#C1_0_d7e70ab93f" y="255.398946"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="243.240036" xlink:href="#C1_0_92e804cb43" y="255.398946"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="287.958998" xlink:href="#C1_0_d7e70ab93f" y="255.39888"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="287.958998" xlink:href="#C1_0_92e804cb43" y="255.39888"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="237.990041" xlink:href="#C1_0_d7e70ab93f" y="255.392631"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="237.990041" xlink:href="#C1_0_92e804cb43" y="255.392631"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="293.208977" xlink:href="#C1_0_d7e70ab93f" y="255.386258"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="293.208977" xlink:href="#C1_0_92e804cb43" y="255.386258"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="232.835552" xlink:href="#C1_0_d7e70ab93f" y="254.550737"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="232.835552" xlink:href="#C1_0_92e804cb43" y="254.550737"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="298.35754" xlink:href="#C1_0_d7e70ab93f" y="254.518889"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="298.35754" xlink:href="#C1_0_92e804cb43" y="254.518889"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="267.290088" xlink:href="#C1_0_d7e70ab93f" y="251.818608"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="267.290088" xlink:href="#C1_0_92e804cb43" y="251.818608"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="272.540085" xlink:href="#C1_0_d7e70ab93f" y="251.814102"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="272.540085" xlink:href="#C1_0_92e804cb43" y="251.814102"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="228.707259" xlink:href="#C1_0_d7e70ab93f" y="251.811551"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="228.707259" xlink:href="#C1_0_92e804cb43" y="251.811551"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="302.512566" xlink:href="#C1_0_d7e70ab93f" y="251.808685"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="302.512566" xlink:href="#C1_0_92e804cb43" y="251.808685"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="223.457266" xlink:href="#C1_0_d7e70ab93f" y="251.804133"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="223.457266" xlink:href="#C1_0_92e804cb43" y="251.804133"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="262.253025" xlink:href="#C1_0_d7e70ab93f" y="250.568662"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="262.253025" xlink:href="#C1_0_92e804cb43" y="250.568662"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="257.003118" xlink:href="#C1_0_d7e70ab93f" y="250.542262"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="257.003118" xlink:href="#C1_0_92e804cb43" y="250.542262"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="251.902653" xlink:href="#C1_0_d7e70ab93f" y="249.491581"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="251.902653" xlink:href="#C1_0_92e804cb43" y="249.491581"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_d7e70ab93f" y="246.069025"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_92e804cb43" y="246.069025"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="259.914095" xlink:href="#C1_0_d7e70ab93f" y="243.556743"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="259.914095" xlink:href="#C1_0_92e804cb43" y="243.556743"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="268.53352" xlink:href="#C1_0_d7e70ab93f" y="243.517426"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="268.53352" xlink:href="#C1_0_92e804cb43" y="243.517426"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_d7e70ab93f" y="241.198577"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_92e804cb43" y="241.198577"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="272.999865" xlink:href="#C1_0_d7e70ab93f" y="241.187009"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="272.999865" xlink:href="#C1_0_92e804cb43" y="241.187009"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="255.482088" xlink:href="#C1_0_d7e70ab93f" y="241.180026"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="255.482088" xlink:href="#C1_0_92e804cb43" y="241.180026"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="250.232089" xlink:href="#C1_0_d7e70ab93f" y="241.177939"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="250.232089" xlink:href="#C1_0_92e804cb43" y="241.177939"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="278.249853" xlink:href="#C1_0_d7e70ab93f" y="241.177174"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="278.249853" xlink:href="#C1_0_92e804cb43" y="241.177174"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="245.017601" xlink:href="#C1_0_d7e70ab93f" y="240.663103"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="245.017601" xlink:href="#C1_0_92e804cb43" y="240.663103"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_d7e70ab93f" y="235.610721"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_92e804cb43" y="235.610721"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="258.995115" xlink:href="#C1_0_d7e70ab93f" y="235.415056"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="258.995115" xlink:href="#C1_0_92e804cb43" y="235.415056"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="269.44962" xlink:href="#C1_0_d7e70ab93f" y="235.061864"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="269.44962" xlink:href="#C1_0_92e804cb43" y="235.061864"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="274.69962" xlink:href="#C1_0_d7e70ab93f" y="235.061728"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="274.69962" xlink:href="#C1_0_92e804cb43" y="235.061728"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="254.476913" xlink:href="#C1_0_d7e70ab93f" y="233.157061"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="254.476913" xlink:href="#C1_0_92e804cb43" y="233.157061"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="249.226939" xlink:href="#C1_0_d7e70ab93f" y="233.143013"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="249.226939" xlink:href="#C1_0_92e804cb43" y="233.143013"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="279.430393" xlink:href="#C1_0_d7e70ab93f" y="233.139163"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="279.430393" xlink:href="#C1_0_92e804cb43" y="233.139163"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="244.044893" xlink:href="#C1_0_d7e70ab93f" y="232.431942"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="244.044893" xlink:href="#C1_0_92e804cb43" y="232.431942"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_d7e70ab93f" y="226.81512"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_92e804cb43" y="226.81512"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="258.990008" xlink:href="#C1_0_d7e70ab93f" y="226.80728"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="258.990008" xlink:href="#C1_0_92e804cb43" y="226.80728"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="269.48994" xlink:href="#C1_0_d7e70ab93f" y="226.793858"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="269.48994" xlink:href="#C1_0_92e804cb43" y="226.793858"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="253.740118" xlink:href="#C1_0_d7e70ab93f" y="226.778611"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="253.740118" xlink:href="#C1_0_92e804cb43" y="226.778611"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="274.739907" xlink:href="#C1_0_d7e70ab93f" y="226.778235"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="274.739907" xlink:href="#C1_0_92e804cb43" y="226.778235"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="248.49012" xlink:href="#C1_0_d7e70ab93f" y="226.775105"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="248.49012" xlink:href="#C1_0_92e804cb43" y="226.775105"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="279.989881" xlink:href="#C1_0_d7e70ab93f" y="226.764233"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="279.989881" xlink:href="#C1_0_92e804cb43" y="226.764233"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="244.028666" xlink:href="#C1_0_d7e70ab93f" y="224.438013"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="244.028666" xlink:href="#C1_0_92e804cb43" y="224.438013"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="284.454966" xlink:href="#C1_0_d7e70ab93f" y="224.432093"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="284.454966" xlink:href="#C1_0_92e804cb43" y="224.432093"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="238.778674" xlink:href="#C1_0_d7e70ab93f" y="224.430566"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="238.778674" xlink:href="#C1_0_92e804cb43" y="224.430566"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="289.704966" xlink:href="#C1_0_d7e70ab93f" y="224.430269"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="289.704966" xlink:href="#C1_0_92e804cb43" y="224.430269"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="233.528696" xlink:href="#C1_0_d7e70ab93f" y="224.417648"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="233.528696" xlink:href="#C1_0_92e804cb43" y="224.417648"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="294.954929" xlink:href="#C1_0_d7e70ab93f" y="224.413712"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="294.954929" xlink:href="#C1_0_92e804cb43" y="224.413712"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="228.278699" xlink:href="#C1_0_d7e70ab93f" y="224.412765"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="228.278699" xlink:href="#C1_0_92e804cb43" y="224.412765"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="300.204887" xlink:href="#C1_0_d7e70ab93f" y="224.396036"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="300.204887" xlink:href="#C1_0_92e804cb43" y="224.396036"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="223.22843" xlink:href="#C1_0_d7e70ab93f" y="223.201424"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="223.22843" xlink:href="#C1_0_92e804cb43" y="223.201424"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="305.25821" xlink:href="#C1_0_d7e70ab93f" y="223.19381"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="305.25821" xlink:href="#C1_0_92e804cb43" y="223.19381"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="220.4928" xlink:href="#C1_0_d7e70ab93f" y="223.189148"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="220.4928" xlink:href="#C1_0_92e804cb43" y="223.189148"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="263.526439" xlink:href="#C1_0_d7e70ab93f" y="222.422431"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="263.526439" xlink:href="#C1_0_92e804cb43" y="222.422431"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="268.631655" xlink:href="#C1_0_d7e70ab93f" y="222.419677"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="268.631655" xlink:href="#C1_0_92e804cb43" y="222.419677"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="258.375763" xlink:href="#C1_0_d7e70ab93f" y="221.564055"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="258.375763" xlink:href="#C1_0_92e804cb43" y="221.564055"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_d7e70ab93f" y="217.275239"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_92e804cb43" y="217.275239"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="259.004415" xlink:href="#C1_0_d7e70ab93f" y="216.946904"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="259.004415" xlink:href="#C1_0_92e804cb43" y="216.946904"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="269.27673" xlink:href="#C1_0_d7e70ab93f" y="216.024336"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="269.27673" xlink:href="#C1_0_92e804cb43" y="216.024336"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="254.023025" xlink:href="#C1_0_d7e70ab93f" y="215.546836"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="254.023025" xlink:href="#C1_0_92e804cb43" y="215.546836"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="274.359389" xlink:href="#C1_0_d7e70ab93f" y="214.913814"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="274.359389" xlink:href="#C1_0_92e804cb43" y="214.913814"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="249.116212" xlink:href="#C1_0_d7e70ab93f" y="213.970083"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="249.116212" xlink:href="#C1_0_92e804cb43" y="213.970083"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="279.488977" xlink:href="#C1_0_d7e70ab93f" y="213.969654"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="279.488977" xlink:href="#C1_0_92e804cb43" y="213.969654"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="243.86622" xlink:href="#C1_0_d7e70ab93f" y="213.961969"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="243.86622" xlink:href="#C1_0_92e804cb43" y="213.961969"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_d7e70ab93f" y="212.533043"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_92e804cb43" y="212.533043"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="258.51585" xlink:href="#C1_0_d7e70ab93f" y="212.532311"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="258.51585" xlink:href="#C1_0_92e804cb43" y="212.532311"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="268.602196" xlink:href="#C1_0_d7e70ab93f" y="210.065925"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="268.602196" xlink:href="#C1_0_92e804cb43" y="210.065925"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="273.852183" xlink:href="#C1_0_d7e70ab93f" y="210.056094"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="273.852183" xlink:href="#C1_0_92e804cb43" y="210.056094"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="254.163461" xlink:href="#C1_0_d7e70ab93f" y="210.05287"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="254.163461" xlink:href="#C1_0_92e804cb43" y="210.05287"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="246.688119" xlink:href="#C1_0_d7e70ab93f" y="210.038947"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="246.688119" xlink:href="#C1_0_92e804cb43" y="210.038947"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="281.913812" xlink:href="#C1_0_d7e70ab93f" y="210.037084"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="281.913812" xlink:href="#C1_0_92e804cb43" y="210.037084"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="241.438122" xlink:href="#C1_0_d7e70ab93f" y="210.034042"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="241.438122" xlink:href="#C1_0_92e804cb43" y="210.034042"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="287.163804" xlink:href="#C1_0_d7e70ab93f" y="210.029456"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="287.163804" xlink:href="#C1_0_92e804cb43" y="210.029456"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="236.188125" xlink:href="#C1_0_d7e70ab93f" y="210.029435"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="236.188125" xlink:href="#C1_0_92e804cb43" y="210.029435"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="292.413804" xlink:href="#C1_0_d7e70ab93f" y="210.029332"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="292.413804" xlink:href="#C1_0_92e804cb43" y="210.029332"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="230.938125" xlink:href="#C1_0_d7e70ab93f" y="210.029281"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="230.938125" xlink:href="#C1_0_92e804cb43" y="210.029281"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="297.663801" xlink:href="#C1_0_d7e70ab93f" y="210.024879"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="297.663801" xlink:href="#C1_0_92e804cb43" y="210.024879"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="225.688128" xlink:href="#C1_0_d7e70ab93f" y="210.024603"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="225.688128" xlink:href="#C1_0_92e804cb43" y="210.024603"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_d7e70ab93f" y="206.456861"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_92e804cb43" y="206.456861"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="258.990055" xlink:href="#C1_0_d7e70ab93f" y="206.43648"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="258.990055" xlink:href="#C1_0_92e804cb43" y="206.43648"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="269.268046" xlink:href="#C1_0_d7e70ab93f" y="205.181284"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="269.268046" xlink:href="#C1_0_92e804cb43" y="205.181284"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="254.257564" xlink:href="#C1_0_d7e70ab93f" y="204.516936"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="254.257564" xlink:href="#C1_0_92e804cb43" y="204.516936"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="274.456441" xlink:href="#C1_0_d7e70ab93f" y="204.50404"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="274.456441" xlink:href="#C1_0_92e804cb43" y="204.50404"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="249.203059" xlink:href="#C1_0_d7e70ab93f" y="203.318259"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="249.203059" xlink:href="#C1_0_92e804cb43" y="203.318259"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="279.514606" xlink:href="#C1_0_d7e70ab93f" y="203.316429"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="279.514606" xlink:href="#C1_0_92e804cb43" y="203.316429"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="243.964867" xlink:href="#C1_0_d7e70ab93f" y="203.021057"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="243.964867" xlink:href="#C1_0_92e804cb43" y="203.021057"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="284.752771" xlink:href="#C1_0_d7e70ab93f" y="203.018884"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="284.752771" xlink:href="#C1_0_92e804cb43" y="203.018884"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="238.72024" xlink:href="#C1_0_d7e70ab93f" y="202.820506"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="238.72024" xlink:href="#C1_0_92e804cb43" y="202.820506"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_d7e70ab93f" y="200.071485"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_92e804cb43" y="200.071485"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="269.489988" xlink:href="#C1_0_d7e70ab93f" y="200.061978"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="269.489988" xlink:href="#C1_0_92e804cb43" y="200.061978"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="259.156659" xlink:href="#C1_0_d7e70ab93f" y="198.963192"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="259.156659" xlink:href="#C1_0_92e804cb43" y="198.963192"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="274.243379" xlink:href="#C1_0_d7e70ab93f" y="198.179627"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="274.243379" xlink:href="#C1_0_92e804cb43" y="198.179627"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="253.989996" xlink:href="#C1_0_d7e70ab93f" y="198.17632"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="253.989996" xlink:href="#C1_0_92e804cb43" y="198.17632"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="279.493367" xlink:href="#C1_0_d7e70ab93f" y="198.170092"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="279.493367" xlink:href="#C1_0_92e804cb43" y="198.170092"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="248.740002" xlink:href="#C1_0_d7e70ab93f" y="198.170038"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="248.740002" xlink:href="#C1_0_92e804cb43" y="198.170038"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="284.743313" xlink:href="#C1_0_d7e70ab93f" y="198.14992"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="284.743313" xlink:href="#C1_0_92e804cb43" y="198.14992"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="243.490061" xlink:href="#C1_0_d7e70ab93f" y="198.148897"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="243.490061" xlink:href="#C1_0_92e804cb43" y="198.148897"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="289.993312" xlink:href="#C1_0_d7e70ab93f" y="198.148057"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="289.993312" xlink:href="#C1_0_92e804cb43" y="198.148057"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="238.24008" xlink:href="#C1_0_d7e70ab93f" y="198.137035"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="238.24008" xlink:href="#C1_0_92e804cb43" y="198.137035"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="295.243035" xlink:href="#C1_0_d7e70ab93f" y="198.102436"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="295.243035" xlink:href="#C1_0_92e804cb43" y="198.102436"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="262.756765" xlink:href="#C1_0_d7e70ab93f" y="195.818282"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="262.756765" xlink:href="#C1_0_92e804cb43" y="195.818282"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="268.006663" xlink:href="#C1_0_d7e70ab93f" y="195.790641"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="268.006663" xlink:href="#C1_0_92e804cb43" y="195.790641"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="233.786179" xlink:href="#C1_0_d7e70ab93f" y="195.789691"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="233.786179" xlink:href="#C1_0_92e804cb43" y="195.789691"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="299.71932" xlink:href="#C1_0_d7e70ab93f" y="195.785661"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="299.71932" xlink:href="#C1_0_92e804cb43" y="195.785661"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="257.725508" xlink:href="#C1_0_d7e70ab93f" y="194.551767"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="257.725508" xlink:href="#C1_0_92e804cb43" y="194.551767"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_d7e70ab93f" y="189.413314"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_92e804cb43" y="189.413314"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="259.018616" xlink:href="#C1_0_d7e70ab93f" y="188.951013"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="259.018616" xlink:href="#C1_0_92e804cb43" y="188.951013"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="269.459935" xlink:href="#C1_0_d7e70ab93f" y="188.939481"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="269.459935" xlink:href="#C1_0_92e804cb43" y="188.939481"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="274.709933" xlink:href="#C1_0_d7e70ab93f" y="188.936081"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="274.709933" xlink:href="#C1_0_92e804cb43" y="188.936081"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="253.949709" xlink:href="#C1_0_d7e70ab93f" y="187.796529"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="253.949709" xlink:href="#C1_0_92e804cb43" y="187.796529"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="279.777146" xlink:href="#C1_0_d7e70ab93f" y="187.776306"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="279.777146" xlink:href="#C1_0_92e804cb43" y="187.776306"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="248.800065" xlink:href="#C1_0_d7e70ab93f" y="186.933748"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="248.800065" xlink:href="#C1_0_92e804cb43" y="186.933748"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="284.928831" xlink:href="#C1_0_d7e70ab93f" y="186.922261"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="284.928831" xlink:href="#C1_0_92e804cb43" y="186.922261"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="243.609746" xlink:href="#C1_0_d7e70ab93f" y="186.2671"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="243.609746" xlink:href="#C1_0_92e804cb43" y="186.2671"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="262.16309" xlink:href="#C1_0_d7e70ab93f" y="185.341185"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="262.16309" xlink:href="#C1_0_92e804cb43" y="185.341185"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="267.213582" xlink:href="#C1_0_d7e70ab93f" y="184.130507"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="267.213582" xlink:href="#C1_0_92e804cb43" y="184.130507"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="257.194757" xlink:href="#C1_0_d7e70ab93f" y="183.908409"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="257.194757" xlink:href="#C1_0_92e804cb43" y="183.908409"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="272.456984" xlink:href="#C1_0_d7e70ab93f" y="183.908283"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="272.456984" xlink:href="#C1_0_92e804cb43" y="183.908283"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="288.777358" xlink:href="#C1_0_d7e70ab93f" y="183.906492"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="288.777358" xlink:href="#C1_0_92e804cb43" y="183.906492"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="252.059442" xlink:href="#C1_0_d7e70ab93f" y="182.986723"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="252.059442" xlink:href="#C1_0_92e804cb43" y="182.986723"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="277.588098" xlink:href="#C1_0_d7e70ab93f" y="182.970059"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="277.588098" xlink:href="#C1_0_92e804cb43" y="182.970059"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="247.148419" xlink:href="#C1_0_d7e70ab93f" y="181.419343"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="247.148419" xlink:href="#C1_0_92e804cb43" y="181.419343"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="282.504498" xlink:href="#C1_0_d7e70ab93f" y="181.414748"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="282.504498" xlink:href="#C1_0_92e804cb43" y="181.414748"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="241.898557" xlink:href="#C1_0_d7e70ab93f" y="181.3872"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="241.898557" xlink:href="#C1_0_92e804cb43" y="181.3872"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="236.648563" xlink:href="#C1_0_d7e70ab93f" y="181.380799"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="236.648563" xlink:href="#C1_0_92e804cb43" y="181.380799"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_d7e70ab93f" y="179.043565"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_92e804cb43" y="179.043565"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="258.990003" xlink:href="#C1_0_d7e70ab93f" y="179.038821"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="258.990003" xlink:href="#C1_0_92e804cb43" y="179.038821"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="267.925797" xlink:href="#C1_0_d7e70ab93f" y="175.886138"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="267.925797" xlink:href="#C1_0_92e804cb43" y="175.886138"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="261.97366" xlink:href="#C1_0_d7e70ab93f" y="175.044137"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="261.97366" xlink:href="#C1_0_92e804cb43" y="175.044137"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="256.723661" xlink:href="#C1_0_d7e70ab93f" y="175.040548"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="256.723661" xlink:href="#C1_0_92e804cb43" y="175.040548"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="273.078775" xlink:href="#C1_0_d7e70ab93f" y="175.037672"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="273.078775" xlink:href="#C1_0_92e804cb43" y="175.037672"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="251.473694" xlink:href="#C1_0_d7e70ab93f" y="175.024813"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="251.473694" xlink:href="#C1_0_92e804cb43" y="175.024813"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="278.328506" xlink:href="#C1_0_d7e70ab93f" y="174.992799"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="278.328506" xlink:href="#C1_0_92e804cb43" y="174.992799"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="246.237125" xlink:href="#C1_0_d7e70ab93f" y="174.707864"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="246.237125" xlink:href="#C1_0_92e804cb43" y="174.707864"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="283.566852" xlink:href="#C1_0_d7e70ab93f" y="174.69753"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="283.566852" xlink:href="#C1_0_92e804cb43" y="174.69753"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="240.987151" xlink:href="#C1_0_d7e70ab93f" y="174.693904"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="240.987151" xlink:href="#C1_0_92e804cb43" y="174.693904"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="288.816849" xlink:href="#C1_0_d7e70ab93f" y="174.692786"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="288.816849" xlink:href="#C1_0_92e804cb43" y="174.692786"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="236.634487" xlink:href="#C1_0_d7e70ab93f" y="172.214809"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="236.634487" xlink:href="#C1_0_92e804cb43" y="172.214809"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="293.15579" xlink:href="#C1_0_d7e70ab93f" y="172.19659"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="293.15579" xlink:href="#C1_0_92e804cb43" y="172.19659"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_d7e70ab93f" y="170.758267"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_92e804cb43" y="170.758267"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="259.187695" xlink:href="#C1_0_d7e70ab93f" y="169.552995"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="259.187695" xlink:href="#C1_0_92e804cb43" y="169.552995"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="269.284939" xlink:href="#C1_0_d7e70ab93f" y="169.531185"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="269.284939" xlink:href="#C1_0_92e804cb43" y="169.531185"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="254.763381" xlink:href="#C1_0_d7e70ab93f" y="167.166077"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="254.763381" xlink:href="#C1_0_92e804cb43" y="167.166077"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="273.290803" xlink:href="#C1_0_d7e70ab93f" y="166.665277"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="273.290803" xlink:href="#C1_0_92e804cb43" y="166.665277"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_d7e70ab93f" y="164.681466"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_92e804cb43" y="164.681466"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="259.101496" xlink:href="#C1_0_d7e70ab93f" y="164.668857"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="259.101496" xlink:href="#C1_0_92e804cb43" y="164.668857"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="277.969594" xlink:href="#C1_0_d7e70ab93f" y="164.654027"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="277.969594" xlink:href="#C1_0_92e804cb43" y="164.654027"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="269.281149" xlink:href="#C1_0_d7e70ab93f" y="163.443325"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="269.281149" xlink:href="#C1_0_92e804cb43" y="163.443325"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="263.189367" xlink:href="#C1_0_d7e70ab93f" y="160.337324"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="263.189367" xlink:href="#C1_0_92e804cb43" y="160.337324"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="258.151866" xlink:href="#C1_0_d7e70ab93f" y="160.308161"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="258.151866" xlink:href="#C1_0_92e804cb43" y="160.308161"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="270.596639" xlink:href="#C1_0_d7e70ab93f" y="159.150938"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="270.596639" xlink:href="#C1_0_92e804cb43" y="159.150938"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="253.085036" xlink:href="#C1_0_d7e70ab93f" y="159.147192"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="253.085036" xlink:href="#C1_0_92e804cb43" y="159.147192"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="265.620942" xlink:href="#C1_0_d7e70ab93f" y="156.407724"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="265.620942" xlink:href="#C1_0_92e804cb43" y="156.407724"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="260.59086" xlink:href="#C1_0_d7e70ab93f" y="156.381842"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="260.59086" xlink:href="#C1_0_92e804cb43" y="156.381842"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="274.676997" xlink:href="#C1_0_d7e70ab93f" y="156.360992"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="274.676997" xlink:href="#C1_0_92e804cb43" y="156.360992"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="268.641423" xlink:href="#C1_0_d7e70ab93f" y="152.781196"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="268.641423" xlink:href="#C1_0_92e804cb43" y="152.781196"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="257.54481" xlink:href="#C1_0_d7e70ab93f" y="152.7706"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="257.54481" xlink:href="#C1_0_92e804cb43" y="152.7706"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_d7e70ab93f" y="149.465622"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_92e804cb43" y="149.465622"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="259.649718" xlink:href="#C1_0_d7e70ab93f" y="147.313817"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="259.649718" xlink:href="#C1_0_92e804cb43" y="147.313817"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="268.826854" xlink:href="#C1_0_d7e70ab93f" y="147.308609"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="268.826854" xlink:href="#C1_0_92e804cb43" y="147.308609"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="274.076634" xlink:href="#C1_0_d7e70ab93f" y="147.268016"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="274.076634" xlink:href="#C1_0_92e804cb43" y="147.268016"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_d7e70ab93f" y="143.569627"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_92e804cb43" y="143.569627"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="259.163522" xlink:href="#C1_0_d7e70ab93f" y="142.439121"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="259.163522" xlink:href="#C1_0_92e804cb43" y="142.439121"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="268.438812" xlink:href="#C1_0_d7e70ab93f" y="140.90799"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="268.438812" xlink:href="#C1_0_92e804cb43" y="140.90799"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_d7e70ab93f" y="127.733366"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_92e804cb43" y="127.733366"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="261.113693" xlink:href="#C1_0_d7e70ab93f" y="124.171381"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="261.113693" xlink:href="#C1_0_92e804cb43" y="124.171381"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_d7e70ab93f" y="116.087931"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_92e804cb43" y="116.087931"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_d7e70ab93f" y="101.918938"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_92e804cb43" y="101.918938"/> </g> </g> <g id="PathCollection_5"> @@ -2039,277 +2039,277 @@ C -2.236584 -1.29895 -2.5 -0.663008 -2.5 0 C -2.5 0.663008 -2.236584 1.29895 -1.767767 1.767767 C -1.29895 2.236584 -0.663008 2.5 0 2.5 z -" id="C2_0_e0b3868871"/> +" id="C2_0_f2b1f24b3b"/> </defs> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_e0b3868871" y="318.553243"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_f2b1f24b3b" y="318.553243"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_e0b3868871" y="308.206677"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_f2b1f24b3b" y="308.206677"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="393.837939" xlink:href="#C2_0_e0b3868871" y="305.689681"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="393.837939" xlink:href="#C2_0_f2b1f24b3b" y="305.689681"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="401.57735" xlink:href="#C2_0_e0b3868871" y="304.840763"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="401.57735" xlink:href="#C2_0_f2b1f24b3b" y="304.840763"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_e0b3868871" y="293.851516"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_f2b1f24b3b" y="293.851516"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="392.910014" xlink:href="#C2_0_e0b3868871" y="293.841403"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="392.910014" xlink:href="#C2_0_f2b1f24b3b" y="293.841403"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="403.409756" xlink:href="#C2_0_e0b3868871" y="293.808801"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="403.409756" xlink:href="#C2_0_f2b1f24b3b" y="293.808801"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="389.848692" xlink:href="#C2_0_e0b3868871" y="290.239384"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="389.848692" xlink:href="#C2_0_f2b1f24b3b" y="290.239384"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_e0b3868871" y="271.595185"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_f2b1f24b3b" y="271.595185"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="392.910235" xlink:href="#C2_0_e0b3868871" y="271.553233"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="392.910235" xlink:href="#C2_0_f2b1f24b3b" y="271.553233"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="403.409764" xlink:href="#C2_0_e0b3868871" y="271.553172"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="403.409764" xlink:href="#C2_0_f2b1f24b3b" y="271.553172"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_e0b3868871" y="265.228305"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_f2b1f24b3b" y="265.228305"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="392.910066" xlink:href="#C2_0_e0b3868871" y="265.206087"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="392.910066" xlink:href="#C2_0_f2b1f24b3b" y="265.206087"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="403.409803" xlink:href="#C2_0_e0b3868871" y="265.189848"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="403.409803" xlink:href="#C2_0_f2b1f24b3b" y="265.189848"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="408.659802" xlink:href="#C2_0_e0b3868871" y="265.187392"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="408.659802" xlink:href="#C2_0_f2b1f24b3b" y="265.187392"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="387.672579" xlink:href="#C2_0_e0b3868871" y="264.900143"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="387.672579" xlink:href="#C2_0_f2b1f24b3b" y="264.900143"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="413.126275" xlink:href="#C2_0_e0b3868871" y="262.85715"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="413.126275" xlink:href="#C2_0_f2b1f24b3b" y="262.85715"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="383.025001" xlink:href="#C2_0_e0b3868871" y="262.837923"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="383.025001" xlink:href="#C2_0_f2b1f24b3b" y="262.837923"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="418.36266" xlink:href="#C2_0_e0b3868871" y="262.538035"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="418.36266" xlink:href="#C2_0_f2b1f24b3b" y="262.538035"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="377.787033" xlink:href="#C2_0_e0b3868871" y="262.537915"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="377.787033" xlink:href="#C2_0_f2b1f24b3b" y="262.537915"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="423.499105" xlink:href="#C2_0_e0b3868871" y="261.62085"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="423.499105" xlink:href="#C2_0_f2b1f24b3b" y="261.62085"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_e0b3868871" y="253.305883"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_f2b1f24b3b" y="253.305883"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="393.015157" xlink:href="#C2_0_e0b3868871" y="252.422909"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="393.015157" xlink:href="#C2_0_f2b1f24b3b" y="252.422909"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="403.084825" xlink:href="#C2_0_e0b3868871" y="251.769702"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="403.084825" xlink:href="#C2_0_f2b1f24b3b" y="251.769702"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="388.19808" xlink:href="#C2_0_e0b3868871" y="250.659809"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="388.19808" xlink:href="#C2_0_f2b1f24b3b" y="250.659809"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_e0b3868871" y="248.469722"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_f2b1f24b3b" y="248.469722"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="406.131089" xlink:href="#C2_0_e0b3868871" y="248.158589"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="406.131089" xlink:href="#C2_0_f2b1f24b3b" y="248.158589"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_e0b3868871" y="241.773311"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_f2b1f24b3b" y="241.773311"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="393.852899" xlink:href="#C2_0_e0b3868871" y="239.23809"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="393.852899" xlink:href="#C2_0_f2b1f24b3b" y="239.23809"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_e0b3868871" y="236.593294"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_f2b1f24b3b" y="236.593294"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="403.40999" xlink:href="#C2_0_e0b3868871" y="236.584609"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="403.40999" xlink:href="#C2_0_f2b1f24b3b" y="236.584609"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="393.604555" xlink:href="#C2_0_e0b3868871" y="234.389318"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="393.604555" xlink:href="#C2_0_f2b1f24b3b" y="234.389318"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="407.602358" xlink:href="#C2_0_e0b3868871" y="233.915738"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="407.602358" xlink:href="#C2_0_f2b1f24b3b" y="233.915738"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_e0b3868871" y="231.715173"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_f2b1f24b3b" y="231.715173"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="403.223143" xlink:href="#C2_0_e0b3868871" y="230.542788"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="403.223143" xlink:href="#C2_0_f2b1f24b3b" y="230.542788"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="396.367135" xlink:href="#C2_0_e0b3868871" y="227.547891"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="396.367135" xlink:href="#C2_0_f2b1f24b3b" y="227.547891"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_e0b3868871" y="222.038458"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_f2b1f24b3b" y="222.038458"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="393.593566" xlink:href="#C2_0_e0b3868871" y="219.850762"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="393.593566" xlink:href="#C2_0_f2b1f24b3b" y="219.850762"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="402.72461" xlink:href="#C2_0_e0b3868871" y="219.848048"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="402.72461" xlink:href="#C2_0_f2b1f24b3b" y="219.848048"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="407.974561" xlink:href="#C2_0_e0b3868871" y="219.8288"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="407.974561" xlink:href="#C2_0_f2b1f24b3b" y="219.8288"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="389.003287" xlink:href="#C2_0_e0b3868871" y="217.698952"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="389.003287" xlink:href="#C2_0_f2b1f24b3b" y="217.698952"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_e0b3868871" y="216.600018"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_f2b1f24b3b" y="216.600018"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="403.744095" xlink:href="#C2_0_e0b3868871" y="215.498616"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="403.744095" xlink:href="#C2_0_f2b1f24b3b" y="215.498616"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="393.668006" xlink:href="#C2_0_e0b3868871" y="214.305031"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="393.668006" xlink:href="#C2_0_f2b1f24b3b" y="214.305031"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="388.72135" xlink:href="#C2_0_e0b3868871" y="212.819723"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="388.72135" xlink:href="#C2_0_f2b1f24b3b" y="212.819723"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="407.926716" xlink:href="#C2_0_e0b3868871" y="212.818859"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="407.926716" xlink:href="#C2_0_f2b1f24b3b" y="212.818859"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="383.472614" xlink:href="#C2_0_e0b3868871" y="212.722431"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="383.472614" xlink:href="#C2_0_f2b1f24b3b" y="212.722431"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_e0b3868871" y="210.012585"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_f2b1f24b3b" y="210.012585"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="403.384385" xlink:href="#C2_0_e0b3868871" y="209.575135"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="403.384385" xlink:href="#C2_0_f2b1f24b3b" y="209.575135"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="394.27286" xlink:href="#C2_0_e0b3868871" y="207.032348"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="394.27286" xlink:href="#C2_0_f2b1f24b3b" y="207.032348"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="407.039123" xlink:href="#C2_0_e0b3868871" y="206.392062"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="407.039123" xlink:href="#C2_0_f2b1f24b3b" y="206.392062"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_e0b3868871" y="202.28269"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_f2b1f24b3b" y="202.28269"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="392.972654" xlink:href="#C2_0_e0b3868871" y="201.599742"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="392.972654" xlink:href="#C2_0_f2b1f24b3b" y="201.599742"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="402.243918" xlink:href="#C2_0_e0b3868871" y="199.496462"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="402.243918" xlink:href="#C2_0_f2b1f24b3b" y="199.496462"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="407.485358" xlink:href="#C2_0_e0b3868871" y="199.243376"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="407.485358" xlink:href="#C2_0_f2b1f24b3b" y="199.243376"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="397.764241" xlink:href="#C2_0_e0b3868871" y="197.184366"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="397.764241" xlink:href="#C2_0_f2b1f24b3b" y="197.184366"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="392.601705" xlink:href="#C2_0_e0b3868871" y="196.378407"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="392.601705" xlink:href="#C2_0_f2b1f24b3b" y="196.378407"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="403.164253" xlink:href="#C2_0_e0b3868871" y="195.131288"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="403.164253" xlink:href="#C2_0_f2b1f24b3b" y="195.131288"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="408.398208" xlink:href="#C2_0_e0b3868871" y="194.784913"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="408.398208" xlink:href="#C2_0_f2b1f24b3b" y="194.784913"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="388.392795" xlink:href="#C2_0_e0b3868871" y="193.72817"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="388.392795" xlink:href="#C2_0_f2b1f24b3b" y="193.72817"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="383.188077" xlink:href="#C2_0_e0b3868871" y="193.14709"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="383.188077" xlink:href="#C2_0_f2b1f24b3b" y="193.14709"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_e0b3868871" y="191.204515"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_f2b1f24b3b" y="191.204515"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="395.06723" xlink:href="#C2_0_e0b3868871" y="187.621708"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="395.06723" xlink:href="#C2_0_f2b1f24b3b" y="187.621708"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="400.163611" xlink:href="#C2_0_e0b3868871" y="186.556984"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="400.163611" xlink:href="#C2_0_f2b1f24b3b" y="186.556984"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="405.267767" xlink:href="#C2_0_e0b3868871" y="185.519166"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="405.267767" xlink:href="#C2_0_f2b1f24b3b" y="185.519166"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="391.306287" xlink:href="#C2_0_e0b3868871" y="184.528132"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="391.306287" xlink:href="#C2_0_f2b1f24b3b" y="184.528132"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="386.056289" xlink:href="#C2_0_e0b3868871" y="184.52442"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="386.056289" xlink:href="#C2_0_f2b1f24b3b" y="184.52442"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="410.287316" xlink:href="#C2_0_e0b3868871" y="184.21994"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="410.287316" xlink:href="#C2_0_f2b1f24b3b" y="184.21994"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="380.841427" xlink:href="#C2_0_e0b3868871" y="184.012297"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="380.841427" xlink:href="#C2_0_f2b1f24b3b" y="184.012297"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_e0b3868871" y="180.811369"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_f2b1f24b3b" y="180.811369"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="395.481505" xlink:href="#C2_0_e0b3868871" y="176.997998"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="395.481505" xlink:href="#C2_0_f2b1f24b3b" y="176.997998"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="400.726839" xlink:href="#C2_0_e0b3868871" y="176.811114"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="400.726839" xlink:href="#C2_0_f2b1f24b3b" y="176.811114"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="405.964536" xlink:href="#C2_0_e0b3868871" y="176.507744"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="405.964536" xlink:href="#C2_0_f2b1f24b3b" y="176.507744"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="390.641179" xlink:href="#C2_0_e0b3868871" y="175.280913"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="390.641179" xlink:href="#C2_0_f2b1f24b3b" y="175.280913"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="409.664307" xlink:href="#C2_0_e0b3868871" y="173.361997"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="409.664307" xlink:href="#C2_0_f2b1f24b3b" y="173.361997"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_e0b3868871" y="172.229668"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_f2b1f24b3b" y="172.229668"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="403.409983" xlink:href="#C2_0_e0b3868871" y="172.218341"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="403.409983" xlink:href="#C2_0_f2b1f24b3b" y="172.218341"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="386.969213" xlink:href="#C2_0_e0b3868871" y="172.112013"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="386.969213" xlink:href="#C2_0_f2b1f24b3b" y="172.112013"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="381.723573" xlink:href="#C2_0_e0b3868871" y="171.931355"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="381.723573" xlink:href="#C2_0_f2b1f24b3b" y="171.931355"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="414.632729" xlink:href="#C2_0_e0b3868871" y="171.929442"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="414.632729" xlink:href="#C2_0_f2b1f24b3b" y="171.929442"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="393.268716" xlink:href="#C2_0_e0b3868871" y="170.618866"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="393.268716" xlink:href="#C2_0_f2b1f24b3b" y="170.618866"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="376.927678" xlink:href="#C2_0_e0b3868871" y="170.127538"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="376.927678" xlink:href="#C2_0_f2b1f24b3b" y="170.127538"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="400.090722" xlink:href="#C2_0_e0b3868871" y="168.106549"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="400.090722" xlink:href="#C2_0_f2b1f24b3b" y="168.106549"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_e0b3868871" y="162.570934"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_f2b1f24b3b" y="162.570934"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="393.291767" xlink:href="#C2_0_e0b3868871" y="160.911074"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="393.291767" xlink:href="#C2_0_f2b1f24b3b" y="160.911074"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="401.694271" xlink:href="#C2_0_e0b3868871" y="159.292268"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="401.694271" xlink:href="#C2_0_f2b1f24b3b" y="159.292268"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_e0b3868871" y="153.37535"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_f2b1f24b3b" y="153.37535"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="392.922518" xlink:href="#C2_0_e0b3868871" y="153.069347"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="392.922518" xlink:href="#C2_0_f2b1f24b3b" y="153.069347"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_e0b3868871" y="147.554881"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_f2b1f24b3b" y="147.554881"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="394.107509" xlink:href="#C2_0_e0b3868871" y="144.736114"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="394.107509" xlink:href="#C2_0_f2b1f24b3b" y="144.736114"/> </g> </g> <g id="PathCollection_6"> @@ -2324,259 +2324,259 @@ C -2.236584 -1.29895 -2.5 -0.663008 -2.5 0 C -2.5 0.663008 -2.236584 1.29895 -1.767767 1.767767 C -1.29895 2.236584 -0.663008 2.5 0 2.5 z -" id="C3_0_acbc77e91b"/> +" id="C3_0_761689fab7"/> </defs> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_acbc77e91b" y="280.449976"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_761689fab7" y="280.449976"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_acbc77e91b" y="275.459547"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_761689fab7" y="275.459547"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_acbc77e91b" y="269.772241"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_761689fab7" y="269.772241"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="482.192459" xlink:href="#C3_0_acbc77e91b" y="269.636549"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="482.192459" xlink:href="#C3_0_761689fab7" y="269.636549"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_acbc77e91b" y="260.330797"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_761689fab7" y="260.330797"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_acbc77e91b" y="255.45697"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_761689fab7" y="255.45697"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="482.190509" xlink:href="#C3_0_acbc77e91b" y="255.395219"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="482.190509" xlink:href="#C3_0_761689fab7" y="255.395219"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="490.412756" xlink:href="#C3_0_acbc77e91b" y="251.802423"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="490.412756" xlink:href="#C3_0_761689fab7" y="251.802423"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="495.662744" xlink:href="#C3_0_acbc77e91b" y="251.792999"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="495.662744" xlink:href="#C3_0_761689fab7" y="251.792999"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_acbc77e91b" y="241.007548"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_761689fab7" y="241.007548"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_acbc77e91b" y="235.072473"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_761689fab7" y="235.072473"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="482.678536" xlink:href="#C3_0_acbc77e91b" y="233.20473"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="482.678536" xlink:href="#C3_0_761689fab7" y="233.20473"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="490.529516" xlink:href="#C3_0_acbc77e91b" y="231.487664"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="490.529516" xlink:href="#C3_0_761689fab7" y="231.487664"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_acbc77e91b" y="226.845842"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_761689fab7" y="226.845842"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="482.196559" xlink:href="#C3_0_acbc77e91b" y="226.624271"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="482.196559" xlink:href="#C3_0_761689fab7" y="226.624271"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="491.879044" xlink:href="#C3_0_acbc77e91b" y="224.478511"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="491.879044" xlink:href="#C3_0_761689fab7" y="224.478511"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="497.128268" xlink:href="#C3_0_acbc77e91b" y="224.402285"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="497.128268" xlink:href="#C3_0_761689fab7" y="224.402285"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="477.757705" xlink:href="#C3_0_acbc77e91b" y="224.256686"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="477.757705" xlink:href="#C3_0_761689fab7" y="224.256686"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_acbc77e91b" y="221.604922"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_761689fab7" y="221.604922"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_acbc77e91b" y="216.539783"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_761689fab7" y="216.539783"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="482.317312" xlink:href="#C3_0_acbc77e91b" y="215.569272"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="482.317312" xlink:href="#C3_0_761689fab7" y="215.569272"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="489.790658" xlink:href="#C3_0_acbc77e91b" y="212.57522"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="489.790658" xlink:href="#C3_0_761689fab7" y="212.57522"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="495.035982" xlink:href="#C3_0_acbc77e91b" y="212.388117"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="495.035982" xlink:href="#C3_0_761689fab7" y="212.388117"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="478.665687" xlink:href="#C3_0_acbc77e91b" y="212.383652"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="478.665687" xlink:href="#C3_0_761689fab7" y="212.383652"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="484.789587" xlink:href="#C3_0_acbc77e91b" y="211.226127"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="484.789587" xlink:href="#C3_0_761689fab7" y="211.226127"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="499.480697" xlink:href="#C3_0_acbc77e91b" y="210.028388"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="499.480697" xlink:href="#C3_0_761689fab7" y="210.028388"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="474.224512" xlink:href="#C3_0_acbc77e91b" y="210.019172"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="474.224512" xlink:href="#C3_0_761689fab7" y="210.019172"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="504.727517" xlink:href="#C3_0_acbc77e91b" y="209.874073"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="504.727517" xlink:href="#C3_0_761689fab7" y="209.874073"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_acbc77e91b" y="205.316997"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_761689fab7" y="205.316997"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="482.19088" xlink:href="#C3_0_acbc77e91b" y="205.235818"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="482.19088" xlink:href="#C3_0_761689fab7" y="205.235818"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="492.617134" xlink:href="#C3_0_acbc77e91b" y="204.580849"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="492.617134" xlink:href="#C3_0_761689fab7" y="204.580849"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="477.433755" xlink:href="#C3_0_acbc77e91b" y="203.360206"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="477.433755" xlink:href="#C3_0_761689fab7" y="203.360206"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="497.527712" xlink:href="#C3_0_acbc77e91b" y="203.012476"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="497.527712" xlink:href="#C3_0_761689fab7" y="203.012476"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="472.200882" xlink:href="#C3_0_acbc77e91b" y="203.002364"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="472.200882" xlink:href="#C3_0_761689fab7" y="203.002364"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_acbc77e91b" y="200.855936"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_761689fab7" y="200.855936"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="492.604299" xlink:href="#C3_0_acbc77e91b" y="200.058073"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="492.604299" xlink:href="#C3_0_761689fab7" y="200.058073"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="483.265011" xlink:href="#C3_0_acbc77e91b" y="198.167702"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="483.265011" xlink:href="#C3_0_761689fab7" y="198.167702"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="478.015081" xlink:href="#C3_0_acbc77e91b" y="198.144923"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="478.015081" xlink:href="#C3_0_761689fab7" y="198.144923"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="497.337152" xlink:href="#C3_0_acbc77e91b" y="198.139164"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="497.337152" xlink:href="#C3_0_761689fab7" y="198.139164"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="489.274817" xlink:href="#C3_0_acbc77e91b" y="196.701698"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="489.274817" xlink:href="#C3_0_761689fab7" y="196.701698"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="473.577138" xlink:href="#C3_0_acbc77e91b" y="195.776119"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="473.577138" xlink:href="#C3_0_761689fab7" y="195.776119"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="501.677634" xlink:href="#C3_0_acbc77e91b" y="195.644881"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="501.677634" xlink:href="#C3_0_761689fab7" y="195.644881"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_acbc77e91b" y="189.721571"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_761689fab7" y="189.721571"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="482.19001" xlink:href="#C3_0_acbc77e91b" y="189.713094"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="482.19001" xlink:href="#C3_0_761689fab7" y="189.713094"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="492.677182" xlink:href="#C3_0_acbc77e91b" y="189.411929"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="492.677182" xlink:href="#C3_0_761689fab7" y="189.411929"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="477.008309" xlink:href="#C3_0_acbc77e91b" y="189.000231"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="477.008309" xlink:href="#C3_0_761689fab7" y="189.000231"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="496.621844" xlink:href="#C3_0_acbc77e91b" y="186.486097"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="496.621844" xlink:href="#C3_0_761689fab7" y="186.486097"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_acbc77e91b" y="185.186955"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_761689fab7" y="185.186955"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="492.325384" xlink:href="#C3_0_acbc77e91b" y="183.938031"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="492.325384" xlink:href="#C3_0_761689fab7" y="183.938031"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="482.414748" xlink:href="#C3_0_acbc77e91b" y="183.903549"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="482.414748" xlink:href="#C3_0_761689fab7" y="183.903549"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="477.164752" xlink:href="#C3_0_acbc77e91b" y="183.897578"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="477.164752" xlink:href="#C3_0_761689fab7" y="183.897578"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="500.766196" xlink:href="#C3_0_acbc77e91b" y="183.764262"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="500.766196" xlink:href="#C3_0_761689fab7" y="183.764262"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="471.917198" xlink:href="#C3_0_acbc77e91b" y="183.762274"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="471.917198" xlink:href="#C3_0_761689fab7" y="183.762274"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="505.259488" xlink:href="#C3_0_acbc77e91b" y="181.471088"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="505.259488" xlink:href="#C3_0_761689fab7" y="181.471088"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="467.475677" xlink:href="#C3_0_acbc77e91b" y="181.398257"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="467.475677" xlink:href="#C3_0_761689fab7" y="181.398257"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_acbc77e91b" y="177.877542"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_761689fab7" y="177.877542"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="482.190076" xlink:href="#C3_0_acbc77e91b" y="177.853726"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="482.190076" xlink:href="#C3_0_761689fab7" y="177.853726"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="492.596932" xlink:href="#C3_0_acbc77e91b" y="177.046387"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="492.596932" xlink:href="#C3_0_761689fab7" y="177.046387"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="477.486685" xlink:href="#C3_0_acbc77e91b" y="175.883828"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="477.486685" xlink:href="#C3_0_761689fab7" y="175.883828"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="497.317281" xlink:href="#C3_0_acbc77e91b" y="175.105635"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="497.317281" xlink:href="#C3_0_761689fab7" y="175.105635"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_acbc77e91b" y="170.79334"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_761689fab7" y="170.79334"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="482.396269" xlink:href="#C3_0_acbc77e91b" y="169.562723"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="482.396269" xlink:href="#C3_0_761689fab7" y="169.562723"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="492.011179" xlink:href="#C3_0_acbc77e91b" y="168.612723"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="492.011179" xlink:href="#C3_0_761689fab7" y="168.612723"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="478.098653" xlink:href="#C3_0_acbc77e91b" y="167.016047"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="478.098653" xlink:href="#C3_0_761689fab7" y="167.016047"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_acbc77e91b" y="165.774909"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_761689fab7" y="165.774909"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="482.556218" xlink:href="#C3_0_acbc77e91b" y="164.673669"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="482.556218" xlink:href="#C3_0_761689fab7" y="164.673669"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_acbc77e91b" y="161.234091"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_761689fab7" y="161.234091"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="492.591086" xlink:href="#C3_0_acbc77e91b" y="160.377472"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="492.591086" xlink:href="#C3_0_761689fab7" y="160.377472"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="482.346137" xlink:href="#C3_0_acbc77e91b" y="160.160809"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="482.346137" xlink:href="#C3_0_761689fab7" y="160.160809"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="497.632716" xlink:href="#C3_0_acbc77e91b" y="159.140727"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="497.632716" xlink:href="#C3_0_761689fab7" y="159.140727"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="477.764945" xlink:href="#C3_0_acbc77e91b" y="157.99523"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="477.764945" xlink:href="#C3_0_761689fab7" y="157.99523"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_acbc77e91b" y="156.391807"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_761689fab7" y="156.391807"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="501.736161" xlink:href="#C3_0_acbc77e91b" y="156.375038"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="501.736161" xlink:href="#C3_0_761689fab7" y="156.375038"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="472.949087" xlink:href="#C3_0_acbc77e91b" y="156.229756"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="472.949087" xlink:href="#C3_0_761689fab7" y="156.229756"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="484.398426" xlink:href="#C3_0_acbc77e91b" y="152.777876"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="484.398426" xlink:href="#C3_0_761689fab7" y="152.777876"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="490.220912" xlink:href="#C3_0_acbc77e91b" y="152.631089"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="490.220912" xlink:href="#C3_0_761689fab7" y="152.631089"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="493.904978" xlink:href="#C3_0_acbc77e91b" y="149.472223"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="493.904978" xlink:href="#C3_0_761689fab7" y="149.472223"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_acbc77e91b" y="147.347022"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_761689fab7" y="147.347022"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="482.197045" xlink:href="#C3_0_acbc77e91b" y="147.117396"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="482.197045" xlink:href="#C3_0_761689fab7" y="147.117396"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="497.257321" xlink:href="#C3_0_acbc77e91b" y="146.060001"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="497.257321" xlink:href="#C3_0_761689fab7" y="146.060001"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_acbc77e91b" y="140.975908"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_761689fab7" y="140.975908"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="482.190784" xlink:href="#C3_0_acbc77e91b" y="140.899286"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="482.190784" xlink:href="#C3_0_761689fab7" y="140.899286"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="492.68895" xlink:href="#C3_0_acbc77e91b" y="140.887246"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="492.68895" xlink:href="#C3_0_761689fab7" y="140.887246"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="477.01552" xlink:href="#C3_0_acbc77e91b" y="140.153818"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="477.01552" xlink:href="#C3_0_761689fab7" y="140.153818"/> </g> </g> <g id="PathCollection_7"> @@ -2591,103 +2591,103 @@ C -2.236584 -1.29895 -2.5 -0.663008 -2.5 0 C -2.5 0.663008 -2.236584 1.29895 -1.767767 1.767767 C -1.29895 2.236584 -0.663008 2.5 0 2.5 z -" id="C4_0_07995dbc83"/> +" id="C4_0_d190ea9174"/> </defs> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_07995dbc83" y="340.777665"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_d190ea9174" y="340.777665"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_07995dbc83" y="334.732129"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_d190ea9174" y="334.732129"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_07995dbc83" y="308.670776"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_d190ea9174" y="308.670776"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_07995dbc83" y="294.28087"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_d190ea9174" y="294.28087"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_07995dbc83" y="280.039095"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_d190ea9174" y="280.039095"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="616.110019" xlink:href="#C4_0_07995dbc83" y="280.027168"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="616.110019" xlink:href="#C4_0_d190ea9174" y="280.027168"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="626.609957" xlink:href="#C4_0_07995dbc83" y="280.021079"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="626.609957" xlink:href="#C4_0_d190ea9174" y="280.021079"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_07995dbc83" y="273.941421"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_d190ea9174" y="273.941421"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_07995dbc83" y="265.654057"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_d190ea9174" y="265.654057"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="618.288247" xlink:href="#C4_0_07995dbc83" y="262.058378"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="618.288247" xlink:href="#C4_0_d190ea9174" y="262.058378"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_07995dbc83" y="248.935194"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_d190ea9174" y="248.935194"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="616.110107" xlink:href="#C4_0_07995dbc83" y="248.9069"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="616.110107" xlink:href="#C4_0_d190ea9174" y="248.9069"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="626.609767" xlink:href="#C4_0_07995dbc83" y="248.893435"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="626.609767" xlink:href="#C4_0_d190ea9174" y="248.893435"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_07995dbc83" y="243.396163"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_d190ea9174" y="243.396163"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_07995dbc83" y="238.959995"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_d190ea9174" y="238.959995"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="616.633467" xlink:href="#C4_0_07995dbc83" y="237.030008"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="616.633467" xlink:href="#C4_0_d190ea9174" y="237.030008"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="626.057533" xlink:href="#C4_0_07995dbc83" y="236.980151"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="626.057533" xlink:href="#C4_0_d190ea9174" y="236.980151"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_07995dbc83" y="227.765285"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_d190ea9174" y="227.765285"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="617.903032" xlink:href="#C4_0_07995dbc83" y="224.428352"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="617.903032" xlink:href="#C4_0_d190ea9174" y="224.428352"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_07995dbc83" y="220.236148"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_d190ea9174" y="220.236148"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_07995dbc83" y="213.258957"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_d190ea9174" y="213.258957"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="616.791913" xlink:href="#C4_0_07995dbc83" y="211.073723"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="616.791913" xlink:href="#C4_0_d190ea9174" y="211.073723"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_07995dbc83" y="207.483215"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_d190ea9174" y="207.483215"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="617.301235" xlink:href="#C4_0_07995dbc83" y="204.670895"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="617.301235" xlink:href="#C4_0_d190ea9174" y="204.670895"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_07995dbc83" y="199.632282"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_d190ea9174" y="199.632282"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="616.168703" xlink:href="#C4_0_07995dbc83" y="198.971089"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="616.168703" xlink:href="#C4_0_d190ea9174" y="198.971089"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="626.370515" xlink:href="#C4_0_07995dbc83" y="198.308418"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="626.370515" xlink:href="#C4_0_d190ea9174" y="198.308418"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_07995dbc83" y="193.584019"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_d190ea9174" y="193.584019"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="616.630728" xlink:href="#C4_0_07995dbc83" y="191.658824"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="616.630728" xlink:href="#C4_0_d190ea9174" y="191.658824"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_07995dbc83" y="188.072881"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_d190ea9174" y="188.072881"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="616.246801" xlink:href="#C4_0_07995dbc83" y="187.067312"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="616.246801" xlink:href="#C4_0_d190ea9174" y="187.067312"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_07995dbc83" y="153.818469"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_d190ea9174" y="153.818469"/> </g> </g> <g id="PathCollection_8"> @@ -2702,85 +2702,85 @@ C -2.236584 -1.29895 -2.5 -0.663008 -2.5 0 C -2.5 0.663008 -2.236584 1.29895 -1.767767 1.767767 C -1.29895 2.236584 -0.663008 2.5 0 2.5 z -" id="C5_0_0f975b7987"/> +" id="C5_0_a3035d72c9"/> </defs> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="710.64" xlink:href="#C5_0_0f975b7987" y="290.143163"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="710.64" xlink:href="#C5_0_a3035d72c9" y="290.143163"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="710.64" xlink:href="#C5_0_0f975b7987" y="271.05296"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="710.64" xlink:href="#C5_0_a3035d72c9" y="271.05296"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="705.390139" xlink:href="#C5_0_0f975b7987" y="271.020684"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="705.390139" xlink:href="#C5_0_a3035d72c9" y="271.020684"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="715.889732" xlink:href="#C5_0_0f975b7987" y="271.008152"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="715.889732" xlink:href="#C5_0_a3035d72c9" y="271.008152"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="710.64" xlink:href="#C5_0_0f975b7987" y="256.669448"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="710.64" xlink:href="#C5_0_a3035d72c9" y="256.669448"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="705.39031" xlink:href="#C5_0_0f975b7987" y="256.621285"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="705.39031" xlink:href="#C5_0_a3035d72c9" y="256.621285"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="715.888557" xlink:href="#C5_0_0f975b7987" y="256.565491"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="715.888557" xlink:href="#C5_0_a3035d72c9" y="256.565491"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="702.293321" xlink:href="#C5_0_0f975b7987" y="253.041079"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="702.293321" xlink:href="#C5_0_a3035d72c9" y="253.041079"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="710.64" xlink:href="#C5_0_0f975b7987" y="242.397727"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="710.64" xlink:href="#C5_0_a3035d72c9" y="242.397727"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="710.64" xlink:href="#C5_0_0f975b7987" y="236.301304"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="710.64" xlink:href="#C5_0_a3035d72c9" y="236.301304"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="705.390018" xlink:href="#C5_0_0f975b7987" y="236.289651"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="705.390018" xlink:href="#C5_0_a3035d72c9" y="236.289651"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="710.64" xlink:href="#C5_0_0f975b7987" y="225.669755"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="710.64" xlink:href="#C5_0_a3035d72c9" y="225.669755"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="706.662207" xlink:href="#C5_0_0f975b7987" y="222.776093"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="706.662207" xlink:href="#C5_0_a3035d72c9" y="222.776093"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="710.64" xlink:href="#C5_0_0f975b7987" y="215.218066"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="710.64" xlink:href="#C5_0_a3035d72c9" y="215.218066"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="705.677351" xlink:href="#C5_0_0f975b7987" y="213.771309"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="705.677351" xlink:href="#C5_0_a3035d72c9" y="213.771309"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="712.988203" xlink:href="#C5_0_0f975b7987" y="211.252465"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="712.988203" xlink:href="#C5_0_a3035d72c9" y="211.252465"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="708.247768" xlink:href="#C5_0_0f975b7987" y="209.346952"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="708.247768" xlink:href="#C5_0_a3035d72c9" y="209.346952"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="716.10116" xlink:href="#C5_0_0f975b7987" y="207.68215"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="716.10116" xlink:href="#C5_0_a3035d72c9" y="207.68215"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="710.64" xlink:href="#C5_0_0f975b7987" y="197.198782"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="710.64" xlink:href="#C5_0_a3035d72c9" y="197.198782"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="710.64" xlink:href="#C5_0_0f975b7987" y="190.657145"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="710.64" xlink:href="#C5_0_a3035d72c9" y="190.657145"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="705.420436" xlink:href="#C5_0_0f975b7987" y="190.180405"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="705.420436" xlink:href="#C5_0_a3035d72c9" y="190.180405"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="710.64" xlink:href="#C5_0_0f975b7987" y="178.222524"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="710.64" xlink:href="#C5_0_a3035d72c9" y="178.222524"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="710.64" xlink:href="#C5_0_0f975b7987" y="169.847016"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="710.64" xlink:href="#C5_0_a3035d72c9" y="169.847016"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="705.390063" xlink:href="#C5_0_0f975b7987" y="169.82537"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="705.390063" xlink:href="#C5_0_a3035d72c9" y="169.82537"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="715.360099" xlink:href="#C5_0_0f975b7987" y="167.90583"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="715.360099" xlink:href="#C5_0_a3035d72c9" y="167.90583"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="710.64" xlink:href="#C5_0_0f975b7987" y="161.551986"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="710.64" xlink:href="#C5_0_a3035d72c9" y="161.551986"/> </g> </g> <g id="text_10"> @@ -3002,10 +3002,10 @@ C -3.464901 -2.012324 -3.872983 -1.027127 -3.872983 0 C -3.872983 1.027127 -3.464901 2.012324 -2.738613 2.738613 C -2.012324 3.464901 -1.027127 3.872983 0 3.872983 z -" id="ma1779d3c2f" style="stroke:#4c72b0;"/> +" id="m9b16eda43b" style="stroke:#4c72b0;"/> </defs> <g> - <use style="fill:#4c72b0;stroke:#4c72b0;" x="665.974063" xlink:href="#ma1779d3c2f" y="83.487344"/> + <use style="fill:#4c72b0;stroke:#4c72b0;" x="665.974063" xlink:href="#m9b16eda43b" y="83.487344"/> </g> </g> <g id="text_14"> @@ -3038,10 +3038,10 @@ C -3.464901 -2.012324 -3.872983 -1.027127 -3.872983 0 C -3.872983 1.027127 -3.464901 2.012324 -2.738613 2.738613 C -2.012324 3.464901 -1.027127 3.872983 0 3.872983 z -" id="mf97074e5f0" style="stroke:#dd8452;"/> +" id="mc9e39b2edf" style="stroke:#dd8452;"/> </defs> <g> - <use style="fill:#dd8452;stroke:#dd8452;" x="665.974063" xlink:href="#mf97074e5f0" y="99.310156"/> + <use style="fill:#dd8452;stroke:#dd8452;" x="665.974063" xlink:href="#mc9e39b2edf" y="99.310156"/> </g> </g> <g id="text_15"> @@ -3063,7 +3063,7 @@ z </g> </g> <defs> - <clipPath id="p37bf755c76"> + <clipPath id="p19cc01c258"> <rect height="326.16" width="669.6" x="108" y="51.84"/> </clipPath> </defs> diff --git a/usage/flu_dataset_log_scale_in_axes_strip.svg b/usage/flu_dataset_log_scale_in_axes_strip.svg index c2160f1..9c37087 100644 --- a/usage/flu_dataset_log_scale_in_axes_strip.svg +++ b/usage/flu_dataset_log_scale_in_axes_strip.svg @@ -115,7 +115,7 @@ Q 16.703125 10.84375 15.234375 8.453125 z " id="ArialMT-50"/> </defs> - <g style="fill:#262626;" transform="translate(185.204637 395.373594)scale(0.11 -0.11)"> + <g style="fill:#262626;" transform="translate(185.158394 395.373594)scale(0.11 -0.11)"> <use xlink:href="#ArialMT-72"/> <use x="72.216797" xlink:href="#ArialMT-51"/> <use x="127.832031" xlink:href="#ArialMT-78"/> @@ -139,7 +139,7 @@ L 37.25 71.875 z " id="ArialMT-49"/> </defs> - <g style="fill:#262626;" transform="translate(429.472379 395.373594)scale(0.11 -0.11)"> + <g style="fill:#262626;" transform="translate(429.107034 395.373594)scale(0.11 -0.11)"> <use xlink:href="#ArialMT-72"/> <use x="72.216797" xlink:href="#ArialMT-49"/> <use x="127.832031" xlink:href="#ArialMT-78"/> @@ -345,7 +345,7 @@ L 16.796875 33.0625 z " id="ArialMT-66"/> </defs> - <g style="fill:#262626;" transform="translate(660.284028 395.373594)scale(0.11 -0.11)"> + <g style="fill:#262626;" transform="translate(659.59958 395.373594)scale(0.11 -0.11)"> <use xlink:href="#ArialMT-73"/> <use x="27.783203" xlink:href="#ArialMT-110"/> <use x="83.398438" xlink:href="#ArialMT-102"/> @@ -510,7 +510,7 @@ z <g id="matplotlib.axis_2"> <g id="ytick_1"> <g id="line2d_1"> - <path clip-path="url(#p24d2ef447f)" d="M 108 355.173964 + <path clip-path="url(#p078fcafce0)" d="M 108 355.173964 L 777.6 355.173964 " style="fill:none;stroke:#cccccc;stroke-linecap:round;"/> </g> @@ -584,7 +584,7 @@ z </g> <g id="ytick_2"> <g id="line2d_2"> - <path clip-path="url(#p24d2ef447f)" d="M 108 258.658399 + <path clip-path="url(#p078fcafce0)" d="M 108 258.658399 L 777.6 258.658399 " style="fill:none;stroke:#cccccc;stroke-linecap:round;"/> </g> @@ -619,7 +619,7 @@ z </g> <g id="ytick_3"> <g id="line2d_3"> - <path clip-path="url(#p24d2ef447f)" d="M 108 162.142834 + <path clip-path="url(#p078fcafce0)" d="M 108 162.142834 L 777.6 162.142834 " style="fill:none;stroke:#cccccc;stroke-linecap:round;"/> </g> @@ -635,7 +635,7 @@ L 777.6 162.142834 </g> <g id="ytick_4"> <g id="line2d_4"> - <path clip-path="url(#p24d2ef447f)" d="M 108 65.627269 + <path clip-path="url(#p078fcafce0)" d="M 108 65.627269 L 777.6 65.627269 " style="fill:none;stroke:#cccccc;stroke-linecap:round;"/> </g> @@ -790,24 +790,24 @@ z <g id="PathCollection_1"/> <g id="PathCollection_2"/> <g id="line2d_5"> - <path clip-path="url(#p24d2ef447f)" d="M 394.679924 118.96632 -L 394.679924 113.094863 -L 492.387021 113.094863 -L 492.387021 118.96632 + <path clip-path="url(#p078fcafce0)" d="M 394.3784 118.96632 +L 394.3784 113.094863 +L 491.957856 113.094863 +L 491.957856 118.96632 " style="fill:none;stroke:#333333;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_6"> - <path clip-path="url(#p24d2ef447f)" d="M 150.412182 80.157043 -L 150.412182 74.285587 -L 248.119279 74.285587 -L 248.119279 80.157043 + <path clip-path="url(#p078fcafce0)" d="M 150.429759 80.157043 +L 150.429759 74.285587 +L 248.009215 74.285587 +L 248.009215 80.157043 " style="fill:none;stroke:#333333;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_7"> - <path clip-path="url(#p24d2ef447f)" d="M 638.947667 132.836253 -L 638.947667 126.964797 -L 736.654764 126.964797 -L 736.654764 132.836253 + <path clip-path="url(#p078fcafce0)" d="M 638.32704 132.836253 +L 638.32704 126.964797 +L 735.906496 126.964797 +L 735.906496 132.836253 " style="fill:none;stroke:#333333;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="patch_3"> @@ -842,589 +842,589 @@ C -2.236584 -1.29895 -2.5 -0.663008 -2.5 0 C -2.5 0.663008 -2.236584 1.29895 -1.767767 1.767767 C -1.29895 2.236584 -0.663008 2.5 0 2.5 z -" id="C0_0_2a88d1c6b0"/> +" id="C0_0_af5b03a438"/> </defs> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="147.060601" xlink:href="#C0_0_2a88d1c6b0" y="234.464917"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="147.384047" xlink:href="#C0_0_af5b03a438" y="234.464917"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="153.728993" xlink:href="#C0_0_2a88d1c6b0" y="234.469355"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="156.319934" xlink:href="#C0_0_af5b03a438" y="234.469355"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="142.975993" xlink:href="#C0_0_2a88d1c6b0" y="186.01941"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="143.707055" xlink:href="#C0_0_af5b03a438" y="186.01941"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="145.058984" xlink:href="#C0_0_2a88d1c6b0" y="217.469375"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="158.997726" xlink:href="#C0_0_af5b03a438" y="217.469375"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="157.886778" xlink:href="#C0_0_2a88d1c6b0" y="181.602768"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="146.26079" xlink:href="#C0_0_af5b03a438" y="181.602768"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="141.682206" xlink:href="#C0_0_2a88d1c6b0" y="206.660532"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="157.257495" xlink:href="#C0_0_af5b03a438" y="206.660532"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="155.013194" xlink:href="#C0_0_2a88d1c6b0" y="217.469308"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="158.3782" xlink:href="#C0_0_af5b03a438" y="217.469308"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="156.076163" xlink:href="#C0_0_2a88d1c6b0" y="159.361091"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="152.719874" xlink:href="#C0_0_af5b03a438" y="159.361091"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="152.644414" xlink:href="#C0_0_2a88d1c6b0" y="200.097852"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="142.68083" xlink:href="#C0_0_af5b03a438" y="200.097852"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="162.204833" xlink:href="#C0_0_2a88d1c6b0" y="237.168751"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="156.736524" xlink:href="#C0_0_af5b03a438" y="237.168751"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="144.539814" xlink:href="#C0_0_2a88d1c6b0" y="204.470561"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="142.793119" xlink:href="#C0_0_af5b03a438" y="204.470561"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="145.589841" xlink:href="#C0_0_2a88d1c6b0" y="170.604131"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="139.50113" xlink:href="#C0_0_af5b03a438" y="170.604131"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="161.521618" xlink:href="#C0_0_2a88d1c6b0" y="148.828987"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="154.13707" xlink:href="#C0_0_af5b03a438" y="148.828987"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="160.930571" xlink:href="#C0_0_2a88d1c6b0" y="161.142385"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="152.786131" xlink:href="#C0_0_af5b03a438" y="161.142385"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="142.043267" xlink:href="#C0_0_2a88d1c6b0" y="261.123162"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="151.352819" xlink:href="#C0_0_af5b03a438" y="261.123162"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="162.581739" xlink:href="#C0_0_2a88d1c6b0" y="213.05278"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="154.503085" xlink:href="#C0_0_af5b03a438" y="213.05278"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="156.507004" xlink:href="#C0_0_2a88d1c6b0" y="246.52338"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="153.029995" xlink:href="#C0_0_af5b03a438" y="246.52338"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="139.967855" xlink:href="#C0_0_2a88d1c6b0" y="202.518813"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="141.345687" xlink:href="#C0_0_af5b03a438" y="202.518813"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="149.960264" xlink:href="#C0_0_2a88d1c6b0" y="263.518057"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="142.718516" xlink:href="#C0_0_af5b03a438" y="263.518057"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="139.778789" xlink:href="#C0_0_2a88d1c6b0" y="145.788884"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="157.425812" xlink:href="#C0_0_af5b03a438" y="145.788884"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="144.746448" xlink:href="#C0_0_2a88d1c6b0" y="199.222027"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="138.52142" xlink:href="#C0_0_af5b03a438" y="199.222027"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="139.087382" xlink:href="#C0_0_2a88d1c6b0" y="198.762645"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="158.958474" xlink:href="#C0_0_af5b03a438" y="198.762645"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="150.118409" xlink:href="#C0_0_2a88d1c6b0" y="211.870604"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="138.610994" xlink:href="#C0_0_af5b03a438" y="211.870604"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="147.871271" xlink:href="#C0_0_2a88d1c6b0" y="203.015022"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="149.926606" xlink:href="#C0_0_af5b03a438" y="203.015022"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="158.517174" xlink:href="#C0_0_2a88d1c6b0" y="246.521469"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="162.583694" xlink:href="#C0_0_af5b03a438" y="246.521469"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="147.357868" xlink:href="#C0_0_2a88d1c6b0" y="185.060335"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="149.861194" xlink:href="#C0_0_af5b03a438" y="185.060335"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="148.132868" xlink:href="#C0_0_2a88d1c6b0" y="199.222339"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="150.875252" xlink:href="#C0_0_af5b03a438" y="199.222339"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="159.911231" xlink:href="#C0_0_2a88d1c6b0" y="292.574965"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="151.180541" xlink:href="#C0_0_af5b03a438" y="292.574965"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="152.844267" xlink:href="#C0_0_2a88d1c6b0" y="278.118541"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="153.250301" xlink:href="#C0_0_af5b03a438" y="278.118541"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="162.550598" xlink:href="#C0_0_2a88d1c6b0" y="288.925258"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="144.932173" xlink:href="#C0_0_af5b03a438" y="288.925258"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="144.998518" xlink:href="#C0_0_2a88d1c6b0" y="197.117394"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="141.24368" xlink:href="#C0_0_af5b03a438" y="197.117394"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="158.082028" xlink:href="#C0_0_2a88d1c6b0" y="293.722513"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="148.861059" xlink:href="#C0_0_af5b03a438" y="293.722513"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="151.318422" xlink:href="#C0_0_2a88d1c6b0" y="244.997663"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="160.2273" xlink:href="#C0_0_af5b03a438" y="244.997663"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="158.55584" xlink:href="#C0_0_2a88d1c6b0" y="192.062031"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="149.797892" xlink:href="#C0_0_af5b03a438" y="192.062031"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="140.910831" xlink:href="#C0_0_2a88d1c6b0" y="236.731089"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="149.177982" xlink:href="#C0_0_af5b03a438" y="236.731089"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="140.983768" xlink:href="#C0_0_2a88d1c6b0" y="326.564074"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="156.055941" xlink:href="#C0_0_af5b03a438" y="326.564074"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="140.314076" xlink:href="#C0_0_2a88d1c6b0" y="206.072603"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="148.029634" xlink:href="#C0_0_af5b03a438" y="206.072603"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="162.242098" xlink:href="#C0_0_2a88d1c6b0" y="287.636006"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="146.603045" xlink:href="#C0_0_af5b03a438" y="287.636006"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="148.970788" xlink:href="#C0_0_2a88d1c6b0" y="186.019413"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="153.228295" xlink:href="#C0_0_af5b03a438" y="186.019413"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="160.718673" xlink:href="#C0_0_2a88d1c6b0" y="234.463233"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="142.488332" xlink:href="#C0_0_af5b03a438" y="234.463233"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="148.875736" xlink:href="#C0_0_2a88d1c6b0" y="204.465046"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="157.532766" xlink:href="#C0_0_af5b03a438" y="204.465046"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="147.786354" xlink:href="#C0_0_2a88d1c6b0" y="249.057643"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="149.075558" xlink:href="#C0_0_af5b03a438" y="249.057643"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="141.687814" xlink:href="#C0_0_2a88d1c6b0" y="164.607881"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="146.980851" xlink:href="#C0_0_af5b03a438" y="164.607881"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="157.625102" xlink:href="#C0_0_2a88d1c6b0" y="196.553038"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="150.781267" xlink:href="#C0_0_af5b03a438" y="196.553038"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="141.588186" xlink:href="#C0_0_2a88d1c6b0" y="156.468577"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="143.965871" xlink:href="#C0_0_af5b03a438" y="156.468577"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="161.339998" xlink:href="#C0_0_2a88d1c6b0" y="222.407449"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="150.09534" xlink:href="#C0_0_af5b03a438" y="222.407449"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="146.873728" xlink:href="#C0_0_2a88d1c6b0" y="280.51449"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="144.404388" xlink:href="#C0_0_af5b03a438" y="280.51449"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="138.436364" xlink:href="#C0_0_2a88d1c6b0" y="193.661519"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="148.761673" xlink:href="#C0_0_af5b03a438" y="193.661519"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="153.227196" xlink:href="#C0_0_2a88d1c6b0" y="280.514507"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="140.397152" xlink:href="#C0_0_af5b03a438" y="280.514507"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="149.526654" xlink:href="#C0_0_2a88d1c6b0" y="163.008982"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="139.1638" xlink:href="#C0_0_af5b03a438" y="163.008982"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="144.694992" xlink:href="#C0_0_2a88d1c6b0" y="307.172431"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="148.797431" xlink:href="#C0_0_af5b03a438" y="307.172431"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="160.631469" xlink:href="#C0_0_2a88d1c6b0" y="240.518048"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="148.936914" xlink:href="#C0_0_af5b03a438" y="240.518048"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="154.109667" xlink:href="#C0_0_2a88d1c6b0" y="226.166435"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="139.513321" xlink:href="#C0_0_af5b03a438" y="226.166435"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="157.974153" xlink:href="#C0_0_2a88d1c6b0" y="199.223789"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="147.412371" xlink:href="#C0_0_af5b03a438" y="199.223789"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="138.949906" xlink:href="#C0_0_2a88d1c6b0" y="307.172214"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="159.539624" xlink:href="#C0_0_af5b03a438" y="307.172214"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="150.25069" xlink:href="#C0_0_2a88d1c6b0" y="271.928388"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="145.317523" xlink:href="#C0_0_af5b03a438" y="271.928388"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="152.664616" xlink:href="#C0_0_2a88d1c6b0" y="196.99711"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="158.404104" xlink:href="#C0_0_af5b03a438" y="196.99711"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="143.546974" xlink:href="#C0_0_2a88d1c6b0" y="295.926683"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="143.351895" xlink:href="#C0_0_af5b03a438" y="295.926683"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="157.664136" xlink:href="#C0_0_2a88d1c6b0" y="191.114521"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="150.510906" xlink:href="#C0_0_af5b03a438" y="191.114521"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="155.430434" xlink:href="#C0_0_2a88d1c6b0" y="194.87592"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="153.889938" xlink:href="#C0_0_af5b03a438" y="194.87592"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="147.302341" xlink:href="#C0_0_2a88d1c6b0" y="364.655781"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="159.276902" xlink:href="#C0_0_af5b03a438" y="364.655781"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="149.465308" xlink:href="#C0_0_2a88d1c6b0" y="197.452892"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="141.670173" xlink:href="#C0_0_af5b03a438" y="197.452892"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="147.121025" xlink:href="#C0_0_2a88d1c6b0" y="181.294506"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="148.233078" xlink:href="#C0_0_af5b03a438" y="181.294506"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="141.270149" xlink:href="#C0_0_2a88d1c6b0" y="317.974952"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="157.678147" xlink:href="#C0_0_af5b03a438" y="317.974952"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="154.585335" xlink:href="#C0_0_2a88d1c6b0" y="269.984711"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="147.294209" xlink:href="#C0_0_af5b03a438" y="269.984711"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="147.61082" xlink:href="#C0_0_2a88d1c6b0" y="164.961136"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="158.442183" xlink:href="#C0_0_af5b03a438" y="164.961136"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="143.649969" xlink:href="#C0_0_2a88d1c6b0" y="202.517314"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="157.570921" xlink:href="#C0_0_af5b03a438" y="202.517314"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="142.022999" xlink:href="#C0_0_2a88d1c6b0" y="292.574949"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="151.838554" xlink:href="#C0_0_af5b03a438" y="292.574949"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="154.283824" xlink:href="#C0_0_2a88d1c6b0" y="171.426949"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="146.673253" xlink:href="#C0_0_af5b03a438" y="171.426949"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="149.079039" xlink:href="#C0_0_2a88d1c6b0" y="192.419882"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="150.847812" xlink:href="#C0_0_af5b03a438" y="192.419882"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="144.040599" xlink:href="#C0_0_2a88d1c6b0" y="195.839019"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="145.619372" xlink:href="#C0_0_af5b03a438" y="195.839019"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="151.537642" xlink:href="#C0_0_2a88d1c6b0" y="258.578438"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="151.563431" xlink:href="#C0_0_af5b03a438" y="258.578438"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="155.912469" xlink:href="#C0_0_2a88d1c6b0" y="249.071341"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="142.147195" xlink:href="#C0_0_af5b03a438" y="249.071341"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="161.424916" xlink:href="#C0_0_2a88d1c6b0" y="231.577382"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="143.816134" xlink:href="#C0_0_af5b03a438" y="231.577382"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="159.442792" xlink:href="#C0_0_2a88d1c6b0" y="288.92293"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="161.670412" xlink:href="#C0_0_af5b03a438" y="288.92293"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="153.839016" xlink:href="#C0_0_2a88d1c6b0" y="196.996728"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="149.388348" xlink:href="#C0_0_af5b03a438" y="196.996728"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="142.41653" xlink:href="#C0_0_2a88d1c6b0" y="248.570576"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="157.82596" xlink:href="#C0_0_af5b03a438" y="248.570576"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="152.904397" xlink:href="#C0_0_2a88d1c6b0" y="161.149028"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="161.27347" xlink:href="#C0_0_af5b03a438" y="161.149028"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="142.754295" xlink:href="#C0_0_2a88d1c6b0" y="150.012056"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="151.687053" xlink:href="#C0_0_af5b03a438" y="150.012056"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="146.252225" xlink:href="#C0_0_2a88d1c6b0" y="211.011584"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="153.792147" xlink:href="#C0_0_af5b03a438" y="211.011584"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="146.916709" xlink:href="#C0_0_2a88d1c6b0" y="263.516866"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="156.430205" xlink:href="#C0_0_af5b03a438" y="263.516866"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="147.354924" xlink:href="#C0_0_2a88d1c6b0" y="269.978278"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="161.750727" xlink:href="#C0_0_af5b03a438" y="269.978278"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="159.038186" xlink:href="#C0_0_2a88d1c6b0" y="274.69867"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="153.586616" xlink:href="#C0_0_af5b03a438" y="274.69867"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="146.835508" xlink:href="#C0_0_2a88d1c6b0" y="171.425088"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="147.166331" xlink:href="#C0_0_af5b03a438" y="171.425088"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="146.92867" xlink:href="#C0_0_2a88d1c6b0" y="181.959734"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="161.412731" xlink:href="#C0_0_af5b03a438" y="181.959734"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="157.273868" xlink:href="#C0_0_2a88d1c6b0" y="166.030933"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="160.349209" xlink:href="#C0_0_af5b03a438" y="166.030933"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="157.569593" xlink:href="#C0_0_2a88d1c6b0" y="307.172933"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="156.271312" xlink:href="#C0_0_af5b03a438" y="307.172933"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="161.53809" xlink:href="#C0_0_2a88d1c6b0" y="204.472191"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="155.112727" xlink:href="#C0_0_af5b03a438" y="204.472191"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="154.943557" xlink:href="#C0_0_2a88d1c6b0" y="188.412611"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="141.475733" xlink:href="#C0_0_af5b03a438" y="188.412611"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="144.79596" xlink:href="#C0_0_2a88d1c6b0" y="169.150455"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="161.580107" xlink:href="#C0_0_af5b03a438" y="169.150455"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="138.787223" xlink:href="#C0_0_2a88d1c6b0" y="292.556326"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="153.016583" xlink:href="#C0_0_af5b03a438" y="292.556326"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="140.933693" xlink:href="#C0_0_2a88d1c6b0" y="190.002095"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="143.433511" xlink:href="#C0_0_af5b03a438" y="190.002095"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="152.878676" xlink:href="#C0_0_2a88d1c6b0" y="152.896666"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="160.626923" xlink:href="#C0_0_af5b03a438" y="152.896666"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="146.383385" xlink:href="#C0_0_2a88d1c6b0" y="220.011786"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="147.424477" xlink:href="#C0_0_af5b03a438" y="220.011786"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="145.250027" xlink:href="#C0_0_2a88d1c6b0" y="171.418987"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="139.708714" xlink:href="#C0_0_af5b03a438" y="171.418987"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="148.825657" xlink:href="#C0_0_2a88d1c6b0" y="257.697832"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="148.189933" xlink:href="#C0_0_af5b03a438" y="257.697832"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="141.757968" xlink:href="#C0_0_2a88d1c6b0" y="193.349058"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="140.811951" xlink:href="#C0_0_af5b03a438" y="193.349058"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="148.174666" xlink:href="#C0_0_2a88d1c6b0" y="292.577062"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="146.650705" xlink:href="#C0_0_af5b03a438" y="292.577062"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="146.108432" xlink:href="#C0_0_2a88d1c6b0" y="136.751329"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="143.393213" xlink:href="#C0_0_af5b03a438" y="136.751329"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="155.283425" xlink:href="#C0_0_2a88d1c6b0" y="205.410931"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="144.519781" xlink:href="#C0_0_af5b03a438" y="205.410931"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="159.384261" xlink:href="#C0_0_2a88d1c6b0" y="173.812482"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="145.901263" xlink:href="#C0_0_af5b03a438" y="173.812482"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="155.429041" xlink:href="#C0_0_2a88d1c6b0" y="215.077353"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="138.66636" xlink:href="#C0_0_af5b03a438" y="215.077353"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="159.555641" xlink:href="#C0_0_2a88d1c6b0" y="263.513214"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="138.436364" xlink:href="#C0_0_af5b03a438" y="263.513214"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="145.199378" xlink:href="#C0_0_2a88d1c6b0" y="239.386134"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="140.403664" xlink:href="#C0_0_af5b03a438" y="239.386134"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="144.782713" xlink:href="#C0_0_2a88d1c6b0" y="186.546072"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="154.308511" xlink:href="#C0_0_af5b03a438" y="186.546072"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="150.03091" xlink:href="#C0_0_2a88d1c6b0" y="210.640421"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="161.992634" xlink:href="#C0_0_af5b03a438" y="210.640421"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="138.812507" xlink:href="#C0_0_2a88d1c6b0" y="197.116231"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="144.834949" xlink:href="#C0_0_af5b03a438" y="197.116231"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="148.290301" xlink:href="#C0_0_2a88d1c6b0" y="251.454711"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="145.577076" xlink:href="#C0_0_af5b03a438" y="251.454711"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="141.310817" xlink:href="#C0_0_2a88d1c6b0" y="169.146493"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="154.492914" xlink:href="#C0_0_af5b03a438" y="169.146493"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="155.253548" xlink:href="#C0_0_2a88d1c6b0" y="156.000344"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="151.712636" xlink:href="#C0_0_af5b03a438" y="156.000344"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="138.571849" xlink:href="#C0_0_2a88d1c6b0" y="278.111605"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="156.859975" xlink:href="#C0_0_af5b03a438" y="278.111605"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="161.91792" xlink:href="#C0_0_2a88d1c6b0" y="199.648095"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="152.633655" xlink:href="#C0_0_af5b03a438" y="199.648095"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="153.636897" xlink:href="#C0_0_2a88d1c6b0" y="246.521571"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="148.540912" xlink:href="#C0_0_af5b03a438" y="246.521571"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="154.861214" xlink:href="#C0_0_2a88d1c6b0" y="188.40958"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="144.146144" xlink:href="#C0_0_af5b03a438" y="188.40958"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="142.491003" xlink:href="#C0_0_2a88d1c6b0" y="211.002179"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="142.401129" xlink:href="#C0_0_af5b03a438" y="211.002179"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="158.600038" xlink:href="#C0_0_2a88d1c6b0" y="232.060628"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="142.137987" xlink:href="#C0_0_af5b03a438" y="232.060628"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="140.747655" xlink:href="#C0_0_2a88d1c6b0" y="211.855489"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="152.06608" xlink:href="#C0_0_af5b03a438" y="211.855489"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="159.948904" xlink:href="#C0_0_2a88d1c6b0" y="151.208438"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="138.88828" xlink:href="#C0_0_af5b03a438" y="151.208438"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="158.672272" xlink:href="#C0_0_2a88d1c6b0" y="252.600066"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="147.99149" xlink:href="#C0_0_af5b03a438" y="252.600066"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="153.011734" xlink:href="#C0_0_2a88d1c6b0" y="271.924558"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="160.654869" xlink:href="#C0_0_af5b03a438" y="271.924558"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="148.842531" xlink:href="#C0_0_2a88d1c6b0" y="179.059415"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="149.069864" xlink:href="#C0_0_af5b03a438" y="179.059415"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="151.3317" xlink:href="#C0_0_2a88d1c6b0" y="249.065509"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="154.346603" xlink:href="#C0_0_af5b03a438" y="249.065509"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="149.740745" xlink:href="#C0_0_2a88d1c6b0" y="263.520399"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="141.220871" xlink:href="#C0_0_af5b03a438" y="263.520399"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="144.611396" xlink:href="#C0_0_2a88d1c6b0" y="263.52343"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="140.956356" xlink:href="#C0_0_af5b03a438" y="263.52343"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="143.521923" xlink:href="#C0_0_2a88d1c6b0" y="261.124522"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="160.717068" xlink:href="#C0_0_af5b03a438" y="261.124522"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="142.982425" xlink:href="#C0_0_2a88d1c6b0" y="307.174143"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="147.323679" xlink:href="#C0_0_af5b03a438" y="307.174143"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="161.033468" xlink:href="#C0_0_2a88d1c6b0" y="200.477101"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="151.793433" xlink:href="#C0_0_af5b03a438" y="200.477101"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="147.722977" xlink:href="#C0_0_2a88d1c6b0" y="269.986637"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="140.252609" xlink:href="#C0_0_af5b03a438" y="269.986637"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="150.856048" xlink:href="#C0_0_2a88d1c6b0" y="320.750698"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="155.47779" xlink:href="#C0_0_af5b03a438" y="320.750698"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="146.556013" xlink:href="#C0_0_2a88d1c6b0" y="278.124431"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="149.053455" xlink:href="#C0_0_af5b03a438" y="278.124431"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="161.673047" xlink:href="#C0_0_2a88d1c6b0" y="273.978125"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="144.265826" xlink:href="#C0_0_af5b03a438" y="273.978125"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="152.082481" xlink:href="#C0_0_2a88d1c6b0" y="235.616775"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="159.079075" xlink:href="#C0_0_af5b03a438" y="235.616775"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="145.902122" xlink:href="#C0_0_2a88d1c6b0" y="210.352087"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="158.068443" xlink:href="#C0_0_af5b03a438" y="210.352087"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="138.679104" xlink:href="#C0_0_2a88d1c6b0" y="199.663269"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="139.72174" xlink:href="#C0_0_af5b03a438" y="199.663269"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="148.26969" xlink:href="#C0_0_2a88d1c6b0" y="164.976889"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="144.302111" xlink:href="#C0_0_af5b03a438" y="164.976889"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="152.836815" xlink:href="#C0_0_2a88d1c6b0" y="146.019431"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="144.638777" xlink:href="#C0_0_af5b03a438" y="146.019431"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="152.426918" xlink:href="#C0_0_2a88d1c6b0" y="263.537703"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="159.055351" xlink:href="#C0_0_af5b03a438" y="263.537703"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="142.416817" xlink:href="#C0_0_2a88d1c6b0" y="211.006731"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="160.205048" xlink:href="#C0_0_af5b03a438" y="211.006731"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="153.3067" xlink:href="#C0_0_2a88d1c6b0" y="148.861019"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="150.968999" xlink:href="#C0_0_af5b03a438" y="148.861019"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="152.086506" xlink:href="#C0_0_2a88d1c6b0" y="242.889012"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="150.013456" xlink:href="#C0_0_af5b03a438" y="242.889012"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="143.634083" xlink:href="#C0_0_2a88d1c6b0" y="240.262686"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="138.659046" xlink:href="#C0_0_af5b03a438" y="240.262686"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="152.888113" xlink:href="#C0_0_2a88d1c6b0" y="229.527227"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="150.585288" xlink:href="#C0_0_af5b03a438" y="229.527227"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="153.931949" xlink:href="#C0_0_2a88d1c6b0" y="263.524022"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="155.072047" xlink:href="#C0_0_af5b03a438" y="263.524022"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="156.698164" xlink:href="#C0_0_2a88d1c6b0" y="199.418433"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="160.92629" xlink:href="#C0_0_af5b03a438" y="199.418433"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="161.473311" xlink:href="#C0_0_2a88d1c6b0" y="257.269739"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="142.238979" xlink:href="#C0_0_af5b03a438" y="257.269739"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="157.028195" xlink:href="#C0_0_2a88d1c6b0" y="288.935701"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="144.811952" xlink:href="#C0_0_af5b03a438" y="288.935701"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="159.767095" xlink:href="#C0_0_2a88d1c6b0" y="249.077259"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="139.111581" xlink:href="#C0_0_af5b03a438" y="249.077259"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="142.34957" xlink:href="#C0_0_2a88d1c6b0" y="196.84005"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="145.95522" xlink:href="#C0_0_af5b03a438" y="196.84005"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="157.140446" xlink:href="#C0_0_2a88d1c6b0" y="263.531794"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="160.646619" xlink:href="#C0_0_af5b03a438" y="263.531794"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="161.055627" xlink:href="#C0_0_2a88d1c6b0" y="211.011912"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="162.535656" xlink:href="#C0_0_af5b03a438" y="211.011912"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="156.812933" xlink:href="#C0_0_2a88d1c6b0" y="253.288159"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="143.256002" xlink:href="#C0_0_af5b03a438" y="253.288159"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="144.04369" xlink:href="#C0_0_2a88d1c6b0" y="217.482077"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="147.182743" xlink:href="#C0_0_af5b03a438" y="217.482077"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="156.572331" xlink:href="#C0_0_2a88d1c6b0" y="249.077437"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="161.523331" xlink:href="#C0_0_af5b03a438" y="249.077437"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="158.332349" xlink:href="#C0_0_2a88d1c6b0" y="307.185539"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="157.423385" xlink:href="#C0_0_af5b03a438" y="307.185539"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="159.330613" xlink:href="#C0_0_2a88d1c6b0" y="154.424025"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="162.548681" xlink:href="#C0_0_af5b03a438" y="154.424025"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="140.208615" xlink:href="#C0_0_2a88d1c6b0" y="192.146247"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="160.065648" xlink:href="#C0_0_af5b03a438" y="192.146247"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="150.967225" xlink:href="#C0_0_2a88d1c6b0" y="200.78484"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="158.695069" xlink:href="#C0_0_af5b03a438" y="200.78484"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="138.535192" xlink:href="#C0_0_2a88d1c6b0" y="307.171497"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="159.030584" xlink:href="#C0_0_af5b03a438" y="307.171497"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="142.703322" xlink:href="#C0_0_2a88d1c6b0" y="317.984868"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="153.194575" xlink:href="#C0_0_af5b03a438" y="317.984868"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="140.435589" xlink:href="#C0_0_2a88d1c6b0" y="182.812804"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="146.287527" xlink:href="#C0_0_af5b03a438" y="182.812804"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="161.393653" xlink:href="#C0_0_2a88d1c6b0" y="292.584627"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="144.868187" xlink:href="#C0_0_af5b03a438" y="292.584627"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="154.259142" xlink:href="#C0_0_2a88d1c6b0" y="225.116796"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="155.954442" xlink:href="#C0_0_af5b03a438" y="225.116796"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="161.836345" xlink:href="#C0_0_2a88d1c6b0" y="159.366585"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="148.357881" xlink:href="#C0_0_af5b03a438" y="159.366585"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="153.144964" xlink:href="#C0_0_2a88d1c6b0" y="156.018214"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="161.821635" xlink:href="#C0_0_af5b03a438" y="156.018214"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="157.287056" xlink:href="#C0_0_2a88d1c6b0" y="278.130212"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="158.574791" xlink:href="#C0_0_af5b03a438" y="278.130212"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="153.362297" xlink:href="#C0_0_2a88d1c6b0" y="207.372381"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="160.998279" xlink:href="#C0_0_af5b03a438" y="207.372381"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="159.552743" xlink:href="#C0_0_2a88d1c6b0" y="291.703424"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="154.184456" xlink:href="#C0_0_af5b03a438" y="291.703424"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="155.765416" xlink:href="#C0_0_2a88d1c6b0" y="293.73309"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="140.496466" xlink:href="#C0_0_af5b03a438" y="293.73309"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="144.456313" xlink:href="#C0_0_2a88d1c6b0" y="268.237954"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="154.139575" xlink:href="#C0_0_af5b03a438" y="268.237954"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="159.571159" xlink:href="#C0_0_2a88d1c6b0" y="261.134578"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="161.580194" xlink:href="#C0_0_af5b03a438" y="261.134578"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="150.166371" xlink:href="#C0_0_2a88d1c6b0" y="176.361807"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="157.141028" xlink:href="#C0_0_af5b03a438" y="176.361807"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="139.233875" xlink:href="#C0_0_2a88d1c6b0" y="188.417584"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="155.561754" xlink:href="#C0_0_af5b03a438" y="188.417584"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="143.021597" xlink:href="#C0_0_2a88d1c6b0" y="293.720295"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="149.363576" xlink:href="#C0_0_af5b03a438" y="293.720295"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="143.890492" xlink:href="#C0_0_2a88d1c6b0" y="197.375617"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="146.512782" xlink:href="#C0_0_af5b03a438" y="197.375617"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="160.474152" xlink:href="#C0_0_2a88d1c6b0" y="234.476448"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="149.952187" xlink:href="#C0_0_af5b03a438" y="234.476448"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="154.369352" xlink:href="#C0_0_2a88d1c6b0" y="190.97031"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="138.905186" xlink:href="#C0_0_af5b03a438" y="190.97031"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="141.943243" xlink:href="#C0_0_2a88d1c6b0" y="307.176938"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="159.088023" xlink:href="#C0_0_af5b03a438" y="307.176938"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="151.113188" xlink:href="#C0_0_2a88d1c6b0" y="193.675276"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="138.557088" xlink:href="#C0_0_af5b03a438" y="193.675276"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="155.950648" xlink:href="#C0_0_2a88d1c6b0" y="326.568957"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="139.493833" xlink:href="#C0_0_af5b03a438" y="326.568957"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="159.531119" xlink:href="#C0_0_2a88d1c6b0" y="211.017531"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="146.092482" xlink:href="#C0_0_af5b03a438" y="211.017531"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="152.040567" xlink:href="#C0_0_2a88d1c6b0" y="167.00589"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="138.937364" xlink:href="#C0_0_af5b03a438" y="167.00589"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="139.968931" xlink:href="#C0_0_2a88d1c6b0" y="335.61745"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="142.151369" xlink:href="#C0_0_af5b03a438" y="335.61745"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="161.062189" xlink:href="#C0_0_2a88d1c6b0" y="240.92954"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="162.220076" xlink:href="#C0_0_af5b03a438" y="240.92954"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="143.707986" xlink:href="#C0_0_2a88d1c6b0" y="250.523607"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="144.323115" xlink:href="#C0_0_af5b03a438" y="250.523607"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="141.303205" xlink:href="#C0_0_2a88d1c6b0" y="185.065536"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="144.448734" xlink:href="#C0_0_af5b03a438" y="185.065536"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="139.400884" xlink:href="#C0_0_2a88d1c6b0" y="211.02224"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="151.66963" xlink:href="#C0_0_af5b03a438" y="211.02224"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="153.628071" xlink:href="#C0_0_2a88d1c6b0" y="307.18565"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="141.480561" xlink:href="#C0_0_af5b03a438" y="307.18565"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="150.95575" xlink:href="#C0_0_2a88d1c6b0" y="278.124"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="147.987779" xlink:href="#C0_0_af5b03a438" y="278.124"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="154.442719" xlink:href="#C0_0_2a88d1c6b0" y="259.107884"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="159.881403" xlink:href="#C0_0_af5b03a438" y="259.107884"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="151.572956" xlink:href="#C0_0_2a88d1c6b0" y="173.96873"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="152.128647" xlink:href="#C0_0_af5b03a438" y="173.96873"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="145.647245" xlink:href="#C0_0_2a88d1c6b0" y="286.319309"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="149.540781" xlink:href="#C0_0_af5b03a438" y="286.319309"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="157.132116" xlink:href="#C0_0_2a88d1c6b0" y="278.125323"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="151.313386" xlink:href="#C0_0_af5b03a438" y="278.125323"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="148.712155" xlink:href="#C0_0_2a88d1c6b0" y="211.007792"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="154.006386" xlink:href="#C0_0_af5b03a438" y="211.007792"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="145.900809" xlink:href="#C0_0_2a88d1c6b0" y="205.412973"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="142.844745" xlink:href="#C0_0_af5b03a438" y="205.412973"/> </g> </g> <g id="PathCollection_4"> @@ -1439,592 +1439,592 @@ C -2.236584 -1.29895 -2.5 -0.663008 -2.5 0 C -2.5 0.663008 -2.236584 1.29895 -1.767767 1.767767 C -1.29895 2.236584 -0.663008 2.5 0 2.5 z -" id="C1_0_82f4c9fa18"/> +" id="C1_0_f6794d661a"/> </defs> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="258.379252" xlink:href="#C1_0_82f4c9fa18" y="207.50588"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="246.094122" xlink:href="#C1_0_f6794d661a" y="207.50588"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="247.850998" xlink:href="#C1_0_82f4c9fa18" y="253.544249"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="238.182833" xlink:href="#C1_0_f6794d661a" y="253.544249"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="253.281851" xlink:href="#C1_0_82f4c9fa18" y="149.397766"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="255.283893" xlink:href="#C1_0_f6794d661a" y="149.397766"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="240.56049" xlink:href="#C1_0_82f4c9fa18" y="249.90831"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="255.015429" xlink:href="#C1_0_f6794d661a" y="249.90831"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="241.408416" xlink:href="#C1_0_82f4c9fa18" y="193.052371"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="250.339336" xlink:href="#C1_0_f6794d661a" y="193.052371"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="259.934517" xlink:href="#C1_0_82f4c9fa18" y="200.692182"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="240.075278" xlink:href="#C1_0_f6794d661a" y="200.692182"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="239.254965" xlink:href="#C1_0_82f4c9fa18" y="230.96294"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="236.212083" xlink:href="#C1_0_f6794d661a" y="230.96294"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="236.756359" xlink:href="#C1_0_82f4c9fa18" y="207.506037"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="256.243371" xlink:href="#C1_0_f6794d661a" y="207.506037"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="254.206824" xlink:href="#C1_0_82f4c9fa18" y="180.992598"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="241.909136" xlink:href="#C1_0_f6794d661a" y="180.992598"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="242.571711" xlink:href="#C1_0_82f4c9fa18" y="213.590998"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="246.136914" xlink:href="#C1_0_f6794d661a" y="213.590998"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="240.011511" xlink:href="#C1_0_82f4c9fa18" y="213.106325"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="256.281903" xlink:href="#C1_0_f6794d661a" y="213.106325"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="252.595156" xlink:href="#C1_0_82f4c9fa18" y="191.79585"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="251.289133" xlink:href="#C1_0_f6794d661a" y="191.79585"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="250.314214" xlink:href="#C1_0_82f4c9fa18" y="171.643345"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="250.939533" xlink:href="#C1_0_f6794d661a" y="171.643345"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="242.541666" xlink:href="#C1_0_82f4c9fa18" y="172.84981"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="247.688815" xlink:href="#C1_0_f6794d661a" y="172.84981"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="249.449495" xlink:href="#C1_0_82f4c9fa18" y="137.346191"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="242.59829" xlink:href="#C1_0_f6794d661a" y="137.346191"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="259.575594" xlink:href="#C1_0_82f4c9fa18" y="210.046514"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="240.168943" xlink:href="#C1_0_f6794d661a" y="210.046514"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="250.967417" xlink:href="#C1_0_82f4c9fa18" y="195.448058"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="244.072482" xlink:href="#C1_0_f6794d661a" y="195.448058"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="245.311474" xlink:href="#C1_0_82f4c9fa18" y="207.506058"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="256.037815" xlink:href="#C1_0_f6794d661a" y="207.506058"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="257.977537" xlink:href="#C1_0_82f4c9fa18" y="155.859397"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="251.371631" xlink:href="#C1_0_f6794d661a" y="155.859397"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="249.741509" xlink:href="#C1_0_82f4c9fa18" y="167.644903"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="254.242134" xlink:href="#C1_0_f6794d661a" y="167.644903"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="251.885075" xlink:href="#C1_0_82f4c9fa18" y="197.39849"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="244.677001" xlink:href="#C1_0_f6794d661a" y="197.39849"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="245.411996" xlink:href="#C1_0_82f4c9fa18" y="207.505933"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="244.494458" xlink:href="#C1_0_f6794d661a" y="207.505933"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="249.876478" xlink:href="#C1_0_82f4c9fa18" y="230.966846"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="238.804304" xlink:href="#C1_0_f6794d661a" y="230.966846"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="249.267529" xlink:href="#C1_0_82f4c9fa18" y="222.105446"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="251.949688" xlink:href="#C1_0_f6794d661a" y="222.105446"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="250.386981" xlink:href="#C1_0_82f4c9fa18" y="224.506452"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="252.156096" xlink:href="#C1_0_f6794d661a" y="224.506452"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="257.542629" xlink:href="#C1_0_82f4c9fa18" y="195.44702"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="253.905468" xlink:href="#C1_0_f6794d661a" y="195.44702"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="241.847823" xlink:href="#C1_0_82f4c9fa18" y="203.859105"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="245.805328" xlink:href="#C1_0_f6794d661a" y="203.859105"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="251.643648" xlink:href="#C1_0_82f4c9fa18" y="253.550718"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="256.576606" xlink:href="#C1_0_f6794d661a" y="253.550718"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="253.504963" xlink:href="#C1_0_82f4c9fa18" y="210.047256"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="240.61208" xlink:href="#C1_0_f6794d661a" y="210.047256"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="247.624951" xlink:href="#C1_0_82f4c9fa18" y="232.914387"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="246.226675" xlink:href="#C1_0_f6794d661a" y="232.914387"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="243.410223" xlink:href="#C1_0_82f4c9fa18" y="184.051726"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="258.081928" xlink:href="#C1_0_f6794d661a" y="184.051726"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="259.218639" xlink:href="#C1_0_82f4c9fa18" y="271.697041"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="241.362785" xlink:href="#C1_0_f6794d661a" y="271.697041"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="255.752941" xlink:href="#C1_0_82f4c9fa18" y="287.550138"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="245.454037" xlink:href="#C1_0_f6794d661a" y="287.550138"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="246.691099" xlink:href="#C1_0_82f4c9fa18" y="171.991517"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="239.216934" xlink:href="#C1_0_f6794d661a" y="171.991517"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="245.677676" xlink:href="#C1_0_82f4c9fa18" y="289.813077"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="242.0264" xlink:href="#C1_0_f6794d661a" y="289.813077"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="238.009157" xlink:href="#C1_0_82f4c9fa18" y="258.492772"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="248.758797" xlink:href="#C1_0_f6794d661a" y="258.492772"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="255.409151" xlink:href="#C1_0_82f4c9fa18" y="207.501413"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="254.971773" xlink:href="#C1_0_f6794d661a" y="207.501413"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="256.322911" xlink:href="#C1_0_82f4c9fa18" y="265.61407"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="252.603318" xlink:href="#C1_0_f6794d661a" y="265.61407"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="253.535011" xlink:href="#C1_0_82f4c9fa18" y="171.639666"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="248.915537" xlink:href="#C1_0_f6794d661a" y="171.639666"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="248.705316" xlink:href="#C1_0_82f4c9fa18" y="178.461475"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="244.662664" xlink:href="#C1_0_f6794d661a" y="178.461475"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="243.204981" xlink:href="#C1_0_82f4c9fa18" y="222.123213"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="240.596198" xlink:href="#C1_0_f6794d661a" y="222.123213"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="254.869873" xlink:href="#C1_0_82f4c9fa18" y="176.05467"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="237.829611" xlink:href="#C1_0_f6794d661a" y="176.05467"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="242.275766" xlink:href="#C1_0_82f4c9fa18" y="163.998748"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="240.829997" xlink:href="#C1_0_f6794d661a" y="163.998748"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="243.833897" xlink:href="#C1_0_82f4c9fa18" y="143.801762"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="260.156396" xlink:href="#C1_0_f6794d661a" y="143.801762"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="242.344419" xlink:href="#C1_0_82f4c9fa18" y="220.082416"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="247.596949" xlink:href="#C1_0_f6794d661a" y="220.082416"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="246.956265" xlink:href="#C1_0_82f4c9fa18" y="241.497113"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="246.234432" xlink:href="#C1_0_f6794d661a" y="241.497113"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="243.153733" xlink:href="#C1_0_82f4c9fa18" y="200.69404"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="247.471406" xlink:href="#C1_0_f6794d661a" y="200.69404"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="259.925785" xlink:href="#C1_0_82f4c9fa18" y="220.085211"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="257.72277" xlink:href="#C1_0_f6794d661a" y="220.085211"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="258.022917" xlink:href="#C1_0_82f4c9fa18" y="268.153783"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="252.2602" xlink:href="#C1_0_f6794d661a" y="268.153783"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="241.535526" xlink:href="#C1_0_82f4c9fa18" y="171.988598"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="249.550905" xlink:href="#C1_0_f6794d661a" y="171.988598"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="246.972551" xlink:href="#C1_0_82f4c9fa18" y="222.106407"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="251.150873" xlink:href="#C1_0_f6794d661a" y="222.106407"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="248.354133" xlink:href="#C1_0_82f4c9fa18" y="233.471627"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="260.118044" xlink:href="#C1_0_f6794d661a" y="233.471627"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="241.121248" xlink:href="#C1_0_82f4c9fa18" y="202.585053"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="253.691718" xlink:href="#C1_0_f6794d661a" y="202.585053"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="246.162138" xlink:href="#C1_0_82f4c9fa18" y="249.918429"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="259.128552" xlink:href="#C1_0_f6794d661a" y="249.918429"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="251.462552" xlink:href="#C1_0_82f4c9fa18" y="211.497648"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="240.52906" xlink:href="#C1_0_f6794d661a" y="211.497648"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="242.691366" xlink:href="#C1_0_82f4c9fa18" y="211.505885"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="253.030406" xlink:href="#C1_0_f6794d661a" y="211.505885"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="240.791723" xlink:href="#C1_0_82f4c9fa18" y="200.390168"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="237.810066" xlink:href="#C1_0_f6794d661a" y="200.390168"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="254.64912" xlink:href="#C1_0_82f4c9fa18" y="247.561017"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="242.260588" xlink:href="#C1_0_f6794d661a" y="247.561017"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="240.04065" xlink:href="#C1_0_82f4c9fa18" y="184.063387"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="236.739356" xlink:href="#C1_0_f6794d661a" y="184.063387"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="253.187488" xlink:href="#C1_0_82f4c9fa18" y="201.910724"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="257.001252" xlink:href="#C1_0_f6794d661a" y="201.910724"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="239.892017" xlink:href="#C1_0_82f4c9fa18" y="258.1914"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="244.68886" xlink:href="#C1_0_f6794d661a" y="258.1914"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="253.716625" xlink:href="#C1_0_82f4c9fa18" y="244.087044"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="256.303264" xlink:href="#C1_0_f6794d661a" y="244.087044"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="244.541493" xlink:href="#C1_0_82f4c9fa18" y="183.386724"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="248.583029" xlink:href="#C1_0_f6794d661a" y="183.386724"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="253.082695" xlink:href="#C1_0_82f4c9fa18" y="278.975137"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="240.825724" xlink:href="#C1_0_f6794d661a" y="278.975137"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="258.763335" xlink:href="#C1_0_82f4c9fa18" y="163.490425"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="250.410914" xlink:href="#C1_0_f6794d661a" y="163.490425"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="253.603683" xlink:href="#C1_0_82f4c9fa18" y="161.44896"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="245.048643" xlink:href="#C1_0_f6794d661a" y="161.44896"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="239.982136" xlink:href="#C1_0_82f4c9fa18" y="155.863199"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="240.884275" xlink:href="#C1_0_f6794d661a" y="155.863199"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="251.264943" xlink:href="#C1_0_82f4c9fa18" y="207.501133"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="254.297786" xlink:href="#C1_0_f6794d661a" y="207.501133"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="250.365584" xlink:href="#C1_0_82f4c9fa18" y="178.433513"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="251.591911" xlink:href="#C1_0_f6794d661a" y="178.433513"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="251.590567" xlink:href="#C1_0_82f4c9fa18" y="178.427016"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="249.190229" xlink:href="#C1_0_f6794d661a" y="178.427016"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="251.802738" xlink:href="#C1_0_82f4c9fa18" y="233.273023"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="256.337985" xlink:href="#C1_0_f6794d661a" y="233.273023"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="257.942809" xlink:href="#C1_0_82f4c9fa18" y="248.627482"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="237.841219" xlink:href="#C1_0_f6794d661a" y="248.627482"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="250.527294" xlink:href="#C1_0_82f4c9fa18" y="222.088465"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="238.131314" xlink:href="#C1_0_f6794d661a" y="222.088465"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="255.099916" xlink:href="#C1_0_82f4c9fa18" y="201.897635"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="241.706771" xlink:href="#C1_0_f6794d661a" y="201.897635"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="252.681877" xlink:href="#C1_0_82f4c9fa18" y="249.915839"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="240.919088" xlink:href="#C1_0_f6794d661a" y="249.915839"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="235.970132" xlink:href="#C1_0_82f4c9fa18" y="180.040151"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="244.096836" xlink:href="#C1_0_f6794d661a" y="180.040151"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="248.750732" xlink:href="#C1_0_82f4c9fa18" y="238.59991"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="239.128033" xlink:href="#C1_0_f6794d661a" y="238.59991"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="256.992603" xlink:href="#C1_0_82f4c9fa18" y="146.032382"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="256.853443" xlink:href="#C1_0_f6794d661a" y="146.032382"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="244.922454" xlink:href="#C1_0_82f4c9fa18" y="149.38701"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="254.542932" xlink:href="#C1_0_f6794d661a" y="149.38701"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="244.432822" xlink:href="#C1_0_82f4c9fa18" y="224.492239"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="246.875296" xlink:href="#C1_0_f6794d661a" y="224.492239"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="260.006577" xlink:href="#C1_0_82f4c9fa18" y="224.506833"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="242.851065" xlink:href="#C1_0_f6794d661a" y="224.506833"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="247.460189" xlink:href="#C1_0_82f4c9fa18" y="184.918602"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="247.258302" xlink:href="#C1_0_f6794d661a" y="184.918602"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="246.05976" xlink:href="#C1_0_82f4c9fa18" y="252.663849"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="248.662857" xlink:href="#C1_0_f6794d661a" y="252.663849"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="240.869412" xlink:href="#C1_0_82f4c9fa18" y="153.031341"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="257.97259" xlink:href="#C1_0_f6794d661a" y="153.031341"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="248.055513" xlink:href="#C1_0_82f4c9fa18" y="171.975546"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="254.173867" xlink:href="#C1_0_f6794d661a" y="171.975546"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="251.274615" xlink:href="#C1_0_82f4c9fa18" y="176.049856"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="245.272048" xlink:href="#C1_0_f6794d661a" y="176.049856"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="240.338251" xlink:href="#C1_0_82f4c9fa18" y="268.154761"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="250.048883" xlink:href="#C1_0_f6794d661a" y="268.154761"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="249.971481" xlink:href="#C1_0_82f4c9fa18" y="197.38884"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="242.563136" xlink:href="#C1_0_f6794d661a" y="197.38884"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="254.302549" xlink:href="#C1_0_82f4c9fa18" y="186.099251"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="247.123948" xlink:href="#C1_0_f6794d661a" y="186.099251"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="237.610395" xlink:href="#C1_0_82f4c9fa18" y="153.052503"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="239.687678" xlink:href="#C1_0_f6794d661a" y="153.052503"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="242.188487" xlink:href="#C1_0_82f4c9fa18" y="224.54389"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="245.92367" xlink:href="#C1_0_f6794d661a" y="224.54389"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="242.588526" xlink:href="#C1_0_82f4c9fa18" y="195.434979"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="251.74373" xlink:href="#C1_0_f6794d661a" y="195.434979"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="256.111274" xlink:href="#C1_0_82f4c9fa18" y="161.464013"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="258.764094" xlink:href="#C1_0_f6794d661a" y="161.464013"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="249.208031" xlink:href="#C1_0_82f4c9fa18" y="193.04828"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="258.285678" xlink:href="#C1_0_f6794d661a" y="193.04828"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="258.558145" xlink:href="#C1_0_82f4c9fa18" y="186.0958"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="251.073876" xlink:href="#C1_0_f6794d661a" y="186.0958"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="250.710288" xlink:href="#C1_0_82f4c9fa18" y="264.735803"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="242.682328" xlink:href="#C1_0_f6794d661a" y="264.735803"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="241.517366" xlink:href="#C1_0_82f4c9fa18" y="200.392373"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="249.524376" xlink:href="#C1_0_f6794d661a" y="200.392373"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="256.88586" xlink:href="#C1_0_82f4c9fa18" y="143.842965"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="239.822537" xlink:href="#C1_0_f6794d661a" y="143.842965"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="237.862394" xlink:href="#C1_0_82f4c9fa18" y="195.446167"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="239.374956" xlink:href="#C1_0_f6794d661a" y="195.446167"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="236.958222" xlink:href="#C1_0_82f4c9fa18" y="207.513801"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="257.654684" xlink:href="#C1_0_f6794d661a" y="207.513801"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="253.008381" xlink:href="#C1_0_82f4c9fa18" y="186.580203"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="248.611767" xlink:href="#C1_0_f6794d661a" y="186.580203"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="250.085054" xlink:href="#C1_0_82f4c9fa18" y="253.570174"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="243.333949" xlink:href="#C1_0_f6794d661a" y="253.570174"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="259.836948" xlink:href="#C1_0_82f4c9fa18" y="241.537021"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="258.724275" xlink:href="#C1_0_f6794d661a" y="241.537021"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="253.924423" xlink:href="#C1_0_82f4c9fa18" y="160.22007"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="242.255379" xlink:href="#C1_0_f6794d661a" y="160.22007"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="256.382545" xlink:href="#C1_0_82f4c9fa18" y="239.143428"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="254.189263" xlink:href="#C1_0_f6794d661a" y="239.143428"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="251.707675" xlink:href="#C1_0_82f4c9fa18" y="169.104835"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="238.735356" xlink:href="#C1_0_f6794d661a" y="169.104835"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="241.675864" xlink:href="#C1_0_82f4c9fa18" y="258.507112"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="235.940039" xlink:href="#C1_0_f6794d661a" y="258.507112"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="244.466881" xlink:href="#C1_0_82f4c9fa18" y="186.110956"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="250.443807" xlink:href="#C1_0_f6794d661a" y="186.110956"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="252.121767" xlink:href="#C1_0_82f4c9fa18" y="120.358149"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="247.848001" xlink:href="#C1_0_f6794d661a" y="120.358149"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="244.132684" xlink:href="#C1_0_82f4c9fa18" y="222.123514"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="254.562908" xlink:href="#C1_0_f6794d661a" y="222.123514"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="257.174049" xlink:href="#C1_0_82f4c9fa18" y="193.081391"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="237.101102" xlink:href="#C1_0_f6794d661a" y="193.081391"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="258.748591" xlink:href="#C1_0_82f4c9fa18" y="171.995159"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="254.342073" xlink:href="#C1_0_f6794d661a" y="171.995159"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="247.30306" xlink:href="#C1_0_82f4c9fa18" y="178.46614"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="244.727938" xlink:href="#C1_0_f6794d661a" y="178.46614"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="248.181746" xlink:href="#C1_0_82f4c9fa18" y="253.570175"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="248.722821" xlink:href="#C1_0_f6794d661a" y="253.570175"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="257.373042" xlink:href="#C1_0_82f4c9fa18" y="239.122481"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="253.220497" xlink:href="#C1_0_f6794d661a" y="239.122481"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="252.161833" xlink:href="#C1_0_82f4c9fa18" y="138.900323"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="257.12959" xlink:href="#C1_0_f6794d661a" y="138.900323"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="243.413596" xlink:href="#C1_0_82f4c9fa18" y="271.721863"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="248.472292" xlink:href="#C1_0_f6794d661a" y="271.721863"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="247.154204" xlink:href="#C1_0_82f4c9fa18" y="249.923002"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="251.094287" xlink:href="#C1_0_f6794d661a" y="249.923002"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="242.700389" xlink:href="#C1_0_82f4c9fa18" y="166.399388"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="239.608628" xlink:href="#C1_0_f6794d661a" y="166.399388"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="251.882565" xlink:href="#C1_0_82f4c9fa18" y="180.990652"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="242.041104" xlink:href="#C1_0_f6794d661a" y="180.990652"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="253.812388" xlink:href="#C1_0_82f4c9fa18" y="249.91293"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="257.851049" xlink:href="#C1_0_f6794d661a" y="249.91293"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="237.048773" xlink:href="#C1_0_82f4c9fa18" y="253.586027"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="260.052025" xlink:href="#C1_0_f6794d661a" y="253.586027"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="254.488895" xlink:href="#C1_0_82f4c9fa18" y="253.578104"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="241.787346" xlink:href="#C1_0_f6794d661a" y="253.578104"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="244.277306" xlink:href="#C1_0_82f4c9fa18" y="268.18575"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="242.785547" xlink:href="#C1_0_f6794d661a" y="268.18575"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="253.029259" xlink:href="#C1_0_82f4c9fa18" y="207.533096"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="257.383617" xlink:href="#C1_0_f6794d661a" y="207.533096"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="237.05314" xlink:href="#C1_0_82f4c9fa18" y="239.131686"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="255.954924" xlink:href="#C1_0_f6794d661a" y="239.131686"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="243.583859" xlink:href="#C1_0_82f4c9fa18" y="169.123327"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="256.660227" xlink:href="#C1_0_f6794d661a" y="169.123327"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="237.199561" xlink:href="#C1_0_82f4c9fa18" y="230.981106"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="253.375578" xlink:href="#C1_0_f6794d661a" y="230.981106"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="238.832207" xlink:href="#C1_0_82f4c9fa18" y="310.800003"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="239.424911" xlink:href="#C1_0_f6794d661a" y="310.800003"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="240.281005" xlink:href="#C1_0_82f4c9fa18" y="222.125064"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="249.343377" xlink:href="#C1_0_f6794d661a" y="222.125064"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="259.358059" xlink:href="#C1_0_82f4c9fa18" y="285.451085"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="244.287032" xlink:href="#C1_0_f6794d661a" y="285.451085"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="256.467924" xlink:href="#C1_0_82f4c9fa18" y="219.21394"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="238.650188" xlink:href="#C1_0_f6794d661a" y="219.21394"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="250.187038" xlink:href="#C1_0_82f4c9fa18" y="180.057066"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="245.933532" xlink:href="#C1_0_f6794d661a" y="180.057066"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="240.165863" xlink:href="#C1_0_82f4c9fa18" y="222.131073"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="249.387768" xlink:href="#C1_0_f6794d661a" y="222.131073"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="244.30431" xlink:href="#C1_0_82f4c9fa18" y="220.868179"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="243.650988" xlink:href="#C1_0_f6794d661a" y="220.868179"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="258.499304" xlink:href="#C1_0_82f4c9fa18" y="171.94305"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="252.576129" xlink:href="#C1_0_f6794d661a" y="171.94305"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="259.50988" xlink:href="#C1_0_82f4c9fa18" y="123.973646"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="245.055722" xlink:href="#C1_0_f6794d661a" y="123.973646"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="251.818792" xlink:href="#C1_0_82f4c9fa18" y="253.57659"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="258.271907" xlink:href="#C1_0_f6794d661a" y="253.57659"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="245.586634" xlink:href="#C1_0_82f4c9fa18" y="195.39986"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="247.881776" xlink:href="#C1_0_f6794d661a" y="195.39986"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="244.260205" xlink:href="#C1_0_82f4c9fa18" y="207.543075"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="244.830856" xlink:href="#C1_0_f6794d661a" y="207.543075"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="249.581326" xlink:href="#C1_0_82f4c9fa18" y="112.153263"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="250.832433" xlink:href="#C1_0_f6794d661a" y="112.153263"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="259.664889" xlink:href="#C1_0_82f4c9fa18" y="232.914524"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="259.171605" xlink:href="#C1_0_f6794d661a" y="232.914524"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="254.113023" xlink:href="#C1_0_82f4c9fa18" y="214.860694"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="242.932513" xlink:href="#C1_0_f6794d661a" y="214.860694"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="237.050506" xlink:href="#C1_0_82f4c9fa18" y="248.654279"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="259.907996" xlink:href="#C1_0_f6794d661a" y="248.654279"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="255.774791" xlink:href="#C1_0_82f4c9fa18" y="195.468533"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="257.452508" xlink:href="#C1_0_f6794d661a" y="195.468533"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="256.928251" xlink:href="#C1_0_82f4c9fa18" y="214.527426"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="241.193256" xlink:href="#C1_0_f6794d661a" y="214.527426"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="242.931765" xlink:href="#C1_0_82f4c9fa18" y="230.245093"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="241.544543" xlink:href="#C1_0_f6794d661a" y="230.245093"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="251.561055" xlink:href="#C1_0_82f4c9fa18" y="220.863447"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="243.046826" xlink:href="#C1_0_f6794d661a" y="220.863447"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="258.702734" xlink:href="#C1_0_82f4c9fa18" y="193.053335"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="247.020449" xlink:href="#C1_0_f6794d661a" y="193.053335"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="255.558219" xlink:href="#C1_0_82f4c9fa18" y="182.446898"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="238.62372" xlink:href="#C1_0_f6794d661a" y="182.446898"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="257.124961" xlink:href="#C1_0_82f4c9fa18" y="253.557128"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="258.845369" xlink:href="#C1_0_f6794d661a" y="253.557128"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="252.780081" xlink:href="#C1_0_82f4c9fa18" y="207.529824"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="249.089013" xlink:href="#C1_0_f6794d661a" y="207.529824"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="258.133333" xlink:href="#C1_0_82f4c9fa18" y="264.752642"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="251.840956" xlink:href="#C1_0_f6794d661a" y="264.752642"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="250.348547" xlink:href="#C1_0_82f4c9fa18" y="224.503275"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="246.209373" xlink:href="#C1_0_f6794d661a" y="224.503275"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="248.947053" xlink:href="#C1_0_82f4c9fa18" y="253.55706"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="241.004185" xlink:href="#C1_0_f6794d661a" y="253.55706"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="241.86021" xlink:href="#C1_0_82f4c9fa18" y="268.156745"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="251.733384" xlink:href="#C1_0_f6794d661a" y="268.156745"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="259.504468" xlink:href="#C1_0_82f4c9fa18" y="153.078774"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="249.795698" xlink:href="#C1_0_f6794d661a" y="153.078774"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="257.110999" xlink:href="#C1_0_82f4c9fa18" y="200.188809"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="241.631853" xlink:href="#C1_0_f6794d661a" y="200.188809"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="246.258515" xlink:href="#C1_0_82f4c9fa18" y="181.218033"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="255.722131" xlink:href="#C1_0_f6794d661a" y="181.218033"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="252.720844" xlink:href="#C1_0_82f4c9fa18" y="268.192474"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="252.511781" xlink:href="#C1_0_f6794d661a" y="268.192474"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="237.477009" xlink:href="#C1_0_82f4c9fa18" y="220.875907"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="239.655477" xlink:href="#C1_0_f6794d661a" y="220.875907"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="236.124965" xlink:href="#C1_0_82f4c9fa18" y="143.848251"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="255.613208" xlink:href="#C1_0_f6794d661a" y="143.848251"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="258.487344" xlink:href="#C1_0_82f4c9fa18" y="253.560292"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="240.031487" xlink:href="#C1_0_f6794d661a" y="253.560292"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="241.038911" xlink:href="#C1_0_82f4c9fa18" y="224.52231"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="249.407197" xlink:href="#C1_0_f6794d661a" y="224.52231"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="252.643726" xlink:href="#C1_0_82f4c9fa18" y="97.771411"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="248.087805" xlink:href="#C1_0_f6794d661a" y="97.771411"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="251.967058" xlink:href="#C1_0_82f4c9fa18" y="140.047813"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="243.026387" xlink:href="#C1_0_f6794d661a" y="140.047813"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="253.680507" xlink:href="#C1_0_82f4c9fa18" y="222.110402"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="256.489812" xlink:href="#C1_0_f6794d661a" y="222.110402"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="254.282556" xlink:href="#C1_0_82f4c9fa18" y="211.505449"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="251.073362" xlink:href="#C1_0_f6794d661a" y="211.505449"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="249.843732" xlink:href="#C1_0_82f4c9fa18" y="281.728657"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="250.747585" xlink:href="#C1_0_f6794d661a" y="281.728657"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="240.221786" xlink:href="#C1_0_82f4c9fa18" y="300.758372"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="240.584159" xlink:href="#C1_0_f6794d661a" y="300.758372"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="248.869938" xlink:href="#C1_0_82f4c9fa18" y="252.696176"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="252.337546" xlink:href="#C1_0_f6794d661a" y="252.696176"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="256.705209" xlink:href="#C1_0_82f4c9fa18" y="268.160248"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="243.575031" xlink:href="#C1_0_f6794d661a" y="268.160248"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="248.336187" xlink:href="#C1_0_82f4c9fa18" y="195.468478"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="243.470738" xlink:href="#C1_0_f6794d661a" y="195.468478"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="251.424509" xlink:href="#C1_0_82f4c9fa18" y="166.421525"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="237.478428" xlink:href="#C1_0_f6794d661a" y="166.421525"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="249.762694" xlink:href="#C1_0_82f4c9fa18" y="254.74119"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="260.063666" xlink:href="#C1_0_f6794d661a" y="254.74119"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="250.402907" xlink:href="#C1_0_82f4c9fa18" y="196.273547"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="249.749826" xlink:href="#C1_0_f6794d661a" y="196.273547"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="252.583235" xlink:href="#C1_0_82f4c9fa18" y="207.510713"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="239.645129" xlink:href="#C1_0_f6794d661a" y="207.510713"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="239.201727" xlink:href="#C1_0_82f4c9fa18" y="171.63853"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="255.408884" xlink:href="#C1_0_f6794d661a" y="171.63853"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="253.584778" xlink:href="#C1_0_82f4c9fa18" y="239.124599"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="241.551166" xlink:href="#C1_0_f6794d661a" y="239.124599"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="239.746061" xlink:href="#C1_0_82f4c9fa18" y="180.99247"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="253.657239" xlink:href="#C1_0_f6794d661a" y="180.99247"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="256.518422" xlink:href="#C1_0_82f4c9fa18" y="258.514464"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="248.847057" xlink:href="#C1_0_f6794d661a" y="258.514464"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="238.00373" xlink:href="#C1_0_82f4c9fa18" y="207.515692"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="237.949501" xlink:href="#C1_0_f6794d661a" y="207.515692"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="245.202313" xlink:href="#C1_0_82f4c9fa18" y="157.067408"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="252.983725" xlink:href="#C1_0_f6794d661a" y="157.067408"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="252.639654" xlink:href="#C1_0_82f4c9fa18" y="296.59294"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="257.586564" xlink:href="#C1_0_f6794d661a" y="296.59294"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="244.352236" xlink:href="#C1_0_82f4c9fa18" y="184.939129"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="245.766564" xlink:href="#C1_0_f6794d661a" y="184.939129"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="252.431817" xlink:href="#C1_0_82f4c9fa18" y="203.879793"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="251.702575" xlink:href="#C1_0_f6794d661a" y="203.879793"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="239.319924" xlink:href="#C1_0_82f4c9fa18" y="161.476812"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="251.839272" xlink:href="#C1_0_f6794d661a" y="161.476812"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="256.393049" xlink:href="#C1_0_82f4c9fa18" y="157.037807"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="240.664193" xlink:href="#C1_0_f6794d661a" y="157.037807"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="253.633429" xlink:href="#C1_0_82f4c9fa18" y="239.121704"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="245.897539" xlink:href="#C1_0_f6794d661a" y="239.121704"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="251.37943" xlink:href="#C1_0_82f4c9fa18" y="195.478211"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="248.704589" xlink:href="#C1_0_f6794d661a" y="195.478211"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="238.811921" xlink:href="#C1_0_82f4c9fa18" y="212.463793"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="251.190932" xlink:href="#C1_0_f6794d661a" y="212.463793"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="236.718956" xlink:href="#C1_0_82f4c9fa18" y="171.653835"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="240.410848" xlink:href="#C1_0_f6794d661a" y="171.653835"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="257.770529" xlink:href="#C1_0_82f4c9fa18" y="264.281748"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="247.082909" xlink:href="#C1_0_f6794d661a" y="264.281748"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="245.479865" xlink:href="#C1_0_82f4c9fa18" y="224.535933"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="248.19913" xlink:href="#C1_0_f6794d661a" y="224.535933"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="245.82339" xlink:href="#C1_0_82f4c9fa18" y="195.474854"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="239.345103" xlink:href="#C1_0_f6794d661a" y="195.474854"/> </g> </g> <g id="PathCollection_5"> @@ -2039,277 +2039,277 @@ C -2.236584 -1.29895 -2.5 -0.663008 -2.5 0 C -2.5 0.663008 -2.236584 1.29895 -1.767767 1.767767 C -1.29895 2.236584 -0.663008 2.5 0 2.5 z -" id="C2_0_e29aafa51a"/> +" id="C2_0_832c39148d"/> </defs> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="390.84124" xlink:href="#C2_0_e29aafa51a" y="149.690246"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="390.475999" xlink:href="#C2_0_832c39148d" y="149.690246"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="405.346413" xlink:href="#C2_0_e29aafa51a" y="207.044912"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="397.965421" xlink:href="#C2_0_832c39148d" y="207.044912"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="393.88288" xlink:href="#C2_0_e29aafa51a" y="246.208"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="393.484872" xlink:href="#C2_0_832c39148d" y="246.208"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="406.578652" xlink:href="#C2_0_e29aafa51a" y="170.287749"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="390.724552" xlink:href="#C2_0_832c39148d" y="170.287749"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="389.621541" xlink:href="#C2_0_e29aafa51a" y="248.746795"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="382.201112" xlink:href="#C2_0_832c39148d" y="248.746795"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="387.448552" xlink:href="#C2_0_e29aafa51a" y="210.338243"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="391.878186" xlink:href="#C2_0_832c39148d" y="210.338243"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="405.604467" xlink:href="#C2_0_e29aafa51a" y="207.488934"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="401.75448" xlink:href="#C2_0_832c39148d" y="207.488934"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="389.459367" xlink:href="#C2_0_e29aafa51a" y="231.75118"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="392.341688" xlink:href="#C2_0_832c39148d" y="231.75118"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="396.200572" xlink:href="#C2_0_e29aafa51a" y="183.680963"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="397.036616" xlink:href="#C2_0_832c39148d" y="183.680963"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="385.15736" xlink:href="#C2_0_e29aafa51a" y="168.833673"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="398.36239" xlink:href="#C2_0_832c39148d" y="168.833673"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="406.122402" xlink:href="#C2_0_e29aafa51a" y="219.69547"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="404.938844" xlink:href="#C2_0_832c39148d" y="219.69547"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="383.552829" xlink:href="#C2_0_e29aafa51a" y="225.28767"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="398.746544" xlink:href="#C2_0_832c39148d" y="225.28767"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="405.77538" xlink:href="#C2_0_e29aafa51a" y="228.327559"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="397.640305" xlink:href="#C2_0_832c39148d" y="228.327559"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="388.489197" xlink:href="#C2_0_e29aafa51a" y="168.835615"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="387.393832" xlink:href="#C2_0_832c39148d" y="168.835615"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="399.468616" xlink:href="#C2_0_e29aafa51a" y="263.201059"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="388.367034" xlink:href="#C2_0_832c39148d" y="263.201059"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="387.503315" xlink:href="#C2_0_e29aafa51a" y="210.337366"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="393.12258" xlink:href="#C2_0_832c39148d" y="210.337366"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="386.225279" xlink:href="#C2_0_e29aafa51a" y="260.803466"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="385.188585" xlink:href="#C2_0_832c39148d" y="260.803466"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="390.749613" xlink:href="#C2_0_e29aafa51a" y="181.098047"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="382.219913" xlink:href="#C2_0_832c39148d" y="181.098047"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="392.403273" xlink:href="#C2_0_e29aafa51a" y="203.81402"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="392.134443" xlink:href="#C2_0_832c39148d" y="203.81402"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="386.084369" xlink:href="#C2_0_e29aafa51a" y="141.231824"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="390.362719" xlink:href="#C2_0_832c39148d" y="141.231824"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="389.526552" xlink:href="#C2_0_e29aafa51a" y="144.092936"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="404.109521" xlink:href="#C2_0_832c39148d" y="144.092936"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="383.975172" xlink:href="#C2_0_e29aafa51a" y="317.660183"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="406.491078" xlink:href="#C2_0_832c39148d" y="317.660183"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="393.41676" xlink:href="#C2_0_e29aafa51a" y="260.803345"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="398.879198" xlink:href="#C2_0_832c39148d" y="260.803345"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="398.153593" xlink:href="#C2_0_e29aafa51a" y="169.018987"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="403.645853" xlink:href="#C2_0_832c39148d" y="169.018987"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="395.7849" xlink:href="#C2_0_e29aafa51a" y="261.107859"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="387.422005" xlink:href="#C2_0_832c39148d" y="261.107859"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="393.088997" xlink:href="#C2_0_e29aafa51a" y="188.398313"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="402.909115" xlink:href="#C2_0_832c39148d" y="188.398313"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="404.482785" xlink:href="#C2_0_e29aafa51a" y="196.814829"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="399.962593" xlink:href="#C2_0_832c39148d" y="196.814829"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="394.651677" xlink:href="#C2_0_e29aafa51a" y="169.138409"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="401.18723" xlink:href="#C2_0_832c39148d" y="169.138409"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="395.221958" xlink:href="#C2_0_e29aafa51a" y="217.452616"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="405.58835" xlink:href="#C2_0_832c39148d" y="217.452616"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="395.822051" xlink:href="#C2_0_e29aafa51a" y="150.000846"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="399.980208" xlink:href="#C2_0_832c39148d" y="150.000846"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="387.132602" xlink:href="#C2_0_e29aafa51a" y="164.953349"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="386.345534" xlink:href="#C2_0_832c39148d" y="164.953349"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="402.280259" xlink:href="#C2_0_e29aafa51a" y="204.463925"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="397.447235" xlink:href="#C2_0_832c39148d" y="204.463925"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="387.228369" xlink:href="#C2_0_e29aafa51a" y="239.726796"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="396.305339" xlink:href="#C2_0_832c39148d" y="239.726796"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="385.624783" xlink:href="#C2_0_e29aafa51a" y="181.308809"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="382.908606" xlink:href="#C2_0_832c39148d" y="181.308809"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="398.622557" xlink:href="#C2_0_e29aafa51a" y="263.495117"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="388.163306" xlink:href="#C2_0_832c39148d" y="263.495117"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="389.655695" xlink:href="#C2_0_e29aafa51a" y="250.536382"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="391.681741" xlink:href="#C2_0_832c39148d" y="250.536382"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="394.237926" xlink:href="#C2_0_e29aafa51a" y="173.978374"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="384.30987" xlink:href="#C2_0_832c39148d" y="173.978374"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="389.015028" xlink:href="#C2_0_e29aafa51a" y="249.873362"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="397.423924" xlink:href="#C2_0_832c39148d" y="249.873362"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="390.194567" xlink:href="#C2_0_e29aafa51a" y="190.95988"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="404.836804" xlink:href="#C2_0_832c39148d" y="190.95988"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="404.660691" xlink:href="#C2_0_e29aafa51a" y="181.621632"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="389.189746" xlink:href="#C2_0_832c39148d" y="181.621632"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="397.577362" xlink:href="#C2_0_e29aafa51a" y="288.920969"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="391.157691" xlink:href="#C2_0_832c39148d" y="288.920969"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="386.870298" xlink:href="#C2_0_e29aafa51a" y="269.996681"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="387.354986" xlink:href="#C2_0_832c39148d" y="269.996681"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="406.646185" xlink:href="#C2_0_e29aafa51a" y="177.849032"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="401.82034" xlink:href="#C2_0_832c39148d" y="177.849032"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="385.814516" xlink:href="#C2_0_e29aafa51a" y="263.534152"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="394.00015" xlink:href="#C2_0_832c39148d" y="263.534152"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="397.667787" xlink:href="#C2_0_e29aafa51a" y="181.617863"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="403.797191" xlink:href="#C2_0_832c39148d" y="181.617863"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="397.07881" xlink:href="#C2_0_e29aafa51a" y="173.480755"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="395.449402" xlink:href="#C2_0_832c39148d" y="173.480755"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="400.978708" xlink:href="#C2_0_e29aafa51a" y="213.057381"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="386.941419" xlink:href="#C2_0_832c39148d" y="213.057381"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="403.436157" xlink:href="#C2_0_e29aafa51a" y="167.503409"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="397.79978" xlink:href="#C2_0_832c39148d" y="167.503409"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="397.997283" xlink:href="#C2_0_e29aafa51a" y="269.954036"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="387.778748" xlink:href="#C2_0_832c39148d" y="269.954036"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="393.433251" xlink:href="#C2_0_e29aafa51a" y="173.788682"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="387.307693" xlink:href="#C2_0_832c39148d" y="173.788682"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="393.550647" xlink:href="#C2_0_e29aafa51a" y="292.587365"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="404.853782" xlink:href="#C2_0_832c39148d" y="292.587365"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="386.15666" xlink:href="#C2_0_e29aafa51a" y="211.845864"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="387.110946" xlink:href="#C2_0_832c39148d" y="211.845864"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="395.829875" xlink:href="#C2_0_e29aafa51a" y="159.334573"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="402.396676" xlink:href="#C2_0_832c39148d" y="159.334573"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="391.076298" xlink:href="#C2_0_e29aafa51a" y="234.468961"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="384.700241" xlink:href="#C2_0_832c39148d" y="234.468961"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="392.892075" xlink:href="#C2_0_e29aafa51a" y="292.577101"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="401.454188" xlink:href="#C2_0_832c39148d" y="292.577101"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="403.328802" xlink:href="#C2_0_e29aafa51a" y="169.126912"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="388.92847" xlink:href="#C2_0_832c39148d" y="169.126912"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="384.642931" xlink:href="#C2_0_e29aafa51a" y="184.761682"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="395.76762" xlink:href="#C2_0_832c39148d" y="184.761682"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="406.646808" xlink:href="#C2_0_e29aafa51a" y="192.032499"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="387.919868" xlink:href="#C2_0_832c39148d" y="192.032499"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="387.428216" xlink:href="#C2_0_e29aafa51a" y="182.627554"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="390.590157" xlink:href="#C2_0_832c39148d" y="182.627554"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="388.185244" xlink:href="#C2_0_e29aafa51a" y="193.649931"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="399.092373" xlink:href="#C2_0_832c39148d" y="193.649931"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="387.248514" xlink:href="#C2_0_e29aafa51a" y="199.642913"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="405.863955" xlink:href="#C2_0_832c39148d" y="199.642913"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="396.628793" xlink:href="#C2_0_e29aafa51a" y="198.949705"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="385.575148" xlink:href="#C2_0_832c39148d" y="198.949705"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="384.284636" xlink:href="#C2_0_e29aafa51a" y="251.432621"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="385.166954" xlink:href="#C2_0_832c39148d" y="251.432621"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="392.275744" xlink:href="#C2_0_e29aafa51a" y="214.175328"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="391.077134" xlink:href="#C2_0_832c39148d" y="214.175328"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="406.65758" xlink:href="#C2_0_e29aafa51a" y="210.239489"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="384.744497" xlink:href="#C2_0_832c39148d" y="210.239489"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="393.260896" xlink:href="#C2_0_e29aafa51a" y="196.55794"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="395.349085" xlink:href="#C2_0_832c39148d" y="196.55794"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="389.712695" xlink:href="#C2_0_e29aafa51a" y="156.006652"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="384.327311" xlink:href="#C2_0_832c39148d" y="156.006652"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="406.594033" xlink:href="#C2_0_e29aafa51a" y="263.492625"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="390.050536" xlink:href="#C2_0_832c39148d" y="263.492625"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="389.42298" xlink:href="#C2_0_e29aafa51a" y="172.235492"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="394.096075" xlink:href="#C2_0_832c39148d" y="172.235492"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="385.888469" xlink:href="#C2_0_e29aafa51a" y="215.290772"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="382.555439" xlink:href="#C2_0_832c39148d" y="215.290772"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="405.930654" xlink:href="#C2_0_e29aafa51a" y="292.544009"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="406.385103" xlink:href="#C2_0_832c39148d" y="292.544009"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="395.917735" xlink:href="#C2_0_e29aafa51a" y="304.603374"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="387.370058" xlink:href="#C2_0_832c39148d" y="304.603374"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="398.618192" xlink:href="#C2_0_e29aafa51a" y="167.004699"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="383.888157" xlink:href="#C2_0_832c39148d" y="167.004699"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="391.897436" xlink:href="#C2_0_e29aafa51a" y="259.872502"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="405.872309" xlink:href="#C2_0_832c39148d" y="259.872502"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="394.545991" xlink:href="#C2_0_e29aafa51a" y="269.954098"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="405.243677" xlink:href="#C2_0_832c39148d" y="269.954098"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="395.492362" xlink:href="#C2_0_e29aafa51a" y="307.158181"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="383.197247" xlink:href="#C2_0_832c39148d" y="307.158181"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="395.793623" xlink:href="#C2_0_e29aafa51a" y="192.384078"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="403.445135" xlink:href="#C2_0_832c39148d" y="192.384078"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="387.666471" xlink:href="#C2_0_e29aafa51a" y="237.153489"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="382.690067" xlink:href="#C2_0_832c39148d" y="237.153489"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="404.107674" xlink:href="#C2_0_e29aafa51a" y="232.231875"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="396.516601" xlink:href="#C2_0_832c39148d" y="232.231875"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="406.10931" xlink:href="#C2_0_e29aafa51a" y="234.460146"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="385.60786" xlink:href="#C2_0_832c39148d" y="234.460146"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="382.755137" xlink:href="#C2_0_e29aafa51a" y="217.472154"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="386.807757" xlink:href="#C2_0_832c39148d" y="217.472154"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="392.761503" xlink:href="#C2_0_e29aafa51a" y="229.517557"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="394.427127" xlink:href="#C2_0_832c39148d" y="229.517557"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="387.366921" xlink:href="#C2_0_e29aafa51a" y="246.523807"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="397.382376" xlink:href="#C2_0_832c39148d" y="246.523807"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="383.203214" xlink:href="#C2_0_e29aafa51a" y="190.370071"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="404.475197" xlink:href="#C2_0_832c39148d" y="190.370071"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="390.607014" xlink:href="#C2_0_e29aafa51a" y="157.649777"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="386.150771" xlink:href="#C2_0_832c39148d" y="157.649777"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="393.449033" xlink:href="#C2_0_e29aafa51a" y="263.511599"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="390.416439" xlink:href="#C2_0_832c39148d" y="263.511599"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="399.232117" xlink:href="#C2_0_e29aafa51a" y="217.474908"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="395.497037" xlink:href="#C2_0_832c39148d" y="217.474908"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="399.286681" xlink:href="#C2_0_e29aafa51a" y="194.467998"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="399.792777" xlink:href="#C2_0_832c39148d" y="194.467998"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="385.211454" xlink:href="#C2_0_e29aafa51a" y="303.741702"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="383.287595" xlink:href="#C2_0_832c39148d" y="303.741702"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="393.815193" xlink:href="#C2_0_e29aafa51a" y="261.127375"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="383.289701" xlink:href="#C2_0_832c39148d" y="261.127375"/> </g> </g> <g id="PathCollection_6"> @@ -2324,259 +2324,259 @@ C -2.236584 -1.29895 -2.5 -0.663008 -2.5 0 C -2.5 0.663008 -2.236584 1.29895 -1.767767 1.767767 C -1.29895 2.236584 -0.663008 2.5 0 2.5 z -" id="C3_0_8e4775d9e6"/> +" id="C3_0_947d114575"/> </defs> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="484.468514" xlink:href="#C3_0_8e4775d9e6" y="149.245404"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="482.642807" xlink:href="#C3_0_947d114575" y="149.245404"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="482.754636" xlink:href="#C3_0_8e4775d9e6" y="182.290351"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="481.498165" xlink:href="#C3_0_947d114575" y="182.290351"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="480.648449" xlink:href="#C3_0_8e4775d9e6" y="207.348341"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="489.850049" xlink:href="#C3_0_947d114575" y="207.348341"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="493.587132" xlink:href="#C3_0_8e4775d9e6" y="163.846465"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="493.24087" xlink:href="#C3_0_947d114575" y="163.846465"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="482.651815" xlink:href="#C3_0_8e4775d9e6" y="238.949529"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="485.118408" xlink:href="#C3_0_947d114575" y="238.949529"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="503.345121" xlink:href="#C3_0_8e4775d9e6" y="180.844268"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="496.060057" xlink:href="#C3_0_947d114575" y="180.844268"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="503.643901" xlink:href="#C3_0_8e4775d9e6" y="229.28663"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="484.159309" xlink:href="#C3_0_947d114575" y="229.28663"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="498.201577" xlink:href="#C3_0_8e4775d9e6" y="209.895621"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="483.579613" xlink:href="#C3_0_947d114575" y="209.895621"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="500.462082" xlink:href="#C3_0_8e4775d9e6" y="186.884487"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="488.988825" xlink:href="#C3_0_947d114575" y="186.884487"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="498.614771" xlink:href="#C3_0_8e4775d9e6" y="152.898133"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="496.006158" xlink:href="#C3_0_947d114575" y="152.898133"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="493.207548" xlink:href="#C3_0_8e4775d9e6" y="221.947021"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="489.82018" xlink:href="#C3_0_947d114575" y="221.947021"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="498.883095" xlink:href="#C3_0_8e4775d9e6" y="192.905385"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="496.084735" xlink:href="#C3_0_947d114575" y="192.905385"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="495.87073" xlink:href="#C3_0_8e4775d9e6" y="214.114189"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="494.178335" xlink:href="#C3_0_947d114575" y="214.114189"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="502.743843" xlink:href="#C3_0_8e4775d9e6" y="156.888241"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="489.193843" xlink:href="#C3_0_947d114575" y="156.888241"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="480.564122" xlink:href="#C3_0_8e4775d9e6" y="224.350175"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="490.264074" xlink:href="#C3_0_947d114575" y="224.350175"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="494.082118" xlink:href="#C3_0_8e4775d9e6" y="180.846286"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="496.495753" xlink:href="#C3_0_947d114575" y="180.846286"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="490.401903" xlink:href="#C3_0_8e4775d9e6" y="209.900153"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="501.257731" xlink:href="#C3_0_947d114575" y="209.900153"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="504.435815" xlink:href="#C3_0_8e4775d9e6" y="162.586682"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="491.668964" xlink:href="#C3_0_947d114575" y="162.586682"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="484.804459" xlink:href="#C3_0_8e4775d9e6" y="193.978079"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="499.069394" xlink:href="#C3_0_947d114575" y="193.978079"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="489.856168" xlink:href="#C3_0_8e4775d9e6" y="136.580688"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="484.655298" xlink:href="#C3_0_947d114575" y="136.580688"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="503.12363" xlink:href="#C3_0_8e4775d9e6" y="143.648879"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="490.762288" xlink:href="#C3_0_947d114575" y="143.648879"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="503.746342" xlink:href="#C3_0_8e4775d9e6" y="268.008621"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="503.865569" xlink:href="#C3_0_947d114575" y="268.008621"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="496.022779" xlink:href="#C3_0_8e4775d9e6" y="154.690129"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="501.94414" xlink:href="#C3_0_947d114575" y="154.690129"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="489.531087" xlink:href="#C3_0_8e4775d9e6" y="200.736617"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="480.990598" xlink:href="#C3_0_947d114575" y="200.736617"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="490.414802" xlink:href="#C3_0_8e4775d9e6" y="186.893091"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="488.106325" xlink:href="#C3_0_947d114575" y="186.893091"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="500.874879" xlink:href="#C3_0_8e4775d9e6" y="166.431399"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="496.360536" xlink:href="#C3_0_947d114575" y="166.431399"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="489.759249" xlink:href="#C3_0_8e4775d9e6" y="174.846957"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="482.564612" xlink:href="#C3_0_947d114575" y="174.846957"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="488.083892" xlink:href="#C3_0_8e4775d9e6" y="153.062619"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="496.25633" xlink:href="#C3_0_947d114575" y="153.062619"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="481.327631" xlink:href="#C3_0_8e4775d9e6" y="161.468898"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="496.917684" xlink:href="#C3_0_947d114575" y="161.468898"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="499.433094" xlink:href="#C3_0_8e4775d9e6" y="232.925293"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="498.124995" xlink:href="#C3_0_947d114575" y="232.925293"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="501.056995" xlink:href="#C3_0_8e4775d9e6" y="180.987666"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="483.935009" xlink:href="#C3_0_947d114575" y="180.987666"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="485.418008" xlink:href="#C3_0_8e4775d9e6" y="157.977647"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="485.348265" xlink:href="#C3_0_947d114575" y="157.977647"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="480.63167" xlink:href="#C3_0_8e4775d9e6" y="253.616023"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="501.726635" xlink:href="#C3_0_947d114575" y="253.616023"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="502.891607" xlink:href="#C3_0_8e4775d9e6" y="249.897009"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="482.113209" xlink:href="#C3_0_947d114575" y="249.897009"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="493.907014" xlink:href="#C3_0_8e4775d9e6" y="180.981605"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="488.976452" xlink:href="#C3_0_947d114575" y="180.981605"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="503.613412" xlink:href="#C3_0_8e4775d9e6" y="273.919096"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="486.52526" xlink:href="#C3_0_947d114575" y="273.919096"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="490.099284" xlink:href="#C3_0_8e4775d9e6" y="268.146351"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="481.746383" xlink:href="#C3_0_947d114575" y="268.146351"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="496.27247" xlink:href="#C3_0_8e4775d9e6" y="193.038595"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="489.402808" xlink:href="#C3_0_947d114575" y="193.038595"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="482.787592" xlink:href="#C3_0_8e4775d9e6" y="278.984496"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="495.742904" xlink:href="#C3_0_947d114575" y="278.984496"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="497.515676" xlink:href="#C3_0_8e4775d9e6" y="172.847465"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="495.92895" xlink:href="#C3_0_947d114575" y="172.847465"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="486.253493" xlink:href="#C3_0_8e4775d9e6" y="143.881955"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="494.399192" xlink:href="#C3_0_947d114575" y="143.881955"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="490.992131" xlink:href="#C3_0_8e4775d9e6" y="195.442985"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="489.845999" xlink:href="#C3_0_947d114575" y="195.442985"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="484.906943" xlink:href="#C3_0_8e4775d9e6" y="149.394395"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="485.468309" xlink:href="#C3_0_947d114575" y="149.394395"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="487.853034" xlink:href="#C3_0_8e4775d9e6" y="155.852835"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="481.592737" xlink:href="#C3_0_947d114575" y="155.852835"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="503.389649" xlink:href="#C3_0_8e4775d9e6" y="200.373399"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="485.56097" xlink:href="#C3_0_947d114575" y="200.373399"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="490.272519" xlink:href="#C3_0_8e4775d9e6" y="142.575599"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="493.624116" xlink:href="#C3_0_947d114575" y="142.575599"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="492.579434" xlink:href="#C3_0_8e4775d9e6" y="157.108159"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="491.704913" xlink:href="#C3_0_947d114575" y="157.108159"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="499.437824" xlink:href="#C3_0_8e4775d9e6" y="253.553345"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="497.182347" xlink:href="#C3_0_947d114575" y="253.553345"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="499.669222" xlink:href="#C3_0_8e4775d9e6" y="201.975598"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="482.376175" xlink:href="#C3_0_947d114575" y="201.975598"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="495.295186" xlink:href="#C3_0_8e4775d9e6" y="172.057581"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="483.5617" xlink:href="#C3_0_947d114575" y="172.057581"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="499.19217" xlink:href="#C3_0_8e4775d9e6" y="207.49562"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="498.21515" xlink:href="#C3_0_947d114575" y="207.49562"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="502.676055" xlink:href="#C3_0_8e4775d9e6" y="195.43714"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="481.272222" xlink:href="#C3_0_947d114575" y="195.43714"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="496.997314" xlink:href="#C3_0_8e4775d9e6" y="186.160914"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="500.774194" xlink:href="#C3_0_947d114575" y="186.160914"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="486.169542" xlink:href="#C3_0_8e4775d9e6" y="165.467127"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="499.939626" xlink:href="#C3_0_947d114575" y="165.467127"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="492.143969" xlink:href="#C3_0_8e4775d9e6" y="137.415129"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="487.242776" xlink:href="#C3_0_947d114575" y="137.415129"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="503.078155" xlink:href="#C3_0_8e4775d9e6" y="222.172179"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="499.125875" xlink:href="#C3_0_947d114575" y="222.172179"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="490.374768" xlink:href="#C3_0_8e4775d9e6" y="174.027489"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="483.291365" xlink:href="#C3_0_947d114575" y="174.027489"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="502.677992" xlink:href="#C3_0_8e4775d9e6" y="181.022666"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="487.635458" xlink:href="#C3_0_947d114575" y="181.022666"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="502.918885" xlink:href="#C3_0_8e4775d9e6" y="210.090067"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="503.739559" xlink:href="#C3_0_947d114575" y="210.090067"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="483.206813" xlink:href="#C3_0_8e4775d9e6" y="200.383663"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="500.62283" xlink:href="#C3_0_947d114575" y="200.383663"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="491.788259" xlink:href="#C3_0_8e4775d9e6" y="258.563069"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="498.261694" xlink:href="#C3_0_947d114575" y="258.563069"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="483.996538" xlink:href="#C3_0_8e4775d9e6" y="208.720706"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="484.946676" xlink:href="#C3_0_947d114575" y="208.720706"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="483.991053" xlink:href="#C3_0_8e4775d9e6" y="178.518662"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="495.292717" xlink:href="#C3_0_947d114575" y="178.518662"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="494.442289" xlink:href="#C3_0_8e4775d9e6" y="186.578797"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="495.226955" xlink:href="#C3_0_947d114575" y="186.578797"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="494.090316" xlink:href="#C3_0_8e4775d9e6" y="137.337355"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="489.204008" xlink:href="#C3_0_947d114575" y="137.337355"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="500.209518" xlink:href="#C3_0_8e4775d9e6" y="167.680504"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="491.528342" xlink:href="#C3_0_947d114575" y="167.680504"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="491.545529" xlink:href="#C3_0_8e4775d9e6" y="202.722805"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="495.554183" xlink:href="#C3_0_947d114575" y="202.722805"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="502.646344" xlink:href="#C3_0_8e4775d9e6" y="224.575075"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="493.016723" xlink:href="#C3_0_947d114575" y="224.575075"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="498.392112" xlink:href="#C3_0_8e4775d9e6" y="202.640406"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="491.705303" xlink:href="#C3_0_947d114575" y="202.640406"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="492.124383" xlink:href="#C3_0_8e4775d9e6" y="146.039082"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="490.434564" xlink:href="#C3_0_947d114575" y="146.039082"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="504.285576" xlink:href="#C3_0_8e4775d9e6" y="249.906575"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="488.495378" xlink:href="#C3_0_947d114575" y="249.906575"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="498.112434" xlink:href="#C3_0_8e4775d9e6" y="231.029491"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="499.422787" xlink:href="#C3_0_947d114575" y="231.029491"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="483.606158" xlink:href="#C3_0_8e4775d9e6" y="174.871131"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="492.959149" xlink:href="#C3_0_947d114575" y="174.871131"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="503.110589" xlink:href="#C3_0_8e4775d9e6" y="198.194725"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="481.061928" xlink:href="#C3_0_947d114575" y="198.194725"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="502.134113" xlink:href="#C3_0_8e4775d9e6" y="219.255421"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="483.090312" xlink:href="#C3_0_947d114575" y="219.255421"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="495.491828" xlink:href="#C3_0_8e4775d9e6" y="178.444737"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="481.707434" xlink:href="#C3_0_947d114575" y="178.444737"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="502.256086" xlink:href="#C3_0_8e4775d9e6" y="213.129098"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="501.416026" xlink:href="#C3_0_947d114575" y="213.129098"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="486.322971" xlink:href="#C3_0_8e4775d9e6" y="207.504974"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="502.828976" xlink:href="#C3_0_947d114575" y="207.504974"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="486.730442" xlink:href="#C3_0_8e4775d9e6" y="197.384877"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="495.866076" xlink:href="#C3_0_947d114575" y="197.384877"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="491.023196" xlink:href="#C3_0_8e4775d9e6" y="153.045598"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="480.372062" xlink:href="#C3_0_947d114575" y="153.045598"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="496.067059" xlink:href="#C3_0_8e4775d9e6" y="195.466107"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="479.983803" xlink:href="#C3_0_947d114575" y="195.466107"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="487.096181" xlink:href="#C3_0_8e4775d9e6" y="137.325134"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="493.137643" xlink:href="#C3_0_947d114575" y="137.325134"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="487.523493" xlink:href="#C3_0_8e4775d9e6" y="183.609011"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="496.77052" xlink:href="#C3_0_947d114575" y="183.609011"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="496.161036" xlink:href="#C3_0_8e4775d9e6" y="222.094808"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="497.659624" xlink:href="#C3_0_947d114575" y="222.094808"/> </g> </g> <g id="PathCollection_7"> @@ -2591,103 +2591,103 @@ C -2.236584 -1.29895 -2.5 -0.663008 -2.5 0 C -2.5 0.663008 -2.236584 1.29895 -1.767767 1.767767 C -1.29895 2.236584 -0.663008 2.5 0 2.5 z -" id="C4_0_cbaeea66ea"/> +" id="C4_0_56c013b1a9"/> </defs> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="639.27349" xlink:href="#C4_0_cbaeea66ea" y="217.866084"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="626.312302" xlink:href="#C4_0_56c013b1a9" y="217.866084"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="639.176222" xlink:href="#C4_0_cbaeea66ea" y="196.952689"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="632.392078" xlink:href="#C4_0_56c013b1a9" y="196.952689"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="627.715185" xlink:href="#C4_0_cbaeea66ea" y="234.86163"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="647.480004" xlink:href="#C4_0_56c013b1a9" y="234.86163"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="631.41688" xlink:href="#C4_0_cbaeea66ea" y="225.50833"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="641.107262" xlink:href="#C4_0_56c013b1a9" y="225.50833"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="648.179464" xlink:href="#C4_0_cbaeea66ea" y="334.082122"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="626.198008" xlink:href="#C4_0_56c013b1a9" y="334.082122"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="637.633266" xlink:href="#C4_0_cbaeea66ea" y="184.198957"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="639.538056" xlink:href="#C4_0_56c013b1a9" y="184.198957"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="631.069641" xlink:href="#C4_0_cbaeea66ea" y="272.378164"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="628.672093" xlink:href="#C4_0_56c013b1a9" y="272.378164"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="639.080733" xlink:href="#C4_0_cbaeea66ea" y="260.316603"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="637.068338" xlink:href="#C4_0_56c013b1a9" y="260.316603"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="637.450074" xlink:href="#C4_0_cbaeea66ea" y="241.374028"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="636.764606" xlink:href="#C4_0_56c013b1a9" y="241.374028"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="638.629363" xlink:href="#C4_0_cbaeea66ea" y="246.996271"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="644.076285" xlink:href="#C4_0_56c013b1a9" y="246.996271"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="649.503452" xlink:href="#C4_0_cbaeea66ea" y="190.813563"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="640.23112" xlink:href="#C4_0_56c013b1a9" y="190.813563"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="646.603658" xlink:href="#C4_0_cbaeea66ea" y="195.608936"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="631.816717" xlink:href="#C4_0_56c013b1a9" y="195.608936"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="642.346318" xlink:href="#C4_0_cbaeea66ea" y="307.629253"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="643.711389" xlink:href="#C4_0_56c013b1a9" y="307.629253"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="640.499877" xlink:href="#C4_0_cbaeea66ea" y="208.566013"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="643.644886" xlink:href="#C4_0_56c013b1a9" y="208.566013"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="640.924142" xlink:href="#C4_0_cbaeea66ea" y="236.871217"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="641.476195" xlink:href="#C4_0_56c013b1a9" y="236.871217"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="647.927393" xlink:href="#C4_0_cbaeea66ea" y="204.921565"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="644.665343" xlink:href="#C4_0_56c013b1a9" y="204.921565"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="635.139978" xlink:href="#C4_0_cbaeea66ea" y="222.121266"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="642.248136" xlink:href="#C4_0_56c013b1a9" y="222.121266"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="640.864215" xlink:href="#C4_0_cbaeea66ea" y="185.219633"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="649.311864" xlink:href="#C4_0_56c013b1a9" y="185.219633"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="649.849965" xlink:href="#C4_0_cbaeea66ea" y="150.450621"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="640.638034" xlink:href="#C4_0_56c013b1a9" y="150.450621"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="644.243909" xlink:href="#C4_0_cbaeea66ea" y="278.555335"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="630.300969" xlink:href="#C4_0_56c013b1a9" y="278.555335"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="631.351261" xlink:href="#C4_0_cbaeea66ea" y="202.066996"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="642.517633" xlink:href="#C4_0_56c013b1a9" y="202.066996"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="627.130877" xlink:href="#C4_0_cbaeea66ea" y="210.784076"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="630.742367" xlink:href="#C4_0_56c013b1a9" y="210.784076"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="631.054337" xlink:href="#C4_0_cbaeea66ea" y="278.567443"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="631.175821" xlink:href="#C4_0_56c013b1a9" y="278.567443"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="646.860781" xlink:href="#C4_0_cbaeea66ea" y="246.967553"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="646.416406" xlink:href="#C4_0_56c013b1a9" y="246.967553"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="640.558154" xlink:href="#C4_0_cbaeea66ea" y="293.023169"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="644.189262" xlink:href="#C4_0_56c013b1a9" y="293.023169"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="648.292345" xlink:href="#C4_0_cbaeea66ea" y="188.859446"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="645.952417" xlink:href="#C4_0_56c013b1a9" y="188.859446"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="641.973182" xlink:href="#C4_0_cbaeea66ea" y="196.281563"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="627.07897" xlink:href="#C4_0_56c013b1a9" y="196.281563"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="643.628242" xlink:href="#C4_0_cbaeea66ea" y="263.9663"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="628.758756" xlink:href="#C4_0_56c013b1a9" y="263.9663"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="640.521946" xlink:href="#C4_0_cbaeea66ea" y="234.912236"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="642.339311" xlink:href="#C4_0_56c013b1a9" y="234.912236"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="640.988051" xlink:href="#C4_0_cbaeea66ea" y="340.218479"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="637.099455" xlink:href="#C4_0_56c013b1a9" y="340.218479"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="635.31085" xlink:href="#C4_0_cbaeea66ea" y="246.953886"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="630.691432" xlink:href="#C4_0_56c013b1a9" y="246.953886"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="627.729206" xlink:href="#C4_0_cbaeea66ea" y="278.549155"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="637.948529" xlink:href="#C4_0_56c013b1a9" y="278.549155"/> </g> </g> <g id="PathCollection_8"> @@ -2702,85 +2702,85 @@ C -2.236584 -1.29895 -2.5 -0.663008 -2.5 0 C -2.5 0.663008 -2.236584 1.29895 -1.767767 1.767767 C -1.29895 2.236584 -0.663008 2.5 0 2.5 z -" id="C5_0_25efba35e2"/> +" id="C5_0_210b45a4e1"/> </defs> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="733.84608" xlink:href="#C5_0_25efba35e2" y="205.123489"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="740.338326" xlink:href="#C5_0_210b45a4e1" y="205.123489"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="741.282928" xlink:href="#C5_0_25efba35e2" y="269.413549"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="733.60375" xlink:href="#C5_0_210b45a4e1" y="269.413549"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="741.367481" xlink:href="#C5_0_25efba35e2" y="187.358818"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="737.572804" xlink:href="#C5_0_210b45a4e1" y="187.358818"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="729.180363" xlink:href="#C5_0_25efba35e2" y="187.84272"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="724.324087" xlink:href="#C5_0_210b45a4e1" y="187.84272"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="725.711059" xlink:href="#C5_0_25efba35e2" y="234.160756"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="725.322845" xlink:href="#C5_0_210b45a4e1" y="234.160756"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="729.710243" xlink:href="#C5_0_25efba35e2" y="251.163838"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="747.163636" xlink:href="#C5_0_210b45a4e1" y="251.163838"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="747.163636" xlink:href="#C5_0_25efba35e2" y="164.749615"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="732.435562" xlink:href="#C5_0_210b45a4e1" y="164.749615"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="745.76086" xlink:href="#C5_0_25efba35e2" y="254.741197"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="742.964507" xlink:href="#C5_0_210b45a4e1" y="254.741197"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="730.516978" xlink:href="#C5_0_25efba35e2" y="166.697991"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="738.025184" xlink:href="#C5_0_210b45a4e1" y="166.697991"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="741.103652" xlink:href="#C5_0_25efba35e2" y="175.221295"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="731.217025" xlink:href="#C5_0_210b45a4e1" y="175.221295"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="731.263442" xlink:href="#C5_0_25efba35e2" y="240.360593"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="729.212931" xlink:href="#C5_0_210b45a4e1" y="240.360593"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="729.14624" xlink:href="#C5_0_25efba35e2" y="208.74744"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="727.449826" xlink:href="#C5_0_210b45a4e1" y="208.74744"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="743.312936" xlink:href="#C5_0_25efba35e2" y="234.172585"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="724.192658" xlink:href="#C5_0_210b45a4e1" y="234.172585"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="741.709241" xlink:href="#C5_0_25efba35e2" y="166.719963"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="727.035455" xlink:href="#C5_0_210b45a4e1" y="166.719963"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="729.359299" xlink:href="#C5_0_25efba35e2" y="206.813301"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="737.130465" xlink:href="#C5_0_210b45a4e1" y="206.813301"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="732.050251" xlink:href="#C5_0_25efba35e2" y="212.772616"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="737.721393" xlink:href="#C5_0_210b45a4e1" y="212.772616"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="724.854083" xlink:href="#C5_0_25efba35e2" y="158.300318"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="744.54167" xlink:href="#C5_0_210b45a4e1" y="158.300318"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="744.956943" xlink:href="#C5_0_25efba35e2" y="223.381319"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="730.20411" xlink:href="#C5_0_210b45a4e1" y="223.381319"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="739.807474" xlink:href="#C5_0_25efba35e2" y="220.444186"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="724.902241" xlink:href="#C5_0_210b45a4e1" y="220.444186"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="728.761014" xlink:href="#C5_0_25efba35e2" y="288.823302"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="728.259834" xlink:href="#C5_0_210b45a4e1" y="288.823302"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="743.595589" xlink:href="#C5_0_25efba35e2" y="269.400828"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="732.049746" xlink:href="#C5_0_210b45a4e1" y="269.400828"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="727.620926" xlink:href="#C5_0_25efba35e2" y="254.797829"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="724.479442" xlink:href="#C5_0_210b45a4e1" y="254.797829"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="740.025616" xlink:href="#C5_0_25efba35e2" y="211.304124"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="729.409629" xlink:href="#C5_0_210b45a4e1" y="211.304124"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="732.648138" xlink:href="#C5_0_25efba35e2" y="194.48263"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="735.811378" xlink:href="#C5_0_210b45a4e1" y="194.48263"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="731.495627" xlink:href="#C5_0_25efba35e2" y="254.846716"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="737.565829" xlink:href="#C5_0_210b45a4e1" y="254.846716"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="725.983054" xlink:href="#C5_0_25efba35e2" y="269.446309"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="742.497576" xlink:href="#C5_0_210b45a4e1" y="269.446309"/> </g> </g> <g id="text_10"> @@ -2805,7 +2805,7 @@ Q 8.796875 56.9375 3.125 58.453125 z " id="ArialMT-42"/> </defs> - <g style="fill:#262626;" transform="translate(436.527535 109.709863)scale(0.12 -0.12)"> + <g style="fill:#262626;" transform="translate(436.16219 109.709863)scale(0.12 -0.12)"> <use xlink:href="#ArialMT-42"/> <use x="38.916016" xlink:href="#ArialMT-42"/> <use x="77.832031" xlink:href="#ArialMT-42"/> @@ -2813,14 +2813,14 @@ z </g> <g id="text_11"> <!-- ** --> - <g style="fill:#262626;" transform="translate(194.595105 70.900587)scale(0.12 -0.12)"> + <g style="fill:#262626;" transform="translate(194.548862 70.900587)scale(0.12 -0.12)"> <use xlink:href="#ArialMT-42"/> <use x="38.916016" xlink:href="#ArialMT-42"/> </g> </g> <g id="text_12"> <!-- ns --> - <g style="fill:#262626;" transform="translate(681.464653 123.579797)scale(0.12 -0.12)"> + <g style="fill:#262626;" transform="translate(680.780205 123.579797)scale(0.12 -0.12)"> <use xlink:href="#ArialMT-110"/> <use x="55.615234" xlink:href="#ArialMT-115"/> </g> @@ -2910,10 +2910,10 @@ C -3.464901 -2.012324 -3.872983 -1.027127 -3.872983 0 C -3.872983 1.027127 -3.464901 2.012324 -2.738613 2.738613 C -2.012324 3.464901 -1.027127 3.872983 0 3.872983 z -" id="m332170e5fd" style="stroke:#4c72b0;"/> +" id="md5fa658d96" style="stroke:#4c72b0;"/> </defs> <g> - <use style="fill:#4c72b0;stroke:#4c72b0;" x="665.974063" xlink:href="#m332170e5fd" y="83.487344"/> + <use style="fill:#4c72b0;stroke:#4c72b0;" x="665.974063" xlink:href="#md5fa658d96" y="83.487344"/> </g> </g> <g id="text_14"> @@ -2946,10 +2946,10 @@ C -3.464901 -2.012324 -3.872983 -1.027127 -3.872983 0 C -3.872983 1.027127 -3.464901 2.012324 -2.738613 2.738613 C -2.012324 3.464901 -1.027127 3.872983 0 3.872983 z -" id="m8557ba0fcf" style="stroke:#dd8452;"/> +" id="mcda2f6a9ba" style="stroke:#dd8452;"/> </defs> <g> - <use style="fill:#dd8452;stroke:#dd8452;" x="665.974063" xlink:href="#m8557ba0fcf" y="99.310156"/> + <use style="fill:#dd8452;stroke:#dd8452;" x="665.974063" xlink:href="#mcda2f6a9ba" y="99.310156"/> </g> </g> <g id="text_15"> @@ -2971,7 +2971,7 @@ z </g> </g> <defs> - <clipPath id="p24d2ef447f"> + <clipPath id="p078fcafce0"> <rect height="326.16" width="669.6" x="108" y="51.84"/> </clipPath> </defs>	Create Annotation, rename box_pairs and box_ to group In this PR, more responsibility is extracted from `Annotator` towards an `Annotation` class. `Annotation` objects now link the plot structures (boxes for boxplots) to the statistical result or custom annotation. Also, it was time to rename the variables from `box_` to `group_` as the package is definitely not only about boxplots.		2021-07-21T18:29:52	[]	[]
trevismd/statannotations	trevismd__statannotations-13	0005f989961a57d712ccecb671499d5ab4068834	diff --git a/CHANGELOG.md b/CHANGELOG.md index b3c5b18..dbf1858 100644 --- a/CHANGELOG.md +++ b/CHANGELOG.md @@ -1,5 +1,14 @@ -## v3 -### v3.0 +## v0.3 +### v0.3.2 + - Fix `simple` format outputs + - Fix `ImportError` when applying a multiple comparison correction without statsmodels. + - Multiple comparison correction is `None` by default, as `statsmodels` is an additional dependency. + +### v0.3.1 + - Added support of functions returning more than the two expected values when used in `StatTest` + - Fix version numbers in CHANGELOG + +### v0.3.0 (Explicitly Requires Python 3.6) - Refactoring with implementation of `StatTest` @@ -9,6 +18,6 @@ ([#5](https://github.com/trevismd/statannotations/pull/5) by [JosephLalli](https://github.com/JosephLalli), fixes https://github.com/webermarcolivier/statannot/issues/21). - Add this CHANGELOG -## v2 -### v2.8 - - Fix bug on group/box named 0, fixes https://github.com/trevismd/statannotations/issues/10, originally in https://github.com/webermarcolivier/statannot/issues/78. Independently fixed in https://github.com/webermarcolivier/statannot/pull/73 before this issue. \ No newline at end of file +## v0.2 +### v0.2.8 + - Fix bug on group/box named 0, fixes https://github.com/trevismd/statannotations/issues/10, originally in https://github.com/webermarcolivier/statannot/issues/78. Independently fixed in https://github.com/webermarcolivier/statannot/pull/73 before this issue. diff --git a/coverage.svg b/coverage.svg index eb83f3f..dd6df12 100644 --- a/coverage.svg +++ b/coverage.svg @@ -9,13 +9,13 @@ </mask> <g mask="url(#a)"> <path fill="#555" d="M0 0h63v20H0z"/> - <path fill="#fe7d37" d="M63 0h36v20H63z"/> + <path fill="#dfb317" d="M63 0h36v20H63z"/> <path fill="url(#b)" d="M0 0h99v20H0z"/> </g> <g fill="#fff" text-anchor="middle" font-family="DejaVu Sans,Verdana,Geneva,sans-serif" font-size="11"> <text x="31.5" y="15" fill="#010101" fill-opacity=".3">coverage</text> <text x="31.5" y="14">coverage</text> - <text x="80" y="15" fill="#010101" fill-opacity=".3">59%</text> - <text x="80" y="14">59%</text> + <text x="80" y="15" fill="#010101" fill-opacity=".3">61%</text> + <text x="80" y="14">61%</text> </g> </svg> diff --git a/statannotations/__init__.py b/statannotations/__init__.py index 924768d..e099ae2 100644 --- a/statannotations/__init__.py +++ b/statannotations/__init__.py @@ -1,4 +1,4 @@ -__all__ = ["__version__", "stats", "statannotations"] +__all__ = ["__version__", "stats", "statannotations", "add_stat_annotation"] from .statannotations import add_stat_annotation from ._version import __version__ diff --git a/statannotations/_version.py b/statannotations/_version.py index 493f741..f9aa3e1 100644 --- a/statannotations/_version.py +++ b/statannotations/_version.py @@ -1,1 +1,1 @@ -__version__ = "0.3.0" +__version__ = "0.3.2" diff --git a/statannotations/statannotations.py b/statannotations/statannotations.py index b46745d..34f64f0 100644 --- a/statannotations/statannotations.py +++ b/statannotations/statannotations.py @@ -29,7 +29,7 @@ def add_stat_annotation(ax, plot='boxplot', data=None, x=None, y=None, text_annot_custom=None, loc='inside', show_test_name=True, pvalue_thresholds=DEFAULT, stats_params: dict = None, - comparisons_correction='bonferroni', + comparisons_correction=None, num_comparisons='auto', use_fixed_offset=False, line_offset_to_box=None, line_offset=None, line_height=0.02, text_offset=1, color='0.2', @@ -52,7 +52,7 @@ def add_stat_annotation(ax, plot='boxplot', data=None, x=None, y=None, :param data: seaborn plot's data :param x: seaborn plot's x :param y: seaborn plot's y - :param test: StatResult or name from list of scipy functions supported + :param test: `StatTest` instance or name from list of scipy functions supported :param hue: seaborn plot's hue :param order: seaborn plot's order :param hue_order: seaborn plot's hue_order @@ -65,7 +65,7 @@ def add_stat_annotation(ax, plot='boxplot', data=None, x=None, y=None, :param test_short_name: How the test name should show on the plot, if show_test_name is True (default). Default is the full test name - :param pvalue_format_string: defaults to `"{.3e}"` + :param pvalue_format_string: defaults to `"{:.3e}"` or `"{:.2f}"` for `"simple"` format :param pvalue_thresholds: list of lists, or tuples. Default is: For "star" text_format: `[[1e-4, "**"], [1e-3, ""], [1e-2, ""], [0.05, ""], [1, "ns"]]`. @@ -186,7 +186,7 @@ def get_box_data(b_plotter, box_name): else: pvalue_thresholds = [[1e-5, "1e-5"], [1e-4, "1e-4"], [1e-3, "0.001"], [1e-2, "0.01"], - [5e-2, "0.05"], [1, "ns"]] + [5e-2, "0.05"]] if stats_params is None: stats_params = dict() @@ -233,7 +233,12 @@ def get_box_data(b_plotter, box_name): pass elif isinstance(comparisons_correction, str): check_valid_correction_name(comparisons_correction) - comparisons_correction = ComparisonsCorrection(comparisons_correction) + try: + comparisons_correction = ComparisonsCorrection(comparisons_correction) + except ImportError as e: + raise ImportError(f"{e} Please install statsmodels or pass " + f"`comparisons_correction=None`.") + elif not(isinstance(comparisons_correction, ComparisonsCorrection)): raise ValueError("comparisons_correction must be a statmodels " "method name or a ComparisonCorrection instance") @@ -465,7 +470,10 @@ def get_transform_func(ax, kind): text = pval_annotation_text(result, pvalue_thresholds) elif text_format == 'simple': if show_test_name: - test_short_name = show_test_name and test_short_name or test + if test_short_name is None: + test_short_name = (test + if isinstance(test, str) + else test.short_name) else: test_short_name = "" text = simple_text(result, simple_format_string, diff --git a/statannotations/stats/ComparisonsCorrection.py b/statannotations/stats/ComparisonsCorrection.py index 3664649..c23a62c 100644 --- a/statannotations/stats/ComparisonsCorrection.py +++ b/statannotations/stats/ComparisonsCorrection.py @@ -6,12 +6,9 @@ try: from statsmodels.stats.multitest import multipletests -except ImportError as statsmodel_import_error: +except ImportError: multipletests = None -else: - statsmodel_import_error = None - # For user convenience methods_names = {'bonferroni': 'Bonferroni', 'bonf': 'Bonferroni', @@ -84,7 +81,9 @@ def __init__(self, method: Union[str, callable], alpha: float = 0.05, if isinstance(method, str): if multipletests is None: - raise ImportError(statsmodel_import_error) + raise ImportError("The statsmodels package is required to use " + "one of the multiple comparisons correction " + "methods proposed in statannotations.") self.name = name if name is not None else method self.method, self.type = get_correction_parameters(method) diff --git a/statannotations/stats/StatTest.py b/statannotations/stats/StatTest.py index 2898680..ef97e12 100644 --- a/statannotations/stats/StatTest.py +++ b/statannotations/stats/StatTest.py @@ -83,7 +83,9 @@ def __init__(self, func: Callable, test_long_name: str, Wrapper for any statistical function to use with statannotations. The function must be callable with two 1D collections of data points as first and second arguments, can have any number of additional arguments - , and must return a) a statistical value, and b) a pvalue. + , and must at least return, in order, + a) a statistical value, and + b) a pvalue. :param func: Function to be called @@ -117,7 +119,11 @@ def __call__(self, box_data1, box_data2, alpha=0.05, *stat_params): stat, pval = self._func(box_data1, box_data2, self.args, - {self.kwargs, stat_params}) + {self.kwargs, stat_params})[:2] return StatResult(self._test_long_name, self._test_short_name, self._stat_name, stat, pval, alpha=alpha) + + @property + def short_name(self): + return self._test_short_name	Support functions with more returned values in `StatTest` This fixes the bug in the situation where the function passed to StatTest would return more than two_values. And includes minor changes from master branch		2021-06-17T07:15:42	[]	[]
RalphTro/epcis-event-hash-generator	RalphTro__epcis-event-hash-generator-61	e38f16a390d5caa71bdfe34addabdc2f92831896	diff --git a/epcis_event_hash_generator/hash_generator.py b/epcis_event_hash_generator/hash_generator.py index 050de39..14ac917 100644 --- a/epcis_event_hash_generator/hash_generator.py +++ b/epcis_event_hash_generator/hash_generator.py @@ -58,23 +58,25 @@ def _fix_time_stamp_format(timestamp): def _child_to_pre_hash_string(child, sub_child_order): + logging.debug("Processing '%s'", child) text = "" grand_child_text = "" if sub_child_order: grand_child_text = _recurse_through_children_in_order(child[2], sub_child_order) if child[1]: text = child[1].strip() - if child[0].lower().find("time") > 0 and child[0].lower().find("offset") < 0: + if child[0].lower().find("time") >= 0 and child[0].lower().find("offset") < 0: text = _fix_time_stamp_format(text) else: text = _canonize_value(text) if text: text = "=" + text - logging.debug("Adding text '%s'", text) if text or grand_child_text: - return child[0] + text + grand_child_text + re = child[0] + text + grand_child_text + logging.debug("pre hash string element: '%s'", text) + return re return "" @@ -129,6 +131,7 @@ def _canonize_value(text): if converted: logging.debug("Converted %s to %s", text, converted) return converted + logging.debug("No canonical form for '%s'", text) return text	Date format with pre-hash string Hi, In rule 8 of the algorithm, you say: ``` All timestamps SHALL be expressed in UTC; the zero UTC offset SHALL be expressed with the capital letter 'Z'. ``` which means that timestamps can have a time offset, right? This is indeed the case of the `time` field of a `sensorMetadata`. If you look at the example `epcisDocWithSensorDataObjectEvent.jsonld`, the pre-hash string is: ``` sensorMetadatatime=2019-04-02T14:05:00.000+01:00 ``` (It keeps the offset so everything is working as expected) However, when you look at the pre-hash string of the last event in `SensorDataExamples.xml`, we have this field: ``` <declarationTime>2020-09-29T15:00:00.000+02:00</declarationTime> ``` But in the pre-hash string, it converts the date like this: ``` declarationTime=2020-09-29T13:00:00.000 ``` Is it the expected output? If it is the case, do I miss something in the algorithm? Thanks a lot.	Dear @clementh59 , Good catch! You are correct. All timestamps should be converted to zero UTC. I hope you do not need a fix urgently? ;-) Kind regards and thanks again; Ralph Dear @RalphTro , I see, thanks! No, that's ok, I just want to be sure I understand the algorithm correctly :) Regards, Clément	2021-10-18T21:18:38	[]	[]
RalphTro/epcis-event-hash-generator	RalphTro__epcis-event-hash-generator-45	ce6ed045f5ec5147bcb39a9a120002996b06a8f2	diff --git a/.flake8 b/.flake8 index fe83fbf..8318a8f 100644 --- a/.flake8 +++ b/.flake8 @@ -5,3 +5,6 @@ exclude = context.py max-complexity = 10 max-line-length = 120 +per-file-ignores = + tests/dl_normaliser_test.py:E501 + epcis_event_hash_generator/dl_normaliser.py:E501,C901 \ No newline at end of file diff --git a/epcis_event_hash_generator/dl_normaliser.py b/epcis_event_hash_generator/dl_normaliser.py new file mode 100644 index 0000000..75efdcd --- /dev/null +++ b/epcis_event_hash_generator/dl_normaliser.py @@ -0,0 +1,445 @@ +""" +GS1 Digital Link Normaliser. + +This script accepts any valid URI scheme accommodating a GS1 ID, +i.e. EPC URIs, EPC Class URIs, EPC ID Pattern URIs, or GS1 Digital Link URIs. +It converts all of them into one normalised form, meaning that it +(a) converts it into a canonical GS1 DL URI, +(b) ensures that it only contains the most fine-granular ID level, +(c) strips off any further attributes. + +.. module:: dl_normaliser + :synopsis: Normalises the gs1 id formats to digital link for https://github.com/RalphTro/epcis-event-hash-generator/ + +.. moduleauthor:: Ralph Troeger <[email protected]> + +Copyright 2019-2021 Ralph Troeger + +This program is free software: you can redistribute it and/or modify +it under the terms given in the LICENSE file. + +This program is distributed in the hope that it will be useful, but +WITHOUT ANY WARRANTY; without even the implied warranty of +MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the LICENSE +file for details. + +""" + +from re import match +from gtin import GTIN +import logging + + +def __web_uri_percent_encoder(input): + """ + Function percent-encodes URL-unsafe characters in GS1 Digital Link URIs. + + Table 7-1 in the GS1 Digital Link Standard requires + the following symbols to be percent-encoded: + '!', space, '#', '%', '&', '(', ')', '', '+', ',', '/', ':' + EPC URIs already prohibit some of them (e.g. '#') + Function 'webURIPercentEncoder' is called to ensure that + data elements accommodating these symbols are percent-encoded. + + Parameters + ---------- + input : str + Character requiring percent-encoding. + + Returns + ------- + str + Percent-encoded equivalent of character. + """ + return ( + input.replace('!', '%21') + .replace('(', '%28') + .replace(')', '%29') + .replace('', '%2A') + .replace('+', '%2B') + .replace(',', '%2C') + .replace(':', '%3A') + ) + + +def normaliser(uri): + """ + Function converts any standard URI conveying a GS1 Key in Canonical GS1 DL URI. + + Function 'normaliser' expects any URI to be used in EPCIS events + that convey a GS1 key, i.e. EPC URIs, EPC Class URIs, + EPC ID Pattern URIs, or GS1 Digital Link URIs. + It returns a corresponding, constrained version of a + canonical GS1 Digital Link URI, i.e. with + the lowest level of identification and without CPV/query string. + + Parameters + ---------- + uri : str + Valid EPC URI, EPC Pattern URI, EPC Class URI, GS1 Digital Link URI. + + Returns + ------- + str + Constrained, canonicalised GS1 Digital Link URI equivalent. + None + """ + if not isinstance(uri, str): + logging.warning("dl normaliser called with non-string argument") + return None + + try: + partition = uri.index('.') + except ValueError: + logging.debug("No '.' in %s. Not a normalisable uri.", uri) + return None + + # EPC URIs + if match( + r'^urn:epc:id:sgtin:((\d{6}\.\d{7})\|(\d{7}\.\d{6})\|(\d{8}\.\d{5})\|(\d{9}\.\d{4})\|(\d{10}\.\d{3})\|(\d{11}\.\d{2})\|(\d{12}\.\d{1}))\.(\%2[125-9A-Fa-f]\|\%3[0-9A-Fa-f]\|\%4[1-9A-Fa-f]\|\%5[0-9AaFf]\|\%6[1-9A-Fa-f]\|\%7[0-9Aa]\|[!\')(+,.0-9:;=A-Za-z_-]){1,20}$', + uri) is not None: + gs1companyprefix = uri[17:partition] + itemref = uri[(partition + 1):(partition + 1 + (13 - len(gs1companyprefix)))] + rawGTIN = itemref[0:1] + gs1companyprefix + itemref[1:] + serial = uri[32:] + return ('https://id.gs1.org/01/' + str(GTIN(raw=rawGTIN)) + '/21/' + __web_uri_percent_encoder(serial)) + + if match( + r'^urn:epc:id:sscc:((\d{6}\.\d{11}$)\|(\d{7}\.\d{10}$)\|(\d{8}\.\d{9}$)\|(\d{9}\.\d{8}$)\|(\d{10}\.\d{7}$)\|(\d{11}\.\d{6}$)\|(\d{12}\.\d{5}$))', + uri) is not None: + gs1companyprefix = uri[16:partition] + serialref = uri[(partition + 1):] + rawSSCC = uri[(partition + 1):(partition + 2)] + gs1companyprefix + serialref + return ('https://id.gs1.org/00/' + str(GTIN(raw=rawSSCC))) + + if match( + r'^urn:epc:id:sgln:((\d{6}\.\d{6})\|(\d{7}\.\d{5})\|(\d{8}\.\d{4})\|(\d{9}\.\d{3})\|(\d{10}\.\d{2})\|(\d{11}\.\d{1})\|(\d{12}\.))\.(\%2[125-9A-Fa-f]\|\%3[0-9A-Fa-f]\|\%4[1-9A-Fa-f]\|\%5[0-9AaFf]\|\%6[1-9A-Fa-f]\|\%7[0-9Aa]\|[!\')(+,.0-9:;=A-Za-z_-]){1,20}$', + uri) is not None: + gs1companyprefix = uri[16:partition] + locationref = uri[(partition + 1):(partition + 1 + + (12 - len(gs1companyprefix)))] + rawGLN = gs1companyprefix + locationref + extension = uri[30:] + if extension == '0': + return ('https://id.gs1.org/414/' + str(GTIN(raw=rawGLN))) + else: + return ('https://id.gs1.org/414/' + str(GTIN(raw=rawGLN)) + '/254/' + __web_uri_percent_encoder(extension)) + + if match( + r'^urn:epc:id:grai:(([\d]{6}\.[\d]{6})\|([\d]{7}\.[\d]{5})\|([\d]{8}\.[\d]{4})\|([\d]{9}\.[\d]{3})\|([\d]{10}\.[\d]{2})\|([\d]{11}\.[\d]{1})\|([\d]{12}\.\.))\.(\%2[125-9A-Fa-f]\|\%3[0-9A-Fa-f]\|\%4[1-9A-Fa-f]\|\%5[0-9AaFf]\|\%6[1-9A-Fa-f]\|\%7[0-9Aa]\|[!\')(+,.0-9:;=A-Za-z_-]){1,16}$', + uri) is not None: + gs1companyprefix = uri[16:partition] + assetref = uri[(partition + 1):(partition + 1 + (12 - len(gs1companyprefix)))] + rawGRAI = '0' + gs1companyprefix + assetref + serial = uri[30:] + return ('https://id.gs1.org/8003/' + str(GTIN(raw=rawGRAI)) + __web_uri_percent_encoder(serial)) + + if match( + r'^urn:epc:id:giai:(([\d]{6}\.(\%2[125-9A-Fa-f]\|\%3[0-9A-Fa-f]\|\%4[1-9A-Fa-f]\|\%5[0-9AaFf]\|\%6[1-9A-Fa-f]\|\%7[0-9Aa]\|[!\')(+,.0-9:;=A-Za-z_-]){1,24})\|([\d]{7}\.(\%2[125-9A-Fa-f]\|\%3[0-9A-Fa-f]\|\%4[1-9A-Fa-f]\|\%5[0-9AaFf]\|\%6[1-9A-Fa-f]\|\%7[0-9Aa]\|[!\')(+,.0-9:;=A-Za-z_-]){1,23})\|([\d]{8}\.(\%2[125-9A-Fa-f]\|\%3[0-9A-Fa-f]\|\%4[1-9A-Fa-f]\|\%5[0-9AaFf]\|\%6[1-9A-Fa-f]\|\%7[0-9Aa]\|[!\')(+,.0-9:;=A-Za-z_-]){1,22})\|([\d]{9}\.(\%2[125-9A-Fa-f]\|\%3[0-9A-Fa-f]\|\%4[1-9A-Fa-f]\|\%5[0-9AaFf]\|\%6[1-9A-Fa-f]\|\%7[0-9Aa]\|[!\')(+,.0-9:;=A-Za-z_-]){1,21})\|([\d]{10}\.(\%2[125-9A-Fa-f]\|\%3[0-9A-Fa-f]\|\%4[1-9A-Fa-f]\|\%5[0-9AaFf]\|\%6[1-9A-Fa-f]\|\%7[0-9Aa]\|[!\')(+,.0-9:;=A-Za-z_-]){1,20})\|([\d]{11}\.(\%2[125-9A-Fa-f]\|\%3[0-9A-Fa-f]\|\%4[1-9A-Fa-f]\|\%5[0-9AaFf]\|\%6[1-9A-Fa-f]\|\%7[0-9Aa]\|[!\')(+,.0-9:;=A-Za-z_-]){1,19})\|([\d]{12}\.(\%2[125-9A-Fa-f]\|\%3[0-9A-Fa-f]\|\%4[1-9A-Fa-f]\|\%5[0-9AaFf]\|\%6[1-9A-Fa-f]\|\%7[0-9Aa]\|[!\')(+,.0-9:;=A-Za-z_-]){1,18}))$', + uri) is not None: + gs1companyprefix = uri[16:partition] + assetref = uri[(partition + 1):] + return ('https://id.gs1.org/8004/' + gs1companyprefix + __web_uri_percent_encoder(assetref)) + + if match( + r'^urn:epc:id:gsrn:(([\d]{6}\.[\d]{11}$)\|([\d]{7}\.[\d]{10}$)\|([\d]{8}\.[\d]{9}$)\|([\d]{9}\.[\d]{8}$)\|([\d]{10}\.[\d]{7}$)\|([\d]{11}\.[\d]{6}$)\|([\d]{12}\.[\d]{5}$))', + uri) is not None: + gs1companyprefix = uri[16:partition] + serviceref = uri[(partition + 1):] + rawGSRN = gs1companyprefix + serviceref + return ('https://id.gs1.org/8018/' + str(GTIN(raw=rawGSRN))) + + if match( + r'^urn:epc:id:gsrnp:(([\d]{6}\.[\d]{11}$)\|([\d]{7}\.[\d]{10}$)\|([\d]{8}\.[\d]{9}$)\|([\d]{9}\.[\d]{8}$)\|([\d]{10}\.[\d]{7}$)\|([\d]{11}\.[\d]{6}$)\|([\d]{12}\.[\d]{5}$))', + uri) is not None: + gs1companyprefix = uri[17:partition] + serviceref = uri[(partition + 1):] + rawGSRNP = gs1companyprefix + serviceref + return ('https://id.gs1.org/8017/' + str(GTIN(raw=rawGSRNP))) + + if match( + r'^urn:epc:id:gdti:(([\d]{6}\.[\d]{6})\|([\d]{7}\.[\d]{5})\|([\d]{8}\.[\d]{4})\|([\d]{9}\.[\d]{3})\|([\d]{10}\.[\d]{2})\|([\d]{11}\.[\d]{1})\|([\d]{12}\.\.))(\%2[125-9A-Fa-f]\|\%3[0-9A-Fa-f]\|\%4[1-9A-Fa-f]\|\%5[0-9AaFf]\|\%6[1-9A-Fa-f]\|\%7[0-9Aa]\|[!\')(+,.0-9:;=A-Za-z_-]){1,20}$', + uri) is not None: + gs1companyprefix = uri[16:partition] + documenttype = uri[(partition + 1):(partition + 1 + + (12 - len(gs1companyprefix)))] + rawGDTI = gs1companyprefix + documenttype + serial = uri[30:] + return ('https://id.gs1.org/253/' + str(GTIN(raw=rawGDTI)) + __web_uri_percent_encoder(serial)) + + if match( + r'^urn:epc:id:cpi:((\d{6}\.(\%2[3dfDF]\|\%3[0-9]\|\%4[1-9A-Fa-f]\|\%5[0-9Aa]\|[0-9A-Z-]){1,24})\|(\d{7}\.(\%2[3dfDF]\|\%3[0-9]\|\%4[1-9A-Fa-f]\|\%5[0-9Aa]\|[0-9A-Z-]){1,23})\|(\d{8}\.(\%2[3dfDF]\|\%3[0-9]\|\%4[1-9A-Fa-f]\|\%5[0-9Aa]\|[0-9A-Z-]){1,22})\|(\d{9}\.(\%2[3dfDF]\|\%3[0-9]\|\%4[1-9A-Fa-f]\|\%5[0-9Aa]\|[0-9A-Z-]){1,21})\|(\d{10}\.(\%2[3dfDF]\|\%3[0-9]\|\%4[1-9A-Fa-f]\|\%5[0-9Aa]\|[0-9A-Z-]){1,20})\|(\d{11}\.(\%2[3dfDF]\|\%3[0-9]\|\%4[1-9A-Fa-f]\|\%5[0-9Aa]\|[0-9A-Z-]){1,19})\|(\d{12}\.(\%2[3dfDF]\|\%3[0-9]\|\%4[1-9A-Fa-f]\|\%5[0-9Aa]\|[0-9A-Z-]){1,18}))\.[\d]{1,12}$', + uri) is not None: + gs1companyprefix = uri[15:partition] + separator = uri.rfind('.') + cpref = uri[(partition + 1):(separator)] + rawCPI = gs1companyprefix + cpref + serial = uri[(separator + 1):] + return ('https://id.gs1.org/8010/' + __web_uri_percent_encoder(rawCPI) + '/8011/' + serial) + + if match( + r'^urn:epc:id:sgcn:(([\d]{6}\.[\d]{6})\|([\d]{7}\.[\d]{5})\|([\d]{8}\.[\d]{4})\|([\d]{9}\.[\d]{3})\|([\d]{10}\.[\d]{2})\|([\d]{11}\.[\d]{1})\|([\d]{12}\.))\.[\d]{1,12}$', + uri) is not None: + gs1companyprefix = uri[16:partition] + couponref = uri[(partition + 1):(partition + 1 + (12 - len(gs1companyprefix)))] + rawSGCN = gs1companyprefix + couponref + serial = uri[30:] + return ('https://id.gs1.org/255/' + str(GTIN(raw=rawSGCN)) + serial) + + if match( + r'^urn:epc:id:ginc:([\d]{6}\.(\%2[125-9A-Fa-f]\|\%3[0-9A-Fa-f]\|\%4[1-9A-Fa-f]\|\%5[0-9AaFf]\|\%6[1-9A-Fa-f]\|\%7[0-9Aa]\|[!\')(+,.0-9:;=A-Za-z_-]){1,24}\|[\d]{7}\.(\%2[125-9A-Fa-f]\|\%3[0-9A-Fa-f]\|\%4[1-9A-Fa-f]\|\%5[0-9AaFf]\|\%6[1-9A-Fa-f]\|\%7[0-9Aa]\|[!\')(+,.0-9:;=A-Za-z_-]){1,23}\|[\d]{8}\.(\%2[125-9A-Fa-f]\|\%3[0-9A-Fa-f]\|\%4[1-9A-Fa-f]\|\%5[0-9AaFf]\|\%6[1-9A-Fa-f]\|\%7[0-9Aa]\|[!\')(+,.0-9:;=A-Za-z_-]){1,22}\|[\d]{9}\.(\%2[125-9A-Fa-f]\|\%3[0-9A-Fa-f]\|\%4[1-9A-Fa-f]\|\%5[0-9AaFf]\|\%6[1-9A-Fa-f]\|\%7[0-9Aa]\|[!\')(+,.0-9:;=A-Za-z_-]){1,21}\|[\d]{10}\.(\%2[125-9A-Fa-f]\|\%3[0-9A-Fa-f]\|\%4[1-9A-Fa-f]\|\%5[0-9AaFf]\|\%6[1-9A-Fa-f]\|\%7[0-9Aa]\|[!\')(+,.0-9:;=A-Za-z_-]){1,20}\|[\d]{11}\.(\%2[125-9A-Fa-f]\|\%3[0-9A-Fa-f]\|\%4[1-9A-Fa-f]\|\%5[0-9AaFf]\|\%6[1-9A-Fa-f]\|\%7[0-9Aa]\|[!\')(+,.0-9:;=A-Za-z_-]){1,19}\|[\d]{12}\.(\%2[125-9A-Fa-f]\|\%3[0-9A-Fa-f]\|\%4[1-9A-Fa-f]\|\%5[0-9AaFf]\|\%6[1-9A-Fa-f]\|\%7[0-9Aa]\|[!\')(+,.0-9:;=A-Za-z_-]){1,18})$', + uri) is not None: + gs1companyprefix = uri[16:partition] + consignmentref = uri[(partition + 1):] + return ('https://id.gs1.org/401/' + gs1companyprefix + __web_uri_percent_encoder(consignmentref)) + + if match( + r'^urn:epc:id:gsin:(([\d]{6}\.[\d]{10}$)\|([\d]{7}\.[\d]{9}$)\|([\d]{8}\.[\d]{8}$)\|([\d]{9}\.[\d]{7}$)\|([\d]{10}\.[\d]{6}$)\|([\d]{11}\.[\d]{5}$)\|([\d]{12}\.[\d]{4}$))', + uri) is not None: + gs1companyprefix = uri[16:partition] + shipperref = uri[(partition + 1):] + rawGSIN = gs1companyprefix + shipperref + return ('https://id.gs1.org/402/' + str(GTIN(raw=rawGSIN))) + + if match( + r'^urn:epc:id:itip:(([\d]{6}\.[\d]{7})\|([\d]{7}\.[\d]{6})\|([\d]{8}\.[\d]{5})\|([\d]{9}\.[\d]{4})\|([\d]{10}\.[\d]{3})\|([\d]{11}\.[\d]{2})\|([\d]{12}\.[\d]{1}))\.[\d]{2}\.[\d]{2}\.(\%2[125-9A-Fa-f]\|\%3[0-9A-Fa-f]\|\%4[1-9A-Fa-f]\|\%5[0-9AaFf]\|\%6[1-9A-Fa-f]\|\%7[0-9Aa]\|[!\')(+,.0-9:;=A-Za-z_-]){1,20}$', + uri) is not None: + gs1companyprefix = uri[16:partition] + itemref = uri[(partition + 1):(partition + 1 + (13 - len(gs1companyprefix)))] + rawGTIN = itemref[0:1] + gs1companyprefix + itemref[1:] + piece = uri[31:33] + total = uri[34:36] + serial = uri[37:] + return ('https://id.gs1.org/8006/' + str(GTIN(raw=rawGTIN)) + + piece + total + '/21/' + __web_uri_percent_encoder(serial)) + + if match( + r'^urn:epc:id:upui:((\d{6}\.\d{7})\|(\d{7}\.\d{6})\|(\d{8}\.\d{5})\|(\d{9}\.\d{4})\|(\d{10}\.\d{3})\|(\d{11}\.\d{2})\|(\d{12}\.\d{1}))\.(\%2[125-9A-Fa-f]\|\%3[0-9A-Fa-f]\|\%4[1-9A-Fa-f]\|\%5[0-9AaFf]\|\%6[1-9A-Fa-f]\|\%7[0-9Aa]\|[!\')(+,.0-9:;=A-Za-z_-]){1,28}$', + uri) is not None: + gs1companyprefix = uri[16:partition] + itemref = uri[(partition + 1):(partition + 1 + (13 - len(gs1companyprefix)))] + rawGTIN = itemref[0:1] + gs1companyprefix + itemref[1:] + serial = uri[31:] + return ('https://id.gs1.org/01/' + str(GTIN(raw=rawGTIN)) + '/235/' + __web_uri_percent_encoder(serial)) + + if match( + r'^urn:epc:id:pgln:(([\d]{6}\.[\d]{6})\|([\d]{7}\.[\d]{5})\|([\d]{8}\.[\d]{4})\|([\d]{9}\.[\d]{3})\|([\d]{10}\.[\d]{2})\|([\d]{11}\.[\d]{1})\|([\d]{12}\.))$', + uri) is not None: + gs1companyprefix = uri[16:partition] + partyref = uri[(partition + 1):(partition + 1 + (12 - len(gs1companyprefix)))] + rawGLN = gs1companyprefix + partyref + return ('https://id.gs1.org/417/' + str(GTIN(raw=rawGLN))) + + # EPC Class URIs + if match( + r'^urn:epc:class:lgtin:(([\d]{6}\.[\d]{7})\|([\d]{7}\.[\d]{6})\|([\d]{8}\.[\d]{5})\|([\d]{9}\.[\d]{4})\|([\d]{10}\.[\d]{3})\|([\d]{11}\.[\d]{2})\|([\d]{12}\.[\d]{1}))\.(\%2[125-9A-Fa-f]\|\%3[0-9A-Fa-f]\|\%4[1-9A-Fa-f]\|\%5[0-9AaFf]\|\%6[1-9A-Fa-f]\|\%7[0-9Aa]\|[!\')(+,.0-9:;=A-Za-z_-]){1,20}$', + uri) is not None: + gs1companyprefix = uri[20:partition] + itemref = uri[(partition + 1):(partition + 1 + (13 - len(gs1companyprefix)))] + rawGTIN = itemref[0:1] + gs1companyprefix + itemref[1:] + lot = uri[35:] + return ('https://id.gs1.org/01/' + str(GTIN(raw=rawGTIN)) + '/10/' + __web_uri_percent_encoder(lot)) + + # EPC ID Pattern URIs + if match( + r'^urn:epc:idpat:sgtin:((\d{6}\.\d{7})\|(\d{7}\.\d{6})\|(\d{8}\.\d{5})\|(\d{9}\.\d{4})\|(\d{10}\.\d{3})\|(\d{11}\.\d{2})\|(\d{12}\.\d{1}))\.\$', + uri) is not None: + gs1companyprefix = uri[20:partition] + itemref = uri[(partition + 1):(partition + 1 + (13 - len(gs1companyprefix)))] + rawGTIN = itemref[0:1] + gs1companyprefix + itemref[1:] + return ('https://id.gs1.org/01/' + str(GTIN(raw=rawGTIN))) + + if match( + r'^urn:epc:idpat:grai:(([\d]{6}\.[\d]{6})\|([\d]{7}\.[\d]{5})\|([\d]{8}\.[\d]{4})\|([\d]{9}\.[\d]{3})\|([\d]{10}\.[\d]{2})\|([\d]{11}\.[\d]{1})\|([\d]{12}\.\.))\.\$', + uri) is not None: + gs1companyprefix = uri[19:partition] + assetref = uri[(partition + 1):(partition + 1 + (12 - len(gs1companyprefix)))] + rawGRAI = '0' + gs1companyprefix + assetref + return ('https://id.gs1.org/8003/' + str(GTIN(raw=rawGRAI))) + + if match( + r'^urn:epc:idpat:gdti:(([\d]{6}\.[\d]{6})\|([\d]{7}\.[\d]{5})\|([\d]{8}\.[\d]{4})\|([\d]{9}\.[\d]{3})\|([\d]{10}\.[\d]{2})\|([\d]{11}\.[\d]{1})\|([\d]{12}\.\.))\.\$', + uri) is not None: + gs1companyprefix = uri[19:partition] + documenttype = uri[(partition + 1):(partition + 1 + + (12 - len(gs1companyprefix)))] + rawGDTI = gs1companyprefix + documenttype + return ('https://id.gs1.org/253/' + str(GTIN(raw=rawGDTI))) + + if match( + r'^urn:epc:idpat:sgcn:(([\d]{6}\.[\d]{6})\|([\d]{7}\.[\d]{5})\|([\d]{8}\.[\d]{4})\|([\d]{9}\.[\d]{3})\|([\d]{10}\.[\d]{2})\|([\d]{11}\.[\d]{1})\|([\d]{12}\.\.))\.\$', + uri) is not None: + gs1companyprefix = uri[19:partition] + couponref = uri[(partition + 1):(partition + 1 + (12 - len(gs1companyprefix)))] + rawSGCN = gs1companyprefix + couponref + return ('https://id.gs1.org/255/' + str(GTIN(raw=rawSGCN))) + + if match( + r'^urn:epc:idpat:cpi:((\d{6}\.(\%2[3dfDF]\|\%3[0-9]\|\%4[1-9A-Fa-f]\|\%5[0-9Aa]\|[0-9A-Z-]){1,24})\|(\d{7}\.(\%2[3dfDF]\|\%3[0-9]\|\%4[1-9A-Fa-f]\|\%5[0-9Aa]\|[0-9A-Z-]){1,23})\|(\d{8}\.(\%2[3dfDF]\|\%3[0-9]\|\%4[1-9A-Fa-f]\|\%5[0-9Aa]\|[0-9A-Z-]){1,22})\|(\d{9}\.(\%2[3dfDF]\|\%3[0-9]\|\%4[1-9A-Fa-f]\|\%5[0-9Aa]\|[0-9A-Z-]){1,21})\|(\d{10}\.(\%2[3dfDF]\|\%3[0-9]\|\%4[1-9A-Fa-f]\|\%5[0-9Aa]\|[0-9A-Z-]){1,20})\|(\d{11}\.(\%2[3dfDF]\|\%3[0-9]\|\%4[1-9A-Fa-f]\|\%5[0-9Aa]\|[0-9A-Z-]){1,19})\|(\d{12}\.(\%2[3dfDF]\|\%3[0-9]\|\%4[1-9A-Fa-f]\|\%5[0-9Aa]\|[0-9A-Z-]){1,18}))\.\$', + uri) is not None: + gs1companyprefix = uri[18:partition] + separator = uri.rfind('.') + cpref = uri[(partition + 1):(separator)] + rawCPI = gs1companyprefix + cpref + return ('https://id.gs1.org/8010/' + __web_uri_percent_encoder(rawCPI)) + + if match( + r'^urn:epc:idpat:itip:(([\d]{6}\.[\d]{7})\|([\d]{7}\.[\d]{6})\|([\d]{8}\.[\d]{5})\|([\d]{9}\.[\d]{4})\|([\d]{10}\.[\d]{3})\|([\d]{11}\.[\d]{2})\|([\d]{12}\.[\d]{1}))\.[\d]{2}\.[\d]{2}\.\$', + uri) is not None: + gs1companyprefix = uri[19:partition] + itemref = uri[(partition + 1):(partition + 1 + (13 - len(gs1companyprefix)))] + rawGTIN = itemref[0:1] + gs1companyprefix + itemref[1:] + piece = uri[34:36] + total = uri[37:39] + return ('https://id.gs1.org/8006/' + str(GTIN(raw=rawGTIN)) + piece + total) + + if match( + r'^urn:epc:idpat:upui:((\d{6}\.\d{7})\|(\d{7}\.\d{6})\|(\d{8}\.\d{5})\|(\d{9}\.\d{4})\|(\d{10}\.\d{3})\|(\d{11}\.\d{2})\|(\d{12}\.\d{1}))\.\$', + uri) is not None: + gs1companyprefix = uri[19:partition] + itemref = uri[(partition + 1):(partition + 1 + (13 - len(gs1companyprefix)))] + rawGTIN = itemref[0:1] + gs1companyprefix + itemref[1:] + return ('https://id.gs1.org/01/' + str(GTIN(raw=rawGTIN))) + + # GS1 DL URIs + if match( + r'^https?:(\/\/((([^\/?#])@)?([^\/?#:])(:([^\/?#]))?))?((([^?#])(\/(01\|gtin\|8006\|itip\|8010\|cpid\|414\|gln\|417\|party\|8017\|gsrnp\|8018\|gsrn\|255\|gcn\|00\|sscc\|253\|gdti\|401\|ginc\|402\|gsin\|8003\|grai\|8004\|giai)\/)(\d{4}[^\/]+)(\/[^/]+\/[^/]+)?[/]?(\?([^?\n]))?(#([^\n]))?)\|(\/[A-Za-z_-]{10}$))', + uri) is None: + return None + + # remove query string + if uri.find('?') >= 0: + uri = uri[:uri.index('?')] + + # replace short names for keys/key extensions with AIs + uri = (uri.replace('/gtin/', '/01/') + .replace('/itip/', '/8006/') + .replace('/cpid/', '/8010/') + .replace('/gln/', '/414/') + .replace('/party/', '/417/') + .replace('/gsrnp/', '/8017/') + .replace('/gsrn/', '/8018/') + .replace('/gcn/', '/255/') + .replace('/sscc/', '/00/') + .replace('/gdti/', '/253/') + .replace('/ginc/', '/401/') + .replace('/gsin/', '/402/') + .replace('/grai/', '/8003/') + .replace('/giai/', '/8004/') + .replace('/cpv/', '/22/') + .replace('/lot/', '/10/') + .replace('/ser/', '/21/')) + + # prefix with canonical domain name + if match( + r'^https:\/\/id.gs1.org\/(01\|8006\|8010\|414\|417\|8017\|8018\|255\|00\|253\|401\|402\|8003\|8004)\/(\d{4}[^\/]+)(\/[^\/]+\/[^\/]+)?[\/]?(\?([^?\n]))?(#([^\n]))?\|(\/[A-Za-z_-]{10}$)', + uri) is None: + if uri.find('/00/') != -1: + uri = 'https://id.gs1.org' + uri[(uri.find('/00/')):] + elif uri.find('/01/') != -1: + uri = 'https://id.gs1.org' + uri[(uri.find('/01/')):] + elif uri.find('/253/') != -1: + uri = 'https://id.gs1.org' + uri[(uri.find('/253/')):] + elif uri.find('/255/') != -1: + uri = 'https://id.gs1.org' + uri[(uri.find('/255/')):] + elif uri.find('/401/') != -1: + uri = 'https://id.gs1.org' + uri[(uri.find('/401/')):] + elif uri.find('/402/') != -1: + uri = 'https://id.gs1.org' + uri[(uri.find('/402/')):] + elif uri.find('/414/') != -1: + uri = 'https://id.gs1.org' + uri[(uri.find('/414/')):] + elif uri.find('/417/') != -1: + uri = 'https://id.gs1.org' + uri[(uri.find('/417/')):] + elif uri.find('/8003/') != -1: + uri = 'https://id.gs1.org' + uri[(uri.find('/8003/')):] + elif uri.find('/8004/') != -1: + uri = 'https://id.gs1.org' + uri[(uri.find('/8004/')):] + elif uri.find('/8006/') != -1: + uri = 'https://id.gs1.org' + uri[(uri.find('/8006/')):] + elif uri.find('/8010/') != -1: + uri = 'https://id.gs1.org' + uri[(uri.find('/8010/')):] + elif uri.find('/8017/') != -1: + uri = 'https://id.gs1.org' + uri[(uri.find('/8017/')):] + elif uri.find('/8018/') != -1: + uri = 'https://id.gs1.org' + uri[(uri.find('/8018/')):] + + # ensure that all GTIN formats are padded to 14 digits + if match(r'^https:\/\/id.gs1.org\/01\/\d{14}', uri) is None: + if match(r'^https:\/\/id.gs1.org\/01\/\d{13}', uri) is not None: + uri = uri.replace('/01/', '/01/0') + elif match(r'^https:\/\/id.gs1.org\/01\/\d{12}', uri) is not None: + uri = uri.replace('/01/', '/01/00') + elif match(r'^https:\/\/id.gs1.org\/01\/\d{8}', uri) is not None: + uri = uri.replace('/01/', '/01/000000') + + # remove cpv + x = uri[(uri.find('/22/') + 4):] + + # for 01/8006 only: + if ( + match( + r'https:\/\/id.gs1.org\/8006\/\d{18}\/22\/([\x22\x27\x2D\x2E\x30-\x39\x3B-\x3F\x41-\x5A\x5F\x61-\x7A]{0,20})$', + uri) or match( + r'https:\/\/id.gs1.org\/01\/\d{14}\/22\/([\x22\x27\x2D\x2E\x30-\x39\x3B-\x3F\x41-\x5A\x5F\x61-\x7A]{0,20})$', + uri)) is not None: + uri = (uri[:(uri.find('/22/'))]) + (x[x.find('/'):-1]) + + # for 01/8006 followed by other key qualifiers: + if ( + match( + r'https:\/\/id.gs1.org\/8006\/\d{18}\/22\/([\x2F\x22\x27\x2D\x2E\x30-\x39\x3B-\x3F\x41-\x5A\x5F\x61-\x7A]{0,20})$', + uri) or match( + r'https:\/\/id.gs1.org\/01\/\d{14}\/22\/([\x2F\x22\x27\x2D\x2E\x30-\x39\x3B-\x3F\x41-\x5A\x5F\x61-\x7A]{0,20})$', + uri)) is not None: + uri = (uri[:(uri.find('/22/'))]) + (x[x.find('/'):]) + + # take only lowest ID granularity level (i.e. if serial is present, omit lot) + if (match( + r'https:\/\/id.gs1.org\/8006\/\d{18}\/10\/([\x22\x27\x2D\x2E\x30-\x39\x3B-\x3F\x41-\x5A\x5F\x61-\x7A]{0,20})\/21\/([\x22\x27\x2D\x2E\x30-\x39\x3B-\x3F\x41-\x5A\x5F\x61-\x7A]{0,20})$', + uri) or match( + r'https:\/\/id.gs1.org\/01\/(\d{14})\/10\/([\x22\x27\x2D\x2E\x30-\x39\x3B-\x3F\x41-\x5A\x5F\x61-\x7A]{0,20})\/21\/([\x22\x27\x2D\x2E\x30-\x39\x3B-\x3F\x41-\x5A\x5F\x61-\x7A]{0,20})$', + uri)) is not None: + y = uri[(uri.find('/10/') + 4):] + uri = (uri[:(uri.find('/10/'))]) + (y[y.find('/'):]) + + # ensure that output has a valid syntax + if (match(r'https:\/\/id.gs1.org\/00\/(\d{18})$', uri) or + match( + r'https:\/\/id.gs1.org\/01\/(\d{14})\/21\/([\x22\x27\x2D\x2E\x30-\x39\x3B-\x3F\x41-\x5A\x5F\x61-\x7A]{0,20})$', + uri) or + match( + r'https:\/\/id.gs1.org\/01\/(\d{14})\/10\/([\x22\x27\x2D\x2E\x30-\x39\x3B-\x3F\x41-\x5A\x5F\x61-\x7A]{0,20})$', + uri) or + match(r'https:\/\/id.gs1.org\/01\/(\d{14})$', uri) or + match( + r'https:\/\/id.gs1.org\/01\/(\d{14})\/235\/([\x22\x27\x2D\x2E\x30-\x39\x3B-\x3F\x41-\x5A\x5F\x61-\x7A]{0,28})$', + uri) or + match(r'https:\/\/id.gs1.org\/253\/(\d{13})([\x22\x27\x2D\x2E\x30-\x39\x3B-\x3F\x41-\x5A\x5F\x61-\x7A]{0,17})$', + uri) or + match(r'https:\/\/id.gs1.org\/255\/(\d{13})(\d{0,12})$', uri) or + match(r'https:\/\/id.gs1.org\/401\/([\x22\x27\x2D\x2E\x30-\x39\x3B-\x3F\x41-\x5A\x5F\x61-\x7A]{0,30})$', uri) or + match(r'https:\/\/id.gs1.org\/402\/(\d{17})$', uri) or + match(r'https:\/\/id.gs1.org\/414\/(\d{13})$', uri) or + match( + r'https:\/\/id.gs1.org\/414\/(\d{13})\/254\/([\x22\x27\x2D\x2E\x30-\x39\x3B-\x3F\x41-\x5A\x5F\x61-\x7A]{0,20})$', + uri) or + match(r'https:\/\/id.gs1.org\/417\/(\d{13})$', uri) or + match( + r'https:\/\/id.gs1.org\/8003\/(\d{14})([\x22\x27\x2D\x2E\x30-\x39\x3B-\x3F\x41-\x5A\x5F\x61-\x7A]{0,16})$', + uri) or + match(r'https:\/\/id.gs1.org\/8004\/([\x22\x27\x2D\x2E\x30-\x39\x3B-\x3F\x41-\x5A\x5F\x61-\x7A]{0,30})$', + uri) or + match( + r'https:\/\/id.gs1.org\/8006\/(\d{18})\/21\/([\x22\x27\x2D\x2E\x30-\x39\x3B-\x3F\x41-\x5A\x5F\x61-\x7A]{0,20})$', + uri) or + match( + r'https:\/\/id.gs1.org\/8006\/(\d{18})\/10\/([\x22\x27\x2D\x2E\x30-\x39\x3B-\x3F\x41-\x5A\x5F\x61-\x7A]{0,20})$', + uri) or + match(r'https:\/\/id.gs1.org\/8006\/(\d{18})$', uri) or + match(r'https:\/\/id.gs1.org\/8010\/([\x23\x2D\x2F\x30-\x39\x41-\x5A]{0,30})\/8011/(\d{0,12})$', uri) or + match(r'https:\/\/id.gs1.org\/8010\/([\x23\x2D\x2F\x30-\x39\x41-\x5A]{0,30})$', uri) or + match(r'https:\/\/id.gs1.org\/8017\/(\d{18})$', uri) or + match(r'https:\/\/id.gs1.org\/8018\/(\d{18})$', uri) + ) is not None: # noqa E124 + return (uri) + return None diff --git a/epcis_event_hash_generator/hash_generator.py b/epcis_event_hash_generator/hash_generator.py index c172135..0516090 100644 --- a/epcis_event_hash_generator/hash_generator.py +++ b/epcis_event_hash_generator/hash_generator.py @@ -29,6 +29,7 @@ except ImportError: from context import epcis_event_hash_generator # noqa: F401 +from epcis_event_hash_generator.dl_normaliser import normaliser as dl_normaliser from epcis_event_hash_generator.xml_to_py import event_list_from_epcis_document_xml as read_xml from epcis_event_hash_generator.xml_to_py import event_list_from_epcis_document_xml_str as read_xml_str from epcis_event_hash_generator.json_to_py import event_list_from_epcis_document_json as read_json @@ -127,6 +128,10 @@ def canonize_value(text): """Run a value through all format canonizations""" text = try_format_web_vocabulary(text) text = try_format_numeric(text) + converted = dl_normaliser(text) + if converted: + logging.debug("Converted %s to %s", text, converted) + return converted return text @@ -163,6 +168,7 @@ def generic_child_list_to_prehash_string(children): for child in children: text = child[1].strip() if text: + text = canonize_value(text) text = "=" + text list_of_values.append(child[0] + text + generic_child_list_to_prehash_string(child[2])) @@ -197,6 +203,7 @@ def compute_prehash_from_file(path, enforce=None): events = read_json(path) else: logging.error("Filename '%s' ending not recognized.", path) + return None return compute_prehash_from_events(events) @@ -287,4 +294,4 @@ def calculate_hash(prehash_string_list, hashalg="sha256"): hashValueList.append(hash_string) - return (hashValueList, prehash_string_list) + return hashValueList, prehash_string_list diff --git a/requirements.txt b/requirements.txt index dd70786..13ca11d 100644 --- a/requirements.txt +++ b/requirements.txt @@ -1,2 +1,3 @@ -python_dateutil==2.8.1 -Flask==1.1.2 +python_dateutil>=2.8 +Flask>=1.1 +gtin>=0.1.13 \ No newline at end of file diff --git a/setup.py b/setup.py index bdd42ec..78472a5 100644 --- a/setup.py +++ b/setup.py @@ -29,6 +29,7 @@ }, install_requires=[ 'python_dateutil>=2.8', - 'Flask>=1.1' + 'Flask>=1.1', + 'gtin>=0.1.13' ], )	Digital Link Syntax for EPCs Ralph wrote a separate Normalizer for this		2021-03-02T10:33:55	[]	[]
RalphTro/epcis-event-hash-generator	RalphTro__epcis-event-hash-generator-35	621877b5ad49d0260cedf4c19b45fa2a06f17f58	diff --git a/epcis_event_hash_generator/context.py b/epcis_event_hash_generator/context.py index df8b0df..913be06 100644 --- a/epcis_event_hash_generator/context.py +++ b/epcis_event_hash_generator/context.py @@ -1,5 +1,6 @@ -import epcis_event_hash_generator import os import sys sys.path.insert(0, os.path.abspath(os.path.join(os.path.dirname(__file__), '..'))) + +import epcis_event_hash_generator diff --git a/epcis_event_hash_generator/hash_generator.py b/epcis_event_hash_generator/hash_generator.py index 2686da0..7e7084d 100644 --- a/epcis_event_hash_generator/hash_generator.py +++ b/epcis_event_hash_generator/hash_generator.py @@ -69,7 +69,7 @@ def child_to_pre_hash_string(child, sub_child_order): if child[0].lower().find("time") > 0 and child[0].lower().find("offset") < 0: text = fix_time_stamp_format(text) else: - text = format_if_numeric(text) + text = canonize_value(text) if text: text = "=" + text @@ -123,8 +123,24 @@ def recurse_through_children_in_order(child_list, child_order): return pre_hash -def format_if_numeric(text): - """remove leading/trailing zeros, leading "+", etc. from numbers""" +def canonize_value(text): + """Run a value through all format canonizations""" + text = try_format_web_vocabulary(text) + text = try_format_numeric(text) + return text + + +def try_format_web_vocabulary(text): + """Replace old CBV URNs by new web vocabulary equivalents.""" + return text.replace('urn:epcglobal:cbv:bizstep:', 'https://ns.gs1.org/voc/Bizstep-' + ).replace('urn:epcglobal:cbv:disp:', 'https://ns.gs1.org/voc/Disp-' + ).replace('urn:epcglobal:cbv:btt:', 'https://ns.gs1.org/voc/BTT-' + ).replace('urn:epcglobal:cbv:sdt:', 'https://ns.gs1.org/voc/SDT-' + ).replace('urn:epcglobal:cbv:er:', 'https://ns.gs1.org/voc/ER-') + + +def try_format_numeric(text): + """remove leading/trailing zeros, leading "+", etc. from numbers. Non numeric values are left untouched.""" try: numeric = float(text) if int(numeric) == numeric: # remove trailing .0	GS1 Web Voc Canonicalisation of standard voc elements See # 14: If present, any URN-based standard vocabulary value (starting with ‘urn:epcglobal:cbv’) SHALL be expressed in its corresponding GS1 Web Vocabulary URI equivalent (starting with ‘https://ns.gs1.org’). Example: ‘urn:epcglobal:cbv:bizstep:receiving’ --> ‘https://ns.gs1.org/cbv/BizStep-receiving’	The following need to be replaced: ``` return (inputStr.replace('urn:epcglobal:cbv:bizstep:', 'https://gs1.org/voc/Bizstep-').replace('urn:epcglobal:cbv:disp:', 'https://gs1.org/voc/Disp-').replace('urn:epcglobal:cbv:btt:', 'https://gs1.org/voc/BTT-').replace('urn:epcglobal:cbv:sdt:', 'https://gs1.org/voc/SDT-').replace('urn:epcglobal:cbv:er:', 'https://gs1.org/voc/ER-')) ``` where the url needs to be ns.gs1....	2020-12-08T10:53:54	[]	[]
drasyl/drasyl	drasyl__drasyl-611	89fc8c3527425cb4ae8357f7386a3b9ec2cea1e1	diff --git a/drasyl-extras/src/main/java/org/drasyl/handler/connection/ConnectionHandler.java b/drasyl-extras/src/main/java/org/drasyl/handler/connection/ConnectionHandler.java index 4ca783291a..1a6b0f609d 100644 --- a/drasyl-extras/src/main/java/org/drasyl/handler/connection/ConnectionHandler.java +++ b/drasyl-extras/src/main/java/org/drasyl/handler/connection/ConnectionHandler.java @@ -1054,7 +1054,7 @@ private void segmentArrivesOnListenState(final ChannelHandlerContext ctx, } // RFC 9293: Set RCV.NXT to SEG.SEQ+1, IRS is set to SEG.SEQ, - tcb.rcvNxt(advanceSeq(seg.seq(), seg.len())); + tcb.rcvNxt(advanceSeq(seg.seq(), 1)); tcb.irs(seg.seq()); // RFC 9293: and any other control or text should be queued for processing later. @@ -1206,7 +1206,7 @@ private void segmentArrivesOnSynSentState(final ChannelHandlerContext ctx, if (seg.isSyn()) { // RFC 9293: If the SYN bit is on and the security/compartment is acceptable, // RFC 9293: then RCV.NXT is set to SEG.SEQ+1, IRS is set to SEG.SEQ. - tcb.rcvNxt(advanceSeq(seg.seq(), seg.len())); + tcb.rcvNxt(advanceSeq(seg.seq(), 1)); tcb.irs(seg.seq()); if (seg.isAck()) { @@ -2056,7 +2056,8 @@ private boolean establishedStateProcessing(final ChannelHandlerContext ctx, // RFC 5681: given connection (TCP.UNA from [RFC793]) and (e) the advertised window in the // RFC 5681: incoming acknowledgment equals the advertised window in the last incoming acknowledgment. final boolean isRfc5681Duplicate = !tcb.retransmissionQueue().isEmpty() && - seg.len() == 0 && // SYN and FIN have length >0 by definition + seg.len() == 0 && + !seg.isAck() && !seg.isFin() && seg.ack() == tcb.sndUna() && seg.wnd() == tcb.lastAdvertisedWindow(); diff --git a/drasyl-extras/src/main/java/org/drasyl/handler/connection/ReceiveBuffer.java b/drasyl-extras/src/main/java/org/drasyl/handler/connection/ReceiveBuffer.java index 26ada7dbd8..dea3c9874f 100644 --- a/drasyl-extras/src/main/java/org/drasyl/handler/connection/ReceiveBuffer.java +++ b/drasyl-extras/src/main/java/org/drasyl/handler/connection/ReceiveBuffer.java @@ -131,6 +131,9 @@ public void receive(final ChannelHandlerContext ctx, ReferenceCountUtil.touch(seg, "ReceiveBuffer receive " + seg.toString()); final ByteBuf content = seg.content(); if (content.isReadable()) { + // (T/TCP or TCP Fast Open not implemented; SYN/FIN flag might require special attention) + assert !seg.isSyn() && !seg.isFin() : "not supported (yet)"; + if (head == null) { // first SEG to be added to RCV.WND? // SEG is located at the left edge of our RCV.WND? @@ -197,15 +200,15 @@ else if (greaterThan(seg.seq(), tcb.rcvNxt()) && lessThan(seg.seq(), add(tcb.rcv head::len, () -> head.next ); - tcb.advanceRcvNxt(ctx, head.len()); + tcb.rcvNxt(ctx, add(tcb.rcvNxt(), head.len())); addToHeadBuf(ctx, head.content()); head = head.next; size--; assert head == null \|\| lessThanOrEqualTo(tcb.rcvNxt(), head.seq()) : tcb.rcvNxt() + " must be less than or equal to " + head; } } - else if (seg.len() > 0) { - tcb.advanceRcvNxt(ctx, seg.len()); + else { + tcb.rcvNxt(ctx, seg.nxtSeq()); } } diff --git a/drasyl-extras/src/main/java/org/drasyl/handler/connection/Segment.java b/drasyl-extras/src/main/java/org/drasyl/handler/connection/Segment.java index b7a299506c..7bd161db86 100644 --- a/drasyl-extras/src/main/java/org/drasyl/handler/connection/Segment.java +++ b/drasyl-extras/src/main/java/org/drasyl/handler/connection/Segment.java @@ -316,13 +316,17 @@ public Map<SegmentOption, Object> options() { * @return the length (in segments) of this segment / public int len() { + return content().readableBytes(); + } + + public long nxtSeq() { if (isSyn() \|\| isFin()) { // the SYN and FIN flags each count as one segment (SYN and FIN are never set at the // same time) - return 1 + content().readableBytes(); + return add(seq(), 1 + len()); } else { - return content().readableBytes(); + return add(seq(), len()); } } diff --git a/drasyl-extras/src/main/java/org/drasyl/handler/connection/TransmissionControlBlock.java b/drasyl-extras/src/main/java/org/drasyl/handler/connection/TransmissionControlBlock.java index 3c3e22c6eb..334e149d73 100644 --- a/drasyl-extras/src/main/java/org/drasyl/handler/connection/TransmissionControlBlock.java +++ b/drasyl-extras/src/main/java/org/drasyl/handler/connection/TransmissionControlBlock.java @@ -38,7 +38,6 @@ import static org.drasyl.handler.connection.Segment.MIN_SEQ_NO; import static org.drasyl.handler.connection.Segment.SEG_HDR_SIZE; import static org.drasyl.handler.connection.Segment.add; -import static org.drasyl.handler.connection.Segment.advanceSeq; import static org.drasyl.handler.connection.Segment.sub; import static org.drasyl.util.NumberUtil.max; import static org.drasyl.util.NumberUtil.min; @@ -453,8 +452,8 @@ public ReceiveBuffer receiveBuffer() { Advances SND.NXT and places the {@code seg} on the outgoing segment queue. */ void send(final ChannelHandlerContext ctx, final Segment seg, final ChannelPromise promise) { - if (sndNxt == seg.seq() && seg.len() > 0) { - final long newSndNxt = add(seg.lastSeq(), 1); + if (sndNxt == seg.seq()) { + final long newSndNxt = seg.nxtSeq(); if (LOG.isTraceEnabled()) { LOG.trace("{} Send data [{},{}]. Advance SND.NXT from {} to {} (+{}).", ctx.channel(), seg.seq(), seg.lastSeq(), sndNxt, newSndNxt, Segment.sub(newSndNxt, sndNxt)); } @@ -692,9 +691,8 @@ public long ssthresh() { return ssthresh; } - public void advanceRcvNxt(final ChannelHandlerContext ctx, final int advancement) { - final long newRcvNxt = advanceSeq(rcvNxt, advancement); - if (LOG.isTraceEnabled()) { + public void rcvNxt(final ChannelHandlerContext ctx, final long newRcvNxt) { + if (LOG.isTraceEnabled() && newRcvNxt != rcvNxt) { LOG.trace("{} Advance RCV.NXT from {} to {} (+{}).", ctx.channel(), rcvNxt, newRcvNxt, Segment.sub(newRcvNxt, rcvNxt)); } rcvNxt = newRcvNxt;	Segment#len should not return +1 if SYN or FIN is set		2024-03-16T12:33:22	[]	[]
apriha/snps	apriha__snps-162	7d2abc2423e12dc97c770205079cf0fee689311e	diff --git a/.gitignore b/.gitignore index 4d9ef8d..038d98c 100644 --- a/.gitignore +++ b/.gitignore @@ -109,7 +109,7 @@ output/ resources/ # snps tests -tests/output/ +tests/output/* tests/resources/* !tests/resources/gsa_chrpos_map.txt !tests/resources/gsa_rsid_map.txt @@ -119,3 +119,7 @@ tests/input/discrepant_snps[12].csv tests/input/ftdna.csv.gz tests/input/generic.fa.gz tests/input/testvcf.vcf.gz +!tests/output/ +!tests/output/vcf_qc/ +!tests/output/vcf_chrom_prefix_chr.vcf +!tests/output/vcf_generic.vcf diff --git a/README.rst b/README.rst index 0e219cf..9a4f214 100644 --- a/README.rst +++ b/README.rst @@ -29,6 +29,7 @@ Build / Assembly Detection and Remapping Data Cleaning ````````````` +- Perform quality control (QC) / filter low quality SNPs based on `chip clusters <https://doi.org/10.1016/j.csbj.2021.06.040>`_ - Fix several common issues when loading SNPs - Sort SNPs based on chromosome and position - Deduplicate RSIDs @@ -39,6 +40,7 @@ Data Cleaning Analysis ```````` - Derive sex from SNPs +- Detect deduced genotype / chip array and chip version based on `chip clusters <https://doi.org/10.1016/j.csbj.2021.06.040>`_ - Predict ancestry from SNPs (when installed with `ezancestry <https://github.com/arvkevi/ezancestry>`_) Supported Genotype Files @@ -213,14 +215,14 @@ Ok, so far we've merged the SNPs from two files (ensuring the same build in the identifying discrepancies along the way). Then, we remapped the SNPs to Build 38. Now, let's save the merged and remapped dataset consisting of 1M+ SNPs to a tab-separated values (TSV) file: ->>> saved_snps = s.save("out.txt") +>>> saved_snps = s.to_tsv("out.txt") Saving output/out.txt >>> print(saved_snps) output/out.txt Moreover, let's get the reference sequences for this assembly and save the SNPs as a VCF file: ->>> saved_snps = s.save("out.vcf", vcf=True) +>>> saved_snps = s.to_vcf("out.vcf") Downloading resources/fasta/GRCh38/Homo_sapiens.GRCh38.dna.chromosome.1.fa.gz Downloading resources/fasta/GRCh38/Homo_sapiens.GRCh38.dna.chromosome.2.fa.gz Downloading resources/fasta/GRCh38/Homo_sapiens.GRCh38.dna.chromosome.3.fa.gz diff --git a/docs/snps.rst b/docs/snps.rst index 28ba6b1..d9110f2 100644 --- a/docs/snps.rst +++ b/docs/snps.rst @@ -11,7 +11,7 @@ SNPs :exclude-members: sort_snps, remap_snps, save_snps, snp_count, get_snp_count, not_null_snps, get_summary, get_assembly, get_chromosomes, get_chromosomes_summary, duplicate_snps, discrepant_XY_snps, heterozygous_MT_snps, heterozygous_snps, homozygous_snps, - discrepant_positions, discrepant_genotypes, discrepant_snps, is_valid + discrepant_positions, discrepant_genotypes, discrepant_snps, is_valid, save snps.ensembl ------------ @@ -36,6 +36,7 @@ snps.io.reader :members: :undoc-members: :show-inheritance: + :exclude-members: read_file snps.io.writer ~~~~~~~~~~~~~~ @@ -44,6 +45,7 @@ snps.io.writer :members: :undoc-members: :show-inheritance: + :exclude-members: write_file snps.resources -------------- diff --git a/src/snps/io/reader.py b/src/snps/io/reader.py index 3ea8765..9912f8e 100644 --- a/src/snps/io/reader.py +++ b/src/snps/io/reader.py @@ -42,6 +42,7 @@ import logging import os import re +import warnings import zipfile import zlib @@ -211,31 +212,9 @@ def read(self): @classmethod def read_file(cls, file, only_detect_source, resources, rsids): - """Read `file`. - - Parameters - ---------- - file : str or bytes - path to file to load or bytes to load - only_detect_source : bool - only detect the source of the data - resources : Resources - instance of Resources - rsids : tuple - rsids to extract if loading a VCF file - - Returns - ------- - dict - dict with the following items: - - snps (pandas.DataFrame) - dataframe of parsed SNPs - source (str) - detected source of SNPs - phased (bool) - flag indicating if SNPs are phased - """ + warnings.warn( + "This method will be removed in a future release.", DeprecationWarning + ) r = cls(file, only_detect_source, resources, rsids) return r.read() diff --git a/src/snps/io/writer.py b/src/snps/io/writer.py index 324c7b6..34c7938 100644 --- a/src/snps/io/writer.py +++ b/src/snps/io/writer.py @@ -37,11 +37,13 @@ import datetime import logging +import warnings import numpy as np import pandas as pd import snps +from snps.io import get_empty_snps_dataframe from snps.utils import save_df_as_csv, clean_str logger = logging.getLogger(__name__) @@ -57,6 +59,9 @@ def __init__( vcf=False, atomic=True, vcf_alt_unavailable=".", + vcf_chrom_prefix="", + vcf_qc_only=False, + vcf_qc_filter=False, kwargs, ): """Initialize a `Writer`. @@ -73,6 +78,12 @@ def __init__( atomically write output to a file on local filesystem vcf_alt_unavailable : str representation of VCF ALT allele when ALT is not able to be determined + vcf_chrom_prefix : str + prefix for chromosomes in VCF CHROM column + vcf_qc_only : bool + for VCF, output only SNPs that pass quality control + vcf_qc_filter : bool + for VCF, populate VCF FILTER column based on quality control results kwargs additional parameters to `pandas.DataFrame.to_csv` """ @@ -81,9 +92,21 @@ def __init__( self._vcf = vcf self._atomic = atomic self._vcf_alt_unavailable = vcf_alt_unavailable + self._vcf_chrom_prefix = vcf_chrom_prefix + self._vcf_qc_only = vcf_qc_only + self._vcf_qc_filter = vcf_qc_filter self._kwargs = kwargs def write(self): + """Write SNPs to file or buffer. + + Returns + ------- + str + path to file in output directory if SNPs were saved, else empty str + discrepant_vcf_position : pd.DataFrame + SNPs with discrepant positions discovered while saving VCF + """ if self._vcf: return self._write_vcf() else: @@ -97,38 +120,21 @@ def write_file( vcf=False, atomic=True, vcf_alt_unavailable=".", + vcf_qc_only=False, + vcf_qc_filter=False, kwargs, ): - """Save SNPs to file. - - Parameters - ---------- - snps : SNPs - SNPs to save to file or write to buffer - filename : str or buffer - filename for file to save or buffer to write to - vcf : bool - flag to save file as VCF - atomic : bool - atomically write output to a file on local filesystem - vcf_alt_unavailable : str - representation of VCF ALT allele when ALT is not able to be determined - kwargs - additional parameters to `pandas.DataFrame.to_csv` - - Returns - ------- - str - path to file in output directory if SNPs were saved, else empty str - discrepant_vcf_position : pd.DataFrame - SNPs with discrepant positions discovered while saving VCF - """ + warnings.warn( + "This method will be removed in a future release.", DeprecationWarning + ) w = cls( snps=snps, filename=filename, vcf=vcf, atomic=atomic, vcf_alt_unavailable=vcf_alt_unavailable, + vcf_qc_only=vcf_qc_only, + vcf_qc_filter=vcf_qc_filter, kwargs, ) return w.write() @@ -257,6 +263,12 @@ def _write_vcf(self): "assembly": self._snps.assembly, "chrom": chrom, "snps": pd.DataFrame(df.loc[(df["chrom"] == chrom)]), + "cluster": self._snps.cluster + if self._vcf_qc_only or self._vcf_qc_filter + else "", + "low_quality_snps": self._snps.low_quality + if self._vcf_qc_only or self._vcf_qc_filter + else get_empty_snps_dataframe(), } ) @@ -279,6 +291,10 @@ def _write_vcf(self): discrepant_vcf_position = pd.concat(discrepant_vcf_position) comment += "".join(contigs) + + if self._vcf_qc_filter and self._snps.cluster: + comment += '##FILTER=<ID=lq,Description="Low quality SNP per Lu et al.: https://doi.org/10.1016/j.csbj.2021.06.040">\n' + comment += '##FORMAT=<ID=GT,Number=1,Type=String,Description="Genotype">\n' comment += "#CHROM\tPOS\tID\tREF\tALT\tQUAL\tFILTER\tINFO\tFORMAT\tSAMPLE\n" @@ -302,6 +318,8 @@ def _create_vcf_representation(self, task): assembly = task["assembly"] chrom = task["chrom"] snps = task["snps"] + cluster = task["cluster"] + low_quality_snps = task["low_quality_snps"] if len(snps.loc[snps["genotype"].notnull()]) == 0: return { @@ -315,6 +333,16 @@ def _create_vcf_representation(self, task): contig = f'##contig=<ID={seq.ID},URL={seq.url},length={seq.length},assembly={seq.build},md5={seq.md5},species="{seq.species}">\n' + if self._vcf_qc_only and cluster: + # drop low quality SNPs if SNPs object maps to a cluster + snps = snps.drop(snps.index.intersection(low_quality_snps.index)) + + if self._vcf_qc_filter and cluster: + # initialize filter for all SNPs if SNPs object maps to a cluster, + snps["filter"] = "PASS" + # then indicate SNPs that were identified as low quality + snps.loc[snps.index.intersection(low_quality_snps.index), "filter"] = "lq" + snps = snps.reset_index() df = pd.DataFrame( @@ -346,10 +374,13 @@ def _create_vcf_representation(self, task): } ) - df["CHROM"] = snps["chrom"] + df["CHROM"] = self._vcf_chrom_prefix + snps["chrom"] df["POS"] = snps["pos"] df["ID"] = snps["rsid"] + if self._vcf_qc_filter and cluster: + df["FILTER"] = snps["filter"] + # drop SNPs with discrepant positions (outside reference sequence) discrepant_vcf_position = snps.loc[ (snps.pos - seq.start < 0) \| (snps.pos - seq.start > seq.length - 1) diff --git a/src/snps/resources.py b/src/snps/resources.py index d7ee9a1..0f7c89a 100644 --- a/src/snps/resources.py +++ b/src/snps/resources.py @@ -90,6 +90,8 @@ def _init_resource_attributes(self): self._gsa_chrpos_map = None self._dbsnp_151_37_reverse = None self._opensnp_datadump_filenames = [] + self._chip_clusters = None + self._low_quality_snps = None def get_reference_sequences( self, @@ -235,6 +237,8 @@ def get_all_resources(self): source, target ) resources["gsa_resources"] = self.get_gsa_resources() + resources["chip_clusters"] = self.get_chip_clusters() + resources["low_quality_snps"] = self.get_low_quality_snps() return resources def get_all_reference_sequences(self, kwargs): @@ -269,6 +273,90 @@ def get_gsa_resources(self): "dbsnp_151_37_reverse": self.get_dbsnp_151_37_reverse(), } + def get_chip_clusters(self): + """Get resource for identifying deduced genotype / chip array based on chip clusters. + + Returns + ------- + pandas.DataFrame + + References + ---------- + 1. Chang Lu, Bastian Greshake Tzovaras, Julian Gough, A survey of + direct-to-consumer genotype data, and quality control tool + (GenomePrep) for research, Computational and Structural + Biotechnology Journal, Volume 19, 2021, Pages 3747-3754, ISSN + 2001-0370, https://doi.org/10.1016/j.csbj.2021.06.040. + """ + if self._chip_clusters is None: + chip_clusters_path = self._download_file( + "https://supfam.mrc-lmb.cam.ac.uk/GenomePrep/datadir/the_list.tsv.gz", + "chip_clusters.tsv.gz", + ) + + df = pd.read_csv( + chip_clusters_path, + sep="\t", + names=["locus", "clusters"], + dtype={"locus": object, "clusters": pd.CategoricalDtype(ordered=False)}, + ) + clusters = df.clusters + df = df.locus.str.split(":", expand=True) + df.rename({0: "chrom", 1: "pos"}, axis=1, inplace=True) + df.pos = df.pos.astype(np.uint32) + df.chrom = df.chrom.astype(pd.CategoricalDtype(ordered=False)) + df["clusters"] = clusters + + self._chip_clusters = df + + return self._chip_clusters + + def get_low_quality_snps(self): + """Get listing of low quality SNPs for quality control based on chip clusters. + + Returns + ------- + pandas.DataFrame + + References + ---------- + 1. Chang Lu, Bastian Greshake Tzovaras, Julian Gough, A survey of + direct-to-consumer genotype data, and quality control tool + (GenomePrep) for research, Computational and Structural + Biotechnology Journal, Volume 19, 2021, Pages 3747-3754, ISSN + 2001-0370, https://doi.org/10.1016/j.csbj.2021.06.040. + """ + if self._low_quality_snps is None: + low_quality_snps_path = self._download_file( + "https://supfam.mrc-lmb.cam.ac.uk/GenomePrep/datadir/badalleles.tsv.gz", + "low_quality_snps.tsv.gz", + ) + + df = pd.read_csv( + low_quality_snps_path, + sep="\t", + names=["cluster", "loci"], + ) + + cluster_dfs = [] + for row in df.itertuples(): + loci = row.loci.split(",") + cluster_dfs.append( + pd.DataFrame({"cluster": [row.cluster] * len(loci), "locus": loci}) + ) + + df = pd.concat(cluster_dfs) + df.reset_index(inplace=True, drop=True) + cluster = df.cluster.astype(pd.CategoricalDtype(ordered=False)) + df = df.locus.str.split(":", expand=True) + df.rename({0: "chrom", 1: "pos"}, axis=1, inplace=True) + df.pos = df.pos.astype(np.uint32) + df.chrom = df.chrom.astype(pd.CategoricalDtype(ordered=False)) + df["cluster"] = cluster + self._low_quality_snps = df + + return self._low_quality_snps + def get_dbsnp_151_37_reverse(self): """Get and load RSIDs that are on the reference reverse (-) strand in dbSNP 151 and lower. diff --git a/src/snps/snps.py b/src/snps/snps.py index aa4250f..89e8008 100644 --- a/src/snps/snps.py +++ b/src/snps/snps.py @@ -35,6 +35,7 @@ """ +import copy from itertools import groupby, count import logging import os @@ -103,6 +104,7 @@ def __init__( self._discrepant_XY = get_empty_snps_dataframe() self._heterozygous_MT = get_empty_snps_dataframe() self._discrepant_vcf_position = get_empty_snps_dataframe() + self._low_quality = get_empty_snps_dataframe().index self._discrepant_merge_positions = pd.DataFrame() self._discrepant_merge_genotypes = pd.DataFrame() self._source = [] @@ -112,6 +114,9 @@ def __init__( self._output_dir = output_dir self._resources = Resources(resources_dir=resources_dir) self._parallelizer = Parallelizer(parallelize=parallelize, processes=processes) + self._cluster = "" + self._chip = "" + self._chip_version = "" if file: @@ -218,6 +223,30 @@ def snps(self): """ return self._snps + @property + def snps_qc(self): + """Normalized SNPs, after quality control. + + Any low quality SNPs, identified per + :meth:`identify_low_quality_snps() <snps.snps.SNPs.identify_low_quality_snps>`, + are not included in the result. + + Returns + ------- + pandas.DataFrame + normalized ``snps`` dataframe + """ + if len(self._low_quality) == 0: + # ensure low quality SNPs, if any, are identified + self.identify_low_quality_snps() + + if len(self._low_quality) > 0: + # filter out low quality SNPs + return self._snps.drop(self._low_quality) + else: + # no low quality SNPs to filter + return self._snps + @property def duplicate(self): """Duplicate SNPs. @@ -268,6 +297,22 @@ def discrepant_vcf_position(self): """ return self._discrepant_vcf_position + @property + def low_quality(self): + """SNPs identified as low quality, if any, per + :meth:`identify_low_quality_snps() <snps.snps.SNPs.identify_low_quality_snps>`. + + Returns + ------- + pandas.DataFrame + normalized ``snps`` dataframe + """ + if len(self._low_quality) == 0: + # ensure low quality SNPs, if any, are identified + self.identify_low_quality_snps() + + return self._snps.loc[self._low_quality] + @property def discrepant_merge_positions(self): """SNPs with discrepant positions discovered while merging SNPs. @@ -487,6 +532,59 @@ def phased(self): """ return self._phased + @property + def cluster(self): + """Detected chip cluster, if any, per + :meth:`compute_cluster_overlap <snps.snps.SNPs.compute_cluster_overlap>`. + + Notes + ----- + Refer to :meth:`compute_cluster_overlap <snps.snps.SNPs.compute_cluster_overlap>` + for more details about chip clusters. + + Returns + ------- + str + detected chip cluster, e.g., 'c1', else empty str + """ + if not self._cluster: + self.compute_cluster_overlap() + return self._cluster + + @property + def chip(self): + """Detected deduced genotype / chip array, if any, per + :meth:`compute_cluster_overlap <snps.snps.SNPs.compute_cluster_overlap>`. + + Returns + ------- + str + detected chip array, else empty str + """ + if not self._chip: + self.compute_cluster_overlap() + return self._chip + + @property + def chip_version(self): + """Detected genotype / chip array version, if any, per + :meth:`compute_cluster_overlap <snps.snps.SNPs.compute_cluster_overlap>`. + + Notes + ----- + Chip array version is only applicable to 23andMe (v3, v4, v5) and AncestryDNA + (v1, v2) files. + + Returns + ------- + str + detected chip array version, e.g., 'v4', else empty str + """ + + if not self._chip_version: + self.compute_cluster_overlap() + return self._chip_version + def heterozygous(self, chrom=""): """Get heterozygous SNPs. @@ -585,27 +683,41 @@ def valid(self): return True def save( - self, filename="", vcf=False, atomic=True, vcf_alt_unavailable=".", kwargs + self, + filename="", + vcf=False, + atomic=True, + vcf_alt_unavailable=".", + vcf_qc_only=False, + vcf_qc_filter=False, + kwargs, ): - """Save SNPs to file. - - Parameters - ---------- - filename : str or buffer - filename for file to save or buffer to write to - vcf : bool - flag to save file as VCF - atomic : bool - atomically write output to a file on local filesystem - vcf_alt_unavailable : str - representation of VCF ALT allele when ALT is not able to be determined - kwargs - additional parameters to `pandas.DataFrame.to_csv` + warnings.warn( + "Method `save` has been replaced by `to_csv`, `to_tsv`, and `to_vcf`.", + DeprecationWarning, + ) + return self._save( + filename, + vcf, + atomic, + vcf_alt_unavailable, + vcf_qc_only, + vcf_qc_filter, + kwargs, + ) - Returns - ------- - str - path to file in output directory if SNPs were saved, else empty str + def _save( + self, + filename="", + vcf=False, + atomic=True, + vcf_alt_unavailable=".", + vcf_chrom_prefix="", + vcf_qc_only=False, + vcf_qc_filter=False, + *kwargs, + ): + """Save SNPs to file. References ---------- @@ -615,14 +727,18 @@ def save( if "sep" not in kwargs: kwargs["sep"] = "\t" - path, extra = Writer.write_file( + w = Writer( snps=self, filename=filename, vcf=vcf, atomic=atomic, vcf_alt_unavailable=vcf_alt_unavailable, + vcf_chrom_prefix=vcf_chrom_prefix, + vcf_qc_only=vcf_qc_only, + vcf_qc_filter=vcf_qc_filter, *kwargs, ) + path, extra = w.write() if len(extra) == 1 and not extra[0].empty: self._discrepant_vcf_position = extra[0] @@ -633,11 +749,112 @@ def save( return path + def to_csv(self, filename="", atomic=True, kwargs): + """Output SNPs as comma-separated values. + + Parameters + ---------- + filename : str or buffer + filename for file to save or buffer to write to + atomic : bool + atomically write output to a file on local filesystem + kwargs + additional parameters to `pandas.DataFrame.to_csv` + + Returns + ------- + str + path to file in output directory if SNPs were saved, else empty str + """ + kwargs["sep"] = "," + return self._save(filename=filename, atomic=atomic, kwargs) + + def to_tsv(self, filename="", atomic=True, kwargs): + """Output SNPs as tab-separated values. + + Note that this results in the same default output as `save`. + + Parameters + ---------- + filename : str or buffer + filename for file to save or buffer to write to + atomic : bool + atomically write output to a file on local filesystem + kwargs + additional parameters to `pandas.DataFrame.to_csv` + + Returns + ------- + str + path to file in output directory if SNPs were saved, else empty str + """ + kwargs["sep"] = "\t" + return self._save(filename=filename, atomic=atomic, kwargs) + + def to_vcf( + self, + filename="", + atomic=True, + alt_unavailable=".", + chrom_prefix="", + qc_only=False, + qc_filter=False, + kwargs, + ): + """Output SNPs as Variant Call Format. + + Parameters + ---------- + filename : str or buffer + filename for file to save or buffer to write to + atomic : bool + atomically write output to a file on local filesystem + alt_unavailable : str + representation of ALT allele when ALT is not able to be determined + chrom_prefix : str + prefix for chromosomes in VCF CHROM column + qc_only : bool + output only SNPs that pass quality control + qc_filter : bool + populate FILTER column based on quality control results + kwargs + additional parameters to `pandas.DataFrame.to_csv` + + Returns + ------- + str + path to file in output directory if SNPs were saved, else empty str + + Notes + ----- + Parameters `qc_only` and `qc_filter`, if true, will identify low quality SNPs per + :meth:`identify_low_quality_snps() <snps.snps.SNPs.identify_low_quality_snps>`, + if not done already. Moreover, these parameters have no effect if this SNPs + object does not map to a cluster per + :meth:`compute_cluster_overlap() <snps.snps.SNPs.compute_cluster_overlap>`. + + References + ---------- + 1. The Variant Call Format (VCF) Version 4.2 Specification, 8 Mar 2019, + https://samtools.github.io/hts-specs/VCFv4.2.pdf + """ + return self._save( + filename=filename, + vcf=True, + atomic=atomic, + vcf_alt_unavailable=alt_unavailable, + vcf_chrom_prefix=chrom_prefix, + vcf_qc_only=qc_only, + vcf_qc_filter=qc_filter, + kwargs, + ) + def _filter(self, chrom=""): return self.snps.loc[self.snps.chrom == chrom] if chrom else self.snps def _read_raw_data(self, file, only_detect_source, rsids): - return Reader.read_file(file, only_detect_source, self._resources, rsids) + r = Reader(file, only_detect_source, self._resources, rsids) + return r.read() def _assign_par_snps(self): """Assign PAR SNPs to the X or Y chromosome using SNP position. @@ -1523,8 +1740,11 @@ def remap_snps(self, target_assembly, complement_bases=True): return self.remap(target_assembly, complement_bases) def save_snps(self, filename="", vcf=False, atomic=True, kwargs): - warnings.warn("This method has been renamed to `save`.", DeprecationWarning) - return self.save(filename, vcf, atomic, kwargs) + warnings.warn( + "Method `save_snps` has been replaced by `to_csv`, `to_tsv`, and `to_vcf`.", + DeprecationWarning, + ) + return self._save(filename, vcf, atomic, kwargs) @property def snp_count(self): @@ -1730,3 +1950,149 @@ def max_pop(row): d["ezancestry_df"] = predictions return d + + def compute_cluster_overlap(self, cluster_overlap_threshold=0.95): + """Compute overlap with chip clusters. + + Chip clusters, which are defined in [1]_, are associated with deduced genotype / + chip arrays and DTC companies. + + This method also sets the values returned by the `cluster`, `chip`, and + `chip_version` properties, based on max overlap, if the specified threshold is + satisfied. + + Parameters + ---------- + cluster_overlap_threshold : float + threshold for cluster to overlap this SNPs object, and vice versa, to set + values returned by the `cluster`, `chip`, and `chip_version` properties + + Returns + ------- + pandas.DataFrame + pandas.DataFrame with the following columns: + + `company_composition` + DTC company composition of associated cluster from [1]_ + `chip_base_deduced` + deduced genotype / chip array of associated cluster from [1]_ + `snps_in_cluster` + count of SNPs in cluster + `snps_in_common` + count of SNPs in common with cluster (inner merge with cluster) + `overlap_with_cluster` + percentage overlap of `snps_in_common` with cluster + `overlap_with_self` + percentage overlap of `snps_in_common` with this SNPs object + + References + ---------- + .. [1] Chang Lu, Bastian Greshake Tzovaras, Julian Gough, A survey of + direct-to-consumer genotype data, and quality control tool + (GenomePrep) for research, Computational and Structural + Biotechnology Journal, Volume 19, 2021, Pages 3747-3754, ISSN + 2001-0370, https://doi.org/10.1016/j.csbj.2021.06.040. + """ + + # information from Lu et. al (Ref. [1]_), Table 2 and Fig. 2 + df = pd.DataFrame( + data={ + "cluster_id": ["c1", "c3", "c4", "c5", "v5"], + "company_composition": [ + "23andMe-v4", + "AncestryDNA-v1, FTDNA, MyHeritage", + "23andMe-v3", + "AncestryDNA-v2", + "23andMe-v5, LivingDNA", + ], + "chip_base_deduced": [ + "HTS iSelect HD", + "OmniExpress", + "OmniExpress plus", + "OmniExpress plus", + "Illumina GSAs", + ], + "snps_in_cluster": [0] * 5, + "snps_in_common": [0] * 5, + } + ) + df.set_index("cluster_id", inplace=True) + + if self.build != 37: + to_remap = copy.deepcopy(self) + to_remap.remap(37) # clusters are relative to Build 37 + self_snps = to_remap.snps[["chrom", "pos"]].drop_duplicates() + else: + self_snps = self.snps[["chrom", "pos"]].drop_duplicates() + + chip_clusters = self._resources.get_chip_clusters() + + for cluster in df.index.values: + cluster_snps = chip_clusters.loc[ + chip_clusters.clusters.str.contains(cluster) + ][["chrom", "pos"]] + df.loc[cluster, "snps_in_cluster"] = len(cluster_snps) + df.loc[cluster, "snps_in_common"] = len( + self_snps.merge(cluster_snps, how="inner") + ) + + df["overlap_with_cluster"] = df.snps_in_common / df.snps_in_cluster + df["overlap_with_self"] = df.snps_in_common / len(self_snps) + + max_overlap = df.overlap_with_cluster.idxmax() + + if ( + df.overlap_with_cluster.loc[max_overlap] > cluster_overlap_threshold + and df.overlap_with_self.loc[max_overlap] > cluster_overlap_threshold + ): + self._cluster = max_overlap + self._chip = df.chip_base_deduced.loc[max_overlap] + + company_composition = df.company_composition.loc[max_overlap] + + if self.source in company_composition: + if self.source == "23andMe" or self.source == "AncestryDNA": + i = company_composition.find("v") + self._chip_version = company_composition[i : i + 2] + else: + logger.warning( + "Detected SNPs data source not found in cluster's company composition" + ) + + return df + + def identify_low_quality_snps(self): + """Identify low quality SNPs based on chip clusters. + + Any low quality SNPs are removed from the + :meth:`snps_qc <snps.snps.SNPs.snps_qc>` dataframe and are made + available as :meth:`low_quality <snps.snps.SNPs.low_quality>`. + + Notes + ----- + Chip clusters, which are defined in [1]_, are associated with low quality SNPs. + As such, low quality SNPs will only be identified when this SNPs object corresponds + to a cluster per + :meth:`compute_cluster_overlap() <snps.snps.SNPs.compute_cluster_overlap>`. + """ + if self.build != 37: + to_remap = copy.deepcopy(self) + to_remap.remap(37) # clusters are relative to Build 37 + self_snps = to_remap._snps[["chrom", "pos"]] + else: + self_snps = self._snps[["chrom", "pos"]] + + low_quality_snps = self._resources.get_low_quality_snps() + + if self.cluster: + cluster_snps = low_quality_snps.loc[ + low_quality_snps.cluster.str.contains(self.cluster) + ][["chrom", "pos"]] + # keep index after merge; https://stackoverflow.com/a/11982843 + merged = ( + self_snps.reset_index() + .merge(cluster_snps, how="inner") + .set_index("rsid") + ) + + self._low_quality = merged.index diff --git a/src/snps/utils.py b/src/snps/utils.py index cd36801..5c50bb1 100644 --- a/src/snps/utils.py +++ b/src/snps/utils.py @@ -43,6 +43,7 @@ import os import re import shutil +import tempfile import zipfile from atomicwrites import atomic_write @@ -182,14 +183,19 @@ def save_df_as_csv( df.to_csv(destination, kwargs) destination.seek(0) elif atomic: - with atomic_write(destination, mode="w", overwrite=True) as f: + fd, tmp_path = tempfile.mkstemp(dir=path) + + with open(fd, mode="w") as f: f.write(s) - # https://stackoverflow.com/a/29233924 - df.to_csv(f, kwargs) + + # https://stackoverflow.com/a/29233924 + df.to_csv(tmp_path, mode="a", kwargs) + + os.rename(tmp_path, destination) else: with open(destination, mode="w") as f: f.write(s) - df.to_csv(f, kwargs) + df.to_csv(destination, mode="a", **kwargs) return destination else:	Truncated position values in output files With `pandas` 1.0.0, using `na_rep` with `to_csv` causes position values in output files to be truncated (e.g., files created via `save_df_as_csv`). See https://github.com/pandas-dev/pandas/issues/31447 In the meantime, a workaround would be to pin the version of `pandas` to <=0.25.3.		2022-11-04T03:53:14	[]	[]
apriha/snps	apriha__snps-148	b3ebe4c5985c881314495d8040c28faafd9bb6d9	diff --git a/.github/workflows/ci.yml b/.github/workflows/ci.yml index 203553e..93b2819 100644 --- a/.github/workflows/ci.yml +++ b/.github/workflows/ci.yml @@ -23,7 +23,7 @@ jobs: - uses: actions/setup-python@v2 - name: Install Black run: \| - pip install black==19.10b0 + pip install black - name: Lint with Black run: \| black --check --diff . @@ -111,3 +111,29 @@ jobs: uses: codecov/codecov-action@v1 with: fail_ci_if_error: true + + test-extras: + needs: [test] + runs-on: ${{ matrix.os }} + strategy: + matrix: + os: [ubuntu-latest] + python-version: [3.7, 3.8, 3.9] + + steps: + - uses: actions/checkout@v2 + with: + fetch-depth: 0 + persist-credentials: false + - name: Setup Python ${{ matrix.python-version }} + uses: actions/setup-python@v2 + with: + python-version: ${{ matrix.python-version }} + - name: Install dependencies + run: \| + pip install --upgrade pip setuptools wheel + pip install pytest-cov + pip install .[ezancestry] + - name: Test with pytest + run: \| + pytest --cov=snps tests diff --git a/Pipfile b/Pipfile index 2f151b3..5f0f1c1 100644 --- a/Pipfile +++ b/Pipfile @@ -9,7 +9,7 @@ pytest-cov = "" pytest-watch = "" sphinx = "" sphinx-rtd-theme = "" -black = "==19.10b0" +black = "==21.10b0" matplotlib = "" [packages] diff --git a/Pipfile.lock b/Pipfile.lock index 38791ea..f16991e 100644 --- a/Pipfile.lock +++ b/Pipfile.lock @@ -1,7 +1,7 @@ { "_meta": { "hash": { - "sha256": "11f3af56d80f81e99d56db7de554f1cf85aa21768fb2f867d122e29efdd99f0e" + "sha256": "1cbbd1515d1a00b1feb0abee9bbec0adffdd45f8a1ad655653a67c8de53742ad" }, "pipfile-spec": 6, "requires": { @@ -26,70 +26,87 @@ }, "numpy": { "hashes": [ - "sha256:1676b0a292dd3c99e49305a16d7a9f42a4ab60ec522eac0d3dd20cdf362ac010", - "sha256:16f221035e8bd19b9dc9a57159e38d2dd060b48e93e1d843c49cb370b0f415fd", - "sha256:43909c8bb289c382170e0282158a38cf306a8ad2ff6dfadc447e90f9961bef43", - "sha256:4e465afc3b96dbc80cf4a5273e5e2b1e3451286361b4af70ce1adb2984d392f9", - "sha256:55b745fca0a5ab738647d0e4db099bd0a23279c32b31a783ad2ccea729e632df", - "sha256:5d050e1e4bc9ddb8656d7b4f414557720ddcca23a5b88dd7cff65e847864c400", - "sha256:637d827248f447e63585ca3f4a7d2dfaa882e094df6cfa177cc9cf9cd6cdf6d2", - "sha256:6690080810f77485667bfbff4f69d717c3be25e5b11bb2073e76bb3f578d99b4", - "sha256:66fbc6fed94a13b9801fb70b96ff30605ab0a123e775a5e7a26938b717c5d71a", - "sha256:67d44acb72c31a97a3d5d33d103ab06d8ac20770e1c5ad81bdb3f0c086a56cf6", - "sha256:6ca2b85a5997dabc38301a22ee43c82adcb53ff660b89ee88dded6b33687e1d8", - "sha256:6e51534e78d14b4a009a062641f465cfaba4fdcb046c3ac0b1f61dd97c861b1b", - "sha256:70eb5808127284c4e5c9e836208e09d685a7978b6a216db85960b1a112eeace8", - "sha256:830b044f4e64a76ba71448fce6e604c0fc47a0e54d8f6467be23749ac2cbd2fb", - "sha256:8b7bb4b9280da3b2856cb1fc425932f46fba609819ee1c62256f61799e6a51d2", - "sha256:a9c65473ebc342715cb2d7926ff1e202c26376c0dcaaee85a1fd4b8d8c1d3b2f", - "sha256:c1c09247ccea742525bdb5f4b5ceeacb34f95731647fe55774aa36557dbb5fa4", - "sha256:c5bf0e132acf7557fc9bb8ded8b53bbbbea8892f3c9a1738205878ca9434206a", - "sha256:db250fd3e90117e0312b611574cd1b3f78bec046783195075cbd7ba9c3d73f16", - "sha256:e515c9a93aebe27166ec9593411c58494fa98e5fcc219e47260d9ab8a1cc7f9f", - "sha256:e55185e51b18d788e49fe8305fd73ef4470596b33fc2c1ceb304566b99c71a69", - "sha256:ea9cff01e75a956dbee133fa8e5b68f2f92175233de2f88de3a682dd94deda65", - "sha256:f1452578d0516283c87608a5a5548b0cdde15b99650efdfd85182102ef7a7c17", - "sha256:f39a995e47cb8649673cfa0579fbdd1cdd33ea497d1728a6cb194d6252268e48" - ], - "markers": "python_version >= '3.7'", - "version": "==1.20.3" + "sha256:043e83bfc274649c82a6f09836943e4a4aebe5e33656271c7dbf9621dd58b8ec", + "sha256:160ccc1bed3a8371bf0d760971f09bfe80a3e18646620e9ded0ad159d9749baa", + "sha256:188031f833bbb623637e66006cf75e933e00e7231f67e2b45cf8189612bb5dc3", + "sha256:28f15209fb535dd4c504a7762d3bc440779b0e37d50ed810ced209e5cea60d96", + "sha256:29fb3dcd0468b7715f8ce2c0c2d9bbbaf5ae686334951343a41bd8d155c6ea27", + "sha256:2a6ee9620061b2a722749b391c0d80a0e2ae97290f1b32e28d5a362e21941ee4", + "sha256:300321e3985c968e3ae7fbda187237b225f3ffe6528395a5b7a5407f73cf093e", + "sha256:32437f0b275c1d09d9c3add782516413e98cd7c09e6baf4715cbce781fc29912", + "sha256:3c09418a14471c7ae69ba682e2428cae5b4420a766659605566c0fa6987f6b7e", + "sha256:49c6249260890e05b8111ebfc391ed58b3cb4b33e63197b2ec7f776e45330721", + "sha256:4cc9b512e9fb590797474f58b7f6d1f1b654b3a94f4fa8558b48ca8b3cfc97cf", + "sha256:508b0b513fa1266875524ba8a9ecc27b02ad771fe1704a16314dc1a816a68737", + "sha256:50cd26b0cf6664cb3b3dd161ba0a09c9c1343db064e7c69f9f8b551f5104d654", + "sha256:5c4193f70f8069550a1788bd0cd3268ab7d3a2b70583dfe3b2e7f421e9aace06", + "sha256:5dfe9d6a4c39b8b6edd7990091fea4f852888e41919d0e6722fe78dd421db0eb", + "sha256:63571bb7897a584ca3249c86dd01c10bcb5fe4296e3568b2e9c1a55356b6410e", + "sha256:75621882d2230ab77fb6a03d4cbccd2038511491076e7964ef87306623aa5272", + "sha256:75eb7cadc8da49302f5b659d40ba4f6d94d5045fbd9569c9d058e77b0514c9e4", + "sha256:88a5d6b268e9ad18f3533e184744acdaa2e913b13148160b1152300c949bbb5f", + "sha256:8a10968963640e75cc0193e1847616ab4c718e83b6938ae74dea44953950f6b7", + "sha256:90bec6a86b348b4559b6482e2b684db4a9a7eed1fa054b86115a48d58fbbf62a", + "sha256:98339aa9911853f131de11010f6dd94c8cec254d3d1f7261528c3b3e3219f139", + "sha256:a99a6b067e5190ac6d12005a4d85aa6227c5606fa93211f86b1dafb16233e57d", + "sha256:bffa2eee3b87376cc6b31eee36d05349571c236d1de1175b804b348dc0941e3f", + "sha256:c6c2d535a7beb1f8790aaa98fd089ceab2e3dd7ca48aca0af7dc60e6ef93ffe1", + "sha256:cc14e7519fab2a4ed87d31f99c31a3796e4e1fe63a86ebdd1c5a1ea78ebd5896", + "sha256:dd0482f3fc547f1b1b5d6a8b8e08f63fdc250c58ce688dedd8851e6e26cff0f3", + "sha256:dde972a1e11bb7b702ed0e447953e7617723760f420decb97305e66fb4afc54f", + "sha256:e54af82d68ef8255535a6cdb353f55d6b8cf418a83e2be3569243787a4f4866f", + "sha256:e606e6316911471c8d9b4618e082635cfe98876007556e89ce03d52ff5e8fcf0", + "sha256:f41b018f126aac18583956c54544db437f25c7ee4794bcb23eb38bef8e5e192a", + "sha256:f8f4625536926a155b80ad2bbff44f8cc59e9f2ad14cdda7acf4c135b4dc8ff2", + "sha256:fe52dbe47d9deb69b05084abd4b0df7abb39a3c51957c09f635520abd49b29dd" + ], + "markers": "python_version < '3.10' and platform_machine != 'aarch64' and platform_machine != 'arm64'", + "version": "==1.21.3" }, "pandas": { "hashes": [ - "sha256:167693a80abc8eb28051fbd184c1b7afd13ce2c727a5af47b048f1ea3afefff4", - "sha256:2111c25e69fa9365ba80bbf4f959400054b2771ac5d041ed19415a8b488dc70a", - "sha256:298f0553fd3ba8e002c4070a723a59cdb28eda579f3e243bc2ee397773f5398b", - "sha256:2b063d41803b6a19703b845609c0b700913593de067b552a8b24dd8eeb8c9895", - "sha256:2cb7e8f4f152f27dc93f30b5c7a98f6c748601ea65da359af734dd0cf3fa733f", - "sha256:52d2472acbb8a56819a87aafdb8b5b6d2b3386e15c95bde56b281882529a7ded", - "sha256:612add929bf3ba9d27b436cc8853f5acc337242d6b584203f207e364bb46cb12", - "sha256:649ecab692fade3cbfcf967ff936496b0cfba0af00a55dfaacd82bdda5cb2279", - "sha256:68d7baa80c74aaacbed597265ca2308f017859123231542ff8a5266d489e1858", - "sha256:8d4c74177c26aadcfb4fd1de6c1c43c2bf822b3e0fc7a9b409eeaf84b3e92aaa", - "sha256:971e2a414fce20cc5331fe791153513d076814d30a60cd7348466943e6e909e4", - "sha256:9db70ffa8b280bb4de83f9739d514cd0735825e79eef3a61d312420b9f16b758", - "sha256:b730add5267f873b3383c18cac4df2527ac4f0f0eed1c6cf37fcb437e25cf558", - "sha256:bd659c11a4578af740782288cac141a322057a2e36920016e0fc7b25c5a4b686", - "sha256:c601c6fdebc729df4438ec1f62275d6136a0dd14d332fc0e8ce3f7d2aadb4dd6", - "sha256:d0877407359811f7b853b548a614aacd7dea83b0c0c84620a9a643f180060950" + "sha256:003ba92db58b71a5f8add604a17a059f3068ef4e8c0c365b088468d0d64935fd", + "sha256:10e10a2527db79af6e830c3d5842a4d60383b162885270f8cffc15abca4ba4a9", + "sha256:22808afb8f96e2269dcc5b846decacb2f526dd0b47baebc63d913bf847317c8f", + "sha256:2d1dc09c0013d8faa7474574d61b575f9af6257ab95c93dcf33a14fd8d2c1bab", + "sha256:35c77609acd2e4d517da41bae0c11c70d31c87aae8dd1aabd2670906c6d2c143", + "sha256:372d72a3d8a5f2dbaf566a5fa5fa7f230842ac80f29a931fb4b071502cf86b9a", + "sha256:42493f8ae67918bf129869abea8204df899902287a7f5eaf596c8e54e0ac7ff4", + "sha256:4acc28364863127bca1029fb72228e6f473bb50c32e77155e80b410e2068eeac", + "sha256:5298a733e5bfbb761181fd4672c36d0c627320eb999c59c65156c6a90c7e1b4f", + "sha256:5ba0aac1397e1d7b654fccf263a4798a9e84ef749866060d19e577e927d66e1b", + "sha256:9707bdc1ea9639c886b4d3be6e2a45812c1ac0c2080f94c31b71c9fa35556f9b", + "sha256:a2aa18d3f0b7d538e21932f637fbfe8518d085238b429e4790a35e1e44a96ffc", + "sha256:a388960f979665b447f0847626e40f99af8cf191bce9dc571d716433130cb3a7", + "sha256:a51528192755f7429c5bcc9e80832c517340317c861318fea9cea081b57c9afd", + "sha256:b528e126c13816a4374e56b7b18bfe91f7a7f6576d1aadba5dee6a87a7f479ae", + "sha256:c1aa4de4919358c5ef119f6377bc5964b3a7023c23e845d9db7d9016fa0c5b1c", + "sha256:c2646458e1dce44df9f71a01dc65f7e8fa4307f29e5c0f2f92c97f47a5bf22f5", + "sha256:c2f44425594ae85e119459bb5abb0748d76ef01d9c08583a667e3339e134218e", + "sha256:d47750cf07dee6b55d8423471be70d627314277976ff2edd1381f02d52dbadf9", + "sha256:d99d2350adb7b6c3f7f8f0e5dfb7d34ff8dd4bc0a53e62c445b7e43e163fce63", + "sha256:dd324f8ee05925ee85de0ea3f0d66e1362e8c80799eb4eb04927d32335a3e44a", + "sha256:eaca36a80acaacb8183930e2e5ad7f71539a66805d6204ea88736570b2876a7b", + "sha256:f567e972dce3bbc3a8076e0b675273b4a9e8576ac629149cf8286ee13c259ae5", + "sha256:fe48e4925455c964db914b958f6e7032d285848b7538a5e1b19aeb26ffaea3ec" ], "markers": "python_full_version >= '3.7.1'", - "version": "==1.2.4" + "version": "==1.3.4" }, "python-dateutil": { "hashes": [ - "sha256:73ebfe9dbf22e832286dafa60473e4cd239f8592f699aa5adaf10050e6e1823c", - "sha256:75bb3f31ea686f1197762692a9ee6a7550b59fc6ca3a1f4b5d7e32fb98e2da2a" + "sha256:0123cacc1627ae19ddf3c27a5de5bd67ee4586fbdd6440d9748f8abb483d3e86", + "sha256:961d03dc3453ebbc59dbdea9e4e11c5651520a876d0f4db161e8674aae935da9" ], "markers": "python_version >= '2.7' and python_version not in '3.0, 3.1, 3.2, 3.3'", - "version": "==2.8.1" + "version": "==2.8.2" }, "pytz": { "hashes": [ - "sha256:83a4a90894bf38e243cf052c8b58f381bfe9a7a483f6a9cab140bc7f702ac4da", - "sha256:eb10ce3e7736052ed3623d49975ce333bcd712c7bb19a58b9e2089d4057d0798" + "sha256:3672058bc3453457b622aab7a1c3bfd5ab0bdae451512f6cf25f64ed37f5b87c", + "sha256:acad2d8b20a1af07d4e4c9d2e9285c5ed9104354062f275f3fcd88dcef4f1326" ], - "version": "==2021.1" + "version": "==2021.3" }, "six": { "hashes": [ @@ -112,13 +129,6 @@ ], "version": "==0.7.12" }, - "appdirs": { - "hashes": [ - "sha256:7d5d0167b2b1ba821647616af46a749d1c653740dd0d2415100fe26e27afdf41", - "sha256:a841dacd6b99318a741b166adb07e19ee71a274450e68237b4650ca1055ab128" - ], - "version": "==1.4.4" - }, "attrs": { "hashes": [ "sha256:149e90d6d8ac20db7a955ad60cf0e6881a3f20d37096140088356da6c716b0b1", @@ -137,34 +147,34 @@ }, "black": { "hashes": [ - "sha256:1b30e59be925fafc1ee4565e5e08abef6b03fe455102883820fe5ee2e4734e0b", - "sha256:c2edb73a08e9e0e6f65a0e6af18b059b8b1cdd5bef997d7a0b181df93dc81539" + "sha256:6eb7448da9143ee65b856a5f3676b7dda98ad9abe0f87fce8c59291f15e82a5b", + "sha256:a9952229092e325fe5f3dae56d81f639b23f7131eb840781947e4b2886030f33" ], "index": "pypi", - "version": "==19.10b0" + "version": "==21.10b0" }, "certifi": { "hashes": [ - "sha256:2bbf76fd432960138b3ef6dda3dde0544f27cbf8546c458e60baf371917ba9ee", - "sha256:50b1e4f8446b06f41be7dd6338db18e0990601dce795c2b1686458aa7e8fa7d8" + "sha256:78884e7c1d4b00ce3cea67b44566851c4343c120abd683433ce934a68ea58872", + "sha256:d62a0163eb4c2344ac042ab2bdf75399a71a2d8c7d47eac2e2ee91b9d6339569" ], - "version": "==2021.5.30" + "version": "==2021.10.8" }, - "chardet": { + "charset-normalizer": { "hashes": [ - "sha256:0d6f53a15db4120f2b08c94f11e7d93d2c911ee118b6b30a04ec3ee8310179fa", - "sha256:f864054d66fd9118f2e67044ac8981a54775ec5b67aed0441892edb553d21da5" + "sha256:e019de665e2bcf9c2b64e2e5aa025fa991da8720daa3c1138cadd2fd1856aed0", + "sha256:f7af805c321bfa1ce6714c51f254e0d5bb5e5834039bc17db7ebe3a4cec9492b" ], - "markers": "python_version >= '2.7' and python_version not in '3.0, 3.1, 3.2, 3.3, 3.4'", - "version": "==4.0.0" + "markers": "python_version >= '3'", + "version": "==2.0.7" }, "click": { "hashes": [ - "sha256:8c04c11192119b1ef78ea049e0a6f0463e4c48ef00a30160c704337586f3ad7a", - "sha256:fba402a4a47334742d782209a7c79bc448911afe1149d07bdabdf480b3e2f4b6" + "sha256:353f466495adaeb40b6b5f592f9f91cb22372351c84caeb068132442a4518ef3", + "sha256:410e932b050f5eed773c4cda94de75971c89cdb3155a72a0831139a79e5ecb5b" ], "markers": "python_version >= '3.6'", - "version": "==8.0.1" + "version": "==8.0.3" }, "colorama": { "hashes": [ @@ -175,69 +185,70 @@ "version": "==0.4.4" }, "coverage": { - "hashes": [ - "sha256:004d1880bed2d97151facef49f08e255a20ceb6f9432df75f4eef018fdd5a78c", - "sha256:01d84219b5cdbfc8122223b39a954820929497a1cb1422824bb86b07b74594b6", - "sha256:040af6c32813fa3eae5305d53f18875bedd079960822ef8ec067a66dd8afcd45", - "sha256:06191eb60f8d8a5bc046f3799f8a07a2d7aefb9504b0209aff0b47298333302a", - "sha256:13034c4409db851670bc9acd836243aeee299949bd5673e11844befcb0149f03", - "sha256:13c4ee887eca0f4c5a247b75398d4114c37882658300e153113dafb1d76de529", - "sha256:184a47bbe0aa6400ed2d41d8e9ed868b8205046518c52464fde713ea06e3a74a", - "sha256:18ba8bbede96a2c3dde7b868de9dcbd55670690af0988713f0603f037848418a", - "sha256:1aa846f56c3d49205c952d8318e76ccc2ae23303351d9270ab220004c580cfe2", - "sha256:217658ec7187497e3f3ebd901afdca1af062b42cfe3e0dafea4cced3983739f6", - "sha256:24d4a7de75446be83244eabbff746d66b9240ae020ced65d060815fac3423759", - "sha256:2910f4d36a6a9b4214bb7038d537f015346f413a975d57ca6b43bf23d6563b53", - "sha256:2949cad1c5208b8298d5686d5a85b66aae46d73eec2c3e08c817dd3513e5848a", - "sha256:2a3859cb82dcbda1cfd3e6f71c27081d18aa251d20a17d87d26d4cd216fb0af4", - "sha256:2cafbbb3af0733db200c9b5f798d18953b1a304d3f86a938367de1567f4b5bff", - "sha256:2e0d881ad471768bf6e6c2bf905d183543f10098e3b3640fc029509530091502", - "sha256:30c77c1dc9f253283e34c27935fded5015f7d1abe83bc7821680ac444eaf7793", - "sha256:3487286bc29a5aa4b93a072e9592f22254291ce96a9fbc5251f566b6b7343cdb", - "sha256:372da284cfd642d8e08ef606917846fa2ee350f64994bebfbd3afb0040436905", - "sha256:41179b8a845742d1eb60449bdb2992196e211341818565abded11cfa90efb821", - "sha256:44d654437b8ddd9eee7d1eaee28b7219bec228520ff809af170488fd2fed3e2b", - "sha256:4a7697d8cb0f27399b0e393c0b90f0f1e40c82023ea4d45d22bce7032a5d7b81", - "sha256:51cb9476a3987c8967ebab3f0fe144819781fca264f57f89760037a2ea191cb0", - "sha256:52596d3d0e8bdf3af43db3e9ba8dcdaac724ba7b5ca3f6358529d56f7a166f8b", - "sha256:53194af30d5bad77fcba80e23a1441c71abfb3e01192034f8246e0d8f99528f3", - "sha256:5fec2d43a2cc6965edc0bb9e83e1e4b557f76f843a77a2496cbe719583ce8184", - "sha256:6c90e11318f0d3c436a42409f2749ee1a115cd8b067d7f14c148f1ce5574d701", - "sha256:74d881fc777ebb11c63736622b60cb9e4aee5cace591ce274fb69e582a12a61a", - "sha256:7501140f755b725495941b43347ba8a2777407fc7f250d4f5a7d2a1050ba8e82", - "sha256:796c9c3c79747146ebd278dbe1e5c5c05dd6b10cc3bcb8389dfdf844f3ead638", - "sha256:869a64f53488f40fa5b5b9dcb9e9b2962a66a87dab37790f3fcfb5144b996ef5", - "sha256:8963a499849a1fc54b35b1c9f162f4108017b2e6db2c46c1bed93a72262ed083", - "sha256:8d0a0725ad7c1a0bcd8d1b437e191107d457e2ec1084b9f190630a4fb1af78e6", - "sha256:900fbf7759501bc7807fd6638c947d7a831fc9fdf742dc10f02956ff7220fa90", - "sha256:92b017ce34b68a7d67bd6d117e6d443a9bf63a2ecf8567bb3d8c6c7bc5014465", - "sha256:970284a88b99673ccb2e4e334cfb38a10aab7cd44f7457564d11898a74b62d0a", - "sha256:972c85d205b51e30e59525694670de6a8a89691186012535f9d7dbaa230e42c3", - "sha256:9a1ef3b66e38ef8618ce5fdc7bea3d9f45f3624e2a66295eea5e57966c85909e", - "sha256:af0e781009aaf59e25c5a678122391cb0f345ac0ec272c7961dc5455e1c40066", - "sha256:b6d534e4b2ab35c9f93f46229363e17f63c53ad01330df9f2d6bd1187e5eaacf", - "sha256:b7895207b4c843c76a25ab8c1e866261bcfe27bfaa20c192de5190121770672b", - "sha256:c0891a6a97b09c1f3e073a890514d5012eb256845c451bd48f7968ef939bf4ae", - "sha256:c2723d347ab06e7ddad1a58b2a821218239249a9e4365eaff6649d31180c1669", - "sha256:d1f8bf7b90ba55699b3a5e44930e93ff0189aa27186e96071fac7dd0d06a1873", - "sha256:d1f9ce122f83b2305592c11d64f181b87153fc2c2bbd3bb4a3dde8303cfb1a6b", - "sha256:d314ed732c25d29775e84a960c3c60808b682c08d86602ec2c3008e1202e3bb6", - "sha256:d636598c8305e1f90b439dbf4f66437de4a5e3c31fdf47ad29542478c8508bbb", - "sha256:deee1077aae10d8fa88cb02c845cfba9b62c55e1183f52f6ae6a2df6a2187160", - "sha256:ebe78fe9a0e874362175b02371bdfbee64d8edc42a044253ddf4ee7d3c15212c", - "sha256:f030f8873312a16414c0d8e1a1ddff2d3235655a2174e3648b4fa66b3f2f1079", - "sha256:f0b278ce10936db1a37e6954e15a3730bea96a0997c26d7fee88e6c396c2086d", - "sha256:f11642dddbb0253cc8853254301b51390ba0081750a8ac03f20ea8103f0c56b6" + "extras": [ + "toml" + ], + "hashes": [ + "sha256:0147f7833c41927d84f5af9219d9b32f875c0689e5e74ac8ca3cb61e73a698f9", + "sha256:04a92a6cf9afd99f9979c61348ec79725a9f9342fb45e63c889e33c04610d97b", + "sha256:10ab138b153e4cc408b43792cb7f518f9ee02f4ff55cd1ab67ad6fd7e9905c7e", + "sha256:2e5b9c17a56b8bf0c0a9477fcd30d357deb486e4e1b389ed154f608f18556c8a", + "sha256:326d944aad0189603733d646e8d4a7d952f7145684da973c463ec2eefe1387c2", + "sha256:359a32515e94e398a5c0fa057e5887a42e647a9502d8e41165cf5cb8d3d1ca67", + "sha256:35cd2230e1ed76df7d0081a997f0fe705be1f7d8696264eb508076e0d0b5a685", + "sha256:3b270c6b48d3ff5a35deb3648028ba2643ad8434b07836782b1139cf9c66313f", + "sha256:42a1fb5dee3355df90b635906bb99126faa7936d87dfc97eacc5293397618cb7", + "sha256:479228e1b798d3c246ac89b09897ee706c51b3e5f8f8d778067f38db73ccc717", + "sha256:4cd919057636f63ab299ccb86ea0e78b87812400c76abab245ca385f17d19fb5", + "sha256:4d8b453764b9b26b0dd2afb83086a7c3f9379134e340288d2a52f8a91592394b", + "sha256:51a441011a30d693e71dea198b2a6f53ba029afc39f8e2aeb5b77245c1b282ef", + "sha256:557594a50bfe3fb0b1b57460f6789affe8850ad19c1acf2d14a3e12b2757d489", + "sha256:572f917267f363101eec375c109c9c1118037c7cc98041440b5eabda3185ac7b", + "sha256:62512c0ec5d307f56d86504c58eace11c1bc2afcdf44e3ff20de8ca427ca1d0e", + "sha256:65ad3ff837c89a229d626b8004f0ee32110f9bfdb6a88b76a80df36ccc60d926", + "sha256:666c6b32b69e56221ad1551d377f718ed00e6167c7a1b9257f780b105a101271", + "sha256:6e994003e719458420e14ffb43c08f4c14990e20d9e077cb5cad7a3e419bbb54", + "sha256:72bf437d54186d104388cbae73c9f2b0f8a3e11b6e8d7deb593bd14625c96026", + "sha256:738e823a746841248b56f0f3bd6abf3b73af191d1fd65e4c723b9c456216f0ad", + "sha256:78287731e3601ea5ce9d6468c82d88a12ef8fe625d6b7bdec9b45d96c1ad6533", + "sha256:7833c872718dc913f18e51ee97ea0dece61d9930893a58b20b3daf09bb1af6b6", + "sha256:7e083d32965d2eb6638a77e65b622be32a094fdc0250f28ce6039b0732fbcaa8", + "sha256:8186b5a4730c896cbe1e4b645bdc524e62d874351ae50e1db7c3e9f5dc81dc26", + "sha256:8605add58e6a960729aa40c0fd9a20a55909dd9b586d3e8104cc7f45869e4c6b", + "sha256:977ce557d79577a3dd510844904d5d968bfef9489f512be65e2882e1c6eed7d8", + "sha256:994ce5a7b3d20981b81d83618aa4882f955bfa573efdbef033d5632b58597ba9", + "sha256:9ad5895938a894c368d49d8470fe9f519909e5ebc6b8f8ea5190bd0df6aa4271", + "sha256:9eb0a1923354e0fdd1c8a6f53f5db2e6180d670e2b587914bf2e79fa8acfd003", + "sha256:a00284dbfb53b42e35c7dd99fc0e26ef89b4a34efff68078ed29d03ccb28402a", + "sha256:a11a2c019324fc111485e79d55907e7289e53d0031275a6c8daed30690bc50c0", + "sha256:ab6a0fe4c96f8058d41948ddf134420d3ef8c42d5508b5a341a440cce7a37a1d", + "sha256:ae6de0e41f44794e68d23644636544ed8003ce24845f213b24de097cbf44997f", + "sha256:b1d0a1bce919de0dd8da5cff4e616b2d9e6ebf3bd1410ff645318c3dd615010a", + "sha256:b8e4f15b672c9156c1154249a9c5746e86ac9ae9edc3799ee3afebc323d9d9e0", + "sha256:bbca34dca5a2d60f81326d908d77313816fad23d11b6069031a3d6b8c97a54f9", + "sha256:bf656cd74ff7b4ed7006cdb2a6728150aaad69c7242b42a2a532f77b63ea233f", + "sha256:c40966b683d92869b72ea3c11fd6b99a091fd30e12652727eca117273fc97366", + "sha256:c95257aa2ccf75d3d91d772060538d5fea7f625e48157f8ca44594f94d41cb33", + "sha256:db2797ed7a7e883b9ab76e8e778bb4c859fc2037d6fd0644d8675e64d58d1653", + "sha256:dc5023be1c2a8b0a0ab5e31389e62c28b2453eb31dd069f4b8d1a0f9814d951a", + "sha256:e14bceb1f3ae8a14374be2b2d7bc12a59226872285f91d66d301e5f41705d4d6", + "sha256:e3c4f5211394cd0bf6874ac5d29684a495f9c374919833dcfff0bd6d37f96201", + "sha256:e76f017b6d4140a038c5ff12be1581183d7874e41f1c0af58ecf07748d36a336", + "sha256:e7d5606b9240ed4def9cbdf35be4308047d11e858b9c88a6c26974758d6225ce", + "sha256:f0f80e323a17af63eac6a9db0c9188c10f1fd815c3ab299727150cc0eb92c7a4", + "sha256:fb2fa2f6506c03c48ca42e3fe5a692d7470d290c047ee6de7c0f3e5fa7639ac9", + "sha256:ffa8fee2b1b9e60b531c4c27cf528d6b5d5da46b1730db1f4d6eee56ff282e07" ], - "markers": "python_version >= '2.7' and python_version not in '3.0, 3.1, 3.2, 3.3, 3.4' and python_version < '4'", - "version": "==5.5" + "markers": "python_version >= '3.6'", + "version": "==6.1.1" }, "cycler": { "hashes": [ - "sha256:1d8a5ae1ff6c5cf9b93e8811e581232ad8920aeec647c37316ceac982b08cb2d", - "sha256:cd7b2d1018258d7247a71425e9f26463dfb444d411c39569972f4ce586b0c9d8" + "sha256:3a27e95f763a428a739d2add979fa7494c912a32c17c4c38c4d5f082cad165a3", + "sha256:9c87405839a19696e837b3b818fed3f5f69f16f1eec1a1ad77e043dcea9c772f" ], - "version": "==0.10.0" + "markers": "python_version >= '3.6'", + "version": "==0.11.0" }, "docopt": { "hashes": [ @@ -247,19 +258,19 @@ }, "docutils": { "hashes": [ - "sha256:0c5b78adfbf7762415433f5515cd5c9e762339e23369dbe8000d84a4bf4ab3af", - "sha256:c2de3a60e9e7d07be26b7f2b00ca0309c207e06c100f9cc2a94931fc75a478fc" + "sha256:686577d2e4c32380bb50cbb22f575ed742d58168cee37e99117a854bcd88f125", + "sha256:cf316c8370a737a022b72b56874f6602acf974a37a9fba42ec2876387549fc61" ], "markers": "python_version >= '2.7' and python_version not in '3.0, 3.1, 3.2, 3.3, 3.4'", - "version": "==0.16" + "version": "==0.17.1" }, "idna": { "hashes": [ - "sha256:b307872f855b18632ce0c21c5e45be78c0ea7ae4c15c828c20788b26921eb3f6", - "sha256:b97d804b1e9b523befed77c48dacec60e6dcb0b5391d57af6a65a312a90648c0" + "sha256:84d9dd047ffa80596e0f246e2eab0b391788b0503584e8945f2368256d2735ff", + "sha256:9d643ff0a55b762d5cdb124b8eaa99c66322e2157b69160bc32796e824360e6d" ], - "markers": "python_version >= '2.7' and python_version not in '3.0, 3.1, 3.2, 3.3'", - "version": "==2.10" + "markers": "python_version >= '3'", + "version": "==3.3" }, "imagesize": { "hashes": [ @@ -271,11 +282,11 @@ }, "importlib-metadata": { "hashes": [ - "sha256:833b26fb89d5de469b24a390e9df088d4e52e4ba33b01dc5e0e4f41b81a16c00", - "sha256:b142cc1dd1342f31ff04bb7d022492b09920cb64fed867cd3ea6f80fe3ebd139" + "sha256:b618b6d2d5ffa2f16add5697cf57a46c76a56229b0ed1c438322e4e95645bd15", + "sha256:f284b3e11256ad1e5d03ab86bb2ccd6f5339688ff17a4d797a0fe7df326f23b1" ], "markers": "python_version < '3.8'", - "version": "==4.5.0" + "version": "==4.8.1" }, "iniconfig": { "hashes": [ @@ -286,81 +297,128 @@ }, "jinja2": { "hashes": [ - "sha256:1f06f2da51e7b56b8f238affdd6b4e2c61e39598a378cc49345bc1bd42a978a4", - "sha256:703f484b47a6af502e743c9122595cc812b0271f661722403114f71a79d0f5a4" + "sha256:827a0e32839ab1600d4eb1c4c33ec5a8edfbc5cb42dafa13b81f182f97784b45", + "sha256:8569982d3f0889eed11dd620c706d39b60c36d6d25843961f33f77fb6bc6b20c" ], "markers": "python_version >= '3.6'", - "version": "==3.0.1" + "version": "==3.0.2" }, "kiwisolver": { "hashes": [ - "sha256:0cd53f403202159b44528498de18f9285b04482bab2a6fc3f5dd8dbb9352e30d", - "sha256:1e1bc12fb773a7b2ffdeb8380609f4f8064777877b2225dec3da711b421fda31", - "sha256:225e2e18f271e0ed8157d7f4518ffbf99b9450fca398d561eb5c4a87d0986dd9", - "sha256:232c9e11fd7ac3a470d65cd67e4359eee155ec57e822e5220322d7b2ac84fbf0", - "sha256:31dfd2ac56edc0ff9ac295193eeaea1c0c923c0355bf948fbd99ed6018010b72", - "sha256:33449715e0101e4d34f64990352bce4095c8bf13bed1b390773fc0a7295967b3", - "sha256:401a2e9afa8588589775fe34fc22d918ae839aaaf0c0e96441c0fdbce6d8ebe6", - "sha256:44a62e24d9b01ba94ae7a4a6c3fb215dc4af1dde817e7498d901e229aaf50e4e", - "sha256:50af681a36b2a1dee1d3c169ade9fdc59207d3c31e522519181e12f1b3ba7000", - "sha256:563c649cfdef27d081c84e72a03b48ea9408c16657500c312575ae9d9f7bc1c3", - "sha256:5989db3b3b34b76c09253deeaf7fbc2707616f130e166996606c284395da3f18", - "sha256:5a7a7dbff17e66fac9142ae2ecafb719393aaee6a3768c9de2fd425c63b53e21", - "sha256:5c3e6455341008a054cccee8c5d24481bcfe1acdbc9add30aa95798e95c65621", - "sha256:5f6ccd3dd0b9739edcf407514016108e2280769c73a85b9e59aa390046dbf08b", - "sha256:72c99e39d005b793fb7d3d4e660aed6b6281b502e8c1eaf8ee8346023c8e03bc", - "sha256:78751b33595f7f9511952e7e60ce858c6d64db2e062afb325985ddbd34b5c131", - "sha256:834ee27348c4aefc20b479335fd422a2c69db55f7d9ab61721ac8cd83eb78882", - "sha256:8be8d84b7d4f2ba4ffff3665bcd0211318aa632395a1a41553250484a871d454", - "sha256:950a199911a8d94683a6b10321f9345d5a3a8433ec58b217ace979e18f16e248", - "sha256:a357fd4f15ee49b4a98b44ec23a34a95f1e00292a139d6015c11f55774ef10de", - "sha256:a53d27d0c2a0ebd07e395e56a1fbdf75ffedc4a05943daf472af163413ce9598", - "sha256:acef3d59d47dd85ecf909c359d0fd2c81ed33bdff70216d3956b463e12c38a54", - "sha256:b38694dcdac990a743aa654037ff1188c7a9801ac3ccc548d3341014bc5ca278", - "sha256:b9edd0110a77fc321ab090aaa1cfcaba1d8499850a12848b81be2222eab648f6", - "sha256:c08e95114951dc2090c4a630c2385bef681cacf12636fb0241accdc6b303fd81", - "sha256:c5518d51a0735b1e6cee1fdce66359f8d2b59c3ca85dc2b0813a8aa86818a030", - "sha256:c8fd0f1ae9d92b42854b2979024d7597685ce4ada367172ed7c09edf2cef9cb8", - "sha256:ca3820eb7f7faf7f0aa88de0e54681bddcb46e485beb844fcecbcd1c8bd01689", - "sha256:cf8b574c7b9aa060c62116d4181f3a1a4e821b2ec5cbfe3775809474113748d4", - "sha256:d3155d828dec1d43283bd24d3d3e0d9c7c350cdfcc0bd06c0ad1209c1bbc36d0", - "sha256:f8d6f8db88049a699817fd9178782867bf22283e3813064302ac59f61d95be05", - "sha256:fd34fbbfbc40628200730bc1febe30631347103fc8d3d4fa012c21ab9c11eca9" + "sha256:0007840186bacfaa0aba4466d5890334ea5938e0bb7e28078a0eb0e63b5b59d5", + "sha256:19554bd8d54cf41139f376753af1a644b63c9ca93f8f72009d50a2080f870f77", + "sha256:1d45d1c74f88b9f41062716c727f78f2a59a5476ecbe74956fafb423c5c87a76", + "sha256:1d819553730d3c2724582124aee8a03c846ec4362ded1034c16fb3ef309264e6", + "sha256:2210f28778c7d2ee13f3c2a20a3a22db889e75f4ec13a21072eabb5693801e84", + "sha256:22521219ca739654a296eea6d4367703558fba16f98688bd8ce65abff36eaa84", + "sha256:25405f88a37c5f5bcba01c6e350086d65e7465fd1caaf986333d2a045045a223", + "sha256:2b65bd35f3e06a47b5c30ea99e0c2b88f72c6476eedaf8cfbc8e66adb5479dcf", + "sha256:2ddb500a2808c100e72c075cbb00bf32e62763c82b6a882d403f01a119e3f402", + "sha256:2f8f6c8f4f1cff93ca5058d6ec5f0efda922ecb3f4c5fb76181f327decff98b8", + "sha256:30fa008c172355c7768159983a7270cb23838c4d7db73d6c0f6b60dde0d432c6", + "sha256:3dbb3cea20b4af4f49f84cffaf45dd5f88e8594d18568e0225e6ad9dec0e7967", + "sha256:4116ba9a58109ed5e4cb315bdcbff9838f3159d099ba5259c7c7fb77f8537492", + "sha256:44e6adf67577dbdfa2d9f06db9fbc5639afefdb5bf2b4dfec25c3a7fbc619536", + "sha256:5326ddfacbe51abf9469fe668944bc2e399181a2158cb5d45e1d40856b2a0589", + "sha256:70adc3658138bc77a36ce769f5f183169bc0a2906a4f61f09673f7181255ac9b", + "sha256:72be6ebb4e92520b9726d7146bc9c9b277513a57a38efcf66db0620aec0097e0", + "sha256:7843b1624d6ccca403a610d1277f7c28ad184c5aa88a1750c1a999754e65b439", + "sha256:7ba5a1041480c6e0a8b11a9544d53562abc2d19220bfa14133e0cdd9967e97af", + "sha256:80efd202108c3a4150e042b269f7c78643420cc232a0a771743bb96b742f838f", + "sha256:82f49c5a79d3839bc8f38cb5f4bfc87e15f04cbafa5fbd12fb32c941cb529cfb", + "sha256:83d2c9db5dfc537d0171e32de160461230eb14663299b7e6d18ca6dca21e4977", + "sha256:8d93a1095f83e908fc253f2fb569c2711414c0bfd451cab580466465b235b470", + "sha256:8dc3d842fa41a33fe83d9f5c66c0cc1f28756530cd89944b63b072281e852031", + "sha256:9661a04ca3c950a8ac8c47f53cbc0b530bce1b52f516a1e87b7736fec24bfff0", + "sha256:a498bcd005e8a3fedd0022bb30ee0ad92728154a8798b703f394484452550507", + "sha256:a7a4cf5bbdc861987a7745aed7a536c6405256853c94abc9f3287c3fa401b174", + "sha256:b5074fb09429f2b7bc82b6fb4be8645dcbac14e592128beeff5461dcde0af09f", + "sha256:b6a5431940f28b6de123de42f0eb47b84a073ee3c3345dc109ad550a3307dd28", + "sha256:ba677bcaff9429fd1bf01648ad0901cea56c0d068df383d5f5856d88221fe75b", + "sha256:bcadb05c3d4794eb9eee1dddf1c24215c92fb7b55a80beae7a60530a91060560", + "sha256:bf7eb45d14fc036514c09554bf983f2a72323254912ed0c3c8e697b62c4c158f", + "sha256:c358721aebd40c243894298f685a19eb0491a5c3e0b923b9f887ef1193ddf829", + "sha256:c4550a359c5157aaf8507e6820d98682872b9100ce7607f8aa070b4b8af6c298", + "sha256:c6572c2dab23c86a14e82c245473d45b4c515314f1f859e92608dcafbd2f19b8", + "sha256:cba430db673c29376135e695c6e2501c44c256a81495da849e85d1793ee975ad", + "sha256:dedc71c8eb9c5096037766390172c34fb86ef048b8e8958b4e484b9e505d66bc", + "sha256:e6f5eb2f53fac7d408a45fbcdeda7224b1cfff64919d0f95473420a931347ae9", + "sha256:ec2eba188c1906b05b9b49ae55aae4efd8150c61ba450e6721f64620c50b59eb", + "sha256:ee040a7de8d295dbd261ef2d6d3192f13e2b08ec4a954de34a6fb8ff6422e24c", + "sha256:eedd3b59190885d1ebdf6c5e0ca56828beb1949b4dfe6e5d0256a461429ac386", + "sha256:f441422bb313ab25de7b3dbfd388e790eceb76ce01a18199ec4944b369017009", + "sha256:f8eb7b6716f5b50e9c06207a14172cf2de201e41912ebe732846c02c830455b9", + "sha256:fc4453705b81d03568d5b808ad8f09c77c47534f6ac2e72e733f9ca4714aa75c" ], - "markers": "python_version >= '3.6'", - "version": "==1.3.1" + "markers": "python_version >= '3.7'", + "version": "==1.3.2" }, "markupsafe": { "hashes": [ "sha256:01a9b8ea66f1658938f65b93a85ebe8bc016e6769611be228d797c9d998dd298", "sha256:023cb26ec21ece8dc3907c0e8320058b2e0cb3c55cf9564da612bc325bed5e64", "sha256:0446679737af14f45767963a1a9ef7620189912317d095f2d9ffa183a4d25d2b", + "sha256:04635854b943835a6ea959e948d19dcd311762c5c0c6e1f0e16ee57022669194", "sha256:0717a7390a68be14b8c793ba258e075c6f4ca819f15edfc2a3a027c823718567", "sha256:0955295dd5eec6cb6cc2fe1698f4c6d84af2e92de33fbcac4111913cd100a6ff", + "sha256:0d4b31cc67ab36e3392bbf3862cfbadac3db12bdd8b02a2731f509ed5b829724", "sha256:10f82115e21dc0dfec9ab5c0223652f7197feb168c940f3ef61563fc2d6beb74", + "sha256:168cd0a3642de83558a5153c8bd34f175a9a6e7f6dc6384b9655d2697312a646", "sha256:1d609f577dc6e1aa17d746f8bd3c31aa4d258f4070d61b2aa5c4166c1539de35", + "sha256:1f2ade76b9903f39aa442b4aadd2177decb66525062db244b35d71d0ee8599b6", + "sha256:20dca64a3ef2d6e4d5d615a3fd418ad3bde77a47ec8a23d984a12b5b4c74491a", + "sha256:2a7d351cbd8cfeb19ca00de495e224dea7e7d919659c2841bbb7f420ad03e2d6", + "sha256:2d7d807855b419fc2ed3e631034685db6079889a1f01d5d9dac950f764da3dad", "sha256:2ef54abee730b502252bcdf31b10dacb0a416229b72c18b19e24a4509f273d26", + "sha256:36bc903cbb393720fad60fc28c10de6acf10dc6cc883f3e24ee4012371399a38", + "sha256:37205cac2a79194e3750b0af2a5720d95f786a55ce7df90c3af697bfa100eaac", "sha256:3c112550557578c26af18a1ccc9e090bfe03832ae994343cfdacd287db6a6ae7", + "sha256:3dd007d54ee88b46be476e293f48c85048603f5f516008bee124ddd891398ed6", + "sha256:4296f2b1ce8c86a6aea78613c34bb1a672ea0e3de9c6ba08a960efe0b0a09047", "sha256:47ab1e7b91c098ab893b828deafa1203de86d0bc6ab587b160f78fe6c4011f75", "sha256:49e3ceeabbfb9d66c3aef5af3a60cc43b85c33df25ce03d0031a608b0a8b2e3f", + "sha256:4dc8f9fb58f7364b63fd9f85013b780ef83c11857ae79f2feda41e270468dd9b", "sha256:4efca8f86c54b22348a5467704e3fec767b2db12fc39c6d963168ab1d3fc9135", "sha256:53edb4da6925ad13c07b6d26c2a852bd81e364f95301c66e930ab2aef5b5ddd8", + "sha256:5855f8438a7d1d458206a2466bf82b0f104a3724bf96a1c781ab731e4201731a", "sha256:594c67807fb16238b30c44bdf74f36c02cdf22d1c8cda91ef8a0ed8dabf5620a", + "sha256:5b6d930f030f8ed98e3e6c98ffa0652bdb82601e7a016ec2ab5d7ff23baa78d1", + "sha256:5bb28c636d87e840583ee3adeb78172efc47c8b26127267f54a9c0ec251d41a9", + "sha256:60bf42e36abfaf9aff1f50f52644b336d4f0a3fd6d8a60ca0d054ac9f713a864", "sha256:611d1ad9a4288cf3e3c16014564df047fe08410e628f89805e475368bd304914", + "sha256:6300b8454aa6930a24b9618fbb54b5a68135092bc666f7b06901f897fa5c2fee", + "sha256:63f3268ba69ace99cab4e3e3b5840b03340efed0948ab8f78d2fd87ee5442a4f", "sha256:6557b31b5e2c9ddf0de32a691f2312a32f77cd7681d8af66c2692efdbef84c18", "sha256:693ce3f9e70a6cf7d2fb9e6c9d8b204b6b39897a2c4a1aa65728d5ac97dcc1d8", "sha256:6a7fae0dd14cf60ad5ff42baa2e95727c3d81ded453457771d02b7d2b3f9c0c2", "sha256:6c4ca60fa24e85fe25b912b01e62cb969d69a23a5d5867682dd3e80b5b02581d", + "sha256:6fcf051089389abe060c9cd7caa212c707e58153afa2c649f00346ce6d260f1b", "sha256:7d91275b0245b1da4d4cfa07e0faedd5b0812efc15b702576d103293e252af1b", + "sha256:89c687013cb1cd489a0f0ac24febe8c7a666e6e221b783e53ac50ebf68e45d86", + "sha256:8d206346619592c6200148b01a2142798c989edcb9c896f9ac9722a99d4e77e6", "sha256:905fec760bd2fa1388bb5b489ee8ee5f7291d692638ea5f67982d968366bef9f", "sha256:97383d78eb34da7e1fa37dd273c20ad4320929af65d156e35a5e2d89566d9dfb", "sha256:984d76483eb32f1bcb536dc27e4ad56bba4baa70be32fa87152832cdd9db0833", + "sha256:99df47edb6bda1249d3e80fdabb1dab8c08ef3975f69aed437cb69d0a5de1e28", + "sha256:9f02365d4e99430a12647f09b6cc8bab61a6564363f313126f775eb4f6ef798e", "sha256:a30e67a65b53ea0a5e62fe23682cfe22712e01f453b95233b25502f7c61cb415", "sha256:ab3ef638ace319fa26553db0624c4699e31a28bb2a835c5faca8f8acf6a5a902", + "sha256:aca6377c0cb8a8253e493c6b451565ac77e98c2951c45f913e0b52facdcff83f", + "sha256:add36cb2dbb8b736611303cd3bfcee00afd96471b09cda130da3581cbdc56a6d", "sha256:b2f4bf27480f5e5e8ce285a8c8fd176c0b03e93dcc6646477d4630e83440c6a9", "sha256:b7f2d075102dc8c794cbde1947378051c4e5180d52d276987b8d28a3bd58c17d", + "sha256:baa1a4e8f868845af802979fcdbf0bb11f94f1cb7ced4c4b8a351bb60d108145", "sha256:be98f628055368795d818ebf93da628541e10b75b41c559fdf36d104c5787066", + "sha256:bf5d821ffabf0ef3533c39c518f3357b171a1651c1ff6827325e4489b0e46c3c", + "sha256:c47adbc92fc1bb2b3274c4b3a43ae0e4573d9fbff4f54cd484555edbf030baf1", + "sha256:cdfba22ea2f0029c9261a4bd07e830a8da012291fbe44dc794e488b6c9bb353a", + "sha256:d6c7ebd4e944c85e2c3421e612a7057a2f48d478d79e61800d81468a8d842207", "sha256:d7f9850398e85aba693bb640262d3611788b1f29a79f0c93c565694658f4071f", + "sha256:d8446c54dc28c01e5a2dbac5a25f071f6653e6e40f3a8818e8b45d790fe6ef53", + "sha256:deb993cacb280823246a026e3b2d81c493c53de6acfd5e6bfe31ab3402bb37dd", + "sha256:e0f138900af21926a02425cf736db95be9f4af72ba1bb21453432a07f6082134", + "sha256:e9936f0b261d4df76ad22f8fee3ae83b60d7c3e871292cd42f40b81b70afae85", + "sha256:f0567c4dc99f264f49fe27da5f735f414c4e7e7dd850cfd8e69f0862d7c74ea9", "sha256:f5653a225f31e113b152e56f154ccbe59eeb1c7487b39b9d9f9cdb58e6c79dc5", "sha256:f826e31d18b516f653fe296d967d700fddad5901ae07c622bb3705955e1faa94", "sha256:f8ba0e8349a38d3001fae7eadded3f6606f0da5d748ee53cc1dab1d6527b9509", @@ -372,121 +430,154 @@ }, "matplotlib": { "hashes": [ - "sha256:0bea5ec5c28d49020e5d7923c2725b837e60bc8be99d3164af410eb4b4c827da", - "sha256:1c1779f7ab7d8bdb7d4c605e6ffaa0614b3e80f1e3c8ccf7b9269a22dbc5986b", - "sha256:21b31057bbc5e75b08e70a43cefc4c0b2c2f1b1a850f4a0f7af044eb4163086c", - "sha256:32fa638cc10886885d1ca3d409d4473d6a22f7ceecd11322150961a70fab66dd", - "sha256:3a5c18dbd2c7c366da26a4ad1462fe3e03a577b39e3b503bbcf482b9cdac093c", - "sha256:5826f56055b9b1c80fef82e326097e34dc4af8c7249226b7dd63095a686177d1", - "sha256:6382bc6e2d7e481bcd977eb131c31dee96e0fb4f9177d15ec6fb976d3b9ace1a", - "sha256:6475d0209024a77f869163ec3657c47fed35d9b6ed8bccba8aa0f0099fbbdaa8", - "sha256:6a6a44f27aabe720ec4fd485061e8a35784c2b9ffa6363ad546316dfc9cea04e", - "sha256:7a58f3d8fe8fac3be522c79d921c9b86e090a59637cb88e3bc51298d7a2c862a", - "sha256:7ad19f3fb6145b9eb41c08e7cbb9f8e10b91291396bee21e9ce761bb78df63ec", - "sha256:85f191bb03cb1a7b04b5c2cca4792bef94df06ef473bc49e2818105671766fee", - "sha256:956c8849b134b4a343598305a3ca1bdd3094f01f5efc8afccdebeffe6b315247", - "sha256:a9d8cb5329df13e0cdaa14b3b43f47b5e593ec637f13f14db75bb16e46178b05", - "sha256:b1d5a2cedf5de05567c441b3a8c2651fbde56df08b82640e7f06c8cd91e201f6", - "sha256:b26535b9de85326e6958cdef720ecd10bcf74a3f4371bf9a7e5b2e659c17e153", - "sha256:c541ee5a3287efe066bbe358320853cf4916bc14c00c38f8f3d8d75275a405a9", - "sha256:d8d994cefdff9aaba45166eb3de4f5211adb4accac85cbf97137e98f26ea0219", - "sha256:df815378a754a7edd4559f8c51fc7064f779a74013644a7f5ac7a0c31f875866" + "sha256:01c9de93a2ca0d128c9064f23709362e7fefb34910c7c9e0b8ab0de8258d5eda", + "sha256:41b6e307458988891fcdea2d8ecf84a8c92d53f84190aa32da65f9505546e684", + "sha256:48e1e0859b54d5f2e29bb78ca179fd59b971c6ceb29977fb52735bfd280eb0f5", + "sha256:54a026055d5f8614f184e588f6e29064019a0aa8448450214c0b60926d62d919", + "sha256:556965514b259204637c360d213de28d43a1f4aed1eca15596ce83f768c5a56f", + "sha256:5c988bb43414c7c2b0a31bd5187b4d27fd625c080371b463a6d422047df78913", + "sha256:6a724e3a48a54b8b6e7c4ae38cd3d07084508fa47c410c8757e9db9791421838", + "sha256:6be8df61b1626e1a142c57e065405e869e9429b4a6dab4a324757d0dc4d42235", + "sha256:844a7b0233e4ff7fba57e90b8799edaa40b9e31e300b8d5efc350937fa8b1bea", + "sha256:85f0c9cf724715e75243a7b3087cf4a3de056b55e05d4d76cc58d610d62894f3", + "sha256:a78a3b51f29448c7f4d4575e561f6b0dbb8d01c13c2046ab6c5220eb25c06506", + "sha256:b884715a59fec9ad3b6048ecf3860f3b2ce965e676ef52593d6fa29abcf7d330", + "sha256:b8b53f336a4688cfce615887505d7e41fd79b3594bf21dd300531a4f5b4f746a", + "sha256:c70b6311dda3e27672f1bf48851a0de816d1ca6aaf3d49365fbdd8e959b33d2b", + "sha256:ebfb01a65c3f5d53a8c2a8133fec2b5221281c053d944ae81ff5822a68266617", + "sha256:eeb1859efe7754b1460e1d4991bbd4a60a56f366bc422ef3a9c5ae05f0bc70b5", + "sha256:f15edcb0629a0801738925fe27070480f446fcaa15de65946ff946ad99a59a40", + "sha256:f1c5efc278d996af8a251b2ce0b07bbeccb821f25c8c9846bdcb00ffc7f158aa", + "sha256:f72657f1596199dc1e4e7a10f52a4784ead8a711f4e5b59bea95bdb97cf0e4fd", + "sha256:fc4f526dfdb31c9bd6b8ca06bf9fab663ca12f3ec9cdf4496fb44bc680140318", + "sha256:fcd6f1954943c0c192bfbebbac263f839d7055409f1173f80d8b11a224d236da" ], "index": "pypi", - "version": "==3.4.2" + "version": "==3.4.3" }, - "numpy": { + "mypy-extensions": { "hashes": [ - "sha256:1676b0a292dd3c99e49305a16d7a9f42a4ab60ec522eac0d3dd20cdf362ac010", - "sha256:16f221035e8bd19b9dc9a57159e38d2dd060b48e93e1d843c49cb370b0f415fd", - "sha256:43909c8bb289c382170e0282158a38cf306a8ad2ff6dfadc447e90f9961bef43", - "sha256:4e465afc3b96dbc80cf4a5273e5e2b1e3451286361b4af70ce1adb2984d392f9", - "sha256:55b745fca0a5ab738647d0e4db099bd0a23279c32b31a783ad2ccea729e632df", - "sha256:5d050e1e4bc9ddb8656d7b4f414557720ddcca23a5b88dd7cff65e847864c400", - "sha256:637d827248f447e63585ca3f4a7d2dfaa882e094df6cfa177cc9cf9cd6cdf6d2", - "sha256:6690080810f77485667bfbff4f69d717c3be25e5b11bb2073e76bb3f578d99b4", - "sha256:66fbc6fed94a13b9801fb70b96ff30605ab0a123e775a5e7a26938b717c5d71a", - "sha256:67d44acb72c31a97a3d5d33d103ab06d8ac20770e1c5ad81bdb3f0c086a56cf6", - "sha256:6ca2b85a5997dabc38301a22ee43c82adcb53ff660b89ee88dded6b33687e1d8", - "sha256:6e51534e78d14b4a009a062641f465cfaba4fdcb046c3ac0b1f61dd97c861b1b", - "sha256:70eb5808127284c4e5c9e836208e09d685a7978b6a216db85960b1a112eeace8", - "sha256:830b044f4e64a76ba71448fce6e604c0fc47a0e54d8f6467be23749ac2cbd2fb", - "sha256:8b7bb4b9280da3b2856cb1fc425932f46fba609819ee1c62256f61799e6a51d2", - "sha256:a9c65473ebc342715cb2d7926ff1e202c26376c0dcaaee85a1fd4b8d8c1d3b2f", - "sha256:c1c09247ccea742525bdb5f4b5ceeacb34f95731647fe55774aa36557dbb5fa4", - "sha256:c5bf0e132acf7557fc9bb8ded8b53bbbbea8892f3c9a1738205878ca9434206a", - "sha256:db250fd3e90117e0312b611574cd1b3f78bec046783195075cbd7ba9c3d73f16", - "sha256:e515c9a93aebe27166ec9593411c58494fa98e5fcc219e47260d9ab8a1cc7f9f", - "sha256:e55185e51b18d788e49fe8305fd73ef4470596b33fc2c1ceb304566b99c71a69", - "sha256:ea9cff01e75a956dbee133fa8e5b68f2f92175233de2f88de3a682dd94deda65", - "sha256:f1452578d0516283c87608a5a5548b0cdde15b99650efdfd85182102ef7a7c17", - "sha256:f39a995e47cb8649673cfa0579fbdd1cdd33ea497d1728a6cb194d6252268e48" + "sha256:090fedd75945a69ae91ce1303b5824f428daf5a028d2f6ab8a299250a846f15d", + "sha256:2d82818f5bb3e369420cb3c4060a7970edba416647068eb4c5343488a6c604a8" ], - "markers": "python_version >= '3.7'", - "version": "==1.20.3" + "version": "==0.4.3" + }, + "numpy": { + "hashes": [ + "sha256:043e83bfc274649c82a6f09836943e4a4aebe5e33656271c7dbf9621dd58b8ec", + "sha256:160ccc1bed3a8371bf0d760971f09bfe80a3e18646620e9ded0ad159d9749baa", + "sha256:188031f833bbb623637e66006cf75e933e00e7231f67e2b45cf8189612bb5dc3", + "sha256:28f15209fb535dd4c504a7762d3bc440779b0e37d50ed810ced209e5cea60d96", + "sha256:29fb3dcd0468b7715f8ce2c0c2d9bbbaf5ae686334951343a41bd8d155c6ea27", + "sha256:2a6ee9620061b2a722749b391c0d80a0e2ae97290f1b32e28d5a362e21941ee4", + "sha256:300321e3985c968e3ae7fbda187237b225f3ffe6528395a5b7a5407f73cf093e", + "sha256:32437f0b275c1d09d9c3add782516413e98cd7c09e6baf4715cbce781fc29912", + "sha256:3c09418a14471c7ae69ba682e2428cae5b4420a766659605566c0fa6987f6b7e", + "sha256:49c6249260890e05b8111ebfc391ed58b3cb4b33e63197b2ec7f776e45330721", + "sha256:4cc9b512e9fb590797474f58b7f6d1f1b654b3a94f4fa8558b48ca8b3cfc97cf", + "sha256:508b0b513fa1266875524ba8a9ecc27b02ad771fe1704a16314dc1a816a68737", + "sha256:50cd26b0cf6664cb3b3dd161ba0a09c9c1343db064e7c69f9f8b551f5104d654", + "sha256:5c4193f70f8069550a1788bd0cd3268ab7d3a2b70583dfe3b2e7f421e9aace06", + "sha256:5dfe9d6a4c39b8b6edd7990091fea4f852888e41919d0e6722fe78dd421db0eb", + "sha256:63571bb7897a584ca3249c86dd01c10bcb5fe4296e3568b2e9c1a55356b6410e", + "sha256:75621882d2230ab77fb6a03d4cbccd2038511491076e7964ef87306623aa5272", + "sha256:75eb7cadc8da49302f5b659d40ba4f6d94d5045fbd9569c9d058e77b0514c9e4", + "sha256:88a5d6b268e9ad18f3533e184744acdaa2e913b13148160b1152300c949bbb5f", + "sha256:8a10968963640e75cc0193e1847616ab4c718e83b6938ae74dea44953950f6b7", + "sha256:90bec6a86b348b4559b6482e2b684db4a9a7eed1fa054b86115a48d58fbbf62a", + "sha256:98339aa9911853f131de11010f6dd94c8cec254d3d1f7261528c3b3e3219f139", + "sha256:a99a6b067e5190ac6d12005a4d85aa6227c5606fa93211f86b1dafb16233e57d", + "sha256:bffa2eee3b87376cc6b31eee36d05349571c236d1de1175b804b348dc0941e3f", + "sha256:c6c2d535a7beb1f8790aaa98fd089ceab2e3dd7ca48aca0af7dc60e6ef93ffe1", + "sha256:cc14e7519fab2a4ed87d31f99c31a3796e4e1fe63a86ebdd1c5a1ea78ebd5896", + "sha256:dd0482f3fc547f1b1b5d6a8b8e08f63fdc250c58ce688dedd8851e6e26cff0f3", + "sha256:dde972a1e11bb7b702ed0e447953e7617723760f420decb97305e66fb4afc54f", + "sha256:e54af82d68ef8255535a6cdb353f55d6b8cf418a83e2be3569243787a4f4866f", + "sha256:e606e6316911471c8d9b4618e082635cfe98876007556e89ce03d52ff5e8fcf0", + "sha256:f41b018f126aac18583956c54544db437f25c7ee4794bcb23eb38bef8e5e192a", + "sha256:f8f4625536926a155b80ad2bbff44f8cc59e9f2ad14cdda7acf4c135b4dc8ff2", + "sha256:fe52dbe47d9deb69b05084abd4b0df7abb39a3c51957c09f635520abd49b29dd" + ], + "markers": "python_version < '3.10' and platform_machine != 'aarch64' and platform_machine != 'arm64'", + "version": "==1.21.3" }, "packaging": { "hashes": [ - "sha256:5b327ac1320dc863dca72f4514ecc086f31186744b84a230374cc1fd776feae5", - "sha256:67714da7f7bc052e064859c05c595155bd1ee9f69f76557e21f051443c20947a" + "sha256:096d689d78ca690e4cd8a89568ba06d07ca097e3306a4381635073ca91479966", + "sha256:14317396d1e8cdb122989b916fa2c7e9ca8e2be9e8060a6eff75b6b7b4d8a7e0" ], - "markers": "python_version >= '2.7' and python_version not in '3.0, 3.1, 3.2, 3.3'", - "version": "==20.9" + "markers": "python_version >= '3.6'", + "version": "==21.2" }, "pathspec": { "hashes": [ - "sha256:86379d6b86d75816baba717e64b1a3a3469deb93bb76d613c9ce79edc5cb68fd", - "sha256:aa0cb481c4041bf52ffa7b0d8fa6cd3e88a2ca4879c533c9153882ee2556790d" + "sha256:7d15c4ddb0b5c802d161efc417ec1a2558ea2653c2e8ad9c19098201dc1c993a", + "sha256:e564499435a2673d586f6b2130bb5b95f04a3ba06f81b8f895b651a3c76aabb1" ], - "version": "==0.8.1" + "version": "==0.9.0" }, "pillow": { "hashes": [ - "sha256:01425106e4e8cee195a411f729cff2a7d61813b0b11737c12bd5991f5f14bcd5", - "sha256:031a6c88c77d08aab84fecc05c3cde8414cd6f8406f4d2b16fed1e97634cc8a4", - "sha256:083781abd261bdabf090ad07bb69f8f5599943ddb539d64497ed021b2a67e5a9", - "sha256:0d19d70ee7c2ba97631bae1e7d4725cdb2ecf238178096e8c82ee481e189168a", - "sha256:0e04d61f0064b545b989126197930807c86bcbd4534d39168f4aa5fda39bb8f9", - "sha256:12e5e7471f9b637762453da74e390e56cc43e486a88289995c1f4c1dc0bfe727", - "sha256:22fd0f42ad15dfdde6c581347eaa4adb9a6fc4b865f90b23378aa7914895e120", - "sha256:238c197fc275b475e87c1453b05b467d2d02c2915fdfdd4af126145ff2e4610c", - "sha256:3b570f84a6161cf8865c4e08adf629441f56e32f180f7aa4ccbd2e0a5a02cba2", - "sha256:463822e2f0d81459e113372a168f2ff59723e78528f91f0bd25680ac185cf797", - "sha256:4d98abdd6b1e3bf1a1cbb14c3895226816e666749ac040c4e2554231068c639b", - "sha256:5afe6b237a0b81bd54b53f835a153770802f164c5570bab5e005aad693dab87f", - "sha256:5b70110acb39f3aff6b74cf09bb4169b167e2660dabc304c1e25b6555fa781ef", - "sha256:5cbf3e3b1014dddc45496e8cf38b9f099c95a326275885199f427825c6522232", - "sha256:624b977355cde8b065f6d51b98497d6cd5fbdd4f36405f7a8790e3376125e2bb", - "sha256:63728564c1410d99e6d1ae8e3b810fe012bc440952168af0a2877e8ff5ab96b9", - "sha256:66cc56579fd91f517290ab02c51e3a80f581aba45fd924fcdee01fa06e635812", - "sha256:6c32cc3145928c4305d142ebec682419a6c0a8ce9e33db900027ddca1ec39178", - "sha256:8b56553c0345ad6dcb2e9b433ae47d67f95fc23fe28a0bde15a120f25257e291", - "sha256:8bb1e155a74e1bfbacd84555ea62fa21c58e0b4e7e6b20e4447b8d07990ac78b", - "sha256:95d5ef984eff897850f3a83883363da64aae1000e79cb3c321915468e8c6add5", - "sha256:a013cbe25d20c2e0c4e85a9daf438f85121a4d0344ddc76e33fd7e3965d9af4b", - "sha256:a787ab10d7bb5494e5f76536ac460741788f1fbce851068d73a87ca7c35fc3e1", - "sha256:a7d5e9fad90eff8f6f6106d3b98b553a88b6f976e51fce287192a5d2d5363713", - "sha256:aac00e4bc94d1b7813fe882c28990c1bc2f9d0e1aa765a5f2b516e8a6a16a9e4", - "sha256:b91c36492a4bbb1ee855b7d16fe51379e5f96b85692dc8210831fbb24c43e484", - "sha256:c03c07ed32c5324939b19e36ae5f75c660c81461e312a41aea30acdd46f93a7c", - "sha256:c5236606e8570542ed424849f7852a0ff0bce2c4c8d0ba05cc202a5a9c97dee9", - "sha256:c6b39294464b03457f9064e98c124e09008b35a62e3189d3513e5148611c9388", - "sha256:cb7a09e173903541fa888ba010c345893cd9fc1b5891aaf060f6ca77b6a3722d", - "sha256:d68cb92c408261f806b15923834203f024110a2e2872ecb0bd2a110f89d3c602", - "sha256:dc38f57d8f20f06dd7c3161c59ca2c86893632623f33a42d592f097b00f720a9", - "sha256:e98eca29a05913e82177b3ba3d198b1728e164869c613d76d0de4bde6768a50e", - "sha256:f217c3954ce5fd88303fc0c317af55d5e0204106d86dea17eb8205700d47dec2" + "sha256:066f3999cb3b070a95c3652712cffa1a748cd02d60ad7b4e485c3748a04d9d76", + "sha256:0a0956fdc5defc34462bb1c765ee88d933239f9a94bc37d132004775241a7585", + "sha256:0b052a619a8bfcf26bd8b3f48f45283f9e977890263e4571f2393ed8898d331b", + "sha256:1394a6ad5abc838c5cd8a92c5a07535648cdf6d09e8e2d6df916dfa9ea86ead8", + "sha256:1bc723b434fbc4ab50bb68e11e93ce5fb69866ad621e3c2c9bdb0cd70e345f55", + "sha256:244cf3b97802c34c41905d22810846802a3329ddcb93ccc432870243211c79fc", + "sha256:25a49dc2e2f74e65efaa32b153527fc5ac98508d502fa46e74fa4fd678ed6645", + "sha256:2e4440b8f00f504ee4b53fe30f4e381aae30b0568193be305256b1462216feff", + "sha256:3862b7256046fcd950618ed22d1d60b842e3a40a48236a5498746f21189afbbc", + "sha256:3eb1ce5f65908556c2d8685a8f0a6e989d887ec4057326f6c22b24e8a172c66b", + "sha256:3f97cfb1e5a392d75dd8b9fd274d205404729923840ca94ca45a0af57e13dbe6", + "sha256:493cb4e415f44cd601fcec11c99836f707bb714ab03f5ed46ac25713baf0ff20", + "sha256:4acc0985ddf39d1bc969a9220b51d94ed51695d455c228d8ac29fcdb25810e6e", + "sha256:5503c86916d27c2e101b7f71c2ae2cddba01a2cf55b8395b0255fd33fa4d1f1a", + "sha256:5b7bb9de00197fb4261825c15551adf7605cf14a80badf1761d61e59da347779", + "sha256:5e9ac5f66616b87d4da618a20ab0a38324dbe88d8a39b55be8964eb520021e02", + "sha256:620582db2a85b2df5f8a82ddeb52116560d7e5e6b055095f04ad828d1b0baa39", + "sha256:62cc1afda735a8d109007164714e73771b499768b9bb5afcbbee9d0ff374b43f", + "sha256:70ad9e5c6cb9b8487280a02c0ad8a51581dcbbe8484ce058477692a27c151c0a", + "sha256:72b9e656e340447f827885b8d7a15fc8c4e68d410dc2297ef6787eec0f0ea409", + "sha256:72cbcfd54df6caf85cc35264c77ede902452d6df41166010262374155947460c", + "sha256:792e5c12376594bfcb986ebf3855aa4b7c225754e9a9521298e460e92fb4a488", + "sha256:7b7017b61bbcdd7f6363aeceb881e23c46583739cb69a3ab39cb384f6ec82e5b", + "sha256:81f8d5c81e483a9442d72d182e1fb6dcb9723f289a57e8030811bac9ea3fef8d", + "sha256:82aafa8d5eb68c8463b6e9baeb4f19043bb31fefc03eb7b216b51e6a9981ae09", + "sha256:84c471a734240653a0ec91dec0996696eea227eafe72a33bd06c92697728046b", + "sha256:8c803ac3c28bbc53763e6825746f05cc407b20e4a69d0122e526a582e3b5e153", + "sha256:93ce9e955cc95959df98505e4608ad98281fff037350d8c2671c9aa86bcf10a9", + "sha256:9a3e5ddc44c14042f0844b8cf7d2cd455f6cc80fd7f5eefbe657292cf601d9ad", + "sha256:a4901622493f88b1a29bd30ec1a2f683782e57c3c16a2dbc7f2595ba01f639df", + "sha256:a5a4532a12314149d8b4e4ad8ff09dde7427731fcfa5917ff16d0291f13609df", + "sha256:b8831cb7332eda5dc89b21a7bce7ef6ad305548820595033a4b03cf3091235ed", + "sha256:b8e2f83c56e141920c39464b852de3719dfbfb6e3c99a2d8da0edf4fb33176ed", + "sha256:c70e94281588ef053ae8998039610dbd71bc509e4acbc77ab59d7d2937b10698", + "sha256:c8a17b5d948f4ceeceb66384727dde11b240736fddeda54ca740b9b8b1556b29", + "sha256:d82cdb63100ef5eedb8391732375e6d05993b765f72cb34311fab92103314649", + "sha256:d89363f02658e253dbd171f7c3716a5d340a24ee82d38aab9183f7fdf0cdca49", + "sha256:d99ec152570e4196772e7a8e4ba5320d2d27bf22fdf11743dd882936ed64305b", + "sha256:ddc4d832a0f0b4c52fff973a0d44b6c99839a9d016fe4e6a1cb8f3eea96479c2", + "sha256:e3dacecfbeec9a33e932f00c6cd7996e62f53ad46fbe677577394aaa90ee419a", + "sha256:eb9fc393f3c61f9054e1ed26e6fe912c7321af2f41ff49d3f83d05bacf22cc78" + ], + "markers": "python_version >= '3.6'", + "version": "==8.4.0" + }, + "platformdirs": { + "hashes": [ + "sha256:367a5e80b3d04d2428ffa76d33f124cf11e8fff2acdaa9b43d545f5c7d661ef2", + "sha256:8868bbe3c3c80d42f20156f22e7131d2fb321f5bc86a2a345375c6481a67021d" ], "markers": "python_version >= '3.6'", - "version": "==8.2.0" + "version": "==2.4.0" }, "pluggy": { "hashes": [ - "sha256:15b2acde666561e1298d71b523007ed7364de07029219b604cf808bfa1c765b0", - "sha256:966c145cd83c96502c3c3868f50408687b38434af77734af1e9ca461a4081d2d" + "sha256:4224373bacce55f955a878bf9cfa763c1e360858e330072059e10bad68531159", + "sha256:74134bbf457f031a36d68416e1509f34bd5ccc019f0bcc952c7b909d06b37bd3" ], - "markers": "python_version >= '2.7' and python_version not in '3.0, 3.1, 3.2, 3.3'", - "version": "==0.13.1" + "markers": "python_version >= '3.6'", + "version": "==1.0.0" }, "py": { "hashes": [ @@ -498,11 +589,11 @@ }, "pygments": { "hashes": [ - "sha256:a18f47b506a429f6f4b9df81bb02beab9ca21d0a5fee38ed15aef65f0545519f", - "sha256:d66e804411278594d764fc69ec36ec13d9ae9147193a1740cd34d272ca383b8e" + "sha256:b8e67fe6af78f492b3c4b3e2970c0624cbf08beb1e493b2c99b9fa1b67a20380", + "sha256:f398865f7eb6874156579fdf36bc840a03cab64d1cde9e93d68f46a425ec52c6" ], "markers": "python_version >= '3.5'", - "version": "==2.9.0" + "version": "==2.10.0" }, "pyparsing": { "hashes": [ @@ -514,19 +605,19 @@ }, "pytest": { "hashes": [ - "sha256:50bcad0a0b9c5a72c8e4e7c9855a3ad496ca6a881a3641b4260605450772c54b", - "sha256:91ef2131a9bd6be8f76f1f08eac5c5317221d6ad1e143ae03894b862e8976890" + "sha256:131b36680866a76e6781d13f101efb86cf674ebb9762eb70d3082b6f29889e89", + "sha256:7310f8d27bc79ced999e760ca304d69f6ba6c6649c0b60fb0e04a4a77cacc134" ], "index": "pypi", - "version": "==6.2.4" + "version": "==6.2.5" }, "pytest-cov": { "hashes": [ - "sha256:261bb9e47e65bd099c89c3edf92972865210c36813f80ede5277dceb77a4a62a", - "sha256:261ceeb8c227b726249b376b8526b600f38667ee314f910353fa318caa01f4d7" + "sha256:578d5d15ac4a25e5f961c938b85a05b09fdaae9deef3bb6de9a6e766622ca7a6", + "sha256:e7f0f5b1617d2210a2cabc266dfe2f4c75a8d32fb89eafb7ad9d06f6d076d470" ], "index": "pypi", - "version": "==2.12.1" + "version": "==3.0.0" }, "pytest-watch": { "hashes": [ @@ -537,72 +628,67 @@ }, "python-dateutil": { "hashes": [ - "sha256:73ebfe9dbf22e832286dafa60473e4cd239f8592f699aa5adaf10050e6e1823c", - "sha256:75bb3f31ea686f1197762692a9ee6a7550b59fc6ca3a1f4b5d7e32fb98e2da2a" + "sha256:0123cacc1627ae19ddf3c27a5de5bd67ee4586fbdd6440d9748f8abb483d3e86", + "sha256:961d03dc3453ebbc59dbdea9e4e11c5651520a876d0f4db161e8674aae935da9" ], "markers": "python_version >= '2.7' and python_version not in '3.0, 3.1, 3.2, 3.3'", - "version": "==2.8.1" + "version": "==2.8.2" }, "pytz": { "hashes": [ - "sha256:83a4a90894bf38e243cf052c8b58f381bfe9a7a483f6a9cab140bc7f702ac4da", - "sha256:eb10ce3e7736052ed3623d49975ce333bcd712c7bb19a58b9e2089d4057d0798" + "sha256:3672058bc3453457b622aab7a1c3bfd5ab0bdae451512f6cf25f64ed37f5b87c", + "sha256:acad2d8b20a1af07d4e4c9d2e9285c5ed9104354062f275f3fcd88dcef4f1326" ], - "version": "==2021.1" + "version": "==2021.3" }, "regex": { "hashes": [ - "sha256:01afaf2ec48e196ba91b37451aa353cb7eda77efe518e481707e0515025f0cd5", - "sha256:11d773d75fa650cd36f68d7ca936e3c7afaae41b863b8c387a22aaa78d3c5c79", - "sha256:18c071c3eb09c30a264879f0d310d37fe5d3a3111662438889ae2eb6fc570c31", - "sha256:1e1c20e29358165242928c2de1482fb2cf4ea54a6a6dea2bd7a0e0d8ee321500", - "sha256:281d2fd05555079448537fe108d79eb031b403dac622621c78944c235f3fcf11", - "sha256:314d66636c494ed9c148a42731b3834496cc9a2c4251b1661e40936814542b14", - "sha256:32e65442138b7b76dd8173ffa2cf67356b7bc1768851dded39a7a13bf9223da3", - "sha256:339456e7d8c06dd36a22e451d58ef72cef293112b559010db3d054d5560ef439", - "sha256:3916d08be28a1149fb97f7728fca1f7c15d309a9f9682d89d79db75d5e52091c", - "sha256:3a9cd17e6e5c7eb328517969e0cb0c3d31fd329298dd0c04af99ebf42e904f82", - "sha256:47bf5bf60cf04d72bf6055ae5927a0bd9016096bf3d742fa50d9bf9f45aa0711", - "sha256:4c46e22a0933dd783467cf32b3516299fb98cfebd895817d685130cc50cd1093", - "sha256:4c557a7b470908b1712fe27fb1ef20772b78079808c87d20a90d051660b1d69a", - "sha256:52ba3d3f9b942c49d7e4bc105bb28551c44065f139a65062ab7912bef10c9afb", - "sha256:563085e55b0d4fb8f746f6a335893bda5c2cef43b2f0258fe1020ab1dd874df8", - "sha256:598585c9f0af8374c28edd609eb291b5726d7cbce16be6a8b95aa074d252ee17", - "sha256:619d71c59a78b84d7f18891fe914446d07edd48dc8328c8e149cbe0929b4e000", - "sha256:67bdb9702427ceddc6ef3dc382455e90f785af4c13d495f9626861763ee13f9d", - "sha256:6d1b01031dedf2503631d0903cb563743f397ccaf6607a5e3b19a3d76fc10480", - "sha256:741a9647fcf2e45f3a1cf0e24f5e17febf3efe8d4ba1281dcc3aa0459ef424dc", - "sha256:7c2a1af393fcc09e898beba5dd59196edaa3116191cc7257f9224beaed3e1aa0", - "sha256:7d9884d86dd4dd489e981d94a65cd30d6f07203d90e98f6f657f05170f6324c9", - "sha256:90f11ff637fe8798933fb29f5ae1148c978cccb0452005bf4c69e13db951e765", - "sha256:919859aa909429fb5aa9cf8807f6045592c85ef56fdd30a9a3747e513db2536e", - "sha256:96fcd1888ab4d03adfc9303a7b3c0bd78c5412b2bfbe76db5b56d9eae004907a", - "sha256:97f29f57d5b84e73fbaf99ab3e26134e6687348e95ef6b48cfd2c06807005a07", - "sha256:980d7be47c84979d9136328d882f67ec5e50008681d94ecc8afa8a65ed1f4a6f", - "sha256:a91aa8619b23b79bcbeb37abe286f2f408d2f2d6f29a17237afda55bb54e7aac", - "sha256:ade17eb5d643b7fead300a1641e9f45401c98eee23763e9ed66a43f92f20b4a7", - "sha256:b9c3db21af35e3b3c05764461b262d6f05bbca08a71a7849fd79d47ba7bc33ed", - "sha256:bd28bc2e3a772acbb07787c6308e00d9626ff89e3bfcdebe87fa5afbfdedf968", - "sha256:bf5824bfac591ddb2c1f0a5f4ab72da28994548c708d2191e3b87dd207eb3ad7", - "sha256:c0502c0fadef0d23b128605d69b58edb2c681c25d44574fc673b0e52dce71ee2", - "sha256:c38c71df845e2aabb7fb0b920d11a1b5ac8526005e533a8920aea97efb8ec6a4", - "sha256:ce15b6d103daff8e9fee13cf7f0add05245a05d866e73926c358e871221eae87", - "sha256:d3029c340cfbb3ac0a71798100ccc13b97dddf373a4ae56b6a72cf70dfd53bc8", - "sha256:e512d8ef5ad7b898cdb2d8ee1cb09a8339e4f8be706d27eaa180c2f177248a10", - "sha256:e8e5b509d5c2ff12f8418006d5a90e9436766133b564db0abaec92fd27fcee29", - "sha256:ee54ff27bf0afaf4c3b3a62bcd016c12c3fdb4ec4f413391a90bd38bc3624605", - "sha256:fa4537fb4a98fe8fde99626e4681cc644bdcf2a795038533f9f711513a862ae6", - "sha256:fd45ff9293d9274c5008a2054ecef86a9bfe819a67c7be1afb65e69b405b3042" - ], - "version": "==2021.4.4" + "sha256:0c186691a7995ef1db61205e00545bf161fb7b59cdb8c1201c89b333141c438a", + "sha256:0dcc0e71118be8c69252c207630faf13ca5e1b8583d57012aae191e7d6d28b84", + "sha256:0f7552429dd39f70057ac5d0e897e5bfe211629652399a21671e53f2a9693a4e", + "sha256:129472cd06062fb13e7b4670a102951a3e655e9b91634432cfbdb7810af9d710", + "sha256:13ec99df95003f56edcd307db44f06fbeb708c4ccdcf940478067dd62353181e", + "sha256:1f2b59c28afc53973d22e7bc18428721ee8ca6079becf1b36571c42627321c65", + "sha256:2b20f544cbbeffe171911f6ce90388ad36fe3fad26b7c7a35d4762817e9ea69c", + "sha256:2fb698037c35109d3c2e30f2beb499e5ebae6e4bb8ff2e60c50b9a805a716f79", + "sha256:34d870f9f27f2161709054d73646fc9aca49480617a65533fc2b4611c518e455", + "sha256:391703a2abf8013d95bae39145d26b4e21531ab82e22f26cd3a181ee2644c234", + "sha256:450dc27483548214314640c89a0f275dbc557968ed088da40bde7ef8fb52829e", + "sha256:45b65d6a275a478ac2cbd7fdbf7cc93c1982d613de4574b56fd6972ceadb8395", + "sha256:5095a411c8479e715784a0c9236568ae72509450ee2226b649083730f3fadfc6", + "sha256:530fc2bbb3dc1ebb17f70f7b234f90a1dd43b1b489ea38cea7be95fb21cdb5c7", + "sha256:56f0c81c44638dfd0e2367df1a331b4ddf2e771366c4b9c5d9a473de75e3e1c7", + "sha256:5e9c9e0ce92f27cef79e28e877c6b6988c48b16942258f3bc55d39b5f911df4f", + "sha256:6d7722136c6ed75caf84e1788df36397efdc5dbadab95e59c2bba82d4d808a4c", + "sha256:74d071dbe4b53c602edd87a7476ab23015a991374ddb228d941929ad7c8c922e", + "sha256:7b568809dca44cb75c8ebb260844ea98252c8c88396f9d203f5094e50a70355f", + "sha256:80bb5d2e92b2258188e7dcae5b188c7bf868eafdf800ea6edd0fbfc029984a88", + "sha256:8d1cdcda6bd16268316d5db1038965acf948f2a6f43acc2e0b1641ceab443623", + "sha256:9f665677e46c5a4d288ece12fdedf4f4204a422bb28ff05f0e6b08b7447796d1", + "sha256:a30513828180264294953cecd942202dfda64e85195ae36c265daf4052af0464", + "sha256:a7a986c45d1099a5de766a15de7bee3840b1e0e1a344430926af08e5297cf666", + "sha256:a940ca7e7189d23da2bfbb38973832813eab6bd83f3bf89a977668c2f813deae", + "sha256:ab7c5684ff3538b67df3f93d66bd3369b749087871ae3786e70ef39e601345b0", + "sha256:be04739a27be55631069b348dda0c81d8ea9822b5da10b8019b789e42d1fe452", + "sha256:c0938ddd60cc04e8f1faf7a14a166ac939aac703745bfcd8e8f20322a7373019", + "sha256:cb46b542133999580ffb691baf67410306833ee1e4f58ed06b6a7aaf4e046952", + "sha256:d134757a37d8640f3c0abb41f5e68b7cf66c644f54ef1cb0573b7ea1c63e1509", + "sha256:de557502c3bec8e634246588a94e82f1ee1b9dfcfdc453267c4fb652ff531570", + "sha256:ded0c4a3eee56b57fcb2315e40812b173cafe79d2f992d50015f4387445737fa", + "sha256:e1dae12321b31059a1a72aaa0e6ba30156fe7e633355e445451e4021b8e122b6", + "sha256:eb672217f7bd640411cfc69756ce721d00ae600814708d35c930930f18e8029f", + "sha256:ee684f139c91e69fe09b8e83d18b4d63bf87d9440c1eb2eeb52ee851883b1b29", + "sha256:f3f9a91d3cc5e5b0ddf1043c0ae5fa4852f18a1c0050318baf5fc7930ecc1f9c" + ], + "version": "==2021.10.23" }, "requests": { "hashes": [ - "sha256:27973dd4a904a4f13b263a19c866c13b92a39ed1c964655f025f3f8d3d75b804", - "sha256:c210084e36a42ae6b9219e00e48287def368a26d03a048ddad7bfee44f75871e" + "sha256:6c1246513ecd5ecd4528a0906f910e8f0f9c6b8ec72030dc9fd154dc1a6efd24", + "sha256:b8aa58f8cf793ffd8782d3d8cb19e66ef36f7aba4353eec859e74678b01b07a7" ], - "markers": "python_version >= '2.7' and python_version not in '3.0, 3.1, 3.2, 3.3, 3.4'", - "version": "==2.25.1" + "markers": "python_version >= '2.7' and python_version not in '3.0, 3.1, 3.2, 3.3, 3.4, 3.5'", + "version": "==2.26.0" }, "six": { "hashes": [ @@ -621,19 +707,19 @@ }, "sphinx": { "hashes": [ - "sha256:b5c2ae4120bf00c799ba9b3699bc895816d272d120080fbc967292f29b52b48c", - "sha256:d1cb10bee9c4231f1700ec2e24a91be3f3a3aba066ea4ca9f3bbe47e59d5a1d4" + "sha256:94078db9184491e15bce0a56d9186e0aec95f16ac20b12d00e06d4e36f1058a6", + "sha256:98a535c62a4fcfcc362528592f69b26f7caec587d32cd55688db580be0287ae0" ], "index": "pypi", - "version": "==4.0.2" + "version": "==4.2.0" }, "sphinx-rtd-theme": { "hashes": [ - "sha256:32bd3b5d13dc8186d7a42fc816a23d32e83a4827d7d9882948e7b837c232da5a", - "sha256:4a05bdbe8b1446d77a01e20a23ebc6777c74f43237035e76be89699308987d6f" + "sha256:4d35a56f4508cfee4c4fb604373ede6feae2a306731d533f409ef5c3496fdbd8", + "sha256:eec6d497e4c2195fa0e8b2016b337532b8a699a68bcb22a512870e16925c6a5c" ], "index": "pypi", - "version": "==0.5.2" + "version": "==1.0.0" }, "sphinxcontrib-applehelp": { "hashes": [ @@ -691,6 +777,13 @@ "markers": "python_version >= '2.6' and python_version not in '3.0, 3.1, 3.2, 3.3'", "version": "==0.10.2" }, + "tomli": { + "hashes": [ + "sha256:c6ce0015eb38820eaf32b5db832dbc26deb3dd427bd5f6556cf0acac2c214fee", + "sha256:f04066f68f5554911363063a30b108d2b5a5b1a010aa8b6132af78489fe3aade" + ], + "version": "==1.2.2" + }, "typed-ast": { "hashes": [ "sha256:01ae5f73431d21eead5015997ab41afa53aa1fbe252f9da060be5dad2c730ace", @@ -724,55 +817,62 @@ "sha256:f8afcf15cc511ada719a88e013cec87c11aff7b91f019295eb4530f96fe5ef2f", "sha256:fb1bbeac803adea29cedd70781399c99138358c26d05fcbd23c13016b7f5ec65" ], + "markers": "python_version < '3.8'", "version": "==1.4.3" }, "typing-extensions": { "hashes": [ - "sha256:0ac0f89795dd19de6b97debb0c6af1c70987fd80a2d62d1958f7e56fcc31b497", - "sha256:50b6f157849174217d0656f99dc82fe932884fb250826c18350e159ec6cdf342", - "sha256:779383f6086d90c99ae41cf0ff39aac8a7937a9283ce0a414e5dd782f4c94a84" + "sha256:49f75d16ff11f1cd258e1b988ccff82a3ca5570217d7ad8c5f48205dd99a677e", + "sha256:d8226d10bc02a29bcc81df19a26e56a9647f8b0a6d4a83924139f4a8b01f17b7", + "sha256:f1d25edafde516b146ecd0613dabcc61409817af4766fbbcfb8d1ad4ec441a34" ], "markers": "python_version < '3.8'", - "version": "==3.10.0.0" + "version": "==3.10.0.2" }, "urllib3": { "hashes": [ - "sha256:753a0374df26658f99d826cfe40394a686d05985786d946fbe4165b5148f5a7c", - "sha256:a7acd0977125325f516bda9735fa7142b909a8d01e8b2e4c8108d0984e6e0098" + "sha256:4987c65554f7a2dbf30c18fd48778ef124af6fab771a377103da0585e2336ece", + "sha256:c4fdf4019605b6e5423637e01bc9fe4daef873709a7973e195ceba0a62bbc844" ], "markers": "python_version >= '2.7' and python_version not in '3.0, 3.1, 3.2, 3.3, 3.4' and python_version < '4'", - "version": "==1.26.5" + "version": "==1.26.7" }, "watchdog": { "hashes": [ - "sha256:0237db4d9024859bea27d0efb59fe75eef290833fd988b8ead7a879b0308c2db", - "sha256:104266a778906ae0e971368d368a65c4cd032a490a9fca5ba0b78c6c7ae11720", - "sha256:188145185c08c73c56f1478ccf1f0f0f85101191439679b35b6b100886ce0b39", - "sha256:1a62a4671796dc93d1a7262286217d9e75823c63d4c42782912d39a506d30046", - "sha256:255a32d44bbbe62e52874ff755e2eefe271b150e0ec240ad7718a62a7a7a73c4", - "sha256:3d6405681471ebe0beb3aa083998c4870e48b57f8afdb45ea1b5957cc5cf1014", - "sha256:4b219d46d89cfa49af1d73175487c14a318a74cb8c5442603fd13c6a5b418c86", - "sha256:581e3548159fe7d2a9f377a1fbcb41bdcee46849cca8ab803c7ac2e5e04ec77c", - "sha256:58ebb1095ee493008a7789d47dd62e4999505d82be89fc884d473086fccc6ebd", - "sha256:598d772beeaf9c98d0df946fbabf0c8365dd95ea46a250c224c725fe0c4730bc", - "sha256:668391e6c32742d76e5be5db6bf95c455fa4b3d11e76a77c13b39bccb3a47a72", - "sha256:6ef9fe57162c4c361692620e1d9167574ba1975ee468b24051ca11c9bba6438e", - "sha256:91387ee2421f30b75f7ff632c9d48f76648e56bf346a7c805c0a34187a93aab4", - "sha256:a42e6d652f820b2b94cd03156c62559a2ea68d476476dfcd77d931e7f1012d4a", - "sha256:a6471517315a8541a943c00b45f1d252e36898a3ae963d2d52509b89a50cb2b9", - "sha256:d34ce2261f118ecd57eedeef95fc2a495fc4a40b3ed7b3bf0bd7a8ccc1ab4f8f", - "sha256:edcd9ef3fd460bb8a98eb1fcf99941e9fd9f275f45f1a82cb1359ec92975d647" + "sha256:25fb5240b195d17de949588628fdf93032ebf163524ef08933db0ea1f99bd685", + "sha256:3386b367e950a11b0568062b70cc026c6f645428a698d33d39e013aaeda4cc04", + "sha256:3becdb380d8916c873ad512f1701f8a92ce79ec6978ffde92919fd18d41da7fb", + "sha256:4ae38bf8ba6f39d5b83f78661273216e7db5b00f08be7592062cb1fc8b8ba542", + "sha256:8047da932432aa32c515ec1447ea79ce578d0559362ca3605f8e9568f844e3c6", + "sha256:8f1c00aa35f504197561060ca4c21d3cc079ba29cf6dd2fe61024c70160c990b", + "sha256:922a69fa533cb0c793b483becaaa0845f655151e7256ec73630a1b2e9ebcb660", + "sha256:9693f35162dc6208d10b10ddf0458cc09ad70c30ba689d9206e02cd836ce28a3", + "sha256:a0f1c7edf116a12f7245be06120b1852275f9506a7d90227648b250755a03923", + "sha256:a36e75df6c767cbf46f61a91c70b3ba71811dfa0aca4a324d9407a06a8b7a2e7", + "sha256:aba5c812f8ee8a3ff3be51887ca2d55fb8e268439ed44110d3846e4229eb0e8b", + "sha256:ad6f1796e37db2223d2a3f302f586f74c72c630b48a9872c1e7ae8e92e0ab669", + "sha256:ae67501c95606072aafa865b6ed47343ac6484472a2f95490ba151f6347acfc2", + "sha256:b2fcf9402fde2672545b139694284dc3b665fd1be660d73eca6805197ef776a3", + "sha256:b52b88021b9541a60531142b0a451baca08d28b74a723d0c99b13c8c8d48d604", + "sha256:b7d336912853d7b77f9b2c24eeed6a5065d0a0cc0d3b6a5a45ad6d1d05fb8cd8", + "sha256:bd9ba4f332cf57b2c1f698be0728c020399ef3040577cde2939f2e045b39c1e5", + "sha256:be9be735f827820a06340dff2ddea1fb7234561fa5e6300a62fe7f54d40546a0", + "sha256:cca7741c0fcc765568350cb139e92b7f9f3c9a08c4f32591d18ab0a6ac9e71b6", + "sha256:d0d19fb2441947b58fbf91336638c2b9f4cc98e05e1045404d7a4cb7cddc7a65", + "sha256:e02794ac791662a5eafc6ffeaf9bcc149035a0e48eb0a9d40a8feb4622605a3d", + "sha256:e0f30db709c939cabf64a6dc5babb276e6d823fd84464ab916f9b9ba5623ca15", + "sha256:e92c2d33858c8f560671b448205a268096e17870dcf60a9bb3ac7bfbafb7f5f9" ], "markers": "python_version >= '3.6'", - "version": "==2.1.2" + "version": "==2.1.6" }, "zipp": { "hashes": [ - "sha256:3607921face881ba3e026887d8150cca609d517579abe052ac81fc5aeffdbd76", - "sha256:51cb66cc54621609dd593d1787f286ee42a5c0adbb4b29abea5a63edc3e03098" + "sha256:71c644c5369f4a6e07636f0aa966270449561fcea2e3d6747b8d23efaa9d7832", + "sha256:9fe5ea21568a0a70e50f273397638d39b03353731e6cbbb3fd8502a33fec40bc" ], "markers": "python_version >= '3.6'", - "version": "==3.4.1" + "version": "==3.6.0" } } } diff --git a/README.rst b/README.rst index 7b18ec0..2ec4c67 100644 --- a/README.rst +++ b/README.rst @@ -36,6 +36,11 @@ Data Cleaning - Deduplicate alleles on MT - Assign PAR SNPs to the X or Y chromosome +Analysis +```````` +- Derive sex from SNPs +- Predict ancestry from SNPs (when installed with `ezancestry <https://github.com/arvkevi/ezancestry>`_) + Supported Genotype Files ------------------------ ``snps`` supports `VCF <https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3137218/>`_ files and @@ -72,6 +77,11 @@ Python dependencies) via ``pip``:: $ pip install snps +For `ancestry prediction <https://snps.readthedocs.io/en/stable/snps.html#snps.snps.SNPs.predict_ancestry>`_ +capability, ``snps`` can be installed with `ezancestry <https://github.com/arvkevi/ezancestry>`_:: + + $ pip install snps[ezancestry] + Examples -------- Download Example Data diff --git a/setup.py b/setup.py index b1d49d8..dd5d674 100644 --- a/setup.py +++ b/setup.py @@ -119,6 +119,7 @@ }, keywords="snps dna chromosomes bioinformatics vcf", install_requires=["numpy", "pandas!=1.0.0,!=1.1.0", "atomicwrites"], + extras_require={"ezancestry": ["ezancestry"]}, python_requires=">=3.7.1", platforms=["any"], ) diff --git a/src/snps/io/reader.py b/src/snps/io/reader.py index a43d64e..c869439 100644 --- a/src/snps/io/reader.py +++ b/src/snps/io/reader.py @@ -68,7 +68,7 @@ def get_empty_snps_dataframe(): - """ Get empty dataframe normalized for usage with ``snps``. + """Get empty dataframe normalized for usage with ``snps``. Returns ------- @@ -81,10 +81,10 @@ def get_empty_snps_dataframe(): class Reader: - """ Class for reading and parsing raw data / genotype files. """ + """Class for reading and parsing raw data / genotype files.""" def __init__(self, file="", only_detect_source=False, resources=None, rsids=()): - """ Initialize a `Reader`. + """Initialize a `Reader`. Parameters ---------- @@ -104,7 +104,7 @@ def __init__(self, file="", only_detect_source=False, resources=None, rsids=()): self._rsids = frozenset(rsids) def read(self): - """ Read and parse a raw data / genotype file. + """Read and parse a raw data / genotype file. Returns ------- @@ -195,7 +195,7 @@ def read(self): @classmethod def read_file(cls, file, only_detect_source, resources, rsids): - """ Read `file`. + """Read `file`. Parameters ---------- @@ -362,12 +362,12 @@ def _handle_bytes_data(self, file, include_data=False): @staticmethod def is_zip(bytes_data): - """ Check whether or not a bytes_data file is a valid Zip file.""" + """Check whether or not a bytes_data file is a valid Zip file.""" return zipfile.is_zipfile(io.BytesIO(bytes_data)) @staticmethod def is_gzip(bytes_data): - """ Check whether or not a bytes_data file is a valid gzip file.""" + """Check whether or not a bytes_data file is a valid gzip file.""" return binascii.hexlify(bytes_data[:2]) == b"1f8b" @staticmethod @@ -379,7 +379,7 @@ def _read_line(f, decode): return f.readline() def read_helper(self, source, parser): - """ Generic method to help read files. + """Generic method to help read files. Parameters ---------- @@ -431,7 +431,7 @@ def read_helper(self, source, parser): return {"snps": df, "source": source, "phased": phased, "build": build} def read_23andme(self, file, compression): - """ Read and parse 23andMe file. + """Read and parse 23andMe file. https://www.23andme.com @@ -515,7 +515,7 @@ def parser(): return self.read_helper("23andMe", parser) def read_ftdna(self, file, compression): - """ Read and parse Family Tree DNA (FTDNA) file. + """Read and parse Family Tree DNA (FTDNA) file. https://www.familytreedna.com @@ -602,7 +602,7 @@ def parser(): return self.read_helper("FTDNA", parser) def read_ftdna_famfinder(self, file, compression): - """ Read and parse Family Tree DNA (FTDNA) "famfinder" file. + """Read and parse Family Tree DNA (FTDNA) "famfinder" file. https://www.familytreedna.com @@ -641,7 +641,7 @@ def parser(): return self.read_helper("FTDNA", parser) def read_ancestry(self, file, compression): - """ Read and parse Ancestry.com file. + """Read and parse Ancestry.com file. http://www.ancestry.com @@ -690,7 +690,7 @@ def parser(): return self.read_helper("AncestryDNA", parser) def read_myheritage(self, file, compression): - """ Read and parse MyHeritage file. + """Read and parse MyHeritage file. https://www.myheritage.com @@ -745,7 +745,7 @@ def parser(): return self.read_helper("MyHeritage", parser) def read_livingdna(self, file, compression): - """ Read and parse LivingDNA file. + """Read and parse LivingDNA file. https://livingdna.com/ @@ -777,7 +777,7 @@ def parser(): return self.read_helper("LivingDNA", parser) def read_mapmygenome(self, file, compression, header): - """ Read and parse Mapmygenome file. + """Read and parse Mapmygenome file. https://mapmygenome.in @@ -827,7 +827,7 @@ def parse(rsid_col_name, rsid_col_idx): return self.read_helper("Mapmygenome", parser) def read_genes_for_good(self, file, compression): - """ Read and parse Genes For Good file. + """Read and parse Genes For Good file. https://genesforgood.sph.umich.edu/readme/readme1.2.txt @@ -1008,7 +1008,7 @@ def parser(): return self.read_helper(source, parser) def read_tellmegen(self, file, compression): - """ Read and parse tellmeGen files. + """Read and parse tellmeGen files. https://www.tellmegen.com/ @@ -1058,7 +1058,7 @@ def parser(): return self.read_helper("tellmeGen", parser) def read_gsa(self, data_or_filename, compresion, comments): - """ Read and parse Illumina Global Screening Array files + """Read and parse Illumina Global Screening Array files Parameters @@ -1085,7 +1085,7 @@ def read_gsa(self, data_or_filename, compresion, comments): return self._read_gsa_helper(data_or_filename, source) def read_dnaland(self, file, compression): - """ Read and parse DNA.land files. + """Read and parse DNA.land files. https://dna.land/ @@ -1117,7 +1117,7 @@ def parser(): return self.read_helper("DNA.Land", parser) def read_snps_csv(self, file, comments, compression): - """ Read and parse CSV file generated by ``snps``. + """Read and parse CSV file generated by ``snps``. https://pypi.org/project/snps/ @@ -1178,7 +1178,7 @@ def parse_csv(sep): return self.read_helper(source, parser) def read_generic(self, file, compression, skip=1): - """ Read and parse generic CSV or TSV file. + """Read and parse generic CSV or TSV file. Notes ----- @@ -1244,7 +1244,7 @@ def parse(sep): return self.read_helper("generic", parser) def read_vcf(self, file, compression, provider, rsids=()): - """ Read and parse VCF file. + """Read and parse VCF file. Notes ----- diff --git a/src/snps/io/writer.py b/src/snps/io/writer.py index f4a2fcf..7975899 100644 --- a/src/snps/io/writer.py +++ b/src/snps/io/writer.py @@ -48,10 +48,10 @@ class Writer: - """ Class for writing SNPs to files. """ + """Class for writing SNPs to files.""" def __init__(self, snps=None, filename="", vcf=False, atomic=True, kwargs): - """ Initialize a `Writer`. + """Initialize a `Writer`. Parameters ---------- @@ -80,7 +80,7 @@ def write(self): @classmethod def write_file(cls, snps=None, filename="", vcf=False, atomic=True, kwargs): - """ Save SNPs to file. + """Save SNPs to file. Parameters ---------- @@ -106,7 +106,7 @@ def write_file(cls, snps=None, filename="", vcf=False, atomic=True, kwargs): return w.write() def _write_csv(self): - """ Write SNPs to a CSV file. + """Write SNPs to a CSV file. Returns ------- @@ -147,7 +147,7 @@ def _write_csv(self): ) def _write_vcf(self): - """ Write SNPs to a VCF file. + """Write SNPs to a VCF file. References ---------- diff --git a/src/snps/resources.py b/src/snps/resources.py index a040003..d7ee9a1 100644 --- a/src/snps/resources.py +++ b/src/snps/resources.py @@ -70,10 +70,10 @@ class Resources(metaclass=Singleton): - """ Object used to manage resources required by `snps`. """ + """Object used to manage resources required by `snps`.""" def __init__(self, resources_dir="resources"): - """ Initialize a ``Resources`` object. + """Initialize a ``Resources`` object. Parameters ---------- @@ -122,7 +122,7 @@ def get_reference_sequences( "MT", ), ): - """ Get Homo sapiens reference sequences for `chroms` of `assembly`. + """Get Homo sapiens reference sequences for `chroms` of `assembly`. Notes ----- @@ -163,7 +163,7 @@ def _reference_chroms_available(self, assembly, chroms): return False def get_assembly_mapping_data(self, source_assembly, target_assembly): - """ Get assembly mapping data. + """Get assembly mapping data. Parameters ---------- @@ -182,7 +182,7 @@ def get_assembly_mapping_data(self, source_assembly, target_assembly): ) def download_example_datasets(self): - """ Download example datasets from `openSNP <https://opensnp.org>`_. + """Download example datasets from `openSNP <https://opensnp.org>`_. Per openSNP, "the data is donated into the public domain using `CC0 1.0 <http://creativecommons.org/publicdomain/zero/1.0/>`_." @@ -217,7 +217,7 @@ def download_example_datasets(self): return paths def get_all_resources(self): - """ Get / download all resources used throughout `snps`. + """Get / download all resources used throughout `snps`. Notes ----- @@ -238,7 +238,7 @@ def get_all_resources(self): return resources def get_all_reference_sequences(self, kwargs): - """ Get Homo sapiens reference sequences for Builds 36, 37, and 38 from Ensembl. + """Get Homo sapiens reference sequences for Builds 36, 37, and 38 from Ensembl. Notes ----- @@ -254,7 +254,7 @@ def get_all_reference_sequences(self, kwargs): return self._reference_sequences def get_gsa_resources(self): - """ Get resources for reading Global Screening Array files. + """Get resources for reading Global Screening Array files. https://support.illumina.com/downloads/infinium-global-screening-array-v2-0-product-files.html @@ -270,7 +270,7 @@ def get_gsa_resources(self): } def get_dbsnp_151_37_reverse(self): - """ Get and load RSIDs that are on the reference reverse (-) strand in dbSNP 151 and lower. + """Get and load RSIDs that are on the reference reverse (-) strand in dbSNP 151 and lower. Returns ------- @@ -321,7 +321,7 @@ def get_dbsnp_151_37_reverse(self): return self._dbsnp_151_37_reverse def get_opensnp_datadump_filenames(self): - """ Get filenames internal to the `openSNP <https://opensnp.org>`_ datadump zip. + """Get filenames internal to the `openSNP <https://opensnp.org>`_ datadump zip. Per openSNP, "the data is donated into the public domain using `CC0 1.0 <http://creativecommons.org/publicdomain/zero/1.0/>`_." @@ -350,7 +350,7 @@ def get_opensnp_datadump_filenames(self): return self._opensnp_datadump_filenames def load_opensnp_datadump_file(self, filename): - """ Load the specified file from the openSNP datadump. + """Load the specified file from the openSNP datadump. Per openSNP, "the data is donated into the public domain using `CC0 1.0 <http://creativecommons.org/publicdomain/zero/1.0/>`_." @@ -380,7 +380,7 @@ def load_opensnp_datadump_file(self, filename): @staticmethod def _get_opensnp_datadump_filenames(filename): - """ Get list of filenames internal to the openSNP datadump zip. + """Get list of filenames internal to the openSNP datadump zip. Parameters ---------- @@ -400,7 +400,7 @@ def _get_opensnp_datadump_filenames(filename): @staticmethod def _write_data_to_gzip(f, data): - """ Write `data` to `f` in `gzip` format. + """Write `data` to `f` in `gzip` format. Parameters ---------- @@ -412,7 +412,7 @@ def _write_data_to_gzip(f, data): @staticmethod def _load_assembly_mapping_data(filename): - """ Load assembly mapping data. + """Load assembly mapping data. Parameters ---------- @@ -446,7 +446,7 @@ def _load_assembly_mapping_data(filename): def _get_paths_reference_sequences( self, sub_dir="fasta", assembly="GRCh37", chroms=() ): - """ Get local paths to Homo sapiens reference sequences from Ensembl. + """Get local paths to Homo sapiens reference sequences from Ensembl. Notes ----- @@ -550,7 +550,7 @@ def _create_reference_sequences(self, assembly, chroms, urls, paths): def _get_path_assembly_mapping_data( self, source_assembly, target_assembly, retries=10 ): - """ Get local path to assembly mapping data, downloading if necessary. + """Get local path to assembly mapping data, downloading if necessary. Parameters ---------- @@ -630,7 +630,7 @@ def _download_assembly_mapping_data( os.remove(f_tmp.name) def get_gsa_rsid(self): - """ Get and load GSA RSID map. + """Get and load GSA RSID map. https://support.illumina.com/downloads/infinium-global-screening-array-v2-0-product-files.html @@ -657,7 +657,7 @@ def get_gsa_rsid(self): return self._gsa_rsid_map def get_gsa_chrpos(self): - """ Get and load GSA chromosome position map. + """Get and load GSA chromosome position map. https://support.illumina.com/downloads/infinium-global-screening-array-v2-0-product-files.html @@ -697,7 +697,7 @@ def _get_path_opensnp_datadump(self): ) def _download_file(self, url, filename, compress=False, timeout=30): - """ Download a file to the resources folder. + """Download a file to the resources folder. Download data from `url`, save as `filename`, and optionally compress with gzip. @@ -765,7 +765,7 @@ def _download_file(self, url, filename, compress=False, timeout=30): @staticmethod def _print_download_msg(path): - """ Print download message. + """Print download message. Parameters ---------- @@ -776,10 +776,10 @@ def _print_download_msg(path): class ReferenceSequence: - """ Object used to represent and interact with a reference sequence. """ + """Object used to represent and interact with a reference sequence.""" def __init__(self, ID="", url="", path="", assembly="", species="", taxonomy=""): - """ Initialize a ``ReferenceSequence`` object. + """Initialize a ``ReferenceSequence`` object. Parameters ---------- @@ -818,7 +818,7 @@ def __repr__(self): @property def ID(self): - """ Get reference sequence chromosome. + """Get reference sequence chromosome. Returns ------- @@ -828,7 +828,7 @@ def ID(self): @property def chrom(self): - """ Get reference sequence chromosome. + """Get reference sequence chromosome. Returns ------- @@ -838,7 +838,7 @@ def chrom(self): @property def url(self): - """ Get URL to Ensembl reference sequence. + """Get URL to Ensembl reference sequence. Returns ------- @@ -848,7 +848,7 @@ def url(self): @property def path(self): - """ Get path to local reference sequence. + """Get path to local reference sequence. Returns ------- @@ -858,7 +858,7 @@ def path(self): @property def assembly(self): - """ Get reference sequence assembly. + """Get reference sequence assembly. Returns ------- @@ -868,7 +868,7 @@ def assembly(self): @property def build(self): - """ Get reference sequence build. + """Get reference sequence build. Returns ------- @@ -879,7 +879,7 @@ def build(self): @property def species(self): - """ Get reference sequence species. + """Get reference sequence species. Returns ------- @@ -889,7 +889,7 @@ def species(self): @property def taxonomy(self): - """ Get reference sequence taxonomy. + """Get reference sequence taxonomy. Returns ------- @@ -899,7 +899,7 @@ def taxonomy(self): @property def sequence(self): - """ Get reference sequence. + """Get reference sequence. Returns ------- @@ -910,7 +910,7 @@ def sequence(self): @property def md5(self): - """ Get reference sequence MD5 hash. + """Get reference sequence MD5 hash. Returns ------- @@ -921,7 +921,7 @@ def md5(self): @property def start(self): - """ Get reference sequence start position (1-based). + """Get reference sequence start position (1-based). Returns ------- @@ -932,7 +932,7 @@ def start(self): @property def end(self): - """ Get reference sequence end position (1-based). + """Get reference sequence end position (1-based). Returns ------- @@ -943,7 +943,7 @@ def end(self): @property def length(self): - """ Get reference sequence length. + """Get reference sequence length. Returns ------- @@ -953,7 +953,7 @@ def length(self): return self._sequence.size def clear(self): - """ Clear reference sequence. """ + """Clear reference sequence.""" self._sequence = np.array([], dtype=np.uint8) self._md5 = "" self._start = 0 diff --git a/src/snps/snps.py b/src/snps/snps.py index fa22528..2e73e2b 100644 --- a/src/snps/snps.py +++ b/src/snps/snps.py @@ -68,7 +68,7 @@ def __init__( processes=os.cpu_count(), rsids=(), ): - """ Object used to read, write, and remap genotype / raw data files. + """Object used to read, write, and remap genotype / raw data files. Parameters ---------- @@ -163,6 +163,7 @@ def __init__( if deduplicate_MT_chrom: self._deduplicate_MT_chrom() + else: logger.warning("no SNPs loaded...") @@ -177,7 +178,7 @@ def __repr__(self): @property def source(self): - """ Summary of the SNP data source(s). + """Summary of the SNP data source(s). Returns ------- @@ -188,7 +189,7 @@ def source(self): @property def snps(self): - """ Normalized SNPs. + """Normalized SNPs. Notes ----- @@ -219,7 +220,7 @@ def snps(self): @property def duplicate(self): - """ Duplicate SNPs. + """Duplicate SNPs. A duplicate SNP has the same RSID as another SNP. The first occurrence of the RSID is not considered a duplicate SNP. @@ -233,7 +234,7 @@ def duplicate(self): @property def discrepant_XY(self): - """ Discrepant XY SNPs. + """Discrepant XY SNPs. A discrepant XY SNP is a heterozygous SNP in the non-PAR region of the X or Y chromosome found during deduplication for a detected male genotype. @@ -247,7 +248,7 @@ def discrepant_XY(self): @property def heterozygous_MT(self): - """ Heterozygous SNPs on the MT chromosome found during deduplication. + """Heterozygous SNPs on the MT chromosome found during deduplication. Returns ------- @@ -258,7 +259,7 @@ def heterozygous_MT(self): @property def discrepant_vcf_position(self): - """ SNPs with discrepant positions discovered while saving VCF. + """SNPs with discrepant positions discovered while saving VCF. Returns ------- @@ -269,7 +270,7 @@ def discrepant_vcf_position(self): @property def discrepant_merge_positions(self): - """ SNPs with discrepant positions discovered while merging SNPs. + """SNPs with discrepant positions discovered while merging SNPs. Notes ----- @@ -295,7 +296,7 @@ def discrepant_merge_positions(self): @property def discrepant_merge_genotypes(self): - """ SNPs with discrepant genotypes discovered while merging SNPs. + """SNPs with discrepant genotypes discovered while merging SNPs. Notes ----- @@ -322,7 +323,7 @@ def discrepant_merge_genotypes(self): @property def discrepant_merge_positions_genotypes(self): - """ SNPs with discrepant positions and / or genotypes discovered while merging SNPs. + """SNPs with discrepant positions and / or genotypes discovered while merging SNPs. Notes ----- @@ -351,7 +352,7 @@ def discrepant_merge_positions_genotypes(self): @property def build(self): - """ Build of SNPs. + """Build of SNPs. Returns ------- @@ -361,7 +362,7 @@ def build(self): @property def build_detected(self): - """ Status indicating if build of SNPs was detected. + """Status indicating if build of SNPs was detected. Returns ------- @@ -371,7 +372,7 @@ def build_detected(self): @property def assembly(self): - """ Assembly of SNPs. + """Assembly of SNPs. Returns ------- @@ -388,7 +389,7 @@ def assembly(self): @property def count(self): - """ Count of SNPs. + """Count of SNPs. Returns ------- @@ -398,7 +399,7 @@ def count(self): @property def chromosomes(self): - """ Chromosomes of SNPs. + """Chromosomes of SNPs. Returns ------- @@ -412,7 +413,7 @@ def chromosomes(self): @property def chromosomes_summary(self): - """ Summary of the chromosomes of SNPs. + """Summary of the chromosomes of SNPs. Returns ------- @@ -449,7 +450,7 @@ def as_range(iterable): @property def sex(self): - """ Sex derived from SNPs. + """Sex derived from SNPs. Returns ------- @@ -463,7 +464,7 @@ def sex(self): @property def unannotated_vcf(self): - """ Indicates if VCF file is unannotated. + """Indicates if VCF file is unannotated. Returns ------- @@ -476,7 +477,7 @@ def unannotated_vcf(self): @property def phased(self): - """ Indicates if genotype is phased. + """Indicates if genotype is phased. Returns ------- @@ -485,7 +486,7 @@ def phased(self): return self._phased def heterozygous(self, chrom=""): - """ Get heterozygous SNPs. + """Get heterozygous SNPs. Parameters ---------- @@ -506,7 +507,7 @@ def heterozygous(self, chrom=""): ] def homozygous(self, chrom=""): - """ Get homozygous SNPs. + """Get homozygous SNPs. Parameters ---------- @@ -527,7 +528,7 @@ def homozygous(self, chrom=""): ] def notnull(self, chrom=""): - """ Get not null genotype SNPs. + """Get not null genotype SNPs. Parameters ---------- @@ -545,7 +546,7 @@ def notnull(self, chrom=""): @property def summary(self): - """ Summary of SNPs. + """Summary of SNPs. Returns ------- @@ -567,7 +568,7 @@ def summary(self): @property def valid(self): - """ Determine if ``SNPs`` is valid. + """Determine if ``SNPs`` is valid. ``SNPs`` is valid when the input file has been successfully parsed. @@ -582,7 +583,7 @@ def valid(self): return True def save(self, filename="", vcf=False, atomic=True, kwargs): - """ Save SNPs to file. + """Save SNPs to file. Parameters ---------- @@ -628,7 +629,7 @@ def _read_raw_data(self, file, only_detect_source, rsids): return Reader.read_file(file, only_detect_source, self._resources, rsids) def _assign_par_snps(self): - """ Assign PAR SNPs to the X or Y chromosome using SNP position. + """Assign PAR SNPs to the X or Y chromosome using SNP position. References ----- @@ -702,7 +703,7 @@ def _extract_build(self, item): return int(assembly_name[-2:]) def detect_build(self): - """ Detect build of SNPs. + """Detect build of SNPs. Use the coordinates of common SNPs to identify the build / assembly of a genotype file that is being loaded. @@ -798,7 +799,7 @@ def lookup_build_with_snp_pos(pos, s): return build def get_count(self, chrom=""): - """ Count of SNPs. + """Count of SNPs. Parameters ---------- @@ -817,7 +818,7 @@ def determine_sex( y_snps_not_null_threshold=0.3, chrom="X", ): - """ Determine sex from SNPs using thresholds. + """Determine sex from SNPs using thresholds. Parameters ---------- @@ -907,7 +908,7 @@ def _deduplicate_alleles(self, rsids): ) def _deduplicate_sex_chrom(self, chrom): - """ Deduplicate a chromosome in the non-PAR region. """ + """Deduplicate a chromosome in the non-PAR region.""" discrepant_XY_snps = self._get_non_par_snps(chrom) @@ -925,12 +926,12 @@ def _deduplicate_sex_chrom(self, chrom): self._deduplicate_alleles(non_par_snps) def _deduplicate_XY_chrom(self): - """ Fix chromosome issue where some data providers duplicate male X and Y chromosomes""" + """Fix chromosome issue where some data providers duplicate male X and Y chromosomes""" self._deduplicate_sex_chrom("X") self._deduplicate_sex_chrom("Y") def _deduplicate_MT_chrom(self): - """ Deduplicate MT chromosome. """ + """Deduplicate MT chromosome.""" heterozygous_MT_snps = self._snps.loc[self.heterozygous("MT").index].index # save heterozygous MT SNPs @@ -945,7 +946,7 @@ def _deduplicate_MT_chrom(self): @staticmethod def get_par_regions(build): - """ Get PAR regions for the X and Y chromosomes. + """Get PAR regions for the X and Y chromosomes. Parameters ---------- @@ -998,7 +999,7 @@ def get_par_regions(build): return pd.DataFrame() def sort(self): - """ Sort SNPs based on ordered chromosome list and position. """ + """Sort SNPs based on ordered chromosome list and position.""" sorted_list = sorted( (str(x) for x in self._snps["chrom"].unique()), key=self._natural_sort_key ) @@ -1027,7 +1028,7 @@ def sort(self): self._snps = snps def remap(self, target_assembly, complement_bases=True): - """ Remap SNP coordinates from one assembly to another. + """Remap SNP coordinates from one assembly to another. This method uses the assembly map endpoint of the Ensembl REST API service (via ``Resources``'s ``EnsemblRestClient``) to convert SNP coordinates / positions from one @@ -1146,7 +1147,7 @@ def remap(self, target_assembly, complement_bases=True): return chromosomes_remapped, chromosomes_not_remapped def _remapper(self, task): - """ Remap SNPs for a chromosome. + """Remap SNPs for a chromosome. Parameters ---------- @@ -1269,7 +1270,7 @@ def merge( remap=True, chrom="", ): - """ Merge other ``SNPs`` objects into this ``SNPs`` object. + """Merge other ``SNPs`` objects into this ``SNPs`` object. Parameters ---------- @@ -1611,3 +1612,110 @@ def is_valid(self): DeprecationWarning, ) return self.valid + + def predict_ancestry( + self, + output_directory=None, + write_predictions=False, + models_directory=None, + aisnps_directory=None, + n_components=None, + k=None, + thousand_genomes_directory=None, + samples_directory=None, + algorithm=None, + aisnps_set=None, + ): + """Predict genetic ancestry for SNPs. + + Predictions by `ezancestry <https://github.com/arvkevi/ezancestry>`_. + + Notes + ----- + Populations below are described `here <https://www.internationalgenome.org/faq/what-do-the-population-codes-mean/>`_. + + Parameters + ---------- + various : optional + See the available settings for `predict` at `ezancestry <https://github.com/arvkevi/ezancestry>`_. + + Returns + ------- + dict + dict with the following keys: + + `population_code` (str) + max predicted population for the sample + `population_description` (str) + descriptive name of the population + `population_percent` (float) + predicted probability for the max predicted population + `superpopulation_code` (str) + max predicted super population (continental) for the sample + `superpopulation_description` (str) + descriptive name of the super population + `superpopulation_percent` (float) + predicted probability for the max predicted super population + `ezancestry_df` (pandas.DataFrame) + pandas.DataFrame with the following columns: + + `component1`, `component2`, `component3` + The coordinates of the sample in the dimensionality-reduced component space. Can be + used as (x, y, z,) coordinates for plotting in a 3d scatter plot. + `predicted_population_population` + The max predicted population for the sample. + `ACB`, `ASW`, `BEB`, `CDX`, `CEU`, `CHB`, `CHS`, `CLM`, `ESN`, `FIN`, `GBR`, `GIH`, `GWD`, `IBS`, `ITU`, `JPT`, `KHV`, `LWK`, `MSL`, `MXL`, `PEL`, `PJL`, `PUR`, `STU`, `TSI`, `YRI` + Predicted probabilities for each of the populations. These sum to 1.0. + `predicted_population_superpopulation` + The max predicted super population (continental) for the sample. + `AFR`, `AMR`, `EAS`, `EUR`, `SAS` + Predicted probabilities for each of the super populations. These sum to 1.0. + `population_description`, `superpopulation_name` + Descriptive names of the population and super population. + + """ + if not self.valid: + return {} + + try: + from ezancestry.commands import predict + except ModuleNotFoundError: + raise ModuleNotFoundError( + "Ancestry prediction requires the ezancestry package; please install it using pip install ezancestry" + ) + + def max_pop(row): + popcode = row["predicted_population_population"] + popdesc = row["population_description"] + poppct = row[popcode] + superpopcode = row["predicted_population_superpopulation"] + superpopdesc = row["superpopulation_name"] + superpoppct = row[superpopcode] + + return { + "population_code": popcode, + "population_description": popdesc, + "population_percent": poppct, + "superpopulation_code": superpopcode, + "superpopulation_description": superpopdesc, + "superpopulation_percent": superpoppct, + } + + predictions = predict( + self.snps, + output_directory, + write_predictions, + models_directory, + aisnps_directory, + n_components, + k, + thousand_genomes_directory, + samples_directory, + algorithm, + aisnps_set, + ) + + d = dict(predictions.apply(max_pop, axis=1).iloc[0]) + d["ezancestry_df"] = predictions + + return d diff --git a/src/snps/utils.py b/src/snps/utils.py index a157da6..cd36801 100644 --- a/src/snps/utils.py +++ b/src/snps/utils.py @@ -55,7 +55,7 @@ class Parallelizer: def __init__(self, parallelize=False, processes=os.cpu_count()): - """ Initialize a `Parallelizer`. + """Initialize a `Parallelizer`. Parameters ---------- @@ -68,7 +68,7 @@ def __init__(self, parallelize=False, processes=os.cpu_count()): self._processes = processes def __call__(self, f, tasks): - """ Optionally parallelize execution of a function. + """Optionally parallelize execution of a function. Parameters ---------- @@ -100,7 +100,7 @@ def __call__(cls, args, kwargs): def create_dir(path): - """ Create directory specified by `path` if it doesn't already exist. + """Create directory specified by `path` if it doesn't already exist. Parameters ---------- @@ -124,7 +124,7 @@ def create_dir(path): def save_df_as_csv( df, path, filename, comment="", prepend_info=True, atomic=True, kwargs ): - """ Save dataframe to a CSV file. + """Save dataframe to a CSV file. Parameters ---------- @@ -198,7 +198,7 @@ def save_df_as_csv( def clean_str(s): - """ Clean a string so that it can be used as a Python variable name. + """Clean a string so that it can be used as a Python variable name. Parameters ---------- @@ -216,7 +216,7 @@ def clean_str(s): def zip_file(src, dest, arcname): - """ Zip a file. + """Zip a file. Parameters ---------- @@ -240,7 +240,7 @@ def zip_file(src, dest, arcname): def gzip_file(src, dest): - """ Gzip a file. + """Gzip a file. Parameters ----------	Support current version of `black` Fixes #144.		2021-11-06T03:37:59	[]	[]
apriha/snps	apriha__snps-137	28de7e13dc0f48ef3234d3be7be35d3ff54a6db3	diff --git a/.circleci/config.yml b/.circleci/config.yml new file mode 100644 index 00000000..32c52d1a --- /dev/null +++ b/.circleci/config.yml @@ -0,0 +1,54 @@ +# Python CircleCI 2.1 configuration file +# +# Check https://circleci.com/docs/2.0/language-python/ for more details +# +version: 2.1 +jobs: + unittests: + docker: + - image: circleci/python:3.8.3 + environment: + PIPENV_VENV_IN_PROJECT: "1" + + steps: + - checkout + # increment vX to manually purge cache + - restore_cache: + name: Restore dependencies from cache, if possible + keys: + - v1-venv-{{ checksum "Pipfile.lock" }} + - run: + name: Dependency installation + # pipenv is itself include in circleci docker image + # prefer the sync command to remove unused cached deps + command: pipenv sync --dev + - save_cache: + name: Save server dependencies + paths: + - .venv + key: v1-venv-{{ checksum "Pipfile.lock" }} + - run: + name: Dependency output for debugging + command: pipenv graph + #- run: + # name: Dependency security check + # command: pipenv check + - run: + name: run tests + command: \| + mkdir -p server-test-reports/pytest + pipenv run pytest --cov=snps --cov-report xml:server-test-reports/coverage.xml --junitxml server-test-reports/pytest/tests.xml tests + bash <(curl -s https://codecov.io/bash) + - store_test_results: + # must be a directory with named subdirectories + path: server-test-reports + - store_artifacts: + path: server-test-reports + - run: + name: check linting + command: pipenv run black --check --diff . + +workflows: + test: + jobs: + - "unittests" \ No newline at end of file diff --git a/.gitignore b/.gitignore index 5e4fd224..4d9ef8d9 100644 --- a/.gitignore +++ b/.gitignore @@ -113,6 +113,7 @@ tests/output/ tests/resources/* !tests/resources/gsa_chrpos_map.txt !tests/resources/gsa_rsid_map.txt +!tests/resources/dbsnp_151_37_reverse.txt tests/input/23andme.txt.zip tests/input/discrepant_snps[12].csv tests/input/ftdna.csv.gz diff --git a/src/snps/io/reader.py b/src/snps/io/reader.py index 8581351c..a43d64ee 100644 --- a/src/snps/io/reader.py +++ b/src/snps/io/reader.py @@ -171,7 +171,7 @@ def read(self): d = self.read_mapmygenome(file, compression, first_line) elif "lineage" in first_line or "snps" in first_line: d = self.read_snps_csv(file, comments, compression) - elif "Chromosome" in first_line: + elif "rsid\tChromosome\tposition\tgenotype" == first_line.strip(): d = self.read_tellmegen(file, compression) elif re.match("^#[ \t]rsid[, \t]chr", first_line): d = self.read_generic(file, compression) @@ -183,10 +183,9 @@ def read(self): d = self.read_genes_for_good(file, compression) elif "DNA.Land" in comments: d = self.read_dnaland(file, compression) - elif "CODIGO46" in comments: - d = self.read_codigo46(file) - elif "SANO" in comments: - d = self.read_sano(file) + elif first_line.startswith("[Header]"): + # Global Screening Array, includes SANO and CODIGO46 + d = self.read_gsa(file, compression, comments) # detect build from comments if build was not already detected from `read` method if not d["build"]: @@ -259,31 +258,71 @@ def _extract_comments(self, f, decode=False, include_data=False): return first_line, comments, data def _detect_build_from_comments(self, comments, source): - comments = comments.lower() - if "build 37" in comments: - return 37 - elif "build 36" in comments: - return 36 - - # allow more variations for VCF + # if its a VCF parse these properly if source == "vcf": - if "https://pypi.org/project/snps/" in comments: # remove `snps` version - comments = f"{comments[: comments.find('snps v')]}{comments[comments.find('https://pypi.org/project/snps/'):]}" - if "hg19" in comments: - return 37 - elif "ncbi36" in comments: + for line in comments.split("\n"): + line = line.strip() + if not line: + # skip blanks + continue + assert line.startswith("#"), line + if not line.startswith("##"): + # skip comments but not preamble + continue + if "=" not in line: + # skip lines without key/value in + continue + line = line[2:].strip() + key = line[: line.index("=")] + value = line[line.index("=") + 1 :] + if key.lower() == "contig": + assert value.startswith("<"), value + assert value.endswith(">"), value + parts = value[1:-1].split(",") + for part in parts: + part_key = part[: part.index("=")] + part_value = part[part.index("=") + 1 :] + if part_key.lower() == "assembly": + if "36" in part_value: + return 36 + elif "37" in part_value or "hg19" in part_value: + return 37 + elif "38" in part_value: + return 38 + elif part_key.lower() == "length": + if "249250621" == part_value: + return 37 # length of chromosome 1 + elif "248956422" == part_value: + return 38 # length of chromosome 1 + # couldn't find anything + return 0 + else: + # not a vcf + if "build 36" in comments.lower(): return 36 - elif "grch38" in comments: - return 38 - elif "build 38" in comments: - return 38 - elif "b37" in comments: + elif "build 37" in comments.lower(): return 37 - elif "hg38" in comments: + elif "build 38" in comments.lower(): return 38 - elif "b38" in comments: + # these can cause false positives + # elif "b37" in comments.lower(): + # return 37 + # elif "b38" in comments.lower(): + # return 38 + # elif "hg19" in comments.lower(): + # return 37 + # elif "hg38" in comments.lower(): + # return 38 + elif "grch38" in comments.lower(): return 38 - return 0 + elif "grch37" in comments.lower(): + return 37 + elif "249250621" in comments.lower(): + return 37 # length of chromosome 1 + elif "248956422" in comments.lower(): + return 38 # length of chromosome 1 + else: + return 0 def _handle_bytes_data(self, file, include_data=False): compression = "infer" @@ -323,12 +362,12 @@ def _handle_bytes_data(self, file, include_data=False): @staticmethod def is_zip(bytes_data): - """Check whether or not a bytes_data file is a valid Zip file.""" + """ Check whether or not a bytes_data file is a valid Zip file.""" return zipfile.is_zipfile(io.BytesIO(bytes_data)) @staticmethod def is_gzip(bytes_data): - """Check whether or not a bytes_data file is a valid gzip file.""" + """ Check whether or not a bytes_data file is a valid gzip file.""" return binascii.hexlify(bytes_data[:2]) == b"1f8b" @staticmethod @@ -476,7 +515,7 @@ def parser(): return self.read_helper("23andMe", parser) def read_ftdna(self, file, compression): - """Read and parse Family Tree DNA (FTDNA) file. + """ Read and parse Family Tree DNA (FTDNA) file. https://www.familytreedna.com @@ -623,7 +662,7 @@ def parser(): comment="#", header=0, engine="c", - sep="\s+", + sep=r"\s+", # delim_whitespace=True, # https://stackoverflow.com/a/15026839 na_values=0, names=["rsid", "chrom", "pos", "allele1", "allele2"], @@ -777,7 +816,8 @@ def parse(rsid_col_name, rsid_col_idx): else: df = parse("SNP.Name", 0) - df["genotype"] = df["Allele1...Top"] + df["Allele2...Top"] + # uses Illumina definition of "Plus" from https://emea.support.illumina.com/bulletins/2017/06/how-to-interpret-dna-strand-and-allele-information-for-infinium-.html + df["genotype"] = df["Allele1...Plus"] + df["Allele2...Plus"] df.rename(columns={"Chr": "chrom", "Position": "pos"}, inplace=True) df.index.name = "rsid" df = df[["chrom", "pos", "genotype"]] @@ -818,52 +858,149 @@ def parser(): return self.read_helper("GenesForGood", parser) - def _read_gsa_helper(self, file, source, strand, dtypes, na_values="--"): + def _read_gsa_helper(self, file, source): def parser(): - gsa_resources = self._resources.get_gsa_resources() + # read the comments so we get to the actual data if isinstance(file, str): try: with open(file, "rb") as f: - first_line, comments, data = self._extract_comments( + _, _, data = self._extract_comments( f, decode=True, include_data=True ) except UnicodeDecodeError: # compressed file on filesystem with open(file, "rb") as f: - ( - first_line, - comments, - data, - compression, - ) = self._handle_bytes_data(f.read(), include_data=True) + _, _, data, _ = self._handle_bytes_data( + f.read(), include_data=True + ) else: - first_line, comments, data, compression = self._handle_bytes_data( - file.read(), include_data=True + _, _, data, _ = self._handle_bytes_data(file.read(), include_data=True) + + # turn the data into a pandas dataframe for manipulation + df = pd.read_csv( + io.StringIO(data), + sep="\t", + engine="c", + dtype={ + "Position": NORMALIZED_DTYPES["pos"], + "Chr": NORMALIZED_DTYPES["chrom"], + }, + ) + + # reserve columns we want out + assert "rsid" not in df.columns + assert "chrom" not in df.columns + assert "pos" not in df.columns + assert "genotype" not in df.columns + + # prefer the specified chromosome and position, in prsent + # this uses the gsa names, so needs to be done before rsid + if "Chr" in df.columns and "Position" in df.columns: + # put the chromosome in the right column with the right type + df["chrom"] = df["Chr"] + # put the position in the right column with the right type + df["pos"] = df["Position"] + + else: + # use an external source to map snp names to chromosome and position + df = df.merge( + self._resources.get_gsa_chrpos(), + how="inner", # inner join, everything must have a position + left_on="SNP Name", + right_on="gsaname_chrpos", + suffixes=(None, "_gsa"), ) + # make sure its the right types + df["chrom"] = df["gsachr"].apply(str).astype(NORMALIZED_DTYPES["chrom"]) + df["pos"] = df["gsapos"].astype(NORMALIZED_DTYPES["pos"]) + + # use the given rsid when avaliable, SNP Name when unavaliable + df["rsid"] = df["SNP Name"].astype(NORMALIZED_DTYPES["rsid"]) + if "RsID" in df.columns: + df.loc[df["RsID"] != ".", "rsid"] = df.loc[df["rsid"] != ".", "RsID"] + else: + # if given RSIDs are not avaliable, then use the external mapping to turn + # SNP names into rsids where possible + df = df.merge( + self._resources.get_gsa_rsid(), + how="left", # left-hand join, gsa rsids may be NA + left_on="SNP Name", + right_on="gsaname_rsid", + suffixes=(None, "_gsa_rsid"), + ) + df.loc[~pd.isna(df["gsaname_rsid"]), "rsid"] = df.loc[ + ~pd.isna(df["gsaname_rsid"]), "gsarsid" + ] - data_io = io.StringIO(data) - df = pd.read_csv(data_io, sep="\t", dtype=dtypes, na_values=na_values) + # combine the alleles into genotype + # prefer Plus strand as that is forward reference + if "Allele1 - Plus" in df.columns and "Allele2 - Plus" in df.columns: + df["genotype"] = (df["Allele1 - Plus"] + df["Allele2 - Plus"]).astype( + NORMALIZED_DTYPES["genotype"] + ) + elif ( + "Allele1 - Forward" in df.columns and "Allele2 - Forward" in df.columns + ): + # if strand is forward, need to take reverse complement of some* rsids + # this is because it is Illumina forward, which is dbSNP strand, which + # is reverse reference for some RSIDs before dbSNP 151. + + # load list of reversable rsids + dbsnp151 = self._resources.get_dbsnp_151_37_reverse() + # keep only the rsids + dbsnp151 = dbsnp151.filter(items=("dbsnp151revrsid",), axis=1) + + # add it as an extra column + df = df.merge( + dbsnp151, + how="left", + left_on="rsid", + right_on="dbsnp151revrsid", + suffixes=(None, "_dbsnp151rev"), + ) - def map_rsids(x): - return gsa_resources["rsid_map"].get(x) + # create plus strand columns from the forward alleles and flip them if appropriate + for i in (1, 2): + df[f"Allele{i} - Plus"] = df[f"Allele{i} - Forward"] + df.loc[ + (df[f"Allele{i} - Forward"] == "A") + & (~pd.isna(df["dbsnp151revrsid"])), + f"Allele{i} - Plus", + ] = "T" + df.loc[ + (df[f"Allele{i} - Forward"] == "T") + & (~pd.isna(df["dbsnp151revrsid"])), + f"Allele{i} - Plus", + ] = "A" + df.loc[ + (df[f"Allele{i} - Forward"] == "C") + & (~pd.isna(df["dbsnp151revrsid"])), + f"Allele{i} - Plus", + ] = "G" + df.loc[ + (df[f"Allele{i} - Forward"] == "G") + & (~pd.isna(df["dbsnp151revrsid"])), + f"Allele{i} - Plus", + ] = "C" + + # create a genotype by combining the new plus columns + df["genotype"] = (df["Allele1 - Plus"] + df["Allele2 - Plus"]).astype( + NORMALIZED_DTYPES["genotype"] + ) + else: + raise ValueError("No supported allele column") - def map_chr(x): - chrpos = gsa_resources["chrpos_map"].get(x) - return chrpos.split(":")[0] if chrpos else None + # mark -- genotype as na + df.loc[df["genotype"] == "--", "genotype"] = np.nan - def map_pos(x): - chrpos = gsa_resources["chrpos_map"].get(x) - return chrpos.split(":")[1] if chrpos else None + # keep only the columns we want + df = df.filter(items=("rsid", "chrom", "pos", "genotype"), axis=1) - df["rsid"] = df["SNP Name"].apply(map_rsids) - df["chrom"] = df["SNP Name"].apply(map_chr) - df["pos"] = df["SNP Name"].apply(map_pos) - df["genotype"] = df[f"Allele1 - {strand}"] + df[f"Allele2 - {strand}"] + # discard rows without values df.dropna(subset=["rsid", "chrom", "pos"], inplace=True) - df = df.astype(NORMALIZED_DTYPES) - df = df[["rsid", "chrom", "pos", "genotype"]] + # reindex for the new identifiers df.set_index(["rsid"], inplace=True) return (df,) @@ -887,57 +1024,65 @@ def read_tellmegen(self, file, compression): """ def parser(): - gsa_resources = self._resources.get_gsa_resources() - df = pd.read_csv( file, sep="\t", skiprows=1, na_values="--", names=["rsid", "chrom", "pos", "genotype"], - index_col=0, dtype=NORMALIZED_DTYPES, compression=compression, ) - df.rename(index=gsa_resources["rsid_map"], inplace=True) - return (df,) - return self.read_helper("tellmeGen", parser) + # use the external mapping to turn + # SNP names into rsids where possible + df = df.merge( + self._resources.get_gsa_rsid(), + how="left", # left-hand join, gsa rsids may be NA + left_on="rsid", + right_on="gsaname_rsid", + suffixes=(None, "_gsa_rsid"), + ) + df.loc[~pd.isna(df["gsaname_rsid"]), "rsid"] = df.loc[ + ~pd.isna(df["gsaname_rsid"]), "gsarsid" + ] - def read_codigo46(self, file): - """ Read and parse Codigo46 files. + # keep only the columns we want + df = df.filter(items=("rsid", "chrom", "pos", "genotype"), axis=1) - https://codigo46.com.mx + # reindex for the new identifiers + df.set_index(["rsid"], inplace=True) - Parameters - ---------- - data : str - data string + return (df,) - Returns - ------- - dict - result of `read_helper` - """ - return self._read_gsa_helper(file, "Codigo46", "Plus", {}) + return self.read_helper("tellmeGen", parser) - def read_sano(self, file): - """ Read and parse Sano Genetics files. + def read_gsa(self, data_or_filename, compresion, comments): + """ Read and parse Illumina Global Screening Array files - https://sanogenetics.com Parameters ---------- - data : str - data string + data_or_filename : str or bytes + either the filename to read from or the bytes data itself Returns ------- dict result of `read_helper` """ - dtype = {"Chr": object, "Position": np.uint32} - return self._read_gsa_helper(file, "Sano", "Forward", dtype, na_values="-") + + # pick the source + # ideally we want something more specific than GSA + if "SANO" in comments: + source = "Sano" + elif "CODIGO46" in comments: + source = "Codigo46" + else: + # default to generic global screening array + source = "GSA" + + return self._read_gsa_helper(data_or_filename, source) def read_dnaland(self, file, compression): """ Read and parse DNA.land files. diff --git a/src/snps/resources.py b/src/snps/resources.py index 42bb319b..a0400034 100644 --- a/src/snps/resources.py +++ b/src/snps/resources.py @@ -61,6 +61,7 @@ from atomicwrites import atomic_write import numpy as np +import pandas as pd from snps.ensembl import EnsemblRestClient from snps.utils import create_dir, Singleton @@ -81,8 +82,13 @@ def __init__(self, resources_dir="resources"): """ self._resources_dir = os.path.abspath(resources_dir) self._ensembl_rest_client = EnsemblRestClient() + self._init_resource_attributes() + + def _init_resource_attributes(self): self._reference_sequences = {} - self._gsa_resources = {} + self._gsa_rsid_map = None + self._gsa_chrpos_map = None + self._dbsnp_151_37_reverse = None self._opensnp_datadump_filenames = [] def get_reference_sequences( @@ -256,11 +262,63 @@ def get_gsa_resources(self): ------- dict """ - if not self._gsa_resources: - self._gsa_resources = self._load_gsa_resources( - self._get_path_gsa_rsid_map(), self._get_path_gsa_chrpos_map() + + return { + "rsid_map": self.get_gsa_rsid(), + "chrpos_map": self.get_gsa_chrpos(), + "dbsnp_151_37_reverse": self.get_dbsnp_151_37_reverse(), + } + + def get_dbsnp_151_37_reverse(self): + """ Get and load RSIDs that are on the reference reverse (-) strand in dbSNP 151 and lower. + + Returns + ------- + pandas.DataFrame + + References + ---------- + 1. Sherry ST, Ward MH, Kholodov M, Baker J, Phan L, Smigielski EM, Sirotkin K. + dbSNP: the NCBI database of genetic variation. Nucleic Acids Res. 2001 Jan 1; + 29(1):308-11. + 2. Database of Single Nucleotide Polymorphisms (dbSNP). Bethesda (MD): National Center + for Biotechnology Information, National Library of Medicine. (dbSNP Build ID: 151). + Available from: http://www.ncbi.nlm.nih.gov/SNP/ + """ + if self._dbsnp_151_37_reverse is None: + # download the file from the cloud, if not done already + dbsnp_rev_path = self._download_file( + "https://sano-public.s3.eu-west-2.amazonaws.com/dbsnp151.b37.snps_reverse.txt.gz", + "dbsnp_151_37_reverse.txt.gz", + ) + + # load into pandas + rsids = pd.read_csv( + dbsnp_rev_path, + sep=" ", + header=None, # dont infer header as there isn't one + names=( + "dbsnp151revrsid", + "dbsnp151freqa", + "dbsnp151freqt", + "dbsnp151freqc", + "dbsnp151freqg", + ), + dtype={ + "dbsnp151revrsid": "string", + "dbsnp151freqa": "double", + "dbsnp151freqt": "double", + "dbsnp151freqc": "double", + "dbsnp151freqg": "double", + }, + engine="c", # force c engine for performance + comment="#", # skip the first row ) - return self._gsa_resources + + # store in memory so we don't have to load again + self._dbsnp_151_37_reverse = rsids + + return self._dbsnp_151_37_reverse def get_opensnp_datadump_filenames(self): """ Get filenames internal to the `openSNP <https://opensnp.org>`_ datadump zip. @@ -571,38 +629,66 @@ def _download_assembly_mapping_data( # remove temp file os.remove(f_tmp.name) - def _load_gsa_resources(self, rsid_map, chrpos_map): - d = {} - - with gzip.open(rsid_map, "rb") as f: - gsa_rsid_map = f.read().decode("utf-8") + def get_gsa_rsid(self): + """ Get and load GSA RSID map. - d["rsid_map"] = dict( - (x.split("\t")[0].strip(), x.split("\t")[1].strip()) - for x in gsa_rsid_map.split("\n")[:-1] - ) + https://support.illumina.com/downloads/infinium-global-screening-array-v2-0-product-files.html - with gzip.open(chrpos_map, "rb") as f: - gsa_chrpos_map = f.read().decode("utf-8") + Returns + ------- + pandas.DataFrame + """ + if self._gsa_rsid_map is None: + # download the file from the cloud, if not done already + rsid_path = self._download_file( + "https://sano-public.s3.eu-west-2.amazonaws.com/gsa_rsid_map.txt.gz", + "gsa_rsid_map.txt.gz", + ) + # load into pandas + rsids = pd.read_csv( + rsid_path, + sep=r"\s+", # whitespace separators + header=0, # dont infer header as there isn't one + names=("gsaname_rsid", "gsarsid"), + dtype={"gsaname_rsid": "string", "gsarsid": "string"}, + engine="c", # force c engine for performance + ) + self._gsa_rsid_map = rsids + return self._gsa_rsid_map - d["chrpos_map"] = dict( - (x.split("\t")[0], x.split("\t")[1] + ":" + x.split("\t")[2]) - for x in gsa_chrpos_map.split("\n")[:-1] - ) + def get_gsa_chrpos(self): + """ Get and load GSA chromosome position map. - return d + https://support.illumina.com/downloads/infinium-global-screening-array-v2-0-product-files.html - def _get_path_gsa_rsid_map(self): - return self._download_file( - "https://sano-public.s3.eu-west-2.amazonaws.com/gsa_rsid_map.txt.gz", - "gsa_rsid_map.txt.gz", - ) + Returns + ------- + pandas.DataFrame + """ + if self._gsa_chrpos_map is None: + # download the file from the cloud, if not done already + chrpos_path = self._download_file( + "https://sano-public.s3.eu-west-2.amazonaws.com/gsa_chrpos_map.txt.gz", + "gsa_chrpos_map.txt.gz", + ) + # load into pandas + + chrpos = pd.read_csv( + chrpos_path, + sep=r"\s+", # whitespace separators + header=0, # dont infer header as there isn't one + names=("gsaname_chrpos", "gsachr", "gsapos", "gsacm"), + dtype={ + "gsaname_chrpos": "string", + "gsachr": "category", + "gsapos": "uint32", + "gsacm": "double", + }, + engine="c", # force c engine for performance + ) - def _get_path_gsa_chrpos_map(self): - return self._download_file( - "https://sano-public.s3.eu-west-2.amazonaws.com/gsa_chrpos_map.txt.gz", - "gsa_chrpos_map.txt.gz", - ) + self._gsa_chrpos_map = chrpos + return self._gsa_chrpos_map def _get_path_opensnp_datadump(self): return self._download_file( @@ -645,7 +731,9 @@ def _download_file(self, url, filename, compress=False, timeout=30): # http://stackoverflow.com/a/7244263 with urllib.request.urlopen( url, timeout=timeout - ) as response, atomic_write(destination, mode="wb") as f: + ) as response, atomic_write( + destination, mode="wb", overwrite=True + ) as f: self._print_download_msg(destination) data = response.read() # a `bytes` object @@ -667,6 +755,11 @@ def _download_file(self, url, filename, compress=False, timeout=30): except socket.timeout: logger.warning(f"Timeout downloading {url}") destination = "" + except FileExistsError: + # if the file exists, another process has created it while it was + # being downloaded + # in such a case, the other copy is identical, so ignore this error + pass return destination diff --git a/src/snps/snps.py b/src/snps/snps.py index bddcc43e..ab68670d 100644 --- a/src/snps/snps.py +++ b/src/snps/snps.py @@ -84,11 +84,11 @@ def __init__( name / path of resources directory deduplicate : bool deduplicate RSIDs and make SNPs available as `SNPs.duplicate` - deduplicate_XY_chrom : bool - deduplicate alleles in the non-PAR regions of X and Y for males; see - `SNPs.discrepant_XY` deduplicate_MT_chrom : bool deduplicate alleles on MT; see `SNPs.heterozygous_MT` + deduplicate_XY_chrom : bool or str + deduplicate alleles in the non-PAR regions of X and Y for males; see `SNPs.discrepant_XY` + if a `str` then this is the sex determination method to use X Y or XY parallelize : bool utilize multiprocessing to speedup calculations processes : int @@ -131,6 +131,8 @@ def __init__( d["source"].split(", ") if ", " in d["source"] else [d["source"]] ) self._phased = d["phased"] + self._build = d["build"] + self._build_detected = True if d["build"] else False if not self._snps.empty: self.sort() @@ -138,13 +140,11 @@ def __init__( if deduplicate: self._deduplicate_rsids() - # prefer to use SNP positions to detect build - self._build = self.detect_build() - self._build_detected = True if self._build else False - + # use build detected from `read` method or comments, if any + # otherwise use SNP positions to detect build if not self._build_detected: - # use build detected from `read` method or comments, if any - self._build = d["build"] + self._build = self.detect_build() + self._build_detected = True if self._build else False if not self._build: self._build = 37 # assume Build 37 / GRCh37 if not detected @@ -156,7 +156,9 @@ def __init__( self.sort() if deduplicate_XY_chrom: - if self.determine_sex() == "Male": + if ( + deduplicate_XY_chrom is True and self.determine_sex() == "Male" + ) or self.determine_sex(chrom=deduplicate_XY_chrom) == "Male": self._deduplicate_XY_chrom() if deduplicate_MT_chrom: @@ -1163,58 +1165,68 @@ def _remapper(self, task): pos_end = int(temp["pos"].describe()["max"]) for mapping in mappings["mappings"]: - # skip if mapping is outside of range of SNP positions - if ( - mapping["original"]["end"] < pos_start - or mapping["original"]["start"] > pos_end - ): - continue - orig_range_len = mapping["original"]["end"] - mapping["original"]["start"] - mapped_range_len = mapping["mapped"]["end"] - mapping["mapped"]["start"] + orig_start = mapping["original"]["start"] + orig_end = mapping["original"]["end"] + mapped_start = mapping["mapped"]["start"] + mapped_end = mapping["mapped"]["end"] orig_region = mapping["original"]["seq_region_name"] mapped_region = mapping["mapped"]["seq_region_name"] - if orig_region != mapped_region: - logger.warning("discrepant chroms") - continue - - if orig_range_len != mapped_range_len: - logger.warning( - "discrepant coords" - ) # observed when mapping NCBI36 -> GRCh38 + # skip if mapping is outside of range of SNP positions + if orig_end < pos_start or orig_start > pos_end: continue # find the SNPs that are being remapped for this mapping snp_indices = temp.loc[ ~temp["remapped"] - & (temp["pos"] >= mapping["original"]["start"]) - & (temp["pos"] <= mapping["original"]["end"]) + & (temp["pos"] >= orig_start) + & (temp["pos"] <= orig_end) ].index - if len(snp_indices) > 0: - # remap the SNPs - if mapping["mapped"]["strand"] == -1: - # flip and (optionally) complement since we're mapping to minus strand - diff_from_start = ( - temp.loc[snp_indices, "pos"] - mapping["original"]["start"] - ) - temp.loc[snp_indices, "pos"] = ( - mapping["mapped"]["end"] - diff_from_start - ) + # if there are no snp here, skip + if not len(snp_indices): + continue - if complement_bases: - temp.loc[snp_indices, "genotype"] = temp.loc[ - snp_indices, "genotype" - ].apply(self._complement_bases) - else: - # mapping is on same (plus) strand, so just remap based on offset - offset = mapping["mapped"]["start"] - mapping["original"]["start"] - temp.loc[snp_indices, "pos"] = temp["pos"] + offset + orig_range_len = orig_end - orig_start + mapped_range_len = mapped_end - mapped_start + + # if this would change chromosome, skip + # TODO allow within normal chromosomes + # TODO flatten patches + if orig_region != mapped_region: + logger.warning( + f"discrepant chroms for {len(snp_indices)} SNPs from {orig_region} to {mapped_region}" + ) + continue + + # if there is any stretching or squashing of the region + # observed when mapping NCBI36 -> GRCh38 + # TODO disallow skipping a version when remapping + if orig_range_len != mapped_range_len: + logger.warning( + f"discrepant coords for {len(snp_indices)} SNPs from {orig_region}:{orig_start}-{orig_end} to {mapped_region}:{mapped_start}-{mapped_end}" + ) + continue + + # remap the SNPs + if mapping["mapped"]["strand"] == -1: + # flip and (optionally) complement since we're mapping to minus strand + diff_from_start = temp.loc[snp_indices, "pos"] - orig_start + temp.loc[snp_indices, "pos"] = mapped_end - diff_from_start + + if complement_bases: + temp.loc[snp_indices, "genotype"] = temp.loc[ + snp_indices, "genotype" + ].apply(self._complement_bases) + else: + # mapping is on same (plus) strand, so just remap based on offset + offset = mapped_start - orig_start + temp.loc[snp_indices, "pos"] = temp["pos"] + offset - # mark these SNPs as remapped - temp.loc[snp_indices, "remapped"] = True + # mark these SNPs as remapped + temp.loc[snp_indices, "remapped"] = True return temp	more specific tellmegen detection We've had a few files wrongly identified as TellMeGen when they weren't, which means they failed to parse. This makes the TellMeGen detection more strict, which should mean these files fail to parse as a "unrecognised" type. more specific tellmegen detection We've had a few files wrongly identified as TellMeGen when they weren't, which means they failed to parse. This makes the TellMeGen detection more strict, which should mean these files fail to parse as a "unrecognised" type.		2021-06-27T04:59:07	[]	[]
tira-io/tira	tira-io__tira-646	ac75ad53d6a94230713616253ec468545bfdbe44	diff --git a/.devcontainer/devcontainer.json b/.devcontainer/devcontainer.json index 985c4822..1ec1e985 100644 --- a/.devcontainer/devcontainer.json +++ b/.devcontainer/devcontainer.json @@ -56,6 +56,8 @@ "customizations": { "vscode": { "settings": { + "editor.formatOnSave": true, + "json.format.keepLines": true, "livePreview.portNumber": 3080, "remote.autoForwardPorts": false, "launch": { @@ -93,12 +95,21 @@ "stopAll": true } ] - } + }, + "[python]": { + "editor.codeActionsOnSave": { "source.organizeImports": "explicit" } + }, + "python.formatting.provider": "black", + "isort.args": [ "--profile", "black" ] }, "extensions": [ "Vue.volar", "ms-vscode.live-server", - "ms-python.python" + "ms-python.python", + "ms-python.isort", + "ms-python.black-formatter", + "ms-python.flake8", + "ms-python.mypy-type-checker" ] } }, @@ -108,5 +119,5 @@ // Needed for parts of the python-client that build and run docker containers "privileged": true, - "mounts": ["source=/var/run/docker.sock,target=/var/run/docker.sock,type=bind"] + "mounts": [ "source=/var/run/docker.sock,target=/var/run/docker.sock,type=bind" ] } diff --git a/.github/workflows/linters.yml b/.github/workflows/linters.yml new file mode 100644 index 00000000..a9d31af7 --- /dev/null +++ b/.github/workflows/linters.yml @@ -0,0 +1,83 @@ +name: Linters + +on: + push: {} + workflow_dispatch: {} + +jobs: + mypy: + runs-on: ubuntu-latest + timeout-minutes: 15 + + steps: + - uses: actions/checkout@v4 + - name: Set up Python 3.9 + uses: actions/setup-python@v4 + with: + python-version: 3.9 + - name: Install mypy + run: pip3 install mypy + - name: Run mypy on backend + working-directory: ${{github.workspace}}/application + run: \| + mkdir .mypy_cache + mypy . --disallow-untyped-calls --explicit-package-bases --ignore-missing-imports --install-types --non-interactive --cache-dir=.mypy_cache/ + - name: Run mypy on python-client + working-directory: ${{github.workspace}}/python-client + run: \| + mkdir .mypy_cache + mypy . --disallow-untyped-calls --explicit-package-bases --ignore-missing-imports --install-types --non-interactive --cache-dir=.mypy_cache/ + + flake8: + runs-on: ubuntu-latest + timeout-minutes: 15 + steps: + - uses: actions/checkout@v4 + - name: Set up Python 3.9 + uses: actions/setup-python@v4 + with: + python-version: 3.9 + - name: Install flake8 + run: pip3 install flake8 + - name: Run flake8 on backend + working-directory: ${{github.workspace}}/application + run: flake8 . + - name: Run flake8 on python-client + working-directory: ${{github.workspace}}/python-client + run: flake8 . + + black: + runs-on: ubuntu-latest + timeout-minutes: 15 + steps: + - uses: actions/checkout@v4 + - name: Set up Python 3.9 + uses: actions/setup-python@v4 + with: + python-version: 3.9 + - name: Install black + run: pip3 install black + - name: Run black on backend + working-directory: ${{github.workspace}}/application + run: black --check . + - name: Run black on python-client + working-directory: ${{github.workspace}}/python-client + run: black --check . + + isort: + runs-on: ubuntu-latest + timeout-minutes: 15 + steps: + - uses: actions/checkout@v4 + - name: Set up Python 3.9 + uses: actions/setup-python@v4 + with: + python-version: 3.9 + - name: Install isort + run: pip3 install isort + - name: Run isort on backend + working-directory: ${{github.workspace}}/application + run: isort --check . + - name: Run isort on python-client + working-directory: ${{github.workspace}}/python-client + run: isort --check . \ No newline at end of file diff --git a/Dockerfile.dev b/Dockerfile.dev index 92cc778e..4d7143a1 100644 --- a/Dockerfile.dev +++ b/Dockerfile.dev @@ -38,8 +38,11 @@ RUN apt-get update && apt-get install -y npm yarn ######################################################################################################################## ENV PIP_BREAK_SYSTEM_PACKAGES 1 USER root -RUN apt-get update && apt-get install -y python3 python3-pip python3-dev pkg-config default-libmysqlclient-dev \ +RUN <<EOF +apt-get update && apt-get install -y python3 python3-pip python3-dev pkg-config default-libmysqlclient-dev \ libpcre3-dev +pip3 install black flake8 isort mypy +EOF COPY application/requirements.txt /tmp/application/requirements.txt RUN <<EOF # Create a dummy secret diff --git a/README.md b/README.md index d3e2f3ad..a1b42c73 100644 --- a/README.md +++ b/README.md @@ -1,7 +1,7 @@ <h1 align="center">TIRA</h1> <p align="center"> - <img src="https://github.com/tira-io/tira-branding/raw/master/tira-icons/logo-tira-120x120-transparent.png"> + <img width=120px src="https://assets.tira.io/tira-icons/tira-logo.svg"> <!--h3>Integrated Research Architecture</h3--> <br/> <br/> @@ -14,6 +14,12 @@ <a href="https://tira.io"> <img alt="Deployment" src="https://img.shields.io/badge/dynamic/json?url=https%3A%2F%2Fwww.tira.io%2Fapi%2Fv1%2F&query=%24.version&prefix=v.&label=tira.io"/> </a> + <a href="https://github.com/tira-io/tira/actions/workflows/run-all-tests.yml"> + <img alt="Tests" src="https://github.com/tira-io/tira/actions/workflows/run-all-tests.yml/badge.svg"/> + </a> + <a href="https://github.com/tira-io/tira/actions/workflows/linters.yml"> + <img alt="Linters" src="https://github.com/tira-io/tira/actions/workflows/linters.yml/badge.svg"/> + </a> <br> <a href="https://tira-io.github.io/tira/">Documentation</a>  \|  <a href="./application/">Backend</a>  \|  diff --git a/application/pyproject.toml b/application/pyproject.toml new file mode 100644 index 00000000..a5adcfac --- /dev/null +++ b/application/pyproject.toml @@ -0,0 +1,8 @@ +[tool.black] +line-length = 120 + +[tool.isort] +profile = "black" +multi_line_output = 3 +line_length = 120 +include_trailing_comma = true diff --git a/application/requirements.txt b/application/requirements.txt index ab27b754..29e27b2c 100644 --- a/application/requirements.txt +++ b/application/requirements.txt @@ -1,3 +1,4 @@ +setuptools==71 grpcio>=1.53.2 # grpcio-tools==1.36.1 # still needed? protobuf<4.0dev diff --git a/application/setup.cfg b/application/setup.cfg new file mode 100644 index 00000000..8332de34 --- /dev/null +++ b/application/setup.cfg @@ -0,0 +1,3 @@ +[flake8] +max-line-length = 120 +extend-ignore = E203 diff --git a/python-client/pyproject.toml b/python-client/pyproject.toml index e34796ec..a5adcfac 100644 --- a/python-client/pyproject.toml +++ b/python-client/pyproject.toml @@ -1,2 +1,8 @@ [tool.black] -line-length = 120 \ No newline at end of file +line-length = 120 + +[tool.isort] +profile = "black" +multi_line_output = 3 +line_length = 120 +include_trailing_comma = true diff --git a/python-client/setup.cfg b/python-client/setup.cfg index ad5c0c57..7a6763d4 100644 --- a/python-client/setup.cfg +++ b/python-client/setup.cfg @@ -56,6 +56,3 @@ tira.static_redirects = *.json [flake8] max-line-length = 120 extend-ignore = E203 - -[isort] -profile=black	Disentangle the Frontend and the Backend		2024-05-28T07:10:02	[]	[]
tira-io/tira	tira-io__tira-632	3f5fe190b6fc2851fd3da1bdd5ac639b7379b63c	diff --git a/application/Makefile b/application/Makefile index 775fdfaba..ced022a55 100644 --- a/application/Makefile +++ b/application/Makefile @@ -1,7 +1,7 @@ .PHONY: help setup run-develop build-docker clean -VERSION_APPLICATION=0.0.103 -VERSION_GRPC=0.0.103 +VERSION_APPLICATION=0.0.104 +VERSION_GRPC=0.0.104 .DEFAULT: help help: diff --git a/application/src/django_admin/settings.py b/application/src/django_admin/settings.py index ba6d51988..e7661fccf 100644 --- a/application/src/django_admin/settings.py +++ b/application/src/django_admin/settings.py @@ -350,12 +350,14 @@ def logger_config(log_dir: Path): 'webpage-classification': 'OpenWebSearch/irixys23-tira-submission-template', 'valueeval-2024-human-value-detection': 'touche-webis-de/valueeval24-tira-software-submission-template', 'workshop-on-open-web-search': 'tira-io/wows24-submission-template', + 'nlpbuw-fsu-sose-24': 'webis-de/natural-language-processing-exercises', } REFERENCE_DATASETS = { 'ir-lab-padua-2024': 'ir-lab-padua-2024/longeval-tiny-train-20240315-training', 'ir-benchmarks': 'ir-benchmarks/cranfield-20230107-training', 'workshop-on-open-web-search': 'workshop-on-open-web-search/retrieval-20231027-training', + 'generative-ai-authorship-verification-panclef-2024': 'generative-ai-authorship-verification-panclef-2024/pan24-generative-authorship-tiny-smoke-20240417-training', } CODE_SUBMISSION_REPOSITORY_NAMESPACE = 'tira-io' diff --git a/application/src/tira/frontend-vuetify/src/components/EditTask.vue b/application/src/tira/frontend-vuetify/src/components/EditTask.vue index 8ed2eea41..63499430a 100644 --- a/application/src/tira/frontend-vuetify/src/components/EditTask.vue +++ b/application/src/tira/frontend-vuetify/src/components/EditTask.vue @@ -33,13 +33,13 @@ <p>A very small dataset (e.g., of 5 input instances) that can be used that participants get very fast feedback that their software is compatible with the data format for the task. This is intended to be fully public so that participants can use the smoke test dataset to test their software on their machine</p> <v-divider/> <h3 class="my-1">Component 2 (optional): An Evaluator</h3> - <p>A docker image that uses the ground-truth data and the outputs of some software as input to test if the outputs are valid and to measure their effectiveness. We already have many evaluators for many different scenarious available. Please look <a href="https://github.com/tira-io/tira/wiki/Organizing-Tasks#create-a-docker-image-for-your-evaluator">at the documentation</a> for existing evaluators and a step-by-step guide to build a new evaluator.</p> + <p>A docker image that uses the ground-truth data and the outputs of some software as input to test if the outputs are valid and to measure their effectiveness. We already have many evaluators for many different scenarious available. Please look <a href="https://tira-io.github.io/tira/organizers/organizing-tasks.html#create-a-docker-image-for-your-evaluator">at the documentation</a> for existing evaluators and a step-by-step guide to build a new evaluator.</p> <v-divider/> <h3 class="my-1">Component 3 (optional): Datasets</h3> <p>You can upload as many datasets (consisting of the inputs for software submissions and the ground truth for evaluation) as you like in whatever format you want. TIRA distinguishes between public datasets for which participants can see the outputs to verify that their software submission is correct and test datasets for which all outputs and evaluations are blinded and not visible to participants.</p> <v-divider/> <h3 class="my-1">Component 4 (optional): Baselines</h3> - <p>In the best case, you provide the code, a published docker image, and instructions on how to compile the code into a docker image to simplify participation in your shared tasks. We have some examples on baselines that you can adopt for your shared task <a href="https://github.com/tira-io/tira/wiki/Organizing-Tasks#provide-public-baselines-to-simplify-participation">in the documentation</a>.</p> + <p>In the best case, you provide the code, a published docker image, and instructions on how to compile the code into a docker image to simplify participation in your shared tasks. We have some examples on baselines that you can adopt for your shared task <a href="https://tira-io.github.io/tira/organizers/organizing-tasks.html#provide-public-baselines-to-simplify-participation">in the documentation</a>.</p> </v-card-text> </v-window-item> diff --git a/application/src/tira/frontend-vuetify/src/components/SoftwareDetails.vue b/application/src/tira/frontend-vuetify/src/components/SoftwareDetails.vue index 21ddf8a45..e88d31efc 100644 --- a/application/src/tira/frontend-vuetify/src/components/SoftwareDetails.vue +++ b/application/src/tira/frontend-vuetify/src/components/SoftwareDetails.vue @@ -141,7 +141,7 @@ export default { get('/task/' + this.task_id + '/vm/' + this.run.vm_id + '/run_details/' + this.run.run_id) .then(inject_response(this, {'loading': false})) .catch(() => {this.details_not_visible = true; this.loading = false}) - }, + }, updateTab() { this.component_tab = this.selectedComponentTab } diff --git a/application/src/tira/frontend-vuetify/src/submission-components/CodeSubmission.vue b/application/src/tira/frontend-vuetify/src/submission-components/CodeSubmission.vue index 549e24974..7377dbc07 100644 --- a/application/src/tira/frontend-vuetify/src/submission-components/CodeSubmission.vue +++ b/application/src/tira/frontend-vuetify/src/submission-components/CodeSubmission.vue @@ -3,7 +3,41 @@ <v-card-item v-if="loading"><loading :loading="loading"/></v-card-item> <v-card-item v-if="!loading && disabled"> <v-card-text> - Code submissions are not enabled for this task. Please contact your organizers <a :href="organizer_link">{{organizer}}</a> to enable code submissions. + You can submit to TIRA from your source code repository via GitHub actions that dockerize your code, test the docker image, and upload it to TIRA. This way, you don't need to install Docker on your machine, and our prepared Github action sends the metadata (i.e., the commit hash, the branch, the repository, etc.) to TIRA to ensure that the submission can be linked to the code that produced it. Please do not hesitate to contact your organizers <a :href="organizer_link">{{organizer}}</a> or the <a href="https://www.tira.io/categories">forum</a> in case of questions or problems. + + <br><br> + You can setup a software submission in three steps: + <br> + + <v-stepper :items="['Setup Your Git Repository', 'Add the Github Action', 'Submit']" flat :border=false> + <template v-slot:item.1> + <v-card title="" flat> + <h3>Create your Git repository</h3> + Please create your git repository with your code. The repository can be private or public. + <br><br> + <h3>Add the TIRA_CLIENT_TOKEN secret to your Git repository</h3> + In your repository, go to "Settings" -> "Secrets and variables" -> "Actions" and add a new repository secret with the name TIRA_CLIENT_TOKEN and the value: + <br><br> + <code-snippet title="TIRA_CLIENT_TOKEN" :code="token" expand_message=""/> + </v-card> + </template> + + <template v-slot:item.2> + <v-card title="" flat> + Please download the Github action <a :href="'/data-download/git-repo-template/' + user_id + '/' + task_id + '.zip'" target="_blank">here</a>. + <br> + This zip directory contains the github action in the file ".github/workflows/upload-software-to-tira.yml" that you can add to your git repository. The rest of the zip directory shows exemplary how to connect the code of your repository to your repository, see the README.md in the zip for more details. + </v-card> + </template> + + <template v-slot:item.3> + <v-card title="" flat> + After you have added the Github action (e.g., ".github/workflows/upload-software-to-tira.yml") to your Github repository, please run it by clicking on "Actions" -> "Upload Software to TIRA" -> "Run workflow". + <br><br> + After successful completion of the Github action, your new submission appears under the "Docker" tab above where you can select and run your new submission. + </v-card> + </template> + </v-stepper> </v-card-text> </v-card-item> <v-card-item v-if="!loading && !disabled"> @@ -83,7 +117,8 @@ export default { ssh_repo_url: '', owner_url: '', http_owner_url: '', - disabled: false + disabled: false, + token: '<ADD-YOUR-TOKEN-HERE>' } }, methods: { @@ -106,6 +141,9 @@ export default { get(`/api/get_software_submission_git_repository/${this.task_id}/${this.user_id}`) .then(inject_response(this, {'loading': false})) .catch(reportError("Problem loading the data of the task.", "This might be a short-term hiccup, please try again. We got the following error: ")) + + get('/api/token/' + this.user_id) + .then(inject_response(this)) }, computed: { organizer_link() {return get_link_to_organizer(this.organizer_id)}, diff --git a/application/src/tira/frontend-vuetify/src/submission-components/DockerSubmission.vue b/application/src/tira/frontend-vuetify/src/submission-components/DockerSubmission.vue index 6f6e951a6..d9ecd4019 100644 --- a/application/src/tira/frontend-vuetify/src/submission-components/DockerSubmission.vue +++ b/application/src/tira/frontend-vuetify/src/submission-components/DockerSubmission.vue @@ -114,7 +114,7 @@ export default { .then(inject_response(this, {'loading': false}, true)) .catch(reportError("Problem While Loading the Docker Details of the Task " + this.task_id, "This might be a short-term hiccup, please try again. We got the following error: ")) this.load_re_ranking_datasets() - this.step === '' ? this.tab = docker.docker_softwares[0].display_name : this.tab = "newDockerImage" + this.step === '' ? this.tab = this.docker.docker_softwares[0].display_name : this.tab = "newDockerImage" }, watch: { step(old_value, new_value) { diff --git a/application/src/tira/frontend-vuetify/src/submission-components/UploadSubmission.vue b/application/src/tira/frontend-vuetify/src/submission-components/UploadSubmission.vue index b1d483461..86bfc53fb 100644 --- a/application/src/tira/frontend-vuetify/src/submission-components/UploadSubmission.vue +++ b/application/src/tira/frontend-vuetify/src/submission-components/UploadSubmission.vue @@ -33,34 +33,87 @@ </v-col> </v-row> <v-window v-model="tab" v-if="!loading && role !== 'guest'" :touch="{left: null, right: null}"> - <v-window-item value="newUploadGroup"> - <h2>Create Run Upload Group</h2> - <p>Please click on "Add Upload Group" below to create a new run upload group.</p> - <p>Please use one upload group (you can edit the metadata of an upload group later) per approach. I.e., in TIRA, you can usually run software submissions on different datasets. For manually uploaded runs, we employ the same methodology: Please create one run upload group per approach, so that you can upload "executions" of the same approach on different datasets while maintaining the documentation.</p> - <br> + <v-window-item value="newUploadGroup"> + <h2>New Upload</h2> + <v-stepper :items="['About', 'Specify Metadata', 'Upload']" v-model="stepperModel" hide-actions="true" flat :border=false> + <template v-slot:item.1> + <v-card title="Specify what you want to upload" flat> + <v-radio-group v-model="upload_configuration"> + <v-radio label="I want to upload runs" value="upload-config-1"/> + <v-radio label="I want to upload a model from Hugging Face" value="upload-config-2"/> + </v-radio-group> + </v-card> + </template> - <v-form> - <v-radio-group v-model="upload_type"> - <v-radio label="Rename all uploaded files as configured by the organizers" value="upload-1"/> - <v-radio label="I want to configure the name of my uploaded file manually (Only for expert users, e.g., to normalize all uploaded files of this group.)" value="upload-2"/> - </v-radio-group> + <template v-slot:item.2> + <v-card flat> + <div v-if="upload_configuration === 'upload-config-1'"> + In TIRA, approaches are often evaluated on multiple datasets. Please ensure that you correctly group runs created by the same approach. <span v-if="userUploadGroups.length == 0">You have no upload groups yet.</span> + <span v-if="userUploadGroups.length > 0">You already have uploads for the following approaches: + <v-list density="compact"> + <v-list-item v-for="i in userUploadGroups"> + <v-list-item-title>Upload run(s) for approach <a href="javascript:void(0);" @click="stepperModel=1; tab=i.id">{{i.display_name}}</a></v-list-item-title> + </v-list-item> + </v-list> + </span> - <v-text-field v-if="upload_type === 'upload-2'" v-model="rename_file_to" label="Rename a file after upload to this name " /> - </v-form> + <div> + <v-checkbox label="I want to upload runs for a new approach" v-model="new_approach"></v-checkbox> - <v-btn variant="outlined" @click="addUpload()"> - Add Upload Group - </v-btn> + <div v-if="new_approach"> + <p>You will create a new run upload group. Please ensure that you only create only one run upload group per approach, so that you can upload "executions" of the same approach on different datasets while maintaining the documentation.</p> + <v-form> + <v-radio-group v-model="upload_type"> + <v-radio label="Rename all uploaded files as configured by the organizers" value="upload-1"/> + <v-radio label="I want to configure the name of my uploaded file manually (Only for expert users, e.g., to normalize all uploaded files of this group.)" value="upload-2"/> + </v-radio-group> - <div v-if="!showImportSubmission"> - <br> - If you participate in multiple tasks, you can also <a href="javascript:void(0);" @click="showImportSubmission=true">import your upload group</a>. + <v-text-field v-if="upload_type === 'upload-2'" v-model="rename_file_to" label="Rename a file after upload to this name " /> + </v-form> + + <p>Please click on "Add Upload Group" to create a new run upload group (i.e., if the run that you want to upload does not belong to the upload groups listed above).</p> + + <v-btn variant="outlined" @click="addUpload()">Add Upload Group</v-btn> + + <div v-if="!showImportSubmission"> + <br> + If you participate in multiple tasks, you can also <a href="javascript:void(0);" @click="showImportSubmission=true">import your upload group</a>. + </div> + <div v-if="showImportSubmission"> + <br> + <import-submission submission_type="upload"/> + </div> + </div> </div> - <div v-if="showImportSubmission"> - <br> - <import-submission submission_type="upload"/> + </div> + + <div v-if="upload_configuration === 'upload-config-2'"> + Please specify the model that will be downloaded via huggingface-cli + <v-text-field v-model="hugging_face_model" label="Hugging Face Model " /> + <div v-if="hugging_face_model"> + The following command will be used to download your model: + + <code-snippet title="TIRA will download and mount the model via" :code="'huggingface-cli download ' + hugging_face_model" expand_message=""/> + You can later mount the downloaded model into your software submission. Please click next to upload/import this Hugging Face model to TIRA. </div> + </div> + </v-card> + </template> + + <template v-slot:item.3> + <v-card title="Step Three" flat> + The Huggingface upload is disabled for this task. Please <a :href="contact_organizer">contact</a> the organizer <a :href="link_organizer"> {{ organizer }}</a> to enable this. + </v-card> + </template> + + + <v-stepper-actions @click:prev="stepperModel=1" @click:next="nextStep" :disabled='disableUploadStepper'></v-stepper-actions> + + </v-stepper> + <br> + + </v-window-item> <v-window-item v-for="us in this.all_uploadgroups" :value="us.id"> <loading :loading="description === 'no-description'"/> @@ -80,26 +133,29 @@ </div> <v-form v-if="description !== 'no-description'"> - <v-file-input v-model="fileHandle" - :rules="[v => !!v \|\| 'File is required']" - label="Click to add run file" - ></v-file-input> - <v-autocomplete label="Input Dataset" - :items="datasets" - item-title="display_name" - item-value="dataset_id" - prepend-icon="mdi-file-document-multiple-outline" - v-model="selectedDataset" - variant="underlined" - clearable/> - <v-text-field v-if="'' + rename_to !== 'null' && '' + rename_to !== '' && '' + rename_to !== 'undefined'" v-model="rename_to" label="Uploaded Files are renamed to (immutable for reproducibility)" disabled="true"/> + <h2>Upload new Submissions for Approach {{us.display_name}}</h2> + <v-radio-group v-model="upload_type_next_upload"> + <v-radio :label="'I want to upload a new run for approach ' + us.display_name + ' via the web UI'" value="upload-via-ui"/> + <v-radio :label="'I want to batch upload runs for approach ' + us.display_name + ' via a script'" value="upload-via-script"/> + </v-radio-group> + <div v-if="upload_type_next_upload == 'upload-via-ui'"> + <v-file-input v-model="fileHandle" :rules="[v => !!v \|\| 'File is required']" label="Click to add run file" /> + <v-autocomplete label="Input Dataset" :items="datasets" item-title="display_name" item-value="dataset_id" prepend-icon="mdi-file-document-multiple-outline" v-model="selectedDataset" variant="underlined" clearable/> + <v-text-field v-if="'' + rename_to !== 'null' && '' + rename_to !== '' && '' + rename_to !== 'undefined'" v-model="rename_to" label="Uploaded Files are renamed to (immutable for reproducibility)" disabled="true"/> + </div> </v-form> - <v-btn v-if="description !== 'no-description'" color="primary" :loading="uploading" :disabled="uploading \|\| fileHandle === null \|\| selectedDataset === ''" + <v-btn v-if="description !== 'no-description' && upload_type_next_upload == 'upload-via-ui'" color="primary" :loading="uploading" :disabled="uploading \|\| fileHandle === null \|\| selectedDataset === ''" @click="fileUpload(us.id)">Upload Run</v-btn> + + <div v-if="upload_type_next_upload == 'upload-via-script'"> + <code-snippet title="(1) Make sure that your TIRA client is up to date and authenticated" :code="tira_setup_code" expand_message="(1) Make sure that your TIRA client is up to date and authenticated"/> + + <code-snippet title="Batch upload runs via python" :code="batch_upload_code(us.display_name)" expand_message=""/> + </div> - <h2>Submissions</h2> + <h2>Your Existing Submissions for Approach {{us.display_name}}</h2> <run-list :task_id="task_id" :organizer="organizer" :organizer_id="organizer_id" :vm_id="user_id_for_task" :upload_id="us.id" ref="upload-run-list" /> </v-window-item> </v-window> @@ -108,14 +164,15 @@ <script> import { VAutocomplete } from 'vuetify/components' -import { extractTaskFromCurrentUrl, extractUserFromCurrentUrl, get, inject_response, reportError, extractRole, post_file, reportSuccess, handleModifiedSubmission } from "@/utils"; +import { extractTaskFromCurrentUrl, extractUserFromCurrentUrl, get, inject_response, reportError, extractRole, post_file, reportSuccess, handleModifiedSubmission, get_link_to_organizer, get_contact_link_to_organizer } from "@/utils"; import { Loading, LoginToSubmit, RunList } from "@/components"; import EditSubmissionDetails from "@/submission-components/EditSubmissionDetails.vue"; import ImportSubmission from "./ImportSubmission.vue"; +import CodeSnippet from "../components/CodeSnippet.vue" export default { name: "upload-submission", - components: {EditSubmissionDetails, Loading, VAutocomplete, LoginToSubmit, RunList, ImportSubmission}, + components: {EditSubmissionDetails, Loading, VAutocomplete, LoginToSubmit, RunList, ImportSubmission, CodeSnippet}, props: ['organizer', 'organizer_id'], emits: ['refresh_running_submissions'], data () { @@ -130,7 +187,12 @@ export default { uploadDataset: '', uploadFormError: '', upload_type: 'upload-1', + new_approach: false, + upload_configuration: '', + upload_type_next_upload: '', rename_file_to: '', + hugging_face_model: '', + stepperModel: '', fileHandle: null, description: 'no-description', rename_to: '', @@ -138,6 +200,7 @@ export default { all_uploadgroups: [{"id": null, "display_name": 'loading...'}], selectedDataset: '', showImportSubmission: false, + token: 'YOUR-TOKEN-HERE', datasets: [{"dataset_id": "loading...", "display_name": "loading...",}] } }, @@ -150,6 +213,32 @@ export default { } return ret.concat(this.all_uploadgroups) + }, + userUploadGroups() { + let ret = [] + + for (let i of this.all_uploadgroups) { + if (i.display_name != 'default upload') { + ret.push(i) + } + } + + return ret + }, + link_organizer() {return get_link_to_organizer(this.organizer_id);}, + contact_organizer() {return get_contact_link_to_organizer(this.organizer_id);}, + tira_setup_code() { + return 'pip3 uninstall -y tira\npip3 install tira\ntira-cli login --token ' + this.token + }, + disableUploadStepper() { + if(this.stepperModel == '1' && !this.upload_configuration) { + return 'next'; + } + if (this.stepperModel == '2' && (this.upload_configuration == 'upload-config-1' \|\| !this.hugging_face_model)) { + return 'next' + } + + return false; } }, methods: { @@ -172,6 +261,26 @@ export default { this.description = editedDetails.description handleModifiedSubmission(editedDetails, this.all_uploadgroups) }, + batch_upload_code(display_name) { + return 'from tira.rest_api_client import Client\n' + + 'from pathlib import Path\n' + + 'from tqdm import tqdm\n\n' + + 'tira = Client()\n' + + 'approach = \'' + this.task_id + '/' + this.user_id_for_task + '/' + display_name + '\'\n' + + 'dataset_ids = [' + this.datasets.map((i) => '\'' + i.dataset_id + '\'') + ']\n\n' + + 'for dataset_id in tqdm(dataset_ids):\n' + + ' # assume run to upload is located in a file dataset_id/run\n' + + ' run_file = Path(dataset_id) / \'run\'\n' + + ' tira.upload_run(approach=approach, dataset_id=dataset_id, file_path=run_file);\n' + }, + nextStep() { + if (this.stepperModel == 1) { + this.stepperModel = 2; + } + else if (this.stepperModel == 2 && this.upload_configuration === 'upload-config-2') { + this.stepperModel = 3; + } + }, deleteUpload(id_to_delete) { get(`/task/${this.task_id}/vm/${this.user_id_for_task}/upload-delete/${this.tab}`) .then(message => { @@ -204,6 +313,10 @@ export default { get('/api/submissions-for-task/' + this.task_id + '/' + this.user_id_for_task + '/upload') .then(inject_response(this, {'loading': false}, true)) .catch(reportError("Problem While Loading The Submissions of the Task " + this.task_id, "This might be a short-term hiccup, please try again. We got the following error: ")) + + get('/api/token/' + this.user_id_for_task) + .then(inject_response(this)) + this.tab = this.all_uploadgroups[0].display_name }, watch: { diff --git a/application/src/tira/management/commands/archive_runs_to_zenodo.py b/application/src/tira/management/commands/archive_runs_to_zenodo.py index 5eb8bcc6b..965a40cc5 100644 --- a/application/src/tira/management/commands/archive_runs_to_zenodo.py +++ b/application/src/tira/management/commands/archive_runs_to_zenodo.py @@ -33,11 +33,14 @@ def handle(self, args, options): ], 'trec-core': [ 'wapo-v2-trec-core-2018-20230107-training', 'disks45-nocr-trec8-20230209-training', 'disks45-nocr-trec7-20230209-training', 'disks45-nocr-trec-robust-2004-20230209-training' + ], + 'ir-lab': [ + 'anthology-20240411-training' ] } #we publish document processors only for fully public datasets, query processors can be published on all groups - fully_public_datasets = dataset_groups['trec-recent'] + dataset_groups['tiny-test-collections'] + dataset_groups['trec-medical'] + dataset_groups['clef-labs'] + fully_public_datasets = dataset_groups['trec-recent'] + dataset_groups['tiny-test-collections'] + dataset_groups['trec-medical'] + dataset_groups['clef-labs'] + dataset_groups['ir-lab'] systems = { 'ir-benchmarks': { @@ -47,6 +50,24 @@ def handle(self, args, *options): 'seanmacavaney': { 'DocT5Query': 'doc-t5-query', 'corpus-graph': 'corpus-graph', + }, + 'ows': { + 'pyterrier-anceindex': 'pyterrier-anceindex' + }, + 'ir-lab-sose-2024': { + 'tira-ir-starter': { + 'Index (tira-ir-starter-pyterrier)': 'ir-lab-sose-2024', + 'Index (pyterrier-stanford-lemmatizer)': 'ir-lab-sose-2024', + }, + 'seanmacavaney': { + 'DocT5Query': 'ir-lab-sose-2024', + 'corpus-graph': 'ir-lab-sose-2024', + }, + 'ows': { + 'pyterrier-anceindex': 'ir-lab-sose-2024' + }, + 'naverlabseurope': { + 'Splade (Index)': 'ir-lab-sose-2024' } }, } @@ -105,7 +126,7 @@ def handle(self, args, *options): ret[task_id][user_id][display_name] = {} output_dir = systems[task_id][user_id][display_name] for i in tqdm(fully_public_datasets): - run_id = model.runs(task_id, i, user_id, display_name)[0] + run_id = model.runs(task_id, i, user_id, display_name)[0]['run_id'] target_file = f'{output_dir}/{run_id}.zip' zip_file = zip_run(i, user_id, run_id) @@ -147,4 +168,4 @@ def handle(self, args, *options): print(json.dumps(ret)) - \ No newline at end of file + diff --git a/application/src/tira/templates/tira/git-repo-template/Dockerfile b/application/src/tira/templates/tira/git-repo-template/Dockerfile new file mode 100644 index 000000000..a947e0de5 --- /dev/null +++ b/application/src/tira/templates/tira/git-repo-template/Dockerfile @@ -0,0 +1,9 @@ +FROM python:3 + +ADD script.py /script.py +ADD requirements.txt /requirements.txt + +RUN pip3 install -r /requirements.txt + +ENTRYPOINT [ "python3", "/script.py", "-i", "$inputDataset", "-o", "$outputDir" ] + diff --git a/application/src/tira/templates/tira/git-repo-template/README.md b/application/src/tira/templates/tira/git-repo-template/README.md new file mode 100644 index 000000000..9397db9c8 --- /dev/null +++ b/application/src/tira/templates/tira/git-repo-template/README.md @@ -0,0 +1,27 @@ +# Code Submission Github Repository for Task {{ task_id }} + +This is a git repository template for code submissions to the shared task {{ task_id }} in TIRA. + +This repository explains how to configure a Github action for your code submission. +The corresponding Github action is [.github/workflows/upload-software-to-tira.yml](.github/workflows/upload-software-to-tira.yml). + +You must copy it to the .github/workflows of your own repository and adjust it to your code. + +The rest of this repository explains how the Github action works with a minimal example, so that you can test it on your machine but also directly by running the action in your own repository. + +The github action makes the following steps: + +(1) Install dependencies, i.e., install python, tira (`pip3 install tira`) and Docker. + +(2) Build the Docker image: (you might have to modify the Dockerfile, e.g., add your own code, install other dependencies, etc. + +``` +docker build -t {{ image }} -f Dockerfile . +``` + +(3) Test that the docker image works on a small example dataset and push it to TIRA: + +``` +tira-run --input-dataset {{ input_dataset }} --image {{ image }} +``` + diff --git a/application/src/tira/templates/tira/git-repo-template/github-action.yml b/application/src/tira/templates/tira/git-repo-template/github-action.yml new file mode 100644 index 000000000..f689b6d41 --- /dev/null +++ b/application/src/tira/templates/tira/git-repo-template/github-action.yml @@ -0,0 +1,25 @@ +name: Upload Software to TIRA +on: workflow_dispatch +jobs: + docker-build: + runs-on: ubuntu-latest + timeout-minutes: 45 + steps: + - name: Checkout + uses: actions/checkout@v3 + - name: Set up QEMU + uses: docker/setup-qemu-action@v2 + - name: Set up Docker Buildx + uses: docker/setup-buildx-action@v2 + - name: Set up Dependencies + uses: actions/setup-python@v4 + with: + python-version: 3.8 + - name: Install required python packages + run: \| + python -m pip install --upgrade pip + pip3 install tira + - name: Build, test, and upload image + run: \| + docker build -t {{ image }} -f {% verbatim %}${{ github.workspace }}{% endverbatim %}/Dockerfile {% verbatim %}${{ github.workspace }}{% endverbatim %} + tira-run --tira-client-token {% verbatim %}${{ secrets.TIRA_CLIENT_TOKEN }}{% endverbatim %} --input-dataset {{ input_dataset }} --image {{ image }} --push true --fail-if-output-is-empty diff --git a/application/src/tira/templates/tira/git-repo-template/script.py b/application/src/tira/templates/tira/git-repo-template/script.py new file mode 100644 index 000000000..af43c5d21 --- /dev/null +++ b/application/src/tira/templates/tira/git-repo-template/script.py @@ -0,0 +1,16 @@ +import argparse + +def parse_args(): + parser = argparse.ArgumentParser(prog='script') + parser.add_argument('-i', '--input', required=True, help='the input to the script.') + parser.add_argument('-o', '--output', required=True, help='the output of the script.') + + return parser.parse_args() + +if __name__ == '__main__': + args = parse_args() + + print(f'This is a demo, I ignore the passed input {args.input} and write some content into the output file {args.output}.') + with open(args.output + '/predictions.jsonl', 'w') as f: + f.write('hello world') + print('Done. I wrote "hello world" to {args.output}/predictions.jsonl.') diff --git a/application/src/tira/tira_model.py b/application/src/tira/tira_model.py index 41c17adec..aacf76dd1 100644 --- a/application/src/tira/tira_model.py +++ b/application/src/tira/tira_model.py @@ -129,12 +129,20 @@ def discourse_api_client(): def tira_run_command(image, command, task_id): - input_dataset = f'{task_id}/ADD-DATASET-ID-HERE' - if task_id in settings.REFERENCE_DATASETS: - input_dataset = settings.REFERENCE_DATASETS[task_id] + input_dataset = reference_dataset(task_id) return f'tira-run \\\n --input-dataset {input_dataset} \\\n --image {image} \\\n --command \'{command}\'' +def reference_dataset(task_id): + if task_id in settings.REFERENCE_DATASETS: + return settings.REFERENCE_DATASETS[task_id] + else: + available_datasets = get_datasets_by_task(task_id) + available_datasets = [i['dataset_id'] for i in available_datasets if i['dataset_id'].endswith('-training') and not i['is_confidential'] and not i['is_deprecated']] + if len(available_datasets) > 0: + return f'{task_id}/{available_datasets[0]}' + else: + return f'{task_id}/ADD-DATASET-ID-HERE' def tira_docker_registry_token(docker_software_help): ret = docker_software_help.split('docker login -u ')[1].split(' -p') diff --git a/application/src/tira/urls.py b/application/src/tira/urls.py index f4976ba1f..d4a9e8440 100644 --- a/application/src/tira/urls.py +++ b/application/src/tira/urls.py @@ -12,6 +12,7 @@ path('background_jobs/<str:task_id>/<str:job_id>', views.background_jobs, name='background_jobs'), path('task/<str:task_id>/user/<str:vm_id>/dataset/<str:dataset_id>/download/<str:run_id>.zip', views.download_rundir, name='download_rundir'), + path('data-download/git-repo-template/<str:vm_id>/<str:task_id>.zip', views.download_repo_template, name='download_repo_template'), path('data-download/<str:dataset_type>/<str:input_type>/<str:dataset_id>.zip', views.download_datadir, name='download_datadir'), re_path(r'^frontend-vuetify/.', views.veutify_page, name='vuetify_page'), diff --git a/application/src/tira/views.py b/application/src/tira/views.py index 7d3c106e9..5df2dd840 100644 --- a/application/src/tira/views.py +++ b/application/src/tira/views.py @@ -1,4 +1,5 @@ from django.shortcuts import render, redirect +from django.template.loader import render_to_string from django.http import JsonResponse, FileResponse from django.conf import settings from django.core.cache import cache @@ -17,6 +18,7 @@ import zipfile import json from http import HTTPStatus +import tempfile logger = logging.getLogger("tira") logger.info("Views: Logger active") @@ -188,6 +190,31 @@ def download_rundir(request, task_id, dataset_id, vm_id, run_id): def download_input_rundir(request, task_id, dataset_id, vm_id, run_id): return download_rundir(request, task_id, dataset_id, vm_id, input_run_id) +def download_repo_template(request, task_id, vm_id): + with tempfile.TemporaryDirectory() as tmpdirname: + directory = Path(tmpdirname) / f'git-repo-template-{task_id}' + os.makedirs(directory, exist_ok=True) + os.makedirs(directory / '.github' / 'workflows', exist_ok=True) + context = { + 'task_id': task_id, + 'image': f'registry.webis.de/code-research/tira/tira-user-{vm_id}/github-action-submission:0.0.1', + 'input_dataset': model.reference_dataset(task_id), + } + + with (directory / 'README.md').open('w') as readme, (directory / 'script.py').open('w') as script, (directory / 'requirements.txt').open('w') as requirements, (directory / 'Dockerfile').open('w') as dockerfile, (directory / '.github' / 'workflows' / 'upload-software-to-tira.yml').open('w') as ci: + readme.write(render_to_string('tira/git-repo-template/README.md', context=context)) + dockerfile.write(render_to_string('tira/git-repo-template/Dockerfile', context=context)) + requirements.write('argparse') + script.write(render_to_string('tira/git-repo-template/script.py', context=context)) + ci.write(render_to_string('tira/git-repo-template/github-action.yml', context=context)) + + zipped = Path(tmpdirname) / f'{task_id}.zip' + with zipfile.ZipFile(zipped, "w") as zipf: + for f in directory.rglob(''): + zipf.write(f, arcname=f.relative_to(directory)) + + return FileResponse(open(zipped, "rb"), as_attachment=True, filename=f"git-repo-template-{task_id}.zip") + @check_permissions def download_datadir(request, dataset_type, input_type, dataset_id): input_type = input_type.lower().replace('input', '') diff --git a/python-client/Makefile b/python-client/Makefile index 35e487907..6cc0ca7b6 100644 --- a/python-client/Makefile +++ b/python-client/Makefile @@ -3,7 +3,7 @@ build-pypi-package: run-tests python3 -m build --sdist . python3 -m build --wheel . - twine upload dist/tira-0.0.113-py3-none-any.whl dist/tira-0.0.113.tar.gz + twine upload dist/tira-0.0.125-py3-none-any.whl dist/tira-0.0.125.tar.gz run-tests: docker run -u root --rm -v /var/run/docker.sock:/var/run/docker.sock -v ${PWD}:/app -w /app --entrypoint pytest webis/tira:python-client-dev-0.0.5 diff --git a/python-client/tira/__init__.py b/python-client/tira/__init__.py index 15e9d7b5b..58f610684 100644 --- a/python-client/tira/__init__.py +++ b/python-client/tira/__init__.py @@ -1,1 +1,1 @@ -__version__ = "0.0.113" +__version__ = "0.0.125" diff --git a/python-client/tira/io_utils.py b/python-client/tira/io_utils.py index 3c60bd380..d647b4c17 100644 --- a/python-client/tira/io_utils.py +++ b/python-client/tira/io_utils.py @@ -140,6 +140,35 @@ def parse_prototext_key_values(file_name): yield ret +def to_prototext(m: List[Dict[str, Any]], upper_k: str = "") -> str: + ret = "" + + def _to_prototext(d: Dict[str, Any], upper_k: str = "") -> str: + ret = "" + for k, v in d.items(): + new_k = upper_k + if not new_k: + new_k = k + elif not new_k.endswith(k): + new_k = upper_k + "_" + k + if isinstance(v, dict): + ret += _to_prototext(v, upper_k=new_k) + else: + ret += ( + 'measure{\n key: "' + + str(new_k.replace("_", " ").title()) + + '"\n value: "' + + str(v) + + '"\n}\n' + ) + return ret + + for d in m: + ret += _to_prototext(d, upper_k=upper_k) + + return ret + + def all_environment_variables_for_github_action_or_fail(params): ret = {} diff --git a/python-client/tira/ir_datasets_util.py b/python-client/tira/ir_datasets_util.py index 1f189c49b..28eab76c9 100644 --- a/python-client/tira/ir_datasets_util.py +++ b/python-client/tira/ir_datasets_util.py @@ -31,6 +31,26 @@ def default_text(self): """ return self.title +class DictDocsstore(): + def __init__(self, docs): + self.docs = docs + + def __getitem__(self, item): + return self.docs[item] + + def get(self, item): + return self.docs.get(item) + + def __iter__(self): + return self.docs.__iter__() + + def __len__(self): + return len(self.docs) + + def get_many_iter(self, docids): + for docid in docids: + yield self.get(docid) + def register_dataset_from_re_rank_file(ir_dataset_id, df_re_rank, original_ir_datasets_id=None): """ @@ -80,7 +100,7 @@ def docs_count(self): return sum(1 for _ in self.stream_docs()) def docs_store(self): - return self.__parsed_docs() + return DictDocsstore(self.__parsed_docs()) def __parsed_docs(self): if self.docs is None: @@ -253,7 +273,9 @@ def load(self, dataset_id): docs = self.lazy_docs(lambda: get_download_dir_from_tira(dataset_id, False), None, True) queries = self.lazy_queries(lambda: get_download_dir_from_tira(dataset_id, True), None) qrels = self.lazy_qrels(lambda: get_download_dir_from_tira(dataset_id, True), None) - return Dataset(docs, queries, qrels) + ret = Dataset(docs, queries, qrels) + ret.dataset_id = lambda: dataset_id + return ret def topics_file(self, tira_path): return get_download_dir_from_tira(tira_path, True) + '/queries.xml' diff --git a/python-client/tira/license_agreements.py b/python-client/tira/license_agreements.py new file mode 100644 index 000000000..c97c14f1e --- /dev/null +++ b/python-client/tira/license_agreements.py @@ -0,0 +1,5 @@ +def print_license_agreement(url): + if not url: + return + if url and 'longeval' in url or 'leipzig-topics-small-20240119' in url or 'jena-topics-20231026-test' in url or 'jena-topics-small-20240119-training' in url or 'leipzig-topics-20231025-test' in url or 'training-20231104-training' in url or 'validation-20231104-training' in url: + print('The download is derived from The LongEval Dataset under the "Qwant LongEval Attribution-NonCommercial-ShareAlike License". Hence, the download is also under this License. By using it, you agree to the terms of this license. Please find details at: https://lindat.mff.cuni.cz/repository/xmlui/page/Qwant_LongEval_BY-NC-SA_License') \ No newline at end of file diff --git a/python-client/tira/local_execution_integration.py b/python-client/tira/local_execution_integration.py index a760347fc..e5576e186 100644 --- a/python-client/tira/local_execution_integration.py +++ b/python-client/tira/local_execution_integration.py @@ -91,7 +91,6 @@ def ensure_image_available_locally(self, image, client=None): def extract_entrypoint(self, image): image_name = image - from tira.third_party_integrations import extract_to_be_executed_notebook_from_command_or_none self.ensure_image_available_locally(image_name) image = self.__docker_client().images.get(image_name) ret = image.attrs['Config']['Entrypoint'] @@ -104,13 +103,16 @@ def extract_entrypoint(self, image): ret = json.loads(i) break - ret = ' '.join(ret) - executable = extract_to_be_executed_notebook_from_command_or_none(ret) + return self.make_command_absolute(image_name, ' '.join(ret)) + + def make_command_absolute(self, image_name, command): + from tira.third_party_integrations import extract_to_be_executed_notebook_from_command_or_none + executable = extract_to_be_executed_notebook_from_command_or_none(command) - if not executable or executable.startswith('/'): - return ret + if not executable or executable.startswith('/') or executable.startswith("'/") or executable.startswith('"/'): + return command else: - return ret.replace(executable, (self.docker_image_work_dir(image_name) + '/' + executable).replace('//', '/')) + return command.replace(executable, (self.docker_image_work_dir(image_name) + '/' + executable).replace('//', '/').replace('/./', '/')) def __docker_client(self): try: @@ -165,7 +167,7 @@ def login_docker_client(self, task_name, team_name, client=None): return True - def run(self, identifier=None, image=None, command=None, input_dir=None, output_dir=None, evaluate=False, dry_run=False, docker_software_id_to_output=None, software_id=None, allow_network=False, input_run=None, additional_volumes=None, eval_dir='tira-evaluation'): + def run(self, identifier=None, image=None, command=None, input_dir=None, output_dir=None, evaluate=False, dry_run=False, docker_software_id_to_output=None, software_id=None, allow_network=False, input_run=None, additional_volumes=None, eval_dir='tira-evaluation', gpu_count=0): previous_stages = [] original_args = {'identifier': identifier, 'image': image, 'command': command, 'input_dir': input_dir, 'output_dir': output_dir, 'evaluate': evaluate, 'dry_run': dry_run, 'docker_software_id_to_output': docker_software_id_to_output, 'software_id': software_id} s_id = 'unknown-software-id' @@ -228,7 +230,11 @@ def run(self, identifier=None, image=None, command=None, input_dir=None, output_ if input_run: environment['inputRun'] = '/tira-data/input-run' - container = client.containers.run(image, entrypoint='sh', command=f'-c "{command}; sleep .1"', environment=environment, volumes=volumes, detach=True, remove=True, network_disabled = not allow_network) + device_requests = [] + if gpu_count != 0: + device_requests=[docker.types.DeviceRequest(count=gpu_count, capabilities=[['gpu']])] + + container = client.containers.run(image, entrypoint='sh', command=f'-c "{command}; sleep .1"', environment=environment, volumes=volumes, detach=True, remove=True, network_disabled = not allow_network, device_requests=device_requests) for line in container.attach(stdout=True, stream=True, logs=True): print(line.decode('utf-8')) @@ -307,19 +313,27 @@ def export_submission_from_jupyter_notebook(self, notebook): return None ret = [] - + ret += ['TIRA_COMMAND=/workspace/run-pyterrier-notebook.py --input ${TIRA_INPUT_DIRECTORY} --output ${TIRA_OUTPUT_DIRECTORY} --notebook /workspace/' + notebook.split("/")[-1]] notebook_content = json.load(open(notebook, 'r')) return '\n'.join(ret) + def normalize_image_name(self, image, required_prefix): + if required_prefix and not image.startswith(required_prefix): + ret = (required_prefix + '/' + image[:10]).replace('//', '/') + ':' + (image.split(':')[-1] if ':' in image else 'latest') + return ret.replace('-:', ':').replace('::', ':').replace('/:', ':') + else: + return image + def push_image(self, image, required_prefix=None, task_name=None, team_name=None): client = self.__docker_client() if not self.docker_client_is_authenticated(client): self.login_docker_client(task_name, team_name, client) - + new_image = image if required_prefix and not image.startswith(required_prefix): - new_image = (required_prefix + '/' + image[:10]).replace('//', '/') + ':' + (image.split(':')[-1] if ':' in image else 'latest') + new_image = self.normalize_image_name(image, required_prefix) + print(f'I tag the image "{image}" as "{new_image}" for upload to TIRA (only internal).') client.images.get(image).tag(new_image) image = new_image diff --git a/python-client/tira/pandas_integration.py b/python-client/tira/pandas_integration.py index 5cc596528..da285607f 100644 --- a/python-client/tira/pandas_integration.py +++ b/python-client/tira/pandas_integration.py @@ -1,8 +1,24 @@ +from typing import List, Tuple +PANDAS_DTYPES = {'docno': str, 'doc_id': str, 'id': str, 'qid': str, 'query_id': str, 'queryid': str, 'docno': str, 'doc_id': str, 'docid': str} + class PandasIntegration(): + """Handling of inputs/outputs in TIRA with pandas. All methods here work in the TIRA sandbox without internet connection (when the data is mounted read-only). + """ def __init__(self, tira_client): self.tira_client = tira_client - def from_retriever_submission(self, approach, dataset, previous_stage=None, datasets=None): + def from_retriever_submission(self, approach: str, dataset: str, previous_stage: str=None, datasets: List[str]=None): + """Load a run file as pandas dataframe from tira. Compatible with PyTerrier. + + Args: + approach (str): the approach for which the run should be loaded, in the format 'task/team/software'. + dataset (str): the dataset id, either an tira or ir_datasets id. + previous_stage (str, optional): The previous stage, in case the approach itself is ambigious. Defaults to None. + datasets (List[str], optional): _description_. Defaults to None. + + Returns: + pd.DataFrame: The run file parsed to a pandas DataFrame. + """ import pandas as pd from tira.ir_datasets_util import translate_irds_id_to_tirex task, team, software = approach.split('/') @@ -42,7 +58,20 @@ def __matching_files(self, approach, dataset, file_selection): return sorted(list(ret)) - def transform_queries(self, approach, dataset, file_selection=('/.jsonl', '/.jsonl.gz')): + def transform_queries(self, approach:str, dataset:str, file_selection: Tuple[str,...]=('/.jsonl', '/.jsonl.gz')): + """Load and transform the query processing outputs specified by the approach on the dataset for direct re-use as a PyTerrier query transformation. + + Args: + approach str: the approach for which the run should be loaded, in the format 'task/team/software'. + dataset (str):the dataset id, either an tira or ir_datasets id. + file_selection (Tuple[str,...], optional): The search glob to outputs specified by the approach on the dataset. Defaults to ('/.jsonl', '/.jsonl.gz'). + + Raises: + ValueError: If no approach with the identifier 'approach' was found or if there was an error parsing the outputs. + + Returns: + pd.DataFrame: a DataFrame with the parsed query processing outputs compatible with PyTerrier query transformations. + """ import pandas as pd matching_files = self.__matching_files(approach, dataset, file_selection) @@ -57,6 +86,19 @@ def transform_queries(self, approach, dataset, file_selection=('/.jsonl', '/.j return ret def transform_documents(self, approach, dataset, file_selection=('/.jsonl', '/.jsonl.gz')): + """Load and transform the document processing outputs specified by the approach on the dataset for direct re-use as a PyTerrier document transformation. + + Args: + approach str: the approach for which the run should be loaded, in the format 'task/team/software'. + dataset (str):the dataset id, either an tira or ir_datasets id. + file_selection (Tuple[str,...], optional): The search glob to outputs specified by the approach on the dataset. Defaults to ('/.jsonl', '/.jsonl.gz'). + + Raises: + ValueError: If no approach with the identifier 'approach' was found or if there was an error parsing the outputs. + + Returns: + pd.DataFrame: a DataFrame with the parsed document processing outputs compatible with PyTerrier document transformations. + """ import pandas as pd matching_files = self.__matching_files(approach, dataset, file_selection) if len(matching_files) == 0: @@ -68,3 +110,60 @@ def transform_documents(self, approach, dataset, file_selection=('/.jsonl', '/* ret['docno'] = ret['doc_id'] del ret['doc_id'] return ret + + def __matching_dataset_files(self, task, dataset, truth_dataset, file_selection): + from glob import glob + ret = [] + local_dir = self.tira_client.download_dataset(task, dataset, truth_dataset) + + for glob_entry in file_selection: + for i in glob(local_dir + glob_entry): + ret += [i] + + return ret + + def inputs(self, task, dataset, file_selection=('/.jsonl', '/.jsonl.gz'), dtype=PANDAS_DTYPES): + """Load the inputs to systems for a task from tira. + + Args: + approach str: the approach for which the run should be loaded, in the format 'task/team/software'. + dataset (str):the dataset id, either an tira or ir_datasets id. + file_selection (Tuple[str,...], optional): The search glob to outputs specified by the approach on the dataset. Defaults to ('/.jsonl', '/.jsonl.gz'). + dtype (Tuple[str,...], optional): Transformations of the data types while loading with pandas. Defaults to PANDAS_DTYPES. + + Raises: + ValueError: If the dataset is not public or does not exist. + + Returns: + pd.DataFrame: A DataFrame with all inputs to systems. + """ + import pandas as pd + matching_files = self.__matching_dataset_files(task, dataset, False, file_selection) + + if len(matching_files) == 0: + raise ValueError('Could not find a dataset output. Used file_selection: ' + str(file_selection) + '. Please specify the file_selection to resolve this.') + + return pd.read_json(matching_files[0], lines=True, dtype=dtype) + + def truths(self, task, dataset, file_selection=('/.jsonl', '/.jsonl.gz'), dtype=PANDAS_DTYPES): + """Load the truths, i.e., ground truth labels, for a task from tira. + + Args: + approach str: the approach for which the run should be loaded, in the format 'task/team/software'. + dataset (str):the dataset id, either an tira or ir_datasets id. + file_selection (Tuple[str,...], optional): The search glob to outputs specified by the approach on the dataset. Defaults to ('/.jsonl', '/.jsonl.gz'). + dtype (Tuple[str,...], optional): Transformations of the data types while loading with pandas. Defaults to PANDAS_DTYPES. + + Raises: + ValueError: If the truth is not public or does not exist. + + Returns: + pd.DataFrame: A DataFrame with all the ground truth labels. + """ + import pandas as pd + matching_files = self.__matching_dataset_files(task, dataset, True, file_selection) + + if len(matching_files) == 0: + raise ValueError('Could not find a dataset output. Used file_selection: ' + str(file_selection) + '. Please specify the file_selection to resolve this.') + + return pd.read_json(matching_files[0], lines=True, dtype=dtype) diff --git a/python-client/tira/pyterrier_integration.py b/python-client/tira/pyterrier_integration.py index 4b495d76f..ef2686ed1 100644 --- a/python-client/tira/pyterrier_integration.py +++ b/python-client/tira/pyterrier_integration.py @@ -1,4 +1,6 @@ import os +from pathlib import Path + class PyTerrierIntegration(): def __init__(self, tira_client): @@ -99,7 +101,8 @@ def index(self, approach, dataset): Load an PyTerrier index from TIRA. """ import pyterrier as pt - ret = self.tira_client.get_run_output(approach, dataset) + '/index' + from tira.ir_datasets_util import translate_irds_id_to_tirex + ret = self.tira_client.get_run_output(approach, translate_irds_id_to_tirex(dataset)) + '/index' return pt.IndexFactory.of(os.path.abspath(ret)) def from_submission(self, approach, dataset=None, datasets=None): @@ -111,10 +114,9 @@ def from_submission(self, approach, dataset=None, datasets=None): return TiraRerankingTransformer(approach, self.tira_client, dataset, datasets) def from_retriever_submission(self, approach, dataset, previous_stage=None, datasets=None): - import pyterrier as pt + from tira.pyterrier_util import TiraSourceTransformer ret = self.pd.from_retriever_submission(approach, dataset, previous_stage, datasets) - - return pt.Transformer.from_df(ret) + return TiraSourceTransformer(ret) def transform_queries(self, approach, dataset, file_selection=('/.jsonl', '/.jsonl.gz'), prefix=''): from pyterrier.apply import generic @@ -143,7 +145,85 @@ def __transform_df(df): return generic(__transform_df) + def doc_features(self, approach, dataset, file_selection=('/.jsonl', '/.jsonl.gz')): + import numpy as np + from pyterrier.apply import doc_features + ret = self.pd.transform_documents(approach, dataset, file_selection) + cols = [i for i in ret.columns if i not in ['docno']] + ret = {str(i['docno']): np.array([i[c] for c in cols]) for _, i in ret.iterrows()} + + return doc_features(lambda row: ret[str(row['docno'])]) + + def query_features(self, approach, dataset, file_selection=('/.jsonl', '/.jsonl.gz')): + import numpy as np + from pyterrier.transformer import Transformer + ret = self.pd.transform_queries(approach, dataset, file_selection) + cols = [i for i in ret.columns if i not in ['qid']] + ret = {str(i['qid']): np.array([i[c] for c in cols]) for _, i in ret.iterrows()} + + class ApplyQueryFeatureTransformer(Transformer): + def __repr__(self): + return "tira.pt.query_features()" + + def transform(self, inputRes): + outputRes = inputRes.copy() + + outputRes["features"] = outputRes.apply(lambda i: ret[str(i['qid'])], axis=1) + return outputRes + + return ApplyQueryFeatureTransformer() + def reranker(self, approach, irds_id=None): from tira.pyterrier_util import TiraLocalExecutionRerankingTransformer return TiraLocalExecutionRerankingTransformer(approach, self.tira_client, irds_id=irds_id) +class PyTerrierAnceIntegration(): + """The pyterrier_ance integration to re-use cached ANCE indices. Wraps https://github.com/terrierteam/pyterrier_ance + """ + def __init__(self, tira_client): + self.tira_client = tira_client + + def ance_retrieval(self, dataset:str): + """Load a cached pyterrier_ance.ANCEIndexer submitted as workshop-on-open-web-search/ows/pyterrier-anceindex from tira. + + References (for citation): + https://arxiv.org/pdf/2007.00808.pdf + https://github.com/microsoft/ANCE/ + + Args: + dataset (str): the dataset id, either an tira or ir_datasets id. + + Returns: + pyterrier_ance.ANCERetrieval: the ANCE index. + """ + from tira.ir_datasets_util import translate_irds_id_to_tirex + ance_index = Path(self.tira_client.get_run_output('ir-lab-sose-2024/ows/pyterrier-anceindex', translate_irds_id_to_tirex(dataset))) / 'anceindex' + ance_checkpoint = self.tira_client.load_resource('Passage_ANCE_FirstP_Checkpoint.zip') + import pyterrier_ance + return pyterrier_ance.ANCERetrieval(ance_checkpoint, ance_index) + + +class PyTerrierSpladeIntegration(): + """The pyt_splade integration to re-use cached Splade indices. Wraps https://github.com/cmacdonald/pyt_splade + """ + def __init__(self, tira_client): + self.tira_client = tira_client + + def splade_index(self, dataset:str, approach: str='workshop-on-open-web-search/naverlabseurope/Splade (Index)'): + """Load a cached pyt_splade index submitted as the passed approach (default 'workshop-on-open-web-search/naverlabseurope/Splade (Index)') from tira. + + References (for citation): + https://github.com/naver/splade?tab=readme-ov-file#cite-scroll + ToDo: Ask Thibault what to cite. + + Args: + dataset (str): the dataset id, either an tira or ir_datasets id. + approach (str, optional): the approach id, defaults 'workshop-on-open-web-search/naverlabseurope/Splade (Index)'. + + Returns: + The PyTerrier index suitable for retrieval. + """ + from tira.ir_datasets_util import translate_irds_id_to_tirex + import pyterrier as pt + ret = Path(self.tira_client.get_run_output('ir-lab-sose-2024/naverlabseurope/Splade (Index)', translate_irds_id_to_tirex(dataset))) / 'spladeindex' + return pt.IndexFactory.of(os.path.abspath(ret)) diff --git a/python-client/tira/pyterrier_util.py b/python-client/tira/pyterrier_util.py index 03b512a90..fa46ca3bf 100644 --- a/python-client/tira/pyterrier_util.py +++ b/python-client/tira/pyterrier_util.py @@ -1,4 +1,4 @@ -from pyterrier.transformer import Transformer +from pyterrier.transformer import Transformer, SourceTransformer import json import pandas as pd import tempfile @@ -25,6 +25,29 @@ def merge_runs(topics, run_file): return topics[keeping].merge(df, how='left', left_on=join_on, right_on=join_on) +class TiraSourceTransformer(SourceTransformer): + def __init__(self, rtr, kwargs): + super().__init__(rtr, kwargs) + + + def transform(self, topics): + import numpy as np + if 'docno' not in topics.columns: + return super().transform(topics) + elif 'qid' not in topics.columns: + raise ValueError('The dataframe needs to have a column "qid"') + + keeping = topics.columns + + common_columns = np.intersect1d(topics.columns, self.df.columns) + + # we drop columns in topics that exist in the self.df + drop_columns = [i for i in common_columns if i not in ("qid", "docno")] + if len(drop_columns) > 0: + keeping = topics.columns[~ topics.columns.isin(drop_columns)] + + return topics[keeping].merge(self.df, on=["qid", "docno"]) + class TiraFullRankTransformer(Transformer): """ A Transformer that re-executes some full-rank approach submitted to a shared task in TIRA. @@ -69,6 +92,7 @@ def __init__(self, approach, tira_client, dataset=None, datasets=None, *kwargs) def transform(self, topics): import numpy as np assert "qid" in topics.columns + if 'tira_task' not in topics.columns or 'tira_dataset' not in topics.columns or 'tira_first_stage_run_id' not in topics.columns: if self.datasets: tira_configurations = [{'tira_task': self.task, 'tira_dataset': i, 'tira_first_stage_run_id': None} for i in self.datasets] diff --git a/python-client/tira/rest_api_client.py b/python-client/tira/rest_api_client.py index 9b87b2c6f..a822450ad 100644 --- a/python-client/tira/rest_api_client.py +++ b/python-client/tira/rest_api_client.py @@ -7,14 +7,14 @@ import io import time from random import randint -from tira.pyterrier_integration import PyTerrierIntegration +from tira.pyterrier_integration import PyTerrierIntegration, PyTerrierAnceIntegration, PyTerrierSpladeIntegration from tira.pandas_integration import PandasIntegration from tira.local_execution_integration import LocalExecutionIntegration import logging from .tira_client import TiraClient from typing import Optional, List, Dict, Union from functools import lru_cache -from tira.tira_redirects import redirects, mirror_url, dataset_ir_redirects +from tira.tira_redirects import redirects, mirror_url, dataset_ir_redirects, RESOURCE_REDIRECTS from tqdm import tqdm import hashlib @@ -38,6 +38,8 @@ def __init__(self, base_url: Optional[str]=None, api_key: str=None, failsave_ret self.fail_if_api_key_is_invalid() self.pd = PandasIntegration(self) self.pt = PyTerrierIntegration(self) + self.pt_ance = PyTerrierAnceIntegration(self) + self.pt_splade = PyTerrierSpladeIntegration(self) self.local_execution = LocalExecutionIntegration(self) self.failsave_retries = failsave_retries @@ -45,7 +47,12 @@ def __init__(self, base_url: Optional[str]=None, api_key: str=None, failsave_ret def load_settings(self): try: - return json.load(open(self.tira_cache_dir + '/.tira-settings.json', 'r')) + ret = json.load(open(self.tira_cache_dir + '/.tira-settings.json', 'r')) + if 'api_key' not in ret: + ret['api_key'] = 'no-api-key' + if 'api_user_name' not in ret: + ret['api_user_name'] = 'no-api-key-user' + return ret except Exception: logging.info(f'No settings given in {self.tira_cache_dir}/.tira-settings.json. I will use defaults.') return {'api_key': 'no-api-key', 'api_user_name': 'no-api-key-user'} @@ -53,6 +60,7 @@ def load_settings(self): def update_settings(self, k, v): settings = self.load_settings() settings[k] = v + os.makedirs(self.tira_cache_dir, exist_ok=True) json.dump(settings, open(Path(self.tira_cache_dir) / '.tira-settings.json', 'w+')) if k == 'api_key': @@ -61,10 +69,10 @@ def update_settings(self, k, v): def api_key_is_valid(self): role = self.json_response('/api/role') - if (self.api_user_name is None or self.api_user_name == 'no-api-key-user') and role and 'context' in role and 'user_id' in role['context']: + if (self.api_user_name is None or self.api_user_name == 'no-api-key-user') and role and 'context' in role and 'user_id' in role['context'] and role['context']['user_id']: self.api_user_name = role['context']['user_id'] - return role and 'status' in role and 'role' in role and 0 == role['status'] + return role and 'status' in role and 'role' in role and 0 == role['status'] and 'user_id' in role['context'] and role['context']['user_id'] and 'role' in role['context'] and role['context']['role'] and 'guest' != role['context']['role'] def fail_if_api_key_is_invalid(self): if not self.api_key_is_valid(): @@ -123,6 +131,8 @@ def add_docker_software(self, image, command, tira_vm_id, tira_task_id, code_rep ret = ret.content.decode('utf8') ret = json.loads(ret) assert ret['status'] == 0 + + print(f'Software with name {ret["context"]["display_name"]} was created.') logging.info(f'Software with name {ret["context"]["display_name"]} was created.') logging.info(f'Please visit {self.base_url}/submit/{tira_task_id}/user/{tira_vm_id}/docker-submission to run your software.') @@ -211,6 +221,29 @@ def evaluations(self, task, dataset, join_submissions=True): def run_was_already_executed_on_dataset(self, approach, dataset): return self.get_run_execution_or_none(approach, dataset) is not None + def load_resource(self, resource:str): + """Load a resource (usually a zip) from TIRA/Zenodo. Serves as utikity function in case some additional resources must be loaded. + + Args: + resource (str): The resource identifier + + Raises: + ValueError: If the resource is not known. + + Returns: + str: The path to the downloaded resource. + """ + target_file = f'{self.tira_cache_dir}/raw_resources/{resource}' + if os.path.exists(target_file): + return target_file + + if resource not in RESOURCE_REDIRECTS: + raise ValueError(f'Resource {resource} not supported.') + + os.makedirs(f'{self.tira_cache_dir}/raw_resources', exist_ok=True) + self.download_and_extract_zip(RESOURCE_REDIRECTS[resource], target_file, extract=False) + return target_file + def get_run_output(self, approach, dataset, allow_without_evaluation=False): """ Downloads the run (or uses the cached version) of the specified approach on the specified dataset. @@ -220,7 +253,10 @@ def get_run_output(self, approach, dataset, allow_without_evaluation=False): if mounted_output_in_sandbox: return mounted_output_in_sandbox - task, team, software = approach.split('/') + task, team, software = approach.split('/') + if '/' in dataset: + dataset = dataset.split('/')[-1] + run_execution = self.get_run_execution_or_none(approach, dataset) if run_execution: @@ -276,6 +312,8 @@ def get_run_execution_or_none(self, approach, dataset, previous_stage_run_id=Non return ret[['task', 'dataset', 'team', 'run_id']].iloc[0].to_dict() def download_run(self, task, dataset, software, team=None, previous_stage=None, return_metadata=False): + if '/' in dataset: + dataset = dataset.split('/')[-1] ret = self.get_run_execution_or_none(f'{task}/{team}/{software}', dataset, previous_stage) if not ret: raise ValueError(f'I could not find a run for the filter criteria task="{task}", dataset="{dataset}", software="{software}", team={team}, previous_stage={previous_stage}') @@ -308,6 +346,8 @@ def download_dataset(self, task, dataset, truth_dataset=False): """ if 'TIRA_INPUT_DATASET' in os.environ: return os.environ['TIRA_INPUT_DATASET'] + if '/' in dataset: + dataset = dataset.split('/')[-1] dataset = dataset_ir_redirects(dataset) @@ -383,7 +423,7 @@ def evaluate_run(self, team, dataset, run_id): return ret - def download_and_extract_zip(self, url, target_dir): + def download_and_extract_zip(self, url, target_dir, extract=True): url = redirects(url=url)['urls'][0] if url.split('://')[1].startswith('files.webis.de'): print(f'Download from the Incubator: {url}') @@ -413,10 +453,16 @@ def download_and_extract_zip(self, url, target_dir): for data in r.iter_content(chunk_size=1024): size = response_content.write(data) bar.update(size) - print('Download finished. Extract...') - z = zipfile.ZipFile(response_content) - z.extractall(target_dir) - print('Extraction finished: ', target_dir) + if extract: + print('Download finished. Extract...') + z = zipfile.ZipFile(response_content) + z.extractall(target_dir) + print('Extraction finished: ', target_dir) + else: + print('Download finished. Persist...') + with open(target_dir, 'wb') as file_out: + file_out.write(response_content.getbuffer()) + print('Download finished: ', target_dir) return except Exception as e: @@ -517,6 +563,10 @@ def upload_run(self, file_path: Path, dataset_id: str, approach: str=None, task_ self.fail_if_api_key_is_invalid() + if file_path is None or not file_path.is_file(): + logging.warn(f'The passed file {file_path} does not exist.') + raise ValueError(f'The passed file {file_path} does not exist.') + # TODO: check that task_id and vm_id don't contain illegal characters (e.g., '/') url = f"/task/{task_id}/vm/{vm_id}/upload/{dataset_id}/{upload_id}" logging.info(f"Submitting the runfile at {url}") diff --git a/python-client/tira/third_party_integrations.py b/python-client/tira/third_party_integrations.py index a1a074264..4f5fcf1a8 100644 --- a/python-client/tira/third_party_integrations.py +++ b/python-client/tira/third_party_integrations.py @@ -3,6 +3,7 @@ from tira.io_utils import all_lines_to_pandas import logging from pathlib import Path +import shlex def ensure_pyterrier_is_loaded(boot_packages=("com.github.terrierteam:terrier-prf:-SNAPSHOT", ), packages=(), patch_ir_datasets=True): @@ -150,13 +151,16 @@ def extract_to_be_executed_notebook_from_command_or_none(command:str): return command.split('--notebook')[1].strip().split(' ')[0].strip() if command is not None and '.py' in command: - for arg in command.split(' '): - if arg.endswith('.py'): + for arg in shlex.split(command, posix=False): + if arg.endswith('.py') or arg.endswith('.py"') or arg.endswith('.py\''): return arg if command is not None: - command = command.split(' ')[0] - if command.endswith('.sh') or command.endswith('.bash'): + command = shlex.split(command, posix=False) + if len(command) > 0: + command = command[0] + + if command and (command.endswith('.sh') or command.endswith('.bash') or command.endswith('.sh"') or command.endswith('.sh\'') or command.endswith('.bash"') or command.endswith('.bash\'')): return command return None @@ -263,6 +267,7 @@ def extract_previous_stages_from_docker_image(image:str, command:str = None): if notebook is None: return [] + notebook = shlex.split(notebook)[0] tira = Client() @@ -276,6 +281,11 @@ def extract_previous_stages_from_docker_image(image:str, command:str = None): return extract_previous_stages_from_notebook(Path(local_file)) def load_ir_datasets(): + try: + from ir_datasets.datasets.base import Dataset + except: + return None + # Detect if we are in the TIRA sandbox if 'TIRA_INPUT_DATASET' in os.environ: from tira.ir_datasets_util import static_ir_dataset diff --git a/python-client/tira/tira_cli.py b/python-client/tira/tira_cli.py index ebb3668df..a582866c7 100644 --- a/python-client/tira/tira_cli.py +++ b/python-client/tira/tira_cli.py @@ -23,6 +23,7 @@ def parse_args(): parser_login = subparsers.add_parser('login', help='Login your TIRA client to the tira server.') parser_login.add_argument('--token', required=True, default=None, help='The token to login to the server.') + parser_login.add_argument('--print-docker-auth', required=False, default=False, action='store_true', help='Print the docker credentials as TIRA_DOCKER_REGISTRY_USER and TIRA_DOCKER_REGISTRY_TOKEN to stdout.') args = parser.parse_args() @@ -45,3 +46,5 @@ def main(): print(client.download_dataset(None, args.dataset)) if args.command == 'login': client.login(args.token) + if args.print_docker_auth: + print(client.local_execution.docker_client_is_authenticated()) diff --git a/python-client/tira/tira_redirects.py b/python-client/tira/tira_redirects.py index 761c8a84d..cf0bd06bc 100644 --- a/python-client/tira/tira_redirects.py +++ b/python-client/tira/tira_redirects.py @@ -1,5 +1,4 @@ - - +from tira.license_agreements import print_license_agreement STATIC_REDIRECTS = { 'ir-benchmarks': { @@ -201,6 +200,66 @@ } } } + }, + + "ir-lab-sose-2024": { + "tira-ir-starter": { + "Index (tira-ir-starter-pyterrier)": { + "ir-acl-anthology-20240411-training": { + "run_id": "2024-04-11-19-43-23", + "md5": "ebb5b8f1d8c7ad36612f408da1203ff2", + "urls": ["https://files.webis.de/data-in-production/data-research/tira-zenodo-dump-preparation/ir-lab-sose2024/2024-04-11-19-43-23.zip"] + } + }, + "Index (pyterrier-stanford-lemmatizer)": { + "ir-acl-anthology-20240411-training": { + "run_id": "2024-04-16-11-05-06", + "md5": "70aa7ce8a437ba47ebf9e17d75871a5f", + "urls": ["https://files.webis.de/data-in-production/data-research/tira-zenodo-dump-preparation/ir-lab-sose2024/2024-04-16-11-05-06.zip"] + } + } + }, + "seanmacavaney": { + "DocT5Query": { + "anthology-20240411-training": { + "run_id": "2024-04-09-22-03-26", + "md5": "1a0f65a8f47051db435ec6031106db7b", + "urls": ["https://files.webis.de/data-in-production/data-research/tira-zenodo-dump-preparation/ir-lab-sose2024/2024-04-09-22-03-26.zip"], + } + }, + "corpus-graph": { + "anthology-20240411-training": { + "run_id": "2024-04-09-16-35-50", + "md5": "75f743914d44d7252425844136ea9722", + "urls": ["https://files.webis.de/data-in-production/data-research/tira-zenodo-dump-preparation/ir-lab-sose2024/2024-04-09-16-35-50.zip"], + } + }, + #"some-interface": { + #copy of corpus-graph + # "anthology-20240411-training": { + # "run_id": "2024-04-09-16-35-50", + # "md5": "75f743914d44d7252425844136ea9722" + # } + #} + }, + "ows": { + "pyterrier-anceindex": { + "ir-acl-anthology-20240411-training": { + "run_id": "2024-04-11-19-47-18", + "md5": "1112dc9e60b6dfa4ac80571ad3425200", + "urls": ["https://files.webis.de/data-in-production/data-research/tira-zenodo-dump-preparation/ir-lab-sose2024/2024-04-11-19-47-18.zip"], + } + } + }, + "naverlabseurope": { + "Splade (Index)": { + "ir-acl-anthology-20240411-training": { + "run_id": "2024-04-14-08-40-58", + "md5": "03f73b47664cc1c843529b948094bf81", + "urls": ["https://files.webis.de/data-in-production/data-research/tira-zenodo-dump-preparation/ir-lab-sose2024/2024-04-14-08-40-58.zip"], + } + } + } } } @@ -302,10 +361,20 @@ 'training-20231104-training': 'https://zenodo.org/records/10628882/files/', 'jena-topics-small-20240119-training': 'https://zenodo.org/records/10628882/files/', 'leipzig-topics-small-20240119-training': 'https://zenodo.org/records/10628882/files/', + 'ir-acl-anthology-20240411-training': 'https://files.webis.de/data-in-production/data-research/tira-zenodo-dump-preparation/ir-lab-sose2024/', + 'longeval-2023-06-20240418-training': 'https://files.webis.de/data-in-production/data-research/tira-zenodo-dump-preparation/ir-lab-padua2024/', + 'longeval-2023-08-20240418-training': 'https://files.webis.de/data-in-production/data-research/tira-zenodo-dump-preparation/ir-lab-padua2024/', +} + +RESOURCE_REDIRECTS = { + 'Passage_ANCE_FirstP_Checkpoint.zip': 'https://files.webis.de/data-in-production/data-research/tira-zenodo-dump-preparation/ir-lab-sose2024/Passage_ANCE_FirstP_Checkpoint.zip', + 'custom-terrier-token-processing-1.0-SNAPSHOT-jar-with-dependencies.jar': 'https://files.webis.de/data-in-production/data-research/tira-zenodo-dump-preparation/ir-lab-sose2024/custom-terrier-token-processing-1.0-SNAPSHOT-jar-with-dependencies.jar' } DATASET_ID_REDIRECTS = { - 'longeval-tiny-train-20240315-training': 'training-20231104-training' + 'longeval-tiny-train-20240315-training': 'training-20231104-training', + 'longeval-2023-06-20240422-training': 'longeval-2023-06-20240418-training', + 'longeval-2023-08-20240422-training': 'longeval-2023-08-20240418-training' } @@ -351,6 +420,8 @@ def mirror_url(url): def redirects(approach=None, dataset=None, url=None): default_ret = {'urls': [url]} if url is not None: + print_license_agreement(url) + if '/task/' in url and '/user/' in url and '/dataset/' in url and '/download/' in url and '.zip' in url: #/task/{task}/user/{team}/dataset/{dataset}/download/{run_id}.zip ret = url.split('/task/')[1] diff --git a/python-client/tira/tira_run.py b/python-client/tira/tira_run.py index 1353feb96..ab6002d40 100755 --- a/python-client/tira/tira_run.py +++ b/python-client/tira/tira_run.py @@ -51,7 +51,8 @@ def parse_args(): group.add_argument('--export-submission-from-jupyter-notebook', required=False, default=None, type=str) group.add_argument('--export-submission-environment', nargs='', required=False, default=None, type=str) parser.add_argument('--command', required=False) - parser.add_argument('--verbose', required=False, default=False, type=bool) + parser.add_argument('--verbose', required=False, default=False, action='store_true') + parser.add_argument('--gpus', required=False, default=False, type=str, help='GPU devices to add to the container ("all" or -1 to pass all GPUs, or the number of devices).') parser.add_argument('--evaluate', required=False, default=False, type=bool) parser.add_argument('--evaluation-directory', required=False, default=str(os.path.abspath("tira-evaluation"))) parser.add_argument('--dry-run', required=False, default=False, type=bool) @@ -98,6 +99,7 @@ def parse_args(): args.previous_stages = [] if not args.input_run else [args.input_run] if args.input_run is None and args.input_run_directory is None: args.input_run = extract_previous_stages_from_docker_image(args.image, args.command) + args.command = LocalExecutionIntegration().make_command_absolute(args.image, args.command) args.previous_stages = args.input_run if args.input_run and len(args.input_run) == 1: args.input_run = args.input_run[0] @@ -112,8 +114,17 @@ def parse_args(): if args.tira_client_token: if api_key_valid: args.tira_client_token = rest_client.api_key + elif 'TIRA_CLIENT_TOKEN' in os.environ and os.environ.get('TIRA_CLIENT_TOKEN'): + rest_client = RestClient(api_key=os.environ.get('TIRA_CLIENT_TOKEN')) + api_key_valid = rest_client.api_key_is_valid() + if api_key_valid: + args.tira_client_token = rest_client.api_key + else: + parser.error('The option --tira-client-token (or environment variable TIRA_CLIENT_TOKEN) is required when --push is active.') else: - parser.error('The option --tira-client-token (or environment variable TIRA_CLIENT_TOKEN) is required when --push is active.') + rest_client = RestClient(api_key=args.tira_client_token) + if not rest_client.api_key_is_valid(): + parser.error('The option --tira-client-token (or environment variable TIRA_CLIENT_TOKEN) is required when --push is active.') if not args.tira_task_id and args.input_dataset: args.tira_task_id = args.input_dataset.split('/')[0] @@ -253,7 +264,11 @@ def main(args=None): # command={args.command} ############################################################################################################################ ''') - client.local_execution.run(identifier=args.approach, image=args.image, command=args.command, input_dir=input_dir, output_dir=args.output_directory, dry_run=args.dry_run, allow_network=args.allow_network, input_run=args.input_run, additional_volumes=args.v, evaluate=evaluate, eval_dir=args.evaluation_directory) + gpus = 0 + if args.gpus: + gpus = -1 if args.gpus == 'all' else int(args.gpus) + + client.local_execution.run(identifier=args.approach, image=args.image, command=args.command, input_dir=input_dir, output_dir=args.output_directory, dry_run=args.dry_run, allow_network=args.allow_network, input_run=args.input_run, additional_volumes=args.v, evaluate=evaluate, eval_dir=args.evaluation_directory, gpu_count=gpus) if args.fail_if_output_is_empty and (not os.path.exists(args.output_directory) or not os.listdir(args.output_directory)): raise ValueError('The software produced an empty output directory, it likely failed?') @@ -271,11 +286,12 @@ def main(args=None): for f in os.listdir(args.output_directory): print(f' - {f}: {os.path.getsize(args.output_directory + "/" + f)} bytes') - continue_user = input("Are the outputs correct and should I push the software to TIRA? (y/N) ").lower() + if not args.fail_if_output_is_empty: + continue_user = input("Are the outputs correct and should I push the software to TIRA? (y/N) ").lower() - if continue_user and continue_user.lower() not in ['y', 'yes']: - print('You did not specify yes, I will not push the software.') - return + if continue_user and continue_user.lower() not in ['y', 'yes']: + print('You did not specify yes, I will not push the software.') + return registry_prefix = client.docker_registry() + '/code-research/tira/tira-user-' + args.tira_vm_id + '/' print('Push Docker image')	No flag for --gpu in tira-run for local tests Current behavior Expected behavior		2024-04-22T22:35:02	[]	[]
tira-io/tira	tira-io__tira-584	3971c30c3d3aaa741e01387d5ced757754bc71c4	diff --git a/.devcontainer.json b/.devcontainer.json new file mode 100644 index 000000000..38b68dece --- /dev/null +++ b/.devcontainer.json @@ -0,0 +1,8 @@ +{ + "image": "webis/tira-application:basis-0.0.95", + "customizations": { + "vscode": { + "extensions": ["ms-python.python", "ms-python.vscode-pylance", "ms-toolsai.jupyter"] + } + } +} diff --git a/README.md b/README.md index 535654164..c1d1bbb73 100644 --- a/README.md +++ b/README.md @@ -12,7 +12,7 @@ Components: ## Setup Your Local Development Environment -We use [dev containers](https://code.visualstudio.com/docs/devcontainers/containers) to simplify development. Please install Docker and an IDE with support for dev containers on your machine (we usually use VS Code). +We use [dev containers](https://code.visualstudio.com/docs/devcontainers/containers) to simplify development. You can [open this repository directly in codespaces](https://github.com/codespaces/new/webis-de/tira-io/tree/tira). If you want to develop locally, please install Docker and an IDE with support for dev containers on your machine (we usually use VS Code). First, please clone the repository: ``` @@ -21,7 +21,7 @@ git clone [email protected]:tira-io/tira.git Please open the directory `application` in VS Code, and confirm to use the provided dev container. -If you want to work on production data, please ensure that you can login to ssh.webis.de, and then do the following: +If you want to work on production data (not necessary in most cases, you usually can skip this step), please ensure that you can login to ssh.webis.de, and then do the following: ``` make import-data-from-dump diff --git a/application/.devcontainer.json b/application/.devcontainer.json index b290a47f2..38b68dece 100644 --- a/application/.devcontainer.json +++ b/application/.devcontainer.json @@ -1,5 +1,5 @@ { - "image": "webis/tira-application:basis-0.0.94", + "image": "webis/tira-application:basis-0.0.95", "customizations": { "vscode": { "extensions": ["ms-python.python", "ms-python.vscode-pylance", "ms-toolsai.jupyter"] diff --git a/application/Dockerfile.application b/application/Dockerfile.application index e83bce614..c4e859015 100644 --- a/application/Dockerfile.application +++ b/application/Dockerfile.application @@ -9,7 +9,7 @@ RUN cd /src/tira/frontend-vuetify \ # Only change in case of new / updated dependencies -FROM webis/tira-application:basis-0.0.94 +FROM webis/tira-application:basis-0.0.95 # This Dockerfile ensures that all dependencies do rarely change by starting from a basis image # that contains already all dependencies (so that the minor versions do rarely change, but we diff --git a/application/Dockerfile.application-dev b/application/Dockerfile.application-dev index 9b1916730..f7850169b 100644 --- a/application/Dockerfile.application-dev +++ b/application/Dockerfile.application-dev @@ -1,4 +1,4 @@ -# docker build -t webis/tira-application:basis-0.0.94 -f Dockerfile.application-dev . +# docker build -t webis/tira-application:basis-0.0.95 -f Dockerfile.application-dev . FROM ubuntu:22.04 RUN apt-get update \ @@ -53,3 +53,24 @@ RUN chown -R tira:tira /usr/local && \ RUN pip3 install tira==0.0.97 ir-measures==0.3.1 +RUN apt-get install -y wget curl + +RUN wget 'https://nodejs.org/dist/v20.10.0/node-v20.10.0-linux-x64.tar.xz' \ + && tar -xf node-v20.10.0-linux-x64.tar.xz \ + && rm node-v20.10.0-linux-x64.tar.xz \ + && mv node-v20.10.0-linux-x64/bin/* /usr/local/bin/ \ + && mv node-v20.10.0-linux-x64/lib/node_modules/ /usr/local/lib/ \ + && npm install --global yarn + +ADD src/tira/frontend-vuetify/package.json src/tira/frontend-vuetify/yarn.lock src/tira/frontend-vuetify/vite.config.ts src/tira/frontend-vuetify/jest.config.js src/tira/frontend-vuetify/babel.config.js /tmp-del/ + +RUN cd /tmp-del \ + && yarn create vuetify \ + && yarn install \ + && yarn install --dev \ + && mv node_modules/* /usr/local/lib/node_modules \ + && mv node_modules/.bin /usr/local/lib/node_modules/.bin \ + && mv node_modules/.yarn-integrity /usr/local/lib/node_modules/.yarn-integrity \ + && cd / \ + && rm -R /tmp-del/ + diff --git a/application/Makefile b/application/Makefile index cbdff5cf5..13972707f 100644 --- a/application/Makefile +++ b/application/Makefile @@ -2,8 +2,8 @@ VENV_NAME?=venv PYTHON=${PWD}/${VENV_NAME}/bin/python3 -VERSION_APPLICATION=0.0.94 -VERSION_GRPC=0.0.94 +VERSION_APPLICATION=0.0.96 +VERSION_GRPC=0.0.96 .DEFAULT: help help: @@ -71,10 +71,18 @@ run-develop: && python src/manage.py run_develop vite-build: - docker run -v ${PWD}:/app --platform linux/amd64 --rm -ti -w /app/src/tira/frontend-vuetify --entrypoint yarn webis/tira:vuetify-dev-0.0.1 build + @cd src/tira/frontend-vuetify \ + && yarn build vite-build-light: - docker run -v ${PWD}:/app --platform linux/amd64 --rm -ti -w /app/src/tira/frontend-vuetify --entrypoint yarn webis/tira:vuetify-dev-0.0.1 build-light + @cd src/tira/frontend-vuetify \ + && yarn build-light + +vite-build-docker: + docker run -v ${PWD}:/app --platform linux/amd64 --rm -ti -w /app/src/tira/frontend-vuetify --entrypoint yarn webis/tira-application:basis-0.0.95 build + +vite-build-light-docker: + docker run -v ${PWD}:/app --platform linux/amd64 --rm -ti -w /app/src/tira/frontend-vuetify --entrypoint yarn webis/tira-application:basis-0.0.95 build-light run-docker: diff --git a/application/src/django_admin/settings.py b/application/src/django_admin/settings.py index 4fa6107a6..f88e7bd0f 100644 --- a/application/src/django_admin/settings.py +++ b/application/src/django_admin/settings.py @@ -339,7 +339,11 @@ def logger_config(log_dir: Path): DISCOURSE_API_URL = 'https://www.tira.io' -CODE_SUBMISSION_REFERENCE_REPOSITORY = 'mam10eks/tira-software-submission-template' +CODE_SUBMISSION_REFERENCE_REPOSITORIES = { + 'ir-lab-jena-leipzig-wise-2023': 'tira-io/tira-ir-lab-wise-submission-template', + 'webpage-classification': 'OpenWebSearch/irixys23-tira-submission-template', + 'valueeval-2024-human-value-detection': 'touche-webis-de/valueeval24-tira-software-submission-template', +} CODE_SUBMISSION_REPOSITORY_NAMESPACE = 'tira-io' try: DISRAPTOR_API_KEY = open(DISRAPTOR_SECRET_FILE, "r").read().strip() diff --git a/application/src/tira/data/HybridDatabase.py b/application/src/tira/data/HybridDatabase.py index 3f37d8c5e..992884a72 100644 --- a/application/src/tira/data/HybridDatabase.py +++ b/application/src/tira/data/HybridDatabase.py @@ -19,6 +19,7 @@ logger = logging.getLogger("tira_db") +#SELECT tira_dockersoftware.vm_id, COUNT(DISTINCT(tira_dockersoftware.docker_software_id)) AS 'Software Count', SUM(tira_run.run_id IS NOT NULL) AS 'Executed Runs' FROM tira_dockersoftware LEFT JOIN tira_run ON tira_dockersoftware.docker_software_id = tira_run.docker_software_id where tira_dockersoftware.task_id = 'webpage-classification' GROUP BY tira_dockersoftware.vm_id; class HybridDatabase(object): """ @@ -1048,10 +1049,19 @@ def __link_to_code(build_environment): except: return None - if 'TIRA_JUPYTER_NOTEBOOK' not in build_environment or 'GITHUB_REPOSITORY' not in build_environment or 'GITHUB_WORKFLOW' not in build_environment or 'GITHUB_SHA' not in build_environment: + if 'GITHUB_REPOSITORY' not in build_environment or 'GITHUB_WORKFLOW' not in build_environment or 'GITHUB_SHA' not in build_environment: return None - if build_environment['GITHUB_WORKFLOW'] == ".github/workflows/upload-notebook-submission.yml": + if build_environment['GITHUB_WORKFLOW'] == "Upload Docker Software to TIRA": + if 'TIRA_DOCKER_PATH' not in build_environment: + return None + + return f'https://github.com/{build_environment["GITHUB_REPOSITORY"]}/tree/{build_environment["GITHUB_SHA"]}/{build_environment["TIRA_DOCKER_PATH"]}' + + if build_environment['GITHUB_WORKFLOW'] == ".github/workflows/upload-notebook-submission.yml" or build_environment['GITHUB_WORKFLOW'] == 'Upload Notebook to TIRA': + if 'TIRA_JUPYTER_NOTEBOOK' not in build_environment: + return None + return f'https://github.com/{build_environment["GITHUB_REPOSITORY"]}/tree/{build_environment["GITHUB_SHA"]}/jupyter-notebook-submissions/{build_environment["TIRA_JUPYTER_NOTEBOOK"]}' return None diff --git a/application/src/tira/endpoints/data_api.py b/application/src/tira/endpoints/data_api.py index fa4892bbe..c8fdf6d89 100644 --- a/application/src/tira/endpoints/data_api.py +++ b/application/src/tira/endpoints/data_api.py @@ -439,6 +439,34 @@ def tirex_components(request, context): context['tirex_components'] = settings.TIREX_COMPONENTS return JsonResponse({'status': 0, 'context': context}) +def get_snippet_to_run_components(request): + all_components = settings.TIREX_COMPONENTS + component_ids = request.GET.get('components', 'false') + + # All links with display_name == "Submission in TIREx" have the ID of the component in their link, its a bit ugly, but at the moment we need to extract the ID from there. + # E.g., the ID from the URL "/submissions/ir-benchmarks/ows/query-segmentation-hyp-a" would be ir-benchmarks/ows/query-segmentation-hyp-a + + # Also Ugly: we need to determine which type of processor (query processor, document processor, etc) something is by using its top-level category, e.g., "Query Processing". + + # I think it makes sense to build a small method that uses the settings.TIREX_COMPONENTS as input and produces a mapping form component ID (the thing below "Submission in TIREx") to the properties, e.g., query processor true or false, etc. + + # I think we can hard code everything against ROBUST04, we can switch this later. + dataset_initialization = 'dataset = pt.get_dataset("irds:disks45/nocr/trec-robust-2004")\n' + + additional_variables = '' + + # If we have a query processor, we need to add an additional variable "topics" + # just for this hard coded example: + current_component_is_query_processor = True + if current_component_is_query_processor: + additional_variables += "topics = dataset.get_topics(variant='title')\n" + + component_definitions = "tira.pt.transform_queries('ir-benchmarks/ows/query-segmentation-hyb-i', dataset)\n" + + snippet = (dataset_initialization + additional_variables + component_definitions).strip() + + return JsonResponse({'status': 0, 'context': {'snippet': snippet}}) + @add_context def reranking_datasets(request, context, task_id): diff --git a/application/src/tira/endpoints/vm_api.py b/application/src/tira/endpoints/vm_api.py index 2243e372b..157749a10 100644 --- a/application/src/tira/endpoints/vm_api.py +++ b/application/src/tira/endpoints/vm_api.py @@ -521,6 +521,7 @@ def add_software_submission_git_repository(request, task_id, vm_id): try: data = json.loads(request.body) external_owner = data['external_owner'] + private = not data.get('allow_public_repo', False) disraptor_user = get_disraptor_user(request, allow_unauthenticated_user=False) if not disraptor_user or not type(disraptor_user) == str: @@ -529,7 +530,8 @@ def add_software_submission_git_repository(request, task_id, vm_id): if not model.github_user_exists(external_owner): return JsonResponse({'status': 1, 'message': f"The user '{external_owner}' does not exist on Github, maybe a typo?"}) - software_submission_git_repo = model.get_submission_git_repo(vm_id, task_id, disraptor_user, external_owner) + software_submission_git_repo = model.get_submission_git_repo(vm_id, task_id, disraptor_user, external_owner, + private) return JsonResponse({'status': 0, "context": software_submission_git_repo}) except Exception as e: @@ -541,7 +543,7 @@ def add_software_submission_git_repository(request, task_id, vm_id): @check_permissions def get_software_submission_git_repository(request, task_id, vm_id): try: - if not model.load_docker_data(task_id, vm_id, cache, force_cache_refresh=False): + if task_id not in settings.CODE_SUBMISSION_REFERENCE_REPOSITORIES or not model.load_docker_data(task_id, vm_id, cache, force_cache_refresh=False): return JsonResponse({'status': 0, "context": {'disabled': True}}) return JsonResponse({'status': 0, "context": model.get_submission_git_repo(vm_id, task_id)}) diff --git a/application/src/tira/frontend-vuetify/src/components/RunReviewForm.vue b/application/src/tira/frontend-vuetify/src/components/RunReviewForm.vue index e506df51c..15d634d65 100644 --- a/application/src/tira/frontend-vuetify/src/components/RunReviewForm.vue +++ b/application/src/tira/frontend-vuetify/src/components/RunReviewForm.vue @@ -48,13 +48,13 @@ <v-divider class="py-6" v-if="!loading"/> <h2 v-if="!loading">Output</h2> - <v-tabs v-model="tab" align-tabs="title" v-if="!loading"> + <v-tabs v-model="tab" v-if="!loading" class="my-2"> <v-tab v-for="item in detailed_tabs" :key="item.key" :value="item.key" >{{ item.key }}</v-tab> </v-tabs> <v-window v-if="!loading" v-model="tab" :touch="{left: () => {}, right: () => {}}"> <v-window-item v-for="item in detailed_tabs" :key="item.key" :value="item.key"> <v-card flat> - <v-code tag="pre">{{item.content }}</v-code> + <code-snippet title="Click the icon to copy code" :code="item.content" expand_message=""/> </v-card> </v-window-item> </v-window> @@ -62,12 +62,13 @@ <script lang="ts"> import Loading from './Loading.vue' +import CodeSnippet from "@/components/CodeSnippet.vue"; import MetadataItems from './MetadataItems.vue' import { get, post, reportError, reportSuccess, inject_response, extractDatasetFromCurrentUrl } from '../utils' export default { name: "run-review-form", - components: { Loading, MetadataItems }, + components: { Loading, CodeSnippet, MetadataItems }, emits: ['review-run'], props: ['run_id', 'vm_id', 'dataset_id_from_props', ], data() { diff --git a/application/src/tira/frontend-vuetify/src/submission-components/CodeSubmission.vue b/application/src/tira/frontend-vuetify/src/submission-components/CodeSubmission.vue index cf3e3fc5e..549e24974 100644 --- a/application/src/tira/frontend-vuetify/src/submission-components/CodeSubmission.vue +++ b/application/src/tira/frontend-vuetify/src/submission-components/CodeSubmission.vue @@ -9,8 +9,12 @@ <v-card-item v-if="!loading && !disabled"> <v-card-text v-if="!repo_url"> <v-form ref="form" v-model="valid"> + + <v-row class="d-flex align-center justify-center"> + <v-col cols="6"><v-checkbox v-model="allow_public_repo" label="Allow Public Github Repository (You can change this later. Using a public Github repository has the advantage that Github Actions are free, otherwise, they count towards the bill of organizers.)" /></v-col> + </v-row> <v-row class="d-flex align-center justify-center"> - <v-col cols="6"><v-text-field v-model="new_git_account" label="Your GitHub Account" :rules="[v => !!(v && v.length) \|\| 'Please select your GitHub account.']"/></v-col> + <v-col cols="6"><v-text-field @keydown.enter.prevent="connectToCodeRepo()" v-model="new_git_account" label="Your GitHub Account" :rules="[v => !!(v && v.length) \|\| 'Please select your GitHub account.']"/></v-col> </v-row> <v-row class="d-flex align-center justify-center"> <v-col cols="6"><v-btn block color="primary" :loading="submit_in_progress" @click="connectToCodeRepo()" text="Add Code Repository"/></v-col> @@ -71,6 +75,7 @@ export default { return { loading: true, new_git_account: '', + allow_public_repo: true, valid: false, submit_in_progress: false, repo_url: '', @@ -90,12 +95,9 @@ export default { } this.submit_in_progress = true - post(`/api/add_software_submission_git_repository/${this.task_id}/${this.user_id}`, {"external_owner": this.new_git_account}) + post(`/api/add_software_submission_git_repository/${this.task_id}/${this.user_id}`, {"external_owner": this.new_git_account, "allow_public_repo": this.allow_public_repo}) .then(reportSuccess('Your git repository was created.')) - .then(message => { - this.repo_url = message.context.repo_url - this.owner_url = message.context.owner_url - }) + .then(inject_response(this)) .catch(reportError("Problem while adding your git repository.", "This might be a short-term hiccup, please try again. We got the following error: ")) .then(() => {this.submit_in_progress = false}) }, diff --git a/application/src/tira/git_runner_integration.py b/application/src/tira/git_runner_integration.py index 60c3f2d56..1c946ccb7 100644 --- a/application/src/tira/git_runner_integration.py +++ b/application/src/tira/git_runner_integration.py @@ -28,8 +28,16 @@ def normalize_file(file_content, tira_user_name, task_id): - return file_content.replace('TIRA_USER_FOR_AUTOMATIC_REPLACEMENT', tira_user_name)\ - .replace('TIRA_TASK_ID_FOR_AUTOMATIC_REPLACEMENT', task_id) + default_datasets = {'webpage-classification': 'webpage-classification/tiny-sample-20231023-training', + 'ir-lab-jena-leipzig-wise-2023': 'workshop-on-open-web-search/retrieval-20231027-training', + 'ir-lab-jena-leipzig-sose-2023': 'workshop-on-open-web-search/retrieval-20231027-training', + 'workshop-on-open-web-search': 'workshop-on-open-web-search/retrieval-20231027-training', + 'ir-benchmarks': 'workshop-on-open-web-search/retrieval-20231027-training', + } + + return file_content.replace('TIRA_USER_FOR_AUTOMATIC_REPLACEMENT', tira_user_name) \ + .replace('TIRA_TASK_ID_FOR_AUTOMATIC_REPLACEMENT', task_id) \ + .replace('TIRA_DATASET_FOR_AUTOMATIC_REPLACEMENT', default_datasets.get(task_id, '<TODO-ADD-DATASET>')) def convert_size(size_bytes): @@ -185,7 +193,7 @@ def initialize_user_repository(self, git_repository_id, repo_name, token): 'user_name': repo_name.replace('tira-user-', ''), 'repo_name': repo_name, 'token': token, - 'image_prefix': self.image_registry_prefix +'/' + repo_name +'/' + 'image_prefix': self.image_registry_prefix + '/' + repo_name + '/' }) with tempfile.TemporaryDirectory() as tmp_dir: @@ -1184,7 +1192,7 @@ def get_git_runner_for_software_integration(self, reference_repository_name, use user_repository_namespace, github_user, tira_user_name, dockerhub_token, dockerhub_user, tira_client_token, repository_search_prefix, tira_task_id, tira_code_repository_id, - tira_client_user): + tira_client_user, private): user = self.gitHoster_client.get_user() try: user_repo = user.get_repo(f'{user_repository_namespace}/{user_repository_name}') @@ -1199,18 +1207,18 @@ def get_git_runner_for_software_integration(self, reference_repository_name, use tira_user_name, dockerhub_token, dockerhub_user, tira_client_token, repository_search_prefix, tira_task_id, tira_code_repository_id, - tira_client_user) + tira_client_user, private) def create_software_submission_repository_for_user(self, reference_repository_name, user_repository_name, user_repository_namespace, github_user, tira_user_name, dockerhub_token, dockerhub_user, tira_client_token, repository_search_prefix, tira_task_id, tira_code_repository_id, - tira_client_user): + tira_client_user, private): reference_repo = self.gitHoster_client.get_repo(reference_repository_name) org = self.gitHoster_client.get_organization(user_repository_namespace) repo = org.create_repo(user_repository_name, - f'The repository of user {tira_user_name} for code submissions in TIRA.', private=True) + f'The repository of user {tira_user_name} for code submissions in TIRA.', private=private) repo.add_to_collaborators(github_user, 'admin') repo.create_secret('TIRA_DOCKER_REGISTRY_TOKEN', dockerhub_token) diff --git a/application/src/tira/tira_model.py b/application/src/tira/tira_model.py index 296c3ddd9..08ed6084c 100644 --- a/application/src/tira/tira_model.py +++ b/application/src/tira/tira_model.py @@ -185,7 +185,7 @@ def github_user_exists(user_name): return g.git_user_exists(user_name) -def get_submission_git_repo(vm_id, task_id, disraptor_user=None, external_owner=None): +def get_submission_git_repo(vm_id, task_id, disraptor_user=None, external_owner=None, private=True): user_repository_name = slugify(task_id) + '-' + slugify(vm_id) repository_url = settings.CODE_SUBMISSION_REPOSITORY_NAMESPACE + '/' + user_repository_name ret = model.get_submission_git_repo_or_none(repository_url, vm_id) @@ -196,7 +196,7 @@ def get_submission_git_repo(vm_id, task_id, disraptor_user=None, external_owner= docker_data = load_docker_data(task_id, vm_id, cache, force_cache_refresh=False) docker_registry_user = docker_data['docker_registry_user'] docker_registry_token = docker_data['docker_registry_token'] - reference_repository = settings.CODE_SUBMISSION_REFERENCE_REPOSITORY + reference_repository = settings.CODE_SUBMISSION_REFERENCE_REPOSITORIES[task_id] disraptor_description = disraptor_user + '-repo-' + task_id + '-' + vm_id discourse_api_key = discourse_api_client().generate_api_key(disraptor_user, disraptor_description) @@ -219,6 +219,7 @@ def get_submission_git_repo(vm_id, task_id, disraptor_user=None, external_owner= tira_task_id=task_id, tira_code_repository_id=repository_url, tira_client_user=disraptor_user, + private=private ) ret.confirmed = True @@ -230,7 +231,7 @@ def get_submission_git_repo(vm_id, task_id, disraptor_user=None, external_owner= def git_pipeline_is_enabled_for_task(task_id, cache, force_cache_refresh=False): evaluators_for_task = get_evaluators_for_task(task_id, cache, force_cache_refresh) git_runners_for_task = [i['is_git_runner'] for i in evaluators_for_task] - + # We enable the docker part only if all evaluators use the docker variant. return len(git_runners_for_task) > 0 and all(i for i in git_runners_for_task) diff --git a/application/src/tira/urls.py b/application/src/tira/urls.py index 0dade4af0..b37ec8bc3 100644 --- a/application/src/tira/urls.py +++ b/application/src/tira/urls.py @@ -107,6 +107,7 @@ path('api/registration/add_registration/<str:vm_id>/<str:task_id>', data_api.add_registration, name='add_registration'), path('api/submissions-for-task/<str:task_id>/<str:user_id>/<str:submission_type>', data_api.submissions_for_task, name="submissions_for_task"), path('api/tirex-components', data_api.tirex_components, name='tirex_components'), + path('api/snippets-for-tirex-components', data_api.get_snippet_to_run_components, name='get_snippet_to_run_components'), path('api/re-ranking-datasets/<str:task_id>', data_api.reranking_datasets, name='reranking_datasets'), path('api/submissions-of-user/<str:vm_id>', data_api.submissions_of_user, name='submissions_of_user'), path('api/add_software_submission_git_repository/<str:task_id>/<str:vm_id>', vm_api.add_software_submission_git_repository, name='add_software_submission_git_repository'), diff --git a/application/src/tirex-components.yml b/application/src/tirex-components.yml index cd32d7130..d3d8fd281 100644 --- a/application/src/tirex-components.yml +++ b/application/src/tirex-components.yml @@ -207,6 +207,8 @@ - display_name: Tutorial href: https://github.com/OpenWebSearch/wows-code/blob/main/ecir24/post-hoc-notebooks/keyphrase-extraction.ipynb target: _blank + - display_name: Submission in TIREx + href: /submissions/ir-benchmarks/ows/bce-fo - display_name: Health Classification component_type: [TIREx, Tutorial] @@ -287,7 +289,9 @@ - display_name: Tutorial href: https://github.com/webis-de/ir-pad/tree/public/tutorials#tutorial-8-query-segmentation-in-progress-by-maik target: _blank - + - display_name: Submission in TIREx + href: /submissions/ir-benchmarks/ows/query-segmentation-hyp-a + - display_name: hyp-b focus_type: [Precision] component_type: [TIREx, Tutorial] @@ -295,6 +299,8 @@ - display_name: Tutorial href: https://github.com/webis-de/ir-pad/tree/public/tutorials#tutorial-8-query-segmentation-in-progress-by-maik target: _blank + - display_name: Submission in TIREx + href: /submissions/ir-benchmarks/ows/query-segmentation-hyp-b - display_name: hyp-i focus_type: [Precision] diff --git a/python-client/Makefile b/python-client/Makefile index 308537015..cd986a83a 100644 --- a/python-client/Makefile +++ b/python-client/Makefile @@ -3,7 +3,7 @@ build-pypi-package: run-tests python3 -m build --sdist . python3 -m build --wheel . - twine upload dist/tira-0.0.97-py3-none-any.whl dist/tira-0.0.97.tar.gz + twine upload dist/tira-0.0.102-py3-none-any.whl dist/tira-0.0.102.tar.gz run-tests: docker run -u root --rm -v /var/run/docker.sock:/var/run/docker.sock -v ${PWD}:/app -w /app --entrypoint pytest webis/tira:python-client-dev-0.0.4 diff --git a/python-client/sphinx/development/python-client/workflow.rst b/python-client/sphinx/development/python-client/workflow.rst index 49b14e313..b581dc665 100644 --- a/python-client/sphinx/development/python-client/workflow.rst +++ b/python-client/sphinx/development/python-client/workflow.rst @@ -35,6 +35,13 @@ Unit Tests pytest +Code Coverage +~~~~~~~~~~~~~ +.. code-block:: bash + + coverage run -m pytest tests + coverage xml --omit=tests/** + Generating documentation ~~~~~~~~~~~~~~~~~~~~~~~~ From within the `python-client/sphynx` directory run diff --git a/python-client/tira/__init__.py b/python-client/tira/__init__.py index 625f000df..945bce51f 100644 --- a/python-client/tira/__init__.py +++ b/python-client/tira/__init__.py @@ -1,1 +1,1 @@ -__version__ = "0.0.97" +__version__ = "0.0.102" diff --git a/python-client/tira/io_utils.py b/python-client/tira/io_utils.py index 8091a76ff..3c60bd380 100644 --- a/python-client/tira/io_utils.py +++ b/python-client/tira/io_utils.py @@ -3,18 +3,19 @@ import gzip import json import os -from typing import Any, Iterable, Dict, Union +from typing import Any, List, Iterable, Dict, Union, Generator +from pathlib import Path +import logging def parse_jsonl_line(input: Union[str, bytearray, bytes], load_default_text: bool) -> Dict: - """ Deseralizes the line using JSON deserialization. Optionally strips the 'original_query' and 'original_document' fields from the resulting object and converts the qid and docno fields to strings. :param str \| bytearray \| bytes input: A json-serialized string. - :param bool load_default_text: If true, the origianl_query and original_document fields are removed and the qid and docno - values are converted to strings. + :param bool load_default_text: If true, the original_query and original_document fields are removed and the qid and + docno values are converted to strings. :return: The deserialized and (optionally) processed object. :rtype: dict @@ -43,7 +44,11 @@ def parse_jsonl_line(input: Union[str, bytearray, bytes], load_default_text: boo return obj -def stream_all_lines(input_file, load_default_text): +def stream_all_lines(input_file: Union[str, Iterable[bytes]], load_default_text: bool) -> Generator[Dict, Any, Any]: + """ + .. todo:: add documentation + .. todo:: this function has two semantics: handling a file and handling file-contents + """ if type(input_file) is str: if not os.path.isfile(input_file): return @@ -61,7 +66,11 @@ def stream_all_lines(input_file, load_default_text): yield parse_jsonl_line(line, load_default_text) -def all_lines_to_pandas(input_file: str \| Iterable[str], load_default_text): +def all_lines_to_pandas(input_file: Union[str, Iterable[str]], load_default_text: bool) -> pd.DataFrame: + """ + .. todo:: add documentation + .. todo:: this function has two semantics: handling a file and handling file-contents + """ if type(input_file) is str: if input_file.endswith('.gz'): with gzip.open(input_file, 'rt', encoding='utf-8') as f: @@ -79,17 +88,36 @@ def all_lines_to_pandas(input_file: str \| Iterable[str], load_default_text): return pd.DataFrame(ret) -def __num(s): +def __num(input: str) -> Union[str, int, float]: + """ + Converts the input to an int or float if possible. Returns the inputted string otherwise. + + :param str input: The string that should be converted to a float or int if possible. + :return: The intrepteted input. + :rtype: str \| int \| float + + :Example: + >>> __num("hello world") + "hello world" + >>> __num("-42") + -42 + >>> __num("3.5") + 3.5 + >>> __num("2e-6") + 2e-6 + >>> __num(" -42") + " -42" + """ try: - return int(s) + return int(input) except ValueError: try: - return float(s) + return float(input) except ValueError: - return s + return input -def run_cmd(cmd, ignore_failure=False): +def run_cmd(cmd: List[str], ignore_failure=False): import subprocess exit_code = subprocess.call(cmd) @@ -148,7 +176,7 @@ def all_environment_variables_for_github_action_or_fail(params): return [k + '=' + v for k, v in ret.items()] -def load_output_of_directory(directory, evaluation=False, verbose=False): +def load_output_of_directory(directory: Path, evaluation: bool=False) -> Union[Dict, pd.DataFrame]: files = glob(str(directory) + '/*') if evaluation: @@ -159,8 +187,7 @@ def load_output_of_directory(directory, evaluation=False, verbose=False): files = files[0] - if verbose: - print(f'Read file from {files}') + logging.debug(f'Read file from {files}') if evaluation: ret = {} diff --git a/python-client/tira/local_execution_integration.py b/python-client/tira/local_execution_integration.py index 3db0db4bf..3261bed74 100644 --- a/python-client/tira/local_execution_integration.py +++ b/python-client/tira/local_execution_integration.py @@ -7,6 +7,8 @@ import subprocess import tarfile import logging +from pathlib import Path +import pandas as pd class LocalExecutionIntegration(): @@ -174,7 +176,7 @@ def run(self, identifier=None, image=None, command=None, input_dir=None, output_ container = client.containers.run(image, entrypoint='sh', command=f'-c "{command}; sleep .1"', environment=environment, volumes=volumes, detach=True, remove=True, network_disabled = not allow_network) for line in container.attach(stdout=True, stream=True, logs=True): - logging.info(line.decode('utf-8')) + print(line.decode('utf-8')) if evaluate: evaluation_volumes = {str(eval_dir): {'bind': '/tira-data/eval_output', 'mode': 'rw'}} @@ -196,7 +198,7 @@ def run(self, identifier=None, image=None, command=None, input_dir=None, output_ container = client.containers.run(image, entrypoint='sh', command=f'-c "{command}; sleep .1"', volumes=evaluation_volumes, detach=True, remove=True, network_disabled = not allow_network) for line in container.attach(stdout=True, stream=True, logs=True): - logging.info(line.decode('utf-8'), flush=True) + print(line.decode('utf-8'), flush=True) if evaluate: approach_name = identifier if identifier else f'"{command}"@{image}' diff --git a/python-client/tira/tira_client.py b/python-client/tira/tira_client.py index 34d3cc7f2..11e46778f 100644 --- a/python-client/tira/tira_client.py +++ b/python-client/tira/tira_client.py @@ -1,5 +1,5 @@ from abc import ABC -from typing import TYPE_CHECKING, Literal, overload +from typing import TYPE_CHECKING, Literal, Union, overload if TYPE_CHECKING: from typing import Any, Dict, Optional @@ -75,6 +75,6 @@ def download_run( team=None, previous_stage=None, return_metadata: bool = False, - ) -> "pd.DataFrame \| tuple[pd.DataFrame, str]": + ) -> "Union[pd.DataFrame, tuple[pd.DataFrame, str]]": # .. todo:: typehint pass diff --git a/python-client/tira/tira_run.py b/python-client/tira/tira_run.py index c755d9c11..23b247e81 100755 --- a/python-client/tira/tira_run.py +++ b/python-client/tira/tira_run.py @@ -43,6 +43,7 @@ def parse_args(): parser.add_argument('--tira-vm-id', required=False, default=os.environ.get('TIRA_VM_ID')) parser.add_argument('--tira-task-id', required=False, default=os.environ.get('TIRA_TASK_ID')) parser.add_argument('--tira-code-repository-id', required=False, default=os.environ.get('TIRA_CODE_REPOSITORY_ID')) + parser.add_argument('--fail-if-output-is-empty', required=False, default=False, action='store_true', help='After the execution of the software, fail if the output directory is empty.') args = parser.parse_args() if args.export_submission_from_jupyter_notebook: @@ -165,8 +166,17 @@ def main(): print(f'Done: Evaluation truth for dataset {dataset} is available.') logging.basicConfig(level=logging.DEBUG if args.verbose else logging.INFO) + print(f''' +########################################### TIRA RUN CONFIGURATION ########################################### +# image=${args.image} +# command=${args.command} +############################################################################################################## +''') client.local_execution.run(identifier=args.approach, image=args.image, command=args.command, input_dir=input_dir, output_dir=args.output_directory, dry_run=args.dry_run, allow_network=args.allow_network, input_run=args.input_run, additional_volumes=args.v, evaluate=evaluate, eval_dir=args.evaluation_directory) - + + if args.fail_if_output_is_empty and (not os.path.exists(args.output_directory) or not os.listdir(args.output_directory)): + raise ValueError('The software produced an empty output directory, it likely failed?') + if args.push.lower() == 'true': print('Push Docker image') client.local_execution.push_image(args.image, args.tira_docker_registry_user, args.tira_docker_registry_token)	Changed code display in run-review to code-snippet component. Changed code display in run review to CodeSnippter component. Also checked if we display code in other place without using CodeSnippet component, but didn't find anything. <img width="764" alt="image" src="https://github.com/tira-io/tira/assets/73996300/1f27aef4-29af-4e7a-8a4f-f61517a7416e"> tira-run seems to not print stderr and stdout to the console within Github Actions Learned this during the IRIXYS hackathon		2023-12-01T15:31:37	[]	[]
GoogleCloudPlatform/snapshot-debugger	GoogleCloudPlatform__snapshot-debugger-143	70f8def8d2daf0d209e0620f7fd19a38a283d44e	diff --git a/README.md b/README.md index 8e65f70..51a99d6 100644 --- a/README.md +++ b/README.md @@ -249,6 +249,42 @@ Note: The information printed by the `init` command can be accessed from within your Firebase project. It’s safe to run the `snapshot-dbg-cli init --use-default-rtdb` command multiple times to view this information. +#### Setting up Firebase RTDB in other regions + +By default, `snapshot-dbg-cli init` will create a Firebase Realtime Database in +`us-central1`. It is possible to create and use a database in any region +supported by Firebase Realtime Database. See +[supported RTDB locations][rtdb_locations]. + +Setting up your database in a non-default region comes with some trade-offs: +* As a positive, you get to control where your snapshot data will be stored. + This may be important for compliance reasons. +* As a negative, the vsCode extension and agents will be unable to + automatically find the database. The database URL will need to be provided + explicitly via configuration, see the following for details: + * [Configuring the Java Agent][java_agent_config] + * [Configuring the Python Agent][python_agent_config] + * [Configuring the Node.js Agent][nodejs_agent_config] + * [Configuring the VsCode Extension][extension_config] + +You can set up your database in a non-default location as follows: +``` +snapshot-dbg-cli init --location={YOUR_LOCATION} +``` + +For example, you may want to set up your database in Belgium, and so would run +```snapshot-dbg-cli init --location=europe-west1``` + +Make note of the database URL provided in the command output; you will need to +provide this to your debug agent(s) and the vsCode plugin. + +[rtdb_locations]: https://firebase.google.com/docs/projects/locations#rtdb-locations +[java_agent_config]: https://github.com/GoogleCloudPlatform/cloud-debug-java#snapshot-debugger---firebase-realtime-database-backend +[python_agent_config]: https://github.com/GoogleCloudPlatform/cloud-debug-python/blob/main/README.md#snapshot-debugger---firebase-realtime-database-backend +[nodejs_agent_config]: https://github.com/googleapis/cloud-debug-nodejs/blob/main/README.md#snapshot-debugger---firebase-realtime-database-backend +[extension_config]: https://github.com/GoogleCloudPlatform/snapshot-debugger/blob/main/snapshot_dbg_extension/README.md#configuration + + ## Set up Snapshot Debugger in your Google Cloud project ### Working Samples diff --git a/snapshot_dbg_cli/firebase_management_rest_service.py b/snapshot_dbg_cli/firebase_management_rest_service.py index 90e1c03..4063100 100644 --- a/snapshot_dbg_cli/firebase_management_rest_service.py +++ b/snapshot_dbg_cli/firebase_management_rest_service.py @@ -58,6 +58,9 @@ {response} """ +VALID_LOCATIONS_URL = ( + "https://firebase.google.com/docs/projects/locations#rtdb-locations") + class FirebaseManagementRestService: """Implements a service for making Firebase RTDB management REST requests. @@ -282,6 +285,20 @@ def rtdb_instance_create(self, database_id, location): self._user_output.debug("Got 400:", parsed_error) return DatabaseCreateResponse( DatabaseCreateStatus.FAILED_PRECONDITION) + + if parsed_error["error"]["status"] == "INVALID_ARGUMENT": + print_http_error( + self._user_output, request, err, error_message=error_message) + + self._user_output.error( + "This was attempting to create the database instance. One " + "potential reason for this is if an invalid location was " + "specified, valid locations for RTDBs can be found at " + f"{VALID_LOCATIONS_URL}. To note, valid RTDB locations are a " + "subset of valid Google Cloud regions.") + + raise SilentlyExitError from err + except (TypeError, KeyError, ValueError): pass diff --git a/snapshot_dbg_cli/firebase_types.py b/snapshot_dbg_cli/firebase_types.py index 83f00fc..a5b71cb 100644 --- a/snapshot_dbg_cli/firebase_types.py +++ b/snapshot_dbg_cli/firebase_types.py @@ -17,6 +17,7 @@ """ from enum import Enum +import re FIREBASE_MANAGMENT_API_SERVICE = 'firebase.googleapis.com' FIREBASE_RTDB_MANAGMENT_API_SERVICE = 'firebasedatabase.googleapis.com' @@ -123,12 +124,26 @@ def __init__(self, database_instance): self.database_url = database_instance['databaseUrl'] self.type = database_instance['type'] self.state = database_instance['state'] + self.location = self.extract_location(self.name) + + if self.location is None: + raise ValueError( + f"Failed to extract location from project name '{self.name}'") + except KeyError as e: missing_key = e.args[0] error_message = ('DatabaseInstance is missing expected field ' f"'{missing_key}' instance: {database_instance}") raise ValueError(error_message) from e + @staticmethod + def extract_location(name): + location_search = re.search('/locations/([^/]+)/', name) + if not location_search or len(location_search.groups()) != 1: + return None + + return location_search.groups()[0] + class DatabaseCreateResponse: """Represents the response of a database create request. diff --git a/snapshot_dbg_cli/init_command.py b/snapshot_dbg_cli/init_command.py index 8421f52..b0a99cf 100644 --- a/snapshot_dbg_cli/init_command.py +++ b/snapshot_dbg_cli/init_command.py @@ -192,6 +192,21 @@ state: {db_state} """ +LOCATION_MISMATCH_ERROR_MSG = ( + "ERROR the following database already exists: '{full_database_name}', " + "however its location '{existing_location}' does not match the requested " + "location '{requested_location}'.\n\n" + "The database ID ('{database_id}' in this case) must be unique across " + 'locations and projects.\n\n' + "If you meant to finalize the initialization of '{database_id}' in " + "'{requested_location}' (or verify it is already correctly initialized) " + "rerun the init command and specify '--location={existing_location}'.\n\n" + "If you meant to create a new database in '{requested_location}', given " + "'{database_id}' is already in use, you'll need to use a new name by " + "providing the '--database-id' argument to the init command. Note, even " + "if you delete '{database_id}', the name will remain reserved and will not " + 'be available to be reused in a new location, a new name must be chosen.') + class InitCommand: """This class implements the init command. @@ -222,14 +237,6 @@ def register(self, args_subparsers, required_parsers, common_parsers): # Only some locations are supported, see: # https://firebase.google.com/docs/projects/locations#rtdb-locations - # - # If unsupported location is used, this error occurs - # "error": { - # "code": 400, - # "message": "Request contains an invalid argument.", - # "status": "INVALID_ARGUMENT" - # } - # For now however we only support 'us-central1' parser.add_argument( '-l', '--location', help=LOCATION_HELP, default=DEFAULT_LOCATION) self.args_parser = parser @@ -243,11 +250,6 @@ def cmd(self, args, cli_services): self.permissions_service = cli_services.permissions_service self.project_id = cli_services.project_id - if args.location != DEFAULT_LOCATION: - self.user_output.error('ERROR: Currently the only supported location is ' - f"'{DEFAULT_LOCATION}'") - raise SilentlyExitError - # If the user does not have the required permissions this will emit an error # message and exit. self.permissions_service.check_required_permissions(REQUIRED_PERMISSIONS) @@ -347,6 +349,16 @@ def check_and_handle_database_instance(self, args, firebase_project): if status == DatabaseGetStatus.EXISTS: database_instance = instance_response.database_instance + if args.location != database_instance.location: + self.user_output.error( + LOCATION_MISMATCH_ERROR_MSG.format( + full_database_name=database_instance.name, + existing_location=database_instance.location, + requested_location=args.location, + database_id=database_id)) + + raise SilentlyExitError + elif status == DatabaseGetStatus.DOES_NOT_EXIST: create_response = self.firebase_management_service.rtdb_instance_create( database_id=database_id, location=args.location)	CLI general `location` support Currently the CLI `init` command restricts the `location` to `us-central1`. However, it is possible to create a Firebase RTDB instances in other regions, which can be found at https://firebase.google.com/docs/projects/locations#rtdb-locations. The CLI should be modified to support these other locations as well. This issue was motivated by issue #134 .		2023-05-17T14:49:19	[]	[]
GoogleCloudPlatform/snapshot-debugger	GoogleCloudPlatform__snapshot-debugger-64	59c525e9a2877be7882e2648cdb668c1913d864f	diff --git a/snapshot_dbg_cli/breakpoint_utils.py b/snapshot_dbg_cli/breakpoint_utils.py index 8f39031..f1767a8 100644 --- a/snapshot_dbg_cli/breakpoint_utils.py +++ b/snapshot_dbg_cli/breakpoint_utils.py @@ -131,7 +131,7 @@ def normalize_breakpoint(bp, bpid=None): Returns: The normalized breakpoint on success, None on failure. """ - if bp is None: + if type(bp) is not dict: return None if 'id' not in bp and bpid is not None:	Harden `normalize_breakpoint` utility function The `normalize_breakpoint` utility function should ensure the breakpoint passed in is a dict as expected. This issue came about when using a DB that had some incorrectly formatted data, and the CLI had the following error: ``` python3 -m snapshot_dbg_cli list_snapshots --include-inactive Traceback (most recent call last): File "/usr/lib/python3.10/runpy.py", line 196, in _run_module_as_main return _run_code(code, main_globals, None, File "/usr/lib/python3.10/runpy.py", line 86, in _run_code exec(code, run_globals) File "/snapshot-debugger/snapshot_dbg_cli/__main__.py", line 18, in <module> snapshot_dbg_cli.run_main() File "/snapshot-debugger/snapshot_dbg_cli/__init__.py", line 33, in run_main main() File "/snapshot-debugger/snapshot_dbg_cli/__init__.py", line 28, in main snapshot_dbg_cli.cli_run.run() File "/snapshot-debugger/snapshot_dbg_cli/cli_run.py", line 69, in run args.func(args=args, cli_services=cli_services) File "/snapshot-debugger/snapshot_dbg_cli/list_snapshots_command.py", line 96, in cmd snapshots = debugger_rtdb_service.get_snapshots( File "/snapshot-debugger/snapshot_dbg_cli/snapshot_debugger_rtdb_service.py", line 212, in get_snapshots return self._get_breakpoints(debuggee_id, include_inactive, 'CAPTURE', File "/snapshot-debugger/snapshot_dbg_cli/snapshot_debugger_rtdb_service.py", line 225, in _get_breakpoints breakpoints += self._get_breakpoints_by_path_and_filter( File "/snapshot-debugger/snapshot_dbg_cli/snapshot_debugger_rtdb_service.py", line 236, in _get_breakpoints_by_path_and_filter breakpoints = [ File "/snapshot-debugger/snapshot_dbg_cli/snapshot_debugger_rtdb_service.py", line 238, in <listcomp> if normalize_breakpoint(bp, bpid) and bp['action'] == action and File "/snapshot-debugger/snapshot_dbg_cli/breakpoint_utils.py", line 138, in normalize_breakpoint bp['id'] = bpid TypeError: 'str' object does not support item assignment ```		2022-07-28T16:17:52	[]	[]
GoogleCloudPlatform/snapshot-debugger	GoogleCloudPlatform__snapshot-debugger-62	a73dd5990b3132fb3b57b737af0548dc7544b5ea	diff --git a/snapshot_dbg_cli/snapshot_parser.py b/snapshot_dbg_cli/snapshot_parser.py index 8fe9d4a..e576a6e 100644 --- a/snapshot_dbg_cli/snapshot_parser.py +++ b/snapshot_dbg_cli/snapshot_parser.py @@ -73,9 +73,10 @@ def parse_expressions(self): def parse_locals(self, stack_frame_index): local_variables = [] - if stack_frame_index < len( - self.stack_frames) and 'locals' in self.stack_frames[stack_frame_index]: - local_variables = self.stack_frames[stack_frame_index]['locals'] + if stack_frame_index < len(self.stack_frames): + stack_frame = self.stack_frames[stack_frame_index] + for p in ['arguments', 'locals']: + local_variables.extend(stack_frame.get(p, [])) return self._resolve_variables(local_variables)	The get_snapshot command does not list function arguments The captured snapshot data for functions in the call stack contains captured variable data for both `locals` and `arguments`. Currently the `get_snapshot` command is only displaying the variables classified as `locals`, the `arguments` variables are missing.		2022-07-28T14:43:37	[]	[]
GoogleCloudPlatform/snapshot-debugger	GoogleCloudPlatform__snapshot-debugger-33	9038f87a96813e32f8fdf00dd433ed7ca3558aaf	diff --git a/cli/data_formatter.py b/cli/data_formatter.py index 8546c1a..b1b4860 100644 --- a/cli/data_formatter.py +++ b/cli/data_formatter.py @@ -39,7 +39,7 @@ def build_table(self, headers, values): """ widths = [ max(len(headers[i]), max([len(v[i]) - for v in values])) + for v in values], default=0)) for i in range(len(headers)) ]	List snapshots failure when there are no snapshots to display When the `list_snapshots` command has no snapshots to display, it results in the following failure: ``` File "[...snip]/snapshot-debugger/cli/data_formatter.py", line 41, in <listcomp> max(len(headers[i]), max([len(v[i]) ValueError: max() arg is an empty sequence ```		2022-06-06T18:35:57	[]	[]
ansys/aedt-testing	ansys__aedt-testing-211	5f12d6d10c6a0a50654aa84f7a728a87af85cf94	diff --git a/docs/CONTRIBUTE.md b/docs/CONTRIBUTE.md index 8ce14dc..6235ce2 100644 --- a/docs/CONTRIBUTE.md +++ b/docs/CONTRIBUTE.md @@ -2,13 +2,13 @@ <!-- toc --> -- [Configure your enviornment](#configure-your-enviornment) +- [Configure your environment](#configure-your-environment) - [Run all tests and validations via tox](#run-all-tests-and-validations-via-tox) - [Build package](#build-package) <!-- tocstop --> -## Configure your enviornment +## Configure your environment Install all dependencies ```bash pip install .[test]	Project report for [Project] does not exist Correlation plot is generated when comparing 2022R1 and 2022R2 result. However, the correlation plot is not generated when comparing 2022R1/R2 with 2023R1. Getting "error_exception" with the error message "Project report for [myProject] does not exist".		2022-08-17T13:53:38	[]	[]
ansys/aedt-testing	ansys__aedt-testing-101	2e611333d948ee9dacd37a20660bef7efc8db429	diff --git a/pyproject.toml b/pyproject.toml index f7362c2..cabf824 100644 --- a/pyproject.toml +++ b/pyproject.toml @@ -18,14 +18,13 @@ classifiers = ["License :: OSI Approved :: MIT License"] dynamic = ["version", "description"] dependencies = [ - "pyaedt==0.4.12", + "pyaedt==0.4.16", "Django==3.2.8", ] # development dependencies [project.optional-dependencies] test = [ - "black==21.9b0", "pre-commit==2.15.0", "mypy==0.910", "pytest==6.2.5",	hide pyaedt logs under --debug flag		2021-12-06T13:42:21	[]	[]
ansys/aedt-testing	ansys__aedt-testing-10	347912f0a9634aedcf3880fe58930c96e9a481f6	diff --git a/requirements.txt b/requirements.txt index 53af3bb..f44d0db 100644 --- a/requirements.txt +++ b/requirements.txt @@ -1,1 +1,2 @@ -pyaedt==0.3.27 \ No newline at end of file +pyaedt==0.3.27 +Django==3.2.8 \ No newline at end of file	report progress of the run: queued, running, status (success/fail) consider using HTML report see https://stackoverflow.com/questions/6748559/generating-html-documents-in-python https://github.com/nedbat/coveragepy/blob/2d2d6088b94845ffc4062853aa7230879ee66a44/coverage/html.py#L125 https://stackoverflow.com/questions/37912260/plotting-graph-using-python-and-dispaying-it-using-html		2021-10-26T12:13:03	[]	[]
QEF/qeschema	QEF__qeschema-28	d10f0b240eb12b18c479a03857448d1a3d7bed6a	diff --git a/.travis.yml b/.travis.yml index ad8b443..2207d4e 100644 --- a/.travis.yml +++ b/.travis.yml @@ -4,7 +4,6 @@ python: - "3.7" - "3.8" - "3.9" - - "pypy3" install: - pip install -r requirements-dev.txt script: diff --git a/docs/conf.py b/docs/conf.py index c56dc27..16ff381 100644 --- a/docs/conf.py +++ b/docs/conf.py @@ -22,7 +22,7 @@ author = 'Davide Brunato, Pietro Delugas' # The full version, including alpha/beta/rc tags -release = 'v1.1' +release = '1.2.0' # -- General configuration --------------------------------------------------- diff --git a/qeschema/__init__.py b/qeschema/__init__.py index 31ab244..de06f4c 100644 --- a/qeschema/__init__.py +++ b/qeschema/__init__.py @@ -1,6 +1,5 @@ -# -- coding: utf-8 -- # -# Copyright (c), 2015-2020, Quantum Espresso Foundation and SISSA (Scuola +# Copyright (c), 2015-2021, Quantum Espresso Foundation and SISSA (Scuola # Internazionale Superiore di Studi Avanzati). All rights reserved. # This file is distributed under the terms of the MIT License. See the # file 'LICENSE' in the root directory of the present distribution, or @@ -15,11 +14,11 @@ from .exceptions import QESchemaError, XmlDocumentError from .utils import set_logger -__version__ = '1.1.0' +__version__ = '1.2.0' __all__ = [ 'XmlDocument', 'QeDocument', 'PwDocument', 'PhononDocument', 'NebDocument', 'TdDocument', 'TdSpectrumDocument', 'RawInputConverter', 'PwInputConverter', 'PhononInputConverter', 'TdInputConverter', 'TdSpectrumInputConverter', - 'QESchemaError', 'XmlDocumentError', 'set_logger' + 'NebInputConverter', 'QESchemaError', 'XmlDocumentError', 'set_logger' ] diff --git a/qeschema/cards.py b/qeschema/cards.py index 82ef609..2233bd1 100644 --- a/qeschema/cards.py +++ b/qeschema/cards.py @@ -47,6 +47,23 @@ def get_atomic_species_card(name, kwargs): return lines +def get_positions_units(item): + """ + Read different type of positions and return them together with units. + + :param item: Data item + :return: Atomic positions dict and unit, empty dict and None if not found + """ + + ptypes = ('atomic_positions', 'crystal_positions', 'wyckoff_positions') + units = ('bohr', 'crystal', 'crystal_sg') + for ptype, unit in zip(ptypes, units): + if ptype in item: + return item[ptype], unit + + return {}, None + + def get_atomic_positions_cell_card(name, kwargs): """ Convert XML data to ATOMIC_POSITIONS card @@ -62,23 +79,15 @@ def get_atomic_positions_cell_card(name, kwargs): return [] # Find atoms - atomic_positions = atomic_structure.get('atomic_positions', {}) - wyckoff_positions = atomic_structure.get('wyckoff_positions', {}) + positions, units = get_positions_units(atomic_structure) try: - is_wyckoff = False - if atomic_positions: - atoms = atomic_positions['atom'] - elif wyckoff_positions: - atoms = wyckoff_positions['atom'] - is_wyckoff = True - else: - atoms = [] + atoms = positions['atom'] except KeyError: logger.error("Cannot find any atoms for building ATOMIC_POSITIONS!") return [] - else: - if not isinstance(atoms, list): - atoms = [atoms] + + if not isinstance(atoms, list): + atoms = [atoms] # Check atoms with position constraints free_positions = kwargs.get('free_positions') @@ -89,7 +98,7 @@ def get_atomic_positions_cell_card(name, kwargs): logger.error("ATOMIC_POSITIONS: incorrect number of position constraints!") # Add atomic positions - lines = ['%s %s' % (name, 'crystal_sg' if is_wyckoff else 'bohr')] + lines = ['%s %s' % (name, units)] for k in range(len(atoms)): try: line = '{:4}'.format(atoms[k].get('@name')) @@ -233,7 +242,6 @@ def get_cell_parameters_card(name, kwargs): except KeyError: logger.error("Missing required arguments when building CELL_PARAMETERS card!") return [] - # Add cell parameters card cells = atomic_structure.get('cell', {}) if cells: @@ -262,7 +270,7 @@ def get_qpoints_card(name, kwargs): qplot = kwargs.get('qplot', False) if not qplot and not ldisp: try: - xq = kwargs['xq_dir'] + xq = kwargs['xq'] except KeyError: xq = [0.e0, 0.e0, 0.e0] line = "{:6.4f} {:8.4f} {:8.4f}".format(xq[0], xq[1], xq[2]) @@ -284,6 +292,19 @@ def get_qpoints_card(name, kwargs): return lines +def get_nat_todo_card(name, kwargs): + + natom = kwargs['nat_todo']['@natom'] + if natom == 0: + assert 'atom' not in kwargs['nat_todo'] + return [] + + atoms = kwargs['nat_todo']['atom'] + + assert natom == len(atoms) + return [' '.join(map(str, atoms))] + + def get_climbing_images(name, kwargs): assert isinstance(name, str) try: @@ -332,9 +353,7 @@ def get_neb_images_positions_card(name, kwargs): lines = ['BEGIN_POSITIONS ', 'FIRST_IMAGE '] for pos, item in enumerate(images): - positions = item.get('atomic_positions', - item.get('crystal_positions', - item.get('wyckoff_positions', {}))) + positions, units = get_positions_units(item) atoms = positions.get('atom', []) if pos == 0: @@ -345,30 +364,33 @@ def get_neb_images_positions_card(name, *kwargs): if first_nat != int(item.get('@nat', 0)): logger.error("nat provided in first image differs from number " "of atoms in atomic_positions!!!") + return '' if free_positions and len(free_positions) != 3 first_nat: logger.error("ATOMIC_POSITIONS: incorrect number of position constraints!") + return '' else: if len(atoms) != first_nat: logger.error('found images with differing number of atoms !!!') + return '' if pos < len(images) - 1: lines.append('INTERMEDIATE_IMAGE ') else: lines.append('LAST_IMAGE ') - lines.append('ATOMIC_POSITIONS { bohr }') - - for k in range(len(atoms)): - sp_name = '{:4}'.format(atoms[k]['@name']) - coords = '{:12.8f} {:12.8f} {:12.8f}'.format(atoms[k]['$']) - if k < len(free_positions): - try: - free_pos = '{:4d}{:4d}{:4d}'.format(free_positions[3 * k:3 * k + 3]) - except IndexError: - pass - else: - if sum(free_positions[3 * k + n] for n in range(3)) < 3: - lines.append('%s %s %s' % (sp_name, coords, free_pos)) + lines.append('ATOMIC_POSITIONS { %s }' % units) + + # reshape to 3x3 + free_positions_sq = [free_positions[i:i + 3] + for i in range(0, len(free_positions), 3)] + + for k, atom in enumerate(atoms): + sp_name = '{:4}'.format(atom['@name']) + coords = '{:12.8f} {:12.8f} {:12.8f}'.format(atom['$']) + + if free_positions_sq and sum(free_positions_sq[k]) < 3: + free_pos = '{:4d}{:4d}{:4d}'.format(free_positions_sq[k]) + lines.append('%s %s %s' % (sp_name, coords, free_pos)) else: lines.append('%s %s' % (sp_name, coords)) diff --git a/qeschema/converters.py b/qeschema/converters.py index f669a11..709a1c6 100644 --- a/qeschema/converters.py +++ b/qeschema/converters.py @@ -21,6 +21,10 @@ logger = logging.getLogger('qeschema') +FCP_NAMELIST = 'FCP' +PH_NAT_TODO_CARD = 'ph_nat_todo_card' +OPTIONAL_CARDS = {PH_NAT_TODO_CARD} + def conversion_maps_builder(template_map): """ @@ -79,7 +83,8 @@ def _build_maps(dict_element, path): if isinstance(item, tuple) or isinstance(item, list): try: variant_map[path_key] = _check_variant(item) - logger.debug("Added single-variant mapping: '{}'={}".format(path_key, variant_map[path_key])) + logger.debug("Added single-variant mapping: %r=%r", + path_key, variant_map[path_key]) except TypeError: variants = [] for variant in item: @@ -90,7 +95,8 @@ def _build_maps(dict_element, path): else: raise TypeError("Expect a tuple, list or string! {0}".format(variant)) variant_map[path_key] = tuple(variants) - logger.debug("Added multi-variant mapping: '{}'={}".format(path_key, variant_map[path_key])) + logger.debug("Added multi-variant mapping: %r=%r", + path_key, variant_map[path_key]) invariant_map = BiunivocalMap() variant_map = dict() @@ -116,7 +122,7 @@ class RawInputConverter(Container): """ A Fortran's namelist builder. """ - target_pattern = re.compile(r'(\w+)(?:\[((?:\w+)(?:%\w+))\]\|)') + target_pattern = re.compile(r'(\w+)(?:\[((?:\w+)(?:%\w+))]\|)') """RE pattern to extract Fortran input's namelist/card and name of a parameter""" def __init__(self, invariant_map, variant_map, input_namelists=None, input_cards=None): @@ -191,7 +197,7 @@ def set_parameter(self, path, value): else: namelist = name = None if name is None: - raise ValueError("Wrong value for invariant parameter '{0}'! '{1}'".format(path, target)) + raise ValueError("Wrong value {!r} for invariant parameter {!r}".format(target, path)) self._input[namelist][name] = to_fortran(value) logger.debug("Set {0}[{1}]={2}".format(namelist, name, self._input[namelist][name])) @@ -238,16 +244,21 @@ def get_qe_input(self): _input = self._input lines = [] for namelist in self.input_namelists: + + if namelist == FCP_NAMELIST and not len(_input[namelist]): + # Skip empty &FCP/ section + continue + lines.append('&%s' % namelist) for name, value in sorted(_input[namelist].items(), key=lambda x: x[0].lower()): - logger.debug("Add input for parameter {}[{}] with value {}".format(namelist, name, value)) + logger.debug("Add input for parameter %s[%r] with value %r", namelist, name, value) if isinstance(value, dict): # Variant conversion: apply to_fortran_input function with saved arguments try: to_fortran_input = value['_get_qe_input'] except KeyError: logger.debug( - 'No conversion function for parameter %s[%s], skip ... ' % (namelist, name) + 'No conversion function for parameter %s[%r], skip ... ', namelist, name ) continue @@ -255,7 +266,7 @@ def get_qe_input(self): lines.extend(to_fortran_input(name, value)) else: logger.error( - 'Parameter %s[%s] conversion function is not callable!' % (namelist, name) + 'Parameter %s[%r] conversion function is not callable!', namelist, name ) else: # Simple invariant conversion @@ -265,14 +276,19 @@ def get_qe_input(self): logger.debug("Add card: %s" % card) card_args = _input[card] logger.debug("Card arguments: {0}".format(card_args)) - if '_get_qe_input' not in card_args or \ - not callable(card_args['_get_qe_input']): + + if card not in OPTIONAL_CARDS and \ + ('_get_qe_input' not in card_args or + not callable(card_args['_get_qe_input'])): logger.error("Missing conversion function for card '%s'" % card) + _get_qe_input = card_args.get('_get_qe_input', None) + if callable(_get_qe_input): lines.extend(_get_qe_input(card, card_args)) - else: + elif card not in OPTIONAL_CARDS: logger.error('Card conversion function not found!') + return '\n'.join(lines) def clear_input(self): @@ -325,9 +341,11 @@ class PwInputConverter(RawInputConverter): '$': [ ('SYSTEM[ibrav]', options.set_ibrav_to_zero, None), ("ATOMIC_POSITIONS", cards.get_atomic_positions_cell_card, None), - ("CELL_PARAMETERS", cards.get_cell_parameters_card, None) + ("CELL_PARAMETERS", cards.get_cell_parameters_card, None) ], 'atomic_positions': ('ATOMIC_FORCES', cards.get_atomic_forces_card, None), + 'crystal_positions': ('ATOMIC_FORCES', cards.get_atomic_forces_card, None), + 'wyckoff_positions': ('ATOMIC_FORCES', cards.get_atomic_forces_card, None) }, 'dft': { 'functional': "SYSTEM[input_dft]", @@ -347,7 +365,8 @@ class PwInputConverter(RawInputConverter): 'dftU': { 'lda_plus_u_kind': 'SYSTEM[lda_plus_u_kind]', 'Hubbard_U': { - '$': ('SYSTEM[Hubbard_U]', options.get_specie_related_values, None), + '$': [('SYSTEM[Hubbard_U]', options.get_specie_related_values, None), + ('SYSTEM[lda_plus_u]', options.set_lda_plus_u_flag, None)] }, 'Hubbard_J0': { '$': ('SYSTEM[Hubbard_J0]', options.get_specie_related_values, None), @@ -362,7 +381,7 @@ class PwInputConverter(RawInputConverter): '$': ('SYSTEM[Hubbard_J]', options.get_specie_related_values, None), }, 'starting_ns': { - '$': ('SYSTEM[starting_ns_eigenvalue]', options.get_specie_related_values, None), + '$': ('SYSTEM[starting_ns_eigenvalue]', options.get_specie_related_values, None) }, 'U_projection_type': 'SYSTEM[U_projection_type]', }, @@ -374,6 +393,8 @@ class PwInputConverter(RawInputConverter): 'london_rcut': 'SYSTEM[london_rcut]', 'xdm_a1': 'SYSTEM[xdm_a1]', 'xdm_a2': 'SYSTEM[xdm_a2]', + 'dftd3_version': 'SYSTEM[dftd3_version]', + 'dftd3_threebody': 'SYSTEM[dftd3_threebody]', 'london_c6': { '$': ('SYSTEM[london_c6]', options.get_specie_related_values, None), } @@ -417,7 +438,8 @@ class PwInputConverter(RawInputConverter): 'nr1': "SYSTEM[nr1b]", 'nr2': "SYSTEM[nr2b]", 'nr3': "SYSTEM[nr3b]", - } + }, + 'spline_ps': "SYSTEM[spline_ps]" }, 'electron_control': { 'diagonalization': "ELECTRONS[diagonalization]", @@ -466,6 +488,7 @@ class PwInputConverter(RawInputConverter): 'free_cell': ("CELL_PARAMETERS", cards.get_cell_parameters_card, None), 'fix_volume': ("CELL[cell_dofree]", options.get_cell_dofree, None), 'fix_area': ("CELL[cell_dofree]", options.get_cell_dofree, None), + 'fix_xy': ("CELL[cell_dofree]", options.get_cell_dofree, None), 'isotropic': ("CELL[cell_dofree]", options.get_cell_dofree, None), }, 'symmetry_flags': { @@ -484,8 +507,8 @@ class PwInputConverter(RawInputConverter): 'w': "SYSTEM[esm_w]", 'efield': "SYSTEM[esm_efield]" }, - 'fcp_opt': "CONTROL[lfcpopt]", - 'fcp_mu': "SYSTEM[fcp_mu]" + 'fcp_opt': "CONTROL[lfcp]", + 'fcp_mu': "FCP[fcp_mu]" }, 'ekin_functional': { 'ecfixed': "SYSTEM[ecfixed]", @@ -496,7 +519,8 @@ class PwInputConverter(RawInputConverter): '$': ('ATOMIC_FORCES', cards.get_atomic_forces_card, None) }, 'free_positions': { - '$': [("ATOMIC_POSITIONS", cards.get_atomic_positions_cell_card, None), ("CELL_PARAMETERS",)]}, + '$': [("ATOMIC_POSITIONS", cards.get_atomic_positions_cell_card, None), + ("CELL_PARAMETERS",)]}, 'electric_field': { 'electric_potential': [ ("CONTROL[tefield]", options.get_electric_potential_related), @@ -533,12 +557,15 @@ class PwInputConverter(RawInputConverter): def __init__(self, kwargs): super(PwInputConverter, self).__init__( conversion_maps_builder(self.PW_TEMPLATE_MAP), - input_namelists=('CONTROL', 'SYSTEM', 'ELECTRONS', 'IONS', 'CELL'), + input_namelists=('CONTROL', 'SYSTEM', 'ELECTRONS', 'IONS', 'CELL', + FCP_NAMELIST), input_cards=('ATOMIC_SPECIES', 'ATOMIC_POSITIONS', 'K_POINTS', 'CELL_PARAMETERS', 'ATOMIC_FORCES') ) if 'xml_file' in kwargs: - self._input['CONTROL']['input_xml_schema_file'] = u'\'{}\''.format(os.path.basename(kwargs['xml_file'])) + self._input['CONTROL']['input_xml_schema_file'] = "{!r}".format( + os.path.basename(kwargs['xml_file']) + ) def clear_input(self): super(PwInputConverter, self).clear_input() @@ -549,7 +576,9 @@ class PhononInputConverter(RawInputConverter): Convert to/from Fortran input for Phonon. """ PHONON_TEMPLATE_MAP = { - 'xq': ('qPointsSpecs', cards.get_qpoints_card, None), + 'xq': { + '$': ('qPointsSpecs', cards.get_qpoints_card, None), + }, 'scf_ph': { 'tr2_ph': "INPUTPH[tr2_ph]", 'niter_ph': "INPUTPH[niter_ph]", @@ -608,7 +637,11 @@ class PhononInputConverter(RawInputConverter): 'last_q': "INPUTPH[last_q]", 'start_irr': "INPUTPH[start_irr]", 'last_irr': "INPUTPH[last_irr]", - 'nat_todo': "INPUTPH[nat_todo]", + 'nat_todo': + { + '@natom': 'INPUTPH[nat_todo]', + '$': [(PH_NAT_TODO_CARD, cards.get_nat_todo_card, None)] + }, 'modenum': "INPUTPH[modenum]", 'only_init': "INPUTPH[only_init]", 'ldiag': "INPUTPH[ldiag]", @@ -637,9 +670,9 @@ class PhononInputConverter(RawInputConverter): }, 'q_points': { 'grid': { - 'nq1': "INPUTPH[nq1]", - 'nq2': "INPUTPH[nq2]", - 'nq3': "INPUTPH[nq3]" + '@nq1': "INPUTPH[nq1]", + '@nq2': "INPUTPH[nq2]", + '@nq3': "INPUTPH[nq3]" }, 'q_points_list': ('qPointsSpecs', cards.get_qpoints_card, None), 'nqs': ('qPointsSpecs', cards.get_qpoints_card, None) @@ -650,7 +683,7 @@ def __init__(self, *_kwargs): super(PhononInputConverter, self).__init__( conversion_maps_builder(self.PHONON_TEMPLATE_MAP), input_namelists=('INPUTPH',), - input_cards=('qPointsSpecs',) + input_cards=('qPointsSpecs', PH_NAT_TODO_CARD) ) @@ -670,6 +703,7 @@ class NebInputConverter(RawInputConverter): 'elasticConstMin': "PATH[k_min]", 'pathThreshold': "PATH[path_thr]", 'endImagesOptimizationFlag': "PATH[first_last_opt]", + 'minimumImageFlag': "PATH[minimum_image]", 'temperature': "PATH[temp_req]", 'climbingImage': [ "PATH[CI_scheme]", @@ -677,7 +711,7 @@ class NebInputConverter(RawInputConverter): ], 'useMassesFlag': "PATH[use_masses]", 'useFreezingFlag': "PATH[use_freezing]", - 'constantBiasFlag': "PATH[lfcpopt]", + 'constantBiasFlag': "PATH[lfcp]", 'targetFermiEnergy': "PATH[fcp_mu]", 'totChargeFirst': "PATH[fcp_tot_charge_first]", 'totChargeLast': "PATH[fcp_tot_charge_last]", @@ -711,8 +745,9 @@ def __init__(self, *_kwargs): def get_qe_input(self): """ - Overrides method in RawInputConverter because few lines in between the namelists are requested for - the NEB input. + Overrides method in RawInputConverter because few lines + in between the namelists are requested for the NEB input. + :return: a string containing the text input for NEB calculations """ qe_input = super(NebInputConverter, self).get_qe_input().split('\n') diff --git a/qeschema/documents.py b/qeschema/documents.py index 83b7e46..cf683c8 100644 --- a/qeschema/documents.py +++ b/qeschema/documents.py @@ -21,7 +21,7 @@ except ImportError: yaml = None -from .namespaces import +from .namespaces import XSD_NAMESPACE from .converters import RawInputConverter, PwInputConverter, PhononInputConverter, \ NebInputConverter, TdInputConverter, TdSpectrumInputConverter from .exceptions import XmlDocumentError diff --git a/qeschema/hdf5/__init__.py b/qeschema/hdf5/__init__.py new file mode 100644 index 0000000..1e7bc0f --- /dev/null +++ b/qeschema/hdf5/__init__.py @@ -0,0 +1,18 @@ +# +# Copyright (c), 2021, Quantum Espresso Foundation and SISSA (Scuola +# Internazionale Superiore di Studi Avanzati). All rights reserved. +# This file is distributed under the terms of the MIT License. See the +# file 'LICENSE' in the root directory of the present distribution, or +# http://opensource.org/licenses/MIT. +# +from .charge import read_charge_file_hdf5, get_minus_indexes, \ + get_charge_r, write_charge +from .readutils import get_wf_attributes, get_wavefunctions, \ + get_wfc_miller_indices, read_pseudo_file, create_header, \ + read_postqe_output_file, read_etotv + +__all__ = [ + 'read_charge_file_hdf5', 'get_minus_indexes', 'get_charge_r', 'write_charge', + 'get_wf_attributes', 'get_wavefunctions', 'get_wfc_miller_indices', + 'read_pseudo_file', 'create_header', 'read_postqe_output_file', 'read_etotv' +] diff --git a/qeschema/hdf5/charge.py b/qeschema/hdf5/charge.py new file mode 100644 index 0000000..42246c2 --- /dev/null +++ b/qeschema/hdf5/charge.py @@ -0,0 +1,162 @@ +# +# Copyright (c), 2021, Quantum Espresso Foundation and SISSA (Scuola +# Internazionale Superiore di Studi Avanzati). All rights reserved. +# This file is distributed under the terms of the MIT License. See the +# file 'LICENSE' in the root directory of the present distribution, or +# http://opensource.org/licenses/MIT. +# +import numpy as np + +try: + import h5py +except ImportError: + h5py = None + H5PY_ERR = 'h5py module is missing' + + +def read_charge_file_hdf5(filename): + """ + Reads a PW charge file in HDF5 format. + + :param filename: the name of the HDF5 file to read. + :return: a dictionary describing the content of file \ + keys=[nr, ngm_g, gamma_only, rhog_, MillerIndexes] + """ + + if not h5py: + raise ImportError(H5PY_ERR) + + with h5py.File(filename, "r") as h5f: + MI = h5f.get('MillerIndices')[:] + nr1 = 2 * max(abs(MI[:, 0])) + 1 + nr2 = 2 * max(abs(MI[:, 1])) + 1 + nr3 = 2 * max(abs(MI[:, 2])) + 1 + nr = np.array([nr1, nr2, nr3]) + res = dict(h5f.attrs.items()) + res.update({'MillInd': MI, 'nr_min': nr}) + rhog = h5f['rhotot_g'][:].reshape(res['ngm_g'], 2).dot([1.e0, 1.e0j]) + res['rhotot_g'] = rhog + if 'rhodiff_g' in h5f.keys(): + rhog = h5f['rhodiff_g'][:].reshape(res['ngm_g'], 2).dot([1.e0, 1.e0j]) + res['rhodiff_g'] = rhog + return res + + +def get_minus_indexes(g1, g2, g3): + """ + Used for getting the corresponding minus Miller Indexes. It is meant + to be used for converting Gamma Trick grids and is defined only for + the for i >=0, in the i =0 plan is defined only for j >=0 and when + i=0 j=0 k must be >=0. Out of this domain returns None. + + :param g1: rank 1 array containing first Miller Index + :param g2: rank 1 array containing second Miller Index + :param g3: rank 1 array containing third Miller Index + :return: a rank 2 array with dimension (ngm/2,3) containing mirrored Miller indexes + """ + + def scalar_func(i, j, k): + """ + scalar function to be vectorized + :param i: 1st Miller Index + :param j: 2nd + :param k: 3rd + :return: the mirrored mirror indexes + """ + if i > 0: + return -i, j, k + elif i == 0 and j > 0: + return 0, -j, k + elif i == 0 and j == 0 and k > 0: + return 0, 0, -k + else: + return i, j, k + + vector_func = np.vectorize(scalar_func) + + res = np.array(vector_func(g1, g2, g3)) + return res.transpose() + + +def get_charge_r(filename, nr=None): + """ + Reads a charge file written with QE in HDF5 format. nr = [nr1,nr2,nr3] (the dimensions of + the charge k-points grid) are given as parameter (taken for the xml output file by the caller). + + Notes: In the new format, the values of the charge in the reciprocal space are stored. + Besides, only the values of the charge > cutoff are stored, together with the Miller indexes. + Hence + """ + cdata = read_charge_file_hdf5(filename) + if nr is None: + nr1, nr2, nr3 = cdata['nrmin'] + else: + nr1, nr2, nr3 = nr + gamma_only = 'TRUE' in str(cdata['gamma_only']).upper() + + # Load the total charge + rho_temp = np.zeros([nr1, nr2, nr3], dtype=np.complex128) + for (i, j, k), rho in zip(cdata['MillInd'], cdata['rhotot_g']): + try: + rho_temp[i, j, k] = rho + except IndexError: + pass + + if gamma_only: + rhotot_g = cdata['rhotot_g'].conjugate() + MI = get_minus_indexes( + cdata['MillInd'][:, 0], cdata['MillInd'][:, 1], cdata['MillInd'][:, 2] + ) + print("MI", MI) + for (i, j, k), rho in zip(MI, rhotot_g): + try: + rho_temp[i, j, k] = rho + except IndexError: + pass + + rhotot_r = np.fft.ifftn(rho_temp) * nr1 * nr2 * nr3 + + # Read the charge difference spin up - spin down if present (for magnetic calculations) + if 'rhodiff_g' in cdata.keys(): + rho_temp = np.zeros([nr1, nr2, nr3], dtype=np.complex128) + for (i, j, k), rho in zip(cdata['MillInd'], cdata['rhodiff_g']): + try: + rho_temp[i, j, k] = rho + except IndexError: + pass + + if gamma_only: + rhodiff_g = cdata['rhodiff_g'].conjugate() + for (i, j, k), rho in zip(MI, rhodiff_g): + try: + rho_temp[i, j, k] = rho + except IndexError: + pass + + rhodiff_r = np.fft.ifftn(rho_temp) * nr1 * nr2 * nr3 + return rhotot_r.real, rhodiff_r.real + else: + return rhotot_r.real, None + + +def write_charge(filename, charge, header): + """ + Write the charge or another quantity calculated by postqe into a text file filename. + """ + + fout = open(filename, "w") + + # The header contains some information on the system, the grid nr, etc. + fout.write(header) + nr = charge.shape + count = 0 + # Loop with the order as in QE files + for z in range(0, nr[2]): + for y in range(0, nr[1]): + for x in range(0, nr[0]): + fout.write(" {:.9E}".format(charge[x, y, z])) + count += 1 + if count % 5 == 0: + fout.write("\n") + + fout.close() diff --git a/qeschema/hdf5/readutils.py b/qeschema/hdf5/readutils.py new file mode 100644 index 0000000..dc92a46 --- /dev/null +++ b/qeschema/hdf5/readutils.py @@ -0,0 +1,288 @@ +# +# Copyright (c), 2021, Quantum Espresso Foundation and SISSA (Scuola +# Internazionale Superiore di Studi Avanzati). All rights reserved. +# This file is distributed under the terms of the MIT License. See the +# file 'LICENSE' in the root directory of the present distribution, or +# http://opensource.org/licenses/MIT. +# +""" +A collection of functions for reading different files and quantities. +""" +import numpy as np +from xml.etree import ElementTree + +try: + import h5py +except ImportError: + h5py = None + H5PY_ERR = 'h5py module is missing' + + +# TODO update to the new format +def get_wf_attributes(filename): + """ + Read attributes from a wfc file. + + :param filename: the path to the wfc file + :return: a dictionary with all attributes included reciprocal vectors + """ + + if not h5py: + raise ImportError(H5PY_ERR) + + with h5py.File(filename, "r") as f: + res = dict(f.attrs) + mi_attrs = f.get('MillerIndices').attrs + bg = np.array(mi_attrs.get(x) for x in ['bg1', 'bg2', 'bg3']) + res.update({'bg': bg}) + return res + + +def get_wavefunctions(filename, start_band=None, stop_band=None): + """ + Returns a numpy array with the wave functions for bands from start_band to + stop_band. If not specified starts from 1st band and ends with last one. + Band numbering is Python style starts from 0.abs + + :param filename: path to the wfc file + :param start_band: first band to read, default first band in the file + :param stop_band: last band to read, default last band in the file + :return: a numpy array with shape [nbnd,npw] + """ + + if not h5py: + raise ImportError(H5PY_ERR) + + with h5py.File(filename, "r") as f: + igwx = f.attrs.get('igwx') + if start_band is None: + start_band = 0 + if stop_band is None: + stop_band = f.attrs.get('nbnd') + if stop_band == start_band: + stop_band = start_band + 1 + res = f.get('evc')[start_band:stop_band, :] + + res = np.asarray(x.reshape([igwx, 2]).dot([1.e0, 1.e0j]) for x in res[:]) + return res + + +def get_wfc_miller_indices(filename): + """ + Reads miller indices from the wfc file + + :param filename: path to the wfc file + :return: a np.array of integers with shape [igwx,3] + """ + + if not h5py: + raise ImportError(H5PY_ERR) + + with h5py.File(filename, "r") as f: + res = f.get("MillerIndices")[:, :] + return res + + +def read_pseudo_file(xmlfile): + """ + This function reads a pseudopotential XML-like file in the QE UPF format (text), + returning the content of each tag in a dictionary. The file is read in strings + and completed with a root UPF tag when it lacks, to avoids an XML syntax error. + + TODO: add support for UPF-schema files + """ + def iter_upf_file(): + """ + Creates an iterator over the lines of an UPF file, + inserting the root <UPF> tag when missing. + """ + with open(xmlfile, 'r') as f: + fake_root = None + for line in f: + if fake_root is not None: + line = line.replace('&input', '&input') + line = line.replace('&inputp', '&inputp') + yield line + else: + line = line.strip() + if line.startswith("<UPF") and line[4] in ('>', ' '): + yield line + fake_root = False + elif line: + yield "<UPF>" + yield line + fake_root = True + if fake_root is True: + yield "</UPF>" + + pseudo = {} + psroot = ElementTree.fromstringlist(list(iter_upf_file())) + + # PP_INFO + try: + pp_info = psroot.find('PP_INFO').text + except AttributeError: + pp_info = "" + try: + pp_input = psroot.find('PP_INFO/PP_INPUTFILE').text + except AttributeError: + pp_input = "" + pseudo.update(PP_INFO=dict(INFO=pp_info, PP_INPUT=pp_input)) + + # PP_HEADER + pseudo.update(PP_HEADER=dict(psroot.find('PP_HEADER').items())) + + # PP_MESH + pp_mesh = dict(psroot.find('PP_MESH').items()) + pp_r = np.array([float(x) for x in psroot.find('PP_MESH/PP_R').text.split()]) + pp_rab = np.array([float(x) for x in psroot.find('PP_MESH/PP_RAB').text.split()]) + pp_mesh.update(PP_R=pp_r, PP_RAB=pp_rab) + pseudo.update(PP_MESH=pp_mesh) + + # PP_LOCAL + node = psroot.find('PP_LOCAL') + if node is not None: + pp_local = np.array([x for x in map(float, node.text.split())]) + else: + pp_local = None + pseudo.update(PP_LOCAL=pp_local) + + # PP_RHOATOM + node = psroot.find('PP_RHOATOM') + if node is not None: + pp_rhoatom = np.array([v for v in map(float, node.text.split())]) + else: + pp_rhoatom = None + pseudo.update(PP_RHOATOM=pp_rhoatom) + + # PP_NONLOCAL + node = psroot.find('PP_NONLOCAL') + if node is not None: + betas = list() + dij = None + pp_aug = None + pp_q = None + for el in node: + if 'PP_BETA' in el.tag: + beta = dict(el.items()) + val = np.array(x for x in map(float, el.text.split())) + beta.update(beta=val) + betas.append(beta) + elif 'PP_DIJ' in el.tag: + text = '\n'.join(el.text.strip().split('\n')[1:]) + dij = np.array([x for x in map(float, text.split())]) + elif 'PP_AUGMENTATION' in el.tag: + pp_aug = dict(el.items()) + pp_qijl = list() + pp_qij = list() + for q in el: + if 'PP_QIJL' in q.tag: + qijl = dict(q.items()) + val = np.array(x for x in map(float, q.text.split())) + qijl.update(qijl=val) + pp_qijl.append(qijl) + elif 'PP_QIJ' in q.tag: + qij = dict(q.items()) + val = np.array(x for x in map(float, q.text.split())) + qij.update(qij=val) + pp_qij.append(qij) + elif q.tag == 'PP_Q': + pp_q = np.array(x for x in map(float, q.text.split())) + pp_aug.update(PP_QIJL=pp_qijl, PP_QIJ=pp_qij, PP_Q=pp_q) + pp_nonlocal = dict(PP_BETA=betas, PP_DIJ=dij, PP_AUGMENTATION=pp_aug) + else: + pp_nonlocal = None + + pseudo.update(PP_NONLOCAL=pp_nonlocal) + return pseudo + + +def create_header(prefix, nr, nr_smooth, ibrav, celldms, nat, ntyp, + atomic_species, atomic_positions): + """ + Creates the header lines for the output charge (or potential) text file as in pp.x. + The format is: + + system_prefix + fft_grid (nr1,nr2,nr3) fft_smooth (nr1,nr2,nr3) nat ntyp + ibrav celldms (6 real as in QE) + missing line with respect to pp.x + List of atoms + List of positions for each atom + """ + text = "# " + prefix + "\n" + text += "# {:8d} {:8d} {:8d} {:8d} {:8d} {:8d} {:8d} {:8d}\n".format( + nr[0], nr[1], nr[2], nr_smooth[0], nr_smooth[1], nr_smooth[2], nat, ntyp + ) + text += "# {:6d} {:8E} {:8E} {:8E} {:8E} {:8E} {:8E}\n".format(ibrav, celldms) + + # TODO This line is to be implemented + text += "# " + 4 "XXXX " + "\n" + + ityp = 1 + for typ in atomic_species: + text += "# {:4d} ".format(ityp) + typ["@name"] + "\n" + ityp += 1 + + ipos = 1 + for pos in atomic_positions: + text += "# {:4d} ".format(ipos) + coords = [float(x) for x in pos['$']] + text += " {:9E} {:9E} {:9E} ".format(coords) + text += pos["@name"] + "\n" + ipos += 1 + + return text + + +def read_postqe_output_file(filename): + """ + This function reads the output charge (or other quantity) as the output + format of postqe. + """ + tempcharge = [] + count = 0 + nr = np.zeros(3, dtype=int) + with open(filename, "r") as lines: + for line in lines: + words = line.split() + if count == 1: + nr[0] = int(words[1]) + nr[1] = int(words[2]) + nr[2] = int(words[3]) + if words[0] != '#': # skip the first lines beginning with # + for j in range(len(words)): # len(words)=5 except maybe for the last line + tempcharge.append(float(words[j])) + count += 1 + + charge = np.zeros((nr[0], nr[1], nr[2])) + count = 0 + + # Loops according to the order it is done in QE + for z in range(nr[2]): + for y in range(nr[1]): + for x in range(nr[0]): + charge[x, y, z] = tempcharge[count] + count += 1 + + return charge + + +def read_etotv(fname): + """ + Read cell volumes and the corresponding energies from input file fname* + (1st col, volumes, 2nd col energies). Units must be :math:`a.u.^3` and + :math:`Ryd/cell` + """ + Vx = [] + Ex = [] + + with open(fname, "r") as lines: + for line in lines: + linesplit = line.split() + V = float(linesplit[0]) + E = float(linesplit[1]) + Vx.append(V) + Ex.append(E) + + return np.array(Vx), np.array(Ex) diff --git a/qeschema/options.py b/qeschema/options.py index b6040fb..3ad75a8 100644 --- a/qeschema/options.py +++ b/qeschema/options.py @@ -209,31 +209,26 @@ def get_control_gdir(name, kwargs): def get_cell_dofree(name, kwargs): assert isinstance(name, str) - try: - fix_volume = kwargs['fix_volume'] - except KeyError: - fix_volume = False - try: - fix_area = kwargs['fix_area'] - except KeyError: - fix_area = False - try: - isotropic = kwargs['isotropic'] - except KeyError: - isotropic = False + cell_dofree_str = "cell_dofree = '%s'" + cell_dofree_all = 'all' + ret = [cell_dofree_str % cell_dofree_all] + + map_data = { + 'fix_volume': 'shape', + 'fix_area': '2Dshape', + 'fix_xy': '2Dxy', + 'isotropic': 'volume' + } + + if len(set(kwargs).intersection(map_data)) > 1: + logger.error("only one of %s can be true" % ', '.join(map_data)) + return ret - cell_dofree = "cell_dofree = 'all'" - if (fix_volume and fix_area) or (fix_volume and isotropic) or (fix_area and isotropic): - logger.error("only one of fix_volume fix_area and isotropic can be true") - return [cell_dofree] + for key, val in map_data.items(): + if kwargs.get(key): + return [cell_dofree_str % val] - if fix_volume: - cell_dofree = "cell_dofree = 'shape'" - if fix_area: - cell_dofree = "cell_dofree = '2Dshape'" - if isotropic: - cell_dofree = "cell_dofree = 'volume'" - return [cell_dofree] + return ret def neb_set_system_nat(name, kwargs): @@ -289,7 +284,7 @@ def set_lda_plus_u_flag(name, kwargs): for value in iter(related_data if isinstance(related_data, list) else [related_data]): if value.get('@label') != 'no Hubbard' and value['$'] > 0: - lines.append('lda_plus_u = .t.') + lines.append(f" {name} = .true.") break return lines diff --git a/qeschema/schemas/ph_xmlschema.xsd b/qeschema/schemas/ph_xmlschema.xsd index f8d0f39..21db2bd 100644 --- a/qeschema/schemas/ph_xmlschema.xsd +++ b/qeschema/schemas/ph_xmlschema.xsd @@ -89,14 +89,14 @@ Created by Mauro Palumbo <!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> <!-- INPUTPH TYPE--> <complexType name="inputPHType"> - <sequence> + <all> <element type="qes_ph:scf_phType" name="scf_ph" minOccurs="0" /> <element type="qes_ph:filesType" name="files" minOccurs="0" /> <element type="qes_ph:control_phType" name="control_ph" minOccurs="0" /> <element type="qes_ph:control_jobType" name="control_job" minOccurs="0" /> <element type="qes_ph:control_dielType" name="control_diel" minOccurs="0"/> <element type="qes_ph:control_qplotType" name="control_qplot" - minOccurs="0" /> + minOccurs="0" /> <element type="qes_ph:miscellaneaType" name="miscellanea" minOccurs="0" /> <!-- Irreproducible representation --> <element type="qes_ph:irr_reprType" name="irr_repr" minOccurs="0" /> @@ -107,8 +107,9 @@ Created by Mauro Palumbo <element type="qes_ph:lraman_optionsType" name="lraman_options" minOccurs="0" /> <!-- q-points mesh or list --> + <element name="xq" type="qes_ph:k_pointType" minOccurs="0"/> <element type="qes_ph:q_pointsType" name="q_points" minOccurs="0" /> - </sequence> + </all> </complexType> <!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> <complexType name="scf_phType"> @@ -173,7 +174,7 @@ Created by Mauro Palumbo <complexType name="miscellaneaType"> <sequence> <!-- if amass is not present, should be taked from pw in the code --> - <element type="qes_ph:amassType" name="amass" minOccurs="0" /> + <element type="qes_ph:amassType" name="amass" minOccurs="0" maxOccurs="unbounded"/> <element type="double" name="start_magnetization" minOccurs="0" /> <element type="nonNegativeInteger" name="verbosity" default="0" minOccurs="0" /> @@ -181,6 +182,7 @@ Created by Mauro Palumbo <element type="boolean" name="low_directory_check" default="false" minOccurs="0" /> <element type="boolean" name="nogg" default="false" minOccurs="0" /> + <element type="qes_ph:monkhorst_packType" name="nscf_MPgrid" minOccurs="0"/> </sequence> </complexType> <!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> @@ -202,9 +204,8 @@ Created by Mauro Palumbo <element type="positiveInteger" name="start_irr" default="1" minOccurs="0" /> <element type="positiveInteger" name="last_irr" default="3" - minOccurs="0" /> - <element type="nonNegativeInteger" name="nat_todo" default="0" - minOccurs="0" /> + minOccurs="0" /> + <element type="qes_ph:nat_todoType" name="nat_todo" minOccurs="0" maxOccurs="1"/> <element type="nonNegativeInteger" name="modenum" default="0" minOccurs="0" /> <element type="boolean" name="only_init" default="false" @@ -217,7 +218,9 @@ Created by Mauro Palumbo <complexType name="electron_phonon_optionsType"> <sequence> <element type="qes_ph:electron_phononType" name="electron_phonon" - minOccurs="0" /> + minOccurs="0" /> + <element type="double" name="sigma" default="0.02" minOccurs="0"/> + <element type="integer" name="nsigma" default="10" minOccurs="0"/> <element type="qes_ph:dvscf_starType" name="dvscf_star" minOccurs="0" /> <element type="qes_ph:drho_starType" name="drho_star" minOccurs="0" /> </sequence> @@ -270,10 +273,8 @@ Created by Mauro Palumbo <complexType name="q_pointsType"> <sequence> <!-- monkhorst_pack --> - <element type="qes_ph:q_monkhorst_packType" name="monkhorst_pack" - minOccurs="0" maxOccurs="1"/> + <element type="qes_ph:q_monkhorst_packType" name="grid" minOccurs="0" maxOccurs="1"/> <!-- monkhorst_pack grid, ex. 3 3 3 --> - <element type="positiveInteger" name="nq"/> <!-- Alternatively, a list of nqs q-points... (optionally with weight as attribute) --> <element type="nonNegativeInteger" name="nqs" default="1" minOccurs="0" /> <element type="qes_ph:q_points_listType" name="q_points_list" minOccurs="0" /> @@ -295,6 +296,14 @@ Created by Mauro Palumbo </extension> </simpleContent> </complexType> + <!-- nat_todo --> + <complexType name="nat_todoType"> + <sequence> + <element type="positiveInteger" name="atom" minOccurs="0" maxOccurs="unbounded"/> + </sequence> + <attribute type="nonNegativeInteger" name="natom"/> + </complexType> + <!-- end nat_todo --> <!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> <!-- k points, same as in qes schema --> <complexType name="k_points_IBZType"> diff --git a/qeschema/schemas/qes.xsd b/qeschema/schemas/qes.xsd index 176ecc1..3b7dd7d 100644 --- a/qeschema/schemas/qes.xsd +++ b/qeschema/schemas/qes.xsd @@ -287,6 +287,9 @@ Authors: Antonio Zambon, Paolo Giannozzi, Pietro Delugas <enumeration value="VDW-DF2-C09"/> <enumeration value="VDW-DF2-B86R"/> <enumeration value="RVV10"/> + <enumeration value="BEEF"/> + <enumeration value="SCAN"/> + <enumeration value="R2SCAN"/> </restriction> </simpleType> @@ -456,6 +459,7 @@ Authors: Antonio Zambon, Paolo Giannozzi, Pietro Delugas <element type="qes:basisSetItemType" name="fft_grid" minOccurs="0" /> <element type="qes:basisSetItemType" name="fft_smooth" minOccurs="0" /> <element type="qes:basisSetItemType" name="fft_box" minOccurs="0" /> + <element type="boolean" name="spline_ps" minOccurs="0" maxOccurs="1"/> </sequence> </complexType> <complexType name="basis_setType"> @@ -519,6 +523,7 @@ Authors: Antonio Zambon, Paolo Giannozzi, Pietro Delugas <enumeration value="davidson"/> <enumeration value="cg"/> <enumeration value="ppcg"/> + <enumeration value="paro"/> </restriction> </simpleType> <complexType name="k_points_IBZType"> @@ -588,9 +593,12 @@ Authors: Antonio Zambon, Paolo Giannozzi, Pietro Delugas <element type="double" name="pressure" default="0.0" /> <element type="double" name="wmass" minOccurs="0" /> <element type="double" name="cell_factor" minOccurs="0" /> - <element type="boolean" name="fix_volume" minOccurs="0" /> - <element type="boolean" name="fix_area" minOccurs="0" /> - <element type="boolean" name="isotropic" minOccurs="0" /> + <choice minOccurs="0" maxOccurs="1"> + <element type="boolean" name="fix_volume" /> + <element type="boolean" name="fix_area" /> + <element type="boolean" name="fix_xy" /> + <element type="boolean" name="isotropic" /> + </choice> <element type="qes:integerMatrixType" name="free_cell" minOccurs="0" /> </sequence> </complexType> diff --git a/qeschema/schemas/releases/ph-20210706.xsd b/qeschema/schemas/releases/ph-20210706.xsd new file mode 100644 index 0000000..3311196 --- /dev/null +++ b/qeschema/schemas/releases/ph-20210706.xsd @@ -0,0 +1,629 @@ +<?xml version="1.0"?> +<!-- GENERAL COMMENTS AND NOTES + + 0) qes-> quantum espresso schema created by Antonio Zambon Paol Giannozzi + 1) Date, time, general_infoType, parallel_infoType, statusType, closedType, + k_points_IBZType, atomic_speciesType and atomic_structureType from qes + in parallel_infoType the field "nimages" has been added + 2) more units for some quantities can be added + + OPEN POINTS + + 1) k1, k2, k3 in monkhorst_packType as in qes are boolean. Integer in + QuantumEspresso. Possible problem? + +Created by Mauro Palumbo +--> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> +<schema xmlns="http://www.w3.org/2001/XMLSchema" + xmlns:qes_ph="http://www.quantum-espresso.org/ns/qes/qes_ph_1.0" + targetNamespace="http://www.quantum-espresso.org/ns/qes/qes_ph_1.0"> + + <!-- ESPRESSO (root element) --> + <element name="espressoph" type="qes_ph:espressophType" /> + <complexType name="espressophType"> + <sequence> + <element type="qes_ph:general_infoType" name="general_info" minOccurs="0"/> + <element type="qes_ph:parallel_infoType" name="parallel_info" minOccurs="0"/> + <element type="qes_ph:inputPHType" name="inputPH" /> + <element type="qes_ph:outputPHType" name="outputPH" minOccurs="0" /> + <element type="qes_ph:statusType" name="status" minOccurs="0" /> + <element type="nonNegativeInteger" name="cputime" minOccurs="0" /> + <element type="qes_ph:closedType" name="closed" minOccurs="0" /> + </sequence> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <!-- GENERAL INFO, same as in qes schema --> + <complexType name="general_infoType"> + <sequence> + <element type="qes_ph:formatType" name="format"/> + <element type="qes_ph:creatorType" name="creator"/> + <element type="qes_ph:createdType" name="created"/> + <element type="string" name="job"/> + </sequence> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <complexType name="formatType"> + <simpleContent> + <extension base="string"> + <attribute type="string" name="NAME"/> + <attribute type="string" name="VERSION"/> + </extension> + </simpleContent> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <complexType name="creatorType"> + <simpleContent> + <extension base="string"> + <attribute type="string" name="NAME"/> + <attribute type="string" name="VERSION"/> + </extension> + </simpleContent> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <complexType name="createdType"> + <simpleContent> + <extension base="string"> + <attribute type="string" name="DATE"/> + <attribute type="string" name="TIME"/> + </extension> + </simpleContent> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <!-- PARALLEL INFO, same as in qes schema --> + <complexType name="parallel_infoType"> + <sequence> + <element type="positiveInteger" name="nprocs"/> + <element type="positiveInteger" name="nthreads"/> + <element type="positiveInteger" name="ntasks"/> + <element type="positiveInteger" name="nbgrp"/> + <element type="positiveInteger" name="npool"/> + <element type="positiveInteger" name="ndiag"/> + <element type="positiveInteger" name="nimages" minOccurs="0"/> + </sequence> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <!-- INPUTPH TYPE--> + <complexType name="inputPHType"> + <all> + <element type="qes_ph:scf_phType" name="scf_ph" minOccurs="0" /> + <element type="qes_ph:filesType" name="files" minOccurs="0" /> + <element type="qes_ph:control_phType" name="control_ph" minOccurs="0" /> + <element type="qes_ph:control_jobType" name="control_job" minOccurs="0" /> + <element type="qes_ph:control_dielType" name="control_diel" minOccurs="0"/> + <element type="qes_ph:control_qplotType" name="control_qplot" + minOccurs="0" /> + <element type="qes_ph:miscellaneaType" name="miscellanea" minOccurs="0" /> + <!-- Irreproducible representation --> + <element type="qes_ph:irr_reprType" name="irr_repr" minOccurs="0" /> + <!-- Electron-phonon options --> + <element type="qes_ph:electron_phonon_optionsType" + name="electron_phonon_options" minOccurs="0" /> + <!-- Raman options --> + <element type="qes_ph:lraman_optionsType" name="lraman_options" + minOccurs="0" /> + <!-- q-points mesh or list --> + <element name="xq" type="qes_ph:k_pointType" minOccurs="0"/> + <element type="qes_ph:q_pointsType" name="q_points" minOccurs="0" /> + </all> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <complexType name="scf_phType"> + <sequence> + <element type="double" name="tr2_ph" default="1E-12" minOccurs="0" /> + <element type="positiveInteger" name="niter_ph" default="100" + minOccurs="0" /> + <element type="double" name="alpha_mix" default="0.7" minOccurs="0" /> + <element type="positiveInteger" name="nmix_ph" default="4" + minOccurs="0" /> + </sequence> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <complexType name="filesType"> + <sequence> + <element type="string" name="prefix" default="pwscf" minOccurs="0" /> + <element type="string" name="outdir" default="./" minOccurs="0" /> + <element type="string" name="fildyn" default="matdyn" minOccurs="0" /> + <element type="string" name="fildrho" default="" minOccurs="0" /> + <element type="string" name="fildvscf" default="" minOccurs="0" /> + <element type="boolean" name="lqdir" default="false" minOccurs="0" /> + </sequence> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <complexType name="control_phType"> + <sequence> + <element type="boolean" name="ldisp" default="false" minOccurs="0" /> + <element type="boolean" name="epsil" default="false" minOccurs="0" /> + <element type="boolean" name="trans" default="true" minOccurs="0" /> + <element type="boolean" name="zeu" default="false" minOccurs="0" /> + <element type="boolean" name="zue" default="false" minOccurs="0" /> + <element type="boolean" name="elop" default="false" minOccurs="0" /> + <element type="boolean" name="fpol" default="false" minOccurs="0" /> + <element type="boolean" name="lraman" default="false" minOccurs="0" /> + <element type="boolean" name="search_sym" default="true" minOccurs="0"/> + </sequence> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <complexType name="control_jobType"> + <sequence> + <element type="boolean" name="recover" default="false" minOccurs="0" /> + <element type="double" name="max_seconds" default="1E7" minOccurs="0" /> + </sequence> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <complexType name="control_dielType"> + <sequence> + <element type="boolean" name="lrpa" default="false" minOccurs="0" /> + <element type="boolean" name="lnoloc" default="false" minOccurs="0" /> + </sequence> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <complexType name="control_qplotType"> + <sequence> + <element type="boolean" name="qplot" default="false" minOccurs="0" /> + <element type="boolean" name="q2d" default="false" minOccurs="0" /> + <element type="boolean" name="q_in_band_form" default="false" + minOccurs="0" /> + </sequence> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <complexType name="miscellaneaType"> + <sequence> + <!-- if amass is not present, should be taked from pw in the code --> + <element type="qes_ph:amassType" name="amass" minOccurs="0" maxOccurs="unbounded"/> + <element type="double" name="start_magnetization" minOccurs="0" /> + <element type="nonNegativeInteger" name="verbosity" default="0" + minOccurs="0" /> + <element type="boolean" name="reduce_io" default="false" minOccurs="0"/> + <element type="boolean" name="low_directory_check" default="false" + minOccurs="0" /> + <element type="boolean" name="nogg" default="false" minOccurs="0" /> + <element type="qes_ph:monkhorst_packType" name="nscf_MPgrid" minOccurs="0"/> + </sequence> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <complexType name="amassType"> + <simpleContent> + <extension base="double"> + <attribute type="positiveInteger" name="atom"/> + </extension> + </simpleContent> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <complexType name="irr_reprType"> + <sequence> + <sequence> + <element type="positiveInteger" name="start_q" default="1" + minOccurs="0" /> + <element type="positiveInteger" name="last_q" default="1" + minOccurs="0" /> + <element type="positiveInteger" name="start_irr" default="1" + minOccurs="0" /> + <element type="positiveInteger" name="last_irr" default="3" + minOccurs="0" /> + <element type="nonNegativeInteger" name="nat_todo" default="0" + minOccurs="0" /> + <element type="nonNegativeInteger" name="modenum" default="0" + minOccurs="0" /> + <element type="boolean" name="only_init" default="false" + minOccurs="0" /> + <element type="boolean" name="ldiag" default="false" minOccurs="0" /> + </sequence> + </sequence> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <complexType name="electron_phonon_optionsType"> + <sequence> + <element type="qes_ph:electron_phononType" name="electron_phonon" + minOccurs="0" /> + <element type="double" name="sigma" default="0.02" minOccurs="0"/> + <element type="integer" name="nsigma" default="10" minOccurs="0"/> + <element type="qes_ph:dvscf_starType" name="dvscf_star" minOccurs="0" /> + <element type="qes_ph:drho_starType" name="drho_star" minOccurs="0" /> + </sequence> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <simpleType name="electron_phononType"> + <restriction base="string"> + <enumeration value="" /> + <enumeration value="simple" /> + <enumeration value="interpolated" /> + </restriction> + </simpleType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <complexType name="dvscf_starType"> + <sequence> + <element type="boolean" name="open" default="false" + minOccurs="0" /> + <element type="string" name="dir" minOccurs="0" /> + <element type="string" name="ext" default="dvscf" + minOccurs="0" /> + <element type="string" name="basis" default="cartesian" + minOccurs="0" /> + <element type="boolean" name="pat" default="true" + minOccurs="0" /> + </sequence> + </complexType> + <!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <complexType name="drho_starType"> + <sequence> + <element type="boolean" name="open" default="false" + minOccurs="0" /> + <element type="string" name="dir" minOccurs="0" /> + <element type="string" name="ext" default="drho" + minOccurs="0" /> + <element type="string" name="basis" default="modes" + minOccurs="0" /> + <element type="boolean" name="pat" default="true" + minOccurs="0" /> + </sequence> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <complexType name="lraman_optionsType"> + <sequence> + <element type="double" name="eth_rps" default="1E-9" minOccurs="0" /> + <element type="double" name="eth_ns" default="1E-12" minOccurs="0" /> + <element type="double" name="dek" default="1E-3" minOccurs="0" /> + </sequence> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <complexType name="q_pointsType"> + <sequence> + <!-- monkhorst_pack --> + <element type="qes_ph:q_monkhorst_packType" name="grid" minOccurs="0" maxOccurs="1"/> + <!-- monkhorst_pack grid, ex. 3 3 3 --> + <!-- Alternatively, a list of nqs q-points... (optionally with weight as attribute) --> + <element type="nonNegativeInteger" name="nqs" default="1" minOccurs="0" /> + <element type="qes_ph:q_points_listType" name="q_points_list" minOccurs="0" /> + </sequence> + </complexType> + <!-- q points, similar but not the same as in qes schema (here no offset) --> + <complexType name="q_points_listType"> + <sequence> + <element type="qes_ph:k_pointType" name="q_point" maxOccurs="unbounded" + minOccurs="0"/> + </sequence> + </complexType> + <complexType name="q_monkhorst_packType"> + <simpleContent> + <extension base="string"> + <attribute type="positiveInteger" name="nq1"/> + <attribute type="positiveInteger" name="nq2"/> + <attribute type="positiveInteger" name="nq3"/> + </extension> + </simpleContent> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <!-- k points, same as in qes schema --> + <complexType name="k_points_IBZType"> + <sequence> + <element type="qes_ph:monkhorst_packType" name="monkhorst_pack" minOccurs="0" + maxOccurs="1"/> + <element type="positiveInteger" name="nk"/> + <element type="qes_ph:k_pointType" name="k_point" maxOccurs="unbounded" + minOccurs="0"/> + </sequence> + </complexType> + <complexType name="monkhorst_packType"> + <simpleContent> + <extension base="string"> + <attribute type="positiveInteger" name="nk1"/> + <attribute type="positiveInteger" name="nk2"/> + <attribute type="positiveInteger" name="nk3"/> + <attribute type="boolean" name="k1"/> + <attribute type="boolean" name="k2"/> + <attribute type="boolean" name="k3"/> + </extension> + </simpleContent> + </complexType> + <complexType name="k_pointType"> + <simpleContent> + <extension base="qes_ph:d3vectorType"> + <attribute type="double" name="weight" use="optional"/> + </extension> + </simpleContent> + </complexType> + <!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <!-- CLOSED, same as in qes schema --> + <complexType name="closedType"> + <simpleContent> + <extension base="string"> + <attribute type="string" name="DATE"/> + <attribute type="string" name="TIME"/> + </extension> + </simpleContent> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <!-- STATUS, same as in qes schema --> + <simpleType name="statusType"> + <restriction base="unsignedByte"> + <enumeration value="0" /> + <enumeration value="1" /> + <enumeration value="2" /> + <enumeration value="3" /> + <enumeration value="255"/> + </restriction> + </simpleType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <!-- OUTPUTPH TYPE--> + <complexType name="outputPHType"> + <sequence> + <!-- As in qes --> + <element type="qes_ph:atomic_speciesType" name="atomic_species"/> + <element type="qes_ph:atomic_structureType" name="atomic_structure"/> + <!-- In addition to qes --> + <element type="qes_ph:unit_cell_volumeType" name="unit_cell_volume"/> + <element type="nonNegativeInteger" name="ibrav" + minOccurs="0" /> + <element type="qes_ph:cell_dimensionsType" name="cell_dimensions" + minOccurs="0" /> + <element type="double" name="start_magnetization" minOccurs="0" /> + <element type="nonNegativeInteger" name="number_of_q" minOccurs="0"/> + <element type="nonNegativeInteger" name="number_of_irr_q" + minOccurs="0"/> + <element type="qes_ph:irr_repr_listType" name="irr_repr_list" + minOccurs="0" /> + <element type="qes_ph:dynamical_matType" name="dynamical_mat" + minOccurs="0" /> + <element type="qes_ph:force_constantsType" name="force_constants" + minOccurs="0" /> + <element type="qes_ph:frequenciesType" name="frequencies" + minOccurs="0"/> + </sequence> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <!-- Unit cell volume with unit --> + <complexType name="unit_cell_volumeType"> + <simpleContent> + <extension base="double"> + <attribute type="string" name="unit" use="required"/> + </extension> + </simpleContent> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <!-- Cell dimensions --> + <simpleType name="cell_dimensionsType"> + <restriction> + <simpleType> + <list itemType="double"/> + </simpleType> + <length value="6"/> + </restriction> + </simpleType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <!-- Irr. representations --> + <complexType name="irr_repr_listType"> + <sequence> + <element type="qes_ph:repr_at_qType" name="repr_at_q" maxOccurs="unbounded" /> + </sequence> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <complexType name="repr_at_qType"> + <sequence> + <element type="qes_ph:d3vectorType" name="q_point" /> + <element type="qes_ph:reprType" name="repr" maxOccurs="unbounded" /> + </sequence> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <complexType name="reprType"> + <sequence> + <element type="qes_ph:modeType" name="mode" maxOccurs="unbounded" /> + </sequence> + <attribute type="nonNegativeInteger" name="number" /> + <attribute type="nonNegativeInteger" name="n_modes" /> + <attribute type="string" name="descr" /> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <complexType name="modeType"> + <sequence> + <element type="qes_ph:displacementType" name="displacement" maxOccurs="unbounded"/> + </sequence> + <attribute type="nonNegativeInteger" name="number" /> + </complexType> + +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <!-- Dynamical matrix --> + <complexType name="dynamical_matType"> + <sequence> + <element type="qes_ph:D_elements_symmequivType" + name="D_elements_symmequiv" maxOccurs="unbounded" /> + </sequence> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <!-- Dynamical matrix element (3x3 matrix at a given q point)--> + <complexType name="D_elements_symmequivType"> + <sequence> + <element type="qes_ph:D_qType" name="D_q" maxOccurs="unbounded" /> + </sequence> + <attribute type="nonNegativeInteger" name="number_of_equiv_q" /> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <!-- Dynamical matrix (3x3) at a given q point)--> + <complexType name="D_qType"> + <sequence> + <element type="qes_ph:d3vectorType" name="q_point" /> + <element type="qes_ph:phi_qType" name="phi_q" maxOccurs="unbounded" /> + </sequence> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <!-- Dynamical matrix element (3x3 matrix at a given q point)--> + <complexType name="phi_qType"> + <simpleContent> + <extension base="qes_ph:d3complexDType"> + <attribute type="positiveInteger" name="at1" /> + <attribute type="positiveInteger" name="at2" /> + </extension> + </simpleContent> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <!-- Frequencies --> + <complexType name="frequenciesType"> + <sequence> + <element type="qes_ph:frequencyType" name="frequency" minOccurs="0" + maxOccurs="unbounded"/> + </sequence> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <!-- frequency type, each containing omega (possibly in different units) + and a displacement --> + <complexType name="frequencyType"> + <sequence> + <!-- two frequencies because they can be in cm-1 and THz--> + <element type="qes_ph:omegaType" name="omega" maxOccurs="2"/> + <element type="qes_ph:displacementType" name="displacement" maxOccurs="unbounded"/> + </sequence> + <attribute type="integer" name="index" use="required"/> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <!-- freq type, unit as attribute (THz and cm-1) --> + <complexType name="omegaType"> + <simpleContent> + <extension base="double"> + <attribute type="string" name="unit" use="required"/> + </extension> + </simpleContent> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <complexType name="displacementType"> + <sequence> + <element type="qes_ph:disp_x_y_zType" name="x" maxOccurs="1"/> + <element type="qes_ph:disp_x_y_zType" name="y" maxOccurs="1"/> + <element type="qes_ph:disp_x_y_zType" name="z" maxOccurs="1"/> + </sequence> + <attribute type="integer" name="atom" use="required"/> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <!-- displacement type --> + <simpleType name="disp_x_y_zType"> + <restriction> + <simpleType> + <list itemType="double"/> + </simpleType> + <length value="2"/> + </restriction> + </simpleType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <!-- Force constants--> + <complexType name="force_constantsType"> + <sequence> + <element type="qes_ph:phi_rType" name="phi_r" maxOccurs="unbounded" /> + </sequence> + <attribute type="positiveInteger" name="nq1" /> + <attribute type="positiveInteger" name="nq2" /> + <attribute type="positiveInteger" name="nq3" /> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <!-- Force constant element (3x3 matrix at a given q point)--> + <complexType name="phi_rType"> + <sequence> + <element type="qes_ph:phi_r_elType" name="phi_r_el" + maxOccurs="unbounded" /> + </sequence> + <attribute type="positiveInteger" name="at1" /> + <attribute type="positiveInteger" name="at2" /> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <!-- Force constant element (3x3 matrix at a given q point)--> + <complexType name="phi_r_elType"> + <simpleContent> + <extension base="double"> + <attribute type="positiveInteger" name="m1" /> + <attribute type="positiveInteger" name="m2" /> + <attribute type="positiveInteger" name="m3" /> + </extension> + </simpleContent> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <!-- Input and Output types as in qes --> + <complexType name="atomic_speciesType"> + <sequence> + <element type="qes_ph:specieType" name="specie" maxOccurs="unbounded"/> + </sequence> + <attribute type="positiveInteger" name="ntyp"/> + </complexType> + <complexType name="specieType"> + <sequence> + <element type="double" name="mass" minOccurs="0" /> + <element type="string" name="pseudo_file"/> + <element type="double" name="starting_magnetization" minOccurs="0"/> + </sequence> + <attribute type="string" name="name"/> + </complexType> + <complexType name="atomic_structureType"> + <sequence> + <element type="qes_ph:atomic_positionsType" name="atomic_positions" + minOccurs="0" /> + <element type="qes_ph:wyckoff_positionsType" name="wyckoff_positions" + minOccurs="0" /> + <element type="qes_ph:cellType" name="cell"/> + </sequence> + <attribute type="positiveInteger" name="nat" use="optional"/> + <attribute type="double" name="alat" use="optional"/> + </complexType> + <complexType name="atomic_positionsType"> + <sequence> + <element type="qes_ph:atomType" name="atom" maxOccurs="unbounded" /> + </sequence> + </complexType> + <complexType name="atomType"> + <simpleContent> + <extension base="qes_ph:d3vectorType"> + <attribute type="string" name="name" /> + <attribute type="string" name="position" use="optional" /> + </extension> + </simpleContent> + </complexType> + <complexType name="wyckoff_positionsType"> + <sequence> + <element type="qes_ph:atomType" name="atom"/> + </sequence> + <attribute type="unsignedByte" name="space_group" use="optional"/> + <attribute type="string" name="more_options" use="optional"/> + </complexType> + <complexType name="cellType"> + <sequence> + <element type="qes_ph:d3vectorType" name="a1"/> + <element type="qes_ph:d3vectorType" name="a2"/> + <element type="qes_ph:d3vectorType" name="a3"/> + </sequence> + </complexType> + + <!-- 3x3 real matrix type --> + <simpleType name="d3realDType"> + <restriction> + <simpleType> + <list itemType="double"/> + </simpleType> + <length value="9"/> + </restriction> + </simpleType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <!-- 3x3 complex matrix type --> + <simpleType name="d3complexDType"> + <restriction> + <simpleType> + <list itemType="double"/> + </simpleType> + <length value="18"/> + </restriction> + </simpleType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <!-- FPMD TYPES (ref: quantum-simulation.org) --> + <simpleType name="d3vectorType"> + <restriction> + <simpleType> + <list itemType="double"/> + </simpleType> + <length value="3"/> + </restriction> + </simpleType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> +</schema> diff --git a/requirements-dev.txt b/requirements-dev.txt index 52d8698..8b6d1b9 100644 --- a/requirements-dev.txt +++ b/requirements-dev.txt @@ -1,9 +1,12 @@ # Requirements for setup a development environment for the qeschema package. setuptools tox +flake8 coverage xmlschema>=1.3.0 pyyaml +numpy +h5py Sphinx sphinx_rtd_theme -e . diff --git a/setup.py b/setup.py index 216c1b5..86d6604 100644 --- a/setup.py +++ b/setup.py @@ -15,8 +15,8 @@ setup( name='qeschema', - version='1.1.0', - install_requires=['xmlschema>=1.3.0', 'pyyaml'], + version='1.2.0', + install_requires=['xmlschema>=1.3.0', 'pyyaml', 'numpy'], packages=['qeschema'], package_data={'qeschema': ['schemas/*.xsd']}, scripts = ['scripts/xml2qeinput.py', 'scripts/yaml2qeinput.py'], @@ -37,8 +37,8 @@ 'Programming Language :: Python :: 3.7', 'Programming Language :: Python :: 3.8', 'Programming Language :: Python :: 3.9', + 'Programming Language :: Python :: 3.10', 'Programming Language :: Python :: Implementation :: CPython', - 'Programming Language :: Python :: Implementation :: PyPy', 'Topic :: Scientific/Engineering :: Physics', 'Topic :: Utilities', ] diff --git a/tox.ini b/tox.ini index 07297d1..ae9347a 100644 --- a/tox.ini +++ b/tox.ini @@ -4,7 +4,7 @@ # and then run "tox" from this directory. [tox] -envlist = py{36,37,38,39}, pypy3, docs, flake8, coverage +envlist = py{36,37,38,39,310}, docs, flake8, coverage skip_missing_interpreters = true toxworkdir = {homedir}/.tox/qeschema @@ -12,6 +12,8 @@ toxworkdir = {homedir}/.tox/qeschema deps = xmlschema>=1.3.0 pyyaml + numpy + h5py docs: Sphinx docs: sphinx_rtd_theme flake8: flake8 @@ -30,7 +32,7 @@ max-line-length = 119 [testenv:flake8] commands = - flake8 --ignore=F401,F403,F405,F811,F821 qeschema + flake8 qeschema [testenv:coverage] commands =	Propagate dftd3_version, dftd3_threebody from XML to .in Also fix a test.		2021-11-30T14:39:50	[]	[]
QEF/qeschema	QEF__qeschema-12	3185274c496dd4a5ef657b6718cb90fe08999765	diff --git a/qeschema/cards.py b/qeschema/cards.py index 82ef609..cc055e9 100644 --- a/qeschema/cards.py +++ b/qeschema/cards.py @@ -233,7 +233,6 @@ def get_cell_parameters_card(name, kwargs): except KeyError: logger.error("Missing required arguments when building CELL_PARAMETERS card!") return [] - # Add cell parameters card cells = atomic_structure.get('cell', {}) if cells: @@ -262,7 +261,7 @@ def get_qpoints_card(name, kwargs): qplot = kwargs.get('qplot', False) if not qplot and not ldisp: try: - xq = kwargs['xq_dir'] + xq = kwargs['xq'] except KeyError: xq = [0.e0, 0.e0, 0.e0] line = "{:6.4f} {:8.4f} {:8.4f}".format(xq[0], xq[1], xq[2]) diff --git a/qeschema/converters.py b/qeschema/converters.py index 4e7fae4..73691a6 100644 --- a/qeschema/converters.py +++ b/qeschema/converters.py @@ -327,7 +327,7 @@ class PwInputConverter(RawInputConverter): '$': [ ('SYSTEM[ibrav]', options.set_ibrav_to_zero, None), ("ATOMIC_POSITIONS", cards.get_atomic_positions_cell_card, None), - ("CELL_PARAMETERS", cards.get_cell_parameters_card, None) + ("CELL_PARAMETERS", cards.get_cell_parameters_card, None) ], 'atomic_positions': ('ATOMIC_FORCES', cards.get_atomic_forces_card, None), }, @@ -349,7 +349,8 @@ class PwInputConverter(RawInputConverter): 'dftU': { 'lda_plus_u_kind': 'SYSTEM[lda_plus_u_kind]', 'Hubbard_U': { - '$': ('SYSTEM[Hubbard_U]', options.get_specie_related_values, None), + '$': [('SYSTEM[Hubbard_U]', options.get_specie_related_values, None), + ('SYSTEM[lda_plus_u]',options.set_lda_plus_u_flag,None)] }, 'Hubbard_J0': { '$': ('SYSTEM[Hubbard_J0]', options.get_specie_related_values, None), @@ -554,7 +555,9 @@ class PhononInputConverter(RawInputConverter): Convert to/from Fortran input for Phonon. """ PHONON_TEMPLATE_MAP = { - 'xq': ('qPointsSpecs', cards.get_qpoints_card, None), + 'xq': { + '$': ('qPointsSpecs', cards.get_qpoints_card, None), + }, 'scf_ph': { 'tr2_ph': "INPUTPH[tr2_ph]", 'niter_ph': "INPUTPH[niter_ph]", @@ -642,9 +645,9 @@ class PhononInputConverter(RawInputConverter): }, 'q_points': { 'grid': { - 'nq1': "INPUTPH[nq1]", - 'nq2': "INPUTPH[nq2]", - 'nq3': "INPUTPH[nq3]" + '@nq1': "INPUTPH[nq1]", + '@nq2': "INPUTPH[nq2]", + '@nq3': "INPUTPH[nq3]" }, 'q_points_list': ('qPointsSpecs', cards.get_qpoints_card, None), 'nqs': ('qPointsSpecs', cards.get_qpoints_card, None) diff --git a/qeschema/options.py b/qeschema/options.py index b6040fb..e27f45b 100644 --- a/qeschema/options.py +++ b/qeschema/options.py @@ -126,6 +126,7 @@ def set_ibrav_to_zero(name, _kwargs): return [line] + def get_system_eamp(name, kwargs): try: electric_potential = kwargs['electric_potential'] @@ -289,7 +290,7 @@ def set_lda_plus_u_flag(name, *kwargs): for value in iter(related_data if isinstance(related_data, list) else [related_data]): if value.get('@label') != 'no Hubbard' and value['$'] > 0: - lines.append('lda_plus_u = .t.') + lines.append(f" {name} = .true.") break return lines diff --git a/qeschema/schemas/ph_xmlschema.xsd b/qeschema/schemas/ph_xmlschema.xsd index f8d0f39..3311196 100644 --- a/qeschema/schemas/ph_xmlschema.xsd +++ b/qeschema/schemas/ph_xmlschema.xsd @@ -89,14 +89,14 @@ Created by Mauro Palumbo <!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> <!-- INPUTPH TYPE--> <complexType name="inputPHType"> - <sequence> + <all> <element type="qes_ph:scf_phType" name="scf_ph" minOccurs="0" /> <element type="qes_ph:filesType" name="files" minOccurs="0" /> <element type="qes_ph:control_phType" name="control_ph" minOccurs="0" /> <element type="qes_ph:control_jobType" name="control_job" minOccurs="0" /> <element type="qes_ph:control_dielType" name="control_diel" minOccurs="0"/> <element type="qes_ph:control_qplotType" name="control_qplot" - minOccurs="0" /> + minOccurs="0" /> <element type="qes_ph:miscellaneaType" name="miscellanea" minOccurs="0" /> <!-- Irreproducible representation --> <element type="qes_ph:irr_reprType" name="irr_repr" minOccurs="0" /> @@ -107,8 +107,9 @@ Created by Mauro Palumbo <element type="qes_ph:lraman_optionsType" name="lraman_options" minOccurs="0" /> <!-- q-points mesh or list --> + <element name="xq" type="qes_ph:k_pointType" minOccurs="0"/> <element type="qes_ph:q_pointsType" name="q_points" minOccurs="0" /> - </sequence> + </all> </complexType> <!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> <complexType name="scf_phType"> @@ -173,7 +174,7 @@ Created by Mauro Palumbo <complexType name="miscellaneaType"> <sequence> <!-- if amass is not present, should be taked from pw in the code --> - <element type="qes_ph:amassType" name="amass" minOccurs="0" /> + <element type="qes_ph:amassType" name="amass" minOccurs="0" maxOccurs="unbounded"/> <element type="double" name="start_magnetization" minOccurs="0" /> <element type="nonNegativeInteger" name="verbosity" default="0" minOccurs="0" /> @@ -181,6 +182,7 @@ Created by Mauro Palumbo <element type="boolean" name="low_directory_check" default="false" minOccurs="0" /> <element type="boolean" name="nogg" default="false" minOccurs="0" /> + <element type="qes_ph:monkhorst_packType" name="nscf_MPgrid" minOccurs="0"/> </sequence> </complexType> <!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> @@ -217,7 +219,9 @@ Created by Mauro Palumbo <complexType name="electron_phonon_optionsType"> <sequence> <element type="qes_ph:electron_phononType" name="electron_phonon" - minOccurs="0" /> + minOccurs="0" /> + <element type="double" name="sigma" default="0.02" minOccurs="0"/> + <element type="integer" name="nsigma" default="10" minOccurs="0"/> <element type="qes_ph:dvscf_starType" name="dvscf_star" minOccurs="0" /> <element type="qes_ph:drho_starType" name="drho_star" minOccurs="0" /> </sequence> @@ -270,10 +274,8 @@ Created by Mauro Palumbo <complexType name="q_pointsType"> <sequence> <!-- monkhorst_pack --> - <element type="qes_ph:q_monkhorst_packType" name="monkhorst_pack" - minOccurs="0" maxOccurs="1"/> + <element type="qes_ph:q_monkhorst_packType" name="grid" minOccurs="0" maxOccurs="1"/> <!-- monkhorst_pack grid, ex. 3 3 3 --> - <element type="positiveInteger" name="nq"/> <!-- Alternatively, a list of nqs q-points... (optionally with weight as attribute) --> <element type="nonNegativeInteger" name="nqs" default="1" minOccurs="0" /> <element type="qes_ph:q_points_listType" name="q_points_list" minOccurs="0" /> diff --git a/qeschema/schemas/releases/ph-20210706.xsd b/qeschema/schemas/releases/ph-20210706.xsd new file mode 100644 index 0000000..3311196 --- /dev/null +++ b/qeschema/schemas/releases/ph-20210706.xsd @@ -0,0 +1,629 @@ +<?xml version="1.0"?> +<!-- GENERAL COMMENTS AND NOTES + + 0) qes-> quantum espresso schema created by Antonio Zambon Paol Giannozzi + 1) Date, time, general_infoType, parallel_infoType, statusType, closedType, + k_points_IBZType, atomic_speciesType and atomic_structureType from qes + * in parallel_infoType the field "nimages" has been added + 2) more units for some quantities can be added + + OPEN POINTS + + 1) k1, k2, k3 in monkhorst_packType as in qes are boolean. Integer in + QuantumEspresso. Possible problem? + +Created by Mauro Palumbo +--> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> +<schema xmlns="http://www.w3.org/2001/XMLSchema" + xmlns:qes_ph="http://www.quantum-espresso.org/ns/qes/qes_ph_1.0" + targetNamespace="http://www.quantum-espresso.org/ns/qes/qes_ph_1.0"> + + <!-- ESPRESSO (root element) --> + <element name="espressoph" type="qes_ph:espressophType" /> + <complexType name="espressophType"> + <sequence> + <element type="qes_ph:general_infoType" name="general_info" minOccurs="0"/> + <element type="qes_ph:parallel_infoType" name="parallel_info" minOccurs="0"/> + <element type="qes_ph:inputPHType" name="inputPH" /> + <element type="qes_ph:outputPHType" name="outputPH" minOccurs="0" /> + <element type="qes_ph:statusType" name="status" minOccurs="0" /> + <element type="nonNegativeInteger" name="cputime" minOccurs="0" /> + <element type="qes_ph:closedType" name="closed" minOccurs="0" /> + </sequence> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <!-- GENERAL INFO, same as in qes schema --> + <complexType name="general_infoType"> + <sequence> + <element type="qes_ph:formatType" name="format"/> + <element type="qes_ph:creatorType" name="creator"/> + <element type="qes_ph:createdType" name="created"/> + <element type="string" name="job"/> + </sequence> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <complexType name="formatType"> + <simpleContent> + <extension base="string"> + <attribute type="string" name="NAME"/> + <attribute type="string" name="VERSION"/> + </extension> + </simpleContent> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <complexType name="creatorType"> + <simpleContent> + <extension base="string"> + <attribute type="string" name="NAME"/> + <attribute type="string" name="VERSION"/> + </extension> + </simpleContent> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <complexType name="createdType"> + <simpleContent> + <extension base="string"> + <attribute type="string" name="DATE"/> + <attribute type="string" name="TIME"/> + </extension> + </simpleContent> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <!-- PARALLEL INFO, same as in qes schema --> + <complexType name="parallel_infoType"> + <sequence> + <element type="positiveInteger" name="nprocs"/> + <element type="positiveInteger" name="nthreads"/> + <element type="positiveInteger" name="ntasks"/> + <element type="positiveInteger" name="nbgrp"/> + <element type="positiveInteger" name="npool"/> + <element type="positiveInteger" name="ndiag"/> + <element type="positiveInteger" name="nimages" minOccurs="0"/> + </sequence> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <!-- INPUTPH TYPE--> + <complexType name="inputPHType"> + <all> + <element type="qes_ph:scf_phType" name="scf_ph" minOccurs="0" /> + <element type="qes_ph:filesType" name="files" minOccurs="0" /> + <element type="qes_ph:control_phType" name="control_ph" minOccurs="0" /> + <element type="qes_ph:control_jobType" name="control_job" minOccurs="0" /> + <element type="qes_ph:control_dielType" name="control_diel" minOccurs="0"/> + <element type="qes_ph:control_qplotType" name="control_qplot" + minOccurs="0" /> + <element type="qes_ph:miscellaneaType" name="miscellanea" minOccurs="0" /> + <!-- Irreproducible representation --> + <element type="qes_ph:irr_reprType" name="irr_repr" minOccurs="0" /> + <!-- Electron-phonon options --> + <element type="qes_ph:electron_phonon_optionsType" + name="electron_phonon_options" minOccurs="0" /> + <!-- Raman options --> + <element type="qes_ph:lraman_optionsType" name="lraman_options" + minOccurs="0" /> + <!-- q-points mesh or list --> + <element name="xq" type="qes_ph:k_pointType" minOccurs="0"/> + <element type="qes_ph:q_pointsType" name="q_points" minOccurs="0" /> + </all> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <complexType name="scf_phType"> + <sequence> + <element type="double" name="tr2_ph" default="1E-12" minOccurs="0" /> + <element type="positiveInteger" name="niter_ph" default="100" + minOccurs="0" /> + <element type="double" name="alpha_mix" default="0.7" minOccurs="0" /> + <element type="positiveInteger" name="nmix_ph" default="4" + minOccurs="0" /> + </sequence> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <complexType name="filesType"> + <sequence> + <element type="string" name="prefix" default="pwscf" minOccurs="0" /> + <element type="string" name="outdir" default="./" minOccurs="0" /> + <element type="string" name="fildyn" default="matdyn" minOccurs="0" /> + <element type="string" name="fildrho" default="" minOccurs="0" /> + <element type="string" name="fildvscf" default="" minOccurs="0" /> + <element type="boolean" name="lqdir" default="false" minOccurs="0" /> + </sequence> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <complexType name="control_phType"> + <sequence> + <element type="boolean" name="ldisp" default="false" minOccurs="0" /> + <element type="boolean" name="epsil" default="false" minOccurs="0" /> + <element type="boolean" name="trans" default="true" minOccurs="0" /> + <element type="boolean" name="zeu" default="false" minOccurs="0" /> + <element type="boolean" name="zue" default="false" minOccurs="0" /> + <element type="boolean" name="elop" default="false" minOccurs="0" /> + <element type="boolean" name="fpol" default="false" minOccurs="0" /> + <element type="boolean" name="lraman" default="false" minOccurs="0" /> + <element type="boolean" name="search_sym" default="true" minOccurs="0"/> + </sequence> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <complexType name="control_jobType"> + <sequence> + <element type="boolean" name="recover" default="false" minOccurs="0" /> + <element type="double" name="max_seconds" default="1E7" minOccurs="0" /> + </sequence> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <complexType name="control_dielType"> + <sequence> + <element type="boolean" name="lrpa" default="false" minOccurs="0" /> + <element type="boolean" name="lnoloc" default="false" minOccurs="0" /> + </sequence> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <complexType name="control_qplotType"> + <sequence> + <element type="boolean" name="qplot" default="false" minOccurs="0" /> + <element type="boolean" name="q2d" default="false" minOccurs="0" /> + <element type="boolean" name="q_in_band_form" default="false" + minOccurs="0" /> + </sequence> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <complexType name="miscellaneaType"> + <sequence> + <!-- if amass is not present, should be taked from pw in the code --> + <element type="qes_ph:amassType" name="amass" minOccurs="0" maxOccurs="unbounded"/> + <element type="double" name="start_magnetization" minOccurs="0" /> + <element type="nonNegativeInteger" name="verbosity" default="0" + minOccurs="0" /> + <element type="boolean" name="reduce_io" default="false" minOccurs="0"/> + <element type="boolean" name="low_directory_check" default="false" + minOccurs="0" /> + <element type="boolean" name="nogg" default="false" minOccurs="0" /> + <element type="qes_ph:monkhorst_packType" name="nscf_MPgrid" minOccurs="0"/> + </sequence> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <complexType name="amassType"> + <simpleContent> + <extension base="double"> + <attribute type="positiveInteger" name="atom"/> + </extension> + </simpleContent> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <complexType name="irr_reprType"> + <sequence> + <sequence> + <element type="positiveInteger" name="start_q" default="1" + minOccurs="0" /> + <element type="positiveInteger" name="last_q" default="1" + minOccurs="0" /> + <element type="positiveInteger" name="start_irr" default="1" + minOccurs="0" /> + <element type="positiveInteger" name="last_irr" default="3" + minOccurs="0" /> + <element type="nonNegativeInteger" name="nat_todo" default="0" + minOccurs="0" /> + <element type="nonNegativeInteger" name="modenum" default="0" + minOccurs="0" /> + <element type="boolean" name="only_init" default="false" + minOccurs="0" /> + <element type="boolean" name="ldiag" default="false" minOccurs="0" /> + </sequence> + </sequence> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <complexType name="electron_phonon_optionsType"> + <sequence> + <element type="qes_ph:electron_phononType" name="electron_phonon" + minOccurs="0" /> + <element type="double" name="sigma" default="0.02" minOccurs="0"/> + <element type="integer" name="nsigma" default="10" minOccurs="0"/> + <element type="qes_ph:dvscf_starType" name="dvscf_star" minOccurs="0" /> + <element type="qes_ph:drho_starType" name="drho_star" minOccurs="0" /> + </sequence> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <simpleType name="electron_phononType"> + <restriction base="string"> + <enumeration value="" /> + <enumeration value="simple" /> + <enumeration value="interpolated" /> + </restriction> + </simpleType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <complexType name="dvscf_starType"> + <sequence> + <element type="boolean" name="open" default="false" + minOccurs="0" /> + <element type="string" name="dir" minOccurs="0" /> + <element type="string" name="ext" default="dvscf" + minOccurs="0" /> + <element type="string" name="basis" default="cartesian" + minOccurs="0" /> + <element type="boolean" name="pat" default="true" + minOccurs="0" /> + </sequence> + </complexType> + <!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <complexType name="drho_starType"> + <sequence> + <element type="boolean" name="open" default="false" + minOccurs="0" /> + <element type="string" name="dir" minOccurs="0" /> + <element type="string" name="ext" default="drho" + minOccurs="0" /> + <element type="string" name="basis" default="modes" + minOccurs="0" /> + <element type="boolean" name="pat" default="true" + minOccurs="0" /> + </sequence> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <complexType name="lraman_optionsType"> + <sequence> + <element type="double" name="eth_rps" default="1E-9" minOccurs="0" /> + <element type="double" name="eth_ns" default="1E-12" minOccurs="0" /> + <element type="double" name="dek" default="1E-3" minOccurs="0" /> + </sequence> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <complexType name="q_pointsType"> + <sequence> + <!-- monkhorst_pack --> + <element type="qes_ph:q_monkhorst_packType" name="grid" minOccurs="0" maxOccurs="1"/> + <!-- monkhorst_pack grid, ex. 3 3 3 --> + <!-- Alternatively, a list of nqs q-points... (optionally with weight as attribute) --> + <element type="nonNegativeInteger" name="nqs" default="1" minOccurs="0" /> + <element type="qes_ph:q_points_listType" name="q_points_list" minOccurs="0" /> + </sequence> + </complexType> + <!-- q points, similar but not the same as in qes schema (here no offset) --> + <complexType name="q_points_listType"> + <sequence> + <element type="qes_ph:k_pointType" name="q_point" maxOccurs="unbounded" + minOccurs="0"/> + </sequence> + </complexType> + <complexType name="q_monkhorst_packType"> + <simpleContent> + <extension base="string"> + <attribute type="positiveInteger" name="nq1"/> + <attribute type="positiveInteger" name="nq2"/> + <attribute type="positiveInteger" name="nq3"/> + </extension> + </simpleContent> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <!-- k points, same as in qes schema --> + <complexType name="k_points_IBZType"> + <sequence> + <element type="qes_ph:monkhorst_packType" name="monkhorst_pack" minOccurs="0" + maxOccurs="1"/> + <element type="positiveInteger" name="nk"/> + <element type="qes_ph:k_pointType" name="k_point" maxOccurs="unbounded" + minOccurs="0"/> + </sequence> + </complexType> + <complexType name="monkhorst_packType"> + <simpleContent> + <extension base="string"> + <attribute type="positiveInteger" name="nk1"/> + <attribute type="positiveInteger" name="nk2"/> + <attribute type="positiveInteger" name="nk3"/> + <attribute type="boolean" name="k1"/> + <attribute type="boolean" name="k2"/> + <attribute type="boolean" name="k3"/> + </extension> + </simpleContent> + </complexType> + <complexType name="k_pointType"> + <simpleContent> + <extension base="qes_ph:d3vectorType"> + <attribute type="double" name="weight" use="optional"/> + </extension> + </simpleContent> + </complexType> + <!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <!-- CLOSED, same as in qes schema --> + <complexType name="closedType"> + <simpleContent> + <extension base="string"> + <attribute type="string" name="DATE"/> + <attribute type="string" name="TIME"/> + </extension> + </simpleContent> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <!-- STATUS, same as in qes schema --> + <simpleType name="statusType"> + <restriction base="unsignedByte"> + <enumeration value="0" /> + <enumeration value="1" /> + <enumeration value="2" /> + <enumeration value="3" /> + <enumeration value="255"/> + </restriction> + </simpleType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <!-- OUTPUTPH TYPE--> + <complexType name="outputPHType"> + <sequence> + <!-- As in qes --> + <element type="qes_ph:atomic_speciesType" name="atomic_species"/> + <element type="qes_ph:atomic_structureType" name="atomic_structure"/> + <!-- In addition to qes --> + <element type="qes_ph:unit_cell_volumeType" name="unit_cell_volume"/> + <element type="nonNegativeInteger" name="ibrav" + minOccurs="0" /> + <element type="qes_ph:cell_dimensionsType" name="cell_dimensions" + minOccurs="0" /> + <element type="double" name="start_magnetization" minOccurs="0" /> + <element type="nonNegativeInteger" name="number_of_q" minOccurs="0"/> + <element type="nonNegativeInteger" name="number_of_irr_q" + minOccurs="0"/> + <element type="qes_ph:irr_repr_listType" name="irr_repr_list" + minOccurs="0" /> + <element type="qes_ph:dynamical_matType" name="dynamical_mat" + minOccurs="0" /> + <element type="qes_ph:force_constantsType" name="force_constants" + minOccurs="0" /> + <element type="qes_ph:frequenciesType" name="frequencies" + minOccurs="0"/> + </sequence> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <!-- Unit cell volume with unit --> + <complexType name="unit_cell_volumeType"> + <simpleContent> + <extension base="double"> + <attribute type="string" name="unit" use="required"/> + </extension> + </simpleContent> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <!-- Cell dimensions --> + <simpleType name="cell_dimensionsType"> + <restriction> + <simpleType> + <list itemType="double"/> + </simpleType> + <length value="6"/> + </restriction> + </simpleType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <!-- Irr. representations --> + <complexType name="irr_repr_listType"> + <sequence> + <element type="qes_ph:repr_at_qType" name="repr_at_q" maxOccurs="unbounded" /> + </sequence> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <complexType name="repr_at_qType"> + <sequence> + <element type="qes_ph:d3vectorType" name="q_point" /> + <element type="qes_ph:reprType" name="repr" maxOccurs="unbounded" /> + </sequence> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <complexType name="reprType"> + <sequence> + <element type="qes_ph:modeType" name="mode" maxOccurs="unbounded" /> + </sequence> + <attribute type="nonNegativeInteger" name="number" /> + <attribute type="nonNegativeInteger" name="n_modes" /> + <attribute type="string" name="descr" /> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <complexType name="modeType"> + <sequence> + <element type="qes_ph:displacementType" name="displacement" maxOccurs="unbounded"/> + </sequence> + <attribute type="nonNegativeInteger" name="number" /> + </complexType> + +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <!-- Dynamical matrix --> + <complexType name="dynamical_matType"> + <sequence> + <element type="qes_ph:D_elements_symmequivType" + name="D_elements_symmequiv" maxOccurs="unbounded" /> + </sequence> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <!-- Dynamical matrix element (3x3 matrix at a given q point)--> + <complexType name="D_elements_symmequivType"> + <sequence> + <element type="qes_ph:D_qType" name="D_q" maxOccurs="unbounded" /> + </sequence> + <attribute type="nonNegativeInteger" name="number_of_equiv_q" /> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <!-- Dynamical matrix (3x3) at a given q point)--> + <complexType name="D_qType"> + <sequence> + <element type="qes_ph:d3vectorType" name="q_point" /> + <element type="qes_ph:phi_qType" name="phi_q" maxOccurs="unbounded" /> + </sequence> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <!-- Dynamical matrix element (3x3 matrix at a given q point)--> + <complexType name="phi_qType"> + <simpleContent> + <extension base="qes_ph:d3complexDType"> + <attribute type="positiveInteger" name="at1" /> + <attribute type="positiveInteger" name="at2" /> + </extension> + </simpleContent> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <!-- Frequencies --> + <complexType name="frequenciesType"> + <sequence> + <element type="qes_ph:frequencyType" name="frequency" minOccurs="0" + maxOccurs="unbounded"/> + </sequence> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <!-- frequency type, each containing omega (possibly in different units) + and a displacement --> + <complexType name="frequencyType"> + <sequence> + <!-- two frequencies because they can be in cm-1 and THz--> + <element type="qes_ph:omegaType" name="omega" maxOccurs="2"/> + <element type="qes_ph:displacementType" name="displacement" maxOccurs="unbounded"/> + </sequence> + <attribute type="integer" name="index" use="required"/> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <!-- freq type, unit as attribute (THz and cm-1) --> + <complexType name="omegaType"> + <simpleContent> + <extension base="double"> + <attribute type="string" name="unit" use="required"/> + </extension> + </simpleContent> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <complexType name="displacementType"> + <sequence> + <element type="qes_ph:disp_x_y_zType" name="x" maxOccurs="1"/> + <element type="qes_ph:disp_x_y_zType" name="y" maxOccurs="1"/> + <element type="qes_ph:disp_x_y_zType" name="z" maxOccurs="1"/> + </sequence> + <attribute type="integer" name="atom" use="required"/> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <!-- displacement type --> + <simpleType name="disp_x_y_zType"> + <restriction> + <simpleType> + <list itemType="double"/> + </simpleType> + <length value="2"/> + </restriction> + </simpleType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <!-- Force constants--> + <complexType name="force_constantsType"> + <sequence> + <element type="qes_ph:phi_rType" name="phi_r" maxOccurs="unbounded" /> + </sequence> + <attribute type="positiveInteger" name="nq1" /> + <attribute type="positiveInteger" name="nq2" /> + <attribute type="positiveInteger" name="nq3" /> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <!-- Force constant element (3x3 matrix at a given q point)--> + <complexType name="phi_rType"> + <sequence> + <element type="qes_ph:phi_r_elType" name="phi_r_el" + maxOccurs="unbounded" /> + </sequence> + <attribute type="positiveInteger" name="at1" /> + <attribute type="positiveInteger" name="at2" /> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <!-- Force constant element (3x3 matrix at a given q point)--> + <complexType name="phi_r_elType"> + <simpleContent> + <extension base="double"> + <attribute type="positiveInteger" name="m1" /> + <attribute type="positiveInteger" name="m2" /> + <attribute type="positiveInteger" name="m3" /> + </extension> + </simpleContent> + </complexType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <!-- Input and Output types as in qes --> + <complexType name="atomic_speciesType"> + <sequence> + <element type="qes_ph:specieType" name="specie" maxOccurs="unbounded"/> + </sequence> + <attribute type="positiveInteger" name="ntyp"/> + </complexType> + <complexType name="specieType"> + <sequence> + <element type="double" name="mass" minOccurs="0" /> + <element type="string" name="pseudo_file"/> + <element type="double" name="starting_magnetization" minOccurs="0"/> + </sequence> + <attribute type="string" name="name"/> + </complexType> + <complexType name="atomic_structureType"> + <sequence> + <element type="qes_ph:atomic_positionsType" name="atomic_positions" + minOccurs="0" /> + <element type="qes_ph:wyckoff_positionsType" name="wyckoff_positions" + minOccurs="0" /> + <element type="qes_ph:cellType" name="cell"/> + </sequence> + <attribute type="positiveInteger" name="nat" use="optional"/> + <attribute type="double" name="alat" use="optional"/> + </complexType> + <complexType name="atomic_positionsType"> + <sequence> + <element type="qes_ph:atomType" name="atom" maxOccurs="unbounded" /> + </sequence> + </complexType> + <complexType name="atomType"> + <simpleContent> + <extension base="qes_ph:d3vectorType"> + <attribute type="string" name="name" /> + <attribute type="string" name="position" use="optional" /> + </extension> + </simpleContent> + </complexType> + <complexType name="wyckoff_positionsType"> + <sequence> + <element type="qes_ph:atomType" name="atom"/> + </sequence> + <attribute type="unsignedByte" name="space_group" use="optional"/> + <attribute type="string" name="more_options" use="optional"/> + </complexType> + <complexType name="cellType"> + <sequence> + <element type="qes_ph:d3vectorType" name="a1"/> + <element type="qes_ph:d3vectorType" name="a2"/> + <element type="qes_ph:d3vectorType" name="a3"/> + </sequence> + </complexType> + + <!-- 3x3 real matrix type --> + <simpleType name="d3realDType"> + <restriction> + <simpleType> + <list itemType="double"/> + </simpleType> + <length value="9"/> + </restriction> + </simpleType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <!-- 3x3 complex matrix type --> + <simpleType name="d3complexDType"> + <restriction> + <simpleType> + <list itemType="double"/> + </simpleType> + <length value="18"/> + </restriction> + </simpleType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> + <!-- FPMD TYPES (ref: quantum-simulation.org) --> + <simpleType name="d3vectorType"> + <restriction> + <simpleType> + <list itemType="double"/> + </simpleType> + <length value="3"/> + </restriction> + </simpleType> +<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - --> +</schema>	lda_plus_u = .t. is missing from input Try attached xml, it is expected that lda_plus_u = .t. should be present in the .in [FeO_LDAU_standard.txt](https://github.com/QEF/qeschema/files/6468051/FeO_LDAU_standard.txt)	The presence of tag "dftU" implies that lda_plus_u = .t. Related qexsd commit: https://github.com/QEF/qexsd/commit/a6d365433641f2b41d6155bd49b4628d3adcf51d yes I forgot to add the fix here. Thanks Sasha. This is a blocker to switch to qeschema (everything else seems to be working). Hopefully it is resolved soon.	2021-06-29T21:25:13	[]	[]
eclipse-zenoh/zenoh-python	eclipse-zenoh__zenoh-python-78	91ab2c2c837e7cfdeadbda8f365feb44921b126c	diff --git a/.github/workflows/release.yml b/.github/workflows/release.yml index 54861ec8..485465b0 100644 --- a/.github/workflows/release.yml +++ b/.github/workflows/release.yml @@ -144,3 +144,21 @@ jobs: with: name: wheels path: dist + + deploy-wheels: + needs: [macos, windows, linux, linux-cross, linux-armv6] + name: deploy wheels to pypi + runs-on: ubuntu-latest + steps: + - uses: actions/download-artifact@v2 + with: + name: wheels + path: dist + - name: Check dist + run: ls -al ./dist/* + - name: publish + if: github.event_name == 'release' && github.event.action == 'released' + uses: pypa/gh-action-pypi-publish@master + with: + user: __token__ + password: ${{ secrets.PYPI_API_TOKEN }} diff --git a/.gitignore b/.gitignore index 0b9c26bf..f01d859d 100644 --- a/.gitignore +++ b/.gitignore @@ -35,3 +35,6 @@ docs/_* .bak .vscode/ + +/pycache +/__pycache__ diff --git a/.gitmodules b/.gitmodules deleted file mode 100644 index e69de29b..00000000 diff --git a/Cargo.lock b/Cargo.lock index f773e926..46edc929 100644 --- a/Cargo.lock +++ b/Cargo.lock @@ -4,39 +4,35 @@ version = 3 [[package]] name = "aes" -version = "0.7.5" +version = "0.8.1" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "9e8b47f52ea9bae42228d07ec09eb676433d7c4ed1ebdf0f1d1c29ed446f1ab8" +checksum = "bfe0133578c0986e1fe3dfcd4af1cc5b2dd6c3dbf534d69916ce16a2701d40ba" dependencies = [ "cfg-if", "cipher", "cpufeatures", - "opaque-debug 0.3.0", ] [[package]] name = "aho-corasick" -version = "0.7.18" +version = "0.7.19" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "1e37cfd5e7657ada45f742d6e99ca5788580b5c529dc78faf11ece6dc702656f" +checksum = "b4f55bd91a0978cbfd91c457a164bab8b4001c833b7f323132c0a4e1922dd44e" dependencies = [ "memchr", ] [[package]] -name = "ansi_term" -version = "0.12.1" +name = "anyhow" +version = "1.0.65" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "d52a9bb7ec0cf484c551830a7ce27bd20d67eac647e1befb56b0be4ee39a55d2" -dependencies = [ - "winapi", -] +checksum = "98161a4e3e2184da77bb14f02184cdd111e83bbbcc9979dfee3c44b9a85f5602" [[package]] -name = "anyhow" -version = "1.0.56" +name = "array-init" +version = "2.0.1" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "4361135be9122e0870de935d7c439aef945b9f9ddd4199a553b5270b49c82a27" +checksum = "bfb6d71005dc22a708c7496eee5c8dc0300ee47355de6256c3b35b12b5fef596" [[package]] name = "async-attributes" @@ -50,9 +46,9 @@ dependencies = [ [[package]] name = "async-channel" -version = "1.6.1" +version = "1.7.1" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "2114d64672151c0c5eaa5e131ec84a74f06e1e559830dabba01ca30605d66319" +checksum = "e14485364214912d3b19cc3435dde4df66065127f05fa0d75c712f36f12c2f28" dependencies = [ "concurrent-queue", "event-listener", @@ -75,27 +71,27 @@ dependencies = [ [[package]] name = "async-global-executor" -version = "2.0.3" +version = "2.3.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "c026b7e44f1316b567ee750fea85103f87fcb80792b860e979f221259796ca0a" +checksum = "0da5b41ee986eed3f524c380e6d64965aea573882a8907682ad100f7859305ca" dependencies = [ "async-channel", "async-executor", "async-io", - "async-mutex", + "async-lock", "blocking", "futures-lite", - "num_cpus", "once_cell", "tokio", ] [[package]] name = "async-io" -version = "1.6.0" +version = "1.9.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "a811e6a479f2439f0c04038796b5cfb3d2ad56c230e0f2d3f7b04d68cfee607b" +checksum = "83e21f3a490c72b3b0cf44962180e60045de2925d8dff97918f7ee43c8f637c7" dependencies = [ + "autocfg", "concurrent-queue", "futures-lite", "libc", @@ -118,22 +114,14 @@ dependencies = [ "event-listener", ] -[[package]] -name = "async-mutex" -version = "1.4.0" -source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "479db852db25d9dbf6204e6cb6253698f175c15726470f78af0d918e99d6156e" -dependencies = [ - "event-listener", -] - [[package]] name = "async-process" -version = "1.3.0" +version = "1.5.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "83137067e3a2a6a06d67168e49e68a0957d215410473a740cea95a2425c0b7c6" +checksum = "02111fd8655a613c25069ea89fc8d9bb89331fa77486eb3bc059ee757cfa481c" dependencies = [ "async-io", + "autocfg", "blocking", "cfg-if", "event-listener", @@ -157,9 +145,9 @@ dependencies = [ [[package]] name = "async-std" -version = "1.11.0" +version = "1.12.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "52580991739c5cdb36cde8b2a516371c0a3b70dda36d916cc08b82372916808c" +checksum = "62565bb4402e926b29953c785397c6dc0391b7b446e45008b0049eb43cec6f5d" dependencies = [ "async-attributes", "async-channel", @@ -176,7 +164,6 @@ dependencies = [ "kv-log-macro", "log", "memchr", - "num_cpus", "once_cell", "pin-project-lite", "pin-utils", @@ -186,15 +173,15 @@ dependencies = [ [[package]] name = "async-task" -version = "4.2.0" +version = "4.3.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "30696a84d817107fc028e049980e09d5e140e8da8f1caeb17e8e950658a3cea9" +checksum = "7a40729d2133846d9ed0ea60a8b9541bccddab49cd30f0715a1da672fe9a2524" [[package]] name = "async-trait" -version = "0.1.52" +version = "0.1.57" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "061a7acccaa286c011ddc30970520b98fa40e00c9d644633fb26b5fc63a265e3" +checksum = "76464446b8bc32758d7e88ee1a804d9914cd9b1cb264c029899680b0be29826f" dependencies = [ "proc-macro2", "quote", @@ -218,15 +205,6 @@ dependencies = [ "winapi", ] -[[package]] -name = "autocfg" -version = "0.1.8" -source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "0dde43e75fd43e8a1bf86103336bc699aa8d17ad1be60c76c0bdfd4828e19b78" -dependencies = [ - "autocfg 1.1.0", -] - [[package]] name = "autocfg" version = "1.1.0" @@ -241,9 +219,9 @@ checksum = "904dfeac50f3cdaba28fc6f57fdcddb75f49ed61346676a78c4ffe55877802fd" [[package]] name = "base64ct" -version = "1.1.1" +version = "1.5.2" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "e6b4d9b1225d28d360ec6a231d65af1fd99a2a095154c8040689617290569c5c" +checksum = "ea2b2456fd614d856680dcd9fcc660a51a820fa09daef2e49772b56a193c8474" [[package]] name = "bincode" @@ -262,41 +240,13 @@ checksum = "bef38d45163c2f1dde094a7dfd33ccf595c92905c8f8f4fdc18d06fb1037718a" [[package]] name = "block-buffer" -version = "0.7.3" -source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "c0940dc441f31689269e10ac70eb1002a3a1d3ad1390e030043662eb7fe4688b" -dependencies = [ - "block-padding 0.1.5", - "byte-tools", - "byteorder", - "generic-array 0.12.4", -] - -[[package]] -name = "block-buffer" -version = "0.9.0" -source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "4152116fd6e9dadb291ae18fc1ec3575ed6d84c29642d97890f4b4a3417297e4" -dependencies = [ - "block-padding 0.2.1", - "generic-array 0.14.5", -] - -[[package]] -name = "block-padding" -version = "0.1.5" +version = "0.10.3" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "fa79dedbb091f449f1f39e53edf88d5dbe95f895dae6135a8d7b881fb5af73f5" +checksum = "69cce20737498f97b993470a6e536b8523f0af7892a4f928cceb1ac5e52ebe7e" dependencies = [ - "byte-tools", + "generic-array", ] -[[package]] -name = "block-padding" -version = "0.2.1" -source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "8d696c370c750c948ada61c69a0ee2cbbb9c50b1019ddb86d9317157a99c2cae" - [[package]] name = "blocking" version = "1.2.0" @@ -313,15 +263,9 @@ dependencies = [ [[package]] name = "bumpalo" -version = "3.9.1" +version = "3.11.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "a4a45a46ab1f2412e53d3a0ade76ffad2025804294569aae387231a0cd6e0899" - -[[package]] -name = "byte-tools" -version = "0.3.1" -source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "e3b5ca7a04898ad4bcd41c90c5285445ff5b791899bb1b0abdd2a2aa791211d7" +checksum = "c1ad822118d20d2c234f427000d5acc36eabe1e29a348c89b63dd60b13f28e5d" [[package]] name = "byteorder" @@ -331,9 +275,9 @@ checksum = "14c189c53d098945499cdfa7ecc63567cf3886b3332b312a5b4585d8d3a6a610" [[package]] name = "bytes" -version = "1.1.0" +version = "1.2.1" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "c4872d67bab6358e59559027aa3b9157c53d9358c51423c17554809a8858e0f8" +checksum = "ec8a7b6a70fde80372154c65702f00a0f56f3e1c36abbc6c440484be248856db" [[package]] name = "cache-padded" @@ -355,42 +299,52 @@ checksum = "baf1de4339761588bc0619e3cbc0120ee582ebb74b53b4efbf79117bd2da40fd" [[package]] name = "cipher" -version = "0.3.0" +version = "0.4.3" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "7ee52072ec15386f770805afd189a01c8841be8696bed250fa2f13c4c0d6dfb7" +checksum = "d1873270f8f7942c191139cb8a40fd228da6c3fd2fc376d7e92d47aa14aeb59e" dependencies = [ - "generic-array 0.14.5", + "crypto-common", + "inout", ] [[package]] name = "clap" -version = "2.34.0" +version = "3.2.22" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "a0610544180c38b88101fecf2dd634b174a62eef6946f84dfc6a7127512b381c" +checksum = "86447ad904c7fb335a790c9d7fe3d0d971dc523b8ccd1561a520de9a85302750" dependencies = [ - "ansi_term", "atty", "bitflags", + "clap_lex", + "indexmap", "strsim", + "termcolor", "textwrap", - "unicode-width", - "vec_map", +] + +[[package]] +name = "clap_lex" +version = "0.2.4" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "2850f2f5a82cbf437dd5af4d49848fbdfc27c157c3d010345776f952765261c5" +dependencies = [ + "os_str_bytes", ] [[package]] name = "concurrent-queue" -version = "1.2.2" +version = "1.2.4" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "30ed07550be01594c6026cff2a1d7fe9c8f683caa798e12b68694ac9e88286a3" +checksum = "af4780a44ab5696ea9e28294517f1fffb421a83a25af521333c838635509db9c" dependencies = [ "cache-padded", ] [[package]] name = "const-oid" -version = "0.6.2" +version = "0.7.1" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "9d6f2aa4d0537bcc1c74df8755072bd31c1ef1a3a1b85a68e8404a8c353b7b8b" +checksum = "e4c78c047431fee22c1a7bb92e00ad095a02a983affe4d8a72e2a2c62c1b94f3" [[package]] name = "core-foundation" @@ -410,49 +364,48 @@ checksum = "5827cebf4670468b8772dd191856768aedcb1b0278a04f989f7766351917b9dc" [[package]] name = "cpufeatures" -version = "0.2.1" +version = "0.2.5" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "95059428f66df56b63431fdb4e1947ed2190586af5c5a8a8b71122bdf5a7f469" +checksum = "28d997bd5e24a5928dd43e46dc529867e207907fe0b239c3477d924f7f2ca320" dependencies = [ "libc", ] [[package]] name = "crossbeam-utils" -version = "0.8.8" +version = "0.8.11" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "0bf124c720b7686e3c2663cf54062ab0f68a88af2fb6a030e87e30bf721fcb38" +checksum = "51887d4adc7b564537b15adcfb307936f8075dfcd5f00dde9a9f1d29383682bc" dependencies = [ "cfg-if", - "lazy_static", + "once_cell", ] [[package]] name = "crypto-bigint" -version = "0.2.11" +version = "0.3.2" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "f83bd3bb4314701c568e340cd8cf78c975aa0ca79e03d3f6d1677d5b0c9c0c03" +checksum = "03c6a1d5fa1de37e071642dfa44ec552ca5b299adb128fab16138e24b548fd21" dependencies = [ - "generic-array 0.14.5", - "rand_core", + "generic-array", "subtle", ] [[package]] -name = "crypto-mac" -version = "0.11.1" +name = "crypto-common" +version = "0.1.6" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "b1d1a86f49236c215f271d40892d5fc950490551400b02ef360692c29815c714" +checksum = "1bfb12502f3fc46cca1bb51ac28df9d618d813cdc3d2f25b9fe775a34af26bb3" dependencies = [ - "generic-array 0.14.5", - "subtle", + "generic-array", + "typenum", ] [[package]] name = "ctor" -version = "0.1.22" +version = "0.1.23" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "f877be4f7c9f246b183111634f75baa039715e3f46ce860677d3b19a69fb229c" +checksum = "cdffe87e1d521a10f9696f833fe502293ea446d7f256c06128293a4119bdf4cb" dependencies = [ "quote", "syn", @@ -460,47 +413,40 @@ dependencies = [ [[package]] name = "der" -version = "0.4.5" +version = "0.5.1" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "79b71cca7d95d7681a4b3b9cdf63c8dbc3730d0584c2c74e31416d64a90493f4" +checksum = "6919815d73839e7ad218de758883aae3a257ba6759ce7a9992501efbb53d705c" dependencies = [ "const-oid", "crypto-bigint", + "pem-rfc7468", ] [[package]] name = "digest" -version = "0.8.1" -source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "f3d0c8c8752312f9713efd397ff63acb9f85585afbf179282e720e7704954dd5" -dependencies = [ - "generic-array 0.12.4", -] - -[[package]] -name = "digest" -version = "0.9.0" +version = "0.10.5" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "d3dd60d1080a57a05ab032377049e0591415d2b31afd7028356dbf3cc6dcb066" +checksum = "adfbc57365a37acbd2ebf2b64d7e69bb766e2fea813521ed536f5d0520dcf86c" dependencies = [ - "generic-array 0.14.5", + "block-buffer", + "crypto-common", + "subtle", ] [[package]] -name = "dirs-next" -version = "2.0.0" +name = "dirs" +version = "4.0.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "b98cf8ebf19c3d1b223e151f99a4f9f0690dca41414773390fc824184ac833e1" +checksum = "ca3aa72a6f96ea37bbc5aa912f6788242832f75369bdfdadcb0e38423f100059" dependencies = [ - "cfg-if", - "dirs-sys-next", + "dirs-sys", ] [[package]] -name = "dirs-sys-next" -version = "0.1.2" +name = "dirs-sys" +version = "0.3.7" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "4ebda144c4fe02d1f7ea1a7d9641b6fc6b580adcfa024ae48797ecdeb6825b4d" +checksum = "1b1d1d91c932ef41c0f2663aa8b0ca0342d444d842c06914aa0a7e352d0bada6" dependencies = [ "libc", "redox_users", @@ -513,25 +459,31 @@ version = "0.6.0-dev" dependencies = [ "async-std", "env_logger", + "form_urlencoded", "futures", "json5", "log", "pyo3", - "pyo3-asyncio", "serde_json", "uhlc", - "validated_struct 0.1.11", + "validated_struct", "zenoh", "zenoh-buffers", "zenoh-cfg-properties", "zenoh-core", ] +[[package]] +name = "either" +version = "1.8.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "90e5c1c8368803113bf0c9584fc495a58b86dc8a29edbf8fe877d21d9507e797" + [[package]] name = "env_logger" -version = "0.9.0" +version = "0.9.1" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "0b2cf0344971ee6c64c31be0d530793fba457d322dfec2810c453d0ef228f9c3" +checksum = "c90bf5f19754d10198ccb95b70664fc925bd1fc090a0fd9a6ebc54acc8cd6272" dependencies = [ "atty", "humantime", @@ -542,49 +494,52 @@ dependencies = [ [[package]] name = "event-listener" -version = "2.5.2" +version = "2.5.3" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "77f3309417938f28bf8228fcff79a4a37103981e3e186d2ccd19c74b38f4eb71" - -[[package]] -name = "fake-simd" -version = "0.1.2" -source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "e88a8acf291dafb59c2d96e8f59828f3838bb1a70398823ade51a84de6a6deed" +checksum = "0206175f82b8d6bf6652ff7d71a1e27fd2e4efde587fd368662814d6ec1d9ce0" [[package]] name = "fastrand" -version = "1.7.0" +version = "1.8.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "c3fcf0cee53519c866c09b5de1f6c56ff9d647101f81c1964fa632e148896cdf" +checksum = "a7a407cfaa3385c4ae6b23e84623d48c2798d06e3e6a1878f7f59f17b3f86499" dependencies = [ "instant", ] [[package]] name = "fixedbitset" -version = "0.4.1" +version = "0.4.2" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "279fb028e20b3c4c320317955b77c5e0c9701f05a1d309905d6fc702cdc5053e" +checksum = "0ce7134b9999ecaf8bcd65542e436736ef32ddca1b3e06094cb6ec5755203b80" [[package]] name = "flume" -version = "0.10.12" +version = "0.10.14" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "843c03199d0c0ca54bc1ea90ac0d507274c28abcc4f691ae8b4eaa375087c76a" +checksum = "1657b4441c3403d9f7b3409e47575237dac27b1b5726df654a6ecbf92f0f7577" dependencies = [ "futures-core", "futures-sink", "nanorand", "pin-project", - "spin 0.9.2", + "spin 0.9.4", +] + +[[package]] +name = "form_urlencoded" +version = "1.1.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "a9c384f161156f5260c24a097c56119f9be8c798586aecc13afbcbe7b7e26bf8" +dependencies = [ + "percent-encoding", ] [[package]] name = "futures" -version = "0.3.21" +version = "0.3.24" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "f73fe65f54d1e12b726f517d3e2135ca3125a437b6d998caf1962961f7172d9e" +checksum = "7f21eda599937fba36daeb58a22e8f5cee2d14c4a17b5b7739c7c8e5e3b8230c" dependencies = [ "futures-channel", "futures-core", @@ -597,9 +552,9 @@ dependencies = [ [[package]] name = "futures-channel" -version = "0.3.21" +version = "0.3.24" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "c3083ce4b914124575708913bca19bfe887522d6e2e6d0952943f5eac4a74010" +checksum = "30bdd20c28fadd505d0fd6712cdfcb0d4b5648baf45faef7f852afb2399bb050" dependencies = [ "futures-core", "futures-sink", @@ -607,15 +562,15 @@ dependencies = [ [[package]] name = "futures-core" -version = "0.3.21" +version = "0.3.24" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "0c09fd04b7e4073ac7156a9539b57a484a8ea920f79c7c675d05d289ab6110d3" +checksum = "4e5aa3de05362c3fb88de6531e6296e85cde7739cccad4b9dfeeb7f6ebce56bf" [[package]] name = "futures-executor" -version = "0.3.21" +version = "0.3.24" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "9420b90cfa29e327d0429f19be13e7ddb68fa1cccb09d65e5706b8c7a749b8a6" +checksum = "9ff63c23854bee61b6e9cd331d523909f238fc7636290b96826e9cfa5faa00ab" dependencies = [ "futures-core", "futures-task", @@ -624,9 +579,9 @@ dependencies = [ [[package]] name = "futures-io" -version = "0.3.21" +version = "0.3.24" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "fc4045962a5a5e935ee2fdedaa4e08284547402885ab326734432bed5d12966b" +checksum = "bbf4d2a7a308fd4578637c0b17c7e1c7ba127b8f6ba00b29f717e9655d85eb68" [[package]] name = "futures-lite" @@ -645,9 +600,9 @@ dependencies = [ [[package]] name = "futures-macro" -version = "0.3.21" +version = "0.3.24" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "33c1e13800337f4d4d7a316bf45a567dbcb6ffe087f16424852d97e97a91f512" +checksum = "42cd15d1c7456c04dbdf7e88bcd69760d74f3a798d6444e16974b505b0e62f17" dependencies = [ "proc-macro2", "quote", @@ -656,21 +611,21 @@ dependencies = [ [[package]] name = "futures-sink" -version = "0.3.21" +version = "0.3.24" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "21163e139fa306126e6eedaf49ecdb4588f939600f0b1e770f4205ee4b7fa868" +checksum = "21b20ba5a92e727ba30e72834706623d94ac93a725410b6a6b6fbc1b07f7ba56" [[package]] name = "futures-task" -version = "0.3.21" +version = "0.3.24" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "57c66a976bf5909d801bbef33416c41372779507e7a6b3a5e25e4749c58f776a" +checksum = "a6508c467c73851293f390476d4491cf4d227dbabcd4170f3bb6044959b294f1" [[package]] name = "futures-util" -version = "0.3.21" +version = "0.3.24" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "d8b7abd5d659d9b90c8cba917f6ec750a74e2dc23902ef9cd4cc8c8b22e6036a" +checksum = "44fb6cb1be61cc1d2e43b262516aafcf63b241cffdb1d3fa115f91d9c7b09c90" dependencies = [ "futures-channel", "futures-core", @@ -695,18 +650,9 @@ dependencies = [ [[package]] name = "generic-array" -version = "0.12.4" +version = "0.14.6" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "ffdf9f34f1447443d37393cc6c2b8313aebddcd96906caf34e54c68d8e57d7bd" -dependencies = [ - "typenum", -] - -[[package]] -name = "generic-array" -version = "0.14.5" -source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "fd48d33ec7f05fbfa152300fdad764757cbded343c1aa1cff2fbaf4134851803" +checksum = "bff49e947297f3312447abdca79f45f4738097cc82b06e72054d2223f601f1b9" dependencies = [ "typenum", "version_check", @@ -714,28 +660,17 @@ dependencies = [ [[package]] name = "getrandom" -version = "0.2.5" +version = "0.2.7" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "d39cd93900197114fa1fcb7ae84ca742095eed9442088988ae74fa744e930e77" +checksum = "4eb1a864a501629691edf6c15a593b7a51eebaa1e8468e9ddc623de7c9b58ec6" dependencies = [ "cfg-if", "js-sys", "libc", - "wasi 0.10.2+wasi-snapshot-preview1", + "wasi", "wasm-bindgen", ] -[[package]] -name = "ghost" -version = "0.1.2" -source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "1a5bcf1bbeab73aa4cf2fde60a846858dc036163c7c33bec309f8d17de785479" -dependencies = [ - "proc-macro2", - "quote", - "syn", -] - [[package]] name = "git-version" version = "0.3.5" @@ -766,9 +701,9 @@ checksum = "9b919933a397b79c37e33b77bb2aa3dc8eb6e165ad809e58ff75bc7db2e34574" [[package]] name = "gloo-timers" -version = "0.2.3" +version = "0.2.4" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "4d12a7f4e95cfe710f1d624fb1210b7d961a5fb05c4fd942f4feab06e61f590e" +checksum = "5fb7d06c1c8cc2a29bee7ec961009a0b2caa0793ee4900c2ffb348734ba1c8f9" dependencies = [ "futures-channel", "futures-core", @@ -778,9 +713,9 @@ dependencies = [ [[package]] name = "hashbrown" -version = "0.11.2" +version = "0.12.3" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "ab5ef0d4909ef3724cc8cce6ccc8572c5c817592e9285f5464f8e86f8bd3726e" +checksum = "8a9ee70c43aaf417c914396645a0fa852624801b24ebb7ae78fe8272889ac888" [[package]] name = "hermit-abi" @@ -799,12 +734,11 @@ checksum = "7f24254aa9a54b5c858eaee2f5bccdb46aaf0e486a595ed5fd8f86ba55232a70" [[package]] name = "hmac" -version = "0.11.0" +version = "0.12.1" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "2a2a2320eb7ec0ebe8da8f744d7812d9fc4cb4d09344ac01898dbcb6a20ae69b" +checksum = "6c49c37c09c17a53d937dfbb742eb3a961d65a994e6bcdcf37e7399d0cc8ab5e" dependencies = [ - "crypto-mac", - "digest 0.9.0", + "digest", ] [[package]] @@ -824,35 +758,27 @@ checksum = "9a3a5bfb195931eeb336b2a7b4d761daec841b97f947d34394601737a7bba5e4" [[package]] name = "indexmap" -version = "1.8.0" +version = "1.9.1" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "282a6247722caba404c065016bbfa522806e51714c34f5dfc3e4a3a46fcb4223" +checksum = "10a35a97730320ffe8e2d410b5d3b69279b98d2c14bdb8b70ea89ecf7888d41e" dependencies = [ - "autocfg 1.1.0", + "autocfg", "hashbrown", ] [[package]] name = "indoc" -version = "0.3.6" +version = "1.0.7" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "47741a8bc60fb26eb8d6e0238bbb26d8575ff623fdc97b1a2c00c050b9684ed8" -dependencies = [ - "indoc-impl", - "proc-macro-hack", -] +checksum = "adab1eaa3408fb7f0c777a73e7465fd5656136fc93b670eb6df3c88c2c1344e3" [[package]] -name = "indoc-impl" -version = "0.3.6" +name = "inout" +version = "0.1.3" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "ce046d161f000fffde5f432a0d034d0341dc152643b2598ed5bfce44c4f3a8f0" +checksum = "a0c10553d664a4d0bcff9f4215d0aac67a639cc68ef660840afe309b807bc9f5" dependencies = [ - "proc-macro-hack", - "proc-macro2", - "quote", - "syn", - "unindent", + "generic-array", ] [[package]] @@ -865,47 +791,34 @@ dependencies = [ ] [[package]] -name = "inventory" -version = "0.1.11" -source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "f0eb5160c60ba1e809707918ee329adb99d222888155835c6feedba19f6c3fd4" -dependencies = [ - "ctor", - "ghost", - "inventory-impl", -] - -[[package]] -name = "inventory-impl" -version = "0.1.11" +name = "ipnetwork" +version = "0.19.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "7e41b53715c6f0c4be49510bb82dee2c1e51c8586d885abe65396e82ed518548" +checksum = "1f84f1612606f3753f205a4e9a2efd6fe5b4c573a6269b2cc6c3003d44a0d127" dependencies = [ - "proc-macro2", - "quote", - "syn", + "serde", ] [[package]] -name = "ipnetwork" -version = "0.18.0" +name = "itertools" +version = "0.10.5" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "4088d739b183546b239688ddbc79891831df421773df95e236daf7867866d355" +checksum = "b0fd2260e829bddf4cb6ea802289de2f86d6a7a690192fbe91b3f46e0f2c8473" dependencies = [ - "serde", + "either", ] [[package]] name = "itoa" -version = "1.0.1" +version = "1.0.3" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "1aab8fc367588b89dcee83ab0fd66b72b50b72fa1904d7095045ace2b0c81c35" +checksum = "6c8af84674fe1f223a982c933a0ee1086ac4d4052aa0fb8060c12c6ad838e754" [[package]] name = "js-sys" -version = "0.3.56" +version = "0.3.60" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "a38fc24e30fd564ce974c02bf1d337caddff65be6cc4735a1f7eab22a7440f04" +checksum = "49409df3e3bf0856b916e2ceaca09ee28e6871cf7d9ce97a692cacfdb2a25a47" dependencies = [ "wasm-bindgen", ] @@ -923,9 +836,9 @@ dependencies = [ [[package]] name = "keccak" -version = "0.1.0" +version = "0.1.2" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "67c21572b4949434e4fc1e1978b99c5f77064153c59d998bf13ecd96fb5ecba7" +checksum = "f9b7d56ba4a8344d6be9729995e6b06f928af29998cdf79fe390cbf6b1fee838" [[package]] name = "kv-log-macro" @@ -947,9 +860,9 @@ dependencies = [ [[package]] name = "libc" -version = "0.2.120" +version = "0.2.133" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "ad5c14e80759d0939d013e6ca49930e59fc53dd8e5009132f76240c179380c09" +checksum = "c0f80d65747a3e43d1596c7c5492d95d5edddaabd45a7fcdb02b95f644164966" [[package]] name = "libloading" @@ -963,46 +876,35 @@ dependencies = [ [[package]] name = "libm" -version = "0.2.2" -source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "33a33a362ce288760ec6a508b94caaec573ae7d3bbbd91b87aa0bad4456839db" - -[[package]] -name = "linked-hash-map" -version = "0.5.4" +version = "0.2.5" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "7fb9b38af92608140b86b693604b9ffcc5824240a484d1ecd4795bacb2fe88f3" +checksum = "292a948cd991e376cf75541fe5b97a1081d713c618b4f1b9500f8844e49eb565" [[package]] name = "lock_api" -version = "0.4.6" +version = "0.4.9" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "88943dd7ef4a2e5a4bfa2753aaab3013e34ce2533d1996fb18ef591e315e2b3b" +checksum = "435011366fe56583b16cf956f9df0095b405b82d76425bc8981c0e22e60ec4df" dependencies = [ + "autocfg", "scopeguard", ] [[package]] name = "log" -version = "0.4.14" +version = "0.4.17" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "51b9bbe6c47d51fc3e1a9b945965946b4c44142ab8792c50835a980d362c2710" +checksum = "abb12e687cfb44aa40f41fc3978ef76448f9b6038cad6aef4259d3c095a2382e" dependencies = [ "cfg-if", "value-bag", ] -[[package]] -name = "maplit" -version = "1.0.2" -source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "3e2e65a1a2e43cfcb47a895c4c8b10d1f4a61097f9f254f183aee60cad9c651d" - [[package]] name = "memchr" -version = "2.4.1" +version = "2.5.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "308cc39be01b73d0d18f82a0e7b2a3df85245f84af96fdddc5d202d27e47b86a" +checksum = "2dffe52ecf27772e601905b7522cb4ef790d2cc203488bbd0e2fe85fcb74566d" [[package]] name = "memoffset" @@ -1010,43 +912,19 @@ version = "0.6.5" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "5aa361d4faea93603064a027415f07bd8e1d5c88c9fbf68bf56a285428fd79ce" dependencies = [ - "autocfg 1.1.0", -] - -[[package]] -name = "mio" -version = "0.7.14" -source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "8067b404fe97c70829f082dec8bcf4f71225d7eaea1d8645349cb76fa06205cc" -dependencies = [ - "libc", - "log", - "miow", - "ntapi", - "winapi", + "autocfg", ] [[package]] name = "mio" -version = "0.8.2" +version = "0.8.4" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "52da4364ffb0e4fe33a9841a98a3f3014fb964045ce4f7a45a398243c8d6b0c9" +checksum = "57ee1c23c7c63b0c9250c339ffdc69255f110b298b901b9f6c82547b7b87caaf" dependencies = [ "libc", "log", - "miow", - "ntapi", - "wasi 0.11.0+wasi-snapshot-preview1", - "winapi", -] - -[[package]] -name = "miow" -version = "0.3.7" -source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "b9f1c5b025cda876f66ef43a113f91ebc9f4ccef34843000e0adf6ebbab84e21" -dependencies = [ - "winapi", + "wasi", + "windows-sys", ] [[package]] @@ -1060,9 +938,9 @@ dependencies = [ [[package]] name = "nix" -version = "0.22.3" +version = "0.23.1" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "e4916f159ed8e5de0082076562152a76b7a1f64a01fd9d1e0fea002c37624faf" +checksum = "9f866317acbd3a240710c63f065ffb1e4fd466259045ccb504130b7f668f35c6" dependencies = [ "bitflags", "cc", @@ -1073,33 +951,30 @@ dependencies = [ [[package]] name = "nix" -version = "0.23.1" +version = "0.25.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "9f866317acbd3a240710c63f065ffb1e4fd466259045ccb504130b7f668f35c6" +checksum = "e322c04a9e3440c327fca7b6c8a63e6890a32fa2ad689db972425f07e0d22abb" dependencies = [ + "autocfg", "bitflags", - "cc", "cfg-if", "libc", "memoffset", + "pin-utils", ] [[package]] -name = "ntapi" -version = "0.3.7" +name = "no-std-net" +version = "0.6.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "c28774a7fd2fbb4f0babd8237ce554b73af68021b5f695a3cebd6c59bac0980f" -dependencies = [ - "winapi", -] +checksum = "43794a0ace135be66a25d3ae77d41b91615fb68ae937f904090203e81f755b65" [[package]] name = "num-bigint-dig" -version = "0.7.0" +version = "0.8.1" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "4547ee5541c18742396ae2c895d0717d0f886d8823b8399cdaf7b07d63ad0480" +checksum = "566d173b2f9406afbc5510a90925d5a2cd80cae4605631f1212303df265de011" dependencies = [ - "autocfg 0.1.8", "byteorder", "lazy_static", "libm", @@ -1113,32 +988,32 @@ dependencies = [ [[package]] name = "num-integer" -version = "0.1.44" +version = "0.1.45" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "d2cc698a63b549a70bc047073d2949cce27cd1c7b0a4a862d08a8031bc2801db" +checksum = "225d3389fb3509a24c93f5c29eb6bde2586b98d9f016636dff58d7c6f7569cd9" dependencies = [ - "autocfg 1.1.0", + "autocfg", "num-traits", ] [[package]] name = "num-iter" -version = "0.1.42" +version = "0.1.43" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "b2021c8337a54d21aca0d59a92577a029af9431cb59b909b03252b9c164fad59" +checksum = "7d03e6c028c5dc5cac6e2dec0efda81fc887605bb3d884578bb6d6bf7514e252" dependencies = [ - "autocfg 1.1.0", + "autocfg", "num-integer", "num-traits", ] [[package]] name = "num-traits" -version = "0.2.14" +version = "0.2.15" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "9a64b1ec5cda2586e284722486d802acf1f7dbdc623e2bfc57e65ca1cd099290" +checksum = "578ede34cf02f8924ab9447f50c28075b4d3e5b269972345e7e0372b38c6cdcd" dependencies = [ - "autocfg 1.1.0", + "autocfg", "libm", ] @@ -1154,21 +1029,9 @@ dependencies = [ [[package]] name = "once_cell" -version = "1.10.0" -source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "87f3e037eac156d1775da914196f0f37741a274155e34a0b7e427c35d2a2ecb9" - -[[package]] -name = "opaque-debug" -version = "0.2.3" -source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "2839e79665f131bdb5782e51f2c6c9599c133c6098982a54c794358bf432529c" - -[[package]] -name = "opaque-debug" -version = "0.3.0" +version = "1.15.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "624a8340c38c1b80fd549087862da4ba43e08858af025b236e509b6649fc13d5" +checksum = "e82dad04139b71a90c080c8463fe0dc7902db5192d939bd0950f074d014339e1" [[package]] name = "openssl-probe" @@ -1178,13 +1041,19 @@ checksum = "ff011a302c396a5197692431fc1948019154afc178baf7d8e37367442a4601cf" [[package]] name = "ordered-float" -version = "2.10.0" +version = "3.1.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "7940cf2ca942593318d07fcf2596cdca60a85c9e7fab408a5e21a4f9dcd40d87" +checksum = "98ffdb14730ed2ef599c65810c15b000896e21e8776b512de0db0c3d7335cc2a" dependencies = [ "num-traits", ] +[[package]] +name = "os_str_bytes" +version = "6.3.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "9ff7415e9ae3fff1225851df9e0d9e4e5479f947619774677a63572e55e80eff" + [[package]] name = "parking" version = "2.0.0" @@ -1193,77 +1062,63 @@ checksum = "427c3892f9e783d91cc128285287e70a59e206ca452770ece88a76f7a3eddd72" [[package]] name = "parking_lot" -version = "0.11.2" +version = "0.12.1" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "7d17b78036a60663b797adeaee46f5c9dfebb86948d1255007a1d6be0271ff99" +checksum = "3742b2c103b9f06bc9fff0a37ff4912935851bee6d36f3c02bcc755bcfec228f" dependencies = [ - "instant", "lock_api", "parking_lot_core", ] [[package]] name = "parking_lot_core" -version = "0.8.5" +version = "0.9.3" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "d76e8e1493bcac0d2766c42737f34458f1c8c50c0d23bcb24ea953affb273216" +checksum = "09a279cbf25cb0757810394fbc1e359949b59e348145c643a939a525692e6929" dependencies = [ "cfg-if", - "instant", "libc", "redox_syscall", "smallvec", - "winapi", + "windows-sys", ] [[package]] name = "paste" -version = "0.1.18" -source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "45ca20c77d80be666aef2b45486da86238fabe33e38306bd3118fe4af33fa880" -dependencies = [ - "paste-impl", - "proc-macro-hack", -] - -[[package]] -name = "paste" -version = "1.0.6" +version = "1.0.9" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "0744126afe1a6dd7f394cb50a716dbe086cb06e255e53d8d0185d82828358fb5" +checksum = "b1de2e551fb905ac83f73f7aedf2f0cb4a0da7e35efa24a202a936269f1f18e1" [[package]] -name = "paste-impl" -version = "0.1.18" +name = "pem-rfc7468" +version = "0.3.1" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "d95a7db200b97ef370c8e6de0088252f7e0dfff7d047a28528e47456c0fc98b6" +checksum = "01de5d978f34aa4b2296576379fcc416034702fd94117c56ffd8a1a767cefb30" dependencies = [ - "proc-macro-hack", + "base64ct", ] [[package]] -name = "pem-rfc7468" -version = "0.2.4" +name = "percent-encoding" +version = "2.2.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "84e93a3b1cc0510b03020f33f21e62acdde3dcaef432edc95bea377fbd4c2cd4" -dependencies = [ - "base64ct", -] +checksum = "478c572c3d73181ff3c2539045f6eb99e5491218eae919370993b890cdbdd98e" [[package]] name = "pest" -version = "2.1.3" +version = "2.3.1" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "10f4872ae94d7b90ae48754df22fd42ad52ce740b8f370b03da4835417403e53" +checksum = "cb779fcf4bb850fbbb0edc96ff6cf34fd90c4b1a112ce042653280d9a7364048" dependencies = [ + "thiserror", "ucd-trie", ] [[package]] name = "pest_derive" -version = "2.1.0" +version = "2.3.1" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "833d1ae558dc601e9a60366421196a8d94bc0ac980476d0b67e1d0988d72b2d0" +checksum = "502b62a6d0245378b04ffe0a7fb4f4419a4815fce813bd8a0ec89a56e07d67b1" dependencies = [ "pest", "pest_generator", @@ -1271,9 +1126,9 @@ dependencies = [ [[package]] name = "pest_generator" -version = "2.1.3" +version = "2.3.1" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "99b8db626e31e5b81787b9783425769681b347011cc59471e33ea46d2ea0cf55" +checksum = "451e629bf49b750254da26132f1a5a9d11fd8a95a3df51d15c4abd1ba154cb6c" dependencies = [ "pest", "pest_meta", @@ -1284,20 +1139,20 @@ dependencies = [ [[package]] name = "pest_meta" -version = "2.1.3" +version = "2.3.1" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "54be6e404f5317079812fc8f9f5279de376d8856929e21c184ecf6bbd692a11d" +checksum = "bcec162c71c45e269dfc3fc2916eaeb97feab22993a21bcce4721d08cd7801a6" dependencies = [ - "maplit", + "once_cell", "pest", - "sha-1", + "sha1", ] [[package]] name = "petgraph" -version = "0.6.0" +version = "0.6.2" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "4a13a2fa9d0b63e5f22328828741e523766fff0ee9e779316902290dff3f824f" +checksum = "e6d5014253a1331579ce62aa67443b4a658c5e7dd03d4bc6d302b94474888143" dependencies = [ "fixedbitset", "indexmap", @@ -1305,18 +1160,18 @@ dependencies = [ [[package]] name = "pin-project" -version = "1.0.10" +version = "1.0.12" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "58ad3879ad3baf4e44784bc6a718a8698867bb991f8ce24d1bcbe2cfb4c3a75e" +checksum = "ad29a609b6bcd67fee905812e544992d216af9d755757c05ed2d0e15a74c6ecc" dependencies = [ "pin-project-internal", ] [[package]] name = "pin-project-internal" -version = "1.0.10" +version = "1.0.12" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "744b6f092ba29c3650faf274db506afd39944f48420f6c86b17cfe0ee1cb36bb" +checksum = "069bdb1e05adc7a8990dce9cc75370895fbe4e3d58b9b73bf1aee56359344a55" dependencies = [ "proc-macro2", "quote", @@ -1325,9 +1180,9 @@ dependencies = [ [[package]] name = "pin-project-lite" -version = "0.2.8" +version = "0.2.9" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "e280fbe77cc62c91527259e9442153f4688736748d24660126286329742b4c6c" +checksum = "e0a7ae3ac2f1173085d398531c705756c94a4c56843785df85a60c1a0afac116" [[package]] name = "pin-utils" @@ -1337,53 +1192,52 @@ checksum = "8b870d8c151b6f2fb93e84a13146138f05d02ed11c7e7c54f8826aaaf7c9f184" [[package]] name = "pkcs1" -version = "0.2.4" +version = "0.3.3" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "116bee8279d783c0cf370efa1a94632f2108e5ef0bb32df31f051647810a4e2c" +checksum = "a78f66c04ccc83dd4486fd46c33896f4e17b24a7a3a6400dedc48ed0ddd72320" dependencies = [ "der", - "pem-rfc7468", + "pkcs8", "zeroize", ] [[package]] name = "pkcs8" -version = "0.7.6" +version = "0.8.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "ee3ef9b64d26bad0536099c816c6734379e45bbd5f14798def6809e5cc350447" +checksum = "7cabda3fb821068a9a4fab19a683eac3af12edf0f34b94a8be53c4972b8149d0" dependencies = [ "der", - "pem-rfc7468", - "pkcs1", "spki", "zeroize", ] [[package]] name = "pnet" -version = "0.28.0" +version = "0.31.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "4b6d2a0409666964722368ef5fb74b9f93fac11c18bef3308693c16c6733f103" +checksum = "0caaf5b11fd907ff15cf14a4477bfabca4b37ab9e447a4f8dead969a59cdafad" dependencies = [ - "ipnetwork", "pnet_base", "pnet_datalink", "pnet_packet", - "pnet_sys", "pnet_transport", ] [[package]] name = "pnet_base" -version = "0.28.0" +version = "0.31.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "25488cd551a753dcaaa6fffc9f69a7610a412dd8954425bf7ffad5f7d1156fb8" +checksum = "f9d3a993d49e5fd5d4d854d6999d4addca1f72d86c65adf224a36757161c02b6" +dependencies = [ + "no-std-net", +] [[package]] name = "pnet_datalink" -version = "0.28.0" +version = "0.31.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "d4d1f8ab1ef6c914cf51dc5dfe0be64088ea5f3b08bbf5a31abc70356d271198" +checksum = "e466faf03a98ad27f6e15cd27a2b7cc89e73e640a43527742977bc503c37f8aa" dependencies = [ "ipnetwork", "libc", @@ -1394,9 +1248,9 @@ dependencies = [ [[package]] name = "pnet_macros" -version = "0.28.0" +version = "0.31.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "30490e0852e58402b8fae0d39897b08a24f493023a4d6cf56b2e30f31ed57548" +checksum = "48dd52a5211fac27e7acb14cfc9f30ae16ae0e956b7b779c8214c74559cef4c3" dependencies = [ "proc-macro2", "quote", @@ -1406,18 +1260,18 @@ dependencies = [ [[package]] name = "pnet_macros_support" -version = "0.28.0" +version = "0.31.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "d4714e10f30cab023005adce048f2d30dd4ac4f093662abf2220855655ef8f90" +checksum = "89de095dc7739349559913aed1ef6a11e73ceade4897dadc77c5e09de6740750" dependencies = [ "pnet_base", ] [[package]] name = "pnet_packet" -version = "0.28.0" +version = "0.31.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "8588067671d03c9f4254b2e66fecb4d8b93b5d3e703195b84f311cd137e32130" +checksum = "bc3b5111e697c39c8b9795b9fdccbc301ab696699e88b9ea5a4e4628978f495f" dependencies = [ "glob", "pnet_base", @@ -1427,9 +1281,9 @@ dependencies = [ [[package]] name = "pnet_sys" -version = "0.28.0" +version = "0.31.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "d9a3f32b0df45515befd19eed04616f6b56a488da92afc61164ef455e955f07f" +checksum = "328e231f0add6d247d82421bf3790b4b33b39c8930637f428eef24c4c6a90805" dependencies = [ "libc", "winapi", @@ -1437,9 +1291,9 @@ dependencies = [ [[package]] name = "pnet_transport" -version = "0.28.0" +version = "0.31.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "932b2916d693bcc5fa18443dc99142e0a6fd31a6ce75a511868f7174c17e2bce" +checksum = "ff597185e6f1f5671b3122e4dba892a1c73e17c17e723d7669bd9299cbe7f124" dependencies = [ "libc", "pnet_base", @@ -1449,10 +1303,11 @@ dependencies = [ [[package]] name = "polling" -version = "2.2.0" +version = "2.3.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "685404d509889fade3e86fe3a5803bca2ec09b0c0778d5ada6ec8bf7a8de5259" +checksum = "899b00b9c8ab553c743b3e11e87c5c7d423b2a2de229ba95b24a756344748011" dependencies = [ + "autocfg", "cfg-if", "libc", "log", @@ -1474,70 +1329,57 @@ checksum = "dbf0c48bc1d91375ae5c3cd81e3722dff1abcf81a30960240640d223f59fe0e5" [[package]] name = "proc-macro2" -version = "1.0.36" +version = "1.0.43" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "c7342d5883fbccae1cc37a2353b09c87c9b0f3afd73f5fb9bba687a1f733b029" +checksum = "0a2ca2c61bc9f3d74d2886294ab7b9853abd9c1ad903a3ac7815c58989bb7bab" dependencies = [ - "unicode-xid", + "unicode-ident", ] [[package]] name = "pyo3" -version = "0.15.1" +version = "0.17.1" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "7cf01dbf1c05af0a14c7779ed6f3aa9deac9c3419606ac9de537a2d649005720" +checksum = "12f72538a0230791398a0986a6518ebd88abc3fded89007b506ed072acc831e1" dependencies = [ "cfg-if", "indoc", "libc", + "memoffset", "parking_lot", - "paste 0.1.18", "pyo3-build-config", + "pyo3-ffi", "pyo3-macros", "unindent", ] [[package]] -name = "pyo3-asyncio" -version = "0.15.0" +name = "pyo3-build-config" +version = "0.17.1" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "0897c7e36110a32b726b975359b2bbe90c37fcf1266046d3b1c08c616a47a886" +checksum = "fc4cf18c20f4f09995f3554e6bcf9b09bd5e4d6b67c562fdfaafa644526ba479" dependencies = [ - "async-std", - "futures", - "inventory", "once_cell", - "pin-project-lite", - "pyo3", - "pyo3-asyncio-macros", -] - -[[package]] -name = "pyo3-asyncio-macros" -version = "0.15.0" -source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "4daada7260db6d6dbc1f9b50f3e3d21f9eb28c034722c9a61ac05ff56ffd62db" -dependencies = [ - "proc-macro2", - "quote", - "syn", + "target-lexicon", ] [[package]] -name = "pyo3-build-config" -version = "0.15.1" +name = "pyo3-ffi" +version = "0.17.1" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "dbf9e4d128bfbddc898ad3409900080d8d5095c379632fbbfbb9c8cfb1fb852b" +checksum = "a41877f28d8ebd600b6aa21a17b40c3b0fc4dfe73a27b6e81ab3d895e401b0e9" dependencies = [ - "once_cell", + "libc", + "pyo3-build-config", ] [[package]] name = "pyo3-macros" -version = "0.15.1" +version = "0.17.1" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "67701eb32b1f9a9722b4bc54b548ff9d7ebfded011c12daece7b9063be1fd755" +checksum = "2e81c8d4bcc2f216dc1b665412df35e46d12ee8d3d046b381aad05f1fcf30547" dependencies = [ + "proc-macro2", "pyo3-macros-backend", "quote", "syn", @@ -1545,21 +1387,20 @@ dependencies = [ [[package]] name = "pyo3-macros-backend" -version = "0.15.1" +version = "0.17.1" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "f44f09e825ee49a105f2c7b23ebee50886a9aee0746f4dd5a704138a64b0218a" +checksum = "85752a767ee19399a78272cc2ab625cd7d373b2e112b4b13db28de71fa892784" dependencies = [ "proc-macro2", - "pyo3-build-config", "quote", "syn", ] [[package]] name = "quinn" -version = "0.8.1" +version = "0.8.5" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "584865613896a1f644d757e52c45c573441c8b04cac38ac13990b0235203db66" +checksum = "5b435e71d9bfa0d8889927231970c51fb89c58fa63bffcab117c9c7a41e5ef8f" dependencies = [ "bytes", "futures-channel", @@ -1567,7 +1408,7 @@ dependencies = [ "fxhash", "quinn-proto", "quinn-udp", - "rustls 0.20.4", + "rustls 0.20.6", "thiserror", "tokio", "tracing", @@ -1576,15 +1417,15 @@ dependencies = [ [[package]] name = "quinn-proto" -version = "0.8.1" +version = "0.8.4" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "d2b1562bf4998b0c6d1841a4742b7103bb82cdde61374833de826bab9e8ad498" +checksum = "3fce546b9688f767a57530652488420d419a8b1f44a478b451c3d1ab6d992a55" dependencies = [ "bytes", "fxhash", "rand", "ring", - "rustls 0.20.4", + "rustls 0.20.6", "rustls-native-certs", "rustls-pemfile 0.2.1", "slab", @@ -1596,13 +1437,12 @@ dependencies = [ [[package]] name = "quinn-udp" -version = "0.1.1" +version = "0.1.3" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "df185e5e5f7611fa6e628ed8f9633df10114b03bbaecab186ec55822c44ac727" +checksum = "9f832d8958db3e84d2ec93b5eb2272b45aa23cf7f8fe6e79f578896f4e6c231b" dependencies = [ "futures-util", "libc", - "mio 0.7.14", "quinn-proto", "socket2", "tokio", @@ -1611,9 +1451,9 @@ dependencies = [ [[package]] name = "quote" -version = "1.0.16" +version = "1.0.21" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "b4af2ec4714533fcdf07e886f17025ace8b997b9ce51204ee69b6da831c3da57" +checksum = "bbe448f377a7d6961e30f5955f9b8d106c3f5e449d493ee1b125c1d43c2b5179" dependencies = [ "proc-macro2", ] @@ -1641,27 +1481,27 @@ dependencies = [ [[package]] name = "rand_core" -version = "0.6.3" +version = "0.6.4" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "d34f1408f55294453790c48b2f1ebbb1c5b4b7563eb1f418bcfcfdbb06ebb4e7" +checksum = "ec0be4795e2f6a28069bec0b5ff3e2ac9bafc99e6a9a7dc3547996c5c816922c" dependencies = [ "getrandom", ] [[package]] name = "redox_syscall" -version = "0.2.11" +version = "0.2.16" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "8380fe0152551244f0747b1bf41737e0f8a74f97a14ccefd1148187271634f3c" +checksum = "fb5a58c1855b4b6819d59012155603f0b22ad30cad752600aadfcb695265519a" dependencies = [ "bitflags", ] [[package]] name = "redox_users" -version = "0.4.2" +version = "0.4.3" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "7776223e2696f1aa4c6b0170e83212f47296a00424305117d013dfe86fb0fe55" +checksum = "b033d837a7cf162d7993aded9304e30a83213c648b6e389db233191f891e5c2b" dependencies = [ "getrandom", "redox_syscall", @@ -1670,9 +1510,9 @@ dependencies = [ [[package]] name = "regex" -version = "1.5.5" +version = "1.6.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "1a11647b6b25ff05a515cb92c365cec08801e83423a235b51e231e1808747286" +checksum = "4c4eb3267174b8c6c2f654116623910a0fef09c4753f8dd83db29c48a0df988b" dependencies = [ "aho-corasick", "memchr", @@ -1681,9 +1521,9 @@ dependencies = [ [[package]] name = "regex-syntax" -version = "0.6.25" +version = "0.6.27" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "f497285884f3fcff424ffc933e56d7cbca511def0c9831a7f9b5f6153e3cc89b" +checksum = "a3f87b73ce11b1619a3c6332f45341e0047173771e8b8b73f87bfeefb7b56244" [[package]] name = "ring" @@ -1700,22 +1540,32 @@ dependencies = [ "winapi", ] +[[package]] +name = "ringbuffer-spsc" +version = "0.1.6" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "3c98a13710286c8fa1b6b549078247332183534f31910d4c929aa45c03a98e50" +dependencies = [ + "array-init", + "cache-padded", +] + [[package]] name = "rsa" -version = "0.5.0" +version = "0.6.1" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "e05c2603e2823634ab331437001b411b9ed11660fbc4066f3908c84a9439260d" +checksum = "4cf22754c49613d2b3b119f0e5d46e34a2c628a937e3024b8762de4e7d8c710b" dependencies = [ "byteorder", - "digest 0.9.0", - "lazy_static", + "digest", "num-bigint-dig", "num-integer", "num-iter", "num-traits", "pkcs1", "pkcs8", - "rand", + "rand_core", + "smallvec", "subtle", "zeroize", ] @@ -1744,9 +1594,9 @@ dependencies = [ [[package]] name = "rustls" -version = "0.20.4" +version = "0.20.6" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "4fbfeb8d0ddb84706bc597a5574ab8912817c52a397f819e5b614e2265206921" +checksum = "5aab8ee6c7097ed6057f43c187a62418d0c05a4bd5f18b3571db50ee0f9ce033" dependencies = [ "log", "ring", @@ -1756,12 +1606,12 @@ dependencies = [ [[package]] name = "rustls-native-certs" -version = "0.6.1" +version = "0.6.2" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "5ca9ebdfa27d3fc180e42879037b5338ab1c040c06affd00d8338598e7800943" +checksum = "0167bac7a9f490495f3c33013e7722b53cb087ecbe082fb0c6387c96f634ea50" dependencies = [ "openssl-probe", - "rustls-pemfile 0.2.1", + "rustls-pemfile 1.0.1", "schannel", "security-framework", ] @@ -1777,27 +1627,27 @@ dependencies = [ [[package]] name = "rustls-pemfile" -version = "0.3.0" +version = "1.0.1" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "1ee86d63972a7c661d1536fefe8c3c8407321c3df668891286de28abcd087360" +checksum = "0864aeff53f8c05aa08d86e5ef839d3dfcf07aeba2db32f12db0ef716e87bd55" dependencies = [ "base64", ] [[package]] name = "ryu" -version = "1.0.9" +version = "1.0.11" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "73b4b750c782965c211b42f022f59af1fbceabdd026623714f104152f1ec149f" +checksum = "4501abdff3ae82a1c1b477a17252eb69cee9e66eb915c1abaa4f44d873df9f09" [[package]] name = "schannel" -version = "0.1.19" +version = "0.1.20" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "8f05ba609c234e60bee0d547fe94a4c7e9da733d1c962cf6e59efa4cd9c8bc75" +checksum = "88d6731146462ea25d9244b2ed5fd1d716d25c52e4d54aa4fb0f3c4e9854dbe2" dependencies = [ "lazy_static", - "winapi", + "windows-sys", ] [[package]] @@ -1828,9 +1678,9 @@ dependencies = [ [[package]] name = "security-framework" -version = "2.6.1" +version = "2.7.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "2dc14f172faf8a0194a3aded622712b0de276821addc574fa54fc0a1167e10dc" +checksum = "2bc1bb97804af6631813c55739f771071e0f2ed33ee20b68c86ec505d906356c" dependencies = [ "bitflags", "core-foundation", @@ -1851,24 +1701,24 @@ dependencies = [ [[package]] name = "semver" -version = "1.0.6" +version = "1.0.14" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "a4a3381e03edd24287172047536f20cabde766e2cd3e65e6b00fb3af51c4f38d" +checksum = "e25dfac463d778e353db5be2449d1cce89bd6fd23c9f1ea21310ce6e5a1b29c4" [[package]] name = "serde" -version = "1.0.136" +version = "1.0.144" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "ce31e24b01e1e524df96f1c2fdd054405f8d7376249a5110886fb4b658484789" +checksum = "0f747710de3dcd43b88c9168773254e809d8ddbdf9653b84e2554ab219f17860" dependencies = [ "serde_derive", ] [[package]] name = "serde_derive" -version = "1.0.136" +version = "1.0.144" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "08597e7152fcd306f41838ed3e37be9eaeed2b61c42e2117266a554fab4662f9" +checksum = "94ed3a816fb1d101812f83e789f888322c34e291f894f19590dc310963e87a00" dependencies = [ "proc-macro2", "quote", @@ -1877,9 +1727,9 @@ dependencies = [ [[package]] name = "serde_json" -version = "1.0.79" +version = "1.0.85" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "8e8d9fa5c3b304765ce1fd9c4c8a3de2c8db365a5b91be52f186efc675681d95" +checksum = "e55a28e3aaef9d5ce0506d0a14dbba8054ddc7e499ef522dd8b26859ec9d4a44" dependencies = [ "itoa", "ryu", @@ -1888,67 +1738,65 @@ dependencies = [ [[package]] name = "serde_yaml" -version = "0.8.23" +version = "0.9.13" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "a4a521f2940385c165a24ee286aa8599633d162077a54bdcae2a6fd5a7bfa7a0" +checksum = "8613d593412a0deb7bbd8de9d908efff5a0cb9ccd8f62c641e7b2ed2f57291d1" dependencies = [ "indexmap", + "itoa", "ryu", "serde", - "yaml-rust", + "unsafe-libyaml", ] [[package]] -name = "sha-1" -version = "0.8.2" +name = "sha1" +version = "0.10.5" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "f7d94d0bede923b3cea61f3f1ff57ff8cdfd77b400fb8f9998949e0cf04163df" +checksum = "f04293dc80c3993519f2d7f6f511707ee7094fe0c6d3406feb330cdb3540eba3" dependencies = [ - "block-buffer 0.7.3", - "digest 0.8.1", - "fake-simd", - "opaque-debug 0.2.3", + "cfg-if", + "cpufeatures", + "digest", ] [[package]] name = "sha3" -version = "0.9.1" +version = "0.10.5" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "f81199417d4e5de3f04b1e871023acea7389672c4135918f05aa9cbf2f2fa809" +checksum = "e2904bea16a1ae962b483322a1c7b81d976029203aea1f461e51cd7705db7ba9" dependencies = [ - "block-buffer 0.9.0", - "digest 0.9.0", + "digest", "keccak", - "opaque-debug 0.3.0", ] [[package]] name = "shared_memory" -version = "0.12.0" +version = "0.12.4" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "681a9e90340f748af3a1cc52eb2c040eee29f976b763e99ad90fc0c5df6f9791" +checksum = "ba8593196da75d9dc4f69349682bd4c2099f8cde114257d1ef7ef1b33d1aba54" dependencies = [ "cfg-if", "libc", - "nix 0.22.3", + "nix 0.23.1", "rand", - "winapi", + "win-sys", ] [[package]] name = "shellexpand" -version = "2.1.0" +version = "2.1.2" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "83bdb7831b2d85ddf4a7b148aa19d0587eddbe8671a436b7bd1182eaad0f2829" +checksum = "7ccc8076840c4da029af4f87e4e8daeb0fca6b87bbb02e10cb60b791450e11e4" dependencies = [ - "dirs-next", + "dirs", ] [[package]] name = "signal-hook" -version = "0.3.13" +version = "0.3.14" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "647c97df271007dcea485bb74ffdb57f2e683f1306c854f468a0c244badabf2d" +checksum = "a253b5e89e2698464fc26b545c9edceb338e18a89effeeecfea192c3025be29d" dependencies = [ "libc", "signal-hook-registry", @@ -1965,21 +1813,24 @@ dependencies = [ [[package]] name = "slab" -version = "0.4.5" +version = "0.4.7" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "9def91fd1e018fe007022791f865d0ccc9b3a0d5001e01aabb8b40e46000afb5" +checksum = "4614a76b2a8be0058caa9dbbaf66d988527d86d003c11a94fbd335d7661edcef" +dependencies = [ + "autocfg", +] [[package]] name = "smallvec" -version = "1.8.0" +version = "1.9.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "f2dd574626839106c320a323308629dcb1acfc96e32a8cba364ddc61ac23ee83" +checksum = "2fd0db749597d91ff862fd1d55ea87f7855a744a8425a64695b6fca237d1dad1" [[package]] name = "socket2" -version = "0.4.4" +version = "0.4.7" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "66d72b759436ae32898a2af0a14218dbf55efde3feeb170eb623637db85ee1e0" +checksum = "02e2d2db9033d13a1567121ddd7a095ee144db4e1ca1b1bda3419bc0da294ebd" dependencies = [ "libc", "winapi", @@ -1993,19 +1844,20 @@ checksum = "6e63cff320ae2c57904679ba7cb63280a3dc4613885beafb148ee7bf9aa9042d" [[package]] name = "spin" -version = "0.9.2" +version = "0.9.4" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "511254be0c5bcf062b019a6c89c01a664aa359ded62f78aa72c6fc137c0590e5" +checksum = "7f6002a767bff9e83f8eeecf883ecb8011875a21ae8da43bffb817a57e78cc09" dependencies = [ "lock_api", ] [[package]] name = "spki" -version = "0.4.1" +version = "0.5.4" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "5c01a0c15da1b0b0e1494112e7af814a678fec9bd157881b49beac661e9b6f32" +checksum = "44d01ac02a6ccf3e07db148d2be087da624fea0221a16152ed01f0496a6b0a27" dependencies = [ + "base64ct", "der", ] @@ -2023,9 +1875,9 @@ dependencies = [ [[package]] name = "strsim" -version = "0.8.0" +version = "0.10.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "8ea5119cdb4c55b55d432abb513a0429384878c15dde60cc77b1c99de1a95a6a" +checksum = "73473c0e59e6d5812c5dfe2a064a6444949f089e20eec9a2e5506596494e4623" [[package]] name = "subtle" @@ -2035,26 +1887,20 @@ checksum = "6bdef32e8150c2a081110b42772ffe7d7c9032b606bc226c8260fd97e0976601" [[package]] name = "syn" -version = "1.0.89" +version = "1.0.100" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "ea297be220d52398dcc07ce15a209fce436d361735ac1db700cab3b6cdfb9f54" +checksum = "52205623b1b0f064a4e71182c3b18ae902267282930c6d5462c91b859668426e" dependencies = [ "proc-macro2", "quote", - "unicode-xid", + "unicode-ident", ] [[package]] -name = "synstructure" -version = "0.12.6" +name = "target-lexicon" +version = "0.12.4" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "f36bdaa60a83aca3921b5259d5400cbf5e90fc51931376a9bd4a0eb79aa7210f" -dependencies = [ - "proc-macro2", - "quote", - "syn", - "unicode-xid", -] +checksum = "c02424087780c9b71cc96799eaeddff35af2bc513278cda5c99fc1f5d026d3c1" [[package]] name = "termcolor" @@ -2067,27 +1913,24 @@ dependencies = [ [[package]] name = "textwrap" -version = "0.11.0" +version = "0.15.1" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "d326610f408c7a4eb6f51c37c330e496b08506c9457c9d34287ecc38809fb060" -dependencies = [ - "unicode-width", -] +checksum = "949517c0cf1bf4ee812e2e07e08ab448e3ae0d23472aee8a06c985f0c8815b16" [[package]] name = "thiserror" -version = "1.0.30" +version = "1.0.35" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "854babe52e4df1653706b98fcfc05843010039b406875930a70e4d9644e5c417" +checksum = "c53f98874615aea268107765aa1ed8f6116782501d18e53d08b471733bea6c85" dependencies = [ "thiserror-impl", ] [[package]] name = "thiserror-impl" -version = "1.0.30" +version = "1.0.35" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "aa32fd3f627f367fe16f893e2597ae3c05020f8bba2666a4e6ea73d377e5714b" +checksum = "f8b463991b4eab2d801e724172285ec4195c650e8ec79b149e6c2a8e6dd3f783" dependencies = [ "proc-macro2", "quote", @@ -2096,9 +1939,9 @@ dependencies = [ [[package]] name = "tinyvec" -version = "1.5.1" +version = "1.6.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "2c1c1d5a42b6245520c249549ec267180beaffcc0615401ac8e31853d4b6d8d2" +checksum = "87cc5ceb3875bb20c2890005a4e226a4651264a5c75edb2421b52861a0a0cb50" dependencies = [ "tinyvec_macros", ] @@ -2111,13 +1954,15 @@ checksum = "cda74da7e1a664f795bb1f8a87ec406fb89a02522cf6e50620d016add6dbbf5c" [[package]] name = "tokio" -version = "1.17.0" +version = "1.21.1" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "2af73ac49756f3f7c01172e34a23e5d0216f6c32333757c2c61feb2bbff5a5ee" +checksum = "0020c875007ad96677dcc890298f4b942882c5d4eb7cc8f439fc3bf813dc9c95" dependencies = [ + "autocfg", "libc", - "mio 0.8.2", + "mio", "num_cpus", + "once_cell", "pin-project-lite", "socket2", "winapi", @@ -2125,9 +1970,9 @@ dependencies = [ [[package]] name = "tracing" -version = "0.1.32" +version = "0.1.36" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "4a1bdf54a7c28a2bbf701e1d2233f6c77f473486b94bee4f9678da5a148dca7f" +checksum = "2fce9567bd60a67d08a16488756721ba392f24f29006402881e43b19aac64307" dependencies = [ "cfg-if", "pin-project-lite", @@ -2137,9 +1982,9 @@ dependencies = [ [[package]] name = "tracing-attributes" -version = "0.1.20" +version = "0.1.22" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "2e65ce065b4b5c53e73bb28912318cb8c9e9ad3921f1d669eb0e68b4c8143a2b" +checksum = "11c75893af559bc8e10716548bdef5cb2b983f8e637db9d0e15126b61b484ee2" dependencies = [ "proc-macro2", "quote", @@ -2148,11 +1993,11 @@ dependencies = [ [[package]] name = "tracing-core" -version = "0.1.23" +version = "0.1.29" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "aa31669fa42c09c34d94d8165dd2012e8ff3c66aca50f3bb226b68f216f2706c" +checksum = "5aeea4303076558a00714b823f9ad67d58a3bbda1df83d8827d21193156e22f7" dependencies = [ - "lazy_static", + "once_cell", ] [[package]] @@ -2163,15 +2008,15 @@ checksum = "dcf81ac59edc17cc8697ff311e8f5ef2d99fcbd9817b34cec66f90b6c3dfd987" [[package]] name = "ucd-trie" -version = "0.1.3" +version = "0.1.5" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "56dee185309b50d1f11bfedef0fe6d036842e3fb77413abef29f8f8d1c5d4c1c" +checksum = "9e79c4d996edb816c91e4308506774452e55e95c3c9de07b6729e17e15a5ef81" [[package]] name = "uhlc" -version = "0.4.1" +version = "0.5.1" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "9d74cc14aac0f650dae365e42250073bc0f89602bad5751d6159642be37690d6" +checksum = "7908438f98a5824af02b34c2b31fb369c5764ef835d26df0badbb9897fb28245" dependencies = [ "hex", "humantime", @@ -2182,22 +2027,22 @@ dependencies = [ ] [[package]] -name = "unicode-width" -version = "0.1.9" +name = "unicode-ident" +version = "1.0.4" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "3ed742d4ea2bd1176e236172c8429aaf54486e7ac098db29ffe6529e0ce50973" +checksum = "dcc811dc4066ac62f84f11307873c4850cb653bfa9b1719cee2bd2204a4bc5dd" [[package]] -name = "unicode-xid" -version = "0.2.2" +name = "unindent" +version = "0.1.10" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "8ccb82d61f80a663efe1f787a51b16b5a51e3314d6ac365b08639f52387b33f3" +checksum = "58ee9362deb4a96cef4d437d1ad49cffc9b9e92d202b6995674e928ce684f112" [[package]] -name = "unindent" -version = "0.1.8" +name = "unsafe-libyaml" +version = "0.2.4" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "514672a55d7380da379785a4d70ca8386c8883ff7eaae877be4d2081cebe73d8" +checksum = "c1e5fa573d8ac5f1a856f8d7be41d390ee973daf97c806b2c1a465e4e1406e68" [[package]] name = "untrusted" @@ -2218,23 +2063,13 @@ dependencies = [ [[package]] name = "uuid" -version = "0.8.2" +version = "1.1.2" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "bc5cf98d8186244414c848017f0e2676b3fcb46807f6668a97dfe67359a3c4b7" +checksum = "dd6469f4314d5f1ffec476e05f17cc9a78bc7a27a6a857842170bdf8d6f98d2f" dependencies = [ "getrandom", ] -[[package]] -name = "validated_struct" -version = "0.1.11" -source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "85affc53ebb88700a36aefff74b6a4db38dbccf26ec777d40c9e9b471be261f8" -dependencies = [ - "serde", - "validated_struct_macros 0.1.10", -] - [[package]] name = "validated_struct" version = "2.1.0" @@ -2244,19 +2079,7 @@ dependencies = [ "json5", "serde", "serde_json", - "validated_struct_macros 2.1.0", -] - -[[package]] -name = "validated_struct_macros" -version = "0.1.10" -source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "ca689f3addec28c7d67ed3139498d0325253c71bf58ef57c2ba76141f7ce576f" -dependencies = [ - "proc-macro2", - "quote", - "syn", - "unzip-n", + "validated_struct_macros", ] [[package]] @@ -2273,9 +2096,9 @@ dependencies = [ [[package]] name = "value-bag" -version = "1.0.0-alpha.8" +version = "1.0.0-alpha.9" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "79923f7731dc61ebfba3633098bf3ac533bbd35ccd8c57e7088d9a5eebe0263f" +checksum = "2209b78d1249f7e6f3293657c9779fe31ced465df091bbd433a1cf88e916ec55" dependencies = [ "ctor", "version_check", @@ -2299,12 +2122,6 @@ version = "1.1.0" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "9d5b2c62b4012a3e1eca5a7e077d13b3bf498c4073e33ccd58626607748ceeca" -[[package]] -name = "wasi" -version = "0.10.2+wasi-snapshot-preview1" -source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "fd6fbd9a79829dd1ad0cc20627bf1ed606756a7f77edff7b66b7064f9cb327c6" - [[package]] name = "wasi" version = "0.11.0+wasi-snapshot-preview1" @@ -2313,9 +2130,9 @@ checksum = "9c8d87e72b64a3b4db28d11ce29237c246188f4f51057d65a7eab63b7987e423" [[package]] name = "wasm-bindgen" -version = "0.2.79" +version = "0.2.83" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "25f1af7423d8588a3d840681122e72e6a24ddbcb3f0ec385cac0d12d24256c06" +checksum = "eaf9f5aceeec8be17c128b2e93e031fb8a4d469bb9c4ae2d7dc1888b26887268" dependencies = [ "cfg-if", "wasm-bindgen-macro", @@ -2323,13 +2140,13 @@ dependencies = [ [[package]] name = "wasm-bindgen-backend" -version = "0.2.79" +version = "0.2.83" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "8b21c0df030f5a177f3cba22e9bc4322695ec43e7257d865302900290bcdedca" +checksum = "4c8ffb332579b0557b52d268b91feab8df3615f265d5270fec2a8c95b17c1142" dependencies = [ "bumpalo", - "lazy_static", "log", + "once_cell", "proc-macro2", "quote", "syn", @@ -2338,9 +2155,9 @@ dependencies = [ [[package]] name = "wasm-bindgen-futures" -version = "0.4.29" +version = "0.4.33" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "2eb6ec270a31b1d3c7e266b999739109abce8b6c87e4b31fcfcd788b65267395" +checksum = "23639446165ca5a5de86ae1d8896b737ae80319560fbaa4c2887b7da6e7ebd7d" dependencies = [ "cfg-if", "js-sys", @@ -2350,9 +2167,9 @@ dependencies = [ [[package]] name = "wasm-bindgen-macro" -version = "0.2.79" +version = "0.2.83" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "2f4203d69e40a52ee523b2529a773d5ffc1dc0071801c87b3d270b471b80ed01" +checksum = "052be0f94026e6cbc75cdefc9bae13fd6052cdcaf532fa6c45e7ae33a1e6c810" dependencies = [ "quote", "wasm-bindgen-macro-support", @@ -2360,9 +2177,9 @@ dependencies = [ [[package]] name = "wasm-bindgen-macro-support" -version = "0.2.79" +version = "0.2.83" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "bfa8a30d46208db204854cadbb5d4baf5fcf8071ba5bf48190c3e59937962ebc" +checksum = "07bc0c051dc5f23e307b13285f9d75df86bfdf816c5721e573dec1f9b8aa193c" dependencies = [ "proc-macro2", "quote", @@ -2373,15 +2190,15 @@ dependencies = [ [[package]] name = "wasm-bindgen-shared" -version = "0.2.79" +version = "0.2.83" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "3d958d035c4438e28c70e4321a2911302f10135ce78a9c7834c0cab4123d06a2" +checksum = "1c38c045535d93ec4f0b4defec448e4291638ee608530863b1e2ba115d4fff7f" [[package]] name = "web-sys" -version = "0.3.56" +version = "0.3.60" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "c060b319f29dd25724f09a2ba1418f142f539b2be99fbf4d2d5a8f7330afb8eb" +checksum = "bcda906d8be16e728fd5adc5b729afad4e444e106ab28cd1c7256e54fa61510f" dependencies = [ "js-sys", "wasm-bindgen", @@ -2416,6 +2233,15 @@ dependencies = [ "cc", ] +[[package]] +name = "win-sys" +version = "0.3.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "5b7b128a98c1cfa201b09eb49ba285887deb3cbe7466a98850eb1adabb452be5" +dependencies = [ + "windows", +] + [[package]] name = "winapi" version = "0.3.9" @@ -2448,18 +2274,95 @@ source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "712e227841d057c1ee1cd2fb22fa7e5a5461ae8e48fa2ca79ec42cfc1931183f" [[package]] -name = "yaml-rust" -version = "0.4.5" +name = "windows" +version = "0.34.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "56c1936c4cc7a1c9ab21a1ebb602eb942ba868cbd44a99cb7cdc5892335e1c85" +checksum = "45296b64204227616fdbf2614cefa4c236b98ee64dfaaaa435207ed99fe7829f" dependencies = [ - "linked-hash-map", + "windows_aarch64_msvc 0.34.0", + "windows_i686_gnu 0.34.0", + "windows_i686_msvc 0.34.0", + "windows_x86_64_gnu 0.34.0", + "windows_x86_64_msvc 0.34.0", ] +[[package]] +name = "windows-sys" +version = "0.36.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "ea04155a16a59f9eab786fe12a4a450e75cdb175f9e0d80da1e17db09f55b8d2" +dependencies = [ + "windows_aarch64_msvc 0.36.1", + "windows_i686_gnu 0.36.1", + "windows_i686_msvc 0.36.1", + "windows_x86_64_gnu 0.36.1", + "windows_x86_64_msvc 0.36.1", +] + +[[package]] +name = "windows_aarch64_msvc" +version = "0.34.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "17cffbe740121affb56fad0fc0e421804adf0ae00891205213b5cecd30db881d" + +[[package]] +name = "windows_aarch64_msvc" +version = "0.36.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "9bb8c3fd39ade2d67e9874ac4f3db21f0d710bee00fe7cab16949ec184eeaa47" + +[[package]] +name = "windows_i686_gnu" +version = "0.34.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "2564fde759adb79129d9b4f54be42b32c89970c18ebf93124ca8870a498688ed" + +[[package]] +name = "windows_i686_gnu" +version = "0.36.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "180e6ccf01daf4c426b846dfc66db1fc518f074baa793aa7d9b9aaeffad6a3b6" + +[[package]] +name = "windows_i686_msvc" +version = "0.34.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "9cd9d32ba70453522332c14d38814bceeb747d80b3958676007acadd7e166956" + +[[package]] +name = "windows_i686_msvc" +version = "0.36.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "e2e7917148b2812d1eeafaeb22a97e4813dfa60a3f8f78ebe204bcc88f12f024" + +[[package]] +name = "windows_x86_64_gnu" +version = "0.34.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "cfce6deae227ee8d356d19effc141a509cc503dfd1f850622ec4b0f84428e1f4" + +[[package]] +name = "windows_x86_64_gnu" +version = "0.36.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "4dcd171b8776c41b97521e5da127a2d86ad280114807d0b2ab1e462bc764d9e1" + +[[package]] +name = "windows_x86_64_msvc" +version = "0.34.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "d19538ccc21819d01deaf88d6a17eae6596a12e9aafdbb97916fb49896d89de9" + +[[package]] +name = "windows_x86_64_msvc" +version = "0.36.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "c811ca4a8c853ef420abd8592ba53ddbbac90410fab6903b3e79972a631f7680" + [[package]] name = "zenoh" version = "0.6.0-dev.0" -source = "git+https://github.com/eclipse-zenoh/zenoh#a3fecd9909ff0d7bccedbf35fc0f1c7e57c0d34b" +source = "git+https://github.com/eclipse-zenoh/zenoh?branch=master#4d8f680f746c57f16cf9cc15e3b44bf32d3de83e" dependencies = [ "async-global-executor", "async-std", @@ -2468,8 +2371,8 @@ dependencies = [ "env_logger", "event-listener", "flume", + "form_urlencoded", "futures", - "futures-lite", "git-version", "hex", "lazy_static", @@ -2504,7 +2407,7 @@ dependencies = [ [[package]] name = "zenoh-buffers" version = "0.6.0-dev.0" -source = "git+https://github.com/eclipse-zenoh/zenoh#a3fecd9909ff0d7bccedbf35fc0f1c7e57c0d34b" +source = "git+https://github.com/eclipse-zenoh/zenoh?branch=master#4d8f680f746c57f16cf9cc15e3b44bf32d3de83e" dependencies = [ "async-std", "bincode", @@ -2519,7 +2422,7 @@ dependencies = [ [[package]] name = "zenoh-cfg-properties" version = "0.6.0-dev.0" -source = "git+https://github.com/eclipse-zenoh/zenoh#a3fecd9909ff0d7bccedbf35fc0f1c7e57c0d34b" +source = "git+https://github.com/eclipse-zenoh/zenoh?branch=master#4d8f680f746c57f16cf9cc15e3b44bf32d3de83e" dependencies = [ "zenoh-core", "zenoh-macros", @@ -2528,7 +2431,7 @@ dependencies = [ [[package]] name = "zenoh-collections" version = "0.6.0-dev.0" -source = "git+https://github.com/eclipse-zenoh/zenoh#a3fecd9909ff0d7bccedbf35fc0f1c7e57c0d34b" +source = "git+https://github.com/eclipse-zenoh/zenoh?branch=master#4d8f680f746c57f16cf9cc15e3b44bf32d3de83e" dependencies = [ "async-std", "async-trait", @@ -2541,14 +2444,15 @@ dependencies = [ [[package]] name = "zenoh-config" version = "0.6.0-dev.0" -source = "git+https://github.com/eclipse-zenoh/zenoh#a3fecd9909ff0d7bccedbf35fc0f1c7e57c0d34b" +source = "git+https://github.com/eclipse-zenoh/zenoh?branch=master#4d8f680f746c57f16cf9cc15e3b44bf32d3de83e" dependencies = [ "flume", "json5", + "num_cpus", "serde", "serde_json", "serde_yaml", - "validated_struct 2.1.0", + "validated_struct", "zenoh-cfg-properties", "zenoh-core", "zenoh-protocol-core", @@ -2558,16 +2462,17 @@ dependencies = [ [[package]] name = "zenoh-core" version = "0.6.0-dev.0" -source = "git+https://github.com/eclipse-zenoh/zenoh#a3fecd9909ff0d7bccedbf35fc0f1c7e57c0d34b" +source = "git+https://github.com/eclipse-zenoh/zenoh?branch=master#4d8f680f746c57f16cf9cc15e3b44bf32d3de83e" dependencies = [ "anyhow", "lazy_static", + "zenoh-macros", ] [[package]] name = "zenoh-crypto" version = "0.6.0-dev.0" -source = "git+https://github.com/eclipse-zenoh/zenoh#a3fecd9909ff0d7bccedbf35fc0f1c7e57c0d34b" +source = "git+https://github.com/eclipse-zenoh/zenoh?branch=master#4d8f680f746c57f16cf9cc15e3b44bf32d3de83e" dependencies = [ "aes", "hmac", @@ -2580,7 +2485,7 @@ dependencies = [ [[package]] name = "zenoh-link" version = "0.6.0-dev.0" -source = "git+https://github.com/eclipse-zenoh/zenoh#a3fecd9909ff0d7bccedbf35fc0f1c7e57c0d34b" +source = "git+https://github.com/eclipse-zenoh/zenoh?branch=master#4d8f680f746c57f16cf9cc15e3b44bf32d3de83e" dependencies = [ "async-std", "async-trait", @@ -2599,7 +2504,7 @@ dependencies = [ [[package]] name = "zenoh-link-commons" version = "0.6.0-dev.0" -source = "git+https://github.com/eclipse-zenoh/zenoh#a3fecd9909ff0d7bccedbf35fc0f1c7e57c0d34b" +source = "git+https://github.com/eclipse-zenoh/zenoh?branch=master#4d8f680f746c57f16cf9cc15e3b44bf32d3de83e" dependencies = [ "async-std", "async-trait", @@ -2614,15 +2519,16 @@ dependencies = [ [[package]] name = "zenoh-link-quic" version = "0.6.0-dev.0" -source = "git+https://github.com/eclipse-zenoh/zenoh#a3fecd9909ff0d7bccedbf35fc0f1c7e57c0d34b" +source = "git+https://github.com/eclipse-zenoh/zenoh?branch=master#4d8f680f746c57f16cf9cc15e3b44bf32d3de83e" dependencies = [ "async-std", "async-trait", + "futures", "log", "quinn", - "rustls 0.20.4", + "rustls 0.20.6", "rustls-native-certs", - "rustls-pemfile 0.3.0", + "rustls-pemfile 1.0.1", "webpki 0.22.0", "zenoh-cfg-properties", "zenoh-config", @@ -2636,7 +2542,7 @@ dependencies = [ [[package]] name = "zenoh-link-tcp" version = "0.6.0-dev.0" -source = "git+https://github.com/eclipse-zenoh/zenoh#a3fecd9909ff0d7bccedbf35fc0f1c7e57c0d34b" +source = "git+https://github.com/eclipse-zenoh/zenoh?branch=master#4d8f680f746c57f16cf9cc15e3b44bf32d3de83e" dependencies = [ "async-std", "async-trait", @@ -2651,11 +2557,12 @@ dependencies = [ [[package]] name = "zenoh-link-tls" version = "0.6.0-dev.0" -source = "git+https://github.com/eclipse-zenoh/zenoh#a3fecd9909ff0d7bccedbf35fc0f1c7e57c0d34b" +source = "git+https://github.com/eclipse-zenoh/zenoh?branch=master#4d8f680f746c57f16cf9cc15e3b44bf32d3de83e" dependencies = [ "async-rustls", "async-std", "async-trait", + "futures", "log", "zenoh-cfg-properties", "zenoh-config", @@ -2669,7 +2576,7 @@ dependencies = [ [[package]] name = "zenoh-link-udp" version = "0.6.0-dev.0" -source = "git+https://github.com/eclipse-zenoh/zenoh#a3fecd9909ff0d7bccedbf35fc0f1c7e57c0d34b" +source = "git+https://github.com/eclipse-zenoh/zenoh?branch=master#4d8f680f746c57f16cf9cc15e3b44bf32d3de83e" dependencies = [ "async-std", "async-trait", @@ -2686,12 +2593,13 @@ dependencies = [ [[package]] name = "zenoh-link-unixsock_stream" version = "0.6.0-dev.0" -source = "git+https://github.com/eclipse-zenoh/zenoh#a3fecd9909ff0d7bccedbf35fc0f1c7e57c0d34b" +source = "git+https://github.com/eclipse-zenoh/zenoh?branch=master#4d8f680f746c57f16cf9cc15e3b44bf32d3de83e" dependencies = [ "async-std", "async-trait", + "futures", "log", - "nix 0.23.1", + "nix 0.25.0", "uuid", "zenoh-core", "zenoh-link-commons", @@ -2702,7 +2610,7 @@ dependencies = [ [[package]] name = "zenoh-macros" version = "0.6.0-dev.0" -source = "git+https://github.com/eclipse-zenoh/zenoh#a3fecd9909ff0d7bccedbf35fc0f1c7e57c0d34b" +source = "git+https://github.com/eclipse-zenoh/zenoh?branch=master#4d8f680f746c57f16cf9cc15e3b44bf32d3de83e" dependencies = [ "proc-macro2", "quote", @@ -2714,7 +2622,7 @@ dependencies = [ [[package]] name = "zenoh-plugin-trait" version = "0.6.0-dev.0" -source = "git+https://github.com/eclipse-zenoh/zenoh#a3fecd9909ff0d7bccedbf35fc0f1c7e57c0d34b" +source = "git+https://github.com/eclipse-zenoh/zenoh?branch=master#4d8f680f746c57f16cf9cc15e3b44bf32d3de83e" dependencies = [ "libloading", "log", @@ -2727,7 +2635,7 @@ dependencies = [ [[package]] name = "zenoh-protocol" version = "0.6.0-dev.0" -source = "git+https://github.com/eclipse-zenoh/zenoh#a3fecd9909ff0d7bccedbf35fc0f1c7e57c0d34b" +source = "git+https://github.com/eclipse-zenoh/zenoh?branch=master#4d8f680f746c57f16cf9cc15e3b44bf32d3de83e" dependencies = [ "log", "uhlc", @@ -2739,10 +2647,12 @@ dependencies = [ [[package]] name = "zenoh-protocol-core" version = "0.6.0-dev.0" -source = "git+https://github.com/eclipse-zenoh/zenoh#a3fecd9909ff0d7bccedbf35fc0f1c7e57c0d34b" +source = "git+https://github.com/eclipse-zenoh/zenoh?branch=master#4d8f680f746c57f16cf9cc15e3b44bf32d3de83e" dependencies = [ "hex", + "itertools", "lazy_static", + "rand", "serde", "uhlc", "uuid", @@ -2752,12 +2662,12 @@ dependencies = [ [[package]] name = "zenoh-sync" version = "0.6.0-dev.0" -source = "git+https://github.com/eclipse-zenoh/zenoh#a3fecd9909ff0d7bccedbf35fc0f1c7e57c0d34b" +source = "git+https://github.com/eclipse-zenoh/zenoh?branch=master#4d8f680f746c57f16cf9cc15e3b44bf32d3de83e" dependencies = [ "async-std", "event-listener", "flume", - "futures-lite", + "futures", "tokio", "zenoh-core", ] @@ -2765,15 +2675,17 @@ dependencies = [ [[package]] name = "zenoh-transport" version = "0.6.0-dev.0" -source = "git+https://github.com/eclipse-zenoh/zenoh#a3fecd9909ff0d7bccedbf35fc0f1c7e57c0d34b" +source = "git+https://github.com/eclipse-zenoh/zenoh?branch=master#4d8f680f746c57f16cf9cc15e3b44bf32d3de83e" dependencies = [ + "async-executor", "async-global-executor", "async-std", "async-trait", "flume", "log", - "paste 1.0.6", + "paste", "rand", + "ringbuffer-spsc", "rsa", "serde", "zenoh-buffers", @@ -2791,7 +2703,7 @@ dependencies = [ [[package]] name = "zenoh-util" version = "0.6.0-dev.0" -source = "git+https://github.com/eclipse-zenoh/zenoh#a3fecd9909ff0d7bccedbf35fc0f1c7e57c0d34b" +source = "git+https://github.com/eclipse-zenoh/zenoh?branch=master#4d8f680f746c57f16cf9cc15e3b44bf32d3de83e" dependencies = [ "async-std", "clap", @@ -2804,6 +2716,7 @@ dependencies = [ "libloading", "log", "pnet", + "pnet_datalink", "shellexpand", "winapi", "zenoh-cfg-properties", @@ -2815,21 +2728,6 @@ dependencies = [ [[package]] name = "zeroize" -version = "1.4.3" +version = "1.5.7" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "d68d9dcec5f9b43a30d38c49f91dfedfaac384cb8f085faca366c26207dd1619" -dependencies = [ - "zeroize_derive", -] - -[[package]] -name = "zeroize_derive" -version = "1.3.2" -source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "3f8f187641dad4f680d25c4bfc4225b418165984179f26ca76ec4fb6441d3a17" -dependencies = [ - "proc-macro2", - "quote", - "syn", - "synstructure", -] +checksum = "c394b5bd0c6f669e7275d9c20aa90ae064cb22e75a1cad54e1b34088034b149f" diff --git a/Cargo.toml b/Cargo.toml index 9737bfc1..007f47c3 100644 --- a/Cargo.toml +++ b/Cargo.toml @@ -34,17 +34,17 @@ crate-type = ["cdylib"] complete_n = ["zenoh/complete_n"] [dependencies] -zenoh = { git = "https://github.com/eclipse-zenoh/zenoh" } -zenoh-cfg-properties = { git = "https://github.com/eclipse-zenoh/zenoh" } -zenoh-core = { git = "https://github.com/eclipse-zenoh/zenoh" } -zenoh-buffers = { git = "https://github.com/eclipse-zenoh/zenoh" } -validated_struct = "0.1" +async-std = "=1.12.0" +env_logger = "0.9.1" +form_urlencoded = "1.1.0" +futures = "0.3.24" json5 = "0.4.1" -serde_json = "1.0" -async-std = "=1.11.0" -uhlc = "0.4.0" -futures = "0.3.12" -log = "0.4" -env_logger = "0.9.0" -pyo3 = { version = "0.16", features = ["extension-module", "abi3-py37"] } -pyo3-asyncio = { version = "0.16", features = ["attributes", "async-std-runtime"] } +log = "0.4.17" +pyo3 = { version = "0.17.1", features = ["extension-module", "abi3-py37"] } +serde_json = "1.0.85" +uhlc = "0.5.1" +validated_struct = "2.1.0" +zenoh = { git = "https://github.com/eclipse-zenoh/zenoh", branch = "master" } +zenoh-buffers = { git = "https://github.com/eclipse-zenoh/zenoh", branch = "master" } +zenoh-cfg-properties = { git = "https://github.com/eclipse-zenoh/zenoh", branch = "master" } +zenoh-core = { git = "https://github.com/eclipse-zenoh/zenoh", branch = "master" } diff --git a/Jenkinsfile b/Jenkinsfile deleted file mode 100644 index 9322d9e1..00000000 --- a/Jenkinsfile +++ /dev/null @@ -1,119 +0,0 @@ -pipeline { - agent { label 'MacMini' } - options { skipDefaultCheckout() } - parameters { - gitParameter(name: 'GIT_TAG', - type: 'PT_BRANCH_TAG', - description: 'The Git tag to checkout. If not specified "master" will be checkout.', - defaultValue: 'master') - booleanParam(name: 'PUBLISH_RESULTS', - description: 'Publish the resulting wheels to Eclipse download and to pypi.org (if not a branch)', - defaultValue: false) - booleanParam(name: 'BUILD_MACOSX', - description: 'Build wheels for macosx_10_7_x86_64.', - defaultValue: true) - booleanParam(name: 'BUILD_LINUX64', - description: 'Build wheels with manylinux2010-x86-64.', - defaultValue: true) - booleanParam(name: 'BUILD_LINUX32', - description: 'Build wheels with manylinux2010-i686.', - defaultValue: true) - booleanParam(name: 'BUILD_AARCH64', - description: 'Build wheels with manylinux2014-aarch64.', - defaultValue: true) - } - environment { - LABEL = get_label() - DOWNLOAD_DIR="/home/data/httpd/download.eclipse.org/zenoh/zenoh-python/${LABEL}" - } - - stages { - stage('Checkout Git TAG') { - steps { - deleteDir() - checkout([$class: 'GitSCM', - branches: [[name: "${params.GIT_TAG}"]], - doGenerateSubmoduleConfigurations: false, - extensions: [], - gitTool: 'Default', - submoduleCfg: [], - userRemoteConfigs: [[url: 'https://github.com/eclipse-zenoh/zenoh-python.git']] - ]) - } - } - - stage('MacOS wheels') { - when { expression { return params.BUILD_MACOSX }} - steps { - sh ''' - set +x - . ~/.zshenv - env - maturin build --release $MATURIN_PYTHONS_OPT - ''' - } - } - - stage('Manylinux2010-x64 wheels') { - when { expression { return params.BUILD_LINUX64 }} - steps { - sh ''' - docker run --init --rm -v $(pwd):/workdir -w /workdir adlinktech/zenoh-dev-manylinux2010-x86_64-gnu maturin build --release --manylinux 2010 - ''' - } - } - - stage('Manylinux2010-i686 wheels') { - when { expression { return params.BUILD_LINUX32 }} - steps { - sh ''' - docker run --init --rm -v $(pwd):/workdir -w /workdir adlinktech/zenoh-dev-manylinux2010-i686-gnu maturin build --release --manylinux 2010 - ''' - } - } - - stage('Manylinux2014-aarch64 wheels') { - when { expression { return params.BUILD_AARCH64 }} - steps { - sh ''' - docker run --init --rm -v $(pwd):/workdir -w /workdir adlinktech/zenoh-dev-manylinux2014-aarch64-gnu maturin build --release --manylinux 2014 - ''' - } - } - - stage('Deploy to download.eclipse.org') { - when { expression { return params.PUBLISH_RESULTS }} - steps { - sshagent ( ['projects-storage.eclipse.org-bot-ssh']) { - sh ''' - if [[ ${GIT_TAG} == origin/ ]]; then - ssh [email protected] rm -fr ${DOWNLOAD_DIR} - fi - ssh [email protected] mkdir -p ${DOWNLOAD_DIR} - scp target/wheels/.tar.gz [email protected]:${DOWNLOAD_DIR}/ - find target -name ".whl" \| xargs -J FILES scp FILES [email protected]:${DOWNLOAD_DIR}/ - ''' - } - } - } - - stage('Deploy on pypi.org') { - when { expression { return params.PUBLISH_RESULTS && !env.GIT_TAG.startsWith('origin/') }} - steps { - sh ''' - python3 -m twine upload --repository eclipse-zenoh `find target -name ".whl"` target/wheels/.tar.gz - ''' - } - } - } - - post { - success { - archiveArtifacts artifacts: 'target/*/.whl, target/wheels/.tar.gz', fingerprint: true - } - } -} - -def get_label() { - return env.GIT_TAG.startsWith('origin/') ? env.GIT_TAG.minus('origin/') : env.GIT_TAG -} diff --git a/README.md b/README.md index f163413f..eb5c3b14 100644 --- a/README.md +++ b/README.md @@ -48,6 +48,7 @@ It relies on the [zenoh](https://github.com/eclipse-zenoh/zenoh/tree/master/zeno Requirements: Python >= 3.7 * pip >= 19.3.1 + * (Optional) A Python virtual environment (for instance [virtualenv](docs.python.org/3.10/tutorial/venv.html) or [miniconda](https://docs.conda.io/en/latest/miniconda.html)) * [Rust and Cargo](https://doc.rust-lang.org/cargo/getting-started/installation.html). Steps: @@ -56,19 +57,24 @@ Steps: pip install -r requirements-dev.txt ``` - * Ensure your system can find the building tool maturin. - For example, it is placed at _$HOME/.local/bin/maturin_ by default on Ubuntu 20.04. - + * Ensure your system can find the building tool `maturin` (installed by previous step). + For example, it is placed at _$HOME/.local/bin/maturin_ by default on Ubuntu 20.04. ```bash export PATH="$HOME/.local/bin:$PATH" ``` - * Build zenoh-python - ```bash - maturin build --release - ``` + * Build and install zenoh-python: + * With a virtual environment active: + ```bash + maturin develop --release + ``` + * Without one: + ```bash + maturin build --release + pip install ./target/wheels/<there should only be one .whl file here> + ``` + -This will automatically build the zenoh Rust API, as well as the zenoh-python API and install it in your Python environement. ------------------------------- ## Running the Examples diff --git a/docs/conf.py b/docs/conf.py index 9589cd83..e50d3c81 100644 --- a/docs/conf.py +++ b/docs/conf.py @@ -72,7 +72,7 @@ # List of patterns, relative to source directory, that match files and # directories to ignore when looking for source files. # This pattern also affects html_static_path and html_extra_path. -exclude_patterns = ['_build', 'Thumbs.db', '.DS_Store'] +exclude_patterns = ['_build', 'Thumbs.db', '.DS_Store', '*/.rs'] # The name of the Pygments (syntax highlighting) style to use. pygments_style = None diff --git a/docs/index.rst b/docs/index.rst index f9b32f30..6e4b189b 100644 --- a/docs/index.rst +++ b/docs/index.rst @@ -19,23 +19,7 @@ Zenoh API Reference module zenoh ============ .. automodule:: zenoh - :members: init_logger, config_from_file, open, async_open, scout, async_scout - - -AsyncQueryable --------------- -.. autoclass:: zenoh.AsyncQueryable - :members: - -AsyncSession ------------- -.. autoclass:: zenoh.AsyncSession - :members: - -AsyncSubscriber ---------------- -.. autoclass:: zenoh.AsyncSubscriber - :members: + :members: init_logger, open, scout Config ------ @@ -48,11 +32,11 @@ CongestionControl :members: :undoc-members: -ConsolidationMode ------------------ -.. autoclass:: zenoh.ConsolidationMode - :members: - :undoc-members: +.. ConsolidationMode +.. ----------------- +.. .. autoclass:: zenoh.ConsolidationMode +.. :members: +.. :undoc-members: Encoding -------- @@ -70,22 +54,10 @@ KeyExpr .. autoclass:: zenoh.KeyExpr :members: -KnownEncoding -------------- -.. autoclass:: zenoh.KnownEncoding - :members: - :undoc-members: - -PeerId ------- -.. autoclass:: zenoh.PeerId - :members: - -Period ------- -.. autoclass:: zenoh.Period +ZenohId +------- +.. autoclass:: zenoh.ZenohId :members: - :special-members: __init__ Priority -------- @@ -145,27 +117,20 @@ Session .. autoclass:: zenoh.Session :members: -SourceInfo +Subscriber ---------- -.. autoclass:: zenoh.SourceInfo +.. autoclass:: zenoh.Subscriber :members: -SubMode -------- -.. autoclass:: zenoh.SubMode +PullSubscriber +-------------- +.. autoclass:: zenoh.PullSubscriber :members: - :undoc-members: -Subscriber +Publisher ---------- -.. autoclass:: zenoh.Subscriber - :members: - -Target ------- -.. autoclass:: zenoh.Target +.. autoclass:: zenoh.Publisher :members: - :undoc-members: Timestamp --------- @@ -177,41 +142,7 @@ Value .. autoclass:: zenoh.Value :members: -ValueSelector -------------- -.. autoclass:: zenoh.ValueSelector - :members: - -WhatAmI -------- -.. autoclass:: zenoh.WhatAmI - :members: - :undoc-members: - -ZError ------- -.. autoexception:: zenoh.ZError - :members: - - -********* -submodules -******* - -module zenoh.config -=================== -.. autoclass:: zenoh.config - :members: - :undoc-members: - -module zenoh.info -================= -.. autoclass:: zenoh.info - :members: - :undoc-members: - -module zenoh.queryable -====================== -.. autoclass:: zenoh.queryable +Info +----- +.. autoclass:: zenoh.Info :members: - :undoc-members: diff --git a/examples/README.md b/examples/README.md index d93b6876..7a5a2481 100644 --- a/examples/README.md +++ b/examples/README.md @@ -97,7 +97,7 @@ ### z_get Sends a query message for a selector. - The queryables with a matching path or selector (for instance [z_eval](#z_eval) and [z_storage](#z_storage)) + The queryables with a matching path or selector (for instance [z_queryable](#z_queryable) and [z_storage](#z_storage)) will receive this query and reply with paths/values that will be received by the query callback. Typical usage: @@ -109,7 +109,7 @@ python3 z_get.py -s '/demo/' ``` -### z_eval +### z_queryable Creates a queryable function with a key expression. This queryable function will be triggered by each call to a get operation on zenoh @@ -117,11 +117,11 @@ Typical usage: ```bash - python3 z_eval.py + python3 z_queryable.py ``` or ```bash - python3 z_eval.py -k /demo/example/eval -v 'This is the result' + python3 z_queryable.py -k /demo/example/queryable -v 'This is the result' ``` ### z_storage @@ -161,4 +161,4 @@ In `asyncio` directory there are similar examples than thse described above, but leveraging Python's [asyncio](https://docs.python.org/fr/3/library/asyncio.html). -Especially, the `asyncio/z_get_parallel.py` and `asyncio/z_eval.py` examples show how a zenoh application can issue concurent queries and how another zenoh application can concurrently compute replies to those queries. +Especially, the `asyncio/z_get_parallel.py` and `asyncio/z_queryable.py` examples show how a zenoh application can issue concurent queries and how another zenoh application can concurrently compute replies to those queries. \ No newline at end of file diff --git a/examples/asyncio/z_delete.py b/examples/asyncio/z_delete.py deleted file mode 100644 index 46ca5cd5..00000000 --- a/examples/asyncio/z_delete.py +++ /dev/null @@ -1,76 +0,0 @@ -# -# Copyright (c) 2022 ZettaScale Technology -# -# This program and the accompanying materials are made available under the -# terms of the Eclipse Public License 2.0 which is available at -# http://www.eclipse.org/legal/epl-2.0, or the Apache License, Version 2.0 -# which is available at https://www.apache.org/licenses/LICENSE-2.0. -# -# SPDX-License-Identifier: EPL-2.0 OR Apache-2.0 -# -# Contributors: -# ZettaScale Zenoh Team, <[email protected]> -# - -import asyncio -import sys -import time -import argparse -import json -import zenoh -from zenoh import config - - -async def main(): - # --- Command line argument parsing --- --- --- --- --- --- - parser = argparse.ArgumentParser( - prog='z_put', - description='zenoh put example') - parser.add_argument('--mode', '-m', dest='mode', - choices=['peer', 'client'], - type=str, - help='The zenoh session mode.') - parser.add_argument('--connect', '-e', dest='connect', - metavar='ENDPOINT', - action='append', - type=str, - help='Endpoints to connect to.') - parser.add_argument('--listen', '-l', dest='listen', - metavar='ENDPOINT', - action='append', - type=str, - help='Endpoints to listen on.') - parser.add_argument('--key', '-k', dest='key', - default='/demo/example/zenoh-python-put', - type=str, - help='The key expression matching resources to delete.') - parser.add_argument('--config', '-c', dest='config', - metavar='FILE', - type=str, - help='A configuration file.') - - args = parser.parse_args() - conf = zenoh.config_from_file( - args.config) if args.config is not None else zenoh.Config() - if args.mode is not None: - conf.insert_json5(zenoh.config.MODE_KEY, json.dumps(args.mode)) - if args.connect is not None: - conf.insert_json5(zenoh.config.CONNECT_KEY, json.dumps(args.connect)) - if args.listen is not None: - conf.insert_json5(zenoh.config.LISTEN_KEY, json.dumps(args.listen)) - key = args.key - - # zenoh-net code --- --- --- --- --- --- --- --- --- --- --- - - # initiate logging - zenoh.init_logger() - - print("Openning session...") - session = await zenoh.async_open(conf) - - print("Deleting resources matching '{}'...".format(key)) - await session.delete(key) - - await session.close() - -asyncio.run(main()) diff --git a/examples/asyncio/z_eval.py b/examples/asyncio/z_eval.py deleted file mode 100644 index 7d7b8d3d..00000000 --- a/examples/asyncio/z_eval.py +++ /dev/null @@ -1,112 +0,0 @@ -# -# Copyright (c) 2022 ZettaScale Technology -# -# This program and the accompanying materials are made available under the -# terms of the Eclipse Public License 2.0 which is available at -# http://www.eclipse.org/legal/epl-2.0, or the Apache License, Version 2.0 -# which is available at https://www.apache.org/licenses/LICENSE-2.0. -# -# SPDX-License-Identifier: EPL-2.0 OR Apache-2.0 -# -# Contributors: -# ZettaScale Zenoh Team, <[email protected]> -# - -import asyncio -import sys -import time -import argparse -import json -import zenoh -from zenoh import config, Sample -from zenoh.queryable import EVAL - - -async def main(): - # --- Command line argument parsing --- --- --- --- --- --- - parser = argparse.ArgumentParser( - prog='z_eval', - description='zenoh eval example') - parser.add_argument('--mode', '-m', dest='mode', - choices=['peer', 'client'], - type=str, - help='The zenoh session mode.') - parser.add_argument('--connect', '-e', dest='connect', - metavar='ENDPOINT', - action='append', - type=str, - help='Endpoints to connect to.') - parser.add_argument('--listen', '-l', dest='listen', - metavar='ENDPOINT', - action='append', - type=str, - help='Endpoints to listen on.') - parser.add_argument('--key', '-k', dest='key', - default='/demo/example/zenoh-python-eval', - type=str, - help='The key expression matching queries to evaluate.') - parser.add_argument('--value', '-v', dest='value', - default='Eval from Python!', - type=str, - help='The value to reply to queries.') - parser.add_argument('--config', '-c', dest='config', - metavar='FILE', - type=str, - help='A configuration file.') - - args = parser.parse_args() - conf = zenoh.config_from_file( - args.config) if args.config is not None else zenoh.Config() - if args.mode is not None: - conf.insert_json5(zenoh.config.MODE_KEY, json.dumps(args.mode)) - if args.connect is not None: - conf.insert_json5(zenoh.config.CONNECT_KEY, json.dumps(args.connect)) - if args.listen is not None: - conf.insert_json5(zenoh.config.LISTEN_KEY, json.dumps(args.listen)) - key = args.key - value = args.value - - # zenoh-net code --- --- --- --- --- --- --- --- --- --- --- - - # Note: As an example the concrete implementation of the eval callback is implemented here as a coroutine. - # It checks if the query's value_selector (the substring after '?') is a float, and if yes, sleeps for this number of seconds. - # Run example/asyncio/z_get_parallel.py example to see how 3 concurrent get() are executed in parallel in this z_eval.py - async def eval_corouting(query): - selector = query.selector - try: - sleep_time = selector.parse_value_selector().properties.get('sleep') - if sleep_time is not None: - print(" Sleeping {} secs before replying".format( - float(sleep_time))) - await asyncio.sleep(float(sleep_time)) - except Exception as e: - print(" WARN: error in value selector: {}. Ignore it.".format(e)) - print(" Replying to query on {}".format(selector)) - reply = "{} (this is the reply to query on {})".format(value, selector) - query.reply(Sample(key_expr=key, payload=reply.encode())) - - async def eval_callback(query): - print(">> [Queryable ] Received Query '{}'".format(query.selector)) - # schedule a task that will call eval_corouting(query) - asyncio.create_task(eval_corouting(query)) - - # initiate logging - zenoh.init_logger() - - print("Openning session...") - session = await zenoh.async_open(conf) - - print("Creating Queryable on '{}'...".format(key)) - queryable = await session.queryable(key, eval_callback, kind=EVAL) - - print("Enter 'q' to quit......") - c = '\0' - while c != 'q': - c = sys.stdin.read(1) - if c == '': - time.sleep(1) - - await queryable.close() - await session.close() - -asyncio.run(main()) diff --git a/examples/asyncio/z_get.py b/examples/asyncio/z_get.py deleted file mode 100644 index 412f9b52..00000000 --- a/examples/asyncio/z_get.py +++ /dev/null @@ -1,99 +0,0 @@ -# -# Copyright (c) 2022 ZettaScale Technology -# -# This program and the accompanying materials are made available under the -# terms of the Eclipse Public License 2.0 which is available at -# http://www.eclipse.org/legal/epl-2.0, or the Apache License, Version 2.0 -# which is available at https://www.apache.org/licenses/LICENSE-2.0. -# -# SPDX-License-Identifier: EPL-2.0 OR Apache-2.0 -# -# Contributors: -# ZettaScale Zenoh Team, <[email protected]> -# - -import asyncio -import sys -import time -import argparse -import json -import zenoh -from zenoh import config, queryable, QueryTarget, Target - - -async def main(): - # --- Command line argument parsing --- --- --- --- --- --- - parser = argparse.ArgumentParser( - prog='z_get', - description='zenoh get example') - parser.add_argument('--mode', '-m', dest='mode', - choices=['peer', 'client'], - type=str, - help='The zenoh session mode.') - parser.add_argument('--connect', '-e', dest='connect', - metavar='ENDPOINT', - action='append', - type=str, - help='Endpoints to connect to.') - parser.add_argument('--listen', '-l', dest='listen', - metavar='ENDPOINT', - action='append', - type=str, - help='Endpoints to listen on.') - parser.add_argument('--selector', '-s', dest='selector', - default='/demo/example/', - type=str, - help='The selection of resources to query.') - parser.add_argument('--kind', '-k', dest='kind', - choices=['ALL_KINDS', 'STORAGE', 'EVAL'], - default='ALL_KINDS', - type=str, - help='The KIND of queryables to query.') - parser.add_argument('--target', '-t', dest='target', - choices=['ALL', 'BEST_MATCHING', - 'ALL_COMPLETE', 'NONE'], - default='ALL', - type=str, - help='The target queryables of the query.') - parser.add_argument('--config', '-c', dest='config', - metavar='FILE', - type=str, - help='A configuration file.') - - args = parser.parse_args() - conf = zenoh.config_from_file( - args.config) if args.config is not None else zenoh.Config() - if args.mode is not None: - conf.insert_json5(zenoh.config.MODE_KEY, json.dumps(args.mode)) - if args.connect is not None: - conf.insert_json5(zenoh.config.CONNECT_KEY, json.dumps(args.connect)) - if args.listen is not None: - conf.insert_json5(zenoh.config.LISTEN_KEY, json.dumps(args.listen)) - selector = args.selector - kind = { - 'ALL_KINDS': queryable.ALL_KINDS, - 'STORAGE': queryable.STORAGE, - 'EVAL': queryable.EVAL}.get(args.kind) - target = { - 'ALL': Target.All(), - 'BEST_MATCHING': Target.BestMatching(), - 'ALL_COMPLETE': Target.AllComplete(), - 'NONE': Target.No()}.get(args.target) - - # zenoh-net code --- --- --- --- --- --- --- --- --- --- --- - - # initiate logging - zenoh.init_logger() - - print("Openning session...") - session = await zenoh.async_open(conf) - - print("Sending Query '{}'...".format(selector)) - replies = await session.get(selector, target=QueryTarget(kind, target)) - for reply in replies: - print(">> Received ('{}': '{}')" - .format(reply.sample.key_expr, reply.sample.payload.decode("utf-8"))) - - await session.close() - -asyncio.run(main()) diff --git a/examples/asyncio/z_get_parallel.py b/examples/asyncio/z_get_parallel.py deleted file mode 100644 index 8b0b33e9..00000000 --- a/examples/asyncio/z_get_parallel.py +++ /dev/null @@ -1,109 +0,0 @@ -# -# Copyright (c) 2022 ZettaScale Technology -# -# This program and the accompanying materials are made available under the -# terms of the Eclipse Public License 2.0 which is available at -# http://www.eclipse.org/legal/epl-2.0, or the Apache License, Version 2.0 -# which is available at https://www.apache.org/licenses/LICENSE-2.0. -# -# SPDX-License-Identifier: EPL-2.0 OR Apache-2.0 -# -# Contributors: -# ZettaScale Zenoh Team, <[email protected]> -# - -import asyncio -import sys -import time -import argparse -import json -import zenoh -from zenoh import config, queryable, QueryTarget, Target - - -async def main(): - # --- Command line argument parsing --- --- --- --- --- --- - parser = argparse.ArgumentParser( - prog='z_get_parallel', - description='zenoh parallel get example') - parser.add_argument('--mode', '-m', dest='mode', - choices=['peer', 'client'], - type=str, - help='The zenoh session mode.') - parser.add_argument('--connect', '-e', dest='connect', - metavar='ENDPOINT', - action='append', - type=str, - help='Endpoints to connect to.') - parser.add_argument('--listen', '-l', dest='listen', - metavar='ENDPOINT', - action='append', - type=str, - help='Endpoints to listen on.') - parser.add_argument('--selector', '-s', dest='selector', - default='/demo/example/', - type=str, - help='The selection of resources to query.') - parser.add_argument('--kind', '-k', dest='kind', - choices=['ALL_KINDS', 'STORAGE', 'EVAL'], - default='ALL_KINDS', - type=str, - help='The KIND of queryables to query.') - parser.add_argument('--target', '-t', dest='target', - choices=['ALL', 'BEST_MATCHING', - 'ALL_COMPLETE', 'NONE'], - default='ALL', - type=str, - help='The target queryables of the query.') - parser.add_argument('--config', '-c', dest='config', - metavar='FILE', - type=str, - help='A configuration file.') - - args = parser.parse_args() - conf = zenoh.config_from_file( - args.config) if args.config is not None else zenoh.Config() - if args.mode is not None: - conf.insert_json5(zenoh.config.MODE_KEY, json.dumps(args.mode)) - if args.connect is not None: - conf.insert_json5(zenoh.config.CONNECT_KEY, json.dumps(args.connect)) - if args.listen is not None: - conf.insert_json5(zenoh.config.LISTEN_KEY, json.dumps(args.listen)) - selector = args.selector - kind = { - 'ALL_KINDS': queryable.ALL_KINDS, - 'STORAGE': queryable.STORAGE, - 'EVAL': queryable.EVAL}.get(args.kind) - target = { - 'ALL': Target.All(), - 'BEST_MATCHING': Target.BestMatching(), - 'ALL_COMPLETE': Target.AllComplete(), - 'NONE': Target.No()}.get(args.target) - - # zenoh-net code --- --- --- --- --- --- --- --- --- --- --- - - # initiate logging - zenoh.init_logger() - - print("Openning session...") - session = await zenoh.async_open(conf) - - async def do_query(sleep_time): - print("Sending Query '{}?(sleep={})'...".format(selector, sleep_time)) - replies = await session.get("{}?(sleep={})".format(selector, sleep_time), target=QueryTarget(kind, target)) - for reply in replies: - print(">> Received ('{}': '{}')" - .format(reply.sample.key_expr, reply.sample.payload.decode("utf-8"))) - - start = time.time() - await asyncio.gather( - asyncio.create_task(do_query(1)), - asyncio.create_task(do_query(2)), - asyncio.create_task(do_query(3)), - ) - end = time.time() - print(f'Time: {end-start:.2f} sec') - - await session.close() - -asyncio.run(main()) diff --git a/examples/asyncio/z_info.py b/examples/asyncio/z_info.py deleted file mode 100644 index c828e7a4..00000000 --- a/examples/asyncio/z_info.py +++ /dev/null @@ -1,71 +0,0 @@ -# -# Copyright (c) 2022 ZettaScale Technology -# -# This program and the accompanying materials are made available under the -# terms of the Eclipse Public License 2.0 which is available at -# http://www.eclipse.org/legal/epl-2.0, or the Apache License, Version 2.0 -# which is available at https://www.apache.org/licenses/LICENSE-2.0. -# -# SPDX-License-Identifier: EPL-2.0 OR Apache-2.0 -# -# Contributors: -# ZettaScale Zenoh Team, <[email protected]> -# - -import asyncio -import sys -import time -import argparse -import json -import zenoh -from zenoh import config - - -async def main(): - # --- Command line argument parsing --- --- --- --- --- --- - parser = argparse.ArgumentParser( - prog='z_info', - description='zenoh info example') - parser.add_argument('--mode', '-m', dest='mode', - choices=['peer', 'client'], - type=str, - help='The zenoh session mode.') - parser.add_argument('--connect', '-e', dest='connect', - metavar='ENDPOINT', - action='append', - type=str, - help='Endpoints to connect to.') - parser.add_argument('--listen', '-l', dest='listen', - metavar='ENDPOINT', - action='append', - type=str, - help='Endpoints to listen on.') - parser.add_argument('--config', '-c', dest='config', - metavar='FILE', - type=str, - help='A configuration file.') - - args = parser.parse_args() - conf = zenoh.config_from_file( - args.config) if args.config is not None else zenoh.Config() - if args.mode is not None: - conf.insert_json5(zenoh.config.MODE_KEY, json.dumps(args.mode)) - if args.connect is not None: - conf.insert_json5(zenoh.config.CONNECT_KEY, json.dumps(args.connect)) - if args.listen is not None: - conf.insert_json5(zenoh.config.LISTEN_KEY, json.dumps(args.listen)) - # zenoh-net code --- --- --- --- --- --- --- --- --- --- --- - - # initiate logging - zenoh.init_logger() - - print("Openning session...") - session = await zenoh.async_open(conf) - - info = await session.info() - for key in info: - print("{} : {}".format(key, info[key])) - - await session.close() - -asyncio.run(main()) diff --git a/examples/asyncio/z_pub.py b/examples/asyncio/z_pub.py deleted file mode 100644 index 071d1738..00000000 --- a/examples/asyncio/z_pub.py +++ /dev/null @@ -1,96 +0,0 @@ -# -# Copyright (c) 2022 ZettaScale Technology -# -# This program and the accompanying materials are made available under the -# terms of the Eclipse Public License 2.0 which is available at -# http://www.eclipse.org/legal/epl-2.0, or the Apache License, Version 2.0 -# which is available at https://www.apache.org/licenses/LICENSE-2.0. -# -# SPDX-License-Identifier: EPL-2.0 OR Apache-2.0 -# -# Contributors: -# ZettaScale Zenoh Team, <[email protected]> -# - -import asyncio -import sys -import time -import argparse -import itertools -import json -import zenoh -from zenoh import config - - -async def main(): - # --- Command line argument parsing --- --- --- --- --- --- - parser = argparse.ArgumentParser( - prog='z_pub', - description='zenoh pub example') - parser.add_argument('--mode', '-m', dest='mode', - choices=['peer', 'client'], - type=str, - help='The zenoh session mode.') - parser.add_argument('--connect', '-e', dest='connect', - metavar='ENDPOINT', - action='append', - type=str, - help='Endpoints to connect to.') - parser.add_argument('--listen', '-l', dest='listen', - metavar='ENDPOINT', - action='append', - type=str, - help='Endpoints to listen on.') - parser.add_argument('--key', '-k', dest='key', - default='/demo/example/zenoh-python-pub', - type=str, - help='The key expression to publish onto.') - parser.add_argument('--value', '-v', dest='value', - default='Pub from Python!', - type=str, - help='The value to publish.') - parser.add_argument("--iter", dest="iter", type=int, - help="How many puts to perform") - parser.add_argument('--config', '-c', dest='config', - metavar='FILE', - type=str, - help='A configuration file.') - - args = parser.parse_args() - conf = zenoh.config_from_file( - args.config) if args.config is not None else zenoh.Config() - if args.mode is not None: - conf.insert_json5(zenoh.config.MODE_KEY, json.dumps(args.mode)) - if args.connect is not None: - conf.insert_json5(zenoh.config.CONNECT_KEY, json.dumps(args.connect)) - if args.listen is not None: - conf.insert_json5(zenoh.config.LISTEN_KEY, json.dumps(args.listen)) - key = args.key - value = args.value - - # zenoh-net code --- --- --- --- --- --- --- --- --- --- --- - - # initiate logging - zenoh.init_logger() - - print("Openning session...") - session = await zenoh.async_open(conf) - - print("Declaring key expression '{}'...".format(key), end='') - rid = await session.declare_expr(key) - print(" => RId {}".format(rid)) - - print("Declaring publication on '{}'...".format(rid)) - await session.declare_publication(rid) - - for idx in itertools.count() if args.iter is None else range(args.iter): - time.sleep(1) - buf = "[{:4d}] {}".format(idx, value) - print("Putting Data ('{}': '{}')...".format(rid, buf)) - await session.put(rid, bytes(buf, encoding='utf8')) - - await session.undeclare_publication(rid) - await session.undeclare_expr(rid) - await session.close() - -asyncio.run(main()) diff --git a/examples/asyncio/z_put.py b/examples/asyncio/z_put.py deleted file mode 100644 index d31e8f0d..00000000 --- a/examples/asyncio/z_put.py +++ /dev/null @@ -1,81 +0,0 @@ -# -# Copyright (c) 2022 ZettaScale Technology -# -# This program and the accompanying materials are made available under the -# terms of the Eclipse Public License 2.0 which is available at -# http://www.eclipse.org/legal/epl-2.0, or the Apache License, Version 2.0 -# which is available at https://www.apache.org/licenses/LICENSE-2.0. -# -# SPDX-License-Identifier: EPL-2.0 OR Apache-2.0 -# -# Contributors: -# ZettaScale Zenoh Team, <[email protected]> -# - -import asyncio -import sys -import time -import argparse -import json -import zenoh -from zenoh import config - - -async def main(): - # --- Command line argument parsing --- --- --- --- --- --- - parser = argparse.ArgumentParser( - prog='z_put', - description='zenoh put example') - parser.add_argument('--mode', '-m', dest='mode', - choices=['peer', 'client'], - type=str, - help='The zenoh session mode.') - parser.add_argument('--connect', '-e', dest='connect', - metavar='ENDPOINT', - action='append', - type=str, - help='Endpoints to connect to.') - parser.add_argument('--listen', '-l', dest='listen', - metavar='ENDPOINT', - action='append', - type=str, - help='Endpoints to listen on.') - parser.add_argument('--key', '-k', dest='key', - default='/demo/example/zenoh-python-put', - type=str, - help='The key expression to write.') - parser.add_argument('--value', '-v', dest='value', - default='Put from Python!', - type=str, - help='The value to write.') - parser.add_argument('--config', '-c', dest='config', - metavar='FILE', - type=str, - help='A configuration file.') - - args = parser.parse_args() - conf = zenoh.config_from_file( - args.config) if args.config is not None else zenoh.Config() - if args.mode is not None: - conf.insert_json5(zenoh.config.MODE_KEY, json.dumps(args.mode)) - if args.connect is not None: - conf.insert_json5(zenoh.config.CONNECT_KEY, json.dumps(args.connect)) - if args.listen is not None: - conf.insert_json5(zenoh.config.LISTEN_KEY, json.dumps(args.listen)) - key = args.key - value = args.value - - # zenoh-net code --- --- --- --- --- --- --- --- --- --- --- - - # initiate logging - zenoh.init_logger() - - print("Openning session...") - session = await zenoh.async_open(conf) - - print("Putting Data ('{}': '{}')...".format(key, value)) - await session.put(key, value) - - await session.close() - -asyncio.run(main()) diff --git a/examples/asyncio/z_scout.py b/examples/asyncio/z_scout.py deleted file mode 100644 index 4def468b..00000000 --- a/examples/asyncio/z_scout.py +++ /dev/null @@ -1,33 +0,0 @@ -# -# Copyright (c) 2022 ZettaScale Technology -# -# This program and the accompanying materials are made available under the -# terms of the Eclipse Public License 2.0 which is available at -# http://www.eclipse.org/legal/epl-2.0, or the Apache License, Version 2.0 -# which is available at https://www.apache.org/licenses/LICENSE-2.0. -# -# SPDX-License-Identifier: EPL-2.0 OR Apache-2.0 -# -# Contributors: -# ZettaScale Zenoh Team, <[email protected]> -# - -import asyncio -import sys -import time -import argparse -import zenoh -from zenoh import WhatAmI - - -async def main(): - # initiate logging - zenoh.init_logger() - - print("Scouting...") - hellos = await zenoh.async_scout(WhatAmI.Peer \| WhatAmI.Router, 1.0) - - for hello in hellos: - print(hello) - -asyncio.run(main()) diff --git a/examples/asyncio/z_sub.py b/examples/asyncio/z_sub.py deleted file mode 100644 index 006c136c..00000000 --- a/examples/asyncio/z_sub.py +++ /dev/null @@ -1,91 +0,0 @@ -# -# Copyright (c) 2022 ZettaScale Technology -# -# This program and the accompanying materials are made available under the -# terms of the Eclipse Public License 2.0 which is available at -# http://www.eclipse.org/legal/epl-2.0, or the Apache License, Version 2.0 -# which is available at https://www.apache.org/licenses/LICENSE-2.0. -# -# SPDX-License-Identifier: EPL-2.0 OR Apache-2.0 -# -# Contributors: -# ZettaScale Zenoh Team, <[email protected]> -# - -import asyncio -import sys -import time -from datetime import datetime -import argparse -import json -import zenoh -from zenoh import Reliability, SubMode - - -async def main(): - # --- Command line argument parsing --- --- --- --- --- --- - parser = argparse.ArgumentParser( - prog='z_sub', - description='zenoh sub example') - parser.add_argument('--mode', '-m', dest='mode', - choices=['peer', 'client'], - type=str, - help='The zenoh session mode.') - parser.add_argument('--connect', '-e', dest='connect', - metavar='ENDPOINT', - action='append', - type=str, - help='Endpoints to connect to.') - parser.add_argument('--listen', '-l', dest='listen', - metavar='ENDPOINT', - action='append', - type=str, - help='Endpoints to listen on.') - parser.add_argument('--key', '-k', dest='key', - default='/demo/example/', - type=str, - help='The key expression to subscribe to.') - parser.add_argument('--config', '-c', dest='config', - metavar='FILE', - type=str, - help='A configuration file.') - - args = parser.parse_args() - conf = zenoh.config_from_file( - args.config) if args.config is not None else zenoh.Config() - if args.mode is not None: - conf.insert_json5(zenoh.config.MODE_KEY, json.dumps(args.mode)) - if args.connect is not None: - conf.insert_json5(zenoh.config.CONNECT_KEY, json.dumps(args.connect)) - if args.listen is not None: - conf.insert_json5(zenoh.config.LISTEN_KEY, json.dumps(args.listen)) - key = args.key - - # zenoh-net code --- --- --- --- --- --- --- --- --- --- --- - - async def listener(sample): - print(">> [Subscriber] Received {} ('{}': '{}')" - .format(sample.kind, sample.key_expr, sample.payload.decode("utf-8"))) - - # initiate logging - zenoh.init_logger() - - print("Openning session...") - session = await zenoh.async_open(conf) - - print("Creating Subscriber on '{}'...".format(key)) - - sub = await session.subscribe(key, listener, reliability=Reliability.Reliable, mode=SubMode.Push) - - print("Enter 'q' to quit...") - c = '\0' - while c != 'q': - c = sys.stdin.read(1) - if c == '': - time.sleep(1) - - await sub.close() - await session.close() - - -asyncio.run(main()) diff --git a/examples/z_delete.py b/examples/z_delete.py index 9839123c..1be4e87a 100644 --- a/examples/z_delete.py +++ b/examples/z_delete.py @@ -38,7 +38,7 @@ type=str, help='Endpoints to listen on.') parser.add_argument('--key', '-k', dest='key', - default='/demo/example/zenoh-python-put', + default='demo/example/zenoh-python-put', type=str, help='The key expression matching resources to delete.') parser.add_argument('--config', '-c', dest='config', @@ -47,8 +47,7 @@ help='A configuration file.') args = parser.parse_args() -conf = zenoh.config_from_file( - args.config) if args.config is not None else zenoh.Config() +conf = zenoh.Config.from_file(args.config) if args.config is not None else zenoh.Config() if args.mode is not None: conf.insert_json5(zenoh.config.MODE_KEY, json.dumps(args.mode)) if args.connect is not None: diff --git a/examples/z_get.py b/examples/z_get.py index d62cf3af..c38a9413 100644 --- a/examples/z_get.py +++ b/examples/z_get.py @@ -17,7 +17,7 @@ import argparse import json import zenoh -from zenoh import config, queryable, QueryTarget, Target +from zenoh import config, QueryTarget # --- Command line argument parsing --- --- --- --- --- --- parser = argparse.ArgumentParser( @@ -38,14 +38,9 @@ type=str, help='Endpoints to listen on.') parser.add_argument('--selector', '-s', dest='selector', - default='/demo/example/', + default='demo/example/', type=str, help='The selection of resources to query.') -parser.add_argument('--kind', '-k', dest='kind', - choices=['ALL_KINDS', 'STORAGE', 'EVAL'], - default='ALL_KINDS', - type=str, - help='The KIND of queryables to query.') parser.add_argument('--target', '-t', dest='target', choices=['ALL', 'BEST_MATCHING', 'ALL_COMPLETE', 'NONE'], default='ALL', @@ -66,15 +61,10 @@ if args.listen is not None: conf.insert_json5(zenoh.config.LISTEN_KEY, json.dumps(args.listen)) selector = args.selector -kind = { - 'ALL_KINDS': queryable.ALL_KINDS, - 'STORAGE': queryable.STORAGE, - 'EVAL': queryable.EVAL}.get(args.kind) target = { - 'ALL': Target.All(), - 'BEST_MATCHING': Target.BestMatching(), - 'ALL_COMPLETE': Target.AllComplete(), - 'NONE': Target.No()}.get(args.target) + 'ALL': QueryTarget.ALL(), + 'BEST_MATCHING': QueryTarget.BEST_MATCHING(), + 'ALL_COMPLETE': QueryTarget.ALL_COMPLETE(),}.get(args.target) # zenoh-net code --- --- --- --- --- --- --- --- --- --- --- @@ -85,9 +75,14 @@ session = zenoh.open(conf) print("Sending Query '{}'...".format(selector)) -replies = session.get(selector, target=QueryTarget(kind, target)) -for reply in replies: - print(">> Received ('{}': '{}')" - .format(reply.sample.key_expr, reply.sample.payload.decode("utf-8"))) +replies = session.get(selector, zenoh.ListCollector(), target=target) +for reply in replies(): + try: + print(">> Received ('{}': '{}')" + .format(reply.ok.key_expr, reply.ok.payload.decode("utf-8"))) + except: + print(">> Received (ERROR: '{}')" + .format(reply.err.payload.decode("utf-8"))) + session.close() diff --git a/examples/z_info.py b/examples/z_info.py index 9083b1b1..58f4b300 100644 --- a/examples/z_info.py +++ b/examples/z_info.py @@ -17,7 +17,6 @@ import argparse import json import zenoh -from zenoh import config # --- Command line argument parsing --- --- --- --- --- --- parser = argparse.ArgumentParser( @@ -60,7 +59,7 @@ session = zenoh.open(conf) info = session.info() -for key in info: - print("{} : {}".format(key, info[key])) - +print(f"zid: {info.zid()}") +print(f"routers: {info.routers_zid()}") +print(f"peers: {info.peers_zid()}") session.close() diff --git a/examples/z_pub.py b/examples/z_pub.py index d26c790c..377ffd66 100644 --- a/examples/z_pub.py +++ b/examples/z_pub.py @@ -39,7 +39,7 @@ type=str, help='Endpoints to listen on.') parser.add_argument('--key', '-k', dest='key', - default='/demo/example/zenoh-python-pub', + default='demo/example/zenoh-python-pub', type=str, help='The key expression to publish onto.') parser.add_argument('--value', '-v', dest='value', @@ -54,8 +54,7 @@ help='A configuration file.') args = parser.parse_args() -conf = zenoh.config_from_file( - args.config) if args.config is not None else zenoh.Config() +conf = zenoh.Config.from_file(args.config) if args.config is not None else zenoh.Config() if args.mode is not None: conf.insert_json5(zenoh.config.MODE_KEY, json.dumps(args.mode)) if args.connect is not None: @@ -65,27 +64,20 @@ key = args.key value = args.value -# zenoh-net code --- --- --- --- --- --- --- --- --- --- --- - # initiate logging zenoh.init_logger() print("Openning session...") session = zenoh.open(conf) -print("Declaring key expression '{}'...".format(key), end='') -rid = session.declare_expr(key) -print(" => RId {}".format(rid)) - -print("Declaring publication on '{}'...".format(rid)) -session.declare_publication(rid) +print(f"Declaring publication on '{key}'...") +pub = session.declare_publisher(key) for idx in itertools.count() if args.iter is None else range(args.iter): time.sleep(1) - buf = "[{:4d}] {}".format(idx, value) - print("Putting Data ('{}': '{}')...".format(rid, buf)) - session.put(rid, buf) + buf = f"[{idx:4d}] {value}" + print(f"Putting Data ('{key}': '{buf}')...") + pub.put(buf) -session.undeclare_publication(rid) -session.undeclare_expr(rid) +pub.undeclare() session.close() diff --git a/examples/z_pub_thr.py b/examples/z_pub_thr.py index a40d6fb9..f29008e0 100644 --- a/examples/z_pub_thr.py +++ b/examples/z_pub_thr.py @@ -17,7 +17,7 @@ import argparse import json import zenoh -from zenoh import config, CongestionControl +from zenoh import config, CongestionControl, Value # --- Command line argument parsing --- --- --- --- --- --- parser = argparse.ArgumentParser( @@ -46,8 +46,7 @@ help='A configuration file.') args = parser.parse_args() -conf = zenoh.config_from_file( - args.config) if args.config is not None else zenoh.Config() +conf = zenoh.Config.from_file(args.config) if args.config is not None else zenoh.Config() if args.mode is not None: conf.insert_json5(zenoh.config.MODE_KEY, json.dumps(args.mode)) if args.connect is not None: @@ -64,14 +63,11 @@ data = bytearray() for i in range(0, size): data.append(i % 10) -data = bytes(data) -congestion_control = CongestionControl.Block +data = Value(bytes(data)) +congestion_control = CongestionControl.BLOCK() session = zenoh.open(conf) - -rid = session.declare_expr('/test/thr') - -pub = session.declare_publication(rid) +pub = session.declare_publisher('test/thr', congestion_control=congestion_control) while True: - session.put(rid, data, congestion_control=congestion_control) + pub.put(data) diff --git a/examples/z_pull.py b/examples/z_pull.py index 856bbdda..48161b06 100644 --- a/examples/z_pull.py +++ b/examples/z_pull.py @@ -18,7 +18,7 @@ import argparse import json import zenoh -from zenoh import Reliability, SubMode +from zenoh import Reliability # --- Command line argument parsing --- --- --- --- --- --- parser = argparse.ArgumentParser( @@ -39,7 +39,7 @@ type=str, help='Endpoints to listen on.') parser.add_argument('--key', '-k', dest='key', - default='/demo/example/', + default='demo/example/', type=str, help='The key expression matching resources to pull.') parser.add_argument('--config', '-c', dest='config', @@ -62,10 +62,7 @@ def listen(sample): - time = '(not specified)' if sample.source_info is None or sample.timestamp is None else datetime.fromtimestamp( - sample.timestamp.time) - print(">> [Subscriber] Received {} ('{}': '{}')" - .format(sample.kind, sample.key_expr, sample.payload.decode("utf-8"), time)) + print(f">> [Subscriber] Received {sample.kind} ('{sample.key_expr}': '{sample.payload.decode('utf-8')}')") # initiate logging @@ -76,8 +73,7 @@ def listen(sample): print("Creating Subscriber on '{}'...".format(key)) -sub = session.subscribe( - key, listen, reliability=Reliability.Reliable, mode=SubMode.Pull) +sub = session.declare_pull_subscriber(key, listen, reliability=Reliability.RELIABLE()) print("Press <enter> to pull data...") c = '\0' @@ -88,5 +84,5 @@ def listen(sample): else: sub.pull() -sub.close() +sub.undeclare() session.close() diff --git a/examples/z_put.py b/examples/z_put.py index de83001c..68c24f38 100644 --- a/examples/z_put.py +++ b/examples/z_put.py @@ -38,7 +38,7 @@ type=str, help='Endpoints to listen on.') parser.add_argument('--key', '-k', dest='key', - default='/demo/example/zenoh-python-put', + default='demo/example/zenoh-python-put', type=str, help='The key expression to write.') parser.add_argument('--value', '-v', dest='value', diff --git a/examples/z_eval.py b/examples/z_queryable.py similarity index 87% rename from examples/z_eval.py rename to examples/z_queryable.py index cdfe1ce5..62bf85ca 100644 --- a/examples/z_eval.py +++ b/examples/z_queryable.py @@ -18,12 +18,11 @@ import json import zenoh from zenoh import config, Sample -from zenoh.queryable import EVAL # --- Command line argument parsing --- --- --- --- --- --- parser = argparse.ArgumentParser( - prog='z_eval', - description='zenoh eval example') + prog='z_queryable', + description='zenoh queryable example') parser.add_argument('--mode', '-m', dest='mode', choices=['peer', 'client'], type=str, @@ -39,11 +38,11 @@ type=str, help='Endpoints to listen on.') parser.add_argument('--key', '-k', dest='key', - default='/demo/example/zenoh-python-eval', + default='demo/example/zenoh-python-queryable', type=str, - help='The key expression matching queries to evaluate.') + help='The key expression matching queries to reply to.') parser.add_argument('--value', '-v', dest='value', - default='Eval from Python!', + default='Queryable from Python!', type=str, help='The value to reply to queries.') parser.add_argument('--config', '-c', dest='config', @@ -66,9 +65,9 @@ # zenoh-net code --- --- --- --- --- --- --- --- --- --- --- -def eval_callback(query): +def queryable_callback(query): print(">> [Queryable ] Received Query '{}'".format(query.selector)) - query.reply(Sample(key_expr=key, payload=value.encode())) + query.reply(Sample(key, value)) # initiate logging @@ -78,7 +77,7 @@ def eval_callback(query): session = zenoh.open(conf) print("Creating Queryable on '{}'...".format(key)) -queryable = session.queryable(key, eval_callback, kind=EVAL) +queryable = session.declare_queryable(key, queryable_callback) print("Enter 'q' to quit......") c = '\0' @@ -87,5 +86,5 @@ def eval_callback(query): if c == '': time.sleep(1) -queryable.close() +queryable.undeclare() session.close() diff --git a/examples/z_scout.py b/examples/z_scout.py index b910f880..fc2da860 100644 --- a/examples/z_scout.py +++ b/examples/z_scout.py @@ -12,17 +12,16 @@ # ZettaScale Zenoh Team, <[email protected]> # -import sys -import time -import argparse import zenoh -from zenoh import WhatAmI # initiate logging zenoh.init_logger() print("Scouting...") -hellos = zenoh.scout(WhatAmI.Peer \| WhatAmI.Router, 1.0) +scout = zenoh.scout(what = "peer\|router", timeout=1.0) -for hello in hellos: +def dbg(x): + print(x) + return x +for hello in dbg(scout.receiver()): print(hello) diff --git a/examples/z_storage.py b/examples/z_storage.py index 15ad0a3d..0baa4c9b 100644 --- a/examples/z_storage.py +++ b/examples/z_storage.py @@ -17,8 +17,7 @@ import argparse import json import zenoh -from zenoh import Reliability, SampleKind, SubMode, Sample, KeyExpr -from zenoh.queryable import STORAGE +from zenoh import Reliability, SampleKind, Query, Sample, KeyExpr # --- Command line argument parsing --- --- --- --- --- --- parser = argparse.ArgumentParser( @@ -39,7 +38,7 @@ type=str, help='Endpoints to listen on.') parser.add_argument('--key', '-k', dest='key', - default='/demo/example/', + default='demo/example/', type=str, help='The key expression matching resources to store.') parser.add_argument('--config', '-c', dest='config', @@ -63,22 +62,20 @@ store = {} -def listener(sample): +def listener(sample: Sample): print(">> [Subscriber] Received {} ('{}': '{}')" .format(sample.kind, sample.key_expr, sample.payload.decode("utf-8"))) if sample.kind == SampleKind.DELETE: - store.pop(str(sample.key_expr), None) + store.pop(sample.key_expr, None) else: - store[str(sample.key_expr)] = (sample.value, sample.source_info) + store[sample.key_expr] = sample -def query_handler(query): +def query_handler(query: Query): print(">> [Queryable ] Received Query '{}'".format(query.selector)) replies = [] - for stored_name, (data, source_info) in store.items(): - if KeyExpr.intersect(query.key_selector, stored_name): - sample = Sample(stored_name, data) - sample.with_source_info(source_info) + for stored_name, sample in store.items(): + if query.key_expr.intersects(stored_name): query.reply(sample) @@ -89,11 +86,10 @@ def query_handler(query): session = zenoh.open(conf) print("Creating Subscriber on '{}'...".format(key)) -sub = session.subscribe( - key, listener, reliability=Reliability.Reliable, mode=SubMode.Push) +sub = session.declare_subscriber(key, listener, reliability=Reliability.RELIABLE()) print("Creating Queryable on '{}'...".format(key)) -queryable = session.queryable(key, query_handler, kind=STORAGE) +queryable = session.declare_queryable(key, query_handler) print("Enter 'q' to quit......") c = '\0' @@ -102,6 +98,6 @@ def query_handler(query): if c == '': time.sleep(1) -sub.close() -queryable.close() +sub.undeclare() +queryable.undeclare() session.close() diff --git a/examples/z_sub.py b/examples/z_sub.py index 0b87a1cb..d126ea27 100644 --- a/examples/z_sub.py +++ b/examples/z_sub.py @@ -18,7 +18,7 @@ import argparse import json import zenoh -from zenoh import Reliability, SubMode +from zenoh import Reliability, Sample # --- Command line argument parsing --- --- --- --- --- --- parser = argparse.ArgumentParser( @@ -39,7 +39,7 @@ type=str, help='Endpoints to listen on.') parser.add_argument('--key', '-k', dest='key', - default='/demo/example/', + default='demo/example/', type=str, help='The key expression to subscribe to.') parser.add_argument('--config', '-c', dest='config', @@ -61,10 +61,6 @@ # zenoh-net code --- --- --- --- --- --- --- --- --- --- --- -def listener(sample): - print(">> [Subscriber] Received {} ('{}': '{}')" - .format(sample.kind, sample.key_expr, sample.payload.decode("utf-8"))) - # initiate logging zenoh.init_logger() @@ -74,8 +70,15 @@ def listener(sample): print("Creating Subscriber on '{}'...".format(key)) -sub = session.subscribe( - key, listener, reliability=Reliability.Reliable, mode=SubMode.Push) + +def listener(sample: Sample): + print(f">> [Subscriber] Received {sample.kind} ('{sample.key_expr}': '{sample.payload.decode('utf-8')}')") + + +# WARNING, you MUST store the return value in order for the subscription to work!! +# This is because if you don't, the reference counter will reach 0 and the subscription +# will be immediately undeclared. +sub = session.declare_subscriber(key, listener, reliability=Reliability.RELIABLE()) print("Enter 'q' to quit...") c = '\0' @@ -84,5 +87,7 @@ def listener(sample): if c == '': time.sleep(1) -sub.close() +# Cleanup: note that even if you forget it, cleanup will happen automatically when +# the reference counter reaches 0 +sub.undeclare() session.close() diff --git a/examples/z_sub_queued.py b/examples/z_sub_queued.py new file mode 100644 index 00000000..98489643 --- /dev/null +++ b/examples/z_sub_queued.py @@ -0,0 +1,97 @@ +# +# Copyright (c) 2022 ZettaScale Technology +# +# This program and the accompanying materials are made available under the +# terms of the Eclipse Public License 2.0 which is available at +# http://www.eclipse.org/legal/epl-2.0, or the Apache License, Version 2.0 +# which is available at https://www.apache.org/licenses/LICENSE-2.0. +# +# SPDX-License-Identifier: EPL-2.0 OR Apache-2.0 +# +# Contributors: +# ZettaScale Zenoh Team, <[email protected]> +# + +import sys +import time +from datetime import datetime +import argparse +import json +import zenoh +from threading import Thread +from zenoh import Reliability, Sample + +# --- Command line argument parsing --- --- --- --- --- --- +parser = argparse.ArgumentParser( + prog='z_sub', + description='zenoh sub example') +parser.add_argument('--mode', '-m', dest='mode', + choices=['peer', 'client'], + type=str, + help='The zenoh session mode.') +parser.add_argument('--connect', '-e', dest='connect', + metavar='ENDPOINT', + action='append', + type=str, + help='Endpoints to connect to.') +parser.add_argument('--listen', '-l', dest='listen', + metavar='ENDPOINT', + action='append', + type=str, + help='Endpoints to listen on.') +parser.add_argument('--key', '-k', dest='key', + default='demo/example/', + type=str, + help='The key expression to subscribe to.') +parser.add_argument('--config', '-c', dest='config', + metavar='FILE', + type=str, + help='A configuration file.') + +args = parser.parse_args() +conf = zenoh.config_from_file( + args.config) if args.config is not None else zenoh.Config() +if args.mode is not None: + conf.insert_json5(zenoh.config.MODE_KEY, json.dumps(args.mode)) +if args.connect is not None: + conf.insert_json5(zenoh.config.CONNECT_KEY, json.dumps(args.connect)) +if args.listen is not None: + conf.insert_json5(zenoh.config.LISTEN_KEY, json.dumps(args.listen)) +key = args.key + +# zenoh-net code --- --- --- --- --- --- --- --- --- --- --- + + + +# initiate logging +zenoh.init_logger() + +print("Openning session...") +session = zenoh.open(conf) + +print("Creating Subscriber on '{}'...".format(key)) + + +# WARNING, you MUST store the return value in order for the subscription to work!! +# This is because if you don't, the reference counter will reach 0 and the subscription +# will be immediately undeclared. +sub = session.declare_subscriber(key, zenoh.Queue(), reliability=Reliability.RELIABLE()) + +def consumer(): + for sample in sub.receiver: # zenoh.Queue's receiver (the queue itself) is an iterator + print(f">> [Subscriber] Received {sample.kind} ('{sample.key_expr}': '{sample.payload.decode('utf-8')}')") + +t = Thread(target=consumer) +t.start() +print("Enter 'q' to quit...") +c = '\0' +while c != 'q': + c = sys.stdin.read(1) + if c == '': + time.sleep(1) + +# Cleanup: note that even if you forget it, cleanup will happen automatically when +# the reference counter reaches 0 +sub.undeclare() +t.join() +session.close() diff --git a/examples/z_sub_thr.py b/examples/z_sub_thr.py index 86f88050..18a592ef 100644 --- a/examples/z_sub_thr.py +++ b/examples/z_sub_thr.py @@ -14,11 +14,10 @@ import sys import time -import datetime import argparse import json import zenoh -from zenoh import Reliability, SubMode +from zenoh import Reliability # --- Command line argument parsing --- --- --- --- --- --- parser = argparse.ArgumentParser( @@ -38,12 +37,6 @@ action='append', type=str, help='Endpoints to listen on.') -parser.add_argument('--samples', '-s', dest='samples', - default=10, - metavar='NUMBER', - action='append', - type=int, - help='Number of throughput measurements.') parser.add_argument('--number', '-n', dest='number', default=50000, metavar='NUMBER', @@ -56,56 +49,58 @@ help='A configuration file.') args = parser.parse_args() -conf = zenoh.config_from_file( - args.config) if args.config is not None else zenoh.Config() +conf = zenoh.Config.from_file(args.config) if args.config is not None else zenoh.Config() if args.mode is not None: conf.insert_json5(zenoh.config.MODE_KEY, json.dumps(args.mode)) if args.connect is not None: conf.insert_json5(zenoh.config.CONNECT_KEY, json.dumps(args.connect)) if args.listen is not None: conf.insert_json5(zenoh.config.LISTEN_KEY, json.dumps(args.listen)) -m = args.samples n = args.number -# zenoh-net code --- --- --- --- --- --- --- --- --- --- --- - - -def print_stats(start): - stop = datetime.datetime.now() - print("{:.6f} msgs/sec".format(n / (stop - start).total_seconds())) +batch_count = 0 count = 0 start = None -nm = 0 - +global_start = None def listener(sample): - global n, m, count, start, nm + global n, count, batch_count, start, global_start if count == 0: - start = datetime.datetime.now() + start = time.time() + if global_start is None: + global_start = start count += 1 elif count < n: count += 1 else: - print_stats(start) - nm += 1 + stop = time.time() + print(f"{n / (stop - start):.6f} msgs/sec") + batch_count += 1 count = 0 - if nm >= m: - sys.exit(0) +def report(): + global n, m, count, batch_count, global_start + end = time.time() + total = batch_count * n + count + print(f"Received {total} messages in {end - global_start}: averaged {total / (end - global_start):.6f} msgs/sec") # initiate logging zenoh.init_logger() session = zenoh.open(conf) -rid = session.declare_expr('/test/thr') - -sub = session.subscribe( - rid, listener, reliablity=Reliability.Reliable, mode=SubMode.Push) +sub = session.declare_subscriber("test/thr", (listener, report), reliability=Reliability.RELIABLE()) -time.sleep(600) +print("Enter 'q' to quit...") +c = '\0' +while c != 'q': + c = sys.stdin.read(1) + if c == '': + time.sleep(1) -session.undeclare_expr(rid) +sub.undeclare() session.close() +# while `sub.undeclare()` only returns once the unsubscription is done (no more callbacks will be queued from that instant), already queued callbacks may still be running in threads that Python can't see. +time.sleep(0.1) \ No newline at end of file diff --git a/rustup-install.sh b/rustup-install.sh deleted file mode 100644 index a4a79a29..00000000 --- a/rustup-install.sh +++ /dev/null @@ -1,662 +0,0 @@ -#!/bin/sh -# shellcheck shell=dash - -# This is just a little script that can be downloaded from the internet to -# install rustup. It just does platform detection, downloads the installer -# and runs it. - -# It runs on Unix shells like {a,ba,da,k,z}sh. It uses the common `local` -# extension. Note: Most shells limit `local` to 1 var per line, contra bash. - -if [ "$KSH_VERSION" = 'Version JM 93t+ 2010-03-05' ]; then - # The version of ksh93 that ships with many illumos systems does not - # support the "local" extension. Print a message rather than fail in - # subtle ways later on: - echo 'rustup does not work with this ksh93 version; please try bash!' >&2 - exit 1 -fi - - -set -u - -# If RUSTUP_UPDATE_ROOT is unset or empty, default it. -RUSTUP_UPDATE_ROOT="${RUSTUP_UPDATE_ROOT:-https://static.rust-lang.org/rustup}" - -#XXX: If you change anything here, please make the same changes in setup_mode.rs -usage() { - cat 1>&2 <<EOF -rustup-init 1.24.3 (c1c769109 2021-05-31) -The installer for rustup - -USAGE: - rustup-init [FLAGS] [OPTIONS] - -FLAGS: - -v, --verbose Enable verbose output - -q, --quiet Disable progress output - -y Disable confirmation prompt. - --no-modify-path Don't configure the PATH environment variable - -h, --help Prints help information - -V, --version Prints version information - -OPTIONS: - --default-host <default-host> Choose a default host triple - --default-toolchain <default-toolchain> Choose a default toolchain to install - --default-toolchain none Do not install any toolchains - --profile [minimal\|default\|complete] Choose a profile - -c, --component <components>... Component name to also install - -t, --target <targets>... Target name to also install -EOF -} - -main() { - downloader --check - need_cmd uname - need_cmd mktemp - need_cmd chmod - need_cmd mkdir - need_cmd rm - need_cmd rmdir - - get_architecture \|\| return 1 - local _arch="$RETVAL" - assert_nz "$_arch" "arch" - - local _ext="" - case "$_arch" in - windows) - _ext=".exe" - ;; - esac - - local _url="${RUSTUP_UPDATE_ROOT}/dist/${_arch}/rustup-init${_ext}" - - local _dir - _dir="$(ensure mktemp -d)" - local _file="${_dir}/rustup-init${_ext}" - - local _ansi_escapes_are_valid=false - if [ -t 2 ]; then - if [ "${TERM+set}" = 'set' ]; then - case "$TERM" in - xterm\|rxvt\|urxvt\|linux\|vt) - _ansi_escapes_are_valid=true - ;; - esac - fi - fi - - # check if we have to use /dev/tty to prompt the user - local need_tty=yes - for arg in "$@"; do - case "$arg" in - -h\|--help) - usage - exit 0 - ;; - -y) - # user wants to skip the prompt -- we don't need /dev/tty - need_tty=no - ;; - ) - ;; - esac - done - - if $_ansi_escapes_are_valid; then - printf "\33[1minfo:\33[0m downloading installer\n" 1>&2 - else - printf '%s\n' 'info: downloading installer' 1>&2 - fi - - ensure mkdir -p "$_dir" - ensure downloader "$_url" "$_file" "$_arch" - ensure chmod u+x "$_file" - if [ ! -x "$_file" ]; then - printf '%s\n' "Cannot execute $_file (likely because of mounting /tmp as noexec)." 1>&2 - printf '%s\n' "Please copy the file to a location where you can execute binaries and run ./rustup-init${_ext}." 1>&2 - exit 1 - fi - - if [ "$need_tty" = "yes" ]; then - # The installer is going to want to ask for confirmation by - # reading stdin. This script was piped into `sh` though and - # doesn't have stdin to pass to its children. Instead we're going - # to explicitly connect /dev/tty to the installer's stdin. - if [ ! -t 1 ]; then - err "Unable to run interactively. Run with -y to accept defaults, --help for additional options" - fi - - ignore "$_file" "$@" < /dev/tty - else - ignore "$_file" "$@" - fi - - local _retval=$? - - ignore rm "$_file" - ignore rmdir "$_dir" - - return "$_retval" -} - -check_proc() { - # Check for /proc by looking for the /proc/self/exe link - # This is only run on Linux - if ! test -L /proc/self/exe ; then - err "fatal: Unable to find /proc/self/exe. Is /proc mounted? Installation cannot proceed without /proc." - fi -} - -get_bitness() { - need_cmd head - # Architecture detection without dependencies beyond coreutils. - # ELF files start out "\x7fELF", and the following byte is - # 0x01 for 32-bit and - # 0x02 for 64-bit. - # The printf builtin on some shells like dash only supports octal - # escape sequences, so we use those. - local _current_exe_head - _current_exe_head=$(head -c 5 /proc/self/exe ) - if [ "$_current_exe_head" = "$(printf '\177ELF\001')" ]; then - echo 32 - elif [ "$_current_exe_head" = "$(printf '\177ELF\002')" ]; then - echo 64 - else - err "unknown platform bitness" - fi -} - -is_host_amd64_elf() { - need_cmd head - need_cmd tail - # ELF e_machine detection without dependencies beyond coreutils. - # Two-byte field at offset 0x12 indicates the CPU, - # but we're interested in it being 0x3E to indicate amd64, or not that. - local _current_exe_machine - _current_exe_machine=$(head -c 19 /proc/self/exe \| tail -c 1) - [ "$_current_exe_machine" = "$(printf '\076')" ] -} - -get_endianness() { - local cputype=$1 - local suffix_eb=$2 - local suffix_el=$3 - - # detect endianness without od/hexdump, like get_bitness() does. - need_cmd head - need_cmd tail - - local _current_exe_endianness - _current_exe_endianness="$(head -c 6 /proc/self/exe \| tail -c 1)" - if [ "$_current_exe_endianness" = "$(printf '\001')" ]; then - echo "${cputype}${suffix_el}" - elif [ "$_current_exe_endianness" = "$(printf '\002')" ]; then - echo "${cputype}${suffix_eb}" - else - err "unknown platform endianness" - fi -} - -get_architecture() { - local _ostype _cputype _bitness _arch _clibtype - _ostype="$(uname -s)" - _cputype="$(uname -m)" - _clibtype="gnu" - - if [ "$_ostype" = Linux ]; then - if [ "$(uname -o)" = Android ]; then - _ostype=Android - fi - if ldd --version 2>&1 \| grep -q 'musl'; then - _clibtype="musl" - fi - fi - - if [ "$_ostype" = Darwin ] && [ "$_cputype" = i386 ]; then - # Darwin `uname -m` lies - if sysctl hw.optional.x86_64 \| grep -q ': 1'; then - _cputype=x86_64 - fi - fi - - if [ "$_ostype" = SunOS ]; then - # Both Solaris and illumos presently announce as "SunOS" in "uname -s" - # so use "uname -o" to disambiguate. We use the full path to the - # system uname in case the user has coreutils uname first in PATH, - # which has historically sometimes printed the wrong value here. - if [ "$(/usr/bin/uname -o)" = illumos ]; then - _ostype=illumos - fi - - # illumos systems have multi-arch userlands, and "uname -m" reports the - # machine hardware name; e.g., "i86pc" on both 32- and 64-bit x86 - # systems. Check for the native (widest) instruction set on the - # running kernel: - if [ "$_cputype" = i86pc ]; then - _cputype="$(isainfo -n)" - fi - fi - - case "$_ostype" in - - Android) - _ostype=linux-android - ;; - - Linux) - check_proc - _ostype=unknown-linux-$_clibtype - _bitness=$(get_bitness) - ;; - - FreeBSD) - _ostype=unknown-freebsd - ;; - - NetBSD) - _ostype=unknown-netbsd - ;; - - DragonFly) - _ostype=unknown-dragonfly - ;; - - Darwin) - _ostype=apple-darwin - ;; - - illumos) - _ostype=unknown-illumos - ;; - - MINGW* \| MSYS* \| CYGWIN) - _ostype=pc-windows-gnu - ;; - - ) - err "unrecognized OS type: $_ostype" - ;; - - esac - - case "$_cputype" in - - i386 \| i486 \| i686 \| i786 \| x86) - _cputype=i686 - ;; - - xscale \| arm) - _cputype=arm - if [ "$_ostype" = "linux-android" ]; then - _ostype=linux-androideabi - fi - ;; - - armv6l) - _cputype=arm - if [ "$_ostype" = "linux-android" ]; then - _ostype=linux-androideabi - else - _ostype="${_ostype}eabihf" - fi - ;; - - armv7l \| armv8l) - _cputype=armv7 - if [ "$_ostype" = "linux-android" ]; then - _ostype=linux-androideabi - else - _ostype="${_ostype}eabihf" - fi - ;; - - aarch64 \| arm64) - _cputype=aarch64 - ;; - - x86_64 \| x86-64 \| x64 \| amd64) - _cputype=x86_64 - ;; - - mips) - _cputype=$(get_endianness mips '' el) - ;; - - mips64) - if [ "$_bitness" -eq 64 ]; then - # only n64 ABI is supported for now - _ostype="${_ostype}abi64" - _cputype=$(get_endianness mips64 '' el) - fi - ;; - - ppc) - _cputype=powerpc - ;; - - ppc64) - _cputype=powerpc64 - ;; - - ppc64le) - _cputype=powerpc64le - ;; - - s390x) - _cputype=s390x - ;; - riscv64) - _cputype=riscv64gc - ;; - ) - err "unknown CPU type: $_cputype" - - esac - - # Detect 64-bit linux with 32-bit userland - if [ "${_ostype}" = unknown-linux-gnu ] && [ "${_bitness}" -eq 32 ]; then - case $_cputype in - x86_64) - if [ -n "${RUSTUP_CPUTYPE:-}" ]; then - _cputype="$RUSTUP_CPUTYPE" - else { - # 32-bit executable for amd64 = x32 - if is_host_amd64_elf; then { - echo "This host is running an x32 userland; as it stands, x32 support is poor," 1>&2 - echo "and there isn't a native toolchain -- you will have to install" 1>&2 - echo "multiarch compatibility with i686 and/or amd64, then select one" 1>&2 - echo "by re-running this script with the RUSTUP_CPUTYPE environment variable" 1>&2 - echo "set to i686 or x86_64, respectively." 1>&2 - echo 1>&2 - echo "You will be able to add an x32 target after installation by running" 1>&2 - echo " rustup target add x86_64-unknown-linux-gnux32" 1>&2 - exit 1 - }; else - _cputype=i686 - fi - }; fi - ;; - mips64) - _cputype=$(get_endianness mips '' el) - ;; - powerpc64) - _cputype=powerpc - ;; - aarch64) - _cputype=armv7 - if [ "$_ostype" = "linux-android" ]; then - _ostype=linux-androideabi - else - _ostype="${_ostype}eabihf" - fi - ;; - riscv64gc) - err "riscv64 with 32-bit userland unsupported" - ;; - esac - fi - - # Detect armv7 but without the CPU features Rust needs in that build, - # and fall back to arm. - # See https://github.com/rust-lang/rustup.rs/issues/587. - if [ "$_ostype" = "unknown-linux-gnueabihf" ] && [ "$_cputype" = armv7 ]; then - if ensure grep '^Features' /proc/cpuinfo \| grep -q -v neon; then - # At least one processor does not have NEON. - _cputype=arm - fi - fi - - _arch="${_cputype}-${_ostype}" - - RETVAL="$_arch" -} - -say() { - printf 'rustup: %s\n' "$1" -} - -err() { - say "$1" >&2 - exit 1 -} - -need_cmd() { - if ! check_cmd "$1"; then - err "need '$1' (command not found)" - fi -} - -check_cmd() { - command -v "$1" > /dev/null 2>&1 -} - -assert_nz() { - if [ -z "$1" ]; then err "assert_nz $2"; fi -} - -# Run a command that should never fail. If the command fails execution -# will immediately terminate with an error showing the failing -# command. -ensure() { - if ! "$@"; then err "command failed: $"; fi -} - -# This is just for indicating that commands' results are being -# intentionally ignored. Usually, because it's being executed -# as part of error handling. -ignore() { - "$@" -} - -# This wraps curl or wget. Try curl first, if not installed, -# use wget instead. -downloader() { - local _dld - local _ciphersuites - local _err - local _status - if check_cmd curl; then - _dld=curl - elif check_cmd wget; then - _dld=wget - else - _dld='curl or wget' # to be used in error message of need_cmd - fi - - if [ "$1" = --check ]; then - need_cmd "$_dld" - elif [ "$_dld" = curl ]; then - get_ciphersuites_for_curl - _ciphersuites="$RETVAL" - if [ -n "$_ciphersuites" ]; then - _err=$(curl --proto '=https' --tlsv1.2 --ciphers "$_ciphersuites" --silent --show-error --fail --location "$1" --output "$2" 2>&1) - _status=$? - else - echo "Warning: Not enforcing strong cipher suites for TLS, this is potentially less secure" - if ! check_help_for "$3" curl --proto --tlsv1.2; then - echo "Warning: Not enforcing TLS v1.2, this is potentially less secure" - _err=$(curl --silent --show-error --fail --location "$1" --output "$2" 2>&1) - _status=$? - else - _err=$(curl --proto '=https' --tlsv1.2 --silent --show-error --fail --location "$1" --output "$2" 2>&1) - _status=$? - fi - fi - if [ -n "$_err" ]; then - echo "$_err" >&2 - if echo "$_err" \| grep -q 404$; then - err "installer for platform '$3' not found, this may be unsupported" - fi - fi - return $_status - elif [ "$_dld" = wget ]; then - get_ciphersuites_for_wget - _ciphersuites="$RETVAL" - if [ -n "$_ciphersuites" ]; then - _err=$(wget --https-only --secure-protocol=TLSv1_2 --ciphers "$_ciphersuites" "$1" -O "$2" 2>&1) - _status=$? - else - echo "Warning: Not enforcing strong cipher suites for TLS, this is potentially less secure" - if ! check_help_for "$3" wget --https-only --secure-protocol; then - echo "Warning: Not enforcing TLS v1.2, this is potentially less secure" - _err=$(wget "$1" -O "$2" 2>&1) - _status=$? - else - _err=$(wget --https-only --secure-protocol=TLSv1_2 "$1" -O "$2" 2>&1) - _status=$? - fi - fi - if [ -n "$_err" ]; then - echo "$_err" >&2 - if echo "$_err" \| grep -q ' 404 Not Found$'; then - err "installer for platform '$3' not found, this may be unsupported" - fi - fi - return $_status - else - err "Unknown downloader" # should not reach here - fi -} - -check_help_for() { - local _arch - local _cmd - local _arg - _arch="$1" - shift - _cmd="$1" - shift - - local _category - if "$_cmd" --help \| grep -q 'For all options use the manual or "--help all".'; then - _category="all" - else - _category="" - fi - - case "$_arch" in - - darwin) - if check_cmd sw_vers; then - case $(sw_vers -productVersion) in - 10.) - # If we're running on macOS, older than 10.13, then we always - # fail to find these options to force fallback - if [ "$(sw_vers -productVersion \| cut -d. -f2)" -lt 13 ]; then - # Older than 10.13 - echo "Warning: Detected macOS platform older than 10.13" - return 1 - fi - ;; - 11.) - # We assume Big Sur will be OK for now - ;; - ) - # Unknown product version, warn and continue - echo "Warning: Detected unknown macOS major version: $(sw_vers -productVersion)" - echo "Warning TLS capabilities detection may fail" - ;; - esac - fi - ;; - - esac - - for _arg in "$@"; do - if ! "$_cmd" --help $_category \| grep -q -- "$_arg"; then - return 1 - fi - done - - true # not strictly needed -} - -# Return cipher suite string specified by user, otherwise return strong TLS 1.2-1.3 cipher suites -# if support by local tools is detected. Detection currently supports these curl backends: -# GnuTLS and OpenSSL (possibly also LibreSSL and BoringSSL). Return value can be empty. -get_ciphersuites_for_curl() { - if [ -n "${RUSTUP_TLS_CIPHERSUITES-}" ]; then - # user specified custom cipher suites, assume they know what they're doing - RETVAL="$RUSTUP_TLS_CIPHERSUITES" - return - fi - - local _openssl_syntax="no" - local _gnutls_syntax="no" - local _backend_supported="yes" - if curl -V \| grep -q ' OpenSSL/'; then - _openssl_syntax="yes" - elif curl -V \| grep -iq ' LibreSSL/'; then - _openssl_syntax="yes" - elif curl -V \| grep -iq ' BoringSSL/'; then - _openssl_syntax="yes" - elif curl -V \| grep -iq ' GnuTLS/'; then - _gnutls_syntax="yes" - else - _backend_supported="no" - fi - - local _args_supported="no" - if [ "$_backend_supported" = "yes" ]; then - # "unspecified" is for arch, allows for possibility old OS using macports, homebrew, etc. - if check_help_for "notspecified" "curl" "--tlsv1.2" "--ciphers" "--proto"; then - _args_supported="yes" - fi - fi - - local _cs="" - if [ "$_args_supported" = "yes" ]; then - if [ "$_openssl_syntax" = "yes" ]; then - _cs=$(get_strong_ciphersuites_for "openssl") - elif [ "$_gnutls_syntax" = "yes" ]; then - _cs=$(get_strong_ciphersuites_for "gnutls") - fi - fi - - RETVAL="$_cs" -} - -# Return cipher suite string specified by user, otherwise return strong TLS 1.2-1.3 cipher suites -# if support by local tools is detected. Detection currently supports these wget backends: -# GnuTLS and OpenSSL (possibly also LibreSSL and BoringSSL). Return value can be empty. -get_ciphersuites_for_wget() { - if [ -n "${RUSTUP_TLS_CIPHERSUITES-}" ]; then - # user specified custom cipher suites, assume they know what they're doing - RETVAL="$RUSTUP_TLS_CIPHERSUITES" - return - fi - - local _cs="" - if wget -V \| grep -q '\-DHAVE_LIBSSL'; then - # "unspecified" is for arch, allows for possibility old OS using macports, homebrew, etc. - if check_help_for "notspecified" "wget" "TLSv1_2" "--ciphers" "--https-only" "--secure-protocol"; then - _cs=$(get_strong_ciphersuites_for "openssl") - fi - elif wget -V \| grep -q '\-DHAVE_LIBGNUTLS'; then - # "unspecified" is for arch, allows for possibility old OS using macports, homebrew, etc. - if check_help_for "notspecified" "wget" "TLSv1_2" "--ciphers" "--https-only" "--secure-protocol"; then - _cs=$(get_strong_ciphersuites_for "gnutls") - fi - fi - - RETVAL="$_cs" -} - -# Return strong TLS 1.2-1.3 cipher suites in OpenSSL or GnuTLS syntax. TLS 1.2 -# excludes non-ECDHE and non-AEAD cipher suites. DHE is excluded due to bad -# DH params often found on servers (see RFC 7919). Sequence matches or is -# similar to Firefox 68 ESR with weak cipher suites disabled via about:config. -# $1 must be openssl or gnutls. -get_strong_ciphersuites_for() { - if [ "$1" = "openssl" ]; then - # OpenSSL is forgiving of unknown values, no problems with TLS 1.3 values on versions that don't support it yet. - echo "TLS_AES_128_GCM_SHA256:TLS_CHACHA20_POLY1305_SHA256:TLS_AES_256_GCM_SHA384:ECDHE-ECDSA-AES128-GCM-SHA256:ECDHE-RSA-AES128-GCM-SHA256:ECDHE-ECDSA-CHACHA20-POLY1305:ECDHE-RSA-CHACHA20-POLY1305:ECDHE-ECDSA-AES256-GCM-SHA384:ECDHE-RSA-AES256-GCM-SHA384" - elif [ "$1" = "gnutls" ]; then - # GnuTLS isn't forgiving of unknown values, so this may require a GnuTLS version that supports TLS 1.3 even if wget doesn't. - # Begin with SECURE128 (and higher) then remove/add to build cipher suites. Produces same 9 cipher suites as OpenSSL but in slightly different order. - echo "SECURE128:-VERS-SSL3.0:-VERS-TLS1.0:-VERS-TLS1.1:-VERS-DTLS-ALL:-CIPHER-ALL:-MAC-ALL:-KX-ALL:+AEAD:+ECDHE-ECDSA:+ECDHE-RSA:+AES-128-GCM:+CHACHA20-POLY1305:+AES-256-GCM" - fi -} - -main "$@" \|\| exit 1 diff --git a/src/async_session.rs b/src/async_session.rs deleted file mode 100644 index bcd1d3f5..00000000 --- a/src/async_session.rs +++ /dev/null @@ -1,772 +0,0 @@ -// -// Copyright (c) 2017, 2022 ZettaScale Technology -// -// This program and the accompanying materials are made available under the -// terms of the Eclipse Public License 2.0 which is available at -// http://www.eclipse.org/legal/epl-2.0, or the Apache License, Version 2.0 -// which is available at https://www.apache.org/licenses/LICENSE-2.0. -// -// SPDX-License-Identifier: EPL-2.0 OR Apache-2.0 -// -// Contributors: -// ZettaScale Zenoh team, <[email protected]> -// -use super::async_types::{AsyncQueryable, AsyncSubscriber}; -use super::encoding::Encoding; -use super::sample_kind::SampleKind; -use super::types::{ - zkey_expr_of_pyany, zvalue_of_pyany, CongestionControl, KeyExpr, Period, Priority, Query, - QueryConsolidation, QueryTarget, Reliability, Reply, Sample, SubMode, ZnSubOps, -}; -use super::{to_pyerr, ZError}; -use async_std::channel::bounded; -use async_std::sync::Arc; -use async_std::task; -use futures::prelude::; -use futures::select; -use log::warn; -use pyo3::exceptions; -use pyo3::prelude::; -use pyo3::types::{IntoPyDict, PyDict, PyTuple}; -use pyo3_asyncio::async_std::future_into_py; -use std::collections::HashMap; -use zenoh::prelude::{ExprId, KeyExpr as ZKeyExpr, Selector, ZFuture, ZInt}; - -/// A zenoh session to be used with asyncio. -#[pyclass] -pub struct AsyncSession { - s: Option<Arc<zenoh::Session>>, -} - -#[pymethods] -impl AsyncSession { - // NOTE: See https://github.com/awestlake87/pyo3-asyncio/issues/50 for the options - // to implement asyncronous methods with async-pyo3. - // Here we choosed to wrap the Session in an Arc, and move a clone of it in each `future_into_py()` call. - // Similarly, each argument coming from Python is converted to Rust and cloned (or Arc-wrapped) before - // moving into the `future_into_py()` call. - - /// Close the zenoh Session. - pub fn close<'p>(&mut self, py: Python<'p>) -> PyResult<&'p PyAny> { - // NOTE: should be sufficient to take the Arc<Session>. Once all arcs are dropped, Session will close. - // Still, we should provide a wait to await for the actual closure... - let _s = self.try_take()?; - future_into_py(py, async move { Ok(()) }) - } - - /// Get informations about the zenoh Session. - /// - /// This method is a coroutine. - /// - /// :rtype: dict[str, str]* - /// - /// :Example: - /// - /// >>> import asyncio, zenoh - /// >>> async def main(): - /// >>> s = await zenoh.async_open() - /// >>> info = await s.info() - /// >>> for key in info: - /// >>> print("{} : {}".format(key, info[key])) - /// >>> - /// >>> asyncio.run(main()) - pub fn info<'p>(&self, py: Python<'p>) -> PyResult<&'p PyAny> { - let s = self.try_clone()?; - future_into_py(py, async move { - use zenoh_cfg_properties::KeyTranscoder; - let result = s - .info() - .map(\|props\| { - let hashmap: HashMap<String, String> = props - .0 - .into_iter() - .filter_map(\|(k, v)\| zenoh::info::InfoTranscoder::decode(k).map(\|k\| (k, v))) - .collect(); - Python::with_gil(\|py\| hashmap.into_py_dict(py).to_object(py)) - }) - .await; - Ok(result) - }) - } - - /// Put data. - /// - /// The key_expr parameter also accepts the following types that can be converted to a :class:`KeyExpr`: - /// - /// * int for a mapped key expression - /// * str for a literal key expression - /// * (int, str) for a mapped key expression with suffix - /// - /// This method is a coroutine. - /// - /// :param key_expr: The key expression matching resources to write - /// :type key_expr: :class:`KeyExpr` - /// :param value: The value to write - /// :type value: any type convertible to a :class:`Value` - /// :param \*kwargs: - /// See below - /// - /// :Keyword Arguments: - /// encoding (:class:`Encoding`) -- - /// Set the encoding of the written data - /// * kind ( int ) -- - /// Set the kind of the written data - /// * congestion_control (:class:`CongestionControl`) -- - /// Set the congestion control to apply when routing the data - /// * priority (:class:`Priority`) -- - /// Set the priority of the written data - /// * local_routing ( bool ) -- - /// Enable or disable local routing - /// - /// :raise: :class:`ZError` - /// - /// :Examples: - /// - /// >>> import asyncio, zenoh - /// >>> async def main(): - /// >>> s = await zenoh.async_open() - /// >>> await s.put('/key/expression', bytes('value', encoding='utf8')) - /// >>> - /// >>> asyncio.run(main()) - #[pyo3(text_signature = "(self, key_expr, value, kwargs)")] - #[args(kwargs = "")] - pub fn put<'p>( - &self, - key_expr: &PyAny, - value: &PyAny, - kwargs: Option<&PyDict>, - py: Python<'p>, - ) -> PyResult<&'p PyAny> { - let s = self.try_clone()?; - let k = zkey_expr_of_pyany(key_expr)?.to_owned(); - let mut v = zvalue_of_pyany(value)?; - let mut encoding: Option<Encoding> = None; - let mut kind: Option<SampleKind> = None; - let mut congestion_control: Option<CongestionControl> = None; - let mut priority: Option<Priority> = None; - let mut local_routing: Option<bool> = None; - if let Some(kwargs) = kwargs { - if let Some(e) = kwargs.get_item("encoding") { - encoding = e.extract().ok() - } - if let Some(k) = kwargs.get_item("kind") { - kind = k.extract().ok() - } - if let Some(cc) = kwargs.get_item("congestion_control") { - congestion_control = cc.extract().ok() - } - if let Some(p) = kwargs.get_item("priority") { - priority = p.extract().ok() - } - if let Some(lr) = kwargs.get_item("local_routing") { - local_routing = lr.extract().ok() - } - } - if let Some(encoding) = encoding { - v.encoding = encoding.into(); - } - future_into_py(py, async move { - let mut writer = s - .put(k, v) - .kind(kind.unwrap_or_default().kind) - .congestion_control(congestion_control.unwrap_or_default().cc) - .priority(priority.unwrap_or_default().p); - if let Some(local_routing) = local_routing { - writer = writer.local_routing(local_routing); - } - writer.await.map_err(to_pyerr) - }) - } - - /// Delete data. - /// - /// The key_expr parameter also accepts the following types that can be converted to a :class:`KeyExpr`: - /// - /// * int for a mapped key expression - /// * str for a literal key expression - /// * (int, str) for a mapped key expression with suffix - /// - /// This method is a coroutine. - /// - /// :param key_expr: The key expression matching resources to delete - /// :type key_expr: :class:`KeyExpr` - /// :param \*kwargs: - /// See below - /// - /// :Keyword Arguments: - /// congestion_control (:class:`CongestionControl`) -- - /// Set the congestion control to apply when routing the data - /// * priority (:class:`Priority`) -- - /// Set the priority of the written data - /// * local_routing ( bool ) -- - /// Enable or disable local routing - /// - /// :raise: :class:`ZError` - /// - /// :Examples: - /// - /// >>> import asyncio, zenoh - /// >>> async def main(): - /// >>> s = await zenoh.async_open() - /// >>> await s.delete('/key/expression') - /// >>> - /// >>> asyncio.run(main()) - #[pyo3(text_signature = "(self, key_expr, kwargs)")] - #[args(kwargs = "")] - pub fn delete<'p>( - &self, - key_expr: &PyAny, - kwargs: Option<&PyDict>, - py: Python<'p>, - ) -> PyResult<&'p PyAny> { - let s = self.try_clone()?; - let k = zkey_expr_of_pyany(key_expr)?.to_owned(); - let mut congestion_control: Option<CongestionControl> = None; - let mut priority: Option<Priority> = None; - let mut local_routing: Option<bool> = None; - if let Some(kwargs) = kwargs { - if let Some(cc) = kwargs.get_item("congestion_control") { - congestion_control = cc.extract().ok() - } - if let Some(p) = kwargs.get_item("priority") { - priority = p.extract().ok() - } - if let Some(lr) = kwargs.get_item("local_routing") { - local_routing = lr.extract().ok() - } - } - future_into_py(py, async move { - let mut writer = s - .delete(k) - .congestion_control(congestion_control.unwrap_or_default().cc) - .priority(priority.unwrap_or_default().p); - if let Some(local_routing) = local_routing { - writer = writer.local_routing(local_routing); - } - writer.await.map_err(to_pyerr) - }) - } - - /// Associate a numerical Id with the given key expression. - /// - /// This numerical Id will be used on the network to save bandwidth and - /// ease the retrieval of the concerned resource in the routing tables. - /// - /// The key_expr parameter also accepts the following types that can be converted to a :class:`KeyExpr`: - /// - /// * int for a mapped key expression - /// * str for a literal key expression - /// * (int, str) for a mapped key expression with suffix - /// - /// This method is a coroutine. - /// - /// :param key_expr: The key expression to map to a numerical Id - /// :type key_expr: :class:`KeyExpr` - /// :rtype: int - /// :raise: :class:`ZError` - /// - /// :Examples: - /// - /// >>> import asyncio, zenoh - /// >>> async def main(): - /// >>> s = await zenoh.async_open() - /// >>> rid = await s.declare_expr('/key/expression') - /// >>> - /// >>> asyncio.run(main()) - #[pyo3(text_signature = "(self, key_expr)")] - pub fn declare_expr<'p>(&self, key_expr: &PyAny, py: Python<'p>) -> PyResult<&'p PyAny> { - let s = self.try_clone()?; - let k = zkey_expr_of_pyany(key_expr)?.to_owned(); - future_into_py(py, async move { s.declare_expr(k).await.map_err(to_pyerr) }) - } - - /// Undeclare the numerical Id/key expression association previously declared - /// with :meth:`declare_expr`. - /// - /// This method is a coroutine. - /// - /// :param rid: The numerical Id to unmap - /// :type rid: ExprId - /// :raise: :class:`ZError` - /// - /// :Examples: - /// - /// >>> import asyncio, zenoh - /// >>> async def main(): - /// >>> s = await zenoh.async_open() - /// >>> rid = await s.declare_expr('/key/expression') - /// >>> await s.undeclare_expr(rid) - /// >>> - /// >>> asyncio.run(main()) - #[pyo3(text_signature = "(self, rid)")] - pub fn undeclare_expr<'p>(&self, rid: ExprId, py: Python<'p>) -> PyResult<&'p PyAny> { - let s = self.try_clone()?; - future_into_py( - py, - async move { s.undeclare_expr(rid).wait().map_err(to_pyerr) }, - ) - } - - /// Declare a publication for the given key expression. - /// - /// Written expressions that match the given key expression will only be sent on the network - /// if matching subscribers exist in the system. - /// - /// The key_expr parameter also accepts the following types that can be converted to a :class:`KeyExpr`: - /// - /// * int for a mapped key expression - /// * str for a literal key expression - /// * (int, str) for a mapped key expression with suffix - /// - /// This method is a coroutine. - /// - /// :param key_expr: The key expression to publish - /// :type key_expr: :class:`KeyExpr` - /// :raise: :class:`ZError` - /// - /// :Examples: - /// - /// >>> import asyncio, zenoh - /// >>> async def main(): - /// >>> s = await zenoh.async_open() - /// >>> rid = await s.declare_publication('/key/expression') - /// >>> await s.put('/key/expression', bytes('value', encoding='utf8')) - /// >>> - /// >>> asyncio.run(main()) - #[pyo3(text_signature = "(self, key_expr)")] - fn declare_publication<'p>(&self, key_expr: &PyAny, py: Python<'p>) -> PyResult<&'p PyAny> { - let s = self.try_clone()?; - let k = zkey_expr_of_pyany(key_expr)?.to_owned(); - future_into_py(py, async move { - s.declare_publication(k).await.map_err(to_pyerr) - }) - } - - #[pyo3(text_signature = "(self, key_expr)")] - fn undeclare_publication<'p>(&self, key_expr: &PyAny, py: Python<'p>) -> PyResult<&'p PyAny> { - let s = self.try_clone()?; - let k = zkey_expr_of_pyany(key_expr)?.to_owned(); - future_into_py(py, async move { - s.undeclare_publication(k).await.map_err(to_pyerr) - }) - } - - /// Create an AsyncSubscriber for the given key expression. - /// - /// The key_expr parameter also accepts the following types that can be converted to a :class:`KeyExpr`: - /// - /// * int for a mapped key expression - /// * str for a literal key expression - /// * (int, str) for a mapped key expression with suffix - /// - /// This method is a coroutine. - /// - /// :param key_expr: The key expression to subscribe - /// :type key_expr: :class:`KeyExpr` - /// :param callback: the subscription callback (must be a coroutine) - /// :type callback: async function(:class:`Sample`) - /// :param \*kwargs: - /// See below - /// - /// :Keyword Arguments: - /// reliability (:class:`Reliability`) -- - /// Set the subscription reliability (BestEffort by default) - /// * mode (:class:`SubMode`) -- - /// Set the subscription mode (Push by default) - /// * period (:class:`Period`) -- - /// Set the subscription period - /// * local ( bool ) -- - /// If true make the subscription local only (false by default) - /// - /// :rtype: :class:`Subscriber` - /// :raise: :class:`ZError` - /// - /// :Examples: - /// - /// >>> import asyncio, zenoh - /// >>> from zenoh import Reliability, SubMode - /// >>> - /// >>> async def callback(sample): - /// >>> print("Received : {}".format(sample)) - /// >>> - /// >>> async def main(): - /// >>> s = await zenoh.async_open() - /// >>> sub = await s.subscribe('/key/expression', callback, - /// ... reliability=Reliability.Reliable, - /// ... mode=SubMode.Push) - /// >>> await asycio.sleep(60) - /// >>> - /// >>> asyncio.run(main()) - #[pyo3(text_signature = "(self, key_expr, callback, kwargs)")] - #[args(kwargs = "")] - fn subscribe<'p>( - &self, - key_expr: &PyAny, - callback: &PyAny, - kwargs: Option<&PyDict>, - py: Python<'p>, - ) -> PyResult<&'p PyAny> { - let s = self.try_clone()?; - let k = zkey_expr_of_pyany(key_expr)?.to_owned(); - // Note: callback cannot be passed as such in task below because it's not Send - let cb_obj: Py<PyAny> = callback.into(); - // note: extract from kwargs here because it's not Send and cannot be moved into future_into_py(py, F) - let mut reliability: Option<Reliability> = None; - let mut mode: Option<SubMode> = None; - let mut period: Option<Period> = None; - let mut local = false; - if let Some(kwargs) = kwargs { - if let Some(r) = kwargs.get_item("reliability") { - reliability = Some(r.extract()?); - } - if let Some(m) = kwargs.get_item("mode") { - mode = Some(m.extract()?); - } - if let Some(p) = kwargs.get_item("period") { - period = Some(p.extract()?) - } - if let Some(p) = kwargs.get_item("local") { - local = p.extract::<bool>()?; - } - } - - future_into_py(py, async move { - // note: create SubscriberBuilder in this async block since its lifetime is bound to s which is moved here. - let mut sub_builder = s.subscribe(&k); - if let Some(r) = reliability { - sub_builder = sub_builder.reliability(r.r); - } - if let Some(m) = mode { - sub_builder = sub_builder.mode(m.m); - } - if let Some(p) = period { - sub_builder = sub_builder.period(Some(p.p)); - } - if local { - sub_builder = sub_builder.local(); - } - - let zn_sub = sub_builder.await.map_err(to_pyerr)?; - // Note: workaround to allow moving of zn_sub into the task below. - // Otherwise, s is moved also, but can't because it doesn't have 'static lifetime. - let mut static_zn_sub = unsafe { - std::mem::transmute::< - zenoh::subscriber::Subscriber<'_>, - zenoh::subscriber::Subscriber<'static>, - >(zn_sub) - }; - - let (unregister_tx, unregister_rx) = bounded::<ZnSubOps>(8); - // Note: This is done to ensure that even if the call-back into Python - // does any blocking call we do not incour the risk of blocking - // any of the task resolving futures. - let loop_handle = task::spawn_blocking(move \|\| { - Python::with_gil(\|py\| { - // Run a Python event loop in this task, to allow coroutines execution within the callback - match pyo3_asyncio::async_std::run(py, async move { - loop { - select!( - s = static_zn_sub.receiver().next().fuse() => { - // call the async callback and transform the resulting Python awaitable into a Rust future - let future = match Python::with_gil(\|py\| { - let cb_args = PyTuple::new(py, &[Sample { s: s.unwrap() }]); - cb_obj.as_ref(py).call1(cb_args).and_then(pyo3_asyncio::async_std::into_future) - }) { - Ok(f) => f, - Err(e) => { warn!("Error calling async queryable callback: {}", e); continue } - }; - // await the future (by default callbacks are executed in sequence) - if let Err(e) = future.await { - warn!("Error suring axecution of async queryable callback: {}", e); - } - }, - op = unregister_rx.recv().fuse() => { - match op { - Ok(ZnSubOps::Pull) => { - if let Err(e) = static_zn_sub.pull().await { - warn!("Error pulling the subscriber: {}", e); - } - }, - Ok(ZnSubOps::Unregister) => { - if let Err(e) = static_zn_sub.close().await { - warn!("Error undeclaring subscriber: {}", e); - } - return Ok(()) - }, - _ => return Ok(()) - } - } - ) - } - }) { - Ok(()) => warn!("Queryable loop running"), - Err(e) => warn!("Failed to start Queryable loop: {}", e), - } - }) - }); - Ok(AsyncSubscriber { - unregister_tx, - loop_handle: Some(loop_handle), - }) - }) - } - - /// Create an AsyncQueryable for the given key expression. - /// - /// The key_expr parameter also accepts the following types that can be converted to a :class:`KeyExpr`: - /// - /// * int for a mapped key expression - /// * str for a literal key expression - /// * (int, str) for a mapped key expression with suffix - /// - /// This method is a coroutine. - /// - /// :param key_expr: The key expression the Queryable will reply to - /// :type key_expr: :class:`KeyExpr` - /// :param callback: the queryable callback (must be a coroutine) - /// :type callback: async function(:class:`Query`) - /// :param \*kwargs: - /// See below - /// - /// :Keyword Arguments: - /// kind ( int ) -- - /// Set the queryable kind. This must be a mask of constants defined in :mod:`zenoh.queryable`) - /// (`queryable.EVAL` by default) - /// * complete ( bool ) -- - /// Set the queryable completeness (true by default) - /// - /// :rtype: :class:`Queryable` - /// :raise: :class:`ZError` - /// - /// :Examples: - /// - /// >>> import asyncio, zenoh - /// >>> from zenoh import Sample, queryable - /// >>> - /// >>> async def callback(query): - /// ... print("Received : {}".format(query)) - /// ... query.reply(Sample('/key/expression', bytes('value', encoding='utf8'))) - /// >>> - /// >>> async def main(): - /// >>> s = await zenoh.async_open() - /// >>> q = await s.queryable('/key/expression', queryable.EVAL, callback) - /// >>> await asycio.sleep(60) - /// >>> - /// >>> asyncio.run(main()) - #[pyo3(text_signature = "(self, key_expr, callback, kwargs)")] - #[args(kwargs = "")] - fn queryable<'p>( - &self, - key_expr: &PyAny, - callback: &PyAny, - kwargs: Option<&PyDict>, - py: Python<'p>, - ) -> PyResult<&'p PyAny> { - let s = self.try_clone()?; - let k = zkey_expr_of_pyany(key_expr)?.to_owned(); - // Note: callback cannot be passed as such in task below because it's not Send - let cb_obj: Py<PyAny> = callback.into(); - // note: extract from kwargs here because it's not Send and cannot be moved into future_into_py(py, F) - let mut kind: Option<ZInt> = None; - let mut complete: Option<bool> = None; - if let Some(kwargs) = kwargs { - if let Some(k) = kwargs.get_item("kind") { - kind = Some(k.extract()?); - } - if let Some(p) = kwargs.get_item("local") { - complete = Some(p.extract::<bool>()?); - } - } - - future_into_py(py, async move { - let mut builder = s.queryable(k); - if let Some(k) = kind { - builder = builder.kind(k); - } - if let Some(c) = complete { - builder = builder.complete(c); - } - let zn_quer = builder.await.map_err(to_pyerr)?; - // Note: workaround to allow moving of zn_quer into the task below. - // Otherwise, s is moved also, but can't because it doesn't have 'static lifetime. - let mut zn_quer = unsafe { - std::mem::transmute::< - zenoh::queryable::Queryable<'_>, - zenoh::queryable::Queryable<'static>, - >(zn_quer) - }; - - let (unregister_tx, unregister_rx) = bounded::<bool>(1); - // Note: This is done to ensure that even if the call-back into Python - // does any blocking call we do not incour the risk of blocking - // any of the task resolving futures. - let loop_handle = task::spawn_blocking(move \|\| { - Python::with_gil(\|py\| { - // Run a Python event loop in this task, to allow coroutines execution within the callback - match pyo3_asyncio::async_std::run(py, async move { - loop { - select!( - q = zn_quer.receiver().next().fuse() => { - // call the async callback and transform the resulting Python awaitable into a Rust future - let future = match Python::with_gil(\|py\| { - let cb_args = PyTuple::new(py, &[Query { q: async_std::sync::Arc::new(q.unwrap()) }]); - cb_obj.as_ref(py).call1(cb_args).and_then(pyo3_asyncio::async_std::into_future) - }) { - Ok(f) => f, - Err(e) => { warn!("Error calling async queryable callback: {}", e); continue } - }; - // await the future (by default callbacks are executed in sequence) - if let Err(e) = future.await { - warn!("Error suring axecution of async queryable callback: {}", e); - } - }, - _ = unregister_rx.recv().fuse() => { - if let Err(e) = zn_quer.close().await { - warn!("Error undeclaring queryable: {}", e); - } - return Ok(()) - } - ) - } - }) { - Ok(()) => warn!("Queryable loop running"), - Err(e) => warn!("Failed to start Queryable loop: {}", e), - } - }) - }); - Ok(AsyncQueryable { - unregister_tx, - loop_handle: Some(loop_handle), - }) - }) - } - - /// Query data from the matching queryables in the system. - /// - /// Replies are collected in a list. - /// - /// The selector parameter also accepts the following types that can be converted to a :class:`Selector`: - /// - /// * KeyExpr for a key expression with no value selector - /// * int for a key expression id with no value selector - /// * str for a litteral selector - /// - /// This method is a coroutine. - /// - /// :param selector: The selection of resources to query - /// :type selector: :class:`Selector` - /// :param \*kwargs: - /// See below - /// - /// :Keyword Arguments: - /// target (:class:`QueryTarget`) -- - /// Set the kind of queryables that should be target of this query - /// * consolidation (:class:`QueryConsolidation`) -- - /// Set the consolidation mode of the query - /// * local_routing ( bool ) -- - /// Enable or disable local routing - /// - /// :rtype: [:class:`Reply`] - /// :raise: :class:`ZError` - /// - /// :Examples: - /// - /// >>> import asyncio, zenoh - /// >>> - /// >>> async def main(): - /// >>> s = await zenoh.async_open() - /// >>> replies = await s.get('/key/selector?value_selector') - /// >>> for reply in replies: - /// ... print("Received : {}".format(reply.sample)) - /// >>> - /// >>> asyncio.run(main()) - #[pyo3(text_signature = "(self, selector, kwargs)")] - #[args(kwargs = "")] - fn get<'p>( - &self, - selector: &PyAny, - kwargs: Option<&PyDict>, - py: Python<'p>, - ) -> PyResult<&'p PyAny> { - let s = self.try_clone()?; - - let selector: Selector = match selector.get_type().name()? { - "KeyExpr" => { - let key_expr: PyRef<KeyExpr> = selector.extract()?; - key_expr.inner.clone().into() - } - "int" => { - let id: u64 = selector.extract()?; - ZKeyExpr::from(id).into() - } - "str" => { - let name: &str = selector.extract()?; - Selector::from(name) - } - x => { - return Err(PyErr::new::<exceptions::PyValueError, _>(format!( - "Cannot convert type '{}' to a zenoh Selector", - x - ))) - } - } - .to_owned(); - // note: extract from kwargs here because it's not Send and cannot be moved into future_into_py(py, F) - let mut target: Option<QueryTarget> = None; - let mut consolidation: Option<QueryConsolidation> = None; - let mut local_routing: Option<bool> = None; - if let Some(kwargs) = kwargs { - if let Some(arg) = kwargs.get_item("target") { - target = Some(arg.extract::<QueryTarget>()?); - } - if let Some(arg) = kwargs.get_item("consolidation") { - consolidation = Some(arg.extract::<QueryConsolidation>()?); - } - if let Some(arg) = kwargs.get_item("local_routing") { - local_routing = Some(arg.extract::<bool>()?); - } - } - - future_into_py(py, async move { - let mut getter = s.get(selector); - if let Some(t) = target { - getter = getter.target(t.t); - } - if let Some(c) = consolidation { - getter = getter.consolidation(c.c); - } - if let Some(lr) = local_routing { - getter = getter.local_routing(lr); - } - let mut reply_rcv = getter.await.map_err(to_pyerr)?; - let mut replies: Vec<Reply> = Vec::new(); - - while let Some(reply) = reply_rcv.next().await { - replies.push(Reply { r: reply }); - } - Ok(replies) - }) - } -} - -impl AsyncSession { - pub(crate) fn new(s: zenoh::Session) -> Self { - AsyncSession { - s: Some(s.into_arc()), - } - } - - #[inline] - fn try_clone(&self) -> PyResult<Arc<zenoh::Session>> { - self.s - .clone() - .ok_or_else(\|\| PyErr::new::<ZError, _>("zenoh session was closed")) - } - - #[inline] - fn try_take(&mut self) -> PyResult<Arc<zenoh::Session>> { - self.s - .take() - .ok_or_else(\|\| PyErr::new::<ZError, _>("zenoh session was closed")) - } -} diff --git a/src/async_types.rs b/src/async_types.rs deleted file mode 100644 index d9bec576..00000000 --- a/src/async_types.rs +++ /dev/null @@ -1,87 +0,0 @@ -// -// Copyright (c) 2017, 2022 ZettaScale Technology Inc. -// -// This program and the accompanying materials are made available under the -// terms of the Eclipse Public License 2.0 which is available at -// http://www.eclipse.org/legal/epl-2.0, or the Apache License, Version 2.0 -// which is available at https://www.apache.org/licenses/LICENSE-2.0. -// -// SPDX-License-Identifier: EPL-2.0 OR Apache-2.0 -// -// Contributors: -// ZettaScale Zenoh team, <[email protected]> -// -use crate::types::; -use async_std::channel::Sender; -use log::warn; -use pyo3::prelude::; -use pyo3_asyncio::async_std::future_into_py; - -/// A subscriber to be used with asyncio. -#[pyclass] -pub(crate) struct AsyncSubscriber { - pub(crate) unregister_tx: Sender<ZnSubOps>, - pub(crate) loop_handle: Option<async_std::task::JoinHandle<()>>, -} - -#[pymethods] -impl AsyncSubscriber { - /// Pull available data for a pull-mode subscriber. - /// - /// This method is a coroutine. - fn pull<'p>(&self, py: Python<'p>) -> PyResult<&'p PyAny> { - let s = self.unregister_tx.clone(); - future_into_py(py, async move { - if let Err(e) = s.send(ZnSubOps::Pull).await { - warn!("Error in Subscriber::pull() : {}", e); - } - Ok(()) - }) - } - - /// Close the subscriber. - /// - /// This method is a coroutine. - fn close<'p>(&mut self, py: Python<'p>) -> PyResult<&'p PyAny> { - if let Some(handle) = self.loop_handle.take() { - let s = self.unregister_tx.clone(); - future_into_py(py, async move { - if let Err(e) = s.send(ZnSubOps::Unregister).await { - warn!("Error in Subscriber::close() : {}", e); - } - handle.await; - Ok(()) - }) - } else { - future_into_py(py, async move { Ok(()) }) - } - } -} - -/// A queryable to be used with asyncio. -#[pyclass] -pub(crate) struct AsyncQueryable { - pub(crate) unregister_tx: Sender<bool>, - pub(crate) loop_handle: Option<async_std::task::JoinHandle<()>>, -} - -#[pymethods] -impl AsyncQueryable { - /// Close the queryable. - /// - /// This method is a coroutine. - fn close<'p>(&mut self, py: Python<'p>) -> PyResult<&'p PyAny> { - if let Some(handle) = self.loop_handle.take() { - let s = self.unregister_tx.clone(); - future_into_py(py, async move { - if let Err(e) = s.send(true).await { - warn!("Error in Queryable::close() : {}", e); - } - handle.await; - Ok(()) - }) - } else { - future_into_py(py, async move { Ok(()) }) - } - } -} diff --git a/src/closures.rs b/src/closures.rs new file mode 100644 index 00000000..43e9eda6 --- /dev/null +++ b/src/closures.rs @@ -0,0 +1,98 @@ +// +// Copyright (c) 2017, 2022 ZettaScale Technology Inc. +// +// This program and the accompanying materials are made available under the +// terms of the Eclipse Public License 2.0 which is available at +// http://www.eclipse.org/legal/epl-2.0, or the Apache License, Version 2.0 +// which is available at https://www.apache.org/licenses/LICENSE-2.0. +// +// SPDX-License-Identifier: EPL-2.0 OR Apache-2.0 +// +// Contributors: +// ZettaScale Zenoh team, <[email protected]> +// +use std::{convert::TryFrom, sync::Arc}; + +use pyo3::{prelude::, types::PyTuple}; +use zenoh::prelude::IntoCallbackReceiverPair; + +trait CallbackUnwrap { + type Output; + fn cb_unwrap(self) -> Self::Output; +} +impl<T> CallbackUnwrap for PyResult<T> { + type Output = T; + fn cb_unwrap(self) -> Self::Output { + match self { + Ok(o) => o, + Err(e) => Python::with_gil(\|py\| { + if let Some(trace) = e.traceback(py).and_then(\|trace\| trace.format().ok()) { + panic!("Exception thrown in callback: {}.\n{}", e, trace) + } else { + panic!("Exception thrown in callback: {}.", e,) + } + }), + } + } +} + +pub(crate) struct PyClosure<I> { + pub(crate) pycall: Py<PyAny>, + pub(crate) drop: Option<Py<PyAny>>, + _marker: std::marker::PhantomData<I>, +} +impl<I> TryFrom<&PyAny> for PyClosure<I> { + type Error = PyErr; + fn try_from(value: &PyAny) -> Result<Self, Self::Error> { + Python::with_gil(\|py\| { + let pycall = match value.getattr("call") { + Ok(value) => value.into_py(py), + Err(e) => return Err(e), + }; + let drop = match value.getattr("drop") { + Ok(value) => { + if value.is_none() { + None + } else { + Some(value.into_py(py)) + } + } + Err(e) => return Err(e), + }; + Ok(PyClosure { + pycall, + drop, + _marker: std::marker::PhantomData, + }) + }) + } +} +impl<I: IntoPy<Py<PyTuple>>> PyClosure<I> { + pub fn call(&self, args: I) -> PyResult<PyObject> { + Python::with_gil(\|py\| self.pycall.call1(py, args)) + } +} +impl<I> Drop for PyClosure<I> { + fn drop(&mut self) { + if let Some(drop) = self.drop.take() { + Python::with_gil(\|py\| drop.call0(py)).unwrap(); + } + } +} +impl<T, I> IntoCallbackReceiverPair<'static, T> for PyClosure<(I,)> +where + T: Into<I>, + I: Send + Sync + 'static, + (I,): IntoPy<Py<PyTuple>>, +{ + type Receiver = (); + + fn into_cb_receiver_pair(self) -> (zenoh::handlers::Callback<'static, T>, Self::Receiver) { + ( + Arc::new(move \|reply\| { + self.call((reply.into(),)).cb_unwrap(); + }), + (), + ) + } +} diff --git a/src/config.rs b/src/config.rs new file mode 100644 index 00000000..7d8f0e8d --- /dev/null +++ b/src/config.rs @@ -0,0 +1,69 @@ +// +// Copyright (c) 2017, 2022 ZettaScale Technology Inc. +// +// This program and the accompanying materials are made available under the +// terms of the Eclipse Public License 2.0 which is available at +// http://www.eclipse.org/legal/epl-2.0, or the Apache License, Version 2.0 +// which is available at https://www.apache.org/licenses/LICENSE-2.0. +// +// SPDX-License-Identifier: EPL-2.0 OR Apache-2.0 +// +// Contributors: +// ZettaScale Zenoh team, <[email protected]> +// + +#![allow(clippy::borrow_deref_ref)] // false positives with pyo3 macros + +use pyo3::prelude::; +use validated_struct::ValidatedMap; +use zenoh::config::Config; +use zenoh_core::zerror; + +use crate::{ToPyErr, ToPyResult}; + +#[pyclass(subclass)] +pub struct _Config(pub(crate) Option<Config>); + +#[pymethods] +impl _Config { + #[allow(clippy::new_without_default)] + #[new] + pub fn new() -> Self { + _Config(Some(Default::default())) + } + #[staticmethod] + pub fn from_file(expr: &str) -> PyResult<Self> { + match Config::from_file(expr) { + Ok(k) => Ok(Self(Some(k))), + Err(e) => Err(e.to_pyerr()), + } + } + #[staticmethod] + pub fn from_json5(expr: &str) -> PyResult<Self> { + match Config::from_deserializer(&mut json5::Deserializer::from_str(expr).to_pyres()?) { + Ok(k) => Ok(Self(Some(k))), + Err(Ok(_)) => Err(zenoh_core::zerror!( + "{} did parse into a config, but invalid values were found", + expr, + ) + .to_pyerr()), + Err(Err(e)) => Err(e.to_pyerr()), + } + } + + pub fn get_json(&self, path: &str) -> PyResult<String> { + self.0 + .as_ref() + .ok_or_else(\|\| zerror!("Attempted to use a moved configuration").to_pyerr())? + .get_json(path) + .to_pyres() + } + + pub fn insert_json5(&mut self, path: &str, value: &str) -> PyResult<()> { + self.0 + .as_mut() + .ok_or_else(\|\| zerror!("Attempted to use a moved configuration").to_pyerr())? + .insert_json5(path, value) + .to_pyres() + } +} diff --git a/src/encoding.rs b/src/encoding.rs deleted file mode 100644 index 53d58542..00000000 --- a/src/encoding.rs +++ /dev/null @@ -1,359 +0,0 @@ -// -// Copyright (c) 2017, 2022 ZettaScale Technology Inc. -// -// This program and the accompanying materials are made available under the -// terms of the Eclipse Public License 2.0 which is available at -// http://www.eclipse.org/legal/epl-2.0, or the Apache License, Version 2.0 -// which is available at https://www.apache.org/licenses/LICENSE-2.0. -// -// SPDX-License-Identifier: EPL-2.0 OR Apache-2.0 -// -// Contributors: -// ZettaScale Zenoh team, <[email protected]> -// -use pyo3::prelude::; -use pyo3::PyObjectProtocol; -use zenoh::prelude::{Encoding as ZEncoding, KnownEncoding as ZKnownEncoding}; - -/// An encoding known by zenoh and which maps to an integer for wire-efficiency. -#[pyclass] -#[derive(Clone)] -pub struct KnownEncoding { - inner: ZKnownEncoding, -} - -#[allow(non_snake_case)] -#[pymethods] -impl KnownEncoding { - #[classattr] - fn Empty() -> Self { - KnownEncoding { - inner: ZKnownEncoding::Empty, - } - } - - #[classattr] - fn AppOctetStream() -> Self { - KnownEncoding { - inner: ZKnownEncoding::AppOctetStream, - } - } - - #[classattr] - fn AppCustom() -> Self { - KnownEncoding { - inner: ZKnownEncoding::AppCustom, - } - } - - #[classattr] - fn TextPlain() -> Self { - KnownEncoding { - inner: ZKnownEncoding::TextPlain, - } - } - - #[classattr] - fn AppProperties() -> Self { - KnownEncoding { - inner: ZKnownEncoding::AppProperties, - } - } - - #[classattr] - fn AppJson() -> Self { - KnownEncoding { - inner: ZKnownEncoding::AppJson, - } - } - - #[classattr] - fn AppSql() -> Self { - KnownEncoding { - inner: ZKnownEncoding::AppSql, - } - } - - #[classattr] - fn AppInteger() -> Self { - KnownEncoding { - inner: ZKnownEncoding::AppInteger, - } - } - - #[classattr] - fn AppFloat() -> Self { - KnownEncoding { - inner: ZKnownEncoding::AppFloat, - } - } - - #[classattr] - fn AppXml() -> Self { - KnownEncoding { - inner: ZKnownEncoding::AppXml, - } - } - - #[classattr] - fn AppXhtmlXml() -> Self { - KnownEncoding { - inner: ZKnownEncoding::AppXhtmlXml, - } - } - - #[classattr] - fn AppXWwwFormUrlencoded() -> Self { - KnownEncoding { - inner: ZKnownEncoding::AppXWwwFormUrlencoded, - } - } - - #[classattr] - fn TextJson() -> Self { - KnownEncoding { - inner: ZKnownEncoding::TextJson, - } - } - - #[classattr] - fn TextHtml() -> Self { - KnownEncoding { - inner: ZKnownEncoding::TextHtml, - } - } - - #[classattr] - fn TextXml() -> Self { - KnownEncoding { - inner: ZKnownEncoding::TextXml, - } - } - - #[classattr] - fn TextCss() -> Self { - KnownEncoding { - inner: ZKnownEncoding::TextCss, - } - } - - #[classattr] - fn TextCsv() -> Self { - KnownEncoding { - inner: ZKnownEncoding::TextCsv, - } - } - - #[classattr] - fn TextJavascript() -> Self { - KnownEncoding { - inner: ZKnownEncoding::TextJavascript, - } - } - - #[classattr] - fn ImageJpeg() -> Self { - KnownEncoding { - inner: ZKnownEncoding::ImageJpeg, - } - } - - #[classattr] - fn ImagePng() -> Self { - KnownEncoding { - inner: ZKnownEncoding::ImagePng, - } - } - - #[classattr] - fn ImageGif() -> Self { - KnownEncoding { - inner: ZKnownEncoding::ImageGif, - } - } -} -impl From<ZKnownEncoding> for KnownEncoding { - fn from(e: ZKnownEncoding) -> Self { - KnownEncoding { inner: e } - } -} -impl From<KnownEncoding> for ZKnownEncoding { - fn from(e: KnownEncoding) -> Self { - e.inner - } -} - -// zenoh.encoding (simulate the package as a class, and consts as class attributes) -/// A zenoh encoding is a HTTP Mime type represented, for wire efficiency, -/// as an integer prefix (that maps to a string) and a string suffix. -#[allow(non_camel_case_types)] -#[pyclass] -#[derive(Debug, Default, Clone, PartialEq)] -pub struct Encoding { - pub(crate) e: ZEncoding, -} -impl From<ZEncoding> for Encoding { - fn from(e: ZEncoding) -> Self { - Encoding { e } - } -} -impl From<Encoding> for ZEncoding { - fn from(e: Encoding) -> Self { - e.e - } -} - -#[allow(non_snake_case)] -#[pymethods] -impl Encoding { - #[new] - fn new(prefix: KnownEncoding, suffix: Option<String>) -> Self { - if let Some(suffix) = suffix { - Encoding { - e: ZEncoding::WithSuffix(prefix.inner, suffix.into()), - } - } else { - Encoding { - e: ZEncoding::Exact(prefix.inner), - } - } - } - - /// the Encoding's prefix. - /// - /// :type: :class:`KnownEncoding` - #[getter] - fn prefix(&self) -> PyResult<KnownEncoding> { - Ok((self.e.prefix()).into()) - } - - /// the Encoding's suffix. - /// - /// :type: str - #[getter] - fn get_suffix(&self) -> PyResult<&str> { - Ok(self.e.suffix()) - } - - /// Returns a copy of this Encoding, but changing its suffix. - /// - /// :param suffix: The new suffix - /// :type suffix: str - /// :rtype: :class:`Encoding` - fn with_suffix(&self, suffix: String) -> Self { - self.e.clone().with_suffix(suffix).into() - } - - #[classattr] - pub fn EMPTY() -> Self { - ZEncoding::EMPTY.into() - } - - #[classattr] - pub fn APP_OCTET_STREAM() -> Self { - ZEncoding::APP_OCTET_STREAM.into() - } - - #[classattr] - pub fn APP_CUSTOM() -> Self { - ZEncoding::APP_CUSTOM.into() - } - - #[classattr] - pub fn TEXT_PLAIN() -> Self { - ZEncoding::TEXT_PLAIN.into() - } - - #[classattr] - pub fn STRING() -> Self { - ZEncoding::STRING.into() - } - - #[classattr] - pub fn APP_PROPERTIES() -> Self { - ZEncoding::APP_PROPERTIES.into() - } - #[classattr] - pub fn APP_JSON() -> Self { - ZEncoding::APP_JSON.into() - } - #[classattr] - pub fn APP_SQL() -> Self { - ZEncoding::APP_SQL.into() - } - #[classattr] - pub fn APP_INTEGER() -> Self { - ZEncoding::APP_INTEGER.into() - } - #[classattr] - pub fn APP_FLOAT() -> Self { - ZEncoding::APP_FLOAT.into() - } - #[classattr] - pub fn APP_XML() -> Self { - ZEncoding::APP_XML.into() - } - #[classattr] - pub fn APP_XHTML_XML() -> Self { - ZEncoding::APP_XHTML_XML.into() - } - #[classattr] - pub fn APP_X_WWW_FORM_URLENCODED() -> Self { - ZEncoding::APP_X_WWW_FORM_URLENCODED.into() - } - #[classattr] - pub fn TEXT_JSON() -> Self { - ZEncoding::TEXT_JSON.into() - } - #[classattr] - pub fn TEXT_HTML() -> Self { - ZEncoding::TEXT_HTML.into() - } - #[classattr] - pub fn TEXT_XML() -> Self { - ZEncoding::TEXT_XML.into() - } - #[classattr] - pub fn TEXT_CSS() -> Self { - ZEncoding::TEXT_CSS.into() - } - #[classattr] - pub fn TEXT_CSV() -> Self { - ZEncoding::TEXT_CSV.into() - } - #[classattr] - pub fn TEXT_JAVASCRIPT() -> Self { - ZEncoding::TEXT_JAVASCRIPT.into() - } - #[classattr] - pub fn IMG_JPG() -> Self { - ZEncoding::IMG_JPG.into() - } - #[classattr] - pub fn IMG_PNG() -> Self { - ZEncoding::IMG_PNG.into() - } - #[classattr] - pub fn IMG_GIF() -> Self { - ZEncoding::IMG_GIF.into() - } - #[classattr] - pub fn DEFAULT() -> Self { - Self::default() - } - #[staticmethod] - fn from_str(s: String) -> Self { - ZEncoding::from(s).into() - } -} -#[pyproto] -impl PyObjectProtocol for Encoding { - fn __str__(&self) -> PyResult<String> { - Ok(self.e.to_string()) - } -} -impl std::fmt::Display for Encoding { - fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result { - write!(f, "{}", self.e) - } -} diff --git a/src/enums.rs b/src/enums.rs new file mode 100644 index 00000000..3209d696 --- /dev/null +++ b/src/enums.rs @@ -0,0 +1,272 @@ +// +// Copyright (c) 2017, 2022 ZettaScale Technology Inc. +// +// This program and the accompanying materials are made available under the +// terms of the Eclipse Public License 2.0 which is available at +// http://www.eclipse.org/legal/epl-2.0, or the Apache License, Version 2.0 +// which is available at https://www.apache.org/licenses/LICENSE-2.0. +// +// SPDX-License-Identifier: EPL-2.0 OR Apache-2.0 +// +// Contributors: +// ZettaScale Zenoh team, <[email protected]> +// +use crate::ToPyErr; +use pyo3::prelude::; +use zenoh::prelude::{Encoding, KnownEncoding, Priority, SampleKind}; +use zenoh::publication::CongestionControl; +use zenoh::query::{ConsolidationMode, QueryTarget}; +use zenoh::subscriber::Reliability; + +#[macro_export] +macro_rules! derive_richcmp { + ($tyname: expr) => { + fn __richcmp__(&self, other: &Self, op: pyo3::pyclass::CompareOp) -> PyResult<bool> { + match op { + pyo3::pyclass::CompareOp::Eq => Ok(self == other), + pyo3::pyclass::CompareOp::Ne => Ok(self != other), + _ => Err(zenoh_core::zerror!("{} does not support comparison", $tyname).to_pyerr()), + } + } + }; + () => { + fn __richcmp__(&self, other: &Self, op: pyo3::pyclass::CompareOp) -> bool { + match op { + pyo3::pyclass::CompareOp::Lt => self < other, + pyo3::pyclass::CompareOp::Le => self <= other, + pyo3::pyclass::CompareOp::Eq => self == other, + pyo3::pyclass::CompareOp::Ne => self != other, + pyo3::pyclass::CompareOp::Gt => self > other, + pyo3::pyclass::CompareOp::Ge => self >= other, + } + } + }; +} + +#[pyclass(subclass)] +#[repr(transparent)] +#[derive(Clone, Debug, PartialEq, Eq)] +pub struct _Encoding(pub(crate) Encoding); +#[pymethods] +impl _Encoding { + #[new] + pub fn new(this: Self) -> Self { + this + } + derive_richcmp!("Encoding"); + #[classattr] + pub const EMPTY: Self = Self(Encoding::Exact(KnownEncoding::Empty)); + #[classattr] + pub const APP_OCTET_STREAM: Self = Self(Encoding::Exact(KnownEncoding::AppOctetStream)); + #[classattr] + pub const APP_CUSTOM: Self = Self(Encoding::Exact(KnownEncoding::AppCustom)); + #[classattr] + pub const TEXT_PLAIN: Self = Self(Encoding::Exact(KnownEncoding::TextPlain)); + #[classattr] + pub const APP_PROPERTIES: Self = Self(Encoding::Exact(KnownEncoding::AppProperties)); + #[classattr] + pub const APP_JSON: Self = Self(Encoding::Exact(KnownEncoding::AppJson)); + #[classattr] + pub const APP_SQL: Self = Self(Encoding::Exact(KnownEncoding::AppSql)); + #[classattr] + pub const APP_INTEGER: Self = Self(Encoding::Exact(KnownEncoding::AppInteger)); + #[classattr] + pub const APP_FLOAT: Self = Self(Encoding::Exact(KnownEncoding::AppFloat)); + #[classattr] + pub const APP_XML: Self = Self(Encoding::Exact(KnownEncoding::AppXml)); + #[classattr] + pub const APP_XHTML_XML: Self = Self(Encoding::Exact(KnownEncoding::AppXhtmlXml)); + #[classattr] + pub const APP_X_WWW_FORM_URLENCODED: Self = + Self(Encoding::Exact(KnownEncoding::AppXWwwFormUrlencoded)); + #[classattr] + pub const TEXT_JSON: Self = Self(Encoding::Exact(KnownEncoding::TextJson)); + #[classattr] + pub const TEXT_HTML: Self = Self(Encoding::Exact(KnownEncoding::TextHtml)); + #[classattr] + pub const TEXT_XML: Self = Self(Encoding::Exact(KnownEncoding::TextXml)); + #[classattr] + pub const TEXT_CSS: Self = Self(Encoding::Exact(KnownEncoding::TextCss)); + #[classattr] + pub const TEXT_CSV: Self = Self(Encoding::Exact(KnownEncoding::TextCsv)); + #[classattr] + pub const TEXT_JAVASCRIPT: Self = Self(Encoding::Exact(KnownEncoding::TextJavascript)); + #[classattr] + pub const IMAGE_JPEG: Self = Self(Encoding::Exact(KnownEncoding::ImageJpeg)); + #[classattr] + pub const IMAGE_PNG: Self = Self(Encoding::Exact(KnownEncoding::ImagePng)); + #[classattr] + pub const IMAGE_GIF: Self = Self(Encoding::Exact(KnownEncoding::ImageGif)); + #[staticmethod] + pub fn from_str(s: String) -> Self { + Self(s.into()) + } + pub fn __str__(&self) -> String { + self.0.to_string() + } + pub fn append(&mut self, suffix: String) { + unsafe { + let mut tmp = std::ptr::read(&self.0); + tmp = tmp.with_suffix(suffix); + std::ptr::write(&mut self.0, tmp); + } + } + pub fn equals(&self, other: &Self) -> bool { + self == other + } +} + +#[pyclass(subclass)] +#[derive(Clone, PartialEq, Eq)] +pub struct _Priority(pub(crate) Priority); +#[pymethods] +impl _Priority { + #[new] + pub fn new(this: Self) -> Self { + this + } + derive_richcmp!(); + #[classattr] + pub const REAL_TIME: Self = Self(Priority::RealTime); + #[classattr] + pub const INTERACTIVE_HIGH: Self = Self(Priority::InteractiveHigh); + #[classattr] + pub const INTERACTIVE_LOW: Self = Self(Priority::InteractiveLow); + #[classattr] + pub const DATA_HIGH: Self = Self(Priority::DataHigh); + #[classattr] + pub const DATA: Self = Self(Priority::Data); + #[classattr] + pub const DATA_LOW: Self = Self(Priority::DataLow); + #[classattr] + pub const BACKGROUND: Self = Self(Priority::Background); + pub fn __str__(&self) -> &'static str { + match self.0 { + Priority::RealTime => "REAL_TIME", + Priority::InteractiveHigh => "INTERACTIVE_HIGH", + Priority::InteractiveLow => "INTERACTIVE_LOW", + Priority::DataHigh => "DATA_HIGH", + Priority::Data => "DATA", + Priority::DataLow => "DATA_LOW", + Priority::Background => "BACKGROUND", + } + } +} +impl std::cmp::PartialOrd for _Priority { + fn partial_cmp(&self, other: &Self) -> Option<std::cmp::Ordering> { + (self.0 as u8).partial_cmp(&(other.0 as u8)) + } +} + +#[pyclass(subclass)] +#[derive(Clone, PartialEq, Eq)] +pub struct _SampleKind(pub(crate) SampleKind); +#[pymethods] +impl _SampleKind { + #[new] + pub fn new(this: Self) -> Self { + this + } + derive_richcmp!("SampleKind"); + #[classattr] + pub const PUT: Self = Self(SampleKind::Put); + #[classattr] + pub const DELETE: Self = Self(SampleKind::Delete); + pub fn __str__(&self) -> &'static str { + match self.0 { + SampleKind::Put => "PUT", + SampleKind::Delete => "DELETE", + } + } +} + +#[pyclass(subclass)] +#[derive(Clone, PartialEq, Eq)] +pub struct _CongestionControl(pub(crate) CongestionControl); +#[pymethods] +impl _CongestionControl { + #[new] + pub fn new(this: Self) -> Self { + this + } + derive_richcmp!("CongestionControl"); + #[classattr] + pub const BLOCK: Self = Self(CongestionControl::Block); + #[classattr] + pub const DROP: Self = Self(CongestionControl::Drop); + pub fn __str__(&self) -> &'static str { + match self.0 { + CongestionControl::Block => "BLOCK", + CongestionControl::Drop => "DROP", + } + } +} + +#[pyclass(subclass)] +#[derive(Clone, PartialEq, Eq)] +pub struct _Reliability(pub(crate) Reliability); +#[pymethods] +impl _Reliability { + #[new] + pub fn new(this: Self) -> Self { + this + } + derive_richcmp!("Reliability"); + #[classattr] + pub const BEST_EFFORT: Self = Self(Reliability::BestEffort); + #[classattr] + pub const RELIABLE: Self = Self(Reliability::Reliable); + pub fn __str__(&self) -> &'static str { + match self.0 { + Reliability::BestEffort => "BEST_EFFORT", + Reliability::Reliable => "RELIABLE", + } + } +} + +#[pyclass(subclass)] +#[derive(Clone, PartialEq, Eq)] +pub struct _QueryTarget(pub(crate) QueryTarget); +#[pymethods] +impl _QueryTarget { + #[new] + pub fn new(this: Self) -> Self { + this + } + derive_richcmp!("QueryTarget"); + #[classattr] + pub const BEST_MATCHING: Self = Self(QueryTarget::BestMatching); + #[classattr] + pub const ALL: Self = Self(QueryTarget::All); + #[classattr] + pub const ALL_COMPLETE: Self = Self(QueryTarget::AllComplete); + pub fn __str__(&self) -> &'static str { + match self.0 { + QueryTarget::BestMatching => "BEST_MATCHING", + QueryTarget::All => "ALL", + QueryTarget::AllComplete => "ALL_COMPLETE", + #[cfg(feature = "complete_n")] + QueryTarget::Complete(_) => "COMPLETE_N", + } + } +} + +#[pyclass(subclass)] +#[derive(Clone, PartialEq, Eq)] +pub struct _QueryConsolidation(pub(crate) Option<ConsolidationMode>); +#[pymethods] +impl _QueryConsolidation { + #[new] + pub fn new(this: Self) -> Self { + this + } + derive_richcmp!("QueryConsolidation"); + #[classattr] + pub const AUTO: Self = Self(None); + #[classattr] + pub const NONE: Self = Self(Some(ConsolidationMode::None)); + #[classattr] + pub const MONOTONIC: Self = Self(Some(ConsolidationMode::Monotonic)); + #[classattr] + pub const LATEST: Self = Self(Some(ConsolidationMode::Latest)); +} diff --git a/src/keyexpr.rs b/src/keyexpr.rs new file mode 100644 index 00000000..a53932c4 --- /dev/null +++ b/src/keyexpr.rs @@ -0,0 +1,119 @@ +// +// Copyright (c) 2017, 2022 ZettaScale Technology Inc. +// +// This program and the accompanying materials are made available under the +// terms of the Eclipse Public License 2.0 which is available at +// http://www.eclipse.org/legal/epl-2.0, or the Apache License, Version 2.0 +// which is available at https://www.apache.org/licenses/LICENSE-2.0. +// +// SPDX-License-Identifier: EPL-2.0 OR Apache-2.0 +// +// Contributors: +// ZettaScale Zenoh team, <[email protected]> +// + +use pyo3::prelude::; +use std::{ + collections::{hash_map::DefaultHasher, HashMap}, + convert::{TryFrom, TryInto}, +}; +use zenoh::prelude::{sync::SyncResolve, KeyExpr, Selector}; + +use crate::{session::_Session, ToPyErr}; + +#[pyclass(subclass)] +#[derive(Clone)] +pub struct _KeyExpr(pub(crate) KeyExpr<'static>); + +#[pymethods] +impl _KeyExpr { + #[new] + pub fn pynew(this: Self) -> Self { + this + } + #[staticmethod] + pub fn new(expr: String) -> PyResult<Self> { + match expr.try_into() { + Ok(k) => Ok(Self(k)), + Err(e) => Err(e.to_pyerr()), + } + } + #[staticmethod] + pub fn autocanonize(expr: String) -> PyResult<Self> { + match KeyExpr::autocanonize(expr) { + Ok(k) => Ok(Self(k)), + Err(e) => Err(e.to_pyerr()), + } + } + + pub fn intersects(&self, other: &Self) -> bool { + self.0.intersects(&other.0) + } + + pub fn includes(&self, other: &Self) -> bool { + self.0.includes(&other.0) + } + + pub fn equals(&self, other: &Self) -> bool { + self.0 == other.0 + } + + pub fn undeclare(&self, session: &_Session) -> PyResult<()> { + session + .0 + .undeclare(self.0.clone()) + .res_sync() + .map_err(\|e\| e.to_pyerr()) + } + + pub fn __str__(&self) -> &str { + self.0.as_str() + } + + pub fn __hash__(&self) -> isize { + use std::hash::; + let mut hasher: DefaultHasher = BuildHasherDefault::default().build_hasher(); + self.0.hash(&mut hasher); + hasher.finish() as isize + } + + pub fn __eq__(&self, other: &Self) -> bool { + self.0 == other.0 + } +} + +#[pyclass(subclass)] +#[derive(Clone)] +pub struct _Selector(pub(crate) Selector<'static>); +#[pymethods] +impl _Selector { + #[new] + pub fn pynew(this: Self) -> Self { + this + } + #[staticmethod] + pub fn new(expr: String) -> PyResult<Self> { + match Selector::try_from(expr) { + Ok(o) => Ok(_Selector(o)), + Err(e) => Err(e.to_pyerr()), + } + } + #[getter] + pub fn key_expr(&self) -> _KeyExpr { + _KeyExpr(self.0.key_expr.clone()) + } + #[getter] + pub fn get_parameters(&self) -> &str { + self.0.parameters() + } + #[setter] + pub fn set_parameters(&mut self, parameters: String) { + self.0.set_parameters(parameters) + } + pub fn decode_parameters(&self) -> PyResult<HashMap<String, String>> { + self.0.parameters_map().map_err(\|e\| e.to_pyerr()) + } + pub fn __str__(&self) -> String { + self.0.to_string() + } +} diff --git a/src/lib.rs b/src/lib.rs index 7625fde2..d9abd5d7 100644 --- a/src/lib.rs +++ b/src/lib.rs @@ -11,399 +11,105 @@ // Contributors: // ZettaScale Zenoh team, <[email protected]> // -use async_std::prelude::FutureExt; -use async_std::task; -use encoding::KnownEncoding; -use futures::prelude::; -use pyo3::prelude::; -use pyo3::{create_exception, wrap_pyfunction}; -use pyo3_asyncio::async_std::future_into_py; -use zenoh::config::{Config as ZConfig, Notifier}; -use zenoh::prelude::ValidatedMap; -use zenoh_core::zerror; - -pub(crate) mod types; -pub(crate) use types::; +use pyo3::{prelude::, types::PyDict, ToPyObject}; +mod closures; +mod config; +mod enums; +mod keyexpr; +mod queryable; mod session; -use session::; -mod async_types; -mod encoding; -mod sample_kind; -use async_types::; -mod async_session; -use async_session::; +mod value; -create_exception!(zenoh, ZError, pyo3::exceptions::PyException); +pyo3::create_exception!(zenoh, ZError, pyo3::exceptions::PyException); -fn to_pyerr(err: zenoh_core::Error) -> PyErr { - PyErr::new::<ZError, _>(err.to_string()) -} -/// The zenoh API. -/// -/// Examples of use: -/// ^^^^^^^^^^^^^^^^ -/// -/// Publish -/// """"""" -/// -/// >>> import zenoh -/// >>> s = zenoh.open() -/// >>> s.put('/resource/name', bytes('value', 'utf8')) -/// -/// Subscribe -/// """"""""" -/// -/// >>> import zenoh -/// >>> def listener(sample): -/// ... print("Received: {} = {}".format(sample.key_expr, sample.payload.decode("utf-8"))) -/// >>> -/// >>> s = zenoh.open() -/// >>> sub = s.subscribe('/resource/', listener) -/// -/// Get -/// """ -/// -/// >>> import zenoh -/// >>> s = zenoh.open() -/// >>> for reply in s.get('/resource/name'): -/// ... print("Received: {} = {}".format(reply.sample.key_expr, reply.sample.payload.decode('utf-8')) -#[pymodule] -pub fn zenoh(py: Python, m: &PyModule) -> PyResult<()> { - m.add_class::<config>()?; - // force addition of "zenoh.config" module - // (see https://github.com/PyO3/pyo3/issues/759#issuecomment-653964601) - py.run( - "\ -import sys -sys.modules['zenoh.config'] = config - ", - None, - Some(m.dict()), - )?; - - m.add_class::<info>()?; - // force addition of "zenoh.info" module - // (see https://github.com/PyO3/pyo3/issues/759#issuecomment-653964601) - py.run( - "\ -import sys -sys.modules['zenoh.info'] = info - ", - None, - Some(m.dict()), - )?; - - m.add_class::<queryable>()?; - // force addition of "zenoh.queryable" module - // (see https://github.com/PyO3/pyo3/issues/759#issuecomment-653964601) - py.run( - "\ -import sys -sys.modules['zenoh.queryable'] = queryable - ", - None, - Some(m.dict()), - )?; - - m.add_class::<AsyncQueryable>()?; - m.add_class::<AsyncSession>()?; - m.add_class::<AsyncSubscriber>()?; - m.add_class::<Config>()?; - m.add_class::<CongestionControl>()?; - m.add_class::<ConsolidationMode>()?; - m.add_class::<encoding::Encoding>()?; - m.add_class::<Hello>()?; - m.add_class::<KeyExpr>()?; - m.add_class::<KnownEncoding>()?; - m.add_class::<PeerId>()?; - m.add_class::<Period>()?; - m.add_class::<Priority>()?; - m.add_class::<Query>()?; - m.add_class::<Queryable>()?; - m.add_class::<QueryConsolidation>()?; - m.add_class::<QueryTarget>()?; - m.add_class::<Reliability>()?; - m.add_class::<Reply>()?; - m.add_class::<Sample>()?; - m.add_class::<sample_kind::SampleKind>()?; - m.add_class::<Selector>()?; - m.add_class::<Session>()?; - m.add_class::<SourceInfo>()?; - m.add_class::<SubMode>()?; - m.add_class::<Subscriber>()?; - m.add_class::<Target>()?; - m.add_class::<Timestamp>()?; - m.add_class::<Value>()?; - m.add_class::<ValueSelector>()?; - m.add_class::<WhatAmI>()?; - m.add("ZError", py.get_type::<ZError>())?; - m.add_wrapped(wrap_pyfunction!(init_logger))?; - m.add_wrapped(wrap_pyfunction!(config_from_file))?; - m.add_wrapped(wrap_pyfunction!(open))?; - m.add_wrapped(wrap_pyfunction!(async_open))?; - m.add_wrapped(wrap_pyfunction!(scout))?; - m.add_wrapped(wrap_pyfunction!(async_scout))?; - Ok(()) -} -/// Initialize the logger used by the Rust implementation of this API. -/// -/// Once initialized, you can configure the logs displayed by the API using the ``RUST_LOG`` environment variable. -/// For instance, start python with the debug* logs available:: -/// -/// $ RUST_LOG=debug python -/// -/// More details on the RUST_LOG configuration on https://docs.rs/env_logger/latest/env_logger -/// -#[pyfunction] -fn init_logger() { - env_logger::init(); +pub(crate) trait ToPyErr { + fn to_pyerr(self) -> PyErr; } - -/// Parse a configuration file for zenoh, returning a Config object. -/// -/// :param path: The path to the config file. -/// :rtype: :class:`Config` -#[pyfunction] -fn config_from_file(path: &str) -> PyResult<Config> { - Config::from_file(path) +impl<E: std::error::Error> ToPyErr for E { + fn to_pyerr(self) -> PyErr { + PyErr::new::<ZError, _>(self.to_string()) + } } - -/// The main configuration structure for Zenoh. -/// -/// To construct a configuration, we advise that you use a configuration file -/// (JSON, JSON5 and YAML are currently supported, please use the proper extension for your format as the deserializer will be picked according to it). -/// A Config object can then be amended calling :func:`Config.insert_json5`. -/// -/// :Example: -/// -/// >>> import zenoh, json -/// >>> conf = zenoh.Config.from_file('zenoh-config.json5') -/// >>> conf.insert_json5(zenoh.config.MODE_KEY, json.dumps('client')) -/// >>> print(conf.json()) -#[pyclass] -#[derive(Debug, Clone)] -pub struct Config { - inner: ZConfig, +pub(crate) trait ToPyResult<T> { + fn to_pyres(self) -> Result<T, PyErr>; } -#[pymethods] -impl Config { - /// Constructor of a default configuration. - #[new] - pub fn new() -> Self { - Config { - inner: ZConfig::default(), - } - } - - /// Parse a configuration file for zenoh, returning a Config object. - /// - /// :param path: The path to the config file. - /// :rtype: :class:`Config` - /// :raise: :class:`ZError` - pub fn insert_json5(&mut self, key: &str, value: &str) -> PyResult<()> { - match self.inner.insert_json5(key, value) { - Ok(()) => Ok(()), - Err(e) => Err(PyErr::new::<pyo3::exceptions::PyException, _>( - e.to_string(), - )), - } - } - - /// Returns the config as a JSON string - /// - /// :rtype: str - pub fn json(&self) -> String { - serde_json::to_string(&self.inner).unwrap() - } - - /// Parse a JSON5 string configuration for zenoh, returning a Config object. - /// - /// :param input: The configuration as a JSON5 string. - /// :rtype: :class:`Config` - /// :raise: :class:`ZError` - #[staticmethod] - pub fn from_json5(input: &str) -> PyResult<Self> { - let mut d = match json5::Deserializer::from_str(input) { - Ok(d) => d, - Err(e) => return Err(to_pyerr(zerror!(e).into())), - }; - match ZConfig::from_deserializer(&mut d) { - Ok(inner) => Ok(Config { inner }), - Err(e) => Err(to_pyerr( - match e { - Ok(c) => zerror!("invalid configuration: {:?}", c), - Err(e) => zerror!(e), - } - .into(), - )), - } - } - - /// Parse a configuration file for zenoh, returning a Config object. - /// - /// :param path: The path to the config file. - /// :rtype: :class:`Config` - /// :raise: :class:`ZError` - #[staticmethod] - pub fn from_file(path: &str) -> PyResult<Self> { - match ZConfig::from_file(path) { - Ok(inner) => Ok(Config { inner }), - Err(e) => Err(to_pyerr(e)), - } +impl<T, E: ToPyErr> ToPyResult<T> for Result<T, E> { + fn to_pyres(self) -> Result<T, PyErr> { + self.map_err(ToPyErr::to_pyerr) } } -impl Default for Config { - fn default() -> Self { - Self::new() +enum ExtractError { + Unavailable(Option<PyErr>), + Other(PyErr), +} +impl From<PyErr> for ExtractError { + fn from(e: PyErr) -> Self { + Self::Other(e) } } - -#[pyclass] -#[derive(Clone)] -pub struct ConfigNotifier { - inner: Notifier<ZConfig>, +pub(crate) trait PyExtract<K> { + fn extract_item<'a, V: FromPyObject<'a>>(&'a self, key: K) -> Result<V, ExtractError>; } -#[pymethods] -impl ConfigNotifier { - pub fn insert_json5(&mut self, key: &str, value: &str) -> PyResult<()> { - match self.inner.insert_json5(key, value) { - Ok(()) => Ok(()), - Err(e) => Err(PyErr::new::<pyo3::exceptions::PyException, _>( - e.to_string(), - )), +impl<K: ToPyObject> PyExtract<K> for PyAny { + fn extract_item<'a, V: FromPyObject<'a>>(&'a self, key: K) -> Result<V, ExtractError> { + match self.get_item(key) { + Ok(item) => Ok(item.extract::<V>()?), + Err(e) => Err(ExtractError::Unavailable(Some(e))), } } - pub fn json(&self) -> String { - serde_json::to_string(&self.inner.lock()).unwrap() - } } - -/// Open a zenoh Session. -/// -/// :param config: The configuration of the zenoh session -/// :type config: :class:`Config`, optional -/// :rtype: :class:`Session` -/// :raise: :class:`ZError` -/// -/// :Example: -/// -/// >>> import zenoh -/// >>> z = zenoh.open(zenoh.config::peer()) -#[pyfunction] -#[pyo3(text_signature = "(config)")] -fn open(config: Option<Config>) -> PyResult<Session> { - let s = task::block_on(zenoh::open(config.unwrap_or_default().inner)).map_err(to_pyerr)?; - Ok(Session::new(s)) +impl<K: ToPyObject> PyExtract<K> for PyDict { + fn extract_item<'a, V: FromPyObject<'a>>(&'a self, key: K) -> Result<V, ExtractError> { + match self.get_item(key) { + Some(item) => Ok(item.extract::<V>()?), + None => Err(ExtractError::Unavailable(None)), + } + } } -/// Coroutine to open a zenoh AsyncSession (similar to a Session, but for asyncio usage). -/// -/// :param config: The configuration of the zenoh session -/// :type config: :class:`Config`, optional -/// :rtype: :class:`AsyncSession` -/// :raise: :class:`ZError` -/// -/// :Example: -/// -/// >>> import asyncio, zenoh -/// >>> async def main(): -/// >>> z = await zenoh.async_open() -/// >>> -/// >>> asyncio.run(main()) -#[pyfunction] -#[pyo3(text_signature = "(config)")] -fn async_open(py: Python, config: Option<Config>) -> PyResult<&PyAny> { - future_into_py(py, async { - let s = zenoh::open(config.unwrap_or_default().inner) - .await - .map_err(to_pyerr)?; - Ok(AsyncSession::new(s)) - }) +#[pymodule] +fn zenoh(_py: Python, m: &PyModule) -> PyResult<()> { + m.add_class::<config::_Config>()?; + m.add_class::<keyexpr::_KeyExpr>()?; + m.add_class::<keyexpr::_Selector>()?; + m.add_class::<session::_Session>()?; + m.add_class::<session::_Publisher>()?; + m.add_class::<session::_Subscriber>()?; + m.add_class::<session::_PullSubscriber>()?; + m.add_class::<session::_Scout>()?; + m.add_class::<queryable::_Query>()?; + m.add_class::<queryable::_Queryable>()?; + m.add_class::<value::_Value>()?; + m.add_class::<value::_Sample>()?; + m.add_class::<value::_Reply>()?; + m.add_class::<value::_Timestamp>()?; + m.add_class::<value::_Hello>()?; + m.add_class::<value::_ZenohId>()?; + m.add_class::<enums::_CongestionControl>()?; + m.add_class::<enums::_Encoding>()?; + m.add_class::<enums::_Priority>()?; + m.add_class::<enums::_SampleKind>()?; + m.add_class::<enums::_Reliability>()?; + m.add_class::<enums::_QueryConsolidation>()?; + m.add_class::<enums::_QueryTarget>()?; + m.add_wrapped(wrap_pyfunction!(init_logger))?; + m.add_wrapped(wrap_pyfunction!(session::scout))?; + Ok(()) } -/// Scout for routers and/or peers. +/// Initialize the logger used by the Rust implementation of this API. /// -/// Sends scout messages for a specified duration and returns -/// a list of received :class:`Hello` messages. +/// Once initialized, you can configure the logs displayed by the API using the ``RUST_LOG`` environment variable. +/// For instance, start python with the debug* logs available:: /// -/// :param whatami: The kind of zenoh process to scout for -/// :type whatami: int -/// :param scout_duration: the duration of scout (in seconds) -/// :type scout_duration: float -/// :param config: The configuration to use for scouting -/// :type config: :class:`Config`, optional -/// :rtype: list of :class:`Hello` -/// :raise: :class:`ZError` +/// $ RUST_LOG=debug python /// -/// :Example: +/// More details on the RUST_LOG configuration on https://docs.rs/env_logger/latest/env_logger /// -/// >>> import zenoh -/// >>> hellos = zenoh.scout(WhatAmI.Peer \| WhatAmI.Router, 1.0) -/// >>> for hello in hellos: -/// >>> print(hello) #[pyfunction] -#[pyo3(text_signature = "(whatami, scout_duration, config)")] -fn scout(whatami: WhatAmI, scout_duration: f64, config: Option<Config>) -> PyResult<Vec<Hello>> { - task::block_on(async move { - let mut result = Vec::<Hello>::new(); - let mut receiver = zenoh::scout(whatami, config.unwrap_or_default().inner) - .await - .unwrap(); - let scout = async { - while let Some(h) = receiver.next().await { - result.push(Hello { h }) - } - }; - let timeout = async_std::task::sleep(std::time::Duration::from_secs_f64(scout_duration)); - FutureExt::race(scout, timeout).await; - Ok(result) - }) +fn init_logger() { + let _ = env_logger::try_init(); } -/// Coroutine to scout for routers and/or peers. -/// -/// Sends scout messages for a specified duration and returns -/// a list of received :class:`Hello` messages. -/// -/// :param whatami: The kind of zenoh process to scout for -/// :type whatami: int -/// :param scout_duration: the duration of scout (in seconds) -/// :type scout_duration: float -/// :param config: The configuration to use for scouting -/// :type config: :class:`Config`, optional -/// :rtype: list of :class:`Hello` -/// :raise: :class:`ZError` -/// -/// :Example: -/// -/// >>> import asyncio, zenoh -/// >>> async def main(): -/// >>> hellos = await zenoh.async_scout(WhatAmI.Peer \| WhatAmI.Router, 1.0) -/// >>> for hello in hellos: -/// >>> print(hello) -/// >>> -/// >>> asyncio.run(main()) -#[pyfunction] -#[pyo3(text_signature = "(whatami, scout_duration, config)")] -fn async_scout( - whatami: WhatAmI, - scout_duration: f64, - config: Option<Config>, - py: Python, -) -> PyResult<&PyAny> { - future_into_py(py, async move { - let mut result = Vec::<Hello>::new(); - let mut receiver = zenoh::scout(whatami, config.unwrap_or_default().inner) - .await - .unwrap(); - let scout = async { - while let Some(h) = receiver.next().await { - result.push(Hello { h }) - } - }; - let timeout = async_std::task::sleep(std::time::Duration::from_secs_f64(scout_duration)); - FutureExt::race(scout, timeout).await; - Ok(result) - }) -} +pub(crate) use value::PyAnyToValue; diff --git a/src/queryable.rs b/src/queryable.rs new file mode 100644 index 00000000..c0f6aa28 --- /dev/null +++ b/src/queryable.rs @@ -0,0 +1,84 @@ +// +// Copyright (c) 2017, 2022 ZettaScale Technology Inc. +// +// This program and the accompanying materials are made available under the +// terms of the Eclipse Public License 2.0 which is available at +// http://www.eclipse.org/legal/epl-2.0, or the Apache License, Version 2.0 +// which is available at https://www.apache.org/licenses/LICENSE-2.0. +// +// SPDX-License-Identifier: EPL-2.0 OR Apache-2.0 +// +// Contributors: +// ZettaScale Zenoh team, <[email protected]> +// +use std::{collections::HashMap, sync::Arc}; + +use pyo3::prelude::; +use zenoh::{ + queryable::{Query, Queryable}, + selector::Parameters, +}; +use zenoh_core::SyncResolve; + +use crate::{ + keyexpr::{_KeyExpr, _Selector}, + value::_Sample, + ToPyErr, +}; + +#[pyclass(subclass)] +#[derive(Clone)] +pub struct _Query(pub(crate) Arc<Query>); +#[pymethods] +impl _Query { + #[new] + pub fn pynew(this: Self) -> Self { + this + } + #[getter] + pub fn key_expr(&self) -> _KeyExpr { + _KeyExpr(self.0.key_expr().clone()) + } + #[getter] + pub fn parameters(&self) -> &str { + self.0.parameters() + } + pub fn decode_parameters(&self) -> PyResult<HashMap<String, String>> { + let mut res = HashMap::new(); + for (k, v) in self.0.parameters().decode() { + let k = k.into_owned(); + match res.entry(k) { + std::collections::hash_map::Entry::Occupied(e) => { + return Err(zenoh_core::zerror!( + "Detected duplicate key {} in value selector {}", + e.key(), + self.0.parameters() + ) + .to_pyerr()) + } + std::collections::hash_map::Entry::Vacant(e) => { + e.insert(v.into_owned()); + } + } + } + Ok(res) + } + #[getter] + pub fn selector(&self) -> _Selector { + _Selector(self.0.selector().clone().into_owned()) + } + pub fn reply(&self, sample: _Sample) -> PyResult<()> { + self.0 + .reply(Ok(sample.into())) + .res_sync() + .map_err(\|e\| e.to_pyerr()) + } +} +impl From<Query> for _Query { + fn from(q: Query) -> Self { + Self(Arc::new(q)) + } +} + +#[pyclass(subclass)] +pub struct _Queryable(pub(crate) Queryable<'static, ()>); diff --git a/src/sample_kind.rs b/src/sample_kind.rs deleted file mode 100644 index fd3dcdb7..00000000 --- a/src/sample_kind.rs +++ /dev/null @@ -1,81 +0,0 @@ -// -// Copyright (c) 2017, 2022 ZettaScale Technology Inc. -// -// This program and the accompanying materials are made available under the -// terms of the Eclipse Public License 2.0 which is available at -// http://www.eclipse.org/legal/epl-2.0, or the Apache License, Version 2.0 -// which is available at https://www.apache.org/licenses/LICENSE-2.0. -// -// SPDX-License-Identifier: EPL-2.0 OR Apache-2.0 -// -// Contributors: -// ZettaScale Zenoh team, <[email protected]> -// -use pyo3::class::basic::CompareOp; -use pyo3::prelude::; -use pyo3::PyObjectProtocol; -use zenoh::prelude::SampleKind as ZSampleKind; - -// zenoh.sample_kind (simulate the package as a class, and consts as class attributes) -/// Constants defining the different data kinds. -#[allow(non_camel_case_types)] -#[pyclass] -#[derive(Debug, Clone, PartialEq, Default)] -pub struct SampleKind { - pub kind: ZSampleKind, -} -impl From<ZSampleKind> for SampleKind { - fn from(kind: ZSampleKind) -> Self { - SampleKind { kind } - } -} -impl From<SampleKind> for ZSampleKind { - fn from(kind: SampleKind) -> Self { - kind.kind - } -} - -#[allow(non_snake_case)] -#[pymethods] -impl SampleKind { - #[classattr] - pub fn PUT() -> Self { - ZSampleKind::Put.into() - } - - #[classattr] - pub fn PATCH() -> Self { - ZSampleKind::Patch.into() - } - - #[classattr] - pub fn DELETE() -> Self { - ZSampleKind::Delete.into() - } - - #[classattr] - pub fn DEFAULT() -> Self { - ZSampleKind::default().into() - } -} - -#[pyproto] -impl PyObjectProtocol for SampleKind { - fn __str__(&self) -> PyResult<String> { - Ok(self.to_string()) - } - - fn __richcmp__(&'p self, other: PyRef<'p, SampleKind>, op: CompareOp) -> PyResult<PyObject> { - match op { - CompareOp::Eq => Ok(self.eq(&other).into_py(other.py())), - CompareOp::Ne => Ok((!self.eq(&other)).into_py(other.py())), - _ => Ok(other.py().NotImplemented()), - } - } -} - -impl std::fmt::Display for SampleKind { - fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result { - write!(f, "{}", self.kind) - } -} diff --git a/src/session.rs b/src/session.rs index 049d032c..6c9ecacf 100644 --- a/src/session.rs +++ b/src/session.rs @@ -1,5 +1,3 @@ -use crate::ConfigNotifier; - // // Copyright (c) 2017, 2022 ZettaScale Technology Inc. // @@ -13,637 +11,294 @@ use crate::ConfigNotifier; // Contributors: // ZettaScale Zenoh team, <[email protected]> // -use super::encoding::Encoding; -use super::sample_kind::SampleKind; -use super::types::{ - zkey_expr_of_pyany, zvalue_of_pyany, CongestionControl, KeyExpr, Period, Priority, Query, - QueryConsolidation, QueryTarget, Queryable, Reliability, Reply, Sample, SubMode, Subscriber, - ZnSubOps, -}; -use super::{to_pyerr, ZError}; -use async_std::channel::bounded; -use async_std::task; -use futures::prelude::; -use futures::select; -use log::warn; -use pyo3::exceptions; -use pyo3::prelude::; -use pyo3::types::{IntoPyDict, PyDict, PyList, PyTuple}; -use std::collections::HashMap; -use zenoh::prelude::{ExprId, KeyExpr as ZKeyExpr, Selector, ZFuture, ZInt}; -/// A zenoh session. -#[pyclass] -pub struct Session { - s: Option<zenoh::Session>, -} +#![allow(clippy::borrow_deref_ref)] // false positives with pyo3 macros -#[pymethods] -impl Session { - /// Close the zenoh Session. - pub fn close(&mut self) -> PyResult<()> { - let s = self.take()?; - s.close().wait().map_err(to_pyerr) - } +use std::convert::TryInto; +use std::sync::Arc; - /// Get informations about the zenoh Session. - /// - /// :rtype: dict[str, str] - /// - /// :Example: - /// - /// >>> import zenoh - /// >>> s = zenoh.open() - /// >>> info = s.info() - /// >>> for key in info: - /// >>> print("{} : {}".format(key, info[key])) - pub fn info(&self, py: Python) -> PyResult<PyObject> { - use zenoh_cfg_properties::KeyTranscoder; - let s = self.as_ref()?; - let props = s.info().wait(); - let pydict: HashMap<String, String> = props - .0 - .into_iter() - .filter_map(\|(k, v)\| zenoh::info::InfoTranscoder::decode(k).map(\|k\| (k, v))) - .collect(); - Ok(pydict.into_py_dict(py).to_object(py)) - } +use pyo3::{prelude::, types::PyDict}; +use zenoh::config::whatami::{WhatAmI, WhatAmIMatcher}; +use zenoh::prelude::SessionDeclarations; +use zenoh::publication::Publisher; +use zenoh::scouting::Scout; +use zenoh::subscriber::{PullSubscriber, Subscriber}; +use zenoh::Session; +use zenoh_core::SyncResolve; - /// Get informations about the zenoh Session. - /// - /// :rtype: dict {str: str} - /// - /// :Example: - /// - /// >>> import zenoh - /// >>> s = zenoh.open() - /// >>> info = s.info() - /// >>> for key in info: - /// >>> print("{} : {}".format(key, info[key])) - /// - /// - /// Get the current configuration of the zenoh Session. - /// - /// The returned ConfigNotifier can be used to read the current - /// zenoh configuration through the json function or - /// modify the zenoh configuration through the insert_json5 funtion. - /// - /// :rtype: dict {str: str} - /// - /// :Example: - /// - /// >>> import zenoh - /// >>> s = zenoh.open() - /// >>> config = s.config() - /// >>> config.insert_json5("connect/endpoints", "[\"tcp/10.10.10.10:7448\"]") - /// - pub fn config(&self) -> PyResult<ConfigNotifier> { - Ok(ConfigNotifier { - inner: self.as_ref()?.config().clone(), - }) - } +use crate::closures::PyClosure; +use crate::config::_Config; +use crate::enums::{ + _CongestionControl, _Priority, _QueryConsolidation, _QueryTarget, _Reliability, _SampleKind, +}; +use crate::keyexpr::{_KeyExpr, _Selector}; +use crate::queryable::{_Query, _Queryable}; +use crate::value::{_Hello, _Reply, _Sample, _Value, _ZenohId}; +use crate::{PyAnyToValue, PyExtract, ToPyErr}; - /// Put data. - /// - /// The key_expr* parameter also accepts the following types that can be converted to a :class:`KeyExpr`: - /// - /// * int for a mapped key expression - /// * str for a literal key expression - /// * (int, str) for a mapped key expression with suffix - /// - /// :param key_expr: The key expression matching resources to write - /// :type key_expr: :class:`KeyExpr` - /// :param value: The value to write - /// :type value: any type convertible to a :class:`Value` - /// :param \*kwargs: - /// See below - /// - /// :Keyword Arguments: - /// encoding (:class:`Encoding`) -- - /// Set the encoding of the written data - /// * kind ( int ) -- - /// Set the kind of the written data - /// * congestion_control (:class:`CongestionControl`) -- - /// Set the congestion control to apply when routing the data - /// * priority (:class:`Priority`) -- - /// Set the priority of the written data - /// * local_routing ( bool ) -- - /// Enable or disable local routing - /// - /// :raise: :class:`ZError` - /// - /// :Examples: - /// - /// >>> import zenoh - /// >>> s = zenoh.open() - /// >>> s.put('/key/expression', 'value') - #[pyo3(text_signature = "(self, key_expr, value, kwargs)")] +#[pyclass(subclass)] +#[derive(Clone)] +pub struct _Session(pub(crate) Arc<Session>); + +#[pymethods] +impl _Session { + #[new] + pub fn new(config: Option<&mut crate::config::_Config>) -> PyResult<Self> { + let config = config.and_then(\|c\| c.0.take()).unwrap_or_default(); + let session = zenoh::open(config).res_sync().map_err(\|e\| e.to_pyerr())?; + Ok(_Session(Arc::new(session))) + } #[args(kwargs = "")] - pub fn put(&self, key_expr: &PyAny, value: &PyAny, kwargs: Option<&PyDict>) -> PyResult<()> { - let s = self.as_ref()?; - let k = zkey_expr_of_pyany(key_expr)?; - let mut v = zvalue_of_pyany(value)?; - let mut encoding: Option<Encoding> = None; - let mut kind: Option<SampleKind> = None; - let mut congestion_control: Option<CongestionControl> = None; - let mut priority: Option<Priority> = None; - let mut local_routing: Option<bool> = None; + pub fn put( + &self, + key_expr: &crate::keyexpr::_KeyExpr, + value: &PyAny, + kwargs: Option<&PyDict>, + ) -> PyResult<()> { + let s = &self.0; + let k = &key_expr.0; + let v = value.to_value()?; + let mut builder = s.put(k, v); if let Some(kwargs) = kwargs { - if let Some(e) = kwargs.get_item("encoding") { - encoding = e.extract().ok() - } - if let Some(k) = kwargs.get_item("kind") { - kind = k.extract().ok() - } - if let Some(cc) = kwargs.get_item("congestion_control") { - congestion_control = cc.extract().ok() + match kwargs.extract_item::<_SampleKind>("kind") { + Ok(kind) => builder = builder.kind(kind.0), + Err(crate::ExtractError::Other(e)) => return Err(e), + _ => {} } - if let Some(p) = kwargs.get_item("priority") { - priority = p.extract().ok() + match kwargs.extract_item::<_CongestionControl>("congestion_control") { + Ok(congestion_control) => { + builder = builder.congestion_control(congestion_control.0) + } + Err(crate::ExtractError::Other(e)) => return Err(e), + _ => {} } - if let Some(lr) = kwargs.get_item("local_routing") { - local_routing = lr.extract().ok() + match kwargs.extract_item::<_Priority>("priority") { + Ok(priority) => builder = builder.priority(priority.0), + Err(crate::ExtractError::Other(e)) => return Err(e), + _ => {} } } - if let Some(encoding) = encoding { - v.encoding = encoding.into(); - } - let mut writer = s - .put(k, v) - .kind(kind.unwrap_or_default().kind) - .congestion_control(congestion_control.unwrap_or_default().cc) - .priority(priority.unwrap_or_default().p); - if let Some(local_routing) = local_routing { - writer = writer.local_routing(local_routing); - } - writer.wait().map_err(to_pyerr) + builder.res_sync().map_err(\|e\| e.to_pyerr()) } - - /// Delete data. - /// - /// The key_expr parameter also accepts the following types that can be converted to a :class:`KeyExpr`: - /// - /// * int for a mapped key expression - /// * str for a literal key expression - /// * (int, str) for a mapped key expression with suffix - /// - /// :param key_expr: The key expression matching resources to delete - /// :type key_expr: :class:`KeyExpr` - /// :param \*kwargs: - /// See below - /// - /// :Keyword Arguments: - /// congestion_control (:class:`CongestionControl`) -- - /// Set the congestion control to apply when routing the data - /// * priority (:class:`Priority`) -- - /// Set the priority of the written data - /// * local_routing ( bool ) -- - /// Enable or disable local routing - /// - /// :raise: :class:`ZError` - /// - /// :Examples: - /// - /// >>> import zenoh - /// >>> s = zenoh.open() - /// >>> s.delete('/key/expression') - #[pyo3(text_signature = "(self, key_expr, kwargs)")] #[args(kwargs = "")] - pub fn delete(&self, key_expr: &PyAny, kwargs: Option<&PyDict>) -> PyResult<()> { - let s = self.as_ref()?; - let k = zkey_expr_of_pyany(key_expr)?; - let mut congestion_control: Option<CongestionControl> = None; - let mut priority: Option<Priority> = None; - let mut local_routing: Option<bool> = None; + pub fn delete( + &self, + key_expr: &crate::keyexpr::_KeyExpr, + kwargs: Option<&PyDict>, + ) -> PyResult<()> { + let s = &self.0; + let k = &key_expr.0; + let mut builder = s.delete(k); if let Some(kwargs) = kwargs { - if let Some(cc) = kwargs.get_item("congestion_control") { - congestion_control = cc.extract().ok() + match kwargs.extract_item::<_SampleKind>("kind") { + Ok(kind) => builder = builder.kind(kind.0), + Err(crate::ExtractError::Other(e)) => return Err(e), + _ => {} } - if let Some(p) = kwargs.get_item("priority") { - priority = p.extract().ok() + match kwargs.extract_item::<_CongestionControl>("congestion_control") { + Ok(congestion_control) => { + builder = builder.congestion_control(congestion_control.0) + } + Err(crate::ExtractError::Other(e)) => return Err(e), + _ => {} } - if let Some(lr) = kwargs.get_item("local_routing") { - local_routing = lr.extract().ok() + match kwargs.extract_item::<_Priority>("priority") { + Ok(priority) => builder = builder.priority(priority.0), + Err(crate::ExtractError::Other(e)) => return Err(e), + _ => {} } } - let mut writer = s - .delete(k) - .congestion_control(congestion_control.unwrap_or_default().cc) - .priority(priority.unwrap_or_default().p); - if let Some(local_routing) = local_routing { - writer = writer.local_routing(local_routing); - } - writer.wait().map_err(to_pyerr) - } - - /// Associate a numerical Id with the given key expression. - /// - /// This numerical Id will be used on the network to save bandwidth and - /// ease the retrieval of the concerned resource in the routing tables. - /// - /// The key_expr parameter also accepts the following types that can be converted to a :class:`KeyExpr`: - /// - /// * int for a mapped key expression - /// * str for a literal key expression - /// * (int, str) for a mapped key expression with suffix - /// - /// :param key_expr: The key expression to map to a numerical Id - /// :type key_expr: :class:`KeyExpr` - /// :rtype: int - /// :raise: :class:`ZError` - /// - /// :Examples: - /// - /// >>> import zenoh - /// >>> s = zenoh.open() - /// >>> rid = s.declare_expr('/key/expression') - #[pyo3(text_signature = "(self, key_expr)")] - pub fn declare_expr(&self, key_expr: &PyAny) -> PyResult<ExprId> { - let s = self.as_ref()?; - let k = zkey_expr_of_pyany(key_expr)?; - s.declare_expr(&k).wait().map_err(to_pyerr) + builder.res_sync().map_err(\|e\| e.to_pyerr()) } - /// Undeclare the numerical Id/key expression association previously declared - /// with :meth:`declare_expr`. - /// - /// :param rid: The numerical Id to unmap - /// :type rid: :class:`ExprId` - /// :raise: :class:`ZError` - /// - /// :Examples: - /// - /// >>> import zenoh - /// >>> s = zenoh.open() - /// >>> rid = s.declare_expr('/key/expression') - /// >>> s.undeclare_expr(rid) - #[pyo3(text_signature = "(self, rid)")] - pub fn undeclare_expr(&self, rid: ExprId) -> PyResult<()> { - let s = self.as_ref()?; - s.undeclare_expr(rid).wait().map_err(to_pyerr) + #[args(kwargs = "*")] + pub fn get( + &self, + selector: &_Selector, + callback: &PyAny, + kwargs: Option<&PyDict>, + ) -> PyResult<()> { + let callback: PyClosure<(_Reply,)> = <_ as TryInto<_>>::try_into(callback)?; + let mut builder = self.0.get(&selector.0).with(callback); + if let Some(kwargs) = kwargs { + match kwargs.extract_item::<_QueryConsolidation>("consolidation") { + Ok(_QueryConsolidation(Some(value))) => builder = builder.consolidation(value), + Err(crate::ExtractError::Other(e)) => return Err(e), + _ => {} + } + match kwargs.extract_item::<_QueryTarget>("target") { + Ok(value) => builder = builder.target(value.0), + Err(crate::ExtractError::Other(e)) => return Err(e), + _ => {} + } + } + builder.res_sync().map_err(\|e\| e.to_pyerr()) } - /// Declare a publication for the given key expression. - /// - /// Written expressions that match the given key expression will only be sent on the network - /// if matching subscribers exist in the system. - /// - /// The key_expr* parameter also accepts the following types that can be converted to a :class:`KeyExpr`: - /// - /// * int for a mapped key expression - /// * str for a literal key expression - /// * (int, str) for a mapped key expression with suffix - /// - /// :param key_expr: The key expression to publish - /// :type key_expr: :class:`KeyExpr` - /// :raise: :class:`ZError` - /// - /// :Examples: - /// - /// >>> import zenoh - /// >>> s = zenoh.open() - /// >>> rid = s.declare_publication('/key/expression') - /// >>> s.put('/key/expression', bytes('value', encoding='utf8')) - #[pyo3(text_signature = "(self, key_expr)")] - fn declare_publication(&self, key_expr: &PyAny) -> PyResult<()> { - let s = self.as_ref()?; - let k = zkey_expr_of_pyany(key_expr)?; - s.declare_publication(&k).wait().map_err(to_pyerr)?; - Ok(()) - } - #[pyo3(text_signature = "(self, key_expr)")] - fn undeclare_publication(&self, key_expr: &PyAny) -> PyResult<()> { - let s = self.as_ref()?; - let k = zkey_expr_of_pyany(key_expr)?; - s.undeclare_publication(&k).wait().map_err(to_pyerr)?; - Ok(()) + pub fn declare_keyexpr(&self, key_expr: &_KeyExpr) -> PyResult<_KeyExpr> { + match self.0.declare_keyexpr(&key_expr.0).res_sync() { + Ok(k) => Ok(_KeyExpr(k.into_owned())), + Err(e) => Err(e.to_pyerr()), + } } - /// Create a Subscriber for the given key expression. - /// - /// The key_expr parameter also accepts the following types that can be converted to a :class:`KeyExpr`: - /// - /// * int for a mapped key expression - /// * str for a literal key expression - /// * (int, str) for a mapped key expression with suffix - /// - /// :param key_expr: The key expression to subscribe - /// :type key_expr: :class:`KeyExpr` - /// :param callback: the subscription callback - /// :type callback: function(:class:`Sample`) - /// :param \*kwargs: - /// See below - /// - /// :Keyword Arguments: - /// reliability (:class:`Reliability`) -- - /// Set the subscription reliability (BestEffort by default) - /// * mode (:class:`SubMode`) -- - /// Set the subscription mode (Push by default) - /// * period (:class:`Period`) -- - /// Set the subscription period - /// * local ( bool ) -- - /// If true make the subscription local only (false by default) - /// - /// :rtype: :class:`Subscriber` - /// :raise: :class:`ZError` - /// - /// :Examples: - /// - /// >>> import zenoh, time - /// >>> from zenoh import Reliability, SubMode - /// >>> - /// >>> s = zenoh.open() - /// >>> sub = s.subscribe('/key/expression', - /// ... lambda sample: print("Received : {}".format(sample)), - /// ... reliability=Reliability.Reliable, - /// ... mode=SubMode.Push) - /// >>> time.sleep(60) - #[pyo3(text_signature = "(self, key_expr, callback, kwargs)")] #[args(kwargs = "")] - fn subscribe( + pub fn declare_queryable( &self, - key_expr: &PyAny, + key_expr: _KeyExpr, callback: &PyAny, kwargs: Option<&PyDict>, - ) -> PyResult<Subscriber> { - let s = self.as_ref()?; - let k = zkey_expr_of_pyany(key_expr)?; - let mut sub_builder = s.subscribe(&k); + ) -> PyResult<_Queryable> { + let callback: PyClosure<(_Query,)> = <_ as TryInto<_>>::try_into(callback)?; + let mut builder = self.0.declare_queryable(key_expr.0).with(callback); if let Some(kwargs) = kwargs { - if let Some(arg) = kwargs.get_item("reliability") { - sub_builder = sub_builder.reliability(arg.extract::<Reliability>()?.r); + match kwargs.extract_item::<bool>("complete") { + Ok(value) => builder = builder.complete(value), + Err(crate::ExtractError::Other(e)) => return Err(e), + _ => {} } - if let Some(arg) = kwargs.get_item("mode") { - sub_builder = sub_builder.mode(arg.extract::<SubMode>()?.m); - } - if let Some(arg) = kwargs.get_item("period") { - sub_builder = sub_builder.period(Some(arg.extract::<Period>()?.p)); + } + match builder.res_sync() { + Ok(o) => Ok(_Queryable(o)), + Err(e) => Err(e.to_pyerr()), + } + } + + #[args(kwargs = "*")] + pub fn declare_publisher( + &self, + key_expr: _KeyExpr, + kwargs: Option<&PyDict>, + ) -> PyResult<_Publisher> { + let mut builder = self.0.declare_publisher(key_expr.0); + if let Some(kwargs) = kwargs { + match kwargs.extract_item::<_Priority>("priority") { + Ok(value) => builder = builder.priority(value.0), + Err(crate::ExtractError::Other(e)) => return Err(e), + _ => {} } - if let Some(arg) = kwargs.get_item("local") { - if arg.extract::<bool>()? { - sub_builder = sub_builder.local(); - } + match kwargs.extract_item::<_CongestionControl>("congestion_control") { + Ok(value) => builder = builder.congestion_control(value.0), + Err(crate::ExtractError::Other(e)) => return Err(e), + _ => {} } } - let zn_sub = sub_builder.wait().map_err(to_pyerr)?; - // Note: workaround to allow moving of zn_sub into the task below. - // Otherwise, s is moved also, but can't because it doesn't have 'static lifetime. - let mut static_zn_sub = unsafe { - std::mem::transmute::< - zenoh::subscriber::Subscriber<'_>, - zenoh::subscriber::Subscriber<'static>, - >(zn_sub) - }; - - // Note: callback cannot be passed as such in task below because it's not Send - let cb_obj: Py<PyAny> = callback.into(); - - let (unregister_tx, unregister_rx) = bounded::<ZnSubOps>(8); - // Note: This is done to ensure that even if the call-back into Python - // does any blocking call we do not incour the risk of blocking - // any of the task resolving futures. - let loop_handle = task::spawn_blocking(move \|\| { - task::block_on(async move { - loop { - select!( - s = static_zn_sub.receiver().next().fuse() => { - // Acquire Python GIL to call the callback - let gil = Python::acquire_gil(); - let py = gil.python(); - let cb_args = PyTuple::new(py, &[Sample { s: s.unwrap() }]); - if let Err(e) = cb_obj.as_ref(py).call1(cb_args) { - warn!("Error calling subscriber callback:"); - e.print(py); - } - }, - op = unregister_rx.recv().fuse() => { - match op { - Ok(ZnSubOps::Pull) => { - if let Err(e) = static_zn_sub.pull().await { - warn!("Error pulling the subscriber: {}", e); - } - }, - Ok(ZnSubOps::Unregister) => { - if let Err(e) = static_zn_sub.close().await { - warn!("Error undeclaring subscriber: {}", e); - } - return - }, - _ => return - } - } - ) - } - }) - }); - Ok(Subscriber { - unregister_tx, - loop_handle: Some(loop_handle), - }) + match builder.res_sync() { + Ok(o) => Ok(_Publisher(o)), + Err(e) => Err(e.to_pyerr()), + } } - /// Create a Queryable for the given key expression. - /// - /// The key_expr* parameter also accepts the following types that can be converted to a :class:`KeyExpr`: - /// - /// * int for a mapped key expression - /// * str for a literal key expression - /// * (int, str) for a mapped key expression with suffix - /// - /// :param key_expr: The key expression the Queryable will reply to - /// :type key_expr: :class:`KeyExpr` - /// :param callback: the queryable callback - /// :type callback: function(:class:`Query`) - /// :param \*kwargs: - /// See below - /// - /// :Keyword Arguments: - /// kind ( int ) -- - /// Set the queryable kind. This must be a mask of constants defined in :mod:`zenoh.queryable`) - /// (`queryable.EVAL` by default) - /// * complete ( bool ) -- - /// Set the queryable completeness (true by default) - /// - /// :rtype: :class:`Queryable` - /// :raise: :class:`ZError` - /// - /// :Examples: - /// - /// >>> import zenoh, time - /// >>> from zenoh import Sample, queryable - /// >>> def callback(query): - /// ... print("Received : {}".format(query)) - /// ... query.reply(Sample('/key/expression', bytes('value', encoding='utf8'))) - /// >>> - /// >>> s = zenoh.open() - /// >>> q = s.queryable('/key/expression', callback, kind=queryable.EVAL) - /// >>> time.sleep(60) - #[pyo3(text_signature = "(self, key_expr, callback, kwargs)")] #[args(kwargs = "")] - fn queryable( + pub fn declare_subscriber( &self, - key_expr: &PyAny, + key_expr: &_KeyExpr, callback: &PyAny, kwargs: Option<&PyDict>, - ) -> PyResult<Queryable> { - let s = self.as_ref()?; - let k = zkey_expr_of_pyany(key_expr)?; - let mut builder = s.queryable(k); + ) -> PyResult<_Subscriber> { + let callback: PyClosure<(_Sample,)> = <_ as TryInto<_>>::try_into(callback)?; + let mut builder = self.0.declare_subscriber(&key_expr.0).with(callback); if let Some(kwargs) = kwargs { - if let Some(arg) = kwargs.get_item("kind") { - builder = builder.kind(arg.extract::<ZInt>()?); - } - if let Some(arg) = kwargs.get_item("complete") { - builder = builder.complete(arg.extract::<bool>()?); + match kwargs.extract_item::<_Reliability>("reliability") { + Ok(reliabilty) => builder = builder.reliability(reliabilty.0), + Err(crate::ExtractError::Other(e)) => return Err(e), + _ => {} } } - let zn_quer = builder.wait().map_err(to_pyerr)?; - - // Note: workaround to allow moving of zn_quer into the task below. - // Otherwise, s is moved also, but can't because it doesn't have 'static lifetime. - let mut zn_quer = unsafe { - std::mem::transmute::< - zenoh::queryable::Queryable<'_>, - zenoh::queryable::Queryable<'static>, - >(zn_quer) - }; - - // Note: callback cannot be passed as such in task below because it's not Send - let cb_obj: Py<PyAny> = callback.into(); - - let (unregister_tx, unregister_rx) = bounded::<bool>(1); - // Note: This is done to ensure that even if the call-back into Python - // does any blocking call we do not incour the risk of blocking - // any of the task resolving futures. - let loop_handle = task::spawn_blocking(move \|\| { - task::block_on(async move { - loop { - select!( - q = zn_quer.receiver().next().fuse() => { - // Acquire Python GIL to call the callback - let gil = Python::acquire_gil(); - let py = gil.python(); - let cb_args = PyTuple::new(py, &[Query { q: async_std::sync::Arc::new(q.unwrap()) }]); - if let Err(e) = cb_obj.as_ref(py).call1(cb_args) { - warn!("Error calling queryable callback:"); - e.print(py); - } - }, - _ = unregister_rx.recv().fuse() => { - if let Err(e) = zn_quer.close().await { - warn!("Error undeclaring queryable: {}", e); - } - return - } - ) - } - }) - }); - Ok(Queryable { - unregister_tx, - loop_handle: Some(loop_handle), - }) + let subscriber = builder.res().map_err(\|e\| e.to_pyerr())?; + Ok(_Subscriber(subscriber)) } - /// Query data from the matching queryables in the system. - /// - /// Replies are collected in a list. - /// - /// The selector parameter also accepts the following types that can be converted to a :class:`Selector`: - /// - /// * KeyExpr for a key expression with no value selector - /// * int for a key expression id with no value selector - /// * str for a litteral selector - /// - /// :param selector: The selection of resources to query - /// :type selector: :class:`Selector` - /// :param \*kwargs: - /// See below - /// - /// :Keyword Arguments: - /// target (:class:`QueryTarget`) -- - /// Set the kind of queryables that should be target of this query - /// * consolidation (:class:`QueryConsolidation`) -- - /// Set the consolidation mode of the query - /// * local_routing ( bool ) -- - /// Enable or disable local routing - /// - /// :rtype: [:class:`Reply`] - /// :raise: :class:`ZError` - /// - /// :Examples: - /// - /// >>> import zenoh, time - /// >>> - /// >>> s = zenoh.open() - /// >>> replies = s.get('/key/selector?value_selector') - /// >>> for reply in replies: - /// ... print("Received : {}".format(reply.sample)) - #[pyo3(text_signature = "(self, selector, kwargs)")] #[args(kwargs = "")] - fn get(&self, selector: &PyAny, kwargs: Option<&PyDict>) -> PyResult<Py<PyList>> { - let s = self.as_ref()?; - task::block_on(async { - let selector: Selector = match selector.get_type().name()? { - "KeyExpr" => { - let key_expr: PyRef<KeyExpr> = selector.extract()?; - key_expr.inner.clone().into() - } - "int" => { - let id: u64 = selector.extract()?; - ZKeyExpr::from(id).into() - } - "str" => { - let name: &str = selector.extract()?; - Selector::from(name) - } - x => { - return Err(PyErr::new::<exceptions::PyValueError, _>(format!( - "Cannot convert type '{}' to a zenoh Selector", - x - ))) - } - }; - let mut getter = s.get(selector); - if let Some(kwargs) = kwargs { - if let Some(arg) = kwargs.get_item("target") { - getter = getter.target(arg.extract::<QueryTarget>()?.t); - } - if let Some(arg) = kwargs.get_item("consolidation") { - getter = getter.consolidation(arg.extract::<QueryConsolidation>()?.c); - } - if let Some(arg) = kwargs.get_item("local_routing") { - getter = getter.local_routing(arg.extract::<bool>()?); - } - } - let mut replies = getter.wait().map_err(to_pyerr)?; - let gil = Python::acquire_gil(); - let py = gil.python(); - let result = PyList::empty(py); - while let Some(reply) = replies.next().await { - result.append(Reply { r: reply })?; + pub fn declare_pull_subscriber( + &self, + key_expr: &_KeyExpr, + callback: &PyAny, + kwargs: Option<&PyDict>, + ) -> PyResult<_PullSubscriber> { + let callback: PyClosure<(_Sample,)> = <_ as TryInto<_>>::try_into(callback)?; + let mut builder = self + .0 + .declare_subscriber(&key_expr.0) + .pull_mode() + .with(callback); + if let Some(kwargs) = kwargs { + match kwargs.extract_item::<_Reliability>("reliability") { + Ok(reliabilty) => builder = builder.reliability(reliabilty.0), + Err(crate::ExtractError::Other(e)) => return Err(e), + _ => {} } - Ok(result.into()) - }) + } + let subscriber = builder.res().map_err(\|e\| e.to_pyerr())?; + Ok(_PullSubscriber(subscriber)) + } + + pub fn zid(&self) -> _ZenohId { + _ZenohId(self.0.zid()) + } + pub fn routers_zid(&self) -> Vec<_ZenohId> { + self.0 + .info() + .routers_zid() + .res_sync() + .map(_ZenohId) + .collect() + } + pub fn peers_zid(&self) -> Vec<_ZenohId> { + self.0.info().peers_zid().res_sync().map(_ZenohId).collect() } } -impl Session { - pub(crate) fn new(s: zenoh::Session) -> Self { - Session { s: Some(s) } +#[pyclass(subclass)] +#[derive(Clone)] +pub struct _Publisher(Publisher<'static>); +#[pymethods] +impl _Publisher { + #[new] + pub fn pynew(this: Self) -> Self { + this + } + #[getter] + pub fn key_expr(&self) -> _KeyExpr { + _KeyExpr(self.0.key_expr().clone()) + } + pub fn put(&self, value: _Value) -> PyResult<()> { + self.0.put(value).res_sync().map_err(\|e\| e.to_pyerr()) + } + pub fn delete(&self) -> PyResult<()> { + self.0.delete().res_sync().map_err(\|e\| e.to_pyerr()) } +} + +#[pyclass(subclass)] +pub struct _Subscriber(Subscriber<'static, ()>); - #[inline] - fn as_ref(&self) -> PyResult<&zenoh::Session> { - self.s - .as_ref() - .ok_or_else(\|\| PyErr::new::<ZError, _>("zenoh session was closed")) +#[pyclass(subclass)] +pub struct _PullSubscriber(PullSubscriber<'static, ()>); +#[pymethods] +impl _PullSubscriber { + fn pull(&self) -> PyResult<()> { + self.0.pull().res_sync().map_err(\|e\| e.to_pyerr()) } +} + +#[pyclass(subclass)] +pub struct _Scout(Scout<()>); - #[inline] - fn take(&mut self) -> PyResult<zenoh::Session> { - self.s - .take() - .ok_or_else(\|\| PyErr::new::<ZError, _>("zenoh session was closed")) +#[pyfunction] +pub fn scout(callback: &PyAny, config: Option<&_Config>, what: Option<&str>) -> PyResult<_Scout> { + let callback: PyClosure<(_Hello,)> = <_ as TryInto<_>>::try_into(callback)?; + let what: WhatAmIMatcher = match what { + None => WhatAmI::Client \| WhatAmI::Peer \| WhatAmI::Router, + Some(s) => match s.parse() { + Ok(w) => w, + Err(_) => return Err(zenoh_core::zerror!("Couldn't parse `{}` into a WhatAmiMatcher: must be a `\|`-separated list of `peer`, `client` or `router`", s).to_pyerr()) + }, + }; + let config = config.and_then(\|c\| c.0.clone()).unwrap_or_default(); + let scout = zenoh::scout(what, config).with(callback).res_sync(); + match scout { + Ok(scout) => Ok(_Scout(scout)), + Err(e) => Err(e.to_pyerr()), } } diff --git a/src/types.rs b/src/types.rs deleted file mode 100644 index 664d547c..00000000 --- a/src/types.rs +++ /dev/null @@ -1,1565 +0,0 @@ -use std::collections::HashMap; -use std::ops::BitOr; - -// -// Copyright (c) 2017, 2022 ZettaScale Technology Inc. -// -// This program and the accompanying materials are made available under the -// terms of the Eclipse Public License 2.0 which is available at -// http://www.eclipse.org/legal/epl-2.0, or the Apache License, Version 2.0 -// which is available at https://www.apache.org/licenses/LICENSE-2.0. -// -// SPDX-License-Identifier: EPL-2.0 OR Apache-2.0 -// -// Contributors: -// ZettaScale Zenoh team, <[email protected]> -// -use crate::encoding::Encoding; -use crate::sample_kind::SampleKind; -use crate::to_pyerr; -use async_std::channel::Sender; -use async_std::task; -use log::warn; -use pyo3::exceptions; -use pyo3::number::PyNumberOrProtocol; -use pyo3::prelude::; -use pyo3::types::{PyBytes, PyTuple}; -use pyo3::PyObjectProtocol; -use zenoh::config::whatami::WhatAmIMatcher; -use zenoh::config::WhatAmI as ZWhatAmI; -use zenoh::prelude::{ - Encoding as ZEncoding, KeyExpr as ZKeyExpr, KnownEncoding as ZKnownEncoding, - Selector as ZSelector, Value as ZValue, ZInt, -}; -use zenoh_buffers::traits::SplitBuffer; - -// zenoh.config (simulate the package as a class, and consts as class attributes) -// -/// The following constants define the several configuration keys accepted for a zenoh -/// session configuration and the associated accepted values. -#[allow(non_camel_case_types)] -#[pyclass] -pub(crate) struct config {} - -#[allow(non_snake_case)] -#[pymethods] -impl config { - /// The library mode. - /// - /// - Accepted values : `"peer"`, `"client"`. - /// - Default value : `"peer"`. - #[classattr] - pub fn MODE_KEY() -> &'static str { - "mode" - } - - /// The locator of a peer to connect to. - /// - /// - Accepted values : `<locator>` (ex: `"tcp/10.10.10.10:7447"`). - /// - Default value : None. - /// - Multiple values accepted. - #[classattr] - pub fn CONNECT_KEY() -> &'static str { - "connect/endpoints" - } - - /// A locator to listen on. - /// - /// - Accepted values : `<locator>` (ex: `"tcp/10.10.10.10:7447"`). - /// - Default value : None. - /// - Multiple values accepted. - #[classattr] - pub fn LISTEN_KEY() -> &'static str { - "listen/endpoints" - } - - /// The user name to use for authentication. - /// - /// - Accepted values : `<string>`. - /// - Default value : None. - #[classattr] - pub fn USER_KEY() -> &'static str { - "transport/auth/usrpwd/user" - } - - /// The password to use for authentication. - /// - /// - Accepted values : `<string>`. - /// - Default value : None. - #[classattr] - fn PASSWORD_KEY() -> &'static str { - "transport/auth/usrpwd/password" - } - - /// Activates/Desactivates multicast scouting. - /// - /// - Accepted values : `"true"`, `"false"`. - /// - Default value : `"true"`. - #[classattr] - pub fn MULTICAST_SCOUTING_KEY() -> &'static str { - "scouting/multicast/enabled" - } - - /// The network interface to use for multicast scouting. - /// - /// - Accepted values : `"auto"`, `<ip address>`, `<interface name>`. - /// - Default value : `"auto"`. - #[classattr] - pub fn MULTICAST_INTERFACE_KEY() -> &'static str { - "scouting/multicast/interface" - } - - /// The multicast address and ports to use for multicast scouting. - /// - /// - Accepted values : `<ip address>:<port>`. - /// - Default value : `"224.0.0.224:7447"`. - #[classattr] - pub fn MULTICAST_IPV4_ADDRESS_KEY() -> &'static str { - "scouting/multicast/address" - } - - /// In client mode, the period dedicated to scouting a router before failing. - /// - /// - Accepted values : `<float in seconds>`. - /// - Default value : `"3.0"`. - #[classattr] - pub fn SCOUTING_TIMEOUT_KEY() -> &'static str { - "scouting/timeout" - } - - /// In peer mode, the period dedicated to scouting first remote peers before doing anything else. - /// - /// - Accepted values : `<float in seconds>`. - /// - Default value : `"0.2"`. - #[classattr] - pub fn SCOUTING_DELAY_KEY() -> &'static str { - "scouting/delay" - } - - /// Indicates if data messages should be timestamped. - /// - /// - Accepted values : `"true"`, `"false"`. - /// - Default value : `"false"`. - #[classattr] - pub fn ADD_TIMESTAMP_KEY() -> &'static str { - "add_timestamp" - } - - /// Indicates if local writes/queries should reach local subscribers/queryables. - #[classattr] - pub fn LOCAL_ROUTING_KEY() -> &'static str { - "local_routing" - } -} - -// zenoh.info (simulate the package as a class, and consts as class attributes) -/// Constants and helpers to interpret the properties returned by :func:`zenoh.Session.info`. -#[allow(non_camel_case_types)] -#[pyclass] -pub(crate) struct info {} - -#[allow(non_snake_case)] -#[pymethods] -impl info { - #[classattr] - fn ZN_INFO_PID_KEY() -> ZInt { - zenoh::info::ZN_INFO_PID_KEY - } - - #[classattr] - fn ZN_INFO_PEER_PID_KEY() -> ZInt { - zenoh::info::ZN_INFO_PEER_PID_KEY - } - - #[classattr] - fn ZN_INFO_ROUTER_PID_KEY() -> ZInt { - zenoh::info::ZN_INFO_ROUTER_PID_KEY - } -} - -// zenoh.whatami (simulate the package as a class, and consts as class attributes) -/// Constants defining the different zenoh process to look for with :func:`zenoh.scout`. -#[allow(non_camel_case_types)] -#[pyclass] -#[derive(Debug, Clone, Copy, PartialEq, Eq, Hash)] -pub struct WhatAmI { - inner: WhatAmIMatcher, -} -impl From<WhatAmI> for WhatAmIMatcher { - fn from(w: WhatAmI) -> Self { - w.inner - } -} -impl BitOr for WhatAmI { - type Output = Self; - fn bitor(self, rhs: Self) -> Self::Output { - WhatAmI { - inner: self.inner \| rhs.inner, - } - } -} -#[pyproto] -impl pyo3::PyNumberProtocol for WhatAmI -where - <Self as PyNumberOrProtocol<'p>>::Left: BitOr< - <Self as PyNumberOrProtocol<'p>>::Right, - Output = <Self as PyNumberOrProtocol<'p>>::Result, - >, -{ - fn __or__(lhs: Self, rhs: Self) -> Self - where - Self: PyNumberOrProtocol<'p>, - { - lhs \| rhs - } -} - -#[allow(non_snake_case)] -#[pymethods] -impl WhatAmI { - #[classattr] - fn Router() -> Self { - WhatAmI { - inner: ZWhatAmI::Router.into(), - } - } - - #[classattr] - fn Peer() -> Self { - WhatAmI { - inner: ZWhatAmI::Peer.into(), - } - } - - #[classattr] - fn Client() -> Self { - WhatAmI { - inner: ZWhatAmI::Client.into(), - } - } -} - -impl std::fmt::Display for WhatAmI { - fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result { - write!(f, "{}", self.inner.to_str()) - } -} - -/// A Hello message received as a response to a :meth:`scout` -#[pyclass] -#[derive(Clone)] -pub(crate) struct Hello { - pub(crate) h: zenoh::scouting::Hello, -} - -#[pymethods] -impl Hello { - /// The PeerId of the Hello message sender - /// - /// :type: :class:`PeerId` or `None` - #[getter] - fn pid(&self) -> Option<PeerId> { - self.h.pid.as_ref().map(\|p\| PeerId { p: p }) - } - - /// The mode of the Hello message sender (bitmask of constants from :class:`whatami`) - /// - /// :type: :class:`whatami` or `None` - #[getter] - fn whatami(&self) -> Option<WhatAmI> { - self.h.whatami.map(\|w\| WhatAmI { inner: w.into() }) - } - - /// The locators list of the Hello message sender - /// - /// :type: list of str or `None` - #[getter] - fn locators(&self) -> Option<Vec<String>> { - self.h - .locators - .as_ref() - .map(\|v\| v.iter().map(\|l\| l.to_string()).collect()) - } -} - -#[pyproto] -impl PyObjectProtocol for Hello { - fn __str__(&self) -> String { - self.h.to_string() - } -} - -// zenoh.resource_name (simulate the package as a class with static methodss) -#[allow(non_camel_case_types)] -#[pyclass] -pub struct KeyExpr { - pub(crate) inner: ZKeyExpr<'static>, -} - -#[allow(non_snake_case)] -#[pymethods] -impl KeyExpr { - /// Return true if both resource names intersect. - /// - /// :param s1: the 1st resource name - /// :type s1: str - /// :param s2: the 2nd resource name - /// :type s2: str - #[staticmethod] - #[pyo3(text_signature = "(s1, s2)")] - fn intersect(s1: &PyAny, s2: &PyAny) -> bool { - let s1 = zkey_expr_of_pyany(s1).unwrap(); - let s2 = zkey_expr_of_pyany(s2).unwrap(); - match (s1.as_id_and_suffix(), s2.as_id_and_suffix()) { - ((s1, _), (s2, _)) if s1 != s2 => false, - ((_, s1), (_, s2)) => zenoh::utils::key_expr::intersect(s1, s2), - } - } -} - -#[pyproto] -impl PyObjectProtocol for KeyExpr { - fn __str__(&self) -> PyResult<String> { - Ok(self.inner.to_string()) - } -} - -impl From<KeyExpr> for ZKeyExpr<'static> { - fn from(r: KeyExpr) -> ZKeyExpr<'static> { - r.inner - } -} - -impl From<ZKeyExpr<'static>> for KeyExpr { - fn from(inner: ZKeyExpr<'static>) -> KeyExpr { - KeyExpr { inner } - } -} - -pub(crate) fn zkey_expr_of_pyany(obj: &PyAny) -> PyResult<ZKeyExpr> { - match obj.get_type().name()? { - "KeyExpr" => { - let rk: PyRef<KeyExpr> = obj.extract()?; - Ok(rk.inner.clone()) - } - "int" => { - let id: u64 = obj.extract()?; - Ok(id.into()) - } - "str" => { - let name: String = obj.extract()?; - Ok(name.into()) - } - "tuple" => { - let tuple: &PyTuple = obj.downcast()?; - if tuple.len() == 2 - && tuple.get_item(0)?.get_type().name()? == "int" - && tuple.get_item(1)?.get_type().name()? == "str" - { - let id: u64 = tuple.get_item(0)?.extract()?; - let suffix: String = tuple.get_item(1)?.extract()?; - Ok(ZKeyExpr::from(id).with_suffix(&suffix).to_owned()) - } else { - Err(PyErr::new::<exceptions::PyValueError, _>(format!( - "Cannot convert type '{:?}' to a zenoh-net KeyExpr", - tuple - ))) - } - } - x => Err(PyErr::new::<exceptions::PyValueError, _>(format!( - "Cannot convert type '{}' to a zenoh-net KeyExpr", - x - ))), - } -} - -/// An expression identifying a selection of resources. -/// -/// A selector is the conjunction of an key expression identifying a set -/// of resource keys and a value selector filtering out the resource values. -/// -/// Structure of a selector::: -/// -/// /s1/s2/..../sn?x>1&y<2&...&z=4(p1=v1;p2=v2;...;pn=vn)[a;b;x;y;...;z] -/// \|key_selector\|\|---------------- value_selector --------------------\| -/// \|--- filter --\| \|---- properties ---\| \|--fragment-\| -/// -/// where: -/// * key_selector: an expression identifying a set of Resources. -/// * filter: a list of `value_selectors` separated by `'&'` allowing to perform filtering on the values -/// associated with the matching keys. Each `value_selector` has the form "`field`-`operator`-`value`" value where: -/// -/// * field is the name of a field in the value (is applicable and is existing. otherwise the `value_selector` is false) -/// * operator is one of a comparison operators: `<` , `>` , `<=` , `>=` , `=` , `!=` -/// * value is the the value to compare the field’s value with -/// -/// * fragment: a list of fields names allowing to return a sub-part of each value. -/// This feature only applies to structured values using a “self-describing” encoding, such as JSON or XML. -/// It allows to select only some fields within the structure. A new structure with only the selected fields -/// will be used in place of the original value. -/// -/// NOTE: the filters and fragments are not yet supported in current zenoh version. -#[allow(non_camel_case_types)] -#[pyclass] -pub struct Selector { - pub(crate) s: ZSelector<'static>, -} - -#[allow(non_snake_case)] -#[pymethods] -impl Selector { - /// The part of this selector identifying which keys should be part of the selection. - /// I.e. all characters before `?`. - /// - /// :type: :class:`KeyExpr` - #[getter] - fn key_selector(&self) -> KeyExpr { - KeyExpr { - inner: self.s.key_selector.to_owned(), - } - } - - /// the part of this selector identifying which values should be part of the selection. - /// I.e. all characters starting from `?`. - /// - /// :type: str - #[getter] - fn value_selector(&self) -> &str { - self.s.value_selector.as_ref() - } - - /// Parses the `value_selector` part of this `Selector`. - /// - /// :rtype: :class:`ValueSelector` - fn parse_value_selector(&self) -> PyResult<ValueSelector> { - let zvs = self.s.parse_value_selector().map_err(to_pyerr)?; - Ok(ValueSelector { - filter: zvs.filter.to_owned(), - properties: zvs.properties.0, - fragment: zvs.fragment.map(\|cow\| cow.to_owned()), - }) - } -} - -#[pyproto] -impl PyObjectProtocol for Selector { - fn __str__(&self) -> String { - self.s.to_string() - } -} - -impl From<Selector> for ZSelector<'static> { - fn from(s: Selector) -> ZSelector<'static> { - s.s - } -} - -impl From<ZSelector<'static>> for Selector { - fn from(s: ZSelector<'static>) -> Selector { - Selector { s } - } -} - -/// A class that can be used to help decoding or encoding the `value_selector` part of a :class:`Selector`. -#[allow(non_camel_case_types)] -#[pyclass] -pub struct ValueSelector { - pub(crate) filter: String, - pub(crate) properties: HashMap<String, String>, - pub(crate) fragment: Option<String>, -} - -#[allow(non_snake_case)] -#[pymethods] -impl ValueSelector { - /// the filter part of this `ValueSelector`, if any (all characters after `?` and before `(` or `[`) - /// - /// :type: str - #[getter] - fn filter(&self) -> &str { - self.filter.as_ref() - } - - /// the properties part of this `ValueSelector`) (all characters between ``( )`` and after `?`) - /// - /// :type: str - #[getter] - fn properties(&self) -> HashMap<String, String> { - self.properties.clone() - } - - /// the filter part of this `ValueSelector`, if any (all characters after `?` and before `(` or `[`) - /// - /// :type: str - #[getter] - fn fragment(&self) -> Option<&str> { - self.fragment.as_ref().map(\|s\| s.as_ref()) - } -} - -/// A Peer id -#[pyclass] -pub(crate) struct PeerId { - pub(crate) p: zenoh::prelude::PeerId, -} - -#[pyproto] -impl PyObjectProtocol for PeerId { - fn __str__(&self) -> String { - self.p.to_string() - } -} - -/// A zenoh Value, consisting of a payload (bytes) and an :class:`Encoding`. -/// -/// It can be created directly from the supported primitive types. -/// The value is automatically encoded in the payload and the Encoding is set accordingly. -/// -/// Or it can be created from a tuple (payload, encoding), where: -/// -/// - payload has type bytes or str (the string is automatically converted into bytes) -/// - encoding has type :class:`Encoding` -/// -/// :Examples: -/// -/// >>> import json, zenoh -/// >>> from zenoh import Encoding, Value -/// >>> -/// >>> string_value = Value('Hello World!') -/// >>> int_value = Value(42) -/// >>> float_value = Value(3.14) -/// >>> bytes_value = Value(b'\x48\x69\x21') -/// >>> properties_value = Value({'p1': 'v1', 'p2': 'v2'}) -/// >>> -/// >>> json_value = Value((json.dumps(['foo', {'bar': ('baz', None, 1.0, 2)}]), Encoding.TEXT_JSON)) -/// >>> xml_value = Value(('<foo>bar</foo>', Encoding.TEXT_XML)) -/// >>> custom_value = Value((b'\x48\x69\x21', Encoding.APP_CUSTOM.with_suffix('my_encoding'))) -#[pyclass] -#[derive(Clone)] -pub struct Value { - pub(crate) v: ZValue, -} -impl From<Value> for ZValue { - fn from(v: Value) -> Self { - v.v - } -} -impl From<ZValue> for Value { - fn from(v: ZValue) -> Self { - Value { v } - } -} - -trait IntoPyAlt<U> { - fn into_py_alt(self, py: Python) -> U; -} - -impl IntoPyAlt<PyObject> for serde_json::Value { - fn into_py_alt(self, py: Python) -> PyObject { - match self { - serde_json::Value::Null => py.None(), - serde_json::Value::Bool(v) => v.into_py(py), - serde_json::Value::Number(v) => v.into_py_alt(py), - serde_json::Value::String(v) => v.into_py(py), - serde_json::Value::Array(a) => a - .into_iter() - .map(\|v\| v.into_py_alt(py)) - .collect::<Vec<_>>() - .into_py(py), - serde_json::Value::Object(m) => m - .into_iter() - .map(\|(k, v)\| (k, v.into_py_alt(py))) - .collect::<std::collections::HashMap<_, _>>() - .into_py(py), - } - } -} -impl IntoPyAlt<PyObject> for serde_json::Number { - fn into_py_alt(self, py: Python) -> PyObject { - if let Some(v) = self.as_u64() { - return v.into_py(py); - } - if let Some(v) = self.as_i64() { - return v.into_py(py); - } - if let Some(v) = self.as_f64() { - return v.into_py(py); - } - unreachable!() - } -} - -#[allow(non_snake_case)] -#[pymethods] -impl Value { - #[new] - fn new(any: &PyAny) -> PyResult<Self> { - Ok(Value { - v: zvalue_of_pyany(any)?, - }) - } - - /// the payload the Value. - /// - /// :type: bytes - #[getter] - fn payload<'a>(&self, py: Python<'a>) -> &'a PyBytes { - PyBytes::new(py, self.v.payload.contiguous().as_ref()) - } - - /// the encoding of the Value. - /// - /// :type: :class:`Encoding` - #[getter] - fn encoding(&self) -> PyResult<Encoding> { - Ok(self.v.encoding.clone().into()) - } - - /// Try to decode the value's payload according to it's encoding, and return a typed object or primitive. - /// - /// :rtype: depend on the encoding flag (e.g. str for a StringUtf8 Value, int for an Integer Value ...) - fn decode(&self, py: Python) -> PyResult<PyObject> { - match self.v.encoding.prefix() { - ZKnownEncoding::Empty \| ZKnownEncoding::AppOctetStream => { - Ok(self.v.payload.contiguous().into_py(py)) - } - ZKnownEncoding::TextPlain => { - Ok(String::from_utf8_lossy(&self.v.payload.contiguous()).into_py(py)) - } - ZKnownEncoding::AppProperties => self - .v - .as_properties() - .map(\|v\| v.0.into_py(py)) - .ok_or_else(\|\| { - exceptions::PyTypeError::new_err( - "Failed to decode Value's payload as Properties", - ) - }), - ZKnownEncoding::AppJson \| ZKnownEncoding::TextJson => self - .v - .as_json() - .map(\|v: serde_json::Value\| v.into_py_alt(py)) - .ok_or_else(\|\| { - exceptions::PyTypeError::new_err("Failed to decode Value's payload as JSON") - }), - ZKnownEncoding::AppInteger => self - .v - .as_integer() - .map(\|v: i64\| v.into_py(py)) - .ok_or_else(\|\| { - exceptions::PyTypeError::new_err("Failed to decode Value's payload as Integer") - }), - ZKnownEncoding::AppFloat => { - self.v - .as_float() - .map(\|v: f64\| v.into_py(py)) - .ok_or_else(\|\| { - exceptions::PyTypeError::new_err( - "Failed to decode Value's payload as Float", - ) - }) - } - _ => Err(exceptions::PyTypeError::new_err(format!( - "Don't know how to decode Value's payload with encoding: {}", - self.v.encoding - ))), - } - } -} - -#[pyproto] -impl PyObjectProtocol for Value { - fn __str__(&self) -> PyResult<String> { - Ok(format!("{:?}", self.v)) - } - - fn __repr__(&self) -> PyResult<String> { - self.__str__() - } -} - -pub(crate) fn zvalue_of_pyany(obj: &PyAny) -> PyResult<ZValue> { - match obj.get_type().name()? { - "Value" => { - let v: Value = obj.extract()?; - Ok(v.v) - } - "bytes" => { - let buf: &[u8] = obj.extract()?; - Ok(Vec::from(buf).into()) - } - "str" => { - let s: String = obj.extract()?; - Ok(s.into()) - } - "dict" => { - let props: HashMap<String, String> = obj.extract()?; - Ok(zenoh::prelude::Properties::from(props).into()) - } - "int" => { - let i: i64 = obj.extract()?; - Ok(i.into()) - } - "float" => { - let f: f64 = obj.extract()?; - Ok(f.into()) - } - "tuple" => { - let tuple: &PyTuple = obj.downcast()?; - if tuple.len() == 2 - && (tuple.get_item(0)?.get_type().name()? == "bytes" - \|\| tuple.get_item(0)?.get_type().name()? == "str") - && (tuple.get_item(1)?.get_type().name()? == "str" - \|\| tuple.get_item(1)?.get_type().name()? == "Encoding") - { - let buf: &[u8] = if tuple.get_item(0)?.get_type().name()? == "bytes" { - tuple.get_item(0)?.extract()? - } else { - tuple.get_item(0)?.extract::<&str>()?.as_bytes() - }; - let encoding_descr: ZEncoding = if tuple.get_item(1)?.get_type().name()? == "str" { - tuple.get_item(1)?.extract::<String>()?.into() - } else { - tuple.get_item(1)?.extract::<Encoding>()?.e - }; - Ok(ZValue::new(Vec::from(buf).into()).encoding(encoding_descr)) - } else { - Err(PyErr::new::<exceptions::PyValueError, _>(format!( - "Cannot convert type '{:?}' to a zenoh Value", - tuple - ))) - } - } - x => Err(PyErr::new::<exceptions::PyValueError, _>(format!( - "Cannot convert type '{}' to a zenoh Value", - x - ))), - } -} - -/// A Timestamp composed of a time and the identifier of the timestamp source. -#[pyclass] -#[derive(Debug, Clone, Copy)] -pub(crate) struct Timestamp { - pub(crate) t: zenoh::time::Timestamp, -} - -#[pymethods] -impl Timestamp { - /// The time in seconds since the UNIX EPOCH (January 1, 1970, 00:00:00 (UTC)) - /// as a floating point number. - /// - /// :type: float - #[getter] - fn time(&self) -> f64 { - self.t.get_time().to_duration().as_secs_f64() - } - - /// The identifier of the timestamp source - /// - /// :type: bytes - #[getter] - fn id(&self) -> &[u8] { - self.t.get_id().as_slice() - } -} - -#[pyproto] -impl PyObjectProtocol for Timestamp { - fn __str__(&self) -> String { - self.t.to_string() - } -} - -/// Some informations about the associated data -#[pyclass] -#[derive(Clone)] -pub(crate) struct SourceInfo { - pub(crate) i: zenoh::prelude::SourceInfo, -} - -#[pymethods] -impl SourceInfo { - /// The :class:`PeerId` of the data source. - /// - /// :type: :class:`PeerId` or `None` - #[getter] - fn source_id(&self) -> Option<PeerId> { - self.i.source_id.as_ref().map(\|p\| PeerId { p: p }) - } - - /// The source sequence number of the data. - /// - /// :type: int or `None` - #[getter] - fn source_sn(&self) -> Option<ZInt> { - self.i.source_sn - } - - /// The :class:`PeerId` of the 1st router that routed the data. - /// - /// :type: :class:`PeerId` or `None` - #[getter] - fn first_router_id(&self) -> Option<PeerId> { - self.i.first_router_id.as_ref().map(\|p\| PeerId { p: p }) - } - - /// The first router sequence number of the data. - /// - /// :type: int or `None` - #[getter] - fn first_router_sn(&self) -> Option<ZInt> { - self.i.first_router_sn - } -} - -/// A zenoh sample. -/// -/// :param key_expr: the resource name -/// :type key_expr: str -/// :param payload: the data payload -/// :type payload: bytes -/// :param source_info: some information about the data -/// :type source_info: SourceInfo, optional -#[pyclass] -#[pyo3(text_signature = "(key_expr, payload, source_info=None)")] -#[derive(Clone)] -pub(crate) struct Sample { - pub(crate) s: zenoh::prelude::Sample, -} - -impl pyo3::conversion::ToPyObject for Sample { - fn to_object(&self, py: Python) -> pyo3::PyObject { - pyo3::IntoPy::into_py(pyo3::Py::new(py, self.clone()).unwrap(), py) - } -} - -#[pymethods] -impl Sample { - #[new] - fn new(key_expr: &PyAny, payload: &PyAny) -> Self { - let key_expr = zkey_expr_of_pyany(key_expr).unwrap(); - let payload = zvalue_of_pyany(payload).unwrap(); - Sample { - s: zenoh::prelude::Sample::new(key_expr.to_owned(), payload), - } - } - pub fn with_timestamp(&mut self, timestamp: Timestamp) { - unsafe { - let s = std::ptr::read(self); - let s = s.s.with_timestamp(timestamp.t); - std::ptr::write(self, Sample { s }); - } - } - pub fn with_source_info(&mut self, info: SourceInfo) { - unsafe { - let s = std::ptr::read(self); - let s = s.s.with_source_info(info.i); - std::ptr::write(self, Sample { s }); - } - } - - /// The resource name - /// - /// :type: str - #[getter] - fn key_expr(&self) -> KeyExpr { - self.s.key_expr.to_owned().into() - } - - /// The data payload - /// - /// DEPRECATED: use the strictly equivalent code: `sample.value.payload` - /// - /// :type: bytes - #[getter] - fn payload<'a>(&self, py: Python<'a>) -> &'a PyBytes { - PyBytes::new(py, self.s.value.payload.contiguous().as_ref()) - } - - /// The data payload - /// - /// :type: bytes - #[getter] - fn value(&self) -> Value { - Value { - v: self.s.value.clone(), - } - } - - /// The data payload - /// - /// :type: bytes - #[getter] - fn kind(&self) -> SampleKind { - self.s.kind.into() - } - - /// Some information about the data - /// - /// :type: :class:`SourceInfo` or `None` - #[getter] - fn source_info(&self) -> Option<SourceInfo> { - Some(SourceInfo { - i: self.s.source_info.clone(), - }) - } - - /// The timestamp - /// - /// :type: :class:`Timestamp` or `None` - #[getter] - fn timestamp(&self) -> Option<Timestamp> { - self.s.timestamp.map(\|t\| Timestamp { t }) - } -} - -#[pyproto] -impl PyObjectProtocol for Sample { - fn __str__(&self) -> String { - format!("{:?}", self.s) - } - - fn __repr__(&self) -> String { - self.__str__() - } -} - -// zenoh.Reliability (simulate the enum as a class with static methods for the cases, -// waiting for https://github.com/PyO3/pyo3/issues/834 to be fixed) -// -/// The kind of reliability -#[pyclass] -#[derive(Clone, Copy, PartialEq, Default)] -pub(crate) struct Reliability { - pub(crate) r: zenoh::subscriber::Reliability, -} - -#[allow(non_snake_case)] -#[pymethods] -impl Reliability { - #[classattr] - fn BestEffort() -> Reliability { - Reliability { - r: zenoh::subscriber::Reliability::BestEffort, - } - } - - #[classattr] - fn Reliable() -> Reliability { - Reliability { - r: zenoh::subscriber::Reliability::Reliable, - } - } -} - -// zenoh.SubMode (simulate the enum as a class with static methods for the cases, -// waiting for https://github.com/PyO3/pyo3/issues/834 to be fixed) -// -/// The subscription mode. -#[pyclass] -#[derive(Clone, Default)] -pub(crate) struct SubMode { - pub(crate) m: zenoh::subscriber::SubMode, -} - -#[allow(non_snake_case)] -#[pymethods] -impl SubMode { - #[classattr] - fn Push() -> SubMode { - SubMode { - m: zenoh::subscriber::SubMode::Push, - } - } - - #[classattr] - fn Pull() -> SubMode { - SubMode { - m: zenoh::subscriber::SubMode::Pull, - } - } -} - -/// A time period. -#[pyclass] -#[pyo3(text_signature = "(origin, period, duration)")] -#[derive(Clone)] -pub(crate) struct Period { - pub(crate) p: zenoh::time::Period, -} - -#[pymethods] -impl Period { - #[new] - fn new(origin: ZInt, period: ZInt, duration: ZInt) -> Period { - Period { - p: zenoh::time::Period { - origin, - period, - duration, - }, - } - } -} - -pub(crate) enum ZnSubOps { - Pull, - Unregister, -} - -/// A subscriber -#[pyclass] -pub(crate) struct Subscriber { - pub(crate) unregister_tx: Sender<ZnSubOps>, - pub(crate) loop_handle: Option<async_std::task::JoinHandle<()>>, -} - -#[pymethods] -impl Subscriber { - /// Pull available data for a pull-mode subscriber. - fn pull(&self) { - task::block_on(async { - if let Err(e) = self.unregister_tx.send(ZnSubOps::Pull).await { - warn!("Error in Subscriber::pull() : {}", e); - } - }); - } - - /// Close the subscriber. - fn close(&mut self) { - if let Some(handle) = self.loop_handle.take() { - task::block_on(async { - if let Err(e) = self.unregister_tx.send(ZnSubOps::Unregister).await { - warn!("Error in Subscriber::close() : {}", e); - } - handle.await; - }); - } - } -} - -// zenoh.queryable (simulate the package as a class, and consts as class attributes) -// -/// Constants defining the different modes of a zenoh :class:`zenoh.Queryable`. -#[allow(non_camel_case_types)] -#[pyclass] -pub(crate) struct queryable {} - -#[allow(non_snake_case)] -#[pymethods] -impl queryable { - #[classattr] - fn ALL_KINDS() -> ZInt { - zenoh::queryable::ALL_KINDS - } - - #[classattr] - fn STORAGE() -> ZInt { - zenoh::queryable::STORAGE - } - - #[classattr] - fn EVAL() -> ZInt { - zenoh::queryable::EVAL - } -} - -/// Type received by a queryable callback. See :meth:`Session.register_queryable`. -#[pyclass] -#[derive(Clone)] -pub(crate) struct Query { - pub(crate) q: async_std::sync::Arc<zenoh::queryable::Query>, -} - -impl pyo3::conversion::ToPyObject for Query { - fn to_object(&self, py: Python) -> pyo3::PyObject { - pyo3::IntoPy::into_py(pyo3::Py::new(py, self.clone()).unwrap(), py) - } -} - -#[pymethods] -impl Query { - /// The key_selector of the query - /// - /// :type: :class:`Selector` - #[getter] - fn selector(&self) -> Selector { - self.q.selector().to_owned().into() - } - - /// The key_selector of the query - /// - /// :type: :class:`KeyExpr` - #[getter] - fn key_selector(&self) -> KeyExpr { - self.q.key_selector().to_owned().into() - } - - /// The value_selector of the query - /// - /// :type: str - #[getter] - fn value_selector(&self) -> String { - self.q.value_selector().to_string() - } - - /// Send a reply to the query - /// - /// :param sample: the reply sample - /// :type: Sample - #[pyo3(text_signature = "(self, sample)")] - fn reply(&self, sample: Sample) { - self.q.reply(sample.s); - } -} - -/// An entity able to reply to queries. -#[pyclass] -pub(crate) struct Queryable { - pub(crate) unregister_tx: Sender<bool>, - pub(crate) loop_handle: Option<async_std::task::JoinHandle<()>>, -} - -#[pymethods] -impl Queryable { - /// Close the queryable. - fn close(&mut self) { - if let Some(handle) = self.loop_handle.take() { - task::block_on(async { - if let Err(e) = self.unregister_tx.send(true).await { - warn!("Error in Queryable::close() : {}", e); - } - handle.await; - }); - } - } -} - -// zenoh.Target (simulate the enum as a class with static methods for the cases, -// waiting for https://github.com/PyO3/pyo3/issues/834 to be fixed) -// -/// The queryables that should be target of a :class:`Query` -#[pyclass] -#[derive(Clone)] -pub(crate) struct Target { - pub(crate) t: zenoh::query::Target, -} - -#[allow(non_snake_case)] -#[pymethods] -impl Target { - #[staticmethod] - fn BestMatching() -> Target { - Target { - t: zenoh::query::Target::BestMatching, - } - } - - #[cfg(features = "complete_n")] - #[staticmethod] - #[pyo3(text_signature = "(n)")] - fn Complete(n: ZInt) -> Target { - Target { - t: zenoh::query::Target::Complete { n }, - } - } - - #[staticmethod] - fn All() -> Target { - Target { - t: zenoh::query::Target::All, - } - } - - #[staticmethod] - fn AllComplete() -> Target { - Target { - t: zenoh::query::Target::AllComplete, - } - } - - #[staticmethod] - fn No() -> Target { - Target { - t: zenoh::query::Target::None, - } - } -} - -/// The queryables that should be target of a :class:`Query`. -/// -/// :param kind: the kind of queryable (one constant from :class:`queryable`) -/// :type kind: int, optional -/// :param target: a characteristic of the queryable. -/// :type target: Target, optional -#[pyclass] -#[pyo3(text_signature = "(kind=None, target=None)")] -#[derive(Clone, Default)] -pub(crate) struct QueryTarget { - pub(crate) t: zenoh::query::QueryTarget, -} - -#[pymethods] -impl QueryTarget { - #[new] - fn new(kind: Option<ZInt>, target: Option<Target>) -> QueryTarget { - let mut t = zenoh::query::QueryTarget::default(); - if let Some(k) = kind { - t.kind = k; - } - if let Some(target) = target { - t.target = target.t; - } - QueryTarget { t } - } -} - -// zenoh.QueryConsolidation (simulate the enum as a class with static methods for the cases, -// waiting for https://github.com/PyO3/pyo3/issues/834 to be fixed) -// -/// The kind of consolidation that should be applied on replies to a :meth:`Session.get`. -#[pyclass] -#[derive(Clone)] -pub(crate) struct ConsolidationMode { - pub(crate) c: zenoh::query::ConsolidationMode, -} - -#[allow(non_snake_case)] -#[pymethods] -impl ConsolidationMode { - #[classattr] - fn No() -> ConsolidationMode { - ConsolidationMode { - c: zenoh::query::ConsolidationMode::None, - } - } - - #[classattr] - fn Lazy() -> ConsolidationMode { - ConsolidationMode { - c: zenoh::query::ConsolidationMode::Lazy, - } - } - - #[classattr] - fn Full() -> ConsolidationMode { - ConsolidationMode { - c: zenoh::query::ConsolidationMode::Full, - } - } -} - -// zenoh.ConsolidationStrategy -// -/// The kind of consolidation that should be applied on replies to a :meth:`Session.get` -/// at the different stages of the reply process. -/// -/// :param first_routers: the consolidation mode to apply on first routers of the replies routing path (default: :attr:`ConsolidationMode.Lazy`) -/// :type first_routers: ConsolidationMode, optional -/// :param last_router: the consolidation mode to apply on last router of the replies routing path (default: :attr:`ConsolidationMode.Lazy`) -/// :type last_router: ConsolidationMode, optional -/// :param reception: the consolidation mode to apply at reception of the replies (default: :attr:`ConsolidationMode.Full`) -/// :type reception: ConsolidationMode, optional -#[pyclass] -#[pyo3(text_signature = "(first_routers=None, last_router=None, reception=None)")] -#[derive(Clone, Default)] -pub(crate) struct ConsolidationStrategy { - pub(crate) c: zenoh::query::ConsolidationStrategy, -} - -#[pymethods] -impl ConsolidationStrategy { - #[new] - fn new( - first_routers: Option<ConsolidationMode>, - last_router: Option<ConsolidationMode>, - reception: Option<ConsolidationMode>, - ) -> ConsolidationStrategy { - let mut c = zenoh::query::ConsolidationStrategy::default(); - if let Some(f) = first_routers { - c.first_routers = f.c; - } - if let Some(l) = last_router { - c.last_router = l.c; - } - if let Some(r) = reception { - c.reception = r.c; - } - ConsolidationStrategy { c } - } - - /// No consolidation performed. - /// - /// This is usefull when querying timeseries data bases or - /// when using quorums. - #[staticmethod] - fn none() -> Self { - Self { - c: zenoh::query::ConsolidationStrategy::none(), - } - } - - /// Lazy consolidation performed at all stages. - /// - /// This strategy offers the best latency. Replies are directly - /// transmitted to the application when received without needing - /// to wait for all replies. - /// - /// This mode does not garantie that there will be no duplicates. - #[staticmethod] - pub fn lazy() -> Self { - Self { - c: zenoh::query::ConsolidationStrategy::lazy(), - } - } - - /// Full consolidation performed at reception. - /// - /// This is the default strategy. It offers the best latency while - /// garantying that there will be no duplicates. - #[staticmethod] - pub fn reception() -> Self { - Self { - c: zenoh::query::ConsolidationStrategy::reception(), - } - } - - /// Full consolidation performed on last router and at reception. - /// - /// This mode offers a good latency while optimizing bandwidth on - /// the last transport link between the router and the application. - #[staticmethod] - pub fn last_router() -> Self { - Self { - c: zenoh::query::ConsolidationStrategy::last_router(), - } - } - - /// Full consolidation performed everywhere. - /// - /// This mode optimizes bandwidth on all links in the system - /// but will provide a very poor latency. - #[staticmethod] - pub fn full() -> Self { - Self { - c: zenoh::query::ConsolidationStrategy::full(), - } - } -} - -// zenoh.QueryConsolidation (simulate the enum as a class with static methods for the cases, -// waiting for https://github.com/PyO3/pyo3/issues/834 to be fixed) -// -/// The replies consolidation strategy to apply on replies to a :meth:`Session.get`. -#[pyclass] -#[pyo3(text_signature = "(first_routers=None, last_router=None, reception=None)")] -#[derive(Clone, Default)] -pub(crate) struct QueryConsolidation { - pub(crate) c: zenoh::query::QueryConsolidation, -} - -#[allow(non_snake_case)] -#[pymethods] -impl QueryConsolidation { - /// Automatic query consolidation strategy selection. - /// - /// A query consolidation strategy will automatically be selected depending - /// the query selector. If the selector contains time range properties, - /// no consolidation is performed. Otherwise the reception strategy is used. - #[staticmethod] - fn Auto() -> Self { - QueryConsolidation { - c: zenoh::query::QueryConsolidation::Auto, - } - } - - /// User defined query consolidation strategy. - #[staticmethod] - fn Manual(strategy: ConsolidationStrategy) -> Self { - QueryConsolidation { - c: zenoh::query::QueryConsolidation::Manual(strategy.c), - } - } - - /// No consolidation performed. - /// - /// This is usefull when querying timeseries data bases or - /// when using quorums. - #[staticmethod] - fn none() -> Self { - Self { - c: zenoh::query::QueryConsolidation::none(), - } - } - - /// Lazy consolidation performed at all stages. - /// - /// This strategy offers the best latency. Replies are directly - /// transmitted to the application when received without needing - /// to wait for all replies. - /// - /// This mode does not garantie that there will be no duplicates. - #[staticmethod] - pub fn lazy() -> Self { - Self { - c: zenoh::query::QueryConsolidation::lazy(), - } - } - - /// Full consolidation performed at reception. - /// - /// This is the default strategy. It offers the best latency while - /// garantying that there will be no duplicates. - #[staticmethod] - pub fn reception() -> Self { - Self { - c: zenoh::query::QueryConsolidation::reception(), - } - } - - /// Full consolidation performed on last router and at reception. - /// - /// This mode offers a good latency while optimizing bandwidth on - /// the last transport link between the router and the application. - #[staticmethod] - pub fn last_router() -> Self { - Self { - c: zenoh::query::QueryConsolidation::last_router(), - } - } - - /// Full consolidation performed everywhere. - /// - /// This mode optimizes bandwidth on all links in the system - /// but will provide a very poor latency. - #[staticmethod] - pub fn full() -> Self { - Self { - c: zenoh::query::QueryConsolidation::full(), - } - } -} - -/// Type received by a get callback. See :meth:`Session.get`. -#[pyclass] -#[derive(Clone)] -pub(crate) struct Reply { - pub(crate) r: zenoh::query::Reply, -} - -impl pyo3::conversion::ToPyObject for Reply { - fn to_object(&self, py: Python) -> pyo3::PyObject { - pyo3::IntoPy::into_py(pyo3::Py::new(py, self.clone()).unwrap(), py) - } -} - -#[pymethods] -impl Reply { - /// The sample - /// - /// :type: Sample - #[getter] - fn sample(&self) -> Sample { - Sample { - s: self.r.sample.clone(), - } - } - - /// The kind of reply source - /// - /// :type: int - #[getter] - fn replier_kind(&self) -> ZInt { - self.r.replier_kind - } - - /// The identifier of reply source - /// - /// :type: PeerId - fn replier_id(&self) -> PeerId { - PeerId { - p: self.r.replier_id, - } - } -} - -// zenoh.CongestionControl (simulate the enum as a class with static methods for the cases, -// waiting for https://github.com/PyO3/pyo3/issues/834 to be fixed) -// -/// The kind of congestion control. -#[pyclass] -#[derive(Clone, Default)] -pub struct CongestionControl { - pub(crate) cc: zenoh::publication::CongestionControl, -} - -#[allow(non_snake_case)] -#[pymethods] -impl CongestionControl { - #[classattr] - fn Drop() -> CongestionControl { - CongestionControl { - cc: zenoh::publication::CongestionControl::Drop, - } - } - #[classattr] - fn Block() -> CongestionControl { - CongestionControl { - cc: zenoh::publication::CongestionControl::Block, - } - } -} - -#[pyclass] -#[derive(Clone, Default)] -pub struct Priority { - pub(crate) p: zenoh::prelude::Priority, -} - -#[allow(non_snake_case)] -#[pymethods] -impl Priority { - #[classattr] - fn Background() -> Self { - Priority { - p: zenoh::prelude::Priority::Background, - } - } - #[classattr] - fn Data() -> Self { - Priority { - p: zenoh::prelude::Priority::Data, - } - } - #[classattr] - fn DataHigh() -> Self { - Priority { - p: zenoh::prelude::Priority::DataHigh, - } - } - #[classattr] - fn DataLow() -> Self { - Priority { - p: zenoh::prelude::Priority::DataLow, - } - } - #[classattr] - fn InteractiveHigh() -> Self { - Priority { - p: zenoh::prelude::Priority::InteractiveHigh, - } - } - #[classattr] - fn InteractiveLow() -> Self { - Priority { - p: zenoh::prelude::Priority::InteractiveLow, - } - } - #[classattr] - fn RealTime() -> Self { - Priority { - p: zenoh::prelude::Priority::RealTime, - } - } -} diff --git a/src/value.rs b/src/value.rs new file mode 100644 index 00000000..2fa29c44 --- /dev/null +++ b/src/value.rs @@ -0,0 +1,332 @@ +// Copyright (c) 2017, 2022 ZettaScale Technology Inc. + +// This program and the accompanying materials are made available under the +// terms of the Eclipse Public License 2.0 which is available at +// http://www.eclipse.org/legal/epl-2.0, or the Apache License, Version 2.0 +// which is available at https://www.apache.org/licenses/LICENSE-2.0. + +// SPDX-License-Identifier: EPL-2.0 OR Apache-2.0 + +// Contributors: +// ZettaScale Zenoh team, <[email protected]> + +use pyo3::{prelude::, types::PyBytes}; +use uhlc::Timestamp; +use zenoh::{ + prelude::{Encoding, KeyExpr, Sample, Value, ZenohId}, + query::Reply, + scouting::Hello, +}; +use zenoh_buffers::{SplitBuffer, ZBuf}; + +use crate::{ + enums::{_Encoding, _SampleKind}, + keyexpr::_KeyExpr, + ToPyErr, +}; + +#[derive(Clone)] +pub(crate) enum Payload { + Zenoh(ZBuf), + Python(Py<PyBytes>), +} +impl Payload { + pub(crate) fn into_zbuf(self) -> ZBuf { + match self { + Payload::Zenoh(buf) => buf, + Payload::Python(buf) => Python::with_gil(\|py\| ZBuf::from(buf.as_bytes(py).to_owned())), + } + } + pub(crate) fn into_pybytes(self) -> Py<PyBytes> { + match self { + Payload::Zenoh(buf) => { + Python::with_gil(\|py\| Py::from(PyBytes::new(py, buf.contiguous().as_ref()))) + } + Payload::Python(buf) => buf, + } + } +} +impl From<ZBuf> for Payload { + fn from(buf: ZBuf) -> Self { + Payload::Zenoh(buf) + } +} +impl From<Py<PyBytes>> for Payload { + fn from(buf: Py<PyBytes>) -> Self { + Payload::Python(buf) + } +} +#[pyclass(subclass)] +#[derive(Clone)] +pub struct _Value { + pub(crate) payload: Payload, + pub(crate) encoding: Encoding, +} +#[pymethods] +impl _Value { + #[new] + pub fn pynew(this: Self) -> Self { + this + } + #[staticmethod] + pub fn new(payload: Py<PyBytes>, encoding: Option<_Encoding>) -> Self { + Self { + payload: payload.into(), + encoding: encoding.map(\|e\| e.0).unwrap_or(Encoding::EMPTY), + } + } + #[getter] + pub fn payload(&mut self) -> Py<PyBytes> { + if let Payload::Python(buf) = &self.payload { + return buf.clone(); + } + let payload = unsafe { std::ptr::read(&self.payload) }; + let buf = payload.into_pybytes(); + unsafe { std::ptr::write(&mut self.payload, Payload::Python(buf.clone())) }; + buf + } + pub fn with_payload(&mut self, payload: Py<PyBytes>) { + self.payload = Payload::Python(payload) + } + #[getter] + pub fn encoding(&self) -> _Encoding { + _Encoding(self.encoding.clone()) + } + pub fn with_encoding(&mut self, encoding: _Encoding) { + self.encoding = encoding.0; + } +} +impl From<Value> for _Value { + fn from(value: Value) -> Self { + _Value { + payload: value.payload.into(), + encoding: value.encoding, + } + } +} +impl From<_Value> for Value { + fn from(value: _Value) -> Self { + Value::new(value.payload.into_zbuf()).encoding(value.encoding) + } +} + +pub(crate) trait PyAnyToValue { + fn to_value(self) -> PyResult<Value>; +} +impl PyAnyToValue for &PyAny { + fn to_value(self) -> PyResult<Value> { + let encoding: _Encoding = self.getattr("encoding")?.extract()?; + let payload: &PyBytes = self.getattr("payload")?.extract()?; + Ok(Value::new(ZBuf::from(payload.as_bytes().to_owned())).encoding(encoding.0)) + } +} + +#[pyclass(subclass)] +#[derive(Clone)] +pub struct _Sample { + key_expr: KeyExpr<'static>, + value: _Value, + kind: _SampleKind, + timestamp: Option<_Timestamp>, +} +impl From<Sample> for _Sample { + fn from(sample: Sample) -> Self { + let Sample { + key_expr, + value, + kind, + timestamp, + .. + } = sample; + _Sample { + key_expr, + value: value.into(), + kind: _SampleKind(kind), + timestamp: timestamp.map(_Timestamp), + } + } +} + +#[pyclass(subclass)] +#[derive(Clone, Copy, PartialEq, Eq)] +pub struct _ZenohId(pub(crate) ZenohId); +#[pymethods] +impl _ZenohId { + #[new] + pub fn pynew(this: Self) -> Self { + this + } + pub fn __str__(&self) -> String { + self.0.to_string() + } +} + +#[pyclass(subclass)] +#[derive(Clone, Copy, PartialEq, Eq, PartialOrd, Ord)] +pub struct _Timestamp(Timestamp); +#[pymethods] +impl _Timestamp { + #[new] + pub fn pynew(this: Self) -> Self { + this + } + crate::derive_richcmp!(); + #[getter] + pub fn seconds_since_unix_epoch(&self) -> PyResult<f64> { + match self + .0 + .get_time() + .to_system_time() + .duration_since(std::time::UNIX_EPOCH) + { + Ok(o) => Ok(o.as_secs_f64()), + Err(e) => Err(e.to_pyerr()), + } + } +} + +#[pymethods] +impl _Sample { + #[new] + pub fn pynew(this: Self) -> Self { + this + } + #[getter] + pub fn value(&self) -> _Value { + self.value.clone() + } + #[getter] + pub fn key_expr(&self) -> _KeyExpr { + _KeyExpr(self.key_expr.clone()) + } + #[getter] + pub fn payload(&mut self) -> Py<PyBytes> { + if let Payload::Python(buf) = &self.value.payload { + return buf.clone(); + } + let payload = unsafe { std::ptr::read(&self.value.payload) }; + let buf = payload.into_pybytes(); + unsafe { std::ptr::write(&mut self.value.payload, Payload::Python(buf.clone())) }; + buf + } + #[getter] + pub fn encoding(&self) -> _Encoding { + _Encoding(self.value.encoding.clone()) + } + #[getter] + pub fn kind(&self) -> _SampleKind { + self.kind.clone() + } + #[getter] + pub fn timestamp(&self) -> Option<_Timestamp> { + self.timestamp + } + #[staticmethod] + pub fn new( + key_expr: _KeyExpr, + value: _Value, + kind: _SampleKind, + timestamp: Option<_Timestamp>, + ) -> Self { + _Sample { + key_expr: key_expr.0, + value, + kind, + timestamp, + } + } +} + +impl From<_Sample> for Sample { + fn from(sample: _Sample) -> Self { + let _Sample { + key_expr, + value, + kind, + timestamp, + } = sample; + let mut sample = Sample::new(key_expr, value); + sample.kind = kind.0; + sample.timestamp = timestamp.map(\|t\| t.0); + sample + } +} + +#[pyclass(subclass)] +#[derive(Clone)] +pub struct _Reply { + #[pyo3(get)] + pub replier_id: _ZenohId, + pub reply: Result<_Sample, _Value>, +} +#[pymethods] +impl _Reply { + #[new] + pub fn pynew(this: Self) -> Self { + this + } + #[getter] + pub fn ok(&self) -> PyResult<_Sample> { + match &self.reply { + Ok(o) => Ok(o.clone()), + Err(_) => Err(zenoh_core::zerror!("Called `Reply.ok` on a non-ok reply.").to_pyerr()), + } + } + #[getter] + pub fn err(&self) -> PyResult<_Value> { + match &self.reply { + Err(o) => Ok(o.clone()), + Ok(_) => Err(zenoh_core::zerror!("Called `Reply.err` on a non-err reply.").to_pyerr()), + } + } +} +impl From<Reply> for _Reply { + fn from(reply: Reply) -> Self { + _Reply { + replier_id: _ZenohId(reply.replier_id), + reply: match reply.sample { + Ok(o) => Ok(o.into()), + Err(e) => Err(e.into()), + }, + } + } +} + +#[pyclass(subclass)] +#[derive(Clone)] +pub struct _Hello(pub(crate) Hello); +#[pymethods] +impl _Hello { + #[new] + pub fn pynew(this: Self) -> Self { + this + } + #[getter] + pub fn zid(&self) -> Option<_ZenohId> { + self.0.zid.map(_ZenohId) + } + #[getter] + pub fn whatami(&self) -> Option<&'static str> { + match self.0.whatami { + Some(zenoh::config::whatami::WhatAmI::Client) => Some("client"), + Some(zenoh::config::whatami::WhatAmI::Peer) => Some("peer"), + Some(zenoh::config::whatami::WhatAmI::Router) => Some("router"), + None => None, + } + } + #[getter] + pub fn locators(&self) -> Vec<String> { + match &self.0.locators { + Some(locators) => locators.iter().map(\|l\| l.to_string()).collect(), + None => Vec::new(), + } + } + pub fn __str__(&self) -> String { + self.0.to_string() + } +} +impl From<Hello> for _Hello { + fn from(h: Hello) -> Self { + _Hello(h) + } +} diff --git a/zenoh/__init__.py b/zenoh/__init__.py new file mode 100644 index 00000000..266d4c60 --- /dev/null +++ b/zenoh/__init__.py @@ -0,0 +1,44 @@ +# +# Copyright (c) 2022 ZettaScale Technology +# +# This program and the accompanying materials are made available under the +# terms of the Eclipse Public License 2.0 which is available at +# http://www.eclipse.org/legal/epl-2.0, or the Apache License, Version 2.0 +# which is available at https://www.apache.org/licenses/LICENSE-2.0. +# +# SPDX-License-Identifier: EPL-2.0 OR Apache-2.0 +# +# Contributors: +# ZettaScale Zenoh Team, <[email protected]> +# +from .zenoh import init_logger, scout as _scout +from .keyexpr import IntoKeyExpr, IntoSelector, KeyExpr, Selector +from .config import Config +from .session import Session, Publisher, Subscriber, PullSubscriber, Info +from .enums import CongestionControl, Encoding, Priority, QueryConsolidation, QueryTarget, Reliability, SampleKind +from .value import Hello, Value, IntoValue, IValue, Sample, IntoSample, ZenohId, Timestamp, Reply +from .closures import Closure, IClosure, IntoClosure, Handler, IHandler, IntoHandler, ListCollector, Queue +from .queryable import Queryable, Query +from typing import Any + +def open(args, *kwargs): + return Session(args, *kwargs) + +class Scout: + def __init__(self, inner, receiver): + self._inner_ = inner + self.receiver = receiver + + def stop(self): + self._inner_ = None + +def scout(handler: IntoHandler[Hello, Any, Any] = None, what: str = None, config: Config = None, timeout=None): + from threading import Timer + if handler is None: + handler = ListCollector() + handler = Handler(handler, lambda x: Hello._upgrade_) + scout = _scout(handler.closure, config, what) + scout = Scout(scout, handler.receiver) + if timeout: + Timer(timeout, lambda: scout.stop()).start() + return scout \ No newline at end of file diff --git a/zenoh/closures.py b/zenoh/closures.py new file mode 100644 index 00000000..db2ccbb7 --- /dev/null +++ b/zenoh/closures.py @@ -0,0 +1,268 @@ +# +# Copyright (c) 2022 ZettaScale Technology +# +# This program and the accompanying materials are made available under the +# terms of the Eclipse Public License 2.0 which is available at +# http://www.eclipse.org/legal/epl-2.0, or the Apache License, Version 2.0 +# which is available at https://www.apache.org/licenses/LICENSE-2.0. +# +# SPDX-License-Identifier: EPL-2.0 OR Apache-2.0 +# +# Contributors: +# ZettaScale Zenoh Team, <[email protected]> +# +import abc +from typing import Generic, Callable, Union, Any, TypeVar, Tuple, List +from threading import Condition, Thread +from collections import deque +import time + +In = TypeVar("In") +Out = TypeVar("Out") +Receiver = TypeVar("Receiver") +CallbackCall = Callable[[In], Out] +CallbackDrop = Callable[[], None] + +class IClosure(Generic[In, Out]): + """ + A Closure is a pair of a `call` function that will be used as a callback, + and a `drop` function that will be called when the closure is destroyed. + """ + @property + @abc.abstractmethod + def call(self) -> Callable[[In], Out]: + """ + Returns the closure's call function as a lambda. + """ + ... + @property + @abc.abstractmethod + def drop(self) -> Callable[[], None]: + """ + Returns the closure's destructor as a lambda. + """ + ... + def __enter__(self): + drop = self.drop + if drop is not None: + drop() + def __exit__(self, args): + drop = self.drop + if drop is not None: + drop() + +class IHandler(Generic[In, Out, Receiver]): + """ + A Handler is a value that may be converted into a callback closure for zenoh to use on one side, while possibly providing a receiver for the data that zenoh would provide through that callback. + """ + @property + @abc.abstractmethod + def closure(self) -> IClosure[In, Out]: + """ + The part of the handler that should be passed as a callback to a zenoh function. + """ + ... + @property + @abc.abstractmethod + def receiver(self) -> Receiver: + "The part of the handler that should be used as the receiver when the handler is channel-like." + ... + +IntoClosure = Union[IHandler[In, Out, Any], IClosure[In, Out], Tuple[CallbackCall, CallbackDrop], CallbackCall] +class Closure(IClosure, Generic[In, Out]): + """ + A Closure is a pair of a `call` function that will be used as a callback, + and a `drop` function that will be called when the closure is destroyed. + """ + def __init__(self, closure: IntoClosure[In, Out], type_adaptor: Callable[[Any], In] = None): + _call_ = None + self._drop_ = lambda: None + if isinstance(closure, IHandler): + closure = closure.closure + # dev-note: do not elif here, the next if will catch the obtained closure. + if isinstance(closure, IClosure): + _call_ = closure.call + self._drop_ = closure.drop + elif isinstance(closure, tuple): + _call_, self._drop_ = closure + elif callable(closure): + _call_ = closure + else: + raise TypeError("Unexpected type as input for zenoh.Closure") + if type_adaptor is not None: + self._call_ = lambda args: _call_(type_adaptor(args)) + else: + self._call_ = _call_ + @property + def call(self) -> Callable[[In], Out]: + return self._call_ + + @property + def drop(self) -> Callable[[], None]: + return self._drop_ + +IntoHandler = Union[IHandler[In, Out, Receiver], IClosure[In, Out], Tuple[IClosure, Receiver], Tuple[CallbackCall,CallbackDrop, Receiver], Tuple[CallbackCall,CallbackDrop], CallbackCall] +class Handler(IHandler, Generic[In, Out, Receiver]): + """ + A Handler is a value that may be converted into a callback closure for zenoh to use on one side, while possibly providing a receiver for the data that zenoh would provide through that callback. + """ + def __init__(self, input: IntoHandler[In, Out, Receiver], type_adaptor: Callable[[Any], In] = None): + self._receiver_ = None + if isinstance(input, IHandler): + self._receiver_ = input.receiver + self._closure_ = input.closure + elif isinstance(input, IClosure): + self._closure_ = input + elif isinstance(input, tuple): + if isinstance(input[0], IClosure): + self._closure_, self._receiver_ = input + elif len(input) == 3: + call, drop, self._receiver_ = input + self._closure_ = (call, drop) + else: + self._closure_ = input + else: + self._closure_ = input + self._closure_ = Closure(self._closure_, type_adaptor) + + @property + def closure(self) -> IClosure[In, Out]: + return self._closure_ + @property + def receiver(self) -> Receiver: + return self._receiver_ + + + +class ListCollector(IHandler[In, None, Callable[[],List[In]]], Generic[In]): + """ + A simple collector that aggregates values into a list. + + When used as a handler, it provides a callback that appends elements to a list, + and provides a function that will await the closing of the callback before returning said list. + """ + def __init__(self, timeout=None): + self._vec_ = [] + self._cv_ = Condition() + self._done_ = False + self.timeout = timeout + + @property + def closure(self): + def call(x): + self._vec_.append(x) + def drop(): + with self._cv_: + self._done_ = True + self._cv_.notify() + return Closure((call, drop)) + + @property + def receiver(self): + def wait(): + with self._cv_: + if not self._done_: + self._cv_.wait(timeout=self.timeout) + return self._vec_ + return wait + +class Queue(IHandler[In, None, 'Queue'], Generic[In]): + """ + A simple single-producer, single-consumer queue implementation. + + When used as a handler, it provides itself as the receiver, and will provide a + callback that appends elements to the queue. + """ + def __init__(self, timeout=None): + self._vec_ = deque() + self._cv_ = Condition() + self._done_ = False + + @property + def closure(self) -> IClosure[In, None]: + def call(x): self.put(x) + def drop(): self.close() + return Closure((call, drop)) + + @property + def receiver(self) -> 'Queue': + return self + + def put(self, value): + """ + Puts one element on the queue. + """ + with self._cv_: + self._vec_.append(value) + self._cv_.notify() + + + def get(self, timeout=None): + """ + Gets one element from the queue. + + Raises a `StopIteration` exception if the queue was closed before the timeout ran out. + Raises a `TimeoutError` if the timeout ran out. + """ + try: + return self._vec_.pop() + except IndexError: + pass + if self._done_: + raise StopIteration() + with self._cv_: + self._cv_.wait(timeout=timeout) + try: + return self._vec_.pop() + except IndexError: + pass + if self._done_: + raise StopIteration() + else: + raise TimeoutError() + + def close(self): + with self._cv_: + self._done_ = True + self._cv_.notify() + + def get_remaining(self, timeout=None) -> List[In]: + """ + Awaits the closing of the queue, returning the remaining queued values in a list. + The values inserted into the queue up until this happens will be available through `get`. + + Raises a `TimeoutError` if the timeout in seconds provided was exceeded before closing. + """ + end = (time.time() + timeout) if timeout is not None else None + while not self._done_: + with self._cv_: + self._cv_.wait(timeout=(timeout - time.time()) if timeout else None) + if self._done_: + return + elif time.time() >= end: + raise TimeoutError() + return list(self._vec_) + + def __iter__(self): + return self + def __next__(self): + return self.get() + +if __name__ == "__main__": + def get(collector): + import time + def target(): + with Closure(collector) as closure: + closure = Closure(collector) + closure.call('hi') + closure.call('there') + Thread(target=target).start() + + collector = ListCollector() + get(collector) + print(collector.receiver()) + assert collector.receiver() == ["hi", "there"] + + print("done") + + diff --git a/zenoh/config.py b/zenoh/config.py new file mode 100644 index 00000000..9f3007e1 --- /dev/null +++ b/zenoh/config.py @@ -0,0 +1,61 @@ +# +# Copyright (c) 2022 ZettaScale Technology +# +# This program and the accompanying materials are made available under the +# terms of the Eclipse Public License 2.0 which is available at +# http://www.eclipse.org/legal/epl-2.0, or the Apache License, Version 2.0 +# which is available at https://www.apache.org/licenses/LICENSE-2.0. +# +# SPDX-License-Identifier: EPL-2.0 OR Apache-2.0 +# +# Contributors: +# ZettaScale Zenoh Team, <[email protected]> +# +from typing import Union +from .zenoh import _Config +import json + +class Config(_Config): + def __init__(self): + super().__init__() + @staticmethod + def from_file(filename: str): + """ + Reads the configuration from a file. + The file's extension must be json, json5 or yaml. + """ + c = super(Config).from_file(filename) + return c + @staticmethod + def from_obj(obj): + """ + Reads the configuration from `obj` as if it was a JSON file. + """ + c = Config.from_json5(json.dumps(obj)) + return c + @staticmethod + def from_json5(json: str): + """ + Reads the configuration from a JSON5 string. + + JSON5 is a superset of JSON, so any JSON string is a valid input for this function. + """ + c = super(Config).from_json5(json) + return c + + def get_json(self, path: str) -> str: + """ + Returns the part of the configuration at `path`, + in a JSON-serialized form. + """ + return super().get_json(path) + + def insert_json5(self, path: str, value: str) -> str: + """ + Inserts the provided value (read from JSON) at the given path in the configuration. + """ + return super().insert_json5(path, value) + +MODE_KEY = "mode" +CONNECT_KEY = "connect/endpoints" +LISTEN_KEY = "listen/endpoints" \ No newline at end of file diff --git a/zenoh/enums.py b/zenoh/enums.py new file mode 100644 index 00000000..48753a68 --- /dev/null +++ b/zenoh/enums.py @@ -0,0 +1,231 @@ +# +# Copyright (c) 2022 ZettaScale Technology +# +# This program and the accompanying materials are made available under the +# terms of the Eclipse Public License 2.0 which is available at +# http://www.eclipse.org/legal/epl-2.0, or the Apache License, Version 2.0 +# which is available at https://www.apache.org/licenses/LICENSE-2.0. +# +# SPDX-License-Identifier: EPL-2.0 OR Apache-2.0 +# +# Contributors: +# ZettaScale Zenoh Team, <[email protected]> +# +from .zenoh import _Encoding, _SampleKind, _CongestionControl, _Priority, _Reliability, _QueryTarget, _QueryConsolidation + +class Priority(_Priority): + """ + The priority of a sending operation. + + They are ordered à la Linux priority: + `Priority.REAL_TIME() < Priority.INTERACTIVE_HIGH() < Priority.INTERACTIVE_LOW() < Priority.DATA() < Priority.BACKGROUND()` + """ + def __new__(cls, inner: _SampleKind): + return super().__new__(cls, inner) + @staticmethod + def REAL_TIME() -> 'Priority': + return Priority(_Priority.REAL_TIME) + @staticmethod + def INTERACTIVE_HIGH() -> 'Priority': + return Priority(_Priority.INTERACTIVE_HIGH) + @staticmethod + def INTERACTIVE_LOW() -> 'Priority': + return Priority(_Priority.INTERACTIVE_LOW) + @staticmethod + def DATA_HIGH() -> 'Priority': + return Priority(_Priority.DATA_HIGH) + @staticmethod + def DATA() -> 'Priority': + return Priority(_Priority.DATA) + @staticmethod + def DATA_LOW() -> 'Priority': + return Priority(_Priority.DATA_LOW) + @staticmethod + def BACKGROUND() -> 'Priority': + return Priority(_Priority.BACKGROUND) + def __eq__(self, other) -> bool: + return super().__eq__(other) + def __ne__(self, other) -> bool: + return not self.__eq__(other) + def __eq__(self, other) -> bool: + return super().__eq__(other) + def __lt__(self, other) -> bool: + return super().__lt__(other) + def __le__(self, other) -> bool: + return super().__le__(other) + def __gt__(self, other) -> bool: + return super().__gt__(other) + def __ge__(self, other) -> bool: + return super().__ge__(other) + +class SampleKind(_SampleKind): + "Similar to an HTTP METHOD: only PUT and DELETE are currently supported." + def __new__(cls, inner: _SampleKind): + return super().__new__(cls, inner) + @staticmethod + def PUT() -> 'SampleKind': + return SampleKind(_SampleKind.PUT) + @staticmethod + def DELETE() -> 'SampleKind': + return SampleKind(_SampleKind.DELETE) + def __eq__(self, other) -> bool: + return super().__eq__(other) + def __ne__(self, other) -> bool: + return not self.__eq__(other) + +class CongestionControl(_CongestionControl): + """ + Defines the network's behaviour regarding a message when heavily congested. + """ + def __new__(cls, inner: _CongestionControl): + return super().__new__(cls, inner) + @staticmethod + def DROP() -> 'CongestionControl': + "Allows the message to be dropped if all buffers are full." + return CongestionControl(_CongestionControl.DROP) + @staticmethod + def BLOCK() -> 'CongestionControl': + """ + Prevents the message from being dropped at all cost. + In the face of heavy congestion on a part of the network, this could result in your publisher node blocking. + """ + return CongestionControl(_CongestionControl.BLOCK) + def __eq__(self, other) -> bool: + return super().__eq__(other) + def __ne__(self, other) -> bool: + return not self.__eq__(other) + +class Encoding(_Encoding): + def __new__(cls, inner: _Encoding): + return super().__new__(cls, inner) + @staticmethod + def from_str(s: str) -> 'Encoding': + return super(Encoding, Encoding).from_str(s) + def append(self, s: str): + super().append(s) + @staticmethod + def EMPTY() -> 'Encoding': + return Encoding(_Encoding.EMPTY ) + @staticmethod + def APP_OCTET_STREAM() -> 'Encoding': + return Encoding(_Encoding.APP_OCTET_STREAM) + @staticmethod + def APP_CUSTOM() -> 'Encoding': + return Encoding(_Encoding.APP_CUSTOM) + @staticmethod + def TEXT_PLAIN() -> 'Encoding': + return Encoding(_Encoding.TEXT_PLAIN) + @staticmethod + def APP_PROPERTIES() -> 'Encoding': + return Encoding(_Encoding.APP_PROPERTIES) + @staticmethod + def APP_JSON() -> 'Encoding': + return Encoding(_Encoding.APP_JSON) + @staticmethod + def APP_SQL() -> 'Encoding': + return Encoding(_Encoding.APP_SQL) + @staticmethod + def APP_INTEGER() -> 'Encoding': + return Encoding(_Encoding.APP_INTEGER) + @staticmethod + def APP_FLOAT() -> 'Encoding': + return Encoding(_Encoding.APP_FLOAT) + @staticmethod + def APP_XML() -> 'Encoding': + return Encoding(_Encoding.APP_XML) + @staticmethod + def APP_XHTML_XML() -> 'Encoding': + return Encoding(_Encoding.APP_XHTML_XML) + @staticmethod + def APP_X_WWW_FORM_URLENCODED() -> 'Encoding': + return Encoding(_Encoding.APP_X_WWW_FORM_URLENCODED) + @staticmethod + def TEXT_JSON() -> 'Encoding': + return Encoding(_Encoding.TEXT_JSON) + @staticmethod + def TEXT_HTML() -> 'Encoding': + return Encoding(_Encoding.TEXT_HTML) + @staticmethod + def TEXT_XML() -> 'Encoding': + return Encoding(_Encoding.TEXT_XML) + @staticmethod + def TEXT_CSS() -> 'Encoding': + return Encoding(_Encoding.TEXT_CSS) + @staticmethod + def TEXT_CSV() -> 'Encoding': + return Encoding(_Encoding.TEXT_CSV) + @staticmethod + def TEXT_JAVASCRIPT() -> 'Encoding': + return Encoding(_Encoding.TEXT_JAVASCRIPT) + @staticmethod + def IMAGE_JPEG() -> 'Encoding': + return Encoding(_Encoding.IMAGE_JPEG) + @staticmethod + def IMAGE_PNG() -> 'Encoding': + return Encoding(_Encoding.IMAGE_PNG) + @staticmethod + def IMAGE_GIF() -> 'Encoding': + return Encoding(_Encoding.IMAGE_GIF) + def __eq__(self, other) -> bool: + return super().__eq__(other) + def __ne__(self, other) -> bool: + return not self.__eq__(other) + +class Reliability(_Reliability): + "Used by subscribers to inform the network of the reliability it wishes to obtain." + def __new__(cls, inner: _Reliability): + return super().__new__(cls, inner) + @staticmethod + def BEST_EFFORT() -> 'CongestionControl': + "Informs the network that dropping some messages is acceptable" + return Reliability(_Reliability.BEST_EFFORT) + @staticmethod + def RELIABLE() -> 'CongestionControl': + """ + Informs the network that this subscriber wishes for all publications to reliably reach it. + + Note that if a publisher puts a sample with the `CongestionControl.DROP()` option, this reliability + requirement may be infringed to prevent slow readers from blocking the network. + """ + return Reliability(_Reliability.RELIABLE) + def __eq__(self, other) -> bool: + return super().__eq__(other) + def __ne__(self, other) -> bool: + return not self.__eq__(other) + +class QueryTarget(_QueryTarget): + def __new__(cls, inner: _QueryTarget): + return super().__new__(cls, inner) + @staticmethod + def BEST_MATCHING() -> 'QueryTarget': + return QueryTarget(_QueryTarget.BEST_MATCHING) + @staticmethod + def ALL() -> 'QueryTarget': + return QueryTarget(_QueryTarget.ALL) + @staticmethod + def ALL_COMPLETE() -> 'QueryTarget': + return QueryTarget(_QueryTarget.ALL_COMPLETE) + def __eq__(self, other) -> bool: + return super().__eq__(other) + def __ne__(self, other) -> bool: + return not self.__eq__(other) + +class QueryConsolidation(_QueryConsolidation): + def __new__(cls, inner: _QueryConsolidation): + return super().__new__(cls, inner) + @staticmethod + def AUTO() -> 'QueryConsolidation': + return QueryConsolidation(_QueryConsolidation.AUTO) + @staticmethod + def NONE() -> 'QueryConsolidation': + return QueryConsolidation(_QueryConsolidation.NONE) + @staticmethod + def MONOTONIC() -> 'QueryConsolidation': + return QueryConsolidation(_QueryConsolidation.MONOTONIC) + @staticmethod + def LATEST() -> 'QueryConsolidation': + return QueryConsolidation(_QueryConsolidation.LATEST) + def __eq__(self, other) -> bool: + return super().__eq__(other) + def __ne__(self, other) -> bool: + return not self.__eq__(other) \ No newline at end of file diff --git a/zenoh/keyexpr.py b/zenoh/keyexpr.py new file mode 100644 index 00000000..b308e232 --- /dev/null +++ b/zenoh/keyexpr.py @@ -0,0 +1,182 @@ +# +# Copyright (c) 2022 ZettaScale Technology +# +# This program and the accompanying materials are made available under the +# terms of the Eclipse Public License 2.0 which is available at +# http://www.eclipse.org/legal/epl-2.0, or the Apache License, Version 2.0 +# which is available at https://www.apache.org/licenses/LICENSE-2.0. +# +# SPDX-License-Identifier: EPL-2.0 OR Apache-2.0 +# +# Contributors: +# ZettaScale Zenoh Team, <[email protected]> +# +from typing import Union, Dict +from .zenoh import _KeyExpr, _Selector + +IntoKeyExpr = Union['KeyExpr', _KeyExpr, str] + +class KeyExpr(_KeyExpr): + """ + Zenoh's address space is designed around keys which serve as the names of ressources. + + Keys are slash-separated lists of non-empty UTF8 strings. They may not contain the following characters: `$#?`. + + Zenoh's operations are executed on key expressions, a small language that allows the definition + of sets of keys via the use of wildcards: + - `` is the single-chunk wildcard, and will match any chunk: "a//c" will match "a/b/c", "a/hello/c", etc... + - `` is the 0 or more chunks wildcard: "a//c" matches "a/c", "a/b/c", "a/b/hello/c"... + - `$` is the subchunk wildcard, it will match any amount of non-/ characters: "a/b$" matches "a/b", "a/because", "a/blue"... but not "a/c" nor "a/blue/c" + + To allow for better performance and gain the property that two key expressions define the same + set if and only if they are the same string, the rules of canon form are mandatory for a key + expression to be propagated by a Zenoh network: + - `/` may not exist, as it could always be replaced by the shorter ``, + - `/` may not exist, and must be written as its equivalent `/` instead, + - `$` may not exist alone in a chunk, as it must be written `` instead. + + The `KeyExpr.autocanonize` constructor exists to correct eventual infrigements of the canonization rules. + + A KeyExpr is a string that has been validated to be a valid Key Expression. + """ + def __new__(cls, expr: IntoKeyExpr): + """ + The default constructor for KeyExpr will ensure that the passed expression is valid. + It won't however try to correct expressions that aren't canon. + + You may use `KeyExpr.autocanonize(expr)` instead if you are unsure if the expression + you will use for construction will be canon. + + Raises a zenoh.ZError exception if `expr` is not a valid key expression. + """ + if isinstance(expr, KeyExpr): + return expr + elif isinstance(expr, _KeyExpr): + return _KeyExpr.__new__(cls, expr) + else: + return _KeyExpr.__new__(cls, _KeyExpr.new(expr)) + + def _upgrade_(this: _KeyExpr) -> 'KeyExpr': + return _KeyExpr.__new__(KeyExpr, expr) + + @staticmethod + def autocanonize(expr: str) -> 'KeyExpr': + """ + This alternative constructor for key expressions will attempt to canonize the passed + expression before checking if it is valid. + + Raises a zenoh.ZError exception if `expr` is not a valid key expression. + """ + if isinstance(expr, KeyExpr): + return expr + else: + e = _KeyExpr.autocanonize(expr) + return KeyExpr(e.as_str()) + + def intersects(self, other: 'KeyExpr') -> bool: + """ + This method returns `True` if there exists at least one key that belongs to both sets + defined by `self` and `other`. + """ + return super().intersects(other) + + def includes(self, other: 'KeyExpr') -> bool: + """ + This method returns `True` if all of the keys defined by `other` also belong to the set + defined by `self`. + """ + return super().includes(other) + + def undeclare(self, session: 'Session'): + """ + Undeclares a key expression previously declared on the session. + """ + super().undeclare(session) + + def __eq__(self, other: 'KeyExpr') -> bool: + """ + Corresponds to set equality. + """ + return super().__eq__(other) + + def __truediv__(self, other: IntoKeyExpr) -> 'KeyExpr': + """ + Joins two key expressions with a `/`. + + Raises a zenoh.ZError exception if `other` is not a valid key expression. + """ + return KeyExpr.autocanonize(f"{self}/{other}") + + def __str__(self): + return super().__str__() + + def __hash__(self): + return super().__hash__() + +IntoSelector = Union['Selector', _Selector, IntoKeyExpr] +class Selector(_Selector): + """ + A selector is the combination of a [Key Expression](crate::prelude::KeyExpr), which defines the + set of keys that are relevant to an operation, and a `parameters`, a set of key-value pairs + with a few uses: + specifying arguments to a queryable, allowing the passing of Remote Procedure Call parameters + * filtering by value, + * filtering by metadata, such as the timestamp of a value, + + When in string form, selectors look a lot like a URI, with similar semantics: + * the `key_expr` before the first `?` must be a valid key expression. + * the `parameters` after the first `?` should be encoded like the query section of a URL: + * key-value pairs are separated by `&`, + * the key and value are separated by the first `=`, + * in the absence of `=`, the value is considered to be the empty string, + * both key and value should use percent-encoding to escape characters, + * defining a value for the same key twice is considered undefined behavior. + + Zenoh intends to standardize the usage of a set of keys. To avoid conflicting with RPC parameters, + the Zenoh team has settled on reserving the set of keys that start with non-alphanumeric characters. + + This document will summarize the standardized keys for which Zenoh provides helpers to facilitate + coherent behavior for some operations. + + Queryable implementers are encouraged to prefer these standardized keys when implementing their + associated features, and to prefix their own keys to avoid having conflicting keys with other + queryables. + + Here are the currently standardized keys for Zenoh: + * `_time`: used to express interest in only values dated within a certain time range, values for + this key must be readable by the [Zenoh Time DSL](zenoh_util::time_range::TimeRange) for the value to be considered valid. + * `_filter`: TBD Zenoh intends to provide helper tools to allow the value associated with + this key to be treated as a predicate that the value should fulfill before being returned. + A DSL will be designed by the Zenoh team to express these predicates. + """ + def __new__(cls, selector: IntoSelector): + if isinstance(selector, Selector): + return selector + if isinstance(selector, _Selector): + return Selector._upgrade_(selector) + return Selector._upgrade_(super().new(str(selector))) + @staticmethod + def _upgrade_(this: _Selector) -> 'Selector': + return _Selector.__new__(Selector, this) + @property + def key_expr(self) -> KeyExpr: + "The key expression part of the selector." + return KeyExpr(super().key_expr) + @property + def parameters(self): + "The value selector part of the selector." + return super().parameters + @parameters.setter + def set_parameters(self, parameters: str): + super().parameters = parameters + def decode_parameters(self) -> Dict[str, str]: + """ + Decodes the value selector part of the selector. + + Raises a ZError if some keys were duplicated: duplicated keys are considered undefined behaviour, + but we encourage you to refuse to process incoming messages with duplicated keys, as they might be + attempting to use HTTP Parameter Pollution like exploits. + """ + return super().decode_parameters() + def __str__(self): + return super().__str__() \ No newline at end of file diff --git a/zenoh/queryable.py b/zenoh/queryable.py new file mode 100644 index 00000000..774ad582 --- /dev/null +++ b/zenoh/queryable.py @@ -0,0 +1,70 @@ +# +# Copyright (c) 2022 ZettaScale Technology +# +# This program and the accompanying materials are made available under the +# terms of the Eclipse Public License 2.0 which is available at +# http://www.eclipse.org/legal/epl-2.0, or the Apache License, Version 2.0 +# which is available at https://www.apache.org/licenses/LICENSE-2.0. +# +# SPDX-License-Identifier: EPL-2.0 OR Apache-2.0 +# +# Contributors: +# ZettaScale Zenoh Team, <[email protected]> +# +from typing import Dict + +from .zenoh import _Query, _Queryable +from .keyexpr import KeyExpr, Selector +from .value import Sample + +class Queryable: + """ + A handle to a queryable. + + Its main purpose is to keep the queryable active as long as it exists. + + When constructed through `Session.declare_queryable(session, keyexpr, handler)`, it exposes `handler`'s receiver + through `self.receiver`. + """ + def __init__(self, inner: _Queryable, receiver): + self._inner_ = inner + self.receiver = receiver + + def undeclare(self): + "Stops the queryable." + self._inner_ = None + +class Query(_Query): + def __new__(cls, inner: _Query): + return super().__new__(cls, inner) + @property + def key_expr(self) -> KeyExpr: + "The query's targeted key expression" + return KeyExpr(super().key_expr) + @property + def parameters(self) -> str: + """ + The query's value selector. + If you'd rather not bother with parsing it yourself, use `self.decode_parameters()` instead. + """ + return super().parameters + + def decode_parameters(self) -> Dict[str, str]: + """ + Decodes the value selector into a dictionary. + + Raises a ZError if duplicate keys are found, as they might otherwise be used for HTTP Parameter Pollution like attacks. + """ + return super().decode_parameters + @property + def selector(self) -> Selector: + """ + The query's selector as a whole. + """ + return Selector._upgrade_(super().selector) + def reply(self, sample: Sample): + """ + Allows you to reply to a query. + You may send any amount of replies to a single query, including 0. + """ + super().reply(sample) \ No newline at end of file diff --git a/zenoh/session.py b/zenoh/session.py new file mode 100644 index 00000000..17271784 --- /dev/null +++ b/zenoh/session.py @@ -0,0 +1,245 @@ +# +# Copyright (c) 2022 ZettaScale Technology +# +# This program and the accompanying materials are made available under the +# terms of the Eclipse Public License 2.0 which is available at +# http://www.eclipse.org/legal/epl-2.0, or the Apache License, Version 2.0 +# which is available at https://www.apache.org/licenses/LICENSE-2.0. +# +# SPDX-License-Identifier: EPL-2.0 OR Apache-2.0 +# +# Contributors: +# ZettaScale Zenoh Team, <[email protected]> +# +from typing import Union, Any, List + +from .zenoh import _Session, _Config, _Publisher, _Subscriber, _PullSubscriber + +from .keyexpr import KeyExpr, IntoKeyExpr, Selector, IntoSelector +from .config import Config +from .closures import IntoHandler, Handler, Receiver +from .enums import * +from .value import IntoValue, Value, Sample, Reply, ZenohId +from .queryable import Queryable, Query + + +class Publisher: + "Use `Publisher`s (constructed with `Session.declare_publisher`) when you want to send values often for the same key expression, as declaring them informs Zenoh that this is you intent, and optimizations will be set up to do so." + + def __init__(self, p: _Publisher): + self._inner_ = p + + def put(self, value: IntoValue, encoding: Encoding = None): + "An optimised version of `session.put(self.key_expr, value, encoding=encoding)" + self._inner_.put(Value(value, encoding)) + + def delete(self): + "An optimised version of `session.delete(self.key_expr)" + self._inner_.delete() + + @property + def key_expr(self) -> KeyExpr: + "This `Publisher`'s key expression" + return KeyExpr(self._inner_.key_expr) + + def undeclare(self): + "Stops the publisher." + self._inner_ = None + + +class Subscriber: + """ + A handle to a subscription. + + Its main purpose is to keep the subscription active as long as it exists. + + When constructed through `Session.declare_subscriber(session, keyexpr, handler)`, it exposes `handler`'s receiver + through `self.receiver`. + """ + + def __init__(self, s: _Subscriber, receiver=None): + self._subscriber_ = s + self.receiver = receiver + + def undeclare(self): + "Undeclares the subscription" + self._subscriber_ = None + + +class PullSubscriber: + """ + A handle to a pull subscription. + + Its main purpose is to keep the subscription active as long as it exists. + + When constructed through `Session.declare_pull_subscriber(session, keyexpr, handler)`, it exposes `handler`'s receiver + through `self.receiver`. + + Calling `self.pull()` will prompt the Zenoh network to send a new sample when available. + """ + + def __init__(self, s: _PullSubscriber, receiver=None): + self._subscriber_ = s + self.receiver = receiver + + def pull(self): + """ + Prompts the Zenoh network to send a new sample if available. + Note that this sample will not be returned by this function, but provided to the handler's callback. + """ + self._subscriber_.pull() + + def undeclare(self): + "Undeclares the subscription" + self._subscriber_ = None + + +class Session(_Session): + """ + A Zenoh Session, the core interraction point with a Zenoh network. + """ + def __new__(cls, config: Union[Config, Any] = None): + if config is None: + return super().__new__(cls) + elif isinstance(config, _Config): + return super().__new__(cls, config) + else: + return super().__new__(cls, Config.from_obj(config)) + + def put(self, keyexpr: IntoKeyExpr, value: IntoValue, encoding=None, + priority: Priority = None, congestion_control: CongestionControl = None, + sample_kind: SampleKind = None): + """ + Sends a value over Zenoh. + """ + value = Value(value, encoding) + keyexpr = KeyExpr(keyexpr) + kwargs = dict() + if priority is not None: + kwargs['priority'] = priority + if congestion_control is not None: + kwargs['congestion_control'] = congestion_control + if sample_kind is not None: + kwargs['sample_kind'] = sample_kind + return super().put(keyexpr, value, kwargs) + + def delete(self, keyexpr: IntoKeyExpr, + priority: Priority = None, congestion_control: CongestionControl = None): + """ + Deletes a value. + """ + keyexpr = KeyExpr(keyexpr) + kwargs = dict() + if priority is not None: + kwargs['priority'] = priority + if congestion_control is not None: + kwargs['congestion_control'] = congestion_control + return super().delete(keyexpr, kwargs) + + def get(self, selector: IntoSelector, handler: IntoHandler[Reply, Any, Receiver], consolidation: QueryConsolidation = None, target: QueryTarget = None) -> Receiver: + """ + Emits a query. + """ + handler = Handler(handler, lambda x: Reply(x)) + kwargs = dict() + if consolidation is not None: + kwargs["conconsolidation"] = consolidation + if target is not None: + kwargs["target"] = target + super().get(Selector(selector), handler.closure, kwargs) + return handler.receiver + + def declare_keyexpr(self, keyexpr: IntoKeyExpr) -> KeyExpr: + """ + Informs Zenoh that you intend to use the provided Key Expression repeatedly. + + This function returns an optimized representation of the passed `keyexpr`. + """ + return KeyExpr(super().declare_keyexpr(KeyExpr(keyexpr))) + + def declare_queryable(self, keyexpr: IntoKeyExpr, handler: IntoHandler[Query, Any, Any], complete: bool = None): + """ + Declares a queryable, which will receive queries intersecting with `keyexpr`. + + These queries are passed to the `handler` as instances of the `Query`class, which lets you respond when applicatble. + + The `handler`'s receiver is returned as the `receiver` field of the return value. + + IMPORTANT: due to how RAII and Python work, you MUST bind this function's return value to a variable in order for it to function as expected. + This is because as soon as a value is no longer referenced in Python, that value's destructor will run, which will undeclare your queryable, stopping it immediately. + """ + handler = Handler(handler, lambda x: Query(x)) + kwargs = dict() + if complete is not None: + kwargs['complete'] = complete + inner = super().declare_queryable(KeyExpr(keyexpr), handler.closure, kwargs) + return Queryable(inner, handler.receiver) + + def declare_publisher(self, keyexpr: IntoKeyExpr, priority: Priority = None, congestion_control: CongestionControl = None): + """ + Declares a publisher, which you may use to send values repeatedly onto a same key expression. + """ + kwargs = dict() + if priority is not None: + kwargs['priority'] = priority + if congestion_control is not None: + kwargs['congestion_control'] = congestion_control + return Publisher(super().declare_publisher(KeyExpr(keyexpr), kwargs)) + + def declare_subscriber(self, keyexpr: IntoKeyExpr, handler: IntoHandler[Sample, Any, Any], reliability: Reliability = None) -> Subscriber: + """ + Declares a subscriber, which will receive any published sample with a key expression intersecting `keyexpr`. + + These samples are passed to the `handler`'s closure as instances of the `Sample` class. + + The `handler`'s receiver is returned as the `receiver` field of the return value. + + IMPORTANT: due to how RAII and Python work, you MUST bind this function's return value to a variable in order for it to function as expected. + This is because as soon as a value is no longer referenced in Python, that value's destructor will run, which will undeclare your subscriber, deactivating the subscription immediately. + """ + handler = Handler(handler, lambda x: Sample._upgrade_(x)) + kwargs = dict() + if reliability is not None: + kwargs['reliability'] = reliability + s = super().declare_subscriber(KeyExpr(keyexpr), handler.closure, kwargs) + return Subscriber(s, handler.receiver) + + def declare_pull_subscriber(self, keyexpr: IntoKeyExpr, handler: IntoHandler[Sample, Any, Any], reliability: Reliability = None) -> PullSubscriber: + """ + Declares a pull-mode subscriber, which will receive a single published sample with a key expression intersecting `keyexpr` any time its `pull` method is called. + + These samples are passed to the `handler`'s closure as instances of the `Sample` class. + + The `handler`'s receiver is returned as the `receiver` field of the return value. + """ + handler = Handler(handler, lambda x: Sample._upgrade_(x)) + kwargs = dict() + if reliability is not None: + kwargs['reliability'] = reliability + s = super().declare_pull_subscriber(KeyExpr(keyexpr), handler.closure, **kwargs) + return PullSubscriber(s, handler.receiver) + + def close(self): + "Closes the Session" + pass + + def info(self): + "Returns an accessor for informations about this Session" + return Info(self) + + +class Info: + def __init__(self, session: _Session): + self.session = session + + def zid(self) -> ZenohId: + "Returns this Zenoh Session's identifier" + return ZenohId._upgrade_(self.session.zid()) + + def routers_zid(self) -> List[ZenohId]: + "Returns the neighbooring routers' identifiers" + return [ZenohId._upgrade_(id) for id in self.session.routers_zid()] + + def peers_zid(self) -> List[ZenohId]: + "Returns the neighbooring peers' identifiers" + return [ZenohId._upgrade_(id) for id in self.session.peers_zid()] diff --git a/zenoh/value.py b/zenoh/value.py new file mode 100644 index 00000000..1bc6c44a --- /dev/null +++ b/zenoh/value.py @@ -0,0 +1,205 @@ +# +# Copyright (c) 2022 ZettaScale Technology +# +# This program and the accompanying materials are made available under the +# terms of the Eclipse Public License 2.0 which is available at +# http://www.eclipse.org/legal/epl-2.0, or the Apache License, Version 2.0 +# which is available at https://www.apache.org/licenses/LICENSE-2.0. +# +# SPDX-License-Identifier: EPL-2.0 OR Apache-2.0 +# +# Contributors: +# ZettaScale Zenoh Team, <[email protected]> +# +import abc +from typing import Union, Tuple, Optional, List +import json + +from .enums import Encoding, SampleKind +from .zenoh import _Value, _Encoding, _Sample, _SampleKind, _Reply, _ZenohId, _Timestamp, _Hello +from .keyexpr import KeyExpr, IntoKeyExpr + +class IValue: + "The IValue interface exposes how to recover a value's payload in a binary-serialized format, as well as that format's encoding." + @property + @abc.abstractmethod + def payload(self) -> bytes: + ... + + @property + @abc.abstractmethod + def encoding(self) -> Encoding: + ... + +IntoValue = Union[IValue, bytes, str, int, float, object] + +class Value(_Value, IValue): + """ + A Value is a pair of a binary payload, and a mime-type-like encoding string. + + When constructed with `encoding==None`, the encoding will be selected depending on the payload's type. + """ + def __new__(cls, payload: IntoValue, encoding: Encoding=None): + if encoding is None: + if isinstance(payload, Value): + return payload + return Value.autoencode(payload) + else: + if not isinstance(payload, bytes): + raise TypeError("`encoding` was passed, but `payload` is not of type `bytes`") + return Value.new(payload, encoding) + + @staticmethod + def autoencode(value: IntoValue) -> 'Value': + if isinstance(value, IValue): + return Value.new(value.payload, value.encoding) + if isinstance(value, bytes): + return Value.new(value, Encoding.APP_OCTET_STREAM()) + if isinstance(value, str): + return Value.new(value.encode(), Encoding.TEXT_PLAIN()) + if isinstance(value, int): + return Value.new(f"{value}".encode(), Encoding.APP_INTEGER()) + if isinstance(value, float): + return Value.new(f"{value}".encode(), Encoding.APP_FLOAT()) + return Value.new(json.dumps(value).encode(), Encoding.APP_JSON()) + + @staticmethod + def new(payload: bytes, encoding: Encoding = None) -> 'Value': + return Value._upgrade_(_Value.new(payload, encoding)) + + @property + def payload(self) -> bytes: + return super().payload + + @payload.setter + def payload(self, payload: bytes): + super().with_payload(payload) + + @property + def encoding(self) -> Encoding: + return Encoding(super().encoding) + + @encoding.setter + def encoding(self, encoding: Encoding): + super().with_encoding(encoding) + + @staticmethod + def _upgrade_(inner: _Value) -> 'Value': + if isinstance(inner, Value): + return inner + return _Value.__new__(Value, inner) + +class ZenohId(_ZenohId): + """A Zenoh UUID""" + @staticmethod + def _upgrade_(this: _ZenohId) -> 'ZenohId': + return _ZenohId.__new__(ZenohId, this) + def __str__(self) -> str: + return super().__str__() + def __repr__(self) -> str: + return str(self) + +class Timestamp(_Timestamp): + """ + A timestamp taken from the Zenoh HLC (Hybrid Logical Clock). + + These timestamps are guaranteed to be unique, as each machine annotates its perceived time with a UUID, which is used as the least significant part of the comparison operation. + """ + @staticmethod + def _upgrade_(this: _Timestamp) -> 'Timestamp': + return _Timestamp.__new__(Timestamp, this) + @property + def seconds_since_unix_epoch(self) -> float: + """ + Returns the number of seconds since the Unix Epoch. + + You shouldn't use this for comparison though, and rely on comparison operators between members of this class. + """ + return super().seconds_since_unix_epoch + + +IntoSample = Union[_Sample, Tuple[IntoKeyExpr, IntoValue, SampleKind], Tuple[KeyExpr, IntoValue]] +class Sample(_Sample): + """ + A KeyExpr-Value pair, annotated with the kind (PUT or DELETE) of publication used to emit it and a timestamp. + """ + def __new__(cls, key: IntoKeyExpr, value: IntoValue, kind: SampleKind = None, timestamp: Timestamp = None): + kind = _SampleKind.PUT if kind is None else kind + return Sample._upgrade_(super().new(KeyExpr(key), Value(value), kind, timestamp)) + @property + def key_expr(self) -> KeyExpr: + "The sample's key expression" + return KeyExpr(super().key_expr) + @property + def value(self) -> Value: + "The sample's value" + return Value._upgrade_(super().value) + @property + def payload(self) -> bytes: + "A shortcut to `self.value.payload`" + return super().payload + @property + def encoding(self) -> Encoding: + "A shortcut to `self.value.encoding`" + return Encoding(super().encoding) + @property + def kind(self) -> SampleKind: + "The sample's kind" + return SampleKind(super().kind) + @property + def timestamp(self) -> Optional[Timestamp]: + "The sample's timestamp. May be None." + ts = super().timestamp + return None if ts is None else Timestamp._upgrade_(ts) + @staticmethod + def _upgrade_(inner: _Sample) -> 'Sample': + if isinstance(inner, Sample): + return inner + return _Sample.__new__(Sample, inner) + +class Reply(_Reply): + def __new__(cls, inner: _Reply): + return super().__new__(cls, inner) + @property + def replier_id(self) -> ZenohId: + "The reply's sender's id." + return ZenohId._upgrade_(super().replier_id) + @property + def ok(self) -> Sample: + """ + The reply's inner data sample. + + Raises a ZError if the `self` is actually an `err` reply. + """ + return Sample._upgrade_(super().ok) + @property + def err(self) -> Value: + """ + The reply's error value. + + Raises a ZError if the `self` is actually an `ok` reply. + """ + return Value._upgrade_(super().err) + +class Hello(_Hello): + "Represents a single Zenoh node discovered through scouting." + @property + def zid(self) -> ZenohId: + "The node's Zenoh UUID." + zid = super().zid + return None if zid is None else ZenohId._upgrade_(zid) + @property + def whatami(self) -> str: + "The node's type, returning either None, 'peer', 'router', or 'client'." + return super().whatami + @property + def locators(self) -> List[str]: + "The locators through which this node may be adressed." + return super().locators + @staticmethod + def _upgrade_(inner: _Hello) -> 'Sample': + if isinstance(inner, Hello): + return inner + return _Hello.__new__(Hello, inner) + def __str__(self): + return super().__str__() \ No newline at end of file	Point to master branch This is a preparatory PR. It will be merged once https://github.com/eclipse-zenoh/zenoh/pull/347 is merged and `Cargo.lock` is synced.		2022-09-22T14:18:12	[]	[]
jefflester/minitrino	jefflester__minitrino-36	a71ce5a2cb0d596e0fdf3d0b419c7ee2be6616c7	diff --git a/cli/minitrino/cmd/cmd_provision.py b/cli/minitrino/cmd/cmd_provision.py index 3b57e653..f47e60d9 100644 --- a/cli/minitrino/cmd/cmd_provision.py +++ b/cli/minitrino/cmd/cmd_provision.py @@ -77,6 +77,7 @@ def cli(ctx, modules, no_rollback, docker_native): utils.check_daemon(ctx.docker_client) utils.check_lib(ctx) + utils.check_starburst_ver(ctx) modules = append_running_modules(modules) check_compatibility(modules) check_enterprise(modules) @@ -391,12 +392,14 @@ def check_dup_configs(ctx): for i, config in enumerate(configs): if config.startswith("#"): continue - config = utils.parse_key_value_pair(config, err_type=err.UserError) + config = utils.parse_key_value_pair( + config, err_type=err.UserError, key_to_upper=False + ) if config is None: continue if i + 1 != len(configs): next_config = utils.parse_key_value_pair( - configs[i + 1], err_type=err.UserError + configs[i + 1], err_type=err.UserError, key_to_upper=False ) if config[0] == next_config[0]: duplicates.extend(["=".join(config), "=".join(next_config)]) @@ -506,7 +509,7 @@ def append_configs(user_configs, current_configs, filename): if filename == TRINO_CONFIG: for user_config in user_configs: user_config = utils.parse_key_value_pair( - user_config, err_type=err.UserError + user_config, err_type=err.UserError, key_to_upper=False ) if user_config is None: continue @@ -514,7 +517,7 @@ def append_configs(user_configs, current_configs, filename): if current_config.startswith("#"): continue current_config = utils.parse_key_value_pair( - current_config, err_type=err.UserError + current_config, err_type=err.UserError, key_to_upper=False ) if current_config is None: continue diff --git a/cli/minitrino/cmd/cmd_snapshot.py b/cli/minitrino/cmd/cmd_snapshot.py index bcecfbbf..5ea62976 100644 --- a/cli/minitrino/cmd/cmd_snapshot.py +++ b/cli/minitrino/cmd/cmd_snapshot.py @@ -205,8 +205,8 @@ def build_command_string(ctx, modules=[]): if modules: options = "" for module in modules: - options += f" {module}" - option_string = f"--module {options}" + options += f"--module {module} " + option_string = f"{options}" bash_source = '"${BASH_SOURCE%/}"' command_string = ( diff --git a/cli/minitrino/utils.py b/cli/minitrino/utils.py index 6bbcbcc8..0f952a24 100644 --- a/cli/minitrino/utils.py +++ b/cli/minitrino/utils.py @@ -240,6 +240,23 @@ def check_lib(ctx): ctx.minitrino_lib_dir +def check_starburst_ver(ctx): + """Checks if a proper Starburst version is provided.""" + + starburst_ver = ctx.env.get_var("STARBURST_VER", "") + error_msg = ( + f"Provided Starburst version '{starburst_ver}' is invalid. " + f"The provided version must be 354-e or higher." + ) + + try: + starburst_ver_int = int(starburst_ver[0:3]) + if starburst_ver_int < 354 or "-e" not in starburst_ver: + raise err.UserError(error_msg) + except: + raise err.UserError(error_msg) + + def generate_identifier(identifiers=None): """Returns an 'object identifier' string used for creating log messages, e.g. '[ID: 12345] [Name: trino]'. @@ -263,7 +280,9 @@ def generate_identifier(identifiers=None): return " ".join(identifier) -def parse_key_value_pair(key_value_pair, err_type=err.MinitrinoError): +def parse_key_value_pair( + key_value_pair, err_type=err.MinitrinoError, key_to_upper=True +): """Parses a key-value pair in string form and returns the resulting pair as both a 2-element list. If the string cannot be split by "=", a MinitrinoError is raised. @@ -273,6 +292,7 @@ def parse_key_value_pair(key_value_pair, err_type=err.MinitrinoError): `"TRINO=354-e"`. - `err_type`: The exception to raise if an "=" delimiter is not in the key-value pair. Defaults to `MinitrinoError`. + - `key_to_upper`: If `True`, the key will be forced to uppercase. ### Return Values - A list `[k, v]`, but will return `None` if the stripped input is an empty @@ -298,6 +318,8 @@ def parse_key_value_pair(key_value_pair, err_type=err.MinitrinoError): key_value_pair[i] = key_value_pair[i].strip() if not key_value_pair[0]: raise err_type(err_msg) + elif key_to_upper: + key_value_pair[0] = key_value_pair[0].upper() if not len(key_value_pair) == 2: raise err_type(err_msg) diff --git a/cli/setup.py b/cli/setup.py index 758a31e9..673d7dac 100644 --- a/cli/setup.py +++ b/cli/setup.py @@ -10,7 +10,7 @@ setup( name="minitrino", - version="2.0.0", + version="2.0.1", description="A command line tool that makes it easy to run modular Trino environments locally.", long_description=README, long_description_content_type="text/markdown", diff --git a/lib/minitrino.env b/lib/minitrino.env index 18716b81..28d63dbd 100644 --- a/lib/minitrino.env +++ b/lib/minitrino.env @@ -2,9 +2,11 @@ COMPOSE_PROJECT_NAME=minitrino STARBURST_VER=354-e +DB2_VER=11.5.4.0 ELASTICSEARCH_VER=7.6.2 MYSQL_VER=5 OPEN_LDAP_VER=1.3.0 +ORACLE_VER=12.2.0.1 POSTGRES_VER=11 POSTGRES_DELTA_LAKE_VER=11 POSTGRES_HIVE_MINIO_VER=11 @@ -13,3 +15,7 @@ POSTGRES_EVENT_LOGGER_VER=11 POSTGRES_RANGER_VER=12.2 SEP_RANGER_VER=2.0.49 SQLSERVER_VER=2017-latest + +METASTORE_HIVE_S3_PORT=9085 +METASTORE_DELTA_LAKE_PORT=9084 +MINIO_DELTA_LAKE_PORT=9001 diff --git a/lib/modules/catalog/db2/db2.yml b/lib/modules/catalog/db2/db2.yml new file mode 100644 index 00000000..77fc2651 --- /dev/null +++ b/lib/modules/catalog/db2/db2.yml @@ -0,0 +1,22 @@ +version: "3.8" +services: + + trino: + volumes: + - "./modules/catalog/db2/resources/trino/db2.properties:/etc/starburst/catalog/db2.properties" + + db2: + image: "ibmcom/db2:${DB2_VER}" + container_name: "db2" + labels: + - "com.starburst.tests=minitrino" + - "com.starburst.tests.module.db2=catalog-db2" + env_file: + - "./modules/catalog/db2/resources/db2/db2.env" + ports: + - "50000:50000" + volumes: + - ./modules/catalog/db2/resources/db2/db2-storage/:/database + privileged: true + tty: true + stdin_open: true diff --git a/lib/modules/catalog/db2/metadata.json b/lib/modules/catalog/db2/metadata.json new file mode 100644 index 00000000..65f17414 --- /dev/null +++ b/lib/modules/catalog/db2/metadata.json @@ -0,0 +1,5 @@ +{ + "description": "Creates a Db2 catalog using the Db2 connector.", + "incompatibleModules": [], + "enterprise": true +} \ No newline at end of file diff --git a/lib/modules/catalog/db2/readme.md b/lib/modules/catalog/db2/readme.md new file mode 100644 index 00000000..020948d2 --- /dev/null +++ b/lib/modules/catalog/db2/readme.md @@ -0,0 +1,10 @@ +# Db2 Connector Module + +This module provisions a standalone Db2 service. + +## Usage + + minitrino --env STARBURST_VER=<ver> provision --module db2 + docker exec -it trino bash + trino-cli + trino> show schemas from db2; diff --git a/lib/modules/catalog/db2/resources/db2/db2-storage/.gitkeep b/lib/modules/catalog/db2/resources/db2/db2-storage/.gitkeep new file mode 100644 index 00000000..e69de29b diff --git a/lib/modules/catalog/db2/resources/db2/db2.env b/lib/modules/catalog/db2/resources/db2/db2.env new file mode 100644 index 00000000..7e304ad5 --- /dev/null +++ b/lib/modules/catalog/db2/resources/db2/db2.env @@ -0,0 +1,3 @@ +DBNAME=minitrino +DB2INST1_PASSWORD=trinoRocks15 +LICENSE=accept diff --git a/lib/modules/catalog/db2/resources/trino/db2.properties b/lib/modules/catalog/db2/resources/trino/db2.properties new file mode 100644 index 00000000..983cd676 --- /dev/null +++ b/lib/modules/catalog/db2/resources/trino/db2.properties @@ -0,0 +1,4 @@ +connector.name=db2 +connection-url=jdbc:db2://db2:50000/minitrino +connection-user=DB2INST1 +connection-password=trinoRocks15 diff --git a/lib/modules/catalog/delta-lake/delta-lake.yml b/lib/modules/catalog/delta-lake/delta-lake.yml index e8dd24b4..e37860d5 100644 --- a/lib/modules/catalog/delta-lake/delta-lake.yml +++ b/lib/modules/catalog/delta-lake/delta-lake.yml @@ -37,13 +37,13 @@ services: "./opt/bin/start-hive-metastore.sh", ] ports: - - "9083:9084" + - "${METASTORE_DELTA_LAKE_PORT}:9083" minio-delta-lake: image: "minio/minio" container_name: "minio-delta-lake" ports: - - "9000:9001" + - "${MINIO_DELTA_LAKE_PORT}:9000" volumes: - "./modules/catalog/delta-lake/resources/minio/data/:/data" labels: diff --git a/lib/modules/catalog/delta-lake/readme.md b/lib/modules/catalog/delta-lake/readme.md index 47b36490..eb6eedbc 100644 --- a/lib/modules/catalog/delta-lake/readme.md +++ b/lib/modules/catalog/delta-lake/readme.md @@ -1,14 +1,20 @@ # Delta-Lake Module + This module uses the Delta Lake connector. There is no Spark backend, so tables -need to be created via CTAS queries from Trino. Example: +need to be created via `CREATE TABLE AS ...` queries from Trino. Example: -``` -CREATE TABLE delta.default.customer -WITH ( - location = 's3a://sample-bucket/default/' -) -AS SELECT FROM tpch.tiny.customer; -``` + CREATE TABLE delta.default.customer + WITH ( + location = 's3a://sample-bucket/default/' + ) + AS SELECT * FROM tpch.tiny.customer; This will create the table `delta.default.customer` and a corresponding `_delta_log` directory in the backing MinIO object storage. + +## Usage + + minitrino --env STARBURST_VER=<ver> provision --module delta-lake + docker exec -it trino bash + trino-cli + trino> show schemas from delta; diff --git a/lib/modules/catalog/elasticsearch/readme.md b/lib/modules/catalog/elasticsearch/readme.md index 4758bc50..6c275a6f 100644 --- a/lib/modules/catalog/elasticsearch/readme.md +++ b/lib/modules/catalog/elasticsearch/readme.md @@ -1,41 +1,49 @@ # Elasticsearch Connector Module -This module contains an ES container with some preloaded data. It contains: a schema (ES mapping), a table (ES doc mapping), and data (ES docs). + +This module contains an ES container with some preloaded data. It contains: a +schema (ES mapping), a table (ES doc mapping), and data (ES docs). ## Loading your own data + Since port 9200 is exposed on localhost you can add your own data like this: -```bash -# Create user index -curl -XPUT http://localhost:9200/user?pretty=true; - -# Create user mapping -curl -XPUT http://localhost:9200/user/_mapping/profile?include_type_name=true -H 'Content-Type: application/json' -d ' -{ - "profile" : { - "properties" : { - "full_name" : { "type" : "text", "store" : true }, - "bio" : { "type" : "text", "store" : true }, - "age" : { "type" : "integer" }, - "location" : { "type" : "geo_point" }, - "enjoys_coffee" : { "type" : "boolean" }, - "created_on" : { - "type" : "date", - "format" : "date_time" + # Create user index + curl -XPUT http://localhost:9200/user?pretty=true; + + # Create user mapping + curl -XPUT http://localhost:9200/user/_mapping/profile?include_type_name=true -H 'Content-Type: application/json' -d ' + { + "profile" : { + "properties" : { + "full_name" : { "type" : "text", "store" : true }, + "bio" : { "type" : "text", "store" : true }, + "age" : { "type" : "integer" }, + "location" : { "type" : "geo_point" }, + "enjoys_coffee" : { "type" : "boolean" }, + "created_on" : { + "type" : "date", + "format" : "date_time" + } } } } -} -'; - -# Create user profile records -curl -XPOST http://localhost:9200/user/profile/1?pretty=true -H 'Content-Type: application/json' -d ' -{ - "full_name" : "Andrew Puch", - "bio" : "My name is Andrew. I have a short bio.", - "age" : 26, - "location" : "41.1246110,-73.4232880", - "enjoys_coffee" : true, - "created_on" : "2015-05-02T14:45:10.000-04:00" -} -'; -``` \ No newline at end of file + '; + + # Create user profile records + curl -XPOST http://localhost:9200/user/profile/1?pretty=true -H 'Content-Type: application/json' -d ' + { + "full_name" : "Andrew Puch", + "bio" : "My name is Andrew. I have a short bio.", + "age" : 26, + "location" : "41.1246110,-73.4232880", + "enjoys_coffee" : true, + "created_on" : "2015-05-02T14:45:10.000-04:00" + } + '; + +## Usage + + minitrino --env STARBURST_VER=<ver> provision --module elasticsearch + docker exec -it trino bash + trino-cli + trino> show schemas from elasticsearch; diff --git a/lib/modules/catalog/hive-minio/readme.md b/lib/modules/catalog/hive-minio/readme.md index 066fb8a1..b189ed58 100644 --- a/lib/modules/catalog/hive-minio/readme.md +++ b/lib/modules/catalog/hive-minio/readme.md @@ -1,18 +1,32 @@ # Hive-Minio Module -This module uses Minio as a local implementation of S3 object storage. You can write data to this service, and the files will be written to your machine. You can read more about Minio [here](https://docs.min.io/docs/minio-docker-quickstart-guide.html). This module also uses a Hive metastore container along with a Postgres container for the metastore's backend storage. -You can access the Minio UI at `http://localhost:9000` with `access-key` and `secret-key` for credentials. +This module uses Minio as a local implementation of S3 server. You can write +data to this service and the files will be written to your machine. You can read +more about Minio +[here](https://docs.min.io/docs/minio-docker-quickstart-guide.html). This module +also uses a Hive metastore container along with a Postgres container for the +metastore's backend storage. + +You can access the Minio UI at `http://localhost:9000` with `access-key` and +`secret-key` for credentials. You can create a table with ORC data with Trino very quickly: -``` -trino> create schema hive_minio.tiny with (location='s3a://sample-bucket/tiny/'); -CREATE SCHEMA + trino> create schema hive_minio.tiny with (location='s3a://sample-bucket/tiny/'); + CREATE SCHEMA -trino> create table hive_minio.tiny.customer as select * from tpch.tiny.customer; -CREATE TABLE: 1500 rows -``` + trino> create table hive_minio.tiny.customer as select * from tpch.tiny.customer; + CREATE TABLE: 1500 rows You will see the ORC data stored in your local Minio bucket. -Note the [relevant commit](https://github.com/starburstdata/docker-images/commit/6b29c2359a173ca6971267fa05191258b1964c8b#diff-8961ce993089ebecf98d0457b676e626) for the property `S3_PATH_STYLE_ACCESS` in `hms.env`. +Note the [relevant +commit](https://github.com/starburstdata/docker-images/commit/6b29c2359a173ca6971267fa05191258b1964c8b#diff-8961ce993089ebecf98d0457b676e626) +for the property `S3_PATH_STYLE_ACCESS` in `hms.env`. + +## Usage + + minitrino --env STARBURST_VER=<ver> provision --module hive-minio + docker exec -it trino bash + trino-cli + trino> show schemas from hive_minio; diff --git a/lib/modules/catalog/hive-s3/hive-s3.yml b/lib/modules/catalog/hive-s3/hive-s3.yml index 930643b6..c3cd19b4 100644 --- a/lib/modules/catalog/hive-s3/hive-s3.yml +++ b/lib/modules/catalog/hive-s3/hive-s3.yml @@ -42,7 +42,7 @@ services: "./opt/bin/start-hive-metastore.sh", ] ports: - - "9083:9085" + - "${METASTORE_HIVE_S3_PORT}:9083" volumes: postgres-hive-s3-data: diff --git a/lib/modules/catalog/hive-s3/readme.md b/lib/modules/catalog/hive-s3/readme.md index b4765a7b..2b182d42 100644 --- a/lib/modules/catalog/hive-s3/readme.md +++ b/lib/modules/catalog/hive-s3/readme.md @@ -1,2 +1,15 @@ # Hive-S3 Module -This module uses a Hive metastore container along with a Postgres container for the metastore's backend storage. Using Minitrino's configuration, you can input AWS credentials to link the module to a an S3 bucket for data storage, reading, and writing. Removing the volume associated with the Postgres container will effectively remove your metastore's database, so only remove the volume if you are prepared to lose all of your metadata. \ No newline at end of file + +This module uses a Hive metastore container along with a Postgres container for +the metastore's backend storage. Using Minitrino's config file, you can input +AWS credentials to link the module to a an S3 bucket for data storage, reading, +and writing. Removing the volume associated with the Postgres container will +effectively remove your metastore's database, so only remove the volume if you +are prepared to lose all of your metadata. + +## Usage + + minitrino --env STARBURST_VER=<ver> provision --module hive-s3 + docker exec -it trino bash + trino-cli + trino> show schemas from hive_s3; diff --git a/lib/modules/catalog/mysql/mysql.yml b/lib/modules/catalog/mysql/mysql.yml index 0968f4c4..37e41f43 100644 --- a/lib/modules/catalog/mysql/mysql.yml +++ b/lib/modules/catalog/mysql/mysql.yml @@ -15,3 +15,5 @@ services: - "./modules/catalog/mysql/resources/mysql/mysql.env" environment: MINITRINO_BOOTSTRAP: "bootstrap.sh" + ports: + - "3306:3306" diff --git a/lib/modules/catalog/mysql/readme.md b/lib/modules/catalog/mysql/readme.md index b71de207..9a5cf1f7 100644 --- a/lib/modules/catalog/mysql/readme.md +++ b/lib/modules/catalog/mysql/readme.md @@ -1,2 +1,11 @@ -# mysql Connector Module -This module provisions a standalone MySQL service. Other modules that uses MySQL as a backend will need a more unique name to avoid conflicts with this one. +# MySQL Connector Module + +This module provisions a standalone MySQL service. Other modules that uses MySQL +as a backend will need a more unique name to avoid conflicts with this one. + +## Usage + + minitrino --env STARBURST_VER=<ver> provision --module mysql + docker exec -it trino bash + trino-cli + trino> show schemas from mysql; diff --git a/lib/modules/catalog/oracle/metadata.json b/lib/modules/catalog/oracle/metadata.json new file mode 100644 index 00000000..0149edc2 --- /dev/null +++ b/lib/modules/catalog/oracle/metadata.json @@ -0,0 +1,5 @@ +{ + "description": "Creates an Oracle catalog using the standard Oracle connector.", + "incompatibleModules": [], + "enterprise": false +} \ No newline at end of file diff --git a/lib/modules/catalog/oracle/oracle.yml b/lib/modules/catalog/oracle/oracle.yml new file mode 100644 index 00000000..2362e40a --- /dev/null +++ b/lib/modules/catalog/oracle/oracle.yml @@ -0,0 +1,19 @@ +version: "3.8" +services: + + trino: + volumes: + - "./modules/catalog/oracle/resources/trino/oracle.properties:/etc/starburst/catalog/oracle.properties" + + oracle: + image: "store/oracle/database-enterprise:${ORACLE_VER}" + container_name: "oracle" + labels: + - "com.starburst.tests=minitrino" + - "com.starburst.tests.module.oracle=catalog-oracle" + environment: + MINITRINO_BOOTSTRAP: "bootstrap-oracle.sh" + env_file: + - "./modules/catalog/oracle/resources/oracle/ora.conf" + ports: + - "1521:1521" diff --git a/lib/modules/catalog/oracle/readme.md b/lib/modules/catalog/oracle/readme.md new file mode 100644 index 00000000..5d2b5ef5 --- /dev/null +++ b/lib/modules/catalog/oracle/readme.md @@ -0,0 +1,27 @@ +# Oracle Connector Module + +This module provisions a standalone Oracle service. + +## Usage + +You will want to use the `trino` schema within the `oracle` catalog. +Additionally, the Oracle server takes some time to become available after the +container boots up. If you see an `Unable to start the Universal Connection +Pool` error, wait 1-2 minutes and try querying Oracle again. + + # Login with the Docker Hub account used to "purchase" the image + echo "your-password" \| docker login -u <user> --password-stdindocker + minitrino --env STARBURST_VER=<ver> provision --module oracle + docker exec -it trino bash + trino-cli + trino> show schemas from oracle; + trino> create table oracle.trino.test (a int); + +## Image Pull Notes + +To pull this image, you will need to: + +- Create a Docker Hub account and then "purchase" the [developer-tier + image](https://hub.docker.com/_/oracle-database-enterprise-edition). +- Authenticate to your Docker Hub account via the Docker CLI (see `docker login + --help` for more information) diff --git a/lib/modules/catalog/oracle/resources/bootstrap/bootstrap-oracle.sh b/lib/modules/catalog/oracle/resources/bootstrap/bootstrap-oracle.sh new file mode 100755 index 00000000..8c039d20 --- /dev/null +++ b/lib/modules/catalog/oracle/resources/bootstrap/bootstrap-oracle.sh @@ -0,0 +1,80 @@ +#!/usr/bin/env bash + +set -ex + +function health_check() { + su - oracle -c "source /home/oracle/.bashrc; sqlplus sys/Oradoc_db1 as sysdba @/tmp/health-check.sql" +} + +function create_user() { + su - oracle -c "source /home/oracle/.bashrc; sqlplus sys/Oradoc_db1 as sysdba @/tmp/create-user.sql" +} + +function grant_privileges() { + su - oracle -c "source /home/oracle/.bashrc; sqlplus sys/Oradoc_db1 as sysdba @/tmp/grant-privileges.sql" +} + +echo "Creating health check SQL script..." +cat <<EOT >> /tmp/health-check.sql +SELECT INSTANCE_NAME, STATUS, DATABASE_STATUS FROM V\$INSTANCE; +exit; +EOT + +echo "Creating user-creation SQL script..." +cat <<EOT >> /tmp/create-user.sql +ALTER SESSION SET "_ORACLE_SCRIPT"=true; +CREATE USER trino IDENTIFIED BY trinoRocks15; +exit; +EOT + +echo "Creating user-privilege SQL script..." +cat <<EOT >> /tmp/grant-privileges.sql +ALTER SESSION SET "_ORACLE_SCRIPT"=true; +GRANT CONNECT, RESOURCE, DBA TO trino; +GRANT CREATE SESSION, CREATE TABLE TO trino; +GRANT UNLIMITED TABLESPACE TO trino; +GRANT ALL PRIVILEGES TO trino; +exit; +EOT + +echo "Performing health checks..." +COUNTER=0 && set +e +while [[ "${COUNTER}" -lt 121 ]]; do + if health_check \| grep -q "trinosid\\|OPEN\\|ACTIVE"; then + break + elif [[ "${COUNTER}" == 121 ]]; then + echo "Database is not up and running and timed out after approx. 2 minutes. Exiting" + exit 1 + else + sleep 1 + ((COUNTER++)) + fi +done + +echo "Creating user..." +COUNTER=0 && set +e +while [[ "${COUNTER}" -lt 61 ]]; do + if create_user \| grep -q "ERROR"; then + sleep 1 + ((COUNTER++)) + elif [[ "${COUNTER}" == 61 ]]; then + echo "User creation failed after approx. 1 minute. Exiting" + exit 1 + else + break + fi +done + +echo "Granting user privileges..." +COUNTER=0 && set +e +while [[ "${COUNTER}" -lt 61 ]]; do + if grant_privileges \| grep -q "ERROR"; then + sleep 1 + ((COUNTER++)) + elif [[ "${COUNTER}" == 61 ]]; then + echo "Privilege grant failed after approx. 1 minute. Exiting" + exit 1 + else + break + fi +done diff --git a/lib/modules/catalog/oracle/resources/oracle/ora.conf b/lib/modules/catalog/oracle/resources/oracle/ora.conf new file mode 100644 index 00000000..64e7c22a --- /dev/null +++ b/lib/modules/catalog/oracle/resources/oracle/ora.conf @@ -0,0 +1,3 @@ +DB_PDB=trino +DB_SID=trinosid +DB_MEMORY=2GB diff --git a/lib/modules/catalog/oracle/resources/trino/oracle.properties b/lib/modules/catalog/oracle/resources/trino/oracle.properties new file mode 100644 index 00000000..0c81e14a --- /dev/null +++ b/lib/modules/catalog/oracle/resources/trino/oracle.properties @@ -0,0 +1,4 @@ +connector.name=oracle +connection-url=jdbc:oracle:thin:@(DESCRIPTION=(ADDRESS=(PROTOCOL=TCP)(HOST=oracle)(PORT=1521))(CONNECT_DATA=(SERVER=DEDICATED)(SERVICE_NAME=trinosid.localdomain))) +connection-user=trino +connection-password=trinoRocks15 diff --git a/lib/modules/catalog/postgres/postgres.yml b/lib/modules/catalog/postgres/postgres.yml index 229ab71e..cb0d2117 100644 --- a/lib/modules/catalog/postgres/postgres.yml +++ b/lib/modules/catalog/postgres/postgres.yml @@ -13,3 +13,5 @@ services: - "com.starburst.tests.module.postgres=catalog-postgres" env_file: - "./modules/catalog/postgres/resources/postgres/postgres.env" + ports: + - "5432:5432" diff --git a/lib/modules/catalog/postgres/readme.md b/lib/modules/catalog/postgres/readme.md index 66b108fb..d3ea8c18 100644 --- a/lib/modules/catalog/postgres/readme.md +++ b/lib/modules/catalog/postgres/readme.md @@ -1,2 +1,12 @@ # Postgres Connector Module -This module provisions a standalone Postgres service. It is named uniquely to avoid conflicts with other modules that may use Trino as a backend, such as the `hive-s3` module. \ No newline at end of file + +This module provisions a standalone Postgres service. It is named uniquely to +avoid conflicts with other modules that may use Trino as a backend, such as the +`hive-s3` module. + +## Usage + + minitrino --env STARBURST_VER=<ver> provision --module postgres + docker exec -it trino bash + trino-cli + trino> show schemas from postgres; diff --git a/lib/modules/catalog/snowflake-distributed/readme.md b/lib/modules/catalog/snowflake-distributed/readme.md index 2e18ff4f..bc50901d 100644 --- a/lib/modules/catalog/snowflake-distributed/readme.md +++ b/lib/modules/catalog/snowflake-distributed/readme.md @@ -1,2 +1,12 @@ # Snowflake Distributed Connector Module -Placeholder readme \ No newline at end of file + +This module hooks up to an externally-hosted Snowflake service and leverages the +[distributed Snowflake +connector](https://docs.starburst.io/latest/connector/starburst-snowflake.html). + +## Usage + + minitrino --env STARBURST_VER=<ver> provision --module snowflake-distributed + docker exec -it trino bash + trino-cli + trino> show schemas from snowflake_distributed; diff --git a/lib/modules/catalog/snowflake-jdbc/readme.md b/lib/modules/catalog/snowflake-jdbc/readme.md index 35b2507d..aa777d58 100644 --- a/lib/modules/catalog/snowflake-jdbc/readme.md +++ b/lib/modules/catalog/snowflake-jdbc/readme.md @@ -1,2 +1,12 @@ # Snowflake JDBC Connector Module -Placeholder readme \ No newline at end of file + +This module hooks up to an externally-hosted Snowflake service and leverages the +[JDBC Snowflake +connector](https://docs.starburst.io/latest/connector/starburst-snowflake.html). + +## Usage + + minitrino --env STARBURST_VER=<ver> provision --module snowflake-jdbc + docker exec -it trino bash + trino-cli + trino> show schemas from snowflake_jdbc; diff --git a/lib/modules/catalog/sqlserver/readme.md b/lib/modules/catalog/sqlserver/readme.md index 5906cca6..cdd757fa 100644 --- a/lib/modules/catalog/sqlserver/readme.md +++ b/lib/modules/catalog/sqlserver/readme.md @@ -1,6 +1,15 @@ # SQL Server Connector Module + This module provisions a standalone SQL Server service. Default database created is `master`. -Note that the 2017 version of SQL Server is used by default, as previous versions were only available on Windows and do not have Docker containers. +Note that the 2017 version of SQL Server is used by default, as previous +versions were only available on Windows and do not have Docker containers. + +## Usage + + minitrino --env STARBURST_VER=<ver> provision --module sqlserver + docker exec -it trino bash + trino-cli + trino> show schemas from sqlserver; diff --git a/lib/modules/security/event-logger/readme.md b/lib/modules/security/event-logger/readme.md index 2e169591..81a3e983 100644 --- a/lib/modules/security/event-logger/readme.md +++ b/lib/modules/security/event-logger/readme.md @@ -1,2 +1,13 @@ # Event Logger Module -Placeholder readme \ No newline at end of file + +This module configures and deploys the necessary components for the [Starburst +event logger](https://docs.starburst.io/latest/security/event-logger.html). This +feature logs query history and persists the data to an external database. This +is a prerequisite for Starburst Insights query history. + +## Usage + + minitrino --env STARBURST_VER=<ver> provision --module event-logger + docker exec -it trino bash + trino-cli + trino> show schemas from postgres_event_logger; diff --git a/lib/modules/security/file-access-control/readme.md b/lib/modules/security/file-access-control/readme.md index a4f896ae..3c87f945 100644 --- a/lib/modules/security/file-access-control/readme.md +++ b/lib/modules/security/file-access-control/readme.md @@ -1,31 +1,18 @@ # File Access Control Module + A module which utilizes Trino's [file-based system access control plugin](https://docs.starburst.io/latest/security/file-system-access-control.html). This also makes used of the [file-based group provider](https://docs.starburst.io/latest/security/group-file.html). -## Sample Usage -To provision this module, run: - -```shell -minitrino provision --module file-access-control -``` - -You will need to supply a username to the Trino CLI in order to map to a group -(see `group.txt` for which users belong to which groups). Example: - -```shell -trino-cli --user admin-2 -trino-cli --user metadata-1 -trino-cli --user platform -``` - ## Policies + The access policy is located in the `rules.json` file which defines groups of users that map to a certain access control permission. The users for the groups are defined in the `group.txt` file. -- Users in the `platform-admins` group have full access to all objects within Trino +- Users in the `platform-admins` group have full access to all objects within + Trino - Users in the `metadata-users` group only have access to the tables within the `system.metadata` schema - Users in the `platform-users` group only have access to the tables within the @@ -33,3 +20,18 @@ are defined in the `group.txt` file. You can modify this module to further specify access control permissions to other catalogs provisioned with other Minitrino modules. + +## Usage + + minitrino --env STARBURST_VER=<ver> provision --module file-access-control + docker exec -it trino bash + trino-cli --user admin + trino> show schemas from tpch; + +You will need to supply a username to the Trino CLI in order to map to a group +(see `lib/modules/security/file-access-control/resources/trino/group.txt` for +which users belong to which groups). Example: + + trino-cli --user admin-2 + trino-cli --user metadata-1 + trino-cli --user platform diff --git a/lib/modules/security/ldap/readme.md b/lib/modules/security/ldap/readme.md index d22547f0..76034daf 100644 --- a/lib/modules/security/ldap/readme.md +++ b/lib/modules/security/ldap/readme.md @@ -1,24 +1,29 @@ # LDAP Module + This module provisions an LDAP server for authenticating users in Trino. This -also enables SSL / TLS between the LDAP server and Trino, and between Trino -and clients. It is compatible with other security modules like system-ranger -and event-logger, but is mutually-exclusive of the password-file module. +also enables SSL / TLS between the LDAP server and Trino, and between Trino and +clients. It is compatible with other security modules like system-ranger and +event-logger, but is mutually-exclusive of the password-file module. -## Requirements -- N/A +## Usage -## Sample Usage -To provision this module, run: + minitrino --env STARBURST_VER=<ver> provision --module ldap + docker exec -it trino bash + + trino-cli --server https://trino:8443 \ + --truststore-path /etc/starburst/ssl/truststore.jks --truststore-password trinoRocks15 \ + --keystore-path /etc/starburst/ssl/keystore.jks --keystore-password trinoRocks15 \ + --user bob --password -```shell -minitrino provision --module ldap -``` + trino> show schemas from tpch; ## Default Usernames and Passwords + - alice / trinoRocks15 - bob / trinoRocks15 ## Client Keystore and Truststore + The Java keystore and truststore needed for clients and drivers to securely connect to Trino are located in a volume mount `~/.minitrino/ssl`. These two files are transient and will be automatically replaced whenever Minitrino is @@ -28,27 +33,24 @@ provisioned with a security module that enables SSL. Via Docker: -``` -docker exec -it trino trino-cli --server https://trino:8443 \ - --truststore-path /etc/starburst/ssl/truststore.jks --truststore-password trinoRocks15 \ - --keystore-path /etc/starburst/ssl/keystore.jks --keystore-password trinoRocks15 \ - --user bob --password -``` + docker exec -it trino trino-cli --server https://trino:8443 \ + --truststore-path /etc/starburst/ssl/truststore.jks --truststore-password trinoRocks15 \ + --keystore-path /etc/starburst/ssl/keystore.jks --keystore-password trinoRocks15 \ + --user bob --password -Via Host Machine: +Via host machine: -``` -trino-cli-xxx-executable.jar --server https://localhost:8443 \ - --truststore-path ~/.minitrino/ssl/truststore.jks --truststore-password trinoRocks15 \ - --keystore-path ~/.minitrino/ssl/keystore.jks --keystore-password trinoRocks15 \ - --user bob --password -``` + trino-cli-xxx-executable.jar --server https://localhost:8443 \ + --truststore-path ~/.minitrino/ssl/truststore.jks --truststore-password trinoRocks15 \ + --keystore-path ~/.minitrino/ssl/keystore.jks --keystore-password trinoRocks15 \ + --user bob --password Note that the CLI will prompt you for the password. ## Accessing the Trino Web UI -Open a web browser and go to https://localhost:8443 and log in with a valid LDAP -username and password. + +Open a web browser and go to <https://localhost:8443> and log in with a valid +LDAP username and password. To have the browser accept the self-signed certificate, do the following: @@ -61,32 +63,28 @@ Risk and Continue. this website. ## Adding a New User to LDAP + 1. Open a shell to the LDAP container -``` -docker exec -it ldap bash -``` + docker exec -it ldap bash 2. Create an LDIF file with the new user information: -``` -cat << EOF > jeff.ldif -dn: uid=jeff,dc=example,dc=com -changetype: add -uid: jeff -objectClass: inetOrgPerson -objectClass: organizationalPerson -objectClass: person -objectClass: top -cn: jeff -sn: jeff -mail: [email protected] -userPassword: trinoRocks15 -EOF -``` + cat << EOF > jeff.ldif + dn: uid=jeff,dc=example,dc=com + changetype: add + uid: jeff + objectClass: inetOrgPerson + objectClass: organizationalPerson + objectClass: person + objectClass: top + cn: jeff + sn: jeff + mail: [email protected] + userPassword: trinoRocks15 + EOF 3. Use the ldapmodify tool to add the new user -``` -ldapmodify -x -D "cn=admin,dc=example,dc=com" -w trinoRocks15 -H ldaps://ldap:636 -f jeff.ldif -``` + ldapmodify -x -D "cn=admin,dc=example,dc=com" \ + -w trinoRocks15 -H ldaps://ldap:636 -f jeff.ldif diff --git a/lib/modules/security/password-file/readme.md b/lib/modules/security/password-file/readme.md index 486f4517..b4f84c98 100644 --- a/lib/modules/security/password-file/readme.md +++ b/lib/modules/security/password-file/readme.md @@ -1,24 +1,29 @@ # Password File Authentication Module -This module configures Trino to authenticate users with a password file. It -also enables SSL / TLS between Trino and clients. It is compatible with other + +This module configures Trino to authenticate users with a password file. It also +enables SSL / TLS between Trino and clients. It is compatible with other security modules like system-ranger and event-logger, but is mutually-exclusive of the ldap module. -## Requirements -- N/A +## Usage -## Sample Usage -To provision this module, run: + minitrino --env STARBURST_VER=<ver> provision --module password-file + docker exec -it trino bash -```shell -minitrino provision --module ldap -``` + trino-cli --server https://trino:8443 \ + --truststore-path /etc/starburst/ssl/truststore.jks --truststore-password trinoRocks15 \ + --keystore-path /etc/starburst/ssl/keystore.jks --keystore-password trinoRocks15 \ + --user bob --password + + trino> show schemas from tpch; ## Default Usernames and Passwords + - alice / trinoRocks15 - bob / trinoRocks15 ## Client Keystore and Truststore + The Java keystore and truststore needed for clients and drivers to securely connect to Trino are located in a volume mount `~/.minitrino/ssl`. These two files are transient and will be automatically replaced whenever Minitrino is @@ -28,26 +33,23 @@ provisioned with a security module that enables SSL. Via Docker: -``` -docker exec -it trino trino-cli --server https://trino:8443 \ - --truststore-path /etc/starburst/ssl/truststore.jks --truststore-password trinoRocks15 \ - --keystore-path /etc/starburst/ssl/keystore.jks --keystore-password trinoRocks15 \ - --user bob --password -``` + docker exec -it trino trino-cli --server https://trino:8443 \ + --truststore-path /etc/starburst/ssl/truststore.jks --truststore-password trinoRocks15 \ + --keystore-path /etc/starburst/ssl/keystore.jks --keystore-password trinoRocks15 \ + --user bob --password Via Host Machine: -``` -trino-cli-xxx-executable.jar --server https://localhost:8443 \ - --truststore-path ~/.minitrino/ssl/truststore.jks --truststore-password trinoRocks15 \ - --keystore-path ~/.minitrino/ssl/keystore.jks --keystore-password trinoRocks15 \ - --user bob --password -``` + trino-cli-xxx-executable.jar --server https://localhost:8443 \ + --truststore-path ~/.minitrino/ssl/truststore.jks --truststore-password trinoRocks15 \ + --keystore-path ~/.minitrino/ssl/keystore.jks --keystore-password trinoRocks15 \ + --user bob --password Note that the CLI will prompt you for the password. ## Accessing the Trino Web UI -Open a web browser and go to https://localhost:8443 and log in with a valid + +Open a web browser and go to <https://localhost:8443> and log in with a valid username and password. To have the browser accept the self-signed certificate, do the following: @@ -64,6 +66,4 @@ this website*. Example with username `jeff` and password `trinoRocks15` -``` -docker exec trino htpasswd -bB -C 10 /etc/starburst/password.db jeff trinoRocks15 -``` + docker exec trino htpasswd -bB -C 10 /etc/starburst/password.db jeff trinoRocks15 diff --git a/lib/modules/security/system-ranger/readme.md b/lib/modules/security/system-ranger/readme.md index d3a7eb86..8da2d428 100644 --- a/lib/modules/security/system-ranger/readme.md +++ b/lib/modules/security/system-ranger/readme.md @@ -1,21 +1,26 @@ # System-Level Ranger Access Control Module -This module provisions a Ranger server and Postgres database for Ranger storage. It is an unsecured deployment without SSL or an authentication mechanism. -## Requirements -- Starburst Data license -- Access to Starburst Data Harbor repository for Docker images +This module provisions a Ranger server and Postgres database for Ranger storage. +It is an unsecured deployment without SSL or an authentication mechanism. + +## Usage + + minitrino --env STARBURST_VER=<ver> provision --module system-ranger + docker exec -it trino bash + trino-cli --user bob + trino> select from tpch.sf100.customer limit 1; -## Sample Usage -To provision this module, run: +## Requirements -```shell -minitrino provision --module system-ranger -``` +- Starburst Data license +- Access to Starburst Data Harbor repository for Docker images ## Policies + - Bob: admin access to TPCH `sf100` schema - Alice: admin access to TPCH `sf10` schema ## Accessing Ranger + - Go to `localhost:6080` -- Sign in with user: admin and pass: trinoRocks15 +- Sign in with user: `admin` and password: `trinoRocks15` diff --git a/lib/version b/lib/version index 227cea21..38f77a65 100644 --- a/lib/version +++ b/lib/version @@ -1,1 +1,1 @@ -2.0.0 +2.0.1 diff --git a/readme.md b/readme.md index 26f4b278..bed4fcdc 100644 --- a/readme.md +++ b/readme.md @@ -1,40 +1,100 @@ # Minitrino + A command line tool that makes it easy to run modular Trino environments locally. Compatible with Staburst version 354-e and later. -[![PyPI version](https://badge.fury.io/py/minitrino.svg)](https://badge.fury.io/py/minitrino) -[![Build Status](https://travis-ci.org/jefflester/minitrino.svg?branch=master)](https://travis-ci.org/jefflester/minitrino) -[![Trino Slack](https://img.shields.io/static/v1?logo=slack&logoColor=959DA5&label=Slack&labelColor=333a41&message=join%20conversation&color=3AC358)](https://trinodb.io/slack.html) +[![PyPI +version](https://badge.fury.io/py/minitrino.svg)](https://badge.fury.io/py/minitrino) +[![Build +Status](https://travis-ci.org/jefflester/minitrino.svg?branch=master)](https://app.travis-ci.com/jefflester/minitrino.svg?branch=master) +[![Trino +Slack](https://img.shields.io/static/v1?logo=slack&logoColor=959DA5&label=Slack&labelColor=333a41&message=join%20conversation&color=3AC358)](https://trinodb.io/slack.html) ----- -Latest Stable Release: 2.0.0 +Latest Stable Release: 2.0.1 ----- ## Overview -- [Requirements](#requirements) -- [Installation](#installation) -- [CLI](#cli) - - [Top-Level CLI Options](#top-level-cli-options) - - [Provisioning Environments](#provisioning-environments) - - [Removing Resources](#removing-resources) - - [Shutting Down Environments](#shutting-down-environments) - - [Taking Environment Snapshots](#taking-environment-snapshots) - - [Manage User Configuration](#manage-user-configuration) - - [Install the Library](#install-the-library) - - [Display Module Metadata](#display-module-metadata) - - [Display CLI Version](#display-cli-version) - - [Pointing the CLI to the Minitrino Library](#pointing-the-cli-to-the-minitrino-library) -- [Minitrino Configuration File](#minitrino-configuration-file) -- [Project Structure](#project-structure) -- [Adding New Modules (Tutorial)](#adding-new-modules-tutorial) -- [Troubleshooting](#troubleshooting) -- [Reporting Bugs and Contributing](#reporting-bugs-and-contributing) + +- [Minitrino](#minitrino) + - [Overview](#overview) + - [Requirements](#requirements) + - [Installation](#installation) + - [End Users](#end-users) + - [Developers](#developers) + - [CLI](#cli) + - [Top-Level CLI Options](#top-level-cli-options) + - [Provisioning Environments](#provisioning-environments) + - [Environment Variables](#environment-variables) + - [Using Licensed Starburst Features](#using-licensed-starburst-features) + - [Removing Resources](#removing-resources) + - [Shutting Down Environments](#shutting-down-environments) + - [Taking Environment Snapshots](#taking-environment-snapshots) + - [Manage User Configuration](#manage-user-configuration) + - [Install the Library](#install-the-library) + - [Display Module Metadata](#display-module-metadata) + - [Display Minitrino Versions](#display-minitrino-versions) + - [Pointing the CLI to the Minitrino Library](#pointing-the-cli-to-the-minitrino-library) + - [Minitrino Configuration File](#minitrino-configuration-file) + - [[CLI] Section](#cli-section) + - [[DOCKER] Section](#docker-section) + - [[TRINO] Section](#trino-section) + - [[MODULES] Section](#modules-section) + - [Project Structure](#project-structure) + - [Trino Dockerfile](#trino-dockerfile) + - [Adding New Modules (Tutorial)](#adding-new-modules-tutorial) + - [Create the Module Directory](#create-the-module-directory) + - [Add Trino Resources](#add-trino-resources) + - [Add the Docker Compose YAML](#add-the-docker-compose-yaml) + - [Add a Metadata File](#add-a-metadata-file) + - [Add a Readme File](#add-a-readme-file) + - [Review Progress](#review-progress) + - [Configure the Docker Compose YAML File](#configure-the-docker-compose-yaml-file) + - [Important Implementation Details: Paths and Labels](#important-implementation-details-paths-and-labels) + - [Path References for Volumes and Build Contexts](#path-references-for-volumes-and-build-contexts) + - [Minitrino Docker Labels](#minitrino-docker-labels) + - [Test the New Catalog](#test-the-new-catalog) + - [Customizing Images](#customizing-images) + - [Bootstrap Scripts](#bootstrap-scripts) + - [Managing Trino's `config.properties` File](#managing-trinos-configproperties-file) + - [Troubleshooting](#troubleshooting) + - [Reporting Bugs and Contributing](#reporting-bugs-and-contributing) + Library](#pointing-the-cli-to-the-minitrino-library) + - [Minitrino Configuration File](#minitrino-configuration-file) + - [[CLI] Section](#cli-section) + - [[DOCKER] Section](#docker-section) + - [[TRINO] Section](#trino-section) + - [[MODULES] Section](#modules-section) + - [Project Structure](#project-structure) + - [Trino Dockerfile](#trino-dockerfile) + - [Adding New Modules (Tutorial)](#adding-new-modules-tutorial) + - [Create the Module Directory](#create-the-module-directory) + - [Add Trino Resources](#add-trino-resources) + - [Add the Docker Compose YAML](#add-the-docker-compose-yaml) + - [Add a Metadata File](#add-a-metadata-file) + - [Add a Readme File](#add-a-readme-file) + - [Review Progress](#review-progress) + - [Configure the Docker Compose YAML + File](#configure-the-docker-compose-yaml-file) + - [Important Implementation Details: Paths and + Labels](#important-implementation-details-paths-and-labels) + - [Path References for Volumes and Build + Contexts](#path-references-for-volumes-and-build-contexts) + - [Minitrino Docker Labels](#minitrino-docker-labels) + - [Test the New Catalog](#test-the-new-catalog) + - [Customizing Images](#customizing-images) + - [Bootstrap Scripts](#bootstrap-scripts) + - [Managing Trino's `config.properties` + File](#managing-trinos-configproperties-file) + - [Troubleshooting](#troubleshooting) + - [Reporting Bugs and Contributing](#reporting-bugs-and-contributing) ----- ## Requirements + - Docker 19.03.0+ - Docker Compose (1.29.0+) - Python 3.8+ @@ -46,17 +106,20 @@ locally. Compatible with Staburst version 354-e and later. ## Installation ### End Users + Minitrino is available on PyPI and the library is available for public download -on GitHub. To install the Minitrino CLI, run `pip install minitrino`. To -install the library, run `minitrino lib_install`. +on GitHub. To install the Minitrino CLI, run `pip install minitrino`. To install +the library, run `minitrino lib_install`. ### Developers + In the project's root, run `./install.sh` to install the Minitrino CLI. If you encounter errors during installation, try running `sudo -H ./install.sh -v`. ----- ## CLI + Minitrino is built with [Click](https://click.palletsprojects.com/en/7.x/), a popular, open-source toolkit used to build Python-based CLIs. @@ -65,8 +128,9 @@ options can be specified with a shorthand alternative, which is the first letter of each option, i.e. `--module` can be `-m`. ### Top-Level CLI Options + You can get help, enable verbose output, and change the runtime library -directory for any command. +directory for any command. ``` Usage: minitrino [OPTIONS] COMMAND [ARGS]... @@ -87,6 +151,7 @@ Options: ``` ### Provisioning Environments + You can provision an environment via the `provision` command. ``` @@ -152,6 +217,7 @@ Using the structure of the Minitrino library, it is able to merge multiple Docker Compose files together. #### Environment Variables + Environment variables passed to Docker containers are sourced through two locations. The first is from the `minitrino.env` file in the library root. These variables define the versions of the provisioned Docker services. The second is @@ -163,10 +229,11 @@ Any existing environment variable can be overridden with the top-level `--env` option, and any unset variable can be set with it. #### Using Licensed Starburst Features + If you are using licensed features, you will need to provide a path to a valid Starburst license. This can be set via `minitrino config` or provided via the `--env` option at command runtime. The variable for this is -`STARBURST_LIC_PATH`. +`STARBURST_LIC_PATH`. Additionally, you need to uncomment the volume mount in the library's root `docker-compose.yml` file: @@ -179,6 +246,7 @@ Additionally, you need to uncomment the volume mount in the library's root ``` ### Removing Resources + You can remove resources with the `remove` command. ``` @@ -203,8 +271,8 @@ Notes: - Named volumes tied to any existing container cannot be forcibly removed, neither by Minitrino nor by the Docker CLI/SDK. - Images tied to stopped containers can be forcibly removed, but any image tied - to a running container cannot be forcibly removed, neither by Minitrino nor - by the Docker CLI. + to a running container cannot be forcibly removed, neither by Minitrino nor by + the Docker CLI. - You can find a module's label key by looking at the module's `docker-compose.yml` file in the Minitrino library. @@ -220,6 +288,7 @@ minitrino -v remove \ This will only remove volumes associated to the Postgres catalog module. ### Shutting Down Environments + You can shut down an active environment with the `down` command. ``` @@ -244,8 +313,9 @@ minitrino -v down ``` ### Taking Environment Snapshots + You can capture snapshots for both active and inactive environments with the -`snapshot` command. +`snapshot` command. ``` Usage: minitrino snapshot [OPTIONS] @@ -280,7 +350,7 @@ Options: --help Show this message and exit. ``` -Notes: +Notes: - Minitrino records the original `provision` command and places it in the snapshot file as `provision-snapshot.sh`; this can be directly executed. This @@ -299,7 +369,8 @@ minitrino snapshot -n super-cool-env -m hive-s3 -m elasticsearch -m ldap ``` ### Manage User Configuration -You can manage Minitrino configuration with the `config` command. + +You can manage Minitrino configuration with the `config` command. ``` Usage: minitrino config [OPTIONS] @@ -317,6 +388,7 @@ Options: ``` ### Install the Library + You can install the Minitrino library with the `lib_install` command. Note that it is best practice to have the library version match the CLI version. You can check these versions with `minitrino version`. @@ -332,7 +404,8 @@ Options: ``` ### Display Module Metadata -You can see Minitrino module metadata with the `modules` command. + +You can see Minitrino module metadata with the `modules` command. ``` Usage: minitrino modules [OPTIONS] @@ -349,8 +422,9 @@ Options: ``` ### Display Minitrino Versions + You can display the Minitrino CLI and library versions with the `version` -command. +command. ``` Usage: minitrino version [OPTIONS] @@ -362,6 +436,7 @@ Options: ``` ### Pointing the CLI to the Minitrino Library + The Minitrino CLI should always point to a compatible library with the expected structure. The library directory can be set one of four ways, listed below in the order of precedence: @@ -382,11 +457,13 @@ pointer to the library in Minitrino's configuration via the `LIB_PATH` config. ----- ## Minitrino Configuration File + Sticky configuration is set in `~/.minitrino/minitrino.cfg`. The sections in this file each serve a separate purpose. ### [CLI] Section -These configs allow the user to customize the behavior of Minitrino. + +These configs allow the user to customize the behavior of Minitrino. - LIB_PATH: The filesystem path of the Minitrino library (specifically to the `lib/` directory). @@ -394,17 +471,19 @@ These configs allow the user to customize the behavior of Minitrino. "nano", etc. Defaults to the shell's default editor. ### [DOCKER] Section + These configs allow the user to customize how Minitrino uses Docker. - DOCKER_HOST: A URL pointing to an accessible Docker host. This is automatically detected by Docker otherwise. ### [TRINO] Section + These configs allow the user to propagate config to the Trino container. Since many modules can append to Trino's core files, the supported way to make propagate changes to these Trino files is with these configs. -- CONFIG: Configuration for Trino's `config.properties` file. +- CONFIG: Configuration for Trino's `config.properties` file. - JVM_CONFIG: Configuration for Trino's `jvm.config` file. A multiline example of this section (note the indentation): @@ -419,6 +498,7 @@ JVM_CONFIG= ``` ### [MODULES] Section + This section sets environment variables passed to containers provisioned by Minitrino. Environment variables are only passed to a container if the variable is specified in the module's `docker-compose.yml` file. @@ -446,9 +526,11 @@ Variables propagated to the Trino container are supported by Trino secrets. ----- -## Project Structure +## Project Structure + The library is built around Docker Compose files and utilizes Docker's ability -to [extend Compose files](https://docs.docker.com/compose/extends/#multiple-compose-files). +to [extend Compose +files](https://docs.docker.com/compose/extends/#multiple-compose-files). The Starburst Trino service is defined in a Compose file at the library root, and all other services look up in the directory tree to reference the parent @@ -514,19 +596,21 @@ services: Notice that the volume mount is not relative to the `lib/modules/catalog/postgres/` directory––it is relative to the parent directory which houses the top-level `docker-compose.yml` file. Also, notice the -labels––these labels will be used to identify Docker resources tied to -Minitrino modules so that the CLI commands actually work. +labels––these labels will be used to identify Docker resources tied to Minitrino +modules so that the CLI commands actually work. ### Trino Dockerfile -Minitrino modifies the Starburst Trino Docker image by adding the Trino CLI -to the image as well as by providing `sudo` to the `trino` user. This is -required for certain bootstrap scripts (i.e. using `yum` to install packages in -a Trino container for a module). This image is compatible with Starburst Trino -images back to Starburst Trino version `332-e.0`. + +Minitrino modifies the Starburst Trino Docker image by adding the Trino CLI to +the image as well as by providing `sudo` to the `trino` user. This is required +for certain bootstrap scripts (i.e. using `yum` to install packages in a Trino +container for a module). This image is compatible with Starburst Trino images +back to Starburst Trino version `332-e.0`. ----- ## Adding New Modules (Tutorial) + Adding new modules is relatively simple, but there are a few important guidelines to follow to ensure compatibility with the Minitrino CLI. The design rules are the same for both catalogs and security modules. The example below @@ -534,6 +618,7 @@ demonstrates the process of creating a new catalog module for a Postgres service. ### Create the Module Directory + Create the module's directory in the `lib/modules/catalog/` directory: ```sh @@ -541,11 +626,12 @@ mkdir lib/modules/catalog/postgres/ cd lib/modules/catalog/postgres/ ``` -### Add Trino Resources +### Add Trino Resources + All resources for a module go inside of a `resources/` directory within the module. Inside this directory, place Trino-specific resources into a `trino/` directory, then mount the resources to the Trino service defined in the root -`docker-compose.yml` file. +`docker-compose.yml` file. ```sh mkdir -p resources/trino/ @@ -565,11 +651,12 @@ EOF" ----- Note: Passwords should always be `trinoRocks15` for consistency throughout -modules. +modules. ----- ### Add the Docker Compose YAML + In `lib/modules/catalog/postgres/`, add a Docker Compose file: ```sh @@ -598,10 +685,11 @@ POSTGRES_DB=minitrino EOF" ``` -This file will initialize Postgres with a database `minitrino`, a user -`trino`, and a password `trinoRocks15`. +This file will initialize Postgres with a database `minitrino`, a user `trino`, +and a password `trinoRocks15`. ### Add a Metadata File + The `metadata.json` file allows Minitrino to obtain key information for the module. It is required for a module to work with the CLI. @@ -622,6 +710,7 @@ alongside the given module. The `` wildcard is a supported convention if the module is incompatible with all other modules. ### Add a Readme File + This step is not required for personal development, but it is required to commit a module to the Minitrino repository. @@ -633,7 +722,8 @@ touch readme.md This file should contain an overview of the module. -### Review Progress +### Review Progress + The resulting directory tree should look like this (from the `/modules/catalog/` directory): @@ -650,6 +740,7 @@ postgres ``` ### Configure the Docker Compose YAML File + We will now define the `postgres.yml` Docker Compose file. Set it up as follows, and read the important notes after: @@ -674,9 +765,11 @@ services: ``` ### Important Implementation Details: Paths and Labels + We can observe a few things about the Compose file we just defined. #### Path References for Volumes and Build Contexts + First, the volumes we mount are not relative to the Compose file itself, they are relative to the base `docker-compose.yml` file in the library root. This is because the CLI extends Compose files, meaning that all path references in child @@ -685,19 +778,20 @@ Compose files need to be relative to the positioning of the parent Compose file. The base Compose file is determined when you execute a Docker Compose command––the first Compose file referenced in the command becomes the base file, and that happens to be the `docker-compose.yml` file in the library root. This -is how Minitrino constructs these commands. +is how Minitrino constructs these commands. If this is confusing, you can read more about extending Compose files on the -[Docker docs](https://docs.docker.com/compose/extends/#multiple-compose-files). +[Docker docs](https://docs.docker.com/compose/extends/#multiple-compose-files). #### Minitrino Docker Labels + Secondly, notice how we applied sets of labels to the Postgres service. These labels tell the CLI which resources to target when executing commands. -In general, there is no need to apply labels to the Trino service since they -are already applied in the parent Compose file unless* the module is an -extension of the Trino service itself (i.e. the Snowflake modules). Labels -should always be applied to: +In general, there is no need to apply labels to the Trino service since they are +already applied in the parent Compose file unless the module is an extension +of the Trino service itself (i.e. the Snowflake modules). Labels should always +be applied to: - Docker services (AKA the resulting container) - Named volumes @@ -753,11 +847,12 @@ volumes: actually define any new services/containers. See the Snowflake catalog modules for an example of this. For these modules, the only label requirement is to add the module-specific label to the Trino service in the relevant -`docker-compose.yml` file +`docker-compose.yml` file ----- ### Test the New Catalog + We are all finished up. We can test our new catalog through the Minitrino CLI: ```sh @@ -773,6 +868,7 @@ trino> show catalogs; ``` ### Customizing Images + If you need to build an image from a local Dockerfile, you can do so and structure the Compose file accordingly. See the library's root `docker-compose.yml` file for an example of this. Path references for volumes @@ -780,12 +876,12 @@ and the image build context will follow the same convention as volume mount paths described earlier. ### Bootstrap Scripts -Minitrino supports container bootstrap scripts. These scripts do not -replace the entrypoint (or default command) for a given container. The script -is copied from the Minitrino library to the container, executed, and then -removed from the container. Containers are restarted after each bootstrap script -execution, so the bootstrap scripts themselves should not restart the -container's service. + +Minitrino supports container bootstrap scripts. These scripts do not replace +the entrypoint (or default command) for a given container. The script is copied +from the Minitrino library to the container, executed, and then removed from the +container. Containers are restarted after each bootstrap script execution, so +the bootstrap scripts themselves should not restart the container's service. If a bootstrap script has already executed in a container and the volume associated with the container still exists, Minitrino will not re-execute the @@ -813,6 +909,7 @@ services: The `elasticsearch` module is a good example of this. ### Managing Trino's `config.properties` File + Many modules can change the Trino `config.properties` and `jvm.config` files. Because of this, there are two supported ways to modify these files with Minitrino. @@ -823,8 +920,8 @@ This will propagate the config to the Trino container when it is provisioned. Generally speaking, this can be used for any type of configuration (i.e. memory configuration) that is unlikely to be modified by any module. This also applies to the `jvm.config` file, which has identical support via the `JVM_CONFIG` -variable. If there are duplicate configs in either file, Minitrino will warn -the user. +variable. If there are duplicate configs in either file, Minitrino will warn the +user. To set these configs, your configuration file should look like: @@ -884,12 +981,13 @@ as possible. ----- ## Reporting Bugs and Contributing + To report bugs, please file a GitHub issue on the [Minitrino repository](https://github.com/jefflester/minitrino). Bug reports should: - Contain any relevant log messages (if the bug is tied to a command, running with the `-v` flag will make debugging easier) -- Describe what the expected outcome is +- Describe what the expected outcome is - Describe the proposed code fix (optional) Contributors have two options: @@ -897,7 +995,7 @@ Contributors have two options: 1. Fork the repository, then make a PR to merge your changes 2. If you have been added as a contributor, you can go with the method above or you can create a feature branch, then submit a PR for that feature branch - when it is ready to be merged. + when it is ready to be merged. In either case, please provide a comprehensive description of your changes with the PR. diff --git a/release-notes/2.0.1.md b/release-notes/2.0.1.md new file mode 100644 index 00000000..6f15ee47 --- /dev/null +++ b/release-notes/2.0.1.md @@ -0,0 +1,27 @@ +# Minitrino Release Notes: 2.0.1 + +## Release Overview + +- [Release Overview](#release-overview) +- [CLI Changes and Additions](#cli-changes-and-additions) +- [Library Changes and Additions](#library-changes-and-additions) +- [Other](#other) + +## CLI Changes and Additions + +- Check for valid SEP version before provisioning modules + +## Library Changes and Additions + +- Add Db2 catalog module +- Add Oracle catalog module (address part of issue #3) +- Exposed all catalog modules on a host port for alternative client access + +## Other + +- Fix conflicting port issue with modules that use overlapping services + (resolves issue #35) +- Standardize module readmes and complete all incomplete readmes (resolves issue + #8) +- Fix issue with snapshots where the incorrect CLI syntax is written to the + provision command	Populate module readmes The readme documents for each module should be accompanied with a high-level description of what the module does and any important instructions for interacting with the module. Will close this issue once all modules have this and it becomes a standard process for new module PRs.		2021-08-20T07:03:08	[]	[]
securefederatedai/openfl	securefederatedai__openfl-991	72312932edb198b2de04aee395ebd688ba9895c4	diff --git a/openfl/experimental/component/aggregator/__init__.py b/openfl/experimental/component/aggregator/__init__.py index 6686ce37b8..c2af4cc2ac 100644 --- a/openfl/experimental/component/aggregator/__init__.py +++ b/openfl/experimental/component/aggregator/__init__.py @@ -3,4 +3,4 @@ """openfl.experimental.component.aggregator package.""" # FIXME: Too much recursion. -from openfl.experimental.component.aggregator import Aggregator +from openfl.experimental.component.aggregator.aggregator import Aggregator diff --git a/openfl/experimental/interface/cli/collaborator.py b/openfl/experimental/interface/cli/collaborator.py index c5c6d06ac3..c5f8c924ee 100644 --- a/openfl/experimental/interface/cli/collaborator.py +++ b/openfl/experimental/interface/cli/collaborator.py @@ -75,7 +75,6 @@ def start_(plan, collaborator_name, secure, data_config="plan/data.yaml"): import yaml from yaml.loader import SafeLoader - collaborator_name = collaborator_name.lower() with open(data_config, "r") as f: data = yaml.load(f, Loader=SafeLoader) if data.get(collaborator_name, None) is None: @@ -118,7 +117,7 @@ def generate_cert_request(collaborator_name, silent, skip_package): from openfl.cryptography.participant import generate_csr from openfl.experimental.interface.cli.cli_helper import CERT_DIR - common_name = f"{collaborator_name}".lower() + common_name = f"{collaborator_name}" subject_alternative_name = f"DNS:{common_name}" file_name = f"col_{common_name}" @@ -282,7 +281,7 @@ def certify(collaborator_name, silent, request_pkg=None, import_=False): from openfl.experimental.interface.cli.cli_helper import CERT_DIR from openfl.utilities.utils import rmtree - common_name = f"{collaborator_name}".lower() + common_name = f"{collaborator_name}" if not import_: if request_pkg: diff --git a/openfl/experimental/interface/participants.py b/openfl/experimental/interface/participants.py index 4f48ae61fd..49d678a426 100644 --- a/openfl/experimental/interface/participants.py +++ b/openfl/experimental/interface/participants.py @@ -9,7 +9,7 @@ class Participant: def __init__(self, name: str = ""): self.private_attributes = {} - self._name = name.lower() + self._name = name @property def name(self): @@ -17,7 +17,7 @@ def name(self): @name.setter def name(self, name: str): - self._name = name.lower() + self._name = name def private_attributes(self, attrs: Dict[str, Any]) -> None: """	Issue observed in Workflow API Global DP & Vertical FL tutorials Describe the bug Running Workflow interface Global DP tutorials leads to unexpected KeyError exception. Similar issue is observed with Workflow interface Vertical_FL tutorial: Workflow_Interface_Vertical_FL.ipynb To Reproduce Steps to reproduce the error with Global DP tutorial: 1. Go to: openfl-tutorials/experimental/Global_DP/Workflow_Interface_Mnist_Implementation_1.py 2. Run tutorial with `python Workflow_Interface_Mnist_Implementation_1.py --config_path test_config.yml` 3. See error: KeyError: 'portland' Expected behavior Tutorial should run successfully without any error. Desktop (please complete the following information): - OS: Ubuntu (WSL) - OS Version: 22.04.3 LTS - Python: 3.8.19 [KeyError_logs.txt](https://github.com/user-attachments/files/15513351/KeyError_logs.txt)		2024-06-12T09:42:47	[]	[]
securefederatedai/openfl	securefederatedai__openfl-550	88c437bc27bc2eafa68fb025fe6f07329dcfbe26	diff --git a/.github/workflows/taskrunner.yml b/.github/workflows/taskrunner.yml index 3545fd4c0b..510fb21b94 100644 --- a/.github/workflows/taskrunner.yml +++ b/.github/workflows/taskrunner.yml @@ -29,4 +29,4 @@ jobs: pip install . - name: Test TaskRunner API run: \| - bash tests/github/test_hello_federation.sh keras_cnn_mnist aggregator col1 col2 $(hostname --all-fqdns \| awk '{print $1}') --rounds-to-train 3 + bash tests/github/test_hello_federation.sh keras_cnn_mnist aggregator col1 col2 $(hostname --all-fqdns \| awk '{print $1}') --rounds-to-train 3 --save-model output_model diff --git a/docs/running_the_federation.rst b/docs/running_the_federation.rst index ebe818cdec..49ea72b796 100644 --- a/docs/running_the_federation.rst +++ b/docs/running_the_federation.rst @@ -782,10 +782,10 @@ However, continue with the following procedure for details in creating a federat .. _creating_workspaces: -STEP 1: Create a Workspace on the Aggregator +STEP 1: Create a Workspace ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ -1. Start a Python 3.8 (>=3.6, <3.9) virtual environment and confirm \|productName\| is available. +1. Start a Python 3.8 (>=3.6, <3.11) virtual environment and confirm \|productName\| is available. .. code-block:: python @@ -1112,6 +1112,27 @@ STEP 3: Start the Federation When the last round of training is completed, the Aggregator stores the final weights in the protobuf file that was specified in the YAML file, which in this example is located at save/${WORKSPACE_TEMPLATE}_latest.pbuf. +Post Experiment +^^^^^^^^^^^^^^^ + +Experiment owners may access the final model in its native format. +Among other training artifacts, the aggregator creates the last and best aggregated (highest validation score) model snapshots. One may convert a snapshot to the native format and save the model to disk by calling the following command from the workspace: + +.. code-block:: console + + fx model save -i model_protobuf_path.pth -o save_model_path + +In order for this command to succeed, the TaskRunner used in the experiment must implement a :code:`save_native()` method. + +Another way to access the trained model is by calling the API command directly from a Python script: + +.. code-block:: python + + from openfl import get_model + model = get_model(plan_config, cols_config, data_config, model_protobuf_path) + +In fact, the :code:`get_model()` method returns a TaskRunner object loaded with the chosen model snapshot. Users may utilize the linked model as a regular Python object. + .. _running_the_federation_docker: diff --git a/openfl/__init__.py b/openfl/__init__.py index e29f0839b7..91cfdfc81f 100644 --- a/openfl/__init__.py +++ b/openfl/__init__.py @@ -1,5 +1,6 @@ # Copyright (C) 2020-2021 Intel Corporation # SPDX-License-Identifier: Apache-2.0 """openfl base package.""" -# flake8: noqa from .__version__ import __version__ +# flake8: noqa +from .interface.model import get_model diff --git a/openfl/interface/model.py b/openfl/interface/model.py new file mode 100644 index 0000000000..94468b797e --- /dev/null +++ b/openfl/interface/model.py @@ -0,0 +1,109 @@ +# Copyright (C) 2020-2022 Intel Corporation +# SPDX-License-Identifier: Apache-2.0 +"""Model CLI module.""" + +from click import confirm +from click import group +from click import option +from click import pass_context +from click import style +from click import Path as ClickPath +from logging import getLogger +from pathlib import Path + +from openfl.federated import Plan +from openfl.federated import TaskRunner +from openfl.protocols import utils +from openfl.pipelines import NoCompressionPipeline +from openfl.utilities.workspace import set_directory + +logger = getLogger(__name__) + + +@group() +@pass_context +def model(context): + """Manage Federated Learning Models.""" + context.obj['group'] = 'model' + + [email protected](name='save') +@pass_context +@option('-i', '--input', 'model_protobuf_path', required=True, + help='The model protobuf to convert', + type=ClickPath(exists=True)) +@option('-o', '--output', 'output_filepath', required=False, + help='Filename the model will be saved to in native format', + default='output_model', type=ClickPath(writable=True)) +@option('-p', '--plan-config', required=False, + help='Federated learning plan [plan/plan.yaml]', + default='plan/plan.yaml', type=ClickPath(exists=True)) +@option('-c', '--cols-config', required=False, + help='Authorized collaborator list [plan/cols.yaml]', + default='plan/cols.yaml', type=ClickPath(exists=True)) +@option('-d', '--data-config', required=False, + help='The data set/shard configuration file [plan/data.yaml]', + default='plan/data.yaml', type=ClickPath(exists=True)) +def save_(context, plan_config, cols_config, data_config, model_protobuf_path, output_filepath): + """ + Save the model in native format (PyTorch / Keras). + """ + output_filepath = Path(output_filepath).absolute() + if output_filepath.exists(): + if not confirm(style( + 'Do you want to overwrite the {}?'.format(output_filepath), fg='red', bold=True + )): + logger.info('Exiting') + context.obj['fail'] = True + return + + task_runner = get_model(plan_config, cols_config, data_config, model_protobuf_path) + + task_runner.save_native(output_filepath) + logger.info(f'Saved model in native format: 🠆 {output_filepath}') + + +def get_model( + plan_config: str, + cols_config: str, + data_config: str, + model_protobuf_path: str +) -> TaskRunner: + """ + Initialize TaskRunner and load it with provided model.pbuf. + + Contrary to its name, this function returns a TaskRunner instance. + The reason for this behavior is the flexibility of the TaskRunner interface and + the diversity of the ways we store models in our template workspaces. + """ + + # Here we change cwd to the experiment workspace folder + # because plan.yaml usually contains relative paths to components. + workspace_path = Path(plan_config).resolve().parent.parent + plan_config = Path(plan_config).resolve().relative_to(workspace_path) + cols_config = Path(cols_config).resolve().relative_to(workspace_path) + data_config = Path(data_config).resolve().relative_to(workspace_path) + + with set_directory(workspace_path): + plan = Plan.parse( + plan_config_path=plan_config, + cols_config_path=cols_config, + data_config_path=data_config + ) + collaborator_name = list(plan.cols_data_paths)[0] + data_loader = plan.get_data_loader(collaborator_name) + task_runner = plan.get_task_runner(data_loader=data_loader) + + model_protobuf_path = Path(model_protobuf_path).resolve() + logger.info(f'Loading OpenFL model protobuf: 🠆 {model_protobuf_path}') + + model_protobuf = utils.load_proto(model_protobuf_path) + + tensor_dict, _ = utils.deconstruct_model_proto(model_protobuf, NoCompressionPipeline()) + + # This may break for multiple models. + # task_runner.set_tensor_dict will need to handle multiple models + task_runner.set_tensor_dict(tensor_dict, with_opt_vars=False) + + del task_runner.data_loader + return task_runner diff --git a/openfl/utilities/workspace.py b/openfl/utilities/workspace.py index 08847f76aa..9afaa748d5 100644 --- a/openfl/utilities/workspace.py +++ b/openfl/utilities/workspace.py @@ -8,6 +8,7 @@ import shutil import sys import time +from contextlib import contextmanager from pathlib import Path from subprocess import check_call from sys import executable @@ -126,3 +127,23 @@ def _is_package_versioned(package: str) -> bool: and package not in ['pkg-resources==0.0.0', 'pkg_resources==0.0.0'] and '-e ' not in package ) + + +@contextmanager +def set_directory(path: Path): + """ + Sets provided path as the cwd within the context. + + Args: + path (Path): The path to the cwd + + Yields: + None + """ + + origin = Path().absolute() + try: + os.chdir(path) + yield + finally: + os.chdir(origin)	Load a trained pbuf model into general pytorch env Is your feature request related to a problem? Please describe. I would like to test the trained model externally, using an independent pytorch-based script that compares several models. The pbuf format is serialized differently that the regular .pth torch snapshots and can't be loaded using model.load_state_dict. Describe the solution you'd like It would be great if you can provide either a tool for converting pbuf to pth, or a flag enabling to save the model in pth format during the learning rounds. Thanks a lot.		2022-10-27T21:59:45	[]	[]

GAA-UAM/scikit-fda

GAA-UAM__scikit-fda-510

e69154b77d9e0fc260dd705de44c01fdf468e202

diff --git a/skfda/preprocessing/dim_reduction/_fpca.py b/skfda/preprocessing/dim_reduction/_fpca.py index d82debbaa..fba44acd3 100644 --- a/skfda/preprocessing/dim_reduction/_fpca.py +++ b/skfda/preprocessing/dim_reduction/_fpca.py @@ -1,8 +1,8 @@ """Functional Principal Component Analysis Module.""" - from __future__ import annotations -from typing import Callable, Optional, TypeVar, Union +import warnings +from typing import Callable, TypeVar import numpy as np import scipy.integrate @@ -34,7 +34,13 @@ class FPCA( # noqa: WPS230 (too many public attributes) Parameters: n_components: Number of principal components to keep from - functional principal component analysis. Defaults to 3. + functional principal component analysis. + + .. versionchanged:: 0.9 + In future versions, it will default to the maximum number + of components that can be extracted. + Currently, it still defaults to 3 but do not assume this + behavior as it will change. centering: Set to ``False`` when the functional data is already known to be centered and there is no need to center it. Otherwise, the mean of the functional data object is calculated and the data @@ -86,13 +92,23 @@ class FPCA( # noqa: WPS230 (too many public attributes) def __init__( self, - n_components: int = 3, + n_components: int | None = None, *, centering: bool = True, - regularization: Optional[L2Regularization[FData]] = None, - components_basis: Optional[Basis] = None, - _weights: Optional[Union[ArrayLike, WeightsCallable]] = None, + regularization: L2Regularization[FData] | None = None, + components_basis: Basis | None = None, + _weights: ArrayLike | WeightsCallable | None = None, ) -> None: + + if n_components is None: + warnings.warn( + "The default value of n_components will change in a future " + "version to the maximum number of components that can be " + "extracted. Update your code to specify explicitly the " + "number of components to avoid this warning.", + DeprecationWarning, + ) + n_components = 3 self.n_components = n_components self.centering = centering self.regularization = regularization

FPCA default number of components is wrong **Describe the bug** By default, it is natural that FPCA returns as many components as possible. The current behaviour is to return only 3, which is not a good default. We should deprecate it and change it in a future version.

2023-02-01T15:45:29

0.0

[]

GAA-UAM/scikit-fda

GAA-UAM__scikit-fda-502

0b578efd1dc78a56124a7e310a35d30116398429

diff --git a/skfda/misc/_math.py b/skfda/misc/_math.py index 0cb71737b..9b5f0b5a8 100644 --- a/skfda/misc/_math.py +++ b/skfda/misc/_math.py @@ -468,7 +468,7 @@ def _inner_product_integrate( len_arg2 = len(arg2) else: # If the arguments are callables, we need to pass the domain range - # explicitly. This is used internally for computing the gramian + # explicitly. This is used internally for computing the gram # matrix of operators. assert _domain_range is not None domain_range = _domain_range diff --git a/skfda/misc/operators/__init__.py b/skfda/misc/operators/__init__.py index 605b3a933..482a3a5cb 100644 --- a/skfda/misc/operators/__init__.py +++ b/skfda/misc/operators/__init__.py @@ -12,8 +12,8 @@ "_operators": [ "MatrixOperator", "Operator", - "gramian_matrix", - "gramian_matrix_optimization", + "gram_matrix", + "gram_matrix_optimization", ], "_srvf": ["SRSF"], }, @@ -28,7 +28,7 @@ from ._operators import ( MatrixOperator as MatrixOperator, Operator as Operator, - gramian_matrix as gramian_matrix, - gramian_matrix_optimization as gramian_matrix_optimization, + gram_matrix as gram_matrix, + gram_matrix_optimization as gram_matrix_optimization, ) from ._srvf import SRSF as SRSF diff --git a/skfda/misc/operators/_identity.py b/skfda/misc/operators/_identity.py index 11f2084a2..43be259cf 100644 --- a/skfda/misc/operators/_identity.py +++ b/skfda/misc/operators/_identity.py @@ -7,7 +7,7 @@ from ...representation import FDataGrid from ...representation.basis import Basis from ...typing._numpy import NDArrayFloat -from ._operators import InputType, Operator, gramian_matrix_optimization +from ._operators import InputType, Operator, gram_matrix_optimization T = TypeVar("T", bound=InputType) @@ -28,7 +28,7 @@ def __call__(self, f: T) -> T: # noqa: D102 return f -@gramian_matrix_optimization.register +@gram_matrix_optimization.register def basis_penalty_matrix_optimized( linear_operator: Identity[Any], basis: Basis, @@ -37,7 +37,7 @@ def basis_penalty_matrix_optimized( return basis.gram_matrix() -@gramian_matrix_optimization.register +@gram_matrix_optimization.register def fdatagrid_penalty_matrix_optimized( linear_operator: Identity[Any], basis: FDataGrid, diff --git a/skfda/misc/operators/_linear_differential_operator.py b/skfda/misc/operators/_linear_differential_operator.py index 02b6f190e..776b4aa72 100644 --- a/skfda/misc/operators/_linear_differential_operator.py +++ b/skfda/misc/operators/_linear_differential_operator.py @@ -18,7 +18,7 @@ ) from ...typing._base import DomainRangeLike from ...typing._numpy import NDArrayFloat -from ._operators import Operator, gramian_matrix_optimization +from ._operators import Operator, gram_matrix_optimization Order = int @@ -218,12 +218,12 @@ def applied_linear_diff_op( ############################################################# # -# Optimized implementations of gramian matrix for each basis. +# Optimized implementations of gram matrix for each basis. # ############################################################# -@gramian_matrix_optimization.register +@gram_matrix_optimization.register def constant_penalty_matrix_optimized( linear_operator: LinearDifferentialOperator, basis: ConstantBasis, @@ -299,7 +299,7 @@ def _monomial_evaluate_constant_linear_diff_op( return polynomials # type: ignore[no-any-return] -@gramian_matrix_optimization.register +@gram_matrix_optimization.register def monomial_penalty_matrix_optimized( linear_operator: LinearDifferentialOperator, basis: MonomialBasis, @@ -427,7 +427,7 @@ def _fourier_penalty_matrix_optimized_orthonormal( return penalty_matrix -@gramian_matrix_optimization.register +@gram_matrix_optimization.register def fourier_penalty_matrix_optimized( linear_operator: LinearDifferentialOperator, basis: FourierBasis, @@ -445,7 +445,7 @@ def fourier_penalty_matrix_optimized( return _fourier_penalty_matrix_optimized_orthonormal(basis, weights) -@gramian_matrix_optimization.register +@gram_matrix_optimization.register def bspline_penalty_matrix_optimized( linear_operator: LinearDifferentialOperator, basis: BSplineBasis, @@ -576,7 +576,7 @@ def bspline_penalty_matrix_optimized( return penalty_matrix -@gramian_matrix_optimization.register +@gram_matrix_optimization.register def fdatagrid_penalty_matrix_optimized( linear_operator: LinearDifferentialOperator, basis: FDataGrid, diff --git a/skfda/misc/operators/_operators.py b/skfda/misc/operators/_operators.py index 770401d7a..221c4ac89 100644 --- a/skfda/misc/operators/_operators.py +++ b/skfda/misc/operators/_operators.py @@ -37,26 +37,26 @@ def __call__(self, vector: OperatorInput) -> OperatorOutput: @multimethod.multidispatch -def gramian_matrix_optimization( +def gram_matrix_optimization( linear_operator: Any, basis: OperatorInput, ) -> NDArrayFloat: """ - Efficient implementation of gramian_matrix. + Efficient implementation of gram_matrix. Generic function that can be subclassed for different combinations of operator and basis in order to provide a more efficient implementation - for the gramian matrix. + for the gram matrix. """ return NotImplemented -def gramian_matrix_numerical( +def gram_matrix_numerical( linear_operator: Operator[OperatorInput, OutputType], basis: OperatorInput, ) -> NDArrayFloat: """ - Return the gramian matrix given a basis, computed numerically. + Return the gram matrix given a basis, computed numerically. This method should work for every linear operator. @@ -70,19 +70,19 @@ def gramian_matrix_numerical( return inner_product_matrix(evaluated_basis, _domain_range=domain_range) -def gramian_matrix( +def gram_matrix( linear_operator: Operator[OperatorInput, OutputType], basis: OperatorInput, ) -> NDArrayFloat: r""" - Return the gramian matrix given a basis. + Return the gram matrix given a basis. - The gramian operator of a linear operator :math:`\Gamma` is + The gram operator of a linear operator :math:`\Gamma` is .. math:: G = \Gamma*\Gamma - This method evaluates that gramian operator in a given basis, + This method evaluates that gram operator in a given basis, which is necessary for performing Tikhonov regularization, among other things. @@ -91,11 +91,11 @@ def gramian_matrix( """ # Try to use a more efficient implementation - matrix = gramian_matrix_optimization(linear_operator, basis) + matrix = gram_matrix_optimization(linear_operator, basis) if matrix is not NotImplemented: return matrix - return gramian_matrix_numerical(linear_operator, basis) + return gram_matrix_numerical(linear_operator, basis) class MatrixOperator(Operator[NDArrayFloat, NDArrayFloat]): diff --git a/skfda/misc/regularization/_regularization.py b/skfda/misc/regularization/_regularization.py index 47c3a5432..d8ca23788 100644 --- a/skfda/misc/regularization/_regularization.py +++ b/skfda/misc/regularization/_regularization.py @@ -11,7 +11,7 @@ from ...representation import FData from ...representation.basis import Basis from ...typing._numpy import NDArrayFloat -from ..operators import Identity, Operator, gramian_matrix +from ..operators import Identity, Operator, gram_matrix from ..operators._operators import OperatorInput @@ -101,7 +101,7 @@ def penalty_matrix( else self.linear_operator ) - return self.regularization_parameter * gramian_matrix( + return self.regularization_parameter * gram_matrix( linear_operator, basis, )

Rename Gramian matrix to Gram matrix. In some places we use the terminology Gramian matrix (even in function names). We should use always the name Gram matrix for consistency.

I'd like to take this up. I would really appreciate some clarification on the same Sure! There are some places in which we used the terminology "Gramian matrix" instead of the more common "Gram matrix": https://github.com/GAA-UAM/scikit-fda/search?q=gramian In particular, the functions `gramian_matrix`, `gramian_matrix_numerical` and `gramian_matrix_optimization` should be renamed, as well as the related comments. As far as I know, nobody is using these functions in other projects (although `gramian_matrix_optimization` was exposed to be able to provide optimized implementations for particular combinations of basis and operator), so no deprecation would be needed. I assign you to this issue, please tell me in case you have any other doubt. So Basically we have to change Gramian_matrix,gramian_matrix_numerical to Gram matrix to be more consistent. Wherein gramian _matrix_optimization is able to provide optimized implementation so it doesn't need to be changed right?

2022-12-24T11:07:24

0.0

[]

GAA-UAM/scikit-fda

GAA-UAM__scikit-fda-436

ebab51dbf57f3f76c0df615d830868425366f425

diff --git a/skfda/misc/hat_matrix.py b/skfda/misc/hat_matrix.py index fc94adc6b..81694bd47 100644 --- a/skfda/misc/hat_matrix.py +++ b/skfda/misc/hat_matrix.py @@ -11,20 +11,21 @@ import abc import math -from typing import Callable, TypeVar, Union, overload +from typing import Callable, Final, TypeVar, Union, overload import numpy as np from .._utils._sklearn_adapter import BaseEstimator from ..representation._functional_data import FData from ..representation.basis import FDataBasis -from ..typing._base import GridPointsLike from ..typing._numpy import NDArrayFloat from . import kernels -Input = TypeVar("Input", bound=Union[FData, GridPointsLike]) +Input = TypeVar("Input", bound=Union[FData, NDArrayFloat]) Prediction = TypeVar("Prediction", bound=Union[NDArrayFloat, FData]) +DEFAULT_BANDWIDTH_PERCENTILE: Final = 15 + class HatMatrix( BaseEstimator, @@ -52,8 +53,8 @@ def __call__( self, *, delta_x: NDArrayFloat, - X_train: Input | None = None, - X: Input | None = None, + X_train: Input, + X: Input, y_train: None = None, weights: NDArrayFloat | None = None, _cv: bool = False, @@ -65,8 +66,8 @@ def __call__( self, *, delta_x: NDArrayFloat, - X_train: Input | None = None, - X: Input | None = None, + X_train: Input, + X: Input, y_train: Prediction | None = None, weights: NDArrayFloat | None = None, _cv: bool = False, @@ -77,8 +78,8 @@ def __call__( self, *, delta_x: NDArrayFloat, - X_train: Input | None = None, - X: Input | None = None, + X_train: Input, + X: Input, y_train: NDArrayFloat | FData | None = None, weights: NDArrayFloat | None = None, _cv: bool = False, @@ -186,7 +187,7 @@ def _hat_matrix_function_not_normalized( ) -> NDArrayFloat: bandwidth = ( - np.percentile(np.abs(delta_x), 15) + float(np.percentile(delta_x, DEFAULT_BANDWIDTH_PERCENTILE)) if self.bandwidth is None else self.bandwidth ) @@ -278,8 +279,8 @@ def __call__( self, *, delta_x: NDArrayFloat, - X_train: FData | GridPointsLike | None = None, - X: FData | GridPointsLike | None = None, + X_train: FData | NDArrayFloat, + X: FData | NDArrayFloat, y_train: None = None, weights: NDArrayFloat | None = None, _cv: bool = False, @@ -291,8 +292,8 @@ def __call__( self, *, delta_x: NDArrayFloat, - X_train: FData | GridPointsLike | None = None, - X: FData | GridPointsLike | None = None, + X_train: FData | NDArrayFloat, + X: FData | NDArrayFloat, y_train: Prediction | None = None, weights: NDArrayFloat | None = None, _cv: bool = False, @@ -303,19 +304,37 @@ def __call__( # noqa: D102 self, *, delta_x: NDArrayFloat, - X_train: FData | GridPointsLike | None = None, - X: FData | GridPointsLike | None = None, + X_train: FData | NDArrayFloat, + X: FData | NDArrayFloat, y_train: NDArrayFloat | FData | None = None, weights: NDArrayFloat | None = None, _cv: bool = False, ) -> NDArrayFloat | FData: bandwidth = ( - np.percentile(np.abs(delta_x), 15) + float(np.percentile(delta_x, DEFAULT_BANDWIDTH_PERCENTILE)) if self.bandwidth is None else self.bandwidth ) + # Smoothing for functions of one variable + if not isinstance(X_train, FDataBasis) and X_train[0].shape[0] == 1: + delta_x = np.subtract.outer( + X.flatten(), + X_train.flatten(), + ) + + assert y_train is None + + return super().__call__( # noqa: WPS503 + delta_x=delta_x, + X_train=X_train, + X=X, + y_train=y_train, + weights=weights, + _cv=_cv, + ) + # Regression if isinstance(X_train, FData): assert isinstance(X, FData) @@ -326,44 +345,41 @@ def __call__( # noqa: D102 ): raise ValueError("Only FDataBasis is supported for now.") + m1 = X_train.coefficients + m2 = X.coefficients + if y_train is None: y_train = np.identity(X_train.n_samples) - m1 = X_train.coefficients - m2 = X.coefficients + else: + m1 = X_train + m2 = X + + if y_train is None: + y_train = np.identity(X_train.shape[0]) - # Subtract previous matrices obtaining a 3D matrix - # The i-th element contains the matrix X_train - X[i] - C = m1 - m2[:, np.newaxis] + # Subtract previous matrices obtaining a 3D matrix + # The i-th element contains the matrix X_train - X[i] + C = m1 - m2[:, np.newaxis] + # Inner product matrix only is applicable in regression + if isinstance(X_train, FDataBasis): inner_product_matrix = X_train.basis.inner_product_matrix() # Calculate new coefficients taking into account cross-products # if the basis is orthonormal, C would not change C = C @ inner_product_matrix # noqa: WPS350 - # Adding a column of ones in the first position of all matrices - dims = (C.shape[0], C.shape[1], 1) - C = np.concatenate((np.ones(dims), C), axis=-1) - - return self._solve_least_squares( - delta_x=delta_x, - coefs=C, - y_train=y_train, - bandwidth=bandwidth, - ) - - # Smoothing - else: + # Adding a column of ones in the first position of all matrices + dims = (C.shape[0], C.shape[1], 1) + C = np.concatenate((np.ones(dims), C), axis=-1) - return super().__call__( # type: ignore[misc, type-var] # noqa: WPS503 - delta_x=delta_x, - X_train=X_train, - X=X, - y_train=y_train, # type: ignore[arg-type] - weights=weights, - _cv=_cv, - ) + return self._solve_least_squares( + delta_x=delta_x, + coefs=C, + y_train=y_train, + bandwidth=bandwidth, + ) def _solve_least_squares( self, @@ -485,7 +501,7 @@ def _hat_matrix_function_not_normalized( # For each row in the distances matrix, it calculates the furthest # point within the k nearest neighbours vec = np.sort( - np.abs(delta_x), + delta_x, axis=1, )[:, n_neighbors - 1] + tol diff --git a/skfda/preprocessing/smoothing/_kernel_smoothers.py b/skfda/preprocessing/smoothing/_kernel_smoothers.py index c07489235..e9b56f44d 100644 --- a/skfda/preprocessing/smoothing/_kernel_smoothers.py +++ b/skfda/preprocessing/smoothing/_kernel_smoothers.py @@ -9,9 +9,11 @@ import numpy as np -from ..._utils._utils import _to_grid_points +from ..._utils._utils import _cartesian_product, _to_grid_points from ...misc.hat_matrix import HatMatrix, NadarayaWatsonHatMatrix +from ...misc.metrics import PairwiseMetric, l2_distance from ...typing._base import GridPointsLike +from ...typing._metric import Metric from ...typing._numpy import NDArrayFloat from ._linear import _LinearSmoother @@ -114,10 +116,12 @@ def __init__( *, weights: Optional[NDArrayFloat] = None, output_points: Optional[GridPointsLike] = None, + metric: Metric[NDArrayFloat] = l2_distance, ): self.kernel_estimator = kernel_estimator self.weights = weights self.output_points = output_points + self.metric = metric self._cv = False # For testing purposes only def _hat_matrix( @@ -126,18 +130,18 @@ def _hat_matrix( output_points: GridPointsLike, ) -> NDArrayFloat: - input_points = _to_grid_points(input_points) - output_points = _to_grid_points(output_points) + input_points = _cartesian_product(_to_grid_points(input_points)) + output_points = _cartesian_product(_to_grid_points(output_points)) if self.kernel_estimator is None: self.kernel_estimator = NadarayaWatsonHatMatrix() - delta_x = np.subtract.outer(output_points[0], input_points[0]) + delta_x = PairwiseMetric(self.metric)(output_points, input_points) return self.kernel_estimator( delta_x=delta_x, weights=self.weights, - X_train=input_points, - X=output_points, + X_train=np.array(input_points), + X=np.array(output_points), _cv=self._cv, ) diff --git a/skfda/preprocessing/smoothing/_linear.py b/skfda/preprocessing/smoothing/_linear.py index 8759704c9..09c6fadcc 100644 --- a/skfda/preprocessing/smoothing/_linear.py +++ b/skfda/preprocessing/smoothing/_linear.py @@ -28,6 +28,7 @@ class _LinearSmoother( ``hat_matrix`` to define the smoothing or 'hat' matrix. """ + input_points_: GridPoints output_points_: GridPoints @@ -116,9 +117,25 @@ def transform( ) ) + dims = [len(e) for e in self.output_points_] + dims += [len(e) for e in self.input_points_] + + hat_matrix = np.reshape( + self.hat_matrix_, + dims, + ) + + data_matrix = np.einsum( + hat_matrix, + [Ellipsis, *range(1, len(self.output_points_) + 1)], + X.data_matrix, + [0, *range(1, len(self.output_points_) + 2)], + [0, Ellipsis, len(self.output_points_) + 1], + ) + # The matrix is cached return X.copy( - data_matrix=self.hat_matrix_ @ X.data_matrix, + data_matrix=data_matrix, grid_points=self.output_points_, )

Smoothing in several dimensions The current smoothers only work for one-dimensional functions. It should be reasonably easy to extend them to several dimensions.

2022-03-28T15:25:31

0.0

[]

GAA-UAM/scikit-fda

GAA-UAM__scikit-fda-430

c47ec786f903dc35dea466c341a6e947a5e7faa6

diff --git a/examples/plot_kernel_regression.py b/examples/plot_kernel_regression.py index 39273791e..d3cb895bb 100644 --- a/examples/plot_kernel_regression.py +++ b/examples/plot_kernel_regression.py @@ -109,8 +109,8 @@ print('Score NW:', nw_res) ############################################################################## -# For Local Linear Regression, FDataBasis representation with an orthonormal -# basis should be used (for the previous cases it is possible to use either +# For Local Linear Regression, FDataBasis representation with a basis should be +# used (for the previous cases it is possible to use either # FDataGrid or FDataBasis). # # For basis, Fourier basis with 10 elements has been selected. Note that the diff --git a/skfda/misc/hat_matrix.py b/skfda/misc/hat_matrix.py index ae0e9e693..67f51810e 100644 --- a/skfda/misc/hat_matrix.py +++ b/skfda/misc/hat_matrix.py @@ -14,10 +14,9 @@ import numpy as np from sklearn.base import BaseEstimator, RegressorMixin -from skfda.representation._functional_data import FData -from skfda.representation.basis import FDataBasis - -from ..representation._typing import GridPoints, GridPointsLike +from ..representation._functional_data import FData +from ..representation._typing import GridPoints, GridPointsLike, NDArrayFloat +from ..representation.basis import FDataBasis from . import kernels @@ -36,7 +35,7 @@ def __init__( self, *, bandwidth: Optional[float] = None, - kernel: Callable[[np.ndarray], np.ndarray] = kernels.normal, + kernel: Callable[[NDArrayFloat], NDArrayFloat] = kernels.normal, ): self.bandwidth = bandwidth self.kernel = kernel @@ -44,13 +43,13 @@ def __init__( def __call__( self, *, - delta_x: np.ndarray, + delta_x: NDArrayFloat, X_train: Optional[Union[FData, GridPointsLike]] = None, X: Optional[Union[FData, GridPointsLike]] = None, - y_train: Optional[np.ndarray] = None, - weights: Optional[np.ndarray] = None, + y_train: Optional[NDArrayFloat] = None, + weights: Optional[NDArrayFloat] = None, _cv: bool = False, - ) -> np.ndarray: + ) -> NDArrayFloat: r""" Calculate the hat matrix or the prediction. @@ -99,8 +98,8 @@ def __call__( def _hat_matrix_function_not_normalized( self, *, - delta_x: np.ndarray, - ) -> np.ndarray: + delta_x: NDArrayFloat, + ) -> NDArrayFloat: pass @@ -141,8 +140,8 @@ class NadarayaWatsonHatMatrix(HatMatrix): def _hat_matrix_function_not_normalized( self, *, - delta_x: np.ndarray, - ) -> np.ndarray: + delta_x: NDArrayFloat, + ) -> NDArrayFloat: if self.bandwidth is None: percentage = 15 @@ -185,7 +184,7 @@ class LocalLinearRegressionHatMatrix(HatMatrix): For **kernel regression** algorithm: Given functional data, :math:`(X_1, X_2, ..., X_n)` where each function - is expressed in a orthonormal basis with :math:`J` elements and scalar + is expressed in a basis with :math:`J` elements and scalar response :math:`Y = (y_1, y_2, ..., y_n)`. It is desired to estimate the values @@ -222,13 +221,13 @@ class LocalLinearRegressionHatMatrix(HatMatrix): def __call__( # noqa: D102 self, *, - delta_x: np.ndarray, + delta_x: NDArrayFloat, X_train: Optional[Union[FDataBasis, GridPoints]] = None, X: Optional[Union[FDataBasis, GridPoints]] = None, - y_train: Optional[np.ndarray] = None, - weights: Optional[np.ndarray] = None, + y_train: Optional[NDArrayFloat] = None, + weights: Optional[NDArrayFloat] = None, _cv: bool = False, - ) -> np.ndarray: + ) -> NDArrayFloat: if self.bandwidth is None: percentage = 15 @@ -243,10 +242,23 @@ def __call__( # noqa: D102 m1 = X_train.coefficients m2 = X.coefficients + # Subtract previous matrices obtaining a 3D matrix + # The i-th element contains the matrix X_train - X[i] + C = m1 - m2[:, np.newaxis] + + inner_product_matrix = X_train.basis.inner_product_matrix() + + # Calculate new coefficients taking into account cross-products + # if the basis is orthonormal, C would not change + C = C @ inner_product_matrix + + # Adding a column of ones in the first position of all matrices + dims = (C.shape[0], C.shape[1], 1) + C = np.concatenate((np.ones(dims), C), axis=-1) + return self._solve_least_squares( delta_x=delta_x, - m1=m1, - m2=m2, + coefs=C, y_train=y_train, ) @@ -264,39 +276,16 @@ def __call__( # noqa: D102 def _solve_least_squares( self, - delta_x: np.ndarray, - m1: np.ndarray, - m2: np.ndarray, - y_train: np.ndarray, - ) -> np.ndarray: + delta_x: NDArrayFloat, + coefs: NDArrayFloat, + y_train: NDArrayFloat, + ) -> NDArrayFloat: W = np.sqrt(self.kernel(delta_x / self.bandwidth)) - # Adding a column of ones to m1 - m1 = np.concatenate( - ( - np.ones(m1.shape[0])[:, np.newaxis], - m1, - ), - axis=1, - ) - - # Adding a column of zeros to m2 - m2 = np.concatenate( - ( - np.zeros(m2.shape[0])[:, np.newaxis], - m2, - ), - axis=1, - ) - - # Subtract previous matrices obtaining a 3D matrix - # The i-th element contains the matrix X_train - X[i] - C = m1 - m2[:, np.newaxis] - # A x = b - # Where x = (a, b_1, ..., b_J) - A = (C.T * W.T).T + # Where x = (a, b_1, ..., b_J). + A = (coefs.T * W.T).T b = np.einsum('ij, j... -> ij...', W, y_train) # For Ax = b calculates x that minimize the square error @@ -312,8 +301,8 @@ def _solve_least_squares( def _hat_matrix_function_not_normalized( self, *, - delta_x: np.ndarray, - ) -> np.ndarray: + delta_x: NDArrayFloat, + ) -> NDArrayFloat: if self.bandwidth is None: percentage = 15 @@ -369,7 +358,7 @@ def __init__( self, *, n_neighbors: Optional[int] = None, - kernel: Callable[[np.ndarray], np.ndarray] = kernels.uniform, + kernel: Callable[[NDArrayFloat], NDArrayFloat] = kernels.uniform, ): self.n_neighbors = n_neighbors self.kernel = kernel @@ -377,8 +366,8 @@ def __init__( def _hat_matrix_function_not_normalized( self, *, - delta_x: np.ndarray, - ) -> np.ndarray: + delta_x: NDArrayFloat, + ) -> NDArrayFloat: input_points_len = delta_x.shape[1] diff --git a/skfda/ml/regression/_kernel_regression.py b/skfda/ml/regression/_kernel_regression.py index 2834cae86..aecc64c60 100644 --- a/skfda/ml/regression/_kernel_regression.py +++ b/skfda/ml/regression/_kernel_regression.py @@ -6,10 +6,10 @@ from sklearn.base import BaseEstimator, RegressorMixin from sklearn.utils.validation import check_is_fitted -from skfda.misc.hat_matrix import HatMatrix, NadarayaWatsonHatMatrix -from skfda.misc.metrics import PairwiseMetric, l2_distance -from skfda.misc.metrics._typing import Metric -from skfda.representation._functional_data import FData +from ...misc.hat_matrix import HatMatrix, NadarayaWatsonHatMatrix +from ...misc.metrics import PairwiseMetric, l2_distance +from ...misc.metrics._typing import Metric +from ...representation._functional_data import FData class KernelRegression(

Allow non orthonormal basis in local linear regression _Originally posted by @vnmabus in https://github.com/GAA-UAM/scikit-fda/pull/417#r800112765_

2022-03-04T03:06:38

0.0

[]

GAA-UAM/scikit-fda

GAA-UAM__scikit-fda-429

46fc684ff9d96bc282e8646d7ac0e26c4f107ee4

diff --git a/docs/modules/misc.rst b/docs/modules/misc.rst index c1a53d05a..6d1669f9c 100644 --- a/docs/modules/misc.rst +++ b/docs/modules/misc.rst @@ -57,4 +57,4 @@ functional data: misc/operators misc/regularization misc/hat_matrix - misc/score_functions + misc/scoring diff --git a/docs/modules/misc/scoring.rst b/docs/modules/misc/scoring.rst new file mode 100644 index 000000000..b1b5e2fd2 --- /dev/null +++ b/docs/modules/misc/scoring.rst @@ -0,0 +1,18 @@ +Scoring methods for regression with functional response. +======================================================== + +The functions in this module are a generalization for functional data of +the regression metrics of the sklearn library +(https://scikit-learn.org/stable/modules/model_evaluation.html#regression-metrics). +Only scores that support multioutput are included. + + +.. autosummary:: + :toctree: autosummary + + skfda.misc.scoring.explained_variance_score + skfda.misc.scoring.mean_absolute_error + skfda.misc.scoring.mean_absolute_percentage_error + skfda.misc.scoring.mean_squared_error + skfda.misc.scoring.mean_squared_log_error + skfda.misc.scoring.r2_score diff --git a/skfda/_utils/_sklearn_adapter.py b/skfda/_utils/_sklearn_adapter.py index 3c23aa330..f3eaecb0b 100644 --- a/skfda/_utils/_sklearn_adapter.py +++ b/skfda/_utils/_sklearn_adapter.py @@ -184,8 +184,10 @@ def score( # noqa: D102 y: TargetPrediction, sample_weight: NDArrayFloat | None = None, ) -> float: - return super().score( # type: ignore[no-any-return] - X, + from ..misc.scoring import r2_score + y_pred = self.predict(X) + return r2_score( y, + y_pred, sample_weight=sample_weight, ) diff --git a/skfda/exploratory/stats/_stats.py b/skfda/exploratory/stats/_stats.py index 27eeac78b..ac89a992b 100644 --- a/skfda/exploratory/stats/_stats.py +++ b/skfda/exploratory/stats/_stats.py @@ -18,25 +18,32 @@ T = TypeVar('T', bound=Union[NDArrayFloat, FData]) -def mean(X: F) -> F: +def mean( + X: F, + weights: NDArrayFloat | None = None, +) -> F: """ Compute the mean of all the samples in a FData object. Args: X: Object containing all the samples whose mean is wanted. - + weights: Sample weight. By default, uniform weight are used. Returns: Mean of all the samples in the original object, as a :term:`functional data object` with just one sample. """ - return X.mean() + if weights is None: + return X.mean() + + weight = (1 / np.sum(weights)) * weights + return (X * weight).sum() def var(X: FData) -> FDataGrid: """ - Compute the variance of a set of samples in a FDataGrid object. + Compute the variance of a set of samples in a FData object. Args: X: Object containing all the set of samples whose variance is desired. diff --git a/skfda/misc/__init__.py b/skfda/misc/__init__.py index cdd16e6d4..6a977724a 100644 --- a/skfda/misc/__init__.py +++ b/skfda/misc/__init__.py @@ -13,6 +13,7 @@ "metrics", "operators", "regularization", + "scoring", "validation", ], submod_attrs={ diff --git a/skfda/misc/scoring.py b/skfda/misc/scoring.py new file mode 100644 index 000000000..c57902668 --- /dev/null +++ b/skfda/misc/scoring.py @@ -0,0 +1,1073 @@ +"""Scoring methods for FData.""" +from __future__ import annotations + +import math +import warnings +from functools import singledispatch +from typing import Callable, Optional, TypeVar, Union, overload + +import numpy as np +import sklearn.metrics +from typing_extensions import Literal, Protocol + +from .._utils import nquad_vec +from ..representation import FData, FDataBasis, FDataGrid +from ..representation._functional_data import EvalPointsType +from ..typing._numpy import NDArrayFloat + +DataType = TypeVar('DataType') + +MultiOutputType = Literal['uniform_average', 'raw_values'] + + +class _InfiniteScoreError(Exception): + """Exception for skipping integral on infinite value.""" + + +class ScoreFunction(Protocol): + """Type definition for score functions.""" + + @overload + def __call__( + self, + y_true: DataType, + y_pred: DataType, + *, + sample_weight: NDArrayFloat | None = None, + multioutput: Literal['uniform_average'] = 'uniform_average', + ) -> float: + pass # noqa: WPS428 + + @overload + def __call__( # noqa: D102 + self, + y_true: DataType, + y_pred: DataType, + *, + sample_weight: NDArrayFloat | None = None, + multioutput: Literal['raw_values'], + ) -> DataType: + pass # noqa: WPS428 + + def __call__( # noqa: D102 + self, + y_true: DataType, + y_pred: DataType, + *, + sample_weight: NDArrayFloat | None = None, + multioutput: MultiOutputType = 'uniform_average', + ) -> float | DataType: + pass # noqa: WPS428 + + +def _domain_measure(fd: FData) -> float: + measure = 1.0 + for interval in fd.domain_range: + measure = measure * (interval[1] - interval[0]) + return measure + + +def _var( + x: FDataGrid, + weights: NDArrayFloat | None = None, +) -> FDataGrid: + from ..exploratory.stats import mean, var + + if weights is None: + return var(x) + + return mean( # type: ignore[no-any-return] + np.power(x - mean(x, weights=weights), 2), + weights=weights, + ) + + +def _multioutput_score_basis( + y_true: FDataBasis, + multioutput: MultiOutputType, + integrand: Callable[[NDArrayFloat], NDArrayFloat], +) -> float: + + if multioutput != "uniform_average": + raise ValueError( + f"Only \"uniform_average\" is supported for \"multioutput\" when " + f"the input is a FDatabasis: received {multioutput} instead", + ) + + try: + integral = nquad_vec( + integrand, + y_true.domain_range, + ) + except _InfiniteScoreError: + return -math.inf + + # If the dimension of the codomain is > 1, + # the mean of the scores is taken + return float(np.mean(integral) / _domain_measure(y_true)) + + +def _multioutput_score_grid( + score: FDataGrid, + multioutput: MultiOutputType, + squared: bool = True, +) -> float | FDataGrid: + + if not squared: + score = np.sqrt(score) + + if multioutput == 'raw_values': + return score + + # Score only contains 1 function + # If the dimension of the codomain is > 1, + # the mean of the scores is taken + return float(np.mean(score.integrate()[0]) / _domain_measure(score)) + + +@overload +def explained_variance_score( + y_true: DataType, + y_pred: DataType, + *, + sample_weight: NDArrayFloat | None = None, + multioutput: Literal['uniform_average'] = 'uniform_average', +) -> float: + pass # noqa: WPS428 + + +@overload +def explained_variance_score( + y_true: DataType, + y_pred: DataType, + *, + sample_weight: NDArrayFloat | None = None, + multioutput: Literal['raw_values'], +) -> DataType: + pass # noqa: WPS428 + + +@singledispatch +def explained_variance_score( + y_true: DataType, + y_pred: DataType, + *, + sample_weight: NDArrayFloat | None = None, + multioutput: MultiOutputType = 'uniform_average', +) -> float | DataType: + r"""Explained variance score for :class:`~skfda.representation.FData`. + + With :math:`y\_true = (X_1, X_2, ..., X_n)` being the real values, + :math:`t\_pred = (\hat{X}_1, \hat{X}_2, ..., \hat{X}_n)` being the + estimated and :math:`sample\_weight = (w_1, w_2, ..., w_n)`, the score is + calculated as + + .. math:: + EV(y\_true, y\_pred)(t) = 1 - + \frac{Var(y\_true(t) - y\_pred(t), sample\_weight)} + {Var(y\_true(t), sample\_weight)} + + where :math:`Var` is a weighted variance. + + Weighted variance is defined as below + + .. math:: + Var(y\_true, sample\_weight)(t) = \sum_{i=1}^n w_i + (X_i(t) - Mean(fd(t), sample\_weight))^2. + + Here, :math:`Mean` is a weighted mean. + + For :math:`y\_true` and :math:`y\_pred` of type + :class:`~skfda.representation.FDataGrid`, :math:`EV` is + also a :class:`~skfda.representation.FDataGrid` object with + the same grid points. + + If multioutput = 'raw_values', the function :math:`EV` is returned. + Otherwise, if multioutput = 'uniform_average', the mean of :math:`EV` is + calculated: + + .. math:: + mean(EV) = \frac{1}{V}\int_{D} EV(t) dt + + where :math:`D` is the function domain and :math:`V` the volume of that + domain. + + For :class:`~skfda.representation.FDataBasis` only + 'uniform_average' is available. + + If :math:`y\_true` and :math:`y\_pred` are numpy arrays, sklearn function + is called. + + The best possible score is 1.0, lower values are worse. + + Args: + y_true: Correct target values. + y_pred: Estimated values. + sample_weight: Sample weights. By default, uniform weights + are taken. + multioutput: Defines format of the return. + + Returns: + Explained variance score. + + If multioutput = 'uniform_average' or + :math:`y\_pred` and :math:`y\_true` are + :class:`~skfda.representation.FDataBasis` objects, float is returned. + + If both :math:`y\_pred` and :math:`y\_true` are + :class:`~skfda.representation.FDataGrid` + objects and multioutput = 'raw_values', + :class:`~skfda.representation.FDataGrid` is returned. + + If both :math:`y\_pred` and :math:`y\_true` are ndarray and + multioutput = 'raw_values', ndarray. + + """ + return ( # type: ignore [no-any-return] + sklearn.metrics.explained_variance_score( + y_true, + y_pred, + sample_weight=sample_weight, + multioutput=multioutput, + ) + ) + + +@explained_variance_score.register # type: ignore[attr-defined, misc] +def _explained_variance_score_fdatagrid( + y_true: FDataGrid, + y_pred: FDataGrid, + *, + sample_weight: Optional[NDArrayFloat] = None, + multioutput: MultiOutputType = 'uniform_average', +) -> Union[float, FDataGrid]: + + num = _var(y_true - y_pred, weights=sample_weight) + den = _var(y_true, weights=sample_weight) + + # Divisions by zero allowed + with np.errstate(divide='ignore', invalid='ignore'): + score = 1 - num / den + + # 0 / 0 divisions should be 0 in this context, and the score, 1 + score.data_matrix[np.isnan(score.data_matrix)] = 1 + + return _multioutput_score_grid(score, multioutput) + + +@explained_variance_score.register # type: ignore[attr-defined, misc] +def _explained_variance_score_fdatabasis( + y_true: FDataBasis, + y_pred: FDataBasis, + *, + sample_weight: Optional[NDArrayFloat] = None, + multioutput: MultiOutputType = 'uniform_average', +) -> float: + + def _ev_func(x: EvalPointsType) -> NDArrayFloat: # noqa: WPS430 + num = np.average( + ( + (y_true(x) - y_pred(x)) + - np.average( + y_true(x) - y_pred(x), + weights=sample_weight, + axis=0, + ) + ) ** 2, + weights=sample_weight, + axis=0, + ) + + den = np.average( + ( + y_true(x) + - np.average(y_true(x), weights=sample_weight, axis=0) + ) ** 2, + weights=sample_weight, + axis=0, + ) + + # Divisions by zero allowed + with np.errstate(divide='ignore', invalid='ignore'): + score = 1 - num / den + + # 0/0 case, the score is 1. + score[np.isnan(score)] = 1 + + # r/0 case, r!= 0. Return -inf outside this function + if np.any(np.isinf(score)): + raise _InfiniteScoreError + + # Score only contains 1 input point + assert score.shape[0] == 1 + return score[0] # type: ignore [no-any-return] + + return _multioutput_score_basis(y_true, multioutput, _ev_func) + + +@overload +def mean_absolute_error( + y_true: DataType, + y_pred: DataType, + *, + sample_weight: NDArrayFloat | None = None, + multioutput: Literal['uniform_average'] = 'uniform_average', +) -> float: + pass # noqa: WPS428 + + +@overload +def mean_absolute_error( + y_true: DataType, + y_pred: DataType, + *, + sample_weight: NDArrayFloat | None = None, + multioutput: Literal['raw_values'], +) -> DataType: + pass # noqa: WPS428 + + +@singledispatch +def mean_absolute_error( + y_true: DataType, + y_pred: DataType, + *, + sample_weight: NDArrayFloat | None = None, + multioutput: MultiOutputType = 'uniform_average', +) -> float | DataType: + r"""Mean Absolute Error for :class:`~skfda.representation.FData`. + + With :math:`y\_true = (X_1, X_2, ..., X_n)` being the real values, + :math:`t\_pred = (\hat{X}_1, \hat{X}_2, ..., \hat{X}_n)` being the + estimated and :math:`sample\_weight = (w_1, w_2, ..., w_n)`, the error is + calculated as + + .. math:: + MAE(y\_true, y\_pred)(t) = \frac{1}{\sum w_i} + \sum_{i=1}^n w_i|X_i(t) - \hat{X}_i(t)| + + For :math:`y\_true` and :math:`y\_pred` of type + :class:`~skfda.representation.FDataGrid`, :math:`MAE` is + also a :class:`~skfda.representation.FDataGrid` object with + the same grid points. + + If multioutput = 'raw_values', the function :math:`MAE` is returned. + Otherwise, if multioutput = 'uniform_average', the mean of :math:`MAE` is + calculated: + + .. math:: + mean(MAE) = \frac{1}{V}\int_{D} MAE(t) dt + + where :math:`D` is the function domain and :math:`V` the volume of that + domain. + + For :class:`~skfda.representation.FDataBasis` only + 'uniform_average' is available. + + If :math:`y\_true` and :math:`y\_pred` are numpy arrays, sklearn function + is called. + + Args: + y_true: Correct target values. + y_pred: Estimated values. + sample_weight: Sample weights. By default, uniform weights + are taken. + multioutput: Defines format of the return. + + Returns: + Mean absolute error. + + If multioutput = 'uniform_average' or + :math:`y\_pred` and :math:`y\_true` are + :class:`~skfda.representation.FDataBasis` objects, float is returned. + + If both :math:`y\_pred` and :math:`y\_true` are + :class:`~skfda.representation.FDataGrid` + objects and multioutput = 'raw_values', + :class:`~skfda.representation.FDataGrid` is returned. + + If both :math:`y\_pred` and :math:`y\_true` are ndarray and + multioutput = 'raw_values', ndarray. + + """ + return sklearn.metrics.mean_absolute_error( # type: ignore [no-any-return] + y_true, + y_pred, + sample_weight=sample_weight, + multioutput=multioutput, + ) + + +@mean_absolute_error.register # type: ignore[attr-defined, misc] +def _mean_absolute_error_fdatagrid( + y_true: FDataGrid, + y_pred: FDataGrid, + *, + sample_weight: Optional[NDArrayFloat] = None, + multioutput: MultiOutputType = 'uniform_average', +) -> Union[float, FDataGrid]: + from ..exploratory.stats import mean + + error = mean(np.abs(y_true - y_pred), weights=sample_weight) + return _multioutput_score_grid(error, multioutput) + + +@mean_absolute_error.register # type: ignore[attr-defined, misc] +def _mean_absolute_error_fdatabasis( + y_true: FDataBasis, + y_pred: FDataBasis, + *, + sample_weight: Optional[NDArrayFloat] = None, + multioutput: MultiOutputType = 'uniform_average', +) -> float: + + def _mae_func(x: EvalPointsType) -> NDArrayFloat: # noqa: WPS430 + error = np.average( + np.abs(y_true(x) - y_pred(x)), + weights=sample_weight, + axis=0, + ) + + # Error only contains 1 input point + assert error.shape[0] == 1 + return error[0] # type: ignore [no-any-return] + + return _multioutput_score_basis(y_true, multioutput, _mae_func) + + +@overload +def mean_absolute_percentage_error( + y_true: DataType, + y_pred: DataType, + *, + sample_weight: NDArrayFloat | None = None, + multioutput: Literal['uniform_average'] = 'uniform_average', +) -> float: + pass # noqa: WPS428 + + +@overload +def mean_absolute_percentage_error( + y_true: DataType, + y_pred: DataType, + *, + sample_weight: NDArrayFloat | None = None, + multioutput: Literal['raw_values'], +) -> DataType: + pass # noqa: WPS428 + + +@singledispatch +def mean_absolute_percentage_error( + y_true: DataType, + y_pred: DataType, + *, + sample_weight: NDArrayFloat | None = None, + multioutput: MultiOutputType = 'uniform_average', +) -> float | DataType: + r"""Mean Absolute Percentage Error for :class:`~skfda.representation.FData`. + + With :math:`y\_true = (X_1, X_2, ..., X_n)` being the real values, + :math:`t\_pred = (\hat{X}_1, \hat{X}_2, ..., \hat{X}_n)` being the + estimated and :math:`sample\_weight = (w_1, w_2, ..., w_n)`, the error is + calculated as + + .. math:: + MAPE(y\_true, y\_pred)(t) = \frac{1}{\sum w_i} + \sum_{i=1}^n w_i\frac{|X_i(t) - \hat{X}_i(t)|}{|X_i(t)|} + + For :math:`y\_true` and :math:`y\_pred` of type + :class:`~skfda.representation.FDataGrid`, :math:`MAPE` is + also a :class:`~skfda.representation.FDataGrid` object with + the same grid points. + + If multioutput = 'raw_values', the function :math:`MAPE` is returned. + Otherwise, if multioutput = 'uniform_average', the mean of :math:`MAPE` is + calculated: + + .. math:: + mean(MAPE) = \frac{1}{V}\int_{D} MAPE(t) dt + + where :math:`D` is the function domain and :math:`V` the volume of that + domain. + + For :class:`~skfda.representation.FDataBasis` only + 'uniform_average' is available. + + If :math:`y\_true` and :math:`y\_pred` are numpy arrays, sklearn function + is called. + + This function should not be used if for some :math:`t` and some :math:`i`, + :math:`X_i(t) = 0`. + + Args: + y_true: Correct target values. + y_pred: Estimated values. + sample_weight: Sample weights. By default, uniform weights + are taken. + multioutput: Defines format of the return. + + Returns: + Mean absolute percentage error. + + If multioutput = 'uniform_average' or + :math:`y\_pred` and :math:`y\_true` are + :class:`~skfda.representation.FDataBasis` objects, float is returned. + + If both :math:`y\_pred` and :math:`y\_true` are + :class:`~skfda.representation.FDataGrid` + objects and multioutput = 'raw_values', + :class:`~skfda.representation.FDataGrid` is returned. + + If both :math:`y\_pred` and :math:`y\_true` are ndarray and + multioutput = 'raw_values', ndarray. + + """ + return ( # type: ignore [no-any-return] + sklearn.metrics.mean_absolute_percentage_error( + y_true, + y_pred, + sample_weight=sample_weight, + multioutput=multioutput, + ) + ) + + +@mean_absolute_percentage_error.register # type: ignore[attr-defined, misc] +def _mean_absolute_percentage_error_fdatagrid( + y_true: FDataGrid, + y_pred: FDataGrid, + *, + sample_weight: Optional[NDArrayFloat] = None, + multioutput: MultiOutputType = 'uniform_average', +) -> Union[float, FDataGrid]: + from ..exploratory.stats import mean + + epsilon = np.finfo(np.float64).eps + + if np.any(np.abs(y_true.data_matrix) < epsilon): + warnings.warn('Zero denominator', RuntimeWarning) + + mape = np.abs(y_pred - y_true) / np.maximum(np.abs(y_true), epsilon) + + error = mean(mape, weights=sample_weight) + return _multioutput_score_grid(error, multioutput) + + +@mean_absolute_percentage_error.register # type: ignore[attr-defined, misc] +def _mean_absolute_percentage_error_fdatabasis( + y_true: FDataBasis, + y_pred: FDataBasis, + *, + sample_weight: Optional[NDArrayFloat] = None, + multioutput: MultiOutputType = 'uniform_average', +) -> float: + + def _mape_func(x: EvalPointsType) -> NDArrayFloat: # noqa: WPS430 + + epsilon = np.finfo(np.float64).eps + if np.any(np.abs(y_true(x)) < epsilon): + warnings.warn('Zero denominator', RuntimeWarning) + + error = np.average( + ( + np.abs(y_true(x) - y_pred(x)) + / np.maximum(np.abs(y_true(x)), epsilon) + ), + weights=sample_weight, + axis=0, + ) + + # Error only contains 1 input point + assert error.shape[0] == 1 + return error[0] # type: ignore [no-any-return] + + return _multioutput_score_basis(y_true, multioutput, _mape_func) + + +@overload +def mean_squared_error( + y_true: DataType, + y_pred: DataType, + *, + sample_weight: NDArrayFloat | None = None, + multioutput: Literal['uniform_average'] = 'uniform_average', + squared: bool = True, +) -> float: + pass # noqa: WPS428 + + +@overload +def mean_squared_error( + y_true: DataType, + y_pred: DataType, + *, + sample_weight: NDArrayFloat | None = None, + multioutput: Literal['raw_values'], + squared: bool = True, +) -> DataType: + pass # noqa: WPS428 + + +@singledispatch +def mean_squared_error( + y_true: DataType, + y_pred: DataType, + *, + sample_weight: NDArrayFloat | None = None, + multioutput: MultiOutputType = 'uniform_average', + squared: bool = True, +) -> float | DataType: + r"""Mean Squared Error for :class:`~skfda.representation.FData`. + + With :math:`y\_true = (X_1, X_2, ..., X_n)` being the real values, + :math:`t\_pred = (\hat{X}_1, \hat{X}_2, ..., \hat{X}_n)` being the + estimated and :math:`sample\_weight = (w_1, w_2, ..., w_n)`, the error is + calculated as + + .. math:: + MSE(y\_true, y\_pred)(t) = \frac{1}{\sum w_i} + \sum_{i=1}^n w_i(X_i(t) - \hat{X}_i(t))^2 + + For :math:`y\_true` and :math:`y\_pred` of type + :class:`~skfda.representation.FDataGrid`, :math:`MSE` is + also a :class:`~skfda.representation.FDataGrid` object with + the same grid points. + + If multioutput = 'raw_values', the function :math:`MSE` is returned. + Otherwise, if multioutput = 'uniform_average', the mean of :math:`MSE` is + calculated: + + .. math:: + mean(MSE) = \frac{1}{V}\int_{D} MSE(t) dt + + where :math:`D` is the function domain and :math:`V` the volume of that + domain. + + For :class:`~skfda.representation.FDataBasis` only + 'uniform_average' is available. + + If :math:`y\_true` and :math:`y\_pred` are numpy arrays, sklearn function + is called. + + Args: + y_true: Correct target values. + y_pred: Estimated values. + sample_weight: Sample weights. By default, uniform weights + are taken. + multioutput: Defines format of the return. + squared: If True returns MSE value, if False returns RMSE value. + + Returns: + Mean squared error. + + If multioutput = 'uniform_average' or + :math:`y\_pred` and :math:`y\_true` are + :class:`~skfda.representation.FDataBasis` objects, float is returned. + + If both :math:`y\_pred` and :math:`y\_true` are + :class:`~skfda.representation.FDataGrid` + objects and multioutput = 'raw_values', + :class:`~skfda.representation.FDataGrid` is returned. + + If both :math:`y\_pred` and :math:`y\_true` are ndarray and + multioutput = 'raw_values', ndarray. + + """ + return sklearn.metrics.mean_squared_error( # type: ignore [no-any-return] + y_true, + y_pred, + sample_weight=sample_weight, + multioutput=multioutput, + squared=squared, + ) + + +@mean_squared_error.register # type: ignore[attr-defined, misc] +def _mean_squared_error_fdatagrid( + y_true: FDataGrid, + y_pred: FDataGrid, + *, + sample_weight: Optional[NDArrayFloat] = None, + multioutput: MultiOutputType = 'uniform_average', + squared: bool = True, +) -> Union[float, FDataGrid]: + from ..exploratory.stats import mean + + error: FDataGrid = mean( + np.power(y_true - y_pred, 2), + weights=sample_weight, + ) + + return _multioutput_score_grid(error, multioutput, squared=squared) + + +@mean_squared_error.register # type: ignore[attr-defined, misc] +def _mean_squared_error_fdatabasis( + y_true: FDataBasis, + y_pred: FDataBasis, + *, + sample_weight: Optional[NDArrayFloat] = None, + squared: bool = True, + multioutput: MultiOutputType = 'uniform_average', +) -> float: + + def _mse_func(x: EvalPointsType) -> NDArrayFloat: # noqa: WPS430 + + error: NDArrayFloat = np.average( + (y_true(x) - y_pred(x)) ** 2, + weights=sample_weight, + axis=0, + ) + + if not squared: + return np.sqrt(error) + + # Error only contains 1 input point + assert error.shape[0] == 1 + return error[0] # type: ignore [no-any-return] + + return _multioutput_score_basis(y_true, multioutput, _mse_func) + + +@overload +def mean_squared_log_error( + y_true: DataType, + y_pred: DataType, + *, + sample_weight: NDArrayFloat | None = None, + multioutput: Literal['uniform_average'] = 'uniform_average', + squared: bool = True, +) -> float: + pass # noqa: WPS428 + + +@overload +def mean_squared_log_error( + y_true: DataType, + y_pred: DataType, + *, + sample_weight: NDArrayFloat | None = None, + multioutput: Literal['raw_values'], + squared: bool = True, +) -> DataType: + pass # noqa: WPS428 + + +@singledispatch +def mean_squared_log_error( + y_true: DataType, + y_pred: DataType, + *, + sample_weight: NDArrayFloat | None = None, + multioutput: MultiOutputType = 'uniform_average', + squared: bool = True, +) -> float | DataType: + r"""Mean Squared Log Error for :class:`~skfda.representation.FData`. + + With :math:`y\_true = (X_1, X_2, ..., X_n)` being the real values, + :math:`t\_pred = (\hat{X}_1, \hat{X}_2, ..., \hat{X}_n)` being the + estimated and :math:`sample\_weight = (w_1, w_2, ..., w_n)`, the error is + calculated as + + .. math:: + MSLE(y\_true, y\_pred)(t) = \frac{1}{\sum w_i} + \sum_{i=1}^n w_i(\log(1 + X_i(t)) - \log(1 + \hat{X}_i(t)))^2 + + where :math:`\log` is the natural logarithm. + + For :math:`y\_true` and :math:`y\_pred` of type + :class:`~skfda.representation.FDataGrid`, :math:`MSLE` is + also a :class:`~skfda.representation.FDataGrid` object with + the same grid points. + + If multioutput = 'raw_values', the function :math:`MSLE` is returned. + Otherwise, if multioutput = 'uniform_average', the mean of :math:`MSLE` is + calculated: + + .. math:: + mean(MSLE) = \frac{1}{V}\int_{D} MSLE(t) dt + + where :math:`D` is the function domain and :math:`V` the volume of that + domain. + + For :class:`~skfda.representation.FDataBasis` only + 'uniform_average' is available. + + If :math:`y\_true` and :math:`y\_pred` are numpy arrays, sklearn function + is called. + + This function should not be used if for some :math:`t` and some :math:`i`, + :math:`X_i(t) < 0`. + + Args: + y_true: Correct target values. + y_pred: Estimated values. + sample_weight: Sample weights. By default, uniform weights + are taken. + multioutput: Defines format of the return. + squared: default True. If False, square root is taken. + + Returns: + Mean squared log error. + + If multioutput = 'uniform_average' or + :math:`y\_pred` and :math:`y\_true` are + :class:`~skfda.representation.FDataBasis` objects, float is returned. + + If both :math:`y\_pred` and :math:`y\_true` are + :class:`~skfda.representation.FDataGrid` + objects and multioutput = 'raw_values', + :class:`~skfda.representation.FDataGrid` is returned. + + If both :math:`y\_pred` and :math:`y\_true` are ndarray and + multioutput = 'raw_values', ndarray. + + """ + return ( # type: ignore [no-any-return] + sklearn.metrics.mean_squared_log_error( + y_true, + y_pred, + sample_weight=sample_weight, + multioutput=multioutput, + squared=squared, + ) + ) + + +@mean_squared_log_error.register # type: ignore[attr-defined, misc] +def _mean_squared_log_error_fdatagrid( + y_true: FDataGrid, + y_pred: FDataGrid, + *, + sample_weight: Optional[NDArrayFloat] = None, + multioutput: MultiOutputType = 'uniform_average', + squared: bool = True, +) -> Union[float, FDataGrid]: + + if np.any(y_true.data_matrix < 0) or np.any(y_pred.data_matrix < 0): + raise ValueError( + "Mean Squared Logarithmic Error cannot be used when " + "targets functions have negative values.", + ) + + return mean_squared_error( + np.log1p(y_true), + np.log1p(y_pred), + sample_weight=sample_weight, + multioutput=multioutput, + squared=squared, + ) + + +@mean_squared_log_error.register # type: ignore[attr-defined, misc] +def _mean_squared_log_error_fdatabasis( + y_true: FDataBasis, + y_pred: FDataBasis, + *, + sample_weight: Optional[NDArrayFloat] = None, + squared: bool = True, + multioutput: MultiOutputType = 'uniform_average', +) -> float: + + def _msle_func(x: EvalPointsType) -> NDArrayFloat: # noqa: WPS430 + + y_true_eval = y_true(x) + y_pred_eval = y_pred(x) + + if np.any(y_true_eval < 0) or np.any(y_pred_eval < 0): + raise ValueError( + "Mean Squared Logarithmic Error cannot be used when " + "targets functions have negative values.", + ) + + error: NDArrayFloat = np.average( + (np.log1p(y_true_eval) - np.log1p(y_pred_eval)) ** 2, + weights=sample_weight, + axis=0, + ) + + if not squared: + return np.sqrt(error) + + # Error only contains 1 input point + assert error.shape[0] == 1 + return error[0] # type: ignore [no-any-return] + + return _multioutput_score_basis(y_true, multioutput, _msle_func) + + +@overload +def r2_score( + y_true: DataType, + y_pred: DataType, + *, + sample_weight: NDArrayFloat | None = None, + multioutput: Literal['uniform_average'] = 'uniform_average', +) -> float: + pass # noqa: WPS428 + + +@overload +def r2_score( + y_true: DataType, + y_pred: DataType, + *, + sample_weight: NDArrayFloat | None = None, + multioutput: Literal['raw_values'], +) -> DataType: + pass # noqa: WPS428 + + +@singledispatch +def r2_score( + y_true: DataType, + y_pred: DataType, + *, + sample_weight: NDArrayFloat | None = None, + multioutput: MultiOutputType = 'uniform_average', +) -> float | DataType: + r"""R^2 score for :class:`~skfda.representation.FData`. + + With :math:`y\_true = (X_1, X_2, ..., X_n)` being the real values, + :math:`t\_pred = (\hat{X}_1, \hat{X}_2, ..., \hat{X}_n)` being the + estimated and :math:`sample\_weight = (w_1, w_2, ..., w_n)`, the score is + calculated as + + .. math:: + R^2(y\_true, y\_pred)(t) = 1 - + \frac{\sum_{i=1}^n w_i (X_i(t) - \hat{X}_i(t))^2} + {\sum_{i=1}^n w_i (X_i(t) - Mean(y\_true, sample\_weight)(t))^2} + + where :math:`Mean` is a weighted mean. + + For :math:`y\_true` and :math:`y\_pred` of type + :class:`~skfda.representation.FDataGrid`, :math:`R^2` is + also a :class:`~skfda.representation.FDataGrid` object with + the same grid points. + + If multioutput = 'raw_values', the function :math:`R^2` is returned. + Otherwise, if multioutput = 'uniform_average', the mean of :math:`R^2` is + calculated: + + .. math:: + mean(R^2) = \frac{1}{V}\int_{D} R^2(t) dt + + where :math:`D` is the function domain and :math:`V` the volume of that + domain. + + For :class:`~skfda.representation.FDataBasis` only + 'uniform_average' is available. + + If :math:`y\_true` and :math:`y\_pred` are numpy arrays, sklearn function + is called. + + Args: + y_true: Correct target values. + y_pred: Estimated values. + sample_weight: Sample weights. By default, uniform weights + are taken. + multioutput: Defines format of the return. + + Returns: + R2 score + + If multioutput = 'uniform_average' or + :math:`y\_pred` and :math:`y\_true` are + :class:`~skfda.representation.FDataBasis` objects, float is returned. + + If both :math:`y\_pred` and :math:`y\_true` are + :class:`~skfda.representation.FDataGrid` + objects and multioutput = 'raw_values', + :class:`~skfda.representation.FDataGrid` is returned. + + If both :math:`y\_pred` and :math:`y\_true` are ndarray and + multioutput = 'raw_values', ndarray. + + """ + return sklearn.metrics.r2_score( # type: ignore [no-any-return] + y_true, + y_pred, + sample_weight=sample_weight, + multioutput=multioutput, + ) + + +@r2_score.register # type: ignore[attr-defined, misc] +def _r2_score_fdatagrid( + y_true: FDataGrid, + y_pred: FDataGrid, + *, + sample_weight: Optional[NDArrayFloat] = None, + multioutput: MultiOutputType = 'uniform_average', +) -> Union[float, FDataGrid]: + from ..exploratory.stats import mean + + if y_pred.n_samples < 2: + raise ValueError( + 'R^2 score is not well-defined with less than two samples.', + ) + + ss_res = mean( + np.power(y_true - y_pred, 2), + weights=sample_weight, + ) + + ss_tot = _var(y_true, weights=sample_weight) + + # Divisions by zero allowed + with np.errstate(divide='ignore', invalid='ignore'): + score: FDataGrid = 1 - ss_res / ss_tot + + # 0 / 0 divisions should be 0 in this context and the score, 1 + score.data_matrix[np.isnan(score.data_matrix)] = 1 + + return _multioutput_score_grid(score, multioutput) + + +@r2_score.register # type: ignore[attr-defined, misc] +def _r2_score_fdatabasis( + y_true: FDataBasis, + y_pred: FDataBasis, + *, + sample_weight: Optional[NDArrayFloat] = None, + multioutput: MultiOutputType = 'uniform_average', +) -> float: + + if y_pred.n_samples < 2: + raise ValueError( + 'R^2 score is not well-defined with less than two samples.', + ) + + def _r2_func(x: NDArrayFloat) -> NDArrayFloat: # noqa: WPS430 + ss_res = np.average( + (y_true(x) - y_pred(x)) ** 2, + weights=sample_weight, + axis=0, + ) + + ss_tot = np.average( + ( + y_true(x) + - np.average(y_true(x), weights=sample_weight, axis=0) + ) ** 2, + weights=sample_weight, + axis=0, + ) + + # Divisions by zero allowed + with np.errstate(divide='ignore', invalid='ignore'): + score = 1 - ss_res / ss_tot + + # 0/0 case, the score is 1. + score[np.isnan(score)] = 1 + + # r/0 case, r!= 0. Return -inf outside this function + if np.any(np.isinf(score)): + raise _InfiniteScoreError + + # Score only had 1 input point + assert score.shape[0] == 1 + return score[0] # type: ignore [no-any-return] + + return _multioutput_score_basis(y_true, multioutput, _r2_func) diff --git a/skfda/ml/_neighbors_base.py b/skfda/ml/_neighbors_base.py index 3beb3b44c..f655ff7d8 100644 --- a/skfda/ml/_neighbors_base.py +++ b/skfda/ml/_neighbors_base.py @@ -662,132 +662,3 @@ def _functional_predict( ) return concatenate(iterable) # type: ignore[no-any-return] - - def score( - self, - X: Input, - y: TargetRegression, - sample_weight: NDArrayFloat | None = None, - ) -> float: - r"""Return the coefficient of determination R^2 of the prediction. - - In the multivariate response case, the coefficient :math:`R^2` is - defined as - - .. math:: - 1 - \frac{\sum_{i=1}^{n} (y_i - \hat y_i)^2} - {\sum_{i=1}^{n} (y_i - \frac{1}{n}\sum_{i=1}^{n}y_i)^2} - - where :math:`\hat{y}_i` is the prediction associated to the test sample - :math:`X_i`, and :math:`{y}_i` is the true response. See - :func:`sklearn.metrics.r2_score <sklearn.metrics.r2_score>` for more - information. - - - In the functional case it is returned an extension of the coefficient - of determination :math:`R^2`, defined as - - .. math:: - 1 - \frac{\sum_{i=1}^{n}\int (y_i(t) - \hat{y}_i(t))^2dt} - {\sum_{i=1}^{n} \int (y_i(t)- \frac{1}{n}\sum_{i=1}^{n}y_i(t))^2dt} - - - The best possible score is 1.0 and it can be negative - (because the model can be arbitrarily worse). A constant model that - always predicts the expected value of y, disregarding the input - features, would get a R^2 score of 0.0. - - Args: - X: Test samples to be predicted. - y: True responses of the test samples. - sample_weight: Sample weights. - - Returns: - Coefficient of determination. - - """ - if self._functional: - return self._functional_score(X, y, sample_weight=sample_weight) - - # Default sklearn multivariate score - return super().score(X, y, sample_weight=sample_weight) - - def _functional_score( - self: NeighborsRegressorMixin[Input, TargetRegressionFData], - X: Input, - y: TargetRegressionFData, - sample_weight: NDArrayFloat | None = None, - ) -> float: - r""" - Return an extension of the coefficient of determination R^2. - - The coefficient is defined as - - .. math:: - 1 - \frac{\sum_{i=1}^{n}\int (y_i(t) - \hat{y}_i(t))^2dt} - {\sum_{i=1}^{n} \int (y_i(t)- \frac{1}{n}\sum_{i=1}^{n}y_i(t))^2dt} - - where :math:`\hat{y}_i` is the prediction associated to the test sample - :math:`X_i`, and :math:`{y}_i` is the true response. - - The best possible score is 1.0 and it can be negative - (because the model can be arbitrarily worse). A constant model that - always predicts the expected value of y, disregarding the input - features, would get a R^2 score of 0.0. - - Args: - X: Test samples to be predicted. - y: True responses of the test samples. - sample_weight (array_like, shape = [n_samples], optional): Sample - weights. - - Returns: - Coefficient of determination. - - """ - # TODO: If it is created a module in ml.regression with other - # score metrics, move it. - from scipy.integrate import simps - - if y.dim_codomain != 1 or y.dim_domain != 1: - raise ValueError( - "Score not implemented for multivariate " - "functional data.", - ) - - # Make prediction - pred = self.predict(X) - - u = y - pred - v = y - y.mean() - - # Discretize to integrate and make squares if needed - if not isinstance(u, FDataGrid): - u = u.to_grid() - v = v.to_grid() - - data_u = u.data_matrix[..., 0] - data_v = v.data_matrix[..., 0] - - # Square without allocate more memory - np.square(data_u, out=data_u) - np.square(data_v, out=data_v) - - if sample_weight is not None: - if len(sample_weight) != len(y): - raise ValueError("Must be a weight for each sample.") - - normalized_sample_weight = sample_weight / sample_weight.sum() - data_u_t = data_u.T - data_u_t *= normalized_sample_weight - data_v_t = data_v.T - data_v_t *= normalized_sample_weight - - # Sum and integrate - sum_u = np.sum(data_u, axis=0) - sum_v = np.sum(data_v, axis=0) - - int_u = simps(sum_u, x=u.grid_points[0]) - int_v = simps(sum_v, x=v.grid_points[0]) - - return float(1 - int_u / int_v) diff --git a/skfda/representation/grid.py b/skfda/representation/grid.py index 4df8e6cfe..855923b1b 100644 --- a/skfda/representation/grid.py +++ b/skfda/representation/grid.py @@ -707,6 +707,11 @@ def _get_op_matrix( return other[other_index] + raise ValueError( + f"Invalid dimensions in operator between FDataGrid and Numpy " + f"array: {other.shape}" + ) + elif isinstance(other, FDataGrid): self._check_same_dimensions(other) return other.data_matrix diff --git a/skfda/representation/interpolation.py b/skfda/representation/interpolation.py index b546766d4..1c107564b 100644 --- a/skfda/representation/interpolation.py +++ b/skfda/representation/interpolation.py @@ -183,6 +183,8 @@ def _get_interpolator_1d( fdatagrid.data_matrix, k=self.interpolation_order, axis=1, + # Orders 0 and 1 behave well + check_finite=self.interpolation_order > 1, ) def _get_interpolator_nd(

Scorers for functional response Implement scoring methods for functional response (to be used for example in validation and hyperparameter search).

2022-03-04T01:06:11

0.0

[]

GAA-UAM/scikit-fda

GAA-UAM__scikit-fda-408

49d6ee767950872b2c46c3602ceb3166f0a866dd

diff --git a/skfda/preprocessing/registration/_landmark_registration.py b/skfda/preprocessing/registration/_landmark_registration.py index e96b5296b..cc02eec6d 100644 --- a/skfda/preprocessing/registration/_landmark_registration.py +++ b/skfda/preprocessing/registration/_landmark_registration.py @@ -42,9 +42,8 @@ def landmark_shift_deltas( passed the location will be the result of the the call, the function should be accept as an unique parameter a numpy array with the list of landmarks. - By default it will be used as location :math:`\frac{1}{2}(max( - \text{landmarks})+ min(\text{landmarks}))` wich minimizes the - max shift. + By default it will be used as location the mean of the original + locations of the landmarks. Returns: Array containing the corresponding shifts. @@ -68,7 +67,7 @@ def landmark_shift_deltas( >>> shifts = landmark_shift_deltas(fd, landmarks) >>> shifts.round(3) - array([ 0.25 , -0.25 , -0.231]) + array([ 0.327, -0.173, -0.154]) The registered samples can be obtained with a shift @@ -88,10 +87,7 @@ def landmark_shift_deltas( # Parses location if location is None: - loc_array = ( - np.max(landmarks, axis=0) - + np.min(landmarks, axis=0) - ) / 2 + loc_array = np.mean(landmarks) elif callable(location): loc_array = location(landmarks) else:

Unify selection of target locations for landmark_shift_registration and landmark_elastic_registration In `landmark_shift_registration` the location of the target landmarks is by default selected as the average of the maximum and minimum values for the location in all curves. In `landmark_elastic_registration` those are instead selected as the means of the locations. We should use that latest method for both, for consistency.

2022-01-07T15:08:07

0.0

[]

GAA-UAM/scikit-fda

GAA-UAM__scikit-fda-392

24a3541a01099467bfc27efe5752bfc1ca5f76d9

diff --git a/skfda/representation/_functional_data.py b/skfda/representation/_functional_data.py index 22ac35a8a..9f25ac2cd 100644 --- a/skfda/representation/_functional_data.py +++ b/skfda/representation/_functional_data.py @@ -33,6 +33,7 @@ GridPointsLike, LabelTuple, LabelTupleLike, + NDArrayInt, ) from .evaluator import Evaluator from .extrapolation import ExtrapolationLike, _parse_extrapolation @@ -1049,6 +1050,24 @@ def __array__(self, *args: Any, **kwargs: Any) -> np.ndarray: return array + def __array_ufunc__( + self, + ufunc: Any, + method: str, + *inputs: Any, + **kwargs: Any, + ) -> Any: + """Prevent NumPy from converting to array just to do operations.""" + + # Make normal multiplication by scalar use the __mul__ method + if ufunc == np.multiply and method == "__call__" and len(inputs) == 2: + if isinstance(inputs[0], np.ndarray): + inputs = inputs[::-1] + + return inputs[0] * inputs[1] + + return NotImplemented + ##################################################################### # Pandas ExtensionArray methods ##################################################################### @@ -1100,9 +1119,17 @@ def _from_factorized(cls, values: Any, original: Any) -> NoReturn: "Factorization does not make sense for functional data", ) + @abstractmethod + def _take_allow_fill( + self: T, + indices: NDArrayInt, + fill_value: T, + ) -> T: + pass + def take( self: T, - indices: Sequence[int], + indices: Union[int, Sequence[int], NDArrayInt], allow_fill: bool = False, fill_value: Optional[T] = None, axis: int = 0, @@ -1148,28 +1175,44 @@ def take( numpy.take pandas.api.extensions.take """ - from pandas.core.algorithms import take - # The axis parameter must exist, because sklearn tries to use take # instead of __getitem__ if axis != 0: raise ValueError(f"Axis must be 0, not {axis}") - # If the ExtensionArray is backed by an ndarray, then - # just pass that here instead of coercing to object. - data = np.asarray(self) - if allow_fill and fill_value is None: + arr_indices = np.atleast_1d(indices) + + if fill_value is None: fill_value = self.dtype.na_value - # fill value should always be translated from the scalar - # type for the array, to the physical storage type for - # the data, before passing to take. - result = take( - data, - indices, - fill_value=fill_value, - allow_fill=allow_fill, - ) - return self._from_sequence(result, dtype=self.dtype) + + non_empty_take_msg = "cannot do a non-empty take from an empty axes" + + if allow_fill: + if (arr_indices < -1).any(): + raise ValueError("Invalid indexes") + + positive_mask = arr_indices >= 0 + if len(self) == 0 and positive_mask.any(): + raise IndexError(non_empty_take_msg) + + sample_names = np.zeros(len(arr_indices), dtype=object) + result = self._take_allow_fill(arr_indices, fill_value) + + sample_names[positive_mask] = np.array(self.sample_names)[ + arr_indices[positive_mask] + ] + + if fill_value is not self.dtype.na_value: + sample_names[~positive_mask] = fill_value.sample_names[0] + + result.sample_names = tuple(sample_names) + else: + if len(self) == 0 and len(arr_indices) != 0: + raise IndexError(non_empty_take_msg) + + result = self[arr_indices] + + return result @classmethod def _concat_same_type( @@ -1198,6 +1241,7 @@ def astype(self, dtype: Any, copy: bool = True) -> Any: if copy: new_obj = self.copy() return new_obj + return super().astype(dtype) def _reduce(self, name: str, skipna: bool = True, **kwargs: Any) -> Any: diff --git a/skfda/representation/basis/_fdatabasis.py b/skfda/representation/basis/_fdatabasis.py index a9071e7f4..fa28d94ea 100644 --- a/skfda/representation/basis/_fdatabasis.py +++ b/skfda/representation/basis/_fdatabasis.py @@ -23,7 +23,13 @@ from ..._utils import _check_array_key, _int_to_real, constants from .. import grid from .._functional_data import FData -from .._typing import ArrayLike, DomainRange, GridPointsLike, LabelTupleLike +from .._typing import ( + ArrayLike, + DomainRange, + GridPointsLike, + LabelTupleLike, + NDArrayInt, +) from ..extrapolation import ExtrapolationLike from . import Basis @@ -866,6 +872,27 @@ def __rtruediv__(self: T, other: Union[np.ndarray, float]) -> T: ##################################################################### # Pandas ExtensionArray methods ##################################################################### + def _take_allow_fill( + self: T, + indices: NDArrayInt, + fill_value: T, + ) -> T: + result = self.copy() + result.coefficients = np.full( + (len(indices),) + self.coefficients.shape[1:], + np.nan, + ) + + positive_mask = indices >= 0 + result.coefficients[positive_mask] = self.coefficients[ + indices[positive_mask] + ] + + if fill_value is not self.dtype.na_value: + result.coefficients[~positive_mask] = fill_value.coefficients[0] + + return result + @property def dtype(self) -> FDataBasisDType: """The dtype for this extension array, FDataGridDType""" diff --git a/skfda/representation/grid.py b/skfda/representation/grid.py index 9f3fb1bef..2a1ef1092 100644 --- a/skfda/representation/grid.py +++ b/skfda/representation/grid.py @@ -25,9 +25,8 @@ import findiff import numpy as np import pandas.api.extensions -from matplotlib.figure import Figure - import scipy.stats.mstats +from matplotlib.figure import Figure from .._utils import ( _check_array_key, @@ -44,6 +43,7 @@ GridPoints, GridPointsLike, LabelTupleLike, + NDArrayInt, ) from .basis import Basis from .evaluator import Evaluator @@ -1260,7 +1260,7 @@ def __array_ufunc__( new_inputs = [ i.data_matrix if isinstance(i, FDataGrid) - else i for i in inputs + else self._get_op_matrix(i) for i in inputs ] outputs = kwargs.pop('out', None) @@ -1292,6 +1292,28 @@ def __array_ufunc__( ##################################################################### # Pandas ExtensionArray methods ##################################################################### + + def _take_allow_fill( + self: T, + indices: NDArrayInt, + fill_value: T, + ) -> T: + result = self.copy() + result.data_matrix = np.full( + (len(indices),) + self.data_matrix.shape[1:], + np.nan, + ) + + positive_mask = indices >= 0 + result.data_matrix[positive_mask] = self.data_matrix[ + indices[positive_mask] + ] + + if fill_value is not self.dtype.na_value: + result.data_matrix[~positive_mask] = fill_value.data_matrix[0] + + return result + @property def dtype(self) -> FDataGridDType: """The dtype for this extension array, FDataGridDType"""

Check conmutativity of operations between FDataGrid and numpy arrays Operations between FDataGrid objects and numpy arrays should not depend on the order. Currently if the array is the first term, the ufunc is dispatched to the FDataGrid `__array_ufunc__` and there are errors.

2021-11-11T00:19:50

0.0

[]

GAA-UAM/scikit-fda

GAA-UAM__scikit-fda-387

bd907fdb9cbf0401d2b46ddf4106d0ae73b188dd

diff --git a/skfda/preprocessing/dim_reduction/feature_extraction/__init__.py b/skfda/preprocessing/dim_reduction/feature_extraction/__init__.py index 16355e236..132cc8f96 100644 --- a/skfda/preprocessing/dim_reduction/feature_extraction/__init__.py +++ b/skfda/preprocessing/dim_reduction/feature_extraction/__init__.py @@ -1,3 +1,4 @@ """Feature extraction.""" from ._ddg_transformer import DDGTransformer +from ._fda_feature_union import FDAFeatureUnion from ._fpca import FPCA diff --git a/skfda/preprocessing/dim_reduction/feature_extraction/_fda_feature_union.py b/skfda/preprocessing/dim_reduction/feature_extraction/_fda_feature_union.py new file mode 100644 index 000000000..e772be898 --- /dev/null +++ b/skfda/preprocessing/dim_reduction/feature_extraction/_fda_feature_union.py @@ -0,0 +1,124 @@ +"""Feature extraction union for dimensionality reduction.""" +from __future__ import annotations + +from typing import Union + +from numpy import ndarray +from pandas import DataFrame +from sklearn.pipeline import FeatureUnion + +from ....representation import FData + + +class FDAFeatureUnion(FeatureUnion): # type: ignore + """Concatenates results of multiple functional transformer objects. + + This estimator applies a list of transformer objects in parallel to the + input data, then concatenates the results (They can be either FDataGrid + and FDataBasis objects or multivariate data itself).This is useful to + combine several feature extraction mechanisms into a single transformer. + Parameters of the transformers may be set using its name and the parameter + name separated by a '__'. A transformer may be replaced entirely by + setting the parameter with its name to another transformer, + or removed by setting to 'drop'. + + Parameters: + transformer_list: list of tuple + List of tuple containing `(str, transformer)`. The first element + of the tuple is name affected to the transformer while the + second element is a scikit-learn transformer instance. + The transformer instance can also be `"drop"` for it to be + ignored. + n_jobs: int + Number of jobs to run in parallel. + ``None`` means 1 unless in a :obj:`joblib.parallel_backend` + context. + ``-1`` means using all processors. + The default value is None + transformer_weights: dict + Multiplicative weights for features per transformer. + Keys are transformer names, values the weights. + Raises ValueError if key not present in ``transformer_list``. + verbose: bool + If True, the time elapsed while fitting each transformer will be + printed as it is completed. By default the value is False + array_output: bool + indicates if the transformed data is requested to be a NumPy array + output. By default the value is False. + + Examples: + Firstly we will import the Berkeley Growth Study data set + >>> from skfda.datasets import fetch_growth + >>> X,y = fetch_growth(return_X_y=True) + + Then we need to import the transformers we want to use. In our case we + will use Generalized depth-versus-depth transformer. + Evaluation Transformer returns the original curve, and as it is helpful, + we will concatenate it to the already metioned transformer. + >>> from skfda.preprocessing.dim_reduction.feature_extraction import ( + ... FDAFeatureUnion, + ... ) + >>> from skfda.preprocessing.dim_reduction.feature_extraction import ( + ... DDGTransformer, + ... ) + >>> from skfda.exploratory.depth import ModifiedBandDepth + >>> from skfda.representation import EvaluationTransformer + >>> import numpy as np + + Finally we apply fit and transform. + >>> union = FDAFeatureUnion( + ... [ + ... ( + ... 'ddgtransformer', + ... DDGTransformer(depth_method=[ModifiedBandDepth()]), + ... ), + ... ("eval", EvaluationTransformer()), + ... ], + ... array_output=True, + ... ) + >>> np.around(union.fit_transform(X,y), decimals=2) + array([[ 2.100e-01, 9.000e-02, 8.130e+01, ..., 1.938e+02, 1.943e+02, + 1.951e+02], + [ 4.600e-01, 3.800e-01, 7.620e+01, ..., 1.761e+02, 1.774e+02, + 1.787e+02], + [ 2.000e-01, 3.300e-01, 7.680e+01, ..., 1.709e+02, 1.712e+02, + 1.715e+02], + ..., + [ 3.900e-01, 5.100e-01, 6.860e+01, ..., 1.660e+02, 1.663e+02, + 1.668e+02], + [ 2.600e-01, 2.700e-01, 7.990e+01, ..., 1.683e+02, 1.684e+02, + 1.686e+02], + [ 3.300e-01, 3.200e-01, 7.610e+01, ..., 1.686e+02, 1.689e+02, + 1.692e+02]]) + """ + + def __init__( + self, + transformer_list: list, # type: ignore + *, + n_jobs: int = 1, + transformer_weights: dict = None, # type: ignore + verbose: bool = False, + array_output: bool = False, + ) -> None: + self.array_output = array_output + super().__init__( + transformer_list, + n_jobs=n_jobs, + transformer_weights=transformer_weights, + verbose=verbose, + ) + + def _hstack(self, Xs: ndarray) -> Union[DataFrame, ndarray]: + + if self.array_output: + for i in Xs: + if isinstance(i, FData): + raise TypeError( + "There are transformed instances of FDataGrid or " + "FDataBasis that can't be concatenated on a NumPy " + "array.", + ) + return super()._hstack(Xs) + + return DataFrame({'Transformed data': Xs})

Join (possibly functional) features The existing scikit-learn class [`FeatureUnion`](https://scikit-learn.org/stable/modules/generated/sklearn.pipeline.FeatureUnion.html) is already able to apply transformations and join features that are multivariate in a new NumPy array. However, features that are functions cannot be in a NumPy array. A possible solution is to subclass `FeatureUnion` and join everything in a Pandas `DataFrame` (which accepts `FDataGrid` and `FDataBasis` as columns). We should also have a flag parameter that optionally allow the user to request NumPy array output (and an exception if that is not possible). Thanks to https://github.com/scikit-learn/scikit-learn/pull/17868 it seems that we will only need to override the `_hstack` method in the subclass to join the results.

2021-11-03T18:48:13

0.0

[]

GAA-UAM/scikit-fda

GAA-UAM__scikit-fda-364

7c6f33db95f94e67def9424ee745dbd1326e256a

diff --git a/skfda/misc/covariances.py b/skfda/misc/covariances.py index 7acda93fa..5358a57fa 100644 --- a/skfda/misc/covariances.py +++ b/skfda/misc/covariances.py @@ -9,7 +9,6 @@ from matplotlib.figure import Figure from scipy.special import gamma, kv -from ..exploratory.visualization._utils import _create_figure, _figure_to_svg from ..representation._typing import ArrayLike, NDArrayFloat @@ -74,6 +73,8 @@ def __call__(self, x: ArrayLike, y: ArrayLike) -> NDArrayFloat: def heatmap(self, limits: Tuple[float, float] = (-1, 1)) -> Figure: """Return a heatmap plot of the covariance function.""" + from ..exploratory.visualization._utils import _create_figure + x = np.linspace(*limits, 1000) cov_matrix = self(x, x) @@ -130,6 +131,8 @@ def _repr_latex_(self) -> str: return fr"$\displaystyle {self._latex_content()}$" def _repr_html_(self) -> str: + from ..exploratory.visualization._utils import _figure_to_svg + fig = self.heatmap() heatmap = _figure_to_svg(fig) plt.close(fig) diff --git a/skfda/misc/metrics/_fisher_rao.py b/skfda/misc/metrics/_fisher_rao.py index f63bb80f8..469ed0479 100644 --- a/skfda/misc/metrics/_fisher_rao.py +++ b/skfda/misc/metrics/_fisher_rao.py @@ -7,7 +7,6 @@ from typing_extensions import Final from ..._utils import normalize_scale, normalize_warping -from ...preprocessing.registration import FisherRaoElasticRegistration from ...representation import FData, FDataGrid from ...representation._typing import NDArrayFloat from ..operators import SRSF @@ -191,6 +190,8 @@ def fisher_rao_amplitude_distance( .. footbibliography:: """ + from ...preprocessing.registration import FisherRaoElasticRegistration + fdata1, fdata2 = _cast_to_grid( fdata1, fdata2, @@ -285,6 +286,8 @@ def fisher_rao_phase_distance( .. footbibliography:: """ + from ...preprocessing.registration import FisherRaoElasticRegistration + fdata1, fdata2 = _cast_to_grid( fdata1, fdata2, diff --git a/skfda/preprocessing/dim_reduction/feature_extraction/_fpca.py b/skfda/preprocessing/dim_reduction/feature_extraction/_fpca.py index f551fc0e0..6e2f7cabe 100644 --- a/skfda/preprocessing/dim_reduction/feature_extraction/_fpca.py +++ b/skfda/preprocessing/dim_reduction/feature_extraction/_fpca.py @@ -5,11 +5,13 @@ from typing import Callable, Optional, TypeVar, Union import numpy as np +import scipy.integrate +from scipy.linalg import solve_triangular from sklearn.base import BaseEstimator, TransformerMixin from sklearn.decomposition import PCA -from scipy.linalg import solve_triangular - +from ....misc import inner_product_matrix +from ....misc.metrics import l2_norm from ....misc.regularization import ( TikhonovRegularization, compute_penalty_matrix, @@ -151,6 +153,8 @@ def _fit_basis( else X.basis.n_basis ) n_samples = X.n_samples + # necessary in inverse_transform + self.n_samples_ = X.n_samples # check that the number of components is smaller than the sample size if self.n_components > X.n_samples: @@ -326,6 +330,9 @@ def _fit_grid( # get the number of samples and the number of points of descretization n_samples, n_points_discretization = fd_data.shape + # necessary for inverse_transform + self.n_samples_ = n_samples + # if centering is True then subtract the mean function to each function # in FDataBasis X = self._center_if_necessary(X) @@ -336,9 +343,8 @@ def _fit_grid( # in trapezoidal rule, suppose \deltax_k = x_k - x_{k-1}, the # weight vector is as follows: [\deltax_1/2, \deltax_1/2 + # \deltax_2/2, \deltax_2/2 + \deltax_3/2, ... , \deltax_n/2] - differences = np.diff(X.grid_points[0]) - differences = np.concatenate(((0,), differences, (0,))) - self.weights = (differences[:-1] + differences[1:]) / 2 + identity = np.eye(len(X.grid_points[0])) + self.weights = scipy.integrate.simps(identity, X.grid_points[0]) elif callable(self.weights): self.weights = self.weights(X.grid_points[0]) # if its a FDataGrid then we need to reduce the dimension to 1-D @@ -382,6 +388,7 @@ def _fit_grid( ), sample_names=(None,) * self.n_components, ) + self.explained_variance_ratio_ = pca.explained_variance_ratio_ self.explained_variance_ = pca.explained_variance_ @@ -408,6 +415,7 @@ def _transform_grid( return ( X.data_matrix.reshape(X.data_matrix.shape[:-1]) + * self.weights @ np.transpose( self.components_.data_matrix.reshape( self.components_.data_matrix.shape[:-1], @@ -480,3 +488,61 @@ def fit_transform( """ return self.fit(X, y).transform(X, y) + + def inverse_transform( + self, + pc_scores: np.ndarray, + ) -> FData: + """ + Compute the recovery from the fitted principal components scores. + + In other words, + it maps ``pc_scores``, from the fitted functional PCs' space, + back to the input functional space. + ``pc_scores`` might be an array returned by ``transform`` method. + + Args: + pc_scores: ndarray (n_samples, n_components). + + Returns: + A FData object. + + """ + # check the instance is fitted. + + # input format check: + if isinstance(pc_scores, np.ndarray): + if pc_scores.ndim == 1: + pc_scores = pc_scores[np.newaxis, :] + + if pc_scores.shape[1] != self.n_components: + raise AttributeError( + "pc_scores must be a numpy array " + "with n_samples rows and n_components columns.", + ) + else: + raise AttributeError("pc_scores is not a numpy array.") + + # inverse_transform is slightly different whether + # .fit was applied to FDataGrid or FDataBasis object + # Does not work (boundary problem in x_hat and bias reconstruction) + if isinstance(self.components_, FDataGrid): + + additional_args = { + "data_matrix": np.einsum( + 'nc,c...->n...', + pc_scores, + self.components_.data_matrix, + ), + } + + elif isinstance(self.components_, FDataBasis): + + additional_args = { + "coefficients": pc_scores @ self.components_.coefficients, + } + + return self.mean_.copy( + **additional_args, + sample_names=(None,) * len(pc_scores), + ) + self.mean_

Implement inverse_transform for FPCA FPCA should have an `inverse_transform` method to convert the data in the space generated by the eigenfunctions back into the original space.

2021-09-10T16:25:47

0.0

[]

monarch-initiative/embiggen

monarch-initiative__embiggen-301

b743f7ff87ef2f6aed5db9ba5c6c43ee2e32bb8d

diff --git a/.gitignore b/.gitignore index b8107f2c..e01ebf96 100644 --- a/.gitignore +++ b/.gitignore @@ -37,4 +37,8 @@ embeddings *.webm model.json -model.pkl.gz \ No newline at end of file +model.pkl.gz +okapi_tfidf_weighted_textual_embedding +precomputed_word_embedding +embedding +*.ipynb \ No newline at end of file diff --git a/dist/embiggen-0.11.41.tar.gz b/dist/embiggen-0.11.41.tar.gz new file mode 100644 index 00000000..df4f3b83 Binary files /dev/null and b/dist/embiggen-0.11.41.tar.gz differ diff --git a/dist/embiggen-0.11.42.tar.gz b/dist/embiggen-0.11.42.tar.gz new file mode 100644 index 00000000..4fdbe9a0 Binary files /dev/null and b/dist/embiggen-0.11.42.tar.gz differ diff --git a/dist/embiggen-0.11.43.tar.gz b/dist/embiggen-0.11.43.tar.gz new file mode 100644 index 00000000..cecae322 Binary files /dev/null and b/dist/embiggen-0.11.43.tar.gz differ diff --git a/dist/embiggen-0.11.44.tar.gz b/dist/embiggen-0.11.44.tar.gz new file mode 100644 index 00000000..2df92419 Binary files /dev/null and b/dist/embiggen-0.11.44.tar.gz differ diff --git a/embiggen/__version__.py b/embiggen/__version__.py index e8f9ef5b..436504c8 100644 --- a/embiggen/__version__.py +++ b/embiggen/__version__.py @@ -1,2 +1,2 @@ """Current version of package Embiggen.""" -__version__ = "0.11.41" \ No newline at end of file +__version__ = "0.11.45" \ No newline at end of file diff --git a/embiggen/edge_prediction/edge_prediction_model.py b/embiggen/edge_prediction/edge_prediction_model.py index 4426c63c..26be6ecd 100644 --- a/embiggen/edge_prediction/edge_prediction_model.py +++ b/embiggen/edge_prediction/edge_prediction_model.py @@ -362,13 +362,16 @@ def predict( raise NotImplementedError( "Currently edge features are not supported in edge prediction models." ) - - predictions = super().predict( - graph, - support=support, - node_features=node_features, - node_type_features=node_type_features - ).flatten() + + if graph.has_edges(): + predictions = super().predict( + graph, + support=support, + node_features=node_features, + node_type_features=node_type_features + ).flatten() + else: + predictions = np.array([]) if return_predictions_dataframe: predictions = pd.DataFrame( diff --git a/embiggen/edge_prediction/edge_prediction_sklearn/sklearn_edge_prediction_adapter.py b/embiggen/edge_prediction/edge_prediction_sklearn/sklearn_edge_prediction_adapter.py index 0bf07d27..eba5a201 100644 --- a/embiggen/edge_prediction/edge_prediction_sklearn/sklearn_edge_prediction_adapter.py +++ b/embiggen/edge_prediction/edge_prediction_sklearn/sklearn_edge_prediction_adapter.py @@ -275,6 +275,9 @@ def _predict( edge_features: Optional[List[np.ndarray]] = None The edge features to use. """ + if not graph.has_edges(): + return np.array([]) + sequence = EdgePredictionSequence( graph=graph, graph_used_in_training=graph, @@ -283,6 +286,7 @@ def _predict( use_edge_metrics=False, batch_size=self._prediction_batch_size ) + return np.concatenate([ self._model_instance.predict(self._trasform_graph_into_edge_embedding( graph=edges[0], @@ -326,6 +330,9 @@ def _predict_proba( edge_features: Optional[List[np.ndarray]] = None The edge features to use. """ + if not graph.has_edges(): + return np.array([]) + sequence = EdgePredictionSequence( graph=graph, graph_used_in_training=graph, @@ -334,6 +341,7 @@ def _predict_proba( use_edge_metrics=False, batch_size=self._prediction_batch_size ) + prediction_probabilities = np.concatenate([ self._model_instance.predict_proba(self._trasform_graph_into_edge_embedding( graph=edges[0], diff --git a/embiggen/embedders/__init__.py b/embiggen/embedders/__init__.py index 5e9b2513..42354909 100644 --- a/embiggen/embedders/__init__.py +++ b/embiggen/embedders/__init__.py @@ -3,7 +3,10 @@ from embiggen.embedders.tensorflow_embedders import * from embiggen.embedders.pykeen_embedders import * from embiggen.embedders.non_existent_embedders import * +from embiggen.embedders.pytorch_geometric import * from embiggen.embedders.karateclub_embedders import * +from embiggen.embedders.pecanpy_embedders import * +from embiggen.embedders.fastnode2vec_embedders import * from embiggen.embedders.graph_embedding_pipeline import embed_graph # Export all non-internals. diff --git a/embiggen/embedders/ensmallen_embedders/deepwalk_cbow.py b/embiggen/embedders/ensmallen_embedders/deepwalk_cbow.py index 99d4c46b..0cd1fa86 100644 --- a/embiggen/embedders/ensmallen_embedders/deepwalk_cbow.py +++ b/embiggen/embedders/ensmallen_embedders/deepwalk_cbow.py @@ -27,7 +27,8 @@ def __init__( random_state: int = 42, dtype: str = "f32", ring_bell: bool = False, - enable_cache: bool = False + enable_cache: bool = False, + verbose: bool = True ): """Create new abstract DeepWalk method. @@ -84,6 +85,13 @@ def __init__( enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. + verbose: bool = True + Whether to display the loading bar. + This will only display the loading bar when + running the script in a bash-like environment. + It will not work in Jupyter Notebooks, there + it will appear in the notebook kernel in some + systems but not necessarily. """ super().__init__( embedding_size=embedding_size, @@ -105,7 +113,8 @@ def __init__( dtype=dtype, random_state=random_state, ring_bell=ring_bell, - enable_cache=enable_cache + enable_cache=enable_cache, + verbose=verbose, ) def parameters(self) -> Dict[str, Any]: diff --git a/embiggen/embedders/ensmallen_embedders/deepwalk_glove.py b/embiggen/embedders/ensmallen_embedders/deepwalk_glove.py index f55e6e4f..5b9861a1 100644 --- a/embiggen/embedders/ensmallen_embedders/deepwalk_glove.py +++ b/embiggen/embedders/ensmallen_embedders/deepwalk_glove.py @@ -21,7 +21,8 @@ def __init__( dtype: str = "f32", random_state: int = 42, ring_bell: bool = False, - enable_cache: bool = False + enable_cache: bool = False, + verbose: bool = True ): """Create new abstract DeepWalk method. @@ -69,6 +70,13 @@ def __init__( enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. + verbose: bool = True + Whether to display the loading bar. + This will only display the loading bar when + running the script in a bash-like environment. + It will not work in Jupyter Notebooks, there + it will appear in the notebook kernel in some + systems but not necessarily. """ super().__init__( embedding_size=embedding_size, @@ -86,7 +94,8 @@ def __init__( enable_cache=enable_cache, dtype=dtype, ring_bell=ring_bell, - random_state=random_state + random_state=random_state, + verbose=verbose, ) def parameters(self) -> Dict[str, Any]: diff --git a/embiggen/embedders/ensmallen_embedders/deepwalk_skipgram.py b/embiggen/embedders/ensmallen_embedders/deepwalk_skipgram.py index b00342ec..18b81d2e 100644 --- a/embiggen/embedders/ensmallen_embedders/deepwalk_skipgram.py +++ b/embiggen/embedders/ensmallen_embedders/deepwalk_skipgram.py @@ -27,7 +27,8 @@ def __init__( random_state: int = 42, dtype: str = "f32", ring_bell: bool = False, - enable_cache: bool = False + enable_cache: bool = False, + verbose: bool = True ): """Create new abstract DeepWalk method. @@ -84,6 +85,13 @@ def __init__( enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. + verbose: bool = True + Whether to display the loading bar. + This will only display the loading bar when + running the script in a bash-like environment. + It will not work in Jupyter Notebooks, there + it will appear in the notebook kernel in some + systems but not necessarily. """ super().__init__( embedding_size=embedding_size, @@ -105,7 +113,8 @@ def __init__( dtype=dtype, random_state=random_state, ring_bell=ring_bell, - enable_cache=enable_cache + enable_cache=enable_cache, + verbose=verbose, ) def parameters(self) -> Dict[str, Any]: diff --git a/embiggen/embedders/ensmallen_embedders/node2vec.py b/embiggen/embedders/ensmallen_embedders/node2vec.py index 05094f59..67054945 100644 --- a/embiggen/embedders/ensmallen_embedders/node2vec.py +++ b/embiggen/embedders/ensmallen_embedders/node2vec.py @@ -51,7 +51,11 @@ def __init__( """ model_name = must_be_in_set(self.model_name(), self.MODELS.keys(), "model name") self._model_kwargs = model_kwargs - self._model = Node2VecEnsmallen.MODELS[model_name](**model_kwargs) + self._model = Node2VecEnsmallen.MODELS[model_name]( + embedding_size=embedding_size, + random_state=random_state, + **model_kwargs + ) super().__init__( embedding_size=embedding_size, diff --git a/embiggen/embedders/ensmallen_embedders/node2vec_cbow.py b/embiggen/embedders/ensmallen_embedders/node2vec_cbow.py index 04961df9..78f956f4 100644 --- a/embiggen/embedders/ensmallen_embedders/node2vec_cbow.py +++ b/embiggen/embedders/ensmallen_embedders/node2vec_cbow.py @@ -29,7 +29,8 @@ def __init__( random_state: int = 42, dtype: str = "f32", ring_bell: bool = False, - enable_cache: bool = False + enable_cache: bool = False, + verbose: bool = True ): """Create new abstract Node2Vec method. @@ -100,6 +101,13 @@ def __init__( enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. + verbose: bool = True + Whether to display the loading bar. + This will only display the loading bar when + running the script in a bash-like environment. + It will not work in Jupyter Notebooks, there + it will appear in the notebook kernel in some + systems but not necessarily. """ super().__init__( embedding_size=embedding_size, @@ -123,7 +131,8 @@ def __init__( dtype=dtype, random_state=random_state, ring_bell=ring_bell, - enable_cache=enable_cache + enable_cache=enable_cache, + verbose=verbose, ) def parameters(self) -> Dict[str, Any]: diff --git a/embiggen/embedders/ensmallen_embedders/node2vec_glove.py b/embiggen/embedders/ensmallen_embedders/node2vec_glove.py index a80772b7..1fdb5919 100644 --- a/embiggen/embedders/ensmallen_embedders/node2vec_glove.py +++ b/embiggen/embedders/ensmallen_embedders/node2vec_glove.py @@ -23,7 +23,8 @@ def __init__( dtype: str = "f32", random_state: int = 42, ring_bell: bool = False, - enable_cache: bool = False + enable_cache: bool = False, + verbose: bool = True ): """Create new Node2Vec GloVe model. @@ -85,6 +86,13 @@ def __init__( enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. + verbose: bool = True + Whether to display the loading bar. + This will only display the loading bar when + running the script in a bash-like environment. + It will not work in Jupyter Notebooks, there + it will appear in the notebook kernel in some + systems but not necessarily. """ super().__init__( embedding_size=embedding_size, @@ -104,7 +112,8 @@ def __init__( dtype=dtype, enable_cache=enable_cache, ring_bell=ring_bell, - random_state=random_state + random_state=random_state, + verbose=verbose, ) def parameters(self) -> Dict[str, Any]: diff --git a/embiggen/embedders/ensmallen_embedders/node2vec_skipgram.py b/embiggen/embedders/ensmallen_embedders/node2vec_skipgram.py index 2dd871e6..4350dbcc 100644 --- a/embiggen/embedders/ensmallen_embedders/node2vec_skipgram.py +++ b/embiggen/embedders/ensmallen_embedders/node2vec_skipgram.py @@ -29,7 +29,8 @@ def __init__( random_state: int = 42, dtype: str = "f32", ring_bell: bool = False, - enable_cache: bool = False + enable_cache: bool = False, + verbose: bool = True ): """Create new abstract Node2Vec method. @@ -102,6 +103,13 @@ def __init__( enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. + verbose: bool = True + Whether to display the loading bar. + This will only display the loading bar when + running the script in a bash-like environment. + It will not work in Jupyter Notebooks, there + it will appear in the notebook kernel in some + systems but not necessarily. """ super().__init__( embedding_size=embedding_size, @@ -125,7 +133,8 @@ def __init__( dtype=dtype, random_state=random_state, ring_bell=ring_bell, - enable_cache=enable_cache + enable_cache=enable_cache, + verbose=verbose, ) def parameters(self) -> Dict[str, Any]: diff --git a/embiggen/embedders/fastnode2vec_embedders/__init__.py b/embiggen/embedders/fastnode2vec_embedders/__init__.py new file mode 100644 index 00000000..77f92926 --- /dev/null +++ b/embiggen/embedders/fastnode2vec_embedders/__init__.py @@ -0,0 +1,8 @@ +"""Module with node embedding models based on FastNode2Vec.""" +from embiggen.utils.abstract_models import build_init, AbstractEmbeddingModel + +build_init( + module_library_names=["numba", "fastnode2vec"], + formatted_library_name="FastNode2Vec", + expected_parent_class=AbstractEmbeddingModel +) \ No newline at end of file diff --git a/embiggen/embedders/fastnode2vec_embedders/node2vec.py b/embiggen/embedders/fastnode2vec_embedders/node2vec.py new file mode 100644 index 00000000..a5d13d5b --- /dev/null +++ b/embiggen/embedders/fastnode2vec_embedders/node2vec.py @@ -0,0 +1,219 @@ +"""Node2Vec wrapper for FastNode2Vec numba-based node embedding library.""" +from typing import Dict, Union, Any + +import numpy as np +import pandas as pd +from ensmallen import Graph + +from embiggen.utils.abstract_models import AbstractEmbeddingModel, EmbeddingResult +from fastnode2vec import Graph as FNGraph +from fastnode2vec import Node2Vec +from multiprocessing import cpu_count +from time import time + + +class Node2VecFastNode2Vec(AbstractEmbeddingModel): + """Node2Vec wrapper for FastNode2Vec numba-based node embedding library.""" + + def __init__( + self, + embedding_size: int = 100, + epochs: int = 30, + walk_length: int = 128, + iterations: int = 10, + window_size: int = 5, + learning_rate: float = 0.01, + return_weight: float = 0.25, + explore_weight: float = 4.0, + number_of_workers: Union[int, str] = "auto", + verbose: bool = False, + random_state: int = 42, + ring_bell: bool = False, + enable_cache: bool = False + ): + """Create new wrapper for Node2Vec model from FastNode2Vec library. + + Parameters + ------------------------- + embedding_size: int = 100 + The dimension of the embedding to compute. + epochs: int = 100 + The number of epochs to use to train the model for. + learning_rate: float = 0.01 + Learning rate of the model. + verbose: bool = False + Whether to show the loading bar. + random_state: int = 42 + Random seed to use while training the model + ring_bell: bool = False, + Whether to play a sound when embedding completes. + enable_cache: bool = False + Whether to enable the cache, that is to + store the computed embedding. + """ + self._walk_length = walk_length + self._iterations = iterations + self._window_size = window_size + self._return_weight = return_weight + self._explore_weight = explore_weight + self._epochs = epochs + self._verbose = verbose + self._learning_rate = learning_rate + self._time_required_by_last_embedding = None + if number_of_workers == "auto": + number_of_workers = cpu_count() + self._number_of_workers = number_of_workers + + super().__init__( + embedding_size=embedding_size, + enable_cache=enable_cache, + ring_bell=ring_bell, + random_state=random_state + ) + + @classmethod + def smoke_test_parameters(cls) -> Dict[str, Any]: + """Returns parameters for smoke test.""" + return dict( + **super().smoke_test_parameters(), + window_size=1, + walk_length=2, + iterations=1 + ) + + def parameters(self) -> Dict[str, Any]: + return dict( + **super().parameters(), + **dict( + epochs=self._epochs, + walk_length=self._walk_length, + window_size=self._window_size, + iterations=self._iterations, + return_weight=self._return_weight, + explore_weight=self._explore_weight, + ) + ) + + @classmethod + def model_name(cls) -> str: + """Returns name of the model""" + return "Node2Vec" + + @classmethod + def library_name(cls) -> str: + return "FastNode2Vec" + + @classmethod + def task_name(cls) -> str: + return "Node Embedding" + + def get_time_required_by_last_embedding(self) -> float: + """Returns the time required by last embedding.""" + if self._time_required_by_last_embedding is None: + raise ValueError("You have not yet run an embedding.") + return self._time_required_by_last_embedding + + def _fit_transform( + self, + graph: Graph, + return_dataframe: bool = True, + ) -> Union[np.ndarray, pd.DataFrame, Dict[str, np.ndarray], Dict[str, pd.DataFrame]]: + """Return node embedding""" + + if graph.has_edge_weights(): + edges_iterator = ( + ( + *graph.get_node_names_from_edge_id(edge_id), + graph.get_edge_weight_from_edge_id(edge_id) + ) + for edge_id in range(graph.get_number_of_directed_edges()) + ) + else: + edges_iterator = ( + graph.get_node_names_from_edge_id(edge_id) + for edge_id in range(graph.get_number_of_directed_edges()) + ) + + fn_graph: FNGraph = FNGraph( + edges_iterator, + directed=True, + weighted=graph.has_edge_weights(), + n_edges=graph.get_number_of_directed_edges() + ) + + start = time() + + model: Node2Vec = Node2Vec( + graph=fn_graph, + dim=self._embedding_size, + walk_length=self._walk_length, + context=self._window_size, + p=1.0/self._return_weight, + q=1.0/self._explore_weight, + workers=self._number_of_workers, + batch_walks=self._iterations, + seed=self._random_state, + ) + + model.train( + epochs=self._epochs * self._iterations, + progress_bar=self._verbose + ) + + self._time_required_by_last_embedding = time() - start + + # This library does not provide node embedding + # for disconnected nodes, so we need to patch it up + # if the graph contains disconnected nodes. + if graph.has_singleton_nodes(): + node_embedding = np.zeros( + shape=(graph.get_number_of_nodes(), self._embedding_size), + ) + for node_id in range(graph.get_number_of_nodes()): + node_name = graph.get_node_name_from_node_id(node_id) + if node_name in model.wv: + node_embedding[node_id] = model.wv[node_name] + else: + node_embedding = model.wv[graph.get_node_names()] + + if return_dataframe: + node_embedding = pd.DataFrame( + node_embedding, + index=graph.get_node_names() + ) + + return EmbeddingResult( + embedding_method_name=self.model_name(), + node_embeddings=node_embedding + ) + + @classmethod + def requires_nodes_sorted_by_decreasing_node_degree(cls) -> bool: + return False + + @classmethod + def is_topological(cls) -> bool: + return True + + @classmethod + def can_use_edge_weights(cls) -> bool: + return True + + @classmethod + def requires_edge_weights(cls) -> bool: + return False + + @classmethod + def can_use_node_types(cls) -> bool: + """Returns whether the model can use node types.""" + return False + + @classmethod + def can_use_edge_types(cls) -> bool: + """Returns whether the model can use edge types.""" + return False + + @classmethod + def is_stocastic(cls) -> bool: + """Returns whether the model is stocastic and has therefore a random state.""" + return True diff --git a/embiggen/embedders/karateclub_embedders/abstract_karateclub_embedder.py b/embiggen/embedders/karateclub_embedders/abstract_karateclub_embedder.py index 2da4030b..1e503d43 100644 --- a/embiggen/embedders/karateclub_embedders/abstract_karateclub_embedder.py +++ b/embiggen/embedders/karateclub_embedders/abstract_karateclub_embedder.py @@ -7,6 +7,7 @@ from embiggen.utils.abstract_models import AbstractEmbeddingModel, EmbeddingResult, abstract_class from embiggen.utils.networkx_utils import convert_ensmallen_graph_to_networkx_graph + @abstract_class class AbstractKarateClubEmbedder(AbstractEmbeddingModel): @@ -28,7 +29,7 @@ def task_name(cls) -> str: def _build_model(self) -> Type[Estimator]: """Returnd the built estimator.""" raise NotImplementedError( - f"In the child class {self.__class__.__name__} of {super().__name__.__name__} " + f"In the child class {self.__class__.__name__} of {super().__class__.__name__} " f"implementing the model {self.model_name()} we could not find the method " "called `_build_model`. Please do implement it." ) @@ -61,7 +62,8 @@ def _fit_transform( "It is not clear what to do with this object." ) - model.fit(convert_ensmallen_graph_to_networkx_graph(graph)) + graph_nx = convert_ensmallen_graph_to_networkx_graph(graph) + model.fit(graph_nx) node_embeddings: np.ndarray = model.get_embedding() diff --git a/embiggen/embedders/karateclub_embedders/deep_walk.py b/embiggen/embedders/karateclub_embedders/deep_walk.py index 17d14ba3..e6023ba1 100644 --- a/embiggen/embedders/karateclub_embedders/deep_walk.py +++ b/embiggen/embedders/karateclub_embedders/deep_walk.py @@ -10,10 +10,10 @@ class DeepWalkSkipGramKarateClub(AbstractKarateClubEmbedder): def __init__( self, embedding_size: int = 100, - walk_number: int = 10, - walk_length: int = 80, + iterations: int = 10, + walk_length: int = 128, window_size: int = 5, - epochs: int = 10, + epochs: int = 30, learning_rate: float = 0.05, min_count: int = 1, random_state: int = 42, @@ -26,13 +26,13 @@ def __init__( ---------------------- embedding_size: int = 100 Size of the embedding to use. - walk_number: int = 10 + iterations: int = 10 Number of random walks. Default is 10. - walk_length: int = 80 + walk_length: int = 128 Length of random walks. Default is 80. window_size: int = 5 Matrix power order. Default is 5. - epochs: int = 10 + epochs: int = 30 Number of epochs. Default is 1. learning_rate: float = 0.05 HogWild! learning rate. Default is 0.05. @@ -47,7 +47,7 @@ def __init__( Whether to enable the cache, that is to store the computed embedding. """ - self._walk_number=walk_number + self._iterations=iterations self._walk_length=walk_length self._workers=cpu_count() self._window_size=window_size @@ -65,7 +65,7 @@ def parameters(self) -> Dict[str, Any]: """Returns the parameters used in the model.""" return dict( **super().parameters(), - walk_number=self._walk_number, + iterations=self._iterations, walk_length=self._walk_length, window_size=self._window_size, epochs=self._epochs, @@ -78,7 +78,7 @@ def smoke_test_parameters(cls) -> Dict[str, Any]: """Returns parameters for smoke test.""" return dict( **AbstractKarateClubEmbedder.smoke_test_parameters(), - walk_number=1, + iterations=1, walk_length=8, window_size=1, epochs=1, @@ -87,7 +87,7 @@ def smoke_test_parameters(cls) -> Dict[str, Any]: def _build_model(self) -> DeepWalk: """Return new instance of the DeepWalk model.""" return DeepWalk( - walk_number=self._walk_number, + walk_number=self._iterations, walk_length=self._walk_length, dimensions=self._embedding_size, workers=self._workers, diff --git a/embiggen/embedders/karateclub_embedders/role2vec.py b/embiggen/embedders/karateclub_embedders/role2vec.py index 1a925f4a..901f81c1 100644 --- a/embiggen/embedders/karateclub_embedders/role2vec.py +++ b/embiggen/embedders/karateclub_embedders/role2vec.py @@ -10,8 +10,8 @@ class Role2VecKarateClub(AbstractKarateClubEmbedder): def __init__( self, embedding_size: int = 100, - walk_number: int = 10, - walk_length: int = 80, + iterations: int = 10, + walk_length: int = 128, window_size: int = 5, epochs: int = 10, learning_rate: float = 0.05, @@ -29,9 +29,9 @@ def __init__( ---------------------- embedding_size: int = 100 Size of the embedding to use. - walk_number: int = 10 + iterations: int = 10 Number of random walks. Default is 10. - walk_length: int = 80 + walk_length: int = 128 Length of random walks. Default is 80. window_size: int = 5 Matrix power order. Default is 5. @@ -56,7 +56,7 @@ def __init__( Whether to enable the cache, that is to store the computed embedding. """ - self._walk_number = walk_number + self._iterations = iterations self._walk_length = walk_length self._workers = cpu_count() self._window_size = window_size @@ -77,7 +77,7 @@ def parameters(self) -> Dict[str, Any]: """Returns the parameters used in the model.""" return dict( **super().parameters(), - walk_number=self._walk_number, + iterations=self._iterations, walk_length=self._walk_length, window_size=self._window_size, epochs=self._epochs, @@ -93,7 +93,7 @@ def smoke_test_parameters(cls) -> Dict[str, Any]: """Returns parameters for smoke test.""" return dict( **AbstractKarateClubEmbedder.smoke_test_parameters(), - walk_number=1, + iterations=1, weisfeiler_lehman_hashing_iterations=1, walk_length=8, window_size=2, @@ -103,7 +103,7 @@ def smoke_test_parameters(cls) -> Dict[str, Any]: def _build_model(self) -> Role2Vec: """Return new instance of the Role2Vec model.""" return Role2Vec( - walk_number=self._walk_number, + walk_number=self._iterations, walk_length=self._walk_length, dimensions=self._embedding_size, workers=self._workers, diff --git a/embiggen/embedders/karateclub_embedders/skipgram.py b/embiggen/embedders/karateclub_embedders/skipgram.py index 443aa78b..728d2381 100644 --- a/embiggen/embedders/karateclub_embedders/skipgram.py +++ b/embiggen/embedders/karateclub_embedders/skipgram.py @@ -10,8 +10,8 @@ class SkipGramKarateClub(AbstractKarateClubEmbedder): def __init__( self, embedding_size: int = 100, - walk_number: int = 10, - walk_length: int = 80, + iterations: int = 10, + walk_length: int = 128, window_size: int = 5, p: float = 1.0, q: float = 1.0, @@ -36,9 +36,9 @@ def __init__( ---------------------- embedding_size: int = 100 Size of the embedding to use. - walk_number: int = 10 + iterations: int = 10 Number of random walks. Default is 10. - walk_length: int = 80 + walk_length: int = 128 Length of random walks. Default is 80. window_size: int = 5 Matrix power order. Default is 5. @@ -61,7 +61,7 @@ def __init__( Whether to enable the cache, that is to store the computed embedding. """ - self._walk_number=walk_number + self._iterations=iterations self._walk_length=walk_length self._workers=cpu_count() self._window_size=window_size @@ -81,7 +81,7 @@ def parameters(self) -> Dict[str, Any]: """Returns the parameters used in the model.""" return dict( **super().parameters(), - walk_number=self._walk_number, + iterations=self._iterations, walk_length=self._walk_length, window_size=self._window_size, p=self._p, @@ -96,7 +96,7 @@ def smoke_test_parameters(cls) -> Dict[str, Any]: """Returns parameters for smoke test.""" return dict( **AbstractKarateClubEmbedder.smoke_test_parameters(), - walk_number=1, + iterations=1, walk_length=8, window_size=2, epochs=1, @@ -105,7 +105,7 @@ def smoke_test_parameters(cls) -> Dict[str, Any]: def _build_model(self) -> Node2Vec: """Return new instance of the Node2Vec CBOW model.""" return Node2Vec( - walk_number=self._walk_number, + walk_number=self._iterations, walk_length=self._walk_length, dimensions=self._embedding_size, workers=self._workers, diff --git a/embiggen/embedders/karateclub_embedders/walklets.py b/embiggen/embedders/karateclub_embedders/walklets.py index d56ba3ec..73848ed7 100644 --- a/embiggen/embedders/karateclub_embedders/walklets.py +++ b/embiggen/embedders/karateclub_embedders/walklets.py @@ -10,8 +10,8 @@ class WalkletsSkipGramKarateClub(AbstractKarateClubEmbedder): def __init__( self, embedding_size: int = 100, - walk_number: int = 10, - walk_length: int = 80, + iterations: int = 10, + walk_length: int = 128, window_size: int = 5, epochs: int = 10, learning_rate: float = 0.05, @@ -26,9 +26,9 @@ def __init__( ---------------------- embedding_size: int = 100 Size of the embedding to use. - walk_number: int = 10 + iterations: int = 10 Number of random walks. Default is 10. - walk_length: int = 80 + walk_length: int = 128 Length of random walks. Default is 80. window_size: int = 5 Matrix power order. Default is 5. @@ -47,7 +47,7 @@ def __init__( Whether to enable the cache, that is to store the computed embedding. """ - self._walk_number = walk_number + self._iterations = iterations self._walk_length = walk_length self._workers = cpu_count() self._window_size = window_size @@ -67,7 +67,7 @@ def parameters(self) -> Dict[str, Any]: parameters["embedding_size"] = parameters["embedding_size"] * self._window_size return dict( **parameters, - walk_number=self._walk_number, + iterations=self._iterations, walk_length=self._walk_length, window_size=self._window_size, epochs=self._epochs, @@ -80,7 +80,7 @@ def smoke_test_parameters(cls) -> Dict[str, Any]: """Returns parameters for smoke test.""" return dict( **AbstractKarateClubEmbedder.smoke_test_parameters(), - walk_number=1, + iterations=1, walk_length=8, window_size=1, epochs=1, @@ -89,7 +89,7 @@ def smoke_test_parameters(cls) -> Dict[str, Any]: def _build_model(self) -> Walklets: """Return new instance of the Walklets model.""" return Walklets( - walk_number=self._walk_number, + walk_number=self._iterations, walk_length=self._walk_length, dimensions=self._embedding_size, workers=self._workers, diff --git a/embiggen/embedders/pecanpy_embedders/__init__.py b/embiggen/embedders/pecanpy_embedders/__init__.py new file mode 100644 index 00000000..8370092a --- /dev/null +++ b/embiggen/embedders/pecanpy_embedders/__init__.py @@ -0,0 +1,8 @@ +"""Module with node embedding models based on PecanPy.""" +from embiggen.utils.abstract_models import build_init, AbstractEmbeddingModel + +build_init( + module_library_names=["numba", "pecanpy"], + formatted_library_name="PecanPy", + expected_parent_class=AbstractEmbeddingModel +) \ No newline at end of file diff --git a/embiggen/embedders/pecanpy_embedders/node2vec.py b/embiggen/embedders/pecanpy_embedders/node2vec.py new file mode 100644 index 00000000..e0c787ac --- /dev/null +++ b/embiggen/embedders/pecanpy_embedders/node2vec.py @@ -0,0 +1,216 @@ +"""Node2Vec wrapper for PecanPy numba-based node embedding library.""" +from typing import Dict, Union, Any + +import numpy as np +import pandas as pd +from ensmallen import Graph + +from embiggen.utils.abstract_models import AbstractEmbeddingModel, EmbeddingResult +from pecanpy.node2vec import SparseOTF +from multiprocessing import cpu_count +from time import time + +class Node2VecPecanPy(AbstractEmbeddingModel): + """Node2Vec wrapper for PecanPy numba-based node embedding library.""" + + def __init__( + self, + embedding_size: int = 100, + epochs: int = 30, + walk_length: int = 128, + iterations: int = 10, + window_size: int = 5, + batch_size: int = 128, + learning_rate: float = 0.01, + return_weight: float = 0.25, + explore_weight: float = 4.0, + number_of_workers: Union[int, str] = "auto", + verbose: bool = False, + random_state: int = 42, + ring_bell: bool = False, + enable_cache: bool = False + ): + """Create new wrapper for Node2Vec model from PecanPy library. + + Parameters + ------------------------- + embedding_size: int = 100 + The dimension of the embedding to compute. + epochs: int = 100 + The number of epochs to use to train the model for. + batch_size: int = 2**10 + Size of the training batch. + learning_rate: float = 0.01 + Learning rate of the model. + verbose: bool = False + Whether to show the loading bar. + random_state: int = 42 + Random seed to use while training the model + ring_bell: bool = False, + Whether to play a sound when embedding completes. + enable_cache: bool = False + Whether to enable the cache, that is to + store the computed embedding. + """ + self._walk_length = walk_length + self._iterations = iterations + self._window_size = window_size + self._return_weight = return_weight + self._explore_weight = explore_weight + self._epochs = epochs + self._verbose = verbose + self._batch_size = batch_size + self._learning_rate = learning_rate + self._time_required_by_last_embedding = None + if number_of_workers == "auto": + number_of_workers = cpu_count() + self._number_of_workers = number_of_workers + + super().__init__( + embedding_size=embedding_size, + enable_cache=enable_cache, + ring_bell=ring_bell, + random_state=random_state + ) + + @classmethod + def smoke_test_parameters(cls) -> Dict[str, Any]: + """Returns parameters for smoke test.""" + return dict( + **super().smoke_test_parameters(), + window_size=1, + walk_length=2, + iterations=1 + ) + + def parameters(self) -> Dict[str, Any]: + return dict( + **super().parameters(), + **dict( + epochs=self._epochs, + batch_size=self._batch_size, + walk_length=self._walk_length, + window_size=self._window_size, + iterations=self._iterations, + return_weight=self._return_weight, + explore_weight=self._explore_weight, + ) + ) + + @classmethod + def model_name(cls) -> str: + """Returns name of the model""" + return "Node2Vec" + + @classmethod + def library_name(cls) -> str: + return "PecanPy" + + @classmethod + def task_name(cls) -> str: + return "Node Embedding" + + def get_time_required_by_last_embedding(self) -> float: + """Returns the time required by last embedding.""" + if self._time_required_by_last_embedding is None: + raise ValueError("You have not yet run an embedding.") + return self._time_required_by_last_embedding + + def _fit_transform( + self, + graph: Graph, + return_dataframe: bool = True, + ) -> Union[np.ndarray, pd.DataFrame, Dict[str, np.ndarray], Dict[str, pd.DataFrame]]: + """Return node embedding""" + + model: SparseOTF = SparseOTF( + p=1.0/self._return_weight, + q=1.0/self._explore_weight, + workers=self._number_of_workers, + verbose=self._verbose, + ) + + # Instead of using `SparseOTF` methods, + # we manipolate directly the object attributes + # in order to avoid a memory peak and making + # the conversion of the Ensmallen graph object + # into the `SparseOTF` more seamless. + + # The `indptr` attribute contains the indices + # of the rows in the CSR representation, which + # are the comulative node degrees. + # This is shifted by a value to the right, and the + # first value is set to zero. + model.indptr = np.zeros(graph.get_number_of_nodes() + 1, dtype=np.int64) + model.indptr[1:] = graph.get_cumulative_node_degrees().astype(np.int64) + + # In the `indices` attribute we need to store the destinations. + model.indices = graph.get_directed_destination_node_ids() + + # We also need to set the `IDlst` attribute, which are + # the node IDs. + model.IDlst = graph.get_node_ids() + + # In model data we need to store the edge weights + # if are present. If the graph is weighted, we use + # the graph edge weights. Otherwise we set all weights + # to one, as the library PecanPy does. + if graph.has_edge_weights(): + model.data = graph.get_directed_edge_weights().astype(np.float64) + else: + model.data = np.ones_like(model.indices, dtype=np.float64) + + start = time() + + node_embedding = model.embed( + dim=self._embedding_size, + num_walks=self._iterations, + walk_length=self._walk_length, + window_size=self._window_size, + epochs=self._epochs, + verbose=self._verbose + ) + + self._time_required_by_last_embedding = time() - start + + if return_dataframe: + node_embedding = pd.DataFrame( + node_embedding, + index=graph.get_node_names() + ) + + return EmbeddingResult( + embedding_method_name=self.model_name(), + node_embeddings=node_embedding + ) + + @classmethod + def requires_nodes_sorted_by_decreasing_node_degree(cls) -> bool: + return False + + @classmethod + def is_topological(cls) -> bool: + return True + + @classmethod + def can_use_edge_weights(cls) -> bool: + return True + + @classmethod + def requires_edge_weights(cls) -> bool: + return False + + @classmethod + def can_use_node_types(cls) -> bool: + """Returns whether the model can use node types.""" + return False + + @classmethod + def can_use_edge_types(cls) -> bool: + """Returns whether the model can use edge types.""" + return False + + @classmethod + def is_stocastic(cls) -> bool: + """Returns whether the model is stocastic and has therefore a random state.""" + return True \ No newline at end of file diff --git a/embiggen/embedders/pykeen_embedders/__init__.py b/embiggen/embedders/pykeen_embedders/__init__.py index affc9d1d..bc15ab46 100644 --- a/embiggen/embedders/pykeen_embedders/__init__.py +++ b/embiggen/embedders/pykeen_embedders/__init__.py @@ -1,4 +1,4 @@ -"""Module with graph embedding models based on PyKEEN.""" +"""Module with node embedding models based on PyKEEN.""" from embiggen.utils.abstract_models import build_init, AbstractEmbeddingModel build_init( diff --git a/embiggen/embedders/pykeen_embedders/pykeen_embedder.py b/embiggen/embedders/pykeen_embedders/pykeen_embedder.py index d0f21d5d..ecf1f198 100644 --- a/embiggen/embedders/pykeen_embedders/pykeen_embedder.py +++ b/embiggen/embedders/pykeen_embedders/pykeen_embedder.py @@ -134,7 +134,7 @@ def _build_model(self, triples_factory: CoreTriplesFactory) -> Type[Model]: The PyKEEN triples factory to use to create the model. """ raise NotImplementedError( - f"In the child class {self.__class__.__name__} of {super().__name__.__name__} " + f"In the child class {self.__class__.__name__} of {super().__class__.__name__} " f"implementing the model {self.model_name()} we could not find the method " "called `_build_model`. Please do implement it." ) @@ -167,7 +167,7 @@ def _extract_embeddings( Whether to return a dataframe of a numpy array. """ raise NotImplementedError( - f"In the child class {self.__class__.__name__} of {super().__name__.__name__} " + f"In the child class {self.__class__.__name__} of {super().__class__.__name__} " f"implementing the model {self.model_name()} we could not find the method " "called `_extract_embeddings`. Please do implement it." ) diff --git a/embiggen/embedders/pytorch_geometric/__init__.py b/embiggen/embedders/pytorch_geometric/__init__.py new file mode 100644 index 00000000..8983b48d --- /dev/null +++ b/embiggen/embedders/pytorch_geometric/__init__.py @@ -0,0 +1,8 @@ +"""Module with node embedding models based on PyTorch Geometric.""" +from embiggen.utils.abstract_models import build_init, AbstractEmbeddingModel + +build_init( + module_library_names=["torch", "torch_geometric", "torch_cluster", "torch_sparse", "torch_scatter"], + formatted_library_name="PyTorch Geometric", + expected_parent_class=AbstractEmbeddingModel +) \ No newline at end of file diff --git a/embiggen/embedders/pytorch_geometric/node2vec.py b/embiggen/embedders/pytorch_geometric/node2vec.py new file mode 100644 index 00000000..7d9d463a --- /dev/null +++ b/embiggen/embedders/pytorch_geometric/node2vec.py @@ -0,0 +1,142 @@ +"""Abstract class for graph embedding models.""" +from typing import Dict, Any + +from embiggen.embedders.pytorch_geometric.pytorch_geometric_embedder import PyTorchGeometricEmbedder +from torch_geometric.nn import Node2Vec +from torch import Tensor +from torch.nn import Module +from torch.optim import Optimizer +from torch import DeviceObjType + + +class Node2VecPyTorchGeometric(PyTorchGeometricEmbedder): + """Abstract class for sequence embedding models.""" + + def __init__( + self, + + embedding_size: int = 100, + epochs: int = 30, + number_of_negative_samples: int = 10, + walk_length: int = 128, + iterations: int = 10, + window_size: int = 5, + batch_size: int = 128, + learning_rate: float = 0.01, + return_weight: float = 0.25, + explore_weight: float = 4.0, + random_state: int = 42, + optimizer: str = "adam", + verbose: bool = False, + ring_bell: bool = False, + enable_cache: bool = False + ): + """Create new PyTorch Geometric Node2Vec model. + + Parameters + ------------------------------- + + """ + self._number_of_negative_samples = number_of_negative_samples + self._embedding_size = embedding_size + self._walk_length = walk_length + self._window_size = window_size + self._iterations = iterations + self._return_weight = return_weight + self._explore_weight = explore_weight + + self._loader = None + + super().__init__( + random_state=random_state, + embedding_size=embedding_size, + epochs=epochs, + batch_size=batch_size, + learning_rate=learning_rate, + optimizer=optimizer, + verbose=verbose, + ring_bell=ring_bell, + enable_cache=enable_cache + ) + + def _build_model( + self, + edge_node_ids: Tensor, + number_of_nodes: int + ) -> Node2Vec: + model = Node2Vec( + edge_index=edge_node_ids, + embedding_dim=self._embedding_size, + walk_length=self._walk_length, + context_size=2*self._window_size, + walks_per_node=self._iterations, + p=1.0/self._return_weight, + q=1.0/self._explore_weight, + num_negative_samples=self._number_of_negative_samples, + num_nodes=number_of_nodes, + ) + + self._loader = model.loader( + batch_size=self._batch_size, + shuffle=True, + num_workers=self._number_of_workers + ) + + return model + + @classmethod + def model_name(cls) -> str: + """Returns name of the model""" + return "Node2Vec" + + @classmethod + def can_use_edge_weights(cls) -> bool: + """Returns whether the model can optionally use edge weights.""" + return False + + def parameters(self) -> Dict[str, Any]: + """Returns parameters of the model.""" + return dict( + **super().parameters(), + walk_length=self._walk_length, + window_size=self._window_size, + iterations=self._iterations, + return_weight=self._return_weight, + explore_weight=self._explore_weight, + number_of_negative_samples=self._number_of_negative_samples, + ) + + @classmethod + def smoke_test_parameters(cls) -> Dict[str, Any]: + """Returns parameters for smoke test.""" + return dict( + **super().smoke_test_parameters(), + number_of_negative_samples=1, + window_size=1, + walk_length=2, + iterations=1 + ) + + def _train_model_step( + self, + model: Module, + optimizer: Optimizer, + device: DeviceObjType + ): + """Train model for a single step. + + Parameters + ------------------ + model: Module + The model to be trained for this step. + optimizer: Optimizer + The optimizer to be trained for this step. + device: DeviceObjType + The device to be used. + """ + model.train() + for pos_rw, neg_rw in self._loader: + optimizer.zero_grad() + loss = model.loss(pos_rw.to(device), neg_rw.to(device)) + loss.backward() + optimizer.step() diff --git a/embiggen/embedders/pytorch_geometric/pytorch_geometric_embedder.py b/embiggen/embedders/pytorch_geometric/pytorch_geometric_embedder.py new file mode 100644 index 00000000..473a9f86 --- /dev/null +++ b/embiggen/embedders/pytorch_geometric/pytorch_geometric_embedder.py @@ -0,0 +1,233 @@ +"""Abstract Torch/PyTorch Geometric Model wrapper for embedding models.""" +from typing import Dict, Union, Any, Type + +import numpy as np +import pandas as pd +from ensmallen import Graph + +from embiggen.utils.pytorch_utils import validate_torch_device +from embiggen.utils.abstract_models import AbstractEmbeddingModel, abstract_class, EmbeddingResult +import torch + +from tqdm.auto import trange + +from multiprocessing import cpu_count +from environments_utils import is_windows, is_macos +from torch import Tensor +from torch.nn import Module +from torch.optim import Optimizer +from torch import DeviceObjType + + +@abstract_class +class PyTorchGeometricEmbedder(AbstractEmbeddingModel): + """Abstract PyTorch Geometric Model wrapper for embedding models.""" + + def __init__( + self, + embedding_size: int = 100, + epochs: int = 100, + batch_size: int = 2**10, + learning_rate: float = 0.01, + number_of_workers: Union[int, str] = "auto", + device: str = "auto", + verbose: bool = False, + random_state: int = 42, + optimizer: str = "adam", + ring_bell: bool = False, + enable_cache: bool = False + ): + """Create new PyTorch Geometric Abstract Embedder model. + + Parameters + ------------------------- + embedding_size: int = 100 + The dimension of the embedding to compute. + epochs: int = 100 + The number of epochs to use to train the model for. + batch_size: int = 2**10 + Size of the training batch. + learning_rate: float = 0.01 + Learning rate of the model. + device: str = "auto" + The devide to use to train the model. + Can either be cpu or cuda. + verbose: bool = False + Whether to show the loading bar. + random_state: int = 42 + Random seed to use while training the model + ring_bell: bool = False, + Whether to play a sound when embedding completes. + enable_cache: bool = False + Whether to enable the cache, that is to + store the computed embedding. + """ + self._epochs = epochs + self._verbose = verbose + self._batch_size = batch_size + self._learning_rate = learning_rate + if number_of_workers == "auto": + number_of_workers = 0 if is_windows() or is_macos() else cpu_count() + self._optimizer = optimizer + self._number_of_workers = number_of_workers + self._device = validate_torch_device(device) + + super().__init__( + embedding_size=embedding_size, + enable_cache=enable_cache, + ring_bell=ring_bell, + random_state=random_state + ) + + @classmethod + def smoke_test_parameters(cls) -> Dict[str, Any]: + """Returns parameters for smoke test.""" + return dict( + embedding_size=10, + epochs=1 + ) + + def parameters(self) -> Dict[str, Any]: + return dict( + **super().parameters(), + **dict( + epochs=self._epochs, + batch_size=self._batch_size, + ) + ) + + @classmethod + def library_name(cls) -> str: + return "PyTorch Geometric" + + @classmethod + def task_name(cls) -> str: + return "Node Embedding" + + def _build_model( + self, + edge_node_ids: Tensor, + number_of_nodes: int + ) -> Type[Module]: + """Build new model for embedding. + + Parameters + ------------------ + edge_node_ids: Tensor + Tuples with the source and destination node ids + number_of_nodes: int + Number of nodes in the graph. + """ + raise NotImplementedError( + f"In the child class {self.__class__.__name__} of {super().__class__.__name__} " + f"implementing the model {self.model_name()} we could not find the method " + "called `_build_model`. Please do implement it." + ) + + def _train_model_step( + self, + model: Module, + optimizer: Optimizer, + device: DeviceObjType + ): + """Train model for a single step. + + Parameters + ------------------ + model: Module + The model to be trained for this step. + optimizer: Optimizer + The optimizer to be trained for this step. + device: DeviceObjType + The device to be used. + """ + raise NotImplementedError( + f"In the child class {self.__class__.__name__} of {super().__class__.__name__} " + f"implementing the model {self.model_name()} we could not find the method " + "called `_train_model_step`. Please do implement it." + ) + + def _fit_transform( + self, + graph: Graph, + return_dataframe: bool = True, + ) -> Union[np.ndarray, pd.DataFrame, Dict[str, np.ndarray], Dict[str, pd.DataFrame]]: + """Return node embedding""" + + torch_device = torch.device(self._device) + + edge_node_ids = torch.LongTensor(np.int64(graph.get_directed_edge_node_ids().T)) + + model = self._build_model( + edge_node_ids=edge_node_ids, + number_of_nodes=graph.get_number_of_nodes() + ) + + # Move the model to gpu if we need to + model.to(torch_device) + + if not issubclass(model.__class__, Module): + raise NotImplementedError( + "The model created with the `_build_model` in the child " + f"class {self.__class__.__name__} for the model {self.model_name()} " + f"in the library {self.library_name()} did not return a " + f"PyTorch Geometric model but an object of type {type(model)}." + ) + + optimizer = torch.optim.Adam( + list(model.parameters()), + lr=self._learning_rate + ) + + for _ in trange( + self._epochs, + dynamic_ncols=True, + desc="Epochs", + disable=not self._verbose, + leave=False + ): + self._train_model_step( + model=model, + optimizer=optimizer, + device=torch_device + ) + + # Extract and return the embedding + node_embedding = model(torch.arange( + graph.get_number_of_nodes(), + device=torch_device + )).cpu().detach().numpy() + + if return_dataframe: + node_embedding = pd.DataFrame( + node_embedding, + index=graph.get_node_names() + ) + + return EmbeddingResult( + embedding_method_name=self.model_name(), + node_embeddings=node_embedding + ) + + @classmethod + def requires_nodes_sorted_by_decreasing_node_degree(cls) -> bool: + return False + + @classmethod + def is_topological(cls) -> bool: + return True + + @classmethod + def can_use_node_types(cls) -> bool: + """Returns whether the model can use node types.""" + return False + + @classmethod + def can_use_edge_types(cls) -> bool: + """Returns whether the model can use edge types.""" + return False + + @classmethod + def is_stocastic(cls) -> bool: + """Returns whether the model is stocastic and has therefore a random state.""" + return True \ No newline at end of file diff --git a/embiggen/embedders/tensorflow_embedders/__init__.py b/embiggen/embedders/tensorflow_embedders/__init__.py index 4d109788..79dcb159 100644 --- a/embiggen/embedders/tensorflow_embedders/__init__.py +++ b/embiggen/embedders/tensorflow_embedders/__init__.py @@ -1,4 +1,4 @@ -"""Module with graph embedding models based on TensorFlow.""" +"""Module with node embedding models based on TensorFlow.""" from embiggen.utils.abstract_models import build_init, AbstractEmbeddingModel build_init( diff --git a/embiggen/embedders/tensorflow_embedders/edge_prediction_based_tensorflow_embedders.py b/embiggen/embedders/tensorflow_embedders/edge_prediction_based_tensorflow_embedders.py index b55422fb..cb6048a5 100644 --- a/embiggen/embedders/tensorflow_embedders/edge_prediction_based_tensorflow_embedders.py +++ b/embiggen/embedders/tensorflow_embedders/edge_prediction_based_tensorflow_embedders.py @@ -30,7 +30,7 @@ def __init__( use_mirrored_strategy: bool = False, activation: str = "sigmoid", loss: str = "binary_crossentropy", - optimizer: str = "nadam", + optimizer: str = "adam", verbose: bool = False, ring_bell: bool = False, enable_cache: bool = False, diff --git a/embiggen/embedders/tensorflow_embedders/siamese.py b/embiggen/embedders/tensorflow_embedders/siamese.py index 16636969..99246008 100644 --- a/embiggen/embedders/tensorflow_embedders/siamese.py +++ b/embiggen/embedders/tensorflow_embedders/siamese.py @@ -32,7 +32,7 @@ def __init__( learning_rate_plateau_patience: int = 5, norm: str = "L2", use_mirrored_strategy: bool = False, - optimizer: str = "nadam", + optimizer: str = "adam", verbose: bool = False, ring_bell: bool = False, enable_cache: bool = False, diff --git a/embiggen/embedders/tensorflow_embedders/tensorflow_embedder.py b/embiggen/embedders/tensorflow_embedders/tensorflow_embedder.py index b33c5fae..9d5b0103 100644 --- a/embiggen/embedders/tensorflow_embedders/tensorflow_embedder.py +++ b/embiggen/embedders/tensorflow_embedders/tensorflow_embedder.py @@ -27,7 +27,7 @@ def __init__( learning_rate_plateau_patience: int = 1, epochs: int = 10, batch_size: int = 2**10, - optimizer: str = "nadam", + optimizer: str = "adam", verbose: bool = False, use_mirrored_strategy: bool = False, ring_bell: bool = False, @@ -129,7 +129,7 @@ def _build_model(self, graph: Graph) -> Model: The graph to build the model for. """ raise NotImplementedError( - f"In the child class {self.__class__.__name__} of {super().__name__.__name__} " + f"In the child class {self.__class__.__name__} of {super().__class__.__name__} " f"implementing the model {self.model_name()} we could not find the method " "called `_build_model`. Please do implement it." ) @@ -145,7 +145,7 @@ def _build_input(self, graph: Graph) -> Tuple[Any]: Whether to show loading bars while building input. """ raise NotImplementedError( - f"In the child class {self.__class__.__name__} of {super().__name__.__name__} " + f"In the child class {self.__class__.__name__} of {super().__class__.__name__} " f"implementing the model {self.model_name()} we could not find the method " "called `_build_input`. Please do implement it." ) diff --git a/embiggen/layers/tensorflow/graph_convolution_layer.py b/embiggen/layers/tensorflow/graph_convolution_layer.py index 06a51eb0..44b4a014 100644 --- a/embiggen/layers/tensorflow/graph_convolution_layer.py +++ b/embiggen/layers/tensorflow/graph_convolution_layer.py @@ -109,6 +109,17 @@ def l2_norm(x): return x super().build(input_shape) + def get_config(self): + config = super(GraphConvolution, self).get_config() + config.update({ + "units": self._units, + "activation": self._activation, + "combiner": self._combiner, + "dropout_rate": self._dropout_rate, + "apply_norm": self._apply_norm, + }) + return config + def call( self, inputs: Tuple[Union[tf.Tensor, List[tf.Tensor], tf.SparseTensor]], diff --git a/embiggen/node_label_prediction/node_label_prediction_model.py b/embiggen/node_label_prediction/node_label_prediction_model.py index 0b02c4d3..eb981a51 100644 --- a/embiggen/node_label_prediction/node_label_prediction_model.py +++ b/embiggen/node_label_prediction/node_label_prediction_model.py @@ -157,7 +157,21 @@ def _evaluate( ) if self.is_binary_prediction_task(): - predictions = prediction_probabilities + if prediction_probabilities.shape[1] == 1: + predictions = prediction_probabilities + elif prediction_probabilities.shape[1] == 2: + predictions = prediction_probabilities[:, 1] + prediction_probabilities = prediction_probabilities[:, 1] + else: + raise NotImplementedError( + f"The model {self.model_name()} as implemented in " + f"the library {self.library_name()} for the task " + f"{self.task_name()} has produced a binary prediction " + f"result with shape {prediction_probabilities.shape}, " + "which is unclear how to handle for evaluation. " + "Please open an issue and pull request to clarify what " + "you expect to happen here." + ) elif self.is_multilabel_prediction_task(): predictions = prediction_probabilities > 0.5 else: diff --git a/embiggen/node_label_prediction/node_label_prediction_tensorflow/gcn.py b/embiggen/node_label_prediction/node_label_prediction_tensorflow/gcn.py index 326b7e7a..b8acca36 100644 --- a/embiggen/node_label_prediction/node_label_prediction_tensorflow/gcn.py +++ b/embiggen/node_label_prediction/node_label_prediction_tensorflow/gcn.py @@ -177,7 +177,7 @@ def _build_model( for units in self._number_of_units_per_head_layer: hidden = Dense( units=units, - activation="ReLU" + activation="relu" )(hidden) output = Dense( diff --git a/embiggen/utils/abstract_gcn.py b/embiggen/utils/abstract_gcn.py index 5cac7202..10bf7c88 100644 --- a/embiggen/utils/abstract_gcn.py +++ b/embiggen/utils/abstract_gcn.py @@ -1,7 +1,7 @@ """Kipf GCN model for node-label prediction.""" from typing import List, Union, Optional, Dict, Any, Type, Tuple -from matplotlib.pyplot import axis +import compress_pickle import numpy as np import pandas as pd from tensorflow.keras.callbacks import EarlyStopping, ReduceLROnPlateau # pylint: disable=import-error,no-name-in-module @@ -14,7 +14,7 @@ from keras_mixed_sequence import Sequence from embiggen.utils.abstract_models import AbstractClassifierModel, abstract_class from embiggen.utils.number_to_ordinal import number_to_ordinal -from embiggen.layers.tensorflow import GraphConvolution, FlatEmbedding +from embiggen.layers.tensorflow import GraphConvolution, FlatEmbedding, EmbeddingLookup, L2Norm from tensorflow.keras.layers import Input, Concatenate import warnings from embiggen.utils.normalize_model_structural_parameters import normalize_model_list_parameter @@ -628,6 +628,9 @@ def _predict_proba( edge_features: Optional[List[np.ndarray]] = None, ) -> pd.DataFrame: """Run predictions on the provided graph.""" + if not graph.has_edges(): + return np.array([]) + if support is None: support = graph @@ -742,4 +745,29 @@ def is_using_edge_weights(self) -> bool: """Returns whether the model is parametrized to use edge weights.""" return not self._use_simmetric_normalized_laplacian - \ No newline at end of file + @classmethod + def load(cls, path: str) -> "Self": + """Load a saved version of the model from the provided path. + + Parameters + ------------------- + path: str + Path from where to load the model. + """ + with tf.keras.utils.custom_object_scope({ + "GraphConvolution": GraphConvolution, + "EmbeddingLookup": EmbeddingLookup, + "FlatEmbedding": FlatEmbedding, + "L2Norm": L2Norm + }): + return compress_pickle.load(path) + + def dump(self, path: str): + """Dump the current model at the provided path. + + Parameters + ------------------- + path: str + Path from where to dump the model. + """ + compress_pickle.dump(self, path) \ No newline at end of file diff --git a/embiggen/utils/abstract_models/abstract_classifier_model.py b/embiggen/utils/abstract_models/abstract_classifier_model.py index e3548115..7a972701 100644 --- a/embiggen/utils/abstract_models/abstract_classifier_model.py +++ b/embiggen/utils/abstract_models/abstract_classifier_model.py @@ -1448,15 +1448,6 @@ def _evaluate_on_single_holdout( **holdouts_kwargs ) - # We enable in the train and test graphs the same - # speedups enabled in the provided graph. - train.enable( - vector_sources=graph.has_sources_tradeoff_enabled(), - vector_destinations=graph.has_destinations_tradeoff_enabled(), - vector_cumulative_node_degrees=graph.has_cumulative_node_degrees_tradeoff_enabled(), - vector_reciprocal_sqrt_degrees=graph.has_reciprocal_sqrt_degrees_tradeoff_enabled() - ) - # We compute the remaining features starting_to_compute_node_features = time.time() holdout_node_features = cls.normalize_node_features( @@ -1556,22 +1547,6 @@ def _evaluate_on_single_holdout( train_of_interest = train test_of_interest = test - # We enable in the train and test graphs of interest the same - # speedups enabled in the provided graph. - train_of_interest.enable( - vector_sources=graph.has_sources_tradeoff_enabled(), - vector_destinations=graph.has_destinations_tradeoff_enabled(), - vector_cumulative_node_degrees=graph.has_cumulative_node_degrees_tradeoff_enabled(), - vector_reciprocal_sqrt_degrees=graph.has_reciprocal_sqrt_degrees_tradeoff_enabled() - ) - - test_of_interest.enable( - vector_sources=graph.has_sources_tradeoff_enabled(), - vector_destinations=graph.has_destinations_tradeoff_enabled(), - vector_cumulative_node_degrees=graph.has_cumulative_node_degrees_tradeoff_enabled(), - vector_reciprocal_sqrt_degrees=graph.has_reciprocal_sqrt_degrees_tradeoff_enabled() - ) - additional_validation_kwargs = cls._prepare_evaluation( graph=graph, support=train, diff --git a/embiggen/utils/abstract_models/auto_init.py b/embiggen/utils/abstract_models/auto_init.py index 4ddeef21..3bde401a 100644 --- a/embiggen/utils/abstract_models/auto_init.py +++ b/embiggen/utils/abstract_models/auto_init.py @@ -7,6 +7,7 @@ from glob import glob from embiggen.utils.abstract_models.model_stub import get_model_or_stub import os +import sys import inspect import traceback @@ -195,8 +196,17 @@ def build_init( expected_parent_class: Type[AbstractModel] The class to check for. """ + # The Python default is 1000. In some situation, + # users may want to modify this and set the global + # tracebacklimit to a value too low for us to execute + # the meta-programming task of initializing GRAPE's + # models. For these reasons, we set it explicitly + # to the default Python value, i.e. `1000` without + # touching the sys variables that the user may have + # customized. + stack_trace = traceback.extract_stack(limit=1000) path_pattern = os.path.join(os.path.dirname(os.path.abspath( - traceback.extract_stack()[-2].filename + stack_trace[-2].filename )), "*.py") # We retrieve the context of the caller. @@ -219,6 +229,7 @@ def build_init( klass.name == expected_parent_class.__name__ ): continue + # If this class has the expected parent. if expected_parent_class.__name__ in get_class_parent_names( path, diff --git a/embiggen/utils/abstract_models/embedding_result.py b/embiggen/utils/abstract_models/embedding_result.py index dd610656..4fa9fe9e 100644 --- a/embiggen/utils/abstract_models/embedding_result.py +++ b/embiggen/utils/abstract_models/embedding_result.py @@ -86,6 +86,14 @@ def __init__( "contains NaN values." ) + if np.isinf(numpy_embedding).any(): + number = np.sum(np.isinf(numpy_embedding)) + raise ValueError( + f"One of the provided {embedding_list_name} " + f"computed with the {embedding_method_name} method " + f"contains {number} infinite values." + ) + if np.isclose(numpy_embedding, 0.0).all(): raise ValueError( f"One of the provided {embedding_list_name} " diff --git a/embiggen/visualizations/graph_visualizer.py b/embiggen/visualizations/graph_visualizer.py index 3c567c8c..3caff43d 100644 --- a/embiggen/visualizations/graph_visualizer.py +++ b/embiggen/visualizations/graph_visualizer.py @@ -2206,8 +2206,8 @@ def _plot_positive_and_negative_edges_metric( types = np.concatenate([ np.zeros( - self._negative_edge_decomposition.shape[0], dtype=np.bool), - np.ones(self._positive_edge_decomposition.shape[0], dtype=np.bool), + self._negative_edge_decomposition.shape[0], dtype=bool), + np.ones(self._positive_edge_decomposition.shape[0], dtype=bool), ]) test_accuracies = [] @@ -4204,6 +4204,7 @@ def plot_node_degree_distribution( figure: Optional[Figure] = None, axes: Optional[Figure] = None, apply_tight_layout: bool = True, + show_title: bool = True, return_caption: bool = True, ) -> Tuple[Figure, Axes]: """Plot the given graph node degree distribution. @@ -4219,6 +4220,8 @@ def plot_node_degree_distribution( apply_tight_layout: bool = True, Whether to apply the tight layout on the matplotlib Figure object. + show_title: bool = True + Wether to show the figure title. return_caption: bool = True, Whether to return a caption for the plot. """ @@ -4234,13 +4237,14 @@ def plot_node_degree_distribution( bins=number_of_buckets, log=True ) - axes.set_ylabel("Counts (log scale)") - axes.set_xlabel("Degrees") + axes.set_ylabel("Count (log scale)") + axes.set_xlabel("Degree") if self._show_graph_name: title = "Degree distribution of graph {}".format(self._graph_name) else: title = "Degree distribution" - axes.set_title(title) + if show_title: + axes.set_title(title) if apply_tight_layout: figure.tight_layout() diff --git a/model.pkl b/model.pkl new file mode 100644 index 00000000..9dca220f Binary files /dev/null and b/model.pkl differ diff --git a/setup.py b/setup.py index 151159bc..40b868f8 100644 --- a/setup.py +++ b/setup.py @@ -72,7 +72,7 @@ def find_version(*file_paths): "ddd_subplots>=1.0.23", "sanitize_ml_labels>=1.0.50", "keras_mixed_sequence>=1.0.28", - "ensmallen>=0.8.31", + "ensmallen>=0.8.38", "environments_utils>=1.0.6", "compress_pickle>=2.1.0", "validate_version_code",

Exposing more APIs from Ensmallen in Embiggen for edge prediction

2023-04-11T12:58:47

0.0

[]

monarch-initiative/embiggen

monarch-initiative__embiggen-300

c868d6264c8728cbbf647d75d57b03b570ee8056

diff --git a/dist/embiggen-0.11.40.tar.gz b/dist/embiggen-0.11.40.tar.gz new file mode 100644 index 00000000..7d1c04fd Binary files /dev/null and b/dist/embiggen-0.11.40.tar.gz differ diff --git a/embiggen/__version__.py b/embiggen/__version__.py index 819b838c..e8f9ef5b 100644 --- a/embiggen/__version__.py +++ b/embiggen/__version__.py @@ -1,2 +1,2 @@ """Current version of package Embiggen.""" -__version__ = "0.11.39" \ No newline at end of file +__version__ = "0.11.41" \ No newline at end of file diff --git a/embiggen/edge_prediction/edge_prediction_evaluation.py b/embiggen/edge_prediction/edge_prediction_evaluation.py index 3b7280ff..9c989521 100644 --- a/embiggen/edge_prediction/edge_prediction_evaluation.py +++ b/embiggen/edge_prediction/edge_prediction_evaluation.py @@ -23,6 +23,12 @@ def edge_prediction_evaluation( repositories: Optional[Union[str, List[str]]] = None, versions: Optional[Union[str, List[str]]] = None, validation_sample_only_edges_with_heterogeneous_node_types: bool = False, + source_node_types_names: Optional[List[str]] = None, + destination_node_types_names: Optional[List[str]] = None, + source_edge_types_names: Optional[List[str]] = None, + destination_edge_types_names: Optional[List[str]] = None, + source_nodes_prefixes: Optional[List[str]] = None, + destination_nodes_prefixes: Optional[List[str]] = None, validation_unbalance_rates: Tuple[float] = (1.0, ), use_scale_free_distribution: bool = True, enable_cache: bool = False, @@ -98,6 +104,24 @@ def edge_prediction_evaluation( validation_sample_only_edges_with_heterogeneous_node_types: bool = False Whether to sample negative edges exclusively between nodes with different node types. This can be useful when executing a bipartite edge prediction task. + source_node_types_names: Optional[List[str]] + Node type names of the nodes to be samples as sources. + If a node has any of the provided node types, it can be sampled as a source node. + destination_node_types_names: Optional[List[str]] + Node type names of the nodes to be samples as destinations. + If a node has any of the provided node types, it can be sampled as a destination node. + source_edge_types_names: Optional[List[str]] + Edge type names of the nodes to be samples as sources. + If a node has any of the provided edge types, it can be sampled as a source node. + destination_edge_types_names: Optional[List[str]] + Edge type names of the nodes to be samples as destinations. + If a node has any of the provided edge types, it can be sampled as a destination node. + source_nodes_prefixes: Optional[List[str]] + Prefixes of the nodes names to be samples as sources. + If a node starts with any of the provided prefixes, it can be sampled as a source node. + destination_nodes_prefixes: Optional[List[str]] + Prefixes of the nodes names to be samples as destinations. + If a node starts with any of the provided prefixes, it can be sampled as a destinations node. validation_unbalance_rates: Tuple[float] = (1.0, ) Unbalance rate for the non-existent graphs generation. use_scale_free_distribution: bool = True @@ -154,6 +178,12 @@ def edge_prediction_evaluation( slurm_node_id_variable=slurm_node_id_variable, verbose=verbose, validation_sample_only_edges_with_heterogeneous_node_types=validation_sample_only_edges_with_heterogeneous_node_types, + source_node_types_names=source_node_types_names, + destination_node_types_names=destination_node_types_names, + source_edge_types_names=source_edge_types_names, + destination_edge_types_names=destination_edge_types_names, + source_nodes_prefixes=source_nodes_prefixes, + destination_nodes_prefixes=destination_nodes_prefixes, validation_unbalance_rates=validation_unbalance_rates, use_scale_free_distribution=use_scale_free_distribution ) diff --git a/embiggen/edge_prediction/edge_prediction_model.py b/embiggen/edge_prediction/edge_prediction_model.py index 0edf5dda..4426c63c 100644 --- a/embiggen/edge_prediction/edge_prediction_model.py +++ b/embiggen/edge_prediction/edge_prediction_model.py @@ -101,6 +101,12 @@ def __iterate_negative_graphs( random_state: int, verbose: bool, validation_sample_only_edges_with_heterogeneous_node_types: bool, + source_node_types_names: Optional[List[str]], + destination_node_types_names: Optional[List[str]], + source_edge_types_names: Optional[List[str]], + destination_edge_types_names: Optional[List[str]], + source_nodes_prefixes: Optional[List[str]], + destination_nodes_prefixes: Optional[List[str]], validation_unbalance_rates: Tuple[float], use_scale_free_distribution: bool ) -> Iterator[Tuple[Graph]]: @@ -136,6 +142,12 @@ def __iterate_negative_graphs( random_state=random_state*(i+1), sample_only_edges_with_heterogeneous_node_types=validation_sample_only_edges_with_heterogeneous_node_types, use_scale_free_distribution=use_scale_free_distribution, + source_node_types_names=source_node_types_names, + destination_node_types_names=destination_node_types_names, + source_edge_types_names=source_edge_types_names, + destination_edge_types_names=destination_edge_types_names, + source_nodes_prefixes=source_nodes_prefixes, + destination_nodes_prefixes=destination_nodes_prefixes, support=support, graph_to_avoid=graph ).random_holdout( @@ -164,6 +176,12 @@ def _prepare_evaluation( random_state: int = 42, verbose: bool = True, validation_sample_only_edges_with_heterogeneous_node_types: bool = False, + source_node_types_names: Optional[List[str]] = None, + destination_node_types_names: Optional[List[str]] = None, + source_edge_types_names: Optional[List[str]] = None, + destination_edge_types_names: Optional[List[str]] = None, + source_nodes_prefixes: Optional[List[str]] = None, + destination_nodes_prefixes: Optional[List[str]] = None, validation_unbalance_rates: Tuple[float] = (1.0, ), use_scale_free_distribution: bool = True ) -> Dict[str, Any]: @@ -178,6 +196,12 @@ def _prepare_evaluation( random_state=random_state, verbose=verbose, validation_sample_only_edges_with_heterogeneous_node_types=validation_sample_only_edges_with_heterogeneous_node_types, + source_node_types_names=source_node_types_names, + destination_node_types_names=destination_node_types_names, + source_edge_types_names=source_edge_types_names, + destination_edge_types_names=destination_edge_types_names, + source_nodes_prefixes=source_nodes_prefixes, + destination_nodes_prefixes=destination_nodes_prefixes, validation_unbalance_rates=validation_unbalance_rates, use_scale_free_distribution=use_scale_free_distribution )) @@ -197,6 +221,12 @@ def _evaluate( verbose: bool = True, negative_graphs: Optional[List[Tuple[Graph]]] = None, validation_sample_only_edges_with_heterogeneous_node_types: bool = False, + source_node_types_names: Optional[List[str]] = None, + destination_node_types_names: Optional[List[str]] = None, + source_edge_types_names: Optional[List[str]] = None, + destination_edge_types_names: Optional[List[str]] = None, + source_nodes_prefixes: Optional[List[str]] = None, + destination_nodes_prefixes: Optional[List[str]] = None, validation_unbalance_rates: Tuple[float] = (1.0, ), use_scale_free_distribution: bool = True, ) -> List[Dict[str, Any]]: @@ -232,6 +262,12 @@ def _evaluate( random_state=random_state, verbose=verbose, validation_sample_only_edges_with_heterogeneous_node_types=validation_sample_only_edges_with_heterogeneous_node_types, + source_node_types_names=source_node_types_names, + destination_node_types_names=destination_node_types_names, + source_edge_types_names=source_edge_types_names, + destination_edge_types_names=destination_edge_types_names, + source_nodes_prefixes=source_nodes_prefixes, + destination_nodes_prefixes=destination_nodes_prefixes, validation_unbalance_rates=validation_unbalance_rates, use_scale_free_distribution=use_scale_free_distribution ) if negative_graphs is None else negative_graphs diff --git a/embiggen/embedders/ensmallen_embedders/rubicone.py b/embiggen/embedders/ensmallen_embedders/rubicone.py new file mode 100644 index 00000000..d01807d1 --- /dev/null +++ b/embiggen/embedders/ensmallen_embedders/rubicone.py @@ -0,0 +1,122 @@ +"""Module providing RUBICONE implementation.""" +from typing import Optional, Dict, Any +from ensmallen import Graph +import pandas as pd +from ensmallen import models +from embiggen.embedders.ensmallen_embedders.ensmallen_embedder import EnsmallenEmbedder +from embiggen.utils import EmbeddingResult + + +class RUBICONE(EnsmallenEmbedder): + """Class implementing the RUBICONE algorithm.""" + + def __init__( + self, + embedding_size: int = 100, + dtype: Optional[str] = "u8", + number_of_convolutions: int = 2, + path: Optional[str] = None, + verbose: bool = False, + ring_bell: bool = False, + enable_cache: bool = False + ): + """Create new RUBICONE method. + + Parameters + -------------------------- + embedding_size: int = 100 + Dimension of the embedding. + dtype: Optional[str] = "u8" + Dtype to use for the embedding. + number_of_convolutions: int = 2 + Number of convolutions to execute + path: Optional[str] = None + Path where to store the mmap-ed embedding. + This parameter is necessary to embed very large graphs. + verbose: bool = False + Whether to show loading bars. + ring_bell: bool = False, + Whether to play a sound when embedding completes. + enable_cache: bool = False + Whether to enable the cache, that is to + store the computed embedding. + """ + self._dtype = dtype + self._verbose = verbose + self._number_of_convolutions = number_of_convolutions + self._path = path + self._model = models.RUBICONE( + embedding_size=embedding_size, + verbose=self._verbose, + number_of_convolutions=self._number_of_convolutions, + path=self._path + ) + + super().__init__( + embedding_size=embedding_size, + ring_bell=ring_bell, + enable_cache=enable_cache + ) + + def parameters(self) -> Dict[str, Any]: + """Returns parameters of the model.""" + return dict( + **super().parameters(), + **dict( + dtype=self._dtype, + number_of_convolutions=self._number_of_convolutions, + path=self._path, + ) + ) + + @classmethod + def smoke_test_parameters(cls) -> Dict[str, Any]: + """Returns parameters for smoke test.""" + return dict( + embedding_size=5, + ) + + def _fit_transform( + self, + graph: Graph, + return_dataframe: bool = True, + ) -> EmbeddingResult: + """Return node embedding.""" + node_embedding = self._model.fit_transform( + graph, + dtype=self._dtype, + ) + if return_dataframe: + node_embedding = pd.DataFrame( + node_embedding, + index=graph.get_node_names() + ) + return EmbeddingResult( + embedding_method_name=self.model_name(), + node_embeddings=node_embedding + ) + + @classmethod + def model_name(cls) -> str: + """Returns name of the model.""" + return "RUBICONE" + + @classmethod + def can_use_node_types(cls) -> bool: + """Returns whether the model can optionally use node types.""" + return False + + @classmethod + def can_use_edge_weights(cls) -> bool: + """Returns whether the model can optionally use edge weights.""" + return False + + @classmethod + def can_use_edge_types(cls) -> bool: + """Returns whether the model can optionally use edge types.""" + return False + + @classmethod + def is_stocastic(cls) -> bool: + """Returns whether the model is stocastic and has therefore a random state.""" + return False \ No newline at end of file diff --git a/embiggen/embedders/ensmallen_embedders/ruine.py b/embiggen/embedders/ensmallen_embedders/ruine.py new file mode 100644 index 00000000..9393b70e --- /dev/null +++ b/embiggen/embedders/ensmallen_embedders/ruine.py @@ -0,0 +1,122 @@ +"""Module providing RUINE implementation.""" +from typing import Optional, Dict, Any +from ensmallen import Graph +import pandas as pd +from ensmallen import models +from embiggen.embedders.ensmallen_embedders.ensmallen_embedder import EnsmallenEmbedder +from embiggen.utils import EmbeddingResult + + +class RUINE(EnsmallenEmbedder): + """Class implementing the RUINE algorithm.""" + + def __init__( + self, + embedding_size: int = 100, + dtype: Optional[str] = "u8", + number_of_convolutions: int = 2, + path: Optional[str] = None, + verbose: bool = False, + ring_bell: bool = False, + enable_cache: bool = False + ): + """Create new RUINE method. + + Parameters + -------------------------- + embedding_size: int = 100 + Dimension of the embedding. + dtype: Optional[str] = "u8" + Dtype to use for the embedding. + number_of_convolutions: int = 2 + Number of convolutions to execute + path: Optional[str] = None + Path where to store the mmap-ed embedding. + This parameter is necessary to embed very large graphs. + verbose: bool = False + Whether to show loading bars. + ring_bell: bool = False, + Whether to play a sound when embedding completes. + enable_cache: bool = False + Whether to enable the cache, that is to + store the computed embedding. + """ + self._dtype = dtype + self._verbose = verbose + self._number_of_convolutions = number_of_convolutions + self._path = path + self._model = models.RUINE( + embedding_size=embedding_size, + verbose=self._verbose, + number_of_convolutions=self._number_of_convolutions, + path=self._path + ) + + super().__init__( + embedding_size=embedding_size, + ring_bell=ring_bell, + enable_cache=enable_cache + ) + + def parameters(self) -> Dict[str, Any]: + """Returns parameters of the model.""" + return dict( + **super().parameters(), + **dict( + dtype=self._dtype, + number_of_convolutions=self._number_of_convolutions, + path=self._path, + ) + ) + + @classmethod + def smoke_test_parameters(cls) -> Dict[str, Any]: + """Returns parameters for smoke test.""" + return dict( + embedding_size=5, + ) + + def _fit_transform( + self, + graph: Graph, + return_dataframe: bool = True, + ) -> EmbeddingResult: + """Return node embedding.""" + node_embedding = self._model.fit_transform( + graph, + dtype=self._dtype, + ) + if return_dataframe: + node_embedding = pd.DataFrame( + node_embedding, + index=graph.get_node_names() + ) + return EmbeddingResult( + embedding_method_name=self.model_name(), + node_embeddings=node_embedding + ) + + @classmethod + def model_name(cls) -> str: + """Returns name of the model.""" + return "RUINE" + + @classmethod + def can_use_node_types(cls) -> bool: + """Returns whether the model can optionally use node types.""" + return False + + @classmethod + def can_use_edge_weights(cls) -> bool: + """Returns whether the model can optionally use edge weights.""" + return False + + @classmethod + def can_use_edge_types(cls) -> bool: + """Returns whether the model can optionally use edge types.""" + return False + + @classmethod + def is_stocastic(cls) -> bool: + """Returns whether the model is stocastic and has therefore a random state.""" + return False \ No newline at end of file diff --git a/embiggen/node_label_prediction/__init__.py b/embiggen/node_label_prediction/__init__.py index e4c95469..2b894472 100644 --- a/embiggen/node_label_prediction/__init__.py +++ b/embiggen/node_label_prediction/__init__.py @@ -1,5 +1,6 @@ """Submodule providing models for edge-label prediction.""" from embiggen.node_label_prediction.node_label_prediction_sklearn import * +from embiggen.node_label_prediction.node_label_prediction_ensmallen import * from embiggen.node_label_prediction.node_label_prediction_tensorflow import * from embiggen.node_label_prediction.node_label_prediction_evaluation import node_label_prediction_evaluation diff --git a/embiggen/node_label_prediction/node_label_prediction_ensmallen/__init__.py b/embiggen/node_label_prediction/node_label_prediction_ensmallen/__init__.py new file mode 100644 index 00000000..81c097da --- /dev/null +++ b/embiggen/node_label_prediction/node_label_prediction_ensmallen/__init__.py @@ -0,0 +1,9 @@ +"""Submodule providing node embedding models implemented in Ensmallen in Rust.""" +from embiggen.node_label_prediction.node_label_prediction_model import AbstractNodeLabelPredictionModel +from embiggen.utils.abstract_models import build_init + +build_init( + module_library_names="ensmallen", + formatted_library_name="Ensmallen", + expected_parent_class=AbstractNodeLabelPredictionModel +) diff --git a/embiggen/node_label_prediction/node_label_prediction_ensmallen/distance_based_perceptron.py b/embiggen/node_label_prediction/node_label_prediction_ensmallen/distance_based_perceptron.py new file mode 100644 index 00000000..a62133b6 --- /dev/null +++ b/embiggen/node_label_prediction/node_label_prediction_ensmallen/distance_based_perceptron.py @@ -0,0 +1,245 @@ +"""Module providing Perceptron for edge prediction.""" +from typing import Optional, Dict, Any, List +from ensmallen import Graph +import numpy as np +from ensmallen import models +import compress_json +import json +from embiggen.node_label_prediction.node_label_prediction_model import AbstractNodeLabelPredictionModel + + +class DistanceBasedPerceptronNodeLabelPrediction(AbstractNodeLabelPredictionModel): + """Distance-based Perceptron model for edge prediction.""" + + def __init__( + self, + number_of_centroids_per_class: int = 1, + number_of_epochs: int = 100, + learning_rate: float = 0.01, + first_order_decay_factor: float = 0.9, + second_order_decay_factor: float = 0.999, + random_state: int = 42, + verbose: bool = True + ): + """Create new Distance-based Perceptron object. + + Parameters + ------------------------ + number_of_centroids_per_class: int = 1 + Number of centroids to compute for each of the classes. + number_of_epochs: int = 100 + The number of epochs to train the model for. By default, 100. + learning_rate: float = 0.01 + Learning rate to use while training the model. By default 0.001. + first_order_decay_factor: float = 0.9 + First order decay factor for the first order momentum. + By default 0.9. + second_order_decay_factor: float = 0.999 + Second order decay factor for the second order momentum. + By default 0.999. + random_state: int = 42 + The random state to reproduce the model initialization and training. By default, 42. + verbose: bool = True + Whether to show epochs loading bar. + """ + super().__init__(random_state=random_state) + + self._model_kwargs = dict( + number_of_centroids_per_class=number_of_centroids_per_class, + number_of_epochs=number_of_epochs, + learning_rate=learning_rate, + first_order_decay_factor=first_order_decay_factor, + second_order_decay_factor=second_order_decay_factor, + ) + self._verbose = verbose + self._model = models.DistanceNodeLabelPredictionPerceptron( + **self._model_kwargs, + random_state=random_state + ) + + def parameters(self) -> Dict[str, Any]: + """Returns parameters used for this model.""" + return dict( + **super().parameters(), + **self._model_kwargs + ) + + def clone(self) -> "DistanceBasedPerceptronNodeLabelPrediction": + return DistanceBasedPerceptronNodeLabelPrediction(**self.parameters()) + + @classmethod + def smoke_test_parameters(cls) -> Dict[str, Any]: + """Returns parameters for smoke test.""" + return dict( + number_of_epochs=1, + ) + + def _fit( + self, + graph: Graph, + support: Optional[Graph] = None, + node_features: Optional[List[np.ndarray]] = None, + node_type_features: Optional[List[np.ndarray]] = None, + edge_features: Optional[List[np.ndarray]] = None, + ): + """Run fitting on the provided graph. + + Parameters + -------------------- + graph: Graph + The graph to run predictions on. + support: Optional[Graph] = None + The graph describiding the topological structure that + includes also the above graph. This parameter + is mostly useful for topological classifiers + such as Graph Convolutional Networks. + node_features: Optional[List[np.ndarray]] = None + The node features to use. + node_type_features: Optional[List[np.ndarray]] = None + The node type features to use. + edge_features: Optional[List[np.ndarray]] = None + The edge features to use. + """ + new_node_features = [] + if node_features is not None: + for node_feature in node_features: + if not node_feature.data.c_contiguous: + node_feature = np.ascontiguousarray(node_feature) + new_node_features.append(node_feature) + + self._model.fit( + graph=graph, + node_features=new_node_features, + verbose=self._verbose, + ) + + def _predict( + self, + graph: Graph, + support: Optional[Graph] = None, + node_features: Optional[List[np.ndarray]] = None, + node_type_features: Optional[List[np.ndarray]] = None, + edge_features: Optional[List[np.ndarray]] = None, + ) -> np.ndarray: + """Run prediction on the provided graph. + + Parameters + -------------------- + graph: Graph + The graph to run predictions on. + support: Optional[Graph] = None + The graph describiding the topological structure that + includes also the above graph. This parameter + is mostly useful for topological classifiers + such as Graph Convolutional Networks. + node_features: Optional[List[np.ndarray]] + The node features to use. + node_type_features: Optional[List[np.ndarray]] = None + The node type features to use. + edge_features: Optional[List[np.ndarray]] = None + The edge features to use. + """ + return self._predict_proba( + graph=graph, + support=support, + node_features=node_features, + node_type_features=node_type_features, + edge_features=edge_features + ).argmax(axis=1) + + def _predict_proba( + self, + graph: Graph, + support: Optional[Graph] = None, + node_features: Optional[List[np.ndarray]] = None, + node_type_features: Optional[List[np.ndarray]] = None, + edge_features: Optional[List[np.ndarray]] = None, + ) -> np.ndarray: + """Run prediction on the provided graph. + + Parameters + -------------------- + graph: Graph + The graph to run predictions on. + support: Optional[Graph] = None + The graph describiding the topological structure that + includes also the above graph. This parameter + is mostly useful for topological classifiers + such as Graph Convolutional Networks. + node_features: Optional[List[np.ndarray]] = None + The node features to use. + node_type_features: Optional[List[np.ndarray]] = None + The node type features to use. + edge_features: Optional[List[np.ndarray]] = None + The edge features to use. + """ + new_node_features = [] + if node_features is not None: + for node_feature in node_features: + if not node_feature.data.c_contiguous: + node_feature = np.ascontiguousarray(node_feature) + new_node_features.append(node_feature) + + return self._model.predict( + graph=graph, + node_features=new_node_features, + ) + + @classmethod + def can_use_edge_weights(cls) -> bool: + """Returns whether the model can optionally use edge weights.""" + return False + + @classmethod + def can_use_edge_types(cls) -> bool: + """Returns whether the model can optionally use edge types.""" + return False + + @classmethod + def model_name(cls) -> str: + return "Distance-based Perceptron" + + @classmethod + def library_name(cls) -> str: + return "Ensmallen" + + @classmethod + def load(cls, path: str) -> "Self": + """Load a saved version of the model from the provided path. + + Parameters + ------------------- + path: str + Path from where to load the model. + """ + data = compress_json.load(path) + model = DistanceBasedPerceptronNodeLabelPrediction(**data["parameters"]) + model._model = models.DistanceNodeLabelPredictionPerceptron.loads( + json.dumps(data["inner_model"]) + ) + for key, value in data["metadata"].items(): + model.__setattr__(key, value) + return model + + def dumps(self) -> Dict[str, Any]: + """Dumps the current model as dictionary.""" + return dict( + parameters=self.parameters(), + inner_model=json.loads(self._model.dumps()), + metadata=dict( + _fitting_was_executed=self._fitting_was_executed + ) + ) + + def dump(self, path: str): + """Dump the current model at the provided path. + + Parameters + ------------------- + path: str + Path from where to dump the model. + """ + compress_json.dump( + self.dumps(), + path + ) diff --git a/embiggen/node_label_prediction/node_label_prediction_ensmallen/perceptron.py b/embiggen/node_label_prediction/node_label_prediction_ensmallen/perceptron.py new file mode 100644 index 00000000..6b2b254e --- /dev/null +++ b/embiggen/node_label_prediction/node_label_prediction_ensmallen/perceptron.py @@ -0,0 +1,241 @@ +"""Module providing Perceptron for edge prediction.""" +from typing import Optional, Dict, Any, List +from ensmallen import Graph +import numpy as np +from ensmallen import models +import compress_json +import json +from embiggen.node_label_prediction.node_label_prediction_model import AbstractNodeLabelPredictionModel + + +class PerceptronNodeLabelPrediction(AbstractNodeLabelPredictionModel): + """Perceptron model for edge prediction.""" + + def __init__( + self, + number_of_epochs: int = 100, + learning_rate: float = 0.01, + first_order_decay_factor: float = 0.9, + second_order_decay_factor: float = 0.999, + random_state: int = 42, + verbose: bool = True + ): + """Create new Perceptron object. + + Parameters + ------------------------ + number_of_epochs: int = 100 + The number of epochs to train the model for. By default, 100. + learning_rate: float = 0.01 + Learning rate to use while training the model. By default 0.001. + first_order_decay_factor: float = 0.9 + First order decay factor for the first order momentum. + By default 0.9. + second_order_decay_factor: float = 0.999 + Second order decay factor for the second order momentum. + By default 0.999. + random_state: int = 42 + The random state to reproduce the model initialization and training. By default, 42. + verbose: bool = True + Whether to show epochs loading bar. + """ + super().__init__(random_state=random_state) + + self._model_kwargs = dict( + number_of_epochs=number_of_epochs, + learning_rate=learning_rate, + first_order_decay_factor=first_order_decay_factor, + second_order_decay_factor=second_order_decay_factor, + ) + self._verbose = verbose + self._model = models.NodeLabelPredictionPerceptron( + **self._model_kwargs, + random_state=random_state + ) + + def parameters(self) -> Dict[str, Any]: + """Returns parameters used for this model.""" + return dict( + **super().parameters(), + **self._model_kwargs + ) + + def clone(self) -> "PerceptronNodeLabelPrediction": + return PerceptronNodeLabelPrediction(**self.parameters()) + + @classmethod + def smoke_test_parameters(cls) -> Dict[str, Any]: + """Returns parameters for smoke test.""" + return dict( + number_of_epochs=1, + ) + + def _fit( + self, + graph: Graph, + support: Optional[Graph] = None, + node_features: Optional[List[np.ndarray]] = None, + node_type_features: Optional[List[np.ndarray]] = None, + edge_features: Optional[List[np.ndarray]] = None, + ): + """Run fitting on the provided graph. + + Parameters + -------------------- + graph: Graph + The graph to run predictions on. + support: Optional[Graph] = None + The graph describiding the topological structure that + includes also the above graph. This parameter + is mostly useful for topological classifiers + such as Graph Convolutional Networks. + node_features: Optional[List[np.ndarray]] = None + The node features to use. + node_type_features: Optional[List[np.ndarray]] = None + The node type features to use. + edge_features: Optional[List[np.ndarray]] = None + The edge features to use. + """ + new_node_features = [] + if node_features is not None: + for node_feature in node_features: + if not node_feature.data.c_contiguous: + node_feature = np.ascontiguousarray(node_feature) + new_node_features.append(node_feature) + + self._model.fit( + graph=graph, + node_features=new_node_features, + verbose=self._verbose, + ) + + def _predict( + self, + graph: Graph, + support: Optional[Graph] = None, + node_features: Optional[List[np.ndarray]] = None, + node_type_features: Optional[List[np.ndarray]] = None, + edge_features: Optional[List[np.ndarray]] = None, + ) -> np.ndarray: + """Run prediction on the provided graph. + + Parameters + -------------------- + graph: Graph + The graph to run predictions on. + support: Optional[Graph] = None + The graph describiding the topological structure that + includes also the above graph. This parameter + is mostly useful for topological classifiers + such as Graph Convolutional Networks. + node_features: Optional[List[np.ndarray]] + The node features to use. + node_type_features: Optional[List[np.ndarray]] = None + The node type features to use. + edge_features: Optional[List[np.ndarray]] = None + The edge features to use. + """ + return self._predict_proba( + graph=graph, + support=support, + node_features=node_features, + node_type_features=node_type_features, + edge_features=edge_features + ).argmax(axis=1) + + def _predict_proba( + self, + graph: Graph, + support: Optional[Graph] = None, + node_features: Optional[List[np.ndarray]] = None, + node_type_features: Optional[List[np.ndarray]] = None, + edge_features: Optional[List[np.ndarray]] = None, + ) -> np.ndarray: + """Run prediction on the provided graph. + + Parameters + -------------------- + graph: Graph + The graph to run predictions on. + support: Optional[Graph] = None + The graph describiding the topological structure that + includes also the above graph. This parameter + is mostly useful for topological classifiers + such as Graph Convolutional Networks. + node_features: Optional[List[np.ndarray]] = None + The node features to use. + node_type_features: Optional[List[np.ndarray]] = None + The node type features to use. + edge_features: Optional[List[np.ndarray]] = None + The edge features to use. + """ + new_node_features = [] + if node_features is not None: + for node_feature in node_features: + if not node_feature.data.c_contiguous: + node_feature = np.ascontiguousarray(node_feature) + new_node_features.append(node_feature) + + return self._model.predict( + graph=graph, + node_features=new_node_features, + ) + + @classmethod + def can_use_edge_weights(cls) -> bool: + """Returns whether the model can optionally use edge weights.""" + return False + + @classmethod + def can_use_edge_types(cls) -> bool: + """Returns whether the model can optionally use edge types.""" + return False + + @classmethod + def model_name(cls) -> str: + return "Perceptron" + + @classmethod + def library_name(cls) -> str: + return "Ensmallen" + + @classmethod + def load(cls, path: str) -> "Self": + """Load a saved version of the model from the provided path. + + Parameters + ------------------- + path: str + Path from where to load the model. + """ + data = compress_json.load(path) + model = PerceptronNodeLabelPrediction(**data["parameters"]) + model._model = models.NodeLabelPredictionPerceptron.loads( + json.dumps(data["inner_model"]) + ) + for key, value in data["metadata"].items(): + model.__setattr__(key, value) + return model + + def dumps(self) -> Dict[str, Any]: + """Dumps the current model as dictionary.""" + return dict( + parameters=self.parameters(), + inner_model=json.loads(self._model.dumps()), + metadata=dict( + _fitting_was_executed=self._fitting_was_executed + ) + ) + + def dump(self, path: str): + """Dump the current model at the provided path. + + Parameters + ------------------- + path: str + Path from where to dump the model. + """ + compress_json.dump( + self.dumps(), + path + ) diff --git a/embiggen/similarities/dag_resnik.py b/embiggen/similarities/dag_resnik.py index c78ba754..5a765993 100644 --- a/embiggen/similarities/dag_resnik.py +++ b/embiggen/similarities/dag_resnik.py @@ -107,6 +107,12 @@ def get_similarities_from_bipartite_graph_node_ids( Whether to return the node names or node IDs associated to the scores. By default we return the node ids, which require much less memory. """ + if isinstance(source_node_ids, int): + source_node_ids = [source_node_ids] + + if isinstance(destination_node_ids, int): + destination_node_ids = [destination_node_ids] + return self._normalize_output( *self._model.get_node_ids_and_similarity_from_node_ids( first_node_ids=source_node_ids, @@ -142,6 +148,12 @@ def get_similarities_from_bipartite_graph_node_names( Whether to return the node names or node IDs associated to the scores. By default we return the node ids, which require much less memory. """ + if isinstance(source_node_names, str): + source_node_names = [source_node_names] + + if isinstance(destination_node_names, str): + destination_node_names = [destination_node_names] + return self._normalize_output( *self._model.get_node_ids_and_similarity_from_node_names( first_node_names=source_node_names, @@ -177,6 +189,12 @@ def get_similarities_from_bipartite_graph_node_prefixes( Whether to return the node names or node IDs associated to the scores. By default we return the node ids, which require much less memory. """ + if isinstance(source_node_prefixes, str): + source_node_prefixes = [source_node_prefixes] + + if isinstance(destination_node_prefixes, str): + destination_node_prefixes = [destination_node_prefixes] + return self._normalize_output( *self._model.get_node_ids_and_similarity_from_node_prefixes( first_node_prefixes=source_node_prefixes, @@ -212,6 +230,12 @@ def get_similarities_from_bipartite_graph_node_type_ids( Whether to return the node names or node IDs associated to the scores. By default we return the node ids, which require much less memory. """ + if isinstance(source_node_type_ids, int): + source_node_type_ids = [source_node_type_ids] + + if isinstance(destination_node_type_ids, int): + destination_node_type_ids = [destination_node_type_ids] + return self._normalize_output( *self._model.get_node_ids_and_similarity_from_node_type_ids( first_node_type_ids=source_node_type_ids, @@ -247,6 +271,12 @@ def get_similarities_from_bipartite_graph_node_type_names( Whether to return the node names or node IDs associated to the scores. By default we return the node ids, which require much less memory. """ + if isinstance(source_node_type_names, str): + source_node_type_names = [source_node_type_names] + + if isinstance(destination_node_type_names, str): + destination_node_type_names = [destination_node_type_names] + return self._normalize_output( *self._model.get_node_ids_and_similarity_from_node_type_names( first_node_type_names=source_node_type_names, diff --git a/embiggen/utils/abstract_models/embedding_result.py b/embiggen/utils/abstract_models/embedding_result.py index 4dd922f4..dd610656 100644 --- a/embiggen/utils/abstract_models/embedding_result.py +++ b/embiggen/utils/abstract_models/embedding_result.py @@ -62,12 +62,15 @@ def __init__( f"computed with the {embedding_method_name} method is neither a " f"numpy array or a pandas DataFrame, but a `{type(embedding)}` object." ) + if embedding.shape[0] == 0: raise ValueError( "One of the provided {embedding_list_name} " f"computed with the {embedding_method_name} method " "is empty." ) + + # If the embedding size is too big, we skip the checking step. if embedding.shape[0] > 1_000_000: continue diff --git a/embiggen/visualizations/graph_visualizer.py b/embiggen/visualizations/graph_visualizer.py index 2043217b..3c567c8c 100644 --- a/embiggen/visualizations/graph_visualizer.py +++ b/embiggen/visualizations/graph_visualizer.py @@ -4237,9 +4237,9 @@ def plot_node_degree_distribution( axes.set_ylabel("Counts (log scale)") axes.set_xlabel("Degrees") if self._show_graph_name: - title = "Degrees distribution of graph {}".format(self._graph_name) + title = "Degree distribution of graph {}".format(self._graph_name) else: - title = "Degrees distribution" + title = "Degree distribution" axes.set_title(title) if apply_tight_layout: figure.tight_layout() @@ -4248,7 +4248,7 @@ def plot_node_degree_distribution( return self._handle_notebook_display(figure, axes) caption = ( - "Node degrees distribution. Node degrees are on the " + "Node degree distribution. Node degrees are on the " "horizontal axis and node counts are on the vertical axis on a logarithmic scale." ) diff --git a/setup.py b/setup.py index ae9a18ea..151159bc 100644 --- a/setup.py +++ b/setup.py @@ -64,14 +64,15 @@ def find_version(*file_paths): 'numpy', 'pandas', "tqdm", + "humanize", "matplotlib>=3.5.2", "scikit-learn", "dict_hash>=1.1.29", "userinput>=1.0.19", "ddd_subplots>=1.0.23", - "sanitize_ml_labels>=1.0.45", + "sanitize_ml_labels>=1.0.50", "keras_mixed_sequence>=1.0.28", - "ensmallen>=0.8.27", + "ensmallen>=0.8.31", "environments_utils>=1.0.6", "compress_pickle>=2.1.0", "validate_version_code",

Brought develop up to speed

2022-11-11T14:25:57

0.0

[]

monarch-initiative/embiggen

monarch-initiative__embiggen-297

3ee8055ad4dce31a1cc04bc071116155bdbcd57d

diff --git a/.gitignore b/.gitignore index 6634e04f..b8107f2c 100644 --- a/.gitignore +++ b/.gitignore @@ -34,4 +34,7 @@ embedding.npy build graphs embeddings -*.webm \ No newline at end of file +*.webm + +model.json +model.pkl.gz \ No newline at end of file diff --git a/README.rst b/README.rst index 87387cda..4d3e7707 100644 --- a/README.rst +++ b/README.rst @@ -2,11 +2,11 @@ ========================================================================================= |pip| |downloads| |tutorials| |documentation| |python_version| |DOI| |license| |telegram| |discord| |twitter| -Embiggen is the graph machine learning submodule of the `🍇 GraPE <https://github.com/AnacletoLAB/grape>`_ library. +Embiggen is the graph machine learning submodule of the `🍇 GRAPE <https://github.com/AnacletoLAB/grape>`_ library. How to install Embiggen ------------------------- -To install the complete GraPE library, do run: +To install the complete GRAPE library, do run: .. code:: bash @@ -19,14 +19,14 @@ Instead, to exclusively install the Embiggen package, you can run: pip install embiggen -Cite GraPE +Cite GRAPE ---------------------------------------------- Please cite the following paper if it was useful for your research: .. code:: bib @misc{cappelletti2021grape, - title={GraPE: fast and scalable Graph Processing and Embedding}, + title={GRAPE: fast and scalable Graph Processing and Embedding}, author={Luca Cappelletti and Tommaso Fontana and Elena Casiraghi and Vida Ravanmehr and Tiffany J. Callahan and Marcin P. Joachimiak and Christopher J. Mungall and Peter N. Robinson and Justin Reese and Giorgio Valentini}, year={2021}, eprint={2110.06196}, @@ -58,7 +58,7 @@ Please cite the following paper if it was useful for your research: :target: https://doi.org/10.48550/arXiv.2110.06196 :alt: DOI -.. |python_version| image:: https://img.shields.io/badge/Python-3.6+-blue.svg +.. |python_version| image:: https://img.shields.io/badge/Python-3.7+-blue.svg :target: https://pypi.org/project/embiggen/#history :alt: Supported Python versions @@ -72,4 +72,4 @@ Please cite the following paper if it was useful for your research: .. |twitter| image:: https://badges.aleen42.com/src/twitter.svg :target: https://twitter.com/grapelib - :alt: Twitter \ No newline at end of file + :alt: Twitter diff --git a/dist/embiggen-0.11.11.tar.gz b/dist/embiggen-0.11.11.tar.gz new file mode 100644 index 00000000..3b80024f Binary files /dev/null and b/dist/embiggen-0.11.11.tar.gz differ diff --git a/dist/embiggen-0.11.16.tar.gz b/dist/embiggen-0.11.16.tar.gz new file mode 100644 index 00000000..b8076636 Binary files /dev/null and b/dist/embiggen-0.11.16.tar.gz differ diff --git a/dist/embiggen-0.11.17.tar.gz b/dist/embiggen-0.11.17.tar.gz new file mode 100644 index 00000000..2313906e Binary files /dev/null and b/dist/embiggen-0.11.17.tar.gz differ diff --git a/dist/embiggen-0.11.18.tar.gz b/dist/embiggen-0.11.18.tar.gz new file mode 100644 index 00000000..439b6f68 Binary files /dev/null and b/dist/embiggen-0.11.18.tar.gz differ diff --git a/dist/embiggen-0.11.19.tar.gz b/dist/embiggen-0.11.19.tar.gz new file mode 100644 index 00000000..67bae59b Binary files /dev/null and b/dist/embiggen-0.11.19.tar.gz differ diff --git a/dist/embiggen-0.11.20.tar.gz b/dist/embiggen-0.11.20.tar.gz new file mode 100644 index 00000000..70b73659 Binary files /dev/null and b/dist/embiggen-0.11.20.tar.gz differ diff --git a/dist/embiggen-0.11.21.tar.gz b/dist/embiggen-0.11.21.tar.gz new file mode 100644 index 00000000..d1361443 Binary files /dev/null and b/dist/embiggen-0.11.21.tar.gz differ diff --git a/dist/embiggen-0.11.22.tar.gz b/dist/embiggen-0.11.22.tar.gz new file mode 100644 index 00000000..1518eac1 Binary files /dev/null and b/dist/embiggen-0.11.22.tar.gz differ diff --git a/dist/embiggen-0.11.23.tar.gz b/dist/embiggen-0.11.23.tar.gz new file mode 100644 index 00000000..42de56f0 Binary files /dev/null and b/dist/embiggen-0.11.23.tar.gz differ diff --git a/dist/embiggen-0.11.24.tar.gz b/dist/embiggen-0.11.24.tar.gz new file mode 100644 index 00000000..ea524493 Binary files /dev/null and b/dist/embiggen-0.11.24.tar.gz differ diff --git a/dist/embiggen-0.11.25.tar.gz b/dist/embiggen-0.11.25.tar.gz new file mode 100644 index 00000000..323bb085 Binary files /dev/null and b/dist/embiggen-0.11.25.tar.gz differ diff --git a/dist/embiggen-0.11.26.tar.gz b/dist/embiggen-0.11.26.tar.gz new file mode 100644 index 00000000..0690228b Binary files /dev/null and b/dist/embiggen-0.11.26.tar.gz differ diff --git a/dist/embiggen-0.11.27.tar.gz b/dist/embiggen-0.11.27.tar.gz new file mode 100644 index 00000000..e3969a24 Binary files /dev/null and b/dist/embiggen-0.11.27.tar.gz differ diff --git a/dist/embiggen-0.11.28.tar.gz b/dist/embiggen-0.11.28.tar.gz new file mode 100644 index 00000000..e9d94564 Binary files /dev/null and b/dist/embiggen-0.11.28.tar.gz differ diff --git a/dist/embiggen-0.11.29.tar.gz b/dist/embiggen-0.11.29.tar.gz new file mode 100644 index 00000000..31fabc70 Binary files /dev/null and b/dist/embiggen-0.11.29.tar.gz differ diff --git a/dist/embiggen-0.11.30.tar.gz b/dist/embiggen-0.11.30.tar.gz new file mode 100644 index 00000000..d49052f3 Binary files /dev/null and b/dist/embiggen-0.11.30.tar.gz differ diff --git a/dist/embiggen-0.11.31.tar.gz b/dist/embiggen-0.11.31.tar.gz new file mode 100644 index 00000000..6a8086e4 Binary files /dev/null and b/dist/embiggen-0.11.31.tar.gz differ diff --git a/dist/embiggen-0.11.32.tar.gz b/dist/embiggen-0.11.32.tar.gz new file mode 100644 index 00000000..c2042520 Binary files /dev/null and b/dist/embiggen-0.11.32.tar.gz differ diff --git a/dist/embiggen-0.11.33.tar.gz b/dist/embiggen-0.11.33.tar.gz new file mode 100644 index 00000000..21fe8809 Binary files /dev/null and b/dist/embiggen-0.11.33.tar.gz differ diff --git a/dist/embiggen-0.11.34.tar.gz b/dist/embiggen-0.11.34.tar.gz new file mode 100644 index 00000000..ef19be77 Binary files /dev/null and b/dist/embiggen-0.11.34.tar.gz differ diff --git a/dist/embiggen-0.11.35.tar.gz b/dist/embiggen-0.11.35.tar.gz new file mode 100644 index 00000000..7b6e0ce5 Binary files /dev/null and b/dist/embiggen-0.11.35.tar.gz differ diff --git a/dist/embiggen-0.11.36.tar.gz b/dist/embiggen-0.11.36.tar.gz new file mode 100644 index 00000000..db48726d Binary files /dev/null and b/dist/embiggen-0.11.36.tar.gz differ diff --git a/dist/embiggen-0.11.37.tar.gz b/dist/embiggen-0.11.37.tar.gz new file mode 100644 index 00000000..761805dc Binary files /dev/null and b/dist/embiggen-0.11.37.tar.gz differ diff --git a/dist/embiggen-0.11.38.tar.gz b/dist/embiggen-0.11.38.tar.gz new file mode 100644 index 00000000..53c56976 Binary files /dev/null and b/dist/embiggen-0.11.38.tar.gz differ diff --git a/embiggen/__version__.py b/embiggen/__version__.py index e1fd9917..6c3f5032 100644 --- a/embiggen/__version__.py +++ b/embiggen/__version__.py @@ -1,2 +1,2 @@ """Current version of package embiggen""" -__version__ = "0.11.15" +__version__ = "0.11.38" \ No newline at end of file diff --git a/embiggen/edge_label_prediction/edge_label_prediction_evaluation.py b/embiggen/edge_label_prediction/edge_label_prediction_evaluation.py index dec73085..57595239 100644 --- a/embiggen/edge_label_prediction/edge_label_prediction_evaluation.py +++ b/embiggen/edge_label_prediction/edge_label_prediction_evaluation.py @@ -18,13 +18,16 @@ def edge_label_prediction_evaluation( library_names: Optional[Union[str, List[str]]] = None, graph_callback: Optional[Callable[[Graph], Graph]] = None, subgraph_of_interest: Optional[Graph] = None, + use_subgraph_as_support: bool = False, number_of_holdouts: int = 10, random_state: int = 42, repositories: Optional[Union[str, List[str]]] = None, versions: Optional[Union[str, List[str]]] = None, enable_cache: bool = False, - precompute_constant_automatic_stocastic_features: bool = False, + precompute_constant_stocastic_features: bool = False, smoke_test: bool = False, + number_of_slurm_nodes: Optional[int] = None, + slurm_node_id_variable: str = "SLURM_GRAPE_ID", verbose: bool = True ) -> pd.DataFrame: """Execute edge-label prediction evaluation pipeline for all provided models and graphs. @@ -53,7 +56,12 @@ def edge_label_prediction_evaluation( For instance this may be used for filtering the uncertain edges in graphs such as STRING PPIs. subgraph_of_interest: Optional[Graph] = None - The subgraph of interest to focus the task on. + Optional subgraph where to focus the task. + This is applied to the train and test graph + after the desired holdout schema is applied. + use_subgraph_as_support: bool = False + Whether to use the provided subgraph as support or + to use the train graph (not filtered by the subgraph). number_of_holdouts: int = 10 The number of holdouts to execute. random_state: int = 42 @@ -65,7 +73,7 @@ def edge_label_prediction_evaluation( Graph versions to retrieve. enable_cache: bool = False Whether to enable the cache. - precompute_constant_automatic_stocastic_features: bool = False + precompute_constant_stocastic_features: bool = False Whether to precompute once the constant automatic stocastic features before starting the embedding loop. This means that, when left set to false, while the features will be computed @@ -79,6 +87,12 @@ def edge_label_prediction_evaluation( and therefore use the smoke test configurations for the provided model names and feature names. This parameter will also turn off the cache. + number_of_slurm_nodes: Optional[int] = None + Number of SLURM nodes to consider as available. + This variable is used to parallelize the holdouts accordingly. + slurm_node_id_variable: str = "SLURM_GRAPE_ID" + Name of the system variable to use as SLURM node id. + It must be set in the slurm bash script. verbose: bool = True Whether to show loading bars """ @@ -94,12 +108,15 @@ def edge_label_prediction_evaluation( library_names=library_names, graph_callback=graph_callback, subgraph_of_interest=subgraph_of_interest, + use_subgraph_as_support=use_subgraph_as_support, number_of_holdouts=number_of_holdouts, random_state=random_state, repositories=repositories, versions=versions, enable_cache=enable_cache, - precompute_constant_automatic_stocastic_features=precompute_constant_automatic_stocastic_features, + precompute_constant_stocastic_features=precompute_constant_stocastic_features, smoke_test=smoke_test, + number_of_slurm_nodes=number_of_slurm_nodes, + slurm_node_id_variable=slurm_node_id_variable, verbose=verbose ) diff --git a/embiggen/edge_label_prediction/edge_label_prediction_sklearn/sklearn_edge_label_prediction_adapter.py b/embiggen/edge_label_prediction/edge_label_prediction_sklearn/sklearn_edge_label_prediction_adapter.py index 013d081c..cf98e7fa 100644 --- a/embiggen/edge_label_prediction/edge_label_prediction_sklearn/sklearn_edge_label_prediction_adapter.py +++ b/embiggen/edge_label_prediction/edge_label_prediction_sklearn/sklearn_edge_label_prediction_adapter.py @@ -3,6 +3,7 @@ from typing import Type, List, Dict, Optional, Any import numpy as np import copy +import compress_pickle from ensmallen import Graph from embiggen.utils.sklearn_utils import must_be_an_sklearn_classifier_model from embiggen.embedding_transformers import EdgeLabelPredictionTransformer, GraphTransformer @@ -295,4 +296,25 @@ def can_use_edge_weights(cls) -> bool: @classmethod def can_use_node_types(cls) -> bool: """Returns whether the model can optionally use node types.""" - return False \ No newline at end of file + return False + + @classmethod + def load(cls, path: str) -> "Self": + """Load a saved version of the model from the provided path. + + Parameters + ------------------- + path: str + Path from where to load the model. + """ + return compress_pickle.load(path) + + def dump(self, path: str): + """Dump the current model at the provided path. + + Parameters + ------------------- + path: str + Path from where to dump the model. + """ + compress_pickle.dump(self, path) \ No newline at end of file diff --git a/embiggen/edge_prediction/edge_prediction_ensmallen/perceptron.py b/embiggen/edge_prediction/edge_prediction_ensmallen/perceptron.py index 5407d194..ea631a63 100644 --- a/embiggen/edge_prediction/edge_prediction_ensmallen/perceptron.py +++ b/embiggen/edge_prediction/edge_prediction_ensmallen/perceptron.py @@ -3,6 +3,8 @@ from ensmallen import Graph import numpy as np from ensmallen import models +import compress_json +import json from embiggen.embedding_transformers import NodeTransformer from embiggen.edge_prediction.edge_prediction_model import AbstractEdgePredictionModel @@ -17,9 +19,9 @@ def __init__( cooccurrence_iterations: int = 100, cooccurrence_window_size: int = 10, number_of_epochs: int = 100, - number_of_edges_per_mini_batch: int = 256, + number_of_edges_per_mini_batch: int = 4096, sample_only_edges_with_heterogeneous_node_types: bool = False, - learning_rate: float = 0.001, + learning_rate: float = 0.01, first_order_decay_factor: float = 0.9, second_order_decay_factor: float = 0.999, avoid_false_negatives: bool = False, @@ -68,7 +70,7 @@ def __init__( sample_only_edges_with_heterogeneous_node_types: bool = False Whether to sample negative edges only with source and destination nodes that have different node types. By default false. - learning_rate: float = 0.001 + learning_rate: float = 0.01 Learning rate to use while training the model. By default 0.001. first_order_decay_factor: float = 0.9 First order decay factor for the first order momentum. @@ -116,7 +118,10 @@ def __init__( def parameters(self) -> Dict[str, Any]: """Returns parameters used for this model.""" - return self._model_kwargs + return dict( + **super().parameters(), + **self._model_kwargs + ) def clone(self) -> "PerceptronEdgePrediction": return PerceptronEdgePrediction(**self.parameters()) @@ -125,7 +130,8 @@ def clone(self) -> "PerceptronEdgePrediction": def smoke_test_parameters(cls) -> Dict[str, Any]: """Returns parameters for smoke test.""" return dict( - number_of_epochs=1 + number_of_epochs=1, + edge_features="Degree" ) def _fit( @@ -176,7 +182,6 @@ def _fit( node_feature = np.ascontiguousarray(node_feature) new_node_features.append(node_feature) - self._model.fit( graph=graph, node_features=new_node_features, @@ -279,3 +284,44 @@ def model_name(cls) -> str: @classmethod def library_name(cls) -> str: return "Ensmallen" + + @classmethod + def load(cls, path: str) -> "Self": + """Load a saved version of the model from the provided path. + + Parameters + ------------------- + path: str + Path from where to load the model. + """ + data = compress_json.load(path) + model = PerceptronEdgePrediction(**data["parameters"]) + model._model = models.EdgePredictionPerceptron.loads( + json.dumps(data["inner_model"]) + ) + for key, value in data["metadata"].items(): + model.__setattr__(key, value) + return model + + def dumps(self) -> Dict[str, Any]: + """Dumps the current model as dictionary.""" + return dict( + parameters=self.parameters(), + inner_model=json.loads(self._model.dumps()), + metadata=dict( + _fitting_was_executed=self._fitting_was_executed + ) + ) + + def dump(self, path: str): + """Dump the current model at the provided path. + + Parameters + ------------------- + path: str + Path from where to dump the model. + """ + compress_json.dump( + self.dumps(), + path + ) diff --git a/embiggen/edge_prediction/edge_prediction_evaluation.py b/embiggen/edge_prediction/edge_prediction_evaluation.py index c8bab4f3..3b7280ff 100644 --- a/embiggen/edge_prediction/edge_prediction_evaluation.py +++ b/embiggen/edge_prediction/edge_prediction_evaluation.py @@ -17,6 +17,7 @@ def edge_prediction_evaluation( library_names: Optional[Union[str, List[str]]] = None, graph_callback: Optional[Callable[[Graph], Graph]] = None, subgraph_of_interest: Optional[Graph] = None, + use_subgraph_as_support: bool = False, number_of_holdouts: int = 10, random_state: int = 42, repositories: Optional[Union[str, List[str]]] = None, @@ -25,8 +26,10 @@ def edge_prediction_evaluation( validation_unbalance_rates: Tuple[float] = (1.0, ), use_scale_free_distribution: bool = True, enable_cache: bool = False, - precompute_constant_automatic_stocastic_features: bool = False, + precompute_constant_stocastic_features: bool = False, smoke_test: bool = False, + number_of_slurm_nodes: Optional[int] = None, + slurm_node_id_variable: str = "SLURM_GRAPE_ID", verbose: bool = True ) -> pd.DataFrame: """Execute edge prediction evaluation pipeline for all provided models and graphs. @@ -77,7 +80,12 @@ def edge_prediction_evaluation( For instance this may be used for filtering the uncertain edges in graphs such as STRING PPIs. subgraph_of_interest: Optional[Graph] = None - The subgraph of interest to focus the task on. + Optional subgraph where to focus the task. + This is applied to the train and test graph + after the desired holdout schema is applied. + use_subgraph_as_support: bool = False + Whether to use the provided subgraph as support or + to use the train graph (not filtered by the subgraph). number_of_holdouts: int = 10 The number of holdouts to execute. random_state: int = 42 @@ -100,7 +108,7 @@ def edge_prediction_evaluation( in the model performance. enable_cache: bool = False Whether to enable the cache. - precompute_constant_automatic_stocastic_features: bool = False + precompute_constant_stocastic_features: bool = False Whether to precompute once the constant automatic stocastic features before starting the embedding loop. This means that, when left set to false, while the features will be computed @@ -114,6 +122,12 @@ def edge_prediction_evaluation( and therefore use the smoke test configurations for the provided model names and feature names. This parameter will also turn off the cache. + number_of_slurm_nodes: Optional[int] = None + Number of SLURM nodes to consider as available. + This variable is used to parallelize the holdouts accordingly. + slurm_node_id_variable: str = "SLURM_GRAPE_ID" + Name of the system variable to use as SLURM node id. + It must be set in the slurm bash script. verbose: bool = True Whether to show loading bars """ @@ -128,13 +142,16 @@ def edge_prediction_evaluation( library_names=library_names, graph_callback=graph_callback, subgraph_of_interest=subgraph_of_interest, + use_subgraph_as_support=use_subgraph_as_support, number_of_holdouts=number_of_holdouts, random_state=random_state, repositories=repositories, versions=versions, enable_cache=enable_cache, - precompute_constant_automatic_stocastic_features=precompute_constant_automatic_stocastic_features, + precompute_constant_stocastic_features=precompute_constant_stocastic_features, smoke_test=smoke_test, + number_of_slurm_nodes=number_of_slurm_nodes, + slurm_node_id_variable=slurm_node_id_variable, verbose=verbose, validation_sample_only_edges_with_heterogeneous_node_types=validation_sample_only_edges_with_heterogeneous_node_types, validation_unbalance_rates=validation_unbalance_rates, diff --git a/embiggen/edge_prediction/edge_prediction_model.py b/embiggen/edge_prediction/edge_prediction_model.py index 79950a1b..0edf5dda 100644 --- a/embiggen/edge_prediction/edge_prediction_model.py +++ b/embiggen/edge_prediction/edge_prediction_model.py @@ -123,13 +123,15 @@ def __iterate_negative_graphs( ) train_size = ( - train.get_number_of_edges() / (train.get_number_of_edges() + test.get_number_of_edges()) + train.get_number_of_edges() / (train.get_number_of_edges() + + test.get_number_of_edges()) ) return ( sampler_graph.sample_negative_graph( number_of_negative_samples=int( - math.ceil(sampler_graph.get_number_of_edges()*unbalance_rate) + math.ceil(sampler_graph.get_number_of_edges() + * unbalance_rate) ), random_state=random_state*(i+1), sample_only_edges_with_heterogeneous_node_types=validation_sample_only_edges_with_heterogeneous_node_types, @@ -201,6 +203,10 @@ def _evaluate( """Return model evaluation on the provided graphs.""" performance = [] + train_size = ( + train.get_number_of_directed_edges() / graph.get_number_of_directed_edges() + ) + train_predic_proba = self.predict_proba( train, support=support, @@ -265,6 +271,7 @@ def _evaluate( performance.append({ "evaluation_mode": evaluation_mode, + "train_size": train_size, "validation_unbalance_rate": unbalance_rate, "use_scale_free_distribution": use_scale_free_distribution, "validation_sample_only_edges_with_heterogeneous_node_types": validation_sample_only_edges_with_heterogeneous_node_types, @@ -287,8 +294,9 @@ def predict( node_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, node_type_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, edge_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, - return_predictions_dataframe: bool = False - ) -> np.ndarray: + return_predictions_dataframe: bool = False, + return_node_names: bool = True + ) -> Union[np.ndarray, pd.DataFrame]: """Execute predictions on the provided graph. Parameters @@ -309,6 +317,10 @@ def predict( return_predictions_dataframe: bool = False Whether to return a pandas DataFrame, which as indices has the node IDs. By default, a numpy array with the predictions is returned as it weights much less. + return_node_names: bool = True + Whether to return node names when returning the prediction DataFrame. + This value is ignored when the values to be returned the user has not + requested for a prediction dataframe to be returned. """ if edge_features is not None: raise NotImplementedError( @@ -326,8 +338,8 @@ def predict( predictions = pd.DataFrame( { "predictions": predictions, - "sources": graph.get_directed_source_node_ids(), - "destinations": graph.get_directed_destination_node_ids(), + "sources": graph.get_source_names(directed=True) if return_node_names else graph.get_directed_source_node_ids(), + "destinations": graph.get_destination_names(directed=True) if return_node_names else graph.get_directed_destination_node_ids(), }, ) @@ -342,8 +354,9 @@ def predict_bipartite_graph_from_edge_node_ids( node_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, node_type_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, edge_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, - return_predictions_dataframe: bool = False - ) -> np.ndarray: + return_predictions_dataframe: bool = False, + return_node_names: bool = True + ) -> Union[np.ndarray, pd.DataFrame]: """Execute predictions on the provided graph bipartite portion. Parameters @@ -368,6 +381,10 @@ def predict_bipartite_graph_from_edge_node_ids( return_predictions_dataframe: bool = False Whether to return a pandas DataFrame, which as indices has the node IDs. By default, a numpy array with the predictions is returned as it weights much less. + return_node_names: bool = True + Whether to return node names when returning the prediction DataFrame. + This value is ignored when the values to be returned the user has not + requested for a prediction dataframe to be returned. """ return self.predict( graph.build_bipartite_graph_from_edge_node_ids( @@ -379,7 +396,8 @@ def predict_bipartite_graph_from_edge_node_ids( node_features=node_features, node_type_features=node_type_features, edge_features=edge_features, - return_predictions_dataframe=return_predictions_dataframe + return_predictions_dataframe=return_predictions_dataframe, + return_node_names=return_node_names ) def predict_bipartite_graph_from_edge_node_names( @@ -391,8 +409,9 @@ def predict_bipartite_graph_from_edge_node_names( node_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, node_type_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, edge_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, - return_predictions_dataframe: bool = False - ) -> np.ndarray: + return_predictions_dataframe: bool = False, + return_node_names: bool = True + ) -> Union[np.ndarray, pd.DataFrame]: """Execute predictions on the provided graph bipartite portion. Parameters @@ -417,6 +436,10 @@ def predict_bipartite_graph_from_edge_node_names( return_predictions_dataframe: bool = False Whether to return a pandas DataFrame, which as indices has the node IDs. By default, a numpy array with the predictions is returned as it weights much less. + return_node_names: bool = True + Whether to return node names when returning the prediction DataFrame. + This value is ignored when the values to be returned the user has not + requested for a prediction dataframe to be returned. """ return self.predict( graph.build_bipartite_graph_from_edge_node_names( @@ -428,7 +451,8 @@ def predict_bipartite_graph_from_edge_node_names( node_features=node_features, node_type_features=node_type_features, edge_features=edge_features, - return_predictions_dataframe=return_predictions_dataframe + return_predictions_dataframe=return_predictions_dataframe, + return_node_names=return_node_names ) def predict_bipartite_graph_from_edge_node_prefixes( @@ -440,8 +464,9 @@ def predict_bipartite_graph_from_edge_node_prefixes( node_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, node_type_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, edge_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, - return_predictions_dataframe: bool = False - ) -> np.ndarray: + return_predictions_dataframe: bool = False, + return_node_names: bool = True + ) -> Union[np.ndarray, pd.DataFrame]: """Execute predictions on the provided graph bipartite portion. Parameters @@ -466,6 +491,10 @@ def predict_bipartite_graph_from_edge_node_prefixes( return_predictions_dataframe: bool = False Whether to return a pandas DataFrame, which as indices has the node IDs. By default, a numpy array with the predictions is returned as it weights much less. + return_node_names: bool = True + Whether to return node names when returning the prediction DataFrame. + This value is ignored when the values to be returned the user has not + requested for a prediction dataframe to be returned. """ return self.predict( graph.build_bipartite_graph_from_edge_node_prefixes( @@ -477,7 +506,8 @@ def predict_bipartite_graph_from_edge_node_prefixes( node_features=node_features, node_type_features=node_type_features, edge_features=edge_features, - return_predictions_dataframe=return_predictions_dataframe + return_predictions_dataframe=return_predictions_dataframe, + return_node_names=return_node_names ) def predict_bipartite_graph_from_edge_node_types( @@ -489,8 +519,9 @@ def predict_bipartite_graph_from_edge_node_types( node_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, node_type_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, edge_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, - return_predictions_dataframe: bool = False - ) -> np.ndarray: + return_predictions_dataframe: bool = False, + return_node_names: bool = True + ) -> Union[np.ndarray, pd.DataFrame]: """Execute predictions on the provided graph bipartite portion. Parameters @@ -515,6 +546,10 @@ def predict_bipartite_graph_from_edge_node_types( return_predictions_dataframe: bool = False Whether to return a pandas DataFrame, which as indices has the node IDs. By default, a numpy array with the predictions is returned as it weights much less. + return_node_names: bool = True + Whether to return node names when returning the prediction DataFrame. + This value is ignored when the values to be returned the user has not + requested for a prediction dataframe to be returned. """ return self.predict( graph.build_bipartite_graph_from_edge_node_types( @@ -526,7 +561,8 @@ def predict_bipartite_graph_from_edge_node_types( node_features=node_features, node_type_features=node_type_features, edge_features=edge_features, - return_predictions_dataframe=return_predictions_dataframe + return_predictions_dataframe=return_predictions_dataframe, + return_node_names=return_node_names ) def predict_clique_graph_from_node_ids( @@ -537,8 +573,9 @@ def predict_clique_graph_from_node_ids( node_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, node_type_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, edge_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, - return_predictions_dataframe: bool = False - ) -> np.ndarray: + return_predictions_dataframe: bool = False, + return_node_names: bool = True + ) -> Union[np.ndarray, pd.DataFrame]: """Execute predictions on the provided graph bipartite portion. Parameters @@ -561,6 +598,10 @@ def predict_clique_graph_from_node_ids( return_predictions_dataframe: bool = False Whether to return a pandas DataFrame, which as indices has the node IDs. By default, a numpy array with the predictions is returned as it weights much less. + return_node_names: bool = True + Whether to return node names when returning the prediction DataFrame. + This value is ignored when the values to be returned the user has not + requested for a prediction dataframe to be returned. """ return self.predict( graph.build_clique_graph_from_node_ids( @@ -571,7 +612,8 @@ def predict_clique_graph_from_node_ids( node_features=node_features, node_type_features=node_type_features, edge_features=edge_features, - return_predictions_dataframe=return_predictions_dataframe + return_predictions_dataframe=return_predictions_dataframe, + return_node_names=return_node_names ) def predict_clique_graph_from_node_names( @@ -582,8 +624,9 @@ def predict_clique_graph_from_node_names( node_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, node_type_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, edge_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, - return_predictions_dataframe: bool = False - ) -> np.ndarray: + return_predictions_dataframe: bool = False, + return_node_names: bool = True + ) -> Union[np.ndarray, pd.DataFrame]: """Execute predictions on the provided graph bipartite portion. Parameters @@ -606,6 +649,10 @@ def predict_clique_graph_from_node_names( return_predictions_dataframe: bool = False Whether to return a pandas DataFrame, which as indices has the node IDs. By default, a numpy array with the predictions is returned as it weights much less. + return_node_names: bool = True + Whether to return node names when returning the prediction DataFrame. + This value is ignored when the values to be returned the user has not + requested for a prediction dataframe to be returned. """ return self.predict( graph.build_clique_graph_from_node_names( @@ -616,7 +663,8 @@ def predict_clique_graph_from_node_names( node_features=node_features, node_type_features=node_type_features, edge_features=edge_features, - return_predictions_dataframe=return_predictions_dataframe + return_predictions_dataframe=return_predictions_dataframe, + return_node_names=return_node_names ) def predict_clique_graph_from_node_prefixes( @@ -627,8 +675,9 @@ def predict_clique_graph_from_node_prefixes( node_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, node_type_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, edge_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, - return_predictions_dataframe: bool = False - ) -> np.ndarray: + return_predictions_dataframe: bool = False, + return_node_names: bool = True + ) -> Union[np.ndarray, pd.DataFrame]: """Execute predictions on the provided graph bipartite portion. Parameters @@ -651,6 +700,10 @@ def predict_clique_graph_from_node_prefixes( return_predictions_dataframe: bool = False Whether to return a pandas DataFrame, which as indices has the node IDs. By default, a numpy array with the predictions is returned as it weights much less. + return_node_names: bool = True + Whether to return node names when returning the prediction DataFrame. + This value is ignored when the values to be returned the user has not + requested for a prediction dataframe to be returned. """ return self.predict( graph.build_clique_graph_from_node_prefixes( @@ -661,26 +714,28 @@ def predict_clique_graph_from_node_prefixes( node_features=node_features, node_type_features=node_type_features, edge_features=edge_features, - return_predictions_dataframe=return_predictions_dataframe + return_predictions_dataframe=return_predictions_dataframe, + return_node_names=return_node_names ) - def predict_clique_graph_from_node_types( + def predict_clique_graph_from_node_type_names( self, graph: Graph, - node_types: List[str], + node_type_names: List[str], support: Optional[Graph] = None, node_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, node_type_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, edge_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, - return_predictions_dataframe: bool = False - ) -> np.ndarray: + return_predictions_dataframe: bool = False, + return_node_names: bool = True + ) -> Union[np.ndarray, pd.DataFrame]: """Execute predictions on the provided graph bipartite portion. Parameters -------------------- graph: Graph The graph from which to extract the edges. - node_types: List[str] + node_type_names: List[str] The node prefixes of the bipartite graph. support: Optional[Graph] = None The graph describiding the topological structure that @@ -696,17 +751,22 @@ def predict_clique_graph_from_node_types( return_predictions_dataframe: bool = False Whether to return a pandas DataFrame, which as indices has the node IDs. By default, a numpy array with the predictions is returned as it weights much less. + return_node_names: bool = True + Whether to return node names when returning the prediction DataFrame. + This value is ignored when the values to be returned the user has not + requested for a prediction dataframe to be returned. """ return self.predict( - graph.build_clique_graph_from_node_types( - node_types=node_types, + graph.build_clique_graph_from_node_type_names( + node_type_names=node_type_names, directed=True ), support=support, node_features=node_features, node_type_features=node_type_features, edge_features=edge_features, - return_predictions_dataframe=return_predictions_dataframe + return_predictions_dataframe=return_predictions_dataframe, + return_node_names=return_node_names ) def predict_proba( @@ -716,8 +776,9 @@ def predict_proba( node_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, node_type_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, edge_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, - return_predictions_dataframe: bool = False - ) -> np.ndarray: + return_predictions_dataframe: bool = False, + return_node_names: bool = True + ) -> Union[np.ndarray, pd.DataFrame]: """Execute predictions on the provided graph. Parameters @@ -738,6 +799,10 @@ def predict_proba( return_predictions_dataframe: bool = False Whether to return a pandas DataFrame, which as indices has the node IDs. By default, a numpy array with the predictions is returned as it weights much less. + return_node_names: bool = True + Whether to return node names when returning the prediction DataFrame. + This value is ignored when the values to be returned the user has not + requested for a prediction dataframe to be returned. """ if edge_features is not None: raise NotImplementedError( @@ -760,8 +825,8 @@ def predict_proba( predictions = pd.DataFrame( { "predictions": predictions, - "sources": graph.get_directed_source_node_ids(), - "destinations": graph.get_directed_destination_node_ids(), + "sources": graph.get_source_names(directed=True) if return_node_names else graph.get_directed_source_node_ids(), + "destinations": graph.get_destination_names(directed=True) if return_node_names else graph.get_directed_destination_node_ids(), }, ) @@ -776,8 +841,9 @@ def predict_proba_bipartite_graph_from_edge_node_ids( node_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, node_type_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, edge_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, - return_predictions_dataframe: bool = False - ) -> np.ndarray: + return_predictions_dataframe: bool = False, + return_node_names: bool = True + ) -> Union[np.ndarray, pd.DataFrame]: """Execute predictions probabilities on the provided graph bipartite portion. Parameters @@ -802,6 +868,10 @@ def predict_proba_bipartite_graph_from_edge_node_ids( return_predictions_dataframe: bool = False Whether to return a pandas DataFrame, which as indices has the node IDs. By default, a numpy array with the predictions is returned as it weights much less. + return_node_names: bool = True + Whether to return node names when returning the prediction DataFrame. + This value is ignored when the values to be returned the user has not + requested for a prediction dataframe to be returned. """ return self.predict_proba( graph.build_bipartite_graph_from_edge_node_ids( @@ -813,7 +883,8 @@ def predict_proba_bipartite_graph_from_edge_node_ids( node_features=node_features, node_type_features=node_type_features, edge_features=edge_features, - return_predictions_dataframe=return_predictions_dataframe + return_predictions_dataframe=return_predictions_dataframe, + return_node_names=return_node_names ) def predict_proba_bipartite_graph_from_edge_node_names( @@ -825,8 +896,9 @@ def predict_proba_bipartite_graph_from_edge_node_names( node_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, node_type_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, edge_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, - return_predictions_dataframe: bool = False - ) -> np.ndarray: + return_predictions_dataframe: bool = False, + return_node_names: bool = True + ) -> Union[np.ndarray, pd.DataFrame]: """Execute predictions probabilities on the provided graph bipartite portion. Parameters @@ -851,6 +923,10 @@ def predict_proba_bipartite_graph_from_edge_node_names( return_predictions_dataframe: bool = False Whether to return a pandas DataFrame, which as indices has the node IDs. By default, a numpy array with the predictions is returned as it weights much less. + return_node_names: bool = True + Whether to return node names when returning the prediction DataFrame. + This value is ignored when the values to be returned the user has not + requested for a prediction dataframe to be returned. """ return self.predict_proba( graph.build_bipartite_graph_from_edge_node_names( @@ -862,7 +938,8 @@ def predict_proba_bipartite_graph_from_edge_node_names( node_features=node_features, node_type_features=node_type_features, edge_features=edge_features, - return_predictions_dataframe=return_predictions_dataframe + return_predictions_dataframe=return_predictions_dataframe, + return_node_names=return_node_names ) def predict_proba_bipartite_graph_from_edge_node_prefixes( @@ -874,8 +951,9 @@ def predict_proba_bipartite_graph_from_edge_node_prefixes( node_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, node_type_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, edge_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, - return_predictions_dataframe: bool = False - ) -> np.ndarray: + return_predictions_dataframe: bool = False, + return_node_names: bool = True + ) -> Union[np.ndarray, pd.DataFrame]: """Execute predictions probabilities on the provided graph bipartite portion. Parameters @@ -900,6 +978,10 @@ def predict_proba_bipartite_graph_from_edge_node_prefixes( return_predictions_dataframe: bool = False Whether to return a pandas DataFrame, which as indices has the node IDs. By default, a numpy array with the predictions is returned as it weights much less. + return_node_names: bool = True + Whether to return node names when returning the prediction DataFrame. + This value is ignored when the values to be returned the user has not + requested for a prediction dataframe to be returned. """ return self.predict_proba( graph.build_bipartite_graph_from_edge_node_prefixes( @@ -911,7 +993,8 @@ def predict_proba_bipartite_graph_from_edge_node_prefixes( node_features=node_features, node_type_features=node_type_features, edge_features=edge_features, - return_predictions_dataframe=return_predictions_dataframe + return_predictions_dataframe=return_predictions_dataframe, + return_node_names=return_node_names ) def predict_proba_bipartite_graph_from_edge_node_types( @@ -923,8 +1006,9 @@ def predict_proba_bipartite_graph_from_edge_node_types( node_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, node_type_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, edge_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, - return_predictions_dataframe: bool = False - ) -> np.ndarray: + return_predictions_dataframe: bool = False, + return_node_names: bool = True + ) -> Union[np.ndarray, pd.DataFrame]: """Execute predictions probabilities on the provided graph bipartite portion. Parameters @@ -949,6 +1033,10 @@ def predict_proba_bipartite_graph_from_edge_node_types( return_predictions_dataframe: bool = False Whether to return a pandas DataFrame, which as indices has the node IDs. By default, a numpy array with the predictions is returned as it weights much less. + return_node_names: bool = True + Whether to return node names when returning the prediction DataFrame. + This value is ignored when the values to be returned the user has not + requested for a prediction dataframe to be returned. """ return self.predict_proba( graph.build_bipartite_graph_from_edge_node_types( @@ -960,7 +1048,8 @@ def predict_proba_bipartite_graph_from_edge_node_types( node_features=node_features, node_type_features=node_type_features, edge_features=edge_features, - return_predictions_dataframe=return_predictions_dataframe + return_predictions_dataframe=return_predictions_dataframe, + return_node_names=return_node_names ) def predict_proba_clique_graph_from_node_ids( @@ -971,8 +1060,9 @@ def predict_proba_clique_graph_from_node_ids( node_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, node_type_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, edge_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, - return_predictions_dataframe: bool = False - ) -> np.ndarray: + return_predictions_dataframe: bool = False, + return_node_names: bool = True + ) -> Union[np.ndarray, pd.DataFrame]: """Execute predictions probabilities on the provided graph bipartite portion. Parameters @@ -995,6 +1085,10 @@ def predict_proba_clique_graph_from_node_ids( return_predictions_dataframe: bool = False Whether to return a pandas DataFrame, which as indices has the node IDs. By default, a numpy array with the predictions is returned as it weights much less. + return_node_names: bool = True + Whether to return node names when returning the prediction DataFrame. + This value is ignored when the values to be returned the user has not + requested for a prediction dataframe to be returned. """ return self.predict_proba( graph.build_clique_graph_from_node_ids( @@ -1005,7 +1099,8 @@ def predict_proba_clique_graph_from_node_ids( node_features=node_features, node_type_features=node_type_features, edge_features=edge_features, - return_predictions_dataframe=return_predictions_dataframe + return_predictions_dataframe=return_predictions_dataframe, + return_node_names=return_node_names ) def predict_proba_clique_graph_from_node_names( @@ -1016,8 +1111,9 @@ def predict_proba_clique_graph_from_node_names( node_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, node_type_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, edge_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, - return_predictions_dataframe: bool = False - ) -> np.ndarray: + return_predictions_dataframe: bool = False, + return_node_names: bool = True + ) -> Union[np.ndarray, pd.DataFrame]: """Execute predictions probabilities on the provided graph bipartite portion. Parameters @@ -1040,6 +1136,10 @@ def predict_proba_clique_graph_from_node_names( return_predictions_dataframe: bool = False Whether to return a pandas DataFrame, which as indices has the node IDs. By default, a numpy array with the predictions is returned as it weights much less. + return_node_names: bool = True + Whether to return node names when returning the prediction DataFrame. + This value is ignored when the values to be returned the user has not + requested for a prediction dataframe to be returned. """ return self.predict_proba( graph.build_clique_graph_from_node_names( @@ -1050,7 +1150,8 @@ def predict_proba_clique_graph_from_node_names( node_features=node_features, node_type_features=node_type_features, edge_features=edge_features, - return_predictions_dataframe=return_predictions_dataframe + return_predictions_dataframe=return_predictions_dataframe, + return_node_names=return_node_names ) def predict_proba_clique_graph_from_node_prefixes( @@ -1061,8 +1162,9 @@ def predict_proba_clique_graph_from_node_prefixes( node_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, node_type_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, edge_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, - return_predictions_dataframe: bool = False - ) -> np.ndarray: + return_predictions_dataframe: bool = False, + return_node_names: bool = True + ) -> Union[np.ndarray, pd.DataFrame]: """Execute predictions probabilities on the provided graph bipartite portion. Parameters @@ -1085,6 +1187,10 @@ def predict_proba_clique_graph_from_node_prefixes( return_predictions_dataframe: bool = False Whether to return a pandas DataFrame, which as indices has the node IDs. By default, a numpy array with the predictions is returned as it weights much less. + return_node_names: bool = True + Whether to return node names when returning the prediction DataFrame. + This value is ignored when the values to be returned the user has not + requested for a prediction dataframe to be returned. """ return self.predict_proba( graph.build_clique_graph_from_node_prefixes( @@ -1095,26 +1201,28 @@ def predict_proba_clique_graph_from_node_prefixes( node_features=node_features, node_type_features=node_type_features, edge_features=edge_features, - return_predictions_dataframe=return_predictions_dataframe + return_predictions_dataframe=return_predictions_dataframe, + return_node_names=return_node_names ) - def predict_proba_clique_graph_from_node_types( + def predict_proba_clique_graph_from_node_type_names( self, graph: Graph, - node_types: List[str], + node_type_names: List[str], support: Optional[Graph] = None, node_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, node_type_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, edge_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, - return_predictions_dataframe: bool = False - ) -> np.ndarray: + return_predictions_dataframe: bool = False, + return_node_names: bool = True + ) -> Union[np.ndarray, pd.DataFrame]: """Execute predictions probabilities on the provided graph bipartite portion. Parameters -------------------- graph: Graph The graph from which to extract the edges. - node_types: List[str] + node_type_names: List[str] The node prefixes of the bipartite graph. support: Optional[Graph] = None The graph describiding the topological structure that @@ -1130,17 +1238,22 @@ def predict_proba_clique_graph_from_node_types( return_predictions_dataframe: bool = False Whether to return a pandas DataFrame, which as indices has the node IDs. By default, a numpy array with the predictions is returned as it weights much less. + return_node_names: bool = True + Whether to return node names when returning the prediction DataFrame. + This value is ignored when the values to be returned the user has not + requested for a prediction dataframe to be returned. """ return self.predict_proba( - graph.build_clique_graph_from_node_types( - node_types=node_types, + graph.build_clique_graph_from_node_type_names( + node_type_names=node_type_names, directed=True ), support=support, node_features=node_features, node_type_features=node_type_features, edge_features=edge_features, - return_predictions_dataframe=return_predictions_dataframe + return_predictions_dataframe=return_predictions_dataframe, + return_node_names=return_node_names ) def fit( diff --git a/embiggen/edge_prediction/edge_prediction_sklearn/sklearn_edge_prediction_adapter.py b/embiggen/edge_prediction/edge_prediction_sklearn/sklearn_edge_prediction_adapter.py index 0ae0d0ea..0bf07d27 100644 --- a/embiggen/edge_prediction/edge_prediction_sklearn/sklearn_edge_prediction_adapter.py +++ b/embiggen/edge_prediction/edge_prediction_sklearn/sklearn_edge_prediction_adapter.py @@ -4,6 +4,7 @@ from typing import Type, List, Optional, Dict, Any, Union import numpy as np import math +import compress_pickle import copy from ensmallen import Graph from embiggen.sequences.generic_sequences import EdgePredictionSequence @@ -373,3 +374,23 @@ def can_use_edge_types(cls) -> bool: """Returns whether the model can optionally use edge types.""" return False + @classmethod + def load(cls, path: str) -> "Self": + """Load a saved version of the model from the provided path. + + Parameters + ------------------- + path: str + Path from where to load the model. + """ + return compress_pickle.load(path) + + def dump(self, path: str): + """Dump the current model at the provided path. + + Parameters + ------------------- + path: str + Path from where to dump the model. + """ + compress_pickle.dump(self, path) \ No newline at end of file diff --git a/embiggen/embedders/ensmallen_embedders/deepwalk_cbow.py b/embiggen/embedders/ensmallen_embedders/deepwalk_cbow.py index 44d72145..99d4c46b 100644 --- a/embiggen/embedders/ensmallen_embedders/deepwalk_cbow.py +++ b/embiggen/embedders/ensmallen_embedders/deepwalk_cbow.py @@ -25,6 +25,8 @@ def __init__( normalize_learning_rate_by_degree: Optional[bool] = False, use_scale_free_distribution: Optional[bool] = True, random_state: int = 42, + dtype: str = "f32", + ring_bell: bool = False, enable_cache: bool = False ): """Create new abstract DeepWalk method. @@ -73,8 +75,12 @@ def __init__( Divide the learning rate by the degree of the central node. By default false. use_scale_free_distribution: Optional[bool] = True Sample negatives proportionally to their degree. By default true. + dtype: str = "f32" + The data type to be employed, by default f32. random_state: int = 42 The random state to reproduce the training sequence. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -96,7 +102,9 @@ def __init__( stochastic_downsample_by_degree=stochastic_downsample_by_degree, normalize_learning_rate_by_degree=normalize_learning_rate_by_degree, use_scale_free_distribution=use_scale_free_distribution, + dtype=dtype, random_state=random_state, + ring_bell=ring_bell, enable_cache=enable_cache ) diff --git a/embiggen/embedders/ensmallen_embedders/deepwalk_glove.py b/embiggen/embedders/ensmallen_embedders/deepwalk_glove.py index 715bfbd1..f55e6e4f 100644 --- a/embiggen/embedders/ensmallen_embedders/deepwalk_glove.py +++ b/embiggen/embedders/ensmallen_embedders/deepwalk_glove.py @@ -9,18 +9,18 @@ def __init__( self, embedding_size: int = 100, alpha: float = 0.75, - epochs: int = 30, - clipping_value: float = 6.0, - walk_length: int = 128, - iterations: int = 10, + epochs: int = 100, + walk_length: int = 512, window_size: int = 5, max_neighbours: Optional[int] = 100, - learning_rate: float = 0.001, - learning_rate_decay: float = 0.9, + learning_rate: float = 0.05, + learning_rate_decay: float = 0.99, central_nodes_embedding_path: Optional[str] = None, contextual_nodes_embedding_path: Optional[str] = None, normalize_by_degree: bool = False, + dtype: str = "f32", random_state: int = 42, + ring_bell: bool = False, enable_cache: bool = False ): """Create new abstract DeepWalk method. @@ -36,13 +36,8 @@ def __init__( window_size: int = 10 Window size for the local context. On the borders the window size is trimmed. - clipping_value: float = 6.0 - Value at which we clip the dot product, mostly for numerical stability issues. - By default, `6.0`, where the loss is already close to zero. - walk_length: int = 128 + walk_length: int = 512 Maximal length of the walks. - iterations: int = 10 - Number of iterations of the single walks. window_size: int = 5 Window size for the local context. On the borders the window size is trimmed. @@ -50,7 +45,7 @@ def __init__( Number of maximum neighbours to consider when using approximated walks. By default, None, we execute exact random walks. This is mainly useful for graphs containing nodes with high degrees. - learning_rate: float = 0.001 + learning_rate: float = 0.05 The learning rate to use to train the DeepWalk model. By default 0.01. central_nodes_embedding_path: Optional[str] = None Path where to mmap and store the central nodes embedding. @@ -60,13 +55,17 @@ def __init__( Path where to mmap and store the central nodes embedding. This is necessary to embed large graphs whose embedding will not fit into the available main memory. - learning_rate_decay: float = 0.9 + learning_rate_decay: float = 0.99 Factor to reduce the learning rate for at each epoch. By default 0.9. normalize_by_degree: bool = False Whether to normalize the random walk by the node degree of the destination node degrees. + dtype: str = "f32" + The data type to be employed, by default f32. random_state: int = 42 The random state to reproduce the training sequence. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -75,9 +74,8 @@ def __init__( embedding_size=embedding_size, alpha=alpha, epochs=epochs, - clipping_value=clipping_value, walk_length=walk_length, - iterations=iterations, + iterations=1, window_size=window_size, max_neighbours=max_neighbours, learning_rate=learning_rate, @@ -86,6 +84,8 @@ def __init__( contextual_nodes_embedding_path=contextual_nodes_embedding_path, normalize_by_degree=normalize_by_degree, enable_cache=enable_cache, + dtype=dtype, + ring_bell=ring_bell, random_state=random_state ) @@ -96,7 +96,8 @@ def parameters(self) -> Dict[str, Any]: "explore_weight", "change_node_type_weight", "change_edge_type_weight", - "number_of_negative_samples" + "number_of_negative_samples", + "iterations" ] return dict( **{ diff --git a/embiggen/embedders/ensmallen_embedders/deepwalk_skipgram.py b/embiggen/embedders/ensmallen_embedders/deepwalk_skipgram.py index 6d160ea6..b00342ec 100644 --- a/embiggen/embedders/ensmallen_embedders/deepwalk_skipgram.py +++ b/embiggen/embedders/ensmallen_embedders/deepwalk_skipgram.py @@ -25,6 +25,8 @@ def __init__( normalize_learning_rate_by_degree: Optional[bool] = False, use_scale_free_distribution: Optional[bool] = True, random_state: int = 42, + dtype: str = "f32", + ring_bell: bool = False, enable_cache: bool = False ): """Create new abstract DeepWalk method. @@ -73,8 +75,12 @@ def __init__( Divide the learning rate by the degree of the central node. By default false. use_scale_free_distribution: Optional[bool] = True Sample negatives proportionally to their degree. By default true. + dtype: str = "f32" + The data type to be employed, by default f32. random_state: int = 42 The random state to reproduce the training sequence. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -96,7 +102,9 @@ def __init__( stochastic_downsample_by_degree=stochastic_downsample_by_degree, normalize_learning_rate_by_degree=normalize_learning_rate_by_degree, use_scale_free_distribution=use_scale_free_distribution, + dtype=dtype, random_state=random_state, + ring_bell=ring_bell, enable_cache=enable_cache ) diff --git a/embiggen/embedders/ensmallen_embedders/degree_spine.py b/embiggen/embedders/ensmallen_embedders/degree_spine.py index e8bc7995..cd86b344 100644 --- a/embiggen/embedders/ensmallen_embedders/degree_spine.py +++ b/embiggen/embedders/ensmallen_embedders/degree_spine.py @@ -17,6 +17,7 @@ def __init__( maximum_depth: Optional[int] = None, path: Optional[str] = None, verbose: bool = False, + ring_bell: bool = False, enable_cache: bool = False ): """Create new Degree-based SPINE method. @@ -34,6 +35,8 @@ def __init__( This parameter is necessary to embed very large graphs. verbose: bool = False Whether to show loading bars. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -51,19 +54,21 @@ def __init__( super().__init__( embedding_size=embedding_size, + ring_bell=ring_bell, enable_cache=enable_cache ) def parameters(self) -> Dict[str, Any]: """Returns parameters of the model.""" - return { + return dict( **super().parameters(), **dict( dtype=self._dtype, maximum_depth=self._maximum_depth, path=self._path, ) - } + ) + @classmethod def smoke_test_parameters(cls) -> Dict[str, Any]: """Returns parameters for smoke test.""" diff --git a/embiggen/embedders/ensmallen_embedders/degree_wine.py b/embiggen/embedders/ensmallen_embedders/degree_wine.py index 5b848a16..28b03be0 100644 --- a/embiggen/embedders/ensmallen_embedders/degree_wine.py +++ b/embiggen/embedders/ensmallen_embedders/degree_wine.py @@ -14,9 +14,10 @@ def __init__( self, embedding_size: int = 100, dtype: Optional[str] = "u8", - walk_length: int = 2, + window_size: int = 2, path: Optional[str] = None, verbose: bool = False, + ring_bell: bool = False, enable_cache: bool = False ): """Create new Degree-based WINE method. @@ -27,44 +28,49 @@ def __init__( Dimension of the embedding. dtype: Optional[str] = "u8" Dtype to use for the embedding. - walk_length: int = 2 - Length of the random walk. + window_size: int = 2 + Size of the co-occurrence window. + Do note that for `window_size = 2` we will use the Two-Hop WINE version, which is more efficient. By default 2, to capture exclusively the immediate context. path: Optional[str] = None Path where to store the mmap-ed embedding. This parameter is necessary to embed very large graphs. verbose: bool = False Whether to show loading bars. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. """ self._dtype = dtype self._verbose = verbose - self._walk_length = walk_length + self._window_size = window_size self._path = path self._model = models.DegreeWINE( embedding_size=embedding_size, verbose=self._verbose, - walk_length=self._walk_length, + window_size=self._window_size, path=self._path ) super().__init__( embedding_size=embedding_size, + ring_bell=ring_bell, enable_cache=enable_cache ) def parameters(self) -> Dict[str, Any]: """Returns parameters of the model.""" - return { + return dict( **super().parameters(), **dict( dtype=self._dtype, - walk_length=self._walk_length, + window_size=self._window_size, path=self._path, ) - } + ) + @classmethod def smoke_test_parameters(cls) -> Dict[str, Any]: """Returns parameters for smoke test.""" diff --git a/embiggen/embedders/ensmallen_embedders/ensmallen_embedder.py b/embiggen/embedders/ensmallen_embedders/ensmallen_embedder.py index 7232edeb..213c8fb7 100644 --- a/embiggen/embedders/ensmallen_embedders/ensmallen_embedder.py +++ b/embiggen/embedders/ensmallen_embedders/ensmallen_embedder.py @@ -13,6 +13,7 @@ def __init__( self, random_state: Optional[int] = None, embedding_size: int = 100, + ring_bell: bool = False, enable_cache: bool = False ): @@ -24,6 +25,8 @@ def __init__( Random state to reproduce the embeddings. embedding_size: int = 100 Dimension of the embedding. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -31,6 +34,7 @@ def __init__( super().__init__( random_state=random_state, embedding_size=embedding_size, + ring_bell=ring_bell, enable_cache=enable_cache, ) diff --git a/embiggen/embedders/ensmallen_embedders/first_order_line.py b/embiggen/embedders/ensmallen_embedders/first_order_line.py index e5bd91b8..fbafe403 100644 --- a/embiggen/embedders/ensmallen_embedders/first_order_line.py +++ b/embiggen/embedders/ensmallen_embedders/first_order_line.py @@ -17,9 +17,12 @@ def __init__( learning_rate: float = 0.05, learning_rate_decay: float = 0.9, avoid_false_negatives: bool = False, + use_scale_free_distribution: bool = True, node_embedding_path: Optional[str] = None, + dtype: str = "f32", random_state: int = 42, verbose: bool = False, + ring_bell: bool = False, enable_cache: bool = False ): """Create new abstract Node2Vec method. @@ -31,20 +34,26 @@ def __init__( epochs: int = 100 The number of epochs to run the model for, by default 10. learning_rate: float = 0.05 - The learning rate to update the gradient, by default 0.01. + The learning rate to update the gradient, by default 0.05. learning_rate_decay: float = 0.9 Factor to reduce the learning rate for at each epoch. By default 0.9. avoid_false_negatives: bool = False Whether to avoid sampling false negatives. This may cause a slower training. + use_scale_free_distribution: bool = True + Whether to train model using a scale free distribution for the negatives. node_embedding_path: Optional[str] = None Path where to mmap and store the nodes embedding. This is necessary to embed large graphs whose embedding will not fit into the available main memory. + dtype: str = "f32" + The data type to be employed, by default f32. random_state: int = 42 Random state to reproduce the embeddings. verbose: bool = False Whether to show loading bars. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -54,7 +63,9 @@ def __init__( learning_rate=learning_rate, learning_rate_decay=learning_rate_decay, avoid_false_negatives=avoid_false_negatives, + use_scale_free_distribution=use_scale_free_distribution, node_embedding_path=node_embedding_path, + dtype=dtype, verbose=verbose ) @@ -67,6 +78,7 @@ def __init__( super().__init__( embedding_size=embedding_size, enable_cache=enable_cache, + ring_bell=ring_bell, random_state=random_state ) diff --git a/embiggen/embedders/ensmallen_embedders/geometric_laplacian_eigenmaps.py b/embiggen/embedders/ensmallen_embedders/geometric_laplacian_eigenmaps.py index 7862ec90..0b36f344 100644 --- a/embiggen/embedders/ensmallen_embedders/geometric_laplacian_eigenmaps.py +++ b/embiggen/embedders/ensmallen_embedders/geometric_laplacian_eigenmaps.py @@ -15,6 +15,7 @@ class GLEEEnsmallen(EnsmallenEmbedder): def __init__( self, embedding_size: int = 100, + ring_bell: bool = False, enable_cache: bool = False ): """Create new GLEE method. @@ -23,12 +24,15 @@ def __init__( -------------------------- embedding_size: int = 100 Dimension of the embedding. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. """ super().__init__( embedding_size=embedding_size, + ring_bell=ring_bell, enable_cache=enable_cache, ) diff --git a/embiggen/embedders/ensmallen_embedders/hope.py b/embiggen/embedders/ensmallen_embedders/hope.py index ed85a408..38aa8dbc 100644 --- a/embiggen/embedders/ensmallen_embedders/hope.py +++ b/embiggen/embedders/ensmallen_embedders/hope.py @@ -20,6 +20,7 @@ def __init__( metric: str = "Neighbours Intersection size", root_node_name: Optional[str] = None, verbose: bool = False, + ring_bell: bool = False, enable_cache: bool = False ): """Create new HOPE method. @@ -47,6 +48,8 @@ def __init__( the Jaccard index is selected. verbose: bool = False Whether to show loading bars. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -71,6 +74,7 @@ def __init__( super().__init__( embedding_size=embedding_size, + ring_bell=ring_bell, enable_cache=enable_cache ) diff --git a/embiggen/embedders/ensmallen_embedders/laplacian_eigenmaps.py b/embiggen/embedders/ensmallen_embedders/laplacian_eigenmaps.py index 899170b3..416923e4 100644 --- a/embiggen/embedders/ensmallen_embedders/laplacian_eigenmaps.py +++ b/embiggen/embedders/ensmallen_embedders/laplacian_eigenmaps.py @@ -14,6 +14,7 @@ class LaplacianEigenmapsEnsmallen(EnsmallenEmbedder): def __init__( self, embedding_size: int = 100, + ring_bell: bool = False, enable_cache: bool = False ): """Create new Laplacian Eigenmaps method. @@ -22,12 +23,15 @@ def __init__( -------------------------- embedding_size: int = 100 Dimension of the embedding. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. """ super().__init__( embedding_size=embedding_size, + ring_bell=ring_bell, enable_cache=enable_cache, ) diff --git a/embiggen/embedders/ensmallen_embedders/netmf.py b/embiggen/embedders/ensmallen_embedders/netmf.py deleted file mode 100644 index af658a33..00000000 --- a/embiggen/embedders/ensmallen_embedders/netmf.py +++ /dev/null @@ -1,139 +0,0 @@ -"""Module providing NetMF implementation.""" -from typing import Optional, Dict, Any -from ensmallen import Graph -import pandas as pd -import numpy as np -from scipy.sparse import coo_matrix -from sklearn.decomposition import TruncatedSVD -from embiggen.embedders.ensmallen_embedders.ensmallen_embedder import EnsmallenEmbedder -from embiggen.utils import EmbeddingResult - - -class NetMFEnsmallen(EnsmallenEmbedder): - """Class implementing the NetMF algorithm.""" - - def __init__( - self, - embedding_size: int = 100, - walk_length: int = 128, - iterations: int = 10, - window_size: int = 10, - max_neighbours: Optional[int] = 100, - random_state: int = 42, - enable_cache: bool = False - ): - """Create new NetMF method. - - Parameters - -------------------------- - embedding_size: int = 100 - Dimension of the embedding. - walk_length: int = 128 - Maximal length of the walks. - iterations: int = 10 - Number of iterations of the single walks. - window_size: int = 10 - Window size for the local context. - On the borders the window size is trimmed. - max_neighbours: Optional[int] = 100 - Number of maximum neighbours to consider when using approximated walks. - By default, None, we execute exact random walks. - This is mainly useful for graphs containing nodes with high degrees. - random_state: int = 42 - The random state to reproduce the training sequence. - enable_cache: bool = False - Whether to enable the cache, that is to - store the computed embedding. - """ - self._walk_parameters = dict( - walk_length=walk_length, - iterations=iterations, - window_size=window_size, - max_neighbours=max_neighbours, - ) - - super().__init__( - embedding_size=embedding_size, - enable_cache=enable_cache, - random_state=random_state - ) - - @classmethod - def smoke_test_parameters(cls) -> Dict[str, Any]: - """Returns parameters for smoke test.""" - return dict( - **EnsmallenEmbedder.smoke_test_parameters(), - window_size=1, - walk_length=4, - iterations=1, - max_neighbours=10, - ) - - def parameters(self) -> Dict[str, Any]: - """Returns parameters of the model.""" - return dict( - **super().parameters(), - **self._walk_parameters - ) - - def _fit_transform( - self, - graph: Graph, - return_dataframe: bool = True, - ) -> EmbeddingResult: - """Return node embedding.""" - edges, weights = graph.get_log_normalized_cooccurrence_coo_matrix( - **self._walk_parameters - ) - - coo = coo_matrix( - (weights, (edges[:, 0], edges[:, 1])), - shape=( - graph.get_number_of_nodes(), - graph.get_number_of_nodes() - ), - dtype=np.float32 - ) - - model = TruncatedSVD( - n_components=self._embedding_size, - random_state=self._random_state - ) - model.fit(coo) - embedding = model.transform(coo) - - if return_dataframe: - node_names = graph.get_node_names() - embedding = pd.DataFrame( - embedding, - index=node_names - ) - return EmbeddingResult( - embedding_method_name=self.model_name(), - node_embeddings=embedding - ) - - @classmethod - def model_name(cls) -> str: - """Returns name of the model.""" - return "NetMF" - - @classmethod - def can_use_edge_weights(cls) -> bool: - """Returns whether the model can optionally use edge weights.""" - return False - - @classmethod - def can_use_node_types(cls) -> bool: - """Returns whether the model can optionally use node types.""" - return False - - @classmethod - def can_use_edge_types(cls) -> bool: - """Returns whether the model can optionally use edge types.""" - return False - - @classmethod - def is_stocastic(cls) -> bool: - """Returns whether the model is stocastic and has therefore a random state.""" - return True diff --git a/embiggen/embedders/ensmallen_embedders/node2vec.py b/embiggen/embedders/ensmallen_embedders/node2vec.py index 31b64caa..05094f59 100644 --- a/embiggen/embedders/ensmallen_embedders/node2vec.py +++ b/embiggen/embedders/ensmallen_embedders/node2vec.py @@ -29,6 +29,7 @@ def __init__( self, embedding_size: int = 100, random_state: int = 42, + ring_bell: bool = False, enable_cache: bool = False, **model_kwargs: Dict ): @@ -40,6 +41,8 @@ def __init__( Dimension of the embedding. random_state: int = 42 The random state to reproduce the training sequence. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -53,6 +56,7 @@ def __init__( super().__init__( embedding_size=embedding_size, enable_cache=enable_cache, + ring_bell=ring_bell, random_state=random_state ) @@ -64,7 +68,6 @@ def smoke_test_parameters(cls) -> Dict[str, Any]: epochs=1, window_size=1, walk_length=4, - iterations=1, max_neighbours=10, ) diff --git a/embiggen/embedders/ensmallen_embedders/node2vec_cbow.py b/embiggen/embedders/ensmallen_embedders/node2vec_cbow.py index 4fc3203b..04961df9 100644 --- a/embiggen/embedders/ensmallen_embedders/node2vec_cbow.py +++ b/embiggen/embedders/ensmallen_embedders/node2vec_cbow.py @@ -27,6 +27,8 @@ def __init__( normalize_learning_rate_by_degree: Optional[bool] = False, use_scale_free_distribution: Optional[bool] = True, random_state: int = 42, + dtype: str = "f32", + ring_bell: bool = False, enable_cache: bool = False ): """Create new abstract Node2Vec method. @@ -89,8 +91,12 @@ def __init__( Divide the learning rate by the degree of the central node. By default false. use_scale_free_distribution: Optional[bool] = True Sample negatives proportionally to their degree. By default true. + dtype: str = "f32" + The data type to be employed, by default f32. random_state: int = 42 The random state to reproduce the training sequence. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -114,7 +120,9 @@ def __init__( stochastic_downsample_by_degree=stochastic_downsample_by_degree, normalize_learning_rate_by_degree=normalize_learning_rate_by_degree, use_scale_free_distribution=use_scale_free_distribution, + dtype=dtype, random_state=random_state, + ring_bell=ring_bell, enable_cache=enable_cache ) diff --git a/embiggen/embedders/ensmallen_embedders/node2vec_glove.py b/embiggen/embedders/ensmallen_embedders/node2vec_glove.py index 806cb8c6..a80772b7 100644 --- a/embiggen/embedders/ensmallen_embedders/node2vec_glove.py +++ b/embiggen/embedders/ensmallen_embedders/node2vec_glove.py @@ -9,20 +9,20 @@ def __init__( self, embedding_size: int = 100, alpha: float = 0.75, - epochs: int = 30, - clipping_value: float = 6.0, - walk_length: int = 128, - iterations: int = 10, + epochs: int = 100, + walk_length: int = 512, window_size: int = 5, return_weight: float = 0.25, explore_weight: float = 4.0, max_neighbours: Optional[int] = 100, - learning_rate: float = 0.001, + learning_rate: float = 0.05, learning_rate_decay: float = 0.9, central_nodes_embedding_path: Optional[str] = None, contextual_nodes_embedding_path: Optional[str] = None, normalize_by_degree: bool = False, + dtype: str = "f32", random_state: int = 42, + ring_bell: bool = False, enable_cache: bool = False ): """Create new Node2Vec GloVe model. @@ -38,13 +38,8 @@ def __init__( window_size: int = 10 Window size for the local context. On the borders the window size is trimmed. - clipping_value: float = 6.0 - Value at which we clip the dot product, mostly for numerical stability issues. - By default, `6.0`, where the loss is already close to zero. - walk_length: int = 128 + walk_length: int = 512 Maximal length of the walks. - iterations: int = 10 - Number of iterations of the single walks. window_size: int = 5 Window size for the local context. On the borders the window size is trimmed. @@ -81,8 +76,12 @@ def __init__( normalize_by_degree: bool = False Whether to normalize the random walk by the node degree of the destination node degrees. + dtype: str = "f32" + The data type to be employed, by default f32. random_state: int = 42 The random state to reproduce the training sequence. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -91,9 +90,8 @@ def __init__( embedding_size=embedding_size, alpha=alpha, epochs=epochs, - clipping_value=clipping_value, walk_length=walk_length, - iterations=iterations, + iterations=1, window_size=window_size, return_weight=return_weight, explore_weight=explore_weight, @@ -103,7 +101,9 @@ def __init__( central_nodes_embedding_path=central_nodes_embedding_path, contextual_nodes_embedding_path=contextual_nodes_embedding_path, normalize_by_degree=normalize_by_degree, + dtype=dtype, enable_cache=enable_cache, + ring_bell=ring_bell, random_state=random_state ) @@ -112,7 +112,8 @@ def parameters(self) -> Dict[str, Any]: removed = [ "change_node_type_weight", "change_edge_type_weight", - "number_of_negative_samples" + "number_of_negative_samples", + "iterations" ] return dict( **{ diff --git a/embiggen/embedders/ensmallen_embedders/node2vec_skipgram.py b/embiggen/embedders/ensmallen_embedders/node2vec_skipgram.py index b3a1f50e..2dd871e6 100644 --- a/embiggen/embedders/ensmallen_embedders/node2vec_skipgram.py +++ b/embiggen/embedders/ensmallen_embedders/node2vec_skipgram.py @@ -27,6 +27,8 @@ def __init__( normalize_learning_rate_by_degree: Optional[bool] = False, use_scale_free_distribution: Optional[bool] = True, random_state: int = 42, + dtype: str = "f32", + ring_bell: bool = False, enable_cache: bool = False ): """Create new abstract Node2Vec method. @@ -91,8 +93,12 @@ def __init__( Divide the learning rate by the degree of the central node. By default false. use_scale_free_distribution: Optional[bool] = True Sample negatives proportionally to their degree. By default true. + dtype: str = "f32" + The data type to be employed, by default f32. random_state: int = 42 The random state to reproduce the training sequence. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -116,7 +122,9 @@ def __init__( stochastic_downsample_by_degree=stochastic_downsample_by_degree, normalize_learning_rate_by_degree=normalize_learning_rate_by_degree, use_scale_free_distribution=use_scale_free_distribution, + dtype=dtype, random_state=random_state, + ring_bell=ring_bell, enable_cache=enable_cache ) diff --git a/embiggen/embedders/ensmallen_embedders/node_label_spine.py b/embiggen/embedders/ensmallen_embedders/node_label_spine.py index 6bcc7230..ec14e85c 100644 --- a/embiggen/embedders/ensmallen_embedders/node_label_spine.py +++ b/embiggen/embedders/ensmallen_embedders/node_label_spine.py @@ -16,6 +16,7 @@ def __init__( maximum_depth: Optional[int] = None, path: Optional[str] = None, verbose: bool = False, + ring_bell: bool = False, enable_cache: bool = False ): """Create new Node-label-based SPINE method. @@ -31,6 +32,8 @@ def __init__( This parameter is necessary to embed very large graphs. verbose: bool = False Whether to show loading bars. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -46,6 +49,7 @@ def __init__( ) super().__init__( + ring_bell=ring_bell, enable_cache=enable_cache ) diff --git a/embiggen/embedders/ensmallen_embedders/node_label_wine.py b/embiggen/embedders/ensmallen_embedders/node_label_wine.py index 109ac40c..4401c1ed 100644 --- a/embiggen/embedders/ensmallen_embedders/node_label_wine.py +++ b/embiggen/embedders/ensmallen_embedders/node_label_wine.py @@ -13,9 +13,10 @@ class NodeLabelWINE(EnsmallenEmbedder): def __init__( self, dtype: Optional[str] = "u8", - walk_length: Optional[int] = None, + window_size: int = 2, path: Optional[str] = None, verbose: bool = False, + ring_bell: bool = False, enable_cache: bool = False ): """Create new Node-label-based WINE method. @@ -24,29 +25,33 @@ def __init__( -------------------------- dtype: Optional[str] = "u8" Dtype to use for the embedding. - walk_length: int = 2 - Length of the random walk. + window_size: int = 2 + Size of the co-occurrence window. + Do note that for `window_size = 2` we will use the Two-Hop WINE version, which is more efficient. By default 2, to capture exclusively the immediate context. path: Optional[str] = None Path where to store the mmap-ed embedding. This parameter is necessary to embed very large graphs. verbose: bool = False Whether to show loading bars. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. """ self._dtype = dtype self._verbose = verbose - self._walk_length = walk_length + self._window_size = window_size self._path = path self._model = models.NodeLabelWINE( verbose=self._verbose, - walk_length=self._walk_length, + window_size=self._window_size, path=self._path ) super().__init__( + ring_bell=ring_bell, enable_cache=enable_cache ) @@ -60,7 +65,7 @@ def parameters(self) -> Dict[str, Any]: }, **dict( dtype=self._dtype, - walk_length=self._walk_length, + window_size=self._window_size, path=self._path, ) ) diff --git a/embiggen/embedders/ensmallen_embedders/score_spine.py b/embiggen/embedders/ensmallen_embedders/score_spine.py index cf018f30..dd91854c 100644 --- a/embiggen/embedders/ensmallen_embedders/score_spine.py +++ b/embiggen/embedders/ensmallen_embedders/score_spine.py @@ -19,6 +19,7 @@ def __init__( maximum_depth: Optional[int] = None, path: Optional[str] = None, verbose: bool = False, + ring_bell: bool = False, enable_cache: bool = False ): """Create new Score-based SPINE method. @@ -38,6 +39,8 @@ def __init__( This parameter is necessary to embed very large graphs. verbose: bool = False Whether to show loading bars. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -56,19 +59,20 @@ def __init__( super().__init__( embedding_size=embedding_size, + ring_bell=ring_bell, enable_cache=enable_cache ) def parameters(self) -> Dict[str, Any]: """Returns parameters of the model.""" - return { + return dict( **super().parameters(), **dict( dtype=self._dtype, maximum_depth=self._maximum_depth, path=self._path, ) - } + ) @classmethod def smoke_test_parameters(cls) -> Dict[str, Any]: diff --git a/embiggen/embedders/ensmallen_embedders/score_wine.py b/embiggen/embedders/ensmallen_embedders/score_wine.py index 35f9b2a9..ce6af544 100644 --- a/embiggen/embedders/ensmallen_embedders/score_wine.py +++ b/embiggen/embedders/ensmallen_embedders/score_wine.py @@ -16,9 +16,10 @@ def __init__( scores: Optional[np.ndarray] = None, embedding_size: int = 100, dtype: Optional[str] = "u8", - walk_length: Optional[int] = None, + window_size: int = 2, path: Optional[str] = None, verbose: bool = False, + ring_bell: bool = False, enable_cache: bool = False ): """Create new Score-based WINE method. @@ -31,45 +32,49 @@ def __init__( Dimension of the embedding. dtype: Optional[str] = "u8" Dtype to use for the embedding. - walk_length: int = 2 - Length of the random walk. + window_size: int = 2 + Size of the co-occurrence window. + Do note that for `window_size = 2` we will use the Two-Hop WINE version, which is more efficient. By default 2, to capture exclusively the immediate context. path: Optional[str] = None Path where to store the mmap-ed embedding. This parameter is necessary to embed very large graphs. verbose: bool = False Whether to show loading bars. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. """ self._dtype = dtype self._verbose = verbose - self._walk_length = walk_length + self._window_size = window_size self._path = path self._scores = scores self._model = models.ScoreWINE( embedding_size=embedding_size, verbose=self._verbose, - walk_length=self._walk_length, + window_size=self._window_size, path=self._path ) super().__init__( embedding_size=embedding_size, + ring_bell=ring_bell, enable_cache=enable_cache ) def parameters(self) -> Dict[str, Any]: """Returns parameters of the model.""" - return { + return dict( **super().parameters(), **dict( dtype=self._dtype, - walk_length=self._walk_length, + window_size=self._window_size, path=self._path, ) - } + ) @classmethod def smoke_test_parameters(cls) -> Dict[str, Any]: diff --git a/embiggen/embedders/ensmallen_embedders/second_order_line.py b/embiggen/embedders/ensmallen_embedders/second_order_line.py index 19858731..439cffd2 100644 --- a/embiggen/embedders/ensmallen_embedders/second_order_line.py +++ b/embiggen/embedders/ensmallen_embedders/second_order_line.py @@ -15,13 +15,16 @@ def __init__( self, embedding_size: int = 100, epochs: int = 100, - learning_rate: float = 0.01, + learning_rate: float = 0.05, learning_rate_decay: float = 0.9, avoid_false_negatives: bool = False, + use_scale_free_distribution: bool = True, node_embedding_path: Optional[str] = None, contextual_node_embedding_path: Optional[str] = None, + dtype: str = "f32", random_state: int = 42, verbose: bool = False, + ring_bell: bool = False, enable_cache: bool = False ): """Create new abstract Node2Vec method. @@ -32,13 +35,15 @@ def __init__( Dimension of the embedding. epochs: int = 100 The number of epochs to run the model for, by default 10. - learning_rate: float = 0.01 - The learning rate to update the gradient, by default 0.01. + learning_rate: float = 0.05 + The learning rate to update the gradient, by default 0.05. learning_rate_decay: float = 0.9 Factor to reduce the learning rate for at each epoch. By default 0.9. avoid_false_negatives: bool = False Whether to avoid sampling false negatives. This may cause a slower training. + use_scale_free_distribution: bool = True + Whether to train model using a scale free distribution for the negatives. node_embedding_path: Optional[str] = None Path where to mmap and store the nodes embedding. This is necessary to embed large graphs whose embedding will not @@ -47,10 +52,14 @@ def __init__( Path where to mmap and store the contextual nodes embedding. This is necessary to embed large graphs whose embedding will not fit into the available main memory. + dtype: str = "f32" + The data type to be employed, by default f32. random_state: int = 42 Random state to reproduce the embeddings. verbose: bool = False Whether to show loading bars. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -60,8 +69,10 @@ def __init__( learning_rate=learning_rate, learning_rate_decay=learning_rate_decay, avoid_false_negatives=avoid_false_negatives, + use_scale_free_distribution=use_scale_free_distribution, node_embedding_path=node_embedding_path, contextual_node_embedding_path=contextual_node_embedding_path, + dtype=dtype, verbose=verbose ) @@ -74,6 +85,7 @@ def __init__( super().__init__( embedding_size=embedding_size, enable_cache=enable_cache, + ring_bell=ring_bell, random_state=random_state ) diff --git a/embiggen/embedders/ensmallen_embedders/siamese_model.py b/embiggen/embedders/ensmallen_embedders/siamese_model.py index 72615225..0a614d6d 100644 --- a/embiggen/embedders/ensmallen_embedders/siamese_model.py +++ b/embiggen/embedders/ensmallen_embedders/siamese_model.py @@ -12,7 +12,6 @@ class SiameseEnsmallen(EnsmallenEmbedder): models = { "TransE": models.TransE, - "TransH": models.TransH, "Unstructured": models.Unstructured, "Structured Embedding": models.StructuredEmbedding, } @@ -25,8 +24,10 @@ def __init__( learning_rate: float = 0.1, learning_rate_decay: float = 0.9, node_embedding_path: Optional[str] = None, + dtype: str = "f32", random_state: int = 42, verbose: bool = False, + ring_bell: bool = False, enable_cache: bool = False, **paths: Dict[str, str] ): @@ -49,10 +50,14 @@ def __init__( Path where to mmap and store the nodes embedding. This is necessary to embed large graphs whose embedding will not fit into the available main memory. + dtype: str = "f32" + The data type to be employed, by default f32. random_state: int = 42 Random state to reproduce the embeddings. verbose: bool = False Whether to show loading bars. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -64,6 +69,7 @@ def __init__( learning_rate_decay=learning_rate_decay, node_embedding_path=node_embedding_path, verbose=verbose, + dtype=dtype, **paths ) @@ -82,6 +88,7 @@ def __init__( super().__init__( embedding_size=embedding_size, enable_cache=enable_cache, + ring_bell=ring_bell, random_state=random_state ) diff --git a/embiggen/embedders/ensmallen_embedders/sociodim.py b/embiggen/embedders/ensmallen_embedders/sociodim.py index 330153c5..1fd3a5cc 100644 --- a/embiggen/embedders/ensmallen_embedders/sociodim.py +++ b/embiggen/embedders/ensmallen_embedders/sociodim.py @@ -15,6 +15,7 @@ def __init__( self, embedding_size: int = 100, use_sparse_reduce: bool = True, + ring_bell: bool = False, enable_cache: bool = False ): """Create new SocioDim method. @@ -29,6 +30,8 @@ def __init__( For both reduce mechanisms, we are using LAPACK implementations. For some currently unknown reason, their implementation using a sparse reduce, even on dense matrices, yields better results. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -36,6 +39,7 @@ def __init__( self._use_sparse_reduce = use_sparse_reduce super().__init__( embedding_size=embedding_size, + ring_bell=ring_bell, enable_cache=enable_cache, ) diff --git a/embiggen/embedders/ensmallen_embedders/structured_embedding.py b/embiggen/embedders/ensmallen_embedders/structured_embedding.py index ccf0334b..6c876ace 100644 --- a/embiggen/embedders/ensmallen_embedders/structured_embedding.py +++ b/embiggen/embedders/ensmallen_embedders/structured_embedding.py @@ -19,8 +19,10 @@ def __init__( node_embedding_path: Optional[str] = None, source_edge_type_embedding_path: Optional[str] = None, destination_edge_type_embedding_path: Optional[str] = None, + dtype: str = "f32", random_state: int = 42, verbose: bool = False, + ring_bell: bool = False, enable_cache: bool = False ): """Create new Structured Embedding method. @@ -52,10 +54,14 @@ def __init__( Path where to mmap and store the destination edge type embedding. This is necessary to embed large graphs whose embedding will not fit into the available main memory. + dtype: str = "f32" + The data type to be employed, by default f32. random_state: int = 42 Random state to reproduce the embeddings. verbose: bool = False Whether to show loading bars. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -69,8 +75,10 @@ def __init__( node_embedding_path=node_embedding_path, source_edge_type_embedding_path=source_edge_type_embedding_path, destination_edge_type_embedding_path=destination_edge_type_embedding_path, + dtype=dtype, random_state=random_state, verbose=verbose, + ring_bell=ring_bell, enable_cache=enable_cache, ) diff --git a/embiggen/embedders/ensmallen_embedders/transe.py b/embiggen/embedders/ensmallen_embedders/transe.py index 1a135d14..8736037f 100644 --- a/embiggen/embedders/ensmallen_embedders/transe.py +++ b/embiggen/embedders/ensmallen_embedders/transe.py @@ -18,8 +18,10 @@ def __init__( learning_rate_decay: float = 0.9, node_embedding_path: Optional[str] = None, edge_type_embedding_path: Optional[str] = None, + dtype: str = "f32", random_state: int = 42, verbose: bool = False, + ring_bell: bool = False, enable_cache: bool = False ): """Create new TransE method. @@ -47,10 +49,14 @@ def __init__( Path where to mmap and store the edge type embedding. This is necessary to embed large graphs whose embedding will not fit into the available main memory. + dtype: str = "f32" + The data type to be employed, by default f32. random_state: int = 42 Random state to reproduce the embeddings. verbose: bool = False Whether to show loading bars. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -63,8 +69,10 @@ def __init__( learning_rate_decay=learning_rate_decay, node_embedding_path=node_embedding_path, edge_type_embedding_path=edge_type_embedding_path, + dtype=dtype, random_state=random_state, verbose=verbose, + ring_bell=ring_bell, enable_cache=enable_cache, ) diff --git a/embiggen/embedders/ensmallen_embedders/transh.py b/embiggen/embedders/ensmallen_embedders/transh.py deleted file mode 100644 index 2de65834..00000000 --- a/embiggen/embedders/ensmallen_embedders/transh.py +++ /dev/null @@ -1,116 +0,0 @@ -"""Module providing TransH implementation.""" -from typing import Optional -from ensmallen import Graph -import pandas as pd -from embiggen.embedders.ensmallen_embedders.siamese_model import SiameseEnsmallen -from embiggen.utils import EmbeddingResult - - -class TransHEnsmallen(SiameseEnsmallen): - """Class implementing the TransH algorithm.""" - - def __init__( - self, - embedding_size: int = 100, - relu_bias: float = 1.0, - epochs: int = 100, - learning_rate: float = 0.1, - learning_rate_decay: float = 0.9, - node_embedding_path: Optional[str] = None, - mult_edge_type_embedding_path: Optional[str] = None, - bias_edge_type_embedding_path: Optional[str] = None, - random_state: int = 42, - verbose: bool = False, - enable_cache: bool = False - ): - """Create new TransH method. - - Parameters - -------------------------- - model_name: str - The model to instantiate. - embedding_size: int = 100 - Dimension of the embedding. - relu_bias: float = 1.0 - Bias to use for the relu. - In the TransH paper it is called gamma. - epochs: int = 100 - The number of epochs to run the model for, by default 10. - learning_rate: float = 0.05 - The learning rate to update the gradient, by default 0.01. - learning_rate_decay: float = 0.9 - Factor to reduce the learning rate for at each epoch. By default 0.9. - node_embedding_path: Optional[str] = None - Path where to mmap and store the nodes embedding. - This is necessary to embed large graphs whose embedding will not - fit into the available main memory. - mult_edge_type_embedding_path: Optional[str] = None - Path where to mmap and store the multiplicative edge type embedding. - This is necessary to embed large graphs whose embedding will not - fit into the available main memory. - bias_edge_type_embedding_path: Optional[str] = None - Path where to mmap and store the bias edge type embedding. - This is necessary to embed large graphs whose embedding will not - fit into the available main memory. - random_state: int = 42 - Random state to reproduce the embeddings. - verbose: bool = False - Whether to show loading bars. - enable_cache: bool = False - Whether to enable the cache, that is to - store the computed embedding. - """ - super().__init__( - embedding_size=embedding_size, - relu_bias=relu_bias, - epochs=epochs, - learning_rate=learning_rate, - learning_rate_decay=learning_rate_decay, - node_embedding_path=node_embedding_path, - mult_edge_type_embedding_path=mult_edge_type_embedding_path, - bias_edge_type_embedding_path=bias_edge_type_embedding_path, - random_state=random_state, - verbose=verbose, - enable_cache=enable_cache, - ) - - def _fit_transform( - self, - graph: Graph, - return_dataframe: bool = True, - ) -> EmbeddingResult: - """Return node embedding.""" - node_embedding, mult_edge_type_embedding, bias_edge_type_embedding = self._model.fit_transform( - graph, - ) - if return_dataframe: - node_embedding = pd.DataFrame( - node_embedding, - index=graph.get_node_names() - ) - mult_edge_type_embedding = pd.DataFrame( - mult_edge_type_embedding, - index=graph.get_unique_edge_type_names() - ) - bias_edge_type_embedding = pd.DataFrame( - bias_edge_type_embedding, - index=graph.get_unique_edge_type_names() - ) - - return EmbeddingResult( - embedding_method_name=self.model_name(), - node_embeddings=node_embedding, - edge_type_embeddings=[ - mult_edge_type_embedding, - bias_edge_type_embedding - ] - ) - - @classmethod - def model_name(cls) -> str: - """Returns name of the model.""" - return "TransH" - - @classmethod - def requires_edge_types(cls) -> bool: - return True \ No newline at end of file diff --git a/embiggen/embedders/ensmallen_embedders/unstructured.py b/embiggen/embedders/ensmallen_embedders/unstructured.py index edcaf5f0..dd5b2403 100644 --- a/embiggen/embedders/ensmallen_embedders/unstructured.py +++ b/embiggen/embedders/ensmallen_embedders/unstructured.py @@ -17,8 +17,10 @@ def __init__( learning_rate: float = 0.01, learning_rate_decay: float = 0.9, node_embedding_path: Optional[str] = None, + dtype: str = "f32", random_state: int = 42, verbose: bool = False, + ring_bell: bool = False, enable_cache: bool = False ): """Create new abstract Node2Vec method. @@ -42,10 +44,14 @@ def __init__( Path where to mmap and store the nodes embedding. This is necessary to embed large graphs whose embedding will not fit into the available main memory. + dtype: str = "f32" + The data type to be employed, by default f32. random_state: int = 42 Random state to reproduce the embeddings. verbose: bool = False Whether to show loading bars. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -57,8 +63,10 @@ def __init__( learning_rate=learning_rate, learning_rate_decay=learning_rate_decay, node_embedding_path=node_embedding_path, + dtype=dtype, random_state=random_state, verbose=verbose, + ring_bell=ring_bell, enable_cache=enable_cache, ) diff --git a/embiggen/embedders/ensmallen_embedders/walklets.py b/embiggen/embedders/ensmallen_embedders/walklets.py index 0f3a4ddd..fa9e4c14 100644 --- a/embiggen/embedders/ensmallen_embedders/walklets.py +++ b/embiggen/embedders/ensmallen_embedders/walklets.py @@ -29,6 +29,8 @@ def __init__( normalize_learning_rate_by_degree: Optional[bool] = False, use_scale_free_distribution: Optional[bool] = True, random_state: int = 42, + dtype: str = "f32", + ring_bell: bool = False, enable_cache: bool = False ): """Create new abstract Node2Vec method. @@ -97,8 +99,12 @@ def __init__( Divide the learning rate by the degree of the central node. By default false. use_scale_free_distribution: Optional[bool] = True Sample negatives proportionally to their degree. By default true. + dtype: str = "f32" + The data type to be employed, by default f32. random_state: int = 42 The random state to reproduce the training sequence. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -123,7 +129,9 @@ def __init__( stochastic_downsample_by_degree=stochastic_downsample_by_degree, normalize_learning_rate_by_degree=normalize_learning_rate_by_degree, use_scale_free_distribution=use_scale_free_distribution, + dtype=dtype, random_state=random_state, + ring_bell=ring_bell, enable_cache=enable_cache ) diff --git a/embiggen/embedders/ensmallen_embedders/walklets_cbow.py b/embiggen/embedders/ensmallen_embedders/walklets_cbow.py index 99d73975..066667bf 100644 --- a/embiggen/embedders/ensmallen_embedders/walklets_cbow.py +++ b/embiggen/embedders/ensmallen_embedders/walklets_cbow.py @@ -27,6 +27,8 @@ def __init__( normalize_learning_rate_by_degree: Optional[bool] = False, use_scale_free_distribution: Optional[bool] = True, random_state: int = 42, + dtype: str = "f32", + ring_bell: bool = False, enable_cache: bool = False ): """Create new abstract Node2Vec method. @@ -93,8 +95,12 @@ def __init__( Divide the learning rate by the degree of the central node. By default false. use_scale_free_distribution: Optional[bool] = True Sample negatives proportionally to their degree. By default true. + dtype: str = "f32" + The data type to be employed, by default f32. random_state: int = 42 The random state to reproduce the training sequence. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -118,7 +124,9 @@ def __init__( stochastic_downsample_by_degree=stochastic_downsample_by_degree, normalize_learning_rate_by_degree=normalize_learning_rate_by_degree, use_scale_free_distribution=use_scale_free_distribution, + dtype=dtype, random_state=random_state, + ring_bell=ring_bell, enable_cache=enable_cache ) diff --git a/embiggen/embedders/ensmallen_embedders/walklets_glove.py b/embiggen/embedders/ensmallen_embedders/walklets_glove.py index 64fb7f4b..ef840f99 100644 --- a/embiggen/embedders/ensmallen_embedders/walklets_glove.py +++ b/embiggen/embedders/ensmallen_embedders/walklets_glove.py @@ -9,14 +9,13 @@ class WalkletsGloVeEnsmallen(WalkletsEnsmallen): def __init__( self, embedding_size: int = 100, - epochs: int = 30, - walk_length: int = 128, - iterations: int = 10, + epochs: int = 100, + walk_length: int = 512, window_size: int = 4, return_weight: float = 1.0, explore_weight: float = 1.0, max_neighbours: Optional[int] = 100, - learning_rate: float = 0.001, + learning_rate: float = 0.05, learning_rate_decay: float = 0.9, central_nodes_embedding_path: Optional[str] = None, contextual_nodes_embedding_path: Optional[str] = None, @@ -26,6 +25,8 @@ def __init__( normalize_learning_rate_by_degree: Optional[bool] = False, use_scale_free_distribution: Optional[bool] = True, random_state: int = 42, + dtype: str = "f32", + ring_bell: bool = False, enable_cache: bool = False ): """Create new abstract Node2Vec method. @@ -34,12 +35,10 @@ def __init__( -------------------------- embedding_size: int = 100 Dimension of the embedding. - epochs: int = 30 + epochs: int = 100 Number of epochs to train the model for. walk_length: int = 128 Maximal length of the walks. - iterations: int = 10 - Number of iterations of the single walks. window_size: int = 4 Window size for the local context. On the borders the window size is trimmed. @@ -88,8 +87,12 @@ def __init__( Divide the learning rate by the degree of the central node. By default false. use_scale_free_distribution: Optional[bool] = True Sample negatives proportionally to their degree. By default true. + dtype: str = "f32" + The data type to be employed, by default f32. random_state: int = 42 The random state to reproduce the training sequence. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -99,7 +102,7 @@ def __init__( epochs=epochs, alpha=alpha, walk_length=walk_length, - iterations=iterations, + iterations=1, window_size=window_size, return_weight=return_weight, explore_weight=explore_weight, @@ -112,7 +115,9 @@ def __init__( stochastic_downsample_by_degree=stochastic_downsample_by_degree, normalize_learning_rate_by_degree=normalize_learning_rate_by_degree, use_scale_free_distribution=use_scale_free_distribution, + dtype=dtype, random_state=random_state, + ring_bell=ring_bell, enable_cache=enable_cache ) @@ -125,7 +130,8 @@ def parameters(self) -> Dict[str, Any]: """Returns parameters for smoke test.""" removed = [ "number_of_negative_samples", - "clipping_value" + "clipping_value", + "iterations" ] return dict( **{ diff --git a/embiggen/embedders/ensmallen_embedders/walklets_skipgram.py b/embiggen/embedders/ensmallen_embedders/walklets_skipgram.py index 73ce0c94..8f741780 100644 --- a/embiggen/embedders/ensmallen_embedders/walklets_skipgram.py +++ b/embiggen/embedders/ensmallen_embedders/walklets_skipgram.py @@ -27,6 +27,8 @@ def __init__( normalize_learning_rate_by_degree: Optional[bool] = False, use_scale_free_distribution: Optional[bool] = True, random_state: int = 42, + dtype: str = "f32", + ring_bell: bool = False, enable_cache: bool = False ): """Create new abstract Node2Vec method. @@ -93,8 +95,12 @@ def __init__( Divide the learning rate by the degree of the central node. By default false. use_scale_free_distribution: Optional[bool] = True Sample negatives proportionally to their degree. By default true. + dtype: str = "f32" + The data type to be employed, by default f32. random_state: int = 42 The random state to reproduce the training sequence. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -118,7 +124,9 @@ def __init__( stochastic_downsample_by_degree=stochastic_downsample_by_degree, normalize_learning_rate_by_degree=normalize_learning_rate_by_degree, use_scale_free_distribution=use_scale_free_distribution, + dtype=dtype, random_state=random_state, + ring_bell=ring_bell, enable_cache=enable_cache ) diff --git a/embiggen/embedders/ensmallen_embedders/weighted_spine.py b/embiggen/embedders/ensmallen_embedders/weighted_spine.py index 8cb06df0..b428a932 100644 --- a/embiggen/embedders/ensmallen_embedders/weighted_spine.py +++ b/embiggen/embedders/ensmallen_embedders/weighted_spine.py @@ -15,6 +15,7 @@ def __init__( embedding_size: int = 100, use_edge_weights_as_probabilities: bool = False, verbose: bool = False, + ring_bell: bool = False, enable_cache: bool = False ): """Create new abstract Node2Vec method. @@ -27,6 +28,8 @@ def __init__( Whether to treat the weights as probabilities. verbose: bool = False Whether to show loading bars. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -39,6 +42,7 @@ def __init__( super().__init__( embedding_size=embedding_size, + ring_bell=ring_bell, enable_cache=enable_cache ) diff --git a/embiggen/embedders/graph_embedding_pipeline.py b/embiggen/embedders/graph_embedding_pipeline.py index 34713828..53a9489a 100644 --- a/embiggen/embedders/graph_embedding_pipeline.py +++ b/embiggen/embedders/graph_embedding_pipeline.py @@ -34,7 +34,8 @@ def embed_graph( Graph version to retrieve. library_name: Optional[str] = None The library from where to retrieve the embedding model. - enable_cache: bool = False + ring_bell: bool = False, + enable_cache: bool = False Whether to enable the cache. smoke_test: bool = False Whether this run should be considered a smoke test diff --git a/embiggen/embedders/karateclub_embedders/boostne.py b/embiggen/embedders/karateclub_embedders/boostne.py index 461fee04..863b8cfe 100644 --- a/embiggen/embedders/karateclub_embedders/boostne.py +++ b/embiggen/embedders/karateclub_embedders/boostne.py @@ -13,6 +13,7 @@ def __init__( order: int = 2, alpha: float = 0.01, random_state: int = 42, + ring_bell: bool = False, enable_cache: bool = False ): """Return a new BoostNE embedding model. @@ -30,6 +31,8 @@ def __init__( random_state: int = 42 Random state to use for the stocastic portions of the embedding algorithm. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -40,6 +43,7 @@ def __init__( super().__init__( embedding_size=embedding_size, enable_cache=enable_cache, + ring_bell=ring_bell, random_state=random_state ) diff --git a/embiggen/embedders/karateclub_embedders/deep_walk.py b/embiggen/embedders/karateclub_embedders/deep_walk.py index db084890..17d14ba3 100644 --- a/embiggen/embedders/karateclub_embedders/deep_walk.py +++ b/embiggen/embedders/karateclub_embedders/deep_walk.py @@ -17,6 +17,7 @@ def __init__( learning_rate: float = 0.05, min_count: int = 1, random_state: int = 42, + ring_bell: bool = False, enable_cache: bool = False ): """Return a new DeepWalk embedding model. @@ -40,6 +41,8 @@ def __init__( random_state: int = 42 Random state to use for the stocastic portions of the embedding algorithm. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -54,6 +57,7 @@ def __init__( super().__init__( embedding_size=embedding_size, enable_cache=enable_cache, + ring_bell=ring_bell, random_state=random_state ) diff --git a/embiggen/embedders/karateclub_embedders/geometric_laplacian_eigenmaps.py b/embiggen/embedders/karateclub_embedders/geometric_laplacian_eigenmaps.py index eb6982a3..e55a0770 100644 --- a/embiggen/embedders/karateclub_embedders/geometric_laplacian_eigenmaps.py +++ b/embiggen/embedders/karateclub_embedders/geometric_laplacian_eigenmaps.py @@ -10,6 +10,7 @@ def __init__( self, embedding_size: int = 100, random_state: int = 42, + ring_bell: bool = False, enable_cache: bool = False ): """Return a new GLEE embedding model. @@ -21,6 +22,8 @@ def __init__( random_state: int = 42 Random state to use for the stocastic portions of the embedding algorithm. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -28,6 +31,7 @@ def __init__( super().__init__( embedding_size=embedding_size, enable_cache=enable_cache, + ring_bell=ring_bell, random_state=random_state ) diff --git a/embiggen/embedders/karateclub_embedders/grarep.py b/embiggen/embedders/karateclub_embedders/grarep.py index 1c67feeb..c0faa166 100644 --- a/embiggen/embedders/karateclub_embedders/grarep.py +++ b/embiggen/embedders/karateclub_embedders/grarep.py @@ -12,6 +12,7 @@ def __init__( iteration: int = 10, order: int = 5, random_state: int = 42, + ring_bell: bool = False, enable_cache: bool = False ): """Return a new GraRep embedding model. @@ -27,6 +28,8 @@ def __init__( random_state: int = 42 Random state to use for the stocastic portions of the embedding algorithm. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -36,6 +39,7 @@ def __init__( super().__init__( embedding_size=embedding_size, enable_cache=enable_cache, + ring_bell=ring_bell, random_state=random_state ) diff --git a/embiggen/embedders/karateclub_embedders/hope.py b/embiggen/embedders/karateclub_embedders/hope.py index ef74dad4..c41b4178 100644 --- a/embiggen/embedders/karateclub_embedders/hope.py +++ b/embiggen/embedders/karateclub_embedders/hope.py @@ -10,6 +10,7 @@ def __init__( self, embedding_size: int = 100, random_state: int = 42, + ring_bell: bool = False, enable_cache: bool = False ): """Return a new HOPE embedding model. @@ -21,6 +22,8 @@ def __init__( random_state: int = 42 Random state to use for the stocastic portions of the embedding algorithm. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -29,6 +32,7 @@ def __init__( super().__init__( embedding_size=embedding_size, enable_cache=enable_cache, + ring_bell=ring_bell, random_state=random_state ) diff --git a/embiggen/embedders/karateclub_embedders/laplacian_eigenmaps.py b/embiggen/embedders/karateclub_embedders/laplacian_eigenmaps.py index e360ee8d..b1fc3233 100644 --- a/embiggen/embedders/karateclub_embedders/laplacian_eigenmaps.py +++ b/embiggen/embedders/karateclub_embedders/laplacian_eigenmaps.py @@ -10,6 +10,7 @@ def __init__( self, embedding_size: int = 100, random_state: int = 42, + ring_bell: bool = False, enable_cache: bool = False ): """Return a new LaplacianEigenmaps embedding model. @@ -21,6 +22,8 @@ def __init__( random_state: int = 42 Random state to use for the stocastic portions of the embedding algorithm. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -28,6 +31,7 @@ def __init__( super().__init__( embedding_size=embedding_size, enable_cache=enable_cache, + ring_bell=ring_bell, random_state=random_state ) diff --git a/embiggen/embedders/karateclub_embedders/netmf.py b/embiggen/embedders/karateclub_embedders/netmf.py index 86ae41c8..6afcff3e 100644 --- a/embiggen/embedders/karateclub_embedders/netmf.py +++ b/embiggen/embedders/karateclub_embedders/netmf.py @@ -13,6 +13,7 @@ def __init__( order: int = 2, negative_samples: int = 1, random_state: int = 42, + ring_bell: bool = False, enable_cache: bool = False ): """Return a new NetMF embedding model. @@ -30,6 +31,8 @@ def __init__( random_state: int = 42 Random state to use for the stocastic portions of the embedding algorithm. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -40,6 +43,7 @@ def __init__( super().__init__( embedding_size=embedding_size, enable_cache=enable_cache, + ring_bell=ring_bell, random_state=random_state ) diff --git a/embiggen/embedders/karateclub_embedders/nmfadmm.py b/embiggen/embedders/karateclub_embedders/nmfadmm.py index 8026e687..80944645 100644 --- a/embiggen/embedders/karateclub_embedders/nmfadmm.py +++ b/embiggen/embedders/karateclub_embedders/nmfadmm.py @@ -12,6 +12,7 @@ def __init__( iterations: int = 100, rho: float = 1.0, random_state: int = 42, + ring_bell: bool = False, enable_cache: bool = False ): """Return a new NMFADMM embedding model. @@ -27,6 +28,8 @@ def __init__( random_state: int = 42 Random state to use for the stocastic portions of the embedding algorithm. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -37,6 +40,7 @@ def __init__( super().__init__( embedding_size=embedding_size, enable_cache=enable_cache, + ring_bell=ring_bell, random_state=random_state ) diff --git a/embiggen/embedders/karateclub_embedders/node_sketch.py b/embiggen/embedders/karateclub_embedders/node_sketch.py index 0829eea7..fb4b27ce 100644 --- a/embiggen/embedders/karateclub_embedders/node_sketch.py +++ b/embiggen/embedders/karateclub_embedders/node_sketch.py @@ -12,6 +12,7 @@ def __init__( iterations: int = 10, decay: float = 0.01, random_state: int = 42, + ring_bell: bool = False, enable_cache: bool = False ): """Return a new NodeSketch embedding model. @@ -27,6 +28,8 @@ def __init__( random_state: int = 42 Random state to use for the stocastic portions of the embedding algorithm. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -37,6 +40,7 @@ def __init__( super().__init__( embedding_size=embedding_size, enable_cache=enable_cache, + ring_bell=ring_bell, random_state=random_state ) diff --git a/embiggen/embedders/karateclub_embedders/randne.py b/embiggen/embedders/karateclub_embedders/randne.py index bee120c0..84e027fc 100644 --- a/embiggen/embedders/karateclub_embedders/randne.py +++ b/embiggen/embedders/karateclub_embedders/randne.py @@ -11,6 +11,7 @@ def __init__( embedding_size: int = 100, alphas: Union[List[float], Tuple[float]] = (0.5, 0.5), random_state: int = 42, + ring_bell: bool = False, enable_cache: bool = False ): """Return a new RandNE embedding model. @@ -24,6 +25,8 @@ def __init__( random_state: int = 42 Random state to use for the stocastic portions of the embedding algorithm. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -32,6 +35,7 @@ def __init__( super().__init__( embedding_size=embedding_size, enable_cache=enable_cache, + ring_bell=ring_bell, random_state=random_state ) diff --git a/embiggen/embedders/karateclub_embedders/role2vec.py b/embiggen/embedders/karateclub_embedders/role2vec.py index 3f6ad6a4..1a925f4a 100644 --- a/embiggen/embedders/karateclub_embedders/role2vec.py +++ b/embiggen/embedders/karateclub_embedders/role2vec.py @@ -20,6 +20,7 @@ def __init__( weisfeiler_lehman_hashing_iterations: int = 2, erase_base_features: bool = False, random_state: int = 42, + ring_bell: bool = False, enable_cache: bool = False ): """Return a new Role2Vec embedding model. @@ -49,6 +50,8 @@ def __init__( random_state: int = 42 Random state to use for the stocastic portions of the embedding algorithm. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -66,6 +69,7 @@ def __init__( super().__init__( embedding_size=embedding_size, enable_cache=enable_cache, + ring_bell=ring_bell, random_state=random_state ) diff --git a/embiggen/embedders/karateclub_embedders/skipgram.py b/embiggen/embedders/karateclub_embedders/skipgram.py index bb9cb317..443aa78b 100644 --- a/embiggen/embedders/karateclub_embedders/skipgram.py +++ b/embiggen/embedders/karateclub_embedders/skipgram.py @@ -19,6 +19,7 @@ def __init__( learning_rate: float = 0.05, min_count: int = 1, random_state: int = 42, + ring_bell: bool = False, enable_cache: bool = False ): """Return a new Node2Vec CBOW embedding model. @@ -54,6 +55,8 @@ def __init__( random_state: int = 42 Random state to use for the stocastic portions of the embedding algorithm. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -70,6 +73,7 @@ def __init__( super().__init__( embedding_size=embedding_size, enable_cache=enable_cache, + ring_bell=ring_bell, random_state=random_state ) diff --git a/embiggen/embedders/karateclub_embedders/sociodim.py b/embiggen/embedders/karateclub_embedders/sociodim.py index 4ccac64f..e8b5777a 100644 --- a/embiggen/embedders/karateclub_embedders/sociodim.py +++ b/embiggen/embedders/karateclub_embedders/sociodim.py @@ -9,6 +9,7 @@ def __init__( self, embedding_size: int = 100, random_state: int = 42, + ring_bell: bool = False, enable_cache: bool = False ): """Return a new SocioDim embedding model. @@ -20,6 +21,8 @@ def __init__( random_state: int = 42 Random state to use for the stocastic portions of the embedding algorithm. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -27,6 +30,7 @@ def __init__( super().__init__( embedding_size=embedding_size, enable_cache=enable_cache, + ring_bell=ring_bell, random_state=random_state ) diff --git a/embiggen/embedders/karateclub_embedders/walklets.py b/embiggen/embedders/karateclub_embedders/walklets.py index 55b681ef..d56ba3ec 100644 --- a/embiggen/embedders/karateclub_embedders/walklets.py +++ b/embiggen/embedders/karateclub_embedders/walklets.py @@ -17,6 +17,7 @@ def __init__( learning_rate: float = 0.05, min_count: int = 1, random_state: int = 42, + ring_bell: bool = False, enable_cache: bool = False ): """Return a new Walklets embedding model. @@ -40,6 +41,8 @@ def __init__( random_state: int = 42 Random state to use for the stocastic portions of the embedding algorithm. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -54,6 +57,7 @@ def __init__( super().__init__( embedding_size=embedding_size // self._window_size, enable_cache=enable_cache, + ring_bell=ring_bell, random_state=random_state ) diff --git a/embiggen/embedders/pykeen_embedders/__init__.py b/embiggen/embedders/pykeen_embedders/__init__.py index 47705d01..affc9d1d 100644 --- a/embiggen/embedders/pykeen_embedders/__init__.py +++ b/embiggen/embedders/pykeen_embedders/__init__.py @@ -1,8 +1,8 @@ -"""Module with graph embedding models based on PyKeen.""" +"""Module with graph embedding models based on PyKEEN.""" from embiggen.utils.abstract_models import build_init, AbstractEmbeddingModel build_init( module_library_names=["torch", "pykeen"], - formatted_library_name="PyKeen", + formatted_library_name="PyKEEN", expected_parent_class=AbstractEmbeddingModel ) \ No newline at end of file diff --git a/embiggen/embedders/pykeen_embedders/auto_sf.py b/embiggen/embedders/pykeen_embedders/auto_sf.py index eabe63f8..d9abdf8d 100644 --- a/embiggen/embedders/pykeen_embedders/auto_sf.py +++ b/embiggen/embedders/pykeen_embedders/auto_sf.py @@ -1,12 +1,12 @@ -"""Submodule providing wrapper for PyKeen's AutoSF model.""" +"""Submodule providing wrapper for PyKEEN's AutoSF model.""" from typing import Union, Type, Dict, Any from pykeen.training import TrainingLoop from pykeen.models import AutoSF -from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKeen +from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKEEN from pykeen.triples import CoreTriplesFactory -class AutoSFPyKeen(EntityRelationEmbeddingModelPyKeen): +class AutoSFPyKEEN(EntityRelationEmbeddingModelPyKEEN): def __init__( self, @@ -18,14 +18,15 @@ def __init__( ] = "Stochastic Local Closed World Assumption", verbose: bool = False, random_state: int = 42, + ring_bell: bool = False, enable_cache: bool = False ): - """Create new PyKeen AutoSF model. + """Create new PyKEEN AutoSF model. Details ------------------------- This is a wrapper of the AutoSF implementation from the - PyKeen library. Please refer to the PyKeen library documentation + PyKEEN library. Please refer to the PyKEEN library documentation for details and posssible errors regarding this model. Parameters @@ -51,6 +52,8 @@ def __init__( Whether to show loading bars. random_state: int = 42 Random seed to use while training the model + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -63,6 +66,7 @@ def __init__( training_loop=training_loop, verbose=verbose, random_state=random_state, + ring_bell=ring_bell, enable_cache=enable_cache ) diff --git a/embiggen/embedders/pykeen_embedders/boxe.py b/embiggen/embedders/pykeen_embedders/boxe.py index babee384..daf0745b 100644 --- a/embiggen/embedders/pykeen_embedders/boxe.py +++ b/embiggen/embedders/pykeen_embedders/boxe.py @@ -1,12 +1,12 @@ -"""Submodule providing wrapper for PyKeen's BoxE model.""" +"""Submodule providing wrapper for PyKEEN's BoxE model.""" from typing import Union, Type, Dict, Any, Optional from pykeen.training import TrainingLoop from pykeen.models import BoxE -from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKeen +from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKEEN from pykeen.triples import CoreTriplesFactory -class BoxEPyKeen(EntityRelationEmbeddingModelPyKeen): +class BoxEPyKEEN(EntityRelationEmbeddingModelPyKEEN): def __init__( self, @@ -20,14 +20,15 @@ def __init__( ] = "Stochastic Local Closed World Assumption", verbose: bool = False, random_state: int = 42, + ring_bell: bool = False, enable_cache: bool = False ): - """Create new PyKeen BoxE model. + """Create new PyKEEN BoxE model. Details ------------------------- This is a wrapper of the BoxE implementation from the - PyKeen library. Please refer to the PyKeen library documentation + PyKEEN library. Please refer to the PyKEEN library documentation for details and posssible errors regarding this model. Parameters @@ -59,6 +60,8 @@ def __init__( Whether to show loading bars. random_state: int = 42 Random seed to use while training the model + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -73,6 +76,7 @@ def __init__( training_loop=training_loop, verbose=verbose, random_state=random_state, + ring_bell=ring_bell, enable_cache=enable_cache ) diff --git a/embiggen/embedders/pykeen_embedders/complex.py b/embiggen/embedders/pykeen_embedders/complex.py index 5258ff2e..e9ec1297 100644 --- a/embiggen/embedders/pykeen_embedders/complex.py +++ b/embiggen/embedders/pykeen_embedders/complex.py @@ -1,12 +1,12 @@ -"""Submodule providing wrapper for PyKeen's ComplEx model.""" +"""Submodule providing wrapper for PyKEEN's ComplEx model.""" from typing import Union, Type, Dict, Any, Optional from pykeen.training import TrainingLoop from pykeen.models import ComplEx -from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKeen +from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKEEN from pykeen.triples import CoreTriplesFactory -class ComplExPyKeen(EntityRelationEmbeddingModelPyKeen): +class ComplExPyKEEN(EntityRelationEmbeddingModelPyKEEN): @classmethod def model_name(cls) -> str: diff --git a/embiggen/embedders/pykeen_embedders/conve.py b/embiggen/embedders/pykeen_embedders/conve.py index e6f389d7..2c21f111 100644 --- a/embiggen/embedders/pykeen_embedders/conve.py +++ b/embiggen/embedders/pykeen_embedders/conve.py @@ -1,12 +1,12 @@ -"""Submodule providing wrapper for PyKeen's ConvE model.""" +"""Submodule providing wrapper for PyKEEN's ConvE model.""" from typing import Union, Type, Dict, Any, Optional from pykeen.training import TrainingLoop from pykeen.models import ConvE -from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKeen +from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKEEN from pykeen.triples import CoreTriplesFactory -class ConvEPyKeen(EntityRelationEmbeddingModelPyKeen): +class ConvEPyKEEN(EntityRelationEmbeddingModelPyKEEN): def __init__( self, @@ -27,14 +27,15 @@ def __init__( ] = "Stochastic Local Closed World Assumption", verbose: bool = False, random_state: int = 42, + ring_bell: bool = False, enable_cache: bool = False ): - """Create new PyKeen ConvE model. + """Create new PyKEEN ConvE model. Details ------------------------- This is a wrapper of the ConvE implementation from the - PyKeen library. Please refer to the PyKeen library documentation + PyKEEN library. Please refer to the PyKEEN library documentation for details and posssible errors regarding this model. Parameters @@ -81,6 +82,8 @@ def __init__( Whether to show loading bars. random_state: int = 42 Random seed to use while training the model + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -102,6 +105,7 @@ def __init__( training_loop=training_loop, verbose=verbose, random_state=random_state, + ring_bell=ring_bell, enable_cache=enable_cache ) @@ -153,3 +157,8 @@ def _build_model( apply_batch_normalization=self._apply_batch_normalization, random_seed=self._random_state ) + + @classmethod + def _create_inverse_triples(cls) -> bool: + """Returns whether the class is expected to create inverse triples.""" + return True \ No newline at end of file diff --git a/embiggen/embedders/pykeen_embedders/crosse.py b/embiggen/embedders/pykeen_embedders/crosse.py index 8aad0b05..a92ebb15 100644 --- a/embiggen/embedders/pykeen_embedders/crosse.py +++ b/embiggen/embedders/pykeen_embedders/crosse.py @@ -1,12 +1,12 @@ -"""Submodule providing wrapper for PyKeen's CrossE model.""" +"""Submodule providing wrapper for PyKEEN's CrossE model.""" from typing import Union, Type, Dict, Any, Optional from pykeen.training import TrainingLoop from pykeen.models import CrossE -from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKeen +from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKEEN from pykeen.triples import CoreTriplesFactory -class CrossEPyKeen(EntityRelationEmbeddingModelPyKeen): +class CrossEPyKEEN(EntityRelationEmbeddingModelPyKEEN): def __init__( self, @@ -18,14 +18,15 @@ def __init__( ] = "Stochastic Local Closed World Assumption", verbose: bool = False, random_state: int = 42, + ring_bell: bool = False, enable_cache: bool = False ): - """Create new PyKeen CrossE model. + """Create new PyKEEN CrossE model. Details ------------------------- This is a wrapper of the CrossE implementation from the - PyKeen library. Please refer to the PyKeen library documentation + PyKEEN library. Please refer to the PyKEEN library documentation for details and posssible errors regarding this model. Parameters @@ -51,6 +52,8 @@ def __init__( Whether to show loading bars. random_state: int = 42 Random seed to use while training the model + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -63,6 +66,7 @@ def __init__( training_loop=training_loop, verbose=verbose, random_state=random_state, + ring_bell=ring_bell, enable_cache=enable_cache ) diff --git a/embiggen/embedders/pykeen_embedders/distma.py b/embiggen/embedders/pykeen_embedders/distma.py index cfa6d0db..68782adb 100644 --- a/embiggen/embedders/pykeen_embedders/distma.py +++ b/embiggen/embedders/pykeen_embedders/distma.py @@ -1,12 +1,12 @@ -"""Submodule providing wrapper for PyKeen's DistMA model.""" +"""Submodule providing wrapper for PyKEEN's DistMA model.""" from typing import Union, Type, Dict, Any, Optional from pykeen.training import TrainingLoop from pykeen.models import DistMA -from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKeen +from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKEEN from pykeen.triples import CoreTriplesFactory -class DistMAPyKeen(EntityRelationEmbeddingModelPyKeen): +class DistMAPyKEEN(EntityRelationEmbeddingModelPyKEEN): @classmethod def model_name(cls) -> str: diff --git a/embiggen/embedders/pykeen_embedders/distmult.py b/embiggen/embedders/pykeen_embedders/distmult.py index c628b683..7911647f 100644 --- a/embiggen/embedders/pykeen_embedders/distmult.py +++ b/embiggen/embedders/pykeen_embedders/distmult.py @@ -1,12 +1,12 @@ -"""Submodule providing wrapper for PyKeen's DistMult model.""" +"""Submodule providing wrapper for PyKEEN's DistMult model.""" from typing import Union, Type, Dict, Any, Optional from pykeen.training import TrainingLoop from pykeen.models import DistMult -from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKeen +from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKEEN from pykeen.triples import CoreTriplesFactory -class DistMultPyKeen(EntityRelationEmbeddingModelPyKeen): +class DistMultPyKEEN(EntityRelationEmbeddingModelPyKEEN): @classmethod def model_name(cls) -> str: diff --git a/embiggen/embedders/pykeen_embedders/entity_relation_embedding_model_pykeen.py b/embiggen/embedders/pykeen_embedders/entity_relation_embedding_model_pykeen.py index d3459a61..40958503 100644 --- a/embiggen/embedders/pykeen_embedders/entity_relation_embedding_model_pykeen.py +++ b/embiggen/embedders/pykeen_embedders/entity_relation_embedding_model_pykeen.py @@ -1,14 +1,22 @@ -"""Submodule providing wrapper for PyKeen's TransE model.""" -from typing import Union +"""Submodule providing wrapper for PyKEEN's TransE model.""" +from typing import Union, Type, List from ensmallen import Graph -from pykeen.models import EntityRelationEmbeddingModel, ERModel -from embiggen.embedders.pykeen_embedders.pykeen_embedder import PyKeenEmbedder +from pykeen.models import ERModel +from pykeen.nn.representation import Representation +from embiggen.embedders.pykeen_embedders.pykeen_embedder import PyKEENEmbedder import pandas as pd from embiggen.utils.abstract_models import abstract_class, EmbeddingResult +try: + from pykeen.models import EntityRelationEmbeddingModel +except ImportError: + # The following is just to patch the removal of the + # class EntityRelationEmbeddingModel in PyKEEN. + class EntityRelationEmbeddingModel: + pass @abstract_class -class EntityRelationEmbeddingModelPyKeen(PyKeenEmbedder): +class EntityRelationEmbeddingModelPyKEEN(PyKEENEmbedder): def _extract_embeddings( self, @@ -28,11 +36,11 @@ def _extract_embeddings( Whether to return a dataframe of a numpy array. """ if isinstance(model, EntityRelationEmbeddingModel): - node_embeddings = [model.entity_embeddings] - edge_type_embeddings = [model.relation_embeddings] + node_embeddings: List[Type[Representation]] = [model.entity_embeddings] + edge_type_embeddings: List[Type[Representation]] = [model.relation_embeddings] elif isinstance(model, ERModel): - node_embeddings = model.entity_representations - edge_type_embeddings = model.relation_representations + node_embeddings: List[Type[Representation]] = model.entity_representations + edge_type_embeddings: List[Type[Representation]] = model.relation_representations else: raise NotImplementedError( f"The provided model has type {type(model)}, which " @@ -41,13 +49,27 @@ def _extract_embeddings( ) node_embeddings = [ - node_embedding._embeddings.weight.cpu().detach().numpy() - for node_embedding in node_embeddings + array + for array in ( + node_embedding().cpu().detach().numpy().reshape(( + graph.get_number_of_nodes(), + -1 + )) + for node_embedding in node_embeddings + ) + if array.size > 0 ] edge_type_embeddings = [ - edge_type_embedding._embeddings.weight.cpu().detach().numpy() - for edge_type_embedding in edge_type_embeddings + array + for array in ( + edge_type_embedding().cpu().detach().numpy().reshape(( + graph.get_number_of_edge_types(), + -1 + )) + for edge_type_embedding in edge_type_embeddings + ) + if array.size > 0 ] if return_dataframe: diff --git a/embiggen/embedders/pykeen_embedders/ermlp.py b/embiggen/embedders/pykeen_embedders/ermlp.py index f0788f15..125eaf9d 100644 --- a/embiggen/embedders/pykeen_embedders/ermlp.py +++ b/embiggen/embedders/pykeen_embedders/ermlp.py @@ -1,12 +1,12 @@ -"""Submodule providing wrapper for PyKeen's ERMLP model.""" +"""Submodule providing wrapper for PyKEEN's ERMLP model.""" from typing import Union, Type, Dict, Any, Optional from pykeen.training import TrainingLoop from pykeen.models import ERMLP -from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKeen +from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKEEN from pykeen.triples import CoreTriplesFactory -class ERMLPPyKeen(EntityRelationEmbeddingModelPyKeen): +class ERMLPPyKEEN(EntityRelationEmbeddingModelPyKEEN): def __init__( self, @@ -18,14 +18,15 @@ def __init__( ] = "Stochastic Local Closed World Assumption", verbose: bool = False, random_state: int = 42, + ring_bell: bool = False, enable_cache: bool = False ): - """Create new PyKeen ERMLP model. + """Create new PyKEEN ERMLP model. Details ------------------------- This is a wrapper of the ERMLP implementation from the - PyKeen library. Please refer to the PyKeen library documentation + PyKEEN library. Please refer to the PyKEEN library documentation for details and posssible errors regarding this model. Parameters @@ -51,6 +52,8 @@ def __init__( Whether to show loading bars. random_state: int = 42 Random seed to use while training the model + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -63,6 +66,7 @@ def __init__( training_loop=training_loop, verbose=verbose, random_state=random_state, + ring_bell=ring_bell, enable_cache=enable_cache ) @@ -70,7 +74,7 @@ def __init__( def smoke_test_parameters(cls) -> Dict[str, Any]: """Returns parameters for smoke test.""" return dict( - **EntityRelationEmbeddingModelPyKeen.smoke_test_parameters(), + **EntityRelationEmbeddingModelPyKEEN.smoke_test_parameters(), hidden_dim=5 ) diff --git a/embiggen/embedders/pykeen_embedders/ermlpe.py b/embiggen/embedders/pykeen_embedders/ermlpe.py index 638bc0c4..4ebdd88a 100644 --- a/embiggen/embedders/pykeen_embedders/ermlpe.py +++ b/embiggen/embedders/pykeen_embedders/ermlpe.py @@ -1,12 +1,12 @@ -"""Submodule providing wrapper for PyKeen's ERMLPE model.""" +"""Submodule providing wrapper for PyKEEN's ERMLPE model.""" from typing import Union, Type, Dict, Any, Optional from pykeen.training import TrainingLoop from pykeen.models import ERMLPE -from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKeen +from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKEEN from pykeen.triples import CoreTriplesFactory -class ERMLPEPyKeen(EntityRelationEmbeddingModelPyKeen): +class ERMLPEPyKEEN(EntityRelationEmbeddingModelPyKEEN): def __init__( self, @@ -18,14 +18,15 @@ def __init__( ] = "Stochastic Local Closed World Assumption", verbose: bool = False, random_state: int = 42, + ring_bell: bool = False, enable_cache: bool = False ): - """Create new PyKeen ERMLPE model. + """Create new PyKEEN ERMLPE model. Details ------------------------- This is a wrapper of the ERMLPE implementation from the - PyKeen library. Please refer to the PyKeen library documentation + PyKEEN library. Please refer to the PyKEEN library documentation for details and posssible errors regarding this model. Parameters @@ -51,6 +52,8 @@ def __init__( Whether to show loading bars. random_state: int = 42 Random seed to use while training the model + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -63,6 +66,7 @@ def __init__( training_loop=training_loop, verbose=verbose, random_state=random_state, + ring_bell=ring_bell, enable_cache=enable_cache ) @@ -70,7 +74,7 @@ def __init__( def smoke_test_parameters(cls) -> Dict[str, Any]: """Returns parameters for smoke test.""" return dict( - **EntityRelationEmbeddingModelPyKeen.smoke_test_parameters(), + **EntityRelationEmbeddingModelPyKEEN.smoke_test_parameters(), hidden_dim=5 ) diff --git a/embiggen/embedders/pykeen_embedders/hole.py b/embiggen/embedders/pykeen_embedders/hole.py index 901ac5fa..9a4a86b3 100644 --- a/embiggen/embedders/pykeen_embedders/hole.py +++ b/embiggen/embedders/pykeen_embedders/hole.py @@ -1,12 +1,12 @@ -"""Submodule providing wrapper for PyKeen's HolE model.""" +"""Submodule providing wrapper for PyKEEN's HolE model.""" from typing import Union, Type, Dict, Any from pykeen.training import TrainingLoop from pykeen.models import HolE -from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKeen +from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKEEN from pykeen.triples import CoreTriplesFactory -class HolEPyKeen(EntityRelationEmbeddingModelPyKeen): +class HolEPyKEEN(EntityRelationEmbeddingModelPyKEEN): @classmethod def model_name(cls) -> str: diff --git a/embiggen/embedders/pykeen_embedders/cp.py b/embiggen/embedders/pykeen_embedders/node_piece.py similarity index 64% rename from embiggen/embedders/pykeen_embedders/cp.py rename to embiggen/embedders/pykeen_embedders/node_piece.py index 885eb52a..567ac58a 100644 --- a/embiggen/embedders/pykeen_embedders/cp.py +++ b/embiggen/embedders/pykeen_embedders/node_piece.py @@ -1,39 +1,42 @@ -"""Submodule providing wrapper for PyKeen's CP model.""" -from typing import Union, Type, Dict, Any, Optional +"""Submodule providing wrapper for PyKEEN's NodePiece model.""" +from typing import Union, Type, Dict, Any, List from pykeen.training import TrainingLoop -from pykeen.models import CP -from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKeen +from pykeen.models import NodePiece +from ensmallen import Graph +from embiggen.utils.abstract_models import EmbeddingResult +from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKEEN from pykeen.triples import CoreTriplesFactory -class CPPyKeen(EntityRelationEmbeddingModelPyKeen): +class NodePiecePyKEEN(EntityRelationEmbeddingModelPyKEEN): def __init__( self, - embedding_size: int = 256, - rank: int = 64, + embedding_size: int = 64, + num_tokens: Union[int, List[int]] = 2, epochs: int = 100, batch_size: int = 2**10, training_loop: Union[str, Type[TrainingLoop] ] = "Stochastic Local Closed World Assumption", verbose: bool = False, random_state: int = 42, + ring_bell: bool = False, enable_cache: bool = False ): - """Create new PyKeen CP model. + """Create new PyKEEN NodePiece model. Details ------------------------- - This is a wrapper of the CP implementation from the - PyKeen library. Please refer to the PyKeen library documentation + This is a wrapper of the NodePiece implementation from the + PyKEEN library. Please refer to the PyKEEN library documentation for details and posssible errors regarding this model. Parameters ------------------------- - embedding_size: int = 256 + embedding_size: int = 64 The dimension of the embedding to compute. - rank: int = 64 - The hidden dropout rate + num_tokens: Union[int, List[int]] = 2 + The number of relations to use to represent each entity, cf. epochs: int = 100 The number of epochs to use to train the model for. batch_size: int = 2**10 @@ -51,11 +54,13 @@ def __init__( Whether to show loading bars. random_state: int = 42 Random seed to use while training the model + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. """ - self._rank = rank + self._num_tokens = num_tokens super().__init__( embedding_size=embedding_size, epochs=epochs, @@ -63,36 +68,40 @@ def __init__( training_loop=training_loop, verbose=verbose, random_state=random_state, + ring_bell=ring_bell, enable_cache=enable_cache ) def parameters(self) -> Dict[str, Any]: return dict( **super().parameters(), - **dict( - rank=self._rank, - ) + num_tokens=self._num_tokens, ) @classmethod def model_name(cls) -> str: """Return name of the model.""" - return "CP" + return "NodePiece" def _build_model( self, triples_factory: CoreTriplesFactory - ) -> CP: - """Build new CP model for embedding. + ) -> NodePiece: + """Build new NodePiece model for embedding. Parameters ------------------ graph: Graph The graph to build the model for. """ - return CP( + return NodePiece( triples_factory=triples_factory, + num_tokens=self._num_tokens, embedding_dim=self._embedding_size, - rank=self._rank, random_seed=self._random_state ) + + @classmethod + def _create_inverse_triples(cls) -> bool: + """Returns whether the class is expected to create inverse triples.""" + return True \ No newline at end of file diff --git a/embiggen/embedders/pykeen_embedders/pairre.py b/embiggen/embedders/pykeen_embedders/pairre.py index 2002ba30..13c13129 100644 --- a/embiggen/embedders/pykeen_embedders/pairre.py +++ b/embiggen/embedders/pykeen_embedders/pairre.py @@ -1,12 +1,12 @@ -"""Submodule providing wrapper for PyKeen's PairRE model.""" +"""Submodule providing wrapper for PyKEEN's PairRE model.""" from typing import Union, Type, Dict, Any, Optional from pykeen.training import TrainingLoop from pykeen.models import PairRE -from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKeen +from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKEEN from pykeen.triples import CoreTriplesFactory -class PairREPyKeen(EntityRelationEmbeddingModelPyKeen): +class PairREPyKEEN(EntityRelationEmbeddingModelPyKEEN): def __init__( self, @@ -19,14 +19,15 @@ def __init__( ] = "Stochastic Local Closed World Assumption", verbose: bool = False, random_state: int = 42, + ring_bell: bool = False, enable_cache: bool = False ): - """Create new PyKeen PairRE model. + """Create new PyKEEN PairRE model. Details ------------------------- This is a wrapper of the PairRE implementation from the - PyKeen library. Please refer to the PyKeen library documentation + PyKEEN library. Please refer to the PyKEEN library documentation for details and posssible errors regarding this model. Parameters @@ -54,6 +55,8 @@ def __init__( Whether to show loading bars. random_state: int = 42 Random seed to use while training the model + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -67,6 +70,7 @@ def __init__( training_loop=training_loop, verbose=verbose, random_state=random_state, + ring_bell=ring_bell, enable_cache=enable_cache ) diff --git a/embiggen/embedders/pykeen_embedders/proje.py b/embiggen/embedders/pykeen_embedders/proje.py index e677551f..3a3f3adc 100644 --- a/embiggen/embedders/pykeen_embedders/proje.py +++ b/embiggen/embedders/pykeen_embedders/proje.py @@ -1,12 +1,12 @@ -"""Submodule providing wrapper for PyKeen's ProjE model.""" +"""Submodule providing wrapper for PyKEEN's ProjE model.""" from typing import Union, Type, Dict, Any, Optional from pykeen.training import TrainingLoop from pykeen.models import ProjE -from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKeen +from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKEEN from pykeen.triples import CoreTriplesFactory -class ProjEPyKeen(EntityRelationEmbeddingModelPyKeen): +class ProjEPyKEEN(EntityRelationEmbeddingModelPyKEEN): @classmethod def model_name(cls) -> str: diff --git a/embiggen/embedders/pykeen_embedders/pykeen_embedder.py b/embiggen/embedders/pykeen_embedders/pykeen_embedder.py index ea975a50..d0f21d5d 100644 --- a/embiggen/embedders/pykeen_embedders/pykeen_embedder.py +++ b/embiggen/embedders/pykeen_embedders/pykeen_embedder.py @@ -1,10 +1,11 @@ -"""Abstract Torch/PyKeen Model wrapper for embedding models.""" +"""Abstract Torch/PyKEEN Model wrapper for embedding models.""" from typing import Dict, Union, Tuple, Any, Type import numpy as np import pandas as pd from ensmallen import Graph import inspect +from inspect import getfullargspec from embiggen.utils.pytorch_utils import validate_torch_device from embiggen.utils.abstract_models import AbstractEmbeddingModel, abstract_class, EmbeddingResult @@ -16,8 +17,8 @@ @abstract_class -class PyKeenEmbedder(AbstractEmbeddingModel): - """Abstract Torch/PyKeen Model wrapper for embedding models.""" +class PyKEENEmbedder(AbstractEmbeddingModel): + """Abstract Torch/PyKEEN Model wrapper for embedding models.""" SUPPORTED_TRAINING_LOOPS = { "Stochastic Local Closed World Assumption": SLCWATrainingLoop, @@ -34,9 +35,10 @@ def __init__( ] = "Stochastic Local Closed World Assumption", verbose: bool = False, random_state: int = 42, + ring_bell: bool = False, enable_cache: bool = False ): - """Create new PyKeen Abstract Embedder model. + """Create new PyKEEN Abstract Embedder model. Parameters ------------------------- @@ -59,18 +61,20 @@ def __init__( Whether to show the loading bar. random_state: int = 42 Random seed to use while training the model + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. """ if isinstance(training_loop, str): - if training_loop in PyKeenEmbedder.SUPPORTED_TRAINING_LOOPS: - training_loop = PyKeenEmbedder.SUPPORTED_TRAINING_LOOPS[training_loop] + if training_loop in PyKEENEmbedder.SUPPORTED_TRAINING_LOOPS: + training_loop = PyKEENEmbedder.SUPPORTED_TRAINING_LOOPS[training_loop] else: raise ValueError( f"The provided training loop name {training_loop} is not " "a supported training loop name. " - f"The supported names are {format_list(PyKeenEmbedder.SUPPORTED_TRAINING_LOOPS)}." + f"The supported names are {format_list(PyKEENEmbedder.SUPPORTED_TRAINING_LOOPS)}." ) if not inspect.isclass(training_loop): @@ -92,6 +96,7 @@ def __init__( super().__init__( embedding_size=embedding_size, enable_cache=enable_cache, + ring_bell=ring_bell, random_state=random_state ) @@ -114,7 +119,7 @@ def parameters(self) -> Dict[str, Any]: @classmethod def library_name(cls) -> str: - return "PyKeen" + return "PyKEEN" @classmethod def task_name(cls) -> str: @@ -126,7 +131,7 @@ def _build_model(self, triples_factory: CoreTriplesFactory) -> Type[Model]: Parameters ------------------ triples_factory: CoreTriplesFactory - The PyKeen triples factory to use to create the model. + The PyKEEN triples factory to use to create the model. """ raise NotImplementedError( f"In the child class {self.__class__.__name__} of {super().__name__.__name__} " @@ -167,6 +172,11 @@ def _extract_embeddings( "called `_extract_embeddings`. Please do implement it." ) + @classmethod + def _create_inverse_triples(cls) -> bool: + """Returns whether the class is expected to create inverse triples.""" + return False + def _fit_transform( self, graph: Graph, @@ -176,14 +186,22 @@ def _fit_transform( torch_device = torch.device(self._device) - triples_factory = CoreTriplesFactory( - torch.IntTensor(graph.get_directed_edge_triples_ids().astype(np.int32)), - num_entities=graph.get_number_of_nodes(), - num_relations=graph.get_number_of_edge_types(), - entity_ids=graph.get_node_ids(), - relation_ids=graph.get_unique_edge_type_ids(), - create_inverse_triples=False, - ) + if "entity_ids" in getfullargspec(CoreTriplesFactory).args: + triples_factory = CoreTriplesFactory( + torch.IntTensor(graph.get_directed_edge_triples_ids().astype(np.int64)), + num_entities=graph.get_number_of_nodes(), + num_relations=graph.get_number_of_edge_types(), + entity_ids=graph.get_node_ids().astype(np.int64), + relation_ids=graph.get_unique_edge_type_ids().astype(np.int64), + create_inverse_triples=self._create_inverse_triples(), + ) + else: + triples_factory = CoreTriplesFactory( + torch.IntTensor(graph.get_directed_edge_triples_ids().astype(np.int64)), + num_entities=graph.get_number_of_nodes(), + num_relations=graph.get_number_of_edge_types(), + create_inverse_triples=self._create_inverse_triples(), + ) batch_size = min( self._batch_size, @@ -197,7 +215,7 @@ def _fit_transform( "The model created with the `_build_model` in the child " f"class {self.__class__.__name__} for the model {self.model_name()} " f"in the library {self.library_name()} did not return a " - f"PyKeen model but an object of type {type(model)}." + f"PyKEEN model but an object of type {type(model)}." ) # Move the model to gpu if we need to diff --git a/embiggen/embedders/pykeen_embedders/quate.py b/embiggen/embedders/pykeen_embedders/quate.py index ca9caff0..9d09197c 100644 --- a/embiggen/embedders/pykeen_embedders/quate.py +++ b/embiggen/embedders/pykeen_embedders/quate.py @@ -1,12 +1,12 @@ -"""Submodule providing wrapper for PyKeen's QuatE model.""" +"""Submodule providing wrapper for PyKEEN's QuatE model.""" from typing import Union, Type, Dict, Any, Optional from pykeen.training import TrainingLoop from pykeen.models import QuatE -from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKeen +from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKEEN from pykeen.triples import CoreTriplesFactory -class QuatEPyKeen(EntityRelationEmbeddingModelPyKeen): +class QuatEPyKEEN(EntityRelationEmbeddingModelPyKEEN): @classmethod def model_name(cls) -> str: diff --git a/embiggen/embedders/pykeen_embedders/rescal.py b/embiggen/embedders/pykeen_embedders/rescal.py index 40071b24..adb9b3ed 100644 --- a/embiggen/embedders/pykeen_embedders/rescal.py +++ b/embiggen/embedders/pykeen_embedders/rescal.py @@ -1,12 +1,12 @@ -"""Submodule providing wrapper for PyKeen's RESCAL model.""" +"""Submodule providing wrapper for PyKEEN's RESCAL model.""" from typing import Union, Type, Dict, Any from pykeen.training import TrainingLoop from pykeen.models import RESCAL -from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKeen +from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKEEN from pykeen.triples import CoreTriplesFactory -class RESCALPyKeen(EntityRelationEmbeddingModelPyKeen): +class RESCALPyKEEN(EntityRelationEmbeddingModelPyKEEN): @classmethod def model_name(cls) -> str: diff --git a/embiggen/embedders/pykeen_embedders/rotate.py b/embiggen/embedders/pykeen_embedders/rotate.py index 0a46929c..aa133e58 100644 --- a/embiggen/embedders/pykeen_embedders/rotate.py +++ b/embiggen/embedders/pykeen_embedders/rotate.py @@ -1,12 +1,12 @@ -"""Submodule providing wrapper for PyKeen's RotatE model.""" +"""Submodule providing wrapper for PyKEEN's RotatE model.""" from typing import Union, Type, Dict, Any from pykeen.training import TrainingLoop from pykeen.models import RotatE -from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKeen +from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKEEN from pykeen.triples import CoreTriplesFactory -class RotatEPyKeen(EntityRelationEmbeddingModelPyKeen): +class RotatEPyKEEN(EntityRelationEmbeddingModelPyKEEN): @classmethod def model_name(cls) -> str: diff --git a/embiggen/embedders/pykeen_embedders/toruse.py b/embiggen/embedders/pykeen_embedders/toruse.py index c0d101b4..2ab6bb17 100644 --- a/embiggen/embedders/pykeen_embedders/toruse.py +++ b/embiggen/embedders/pykeen_embedders/toruse.py @@ -1,12 +1,12 @@ -"""Submodule providing wrapper for PyKeen's TorusE model.""" +"""Submodule providing wrapper for PyKEEN's TorusE model.""" from typing import Union, Type, Dict, Any, Optional from pykeen.training import TrainingLoop from pykeen.models import TorusE -from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKeen +from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKEEN from pykeen.triples import CoreTriplesFactory -class TorusEPyKeen(EntityRelationEmbeddingModelPyKeen): +class TorusEPyKEEN(EntityRelationEmbeddingModelPyKEEN): def __init__( self, @@ -19,14 +19,15 @@ def __init__( ] = "Stochastic Local Closed World Assumption", verbose: bool = False, random_state: int = 42, + ring_bell: bool = False, enable_cache: bool = False ): - """Create new PyKeen TorusE model. + """Create new PyKEEN TorusE model. Details ------------------------- This is a wrapper of the TorusE implementation from the - PyKeen library. Please refer to the PyKeen library documentation + PyKEEN library. Please refer to the PyKEEN library documentation for details and posssible errors regarding this model. Parameters @@ -54,6 +55,8 @@ def __init__( Whether to show loading bars. random_state: int = 42 Random seed to use while training the model + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -67,6 +70,7 @@ def __init__( training_loop=training_loop, verbose=verbose, random_state=random_state, + ring_bell=ring_bell, enable_cache=enable_cache ) diff --git a/embiggen/embedders/pykeen_embedders/transd.py b/embiggen/embedders/pykeen_embedders/transd.py index 1b8c4c39..a8c0510d 100644 --- a/embiggen/embedders/pykeen_embedders/transd.py +++ b/embiggen/embedders/pykeen_embedders/transd.py @@ -1,12 +1,12 @@ -"""Submodule providing wrapper for PyKeen's TransD model.""" +"""Submodule providing wrapper for PyKEEN's TransD model.""" from typing import Union, Type, Dict, Any, Optional from pykeen.training import TrainingLoop from pykeen.models import TransD -from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKeen +from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKEEN from pykeen.triples import CoreTriplesFactory -class TransDPyKeen(EntityRelationEmbeddingModelPyKeen): +class TransDPyKEEN(EntityRelationEmbeddingModelPyKEEN): def __init__( self, @@ -18,14 +18,15 @@ def __init__( ] = "Stochastic Local Closed World Assumption", verbose: bool = False, random_state: int = 42, + ring_bell: bool = False, enable_cache: bool = False ): - """Create new PyKeen TransD model. + """Create new PyKEEN TransD model. Details ------------------------- This is a wrapper of the TransD implementation from the - PyKeen library. Please refer to the PyKeen library documentation + PyKEEN library. Please refer to the PyKEEN library documentation for details and posssible errors regarding this model. Parameters @@ -51,6 +52,8 @@ def __init__( Whether to show loading bars. random_state: int = 42 Random seed to use while training the model + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -63,6 +66,7 @@ def __init__( training_loop=training_loop, verbose=verbose, random_state=random_state, + ring_bell=ring_bell, enable_cache=enable_cache ) @@ -70,7 +74,7 @@ def __init__( def smoke_test_parameters(cls) -> Dict[str, Any]: """Returns parameters for smoke test.""" return dict( - **EntityRelationEmbeddingModelPyKeen.smoke_test_parameters(), + **EntityRelationEmbeddingModelPyKEEN.smoke_test_parameters(), relation_dim=5, ) diff --git a/embiggen/embedders/pykeen_embedders/transe.py b/embiggen/embedders/pykeen_embedders/transe.py index 3435e5bc..f460ba20 100644 --- a/embiggen/embedders/pykeen_embedders/transe.py +++ b/embiggen/embedders/pykeen_embedders/transe.py @@ -1,12 +1,12 @@ -"""Submodule providing wrapper for PyKeen's TransE model.""" +"""Submodule providing wrapper for PyKEEN's TransE model.""" from typing import Union, Type, Dict, Any from pykeen.training import TrainingLoop from pykeen.models import TransE -from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKeen +from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKEEN from pykeen.triples import CoreTriplesFactory -class TransEPyKeen(EntityRelationEmbeddingModelPyKeen): +class TransEPyKEEN(EntityRelationEmbeddingModelPyKEEN): def __init__( self, @@ -18,14 +18,15 @@ def __init__( ] = "Stochastic Local Closed World Assumption", verbose: bool = False, random_state: int = 42, + ring_bell: bool = False, enable_cache: bool = False ): - """Create new PyKeen TransE model. + """Create new PyKEEN TransE model. Details ------------------------- This is a wrapper of the TransE implementation from the - PyKeen library. Please refer to the PyKeen library documentation + PyKEEN library. Please refer to the PyKEEN library documentation for details and posssible errors regarding this model. Parameters @@ -51,6 +52,8 @@ def __init__( Whether to show loading bars. random_state: int = 42 Random seed to use while training the model + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -63,6 +66,7 @@ def __init__( training_loop=training_loop, verbose=verbose, random_state=random_state, + ring_bell=ring_bell, enable_cache=enable_cache ) @@ -70,7 +74,7 @@ def __init__( def smoke_test_parameters(cls) -> Dict[str, Any]: """Returns parameters for smoke test.""" return dict( - **EntityRelationEmbeddingModelPyKeen.smoke_test_parameters(), + **EntityRelationEmbeddingModelPyKEEN.smoke_test_parameters(), scoring_fct_norm=1 ) diff --git a/embiggen/embedders/pykeen_embedders/transf.py b/embiggen/embedders/pykeen_embedders/transf.py index f6d4cf1f..a7739700 100644 --- a/embiggen/embedders/pykeen_embedders/transf.py +++ b/embiggen/embedders/pykeen_embedders/transf.py @@ -1,12 +1,12 @@ -"""Submodule providing wrapper for PyKeen's TransF model.""" +"""Submodule providing wrapper for PyKEEN's TransF model.""" from typing import Union, Type, Dict, Any from pykeen.training import TrainingLoop from pykeen.models import TransF -from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKeen +from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKEEN from pykeen.triples import CoreTriplesFactory -class TransFPyKeen(EntityRelationEmbeddingModelPyKeen): +class TransFPyKEEN(EntityRelationEmbeddingModelPyKEEN): @classmethod def model_name(cls) -> str: diff --git a/embiggen/embedders/pykeen_embedders/transh.py b/embiggen/embedders/pykeen_embedders/transh.py index 1fddcaf6..a8935375 100644 --- a/embiggen/embedders/pykeen_embedders/transh.py +++ b/embiggen/embedders/pykeen_embedders/transh.py @@ -1,12 +1,12 @@ -"""Submodule providing wrapper for PyKeen's TransH model.""" +"""Submodule providing wrapper for PyKEEN's TransH model.""" from typing import Union, Type, Dict, Any from pykeen.training import TrainingLoop from pykeen.models import TransH -from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKeen +from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKEEN from pykeen.triples import CoreTriplesFactory -class TransHPyKeen(EntityRelationEmbeddingModelPyKeen): +class TransHPyKEEN(EntityRelationEmbeddingModelPyKEEN): def __init__( self, @@ -18,14 +18,15 @@ def __init__( ] = "Stochastic Local Closed World Assumption", verbose: bool = False, random_state: int = 42, + ring_bell: bool = False, enable_cache: bool = False ): - """Create new PyKeen TransH model. + """Create new PyKEEN TransH model. Details ------------------------- This is a wrapper of the TransH implementation from the - PyKeen library. Please refer to the PyKeen library documentation + PyKEEN library. Please refer to the PyKEEN library documentation for details and posssible errors regarding this model. Parameters @@ -51,6 +52,8 @@ def __init__( Whether to show loading bars. random_state: int = 42 Random seed to use while training the model + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -63,6 +66,7 @@ def __init__( training_loop=training_loop, verbose=verbose, random_state=random_state, + ring_bell=ring_bell, enable_cache=enable_cache ) @@ -70,7 +74,7 @@ def __init__( def smoke_test_parameters(cls) -> Dict[str, Any]: """Returns parameters for smoke test.""" return dict( - **EntityRelationEmbeddingModelPyKeen.smoke_test_parameters(), + **EntityRelationEmbeddingModelPyKEEN.smoke_test_parameters(), scoring_fct_norm=1 ) diff --git a/embiggen/embedders/pykeen_embedders/transr.py b/embiggen/embedders/pykeen_embedders/transr.py index 3f8b5637..4162bcc4 100644 --- a/embiggen/embedders/pykeen_embedders/transr.py +++ b/embiggen/embedders/pykeen_embedders/transr.py @@ -1,12 +1,12 @@ -"""Submodule providing wrapper for PyKeen's TransR model.""" +"""Submodule providing wrapper for PyKEEN's TransR model.""" from typing import Union, Type, Dict, Any from pykeen.training import TrainingLoop from pykeen.models import TransR -from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKeen +from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKEEN from pykeen.triples import CoreTriplesFactory -class TransRPyKeen(EntityRelationEmbeddingModelPyKeen): +class TransRPyKEEN(EntityRelationEmbeddingModelPyKEEN): def __init__( self, @@ -19,14 +19,15 @@ def __init__( ] = "Stochastic Local Closed World Assumption", verbose: bool = False, random_state: int = 42, + ring_bell: bool = False, enable_cache: bool = False ): - """Create new PyKeen TransR model. + """Create new PyKEEN TransR model. Details ------------------------- This is a wrapper of the TransR implementation from the - PyKeen library. Please refer to the PyKeen library documentation + PyKEEN library. Please refer to the PyKEEN library documentation for details and posssible errors regarding this model. Parameters @@ -54,6 +55,8 @@ def __init__( Whether to show loading bars. random_state: int = 42 Random seed to use while training the model + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -67,6 +70,7 @@ def __init__( training_loop=training_loop, verbose=verbose, random_state=random_state, + ring_bell=ring_bell, enable_cache=enable_cache ) @@ -74,7 +78,7 @@ def __init__( def smoke_test_parameters(cls) -> Dict[str, Any]: """Returns parameters for smoke test.""" return dict( - **EntityRelationEmbeddingModelPyKeen.smoke_test_parameters(), + **EntityRelationEmbeddingModelPyKEEN.smoke_test_parameters(), scoring_fct_norm=1, relation_dim=5 ) diff --git a/embiggen/embedders/pykeen_embedders/tucker.py b/embiggen/embedders/pykeen_embedders/tucker.py index 70cc9b00..1f75927d 100644 --- a/embiggen/embedders/pykeen_embedders/tucker.py +++ b/embiggen/embedders/pykeen_embedders/tucker.py @@ -1,12 +1,12 @@ -"""Submodule providing wrapper for PyKeen's TuckER model.""" +"""Submodule providing wrapper for PyKEEN's TuckER model.""" from typing import Union, Type, Dict, Any, Optional from pykeen.training import TrainingLoop from pykeen.models import TuckER -from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKeen +from embiggen.embedders.pykeen_embedders.entity_relation_embedding_model_pykeen import EntityRelationEmbeddingModelPyKEEN from pykeen.triples import CoreTriplesFactory -class TuckERPyKeen(EntityRelationEmbeddingModelPyKeen): +class TuckERPyKEEN(EntityRelationEmbeddingModelPyKEEN): def __init__( self, @@ -22,14 +22,15 @@ def __init__( ] = "Stochastic Local Closed World Assumption", verbose: bool = False, random_state: int = 42, + ring_bell: bool = False, enable_cache: bool = False ): - """Create new PyKeen TuckER model. + """Create new PyKEEN TuckER model. Details ------------------------- This is a wrapper of the TuckER implementation from the - PyKeen library. Please refer to the PyKeen library documentation + PyKEEN library. Please refer to the PyKEEN library documentation for details and posssible errors regarding this model. Parameters @@ -63,6 +64,8 @@ def __init__( Whether to show loading bars. random_state: int = 42 Random seed to use while training the model + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -79,6 +82,7 @@ def __init__( training_loop=training_loop, verbose=verbose, random_state=random_state, + ring_bell=ring_bell, enable_cache=enable_cache ) diff --git a/embiggen/embedders/tensorflow_embedders/edge_prediction_based_tensorflow_embedders.py b/embiggen/embedders/tensorflow_embedders/edge_prediction_based_tensorflow_embedders.py index b949b6b8..b55422fb 100644 --- a/embiggen/embedders/tensorflow_embedders/edge_prediction_based_tensorflow_embedders.py +++ b/embiggen/embedders/tensorflow_embedders/edge_prediction_based_tensorflow_embedders.py @@ -32,6 +32,7 @@ def __init__( loss: str = "binary_crossentropy", optimizer: str = "nadam", verbose: bool = False, + ring_bell: bool = False, enable_cache: bool = False, random_state: int = 42 ): @@ -77,6 +78,8 @@ def __init__( The optimizer to be used during the training of the model. verbose: bool = False Whether to show the loading bar while training the model. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -99,6 +102,7 @@ def __init__( verbose=verbose, use_mirrored_strategy=use_mirrored_strategy, enable_cache=enable_cache, + ring_bell=ring_bell, random_state=random_state ) diff --git a/embiggen/embedders/tensorflow_embedders/node2vec.py b/embiggen/embedders/tensorflow_embedders/node2vec.py index bd61f805..4fadd1ed 100644 --- a/embiggen/embedders/tensorflow_embedders/node2vec.py +++ b/embiggen/embedders/tensorflow_embedders/node2vec.py @@ -38,6 +38,7 @@ def __init__( optimizer: str = "nadam", verbose: bool = False, use_mirrored_strategy: bool = False, + ring_bell: bool = False, enable_cache: bool = False ): """Create new abstract Node2Vec model. @@ -109,6 +110,8 @@ def __init__( Whether to show loading bars. use_mirrored_strategy: bool = False Whether to use mirrored strategy. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -136,6 +139,7 @@ def __init__( optimizer=optimizer, verbose=verbose, use_mirrored_strategy=use_mirrored_strategy, + ring_bell=ring_bell, enable_cache=enable_cache ) diff --git a/embiggen/embedders/tensorflow_embedders/siamese.py b/embiggen/embedders/tensorflow_embedders/siamese.py index 7753ec7b..16636969 100644 --- a/embiggen/embedders/tensorflow_embedders/siamese.py +++ b/embiggen/embedders/tensorflow_embedders/siamese.py @@ -34,6 +34,7 @@ def __init__( use_mirrored_strategy: bool = False, optimizer: str = "nadam", verbose: bool = False, + ring_bell: bool = False, enable_cache: bool = False, random_state: int = 42 ): @@ -72,6 +73,8 @@ def __init__( The optimizer to be used during the training of the model. verbose: bool = False Whether to show loading bars. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -92,6 +95,7 @@ def __init__( verbose=verbose, use_mirrored_strategy=use_mirrored_strategy, enable_cache=enable_cache, + ring_bell=ring_bell, random_state=random_state ) diff --git a/embiggen/embedders/tensorflow_embedders/tensorflow_embedder.py b/embiggen/embedders/tensorflow_embedders/tensorflow_embedder.py index d759f698..b33c5fae 100644 --- a/embiggen/embedders/tensorflow_embedders/tensorflow_embedder.py +++ b/embiggen/embedders/tensorflow_embedders/tensorflow_embedder.py @@ -30,6 +30,7 @@ def __init__( optimizer: str = "nadam", verbose: bool = False, use_mirrored_strategy: bool = False, + ring_bell: bool = False, enable_cache: bool = False, random_state: int = 42 ): @@ -61,6 +62,8 @@ def __init__( Whether to show loading bars. use_mirrored_strategy: bool = False Whether to use mirrored strategy. + ring_bell: bool = False, + Whether to play a sound when embedding completes. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -85,6 +88,7 @@ def __init__( super().__init__( embedding_size=embedding_size, enable_cache=enable_cache, + ring_bell=ring_bell, random_state=random_state, ) diff --git a/embiggen/node_label_prediction/node_label_prediction_evaluation.py b/embiggen/node_label_prediction/node_label_prediction_evaluation.py index aea374f1..a126edd7 100644 --- a/embiggen/node_label_prediction/node_label_prediction_evaluation.py +++ b/embiggen/node_label_prediction/node_label_prediction_evaluation.py @@ -16,13 +16,16 @@ def node_label_prediction_evaluation( library_names: Optional[Union[str, List[str]]] = None, graph_callback: Optional[Callable[[Graph], Graph]] = None, subgraph_of_interest: Optional[Graph] = None, + use_subgraph_as_support: bool = False, number_of_holdouts: int = 10, random_state: int = 42, repositories: Optional[Union[str, List[str]]] = None, versions: Optional[Union[str, List[str]]] = None, enable_cache: bool = False, - precompute_constant_automatic_stocastic_features: bool = False, + precompute_constant_stocastic_features: bool = False, smoke_test: bool = False, + number_of_slurm_nodes: Optional[int] = None, + slurm_node_id_variable: str = "SLURM_GRAPE_ID", verbose: bool = True ) -> pd.DataFrame: """Execute node-label prediction evaluation pipeline for all provided models and graphs. @@ -49,7 +52,12 @@ def node_label_prediction_evaluation( For instance this may be used for filtering the uncertain edges in graphs such as STRING PPIs. subgraph_of_interest: Optional[Graph] = None - The subgraph of interest to focus the task on. + Optional subgraph where to focus the task. + This is applied to the train and test graph + after the desired holdout schema is applied. + use_subgraph_as_support: bool = False + Whether to use the provided subgraph as support or + to use the train graph (not filtered by the subgraph). number_of_holdouts: int = 10 The number of holdouts to execute. random_state: int = 42 @@ -61,7 +69,7 @@ def node_label_prediction_evaluation( Graph versions to retrieve. enable_cache: bool = False Whether to enable the cache. - precompute_constant_automatic_stocastic_features: bool = False + precompute_constant_stocastic_features: bool = False Whether to precompute once the constant automatic stocastic features before starting the embedding loop. This means that, when left set to false, while the features will be computed @@ -75,6 +83,12 @@ def node_label_prediction_evaluation( and therefore use the smoke test configurations for the provided model names and feature names. This parameter will also turn off the cache. + number_of_slurm_nodes: Optional[int] = None + Number of SLURM nodes to consider as available. + This variable is used to parallelize the holdouts accordingly. + slurm_node_id_variable: str = "SLURM_GRAPE_ID" + Name of the system variable to use as SLURM node id. + It must be set in the slurm bash script. verbose: bool = True Whether to show loading bars """ @@ -88,12 +102,15 @@ def node_label_prediction_evaluation( library_names=library_names, graph_callback=graph_callback, subgraph_of_interest=subgraph_of_interest, + use_subgraph_as_support=use_subgraph_as_support, number_of_holdouts=number_of_holdouts, random_state=random_state, repositories=repositories, versions=versions, enable_cache=enable_cache, - precompute_constant_automatic_stocastic_features=precompute_constant_automatic_stocastic_features, + precompute_constant_stocastic_features=precompute_constant_stocastic_features, smoke_test=smoke_test, + number_of_slurm_nodes=number_of_slurm_nodes, + slurm_node_id_variable=slurm_node_id_variable, verbose=verbose ) diff --git a/embiggen/node_label_prediction/node_label_prediction_sklearn/sklearn_node_label_prediction_adapter.py b/embiggen/node_label_prediction/node_label_prediction_sklearn/sklearn_node_label_prediction_adapter.py index 9b73f9f4..12ae5026 100644 --- a/embiggen/node_label_prediction/node_label_prediction_sklearn/sklearn_node_label_prediction_adapter.py +++ b/embiggen/node_label_prediction/node_label_prediction_sklearn/sklearn_node_label_prediction_adapter.py @@ -2,6 +2,7 @@ from sklearn.base import ClassifierMixin from typing import Type, List, Dict, Optional, Any import numpy as np +import compress_pickle import copy from ensmallen import Graph from embiggen.embedding_transformers import NodeLabelPredictionTransformer, NodeTransformer @@ -224,3 +225,23 @@ def can_use_edge_types(cls) -> bool: """Returns whether the model can optionally use edge types.""" return False + @classmethod + def load(cls, path: str) -> "Self": + """Load a saved version of the model from the provided path. + + Parameters + ------------------- + path: str + Path from where to load the model. + """ + return compress_pickle.load(path) + + def dump(self, path: str): + """Dump the current model at the provided path. + + Parameters + ------------------- + path: str + Path from where to dump the model. + """ + compress_pickle.dump(self, path) \ No newline at end of file diff --git a/embiggen/similarities/dag_resnik.py b/embiggen/similarities/dag_resnik.py index e4d368ff..c78ba754 100644 --- a/embiggen/similarities/dag_resnik.py +++ b/embiggen/similarities/dag_resnik.py @@ -1,4 +1,4 @@ -from typing import List, Optional, Union, Dict +from typing import List, Optional, Union, Dict, Tuple import pandas as pd import numpy as np from ensmallen import models, Graph @@ -7,7 +7,9 @@ class DAGResnik: def __init__(self, verbose: bool = True): + """Create new Resnik similarity model.""" self._model = models.DAGResnik(verbose) + self._graph = None def fit( self, @@ -26,370 +28,381 @@ def fit( node_frequencies: Optional[np.ndarray] = None Optional vector of node frequencies. """ + self._graph = graph self._model.fit( graph, node_counts=node_counts, node_frequencies=node_frequencies ) - def get_similarity_from_node_id( + def _normalize_output( self, - first_node_id: int, - second_node_id: int - ) -> float: - """Return the similarity between the two provided nodes. - - Arguments - -------------------- - first_node_id: int - The first node for which to compute the similarity. - second_node_id: int - The second node for which to compute the similarity. - """ - return self._model.get_similarity_from_node_id(first_node_id, second_node_id) - - def get_similarity_from_node_ids( - self, - first_node_ids: List[int], - second_node_ids: List[int] - ) -> np.ndarray: - """Return the similarity between the two provided nodes. - - Arguments - -------------------- - first_node_ids: List[int] - The first node for which to compute the similarity. - second_node_ids: List[int] - The second node for which to compute the similarity. - """ - return self._model.get_similarity_from_node_ids(first_node_ids, second_node_ids) - - def get_similarity_from_node_name( - self, - first_node_name: str, - second_node_name: str - ) -> float: - """Return the similarity between the two provided nodes. - - Arguments - -------------------- - first_node_name: str - The first node for which to compute the similarity. - second_node_name: str - The second node for which to compute the similarity. - """ - return self._model.get_similarity_from_node_name(first_node_name, second_node_name) - - def get_similarity_from_node_names( - self, - first_node_names: List[str], - second_node_names: List[str] - ) -> np.ndarray: - """Return the similarity between the two provided nodes. - - Arguments - -------------------- - first_node_names: List[str] - The first node for which to compute the similarity. - second_node_names: List[str] - The second node for which to compute the similarity. - """ - return self._model.get_similarity_from_node_names(first_node_names, second_node_names) - - def get_pairwise_similarities( - self, - graph: Optional[Graph] = None, - return_similarities_dataframe: bool = False - ) -> Union[pd.DataFrame, np.ndarray]: - """Execute similarities on the provided graph. - + edge_node_ids: np.ndarray, + similarities: np.ndarray, + return_similarities_dataframe: bool = False, + return_node_names: bool = False + ) -> Union[pd.DataFrame, Tuple[np.ndarray, np.ndarray]]: + """Normalize output to provided standard. + Parameters - -------------------- - graph: Optional[Graph] = None - The graph to run similarities on. + --------------------- + edge_node_ids: np.ndarray + The edge node IDs composing the edges. + similarities: np.ndarray + Resnik similarity scores. return_similarities_dataframe: bool = False - Whether to return a pandas DataFrame, which as indices has the node IDs. - By default, a numpy array with the similarities is returned as it weights much less. + Whether to return the data as a DataFrame. + Do note that this will require more RAM. + return_node_names: bool = False + Whether to return the node names. + Do note that this will require SIGNIFICANTLY more RAM. """ - similarities = self._model.get_pairwise_similarities() - - if return_similarities_dataframe and graph is not None: - similarities = pd.DataFrame( - similarities, - columns=graph.get_node_names(), - index=graph.get_node_names(), + if return_node_names and not return_similarities_dataframe: + raise NotImplementedError( + "It is not currently supported to return the node names " + "when it is not requested to return a pandas DataFrame. " + "This is not supported as the RAM requirements for common " + "use cases are large enough to be unfeaseable on most systems." ) - return similarities + if not return_similarities_dataframe: + return (edge_node_ids, similarities) - def get_similarities_from_graph( - self, - graph: Graph, - return_similarities_dataframe: bool = False - ) -> Union[pd.DataFrame, np.ndarray]: - """Execute similarities on the provided graph. - - Parameters - -------------------- - graph: Graph - The graph to run similarities on. - node_frequencies: Optional[np.ndarray] - Optional vector of node frequencies. - return_similarities_dataframe: bool = False - Whether to return a pandas DataFrame, which as indices has the node IDs. - By default, a numpy array with the similarities is returned as it weights much less. - """ - similarities = self._model.get_similarities_from_graph( - graph, - ) - - if return_similarities_dataframe: - similarities = pd.DataFrame( - { - "similarities": similarities, - "sources": graph.get_directed_source_node_ids(), - "destinations": graph.get_directed_destination_node_ids(), - }, - ) - - return similarities + return pd.DataFrame({ + "source": ( + self._graph.get_node_names_from_node_ids(edge_node_ids[:, 0]) + if return_node_names + else edge_node_ids[:, 0] + ), + "destination": ( + self._graph.get_node_names_from_node_ids(edge_node_ids[:, 1]) + if return_node_names + else edge_node_ids[:, 1] + ), + "resnik_score": similarities + }) - def get_similarities_from_bipartite_graph_from_edge_node_ids( + def get_similarities_from_bipartite_graph_node_ids( self, - graph: Graph, - source_node_ids: List[int], - destination_node_ids: List[int], - return_similarities_dataframe: bool = False - ) -> Union[pd.DataFrame, np.ndarray]: - """Execute similarities probabilities on the provided graph bipartite portion. + source_node_ids: List[str], + destination_node_ids: List[str], + minimum_similarity: Optional[float] = 0.0, + return_similarities_dataframe: bool = False, + return_node_names: bool = False + ) -> Union[pd.DataFrame, Tuple[np.ndarray, np.ndarray]]: + """Execute similarities probabilities on the bipartite portion Parameters -------------------- - graph: Graph - The graph from which to extract the edges. - node_frequencies: Optional[np.ndarray] - Optional vector of node frequencies. source_node_ids: List[int] - The source nodes of the bipartite graph. + The source node ids defining a bipartite graph. destination_node_ids: List[int] - The destination nodes of the bipartite graph. + The destination node ids defining a bipartite graph. + minimum_similarity: Optional[float] = 0.0 + Minimum similarity to be kept. Values below this amount are filtered. return_similarities_dataframe: bool = False Whether to return a pandas DataFrame, which as indices has the node IDs. - By default, a numpy array with the similarities is returned as it weights much less. + By default, a numpy array with the similarities is returned as requires much less RAM. + return_node_names: bool = False + Whether to return the node names or node IDs associated to the scores. + By default we return the node ids, which require much less memory. """ - return self.get_similarities_from_graph( - graph.build_bipartite_graph_from_edge_node_ids( - source_node_ids=source_node_ids, - destination_node_ids=destination_node_ids, - directed=True + return self._normalize_output( + *self._model.get_node_ids_and_similarity_from_node_ids( + first_node_ids=source_node_ids, + second_node_ids=destination_node_ids, + minimum_similarity=minimum_similarity ), - return_similarities_dataframe=return_similarities_dataframe + return_similarities_dataframe=return_similarities_dataframe, + return_node_names=return_node_names ) - def get_similarities_from_bipartite_graph_from_edge_node_names( + def get_similarities_from_bipartite_graph_node_names( self, - graph: Graph, source_node_names: List[str], destination_node_names: List[str], - return_similarities_dataframe: bool = False - ) -> Union[pd.DataFrame, np.ndarray]: - """Execute similarities probabilities on the provided graph bipartite portion. + minimum_similarity: Optional[float] = 0.0, + return_similarities_dataframe: bool = False, + return_node_names: bool = False + ) -> Union[pd.DataFrame, Tuple[np.ndarray, np.ndarray]]: + """Execute similarities probabilities on the bipartite portion Parameters -------------------- - graph: Graph - The graph from which to extract the edges. - node_frequencies: Optional[np.ndarray] - Optional vector of node frequencies. source_node_names: List[str] - The source nodes of the bipartite graph. + The source node names defining a bipartite graph. destination_node_names: List[str] - The destination nodes of the bipartite graph. + The destination node names defining a bipartite graph. + minimum_similarity: Optional[float] = 0.0 + Minimum similarity to be kept. Values below this amount are filtered. return_similarities_dataframe: bool = False Whether to return a pandas DataFrame, which as indices has the node IDs. - By default, a numpy array with the similarities is returned as it weights much less. + By default, a numpy array with the similarities is returned as requires much less RAM. + return_node_names: bool = False + Whether to return the node names or node IDs associated to the scores. + By default we return the node ids, which require much less memory. """ - return self.get_similarities_from_graph( - graph.build_bipartite_graph_from_edge_node_names( - source_node_names=source_node_names, - destination_node_names=destination_node_names, - directed=True + return self._normalize_output( + *self._model.get_node_ids_and_similarity_from_node_names( + first_node_names=source_node_names, + second_node_names=destination_node_names, + minimum_similarity=minimum_similarity ), - return_similarities_dataframe=return_similarities_dataframe + return_similarities_dataframe=return_similarities_dataframe, + return_node_names=return_node_names ) - def get_similarities_from_bipartite_graph_from_edge_node_prefixes( + def get_similarities_from_bipartite_graph_node_prefixes( self, - graph: Graph, source_node_prefixes: List[str], destination_node_prefixes: List[str], - return_similarities_dataframe: bool = False - ) -> Union[pd.DataFrame, np.ndarray]: - """Execute similarities probabilities on the provided graph bipartite portion. + minimum_similarity: Optional[float] = 0.0, + return_similarities_dataframe: bool = False, + return_node_names: bool = False + ) -> Union[pd.DataFrame, Tuple[np.ndarray, np.ndarray]]: + """Execute similarities probabilities on the bipartite portion Parameters -------------------- - graph: Graph - The graph from which to extract the edges. - node_frequencies: Optional[np.ndarray] - Optional vector of node frequencies. source_node_prefixes: List[str] - The source node prefixes of the bipartite graph. + The source node prefixes defining a bipartite graph. destination_node_prefixes: List[str] - The destination node prefixes of the bipartite graph. + The destination node prefixes defining a bipartite graph. + minimum_similarity: Optional[float] = 0.0 + Minimum similarity to be kept. Values below this amount are filtered. return_similarities_dataframe: bool = False Whether to return a pandas DataFrame, which as indices has the node IDs. - By default, a numpy array with the similarities is returned as it weights much less. + By default, a numpy array with the similarities is returned as requires much less RAM. + return_node_names: bool = False + Whether to return the node names or node IDs associated to the scores. + By default we return the node ids, which require much less memory. """ - return self.get_similarities_from_graph( - graph.build_bipartite_graph_from_edge_node_prefixes( - source_node_prefixes=source_node_prefixes, - destination_node_prefixes=destination_node_prefixes, - directed=True + return self._normalize_output( + *self._model.get_node_ids_and_similarity_from_node_prefixes( + first_node_prefixes=source_node_prefixes, + second_node_prefixes=destination_node_prefixes, + minimum_similarity=minimum_similarity ), - return_similarities_dataframe=return_similarities_dataframe + return_similarities_dataframe=return_similarities_dataframe, + return_node_names=return_node_names ) - def get_similarities_from_bipartite_graph_from_edge_node_types( + def get_similarities_from_bipartite_graph_node_type_ids( self, - graph: Graph, - source_node_types: List[str], - destination_node_types: List[str], + source_node_type_ids: List[int], + destination_node_type_ids: List[int], + minimum_similarity: Optional[float] = 0.0, + return_similarities_dataframe: bool = False, + return_node_names: bool = False + ) -> Union[pd.DataFrame, Tuple[np.ndarray, np.ndarray]]: + """Execute similarities probabilities on the bipartite portion + + Parameters + -------------------- + source_node_type_ids: List[int] + The source node type ids defining a bipartite graph. + destination_node_type_ids: List[int] + The destination node type ids defining a bipartite graph. + minimum_similarity: Optional[float] = 0.0 + Minimum similarity to be kept. Values below this amount are filtered. return_similarities_dataframe: bool = False - ) -> Union[pd.DataFrame, np.ndarray]: + Whether to return a pandas DataFrame, which as indices has the node IDs. + By default, a numpy array with the similarities is returned as requires much less RAM. + return_node_names: bool = False + Whether to return the node names or node IDs associated to the scores. + By default we return the node ids, which require much less memory. + """ + return self._normalize_output( + *self._model.get_node_ids_and_similarity_from_node_type_ids( + first_node_type_ids=source_node_type_ids, + second_node_type_ids=destination_node_type_ids, + minimum_similarity=minimum_similarity + ), + return_similarities_dataframe=return_similarities_dataframe, + return_node_names=return_node_names + ) + + def get_similarities_from_bipartite_graph_node_type_names( + self, + source_node_type_names: List[str], + destination_node_type_names: List[str], + minimum_similarity: Optional[float] = 0.0, + return_similarities_dataframe: bool = False, + return_node_names: bool = False + ) -> Union[pd.DataFrame, Tuple[np.ndarray, np.ndarray]]: """Execute similarities probabilities on the provided graph bipartite portion. Parameters -------------------- - graph: Graph - The graph from which to extract the edges. - node_frequencies: Optional[np.ndarray] - Optional vector of node frequencies. - source_node_types: List[str] - The source node prefixes of the bipartite graph. - destination_node_types: List[str] - The destination node prefixes of the bipartite graph. + source_node_type_names: List[str] + The source node type names defining a bipartite graph. + destination_node_type_names: List[str] + The destination node type names defining a bipartite graph. + minimum_similarity: Optional[float] = 0.0 + Minimum similarity to be kept. Values below this amount are filtered. return_similarities_dataframe: bool = False Whether to return a pandas DataFrame, which as indices has the node IDs. - By default, a numpy array with the similarities is returned as it weights much less. + By default, a numpy array with the similarities is returned as requires much less RAM. + return_node_names: bool = False + Whether to return the node names or node IDs associated to the scores. + By default we return the node ids, which require much less memory. """ - return self.get_similarities_from_graph( - graph.build_bipartite_graph_from_edge_node_types( - source_node_types=source_node_types, - destination_node_types=destination_node_types, - directed=True + return self._normalize_output( + *self._model.get_node_ids_and_similarity_from_node_type_names( + first_node_type_names=source_node_type_names, + second_node_type_names=destination_node_type_names, + minimum_similarity=minimum_similarity ), - return_similarities_dataframe=return_similarities_dataframe + return_similarities_dataframe=return_similarities_dataframe, + return_node_names=return_node_names ) - def get_similarities_from_clique_graph_from_node_ids( + def get_similarities_from_clique_graph_node_ids( self, - graph: Graph, - node_ids: List[int], - return_similarities_dataframe: bool = False - ) -> Union[pd.DataFrame, np.ndarray]: - """Execute similarities probabilities on the provided graph bipartite portion. + node_ids: List[str], + minimum_similarity: Optional[float] = 0.0, + return_similarities_dataframe: bool = False, + return_node_names: bool = False + ) -> Union[pd.DataFrame, Tuple[np.ndarray, np.ndarray]]: + """Execute similarities probabilities on the clique portion Parameters -------------------- - graph: Graph - The graph from which to extract the edges. - node_frequencies: Optional[np.ndarray] - Optional vector of node frequencies. node_ids: List[int] - The nodes of the bipartite graph. + The node type ids defining a clique. + minimum_similarity: Optional[float] = 0.0 + Minimum similarity to be kept. Values below this amount are filtered. return_similarities_dataframe: bool = False Whether to return a pandas DataFrame, which as indices has the node IDs. - By default, a numpy array with the similarities is returned as it weights much less. + By default, a numpy array with the similarities is returned as requires much less RAM. + return_node_names: bool = False + Whether to return the node names or node IDs associated to the scores. + By default we return the node ids, which require much less memory. """ - return self.get_similarities_from_graph( - graph.build_clique_graph_from_node_ids( - node_ids=node_ids, - directed=True - ), - return_similarities_dataframe=return_similarities_dataframe + return self.get_similarities_from_bipartite_graph_node_ids( + source_node_ids=node_ids, + destination_node_ids=node_ids, + minimum_similarity=minimum_similarity, + return_similarities_dataframe=return_similarities_dataframe, + return_node_names=return_node_names ) - def get_similarities_from_clique_graph_from_node_names( + def get_similarities_from_clique_graph_node_names( self, - graph: Graph, node_names: List[str], - return_similarities_dataframe: bool = False - ) -> Union[pd.DataFrame, np.ndarray]: - """Execute similarities probabilities on the provided graph bipartite portion. + minimum_similarity: Optional[float] = 0.0, + return_similarities_dataframe: bool = False, + return_node_names: bool = False + ) -> Union[pd.DataFrame, Tuple[np.ndarray, np.ndarray]]: + """Execute similarities probabilities on the clique portion Parameters -------------------- - graph: Graph - The graph from which to extract the edges. - node_frequencies: Optional[np.ndarray] - Optional vector of node frequencies. node_names: List[str] - The nodes of the bipartite graph. + The node type ids defining a clique. + minimum_similarity: Optional[float] = 0.0 + Minimum similarity to be kept. Values below this amount are filtered. return_similarities_dataframe: bool = False Whether to return a pandas DataFrame, which as indices has the node IDs. - By default, a numpy array with the similarities is returned as it weights much less. + By default, a numpy array with the similarities is returned as requires much less RAM. + return_node_names: bool = False + Whether to return the node names or node IDs associated to the scores. + By default we return the node ids, which require much less memory. """ - return self.get_similarities_from_graph( - graph.build_clique_graph_from_node_names( - node_names=node_names, - directed=True - ), - return_similarities_dataframe=return_similarities_dataframe + return self.get_similarities_from_bipartite_graph_node_names( + source_node_names=node_names, + destination_node_names=node_names, + minimum_similarity=minimum_similarity, + return_similarities_dataframe=return_similarities_dataframe, + return_node_names=return_node_names ) - def get_similarities_from_clique_graph_from_node_prefixes( + def get_similarities_from_clique_graph_node_prefixes( self, - graph: Graph, node_prefixes: List[str], - return_similarities_dataframe: bool = False - ) -> Union[pd.DataFrame, np.ndarray]: - """Execute similarities probabilities on the provided graph bipartite portion. + minimum_similarity: Optional[float] = 0.0, + return_similarities_dataframe: bool = False, + return_node_names: bool = False + ) -> Union[pd.DataFrame, Tuple[np.ndarray, np.ndarray]]: + """Execute similarities probabilities on the clique portion Parameters -------------------- - graph: Graph - The graph from which to extract the edges. - node_frequencies: Optional[np.ndarray] - Optional vector of node frequencies. node_prefixes: List[str] - The node prefixes of the bipartite graph. + The node type ids defining a clique. + minimum_similarity: Optional[float] = 0.0 + Minimum similarity to be kept. Values below this amount are filtered. return_similarities_dataframe: bool = False Whether to return a pandas DataFrame, which as indices has the node IDs. - By default, a numpy array with the similarities is returned as it weights much less. + By default, a numpy array with the similarities is returned as requires much less RAM. + return_node_names: bool = False + Whether to return the node names or node IDs associated to the scores. + By default we return the node ids, which require much less memory. """ - return self.get_similarities_from_graph( - graph.build_clique_graph_from_node_prefixes( - node_prefixes=node_prefixes, - directed=True - ), - return_similarities_dataframe=return_similarities_dataframe + return self.get_similarities_from_bipartite_graph_node_prefixes( + source_node_prefixes=node_prefixes, + destination_node_prefixes=node_prefixes, + minimum_similarity=minimum_similarity, + return_similarities_dataframe=return_similarities_dataframe, + return_node_names=return_node_names ) - def get_similarities_from_clique_graph_from_node_types( + def get_similarities_from_clique_graph_node_type_ids( self, - graph: Graph, - node_types: List[str], + node_type_ids: List[int], + minimum_similarity: Optional[float] = 0.0, + return_similarities_dataframe: bool = False, + return_node_names: bool = False + ) -> Union[pd.DataFrame, Tuple[np.ndarray, np.ndarray]]: + """Execute similarities probabilities on the clique portion + + Parameters + -------------------- + node_type_ids: List[int] + The node type ids defining a clique. + minimum_similarity: Optional[float] = 0.0 + Minimum similarity to be kept. Values below this amount are filtered. return_similarities_dataframe: bool = False - ) -> Union[pd.DataFrame, np.ndarray]: - """Execute similarities probabilities on the provided graph bipartite portion. + Whether to return a pandas DataFrame, which as indices has the node IDs. + By default, a numpy array with the similarities is returned as requires much less RAM. + return_node_names: bool = False + Whether to return the node names or node IDs associated to the scores. + By default we return the node ids, which require much less memory. + """ + return self.get_similarities_from_bipartite_graph_node_type_ids( + source_node_type_ids=node_type_ids, + destination_node_type_ids=node_type_ids, + minimum_similarity=minimum_similarity, + return_similarities_dataframe=return_similarities_dataframe, + return_node_names=return_node_names + ) + + def get_similarities_from_clique_graph_node_type_names( + self, + node_type_names: List[str], + minimum_similarity: Optional[float] = 0.0, + return_similarities_dataframe: bool = False, + return_node_names: bool = False + ) -> Union[pd.DataFrame, Tuple[np.ndarray, np.ndarray]]: + """Execute similarities probabilities on the provided graph clique portion. Parameters -------------------- - graph: Graph - The graph from which to extract the edges. - node_frequencies: Optional[np.ndarray] - Optional vector of node frequencies. - node_types: List[str] - The node prefixes of the bipartite graph. + node_type_names: List[str] + The node type names defining a clique. + minimum_similarity: Optional[float] = 0.0 + Minimum similarity to be kept. Values below this amount are filtered. return_similarities_dataframe: bool = False Whether to return a pandas DataFrame, which as indices has the node IDs. - By default, a numpy array with the similarities is returned as it weights much less. + By default, a numpy array with the similarities is returned as requires much less RAM. + return_node_names: bool = False + Whether to return the node names or node IDs associated to the scores. + By default we return the node ids, which require much less memory. """ - return self.get_similarities_from_graph( - graph.build_clique_graph_from_node_types( - node_types=node_types, - directed=True - ), - return_similarities_dataframe=return_similarities_dataframe + return self.get_similarities_from_bipartite_graph_node_type_names( + source_node_type_names=node_type_names, + destination_node_type_names=node_type_names, + minimum_similarity=minimum_similarity, + return_similarities_dataframe=return_similarities_dataframe, + return_node_names=return_node_names ) diff --git a/embiggen/utils/abstract_models/abstract_classifier_model.py b/embiggen/utils/abstract_models/abstract_classifier_model.py index c131c9bb..e3548115 100644 --- a/embiggen/utils/abstract_models/abstract_classifier_model.py +++ b/embiggen/utils/abstract_models/abstract_classifier_model.py @@ -3,9 +3,13 @@ from ensmallen import Graph, express_measures import numpy as np import pandas as pd +import os +import platform import time +import json from userinput.utils import must_be_in_set from tqdm.auto import trange, tqdm +from environments_utils import get_slurm_node_id, must_be_in_slurm_node from embiggen.utils.abstract_models.list_formatting import format_list from cache_decorator import Cache @@ -156,6 +160,35 @@ def get_available_evaluation_schemas(cls) -> List[str]: "in the child classes of abstract model." )) + @classmethod + def load(cls, path: str) -> "Self": + """Load a saved version of the model from the provided path. + + Parameters + ------------------- + path: str + Path from where to load the model. + """ + raise NotImplementedError(( + f"The `load` was not implemented for the {cls.model_name()} " + f"for the {cls.task_name()} task as made available from the " + f"{cls.library_name()} library." + )) + + def dump(self, path: str): + """Dump the current model at the provided path. + + Parameters + ------------------- + path: str + Path from where to dump the model. + """ + raise NotImplementedError(( + f"The `dump` was not implemented for the {self.model_name()} " + f"for the {self.task_name()} task as made available from the " + f"{self.library_name()} library." + )) + @classmethod def normalize_node_feature( cls, @@ -165,7 +198,7 @@ def normalize_node_feature( allow_automatic_feature: bool = True, skip_evaluation_biased_feature: bool = False, smoke_test: bool = False, - precompute_constant_automatic_stocastic_features: bool = False + precompute_constant_stocastic_features: bool = False ) -> List[np.ndarray]: """Normalizes the provided node features and validates them. @@ -190,7 +223,7 @@ def normalize_node_feature( Whether this run should be considered a smoke test and therefore use the smoke test configurations for the provided model names and feature names. - precompute_constant_automatic_stocastic_features: bool = False + precompute_constant_stocastic_features: bool = False Whether to precompute once the constant automatic stocastic features before starting the embedding loop. This means that, when left set to false, while the features will be computed @@ -230,7 +263,7 @@ def normalize_node_feature( cls.task_involves_edge_weights() and node_feature.can_use_edge_weights() and node_feature.is_using_edge_weights() or cls.task_involves_topology() and node_feature.is_topological() ) or - not precompute_constant_automatic_stocastic_features and node_feature.is_stocastic() + not precompute_constant_stocastic_features and node_feature.is_stocastic() ): yield node_feature return None @@ -298,7 +331,7 @@ def normalize_node_features( allow_automatic_feature: bool = True, skip_evaluation_biased_feature: bool = False, smoke_test: bool = False, - precompute_constant_automatic_stocastic_features: bool = False + precompute_constant_stocastic_features: bool = False ) -> List[np.ndarray]: """Normalizes the provided node features and validates them. @@ -323,7 +356,7 @@ def normalize_node_features( Whether this run should be considered a smoke test and therefore use the smoke test configurations for the provided model names and feature names. - precompute_constant_automatic_stocastic_features: bool = False + precompute_constant_stocastic_features: bool = False Whether to precompute once the constant automatic stocastic features before starting the embedding loop. This means that, when left set to false, while the features will be computed @@ -349,7 +382,7 @@ def normalize_node_features( allow_automatic_feature=allow_automatic_feature, skip_evaluation_biased_feature=skip_evaluation_biased_feature, smoke_test=smoke_test, - precompute_constant_automatic_stocastic_features=precompute_constant_automatic_stocastic_features + precompute_constant_stocastic_features=precompute_constant_stocastic_features ) ] @@ -362,7 +395,7 @@ def normalize_node_type_feature( allow_automatic_feature: bool = True, skip_evaluation_biased_feature: bool = False, smoke_test: bool = False, - precompute_constant_automatic_stocastic_features: bool = False + precompute_constant_stocastic_features: bool = False ) -> List[np.ndarray]: """Normalizes the provided node type features and validates them. @@ -387,7 +420,7 @@ def normalize_node_type_feature( Whether this run should be considered a smoke test and therefore use the smoke test configurations for the provided model names and feature names. - precompute_constant_automatic_stocastic_features: bool = False + precompute_constant_stocastic_features: bool = False Whether to precompute once the constant automatic stocastic features before starting the embedding loop. This means that, when left set to false, while the features will be computed @@ -427,11 +460,11 @@ def normalize_node_type_feature( cls.task_involves_edge_weights() and node_type_feature.is_using_edge_weights() or cls.task_involves_topology() and node_type_feature.is_topological() ) or - not precompute_constant_automatic_stocastic_features and node_type_feature.is_stocastic() + not precompute_constant_stocastic_features and node_type_feature.is_stocastic() ): yield node_type_feature return None - + if node_type_feature.is_stocastic(): node_type_feature.set_random_state(random_state) @@ -495,7 +528,7 @@ def normalize_node_type_features( allow_automatic_feature: bool = True, skip_evaluation_biased_feature: bool = False, smoke_test: bool = False, - precompute_constant_automatic_stocastic_features: bool = False + precompute_constant_stocastic_features: bool = False ) -> List[np.ndarray]: """Normalizes the provided node type features and validates them. @@ -520,7 +553,7 @@ def normalize_node_type_features( Whether this run should be considered a smoke test and therefore use the smoke test configurations for the provided model names and feature names. - precompute_constant_automatic_stocastic_features: bool = False + precompute_constant_stocastic_features: bool = False Whether to precompute once the constant automatic stocastic features before starting the embedding loop. This means that, when left set to false, while the features will be computed @@ -546,7 +579,7 @@ def normalize_node_type_features( allow_automatic_feature=allow_automatic_feature, skip_evaluation_biased_feature=skip_evaluation_biased_feature, smoke_test=smoke_test, - precompute_constant_automatic_stocastic_features=precompute_constant_automatic_stocastic_features + precompute_constant_stocastic_features=precompute_constant_stocastic_features ) ] @@ -559,7 +592,7 @@ def normalize_edge_feature( allow_automatic_feature: bool = True, skip_evaluation_biased_feature: bool = False, smoke_test: bool = False, - precompute_constant_automatic_stocastic_features: bool = False + precompute_constant_stocastic_features: bool = False ) -> List[np.ndarray]: """Normalizes the provided edge features and validates them. @@ -584,7 +617,7 @@ def normalize_edge_feature( Whether this run should be considered a smoke test and therefore use the smoke test configurations for the provided model names and feature names. - precompute_constant_automatic_stocastic_features: bool = False + precompute_constant_stocastic_features: bool = False Whether to precompute once the constant automatic stocastic features before starting the embedding loop. This means that, when left set to false, while the features will be computed @@ -624,7 +657,7 @@ def normalize_edge_feature( cls.task_involves_edge_weights() and edge_feature.is_using_edge_weights() or cls.task_involves_topology() and edge_feature.is_topological() ) or - not precompute_constant_automatic_stocastic_features and edge_feature.is_stocastic() + not precompute_constant_stocastic_features and edge_feature.is_stocastic() ): yield edge_feature return None @@ -692,7 +725,7 @@ def normalize_edge_features( allow_automatic_feature: bool = True, skip_evaluation_biased_feature: bool = False, smoke_test: bool = False, - precompute_constant_automatic_stocastic_features: bool = False + precompute_constant_stocastic_features: bool = False ) -> List[np.ndarray]: """Normalizes the provided edge features and validates them. @@ -717,7 +750,7 @@ def normalize_edge_features( Whether this run should be considered a smoke test and therefore use the smoke test configurations for the provided model names and feature names. - precompute_constant_automatic_stocastic_features: bool = False + precompute_constant_stocastic_features: bool = False Whether to precompute once the constant automatic stocastic features before starting the embedding loop. This means that, when left set to false, while the features will be computed @@ -743,7 +776,7 @@ def normalize_edge_features( allow_automatic_feature=allow_automatic_feature, skip_evaluation_biased_feature=skip_evaluation_biased_feature, smoke_test=smoke_test, - precompute_constant_automatic_stocastic_features=precompute_constant_automatic_stocastic_features + precompute_constant_stocastic_features=precompute_constant_stocastic_features ) ] @@ -867,7 +900,7 @@ def predict( if support is None: support = graph - + try: predictions = self._predict( graph=graph, @@ -897,7 +930,7 @@ def predict( f"implemented using the {self.library_name()} for the {self.task_name()} task. " f"Specifically, the class of the model is {self.__class__.__name__}." ) from e - + return predictions def predict_proba( @@ -982,7 +1015,7 @@ def predict_proba( "returned a vector of prediction probabilities with " f"shape {predictions.shape}." ) - + return predictions def evaluate_predictions( @@ -1230,25 +1263,35 @@ def iterate_classifier_models( cache_path="{cache_dir}/{self.task_name()}/{graph.get_name()}/holdout_{holdout_number}/{self.model_name()}/{self.library_name()}/{_hash}.csv.gz", cache_dir="experiments", enable_cache_arg_name="enable_cache", - args_to_ignore=["verbose", "smoke_test", "train_of_interest", "test_of_interest", "train",], + args_to_ignore=[ + "verbose", + "smoke_test", + "train_of_interest", + "test_of_interest", + "train", + "metadata" + ], capture_enable_cache_arg_name=True, use_approximated_hash=True ) - def __train_and_evaluate_model( + def _train_and_evaluate_model( self, graph: Graph, train_of_interest: Graph, test_of_interest: Graph, train: Graph, subgraph_of_interest: Graph, + use_subgraph_as_support: bool, node_features: Optional[List[np.ndarray]], node_type_features: Optional[List[np.ndarray]], edge_features: Optional[List[np.ndarray]], random_state: int, holdout_number: int, evaluation_schema: str, - automatic_features_names: List[str], - automatic_features_parameters: Dict[str, Any], + holdouts_kwargs: Dict[str, Any], + features_names: List[str], + features_parameters: Dict[str, Any], + metadata: Dict[str, Any], **validation_kwargs ) -> pd.DataFrame: """Run inner training and evaluation.""" @@ -1258,7 +1301,7 @@ def __train_and_evaluate_model( training_start = time.time() self.fit( graph=train_of_interest, - support=train, + support=train_of_interest if use_subgraph_as_support else train, node_features=node_features, node_type_features=node_type_features, edge_features=edge_features @@ -1271,7 +1314,7 @@ def __train_and_evaluate_model( # We add the newly computed performance. model_performance = pd.DataFrame(self._evaluate( graph=graph, - support=train, + support=train_of_interest if use_subgraph_as_support else train, train=train_of_interest, test=test_of_interest, node_features=node_features, @@ -1281,7 +1324,7 @@ def __train_and_evaluate_model( random_state=random_state * holdout_number, verbose=False, **validation_kwargs - )) + )).reset_index(drop=True) except RuntimeError as e: raise e except Exception as e: @@ -1302,26 +1345,34 @@ def __train_and_evaluate_model( model_performance["nodes_number"] = graph.get_number_of_nodes() model_performance["edges_number"] = graph.get_number_of_directed_edges() model_performance["evaluation_schema"] = evaluation_schema + model_performance["holdout_number"] = holdout_number + model_performance["holdouts_kwargs"] = json.dumps(holdouts_kwargs) + model_performance["use_subgraph_as_support"] = use_subgraph_as_support + + for column_name, column_value in metadata.items(): + model_performance[column_name] = column_value + + df_model_parameters = pd.DataFrame( + dict(), index=model_performance.index) + df_features_parameters = pd.DataFrame( + dict(), index=model_performance.index) for parameter_name, parameter_value in self.parameters().items(): - if ("Model", parameter_name) in model_performance.columns: - raise ValueError( - "There has been a collision between the column names used in " - "the model performance report and the parameter names " - f" of one of the classifiers {self.model_name()}." - f"The parameter that has caused the collision is {parameter}. " - "Please do change the name of the parameter in your model." - ) if isinstance(parameter_value, (list, tuple)): parameter_value = str(parameter_value) - model_performance[("Model", parameter_name)] = parameter_value - - model_performance["automatic_features_names"] = format_list( - automatic_features_names + df_model_parameters[parameter_name] = parameter_value + + df_model_parameters.columns = [ + ["model_parameters"] * len(df_model_parameters.columns), + df_model_parameters.columns + ] + + model_performance["features_names"] = format_list( + features_names ) - - for parameter, value in enumerate(automatic_features_parameters.items()): - if parameter in model_performance.columns: + + for parameter, value in features_parameters.items(): + if parameter in df_features_parameters.columns: raise ValueError( "There has been a collision between the parameters used in " "one of the embedding models and the parameter " @@ -1329,7 +1380,21 @@ def __train_and_evaluate_model( f"The parameter that has caused the collision is {parameter}. " "Please do change the name of the parameter in your model." ) - model_performance[parameter] = str(value) + df_features_parameters[parameter] = str(value) + + df_features_parameters.columns = [ + ["features_parameters"] * len(df_features_parameters.columns), + df_features_parameters.columns + ] + + model_performance = pd.concat( + [ + model_performance, + df_model_parameters, + df_features_parameters + ], + axis=1 + ) return model_performance @@ -1338,16 +1403,22 @@ def __train_and_evaluate_model( cache_path="{cache_dir}/{cls.task_name()}/{graph.get_name()}/holdout_{holdout_number}/{_hash}.csv.gz", cache_dir="experiments", enable_cache_arg_name="enable_cache", - args_to_ignore=["verbose", "smoke_test", "number_of_holdouts"], + args_to_ignore=[ + "verbose", + "smoke_test", + "number_of_holdouts", + "metadata" + ], capture_enable_cache_arg_name=False, use_approximated_hash=True ) - def __evaluate_on_single_holdout( + def _evaluate_on_single_holdout( cls, models: Union[Type["AbstractClassifierModel"], List[Type["AbstractClassifierModel"]]], library_names: Optional[Union[str, List[str]]], graph: Graph, subgraph_of_interest: Graph, + use_subgraph_as_support: bool, node_features: Optional[List[np.ndarray]], node_type_features: Optional[List[np.ndarray]], edge_features: Optional[List[np.ndarray]], @@ -1359,8 +1430,9 @@ def __evaluate_on_single_holdout( smoke_test: bool, holdouts_kwargs: Dict[str, Any], subgraph_of_interest_has_compatible_nodes: Optional[bool], - automatic_features_names: List[str], - automatic_features_parameters: Dict[str, Any], + features_names: List[str], + features_parameters: Dict[str, Any], + metadata: Dict[str, Any], verbose: bool, **validation_kwargs ) -> pd.DataFrame: @@ -1384,14 +1456,9 @@ def __evaluate_on_single_holdout( vector_cumulative_node_degrees=graph.has_cumulative_node_degrees_tradeoff_enabled(), vector_reciprocal_sqrt_degrees=graph.has_reciprocal_sqrt_degrees_tradeoff_enabled() ) - test.enable( - vector_sources=graph.has_sources_tradeoff_enabled(), - vector_destinations=graph.has_destinations_tradeoff_enabled(), - vector_cumulative_node_degrees=graph.has_cumulative_node_degrees_tradeoff_enabled(), - vector_reciprocal_sqrt_degrees=graph.has_reciprocal_sqrt_degrees_tradeoff_enabled() - ) # We compute the remaining features + starting_to_compute_node_features = time.time() holdout_node_features = cls.normalize_node_features( train, random_state=random_state*(holdout_number+1), @@ -1399,10 +1466,13 @@ def __evaluate_on_single_holdout( allow_automatic_feature=True, skip_evaluation_biased_feature=False, smoke_test=smoke_test, - precompute_constant_automatic_stocastic_features=True + precompute_constant_stocastic_features=True ) + time_required_to_compute_node_features = time.time() - \ + starting_to_compute_node_features # We compute the remaining features + starting_to_compute_node_type_features = time.time() holdout_node_type_features = cls.normalize_node_type_features( train, random_state=random_state*(holdout_number+1), @@ -1410,12 +1480,15 @@ def __evaluate_on_single_holdout( allow_automatic_feature=True, skip_evaluation_biased_feature=False, smoke_test=smoke_test, - precompute_constant_automatic_stocastic_features=True + precompute_constant_stocastic_features=True ) + time_required_to_compute_node_type_features = time.time( + ) - starting_to_compute_node_type_features # We execute the same thing as described above, # but now for the edge features instead that for # the node features. + starting_to_compute_edge_features = time.time() holdout_edge_features = cls.normalize_edge_features( train, random_state=random_state*(holdout_number+1), @@ -1423,8 +1496,10 @@ def __evaluate_on_single_holdout( allow_automatic_feature=True, skip_evaluation_biased_feature=False, smoke_test=smoke_test, - precompute_constant_automatic_stocastic_features=True + precompute_constant_stocastic_features=True ) + time_required_to_compute_edge_features = time.time() - \ + starting_to_compute_edge_features if subgraph_of_interest is not None: # First we align the train and test graph to have @@ -1439,22 +1514,6 @@ def __evaluate_on_single_holdout( node_names_to_keep_from_graph=subgraph_of_interest ) - # We enable in the train and test graphs of interest the same - # speedups enabled in the provided graph. - train.enable( - vector_sources=graph.has_sources_tradeoff_enabled(), - vector_destinations=graph.has_destinations_tradeoff_enabled(), - vector_cumulative_node_degrees=graph.has_cumulative_node_degrees_tradeoff_enabled(), - vector_reciprocal_sqrt_degrees=graph.has_reciprocal_sqrt_degrees_tradeoff_enabled() - ) - - test.enable( - vector_sources=graph.has_sources_tradeoff_enabled(), - vector_destinations=graph.has_destinations_tradeoff_enabled(), - vector_cumulative_node_degrees=graph.has_cumulative_node_degrees_tradeoff_enabled(), - vector_reciprocal_sqrt_degrees=graph.has_reciprocal_sqrt_degrees_tradeoff_enabled() - ) - # We adjust the node features to only include the node features # that the subgraph of interest allows us to use. if holdout_node_features is not None: @@ -1493,25 +1552,26 @@ def __evaluate_on_single_holdout( "subgraph of interest, does not have any more edges which are " f"essential when running a {cls.task_name()} task." ) - # We enable in the train and test graphs of interest the same - # speedups enabled in the provided graph. - train_of_interest.enable( - vector_sources=graph.has_sources_tradeoff_enabled(), - vector_destinations=graph.has_destinations_tradeoff_enabled(), - vector_cumulative_node_degrees=graph.has_cumulative_node_degrees_tradeoff_enabled(), - vector_reciprocal_sqrt_degrees=graph.has_reciprocal_sqrt_degrees_tradeoff_enabled() - ) - - test_of_interest.enable( - vector_sources=graph.has_sources_tradeoff_enabled(), - vector_destinations=graph.has_destinations_tradeoff_enabled(), - vector_cumulative_node_degrees=graph.has_cumulative_node_degrees_tradeoff_enabled(), - vector_reciprocal_sqrt_degrees=graph.has_reciprocal_sqrt_degrees_tradeoff_enabled() - ) else: train_of_interest = train test_of_interest = test + # We enable in the train and test graphs of interest the same + # speedups enabled in the provided graph. + train_of_interest.enable( + vector_sources=graph.has_sources_tradeoff_enabled(), + vector_destinations=graph.has_destinations_tradeoff_enabled(), + vector_cumulative_node_degrees=graph.has_cumulative_node_degrees_tradeoff_enabled(), + vector_reciprocal_sqrt_degrees=graph.has_reciprocal_sqrt_degrees_tradeoff_enabled() + ) + + test_of_interest.enable( + vector_sources=graph.has_sources_tradeoff_enabled(), + vector_destinations=graph.has_destinations_tradeoff_enabled(), + vector_cumulative_node_degrees=graph.has_cumulative_node_degrees_tradeoff_enabled(), + vector_reciprocal_sqrt_degrees=graph.has_reciprocal_sqrt_degrees_tradeoff_enabled() + ) + additional_validation_kwargs = cls._prepare_evaluation( graph=graph, support=train, @@ -1525,22 +1585,33 @@ def __evaluate_on_single_holdout( time_required_for_setting_up_holdout = time.time() - starting_setting_up_holdout + metadata = dict( + **metadata, + time_required_for_setting_up_holdout=time_required_for_setting_up_holdout, + time_required_to_compute_node_features=time_required_to_compute_node_features, + time_required_to_compute_node_type_features=time_required_to_compute_node_type_features, + time_required_to_compute_edge_features=time_required_to_compute_edge_features + ) + holdout_performance = pd.concat([ - classifier.__train_and_evaluate_model( + classifier._train_and_evaluate_model( graph=graph, train_of_interest=train_of_interest, test_of_interest=test_of_interest, train=train, subgraph_of_interest=subgraph_of_interest, + use_subgraph_as_support=use_subgraph_as_support, node_features=holdout_node_features, node_type_features=holdout_node_type_features, edge_features=holdout_edge_features, random_state=random_state, holdout_number=holdout_number, evaluation_schema=evaluation_schema, + holdouts_kwargs=holdouts_kwargs, enable_cache=enable_cache, - automatic_features_names=automatic_features_names, - automatic_features_parameters=automatic_features_parameters, + features_names=features_names, + features_parameters=features_parameters, + metadata=metadata.copy(), **additional_validation_kwargs, **validation_kwargs, ) @@ -1551,17 +1622,15 @@ def __evaluate_on_single_holdout( ) ]) - holdout_performance["time_required_for_setting_up_holdout"] = time_required_for_setting_up_holdout - return holdout_performance @classmethod @Cache( cache_path="{cache_dir}/{cls.task_name()}/{graph.get_name()}/{_hash}.csv.gz", cache_dir="experiments", - enable_cache_arg_name="enable_cache", + enable_cache_arg_name="enable_top_layer_cache", args_to_ignore=["verbose", "smoke_test"], - capture_enable_cache_arg_name=False, + capture_enable_cache_arg_name=True, use_approximated_hash=True ) def evaluate( @@ -1575,12 +1644,15 @@ def evaluate( node_type_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, edge_features: Optional[Union[str, pd.DataFrame, np.ndarray, List[Union[str, pd.DataFrame, np.ndarray]]]] = None, subgraph_of_interest: Optional[Graph] = None, + use_subgraph_as_support: bool = False, number_of_holdouts: int = 10, random_state: int = 42, verbose: bool = True, enable_cache: bool = False, - precompute_constant_automatic_stocastic_features: bool = False, + precompute_constant_stocastic_features: bool = False, smoke_test: bool = False, + number_of_slurm_nodes: Optional[int] = None, + slurm_node_id_variable: str = "SLURM_GRAPE_ID", **validation_kwargs: Dict ) -> pd.DataFrame: """Execute evaluation on the provided graph. @@ -1606,6 +1678,11 @@ def evaluate( The edge features to use. subgraph_of_interest: Optional[Graph] = None Optional subgraph where to focus the task. + This is applied to the train and test graph + after the desired holdout schema is applied. + use_subgraph_as_support: bool = False + Whether to use the provided subgraph as support or + to use the train graph (not filtered by the subgraph). skip_evaluation_biased_feature: bool = False Whether to skip feature names that are known to be biased when running an holdout. These features should be computed @@ -1618,7 +1695,7 @@ def evaluate( Whether to show a loading bar while computing holdouts. enable_cache: bool = False Whether to enable the cache. - precompute_constant_automatic_stocastic_features: bool = False + precompute_constant_stocastic_features: bool = False Whether to precompute once the constant automatic stocastic features before starting the embedding loop. This means that, when left set to false, while the features will be computed @@ -1631,6 +1708,12 @@ def evaluate( Whether this run should be considered a smoke test and therefore use the smoke test configurations for the provided model names and feature names. + number_of_slurm_nodes: Optional[int] = None + Number of SLURM nodes to consider as available. + This variable is used to parallelize the holdouts accordingly. + slurm_node_id_variable: str = "SLURM_GRAPE_ID" + Name of the system variable to use as SLURM node id. + It must be set in the slurm bash script. **validation_kwargs: Dict kwargs to be forwarded to the model `_evaluate` method. """ @@ -1667,11 +1750,41 @@ def evaluate( subgraph_of_interest ) else: + if use_subgraph_as_support: + raise ValueError( + "No subgraph of interest was provided but " + "it has been requested to use the subgraph " + "of interest as support. It is not clear " + "how to proceed." + ) subgraph_of_interest_has_compatible_nodes = None + if number_of_slurm_nodes is not None: + must_be_in_slurm_node() + if not isinstance(number_of_slurm_nodes, int) or number_of_slurm_nodes <= 0: + raise ValueError( + "The number of SLURM nodes must be a positive integer value." + ) + if number_of_holdouts > number_of_slurm_nodes: + raise ValueError( + ( + "Please be advised that you are currently running an excessive " + "parametrization of the SLURM cluster. We currently can only parallelize " + "the execution of different holdouts. " + "The number of holdouts requested are {number_of_holdouts} but you are " + "currently using {number_of_slurm_nodes} SLURM nodes! " + "Possibly, you are currently running a task such as a grid search " + "and therefore you intend us to parallelize only the sub-segment of SLURM " + "nodes necessary to run the holdouts." + ).format( + number_of_holdouts=number_of_holdouts, + number_of_slurm_nodes=number_of_slurm_nodes + ) + ) + # Retrieve the set of provided automatic features parameters # so we can put them in the report. - automatic_features_parameters = { + features_parameters = { parameter_name: value for features in ( node_features @@ -1691,7 +1804,7 @@ def evaluate( # Retrieve the set of provided automatic features names # so we can put them in the report. - automatic_features_names = { + features_names = list({ feature.model_name() for features in ( node_features @@ -1706,7 +1819,7 @@ def evaluate( ) for feature in features if issubclass(feature.__class__, AbstractEmbeddingModel) - } + }) # We normalize and/or compute the node features, having # the care of skipping the features that induce bias when @@ -1714,49 +1827,91 @@ def evaluate( # This way we compute only once the features that do not # cause biases for this task, while recomputing those # that cause biases at each holdout, avoiding said biases. + starting_to_compute_constant_node_features = time.time() node_features = cls.normalize_node_features( graph, random_state=random_state, node_features=node_features, allow_automatic_feature=True, skip_evaluation_biased_feature=True, - precompute_constant_automatic_stocastic_features=precompute_constant_automatic_stocastic_features, + precompute_constant_stocastic_features=precompute_constant_stocastic_features, smoke_test=smoke_test ) + time_required_to_compute_constant_node_features = time.time( + ) - starting_to_compute_constant_node_features # We execute the same thing as described above, # but now for the node type features instead that for # the node features. + starting_to_compute_constant_node_type_features = time.time() node_type_features = cls.normalize_node_type_features( graph, random_state=random_state, node_type_features=node_type_features, allow_automatic_feature=True, skip_evaluation_biased_feature=True, - precompute_constant_automatic_stocastic_features=precompute_constant_automatic_stocastic_features, + precompute_constant_stocastic_features=precompute_constant_stocastic_features, smoke_test=smoke_test ) + time_required_to_compute_constant_node_type_features = time.time( + ) - starting_to_compute_constant_node_type_features # We execute the same thing as described above, # but now for the edge features instead that for # the node features. + starting_to_compute_constant_edge_features = time.time() edge_features = cls.normalize_edge_features( graph, random_state=random_state, edge_features=edge_features, allow_automatic_feature=True, skip_evaluation_biased_feature=True, - precompute_constant_automatic_stocastic_features=precompute_constant_automatic_stocastic_features, + precompute_constant_stocastic_features=precompute_constant_stocastic_features, smoke_test=smoke_test ) + time_required_to_compute_constant_edge_features = time.time( + ) - starting_to_compute_constant_edge_features + + metadata = dict( + number_of_threads=os.cpu_count(), + python_version=platform.python_version(), + platform=platform.platform(), + number_of_holdouts=number_of_holdouts, + number_of_slurm_nodes=number_of_slurm_nodes, + time_required_to_compute_constant_node_features=time_required_to_compute_constant_node_features, + time_required_to_compute_constant_node_type_features=time_required_to_compute_constant_node_type_features, + time_required_to_compute_constant_edge_features=time_required_to_compute_constant_edge_features, + ) + + if number_of_slurm_nodes is not None: + if slurm_node_id_variable not in os.environ: + raise ValueError( + ( + "Please do be advised that you have not provided " + "the {slurm_node_id_variable} variable but you have provided specified that " + "we should parallelize the holdouts across {number_of_slurm_nodes} nodes. " + "Please do make available this variable in the " + "slurm bash script by using the `export` flag " + "in a script similar to the following:\n" + "`srun --export=ALL,{slurm_node_id_variable}=$node_id` python3 your_script.py\n" + "You can learn more about this in the library tutorials." + ).format( + slurm_node_id_variable=slurm_node_id_variable, + number_of_slurm_nodes=number_of_slurm_nodes + ) + ) + slurm_node_id = int(os.environ[slurm_node_id_variable]) + metadata["slurm_node_id"] = slurm_node_id + metadata["number_of_slurm_nodes"] = number_of_slurm_nodes # We start to iterate on the holdouts. performance = pd.concat([ - cls.__evaluate_on_single_holdout( + cls._evaluate_on_single_holdout( models=models, library_names=library_names, graph=graph, subgraph_of_interest=subgraph_of_interest, + use_subgraph_as_support=use_subgraph_as_support, node_features=node_features, node_type_features=node_type_features, edge_features=edge_features, @@ -1768,31 +1923,35 @@ def evaluate( smoke_test=smoke_test, holdouts_kwargs=holdouts_kwargs, subgraph_of_interest_has_compatible_nodes=subgraph_of_interest_has_compatible_nodes, - verbose=verbose, - automatic_features_names=automatic_features_names, - automatic_features_parameters=automatic_features_parameters, + verbose=verbose and (number_of_slurm_nodes is None or slurm_node_id==0), + features_names=features_names, + features_parameters=features_parameters, + metadata=metadata.copy(), **validation_kwargs ) for holdout_number in trange( number_of_holdouts, - disable=not verbose, + disable=not (verbose and (number_of_slurm_nodes is None or slurm_node_id==0)), leave=False, dynamic_ncols=True, desc=f"Evaluating on {graph.get_name()}" ) + if ( + number_of_slurm_nodes is None or + ( + # We need to also mode the number of SLURM node IDs + # because the user may be parallelizing across many + # diffent nodes in contexts such as wide grid searches. + slurm_node_id % number_of_slurm_nodes + ) == ( + # We need to mode the holdout number as the number + # of holdouts may exceed the number of available SLURM + # nodes that the user has made available to this pipeline. + holdout_number % number_of_slurm_nodes + ) + ) ]) - # Adding to the report the informations relative to - # the whole validation run, which are NOT necessary - # to make unique the cache hash of the single holdouts - # or the single models. - # Be extremely weary of adding informations at this - # high level, as often the best place should be - # in the core loop, where the actual model is trained, - # as often such information changes how the model - # is trained. - performance["number_of_holdouts"] = number_of_holdouts - # We save the constant values for this model # execution. return performance diff --git a/embiggen/utils/abstract_models/abstract_embedding_model.py b/embiggen/utils/abstract_models/abstract_embedding_model.py index 6f4d7a00..24f22542 100644 --- a/embiggen/utils/abstract_models/abstract_embedding_model.py +++ b/embiggen/utils/abstract_models/abstract_embedding_model.py @@ -3,6 +3,7 @@ from ensmallen import Graph from ensmallen.datasets import get_dataset import warnings +from ringbell import RingBell from cache_decorator import Cache from embiggen.utils.abstract_models.abstract_model import AbstractModel, abstract_class from embiggen.utils.abstract_models.embedding_result import EmbeddingResult @@ -16,6 +17,7 @@ def __init__( self, embedding_size: int, enable_cache: bool = False, + ring_bell: bool = False, random_state: Optional[int] = None ): """Create new embedding model. @@ -27,6 +29,8 @@ def __init__( enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. + ring_bell: bool = False + Whether to play a sound when embedding completes. random_state: Optional[int] = None The random state to use if the model is stocastic. """ @@ -38,6 +42,10 @@ def __init__( ) self._embedding_size = embedding_size self._enable_cache = enable_cache + self._ring_bell = RingBell( + verbose=ring_bell, + sample="positive_notification" + ) def parameters(self) -> Dict[str, Any]: """Returns parameters of the embedding model.""" @@ -189,6 +197,8 @@ def _cached_fit_transform( f"but returns an object of type {type(result)}." ) + self._ring_bell.play() + return result def fit_transform( @@ -227,6 +237,18 @@ def fit_transform( repository=repository, version=version )() + if return_dataframe and graph.get_number_of_nodes() > 100_000_000: + raise ValueError( + ( + "We cowardly refuse to execute this embedding with the " + "added requirement to also return the dataframe version " + "of this graph. This graph has {number_of_nodes}, and " + "creating a Dataframe would most likely cause an OOM on " + "your system." + ).format( + number_of_nodes=graph.get_number_of_nodes() + ) + ) return self._cached_fit_transform( graph=graph, return_dataframe=return_dataframe, diff --git a/embiggen/utils/abstract_models/embedding_result.py b/embiggen/utils/abstract_models/embedding_result.py index 94e6e490..4dd922f4 100644 --- a/embiggen/utils/abstract_models/embedding_result.py +++ b/embiggen/utils/abstract_models/embedding_result.py @@ -55,6 +55,7 @@ def __init__( if embedding_list is None: continue for embedding in embedding_list: + if not isinstance(embedding, (np.ndarray, pd.DataFrame)): raise ValueError( f"One of the provided {embedding_list_name} " @@ -67,6 +68,8 @@ def __init__( f"computed with the {embedding_method_name} method " "is empty." ) + if embedding.shape[0] > 1_000_000: + continue if isinstance(embedding, pd.DataFrame): numpy_embedding = embedding.to_numpy() diff --git a/embiggen/utils/abstract_models/model_stub.py b/embiggen/utils/abstract_models/model_stub.py index cfaa7f75..a18de17a 100644 --- a/embiggen/utils/abstract_models/model_stub.py +++ b/embiggen/utils/abstract_models/model_stub.py @@ -91,7 +91,7 @@ def get_model_or_stub( # If effectively the error is that we cannot load the desired # library name, we catch this and re-raise it. if any( - f"No module named '{module_library_name}'" == str(e) + str(e).startswith(f"No module named '{module_library_name}") for module_library_name in module_library_names ): class StubClass(parent_class): diff --git a/embiggen/utils/pipeline.py b/embiggen/utils/pipeline.py index 2a50c728..72fd468f 100644 --- a/embiggen/utils/pipeline.py +++ b/embiggen/utils/pipeline.py @@ -95,7 +95,7 @@ def iterate_graphs( ): if isinstance(graph, str): graph = get_dataset( - name=graph, + graph_name=graph, repository=repository, version=version )() @@ -116,13 +116,16 @@ def classification_evaluation_pipeline( library_names: Optional[Union[str, List[str]]] = None, graph_callback: Optional[Callable[[Graph], Graph]] = None, subgraph_of_interest: Optional[Graph] = None, + use_subgraph_as_support: bool = False, number_of_holdouts: int = 10, random_state: int = 42, repositories: Optional[Union[str, List[str]]] = None, versions: Optional[Union[str, List[str]]] = None, enable_cache: bool = False, - precompute_constant_automatic_stocastic_features: bool = False, + precompute_constant_stocastic_features: bool = False, smoke_test: bool = False, + number_of_slurm_nodes: Optional[int] = None, + slurm_node_id_variable: str = "SLURM_GRAPE_ID", **evaluation_kwargs ) -> pd.DataFrame: """Execute classification pipeline for all provided models and graphs. @@ -151,7 +154,12 @@ def classification_evaluation_pipeline( For instance this may be used for filtering the uncertain edges in graphs such as STRING PPIs. subgraph_of_interest: Optional[Graph] = None - The subgraph of interest to focus the task on. + Optional subgraph where to focus the task. + This is applied to the train and test graph + after the desired holdout schema is applied. + use_subgraph_as_support: bool = False + Whether to use the provided subgraph as support or + to use the train graph (not filtered by the subgraph). number_of_holdouts: int = 10 The number of holdouts to execute. random_state: int = 42 @@ -163,7 +171,7 @@ def classification_evaluation_pipeline( Graph versions to retrieve. enable_cache: bool = False Whether to enable the cache. - precompute_constant_automatic_stocastic_features: bool = False + precompute_constant_stocastic_features: bool = False Whether to precompute once the constant automatic stocastic features before starting the embedding loop. This means that, when left set to false, while the features will be computed @@ -177,9 +185,16 @@ def classification_evaluation_pipeline( and therefore use the smoke test configurations for the provided model names and feature names. This parameter will also turn off the cache. + number_of_slurm_nodes: Optional[int] = None + Number of SLURM nodes to consider as available. + This variable is used to parallelize the holdouts accordingly. + slurm_node_id_variable: str = "SLURM_GRAPE_ID" + Name of the system variable to use as SLURM node id. + It must be set in the slurm bash script. **evaluation_kwargs: Dict Keyword arguments to forward to evaluation. """ + enable_cache = enable_cache and not smoke_test return pd.concat([ expected_parent_class.evaluate( models=models, @@ -191,11 +206,22 @@ def classification_evaluation_pipeline( node_type_features=node_type_features, edge_features=edge_features, subgraph_of_interest=subgraph_of_interest, + use_subgraph_as_support=use_subgraph_as_support, number_of_holdouts=number_of_holdouts, random_state=random_state, - enable_cache=enable_cache and not smoke_test, - precompute_constant_automatic_stocastic_features=precompute_constant_automatic_stocastic_features, + enable_cache=enable_cache, + # We need this second layer of cache handled separately as + # the different SLURM nodes will try to write simultaneously + # on the same cache files. We could add as additional cache seeds + # also the SLURM node ids and avoid this issue, but then the cache + # would only be valid for that specific cluster and it would + # not be possible to reuse it in other settings such as during + # a reproduction using cache on a notebook. + enable_top_layer_cache=enable_cache and number_of_slurm_nodes is None, + precompute_constant_stocastic_features=precompute_constant_stocastic_features, smoke_test=smoke_test, + number_of_slurm_nodes=number_of_slurm_nodes, + slurm_node_id_variable=slurm_node_id_variable, **evaluation_kwargs ) for graph in iterate_graphs( diff --git a/embiggen/visualizations/graph_visualizer.py b/embiggen/visualizations/graph_visualizer.py index f26cd77a..2043217b 100644 --- a/embiggen/visualizations/graph_visualizer.py +++ b/embiggen/visualizations/graph_visualizer.py @@ -27,9 +27,10 @@ from sklearn.decomposition import PCA from userinput.utils import must_be_in_set from sklearn.metrics import balanced_accuracy_score -from sklearn.model_selection import ShuffleSplit +from sklearn.model_selection import ShuffleSplit, StratifiedShuffleSplit import itertools from embiggen.utils.abstract_models.abstract_embedding_model import AbstractEmbeddingModel +from ensmallen.datasets.graph_retrieval import normalize_node_name from embiggen.utils.abstract_models.embedding_result import EmbeddingResult @@ -95,7 +96,7 @@ def __init__( fps: int = 24, node_embedding_method_name: str = "auto", edge_embedding_method: str = "Concatenate", - minimum_node_degree: int = 1, + minimum_node_degree: int = 0, maximum_node_degree: Optional[int] = None, only_from_same_component: bool = True, sample_only_edges_with_heterogeneous_node_types: bool = False, @@ -202,7 +203,7 @@ def __init__( Using this parameter will raise an exception when the provided graph wither does not have node types or has exclusively constant node types. - minimum_node_degree: Optional[int] = 1 + minimum_node_degree: Optional[int] = 0 The minimum node degree of either the source or destination node to be sampled. maximum_node_degree: Optional[int] = None @@ -464,11 +465,14 @@ def _handle_notebook_display( """ # This is a visualization run for rotation. if len(args) < 2: - return None - figure, axes = args[:2] + figure = None + axes = None + else: + figure, axes = args[:2] if is_notebook() and self._automatically_display_on_notebooks: from IPython.display import display, HTML - display(figure) + if figure is not None: + display(figure) if caption is not None: display(HTML( '{}'.format( @@ -476,7 +480,7 @@ def _handle_notebook_display( ) )) plt.close() - elif caption is None: + elif caption is None or self._rotate: return (figure, axes, *args[2:]) else: return (figure, axes, *args[2:], caption) @@ -515,7 +519,7 @@ def get_heatmaps_comments(self, letters: Optional[List[str]] = None) -> str: " In the heatmap{plural}, {letters}" "low and high values appear in red and blue hues, respectively. " "Intermediate values appear in either a yellow or cyan hue. " - "The values are on a logarithmic scale." + "The values are on a logarithmic scale" ).format( plural=plural, letters="{}, ".format( @@ -547,8 +551,9 @@ def get_non_existing_edges_sampling_description(self) -> str: def get_decomposition_method(self) -> Callable: # Adding a warning for when decomposing methods that # embed nodes using a cosine similarity / distance approach - # in order to avoid false negatives. - if self._decomposition_method in ("UMAP", "TSNE") and self._node_embedding_method_name in ( + # in order to avoid false negatives, that is bad TSNE decompositions + # while the embedding is actually good. + if self._n_components < 3 and self._decomposition_method in ("UMAP", "TSNE") and self._node_embedding_method_name in ( "Node2Vec GloVe", "DeepWalk GloVe", "First-order LINE" ): metric = self._decomposition_kwargs.get("metric") @@ -736,6 +741,7 @@ def _get_node_embedding( f"to have an index curresponding to the node name `{node_name}`, " f"but we have found `{node_embedding.index[node_id]}`." ) + node_embedding = node_embedding.to_numpy() return node_embedding @@ -795,6 +801,16 @@ def decompose(self, X: np.ndarray) -> np.ndarray: "The vector to decompose has less components than " "the decomposition target." ) + # Some embedding method have complex values. + # Such values are, of course, not supported by UMAP, TSNE or PCA. + # For such cases, we need to convert the complex value into a real + # value. If the user desires to use some different approach, it can + # be applied to the embedding before providing it to this visualization tool. + if "complex" in str(X.dtype): + X = np.hstack([ + np.real(X), + np.imag(X) + ]) if self._decomposition_method == "TSNE" and X.shape[1] > 50 and self._graph.get_number_of_nodes() > 50: X = PCA( n_components=50, @@ -802,6 +818,28 @@ def decompose(self, X: np.ndarray) -> np.ndarray: ).fit_transform(X) return self.get_decomposition_method()(X) + def _normalize_label( + self, + labels: List[str] + ) -> List[str]: + last_element = labels[-1] + if last_element.lower().startswith("other"): + labels = [ + label + for label in sanitize_ml_labels([ + normalize_node_name(label) + for label in labels[:-1] + ]) + ] + labels.append(last_element) + return [ + label + for label in sanitize_ml_labels([ + normalize_node_name(label) + for label in labels + ]) + ] + def _set_legend( self, axes: Axes, @@ -827,13 +865,15 @@ def _set_legend( len(label) > 20 for label in labels ) else self._number_of_columns_in_legend + + labels = [ + "{}...".format(label[:20]) + if len(label) > 20 and number_of_columns == 2 else label + for label in self._normalize_label(labels) + ] legend = axes.legend( handles=handles, - labels=[ - "{}...".format(label[:20]) - if len(label) > 20 and number_of_columns == 2 else label - for label in sanitize_ml_labels(labels) - ], + labels=labels, loc=loc, ncol=number_of_columns, prop={'size': 8}, @@ -1054,8 +1094,6 @@ def _get_figure_and_axes( }, squeeze=False ) - for ax in axes.flatten(): - ax.axis('equal') if axes.size == 1: axes = axes.flatten()[0] else: @@ -1066,8 +1104,6 @@ def _get_figure_and_axes( }, squeeze=False ) - for ax in axes.flatten(): - ax.axis('auto') if axes.size == 1: axes = axes.flatten()[0] figure.patch.set_facecolor("white") @@ -1197,6 +1233,11 @@ def _plot_scatter( **kwargs ) + if self._n_components == 2: + axes.axis('equal') + else: + axes.axis('auto') + scatter_kwargs = { **GraphVisualizer.DEFAULT_SCATTER_KWARGS, **( @@ -1356,7 +1397,7 @@ def _plot_scatter( color_name=color_name, quotations="\'" if "other" not in label.lower() else "", ) - for color_name, label in zip(color_names_to_be_used, labels) + for color_name, label in zip(color_names_to_be_used, self._normalize_label(labels)) ]) return_values = (*return_values, caption) @@ -1530,10 +1571,6 @@ def _plot_types( if k < number_of_types: type_labels.append(other_label.format(number_of_types - k)) - type_labels = sanitize_ml_labels( - type_labels[:k] - ) + sanitize_ml_labels(type_labels[k:]) - result = self._wrapped_plot_scatter(**{ **dict( return_caption=return_caption, @@ -1566,11 +1603,16 @@ def _plot_types( test_accuracies = [] - for train_indices, test_indices in ShuffleSplit( + if min(Counter(types).values()) == 1: + SplitterClass = ShuffleSplit + else: + SplitterClass = StratifiedShuffleSplit + + for train_indices, test_indices in SplitterClass( n_splits=self._number_of_holdouts_for_cluster_comments, test_size=0.3, random_state=self._random_state - ).split(points): + ).split(points, types): model = DecisionTreeClassifier(max_depth=5) @@ -1988,7 +2030,7 @@ def plot_positive_and_negative_edges( returned_values = self._plot_types( points=points, title=self._get_complete_title( - "Existent & non-existent edges", + "Edge prediction", show_edge_embedding=True ), types=types, @@ -2170,11 +2212,16 @@ def _plot_positive_and_negative_edges_metric( test_accuracies = [] - for train_indices, test_indices in ShuffleSplit( + if min(Counter(types).values()) == 1: + SplitterClass = ShuffleSplit + else: + SplitterClass = StratifiedShuffleSplit + + for train_indices, test_indices in SplitterClass( n_splits=self._number_of_holdouts_for_cluster_comments, test_size=0.3, random_state=self._random_state - ).split(edge_metrics): + ).split(edge_metrics, types): model = DecisionTreeClassifier(max_depth=5) @@ -2262,7 +2309,7 @@ def _plot_positive_and_negative_edges_metric_histogram( edge_metrics = np.concatenate(( edge_metric_callback(subgraph=self._negative_graph), edge_metric_callback(subgraph=self._positive_graph), - )) + sys.float_info.epsilon + )) axes.hist( [ @@ -2273,6 +2320,7 @@ def _plot_positive_and_negative_edges_metric_histogram( log=True, label=["Non-existent", "Existent"] ) + axes.set_xlim(edge_metrics.min(), edge_metrics.max()) axes.set_ylabel("Counts (log scale)") axes.set_xlabel(metric_name) axes.legend(loc='best', prop={'size': 8},) @@ -3819,7 +3867,7 @@ def _plot_positive_and_negative_edges_distance_histogram( aligned_mapping=True, include_both_undirected_edges=False ) - graph_transformer.fit(node_features.astype(np.float32)) + graph_transformer.fit(node_features) return self._plot_positive_and_negative_edges_metric_histogram( metric_name=distance_name, @@ -3868,7 +3916,7 @@ def _plot_positive_and_negative_edges_distance( include_both_undirected_edges=False ) - graph_transformer.fit(node_features.astype(np.float32)) + graph_transformer.fit(node_features) return self._plot_positive_and_negative_edges_metric( metric_name=distance_name, @@ -4262,119 +4310,29 @@ def plot_edge_weight_distribution( return self._handle_notebook_display(figure, axes, caption=caption) - def fit_and_plot_all( + def _fit_and_plot_all( self, - node_embedding: Union[pd.DataFrame, np.ndarray, str, EmbeddingResult], - number_of_columns: int = 4, - show_letters: bool = True, - include_distribution_plots: bool = True, - **node_embedding_kwargs: Dict + points: List[np.ndarray], + nrows: int, + ncols: int, + plotting_callbacks: List[Callable], + show_letters: bool, ) -> Tuple[Figure, Axes]: """Fits and plots all available features of the graph. Parameters ------------------------- - node_embedding: Union[pd.DataFrame, np.ndarray, str] - Embedding of the graph nodes. - If a string is provided, we will run the node embedding - from one of the available methods. - number_of_columns: int = 4 - Number of columns to use for the layout. - show_letters: bool = True - Whether to show letters on the top left of the subplots. - include_distribution_plots: bool = True - Whether to include the distribution plots for the degrees - and the edge weights, if they are present. - **node_embedding_kwargs: Dict Kwargs to be forwarded to the node embedding algorithm. """ - node_embedding = self._get_node_embedding( - node_embedding, - **node_embedding_kwargs - ) - self.fit_nodes(node_embedding, **node_embedding_kwargs) - self.fit_negative_and_positive_edges( - node_embedding, **node_embedding_kwargs) - - node_scatter_plot_methods_to_call = [] - distribution_plot_methods_to_call = [] - - node_scatter_plot_methods_to_call.append( - self.plot_node_degrees, - ) - distribution_plot_methods_to_call.append( - self.plot_node_degree_distribution, - ) - - def plot_distance_wrapper(plot_distance): - @functools.wraps(plot_distance) - def wrapped_plot_distance(**kwargs): - return plot_distance( - node_features=node_embedding, - **kwargs - ) - return wrapped_plot_distance - - edge_scatter_plot_methods_to_call = [ - self.plot_positive_and_negative_edges, - plot_distance_wrapper( - self.plot_positive_and_negative_edges_euclidean_distance), - plot_distance_wrapper( - self.plot_positive_and_negative_edges_cosine_similarity), - self.plot_positive_and_negative_edges_adamic_adar, - self.plot_positive_and_negative_edges_jaccard_coefficient, - self.plot_positive_and_negative_edges_preferential_attachment, - self.plot_positive_and_negative_edges_resource_allocation_index - ] - - distribution_plot_methods_to_call = [ - plot_distance_wrapper( - self.plot_positive_and_negative_edges_euclidean_distance_histogram), - plot_distance_wrapper( - self.plot_positive_and_negative_edges_cosine_similarity_histogram), - self.plot_positive_and_negative_adamic_adar_histogram, - self.plot_positive_and_negative_jaccard_coefficient_histogram, - self.plot_positive_and_negative_preferential_attachment_histogram, - self.plot_positive_and_negative_resource_allocation_index_histogram + decompositions_backup = [ + self._node_decomposition, + self._positive_edge_decomposition, + self._negative_edge_decomposition ] - if self._graph.has_node_types() and not self._graph.has_homogeneous_node_types(): - node_scatter_plot_methods_to_call.append( - self.plot_node_types - ) - - if self._graph.has_node_ontologies() and not self._graph.has_homogeneous_node_ontologies(): - node_scatter_plot_methods_to_call.append( - self.plot_node_ontologies - ) - - if not self._support.is_connected() and not self._support.is_directed(): - node_scatter_plot_methods_to_call.append( - self.plot_connected_components - ) - - if self._positive_graph.has_edge_types() and not self._positive_graph.has_homogeneous_edge_types(): - edge_scatter_plot_methods_to_call.append( - self.plot_edge_types - ) - - if self._positive_graph.has_edge_weights() and not self._positive_graph.has_constant_edge_weights(): - edge_scatter_plot_methods_to_call.append( - self.plot_edge_weights - ) - distribution_plot_methods_to_call.append( - self.plot_edge_weight_distribution - ) - - if not include_distribution_plots: - distribution_plot_methods_to_call = [] - - number_of_total_plots = len(node_scatter_plot_methods_to_call) + len( - edge_scatter_plot_methods_to_call - ) + len(distribution_plot_methods_to_call) - nrows = max( - int(math.ceil(number_of_total_plots / number_of_columns)), 1) - ncols = min(number_of_columns, number_of_total_plots) + (self._node_decomposition, + self._positive_edge_decomposition, + self._negative_edge_decomposition) = points figure, axes = self._get_figure_and_axes( nrows=nrows, @@ -4383,6 +4341,7 @@ def wrapped_plot_distance(**kwargs): dpi=96 ) figure.patch.set_facecolor("white") + number_of_total_plots = len(plotting_callbacks) flat_axes = np.array(axes).flatten() @@ -4413,20 +4372,17 @@ def wrapped_plot_distance(**kwargs): for ax, plot_callback, letter in zip( flat_axes, - itertools.chain( - node_scatter_plot_methods_to_call, - edge_scatter_plot_methods_to_call, - distribution_plot_methods_to_call - ), + itertools.chain(plotting_callbacks), "abcdefghjkilmnopqrstuvwxyz" ): inspect.signature(plot_callback).parameters - _, _, caption = plot_callback( + figure, axes, caption = plot_callback( figure=figure, axes=ax, **(dict(loc="lower center") if "loc" in inspect.signature(plot_callback).parameters else dict()), apply_tight_layout=False ) + if "heatmap" in caption.lower(): heatmaps_letters.append(letter) if "accuracy" in caption.lower(): @@ -4434,14 +4390,12 @@ def wrapped_plot_distance(**kwargs): complete_caption += f" ({letter}) {caption}" if show_letters: - if self._n_components == 3: - additional_kwargs = dict(z=0.0) + if self._n_components >= 3: + additional_kwargs = dict(z=0, y=0, x=0) else: - additional_kwargs = dict() + additional_kwargs = dict(y=1.1, x=0) ax.text( - x=0.0, - y=1.1, s=letter, size=18, color="black", @@ -4488,9 +4442,191 @@ def wrapped_plot_distance(**kwargs): figure.tight_layout() self._show_graph_name = show_name_backup + (self._node_decomposition, + self._positive_edge_decomposition, + self._negative_edge_decomposition) = decompositions_backup return self._handle_notebook_display( figure, - axes, + flat_axes, caption=complete_caption ) + + def fit_and_plot_all( + self, + node_embedding: Union[pd.DataFrame, np.ndarray, str, EmbeddingResult], + number_of_columns: int = 4, + show_letters: bool = True, + include_distribution_plots: bool = True, + **node_embedding_kwargs: Dict + ) -> Tuple[Figure, Axes]: + """Fits and plots all available features of the graph. + + Parameters + ------------------------- + node_embedding: Union[pd.DataFrame, np.ndarray, str] + Embedding of the graph nodes. + If a string is provided, we will run the node embedding + from one of the available methods. + number_of_columns: int = 4 + Number of columns to use for the layout. + show_letters: bool = True + Whether to show letters on the top left of the subplots. + include_distribution_plots: bool = True + Whether to include the distribution plots for the degrees + and the edge weights, if they are present. + **node_embedding_kwargs: Dict + Kwargs to be forwarded to the node embedding algorithm. + """ + node_embedding = self._get_node_embedding( + node_embedding, + **node_embedding_kwargs + ) + self.fit_nodes(node_embedding, **node_embedding_kwargs) + self.fit_negative_and_positive_edges( + node_embedding, **node_embedding_kwargs) + + node_scatter_plot_methods_to_call = [] + distribution_plot_methods_to_call = [] + + node_scatter_plot_methods_to_call.append( + self.plot_node_degrees, + ) + distribution_plot_methods_to_call.append( + self.plot_node_degree_distribution, + ) + + def plot_distance_wrapper(plot_distance): + @functools.wraps(plot_distance) + def wrapped_plot_distance(**kwargs): + return plot_distance( + node_features=node_embedding, + **kwargs + ) + return wrapped_plot_distance + + edge_scatter_plot_methods_to_call = [ + self.plot_positive_and_negative_edges, + plot_distance_wrapper( + self.plot_positive_and_negative_edges_euclidean_distance), + plot_distance_wrapper( + self.plot_positive_and_negative_edges_cosine_similarity), + self.plot_positive_and_negative_edges_adamic_adar, + self.plot_positive_and_negative_edges_jaccard_coefficient, + self.plot_positive_and_negative_edges_preferential_attachment, + self.plot_positive_and_negative_edges_resource_allocation_index + ] + + distribution_plot_methods_to_call = [ + plot_distance_wrapper( + self.plot_positive_and_negative_edges_euclidean_distance_histogram), + plot_distance_wrapper( + self.plot_positive_and_negative_edges_cosine_similarity_histogram), + self.plot_positive_and_negative_adamic_adar_histogram, + self.plot_positive_and_negative_jaccard_coefficient_histogram, + self.plot_positive_and_negative_preferential_attachment_histogram, + self.plot_positive_and_negative_resource_allocation_index_histogram + ] + + if self._graph.has_node_types() and not self._graph.has_homogeneous_node_types(): + node_scatter_plot_methods_to_call.append( + self.plot_node_types + ) + + if self._graph.has_node_ontologies() and not self._graph.has_homogeneous_node_ontologies(): + node_scatter_plot_methods_to_call.append( + self.plot_node_ontologies + ) + + if not self._support.is_connected() and not self._support.is_directed(): + node_scatter_plot_methods_to_call.append( + self.plot_connected_components + ) + + if self._positive_graph.has_edge_types() and not self._positive_graph.has_homogeneous_edge_types(): + edge_scatter_plot_methods_to_call.append( + self.plot_edge_types + ) + + if self._positive_graph.has_edge_weights() and not self._positive_graph.has_constant_edge_weights(): + edge_scatter_plot_methods_to_call.append( + self.plot_edge_weights + ) + distribution_plot_methods_to_call.append( + self.plot_edge_weight_distribution + ) + + if not include_distribution_plots or self._rotate or self._n_components > 2: + distribution_plot_methods_to_call = [] + + plotting_callbacks = [ + callback + for callbacks in ( + node_scatter_plot_methods_to_call, + edge_scatter_plot_methods_to_call, + distribution_plot_methods_to_call + ) + for callback in callbacks + ] + + number_of_total_plots = len(plotting_callbacks) + + nrows = max( + int(math.ceil(number_of_total_plots / number_of_columns)), 1) + ncols = min(number_of_columns, number_of_total_plots) + + if self._rotate: + path = "fit_and_plot_all.webm" + display_backup = self._automatically_display_on_notebooks + self._automatically_display_on_notebooks = False + rotate_backup = self._rotate + self._rotate = False + rotate( + self._fit_and_plot_all, + path=path, + points=[ + self._node_decomposition, + self._positive_edge_decomposition, + self._negative_edge_decomposition + ], + duration=self._duration, + fps=self._fps, + verbose=self._verbose, + nrows = nrows, + ncols = ncols, + plotting_callbacks=plotting_callbacks, + show_letters=show_letters, + ) + to_display = display_video_at_path( + path, + width="100%", + height=None + ) + figure, axes, complete_caption = self._fit_and_plot_all( + points=[ + self._node_decomposition, + self._positive_edge_decomposition, + self._negative_edge_decomposition + ], + nrows = nrows, + ncols = ncols, + plotting_callbacks=plotting_callbacks, + show_letters=show_letters + ) + self._rotate = rotate_backup + self._automatically_display_on_notebooks = display_backup + return self._handle_notebook_display( + to_display, None, + caption=complete_caption + ) + return self._fit_and_plot_all( + points=[ + self._node_decomposition, + self._positive_edge_decomposition, + self._negative_edge_decomposition + ], + nrows = nrows, + ncols = ncols, + plotting_callbacks=plotting_callbacks, + show_letters=show_letters + ) diff --git a/setup.py b/setup.py index 599276ae..3c85d921 100644 --- a/setup.py +++ b/setup.py @@ -8,7 +8,7 @@ def readme(): - with open('README.rst') as f: + with open('README.rst', encoding="utf8") as f: return f.read() @@ -24,7 +24,7 @@ def readme(): def read(*parts): - with copen(os.path.join(here, *parts), 'r') as fp: + with copen(os.path.join(here, *parts), 'r', encoding="utf8") as fp: return fp.read() @@ -66,13 +66,17 @@ def find_version(*file_paths): "tqdm", "matplotlib>=3.5.2", "scikit-learn", + "dict_hash>=1.1.29", "userinput>=1.0.19", - "ddd_subplots>=1.0.20", - "sanitize_ml_labels>=1.0.38", + "ddd_subplots>=1.0.23", + "sanitize_ml_labels>=1.0.45", "keras_mixed_sequence>=1.0.28", - "ensmallen>=0.8.8", + "ensmallen>=0.8.24", + "environments_utils>=1.0.6", + "compress_pickle>=2.1.0", "validate_version_code", - "cache_decorator>=2.1.8", + "ringbell>=1.0.2", + "cache_decorator>=2.1.11", "packaging" ], tests_require=test_deps,

Fix typo in `environments_utils` dependency

2022-10-04T18:57:08

0.0

[]

monarch-initiative/embiggen

monarch-initiative__embiggen-283

42500d5d18503b5ef3a109daf915d49c2064ba03

diff --git a/.gitignore b/.gitignore index cbec2be7..6634e04f 100644 --- a/.gitignore +++ b/.gitignore @@ -33,4 +33,5 @@ htmlcov embedding.npy build graphs -embeddings \ No newline at end of file +embeddings +*.webm \ No newline at end of file diff --git a/README.rst b/README.rst index 5a6d6464..87387cda 100644 --- a/README.rst +++ b/README.rst @@ -1,8 +1,8 @@ -Embiggen +🍇 Embiggen ========================================================================================= |pip| |downloads| |tutorials| |documentation| |python_version| |DOI| |license| |telegram| |discord| |twitter| -Embiggen is the graph machine learning submodule of the `GraPE <https://github.com/AnacletoLAB/grape>`_ library. +Embiggen is the graph machine learning submodule of the `🍇 GraPE <https://github.com/AnacletoLAB/grape>`_ library. How to install Embiggen ------------------------- diff --git a/dist/embiggen-0.11.12.tar.gz b/dist/embiggen-0.11.12.tar.gz new file mode 100644 index 00000000..ccc2cc1b Binary files /dev/null and b/dist/embiggen-0.11.12.tar.gz differ diff --git a/embiggen/__version__.py b/embiggen/__version__.py index c084d6b1..e1fd9917 100644 --- a/embiggen/__version__.py +++ b/embiggen/__version__.py @@ -1,2 +1,2 @@ """Current version of package embiggen""" -__version__ = "0.11.11" +__version__ = "0.11.15" diff --git a/embiggen/edge_label_prediction/edge_label_prediction_evaluation.py b/embiggen/edge_label_prediction/edge_label_prediction_evaluation.py index 254662b0..dec73085 100644 --- a/embiggen/edge_label_prediction/edge_label_prediction_evaluation.py +++ b/embiggen/edge_label_prediction/edge_label_prediction_evaluation.py @@ -13,6 +13,7 @@ def edge_label_prediction_evaluation( models: Union[Type[AbstractEdgeLabelPredictionModel], List[Type[AbstractEdgeLabelPredictionModel]]], evaluation_schema: str = "Stratified Monte Carlo", node_features: Optional[Union[str, pd.DataFrame, np.ndarray, Type[AbstractEmbeddingModel], List[Union[str, pd.DataFrame, np.ndarray, Type[AbstractEmbeddingModel]]]]] = None, + node_type_features: Optional[Union[str, pd.DataFrame, np.ndarray, Type[AbstractEmbeddingModel], List[Union[str, pd.DataFrame, np.ndarray, Type[AbstractEmbeddingModel]]]]] = None, edge_features: Optional[Union[str, pd.DataFrame, np.ndarray, List[Union[str, pd.DataFrame, np.ndarray]]]] = None, library_names: Optional[Union[str, List[str]]] = None, graph_callback: Optional[Callable[[Graph], Graph]] = None, @@ -40,6 +41,8 @@ def edge_label_prediction_evaluation( The evaluation schema to follow. node_features: Optional[Union[str, pd.DataFrame, np.ndarray, Type[AbstractEmbeddingModel], List[Union[str, pd.DataFrame, np.ndarray, Type[AbstractEmbeddingModel]]]]] = None The node features to use. + node_type_features: Optional[Union[str, pd.DataFrame, np.ndarray, Type[AbstractEmbeddingModel], List[Union[str, pd.DataFrame, np.ndarray, Type[AbstractEmbeddingModel]]]]] = None + The node type features to use. edge_features: Optional[Union[str, pd.DataFrame, np.ndarray, List[Union[str, pd.DataFrame, np.ndarray]]]] = None The edge features to use. library_names: Optional[Union[str, List[str]]] = None @@ -86,6 +89,7 @@ def edge_label_prediction_evaluation( models=models, expected_parent_class=AbstractEdgeLabelPredictionModel, node_features=node_features, + node_type_features=node_type_features, edge_features=edge_features, library_names=library_names, graph_callback=graph_callback, diff --git a/embiggen/edge_label_prediction/edge_label_prediction_model.py b/embiggen/edge_label_prediction/edge_label_prediction_model.py index aa21c278..00d82516 100644 --- a/embiggen/edge_label_prediction/edge_label_prediction_model.py +++ b/embiggen/edge_label_prediction/edge_label_prediction_model.py @@ -3,7 +3,7 @@ import pandas as pd import numpy as np from ensmallen import Graph -from embiggen.utils.abstract_models import AbstractClassifierModel, format_list +from embiggen.utils.abstract_models import AbstractClassifierModel class AbstractEdgeLabelPredictionModel(AbstractClassifierModel): @@ -20,7 +20,7 @@ def __init__(self, random_state: Optional[int] = None): self._is_binary_prediction_task = None self._is_multilabel_prediction_task = None super().__init__(random_state=random_state) - + @classmethod def requires_edge_types(cls) -> bool: """Returns whether this method requires node types.""" @@ -48,7 +48,9 @@ def get_available_evaluation_schemas(cls) -> List[str]: """Returns available evaluation schemas for this task.""" return [ "Stratified Monte Carlo", - "Stratified Kfold" + "Stratified Kfold", + "Kfold", + "Monte Carlo", ] @classmethod @@ -78,23 +80,26 @@ def split_graph_following_evaluation_schema( holdouts_kwargs: Dict[str, Any] The kwargs to be forwarded to the holdout method. """ - if evaluation_schema == "Stratified Monte Carlo": + if evaluation_schema in ("Monte Carlo", "Stratified Monte Carlo"): return graph.get_edge_label_holdout_graphs( **holdouts_kwargs, - use_stratification=True, + use_stratification="Stratified" in evaluation_schema, random_state=random_state+holdout_number, ) - if evaluation_schema == "Stratified Kfold": + if evaluation_schema in ("Kfold", "Stratified Kfold"): return graph.get_edge_label_kfold( k=number_of_holdouts, k_index=holdout_number, - use_stratification=True, + use_stratification="Stratified" in evaluation_schema, random_state=random_state, ) - raise ValueError( - f"The requested evaluation schema `{evaluation_schema}` " - "is not available. The available evaluation schemas " - f"are: {format_list(cls.get_available_evaluation_schemas())}." + super().split_graph_following_evaluation_schema( + graph=graph, + evaluation_schema=evaluation_schema, + random_state=random_state, + holdout_number=holdout_number, + number_of_holdouts=number_of_holdouts, + **holdouts_kwargs, ) @classmethod @@ -127,6 +132,7 @@ def _evaluate( ) -> List[Dict[str, Any]]: """Return model evaluation on the provided graphs.""" train_size = train.get_number_of_known_edge_types() / graph.get_number_of_known_edge_types() + performance = [] for evaluation_mode, evaluation_graph in ( ("train", train), @@ -139,9 +145,8 @@ def _evaluate( node_type_features=node_type_features, edge_features=edge_features )[evaluation_graph.get_edge_ids_with_known_edge_types()] - + if self.is_binary_prediction_task(): - prediction_probabilities = prediction_probabilities[:, 1] predictions = prediction_probabilities labels = evaluation_graph.get_known_edge_type_ids() == 1 elif self.is_multilabel_prediction_task(): @@ -158,7 +163,7 @@ def _evaluate( performance.append({ "evaluation_mode": evaluation_mode, "train_size": train_size, - "known_edges_number": graph.get_number_of_known_node_types(), + "known_edges_number": graph.get_number_of_known_edge_types(), **self.evaluate_predictions( labels, predictions, @@ -197,15 +202,26 @@ def fit( edge_features: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None The edge features to use. """ - if node_type_features is not None: - raise NotImplementedError( - "Support for node type features is not currently available for any " - "of the edge-label prediction models." + non_zero_edge_types = sum([ + 1 + for count in graph.get_edge_type_names_counts_hashmap().values() + if count > 0 + ]) + + if non_zero_edge_types < 2: + raise ValueError( + "The provided training graph has less than two non-zero edge types. " + "It is unclear how to proceeed." ) - self._is_binary_prediction_task = graph.get_number_of_edge_types() == 2 + self._is_binary_prediction_task = non_zero_edge_types == 2 self._is_multilabel_prediction_task = graph.is_multigraph() + if self._is_multilabel_prediction_task: + raise ValueError( + "Currently we do not support multi-label edge prediction." + ) + super().fit( graph=graph, support=support, @@ -232,4 +248,4 @@ def task_involves_node_types(cls) -> bool: @classmethod def task_involves_topology(cls) -> bool: """Returns whether the model task involves topology.""" - return False \ No newline at end of file + return False diff --git a/embiggen/edge_label_prediction/edge_label_prediction_sklearn/decision_tree_edge_label_prediction.py b/embiggen/edge_label_prediction/edge_label_prediction_sklearn/decision_tree_edge_label_prediction.py index f03bfeb5..107f6365 100644 --- a/embiggen/edge_label_prediction/edge_label_prediction_sklearn/decision_tree_edge_label_prediction.py +++ b/embiggen/edge_label_prediction/edge_label_prediction_sklearn/decision_tree_edge_label_prediction.py @@ -18,10 +18,10 @@ def __init__( max_features=None, max_leaf_nodes=None, min_impurity_decrease=0., - min_impurity_split=None, - class_weight=None, + class_weight="balanced", ccp_alpha=0.0, edge_embedding_method: str = "Concatenate", + use_edge_metrics: bool = False, random_state: int = 42 ): """Create the Decision Tree for Edge Label Prediction.""" @@ -34,7 +34,6 @@ def __init__( self._max_features = max_features self._max_leaf_nodes = max_leaf_nodes self._min_impurity_decrease = min_impurity_decrease - self._min_impurity_split = min_impurity_split self._random_state = random_state self._class_weight = class_weight self._ccp_alpha = ccp_alpha @@ -50,13 +49,13 @@ def __init__( max_features=max_features, max_leaf_nodes=max_leaf_nodes, min_impurity_decrease=min_impurity_decrease, - min_impurity_split=min_impurity_split, random_state=random_state, class_weight=class_weight, ccp_alpha=ccp_alpha, ), - edge_embedding_method, - random_state + edge_embedding_method=edge_embedding_method, + use_edge_metrics=use_edge_metrics, + random_state=random_state ) @classmethod @@ -80,7 +79,6 @@ def parameters(self) -> Dict[str, Any]: max_features=self._max_features, max_leaf_nodes=self._max_leaf_nodes, min_impurity_decrease=self._min_impurity_decrease, - min_impurity_split=self._min_impurity_split, random_state=self._random_state, class_weight=self._class_weight, ccp_alpha=self._ccp_alpha, diff --git a/embiggen/edge_label_prediction/edge_label_prediction_sklearn/extra_trees_edge_label_prediction.py b/embiggen/edge_label_prediction/edge_label_prediction_sklearn/extra_trees_edge_label_prediction.py index d9f0d961..df0c2b93 100644 --- a/embiggen/edge_label_prediction/edge_label_prediction_sklearn/extra_trees_edge_label_prediction.py +++ b/embiggen/edge_label_prediction/edge_label_prediction_sklearn/extra_trees_edge_label_prediction.py @@ -20,16 +20,16 @@ def __init__( max_features="auto", max_leaf_nodes=None, min_impurity_decrease=0., - min_impurity_split=None, bootstrap=True, oob_score=False, n_jobs=-1, verbose=0, warm_start=False, - class_weight=None, + class_weight="balanced", ccp_alpha=0.0, max_samples=None, edge_embedding_method: str = "Concatenate", + use_edge_metrics: bool = False, random_state: int = 42 ): """Create the Extra Trees for Edge Label Prediction.""" @@ -42,7 +42,6 @@ def __init__( self._max_features = max_features self._max_leaf_nodes = max_leaf_nodes self._min_impurity_decrease = min_impurity_decrease - self._min_impurity_split = min_impurity_split self._bootstrap = bootstrap self._oob_score = oob_score self._n_jobs = cpu_count() if n_jobs == -1 else n_jobs @@ -64,7 +63,6 @@ def __init__( max_features=max_features, max_leaf_nodes=max_leaf_nodes, min_impurity_decrease=min_impurity_decrease, - min_impurity_split=min_impurity_split, bootstrap=bootstrap, oob_score=oob_score, n_jobs=n_jobs, @@ -75,8 +73,9 @@ def __init__( ccp_alpha=ccp_alpha, max_samples=max_samples ), - edge_embedding_method, - random_state + edge_embedding_method=edge_embedding_method, + use_edge_metrics=use_edge_metrics, + random_state=random_state ) @classmethod @@ -101,7 +100,6 @@ def parameters(self) -> Dict[str, Any]: max_features = self._max_features, max_leaf_nodes = self._max_leaf_nodes, min_impurity_decrease = self._min_impurity_decrease, - min_impurity_split = self._min_impurity_split, bootstrap = self._bootstrap, oob_score = self._oob_score, n_jobs = self._n_jobs, diff --git a/embiggen/edge_label_prediction/edge_label_prediction_sklearn/gradient_boosting_edge_label_prediction.py b/embiggen/edge_label_prediction/edge_label_prediction_sklearn/gradient_boosting_edge_label_prediction.py index cddb60e5..bdb243d0 100644 --- a/embiggen/edge_label_prediction/edge_label_prediction_sklearn/gradient_boosting_edge_label_prediction.py +++ b/embiggen/edge_label_prediction/edge_label_prediction_sklearn/gradient_boosting_edge_label_prediction.py @@ -19,7 +19,6 @@ def __init__( min_weight_fraction_leaf=0., max_depth=3, min_impurity_decrease=0., - min_impurity_split=None, init=None, max_features=None, verbose=0, @@ -30,25 +29,25 @@ def __init__( tol=1e-4, ccp_alpha=0.0, edge_embedding_method: str = "Concatenate", + use_edge_metrics: bool = False, random_state: int = 42 ): """Create the Gradient Boosting for Edge Label Prediction.""" self._loss=loss self._learning_rate=learning_rate - self._n_estimators=n_estimators, + self._n_estimators=n_estimators self._criterion=criterion - self._min_samples_split=min_samples_split, - self._min_samples_leaf=min_samples_leaf, - self._min_weight_fraction_leaf=min_weight_fraction_leaf, + self._min_samples_split=min_samples_split + self._min_samples_leaf=min_samples_leaf + self._min_weight_fraction_leaf=min_weight_fraction_leaf self._max_depth=max_depth self._init=init - self._subsample=subsample, - self._max_features=max_features, - self._max_leaf_nodes=max_leaf_nodes, - self._min_impurity_decrease=min_impurity_decrease, - self._min_impurity_split=min_impurity_split, + self._subsample=subsample + self._max_features=max_features + self._max_leaf_nodes=max_leaf_nodes + self._min_impurity_decrease=min_impurity_decrease self._warm_start=warm_start - self._validation_fraction=validation_fraction, + self._validation_fraction=validation_fraction self._n_iter_no_change=n_iter_no_change self._tol=tol self._ccp_alpha=ccp_alpha @@ -63,12 +62,12 @@ def __init__( random_state=random_state, verbose=verbose, max_leaf_nodes=max_leaf_nodes, min_impurity_decrease=min_impurity_decrease, - min_impurity_split=min_impurity_split, warm_start=warm_start, validation_fraction=validation_fraction, n_iter_no_change=n_iter_no_change, tol=tol, ccp_alpha=ccp_alpha ), - edge_embedding_method, - random_state + edge_embedding_method=edge_embedding_method, + use_edge_metrics=use_edge_metrics, + random_state=random_state ) @classmethod @@ -97,7 +96,6 @@ def parameters(self) -> Dict[str, Any]: max_features=self._max_features, max_leaf_nodes=self._max_leaf_nodes, min_impurity_decrease=self._min_impurity_decrease, - min_impurity_split=self._min_impurity_split, warm_start=self._warm_start, validation_fraction=self._validation_fraction, n_iter_no_change=self._n_iter_no_change, diff --git a/embiggen/edge_label_prediction/edge_label_prediction_sklearn/mlp_edge_label_prediction.py b/embiggen/edge_label_prediction/edge_label_prediction_sklearn/mlp_edge_label_prediction.py index 1e825e1d..f76f39f0 100644 --- a/embiggen/edge_label_prediction/edge_label_prediction_sklearn/mlp_edge_label_prediction.py +++ b/embiggen/edge_label_prediction/edge_label_prediction_sklearn/mlp_edge_label_prediction.py @@ -32,6 +32,7 @@ def __init__( n_iter_no_change=10, max_fun=15000, edge_embedding_method: str = "Concatenate", + use_edge_metrics: bool = False, random_state: int = 42 ): """Create the MLP for Edge Label Prediction.""" @@ -82,8 +83,9 @@ def __init__( n_iter_no_change=n_iter_no_change, max_fun=max_fun ), - edge_embedding_method, - random_state + edge_embedding_method=edge_embedding_method, + use_edge_metrics=use_edge_metrics, + random_state=random_state ) @classmethod diff --git a/embiggen/edge_label_prediction/edge_label_prediction_sklearn/random_forest_edge_label_prediction.py b/embiggen/edge_label_prediction/edge_label_prediction_sklearn/random_forest_edge_label_prediction.py index 1fe4170c..551a7554 100644 --- a/embiggen/edge_label_prediction/edge_label_prediction_sklearn/random_forest_edge_label_prediction.py +++ b/embiggen/edge_label_prediction/edge_label_prediction_sklearn/random_forest_edge_label_prediction.py @@ -17,19 +17,19 @@ def __init__( min_samples_split: int = 2, min_samples_leaf: int = 1, min_weight_fraction_leaf: float = 0., - max_features="auto", + max_features="sqrt", max_leaf_nodes=None, min_impurity_decrease=0., - min_impurity_split=None, bootstrap=True, oob_score=False, n_jobs=-1, verbose=0, warm_start=False, - class_weight=None, + class_weight="balanced", ccp_alpha=0.0, max_samples=None, edge_embedding_method: str = "Concatenate", + use_edge_metrics: bool = False, random_state: int = 42 ): """Create the Random Forest for Edge Label Prediction.""" @@ -42,7 +42,6 @@ def __init__( self._max_features = max_features self._max_leaf_nodes = max_leaf_nodes self._min_impurity_decrease = min_impurity_decrease - self._min_impurity_split = min_impurity_split self._bootstrap = bootstrap self._oob_score = oob_score self._n_jobs = cpu_count() if n_jobs == -1 else n_jobs @@ -64,7 +63,6 @@ def __init__( max_features=max_features, max_leaf_nodes=max_leaf_nodes, min_impurity_decrease=min_impurity_decrease, - min_impurity_split=min_impurity_split, bootstrap=bootstrap, oob_score=oob_score, n_jobs=n_jobs, @@ -75,8 +73,9 @@ def __init__( ccp_alpha=ccp_alpha, max_samples=max_samples ), - edge_embedding_method, - random_state + edge_embedding_method=edge_embedding_method, + use_edge_metrics=use_edge_metrics, + random_state=random_state ) @classmethod @@ -101,7 +100,6 @@ def parameters(self) -> Dict[str, Any]: max_features = self._max_features, max_leaf_nodes = self._max_leaf_nodes, min_impurity_decrease = self._min_impurity_decrease, - min_impurity_split = self._min_impurity_split, bootstrap = self._bootstrap, oob_score = self._oob_score, n_jobs = self._n_jobs, diff --git a/embiggen/edge_label_prediction/edge_label_prediction_sklearn/sklearn_edge_label_prediction_adapter.py b/embiggen/edge_label_prediction/edge_label_prediction_sklearn/sklearn_edge_label_prediction_adapter.py index aafda249..013d081c 100644 --- a/embiggen/edge_label_prediction/edge_label_prediction_sklearn/sklearn_edge_label_prediction_adapter.py +++ b/embiggen/edge_label_prediction/edge_label_prediction_sklearn/sklearn_edge_label_prediction_adapter.py @@ -16,6 +16,7 @@ def __init__( self, model_instance: Type[ClassifierMixin], edge_embedding_method: str = "Concatenate", + use_edge_metrics: bool = False, random_state: int = 42 ): """Create the adapter for Sklearn object. @@ -26,6 +27,13 @@ def __init__( The class instance to be adapted into edge-label prediction. edge_embedding_method: str = "Concatenate" The method to use to compute the edges. + use_edge_metrics: bool = False + Whether to use the edge metrics from traditional edge prediction. + These metrics currently include: + - Adamic Adar + - Jaccard Coefficient + - Resource allocation index + - Preferential attachment random_state: int = 42 The random state to use to reproduce the training. @@ -38,6 +46,7 @@ def __init__( must_be_an_sklearn_classifier_model(model_instance) self._model_instance = model_instance self._edge_embedding_method = edge_embedding_method + self._use_edge_metrics = use_edge_metrics # We want to mask the decorator class name self.__class__.__name__ = model_instance.__class__.__name__ self.__class__.__doc__ = model_instance.__class__.__doc__ @@ -46,6 +55,7 @@ def parameters(self) -> Dict[str, Any]: """Returns parameters used for this model.""" return { "edge_embedding_method": self._edge_embedding_method, + "use_edge_metrics": self._use_edge_metrics, **super().parameters() } @@ -56,6 +66,7 @@ def clone(self) -> Type["SklearnEdgeLabelPredictionAdapter"]: def _trasform_graph_into_edge_embedding( self, graph: Graph, + support: Optional[Graph] = None, node_features: Optional[List[np.ndarray]] = None, node_type_features: Optional[List[np.ndarray]] = None, edge_features: Optional[List[np.ndarray]] = None, @@ -67,6 +78,11 @@ def _trasform_graph_into_edge_embedding( graph: Graph The graph whose edges are to be embedded and predicted. It can either be an Graph or a list of lists of edges. + support: Optional[Graph] = None + The graph describiding the topological structure that + includes also the above graph. This parameter + is mostly useful for topological classifiers + such as Graph Convolutional Networks. node_features: np.ndarray The node features to be used in the training of the model. node_type_features: np.ndarray @@ -83,13 +99,30 @@ def _trasform_graph_into_edge_embedding( """ gt = GraphTransformer( method=self._edge_embedding_method, - aligned_node_mapping=True + aligned_mapping=True + ) + + gt.fit( + node_features, + node_type_feature=node_type_features ) - gt.fit(node_features) + if self._use_edge_metrics: + edge_metrics = support.get_all_edge_metrics( + normalize=True, + subgraph=graph, + ) + if edge_features is not None: + edge_features = np.hstack([ + *edge_features, + edge_metrics + ]) + else: + edge_features = edge_metrics return gt.transform( graph=graph, + node_types=graph, edge_features=edge_features ) @@ -129,16 +162,31 @@ def _fit( """ lpt = EdgeLabelPredictionTransformer( method=self._edge_embedding_method, - aligned_node_mapping=True + aligned_mapping=True ) - lpt.fit(node_features) + lpt.fit( + node_features, + node_type_feature=node_type_features + ) + + if self._use_edge_metrics: + edge_metrics = support.get_all_edge_metrics( + normalize=True, + subgraph=graph, + ) + if edge_features is not None: + edge_features = np.hstack([ + *edge_features, + edge_metrics + ]) + else: + edge_features = edge_metrics self._model_instance.fit(*lpt.transform( graph=graph, edge_features=edge_features, behaviour_for_unknown_edge_labels="drop", - random_state=self._random_state )) def _predict_proba( @@ -175,13 +223,24 @@ def _predict_proba( ValueError If the two graphs do not share the same node vocabulary. """ - return self._model_instance.predict_proba(self._trasform_graph_into_edge_embedding( - graph=graph, - node_features=node_features, - edge_features=edge_features, - )) + prediction_probabilities = self._model_instance.predict_proba( + self._trasform_graph_into_edge_embedding( + graph=graph, + support=support, + node_features=node_features, + node_type_features=node_type_features, + edge_features=edge_features, + ) + ) + # In the majority but not totality of sklearn models, + # the predictions of binary models are returned as + # a couple of vectors for the positive and negative class. + if self.is_binary_prediction_task() and prediction_probabilities.shape[1] == 2: + prediction_probabilities = prediction_probabilities[:, 1] + return prediction_probabilities + - def predict( + def _predict( self, graph: Graph, support: Optional[Graph] = None, @@ -217,7 +276,9 @@ def predict( """ return self._model_instance.predict(self._trasform_graph_into_edge_embedding( graph=graph, + support=support, node_features=node_features, + node_type_features=node_type_features, edge_features=edge_features, )) diff --git a/embiggen/edge_label_prediction/edge_label_prediction_tensorflow/gcn.py b/embiggen/edge_label_prediction/edge_label_prediction_tensorflow/gcn.py index 1747394d..80264125 100644 --- a/embiggen/edge_label_prediction/edge_label_prediction_tensorflow/gcn.py +++ b/embiggen/edge_label_prediction/edge_label_prediction_tensorflow/gcn.py @@ -1,9 +1,9 @@ """GCN model for edge-label prediction.""" -from typing import List, Union, Optional, Dict, Type, Any +from typing import List, Union, Optional, Dict, Type import numpy as np from ensmallen import Graph - +import tensorflow as tf from tensorflow.keras.optimizers import Optimizer from embiggen.utils.abstract_edge_gcn import AbstractEdgeGCN, abstract_class from embiggen.edge_label_prediction.edge_label_prediction_model import AbstractEdgeLabelPredictionModel @@ -25,6 +25,7 @@ def __init__( number_of_units_per_ffnn_head_layer: Union[int, List[int]] = 128, dropout_rate: float = 0.3, apply_norm: bool = False, + combiner: str = "mean", edge_embedding_method: str = "Concatenate", optimizer: Union[str, Type[Optimizer]] = "adam", early_stopping_min_delta: float = 0.0001, @@ -47,6 +48,7 @@ def __init__( handling_multi_graph: str = "warn", node_feature_names: Optional[List[str]] = None, node_type_feature_names: Optional[List[str]] = None, + edge_feature_names: Optional[List[str]] = None, verbose: bool = False ): """Create new Kipf GCN object. @@ -77,6 +79,13 @@ def __init__( apply_norm: bool = False Whether to normalize the output of the convolution operations, after applying the level activations. + combiner: str = "mean" + A string specifying the reduction op. + Currently "mean", "sqrtn" and "sum" are supported. + "sum" computes the weighted sum of the embedding results for each row. + "mean" is the weighted sum divided by the total weight. + "sqrtn" is the weighted sum divided by the square root of the sum of the squares of the weights. + Defaults to mean. edge_embedding_method: str = "Concatenate" The edge embedding method to use to put togheter the source and destination node features, which includes: @@ -173,6 +182,9 @@ def __init__( node_type_feature_names: Optional[List[str]] = None Names of the node type features. This is used as the layer names. + edge_feature_names: Optional[List[str]] = None + Names of the edge features. + This is used as the layer names. verbose: bool = False Whether to show loading bars. """ @@ -187,6 +199,7 @@ def __init__( number_of_units_per_ffnn_head_layer=number_of_units_per_ffnn_head_layer, dropout_rate=dropout_rate, apply_norm=apply_norm, + combiner=combiner, edge_embedding_method=edge_embedding_method, optimizer=optimizer, early_stopping_min_delta=early_stopping_min_delta, @@ -209,6 +222,7 @@ def __init__( handling_multi_graph=handling_multi_graph, node_feature_names=node_feature_names, node_type_feature_names=node_type_feature_names, + edge_feature_names=edge_feature_names, verbose=verbose, ) AbstractEdgeLabelPredictionModel.__init__(self, random_state=random_state) @@ -245,4 +259,6 @@ def _get_class_weights(self, graph: Graph) -> Dict[int, float]: def get_output_classes(self, graph: Graph) -> int: """Returns number of output classes.""" + if self.is_binary_prediction_task(): + return 1 return graph.get_number_of_edge_types() diff --git a/embiggen/edge_label_prediction/edge_label_prediction_tensorflow/gnn.py b/embiggen/edge_label_prediction/edge_label_prediction_tensorflow/gnn.py index e8adf840..e68d4734 100644 --- a/embiggen/edge_label_prediction/edge_label_prediction_tensorflow/gnn.py +++ b/embiggen/edge_label_prediction/edge_label_prediction_tensorflow/gnn.py @@ -14,7 +14,6 @@ def __init__( number_of_head_layers: int = 1, number_of_units_per_body_layer: Union[int, List[int]] = 128, number_of_units_per_head_layer: Union[int, List[int]] = 128, - dropout_rate: float = 0.3, edge_embedding_method: str = "Concatenate", optimizer: Union[str, Type[Optimizer]] = "adam", early_stopping_min_delta: float = 0.0001, @@ -35,6 +34,7 @@ def __init__( node_type_embedding_size: int = 50, node_feature_names: Optional[List[str]] = None, node_type_feature_names: Optional[List[str]] = None, + edge_feature_names: Optional[List[str]] = None, verbose: bool = False ): """Create new GraphSAGE object. @@ -55,9 +55,6 @@ def __init__( Number of units per ffnn body layer. number_of_units_per_ffnn_head_layer: Union[int, List[int]] = 128 Number of units per ffnn head layer. - dropout_rate: float = 0.3 - Float between 0 and 1. - Fraction of the input units to dropout. edge_embedding_method: str = "Concatenate" The edge embedding method to use to put togheter the source and destination node features, which includes: @@ -146,6 +143,9 @@ def __init__( node_type_feature_names: Optional[List[str]] = None Names of the node type features. This is used as the layer names. + edge_feature_names: Optional[List[str]] = None + Names of the edge features. + This is used as the layer names. verbose: bool = False Whether to show loading bars. """ @@ -157,7 +157,6 @@ def __init__( number_of_ffnn_head_layers=number_of_head_layers, number_of_units_per_ffnn_body_layer=number_of_units_per_body_layer, number_of_units_per_ffnn_head_layer=number_of_units_per_head_layer, - dropout_rate=dropout_rate, edge_embedding_method=edge_embedding_method, optimizer=optimizer, early_stopping_min_delta=early_stopping_min_delta, @@ -178,6 +177,7 @@ def __init__( node_type_embedding_size=node_type_embedding_size, node_feature_names=node_feature_names, node_type_feature_names=node_type_feature_names, + edge_feature_names=edge_feature_names, verbose=verbose, ) @@ -206,14 +206,17 @@ def parameters(self) -> Dict[str, Any]: "number_of_units_per_graph_convolution_layers", "handling_multi_graph", "number_of_units_per_ffnn_body_layer", - "number_of_units_per_ffnn_head_layer" + "number_of_units_per_ffnn_head_layer", + "combiner", + "apply_norm", + "dropout_rate" ] return dict( number_of_units_per_body_layer=self._number_of_units_per_ffnn_body_layer, number_of_units_per_head_layer=self._number_of_units_per_ffnn_head_layer, **{ key: value - for key, value in super().smoke_test_parameters().items() + for key, value in super().parameters().items() if key not in removed } ) diff --git a/embiggen/edge_label_prediction/edge_label_prediction_tensorflow/graph_sage.py b/embiggen/edge_label_prediction/edge_label_prediction_tensorflow/graph_sage.py index 7718701b..3e27699f 100644 --- a/embiggen/edge_label_prediction/edge_label_prediction_tensorflow/graph_sage.py +++ b/embiggen/edge_label_prediction/edge_label_prediction_tensorflow/graph_sage.py @@ -39,6 +39,7 @@ def __init__( handling_multi_graph: str = "warn", node_feature_names: Optional[List[str]] = None, node_type_feature_names: Optional[List[str]] = None, + edge_feature_names: Optional[List[str]] = None, verbose: bool = False ): """Create new GraphSAGE object. @@ -160,6 +161,9 @@ def __init__( node_type_feature_names: Optional[List[str]] = None Names of the node type features. This is used as the layer names. + edge_feature_names: Optional[List[str]] = None + Names of the edge features. + This is used as the layer names. verbose: bool = False Whether to show loading bars. """ @@ -173,6 +177,7 @@ def __init__( number_of_units_per_ffnn_head_layer=number_of_units_per_ffnn_head_layer, dropout_rate=dropout_rate, apply_norm=True, + combiner="mean", edge_embedding_method=edge_embedding_method, optimizer=optimizer, early_stopping_min_delta=early_stopping_min_delta, @@ -195,6 +200,7 @@ def __init__( handling_multi_graph=handling_multi_graph, node_feature_names=node_feature_names, node_type_feature_names=node_type_feature_names, + edge_feature_names=edge_feature_names, verbose=verbose, ) @@ -202,7 +208,8 @@ def parameters(self) -> Dict[str, Any]: """Returns parameters for smoke test.""" removed = [ "apply_norm", - "use_simmetric_normalized_laplacian" + "use_simmetric_normalized_laplacian", + "combiner" ] return dict( **{ diff --git a/embiggen/edge_label_prediction/edge_label_prediction_tensorflow/kipf_gcn.py b/embiggen/edge_label_prediction/edge_label_prediction_tensorflow/kipf_gcn.py index 341b1783..0e40c886 100644 --- a/embiggen/edge_label_prediction/edge_label_prediction_tensorflow/kipf_gcn.py +++ b/embiggen/edge_label_prediction/edge_label_prediction_tensorflow/kipf_gcn.py @@ -1,5 +1,5 @@ """Kipf GCN model for edge-label prediction.""" -from typing import List, Union, Optional, Type +from typing import List, Union, Optional, Type, Dict, Any from tensorflow.keras.optimizers import Optimizer from embiggen.edge_label_prediction.edge_label_prediction_tensorflow.gcn import GCNEdgeLabelPrediction @@ -40,6 +40,7 @@ def __init__( handling_multi_graph: str = "warn", node_feature_names: Optional[List[str]] = None, node_type_feature_names: Optional[List[str]] = None, + edge_feature_names: Optional[List[str]] = None, verbose: bool = False ): """Create new Kipf GCN object. @@ -164,6 +165,9 @@ def __init__( node_type_feature_names: Optional[List[str]] = None Names of the node type features. This is used as the layer names. + edge_feature_names: Optional[List[str]] = None + Names of the edge features. + This is used as the layer names. verbose: bool = False Whether to show loading bars. """ @@ -177,6 +181,7 @@ def __init__( number_of_units_per_ffnn_head_layer=number_of_units_per_ffnn_head_layer, dropout_rate=dropout_rate, apply_norm=apply_norm, + combiner="sum", edge_embedding_method=edge_embedding_method, optimizer=optimizer, early_stopping_min_delta=early_stopping_min_delta, @@ -199,9 +204,23 @@ def __init__( handling_multi_graph=handling_multi_graph, node_feature_names=node_feature_names, node_type_feature_names=node_type_feature_names, + edge_feature_names=edge_feature_names, verbose=verbose, ) + def parameters(self) -> Dict[str, Any]: + """Returns parameters for smoke test.""" + removed = [ + "combiner" + ] + return dict( + **{ + key: value + for key, value in super().parameters().items() + if key not in removed + } + ) + @classmethod def model_name(cls) -> str: return "Kipf GCN" \ No newline at end of file diff --git a/embiggen/edge_prediction/edge_prediction_ensmallen/perceptron.py b/embiggen/edge_prediction/edge_prediction_ensmallen/perceptron.py index 87b8b030..5407d194 100644 --- a/embiggen/edge_prediction/edge_prediction_ensmallen/perceptron.py +++ b/embiggen/edge_prediction/edge_prediction_ensmallen/perceptron.py @@ -3,6 +3,7 @@ from ensmallen import Graph import numpy as np from ensmallen import models +from embiggen.embedding_transformers import NodeTransformer from embiggen.edge_prediction.edge_prediction_model import AbstractEdgePredictionModel @@ -11,15 +12,18 @@ class PerceptronEdgePrediction(AbstractEdgePredictionModel): def __init__( self, - edge_features: Union[str, List[str]] = "JaccardCoefficient", - edge_embeddings: Optional[List[str]] = None, + edge_features: Optional[Union[str, List[str]]] = "JaccardCoefficient", + edge_embeddings: Optional[Union[str, List[str]]] = None, cooccurrence_iterations: int = 100, cooccurrence_window_size: int = 10, number_of_epochs: int = 100, - number_of_edges_per_mini_batch: int = 4096, + number_of_edges_per_mini_batch: int = 256, sample_only_edges_with_heterogeneous_node_types: bool = False, learning_rate: float = 0.001, - learning_rate_decay: float = 0.99, + first_order_decay_factor: float = 0.9, + second_order_decay_factor: float = 0.999, + avoid_false_negatives: bool = False, + use_scale_free_distribution: bool = True, random_state: int = 42, verbose: bool = True ): @@ -27,7 +31,7 @@ def __init__( Parameters ------------------------ - edge_features: Union[str, List[str]] = "JaccardCoefficient" + edge_features: Optional[Union[str, List[str]]] = "JaccardCoefficient" The edge features to compute for each edge. Zero or more edge features can be used at once. The currently supported edge features are: @@ -37,7 +41,7 @@ def __init__( - Cooccurrence, - ResourceAllocationIndex, - PreferentialAttachment, - edge_embeddings: Optional[List[str]] = None + edge_embeddings: Optional[Union[str, List[str]]] = None The embedding methods to use for the provided node features. Zero or more edge emmbedding methods can be used at once. The currently supported edge embedding are: @@ -59,23 +63,37 @@ def __init__( By default 100. number_of_epochs: int = 100 The number of epochs to train the model for. By default, 100. - number_of_edges_per_mini_batch: int = 1024 - The number of samples to include for each mini-batch. By default 1024. + number_of_edges_per_mini_batch: int = 4096 + The number of samples to include for each mini-batch. By default 4096. sample_only_edges_with_heterogeneous_node_types: bool = False Whether to sample negative edges only with source and destination nodes that have different node types. By default false. learning_rate: float = 0.001 Learning rate to use while training the model. By default 0.001. + first_order_decay_factor: float = 0.9 + First order decay factor for the first order momentum. + By default 0.9. + second_order_decay_factor: float = 0.999 + Second order decay factor for the second order momentum. + By default 0.999. + avoid_false_negatives: bool = False + Whether to avoid sampling false negatives. + This may cause a slower training. + use_scale_free_distribution: bool = True + Whether to train model using a scale free distribution for the negatives. random_state: int = 42 The random state to reproduce the model initialization and training. By default, 42. verbose: bool = True Whether to show epochs loading bar. """ super().__init__(random_state=random_state) - + if isinstance(edge_features, str): edge_features = [edge_features] + if isinstance(edge_embeddings, str): + edge_embeddings = [edge_embeddings] + self._model_kwargs = dict( edge_features=edge_features, edge_embeddings=edge_embeddings, @@ -85,7 +103,10 @@ def __init__( number_of_edges_per_mini_batch=number_of_edges_per_mini_batch, sample_only_edges_with_heterogeneous_node_types=sample_only_edges_with_heterogeneous_node_types, learning_rate=learning_rate, - learning_rate_decay=learning_rate_decay + first_order_decay_factor=first_order_decay_factor, + second_order_decay_factor=second_order_decay_factor, + avoid_false_negatives=avoid_false_negatives, + use_scale_free_distribution=use_scale_free_distribution, ) self._verbose = verbose self._model = models.EdgePredictionPerceptron( @@ -136,12 +157,26 @@ def _fit( new_node_features = [] if node_features is not None: for node_feature in node_features: - if node_feature.dtype != np.float32: - node_feature = node_feature.astype(np.float32) if not node_feature.data.c_contiguous: node_feature = np.ascontiguousarray(node_feature) new_node_features.append(node_feature) + if node_type_features is not None: + for node_type_feature in node_type_features: + node_transformer = NodeTransformer( + aligned_mapping=True + ) + node_transformer.fit( + node_type_feature=node_type_feature + ) + node_feature = node_transformer.transform( + node_types=graph + ) + if not node_feature.data.c_contiguous: + node_feature = np.ascontiguousarray(node_feature) + new_node_features.append(node_feature) + + self._model.fit( graph=graph, node_features=new_node_features, @@ -212,8 +247,6 @@ def _predict_proba( new_node_features = [] if node_features is not None: for node_feature in node_features: - if node_feature.dtype != np.float32: - node_feature = node_feature.astype(np.float32) if not node_feature.data.c_contiguous: node_feature = np.ascontiguousarray(node_feature) new_node_features.append(node_feature) diff --git a/embiggen/edge_prediction/edge_prediction_evaluation.py b/embiggen/edge_prediction/edge_prediction_evaluation.py index 4d50a0f8..c8bab4f3 100644 --- a/embiggen/edge_prediction/edge_prediction_evaluation.py +++ b/embiggen/edge_prediction/edge_prediction_evaluation.py @@ -23,7 +23,7 @@ def edge_prediction_evaluation( versions: Optional[Union[str, List[str]]] = None, validation_sample_only_edges_with_heterogeneous_node_types: bool = False, validation_unbalance_rates: Tuple[float] = (1.0, ), - use_zipfian_sampling: bool = True, + use_scale_free_distribution: bool = True, enable_cache: bool = False, precompute_constant_automatic_stocastic_features: bool = False, smoke_test: bool = False, @@ -92,10 +92,10 @@ def edge_prediction_evaluation( This can be useful when executing a bipartite edge prediction task. validation_unbalance_rates: Tuple[float] = (1.0, ) Unbalance rate for the non-existent graphs generation. - use_zipfian_sampling: bool = True - Whether to use the zipfian sampling of the NEGATIVE edges for the EVALUATION + use_scale_free_distribution: bool = True + Whether to use the scale free sampling of the NEGATIVE edges for the EVALUATION of the edge prediction performance of the provided models. - Please DO BE ADVISED that not using a zipfian sampling for the negative + Please DO BE ADVISED that not using a scale free sampling for the negative edges is a poor choice and will cause a significant positive bias in the model performance. enable_cache: bool = False @@ -138,5 +138,5 @@ def edge_prediction_evaluation( verbose=verbose, validation_sample_only_edges_with_heterogeneous_node_types=validation_sample_only_edges_with_heterogeneous_node_types, validation_unbalance_rates=validation_unbalance_rates, - use_zipfian_sampling=use_zipfian_sampling + use_scale_free_distribution=use_scale_free_distribution ) diff --git a/embiggen/edge_prediction/edge_prediction_model.py b/embiggen/edge_prediction/edge_prediction_model.py index 6e015de8..79950a1b 100644 --- a/embiggen/edge_prediction/edge_prediction_model.py +++ b/embiggen/edge_prediction/edge_prediction_model.py @@ -6,7 +6,7 @@ import math from ensmallen import Graph from tqdm.auto import tqdm -from embiggen.utils.abstract_models import AbstractClassifierModel, AbstractEmbeddingModel, abstract_class, format_list +from embiggen.utils.abstract_models import AbstractClassifierModel, abstract_class, format_list @abstract_class @@ -82,10 +82,13 @@ def split_graph_following_evaluation_schema( random_state=random_state, verbose=False ) - raise ValueError( - f"The requested evaluation schema `{evaluation_schema}` " - "is not available. The available evaluation schemas " - f"are: {format_list(cls.get_available_evaluation_schemas())}." + super().split_graph_following_evaluation_schema( + graph=graph, + evaluation_schema=evaluation_schema, + random_state=random_state, + holdout_number=holdout_number, + number_of_holdouts=number_of_holdouts, + **holdouts_kwargs, ) @staticmethod @@ -99,7 +102,7 @@ def __iterate_negative_graphs( verbose: bool, validation_sample_only_edges_with_heterogeneous_node_types: bool, validation_unbalance_rates: Tuple[float], - use_zipfian_sampling: bool + use_scale_free_distribution: bool ) -> Iterator[Tuple[Graph]]: """Return iterator over the negative graphs for evaluation.""" if subgraph_of_interest is None: @@ -107,9 +110,9 @@ def __iterate_negative_graphs( else: sampler_graph = subgraph_of_interest - if not use_zipfian_sampling: + if not use_scale_free_distribution: warnings.warn( - "Please do be advised that you have DISABLED the use of zipfian sampling " + "Please do be advised that you have DISABLED the use of scale free sampling " "for the negative edges for the EVALUATION (not the training) " "of a model. This is a POOR CHOICE as it will introduce a positive bias " "as edges sampled uniformely have a significantly different node degree " @@ -130,7 +133,7 @@ def __iterate_negative_graphs( ), random_state=random_state*(i+1), sample_only_edges_with_heterogeneous_node_types=validation_sample_only_edges_with_heterogeneous_node_types, - use_zipfian_sampling=use_zipfian_sampling, + use_scale_free_distribution=use_scale_free_distribution, support=support, graph_to_avoid=graph ).random_holdout( @@ -160,7 +163,7 @@ def _prepare_evaluation( verbose: bool = True, validation_sample_only_edges_with_heterogeneous_node_types: bool = False, validation_unbalance_rates: Tuple[float] = (1.0, ), - use_zipfian_sampling: bool = True + use_scale_free_distribution: bool = True ) -> Dict[str, Any]: """Return additional custom parameters for the current holdout.""" return dict( @@ -174,7 +177,7 @@ def _prepare_evaluation( verbose=verbose, validation_sample_only_edges_with_heterogeneous_node_types=validation_sample_only_edges_with_heterogeneous_node_types, validation_unbalance_rates=validation_unbalance_rates, - use_zipfian_sampling=use_zipfian_sampling + use_scale_free_distribution=use_scale_free_distribution )) ) @@ -193,7 +196,7 @@ def _evaluate( negative_graphs: Optional[List[Tuple[Graph]]] = None, validation_sample_only_edges_with_heterogeneous_node_types: bool = False, validation_unbalance_rates: Tuple[float] = (1.0, ), - use_zipfian_sampling: bool = True, + use_scale_free_distribution: bool = True, ) -> List[Dict[str, Any]]: """Return model evaluation on the provided graphs.""" performance = [] @@ -206,9 +209,6 @@ def _evaluate( edge_features=edge_features ) - if len(train_predic_proba.shape) > 1 and train_predic_proba.shape[1] > 1: - train_predic_proba = train_predic_proba[:, 1] - test_predict_proba = self.predict_proba( test, support=support, @@ -217,9 +217,6 @@ def _evaluate( edge_features=edge_features ) - if len(test_predict_proba.shape) > 1 and test_predict_proba.shape[1] > 1: - test_predict_proba = test_predict_proba[:, 1] - negative_graph_iterator = self.__iterate_negative_graphs( graph=graph, train=train, @@ -230,7 +227,7 @@ def _evaluate( verbose=verbose, validation_sample_only_edges_with_heterogeneous_node_types=validation_sample_only_edges_with_heterogeneous_node_types, validation_unbalance_rates=validation_unbalance_rates, - use_zipfian_sampling=use_zipfian_sampling + use_scale_free_distribution=use_scale_free_distribution ) if negative_graphs is None else negative_graphs for unbalance_rate, (negative_train, negative_test) in tqdm( @@ -269,6 +266,7 @@ def _evaluate( performance.append({ "evaluation_mode": evaluation_mode, "validation_unbalance_rate": unbalance_rate, + "use_scale_free_distribution": use_scale_free_distribution, "validation_sample_only_edges_with_heterogeneous_node_types": validation_sample_only_edges_with_heterogeneous_node_types, **self.evaluate_predictions( labels, @@ -322,7 +320,7 @@ def predict( support=support, node_features=node_features, node_type_features=node_type_features - ) + ).flatten() if return_predictions_dataframe: predictions = pd.DataFrame( @@ -751,7 +749,12 @@ def predict_proba( support=support, node_features=node_features, node_type_features=node_type_features - ) + ).flatten() + + if np.isnan(predictions).any(): + raise ValueError( + "There are NaN values in the predicted probabilities!" + ) if return_predictions_dataframe: predictions = pd.DataFrame( diff --git a/embiggen/edge_prediction/edge_prediction_sklearn/decision_tree_edge_prediction.py b/embiggen/edge_prediction/edge_prediction_sklearn/decision_tree_edge_prediction.py index 274f2eaa..b4023b56 100644 --- a/embiggen/edge_prediction/edge_prediction_sklearn/decision_tree_edge_prediction.py +++ b/embiggen/edge_prediction/edge_prediction_sklearn/decision_tree_edge_prediction.py @@ -11,20 +11,19 @@ def __init__( self, criterion="gini", splitter="best", - max_depth=None, + max_depth=10, min_samples_split=2, min_samples_leaf=1, min_weight_fraction_leaf=0., max_features=None, max_leaf_nodes=None, min_impurity_decrease=0., - class_weight=None, ccp_alpha=0.0, edge_embedding_method: str = "Concatenate", training_unbalance_rate: float = 1.0, training_sample_only_edges_with_heterogeneous_node_types: bool = False, use_edge_metrics: bool = False, - use_zipfian_sampling: bool = True, + use_scale_free_distribution: bool = True, prediction_batch_size: int = 2**12, random_state: int = 42 ): @@ -39,7 +38,6 @@ def __init__( self._max_leaf_nodes = max_leaf_nodes self._min_impurity_decrease = min_impurity_decrease self._random_state = random_state - self._class_weight = class_weight self._ccp_alpha = ccp_alpha super().__init__( @@ -54,13 +52,12 @@ def __init__( max_leaf_nodes=max_leaf_nodes, min_impurity_decrease=min_impurity_decrease, random_state=random_state, - class_weight=class_weight, ccp_alpha=ccp_alpha, ), edge_embedding_method=edge_embedding_method, training_unbalance_rate=training_unbalance_rate, use_edge_metrics=use_edge_metrics, - use_zipfian_sampling=use_zipfian_sampling, + use_scale_free_distribution=use_scale_free_distribution, training_sample_only_edges_with_heterogeneous_node_types=training_sample_only_edges_with_heterogeneous_node_types, prediction_batch_size=prediction_batch_size, random_state=random_state @@ -81,7 +78,6 @@ def parameters(self) -> Dict[str, Any]: max_leaf_nodes=self._max_leaf_nodes, min_impurity_decrease=self._min_impurity_decrease, random_state=self._random_state, - class_weight=self._class_weight, ccp_alpha=self._ccp_alpha, ) } diff --git a/embiggen/edge_prediction/edge_prediction_sklearn/extra_trees_edge_prediction.py b/embiggen/edge_prediction/edge_prediction_sklearn/extra_trees_edge_prediction.py index 77259a02..7b93da62 100644 --- a/embiggen/edge_prediction/edge_prediction_sklearn/extra_trees_edge_prediction.py +++ b/embiggen/edge_prediction/edge_prediction_sklearn/extra_trees_edge_prediction.py @@ -19,20 +19,18 @@ def __init__( max_features="auto", max_leaf_nodes=None, min_impurity_decrease=0., - min_impurity_split=None, bootstrap=True, oob_score=False, n_jobs=-1, verbose=0, warm_start=False, - class_weight=None, ccp_alpha=0.0, max_samples=None, edge_embedding_method: str = "Concatenate", training_unbalance_rate: float = 1.0, training_sample_only_edges_with_heterogeneous_node_types: bool = False, use_edge_metrics: bool = False, - use_zipfian_sampling: bool = True, + use_scale_free_distribution: bool = True, prediction_batch_size: int = 2**12, random_state: int = 42 ): @@ -46,14 +44,12 @@ def __init__( self._max_features = max_features self._max_leaf_nodes = max_leaf_nodes self._min_impurity_decrease = min_impurity_decrease - self._min_impurity_split = min_impurity_split self._bootstrap = bootstrap self._oob_score = oob_score self._n_jobs = n_jobs self._random_state = random_state self._verbose = verbose self._warm_start = warm_start - self._class_weight = class_weight self._ccp_alpha = ccp_alpha self._max_samples = max_samples @@ -68,21 +64,19 @@ def __init__( max_features=max_features, max_leaf_nodes=max_leaf_nodes, min_impurity_decrease=min_impurity_decrease, - min_impurity_split=min_impurity_split, bootstrap=bootstrap, oob_score=oob_score, n_jobs=n_jobs, random_state=random_state, verbose=verbose, warm_start=warm_start, - class_weight=class_weight, ccp_alpha=ccp_alpha, max_samples=max_samples ), edge_embedding_method=edge_embedding_method, training_unbalance_rate=training_unbalance_rate, use_edge_metrics=use_edge_metrics, - use_zipfian_sampling=use_zipfian_sampling, + use_scale_free_distribution=use_scale_free_distribution, training_sample_only_edges_with_heterogeneous_node_types=training_sample_only_edges_with_heterogeneous_node_types, prediction_batch_size=prediction_batch_size, random_state=random_state @@ -110,14 +104,12 @@ def parameters(self) -> Dict[str, Any]: max_features = self._max_features, max_leaf_nodes = self._max_leaf_nodes, min_impurity_decrease = self._min_impurity_decrease, - min_impurity_split = self._min_impurity_split, bootstrap = self._bootstrap, oob_score = self._oob_score, n_jobs = self._n_jobs, random_state = self._random_state, verbose = self._verbose, warm_start = self._warm_start, - class_weight = self._class_weight, ccp_alpha = self._ccp_alpha, max_samples = self._max_samples, ) diff --git a/embiggen/edge_prediction/edge_prediction_sklearn/gradient_boosting_edge_prediction.py b/embiggen/edge_prediction/edge_prediction_sklearn/gradient_boosting_edge_prediction.py index e94206c6..63f9b56f 100644 --- a/embiggen/edge_prediction/edge_prediction_sklearn/gradient_boosting_edge_prediction.py +++ b/embiggen/edge_prediction/edge_prediction_sklearn/gradient_boosting_edge_prediction.py @@ -19,7 +19,6 @@ def __init__( min_weight_fraction_leaf=0., max_depth=3, min_impurity_decrease=0., - min_impurity_split=None, init=None, max_features=None, verbose=0, @@ -33,27 +32,26 @@ def __init__( training_unbalance_rate: float = 1.0, training_sample_only_edges_with_heterogeneous_node_types: bool = False, use_edge_metrics: bool = False, - use_zipfian_sampling: bool = True, + use_scale_free_distribution: bool = True, prediction_batch_size: int = 2**12, random_state: int = 42 ): """Create the Gradient Boosting for Edge Prediction.""" self._loss=loss self._learning_rate=learning_rate - self._n_estimators=n_estimators, + self._n_estimators=n_estimators self._criterion=criterion - self._min_samples_split=min_samples_split, - self._min_samples_leaf=min_samples_leaf, - self._min_weight_fraction_leaf=min_weight_fraction_leaf, + self._min_samples_split=min_samples_split + self._min_samples_leaf=min_samples_leaf + self._min_weight_fraction_leaf=min_weight_fraction_leaf self._max_depth=max_depth self._init=init - self._subsample=subsample, - self._max_features=max_features, - self._max_leaf_nodes=max_leaf_nodes, - self._min_impurity_decrease=min_impurity_decrease, - self._min_impurity_split=min_impurity_split, + self._subsample=subsample + self._max_features=max_features + self._max_leaf_nodes=max_leaf_nodes + self._min_impurity_decrease=min_impurity_decrease self._warm_start=warm_start - self._validation_fraction=validation_fraction, + self._validation_fraction=validation_fraction self._n_iter_no_change=n_iter_no_change self._tol=tol self._ccp_alpha=ccp_alpha @@ -68,14 +66,13 @@ def __init__( random_state=random_state, verbose=verbose, max_leaf_nodes=max_leaf_nodes, min_impurity_decrease=min_impurity_decrease, - min_impurity_split=min_impurity_split, warm_start=warm_start, validation_fraction=validation_fraction, n_iter_no_change=n_iter_no_change, tol=tol, ccp_alpha=ccp_alpha ), edge_embedding_method=edge_embedding_method, training_unbalance_rate=training_unbalance_rate, use_edge_metrics=use_edge_metrics, - use_zipfian_sampling=use_zipfian_sampling, + use_scale_free_distribution=use_scale_free_distribution, training_sample_only_edges_with_heterogeneous_node_types=training_sample_only_edges_with_heterogeneous_node_types, prediction_batch_size=prediction_batch_size, random_state=random_state @@ -99,7 +96,6 @@ def parameters(self) -> Dict[str, Any]: max_features=self._max_features, max_leaf_nodes=self._max_leaf_nodes, min_impurity_decrease=self._min_impurity_decrease, - min_impurity_split=self._min_impurity_split, warm_start=self._warm_start, validation_fraction=self._validation_fraction, n_iter_no_change=self._n_iter_no_change, diff --git a/embiggen/edge_prediction/edge_prediction_sklearn/mlp_edge_prediction.py b/embiggen/edge_prediction/edge_prediction_sklearn/mlp_edge_prediction.py index 40b3d614..82ae8c52 100644 --- a/embiggen/edge_prediction/edge_prediction_sklearn/mlp_edge_prediction.py +++ b/embiggen/edge_prediction/edge_prediction_sklearn/mlp_edge_prediction.py @@ -35,7 +35,7 @@ def __init__( training_unbalance_rate: float = 1.0, training_sample_only_edges_with_heterogeneous_node_types: bool = False, use_edge_metrics: bool = False, - use_zipfian_sampling: bool = True, + use_scale_free_distribution: bool = True, prediction_batch_size: int = 2**12, random_state: int = 42 ): @@ -90,7 +90,7 @@ def __init__( edge_embedding_method=edge_embedding_method, training_unbalance_rate=training_unbalance_rate, use_edge_metrics=use_edge_metrics, - use_zipfian_sampling=use_zipfian_sampling, + use_scale_free_distribution=use_scale_free_distribution, training_sample_only_edges_with_heterogeneous_node_types=training_sample_only_edges_with_heterogeneous_node_types, prediction_batch_size=prediction_batch_size, random_state=random_state diff --git a/embiggen/edge_prediction/edge_prediction_sklearn/random_forest_edge_prediction.py b/embiggen/edge_prediction/edge_prediction_sklearn/random_forest_edge_prediction.py index 39b6cb10..74511f36 100644 --- a/embiggen/edge_prediction/edge_prediction_sklearn/random_forest_edge_prediction.py +++ b/embiggen/edge_prediction/edge_prediction_sklearn/random_forest_edge_prediction.py @@ -16,7 +16,7 @@ def __init__( min_samples_split: int = 2, min_samples_leaf: int = 1, min_weight_fraction_leaf: float = 0., - max_features="auto", + max_features="sqrt", max_leaf_nodes=None, min_impurity_decrease=0., bootstrap=True, @@ -24,14 +24,13 @@ def __init__( n_jobs=-1, verbose=0, warm_start=False, - class_weight=None, ccp_alpha=0.0, max_samples=None, edge_embedding_method: str = "Concatenate", training_unbalance_rate: float = 1.0, training_sample_only_edges_with_heterogeneous_node_types: bool = False, use_edge_metrics: bool = False, - use_zipfian_sampling: bool = True, + use_scale_free_distribution: bool = True, prediction_batch_size: int = 2**12, random_state: int = 42 ): @@ -51,7 +50,6 @@ def __init__( self._random_state = random_state self._verbose = verbose self._warm_start = warm_start - self._class_weight = class_weight self._ccp_alpha = ccp_alpha self._max_samples = max_samples @@ -72,14 +70,13 @@ def __init__( random_state=random_state, verbose=verbose, warm_start=warm_start, - class_weight=class_weight, ccp_alpha=ccp_alpha, max_samples=max_samples ), edge_embedding_method=edge_embedding_method, training_unbalance_rate=training_unbalance_rate, use_edge_metrics=use_edge_metrics, - use_zipfian_sampling=use_zipfian_sampling, + use_scale_free_distribution=use_scale_free_distribution, training_sample_only_edges_with_heterogeneous_node_types=training_sample_only_edges_with_heterogeneous_node_types, prediction_batch_size=prediction_batch_size, random_state=random_state @@ -113,7 +110,6 @@ def parameters(self) -> Dict[str, Any]: random_state = self._random_state, verbose = self._verbose, warm_start = self._warm_start, - class_weight = self._class_weight, ccp_alpha = self._ccp_alpha, max_samples = self._max_samples, ) diff --git a/embiggen/edge_prediction/edge_prediction_sklearn/sklearn_edge_prediction_adapter.py b/embiggen/edge_prediction/edge_prediction_sklearn/sklearn_edge_prediction_adapter.py index 94bb955b..0ae0d0ea 100644 --- a/embiggen/edge_prediction/edge_prediction_sklearn/sklearn_edge_prediction_adapter.py +++ b/embiggen/edge_prediction/edge_prediction_sklearn/sklearn_edge_prediction_adapter.py @@ -24,8 +24,8 @@ def __init__( edge_embedding_method: str = "Concatenate", training_unbalance_rate: float = 1.0, training_sample_only_edges_with_heterogeneous_node_types: bool = False, + use_scale_free_distribution: bool = True, use_edge_metrics: bool = False, - use_zipfian_sampling: bool = True, prediction_batch_size: int = 2**15, random_state: int = 42 ): @@ -43,7 +43,7 @@ def __init__( Whether to sample negative edges exclusively between nodes with different node types to generate the negative edges used during the training of the model. This can be useful when executing a bipartite edge prediction task. - use_zipfian_sampling: bool = True + use_scale_free_distribution: bool = True Whether to sample the negative edges for the TRAINING of the model using a zipfian-like distribution that follows the degree distribution of the graph. This is generally useful, as these negative edges are less @@ -76,9 +76,8 @@ def __init__( self._training_unbalance_rate = training_unbalance_rate self._prediction_batch_size = prediction_batch_size self._use_edge_metrics = use_edge_metrics - self._use_zipfian_sampling = use_zipfian_sampling + self._use_scale_free_distribution = use_scale_free_distribution self._training_sample_only_edges_with_heterogeneous_node_types = training_sample_only_edges_with_heterogeneous_node_types - self._support = None # We want to mask the decorator class name self.__class__.__name__ = model_instance.__class__.__name__ self.__class__.__doc__ = model_instance.__class__.__doc__ @@ -91,7 +90,7 @@ def parameters(self) -> Dict[str, Any]: "training_unbalance_rate": self._training_unbalance_rate, "prediction_batch_size": self._prediction_batch_size, "use_edge_metrics": self._use_edge_metrics, - "use_zipfian_sampling": self._use_zipfian_sampling, + "use_scale_free_distribution": self._use_scale_free_distribution, **super().parameters() } @@ -108,6 +107,8 @@ def _trasform_graph_into_edge_embedding( self, graph: Union[Graph, np.ndarray], node_features: List[np.ndarray], + support: Optional[Graph] = None, + node_types: Optional[Union[Graph, List[Optional[List[str]]], List[Optional[List[int]]]]] = None, node_type_features: Optional[List[np.ndarray]] = None, ) -> np.ndarray: """Transforms the provided data into an Sklearn-compatible numpy array. @@ -119,6 +120,16 @@ def _trasform_graph_into_edge_embedding( It can either be an Graph or a list of lists of edges. node_features: List[np.ndarray] The node features to be used in the training of the model. + support: Optional[Graph] = None + The graph describiding the topological structure that + includes also the above graph. This parameter + is mostly useful for topological classifiers + such as Graph Convolutional Networks. + node_types: Optional[Union[Graph, List[Optional[List[str]]], List[Optional[List[int]]]]] = None, + List of node types whose embedding is to be returned. + This can be either a list of strings, or a graph, or if the + aligned_mapping is setted, then this methods also accepts + a list of ints. node_type_features: Optional[List[np.ndarray]] = None, The node type features to be used in the training of the model. @@ -132,25 +143,20 @@ def _trasform_graph_into_edge_embedding( ValueError If the two graphs do not share the same node vocabulary. """ - if node_type_features is not None: - raise NotImplementedError( - "Support for node type features is not currently available for any " - "of the edge prediction models from the Sklearn library." - ) gt = GraphTransformer( method=self._edge_embedding_method, - aligned_node_mapping=True + aligned_mapping=True ) if self._use_edge_metrics: if isinstance(graph, Graph): - edge_features = self._support.get_all_edge_metrics( + edge_features = support.get_all_edge_metrics( normalize=True, subgraph=graph, ) elif isinstance(graph, tuple): - edge_features = self._support.get_all_edge_metrics_from_node_ids( + edge_features = support.get_all_edge_metrics_from_node_ids( *graph, normalize=True, ) @@ -161,10 +167,14 @@ def _trasform_graph_into_edge_embedding( else: edge_features = None - gt.fit(node_features) + gt.fit( + node_features, + node_type_feature=node_type_features + ) return gt.transform( graph=graph, + node_types=node_types, edge_features=edge_features ) @@ -196,10 +206,13 @@ def _fit( """ lpt = EdgePredictionTransformer( method=self._edge_embedding_method, - aligned_node_mapping=True + aligned_mapping=True ) - lpt.fit(node_features) + lpt.fit( + node_features, + node_type_feature=node_type_features + ) if support is None: support = graph @@ -211,11 +224,11 @@ def _fit( ), random_state=self._random_state, sample_only_edges_with_heterogeneous_node_types=self._training_sample_only_edges_with_heterogeneous_node_types, - use_zipfian_sampling=self._use_zipfian_sampling + use_scale_free_distribution=self._use_scale_free_distribution ) if self._use_edge_metrics: - self._support = support + support = support edge_features = np.vstack(( support.get_all_edge_metrics( normalize=True, @@ -272,7 +285,9 @@ def _predict( return np.concatenate([ self._model_instance.predict(self._trasform_graph_into_edge_embedding( graph=edges[0], + support=support, node_features=node_features, + node_types=graph, node_type_features=node_type_features, )) for edges in tqdm( @@ -318,10 +333,12 @@ def _predict_proba( use_edge_metrics=False, batch_size=self._prediction_batch_size ) - return np.concatenate([ + prediction_probabilities = np.concatenate([ self._model_instance.predict_proba(self._trasform_graph_into_edge_embedding( graph=edges[0], + support=support, node_features=node_features, + node_types=graph, node_type_features=node_type_features, )) for edges in tqdm( @@ -333,6 +350,14 @@ def _predict_proba( ) ]) + # In the majority but not totality of sklearn models, + # the predictions of binary models are returned as + # a couple of vectors for the positive and negative class. + if len(prediction_probabilities.shape) > 1 and prediction_probabilities.shape[1] > 1: + prediction_probabilities = prediction_probabilities[:, 1] + + return prediction_probabilities + @classmethod def can_use_edge_weights(cls) -> bool: """Returns whether the model can optionally use edge weights.""" diff --git a/embiggen/edge_prediction/edge_prediction_tensorflow/gcn.py b/embiggen/edge_prediction/edge_prediction_tensorflow/gcn.py index 7e1b2dc5..11b7c30f 100644 --- a/embiggen/edge_prediction/edge_prediction_tensorflow/gcn.py +++ b/embiggen/edge_prediction/edge_prediction_tensorflow/gcn.py @@ -27,6 +27,7 @@ def __init__( number_of_units_per_ffnn_head_layer: Union[int, List[int]] = 128, dropout_rate: float = 0.2, apply_norm: bool = False, + combiner: str = "mean", edge_embedding_method: str = "Concatenate", optimizer: Union[str, Optimizer] = "adam", early_stopping_min_delta: float = 0.0001, @@ -81,6 +82,13 @@ def __init__( apply_norm: bool = False Whether to normalize the output of the convolution operations, after applying the level activations. + combiner: str = "mean" + A string specifying the reduction op. + Currently "mean", "sqrtn" and "sum" are supported. + "sum" computes the weighted sum of the embedding results for each row. + "mean" is the weighted sum divided by the total weight. + "sqrtn" is the weighted sum divided by the square root of the sum of the squares of the weights. + Defaults to mean. edge_embedding_method: str = "Concatenate" The edge embedding method to use to put togheter the source and destination node features, which includes: @@ -186,6 +194,7 @@ def __init__( number_of_units_per_ffnn_body_layer=number_of_units_per_ffnn_body_layer, number_of_units_per_ffnn_head_layer=number_of_units_per_ffnn_head_layer, dropout_rate=dropout_rate, + combiner=combiner, apply_norm=apply_norm, edge_embedding_method=edge_embedding_method, optimizer=optimizer, diff --git a/embiggen/edge_prediction/edge_prediction_tensorflow/gnn.py b/embiggen/edge_prediction/edge_prediction_tensorflow/gnn.py index a875709e..cda0e418 100644 --- a/embiggen/edge_prediction/edge_prediction_tensorflow/gnn.py +++ b/embiggen/edge_prediction/edge_prediction_tensorflow/gnn.py @@ -196,14 +196,17 @@ def parameters(self) -> Dict[str, Any]: "number_of_units_per_graph_convolution_layers", "handling_multi_graph", "number_of_units_per_ffnn_body_layer", - "number_of_units_per_ffnn_head_layer" + "number_of_units_per_ffnn_head_layer", + "combiner", + "apply_norm", + "dropout_rate" ] return dict( number_of_units_per_body_layer=self._number_of_units_per_ffnn_body_layer, number_of_units_per_head_layer=self._number_of_units_per_ffnn_head_layer, **{ key: value - for key, value in super().smoke_test_parameters().items() + for key, value in super().parameters().items() if key not in removed } ) diff --git a/embiggen/edge_prediction/edge_prediction_tensorflow/graph_sage.py b/embiggen/edge_prediction/edge_prediction_tensorflow/graph_sage.py index 67c1f9e1..d9923aa0 100644 --- a/embiggen/edge_prediction/edge_prediction_tensorflow/graph_sage.py +++ b/embiggen/edge_prediction/edge_prediction_tensorflow/graph_sage.py @@ -170,6 +170,7 @@ def __init__( number_of_units_per_ffnn_body_layer=number_of_units_per_ffnn_body_layer, number_of_units_per_ffnn_head_layer=number_of_units_per_ffnn_head_layer, dropout_rate=dropout_rate, + combiner="mean", apply_norm=True, edge_embedding_method=edge_embedding_method, optimizer=optimizer, @@ -202,7 +203,8 @@ def parameters(self) -> Dict[str, Any]: """Returns parameters for smoke test.""" removed = [ "apply_norm", - "use_simmetric_normalized_laplacian" + "use_simmetric_normalized_laplacian", + "combiner" ] return dict( **{ diff --git a/embiggen/edge_prediction/edge_prediction_tensorflow/kipf_gcn.py b/embiggen/edge_prediction/edge_prediction_tensorflow/kipf_gcn.py index 01f1b19c..e82a3c04 100644 --- a/embiggen/edge_prediction/edge_prediction_tensorflow/kipf_gcn.py +++ b/embiggen/edge_prediction/edge_prediction_tensorflow/kipf_gcn.py @@ -1,5 +1,5 @@ """Kipf GCN model for edge prediction.""" -from typing import List, Union, Optional, Type +from typing import List, Union, Optional, Type, Dict, Any from tensorflow.keras.optimizers import Optimizer from embiggen.edge_prediction.edge_prediction_tensorflow.gcn import GCNEdgePrediction @@ -174,6 +174,7 @@ def __init__( number_of_units_per_ffnn_body_layer=number_of_units_per_ffnn_body_layer, number_of_units_per_ffnn_head_layer=number_of_units_per_ffnn_head_layer, dropout_rate=dropout_rate, + combiner="sum", apply_norm=apply_norm, edge_embedding_method=edge_embedding_method, optimizer=optimizer, @@ -202,6 +203,19 @@ def __init__( verbose=verbose, ) + def parameters(self) -> Dict[str, Any]: + """Returns parameters for smoke test.""" + removed = [ + "combiner" + ] + return dict( + **{ + key: value + for key, value in super().parameters().items() + if key not in removed + } + ) + @classmethod def model_name(cls) -> str: return "Kipf GCN" \ No newline at end of file diff --git a/embiggen/embedders/__init__.py b/embiggen/embedders/__init__.py index 089bdf62..5e9b2513 100644 --- a/embiggen/embedders/__init__.py +++ b/embiggen/embedders/__init__.py @@ -2,6 +2,7 @@ from embiggen.embedders.ensmallen_embedders import * from embiggen.embedders.tensorflow_embedders import * from embiggen.embedders.pykeen_embedders import * +from embiggen.embedders.non_existent_embedders import * from embiggen.embedders.karateclub_embedders import * from embiggen.embedders.graph_embedding_pipeline import embed_graph diff --git a/embiggen/embedders/ensmallen_embedders/deepwalk_cbow.py b/embiggen/embedders/ensmallen_embedders/deepwalk_cbow.py new file mode 100644 index 00000000..44d72145 --- /dev/null +++ b/embiggen/embedders/ensmallen_embedders/deepwalk_cbow.py @@ -0,0 +1,123 @@ +"""Module providing DeepWalk CBOW model implementation.""" +from typing import Optional, Dict, Any +from embiggen.embedders.ensmallen_embedders.node2vec import Node2VecEnsmallen + + +class DeepWalkCBOWEnsmallen(Node2VecEnsmallen): + """Class providing DeepWalk CBOW implemeted in Rust from Ensmallen.""" + + def __init__( + self, + embedding_size: int = 100, + epochs: int = 30, + clipping_value: float = 6.0, + number_of_negative_samples: int = 10, + walk_length: int = 128, + iterations: int = 10, + window_size: int = 5, + max_neighbours: Optional[int] = 100, + learning_rate: float = 0.01, + learning_rate_decay: float = 0.9, + central_nodes_embedding_path: Optional[str] = None, + contextual_nodes_embedding_path: Optional[str] = None, + normalize_by_degree: bool = False, + stochastic_downsample_by_degree: Optional[bool] = False, + normalize_learning_rate_by_degree: Optional[bool] = False, + use_scale_free_distribution: Optional[bool] = True, + random_state: int = 42, + enable_cache: bool = False + ): + """Create new abstract DeepWalk method. + + Parameters + -------------------------- + embedding_size: int = 100 + Dimension of the embedding. + epochs: int = 30 + Number of epochs to train the model for. + clipping_value: float = 6.0 + Value at which we clip the dot product, mostly for numerical stability issues. + By default, `6.0`, where the loss is already close to zero. + number_of_negative_samples: int = 10 + The number of negative classes to randomly sample per batch. + This single sample of negative classes is evaluated for each element in the batch. + walk_length: int = 128 + Maximal length of the walks. + iterations: int = 10 + Number of iterations of the single walks. + window_size: int = 5 + Window size for the local context. + On the borders the window size is trimmed. + max_neighbours: Optional[int] = 100 + Number of maximum neighbours to consider when using approximated walks. + By default, None, we execute exact random walks. + This is mainly useful for graphs containing nodes with high degrees. + learning_rate: float = 0.01 + The learning rate to use to train the Node2Vec model. By default 0.01. + learning_rate_decay: float = 0.9 + Factor to reduce the learning rate for at each epoch. By default 0.9. + central_nodes_embedding_path: Optional[str] = None + Path where to mmap and store the central nodes embedding. + This is necessary to embed large graphs whose embedding will not + fit into the available main memory. + contextual_nodes_embedding_path: Optional[str] = None + Path where to mmap and store the central nodes embedding. + This is necessary to embed large graphs whose embedding will not + fit into the available main memory. + normalize_by_degree: bool = False + Whether to normalize the random walk by the node degree + of the destination node degrees. + stochastic_downsample_by_degree: Optional[bool] = False + Randomly skip samples with probability proportional to the degree of the central node. By default false. + normalize_learning_rate_by_degree: Optional[bool] = False + Divide the learning rate by the degree of the central node. By default false. + use_scale_free_distribution: Optional[bool] = True + Sample negatives proportionally to their degree. By default true. + random_state: int = 42 + The random state to reproduce the training sequence. + enable_cache: bool = False + Whether to enable the cache, that is to + store the computed embedding. + """ + super().__init__( + embedding_size=embedding_size, + epochs=epochs, + clipping_value=clipping_value, + number_of_negative_samples=number_of_negative_samples, + walk_length=walk_length, + iterations=iterations, + window_size=window_size, + max_neighbours=max_neighbours, + learning_rate=learning_rate, + learning_rate_decay=learning_rate_decay, + central_nodes_embedding_path=central_nodes_embedding_path, + contextual_nodes_embedding_path=contextual_nodes_embedding_path, + normalize_by_degree=normalize_by_degree, + stochastic_downsample_by_degree=stochastic_downsample_by_degree, + normalize_learning_rate_by_degree=normalize_learning_rate_by_degree, + use_scale_free_distribution=use_scale_free_distribution, + random_state=random_state, + enable_cache=enable_cache + ) + + def parameters(self) -> Dict[str, Any]: + """Returns parameters for smoke test.""" + removed = [ + "return_weight", + "explore_weight", + "change_node_type_weight", + "change_edge_type_weight", + "alpha" + ] + return dict( + **{ + key: value + for key, value in super().parameters().items() + if key not in removed + } + ) + + @classmethod + def model_name(cls) -> str: + """Returns name of the model.""" + return "DeepWalk CBOW" \ No newline at end of file diff --git a/embiggen/embedders/ensmallen_embedders/deepwalk_glove.py b/embiggen/embedders/ensmallen_embedders/deepwalk_glove.py new file mode 100644 index 00000000..715bfbd1 --- /dev/null +++ b/embiggen/embedders/ensmallen_embedders/deepwalk_glove.py @@ -0,0 +1,112 @@ +"""Module providing DeepWalk GloVe model implementation.""" +from typing import Optional, Dict, Any +from embiggen.embedders.ensmallen_embedders.node2vec import Node2VecEnsmallen + +class DeepWalkGloVeEnsmallen(Node2VecEnsmallen): + """Class providing DeepWalk GloVe implemeted in Rust from Ensmallen.""" + + def __init__( + self, + embedding_size: int = 100, + alpha: float = 0.75, + epochs: int = 30, + clipping_value: float = 6.0, + walk_length: int = 128, + iterations: int = 10, + window_size: int = 5, + max_neighbours: Optional[int] = 100, + learning_rate: float = 0.001, + learning_rate_decay: float = 0.9, + central_nodes_embedding_path: Optional[str] = None, + contextual_nodes_embedding_path: Optional[str] = None, + normalize_by_degree: bool = False, + random_state: int = 42, + enable_cache: bool = False + ): + """Create new abstract DeepWalk method. + + Parameters + -------------------------- + embedding_size: int = 100 + Dimension of the embedding. + alpha: float = 0.75 + Alpha parameter for GloVe's loss. + epochs: int = 100 + Number of epochs to train the model for. + window_size: int = 10 + Window size for the local context. + On the borders the window size is trimmed. + clipping_value: float = 6.0 + Value at which we clip the dot product, mostly for numerical stability issues. + By default, `6.0`, where the loss is already close to zero. + walk_length: int = 128 + Maximal length of the walks. + iterations: int = 10 + Number of iterations of the single walks. + window_size: int = 5 + Window size for the local context. + On the borders the window size is trimmed. + max_neighbours: Optional[int] = 100 + Number of maximum neighbours to consider when using approximated walks. + By default, None, we execute exact random walks. + This is mainly useful for graphs containing nodes with high degrees. + learning_rate: float = 0.001 + The learning rate to use to train the DeepWalk model. By default 0.01. + central_nodes_embedding_path: Optional[str] = None + Path where to mmap and store the central nodes embedding. + This is necessary to embed large graphs whose embedding will not + fit into the available main memory. + contextual_nodes_embedding_path: Optional[str] = None + Path where to mmap and store the central nodes embedding. + This is necessary to embed large graphs whose embedding will not + fit into the available main memory. + learning_rate_decay: float = 0.9 + Factor to reduce the learning rate for at each epoch. By default 0.9. + normalize_by_degree: bool = False + Whether to normalize the random walk by the node degree + of the destination node degrees. + random_state: int = 42 + The random state to reproduce the training sequence. + enable_cache: bool = False + Whether to enable the cache, that is to + store the computed embedding. + """ + super().__init__( + embedding_size=embedding_size, + alpha=alpha, + epochs=epochs, + clipping_value=clipping_value, + walk_length=walk_length, + iterations=iterations, + window_size=window_size, + max_neighbours=max_neighbours, + learning_rate=learning_rate, + learning_rate_decay=learning_rate_decay, + central_nodes_embedding_path=central_nodes_embedding_path, + contextual_nodes_embedding_path=contextual_nodes_embedding_path, + normalize_by_degree=normalize_by_degree, + enable_cache=enable_cache, + random_state=random_state + ) + + def parameters(self) -> Dict[str, Any]: + """Returns parameters for smoke test.""" + removed = [ + "return_weight", + "explore_weight", + "change_node_type_weight", + "change_edge_type_weight", + "number_of_negative_samples" + ] + return dict( + **{ + key: value + for key, value in super().parameters().items() + if key not in removed + } + ) + + @classmethod + def model_name(cls) -> str: + """Returns name of the model.""" + return "DeepWalk GloVe" \ No newline at end of file diff --git a/embiggen/embedders/ensmallen_embedders/deepwalk_skipgram.py b/embiggen/embedders/ensmallen_embedders/deepwalk_skipgram.py new file mode 100644 index 00000000..6d160ea6 --- /dev/null +++ b/embiggen/embedders/ensmallen_embedders/deepwalk_skipgram.py @@ -0,0 +1,123 @@ +"""Module providing DeepWalk SkipGram model implementation.""" +from typing import Optional, Dict, Any +from embiggen.embedders.ensmallen_embedders.node2vec import Node2VecEnsmallen + + +class DeepWalkSkipGramEnsmallen(Node2VecEnsmallen): + """Class providing DeepWalk SkipGram implemeted in Rust from Ensmallen.""" + + def __init__( + self, + embedding_size: int = 100, + epochs: int = 30, + clipping_value: float = 6.0, + number_of_negative_samples: int = 10, + walk_length: int = 128, + iterations: int = 10, + window_size: int = 5, + max_neighbours: Optional[int] = 100, + learning_rate: float = 0.01, + learning_rate_decay: float = 0.9, + central_nodes_embedding_path: Optional[str] = None, + contextual_nodes_embedding_path: Optional[str] = None, + normalize_by_degree: bool = False, + stochastic_downsample_by_degree: Optional[bool] = False, + normalize_learning_rate_by_degree: Optional[bool] = False, + use_scale_free_distribution: Optional[bool] = True, + random_state: int = 42, + enable_cache: bool = False + ): + """Create new abstract DeepWalk method. + + Parameters + -------------------------- + embedding_size: int = 100 + Dimension of the embedding. + epochs: int = 30 + Number of epochs to train the model for. + clipping_value: float = 6.0 + Value at which we clip the dot product, mostly for numerical stability issues. + By default, `6.0`, where the loss is already close to zero. + number_of_negative_samples: int = 10 + The number of negative classes to randomly sample per batch. + This single sample of negative classes is evaluated for each element in the batch. + walk_length: int = 128 + Maximal length of the walks. + iterations: int = 10 + Number of iterations of the single walks. + window_size: int = 5 + Window size for the local context. + On the borders the window size is trimmed. + max_neighbours: Optional[int] = 100 + Number of maximum neighbours to consider when using approximated walks. + By default, None, we execute exact random walks. + This is mainly useful for graphs containing nodes with high degrees. + learning_rate: float = 0.01 + The learning rate to use to train the Node2Vec model. By default 0.01. + learning_rate_decay: float = 0.9 + Factor to reduce the learning rate for at each epoch. By default 0.9. + central_nodes_embedding_path: Optional[str] = None + Path where to mmap and store the central nodes embedding. + This is necessary to embed large graphs whose embedding will not + fit into the available main memory. + contextual_nodes_embedding_path: Optional[str] = None + Path where to mmap and store the central nodes embedding. + This is necessary to embed large graphs whose embedding will not + fit into the available main memory. + normalize_by_degree: bool = False + Whether to normalize the random walk by the node degree + of the destination node degrees. + stochastic_downsample_by_degree: Optional[bool] = False + Randomly skip samples with probability proportional to the degree of the central node. By default false. + normalize_learning_rate_by_degree: Optional[bool] = False + Divide the learning rate by the degree of the central node. By default false. + use_scale_free_distribution: Optional[bool] = True + Sample negatives proportionally to their degree. By default true. + random_state: int = 42 + The random state to reproduce the training sequence. + enable_cache: bool = False + Whether to enable the cache, that is to + store the computed embedding. + """ + super().__init__( + embedding_size=embedding_size, + epochs=epochs, + clipping_value=clipping_value, + number_of_negative_samples=number_of_negative_samples, + walk_length=walk_length, + iterations=iterations, + window_size=window_size, + max_neighbours=max_neighbours, + learning_rate=learning_rate, + learning_rate_decay=learning_rate_decay, + central_nodes_embedding_path=central_nodes_embedding_path, + contextual_nodes_embedding_path=contextual_nodes_embedding_path, + normalize_by_degree=normalize_by_degree, + stochastic_downsample_by_degree=stochastic_downsample_by_degree, + normalize_learning_rate_by_degree=normalize_learning_rate_by_degree, + use_scale_free_distribution=use_scale_free_distribution, + random_state=random_state, + enable_cache=enable_cache + ) + + def parameters(self) -> Dict[str, Any]: + """Returns parameters for smoke test.""" + removed = [ + "return_weight", + "explore_weight", + "change_node_type_weight", + "change_edge_type_weight", + "alpha" + ] + return dict( + **{ + key: value + for key, value in super().parameters().items() + if key not in removed + } + ) + + @classmethod + def model_name(cls) -> str: + """Returns name of the model.""" + return "DeepWalk SkipGram" \ No newline at end of file diff --git a/embiggen/embedders/ensmallen_embedders/spine.py b/embiggen/embedders/ensmallen_embedders/degree_spine.py similarity index 63% rename from embiggen/embedders/ensmallen_embedders/spine.py rename to embiggen/embedders/ensmallen_embedders/degree_spine.py index 4b0eca3d..e8bc7995 100644 --- a/embiggen/embedders/ensmallen_embedders/spine.py +++ b/embiggen/embedders/ensmallen_embedders/degree_spine.py @@ -1,35 +1,53 @@ -"""Module providing SPINE implementation.""" -from typing import Optional, Dict, Any, Union +"""Module providing Degree-based SPINE implementation.""" +from typing import Optional, Dict, Any from ensmallen import Graph -import numpy as np import pandas as pd from ensmallen import models -from embiggen.utils.abstract_models import AbstractEmbeddingModel, EmbeddingResult +from embiggen.embedders.ensmallen_embedders.ensmallen_embedder import EnsmallenEmbedder +from embiggen.utils import EmbeddingResult -class SPINE(AbstractEmbeddingModel): - """Class implementing the SPINE algorithm.""" +class DegreeSPINE(EnsmallenEmbedder): + """Class implementing the Degree-based SPINE algorithm.""" def __init__( self, embedding_size: int = 100, dtype: Optional[str] = "u8", + maximum_depth: Optional[int] = None, + path: Optional[str] = None, + verbose: bool = False, enable_cache: bool = False ): - """Create new SPINE method. + """Create new Degree-based SPINE method. Parameters -------------------------- embedding_size: int = 100 Dimension of the embedding. dtype: Optional[str] = "u8" - Dtype to use for the embedding. Note that an improper dtype may cause overflows. + Dtype to use for the embedding. + maximum_depth: Optional[int] = None + Maximum depth of the shortest path. + path: Optional[str] = None + Path where to store the mmap-ed embedding. + This parameter is necessary to embed very large graphs. + verbose: bool = False + Whether to show loading bars. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. """ self._dtype = dtype - self._model = models.SPINE(embedding_size=embedding_size) + self._verbose = verbose + self._maximum_depth = maximum_depth + self._path = path + self._model = models.DegreeSPINE( + embedding_size=embedding_size, + verbose=self._verbose, + maximum_depth=self._maximum_depth, + path=self._path + ) super().__init__( embedding_size=embedding_size, @@ -41,7 +59,9 @@ def parameters(self) -> Dict[str, Any]: return { **super().parameters(), **dict( - dtype=self._dtype + dtype=self._dtype, + maximum_depth=self._maximum_depth, + path=self._path, ) } @classmethod @@ -55,14 +75,12 @@ def _fit_transform( self, graph: Graph, return_dataframe: bool = True, - verbose: bool = True ) -> EmbeddingResult: """Return node embedding.""" node_embedding = self._model.fit_transform( graph, dtype=self._dtype, - verbose=verbose, - ).T + ) if return_dataframe: node_embedding = pd.DataFrame( node_embedding, @@ -73,26 +91,10 @@ def _fit_transform( node_embeddings=node_embedding ) - @classmethod - def task_name(cls) -> str: - return "Node Embedding" - @classmethod def model_name(cls) -> str: """Returns name of the model.""" - return "SPINE" - - @classmethod - def library_name(cls) -> str: - return "Ensmallen" - - @classmethod - def requires_nodes_sorted_by_decreasing_node_degree(cls) -> bool: - return False - - @classmethod - def is_topological(cls) -> bool: - return True + return "Degree-based SPINE" @classmethod def can_use_node_types(cls) -> bool: diff --git a/embiggen/embedders/ensmallen_embedders/degree_wine.py b/embiggen/embedders/ensmallen_embedders/degree_wine.py new file mode 100644 index 00000000..5b848a16 --- /dev/null +++ b/embiggen/embedders/ensmallen_embedders/degree_wine.py @@ -0,0 +1,118 @@ +"""Module providing Degree-based WINE implementation.""" +from typing import Optional, Dict, Any +from ensmallen import Graph +import pandas as pd +from ensmallen import models +from embiggen.embedders.ensmallen_embedders.ensmallen_embedder import EnsmallenEmbedder +from embiggen.utils import EmbeddingResult + + +class DegreeWINE(EnsmallenEmbedder): + """Class implementing the Degree-based WINE algorithm.""" + + def __init__( + self, + embedding_size: int = 100, + dtype: Optional[str] = "u8", + walk_length: int = 2, + path: Optional[str] = None, + verbose: bool = False, + enable_cache: bool = False + ): + """Create new Degree-based WINE method. + + Parameters + -------------------------- + embedding_size: int = 100 + Dimension of the embedding. + dtype: Optional[str] = "u8" + Dtype to use for the embedding. + walk_length: int = 2 + Length of the random walk. + By default 2, to capture exclusively the immediate context. + path: Optional[str] = None + Path where to store the mmap-ed embedding. + This parameter is necessary to embed very large graphs. + verbose: bool = False + Whether to show loading bars. + enable_cache: bool = False + Whether to enable the cache, that is to + store the computed embedding. + """ + self._dtype = dtype + self._verbose = verbose + self._walk_length = walk_length + self._path = path + self._model = models.DegreeWINE( + embedding_size=embedding_size, + verbose=self._verbose, + walk_length=self._walk_length, + path=self._path + ) + + super().__init__( + embedding_size=embedding_size, + enable_cache=enable_cache + ) + + def parameters(self) -> Dict[str, Any]: + """Returns parameters of the model.""" + return { + **super().parameters(), + **dict( + dtype=self._dtype, + walk_length=self._walk_length, + path=self._path, + ) + } + @classmethod + def smoke_test_parameters(cls) -> Dict[str, Any]: + """Returns parameters for smoke test.""" + return dict( + embedding_size=5, + ) + + def _fit_transform( + self, + graph: Graph, + return_dataframe: bool = True, + ) -> EmbeddingResult: + """Return node embedding.""" + node_embedding = self._model.fit_transform( + graph, + dtype=self._dtype, + ) + if return_dataframe: + node_embedding = pd.DataFrame( + node_embedding, + index=graph.get_node_names() + ) + return EmbeddingResult( + embedding_method_name=self.model_name(), + node_embeddings=node_embedding + ) + + @classmethod + def model_name(cls) -> str: + """Returns name of the model.""" + return "Degree-based WINE" + + @classmethod + def can_use_node_types(cls) -> bool: + """Returns whether the model can optionally use node types.""" + return False + + @classmethod + def can_use_edge_weights(cls) -> bool: + """Returns whether the model can optionally use edge weights.""" + return False + + @classmethod + def can_use_edge_types(cls) -> bool: + """Returns whether the model can optionally use edge types.""" + return False + + @classmethod + def is_stocastic(cls) -> bool: + """Returns whether the model is stocastic and has therefore a random state.""" + return False \ No newline at end of file diff --git a/embiggen/embedders/ensmallen_embedders/ensmallen_embedder.py b/embiggen/embedders/ensmallen_embedders/ensmallen_embedder.py new file mode 100644 index 00000000..7232edeb --- /dev/null +++ b/embiggen/embedders/ensmallen_embedders/ensmallen_embedder.py @@ -0,0 +1,51 @@ +"""Module providing SocioDim implementation.""" +from typing import Optional, Dict, Any +from ensmallen import Graph +import pandas as pd +import numpy as np +from embiggen.utils.abstract_models import AbstractEmbeddingModel, abstract_class + +@abstract_class +class EnsmallenEmbedder(AbstractEmbeddingModel): + """Class implementing the Ensmallen Embedder algorithm.""" + + def __init__( + self, + random_state: Optional[int] = None, + embedding_size: int = 100, + enable_cache: bool = False + ): + + """Create new EnsmallenEmbedder method. + + Parameters + -------------------------- + random_state: Optional[int] = None + Random state to reproduce the embeddings. + embedding_size: int = 100 + Dimension of the embedding. + enable_cache: bool = False + Whether to enable the cache, that is to + store the computed embedding. + """ + super().__init__( + random_state=random_state, + embedding_size=embedding_size, + enable_cache=enable_cache, + ) + + @classmethod + def task_name(cls) -> str: + return "Node Embedding" + + @classmethod + def library_name(cls) -> str: + return "Ensmallen" + + @classmethod + def requires_nodes_sorted_by_decreasing_node_degree(cls) -> bool: + return False + + @classmethod + def is_topological(cls) -> bool: + return True \ No newline at end of file diff --git a/embiggen/embedders/ensmallen_embedders/first_order_line.py b/embiggen/embedders/ensmallen_embedders/first_order_line.py new file mode 100644 index 00000000..e5bd91b8 --- /dev/null +++ b/embiggen/embedders/ensmallen_embedders/first_order_line.py @@ -0,0 +1,131 @@ +"""Module providing first-order LINE implementation.""" +from typing import Dict, Any, Optional +from ensmallen import Graph +import pandas as pd +from ensmallen import models +from embiggen.embedders.ensmallen_embedders.ensmallen_embedder import EnsmallenEmbedder +from embiggen.utils import EmbeddingResult + + +class FirstOrderLINEEnsmallen(EnsmallenEmbedder): + """Class implementing the first-order LINE algorithm.""" + + def __init__( + self, + embedding_size: int = 100, + epochs: int = 100, + learning_rate: float = 0.05, + learning_rate_decay: float = 0.9, + avoid_false_negatives: bool = False, + node_embedding_path: Optional[str] = None, + random_state: int = 42, + verbose: bool = False, + enable_cache: bool = False + ): + """Create new abstract Node2Vec method. + + Parameters + -------------------------- + embedding_size: int = 100 + Dimension of the embedding. + epochs: int = 100 + The number of epochs to run the model for, by default 10. + learning_rate: float = 0.05 + The learning rate to update the gradient, by default 0.01. + learning_rate_decay: float = 0.9 + Factor to reduce the learning rate for at each epoch. By default 0.9. + avoid_false_negatives: bool = False + Whether to avoid sampling false negatives. + This may cause a slower training. + node_embedding_path: Optional[str] = None + Path where to mmap and store the nodes embedding. + This is necessary to embed large graphs whose embedding will not + fit into the available main memory. + random_state: int = 42 + Random state to reproduce the embeddings. + verbose: bool = False + Whether to show loading bars. + enable_cache: bool = False + Whether to enable the cache, that is to + store the computed embedding. + """ + self._kwargs = dict( + epochs=epochs, + learning_rate=learning_rate, + learning_rate_decay=learning_rate_decay, + avoid_false_negatives=avoid_false_negatives, + node_embedding_path=node_embedding_path, + verbose=verbose + ) + + self._model = models.FirstOrderLINE( + **self._kwargs, + embedding_size=embedding_size, + random_state=random_state + ) + + super().__init__( + embedding_size=embedding_size, + enable_cache=enable_cache, + random_state=random_state + ) + + def parameters(self) -> Dict[str, Any]: + """Returns parameters of the model.""" + return dict( + **super().parameters(), + **self._kwargs, + ) + + @classmethod + def smoke_test_parameters(cls) -> Dict[str, Any]: + """Returns parameters for smoke test.""" + return dict( + embedding_size=5, + epochs=1 + ) + + def _fit_transform( + self, + graph: Graph, + return_dataframe: bool = True, + ) -> EmbeddingResult: + """Return node embedding.""" + node_embedding = self._model.fit_transform( + graph + ) + if return_dataframe: + node_embedding = pd.DataFrame( + node_embedding, + index=graph.get_node_names() + ) + + return EmbeddingResult( + embedding_method_name=self.model_name(), + node_embeddings= node_embedding, + ) + + @classmethod + def model_name(cls) -> str: + """Returns name of the model.""" + return "First-order LINE" + + @classmethod + def can_use_edge_weights(cls) -> bool: + """Returns whether the model can optionally use edge weights.""" + return False + + @classmethod + def can_use_node_types(cls) -> bool: + """Returns whether the model can optionally use node types.""" + return False + + @classmethod + def can_use_edge_types(cls) -> bool: + """Returns whether the model can optionally use edge types.""" + return False + + @classmethod + def is_stocastic(cls) -> bool: + """Returns whether the model is stocastic and has therefore a random state.""" + return True \ No newline at end of file diff --git a/embiggen/embedders/ensmallen_embedders/geometric_laplacian_eigenmaps.py b/embiggen/embedders/ensmallen_embedders/geometric_laplacian_eigenmaps.py index 15d487b4..7862ec90 100644 --- a/embiggen/embedders/ensmallen_embedders/geometric_laplacian_eigenmaps.py +++ b/embiggen/embedders/ensmallen_embedders/geometric_laplacian_eigenmaps.py @@ -5,10 +5,11 @@ import numpy as np from scipy.sparse import coo_matrix from scipy.sparse.linalg import eigsh -from embiggen.utils.abstract_models import AbstractEmbeddingModel, EmbeddingResult +from embiggen.embedders.ensmallen_embedders.ensmallen_embedder import EnsmallenEmbedder +from embiggen.utils import EmbeddingResult -class GLEEEnsmallen(AbstractEmbeddingModel): +class GLEEEnsmallen(EnsmallenEmbedder): """Class implementing the GLEE algorithm.""" def __init__( @@ -35,7 +36,6 @@ def _fit_transform( self, graph: Graph, return_dataframe: bool = True, - verbose: bool = True ) -> EmbeddingResult: """Return node embedding.""" edges, weights = graph.get_symmetric_normalized_laplacian_coo_matrix() @@ -67,27 +67,11 @@ def _fit_transform( node_embeddings=embedding ) - @classmethod - def task_name(cls) -> str: - return "Node Embedding" - @classmethod def model_name(cls) -> str: """Returns name of the model.""" return "GLEE" - @classmethod - def library_name(cls) -> str: - return "Ensmallen" - - @classmethod - def requires_nodes_sorted_by_decreasing_node_degree(cls) -> bool: - return False - - @classmethod - def is_topological(cls) -> bool: - return True - @classmethod def can_use_edge_weights(cls) -> bool: """Returns whether the model can optionally use edge weights.""" diff --git a/embiggen/embedders/ensmallen_embedders/glove.py b/embiggen/embedders/ensmallen_embedders/glove.py deleted file mode 100644 index 5f109884..00000000 --- a/embiggen/embedders/ensmallen_embedders/glove.py +++ /dev/null @@ -1,239 +0,0 @@ -"""Module providing GloVe model implementation.""" -from typing import Optional, Dict, Any, Union -from ensmallen import Graph -from ensmallen import models -import numpy as np -import pandas as pd -from embiggen.utils.abstract_models import AbstractEmbeddingModel, EmbeddingResult - - -class GloVeEnsmallen(AbstractEmbeddingModel): - """Class providing GloVe implemeted in Rust from Ensmallen.""" - - def __init__( - self, - embedding_size: int = 100, - alpha: float = 0.75, - epochs: int = 500, - clipping_value: float = 6.0, - walk_length: int = 128, - iterations: int = 10, - window_size: int = 10, - return_weight: float = 1.0, - explore_weight: float = 1.0, - change_node_type_weight: float = 1.0, - change_edge_type_weight: float = 1.0, - max_neighbours: Optional[int] = 100, - learning_rate: float = 0.01, - learning_rate_decay: float = 0.9, - normalize_by_degree: bool = False, - random_state: int = 42, - enable_cache: bool = False - ): - """Create new abstract Node2Vec method. - - Parameters - -------------------------- - embedding_size: int = 100 - Dimension of the embedding. - alpha: float = 0.75 - Alpha parameter for GloVe's loss. - epochs: int = 100 - Number of epochs to train the model for. - window_size: int = 10 - Window size for the local context. - On the borders the window size is trimmed. - clipping_value: float = 6.0 - Value at which we clip the dot product, mostly for numerical stability issues. - By default, `6.0`, where the loss is already close to zero. - walk_length: int = 128 - Maximal length of the walks. - iterations: int = 10 - Number of iterations of the single walks. - window_size: int = 10 - Window size for the local context. - On the borders the window size is trimmed. - return_weight: float = 1.0 - Weight on the probability of returning to the same node the walk just came from - Having this higher tends the walks to be - more like a Breadth-First Search. - Having this very high (> 2) makes search very local. - Equal to the inverse of p in the Node2Vec paper. - explore_weight: float = 1.0 - Weight on the probability of visiting a neighbor node - to the one we're coming from in the random walk - Having this higher tends the walks to be - more like a Depth-First Search. - Having this very high makes search more outward. - Having this very low makes search very local. - Equal to the inverse of q in the Node2Vec paper. - change_node_type_weight: float = 1.0 - Weight on the probability of visiting a neighbor node of a - different type than the previous node. This only applies to - colored graphs, otherwise it has no impact. - change_edge_type_weight: float = 1.0 - Weight on the probability of visiting a neighbor edge of a - different type than the previous edge. This only applies to - multigraphs, otherwise it has no impact. - max_neighbours: Optional[int] = 100 - Number of maximum neighbours to consider when using approximated walks. - By default, None, we execute exact random walks. - This is mainly useful for graphs containing nodes with high degrees. - learning_rate: float = 0.01 - The learning rate to use to train the Node2Vec model. By default 0.01. - learning_rate_decay: float = 0.9 - Factor to reduce the learning rate for at each epoch. By default 0.9. - normalize_by_degree: bool = False - Whether to normalize the random walk by the node degree - of the destination node degrees. - random_state: int = 42 - The random state to reproduce the training sequence. - enable_cache: bool = False - Whether to enable the cache, that is to - store the computed embedding. - """ - - self._epochs = epochs - self._learning_rate = learning_rate - self._learning_rate_decay = learning_rate_decay - - self._model_kwargs = dict( - alpha=alpha, - embedding_size=embedding_size, - clipping_value=clipping_value, - walk_length=walk_length, - iterations=iterations, - window_size=window_size, - return_weight=return_weight, - explore_weight=explore_weight, - change_edge_type_weight=change_edge_type_weight, - change_node_type_weight=change_node_type_weight, - max_neighbours=max_neighbours, - normalize_by_degree=normalize_by_degree, - ) - - self._model = models.GloVe( - **self._model_kwargs, - random_state=random_state - ) - - super().__init__( - embedding_size=embedding_size, - enable_cache=enable_cache, - random_state=random_state - ) - - @classmethod - def smoke_test_parameters(cls) -> Dict[str, Any]: - """Returns parameters for smoke test.""" - return dict( - embedding_size=5, - epochs=1, - window_size=1, - walk_length=4, - iterations=1, - max_neighbours=10, - ) - - def parameters(self) -> Dict[str, Any]: - """Returns parameters of the model.""" - return { - **super().parameters(), - **self._model_kwargs, - **dict( - epochs=self._epochs, - learning_rate=self._learning_rate, - learning_rate_decay=self._learning_rate_decay, - ) - } - - @classmethod - def task_name(cls) -> str: - return "Node Embedding" - - @classmethod - def library_name(cls) -> str: - return "Ensmallen" - - def requires_nodes_sorted_by_decreasing_node_degree(self) -> bool: - return False - - def is_topological(self) -> bool: - return True - - def _fit_transform( - self, - graph: Graph, - return_dataframe: bool = True, - verbose: bool = True - ) -> EmbeddingResult: - """Return node embedding.""" - node_embedding = self._model.fit_transform( - graph, - epochs=self._epochs, - learning_rate=self._learning_rate, - learning_rate_decay=self._learning_rate_decay, - verbose=verbose - ) - if return_dataframe: - node_embedding = pd.DataFrame( - node_embedding, - index=graph.get_node_names() - ) - return EmbeddingResult( - embedding_method_name=self.model_name(), - node_embeddings=node_embedding - ) - - @classmethod - def requires_edge_weights(cls) -> bool: - return False - - @classmethod - def requires_positive_edge_weights(cls) -> bool: - return True - - @classmethod - def model_name(cls) -> str: - """Returns name of the model.""" - return "GloVe" - - @classmethod - def requires_node_types(cls) -> bool: - return False - - @classmethod - def requires_edge_types(cls) -> bool: - return False - - @classmethod - def can_use_edge_weights(cls) -> bool: - """Returns whether the model can optionally use edge weights.""" - return True - - def is_using_edge_weights(self) -> bool: - """Returns whether the model is parametrized to use edge weights.""" - return True - - @classmethod - def can_use_node_types(cls) -> bool: - """Returns whether the model can optionally use node types.""" - return True - - def is_using_node_types(self) -> bool: - """Returns whether the model is parametrized to use node types.""" - return self._model_kwargs["change_node_type_weight"] != 1.0 - - @classmethod - def can_use_edge_types(cls) -> bool: - """Returns whether the model can optionally use edge types.""" - return True - - def is_using_edge_types(self) -> bool: - """Returns whether the model is parametrized to use edge types.""" - return self._model_kwargs["change_edge_type_weight"] != 1.0 - - @classmethod - def is_stocastic(cls) -> bool: - """Returns whether the model is stocastic and has therefore a random state.""" - return True \ No newline at end of file diff --git a/embiggen/embedders/ensmallen_embedders/hope.py b/embiggen/embedders/ensmallen_embedders/hope.py index c6ec985f..ed85a408 100644 --- a/embiggen/embedders/ensmallen_embedders/hope.py +++ b/embiggen/embedders/ensmallen_embedders/hope.py @@ -6,17 +6,20 @@ from scipy.sparse import coo_matrix from scipy.sparse.linalg import svds as sparse_svds from sklearn.utils.extmath import randomized_svd -from embiggen.utils.abstract_models import AbstractEmbeddingModel, EmbeddingResult, format_list +from userinput.utils import must_be_in_set +from embiggen.embedders.ensmallen_embedders.ensmallen_embedder import EnsmallenEmbedder +from embiggen.utils import EmbeddingResult -class HOPEEnsmallen(AbstractEmbeddingModel): +class HOPEEnsmallen(EnsmallenEmbedder): """Class implementing the HOPE algorithm.""" def __init__( self, embedding_size: int = 100, - metric: str = "Jaccard", + metric: str = "Neighbours Intersection size", root_node_name: Optional[str] = None, + verbose: bool = False, enable_cache: bool = False ): """Create new HOPE method. @@ -25,7 +28,7 @@ def __init__( -------------------------- embedding_size: int = 100 Dimension of the embedding. - metric: str = "Jaccard" + metric: str = "Neighbours Intersection size" The metric to use. You can either use: - Jaccard @@ -38,18 +41,17 @@ def __init__( - Left Normalized Laplacian - Right Normalized Laplacian - Symmetric Normalized Laplacian + - Resnik root_node_name: Optional[str] = None Root node to use when the ancestors mode for the Jaccard index is selected. + verbose: bool = False + Whether to show loading bars. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. """ - if metric not in self.get_available_metrics(): - raise ValueError( - f"The provided metric {metric} is not a supported " - f"metric. The supported metrics are: {format_list(self.get_available_metrics())}." - ) + metric = must_be_in_set(metric, self.get_available_metrics(), "metric") ancestral_metric = ("Ancestors Jaccard", "Ancestors size") if root_node_name is None and metric in ancestral_metric: raise ValueError( @@ -65,6 +67,7 @@ def __init__( self._metric = metric self._root_node_name = root_node_name + self._verbose = verbose super().__init__( embedding_size=embedding_size, @@ -73,19 +76,21 @@ def __init__( def parameters(self) -> Dict[str, Any]: """Returns parameters of the model.""" - return { + return dict( **super().parameters(), **dict( metric=self._metric, + verbose=self._verbose, root_node_name=self._root_node_name ) - } + ) @classmethod def get_available_metrics(cls) -> List[str]: """Returns list of the available metrics.""" return [ "Jaccard", + "Shortest Paths", "Neighbours Intersection size", "Ancestors Jaccard", "Ancestors size", @@ -102,7 +107,6 @@ def _fit_transform( self, graph: Graph, return_dataframe: bool = True, - verbose: bool = True ) -> EmbeddingResult: """Return node embedding.""" matrix = None @@ -110,6 +114,8 @@ def _fit_transform( edges, weights = graph.get_jaccard_coo_matrix() elif self._metric == "Laplacian": edges, weights = graph.get_laplacian_coo_matrix() + elif self._metric == "Shortest Paths": + matrix = graph.get_shortest_paths_matrix() elif self._metric == "Modularity": matrix = graph.get_dense_modularity_matrix() elif self._metric == "Left Normalized Laplacian": @@ -126,7 +132,7 @@ def _fit_transform( src_node_name=self._root_node_name, compute_predecessors=True ), - verbose=verbose + verbose=self._verbose ) elif self._metric == "Ancestors size": matrix = graph.get_shared_ancestors_size_adjacency_matrix( @@ -134,18 +140,13 @@ def _fit_transform( src_node_name=self._root_node_name, compute_predecessors=True ), - verbose=verbose + verbose=self._verbose ) elif self._metric == "Adamic-Adar": edges, weights = graph.get_adamic_adar_coo_matrix() elif self._metric == "Adjacency": edges, weights = graph.get_directed_edge_node_ids(), np.ones( graph.get_number_of_directed_edges()) - else: - raise NotImplementedError( - f"The provided metric {self._metric} " - "is not currently supported." - ) if matrix is None: matrix = coo_matrix( @@ -186,27 +187,11 @@ def _fit_transform( node_embeddings=[left_embedding, right_embedding] ) - @classmethod - def task_name(cls) -> str: - return "Node Embedding" - @classmethod def model_name(cls) -> str: """Returns name of the model.""" return "HOPE" - @classmethod - def library_name(cls) -> str: - return "Ensmallen" - - @classmethod - def requires_nodes_sorted_by_decreasing_node_degree(cls) -> bool: - return False - - @classmethod - def is_topological(cls) -> bool: - return True - @classmethod def can_use_edge_weights(cls) -> bool: """Returns whether the model can optionally use edge weights.""" diff --git a/embiggen/embedders/ensmallen_embedders/laplacian_eigenmaps.py b/embiggen/embedders/ensmallen_embedders/laplacian_eigenmaps.py new file mode 100644 index 00000000..899170b3 --- /dev/null +++ b/embiggen/embedders/ensmallen_embedders/laplacian_eigenmaps.py @@ -0,0 +1,93 @@ +"""Module providing Laplacian Eigenmaps implementation.""" +from ensmallen import Graph +import pandas as pd +import numpy as np +from scipy.sparse import coo_matrix +from scipy.sparse.linalg import eigsh +from embiggen.embedders.ensmallen_embedders.ensmallen_embedder import EnsmallenEmbedder +from embiggen.utils import EmbeddingResult + + +class LaplacianEigenmapsEnsmallen(EnsmallenEmbedder): + """Class implementing the Laplacian Eigenmaps algorithm.""" + + def __init__( + self, + embedding_size: int = 100, + enable_cache: bool = False + ): + """Create new Laplacian Eigenmaps method. + + Parameters + -------------------------- + embedding_size: int = 100 + Dimension of the embedding. + enable_cache: bool = False + Whether to enable the cache, that is to + store the computed embedding. + """ + super().__init__( + embedding_size=embedding_size, + enable_cache=enable_cache, + ) + + def _fit_transform( + self, + graph: Graph, + return_dataframe: bool = True, + ) -> EmbeddingResult: + """Return node embedding.""" + edges, weights = graph.get_symmetric_normalized_laplacian_coo_matrix() + + coo = coo_matrix( + (weights, (edges[:, 0], edges[:, 1])), + shape=( + graph.get_number_of_nodes(), + graph.get_number_of_nodes() + ), + dtype=np.float32 + ) + + embedding = eigsh( + coo, + k=self._embedding_size, + which="SM", + maxiter=graph.get_number_of_nodes()*100, + return_eigenvectors=True + )[1] + + if return_dataframe: + node_names = graph.get_node_names() + embedding = pd.DataFrame( + embedding, + index=node_names + ) + return EmbeddingResult( + embedding_method_name=self.model_name(), + node_embeddings=embedding + ) + + @classmethod + def model_name(cls) -> str: + """Returns name of the model.""" + return "Laplacian Eigenmaps" + + @classmethod + def can_use_edge_weights(cls) -> bool: + """Returns whether the model can optionally use edge weights.""" + return False + + @classmethod + def can_use_node_types(cls) -> bool: + """Returns whether the model can optionally use node types.""" + return False + + @classmethod + def can_use_edge_types(cls) -> bool: + """Returns whether the model can optionally use edge types.""" + return False + + @classmethod + def is_stocastic(cls) -> bool: + """Returns whether the model is stocastic and has therefore a random state.""" + return False diff --git a/embiggen/embedders/ensmallen_embedders/netmf.py b/embiggen/embedders/ensmallen_embedders/netmf.py index c33665ff..af658a33 100644 --- a/embiggen/embedders/ensmallen_embedders/netmf.py +++ b/embiggen/embedders/ensmallen_embedders/netmf.py @@ -5,10 +5,11 @@ import numpy as np from scipy.sparse import coo_matrix from sklearn.decomposition import TruncatedSVD -from embiggen.utils.abstract_models import AbstractEmbeddingModel, EmbeddingResult +from embiggen.embedders.ensmallen_embedders.ensmallen_embedder import EnsmallenEmbedder +from embiggen.utils import EmbeddingResult -class NetMFEnsmallen(AbstractEmbeddingModel): +class NetMFEnsmallen(EnsmallenEmbedder): """Class implementing the NetMF algorithm.""" def __init__( @@ -17,10 +18,6 @@ def __init__( walk_length: int = 128, iterations: int = 10, window_size: int = 10, - return_weight: float = 1.0, - explore_weight: float = 1.0, - change_node_type_weight: float = 1.0, - change_edge_type_weight: float = 1.0, max_neighbours: Optional[int] = 100, random_state: int = 42, enable_cache: bool = False @@ -38,28 +35,6 @@ def __init__( window_size: int = 10 Window size for the local context. On the borders the window size is trimmed. - return_weight: float = 1.0 - Weight on the probability of returning to the same node the walk just came from - Having this higher tends the walks to be - more like a Breadth-First Search. - Having this very high (> 2) makes search very local. - Equal to the inverse of p in the Node2Vec paper. - explore_weight: float = 1.0 - Weight on the probability of visiting a neighbor node - to the one we're coming from in the random walk - Having this higher tends the walks to be - more like a Depth-First Search. - Having this very high makes search more outward. - Having this very low makes search very local. - Equal to the inverse of q in the Node2Vec paper. - change_node_type_weight: float = 1.0 - Weight on the probability of visiting a neighbor node of a - different type than the previous node. This only applies to - colored graphs, otherwise it has no impact. - change_edge_type_weight: float = 1.0 - Weight on the probability of visiting a neighbor edge of a - different type than the previous edge. This only applies to - multigraphs, otherwise it has no impact. max_neighbours: Optional[int] = 100 Number of maximum neighbours to consider when using approximated walks. By default, None, we execute exact random walks. @@ -74,10 +49,6 @@ def __init__( walk_length=walk_length, iterations=iterations, window_size=window_size, - return_weight=return_weight, - explore_weight=explore_weight, - change_node_type_weight=change_node_type_weight, - change_edge_type_weight=change_edge_type_weight, max_neighbours=max_neighbours, ) @@ -91,7 +62,7 @@ def __init__( def smoke_test_parameters(cls) -> Dict[str, Any]: """Returns parameters for smoke test.""" return dict( - **AbstractEmbeddingModel.smoke_test_parameters(), + **EnsmallenEmbedder.smoke_test_parameters(), window_size=1, walk_length=4, iterations=1, @@ -109,7 +80,6 @@ def _fit_transform( self, graph: Graph, return_dataframe: bool = True, - verbose: bool = True ) -> EmbeddingResult: """Return node embedding.""" edges, weights = graph.get_log_normalized_cooccurrence_coo_matrix( @@ -143,27 +113,11 @@ def _fit_transform( node_embeddings=embedding ) - @classmethod - def task_name(cls) -> str: - return "Node Embedding" - @classmethod def model_name(cls) -> str: """Returns name of the model.""" return "NetMF" - @classmethod - def library_name(cls) -> str: - return "Ensmallen" - - @classmethod - def requires_nodes_sorted_by_decreasing_node_degree(cls) -> bool: - return False - - @classmethod - def is_topological(cls) -> bool: - return True - @classmethod def can_use_edge_weights(cls) -> bool: """Returns whether the model can optionally use edge weights.""" diff --git a/embiggen/embedders/ensmallen_embedders/node2vec.py b/embiggen/embedders/ensmallen_embedders/node2vec.py index 36ce0781..31b64caa 100644 --- a/embiggen/embedders/ensmallen_embedders/node2vec.py +++ b/embiggen/embedders/ensmallen_embedders/node2vec.py @@ -1,26 +1,34 @@ """Module providing abstract Node2Vec implementation.""" -from typing import Optional, Union, Dict, Any +from typing import Dict, Any from ensmallen import Graph -import numpy as np import pandas as pd +from userinput.utils import must_be_in_set from ensmallen import models -from embiggen.utils.abstract_models import abstract_class, AbstractEmbeddingModel, EmbeddingResult +from embiggen.utils.abstract_models import abstract_class +from embiggen.embedders.ensmallen_embedders.ensmallen_embedder import EnsmallenEmbedder +from embiggen.utils import EmbeddingResult @abstract_class -class Node2VecEnsmallen(AbstractEmbeddingModel): +class Node2VecEnsmallen(EnsmallenEmbedder): """Abstract class for Node2Vec algorithms.""" MODELS = { - "cbow": models.CBOW, - "skipgram": models.SkipGram, - "walkletscbow": models.WalkletsCBOW, - "walkletsskipgram": models.WalkletsSkipGram, + "DeepWalk CBOW": models.CBOW, + "DeepWalk SkipGram": models.SkipGram, + "DeepWalk GloVe": models.GloVe, + "Node2Vec CBOW": models.CBOW, + "Node2Vec SkipGram": models.SkipGram, + "Node2Vec GloVe": models.GloVe, + "Walklets CBOW": models.WalkletsCBOW, + "Walklets SkipGram": models.WalkletsSkipGram, + "Walklets GloVe": models.WalkletsGloVe, } def __init__( self, - model_name: str, + embedding_size: int = 100, + random_state: int = 42, enable_cache: bool = False, **model_kwargs: Dict ): @@ -28,77 +36,56 @@ def __init__( Parameters -------------------------- - model_name: str - The ensmallen graph embedding model to use. - Can either be the SkipGram, WalkletsSkipGram, CBOW and WalkletsCBOW models. + embedding_size: int = 100 + Dimension of the embedding. + random_state: int = 42 + The random state to reproduce the training sequence. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. model_kwargs: Dict Further parameters to forward to the model. """ - if model_name.lower() not in self.MODELS: - raise ValueError( - ( - "The provided model name {} is not supported. " - "The supported models are `CBOW` and `SkipGram`." - ).format(model_name) - ) - + model_name = must_be_in_set(self.model_name(), self.MODELS.keys(), "model name") self._model_kwargs = model_kwargs - - self._model = Node2VecEnsmallen.MODELS[model_name.lower()](**model_kwargs) + self._model = Node2VecEnsmallen.MODELS[model_name](**model_kwargs) super().__init__( - embedding_size=model_kwargs["embedding_size"], + embedding_size=embedding_size, enable_cache=enable_cache, - random_state=model_kwargs["random_state"] + random_state=random_state ) @classmethod def smoke_test_parameters(cls) -> Dict[str, Any]: """Returns parameters for smoke test.""" return dict( - **AbstractEmbeddingModel.smoke_test_parameters(), + **EnsmallenEmbedder.smoke_test_parameters(), epochs=1, window_size=1, walk_length=4, iterations=1, max_neighbours=10, - number_of_negative_samples=1 ) def parameters(self) -> Dict[str, Any]: """Returns parameters of the model.""" - return { + return dict( **super().parameters(), **self._model_kwargs - } - - @classmethod - def task_name(cls) -> str: - return "Node Embedding" - - @classmethod - def library_name(cls) -> str: - return "Ensmallen" - - def requires_nodes_sorted_by_decreasing_node_degree(self) -> bool: - return False - - def is_topological(self) -> bool: - return True + ) def _fit_transform( self, graph: Graph, return_dataframe: bool = True, - verbose: bool = True ) -> EmbeddingResult: """Return node embedding.""" node_embeddings = self._model.fit_transform(graph) - if not isinstance(node_embeddings, list): - node_embeddings = [node_embeddings] + + if "CBOW" in self.model_name(): + node_embeddings = list(reversed(node_embeddings)) + if return_dataframe: node_names = graph.get_node_names() node_embeddings = [ @@ -137,7 +124,7 @@ def can_use_node_types(cls) -> bool: def is_using_node_types(self) -> bool: """Returns whether the model is parametrized to use node types.""" - return self._model_kwargs["change_node_type_weight"] != 1.0 + return "change_node_type_weight" in self._model_kwargs and self._model_kwargs["change_node_type_weight"] != 1.0 @classmethod def can_use_edge_types(cls) -> bool: @@ -146,9 +133,17 @@ def can_use_edge_types(cls) -> bool: def is_using_edge_types(self) -> bool: """Returns whether the model is parametrized to use edge types.""" - return self._model_kwargs["change_edge_type_weight"] != 1.0 + return "change_edge_type_weight" in self._model_kwargs and self._model_kwargs["change_edge_type_weight"] != 1.0 @classmethod def is_stocastic(cls) -> bool: """Returns whether the model is stocastic and has therefore a random state.""" - return True \ No newline at end of file + return True + + @classmethod + def requires_node_types(cls) -> bool: + return False + + @classmethod + def requires_edge_types(cls) -> bool: + return False \ No newline at end of file diff --git a/embiggen/embedders/ensmallen_embedders/cbow.py b/embiggen/embedders/ensmallen_embedders/node2vec_cbow.py similarity index 70% rename from embiggen/embedders/ensmallen_embedders/cbow.py rename to embiggen/embedders/ensmallen_embedders/node2vec_cbow.py index 8e0e58a8..4fc3203b 100644 --- a/embiggen/embedders/ensmallen_embedders/cbow.py +++ b/embiggen/embedders/ensmallen_embedders/node2vec_cbow.py @@ -1,31 +1,31 @@ -"""Module providing CBOW model implementation.""" -from typing import Optional +"""Module providing Node2Vec CBOW model implementation.""" +from typing import Optional, Dict, Any from embiggen.embedders.ensmallen_embedders.node2vec import Node2VecEnsmallen -class CBOWEnsmallen(Node2VecEnsmallen): - """Class providing CBOW implemeted in Rust from Ensmallen.""" +class Node2VecCBOWEnsmallen(Node2VecEnsmallen): + """Class providing Node2Vec CBOW implemeted in Rust from Ensmallen.""" def __init__( self, embedding_size: int = 100, - epochs: int = 10, + epochs: int = 30, clipping_value: float = 6.0, number_of_negative_samples: int = 10, walk_length: int = 128, iterations: int = 10, - window_size: int = 10, - return_weight: float = 1.0, - explore_weight: float = 1.0, - change_node_type_weight: float = 1.0, - change_edge_type_weight: float = 1.0, + window_size: int = 5, + return_weight: float = 0.25, + explore_weight: float = 4.0, max_neighbours: Optional[int] = 100, learning_rate: float = 0.01, learning_rate_decay: float = 0.9, + central_nodes_embedding_path: Optional[str] = None, + contextual_nodes_embedding_path: Optional[str] = None, normalize_by_degree: bool = False, stochastic_downsample_by_degree: Optional[bool] = False, normalize_learning_rate_by_degree: Optional[bool] = False, - use_zipfian_sampling: Optional[bool] = True, + use_scale_free_distribution: Optional[bool] = True, random_state: int = 42, enable_cache: bool = False ): @@ -35,7 +35,7 @@ def __init__( -------------------------- embedding_size: int = 100 Dimension of the embedding. - epochs: int = 10 + epochs: int = 30 Number of epochs to train the model for. clipping_value: float = 6.0 Value at which we clip the dot product, mostly for numerical stability issues. @@ -47,16 +47,16 @@ def __init__( Maximal length of the walks. iterations: int = 10 Number of iterations of the single walks. - window_size: int = 10 + window_size: int = 5 Window size for the local context. On the borders the window size is trimmed. - return_weight: float = 1.0 + return_weight: float = 0.25 Weight on the probability of returning to the same node the walk just came from Having this higher tends the walks to be more like a Breadth-First Search. Having this very high (> 2) makes search very local. Equal to the inverse of p in the Node2Vec paper. - explore_weight: float = 1.0 + explore_weight: float = 4.0 Weight on the probability of visiting a neighbor node to the one we're coming from in the random walk Having this higher tends the walks to be @@ -64,14 +64,6 @@ def __init__( Having this very high makes search more outward. Having this very low makes search very local. Equal to the inverse of q in the Node2Vec paper. - change_node_type_weight: float = 1.0 - Weight on the probability of visiting a neighbor node of a - different type than the previous node. This only applies to - colored graphs, otherwise it has no impact. - change_edge_type_weight: float = 1.0 - Weight on the probability of visiting a neighbor edge of a - different type than the previous edge. This only applies to - multigraphs, otherwise it has no impact. max_neighbours: Optional[int] = 100 Number of maximum neighbours to consider when using approximated walks. By default, None, we execute exact random walks. @@ -80,6 +72,14 @@ def __init__( The learning rate to use to train the Node2Vec model. By default 0.01. learning_rate_decay: float = 0.9 Factor to reduce the learning rate for at each epoch. By default 0.9. + central_nodes_embedding_path: Optional[str] = None + Path where to mmap and store the central nodes embedding. + This is necessary to embed large graphs whose embedding will not + fit into the available main memory. + contextual_nodes_embedding_path: Optional[str] = None + Path where to mmap and store the central nodes embedding. + This is necessary to embed large graphs whose embedding will not + fit into the available main memory. normalize_by_degree: bool = False Whether to normalize the random walk by the node degree of the destination node degrees. @@ -87,7 +87,7 @@ def __init__( Randomly skip samples with probability proportional to the degree of the central node. By default false. normalize_learning_rate_by_degree: Optional[bool] = False Divide the learning rate by the degree of the central node. By default false. - use_zipfian_sampling: Optional[bool] = True + use_scale_free_distribution: Optional[bool] = True Sample negatives proportionally to their degree. By default true. random_state: int = 42 The random state to reproduce the training sequence. @@ -96,7 +96,6 @@ def __init__( store the computed embedding. """ super().__init__( - model_name="CBOW", embedding_size=embedding_size, epochs=epochs, clipping_value=clipping_value, @@ -106,28 +105,35 @@ def __init__( window_size=window_size, return_weight=return_weight, explore_weight=explore_weight, - change_edge_type_weight=change_edge_type_weight, - change_node_type_weight=change_node_type_weight, max_neighbours=max_neighbours, learning_rate=learning_rate, learning_rate_decay=learning_rate_decay, + central_nodes_embedding_path=central_nodes_embedding_path, + contextual_nodes_embedding_path=contextual_nodes_embedding_path, normalize_by_degree=normalize_by_degree, stochastic_downsample_by_degree=stochastic_downsample_by_degree, normalize_learning_rate_by_degree=normalize_learning_rate_by_degree, - use_zipfian_sampling=use_zipfian_sampling, + use_scale_free_distribution=use_scale_free_distribution, random_state=random_state, enable_cache=enable_cache ) + def parameters(self) -> Dict[str, Any]: + """Returns parameters for smoke test.""" + removed = [ + "change_node_type_weight", + "change_edge_type_weight", + "alpha" + ] + return dict( + **{ + key: value + for key, value in super().parameters().items() + if key not in removed + } + ) + @classmethod def model_name(cls) -> str: """Returns name of the model.""" - return "CBOW" - - @classmethod - def requires_node_types(cls) -> bool: - return False - - @classmethod - def requires_edge_types(cls) -> bool: - return False \ No newline at end of file + return "Node2Vec CBOW" \ No newline at end of file diff --git a/embiggen/embedders/ensmallen_embedders/node2vec_glove.py b/embiggen/embedders/ensmallen_embedders/node2vec_glove.py new file mode 100644 index 00000000..806cb8c6 --- /dev/null +++ b/embiggen/embedders/ensmallen_embedders/node2vec_glove.py @@ -0,0 +1,128 @@ +"""Module providing Node2Vec GloVe model implementation.""" +from typing import Optional, Dict, Any +from embiggen.embedders.ensmallen_embedders.node2vec import Node2VecEnsmallen + +class Node2VecGloVeEnsmallen(Node2VecEnsmallen): + """Class providing Node2Vec GloVe implemeted in Rust from Ensmallen.""" + + def __init__( + self, + embedding_size: int = 100, + alpha: float = 0.75, + epochs: int = 30, + clipping_value: float = 6.0, + walk_length: int = 128, + iterations: int = 10, + window_size: int = 5, + return_weight: float = 0.25, + explore_weight: float = 4.0, + max_neighbours: Optional[int] = 100, + learning_rate: float = 0.001, + learning_rate_decay: float = 0.9, + central_nodes_embedding_path: Optional[str] = None, + contextual_nodes_embedding_path: Optional[str] = None, + normalize_by_degree: bool = False, + random_state: int = 42, + enable_cache: bool = False + ): + """Create new Node2Vec GloVe model. + + Parameters + -------------------------- + embedding_size: int = 100 + Dimension of the embedding. + alpha: float = 0.75 + Alpha parameter for GloVe's loss. + epochs: int = 100 + Number of epochs to train the model for. + window_size: int = 10 + Window size for the local context. + On the borders the window size is trimmed. + clipping_value: float = 6.0 + Value at which we clip the dot product, mostly for numerical stability issues. + By default, `6.0`, where the loss is already close to zero. + walk_length: int = 128 + Maximal length of the walks. + iterations: int = 10 + Number of iterations of the single walks. + window_size: int = 5 + Window size for the local context. + On the borders the window size is trimmed. + return_weight: float = 0.25 + Weight on the probability of returning to the same node the walk just came from + Having this higher tends the walks to be + more like a Breadth-First Search. + Having this very high (> 2) makes search very local. + Equal to the inverse of p in the Node2Vec paper. + explore_weight: float = 4.0 + Weight on the probability of visiting a neighbor node + to the one we're coming from in the random walk + Having this higher tends the walks to be + more like a Depth-First Search. + Having this very high makes search more outward. + Having this very low makes search very local. + Equal to the inverse of q in the Node2Vec paper. + max_neighbours: Optional[int] = 100 + Number of maximum neighbours to consider when using approximated walks. + By default, None, we execute exact random walks. + This is mainly useful for graphs containing nodes with high degrees. + learning_rate: float = 0.001 + The learning rate to use to train the Node2Vec model. By default 0.01. + learning_rate_decay: float = 0.9 + Factor to reduce the learning rate for at each epoch. By default 0.9. + central_nodes_embedding_path: Optional[str] = None + Path where to mmap and store the central nodes embedding. + This is necessary to embed large graphs whose embedding will not + fit into the available main memory. + contextual_nodes_embedding_path: Optional[str] = None + Path where to mmap and store the central nodes embedding. + This is necessary to embed large graphs whose embedding will not + fit into the available main memory. + normalize_by_degree: bool = False + Whether to normalize the random walk by the node degree + of the destination node degrees. + random_state: int = 42 + The random state to reproduce the training sequence. + enable_cache: bool = False + Whether to enable the cache, that is to + store the computed embedding. + """ + super().__init__( + embedding_size=embedding_size, + alpha=alpha, + epochs=epochs, + clipping_value=clipping_value, + walk_length=walk_length, + iterations=iterations, + window_size=window_size, + return_weight=return_weight, + explore_weight=explore_weight, + max_neighbours=max_neighbours, + learning_rate=learning_rate, + learning_rate_decay=learning_rate_decay, + central_nodes_embedding_path=central_nodes_embedding_path, + contextual_nodes_embedding_path=contextual_nodes_embedding_path, + normalize_by_degree=normalize_by_degree, + enable_cache=enable_cache, + random_state=random_state + ) + + def parameters(self) -> Dict[str, Any]: + """Returns parameters for smoke test.""" + removed = [ + "change_node_type_weight", + "change_edge_type_weight", + "number_of_negative_samples" + ] + return dict( + **{ + key: value + for key, value in super().parameters().items() + if key not in removed + } + ) + + @classmethod + def model_name(cls) -> str: + """Returns name of the model.""" + return "Node2Vec GloVe" \ No newline at end of file diff --git a/embiggen/embedders/ensmallen_embedders/skipgram.py b/embiggen/embedders/ensmallen_embedders/node2vec_skipgram.py similarity index 69% rename from embiggen/embedders/ensmallen_embedders/skipgram.py rename to embiggen/embedders/ensmallen_embedders/node2vec_skipgram.py index 89ef432d..b3a1f50e 100644 --- a/embiggen/embedders/ensmallen_embedders/skipgram.py +++ b/embiggen/embedders/ensmallen_embedders/node2vec_skipgram.py @@ -1,31 +1,31 @@ -"""Module providing SkipGram model implementation.""" -from typing import Optional +"""Module providing Node2Vec SkipGram model implementation.""" +from typing import Optional, Dict, Any from embiggen.embedders.ensmallen_embedders.node2vec import Node2VecEnsmallen -class SkipGramEnsmallen(Node2VecEnsmallen): - """Class providing SkipGram implemeted in Rust from Ensmallen.""" +class Node2VecSkipGramEnsmallen(Node2VecEnsmallen): + """Class providing Node2Vec SkipGram implemeted in Rust from Ensmallen.""" def __init__( self, embedding_size: int = 100, - epochs: int = 10, + epochs: int = 30, clipping_value: float = 6.0, number_of_negative_samples: int = 10, walk_length: int = 128, iterations: int = 10, - window_size: int = 10, - return_weight: float = 1.0, - explore_weight: float = 1.0, - change_node_type_weight: float = 1.0, - change_edge_type_weight: float = 1.0, + window_size: int = 5, + return_weight: float = 0.25, + explore_weight: float = 4.0, max_neighbours: Optional[int] = 100, learning_rate: float = 0.01, learning_rate_decay: float = 0.9, + central_nodes_embedding_path: Optional[str] = None, + contextual_nodes_embedding_path: Optional[str] = None, normalize_by_degree: bool = False, stochastic_downsample_by_degree: Optional[bool] = False, normalize_learning_rate_by_degree: Optional[bool] = False, - use_zipfian_sampling: Optional[bool] = True, + use_scale_free_distribution: Optional[bool] = True, random_state: int = 42, enable_cache: bool = False ): @@ -35,11 +35,8 @@ def __init__( -------------------------- embedding_size: int = 100 Dimension of the embedding. - epochs: int = 10 + epochs: int = 30 Number of epochs to train the model for. - window_size: int = 10 - Window size for the local context. - On the borders the window size is trimmed. clipping_value: float = 6.0 Value at which we clip the dot product, mostly for numerical stability issues. By default, `6.0`, where the loss is already close to zero. @@ -50,16 +47,16 @@ def __init__( Maximal length of the walks. iterations: int = 10 Number of iterations of the single walks. - window_size: int = 10 + window_size: int = 5 Window size for the local context. On the borders the window size is trimmed. - return_weight: float = 1.0 + return_weight: float = 0.25 Weight on the probability of returning to the same node the walk just came from Having this higher tends the walks to be more like a Breadth-First Search. Having this very high (> 2) makes search very local. Equal to the inverse of p in the Node2Vec paper. - explore_weight: float = 1.0 + explore_weight: float = 4.0 Weight on the probability of visiting a neighbor node to the one we're coming from in the random walk Having this higher tends the walks to be @@ -67,14 +64,6 @@ def __init__( Having this very high makes search more outward. Having this very low makes search very local. Equal to the inverse of q in the Node2Vec paper. - change_node_type_weight: float = 1.0 - Weight on the probability of visiting a neighbor node of a - different type than the previous node. This only applies to - colored graphs, otherwise it has no impact. - change_edge_type_weight: float = 1.0 - Weight on the probability of visiting a neighbor edge of a - different type than the previous edge. This only applies to - multigraphs, otherwise it has no impact. max_neighbours: Optional[int] = 100 Number of maximum neighbours to consider when using approximated walks. By default, None, we execute exact random walks. @@ -83,6 +72,16 @@ def __init__( The learning rate to use to train the Node2Vec model. By default 0.01. learning_rate_decay: float = 0.9 Factor to reduce the learning rate for at each epoch. By default 0.9. + central_nodes_embedding_path: Optional[str] = None + Path where to mmap and store the central nodes embedding. + + This is necessary to embed large graphs whose embedding will not + fit into the available main memory. + contextual_nodes_embedding_path: Optional[str] = None + Path where to mmap and store the central nodes embedding. + + This is necessary to embed large graphs whose embedding will not + fit into the available main memory. normalize_by_degree: bool = False Whether to normalize the random walk by the node degree of the destination node degrees. @@ -90,7 +89,7 @@ def __init__( Randomly skip samples with probability proportional to the degree of the central node. By default false. normalize_learning_rate_by_degree: Optional[bool] = False Divide the learning rate by the degree of the central node. By default false. - use_zipfian_sampling: Optional[bool] = True + use_scale_free_distribution: Optional[bool] = True Sample negatives proportionally to their degree. By default true. random_state: int = 42 The random state to reproduce the training sequence. @@ -99,7 +98,6 @@ def __init__( store the computed embedding. """ super().__init__( - model_name="SkipGram", embedding_size=embedding_size, epochs=epochs, clipping_value=clipping_value, @@ -109,28 +107,35 @@ def __init__( window_size=window_size, return_weight=return_weight, explore_weight=explore_weight, - change_edge_type_weight=change_edge_type_weight, - change_node_type_weight=change_node_type_weight, max_neighbours=max_neighbours, learning_rate=learning_rate, learning_rate_decay=learning_rate_decay, + central_nodes_embedding_path=central_nodes_embedding_path, + contextual_nodes_embedding_path=contextual_nodes_embedding_path, normalize_by_degree=normalize_by_degree, stochastic_downsample_by_degree=stochastic_downsample_by_degree, normalize_learning_rate_by_degree=normalize_learning_rate_by_degree, - use_zipfian_sampling=use_zipfian_sampling, + use_scale_free_distribution=use_scale_free_distribution, random_state=random_state, - enable_cache=enable_cache, + enable_cache=enable_cache ) + def parameters(self) -> Dict[str, Any]: + """Returns parameters for smoke test.""" + removed = [ + "change_node_type_weight", + "change_edge_type_weight", + "alpha" + ] + return dict( + **{ + key: value + for key, value in super().parameters().items() + if key not in removed + } + ) + @classmethod def model_name(cls) -> str: """Returns name of the model.""" - return "SkipGram" - - @classmethod - def requires_node_types(cls) -> bool: - return False - - @classmethod - def requires_edge_types(cls) -> bool: - return False \ No newline at end of file + return "Node2Vec SkipGram" \ No newline at end of file diff --git a/embiggen/embedders/ensmallen_embedders/node_label_spine.py b/embiggen/embedders/ensmallen_embedders/node_label_spine.py new file mode 100644 index 00000000..6bcc7230 --- /dev/null +++ b/embiggen/embedders/ensmallen_embedders/node_label_spine.py @@ -0,0 +1,119 @@ +"""Module providing Node-label-based SPINE implementation.""" +from typing import Optional, Dict, Any +from ensmallen import Graph +import pandas as pd +from ensmallen import models +from embiggen.embedders.ensmallen_embedders.ensmallen_embedder import EnsmallenEmbedder +from embiggen.utils import EmbeddingResult + + +class NodeLabelSPINE(EnsmallenEmbedder): + """Class implementing the Node-label-based SPINE algorithm.""" + + def __init__( + self, + dtype: Optional[str] = "u8", + maximum_depth: Optional[int] = None, + path: Optional[str] = None, + verbose: bool = False, + enable_cache: bool = False + ): + """Create new Node-label-based SPINE method. + + Parameters + -------------------------- + dtype: Optional[str] = "u8" + Dtype to use for the embedding. + maximum_depth: Optional[int] = None + Maximum depth of the shortest path. + path: Optional[str] = None + Path where to store the mmap-ed embedding. + This parameter is necessary to embed very large graphs. + verbose: bool = False + Whether to show loading bars. + enable_cache: bool = False + Whether to enable the cache, that is to + store the computed embedding. + """ + self._dtype = dtype + self._verbose = verbose + self._maximum_depth = maximum_depth + self._path = path + self._model = models.NodeLabelSPINE( + verbose=self._verbose, + maximum_depth=self._maximum_depth, + path=self._path + ) + + super().__init__( + enable_cache=enable_cache + ) + + def parameters(self) -> Dict[str, Any]: + """Returns parameters of the model.""" + return dict( + **{ + key: value + for key, value in super().parameters().items() + if key != "embedding_size" + }, + **dict( + dtype=self._dtype, + maximum_depth=self._maximum_depth, + path=self._path, + ) + ) + @classmethod + def smoke_test_parameters(cls) -> Dict[str, Any]: + """Returns parameters for smoke test.""" + return dict() + + def _fit_transform( + self, + graph: Graph, + return_dataframe: bool = True, + ) -> EmbeddingResult: + """Return node embedding.""" + node_embedding = self._model.fit_transform( + graph, + dtype=self._dtype, + ) + if return_dataframe: + node_embedding = pd.DataFrame( + node_embedding, + index=graph.get_node_names() + ) + return EmbeddingResult( + embedding_method_name=self.model_name(), + node_embeddings=node_embedding + ) + + @classmethod + def model_name(cls) -> str: + """Returns name of the model.""" + return "Node-label-based SPINE" + + @classmethod + def requires_node_types(cls) -> bool: + """Returns whether the model requires node types.""" + return True + + @classmethod + def get_minimum_required_number_of_node_types(cls) -> int: + """Requires minimum number of required node types.""" + return 2 + + @classmethod + def can_use_edge_weights(cls) -> bool: + """Returns whether the model can optionally use edge weights.""" + return False + + @classmethod + def can_use_edge_types(cls) -> bool: + """Returns whether the model can optionally use edge types.""" + return False + + @classmethod + def is_stocastic(cls) -> bool: + """Returns whether the model is stocastic and has therefore a random state.""" + return False \ No newline at end of file diff --git a/embiggen/embedders/ensmallen_embedders/node_label_wine.py b/embiggen/embedders/ensmallen_embedders/node_label_wine.py new file mode 100644 index 00000000..109ac40c --- /dev/null +++ b/embiggen/embedders/ensmallen_embedders/node_label_wine.py @@ -0,0 +1,120 @@ +"""Module providing Node-label-based WINE implementation.""" +from typing import Optional, Dict, Any +from ensmallen import Graph +import pandas as pd +from ensmallen import models +from embiggen.embedders.ensmallen_embedders.ensmallen_embedder import EnsmallenEmbedder +from embiggen.utils import EmbeddingResult + + +class NodeLabelWINE(EnsmallenEmbedder): + """Class implementing the Node-label-based WINE algorithm.""" + + def __init__( + self, + dtype: Optional[str] = "u8", + walk_length: Optional[int] = None, + path: Optional[str] = None, + verbose: bool = False, + enable_cache: bool = False + ): + """Create new Node-label-based WINE method. + + Parameters + -------------------------- + dtype: Optional[str] = "u8" + Dtype to use for the embedding. + walk_length: int = 2 + Length of the random walk. + By default 2, to capture exclusively the immediate context. + path: Optional[str] = None + Path where to store the mmap-ed embedding. + This parameter is necessary to embed very large graphs. + verbose: bool = False + Whether to show loading bars. + enable_cache: bool = False + Whether to enable the cache, that is to + store the computed embedding. + """ + self._dtype = dtype + self._verbose = verbose + self._walk_length = walk_length + self._path = path + self._model = models.NodeLabelWINE( + verbose=self._verbose, + walk_length=self._walk_length, + path=self._path + ) + + super().__init__( + enable_cache=enable_cache + ) + + def parameters(self) -> Dict[str, Any]: + """Returns parameters of the model.""" + return dict( + **{ + key: value + for key, value in super().parameters().items() + if key != "embedding_size" + }, + **dict( + dtype=self._dtype, + walk_length=self._walk_length, + path=self._path, + ) + ) + @classmethod + def smoke_test_parameters(cls) -> Dict[str, Any]: + """Returns parameters for smoke test.""" + return dict() + + def _fit_transform( + self, + graph: Graph, + return_dataframe: bool = True, + ) -> EmbeddingResult: + """Return node embedding.""" + node_embedding = self._model.fit_transform( + graph, + dtype=self._dtype, + ) + if return_dataframe: + node_embedding = pd.DataFrame( + node_embedding, + index=graph.get_node_names() + ) + return EmbeddingResult( + embedding_method_name=self.model_name(), + node_embeddings=node_embedding + ) + + @classmethod + def model_name(cls) -> str: + """Returns name of the model.""" + return "Node-label-based WINE" + + @classmethod + def requires_node_types(cls) -> bool: + """Returns whether the model requires node types.""" + return True + + @classmethod + def get_minimum_required_number_of_node_types(cls) -> int: + """Requires minimum number of required node types.""" + return 2 + + @classmethod + def can_use_edge_weights(cls) -> bool: + """Returns whether the model can optionally use edge weights.""" + return False + + @classmethod + def can_use_edge_types(cls) -> bool: + """Returns whether the model can optionally use edge types.""" + return False + + @classmethod + def is_stocastic(cls) -> bool: + """Returns whether the model is stocastic and has therefore a random state.""" + return False \ No newline at end of file diff --git a/embiggen/embedders/ensmallen_embedders/score_spine.py b/embiggen/embedders/ensmallen_embedders/score_spine.py new file mode 100644 index 00000000..cf018f30 --- /dev/null +++ b/embiggen/embedders/ensmallen_embedders/score_spine.py @@ -0,0 +1,125 @@ +"""Module providing Score-based SPINE implementation.""" +from typing import Optional, Dict, Any +from ensmallen import Graph +import pandas as pd +import numpy as np +from ensmallen import models +from embiggen.embedders.ensmallen_embedders.ensmallen_embedder import EnsmallenEmbedder +from embiggen.utils import EmbeddingResult + + +class ScoreSPINE(EnsmallenEmbedder): + """Class implementing the Score-based SPINE algorithm.""" + + def __init__( + self, + scores: Optional[np.ndarray] = None, + embedding_size: int = 100, + dtype: Optional[str] = "u8", + maximum_depth: Optional[int] = None, + path: Optional[str] = None, + verbose: bool = False, + enable_cache: bool = False + ): + """Create new Score-based SPINE method. + + Parameters + -------------------------- + scores: Optional[np.ndarray] = None + Numpy array to be used to sort the anchor nodes. + embedding_size: int = 100 + Dimension of the embedding. + dtype: Optional[str] = "u8" + Dtype to use for the embedding. + maximum_depth: Optional[int] = None + Maximum depth of the shortest path. + path: Optional[str] = None + Path where to store the mmap-ed embedding. + This parameter is necessary to embed very large graphs. + verbose: bool = False + Whether to show loading bars. + enable_cache: bool = False + Whether to enable the cache, that is to + store the computed embedding. + """ + self._dtype = dtype + self._verbose = verbose + self._maximum_depth = maximum_depth + self._path = path + self._scores = scores + self._model = models.ScoreSPINE( + embedding_size=embedding_size, + verbose=self._verbose, + maximum_depth=self._maximum_depth, + path=self._path + ) + + super().__init__( + embedding_size=embedding_size, + enable_cache=enable_cache + ) + + def parameters(self) -> Dict[str, Any]: + """Returns parameters of the model.""" + return { + **super().parameters(), + **dict( + dtype=self._dtype, + maximum_depth=self._maximum_depth, + path=self._path, + ) + } + + @classmethod + def smoke_test_parameters(cls) -> Dict[str, Any]: + """Returns parameters for smoke test.""" + return dict( + embedding_size=5, + ) + + def _fit_transform( + self, + graph: Graph, + return_dataframe: bool = True, + ) -> EmbeddingResult: + """Return node embedding.""" + node_embedding = self._model.fit_transform( + scores=(np.ones(graph.get_number_of_nodes()) + if self._scores is None else self._scores).astype(np.float32), + graph=graph, + dtype=self._dtype, + ) + if return_dataframe: + node_embedding = pd.DataFrame( + node_embedding, + index=graph.get_node_names() + ) + return EmbeddingResult( + embedding_method_name=self.model_name(), + node_embeddings=node_embedding + ) + + @classmethod + def model_name(cls) -> str: + """Returns name of the model.""" + return "Score-based SPINE" + + @classmethod + def can_use_node_types(cls) -> bool: + """Returns whether the model can optionally use node types.""" + return False + + @classmethod + def can_use_edge_weights(cls) -> bool: + """Returns whether the model can optionally use edge weights.""" + return False + + @classmethod + def can_use_edge_types(cls) -> bool: + """Returns whether the model can optionally use edge types.""" + return False + + @classmethod + def is_stocastic(cls) -> bool: + """Returns whether the model is stocastic and has therefore a random state.""" + return False diff --git a/embiggen/embedders/ensmallen_embedders/score_wine.py b/embiggen/embedders/ensmallen_embedders/score_wine.py new file mode 100644 index 00000000..35f9b2a9 --- /dev/null +++ b/embiggen/embedders/ensmallen_embedders/score_wine.py @@ -0,0 +1,126 @@ +"""Module providing Score-based WINE implementation.""" +from typing import Optional, Dict, Any +from ensmallen import Graph +import pandas as pd +import numpy as np +from ensmallen import models +from embiggen.embedders.ensmallen_embedders.ensmallen_embedder import EnsmallenEmbedder +from embiggen.utils import EmbeddingResult + + +class ScoreWINE(EnsmallenEmbedder): + """Class implementing the Score-based WINE algorithm.""" + + def __init__( + self, + scores: Optional[np.ndarray] = None, + embedding_size: int = 100, + dtype: Optional[str] = "u8", + walk_length: Optional[int] = None, + path: Optional[str] = None, + verbose: bool = False, + enable_cache: bool = False + ): + """Create new Score-based WINE method. + + Parameters + -------------------------- + scores: Optional[np.ndarray] = None + Numpy array to be used to sort the anchor nodes. + embedding_size: int = 100 + Dimension of the embedding. + dtype: Optional[str] = "u8" + Dtype to use for the embedding. + walk_length: int = 2 + Length of the random walk. + By default 2, to capture exclusively the immediate context. + path: Optional[str] = None + Path where to store the mmap-ed embedding. + This parameter is necessary to embed very large graphs. + verbose: bool = False + Whether to show loading bars. + enable_cache: bool = False + Whether to enable the cache, that is to + store the computed embedding. + """ + self._dtype = dtype + self._verbose = verbose + self._walk_length = walk_length + self._path = path + self._scores = scores + self._model = models.ScoreWINE( + embedding_size=embedding_size, + verbose=self._verbose, + walk_length=self._walk_length, + path=self._path + ) + + super().__init__( + embedding_size=embedding_size, + enable_cache=enable_cache + ) + + def parameters(self) -> Dict[str, Any]: + """Returns parameters of the model.""" + return { + **super().parameters(), + **dict( + dtype=self._dtype, + walk_length=self._walk_length, + path=self._path, + ) + } + + @classmethod + def smoke_test_parameters(cls) -> Dict[str, Any]: + """Returns parameters for smoke test.""" + return dict( + embedding_size=5, + ) + + def _fit_transform( + self, + graph: Graph, + return_dataframe: bool = True, + ) -> EmbeddingResult: + """Return node embedding.""" + node_embedding = self._model.fit_transform( + scores=(np.ones(graph.get_number_of_nodes()) + if self._scores is None else self._scores).astype(np.float32), + graph=graph, + dtype=self._dtype, + ) + if return_dataframe: + node_embedding = pd.DataFrame( + node_embedding, + index=graph.get_node_names() + ) + return EmbeddingResult( + embedding_method_name=self.model_name(), + node_embeddings=node_embedding + ) + + @classmethod + def model_name(cls) -> str: + """Returns name of the model.""" + return "Score-based WINE" + + @classmethod + def can_use_node_types(cls) -> bool: + """Returns whether the model can optionally use node types.""" + return False + + @classmethod + def can_use_edge_weights(cls) -> bool: + """Returns whether the model can optionally use edge weights.""" + return False + + @classmethod + def can_use_edge_types(cls) -> bool: + """Returns whether the model can optionally use edge types.""" + return False + + @classmethod + def is_stocastic(cls) -> bool: + """Returns whether the model is stocastic and has therefore a random state.""" + return False diff --git a/embiggen/embedders/ensmallen_embedders/second_order_line.py b/embiggen/embedders/ensmallen_embedders/second_order_line.py new file mode 100644 index 00000000..19858731 --- /dev/null +++ b/embiggen/embedders/ensmallen_embedders/second_order_line.py @@ -0,0 +1,140 @@ +"""Module providing second-order LINE implementation.""" +from typing import Dict, Any, Optional +from ensmallen import Graph +import numpy as np +import pandas as pd +from ensmallen import models +from embiggen.embedders.ensmallen_embedders.ensmallen_embedder import EnsmallenEmbedder +from embiggen.utils import EmbeddingResult + + +class SecondOrderLINEEnsmallen(EnsmallenEmbedder): + """Class implementing the second-order LINE algorithm.""" + + def __init__( + self, + embedding_size: int = 100, + epochs: int = 100, + learning_rate: float = 0.01, + learning_rate_decay: float = 0.9, + avoid_false_negatives: bool = False, + node_embedding_path: Optional[str] = None, + contextual_node_embedding_path: Optional[str] = None, + random_state: int = 42, + verbose: bool = False, + enable_cache: bool = False + ): + """Create new abstract Node2Vec method. + + Parameters + -------------------------- + embedding_size: int = 100 + Dimension of the embedding. + epochs: int = 100 + The number of epochs to run the model for, by default 10. + learning_rate: float = 0.01 + The learning rate to update the gradient, by default 0.01. + learning_rate_decay: float = 0.9 + Factor to reduce the learning rate for at each epoch. By default 0.9. + avoid_false_negatives: bool = False + Whether to avoid sampling false negatives. + This may cause a slower training. + node_embedding_path: Optional[str] = None + Path where to mmap and store the nodes embedding. + This is necessary to embed large graphs whose embedding will not + fit into the available main memory. + contextual_node_embedding_path: Optional[str] = None + Path where to mmap and store the contextual nodes embedding. + This is necessary to embed large graphs whose embedding will not + fit into the available main memory. + random_state: int = 42 + Random state to reproduce the embeddings. + verbose: bool = False + Whether to show loading bars. + enable_cache: bool = False + Whether to enable the cache, that is to + store the computed embedding. + """ + self._kwargs = dict( + epochs=epochs, + learning_rate=learning_rate, + learning_rate_decay=learning_rate_decay, + avoid_false_negatives=avoid_false_negatives, + node_embedding_path=node_embedding_path, + contextual_node_embedding_path=contextual_node_embedding_path, + verbose=verbose + ) + + self._model = models.SecondOrderLINE( + embedding_size=embedding_size, + random_state=random_state, + **self._kwargs + ) + + super().__init__( + embedding_size=embedding_size, + enable_cache=enable_cache, + random_state=random_state + ) + + def parameters(self) -> Dict[str, Any]: + """Returns parameters of the model.""" + return dict( + **super().parameters(), + **self._kwargs + ) + + @classmethod + def smoke_test_parameters(cls) -> Dict[str, Any]: + """Returns parameters for smoke test.""" + return dict( + embedding_size=5, + epochs=1 + ) + + def _fit_transform( + self, + graph: Graph, + return_dataframe: bool = True, + ) -> EmbeddingResult: + """Return node embedding.""" + node_embeddings = self._model.fit_transform(graph) + if return_dataframe: + node_names = graph.get_node_names() + node_embeddings = [ + pd.DataFrame( + node_embedding, + index=node_names + ) + for node_embedding in node_embeddings + ] + + return EmbeddingResult( + embedding_method_name=self.model_name(), + node_embeddings=node_embeddings, + ) + + @classmethod + def model_name(cls) -> str: + """Returns name of the model.""" + return "Second-order LINE" + + @classmethod + def can_use_edge_weights(cls) -> bool: + """Returns whether the model can optionally use edge weights.""" + return False + + @classmethod + def can_use_node_types(cls) -> bool: + """Returns whether the model can optionally use node types.""" + return False + + @classmethod + def can_use_edge_types(cls) -> bool: + """Returns whether the model can optionally use edge types.""" + return False + + @classmethod + def is_stocastic(cls) -> bool: + """Returns whether the model is stocastic and has therefore a random state.""" + return True \ No newline at end of file diff --git a/embiggen/embedders/ensmallen_embedders/siamese_model.py b/embiggen/embedders/ensmallen_embedders/siamese_model.py new file mode 100644 index 00000000..72615225 --- /dev/null +++ b/embiggen/embedders/ensmallen_embedders/siamese_model.py @@ -0,0 +1,116 @@ +"""Module providing Siamese implementation.""" +from typing import Dict, Any, Optional +from ensmallen import models +from userinput.utils import must_be_in_set +from embiggen.embedders.ensmallen_embedders.ensmallen_embedder import EnsmallenEmbedder +from embiggen.utils import abstract_class + + +@abstract_class +class SiameseEnsmallen(EnsmallenEmbedder): + """Class implementing the Siamese algorithm.""" + + models = { + "TransE": models.TransE, + "TransH": models.TransH, + "Unstructured": models.Unstructured, + "Structured Embedding": models.StructuredEmbedding, + } + + def __init__( + self, + embedding_size: int = 100, + relu_bias: float = 1.0, + epochs: int = 100, + learning_rate: float = 0.1, + learning_rate_decay: float = 0.9, + node_embedding_path: Optional[str] = None, + random_state: int = 42, + verbose: bool = False, + enable_cache: bool = False, + **paths: Dict[str, str] + ): + """Create new abstract Node2Vec method. + + Parameters + -------------------------- + embedding_size: int = 100 + Dimension of the embedding. + relu_bias: float = 1.0 + Bias to use for the relu. + In the Siamese paper it is called gamma. + epochs: int = 100 + The number of epochs to run the model for, by default 10. + learning_rate: float = 0.05 + The learning rate to update the gradient, by default 0.01. + learning_rate_decay: float = 0.9 + Factor to reduce the learning rate for at each epoch. By default 0.9. + node_embedding_path: Optional[str] = None + Path where to mmap and store the nodes embedding. + This is necessary to embed large graphs whose embedding will not + fit into the available main memory. + random_state: int = 42 + Random state to reproduce the embeddings. + verbose: bool = False + Whether to show loading bars. + enable_cache: bool = False + Whether to enable the cache, that is to + store the computed embedding. + """ + self._kwargs = dict( + relu_bias=relu_bias, + epochs=epochs, + learning_rate=learning_rate, + learning_rate_decay=learning_rate_decay, + node_embedding_path=node_embedding_path, + verbose=verbose, + **paths + ) + + self._model_name = must_be_in_set( + self.model_name(), + SiameseEnsmallen.models, + "Siamese models" + ) + + self._model = SiameseEnsmallen.models[self._model_name]( + embedding_size=embedding_size, + random_state=random_state, + **self._kwargs + ) + + super().__init__( + embedding_size=embedding_size, + enable_cache=enable_cache, + random_state=random_state + ) + + def parameters(self) -> Dict[str, Any]: + """Returns parameters of the model.""" + return dict( + **super().parameters(), + **self._kwargs, + ) + + @classmethod + def smoke_test_parameters(cls) -> Dict[str, Any]: + """Returns parameters for smoke test.""" + return dict( + embedding_size=5, + epochs=1 + ) + + @classmethod + def can_use_edge_weights(cls) -> bool: + """Returns whether the model can optionally use edge weights.""" + return False + + @classmethod + def can_use_node_types(cls) -> bool: + """Returns whether the model can optionally use node types.""" + return False + + @classmethod + def is_stocastic(cls) -> bool: + """Returns whether the model is stocastic and has therefore a random state.""" + return True diff --git a/embiggen/embedders/ensmallen_embedders/sociodim.py b/embiggen/embedders/ensmallen_embedders/sociodim.py index aa8f3923..330153c5 100644 --- a/embiggen/embedders/ensmallen_embedders/sociodim.py +++ b/embiggen/embedders/ensmallen_embedders/sociodim.py @@ -1,18 +1,20 @@ """Module providing SocioDim implementation.""" -from typing import Optional, Dict, Any +from typing import Dict, Any from ensmallen import Graph import pandas as pd -import numpy as np from scipy.linalg import eigh -from embiggen.utils.abstract_models import AbstractEmbeddingModel, EmbeddingResult +from scipy.sparse.linalg import eigsh +from embiggen.embedders.ensmallen_embedders.ensmallen_embedder import EnsmallenEmbedder +from embiggen.utils import EmbeddingResult -class SocioDimEnsmallen(AbstractEmbeddingModel): +class SocioDimEnsmallen(EnsmallenEmbedder): """Class implementing the SocioDim algorithm.""" def __init__( self, embedding_size: int = 100, + use_sparse_reduce: bool = True, enable_cache: bool = False ): """Create new SocioDim method. @@ -21,30 +23,53 @@ def __init__( -------------------------- embedding_size: int = 100 Dimension of the embedding. + use_sparse_reduce: bool = True + Whether to use the sparse reduce or the dense reduce for + the computation of eigenvectors. + For both reduce mechanisms, we are using LAPACK implementations. + For some currently unknown reason, their implementation using + a sparse reduce, even on dense matrices, yields better results. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. """ + self._use_sparse_reduce = use_sparse_reduce super().__init__( embedding_size=embedding_size, enable_cache=enable_cache, ) + def parameters(self) -> Dict[str, Any]: + """Returns parameters of the model.""" + return { + **super().parameters(), + **dict( + use_sparse_reduce=self._use_sparse_reduce, + ) + } + def _fit_transform( self, graph: Graph, return_dataframe: bool = True, - verbose: bool = True ) -> EmbeddingResult: """Return node embedding.""" - number_of_nodes = graph.get_number_of_nodes() - embedding = eigh( - graph.get_dense_modularity_matrix(), - eigvals=( - number_of_nodes-self._embedding_size, - number_of_nodes-1 - ) - )[1] + if self._use_sparse_reduce: + embedding = eigsh( + graph.get_dense_modularity_matrix(), + k=self._embedding_size, + which="LM", + return_eigenvectors=True + )[1] + else: + number_of_nodes = graph.get_number_of_nodes() + embedding = eigh( + graph.get_dense_modularity_matrix(), + eigvals=( + number_of_nodes-self._embedding_size, + number_of_nodes-1 + ) + )[1] if return_dataframe: node_names = graph.get_node_names() @@ -57,27 +82,11 @@ def _fit_transform( node_embeddings=embedding ) - @classmethod - def task_name(cls) -> str: - return "Node Embedding" - @classmethod def model_name(cls) -> str: """Returns name of the model.""" return "SocioDim" - @classmethod - def library_name(cls) -> str: - return "Ensmallen" - - @classmethod - def requires_nodes_sorted_by_decreasing_node_degree(cls) -> bool: - return False - - @classmethod - def is_topological(cls) -> bool: - return True - @classmethod def can_use_edge_weights(cls) -> bool: """Returns whether the model can optionally use edge weights.""" diff --git a/embiggen/embedders/ensmallen_embedders/structured_embedding.py b/embiggen/embedders/ensmallen_embedders/structured_embedding.py new file mode 100644 index 00000000..ccf0334b --- /dev/null +++ b/embiggen/embedders/ensmallen_embedders/structured_embedding.py @@ -0,0 +1,116 @@ +"""Module providing Structured Embedding implementation.""" +from typing import Optional +from ensmallen import Graph +import pandas as pd +from embiggen.embedders.ensmallen_embedders.siamese_model import SiameseEnsmallen +from embiggen.utils import EmbeddingResult + + +class StructuredEmbeddingEnsmallen(SiameseEnsmallen): + """Class implementing the Structured Embedding algorithm.""" + + def __init__( + self, + embedding_size: int = 100, + relu_bias: float = 1.0, + epochs: int = 100, + learning_rate: float = 0.05, + learning_rate_decay: float = 0.9, + node_embedding_path: Optional[str] = None, + source_edge_type_embedding_path: Optional[str] = None, + destination_edge_type_embedding_path: Optional[str] = None, + random_state: int = 42, + verbose: bool = False, + enable_cache: bool = False + ): + """Create new Structured Embedding method. + + Parameters + -------------------------- + model_name: str + The model to instantiate. + embedding_size: int = 100 + Dimension of the embedding. + relu_bias: float = 1.0 + Bias to use for the relu. + In the StructuredEmbedding paper it is called gamma. + epochs: int = 100 + The number of epochs to run the model for, by default 10. + learning_rate: float = 0.05 + The learning rate to update the gradient, by default 0.01. + learning_rate_decay: float = 0.9 + Factor to reduce the learning rate for at each epoch. By default 0.9. + node_embedding_path: Optional[str] = None + Path where to mmap and store the nodes embedding. + This is necessary to embed large graphs whose embedding will not + fit into the available main memory. + source_edge_type_embedding_path: Optional[str] = None + Path where to mmap and store the source edge type embedding. + This is necessary to embed large graphs whose embedding will not + fit into the available main memory. + destination_edge_type_embedding_path: Optional[str] = None + Path where to mmap and store the destination edge type embedding. + This is necessary to embed large graphs whose embedding will not + fit into the available main memory. + random_state: int = 42 + Random state to reproduce the embeddings. + verbose: bool = False + Whether to show loading bars. + enable_cache: bool = False + Whether to enable the cache, that is to + store the computed embedding. + """ + super().__init__( + embedding_size=embedding_size, + relu_bias=relu_bias, + epochs=epochs, + learning_rate=learning_rate, + learning_rate_decay=learning_rate_decay, + node_embedding_path=node_embedding_path, + source_edge_type_embedding_path=source_edge_type_embedding_path, + destination_edge_type_embedding_path=destination_edge_type_embedding_path, + random_state=random_state, + verbose=verbose, + enable_cache=enable_cache, + ) + + def _fit_transform( + self, + graph: Graph, + return_dataframe: bool = True, + ) -> EmbeddingResult: + """Return node embedding.""" + node_embedding, source_edge_type_embedding, destination_edge_type_embedding = self._model.fit_transform( + graph, + ) + if return_dataframe: + node_embedding = pd.DataFrame( + node_embedding, + index=graph.get_node_names() + ) + source_edge_type_embedding = pd.DataFrame( + source_edge_type_embedding, + index=graph.get_unique_edge_type_names() + ) + destination_edge_type_embedding = pd.DataFrame( + destination_edge_type_embedding, + index=graph.get_unique_edge_type_names() + ) + + return EmbeddingResult( + embedding_method_name=self.model_name(), + node_embeddings=node_embedding, + edge_type_embeddings=[ + source_edge_type_embedding, + destination_edge_type_embedding + ] + ) + + @classmethod + def model_name(cls) -> str: + """Returns name of the model.""" + return "Structured Embedding" + + @classmethod + def requires_edge_types(cls) -> bool: + return True \ No newline at end of file diff --git a/embiggen/embedders/ensmallen_embedders/transe.py b/embiggen/embedders/ensmallen_embedders/transe.py index 4069762f..1a135d14 100644 --- a/embiggen/embedders/ensmallen_embedders/transe.py +++ b/embiggen/embedders/ensmallen_embedders/transe.py @@ -1,101 +1,81 @@ """Module providing TransE implementation.""" -from typing import Dict, Any, Union +from typing import Optional from ensmallen import Graph -import numpy as np import pandas as pd -from ensmallen import models -from embiggen.utils.abstract_models import AbstractEmbeddingModel, EmbeddingResult +from embiggen.embedders.ensmallen_embedders.siamese_model import SiameseEnsmallen +from embiggen.utils import EmbeddingResult -class TransEEnsmallen(AbstractEmbeddingModel): +class TransEEnsmallen(SiameseEnsmallen): """Class implementing the TransE algorithm.""" def __init__( self, embedding_size: int = 100, - renormalize: bool = True, relu_bias: float = 1.0, epochs: int = 100, - learning_rate: float = 0.01, + learning_rate: float = 0.05, learning_rate_decay: float = 0.9, + node_embedding_path: Optional[str] = None, + edge_type_embedding_path: Optional[str] = None, random_state: int = 42, + verbose: bool = False, enable_cache: bool = False ): - """Create new abstract Node2Vec method. + """Create new TransE method. Parameters -------------------------- + model_name: str + The model to instantiate. embedding_size: int = 100 Dimension of the embedding. - renormalize: bool = True - Whether to renormalize at each loop, by default true. relu_bias: float = 1.0 Bias to use for the relu. In the TransE paper it is called gamma. epochs: int = 100 The number of epochs to run the model for, by default 10. - learning_rate: float = 0.01 + learning_rate: float = 0.05 The learning rate to update the gradient, by default 0.01. learning_rate_decay: float = 0.9 Factor to reduce the learning rate for at each epoch. By default 0.9. + node_embedding_path: Optional[str] = None + Path where to mmap and store the nodes embedding. + This is necessary to embed large graphs whose embedding will not + fit into the available main memory. + edge_type_embedding_path: Optional[str] = None + Path where to mmap and store the edge type embedding. + This is necessary to embed large graphs whose embedding will not + fit into the available main memory. random_state: int = 42 Random state to reproduce the embeddings. + verbose: bool = False + Whether to show loading bars. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. """ - self._renormalize = renormalize - self._relu_bias = relu_bias - self._epochs = epochs - self._learning_rate = learning_rate - self._learning_rate_decay = learning_rate_decay - - self._model = models.TransE( - embedding_size=embedding_size, - renormalize=renormalize, - random_state=random_state - ) - super().__init__( embedding_size=embedding_size, + relu_bias=relu_bias, + epochs=epochs, + learning_rate=learning_rate, + learning_rate_decay=learning_rate_decay, + node_embedding_path=node_embedding_path, + edge_type_embedding_path=edge_type_embedding_path, + random_state=random_state, + verbose=verbose, enable_cache=enable_cache, - random_state=random_state - ) - - def parameters(self) -> Dict[str, Any]: - """Returns parameters of the model.""" - return { - **super().parameters(), - **dict( - renormalize=self._renormalize, - relu_bias=self._relu_bias, - epochs=self._epochs, - learning_rate=self._learning_rate, - learning_rate_decay=self._learning_rate_decay, - ) - } - - @classmethod - def smoke_test_parameters(cls) -> Dict[str, Any]: - """Returns parameters for smoke test.""" - return dict( - embedding_size=5, - epochs=1 ) def _fit_transform( self, graph: Graph, return_dataframe: bool = True, - verbose: bool = True ) -> EmbeddingResult: """Return node embedding.""" node_embedding, edge_type_embedding = self._model.fit_transform( graph, - epochs=self._epochs, - learning_rate=self._learning_rate, - learning_rate_decay=self._learning_rate_decay, - verbose=verbose, ) if return_dataframe: node_embedding = pd.DataFrame( @@ -109,51 +89,15 @@ def _fit_transform( return EmbeddingResult( embedding_method_name=self.model_name(), - node_embeddings= node_embedding, - edge_type_embeddings= edge_type_embedding, + node_embeddings=node_embedding, + edge_type_embeddings=edge_type_embedding, ) - @classmethod - def task_name(cls) -> str: - return "Node Embedding" - @classmethod def model_name(cls) -> str: """Returns name of the model.""" return "TransE" - @classmethod - def library_name(cls) -> str: - return "Ensmallen" - - @classmethod - def requires_nodes_sorted_by_decreasing_node_degree(cls) -> bool: - return False - - @classmethod - def is_topological(cls) -> bool: - return True - @classmethod def requires_edge_types(cls) -> bool: return True - - @classmethod - def can_use_edge_weights(cls) -> bool: - """Returns whether the model can optionally use edge weights.""" - return False - - @classmethod - def can_use_node_types(cls) -> bool: - """Returns whether the model can optionally use node types.""" - return False - - @classmethod - def task_involves_edge_types(cls) -> bool: - """Returns whether the model task involves edge types.""" - return True - - @classmethod - def is_stocastic(cls) -> bool: - """Returns whether the model is stocastic and has therefore a random state.""" - return True \ No newline at end of file diff --git a/embiggen/embedders/ensmallen_embedders/transh.py b/embiggen/embedders/ensmallen_embedders/transh.py new file mode 100644 index 00000000..2de65834 --- /dev/null +++ b/embiggen/embedders/ensmallen_embedders/transh.py @@ -0,0 +1,116 @@ +"""Module providing TransH implementation.""" +from typing import Optional +from ensmallen import Graph +import pandas as pd +from embiggen.embedders.ensmallen_embedders.siamese_model import SiameseEnsmallen +from embiggen.utils import EmbeddingResult + + +class TransHEnsmallen(SiameseEnsmallen): + """Class implementing the TransH algorithm.""" + + def __init__( + self, + embedding_size: int = 100, + relu_bias: float = 1.0, + epochs: int = 100, + learning_rate: float = 0.1, + learning_rate_decay: float = 0.9, + node_embedding_path: Optional[str] = None, + mult_edge_type_embedding_path: Optional[str] = None, + bias_edge_type_embedding_path: Optional[str] = None, + random_state: int = 42, + verbose: bool = False, + enable_cache: bool = False + ): + """Create new TransH method. + + Parameters + -------------------------- + model_name: str + The model to instantiate. + embedding_size: int = 100 + Dimension of the embedding. + relu_bias: float = 1.0 + Bias to use for the relu. + In the TransH paper it is called gamma. + epochs: int = 100 + The number of epochs to run the model for, by default 10. + learning_rate: float = 0.05 + The learning rate to update the gradient, by default 0.01. + learning_rate_decay: float = 0.9 + Factor to reduce the learning rate for at each epoch. By default 0.9. + node_embedding_path: Optional[str] = None + Path where to mmap and store the nodes embedding. + This is necessary to embed large graphs whose embedding will not + fit into the available main memory. + mult_edge_type_embedding_path: Optional[str] = None + Path where to mmap and store the multiplicative edge type embedding. + This is necessary to embed large graphs whose embedding will not + fit into the available main memory. + bias_edge_type_embedding_path: Optional[str] = None + Path where to mmap and store the bias edge type embedding. + This is necessary to embed large graphs whose embedding will not + fit into the available main memory. + random_state: int = 42 + Random state to reproduce the embeddings. + verbose: bool = False + Whether to show loading bars. + enable_cache: bool = False + Whether to enable the cache, that is to + store the computed embedding. + """ + super().__init__( + embedding_size=embedding_size, + relu_bias=relu_bias, + epochs=epochs, + learning_rate=learning_rate, + learning_rate_decay=learning_rate_decay, + node_embedding_path=node_embedding_path, + mult_edge_type_embedding_path=mult_edge_type_embedding_path, + bias_edge_type_embedding_path=bias_edge_type_embedding_path, + random_state=random_state, + verbose=verbose, + enable_cache=enable_cache, + ) + + def _fit_transform( + self, + graph: Graph, + return_dataframe: bool = True, + ) -> EmbeddingResult: + """Return node embedding.""" + node_embedding, mult_edge_type_embedding, bias_edge_type_embedding = self._model.fit_transform( + graph, + ) + if return_dataframe: + node_embedding = pd.DataFrame( + node_embedding, + index=graph.get_node_names() + ) + mult_edge_type_embedding = pd.DataFrame( + mult_edge_type_embedding, + index=graph.get_unique_edge_type_names() + ) + bias_edge_type_embedding = pd.DataFrame( + bias_edge_type_embedding, + index=graph.get_unique_edge_type_names() + ) + + return EmbeddingResult( + embedding_method_name=self.model_name(), + node_embeddings=node_embedding, + edge_type_embeddings=[ + mult_edge_type_embedding, + bias_edge_type_embedding + ] + ) + + @classmethod + def model_name(cls) -> str: + """Returns name of the model.""" + return "TransH" + + @classmethod + def requires_edge_types(cls) -> bool: + return True \ No newline at end of file diff --git a/embiggen/embedders/ensmallen_embedders/unstructured.py b/embiggen/embedders/ensmallen_embedders/unstructured.py new file mode 100644 index 00000000..edcaf5f0 --- /dev/null +++ b/embiggen/embedders/ensmallen_embedders/unstructured.py @@ -0,0 +1,92 @@ +"""Module providing Unstructured implementation.""" +from typing import Optional +from ensmallen import Graph +import pandas as pd +from embiggen.embedders.ensmallen_embedders.siamese_model import SiameseEnsmallen +from embiggen.utils import EmbeddingResult + + +class UnstructuredEnsmallen(SiameseEnsmallen): + """Class implementing the Unstructured algorithm.""" + + def __init__( + self, + embedding_size: int = 100, + relu_bias: float = 1.0, + epochs: int = 100, + learning_rate: float = 0.01, + learning_rate_decay: float = 0.9, + node_embedding_path: Optional[str] = None, + random_state: int = 42, + verbose: bool = False, + enable_cache: bool = False + ): + """Create new abstract Node2Vec method. + + Parameters + -------------------------- + model_name: str + The model to instantiate. + embedding_size: int = 100 + Dimension of the embedding. + relu_bias: float = 1.0 + Bias to use for the relu. + In the Unstructured paper it is called gamma. + epochs: int = 100 + The number of epochs to run the model for, by default 100. + learning_rate: float = 0.01 + The learning rate to update the gradient, by default 0.01. + learning_rate_decay: float = 0.9 + Factor to reduce the learning rate for at each epoch. By default 0.9. + node_embedding_path: Optional[str] = None + Path where to mmap and store the nodes embedding. + This is necessary to embed large graphs whose embedding will not + fit into the available main memory. + random_state: int = 42 + Random state to reproduce the embeddings. + verbose: bool = False + Whether to show loading bars. + enable_cache: bool = False + Whether to enable the cache, that is to + store the computed embedding. + """ + super().__init__( + embedding_size=embedding_size, + relu_bias=relu_bias, + epochs=epochs, + learning_rate=learning_rate, + learning_rate_decay=learning_rate_decay, + node_embedding_path=node_embedding_path, + random_state=random_state, + verbose=verbose, + enable_cache=enable_cache, + ) + + def _fit_transform( + self, + graph: Graph, + return_dataframe: bool = True, + ) -> EmbeddingResult: + """Return node embedding.""" + node_embedding = self._model.fit_transform( + graph, + ) + if return_dataframe: + node_embedding = pd.DataFrame( + node_embedding, + index=graph.get_node_names() + ) + + return EmbeddingResult( + embedding_method_name=self.model_name(), + node_embeddings=node_embedding, + ) + + @classmethod + def model_name(cls) -> str: + """Returns name of the model.""" + return "Unstructured" + + @classmethod + def can_use_edge_types(cls) -> bool: + return False \ No newline at end of file diff --git a/embiggen/embedders/ensmallen_embedders/walklets.py b/embiggen/embedders/ensmallen_embedders/walklets.py new file mode 100644 index 00000000..0f3a4ddd --- /dev/null +++ b/embiggen/embedders/ensmallen_embedders/walklets.py @@ -0,0 +1,142 @@ +"""Module providing WalkletsEnsmallen model implementation.""" +from typing import Optional, Dict, Any +from embiggen.embedders.ensmallen_embedders.node2vec import Node2VecEnsmallen +from embiggen.utils.abstract_models.abstract_model import abstract_class + +@abstract_class +class WalkletsEnsmallen(Node2VecEnsmallen): + """Class providing WalkletsEnsmallen implemeted in Rust from Ensmallen.""" + + def __init__( + self, + embedding_size: int = 100, + epochs: int = 10, + clipping_value: float = 6.0, + number_of_negative_samples: int = 10, + walk_length: int = 128, + iterations: int = 10, + window_size: int = 4, + return_weight: float = 1.0, + explore_weight: float = 1.0, + max_neighbours: Optional[int] = 100, + learning_rate: float = 0.01, + learning_rate_decay: float = 0.9, + central_nodes_embedding_path: Optional[str] = None, + contextual_nodes_embedding_path: Optional[str] = None, + alpha: float = 0.75, + normalize_by_degree: bool = False, + stochastic_downsample_by_degree: Optional[bool] = False, + normalize_learning_rate_by_degree: Optional[bool] = False, + use_scale_free_distribution: Optional[bool] = True, + random_state: int = 42, + enable_cache: bool = False + ): + """Create new abstract Node2Vec method. + + Parameters + -------------------------- + embedding_size: int = 100 + Dimension of the embedding. + epochs: int = 30 + Number of epochs to train the model for. + clipping_value: float = 6.0 + Value at which we clip the dot product, mostly for numerical stability issues. + By default, `6.0`, where the loss is already close to zero. + number_of_negative_samples: int = 10 + The number of negative classes to randomly sample per batch. + This single sample of negative classes is evaluated for each element in the batch. + walk_length: int = 128 + Maximal length of the walks. + iterations: int = 10 + Number of iterations of the single walks. + window_size: int = 4 + Window size for the local context. + On the borders the window size is trimmed. + return_weight: float = 1.0 + Weight on the probability of returning to the same node the walk just came from + Having this higher tends the walks to be + more like a Breadth-First Search. + Having this very high (> 2) makes search very local. + Equal to the inverse of p in the Node2Vec paper. + explore_weight: float = 1.0 + Weight on the probability of visiting a neighbor node + to the one we're coming from in the random walk + Having this higher tends the walks to be + more like a Depth-First Search. + Having this very high makes search more outward. + Having this very low makes search very local. + Equal to the inverse of q in the Node2Vec paper. + max_neighbours: Optional[int] = 100 + Number of maximum neighbours to consider when using approximated walks. + By default, None, we execute exact random walks. + This is mainly useful for graphs containing nodes with high degrees. + learning_rate: float = 0.01 + The learning rate to use to train the Node2Vec model. By default 0.01. + learning_rate_decay: float = 0.9 + Factor to reduce the learning rate for at each epoch. By default 0.9. + central_nodes_embedding_path: Optional[str] = None + Path where to mmap and store the central nodes embedding. + If provided, we expect the path to contain the substring `{window_size}` which + will be replaced with the i-th window size embedding that is being computed. + This is necessary to embed large graphs whose embedding will not + fit into the available main memory. + contextual_nodes_embedding_path: Optional[str] = None + Path where to mmap and store the central nodes embedding. + If provided, we expect the path to contain the substring `{window_size}` which + will be replaced with the i-th window size embedding that is being computed. + This is necessary to embed large graphs whose embedding will not + fit into the available main memory. + alpha: float = 0.75 + Alpha parameter for GloVe's loss. + normalize_by_degree: bool = False + Whether to normalize the random walk by the node degree + of the destination node degrees. + stochastic_downsample_by_degree: Optional[bool] = False + Randomly skip samples with probability proportional to the degree of the central node. By default false. + normalize_learning_rate_by_degree: Optional[bool] = False + Divide the learning rate by the degree of the central node. By default false. + use_scale_free_distribution: Optional[bool] = True + Sample negatives proportionally to their degree. By default true. + random_state: int = 42 + The random state to reproduce the training sequence. + enable_cache: bool = False + Whether to enable the cache, that is to + store the computed embedding. + """ + super().__init__( + embedding_size=embedding_size // window_size, + epochs=epochs, + clipping_value=clipping_value, + number_of_negative_samples=number_of_negative_samples, + walk_length=walk_length, + iterations=iterations, + window_size=window_size, + return_weight=return_weight, + explore_weight=explore_weight, + max_neighbours=max_neighbours, + learning_rate=learning_rate, + learning_rate_decay=learning_rate_decay, + central_nodes_embedding_path=central_nodes_embedding_path, + contextual_nodes_embedding_path=contextual_nodes_embedding_path, + alpha=alpha, + normalize_by_degree=normalize_by_degree, + stochastic_downsample_by_degree=stochastic_downsample_by_degree, + normalize_learning_rate_by_degree=normalize_learning_rate_by_degree, + use_scale_free_distribution=use_scale_free_distribution, + random_state=random_state, + enable_cache=enable_cache + ) + + def parameters(self) -> Dict[str, Any]: + """Returns parameters of the model.""" + parameters = super().parameters() + parameters["embedding_size"] = parameters["embedding_size"] * parameters["window_size"] + return parameters + + @classmethod + def requires_node_types(cls) -> bool: + return False + + @classmethod + def requires_edge_types(cls) -> bool: + return False \ No newline at end of file diff --git a/embiggen/embedders/ensmallen_embedders/walklets_cbow.py b/embiggen/embedders/ensmallen_embedders/walklets_cbow.py index 51e8ccfe..99d73975 100644 --- a/embiggen/embedders/ensmallen_embedders/walklets_cbow.py +++ b/embiggen/embedders/ensmallen_embedders/walklets_cbow.py @@ -1,31 +1,31 @@ """Module providing WalkletsCBOW model implementation.""" -from typing import Optional -from embiggen.embedders.ensmallen_embedders.node2vec import Node2VecEnsmallen +from typing import Optional, Dict, Any +from embiggen.embedders.ensmallen_embedders.walklets import WalkletsEnsmallen -class WalkletsCBOWEnsmallen(Node2VecEnsmallen): +class WalkletsCBOWEnsmallen(WalkletsEnsmallen): """Class providing WalkletsCBOW implemeted in Rust from Ensmallen.""" def __init__( self, embedding_size: int = 100, - epochs: int = 10, + epochs: int = 30, clipping_value: float = 6.0, number_of_negative_samples: int = 10, walk_length: int = 128, iterations: int = 10, - window_size: int = 10, + window_size: int = 4, return_weight: float = 1.0, explore_weight: float = 1.0, - change_node_type_weight: float = 1.0, - change_edge_type_weight: float = 1.0, max_neighbours: Optional[int] = 100, learning_rate: float = 0.01, learning_rate_decay: float = 0.9, + central_nodes_embedding_path: Optional[str] = None, + contextual_nodes_embedding_path: Optional[str] = None, normalize_by_degree: bool = False, stochastic_downsample_by_degree: Optional[bool] = False, normalize_learning_rate_by_degree: Optional[bool] = False, - use_zipfian_sampling: Optional[bool] = True, + use_scale_free_distribution: Optional[bool] = True, random_state: int = 42, enable_cache: bool = False ): @@ -35,7 +35,7 @@ def __init__( -------------------------- embedding_size: int = 100 Dimension of the embedding. - epochs: int = 10 + epochs: int = 30 Number of epochs to train the model for. clipping_value: float = 6.0 Value at which we clip the dot product, mostly for numerical stability issues. @@ -47,7 +47,7 @@ def __init__( Maximal length of the walks. iterations: int = 10 Number of iterations of the single walks. - window_size: int = 10 + window_size: int = 4 Window size for the local context. On the borders the window size is trimmed. return_weight: float = 1.0 @@ -64,14 +64,6 @@ def __init__( Having this very high makes search more outward. Having this very low makes search very local. Equal to the inverse of q in the Node2Vec paper. - change_node_type_weight: float = 1.0 - Weight on the probability of visiting a neighbor node of a - different type than the previous node. This only applies to - colored graphs, otherwise it has no impact. - change_edge_type_weight: float = 1.0 - Weight on the probability of visiting a neighbor edge of a - different type than the previous edge. This only applies to - multigraphs, otherwise it has no impact. max_neighbours: Optional[int] = 100 Number of maximum neighbours to consider when using approximated walks. By default, None, we execute exact random walks. @@ -80,6 +72,18 @@ def __init__( The learning rate to use to train the Node2Vec model. By default 0.01. learning_rate_decay: float = 0.9 Factor to reduce the learning rate for at each epoch. By default 0.9. + central_nodes_embedding_path: Optional[str] = None + Path where to mmap and store the central nodes embedding. + If provided, we expect the path to contain the substring `{window_size}` which + will be replaced with the i-th window size embedding that is being computed. + This is necessary to embed large graphs whose embedding will not + fit into the available main memory. + contextual_nodes_embedding_path: Optional[str] = None + Path where to mmap and store the central nodes embedding. + If provided, we expect the path to contain the substring `{window_size}` which + will be replaced with the i-th window size embedding that is being computed. + This is necessary to embed large graphs whose embedding will not + fit into the available main memory. normalize_by_degree: bool = False Whether to normalize the random walk by the node degree of the destination node degrees. @@ -87,7 +91,7 @@ def __init__( Randomly skip samples with probability proportional to the degree of the central node. By default false. normalize_learning_rate_by_degree: Optional[bool] = False Divide the learning rate by the degree of the central node. By default false. - use_zipfian_sampling: Optional[bool] = True + use_scale_free_distribution: Optional[bool] = True Sample negatives proportionally to their degree. By default true. random_state: int = 42 The random state to reproduce the training sequence. @@ -96,8 +100,7 @@ def __init__( store the computed embedding. """ super().__init__( - model_name="WalkletsCBOW", - embedding_size=embedding_size // window_size, + embedding_size=embedding_size, epochs=epochs, clipping_value=clipping_value, number_of_negative_samples=number_of_negative_samples, @@ -106,28 +109,33 @@ def __init__( window_size=window_size, return_weight=return_weight, explore_weight=explore_weight, - change_edge_type_weight=change_edge_type_weight, - change_node_type_weight=change_node_type_weight, max_neighbours=max_neighbours, learning_rate=learning_rate, learning_rate_decay=learning_rate_decay, + central_nodes_embedding_path=central_nodes_embedding_path, + contextual_nodes_embedding_path=contextual_nodes_embedding_path, normalize_by_degree=normalize_by_degree, stochastic_downsample_by_degree=stochastic_downsample_by_degree, normalize_learning_rate_by_degree=normalize_learning_rate_by_degree, - use_zipfian_sampling=use_zipfian_sampling, + use_scale_free_distribution=use_scale_free_distribution, random_state=random_state, enable_cache=enable_cache ) - + + def parameters(self) -> Dict[str, Any]: + """Returns parameters for smoke test.""" + removed = [ + "alpha", + ] + return dict( + **{ + key: value + for key, value in super().parameters().items() + if key not in removed + } + ) + @classmethod def model_name(cls) -> str: """Returns name of the model.""" return "Walklets CBOW" - - @classmethod - def requires_node_types(cls) -> bool: - return False - - @classmethod - def requires_edge_types(cls) -> bool: - return False \ No newline at end of file diff --git a/embiggen/embedders/ensmallen_embedders/walklets_glove.py b/embiggen/embedders/ensmallen_embedders/walklets_glove.py new file mode 100644 index 00000000..64fb7f4b --- /dev/null +++ b/embiggen/embedders/ensmallen_embedders/walklets_glove.py @@ -0,0 +1,136 @@ +"""Module providing WalkletsGloVe model implementation.""" +from typing import Optional, Dict, Any +from embiggen.embedders.ensmallen_embedders.walklets import WalkletsEnsmallen + + +class WalkletsGloVeEnsmallen(WalkletsEnsmallen): + """Class providing WalkletsGloVe implemeted in Rust from Ensmallen.""" + + def __init__( + self, + embedding_size: int = 100, + epochs: int = 30, + walk_length: int = 128, + iterations: int = 10, + window_size: int = 4, + return_weight: float = 1.0, + explore_weight: float = 1.0, + max_neighbours: Optional[int] = 100, + learning_rate: float = 0.001, + learning_rate_decay: float = 0.9, + central_nodes_embedding_path: Optional[str] = None, + contextual_nodes_embedding_path: Optional[str] = None, + alpha: float = 0.75, + normalize_by_degree: bool = False, + stochastic_downsample_by_degree: Optional[bool] = False, + normalize_learning_rate_by_degree: Optional[bool] = False, + use_scale_free_distribution: Optional[bool] = True, + random_state: int = 42, + enable_cache: bool = False + ): + """Create new abstract Node2Vec method. + + Parameters + -------------------------- + embedding_size: int = 100 + Dimension of the embedding. + epochs: int = 30 + Number of epochs to train the model for. + walk_length: int = 128 + Maximal length of the walks. + iterations: int = 10 + Number of iterations of the single walks. + window_size: int = 4 + Window size for the local context. + On the borders the window size is trimmed. + return_weight: float = 1.0 + Weight on the probability of returning to the same node the walk just came from + Having this higher tends the walks to be + more like a Breadth-First Search. + Having this very high (> 2) makes search very local. + Equal to the inverse of p in the Node2Vec paper. + explore_weight: float = 1.0 + Weight on the probability of visiting a neighbor node + to the one we're coming from in the random walk + Having this higher tends the walks to be + more like a Depth-First Search. + Having this very high makes search more outward. + Having this very low makes search very local. + Equal to the inverse of q in the Node2Vec paper. + max_neighbours: Optional[int] = 100 + Number of maximum neighbours to consider when using approximated walks. + By default, None, we execute exact random walks. + This is mainly useful for graphs containing nodes with high degrees. + learning_rate: float = 0.001 + The learning rate to use to train the Node2Vec model. By default 0.01. + learning_rate_decay: float = 0.9 + Factor to reduce the learning rate for at each epoch. By default 0.9. + central_nodes_embedding_path: Optional[str] = None + Path where to mmap and store the central nodes embedding. + If provided, we expect the path to contain the substring `{window_size}` which + will be replaced with the i-th window size embedding that is being computed. + This is necessary to embed large graphs whose embedding will not + fit into the available main memory. + contextual_nodes_embedding_path: Optional[str] = None + Path where to mmap and store the central nodes embedding. + If provided, we expect the path to contain the substring `{window_size}` which + will be replaced with the i-th window size embedding that is being computed. + This is necessary to embed large graphs whose embedding will not + fit into the available main memory. + alpha: float = 0.75 + Alpha parameter for GloVe's loss. + normalize_by_degree: bool = False + Whether to normalize the random walk by the node degree + of the destination node degrees. + stochastic_downsample_by_degree: Optional[bool] = False + Randomly skip samples with probability proportional to the degree of the central node. By default false. + normalize_learning_rate_by_degree: Optional[bool] = False + Divide the learning rate by the degree of the central node. By default false. + use_scale_free_distribution: Optional[bool] = True + Sample negatives proportionally to their degree. By default true. + random_state: int = 42 + The random state to reproduce the training sequence. + enable_cache: bool = False + Whether to enable the cache, that is to + store the computed embedding. + """ + super().__init__( + embedding_size=embedding_size, + epochs=epochs, + alpha=alpha, + walk_length=walk_length, + iterations=iterations, + window_size=window_size, + return_weight=return_weight, + explore_weight=explore_weight, + max_neighbours=max_neighbours, + learning_rate=learning_rate, + learning_rate_decay=learning_rate_decay, + central_nodes_embedding_path=central_nodes_embedding_path, + contextual_nodes_embedding_path=contextual_nodes_embedding_path, + normalize_by_degree=normalize_by_degree, + stochastic_downsample_by_degree=stochastic_downsample_by_degree, + normalize_learning_rate_by_degree=normalize_learning_rate_by_degree, + use_scale_free_distribution=use_scale_free_distribution, + random_state=random_state, + enable_cache=enable_cache + ) + + @classmethod + def model_name(cls) -> str: + """Returns name of the model.""" + return "Walklets GloVe" + + def parameters(self) -> Dict[str, Any]: + """Returns parameters for smoke test.""" + removed = [ + "number_of_negative_samples", + "clipping_value" + ] + return dict( + **{ + key: value + for key, value in super().parameters().items() + if key not in removed + } + ) \ No newline at end of file diff --git a/embiggen/embedders/ensmallen_embedders/walklets_skipgram.py b/embiggen/embedders/ensmallen_embedders/walklets_skipgram.py index 2cd9bedf..73ce0c94 100644 --- a/embiggen/embedders/ensmallen_embedders/walklets_skipgram.py +++ b/embiggen/embedders/ensmallen_embedders/walklets_skipgram.py @@ -1,31 +1,31 @@ """Module providing WalkletsSkipGram model implementation.""" -from typing import Optional -from embiggen.embedders.ensmallen_embedders.node2vec import Node2VecEnsmallen +from typing import Optional, Dict, Any +from embiggen.embedders.ensmallen_embedders.walklets import WalkletsEnsmallen -class WalkletsSkipGramEnsmallen(Node2VecEnsmallen): +class WalkletsSkipGramEnsmallen(WalkletsEnsmallen): """Class providing WalkletsSkipGram implemeted in Rust from Ensmallen.""" def __init__( self, embedding_size: int = 100, - epochs: int = 10, + epochs: int = 30, clipping_value: float = 6.0, number_of_negative_samples: int = 10, walk_length: int = 128, iterations: int = 10, - window_size: int = 10, + window_size: int = 4, return_weight: float = 1.0, explore_weight: float = 1.0, - change_node_type_weight: float = 1.0, - change_edge_type_weight: float = 1.0, max_neighbours: Optional[int] = 100, learning_rate: float = 0.01, learning_rate_decay: float = 0.9, + central_nodes_embedding_path: Optional[str] = None, + contextual_nodes_embedding_path: Optional[str] = None, normalize_by_degree: bool = False, stochastic_downsample_by_degree: Optional[bool] = False, normalize_learning_rate_by_degree: Optional[bool] = False, - use_zipfian_sampling: Optional[bool] = True, + use_scale_free_distribution: Optional[bool] = True, random_state: int = 42, enable_cache: bool = False ): @@ -35,7 +35,7 @@ def __init__( -------------------------- embedding_size: int = 100 Dimension of the embedding. - epochs: int = 10 + epochs: int = 30 Number of epochs to train the model for. clipping_value: float = 6.0 Value at which we clip the dot product, mostly for numerical stability issues. @@ -47,7 +47,7 @@ def __init__( Maximal length of the walks. iterations: int = 10 Number of iterations of the single walks. - window_size: int = 10 + window_size: int = 4 Window size for the local context. On the borders the window size is trimmed. return_weight: float = 1.0 @@ -64,14 +64,6 @@ def __init__( Having this very high makes search more outward. Having this very low makes search very local. Equal to the inverse of q in the Node2Vec paper. - change_node_type_weight: float = 1.0 - Weight on the probability of visiting a neighbor node of a - different type than the previous node. This only applies to - colored graphs, otherwise it has no impact. - change_edge_type_weight: float = 1.0 - Weight on the probability of visiting a neighbor edge of a - different type than the previous edge. This only applies to - multigraphs, otherwise it has no impact. max_neighbours: Optional[int] = 100 Number of maximum neighbours to consider when using approximated walks. By default, None, we execute exact random walks. @@ -80,6 +72,18 @@ def __init__( The learning rate to use to train the Node2Vec model. By default 0.01. learning_rate_decay: float = 0.9 Factor to reduce the learning rate for at each epoch. By default 0.9. + central_nodes_embedding_path: Optional[str] = None + Path where to mmap and store the central nodes embedding. + If provided, we expect the path to contain the substring `{window_size}` which + will be replaced with the i-th window size embedding that is being computed. + This is necessary to embed large graphs whose embedding will not + fit into the available main memory. + contextual_nodes_embedding_path: Optional[str] = None + Path where to mmap and store the central nodes embedding. + If provided, we expect the path to contain the substring `{window_size}` which + will be replaced with the i-th window size embedding that is being computed. + This is necessary to embed large graphs whose embedding will not + fit into the available main memory. normalize_by_degree: bool = False Whether to normalize the random walk by the node degree of the destination node degrees. @@ -87,7 +91,7 @@ def __init__( Randomly skip samples with probability proportional to the degree of the central node. By default false. normalize_learning_rate_by_degree: Optional[bool] = False Divide the learning rate by the degree of the central node. By default false. - use_zipfian_sampling: Optional[bool] = True + use_scale_free_distribution: Optional[bool] = True Sample negatives proportionally to their degree. By default true. random_state: int = 42 The random state to reproduce the training sequence. @@ -96,8 +100,7 @@ def __init__( store the computed embedding. """ super().__init__( - model_name="WalkletsSkipGram", - embedding_size=embedding_size // window_size, + embedding_size=embedding_size, epochs=epochs, clipping_value=clipping_value, number_of_negative_samples=number_of_negative_samples, @@ -106,28 +109,33 @@ def __init__( window_size=window_size, return_weight=return_weight, explore_weight=explore_weight, - change_edge_type_weight=change_edge_type_weight, - change_node_type_weight=change_node_type_weight, max_neighbours=max_neighbours, learning_rate=learning_rate, learning_rate_decay=learning_rate_decay, + central_nodes_embedding_path=central_nodes_embedding_path, + contextual_nodes_embedding_path=contextual_nodes_embedding_path, normalize_by_degree=normalize_by_degree, stochastic_downsample_by_degree=stochastic_downsample_by_degree, normalize_learning_rate_by_degree=normalize_learning_rate_by_degree, - use_zipfian_sampling=use_zipfian_sampling, + use_scale_free_distribution=use_scale_free_distribution, random_state=random_state, enable_cache=enable_cache ) - + + def parameters(self) -> Dict[str, Any]: + """Returns parameters for smoke test.""" + removed = [ + "alpha", + ] + return dict( + **{ + key: value + for key, value in super().parameters().items() + if key not in removed + } + ) + @classmethod def model_name(cls) -> str: """Returns name of the model.""" return "Walklets SkipGram" - - @classmethod - def requires_node_types(cls) -> bool: - return False - - @classmethod - def requires_edge_types(cls) -> bool: - return False \ No newline at end of file diff --git a/embiggen/embedders/ensmallen_embedders/weighted_spine.py b/embiggen/embedders/ensmallen_embedders/weighted_spine.py index c7ad70e8..8cb06df0 100644 --- a/embiggen/embedders/ensmallen_embedders/weighted_spine.py +++ b/embiggen/embedders/ensmallen_embedders/weighted_spine.py @@ -3,16 +3,18 @@ from ensmallen import Graph import pandas as pd from ensmallen import models -from embiggen.utils.abstract_models import AbstractEmbeddingModel, EmbeddingResult +from embiggen.embedders.ensmallen_embedders.ensmallen_embedder import EnsmallenEmbedder +from embiggen.utils import EmbeddingResult -class WeightedSPINE(AbstractEmbeddingModel): +class WeightedSPINE(EnsmallenEmbedder): """Abstract class for Node2Vec algorithms.""" def __init__( self, embedding_size: int = 100, use_edge_weights_as_probabilities: bool = False, + verbose: bool = False, enable_cache: bool = False ): """Create new abstract Node2Vec method. @@ -23,10 +25,13 @@ def __init__( Dimension of the embedding. use_edge_weights_as_probabilities: bool = False Whether to treat the weights as probabilities. + verbose: bool = False + Whether to show loading bars. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. """ + self._verbose = verbose self._model = models.WeightedSPINE( embedding_size=embedding_size, use_edge_weights_as_probabilities=use_edge_weights_as_probabilities, @@ -48,12 +53,11 @@ def _fit_transform( self, graph: Graph, return_dataframe: bool = True, - verbose: bool = True ) -> EmbeddingResult: """Return node embedding.""" node_embedding = self._model.fit_transform( graph, - verbose=verbose, + verbose=self._verbose, ).T if return_dataframe: node_embedding = pd.DataFrame( @@ -65,27 +69,11 @@ def _fit_transform( node_embeddings=node_embedding ) - @classmethod - def task_name(cls) -> str: - return "Node Embedding" - @classmethod def model_name(cls) -> str: """Returns name of the model.""" return "WeightedSPINE" - @classmethod - def library_name(cls) -> str: - return "Ensmallen" - - @classmethod - def requires_nodes_sorted_by_decreasing_node_degree(cls) -> bool: - return False - - @classmethod - def is_topological(cls) -> bool: - return True - @classmethod def requires_edge_weights(cls) -> bool: return True diff --git a/embiggen/embedders/graph_embedding_pipeline.py b/embiggen/embedders/graph_embedding_pipeline.py index d1c9ba42..34713828 100644 --- a/embiggen/embedders/graph_embedding_pipeline.py +++ b/embiggen/embedders/graph_embedding_pipeline.py @@ -15,7 +15,6 @@ def embed_graph( library_name: Optional[str] = None, smoke_test: bool = False, return_dataframe: bool = True, - verbose: bool = True, **kwargs: Dict ) -> EmbeddingResult: """Return embedding of the provided graph. @@ -45,14 +44,11 @@ def embed_graph( Path where to store the cache if it is enabled. return_dataframe: bool = True Whether to return a pandas DataFrame with the embedding. - verbose: bool - Whether to show loading bars. **kwargs: Dict Kwargs to forward to the embedding model creation. If a model name was NOT provided, an exception will be raised as it is unclear how to behave. """ - graph = next(iterate_graphs( graphs=graph, repositories=repository, @@ -64,7 +60,7 @@ def embed_graph( model_name=embedding_model, library_name=library_name )(**kwargs) - elif kwargs is not None: + elif kwargs: raise ValueError( "Please be advised that even though you have provided yourself " "the embedding model, you have also provided the kwargs which " @@ -80,9 +76,7 @@ def embed_graph( if smoke_test: try: - embedding_model = embedding_model.__class__( - **embedding_model.smoke_test_parameters() - ) + embedding_model = embedding_model.into_smoke_test() except Exception as e: raise ValueError( "An exception was raised while trying to create " @@ -100,7 +94,6 @@ def embed_graph( return embedding_model.fit_transform( graph, return_dataframe=return_dataframe, - verbose=verbose ) except Exception as e: raise ValueError( @@ -109,5 +102,5 @@ def embed_graph( f"using the model called {embedding_model.model_name()} " f"from the library {library_name}, specifically " f"implemented in the class {embedding_model.__class__.__name__}. " - f"The body of the exception was: {str(e)}." + f"The body of the exception was: {str(e)}" ) from e \ No newline at end of file diff --git a/embiggen/embedders/karateclub_embedders/boostne.py b/embiggen/embedders/karateclub_embedders/boostne.py index 3cca9296..461fee04 100644 --- a/embiggen/embedders/karateclub_embedders/boostne.py +++ b/embiggen/embedders/karateclub_embedders/boostne.py @@ -8,7 +8,7 @@ class BoostNEKarateClub(AbstractKarateClubEmbedder): def __init__( self, - embedding_size: int = 128, + embedding_size: int = 100, iterations: int = 16, order: int = 2, alpha: float = 0.01, @@ -19,7 +19,7 @@ def __init__( Parameters ---------------------- - embedding_size: int = 128 + embedding_size: int = 100 Size of the embedding to use. iterations: int = 16 Number of boosting iterations. Default is 16. diff --git a/embiggen/embedders/karateclub_embedders/deep_walk.py b/embiggen/embedders/karateclub_embedders/deep_walk.py index 36f858d0..db084890 100644 --- a/embiggen/embedders/karateclub_embedders/deep_walk.py +++ b/embiggen/embedders/karateclub_embedders/deep_walk.py @@ -5,11 +5,11 @@ from embiggen.embedders.karateclub_embedders.abstract_karateclub_embedder import AbstractKarateClubEmbedder -class DeepWalkKarateClub(AbstractKarateClubEmbedder): +class DeepWalkSkipGramKarateClub(AbstractKarateClubEmbedder): def __init__( self, - embedding_size: int = 128, + embedding_size: int = 100, walk_number: int = 10, walk_length: int = 80, window_size: int = 5, @@ -23,7 +23,7 @@ def __init__( Parameters ---------------------- - embedding_size: int = 128 + embedding_size: int = 100 Size of the embedding to use. walk_number: int = 10 Number of random walks. Default is 10. @@ -97,7 +97,7 @@ def _build_model(self) -> DeepWalk: @classmethod def model_name(cls) -> str: """Returns name of the model""" - return "DeepWalk" + return "DeepWalk SkipGram" @classmethod def requires_nodes_sorted_by_decreasing_node_degree(cls) -> bool: diff --git a/embiggen/embedders/karateclub_embedders/geometric_laplacian_eigenmaps.py b/embiggen/embedders/karateclub_embedders/geometric_laplacian_eigenmaps.py index 69793eb3..eb6982a3 100644 --- a/embiggen/embedders/karateclub_embedders/geometric_laplacian_eigenmaps.py +++ b/embiggen/embedders/karateclub_embedders/geometric_laplacian_eigenmaps.py @@ -8,7 +8,7 @@ class GLEEKarateClub(AbstractKarateClubEmbedder): def __init__( self, - embedding_size: int = 128, + embedding_size: int = 100, random_state: int = 42, enable_cache: bool = False ): @@ -16,7 +16,7 @@ def __init__( Parameters ---------------------- - embedding_size: int = 128 + embedding_size: int = 100 Size of the embedding to use. random_state: int = 42 Random state to use for the stocastic diff --git a/embiggen/embedders/karateclub_embedders/grarep.py b/embiggen/embedders/karateclub_embedders/grarep.py index 052510f0..1c67feeb 100644 --- a/embiggen/embedders/karateclub_embedders/grarep.py +++ b/embiggen/embedders/karateclub_embedders/grarep.py @@ -8,7 +8,7 @@ class GraRepKarateClub(AbstractKarateClubEmbedder): def __init__( self, - embedding_size: int = 128, + embedding_size: int = 100, iteration: int = 10, order: int = 5, random_state: int = 42, @@ -18,7 +18,7 @@ def __init__( Parameters ---------------------- - embedding_size: int = 128 + embedding_size: int = 100 Size of the embedding to use. iteration: int = 10 Number of SVD iterations. Default is 10. diff --git a/embiggen/embedders/karateclub_embedders/hope.py b/embiggen/embedders/karateclub_embedders/hope.py index c3ef948d..ef74dad4 100644 --- a/embiggen/embedders/karateclub_embedders/hope.py +++ b/embiggen/embedders/karateclub_embedders/hope.py @@ -8,7 +8,7 @@ class HOPEKarateClub(AbstractKarateClubEmbedder): def __init__( self, - embedding_size: int = 128, + embedding_size: int = 100, random_state: int = 42, enable_cache: bool = False ): @@ -16,7 +16,7 @@ def __init__( Parameters ---------------------- - embedding_size: int = 128 + embedding_size: int = 100 Size of the embedding to use. random_state: int = 42 Random state to use for the stocastic diff --git a/embiggen/embedders/karateclub_embedders/laplacian_eigenmaps.py b/embiggen/embedders/karateclub_embedders/laplacian_eigenmaps.py index 5156a315..e360ee8d 100644 --- a/embiggen/embedders/karateclub_embedders/laplacian_eigenmaps.py +++ b/embiggen/embedders/karateclub_embedders/laplacian_eigenmaps.py @@ -8,7 +8,7 @@ class LaplacianEigenmapsKarateClub(AbstractKarateClubEmbedder): def __init__( self, - embedding_size: int = 128, + embedding_size: int = 100, random_state: int = 42, enable_cache: bool = False ): @@ -16,7 +16,7 @@ def __init__( Parameters ---------------------- - embedding_size: int = 128 + embedding_size: int = 100 Size of the embedding to use. random_state: int = 42 Random state to use for the stocastic diff --git a/embiggen/embedders/karateclub_embedders/netmf.py b/embiggen/embedders/karateclub_embedders/netmf.py index adb714fe..86ae41c8 100644 --- a/embiggen/embedders/karateclub_embedders/netmf.py +++ b/embiggen/embedders/karateclub_embedders/netmf.py @@ -8,7 +8,7 @@ class NetMFKarateClub(AbstractKarateClubEmbedder): def __init__( self, - embedding_size: int = 128, + embedding_size: int = 100, iteration: int = 10, order: int = 2, negative_samples: int = 1, @@ -19,7 +19,7 @@ def __init__( Parameters ---------------------- - embedding_size: int = 128 + embedding_size: int = 100 Size of the embedding to use. iteration: int = 10 Number of SVD iterations. Default is 10. diff --git a/embiggen/embedders/karateclub_embedders/nmfadmm.py b/embiggen/embedders/karateclub_embedders/nmfadmm.py index bf4c6b80..8026e687 100644 --- a/embiggen/embedders/karateclub_embedders/nmfadmm.py +++ b/embiggen/embedders/karateclub_embedders/nmfadmm.py @@ -8,7 +8,7 @@ class NMFADMMKarateClub(AbstractKarateClubEmbedder): def __init__( self, - embedding_size: int = 128, + embedding_size: int = 100, iterations: int = 100, rho: float = 1.0, random_state: int = 42, @@ -18,7 +18,7 @@ def __init__( Parameters ---------------------- - embedding_size: int = 128 + embedding_size: int = 100 Size of the embedding to use. iterations: int = 100 Number of SVD iterationss. Default is 10. diff --git a/embiggen/embedders/karateclub_embedders/node_sketch.py b/embiggen/embedders/karateclub_embedders/node_sketch.py index d5082873..0829eea7 100644 --- a/embiggen/embedders/karateclub_embedders/node_sketch.py +++ b/embiggen/embedders/karateclub_embedders/node_sketch.py @@ -8,7 +8,7 @@ class NodeSketchKarateClub(AbstractKarateClubEmbedder): def __init__( self, - embedding_size: int = 128, + embedding_size: int = 100, iterations: int = 10, decay: float = 0.01, random_state: int = 42, @@ -18,7 +18,7 @@ def __init__( Parameters ---------------------- - embedding_size: int = 128 + embedding_size: int = 100 Size of the embedding to use. iterations: int = 10 Number of SVD iterationss. Default is 10. diff --git a/embiggen/embedders/karateclub_embedders/randne.py b/embiggen/embedders/karateclub_embedders/randne.py index 757b3183..bee120c0 100644 --- a/embiggen/embedders/karateclub_embedders/randne.py +++ b/embiggen/embedders/karateclub_embedders/randne.py @@ -8,7 +8,7 @@ class RandNEKarateClub(AbstractKarateClubEmbedder): def __init__( self, - embedding_size: int = 128, + embedding_size: int = 100, alphas: Union[List[float], Tuple[float]] = (0.5, 0.5), random_state: int = 42, enable_cache: bool = False @@ -17,7 +17,7 @@ def __init__( Parameters ---------------------- - embedding_size: int = 128 + embedding_size: int = 100 Size of the embedding to use. alphas: Union[List[float], Tuple[float]] = (0.5, 0.5) Smoothing weights for adjacency matrix powers. Default is [0.5, 0.5]. diff --git a/embiggen/embedders/karateclub_embedders/role2vec.py b/embiggen/embedders/karateclub_embedders/role2vec.py index 866b8db4..3f6ad6a4 100644 --- a/embiggen/embedders/karateclub_embedders/role2vec.py +++ b/embiggen/embedders/karateclub_embedders/role2vec.py @@ -9,7 +9,7 @@ class Role2VecKarateClub(AbstractKarateClubEmbedder): def __init__( self, - embedding_size: int = 128, + embedding_size: int = 100, walk_number: int = 10, walk_length: int = 80, window_size: int = 5, @@ -26,7 +26,7 @@ def __init__( Parameters ---------------------- - embedding_size: int = 128 + embedding_size: int = 100 Size of the embedding to use. walk_number: int = 10 Number of random walks. Default is 10. diff --git a/embiggen/embedders/karateclub_embedders/skipgram.py b/embiggen/embedders/karateclub_embedders/skipgram.py index 2c275b66..bb9cb317 100644 --- a/embiggen/embedders/karateclub_embedders/skipgram.py +++ b/embiggen/embedders/karateclub_embedders/skipgram.py @@ -9,7 +9,7 @@ class SkipGramKarateClub(AbstractKarateClubEmbedder): def __init__( self, - embedding_size: int = 128, + embedding_size: int = 100, walk_number: int = 10, walk_length: int = 80, window_size: int = 5, @@ -33,7 +33,7 @@ def __init__( Parameters ---------------------- - embedding_size: int = 128 + embedding_size: int = 100 Size of the embedding to use. walk_number: int = 10 Number of random walks. Default is 10. @@ -117,7 +117,7 @@ def _build_model(self) -> Node2Vec: @classmethod def model_name(cls) -> str: """Returns name of the model""" - return "SkipGram" + return "Node2Vec SkipGram" @classmethod def requires_nodes_sorted_by_decreasing_node_degree(cls) -> bool: diff --git a/embiggen/embedders/karateclub_embedders/sociodim.py b/embiggen/embedders/karateclub_embedders/sociodim.py index 97f3c730..4ccac64f 100644 --- a/embiggen/embedders/karateclub_embedders/sociodim.py +++ b/embiggen/embedders/karateclub_embedders/sociodim.py @@ -1,5 +1,4 @@ """Wrapper for SocioDim model provided from the Karate Club package.""" -from typing import Dict, Any from karateclub.node_embedding import SocioDim from embiggen.embedders.karateclub_embedders.abstract_karateclub_embedder import AbstractKarateClubEmbedder @@ -8,7 +7,7 @@ class SocioDimKarateClub(AbstractKarateClubEmbedder): def __init__( self, - embedding_size: int = 128, + embedding_size: int = 100, random_state: int = 42, enable_cache: bool = False ): @@ -16,7 +15,7 @@ def __init__( Parameters ---------------------- - embedding_size: int = 128 + embedding_size: int = 100 Size of the embedding to use. random_state: int = 42 Random state to use for the stocastic diff --git a/embiggen/embedders/karateclub_embedders/walklets.py b/embiggen/embedders/karateclub_embedders/walklets.py index 74381d48..55b681ef 100644 --- a/embiggen/embedders/karateclub_embedders/walklets.py +++ b/embiggen/embedders/karateclub_embedders/walklets.py @@ -9,7 +9,7 @@ class WalkletsSkipGramKarateClub(AbstractKarateClubEmbedder): def __init__( self, - embedding_size: int = 128, + embedding_size: int = 100, walk_number: int = 10, walk_length: int = 80, window_size: int = 5, @@ -23,7 +23,7 @@ def __init__( Parameters ---------------------- - embedding_size: int = 128 + embedding_size: int = 100 Size of the embedding to use. walk_number: int = 10 Number of random walks. Default is 10. @@ -52,16 +52,17 @@ def __init__( self._learning_rate = learning_rate self._min_count = min_count super().__init__( - embedding_size=embedding_size, + embedding_size=embedding_size // self._window_size, enable_cache=enable_cache, random_state=random_state ) def parameters(self) -> Dict[str, Any]: """Returns the parameters used in the model.""" + parameters = super().parameters() + parameters["embedding_size"] = parameters["embedding_size"] * self._window_size return dict( - **super().parameters(), - + **parameters, walk_number=self._walk_number, walk_length=self._walk_length, window_size=self._window_size, @@ -77,7 +78,7 @@ def smoke_test_parameters(cls) -> Dict[str, Any]: **AbstractKarateClubEmbedder.smoke_test_parameters(), walk_number=1, walk_length=8, - window_size=2, + window_size=1, epochs=1, ) diff --git a/embiggen/embedders/non_existent_embedders/__init__.py b/embiggen/embedders/non_existent_embedders/__init__.py new file mode 100644 index 00000000..f122946f --- /dev/null +++ b/embiggen/embedders/non_existent_embedders/__init__.py @@ -0,0 +1,8 @@ +"""Submodule to test whether the metaprogramming is working properly.""" +from embiggen.utils.abstract_models import build_init, AbstractEmbeddingModel + +build_init( + module_library_names="non_existent_module", + formatted_library_name="Non Existent Module", + expected_parent_class=AbstractEmbeddingModel +) diff --git a/embiggen/embedders/non_existent_embedders/non_existent_model.py b/embiggen/embedders/non_existent_embedders/non_existent_model.py new file mode 100644 index 00000000..46ac040d --- /dev/null +++ b/embiggen/embedders/non_existent_embedders/non_existent_model.py @@ -0,0 +1,14 @@ +"""A model strictly for validating the meta programming.""" +from embiggen.utils.abstract_models import AbstractEmbeddingModel + + +class NonExistentModel(AbstractEmbeddingModel): + """A model strictly for validating the meta programming.""" + + @classmethod + def model_name(cls) -> str: + """Returns name of the model.""" + return "Non Existent Model" + +import non_existent_module + \ No newline at end of file diff --git a/embiggen/embedders/pykeen_embedders/auto_sf.py b/embiggen/embedders/pykeen_embedders/auto_sf.py index a0afe4f4..eabe63f8 100644 --- a/embiggen/embedders/pykeen_embedders/auto_sf.py +++ b/embiggen/embedders/pykeen_embedders/auto_sf.py @@ -16,6 +16,7 @@ def __init__( batch_size: int = 2**10, training_loop: Union[str, Type[TrainingLoop] ] = "Stochastic Local Closed World Assumption", + verbose: bool = False, random_state: int = 42, enable_cache: bool = False ): @@ -46,6 +47,8 @@ def __init__( Can either be: - Stochastic Local Closed World Assumption - Local Closed World Assumption + verbose: bool = False + Whether to show loading bars. random_state: int = 42 Random seed to use while training the model enable_cache: bool = False @@ -58,17 +61,18 @@ def __init__( epochs=epochs, batch_size=batch_size, training_loop=training_loop, + verbose=verbose, random_state=random_state, enable_cache=enable_cache ) def parameters(self) -> Dict[str, Any]: - return { + return dict( **super().parameters(), **dict( num_components=self._num_components ) - } + ) @classmethod def model_name(cls) -> str: @@ -89,5 +93,6 @@ def _build_model( return AutoSF( triples_factory=triples_factory, embedding_dim=self._embedding_size, - num_components=self._num_components + num_components=self._num_components, + random_seed=self._random_state ) diff --git a/embiggen/embedders/pykeen_embedders/boxe.py b/embiggen/embedders/pykeen_embedders/boxe.py index 3ff64227..babee384 100644 --- a/embiggen/embedders/pykeen_embedders/boxe.py +++ b/embiggen/embedders/pykeen_embedders/boxe.py @@ -18,6 +18,7 @@ def __init__( batch_size: int = 2**10, training_loop: Union[str, Type[TrainingLoop] ] = "Stochastic Local Closed World Assumption", + verbose: bool = False, random_state: int = 42, enable_cache: bool = False ): @@ -54,6 +55,8 @@ def __init__( Can either be: - Stochastic Local Closed World Assumption - Local Closed World Assumption + verbose: bool = False + Whether to show loading bars. random_state: int = 42 Random seed to use while training the model enable_cache: bool = False @@ -68,19 +71,20 @@ def __init__( epochs=epochs, batch_size=batch_size, training_loop=training_loop, + verbose=verbose, random_state=random_state, enable_cache=enable_cache ) def parameters(self) -> Dict[str, Any]: - return { + return dict( **super().parameters(), **dict( tanh_map=self._tanh_map, p=self._p, power_norm=self._power_norm, ) - } + ) @classmethod def model_name(cls) -> str: @@ -104,4 +108,5 @@ def _build_model( tanh_map=self._tanh_map, p=self._p, power_norm=self._power_norm, + random_seed=self._random_state ) diff --git a/embiggen/embedders/pykeen_embedders/complex.py b/embiggen/embedders/pykeen_embedders/complex.py index 47a87819..5258ff2e 100644 --- a/embiggen/embedders/pykeen_embedders/complex.py +++ b/embiggen/embedders/pykeen_embedders/complex.py @@ -8,56 +8,6 @@ class ComplExPyKeen(EntityRelationEmbeddingModelPyKeen): - def __init__( - self, - embedding_size: int = 256, - epochs: int = 100, - batch_size: int = 2**10, - training_loop: Union[str, Type[TrainingLoop] - ] = "Stochastic Local Closed World Assumption", - random_state: int = 42, - enable_cache: bool = False - ): - """Create new PyKeen ComplEx model. - - Details - ------------------------- - This is a wrapper of the ComplEx implementation from the - PyKeen library. Please refer to the PyKeen library documentation - for details and posssible errors regarding this model. - - Parameters - ------------------------- - embedding_size: int = 256 - The dimension of the embedding to compute. - epochs: int = 100 - The number of epochs to use to train the model for. - batch_size: int = 2**10 - Size of the training batch. - device: str = "auto" - The devide to use to train the model. - Can either be cpu or cuda. - training_loop: Union[str, Type[TrainingLoop] - ] = "Stochastic Local Closed World Assumption" - The training loop to use to train the model. - Can either be: - - Stochastic Local Closed World Assumption - - Local Closed World Assumption - random_state: int = 42 - Random seed to use while training the model - enable_cache: bool = False - Whether to enable the cache, that is to - store the computed embedding. - """ - super().__init__( - embedding_size=embedding_size, - epochs=epochs, - batch_size=batch_size, - training_loop=training_loop, - random_state=random_state, - enable_cache=enable_cache - ) - @classmethod def model_name(cls) -> str: """Return name of the model.""" @@ -77,4 +27,5 @@ def _build_model( return ComplEx( triples_factory=triples_factory, embedding_dim=self._embedding_size, + random_seed=self._random_state ) diff --git a/embiggen/embedders/pykeen_embedders/conve.py b/embiggen/embedders/pykeen_embedders/conve.py index 1cf0884a..e6f389d7 100644 --- a/embiggen/embedders/pykeen_embedders/conve.py +++ b/embiggen/embedders/pykeen_embedders/conve.py @@ -25,6 +25,7 @@ def __init__( batch_size: int = 2**10, training_loop: Union[str, Type[TrainingLoop] ] = "Stochastic Local Closed World Assumption", + verbose: bool = False, random_state: int = 42, enable_cache: bool = False ): @@ -76,6 +77,8 @@ def __init__( Can either be: - Stochastic Local Closed World Assumption - Local Closed World Assumption + verbose: bool = False + Whether to show loading bars. random_state: int = 42 Random seed to use while training the model enable_cache: bool = False @@ -97,12 +100,13 @@ def __init__( epochs=epochs, batch_size=batch_size, training_loop=training_loop, + verbose=verbose, random_state=random_state, enable_cache=enable_cache ) def parameters(self) -> Dict[str, Any]: - return { + return dict( **super().parameters(), **dict( input_channels=self._input_channels, @@ -116,7 +120,7 @@ def parameters(self) -> Dict[str, Any]: feature_map_dropout=self._feature_map_dropout, apply_batch_normalization=self._apply_batch_normalization, ) - } + ) @classmethod def model_name(cls) -> str: @@ -147,4 +151,5 @@ def _build_model( output_dropout=self._output_dropout, feature_map_dropout=self._feature_map_dropout, apply_batch_normalization=self._apply_batch_normalization, + random_seed=self._random_state ) diff --git a/embiggen/embedders/pykeen_embedders/cp.py b/embiggen/embedders/pykeen_embedders/cp.py index 7933924b..885eb52a 100644 --- a/embiggen/embedders/pykeen_embedders/cp.py +++ b/embiggen/embedders/pykeen_embedders/cp.py @@ -16,6 +16,7 @@ def __init__( batch_size: int = 2**10, training_loop: Union[str, Type[TrainingLoop] ] = "Stochastic Local Closed World Assumption", + verbose: bool = False, random_state: int = 42, enable_cache: bool = False ): @@ -46,6 +47,8 @@ def __init__( Can either be: - Stochastic Local Closed World Assumption - Local Closed World Assumption + verbose: bool = False + Whether to show loading bars. random_state: int = 42 Random seed to use while training the model enable_cache: bool = False @@ -58,17 +61,18 @@ def __init__( epochs=epochs, batch_size=batch_size, training_loop=training_loop, + verbose=verbose, random_state=random_state, enable_cache=enable_cache ) def parameters(self) -> Dict[str, Any]: - return { + return dict( **super().parameters(), **dict( rank=self._rank, ) - } + ) @classmethod def model_name(cls) -> str: @@ -90,4 +94,5 @@ def _build_model( triples_factory=triples_factory, embedding_dim=self._embedding_size, rank=self._rank, + random_seed=self._random_state ) diff --git a/embiggen/embedders/pykeen_embedders/crosse.py b/embiggen/embedders/pykeen_embedders/crosse.py index 9cc5f618..8aad0b05 100644 --- a/embiggen/embedders/pykeen_embedders/crosse.py +++ b/embiggen/embedders/pykeen_embedders/crosse.py @@ -16,6 +16,7 @@ def __init__( batch_size: int = 2**10, training_loop: Union[str, Type[TrainingLoop] ] = "Stochastic Local Closed World Assumption", + verbose: bool = False, random_state: int = 42, enable_cache: bool = False ): @@ -46,6 +47,8 @@ def __init__( Can either be: - Stochastic Local Closed World Assumption - Local Closed World Assumption + verbose: bool = False + Whether to show loading bars. random_state: int = 42 Random seed to use while training the model enable_cache: bool = False @@ -58,17 +61,18 @@ def __init__( epochs=epochs, batch_size=batch_size, training_loop=training_loop, + verbose=verbose, random_state=random_state, enable_cache=enable_cache ) def parameters(self) -> Dict[str, Any]: - return { + return dict( **super().parameters(), **dict( combination_dropout=self._combination_dropout, ) - } + ) @classmethod def model_name(cls) -> str: @@ -90,4 +94,5 @@ def _build_model( triples_factory=triples_factory, embedding_dim=self._embedding_size, combination_dropout=self._combination_dropout, + random_seed=self._random_state ) diff --git a/embiggen/embedders/pykeen_embedders/distma.py b/embiggen/embedders/pykeen_embedders/distma.py index 6c82c557..cfa6d0db 100644 --- a/embiggen/embedders/pykeen_embedders/distma.py +++ b/embiggen/embedders/pykeen_embedders/distma.py @@ -8,56 +8,6 @@ class DistMAPyKeen(EntityRelationEmbeddingModelPyKeen): - def __init__( - self, - embedding_size: int = 256, - epochs: int = 100, - batch_size: int = 2**10, - training_loop: Union[str, Type[TrainingLoop] - ] = "Stochastic Local Closed World Assumption", - random_state: int = 42, - enable_cache: bool = False - ): - """Create new PyKeen DistMA model. - - Details - ------------------------- - This is a wrapper of the DistMA implementation from the - PyKeen library. Please refer to the PyKeen library documentation - for details and posssible errors regarding this model. - - Parameters - ------------------------- - embedding_size: int = 256 - The dimension of the embedding to compute. - epochs: int = 100 - The number of epochs to use to train the model for. - batch_size: int = 2**10 - Size of the training batch. - device: str = "auto" - The devide to use to train the model. - Can either be cpu or cuda. - training_loop: Union[str, Type[TrainingLoop] - ] = "Stochastic Local Closed World Assumption" - The training loop to use to train the model. - Can either be: - - Stochastic Local Closed World Assumption - - Local Closed World Assumption - random_state: int = 42 - Random seed to use while training the model - enable_cache: bool = False - Whether to enable the cache, that is to - store the computed embedding. - """ - super().__init__( - embedding_size=embedding_size, - epochs=epochs, - batch_size=batch_size, - training_loop=training_loop, - random_state=random_state, - enable_cache=enable_cache - ) - @classmethod def model_name(cls) -> str: """Return name of the model.""" @@ -77,4 +27,5 @@ def _build_model( return DistMA( triples_factory=triples_factory, embedding_dim=self._embedding_size, + random_seed=self._random_state ) diff --git a/embiggen/embedders/pykeen_embedders/distmult.py b/embiggen/embedders/pykeen_embedders/distmult.py index a7770b59..c628b683 100644 --- a/embiggen/embedders/pykeen_embedders/distmult.py +++ b/embiggen/embedders/pykeen_embedders/distmult.py @@ -8,56 +8,6 @@ class DistMultPyKeen(EntityRelationEmbeddingModelPyKeen): - def __init__( - self, - embedding_size: int = 256, - epochs: int = 100, - batch_size: int = 2**10, - training_loop: Union[str, Type[TrainingLoop] - ] = "Stochastic Local Closed World Assumption", - random_state: int = 42, - enable_cache: bool = False - ): - """Create new PyKeen DistMult model. - - Details - ------------------------- - This is a wrapper of the DistMult implementation from the - PyKeen library. Please refer to the PyKeen library documentation - for details and posssible errors regarding this model. - - Parameters - ------------------------- - embedding_size: int = 256 - The dimension of the embedding to compute. - epochs: int = 100 - The number of epochs to use to train the model for. - batch_size: int = 2**10 - Size of the training batch. - device: str = "auto" - The devide to use to train the model. - Can either be cpu or cuda. - training_loop: Union[str, Type[TrainingLoop] - ] = "Stochastic Local Closed World Assumption" - The training loop to use to train the model. - Can either be: - - Stochastic Local Closed World Assumption - - Local Closed World Assumption - random_state: int = 42 - Random seed to use while training the model - enable_cache: bool = False - Whether to enable the cache, that is to - store the computed embedding. - """ - super().__init__( - embedding_size=embedding_size, - epochs=epochs, - batch_size=batch_size, - training_loop=training_loop, - random_state=random_state, - enable_cache=enable_cache - ) - @classmethod def model_name(cls) -> str: """Return name of the model.""" @@ -77,4 +27,5 @@ def _build_model( return DistMult( triples_factory=triples_factory, embedding_dim=self._embedding_size, + random_seed=self._random_state ) diff --git a/embiggen/embedders/pykeen_embedders/ermlp.py b/embiggen/embedders/pykeen_embedders/ermlp.py index fe54ab47..f0788f15 100644 --- a/embiggen/embedders/pykeen_embedders/ermlp.py +++ b/embiggen/embedders/pykeen_embedders/ermlp.py @@ -16,6 +16,7 @@ def __init__( batch_size: int = 2**10, training_loop: Union[str, Type[TrainingLoop] ] = "Stochastic Local Closed World Assumption", + verbose: bool = False, random_state: int = 42, enable_cache: bool = False ): @@ -46,6 +47,8 @@ def __init__( Can either be: - Stochastic Local Closed World Assumption - Local Closed World Assumption + verbose: bool = False + Whether to show loading bars. random_state: int = 42 Random seed to use while training the model enable_cache: bool = False @@ -58,6 +61,7 @@ def __init__( epochs=epochs, batch_size=batch_size, training_loop=training_loop, + verbose=verbose, random_state=random_state, enable_cache=enable_cache ) @@ -71,12 +75,12 @@ def smoke_test_parameters(cls) -> Dict[str, Any]: ) def parameters(self) -> Dict[str, Any]: - return { + return dict( **super().parameters(), **dict( hidden_dim=self._hidden_dim ) - } + ) @classmethod def model_name(cls) -> str: @@ -97,5 +101,6 @@ def _build_model( return ERMLP( triples_factory=triples_factory, embedding_dim=self._embedding_size, - hidden_dim=self._hidden_dim + hidden_dim=self._hidden_dim, + random_seed=self._random_state ) diff --git a/embiggen/embedders/pykeen_embedders/ermlpe.py b/embiggen/embedders/pykeen_embedders/ermlpe.py index 2bfb8b31..638bc0c4 100644 --- a/embiggen/embedders/pykeen_embedders/ermlpe.py +++ b/embiggen/embedders/pykeen_embedders/ermlpe.py @@ -16,6 +16,7 @@ def __init__( batch_size: int = 2**10, training_loop: Union[str, Type[TrainingLoop] ] = "Stochastic Local Closed World Assumption", + verbose: bool = False, random_state: int = 42, enable_cache: bool = False ): @@ -46,6 +47,8 @@ def __init__( Can either be: - Stochastic Local Closed World Assumption - Local Closed World Assumption + verbose: bool = False + Whether to show loading bars. random_state: int = 42 Random seed to use while training the model enable_cache: bool = False @@ -58,6 +61,7 @@ def __init__( epochs=epochs, batch_size=batch_size, training_loop=training_loop, + verbose=verbose, random_state=random_state, enable_cache=enable_cache ) @@ -71,12 +75,12 @@ def smoke_test_parameters(cls) -> Dict[str, Any]: ) def parameters(self) -> Dict[str, Any]: - return { + return dict( **super().parameters(), **dict( hidden_dim=self._hidden_dim ) - } + ) @classmethod def model_name(cls) -> str: @@ -97,5 +101,6 @@ def _build_model( return ERMLPE( triples_factory=triples_factory, embedding_dim=self._embedding_size, - hidden_dim=self._hidden_dim + hidden_dim=self._hidden_dim, + random_seed=self._random_state ) diff --git a/embiggen/embedders/pykeen_embedders/hole.py b/embiggen/embedders/pykeen_embedders/hole.py index 3d38369c..901ac5fa 100644 --- a/embiggen/embedders/pykeen_embedders/hole.py +++ b/embiggen/embedders/pykeen_embedders/hole.py @@ -8,56 +8,6 @@ class HolEPyKeen(EntityRelationEmbeddingModelPyKeen): - def __init__( - self, - embedding_size: int = 100, - epochs: int = 100, - batch_size: int = 2**10, - training_loop: Union[str, Type[TrainingLoop] - ] = "Stochastic Local Closed World Assumption", - random_state: int = 42, - enable_cache: bool = False - ): - """Create new PyKeen HolE model. - - Details - ------------------------- - This is a wrapper of the HolE implementation from the - PyKeen library. Please refer to the PyKeen library documentation - for details and posssible errors regarding this model. - - Parameters - ------------------------- - embedding_size: int = 100 - The dimension of the embedding to compute. - epochs: int = 100 - The number of epochs to use to train the model for. - batch_size: int = 2**10 - Size of the training batch. - device: str = "auto" - The devide to use to train the model. - Can either be cpu or cuda. - training_loop: Union[str, Type[TrainingLoop] - ] = "Stochastic Local Closed World Assumption" - The training loop to use to train the model. - Can either be: - - Stochastic Local Closed World Assumption - - Local Closed World Assumption - random_state: int = 42 - Random seed to use while training the model - enable_cache: bool = False - Whether to enable the cache, that is to - store the computed embedding. - """ - super().__init__( - embedding_size=embedding_size, - epochs=epochs, - batch_size=batch_size, - training_loop=training_loop, - random_state=random_state, - enable_cache=enable_cache - ) - @classmethod def model_name(cls) -> str: """Return name of the model.""" @@ -77,4 +27,5 @@ def _build_model( return HolE( triples_factory=triples_factory, embedding_dim=self._embedding_size, + random_seed=self._random_state ) diff --git a/embiggen/embedders/pykeen_embedders/pairre.py b/embiggen/embedders/pykeen_embedders/pairre.py index c5c7cf48..2002ba30 100644 --- a/embiggen/embedders/pykeen_embedders/pairre.py +++ b/embiggen/embedders/pykeen_embedders/pairre.py @@ -17,6 +17,7 @@ def __init__( batch_size: int = 2**10, training_loop: Union[str, Type[TrainingLoop] ] = "Stochastic Local Closed World Assumption", + verbose: bool = False, random_state: int = 42, enable_cache: bool = False ): @@ -49,6 +50,8 @@ def __init__( Can either be: - Stochastic Local Closed World Assumption - Local Closed World Assumption + verbose: bool = False + Whether to show loading bars. random_state: int = 42 Random seed to use while training the model enable_cache: bool = False @@ -62,18 +65,19 @@ def __init__( epochs=epochs, batch_size=batch_size, training_loop=training_loop, + verbose=verbose, random_state=random_state, enable_cache=enable_cache ) def parameters(self) -> Dict[str, Any]: - return { + return dict( **super().parameters(), **dict( p=self._p, power_norm=self._power_norm, ) - } + ) @classmethod def model_name(cls) -> str: @@ -96,4 +100,5 @@ def _build_model( embedding_dim=self._embedding_size, p=self._p, power_norm=self._power_norm, + random_seed=self._random_state ) diff --git a/embiggen/embedders/pykeen_embedders/proje.py b/embiggen/embedders/pykeen_embedders/proje.py index 575c74f1..e677551f 100644 --- a/embiggen/embedders/pykeen_embedders/proje.py +++ b/embiggen/embedders/pykeen_embedders/proje.py @@ -8,56 +8,6 @@ class ProjEPyKeen(EntityRelationEmbeddingModelPyKeen): - def __init__( - self, - embedding_size: int = 50, - epochs: int = 100, - batch_size: int = 2**10, - training_loop: Union[str, Type[TrainingLoop] - ] = "Stochastic Local Closed World Assumption", - random_state: int = 42, - enable_cache: bool = False - ): - """Create new PyKeen ProjE model. - - Details - ------------------------- - This is a wrapper of the ProjE implementation from the - PyKeen library. Please refer to the PyKeen library documentation - for details and posssible errors regarding this model. - - Parameters - ------------------------- - embedding_size: int = 50 - The dimension of the embedding to compute. - epochs: int = 100 - The number of epochs to use to train the model for. - batch_size: int = 2**10 - Size of the training batch. - device: str = "auto" - The devide to use to train the model. - Can either be cpu or cuda. - training_loop: Union[str, Type[TrainingLoop] - ] = "Stochastic Local Closed World Assumption" - The training loop to use to train the model. - Can either be: - - Stochastic Local Closed World Assumption - - Local Closed World Assumption - random_state: int = 42 - Random seed to use while training the model - enable_cache: bool = False - Whether to enable the cache, that is to - store the computed embedding. - """ - super().__init__( - embedding_size=embedding_size, - epochs=epochs, - batch_size=batch_size, - training_loop=training_loop, - random_state=random_state, - enable_cache=enable_cache - ) - @classmethod def model_name(cls) -> str: """Return name of the model.""" @@ -77,4 +27,5 @@ def _build_model( return ProjE( triples_factory=triples_factory, embedding_dim=self._embedding_size, + random_seed=self._random_state ) diff --git a/embiggen/embedders/pykeen_embedders/pykeen_embedder.py b/embiggen/embedders/pykeen_embedders/pykeen_embedder.py index bf761719..ea975a50 100644 --- a/embiggen/embedders/pykeen_embedders/pykeen_embedder.py +++ b/embiggen/embedders/pykeen_embedders/pykeen_embedder.py @@ -1,5 +1,5 @@ """Abstract Torch/PyKeen Model wrapper for embedding models.""" -from typing import Dict, List, Sequence, Union, Optional, Tuple, Any, Type +from typing import Dict, Union, Tuple, Any, Type import numpy as np import pandas as pd @@ -13,7 +13,6 @@ from pykeen.models import Model from pykeen.triples import CoreTriplesFactory from pykeen.training import SLCWATrainingLoop, LCWATrainingLoop, TrainingLoop -from torch.optim import Optimizer @abstract_class @@ -33,6 +32,7 @@ def __init__( device: str = "auto", training_loop: Union[str, Type[TrainingLoop] ] = "Stochastic Local Closed World Assumption", + verbose: bool = False, random_state: int = 42, enable_cache: bool = False ): @@ -55,6 +55,8 @@ def __init__( Can either be: - Stochastic Local Closed World Assumption - Local Closed World Assumption + verbose: bool = False + Whether to show the loading bar. random_state: int = 42 Random seed to use while training the model enable_cache: bool = False @@ -83,6 +85,7 @@ def __init__( self._training_loop = training_loop self._epochs = epochs + self._verbose = verbose self._batch_size = batch_size self._device = validate_torch_device(device) @@ -101,13 +104,13 @@ def smoke_test_parameters(cls) -> Dict[str, Any]: ) def parameters(self) -> Dict[str, Any]: - return { + return dict( **super().parameters(), **dict( epochs=self._epochs, batch_size=self._batch_size, ) - } + ) @classmethod def library_name(cls) -> str: @@ -168,7 +171,6 @@ def _fit_transform( self, graph: Graph, return_dataframe: bool = True, - verbose: bool = True ) -> Union[np.ndarray, pd.DataFrame, Dict[str, np.ndarray], Dict[str, pd.DataFrame]]: """Return node embedding""" @@ -210,10 +212,12 @@ def _fit_transform( triples_factory=triples_factory, num_epochs=self._epochs, batch_size=batch_size, - use_tqdm=True, - use_tqdm_batch=True, + use_tqdm=self._verbose, + use_tqdm_batch=self._verbose, tqdm_kwargs=dict( - disable=not verbose + disable=not self._verbose, + dynamic_ncols=True, + leave=False ) ) diff --git a/embiggen/embedders/pykeen_embedders/quate.py b/embiggen/embedders/pykeen_embedders/quate.py index 98d9ba51..ca9caff0 100644 --- a/embiggen/embedders/pykeen_embedders/quate.py +++ b/embiggen/embedders/pykeen_embedders/quate.py @@ -8,56 +8,6 @@ class QuatEPyKeen(EntityRelationEmbeddingModelPyKeen): - def __init__( - self, - embedding_size: int = 100, - epochs: int = 100, - batch_size: int = 2**10, - training_loop: Union[str, Type[TrainingLoop] - ] = "Stochastic Local Closed World Assumption", - random_state: int = 42, - enable_cache: bool = False - ): - """Create new PyKeen QuatE model. - - Details - ------------------------- - This is a wrapper of the QuatE implementation from the - PyKeen library. Please refer to the PyKeen library documentation - for details and posssible errors regarding this model. - - Parameters - ------------------------- - embedding_size: int = 100 - The dimension of the embedding to compute. - epochs: int = 100 - The number of epochs to use to train the model for. - batch_size: int = 2**10 - Size of the training batch. - device: str = "auto" - The devide to use to train the model. - Can either be cpu or cuda. - training_loop: Union[str, Type[TrainingLoop] - ] = "Stochastic Local Closed World Assumption" - The training loop to use to train the model. - Can either be: - - Stochastic Local Closed World Assumption - - Local Closed World Assumption - random_state: int = 42 - Random seed to use while training the model - enable_cache: bool = False - Whether to enable the cache, that is to - store the computed embedding. - """ - super().__init__( - embedding_size=embedding_size, - epochs=epochs, - batch_size=batch_size, - training_loop=training_loop, - random_state=random_state, - enable_cache=enable_cache - ) - @classmethod def model_name(cls) -> str: """Return name of the model.""" @@ -77,4 +27,5 @@ def _build_model( return QuatE( triples_factory=triples_factory, embedding_dim=self._embedding_size, + random_seed=self._random_state ) diff --git a/embiggen/embedders/pykeen_embedders/rescal.py b/embiggen/embedders/pykeen_embedders/rescal.py index 8376f16d..40071b24 100644 --- a/embiggen/embedders/pykeen_embedders/rescal.py +++ b/embiggen/embedders/pykeen_embedders/rescal.py @@ -8,56 +8,6 @@ class RESCALPyKeen(EntityRelationEmbeddingModelPyKeen): - def __init__( - self, - embedding_size: int = 100, - epochs: int = 100, - batch_size: int = 2**10, - training_loop: Union[str, Type[TrainingLoop] - ] = "Stochastic Local Closed World Assumption", - random_state: int = 42, - enable_cache: bool = False - ): - """Create new PyKeen RESCAL model. - - Details - ------------------------- - This is a wrapper of the RESCAL implementation from the - PyKeen library. Please refer to the PyKeen library documentation - for details and posssible errors regarding this model. - - Parameters - ------------------------- - embedding_size: int = 100 - The dimension of the embedding to compute. - epochs: int = 100 - The number of epochs to use to train the model for. - batch_size: int = 2**10 - Size of the training batch. - device: str = "auto" - The devide to use to train the model. - Can either be cpu or cuda. - training_loop: Union[str, Type[TrainingLoop] - ] = "Stochastic Local Closed World Assumption" - The training loop to use to train the model. - Can either be: - - Stochastic Local Closed World Assumption - - Local Closed World Assumption - random_state: int = 42 - Random seed to use while training the model - enable_cache: bool = False - Whether to enable the cache, that is to - store the computed embedding. - """ - super().__init__( - embedding_size=embedding_size, - epochs=epochs, - batch_size=batch_size, - training_loop=training_loop, - random_state=random_state, - enable_cache=enable_cache - ) - @classmethod def model_name(cls) -> str: """Return name of the model.""" @@ -77,4 +27,5 @@ def _build_model( return RESCAL( triples_factory=triples_factory, embedding_dim=self._embedding_size, + random_seed=self._random_state ) diff --git a/embiggen/embedders/pykeen_embedders/rotate.py b/embiggen/embedders/pykeen_embedders/rotate.py index 2e7c792a..0a46929c 100644 --- a/embiggen/embedders/pykeen_embedders/rotate.py +++ b/embiggen/embedders/pykeen_embedders/rotate.py @@ -8,56 +8,6 @@ class RotatEPyKeen(EntityRelationEmbeddingModelPyKeen): - def __init__( - self, - embedding_size: int = 200, - epochs: int = 100, - batch_size: int = 2**10, - training_loop: Union[str, Type[TrainingLoop] - ] = "Stochastic Local Closed World Assumption", - random_state: int = 42, - enable_cache: bool = False - ): - """Create new PyKeen RotatE model. - - Details - ------------------------- - This is a wrapper of the RotatE implementation from the - PyKeen library. Please refer to the PyKeen library documentation - for details and posssible errors regarding this model. - - Parameters - ------------------------- - embedding_size: int = 200 - The dimension of the embedding to compute. - epochs: int = 100 - The number of epochs to use to train the model for. - batch_size: int = 2**10 - Size of the training batch. - device: str = "auto" - The devide to use to train the model. - Can either be cpu or cuda. - training_loop: Union[str, Type[TrainingLoop] - ] = "Stochastic Local Closed World Assumption" - The training loop to use to train the model. - Can either be: - - Stochastic Local Closed World Assumption - - Local Closed World Assumption - random_state: int = 42 - Random seed to use while training the model - enable_cache: bool = False - Whether to enable the cache, that is to - store the computed embedding. - """ - super().__init__( - embedding_size=embedding_size, - epochs=epochs, - batch_size=batch_size, - training_loop=training_loop, - random_state=random_state, - enable_cache=enable_cache - ) - @classmethod def model_name(cls) -> str: """Return name of the model.""" @@ -77,4 +27,5 @@ def _build_model( return RotatE( triples_factory=triples_factory, embedding_dim=self._embedding_size, + random_seed=self._random_state ) diff --git a/embiggen/embedders/pykeen_embedders/toruse.py b/embiggen/embedders/pykeen_embedders/toruse.py index d3a178a4..c0d101b4 100644 --- a/embiggen/embedders/pykeen_embedders/toruse.py +++ b/embiggen/embedders/pykeen_embedders/toruse.py @@ -17,6 +17,7 @@ def __init__( batch_size: int = 2**10, training_loop: Union[str, Type[TrainingLoop] ] = "Stochastic Local Closed World Assumption", + verbose: bool = False, random_state: int = 42, enable_cache: bool = False ): @@ -49,6 +50,8 @@ def __init__( Can either be: - Stochastic Local Closed World Assumption - Local Closed World Assumption + verbose: bool = False + Whether to show loading bars. random_state: int = 42 Random seed to use while training the model enable_cache: bool = False @@ -62,18 +65,19 @@ def __init__( epochs=epochs, batch_size=batch_size, training_loop=training_loop, + verbose=verbose, random_state=random_state, enable_cache=enable_cache ) def parameters(self) -> Dict[str, Any]: - return { + return dict( **super().parameters(), **dict( p=self._p, power_norm=self._power_norm, ) - } + ) @classmethod def model_name(cls) -> str: @@ -96,4 +100,5 @@ def _build_model( embedding_dim=self._embedding_size, p=self._p, power_norm=self._power_norm, + random_seed=self._random_state ) diff --git a/embiggen/embedders/pykeen_embedders/transd.py b/embiggen/embedders/pykeen_embedders/transd.py index b9fd8974..1b8c4c39 100644 --- a/embiggen/embedders/pykeen_embedders/transd.py +++ b/embiggen/embedders/pykeen_embedders/transd.py @@ -16,6 +16,7 @@ def __init__( batch_size: int = 2**10, training_loop: Union[str, Type[TrainingLoop] ] = "Stochastic Local Closed World Assumption", + verbose: bool = False, random_state: int = 42, enable_cache: bool = False ): @@ -46,6 +47,8 @@ def __init__( Can either be: - Stochastic Local Closed World Assumption - Local Closed World Assumption + verbose: bool = False + Whether to show loading bars. random_state: int = 42 Random seed to use while training the model enable_cache: bool = False @@ -58,6 +61,7 @@ def __init__( epochs=epochs, batch_size=batch_size, training_loop=training_loop, + verbose=verbose, random_state=random_state, enable_cache=enable_cache ) @@ -71,12 +75,12 @@ def smoke_test_parameters(cls) -> Dict[str, Any]: ) def parameters(self) -> Dict[str, Any]: - return { + return dict( **super().parameters(), **dict( relation_dim=self._relation_dim ) - } + ) @classmethod def model_name(cls) -> str: diff --git a/embiggen/embedders/pykeen_embedders/transe.py b/embiggen/embedders/pykeen_embedders/transe.py index f209b01c..3435e5bc 100644 --- a/embiggen/embedders/pykeen_embedders/transe.py +++ b/embiggen/embedders/pykeen_embedders/transe.py @@ -16,6 +16,7 @@ def __init__( batch_size: int = 2**10, training_loop: Union[str, Type[TrainingLoop] ] = "Stochastic Local Closed World Assumption", + verbose: bool = False, random_state: int = 42, enable_cache: bool = False ): @@ -46,6 +47,8 @@ def __init__( Can either be: - Stochastic Local Closed World Assumption - Local Closed World Assumption + verbose: bool = False + Whether to show loading bars. random_state: int = 42 Random seed to use while training the model enable_cache: bool = False @@ -58,6 +61,7 @@ def __init__( epochs=epochs, batch_size=batch_size, training_loop=training_loop, + verbose=verbose, random_state=random_state, enable_cache=enable_cache ) @@ -71,12 +75,12 @@ def smoke_test_parameters(cls) -> Dict[str, Any]: ) def parameters(self) -> Dict[str, Any]: - return { + return dict( **super().parameters(), **dict( scoring_fct_norm=self._scoring_fct_norm ) - } + ) @classmethod def model_name(cls) -> str: @@ -97,5 +101,6 @@ def _build_model( return TransE( triples_factory=triples_factory, embedding_dim=self._embedding_size, - scoring_fct_norm=self._scoring_fct_norm + scoring_fct_norm=self._scoring_fct_norm, + random_seed=self._random_state ) diff --git a/embiggen/embedders/pykeen_embedders/transf.py b/embiggen/embedders/pykeen_embedders/transf.py index c29838bd..f6d4cf1f 100644 --- a/embiggen/embedders/pykeen_embedders/transf.py +++ b/embiggen/embedders/pykeen_embedders/transf.py @@ -8,56 +8,6 @@ class TransFPyKeen(EntityRelationEmbeddingModelPyKeen): - def __init__( - self, - embedding_size: int = 100, - epochs: int = 100, - batch_size: int = 2**10, - training_loop: Union[str, Type[TrainingLoop] - ] = "Stochastic Local Closed World Assumption", - random_state: int = 42, - enable_cache: bool = False - ): - """Create new PyKeen TransF model. - - Details - ------------------------- - This is a wrapper of the TransF implementation from the - PyKeen library. Please refer to the PyKeen library documentation - for details and posssible errors regarding this model. - - Parameters - ------------------------- - embedding_size: int = 100 - The dimension of the embedding to compute. - epochs: int = 100 - The number of epochs to use to train the model for. - batch_size: int = 2**10 - Size of the training batch. - device: str = "auto" - The devide to use to train the model. - Can either be cpu or cuda. - training_loop: Union[str, Type[TrainingLoop] - ] = "Stochastic Local Closed World Assumption" - The training loop to use to train the model. - Can either be: - - Stochastic Local Closed World Assumption - - Local Closed World Assumption - random_state: int = 42 - Random seed to use while training the model - enable_cache: bool = False - Whether to enable the cache, that is to - store the computed embedding. - """ - super().__init__( - embedding_size=embedding_size, - epochs=epochs, - batch_size=batch_size, - training_loop=training_loop, - random_state=random_state, - enable_cache=enable_cache - ) - @classmethod def model_name(cls) -> str: """Return name of the model.""" @@ -77,4 +27,5 @@ def _build_model( return TransF( triples_factory=triples_factory, embedding_dim=self._embedding_size, + random_seed=self._random_state ) diff --git a/embiggen/embedders/pykeen_embedders/transh.py b/embiggen/embedders/pykeen_embedders/transh.py index f2849054..1fddcaf6 100644 --- a/embiggen/embedders/pykeen_embedders/transh.py +++ b/embiggen/embedders/pykeen_embedders/transh.py @@ -16,6 +16,7 @@ def __init__( batch_size: int = 2**10, training_loop: Union[str, Type[TrainingLoop] ] = "Stochastic Local Closed World Assumption", + verbose: bool = False, random_state: int = 42, enable_cache: bool = False ): @@ -46,6 +47,8 @@ def __init__( Can either be: - Stochastic Local Closed World Assumption - Local Closed World Assumption + verbose: bool = False + Whether to show loading bars. random_state: int = 42 Random seed to use while training the model enable_cache: bool = False @@ -58,6 +61,7 @@ def __init__( epochs=epochs, batch_size=batch_size, training_loop=training_loop, + verbose=verbose, random_state=random_state, enable_cache=enable_cache ) @@ -71,12 +75,12 @@ def smoke_test_parameters(cls) -> Dict[str, Any]: ) def parameters(self) -> Dict[str, Any]: - return { + return dict( **super().parameters(), **dict( scoring_fct_norm=self._scoring_fct_norm ) - } + ) @classmethod def model_name(cls) -> str: @@ -97,5 +101,6 @@ def _build_model( return TransH( triples_factory=triples_factory, embedding_dim=self._embedding_size, - scoring_fct_norm=self._scoring_fct_norm + scoring_fct_norm=self._scoring_fct_norm, + random_seed=self._random_state ) diff --git a/embiggen/embedders/pykeen_embedders/transr.py b/embiggen/embedders/pykeen_embedders/transr.py index fa1e48eb..3f8b5637 100644 --- a/embiggen/embedders/pykeen_embedders/transr.py +++ b/embiggen/embedders/pykeen_embedders/transr.py @@ -17,6 +17,7 @@ def __init__( batch_size: int = 2**10, training_loop: Union[str, Type[TrainingLoop] ] = "Stochastic Local Closed World Assumption", + verbose: bool = False, random_state: int = 42, enable_cache: bool = False ): @@ -49,6 +50,8 @@ def __init__( Can either be: - Stochastic Local Closed World Assumption - Local Closed World Assumption + verbose: bool = False + Whether to show loading bars. random_state: int = 42 Random seed to use while training the model enable_cache: bool = False @@ -62,6 +65,7 @@ def __init__( epochs=epochs, batch_size=batch_size, training_loop=training_loop, + verbose=verbose, random_state=random_state, enable_cache=enable_cache ) @@ -76,13 +80,13 @@ def smoke_test_parameters(cls) -> Dict[str, Any]: ) def parameters(self) -> Dict[str, Any]: - return { + return dict( **super().parameters(), **dict( scoring_fct_norm=self._scoring_fct_norm, relation_dim=self._relation_dim ) - } + ) @classmethod def model_name(cls) -> str: @@ -104,5 +108,6 @@ def _build_model( triples_factory=triples_factory, embedding_dim=self._embedding_size, relation_dim=self._relation_dim, - scoring_fct_norm=self._scoring_fct_norm + scoring_fct_norm=self._scoring_fct_norm, + random_seed=self._random_state ) diff --git a/embiggen/embedders/pykeen_embedders/tucker.py b/embiggen/embedders/pykeen_embedders/tucker.py index 99db4d20..70cc9b00 100644 --- a/embiggen/embedders/pykeen_embedders/tucker.py +++ b/embiggen/embedders/pykeen_embedders/tucker.py @@ -20,6 +20,7 @@ def __init__( batch_size: int = 2**10, training_loop: Union[str, Type[TrainingLoop] ] = "Stochastic Local Closed World Assumption", + verbose: bool = False, random_state: int = 42, enable_cache: bool = False ): @@ -58,6 +59,8 @@ def __init__( Can either be: - Stochastic Local Closed World Assumption - Local Closed World Assumption + verbose: bool = False + Whether to show loading bars. random_state: int = 42 Random seed to use while training the model enable_cache: bool = False @@ -74,12 +77,13 @@ def __init__( epochs=epochs, batch_size=batch_size, training_loop=training_loop, + verbose=verbose, random_state=random_state, enable_cache=enable_cache ) def parameters(self) -> Dict[str, Any]: - return { + return dict( **super().parameters(), **dict( relation_dim=self._relation_dim, @@ -88,7 +92,7 @@ def parameters(self) -> Dict[str, Any]: dropout_2=self._dropout_2, apply_batch_normalization=self._apply_batch_normalization ) - } + ) @classmethod def model_name(cls) -> str: @@ -113,5 +117,6 @@ def _build_model( dropout_0=self._dropout_0, dropout_1=self._dropout_1, dropout_2=self._dropout_2, - apply_batch_normalization=self._apply_batch_normalization + apply_batch_normalization=self._apply_batch_normalization, + random_seed=self._random_state ) diff --git a/embiggen/embedders/tensorflow_embedders/cbow.py b/embiggen/embedders/tensorflow_embedders/cbow.py index bbe9eedf..f9f5cb19 100644 --- a/embiggen/embedders/tensorflow_embedders/cbow.py +++ b/embiggen/embedders/tensorflow_embedders/cbow.py @@ -1,5 +1,4 @@ """CBOW model for sequence embedding.""" -from typing import Dict, Union from tensorflow.keras.layers import ( # pylint: disable=import-error,no-name-in-module GlobalAveragePooling1D, Input, Embedding ) @@ -22,7 +21,7 @@ class CBOWTensorFlow(Node2Vec): @classmethod def model_name(cls) -> str: """Returns name of the model.""" - return "CBOW" + return "Node2Vec CBOW" def _build_model(self, graph: Graph) -> Model: """Return CBOW model.""" @@ -53,7 +52,7 @@ def _build_model(self, graph: Graph) -> Model: model = Model( inputs=[contextual_terms, central_terms], outputs=sampled_softmax, - name=self.model_name() + name=self.model_name().replace(" ", "") ) model.compile(optimizer=self._optimizer) diff --git a/embiggen/embedders/tensorflow_embedders/edge_prediction_based_tensorflow_embedders.py b/embiggen/embedders/tensorflow_embedders/edge_prediction_based_tensorflow_embedders.py index 75be62cf..b949b6b8 100644 --- a/embiggen/embedders/tensorflow_embedders/edge_prediction_based_tensorflow_embedders.py +++ b/embiggen/embedders/tensorflow_embedders/edge_prediction_based_tensorflow_embedders.py @@ -31,6 +31,7 @@ def __init__( activation: str = "sigmoid", loss: str = "binary_crossentropy", optimizer: str = "nadam", + verbose: bool = False, enable_cache: bool = False, random_state: int = 42 ): @@ -74,6 +75,8 @@ def __init__( while for the HOPE models this is an Mean squared error. optimizer: str = "nadam" The optimizer to be used during the training of the model. + verbose: bool = False + Whether to show the loading bar while training the model. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -93,19 +96,20 @@ def __init__( epochs=epochs, batch_size=batch_size, optimizer=optimizer, + verbose=verbose, use_mirrored_strategy=use_mirrored_strategy, enable_cache=enable_cache, random_state=random_state ) def parameters(self) -> Dict[str, Any]: - return { + return dict( **super().parameters(), **dict( negative_samples_rate=self._negative_samples_rate, activation=self._activation ) - } + ) def _build_edge_prediction_based_model( self, @@ -192,7 +196,6 @@ def _build_sequence( def _build_input( self, graph: Graph, - verbose: bool ) -> Tuple[np.ndarray]: """Returns values to be fed as input into the model. @@ -200,20 +203,8 @@ def _build_input( ------------------ graph: Graph The graph to build the model for. - verbose: bool - Whether to show loading bars. - Not used in this context. """ - try: - AUTOTUNE = tf.data.AUTOTUNE - except: - AUTOTUNE = tf.data.experimental.AUTOTUNE - - return ( - self._build_sequence(graph) - .into_dataset() - .repeat() - .prefetch(AUTOTUNE), ) + return (self._build_sequence(graph), ) @classmethod def requires_nodes_sorted_by_decreasing_node_degree(cls) -> bool: diff --git a/embiggen/embedders/tensorflow_embedders/first_order_line.py b/embiggen/embedders/tensorflow_embedders/first_order_line.py index c0f13fec..fcc90a45 100644 --- a/embiggen/embedders/tensorflow_embedders/first_order_line.py +++ b/embiggen/embedders/tensorflow_embedders/first_order_line.py @@ -44,7 +44,7 @@ def _build_edge_prediction_based_model( @classmethod def model_name(cls) -> str: """Returns name of the current model.""" - return "First Order LINE" + return "First-order LINE" def _extract_embeddings( self, diff --git a/embiggen/embedders/tensorflow_embedders/node2vec.py b/embiggen/embedders/tensorflow_embedders/node2vec.py index e916883e..bd61f805 100644 --- a/embiggen/embedders/tensorflow_embedders/node2vec.py +++ b/embiggen/embedders/tensorflow_embedders/node2vec.py @@ -17,7 +17,7 @@ class Node2Vec(AbstractRandomWalkBasedEmbedderModel): def __init__( self, - number_of_negative_samples: int = 5, + number_of_negative_samples: int = 10, batch_size: int = 128, embedding_size: int = 100, epochs: int = 10, @@ -35,7 +35,8 @@ def __init__( max_neighbours: int = 100, normalize_by_degree: bool = False, random_state: int = 42, - optimizer: str = "sgd", + optimizer: str = "nadam", + verbose: bool = False, use_mirrored_strategy: bool = False, enable_cache: bool = False ): @@ -43,14 +44,14 @@ def __init__( Parameters ------------------------------- - number_of_negative_samples: int = 5 + number_of_negative_samples: int = 10 The number of negative classes to randomly sample per batch. This single sample of negative classes is evaluated for each element in the batch. batch_size: int = 128 The number of nodes to consider for each walk. embedding_size: int = 100 Dimension of the embedding. - epochs: int = 10 + epochs: int = 50 Number of epochs to train the model for. early_stopping_min_delta: float The minimum variation in the provided patience time @@ -102,8 +103,10 @@ def __init__( of the destination node degrees. random_state: int = 42 The random state to reproduce the training sequence. - optimizer: str = "sgd" + optimizer: str = "nadam" Optimizer to use during the training. + verbose: bool = False + Whether to show loading bars. use_mirrored_strategy: bool = False Whether to use mirrored strategy. enable_cache: bool = False @@ -131,6 +134,7 @@ def __init__( epochs=epochs, batch_size=batch_size, optimizer=optimizer, + verbose=verbose, use_mirrored_strategy=use_mirrored_strategy, enable_cache=enable_cache ) @@ -163,7 +167,6 @@ def _get_steps_per_epoch(self, graph: Graph) -> Tuple[Any]: def _build_input( self, graph: Graph, - verbose: bool ) -> Tuple[np.ndarray]: """Returns values to be fed as input into the model. @@ -171,9 +174,6 @@ def _build_input( ------------------ graph: Graph The graph to build the model for. - verbose: bool - Whether to show loading bars. - Not used in this context. """ try: AUTOTUNE = tf.data.AUTOTUNE diff --git a/embiggen/embedders/tensorflow_embedders/second_order_line.py b/embiggen/embedders/tensorflow_embedders/second_order_line.py index 96c47352..840983d8 100644 --- a/embiggen/embedders/tensorflow_embedders/second_order_line.py +++ b/embiggen/embedders/tensorflow_embedders/second_order_line.py @@ -51,7 +51,7 @@ def _build_edge_prediction_based_model( @classmethod def model_name(cls) -> str: """Returns name of the current model.""" - return "Second Order LINE" + return "Second-order LINE" def _extract_embeddings( self, diff --git a/embiggen/embedders/tensorflow_embedders/siamese.py b/embiggen/embedders/tensorflow_embedders/siamese.py index f460f052..7753ec7b 100644 --- a/embiggen/embedders/tensorflow_embedders/siamese.py +++ b/embiggen/embedders/tensorflow_embedders/siamese.py @@ -2,23 +2,17 @@ from typing import Dict, Tuple, Any, Optional import numpy as np -import pandas as pd import tensorflow as tf from ensmallen import Graph from tensorflow.keras import \ backend as K # pylint: disable=import-error,no-name-in-module -from tensorflow.keras.constraints import \ - UnitNorm # pylint: disable=import-error,no-name-in-module,no-name-in-module -from tensorflow.keras.layers import \ - Embedding # pylint: disable=import-error,no-name-in-module from tensorflow.keras.layers import ( # pylint: disable=import-error,no-name-in-module - Add, - GlobalAveragePooling1D, - Input + Input, ReLU ) from tensorflow.keras.models import Model +from embiggen.layers.tensorflow import FlatEmbedding, ElementWiseL2, ElementWiseL1 from embiggen.utils.abstract_models import abstract_class -from embiggen.sequences.tensorflow_sequences import SiameseSequence, KGSiameseSequence +from embiggen.sequences.tensorflow_sequences import SiameseSequence from embiggen.embedders.tensorflow_embedders.tensorflow_embedder import TensorFlowEmbedder @@ -30,14 +24,16 @@ def __init__( self, embedding_size: int = 100, relu_bias: float = 1.0, - epochs: int = 10, - batch_size: int = 2**10, - early_stopping_min_delta: float = 0.001, - early_stopping_patience: int = 5, - learning_rate_plateau_min_delta: float = 0.001, - learning_rate_plateau_patience: int = 2, + epochs: int = 50, + batch_size: int = 2**8, + early_stopping_min_delta: float = 0.0001, + early_stopping_patience: int = 10, + learning_rate_plateau_min_delta: float = 0.0001, + learning_rate_plateau_patience: int = 5, + norm: str = "L2", use_mirrored_strategy: bool = False, optimizer: str = "nadam", + verbose: bool = False, enable_cache: bool = False, random_state: int = 42 ): @@ -51,9 +47,9 @@ def __init__( It is not possible to provide both at once. relu_bias: float = 1.0 The bias to use for the ReLu. - epochs: int = 10 + epochs: int = 50 Number of epochs to train the model for. - batch_size: int = 2**14 + batch_size: int = 2**8 Batch size to use during the training. early_stopping_min_delta: float = 0.001 The minimum variation in the provided patience time @@ -67,10 +63,15 @@ def __init__( learning_rate_plateau_patience: int = 1 The amount of epochs to wait for better training performance without decreasing the learning rate. + norm: str = "L2" + Normalization to use. + Can either be `L1` or `L2`. use_mirrored_strategy: bool = False Whether to use mirrored strategy. optimizer: str = "nadam" The optimizer to be used during the training of the model. + verbose: bool = False + Whether to show loading bars. enable_cache: bool = False Whether to enable the cache, that is to store the computed embedding. @@ -78,7 +79,7 @@ def __init__( The random state to use if the model is stocastic. """ self._relu_bias = relu_bias - + self._norm = norm super().__init__( embedding_size=embedding_size, early_stopping_min_delta=early_stopping_min_delta, @@ -88,126 +89,92 @@ def __init__( epochs=epochs, batch_size=batch_size, optimizer=optimizer, + verbose=verbose, use_mirrored_strategy=use_mirrored_strategy, enable_cache=enable_cache, random_state=random_state ) - @classmethod - def smoke_test_parameters(cls) -> Dict[str, Any]: - """Returns parameters for smoke test.""" - return dict( - **TensorFlowEmbedder.smoke_test_parameters(), - ) - def parameters(self) -> Dict[str, Any]: - return { + return dict( **super().parameters(), **dict( - relu_bias = self._relu_bias + relu_bias=self._relu_bias, + norm=self._norm ) - } + ) def _build_model(self, graph: Graph): """Return Siamese model.""" # Creating the inputs layers inputs = [ - Input((1,), dtype=tf.int32) - for _ in range(4) + Input((1,), dtype=tf.int32, name=node_name) + for node_name in ( + "Sources", + "Destinations", + "Corrupted Sources", + "Corrupted Destinations", + ) ] - edge_types = Input((1,), dtype=tf.int32) - # Creating the embedding layer for the contexts - node_embedding_layer = Embedding( - input_dim=graph.get_number_of_nodes(), - output_dim=self._embedding_size, + node_embedding_layer = FlatEmbedding( + vocabulary_size=graph.get_number_of_nodes(), + dimension=self._embedding_size, input_length=1, - name="node_embeddings" + name="NodeEmbedding" ) # Get the node embedding node_embeddings = [ - UnitNorm(axis=-1)(node_embedding_layer(node_input)) + node_embedding_layer(node_input) for node_input in inputs ] - if self.requires_node_types(): - max_node_types = graph.get_maximum_multilabel_count() - multilabel = graph.has_multilabel_node_types() - unknown_node_types = graph.has_unknown_node_types() - node_types_offset = int(multilabel or unknown_node_types) - node_type_inputs = [ - Input((max_node_types,), dtype=tf.int32) - for _ in range(4) - ] - - node_type_embedding_layer = Embedding( - input_dim=graph.get_number_of_node_types() + node_types_offset, - output_dim=self._embedding_size, - input_length=max_node_types, - name="node_type_embeddings", - mask_zero=multilabel or unknown_node_types - ) - - node_embeddings = [ - Add()([ - GlobalAveragePooling1D()( - node_type_embedding_layer(node_type_input) - ), - node_embedding - ]) - for node_type_input, node_embedding in zip( - node_type_inputs, - node_embeddings - ) - ] - else: - node_type_inputs = [] - - inputs.extend(node_type_inputs) - inputs.append(edge_types) - - edge_types_number = graph.get_number_of_edge_types() - unknown_edge_types = graph.has_unknown_edge_types() - edge_types_offset = int(unknown_edge_types) - edge_type_embedding = GlobalAveragePooling1D()(Embedding( - input_dim=edge_types_number, - output_dim=self._embedding_size, - input_length=1 + edge_types_offset, - mask_zero=unknown_edge_types, - name="edge_type_embeddings", - )(edge_types)) - ( + edge_types, + regularization, srcs_embedding, dsts_embedding, not_srcs_embedding, not_dsts_embedding, - edge_type_embedding ) = self._build_output( *node_embeddings, - edge_type_embedding, - edge_types + graph ) - loss = K.relu(self._relu_bias + tf.norm( - srcs_embedding + edge_type_embedding - dsts_embedding, - axis=-1 - ) - tf.norm( - not_srcs_embedding + edge_type_embedding - not_dsts_embedding, - axis=-1 - )) + if self._norm == "L2": + norm_layer = ElementWiseL2 + else: + norm_layer = ElementWiseL1 + + if dsts_embedding is not None: + srcs_embedding = norm_layer()([ + srcs_embedding, + dsts_embedding + ]) + + if not_dsts_embedding is not None: + not_srcs_embedding = norm_layer()([ + not_srcs_embedding, + not_dsts_embedding + ]) + + loss = ReLU()( + self._relu_bias + srcs_embedding - not_srcs_embedding + ) + regularization + + if edge_types is not None: + inputs.append(edge_types) # Creating the actual model model = Model( inputs=inputs, outputs=loss, - name=self.model_name() + name=self.model_name().replace(" ", "") ) model.add_loss(loss) - model.compile(optimizer=self._optimizer) return model @@ -218,10 +185,9 @@ def _build_output( dsts_embedding: tf.Tensor, not_srcs_embedding: tf.Tensor, not_dsts_embedding: tf.Tensor, - edge_type_embedding: tf.Tensor, - edge_type_input: Optional[Input] - ) -> Tuple[tf.Tensor, tf.Tensor, tf.Tensor, tf.Tensor, tf.Tensor]: - """Returns the five input tensors, arbitrarily changed. + graph: Graph + ) -> Tuple[Optional[Input], tf.Tensor, tf.Tensor, tf.Tensor, tf.Tensor, tf.Tensor]: + """Returns the inputs, if any, the regularization loss, and the received node embedding arbitrarily modified. Parameters ---------------------- @@ -233,12 +199,9 @@ def _build_output( Embedding of the fake source node. not_dsts_embedding: tf.Tensor Embedding of the fake destination node. - edge_type_embedding: tf.Tensor - Embedding of the edge types. - edge_type_input: Optional[Input] - Input of the edge types. This is not None - only when there is one such input in the - model, when edge types are requested. + graph: Graph + Graph whose structure is to be used to build + the model. """ raise NotImplementedError( "The method `_build_output` should be implemented in the child " @@ -259,7 +222,6 @@ def _get_steps_per_epoch(self, graph: Graph) -> int: def _build_input( self, graph: Graph, - verbose: bool ) -> Tuple[np.ndarray]: """Returns values to be fed as input into the model. @@ -267,29 +229,16 @@ def _build_input( ------------------ graph: Graph The graph to build the model for. - verbose: bool - Whether to show loading bars. - Not used in this context. """ - try: - AUTOTUNE = tf.data.AUTOTUNE - except: - AUTOTUNE = tf.data.experimental.AUTOTUNE - - if self.requires_node_types(): - sequence = KGSiameseSequence( - graph=graph, - batch_size=self._batch_size, - ) - else: - sequence = SiameseSequence( - graph=graph, - batch_size=self._batch_size, - ) + sequence = SiameseSequence( + graph=graph, + batch_size=self._batch_size, + return_edge_types=self.requires_edge_types() + ) return ( sequence.into_dataset() .repeat() - .prefetch(AUTOTUNE), ) + .prefetch(tf.data.AUTOTUNE), ) @classmethod def requires_nodes_sorted_by_decreasing_node_degree(cls) -> bool: @@ -302,4 +251,4 @@ def is_topological(cls) -> bool: @classmethod def can_use_edge_weights(cls) -> bool: """Returns whether the model can optionally use edge weights.""" - return False \ No newline at end of file + return False diff --git a/embiggen/embedders/tensorflow_embedders/skipgram.py b/embiggen/embedders/tensorflow_embedders/skipgram.py index 546af046..5147f941 100644 --- a/embiggen/embedders/tensorflow_embedders/skipgram.py +++ b/embiggen/embedders/tensorflow_embedders/skipgram.py @@ -1,5 +1,4 @@ """SkipGram model for sequence embedding.""" -from typing import Dict, Union from tensorflow.keras.layers import ( # pylint: disable=import-error,no-name-in-module Input, Embedding, Flatten ) @@ -24,7 +23,7 @@ class SkipGramTensorFlow(Node2Vec): @classmethod def model_name(cls) -> str: """Returns name of the model.""" - return "SkipGram" + return "Node2Vec SkipGram" def _build_model(self, graph: Graph) -> Model: """Return SkipGram model.""" @@ -54,7 +53,7 @@ def _build_model(self, graph: Graph) -> Model: model = Model( inputs=[contextual_terms, central_terms], outputs=output, - name=self.model_name() + name=self.model_name().replace(" ", "") ) model.compile(optimizer=self._optimizer) diff --git a/embiggen/embedders/tensorflow_embedders/structured_embedding.py b/embiggen/embedders/tensorflow_embedders/structured_embedding.py new file mode 100644 index 00000000..c1f1be0c --- /dev/null +++ b/embiggen/embedders/tensorflow_embedders/structured_embedding.py @@ -0,0 +1,143 @@ +"""StructuredEmbedding model.""" +import tensorflow as tf +from ensmallen import Graph +import pandas as pd +from tensorflow.keras import Model +from tensorflow.keras.layers import Input, Reshape +from embiggen.embedders.tensorflow_embedders.siamese import Siamese +from embiggen.utils.abstract_models import EmbeddingResult +from embiggen.layers.tensorflow import FlatEmbedding + + +class StructuredEmbeddingTensorFlow(Siamese): + """StructuredEmbedding model.""" + + def _build_output( + self, + srcs_embedding: tf.Tensor, + dsts_embedding: tf.Tensor, + not_srcs_embedding: tf.Tensor, + not_dsts_embedding: tf.Tensor, + graph: Graph + ): + """Returns the five input tensors, unchanged.""" + edge_types = Input((1,), dtype=tf.int32, name="Edge Types") + source_edge_type_embedding = Reshape(( + self._embedding_size, + self._embedding_size + ))(FlatEmbedding( + vocabulary_size=graph.get_number_of_edge_types(), + dimension=self._embedding_size*self._embedding_size, + input_length=1, + mask_zero=graph.has_unknown_edge_types(), + name="SourceEdgeTypeEmbedding", + )(edge_types)) + destination_edge_type_embedding = Reshape(( + self._embedding_size, + self._embedding_size + ))(FlatEmbedding( + vocabulary_size=graph.get_number_of_edge_types(), + dimension=self._embedding_size*self._embedding_size, + input_length=1, + mask_zero=graph.has_unknown_edge_types(), + name="DestinationEdgeTypeEmbedding", + )(edge_types)) + + return ( + edge_types, + 0.0, + tf.einsum( + 'ijk,ij->ij', + source_edge_type_embedding, + srcs_embedding + ), + tf.einsum( + 'ijk,ij->ij', + destination_edge_type_embedding, + dsts_embedding + ), + tf.einsum( + 'ijk,ij->ij', + source_edge_type_embedding, + not_srcs_embedding + ), + tf.einsum( + 'ijk,ij->ij', + destination_edge_type_embedding, + not_dsts_embedding + ), + ) + + @classmethod + def model_name(cls) -> str: + """Returns name of the current model.""" + return "Structured Embedding" + + @classmethod + def requires_edge_types(cls) -> bool: + return True + + def _extract_embeddings( + self, + graph: Graph, + model: Model, + return_dataframe: bool + ) -> EmbeddingResult: + """Returns embedding from the model. + + Parameters + ------------------ + graph: Graph + The graph that was embedded. + model: Model + The Keras model used to embed the graph. + return_dataframe: bool + Whether to return a dataframe of a numpy array. + """ + node_embedding = self.get_layer_weights( + "NodeEmbedding", + model, + drop_first_row=False + ) + source_edge_type_embedding = self.get_layer_weights( + "SourceEdgeTypeEmbedding", + model, + drop_first_row=graph.has_unknown_edge_types() + ) + destination_edge_type_embedding = self.get_layer_weights( + "DestinationEdgeTypeEmbedding", + model, + drop_first_row=graph.has_unknown_edge_types() + ) + + if return_dataframe: + node_embedding = pd.DataFrame( + node_embedding, + index=graph.get_node_names() + ) + + edge_type_names = graph.get_unique_edge_type_names() + + source_edge_type_embedding = pd.DataFrame( + source_edge_type_embedding, + index=edge_type_names + ) + + destination_edge_type_embedding = pd.DataFrame( + destination_edge_type_embedding, + index=edge_type_names + ) + + return EmbeddingResult( + embedding_method_name=self.model_name(), + node_embeddings=node_embedding, + edge_type_embeddings=[ + source_edge_type_embedding, + destination_edge_type_embedding + ] + ) + + @classmethod + def can_use_node_types(cls) -> bool: + """Returns whether the model can optionally use node types.""" + return False \ No newline at end of file diff --git a/embiggen/embedders/tensorflow_embedders/tensorflow_embedder.py b/embiggen/embedders/tensorflow_embedders/tensorflow_embedder.py index 58649702..d759f698 100644 --- a/embiggen/embedders/tensorflow_embedders/tensorflow_embedder.py +++ b/embiggen/embedders/tensorflow_embedders/tensorflow_embedder.py @@ -28,6 +28,7 @@ def __init__( epochs: int = 10, batch_size: int = 2**10, optimizer: str = "nadam", + verbose: bool = False, use_mirrored_strategy: bool = False, enable_cache: bool = False, random_state: int = 42 @@ -56,6 +57,8 @@ def __init__( Batch size to use during the training. optimizer: str = "nadam" Optimizer to use during the training. + verbose: bool = False + Whether to show loading bars. use_mirrored_strategy: bool = False Whether to use mirrored strategy. enable_cache: bool = False @@ -74,6 +77,7 @@ def __init__( self._epochs = epochs self._batch_size = batch_size self._optimizer = optimizer + self._verbose = verbose self._early_stopping_min_delta = early_stopping_min_delta self._early_stopping_patience = early_stopping_patience self._learning_rate_plateau_min_delta = learning_rate_plateau_min_delta @@ -93,19 +97,20 @@ def smoke_test_parameters(cls) -> Dict[str, Any]: ) def parameters(self) -> Dict[str, Any]: - return { + return dict( **super().parameters(), **dict( use_mirrored_strategy=self._use_mirrored_strategy, epochs=self._epochs, batch_size=self._batch_size, optimizer=self._optimizer, + verbose=self._verbose, early_stopping_min_delta=self._early_stopping_min_delta, early_stopping_patience=self._early_stopping_patience, learning_rate_plateau_min_delta=self._learning_rate_plateau_min_delta, learning_rate_plateau_patience=self._learning_rate_plateau_patience, ) - } + ) @classmethod def library_name(cls) -> str: @@ -125,7 +130,7 @@ def _build_model(self, graph: Graph) -> Model: "called `_build_model`. Please do implement it." ) - def _build_input(self, graph: Graph, verbose: bool) -> Tuple[Any]: + def _build_input(self, graph: Graph) -> Tuple[Any]: """Returns values to be fed as input into the model. Parameters @@ -209,7 +214,6 @@ def _fit_transform( self, graph: Graph, return_dataframe: bool = True, - verbose: bool = True ) -> EmbeddingResult: """Return node embedding""" try: @@ -230,7 +234,6 @@ def _fit_transform( # Get the model input training_input = self._build_input( graph, - verbose=verbose ) if not isinstance(training_input, tuple): @@ -242,7 +245,7 @@ def _fit_transform( model.fit( *training_input, epochs=self._epochs, - verbose=traditional_verbose and verbose > 0, + verbose=traditional_verbose and self._verbose > 0, batch_size=( self._batch_size if issubclass(training_input[0].__class__, Sequence) @@ -264,7 +267,7 @@ def _fit_transform( mode="min", ), *((TqdmCallback(verbose=1, leave=False),) - if not traditional_verbose and verbose > 0 else ()), + if not traditional_verbose and self._verbose > 0 else ()), ], ) @@ -278,4 +281,4 @@ def _fit_transform( @classmethod def is_stocastic(cls) -> bool: """Returns whether the model is stocastic and has therefore a random state.""" - return True \ No newline at end of file + return True diff --git a/embiggen/embedders/tensorflow_embedders/transe.py b/embiggen/embedders/tensorflow_embedders/transe.py index 59644d7a..877aceac 100644 --- a/embiggen/embedders/tensorflow_embedders/transe.py +++ b/embiggen/embedders/tensorflow_embedders/transe.py @@ -2,19 +2,40 @@ from ensmallen import Graph import pandas as pd from tensorflow.keras import Model +from tensorflow.keras.layers import Input +import tensorflow as tf from embiggen.embedders.tensorflow_embedders.siamese import Siamese from embiggen.utils.abstract_models import EmbeddingResult - +from embiggen.layers.tensorflow import FlatEmbedding class TransETensorFlow(Siamese): """TransE model.""" def _build_output( self, - *args + srcs_embedding: tf.Tensor, + dsts_embedding: tf.Tensor, + not_srcs_embedding: tf.Tensor, + not_dsts_embedding: tf.Tensor, + graph: Graph ): """Returns the five input tensors, unchanged.""" - return args[:-1] + edge_types = Input((1,), dtype=tf.int32, name="Edge Types") + edge_type_embedding = FlatEmbedding( + vocabulary_size=graph.get_number_of_edge_types(), + dimension=self._embedding_size, + input_length=1, + mask_zero=graph.has_unknown_edge_types(), + name="EdgeTypeEmbedding", + )(edge_types) + return ( + edge_types, + 0.0, + srcs_embedding + edge_type_embedding, + dsts_embedding, + not_srcs_embedding + edge_type_embedding, + not_dsts_embedding, + ) @classmethod def model_name(cls) -> str: @@ -42,36 +63,31 @@ def _extract_embeddings( return_dataframe: bool Whether to return a dataframe of a numpy array. """ + node_embedding = self.get_layer_weights( + "NodeEmbedding", + model, + drop_first_row=False + ) + edge_type_embedding = self.get_layer_weights( + "EdgeTypeEmbedding", + model, + drop_first_row=graph.has_unknown_edge_types() + ) + if return_dataframe: - result = { - layer_name: pd.DataFrame( - self.get_layer_weights( - layer_name, - model, - drop_first_row=drop_first_row - ), - index=names - ) - for layer_name, names, drop_first_row in ( - ("node_embeddings", graph.get_node_names(), False), - ("edge_type_embeddings", graph.get_unique_edge_type_names(), graph.has_unknown_edge_types()) - ) - } - else: - result = { - layer_name: self.get_layer_weights( - layer_name, - model, - drop_first_row=drop_first_row - ) - for layer_name, drop_first_row in ( - ("node_embeddings", False), - ("edge_type_embeddings", graph.has_unknown_edge_types()) - ) - } + node_embedding = pd.DataFrame( + node_embedding, + index=graph.get_node_names() + ) + edge_type_embedding = pd.DataFrame( + edge_type_embedding, + index=graph.get_unique_edge_type_names() + ) + return EmbeddingResult( embedding_method_name=self.model_name(), - **result + node_embeddings=node_embedding, + edge_type_embeddings=edge_type_embedding ) @classmethod @@ -80,6 +96,6 @@ def can_use_node_types(cls) -> bool: return False @classmethod - def task_involves_edge_types(cls) -> bool: - """Returns whether the model task involves edge types.""" + def requires_edge_types(cls) -> bool: + """Returns whether the model requires edge types.""" return True \ No newline at end of file diff --git a/embiggen/embedders/tensorflow_embedders/transh.py b/embiggen/embedders/tensorflow_embedders/transh.py new file mode 100644 index 00000000..9bd8fed6 --- /dev/null +++ b/embiggen/embedders/tensorflow_embedders/transh.py @@ -0,0 +1,146 @@ +"""TransH model.""" +import tensorflow as tf +from ensmallen import Graph +import pandas as pd +from tensorflow.keras import Model +from tensorflow.keras.layers import Input, Dot +from embiggen.embedders.tensorflow_embedders.siamese import Siamese +from embiggen.utils.abstract_models import EmbeddingResult +from embiggen.layers.tensorflow import FlatEmbedding + + +class TransHTensorFlow(Siamese): + """TransH model.""" + + def _build_output( + self, + srcs_embedding: tf.Tensor, + dsts_embedding: tf.Tensor, + not_srcs_embedding: tf.Tensor, + not_dsts_embedding: tf.Tensor, + graph: Graph + ): + """Returns the five input tensors, unchanged.""" + edge_types = Input((1,), dtype=tf.int32, name="Edge Types") + bias_edge_type_embedding = FlatEmbedding( + vocabulary_size=graph.get_number_of_edge_types(), + dimension=self._embedding_size, + input_length=1, + mask_zero=graph.has_unknown_edge_types(), + name="BiasEdgeTypeEmbedding", + )(edge_types) + multiplicative_edge_type_embedding = FlatEmbedding( + vocabulary_size=graph.get_number_of_edge_types(), + dimension=self._embedding_size, + input_length=1, + mask_zero=graph.has_unknown_edge_types(), + name="MultiplicativeEdgeTypeEmbedding", + )(edge_types) + + dot = Dot(axes=-1)([ + bias_edge_type_embedding, + multiplicative_edge_type_embedding + ]) + + return ( + edge_types, + dot * dot / Dot(axes=-1)([ + bias_edge_type_embedding, + bias_edge_type_embedding + ]), + srcs_embedding - multiplicative_edge_type_embedding*Dot(axes=-1)([ + multiplicative_edge_type_embedding, + srcs_embedding + ]) + bias_edge_type_embedding, + dsts_embedding - multiplicative_edge_type_embedding*Dot(axes=-1)([ + multiplicative_edge_type_embedding, + dsts_embedding + ]), + not_srcs_embedding - multiplicative_edge_type_embedding*Dot(axes=-1)([ + multiplicative_edge_type_embedding, + not_srcs_embedding + ]) + bias_edge_type_embedding, + not_dsts_embedding - multiplicative_edge_type_embedding*Dot(axes=-1)([ + multiplicative_edge_type_embedding, + not_dsts_embedding + ]) + ) + + @classmethod + def model_name(cls) -> str: + """Returns name of the current model.""" + return "TransH" + + @classmethod + def requires_edge_types(cls) -> bool: + return True + + def _extract_embeddings( + self, + graph: Graph, + model: Model, + return_dataframe: bool + ) -> EmbeddingResult: + """Returns embedding from the model. + + Parameters + ------------------ + graph: Graph + The graph that was embedded. + model: Model + The Keras model used to embed the graph. + return_dataframe: bool + Whether to return a dataframe of a numpy array. + """ + node_embedding = self.get_layer_weights( + "NodeEmbedding", + model, + drop_first_row=False + ) + bias_edge_type_embedding = self.get_layer_weights( + "BiasEdgeTypeEmbedding", + model, + drop_first_row=graph.has_unknown_edge_types() + ) + multiplicative_edge_type_embedding = self.get_layer_weights( + "MultiplicativeEdgeTypeEmbedding", + model, + drop_first_row=graph.has_unknown_edge_types() + ) + + if return_dataframe: + node_embedding = pd.DataFrame( + node_embedding, + index=graph.get_node_names() + ) + + edge_type_names = graph.get_unique_edge_type_names() + + bias_edge_type_embedding = pd.DataFrame( + bias_edge_type_embedding, + index=edge_type_names + ) + + multiplicative_edge_type_embedding = pd.DataFrame( + multiplicative_edge_type_embedding, + index=edge_type_names + ) + + return EmbeddingResult( + embedding_method_name=self.model_name(), + node_embeddings=node_embedding, + edge_type_embeddings=[ + bias_edge_type_embedding, + multiplicative_edge_type_embedding + ] + ) + + @classmethod + def can_use_node_types(cls) -> bool: + """Returns whether the model can optionally use node types.""" + return False + + @classmethod + def requires_edge_types(cls) -> bool: + """Returns whether the model requires edge types.""" + return True \ No newline at end of file diff --git a/embiggen/embedders/tensorflow_embedders/unstructured.py b/embiggen/embedders/tensorflow_embedders/unstructured.py new file mode 100644 index 00000000..20794f20 --- /dev/null +++ b/embiggen/embedders/tensorflow_embedders/unstructured.py @@ -0,0 +1,77 @@ +"""Unstructured model.""" +from ensmallen import Graph +import pandas as pd +import tensorflow as tf +from tensorflow.keras import Model +from embiggen.embedders.tensorflow_embedders.siamese import Siamese +from embiggen.utils.abstract_models import EmbeddingResult + + +class UnstructuredTensorFlow(Siamese): + """Unstructured model.""" + + def _build_output( + self, + srcs_embedding: tf.Tensor, + dsts_embedding: tf.Tensor, + not_srcs_embedding: tf.Tensor, + not_dsts_embedding: tf.Tensor, + graph: Graph + ): + """Returns the five input tensors, unchanged.""" + return ( + None, + 0.0, + srcs_embedding, + dsts_embedding, + not_srcs_embedding, + not_dsts_embedding + ) + + @classmethod + def model_name(cls) -> str: + """Returns name of the current model.""" + return "Unstructured" + + @classmethod + def can_use_edge_types(cls) -> bool: + return False + + def _extract_embeddings( + self, + graph: Graph, + model: Model, + return_dataframe: bool + ) -> EmbeddingResult: + """Returns embedding from the model. + + Parameters + ------------------ + graph: Graph + The graph that was embedded. + model: Model + The Keras model used to embed the graph. + return_dataframe: bool + Whether to return a dataframe of a numpy array. + """ + node_embedding = self.get_layer_weights( + "NodeEmbedding", + model, + drop_first_row=False + ) + + if return_dataframe: + node_embedding = pd.DataFrame( + node_embedding, + index=graph.get_node_names() + ) + + return EmbeddingResult( + embedding_method_name=self.model_name(), + node_embeddings=node_embedding, + ) + + @classmethod + def can_use_node_types(cls) -> bool: + """Returns whether the model can optionally use node types.""" + return False diff --git a/embiggen/embedding_transformers/__init__.py b/embiggen/embedding_transformers/__init__.py index cd8953ec..a465e9f4 100644 --- a/embiggen/embedding_transformers/__init__.py +++ b/embiggen/embedding_transformers/__init__.py @@ -10,7 +10,6 @@ "EdgeTransformer", "NodeTransformer", "GraphTransformer", - "CorpusTransformer", "EdgePredictionTransformer", "EdgeLabelPredictionTransformer", "NodeLabelPredictionTransformer" diff --git a/embiggen/embedding_transformers/edge_label_prediction_transformer.py b/embiggen/embedding_transformers/edge_label_prediction_transformer.py index dd3de564..4b2210ac 100644 --- a/embiggen/embedding_transformers/edge_label_prediction_transformer.py +++ b/embiggen/embedding_transformers/edge_label_prediction_transformer.py @@ -14,7 +14,7 @@ class EdgeLabelPredictionTransformer: def __init__( self, method: str = "Hadamard", - aligned_node_mapping: bool = False, + aligned_mapping: bool = False, one_hot_encode_labels: bool = False ): """Create new EdgeLabelPredictionTransformer object. @@ -24,7 +24,7 @@ def __init__( method: str = "hadamard" Method to use for the embedding. Can either be 'Hadamard', 'Sum', 'Average', 'L1', 'AbsoluteL1', 'L2' or 'Concatenate'. - aligned_node_mapping: bool = False + aligned_mapping: bool = False This parameter specifies whether the mapping of the embeddings nodes matches the internal node mapping of the given graph. If these two mappings do not match, the generated edge embedding @@ -34,31 +34,39 @@ def __init__( """ self._transformer = GraphTransformer( method=method, - aligned_node_mapping=aligned_node_mapping, + aligned_mapping=aligned_mapping, ) self._one_hot_encode_labels = one_hot_encode_labels - def fit(self, node_feature: Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]): + def fit( + self, + node_feature: Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]], + node_type_feature: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, + ): """Fit the model. Parameters ------------------------- node_feature: Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]], Node feature to use to fit the transformer. + node_type_feature: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None + Node type feature to use to fit the transformer. Raises ------------------------- ValueError If the given method is None there is no need to call the fit method. """ - self._transformer.fit(node_feature) + self._transformer.fit( + node_feature, + node_type_feature=node_type_feature + ) def transform( self, graph: Graph, edge_features: Optional[Union[np.ndarray, List[np.ndarray]]] = None, behaviour_for_unknown_edge_labels: Optional[str] = None, - random_state: int = 42 ) -> Tuple[np.ndarray, np.ndarray]: """Return edge embedding for given graph using provided method. @@ -79,8 +87,6 @@ def transform( By default, we drop these edges. If the behaviour has not been specified and left to None, a warning will be raised to notify the user of this uncertainty. - random_state: int = 42, - The random state to use to shuffle the labels. Raises -------------------------- @@ -104,6 +110,7 @@ def transform( "The provided graph for the edge-label prediction does " "not contain edge-types." ) + if not graph.has_known_edge_types(): raise ValueError( "The provided graph for the edge-label prediction does " @@ -111,15 +118,18 @@ def transform( "an edge type vocabulary but no edge has an edge type " "assigned to it." ) + if graph.has_homogeneous_edge_types(): raise ValueError( "The provided graph for the edge-label prediction contains " "edges of a single type, making predictions pointless." ) + if graph.is_multigraph(): raise NotImplementedError( "Multi-graphs are not currently supported by this class." ) + if graph.has_singleton_edge_types(): warnings.warn( "Please do be advised that this graph contains edges with " @@ -137,6 +147,7 @@ def transform( edge_type_counts.items(), key=lambda x: x[1] ) + if most_common_count > least_common_count * 20: warnings.warn( ( @@ -163,6 +174,7 @@ def transform( edge_embeddings = self._transformer.transform( graph, + node_types=graph, edge_features=edge_features ) @@ -176,12 +188,6 @@ def transform( edge_labels = graph.get_known_edge_type_ids() edge_embeddings = edge_embeddings[graph.get_edges_with_known_edge_types_mask()] else: - edge_labels = graph.get_edge_types_ids() - - numpy_random_state = np.random.RandomState( # pylint: disable=no-member - seed=random_state - ) - - indices = numpy_random_state.permutation(edge_labels.size) + edge_labels = graph.get_directed_edge_type_ids() - return edge_embeddings[indices], edge_labels[indices] + return edge_embeddings, edge_labels diff --git a/embiggen/embedding_transformers/edge_prediction_transformer.py b/embiggen/embedding_transformers/edge_prediction_transformer.py index ded2da1c..4e911089 100644 --- a/embiggen/embedding_transformers/edge_prediction_transformer.py +++ b/embiggen/embedding_transformers/edge_prediction_transformer.py @@ -13,7 +13,7 @@ class EdgePredictionTransformer: def __init__( self, method: str = "Hadamard", - aligned_node_mapping: bool = False, + aligned_mapping: bool = False, ): """Create new EdgePredictionTransformer object. @@ -22,7 +22,7 @@ def __init__( method: str = "hadamard", Method to use for the embedding. Can either be 'Hadamard', 'Sum', 'Average', 'L1', 'AbsoluteL1', 'L2' or 'Concatenate'. - aligned_node_mapping: bool = False, + aligned_mapping: bool = False, This parameter specifies whether the mapping of the embeddings nodes matches the internal node mapping of the given graph. If these two mappings do not match, the generated edge embedding @@ -30,23 +30,32 @@ def __init__( """ self._transformer = GraphTransformer( method=method, - aligned_node_mapping=aligned_node_mapping, + aligned_mapping=aligned_mapping, ) - def fit(self, node_feature: Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]): + def fit( + self, + node_feature: Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]], + node_type_feature: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, + ): """Fit the model. Parameters ------------------------- node_feature: Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]], Node feature to use to fit the transformer. + node_type_feature: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None + Node type feature to use to fit the transformer. Raises ------------------------- ValueError If the given method is None there is no need to call the fit method. """ - self._transformer.fit(node_feature) + self._transformer.fit( + node_feature, + node_type_feature=node_type_feature + ) def transform( self, @@ -116,10 +125,12 @@ def transform( positive_edge_embedding = self._transformer.transform( positive_graph, + node_types=positive_graph, edge_features=positive_edge_features ) negative_edge_embedding = self._transformer.transform( negative_graph, + node_types=negative_graph, edge_features=negative_edge_features ) diff --git a/embiggen/embedding_transformers/edge_transformer.py b/embiggen/embedding_transformers/edge_transformer.py index 3470645c..a9967d56 100644 --- a/embiggen/embedding_transformers/edge_transformer.py +++ b/embiggen/embedding_transformers/edge_transformer.py @@ -3,7 +3,7 @@ import numpy as np import pandas as pd -from userinput.utils import closest +from userinput.utils import must_be_in_set from ensmallen import express_measures from embiggen.utils.abstract_models import format_list from embiggen.embedding_transformers.node_transformer import NodeTransformer @@ -308,18 +308,6 @@ def get_max_edge_embedding( axis=0 ) - -def get_indices_edge_embedding( - source_node_embedding: np.ndarray, - destination_node_embedding: np.ndarray -) -> np.ndarray: - """Placeholder method.""" - return np.vstack(( - source_node_embedding, - destination_node_embedding - )).T - - class EdgeTransformer: """EdgeTransformer class to convert edges to edge embeddings.""" @@ -336,13 +324,12 @@ class EdgeTransformer: "Max": get_max_edge_embedding, "L2Distance": get_l2_distance, "CosineSimilarity": get_cosine_similarity, - None: get_indices_edge_embedding, } def __init__( self, method: str = "Hadamard", - aligned_node_mapping: bool = False, + aligned_mapping: bool = False, ): """Create new EdgeTransformer object. @@ -353,78 +340,53 @@ def __init__( If None is used, we return instead the numeric tuples. Can either be 'Hadamard', 'Min', 'Max', 'Sum', 'Average', 'L1', 'AbsoluteL1', 'SquaredL2', 'L2' or 'Concatenate'. - aligned_node_mapping: bool = False, + aligned_mapping: bool = False, This parameter specifies whether the mapping of the embeddings nodes matches the internal node mapping of the given graph. If these two mappings do not match, the generated edge embedding will be meaningless. """ - if isinstance(method, str) and method.lower() not in [ - None if method_name is None else method_name.lower() - for method_name in EdgeTransformer.methods - ]: - raise ValueError(( - "Given method '{}' is not supported. " - "Supported methods are {}, or alternatively a lambda. " - "Maybe you meant {}?" - ).format( - method, - format_list( - [method for method in EdgeTransformer.methods.keys() if method is not None]), - closest(method, [ - method_name - for method_name in EdgeTransformer.methods - if method_name is not None - ]) - )) + method = must_be_in_set( + method, + self.methods, + "edge embedding method" + ) self._transformer = NodeTransformer( - numeric_node_ids=method is None, - aligned_node_mapping=aligned_node_mapping, + aligned_mapping=aligned_mapping, ) self._method_name = method - if self._method_name is None: - self._method = EdgeTransformer.methods[None] - else: - self._method = { - None if method_name is None else method_name.lower(): callback - for method_name, callback in EdgeTransformer.methods.items() - }[self._method_name.lower()] - - @property - def numeric_node_ids(self) -> bool: - """Return whether the transformer returns numeric node IDs.""" - return self._transformer.numeric_node_ids + self._method = self.methods[method] @property def method(self) -> str: """Return the used edge embedding method.""" return self._method_name - def fit(self, node_feature: Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]): + def fit( + self, + node_feature: Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]], + node_type_feature: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, + ): """Fit the model. Parameters ------------------------- node_feature: Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]], Node feature to use to fit the transformer. - - Raises - ------------------------- - ValueError - If the given method is None there is no need to call the fit method. + node_type_feature: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None + Node type feature to use to fit the transformer. """ - if self._method is None: - raise ValueError( - "There is no need to call the fit when edge method is None, " - "as the transformer will exclusively return the numeric node " - "indices and not any node feature." - ) - self._transformer.fit(node_feature) + self._transformer.fit( + node_feature, + node_type_feature=node_type_feature + ) def transform( self, sources: Union[List[str], List[int]], destinations: Union[List[str], List[int]], + source_node_types: Optional[Union[List[Optional[List[str]]], List[Optional[List[int]]]]] = None, + destination_node_types: Optional[Union[List[Optional[List[str]]], List[Optional[List[int]]]]] = None, edge_features: Optional[Union[np.ndarray, List[np.ndarray]]] = None, ) -> np.ndarray: """Return embedding for given edges using provided method. @@ -435,6 +397,16 @@ def transform( List of source nodes whose embedding is to be returned. destinations:Union[List[str], List[int]] List of destination nodes whose embedding is to be returned. + source_node_types: Optional[Union[List[Optional[List[str]]], List[Optional[List[int]]]]] = None, + List of source node types whose embedding is to be returned. + This can be either a list of strings, or a graph, or if the + aligned_mapping is setted, then this methods also accepts + a list of ints. + destination_node_types: Optional[Union[List[Optional[List[str]]], List[Optional[List[int]]]]] = None, + List of destination node types whose embedding is to be returned. + This can be either a list of strings, or a graph, or if the + aligned_mapping is setted, then this methods also accepts + a list of ints. edge_features: Optional[Union[np.ndarray, List[np.ndarray]]] = None Optional edge features to be used as input concatenated to the obtained edge embedding. The shape must be equal @@ -470,15 +442,17 @@ def transform( "numpy arrays of type uint32, but you have provided objects of type " f"{sources.dtype} and {destinations.dtype}. " ) - if self._transformer._node_feature.dtype != np.float32: - self._transformer._node_feature = self._transformer._node_feature.astype( - np.float32 - ) if not self._transformer._node_feature.data.c_contiguous: self._transformer._node_feature = np.ascontiguousarray( self._transformer._node_feature ) + if self._transformer._node_type_feature is not None: + raise NotImplementedError( + "The node type features are not yet supported for the " + "Cosine Similarity." + ) + edge_embeddings = self._method( embedding=self._transformer._node_feature, source_node_ids=sources, @@ -486,8 +460,8 @@ def transform( ) else: edge_embeddings = self._method( - self._transformer.transform(sources), - self._transformer.transform(destinations) + self._transformer.transform(sources, node_types=source_node_types), + self._transformer.transform(destinations, node_types=destination_node_types) ) if edge_features is not None: diff --git a/embiggen/embedding_transformers/graph_transformer.py b/embiggen/embedding_transformers/graph_transformer.py index 7b314f65..a7595e0f 100644 --- a/embiggen/embedding_transformers/graph_transformer.py +++ b/embiggen/embedding_transformers/graph_transformer.py @@ -13,55 +13,69 @@ class GraphTransformer: def __init__( self, method: str = "Hadamard", - aligned_node_mapping: bool = False, + aligned_mapping: bool = False, + include_both_undirected_edges: bool = True ): """Create new GraphTransformer object. Parameters ------------------------ - method: str = "hadamard", + method: str = "hadamard" Method to use for the embedding. Can either be 'Hadamard', 'Sum', 'Average', 'L1', 'AbsoluteL1', 'L2' or 'Concatenate'. - aligned_node_mapping: bool = False, + aligned_mapping: bool = False This parameter specifies whether the mapping of the embeddings nodes matches the internal node mapping of the given graph. If these two mappings do not match, the generated edge embedding will be meaningless. + include_both_undirected_edges: bool = True + Whether to include both undirected edges when parsing an undirected + graph, that is both the edge from source to destination and the edge + from destination to source. While both edges should be included when + training a model, as the model should learn about these simmetries + in the graph, these edges are not necessary in the context of visualizations + where they create redoundancy. """ self._transformer = EdgeTransformer( method=method, - aligned_node_mapping=aligned_node_mapping, + aligned_mapping=aligned_mapping, ) - self._aligned_node_mapping = aligned_node_mapping - - @property - def numeric_node_ids(self) -> bool: - """Return whether the transformer returns numeric node IDs.""" - return self._transformer.numeric_node_ids + self._include_both_undirected_edges = include_both_undirected_edges + self._aligned_mapping = aligned_mapping @property def method(self) -> str: """Return the used edge embedding method.""" return self._transformer.method - def fit(self, node_feature: Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]): + def fit( + self, + node_feature: Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]], + node_type_feature: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, + ): """Fit the model. Parameters ------------------------- - node_feature: Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]], + node_feature: Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]] Node feature to use to fit the transformer. + node_type_feature: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None + Node type feature to use to fit the transformer. Raises ------------------------- ValueError If the given method is None there is no need to call the fit method. """ - self._transformer.fit(node_feature) + self._transformer.fit( + node_feature, + node_type_feature=node_type_feature + ) def transform( self, graph: Union[Graph, np.ndarray, List[List[str]], List[List[int]]], + node_types: Optional[Union[Graph, List[Optional[List[str]]], List[Optional[List[int]]]]] = None, edge_features: Optional[Union[np.ndarray, List[np.ndarray]]] = None, ) -> np.ndarray: """Return edge embedding for given graph using provided method. @@ -71,6 +85,11 @@ def transform( graph: Union[Graph, np.ndarray, List[List[str]], List[List[int]]], The graph whose edges are to embed. It can either be an Graph or a list of lists of edges. + node_types: Optional[Union[Graph, List[Optional[List[str]]], List[Optional[List[int]]]]] = None, + List of node types whose embedding is to be returned. + This can be either a list of strings, or a graph, or if the + aligned_mapping is setted, then this methods also accepts + a list of ints. edge_features: Optional[Union[np.ndarray, List[np.ndarray]]] = None Optional edge features to be used as input concatenated to the obtained edge embedding. The shape must be equal @@ -86,43 +105,85 @@ def transform( Numpy array of embeddings. """ if isinstance(graph, Graph): - if self._aligned_node_mapping: - graph = ( - graph.get_directed_source_node_ids(), - graph.get_directed_destination_node_ids(), - ) + if self._aligned_mapping: + if graph.is_directed() or self._include_both_undirected_edges: + edge_node_ids = ( + graph.get_directed_source_node_ids(), + graph.get_directed_destination_node_ids(), + ) + else: + edge_node_ids = ( + graph.get_source_node_ids(directed=False), + graph.get_destination_node_ids(directed=False), + ) else: - graph = graph.get_directed_edge_node_names() - if isinstance(graph, List): - graph = np.array(graph) + edge_node_ids = graph.get_directed_edge_node_names() + else: + edge_node_ids = graph + + if isinstance(edge_node_ids, List): + edge_node_ids = np.array(edge_node_ids) if ( - isinstance(graph, tuple) and - len(graph) == 2 and - all(isinstance(e, np.ndarray) for e in graph) + isinstance(edge_node_ids, tuple) and + len(edge_node_ids) == 2 and + all(isinstance(e, np.ndarray) for e in edge_node_ids) ): if ( - len(graph[0].shape) != 1 or - len(graph[1].shape) != 1 or - graph[0].shape[0] == 0 or - graph[1].shape[0] == 0 or - graph[0].shape[0] != graph[1].shape[0] + len(edge_node_ids[0].shape) != 1 or + len(edge_node_ids[1].shape) != 1 or + edge_node_ids[0].shape[0] == 0 or + edge_node_ids[1].shape[0] == 0 or + edge_node_ids[0].shape[0] != edge_node_ids[1].shape[0] ): raise ValueError( "When providing a tuple of numpy arrays containing the source and destination " "node IDs, we expect to receive two arrays both with shape " "with shape (number of edges,). " - f"The ones you have provided have shapes {graph[0].shape} " - f"and {graph[1].shape}." + f"The ones you have provided have shapes {edge_node_ids[0].shape} " + f"and {edge_node_ids[1].shape}." ) - sources = graph[0] - destinations = graph[1] - elif isinstance(graph, np.ndarray): - if len(graph.shape) != 2 or graph.shape[1] != 2 or graph.shape[0] == 0: + sources = edge_node_ids[0] + destinations = edge_node_ids[1] + elif isinstance(edge_node_ids, np.ndarray): + if len(edge_node_ids.shape) != 2 or edge_node_ids.shape[1] != 2 or edge_node_ids.shape[0] == 0: raise ValueError( "When providing a numpy array containing the source and destination " "node IDs representing the graph edges, we expect to receive an array " - f"with shape (number of edges, 2). The one you have provided has shape {graph.shape}." + f"with shape (number of edges, 2). The one you have provided has shape {edge_node_ids.shape}." ) - sources = graph[:, 0] - destinations = graph[:, 1] - return self._transformer.transform(sources, destinations, edge_features=edge_features) + sources = edge_node_ids[:, 0] + destinations = edge_node_ids[:, 1] + + if node_types is not None and self._transformer._transformer._node_type_feature is not None: + if isinstance(node_types, Graph): + if self._aligned_mapping: + source_node_types = [ + node_types.get_node_type_ids_from_node_id(src) + for src in sources + ] + destination_node_types = [ + node_types.get_node_type_ids_from_node_id(dst) + for dst in destinations + ] + else: + source_node_types = [ + node_types.get_node_type_names_from_node_name(src) + for src in sources + ] + destination_node_types = [ + node_types.get_node_type_names_from_node_name(dst) + for dst in destinations + ] + else: + source_node_types, destination_node_types = node_types + else: + source_node_types = None + destination_node_types = None + + return self._transformer.transform( + sources, + destinations, + source_node_types=source_node_types, + destination_node_types=destination_node_types, + edge_features=edge_features + ) diff --git a/embiggen/embedding_transformers/node_label_prediction_transformer.py b/embiggen/embedding_transformers/node_label_prediction_transformer.py index 0aef517c..02b8ee26 100644 --- a/embiggen/embedding_transformers/node_label_prediction_transformer.py +++ b/embiggen/embedding_transformers/node_label_prediction_transformer.py @@ -13,20 +13,20 @@ class NodeLabelPredictionTransformer: def __init__( self, - aligned_node_mapping: bool = False, + aligned_mapping: bool = False, ): """Create new NodeLabelPredictionTransformer object. Parameters ------------------------ - aligned_node_mapping: bool = False + aligned_mapping: bool = False This parameter specifies whether the mapping of the embeddings nodes matches the internal node mapping of the given graph. If these two mappings do not match, the generated node embedding will be meaningless. """ self._transformer = NodeTransformer( - aligned_node_mapping=aligned_node_mapping, + aligned_mapping=aligned_mapping, ) def fit(self, node_feature: Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]): @@ -86,24 +86,25 @@ def transform( """ if not graph.has_node_types(): raise ValueError( - "The provided graph for the node-label prediction does " + f"The provided graph {graph.get_name()} " + "for the node-label prediction does " "not contain node-types." ) if not graph.has_known_node_types(): raise ValueError( - "The provided graph for the node-label prediction does " + "fThe provided graph {graph.get_name()} for the node-label prediction does " "not contain known node-types, that is, it contains " "an node type vocabulary but no node has an node type " "assigned to it." ) if graph.has_homogeneous_node_types(): raise ValueError( - "The provided graph for the node-label prediction contains " + f"The provided graph {graph.get_name()} for the node-label prediction contains " "nodes of a single type, making predictions pointless." ) if graph.has_singleton_node_types(): warnings.warn( - "Please do be advised that this graph contains nodes with " + f"Please do be advised that the {graph.get_name()} graph contains nodes with " "a singleton node type, that is an node type that appears " "only once in the graph. Predictions on such rare node types " "will be unlikely to generalize well." diff --git a/embiggen/embedding_transformers/node_transformer.py b/embiggen/embedding_transformers/node_transformer.py index 1424c6d3..f6444795 100644 --- a/embiggen/embedding_transformers/node_transformer.py +++ b/embiggen/embedding_transformers/node_transformer.py @@ -1,105 +1,134 @@ """NodeTransformer class to convert nodes to edge embeddings.""" -from typing import List, Union +from typing import List, Union, Optional import numpy as np import pandas as pd from ensmallen import Graph + class NodeTransformer: """NodeTransformer class to convert nodes to edge embeddings.""" def __init__( self, - numeric_node_ids: bool = False, - aligned_node_mapping: bool = False, + aligned_mapping: bool = False, ): """Create new NodeTransformer object. Parameters ------------------- - numeric_node_ids: bool = False, - Wether to return the numeric node IDs instead of the node embedding. - aligned_node_mapping: bool = False, + aligned_mapping: bool = False, This parameter specifies whether the mapping of the embeddings nodes matches the internal node mapping of the given graph. If these two mappings do not match, the generated edge embedding will be meaningless. """ - self._numeric_node_ids = numeric_node_ids self._node_feature = None - self._aligned_node_mapping = aligned_node_mapping - - @property - def numeric_node_ids(self) -> bool: - """Return whether the transformer returns numeric node IDs.""" - return self._numeric_node_ids + self._node_type_feature = None + self._aligned_mapping = aligned_mapping - def fit(self, node_feature: Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]): + def fit( + self, + node_feature: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, + node_type_feature: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None, + ): """Fit the model. Parameters ------------------------- - node_feature: Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]], + node_feature: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None Node feature to use to fit the transformer. + node_type_feature: Optional[Union[pd.DataFrame, np.ndarray, List[Union[pd.DataFrame, np.ndarray]]]] = None + Node type feature to use to fit the transformer. """ - if not isinstance(node_feature, list): + if node_feature is not None and not isinstance(node_feature, list): node_feature = [node_feature] - if len(node_feature) == 0: + if node_type_feature is not None and not isinstance(node_type_feature, list): + node_type_feature = [node_type_feature] + + if (node_feature is None or len(node_feature) == 0) and (node_type_feature is None or len(node_type_feature) == 0): raise ValueError( "The provided list of features is empty!" ) # We check if any of the provided node features # is neither a numpy array nor a pandas dataframe. - for nf in node_feature: - if not isinstance(nf, (pd.DataFrame, np.ndarray)): - raise ValueError( - ( - "One of the provided node features is not " - "neither a pandas DataFrame nor a numpy array, but " - "of type {node_feature_type}. It is not clear " - "what to do with this feature." - ).format( - node_feature_type=type(node_feature) + for features, feature_name in ( + (node_type_feature, "node type"), + (node_feature, "node"), + ): + if features is None: + continue + for nf in features: + if not isinstance(nf, (pd.DataFrame, np.ndarray)): + raise ValueError( + ( + f"One of the provided {feature_name} features is not " + "neither a pandas DataFrame nor a numpy array, but " + f"of type {type(nf)}. It is not clear " + "what to do with this feature." + ) ) - ) - # We check if, while the parameters for alignment - # has not been provided, numpy arrays were provided. - # This would be an issue as we cannot check for alignment - # in numpy arrays. - if not self._aligned_node_mapping and any( - isinstance(nf, np.ndarray) - for nf in node_feature - ): - raise ValueError( - "A numpy array feature was provided while the " - "aligned node mapping parameter was set to false. " - "If you intend to specify that you are providing a numpy " - "array node feature that is aligned with the node vocabulary " - "of the graph set the `aligned_node_mapping` parameter " - "to True." - ) + # We check if, while the parameters for alignment + # has not been provided, numpy arrays were provided. + # This would be an issue as we cannot check for alignment + # in numpy arrays. + if not self._aligned_mapping and any( + isinstance(nf, np.ndarray) + for nf in features + ): + raise ValueError( + "A numpy array feature was provided while the " + f"aligned mapping parameter was set to false. " + "If you intend to specify that you are providing a numpy " + f"array {feature_name} feature that is aligned with the vocabulary " + "of the graph set the `aligned_mapping` parameter " + "to True." + ) - if self._aligned_node_mapping: - self._node_feature = np.hstack([ - nf.to_numpy() if isinstance(nf, pd.DataFrame) else nf - for nf in node_feature - ]) - if not self._node_feature.data.c_contiguous: - self._node_feature = np.ascontiguousarray(self._node_feature) + if self._aligned_mapping: + if node_feature is not None: + if len(node_feature) > 1: + self._node_feature = np.hstack([ + nf.to_numpy() if isinstance(nf, pd.DataFrame) else nf + for nf in node_feature + ]) + else: + self._node_feature = node_feature[0] + + if node_type_feature is not None: + if len(node_type_feature) > 1: + self._node_type_feature = np.hstack([ + nf.to_numpy() if isinstance(nf, pd.DataFrame) else nf + for nf in node_type_feature + ]) + else: + self._node_type_feature = node_type_feature[0] else: - self._node_feature = pd.concat(node_feature, axis=1) + if node_feature is not None: + self._node_feature = pd.concat(node_feature, axis=1) + if node_type_feature is not None: + self._node_type_feature = pd.concat(node_type_feature, axis=1) - def transform(self, nodes: Union[Graph, List[str], List[int]]) -> np.ndarray: + def transform( + self, + nodes: Optional[Union[Graph, List[str], List[int]]] = None, + node_types: Optional[Union[Graph, List[Optional[List[str]]], List[Optional[List[int]]]]] = None, + ) -> np.ndarray: """Return embeddings from given node. Parameters -------------------------- - nodes: Union[List[str], List[int]], + nodes: Optional[Union[List[str], List[int]]] = None List of nodes whose embedding is to be returned. - By default this should be a list of strings, if the - aligned_node_mapping is setted, then this methods also accepts + This can be either a list of strings, or a graph, or if the + aligned_mapping is setted, then this methods also accepts + a list of ints. + node_types: Optional[Union[Graph, List[Optional[List[str]]], List[Optional[List[int]]]]] = None, + List of node types whose embedding is to be returned. + This can be either a list of strings, or a graph, or if the + aligned_mapping is setted, then this methods also accepts a list of ints. Raises @@ -111,22 +140,68 @@ def transform(self, nodes: Union[Graph, List[str], List[int]]) -> np.ndarray: -------------------------- Numpy array of embeddings. """ - if self._node_feature is None and not self.numeric_node_ids: + if self._node_feature is None and self._node_type_feature is None: raise ValueError( "Transformer was not fitted yet." ) - if self._aligned_node_mapping: - if self.numeric_node_ids: - return nodes - if isinstance(nodes, Graph): - return self._node_feature - return self._node_feature[nodes] + node_type_features = None + node_features = None - if self.numeric_node_ids: - return self._node_feature.index.get_indexer(nodes) - - if isinstance(nodes, Graph): - nodes = nodes.get_node_names() + if self._aligned_mapping: + if nodes is not None and self._node_feature is not None: + if not isinstance(nodes, (np.ndarray, Graph)): + raise ValueError( + "The provided nodes are not numpy array and the " + "node IDs or Graph are expected to be aligned." + ) + + if isinstance(nodes, Graph): + node_features = self._node_feature + else: + node_features = self._node_feature[nodes] + + if node_types is not None and self._node_type_feature is not None: + if isinstance(node_types, Graph): + node_types = node_types.get_node_type_ids() + node_type_feature_dimensionality = self._node_type_feature.shape[1] + node_type_features = np.vstack([ + np.mean( + self._node_type_feature[node_type_ids], + axis=0 + ) + if node_type_ids is not None + else + np.zeros(shape=node_type_feature_dimensionality) + for node_type_ids in node_types + ]) + else: + if nodes is not None and self._node_feature is not None: + if isinstance(nodes, Graph): + nodes = nodes.get_node_names() + + node_features = self._node_feature.loc[nodes].to_numpy() + + if node_types is not None and self._node_type_feature is not None: + node_type_feature_dimensionality = self._node_type_feature.shape[1] + node_type_features = np.vstack([ + np.mean( + self._node_feature.loc[node_type_names].to_numpy(), + axis=0 + ) + if node_type_names is not None + else + np.zeros(shape=node_type_feature_dimensionality) + for node_type_names in node_types + ]) + + + if node_features is None: + node_features = node_type_features + elif node_type_features is not None : + node_features = np.hstack([ + node_features, + node_type_features + ]) - return self._node_feature.loc[nodes].to_numpy() + return node_features diff --git a/embiggen/layers/__init__.py b/embiggen/layers/__init__.py index e69de29b..b680692d 100644 --- a/embiggen/layers/__init__.py +++ b/embiggen/layers/__init__.py @@ -0,0 +1,1 @@ +__all__ = [] \ No newline at end of file diff --git a/embiggen/layers/tensorflow/graph_convolution_layer.py b/embiggen/layers/tensorflow/graph_convolution_layer.py index a17817b9..06a51eb0 100644 --- a/embiggen/layers/tensorflow/graph_convolution_layer.py +++ b/embiggen/layers/tensorflow/graph_convolution_layer.py @@ -7,6 +7,7 @@ """ from typing import Tuple, Union, Dict, Optional, List import tensorflow as tf +from userinput.utils import must_be_in_set from tensorflow.python.ops import embedding_ops # pylint: disable=import-error,no-name-in-module from tensorflow.keras.layers import Dropout, Layer, Dense # pylint: disable=import-error,no-name-in-module from embiggen.layers.tensorflow.l2_norm import L2Norm @@ -15,10 +16,13 @@ class GraphConvolution(Layer): """Layer implementing graph convolution layer.""" + available_combiner = ["mean", "sqrtn", "sum"] + def __init__( self, units: int, activation: str = "relu", + combiner: str = "mean", dropout_rate: Optional[float] = 0.5, apply_norm: bool = False, **kwargs: Dict @@ -27,18 +31,28 @@ def __init__( Parameters ---------------------- - units: int, + units: int The dimensionality of the output space (i.e. the number of output units). - activation: str = "relu", + activation: str = "relu" Activation function to use. If you don't specify anything, relu is used. - dropout_rate: Optional[float] = 0.5, + combiner: str = "mean" + A string specifying the reduction op. + Currently "mean", "sqrtn" and "sum" are supported. + "sum" computes the weighted sum of the embedding results for each row. + "mean" is the weighted sum divided by the total weight. + "sqrtn" is the weighted sum divided by the square root of the sum of the squares of the weights. + Defaults to mean. + dropout_rate: Optional[float] = 0.5 Float between 0 and 1. Fraction of the input units to drop. If the provided value is either zero or None the dropout rate is not applied. - **kwargs: Dict, + **kwargs: Dict Kwargs to pass to the parent Layer class. """ super().__init__(**kwargs) + combiner = must_be_in_set( + combiner, self.available_combiner, "combiner") + self._combiner = combiner self._units = units self._activation = activation if dropout_rate == 0.0: @@ -46,8 +60,8 @@ def __init__( self._dropout_rate = dropout_rate self._apply_norm = apply_norm self._dense_layers = [] - self._dropout_layer = None - self._l2_norm = None + self._dropout_layers = [] + self._l2_norms = [] def build(self, input_shape) -> None: """Build the Graph Convolution layer. @@ -70,6 +84,7 @@ def build(self, input_shape) -> None: "has a single element. It should contain exactly two elements, " "the adjacency matrix and the node features." ) + for node_feature_shape in input_shape[1:]: dense_layer = Dense( units=self._units, @@ -78,17 +93,19 @@ def build(self, input_shape) -> None: dense_layer.build(node_feature_shape) self._dense_layers.append(dense_layer) - if self._dropout_rate is not None: - self._dropout_layer = Dropout(self._dropout_rate) - self._dropout_layer.build(node_feature_shape) - else: - self._dropout_layer = lambda x: x + if self._dropout_rate is not None: + dropout_layer = Dropout(self._dropout_rate) + dropout_layer.build(node_feature_shape) + else: + def dropout_layer(x): return x + self._dropout_layers.append(dropout_layer) - if self._apply_norm: - self._l2_norm = L2Norm() - self._l2_norm.build(node_feature_shape) - else: - self._l2_norm = lambda x: x + if self._apply_norm: + l2_norm = L2Norm() + l2_norm.build((self._units,)) + else: + def l2_norm(x): return x + self._l2_norms.append(l2_norm) super().build(input_shape) @@ -111,14 +128,16 @@ def call( ) return [ - self._l2_norm(dense(embedding_ops.embedding_lookup_sparse_v2( - self._dropout_layer(node_feature), + dense(embedding_ops.embedding_lookup_sparse_v2( + node_feature, ids, adjacency, - combiner='mean' - ))) - for dense, node_feature in zip( + combiner=self._combiner + )) + for dense, dropout, l2_norm, node_feature in zip( self._dense_layers, + self._dropout_layers, + self._l2_norms, node_features ) ] diff --git a/embiggen/layers/tensorflow/l2_norm.py b/embiggen/layers/tensorflow/l2_norm.py index 2ba0c728..101fdd6d 100644 --- a/embiggen/layers/tensorflow/l2_norm.py +++ b/embiggen/layers/tensorflow/l2_norm.py @@ -6,19 +6,18 @@ class L2Norm(Layer): - """Layer implementing simple wrapped embedding layer plus a flatten.""" + """Layer implementing L2 Norm.""" def __init__( self, **kwargs: Dict ): - """Create new GraphConvolution layer. - """ + """Create new L2 Norm layer.""" super().__init__(**kwargs) self._norm_layer = None def build(self, input_shape) -> None: - """Build the Graph Convolution layer. + """Build the L2 Norm layer. Parameters ------------------------------ diff --git a/embiggen/node_label_prediction/node_label_prediction_model.py b/embiggen/node_label_prediction/node_label_prediction_model.py index 61a95880..0b02c4d3 100644 --- a/embiggen/node_label_prediction/node_label_prediction_model.py +++ b/embiggen/node_label_prediction/node_label_prediction_model.py @@ -4,7 +4,7 @@ import numpy as np import warnings from ensmallen import Graph -from embiggen.utils.abstract_models import AbstractClassifierModel, AbstractEmbeddingModel, abstract_class, format_list +from embiggen.utils.abstract_models import AbstractClassifierModel, abstract_class @abstract_class @@ -95,10 +95,13 @@ def split_graph_following_evaluation_schema( use_stratification="Stratified" in evaluation_schema, random_state=random_state, ) - raise ValueError( - f"The requested evaluation schema `{evaluation_schema}` " - "is not available. The available evaluation schemas " - f"are: {format_list(cls.get_available_evaluation_schemas())}." + super().split_graph_following_evaluation_schema( + graph=graph, + evaluation_schema=evaluation_schema, + random_state=random_state, + holdout_number=holdout_number, + number_of_holdouts=number_of_holdouts, + **holdouts_kwargs, ) @classmethod @@ -130,7 +133,8 @@ def _evaluate( verbose: bool = True, ) -> List[Dict[str, Any]]: """Return model evaluation on the provided graphs.""" - train_size = train.get_number_of_known_node_types() / graph.get_number_of_known_node_types() + train_size = train.get_number_of_known_node_types( + ) / graph.get_number_of_known_node_types() if self.is_multilabel_prediction_task(): labels = graph.get_one_hot_encoded_node_types() @@ -295,7 +299,19 @@ def fit( "of the node-label prediction models." ) - self._is_binary_prediction_task = graph.get_number_of_node_types() == 2 + non_zero_node_types = sum([ + 1 + for count in graph.get_node_type_names_counts_hashmap().values() + if count > 0 + ]) + + if non_zero_node_types < 2: + raise ValueError( + "The provided training graph has less than two non-zero node types. " + "It is unclear how to proceeed." + ) + + self._is_binary_prediction_task = non_zero_node_types == 2 self._is_multilabel_prediction_task = graph.has_multilabel_node_types() node_type_counts = graph.get_node_type_names_counts_hashmap() diff --git a/embiggen/node_label_prediction/node_label_prediction_sklearn/decision_tree_for_node_prediction.py b/embiggen/node_label_prediction/node_label_prediction_sklearn/decision_tree_for_node_prediction.py index 94af1a26..de76f9d6 100644 --- a/embiggen/node_label_prediction/node_label_prediction_sklearn/decision_tree_for_node_prediction.py +++ b/embiggen/node_label_prediction/node_label_prediction_sklearn/decision_tree_for_node_prediction.py @@ -11,15 +11,14 @@ def __init__( self, criterion="gini", splitter="best", - max_depth=None, + max_depth=10, min_samples_split=2, min_samples_leaf=1, min_weight_fraction_leaf=0., max_features=None, max_leaf_nodes=None, min_impurity_decrease=0., - min_impurity_split=None, - class_weight=None, + class_weight="balanced", ccp_alpha=0.0, random_state: int = 42 ): @@ -33,7 +32,6 @@ def __init__( self._max_features = max_features self._max_leaf_nodes = max_leaf_nodes self._min_impurity_decrease = min_impurity_decrease - self._min_impurity_split = min_impurity_split self._random_state = random_state self._class_weight = class_weight self._ccp_alpha = ccp_alpha @@ -49,7 +47,6 @@ def __init__( max_features=max_features, max_leaf_nodes=max_leaf_nodes, min_impurity_decrease=min_impurity_decrease, - min_impurity_split=min_impurity_split, random_state=random_state, class_weight=class_weight, ccp_alpha=ccp_alpha, @@ -71,7 +68,6 @@ def parameters(self) -> Dict[str, Any]: max_features=self._max_features, max_leaf_nodes=self._max_leaf_nodes, min_impurity_decrease=self._min_impurity_decrease, - min_impurity_split=self._min_impurity_split, random_state=self._random_state, class_weight=self._class_weight, ccp_alpha=self._ccp_alpha, diff --git a/embiggen/node_label_prediction/node_label_prediction_sklearn/extra_trees_node_label_prediction.py b/embiggen/node_label_prediction/node_label_prediction_sklearn/extra_trees_node_label_prediction.py index 7cbed2ed..a0a9c328 100644 --- a/embiggen/node_label_prediction/node_label_prediction_sklearn/extra_trees_node_label_prediction.py +++ b/embiggen/node_label_prediction/node_label_prediction_sklearn/extra_trees_node_label_prediction.py @@ -19,13 +19,12 @@ def __init__( max_features="auto", max_leaf_nodes=None, min_impurity_decrease=0., - min_impurity_split=None, bootstrap=True, oob_score=False, n_jobs=-1, verbose=0, warm_start=False, - class_weight=None, + class_weight="balanced", ccp_alpha=0.0, max_samples=None, random_state: int = 42 @@ -40,7 +39,6 @@ def __init__( self._max_features = max_features self._max_leaf_nodes = max_leaf_nodes self._min_impurity_decrease = min_impurity_decrease - self._min_impurity_split = min_impurity_split self._bootstrap = bootstrap self._oob_score = oob_score self._n_jobs = n_jobs @@ -61,7 +59,6 @@ def __init__( max_features=max_features, max_leaf_nodes=max_leaf_nodes, min_impurity_decrease=min_impurity_decrease, - min_impurity_split=min_impurity_split, bootstrap=bootstrap, oob_score=oob_score, n_jobs=n_jobs, @@ -97,7 +94,6 @@ def parameters(self) -> Dict[str, Any]: max_features = self._max_features, max_leaf_nodes = self._max_leaf_nodes, min_impurity_decrease = self._min_impurity_decrease, - min_impurity_split = self._min_impurity_split, bootstrap = self._bootstrap, oob_score = self._oob_score, n_jobs = self._n_jobs, diff --git a/embiggen/node_label_prediction/node_label_prediction_sklearn/gradient_boosting_node_label_prediction.py b/embiggen/node_label_prediction/node_label_prediction_sklearn/gradient_boosting_node_label_prediction.py index 5015f219..697ae628 100644 --- a/embiggen/node_label_prediction/node_label_prediction_sklearn/gradient_boosting_node_label_prediction.py +++ b/embiggen/node_label_prediction/node_label_prediction_sklearn/gradient_boosting_node_label_prediction.py @@ -19,7 +19,6 @@ def __init__( min_weight_fraction_leaf=0., max_depth=3, min_impurity_decrease=0., - min_impurity_split=None, init=None, max_features=None, verbose=0, @@ -34,20 +33,19 @@ def __init__( """Create the Gradient Boosting for node label Prediction.""" self._loss=loss self._learning_rate=learning_rate - self._n_estimators=n_estimators, + self._n_estimators=n_estimators self._criterion=criterion - self._min_samples_split=min_samples_split, - self._min_samples_leaf=min_samples_leaf, - self._min_weight_fraction_leaf=min_weight_fraction_leaf, + self._min_samples_split=min_samples_split + self._min_samples_leaf=min_samples_leaf + self._min_weight_fraction_leaf=min_weight_fraction_leaf self._max_depth=max_depth self._init=init - self._subsample=subsample, - self._max_features=max_features, - self._max_leaf_nodes=max_leaf_nodes, - self._min_impurity_decrease=min_impurity_decrease, - self._min_impurity_split=min_impurity_split, + self._subsample=subsample + self._max_features=max_features + self._max_leaf_nodes=max_leaf_nodes + self._min_impurity_decrease=min_impurity_decrease self._warm_start=warm_start - self._validation_fraction=validation_fraction, + self._validation_fraction=validation_fraction self._n_iter_no_change=n_iter_no_change self._tol=tol self._ccp_alpha=ccp_alpha @@ -62,7 +60,6 @@ def __init__( random_state=random_state, verbose=verbose, max_leaf_nodes=max_leaf_nodes, min_impurity_decrease=min_impurity_decrease, - min_impurity_split=min_impurity_split, warm_start=warm_start, validation_fraction=validation_fraction, n_iter_no_change=n_iter_no_change, tol=tol, ccp_alpha=ccp_alpha ), @@ -71,30 +68,27 @@ def __init__( def parameters(self) -> Dict[str, Any]: """Returns parameters used for this model.""" - return { + return dict( **super().parameters(), - **dict( - loss=self._loss, - learning_rate=self._learning_rate, - n_estimators=self._n_estimators, - criterion=self._criterion, - min_samples_split=self._min_samples_split, - min_samples_leaf=self._min_samples_leaf, - min_weight_fraction_leaf=self._min_weight_fraction_leaf, - max_depth=self._max_depth, - init=self._init, - subsample=self._subsample, - max_features=self._max_features, - max_leaf_nodes=self._max_leaf_nodes, - min_impurity_decrease=self._min_impurity_decrease, - min_impurity_split=self._min_impurity_split, - warm_start=self._warm_start, - validation_fraction=self._validation_fraction, - n_iter_no_change=self._n_iter_no_change, - tol=self._tol, - ccp_alpha=self._ccp_alpha - ) - } + loss=self._loss, + learning_rate=self._learning_rate, + n_estimators=self._n_estimators, + criterion=self._criterion, + min_samples_split=self._min_samples_split, + min_samples_leaf=self._min_samples_leaf, + min_weight_fraction_leaf=self._min_weight_fraction_leaf, + max_depth=self._max_depth, + init=self._init, + subsample=self._subsample, + max_features=self._max_features, + max_leaf_nodes=self._max_leaf_nodes, + min_impurity_decrease=self._min_impurity_decrease, + warm_start=self._warm_start, + validation_fraction=self._validation_fraction, + n_iter_no_change=self._n_iter_no_change, + tol=self._tol, + ccp_alpha=self._ccp_alpha + ) @classmethod def smoke_test_parameters(cls) -> Dict[str, Any]: diff --git a/embiggen/node_label_prediction/node_label_prediction_sklearn/random_forest_node_label_prediction.py b/embiggen/node_label_prediction/node_label_prediction_sklearn/random_forest_node_label_prediction.py index c8235af5..d00a77ab 100644 --- a/embiggen/node_label_prediction/node_label_prediction_sklearn/random_forest_node_label_prediction.py +++ b/embiggen/node_label_prediction/node_label_prediction_sklearn/random_forest_node_label_prediction.py @@ -17,16 +17,15 @@ def __init__( min_samples_split: int = 2, min_samples_leaf: int = 1, min_weight_fraction_leaf: float = 0., - max_features="auto", + max_features="sqrt", max_leaf_nodes=None, min_impurity_decrease=0., - min_impurity_split=None, bootstrap=True, oob_score=False, n_jobs=-1, verbose=0, warm_start=False, - class_weight=None, + class_weight="balanced", ccp_alpha=0.0, max_samples=None, random_state: int = 42 @@ -41,7 +40,6 @@ def __init__( self._max_features = max_features self._max_leaf_nodes = max_leaf_nodes self._min_impurity_decrease = min_impurity_decrease - self._min_impurity_split = min_impurity_split self._bootstrap = bootstrap self._oob_score = oob_score self._n_jobs = cpu_count() if n_jobs == -1 else n_jobs @@ -62,7 +60,6 @@ def __init__( max_features=max_features, max_leaf_nodes=max_leaf_nodes, min_impurity_decrease=min_impurity_decrease, - min_impurity_split=min_impurity_split, bootstrap=bootstrap, oob_score=oob_score, n_jobs=n_jobs, @@ -98,7 +95,6 @@ def parameters(self) -> Dict[str, Any]: max_features = self._max_features, max_leaf_nodes = self._max_leaf_nodes, min_impurity_decrease = self._min_impurity_decrease, - min_impurity_split = self._min_impurity_split, bootstrap = self._bootstrap, oob_score = self._oob_score, n_jobs = self._n_jobs, diff --git a/embiggen/node_label_prediction/node_label_prediction_sklearn/sklearn_node_label_prediction_adapter.py b/embiggen/node_label_prediction/node_label_prediction_sklearn/sklearn_node_label_prediction_adapter.py index d2c5375e..9b73f9f4 100644 --- a/embiggen/node_label_prediction/node_label_prediction_sklearn/sklearn_node_label_prediction_adapter.py +++ b/embiggen/node_label_prediction/node_label_prediction_sklearn/sklearn_node_label_prediction_adapter.py @@ -64,7 +64,7 @@ def _trasform_graph_into_node_embedding( ValueError If the two graphs do not share the same node vocabulary. """ - gt = NodeTransformer(aligned_node_mapping=True) + gt = NodeTransformer(aligned_mapping=True) gt.fit(node_features) return gt.transform(graph,) @@ -103,7 +103,7 @@ def _fit( If the two graphs do not share the same node vocabulary. """ nlpt = NodeLabelPredictionTransformer( - aligned_node_mapping=True + aligned_mapping=True ) nlpt.fit(node_features) @@ -155,10 +155,18 @@ def _predict_proba( )) if self.is_multilabel_prediction_task(): - return np.array([ - class_predictions[:, 1] - for class_predictions in predictions_probabilities - ]).T + if isinstance(predictions_probabilities, np.ndarray): + return predictions_probabilities + if isinstance(predictions_probabilities, list): + return np.array([ + class_predictions[:, 1] + for class_predictions in predictions_probabilities + ]).T + raise NotImplementedError( + f"The model {self.model_name()} from library {self.library_name()} " + f"returned an object of type {type(predictions_probabilities)} during " + "the execution of the predict proba method." + ) return predictions_probabilities diff --git a/embiggen/node_label_prediction/node_label_prediction_tensorflow/gcn.py b/embiggen/node_label_prediction/node_label_prediction_tensorflow/gcn.py index fdbefbf7..326b7e7a 100644 --- a/embiggen/node_label_prediction/node_label_prediction_tensorflow/gcn.py +++ b/embiggen/node_label_prediction/node_label_prediction_tensorflow/gcn.py @@ -28,6 +28,7 @@ def __init__( number_of_units_per_head_layer: Union[int, List[int]] = 128, dropout_rate: float = 0.1, apply_norm: bool = False, + combiner: str = "mean", optimizer: Union[str, Optimizer] = "adam", early_stopping_min_delta: float = 0.0001, early_stopping_patience: int = 30, @@ -68,6 +69,13 @@ def __init__( apply_norm: bool = False Whether to normalize the output of the convolution operations, after applying the level activations. + combiner: str = "mean" + A string specifying the reduction op. + Currently "mean", "sqrtn" and "sum" are supported. + "sum" computes the weighted sum of the embedding results for each row. + "mean" is the weighted sum divided by the total weight. + "sqrtn" is the weighted sum divided by the square root of the sum of the squares of the weights. + Defaults to mean. optimizer: str = "Adam" The optimizer to use while training the model. early_stopping_min_delta: float @@ -126,6 +134,7 @@ def __init__( number_of_units_per_graph_convolution_layers=number_of_units_per_graph_convolution_layers, dropout_rate=dropout_rate, apply_norm=apply_norm, + combiner=combiner, optimizer=optimizer, early_stopping_min_delta=early_stopping_min_delta, early_stopping_patience=early_stopping_patience, @@ -194,6 +203,8 @@ def _build_model( def get_output_classes(self, graph: Graph) -> int: """Returns number of output classes.""" + if self.is_binary_prediction_task(): + return 1 return graph.get_number_of_node_types() def _get_class_weights(self, graph: Graph) -> Dict[int, float]: @@ -249,7 +260,7 @@ def _get_model_training_sample_weights( graph: Graph, ) -> Optional[np.ndarray]: """Returns training output tuple.""" - return graph.get_known_node_types_mask().astype(tf.float32) + return graph.get_known_node_types_mask().astype(np.float32) def _get_model_prediction_input( self, diff --git a/embiggen/node_label_prediction/node_label_prediction_tensorflow/gnn.py b/embiggen/node_label_prediction/node_label_prediction_tensorflow/gnn.py index 82b18338..ca0aae1c 100644 --- a/embiggen/node_label_prediction/node_label_prediction_tensorflow/gnn.py +++ b/embiggen/node_label_prediction/node_label_prediction_tensorflow/gnn.py @@ -12,7 +12,6 @@ def __init__( epochs: int = 1000, number_of_head_layers: int = 1, number_of_units_per_head_layer: Union[int, List[int]] = 128, - dropout_rate: float = 0.1, optimizer: Union[str, Optimizer] = "adam", early_stopping_min_delta: float = 0.0001, early_stopping_patience: int = 30, @@ -45,9 +44,6 @@ def __init__( Number of graph convolution layer. number_of_units_per_hidden_layer: Union[int, List[int]] = 128 Number of units per hidden layer. - dropout_rate: float = 0.3 - Float between 0 and 1. - Fraction of the input units to dropout. optimizer: str = "Adam" The optimizer to use while training the model. early_stopping_min_delta: float @@ -95,7 +91,6 @@ def __init__( number_of_head_layers=number_of_head_layers, number_of_units_per_graph_convolution_layers=0, number_of_units_per_head_layer=number_of_units_per_head_layer, - dropout_rate=dropout_rate, optimizer=optimizer, early_stopping_min_delta=early_stopping_min_delta, early_stopping_patience=early_stopping_patience, @@ -138,13 +133,16 @@ def parameters(self) -> Dict[str, Any]: removed = [ "number_of_units_per_graph_convolution_layers", "handling_multi_graph", - "number_of_units_per_head_layer" + "number_of_units_per_head_layer", + "combiner", + "apply_norm", + "dropout_rate" ] return dict( number_of_units_per_head_layer=self._number_of_units_per_head_layer, **{ key: value - for key, value in super().smoke_test_parameters().items() + for key, value in super().parameters().items() if key not in removed } ) diff --git a/embiggen/node_label_prediction/node_label_prediction_tensorflow/graph_sage.py b/embiggen/node_label_prediction/node_label_prediction_tensorflow/graph_sage.py index 317e8207..c9763e18 100644 --- a/embiggen/node_label_prediction/node_label_prediction_tensorflow/graph_sage.py +++ b/embiggen/node_label_prediction/node_label_prediction_tensorflow/graph_sage.py @@ -110,6 +110,7 @@ def __init__( number_of_units_per_head_layer=number_of_units_per_head_layer, dropout_rate=dropout_rate, apply_norm=True, + combiner="mean", optimizer=optimizer, early_stopping_min_delta=early_stopping_min_delta, early_stopping_patience=early_stopping_patience, @@ -135,6 +136,7 @@ def parameters(self) -> Dict[str, Any]: """Returns parameters for smoke test.""" removed = [ "apply_norm", + "combiner", "use_simmetric_normalized_laplacian" ] return dict( diff --git a/embiggen/node_label_prediction/node_label_prediction_tensorflow/kipf_gcn.py b/embiggen/node_label_prediction/node_label_prediction_tensorflow/kipf_gcn.py index 0c7affa4..a74bdd69 100644 --- a/embiggen/node_label_prediction/node_label_prediction_tensorflow/kipf_gcn.py +++ b/embiggen/node_label_prediction/node_label_prediction_tensorflow/kipf_gcn.py @@ -1,5 +1,5 @@ """GCN model for node-label prediction.""" -from typing import List, Union, Optional +from typing import List, Union, Optional, Dict, Any from tensorflow.keras.optimizers import \ Optimizer # pylint: disable=import-error,no-name-in-module @@ -111,6 +111,7 @@ def __init__( number_of_units_per_head_layer=number_of_units_per_head_layer, dropout_rate=dropout_rate, apply_norm=apply_norm, + combiner="sum", optimizer=optimizer, early_stopping_min_delta=early_stopping_min_delta, early_stopping_patience=early_stopping_patience, @@ -132,6 +133,19 @@ def __init__( verbose=verbose, ) + def parameters(self) -> Dict[str, Any]: + """Returns parameters for smoke test.""" + removed = [ + "combiner", + ] + return dict( + **{ + key: value + for key, value in super().parameters().items() + if key not in removed + } + ) + @classmethod def model_name(cls) -> str: return "Kipf GCN" \ No newline at end of file diff --git a/embiggen/sequences/__init__.py b/embiggen/sequences/__init__.py index 6bb0d6ae..8ed70a22 100644 --- a/embiggen/sequences/__init__.py +++ b/embiggen/sequences/__init__.py @@ -1,1 +1,2 @@ -"""Module with Sequences, mostly for Keras models.""" \ No newline at end of file +"""Module with Sequences, mostly for Keras models.""" +__all__ = [] \ No newline at end of file diff --git a/embiggen/sequences/generic_sequences/edge_prediction_sequence.py b/embiggen/sequences/generic_sequences/edge_prediction_sequence.py index da5c5a61..108ad7e1 100644 --- a/embiggen/sequences/generic_sequences/edge_prediction_sequence.py +++ b/embiggen/sequences/generic_sequences/edge_prediction_sequence.py @@ -35,14 +35,6 @@ def __init__( batch_size: int = 2**10, The batch size to use. """ - if not graph.has_edges(): - raise ValueError( - f"An empty instance of graph {graph.get_name()} was provided!" - ) - if not graph.has_edges(): - raise ValueError( - f"An empty instance of graph {graph_used_in_training.get_name()} was provided!" - ) if not graph.has_compatible_node_vocabularies(graph_used_in_training): raise ValueError( f"The provided graph {graph.get_name()} does not have a node vocabulary " diff --git a/embiggen/sequences/tensorflow_sequences/__init__.py b/embiggen/sequences/tensorflow_sequences/__init__.py index bbfdb2cc..b4843399 100644 --- a/embiggen/sequences/tensorflow_sequences/__init__.py +++ b/embiggen/sequences/tensorflow_sequences/__init__.py @@ -4,19 +4,13 @@ from embiggen.sequences.tensorflow_sequences.gcn_edge_prediction_training_sequence import GCNEdgePredictionTrainingSequence from embiggen.sequences.tensorflow_sequences.gcn_edge_prediction_sequence import GCNEdgePredictionSequence from embiggen.sequences.tensorflow_sequences.gcn_edge_label_prediction_training_sequence import GCNEdgeLabelPredictionTrainingSequence -from embiggen.sequences.tensorflow_sequences.edge_prediction_sequence import EdgePredictionSequence from embiggen.sequences.tensorflow_sequences.siamese_sequence import SiameseSequence -from embiggen.sequences.tensorflow_sequences.kgsiamese_sequence import KGSiameseSequence __all__ = [ "Node2VecSequence", "EdgePredictionTrainingSequence", - "EdgePredictionSequence", "GCNEdgePredictionTrainingSequence", "GCNEdgePredictionSequence", "GCNEdgeLabelPredictionTrainingSequence", "SiameseSequence", - "KGSiameseSequence", - "EdgeJaccardSequence", - "EdgeAdamicAdarSequence", ] diff --git a/embiggen/sequences/tensorflow_sequences/edge_prediction_sequence.py b/embiggen/sequences/tensorflow_sequences/edge_prediction_sequence.py deleted file mode 100644 index 6145f82f..00000000 --- a/embiggen/sequences/tensorflow_sequences/edge_prediction_sequence.py +++ /dev/null @@ -1,147 +0,0 @@ -"""Keras Sequence for running Neural Network on graph edge prediction.""" -from typing import Tuple - -import numpy as np -import tensorflow as tf -from ensmallen import Graph # pylint: disable=no-name-in-module -from keras_mixed_sequence import Sequence -from embiggen.sequences.generic_sequences import EdgePredictionSequence as GenericEdgePredictionSequence -from embiggen.utils.tensorflow_utils import tensorflow_version_is_higher_or_equal_than - - -class EdgePredictionSequence(Sequence): - """Keras Sequence for running Neural Network on graph edge prediction.""" - - def __init__( - self, - graph: Graph, - graph_used_in_training: Graph, - use_node_types: bool, - use_edge_metrics: bool, - batch_size: int = 2**10, - ): - """Create new EdgePredictionSequence object. - - Parameters - -------------------------------- - graph: Graph - The graph whose edges are to be predicted. - graph_used_in_training: Graph - The graph that was used while training the current - edge prediction model. - use_node_types: bool - Whether to return the node types. - use_edge_metrics: bool = False - Whether to return the edge metrics. - batch_size: int = 2**10, - The batch size to use. - """ - self._sequence = GenericEdgePredictionSequence( - graph=graph, - graph_used_in_training=graph_used_in_training, - use_node_types=use_node_types, - use_edge_metrics=use_edge_metrics, - batch_size=batch_size - ) - self._current_index = 0 - super().__init__( - sample_number=graph.get_number_of_directed_edges(), - batch_size=batch_size, - ) - - def __call__(self): - """Return next batch using an infinite generator model.""" - self._current_index += 1 - return (self[self._current_index],) - - def into_dataset(self) -> tf.data.Dataset: - """Return dataset generated out of the current sequence instance. - - Implementative details - --------------------------------- - This method handles the conversion of this Keras Sequence into - a TensorFlow dataset, also handling the proper dispatching according - to what version of TensorFlow is installed in this system. - - Returns - ---------------------------------- - Dataset to be used for the training of a model - """ - - ################################################################# - # Handling kernel creation when TensorFlow is a modern version. # - ################################################################# - - if tensorflow_version_is_higher_or_equal_than("2.5.0"): - input_tensor_specs = [] - - for _ in range(2): - # Shapes of the source and destination node IDs - input_tensor_specs.append(tf.TensorSpec( - shape=(None, ), - dtype=tf.int32 - )) - - if self._use_node_types: - # Shapes of the source and destination node type IDs - input_tensor_specs.append(tf.TensorSpec( - shape=(None, - self._graph.get_maximum_multilabel_count()), - dtype=tf.int32 - )) - - if self._use_edge_metrics: - # Shapes of the edge type IDs - input_tensor_specs.append(tf.TensorSpec( - shape=(None, - self._graph.get_number_of_available_edge_metrics()), - dtype=tf.float32 - )) - - return tf.data.Dataset.from_generator( - self, - output_signature=( - tuple(input_tensor_specs), - ) - ) - - input_tensor_types = [] - input_tensor_shapes = [] - - for _ in range(2): - input_tensor_types.append(tf.int32,) - input_tensor_shapes.append(tf.TensorShape([None, ]),) - - if self._use_node_types: - input_tensor_types.append(tf.int32,) - input_tensor_shapes.append( - tf.TensorShape([None, self._graph.get_maximum_multilabel_count()]),) - - if self._use_edge_metrics: - input_tensor_types.append(tf.float32,) - input_tensor_shapes.append(tf.TensorShape( - [None, self._graph.get_number_of_available_edge_metrics()]),) - - return tf.data.Dataset.from_generator( - self, - output_types=( - tuple(input_tensor_types), - ), - output_shapes=( - tuple(input_tensor_shapes), - ) - ) - - def __getitem__(self, idx: int) -> Tuple[np.ndarray, np.ndarray]: - """Return batch corresponding to given index. - - Parameters - --------------- - idx: int, - Index corresponding to batch to be returned. - - Returns - --------------- - Return Tuple containing X and Y numpy arrays corresponding to given batch index. - """ - return self._sequence[idx] \ No newline at end of file diff --git a/embiggen/sequences/tensorflow_sequences/edge_prediction_training_sequence.py b/embiggen/sequences/tensorflow_sequences/edge_prediction_training_sequence.py index 65b39c94..0cbc9d53 100644 --- a/embiggen/sequences/tensorflow_sequences/edge_prediction_training_sequence.py +++ b/embiggen/sequences/tensorflow_sequences/edge_prediction_training_sequence.py @@ -56,10 +56,6 @@ def __init__( random_state: int = 42, The random_state to use to make extraction reproducible. """ - if not graph.has_edges(): - raise ValueError( - f"An empty instance of graph {graph.get_name()} was provided!" - ) self._graph = graph self._negative_samples_rate = negative_samples_rate self._avoid_false_negatives = avoid_false_negatives @@ -74,101 +70,6 @@ def __init__( batch_size=batch_size, ) - def __call__(self): - """Return next batch using an infinite generator model.""" - self._current_index += 1 - return (self[self._current_index],) - - def into_dataset(self) -> tf.data.Dataset: - """Return dataset generated out of the current sequence instance. - - Implementative details - --------------------------------- - This method handles the conversion of this Keras Sequence into - a TensorFlow dataset, also handling the proper dispatching according - to what version of TensorFlow is installed in this system. - - Returns - ---------------------------------- - Dataset to be used for the training of a model - """ - - ################################################################# - # Handling kernel creation when TensorFlow is a modern version. # - ################################################################# - - if tensorflow_version_is_higher_or_equal_than("2.5.0"): - input_tensor_specs = [] - - for _ in range(2): - # Shapes of the source and destination node IDs - input_tensor_specs.append(tf.TensorSpec( - shape=(self._batch_size, ), - dtype=tf.int32 - )) - - if self._use_node_types: - # Shapes of the source and destination node type IDs - input_tensor_specs.append(tf.TensorSpec( - shape=(self._batch_size, - self._graph.get_maximum_multilabel_count()), - dtype=tf.int32 - )) - - if self._use_edge_metrics: - # Shapes of the edge type IDs - input_tensor_specs.append(tf.TensorSpec( - shape=(self._batch_size, - self._graph.get_number_of_available_edge_metrics()), - dtype=tf.float64 - )) - - return tf.data.Dataset.from_generator( - self, - output_signature=( - ( - *input_tensor_specs, - ), - tf.TensorSpec( - shape=(self._batch_size,), - dtype=tf.bool - ) - ) - ) - - input_tensor_types = [] - input_tensor_shapes = [] - - for _ in range(2): - input_tensor_types.append(tf.int32,) - input_tensor_shapes.append(tf.TensorShape([self._batch_size, ]),) - - if self._use_node_types: - input_tensor_types.append(tf.int32,) - input_tensor_shapes.append( - tf.TensorShape([self._batch_size, self._graph.get_maximum_multilabel_count()]),) - - if self._use_edge_metrics: - input_tensor_types.append(tf.float64,) - input_tensor_shapes.append(tf.TensorShape( - [self._batch_size, self._graph.get_number_of_available_edge_metrics()]),) - - return tf.data.Dataset.from_generator( - self, - output_types=( - ( - *input_tensor_types, - ), - tf.bool - ), - output_shapes=( - ( - *input_tensor_shapes, - ), - tf.TensorShape([self._batch_size, ]), - ) - ) - def __getitem__(self, idx: int) -> Tuple[np.ndarray, np.ndarray]: """Return batch corresponding to given index. diff --git a/embiggen/sequences/tensorflow_sequences/gcn_edge_label_prediction_training_sequence.py b/embiggen/sequences/tensorflow_sequences/gcn_edge_label_prediction_training_sequence.py index 67d1c600..e1ce3577 100644 --- a/embiggen/sequences/tensorflow_sequences/gcn_edge_label_prediction_training_sequence.py +++ b/embiggen/sequences/tensorflow_sequences/gcn_edge_label_prediction_training_sequence.py @@ -49,7 +49,11 @@ def __init__( When the graph has multilabel node types, we will average the features. edge_features: Optional[List[np.ndarray]] = None, - + The edge features to be used. + For instance, these could be BERT embeddings of the + description of the edges. + When the graph has multilabel edges, + we will average the features. use_edge_metrics: bool = False Whether to return the edge metrics. """ @@ -74,27 +78,10 @@ def __init__( # The index in the returned sequence that contains the # edge label is 2 (source and destination nodes). - self._edge_label_index = 2 - - def __call__(self): - """Return next batch using an infinite generator model.""" - self._current_index += 1 - return (self[self._current_index],) - - def into_dataset(self) -> tf.data.Dataset: - """Return dataset generated out of the current sequence instance. - - Implementative details - --------------------------------- - This method handles the conversion of this Keras Sequence into - a TensorFlow dataset, also handling the proper dispatching according - to what version of TensorFlow is installed in this system. - - Returns - ---------------------------------- - Dataset to be used for the training of a model - """ - raise NotImplementedError("TODO!") + if return_node_types: + self._edge_label_index = 4 + else: + self._edge_label_index = 2 def __getitem__(self, idx: int): """Return batch corresponding to given index. diff --git a/embiggen/sequences/tensorflow_sequences/gcn_edge_prediction_sequence.py b/embiggen/sequences/tensorflow_sequences/gcn_edge_prediction_sequence.py index 7418ad20..c7d1ceef 100644 --- a/embiggen/sequences/tensorflow_sequences/gcn_edge_prediction_sequence.py +++ b/embiggen/sequences/tensorflow_sequences/gcn_edge_prediction_sequence.py @@ -54,34 +54,43 @@ def __init__( use_edge_metrics: bool = False Whether to return the edge metrics. """ - if not graph.has_edges(): + if ( + return_node_types or + node_type_features is not None + ) and graph.has_unknown_node_types(): raise ValueError( - f"An empty instance of graph {graph.get_name()} was provided!" + f"The provided graph {graph.get_name()} " + "contains unknown node types but node types " + "have been requested for the sequence." ) + self._graph = graph self._kernel = kernel if node_features is None: node_features = [] - self._node_features = node_features + self._node_features = [ + node_feature.astype(np.float32) + for node_feature in node_features + ] if return_node_ids: self._node_ids = graph.get_node_ids() else: self._node_ids = None - + if return_node_types or node_type_features is not None: if graph.has_multilabel_node_types(): node_types = graph.get_one_hot_encoded_node_types() else: node_types = graph.get_single_label_node_type_ids() - + self._node_types = node_types if return_node_types else None if node_type_features is not None: if self._graph.has_multilabel_node_types(): self._node_type_features = [] - minus_node_types = node_type_feature[node_types - 1] - node_types_mask = node_types==0 - for node_type_feature in self._node_type_features: + minus_node_types = node_types - 1 + node_types_mask = node_types == 0 + for node_type_feature in node_type_features: self._node_type_features.append(np.ma.array( node_type_feature[minus_node_types], mask=np.repeat( @@ -91,22 +100,12 @@ def __init__( *node_types.shape, node_type_feature.shape[1] )) - ).mean(axis=-1).data) + ).mean(axis=-2).data) else: - if self._graph.has_unknown_node_types(): - self._node_type_features = [] - minus_node_types = node_type_feature[node_types - 1] - node_types_mask = node_types==0 - for node_type_feature in self._node_type_features: - ntf = node_type_feature[minus_node_types] - # Masking the unknown values to zero. - ntf[node_types_mask] = 0.0 - self._node_type_features.append(ntf) - else: - self._node_type_features = [ - node_type_feature[node_types] - for node_type_feature in node_type_features - ] + self._node_type_features = [ + node_type_feature[node_types] + for node_type_feature in node_type_features + ] else: self._node_type_features = [] @@ -119,11 +118,15 @@ def __init__( batch_size=support.get_number_of_nodes() ) - self._edge_features_sequence = None if edge_features is None else VectorSequence( - edge_features, - batch_size=graph.get_number_of_nodes(), - shuffle=False - ) + self._edge_features = edge_features + self._edge_features_sequences = None if edge_features is None else [ + VectorSequence( + edge_feature, + batch_size=graph.get_number_of_nodes(), + shuffle=False + ) + for edge_feature in edge_features + ] self._use_edge_metrics = use_edge_metrics self._current_index = 0 @@ -132,78 +135,6 @@ def __init__( batch_size=graph.get_number_of_nodes(), ) - def __call__(self): - """Return next batch using an infinite generator model.""" - self._current_index += 1 - return (self[self._current_index],) - - def into_dataset(self) -> tf.data.Dataset: - """Return dataset generated out of the current sequence instance. - - Implementative details - --------------------------------- - This method handles the conversion of this Keras Sequence into - a TensorFlow dataset, also handling the proper dispatching according - to what version of TensorFlow is installed in this system. - - Returns - ---------------------------------- - Dataset to be used for the training of a model - """ - - input_tensor_specs = [] - - for _ in range(2): - # Shapes of the source and destination node IDs - input_tensor_specs.append(tf.TensorSpec( - shape=(None, ), - dtype=tf.int32 - )) - - if self._use_edge_metrics: - # Shapes of the edge type IDs - input_tensor_specs.append(tf.TensorSpec( - shape=(None, self._graph.get_number_of_available_edge_metrics()), - dtype=tf.float64 - )) - - # Shapes of the kernel - input_tensor_specs.append(tf.SparseTensorSpec( - shape=(None, None), - dtype=tf.float32 - )) - - if self._node_features is not None: - input_tensor_specs.extend([ - tf.TensorSpec( - shape=(None, node_feature.shape[1]), - dtype=tf.float32 - ) - for node_feature in self._node_features - ]) - - input_tensor_specs.extend([ - tf.TensorSpec( - shape=(None, node_type_feature.shape[1]), - dtype=tf.float32 - ) - for node_type_feature in self._node_type_features - ]) - - return tf.data.Dataset.from_generator( - self, - output_signature=( - ( - *input_tensor_specs, - ), - tf.TensorSpec( - shape=(None,), - dtype=tf.bool - ) - ) - ) - - def __getitem__(self, idx: int) -> Tuple[np.ndarray, np.ndarray]: """Return batch corresponding to given index. @@ -217,27 +148,37 @@ def __getitem__(self, idx: int) -> Tuple[np.ndarray, np.ndarray]: Return Tuple containing X and Y numpy arrays corresponding to given batch index. """ values = self._sequence[idx][0] - edge_features = None if self._edge_features_sequence is None else self._edge_features_sequence[idx] + edge_features = None if self._edge_features_sequences is None else [ + edge_features_sequence[idx] + for edge_features_sequence in self._edge_features_sequences + ] # If necessary, we add the padding as the last batch may be # smaller than the required size (number of nodes). delta = self.batch_size - values[0].shape[0] if delta > 0: values = [ - np.pad(value, (0, delta) if len(value.shape) == 1 else [(0, delta), (0, 0)]) + np.pad(value, (0, delta) if len(value.shape) + == 1 else [(0, delta), (0, 0)]) for value in values if value is not None ] if edge_features is not None: - edge_features = np.pad(edge_features, [(0, delta), (0, 0)]) + edge_features = [ + np.pad(edge_feature, [(0, delta), (0, 0)]) + for edge_feature in edge_features + ] + + if edge_features is None: + edge_features = [] return (tuple([ value for value in ( *values, + *edge_features, self._kernel, *self._node_features, *self._node_type_features, - edge_features, self._node_ids, self._node_types, ) diff --git a/embiggen/sequences/tensorflow_sequences/gcn_edge_prediction_training_sequence.py b/embiggen/sequences/tensorflow_sequences/gcn_edge_prediction_training_sequence.py index 535cd19a..aa9bb2ca 100644 --- a/embiggen/sequences/tensorflow_sequences/gcn_edge_prediction_training_sequence.py +++ b/embiggen/sequences/tensorflow_sequences/gcn_edge_prediction_training_sequence.py @@ -87,11 +87,13 @@ def __init__( if node_features is None: node_features = [] + self._node_features = node_features self._negative_samples_rate = negative_samples_rate self._avoid_false_negatives = avoid_false_negatives self._graph_to_avoid = graph_to_avoid self._random_state = random_state + if return_node_ids: self._node_ids = graph.get_node_ids() else: @@ -105,16 +107,12 @@ def __init__( self._node_types = node_types if return_node_types else None - if node_type_features is not None: - if graph.has_multilabel_node_types(): - node_types = graph.get_one_hot_encoded_node_types() - else: - node_types = graph.get_single_label_node_type_ids() + if node_type_features is not None: if self._graph.has_multilabel_node_types(): self._node_type_features = [] - minus_node_types = node_type_feature[node_types - 1] node_types_mask = node_types==0 - for node_type_feature in self._node_type_features: + minus_node_types = node_types - 1 + for node_type_feature in node_type_features: self._node_type_features.append(np.ma.array( node_type_feature[minus_node_types], mask=np.repeat( @@ -124,13 +122,14 @@ def __init__( *node_types.shape, node_type_feature.shape[1] )) - ).mean(axis=-1).data) + ).mean(axis=-2).data) else: if self._graph.has_unknown_node_types(): self._node_type_features = [] - minus_node_types = node_type_feature[node_types - 1] node_types_mask = node_types==0 - for node_type_feature in self._node_type_features: + minus_node_types = node_types - 1 + minus_node_types[node_types_mask] = 0 + for node_type_feature in node_type_features: ntf = node_type_feature[minus_node_types] # Masking the unknown values to zero. ntf[node_types_mask] = 0.0 @@ -151,78 +150,6 @@ def __init__( batch_size=graph.get_number_of_nodes(), ) - def __call__(self): - """Return next batch using an infinite generator model.""" - self._current_index += 1 - return (self[self._current_index],) - - def into_dataset(self) -> tf.data.Dataset: - """Return dataset generated out of the current sequence instance. - - Implementative details - --------------------------------- - This method handles the conversion of this Keras Sequence into - a TensorFlow dataset, also handling the proper dispatching according - to what version of TensorFlow is installed in this system. - - Returns - ---------------------------------- - Dataset to be used for the training of a model - """ - - input_tensor_specs = [] - - for _ in range(2): - # Shapes of the source and destination node IDs - input_tensor_specs.append(tf.TensorSpec( - shape=(None, ), - dtype=tf.int32 - )) - - if self._use_edge_metrics: - # Shapes of the edge type IDs - input_tensor_specs.append(tf.TensorSpec( - shape=(None, self._graph.get_number_of_available_edge_metrics()), - dtype=tf.float64 - )) - - # Shapes of the kernel - input_tensor_specs.append(tf.SparseTensorSpec( - shape=(None, None), - dtype=tf.float32 - )) - - if self._node_features is not None: - input_tensor_specs.extend([ - tf.TensorSpec( - shape=(None, node_feature.shape[1]), - dtype=tf.float32 - ) - for node_feature in self._node_features - ]) - - input_tensor_specs.extend([ - tf.TensorSpec( - shape=(None, node_type_feature.shape[1]), - dtype=tf.float32 - ) - for node_type_feature in self._node_type_features - ]) - - return tf.data.Dataset.from_generator( - self, - output_signature=( - ( - *input_tensor_specs, - ), - tf.TensorSpec( - shape=(None,), - dtype=tf.bool - ) - ) - ) - - def __getitem__(self, idx: int) -> Tuple[np.ndarray, np.ndarray]: """Return batch corresponding to given index. diff --git a/embiggen/sequences/tensorflow_sequences/kgsiamese_sequence.py b/embiggen/sequences/tensorflow_sequences/kgsiamese_sequence.py deleted file mode 100644 index 7aeaf911..00000000 --- a/embiggen/sequences/tensorflow_sequences/kgsiamese_sequence.py +++ /dev/null @@ -1,128 +0,0 @@ -"""Keras Sequence for running Neural Network on graph edge prediction.""" -from typing import List - -import numpy as np -import tensorflow as tf -from ensmallen import Graph # pylint: disable=no-name-in-module -from keras_mixed_sequence import Sequence -from embiggen.utils.tensorflow_utils import tensorflow_version_is_higher_or_equal_than - - -class KGSiameseSequence(Sequence): - """Keras Sequence for running Siamese Neural Network.""" - - def __init__( - self, - graph: Graph, - batch_size: int = 2**10, - random_state: int = 42 - ): - """Create new EdgePredictionSequence object. - - Parameters - -------------------------------- - graph: Graph, - The graph from which to sample the triples. - batch_size: int = 2**10, - The batch size to use. - random_state: int = 42, - The random_state to use to make extraction reproducible. - """ - self._graph = graph - self._random_state = random_state - self._current_index = 0 - super().__init__( - sample_number=self._graph.get_number_of_directed_edges(), - batch_size=batch_size, - ) - - def __call__(self): - """Return next batch using an infinite generator model.""" - self._current_index += 1 - return (tuple(self[self._current_index]),) - - def into_dataset(self) -> tf.data.Dataset: - """Return dataset generated out of the current sequence instance. - - Implementative details - --------------------------------- - This method handles the conversion of this Keras Sequence into - a TensorFlow dataset, also handling the proper dispatching according - to what version of TensorFlow is installed in this system. - - Returns - ---------------------------------- - Dataset to be used for the training of a model - """ - - ################################################################# - # Handling kernel creation when TensorFlow is a modern version. # - ################################################################# - - if tensorflow_version_is_higher_or_equal_than("2.5.0"): - input_tensor_specs = [] - - # For both the real and fake nodes. - for _ in range(4): - # Shapes of the source and destination node IDs - input_tensor_specs.append(tf.TensorSpec( - shape=(self._batch_size, ), - dtype=tf.uint32 - )) - for _ in range(4): - input_tensor_specs.append( - tf.TensorSpec( - shape=( - self._batch_size, - self._graph.get_maximum_multilabel_count() - ), - dtype=tf.uint32 - ) - ) - - # Shapes of the edge type IDs - input_tensor_specs.append(tf.TensorSpec( - shape=(self._batch_size,), - dtype=tf.uint32 - )) - - return tf.data.Dataset.from_generator( - self, - output_signature=(tuple(input_tensor_specs),) - ) - - input_tensor_types = [] - input_tensor_shapes = [] - - for _ in range(4): - input_tensor_types.append(tf.uint32,) - input_tensor_shapes.append(tf.TensorShape([self._batch_size, ]),) - - for _ in range(4): - input_tensor_types.append(tf.uint32,) - input_tensor_shapes.append( - tf.TensorShape([self._batch_size, self._graph.get_maximum_multilabel_count()]),) - - input_tensor_types.append(tf.uint32,) - input_tensor_shapes.append(tf.TensorShape([self._batch_size, ]),) - - return tf.data.Dataset.from_generator( - self, - output_types=input_tensor_types, - output_shapes=input_tensor_shapes - ) - - def __getitem__(self, idx: int) -> List[np.ndarray]: - """Return batch corresponding to given index to train a Siamese network. - - Parameters - --------------- - idx: int, - Index corresponding to batch to be returned. - """ - random_state = (self._random_state + idx) * (1 + self.elapsed_epochs) - return (self._graph.get_kgsiamese_mini_batch( - random_state, - batch_size=self.batch_size, - use_zipfian_sampling=True - ),) diff --git a/embiggen/sequences/tensorflow_sequences/node_label_prediction_sequence.py b/embiggen/sequences/tensorflow_sequences/node_label_prediction_sequence.py deleted file mode 100644 index 08d70208..00000000 --- a/embiggen/sequences/tensorflow_sequences/node_label_prediction_sequence.py +++ /dev/null @@ -1,58 +0,0 @@ -"""Keras Sequence for running Neural Network on graph node-label embedding.""" -from typing import Tuple, Union, Optional - -import numpy as np -import tensorflow as tf -from ensmallen import Graph -from keras_mixed_sequence import Sequence - - -class NodeLabelPredictionSequence(Sequence): - """Keras Sequence for running Neural Network on graph node-label prediction.""" - - def __init__( - self, - graph: Graph, - max_neighbours: Optional[int] = 100, - include_central_node: bool = False, - return_edge_weights: bool = False, - batch_size: int = 2**8, - elapsed_epochs: int = 0, - random_state: int = 42, - ): - """Create new Node Label Prediction Sequence. - """ - self._graph = graph - self._random_state = random_state - self._include_central_node = include_central_node - self._return_edge_weights = return_edge_weights - self._max_neighbours = max_neighbours - super().__init__( - sample_number=graph.get_number_of_directed_edges(), - batch_size=batch_size, - elapsed_epochs=elapsed_epochs - ) - - def __getitem__(self, idx: int) -> Tuple[Union[tf.RaggedTensor, Tuple[tf.RaggedTensor]], np.ndarray]: - """Return batch corresponding to given index. - - Parameters - --------------- - idx: int, - Index corresponding to batch to be returned. - - Returns - --------------- - Return Tuple containing X and Y numpy arrays corresponding to given batch index. - """ - neighbours, weights, labels = self._graph.get_node_label_prediction_mini_batch( - idx=(self._random_state + idx) * (1 + self.elapsed_epochs), - batch_size=self._batch_size, - include_central_node=self._include_central_node, - return_edge_weights=self._return_edge_weights, - max_neighbours=self._max_neighbours - ) - - if self._return_edge_weights: - return (tf.ragged.constant(neighbours), tf.ragged.constant(weights)), labels - return tf.ragged.constant(neighbours), labels diff --git a/embiggen/sequences/tensorflow_sequences/siamese_sequence.py b/embiggen/sequences/tensorflow_sequences/siamese_sequence.py index 76b92ff5..4b006638 100644 --- a/embiggen/sequences/tensorflow_sequences/siamese_sequence.py +++ b/embiggen/sequences/tensorflow_sequences/siamese_sequence.py @@ -1,4 +1,4 @@ -"""Keras Sequence for running Neural Network on graph edge prediction.""" +"""Keras Sequence for running Siamese Neural Network based on currupted triples sampling.""" from typing import List import numpy as np @@ -15,9 +15,10 @@ def __init__( self, graph: Graph, batch_size: int = 2**10, + return_edge_types: bool = False, random_state: int = 42 ): - """Create new EdgePredictionSequence object. + """Create new SiameseSequence object. Parameters -------------------------------- @@ -25,10 +26,13 @@ def __init__( The graph from which to sample the triples. batch_size: int = 2**10, The batch size to use. + return_edge_types: bool = False + Whether to return the edge types. random_state: int = 42, The random_state to use to make extraction reproducible. """ self._graph = graph + self._return_edge_types = return_edge_types self._random_state = random_state self._current_index = 0 super().__init__( @@ -54,47 +58,15 @@ def into_dataset(self) -> tf.data.Dataset: ---------------------------------- Dataset to be used for the training of a model """ - - ################################################################# - # Handling kernel creation when TensorFlow is a modern version. # - ################################################################# - - if tensorflow_version_is_higher_or_equal_than("2.5.0"): - input_tensor_specs = [] - - # For both the real and fake nodes. - for _ in range(4): - # Shapes of the source and destination node IDs - input_tensor_specs.append(tf.TensorSpec( - shape=(self._batch_size, ), - dtype=tf.uint32 - )) - - # Shapes of the edge type IDs - input_tensor_specs.append(tf.TensorSpec( - shape=(self._batch_size,), - dtype=tf.uint32 - )) - - return tf.data.Dataset.from_generator( - self, - output_signature=(tuple(input_tensor_specs),) - ) - - input_tensor_types = [] - input_tensor_shapes = [] - - for _ in range(4): - input_tensor_types.append(tf.uint32,) - input_tensor_shapes.append(tf.TensorShape([self._batch_size, ]),) - - input_tensor_types.append(tf.uint32,) - input_tensor_shapes.append(tf.TensorShape([self._batch_size, ]),) - return tf.data.Dataset.from_generator( self, - output_types=input_tensor_types, - output_shapes=input_tensor_shapes + output_signature=(tuple([ + tf.TensorSpec( + shape=(self._batch_size, ), + dtype=tf.uint32 + ) + for _ in range(4 + int(self._return_edge_types)) + ]),) ) def __getitem__(self, idx: int) -> List[np.ndarray]: @@ -106,8 +78,14 @@ def __getitem__(self, idx: int) -> List[np.ndarray]: Index corresponding to batch to be returned. """ random_state = (self._random_state + idx) * (1 + self.elapsed_epochs) - return (self._graph.get_siamese_mini_batch( - random_state, - batch_size=self.batch_size, - use_zipfian_sampling=True - ),) + return ( + self._graph.get_siamese_mini_batch_with_edge_types( + random_state, + batch_size=self.batch_size, + ) + if self._return_edge_types + else self._graph.get_siamese_mini_batch( + random_state, + batch_size=self.batch_size, + ), + ) diff --git a/embiggen/similarities/__init__.py b/embiggen/similarities/__init__.py new file mode 100644 index 00000000..d1958920 --- /dev/null +++ b/embiggen/similarities/__init__.py @@ -0,0 +1,5 @@ +"""Submodule providing models for edge prediction.""" +from embiggen.similarities.dag_resnik import DAGResnik + +# Export all non-internals. +__all__ = ["DAGResnik"] diff --git a/embiggen/similarities/dag_resnik.py b/embiggen/similarities/dag_resnik.py new file mode 100644 index 00000000..e4d368ff --- /dev/null +++ b/embiggen/similarities/dag_resnik.py @@ -0,0 +1,395 @@ +from typing import List, Optional, Union, Dict +import pandas as pd +import numpy as np +from ensmallen import models, Graph + + +class DAGResnik: + + def __init__(self, verbose: bool = True): + self._model = models.DAGResnik(verbose) + + def fit( + self, + graph: Graph, + node_counts: Dict[str, float], + node_frequencies: Optional[np.ndarray] = None, + ): + """Fit the Resnik similarity model. + + Parameters + -------------------- + graph: Graph + The graph to run similarities on. + node_counts: Dict[str, float] + Counts to compute the terms frequencies. + node_frequencies: Optional[np.ndarray] = None + Optional vector of node frequencies. + """ + self._model.fit( + graph, + node_counts=node_counts, + node_frequencies=node_frequencies + ) + + def get_similarity_from_node_id( + self, + first_node_id: int, + second_node_id: int + ) -> float: + """Return the similarity between the two provided nodes. + + Arguments + -------------------- + first_node_id: int + The first node for which to compute the similarity. + second_node_id: int + The second node for which to compute the similarity. + """ + return self._model.get_similarity_from_node_id(first_node_id, second_node_id) + + def get_similarity_from_node_ids( + self, + first_node_ids: List[int], + second_node_ids: List[int] + ) -> np.ndarray: + """Return the similarity between the two provided nodes. + + Arguments + -------------------- + first_node_ids: List[int] + The first node for which to compute the similarity. + second_node_ids: List[int] + The second node for which to compute the similarity. + """ + return self._model.get_similarity_from_node_ids(first_node_ids, second_node_ids) + + def get_similarity_from_node_name( + self, + first_node_name: str, + second_node_name: str + ) -> float: + """Return the similarity between the two provided nodes. + + Arguments + -------------------- + first_node_name: str + The first node for which to compute the similarity. + second_node_name: str + The second node for which to compute the similarity. + """ + return self._model.get_similarity_from_node_name(first_node_name, second_node_name) + + def get_similarity_from_node_names( + self, + first_node_names: List[str], + second_node_names: List[str] + ) -> np.ndarray: + """Return the similarity between the two provided nodes. + + Arguments + -------------------- + first_node_names: List[str] + The first node for which to compute the similarity. + second_node_names: List[str] + The second node for which to compute the similarity. + """ + return self._model.get_similarity_from_node_names(first_node_names, second_node_names) + + def get_pairwise_similarities( + self, + graph: Optional[Graph] = None, + return_similarities_dataframe: bool = False + ) -> Union[pd.DataFrame, np.ndarray]: + """Execute similarities on the provided graph. + + Parameters + -------------------- + graph: Optional[Graph] = None + The graph to run similarities on. + return_similarities_dataframe: bool = False + Whether to return a pandas DataFrame, which as indices has the node IDs. + By default, a numpy array with the similarities is returned as it weights much less. + """ + similarities = self._model.get_pairwise_similarities() + + if return_similarities_dataframe and graph is not None: + similarities = pd.DataFrame( + similarities, + columns=graph.get_node_names(), + index=graph.get_node_names(), + ) + + return similarities + + def get_similarities_from_graph( + self, + graph: Graph, + return_similarities_dataframe: bool = False + ) -> Union[pd.DataFrame, np.ndarray]: + """Execute similarities on the provided graph. + + Parameters + -------------------- + graph: Graph + The graph to run similarities on. + node_frequencies: Optional[np.ndarray] + Optional vector of node frequencies. + return_similarities_dataframe: bool = False + Whether to return a pandas DataFrame, which as indices has the node IDs. + By default, a numpy array with the similarities is returned as it weights much less. + """ + similarities = self._model.get_similarities_from_graph( + graph, + ) + + if return_similarities_dataframe: + similarities = pd.DataFrame( + { + "similarities": similarities, + "sources": graph.get_directed_source_node_ids(), + "destinations": graph.get_directed_destination_node_ids(), + }, + ) + + return similarities + + def get_similarities_from_bipartite_graph_from_edge_node_ids( + self, + graph: Graph, + source_node_ids: List[int], + destination_node_ids: List[int], + return_similarities_dataframe: bool = False + ) -> Union[pd.DataFrame, np.ndarray]: + """Execute similarities probabilities on the provided graph bipartite portion. + + Parameters + -------------------- + graph: Graph + The graph from which to extract the edges. + node_frequencies: Optional[np.ndarray] + Optional vector of node frequencies. + source_node_ids: List[int] + The source nodes of the bipartite graph. + destination_node_ids: List[int] + The destination nodes of the bipartite graph. + return_similarities_dataframe: bool = False + Whether to return a pandas DataFrame, which as indices has the node IDs. + By default, a numpy array with the similarities is returned as it weights much less. + """ + return self.get_similarities_from_graph( + graph.build_bipartite_graph_from_edge_node_ids( + source_node_ids=source_node_ids, + destination_node_ids=destination_node_ids, + directed=True + ), + return_similarities_dataframe=return_similarities_dataframe + ) + + def get_similarities_from_bipartite_graph_from_edge_node_names( + self, + graph: Graph, + source_node_names: List[str], + destination_node_names: List[str], + return_similarities_dataframe: bool = False + ) -> Union[pd.DataFrame, np.ndarray]: + """Execute similarities probabilities on the provided graph bipartite portion. + + Parameters + -------------------- + graph: Graph + The graph from which to extract the edges. + node_frequencies: Optional[np.ndarray] + Optional vector of node frequencies. + source_node_names: List[str] + The source nodes of the bipartite graph. + destination_node_names: List[str] + The destination nodes of the bipartite graph. + return_similarities_dataframe: bool = False + Whether to return a pandas DataFrame, which as indices has the node IDs. + By default, a numpy array with the similarities is returned as it weights much less. + """ + return self.get_similarities_from_graph( + graph.build_bipartite_graph_from_edge_node_names( + source_node_names=source_node_names, + destination_node_names=destination_node_names, + directed=True + ), + return_similarities_dataframe=return_similarities_dataframe + ) + + def get_similarities_from_bipartite_graph_from_edge_node_prefixes( + self, + graph: Graph, + source_node_prefixes: List[str], + destination_node_prefixes: List[str], + return_similarities_dataframe: bool = False + ) -> Union[pd.DataFrame, np.ndarray]: + """Execute similarities probabilities on the provided graph bipartite portion. + + Parameters + -------------------- + graph: Graph + The graph from which to extract the edges. + node_frequencies: Optional[np.ndarray] + Optional vector of node frequencies. + source_node_prefixes: List[str] + The source node prefixes of the bipartite graph. + destination_node_prefixes: List[str] + The destination node prefixes of the bipartite graph. + return_similarities_dataframe: bool = False + Whether to return a pandas DataFrame, which as indices has the node IDs. + By default, a numpy array with the similarities is returned as it weights much less. + """ + return self.get_similarities_from_graph( + graph.build_bipartite_graph_from_edge_node_prefixes( + source_node_prefixes=source_node_prefixes, + destination_node_prefixes=destination_node_prefixes, + directed=True + ), + return_similarities_dataframe=return_similarities_dataframe + ) + + def get_similarities_from_bipartite_graph_from_edge_node_types( + self, + graph: Graph, + source_node_types: List[str], + destination_node_types: List[str], + return_similarities_dataframe: bool = False + ) -> Union[pd.DataFrame, np.ndarray]: + """Execute similarities probabilities on the provided graph bipartite portion. + + Parameters + -------------------- + graph: Graph + The graph from which to extract the edges. + node_frequencies: Optional[np.ndarray] + Optional vector of node frequencies. + source_node_types: List[str] + The source node prefixes of the bipartite graph. + destination_node_types: List[str] + The destination node prefixes of the bipartite graph. + return_similarities_dataframe: bool = False + Whether to return a pandas DataFrame, which as indices has the node IDs. + By default, a numpy array with the similarities is returned as it weights much less. + """ + return self.get_similarities_from_graph( + graph.build_bipartite_graph_from_edge_node_types( + source_node_types=source_node_types, + destination_node_types=destination_node_types, + directed=True + ), + return_similarities_dataframe=return_similarities_dataframe + ) + + def get_similarities_from_clique_graph_from_node_ids( + self, + graph: Graph, + node_ids: List[int], + return_similarities_dataframe: bool = False + ) -> Union[pd.DataFrame, np.ndarray]: + """Execute similarities probabilities on the provided graph bipartite portion. + + Parameters + -------------------- + graph: Graph + The graph from which to extract the edges. + node_frequencies: Optional[np.ndarray] + Optional vector of node frequencies. + node_ids: List[int] + The nodes of the bipartite graph. + return_similarities_dataframe: bool = False + Whether to return a pandas DataFrame, which as indices has the node IDs. + By default, a numpy array with the similarities is returned as it weights much less. + """ + return self.get_similarities_from_graph( + graph.build_clique_graph_from_node_ids( + node_ids=node_ids, + directed=True + ), + return_similarities_dataframe=return_similarities_dataframe + ) + + def get_similarities_from_clique_graph_from_node_names( + self, + graph: Graph, + node_names: List[str], + return_similarities_dataframe: bool = False + ) -> Union[pd.DataFrame, np.ndarray]: + """Execute similarities probabilities on the provided graph bipartite portion. + + Parameters + -------------------- + graph: Graph + The graph from which to extract the edges. + node_frequencies: Optional[np.ndarray] + Optional vector of node frequencies. + node_names: List[str] + The nodes of the bipartite graph. + return_similarities_dataframe: bool = False + Whether to return a pandas DataFrame, which as indices has the node IDs. + By default, a numpy array with the similarities is returned as it weights much less. + """ + return self.get_similarities_from_graph( + graph.build_clique_graph_from_node_names( + node_names=node_names, + directed=True + ), + return_similarities_dataframe=return_similarities_dataframe + ) + + def get_similarities_from_clique_graph_from_node_prefixes( + self, + graph: Graph, + node_prefixes: List[str], + return_similarities_dataframe: bool = False + ) -> Union[pd.DataFrame, np.ndarray]: + """Execute similarities probabilities on the provided graph bipartite portion. + + Parameters + -------------------- + graph: Graph + The graph from which to extract the edges. + node_frequencies: Optional[np.ndarray] + Optional vector of node frequencies. + node_prefixes: List[str] + The node prefixes of the bipartite graph. + return_similarities_dataframe: bool = False + Whether to return a pandas DataFrame, which as indices has the node IDs. + By default, a numpy array with the similarities is returned as it weights much less. + """ + return self.get_similarities_from_graph( + graph.build_clique_graph_from_node_prefixes( + node_prefixes=node_prefixes, + directed=True + ), + return_similarities_dataframe=return_similarities_dataframe + ) + + def get_similarities_from_clique_graph_from_node_types( + self, + graph: Graph, + node_types: List[str], + return_similarities_dataframe: bool = False + ) -> Union[pd.DataFrame, np.ndarray]: + """Execute similarities probabilities on the provided graph bipartite portion. + + Parameters + -------------------- + graph: Graph + The graph from which to extract the edges. + node_frequencies: Optional[np.ndarray] + Optional vector of node frequencies. + node_types: List[str] + The node prefixes of the bipartite graph. + return_similarities_dataframe: bool = False + Whether to return a pandas DataFrame, which as indices has the node IDs. + By default, a numpy array with the similarities is returned as it weights much less. + """ + return self.get_similarities_from_graph( + graph.build_clique_graph_from_node_types( + node_types=node_types, + directed=True + ), + return_similarities_dataframe=return_similarities_dataframe + ) diff --git a/embiggen/utils/__init__.py b/embiggen/utils/__init__.py index feceb1e6..0f627c6f 100644 --- a/embiggen/utils/__init__.py +++ b/embiggen/utils/__init__.py @@ -21,7 +21,6 @@ "EmbeddingResult", "AbstractModel", "classification_evaluation_pipeline", - "build_init", "format_list", "get_models_dataframe", "get_available_models_for_node_label_prediction", @@ -30,5 +29,4 @@ "get_available_models_for_node_embedding", "abstract_class", "number_to_ordinal", - "format_list" ] diff --git a/embiggen/utils/abstract_edge_gcn.py b/embiggen/utils/abstract_edge_gcn.py index 46428ab5..ff3f1ce4 100644 --- a/embiggen/utils/abstract_edge_gcn.py +++ b/embiggen/utils/abstract_edge_gcn.py @@ -33,6 +33,7 @@ def __init__( number_of_units_per_ffnn_head_layer: Union[int, List[int]] = 128, dropout_rate: float = 0.3, apply_norm: bool = False, + combiner: str ="sum", edge_embedding_method: str = "Concatenate", optimizer: Union[str, Optimizer] = "adam", early_stopping_min_delta: float = 0.0001, @@ -55,6 +56,7 @@ def __init__( handling_multi_graph: str = "warn", node_feature_names: Optional[List[str]] = None, node_type_feature_names: Optional[List[str]] = None, + edge_feature_names: Optional[List[str]] = None, verbose: bool = True ): """Create new Kipf GCN object. @@ -85,6 +87,13 @@ def __init__( apply_norm: bool = False Whether to normalize the output of the convolution operations, after applying the level activations. + combiner: str = "mean" + A string specifying the reduction op. + Currently "mean", "sqrtn" and "sum" are supported. + "sum" computes the weighted sum of the embedding results for each row. + "mean" is the weighted sum divided by the total weight. + "sqrtn" is the weighted sum divided by the square root of the sum of the squares of the weights. + Defaults to mean. edge_embedding_method: str = "Concatenate" The edge embedding method to use to put togheter the source and destination node features, which includes: @@ -181,6 +190,9 @@ def __init__( node_type_feature_names: Optional[List[str]] = None Names of the node type features. This is used as the layer names. + edge_feature_names: Optional[List[str]] = None + Names of the edge features. + This is used as the layer names. verbose: bool = True Whether to show loading bars. """ @@ -191,6 +203,7 @@ def __init__( number_of_units_per_graph_convolution_layers=number_of_units_per_graph_convolution_layers, dropout_rate=dropout_rate, apply_norm=apply_norm, + combiner=combiner, optimizer=optimizer, early_stopping_min_delta=early_stopping_min_delta, early_stopping_patience=early_stopping_patience, @@ -225,6 +238,7 @@ def __init__( self._edge_embedding_method = edge_embedding_method self._use_edge_metrics = use_edge_metrics + self._edge_feature_names = edge_feature_names self._use_node_types = None @classmethod @@ -263,6 +277,7 @@ def _get_model_prediction_input( return_node_types=self.is_using_node_types(), node_type_features=node_type_features, use_edge_metrics=self._use_edge_metrics, + edge_features=edge_features ) def _get_model_training_output(self, graph: Graph) -> Optional[np.ndarray]: @@ -346,6 +361,31 @@ def _build_model( else: edge_metrics = None + edge_feature_inputs = [] + if edge_features is None: + edge_features = [] + edge_feature_names = self._edge_feature_names + if edge_feature_names is None: + edge_feature_names = [ + f"{number_to_ordinal(i+1)}EdgeFeature" + for i in range(len(edge_features)) + ] + if len(edge_feature_names) != len(edge_features): + raise ValueError( + f"You have provided {len(edge_feature_names)} " + f"edge feature names but you have provided {len(edge_features)} " + "edge features to the model." + ) + for edge_feature, feature_name in zip(edge_features, edge_feature_names): + feature_names.append(feature_name) + edge_feature_input = Input( + shape=edge_feature.shape[1:], + batch_size=nodes_number, + name=feature_name, + ) + edge_feature_inputs.append(edge_feature_input) + source_and_destination_features.append(edge_feature_input) + ffnn_outputs = [] for hidden, feature_name in zip(source_and_destination_features, feature_names): @@ -364,7 +404,8 @@ def _build_model( if self._edge_embedding_method == "Concatenate": hidden = Concatenate( - name="NodeConcatenation" + name="NodeConcatenation", + axis=-1 )(source_and_destination_features) elif self._edge_embedding_method == "Average": hidden = Average( @@ -407,10 +448,11 @@ def _build_model( if len(other_features) > 0: hidden = Concatenate( - name="EdgeFeatures" + name="EdgeFeatures", + axis=-1 )([ hidden, - other_features + *other_features ]) # Building the head of the model. @@ -427,18 +469,21 @@ def _build_model( name="Output" )(hidden) + inputs=[ + input_layer + for input_layer in ( + source_nodes, + destination_nodes, + edge_metrics, + *edge_feature_inputs, + *graph_convolution_model.inputs, + ) + if input_layer is not None + ] + # Building the the model. model = Model( - inputs=[ - input_layer - for input_layer in ( - source_nodes, - destination_nodes, - edge_metrics, - *graph_convolution_model.inputs, - ) - if input_layer is not None - ], + inputs=inputs, outputs=output, name=self.model_name().replace(" ", "_") ) diff --git a/embiggen/utils/abstract_gcn.py b/embiggen/utils/abstract_gcn.py index 70a6521c..5cac7202 100644 --- a/embiggen/utils/abstract_gcn.py +++ b/embiggen/utils/abstract_gcn.py @@ -1,5 +1,6 @@ """Kipf GCN model for node-label prediction.""" from typing import List, Union, Optional, Dict, Any, Type, Tuple +from matplotlib.pyplot import axis import numpy as np import pandas as pd @@ -10,7 +11,7 @@ from ensmallen import Graph import tensorflow as tf -from tensorflow.keras.utils import Sequence +from keras_mixed_sequence import Sequence from embiggen.utils.abstract_models import AbstractClassifierModel, abstract_class from embiggen.utils.number_to_ordinal import number_to_ordinal from embiggen.layers.tensorflow import GraphConvolution, FlatEmbedding @@ -96,7 +97,7 @@ def graph_to_sparse_tensor( return tf.SparseTensor( graph.get_directed_edge_node_ids(), ( - graph.get_edge_weights() + graph.get_directed_edge_weights() if use_weights else tf.ones(graph.get_number_of_directed_edges()) ), @@ -115,6 +116,7 @@ def __init__( number_of_units_per_graph_convolution_layers: Union[int, List[int]] = 128, dropout_rate: float = 0.5, apply_norm: bool = False, + combiner: str = "mean", optimizer: Union[str, Optimizer] = "adam", early_stopping_min_delta: float = 0.001, early_stopping_patience: int = 10, @@ -153,6 +155,13 @@ def __init__( apply_norm: bool = False Whether to normalize the output of the convolution operations, after applying the level activations. + combiner: str = "mean" + A string specifying the reduction op. + Currently "mean", "sqrtn" and "sum" are supported. + "sum" computes the weighted sum of the embedding results for each row. + "mean" is the weighted sum divided by the total weight. + "sqrtn" is the weighted sum divided by the square root of the sum of the squares of the weights. + Defaults to mean. optimizer: str = "Adam" The optimizer to use while training the model. early_stopping_min_delta: float @@ -219,6 +228,7 @@ def __init__( can_be_empty=True ) + self._combiner = combiner self._epochs = epochs self._use_class_weights = use_class_weights self._dropout_rate = dropout_rate @@ -275,6 +285,7 @@ def parameters(self) -> Dict[str, Any]: number_of_units_per_graph_convolution_layers=self._number_of_units_per_graph_convolution_layers, epochs=self._epochs, apply_norm=self._apply_norm, + combiner=self._combiner, use_class_weights=self._use_class_weights, dropout_rate=self._dropout_rate, optimizer=self._optimizer, @@ -419,18 +430,21 @@ def _build_graph_convolution_model( ): if features is not None: if feature_names is None: - feature_names = [None] * len(features) + feature_names = [ + f"{number_to_ordinal(i+1)} {feature_category}" + for i in range(len(features)) + ] if len(feature_names) != len(features): raise ValueError( f"You have provided {len(feature_names)} " f"{feature_category} names but you have provided {len(features)} " - "{feature_category}s to the model." + f"{feature_category}s to the model." ) input_features.extend([ Input( shape=node_feature.shape[1:], batch_size=nodes_number, - name=node_feature_name + name=node_feature_name, ) for node_feature, node_feature_name in zip( features, @@ -482,6 +496,7 @@ def _build_graph_convolution_model( for i, units in enumerate(self._number_of_units_per_graph_convolution_layers): hidden = GraphConvolution( units=units, + combiner=self._combiner, dropout_rate=self._dropout_rate, apply_norm=self._apply_norm, name=f"{number_to_ordinal(i+1)}GraphConvolution" @@ -492,12 +507,16 @@ def _build_graph_convolution_model( inputs=[ input_layer for input_layer in ( - adjacency_matrix, *input_features + adjacency_matrix, + *input_features ) if input_layer is not None ], outputs=( - Concatenate(name="ConcatenatedNodeFeatures")(hidden) + Concatenate( + name="ConcatenatedNodeFeatures", + axis=-1 + )(hidden) if len(hidden) > 1 else hidden[0] ) @@ -561,16 +580,18 @@ def _fit( edge_features=edge_features, ) + model_input = self._get_model_training_input( + graph, + support=support, + edge_features=edge_features, + node_type_features=node_type_features, + node_features=node_features + ) + self.history = self._model.fit( - x=self._get_model_training_input( - graph, - support=support, - edge_features=edge_features, - node_type_features=node_type_features, - node_features=node_features - ), + x=model_input, y=self._get_model_training_output(graph), - sample_weight=self._get_model_training_output(graph), + sample_weight=self._get_model_training_sample_weights(graph), epochs=self._epochs, verbose=traditional_verbose and self._verbose > 0, batch_size=graph.get_number_of_nodes(), @@ -609,14 +630,17 @@ def _predict_proba( """Run predictions on the provided graph.""" if support is None: support = graph + + model_input = self._get_model_prediction_input( + graph, + support, + node_features, + node_type_features, + edge_features, + ) + return self._model.predict( - self._get_model_prediction_input( - graph, - support, - node_features, - node_type_features, - edge_features, - ), + model_input, batch_size=support.get_number_of_nodes(), verbose=False ) diff --git a/embiggen/utils/abstract_models/__init__.py b/embiggen/utils/abstract_models/__init__.py index 53044657..cde3ea73 100644 --- a/embiggen/utils/abstract_models/__init__.py +++ b/embiggen/utils/abstract_models/__init__.py @@ -16,7 +16,6 @@ "AbstractClassifierModel", "AbstractEmbeddingModel", "EmbeddingResult", - "get_standardized_model_map", "abstract_class", "AbstractModel", "get_models_dataframe", @@ -25,5 +24,5 @@ "get_available_models_for_node_label_prediction", "get_available_models_for_node_embedding", "build_init", - "format_list" + "format_list", ] diff --git a/embiggen/utils/abstract_models/abstract_classifier_model.py b/embiggen/utils/abstract_models/abstract_classifier_model.py index 74cec045..c131c9bb 100644 --- a/embiggen/utils/abstract_models/abstract_classifier_model.py +++ b/embiggen/utils/abstract_models/abstract_classifier_model.py @@ -4,6 +4,7 @@ import numpy as np import pandas as pd import time +from userinput.utils import must_be_in_set from tqdm.auto import trange, tqdm from embiggen.utils.abstract_models.list_formatting import format_list from cache_decorator import Cache @@ -238,17 +239,11 @@ def normalize_node_feature( node_feature.set_random_state(random_state) if smoke_test: - node_feature = node_feature.__class__( - **node_feature.smoke_test_parameters() - ) + node_feature = node_feature.into_smoke_test() node_feature = node_feature.fit_transform( graph=graph, - return_dataframe=False, - # If this is an Ensmallen model, we can enable the verbosity - # as it will only show up in the jupyter kernel and it won't bother the - # other loading bars. - verbose="Ensmallen" == node_feature.library_name() + return_dataframe=False ) if isinstance(node_feature, EmbeddingResult): @@ -441,14 +436,11 @@ def normalize_node_type_feature( node_type_feature.set_random_state(random_state) if smoke_test: - node_type_feature = node_type_feature.__class__( - **node_type_feature.smoke_test_parameters() - ) + node_type_feature = node_type_feature.into_smoke_test() node_type_feature = node_type_feature.fit_transform( graph=graph, return_dataframe=False, - verbose=False ) if isinstance(node_type_feature, EmbeddingResult): @@ -641,14 +633,11 @@ def normalize_edge_feature( edge_feature.set_random_state(random_state) if smoke_test: - edge_feature = edge_feature.__class__( - **edge_feature.smoke_test_parameters() - ) + edge_feature = edge_feature.into_smoke_test() edge_feature = edge_feature.fit_transform( graph=graph, return_dataframe=False, - verbose=False ) if isinstance(edge_feature, EmbeddingResult): @@ -670,7 +659,7 @@ def normalize_edge_feature( ) ) - if graph.get_directed_edges_number() != ef.shape[0]: + if graph.get_number_of_directed_edges() != ef.shape[0]: raise ValueError( ( "The provided edge features have {rows_number} rows " @@ -806,6 +795,9 @@ def fit( f"the {self.model_name()} requires edge types." ) + if support is None: + support = graph + try: self._fit( graph=graph, @@ -872,22 +864,29 @@ def predict( "Do call the `.fit` method before the `.predict` " "method." )) + + if support is None: + support = graph + try: - return self._predict( + predictions = self._predict( graph=graph, support=support, node_features=self.normalize_node_features( graph=graph, + random_state=self._random_state, node_features=node_features, allow_automatic_feature=False, ), node_type_features=self.normalize_node_type_features( graph=graph, + random_state=self._random_state, node_type_features=node_type_features, allow_automatic_feature=True, ), edge_features=self.normalize_edge_features( graph=graph, + random_state=self._random_state, edge_features=edge_features, allow_automatic_feature=False, ), @@ -898,6 +897,8 @@ def predict( f"implemented using the {self.library_name()} for the {self.task_name()} task. " f"Specifically, the class of the model is {self.__class__.__name__}." ) from e + + return predictions def predict_proba( self, @@ -934,8 +935,11 @@ def predict_proba( "method." )) + if support is None: + support = graph + try: - return self._predict_proba( + predictions = self._predict_proba( graph=graph, support=support, node_features=self.normalize_node_features( @@ -964,6 +968,23 @@ def predict_proba( f"Specifically, the class of the model is {self.__class__.__name__}." ) from e + if ( + not self.is_binary_prediction_task() and + not self.is_multilabel_prediction_task() and + ( + len(predictions.shape) == 1 or + predictions.shape[1] == 2 + ) + ): + raise ValueError( + "This task is not a binary prediction, " + f"yet the {self.model_name()} model predict proba method has " + "returned a vector of prediction probabilities with " + f"shape {predictions.shape}." + ) + + return predictions + def evaluate_predictions( self, ground_truth: np.ndarray, @@ -1002,7 +1023,7 @@ def evaluate_predictions( metric.__name__: metric( ground_truth, predictions, - average="binary" if self.is_binary_prediction_task() else "macro", + average="macro", zero_division=0 ) for metric in ( @@ -1027,7 +1048,6 @@ def evaluate_prediction_probabilities( prediction_probabilities: np.ndarray The predictions to be evaluated. """ - metrics = [] if self.is_binary_prediction_task(): return { "auroc": express_measures.binary_auroc( @@ -1044,18 +1064,16 @@ def evaluate_prediction_probabilities( def wrapper_roc_auc_score(*args, **kwargs): return roc_auc_score(*args, **kwargs, multi_class="ovr") - metrics.append(wrapper_roc_auc_score) - return { - metric.__name__: metric( + "auroc": wrapper_roc_auc_score( ground_truth, - prediction_probabilities, + prediction_probabilities ) - for metric in metrics } - @staticmethod + @classmethod def split_graph_following_evaluation_schema( + cls, graph: Graph, evaluation_schema: str, holdout_number: int, @@ -1077,9 +1095,10 @@ def split_graph_following_evaluation_schema( holdouts_kwargs: Dict[str, Any] The kwargs to be forwarded to the holdout method. """ - raise NotImplementedError( - "The `split_graph_following_evaluation_schema` method should be implemented " - "in the child classes of abstract classifier model." + must_be_in_set( + evaluation_schema, + cls.get_available_evaluation_schemas(), + f"evaluation schema for {cls.task_name()} and model {cls.model_name()}" ) def _evaluate( @@ -1211,7 +1230,7 @@ def iterate_classifier_models( cache_path="{cache_dir}/{self.task_name()}/{graph.get_name()}/holdout_{holdout_number}/{self.model_name()}/{self.library_name()}/{_hash}.csv.gz", cache_dir="experiments", enable_cache_arg_name="enable_cache", - args_to_ignore=["verbose", "smoke_test"], + args_to_ignore=["verbose", "smoke_test", "train_of_interest", "test_of_interest", "train",], capture_enable_cache_arg_name=True, use_approximated_hash=True ) diff --git a/embiggen/utils/abstract_models/abstract_embedding_model.py b/embiggen/utils/abstract_models/abstract_embedding_model.py index 4e9e75ac..6f4d7a00 100644 --- a/embiggen/utils/abstract_models/abstract_embedding_model.py +++ b/embiggen/utils/abstract_models/abstract_embedding_model.py @@ -41,10 +41,10 @@ def __init__( def parameters(self) -> Dict[str, Any]: """Returns parameters of the embedding model.""" - return { + return dict( **super().parameters(), - "embedding_size": self._embedding_size - } + embedding_size=self._embedding_size + ) @classmethod def smoke_test_parameters(cls) -> Dict[str, Any]: @@ -63,11 +63,15 @@ def requires_nodes_sorted_by_decreasing_node_degree(cls) -> bool: "in the child classes of abstract model." )) + @classmethod + def get_minimum_required_number_of_node_types(cls) -> int: + """Requires minimum number of required node types.""" + return 0 + def _fit_transform( self, graph: Graph, return_dataframe: bool = True, - verbose: bool = True ) -> EmbeddingResult: """Run embedding on the provided graph. @@ -75,8 +79,6 @@ def _fit_transform( -------------------- graph: Graph The graph to run predictions on. - verbose: bool - Whether to show loading bars. """ raise NotImplementedError(( "The `_fit_transform` method must be implemented " @@ -87,13 +89,11 @@ def _fit_transform( cache_path="{cache_dir}/{self.model_name()}/{self.library_name()}/{graph.get_name()}/{_hash}.pkl.gz", cache_dir="embedding", enable_cache_arg_name="self._enable_cache", - args_to_ignore=["verbose"] ) def _cached_fit_transform( self, graph: Graph, return_dataframe: bool = True, - verbose: bool = True ) -> EmbeddingResult: """Execute embedding on the provided graph. @@ -103,8 +103,6 @@ def _cached_fit_transform( The graph to run embedding on. return_dataframe: bool = True Whether to return a pandas DataFrame with the embedding. - verbose: bool = True - Whether to show loading bars. Returns -------------------- @@ -119,7 +117,7 @@ def _cached_fit_transform( if not graph.has_nodes_sorted_by_decreasing_outbound_node_degree(): raise ValueError( f"The given graph {graph.get_name()} does not have the nodes sorted by decreasing " - "order, therefore the negative sampling (which follows a zipfian " + "order, therefore the negative sampling (which follows a scale free " "distribution) would not approximate well the Softmax.\n" "In order to sort the given graph in such a way that the node IDs " "are sorted by decreasing outbound node degrees, you can use " @@ -131,6 +129,14 @@ def _cached_fit_transform( f"The provided graph {graph.get_name()} does not have node types, but " f"the {self.model_name()} requires node types." ) + + if self.requires_node_types() and graph.get_number_of_node_types() <= 1: + raise ValueError( + f"The {self.model_name()} requires the graph to have " + f"at least {self.get_minimum_required_number_of_node_types()} node types, " + f"but the provided one has {graph.get_number_of_node_types()} " + "node types." + ) if self.requires_edge_types() and not graph.has_edge_types(): raise ValueError( @@ -173,7 +179,6 @@ def _cached_fit_transform( result = self._fit_transform( graph=graph, return_dataframe=return_dataframe, - verbose=verbose ) if not isinstance(result, EmbeddingResult): @@ -192,7 +197,6 @@ def fit_transform( repository: Optional[str] = None, version: Optional[str] = None, return_dataframe: bool = True, - verbose: bool = True ) -> EmbeddingResult: """Execute embedding on the provided graph. @@ -212,8 +216,6 @@ def fit_transform( from the ensmallen automatic retrieval. return_dataframe: bool = True Whether to return a pandas DataFrame with the embedding. - verbose: bool = True - Whether to show loading bars. Returns -------------------- @@ -228,5 +230,4 @@ def fit_transform( return self._cached_fit_transform( graph=graph, return_dataframe=return_dataframe, - verbose=verbose ) diff --git a/embiggen/utils/abstract_models/abstract_model.py b/embiggen/utils/abstract_models/abstract_model.py index eb714a29..e38ddd9a 100644 --- a/embiggen/utils/abstract_models/abstract_model.py +++ b/embiggen/utils/abstract_models/abstract_model.py @@ -3,7 +3,7 @@ from embiggen.utils.abstract_models.list_formatting import format_list from typing import Dict, Any, Type, List, Optional from dict_hash import Hashable, sha256 -from userinput.utils import closest +from userinput.utils import must_be_in_set import inspect @@ -15,6 +15,7 @@ def is_not_implemented(method: Callable) -> bool: """Returns whether this method contains a raise for not being implemented.""" return "raise NotImplementedError" in inspect.getsource(method) + def is_implemented(method: Callable) -> bool: """Returns whether this method is implemented.""" return not is_not_implemented(method) @@ -46,9 +47,14 @@ def __init__(self, random_state: Optional[int] = None): f"random state of `{random_state}` was provided." ) + can_use_edge_weights = self.__getattribute__("can_use_edge_weights") + requires_positive_edge_weights = self.__getattribute__( + "requires_positive_edge_weights") + if ( - not self.__getattribute__("can_use_edge_weights")() and - is_implemented(self.__getattribute__("requires_positive_edge_weights")) + is_implemented(can_use_edge_weights) and + not can_use_edge_weights() and + is_implemented(requires_positive_edge_weights) ): raise ValueError( "We have found an useless method in the " @@ -128,6 +134,13 @@ def parameters(self) -> Dict[str, Any]: random_state=self._random_state ) + def into_smoke_test(self) -> "Self": + """Creates new instance with smoke test parameters.""" + return self.__class__(**{ + **self.parameters(), + **self.smoke_test_parameters() + }) + @classmethod def requires_edge_weights(cls) -> bool: """Returns whether the model requires edge weights.""" @@ -153,7 +166,10 @@ def task_involves_edge_weights(cls) -> bool: def can_use_edge_weights(cls) -> bool: """Returns whether the model can optionally use edge weights.""" try: - if cls.requires_edge_weights(): + if ( + is_implemented(cls.requires_edge_weights) and + cls.requires_edge_weights() + ): return True except NotImplementedError: pass @@ -332,9 +348,9 @@ def consistent_hash(self) -> str: """Returns consistent hash describing the model.""" return sha256(dict( **self.parameters(), - model_name= self.model_name(), - library_name= self.library_name(), - task_name= self.task_name(), + model_name=self.model_name(), + library_name=self.library_name(), + task_name=self.task_name(), )) @staticmethod @@ -359,38 +375,16 @@ def set_random_state(self, random_state: int): ) self._random_state = random_state - @staticmethod - def get_model_data( - model_name: str - ) -> Dict[str, Dict]: - """Returns data relative to the registered model data.""" - # We check if the provided string is not an empty string. - if len(model_name) == 0: - raise ValueError( - "The provided model name is empty." - ) - - # We turn this to lowercase in order to allow - # for error in casing of the model, since one may - # write models like `GloVe` also as `Glove` or other - # typos, which are generally easy to make. - lowercase_model_mapping = AbstractModel.get_available_model_names_in_lowercase_mapping() - if model_name.lower() not in lowercase_model_mapping: - raise ValueError( - f"The provided model name `{model_name}` is not available. " - f"Did you mean {closest(model_name, lowercase_model_mapping.values())}?" - ) - # We retrieve the model standard name. - model_name = lowercase_model_mapping[model_name.lower()] - return AbstractModel.MODELS_LIBRARY[model_name] - @staticmethod def get_task_data( model_name: str, task_name: str ) -> Dict[str, Dict]: """Returns data relative to the registered model and task data.""" - model_data = AbstractModel.get_model_data(model_name) + if len(model_name) == 0: + raise ValueError( + "The provided model name is empty." + ) # We check if the provided string is not an empty string. if len(task_name) == 0: @@ -398,25 +392,20 @@ def get_task_data( "The provided task name is empty." ) - # We do a similar check as the one above for the tasks, - # as one may do typos while writig the task name and - # we should always provide the best possible help message. - lowercase_task_mapping = { - t.lower(): t - for t in model_data.keys() - } - if task_name.lower() not in lowercase_task_mapping: - raise ValueError( - f"The provided task name `{task_name}` is not available for " - f"the requested model {model_name}." - f"Did you mean {closest(task_name, lowercase_task_mapping.values())}?" - ) + task_name = must_be_in_set( + task_name, + AbstractModel.MODELS_LIBRARY, + "task name" + ) - # We retrieve the task standard name. - task_name = lowercase_task_mapping[task_name.lower()] + model_name = must_be_in_set( + model_name, + AbstractModel.MODELS_LIBRARY[task_name], + "model name" + ) # We retrieve the task data. - return model_data[task_name] + return AbstractModel.MODELS_LIBRARY[task_name][model_name] @staticmethod def get_library_data( @@ -433,25 +422,18 @@ def get_library_data( "The provided library name is empty." ) - lowercase_libraries_mapping = { - t.lower(): t - for t in task_data.keys() - } - if library_name.lower() not in lowercase_libraries_mapping: - raise ValueError( - f"The provided library name `{library_name}` is not available for " - f"the requested model {model_name}. " - f"Did you mean {closest(library_name, lowercase_libraries_mapping.values())}?" - ) - - # We retrieve the library standard name. - library_name = lowercase_libraries_mapping[library_name.lower()] + library_name = must_be_in_set( + library_name, + task_data.keys(), + "library name" + ) # We retrieve the library data. return task_data[library_name] - @staticmethod + @classmethod def get_model_from_library( + cls, model_name: str, task_name: Optional[str] = None, library_name: Optional[str] = None, @@ -474,18 +456,7 @@ def get_model_from_library( be raised. """ if task_name is None: - task_names = list(AbstractModel.get_model_data(model_name).keys()) - if len(task_names) == 1: - task_name = task_names[0] - else: - formatted_list = format_list(task_names) - raise ValueError( - f"The requested model `{model_name}` is available for " - "multiple tasks and no specific task was requested, " - "so it is unclear which task you intend to execute. " - f"Specifically, the available tasks are {formatted_list}." - "Please do provide a task name to resolve this ambiguity." - ) + task_name = cls.task_name() task_data = AbstractModel.get_task_data(model_name, task_name) @@ -523,19 +494,6 @@ def get_model_from_library( # its class to let the user do whathever they want. return model_class - @staticmethod - def get_available_model_names() -> List[str]: - """Returns list of available model names.""" - return list(AbstractModel.MODELS_LIBRARY.keys()) - - @staticmethod - def get_available_model_names_in_lowercase_mapping() -> Dict[str, str]: - """Returns list of available model names in lowercase.""" - return { - model_name.lower(): model_name - for model_name in AbstractModel.get_available_model_names() - } - @staticmethod def find_available_models( model_name: str, @@ -565,20 +523,20 @@ def register(model_class: Type["AbstractModel"]): model_class: Type["AbstractModel"] The class to register. """ + task_name = model_class.task_name() model_name = model_class.model_name() # If this is the first model of its kind to be registered. - if model_name not in AbstractModel.MODELS_LIBRARY: - AbstractModel.MODELS_LIBRARY[model_name] = {} + if task_name not in AbstractModel.MODELS_LIBRARY: + AbstractModel.MODELS_LIBRARY[task_name] = {} # We retrieve the data for the model to enrich it. # This is NOT a copy, but a reference to the same STATIC object. - model_data = AbstractModel.MODELS_LIBRARY[model_name] + model_data = AbstractModel.MODELS_LIBRARY[task_name] - task_name = model_class.task_name() - if task_name not in model_data: - model_data[task_name] = {} + if model_name not in model_data: + model_data[model_name] = {} - task_data = model_data[task_name] + task_data = model_data[model_name] class_name = model_class.__name__ diff --git a/embiggen/utils/abstract_models/model_stub.py b/embiggen/utils/abstract_models/model_stub.py index 64c0d597..cfaa7f75 100644 --- a/embiggen/utils/abstract_models/model_stub.py +++ b/embiggen/utils/abstract_models/model_stub.py @@ -180,6 +180,11 @@ def can_use_edge_weights(cls) -> bool: """Returns whether the model can optionally use edge weights.""" return None + @classmethod + def is_stocastic(cls) -> bool: + """Returns whether the model can optionally use edge weights.""" + return None + @classmethod def is_topological(cls) -> str: """Returns whether this embedding is based on graph topology.""" @@ -193,7 +198,7 @@ def task_name(cls) -> str: @classmethod def is_available(cls) -> bool: """Returns whether the model class is actually available in the user system.""" - return True + return False model_class = StubClass else: diff --git a/embiggen/utils/networkx_utils.py b/embiggen/utils/networkx_utils.py index 5bc75832..c0c89ce0 100644 --- a/embiggen/utils/networkx_utils.py +++ b/embiggen/utils/networkx_utils.py @@ -25,7 +25,7 @@ def convert_ensmallen_graph_to_networkx_graph( (src_name, dst_name, edge_weight) for (src_name, dst_name), edge_weight in zip( graph.get_directed_edge_node_ids(), - graph.get_edge_weights() + graph.get_directed_edge_weights() ) ]) else: diff --git a/embiggen/utils/normalize_model_structural_parameters.py b/embiggen/utils/normalize_model_structural_parameters.py index 10831f36..1c68067f 100644 --- a/embiggen/utils/normalize_model_structural_parameters.py +++ b/embiggen/utils/normalize_model_structural_parameters.py @@ -1,9 +1,9 @@ """Submodule providing utilities to normalize parameters for modular models.""" -from typing import List, Any, Union, Type +from typing import List, Any, Union, Type, Optional def normalize_model_list_parameter( - candidate_list: Union[Any, List[Any]], + candidate_list: Optional[Union[Any, List[Any]]], elements_number: int, object_type: Type, default_value: Any = None, @@ -104,7 +104,7 @@ def normalize_model_list_parameter( def normalize_model_ragged_list_parameter( - candidate_ragged_list: Union[Any, List[Any], List[List[Any]]], + candidate_ragged_list: Optional[Union[Any, List[Any], List[List[Any]]]], submodules_number: int, layers_per_submodule: List[int], object_type: Type, diff --git a/embiggen/visualizations/graph_visualizer.py b/embiggen/visualizations/graph_visualizer.py index e4487f1f..f26cd77a 100644 --- a/embiggen/visualizations/graph_visualizer.py +++ b/embiggen/visualizations/graph_visualizer.py @@ -1,7 +1,7 @@ """Module with embedding visualization tools.""" import functools from multiprocessing import cpu_count -from typing import Dict, Iterator, List, Tuple, Union, Optional, Callable +from typing import Dict, Iterator, List, Tuple, Union, Optional, Callable, Type import matplotlib import matplotlib.pyplot as plt @@ -22,14 +22,32 @@ from sklearn.tree import DecisionTreeClassifier from matplotlib.legend_handler import HandlerBase, HandlerTuple from matplotlib import collections as mc +from mpl_toolkits.mplot3d.art3d import Line3DCollection from sanitize_ml_labels import sanitize_ml_labels from sklearn.decomposition import PCA +from userinput.utils import must_be_in_set from sklearn.metrics import balanced_accuracy_score from sklearn.model_selection import ShuffleSplit import itertools +from embiggen.utils.abstract_models.abstract_embedding_model import AbstractEmbeddingModel from embiggen.utils.abstract_models.embedding_result import EmbeddingResult +from mpl_toolkits.mplot3d.proj3d import proj_transform +from matplotlib.text import Annotation + +class Annotation3D(Annotation): + '''Annotate the point xyz with text s''' + + def __init__(self, s, xyz, *args, **kwargs): + Annotation.__init__(self,s, xy=(0,0), *args, **kwargs) + self._verts3d = xyz + + def draw(self, renderer): + xs3d, ys3d, zs3d = self._verts3d + xs, ys, zs = proj_transform(xs3d, ys3d, zs3d, renderer.M) + self.xy=(xs,ys) + Annotation.draw(self, renderer) try: from ddd_subplots import subplots as subplots_3d, rotate, display_video_at_path @@ -89,6 +107,7 @@ def __init__( destination_nodes_prefixes: Optional[Union[str, List[str]]] = None, edge_type_names: Optional[List[Optional[str]]] = None, show_graph_name: Union[str, bool] = "auto", + number_of_columns_in_legend: int = 2, show_node_embedding_method: bool = True, show_edge_embedding_method: bool = True, show_separability_considerations_explanation: bool = True, @@ -220,6 +239,8 @@ def __init__( Whether to show the graph name in the plots. By default, it is shown if the graph does not have a trivial name such as `Graph`. + number_of_columns_in_legend: int = 2 + The number of columns to be used with the legend. show_node_embedding_method: bool = True Whether to show the node embedding method. By default, we show it if we can detect it. @@ -288,6 +309,7 @@ def __init__( self._support = support self._subgraph_of_interest = subgraph_of_interest + self._number_of_columns_in_legend = number_of_columns_in_legend if isinstance(source_node_types_names, str): source_node_types_names = [source_node_types_names] @@ -345,10 +367,10 @@ def __init__( ) # We sample the negative edges using the subgraph of interest as base graph - # to follow its zipfian distribution, which may be different from the - # main graph zipfian distribution when particular filters are applied to it. - # For instance, the zipfian distribution of one particular edge type - # may be very different from the whole graph zipfian distribution. + # to follow its scale free distribution, which may be different from the + # main graph scale free distribution when particular filters are applied to it. + # For instance, the scale free distribution of one particular edge type + # may be very different from the whole graph scale free distribution. # Furthermore, we avoid sampling false negatives by passing to the # method also the support graph. try: @@ -358,7 +380,7 @@ def __init__( self._positive_graph.get_number_of_edges() ), random_state=random_state, - use_zipfian_sampling=True, + use_scale_free_distribution=True, graph_to_avoid=self._support, support=self._support, only_from_same_component=only_from_same_component, @@ -408,7 +430,12 @@ def __init__( decomposition_kwargs = {} self._n_components = n_components - self._decomposition_method = decomposition_method + + self._decomposition_method = must_be_in_set( + decomposition_method, + ("PCA", "TSNE", "UMAP"), + "decomposition method" + ) self._decomposition_kwargs = decomposition_kwargs def iterate_subsampled_node_ids(self) -> Iterator[int]: @@ -521,7 +548,9 @@ def get_decomposition_method(self) -> Callable: # Adding a warning for when decomposing methods that # embed nodes using a cosine similarity / distance approach # in order to avoid false negatives. - if self._node_embedding_method_name == "GloVe": + if self._decomposition_method in ("UMAP", "TSNE") and self._node_embedding_method_name in ( + "Node2Vec GloVe", "DeepWalk GloVe", "First-order LINE" + ): metric = self._decomposition_kwargs.get("metric") if metric is not None and metric != "cosine": warnings.warn( @@ -529,7 +558,7 @@ def get_decomposition_method(self) -> Callable: "such as Glove, which embeds nodes using a dot product, it is " "highly suggested to use a `cosine` metric. Using a different " f"metric, such as the one you have provided ({metric}) may lead " - "to unsuccessfull decompositions using UMAP or t-SNE." + "to worse decompositions using UMAP or t-SNE." ) else: # Otherwise we switch to using a cosine metric. @@ -556,7 +585,7 @@ def get_decomposition_method(self) -> Callable: leave=False, dynamic_ncols=True ), - verbose=True, + verbose=self._verbose, ), **self._decomposition_kwargs }).fit_transform @@ -599,6 +628,7 @@ def get_decomposition_method(self) -> Callable: n_jobs=cpu_count(), random_state=self._random_state, verbose=self._verbose, + learning_rate=200, n_iter=400, init="random", method="exact" if self._n_components == 4 else "barnes_hut", @@ -631,14 +661,10 @@ def get_decomposition_method(self) -> Callable: ), **self._decomposition_kwargs }).fit_transform - else: - raise ValueError( - "We currently only support PCA and TSNE decomposition methods." - ) def _get_node_embedding( self, - node_embedding: Optional[Union[pd.DataFrame, np.ndarray, str, EmbeddingResult]] = None, + node_embedding: Optional[Union[pd.DataFrame, np.ndarray, str, EmbeddingResult, Type[AbstractEmbeddingModel]]] = None, **node_embedding_kwargs: Dict ) -> np.ndarray: """Computes the node embedding if it was not otherwise provided. @@ -666,7 +692,10 @@ def _get_node_embedding( ) else: node_embedding = self._node_embedding_method_name - if isinstance(node_embedding, str): + if ( + isinstance(node_embedding, str) or + issubclass(node_embedding.__class__, AbstractEmbeddingModel) + ): node_embedding = embed_graph( graph=self._graph, embedding_model=node_embedding, @@ -678,7 +707,6 @@ def _get_node_embedding( self._has_autodetermined_node_embedding_name = True self._node_embedding_method_name = node_embedding.embedding_method_name node_embedding = np.hstack(node_embedding.get_all_node_embedding()) - elif self._node_embedding_method_name == "auto" or self._has_autodetermined_node_embedding_name: self._has_autodetermined_node_embedding_name = True self._node_embedding_method_name = self.automatically_detect_node_embedding_method( @@ -798,7 +826,7 @@ def _set_legend( number_of_columns = 1 if len(labels) <= 2 and any( len(label) > 20 for label in labels - ) else 2 + ) else self._number_of_columns_in_legend legend = axes.legend( handles=handles, labels=[ @@ -827,9 +855,6 @@ def _set_legend( def automatically_detect_node_embedding_method(self, node_embedding: np.ndarray) -> Optional[str]: """Detect node embedding method using heuristics, where possible.""" - # Rules to detect SPINE embedding - if node_embedding.dtype == "uint8" and node_embedding.min() == 0: - return "SPINE" # Rules to detect TFIDF/BERT embedding if node_embedding.dtype == "float16" and node_embedding.shape[1] == 768: return "TFIDF-weighted BERT" @@ -877,7 +902,7 @@ def fit_nodes( size=self._number_of_subsampled_nodes ) node_transformer = NodeTransformer( - aligned_node_mapping=True + aligned_mapping=True ) node_transformer.fit(node_embedding) node_embedding = node_transformer.transform( @@ -899,7 +924,8 @@ def _get_positive_edges_embedding( """ graph_transformer = GraphTransformer( method=self._edge_embedding_method, - aligned_node_mapping=True + aligned_mapping=True, + include_both_undirected_edges=False ) graph_transformer.fit(node_embedding) return graph_transformer.transform(self._positive_graph) @@ -940,7 +966,8 @@ def _get_negative_edge_embedding( """ graph_transformer = GraphTransformer( method=self._edge_embedding_method, - aligned_node_mapping=True + aligned_mapping=True, + include_both_undirected_edges=False ) graph_transformer.fit(node_embedding) return graph_transformer.transform( @@ -1020,19 +1047,30 @@ def _get_figure_and_axes( "them can be None." )) if self._n_components == 2: - figure, axes = plt.subplots(**{ - **GraphVisualizer.DEFAULT_SUBPLOT_KWARGS, - **kwargs - }) - figure.patch.set_facecolor("white") - axes.axis('equal') + figure, axes = plt.subplots( + **{ + **GraphVisualizer.DEFAULT_SUBPLOT_KWARGS, + **kwargs + }, + squeeze=False + ) + for ax in axes.flatten(): + ax.axis('equal') + if axes.size == 1: + axes = axes.flatten()[0] else: - figure, axes = subplots_3d(**{ - **GraphVisualizer.DEFAULT_SUBPLOT_KWARGS, - **kwargs - }) - figure.patch.set_facecolor("white") - axes.axis('auto') + figure, axes = subplots_3d( + **{ + **GraphVisualizer.DEFAULT_SUBPLOT_KWARGS, + **kwargs + }, + squeeze=False + ) + for ax in axes.flatten(): + ax.axis('auto') + if axes.size == 1: + axes = axes.flatten()[0] + figure.patch.set_facecolor("white") return figure, axes def _get_complete_title( @@ -1208,15 +1246,16 @@ def _plot_scatter( if train_indices is None and test_indices is None: scatter = axes.scatter( *points.T, - **{ + **dict( **dict( c=colors, - edgecolors=None if edgecolors is None else cmap(edgecolors), + edgecolors=None if edgecolors is None else cmap( + edgecolors), marker=train_marker, cmap=cmap, ), **scatter_kwargs - } + ) ) collections.append(scatter) legend_elements += scatter.legend_elements()[0] @@ -1325,19 +1364,6 @@ def _plot_scatter( def _wrapped_plot_scatter(self, **kwargs): if self._rotate: - # These backups are needed for two reasons: - # 1) Processes in python necessarily copy the instance objects for each process - # and this can cause a considerable memery peak to occour. - # 2) Some of the objects considered are not picklable, such as, at the time of writing - # the lambdas used in the graph transformer or the graph object itself. - graph_backup = self._graph - node_embedding = self._node_decomposition - edge_embedding = self._positive_edge_decomposition - negative_edge_embedding = self._negative_edge_decomposition - self._node_decomposition = None - self._positive_edge_decomposition = None - self._negative_edge_decomposition = None - self._graph = None try: kwargs["loc"] = "lower right" path = "{}.{}".format( @@ -1350,20 +1376,15 @@ def _wrapped_plot_scatter(self, **kwargs): path=path, duration=self._duration, fps=self._fps, - verbose=True, + verbose=self._verbose, **kwargs ) except (Exception, KeyboardInterrupt) as e: - self._node_decomposition = node_embedding - self._positive_edge_decomposition = edge_embedding - self._negative_edge_decomposition = negative_edge_embedding - self._graph = graph_backup raise e - self._node_decomposition = node_embedding - self._positive_edge_decomposition = edge_embedding - self._negative_edge_decomposition = negative_edge_embedding - self._graph = graph_backup - return display_video_at_path(path) + to_display = display_video_at_path(path) + if to_display is None: + return () + return to_display return self._plot_scatter(**kwargs) def _plot_types( @@ -1634,22 +1655,46 @@ def plot_edge_segments( add_selfloops_where_missing=False, complete=False, ) + edge_node_ids = np.array([ + [ + np.where(self._subsampled_node_ids == src)[0][0], + np.where(self._subsampled_node_ids == dst)[0][0] + ] + for src, dst in edge_node_ids + if src in self._subsampled_node_ids and dst in self._subsampled_node_ids + ]) else: edge_node_ids = self._graph.get_edge_node_ids( directed=False ) - lines_collection = mc.LineCollection( - self._node_decomposition[edge_node_ids], - linewidths=1, - zorder=0, - **{ - **GraphVisualizer.DEFAULT_EDGES_SCATTER_KWARGS, - **( - {} if scatter_kwargs is None else scatter_kwargs - ) - } - ) + if edge_node_ids.size == 0: + return figure, axes + + if self._n_components == 3: + lines_collection = Line3DCollection( + self._node_decomposition[edge_node_ids], + linewidths=1, + zorder=0, + **{ + **GraphVisualizer.DEFAULT_EDGES_SCATTER_KWARGS, + **( + {} if scatter_kwargs is None else scatter_kwargs + ) + } + ) + else: + lines_collection = mc.LineCollection( + self._node_decomposition[edge_node_ids], + linewidths=1, + zorder=0, + **{ + **GraphVisualizer.DEFAULT_EDGES_SCATTER_KWARGS, + **( + {} if scatter_kwargs is None else scatter_kwargs + ) + } + ) axes.add_collection(lines_collection) return figure, axes @@ -1759,7 +1804,7 @@ def plot_nodes( **kwargs ) - if annotate_nodes: + if annotate_nodes and returned_values: figure, axes = returned_values[:2] self.annotate_nodes( figure=figure, @@ -1775,15 +1820,26 @@ def annotate_nodes( axes: Axes, points: np.ndarray ) -> Tuple[Figure, Axes]: - if self._subsampled_node_ids is not None: - node_names = [ - self._graph.get_node_name_from_node_id(node_id) - for node_id in self._subsampled_node_ids - ] - else: - node_names = self._graph.get_node_names() - for i, txt in enumerate(node_names): - axes.annotate(txt, points[i], fontsize=8, ha="center", va="center") + for node_name, point in zip(( + self._graph.get_node_name_from_node_id(node_id) + for node_id in self.iterate_subsampled_node_ids() + ), points): + if point.size == 3: + axes.add_artist(Annotation3D( + node_name, + point, + fontsize=8, + ha="center", + va="center" + )) + else: + axes.annotate( + node_name, + point, + fontsize=8, + ha="center", + va="center" + ) return (figure, axes) def plot_edges( @@ -2047,6 +2103,16 @@ def _plot_positive_and_negative_edges_metric( edge_metric_callback(subgraph=self._positive_graph), )) + # Filter the edge metrics relative to edges that are not + # to be displayed. + if not self._graph.is_directed(): + edge_metrics = edge_metrics[np.concatenate([ + self._negative_graph.get_directed_source_node_ids( + ) <= self._negative_graph.get_directed_destination_node_ids(), + self._positive_graph.get_directed_source_node_ids( + ) <= self._positive_graph.get_directed_destination_node_ids(), + ])] + points = np.vstack([ self._negative_edge_decomposition, self._positive_edge_decomposition, @@ -2187,6 +2253,11 @@ def _plot_positive_and_negative_edges_metric_histogram( figure, axes = plt.subplots(figsize=(5, 5)) figure.patch.set_facecolor("white") + if self._negative_graph.is_directed() and edge_metrics is None: + number_of_negative_edges = self._negative_graph.get_number_of_directed_edges() + else: + number_of_negative_edges = self._negative_graph.get_number_of_undirected_edges() + if edge_metrics is None: edge_metrics = np.concatenate(( edge_metric_callback(subgraph=self._negative_graph), @@ -2195,8 +2266,8 @@ def _plot_positive_and_negative_edges_metric_histogram( axes.hist( [ - edge_metrics[:self._negative_graph.get_number_of_directed_edges()], - edge_metrics[self._negative_graph.get_number_of_directed_edges():], + edge_metrics[:number_of_negative_edges], + edge_metrics[number_of_negative_edges:], ], bins=10, log=True, @@ -2652,13 +2723,10 @@ def plot_positive_and_negative_edges_resource_allocation_index( def _get_flatten_unknown_node_ontologies(self) -> Tuple[List[str], np.ndarray]: """Returns unique ontologies and node ontologies adjusted for the current instance.""" - if self._subsampled_node_ids is None: - ontology_names = self._graph.get_node_ontologies() - else: - ontology_names = [ - self._graph.get_ontology_from_node_id(node_id) - for node_id in self._subsampled_node_ids - ] + ontology_names = [ + self._graph.get_ontology_from_node_id(node_id) + for node_id in self.iterate_subsampled_node_ids() + ] # The following is needed to normalize the multiple types ontologies_counts = Counter(ontology_names) @@ -2704,13 +2772,6 @@ def _get_flatten_multi_label_and_unknown_node_types(self) -> np.ndarray: node_types_number = self._graph.get_number_of_node_types() unknown_node_types_id = node_types_number - # According to whether the subsampled node IDs were given, - # we iterate on them or on the complete set of nodes of the graph. - if self._subsampled_node_ids is None: - nodes_iterator = range(self._graph.get_number_of_nodes()) - else: - nodes_iterator = self._subsampled_node_ids - # When we have multiple node types for a given node, we set it to # the most common node type of the set. return np.fromiter( @@ -2725,7 +2786,7 @@ def _get_flatten_multi_label_and_unknown_node_types(self) -> np.ndarray: )[0] for node_type_ids in ( self._graph.get_node_type_ids_from_node_id(node_id) - for node_id in nodes_iterator + for node_id in self.iterate_subsampled_node_ids() ) ), dtype=np.uint32 @@ -2743,7 +2804,12 @@ def _get_flatten_unknown_edge_types(self) -> np.ndarray: if edge_type_id is None else edge_type_id - for edge_type_id in self._positive_graph.get_edge_type_ids() + for edge_type_id in ( + self._positive_graph.get_directed_edge_type_ids() + if self._positive_graph.is_directed() + else + self._positive_graph.get_undirected_edge_type_ids() + ) ), dtype=np.uint32 ) @@ -2850,6 +2916,29 @@ def plot_node_types( **kwargs ) + if self._subsampled_node_ids is not None: + if node_type_predictions is not None: + node_type_predictions = node_type_predictions[self._subsampled_node_ids] + + if train_indices is not None: + train_indices = np.fromiter( + ( + np.where(self._subsampled_node_ids == node_id)[0][0] + for node_id in train_indices + if node_id in self._subsampled_node_ids + ), + dtype=train_indices.dtype + ) + if test_indices is not None: + test_indices = np.fromiter( + ( + np.where(self._subsampled_node_ids == node_id)[0][0] + for node_id in test_indices + if node_id in self._subsampled_node_ids + ), + dtype=test_indices.dtype + ) + if annotate_nodes == "auto": annotate_nodes = self._graph.get_number_of_nodes() < 50 and not self._rotate @@ -3267,7 +3356,9 @@ def plot_connected_components( points=self._node_decomposition, ) - if not return_caption: + if not return_caption or self._rotate: + if self._rotate: + return returned_values return self._handle_notebook_display(*returned_values) # TODO! Add caption node abount gaussian ball! @@ -3357,16 +3448,13 @@ def plot_node_degrees( "method before plotting the nodes." ) - if self._subsampled_node_ids is None: - degrees = self._support.get_node_degrees() - else: - degrees = np.fromiter( - ( - self._support.get_node_degree_from_node_id(node_id) - for node_id in self._subsampled_node_ids - ), - dtype=np.uint32 - ) + degrees = np.fromiter( + ( + self._support.get_node_degree_from_node_id(node_id) + for node_id in self.iterate_subsampled_node_ids() + ), + dtype=np.uint32 + ) if annotate_nodes == "auto": annotate_nodes = self._graph.get_number_of_nodes() < 50 and not self._rotate @@ -3644,7 +3732,10 @@ def plot_edge_weights( "method before plotting the nodes." ) - weights = self._positive_graph.get_edge_weights() + if self._positive_graph.is_directed(): + weights = self._positive_graph.get_directed_edge_weights() + else: + weights = self._positive_graph.get_undirected_edge_weights() returned_values = self._wrapped_plot_scatter( points=self._positive_edge_decomposition, @@ -3725,16 +3816,17 @@ def _plot_positive_and_negative_edges_distance_histogram( """ graph_transformer = GraphTransformer( method=distance_callback, - aligned_node_mapping=True + aligned_mapping=True, + include_both_undirected_edges=False ) graph_transformer.fit(node_features.astype(np.float32)) return self._plot_positive_and_negative_edges_metric_histogram( metric_name=distance_name, - edge_metrics=graph_transformer.transform(np.vstack([ - self._negative_graph.get_directed_edge_node_ids(), - self._positive_graph.get_directed_edge_node_ids(), - ])).flatten(), + edge_metrics=np.vstack([ + graph_transformer.transform(self._negative_graph), + graph_transformer.transform(self._positive_graph) + ]).flatten(), figure=figure, axes=axes, apply_tight_layout=apply_tight_layout, @@ -3746,7 +3838,6 @@ def _plot_positive_and_negative_edges_distance( node_features: np.ndarray, distance_name: str, distance_callback: str, - offset: float = 0.0, **kwargs: Dict ): """Plot distances of node features for positive and negative edges. @@ -3759,10 +3850,6 @@ def _plot_positive_and_negative_edges_distance( The title for the heatmap. distance_callback: str The callback to use to compute the distances. - offset: float = 0.0 - The offset to move the distance when it is not a true distance - such as with the cosine similarity and negative value would - not be plottable on a logarithmic scale. **kwargs: Dict Additional kwargs to forward. @@ -3777,17 +3864,18 @@ def _plot_positive_and_negative_edges_distance( """ graph_transformer = GraphTransformer( method=distance_callback, - aligned_node_mapping=True + aligned_mapping=True, + include_both_undirected_edges=False ) graph_transformer.fit(node_features.astype(np.float32)) return self._plot_positive_and_negative_edges_metric( metric_name=distance_name, - edge_metrics=offset + graph_transformer.transform(np.vstack([ - self._negative_graph.get_directed_edge_node_ids(), - self._positive_graph.get_directed_edge_node_ids(), - ])), + edge_metrics=np.vstack([ + graph_transformer.transform(self._negative_graph), + graph_transformer.transform(self._positive_graph) + ]), **kwargs, ) @@ -4005,7 +4093,6 @@ def plot_positive_and_negative_edges_cosine_similarity( node_features=node_features, distance_name="Cosine similarity", distance_callback="CosineSimilarity", - offset=1.0, figure=figure, axes=axes, scatter_kwargs=scatter_kwargs, @@ -4149,7 +4236,7 @@ def plot_edge_weight_distribution( self._graph.get_number_of_directed_edges() // 10 ) axes.hist( - self._graph.get_edge_weights(), + self._graph.get_directed_edge_weights(), bins=number_of_buckets, log=True ) @@ -4212,13 +4299,12 @@ def fit_and_plot_all( node_scatter_plot_methods_to_call = [] distribution_plot_methods_to_call = [] - if self._graph.has_constant_non_zero_node_degrees(): - node_scatter_plot_methods_to_call.append( - self.plot_node_degrees, - ) - distribution_plot_methods_to_call.append( - self.plot_node_degree_distribution, - ) + node_scatter_plot_methods_to_call.append( + self.plot_node_degrees, + ) + distribution_plot_methods_to_call.append( + self.plot_node_degree_distribution, + ) def plot_distance_wrapper(plot_distance): @functools.wraps(plot_distance) @@ -4262,7 +4348,7 @@ def wrapped_plot_distance(**kwargs): self.plot_node_ontologies ) - if not self._support.is_connected(): + if not self._support.is_connected() and not self._support.is_directed(): node_scatter_plot_methods_to_call.append( self.plot_connected_components ) @@ -4290,7 +4376,7 @@ def wrapped_plot_distance(**kwargs): int(math.ceil(number_of_total_plots / number_of_columns)), 1) ncols = min(number_of_columns, number_of_total_plots) - figure, axes = plt.subplots( + figure, axes = self._get_figure_and_axes( nrows=nrows, ncols=ncols, figsize=(5*ncols, 5*nrows), @@ -4348,16 +4434,22 @@ def wrapped_plot_distance(**kwargs): complete_caption += f" ({letter}) {caption}" if show_letters: + if self._n_components == 3: + additional_kwargs = dict(z=0.0) + else: + additional_kwargs = dict() + ax.text( - 0.0, - 1.1, - letter, + x=0.0, + y=1.1, + s=letter, size=18, color="black", weight="bold", horizontalalignment="left", verticalalignment="center", transform=ax.transAxes, + **additional_kwargs ) complete_caption += " " @@ -4378,6 +4470,8 @@ def wrapped_plot_distance(**kwargs): complete_caption += self.get_non_existing_edges_sampling_description() for axis in flat_axes[number_of_total_plots:]: + for spine in axis.spines.values(): + spine.set_visible(False) axis.axis("off") if show_name_backup: @@ -4388,12 +4482,15 @@ def wrapped_plot_distance(**kwargs): ), fontsize=20 ) - figure.tight_layout(rect=[0, 0.03, 1, 0.96]) - else: + if self._n_components != 3: + figure.tight_layout(rect=[0, 0.03, 1, 0.96]) + elif self._n_components != 3: figure.tight_layout() self._show_graph_name = show_name_backup return self._handle_notebook_display( - figure, axes, caption=complete_caption + figure, + axes, + caption=complete_caption ) diff --git a/setup.py b/setup.py index 3a7b6649..077ba6b3 100644 --- a/setup.py +++ b/setup.py @@ -22,9 +22,6 @@ def readme(): "silence_tensorflow" ] -extras = { - 'test': test_deps, -} def read(*parts): with copen(os.path.join(here, *parts), 'r') as fp: @@ -56,11 +53,11 @@ def find_version(*file_paths): description='Graph embedding, machine learning, and visualization library.', long_description=readme(), url='https://github.com/monarch-initiative/embiggen', - keywords='node2vec,word2vec,CBOW,SkipGram,GloVe', + keywords='Graph Representation Learning,LINE,TransE,Node2Vec,DeeWalk', author=", ".join(list(authors.keys())), author_email=", ".join(list(authors.values())), license='BSD3', - python_requires='>=3.6.0', + python_requires='>=3.7.0', packages=find_packages( exclude=['contrib', 'docs', 'tests*', 'notebooks*']), install_requires=[ @@ -69,17 +66,18 @@ def find_version(*file_paths): "tqdm", "matplotlib>=3.5.2", "scikit-learn", - "ddd_subplots>=1.0.19", + "userinput>=1.0.19", + "ddd_subplots>=1.0.20", "sanitize_ml_labels>=1.0.38", "keras_mixed_sequence>=1.0.28", - "extra_keras_metrics>=2.0.7", - "ensmallen>=0.8.3", + "ensmallen>=0.8.8", "validate_version_code", "cache_decorator>=2.1.8", "packaging" ], tests_require=test_deps, include_package_data=True, - zip_safe=False, - extras_require=extras, + extras_require={ + 'test': test_deps, + }, )

Bringing resnik, bug fixes, MMAPs and generalizations to develop

2022-08-02T20:05:14

0.0

[]

emmett-framework/granian

emmett-framework__granian-454

5cd82992e1085539dbff295e804021365f347d3c

diff --git a/.github/workflows/build.yml b/.github/workflows/build.yml index e2f9c11..006a055 100644 --- a/.github/workflows/build.yml +++ b/.github/workflows/build.yml @@ -4,7 +4,7 @@ on: workflow_dispatch env: MATURIN_VERSION: 1.7.4 - PY_ALL: 3.8 3.9 3.10 3.11 3.12 3.13 pypy3.8 pypy3.9 pypy3.10 + PY_ALL: 3.9 3.10 3.11 3.12 3.13 pypy3.9 pypy3.10 jobs: wheels: @@ -23,13 +23,13 @@ jobs: platform: linux target: x86_64 manylinux: auto - interpreter: pypy3.8 pypy3.9 pypy3.10 + interpreter: pypy3.9 pypy3.10 - os: macos target: x86_64 - interpreter: pypy3.8 pypy3.9 pypy3.10 + interpreter: pypy3.9 pypy3.10 - os: macos target: aarch64 - interpreter: 3.8 3.9 pypy3.8 pypy3.9 pypy3.10 + interpreter: 3.9 pypy3.9 pypy3.10 - os: ubuntu platform: linux target: aarch64 @@ -44,7 +44,7 @@ jobs: manylinux: musllinux_1_1 - os: windows target: x86_64 - interpreter: pypy3.8 pypy3.9 pypy3.10 + interpreter: pypy3.9 pypy3.10 exclude: - os: windows target: aarch64 @@ -84,7 +84,7 @@ jobs: matrix: os: [ubuntu-latest, macos-latest, macos-14, windows-latest] manylinux: [auto] - interpreter: ["3.8", "3.9", "3.10", "3.11", "3.12", "3.13"] + interpreter: ["3.9", "3.10", "3.11", "3.12", "3.13"] include: - os: ubuntu-latest platform: linux diff --git a/.github/workflows/release.yml b/.github/workflows/release.yml index f2d3044..47863c3 100644 --- a/.github/workflows/release.yml +++ b/.github/workflows/release.yml @@ -7,7 +7,7 @@ on: env: MATURIN_VERSION: 1.7.4 - PY_ALL: 3.8 3.9 3.10 3.11 3.12 3.13 pypy3.8 pypy3.9 pypy3.10 + PY_ALL: 3.9 3.10 3.11 3.12 3.13 pypy3.9 pypy3.10 jobs: sdist: @@ -46,13 +46,13 @@ jobs: platform: linux target: x86_64 manylinux: auto - interpreter: pypy3.8 pypy3.9 pypy3.10 + interpreter: pypy3.9 pypy3.10 - os: macos target: x86_64 - interpreter: pypy3.8 pypy3.9 pypy3.10 + interpreter: pypy3.9 pypy3.10 - os: macos target: aarch64 - interpreter: pypy3.8 pypy3.9 pypy3.10 + interpreter: pypy3.9 pypy3.10 - os: ubuntu platform: linux target: aarch64 @@ -67,7 +67,7 @@ jobs: manylinux: musllinux_1_1 - os: windows target: x86_64 - interpreter: pypy3.8 pypy3.9 pypy3.10 + interpreter: pypy3.9 pypy3.10 exclude: - os: windows target: aarch64 @@ -102,7 +102,7 @@ jobs: matrix: os: [ubuntu-latest, macos-13, macos-14, windows-latest] manylinux: [auto] - interpreter: ["3.8", "3.9", "3.10", "3.11", "3.12", "3.13"] + interpreter: ["3.9", "3.10", "3.11", "3.12", "3.13"] include: - os: ubuntu-latest platform: linux diff --git a/Cargo.lock b/Cargo.lock index ce9d6b9..2c1ee71 100644 --- a/Cargo.lock +++ b/Cargo.lock @@ -970,9 +970,9 @@ dependencies = [ [[package]] name = "pyo3" -version = "0.22.6" +version = "0.23.2" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "f402062616ab18202ae8319da13fa4279883a2b8a9d9f83f20dbade813ce1884" +checksum = "f54b3d09cbdd1f8c20650b28e7b09e338881482f4aa908a5f61a00c98fba2690" dependencies = [ "anyhow", "cfg-if", @@ -989,9 +989,9 @@ dependencies = [ [[package]] name = "pyo3-build-config" -version = "0.22.6" +version = "0.23.2" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "b14b5775b5ff446dd1056212d778012cbe8a0fbffd368029fd9e25b514479c38" +checksum = "3015cf985888fe66cfb63ce0e321c603706cd541b7aec7ddd35c281390af45d8" dependencies = [ "once_cell", "python3-dll-a", @@ -1000,9 +1000,9 @@ dependencies = [ [[package]] name = "pyo3-ffi" -version = "0.22.6" +version = "0.23.2" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "9ab5bcf04a2cdcbb50c7d6105de943f543f9ed92af55818fd17b660390fc8636" +checksum = "6fca7cd8fd809b5ac4eefb89c1f98f7a7651d3739dfb341ca6980090f554c270" dependencies = [ "libc", "pyo3-build-config", @@ -1010,9 +1010,9 @@ dependencies = [ [[package]] name = "pyo3-log" -version = "0.11.0" +version = "0.12.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "3ac84e6eec1159bc2a575c9ae6723baa6ee9d45873e9bebad1e3ad7e8d28a443" +checksum = "3eb421dc86d38d08e04b927b02424db480be71b777fa3a56f32e2f2a3a1a3b08" dependencies = [ "arc-swap", "log", @@ -1021,9 +1021,9 @@ dependencies = [ [[package]] name = "pyo3-macros" -version = "0.22.6" +version = "0.23.2" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "0fd24d897903a9e6d80b968368a34e1525aeb719d568dba8b3d4bfa5dc67d453" +checksum = "34e657fa5379a79151b6ff5328d9216a84f55dc93b17b08e7c3609a969b73aa0" dependencies = [ "proc-macro2", "pyo3-macros-backend", @@ -1033,9 +1033,9 @@ dependencies = [ [[package]] name = "pyo3-macros-backend" -version = "0.22.6" +version = "0.23.2" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "36c011a03ba1e50152b4b394b479826cad97e7a21eb52df179cd91ac411cbfbe" +checksum = "295548d5ffd95fd1981d2d3cf4458831b21d60af046b729b6fd143b0ba7aee2f" dependencies = [ "heck", "proc-macro2", diff --git a/Cargo.toml b/Cargo.toml index 349c28a..09f7903 100644 --- a/Cargo.toml +++ b/Cargo.toml @@ -43,8 +43,8 @@ pem = "=3.0" percent-encoding = "=2.3" pin-project = "1.1" pkcs8 = { version = "=0.10", features = ["encryption", "pkcs5"] } -pyo3 = { version = "=0.22", features = ["anyhow", "extension-module", "generate-import-lib"] } -pyo3-log = "=0.11" +pyo3 = { version = "=0.23", features = ["anyhow", "extension-module", "generate-import-lib"] } +pyo3-log = "=0.12" rustls-pemfile = "2.2" socket2 = { version = "0.5", features = ["all"] } tls-listener = { version = "=0.10", features = ["rustls"] } diff --git a/README.md b/README.md index 087f68d..1ad6560 100644 --- a/README.md +++ b/README.md @@ -277,7 +277,7 @@ You might test the effect such optimizations cause over your application and dec Granian is currently under active development. -Granian is compatible with Python 3.8 and above versions. +Granian is compatible with Python 3.9 and above versions. ## License diff --git a/pyproject.toml b/pyproject.toml index fb08077..55aaadc 100644 --- a/pyproject.toml +++ b/pyproject.toml @@ -11,7 +11,6 @@ classifiers = [ 'Operating System :: Microsoft :: Windows', 'Operating System :: POSIX :: Linux', 'Programming Language :: Python :: 3', - 'Programming Language :: Python :: 3.8', 'Programming Language :: Python :: 3.9', 'Programming Language :: Python :: 3.10', 'Programming Language :: Python :: 3.11', @@ -32,7 +31,7 @@ dynamic = [ 'version', ] -requires-python = '>=3.8' +requires-python = '>=3.9' dependencies = [ 'click>=8.0.0', 'uvloop>=0.18.0; sys_platform != "win32" and platform_python_implementation == "CPython"', diff --git a/src/asgi/callbacks.rs b/src/asgi/callbacks.rs index d2cc696..e437017 100644 --- a/src/asgi/callbacks.rs +++ b/src/asgi/callbacks.rs @@ -43,14 +43,14 @@ pub(crate) struct CallbackWatcherHTTP { #[pyo3(get)] proto: Py<HTTPProtocol>, #[pyo3(get)] - scope: PyObject, + scope: Py<PyDict>, } impl CallbackWatcherHTTP { pub fn new(py: Python, proto: HTTPProtocol, scope: Bound<PyDict>) -> Self { Self { proto: Py::new(py, proto).unwrap(), - scope: scope.into_py(py), + scope: scope.unbind(), } } } @@ -62,7 +62,7 @@ impl CallbackWatcherHTTP { } fn err(&self, err: Bound<PyAny>) { - callback_impl_done_err!(self, &PyErr::from_value_bound(err)); + callback_impl_done_err!(self, &PyErr::from_value(err)); } } @@ -71,14 +71,14 @@ pub(crate) struct CallbackWatcherWebsocket { #[pyo3(get)] proto: Py<WebsocketProtocol>, #[pyo3(get)] - scope: PyObject, + scope: Py<PyDict>, } impl CallbackWatcherWebsocket { pub fn new(py: Python, proto: WebsocketProtocol, scope: Bound<PyDict>) -> Self { Self { proto: Py::new(py, proto).unwrap(), - scope: scope.into_py(py), + scope: scope.unbind(), } } } @@ -90,7 +90,7 @@ impl CallbackWatcherWebsocket { } fn err(&self, err: Bound<PyAny>) { - callback_impl_done_err!(self, &PyErr::from_value_bound(err)); + callback_impl_done_err!(self, &PyErr::from_value(err)); } } diff --git a/src/asgi/io.rs b/src/asgi/io.rs index f0ccaf1..a4a112f 100644 --- a/src/asgi/io.rs +++ b/src/asgi/io.rs @@ -87,6 +87,8 @@ impl ASGIHTTPProtocol { close: bool, ) -> PyResult<Bound<'p, PyAny>> { let flow_hld = self.flow_tx_waiter.clone(); + let pynone = py.None(); + future_into_py_futlike(self.rt.clone(), py, async move { match tx.send(Ok(body.into())).await { Ok(()) => { @@ -99,7 +101,7 @@ impl ASGIHTTPProtocol { flow_hld.notify_one(); } } - Ok(()) + Ok(pynone) }) } @@ -116,9 +118,9 @@ impl ASGIHTTPProtocol { return future_into_py_futlike(self.rt.clone(), py, async move { let () = flow_hld.notified().await; Python::with_gil(|py| { - let dict = PyDict::new_bound(py); + let dict = PyDict::new(py); dict.set_item(pyo3::intern!(py, "type"), pyo3::intern!(py, "http.disconnect"))?; - Ok(dict.to_object(py)) + Ok(dict.into_any().unbind()) }) }); } @@ -144,18 +146,18 @@ impl ASGIHTTPProtocol { match chunk { Ok(data) => Python::with_gil(|py| { - let dict = PyDict::new_bound(py); + let dict = PyDict::new(py); dict.set_item(pyo3::intern!(py, "type"), pyo3::intern!(py, "http.request"))?; dict.set_item(pyo3::intern!(py, "body"), BytesToPy(data))?; dict.set_item(pyo3::intern!(py, "more_body"), more_body)?; - Ok(dict.to_object(py)) + Ok(dict.into_any().unbind()) }), _ => { flow_hld.notify_one(); Python::with_gil(|py| { - let dict = PyDict::new_bound(py); + let dict = PyDict::new(py); dict.set_item(pyo3::intern!(py, "type"), pyo3::intern!(py, "http.disconnect"))?; - Ok(dict.to_object(py)) + Ok(dict.into_any().unbind()) }) } } @@ -323,6 +325,7 @@ impl ASGIWebsocketProtocol { let accepted = self.accepted.clone(); let rx = self.ws_rx.clone(); let tx = self.ws_tx.clone(); + let pynone = py.None(); future_into_py_iter(self.rt.clone(), py, async move { if let Some(mut upgrade) = upgrade { @@ -339,7 +342,7 @@ impl ASGIWebsocketProtocol { *wtx = Some(tx); *wrx = Some(rx); accepted.store(true, atomic::Ordering::Relaxed); - return Ok(()); + return Ok(pynone); } } } @@ -352,15 +355,16 @@ impl ASGIWebsocketProtocol { fn send_message<'p>(&self, py: Python<'p>, data: Message) -> PyResult<Bound<'p, PyAny>> { let transport = self.ws_tx.clone(); let closed = self.closed.clone(); + let pynone = py.None(); future_into_py_futlike(self.rt.clone(), py, async move { if let Some(ws) = &mut *(transport.lock().await) { match ws.send(data).await { - Ok(()) => return Ok(()), + Ok(()) => return Ok(pynone), _ => { if closed.load(atomic::Ordering::Relaxed) { log::info!("Attempted to write to a closed websocket"); - return Ok(()); + return Ok(pynone); } } }; @@ -374,6 +378,7 @@ impl ASGIWebsocketProtocol { let closed = self.closed.clone(); let ws_rx = self.ws_rx.clone(); let ws_tx = self.ws_tx.clone(); + let pynone = py.None(); future_into_py_iter(self.rt.clone(), py, async move { if let Some(tx) = ws_tx.lock().await.take() { @@ -382,7 +387,7 @@ impl ASGIWebsocketProtocol { .close() .await; } - Ok(()) + Ok(pynone) }) } @@ -416,9 +421,9 @@ impl ASGIWebsocketProtocol { let accepted = accepted.load(atomic::Ordering::Relaxed); if !accepted { return Python::with_gil(|py| { - let dict = PyDict::new_bound(py); + let dict = PyDict::new(py); dict.set_item(pyo3::intern!(py, "type"), pyo3::intern!(py, "websocket.connect"))?; - Ok(dict.to_object(py)) + Ok(dict.into_any().unbind()) }); } @@ -444,7 +449,7 @@ impl ASGIWebsocketProtocol { Ok(ASGIMessageType::WSAccept(subproto)) => self.accept(py, subproto), Ok(ASGIMessageType::WSClose) => self.close(py), Ok(ASGIMessageType::WSMessage(message)) => self.send_message(py, message), - _ => future_into_py_iter::<_, _, PyErr>(self.rt.clone(), py, async { error_message!() }), + _ => future_into_py_iter::<_, _>(self.rt.clone(), py, async { error_message!() }), } } } @@ -549,26 +554,26 @@ fn ws_message_into_rs(py: Python, message: &Bound<PyDict>) -> PyResult<Message> fn ws_message_into_py(message: Message) -> PyResult<PyObject> { match message { Message::Binary(message) => Python::with_gil(|py| { - let dict = PyDict::new_bound(py); + let dict = PyDict::new(py); dict.set_item(pyo3::intern!(py, "type"), pyo3::intern!(py, "websocket.receive"))?; - dict.set_item(pyo3::intern!(py, "bytes"), PyBytes::new_bound(py, &message[..]))?; - Ok(dict.to_object(py)) + dict.set_item(pyo3::intern!(py, "bytes"), PyBytes::new(py, &message[..]))?; + Ok(dict.into_any().unbind()) }), Message::Text(message) => Python::with_gil(|py| { - let dict = PyDict::new_bound(py); + let dict = PyDict::new(py); dict.set_item(pyo3::intern!(py, "type"), pyo3::intern!(py, "websocket.receive"))?; dict.set_item(pyo3::intern!(py, "text"), message)?; - Ok(dict.to_object(py)) + Ok(dict.into_any().unbind()) }), Message::Close(frame) => Python::with_gil(|py| { let close_code: u16 = match frame { Some(frame) => frame.code.into(), _ => 1005, }; - let dict = PyDict::new_bound(py); + let dict = PyDict::new(py); dict.set_item(pyo3::intern!(py, "type"), pyo3::intern!(py, "websocket.disconnect"))?; dict.set_item(pyo3::intern!(py, "code"), close_code)?; - Ok(dict.to_object(py)) + Ok(dict.into_any().unbind()) }), v => { log::warn!("Unsupported websocket message received {:?}", v); diff --git a/src/asgi/utils.rs b/src/asgi/utils.rs index 0c8f2e1..3561892 100644 --- a/src/asgi/utils.rs +++ b/src/asgi/utils.rs @@ -42,13 +42,13 @@ pub(super) fn build_scope<'p>( path: &'p str, query_string: &'p str, ) -> PyResult<Bound<'p, PyDict>> { - let scope = PyDict::new_bound(py); + let scope = PyDict::new(py); scope.set_item( pyo3::intern!(py, "asgi"), ASGI_VERSION .get_or_try_init(py, || { - let rv = PyDict::new_bound(py); + let rv = PyDict::new(py); rv.set_item("version", "3.0")?; rv.set_item("spec_version", "2.3")?; Ok::<PyObject, PyErr>(rv.into()) @@ -59,8 +59,8 @@ pub(super) fn build_scope<'p>( pyo3::intern!(py, "extensions"), ASGI_EXTENSIONS .get_or_try_init(py, || { - let rv = PyDict::new_bound(py); - rv.set_item("http.response.pathsend", PyDict::new_bound(py))?; + let rv = PyDict::new(py); + rv.set_item("http.response.pathsend", PyDict::new(py))?; Ok::<PyObject, PyErr>(rv.into()) })? .bind(py), @@ -71,25 +71,22 @@ pub(super) fn build_scope<'p>( scope.set_item(pyo3::intern!(py, "client"), client)?; scope.set_item(pyo3::intern!(py, "scheme"), scheme)?; scope.set_item(pyo3::intern!(py, "path"), path)?; - scope.set_item(pyo3::intern!(py, "raw_path"), PyBytes::new_bound(py, path.as_bytes()))?; + scope.set_item(pyo3::intern!(py, "raw_path"), PyBytes::new(py, path.as_bytes()))?; scope.set_item( pyo3::intern!(py, "query_string"), - PyBytes::new_bound(py, query_string.as_bytes()), + PyBytes::new(py, query_string.as_bytes()), )?; - let headers = PyList::empty_bound(py); + let headers = PyList::empty(py); for (key, value) in &req.headers { headers.append(( - PyBytes::new_bound(py, key.as_str().as_bytes()), - PyBytes::new_bound(py, value.as_bytes()), + PyBytes::new(py, key.as_str().as_bytes()), + PyBytes::new(py, value.as_bytes()), ))?; } if !req.headers.contains_key(header::HOST) { let host = req.uri.authority().map_or("", Authority::as_str); - headers.insert( - 0, - (PyBytes::new_bound(py, b"host"), PyBytes::new_bound(py, host.as_bytes())), - )?; + headers.insert(0, (PyBytes::new(py, b"host"), PyBytes::new(py, host.as_bytes())))?; } scope.set_item(pyo3::intern!(py, "headers"), headers)?; @@ -136,14 +133,14 @@ pub(super) fn build_scope_ws<'p>( )?; scope.set_item( pyo3::intern!(py, "subprotocols"), - PyList::new_bound( + PyList::new( py, req.headers .get_all("Sec-WebSocket-Protocol") .iter() - .map(|v| PyString::new_bound(py, v.to_str().unwrap())) + .map(|v| PyString::new(py, v.to_str().unwrap())) .collect::<Vec<Bound<PyString>>>(), - ), + )?, )?; Ok(scope) } diff --git a/src/asyncio.rs b/src/asyncio.rs index f6b1b38..3376c8f 100644 --- a/src/asyncio.rs +++ b/src/asyncio.rs @@ -6,7 +6,7 @@ static CONTEXT: GILOnceCell<PyObject> = GILOnceCell::new(); fn contextvars(py: Python) -> PyResult<&Bound<PyAny>> { Ok(CONTEXTVARS - .get_or_try_init(py, || py.import_bound("contextvars").map(Into::into))? + .get_or_try_init(py, || py.import("contextvars").map(Into::into))? .bind(py)) } diff --git a/src/callbacks.rs b/src/callbacks.rs index 2944fbf..d83583b 100644 --- a/src/callbacks.rs +++ b/src/callbacks.rs @@ -78,9 +78,9 @@ impl PyIterAwaitable { } } - pub(crate) fn set_result(&self, result: PyResult<impl IntoPy<PyObject>>) { + pub(crate) fn set_result(&self, result: PyResult<PyObject>) { let mut res = self.result.write().unwrap(); - *res = Some(Python::with_gil(|py| result.map(|v| v.into_py(py)))); + *res = Some(result); } } @@ -138,17 +138,17 @@ impl PyFutureAwaitable { Ok((Py::new(py, self)?, cancel_tx)) } - pub(crate) fn set_result(&self, result: PyResult<impl IntoPy<PyObject>>, aw: Py<PyFutureAwaitable>) { + pub(crate) fn set_result(&self, result: PyResult<PyObject>, aw: Py<PyFutureAwaitable>) { Python::with_gil(|py| { let mut state = self.state.write().unwrap(); if !matches!(&mut *state, PyFutureAwaitableState::Pending) { return; } - *state = PyFutureAwaitableState::Completed(result.map(|v| v.into_py(py))); + *state = PyFutureAwaitableState::Completed(result); let ack = self.ack.read().unwrap(); if let Some((cb, ctx)) = &*ack { - let _ = self.event_loop.clone_ref(py).call_method_bound( + let _ = self.event_loop.clone_ref(py).call_method( py, pyo3::intern!(py, "call_soon_threadsafe"), (cb, aw), @@ -198,7 +198,7 @@ impl PyFutureAwaitable { #[pyo3(signature = (cb, context=None))] fn add_done_callback(pyself: PyRef<'_, Self>, cb: PyObject, context: Option<PyObject>) -> PyResult<()> { let py = pyself.py(); - let kwctx = pyo3::types::PyDict::new_bound(py); + let kwctx = pyo3::types::PyDict::new(py); kwctx.set_item(pyo3::intern!(py, "context"), context)?; let state = pyself.state.read().unwrap(); @@ -211,7 +211,7 @@ impl PyFutureAwaitable { _ => { drop(state); let event_loop = pyself.event_loop.clone_ref(py); - event_loop.call_method_bound(py, pyo3::intern!(py, "call_soon"), (cb, pyself), Some(&kwctx))?; + event_loop.call_method(py, pyo3::intern!(py, "call_soon"), (cb, pyself), Some(&kwctx))?; Ok(()) } } @@ -244,7 +244,7 @@ impl PyFutureAwaitable { drop(ack); drop(state); - let _ = event_loop.call_method_bound(py, pyo3::intern!(py, "call_soon"), (cb, pyself), Some(ctx.bind(py))); + let _ = event_loop.call_method(py, pyo3::intern!(py, "call_soon"), (cb, pyself), Some(ctx.bind(py))); } true diff --git a/src/conversion.rs b/src/conversion.rs index 56c1884..875dbdc 100644 --- a/src/conversion.rs +++ b/src/conversion.rs @@ -1,20 +1,16 @@ -use pyo3::prelude::*; +use pyo3::{prelude::*, IntoPyObjectExt}; use crate::workers::{HTTP1Config, HTTP2Config}; pub(crate) struct BytesToPy(pub hyper::body::Bytes); -impl IntoPy<PyObject> for BytesToPy { - #[inline] - fn into_py(self, py: Python) -> PyObject { - self.0.as_ref().into_py(py) - } -} +impl<'p> IntoPyObject<'p> for BytesToPy { + type Target = PyAny; + type Output = Bound<'p, Self::Target>; + type Error = PyErr; -impl ToPyObject for BytesToPy { - #[inline] - fn to_object(&self, py: Python<'_>) -> PyObject { - self.0.as_ref().into_py(py) + fn into_pyobject(self, py: Python<'p>) -> Result<Self::Output, Self::Error> { + self.0.as_ref().into_bound_py_any(py) } } diff --git a/src/rsgi/callbacks.rs b/src/rsgi/callbacks.rs index b67081b..94a97c7 100644 --- a/src/rsgi/callbacks.rs +++ b/src/rsgi/callbacks.rs @@ -38,14 +38,14 @@ pub(crate) struct CallbackWatcherHTTP { #[pyo3(get)] proto: Py<HTTPProtocol>, #[pyo3(get)] - scope: PyObject, + scope: Py<HTTPScope>, } impl CallbackWatcherHTTP { pub fn new(py: Python, proto: HTTPProtocol, scope: HTTPScope) -> Self { Self { proto: Py::new(py, proto).unwrap(), - scope: scope.into_py(py), + scope: Py::new(py, scope).unwrap(), } } } @@ -57,7 +57,7 @@ impl CallbackWatcherHTTP { } fn err(&self, err: Bound<PyAny>) { - callback_impl_done_err!(self, &PyErr::from_value_bound(err)); + callback_impl_done_err!(self, &PyErr::from_value(err)); } } @@ -66,14 +66,14 @@ pub(crate) struct CallbackWatcherWebsocket { #[pyo3(get)] proto: Py<WebsocketProtocol>, #[pyo3(get)] - scope: PyObject, + scope: Py<WebsocketScope>, } impl CallbackWatcherWebsocket { pub fn new(py: Python, proto: WebsocketProtocol, scope: WebsocketScope) -> Self { Self { proto: Py::new(py, proto).unwrap(), - scope: scope.into_py(py), + scope: Py::new(py, scope).unwrap(), } } } @@ -85,7 +85,7 @@ impl CallbackWatcherWebsocket { } fn err(&self, err: Bound<PyAny>) { - callback_impl_done_err!(self, &PyErr::from_value_bound(err)); + callback_impl_done_err!(self, &PyErr::from_value(err)); } } diff --git a/src/rsgi/io.rs b/src/rsgi/io.rs index cd813ea..cbf0f3a 100644 --- a/src/rsgi/io.rs +++ b/src/rsgi/io.rs @@ -1,9 +1,12 @@ use futures::{sink::SinkExt, StreamExt, TryStreamExt}; use http_body_util::BodyExt; use hyper::body; -use pyo3::prelude::*; -use pyo3::pybacked::PyBackedStr; -use pyo3::types::{PyBytes, PyString}; +use pyo3::{ + prelude::*, + pybacked::PyBackedStr, + types::{PyBytes, PyString}, + IntoPyObjectExt, +}; use std::{ borrow::Cow, sync::{atomic, Arc, Mutex, RwLock}, @@ -40,9 +43,11 @@ impl RSGIHTTPStreamTransport { fn send_bytes<'p>(&self, py: Python<'p>, data: Cow<[u8]>) -> PyResult<Bound<'p, PyAny>> { let transport = self.tx.clone(); let bdata: Box<[u8]> = data.into(); + let pynone = py.None(); + future_into_py_futlike(self.rt.clone(), py, async move { match transport.send(Ok(body::Bytes::from(bdata))).await { - Ok(()) => Ok(()), + Ok(()) => Ok(pynone), _ => error_stream!(), } }) @@ -50,9 +55,11 @@ impl RSGIHTTPStreamTransport { fn send_str<'p>(&self, py: Python<'p>, data: String) -> PyResult<Bound<'p, PyAny>> { let transport = self.tx.clone(); + let pynone = py.None(); + future_into_py_futlike(self.rt.clone(), py, async move { match transport.send(Ok(body::Bytes::from(data))).await { - Ok(()) => Ok(()), + Ok(()) => Ok(pynone), _ => error_stream!(), } }) @@ -88,7 +95,10 @@ impl RSGIHTTPProtocol { if let Some(body) = self.body.lock().unwrap().take() { return future_into_py_iter(self.rt.clone(), py, async move { match body.collect().await { - Ok(data) => Ok(BytesToPy(data.to_bytes())), + Ok(data) => { + let bytes = BytesToPy(data.to_bytes()); + Ok(Python::with_gil(|py| bytes.into_py_any(py))?) + } _ => error_stream!(), } }); @@ -125,7 +135,7 @@ impl RSGIHTTPProtocol { BytesToPy(body::Bytes::new()) } }; - Ok(bytes) + Python::with_gil(|py| bytes.into_py_any(py)) }) } @@ -240,10 +250,12 @@ impl RSGIWebsocketTransport { fn send_bytes<'p>(&self, py: Python<'p>, data: Cow<[u8]>) -> PyResult<Bound<'p, PyAny>> { let transport = self.tx.clone(); let bdata: Box<[u8]> = data.into(); + let pynone = py.None(); + future_into_py_iter(self.rt.clone(), py, async move { if let Ok(mut stream) = transport.try_lock() { return match stream.send(bdata[..].into()).await { - Ok(()) => Ok(()), + Ok(()) => Ok(pynone), _ => error_stream!(), }; } @@ -253,10 +265,12 @@ impl RSGIWebsocketTransport { fn send_str<'p>(&self, py: Python<'p>, data: String) -> PyResult<Bound<'p, PyAny>> { let transport = self.tx.clone(); + let pynone = py.None(); + future_into_py_iter(self.rt.clone(), py, async move { if let Ok(mut stream) = transport.try_lock() { return match stream.send(Message::Text(data)).await { - Ok(()) => Ok(()), + Ok(()) => Ok(pynone), _ => error_stream!(), }; } @@ -379,7 +393,7 @@ impl RSGIWebsocketProtocol { Ok(Python::with_gil(|py| { let pytransport = Py::new(py, RSGIWebsocketTransport::new(rth, stream)).unwrap(); *trx = Some(pytransport.clone_ref(py)); - pytransport + pytransport.into_any() })) } _ => error_proto!(), @@ -394,12 +408,14 @@ impl RSGIWebsocketProtocol { fn message_into_py(message: Message) -> PyResult<PyObject> { match message { Message::Binary(message) => Ok(Python::with_gil(|py| { - WebsocketInboundBytesMessage::new(PyBytes::new_bound(py, &message).unbind()).into_py(py) - })), + WebsocketInboundBytesMessage::new(PyBytes::new(py, &message).unbind()).into_py_any(py) + })?), Message::Text(message) => Ok(Python::with_gil(|py| { - WebsocketInboundTextMessage::new(PyString::new_bound(py, &message).unbind()).into_py(py) - })), - Message::Close(_) => Ok(Python::with_gil(|py| WebsocketInboundCloseMessage::new().into_py(py))), + WebsocketInboundTextMessage::new(PyString::new(py, &message).unbind()).into_py_any(py) + })?), + Message::Close(_) => Ok(Python::with_gil(|py| { + WebsocketInboundCloseMessage::new().into_py_any(py) + })?), v => { log::warn!("Unsupported websocket message received {:?}", v); error_proto!() diff --git a/src/rsgi/mod.rs b/src/rsgi/mod.rs index e27bf4d..6c18eed 100644 --- a/src/rsgi/mod.rs +++ b/src/rsgi/mod.rs @@ -8,8 +8,8 @@ pub(crate) mod serve; mod types; pub(crate) fn init_pymodule(py: Python, module: &Bound<PyModule>) -> PyResult<()> { - module.add("RSGIProtocolError", py.get_type_bound::<errors::RSGIProtocolError>())?; - module.add("RSGIProtocolClosed", py.get_type_bound::<errors::RSGIProtocolClosed>())?; + module.add("RSGIProtocolError", py.get_type::<errors::RSGIProtocolError>())?; + module.add("RSGIProtocolClosed", py.get_type::<errors::RSGIProtocolClosed>())?; module.add_class::<io::RSGIHTTPProtocol>()?; module.add_class::<io::RSGIHTTPStreamTransport>()?; module.add_class::<io::RSGIWebsocketProtocol>()?; diff --git a/src/rsgi/types.rs b/src/rsgi/types.rs index 49d86bd..0f38863 100644 --- a/src/rsgi/types.rs +++ b/src/rsgi/types.rs @@ -68,7 +68,7 @@ impl RSGIHeaders { } fn __iter__<'p>(&self, py: Python<'p>) -> PyResult<Bound<'p, PyIterator>> { - PyIterator::from_bound_object(&PyList::new_bound(py, self.keys())) + PyIterator::from_object(PyList::new(py, self.keys())?.as_any()) } fn __len__(&self) -> usize { @@ -79,7 +79,7 @@ impl RSGIHeaders { fn get(&self, py: Python, key: &str, default: Option<PyObject>) -> Option<PyObject> { match self.inner.get(key) { Some(val) => match val.to_str() { - Ok(string) => Some(PyString::new_bound(py, string).into()), + Ok(string) => Some(PyString::new(py, string).into()), _ => default, }, _ => default, @@ -87,13 +87,13 @@ impl RSGIHeaders { } #[pyo3(signature = (key))] - fn get_all<'p>(&self, py: Python<'p>, key: &'p str) -> Bound<'p, PyList> { - PyList::new_bound( + fn get_all<'p>(&self, py: Python<'p>, key: &'p str) -> PyResult<Bound<'p, PyList>> { + PyList::new( py, self.inner .get_all(key) .iter() - .map(|v| PyString::new_bound(py, v.to_str().unwrap())) + .map(|v| PyString::new(py, v.to_str().unwrap())) .collect::<Vec<Bound<PyString>>>(), ) } diff --git a/src/runtime.rs b/src/runtime.rs index df785b9..aff85c1 100644 --- a/src/runtime.rs +++ b/src/runtime.rs @@ -1,4 +1,6 @@ use pyo3::prelude::*; +#[cfg(windows)] +use pyo3::IntoPyObjectExt; use std::{ future::Future, sync::{Arc, Mutex}, @@ -141,11 +143,10 @@ pub(crate) fn init_runtime_st(blocking_threads: usize, py_loop: Arc<PyObject>) - // but for "quick" operations it's something like 12% faster. #[allow(unused_must_use)] #[cfg(not(target_os = "linux"))] -pub(crate) fn future_into_py_iter<R, F, T>(rt: R, py: Python, fut: F) -> PyResult<Bound<PyAny>> +pub(crate) fn future_into_py_iter<R, F>(rt: R, py: Python, fut: F) -> PyResult<Bound<PyAny>> where R: Runtime + ContextExt + Clone, - F: Future<Output = PyResult<T>> + Send + 'static, - T: IntoPy<PyObject> + Send + 'static, + F: Future<Output = PyResult<PyObject>> + Send + 'static, { let aw = Py::new(py, PyIterAwaitable::new())?; let py_fut = aw.clone_ref(py); @@ -168,11 +169,10 @@ where // MacOS works best with original impl, Windows still needs further analysis. #[cfg(target_os = "linux")] #[inline(always)] -pub(crate) fn future_into_py_iter<R, F, T>(rt: R, py: Python, fut: F) -> PyResult<Bound<PyAny>> +pub(crate) fn future_into_py_iter<R, F>(rt: R, py: Python, fut: F) -> PyResult<Bound<PyAny>> where R: Runtime + ContextExt + Clone, - F: Future<Output = PyResult<T>> + Send + 'static, - T: IntoPy<PyObject> + Send + 'static, + F: Future<Output = PyResult<PyObject>> + Send + 'static, { future_into_py_futlike(rt, py, fut) } @@ -184,11 +184,10 @@ where // and for "long" operations it's something like 6% faster than `future_into_py_iter`. #[allow(unused_must_use)] #[cfg(unix)] -pub(crate) fn future_into_py_futlike<R, F, T>(rt: R, py: Python, fut: F) -> PyResult<Bound<PyAny>> +pub(crate) fn future_into_py_futlike<R, F>(rt: R, py: Python, fut: F) -> PyResult<Bound<PyAny>> where R: Runtime + ContextExt + Clone, - F: Future<Output = PyResult<T>> + Send + 'static, - T: IntoPy<PyObject> + Send + 'static, + F: Future<Output = PyResult<PyObject>> + Send + 'static, { let event_loop = rt.py_event_loop(py); let (aw, cancel_tx) = PyFutureAwaitable::new(event_loop).to_spawn(py)?; @@ -220,11 +219,10 @@ where #[allow(unused_must_use)] #[cfg(windows)] -pub(crate) fn future_into_py_futlike<R, F, T>(rt: R, py: Python, fut: F) -> PyResult<Bound<PyAny>> +pub(crate) fn future_into_py_futlike<R, F>(rt: R, py: Python, fut: F) -> PyResult<Bound<PyAny>> where R: Runtime + ContextExt + Clone, - F: Future<Output = PyResult<T>> + Send + 'static, - T: IntoPy<PyObject> + Send + 'static, + F: Future<Output = PyResult<PyObject>> + Send + 'static, { let event_loop = rt.py_event_loop(py); let event_loop_ref = event_loop.clone_ref(py); @@ -247,8 +245,8 @@ where let _ = rb.run(move || { Python::with_gil(|py| { let (cb, value) = match result { - Ok(val) => (fut_ref.getattr(py, pyo3::intern!(py, "set_result")).unwrap(), val.into_py(py)), - Err(err) => (fut_ref.getattr(py, pyo3::intern!(py, "set_exception")).unwrap(), err.into_py(py)) + Ok(val) => (fut_ref.getattr(py, pyo3::intern!(py, "set_result")).unwrap(), val), + Err(err) => (fut_ref.getattr(py, pyo3::intern!(py, "set_exception")).unwrap(), err.into_py_any(py).unwrap()) }; let _ = event_loop_ref.call_method1(py, pyo3::intern!(py, "call_soon_threadsafe"), (PyFutureResultSetter, cb, value)); drop(fut_ref); @@ -258,7 +256,7 @@ where }, () = cancel_tx.notified() => { let _ = rb.run(move || { - Python::with_gil(|py| { + Python::with_gil(|_| { drop(fut_ref); drop(event_loop_ref); }); @@ -273,35 +271,33 @@ where #[allow(clippy::unnecessary_wraps)] #[inline(always)] pub(crate) fn empty_future_into_py(py: Python) -> PyResult<Bound<PyAny>> { - Ok(PyEmptyAwaitable.into_py(py).into_bound(py)) + Ok(PyEmptyAwaitable.into_pyobject(py)?.into_any()) } #[allow(unused_must_use)] -pub(crate) fn run_until_complete<R, F, T>(rt: R, event_loop: Bound<PyAny>, fut: F) -> PyResult<T> +pub(crate) fn run_until_complete<R, F>(rt: R, event_loop: Bound<PyAny>, fut: F) -> PyResult<()> where R: Runtime + ContextExt + Clone, - F: Future<Output = PyResult<T>> + Send + 'static, - T: Send + Sync + 'static, + F: Future<Output = PyResult<()>> + Send + 'static, { - let py = event_loop.py(); let result_tx = Arc::new(Mutex::new(None)); let result_rx = Arc::clone(&result_tx); let py_fut = event_loop.call_method0("create_future")?; - let loop_tx = event_loop.clone().into_py(py); - let future_tx = py_fut.clone().into_py(py); + let loop_tx = event_loop.clone().unbind(); + let future_tx = py_fut.clone().unbind(); rt.spawn(async move { - let val = fut.await; + let _ = fut.await; if let Ok(mut result) = result_tx.lock() { - *result = Some(val.unwrap()); + *result = Some(()); } // NOTE: we don't care if we block the runtime. // `run_until_complete` is used only for the workers main loop. Python::with_gil(move |py| { let res_method = future_tx.getattr(py, "set_result").unwrap(); - let _ = loop_tx.call_method_bound(py, "call_soon_threadsafe", (res_method, py.None()), None); + let _ = loop_tx.call_method(py, "call_soon_threadsafe", (res_method, py.None()), None); drop(future_tx); drop(loop_tx); }); @@ -309,8 +305,8 @@ where event_loop.call_method1("run_until_complete", (py_fut,))?; - let result = result_rx.lock().unwrap().take().unwrap(); - Ok(result) + result_rx.lock().unwrap().take().unwrap(); + Ok(()) } pub(crate) fn block_on_local<F>(rt: &RuntimeWrapper, local: LocalSet, fut: F) diff --git a/src/tcp.rs b/src/tcp.rs index f604966..dd62f58 100644 --- a/src/tcp.rs +++ b/src/tcp.rs @@ -1,5 +1,4 @@ -use pyo3::prelude::*; -use pyo3::types::PyType; +use pyo3::{prelude::*, types::PyType, IntoPyObjectExt}; use std::net::{IpAddr, SocketAddr, TcpListener}; #[cfg(unix)] @@ -53,23 +52,23 @@ impl ListenerHolder { #[cfg(unix)] pub fn __getstate__(&self, py: Python) -> PyObject { let fd = self.socket.as_raw_fd(); - (fd.into_py(py),).to_object(py) + (fd,).into_py_any(py).unwrap() } #[cfg(windows)] pub fn __getstate__(&self, py: Python) -> PyObject { let fd = self.socket.as_raw_socket(); - (fd.into_py(py),).to_object(py) + (fd,).into_py_any(py).unwrap() } #[cfg(unix)] pub fn get_fd(&self, py: Python) -> PyObject { - self.socket.as_raw_fd().into_py(py).to_object(py) + self.socket.as_raw_fd().into_py_any(py).unwrap() } #[cfg(windows)] pub fn get_fd(&self, py: Python) -> PyObject { - self.socket.as_raw_socket().into_py(py).to_object(py) + self.socket.as_raw_socket().into_py_any(py).unwrap() } } diff --git a/src/utils.rs b/src/utils.rs index 711c95e..4ad10e8 100644 --- a/src/utils.rs +++ b/src/utils.rs @@ -43,7 +43,7 @@ fn trim_end(data: &[u8]) -> &[u8] { #[inline] pub(crate) fn log_application_callable_exception(err: &pyo3::PyErr) { let tb = pyo3::Python::with_gil(|py| { - let tb = match err.traceback_bound(py).map(|t| t.format()) { + let tb = match err.traceback(py).map(|t| t.format()) { Some(Ok(tb)) => tb, _ => String::new(), }; diff --git a/src/wsgi/callbacks.rs b/src/wsgi/callbacks.rs index 49ca894..5946292 100644 --- a/src/wsgi/callbacks.rs +++ b/src/wsgi/callbacks.rs @@ -67,7 +67,7 @@ fn run_callback( let _ = Python::with_gil(|py| -> PyResult<()> { let proto = Py::new(py, WSGIProtocol::new(tx))?; let callback = cbs.get().cb.clone_ref(py); - let environ = PyDict::new_bound(py); + let environ = PyDict::new(py); environ.set_item(pyo3::intern!(py, "SERVER_PROTOCOL"), version)?; environ.set_item(pyo3::intern!(py, "SERVER_NAME"), server.0)?; environ.set_item(pyo3::intern!(py, "SERVER_PORT"), server.1)?; @@ -75,7 +75,7 @@ fn run_callback( environ.set_item(pyo3::intern!(py, "REQUEST_METHOD"), parts.method.as_str())?; environ.set_item( pyo3::intern!(py, "PATH_INFO"), - PyBytes::new_bound(py, &path).call_method1(pyo3::intern!(py, "decode"), (pyo3::intern!(py, "latin1"),))?, + PyBytes::new(py, &path).call_method1(pyo3::intern!(py, "decode"), (pyo3::intern!(py, "latin1"),))?, )?; environ.set_item(pyo3::intern!(py, "QUERY_STRING"), query_string)?; environ.set_item(pyo3::intern!(py, "wsgi.url_scheme"), scheme)?; @@ -92,7 +92,7 @@ fn run_callback( content_len.to_str().unwrap_or_default(), )?; } - environ.update(headers.into_py_dict_bound(py).as_mapping())?; + environ.update(headers.into_py_dict(py).unwrap().as_mapping())?; if let Err(err) = callback.call1(py, (proto.clone_ref(py), environ)) { log_application_callable_exception(&err); diff --git a/src/wsgi/types.rs b/src/wsgi/types.rs index deeb567..9a26233 100644 --- a/src/wsgi/types.rs +++ b/src/wsgi/types.rs @@ -147,7 +147,7 @@ impl WSGIBody { } #[pyo3(signature = (_hint=None))] - fn readlines<'p>(&self, py: Python<'p>, _hint: Option<PyObject>) -> Bound<'p, PyList> { + fn readlines<'p>(&self, py: Python<'p>, _hint: Option<PyObject>) -> PyResult<Bound<'p, PyList>> { let inner = self.inner.clone(); let data = py.allow_threads(|| { self.rt.inner.block_on(async move { @@ -159,8 +159,8 @@ impl WSGIBody { }); let lines: Vec<Bound<PyBytes>> = data .split(|&c| c == LINE_SPLIT) - .map(|item| PyBytes::new_bound(py, item)) + .map(|item| PyBytes::new(py, item)) .collect(); - PyList::new_bound(py, lines) + PyList::new(py, lines) } }

Drop Python 3.8 support As it reached EOL

2024-12-01T23:18:50

0.0

[]

emmett-framework/granian

emmett-framework__granian-452

1332b2a3c6f05c8d68cb99d49a5c18bbb4eef3fc

diff --git a/.github/workflows/build.yml b/.github/workflows/build.yml index cc875242..e2f9c11c 100644 --- a/.github/workflows/build.yml +++ b/.github/workflows/build.yml @@ -118,7 +118,6 @@ jobs: pip install .[test] pip install granian --no-index --no-deps --find-links pgo_wheel --force-reinstall PGO_RUN=y pytest tests - PGO_RUN=y LOOP_OPT=y pytest tests/test_asgi.py tests/test_rsgi.py - name: merge PGO data run: ${{ env.LLVM_PROFDATA }} merge -o ${{ github.workspace }}/merged.profdata ${{ github.workspace }}/profdata - name: Build PGO wheel diff --git a/.github/workflows/release.yml b/.github/workflows/release.yml index b5b7928e..f2d30443 100644 --- a/.github/workflows/release.yml +++ b/.github/workflows/release.yml @@ -139,7 +139,6 @@ jobs: pip install .[test] pip install granian --no-index --no-deps --find-links pgo_wheel --force-reinstall PGO_RUN=y pytest tests - PGO_RUN=y LOOP_OPT=y pytest tests/test_asgi.py tests/test_rsgi.py - name: merge PGO data run: ${{ env.LLVM_PROFDATA }} merge -o ${{ github.workspace }}/merged.profdata ${{ github.workspace }}/profdata - name: Build PGO wheel diff --git a/README.md b/README.md index 0bb22b1f..087f68d1 100644 --- a/README.md +++ b/README.md @@ -100,8 +100,6 @@ Options: GRANIAN_THREADING_MODE; default: (workers)] --loop [auto|asyncio|uvloop] Event loop implementation [env var: GRANIAN_LOOP; default: (auto)] - --opt / --no-opt Enable loop optimizations [env var: - GRANIAN_LOOP_OPT; default: (disabled)] --backlog INTEGER RANGE Maximum number of connections to hold in backlog (globally) [env var: GRANIAN_BACKLOG; default: 1024; x>=128] diff --git a/granian/_futures.py b/granian/_futures.py index 2abeee9f..3084630f 100644 --- a/granian/_futures.py +++ b/granian/_futures.py @@ -1,4 +1,9 @@ -def future_watcher_wrapper(inner): +from asyncio.tasks import _enter_task, _leave_task + +from ._granian import CallbackScheduler as _BaseCBScheduler + + +def _future_watcher_wrapper(inner): async def future_watcher(watcher): try: await inner(watcher.scope, watcher.proto) @@ -8,3 +13,59 @@ async def future_watcher(watcher): watcher.done() return future_watcher + + +class _CBScheduler(_BaseCBScheduler): + __slots__ = [] + + def __init__(self, loop, ctx, cb): + super().__init__() + self._schedule_fn = _cbsched_schedule(loop, ctx, self._run, cb) + + def _waker(self, coro): + def _wake(fut): + self._resume(coro, fut) + + return _wake + + def _resume(self, coro, fut): + try: + fut.result() + except BaseException as exc: + self._throw(coro, exc) + else: + self._run(coro) + + def _run(self, coro): + _enter_task(self._loop, self) + try: + try: + result = coro.send(None) + except (KeyboardInterrupt, SystemExit): + raise + except BaseException: + pass + else: + if getattr(result, '_asyncio_future_blocking', None): + result._asyncio_future_blocking = False + result.add_done_callback(self._waker(coro), context=self._ctx) + elif result is None: + self._loop.call_soon(self._run, coro, context=self._ctx) + finally: + _leave_task(self._loop, self) + + def _throw(self, coro, exc): + _enter_task(self._loop, self) + try: + coro.throw(exc) + except BaseException: + pass + finally: + _leave_task(self._loop, self) + + +def _cbsched_schedule(loop, ctx, run, cb): + def _schedule(watcher): + loop.call_soon_threadsafe(run, cb(watcher), context=ctx) + + return _schedule diff --git a/granian/_granian.pyi b/granian/_granian.pyi index 55c2134a..0072f7c7 100644 --- a/granian/_granian.pyi +++ b/granian/_granian.pyi @@ -107,3 +107,7 @@ class ListenerHolder: @classmethod def from_address(cls, address: str, port: int, backlog: int) -> ListenerHolder: ... def get_fd(self) -> Any: ... + +class CallbackScheduler: + _loop: Any + _ctx: Any diff --git a/granian/cli.py b/granian/cli.py index 7060e5f3..bb5cbdb9 100644 --- a/granian/cli.py +++ b/granian/cli.py @@ -77,7 +77,6 @@ def option(*param_decls: str, cls: Optional[Type[click.Option]] = None, **attrs: help='Threading mode to use', ) @option('--loop', type=EnumType(Loops), default=Loops.auto, help='Event loop implementation') -@option('--opt/--no-opt', 'loop_opt', default=False, help='Enable loop optimizations') @option( '--backlog', type=click.IntRange(128), @@ -256,7 +255,6 @@ def cli( blocking_threads: Optional[int], threading_mode: ThreadModes, loop: Loops, - loop_opt: bool, backlog: int, backpressure: Optional[int], http1_buffer_size: int, @@ -311,7 +309,6 @@ def cli( blocking_threads=blocking_threads, threading_mode=threading_mode, loop=loop, - loop_opt=loop_opt, http=http, websockets=websockets, backlog=backlog, diff --git a/granian/server.py b/granian/server.py index bf658476..1e78f38b 100644 --- a/granian/server.py +++ b/granian/server.py @@ -13,7 +13,7 @@ from pathlib import Path from typing import Any, Callable, Dict, List, Optional, Sequence, Tuple, Type -from ._futures import future_watcher_wrapper +from ._futures import _CBScheduler, _future_watcher_wrapper from ._granian import ASGIWorker, RSGIWorker, WSGIWorker from ._imports import setproctitle, watchfiles from ._internal import load_target @@ -78,7 +78,6 @@ def __init__( blocking_threads: Optional[int] = None, threading_mode: ThreadModes = ThreadModes.workers, loop: Loops = Loops.auto, - loop_opt: bool = False, http: HTTPModes = HTTPModes.auto, websockets: bool = True, backlog: int = 1024, @@ -115,7 +114,6 @@ def __init__( self.threads = max(1, threads) self.threading_mode = threading_mode self.loop = loop - self.loop_opt = loop_opt self.http = http self.websockets = websockets self.backlog = max(128, backlog) @@ -189,7 +187,6 @@ def _spawn_asgi_worker( http1_settings: Optional[HTTP1Settings], http2_settings: Optional[HTTP2Settings], websockets: bool, - loop_opt: bool, log_enabled: bool, log_level: LogLevels, log_config: Dict[str, Any], @@ -207,12 +204,8 @@ def _spawn_asgi_worker( loop = loops.get(loop_impl) sfd = socket.fileno() callback = callback_loader() - shutdown_event = set_loop_signals(loop) - wcallback = _asgi_call_wrap(callback, scope_opts, {}, log_access_fmt) - if not loop_opt: - wcallback = future_watcher_wrapper(wcallback) worker = ASGIWorker( worker_id, @@ -224,11 +217,11 @@ def _spawn_asgi_worker( http1_settings, http2_settings, websockets, - loop_opt, *ssl_ctx, ) serve = getattr(worker, {ThreadModes.runtime: 'serve_rth', ThreadModes.workers: 'serve_wth'}[threading_mode]) - serve(wcallback, loop, contextvars.copy_context(), shutdown_event) + scheduler = _CBScheduler(loop, contextvars.copy_context(), _future_watcher_wrapper(wcallback)) + serve(scheduler, loop, shutdown_event) @staticmethod def _spawn_asgi_lifespan_worker( @@ -245,7 +238,6 @@ def _spawn_asgi_lifespan_worker( http1_settings: Optional[HTTP1Settings], http2_settings: Optional[HTTP2Settings], websockets: bool, - loop_opt: bool, log_enabled: bool, log_level: LogLevels, log_config: Dict[str, Any], @@ -271,10 +263,7 @@ def _spawn_asgi_lifespan_worker( sys.exit(1) shutdown_event = set_loop_signals(loop) - wcallback = _asgi_call_wrap(callback, scope_opts, lifespan_handler.state, log_access_fmt) - if not loop_opt: - wcallback = future_watcher_wrapper(wcallback) worker = ASGIWorker( worker_id, @@ -286,11 +275,11 @@ def _spawn_asgi_lifespan_worker( http1_settings, http2_settings, websockets, - loop_opt, *ssl_ctx, ) serve = getattr(worker, {ThreadModes.runtime: 'serve_rth', ThreadModes.workers: 'serve_wth'}[threading_mode]) - serve(wcallback, loop, contextvars.copy_context(), shutdown_event) + scheduler = _CBScheduler(loop, contextvars.copy_context(), _future_watcher_wrapper(wcallback)) + serve(scheduler, loop, shutdown_event) loop.run_until_complete(lifespan_handler.shutdown()) @staticmethod @@ -308,7 +297,6 @@ def _spawn_rsgi_worker( http1_settings: Optional[HTTP1Settings], http2_settings: Optional[HTTP2Settings], websockets: bool, - loop_opt: bool, log_enabled: bool, log_level: LogLevels, log_config: Dict[str, Any], @@ -334,7 +322,6 @@ def _spawn_rsgi_worker( getattr(target, '__rsgi_del__') if hasattr(target, '__rsgi_del__') else lambda *args, **kwargs: None ) callback = _rsgi_call_wrap(callback, log_access_fmt) - shutdown_event = set_loop_signals(loop) callback_init(loop) @@ -348,16 +335,11 @@ def _spawn_rsgi_worker( http1_settings, http2_settings, websockets, - loop_opt, *ssl_ctx, ) serve = getattr(worker, {ThreadModes.runtime: 'serve_rth', ThreadModes.workers: 'serve_wth'}[threading_mode]) - serve( - future_watcher_wrapper(callback) if not loop_opt else callback, - loop, - contextvars.copy_context(), - shutdown_event, - ) + scheduler = _CBScheduler(loop, contextvars.copy_context(), _future_watcher_wrapper(callback)) + serve(scheduler, loop, shutdown_event) callback_del(loop) @staticmethod @@ -375,7 +357,6 @@ def _spawn_wsgi_worker( http1_settings: Optional[HTTP1Settings], http2_settings: Optional[HTTP2Settings], websockets: bool, - loop_opt: bool, log_enabled: bool, log_level: LogLevels, log_config: Dict[str, Any], @@ -393,14 +374,16 @@ def _spawn_wsgi_worker( loop = loops.get(loop_impl) sfd = socket.fileno() callback = callback_loader() - shutdown_event = set_sync_signals() worker = WSGIWorker( worker_id, sfd, threads, blocking_threads, backpressure, http_mode, http1_settings, http2_settings, *ssl_ctx ) serve = getattr(worker, {ThreadModes.runtime: 'serve_rth', ThreadModes.workers: 'serve_wth'}[threading_mode]) - serve(_wsgi_call_wrap(callback, scope_opts, log_access_fmt), loop, contextvars.copy_context(), shutdown_event) + scheduler = _CBScheduler( + loop, contextvars.copy_context(), _wsgi_call_wrap(callback, scope_opts, log_access_fmt) + ) + serve(scheduler, loop, shutdown_event) shutdown_event.qs.wait() def _init_shared_socket(self): @@ -434,7 +417,6 @@ def _spawn_proc(self, idx, target, callback_loader, socket_loader) -> Worker: self.http1_settings, self.http2_settings, self.websockets, - self.loop_opt, self.log_enabled, self.log_level, self.log_config, diff --git a/pyproject.toml b/pyproject.toml index c80fb897..fb080770 100644 --- a/pyproject.toml +++ b/pyproject.toml @@ -52,6 +52,7 @@ test = [ 'httpx~=0.25.0', 'pytest~=7.4.2', 'pytest-asyncio~=0.21.1', + 'sniffio~=1.3', 'websockets~=11.0', ] all = ['granian[pname,reload]'] @@ -95,6 +96,7 @@ extend-ignore = [ 'E501', # leave line length to black 'N818', # leave to us exceptions naming 'S101', # assert is fine + 'S110', # except pass is fine ] flake8-quotes = { inline-quotes = 'single', multiline-quotes = 'double' } mccabe = { max-complexity = 13 } diff --git a/src/asgi/callbacks.rs b/src/asgi/callbacks.rs index 300a60f3..d2cc696e 100644 --- a/src/asgi/callbacks.rs +++ b/src/asgi/callbacks.rs @@ -1,9 +1,6 @@ use pyo3::prelude::*; use pyo3::types::PyDict; -use std::{ - net::SocketAddr, - sync::{Arc, Mutex}, -}; +use std::{net::SocketAddr, sync::Arc}; use tokio::sync::oneshot; use super::{ @@ -11,50 +8,13 @@ use super::{ utils::{build_scope_http, build_scope_ws, scope_native_parts}, }; use crate::{ - asyncio::PyContext, - callbacks::{ - callback_impl_loop_err, callback_impl_loop_pytask, callback_impl_loop_run, callback_impl_loop_step, - callback_impl_loop_wake, callback_impl_run, callback_impl_run_pytask, CallbackWrapper, - }, + callbacks::ArcCBScheduler, http::{response_500, HTTPResponse}, runtime::RuntimeRef, utils::log_application_callable_exception, ws::{HyperWebsocket, UpgradeData}, }; -#[pyclass(frozen)] -pub(crate) struct CallbackRunnerHTTP { - proto: Py<HTTPProtocol>, - context: PyContext, - cb: PyObject, -} - -impl CallbackRunnerHTTP { - pub fn new(py: Python, cb: CallbackWrapper, proto: HTTPProtocol, scope: Bound<PyDict>) -> Self { - let pyproto = Py::new(py, proto).unwrap(); - Self { - proto: pyproto.clone_ref(py), - context: cb.context, - cb: cb.callback.call1(py, (scope, pyproto)).unwrap(), - } - } - - callback_impl_run!(); -} - -#[pymethods] -impl CallbackRunnerHTTP { - fn _loop_task<'p>(&self, py: Python<'p>) -> PyResult<Bound<'p, PyAny>> { - CallbackTaskHTTP::new( - py, - self.cb.clone_ref(py), - self.proto.clone_ref(py), - self.context.clone(), - )? - .run(py) - } -} - macro_rules! callback_impl_done_http { ($self:expr) => { if let Some(tx) = $self.proto.get().tx() { @@ -63,6 +23,14 @@ macro_rules! callback_impl_done_http { }; } +macro_rules! callback_impl_done_ws { + ($self:expr) => { + if let (Some(tx), res) = $self.proto.get().tx() { + let _ = tx.send(res); + } + }; +} + macro_rules! callback_impl_done_err { ($self:expr, $err:expr) => { $self.done(); @@ -71,211 +39,58 @@ macro_rules! callback_impl_done_err { } #[pyclass(frozen)] -pub(crate) struct CallbackTaskHTTP { - proto: Py<HTTPProtocol>, - context: PyContext, - pycontext: PyObject, - cb: PyObject, -} - -impl CallbackTaskHTTP { - pub fn new(py: Python, cb: PyObject, proto: Py<HTTPProtocol>, context: PyContext) -> PyResult<Self> { - let pyctx = context.context(py); - Ok(Self { - proto, - context, - pycontext: pyctx.call_method0(pyo3::intern!(py, "copy"))?.into(), - cb, - }) - } - - fn done(&self) { - callback_impl_done_http!(self); - } - - fn err(&self, err: &PyErr) { - callback_impl_done_err!(self, err); - } - - callback_impl_loop_run!(); - callback_impl_loop_err!(); -} - -#[pymethods] -impl CallbackTaskHTTP { - fn _loop_step(pyself: PyRef<'_, Self>, py: Python) -> PyResult<()> { - callback_impl_loop_step!(pyself, py) - } - - fn _loop_wake(pyself: PyRef<'_, Self>, py: Python, fut: PyObject) -> PyResult<PyObject> { - callback_impl_loop_wake!(pyself, py, fut) - } -} - -#[pyclass(frozen)] -pub(crate) struct CallbackWrappedRunnerHTTP { +pub(crate) struct CallbackWatcherHTTP { #[pyo3(get)] proto: Py<HTTPProtocol>, - context: PyContext, - cb: PyObject, #[pyo3(get)] scope: PyObject, - pytaskref: Arc<Mutex<Option<PyObject>>>, } -impl CallbackWrappedRunnerHTTP { - pub fn new(py: Python, cb: CallbackWrapper, proto: HTTPProtocol, scope: Bound<PyDict>) -> Self { +impl CallbackWatcherHTTP { + pub fn new(py: Python, proto: HTTPProtocol, scope: Bound<PyDict>) -> Self { Self { proto: Py::new(py, proto).unwrap(), - context: cb.context, - cb: cb.callback.clone_ref(py), scope: scope.into_py(py), - pytaskref: Arc::new(Mutex::new(None)), } } - - callback_impl_run_pytask!(); } #[pymethods] -impl CallbackWrappedRunnerHTTP { - fn _loop_task<'p>(pyself: PyRef<'_, Self>, py: Python<'p>) -> PyResult<Bound<'p, PyAny>> { - callback_impl_loop_pytask!(pyself, py) - } - +impl CallbackWatcherHTTP { fn done(&self) { callback_impl_done_http!(self); - self.pytaskref.lock().unwrap().take(); } fn err(&self, err: Bound<PyAny>) { callback_impl_done_err!(self, &PyErr::from_value_bound(err)); - self.pytaskref.lock().unwrap().take(); - } -} - -#[pyclass(frozen)] -pub(crate) struct CallbackRunnerWebsocket { - proto: Py<WebsocketProtocol>, - context: PyContext, - cb: PyObject, -} - -impl CallbackRunnerWebsocket { - pub fn new(py: Python, cb: CallbackWrapper, proto: WebsocketProtocol, scope: Bound<PyDict>) -> Self { - let pyproto = Py::new(py, proto).unwrap(); - Self { - proto: pyproto.clone_ref(py), - context: cb.context, - cb: cb.callback.call1(py, (scope, pyproto)).unwrap(), - } } - - callback_impl_run!(); -} - -#[pymethods] -impl CallbackRunnerWebsocket { - fn _loop_task<'p>(&self, py: Python<'p>) -> PyResult<Bound<'p, PyAny>> { - CallbackTaskWebsocket::new( - py, - self.cb.clone_ref(py), - self.proto.clone_ref(py), - self.context.clone(), - )? - .run(py) - } -} - -macro_rules! callback_impl_done_ws { - ($self:expr) => { - if let (Some(tx), res) = $self.proto.get().tx() { - let _ = tx.send(res); - } - }; } #[pyclass(frozen)] -pub(crate) struct CallbackTaskWebsocket { - proto: Py<WebsocketProtocol>, - context: PyContext, - pycontext: PyObject, - cb: PyObject, -} - -impl CallbackTaskWebsocket { - pub fn new(py: Python, cb: PyObject, proto: Py<WebsocketProtocol>, context: PyContext) -> PyResult<Self> { - let pyctx = context.context(py); - Ok(Self { - proto, - context, - pycontext: pyctx.call_method0(pyo3::intern!(py, "copy"))?.into(), - cb, - }) - } - - fn done(&self) { - callback_impl_done_ws!(self); - } - - fn err(&self, err: &PyErr) { - callback_impl_done_err!(self, err); - } - - callback_impl_loop_run!(); - callback_impl_loop_err!(); -} - -#[pymethods] -impl CallbackTaskWebsocket { - fn _loop_step(pyself: PyRef<'_, Self>, py: Python) -> PyResult<()> { - callback_impl_loop_step!(pyself, py) - } - - fn _loop_wake(pyself: PyRef<'_, Self>, py: Python, fut: PyObject) -> PyResult<PyObject> { - callback_impl_loop_wake!(pyself, py, fut) - } -} - -#[pyclass(frozen)] -pub(crate) struct CallbackWrappedRunnerWebsocket { +pub(crate) struct CallbackWatcherWebsocket { #[pyo3(get)] proto: Py<WebsocketProtocol>, - context: PyContext, - cb: PyObject, #[pyo3(get)] scope: PyObject, - pytaskref: Arc<Mutex<Option<PyObject>>>, } -impl CallbackWrappedRunnerWebsocket { - pub fn new(py: Python, cb: CallbackWrapper, proto: WebsocketProtocol, scope: Bound<PyDict>) -> Self { +impl CallbackWatcherWebsocket { + pub fn new(py: Python, proto: WebsocketProtocol, scope: Bound<PyDict>) -> Self { Self { proto: Py::new(py, proto).unwrap(), - context: cb.context, - cb: cb.callback.clone_ref(py), scope: scope.into_py(py), - pytaskref: Arc::new(Mutex::new(None)), } } - - callback_impl_run_pytask!(); } #[pymethods] -impl CallbackWrappedRunnerWebsocket { - fn _loop_task<'p>(pyself: PyRef<'_, Self>, py: Python<'p>) -> PyResult<Bound<'p, PyAny>> { - callback_impl_loop_pytask!(pyself, py) - } - +impl CallbackWatcherWebsocket { fn done(&self) { callback_impl_done_ws!(self); - self.pytaskref.lock().unwrap().take(); } fn err(&self, err: Bound<PyAny>) { callback_impl_done_err!(self, &PyErr::from_value_bound(err)); - self.pytaskref.lock().unwrap().take(); } } @@ -302,88 +117,75 @@ impl CallbackWrappedRunnerWebsocket { // } // } -macro_rules! call_impl_http { - ($func_name:ident, $runner:ident) => { - #[inline] - pub(crate) fn $func_name( - cb: CallbackWrapper, - rt: RuntimeRef, - server_addr: SocketAddr, - client_addr: SocketAddr, - scheme: &str, - req: hyper::http::request::Parts, - body: hyper::body::Incoming, - ) -> oneshot::Receiver<HTTPResponse> { - let brt = rt.innerb.clone(); - let (tx, rx) = oneshot::channel(); - let protocol = HTTPProtocol::new(rt, body, tx); - let scheme: Arc<str> = scheme.into(); - - let _ = brt.run(move || { - scope_native_parts!( - req, - server_addr, - client_addr, - path, - query_string, - version, - server, - client - ); - Python::with_gil(|py| { - let scope = - build_scope_http(py, &req, version, server, client, &scheme, &path, query_string).unwrap(); - let _ = $runner::new(py, cb, protocol, scope).run(py); - }); - }); - - rx - } - }; -} - -macro_rules! call_impl_ws { - ($func_name:ident, $runner:ident) => { - #[inline] - pub(crate) fn $func_name( - cb: CallbackWrapper, - rt: RuntimeRef, - server_addr: SocketAddr, - client_addr: SocketAddr, - scheme: &str, - ws: HyperWebsocket, - req: hyper::http::request::Parts, - upgrade: UpgradeData, - ) -> oneshot::Receiver<WebsocketDetachedTransport> { - let brt = rt.innerb.clone(); - let (tx, rx) = oneshot::channel(); - let protocol = WebsocketProtocol::new(rt, tx, ws, upgrade); - let scheme: Arc<str> = scheme.into(); - - let _ = brt.run(move || { - scope_native_parts!( - req, - server_addr, - client_addr, - path, - query_string, - version, - server, - client - ); - Python::with_gil(|py| { - let scope = - build_scope_ws(py, &req, version, server, client, &scheme, &path, query_string).unwrap(); - let _ = $runner::new(py, cb, protocol, scope).run(py); - }); - }); - - rx - } - }; +#[inline] +pub(crate) fn call_http( + cb: ArcCBScheduler, + rt: RuntimeRef, + server_addr: SocketAddr, + client_addr: SocketAddr, + scheme: &str, + req: hyper::http::request::Parts, + body: hyper::body::Incoming, +) -> oneshot::Receiver<HTTPResponse> { + let brt = rt.innerb.clone(); + let (tx, rx) = oneshot::channel(); + let protocol = HTTPProtocol::new(rt, body, tx); + let scheme: Arc<str> = scheme.into(); + + let _ = brt.run(move || { + scope_native_parts!( + req, + server_addr, + client_addr, + path, + query_string, + version, + server, + client + ); + Python::with_gil(|py| { + let scope = build_scope_http(py, &req, version, server, client, &scheme, &path, query_string).unwrap(); + let watcher = Py::new(py, CallbackWatcherHTTP::new(py, protocol, scope)).unwrap(); + cb.get().schedule(py, watcher.as_any()); + }); + }); + + rx +} + +#[inline] +pub(crate) fn call_ws( + cb: ArcCBScheduler, + rt: RuntimeRef, + server_addr: SocketAddr, + client_addr: SocketAddr, + scheme: &str, + ws: HyperWebsocket, + req: hyper::http::request::Parts, + upgrade: UpgradeData, +) -> oneshot::Receiver<WebsocketDetachedTransport> { + let brt = rt.innerb.clone(); + let (tx, rx) = oneshot::channel(); + let protocol = WebsocketProtocol::new(rt, tx, ws, upgrade); + let scheme: Arc<str> = scheme.into(); + + let _ = brt.run(move || { + scope_native_parts!( + req, + server_addr, + client_addr, + path, + query_string, + version, + server, + client + ); + Python::with_gil(|py| { + let scope = build_scope_ws(py, &req, version, server, client, &scheme, &path, query_string).unwrap(); + let watcher = Py::new(py, CallbackWatcherWebsocket::new(py, protocol, scope)).unwrap(); + cb.get().schedule(py, watcher.as_any()); + }); + }); + + rx } - -call_impl_http!(call_http, CallbackRunnerHTTP); -call_impl_http!(call_http_pyw, CallbackWrappedRunnerHTTP); -call_impl_ws!(call_ws, CallbackRunnerWebsocket); -call_impl_ws!(call_ws_pyw, CallbackWrappedRunnerWebsocket); diff --git a/src/asgi/http.rs b/src/asgi/http.rs index 2aed8cbe..40c168c4 100644 --- a/src/asgi/http.rs +++ b/src/asgi/http.rs @@ -3,9 +3,9 @@ use hyper::{header::SERVER as HK_SERVER, http::response::Builder as ResponseBuil use std::net::SocketAddr; use tokio::sync::mpsc; -use super::callbacks::{call_http, call_http_pyw, call_ws, call_ws_pyw}; +use super::callbacks::{call_http, call_ws}; use crate::{ - callbacks::CallbackWrapper, + callbacks::ArcCBScheduler, http::{empty_body, response_500, HTTPRequest, HTTPResponse, HV_SERVER}, runtime::RuntimeRef, ws::{is_upgrade_request as is_ws_upgrade, upgrade_intent as ws_upgrade, UpgradeData}, @@ -32,7 +32,7 @@ macro_rules! handle_request { #[inline] pub(crate) async fn $func_name( rt: RuntimeRef, - callback: CallbackWrapper, + callback: ArcCBScheduler, server_addr: SocketAddr, client_addr: SocketAddr, req: HTTPRequest, @@ -58,7 +58,7 @@ macro_rules! handle_request_with_ws { #[inline] pub(crate) async fn $func_name( rt: RuntimeRef, - callback: CallbackWrapper, + callback: ArcCBScheduler, server_addr: SocketAddr, client_addr: SocketAddr, mut req: HTTPRequest, @@ -154,10 +154,4 @@ macro_rules! handle_request_with_ws { } handle_request!(handle, call_http); -// handle_request!(handle_rtb, call_rtb_http); -handle_request!(handle_pyw, call_http_pyw); -// handle_request!(handle_rtb_pyw, call_rtb_http_pyw); handle_request_with_ws!(handle_ws, call_http, call_ws); -// handle_request_with_ws!(handle_rtb_ws, call_rtb_http, call_rtb_ws); -handle_request_with_ws!(handle_ws_pyw, call_http_pyw, call_ws_pyw); -// handle_request_with_ws!(handle_rtb_ws_pyw, call_rtb_http_pyw, call_rtb_ws_pyw); diff --git a/src/asgi/serve.rs b/src/asgi/serve.rs index 4c8b9c14..dd7d1380 100644 --- a/src/asgi/serve.rs +++ b/src/asgi/serve.rs @@ -1,7 +1,8 @@ use pyo3::prelude::*; -use super::http::{handle, handle_pyw, handle_ws, handle_ws_pyw}; +use super::http::{handle, handle_ws}; +use crate::callbacks::CallbackScheduler; use crate::conversion::{worker_http1_config_from_py, worker_http2_config_from_py}; use crate::workers::{serve_rth, serve_rth_ssl, serve_wth, serve_wth_ssl, WorkerConfig, WorkerSignal, WorkerSignals}; @@ -12,21 +13,13 @@ pub struct ASGIWorker { impl ASGIWorker { serve_rth!(_serve_rth, handle); - serve_rth!(_serve_rth_pyw, handle_pyw); serve_rth!(_serve_rth_ws, handle_ws); - serve_rth!(_serve_rth_ws_pyw, handle_ws_pyw); serve_wth!(_serve_wth, handle); - serve_wth!(_serve_wth_pyw, handle_pyw); serve_wth!(_serve_wth_ws, handle_ws); - serve_wth!(_serve_wth_ws_pyw, handle_ws_pyw); serve_rth_ssl!(_serve_rth_ssl, handle); - serve_rth_ssl!(_serve_rth_ssl_pyw, handle_pyw); serve_rth_ssl!(_serve_rth_ssl_ws, handle_ws); - serve_rth_ssl!(_serve_rth_ssl_ws_pyw, handle_ws_pyw); serve_wth_ssl!(_serve_wth_ssl, handle); - serve_wth_ssl!(_serve_wth_ssl_pyw, handle_pyw); serve_wth_ssl!(_serve_wth_ssl_ws, handle_ws); - serve_wth_ssl!(_serve_wth_ssl_ws_pyw, handle_ws_pyw); } #[pymethods] @@ -43,7 +36,6 @@ impl ASGIWorker { http1_opts=None, http2_opts=None, websockets_enabled=false, - opt_enabled=true, ssl_enabled=false, ssl_cert=None, ssl_key=None, @@ -61,7 +53,6 @@ impl ASGIWorker { http1_opts: Option<PyObject>, http2_opts: Option<PyObject>, websockets_enabled: bool, - opt_enabled: bool, ssl_enabled: bool, ssl_cert: Option<&str>, ssl_key: Option<&str>, @@ -78,7 +69,6 @@ impl ASGIWorker { worker_http1_config_from_py(py, http1_opts)?, worker_http2_config_from_py(py, http2_opts)?, websockets_enabled, - opt_enabled, ssl_enabled, ssl_cert, ssl_key, @@ -87,57 +77,21 @@ impl ASGIWorker { }) } - fn serve_rth( - &self, - callback: PyObject, - event_loop: &Bound<PyAny>, - context: Bound<PyAny>, - signal: Py<WorkerSignal>, - ) { - match ( - self.config.websockets_enabled, - self.config.ssl_enabled, - self.config.opt_enabled, - ) { - (false, false, true) => self._serve_rth(callback, event_loop, context, WorkerSignals::Tokio(signal)), - (false, false, false) => self._serve_rth_pyw(callback, event_loop, context, WorkerSignals::Tokio(signal)), - (true, false, true) => self._serve_rth_ws(callback, event_loop, context, WorkerSignals::Tokio(signal)), - (true, false, false) => self._serve_rth_ws_pyw(callback, event_loop, context, WorkerSignals::Tokio(signal)), - (false, true, true) => self._serve_rth_ssl(callback, event_loop, context, WorkerSignals::Tokio(signal)), - (false, true, false) => { - self._serve_rth_ssl_pyw(callback, event_loop, context, WorkerSignals::Tokio(signal)); - } - (true, true, true) => self._serve_rth_ssl_ws(callback, event_loop, context, WorkerSignals::Tokio(signal)), - (true, true, false) => { - self._serve_rth_ssl_ws_pyw(callback, event_loop, context, WorkerSignals::Tokio(signal)); - } + fn serve_rth(&self, callback: Py<CallbackScheduler>, event_loop: &Bound<PyAny>, signal: Py<WorkerSignal>) { + match (self.config.websockets_enabled, self.config.ssl_enabled) { + (false, false) => self._serve_rth(callback, event_loop, WorkerSignals::Tokio(signal)), + (true, false) => self._serve_rth_ws(callback, event_loop, WorkerSignals::Tokio(signal)), + (false, true) => self._serve_rth_ssl(callback, event_loop, WorkerSignals::Tokio(signal)), + (true, true) => self._serve_rth_ssl_ws(callback, event_loop, WorkerSignals::Tokio(signal)), } } - fn serve_wth( - &self, - callback: PyObject, - event_loop: &Bound<PyAny>, - context: Bound<PyAny>, - signal: Py<WorkerSignal>, - ) { - match ( - self.config.websockets_enabled, - self.config.ssl_enabled, - self.config.opt_enabled, - ) { - (false, false, true) => self._serve_wth(callback, event_loop, context, WorkerSignals::Tokio(signal)), - (false, false, false) => self._serve_wth_pyw(callback, event_loop, context, WorkerSignals::Tokio(signal)), - (true, false, true) => self._serve_wth_ws(callback, event_loop, context, WorkerSignals::Tokio(signal)), - (true, false, false) => self._serve_wth_ws_pyw(callback, event_loop, context, WorkerSignals::Tokio(signal)), - (false, true, true) => self._serve_wth_ssl(callback, event_loop, context, WorkerSignals::Tokio(signal)), - (false, true, false) => { - self._serve_wth_ssl_pyw(callback, event_loop, context, WorkerSignals::Tokio(signal)); - } - (true, true, true) => self._serve_wth_ssl_ws(callback, event_loop, context, WorkerSignals::Tokio(signal)), - (true, true, false) => { - self._serve_wth_ssl_ws_pyw(callback, event_loop, context, WorkerSignals::Tokio(signal)); - } + fn serve_wth(&self, callback: Py<CallbackScheduler>, event_loop: &Bound<PyAny>, signal: Py<WorkerSignal>) { + match (self.config.websockets_enabled, self.config.ssl_enabled) { + (false, false) => self._serve_wth(callback, event_loop, WorkerSignals::Tokio(signal)), + (true, false) => self._serve_wth_ws(callback, event_loop, WorkerSignals::Tokio(signal)), + (false, true) => self._serve_wth_ssl(callback, event_loop, WorkerSignals::Tokio(signal)), + (true, true) => self._serve_wth_ssl_ws(callback, event_loop, WorkerSignals::Tokio(signal)), } } } diff --git a/src/asyncio.rs b/src/asyncio.rs index 7767ed9d..f6b1b38a 100644 --- a/src/asyncio.rs +++ b/src/asyncio.rs @@ -1,46 +1,16 @@ use pyo3::{prelude::*, sync::GILOnceCell}; -use std::{convert::Into, sync::Arc}; +use std::convert::Into; static CONTEXTVARS: GILOnceCell<PyObject> = GILOnceCell::new(); static CONTEXT: GILOnceCell<PyObject> = GILOnceCell::new(); -#[derive(Clone, Debug)] -pub struct PyContext { - event_loop: Arc<PyObject>, - context: Arc<PyObject>, -} - -impl PyContext { - pub fn new(event_loop: Bound<PyAny>) -> Self { - let pynone = event_loop.py().None(); - Self { - event_loop: Arc::new(event_loop.unbind()), - context: Arc::new(pynone), - } - } - - pub fn with_context(self, context: Bound<PyAny>) -> Self { - Self { - context: Arc::new(context.unbind()), - ..self - } - } - - pub fn event_loop<'p>(&self, py: Python<'p>) -> Bound<'p, PyAny> { - self.event_loop.clone_ref(py).into_bound(py) - } - - pub fn context<'p>(&self, py: Python<'p>) -> Bound<'p, PyAny> { - self.context.clone_ref(py).into_bound(py) - } -} - fn contextvars(py: Python) -> PyResult<&Bound<PyAny>> { Ok(CONTEXTVARS .get_or_try_init(py, || py.import_bound("contextvars").map(Into::into))? .bind(py)) } +#[allow(dead_code)] pub(crate) fn empty_context(py: Python) -> PyResult<&Bound<PyAny>> { Ok(CONTEXT .get_or_try_init(py, || { diff --git a/src/callbacks.rs b/src/callbacks.rs index 07a6bfb4..2944fbff 100644 --- a/src/callbacks.rs +++ b/src/callbacks.rs @@ -3,21 +3,48 @@ use pyo3::{exceptions::PyStopIteration, prelude::*}; use std::sync::{atomic, Arc, RwLock}; use tokio::sync::Notify; -use super::asyncio::PyContext; +pub(crate) type ArcCBScheduler = Arc<Py<CallbackScheduler>>; + +#[pyclass(frozen, subclass)] +pub(crate) struct CallbackScheduler { + #[pyo3(get)] + _loop: PyObject, + #[pyo3(get)] + _ctx: PyObject, + schedule_fn: Arc<RwLock<PyObject>>, + pub cb: PyObject, +} -#[derive(Clone)] -pub(crate) struct CallbackWrapper { - pub callback: Arc<PyObject>, - pub context: PyContext, +impl CallbackScheduler { + #[inline] + pub(crate) fn schedule(&self, _py: Python, watcher: &PyObject) { + // // let cb = self.cb.as_ptr(); + let cbarg = watcher.as_ptr(); + let sched = self.schedule_fn.read().unwrap().as_ptr(); + unsafe { + // let coro = pyo3::ffi::PyObject_CallOneArg(cb, cbarg); + pyo3::ffi::PyObject_CallOneArg(sched, cbarg); + } + } } -impl CallbackWrapper { - pub(crate) fn new(callback: PyObject, event_loop: Bound<PyAny>, context: Bound<PyAny>) -> Self { +#[pymethods] +impl CallbackScheduler { + #[new] + fn new(py: Python, event_loop: PyObject, ctx: PyObject, cb: PyObject) -> Self { Self { - callback: Arc::new(callback), - context: PyContext::new(event_loop).with_context(context), + _loop: event_loop, + _ctx: ctx, + schedule_fn: Arc::new(RwLock::new(py.None())), + cb, } } + + #[setter(_schedule_fn)] + fn _set_schedule_fn(&self, val: PyObject) { + let mut guard = self.schedule_fn.write().unwrap(); + *guard = val; + } } #[pyclass(frozen)] @@ -284,135 +311,3 @@ impl PyFutureResultSetter { let _ = target.call1((value,)); } } - -macro_rules! callback_impl_run { - () => { - pub fn run(self, py: Python<'_>) -> PyResult<Bound<PyAny>> { - let event_loop = self.context.event_loop(py); - let target = self.into_py(py).getattr(py, pyo3::intern!(py, "_loop_task"))?; - let kwctx = pyo3::types::PyDict::new_bound(py); - kwctx.set_item(pyo3::intern!(py, "context"), crate::asyncio::empty_context(py)?)?; - event_loop.call_method(pyo3::intern!(py, "call_soon_threadsafe"), (target,), Some(&kwctx)) - } - }; -} - -macro_rules! callback_impl_run_pytask { - () => { - pub fn run(self, py: Python<'_>) -> PyResult<Bound<PyAny>> { - let taskref = self.pytaskref.clone(); - let event_loop = self.context.event_loop(py); - let context = self.context.context(py); - let target = self.into_py(py).getattr(py, pyo3::intern!(py, "_loop_task"))?; - let kwctx = pyo3::types::PyDict::new_bound(py); - { - let mut taskref_guard = taskref.lock().unwrap(); - *taskref_guard = Some(target.clone_ref(py)); - } - kwctx.set_item(pyo3::intern!(py, "context"), context)?; - event_loop.call_method(pyo3::intern!(py, "call_soon_threadsafe"), (target,), Some(&kwctx)) - } - }; -} - -macro_rules! callback_impl_loop_run { - () => { - pub fn run(self, py: Python<'_>) -> PyResult<Bound<PyAny>> { - let context = self.pycontext.clone_ref(py).into_bound(py); - context.call_method1( - pyo3::intern!(py, "run"), - (self.into_py(py).getattr(py, pyo3::intern!(py, "_loop_step"))?,), - ) - } - }; -} - -macro_rules! callback_impl_loop_pytask { - ($pyself:expr, $py:expr) => { - $pyself.context.event_loop($py).call_method1( - pyo3::intern!($py, "create_task"), - ($pyself - .cb - .clone_ref($py) - .into_bound($py) - .call1(($pyself.into_py($py),))?,), - ) - }; -} - -macro_rules! callback_impl_loop_step { - ($pyself:expr, $py:expr) => { - match $pyself.cb.call_method1($py, pyo3::intern!($py, "send"), ($py.None(),)) { - Ok(res) => { - let blocking: bool = match res.getattr($py, pyo3::intern!($py, "_asyncio_future_blocking")) { - Ok(v) => v.extract($py)?, - _ => false, - }; - - let ctx = $pyself.pycontext.clone_ref($py); - let kwctx = pyo3::types::PyDict::new_bound($py); - kwctx.set_item(pyo3::intern!($py, "context"), ctx)?; - - match blocking { - true => { - res.setattr($py, pyo3::intern!($py, "_asyncio_future_blocking"), false)?; - res.call_method_bound( - $py, - pyo3::intern!($py, "add_done_callback"), - ($pyself.into_py($py).getattr($py, pyo3::intern!($py, "_loop_wake"))?,), - Some(&kwctx), - )?; - } - false => { - let event_loop = $pyself.context.event_loop($py); - event_loop.call_method( - pyo3::intern!($py, "call_soon"), - ($pyself.into_py($py).getattr($py, pyo3::intern!($py, "_loop_step"))?,), - Some(&kwctx), - )?; - } - } - Ok(()) - } - Err(err) - if (err.is_instance_of::<pyo3::exceptions::PyStopIteration>($py) - || err.is_instance_of::<pyo3::exceptions::PyStopAsyncIteration>($py) - || err.is_instance_of::<pyo3::exceptions::asyncio::CancelledError>($py)) => - { - $pyself.done(); - Ok(()) - } - Err(err) => { - $pyself.err(&err); - Ok(()) - } - } - }; -} - -macro_rules! callback_impl_loop_wake { - ($pyself:expr, $py:expr, $fut:expr) => { - match $fut.call_method0($py, pyo3::intern!($py, "result")) { - Ok(_) => $pyself.into_py($py).call_method0($py, pyo3::intern!($py, "_loop_step")), - Err(err) => $pyself._loop_err($py, err), - } - }; -} - -macro_rules! callback_impl_loop_err { - () => { - pub fn _loop_err(&self, py: Python, err: PyErr) -> PyResult<PyObject> { - self.err(&err); - let cberr = self.cb.call_method1(py, pyo3::intern!(py, "throw"), (err,)); - cberr - } - }; -} - -pub(crate) use callback_impl_loop_err; -pub(crate) use callback_impl_loop_pytask; -pub(crate) use callback_impl_loop_run; -pub(crate) use callback_impl_loop_step; -pub(crate) use callback_impl_loop_wake; -pub(crate) use callback_impl_run; -pub(crate) use callback_impl_run_pytask; diff --git a/src/lib.rs b/src/lib.rs index d639101b..b26a7349 100644 --- a/src/lib.rs +++ b/src/lib.rs @@ -37,6 +37,7 @@ pub fn get_granian_version() -> &'static str { #[pymodule] fn _granian(py: Python, module: &Bound<PyModule>) -> PyResult<()> { module.add("__version__", get_granian_version())?; + module.add_class::<callbacks::CallbackScheduler>()?; asgi::init_pymodule(module)?; rsgi::init_pymodule(py, module)?; tcp::init_pymodule(module)?; diff --git a/src/rsgi/callbacks.rs b/src/rsgi/callbacks.rs index bdc61f2e..b67081b7 100644 --- a/src/rsgi/callbacks.rs +++ b/src/rsgi/callbacks.rs @@ -1,5 +1,4 @@ use pyo3::prelude::*; -use std::sync::{Arc, Mutex}; use tokio::sync::oneshot; use super::{ @@ -7,49 +6,12 @@ use super::{ types::{PyResponse, PyResponseBody, RSGIHTTPScope as HTTPScope, RSGIWebsocketScope as WebsocketScope}, }; use crate::{ - asyncio::PyContext, - callbacks::{ - callback_impl_loop_err, callback_impl_loop_pytask, callback_impl_loop_run, callback_impl_loop_step, - callback_impl_loop_wake, callback_impl_run, callback_impl_run_pytask, CallbackWrapper, - }, + callbacks::ArcCBScheduler, runtime::RuntimeRef, utils::log_application_callable_exception, ws::{HyperWebsocket, UpgradeData}, }; -#[pyclass(frozen)] -pub(crate) struct CallbackRunnerHTTP { - proto: Py<HTTPProtocol>, - context: PyContext, - cb: PyObject, -} - -impl CallbackRunnerHTTP { - pub fn new(py: Python, cb: CallbackWrapper, proto: HTTPProtocol, scope: HTTPScope) -> Self { - let pyproto = Py::new(py, proto).unwrap(); - Self { - proto: pyproto.clone_ref(py), - context: cb.context, - cb: cb.callback.call1(py, (scope, pyproto)).unwrap(), - } - } - - callback_impl_run!(); -} - -#[pymethods] -impl CallbackRunnerHTTP { - fn _loop_task<'p>(&self, py: Python<'p>) -> PyResult<Bound<'p, PyAny>> { - CallbackTaskHTTP::new( - py, - self.cb.clone_ref(py), - self.proto.clone_ref(py), - self.context.clone(), - )? - .run(py) - } -} - macro_rules! callback_impl_done_http { ($self:expr) => { if let Some(tx) = $self.proto.get().tx() { @@ -58,6 +20,12 @@ macro_rules! callback_impl_done_http { }; } +macro_rules! callback_impl_done_ws { + ($self:expr) => { + let _ = $self.proto.get().close(None); + }; +} + macro_rules! callback_impl_done_err { ($self:expr, $err:expr) => { $self.done(); @@ -66,262 +34,100 @@ macro_rules! callback_impl_done_err { } #[pyclass(frozen)] -pub(crate) struct CallbackTaskHTTP { - proto: Py<HTTPProtocol>, - context: PyContext, - pycontext: PyObject, - cb: PyObject, -} - -impl CallbackTaskHTTP { - pub fn new(py: Python, cb: PyObject, proto: Py<HTTPProtocol>, context: PyContext) -> PyResult<Self> { - let pyctx = context.context(py); - Ok(Self { - proto, - context, - pycontext: pyctx.call_method0(pyo3::intern!(py, "copy"))?.into(), - cb, - }) - } - - fn done(&self) { - callback_impl_done_http!(self); - } - - fn err(&self, err: &PyErr) { - callback_impl_done_err!(self, err); - } - - callback_impl_loop_run!(); - callback_impl_loop_err!(); -} - -#[pymethods] -impl CallbackTaskHTTP { - fn _loop_step(pyself: PyRef<'_, Self>, py: Python) -> PyResult<()> { - callback_impl_loop_step!(pyself, py) - } - - fn _loop_wake(pyself: PyRef<'_, Self>, py: Python, fut: PyObject) -> PyResult<PyObject> { - callback_impl_loop_wake!(pyself, py, fut) - } -} - -#[pyclass(frozen)] -pub(crate) struct CallbackWrappedRunnerHTTP { +pub(crate) struct CallbackWatcherHTTP { #[pyo3(get)] proto: Py<HTTPProtocol>, - context: PyContext, - cb: PyObject, #[pyo3(get)] scope: PyObject, - pytaskref: Arc<Mutex<Option<PyObject>>>, } -impl CallbackWrappedRunnerHTTP { - pub fn new(py: Python, cb: CallbackWrapper, proto: HTTPProtocol, scope: HTTPScope) -> Self { +impl CallbackWatcherHTTP { + pub fn new(py: Python, proto: HTTPProtocol, scope: HTTPScope) -> Self { Self { proto: Py::new(py, proto).unwrap(), - context: cb.context, - cb: cb.callback.clone_ref(py), scope: scope.into_py(py), - pytaskref: Arc::new(Mutex::new(None)), } } - - callback_impl_run_pytask!(); } #[pymethods] -impl CallbackWrappedRunnerHTTP { - fn _loop_task<'p>(pyself: PyRef<'_, Self>, py: Python<'p>) -> PyResult<Bound<'p, PyAny>> { - callback_impl_loop_pytask!(pyself, py) - } - +impl CallbackWatcherHTTP { fn done(&self) { callback_impl_done_http!(self); - self.pytaskref.lock().unwrap().take(); } fn err(&self, err: Bound<PyAny>) { callback_impl_done_err!(self, &PyErr::from_value_bound(err)); - self.pytaskref.lock().unwrap().take(); } } #[pyclass(frozen)] -pub(crate) struct CallbackRunnerWebsocket { - proto: Py<WebsocketProtocol>, - context: PyContext, - cb: PyObject, -} - -impl CallbackRunnerWebsocket { - pub fn new(py: Python, cb: CallbackWrapper, proto: WebsocketProtocol, scope: WebsocketScope) -> Self { - let pyproto = Py::new(py, proto).unwrap(); - Self { - proto: pyproto.clone_ref(py), - context: cb.context, - cb: cb.callback.call1(py, (scope, pyproto)).unwrap(), - } - } - - callback_impl_run!(); -} - -#[pymethods] -impl CallbackRunnerWebsocket { - fn _loop_task<'p>(&self, py: Python<'p>) -> PyResult<Bound<'p, PyAny>> { - CallbackTaskWebsocket::new( - py, - self.cb.clone_ref(py), - self.proto.clone_ref(py), - self.context.clone(), - )? - .run(py) - } -} - -macro_rules! callback_impl_done_ws { - ($self:expr) => { - let _ = $self.proto.get().close(None); - }; -} - -#[pyclass(frozen)] -pub(crate) struct CallbackTaskWebsocket { - proto: Py<WebsocketProtocol>, - context: PyContext, - pycontext: PyObject, - cb: PyObject, -} - -impl CallbackTaskWebsocket { - pub fn new(py: Python, cb: PyObject, proto: Py<WebsocketProtocol>, context: PyContext) -> PyResult<Self> { - let pyctx = context.context(py); - Ok(Self { - proto, - context, - pycontext: pyctx.call_method0(pyo3::intern!(py, "copy"))?.into(), - cb, - }) - } - - fn done(&self) { - callback_impl_done_ws!(self); - } - - fn err(&self, err: &PyErr) { - callback_impl_done_err!(self, err); - } - - callback_impl_loop_run!(); - callback_impl_loop_err!(); -} - -#[pymethods] -impl CallbackTaskWebsocket { - fn _loop_step(pyself: PyRef<'_, Self>, py: Python) -> PyResult<()> { - callback_impl_loop_step!(pyself, py) - } - - fn _loop_wake(pyself: PyRef<'_, Self>, py: Python, fut: PyObject) -> PyResult<PyObject> { - callback_impl_loop_wake!(pyself, py, fut) - } -} - -#[pyclass(frozen)] -pub(crate) struct CallbackWrappedRunnerWebsocket { +pub(crate) struct CallbackWatcherWebsocket { #[pyo3(get)] proto: Py<WebsocketProtocol>, - context: PyContext, - cb: PyObject, #[pyo3(get)] scope: PyObject, - pytaskref: Arc<Mutex<Option<PyObject>>>, } -impl CallbackWrappedRunnerWebsocket { - pub fn new(py: Python, cb: CallbackWrapper, proto: WebsocketProtocol, scope: WebsocketScope) -> Self { +impl CallbackWatcherWebsocket { + pub fn new(py: Python, proto: WebsocketProtocol, scope: WebsocketScope) -> Self { Self { proto: Py::new(py, proto).unwrap(), - context: cb.context, - cb: cb.callback.clone_ref(py), scope: scope.into_py(py), - pytaskref: Arc::new(Mutex::new(None)), } } - - callback_impl_run_pytask!(); } #[pymethods] -impl CallbackWrappedRunnerWebsocket { - fn _loop_task<'p>(pyself: PyRef<'_, Self>, py: Python<'p>) -> PyResult<Bound<'p, PyAny>> { - callback_impl_loop_pytask!(pyself, py) - } - +impl CallbackWatcherWebsocket { fn done(&self) { callback_impl_done_ws!(self); - self.pytaskref.lock().unwrap().take(); } fn err(&self, err: Bound<PyAny>) { callback_impl_done_err!(self, &PyErr::from_value_bound(err)); - self.pytaskref.lock().unwrap().take(); } } -macro_rules! call_impl_http { - ($func_name:ident, $runner:ident) => { - #[inline] - pub(crate) fn $func_name( - cb: CallbackWrapper, - rt: RuntimeRef, - body: hyper::body::Incoming, - scope: HTTPScope, - ) -> oneshot::Receiver<PyResponse> { - let brt = rt.innerb.clone(); - let (tx, rx) = oneshot::channel(); - let protocol = HTTPProtocol::new(rt, tx, body); - - let _ = brt.run(|| { - Python::with_gil(|py| { - let _ = $runner::new(py, cb, protocol, scope).run(py); - }); - }); - - rx - } - }; -} - -macro_rules! call_impl_ws { - ($func_name:ident, $runner:ident) => { - #[inline] - pub(crate) fn $func_name( - cb: CallbackWrapper, - rt: RuntimeRef, - ws: HyperWebsocket, - upgrade: UpgradeData, - scope: WebsocketScope, - ) -> oneshot::Receiver<WebsocketDetachedTransport> { - let brt = rt.innerb.clone(); - let (tx, rx) = oneshot::channel(); - let protocol = WebsocketProtocol::new(rt, tx, ws, upgrade); - - let _ = brt.run(|| { - Python::with_gil(|py| { - let _ = $runner::new(py, cb, protocol, scope).run(py); - }); - }); - - rx - } - }; +#[inline] +pub(crate) fn call_http( + cb: ArcCBScheduler, + rt: RuntimeRef, + body: hyper::body::Incoming, + scope: HTTPScope, +) -> oneshot::Receiver<PyResponse> { + let brt = rt.innerb.clone(); + let (tx, rx) = oneshot::channel(); + let protocol = HTTPProtocol::new(rt, tx, body); + + let _ = brt.run(move || { + Python::with_gil(|py| { + let watcher = Py::new(py, CallbackWatcherHTTP::new(py, protocol, scope)).unwrap(); + cb.get().schedule(py, watcher.as_any()); + }); + }); + + rx +} + +#[inline] +pub(crate) fn call_ws( + cb: ArcCBScheduler, + rt: RuntimeRef, + ws: HyperWebsocket, + upgrade: UpgradeData, + scope: WebsocketScope, +) -> oneshot::Receiver<WebsocketDetachedTransport> { + let brt = rt.innerb.clone(); + let (tx, rx) = oneshot::channel(); + let protocol = WebsocketProtocol::new(rt, tx, ws, upgrade); + + let _ = brt.run(move || { + Python::with_gil(|py| { + let watcher = Py::new(py, CallbackWatcherWebsocket::new(py, protocol, scope)).unwrap(); + cb.get().schedule(py, watcher.as_any()); + }); + }); + + rx } - -call_impl_http!(call_http, CallbackRunnerHTTP); -call_impl_http!(call_http_pyw, CallbackWrappedRunnerHTTP); -call_impl_ws!(call_ws, CallbackRunnerWebsocket); -call_impl_ws!(call_ws_pyw, CallbackWrappedRunnerWebsocket); diff --git a/src/rsgi/http.rs b/src/rsgi/http.rs index 50f52361..ee3278ba 100644 --- a/src/rsgi/http.rs +++ b/src/rsgi/http.rs @@ -4,11 +4,11 @@ use std::net::SocketAddr; use tokio::sync::mpsc; use super::{ - callbacks::{call_http, call_http_pyw, call_ws, call_ws_pyw}, + callbacks::{call_http, call_ws}, types::{PyResponse, RSGIHTTPScope as HTTPScope, RSGIWebsocketScope as WebsocketScope}, }; use crate::{ - callbacks::CallbackWrapper, + callbacks::ArcCBScheduler, http::{empty_body, response_500, HTTPRequest, HTTPResponse, HV_SERVER}, runtime::RuntimeRef, ws::{is_upgrade_request as is_ws_upgrade, upgrade_intent as ws_upgrade, UpgradeData}, @@ -46,7 +46,7 @@ macro_rules! handle_request { #[inline] pub(crate) async fn $func_name( rt: RuntimeRef, - callback: CallbackWrapper, + callback: ArcCBScheduler, server_addr: SocketAddr, client_addr: SocketAddr, req: HTTPRequest, @@ -59,12 +59,30 @@ macro_rules! handle_request { }; } +// macro_rules! handle_request2 { +// ($func_name:ident, $handler:expr) => { +// #[inline] +// pub(crate) async fn $func_name( +// rt: RuntimeRef, +// callback: ArcCBScheduler, +// server_addr: SocketAddr, +// client_addr: SocketAddr, +// req: HTTPRequest, +// scheme: &str, +// ) -> HTTPResponse { +// let (parts, body) = req.into_parts(); +// let scope = build_scope!(HTTPScope, server_addr, client_addr, parts, scheme); +// handle_http_response!($handler, rt, callback, body, scope) +// } +// }; +// } + macro_rules! handle_request_with_ws { ($func_name:ident, $handler_req:expr, $handler_ws:expr) => { #[inline] pub(crate) async fn $func_name( rt: RuntimeRef, - callback: CallbackWrapper, + callback: ArcCBScheduler, server_addr: SocketAddr, client_addr: SocketAddr, mut req: HTTPRequest, @@ -136,6 +154,7 @@ macro_rules! handle_request_with_ws { } handle_request!(handle, call_http); -handle_request!(handle_pyw, call_http_pyw); +// handle_request!(handle_pyw, call_http_pyw); +// handle_request2!(handle2, call_http2); handle_request_with_ws!(handle_ws, call_http, call_ws); -handle_request_with_ws!(handle_ws_pyw, call_http_pyw, call_ws_pyw); +// handle_request_with_ws!(handle_ws_pyw, call_http_pyw, call_ws_pyw); diff --git a/src/rsgi/io.rs b/src/rsgi/io.rs index e83c2246..cd813ea1 100644 --- a/src/rsgi/io.rs +++ b/src/rsgi/io.rs @@ -96,29 +96,37 @@ impl RSGIHTTPProtocol { error_proto!() } - fn __aiter__(pyself: PyRef<'_, Self>) -> PyRef<'_, Self> { - if let Some(body) = pyself.body.lock().unwrap().take() { - let mut stream = pyself.body_stream.blocking_lock(); + fn __aiter__(pyself: Py<Self>, py: Python) -> PyResult<Py<Self>> { + let inner = pyself.get(); + if let Some(body) = inner.body.lock().unwrap().take() { + let mut stream = inner.body_stream.blocking_lock(); *stream = Some(http_body_util::BodyStream::new(body)); + return Ok(pyself.clone_ref(py)); } - pyself + error_proto!() } - fn __anext__<'p>(&self, py: Python<'p>) -> PyResult<Option<Bound<'p, PyAny>>> { + fn __anext__<'p>(&self, py: Python<'p>) -> PyResult<Bound<'p, PyAny>> { + if self.body_stream.blocking_lock().is_none() { + return Err(pyo3::exceptions::PyStopAsyncIteration::new_err("stream exhausted")); + } let body_stream = self.body_stream.clone(); - let pyfut = future_into_py_iter(self.rt.clone(), py, async move { - if let Some(stream) = &mut *body_stream.lock().await { - if let Some(chunk) = stream.next().await { + future_into_py_iter(self.rt.clone(), py, async move { + let guard = &mut *body_stream.lock().await; + let bytes = match guard.as_mut().unwrap().next().await { + Some(chunk) => { let chunk = chunk .map(|buf| buf.into_data().unwrap_or_default()) .unwrap_or(body::Bytes::new()); - return Ok(BytesToPy(chunk)); - }; - return Err(pyo3::exceptions::PyStopAsyncIteration::new_err("stream exhausted")); - } - error_proto!() - })?; - Ok(Some(pyfut)) + BytesToPy(chunk) + } + _ => { + let _ = guard.take(); + BytesToPy(body::Bytes::new()) + } + }; + Ok(bytes) + }) } #[pyo3(signature = (status=200, headers=vec![]))] diff --git a/src/rsgi/serve.rs b/src/rsgi/serve.rs index 1f13dca4..bc904b91 100644 --- a/src/rsgi/serve.rs +++ b/src/rsgi/serve.rs @@ -1,7 +1,8 @@ use pyo3::prelude::*; -use super::http::{handle, handle_pyw, handle_ws, handle_ws_pyw}; +use super::http::{handle, handle_ws}; +use crate::callbacks::CallbackScheduler; use crate::conversion::{worker_http1_config_from_py, worker_http2_config_from_py}; use crate::workers::{serve_rth, serve_rth_ssl, serve_wth, serve_wth_ssl, WorkerConfig, WorkerSignal, WorkerSignals}; @@ -12,21 +13,13 @@ pub struct RSGIWorker { impl RSGIWorker { serve_rth!(_serve_rth, handle); - serve_rth!(_serve_rth_pyw, handle_pyw); serve_rth!(_serve_rth_ws, handle_ws); - serve_rth!(_serve_rth_ws_pyw, handle_ws_pyw); serve_wth!(_serve_wth, handle); - serve_wth!(_serve_wth_pyw, handle_pyw); serve_wth!(_serve_wth_ws, handle_ws); - serve_wth!(_serve_wth_ws_pyw, handle_ws_pyw); serve_rth_ssl!(_serve_rth_ssl, handle); - serve_rth_ssl!(_serve_rth_ssl_pyw, handle_pyw); serve_rth_ssl!(_serve_rth_ssl_ws, handle_ws); - serve_rth_ssl!(_serve_rth_ssl_ws_pyw, handle_ws_pyw); serve_wth_ssl!(_serve_wth_ssl, handle); - serve_wth_ssl!(_serve_wth_ssl_pyw, handle_pyw); serve_wth_ssl!(_serve_wth_ssl_ws, handle_ws); - serve_wth_ssl!(_serve_wth_ssl_ws_pyw, handle_ws_pyw); } #[pymethods] @@ -43,7 +36,6 @@ impl RSGIWorker { http1_opts=None, http2_opts=None, websockets_enabled=false, - opt_enabled=true, ssl_enabled=false, ssl_cert=None, ssl_key=None, @@ -61,7 +53,6 @@ impl RSGIWorker { http1_opts: Option<PyObject>, http2_opts: Option<PyObject>, websockets_enabled: bool, - opt_enabled: bool, ssl_enabled: bool, ssl_cert: Option<&str>, ssl_key: Option<&str>, @@ -78,7 +69,6 @@ impl RSGIWorker { worker_http1_config_from_py(py, http1_opts)?, worker_http2_config_from_py(py, http2_opts)?, websockets_enabled, - opt_enabled, ssl_enabled, ssl_cert, ssl_key, @@ -87,57 +77,21 @@ impl RSGIWorker { }) } - fn serve_rth( - &self, - callback: PyObject, - event_loop: &Bound<PyAny>, - context: Bound<PyAny>, - signal: Py<WorkerSignal>, - ) { - match ( - self.config.websockets_enabled, - self.config.ssl_enabled, - self.config.opt_enabled, - ) { - (false, false, true) => self._serve_rth(callback, event_loop, context, WorkerSignals::Tokio(signal)), - (false, false, false) => self._serve_rth_pyw(callback, event_loop, context, WorkerSignals::Tokio(signal)), - (true, false, true) => self._serve_rth_ws(callback, event_loop, context, WorkerSignals::Tokio(signal)), - (true, false, false) => self._serve_rth_ws_pyw(callback, event_loop, context, WorkerSignals::Tokio(signal)), - (false, true, true) => self._serve_rth_ssl(callback, event_loop, context, WorkerSignals::Tokio(signal)), - (false, true, false) => { - self._serve_rth_ssl_pyw(callback, event_loop, context, WorkerSignals::Tokio(signal)); - } - (true, true, true) => self._serve_rth_ssl_ws(callback, event_loop, context, WorkerSignals::Tokio(signal)), - (true, true, false) => { - self._serve_rth_ssl_ws_pyw(callback, event_loop, context, WorkerSignals::Tokio(signal)); - } + fn serve_rth(&self, callback: Py<CallbackScheduler>, event_loop: &Bound<PyAny>, signal: Py<WorkerSignal>) { + match (self.config.websockets_enabled, self.config.ssl_enabled) { + (false, false) => self._serve_rth(callback, event_loop, WorkerSignals::Tokio(signal)), + (true, false) => self._serve_rth_ws(callback, event_loop, WorkerSignals::Tokio(signal)), + (false, true) => self._serve_rth_ssl(callback, event_loop, WorkerSignals::Tokio(signal)), + (true, true) => self._serve_rth_ssl_ws(callback, event_loop, WorkerSignals::Tokio(signal)), } } - fn serve_wth( - &self, - callback: PyObject, - event_loop: &Bound<PyAny>, - context: Bound<PyAny>, - signal: Py<WorkerSignal>, - ) { - match ( - self.config.websockets_enabled, - self.config.ssl_enabled, - self.config.opt_enabled, - ) { - (false, false, true) => self._serve_wth(callback, event_loop, context, WorkerSignals::Tokio(signal)), - (false, false, false) => self._serve_wth_pyw(callback, event_loop, context, WorkerSignals::Tokio(signal)), - (true, false, true) => self._serve_wth_ws(callback, event_loop, context, WorkerSignals::Tokio(signal)), - (true, false, false) => self._serve_wth_ws_pyw(callback, event_loop, context, WorkerSignals::Tokio(signal)), - (false, true, true) => self._serve_wth_ssl(callback, event_loop, context, WorkerSignals::Tokio(signal)), - (false, true, false) => { - self._serve_wth_ssl_pyw(callback, event_loop, context, WorkerSignals::Tokio(signal)); - } - (true, true, true) => self._serve_wth_ssl_ws(callback, event_loop, context, WorkerSignals::Tokio(signal)), - (true, true, false) => { - self._serve_wth_ssl_ws_pyw(callback, event_loop, context, WorkerSignals::Tokio(signal)); - } + fn serve_wth(&self, callback: Py<CallbackScheduler>, event_loop: &Bound<PyAny>, signal: Py<WorkerSignal>) { + match (self.config.websockets_enabled, self.config.ssl_enabled) { + (false, false) => self._serve_wth(callback, event_loop, WorkerSignals::Tokio(signal)), + (true, false) => self._serve_wth_ws(callback, event_loop, WorkerSignals::Tokio(signal)), + (false, true) => self._serve_wth_ssl(callback, event_loop, WorkerSignals::Tokio(signal)), + (true, true) => self._serve_wth_ssl_ws(callback, event_loop, WorkerSignals::Tokio(signal)), } } } diff --git a/src/workers.rs b/src/workers.rs index 5eb4f316..ebb63a5a 100644 --- a/src/workers.rs +++ b/src/workers.rs @@ -100,7 +100,6 @@ pub(crate) struct WorkerConfig { pub http1_opts: HTTP1Config, pub http2_opts: HTTP2Config, pub websockets_enabled: bool, - pub opt_enabled: bool, pub ssl_enabled: bool, ssl_cert: Option<String>, ssl_key: Option<String>, @@ -118,7 +117,6 @@ impl WorkerConfig { http1_opts: HTTP1Config, http2_opts: HTTP2Config, websockets_enabled: bool, - opt_enabled: bool, ssl_enabled: bool, ssl_cert: Option<&str>, ssl_key: Option<&str>, @@ -134,7 +132,6 @@ impl WorkerConfig { http1_opts, http2_opts, websockets_enabled, - opt_enabled, ssl_enabled, ssl_cert: ssl_cert.map(std::convert::Into::into), ssl_key: ssl_key.map(std::convert::Into::into), @@ -593,9 +590,8 @@ macro_rules! serve_rth { ($func_name:ident, $target:expr) => { fn $func_name( &self, - callback: PyObject, + callback: Py<crate::callbacks::CallbackScheduler>, event_loop: &Bound<PyAny>, - context: Bound<PyAny>, signal: crate::workers::WorkerSignals, ) { pyo3_log::init(); @@ -609,7 +605,7 @@ macro_rules! serve_rth { let http1_opts = self.config.http1_opts.clone(); let http2_opts = self.config.http2_opts.clone(); let backpressure = self.config.backpressure.clone(); - let callback_wrapper = crate::callbacks::CallbackWrapper::new(callback, event_loop.clone(), context); + let callback_wrapper = std::sync::Arc::new(callback); let rt = crate::runtime::init_runtime_mt( self.config.threads, @@ -673,9 +669,9 @@ macro_rules! serve_rth_ssl { ($func_name:ident, $target:expr) => { fn $func_name( &self, - callback: PyObject, + callback: Py<crate::callbacks::CallbackScheduler>, event_loop: &Bound<PyAny>, - context: Bound<PyAny>, + // context: Bound<PyAny>, signal: crate::workers::WorkerSignals, ) { pyo3_log::init(); @@ -690,7 +686,8 @@ macro_rules! serve_rth_ssl { let http2_opts = self.config.http2_opts.clone(); let backpressure = self.config.backpressure.clone(); let tls_cfg = self.config.tls_cfg(); - let callback_wrapper = crate::callbacks::CallbackWrapper::new(callback, event_loop.clone(), context); + // let callback_wrapper = crate::callbacks::CallbackWrapper::new(callback, event_loop.clone(), context); + let callback_wrapper = std::sync::Arc::new(callback); let rt = crate::runtime::init_runtime_mt( self.config.threads, @@ -752,8 +749,9 @@ macro_rules! serve_rth_ssl { } macro_rules! serve_wth_inner { - ($self:expr, $target:expr, $callback:expr, $event_loop:expr, $context:expr, $wid:expr, $workers:expr, $srx:expr) => { - let callback_wrapper = crate::callbacks::CallbackWrapper::new($callback, $event_loop.clone(), $context); + ($self:expr, $target:expr, $callback:expr, $event_loop:expr, $wid:expr, $workers:expr, $srx:expr) => { + // let callback_wrapper = crate::callbacks::CallbackWrapper::new($callback, $event_loop.clone(), $context); + let callback_wrapper = std::sync::Arc::new($callback); let py_loop = std::sync::Arc::new($event_loop.clone().unbind()); for thread_id in 0..$self.config.threads { @@ -807,9 +805,9 @@ macro_rules! serve_wth { ($func_name: ident, $target:expr) => { fn $func_name( &self, - callback: PyObject, + callback: Py<crate::callbacks::CallbackScheduler>, event_loop: &Bound<PyAny>, - context: Bound<PyAny>, + // context: Bound<PyAny>, signal: crate::workers::WorkerSignals, ) { pyo3_log::init(); @@ -819,7 +817,7 @@ macro_rules! serve_wth { let (stx, srx) = tokio::sync::watch::channel(false); let mut workers = vec![]; - crate::workers::serve_wth_inner!(self, $target, callback, event_loop, context, worker_id, workers, srx); + crate::workers::serve_wth_inner!(self, $target, callback, event_loop, worker_id, workers, srx); match signal { crate::workers::WorkerSignals::Tokio(sig) => { @@ -865,8 +863,9 @@ macro_rules! serve_wth { } macro_rules! serve_wth_ssl_inner { - ($self:expr, $target:expr, $callback:expr, $event_loop:expr, $context:expr, $wid:expr, $workers:expr, $srx:expr) => { - let callback_wrapper = crate::callbacks::CallbackWrapper::new($callback, $event_loop.clone(), $context); + ($self:expr, $target:expr, $callback:expr, $event_loop:expr, $wid:expr, $workers:expr, $srx:expr) => { + // let callback_wrapper = crate::callbacks::CallbackWrapper::new($callback, $event_loop.clone(), $context); + let callback_wrapper = std::sync::Arc::new($callback); let py_loop = std::sync::Arc::new($event_loop.clone().unbind()); for thread_id in 0..$self.config.threads { @@ -918,9 +917,9 @@ macro_rules! serve_wth_ssl { ($func_name: ident, $target:expr) => { fn $func_name( &self, - callback: PyObject, + callback: Py<crate::callbacks::CallbackScheduler>, event_loop: &Bound<PyAny>, - context: Bound<PyAny>, + // context: Bound<PyAny>, signal: crate::workers::WorkerSignals, ) { pyo3_log::init(); @@ -930,7 +929,7 @@ macro_rules! serve_wth_ssl { let (stx, srx) = tokio::sync::watch::channel(false); let mut workers = vec![]; - crate::workers::serve_wth_ssl_inner!(self, $target, callback, event_loop, context, worker_id, workers, srx); + crate::workers::serve_wth_ssl_inner!(self, $target, callback, event_loop, worker_id, workers, srx); match signal { crate::workers::WorkerSignals::Tokio(sig) => { diff --git a/src/wsgi/callbacks.rs b/src/wsgi/callbacks.rs index 4a674946..49ca894f 100644 --- a/src/wsgi/callbacks.rs +++ b/src/wsgi/callbacks.rs @@ -9,12 +9,12 @@ use pyo3::{ prelude::*, types::{IntoPyDict, PyBytes, PyDict}, }; -use std::{net::SocketAddr, sync::Arc}; +use std::net::SocketAddr; use tokio::sync::oneshot; use super::{io::WSGIProtocol, types::WSGIBody}; use crate::{ - callbacks::CallbackWrapper, + callbacks::ArcCBScheduler, http::{empty_body, HTTPResponseBody}, runtime::RuntimeRef, utils::log_application_callable_exception, @@ -24,7 +24,7 @@ use crate::{ fn run_callback( rt: RuntimeRef, tx: oneshot::Sender<(u16, HeaderMap, HTTPResponseBody)>, - callback: Arc<PyObject>, + cbs: ArcCBScheduler, mut parts: request::Parts, server_addr: SocketAddr, client_addr: SocketAddr, @@ -66,7 +66,7 @@ fn run_callback( let _ = Python::with_gil(|py| -> PyResult<()> { let proto = Py::new(py, WSGIProtocol::new(tx))?; - let callback = callback.clone_ref(py); + let callback = cbs.get().cb.clone_ref(py); let environ = PyDict::new_bound(py); environ.set_item(pyo3::intern!(py, "SERVER_PROTOCOL"), version)?; environ.set_item(pyo3::intern!(py, "SERVER_NAME"), server.0)?; @@ -108,7 +108,7 @@ fn run_callback( #[inline(always)] pub(crate) fn call_http( rt: RuntimeRef, - cb: CallbackWrapper, + cb: ArcCBScheduler, server_addr: SocketAddr, client_addr: SocketAddr, scheme: &str, @@ -118,7 +118,7 @@ pub(crate) fn call_http( let scheme: std::sync::Arc<str> = scheme.into(); let (tx, rx) = oneshot::channel(); tokio::task::spawn_blocking(move || { - run_callback(rt, tx, cb.callback, req, server_addr, client_addr, &scheme, body); + run_callback(rt, tx, cb, req, server_addr, client_addr, &scheme, body); }); rx } diff --git a/src/wsgi/http.rs b/src/wsgi/http.rs index 030af3bb..287a0e07 100644 --- a/src/wsgi/http.rs +++ b/src/wsgi/http.rs @@ -3,7 +3,7 @@ use std::net::SocketAddr; use super::callbacks::call_http; use crate::{ - callbacks::CallbackWrapper, + callbacks::ArcCBScheduler, http::{response_500, HTTPRequest, HTTPResponse, HTTPResponseBody}, runtime::RuntimeRef, }; @@ -19,7 +19,7 @@ fn build_response(status: u16, pyheaders: hyper::HeaderMap, body: HTTPResponseBo #[inline] pub(crate) async fn handle( rt: RuntimeRef, - callback: CallbackWrapper, + callback: ArcCBScheduler, server_addr: SocketAddr, client_addr: SocketAddr, req: HTTPRequest, diff --git a/src/wsgi/serve.rs b/src/wsgi/serve.rs index 5a295637..f10922ee 100644 --- a/src/wsgi/serve.rs +++ b/src/wsgi/serve.rs @@ -2,6 +2,7 @@ use pyo3::prelude::*; use super::http::handle; +use crate::callbacks::CallbackScheduler; use crate::conversion::{worker_http1_config_from_py, worker_http2_config_from_py}; use crate::workers::{ serve_rth, serve_rth_ssl, serve_wth, serve_wth_ssl, WorkerConfig, WorkerSignalSync, WorkerSignals, @@ -64,7 +65,6 @@ impl WSGIWorker { worker_http1_config_from_py(py, http1_opts)?, worker_http2_config_from_py(py, http2_opts)?, false, - true, ssl_enabled, ssl_cert, ssl_key, @@ -73,29 +73,17 @@ impl WSGIWorker { }) } - fn serve_rth( - &self, - callback: PyObject, - event_loop: &Bound<PyAny>, - context: Bound<PyAny>, - signal: Py<WorkerSignalSync>, - ) { + fn serve_rth(&self, callback: Py<CallbackScheduler>, event_loop: &Bound<PyAny>, signal: Py<WorkerSignalSync>) { match self.config.ssl_enabled { - false => self._serve_rth(callback, event_loop, context, WorkerSignals::Crossbeam(signal)), - true => self._serve_rth_ssl(callback, event_loop, context, WorkerSignals::Crossbeam(signal)), + false => self._serve_rth(callback, event_loop, WorkerSignals::Crossbeam(signal)), + true => self._serve_rth_ssl(callback, event_loop, WorkerSignals::Crossbeam(signal)), } } - fn serve_wth( - &self, - callback: PyObject, - event_loop: &Bound<PyAny>, - context: Bound<PyAny>, - signal: Py<WorkerSignalSync>, - ) { + fn serve_wth(&self, callback: Py<CallbackScheduler>, event_loop: &Bound<PyAny>, signal: Py<WorkerSignalSync>) { match self.config.ssl_enabled { - false => self._serve_wth(callback, event_loop, context, WorkerSignals::Crossbeam(signal)), - true => self._serve_wth_ssl(callback, event_loop, context, WorkerSignals::Crossbeam(signal)), + false => self._serve_wth(callback, event_loop, WorkerSignals::Crossbeam(signal)), + true => self._serve_wth_ssl(callback, event_loop, WorkerSignals::Crossbeam(signal)), } } }

aiohttp / sniffio is unhappy with "--opt" When I use granian with `--opt`, `aiohttp`'s use of [`sniffio`](https://sniffio.readthedocs.io/en/latest/) gets unhappy when trying to detect the async library, without `--opt` it works just fine. How I trigger this is by going to /docs in my `FastAPI` project, which produces an HTTP 500 - Internal Server Error with the below traceback. - Linux amd64 - Python 3.13 with the project managed by `uv` + `pyenv` ```python [ERROR] Application callable raised an exception Traceback (most recent call last): File "./discord_bot/.venv/lib/python3.13/site-packages/granian/asgi.py", line 118, in _http_logger rv = await _runner(scope, proto) File "./discord_bot/.venv/lib/python3.13/site-packages/fastapi/applications.py", line 1054, in __call__ await super().__call__(scope, receive, send) File "./discord_bot/.venv/lib/python3.13/site-packages/starlette/applications.py", line 113, in __call__ await self.middleware_stack(scope, receive, send) File "./discord_bot/.venv/lib/python3.13/site-packages/starlette/middleware/errors.py", line 177, in __call__ response = await self.handler(request, exc) File "./discord_bot/main.py", line 148, in catch_all_handler await async_excepthook(*sys.exc_info()) File "./discord_bot/main.py", line 56, in async_excepthook ryan: discord.User = await discord_client.fetch_user( File "./discord_bot/.venv/lib/python3.13/site-packages/discord/client.py", line 1804, in fetch_user data = await self.http.get_user(user_id) File "./discord_bot/.venv/lib/python3.13/site-packages/discord/http.py", line 286, in request async with self.__session.request( File "./discord_bot/.venv/lib/python3.13/site-packages/aiohttp/client.py", line 1359, in __aenter__ self._resp: _RetType = await self._coro File "./discord_bot/.venv/lib/python3.13/site-packages/aiohttp/client.py", line 579, in _request with timer: File "./discord_bot/.venv/lib/python3.13/site-packages/aiohttp/helpers.py", line 712, in __enter__ raise RuntimeError( RuntimeError: Timeout context manager should be used inside a task 2024-11-30 09:39:34,432 - async_excepthook - ERROR: Exception in AsyncLibraryNotFoundError: unknown async library, or not in async context ```

2024-12-01T13:36:41

0.0

[]

emmett-framework/granian

emmett-framework__granian-296

14f258305cfb824649a26fbc286c19409a69e696

diff --git a/Cargo.lock b/Cargo.lock index 4ff8c8c5..1b8493c3 100644 --- a/Cargo.lock +++ b/Cargo.lock @@ -130,6 +130,12 @@ dependencies = [ "crypto-common", ] +[[package]] +name = "either" +version = "1.12.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "3dca9240753cf90908d7e4aac30f630662b02aebaa1b58a3cadabdb23385b58b" + [[package]] name = "equivalent" version = "1.0.1" @@ -276,6 +282,7 @@ dependencies = [ "http-body-util", "hyper", "hyper-util", + "itertools", "log", "mimalloc", "percent-encoding", @@ -437,6 +444,15 @@ version = "2.0.5" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "b248f5224d1d606005e02c97f5aa4e88eeb230488bcc03bc9ca4d7991399f2b5" +[[package]] +name = "itertools" +version = "0.13.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "413ee7dfc52ee1a4949ceeb7dbc8a33f2d6c088194d9f922fb8318faf1f01186" +dependencies = [ + "either", +] + [[package]] name = "itoa" version = "1.0.11" diff --git a/Cargo.toml b/Cargo.toml index 28f748f7..55b8ca09 100644 --- a/Cargo.toml +++ b/Cargo.toml @@ -36,6 +36,7 @@ futures = "0.3" http-body-util = { version = "=0.1" } hyper = { version = "=1.3", features = ["http1", "http2", "server"] } hyper-util = { version = "=0.1", features = ["server-auto", "tokio"] } +itertools = "0.13" log = "0.4" percent-encoding = "=2.3" pin-project = "1.1" diff --git a/src/wsgi/callbacks.rs b/src/wsgi/callbacks.rs index 216ab809..bfce9676 100644 --- a/src/wsgi/callbacks.rs +++ b/src/wsgi/callbacks.rs @@ -7,6 +7,7 @@ use hyper::{ http::{request, uri::Authority}, Version, }; +use itertools::Itertools; use percent_encoding::percent_decode_str; use pyo3::{ prelude::*, @@ -51,15 +52,20 @@ fn run_callback( let content_type = parts.headers.remove(header::CONTENT_TYPE); let content_len = parts.headers.remove(header::CONTENT_LENGTH); let mut headers = Vec::with_capacity(parts.headers.len()); - for (key, val) in &parts.headers { + for key in parts.headers.keys() { headers.push(( format!("HTTP_{}", key.as_str().replace('-', "_").to_uppercase()), - val.to_str().unwrap_or_default(), + parts + .headers + .get_all(key) + .iter() + .map(|v| v.to_str().unwrap_or_default()) + .join(","), )); } if !parts.headers.contains_key(header::HOST) { let host = parts.uri.authority().map_or("", Authority::as_str); - headers.push(("HTTP_HOST".to_string(), host)); + headers.push(("HTTP_HOST".to_string(), host.to_string())); } Python::with_gil(|py| {

Some cookies are missing on HTTP/2 When running Granian with HTTP/2 some cookies are missing. I created the following test application: ```python def app(environ, start_response): headers = [ ('Content-Type', 'text/html'), ('Set-Cookie', 'csrftoken=abcdef; expires=Mon, 05 May 2025 22:45:47 GMT; HttpOnly; Max-Age=31449600; Path=/; SameSite=Strict; Secure'), ('Set-Cookie', 'sessionid=123456; expires=Mon, 05 May 2025 22:45:47 GMT; HttpOnly; Max-Age=31449600; Path=/; SameSite=Strict; Secure'), ] try: cookies = environ["HTTP_COOKIE"] except KeyError: cookies = "" start_response('200 OK', headers) return [f'<html><body>Cookies: {cookies}</body></html>'.encode()] ``` If I run Granian with HTTP/2 and visis the site with Firefox or Chrome I only see the sessionid cookie, not the csrftoken cookie. With curl it seems to work however. I ran Granian using: ``` granian --http 2 --ssl-keyfile key.pem --ssl-certificate cert.pem --interface wsgi app:app ``` If I run it with HTTP/1 it works without problem however: ``` granian --http 1 --ssl-keyfile key.pem --ssl-certificate cert.pem --interface wsgi app:app ``` It might be caused by https://github.com/hyperium/hyper/issues/2528. I don't know if the Firefox and Chrome split the cookies and curl does not, but if that is the case it would explain why there is no problem with curl but there is a problem with Firefox and Chrome.

@dekkers no, this is a Granian bug on WSGI protocol, thank you for reporting this

2024-05-21T17:35:30

0.0

[]

emmett-framework/granian

emmett-framework__granian-247

bf79551e1ccb7d8ed977574f153990dffa239eae

diff --git a/docs/spec/RSGI.md b/docs/spec/RSGI.md index a33c44c9..ffcf20be 100644 --- a/docs/spec/RSGI.md +++ b/docs/spec/RSGI.md @@ -1,6 +1,6 @@ # RSGI Specification -**Version:** 1.3 +**Version:** 1.4 ## Abstract @@ -92,7 +92,7 @@ And here are descriptions for the upper attributes: - `method`: the HTTP method name, uppercased - `path`: HTTP request target excluding any query string - `query_string`: URL portion after the `?` -- `headers`: a mapping-like object, where keys is the header name, and value is the header value; header names are always lower-case +- `headers`: a mapping-like object, where key is the header name, and value is the header value; header names are always lower-case; a `get_all` method returns a list of all the header values for the given key - `authority`: an optional string containing the relevant pseudo-header (empty on HTTP versions prior to 2) #### HTTP protocol interface @@ -164,7 +164,7 @@ And here are descriptions for the upper attributes: - `method`: the HTTP method name, uppercased - `path`: HTTP request target excluding any query string - `query_string`: URL portion after the `?` -- `headers`: a mapping-like object, where keys is the header name, and value is the header value; header names are always lower-case +- `headers`: a mapping-like object, where key is the header name, and value is the header value; header names are always lower-case; a `get_all` method returns a list of all the header values for the given key - `authority`: an optional string containing the relevant pseudo-header (empty on HTTP versions prior to 2) #### Websocket protocol interface diff --git a/src/rsgi/types.rs b/src/rsgi/types.rs index bdb4c5b9..59c8fc54 100644 --- a/src/rsgi/types.rs +++ b/src/rsgi/types.rs @@ -16,7 +16,7 @@ use tokio_util::io::ReaderStream; use crate::http::{empty_body, response_404, HTTPResponseBody, HV_SERVER}; -const RSGI_PROTO_VERSION: &str = "1.3"; +const RSGI_PROTO_VERSION: &str = "1.4"; #[pyclass(frozen, module = "granian._granian")] #[derive(Clone)] @@ -85,6 +85,18 @@ impl RSGIHeaders { _ => default, } } + + #[pyo3(signature = (key))] + fn get_all<'p>(&self, py: Python<'p>, key: &'p str) -> &'p PyList { + PyList::new( + py, + self.inner + .get_all(key) + .iter() + .map(|v| PyString::new(py, v.to_str().unwrap())) + .collect::<Vec<&PyString>>(), + ) + } } macro_rules! rsgi_scope_cls {

Allow to get all the header values for a single key in `RSGIHeaders`

2024-03-18T22:31:56

0.0

[]

emmett-framework/granian

emmett-framework__granian-189

8b7f9a9b7ecb62ec66ea77b97ed877e063fdf115

diff --git a/docs/spec/RSGI.md b/docs/spec/RSGI.md index 8cac21f2..812c559d 100644 --- a/docs/spec/RSGI.md +++ b/docs/spec/RSGI.md @@ -1,6 +1,6 @@ # RSGI Specification -**Version:** 1.2 +**Version:** 1.3 ## Abstract @@ -79,6 +79,7 @@ class Scope: path: str query_string: str headers: Mapping[str, str] + authority: Optional[str] ``` And here are descriptions for the upper attributes: @@ -92,6 +93,7 @@ And here are descriptions for the upper attributes: - `path`: HTTP request target excluding any query string - `query_string`: URL portion after the `?` - `headers`: a mapping-like object, where keys is the header name, and value is the header value +- `authority`: an optional string containing the relevant pseudo-header (empty on HTTP versions prior to 2) #### HTTP protocol interface @@ -149,6 +151,7 @@ class Scope: path: str query_string: str headers: Mapping[str, str] + authority: Optional[str] ``` And here are descriptions for the upper attributes: @@ -162,6 +165,7 @@ And here are descriptions for the upper attributes: - `path`: HTTP request target excluding any query string - `query_string`: URL portion after the `?` - `headers`: a mapping-like object, where keys is the header name, and value is the header value +- `authority`: an optional string containing the relevant pseudo-header (empty on HTTP versions prior to 2) #### Websocket protocol interface diff --git a/granian/_granian.pyi b/granian/_granian.pyi index 43c9002b..ae1fce4a 100644 --- a/granian/_granian.pyi +++ b/granian/_granian.pyi @@ -24,6 +24,7 @@ class RSGIScope: method: str path: str query_string: str + authority: Optional[str] @property def headers(self) -> RSGIHeaders: ... diff --git a/src/asgi/types.rs b/src/asgi/types.rs index e148af86..7d2a6386 100644 --- a/src/asgi/types.rs +++ b/src/asgi/types.rs @@ -1,4 +1,8 @@ -use hyper::{header::HeaderMap, Uri, Version}; +use hyper::{ + header::{self, HeaderMap}, + http::uri::Authority, + Uri, Version, +}; use percent_encoding::percent_decode_str; use pyo3::{ prelude::*, @@ -105,6 +109,10 @@ impl ASGIScope { PyBytes::new(py, value.as_bytes()), ))?; } + if !self.headers.contains_key(header::HOST) { + let host = self.uri.authority().map_or("", Authority::as_str); + rv.insert(0, (PyBytes::new(py, b"host"), PyBytes::new(py, host.as_bytes())))?; + } Ok(rv) } } diff --git a/src/rsgi/types.rs b/src/rsgi/types.rs index de18e1ac..b6557472 100644 --- a/src/rsgi/types.rs +++ b/src/rsgi/types.rs @@ -1,8 +1,10 @@ +use anyhow::Result; use futures::TryStreamExt; use http_body_util::BodyExt; use hyper::{ body::Bytes, header::{HeaderMap, HeaderName, HeaderValue, SERVER as HK_SERVER}, + http::uri::Authority, Uri, Version, }; use percent_encoding::percent_decode_str; @@ -36,18 +38,18 @@ impl RSGIHeaders { ret } - fn values(&self) -> PyResult<Vec<&str>> { + fn values(&self) -> Result<Vec<&str>> { let mut ret = Vec::with_capacity(self.inner.keys_len()); for val in self.inner.values() { - ret.push(val.to_str().unwrap()); + ret.push(val.to_str()?); } Ok(ret) } - fn items(&self) -> PyResult<Vec<(&str, &str)>> { + fn items(&self) -> Result<Vec<(&str, &str)>> { let mut ret = Vec::with_capacity(self.inner.keys_len()); for (key, val) in &self.inner { - ret.push((key.as_str(), val.to_str().unwrap())); + ret.push((key.as_str(), val.to_str()?)); } Ok(ret) } @@ -102,7 +104,7 @@ impl RSGIScope { Self { proto: proto.to_string(), http_version, - rsgi_version: "1.2".to_string(), + rsgi_version: "1.3".to_string(), scheme: scheme.to_string(), method: method.to_string(), uri, @@ -130,9 +132,14 @@ impl RSGIScope { } } + #[getter(authority)] + fn get_authority(&self) -> Option<String> { + self.uri.authority().map(Authority::to_string) + } + #[getter(path)] - fn get_path(&self) -> Cow<str> { - percent_decode_str(self.uri.path()).decode_utf8().unwrap() + fn get_path(&self) -> Result<Cow<str>> { + Ok(percent_decode_str(self.uri.path()).decode_utf8()?) } #[getter(query_string)] diff --git a/src/wsgi/types.rs b/src/wsgi/types.rs index 15e349a2..544913c0 100644 --- a/src/wsgi/types.rs +++ b/src/wsgi/types.rs @@ -2,7 +2,8 @@ use futures::Stream; use http_body_util::BodyExt; use hyper::{ body::Bytes, - header::{HeaderMap, CONTENT_LENGTH, CONTENT_TYPE}, + header::{HeaderMap, CONTENT_LENGTH, CONTENT_TYPE, HOST}, + http::uri::Authority, Method, Uri, Version, }; use percent_encoding::percent_decode_str; @@ -173,6 +174,10 @@ impl WSGIScope { val.to_str().unwrap_or_default(), )); } + if !self.headers.contains_key(HOST) { + let host = self.uri.authority().map_or("", Authority::as_str); + headers.push(("HTTP_HOST".to_string(), host)); + } ( percent_decode_str(path).decode_utf8().unwrap(),

Missing "host" header in ASGI scope with HTTP/2 Tried granian today with --interface asgi and a starlette application but I'm missing the "host" header in the asgi scope.

Headers are sent by the client, Granian is not removing nor adding anything to request headers. Tested with the following code and can't reproduce your issue: ```python from starlette.applications import Starlette from starlette.responses import PlainTextResponse from starlette.routing import Route def homepage(request): print('scope', request.scope) print('headers', request.headers) print('host', request.headers.get('host')) return PlainTextResponse('Hello, world!') routes = [ Route('/', homepage), ] app = Starlette(debug=True, routes=routes) ``` Request: ``` ❯ curl -v http://localhost:8000 * Trying 127.0.0.1:8000... * Connected to localhost (127.0.0.1) port 8000 (#0) > GET / HTTP/1.1 > Host: localhost:8000 > User-Agent: curl/8.1.2 > Accept: */* > < HTTP/1.1 200 OK < server: granian < content-length: 13 < content-type: text/plain; charset=utf-8 < date: Thu, 25 Jan 2024 09:48:51 GMT < * Connection #0 to host localhost left intact Hello, world! ``` Server logs: ``` ❯ granian --interface asgi t:app [INFO] Starting granian (main PID: 71370) [INFO] Listening at: 127.0.0.1:8000 [INFO] Spawning worker-1 with pid: 71372 [INFO] Started worker-1 [INFO] Started worker-1 runtime-1 scope {'asgi': {'version': '3.0', 'spec_version': '2.3'}, 'extensions': {'http.response.pathsend': {}}, 'type': 'http', 'http_version': '1.1', 'server': ('127.0.0.1', 8000), 'client': ('127.0.0.1', 55309), 'scheme': 'http', 'method': 'GET', 'root_path': '', 'path': '/', 'raw_path': b'/', 'query_string': b'', 'headers': [(b'host', b'localhost:8000'), (b'user-agent', b'curl/8.1.2'), (b'accept', b'*/*')], 'state': {}, 'app': <starlette.applications.Starlette object at 0x1026237d0>, 'starlette.exception_handlers': ({<class 'starlette.exceptions.HTTPException'>: <bound method ExceptionMiddleware.http_exception of <starlette.middleware.exceptions.ExceptionMiddleware object at 0x10273b920>>, <class 'starlette.exceptions.WebSocketException'>: <bound method ExceptionMiddleware.websocket_exception of <starlette.middleware.exceptions.ExceptionMiddleware object at 0x10273b920>>}, {}), 'router': <starlette.routing.Router object at 0x10273b6b0>, 'endpoint': <function homepage at 0x10262ca40>, 'path_params': {}} headers Headers({'host': 'localhost:8000', 'user-agent': 'curl/8.1.2', 'accept': '*/*'}) host localhost:8000 ``` Thanks for your extra fast reply! I did some further investigation, and your example works with http, but as soon as i'm using https, the host header disappears from the scope. > granian --interface asgi --ssl-certificate ssl-cert.pem --ssl-keyfile ssl-key.pem t:app ... scope {'asgi': {'version': '3.0', 'spec_version': '2.3'}, 'extensions': {'http.response.pathsend': {}}, 'type': 'http', 'http_version': '2', 'server': ('127.0.0.1', 8000), 'client': ('127.0.0.1', 21342), 'scheme': 'https', 'method': 'GET', 'root_path': '', 'path': '/', 'raw_path': b'/', 'query_string': b'', 'headers': [(b'user-agent', b'curl/7.76.1'), (b'accept', b'*/*')], 'state': {}, 'app': <starlette.applications.Starlette object at 0x7fd984583bd0>, 'router': <starlette.routing.Router object at 0x7fd984116110>, 'endpoint': <function homepage at 0x7fd984562480>, 'path_params': {}} headers Headers({'user-agent': 'curl/7.76.1', 'accept': '*/*'}) host None Am I doing sth wrong? With hypercorn I get the host header as expected. @kramar11 thank you for the additional details, I'll investigate this asap. @kramar11 after re-checking your comment, I noticed this is HTTP/2. `Host` header is not allowed on HTTP/2, thus is not present in the scope's headers. Now the actual issue here is that Granian doesn't allow access to pseudo-header `:authority` which would contain the relative information. I see Hypercorn automatically translate such pseudo-header content into the `Host` header on HTTP/2, [as per ASGI spec](https://asgi.readthedocs.io/en/latest/specs/www.html#http-connection-scope). Now I don't like this, but now Granian is out of spec for ASGI, so I gonna do the change. But a the same time I need to figure out something for RSGI and WSGI which can incur in the same issue.

2024-01-25T15:35:01

0.0

[]

emmett-framework/granian

emmett-framework__granian-141

9d058bbac55ed1aaf49eac001f6860a387dfe5ac

diff --git a/.github/workflows/build.yml b/.github/workflows/build.yml index c23fade1..5023de4b 100644 --- a/.github/workflows/build.yml +++ b/.github/workflows/build.yml @@ -4,7 +4,7 @@ on: workflow_dispatch env: MATURIN_VERSION: 1.3.0 - PY_ALL: 3.8 3.9 3.10 3.11 pypy3.8 pypy3.9 pypy3.10 + PY_ALL: 3.8 3.9 3.10 3.11 3.12 pypy3.8 pypy3.9 pypy3.10 jobs: wheels: @@ -20,7 +20,7 @@ jobs: - os: ubuntu platform: linux - os: macos - interpreter: 3.8 3.9 3.10 3.11 pypy3.9 pypy3.10 + interpreter: 3.8 3.9 3.10 3.11 3.12 pypy3.9 pypy3.10 - os: ubuntu platform: linux target: aarch64 diff --git a/.github/workflows/lint.yml b/.github/workflows/lint.yml index 08a694ce..64a26d44 100644 --- a/.github/workflows/lint.yml +++ b/.github/workflows/lint.yml @@ -8,7 +8,7 @@ on: env: MATURIN_VERSION: 1.2.3 - PYTHON_VERSION: 3.11 + PYTHON_VERSION: 3.12 jobs: lint: diff --git a/.github/workflows/release.yml b/.github/workflows/release.yml index 91467d8b..767d655a 100644 --- a/.github/workflows/release.yml +++ b/.github/workflows/release.yml @@ -6,7 +6,7 @@ on: env: MATURIN_VERSION: 1.3.0 - PY_ALL: 3.8 3.9 3.10 3.11 pypy3.8 pypy3.9 pypy3.10 + PY_ALL: 3.8 3.9 3.10 3.11 3.12 pypy3.8 pypy3.9 pypy3.10 jobs: sdist: @@ -43,7 +43,7 @@ jobs: - os: ubuntu platform: linux - os: macos - interpreter: 3.8 3.9 3.10 3.11 pypy3.9 pypy3.10 + interpreter: 3.8 3.9 3.10 3.11 3.12 pypy3.9 pypy3.10 - os: ubuntu platform: linux target: aarch64 diff --git a/Makefile b/Makefile index 32b26753..b7e350ad 100644 --- a/Makefile +++ b/Makefile @@ -61,7 +61,7 @@ lint: lint-python lint-rust .PHONY: test test: - pytest -v test + pytest -v tests .PHONY: all all: format build-dev lint test diff --git a/pyproject.toml b/pyproject.toml index fe177f2d..3a8ef2f9 100644 --- a/pyproject.toml +++ b/pyproject.toml @@ -1,7 +1,7 @@ [project] name = 'granian' authors = [ - {name = 'Giovanni Barillari', email = '[email protected]'} + { name = 'Giovanni Barillari', email = '[email protected]' } ] classifiers = [ 'Development Status :: 5 - Production/Stable', @@ -15,11 +15,12 @@ classifiers = [ 'Programming Language :: Python :: 3.9', 'Programming Language :: Python :: 3.10', 'Programming Language :: Python :: 3.11', + 'Programming Language :: Python :: 3.12', 'Programming Language :: Python :: Implementation :: CPython', 'Programming Language :: Python :: Implementation :: PyPy', 'Programming Language :: Python', 'Programming Language :: Rust', - 'Topic :: Internet :: WWW/HTTP' + 'Topic :: Internet :: WWW/HTTP', ] dynamic = [ @@ -27,26 +28,26 @@ dynamic = [ 'keywords', 'license', 'readme', - 'version' + 'version', ] requires-python = '>=3.8' dependencies = [ - 'watchfiles~=0.18', - 'typer~=0.4', - 'uvloop~=0.17.0; sys_platform != "win32" and platform_python_implementation == "CPython"' + 'watchfiles~=0.21', + 'typer~=0.9', + 'uvloop~=0.18.0; sys_platform != "win32" and platform_python_implementation == "CPython"', ] [project.optional-dependencies] lint = [ - 'black~=23.7.0', - 'ruff~=0.0.287' + 'black~=23.9.0', + 'ruff~=0.0.292', ] test = [ - 'httpx~=0.23.0', - 'pytest~=7.1.2', - 'pytest-asyncio~=0.18.3', - 'websockets~=10.3' + 'httpx~=0.25.0', + 'pytest~=7.4.2', + 'pytest-asyncio~=0.21.1', + 'websockets~=11.0', ] [project.urls] @@ -77,7 +78,7 @@ extend-select = [ 'Q', # flake8-quotes 'RUF100', # ruff (unused noqa) 'S', # flake8-bandit - 'W' # pycodestyle + 'W', # pycodestyle ] extend-ignore = [ 'B008', # function calls in args defaults are fine @@ -86,7 +87,7 @@ extend-ignore = [ 'B904', # rising without from is fine 'E501', # leave line length to black 'N818', # leave to us exceptions naming - 'S101' # assert is fine + 'S101', # assert is fine ] flake8-quotes = { inline-quotes = 'single', multiline-quotes = 'double' } mccabe = { max-complexity = 13 } @@ -103,7 +104,7 @@ known-first-party = ['granian', 'tests'] [tool.black] color = true line-length = 120 -target-version = ['py38', 'py39', 'py310', 'py311'] +target-version = ['py38', 'py39', 'py310', 'py311', 'py312'] skip-string-normalization = true # leave this to ruff skip-magic-trailing-comma = true # leave this to ruff

Support for Python 3.12 This depends on: - https://github.com/MagicStack/uvloop/issues/567 - https://github.com/PyO3/pyo3/pull/3493 - related release of pyo3-asyncio (https://github.com/awestlake87/pyo3-asyncio/pull/108) - `watchfiles` package to support 3.12 - PR ref https://github.com/samuelcolvin/watchfiles/pull/248

2023-10-16T10:46:56

0.0

[]

emmett-framework/granian

emmett-framework__granian-124

e28eb81db656fd7577a962a2f394d1ff6fe45986

diff --git a/.github/workflows/lint.yml b/.github/workflows/lint.yml new file mode 100644 index 00000000..41bf0c98 --- /dev/null +++ b/.github/workflows/lint.yml @@ -0,0 +1,32 @@ +name: lint + +on: + pull_request: + types: [opened, synchronize] + branches: + - master + +env: + MATURIN_VERSION: 1.2.3 + PYTHON_VERSION: 3.11 + +jobs: + lint: + runs-on: ubuntu-latest + + steps: + - uses: actions/checkout@v3 + - name: Set up Python ${{ env.PYTHON_VERSION }} + uses: actions/setup-python@v4 + with: + python-version: ${{ env.PYTHON_VERSION }} + - name: Install + run: | + python -m venv .venv + source .venv/bin/activate + pip install maturin==${{ env.MATURIN_VERSION }} + maturin develop --extras=lint + - name: Lint + run: | + source .venv/bin/activate + make lint diff --git a/.pre-commit-config.yaml b/.pre-commit-config.yaml new file mode 100644 index 00000000..883e8a08 --- /dev/null +++ b/.pre-commit-config.yaml @@ -0,0 +1,26 @@ +fail_fast: true + +repos: +- repo: https://github.com/pre-commit/pre-commit-hooks + rev: v4.0.1 + hooks: + - id: check-yaml + - id: check-toml + - id: end-of-file-fixer + - id: trailing-whitespace + - id: check-added-large-files + +- repo: local + hooks: + - id: lint-python + name: Lint Python + entry: make lint-python + types: [python] + language: system + pass_filenames: false + - id: lint-rust + name: Lint Rust + entry: make lint-rust + types: [rust] + language: system + pass_filenames: false diff --git a/.rustfmt.toml b/.rustfmt.toml new file mode 100644 index 00000000..75306517 --- /dev/null +++ b/.rustfmt.toml @@ -0,0 +1,1 @@ +max_width = 120 diff --git a/Makefile b/Makefile new file mode 100644 index 00000000..32b26753 --- /dev/null +++ b/Makefile @@ -0,0 +1,67 @@ +.DEFAULT_GOAL := all +black = black granian tests +ruff = ruff granian tests + +.PHONY: build-dev +build-dev: + @rm -f granian/*.so + maturin develop --extras test + +.PHONY: format +format: + $(black) + $(ruff) --fix --exit-zero + cargo fmt + +.PHONY: lint-python +lint-python: + $(ruff) + $(black) --check --diff + +.PHONY: lint-rust +lint-rust: + cargo fmt --version + cargo fmt --all -- --check + cargo clippy --version + cargo clippy --tests -- \ + -D warnings \ + -W clippy::pedantic \ + -W clippy::dbg_macro \ + -W clippy::print_stdout \ + -A clippy::cast-possible-truncation \ + -A clippy::cast-possible-wrap \ + -A clippy::cast-precision-loss \ + -A clippy::cast-sign-loss \ + -A clippy::declare-interior-mutable-const \ + -A clippy::float-cmp \ + -A clippy::fn-params-excessive-bools \ + -A clippy::if-not-else \ + -A clippy::inline-always \ + -A clippy::manual-let-else \ + -A clippy::match-bool \ + -A clippy::match-same-arms \ + -A clippy::missing-errors-doc \ + -A clippy::missing-panics-doc \ + -A clippy::module-name-repetitions \ + -A clippy::must-use-candidate \ + -A clippy::needless-pass-by-value \ + -A clippy::similar-names \ + -A clippy::single-match-else \ + -A clippy::struct-excessive-bools \ + -A clippy::too-many-arguments \ + -A clippy::too-many-lines \ + -A clippy::type-complexity \ + -A clippy::unnecessary-wraps \ + -A clippy::unused-self \ + -A clippy::used-underscore-binding \ + -A clippy::wrong-self-convention + +.PHONY: lint +lint: lint-python lint-rust + +.PHONY: test +test: + pytest -v test + +.PHONY: all +all: format build-dev lint test diff --git a/granian/__init__.py b/granian/__init__.py index ce03110a..37ebb755 100644 --- a/granian/__init__.py +++ b/granian/__init__.py @@ -1,1 +1,1 @@ -from .server import Granian +from .server import Granian # noqa diff --git a/granian/__main__.py b/granian/__main__.py index c368da26..f1f1f304 100644 --- a/granian/__main__.py +++ b/granian/__main__.py @@ -1,3 +1,4 @@ from granian.cli import cli + cli() diff --git a/granian/__version__.py b/granian/__version__.py index 906d362f..ef7eb44d 100644 --- a/granian/__version__.py +++ b/granian/__version__.py @@ -1,1 +1,1 @@ -__version__ = "0.6.0" +__version__ = '0.6.0' diff --git a/granian/_granian.pyi b/granian/_granian.pyi index 797d95f0..836b1283 100644 --- a/granian/_granian.pyi +++ b/granian/_granian.pyi @@ -1,12 +1,10 @@ -from typing import Any, Dict, List, Tuple, Optional +from typing import Any, Dict, List, Optional, Tuple from ._types import WebsocketMessage - class ASGIScope: def as_dict(self, root_path: str) -> Dict[str, Any]: ... - class RSGIHeaders: def __contains__(self, key: str) -> bool: ... def keys(self) -> List[str]: ... @@ -14,7 +12,6 @@ class RSGIHeaders: def items(self) -> List[Tuple[str]]: ... def get(self, key: str, default: Any = None) -> Any: ... - class RSGIScope: proto: str http_version: str @@ -29,12 +26,10 @@ class RSGIScope: @property def headers(self) -> RSGIHeaders: ... - class RSGIHTTPStreamTransport: async def send_bytes(self, data: bytes): ... async def send_str(self, data: str): ... - class RSGIHTTPProtocol: async def __call__(self) -> bytes: ... def response_empty(self, status: int, headers: List[Tuple[str, str]]): ... @@ -43,25 +38,17 @@ class RSGIHTTPProtocol: def response_file(self, status: int, headers: List[Tuple[str, str]], file: str): ... def response_stream(self, status: int, headers: List[Tuple[str, str]]) -> RSGIHTTPStreamTransport: ... - class RSGIWebsocketTransport: async def receive(self) -> WebsocketMessage: ... async def send_bytes(self, data: bytes): ... async def send_str(self, data: str): ... - class RSGIWebsocketProtocol: async def accept(self) -> RSGIWebsocketTransport: ... def close(self, status: Optional[int]) -> Tuple[int, bool]: ... - -class RSGIProtocolError(RuntimeError): - ... - - -class RSGIProtocolClosed(RuntimeError): - ... - +class RSGIProtocolError(RuntimeError): ... +class RSGIProtocolClosed(RuntimeError): ... class WSGIScope: def to_environ(self, environ: Dict[str, Any]) -> Dict[str, Any]: ... diff --git a/granian/_internal.py b/granian/_internal.py index 26fedbdf..6ebc300a 100644 --- a/granian/_internal.py +++ b/granian/_internal.py @@ -2,22 +2,21 @@ import re import sys import traceback - from types import ModuleType from typing import Callable, List, Optional def get_import_components(path: str) -> List[Optional[str]]: - return (re.split(r":(?![\\/])", path, 1) + [None])[:2] + return (re.split(r':(?![\\/])', path, 1) + [None])[:2] def prepare_import(path: str) -> str: path = os.path.realpath(path) fname, ext = os.path.splitext(path) - if ext == ".py": + if ext == '.py': path = fname - if os.path.basename(path) == "__init__": + if os.path.basename(path) == '__init__': path = os.path.dirname(path) module_name = [] @@ -27,26 +26,22 @@ def prepare_import(path: str) -> str: path, name = os.path.split(path) module_name.append(name) - if not os.path.exists(os.path.join(path, "__init__.py")): + if not os.path.exists(os.path.join(path, '__init__.py')): break if sys.path[0] != path: sys.path.insert(0, path) - return ".".join(module_name[::-1]) + return '.'.join(module_name[::-1]) -def load_module( - module_name: str, - raise_on_failure: bool = True -) -> Optional[ModuleType]: +def load_module(module_name: str, raise_on_failure: bool = True) -> Optional[ModuleType]: try: __import__(module_name) except ImportError: if sys.exc_info()[-1].tb_next: raise RuntimeError( - f"While importing '{module_name}', an ImportError was raised:" - f"\n\n{traceback.format_exc()}" + f"While importing '{module_name}', an ImportError was raised:" f"\n\n{traceback.format_exc()}" ) elif raise_on_failure: raise RuntimeError(f"Could not import '{module_name}'.") @@ -58,9 +53,9 @@ def load_module( def load_target(target: str) -> Callable[..., None]: path, name = get_import_components(target) path = prepare_import(path) if path else None - name = name or "app" + name = name or 'app' module = load_module(path) rv = module - for element in name.split("."): + for element in name.split('.'): rv = getattr(rv, element) return rv diff --git a/granian/_loops.py b/granian/_loops.py index b38f0adc..22145c46 100644 --- a/granian/_loops.py +++ b/granian/_loops.py @@ -2,12 +2,11 @@ import os import signal import sys - from typing import Any, Callable, Dict, Iterable, List, Optional, Tuple class Registry: - __slots__ = ["_data"] + __slots__ = ['_data'] def __init__(self): self._data: Dict[str, Callable[..., Any]] = {} @@ -22,6 +21,7 @@ def register(self, key: str) -> Callable[[], Callable[..., Any]]: def wrap(builder: Callable[..., Any]) -> Callable[..., Any]: self._data[key] = builder return builder + return wrap def get(self, key: str) -> Callable[..., Any]: @@ -31,18 +31,13 @@ def get(self, key: str) -> Callable[..., Any]: raise RuntimeError(f"'{key}' implementation not available.") - class BuilderRegistry(Registry): __slots__ = [] def __init__(self): self._data: Dict[str, Tuple[Callable[..., Any], List[str]]] = {} - def register( - self, - key: str, - packages: Optional[List[str]] = None - ) -> Callable[[], Callable[..., Any]]: + def register(self, key: str, packages: Optional[List[str]] = None) -> Callable[[], Callable[..., Any]]: packages = packages or [] def wrap(builder: Callable[..., Any]) -> Callable[..., Any]: @@ -56,6 +51,7 @@ def wrap(builder: Callable[..., Any]) -> Callable[..., Any]: if implemented: self._data[key] = (builder, loaded_packages) return builder + return wrap def get(self, key: str) -> Callable[..., Any]: diff --git a/granian/asgi.py b/granian/asgi.py index e0b3fd5d..bb955e54 100644 --- a/granian/asgi.py +++ b/granian/asgi.py @@ -1,5 +1,4 @@ import asyncio - from functools import wraps from ._granian import ASGIScope as Scope @@ -22,12 +21,7 @@ def __init__(self, callable): async def handle(self): try: await self.callable( - { - "type": "lifespan", - "asgi": {"version": "3.0", "spec_version": "2.3"} - }, - self.receive, - self.send + {'type': 'lifespan', 'asgi': {'version': '3.0', 'spec_version': '2.3'}}, self.receive, self.send ) except Exception: self.errored = True @@ -43,7 +37,7 @@ async def startup(self): loop = asyncio.get_event_loop() _handler_task = loop.create_task(self.handle()) - await self.event_queue.put({"type": "lifespan.startup"}) + await self.event_queue.put({'type': 'lifespan.startup'}) await self.event_startup.wait() if self.failure_startup or (self.errored and not self.unsupported): @@ -53,7 +47,7 @@ async def shutdown(self): if self.errored: return - await self.event_queue.put({"type": "lifespan.shutdown"}) + await self.event_queue.put({'type': 'lifespan.shutdown'}) await self.event_shutdown.wait() if self.failure_shutdown or (self.errored and not self.unsupported): @@ -89,14 +83,14 @@ def _handle_shutdown_failed(self, message): # self.logger.error(message["message"]) _event_handlers = { - "lifespan.startup.complete": _handle_startup_complete, - "lifespan.startup.failed": _handle_startup_failed, - "lifespan.shutdown.complete": _handle_shutdown_complete, - "lifespan.shutdown.failed": _handle_shutdown_failed + 'lifespan.startup.complete': _handle_startup_complete, + 'lifespan.startup.failed': _handle_startup_failed, + 'lifespan.shutdown.complete': _handle_shutdown_complete, + 'lifespan.shutdown.failed': _handle_shutdown_failed, } async def send(self, message): - handler = self._event_handlers[message["type"]] + handler = self._event_handlers[message['type']] handler(self, message) @@ -108,6 +102,7 @@ def _send_wrapper(proto): @wraps(proto) def send(data): return proto(_noop_coro, data) + return send @@ -116,9 +111,6 @@ def _callback_wrapper(callback, scope_opts): @wraps(callback) def wrapper(scope: Scope, proto): - return callback( - scope.as_dict(root_url_path), - proto.receive, - _send_wrapper(proto.send) - ) + return callback(scope.as_dict(root_url_path), proto.receive, _send_wrapper(proto.send)) + return wrapper diff --git a/granian/cli.py b/granian/cli.py index c37ec96a..0bbbce3b 100644 --- a/granian/cli.py +++ b/granian/cli.py @@ -1,107 +1,55 @@ import json - from pathlib import Path from typing import Optional import typer from .__version__ import __version__ -from .constants import Interfaces, HTTPModes, Loops, ThreadModes +from .constants import HTTPModes, Interfaces, Loops, ThreadModes from .log import LogLevels from .server import Granian -cli = typer.Typer(name="granian", context_settings={"ignore_unknown_options": True}) +cli = typer.Typer(name='granian', context_settings={'ignore_unknown_options': True}) def version_callback(value: bool): if value: - typer.echo(f"{cli.info.name} {__version__}") + typer.echo(f'{cli.info.name} {__version__}') raise typer.Exit() @cli.command() def main( - app: str = typer.Argument(..., help="Application target to serve."), - host: str = typer.Option("127.0.0.1", help="Host address to bind to."), - port: int = typer.Option(8000, help="Port to bind to."), - interface: Interfaces = typer.Option( - Interfaces.RSGI.value, - help="Application interface type." - ), - http: HTTPModes = typer.Option( - HTTPModes.auto.value, - help="HTTP version." - ), - websockets: bool = typer.Option( - True, - "--ws/--no-ws", - help="Enable websockets handling", - show_default="enabled" - ), - workers: int = typer.Option(1, min=1, help="Number of worker processes."), - threads: int = typer.Option(1, min=1, help="Number of threads."), - threading_mode: ThreadModes = typer.Option( - ThreadModes.workers.value, - help="Threading mode to use." - ), - loop: Loops = typer.Option(Loops.auto.value, help="Event loop implementation"), - loop_opt: bool = typer.Option( - False, - "--opt/--no-opt", - help="Enable loop optimizations", - show_default="disabled" - ), - backlog: int = typer.Option( - 1024, - min=128, - help="Maximum number of connections to hold in backlog." - ), - log_level: LogLevels = typer.Option( - LogLevels.info.value, - help="Log level", - case_sensitive=False - ), + app: str = typer.Argument(..., help='Application target to serve.'), + host: str = typer.Option('127.0.0.1', help='Host address to bind to.'), + port: int = typer.Option(8000, help='Port to bind to.'), + interface: Interfaces = typer.Option(Interfaces.RSGI.value, help='Application interface type.'), + http: HTTPModes = typer.Option(HTTPModes.auto.value, help='HTTP version.'), + websockets: bool = typer.Option(True, '--ws/--no-ws', help='Enable websockets handling', show_default='enabled'), + workers: int = typer.Option(1, min=1, help='Number of worker processes.'), + threads: int = typer.Option(1, min=1, help='Number of threads.'), + threading_mode: ThreadModes = typer.Option(ThreadModes.workers.value, help='Threading mode to use.'), + loop: Loops = typer.Option(Loops.auto.value, help='Event loop implementation'), + loop_opt: bool = typer.Option(False, '--opt/--no-opt', help='Enable loop optimizations', show_default='disabled'), + backlog: int = typer.Option(1024, min=128, help='Maximum number of connections to hold in backlog.'), + log_level: LogLevels = typer.Option(LogLevels.info.value, help='Log level', case_sensitive=False), log_config: Optional[Path] = typer.Option( - None, - help="Logging configuration file (json)", - exists=True, - file_okay=True, - dir_okay=False, - readable=True + None, help='Logging configuration file (json)', exists=True, file_okay=True, dir_okay=False, readable=True ), ssl_keyfile: Optional[Path] = typer.Option( - None, - help="SSL key file", - exists=True, - file_okay=True, - dir_okay=False, - readable=True + None, help='SSL key file', exists=True, file_okay=True, dir_okay=False, readable=True ), ssl_certificate: Optional[Path] = typer.Option( - None, - help="SSL certificate file", - exists=True, - file_okay=True, - dir_okay=False, - readable=True - ), - url_path_prefix: Optional[str] = typer.Option( - None, - help="URL path prefix the app is mounted on" + None, help='SSL certificate file', exists=True, file_okay=True, dir_okay=False, readable=True ), + url_path_prefix: Optional[str] = typer.Option(None, help='URL path prefix the app is mounted on'), reload: bool = typer.Option( - False, - "--reload/--no-reload", - help="Enable auto reload on application's files changes" + False, '--reload/--no-reload', help="Enable auto reload on application's files changes" ), _: Optional[bool] = typer.Option( - None, - "--version", - callback=version_callback, - is_eager=True, - help="Shows the version and exit." - ) + None, '--version', callback=version_callback, is_eager=True, help='Shows the version and exit.' + ), ): log_dictconfig = None if log_config: @@ -109,7 +57,7 @@ def main( try: log_dictconfig = json.loads(log_config_file.read()) except Exception: - print("Unable to parse provided logging config.") + print('Unable to parse provided logging config.') raise typer.Exit(1) Granian( @@ -131,5 +79,5 @@ def main( ssl_cert=ssl_certificate, ssl_key=ssl_keyfile, url_path_prefix=url_path_prefix, - reload=reload + reload=reload, ).serve() diff --git a/granian/constants.py b/granian/constants.py index 1579afea..f0b2100d 100644 --- a/granian/constants.py +++ b/granian/constants.py @@ -2,23 +2,23 @@ class Interfaces(str, Enum): - ASGI = "asgi" - RSGI = "rsgi" - WSGI = "wsgi" + ASGI = 'asgi' + RSGI = 'rsgi' + WSGI = 'wsgi' class HTTPModes(str, Enum): - auto = "auto" - http1 = "1" - http2 = "2" + auto = 'auto' + http1 = '1' + http2 = '2' class ThreadModes(str, Enum): - runtime = "runtime" - workers = "workers" + runtime = 'runtime' + workers = 'workers' class Loops(str, Enum): - auto = "auto" - asyncio = "asyncio" - uvloop = "uvloop" + auto = 'auto' + asyncio = 'asyncio' + uvloop = 'uvloop' diff --git a/granian/log.py b/granian/log.py index 378e9a4b..df21b69f 100644 --- a/granian/log.py +++ b/granian/log.py @@ -1,18 +1,17 @@ import copy import logging import logging.config - from enum import Enum from typing import Any, Dict, Optional class LogLevels(str, Enum): - critical = "critical" - error = "error" - warning = "warning" - warn = "warn" - info = "info" - debug = "debug" + critical = 'critical' + error = 'error' + warning = 'warning' + warn = 'warn' + info = 'info' + debug = 'debug' log_levels_map = { @@ -21,27 +20,21 @@ class LogLevels(str, Enum): LogLevels.warning: logging.WARNING, LogLevels.warn: logging.WARN, LogLevels.info: logging.INFO, - LogLevels.debug: logging.DEBUG + LogLevels.debug: logging.DEBUG, } LOGGING_CONFIG = { - "version": 1, - "disable_existing_loggers": False, - "root": {"level": "INFO", "handlers": ["console"]}, - "formatters": { - "generic": { - "()": "logging.Formatter", - "fmt": "[%(levelname)s] %(message)s", - "datefmt": "[%Y-%m-%d %H:%M:%S %z]" + 'version': 1, + 'disable_existing_loggers': False, + 'root': {'level': 'INFO', 'handlers': ['console']}, + 'formatters': { + 'generic': { + '()': 'logging.Formatter', + 'fmt': '[%(levelname)s] %(message)s', + 'datefmt': '[%Y-%m-%d %H:%M:%S %z]', } }, - "handlers": { - "console": { - "formatter": "generic", - "class": "logging.StreamHandler", - "stream": "ext://sys.stdout", - } - } + 'handlers': {'console': {'formatter': 'generic', 'class': 'logging.StreamHandler', 'stream': 'ext://sys.stdout'}}, } logger = logging.getLogger() @@ -51,5 +44,5 @@ def configure_logging(level: LogLevels, config: Optional[Dict[str, Any]] = None) log_config = copy.deepcopy(LOGGING_CONFIG) if config: log_config.update(config) - log_config["root"]["level"] = log_levels_map[level] + log_config['root']['level'] = log_levels_map[level] logging.config.dictConfig(log_config) diff --git a/granian/net.py b/granian/net.py index 26a60d3f..d0b2bc5f 100644 --- a/granian/net.py +++ b/granian/net.py @@ -2,7 +2,5 @@ from ._granian import ListenerHolder as SocketHolder -copyreg.pickle( - SocketHolder, - lambda v: (SocketHolder, v.__getstate__()) -) + +copyreg.pickle(SocketHolder, lambda v: (SocketHolder, v.__getstate__())) diff --git a/granian/rsgi.py b/granian/rsgi.py index 64555e01..4f9ecbba 100644 --- a/granian/rsgi.py +++ b/granian/rsgi.py @@ -2,12 +2,12 @@ from typing import Union from ._granian import ( - RSGIHTTPProtocol as HTTPProtocol, - RSGIWebsocketProtocol as WebsocketProtocol, - RSGIHeaders as Headers, - RSGIScope as Scope, - RSGIProtocolError as ProtocolError, - RSGIProtocolClosed as ProtocolClosed + RSGIHeaders as Headers, # noqa + RSGIHTTPProtocol as HTTPProtocol, # noqa + RSGIProtocolClosed as ProtocolClosed, # noqa + RSGIProtocolError as ProtocolError, # noqa + RSGIScope as Scope, # noqa + RSGIWebsocketProtocol as WebsocketProtocol, # noqa ) diff --git a/granian/server.py b/granian/server.py index a9800571..f0f610e9 100644 --- a/granian/server.py +++ b/granian/server.py @@ -5,7 +5,6 @@ import ssl import sys import threading - from functools import partial from pathlib import Path from typing import Any, Callable, Dict, List, Optional @@ -16,11 +15,12 @@ from ._granian import ASGIWorker, RSGIWorker, WSGIWorker from ._internal import load_target from .asgi import LifespanProtocol, _callback_wrapper as _asgi_call_wrap -from .constants import Interfaces, HTTPModes, Loops, ThreadModes +from .constants import HTTPModes, Interfaces, Loops, ThreadModes from .log import LogLevels, configure_logging, logger from .net import SocketHolder from .wsgi import _callback_wrapper as _wsgi_call_wrap + multiprocessing.allow_connection_pickling() @@ -30,7 +30,7 @@ class Granian: def __init__( self, target: str, - address: str = "127.0.0.1", + address: str = '127.0.0.1', port: int = 8000, interface: Interfaces = Interfaces.RSGI, workers: int = 1, @@ -48,7 +48,7 @@ def __init__( ssl_cert: Optional[Path] = None, ssl_key: Optional[Path] = None, url_path_prefix: Optional[str] = None, - reload: bool = False + reload: bool = False, ): self.target = target self.bind_addr = address @@ -75,11 +75,7 @@ def __init__( self.procs: List[multiprocessing.Process] = [] self.exit_event = threading.Event() - def build_ssl_context( - self, - cert: Optional[Path], - key: Optional[Path] - ): + def build_ssl_context(self, cert: Optional[Path], key: Optional[Path]): if not (cert and key): self.ssl_ctx = (False, None, None) return @@ -108,7 +104,7 @@ def _spawn_asgi_worker( log_level, log_config, ssl_ctx, - scope_opts + scope_opts, ): from granian._loops import loops, set_loop_signals @@ -129,29 +125,12 @@ def _spawn_asgi_worker( wcallback = future_watcher_wrapper(wcallback) worker = ASGIWorker( - worker_id, - sfd, - threads, - pthreads, - http_mode, - http1_buffer_size, - websockets, - loop_opt, - *ssl_ctx - ) - serve = getattr(worker, { - ThreadModes.runtime: "serve_rth", - ThreadModes.workers: "serve_wth" - }[threading_mode]) - serve( - wcallback, - loop, - contextvars.copy_context(), - shutdown_event.wait() + worker_id, sfd, threads, pthreads, http_mode, http1_buffer_size, websockets, loop_opt, *ssl_ctx ) + serve = getattr(worker, {ThreadModes.runtime: 'serve_rth', ThreadModes.workers: 'serve_wth'}[threading_mode]) + serve(wcallback, loop, contextvars.copy_context(), shutdown_event.wait()) loop.run_until_complete(lifespan_handler.shutdown()) - @staticmethod def _spawn_rsgi_worker( worker_id, @@ -168,7 +147,7 @@ def _spawn_rsgi_worker( log_level, log_config, ssl_ctx, - scope_opts + scope_opts, ): from granian._loops import loops, set_loop_signals @@ -176,38 +155,23 @@ def _spawn_rsgi_worker( loop = loops.get(loop_impl) sfd = socket.fileno() target = callback_loader() - callback = ( - getattr(target, '__rsgi__') if hasattr(target, '__rsgi__') else - target - ) + callback = getattr(target, '__rsgi__') if hasattr(target, '__rsgi__') else target callback_init = ( - getattr(target, '__rsgi_init__') if hasattr(target, '__rsgi_init__') else - lambda *args, **kwargs: None + getattr(target, '__rsgi_init__') if hasattr(target, '__rsgi_init__') else lambda *args, **kwargs: None ) shutdown_event = set_loop_signals(loop, [signal.SIGTERM, signal.SIGINT]) callback_init(loop) worker = RSGIWorker( - worker_id, - sfd, - threads, - pthreads, - http_mode, - http1_buffer_size, - websockets, - loop_opt, - *ssl_ctx + worker_id, sfd, threads, pthreads, http_mode, http1_buffer_size, websockets, loop_opt, *ssl_ctx ) - serve = getattr(worker, { - ThreadModes.runtime: "serve_rth", - ThreadModes.workers: "serve_wth" - }[threading_mode]) + serve = getattr(worker, {ThreadModes.runtime: 'serve_rth', ThreadModes.workers: 'serve_wth'}[threading_mode]) serve( future_watcher_wrapper(callback) if not loop_opt else callback, loop, contextvars.copy_context(), - shutdown_event.wait() + shutdown_event.wait(), ) @staticmethod @@ -226,7 +190,7 @@ def _spawn_wsgi_worker( log_level, log_config, ssl_ctx, - scope_opts + scope_opts, ): from granian._loops import loops, set_loop_signals @@ -237,46 +201,20 @@ def _spawn_wsgi_worker( shutdown_event = set_loop_signals(loop, [signal.SIGTERM, signal.SIGINT]) - worker = WSGIWorker( - worker_id, - sfd, - threads, - pthreads, - http_mode, - http1_buffer_size, - *ssl_ctx - ) - serve = getattr(worker, { - ThreadModes.runtime: "serve_rth", - ThreadModes.workers: "serve_wth" - }[threading_mode]) - serve( - _wsgi_call_wrap(callback, scope_opts), - loop, - contextvars.copy_context(), - shutdown_event.wait() - ) + worker = WSGIWorker(worker_id, sfd, threads, pthreads, http_mode, http1_buffer_size, *ssl_ctx) + serve = getattr(worker, {ThreadModes.runtime: 'serve_rth', ThreadModes.workers: 'serve_wth'}[threading_mode]) + serve(_wsgi_call_wrap(callback, scope_opts), loop, contextvars.copy_context(), shutdown_event.wait()) def _init_shared_socket(self): - self._shd = SocketHolder.from_address( - self.bind_addr, - self.bind_port, - self.backlog - ) + self._shd = SocketHolder.from_address(self.bind_addr, self.bind_port, self.backlog) self._sfd = self._shd.get_fd() def signal_handler(self, *args, **kwargs): self.exit_event.set() - def _spawn_proc( - self, - id, - target, - callback_loader, - socket_loader - ) -> multiprocessing.Process: + def _spawn_proc(self, id, target, callback_loader, socket_loader) -> multiprocessing.Process: return multiprocessing.get_context().Process( - name="granian-worker", + name='granian-worker', target=target, args=( id, @@ -293,10 +231,8 @@ def _spawn_proc( self.log_level, self.log_config, self.ssl_ctx, - { - "url_path_prefix": self.url_path_prefix - } - ) + {'url_path_prefix': self.url_path_prefix}, + ), ) def _spawn_workers(self, sock, spawn_target, target_loader): @@ -305,14 +241,11 @@ def socket_loader(): for idx in range(self.workers): proc = self._spawn_proc( - id=idx + 1, - target=spawn_target, - callback_loader=target_loader, - socket_loader=socket_loader + id=idx + 1, target=spawn_target, callback_loader=target_loader, socket_loader=socket_loader ) proc.start() self.procs.append(proc) - logger.info(f"Spawning worker-{idx + 1} with pid: {proc.pid}") + logger.info(f'Spawning worker-{idx + 1} with pid: {proc.pid}') def _stop_workers(self): for proc in self.procs: @@ -321,7 +254,7 @@ def _stop_workers(self): proc.join() def startup(self, spawn_target, target_loader): - logger.info("Starting granian") + logger.info('Starting granian') for sig in self.SIGNALS: signal.signal(sig, self.signal_handler) @@ -329,13 +262,13 @@ def startup(self, spawn_target, target_loader): self._init_shared_socket() sock = socket.socket(fileno=self._sfd) sock.set_inheritable(True) - logger.info(f"Listening at: {self.bind_addr}:{self.bind_port}") + logger.info(f'Listening at: {self.bind_addr}:{self.bind_port}') self._spawn_workers(sock, spawn_target, target_loader) return sock def shutdown(self): - logger.info("Shutting down granian") + logger.info('Shutting down granian') self._stop_workers() def _serve(self, spawn_target, target_loader): @@ -360,12 +293,12 @@ def serve( self, spawn_target: Optional[Callable[..., None]] = None, target_loader: Optional[Callable[..., Callable[..., Any]]] = None, - wrap_loader: bool = True + wrap_loader: bool = True, ): default_spawners = { Interfaces.ASGI: self._spawn_asgi_worker, Interfaces.RSGI: self._spawn_rsgi_worker, - Interfaces.WSGI: self._spawn_wsgi_worker + Interfaces.WSGI: self._spawn_wsgi_worker, } if target_loader: if wrap_loader: @@ -383,8 +316,5 @@ def serve( "Number of workers will now fallback to 1." ) - serve_method = ( - self._serve_with_reloader if self.reload_on_changes else - self._serve - ) + serve_method = self._serve_with_reloader if self.reload_on_changes else self._serve serve_method(spawn_target, target_loader) diff --git a/granian/wsgi.py b/granian/wsgi.py index 6ba53db4..6b73eaed 100644 --- a/granian/wsgi.py +++ b/granian/wsgi.py @@ -1,6 +1,5 @@ import os import sys - from functools import wraps from typing import Any, List, Tuple @@ -14,30 +13,27 @@ def __init__(self): self.status = 200 self.headers = [] - def __call__( - self, - status: str, - headers: List[Tuple[str, str]], - exc_info: Any = None - ): + def __call__(self, status: str, headers: List[Tuple[str, str]], exc_info: Any = None): self.status = int(status.split(' ', 1)[0]) self.headers = headers def _callback_wrapper(callback, scope_opts): basic_env = dict(os.environ) - basic_env.update({ - 'GATEWAY_INTERFACE': 'CGI/1.1', - 'SCRIPT_NAME': scope_opts.get('url_path_prefix') or '', - 'SERVER_SOFTWARE': 'Granian', - 'wsgi.errors': sys.stderr, - 'wsgi.input_terminated': True, - 'wsgi.input': None, - 'wsgi.multiprocess': False, - 'wsgi.multithread': False, - 'wsgi.run_once': False, - 'wsgi.version': (1, 0) - }) + basic_env.update( + { + 'GATEWAY_INTERFACE': 'CGI/1.1', + 'SCRIPT_NAME': scope_opts.get('url_path_prefix') or '', + 'SERVER_SOFTWARE': 'Granian', + 'wsgi.errors': sys.stderr, + 'wsgi.input_terminated': True, + 'wsgi.input': None, + 'wsgi.multiprocess': False, + 'wsgi.multithread': False, + 'wsgi.run_once': False, + 'wsgi.version': (1, 0), + } + ) @wraps(callback) def wrapper(scope: Scope) -> Tuple[int, List[Tuple[str, str]], bytes]: @@ -46,7 +42,7 @@ def wrapper(scope: Scope) -> Tuple[int, List[Tuple[str, str]], bytes]: if isinstance(rv, list): resp_type = 0 - rv = b"".join(rv) + rv = b''.join(rv) else: resp_type = 1 rv = iter(rv) diff --git a/pyproject.toml b/pyproject.toml index d2a79e0b..0f4d07b9 100644 --- a/pyproject.toml +++ b/pyproject.toml @@ -1,65 +1,111 @@ [project] -name = "granian" +name = 'granian' authors = [ - {name = "Giovanni Barillari", email = "[email protected]"} + {name = 'Giovanni Barillari', email = '[email protected]'} ] classifiers = [ - "Development Status :: 5 - Production/Stable", - "Intended Audience :: Developers", - "License :: OSI Approved :: BSD License", - "Operating System :: MacOS", - "Operating System :: Microsoft :: Windows", - "Operating System :: POSIX :: Linux", - "Programming Language :: Python :: 3", - "Programming Language :: Python :: 3.8", - "Programming Language :: Python :: 3.9", - "Programming Language :: Python :: 3.10", - "Programming Language :: Python :: 3.11", - "Programming Language :: Python :: Implementation :: CPython", - "Programming Language :: Python :: Implementation :: PyPy", - "Programming Language :: Python", - "Programming Language :: Rust", - "Topic :: Internet :: WWW/HTTP" + 'Development Status :: 5 - Production/Stable', + 'Intended Audience :: Developers', + 'License :: OSI Approved :: BSD License', + 'Operating System :: MacOS', + 'Operating System :: Microsoft :: Windows', + 'Operating System :: POSIX :: Linux', + 'Programming Language :: Python :: 3', + 'Programming Language :: Python :: 3.8', + 'Programming Language :: Python :: 3.9', + 'Programming Language :: Python :: 3.10', + 'Programming Language :: Python :: 3.11', + 'Programming Language :: Python :: Implementation :: CPython', + 'Programming Language :: Python :: Implementation :: PyPy', + 'Programming Language :: Python', + 'Programming Language :: Rust', + 'Topic :: Internet :: WWW/HTTP' ] dynamic = [ - "description", - "keywords", - "license", - "readme", - "version" + 'description', + 'keywords', + 'license', + 'readme', + 'version' ] -requires-python = ">=3.8" +requires-python = '>=3.8' dependencies = [ - "watchfiles~=0.18", - "typer~=0.4", - "uvloop~=0.17.0; sys_platform != 'win32' and platform_python_implementation == 'CPython'" + 'watchfiles~=0.18', + 'typer~=0.4', + 'uvloop~=0.17.0; sys_platform != "win32" and platform_python_implementation == "CPython"' ] [project.optional-dependencies] +lint = [ + 'black~=23.7.0', + 'ruff~=0.0.287' +] test = [ - "httpx~=0.23.0", - "pytest~=7.1.2", - "pytest-asyncio~=0.18.3", - "websockets~=10.3" + 'httpx~=0.23.0', + 'pytest~=7.1.2', + 'pytest-asyncio~=0.18.3', + 'websockets~=10.3' ] [project.urls] -Homepage = "https://github.com/emmett-framework/granian" -Funding = "https://github.com/sponsors/gi0baro" -Source = "https://github.com/emmett-framework/granian" +Homepage = 'https://github.com/emmett-framework/granian' +Funding = 'https://github.com/sponsors/gi0baro' +Source = 'https://github.com/emmett-framework/granian' [project.scripts] -granian = "granian:cli.cli" +granian = 'granian:cli.cli' [build-system] -requires = ["maturin>=1.1.0,<1.3.0"] -build-backend = "maturin" +requires = ['maturin>=1.1.0,<1.3.0'] +build-backend = 'maturin' [tool.maturin] -module-name = "granian._granian" -bindings = "pyo3" +module-name = 'granian._granian' +bindings = 'pyo3' + +[tool.ruff] +line-length = 120 +extend-select = [ + # E and F are enabled by default + 'B', # flake8-bugbear + 'C4', # flake8-comprehensions + 'C90', # mccabe + 'I', # isort + 'N', # pep8-naming + 'Q', # flake8-quotes + 'RUF100', # ruff (unused noqa) + 'S', # flake8-bandit + 'W' # pycodestyle +] +extend-ignore = [ + 'B008', # function calls in args defaults are fine + 'B009', # getattr with constants is fine + 'B034', # re.split won't confuse us + 'B904', # rising without from is fine + 'E501', # leave line length to black + 'N818', # leave to us exceptions naming + 'S101' # assert is fine +] +flake8-quotes = { inline-quotes = 'single', multiline-quotes = 'double' } +mccabe = { max-complexity = 13 } + +[tool.ruff.isort] +combine-as-imports = true +lines-after-imports = 2 +known-first-party = ['granian', 'tests'] + +[tool.ruff.per-file-ignores] +'granian/_granian.pyi' = ['I001'] +'tests/**' = ['B018', 'S110', 'S501'] + +[tool.black] +color = true +line-length = 120 +target-version = ['py38', 'py39', 'py310', 'py311'] +skip-string-normalization = true # leave this to ruff +skip-magic-trailing-comma = true # leave this to ruff [tool.pytest.ini_options] -asyncio_mode = "auto" +asyncio_mode = 'auto' diff --git a/src/asgi/callbacks.rs b/src/asgi/callbacks.rs index 5af7766f..96c100fb 100644 --- a/src/asgi/callbacks.rs +++ b/src/asgi/callbacks.rs @@ -3,45 +3,33 @@ use pyo3::prelude::*; use pyo3_asyncio::TaskLocals; use tokio::sync::oneshot; +use super::{ + io::{ASGIHTTPProtocol as HTTPProtocol, ASGIWebsocketProtocol as WebsocketProtocol}, + types::ASGIScope as Scope, +}; use crate::{ callbacks::{ - CallbackWrapper, - callback_impl_run, - callback_impl_run_pytask, - callback_impl_loop_run, - callback_impl_loop_pytask, - callback_impl_loop_step, - callback_impl_loop_wake, - callback_impl_loop_err + callback_impl_loop_err, callback_impl_loop_pytask, callback_impl_loop_run, callback_impl_loop_step, + callback_impl_loop_wake, callback_impl_run, callback_impl_run_pytask, CallbackWrapper, }, runtime::RuntimeRef, - ws::{HyperWebsocket, UpgradeData} + ws::{HyperWebsocket, UpgradeData}, }; -use super::{ - io::{ASGIHTTPProtocol as HTTPProtocol, ASGIWebsocketProtocol as WebsocketProtocol}, - types::ASGIScope as Scope -}; - #[pyclass] pub(crate) struct CallbackRunnerHTTP { proto: Py<HTTPProtocol>, context: TaskLocals, - cb: PyObject + cb: PyObject, } impl CallbackRunnerHTTP { - pub fn new( - py: Python, - cb: CallbackWrapper, - proto: HTTPProtocol, - scope: Scope - ) -> Self { + pub fn new(py: Python, cb: CallbackWrapper, proto: HTTPProtocol, scope: Scope) -> Self { let pyproto = Py::new(py, proto).unwrap(); Self { proto: pyproto.clone(), context: cb.context, - cb: cb.callback.call1(py, (scope, pyproto)).unwrap() + cb: cb.callback.call1(py, (scope, pyproto)).unwrap(), } } @@ -64,14 +52,14 @@ macro_rules! callback_impl_done_http { let _ = tx.send(res); } } - } + }; } macro_rules! callback_impl_done_err { ($self:expr, $py:expr) => { log::warn!("Application callable raised an exception"); $self.done($py) - } + }; } #[pyclass] @@ -79,22 +67,17 @@ pub(crate) struct CallbackTaskHTTP { proto: Py<HTTPProtocol>, context: TaskLocals, pycontext: PyObject, - cb: PyObject + cb: PyObject, } impl CallbackTaskHTTP { - pub fn new( - py: Python, - cb: PyObject, - proto: Py<HTTPProtocol>, - context: TaskLocals - ) -> PyResult<Self> { + pub fn new(py: Python, cb: PyObject, proto: Py<HTTPProtocol>, context: TaskLocals) -> PyResult<Self> { let pyctx = context.context(py); Ok(Self { proto, context, pycontext: pyctx.call_method0(pyo3::intern!(py, "copy"))?.into(), - cb + cb, }) } @@ -128,21 +111,16 @@ pub(crate) struct CallbackWrappedRunnerHTTP { context: TaskLocals, cb: PyObject, #[pyo3(get)] - scope: PyObject + scope: PyObject, } impl CallbackWrappedRunnerHTTP { - pub fn new( - py: Python, - cb: CallbackWrapper, - proto: HTTPProtocol, - scope: Scope - ) -> Self { + pub fn new(py: Python, cb: CallbackWrapper, proto: HTTPProtocol, scope: Scope) -> Self { Self { proto: Py::new(py, proto).unwrap(), context: cb.context, cb: cb.callback, - scope: scope.into_py(py) + scope: scope.into_py(py), } } @@ -168,21 +146,16 @@ impl CallbackWrappedRunnerHTTP { pub(crate) struct CallbackRunnerWebsocket { proto: Py<WebsocketProtocol>, context: TaskLocals, - cb: PyObject + cb: PyObject, } impl CallbackRunnerWebsocket { - pub fn new( - py: Python, - cb: CallbackWrapper, - proto: WebsocketProtocol, - scope: Scope - ) -> Self { + pub fn new(py: Python, cb: CallbackWrapper, proto: WebsocketProtocol, scope: Scope) -> Self { let pyproto = Py::new(py, proto).unwrap(); Self { proto: pyproto.clone(), context: cb.context, - cb: cb.callback.call1(py, (scope, pyproto)).unwrap() + cb: cb.callback.call1(py, (scope, pyproto)).unwrap(), } } @@ -203,7 +176,7 @@ macro_rules! callback_impl_done_ws { let _ = tx.send(res); } } - } + }; } #[pyclass] @@ -211,22 +184,17 @@ pub(crate) struct CallbackTaskWebsocket { proto: Py<WebsocketProtocol>, context: TaskLocals, pycontext: PyObject, - cb: PyObject + cb: PyObject, } impl CallbackTaskWebsocket { - pub fn new( - py: Python, - cb: PyObject, - proto: Py<WebsocketProtocol>, - context: TaskLocals - ) -> PyResult<Self> { + pub fn new(py: Python, cb: PyObject, proto: Py<WebsocketProtocol>, context: TaskLocals) -> PyResult<Self> { let pyctx = context.context(py); Ok(Self { proto, context, pycontext: pyctx.call_method0(pyo3::intern!(py, "copy"))?.into(), - cb + cb, }) } @@ -260,21 +228,16 @@ pub(crate) struct CallbackWrappedRunnerWebsocket { context: TaskLocals, cb: PyObject, #[pyo3(get)] - scope: PyObject + scope: PyObject, } impl CallbackWrappedRunnerWebsocket { - pub fn new( - py: Python, - cb: CallbackWrapper, - proto: WebsocketProtocol, - scope: Scope - ) -> Self { + pub fn new(py: Python, cb: CallbackWrapper, proto: WebsocketProtocol, scope: Scope) -> Self { Self { proto: Py::new(py, proto).unwrap(), context: cb.context, cb: cb.callback, - scope: scope.into_py(py) + scope: scope.into_py(py), } } @@ -325,7 +288,7 @@ macro_rules! call_impl_rtb_http { cb: CallbackWrapper, rt: RuntimeRef, req: Request<Body>, - scope: Scope + scope: Scope, ) -> oneshot::Receiver<Response<Body>> { let (tx, rx) = oneshot::channel(); let protocol = HTTPProtocol::new(rt, req, tx); @@ -345,7 +308,7 @@ macro_rules! call_impl_rtt_http { cb: CallbackWrapper, rt: RuntimeRef, req: Request<Body>, - scope: Scope + scope: Scope, ) -> oneshot::Receiver<Response<Body>> { let (tx, rx) = oneshot::channel(); let protocol = HTTPProtocol::new(rt, req, tx); @@ -368,7 +331,7 @@ macro_rules! call_impl_rtb_ws { rt: RuntimeRef, ws: HyperWebsocket, upgrade: UpgradeData, - scope: Scope + scope: Scope, ) -> oneshot::Receiver<bool> { let (tx, rx) = oneshot::channel(); let protocol = WebsocketProtocol::new(rt, tx, ws, upgrade); @@ -389,7 +352,7 @@ macro_rules! call_impl_rtt_ws { rt: RuntimeRef, ws: HyperWebsocket, upgrade: UpgradeData, - scope: Scope + scope: Scope, ) -> oneshot::Receiver<bool> { let (tx, rx) = oneshot::channel(); let protocol = WebsocketProtocol::new(rt, tx, ws, upgrade); diff --git a/src/asgi/errors.rs b/src/asgi/errors.rs index 6b788f60..8a6dda40 100644 --- a/src/asgi/errors.rs +++ b/src/asgi/errors.rs @@ -88,5 +88,5 @@ macro_rules! error_message { } pub(crate) use error_flow; -pub(crate) use error_transport; pub(crate) use error_message; +pub(crate) use error_transport; diff --git a/src/asgi/http.rs b/src/asgi/http.rs index cbf295f7..1fa62773 100644 --- a/src/asgi/http.rs +++ b/src/asgi/http.rs @@ -1,34 +1,22 @@ use hyper::{ - Body, - Request, - Response, - StatusCode, - header::SERVER as HK_SERVER, - http::response::Builder as ResponseBuilder + header::SERVER as HK_SERVER, http::response::Builder as ResponseBuilder, Body, Request, Response, StatusCode, }; use std::net::SocketAddr; use tokio::sync::mpsc; -use crate::{ - callbacks::CallbackWrapper, - http::{HV_SERVER, response_500}, - runtime::RuntimeRef, - ws::{UpgradeData, is_upgrade_request as is_ws_upgrade, upgrade_intent as ws_upgrade} -}; use super::{ callbacks::{ - call_rtb_http, - call_rtb_http_pyw, - call_rtb_ws, - call_rtb_ws_pyw, - call_rtt_http, - call_rtt_http_pyw, - call_rtt_ws, - call_rtt_ws_pyw + call_rtb_http, call_rtb_http_pyw, call_rtb_ws, call_rtb_ws_pyw, call_rtt_http, call_rtt_http_pyw, call_rtt_ws, + call_rtt_ws_pyw, }, - types::ASGIScope as Scope + types::ASGIScope as Scope, +}; +use crate::{ + callbacks::CallbackWrapper, + http::{response_500, HV_SERVER}, + runtime::RuntimeRef, + ws::{is_upgrade_request as is_ws_upgrade, upgrade_intent as ws_upgrade, UpgradeData}, }; - macro_rules! default_scope { ($server_addr:expr, $client_addr:expr, $req:expr, $scheme:expr) => { @@ -39,7 +27,7 @@ macro_rules! default_scope { $req.method().as_ref(), $server_addr, $client_addr, - $req.headers() + $req.headers(), ) }; } @@ -53,7 +41,7 @@ macro_rules! handle_http_response { response_500() } } - } + }; } macro_rules! handle_request { @@ -64,7 +52,7 @@ macro_rules! handle_request { server_addr: SocketAddr, client_addr: SocketAddr, req: Request<Body>, - scheme: &str + scheme: &str, ) -> Response<Body> { let scope = default_scope!(server_addr, client_addr, &req, scheme); handle_http_response!($handler, rt, callback, req, scope) @@ -80,7 +68,7 @@ macro_rules! handle_request_with_ws { server_addr: SocketAddr, client_addr: SocketAddr, req: Request<Body>, - scheme: &str + scheme: &str, ) -> Response<Body> { let mut scope = default_scope!(server_addr, client_addr, &req, scheme); @@ -95,24 +83,20 @@ macro_rules! handle_request_with_ws { rt.inner.spawn(async move { let tx_ref = restx.clone(); - match $handler_ws( - callback, - rth, - ws, - UpgradeData::new(res, restx), - scope - ).await { + match $handler_ws(callback, rth, ws, UpgradeData::new(res, restx), scope).await { Ok(consumed) => { if !consumed { - let _ = tx_ref.send( - ResponseBuilder::new() - .status(StatusCode::FORBIDDEN) - .header(HK_SERVER, HV_SERVER) - .body(Body::from("")) - .unwrap() - ).await; + let _ = tx_ref + .send( + ResponseBuilder::new() + .status(StatusCode::FORBIDDEN) + .header(HK_SERVER, HV_SERVER) + .body(Body::from("")) + .unwrap(), + ) + .await; }; - }, + } _ => { log::error!("ASGI protocol failure"); let _ = tx_ref.send(response_500()).await; @@ -124,10 +108,10 @@ macro_rules! handle_request_with_ws { Some(res) => { resrx.close(); res - }, - _ => response_500() + } + _ => response_500(), } - }, + } Err(err) => { return ResponseBuilder::new() .status(StatusCode::BAD_REQUEST) diff --git a/src/asgi/io.rs b/src/asgi/io.rs index 4b2f25eb..d206aaf2 100644 --- a/src/asgi/io.rs +++ b/src/asgi/io.rs @@ -1,33 +1,34 @@ use bytes::Bytes; -use futures::{sink::SinkExt, stream::{SplitSink, SplitStream, StreamExt}}; +use futures::{ + sink::SinkExt, + stream::{SplitSink, SplitStream, StreamExt}, +}; use hyper::{ - Request, - Response, body::{Body, HttpBody, Sender as BodySender}, - header::{HeaderName, HeaderValue, HeaderMap, SERVER as HK_SERVER} + header::{HeaderMap, HeaderName, HeaderValue, SERVER as HK_SERVER}, + Request, Response, }; use pyo3::prelude::*; use pyo3::types::{PyBytes, PyDict}; use std::sync::Arc; +use tokio::sync::{oneshot, Mutex}; use tokio_tungstenite::WebSocketStream; -use tokio::sync::{Mutex, oneshot}; use tungstenite::Message; +use super::{ + errors::{error_flow, error_message, error_transport, UnsupportedASGIMessage}, + types::ASGIMessageType, +}; use crate::{ http::HV_SERVER, - runtime::{RuntimeRef, future_into_py_iter, future_into_py_futlike}, - ws::{HyperWebsocket, UpgradeData} + runtime::{future_into_py_futlike, future_into_py_iter, RuntimeRef}, + ws::{HyperWebsocket, UpgradeData}, }; -use super::{ - errors::{UnsupportedASGIMessage, error_flow, error_transport, error_message}, - types::ASGIMessageType -}; - const EMPTY_BYTES: Vec<u8> = Vec::new(); const EMPTY_STRING: String = String::new(); -#[pyclass(module="granian._granian")] +#[pyclass(module = "granian._granian")] pub(crate) struct ASGIHTTPProtocol { rt: RuntimeRef, tx: Option<oneshot::Sender<Response<Body>>>, @@ -36,15 +37,11 @@ pub(crate) struct ASGIHTTPProtocol { response_chunked: bool, response_status: Option<i16>, response_headers: Option<HeaderMap>, - body_tx: Option<Arc<Mutex<BodySender>>> + body_tx: Option<Arc<Mutex<BodySender>>>, } impl ASGIHTTPProtocol { - pub fn new( - rt: RuntimeRef, - request: Request<Body>, - tx: oneshot::Sender<Response<Body>> - ) -> Self { + pub fn new(rt: RuntimeRef, request: Request<Body>, tx: oneshot::Sender<Response<Body>>) -> Self { Self { rt, tx: Some(tx), @@ -53,7 +50,7 @@ impl ASGIHTTPProtocol { response_chunked: false, response_status: None, response_headers: None, - body_tx: None + body_tx: None, } } @@ -71,7 +68,7 @@ impl ASGIHTTPProtocol { fn send_body<'p>(&self, py: Python<'p>, tx: Arc<Mutex<BodySender>>, body: Vec<u8>) -> PyResult<&'p PyAny> { future_into_py_futlike(self.rt.clone(), py, async move { let mut tx = tx.lock().await; - match (&mut *tx).send_data(body.into()).await { + match (*tx).send_data(body.into()).await { Ok(_) => Ok(()), Err(err) => { log::warn!("ASGI transport tx error: {:?}", err); @@ -94,18 +91,18 @@ impl ASGIHTTPProtocol { let mut bodym = body_ref.lock().await; let body = &mut *bodym; let mut more_body = false; - let chunk = body.data().await.map_or_else(|| Bytes::new(), |buf| { - buf.map_or_else(|_| Bytes::new(), |buf| { - more_body = !body.is_end_stream(); - buf - }) + let chunk = body.data().await.map_or_else(Bytes::new, |buf| { + buf.map_or_else( + |_| Bytes::new(), + |buf| { + more_body = !body.is_end_stream(); + buf + }, + ) }); Python::with_gil(|py| { let dict = PyDict::new(py); - dict.set_item( - pyo3::intern!(py, "type"), - pyo3::intern!(py, "http.request") - )?; + dict.set_item(pyo3::intern!(py, "type"), pyo3::intern!(py, "http.request"))?; dict.set_item(pyo3::intern!(py, "body"), PyBytes::new(py, &chunk[..]))?; dict.set_item(pyo3::intern!(py, "more_body"), more_body)?; Ok(dict.to_object(py)) @@ -115,16 +112,14 @@ impl ASGIHTTPProtocol { fn send<'p>(&mut self, py: Python<'p>, asyncw: &'p PyAny, data: &'p PyDict) -> PyResult<&'p PyAny> { match adapt_message_type(data) { - Ok(ASGIMessageType::HTTPStart) => { - match self.response_started { - false => { - self.response_status = Some(adapt_status_code(data)?); - self.response_headers = Some(adapt_headers(data)); - self.response_started = true; - asyncw.call0() - }, - true => error_flow!() + Ok(ASGIMessageType::HTTPStart) => match self.response_started { + false => { + self.response_status = Some(adapt_status_code(data)?); + self.response_headers = Some(adapt_headers(data)); + self.response_started = true; + asyncw.call0() } + true => error_flow!(), }, Ok(ASGIMessageType::HTTPBody) => { let (body, more) = adapt_body(data); @@ -133,7 +128,7 @@ impl ASGIHTTPProtocol { let headers = self.response_headers.take().unwrap(); self.send_response(self.response_status.unwrap(), headers, body.into()); asyncw.call0() - }, + } (true, true, false) => { self.response_chunked = true; let headers = self.response_headers.take().unwrap(); @@ -142,37 +137,31 @@ impl ASGIHTTPProtocol { self.body_tx = Some(tx.clone()); self.send_response(self.response_status.unwrap(), headers, body_stream); self.send_body(py, tx, body) - }, - (true, true, true) => { - match self.body_tx.as_mut() { - Some(tx) => { - let tx = tx.clone(); - self.send_body(py, tx, body) - }, - _ => error_flow!() + } + (true, true, true) => match self.body_tx.as_mut() { + Some(tx) => { + let tx = tx.clone(); + self.send_body(py, tx, body) } + _ => error_flow!(), }, - (true, false, true) => { - match self.body_tx.take() { - Some(tx) => { - match body.is_empty() { - false => self.send_body(py, tx, body), - true => asyncw.call0() - } - }, - _ => error_flow!() - } + (true, false, true) => match self.body_tx.take() { + Some(tx) => match body.is_empty() { + false => self.send_body(py, tx, body), + true => asyncw.call0(), + }, + _ => error_flow!(), }, - _ => error_flow!() + _ => error_flow!(), } - }, + } Err(err) => Err(err.into()), - _ => error_message!() + _ => error_message!(), } } } -#[pyclass(module="granian._granian")] +#[pyclass(module = "granian._granian")] pub(crate) struct ASGIWebsocketProtocol { rt: RuntimeRef, tx: Option<oneshot::Sender<bool>>, @@ -181,16 +170,11 @@ pub(crate) struct ASGIWebsocketProtocol { ws_tx: Arc<Mutex<Option<SplitSink<WebSocketStream<hyper::upgrade::Upgraded>, Message>>>>, ws_rx: Arc<Mutex<Option<SplitStream<WebSocketStream<hyper::upgrade::Upgraded>>>>>, accepted: Arc<Mutex<bool>>, - closed: bool + closed: bool, } impl ASGIWebsocketProtocol { - pub fn new( - rt: RuntimeRef, - tx: oneshot::Sender<bool>, - websocket: HyperWebsocket, - upgrade: UpgradeData - ) -> Self { + pub fn new(rt: RuntimeRef, tx: oneshot::Sender<bool>, websocket: HyperWebsocket, upgrade: UpgradeData) -> Self { Self { rt, tx: Some(tx), @@ -199,7 +183,7 @@ impl ASGIWebsocketProtocol { ws_tx: Arc::new(Mutex::new(None)), ws_rx: Arc::new(Mutex::new(None)), accepted: Arc::new(Mutex::new(false)), - closed: false + closed: false, } } @@ -211,7 +195,7 @@ impl ASGIWebsocketProtocol { let tx = self.ws_tx.clone(); let rx = self.ws_rx.clone(); future_into_py_iter(self.rt.clone(), py, async move { - if let Ok(_) = upgrade.send().await { + if (upgrade.send().await).is_ok() { if let Ok(stream) = websocket.await { let mut wtx = tx.lock().await; let mut wrx = rx.lock().await; @@ -220,7 +204,7 @@ impl ASGIWebsocketProtocol { *wtx = Some(tx); *wrx = Some(rx); *accepted = true; - return Ok(()) + return Ok(()); } } error_flow!() @@ -228,18 +212,14 @@ impl ASGIWebsocketProtocol { } #[inline(always)] - fn send_message<'p>( - &self, - py: Python<'p>, - data: &'p PyDict - ) -> PyResult<&'p PyAny> { + fn send_message<'p>(&self, py: Python<'p>, data: &'p PyDict) -> PyResult<&'p PyAny> { let transport = self.ws_tx.clone(); let message = ws_message_into_rs(data); future_into_py_iter(self.rt.clone(), py, async move { if let Ok(message) = message { if let Some(ws) = &mut *(transport.lock().await) { - if let Ok(_) = ws.send(message).await { - return Ok(()) + if (ws.send(message).await).is_ok() { + return Ok(()); } }; }; @@ -253,8 +233,8 @@ impl ASGIWebsocketProtocol { let transport = self.ws_tx.clone(); future_into_py_iter(self.rt.clone(), py, async move { if let Some(ws) = &mut *(transport.lock().await) { - if let Ok(_) = ws.close().await { - return Ok(()) + if (ws.close().await).is_ok() { + return Ok(()); } }; error_flow!() @@ -262,10 +242,7 @@ impl ASGIWebsocketProtocol { } fn consumed(&self) -> bool { - match &self.upgrade { - Some(_) => false, - _ => true - } + self.upgrade.is_none() } pub fn tx(&mut self) -> (Option<oneshot::Sender<bool>>, bool) { @@ -278,44 +255,26 @@ impl ASGIWebsocketProtocol { fn receive<'p>(&mut self, py: Python<'p>) -> PyResult<&'p PyAny> { let transport = self.ws_rx.clone(); let accepted = self.accepted.clone(); - let closed = self.closed.clone(); + let closed = self.closed; future_into_py_futlike(self.rt.clone(), py, async move { let accepted = accepted.lock().await; match (*accepted, closed) { (false, false) => { return Python::with_gil(|py| { let dict = PyDict::new(py); - dict.set_item( - pyo3::intern!(py, "type"), - pyo3::intern!(py, "websocket.connect") - )?; + dict.set_item(pyo3::intern!(py, "type"), pyo3::intern!(py, "websocket.connect"))?; Ok(dict.to_object(py)) }) - }, - (true, false) => {}, - _ => { - return error_flow!() } + (true, false) => {} + _ => return error_flow!(), } if let Some(ws) = &mut *(transport.lock().await) { - loop { - match ws.next().await { - Some(recv) => { - match recv { - Ok(Message::Ping(_)) => { - continue - }, - Ok(message) => { - return ws_message_into_py(message) - }, - _ => { - break - } - } - }, - _ => { - break - } + while let Some(recv) = ws.next().await { + match recv { + Ok(Message::Ping(_)) => continue, + Ok(message) => return ws_message_into_py(message), + _ => break, } } } @@ -325,25 +284,17 @@ impl ASGIWebsocketProtocol { fn send<'p>(&mut self, py: Python<'p>, _asyncw: &'p PyAny, data: &'p PyDict) -> PyResult<&'p PyAny> { match (adapt_message_type(data), self.closed) { - (Ok(ASGIMessageType::WSAccept), _) => { - self.accept(py) - }, - (Ok(ASGIMessageType::WSClose), false) => { - self.close(py) - }, - (Ok(ASGIMessageType::WSMessage), false) => { - self.send_message(py, data) - }, + (Ok(ASGIMessageType::WSAccept), _) => self.accept(py), + (Ok(ASGIMessageType::WSClose), false) => self.close(py), + (Ok(ASGIMessageType::WSMessage), false) => self.send_message(py, data), (Err(err), _) => Err(err.into()), - _ => error_message!() + _ => error_message!(), } } } #[inline(never)] -fn adapt_message_type( - message: &PyDict -) -> Result<ASGIMessageType, UnsupportedASGIMessage> { +fn adapt_message_type(message: &PyDict) -> Result<ASGIMessageType, UnsupportedASGIMessage> { match message.get_item("type") { Some(item) => { let message_type: &str = item.extract()?; @@ -353,20 +304,18 @@ fn adapt_message_type( "websocket.accept" => Ok(ASGIMessageType::WSAccept), "websocket.close" => Ok(ASGIMessageType::WSClose), "websocket.send" => Ok(ASGIMessageType::WSMessage), - _ => error_message!() + _ => error_message!(), } - }, - _ => error_message!() + } + _ => error_message!(), } } #[inline(always)] fn adapt_status_code(message: &PyDict) -> Result<i16, UnsupportedASGIMessage> { match message.get_item("status") { - Some(item) => { - Ok(item.extract()?) - }, - _ => error_message!() + Some(item) => Ok(item.extract()?), + _ => error_message!(), } } @@ -377,34 +326,26 @@ fn adapt_headers(message: &PyDict) -> HeaderMap { match message.get_item("headers") { Some(item) => { let accum: Vec<Vec<&[u8]>> = item.extract().unwrap_or(Vec::new()); - for tup in accum.iter() { - match ( - HeaderName::from_bytes(tup[0]), - HeaderValue::from_bytes(tup[1]) - ) { - (Ok(key), Ok(val)) => { ret.append(key, val); }, - _ => {} + for tup in &accum { + if let (Ok(key), Ok(val)) = (HeaderName::from_bytes(tup[0]), HeaderValue::from_bytes(tup[1])) { + ret.append(key, val); } - }; + } ret - }, - _ => ret + } + _ => ret, } } #[inline(always)] fn adapt_body(message: &PyDict) -> (Vec<u8>, bool) { let body = match message.get_item("body") { - Some(item) => { - item.extract().unwrap_or(EMPTY_BYTES) - }, - _ => EMPTY_BYTES + Some(item) => item.extract().unwrap_or(EMPTY_BYTES), + _ => EMPTY_BYTES, }; let more = match message.get_item("more_body") { - Some(item) => { - item.extract().unwrap_or(false) - }, - _ => false + Some(item) => item.extract().unwrap_or(false), + _ => false, }; (body, more) } @@ -412,22 +353,12 @@ fn adapt_body(message: &PyDict) -> (Vec<u8>, bool) { #[inline(always)] fn ws_message_into_rs(message: &PyDict) -> PyResult<Message> { match (message.get_item("bytes"), message.get_item("text")) { - (Some(item), None) => { - Ok(Message::Binary(item.extract().unwrap_or(EMPTY_BYTES))) - }, - (None, Some(item)) => { - Ok(Message::Text(item.extract().unwrap_or(EMPTY_STRING))) - }, - (Some(itemb), Some(itemt)) => { - match (itemb.extract().unwrap_or(None), itemt.extract().unwrap_or(None)) { - (Some(msgb), None) => { - Ok(Message::Binary(msgb)) - }, - (None, Some(msgt)) => { - Ok(Message::Text(msgt)) - }, - _ => error_flow!() - } + (Some(item), None) => Ok(Message::Binary(item.extract().unwrap_or(EMPTY_BYTES))), + (None, Some(item)) => Ok(Message::Text(item.extract().unwrap_or(EMPTY_STRING))), + (Some(itemb), Some(itemt)) => match (itemb.extract().unwrap_or(None), itemt.extract().unwrap_or(None)) { + (Some(msgb), None) => Ok(Message::Binary(msgb)), + (None, Some(msgt)) => Ok(Message::Text(msgt)), + _ => error_flow!(), }, _ => { error_flow!() @@ -438,41 +369,23 @@ fn ws_message_into_rs(message: &PyDict) -> PyResult<Message> { #[inline(always)] fn ws_message_into_py(message: Message) -> PyResult<PyObject> { match message { - Message::Binary(message) => { - Python::with_gil(|py| { - let dict = PyDict::new(py); - dict.set_item( - pyo3::intern!(py, "type"), - pyo3::intern!(py, "websocket.receive") - )?; - dict.set_item( - pyo3::intern!(py, "bytes"), - PyBytes::new(py, &message[..]) - )?; - Ok(dict.to_object(py)) - }) - }, - Message::Text(message) => { - Python::with_gil(|py| { - let dict = PyDict::new(py); - dict.set_item( - pyo3::intern!(py, "type"), - pyo3::intern!(py, "websocket.receive") - )?; - dict.set_item(pyo3::intern!(py, "text"), message)?; - Ok(dict.to_object(py)) - }) - }, - Message::Close(_) => { - Python::with_gil(|py| { - let dict = PyDict::new(py); - dict.set_item( - pyo3::intern!(py, "type"), - pyo3::intern!(py, "websocket.disconnect") - )?; - Ok(dict.to_object(py)) - }) - }, + Message::Binary(message) => Python::with_gil(|py| { + let dict = PyDict::new(py); + dict.set_item(pyo3::intern!(py, "type"), pyo3::intern!(py, "websocket.receive"))?; + dict.set_item(pyo3::intern!(py, "bytes"), PyBytes::new(py, &message[..]))?; + Ok(dict.to_object(py)) + }), + Message::Text(message) => Python::with_gil(|py| { + let dict = PyDict::new(py); + dict.set_item(pyo3::intern!(py, "type"), pyo3::intern!(py, "websocket.receive"))?; + dict.set_item(pyo3::intern!(py, "text"), message)?; + Ok(dict.to_object(py)) + }), + Message::Close(_) => Python::with_gil(|py| { + let dict = PyDict::new(py); + dict.set_item(pyo3::intern!(py, "type"), pyo3::intern!(py, "websocket.disconnect"))?; + Ok(dict.to_object(py)) + }), v => { log::warn!("Unsupported websocket message received {:?}", v); error_flow!() diff --git a/src/asgi/serve.rs b/src/asgi/serve.rs index a30f15c3..6243cdc6 100644 --- a/src/asgi/serve.rs +++ b/src/asgi/serve.rs @@ -1,29 +1,14 @@ use pyo3::prelude::*; -use crate::{ - workers::{ - WorkerConfig, - serve_rth, - serve_wth, - serve_rth_ssl, - serve_wth_ssl - } -}; use super::http::{ - handle_rtb, - handle_rtb_pyw, - handle_rtt, - handle_rtt_pyw, - handle_rtb_ws, - handle_rtb_ws_pyw, - handle_rtt_ws, - handle_rtt_ws_pyw + handle_rtb, handle_rtb_pyw, handle_rtb_ws, handle_rtb_ws_pyw, handle_rtt, handle_rtt_pyw, handle_rtt_ws, + handle_rtt_ws_pyw, }; +use crate::workers::{serve_rth, serve_rth_ssl, serve_wth, serve_wth_ssl, WorkerConfig}; - -#[pyclass(module="granian._granian")] +#[pyclass(module = "granian._granian")] pub struct ASGIWorker { - config: WorkerConfig + config: WorkerConfig, } impl ASGIWorker { @@ -74,7 +59,7 @@ impl ASGIWorker { opt_enabled: bool, ssl_enabled: bool, ssl_cert: Option<&str>, - ssl_key: Option<&str> + ssl_key: Option<&str>, ) -> PyResult<Self> { Ok(Self { config: WorkerConfig::new( @@ -88,22 +73,16 @@ impl ASGIWorker { opt_enabled, ssl_enabled, ssl_cert, - ssl_key - ) + ssl_key, + ), }) } - fn serve_rth( - &self, - callback: PyObject, - event_loop: &PyAny, - context: &PyAny, - signal_rx: PyObject - ) { + fn serve_rth(&self, callback: PyObject, event_loop: &PyAny, context: &PyAny, signal_rx: PyObject) { match ( self.config.websockets_enabled, self.config.ssl_enabled, - self.config.opt_enabled + self.config.opt_enabled, ) { (false, false, true) => self._serve_rth(callback, event_loop, context, signal_rx), (false, false, false) => self._serve_rth_pyw(callback, event_loop, context, signal_rx), @@ -112,21 +91,15 @@ impl ASGIWorker { (false, true, true) => self._serve_rth_ssl(callback, event_loop, context, signal_rx), (false, true, false) => self._serve_rth_ssl_pyw(callback, event_loop, context, signal_rx), (true, true, true) => self._serve_rth_ssl_ws(callback, event_loop, context, signal_rx), - (true, true, false) => self._serve_rth_ssl_ws_pyw(callback, event_loop, context, signal_rx) + (true, true, false) => self._serve_rth_ssl_ws_pyw(callback, event_loop, context, signal_rx), } } - fn serve_wth( - &self, - callback: PyObject, - event_loop: &PyAny, - context: &PyAny, - signal_rx: PyObject - ) { + fn serve_wth(&self, callback: PyObject, event_loop: &PyAny, context: &PyAny, signal_rx: PyObject) { match ( self.config.websockets_enabled, self.config.ssl_enabled, - self.config.opt_enabled + self.config.opt_enabled, ) { (false, false, true) => self._serve_wth(callback, event_loop, context, signal_rx), (false, false, false) => self._serve_wth_pyw(callback, event_loop, context, signal_rx), @@ -135,7 +108,7 @@ impl ASGIWorker { (false, true, true) => self._serve_wth_ssl(callback, event_loop, context, signal_rx), (false, true, false) => self._serve_wth_ssl_pyw(callback, event_loop, context, signal_rx), (true, true, true) => self._serve_wth_ssl_ws(callback, event_loop, context, signal_rx), - (true, true, false) => self._serve_wth_ssl_ws_pyw(callback, event_loop, context, signal_rx) + (true, true, false) => self._serve_wth_ssl_ws_pyw(callback, event_loop, context, signal_rx), } } } diff --git a/src/asgi/types.rs b/src/asgi/types.rs index 1841d0e6..8e1056e2 100644 --- a/src/asgi/types.rs +++ b/src/asgi/types.rs @@ -1,10 +1,9 @@ -use hyper::{Uri, Version, header::HeaderMap}; +use hyper::{header::HeaderMap, Uri, Version}; use once_cell::sync::OnceCell; use pyo3::prelude::*; use pyo3::types::{PyBytes, PyDict, PyList, PyString}; use std::net::{IpAddr, SocketAddr}; - const SCHEME_HTTPS: &str = "https"; const SCHEME_WS: &str = "ws"; const SCHEME_WSS: &str = "wss"; @@ -17,10 +16,10 @@ pub(crate) enum ASGIMessageType { HTTPBody, WSAccept, WSClose, - WSMessage + WSMessage, } -#[pyclass(module="granian._granian")] +#[pyclass(module = "granian._granian")] pub(crate) struct ASGIScope { http_version: Version, scheme: String, @@ -31,7 +30,7 @@ pub(crate) struct ASGIScope { client_ip: IpAddr, client_port: u16, headers: HeaderMap, - is_websocket: bool + is_websocket: bool, } impl ASGIScope { @@ -42,31 +41,31 @@ impl ASGIScope { method: &str, server: SocketAddr, client: SocketAddr, - headers: &HeaderMap + headers: &HeaderMap, ) -> Self { Self { - http_version: http_version, + http_version, scheme: scheme.to_string(), method: method.to_string(), - uri: uri, + uri, server_ip: server.ip(), server_port: server.port(), client_ip: client.ip(), client_port: client.port(), - headers: headers.to_owned(), - is_websocket: false + headers: headers.clone(), + is_websocket: false, } } pub fn set_websocket(&mut self) { - self.is_websocket = true + self.is_websocket = true; } #[inline(always)] fn py_proto(&self) -> &str { match self.is_websocket { false => "http", - true => "websocket" + true => "websocket", } } @@ -76,7 +75,7 @@ impl ASGIScope { Version::HTTP_10 => "1", Version::HTTP_11 => "1.1", Version::HTTP_2 => "2", - _ => "1" + _ => "1", } } @@ -85,22 +84,20 @@ impl ASGIScope { let scheme = &self.scheme[..]; match self.is_websocket { false => scheme, - true => { - match scheme { - SCHEME_HTTPS => SCHEME_WSS, - _ => SCHEME_WS - } - } + true => match scheme { + SCHEME_HTTPS => SCHEME_WSS, + _ => SCHEME_WS, + }, } } #[inline(always)] fn py_headers<'p>(&self, py: Python<'p>) -> PyResult<&'p PyList> { let rv = PyList::empty(py); - for (key, value) in self.headers.iter() { + for (key, value) in &self.headers { rv.append(( PyBytes::new(py, key.as_str().as_bytes()), - PyBytes::new(py, value.as_bytes()) + PyBytes::new(py, value.as_bytes()), ))?; } Ok(rv) @@ -110,17 +107,10 @@ impl ASGIScope { #[pymethods] impl ASGIScope { fn as_dict<'p>(&self, py: Python<'p>, url_path_prefix: &'p str) -> PyResult<&'p PyAny> { - let ( - path, - query_string, - proto, - http_version, - server, - client, - scheme, - method - ) = py.allow_threads(|| { - let (path, query_string) = self.uri.path_and_query() + let (path, query_string, proto, http_version, server, client, scheme, method) = py.allow_threads(|| { + let (path, query_string) = self + .uri + .path_and_query() .map_or_else(|| ("", ""), |pq| (pq.path(), pq.query().unwrap_or(""))); ( path, @@ -136,19 +126,23 @@ impl ASGIScope { let dict: &PyDict = PyDict::new(py); dict.set_item( pyo3::intern!(py, "asgi"), - ASGI_VERSION.get_or_try_init(|| { - let rv = PyDict::new(py); - rv.set_item("version", "3.0")?; - rv.set_item("spec_version", "2.3")?; - Ok::<PyObject, PyErr>(rv.into()) - })?.as_ref(py) + ASGI_VERSION + .get_or_try_init(|| { + let rv = PyDict::new(py); + rv.set_item("version", "3.0")?; + rv.set_item("spec_version", "2.3")?; + Ok::<PyObject, PyErr>(rv.into()) + })? + .as_ref(py), )?; dict.set_item( pyo3::intern!(py, "extensions"), - ASGI_EXTENSIONS.get_or_try_init(|| { - let rv = PyDict::new(py); - Ok::<PyObject, PyErr>(rv.into()) - })?.as_ref(py) + ASGI_EXTENSIONS + .get_or_try_init(|| { + let rv = PyDict::new(py); + Ok::<PyObject, PyErr>(rv.into()) + })? + .as_ref(py), )?; dict.set_item(pyo3::intern!(py, "type"), proto)?; dict.set_item(pyo3::intern!(py, "http_version"), http_version)?; @@ -160,17 +154,12 @@ impl ASGIScope { dict.set_item(pyo3::intern!(py, "path"), path)?; dict.set_item( pyo3::intern!(py, "raw_path"), - PyString::new(py, path) - .call_method1( - pyo3::intern!(py, "encode"), (pyo3::intern!(py, "ascii"),) - )? + PyString::new(py, path).call_method1(pyo3::intern!(py, "encode"), (pyo3::intern!(py, "ascii"),))?, )?; dict.set_item( pyo3::intern!(py, "query_string"), PyString::new(py, query_string) - .call_method1( - pyo3::intern!(py, "encode"), (pyo3::intern!(py, "latin-1"),) - )? + .call_method1(pyo3::intern!(py, "encode"), (pyo3::intern!(py, "latin-1"),))?, )?; dict.set_item(pyo3::intern!(py, "headers"), self.py_headers(py)?)?; Ok(dict) diff --git a/src/callbacks.rs b/src/callbacks.rs index c6a67ef1..9af137b1 100644 --- a/src/callbacks.rs +++ b/src/callbacks.rs @@ -2,32 +2,27 @@ use once_cell::sync::OnceCell; use pyo3::prelude::*; use pyo3::pyclass::IterNextOutput; - static CONTEXTVARS: OnceCell<PyObject> = OnceCell::new(); static CONTEXT: OnceCell<PyObject> = OnceCell::new(); #[derive(Clone)] pub(crate) struct CallbackWrapper { pub callback: PyObject, - pub context: pyo3_asyncio::TaskLocals + pub context: pyo3_asyncio::TaskLocals, } impl CallbackWrapper { - pub(crate) fn new( - callback: PyObject, - event_loop: &PyAny, - context: &PyAny - ) -> Self { + pub(crate) fn new(callback: PyObject, event_loop: &PyAny, context: &PyAny) -> Self { Self { callback, - context: pyo3_asyncio::TaskLocals::new(event_loop).with_context(context) + context: pyo3_asyncio::TaskLocals::new(event_loop).with_context(context), } } } #[pyclass] pub(crate) struct PyIterAwaitable { - result: Option<PyResult<PyObject>> + result: Option<PyResult<PyObject>>, } impl PyIterAwaitable { @@ -36,7 +31,7 @@ impl PyIterAwaitable { } pub(crate) fn set_result(&mut self, result: PyResult<PyObject>) { - self.result = Some(result) + self.result = Some(result); } } @@ -52,13 +47,11 @@ impl PyIterAwaitable { fn __next__(&mut self, py: Python) -> PyResult<IterNextOutput<PyObject, PyObject>> { match self.result.take() { - Some(res) => { - match res { - Ok(v) => Ok(IterNextOutput::Return(v)), - Err(err) => Err(err) - } + Some(res) => match res { + Ok(v) => Ok(IterNextOutput::Return(v)), + Err(err) => Err(err), }, - _ => Ok(IterNextOutput::Yield(py.None())) + _ => Ok(IterNextOutput::Yield(py.None())), } } } @@ -68,7 +61,7 @@ pub(crate) struct PyFutureAwaitable { py_block: bool, event_loop: PyObject, result: Option<PyResult<PyObject>>, - cb: Option<(PyObject, Py<pyo3::types::PyDict>)> + cb: Option<(PyObject, Py<pyo3::types::PyDict>)>, } impl PyFutureAwaitable { @@ -77,7 +70,7 @@ impl PyFutureAwaitable { event_loop, py_block: true, result: None, - cb: None + cb: None, } } @@ -85,12 +78,9 @@ impl PyFutureAwaitable { pyself.result = Some(result); if let Some((cb, ctx)) = pyself.cb.take() { let py = pyself.py(); - let _ = pyself.event_loop.call_method( - py, - "call_soon_threadsafe", - (cb, &pyself), - Some(ctx.as_ref(py)) - ); + let _ = pyself + .event_loop + .call_method(py, "call_soon_threadsafe", (cb, &pyself), Some(ctx.as_ref(py))); } } } @@ -108,25 +98,22 @@ impl PyFutureAwaitable { #[setter(_asyncio_future_blocking)] fn set_block(&mut self, val: bool) { - self.py_block = val + self.py_block = val; } fn get_loop(&mut self) -> PyObject { self.event_loop.clone() } - fn add_done_callback( - mut pyself: PyRefMut<'_, Self>, - py: Python, - cb: PyObject, - context: PyObject - ) -> PyResult<()> { + fn add_done_callback(mut pyself: PyRefMut<'_, Self>, py: Python, cb: PyObject, context: PyObject) -> PyResult<()> { let kwctx = pyo3::types::PyDict::new(py); kwctx.set_item("context", context)?; match pyself.result { Some(_) => { - pyself.event_loop.call_method(py, "call_soon", (cb, &pyself), Some(kwctx))?; - }, + pyself + .event_loop + .call_method(py, "call_soon", (cb, &pyself), Some(kwctx))?; + } _ => { pyself.cb = Some((cb, kwctx.into_py(py))); } @@ -136,9 +123,9 @@ impl PyFutureAwaitable { fn cancel(mut pyself: PyRefMut<'_, Self>, py: Python) -> bool { if let Some((cb, kwctx)) = pyself.cb.take() { - let _ = pyself.event_loop.call_method( - py, "call_soon", (cb, &pyself), Some(kwctx.as_ref(py)) - ); + let _ = pyself + .event_loop + .call_method(py, "call_soon", (cb, &pyself), Some(kwctx.as_ref(py))); } false } @@ -150,79 +137,69 @@ impl PyFutureAwaitable { pyself } - fn __next__( - mut pyself: PyRefMut<'_, Self> - ) -> PyResult<IterNextOutput<PyRefMut<'_, Self>, PyObject>> { + fn __next__(mut pyself: PyRefMut<'_, Self>) -> PyResult<IterNextOutput<PyRefMut<'_, Self>, PyObject>> { match pyself.result { - Some(_) => { - match pyself.result.take().unwrap() { - Ok(v) => Ok(IterNextOutput::Return(v)), - Err(err) => Err(err) - } + Some(_) => match pyself.result.take().unwrap() { + Ok(v) => Ok(IterNextOutput::Return(v)), + Err(err) => Err(err), }, - _ => Ok(IterNextOutput::Yield(pyself)) + _ => Ok(IterNextOutput::Yield(pyself)), } } } fn contextvars(py: Python) -> PyResult<&PyAny> { Ok(CONTEXTVARS - .get_or_try_init(|| py.import("contextvars").map(|m| m.into()))? + .get_or_try_init(|| py.import("contextvars").map(std::convert::Into::into))? .as_ref(py)) } pub fn empty_pycontext(py: Python) -> PyResult<&PyAny> { Ok(CONTEXT - .get_or_try_init(|| contextvars(py)?.getattr("Context")?.call0().map(|c| c.into()))? + .get_or_try_init(|| { + contextvars(py)? + .getattr("Context")? + .call0() + .map(std::convert::Into::into) + })? .as_ref(py)) } macro_rules! callback_impl_run { () => { - pub fn run<'p>(self, py: Python<'p>) -> PyResult<&'p PyAny> { + pub fn run(self, py: Python<'_>) -> PyResult<&PyAny> { let event_loop = self.context.event_loop(py); let target = self.into_py(py).getattr(py, pyo3::intern!(py, "_loop_task"))?; let kwctx = pyo3::types::PyDict::new(py); kwctx.set_item( pyo3::intern!(py, "context"), - crate::callbacks::empty_pycontext(py)? + crate::callbacks::empty_pycontext(py)?, )?; - event_loop.call_method( - pyo3::intern!(py, "call_soon_threadsafe"), - (target,), - Some(kwctx) - ) + event_loop.call_method(pyo3::intern!(py, "call_soon_threadsafe"), (target,), Some(kwctx)) } }; } macro_rules! callback_impl_run_pytask { () => { - pub fn run<'p>(self, py: Python<'p>) -> PyResult<&'p PyAny> { + pub fn run(self, py: Python<'_>) -> PyResult<&PyAny> { let event_loop = self.context.event_loop(py); let context = self.context.context(py); let target = self.into_py(py).getattr(py, pyo3::intern!(py, "_loop_task"))?; let kwctx = pyo3::types::PyDict::new(py); - kwctx.set_item( - pyo3::intern!(py, "context"), - context - )?; - event_loop.call_method( - pyo3::intern!(py, "call_soon_threadsafe"), - (target,), - Some(kwctx) - ) + kwctx.set_item(pyo3::intern!(py, "context"), context)?; + event_loop.call_method(pyo3::intern!(py, "call_soon_threadsafe"), (target,), Some(kwctx)) } }; } macro_rules! callback_impl_loop_run { () => { - pub fn run<'p>(self, py: Python<'p>) -> PyResult<&'p PyAny> { + pub fn run(self, py: Python<'_>) -> PyResult<&PyAny> { let context = self.pycontext.clone().into_ref(py); context.call_method1( pyo3::intern!(py, "run"), - (self.into_py(py).getattr(py, pyo3::intern!(py, "_loop_step"))?,) + (self.into_py(py).getattr(py, pyo3::intern!(py, "_loop_step"))?,), ) } }; @@ -232,7 +209,7 @@ macro_rules! callback_impl_loop_pytask { ($pyself:expr, $py:expr) => { $pyself.context.event_loop($py).call_method1( pyo3::intern!($py, "create_task"), - ($pyself.cb.clone().into_ref($py).call1(($pyself.into_py($py),))?,) + ($pyself.cb.clone().into_ref($py).call1(($pyself.into_py($py),))?,), ) }; } @@ -241,12 +218,9 @@ macro_rules! callback_impl_loop_step { ($pyself:expr, $py:expr) => { match $pyself.cb.call_method1($py, pyo3::intern!($py, "send"), ($py.None(),)) { Ok(res) => { - let blocking: bool = match res.getattr( - $py, - pyo3::intern!($py, "_asyncio_future_blocking") - ) { + let blocking: bool = match res.getattr($py, pyo3::intern!($py, "_asyncio_future_blocking")) { Ok(v) => v.extract($py)?, - _ => false + _ => false, }; let ctx = $pyself.pycontext.clone(); @@ -255,43 +229,30 @@ macro_rules! callback_impl_loop_step { match blocking { true => { - res.setattr( - $py, - pyo3::intern!($py, "_asyncio_future_blocking"), - false - )?; + res.setattr($py, pyo3::intern!($py, "_asyncio_future_blocking"), false)?; res.call_method( $py, pyo3::intern!($py, "add_done_callback"), - ( - $pyself - .into_py($py) - .getattr($py, pyo3::intern!($py, "_loop_wake"))?, - ), - Some(kwctx) + ($pyself.into_py($py).getattr($py, pyo3::intern!($py, "_loop_wake"))?,), + Some(kwctx), )?; Ok(()) - }, + } false => { let event_loop = $pyself.context.event_loop($py); event_loop.call_method( pyo3::intern!($py, "call_soon"), - ( - $pyself - .into_py($py) - .getattr($py, pyo3::intern!($py, "_loop_step"))?, - ), - Some(kwctx) + ($pyself.into_py($py).getattr($py, pyo3::intern!($py, "_loop_step"))?,), + Some(kwctx), )?; Ok(()) } } - }, + } Err(err) => { - if ( - err.is_instance_of::<pyo3::exceptions::PyStopIteration>($py) || - err.is_instance_of::<pyo3::exceptions::asyncio::CancelledError>($py) - ) { + if (err.is_instance_of::<pyo3::exceptions::PyStopIteration>($py) + || err.is_instance_of::<pyo3::exceptions::asyncio::CancelledError>($py)) + { $pyself.done($py); Ok(()) } else { @@ -307,7 +268,7 @@ macro_rules! callback_impl_loop_wake { ($pyself:expr, $py:expr, $fut:expr) => { match $fut.call_method0($py, pyo3::intern!($py, "result")) { Ok(_) => $pyself.into_py($py).call_method0($py, pyo3::intern!($py, "_loop_step")), - Err(err) => $pyself._loop_err($py, err) + Err(err) => $pyself._loop_err($py, err), } }; } @@ -322,10 +283,10 @@ macro_rules! callback_impl_loop_err { }; } -pub(crate) use callback_impl_run; -pub(crate) use callback_impl_run_pytask; -pub(crate) use callback_impl_loop_run; +pub(crate) use callback_impl_loop_err; pub(crate) use callback_impl_loop_pytask; +pub(crate) use callback_impl_loop_run; pub(crate) use callback_impl_loop_step; pub(crate) use callback_impl_loop_wake; -pub(crate) use callback_impl_loop_err; +pub(crate) use callback_impl_run; +pub(crate) use callback_impl_run_pytask; diff --git a/src/http.rs b/src/http.rs index f4ccef32..e0989d7d 100644 --- a/src/http.rs +++ b/src/http.rs @@ -1,4 +1,7 @@ -use hyper::{Body, Response, header::{HeaderValue, SERVER as HK_SERVER}}; +use hyper::{ + header::{HeaderValue, SERVER as HK_SERVER}, + Body, Response, +}; pub(crate) const HV_SERVER: HeaderValue = HeaderValue::from_static("granian"); diff --git a/src/lib.rs b/src/lib.rs index 327f9df1..97cec0f6 100644 --- a/src/lib.rs +++ b/src/lib.rs @@ -8,8 +8,8 @@ mod callbacks; mod http; mod rsgi; mod runtime; -mod tls; mod tcp; +mod tls; mod utils; mod workers; mod ws; diff --git a/src/rsgi/callbacks.rs b/src/rsgi/callbacks.rs index 09872589..184848c5 100644 --- a/src/rsgi/callbacks.rs +++ b/src/rsgi/callbacks.rs @@ -2,45 +2,33 @@ use pyo3::prelude::*; use pyo3_asyncio::TaskLocals; use tokio::sync::oneshot; +use super::{ + io::{RSGIHTTPProtocol as HTTPProtocol, RSGIWebsocketProtocol as WebsocketProtocol}, + types::{PyResponse, PyResponseBody, RSGIScope as Scope}, +}; use crate::{ callbacks::{ - CallbackWrapper, - callback_impl_run, - callback_impl_run_pytask, - callback_impl_loop_run, - callback_impl_loop_pytask, - callback_impl_loop_step, - callback_impl_loop_wake, - callback_impl_loop_err + callback_impl_loop_err, callback_impl_loop_pytask, callback_impl_loop_run, callback_impl_loop_step, + callback_impl_loop_wake, callback_impl_run, callback_impl_run_pytask, CallbackWrapper, }, runtime::RuntimeRef, - ws::{HyperWebsocket, UpgradeData} + ws::{HyperWebsocket, UpgradeData}, }; -use super::{ - io::{RSGIHTTPProtocol as HTTPProtocol, RSGIWebsocketProtocol as WebsocketProtocol}, - types::{RSGIScope as Scope, PyResponse, PyResponseBody} -}; - #[pyclass] pub(crate) struct CallbackRunnerHTTP { proto: Py<HTTPProtocol>, context: TaskLocals, - cb: PyObject + cb: PyObject, } impl CallbackRunnerHTTP { - pub fn new( - py: Python, - cb: CallbackWrapper, - proto: HTTPProtocol, - scope: Scope - ) -> Self { + pub fn new(py: Python, cb: CallbackWrapper, proto: HTTPProtocol, scope: Scope) -> Self { let pyproto = Py::new(py, proto).unwrap(); Self { proto: pyproto.clone(), context: cb.context, - cb: cb.callback.call1(py, (scope, pyproto)).unwrap() + cb: cb.callback.call1(py, (scope, pyproto)).unwrap(), } } @@ -58,19 +46,17 @@ macro_rules! callback_impl_done_http { ($self:expr, $py:expr) => { if let Ok(mut proto) = $self.proto.as_ref($py).try_borrow_mut() { if let Some(tx) = proto.tx() { - let _ = tx.send( - PyResponse::Body(PyResponseBody::empty(500, Vec::new())) - ); + let _ = tx.send(PyResponse::Body(PyResponseBody::empty(500, Vec::new()))); } } - } + }; } macro_rules! callback_impl_done_err { ($self:expr, $py:expr) => { log::warn!("Application callable raised an exception"); $self.done($py) - } + }; } #[pyclass] @@ -78,18 +64,18 @@ pub(crate) struct CallbackTaskHTTP { proto: Py<HTTPProtocol>, context: TaskLocals, pycontext: PyObject, - cb: PyObject + cb: PyObject, } impl CallbackTaskHTTP { - pub fn new( - py: Python, - cb: PyObject, - proto: Py<HTTPProtocol>, - context: TaskLocals - ) -> PyResult<Self> { + pub fn new(py: Python, cb: PyObject, proto: Py<HTTPProtocol>, context: TaskLocals) -> PyResult<Self> { let pyctx = context.context(py); - Ok(Self { proto, context, pycontext: pyctx.call_method0(pyo3::intern!(py, "copy"))?.into(), cb }) + Ok(Self { + proto, + context, + pycontext: pyctx.call_method0(pyo3::intern!(py, "copy"))?.into(), + cb, + }) } fn done(&self, py: Python) { @@ -122,21 +108,16 @@ pub(crate) struct CallbackWrappedRunnerHTTP { context: TaskLocals, cb: PyObject, #[pyo3(get)] - scope: PyObject + scope: PyObject, } impl CallbackWrappedRunnerHTTP { - pub fn new( - py: Python, - cb: CallbackWrapper, - proto: HTTPProtocol, - scope: Scope - ) -> Self { + pub fn new(py: Python, cb: CallbackWrapper, proto: HTTPProtocol, scope: Scope) -> Self { Self { proto: Py::new(py, proto).unwrap(), context: cb.context, cb: cb.callback, - scope: scope.into_py(py) + scope: scope.into_py(py), } } @@ -162,21 +143,16 @@ impl CallbackWrappedRunnerHTTP { pub(crate) struct CallbackRunnerWebsocket { proto: Py<WebsocketProtocol>, context: TaskLocals, - cb: PyObject + cb: PyObject, } impl CallbackRunnerWebsocket { - pub fn new( - py: Python, - cb: CallbackWrapper, - proto: WebsocketProtocol, - scope: Scope - ) -> Self { + pub fn new(py: Python, cb: CallbackWrapper, proto: WebsocketProtocol, scope: Scope) -> Self { let pyproto = Py::new(py, proto).unwrap(); Self { proto: pyproto.clone(), context: cb.context, - cb: cb.callback.call1(py, (scope, pyproto)).unwrap() + cb: cb.callback.call1(py, (scope, pyproto)).unwrap(), } } @@ -197,7 +173,7 @@ macro_rules! callback_impl_done_ws { let _ = tx.send(res); } } - } + }; } #[pyclass] @@ -205,18 +181,18 @@ pub(crate) struct CallbackTaskWebsocket { proto: Py<WebsocketProtocol>, context: TaskLocals, pycontext: PyObject, - cb: PyObject + cb: PyObject, } impl CallbackTaskWebsocket { - pub fn new( - py: Python, - cb: PyObject, - proto: Py<WebsocketProtocol>, - context: TaskLocals - ) -> PyResult<Self> { + pub fn new(py: Python, cb: PyObject, proto: Py<WebsocketProtocol>, context: TaskLocals) -> PyResult<Self> { let pyctx = context.context(py); - Ok(Self { proto, context, pycontext: pyctx.call_method0(pyo3::intern!(py, "copy"))?.into(), cb }) + Ok(Self { + proto, + context, + pycontext: pyctx.call_method0(pyo3::intern!(py, "copy"))?.into(), + cb, + }) } fn done(&self, py: Python) { @@ -249,21 +225,16 @@ pub(crate) struct CallbackWrappedRunnerWebsocket { context: TaskLocals, cb: PyObject, #[pyo3(get)] - scope: PyObject + scope: PyObject, } impl CallbackWrappedRunnerWebsocket { - pub fn new( - py: Python, - cb: CallbackWrapper, - proto: WebsocketProtocol, - scope: Scope - ) -> Self { + pub fn new(py: Python, cb: CallbackWrapper, proto: WebsocketProtocol, scope: Scope) -> Self { Self { proto: Py::new(py, proto).unwrap(), context: cb.context, cb: cb.callback, - scope: scope.into_py(py) + scope: scope.into_py(py), } } @@ -291,7 +262,7 @@ macro_rules! call_impl_rtb_http { cb: CallbackWrapper, rt: RuntimeRef, req: hyper::Request<hyper::Body>, - scope: Scope + scope: Scope, ) -> oneshot::Receiver<PyResponse> { let (tx, rx) = oneshot::channel(); let protocol = HTTPProtocol::new(rt, tx, req); @@ -311,7 +282,7 @@ macro_rules! call_impl_rtt_http { cb: CallbackWrapper, rt: RuntimeRef, req: hyper::Request<hyper::Body>, - scope: Scope + scope: Scope, ) -> oneshot::Receiver<PyResponse> { let (tx, rx) = oneshot::channel(); let protocol = HTTPProtocol::new(rt, tx, req); @@ -334,7 +305,7 @@ macro_rules! call_impl_rtb_ws { rt: RuntimeRef, ws: HyperWebsocket, upgrade: UpgradeData, - scope: Scope + scope: Scope, ) -> oneshot::Receiver<(i32, bool)> { let (tx, rx) = oneshot::channel(); let protocol = WebsocketProtocol::new(rt, tx, ws, upgrade); @@ -355,7 +326,7 @@ macro_rules! call_impl_rtt_ws { rt: RuntimeRef, ws: HyperWebsocket, upgrade: UpgradeData, - scope: Scope + scope: Scope, ) -> oneshot::Receiver<(i32, bool)> { let (tx, rx) = oneshot::channel(); let protocol = WebsocketProtocol::new(rt, tx, ws, upgrade); diff --git a/src/rsgi/errors.rs b/src/rsgi/errors.rs index 4f7cbbec..fe27eea4 100644 --- a/src/rsgi/errors.rs +++ b/src/rsgi/errors.rs @@ -1,6 +1,5 @@ use pyo3::{create_exception, exceptions::PyRuntimeError}; - create_exception!(_granian, RSGIProtocolError, PyRuntimeError, "RSGIProtocolError"); create_exception!(_granian, RSGIProtocolClosed, PyRuntimeError, "RSGIProtocolClosed"); diff --git a/src/rsgi/http.rs b/src/rsgi/http.rs index f786be62..76b7d66e 100644 --- a/src/rsgi/http.rs +++ b/src/rsgi/http.rs @@ -1,34 +1,22 @@ use hyper::{ - Body, - Request, - Response, - StatusCode, - header::SERVER as HK_SERVER, - http::response::Builder as ResponseBuilder + header::SERVER as HK_SERVER, http::response::Builder as ResponseBuilder, Body, Request, Response, StatusCode, }; use std::net::SocketAddr; use tokio::sync::mpsc; -use crate::{ - callbacks::CallbackWrapper, - http::{HV_SERVER, response_500}, - runtime::RuntimeRef, - ws::{UpgradeData, is_upgrade_request as is_ws_upgrade, upgrade_intent as ws_upgrade} -}; use super::{ callbacks::{ - call_rtb_http, - call_rtb_http_pyw, - call_rtb_ws, - call_rtb_ws_pyw, - call_rtt_http, - call_rtt_http_pyw, - call_rtt_ws, - call_rtt_ws_pyw + call_rtb_http, call_rtb_http_pyw, call_rtb_ws, call_rtb_ws_pyw, call_rtt_http, call_rtt_http_pyw, call_rtt_ws, + call_rtt_ws_pyw, }, - types::{RSGIScope as Scope, PyResponse} + types::{PyResponse, RSGIScope as Scope}, +}; +use crate::{ + callbacks::CallbackWrapper, + http::{response_500, HV_SERVER}, + runtime::RuntimeRef, + ws::{is_upgrade_request as is_ws_upgrade, upgrade_intent as ws_upgrade, UpgradeData}, }; - macro_rules! default_scope { ($server_addr:expr, $client_addr:expr, $req:expr, $scheme:expr) => { @@ -40,7 +28,7 @@ macro_rules! default_scope { $req.method().as_ref(), $server_addr, $client_addr, - $req.headers() + $req.headers(), ) }; } @@ -48,12 +36,8 @@ macro_rules! default_scope { macro_rules! handle_http_response { ($handler:expr, $rt:expr, $callback:expr, $req:expr, $scope:expr) => { match $handler($callback, $rt, $req, $scope).await { - Ok(PyResponse::Body(pyres)) => { - pyres.to_response() - }, - Ok(PyResponse::File(pyres)) => { - pyres.to_response().await - }, + Ok(PyResponse::Body(pyres)) => pyres.to_response(), + Ok(PyResponse::File(pyres)) => pyres.to_response().await, _ => { log::error!("RSGI protocol failure"); response_500() @@ -70,7 +54,7 @@ macro_rules! handle_request { server_addr: SocketAddr, client_addr: SocketAddr, req: Request<Body>, - scheme: &str + scheme: &str, ) -> Response<Body> { let scope = default_scope!(server_addr, client_addr, &req, scheme); handle_http_response!($handler, rt, callback, req, scope) @@ -86,7 +70,7 @@ macro_rules! handle_request_with_ws { server_addr: SocketAddr, client_addr: SocketAddr, req: Request<Body>, - scheme: &str + scheme: &str, ) -> Response<Body> { let mut scope = default_scope!(server_addr, client_addr, &req, scheme); @@ -101,27 +85,23 @@ macro_rules! handle_request_with_ws { rt.inner.spawn(async move { let tx_ref = restx.clone(); - match $handler_ws( - callback, - rth, - ws, - UpgradeData::new(res, restx), - scope - ).await { + match $handler_ws(callback, rth, ws, UpgradeData::new(res, restx), scope).await { Ok((status, consumed)) => { if !consumed { - let _ = tx_ref.send( - ResponseBuilder::new() - .status( - StatusCode::from_u16(status as u16) - .unwrap_or(StatusCode::FORBIDDEN) - ) - .header(HK_SERVER, HV_SERVER) - .body(Body::from("")) - .unwrap() - ).await; + let _ = tx_ref + .send( + ResponseBuilder::new() + .status( + StatusCode::from_u16(status as u16) + .unwrap_or(StatusCode::FORBIDDEN), + ) + .header(HK_SERVER, HV_SERVER) + .body(Body::from("")) + .unwrap(), + ) + .await; } - }, + } _ => { log::error!("RSGI protocol failure"); let _ = tx_ref.send(response_500()).await; @@ -133,10 +113,10 @@ macro_rules! handle_request_with_ws { Some(res) => { resrx.close(); res - }, - _ => response_500() - } - }, + } + _ => response_500(), + }; + } Err(err) => { return ResponseBuilder::new() .status(StatusCode::BAD_REQUEST) @@ -149,7 +129,6 @@ macro_rules! handle_request_with_ws { handle_http_response!($handler_req, rt, callback, req, scope) } - }; } diff --git a/src/rsgi/io.rs b/src/rsgi/io.rs index 4f061484..5b629907 100644 --- a/src/rsgi/io.rs +++ b/src/rsgi/io.rs @@ -1,35 +1,40 @@ use bytes::Bytes; -use futures::{sink::SinkExt, stream::{SplitSink, SplitStream, StreamExt}}; -use hyper::{body::{Body, Sender as BodySender, HttpBody}, Request}; +use futures::{ + sink::SinkExt, + stream::{SplitSink, SplitStream, StreamExt}, +}; +use hyper::{ + body::{Body, HttpBody, Sender as BodySender}, + Request, +}; use pyo3::prelude::*; use pyo3::types::{PyBytes, PyString}; use std::sync::Arc; -use tokio_tungstenite::WebSocketStream; use tokio::sync::{oneshot, Mutex}; +use tokio_tungstenite::WebSocketStream; use tungstenite::Message; -use crate::{ - runtime::{Runtime, RuntimeRef, future_into_py_iter, future_into_py_futlike}, - ws::{HyperWebsocket, UpgradeData} -}; use super::{ errors::{error_proto, error_stream}, - types::{PyResponse, PyResponseBody, PyResponseFile} + types::{PyResponse, PyResponseBody, PyResponseFile}, +}; +use crate::{ + runtime::{future_into_py_futlike, future_into_py_iter, Runtime, RuntimeRef}, + ws::{HyperWebsocket, UpgradeData}, }; - -#[pyclass(module="granian._granian")] +#[pyclass(module = "granian._granian")] pub(crate) struct RSGIHTTPStreamTransport { rt: RuntimeRef, - tx: Arc<Mutex<BodySender>> + tx: Arc<Mutex<BodySender>>, } impl RSGIHTTPStreamTransport { - pub fn new( - rt: RuntimeRef, - transport: BodySender - ) -> Self { - Self { rt: rt, tx: Arc::new(Mutex::new(transport)) } + pub fn new(rt: RuntimeRef, transport: BodySender) -> Self { + Self { + rt, + tx: Arc::new(Mutex::new(transport)), + } } } @@ -41,8 +46,8 @@ impl RSGIHTTPStreamTransport { if let Ok(mut stream) = transport.try_lock() { return match stream.send_data(data.into()).await { Ok(_) => Ok(()), - _ => error_stream!() - } + _ => error_stream!(), + }; } error_proto!() }) @@ -54,31 +59,27 @@ impl RSGIHTTPStreamTransport { if let Ok(mut stream) = transport.try_lock() { return match stream.send_data(data.into()).await { Ok(_) => Ok(()), - _ => error_stream!() - } + _ => error_stream!(), + }; } error_proto!() }) } } -#[pyclass(module="granian._granian")] +#[pyclass(module = "granian._granian")] pub(crate) struct RSGIHTTPProtocol { rt: RuntimeRef, tx: Option<oneshot::Sender<super::types::PyResponse>>, - body: Arc<Mutex<Body>> + body: Arc<Mutex<Body>>, } impl RSGIHTTPProtocol { - pub fn new( - rt: RuntimeRef, - tx: oneshot::Sender<super::types::PyResponse>, - request: Request<Body> - ) -> Self { + pub fn new(rt: RuntimeRef, tx: oneshot::Sender<super::types::PyResponse>, request: Request<Body>) -> Self { Self { rt, tx: Some(tx), - body: Arc::new(Mutex::new(request.into_body())) + body: Arc::new(Mutex::new(request.into_body())), } } @@ -110,14 +111,13 @@ impl RSGIHTTPProtocol { let mut bodym = body_ref.lock().await; let body = &mut *bodym; if body.is_end_stream() { - return Err(pyo3::exceptions::PyStopAsyncIteration::new_err("stream exhausted")) + return Err(pyo3::exceptions::PyStopAsyncIteration::new_err("stream exhausted")); } - let chunk = body.data().await.map_or_else(|| Bytes::new(), |buf| { - buf.unwrap_or_else(|_| Bytes::new()) - }); - Ok(Python::with_gil(|py| { - PyBytes::new(py, &chunk[..]).to_object(py) - })) + let chunk = body + .data() + .await + .map_or_else(Bytes::new, |buf| buf.unwrap_or_else(|_| Bytes::new())); + Ok(Python::with_gil(|py| PyBytes::new(py, &chunk[..]).to_object(py))) })?; Ok(Some(fut)) } @@ -125,36 +125,28 @@ impl RSGIHTTPProtocol { #[pyo3(signature = (status=200, headers=vec![]))] fn response_empty(&mut self, status: u16, headers: Vec<(String, String)>) { if let Some(tx) = self.tx.take() { - let _ = tx.send( - PyResponse::Body(PyResponseBody::empty(status, headers)) - ); + let _ = tx.send(PyResponse::Body(PyResponseBody::empty(status, headers))); } } #[pyo3(signature = (status=200, headers=vec![], body=vec![]))] fn response_bytes(&mut self, status: u16, headers: Vec<(String, String)>, body: Vec<u8>) { if let Some(tx) = self.tx.take() { - let _ = tx.send( - PyResponse::Body(PyResponseBody::from_bytes(status, headers, body)) - ); + let _ = tx.send(PyResponse::Body(PyResponseBody::from_bytes(status, headers, body))); } } - #[pyo3(signature = (status=200, headers=vec![], body="".to_string()))] + #[pyo3(signature = (status=200, headers=vec![], body=String::new()))] fn response_str(&mut self, status: u16, headers: Vec<(String, String)>, body: String) { if let Some(tx) = self.tx.take() { - let _ = tx.send( - PyResponse::Body(PyResponseBody::from_string(status, headers, body)) - ); + let _ = tx.send(PyResponse::Body(PyResponseBody::from_string(status, headers, body))); } } #[pyo3(signature = (status, headers, file))] fn response_file(&mut self, status: u16, headers: Vec<(String, String)>, file: String) { if let Some(tx) = self.tx.take() { - let _ = tx.send( - PyResponse::File(PyResponseFile::new(status, headers, file)) - ); + let _ = tx.send(PyResponse::File(PyResponseFile::new(status, headers, file))); } } @@ -163,34 +155,33 @@ impl RSGIHTTPProtocol { &mut self, py: Python<'p>, status: u16, - headers: Vec<(String, String)> + headers: Vec<(String, String)>, ) -> PyResult<&'p PyAny> { if let Some(tx) = self.tx.take() { let (body_tx, body_stream) = Body::channel(); - let _ = tx.send( - PyResponse::Body(PyResponseBody::new(status, headers, body_stream)) - ); + let _ = tx.send(PyResponse::Body(PyResponseBody::new(status, headers, body_stream))); let trx = Py::new(py, RSGIHTTPStreamTransport::new(self.rt.clone(), body_tx))?; - return Ok(trx.into_ref(py)) + return Ok(trx.into_ref(py)); } error_proto!() } } -#[pyclass(module="granian._granian")] +#[pyclass(module = "granian._granian")] pub(crate) struct RSGIWebsocketTransport { rt: RuntimeRef, tx: Arc<Mutex<SplitSink<WebSocketStream<hyper::upgrade::Upgraded>, Message>>>, - rx: Arc<Mutex<SplitStream<WebSocketStream<hyper::upgrade::Upgraded>>>> + rx: Arc<Mutex<SplitStream<WebSocketStream<hyper::upgrade::Upgraded>>>>, } impl RSGIWebsocketTransport { - pub fn new( - rt: RuntimeRef, - transport: WebSocketStream<hyper::upgrade::Upgraded> - ) -> Self { + pub fn new(rt: RuntimeRef, transport: WebSocketStream<hyper::upgrade::Upgraded>) -> Self { let (tx, rx) = transport.split(); - Self { rt: rt, tx: Arc::new(Mutex::new(tx)), rx: Arc::new(Mutex::new(rx)) } + Self { + rt, + tx: Arc::new(Mutex::new(tx)), + rx: Arc::new(Mutex::new(rx)), + } } pub fn close(&self) { @@ -209,27 +200,14 @@ impl RSGIWebsocketTransport { let transport = self.rx.clone(); future_into_py_futlike(self.rt.clone(), py, async move { if let Ok(mut stream) = transport.try_lock() { - loop { - match stream.next().await { - Some(recv) => { - match recv { - Ok(Message::Ping(_)) => { - continue - }, - Ok(message) => { - return message_into_py(message) - }, - _ => { - break - } - } - }, - _ => { - break - } + while let Some(recv) = stream.next().await { + match recv { + Ok(Message::Ping(_)) => continue, + Ok(message) => return message_into_py(message), + _ => break, } } - return error_stream!() + return error_stream!(); } error_proto!() }) @@ -241,8 +219,8 @@ impl RSGIWebsocketTransport { if let Ok(mut stream) = transport.try_lock() { return match stream.send(Message::Binary(data)).await { Ok(_) => Ok(()), - _ => error_stream!() - } + _ => error_stream!(), + }; } error_proto!() }) @@ -254,22 +232,22 @@ impl RSGIWebsocketTransport { if let Ok(mut stream) = transport.try_lock() { return match stream.send(Message::Text(data)).await { Ok(_) => Ok(()), - _ => error_stream!() - } + _ => error_stream!(), + }; } error_proto!() }) } } -#[pyclass(module="granian._granian")] +#[pyclass(module = "granian._granian")] pub(crate) struct RSGIWebsocketProtocol { rt: RuntimeRef, tx: Option<oneshot::Sender<(i32, bool)>>, websocket: Arc<Mutex<HyperWebsocket>>, upgrade: Option<UpgradeData>, transport: Arc<Mutex<Option<Py<RSGIWebsocketTransport>>>>, - status: i32 + status: i32, } impl RSGIWebsocketProtocol { @@ -277,7 +255,7 @@ impl RSGIWebsocketProtocol { rt: RuntimeRef, tx: oneshot::Sender<(i32, bool)>, websocket: HyperWebsocket, - upgrade: UpgradeData + upgrade: UpgradeData, ) -> Self { Self { rt, @@ -285,15 +263,12 @@ impl RSGIWebsocketProtocol { websocket: Arc::new(Mutex::new(websocket)), upgrade: Some(upgrade), transport: Arc::new(Mutex::new(None)), - status: 0 + status: 0, } } fn consumed(&self) -> bool { - match &self.upgrade { - Some(_) => false, - _ => true - } + self.upgrade.is_none() } pub fn tx(&mut self) -> (Option<oneshot::Sender<(i32, bool)>>, (i32, bool)) { @@ -304,18 +279,20 @@ impl RSGIWebsocketProtocol { enum WebsocketMessageType { Close = 0, Bytes = 1, - Text = 2 + Text = 2, } #[pyclass] struct WebsocketInboundCloseMessage { #[pyo3(get)] - kind: usize + kind: usize, } impl WebsocketInboundCloseMessage { pub fn new() -> Self { - Self { kind: WebsocketMessageType::Close as usize } + Self { + kind: WebsocketMessageType::Close as usize, + } } } @@ -324,12 +301,15 @@ struct WebsocketInboundBytesMessage { #[pyo3(get)] kind: usize, #[pyo3(get)] - data: Py<PyBytes> + data: Py<PyBytes>, } impl WebsocketInboundBytesMessage { - pub fn new(data:Py<PyBytes>) -> Self { - Self { kind: WebsocketMessageType::Bytes as usize, data: data } + pub fn new(data: Py<PyBytes>) -> Self { + Self { + kind: WebsocketMessageType::Bytes as usize, + data, + } } } @@ -338,12 +318,15 @@ struct WebsocketInboundTextMessage { #[pyo3(get)] kind: usize, #[pyo3(get)] - data: Py<PyString> + data: Py<PyString>, } impl WebsocketInboundTextMessage { pub fn new(data: Py<PyString>) -> Self { - Self { kind: WebsocketMessageType::Text as usize, data: data } + Self { + kind: WebsocketMessageType::Text as usize, + data, + } } } @@ -374,23 +357,18 @@ impl RSGIWebsocketProtocol { future_into_py_iter(self.rt.clone(), py, async move { let mut ws = transport.lock().await; match upgrade.send().await { - Ok(_) => { - match (&mut *ws).await { - Ok(stream) => { - let mut trx = itransport.lock().await; - Ok(Python::with_gil(|py| { - let pytransport = Py::new( - py, - RSGIWebsocketTransport::new(rth, stream) - ).unwrap(); - *trx = Some(pytransport.clone()); - pytransport - })) - }, - _ => error_proto!() + Ok(_) => match (&mut *ws).await { + Ok(stream) => { + let mut trx = itransport.lock().await; + Ok(Python::with_gil(|py| { + let pytransport = Py::new(py, RSGIWebsocketTransport::new(rth, stream)).unwrap(); + *trx = Some(pytransport.clone()); + pytransport + })) } + _ => error_proto!(), }, - _ => error_proto!() + _ => error_proto!(), } }) } @@ -399,25 +377,13 @@ impl RSGIWebsocketProtocol { #[inline(always)] fn message_into_py(message: Message) -> PyResult<PyObject> { match message { - Message::Binary(message) => { - Ok(Python::with_gil(|py| { - WebsocketInboundBytesMessage::new( - PyBytes::new(py, &message).into() - ).into_py(py) - })) - }, - Message::Text(message) => { - Ok(Python::with_gil(|py| { - WebsocketInboundTextMessage::new( - PyString::new(py, &message).into() - ).into_py(py) - })) - }, - Message::Close(_) => { - Ok(Python::with_gil(|py| { - WebsocketInboundCloseMessage::new().into_py(py) - })) - } + Message::Binary(message) => Ok(Python::with_gil(|py| { + WebsocketInboundBytesMessage::new(PyBytes::new(py, &message).into()).into_py(py) + })), + Message::Text(message) => Ok(Python::with_gil(|py| { + WebsocketInboundTextMessage::new(PyString::new(py, &message).into()).into_py(py) + })), + Message::Close(_) => Ok(Python::with_gil(|py| WebsocketInboundCloseMessage::new().into_py(py))), v => { log::warn!("Unsupported websocket message received {:?}", v); error_proto!() diff --git a/src/rsgi/serve.rs b/src/rsgi/serve.rs index f361c095..fec2c5f2 100644 --- a/src/rsgi/serve.rs +++ b/src/rsgi/serve.rs @@ -1,28 +1,14 @@ use pyo3::prelude::*; -use crate::{ - workers::{ - WorkerConfig, - serve_rth, - serve_wth, - serve_rth_ssl, - serve_wth_ssl - } -}; use super::http::{ - handle_rtb, - handle_rtb_pyw, - handle_rtt, - handle_rtt_pyw, - handle_rtb_ws, - handle_rtb_ws_pyw, - handle_rtt_ws, - handle_rtt_ws_pyw + handle_rtb, handle_rtb_pyw, handle_rtb_ws, handle_rtb_ws_pyw, handle_rtt, handle_rtt_pyw, handle_rtt_ws, + handle_rtt_ws_pyw, }; +use crate::workers::{serve_rth, serve_rth_ssl, serve_wth, serve_wth_ssl, WorkerConfig}; -#[pyclass(module="granian._granian")] +#[pyclass(module = "granian._granian")] pub struct RSGIWorker { - config: WorkerConfig + config: WorkerConfig, } impl RSGIWorker { @@ -73,7 +59,7 @@ impl RSGIWorker { opt_enabled: bool, ssl_enabled: bool, ssl_cert: Option<&str>, - ssl_key: Option<&str> + ssl_key: Option<&str>, ) -> PyResult<Self> { Ok(Self { config: WorkerConfig::new( @@ -87,22 +73,16 @@ impl RSGIWorker { opt_enabled, ssl_enabled, ssl_cert, - ssl_key - ) + ssl_key, + ), }) } - fn serve_rth( - &self, - callback: PyObject, - event_loop: &PyAny, - context: &PyAny, - signal_rx: PyObject - ) { + fn serve_rth(&self, callback: PyObject, event_loop: &PyAny, context: &PyAny, signal_rx: PyObject) { match ( self.config.websockets_enabled, self.config.ssl_enabled, - self.config.opt_enabled + self.config.opt_enabled, ) { (false, false, true) => self._serve_rth(callback, event_loop, context, signal_rx), (false, false, false) => self._serve_rth_pyw(callback, event_loop, context, signal_rx), @@ -111,21 +91,15 @@ impl RSGIWorker { (false, true, true) => self._serve_rth_ssl(callback, event_loop, context, signal_rx), (false, true, false) => self._serve_rth_ssl_pyw(callback, event_loop, context, signal_rx), (true, true, true) => self._serve_rth_ssl_ws(callback, event_loop, context, signal_rx), - (true, true, false) => self._serve_rth_ssl_ws_pyw(callback, event_loop, context, signal_rx) + (true, true, false) => self._serve_rth_ssl_ws_pyw(callback, event_loop, context, signal_rx), } } - fn serve_wth( - &self, - callback: PyObject, - event_loop: &PyAny, - context: &PyAny, - signal_rx: PyObject - ) { + fn serve_wth(&self, callback: PyObject, event_loop: &PyAny, context: &PyAny, signal_rx: PyObject) { match ( self.config.websockets_enabled, self.config.ssl_enabled, - self.config.opt_enabled + self.config.opt_enabled, ) { (false, false, true) => self._serve_wth(callback, event_loop, context, signal_rx), (false, false, false) => self._serve_wth_pyw(callback, event_loop, context, signal_rx), @@ -134,7 +108,7 @@ impl RSGIWorker { (false, true, true) => self._serve_wth_ssl(callback, event_loop, context, signal_rx), (false, true, false) => self._serve_wth_ssl_pyw(callback, event_loop, context, signal_rx), (true, true, true) => self._serve_wth_ssl_ws(callback, event_loop, context, signal_rx), - (true, true, false) => self._serve_wth_ssl_ws_pyw(callback, event_loop, context, signal_rx) + (true, true, false) => self._serve_wth_ssl_ws_pyw(callback, event_loop, context, signal_rx), } } } diff --git a/src/rsgi/types.rs b/src/rsgi/types.rs index 37123235..66037012 100644 --- a/src/rsgi/types.rs +++ b/src/rsgi/types.rs @@ -1,6 +1,6 @@ use hyper::{ header::{HeaderMap, HeaderName, HeaderValue, SERVER as HK_SERVER}, - Body, Uri, Version + Body, Uri, Version, }; use pyo3::prelude::*; use pyo3::types::PyString; @@ -10,11 +10,10 @@ use tokio_util::codec::{BytesCodec, FramedRead}; use crate::http::HV_SERVER; - -#[pyclass(module="granian._granian")] +#[pyclass(module = "granian._granian")] #[derive(Clone)] pub(crate) struct RSGIHeaders { - inner: HeaderMap + inner: HeaderMap, } impl RSGIHeaders { @@ -29,7 +28,7 @@ impl RSGIHeaders { let mut ret = Vec::with_capacity(self.inner.keys_len()); for key in self.inner.keys() { ret.push(key.as_str()); - }; + } ret } @@ -37,15 +36,15 @@ impl RSGIHeaders { let mut ret = Vec::with_capacity(self.inner.keys_len()); for val in self.inner.values() { ret.push(val.to_str().unwrap()); - }; + } Ok(ret) } fn items(&self) -> PyResult<Vec<(&str, &str)>> { let mut ret = Vec::with_capacity(self.inner.keys_len()); - for (key, val) in self.inner.iter() { + for (key, val) in &self.inner { ret.push((key.as_str(), val.to_str().unwrap())); - }; + } Ok(ret) } @@ -56,18 +55,16 @@ impl RSGIHeaders { #[pyo3(signature = (key, default=None))] fn get(&self, py: Python, key: &str, default: Option<PyObject>) -> Option<PyObject> { match self.inner.get(key) { - Some(val) => { - match val.to_str() { - Ok(string) => Some(PyString::new(py, string).into()), - _ => default - } + Some(val) => match val.to_str() { + Ok(string) => Some(PyString::new(py, string).into()), + _ => default, }, - _ => default + _ => default, } } } -#[pyclass(module="granian._granian")] +#[pyclass(module = "granian._granian")] pub(crate) struct RSGIScope { #[pyo3(get)] proto: String, @@ -84,7 +81,7 @@ pub(crate) struct RSGIScope { #[pyo3(get)] client: String, #[pyo3(get)] - headers: RSGIHeaders + headers: RSGIHeaders, } impl RSGIScope { @@ -96,23 +93,23 @@ impl RSGIScope { method: &str, server: SocketAddr, client: SocketAddr, - headers: &HeaderMap + headers: &HeaderMap, ) -> Self { Self { proto: proto.to_string(), - http_version: http_version, + http_version, rsgi_version: "1.2".to_string(), scheme: scheme.to_string(), method: method.to_string(), - uri: uri, + uri, server: server.to_string(), client: client.to_string(), - headers: RSGIHeaders::new(headers) + headers: RSGIHeaders::new(headers), } } pub fn set_proto(&mut self, value: &str) { - self.proto = value.to_string() + self.proto = value.to_string(); } } @@ -125,7 +122,7 @@ impl RSGIScope { Version::HTTP_11 => "1.1", Version::HTTP_2 => "2", Version::HTTP_3 => "3", - _ => "1" + _ => "1", } } @@ -142,38 +139,36 @@ impl RSGIScope { pub(crate) enum PyResponse { Body(PyResponseBody), - File(PyResponseFile) + File(PyResponseFile), } pub(crate) struct PyResponseBody { status: u16, headers: Vec<(String, String)>, - body: Body + body: Body, } pub(crate) struct PyResponseFile { status: u16, headers: Vec<(String, String)>, - file_path: String + file_path: String, } macro_rules! response_head_from_py { - ($status:expr, $headers:expr, $res:expr) => { - { - let mut rh = hyper::http::HeaderMap::new(); - - rh.insert(HK_SERVER, HV_SERVER); - for (key, value) in $headers { - rh.append( - HeaderName::from_bytes(key.as_bytes()).unwrap(), - HeaderValue::from_str(&value).unwrap() - ); - } - - *$res.status_mut() = $status.try_into().unwrap(); - *$res.headers_mut() = rh; + ($status:expr, $headers:expr, $res:expr) => {{ + let mut rh = hyper::http::HeaderMap::new(); + + rh.insert(HK_SERVER, HV_SERVER); + for (key, value) in $headers { + rh.append( + HeaderName::from_bytes(key.as_bytes()).unwrap(), + HeaderValue::from_str(&value).unwrap(), + ); } - } + + *$res.status_mut() = $status.try_into().unwrap(); + *$res.headers_mut() = rh; + }}; } impl PyResponseBody { @@ -182,18 +177,30 @@ impl PyResponseBody { } pub fn empty(status: u16, headers: Vec<(String, String)>) -> Self { - Self { status, headers, body: Body::empty() } + Self { + status, + headers, + body: Body::empty(), + } } pub fn from_bytes(status: u16, headers: Vec<(String, String)>, body: Vec<u8>) -> Self { - Self { status, headers, body: Body::from(body) } + Self { + status, + headers, + body: Body::from(body), + } } pub fn from_string(status: u16, headers: Vec<(String, String)>, body: String) -> Self { - Self { status, headers, body: Body::from(body) } + Self { + status, + headers, + body: Body::from(body), + } } - pub fn to_response(self) -> hyper::Response::<Body> { + pub fn to_response(self) -> hyper::Response<Body> { let mut res = hyper::Response::<Body>::new(self.body); response_head_from_py!(self.status, &self.headers, res); res @@ -202,10 +209,14 @@ impl PyResponseBody { impl PyResponseFile { pub fn new(status: u16, headers: Vec<(String, String)>, file_path: String) -> Self { - Self { status, headers, file_path } + Self { + status, + headers, + file_path, + } } - pub async fn to_response(&self) -> hyper::Response::<Body> { + pub async fn to_response(&self) -> hyper::Response<Body> { let file = File::open(&self.file_path).await.unwrap(); let stream = FramedRead::new(file, BytesCodec::new()); let mut res = hyper::Response::<Body>::new(Body::wrap_stream(stream)); diff --git a/src/runtime.rs b/src/runtime.rs index 5b5e2198..4ed03ebb 100644 --- a/src/runtime.rs +++ b/src/runtime.rs @@ -1,12 +1,19 @@ use once_cell::unsync::OnceCell as UnsyncOnceCell; -use pyo3_asyncio::TaskLocals; use pyo3::prelude::*; -use std::{future::Future, io, pin::Pin, sync::{Arc, Mutex}}; -use tokio::{runtime::Builder, task::{JoinHandle, LocalSet}}; +use pyo3_asyncio::TaskLocals; +use std::{ + future::Future, + io, + pin::Pin, + sync::{Arc, Mutex}, +}; +use tokio::{ + runtime::Builder, + task::{JoinHandle, LocalSet}, +}; use super::callbacks::{PyFutureAwaitable, PyIterAwaitable}; - tokio::task_local! { static TASK_LOCALS: UnsyncOnceCell<TaskLocals>; } @@ -27,11 +34,7 @@ pub trait Runtime: Send + 'static { } pub trait ContextExt: Runtime { - fn scope<F, R>( - &self, - locals: TaskLocals, - fut: F - ) -> Pin<Box<dyn Future<Output = R> + Send>> + fn scope<F, R>(&self, locals: TaskLocals, fut: F) -> Pin<Box<dyn Future<Output = R> + Send>> where F: Future<Output = R> + Send + 'static; @@ -45,36 +48,34 @@ pub trait SpawnLocalExt: Runtime { } pub trait LocalContextExt: Runtime { - fn scope_local<F, R>( - &self, - locals: TaskLocals, - fut: F - ) -> Pin<Box<dyn Future<Output = R>>> + fn scope_local<F, R>(&self, locals: TaskLocals, fut: F) -> Pin<Box<dyn Future<Output = R>>> where F: Future<Output = R> + 'static; } pub(crate) struct RuntimeWrapper { - rt: tokio::runtime::Runtime + rt: tokio::runtime::Runtime, } impl RuntimeWrapper { pub fn new(blocking_threads: usize) -> Self { - Self { rt: default_runtime(blocking_threads).unwrap() } + Self { + rt: default_runtime(blocking_threads).unwrap(), + } } pub fn with_runtime(rt: tokio::runtime::Runtime) -> Self { - Self { rt: rt } + Self { rt } } pub fn handler(&self) -> RuntimeRef { - RuntimeRef::new(self.rt.handle().to_owned()) + RuntimeRef::new(self.rt.handle().clone()) } } #[derive(Clone)] pub struct RuntimeRef { - pub inner: tokio::runtime::Handle + pub inner: tokio::runtime::Handle, } impl RuntimeRef { @@ -108,11 +109,7 @@ impl Runtime for RuntimeRef { } impl ContextExt for RuntimeRef { - fn scope<F, R>( - &self, - locals: TaskLocals, - fut: F - ) -> Pin<Box<dyn Future<Output = R> + Send>> + fn scope<F, R>(&self, locals: TaskLocals, fut: F) -> Pin<Box<dyn Future<Output = R> + Send>> where F: Future<Output = R> + Send + 'static, { @@ -123,7 +120,7 @@ impl ContextExt for RuntimeRef { } fn get_task_locals() -> Option<TaskLocals> { - match TASK_LOCALS.try_with(|c| c.get().map(|locals| locals.clone())) { + match TASK_LOCALS.try_with(|c| c.get().cloned()) { Ok(locals) => locals, Err(_) => None, } @@ -140,11 +137,7 @@ impl SpawnLocalExt for RuntimeRef { } impl LocalContextExt for RuntimeRef { - fn scope_local<F, R>( - &self, - locals: TaskLocals, - fut: F - ) -> Pin<Box<dyn Future<Output = R>>> + fn scope_local<F, R>(&self, locals: TaskLocals, fut: F) -> Pin<Box<dyn Future<Output = R>>> where F: Future<Output = R> + 'static, { @@ -169,7 +162,7 @@ pub(crate) fn init_runtime_mt(threads: usize, blocking_threads: usize) -> Runtim .max_blocking_threads(blocking_threads) .enable_all() .build() - .unwrap() + .unwrap(), ) } @@ -177,12 +170,8 @@ pub(crate) fn init_runtime_st(blocking_threads: usize) -> RuntimeWrapper { RuntimeWrapper::new(blocking_threads) } -pub(crate) fn into_future( - awaitable: &PyAny, -) -> PyResult<impl Future<Output = PyResult<PyObject>> + Send> { - pyo3_asyncio::into_future_with_locals( - &get_current_locals::<RuntimeRef>(awaitable.py())?, awaitable - ) +pub(crate) fn into_future(awaitable: &PyAny) -> PyResult<impl Future<Output = PyResult<PyObject>> + Send> { + pyo3_asyncio::into_future_with_locals(&get_current_locals::<RuntimeRef>(awaitable.py())?, awaitable) } #[inline] @@ -241,10 +230,7 @@ where rt.spawn(async move { let result = fut.await; Python::with_gil(move |py| { - PyFutureAwaitable::set_result( - py_aw.as_ref(py).borrow_mut(), - result.map(|v| v.into_py(py)) - ); + PyFutureAwaitable::set_result(py_aw.as_ref(py).borrow_mut(), result.map(|v| v.into_py(py))); }); }); @@ -269,11 +255,7 @@ where let rth = rt.handler(); rt.spawn(async move { - let val = rth.scope( - task_locals.clone(), - fut - ) - .await; + let val = rth.scope(task_locals.clone(), fut).await; if let Ok(mut result) = result_tx.lock() { *result = Some(val.unwrap()); } @@ -292,7 +274,7 @@ where pub(crate) fn block_on_local<F>(rt: RuntimeWrapper, local: LocalSet, fut: F) where - F: Future + 'static + F: Future + 'static, { local.block_on(&rt.rt, fut); } diff --git a/src/tcp.rs b/src/tcp.rs index 4aa54eef..7b32fc95 100644 --- a/src/tcp.rs +++ b/src/tcp.rs @@ -9,10 +9,9 @@ use std::os::windows::io::{AsRawSocket, FromRawSocket}; use socket2::{Domain, Protocol, Socket, Type}; - -#[pyclass(module="granian._granian")] +#[pyclass(module = "granian._granian")] pub struct ListenerHolder { - socket: TcpListener + socket: TcpListener, } #[pymethods] @@ -20,28 +19,19 @@ impl ListenerHolder { #[cfg(unix)] #[new] pub fn new(fd: i32) -> PyResult<Self> { - let socket = unsafe { - TcpListener::from_raw_fd(fd) - }; - Ok(Self { socket: socket }) + let socket = unsafe { TcpListener::from_raw_fd(fd) }; + Ok(Self { socket }) } #[cfg(windows)] #[new] pub fn new(fd: u64) -> PyResult<Self> { - let socket = unsafe { - TcpListener::from_raw_socket(fd) - }; - Ok(Self { socket: socket }) + let socket = unsafe { TcpListener::from_raw_socket(fd) }; + Ok(Self { socket }) } #[classmethod] - pub fn from_address( - _cls: &PyType, - address: &str, - port: u16, - backlog: i32 - ) -> PyResult<Self> { + pub fn from_address(_cls: &PyType, address: &str, port: u16, backlog: i32) -> PyResult<Self> { let address: SocketAddr = (address.parse::<IpAddr>()?, port).into(); let socket = Socket::new(Domain::IPV4, Type::STREAM, Some(Protocol::TCP))?; socket.set_reuse_address(true)?; @@ -54,17 +44,13 @@ impl ListenerHolder { #[cfg(unix)] pub fn __getstate__(&self, py: Python) -> PyObject { let fd = self.socket.as_raw_fd(); - ( - fd.into_py(py), - ).to_object(py) + (fd.into_py(py),).to_object(py) } #[cfg(windows)] pub fn __getstate__(&self, py: Python) -> PyObject { let fd = self.socket.as_raw_socket(); - ( - fd.into_py(py), - ).to_object(py) + (fd.into_py(py),).to_object(py) } #[cfg(unix)] @@ -84,7 +70,6 @@ impl ListenerHolder { } } - pub(crate) fn init_pymodule(module: &PyModule) -> PyResult<()> { module.add_class::<ListenerHolder>()?; diff --git a/src/tls.rs b/src/tls.rs index d97d39bf..9355ed2f 100644 --- a/src/tls.rs +++ b/src/tls.rs @@ -1,20 +1,22 @@ use futures::stream::StreamExt; -use hyper::server::{accept, conn::{AddrIncoming, AddrStream}}; +use hyper::server::{ + accept, + conn::{AddrIncoming, AddrStream}, +}; use std::{fs, future, io, iter::Iterator, net::TcpListener, sync::Arc}; use tls_listener::{Error as TlsError, TlsListener}; use tokio_rustls::{ - TlsAcceptor, rustls::{Certificate, PrivateKey, ServerConfig}, - server::TlsStream + server::TlsStream, + TlsAcceptor, }; - pub(crate) type TlsAddrStream = TlsStream<AddrStream>; pub(crate) fn tls_listen( config: Arc<ServerConfig>, - tcp: TcpListener -) -> impl accept::Accept<Conn=TlsAddrStream, Error=TlsError<io::Error, io::Error>> { + tcp: TcpListener, +) -> impl accept::Accept<Conn = TlsAddrStream, Error = TlsError<io::Error, io::Error>> { tcp.set_nonblocking(true).unwrap(); let tcp_listener = tokio::net::TcpListener::from_std(tcp).unwrap(); let incoming = AddrIncoming::from_listener(tcp_listener).unwrap(); @@ -34,30 +36,26 @@ fn tls_error(err: String) -> io::Error { } pub(crate) fn load_certs(filename: &str) -> io::Result<Vec<Certificate>> { - let certfile = fs::File::open(filename) - .map_err(|e| tls_error(format!("failed to open {}: {}", filename, e)))?; + let certfile = fs::File::open(filename).map_err(|e| tls_error(format!("failed to open {filename}: {e}")))?; let mut reader = io::BufReader::new(certfile); - let certs = rustls_pemfile::certs(&mut reader) - .map_err(|_| tls_error("failed to load certificate".into()))?; + let certs = rustls_pemfile::certs(&mut reader).map_err(|_| tls_error("failed to load certificate".into()))?; Ok(certs.into_iter().map(Certificate).collect()) } pub(crate) fn load_private_key(filename: &str) -> io::Result<PrivateKey> { - let keyfile = fs::File::open(filename) - .map_err(|e| tls_error(format!("failed to open {}: {}", filename, e)))?; + let keyfile = fs::File::open(filename).map_err(|e| tls_error(format!("failed to open {filename}: {e}")))?; let mut reader = io::BufReader::new(keyfile); - let keys = rustls_pemfile::read_all(&mut reader) - .map_err(|_| tls_error("failed to load private key".into()))?; + let keys = rustls_pemfile::read_all(&mut reader).map_err(|_| tls_error("failed to load private key".into()))?; if keys.len() != 1 { return Err(tls_error("expected a single private key".into())); } let key = match &keys[0] { - rustls_pemfile::Item::RSAKey(key) => PrivateKey(key.to_vec()), - rustls_pemfile::Item::PKCS8Key(key) => PrivateKey(key.to_vec()), - rustls_pemfile::Item::ECKey(key) => PrivateKey(key.to_vec()), + rustls_pemfile::Item::RSAKey(key) => PrivateKey(key.clone()), + rustls_pemfile::Item::PKCS8Key(key) => PrivateKey(key.clone()), + rustls_pemfile::Item::ECKey(key) => PrivateKey(key.clone()), _ => { return Err(tls_error("failed to load private key".into())); } diff --git a/src/utils.rs b/src/utils.rs index 379ec656..93cf4c3d 100644 --- a/src/utils.rs +++ b/src/utils.rs @@ -1,13 +1,15 @@ pub(crate) fn header_contains_value( headers: &hyper::HeaderMap, header: impl hyper::header::AsHeaderName, - value: impl AsRef<[u8]> + value: impl AsRef<[u8]>, ) -> bool { let value = value.as_ref(); for header in headers.get_all(header) { - if header.as_bytes().split(|&c| c == b',').any( - |x| trim(x).eq_ignore_ascii_case(value) - ) { + if header + .as_bytes() + .split(|&c| c == b',') + .any(|x| trim(x).eq_ignore_ascii_case(value)) + { return true; } } @@ -30,7 +32,7 @@ fn trim_start(data: &[u8]) -> &[u8] { #[inline] fn trim_end(data: &[u8]) -> &[u8] { if let Some(last) = data.iter().rposition(|x| !x.is_ascii_whitespace()) { - &data[..last + 1] + &data[..=last] } else { b"" } diff --git a/src/workers.rs b/src/workers.rs index e49fa3b6..fc7750d2 100644 --- a/src/workers.rs +++ b/src/workers.rs @@ -8,8 +8,8 @@ use std::os::windows::io::FromRawSocket; use super::asgi::serve::ASGIWorker; use super::rsgi::serve::RSGIWorker; -use super::wsgi::serve::WSGIWorker; use super::tls::{load_certs as tls_load_certs, load_private_key as tls_load_pkey}; +use super::wsgi::serve::WSGIWorker; pub(crate) struct WorkerConfig { pub id: i32, @@ -22,7 +22,7 @@ pub(crate) struct WorkerConfig { pub opt_enabled: bool, pub ssl_enabled: bool, ssl_cert: Option<String>, - ssl_key: Option<String> + ssl_key: Option<String>, } impl WorkerConfig { @@ -37,7 +37,7 @@ impl WorkerConfig { opt_enabled: bool, ssl_enabled: bool, ssl_cert: Option<&str>, - ssl_key: Option<&str> + ssl_key: Option<&str>, ) -> Self { Self { id, @@ -49,23 +49,19 @@ impl WorkerConfig { websockets_enabled, opt_enabled, ssl_enabled, - ssl_cert: ssl_cert.map_or(None, |v| Some(v.into())), - ssl_key: ssl_key.map_or(None, |v| Some(v.into())) + ssl_cert: ssl_cert.map(std::convert::Into::into), + ssl_key: ssl_key.map(std::convert::Into::into), } } #[cfg(unix)] pub fn tcp_listener(&self) -> TcpListener { - unsafe { - TcpListener::from_raw_fd(self.socket_fd) - } + unsafe { TcpListener::from_raw_fd(self.socket_fd) } } #[cfg(windows)] pub fn tcp_listener(&self) -> TcpListener { - unsafe { - TcpListener::from_raw_socket(self.socket_fd as u64) - } + unsafe { TcpListener::from_raw_socket(self.socket_fd as u64) } } pub fn tls_cfg(&self) -> tokio_rustls::rustls::ServerConfig { @@ -74,13 +70,13 @@ impl WorkerConfig { .with_no_client_auth() .with_single_cert( tls_load_certs(&self.ssl_cert.clone().unwrap()[..]).unwrap(), - tls_load_pkey(&self.ssl_key.clone().unwrap()[..]).unwrap() + tls_load_pkey(&self.ssl_key.clone().unwrap()[..]).unwrap(), ) .unwrap(); cfg.alpn_protocols = match &self.http_mode[..] { "1" => vec![b"http/1.1".to_vec()], "2" => vec![b"h2".to_vec()], - _ => vec![b"h2".to_vec(), b"http/1.1".to_vec()] + _ => vec![b"h2".to_vec(), b"http/1.1".to_vec()], }; cfg } @@ -102,7 +98,7 @@ pub(crate) struct WorkerExecutor; impl<F> hyper::rt::Executor<F> for WorkerExecutor where - F: std::future::Future + 'static + F: std::future::Future + 'static, { fn execute(&self, fut: F) { tokio::task::spawn_local(fut); @@ -123,14 +119,9 @@ macro_rules! build_service { let rth = rth.clone(); async move { - Ok::<_, std::convert::Infallible>($target( - rth, - callback_wrapper, - local_addr, - remote_addr, - req, - "http" - ).await) + Ok::<_, std::convert::Infallible>( + $target(rth, callback_wrapper, local_addr, remote_addr, req, "http").await, + ) } })) } @@ -153,14 +144,9 @@ macro_rules! build_service_ssl { let rth = rth.clone(); async move { - Ok::<_, std::convert::Infallible>($target( - rth, - callback_wrapper, - local_addr, - remote_addr, - req, - "https" - ).await) + Ok::<_, std::convert::Infallible>( + $target(rth, callback_wrapper, local_addr, remote_addr, req, "https").await, + ) } })) } @@ -170,13 +156,7 @@ macro_rules! build_service_ssl { macro_rules! serve_rth { ($func_name:ident, $target:expr) => { - fn $func_name( - &self, - callback: PyObject, - event_loop: &PyAny, - context: &PyAny, - signal_rx: PyObject - ) { + fn $func_name(&self, callback: PyObject, event_loop: &PyAny, context: &PyAny, signal_rx: PyObject) { pyo3_log::init(); let rt = crate::runtime::init_runtime_mt(self.config.threads, self.config.pthreads); let rth = rt.handler(); @@ -184,34 +164,30 @@ macro_rules! serve_rth { let http1_only = self.config.http_mode == "1"; let http2_only = self.config.http_mode == "2"; let http1_buffer_max = self.config.http1_buffer_max.clone(); - let callback_wrapper = crate::callbacks::CallbackWrapper::new( - callback, event_loop, context - ); + let callback_wrapper = crate::callbacks::CallbackWrapper::new(callback, event_loop, context); let worker_id = self.config.id; log::info!("Started worker-{}", worker_id); - let svc_loop = crate::runtime::run_until_complete( - rt.handler(), - event_loop, - async move { - let service = crate::workers::build_service!( - callback_wrapper, rth, $target - ); - let server = hyper::Server::from_tcp(tcp_listener).unwrap() - .http1_only(http1_only) - .http2_only(http2_only) - .http1_max_buf_size(http1_buffer_max) - .serve(service); - server.with_graceful_shutdown(async move { - Python::with_gil(|py| { - crate::runtime::into_future(signal_rx.as_ref(py)).unwrap() - }).await.unwrap(); - }).await.unwrap(); - log::info!("Stopping worker-{}", worker_id); - Ok(()) - } - ); + let svc_loop = crate::runtime::run_until_complete(rt.handler(), event_loop, async move { + let service = crate::workers::build_service!(callback_wrapper, rth, $target); + let server = hyper::Server::from_tcp(tcp_listener) + .unwrap() + .http1_only(http1_only) + .http2_only(http2_only) + .http1_max_buf_size(http1_buffer_max) + .serve(service); + server + .with_graceful_shutdown(async move { + Python::with_gil(|py| crate::runtime::into_future(signal_rx.as_ref(py)).unwrap()) + .await + .unwrap(); + }) + .await + .unwrap(); + log::info!("Stopping worker-{}", worker_id); + Ok(()) + }); match svc_loop { Ok(_) => {} @@ -226,13 +202,7 @@ macro_rules! serve_rth { macro_rules! serve_rth_ssl { ($func_name:ident, $target:expr) => { - fn $func_name( - &self, - callback: PyObject, - event_loop: &PyAny, - context: &PyAny, - signal_rx: PyObject - ) { + fn $func_name(&self, callback: PyObject, event_loop: &PyAny, context: &PyAny, signal_rx: PyObject) { pyo3_log::init(); let rt = crate::runtime::init_runtime_mt(self.config.threads, self.config.pthreads); let rth = rt.handler(); @@ -241,38 +211,29 @@ macro_rules! serve_rth_ssl { let http2_only = self.config.http_mode == "2"; let http1_buffer_max = self.config.http1_buffer_max.clone(); let tls_cfg = self.config.tls_cfg(); - let callback_wrapper = crate::callbacks::CallbackWrapper::new( - callback, event_loop, context - ); + let callback_wrapper = crate::callbacks::CallbackWrapper::new(callback, event_loop, context); let worker_id = self.config.id; log::info!("Started worker-{}", worker_id); - let svc_loop = crate::runtime::run_until_complete( - rt.handler(), - event_loop, - async move { - let service = crate::workers::build_service_ssl!( - callback_wrapper, rth, $target - ); - let server = hyper::Server::builder( - crate::tls::tls_listen( - std::sync::Arc::new(tls_cfg), tcp_listener - ) - ) - .http1_only(http1_only) - .http2_only(http2_only) - .http1_max_buf_size(http1_buffer_max) - .serve(service); - server.with_graceful_shutdown(async move { - Python::with_gil(|py| { - crate::runtime::into_future(signal_rx.as_ref(py)).unwrap() - }).await.unwrap(); - }).await.unwrap(); - log::info!("Stopping worker-{}", worker_id); - Ok(()) - } - ); + let svc_loop = crate::runtime::run_until_complete(rt.handler(), event_loop, async move { + let service = crate::workers::build_service_ssl!(callback_wrapper, rth, $target); + let server = hyper::Server::builder(crate::tls::tls_listen(std::sync::Arc::new(tls_cfg), tcp_listener)) + .http1_only(http1_only) + .http2_only(http2_only) + .http1_max_buf_size(http1_buffer_max) + .serve(service); + server + .with_graceful_shutdown(async move { + Python::with_gil(|py| crate::runtime::into_future(signal_rx.as_ref(py)).unwrap()) + .await + .unwrap(); + }) + .await + .unwrap(); + log::info!("Stopping worker-{}", worker_id); + Ok(()) + }); match svc_loop { Ok(_) => {} @@ -287,22 +248,14 @@ macro_rules! serve_rth_ssl { macro_rules! serve_wth { ($func_name: ident, $target:expr) => { - fn $func_name( - &self, - callback: PyObject, - event_loop: &PyAny, - context: &PyAny, - signal_rx: PyObject - ) { + fn $func_name(&self, callback: PyObject, event_loop: &PyAny, context: &PyAny, signal_rx: PyObject) { pyo3_log::init(); let rtm = crate::runtime::init_runtime_mt(1, 1); let worker_id = self.config.id; log::info!("Started worker-{}", worker_id); - let callback_wrapper = crate::callbacks::CallbackWrapper::new( - callback, event_loop, context - ); + let callback_wrapper = crate::callbacks::CallbackWrapper::new(callback, event_loop, context); let mut workers = vec![]; let (stx, srx) = tokio::sync::watch::channel(false); @@ -323,38 +276,36 @@ macro_rules! serve_wth { let local = tokio::task::LocalSet::new(); crate::runtime::block_on_local(rt, local, async move { - let service = crate::workers::build_service!( - callback_wrapper, rth, $target - ); - let server = hyper::Server::from_tcp(tcp_listener).unwrap() + let service = crate::workers::build_service!(callback_wrapper, rth, $target); + let server = hyper::Server::from_tcp(tcp_listener) + .unwrap() .executor(crate::workers::WorkerExecutor) .http1_only(http1_only) .http2_only(http2_only) .http1_max_buf_size(http1_buffer_max) .serve(service); - server.with_graceful_shutdown(async move { - srx.changed().await.unwrap(); - }).await.unwrap(); + server + .with_graceful_shutdown(async move { + srx.changed().await.unwrap(); + }) + .await + .unwrap(); log::info!("Stopping worker-{} runtime-{}", worker_id, thread_id + 1); }); })); - }; + } - let main_loop = crate::runtime::run_until_complete( - rtm.handler(), - event_loop, - async move { - Python::with_gil(|py| { - crate::runtime::into_future(signal_rx.as_ref(py)).unwrap() - }).await.unwrap(); - stx.send(true).unwrap(); - log::info!("Stopping worker-{}", worker_id); - while let Some(worker) = workers.pop() { - worker.join().unwrap(); - } - Ok(()) + let main_loop = crate::runtime::run_until_complete(rtm.handler(), event_loop, async move { + Python::with_gil(|py| crate::runtime::into_future(signal_rx.as_ref(py)).unwrap()) + .await + .unwrap(); + stx.send(true).unwrap(); + log::info!("Stopping worker-{}", worker_id); + while let Some(worker) = workers.pop() { + worker.join().unwrap(); } - ); + Ok(()) + }); match main_loop { Ok(_) => {} @@ -369,22 +320,14 @@ macro_rules! serve_wth { macro_rules! serve_wth_ssl { ($func_name: ident, $target:expr) => { - fn $func_name( - &self, - callback: PyObject, - event_loop: &PyAny, - context: &PyAny, - signal_rx: PyObject - ) { + fn $func_name(&self, callback: PyObject, event_loop: &PyAny, context: &PyAny, signal_rx: PyObject) { pyo3_log::init(); let rtm = crate::runtime::init_runtime_mt(1, 1); let worker_id = self.config.id; log::info!("Started worker-{}", worker_id); - let callback_wrapper = crate::callbacks::CallbackWrapper::new( - callback, event_loop, context - ); + let callback_wrapper = crate::callbacks::CallbackWrapper::new(callback, event_loop, context); let mut workers = vec![]; let (stx, srx) = tokio::sync::watch::channel(false); @@ -406,42 +349,36 @@ macro_rules! serve_wth_ssl { let local = tokio::task::LocalSet::new(); crate::runtime::block_on_local(rt, local, async move { - let service = crate::workers::build_service_ssl!( - callback_wrapper, rth, $target - ); - let server = hyper::Server::builder( - crate::tls::tls_listen( - std::sync::Arc::new(tls_cfg), tcp_listener - ) - ) - .executor(crate::workers::WorkerExecutor) - .http1_only(http1_only) - .http2_only(http2_only) - .http1_max_buf_size(http1_buffer_max) - .serve(service); - server.with_graceful_shutdown(async move { - srx.changed().await.unwrap(); - }).await.unwrap(); + let service = crate::workers::build_service_ssl!(callback_wrapper, rth, $target); + let server = + hyper::Server::builder(crate::tls::tls_listen(std::sync::Arc::new(tls_cfg), tcp_listener)) + .executor(crate::workers::WorkerExecutor) + .http1_only(http1_only) + .http2_only(http2_only) + .http1_max_buf_size(http1_buffer_max) + .serve(service); + server + .with_graceful_shutdown(async move { + srx.changed().await.unwrap(); + }) + .await + .unwrap(); log::info!("Stopping worker-{} runtime-{}", worker_id, thread_id + 1); }); })); - }; + } - let main_loop = crate::runtime::run_until_complete( - rtm.handler(), - event_loop, - async move { - Python::with_gil(|py| { - crate::runtime::into_future(signal_rx.as_ref(py)).unwrap() - }).await.unwrap(); - stx.send(true).unwrap(); - log::info!("Stopping worker-{}", worker_id); - while let Some(worker) = workers.pop() { - worker.join().unwrap(); - } - Ok(()) + let main_loop = crate::runtime::run_until_complete(rtm.handler(), event_loop, async move { + Python::with_gil(|py| crate::runtime::into_future(signal_rx.as_ref(py)).unwrap()) + .await + .unwrap(); + stx.send(true).unwrap(); + log::info!("Stopping worker-{}", worker_id); + while let Some(worker) = workers.pop() { + worker.join().unwrap(); } - ); + Ok(()) + }); match main_loop { Ok(_) => {} @@ -457,8 +394,8 @@ macro_rules! serve_wth_ssl { pub(crate) use build_service; pub(crate) use build_service_ssl; pub(crate) use serve_rth; -pub(crate) use serve_wth; pub(crate) use serve_rth_ssl; +pub(crate) use serve_wth; pub(crate) use serve_wth_ssl; pub(crate) fn init_pymodule(module: &PyModule) -> PyResult<()> { diff --git a/src/ws.rs b/src/ws.rs index bd44dded..3f892289 100644 --- a/src/ws.rs +++ b/src/ws.rs @@ -1,24 +1,24 @@ use hyper::{ - Body, - Request, - Response, - StatusCode, header::{CONNECTION, UPGRADE}, - http::response::Builder + http::response::Builder, + Body, Request, Response, StatusCode, }; +use pin_project::pin_project; +use std::{ + future::Future, + pin::Pin, + task::{Context, Poll}, +}; +use tokio::sync::mpsc; +use tokio_tungstenite::WebSocketStream; use tungstenite::{ error::ProtocolError, handshake::derive_accept_key, - protocol::{Role, WebSocketConfig} + protocol::{Role, WebSocketConfig}, }; -use pin_project::pin_project; -use std::{future::Future, pin::Pin, task::{Context, Poll}}; -use tokio_tungstenite::WebSocketStream; -use tokio::sync::mpsc; use super::utils::header_contains_value; - #[pin_project] #[derive(Debug)] pub struct HyperWebsocket { @@ -37,15 +37,9 @@ impl Future for HyperWebsocket { Poll::Ready(x) => x, }; - let upgraded = upgraded.map_err(|_| - tungstenite::Error::Protocol(ProtocolError::HandshakeIncomplete) - )?; + let upgraded = upgraded.map_err(|_| tungstenite::Error::Protocol(ProtocolError::HandshakeIncomplete))?; - let stream = WebSocketStream::from_raw_socket( - upgraded, - Role::Server, - this.config.take(), - ); + let stream = WebSocketStream::from_raw_socket(upgraded, Role::Server, this.config.take()); tokio::pin!(stream); match stream.as_mut().poll(cx) { @@ -58,18 +52,15 @@ impl Future for HyperWebsocket { pub(crate) struct UpgradeData { response_builder: Option<Builder>, response_tx: Option<mpsc::Sender<Response<Body>>>, - pub consumed: bool + pub consumed: bool, } impl UpgradeData { - pub fn new( - response_builder: Builder, - response_tx: mpsc::Sender<Response<Body>>) - -> Self { + pub fn new(response_builder: Builder, response_tx: mpsc::Sender<Response<Body>>) -> Self { Self { response_builder: Some(response_builder), response_tx: Some(response_tx), - consumed: false + consumed: false, } } @@ -79,19 +70,16 @@ impl UpgradeData { Ok(_) => { self.consumed = true; Ok(()) - }, - err => err + } + err => err, } } } #[inline] pub(crate) fn is_upgrade_request(request: &Request) -> bool { - header_contains_value( - request.headers(), CONNECTION, "Upgrade" - ) && header_contains_value( - request.headers(), UPGRADE, "websocket" - ) + header_contains_value(request.headers(), CONNECTION, "Upgrade") + && header_contains_value(request.headers(), UPGRADE, "websocket") } pub(crate) fn upgrade_intent( @@ -100,13 +88,17 @@ pub(crate) fn upgrade_intent( ) -> Result<(Builder, HyperWebsocket), ProtocolError> { let request = request.borrow_mut(); - let key = request.headers() + let key = request + .headers() .get("Sec-WebSocket-Key") .ok_or(ProtocolError::MissingSecWebSocketKey)?; - if request.headers().get("Sec-WebSocket-Version").map( - |v| v.as_bytes() - ) != Some(b"13") { + if request + .headers() + .get("Sec-WebSocket-Version") + .map(hyper::http::HeaderValue::as_bytes) + != Some(b"13") + { return Err(ProtocolError::MissingSecWebSocketVersionHeader); } diff --git a/src/wsgi/callbacks.rs b/src/wsgi/callbacks.rs index ae86cf01..e961827b 100644 --- a/src/wsgi/callbacks.rs +++ b/src/wsgi/callbacks.rs @@ -2,44 +2,35 @@ use hyper::Body; use pyo3::prelude::*; use tokio::task::JoinHandle; +use super::types::{WSGIResponseBodyIter, WSGIScope as Scope}; use crate::callbacks::CallbackWrapper; -use super::types::{WSGIScope as Scope, WSGIResponseBodyIter}; const WSGI_LIST_RESPONSE_BODY: i32 = 0; const WSGI_ITER_RESPONSE_BODY: i32 = 1; - #[inline(always)] -fn run_callback( - callback: PyObject, - scope: Scope -) -> PyResult<(i32, Vec<(String, String)>, Body)> { +fn run_callback(callback: PyObject, scope: Scope) -> PyResult<(i32, Vec<(String, String)>, Body)> { Python::with_gil(|py| { - let (status, headers, body_type, pybody) = callback.call1(py, (scope,))? - .extract::<(i32, Vec<(String, String)>, i32, PyObject)>(py)?; + let (status, headers, body_type, pybody) = + callback + .call1(py, (scope,))? + .extract::<(i32, Vec<(String, String)>, i32, PyObject)>(py)?; let body = match body_type { WSGI_LIST_RESPONSE_BODY => Body::from(pybody.extract::<Vec<u8>>(py)?), WSGI_ITER_RESPONSE_BODY => Body::wrap_stream(WSGIResponseBodyIter::new(pybody)), - _ => Body::empty() + _ => Body::empty(), }; Ok((status, headers, body)) }) } -pub(crate) fn call_rtb_http( - cb: CallbackWrapper, - scope: Scope -) -> PyResult<(i32, Vec<(String, String)>, Body)> { - run_callback(cb.callback.clone(), scope) +pub(crate) fn call_rtb_http(cb: CallbackWrapper, scope: Scope) -> PyResult<(i32, Vec<(String, String)>, Body)> { + run_callback(cb.callback, scope) } pub(crate) fn call_rtt_http( cb: CallbackWrapper, - scope: Scope + scope: Scope, ) -> JoinHandle<PyResult<(i32, Vec<(String, String)>, Body)>> { - let callback = cb.callback.clone(); - - tokio::task::spawn_blocking(move || { - run_callback(callback, scope) - }) + tokio::task::spawn_blocking(move || run_callback(cb.callback, scope)) } diff --git a/src/wsgi/http.rs b/src/wsgi/http.rs index 66cccd75..e74882a7 100644 --- a/src/wsgi/http.rs +++ b/src/wsgi/http.rs @@ -1,21 +1,18 @@ use hyper::{ - Body, - Request, - Response, - header::{SERVER as HK_SERVER, HeaderName, HeaderValue} + header::{HeaderName, HeaderValue, SERVER as HK_SERVER}, + Body, Request, Response, }; use std::net::SocketAddr; +use super::{ + callbacks::{call_rtb_http, call_rtt_http}, + types::WSGIScope as Scope, +}; use crate::{ callbacks::CallbackWrapper, - http::{HV_SERVER, response_500}, + http::{response_500, HV_SERVER}, runtime::RuntimeRef, }; -use super::{ - callbacks::{call_rtb_http, call_rtt_http}, - types::WSGIScope as Scope -}; - #[inline(always)] fn build_response(status: i32, pyheaders: Vec<(String, String)>, body: Body) -> Response<Body> { @@ -26,7 +23,7 @@ fn build_response(status: i32, pyheaders: Vec<(String, String)>, body: Body) -> for (key, val) in pyheaders { headers.append( HeaderName::from_bytes(key.as_bytes()).unwrap(), - HeaderValue::from_str(&val).unwrap() + HeaderValue::from_str(&val).unwrap(), ); } res @@ -38,14 +35,11 @@ pub(crate) async fn handle_rtt( server_addr: SocketAddr, client_addr: SocketAddr, req: Request<Body>, - scheme: &str + scheme: &str, ) -> Response<Body> { - if let Ok(res) = call_rtt_http( - callback, - Scope::new(scheme, server_addr, client_addr, req).await - ).await { + if let Ok(res) = call_rtt_http(callback, Scope::new(scheme, server_addr, client_addr, req).await).await { if let Ok((status, headers, body)) = res { - return build_response(status, headers, body) + return build_response(status, headers, body); } log::warn!("Application callable raised an exception"); } else { @@ -60,12 +54,9 @@ pub(crate) async fn handle_rtb( server_addr: SocketAddr, client_addr: SocketAddr, req: Request<Body>, - scheme: &str + scheme: &str, ) -> Response<Body> { - match call_rtb_http( - callback, - Scope::new(scheme, server_addr, client_addr, req).await - ) { + match call_rtb_http(callback, Scope::new(scheme, server_addr, client_addr, req).await) { Ok((status, headers, body)) => build_response(status, headers, body), _ => { log::warn!("Application callable raised an exception"); diff --git a/src/wsgi/serve.rs b/src/wsgi/serve.rs index 284a4cdf..0b60e8b3 100644 --- a/src/wsgi/serve.rs +++ b/src/wsgi/serve.rs @@ -1,17 +1,11 @@ use pyo3::prelude::*; -use crate::workers::{ - WorkerConfig, - serve_rth, - serve_wth, - serve_rth_ssl, - serve_wth_ssl -}; use super::http::{handle_rtb, handle_rtt}; +use crate::workers::{serve_rth, serve_rth_ssl, serve_wth, serve_wth_ssl, WorkerConfig}; -#[pyclass(module="granian._granian")] +#[pyclass(module = "granian._granian")] pub struct WSGIWorker { - config: WorkerConfig + config: WorkerConfig, } impl WSGIWorker { @@ -46,7 +40,7 @@ impl WSGIWorker { http1_buffer_max: usize, ssl_enabled: bool, ssl_cert: Option<&str>, - ssl_key: Option<&str> + ssl_key: Option<&str>, ) -> PyResult<Self> { Ok(Self { config: WorkerConfig::new( @@ -60,34 +54,22 @@ impl WSGIWorker { true, ssl_enabled, ssl_cert, - ssl_key - ) + ssl_key, + ), }) } - fn serve_rth( - &self, - callback: PyObject, - event_loop: &PyAny, - context: &PyAny, - signal_rx: PyObject - ) { + fn serve_rth(&self, callback: PyObject, event_loop: &PyAny, context: &PyAny, signal_rx: PyObject) { match self.config.ssl_enabled { false => self._serve_rth(callback, event_loop, context, signal_rx), - true => self._serve_rth_ssl(callback, event_loop, context, signal_rx) + true => self._serve_rth_ssl(callback, event_loop, context, signal_rx), } } - fn serve_wth( - &self, - callback: PyObject, - event_loop: &PyAny, - context: &PyAny, - signal_rx: PyObject - ) { + fn serve_wth(&self, callback: PyObject, event_loop: &PyAny, context: &PyAny, signal_rx: PyObject) { match self.config.ssl_enabled { false => self._serve_wth(callback, event_loop, context, signal_rx), - true => self._serve_wth_ssl(callback, event_loop, context, signal_rx) + true => self._serve_wth_ssl(callback, event_loop, context, signal_rx), } } } diff --git a/src/wsgi/types.rs b/src/wsgi/types.rs index 02c24b26..7e5c71f8 100644 --- a/src/wsgi/types.rs +++ b/src/wsgi/types.rs @@ -1,23 +1,21 @@ use futures::Stream; use hyper::{ body::Bytes, - header::{CONTENT_TYPE, CONTENT_LENGTH, HeaderMap}, - Body, - Method, - Request, - Uri, - Version + header::{HeaderMap, CONTENT_LENGTH, CONTENT_TYPE}, + Body, Method, Request, Uri, Version, }; -use pyo3::{prelude::*, types::IntoPyDict}; use pyo3::types::{PyBytes, PyDict, PyList}; -use std::{net::{IpAddr, SocketAddr}, task::{Context, Poll}}; +use pyo3::{prelude::*, types::IntoPyDict}; +use std::{ + net::{IpAddr, SocketAddr}, + task::{Context, Poll}, +}; const LINE_SPLIT: u8 = u8::from_be_bytes(*b"\n"); - #[pyclass(module = "granian._granian")] pub(crate) struct WSGIBody { - inner: Bytes + inner: Bytes, } impl WSGIBody { @@ -36,9 +34,9 @@ impl WSGIBody { match self.inner.iter().position(|&c| c == LINE_SPLIT) { Some(next_split) => { let bytes = self.inner.split_to(next_split); - Some(PyBytes::new(py, &bytes[..])) - }, - _ => None + Some(PyBytes::new(py, &bytes)) + } + _ => None, } } @@ -48,18 +46,16 @@ impl WSGIBody { None => { let bytes = self.inner.split_to(self.inner.len()); PyBytes::new(py, &bytes[..]) - }, - Some(size) => { - match size { - 0 => PyBytes::new(py, b""), - size => { - let limit = self.inner.len(); - let rsize = if size > limit { limit } else { size }; - let bytes = self.inner.split_to(rsize); - PyBytes::new(py, &bytes[..]) - } - } } + Some(size) => match size { + 0 => PyBytes::new(py, b""), + size => { + let limit = self.inner.len(); + let rsize = if size > limit { limit } else { size }; + let bytes = self.inner.split_to(rsize); + PyBytes::new(py, &bytes[..]) + } + }, } } @@ -69,16 +65,17 @@ impl WSGIBody { let bytes = self.inner.split_to(next_split); self.inner = self.inner.slice(1..); PyBytes::new(py, &bytes[..]) - }, - _ => PyBytes::new(py, b"") + } + _ => PyBytes::new(py, b""), } } #[pyo3(signature = (_hint=None))] fn readlines<'p>(&mut self, py: Python<'p>, _hint: Option<PyObject>) -> &'p PyList { - let lines: Vec<&PyBytes> = self.inner + let lines: Vec<&PyBytes> = self + .inner .split(|&c| c == LINE_SPLIT) - .map(|item| PyBytes::new(py, &item[..])) + .map(|item| PyBytes::new(py, item)) .collect(); self.inner.clear(); PyList::new(py, lines) @@ -95,28 +92,19 @@ pub(crate) struct WSGIScope { server_port: u16, client: String, headers: HeaderMap, - body: Bytes + body: Bytes, } impl WSGIScope { - pub async fn new( - scheme: &str, - server: SocketAddr, - client: SocketAddr, - request: Request<Body>, - ) -> Self { + pub async fn new(scheme: &str, server: SocketAddr, client: SocketAddr, request: Request<Body>) -> Self { let http_version = request.version(); - let method = request.method().to_owned(); - let uri = request.uri().to_owned(); - let headers = request.headers().to_owned(); + let method = request.method().clone(); + let uri = request.uri().clone(); + let headers = request.headers().clone(); let body = match method { - Method::HEAD | Method::GET | Method::OPTIONS => { Bytes::new() }, - _ => { - hyper::body::to_bytes(request) - .await - .unwrap_or(Bytes::new()) - } + Method::HEAD | Method::GET | Method::OPTIONS => Bytes::new(), + _ => hyper::body::to_bytes(request).await.unwrap_or(Bytes::new()), }; Self { @@ -128,7 +116,7 @@ impl WSGIScope { server_port: server.port(), client: client.to_string(), headers, - body + body, } } @@ -138,7 +126,7 @@ impl WSGIScope { Version::HTTP_10 => "HTTP/1", Version::HTTP_11 => "HTTP/1.1", Version::HTTP_2 => "HTTP/2", - _ => "HTTP/1" + _ => "HTTP/1", } } } @@ -157,21 +145,21 @@ impl WSGIScope { content_type, content_len, headers, - body + body, ) = py.allow_threads(|| { - let (path, query_string) = self.uri.path_and_query() + let (path, query_string) = self + .uri + .path_and_query() .map_or_else(|| ("", ""), |pq| (pq.path(), pq.query().unwrap_or(""))); let content_type = self.headers.remove(CONTENT_TYPE); let content_len = self.headers.remove(CONTENT_LENGTH); let mut headers = Vec::with_capacity(self.headers.len()); - for (key, val) in self.headers.iter() { - headers.push( - ( - format!("HTTP_{}", key.as_str().replace("-", "_").to_uppercase()), - val.to_str().unwrap_or_default() - ) - ); + for (key, val) in &self.headers { + headers.push(( + format!("HTTP_{}", key.as_str().replace('-', "_").to_uppercase()), + val.to_str().unwrap_or_default(), + )); } ( @@ -185,7 +173,7 @@ impl WSGIScope { content_type, content_len, headers, - WSGIBody::new(self.body.to_owned()) + WSGIBody::new(self.body.clone()), ) }); @@ -202,13 +190,13 @@ impl WSGIScope { if let Some(content_type) = content_type { ret.set_item( pyo3::intern!(py, "CONTENT_TYPE"), - content_type.to_str().unwrap_or_default() + content_type.to_str().unwrap_or_default(), )?; } if let Some(content_len) = content_len { ret.set_item( pyo3::intern!(py, "CONTENT_LENGTH"), - content_len.to_str().unwrap_or_default() + content_len.to_str().unwrap_or_default(), )?; } @@ -219,7 +207,7 @@ impl WSGIScope { } pub(crate) struct WSGIResponseBodyIter { - inner: PyObject + inner: PyObject, } impl WSGIResponseBodyIter { @@ -235,27 +223,20 @@ impl WSGIResponseBodyIter { impl Stream for WSGIResponseBodyIter { type Item = PyResult<Vec<u8>>; - fn poll_next( - self: std::pin::Pin<&mut Self>, - _cx: &mut Context<'_> - ) -> Poll<Option<Self::Item>> { - Python::with_gil(|py| { - match self.inner.call_method0(py, pyo3::intern!(py, "__next__")) { - Ok(chunk_obj) => { - match chunk_obj.extract::<Vec<u8>>(py) { - Ok(chunk) => Poll::Ready(Some(Ok(chunk))), - _ => { - self.close_inner(py); - Poll::Ready(None) - } - } - }, - Err(err) => { - if err.is_instance_of::<pyo3::exceptions::PyStopIteration>(py) { - self.close_inner(py); - } + fn poll_next(self: std::pin::Pin<&mut Self>, _cx: &mut Context<'_>) -> Poll<Option<Self::Item>> { + Python::with_gil(|py| match self.inner.call_method0(py, pyo3::intern!(py, "__next__")) { + Ok(chunk_obj) => match chunk_obj.extract::<Vec<u8>>(py) { + Ok(chunk) => Poll::Ready(Some(Ok(chunk))), + _ => { + self.close_inner(py); Poll::Ready(None) } + }, + Err(err) => { + if err.is_instance_of::<pyo3::exceptions::PyStopIteration>(py) { + self.close_inner(py); + } + Poll::Ready(None) } }) }

Python code formatting and linting Considering the more people contributing to the project, especially having #116, will it be appropriate to establish the formatting and linting rules for Python code? The obvious benefit is that no contributor will doubt what development tools and styles to apply. I will gladly provide a configuration for tools like `black`, `isort`, and `ruff`, for a start at least. While doing #122, I deactivated my `black` and `ruff` plugins because any file save immediately changes too many things I didn’t touch. And those two are the industry standard by now.

@Aeron it definitely makes sense. If you don't mind, I gonna open a PR myself later today so I'll cover also the Rust part. I gonna ask a comment from you on that.

2023-09-07T17:40:31

0.0

[]

emmett-framework/granian

emmett-framework__granian-61

ac12149b271539040d8a85bc80c83667eb712c1a

diff --git a/Cargo.lock b/Cargo.lock index 12cb1e53..b188b290 100644 --- a/Cargo.lock +++ b/Cargo.lock @@ -566,6 +566,8 @@ dependencies = [ [[package]] name = "pyo3-asyncio" version = "0.17.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "1febe3946b26194628f00526929ee6f8559f9e807f811257e94d4c456103be0e" dependencies = [ "futures", "once_cell", diff --git a/Cargo.toml b/Cargo.toml index 1654d752..5601986b 100644 --- a/Cargo.toml +++ b/Cargo.toml @@ -36,7 +36,7 @@ mimalloc = { version = "0.1.34", default-features = false, features = ["local_dy once_cell = "1.5" pin-project = "1.0" pyo3 = { version = "=0.17", features = ["extension-module"] } -pyo3-asyncio = { path = "lib/pyo3-asyncio", version = "0.17", features = ["tokio-runtime"] } +pyo3-asyncio = { version = "=0.17", features = ["tokio-runtime"] } pyo3-log = "=0.7" rustls-pemfile = "1.0" socket2 = { version = "0.4", features = ["all"] } diff --git a/lib/pyo3-asyncio/.githooks/pre-commit b/lib/pyo3-asyncio/.githooks/pre-commit deleted file mode 100755 index a576dfe1..00000000 --- a/lib/pyo3-asyncio/.githooks/pre-commit +++ /dev/null @@ -1,3 +0,0 @@ -#!/bin/bash - -cargo check --all-targets --all-features \ No newline at end of file diff --git a/lib/pyo3-asyncio/.githooks/pre-push b/lib/pyo3-asyncio/.githooks/pre-push deleted file mode 100755 index 19c4269e..00000000 --- a/lib/pyo3-asyncio/.githooks/pre-push +++ /dev/null @@ -1,4 +0,0 @@ -#!/bin/bash - -make clippy -make test \ No newline at end of file diff --git a/lib/pyo3-asyncio/.github/dependabot.yml b/lib/pyo3-asyncio/.github/dependabot.yml deleted file mode 100644 index c1ce3c3b..00000000 --- a/lib/pyo3-asyncio/.github/dependabot.yml +++ /dev/null @@ -1,11 +0,0 @@ -# To get started with Dependabot version updates, you'll need to specify which -# package ecosystems to update and where the package manifests are located. -# Please see the documentation for all configuration options: -# https://help.github.com/github/administering-a-repository/configuration-options-for-dependency-updates - -version: 2 -updates: - - package-ecosystem: "cargo" # See documentation for possible values - directory: "/" # Location of package manifests - schedule: - interval: "weekly" diff --git a/lib/pyo3-asyncio/.github/issue_template.md b/lib/pyo3-asyncio/.github/issue_template.md deleted file mode 100644 index b582e56a..00000000 --- a/lib/pyo3-asyncio/.github/issue_template.md +++ /dev/null @@ -1,18 +0,0 @@ -## 🐛 Bug Reports - -When reporting a bug, please provide the following information. If this is not a bug report you can just discard this template. - -### 🌍 Environment - - - Your operating system and version: - - Your python version: - - How did you install python (e.g. apt or pyenv)? Did you use a virtualenv?: - - Your Rust version (`rustc --version`): - - Your PyO3 version: - - Have you tried using latest PyO3 master (replace `version = "0.x.y"` with `git = "https://github.com/awestlake87/pyo3-asyncio")?`: - -### 💥 Reproducing - -Please provide a [minimal working example](https://stackoverflow.com/help/mcve). This means both the Rust code and the Python. - -Please also write what exact flags are required to reproduce your results. diff --git a/lib/pyo3-asyncio/.github/pull_request_template.md b/lib/pyo3-asyncio/.github/pull_request_template.md deleted file mode 100644 index 5798dc54..00000000 --- a/lib/pyo3-asyncio/.github/pull_request_template.md +++ /dev/null @@ -1,12 +0,0 @@ -Thank you for contributing to pyo3-asyncio! - -Please consider adding the following to your pull request: - - an entry in CHANGELOG.md - - docs to all new functions and / or detail in the guide - - tests for all new or changed functions - -Be aware the CI pipeline will check your pull request for the following: - - Rust tests (Just `cargo test`) - - Rust lints (`make clippy`) - - Rust formatting (`cargo fmt`) - - Python formatting (`black . --check`. You can install black with `pip install black`) diff --git a/lib/pyo3-asyncio/.github/workflows/ci.yml b/lib/pyo3-asyncio/.github/workflows/ci.yml deleted file mode 100644 index d828197d..00000000 --- a/lib/pyo3-asyncio/.github/workflows/ci.yml +++ /dev/null @@ -1,199 +0,0 @@ -name: CI - -on: - push: - branches: - - master - pull_request: - branches: - - master - -env: - CARGO_TERM_COLOR: always - -jobs: - fmt: - runs-on: ubuntu-latest - steps: - - uses: actions/checkout@v2 - - uses: actions/setup-python@v2 - - run: pip install black==22.8.0 - - uses: actions-rs/toolchain@v1 - with: - toolchain: stable - profile: minimal - components: rustfmt - - name: Check python formatting (black) - run: black --check . - - name: Check rust formatting (rustfmt) - run: cargo fmt --all -- --check - - clippy: - runs-on: ubuntu-latest - steps: - - uses: actions/checkout@v2 - - uses: actions-rs/toolchain@v1 - with: - toolchain: stable - profile: minimal - components: clippy - - run: make clippy - - build: - needs: [fmt] # don't wait for clippy as fails rarely and takes longer - name: python${{ matrix.python-version }}-${{ matrix.platform.python-architecture }} ${{ matrix.platform.os }} ${{ matrix.msrv }} - runs-on: ${{ matrix.platform.os }} - strategy: - fail-fast: false # If one platform fails, allow the rest to keep testing. - matrix: - rust: [stable] - python-version: - [ - "3.7", - "3.8", - "3.9", - "3.10", - "3.11-dev", - "pypy-3.7", - "pypy-3.8", - "pypy-3.9", - ] - platform: - [ - { - os: "macOS-latest", - python-architecture: "x64", - rust-target: "x86_64-apple-darwin", - }, - { - os: "ubuntu-latest", - python-architecture: "x64", - rust-target: "x86_64-unknown-linux-gnu", - }, - { - os: "windows-latest", - python-architecture: "x64", - rust-target: "x86_64-pc-windows-msvc", - }, - { - os: "windows-latest", - python-architecture: "x86", - rust-target: "i686-pc-windows-msvc", - }, - ] - exclude: - # PyPy doesn't release 32-bit Windows builds any more - - python-version: pypy-3.7 - platform: { os: "windows-latest", python-architecture: "x86" } - - python-version: pypy-3.8 - platform: { os: "windows-latest", python-architecture: "x86" } - - python-version: pypy-3.9 - platform: { os: "windows-latest", python-architecture: "x86" } - include: - # Test minimal supported Rust version - - rust: 1.48.0 - python-version: "3.10" - platform: - { - os: "ubuntu-latest", - python-architecture: "x64", - rust-target: "x86_64-unknown-linux-gnu", - } - msrv: "MSRV" - - # Test the `nightly` feature - - rust: nightly - python-version: "3.10" - platform: - { - os: "ubuntu-latest", - python-architecture: "x64", - rust-target: "x86_64-unknown-linux-gnu", - } - msrv: "nightly" - extra_features: "nightly" - - steps: - - uses: actions/checkout@v2 - - - name: Set up Python ${{ matrix.python-version }} - uses: actions/setup-python@v2 - with: - python-version: ${{ matrix.python-version }} - architecture: ${{ matrix.platform.python-architecture }} - - - name: Install Rust toolchain - uses: actions-rs/toolchain@v1 - with: - toolchain: ${{ matrix.rust }} - target: ${{ matrix.platform.rust-target }} - profile: minimal - default: true - - - if: matrix.platform.os == 'ubuntu-latest' - name: Prepare LD_LIBRARY_PATH (Ubuntu only) - run: echo LD_LIBRARY_PATH=${pythonLocation}/lib >> $GITHUB_ENV - - - if: matrix.msrv == 'MSRV' - name: Prepare minimal package versions (MSRV only) - run: | - set -x - cargo update -p tokio --precise 1.13.1 - cargo update -p indexmap --precise 1.6.2 - cargo update -p parking_lot:0.12.1 --precise 0.11.2 - cargo update -p async-global-executor --precise 2.2.0 - cargo update -p once_cell --precise 1.14.0 - - - name: Build (no features) - run: cargo build --no-default-features --verbose --target ${{ matrix.platform.rust-target }} - - - name: Build - run: cargo build --features=${{env.features}} --verbose --target ${{ matrix.platform.rust-target }} - - # uvloop doesn't compile under Windows, Python 3.11-dev, and PyPy - - if: ${{ matrix.platform.os != 'windows-latest' && matrix.python-version != '3.11-dev' && !startsWith(matrix.python-version, 'pypy') }} - name: Install pyo3-asyncio test dependencies - run: | - python -m pip install -U uvloop - - - if: ${{ matrix.msrv != 'MSRV' && matrix.python-version != '3.11-dev' && !startsWith(matrix.python-version, 'pypy') }} - name: Test - run: cargo test --all-features --target ${{ matrix.platform.rust-target }} - - - if: ${{ matrix.msrv == 'MSRV' && matrix.python-version != '3.11-dev' && !startsWith(matrix.python-version, 'pypy') }} - name: Test (MSRV, --no-default-features) - run: cargo test --no-default-features --features tokio-runtime,async-std-runtime,attributes,unstable-streams --target ${{ matrix.platform.rust-target }} - - env: - RUST_BACKTRACE: 1 - RUSTFLAGS: "-D warnings" - # TODO: this is a hack to workaround compile_error! warnings about auto-initialize on PyPy - # Once cargo's `resolver = "2"` is stable (~ MSRV Rust 1.52), remove this. - PYO3_CI: 1 - - coverage: - needs: [fmt] - runs-on: ubuntu-latest - steps: - - uses: actions/checkout@v2 - - uses: actions-rs/toolchain@v1 - with: - toolchain: nightly - override: true - profile: minimal - - name: Install pyo3-asyncio test dependencies - run: | - python -m pip install -U uvloop - - uses: actions-rs/cargo@v1 - with: - command: test - args: --all-features - env: - CARGO_INCREMENTAL: 0 - RUSTFLAGS: "-Zprofile -Ccodegen-units=1 -Cinline-threshold=0 -Clink-dead-code -Coverflow-checks=off" - RUSTDOCFLAGS: "-Zprofile -Ccodegen-units=1 -Cinline-threshold=0 -Clink-dead-code -Coverflow-checks=off" - - uses: actions-rs/[email protected] - id: coverage - - uses: codecov/codecov-action@v1 - with: - file: ${{ steps.coverage.outputs.report }} diff --git a/lib/pyo3-asyncio/.github/workflows/guide.yml b/lib/pyo3-asyncio/.github/workflows/guide.yml deleted file mode 100644 index 461328cd..00000000 --- a/lib/pyo3-asyncio/.github/workflows/guide.yml +++ /dev/null @@ -1,61 +0,0 @@ -name: gh-pages - -on: - push: - branches: - - master - release: - types: [published] - -env: - CARGO_TERM_COLOR: always - -jobs: - deploy: - runs-on: ubuntu-latest - outputs: - tag_name: ${{ steps.prepare_tag.outputs.tag_name }} - steps: - - uses: actions/checkout@v2 - - # This adds the docs to gh-pages-build/doc - - name: Build the doc - run: | - cargo doc --no-deps --all-features - mkdir -p gh-pages-build - cp -r target/doc gh-pages-build/doc - echo "<meta http-equiv=refresh content=0;url=pyo3_asyncio/index.html>" > gh-pages-build/doc/index.html - - - name: Prepare tag - id: prepare_tag - run: | - TAG_NAME="${GITHUB_REF##*/}" - echo "::set-output name=tag_name::${TAG_NAME}" - - - name: Deploy - uses: peaceiris/[email protected] - with: - github_token: ${{ secrets.GITHUB_TOKEN }} - publish_dir: ./gh-pages-build/ - destination_dir: ${{ steps.prepare_tag.outputs.tag_name }} - full_commit_message: 'Upload documentation for ${{ steps.prepare_tag.outputs.tag_name }}' - - release: - needs: deploy - runs-on: ubuntu-latest - if: ${{ github.event_name == 'release' }} - steps: - - name: Create latest tag redirect - env: - TAG_NAME: ${{ needs.deploy.outputs.tag_name }} - run: | - mkdir public - echo "<meta http-equiv=refresh content=0;url='$TAG_NAME/'>" > public/index.html - - - name: Deploy - uses: peaceiris/[email protected] - with: - github_token: ${{ secrets.GITHUB_TOKEN }} - publish_dir: ./public/ - full_commit_message: 'Release ${{ needs.deploy.outputs.tag_name }}' - keep_files: true diff --git a/lib/pyo3-asyncio/.gitignore b/lib/pyo3-asyncio/.gitignore deleted file mode 100644 index f2f9e58e..00000000 --- a/lib/pyo3-asyncio/.gitignore +++ /dev/null @@ -1,2 +0,0 @@ -target -Cargo.lock \ No newline at end of file diff --git a/lib/pyo3-asyncio/Cargo.toml b/lib/pyo3-asyncio/Cargo.toml deleted file mode 100644 index 3a612c47..00000000 --- a/lib/pyo3-asyncio/Cargo.toml +++ /dev/null @@ -1,129 +0,0 @@ -[package] -name = "pyo3-asyncio" -description = "PyO3 utilities for Python's Asyncio library" -version = "0.17.0" -authors = ["Andrew J Westlake <[email protected]>"] -readme = "README.md" -keywords = ["pyo3", "python", "ffi", "async", "asyncio"] -homepage = "https://github.com/awestlake87/pyo3-asyncio" -repository = "https://github.com/awestlake87/pyo3-asyncio" -documentation = "https://docs.rs/crate/pyo3-asyncio/" -categories = ["api-bindings", "development-tools::ffi"] -license = "Apache-2.0" -exclude = ["/.gitignore", "/codecov.yml", "/Makefile"] -edition = "2018" - -[features] -async-std-runtime = ["async-std"] -attributes = [] -testing = ["clap", "inventory"] -tokio-runtime = ["tokio"] -unstable-streams = ["async-channel"] -default = [] - -[package.metadata.docs.rs] -features = ["attributes", "testing", "async-std-runtime", "tokio-runtime"] - -[[example]] -name = "async_std" -path = "examples/async_std.rs" -required-features = ["attributes", "async-std-runtime"] - -[[example]] -name = "tokio" -path = "examples/tokio.rs" -required-features = ["attributes", "tokio-runtime"] - -[[example]] -name = "tokio_current_thread" -path = "examples/tokio_current_thread.rs" -required-features = ["attributes", "tokio-runtime"] - -[[example]] -name = "tokio_multi_thread" -path = "examples/tokio_multi_thread.rs" -required-features = ["attributes", "tokio-runtime"] - - -[[test]] -name = "test_async_std_asyncio" -path = "pytests/test_async_std_asyncio.rs" -harness = false -required-features = ["async-std-runtime", "testing", "attributes"] - -[[test]] -name = "test_async_std_run_forever" -path = "pytests/test_async_std_run_forever.rs" -harness = false -required-features = ["async-std-runtime", "testing"] - -[[test]] -name = "test_tokio_current_thread_asyncio" -path = "pytests/test_tokio_current_thread_asyncio.rs" -harness = false -required-features = ["tokio-runtime", "testing", "attributes"] - -[[test]] -name = "test_tokio_current_thread_run_forever" -path = "pytests/test_tokio_current_thread_run_forever.rs" -harness = false -required-features = ["tokio-runtime", "testing"] - -[[test]] -name = "test_tokio_multi_thread_asyncio" -path = "pytests/test_tokio_multi_thread_asyncio.rs" -harness = false -required-features = ["tokio-runtime", "testing", "attributes"] - -[[test]] -name = "test_tokio_multi_thread_run_forever" -path = "pytests/test_tokio_multi_thread_run_forever.rs" -harness = false -required-features = ["tokio-runtime", "testing"] - -[[test]] -name = "test_async_std_uvloop" -path = "pytests/test_async_std_uvloop.rs" -harness = false -required-features = ["async-std-runtime", "testing"] - -[[test]] -name = "test_tokio_current_thread_uvloop" -path = "pytests/test_tokio_current_thread_uvloop.rs" -harness = false -required-features = ["tokio-runtime", "testing"] - -[[test]] -name = "test_tokio_multi_thread_uvloop" -path = "pytests/test_tokio_multi_thread_uvloop.rs" -harness = false -required-features = ["tokio-runtime", "testing"] - - -[[test]] -name = "test_race_condition_regression" -path = "pytests/test_race_condition_regression.rs" -harness = false -required-features = ["async-std-runtime", "testing"] - -[dependencies] -async-channel = { version = "1.6", optional = true } -clap = { version = "3.2", optional = true } -futures = "0.3" -inventory = { version = "0.3", optional = true } -once_cell = "1.14" -pin-project-lite = "0.2" -pyo3 = "0.17" - -[dev-dependencies] -pyo3 = { version = "0.17", features = ["macros"] } - -[dependencies.async-std] -version = "1.12" -features = ["unstable"] -optional = true - -[dependencies.tokio] -version = "1.13" -features = ["full"] -optional = true diff --git a/lib/pyo3-asyncio/Code-of-Conduct.md b/lib/pyo3-asyncio/Code-of-Conduct.md deleted file mode 100644 index c06c6254..00000000 --- a/lib/pyo3-asyncio/Code-of-Conduct.md +++ /dev/null @@ -1,77 +0,0 @@ -# Contributor Covenant Code of Conduct - -## Our Pledge - -In the interest of fostering an open and welcoming environment, we as -contributors and maintainers pledge to making participation in our project and -our community a harassment-free experience for everyone, regardless of age, body -size, disability, ethnicity, sex characteristics, gender identity and expression, -level of experience, education, socio-economic status, nationality, personal -appearance, race, religion, or sexual identity and orientation. - -## Our Standards - -Examples of behavior that contributes to creating a positive environment -include: - -* Using welcoming and inclusive language -* Being respectful of differing viewpoints and experiences -* Gracefully accepting constructive criticism -* Focusing on what is best for the community -* Showing empathy towards other community members - -Examples of unacceptable behavior by participants include: - -* The use of sexualized language or imagery and unwelcome sexual attention or - advances -* Trolling, insulting/derogatory comments, and personal or political attacks -* Public or private harassment -* Publishing others' private information, such as a physical or electronic - address, without explicit permission -* Other conduct which could reasonably be considered inappropriate in a - professional setting - -## Our Responsibilities - -Project maintainers are responsible for clarifying the standards of acceptable -behavior and are expected to take appropriate and fair corrective action in -response to any instances of unacceptable behavior. - -Project maintainers have the right and responsibility to remove, edit, or -reject comments, commits, code, wiki edits, issues, and other contributions -that are not aligned to this Code of Conduct, or to ban temporarily or -permanently any contributor for other behaviors that they deem inappropriate, -threatening, offensive, or harmful. - -## Scope - -This Code of Conduct applies both within project spaces and in public spaces -when an individual is representing the project or its community. Examples of -representing a project or community include using an official project e-mail -address, posting via an official social media account, or acting as an appointed -representative at an online or offline event. Representation of a project may be -further defined and clarified by project maintainers. - -## Enforcement - -Instances of abusive, harassing, or otherwise unacceptable behavior may be -reported by contacting the project team. All -complaints will be reviewed and investigated and will result in a response that -is deemed necessary and appropriate to the circumstances. The project team is -obligated to maintain confidentiality with regard to the reporter of an incident. -Further details of specific enforcement policies may be posted separately. - -Project maintainers who do not follow or enforce the Code of Conduct in good -faith may face temporary or permanent repercussions as determined by other -members of the project's leadership. - -## Attribution - -This Code of Conduct is adapted from the [Contributor Covenant][homepage], version 1.4, -available at https://www.contributor-covenant.org/version/1/4/code-of-conduct.html - -[homepage]: https://www.contributor-covenant.org - -For answers to common questions about this code of conduct, see -https://www.contributor-covenant.org/faq - diff --git a/lib/pyo3-asyncio/Contributing.md b/lib/pyo3-asyncio/Contributing.md deleted file mode 100644 index 2dc01cb6..00000000 --- a/lib/pyo3-asyncio/Contributing.md +++ /dev/null @@ -1,52 +0,0 @@ -# Contributing - -Thank you for your interest in contributing to PyO3! All are welcome - please consider reading our [Code of Conduct](Code-of-Conduct.md) to keep our community positive and inclusive. - -If you are searching for ideas how to contribute, please read the "Getting started contributing" section. Once you've found an issue to contribute to, you may find the section "Writing pull requests" helpful. - -## Getting started contributing - -Please join in with any part of PyO3 which interests you. We use Github issues to record all bugs and ideas. Feel free to request an issue to be assigned to you if you want to work on it. - -The following sections also contain specific ideas on where to start contributing to PyO3. - -### Help users identify bugs - -The [PyO3 Gitter channel](https://gitter.im/PyO3/Lobby) is very active with users who are new to PyO3, and often completely new to Rust. Helping them debug is a great way to get experience with the PyO3 codebase. - -Helping others often reveals bugs, documentation weaknesses, and missing APIs. It's a good idea to open Github issues for these immediately so the resolution can be designed and implemented! - -### Review pull requests - -Everybody is welcome to submit comments on open PRs. Please help ensure new PyO3 APIs are safe, performant, tidy, and easy to use! - -## Writing pull requests - -Here are a few things to note when you are writing PRs. - -### Continuous Integration - -The PyO3 Asyncio repo uses Github Actions. PRs are blocked from merging if CI is not successful. - -Formatting, linting and tests are checked for all Rust and Python code. In addition, all warnings in Rust code are disallowed (using `RUSTFLAGS="-D warnings"`). - -Tests run with all supported Python versions with the latest stable Rust compiler, as well as for Python 3.9 with the minimum supported Rust version. - -### Minimum supported Rust version - -PyO3 aims to make use of up-to-date Rust language features to keep the implementation as efficient as possible. - -However, there will always be support for at least the last few Rust compiler versions, so that users have time to update. - -If your PR needs to bump the minimum supported Rust version, this is acceptable with the following conditions: -- Any changes which require a more recent version than what is [currently available on stable Red Hat Enterprise Linux](https://access.redhat.com/documentation/en-us/red_hat_developer_tools/1/) will be postponed. (This is to allow package managers to update support for newer `rustc` versions; RHEL was arbitrarily picked because their update policy is clear.) -- You might be asked to do extra work to tidy up other parts of the PyO3 codebase which can use the compiler version bump :) - - -## Git Hooks (Recommended) - -Using the project's githooks are recommended to prevent CI from failing for trivial reasons such as build checks or integration tests failing. Enabling the hooks should run `cargo check --all-targets` for every commit and `cargo test` for every push. - -``` -git config core.hookspath .githooks -``` \ No newline at end of file diff --git a/lib/pyo3-asyncio/LICENSE b/lib/pyo3-asyncio/LICENSE deleted file mode 100644 index fca31990..00000000 --- a/lib/pyo3-asyncio/LICENSE +++ /dev/null @@ -1,189 +0,0 @@ - Copyright (c) 2017-present PyO3 Project and Contributors. https://github.com/PyO3 - - Licensed under the Apache License, Version 2.0 (the "License"); - you may not use this file except in compliance with the License. - You may obtain a copy of the License at - - http://www.apache.org/licenses/LICENSE-2.0 - - Unless required by applicable law or agreed to in writing, software - distributed under the License is distributed on an "AS IS" BASIS, - WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. - See the License for the specific language governing permissions and - limitations under the License. - - - Apache License - Version 2.0, January 2004 - http://www.apache.org/licenses/ - - TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION - - 1. Definitions. - - "License" shall mean the terms and conditions for use, reproduction, - and distribution as defined by Sections 1 through 9 of this document. - - "Licensor" shall mean the copyright owner or entity authorized by - the copyright owner that is granting the License. - - "Legal Entity" shall mean the union of the acting entity and all - other entities that control, are controlled by, or are under common - control with that entity. For the purposes of this definition, - "control" means (i) the power, direct or indirect, to cause the - direction or management of such entity, whether by contract or - otherwise, or (ii) ownership of fifty percent (50%) or more of the - outstanding shares, or (iii) beneficial ownership of such entity. - - "You" (or "Your") shall mean an individual or Legal Entity - exercising permissions granted by this License. - - "Source" form shall mean the preferred form for making modifications, - including but not limited to software source code, documentation - source, and configuration files. - - "Object" form shall mean any form resulting from mechanical - transformation or translation of a Source form, including but - not limited to compiled object code, generated documentation, - and conversions to other media types. - - "Work" shall mean the work of authorship, whether in Source or - Object form, made available under the License, as indicated by a - copyright notice that is included in or attached to the work - (an example is provided in the Appendix below). - - "Derivative Works" shall mean any work, whether in Source or Object - form, that is based on (or derived from) the Work and for which the - editorial revisions, annotations, elaborations, or other modifications - represent, as a whole, an original work of authorship. For the purposes - of this License, Derivative Works shall not include works that remain - separable from, or merely link (or bind by name) to the interfaces of, - the Work and Derivative Works thereof. - - "Contribution" shall mean any work of authorship, including - the original version of the Work and any modifications or additions - to that Work or Derivative Works thereof, that is intentionally - submitted to Licensor for inclusion in the Work by the copyright owner - or by an individual or Legal Entity authorized to submit on behalf of - the copyright owner. For the purposes of this definition, "submitted" - means any form of electronic, verbal, or written communication sent - to the Licensor or its representatives, including but not limited to - communication on electronic mailing lists, source code control systems, - and issue tracking systems that are managed by, or on behalf of, the - Licensor for the purpose of discussing and improving the Work, but - excluding communication that is conspicuously marked or otherwise - designated in writing by the copyright owner as "Not a Contribution." - - "Contributor" shall mean Licensor and any individual or Legal Entity - on behalf of whom a Contribution has been received by Licensor and - subsequently incorporated within the Work. - - 2. Grant of Copyright License. Subject to the terms and conditions of - this License, each Contributor hereby grants to You a perpetual, - worldwide, non-exclusive, no-charge, royalty-free, irrevocable - copyright license to reproduce, prepare Derivative Works of, - publicly display, publicly perform, sublicense, and distribute the - Work and such Derivative Works in Source or Object form. - - 3. Grant of Patent License. Subject to the terms and conditions of - this License, each Contributor hereby grants to You a perpetual, - worldwide, non-exclusive, no-charge, royalty-free, irrevocable - (except as stated in this section) patent license to make, have made, - use, offer to sell, sell, import, and otherwise transfer the Work, - where such license applies only to those patent claims licensable - by such Contributor that are necessarily infringed by their - Contribution(s) alone or by combination of their Contribution(s) - with the Work to which such Contribution(s) was submitted. If You - institute patent litigation against any entity (including a - cross-claim or counterclaim in a lawsuit) alleging that the Work - or a Contribution incorporated within the Work constitutes direct - or contributory patent infringement, then any patent licenses - granted to You under this License for that Work shall terminate - as of the date such litigation is filed. - - 4. Redistribution. You may reproduce and distribute copies of the - Work or Derivative Works thereof in any medium, with or without - modifications, and in Source or Object form, provided that You - meet the following conditions: - - (a) You must give any other recipients of the Work or - Derivative Works a copy of this License; and - - (b) You must cause any modified files to carry prominent notices - stating that You changed the files; and - - (c) You must retain, in the Source form of any Derivative Works - that You distribute, all copyright, patent, trademark, and - attribution notices from the Source form of the Work, - excluding those notices that do not pertain to any part of - the Derivative Works; and - - (d) If the Work includes a "NOTICE" text file as part of its - distribution, then any Derivative Works that You distribute must - include a readable copy of the attribution notices contained - within such NOTICE file, excluding those notices that do not - pertain to any part of the Derivative Works, in at least one - of the following places: within a NOTICE text file distributed - as part of the Derivative Works; within the Source form or - documentation, if provided along with the Derivative Works; or, - within a display generated by the Derivative Works, if and - wherever such third-party notices normally appear. The contents - of the NOTICE file are for informational purposes only and - do not modify the License. You may add Your own attribution - notices within Derivative Works that You distribute, alongside - or as an addendum to the NOTICE text from the Work, provided - that such additional attribution notices cannot be construed - as modifying the License. - - You may add Your own copyright statement to Your modifications and - may provide additional or different license terms and conditions - for use, reproduction, or distribution of Your modifications, or - for any such Derivative Works as a whole, provided Your use, - reproduction, and distribution of the Work otherwise complies with - the conditions stated in this License. - - 5. Submission of Contributions. Unless You explicitly state otherwise, - any Contribution intentionally submitted for inclusion in the Work - by You to the Licensor shall be under the terms and conditions of - this License, without any additional terms or conditions. - Notwithstanding the above, nothing herein shall supersede or modify - the terms of any separate license agreement you may have executed - with Licensor regarding such Contributions. - - 6. Trademarks. This License does not grant permission to use the trade - names, trademarks, service marks, or product names of the Licensor, - except as required for reasonable and customary use in describing the - origin of the Work and reproducing the content of the NOTICE file. - - 7. Disclaimer of Warranty. Unless required by applicable law or - agreed to in writing, Licensor provides the Work (and each - Contributor provides its Contributions) on an "AS IS" BASIS, - WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or - implied, including, without limitation, any warranties or conditions - of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A - PARTICULAR PURPOSE. You are solely responsible for determining the - appropriateness of using or redistributing the Work and assume any - risks associated with Your exercise of permissions under this License. - - 8. Limitation of Liability. In no event and under no legal theory, - whether in tort (including negligence), contract, or otherwise, - unless required by applicable law (such as deliberate and grossly - negligent acts) or agreed to in writing, shall any Contributor be - liable to You for damages, including any direct, indirect, special, - incidental, or consequential damages of any character arising as a - result of this License or out of the use or inability to use the - Work (including but not limited to damages for loss of goodwill, - work stoppage, computer failure or malfunction, or any and all - other commercial damages or losses), even if such Contributor - has been advised of the possibility of such damages. - - 9. Accepting Warranty or Additional Liability. While redistributing - the Work or Derivative Works thereof, You may choose to offer, - and charge a fee for, acceptance of support, warranty, indemnity, - or other liability obligations and/or rights consistent with this - License. However, in accepting such obligations, You may act only - on Your own behalf and on Your sole responsibility, not on behalf - of any other Contributor, and only if You agree to indemnify, - defend, and hold each Contributor harmless for any liability - incurred by, or claims asserted against, such Contributor by reason - of your accepting any such warranty or additional liability. diff --git a/lib/pyo3-asyncio/Makefile b/lib/pyo3-asyncio/Makefile deleted file mode 100644 index 63fe41b2..00000000 --- a/lib/pyo3-asyncio/Makefile +++ /dev/null @@ -1,24 +0,0 @@ - -clippy: - @touch src/lib.rs # Touching file to ensure that cargo clippy will re-check the project - cargo clippy --tests -- -Dwarnings - cargo clippy --tests -- -Dwarnings - # for example in examples/*; do cargo clippy --manifest-path $$example/Cargo.toml -- -Dwarnings || exit 1; done - -fmt: - cargo fmt --all -- --check - -lint: fmt clippy - @true - -test: lint - cargo test --all-features - -test-feature-powerset: lint - cargo install cargo-hack - cargo hack test --feature-powerset - -publish: test-feature-powerset - cargo publish --manifest-path pyo3-asyncio-macros/Cargo.toml - sleep 30 # wait for crates.io to update - cargo publish diff --git a/lib/pyo3-asyncio/README.md b/lib/pyo3-asyncio/README.md deleted file mode 100644 index c15e6c69..00000000 --- a/lib/pyo3-asyncio/README.md +++ /dev/null @@ -1,703 +0,0 @@ -# PyO3 Asyncio - -[![Actions Status](https://github.com/awestlake87/pyo3-asyncio/workflows/CI/badge.svg)](https://github.com/awestlake87/pyo3-asyncio/actions) -[![codecov](https://codecov.io/gh/awestlake87/pyo3-asyncio/branch/master/graph/badge.svg)](https://codecov.io/gh/awestlake87/pyo3-asyncio) -[![crates.io](https://img.shields.io/crates/v/pyo3-asyncio)](https://crates.io/crates/pyo3-asyncio) -[![minimum rustc 1.48](https://img.shields.io/badge/rustc-1.48+-blue.svg)](https://rust-lang.github.io/rfcs/2495-min-rust-version.html) - -[Rust](http://www.rust-lang.org/) bindings for [Python](https://www.python.org/)'s [Asyncio Library](https://docs.python.org/3/library/asyncio.html). This crate facilitates interactions between Rust Futures and Python Coroutines and manages the lifecycle of their corresponding event loops. - -- PyO3 Project: [Homepage](https://pyo3.rs/) | [GitHub](https://github.com/PyO3/pyo3) - -- PyO3 Asyncio API Documentation: [stable](https://docs.rs/pyo3-asyncio/) | [master](https://awestlake87.github.io/pyo3-asyncio/master/doc) - -- Guide for Async / Await [stable](https://pyo3.rs/v0.13.2/ecosystem/async-await.html) | [main](https://pyo3.rs/main/ecosystem/async-await.html) - -- Contributing Notes: [github](https://github.com/awestlake87/pyo3-asyncio/blob/master/Contributing.md) - -> PyO3 Asyncio is a _brand new_ part of the broader PyO3 ecosystem. Feel free to open any issues for feature requests or bugfixes for this crate. - -**If you're a new-comer, the best way to get started is to read through the primer below! For `v0.13` and `v0.14` users I highly recommend reading through the [migration section](#migration-guide) to get a general idea of what's changed in `v0.14` and `v0.15`.** - -## Usage - -Like PyO3, PyO3 Asyncio supports the following software versions: - -- Python 3.7 and up (CPython and PyPy) -- Rust 1.48 and up - -## PyO3 Asyncio Primer - -If you are working with a Python library that makes use of async functions or wish to provide -Python bindings for an async Rust library, [`pyo3-asyncio`](https://github.com/awestlake87/pyo3-asyncio) -likely has the tools you need. It provides conversions between async functions in both Python and -Rust and was designed with first-class support for popular Rust runtimes such as -[`tokio`](https://tokio.rs/) and [`async-std`](https://async.rs/). In addition, all async Python -code runs on the default `asyncio` event loop, so `pyo3-asyncio` should work just fine with existing -Python libraries. - -In the following sections, we'll give a general overview of `pyo3-asyncio` explaining how to call -async Python functions with PyO3, how to call async Rust functions from Python, and how to configure -your codebase to manage the runtimes of both. - -### Quickstart - -Here are some examples to get you started right away! A more detailed breakdown -of the concepts in these examples can be found in the following sections. - -#### Rust Applications - -Here we initialize the runtime, import Python's `asyncio` library and run the given future to completion using Python's default `EventLoop` and `async-std`. Inside the future, we convert `asyncio` sleep into a Rust future and await it. - -```toml -# Cargo.toml dependencies -[dependencies] -pyo3 = { version = "0.17" } -pyo3-asyncio = { version = "0.17", features = ["attributes", "async-std-runtime"] } -async-std = "1.9" -``` - -```rust -//! main.rs - -use pyo3::prelude::*; - -#[pyo3_asyncio::async_std::main] -async fn main() -> PyResult<()> { - let fut = Python::with_gil(|py| { - let asyncio = py.import("asyncio")?; - // convert asyncio.sleep into a Rust Future - pyo3_asyncio::async_std::into_future(asyncio.call_method1("sleep", (1.into_py(py),))?) - })?; - - fut.await?; - - Ok(()) -} -``` - -The same application can be written to use `tokio` instead using the `#[pyo3_asyncio::tokio::main]` -attribute. - -```toml -# Cargo.toml dependencies -[dependencies] -pyo3 = { version = "0.17" } -pyo3-asyncio = { version = "0.17", features = ["attributes", "tokio-runtime"] } -tokio = "1.9" -``` - -```rust -//! main.rs - -use pyo3::prelude::*; - -#[pyo3_asyncio::tokio::main] -async fn main() -> PyResult<()> { - let fut = Python::with_gil(|py| { - let asyncio = py.import("asyncio")?; - // convert asyncio.sleep into a Rust Future - pyo3_asyncio::tokio::into_future(asyncio.call_method1("sleep", (1.into_py(py),))?) - })?; - - fut.await?; - - Ok(()) -} -``` - -More details on the usage of this library can be found in the [API docs](https://awestlake87.github.io/pyo3-asyncio/master/doc) and the primer below. - -#### PyO3 Native Rust Modules - -PyO3 Asyncio can also be used to write native modules with async functions. - -Add the `[lib]` section to `Cargo.toml` to make your library a `cdylib` that Python can import. - -```toml -[lib] -name = "my_async_module" -crate-type = ["cdylib"] -``` - -Make your project depend on `pyo3` with the `extension-module` feature enabled and select your -`pyo3-asyncio` runtime: - -For `async-std`: - -```toml -[dependencies] -pyo3 = { version = "0.17", features = ["extension-module"] } -pyo3-asyncio = { version = "0.17", features = ["async-std-runtime"] } -async-std = "1.9" -``` - -For `tokio`: - -```toml -[dependencies] -pyo3 = { version = "0.17", features = ["extension-module"] } -pyo3-asyncio = { version = "0.17", features = ["tokio-runtime"] } -tokio = "1.9" -``` - -Export an async function that makes use of `async-std`: - -```rust -//! lib.rs - -use pyo3::{prelude::*, wrap_pyfunction}; - -#[pyfunction] -fn rust_sleep(py: Python) -> PyResult<&PyAny> { - pyo3_asyncio::async_std::future_into_py(py, async { - async_std::task::sleep(std::time::Duration::from_secs(1)).await; - Ok(()) - }) -} - -#[pymodule] -fn my_async_module(py: Python, m: &PyModule) -> PyResult<()> { - m.add_function(wrap_pyfunction!(rust_sleep, m)?)?; - - Ok(()) -} - -``` - -If you want to use `tokio` instead, here's what your module should look like: - -```rust -//! lib.rs - -use pyo3::{prelude::*, wrap_pyfunction}; - -#[pyfunction] -fn rust_sleep(py: Python) -> PyResult<&PyAny> { - pyo3_asyncio::tokio::future_into_py(py, async { - tokio::time::sleep(std::time::Duration::from_secs(1)).await; - Ok(()) - }) -} - -#[pymodule] -fn my_async_module(py: Python, m: &PyModule) -> PyResult<()> { - m.add_function(wrap_pyfunction!(rust_sleep, m)?)?; - Ok(()) -} -``` - -You can build your module with maturin (see the [Using Rust in Python](https://pyo3.rs/main/#using-rust-from-python) section in the PyO3 guide for setup instructions). After that you should be able to run the Python REPL to try it out. - -```bash -maturin develop && python3 -🔗 Found pyo3 bindings -🐍 Found CPython 3.8 at python3 - Finished dev [unoptimized + debuginfo] target(s) in 0.04s -Python 3.8.5 (default, Jan 27 2021, 15:41:15) -[GCC 9.3.0] on linux -Type "help", "copyright", "credits" or "license" for more information. ->>> import asyncio ->>> ->>> from my_async_module import rust_sleep ->>> ->>> async def main(): ->>> await rust_sleep() ->>> ->>> # should sleep for 1s ->>> asyncio.run(main()) ->>> -``` - -### Awaiting an Async Python Function in Rust - -Let's take a look at a dead simple async Python function: - -```python -# Sleep for 1 second -async def py_sleep(): - await asyncio.sleep(1) -``` - -**Async functions in Python are simply functions that return a `coroutine` object**. For our purposes, -we really don't need to know much about these `coroutine` objects. The key factor here is that calling -an `async` function is _just like calling a regular function_, the only difference is that we have -to do something special with the object that it returns. - -Normally in Python, that something special is the `await` keyword, but in order to await this -coroutine in Rust, we first need to convert it into Rust's version of a `coroutine`: a `Future`. -That's where `pyo3-asyncio` comes in. -[`pyo3_asyncio::into_future`](https://docs.rs/pyo3-asyncio/latest/pyo3_asyncio/fn.into_future.html) -performs this conversion for us: - -```rust no_run -use pyo3::prelude::*; - -#[pyo3_asyncio::tokio::main] -async fn main() -> PyResult<()> { - let future = Python::with_gil(|py| -> PyResult<_> { - // import the module containing the py_sleep function - let example = py.import("example")?; - - // calling the py_sleep method like a normal function - // returns a coroutine - let coroutine = example.call_method0("py_sleep")?; - - // convert the coroutine into a Rust future using the - // tokio runtime - pyo3_asyncio::tokio::into_future(coroutine) - })?; - - // await the future - future.await?; - - Ok(()) -} -``` - -> If you're interested in learning more about `coroutines` and `awaitables` in general, check out the -> [Python 3 `asyncio` docs](https://docs.python.org/3/library/asyncio-task.html) for more information. - -### Awaiting a Rust Future in Python - -Here we have the same async function as before written in Rust using the -[`async-std`](https://async.rs/) runtime: - -```rust -/// Sleep for 1 second -async fn rust_sleep() { - async_std::task::sleep(std::time::Duration::from_secs(1)).await; -} -``` - -Similar to Python, Rust's async functions also return a special object called a -`Future`: - -```rust compile_fail -let future = rust_sleep(); -``` - -We can convert this `Future` object into Python to make it `awaitable`. This tells Python that you -can use the `await` keyword with it. In order to do this, we'll call -[`pyo3_asyncio::async_std::future_into_py`](https://docs.rs/pyo3-asyncio/latest/pyo3_asyncio/async_std/fn.future_into_py.html): - -```rust -use pyo3::prelude::*; - -async fn rust_sleep() { - async_std::task::sleep(std::time::Duration::from_secs(1)).await; -} - -#[pyfunction] -fn call_rust_sleep(py: Python) -> PyResult<&PyAny> { - pyo3_asyncio::async_std::future_into_py(py, async move { - rust_sleep().await; - Ok(()) - }) -} -``` - -In Python, we can call this pyo3 function just like any other async function: - -```python -from example import call_rust_sleep - -async def rust_sleep(): - await call_rust_sleep() -``` - -## Managing Event Loops - -Python's event loop requires some special treatment, especially regarding the main thread. Some of -Python's `asyncio` features, like proper signal handling, require control over the main thread, which -doesn't always play well with Rust. - -Luckily, Rust's event loops are pretty flexible and don't _need_ control over the main thread, so in -`pyo3-asyncio`, we decided the best way to handle Rust/Python interop was to just surrender the main -thread to Python and run Rust's event loops in the background. Unfortunately, since most event loop -implementations _prefer_ control over the main thread, this can still make some things awkward. - -### PyO3 Asyncio Initialization - -Because Python needs to control the main thread, we can't use the convenient proc macros from Rust -runtimes to handle the `main` function or `#[test]` functions. Instead, the initialization for PyO3 has to be done from the `main` function and the main -thread must block on [`pyo3_asyncio::run_forever`](https://docs.rs/pyo3-asyncio/latest/pyo3_asyncio/fn.run_forever.html) or [`pyo3_asyncio::async_std::run_until_complete`](https://docs.rs/pyo3-asyncio/latest/pyo3_asyncio/async_std/fn.run_until_complete.html). - -Because we have to block on one of those functions, we can't use [`#[async_std::main]`](https://docs.rs/async-std/latest/async_std/attr.main.html) or [`#[tokio::main]`](https://docs.rs/tokio/1.1.0/tokio/attr.main.html) -since it's not a good idea to make long blocking calls during an async function. - -> Internally, these `#[main]` proc macros are expanded to something like this: -> -> ```rust compile_fail -> fn main() { -> // your async main fn -> async fn _main_impl() { /* ... */ } -> Runtime::new().block_on(_main_impl()); -> } -> ``` -> -> Making a long blocking call inside the `Future` that's being driven by `block_on` prevents that -> thread from doing anything else and can spell trouble for some runtimes (also this will actually -> deadlock a single-threaded runtime!). Many runtimes have some sort of `spawn_blocking` mechanism -> that can avoid this problem, but again that's not something we can use here since we need it to -> block on the _main_ thread. - -For this reason, `pyo3-asyncio` provides its own set of proc macros to provide you with this -initialization. These macros are intended to mirror the initialization of `async-std` and `tokio` -while also satisfying the Python runtime's needs. - -Here's a full example of PyO3 initialization with the `async-std` runtime: - -```rust no_run -use pyo3::prelude::*; - -#[pyo3_asyncio::async_std::main] -async fn main() -> PyResult<()> { - // PyO3 is initialized - Ready to go - - let fut = Python::with_gil(|py| -> PyResult<_> { - let asyncio = py.import("asyncio")?; - - // convert asyncio.sleep into a Rust Future - pyo3_asyncio::async_std::into_future( - asyncio.call_method1("sleep", (1.into_py(py),))? - ) - })?; - - fut.await?; - - Ok(()) -} -``` - -#### A Note About `asyncio.run` - -In Python 3.7+, the recommended way to run a top-level coroutine with `asyncio` -is with `asyncio.run`. In `v0.13` we recommended against using this function due to initialization issues, but in `v0.14` it's perfectly valid to use this function... with a caveat. - -Since our Rust <--> Python conversions require a reference to the Python event loop, this poses a problem. Imagine we have a PyO3 Asyncio module that defines -a `rust_sleep` function like in previous examples. You might rightfully assume that you can call pass this directly into `asyncio.run` like this: - -```python -import asyncio - -from my_async_module import rust_sleep - -asyncio.run(rust_sleep()) -``` - -You might be surprised to find out that this throws an error: - -```bash -Traceback (most recent call last): - File "example.py", line 5, in <module> - asyncio.run(rust_sleep()) -RuntimeError: no running event loop -``` - -What's happening here is that we are calling `rust_sleep` _before_ the future is -actually running on the event loop created by `asyncio.run`. This is counter-intuitive, but expected behaviour, and unfortunately there doesn't seem to be a good way of solving this problem within PyO3 Asyncio itself. - -However, we can make this example work with a simple workaround: - -```python -import asyncio - -from my_async_module import rust_sleep - -# Calling main will just construct the coroutine that later calls rust_sleep. -# - This ensures that rust_sleep will be called when the event loop is running, -# not before. -async def main(): - await rust_sleep() - -# Run the main() coroutine at the top-level instead -asyncio.run(main()) -``` - -#### Non-standard Python Event Loops - -Python allows you to use alternatives to the default `asyncio` event loop. One -popular alternative is `uvloop`. In `v0.13` using non-standard event loops was -a bit of an ordeal, but in `v0.14` it's trivial. - -#### Using `uvloop` in a PyO3 Asyncio Native Extensions - -```toml -# Cargo.toml - -[lib] -name = "my_async_module" -crate-type = ["cdylib"] - -[dependencies] -pyo3 = { version = "0.17", features = ["extension-module"] } -pyo3-asyncio = { version = "0.17", features = ["tokio-runtime"] } -async-std = "1.9" -tokio = "1.9" -``` - -```rust -//! lib.rs - -use pyo3::{prelude::*, wrap_pyfunction}; - -#[pyfunction] -fn rust_sleep(py: Python) -> PyResult<&PyAny> { - pyo3_asyncio::tokio::future_into_py(py, async { - tokio::time::sleep(std::time::Duration::from_secs(1)).await; - Ok(()) - }) -} - -#[pymodule] -fn my_async_module(_py: Python, m: &PyModule) -> PyResult<()> { - m.add_function(wrap_pyfunction!(rust_sleep, m)?)?; - - Ok(()) -} -``` - -```bash -$ maturin develop && python3 -🔗 Found pyo3 bindings -🐍 Found CPython 3.8 at python3 - Finished dev [unoptimized + debuginfo] target(s) in 0.04s -Python 3.8.8 (default, Apr 13 2021, 19:58:26) -[GCC 7.3.0] :: Anaconda, Inc. on linux -Type "help", "copyright", "credits" or "license" for more information. ->>> import asyncio ->>> import uvloop ->>> ->>> import my_async_module ->>> ->>> uvloop.install() ->>> ->>> async def main(): -... await my_async_module.rust_sleep() -... ->>> asyncio.run(main()) ->>> -``` - -#### Using `uvloop` in Rust Applications - -Using `uvloop` in Rust applications is a bit trickier, but it's still possible -with relatively few modifications. - -Unfortunately, we can't make use of the `#[pyo3_asyncio::<runtime>::main]` attribute with non-standard event loops. This is because the `#[pyo3_asyncio::<runtime>::main]` proc macro has to interact with the Python -event loop before we can install the `uvloop` policy. - -```toml -[dependencies] -async-std = "1.9" -pyo3 = "0.17" -pyo3-asyncio = { version = "0.17", features = ["async-std-runtime"] } -``` - -```rust no_run -//! main.rs - -use pyo3::{prelude::*, types::PyType}; - -fn main() -> PyResult<()> { - pyo3::prepare_freethreaded_python(); - - Python::with_gil(|py| { - let uvloop = py.import("uvloop")?; - uvloop.call_method0("install")?; - - // store a reference for the assertion - let uvloop = PyObject::from(uvloop); - - pyo3_asyncio::async_std::run(py, async move { - // verify that we are on a uvloop.Loop - Python::with_gil(|py| -> PyResult<()> { - assert!(uvloop - .as_ref(py) - .getattr("Loop")? - .downcast::<PyType>() - .unwrap() - .is_instance(pyo3_asyncio::async_std::get_current_loop(py)?)?); - Ok(()) - })?; - - async_std::task::sleep(std::time::Duration::from_secs(1)).await; - - Ok(()) - }) - }) -} -``` - -### Additional Information - -- Managing event loop references can be tricky with pyo3-asyncio. See [Event Loop References and ContextVars](https://awestlake87.github.io/pyo3-asyncio/master/doc/pyo3_asyncio/#event-loop-references-and-contextvars) in the API docs to get a better intuition for how event loop references are managed in this library. -- Testing pyo3-asyncio libraries and applications requires a custom test harness since Python requires control over the main thread. You can find a testing guide in the [API docs for the `testing` module](https://awestlake87.github.io/pyo3-asyncio/master/doc/pyo3_asyncio/testing) - -## Migration Guide - -### Migrating from 0.13 to 0.14 - -So what's changed from `v0.13` to `v0.14`? - -Well, a lot actually. There were some pretty major flaws in the initialization behaviour of `v0.13`. While it would have been nicer to address these issues without changing the public API, I decided it'd be better to break some of the old API rather than completely change the underlying behaviour of the existing functions. I realize this is going to be a bit of a headache, so hopefully this section will help you through it. - -To make things a bit easier, I decided to keep most of the old API alongside the new one (with some deprecation warnings to encourage users to move away from it). It should be possible to use the `v0.13` API alongside the newer `v0.14` API, which should allow you to upgrade your application piecemeal rather than all at once. - -**Before you get started, I personally recommend taking a look at [Event Loop References and ContextVars](https://awestlake87.github.io/pyo3-asyncio/master/doc/pyo3_asyncio/#event-loop-references-and-contextvars) in order to get a better grasp on the motivation behind these changes and the nuance involved in using the new conversions.** - -### 0.14 Highlights - -- Tokio initialization is now lazy. - - No configuration necessary if you're using the multithreaded scheduler - - Calls to `pyo3_asyncio::tokio::init_multithread` or `pyo3_asyncio::tokio::init_multithread_once` can just be removed. - - Calls to `pyo3_asyncio::tokio::init_current_thread` or `pyo3_asyncio::tokio::init_current_thread_once` require some special attention. - - Custom runtime configuration is done by passing a `tokio::runtime::Builder` into `pyo3_asyncio::tokio::init` instead of a `tokio::runtime::Runtime` -- A new, more correct set of functions has been added to replace the `v0.13` conversions. - - `pyo3_asyncio::into_future_with_loop` - - `pyo3_asyncio::<runtime>::future_into_py_with_loop` - - `pyo3_asyncio::<runtime>::local_future_into_py_with_loop` - - `pyo3_asyncio::<runtime>::into_future` - - `pyo3_asyncio::<runtime>::future_into_py` - - `pyo3_asyncio::<runtime>::local_future_into_py` - - `pyo3_asyncio::<runtime>::get_current_loop` -- `pyo3_asyncio::try_init` is no longer required if you're only using `0.14` conversions -- The `ThreadPoolExecutor` is no longer configured automatically at the start. - - Fortunately, this doesn't seem to have much effect on `v0.13` code, it just means that it's now possible to configure the executor manually as you see fit. - -### Upgrading Your Code to 0.14 - -1. Fix PyO3 0.14 initialization. - - PyO3 0.14 feature gated its automatic initialization behaviour behind "auto-initialize". You can either enable the "auto-initialize" behaviour in your project or add a call to `pyo3::prepare_freethreaded_python()` to the start of your program. - - If you're using the `#[pyo3_asyncio::<runtime>::main]` proc macro attributes, then you can skip this step. `#[pyo3_asyncio::<runtime>::main]` will call `pyo3::prepare_freethreaded_python()` at the start regardless of your project's "auto-initialize" feature. -2. Fix the tokio initialization. - - - Calls to `pyo3_asyncio::tokio::init_multithread` or `pyo3_asyncio::tokio::init_multithread_once` can just be removed. - - If you're using the current thread scheduler, you'll need to manually spawn the thread that it runs on during initialization: - - ```rust no_run - let mut builder = tokio::runtime::Builder::new_current_thread(); - builder.enable_all(); - - pyo3_asyncio::tokio::init(builder); - std::thread::spawn(move || { - pyo3_asyncio::tokio::get_runtime().block_on( - futures::future::pending::<()>() - ); - }); - ``` - - - Custom `tokio::runtime::Builder` configs can be passed into `pyo3_asyncio::tokio::init`. The `tokio::runtime::Runtime` will be lazily instantiated on the first call to `pyo3_asyncio::tokio::get_runtime()` - -3. If you're using `pyo3_asyncio::run_forever` in your application, you should switch to a more manual approach. - - > `run_forever` is not the recommended way of running an event loop in Python, so it might be a good idea to move away from it. This function would have needed to change for `0.14`, but since it's considered an edge case, it was decided that users could just manually call it if they need to. - - ```rust - use pyo3::prelude::*; - - fn main() -> PyResult<()> { - pyo3::prepare_freethreaded_python(); - - Python::with_gil(|py| { - let asyncio = py.import("asyncio")?; - - let event_loop = asyncio.call_method0("new_event_loop")?; - asyncio.call_method1("set_event_loop", (event_loop,))?; - - let event_loop_hdl = PyObject::from(event_loop); - - pyo3_asyncio::tokio::get_runtime().spawn(async move { - tokio::time::sleep(std::time::Duration::from_secs(1)).await; - - // Stop the event loop manually - Python::with_gil(|py| { - event_loop_hdl - .as_ref(py) - .call_method1( - "call_soon_threadsafe", - (event_loop_hdl - .as_ref(py) - .getattr("stop") - .unwrap(),), - ) - .unwrap(); - }) - }); - - event_loop.call_method0("run_forever")?; - Ok(()) - }) - } - ``` - -4. Replace conversions with their newer counterparts. - > You may encounter some issues regarding the usage of `get_running_loop` vs `get_event_loop`. For more details on these newer conversions and how they should be used see [Event Loop References and ContextVars](https://awestlake87.github.io/pyo3-asyncio/master/doc/pyo3_asyncio/#event-loop-references-and-contextvars). - - Replace `pyo3_asyncio::into_future` with `pyo3_asyncio::<runtime>::into_future` - - Replace `pyo3_asyncio::<runtime>::into_coroutine` with `pyo3_asyncio::<runtime>::future_into_py` - - Replace `pyo3_asyncio::get_event_loop` with `pyo3_asyncio::<runtime>::get_current_loop` -5. After all conversions have been replaced with their `v0.14` counterparts, `pyo3_asyncio::try_init` can safely be removed. - -> The `v0.13` API has been removed in version `v0.15` - -### Migrating from 0.14 to 0.15+ - -There have been a few changes to the API in order to support proper cancellation from Python and the `contextvars` module. - -- Any instance of `cancellable_future_into_py` and `local_cancellable_future_into_py` conversions can be replaced with their`future_into_py` and `local_future_into_py` counterparts. - > Cancellation support became the default behaviour in 0.15. -- Instances of `*_with_loop` conversions should be replaced with the newer `*_with_locals` conversions. - - ```rust no_run - use pyo3::prelude::*; - - Python::with_gil(|py| -> PyResult<()> { - - // *_with_loop conversions in 0.14 - // - // let event_loop = pyo3_asyncio::get_running_loop(py)?; - // - // let fut = pyo3_asyncio::tokio::future_into_py_with_loop( - // event_loop, - // async move { Ok(Python::with_gil(|py| py.None())) } - // )?; - // - // should be replaced with *_with_locals in 0.15+ - let fut = pyo3_asyncio::tokio::future_into_py_with_locals( - py, - pyo3_asyncio::tokio::get_current_locals(py)?, - async move { Ok(()) } - )?; - - Ok(()) - }); - ``` - -- `scope` and `scope_local` variants now accept `TaskLocals` instead of `event_loop`. You can usually just replace the `event_loop` with `pyo3_asyncio::TaskLocals::new(event_loop).copy_context(py)?`. -- Return types for `future_into_py`, `future_into_py_with_locals` `local_future_into_py`, and `local_future_into_py_with_locals` are now constrained by the bound `IntoPy<PyObject>` instead of requiring the return type to be `PyObject`. This can make the return types for futures more flexible, but inference can also fail when the concrete type is ambiguous (for example when using `into()`). Sometimes the `into()` can just be removed, -- `run`, and `run_until_complete` can now return any `Send + 'static` value. - -### Migrating from 0.15 to 0.16 - -Actually, not much has changed in the API. I'm happy to say that the PyO3 Asyncio is reaching a -pretty stable point in 0.16. For the most part, 0.16 has been about cleanup and removing deprecated -functions from the API. - -PyO3 0.16 comes with a few API changes of its own, but one of the changes that most impacted PyO3 -Asyncio was it's decision to drop support for Python 3.6. PyO3 Asyncio has been using a few -workarounds / hacks to support the pre-3.7 version of Python's asyncio library that are no longer -necessary. PyO3 Asyncio's underlying implementation is now a bit cleaner because of this. - -PyO3 Asyncio 0.15 included some important fixes to the API in order to add support for proper task -cancellation and allow for the preservation / use of contextvars in Python coroutines. This led to -the deprecation of some 0.14 functions that were used for edge cases in favor of some more correct -versions, and those deprecated functions are now removed from the API in 0.16. - -In addition, with PyO3 Asyncio 0.16, the library now has experimental support for conversions from -Python's async generators into a Rust `Stream`. There are currently two versions `v1` and `v2` with -slightly different performance and type signatures, so I'm hoping to get some feedback on which one -works best for downstream users. Just enable the `unstable-streams` feature and you're good to go! - -> The inverse conversion, Rust `Stream` to Python async generator, may come in a later release if -> requested! diff --git a/lib/pyo3-asyncio/codecov.yml b/lib/pyo3-asyncio/codecov.yml deleted file mode 100644 index 4e2196bd..00000000 --- a/lib/pyo3-asyncio/codecov.yml +++ /dev/null @@ -1,9 +0,0 @@ -comment: off - -coverage: - status: - project: - default: - target: auto - # Allow a tiny drop of overall project coverage in PR to reduce spurious failures. - threshold: 0.25% diff --git a/lib/pyo3-asyncio/examples/async_std.rs b/lib/pyo3-asyncio/examples/async_std.rs deleted file mode 100644 index ba7c8e71..00000000 --- a/lib/pyo3-asyncio/examples/async_std.rs +++ /dev/null @@ -1,17 +0,0 @@ -use pyo3::prelude::*; - -#[pyo3_asyncio::async_std::main] -async fn main() -> PyResult<()> { - let fut = Python::with_gil(|py| { - let asyncio = py.import("asyncio")?; - - // convert asyncio.sleep into a Rust Future - pyo3_asyncio::async_std::into_future(asyncio.call_method1("sleep", (1.into_py(py),))?) - })?; - - println!("sleeping for 1s"); - fut.await?; - println!("done"); - - Ok(()) -} diff --git a/lib/pyo3-asyncio/examples/tokio.rs b/lib/pyo3-asyncio/examples/tokio.rs deleted file mode 100644 index 794c43a2..00000000 --- a/lib/pyo3-asyncio/examples/tokio.rs +++ /dev/null @@ -1,17 +0,0 @@ -use pyo3::prelude::*; - -#[pyo3_asyncio::tokio::main] -async fn main() -> PyResult<()> { - let fut = Python::with_gil(|py| { - let asyncio = py.import("asyncio")?; - - // convert asyncio.sleep into a Rust Future - pyo3_asyncio::tokio::into_future(asyncio.call_method1("sleep", (1.into_py(py),))?) - })?; - - println!("sleeping for 1s"); - fut.await?; - println!("done"); - - Ok(()) -} diff --git a/lib/pyo3-asyncio/examples/tokio_current_thread.rs b/lib/pyo3-asyncio/examples/tokio_current_thread.rs deleted file mode 100644 index 5b8819a7..00000000 --- a/lib/pyo3-asyncio/examples/tokio_current_thread.rs +++ /dev/null @@ -1,17 +0,0 @@ -use pyo3::prelude::*; - -#[pyo3_asyncio::tokio::main(flavor = "current_thread")] -async fn main() -> PyResult<()> { - let fut = Python::with_gil(|py| { - let asyncio = py.import("asyncio")?; - - // convert asyncio.sleep into a Rust Future - pyo3_asyncio::tokio::into_future(asyncio.call_method1("sleep", (1.into_py(py),))?) - })?; - - println!("sleeping for 1s"); - fut.await?; - println!("done"); - - Ok(()) -} diff --git a/lib/pyo3-asyncio/examples/tokio_multi_thread.rs b/lib/pyo3-asyncio/examples/tokio_multi_thread.rs deleted file mode 100644 index 0deb6882..00000000 --- a/lib/pyo3-asyncio/examples/tokio_multi_thread.rs +++ /dev/null @@ -1,17 +0,0 @@ -use pyo3::prelude::*; - -#[pyo3_asyncio::tokio::main(flavor = "multi_thread", worker_threads = 10)] -async fn main() -> PyResult<()> { - let fut = Python::with_gil(|py| { - let asyncio = py.import("asyncio")?; - - // convert asyncio.sleep into a Rust Future - pyo3_asyncio::tokio::into_future(asyncio.call_method1("sleep", (1.into_py(py),))?) - })?; - - println!("sleeping for 1s"); - fut.await?; - println!("done"); - - Ok(()) -} diff --git a/lib/pyo3-asyncio/src/async_std.rs b/lib/pyo3-asyncio/src/async_std.rs deleted file mode 100644 index 0dbe04f9..00000000 --- a/lib/pyo3-asyncio/src/async_std.rs +++ /dev/null @@ -1,739 +0,0 @@ -//! <code>async-std-runtime</code> PyO3 Asyncio functions specific to the async-std runtime -//! -//! Items marked with -//! <code>unstable-streams</code> -//! are only available when the `unstable-streams` Cargo feature is enabled: -//! -//! ```toml -//! [dependencies.pyo3-asyncio] -//! version = "0.17" -//! features = ["unstable-streams"] -//! ``` - -use std::{any::Any, cell::RefCell, future::Future, panic::AssertUnwindSafe, pin::Pin}; - -use async_std::task; -use futures::FutureExt; -use pyo3::prelude::*; - -use crate::{ - generic::{self, ContextExt, JoinError, LocalContextExt, Runtime, SpawnLocalExt}, - TaskLocals, -}; - -/// <code>attributes</code> -/// re-exports for macros -#[cfg(feature = "attributes")] -pub mod re_exports { - /// re-export spawn_blocking for use in `#[test]` macro without external dependency - pub use async_std::task::spawn_blocking; -} - -/// <code>attributes</code> Provides the boilerplate for the `async-std` runtime and runs an async fn as main -#[cfg(feature = "attributes")] -pub use pyo3_asyncio_macros::async_std_main as main; - -/// <code>attributes</code> -/// <code>testing</code> -/// Registers an `async-std` test with the `pyo3-asyncio` test harness -#[cfg(all(feature = "attributes", feature = "testing"))] -pub use pyo3_asyncio_macros::async_std_test as test; - -struct AsyncStdJoinErr(Box<dyn Any + Send + 'static>); - -impl JoinError for AsyncStdJoinErr { - fn is_panic(&self) -> bool { - true - } -} - -async_std::task_local! { - static TASK_LOCALS: RefCell<Option<TaskLocals>> = RefCell::new(None); -} - -struct AsyncStdRuntime; - -impl Runtime for AsyncStdRuntime { - type JoinError = AsyncStdJoinErr; - type JoinHandle = task::JoinHandle<Result<(), AsyncStdJoinErr>>; - - fn spawn<F>(fut: F) -> Self::JoinHandle - where - F: Future<Output = ()> + Send + 'static, - { - task::spawn(async move { - AssertUnwindSafe(fut) - .catch_unwind() - .await - .map_err(|e| AsyncStdJoinErr(e)) - }) - } -} - -impl ContextExt for AsyncStdRuntime { - fn scope<F, R>(locals: TaskLocals, fut: F) -> Pin<Box<dyn Future<Output = R> + Send>> - where - F: Future<Output = R> + Send + 'static, - { - let old = TASK_LOCALS.with(|c| c.replace(Some(locals))); - Box::pin(async move { - let result = fut.await; - TASK_LOCALS.with(|c| c.replace(old)); - result - }) - } - - fn get_task_locals() -> Option<TaskLocals> { - match TASK_LOCALS.try_with(|c| c.borrow().clone()) { - Ok(locals) => locals, - Err(_) => None, - } - } -} - -impl SpawnLocalExt for AsyncStdRuntime { - fn spawn_local<F>(fut: F) -> Self::JoinHandle - where - F: Future<Output = ()> + 'static, - { - task::spawn_local(async move { - fut.await; - Ok(()) - }) - } -} - -impl LocalContextExt for AsyncStdRuntime { - fn scope_local<F, R>(locals: TaskLocals, fut: F) -> Pin<Box<dyn Future<Output = R>>> - where - F: Future<Output = R> + 'static, - { - let old = TASK_LOCALS.with(|c| c.replace(Some(locals))); - Box::pin(async move { - let result = fut.await; - TASK_LOCALS.with(|c| c.replace(old)); - result - }) - } -} - -/// Set the task local event loop for the given future -pub async fn scope<F, R>(locals: TaskLocals, fut: F) -> R -where - F: Future<Output = R> + Send + 'static, -{ - AsyncStdRuntime::scope(locals, fut).await -} - -/// Set the task local event loop for the given !Send future -pub async fn scope_local<F, R>(locals: TaskLocals, fut: F) -> R -where - F: Future<Output = R> + 'static, -{ - AsyncStdRuntime::scope_local(locals, fut).await -} - -/// Get the current event loop from either Python or Rust async task local context -/// -/// This function first checks if the runtime has a task-local reference to the Python event loop. -/// If not, it calls [`get_running_loop`](`crate::get_running_loop`) to get the event loop -/// associated with the current OS thread. -pub fn get_current_loop(py: Python) -> PyResult<&PyAny> { - generic::get_current_loop::<AsyncStdRuntime>(py) -} - -/// Either copy the task locals from the current task OR get the current running loop and -/// contextvars from Python. -pub fn get_current_locals(py: Python) -> PyResult<TaskLocals> { - generic::get_current_locals::<AsyncStdRuntime>(py) -} - -/// Run the event loop until the given Future completes -/// -/// The event loop runs until the given future is complete. -/// -/// After this function returns, the event loop can be resumed with [`run_until_complete`] -/// -/// # Arguments -/// * `event_loop` - The Python event loop that should run the future -/// * `fut` - The future to drive to completion -/// -/// # Examples -/// -/// ``` -/// # use std::time::Duration; -/// # -/// # use pyo3::prelude::*; -/// # -/// # pyo3::prepare_freethreaded_python(); -/// # -/// # Python::with_gil(|py| -> PyResult<()> { -/// # let event_loop = py.import("asyncio")?.call_method0("new_event_loop")?; -/// pyo3_asyncio::async_std::run_until_complete(event_loop, async move { -/// async_std::task::sleep(Duration::from_secs(1)).await; -/// Ok(()) -/// })?; -/// # Ok(()) -/// # }).unwrap(); -/// ``` -pub fn run_until_complete<F, T>(event_loop: &PyAny, fut: F) -> PyResult<T> -where - F: Future<Output = PyResult<T>> + Send + 'static, - T: Send + Sync + 'static, -{ - generic::run_until_complete::<AsyncStdRuntime, _, T>(event_loop, fut) -} - -/// Run the event loop until the given Future completes -/// -/// # Arguments -/// * `py` - The current PyO3 GIL guard -/// * `fut` - The future to drive to completion -/// -/// # Examples -/// -/// ```no_run -/// # use std::time::Duration; -/// # -/// # use pyo3::prelude::*; -/// # -/// fn main() { -/// // call this or use pyo3 0.14 "auto-initialize" feature -/// pyo3::prepare_freethreaded_python(); -/// -/// Python::with_gil(|py| { -/// pyo3_asyncio::async_std::run(py, async move { -/// async_std::task::sleep(Duration::from_secs(1)).await; -/// Ok(()) -/// }) -/// .map_err(|e| { -/// e.print_and_set_sys_last_vars(py); -/// }) -/// .unwrap(); -/// }) -/// } -/// ``` -pub fn run<F, T>(py: Python, fut: F) -> PyResult<T> -where - F: Future<Output = PyResult<T>> + Send + 'static, - T: Send + Sync + 'static, -{ - generic::run::<AsyncStdRuntime, F, T>(py, fut) -} - -/// Convert a Rust Future into a Python awaitable -/// -/// If the `asyncio.Future` returned by this conversion is cancelled via `asyncio.Future.cancel`, -/// the Rust future will be cancelled as well (new behaviour in `v0.15`). -/// -/// Python `contextvars` are preserved when calling async Python functions within the Rust future -/// via [`into_future`] (new behaviour in `v0.15`). -/// -/// > Although `contextvars` are preserved for async Python functions, synchronous functions will -/// unfortunately fail to resolve them when called within the Rust future. This is because the -/// function is being called from a Rust thread, not inside an actual Python coroutine context. -/// > -/// > As a workaround, you can get the `contextvars` from the current task locals using -/// [`get_current_locals`] and [`TaskLocals::context`](`crate::TaskLocals::context`), then wrap your -/// synchronous function in a call to `contextvars.Context.run`. This will set the context, call the -/// synchronous function, and restore the previous context when it returns or raises an exception. -/// -/// # Arguments -/// * `py` - PyO3 GIL guard -/// * `locals` - The task locals for the given future -/// * `fut` - The Rust future to be converted -/// -/// # Examples -/// -/// ``` -/// use std::time::Duration; -/// -/// use pyo3::prelude::*; -/// -/// /// Awaitable sleep function -/// #[pyfunction] -/// fn sleep_for<'p>(py: Python<'p>, secs: &'p PyAny) -> PyResult<&'p PyAny> { -/// let secs = secs.extract()?; -/// pyo3_asyncio::async_std::future_into_py_with_locals( -/// py, -/// pyo3_asyncio::async_std::get_current_locals(py)?, -/// async move { -/// async_std::task::sleep(Duration::from_secs(secs)).await; -/// Python::with_gil(|py| Ok(py.None())) -/// } -/// ) -/// } -/// ``` -pub fn future_into_py_with_locals<F, T>(py: Python, locals: TaskLocals, fut: F) -> PyResult<&PyAny> -where - F: Future<Output = PyResult<T>> + Send + 'static, - T: IntoPy<PyObject>, -{ - generic::future_into_py_with_locals::<AsyncStdRuntime, F, T>(py, locals, fut) -} - -/// Convert a Rust Future into a Python awaitable -/// -/// If the `asyncio.Future` returned by this conversion is cancelled via `asyncio.Future.cancel`, -/// the Rust future will be cancelled as well (new behaviour in `v0.15`). -/// -/// Python `contextvars` are preserved when calling async Python functions within the Rust future -/// via [`into_future`] (new behaviour in `v0.15`). -/// -/// > Although `contextvars` are preserved for async Python functions, synchronous functions will -/// unfortunately fail to resolve them when called within the Rust future. This is because the -/// function is being called from a Rust thread, not inside an actual Python coroutine context. -/// > -/// > As a workaround, you can get the `contextvars` from the current task locals using -/// [`get_current_locals`] and [`TaskLocals::context`](`crate::TaskLocals::context`), then wrap your -/// synchronous function in a call to `contextvars.Context.run`. This will set the context, call the -/// synchronous function, and restore the previous context when it returns or raises an exception. -/// -/// # Arguments -/// * `py` - The current PyO3 GIL guard -/// * `fut` - The Rust future to be converted -/// -/// # Examples -/// -/// ``` -/// use std::time::Duration; -/// -/// use pyo3::prelude::*; -/// -/// /// Awaitable sleep function -/// #[pyfunction] -/// fn sleep_for<'p>(py: Python<'p>, secs: &'p PyAny) -> PyResult<&'p PyAny> { -/// let secs = secs.extract()?; -/// pyo3_asyncio::async_std::future_into_py(py, async move { -/// async_std::task::sleep(Duration::from_secs(secs)).await; -/// Ok(()) -/// }) -/// } -/// ``` -pub fn future_into_py<F, T>(py: Python, fut: F) -> PyResult<&PyAny> -where - F: Future<Output = PyResult<T>> + Send + 'static, - T: IntoPy<PyObject>, -{ - generic::future_into_py::<AsyncStdRuntime, _, T>(py, fut) -} - -/// Convert a `!Send` Rust Future into a Python awaitable -/// -/// If the `asyncio.Future` returned by this conversion is cancelled via `asyncio.Future.cancel`, -/// the Rust future will be cancelled as well (new behaviour in `v0.15`). -/// -/// Python `contextvars` are preserved when calling async Python functions within the Rust future -/// via [`into_future`] (new behaviour in `v0.15`). -/// -/// > Although `contextvars` are preserved for async Python functions, synchronous functions will -/// unfortunately fail to resolve them when called within the Rust future. This is because the -/// function is being called from a Rust thread, not inside an actual Python coroutine context. -/// > -/// > As a workaround, you can get the `contextvars` from the current task locals using -/// [`get_current_locals`] and [`TaskLocals::context`](`crate::TaskLocals::context`), then wrap your -/// synchronous function in a call to `contextvars.Context.run`. This will set the context, call the -/// synchronous function, and restore the previous context when it returns or raises an exception. -/// -/// # Arguments -/// * `py` - PyO3 GIL guard -/// * `locals` - The task locals for the given future -/// * `fut` - The Rust future to be converted -/// -/// # Examples -/// -/// ``` -/// use std::{rc::Rc, time::Duration}; -/// -/// use pyo3::prelude::*; -/// -/// /// Awaitable non-send sleep function -/// #[pyfunction] -/// fn sleep_for(py: Python, secs: u64) -> PyResult<&PyAny> { -/// // Rc is non-send so it cannot be passed into pyo3_asyncio::async_std::future_into_py -/// let secs = Rc::new(secs); -/// Ok(pyo3_asyncio::async_std::local_future_into_py_with_locals( -/// py, -/// pyo3_asyncio::async_std::get_current_locals(py)?, -/// async move { -/// async_std::task::sleep(Duration::from_secs(*secs)).await; -/// Python::with_gil(|py| Ok(py.None())) -/// } -/// )?.into()) -/// } -/// -/// # #[cfg(all(feature = "async-std-runtime", feature = "attributes"))] -/// #[pyo3_asyncio::async_std::main] -/// async fn main() -> PyResult<()> { -/// Python::with_gil(|py| { -/// let py_future = sleep_for(py, 1)?; -/// pyo3_asyncio::async_std::into_future(py_future) -/// })? -/// .await?; -/// -/// Ok(()) -/// } -/// # #[cfg(not(all(feature = "async-std-runtime", feature = "attributes")))] -/// # fn main() {} -/// ``` -pub fn local_future_into_py_with_locals<F, T>( - py: Python, - locals: TaskLocals, - fut: F, -) -> PyResult<&PyAny> -where - F: Future<Output = PyResult<T>> + 'static, - T: IntoPy<PyObject>, -{ - generic::local_future_into_py_with_locals::<AsyncStdRuntime, _, T>(py, locals, fut) -} - -/// Convert a `!Send` Rust Future into a Python awaitable -/// -/// If the `asyncio.Future` returned by this conversion is cancelled via `asyncio.Future.cancel`, -/// the Rust future will be cancelled as well (new behaviour in `v0.15`). -/// -/// Python `contextvars` are preserved when calling async Python functions within the Rust future -/// via [`into_future`] (new behaviour in `v0.15`). -/// -/// > Although `contextvars` are preserved for async Python functions, synchronous functions will -/// unfortunately fail to resolve them when called within the Rust future. This is because the -/// function is being called from a Rust thread, not inside an actual Python coroutine context. -/// > -/// > As a workaround, you can get the `contextvars` from the current task locals using -/// [`get_current_locals`] and [`TaskLocals::context`](`crate::TaskLocals::context`), then wrap your -/// synchronous function in a call to `contextvars.Context.run`. This will set the context, call the -/// synchronous function, and restore the previous context when it returns or raises an exception. -/// -/// # Arguments -/// * `py` - The current PyO3 GIL guard -/// * `fut` - The Rust future to be converted -/// -/// # Examples -/// -/// ``` -/// use std::{rc::Rc, time::Duration}; -/// -/// use pyo3::prelude::*; -/// -/// /// Awaitable non-send sleep function -/// #[pyfunction] -/// fn sleep_for(py: Python, secs: u64) -> PyResult<&PyAny> { -/// // Rc is non-send so it cannot be passed into pyo3_asyncio::async_std::future_into_py -/// let secs = Rc::new(secs); -/// pyo3_asyncio::async_std::local_future_into_py(py, async move { -/// async_std::task::sleep(Duration::from_secs(*secs)).await; -/// Ok(()) -/// }) -/// } -/// -/// # #[cfg(all(feature = "async-std-runtime", feature = "attributes"))] -/// #[pyo3_asyncio::async_std::main] -/// async fn main() -> PyResult<()> { -/// Python::with_gil(|py| { -/// let py_future = sleep_for(py, 1)?; -/// pyo3_asyncio::async_std::into_future(py_future) -/// })? -/// .await?; -/// -/// Ok(()) -/// } -/// # #[cfg(not(all(feature = "async-std-runtime", feature = "attributes")))] -/// # fn main() {} -/// ``` -pub fn local_future_into_py<F, T>(py: Python, fut: F) -> PyResult<&PyAny> -where - F: Future<Output = PyResult<T>> + 'static, - T: IntoPy<PyObject>, -{ - generic::local_future_into_py::<AsyncStdRuntime, _, T>(py, fut) -} - -/// Convert a Python `awaitable` into a Rust Future -/// -/// This function converts the `awaitable` into a Python Task using `run_coroutine_threadsafe`. A -/// completion handler sends the result of this Task through a -/// `futures::channel::oneshot::Sender<PyResult<PyObject>>` and the future returned by this function -/// simply awaits the result through the `futures::channel::oneshot::Receiver<PyResult<PyObject>>`. -/// -/// # Arguments -/// * `awaitable` - The Python `awaitable` to be converted -/// -/// # Examples -/// -/// ``` -/// use std::time::Duration; -/// -/// use pyo3::prelude::*; -/// -/// const PYTHON_CODE: &'static str = r#" -/// import asyncio -/// -/// async def py_sleep(duration): -/// await asyncio.sleep(duration) -/// "#; -/// -/// async fn py_sleep(seconds: f32) -> PyResult<()> { -/// let test_mod = Python::with_gil(|py| -> PyResult<PyObject> { -/// Ok( -/// PyModule::from_code( -/// py, -/// PYTHON_CODE, -/// "test_into_future/test_mod.py", -/// "test_mod" -/// )? -/// .into() -/// ) -/// })?; -/// -/// Python::with_gil(|py| { -/// pyo3_asyncio::async_std::into_future( -/// test_mod -/// .call_method1(py, "py_sleep", (seconds.into_py(py),))? -/// .as_ref(py), -/// ) -/// })? -/// .await?; -/// Ok(()) -/// } -/// ``` -pub fn into_future(awaitable: &PyAny) -> PyResult<impl Future<Output = PyResult<PyObject>> + Send> { - generic::into_future::<AsyncStdRuntime>(awaitable) -} - -/// <code>unstable-streams</code> Convert an async generator into a stream -/// -/// **This API is marked as unstable** and is only available when the -/// `unstable-streams` crate feature is enabled. This comes with no -/// stability guarantees, and could be changed or removed at any time. -/// -/// # Arguments -/// * `gen` - The Python async generator to be converted -/// -/// # Examples -/// ``` -/// use pyo3::prelude::*; -/// use futures::{StreamExt, TryStreamExt}; -/// -/// const TEST_MOD: &str = r#" -/// import asyncio -/// -/// async def gen(): -/// for i in range(10): -/// await asyncio.sleep(0.1) -/// yield i -/// "#; -/// -/// # #[cfg(all(feature = "unstable-streams", feature = "attributes"))] -/// # #[pyo3_asyncio::async_std::main] -/// # async fn main() -> PyResult<()> { -/// let stream = Python::with_gil(|py| { -/// let test_mod = PyModule::from_code( -/// py, -/// TEST_MOD, -/// "test_rust_coroutine/test_mod.py", -/// "test_mod", -/// )?; -/// -/// pyo3_asyncio::async_std::into_stream_v1(test_mod.call_method0("gen")?) -/// })?; -/// -/// let vals = stream -/// .map(|item| Python::with_gil(|py| -> PyResult<i32> { Ok(item?.as_ref(py).extract()?) })) -/// .try_collect::<Vec<i32>>() -/// .await?; -/// -/// assert_eq!((0..10).collect::<Vec<i32>>(), vals); -/// -/// Ok(()) -/// # } -/// # #[cfg(not(all(feature = "unstable-streams", feature = "attributes")))] -/// # fn main() {} -/// ``` -#[cfg(feature = "unstable-streams")] -pub fn into_stream_v1<'p>( - gen: &'p PyAny, -) -> PyResult<impl futures::Stream<Item = PyResult<PyObject>> + 'static> { - generic::into_stream_v1::<AsyncStdRuntime>(gen) -} - -/// <code>unstable-streams</code> Convert an async generator into a stream -/// -/// **This API is marked as unstable** and is only available when the -/// `unstable-streams` crate feature is enabled. This comes with no -/// stability guarantees, and could be changed or removed at any time. -/// -/// # Arguments -/// * `locals` - The current task locals -/// * `gen` - The Python async generator to be converted -/// -/// # Examples -/// ``` -/// use pyo3::prelude::*; -/// use futures::{StreamExt, TryStreamExt}; -/// -/// const TEST_MOD: &str = r#" -/// import asyncio -/// -/// async def gen(): -/// for i in range(10): -/// await asyncio.sleep(0.1) -/// yield i -/// "#; -/// -/// # #[cfg(all(feature = "unstable-streams", feature = "attributes"))] -/// # #[pyo3_asyncio::async_std::main] -/// # async fn main() -> PyResult<()> { -/// let stream = Python::with_gil(|py| { -/// let test_mod = PyModule::from_code( -/// py, -/// TEST_MOD, -/// "test_rust_coroutine/test_mod.py", -/// "test_mod", -/// )?; -/// -/// pyo3_asyncio::async_std::into_stream_with_locals_v1( -/// pyo3_asyncio::async_std::get_current_locals(py)?, -/// test_mod.call_method0("gen")? -/// ) -/// })?; -/// -/// let vals = stream -/// .map(|item| Python::with_gil(|py| -> PyResult<i32> { Ok(item?.as_ref(py).extract()?) })) -/// .try_collect::<Vec<i32>>() -/// .await?; -/// -/// assert_eq!((0..10).collect::<Vec<i32>>(), vals); -/// -/// Ok(()) -/// # } -/// # #[cfg(not(all(feature = "unstable-streams", feature = "attributes")))] -/// # fn main() {} -/// ``` -#[cfg(feature = "unstable-streams")] -pub fn into_stream_with_locals_v1<'p>( - locals: TaskLocals, - gen: &'p PyAny, -) -> PyResult<impl futures::Stream<Item = PyResult<PyObject>> + 'static> { - generic::into_stream_with_locals_v1::<AsyncStdRuntime>(locals, gen) -} - -/// <code>unstable-streams</code> Convert an async generator into a stream -/// -/// **This API is marked as unstable** and is only available when the -/// `unstable-streams` crate feature is enabled. This comes with no -/// stability guarantees, and could be changed or removed at any time. -/// -/// # Arguments -/// * `locals` - The current task locals -/// * `gen` - The Python async generator to be converted -/// -/// # Examples -/// ``` -/// use pyo3::prelude::*; -/// use futures::{StreamExt, TryStreamExt}; -/// -/// const TEST_MOD: &str = r#" -/// import asyncio -/// -/// async def gen(): -/// for i in range(10): -/// await asyncio.sleep(0.1) -/// yield i -/// "#; -/// -/// # #[cfg(all(feature = "unstable-streams", feature = "attributes"))] -/// # #[pyo3_asyncio::async_std::main] -/// # async fn main() -> PyResult<()> { -/// let stream = Python::with_gil(|py| { -/// let test_mod = PyModule::from_code( -/// py, -/// TEST_MOD, -/// "test_rust_coroutine/test_mod.py", -/// "test_mod", -/// )?; -/// -/// pyo3_asyncio::async_std::into_stream_with_locals_v2( -/// pyo3_asyncio::async_std::get_current_locals(py)?, -/// test_mod.call_method0("gen")? -/// ) -/// })?; -/// -/// let vals = stream -/// .map(|item| Python::with_gil(|py| -> PyResult<i32> { Ok(item.as_ref(py).extract()?) })) -/// .try_collect::<Vec<i32>>() -/// .await?; -/// -/// assert_eq!((0..10).collect::<Vec<i32>>(), vals); -/// -/// Ok(()) -/// # } -/// # #[cfg(not(all(feature = "unstable-streams", feature = "attributes")))] -/// # fn main() {} -/// ``` -#[cfg(feature = "unstable-streams")] -pub fn into_stream_with_locals_v2<'p>( - locals: TaskLocals, - gen: &'p PyAny, -) -> PyResult<impl futures::Stream<Item = PyObject> + 'static> { - generic::into_stream_with_locals_v2::<AsyncStdRuntime>(locals, gen) -} - -/// <code>unstable-streams</code> Convert an async generator into a stream -/// -/// **This API is marked as unstable** and is only available when the -/// `unstable-streams` crate feature is enabled. This comes with no -/// stability guarantees, and could be changed or removed at any time. -/// -/// # Arguments -/// * `gen` - The Python async generator to be converted -/// -/// # Examples -/// ``` -/// use pyo3::prelude::*; -/// use futures::{StreamExt, TryStreamExt}; -/// -/// const TEST_MOD: &str = r#" -/// import asyncio -/// -/// async def gen(): -/// for i in range(10): -/// await asyncio.sleep(0.1) -/// yield i -/// "#; -/// -/// # #[cfg(all(feature = "unstable-streams", feature = "attributes"))] -/// # #[pyo3_asyncio::async_std::main] -/// # async fn main() -> PyResult<()> { -/// let stream = Python::with_gil(|py| { -/// let test_mod = PyModule::from_code( -/// py, -/// TEST_MOD, -/// "test_rust_coroutine/test_mod.py", -/// "test_mod", -/// )?; -/// -/// pyo3_asyncio::async_std::into_stream_v2(test_mod.call_method0("gen")?) -/// })?; -/// -/// let vals = stream -/// .map(|item| Python::with_gil(|py| -> PyResult<i32> { Ok(item.as_ref(py).extract()?) })) -/// .try_collect::<Vec<i32>>() -/// .await?; -/// -/// assert_eq!((0..10).collect::<Vec<i32>>(), vals); -/// -/// Ok(()) -/// # } -/// # #[cfg(not(all(feature = "unstable-streams", feature = "attributes")))] -/// # fn main() {} -/// ``` -#[cfg(feature = "unstable-streams")] -pub fn into_stream_v2<'p>( - gen: &'p PyAny, -) -> PyResult<impl futures::Stream<Item = PyObject> + 'static> { - generic::into_stream_v2::<AsyncStdRuntime>(gen) -} diff --git a/lib/pyo3-asyncio/src/err.rs b/lib/pyo3-asyncio/src/err.rs deleted file mode 100644 index 95e5d152..00000000 --- a/lib/pyo3-asyncio/src/err.rs +++ /dev/null @@ -1,9 +0,0 @@ -// FIXME - is there a way to document custom PyO3 exceptions? -#[allow(missing_docs)] -mod exceptions { - use pyo3::{create_exception, exceptions::PyException}; - - create_exception!(pyo3_asyncio, RustPanic, PyException); -} - -pub use exceptions::RustPanic; diff --git a/lib/pyo3-asyncio/src/generic.rs b/lib/pyo3-asyncio/src/generic.rs deleted file mode 100644 index 0f522b34..00000000 --- a/lib/pyo3-asyncio/src/generic.rs +++ /dev/null @@ -1,1729 +0,0 @@ -//! Generic implementations of PyO3 Asyncio utilities that can be used for any Rust runtime -//! -//! Items marked with -//! <code>unstable-streams</code> -//! are only available when the `unstable-streams` Cargo feature is enabled: -//! -//! ```toml -//! [dependencies.pyo3-asyncio] -//! version = "0.17" -//! features = ["unstable-streams"] -//! ``` - -use std::{ - future::Future, - marker::PhantomData, - pin::Pin, - sync::{Arc, Mutex}, - task::{Context, Poll}, -}; - -use futures::{ - channel::{mpsc, oneshot}, - SinkExt, -}; -use once_cell::sync::OnceCell; -use pin_project_lite::pin_project; -use pyo3::prelude::*; - -#[allow(deprecated)] -use crate::{ - asyncio, call_soon_threadsafe, close, create_future, dump_err, err::RustPanic, - get_running_loop, into_future_with_locals, TaskLocals, -}; - -/// Generic utilities for a JoinError -pub trait JoinError { - /// Check if the spawned task exited because of a panic - fn is_panic(&self) -> bool; -} - -/// Generic Rust async/await runtime -pub trait Runtime: Send + 'static { - /// The error returned by a JoinHandle after being awaited - type JoinError: JoinError + Send; - /// A future that completes with the result of the spawned task - type JoinHandle: Future<Output = Result<(), Self::JoinError>> + Send; - - /// Spawn a future onto this runtime's event loop - fn spawn<F>(fut: F) -> Self::JoinHandle - where - F: Future<Output = ()> + Send + 'static; -} - -/// Extension trait for async/await runtimes that support spawning local tasks -pub trait SpawnLocalExt: Runtime { - /// Spawn a !Send future onto this runtime's event loop - fn spawn_local<F>(fut: F) -> Self::JoinHandle - where - F: Future<Output = ()> + 'static; -} - -/// Exposes the utilities necessary for using task-local data in the Runtime -pub trait ContextExt: Runtime { - /// Set the task locals for the given future - fn scope<F, R>(locals: TaskLocals, fut: F) -> Pin<Box<dyn Future<Output = R> + Send>> - where - F: Future<Output = R> + Send + 'static; - - /// Get the task locals for the current task - fn get_task_locals() -> Option<TaskLocals>; -} - -/// Adds the ability to scope task-local data for !Send futures -pub trait LocalContextExt: Runtime { - /// Set the task locals for the given !Send future - fn scope_local<F, R>(locals: TaskLocals, fut: F) -> Pin<Box<dyn Future<Output = R>>> - where - F: Future<Output = R> + 'static; -} - -/// Get the current event loop from either Python or Rust async task local context -/// -/// This function first checks if the runtime has a task-local reference to the Python event loop. -/// If not, it calls [`get_running_loop`](crate::get_running_loop`) to get the event loop associated -/// with the current OS thread. -pub fn get_current_loop<R>(py: Python) -> PyResult<&PyAny> -where - R: ContextExt, -{ - if let Some(locals) = R::get_task_locals() { - Ok(locals.event_loop.into_ref(py)) - } else { - get_running_loop(py) - } -} - -/// Either copy the task locals from the current task OR get the current running loop and -/// contextvars from Python. -pub fn get_current_locals<R>(py: Python) -> PyResult<TaskLocals> -where - R: ContextExt, -{ - if let Some(locals) = R::get_task_locals() { - Ok(locals) - } else { - Ok(TaskLocals::with_running_loop(py)?.copy_context(py)?) - } -} - -/// Run the event loop until the given Future completes -/// -/// After this function returns, the event loop can be resumed with [`run_until_complete`] -/// -/// # Arguments -/// * `event_loop` - The Python event loop that should run the future -/// * `fut` - The future to drive to completion -/// -/// # Examples -/// -/// ```no_run -/// # use std::{task::{Context, Poll}, pin::Pin, future::Future}; -/// # -/// # use pyo3_asyncio::{ -/// # TaskLocals, -/// # generic::{JoinError, SpawnLocalExt, ContextExt, LocalContextExt, Runtime} -/// # }; -/// # -/// # struct MyCustomJoinError; -/// # -/// # impl JoinError for MyCustomJoinError { -/// # fn is_panic(&self) -> bool { -/// # unreachable!() -/// # } -/// # } -/// # -/// # struct MyCustomJoinHandle; -/// # -/// # impl Future for MyCustomJoinHandle { -/// # type Output = Result<(), MyCustomJoinError>; -/// # -/// # fn poll(self: Pin<&mut Self>, _cx: &mut Context) -> Poll<Self::Output> { -/// # unreachable!() -/// # } -/// # } -/// # -/// # struct MyCustomRuntime; -/// # -/// # impl Runtime for MyCustomRuntime { -/// # type JoinError = MyCustomJoinError; -/// # type JoinHandle = MyCustomJoinHandle; -/// # -/// # fn spawn<F>(fut: F) -> Self::JoinHandle -/// # where -/// # F: Future<Output = ()> + Send + 'static -/// # { -/// # unreachable!() -/// # } -/// # } -/// # -/// # impl ContextExt for MyCustomRuntime { -/// # fn scope<F, R>(locals: TaskLocals, fut: F) -> Pin<Box<dyn Future<Output = R> + Send>> -/// # where -/// # F: Future<Output = R> + Send + 'static -/// # { -/// # unreachable!() -/// # } -/// # fn get_task_locals() -> Option<TaskLocals> { -/// # unreachable!() -/// # } -/// # } -/// # -/// # use std::time::Duration; -/// # -/// # use pyo3::prelude::*; -/// # -/// # Python::with_gil(|py| -> PyResult<()> { -/// # let event_loop = py.import("asyncio")?.call_method0("new_event_loop")?; -/// # #[cfg(feature = "tokio-runtime")] -/// pyo3_asyncio::generic::run_until_complete::<MyCustomRuntime, _, _>(event_loop, async move { -/// tokio::time::sleep(Duration::from_secs(1)).await; -/// Ok(()) -/// })?; -/// # Ok(()) -/// # }).unwrap(); -/// ``` -pub fn run_until_complete<R, F, T>(event_loop: &PyAny, fut: F) -> PyResult<T> -where - R: Runtime + ContextExt, - F: Future<Output = PyResult<T>> + Send + 'static, - T: Send + Sync + 'static, -{ - let py = event_loop.py(); - let result_tx = Arc::new(Mutex::new(None)); - let result_rx = Arc::clone(&result_tx); - let coro = future_into_py_with_locals::<R, _, ()>( - py, - TaskLocals::new(event_loop).copy_context(py)?, - async move { - let val = fut.await?; - if let Ok(mut result) = result_tx.lock() { - *result = Some(val); - } - Ok(()) - }, - )?; - - event_loop.call_method1("run_until_complete", (coro,))?; - - let result = result_rx.lock().unwrap().take().unwrap(); - Ok(result) -} - -/// Run the event loop until the given Future completes -/// -/// # Arguments -/// * `py` - The current PyO3 GIL guard -/// * `fut` - The future to drive to completion -/// -/// # Examples -/// -/// ```no_run -/// # use std::{task::{Context, Poll}, pin::Pin, future::Future}; -/// # -/// # use pyo3_asyncio::{ -/// # TaskLocals, -/// # generic::{JoinError, SpawnLocalExt, ContextExt, LocalContextExt, Runtime} -/// # }; -/// # -/// # struct MyCustomJoinError; -/// # -/// # impl JoinError for MyCustomJoinError { -/// # fn is_panic(&self) -> bool { -/// # unreachable!() -/// # } -/// # } -/// # -/// # struct MyCustomJoinHandle; -/// # -/// # impl Future for MyCustomJoinHandle { -/// # type Output = Result<(), MyCustomJoinError>; -/// # -/// # fn poll(self: Pin<&mut Self>, _cx: &mut Context) -> Poll<Self::Output> { -/// # unreachable!() -/// # } -/// # } -/// # -/// # struct MyCustomRuntime; -/// # -/// # impl Runtime for MyCustomRuntime { -/// # type JoinError = MyCustomJoinError; -/// # type JoinHandle = MyCustomJoinHandle; -/// # -/// # fn spawn<F>(fut: F) -> Self::JoinHandle -/// # where -/// # F: Future<Output = ()> + Send + 'static -/// # { -/// # unreachable!() -/// # } -/// # } -/// # -/// # impl ContextExt for MyCustomRuntime { -/// # fn scope<F, R>(locals: TaskLocals, fut: F) -> Pin<Box<dyn Future<Output = R> + Send>> -/// # where -/// # F: Future<Output = R> + Send + 'static -/// # { -/// # unreachable!() -/// # } -/// # fn get_task_locals() -> Option<TaskLocals> { -/// # unreachable!() -/// # } -/// # } -/// # -/// # use std::time::Duration; -/// # async fn custom_sleep(_duration: Duration) { } -/// # -/// # use pyo3::prelude::*; -/// # -/// fn main() { -/// Python::with_gil(|py| { -/// pyo3_asyncio::generic::run::<MyCustomRuntime, _, _>(py, async move { -/// custom_sleep(Duration::from_secs(1)).await; -/// Ok(()) -/// }) -/// .map_err(|e| { -/// e.print_and_set_sys_last_vars(py); -/// }) -/// .unwrap(); -/// }) -/// } -/// ``` -pub fn run<R, F, T>(py: Python, fut: F) -> PyResult<T> -where - R: Runtime + ContextExt, - F: Future<Output = PyResult<T>> + Send + 'static, - T: Send + Sync + 'static, -{ - let event_loop = asyncio(py)?.call_method0("new_event_loop")?; - - let result = run_until_complete::<R, F, T>(event_loop, fut); - - close(event_loop)?; - - result -} - -pub fn cancelled(future: &PyAny) -> PyResult<bool> { - future.getattr("cancelled")?.call0()?.is_true() -} - -#[pyclass] -struct CheckedCompletor; - -#[pymethods] -impl CheckedCompletor { - fn __call__(&self, future: &PyAny, complete: &PyAny, value: &PyAny) -> PyResult<()> { - if cancelled(future)? { - return Ok(()); - } - - complete.call1((value,))?; - - Ok(()) - } -} - -pub fn set_result(event_loop: &PyAny, future: &PyAny, result: PyResult<PyObject>) -> PyResult<()> { - let py = event_loop.py(); - let none = py.None().into_ref(py); - - let (complete, val) = match result { - Ok(val) => (future.getattr("set_result")?, val.into_py(py)), - Err(err) => (future.getattr("set_exception")?, err.into_py(py)), - }; - call_soon_threadsafe(event_loop, none, (CheckedCompletor, future, complete, val))?; - - Ok(()) -} - -/// Convert a Python `awaitable` into a Rust Future -/// -/// This function simply forwards the future and the task locals returned by [`get_current_locals`] -/// to [`into_future_with_locals`](`crate::into_future_with_locals`). See -/// [`into_future_with_locals`](`crate::into_future_with_locals`) for more details. -/// -/// # Arguments -/// * `awaitable` - The Python `awaitable` to be converted -/// -/// # Examples -/// -/// ```no_run -/// # use std::{pin::Pin, future::Future, task::{Context, Poll}, time::Duration}; -/// # -/// # use pyo3::prelude::*; -/// # -/// # use pyo3_asyncio::{ -/// # TaskLocals, -/// # generic::{JoinError, SpawnLocalExt, ContextExt, LocalContextExt, Runtime} -/// # }; -/// # -/// # struct MyCustomJoinError; -/// # -/// # impl JoinError for MyCustomJoinError { -/// # fn is_panic(&self) -> bool { -/// # unreachable!() -/// # } -/// # } -/// # -/// # struct MyCustomJoinHandle; -/// # -/// # impl Future for MyCustomJoinHandle { -/// # type Output = Result<(), MyCustomJoinError>; -/// # -/// # fn poll(self: Pin<&mut Self>, _cx: &mut Context) -> Poll<Self::Output> { -/// # unreachable!() -/// # } -/// # } -/// # -/// # struct MyCustomRuntime; -/// # -/// # impl MyCustomRuntime { -/// # async fn sleep(_: Duration) { -/// # unreachable!() -/// # } -/// # } -/// # -/// # impl Runtime for MyCustomRuntime { -/// # type JoinError = MyCustomJoinError; -/// # type JoinHandle = MyCustomJoinHandle; -/// # -/// # fn spawn<F>(fut: F) -> Self::JoinHandle -/// # where -/// # F: Future<Output = ()> + Send + 'static -/// # { -/// # unreachable!() -/// # } -/// # } -/// # -/// # impl ContextExt for MyCustomRuntime { -/// # fn scope<F, R>(locals: TaskLocals, fut: F) -> Pin<Box<dyn Future<Output = R> + Send>> -/// # where -/// # F: Future<Output = R> + Send + 'static -/// # { -/// # unreachable!() -/// # } -/// # fn get_task_locals() -> Option<TaskLocals> { -/// # unreachable!() -/// # } -/// # } -/// # -/// const PYTHON_CODE: &'static str = r#" -/// import asyncio -/// -/// async def py_sleep(duration): -/// await asyncio.sleep(duration) -/// "#; -/// -/// async fn py_sleep(seconds: f32) -> PyResult<()> { -/// let test_mod = Python::with_gil(|py| -> PyResult<PyObject> { -/// Ok( -/// PyModule::from_code( -/// py, -/// PYTHON_CODE, -/// "test_into_future/test_mod.py", -/// "test_mod" -/// )? -/// .into() -/// ) -/// })?; -/// -/// Python::with_gil(|py| { -/// pyo3_asyncio::generic::into_future::<MyCustomRuntime>( -/// test_mod -/// .call_method1(py, "py_sleep", (seconds.into_py(py),))? -/// .as_ref(py), -/// ) -/// })? -/// .await?; -/// Ok(()) -/// } -/// ``` -pub fn into_future<R>( - awaitable: &PyAny, -) -> PyResult<impl Future<Output = PyResult<PyObject>> + Send> -where - R: Runtime + ContextExt, -{ - into_future_with_locals(&get_current_locals::<R>(awaitable.py())?, awaitable) -} - -/// Convert a Rust Future into a Python awaitable with a generic runtime -/// -/// If the `asyncio.Future` returned by this conversion is cancelled via `asyncio.Future.cancel`, -/// the Rust future will be cancelled as well (new behaviour in `v0.15`). -/// -/// Python `contextvars` are preserved when calling async Python functions within the Rust future -/// via [`into_future`] (new behaviour in `v0.15`). -/// -/// > Although `contextvars` are preserved for async Python functions, synchronous functions will -/// unfortunately fail to resolve them when called within the Rust future. This is because the -/// function is being called from a Rust thread, not inside an actual Python coroutine context. -/// > -/// > As a workaround, you can get the `contextvars` from the current task locals using -/// [`get_current_locals`] and [`TaskLocals::context`](`crate::TaskLocals::context`), then wrap your -/// synchronous function in a call to `contextvars.Context.run`. This will set the context, call the -/// synchronous function, and restore the previous context when it returns or raises an exception. -/// -/// # Arguments -/// * `py` - PyO3 GIL guard -/// * `locals` - The task-local data for Python -/// * `fut` - The Rust future to be converted -/// -/// # Examples -/// -/// ```no_run -/// # use std::{task::{Context, Poll}, pin::Pin, future::Future}; -/// # -/// # use pyo3_asyncio::{ -/// # TaskLocals, -/// # generic::{JoinError, SpawnLocalExt, ContextExt, LocalContextExt, Runtime} -/// # }; -/// # -/// # struct MyCustomJoinError; -/// # -/// # impl JoinError for MyCustomJoinError { -/// # fn is_panic(&self) -> bool { -/// # unreachable!() -/// # } -/// # } -/// # -/// # struct MyCustomJoinHandle; -/// # -/// # impl Future for MyCustomJoinHandle { -/// # type Output = Result<(), MyCustomJoinError>; -/// # -/// # fn poll(self: Pin<&mut Self>, _cx: &mut Context) -> Poll<Self::Output> { -/// # unreachable!() -/// # } -/// # } -/// # -/// # struct MyCustomRuntime; -/// # -/// # impl MyCustomRuntime { -/// # async fn sleep(_: Duration) { -/// # unreachable!() -/// # } -/// # } -/// # -/// # impl Runtime for MyCustomRuntime { -/// # type JoinError = MyCustomJoinError; -/// # type JoinHandle = MyCustomJoinHandle; -/// # -/// # fn spawn<F>(fut: F) -> Self::JoinHandle -/// # where -/// # F: Future<Output = ()> + Send + 'static -/// # { -/// # unreachable!() -/// # } -/// # } -/// # -/// # impl ContextExt for MyCustomRuntime { -/// # fn scope<F, R>(locals: TaskLocals, fut: F) -> Pin<Box<dyn Future<Output = R> + Send>> -/// # where -/// # F: Future<Output = R> + Send + 'static -/// # { -/// # unreachable!() -/// # } -/// # fn get_task_locals() -> Option<TaskLocals> { -/// # unreachable!() -/// # } -/// # } -/// # -/// use std::time::Duration; -/// -/// use pyo3::prelude::*; -/// -/// /// Awaitable sleep function -/// #[pyfunction] -/// fn sleep_for<'p>(py: Python<'p>, secs: &'p PyAny) -> PyResult<&'p PyAny> { -/// let secs = secs.extract()?; -/// pyo3_asyncio::generic::future_into_py_with_locals::<MyCustomRuntime, _, _>( -/// py, -/// pyo3_asyncio::generic::get_current_locals::<MyCustomRuntime>(py)?, -/// async move { -/// MyCustomRuntime::sleep(Duration::from_secs(secs)).await; -/// Ok(()) -/// } -/// ) -/// } -/// ``` -pub fn future_into_py_with_locals<R, F, T>( - py: Python, - locals: TaskLocals, - fut: F, -) -> PyResult<&PyAny> -where - R: Runtime + ContextExt, - F: Future<Output = PyResult<T>> + Send + 'static, - T: IntoPy<PyObject>, -{ - let (cancel_tx, cancel_rx) = oneshot::channel(); - - let py_fut = create_future(locals.event_loop.clone().into_ref(py))?; - py_fut.call_method1( - "add_done_callback", - (PyDoneCallback { - cancel_tx: Some(cancel_tx), - },), - )?; - - let future_tx1 = PyObject::from(py_fut); - let future_tx2 = future_tx1.clone(); - - R::spawn(async move { - let locals2 = locals.clone(); - - if let Err(e) = R::spawn(async move { - let result = R::scope( - locals2.clone(), - Cancellable::new_with_cancel_rx(fut, cancel_rx), - ) - .await; - - Python::with_gil(move |py| { - if cancelled(future_tx1.as_ref(py)) - .map_err(dump_err(py)) - .unwrap_or(false) - { - return; - } - - let _ = set_result( - locals2.event_loop(py), - future_tx1.as_ref(py), - result.map(|val| val.into_py(py)), - ) - .map_err(dump_err(py)); - }); - }) - .await - { - if e.is_panic() { - Python::with_gil(move |py| { - if cancelled(future_tx2.as_ref(py)) - .map_err(dump_err(py)) - .unwrap_or(false) - { - return; - } - - let _ = set_result( - locals.event_loop.as_ref(py), - future_tx2.as_ref(py), - Err(RustPanic::new_err("rust future panicked")), - ) - .map_err(dump_err(py)); - }); - } - } - }); - - Ok(py_fut) -} - -pin_project! { - /// Future returned by [`timeout`](timeout) and [`timeout_at`](timeout_at). - #[must_use = "futures do nothing unless you `.await` or poll them"] - #[derive(Debug)] - pub struct Cancellable<T> { - #[pin] - future: T, - #[pin] - cancel_rx: oneshot::Receiver<()>, - - poll_cancel_rx: bool - } -} - -impl<T> Cancellable<T> { - pub fn new_with_cancel_rx(future: T, cancel_rx: oneshot::Receiver<()>) -> Self { - Self { - future, - cancel_rx, - - poll_cancel_rx: true, - } - } -} - -impl<F, T> Future for Cancellable<F> -where - F: Future<Output = PyResult<T>>, - T: IntoPy<PyObject>, -{ - type Output = F::Output; - - fn poll(self: Pin<&mut Self>, cx: &mut Context<'_>) -> Poll<Self::Output> { - let this = self.project(); - - // First, try polling the future - if let Poll::Ready(v) = this.future.poll(cx) { - return Poll::Ready(v); - } - - // Now check for cancellation - if *this.poll_cancel_rx { - match this.cancel_rx.poll(cx) { - Poll::Ready(Ok(())) => { - *this.poll_cancel_rx = false; - // The python future has already been cancelled, so this return value will never - // be used. - Poll::Ready(Err(pyo3::exceptions::PyBaseException::new_err( - "unreachable", - ))) - } - Poll::Ready(Err(_)) => { - *this.poll_cancel_rx = false; - Poll::Pending - } - Poll::Pending => Poll::Pending, - } - } else { - Poll::Pending - } - } -} - -#[pyclass] -pub struct PyDoneCallback { - pub cancel_tx: Option<oneshot::Sender<()>>, -} - -#[pymethods] -impl PyDoneCallback { - pub fn __call__(&mut self, fut: &PyAny) -> PyResult<()> { - let py = fut.py(); - - if cancelled(fut).map_err(dump_err(py)).unwrap_or(false) { - let _ = self.cancel_tx.take().unwrap().send(()); - } - - Ok(()) - } -} - -/// Convert a Rust Future into a Python awaitable with a generic runtime -/// -/// If the `asyncio.Future` returned by this conversion is cancelled via `asyncio.Future.cancel`, -/// the Rust future will be cancelled as well (new behaviour in `v0.15`). -/// -/// Python `contextvars` are preserved when calling async Python functions within the Rust future -/// via [`into_future`] (new behaviour in `v0.15`). -/// -/// > Although `contextvars` are preserved for async Python functions, synchronous functions will -/// unfortunately fail to resolve them when called within the Rust future. This is because the -/// function is being called from a Rust thread, not inside an actual Python coroutine context. -/// > -/// > As a workaround, you can get the `contextvars` from the current task locals using -/// [`get_current_locals`] and [`TaskLocals::context`](`crate::TaskLocals::context`), then wrap your -/// synchronous function in a call to `contextvars.Context.run`. This will set the context, call the -/// synchronous function, and restore the previous context when it returns or raises an exception. -/// -/// # Arguments -/// * `py` - The current PyO3 GIL guard -/// * `fut` - The Rust future to be converted -/// -/// # Examples -/// -/// ```no_run -/// # use std::{task::{Context, Poll}, pin::Pin, future::Future}; -/// # -/// # use pyo3_asyncio::{ -/// # TaskLocals, -/// # generic::{JoinError, SpawnLocalExt, ContextExt, LocalContextExt, Runtime} -/// # }; -/// # -/// # struct MyCustomJoinError; -/// # -/// # impl JoinError for MyCustomJoinError { -/// # fn is_panic(&self) -> bool { -/// # unreachable!() -/// # } -/// # } -/// # -/// # struct MyCustomJoinHandle; -/// # -/// # impl Future for MyCustomJoinHandle { -/// # type Output = Result<(), MyCustomJoinError>; -/// # -/// # fn poll(self: Pin<&mut Self>, _cx: &mut Context) -> Poll<Self::Output> { -/// # unreachable!() -/// # } -/// # } -/// # -/// # struct MyCustomRuntime; -/// # -/// # impl MyCustomRuntime { -/// # async fn sleep(_: Duration) { -/// # unreachable!() -/// # } -/// # } -/// # -/// # impl Runtime for MyCustomRuntime { -/// # type JoinError = MyCustomJoinError; -/// # type JoinHandle = MyCustomJoinHandle; -/// # -/// # fn spawn<F>(fut: F) -> Self::JoinHandle -/// # where -/// # F: Future<Output = ()> + Send + 'static -/// # { -/// # unreachable!() -/// # } -/// # } -/// # -/// # impl ContextExt for MyCustomRuntime { -/// # fn scope<F, R>(locals: TaskLocals, fut: F) -> Pin<Box<dyn Future<Output = R> + Send>> -/// # where -/// # F: Future<Output = R> + Send + 'static -/// # { -/// # unreachable!() -/// # } -/// # fn get_task_locals() -> Option<TaskLocals> { -/// # unreachable!() -/// # } -/// # } -/// # -/// use std::time::Duration; -/// -/// use pyo3::prelude::*; -/// -/// /// Awaitable sleep function -/// #[pyfunction] -/// fn sleep_for<'p>(py: Python<'p>, secs: &'p PyAny) -> PyResult<&'p PyAny> { -/// let secs = secs.extract()?; -/// pyo3_asyncio::generic::future_into_py::<MyCustomRuntime, _, _>(py, async move { -/// MyCustomRuntime::sleep(Duration::from_secs(secs)).await; -/// Ok(()) -/// }) -/// } -/// ``` -pub fn future_into_py<R, F, T>(py: Python, fut: F) -> PyResult<&PyAny> -where - R: Runtime + ContextExt, - F: Future<Output = PyResult<T>> + Send + 'static, - T: IntoPy<PyObject>, -{ - future_into_py_with_locals::<R, F, T>(py, get_current_locals::<R>(py)?, fut) -} - -/// Convert a `!Send` Rust Future into a Python awaitable with a generic runtime and manual -/// specification of task locals. -/// -/// If the `asyncio.Future` returned by this conversion is cancelled via `asyncio.Future.cancel`, -/// the Rust future will be cancelled as well (new behaviour in `v0.15`). -/// -/// Python `contextvars` are preserved when calling async Python functions within the Rust future -/// via [`into_future`] (new behaviour in `v0.15`). -/// -/// > Although `contextvars` are preserved for async Python functions, synchronous functions will -/// unfortunately fail to resolve them when called within the Rust future. This is because the -/// function is being called from a Rust thread, not inside an actual Python coroutine context. -/// > -/// > As a workaround, you can get the `contextvars` from the current task locals using -/// [`get_current_locals`] and [`TaskLocals::context`](`crate::TaskLocals::context`), then wrap your -/// synchronous function in a call to `contextvars.Context.run`. This will set the context, call the -/// synchronous function, and restore the previous context when it returns or raises an exception. -/// -/// # Arguments -/// * `py` - PyO3 GIL guard -/// * `locals` - The task locals for the future -/// * `fut` - The Rust future to be converted -/// -/// # Examples -/// -/// ```no_run -/// # use std::{task::{Context, Poll}, pin::Pin, future::Future}; -/// # -/// # use pyo3_asyncio::{ -/// # TaskLocals, -/// # generic::{JoinError, SpawnLocalExt, ContextExt, LocalContextExt, Runtime} -/// # }; -/// # -/// # struct MyCustomJoinError; -/// # -/// # impl JoinError for MyCustomJoinError { -/// # fn is_panic(&self) -> bool { -/// # unreachable!() -/// # } -/// # } -/// # -/// # struct MyCustomJoinHandle; -/// # -/// # impl Future for MyCustomJoinHandle { -/// # type Output = Result<(), MyCustomJoinError>; -/// # -/// # fn poll(self: Pin<&mut Self>, _cx: &mut Context) -> Poll<Self::Output> { -/// # unreachable!() -/// # } -/// # } -/// # -/// # struct MyCustomRuntime; -/// # -/// # impl MyCustomRuntime { -/// # async fn sleep(_: Duration) { -/// # unreachable!() -/// # } -/// # } -/// # -/// # impl Runtime for MyCustomRuntime { -/// # type JoinError = MyCustomJoinError; -/// # type JoinHandle = MyCustomJoinHandle; -/// # -/// # fn spawn<F>(fut: F) -> Self::JoinHandle -/// # where -/// # F: Future<Output = ()> + Send + 'static -/// # { -/// # unreachable!() -/// # } -/// # } -/// # -/// # impl ContextExt for MyCustomRuntime { -/// # fn scope<F, R>(locals: TaskLocals, fut: F) -> Pin<Box<dyn Future<Output = R> + Send>> -/// # where -/// # F: Future<Output = R> + Send + 'static -/// # { -/// # unreachable!() -/// # } -/// # fn get_task_locals() -> Option<TaskLocals> { -/// # unreachable!() -/// # } -/// # } -/// # -/// # impl SpawnLocalExt for MyCustomRuntime { -/// # fn spawn_local<F>(fut: F) -> Self::JoinHandle -/// # where -/// # F: Future<Output = ()> + 'static -/// # { -/// # unreachable!() -/// # } -/// # } -/// # -/// # impl LocalContextExt for MyCustomRuntime { -/// # fn scope_local<F, R>(locals: TaskLocals, fut: F) -> Pin<Box<dyn Future<Output = R>>> -/// # where -/// # F: Future<Output = R> + 'static -/// # { -/// # unreachable!() -/// # } -/// # } -/// # -/// use std::{rc::Rc, time::Duration}; -/// -/// use pyo3::prelude::*; -/// -/// /// Awaitable sleep function -/// #[pyfunction] -/// fn sleep_for(py: Python, secs: u64) -> PyResult<&PyAny> { -/// // Rc is !Send so it cannot be passed into pyo3_asyncio::generic::future_into_py -/// let secs = Rc::new(secs); -/// -/// pyo3_asyncio::generic::local_future_into_py_with_locals::<MyCustomRuntime, _, _>( -/// py, -/// pyo3_asyncio::generic::get_current_locals::<MyCustomRuntime>(py)?, -/// async move { -/// MyCustomRuntime::sleep(Duration::from_secs(*secs)).await; -/// Ok(()) -/// } -/// ) -/// } -/// ``` -pub fn local_future_into_py_with_locals<R, F, T>( - py: Python, - locals: TaskLocals, - fut: F, -) -> PyResult<&PyAny> -where - R: Runtime + SpawnLocalExt + LocalContextExt, - F: Future<Output = PyResult<T>> + 'static, - T: IntoPy<PyObject>, -{ - let (cancel_tx, cancel_rx) = oneshot::channel(); - - let py_fut = create_future(locals.event_loop.clone().into_ref(py))?; - py_fut.call_method1( - "add_done_callback", - (PyDoneCallback { - cancel_tx: Some(cancel_tx), - },), - )?; - - let future_tx1 = PyObject::from(py_fut); - let future_tx2 = future_tx1.clone(); - - R::spawn_local(async move { - let locals2 = locals.clone(); - - if let Err(e) = R::spawn_local(async move { - let result = R::scope_local( - locals2.clone(), - Cancellable::new_with_cancel_rx(fut, cancel_rx), - ) - .await; - - Python::with_gil(move |py| { - if cancelled(future_tx1.as_ref(py)) - .map_err(dump_err(py)) - .unwrap_or(false) - { - return; - } - - let _ = set_result( - locals2.event_loop.as_ref(py), - future_tx1.as_ref(py), - result.map(|val| val.into_py(py)), - ) - .map_err(dump_err(py)); - }); - }) - .await - { - if e.is_panic() { - Python::with_gil(move |py| { - if cancelled(future_tx2.as_ref(py)) - .map_err(dump_err(py)) - .unwrap_or(false) - { - return; - } - - let _ = set_result( - locals.event_loop.as_ref(py), - future_tx2.as_ref(py), - Err(RustPanic::new_err("Rust future panicked")), - ) - .map_err(dump_err(py)); - }); - } - } - }); - - Ok(py_fut) -} - -/// Convert a `!Send` Rust Future into a Python awaitable with a generic runtime -/// -/// If the `asyncio.Future` returned by this conversion is cancelled via `asyncio.Future.cancel`, -/// the Rust future will be cancelled as well (new behaviour in `v0.15`). -/// -/// Python `contextvars` are preserved when calling async Python functions within the Rust future -/// via [`into_future`] (new behaviour in `v0.15`). -/// -/// > Although `contextvars` are preserved for async Python functions, synchronous functions will -/// unfortunately fail to resolve them when called within the Rust future. This is because the -/// function is being called from a Rust thread, not inside an actual Python coroutine context. -/// > -/// > As a workaround, you can get the `contextvars` from the current task locals using -/// [`get_current_locals`] and [`TaskLocals::context`](`crate::TaskLocals::context`), then wrap your -/// synchronous function in a call to `contextvars.Context.run`. This will set the context, call the -/// synchronous function, and restore the previous context when it returns or raises an exception. -/// -/// # Arguments -/// * `py` - The current PyO3 GIL guard -/// * `fut` - The Rust future to be converted -/// -/// # Examples -/// -/// ```no_run -/// # use std::{task::{Context, Poll}, pin::Pin, future::Future}; -/// # -/// # use pyo3_asyncio::{ -/// # TaskLocals, -/// # generic::{JoinError, SpawnLocalExt, ContextExt, LocalContextExt, Runtime} -/// # }; -/// # -/// # struct MyCustomJoinError; -/// # -/// # impl JoinError for MyCustomJoinError { -/// # fn is_panic(&self) -> bool { -/// # unreachable!() -/// # } -/// # } -/// # -/// # struct MyCustomJoinHandle; -/// # -/// # impl Future for MyCustomJoinHandle { -/// # type Output = Result<(), MyCustomJoinError>; -/// # -/// # fn poll(self: Pin<&mut Self>, _cx: &mut Context) -> Poll<Self::Output> { -/// # unreachable!() -/// # } -/// # } -/// # -/// # struct MyCustomRuntime; -/// # -/// # impl MyCustomRuntime { -/// # async fn sleep(_: Duration) { -/// # unreachable!() -/// # } -/// # } -/// # -/// # impl Runtime for MyCustomRuntime { -/// # type JoinError = MyCustomJoinError; -/// # type JoinHandle = MyCustomJoinHandle; -/// # -/// # fn spawn<F>(fut: F) -> Self::JoinHandle -/// # where -/// # F: Future<Output = ()> + Send + 'static -/// # { -/// # unreachable!() -/// # } -/// # } -/// # -/// # impl ContextExt for MyCustomRuntime { -/// # fn scope<F, R>(locals: TaskLocals, fut: F) -> Pin<Box<dyn Future<Output = R> + Send>> -/// # where -/// # F: Future<Output = R> + Send + 'static -/// # { -/// # unreachable!() -/// # } -/// # fn get_task_locals() -> Option<TaskLocals> { -/// # unreachable!() -/// # } -/// # } -/// # -/// # impl SpawnLocalExt for MyCustomRuntime { -/// # fn spawn_local<F>(fut: F) -> Self::JoinHandle -/// # where -/// # F: Future<Output = ()> + 'static -/// # { -/// # unreachable!() -/// # } -/// # } -/// # -/// # impl LocalContextExt for MyCustomRuntime { -/// # fn scope_local<F, R>(locals: TaskLocals, fut: F) -> Pin<Box<dyn Future<Output = R>>> -/// # where -/// # F: Future<Output = R> + 'static -/// # { -/// # unreachable!() -/// # } -/// # } -/// # -/// use std::{rc::Rc, time::Duration}; -/// -/// use pyo3::prelude::*; -/// -/// /// Awaitable sleep function -/// #[pyfunction] -/// fn sleep_for(py: Python, secs: u64) -> PyResult<&PyAny> { -/// // Rc is !Send so it cannot be passed into pyo3_asyncio::generic::future_into_py -/// let secs = Rc::new(secs); -/// -/// pyo3_asyncio::generic::local_future_into_py::<MyCustomRuntime, _, _>(py, async move { -/// MyCustomRuntime::sleep(Duration::from_secs(*secs)).await; -/// Ok(()) -/// }) -/// } -/// ``` -pub fn local_future_into_py<R, F, T>(py: Python, fut: F) -> PyResult<&PyAny> -where - R: Runtime + ContextExt + SpawnLocalExt + LocalContextExt, - F: Future<Output = PyResult<T>> + 'static, - T: IntoPy<PyObject>, -{ - local_future_into_py_with_locals::<R, F, T>(py, get_current_locals::<R>(py)?, fut) -} - -/// <code>unstable-streams</code> Convert an async generator into a stream -/// -/// **This API is marked as unstable** and is only available when the -/// `unstable-streams` crate feature is enabled. This comes with no -/// stability guarantees, and could be changed or removed at any time. -/// -/// # Arguments -/// * `locals` - The current task locals -/// * `gen` - The Python async generator to be converted -/// -/// # Examples -/// ```no_run -/// # use std::{task::{Context, Poll}, pin::Pin, future::Future}; -/// # -/// # use pyo3_asyncio::{ -/// # TaskLocals, -/// # generic::{JoinError, ContextExt, Runtime} -/// # }; -/// # -/// # struct MyCustomJoinError; -/// # -/// # impl JoinError for MyCustomJoinError { -/// # fn is_panic(&self) -> bool { -/// # unreachable!() -/// # } -/// # } -/// # -/// # struct MyCustomJoinHandle; -/// # -/// # impl Future for MyCustomJoinHandle { -/// # type Output = Result<(), MyCustomJoinError>; -/// # -/// # fn poll(self: Pin<&mut Self>, _cx: &mut Context) -> Poll<Self::Output> { -/// # unreachable!() -/// # } -/// # } -/// # -/// # struct MyCustomRuntime; -/// # -/// # impl Runtime for MyCustomRuntime { -/// # type JoinError = MyCustomJoinError; -/// # type JoinHandle = MyCustomJoinHandle; -/// # -/// # fn spawn<F>(fut: F) -> Self::JoinHandle -/// # where -/// # F: Future<Output = ()> + Send + 'static -/// # { -/// # unreachable!() -/// # } -/// # } -/// # -/// # impl ContextExt for MyCustomRuntime { -/// # fn scope<F, R>(locals: TaskLocals, fut: F) -> Pin<Box<dyn Future<Output = R> + Send>> -/// # where -/// # F: Future<Output = R> + Send + 'static -/// # { -/// # unreachable!() -/// # } -/// # fn get_task_locals() -> Option<TaskLocals> { -/// # unreachable!() -/// # } -/// # } -/// -/// use pyo3::prelude::*; -/// use futures::{StreamExt, TryStreamExt}; -/// -/// const TEST_MOD: &str = r#" -/// import asyncio -/// -/// async def gen(): -/// for i in range(10): -/// await asyncio.sleep(0.1) -/// yield i -/// "#; -/// -/// # async fn test_async_gen() -> PyResult<()> { -/// let stream = Python::with_gil(|py| { -/// let test_mod = PyModule::from_code( -/// py, -/// TEST_MOD, -/// "test_rust_coroutine/test_mod.py", -/// "test_mod", -/// )?; -/// -/// pyo3_asyncio::generic::into_stream_with_locals_v1::<MyCustomRuntime>( -/// pyo3_asyncio::generic::get_current_locals::<MyCustomRuntime>(py)?, -/// test_mod.call_method0("gen")? -/// ) -/// })?; -/// -/// let vals = stream -/// .map(|item| Python::with_gil(|py| -> PyResult<i32> { Ok(item?.as_ref(py).extract()?) })) -/// .try_collect::<Vec<i32>>() -/// .await?; -/// -/// assert_eq!((0..10).collect::<Vec<i32>>(), vals); -/// -/// Ok(()) -/// # } -/// ``` -#[cfg(feature = "unstable-streams")] -pub fn into_stream_with_locals_v1<'p, R>( - locals: TaskLocals, - gen: &'p PyAny, -) -> PyResult<impl futures::Stream<Item = PyResult<PyObject>> + 'static> -where - R: Runtime, -{ - let (tx, rx) = async_channel::bounded(1); - let anext = PyObject::from(gen.getattr("__anext__")?); - - R::spawn(async move { - loop { - let fut = Python::with_gil(|py| -> PyResult<_> { - into_future_with_locals(&locals, anext.as_ref(py).call0()?) - }); - let item = match fut { - Ok(fut) => match fut.await { - Ok(item) => Ok(item), - Err(e) => { - let stop_iter = Python::with_gil(|py| { - e.is_instance_of::<pyo3::exceptions::PyStopAsyncIteration>(py) - }); - - if stop_iter { - // end the iteration - break; - } else { - Err(e) - } - } - }, - Err(e) => Err(e), - }; - - if tx.send(item).await.is_err() { - // receiving side was dropped - break; - } - } - }); - - Ok(rx) -} - -/// <code>unstable-streams</code> Convert an async generator into a stream -/// -/// **This API is marked as unstable** and is only available when the -/// `unstable-streams` crate feature is enabled. This comes with no -/// stability guarantees, and could be changed or removed at any time. -/// -/// # Arguments -/// * `gen` - The Python async generator to be converted -/// -/// # Examples -/// ```no_run -/// # use std::{task::{Context, Poll}, pin::Pin, future::Future}; -/// # -/// # use pyo3_asyncio::{ -/// # TaskLocals, -/// # generic::{JoinError, ContextExt, Runtime} -/// # }; -/// # -/// # struct MyCustomJoinError; -/// # -/// # impl JoinError for MyCustomJoinError { -/// # fn is_panic(&self) -> bool { -/// # unreachable!() -/// # } -/// # } -/// # -/// # struct MyCustomJoinHandle; -/// # -/// # impl Future for MyCustomJoinHandle { -/// # type Output = Result<(), MyCustomJoinError>; -/// # -/// # fn poll(self: Pin<&mut Self>, _cx: &mut Context) -> Poll<Self::Output> { -/// # unreachable!() -/// # } -/// # } -/// # -/// # struct MyCustomRuntime; -/// # -/// # impl Runtime for MyCustomRuntime { -/// # type JoinError = MyCustomJoinError; -/// # type JoinHandle = MyCustomJoinHandle; -/// # -/// # fn spawn<F>(fut: F) -> Self::JoinHandle -/// # where -/// # F: Future<Output = ()> + Send + 'static -/// # { -/// # unreachable!() -/// # } -/// # } -/// # -/// # impl ContextExt for MyCustomRuntime { -/// # fn scope<F, R>(locals: TaskLocals, fut: F) -> Pin<Box<dyn Future<Output = R> + Send>> -/// # where -/// # F: Future<Output = R> + Send + 'static -/// # { -/// # unreachable!() -/// # } -/// # fn get_task_locals() -> Option<TaskLocals> { -/// # unreachable!() -/// # } -/// # } -/// -/// use pyo3::prelude::*; -/// use futures::{StreamExt, TryStreamExt}; -/// -/// const TEST_MOD: &str = r#" -/// import asyncio -/// -/// async def gen(): -/// for i in range(10): -/// await asyncio.sleep(0.1) -/// yield i -/// "#; -/// -/// # async fn test_async_gen() -> PyResult<()> { -/// let stream = Python::with_gil(|py| { -/// let test_mod = PyModule::from_code( -/// py, -/// TEST_MOD, -/// "test_rust_coroutine/test_mod.py", -/// "test_mod", -/// )?; -/// -/// pyo3_asyncio::generic::into_stream_v1::<MyCustomRuntime>(test_mod.call_method0("gen")?) -/// })?; -/// -/// let vals = stream -/// .map(|item| Python::with_gil(|py| -> PyResult<i32> { Ok(item?.as_ref(py).extract()?) })) -/// .try_collect::<Vec<i32>>() -/// .await?; -/// -/// assert_eq!((0..10).collect::<Vec<i32>>(), vals); -/// -/// Ok(()) -/// # } -/// ``` -#[cfg(feature = "unstable-streams")] -pub fn into_stream_v1<'p, R>( - gen: &'p PyAny, -) -> PyResult<impl futures::Stream<Item = PyResult<PyObject>> + 'static> -where - R: Runtime + ContextExt, -{ - into_stream_with_locals_v1::<R>(get_current_locals::<R>(gen.py())?, gen) -} - -fn py_true() -> PyObject { - static TRUE: OnceCell<PyObject> = OnceCell::new(); - TRUE.get_or_init(|| Python::with_gil(|py| true.into_py(py))) - .clone() -} -fn py_false() -> PyObject { - static FALSE: OnceCell<PyObject> = OnceCell::new(); - FALSE - .get_or_init(|| Python::with_gil(|py| false.into_py(py))) - .clone() -} - -trait Sender: Send + 'static { - fn send(&mut self, locals: TaskLocals, item: PyObject) -> PyResult<PyObject>; - fn close(&mut self) -> PyResult<()>; -} - -struct GenericSender<R> -where - R: Runtime, -{ - runtime: PhantomData<R>, - tx: mpsc::Sender<PyObject>, -} - -impl<R> Sender for GenericSender<R> -where - R: Runtime + ContextExt, -{ - fn send(&mut self, locals: TaskLocals, item: PyObject) -> PyResult<PyObject> { - match self.tx.try_send(item.clone()) { - Ok(_) => Ok(py_true()), - Err(e) => { - if e.is_full() { - let mut tx = self.tx.clone(); - Python::with_gil(move |py| { - Ok( - future_into_py_with_locals::<R, _, PyObject>(py, locals, async move { - if tx.flush().await.is_err() { - // receiving side disconnected - return Ok(py_false()); - } - if tx.send(item).await.is_err() { - // receiving side disconnected - return Ok(py_false()); - } - Ok(py_true()) - })? - .into(), - ) - }) - } else { - Ok(py_false()) - } - } - } - } - fn close(&mut self) -> PyResult<()> { - self.tx.close_channel(); - Ok(()) - } -} - -#[pyclass] -struct SenderGlue { - locals: TaskLocals, - tx: Box<dyn Sender>, -} -#[pymethods] -impl SenderGlue { - pub fn send(&mut self, item: PyObject) -> PyResult<PyObject> { - self.tx.send(self.locals.clone(), item) - } - pub fn close(&mut self) -> PyResult<()> { - self.tx.close() - } -} - -#[cfg(feature = "unstable-streams")] -const STREAM_GLUE: &str = r#" -import asyncio - -async def forward(gen, sender): - async for item in gen: - should_continue = sender.send(item) - - if asyncio.iscoroutine(should_continue): - should_continue = await should_continue - - if should_continue: - continue - else: - break - - sender.close() -"#; - -/// <code>unstable-streams</code> Convert an async generator into a stream -/// -/// **This API is marked as unstable** and is only available when the -/// `unstable-streams` crate feature is enabled. This comes with no -/// stability guarantees, and could be changed or removed at any time. -/// -/// # Arguments -/// * `locals` - The current task locals -/// * `gen` - The Python async generator to be converted -/// -/// # Examples -/// ```no_run -/// # use std::{task::{Context, Poll}, pin::Pin, future::Future}; -/// # -/// # use pyo3_asyncio::{ -/// # TaskLocals, -/// # generic::{JoinError, ContextExt, Runtime} -/// # }; -/// # -/// # struct MyCustomJoinError; -/// # -/// # impl JoinError for MyCustomJoinError { -/// # fn is_panic(&self) -> bool { -/// # unreachable!() -/// # } -/// # } -/// # -/// # struct MyCustomJoinHandle; -/// # -/// # impl Future for MyCustomJoinHandle { -/// # type Output = Result<(), MyCustomJoinError>; -/// # -/// # fn poll(self: Pin<&mut Self>, _cx: &mut Context) -> Poll<Self::Output> { -/// # unreachable!() -/// # } -/// # } -/// # -/// # struct MyCustomRuntime; -/// # -/// # impl Runtime for MyCustomRuntime { -/// # type JoinError = MyCustomJoinError; -/// # type JoinHandle = MyCustomJoinHandle; -/// # -/// # fn spawn<F>(fut: F) -> Self::JoinHandle -/// # where -/// # F: Future<Output = ()> + Send + 'static -/// # { -/// # unreachable!() -/// # } -/// # } -/// # -/// # impl ContextExt for MyCustomRuntime { -/// # fn scope<F, R>(locals: TaskLocals, fut: F) -> Pin<Box<dyn Future<Output = R> + Send>> -/// # where -/// # F: Future<Output = R> + Send + 'static -/// # { -/// # unreachable!() -/// # } -/// # fn get_task_locals() -> Option<TaskLocals> { -/// # unreachable!() -/// # } -/// # } -/// -/// use pyo3::prelude::*; -/// use futures::{StreamExt, TryStreamExt}; -/// -/// const TEST_MOD: &str = r#" -/// import asyncio -/// -/// async def gen(): -/// for i in range(10): -/// await asyncio.sleep(0.1) -/// yield i -/// "#; -/// -/// # async fn test_async_gen() -> PyResult<()> { -/// let stream = Python::with_gil(|py| { -/// let test_mod = PyModule::from_code( -/// py, -/// TEST_MOD, -/// "test_rust_coroutine/test_mod.py", -/// "test_mod", -/// )?; -/// -/// pyo3_asyncio::generic::into_stream_with_locals_v2::<MyCustomRuntime>( -/// pyo3_asyncio::generic::get_current_locals::<MyCustomRuntime>(py)?, -/// test_mod.call_method0("gen")? -/// ) -/// })?; -/// -/// let vals = stream -/// .map(|item| Python::with_gil(|py| -> PyResult<i32> { Ok(item.as_ref(py).extract()?) })) -/// .try_collect::<Vec<i32>>() -/// .await?; -/// -/// assert_eq!((0..10).collect::<Vec<i32>>(), vals); -/// -/// Ok(()) -/// # } -/// ``` -#[cfg(feature = "unstable-streams")] -pub fn into_stream_with_locals_v2<'p, R>( - locals: TaskLocals, - gen: &'p PyAny, -) -> PyResult<impl futures::Stream<Item = PyObject> + 'static> -where - R: Runtime + ContextExt, -{ - static GLUE_MOD: OnceCell<PyObject> = OnceCell::new(); - let py = gen.py(); - let glue = GLUE_MOD - .get_or_try_init(|| -> PyResult<PyObject> { - Ok(PyModule::from_code( - py, - STREAM_GLUE, - "pyo3_asyncio/pyo3_asyncio_glue.py", - "pyo3_asyncio_glue", - )? - .into()) - })? - .as_ref(py); - - let (tx, rx) = mpsc::channel(10); - - locals.event_loop(py).call_method1( - "call_soon_threadsafe", - ( - locals.event_loop(py).getattr("create_task")?, - glue.call_method1( - "forward", - ( - gen, - SenderGlue { - locals, - tx: Box::new(GenericSender { - runtime: PhantomData::<R>, - tx, - }), - }, - ), - )?, - ), - )?; - Ok(rx) -} - -/// <code>unstable-streams</code> Convert an async generator into a stream -/// -/// **This API is marked as unstable** and is only available when the -/// `unstable-streams` crate feature is enabled. This comes with no -/// stability guarantees, and could be changed or removed at any time. -/// -/// # Arguments -/// * `gen` - The Python async generator to be converted -/// -/// # Examples -/// ```no_run -/// # use std::{task::{Context, Poll}, pin::Pin, future::Future}; -/// # -/// # use pyo3_asyncio::{ -/// # TaskLocals, -/// # generic::{JoinError, ContextExt, Runtime} -/// # }; -/// # -/// # struct MyCustomJoinError; -/// # -/// # impl JoinError for MyCustomJoinError { -/// # fn is_panic(&self) -> bool { -/// # unreachable!() -/// # } -/// # } -/// # -/// # struct MyCustomJoinHandle; -/// # -/// # impl Future for MyCustomJoinHandle { -/// # type Output = Result<(), MyCustomJoinError>; -/// # -/// # fn poll(self: Pin<&mut Self>, _cx: &mut Context) -> Poll<Self::Output> { -/// # unreachable!() -/// # } -/// # } -/// # -/// # struct MyCustomRuntime; -/// # -/// # impl Runtime for MyCustomRuntime { -/// # type JoinError = MyCustomJoinError; -/// # type JoinHandle = MyCustomJoinHandle; -/// # -/// # fn spawn<F>(fut: F) -> Self::JoinHandle -/// # where -/// # F: Future<Output = ()> + Send + 'static -/// # { -/// # unreachable!() -/// # } -/// # } -/// # -/// # impl ContextExt for MyCustomRuntime { -/// # fn scope<F, R>(locals: TaskLocals, fut: F) -> Pin<Box<dyn Future<Output = R> + Send>> -/// # where -/// # F: Future<Output = R> + Send + 'static -/// # { -/// # unreachable!() -/// # } -/// # fn get_task_locals() -> Option<TaskLocals> { -/// # unreachable!() -/// # } -/// # } -/// -/// use pyo3::prelude::*; -/// use futures::{StreamExt, TryStreamExt}; -/// -/// const TEST_MOD: &str = r#" -/// import asyncio -/// -/// async def gen(): -/// for i in range(10): -/// await asyncio.sleep(0.1) -/// yield i -/// "#; -/// -/// # async fn test_async_gen() -> PyResult<()> { -/// let stream = Python::with_gil(|py| { -/// let test_mod = PyModule::from_code( -/// py, -/// TEST_MOD, -/// "test_rust_coroutine/test_mod.py", -/// "test_mod", -/// )?; -/// -/// pyo3_asyncio::generic::into_stream_v2::<MyCustomRuntime>(test_mod.call_method0("gen")?) -/// })?; -/// -/// let vals = stream -/// .map(|item| Python::with_gil(|py| -> PyResult<i32> { Ok(item.as_ref(py).extract()?) })) -/// .try_collect::<Vec<i32>>() -/// .await?; -/// -/// assert_eq!((0..10).collect::<Vec<i32>>(), vals); -/// -/// Ok(()) -/// # } -/// ``` -#[cfg(feature = "unstable-streams")] -pub fn into_stream_v2<'p, R>( - gen: &'p PyAny, -) -> PyResult<impl futures::Stream<Item = PyObject> + 'static> -where - R: Runtime + ContextExt, -{ - into_stream_with_locals_v2::<R>(get_current_locals::<R>(gen.py())?, gen) -} diff --git a/lib/pyo3-asyncio/src/lib.rs b/lib/pyo3-asyncio/src/lib.rs deleted file mode 100644 index c3c820fc..00000000 --- a/lib/pyo3-asyncio/src/lib.rs +++ /dev/null @@ -1,663 +0,0 @@ -#![warn(missing_docs)] -#![allow(clippy::borrow_deref_ref)] - -//! Rust Bindings to the Python Asyncio Event Loop -//! -//! # Motivation -//! -//! This crate aims to provide a convenient interface to manage the interop between Python and -//! Rust's async/await models. It supports conversions between Rust and Python futures and manages -//! the event loops for both languages. Python's threading model and GIL can make this interop a bit -//! trickier than one might expect, so there are a few caveats that users should be aware of. -//! -//! ## Why Two Event Loops -//! -//! Currently, we don't have a way to run Rust futures directly on Python's event loop. Likewise, -//! Python's coroutines cannot be directly spawned on a Rust event loop. The two coroutine models -//! require some additional assistance from their event loops, so in all likelihood they will need -//! a new _unique_ event loop that addresses the needs of both languages if the coroutines are to -//! be run on the same loop. -//! -//! It's not immediately clear that this would provide worthwhile performance wins either, so in the -//! interest of getting something simple out there to facilitate these conversions, this crate -//! handles the communication between _separate_ Python and Rust event loops. -//! -//! ## Python's Event Loop and the Main Thread -//! -//! Python is very picky about the threads used by the `asyncio` executor. In particular, it needs -//! to have control over the main thread in order to handle signals like CTRL-C correctly. This -//! means that Cargo's default test harness will no longer work since it doesn't provide a method of -//! overriding the main function to add our event loop initialization and finalization. -//! -//! ## Event Loop References and ContextVars -//! -//! One problem that arises when interacting with Python's asyncio library is that the functions we -//! use to get a reference to the Python event loop can only be called in certain contexts. Since -//! PyO3 Asyncio needs to interact with Python's event loop during conversions, the context of these -//! conversions can matter a lot. -//! -//! Likewise, Python's `contextvars` library can require some special treatment. Python functions -//! and coroutines can rely on the context of outer coroutines to function correctly, so this -//! library needs to be able to preserve `contextvars` during conversions. -//! -//! > The core conversions we've mentioned so far in the README should insulate you from these -//! concerns in most cases. For the edge cases where they don't, this section should provide you -//! with the information you need to solve these problems. -//! -//! ### The Main Dilemma -//! -//! Python programs can have many independent event loop instances throughout the lifetime of the -//! application (`asyncio.run` for example creates its own event loop each time it's called for -//! instance), and they can even run concurrent with other event loops. For this reason, the most -//! correct method of obtaining a reference to the Python event loop is via -//! `asyncio.get_running_loop`. -//! -//! `asyncio.get_running_loop` returns the event loop associated with the current OS thread. It can -//! be used inside Python coroutines to spawn concurrent tasks, interact with timers, or in our case -//! signal between Rust and Python. This is all well and good when we are operating on a Python -//! thread, but since Rust threads are not associated with a Python event loop, -//! `asyncio.get_running_loop` will fail when called on a Rust runtime. -//! -//! `contextvars` operates in a similar way, though the current context is not always associated -//! with the current OS thread. Different contexts can be associated with different coroutines even -//! if they run on the same OS thread. -//! -//! ### The Solution -//! -//! A really straightforward way of dealing with this problem is to pass references to the -//! associated Python event loop and context for every conversion. That's why we have a structure -//! called `TaskLocals` and a set of conversions that accept it. -//! -//! `TaskLocals` stores the current event loop, and allows the user to copy the current Python -//! context if necessary. The following conversions will use these references to perform the -//! necessary conversions and restore Python context when needed: -//! -//! - `pyo3_asyncio::into_future_with_locals` - Convert a Python awaitable into a Rust future. -//! - `pyo3_asyncio::<runtime>::future_into_py_with_locals` - Convert a Rust future into a Python -//! awaitable. -//! - `pyo3_asyncio::<runtime>::local_future_into_py_with_locals` - Convert a `!Send` Rust future -//! into a Python awaitable. -//! -//! One clear disadvantage to this approach is that the Rust application has to explicitly track -//! these references. In native libraries, we can't make any assumptions about the underlying event -//! loop, so the only reliable way to make sure our conversions work properly is to store these -//! references at the callsite to use later on. -//! -//! ```rust -//! use pyo3::{wrap_pyfunction, prelude::*}; -//! -//! # #[cfg(feature = "tokio-runtime")] -//! #[pyfunction] -//! fn sleep(py: Python) -> PyResult<&PyAny> { -//! // Construct the task locals structure with the current running loop and context -//! let locals = pyo3_asyncio::TaskLocals::with_running_loop(py)?.copy_context(py)?; -//! -//! // Convert the async move { } block to a Python awaitable -//! pyo3_asyncio::tokio::future_into_py_with_locals(py, locals.clone(), async move { -//! let py_sleep = Python::with_gil(|py| { -//! // Sometimes we need to call other async Python functions within -//! // this future. In order for this to work, we need to track the -//! // event loop from earlier. -//! pyo3_asyncio::into_future_with_locals( -//! &locals, -//! py.import("asyncio")?.call_method1("sleep", (1,))? -//! ) -//! })?; -//! -//! py_sleep.await?; -//! -//! Ok(()) -//! }) -//! } -//! -//! # #[cfg(feature = "tokio-runtime")] -//! #[pymodule] -//! fn my_mod(py: Python, m: &PyModule) -> PyResult<()> { -//! m.add_function(wrap_pyfunction!(sleep, m)?)?; -//! Ok(()) -//! } -//! ``` -//! -//! > A naive solution to this tracking problem would be to cache a global reference to the asyncio -//! event loop that all PyO3 Asyncio conversions can use. In fact this is what we did in PyO3 -//! Asyncio `v0.13`. This works well for applications, but it soon became clear that this is not -//! so ideal for libraries. Libraries usually have no direct control over how the event loop is -//! managed, they're just expected to work with any event loop at any point in the application. -//! This problem is compounded further when multiple event loops are used in the application since -//! the global reference will only point to one. -//! -//! Another disadvantage to this explicit approach that is less obvious is that we can no longer -//! call our `#[pyfunction] fn sleep` on a Rust runtime since `asyncio.get_running_loop` only works -//! on Python threads! It's clear that we need a slightly more flexible approach. -//! -//! In order to detect the Python event loop at the callsite, we need something like -//! `asyncio.get_running_loop` and `contextvars.copy_context` that works for _both Python and Rust_. -//! In Python, `asyncio.get_running_loop` uses thread-local data to retrieve the event loop -//! associated with the current thread. What we need in Rust is something that can retrieve the -//! Python event loop and contextvars associated with the current Rust _task_. -//! -//! Enter `pyo3_asyncio::<runtime>::get_current_locals`. This function first checks task-local data -//! for the `TaskLocals`, then falls back on `asyncio.get_running_loop` and -//! `contextvars.copy_context` if no task locals are found. This way both bases are -//! covered. -//! -//! Now, all we need is a way to store the `TaskLocals` for the Rust future. Since this is a -//! runtime-specific feature, you can find the following functions in each runtime module: -//! -//! - `pyo3_asyncio::<runtime>::scope` - Store the task-local data when executing the given Future. -//! - `pyo3_asyncio::<runtime>::scope_local` - Store the task-local data when executing the given -//! `!Send` Future. -//! -//! With these new functions, we can make our previous example more correct: -//! -//! ```rust no_run -//! use pyo3::prelude::*; -//! -//! # #[cfg(feature = "tokio-runtime")] -//! #[pyfunction] -//! fn sleep(py: Python) -> PyResult<&PyAny> { -//! // get the current event loop through task-local data -//! // OR `asyncio.get_running_loop` and `contextvars.copy_context` -//! let locals = pyo3_asyncio::tokio::get_current_locals(py)?; -//! -//! pyo3_asyncio::tokio::future_into_py_with_locals( -//! py, -//! locals.clone(), -//! // Store the current locals in task-local data -//! pyo3_asyncio::tokio::scope(locals.clone(), async move { -//! let py_sleep = Python::with_gil(|py| { -//! pyo3_asyncio::into_future_with_locals( -//! // Now we can get the current locals through task-local data -//! &pyo3_asyncio::tokio::get_current_locals(py)?, -//! py.import("asyncio")?.call_method1("sleep", (1,))? -//! ) -//! })?; -//! -//! py_sleep.await?; -//! -//! Ok(Python::with_gil(|py| py.None())) -//! }) -//! ) -//! } -//! -//! # #[cfg(feature = "tokio-runtime")] -//! #[pyfunction] -//! fn wrap_sleep(py: Python) -> PyResult<&PyAny> { -//! // get the current event loop through task-local data -//! // OR `asyncio.get_running_loop` and `contextvars.copy_context` -//! let locals = pyo3_asyncio::tokio::get_current_locals(py)?; -//! -//! pyo3_asyncio::tokio::future_into_py_with_locals( -//! py, -//! locals.clone(), -//! // Store the current locals in task-local data -//! pyo3_asyncio::tokio::scope(locals.clone(), async move { -//! let py_sleep = Python::with_gil(|py| { -//! pyo3_asyncio::into_future_with_locals( -//! &pyo3_asyncio::tokio::get_current_locals(py)?, -//! // We can also call sleep within a Rust task since the -//! // locals are stored in task local data -//! sleep(py)? -//! ) -//! })?; -//! -//! py_sleep.await?; -//! -//! Ok(Python::with_gil(|py| py.None())) -//! }) -//! ) -//! } -//! -//! # #[cfg(feature = "tokio-runtime")] -//! #[pymodule] -//! fn my_mod(py: Python, m: &PyModule) -> PyResult<()> { -//! m.add_function(wrap_pyfunction!(sleep, m)?)?; -//! m.add_function(wrap_pyfunction!(wrap_sleep, m)?)?; -//! Ok(()) -//! } -//! ``` -//! -//! Even though this is more correct, it's clearly not more ergonomic. That's why we introduced a -//! set of functions with this functionality baked in: -//! -//! - `pyo3_asyncio::<runtime>::into_future` -//! > Convert a Python awaitable into a Rust future (using -//! `pyo3_asyncio::<runtime>::get_current_locals`) -//! - `pyo3_asyncio::<runtime>::future_into_py` -//! > Convert a Rust future into a Python awaitable (using -//! `pyo3_asyncio::<runtime>::get_current_locals` and `pyo3_asyncio::<runtime>::scope` to set the -//! task-local event loop for the given Rust future) -//! - `pyo3_asyncio::<runtime>::local_future_into_py` -//! > Convert a `!Send` Rust future into a Python awaitable (using -//! `pyo3_asyncio::<runtime>::get_current_locals` and `pyo3_asyncio::<runtime>::scope_local` to -//! set the task-local event loop for the given Rust future). -//! -//! __These are the functions that we recommend using__. With these functions, the previous example -//! can be rewritten to be more compact: -//! -//! ```rust -//! use pyo3::prelude::*; -//! -//! # #[cfg(feature = "tokio-runtime")] -//! #[pyfunction] -//! fn sleep(py: Python) -> PyResult<&PyAny> { -//! pyo3_asyncio::tokio::future_into_py(py, async move { -//! let py_sleep = Python::with_gil(|py| { -//! pyo3_asyncio::tokio::into_future( -//! py.import("asyncio")?.call_method1("sleep", (1,))? -//! ) -//! })?; -//! -//! py_sleep.await?; -//! -//! Ok(Python::with_gil(|py| py.None())) -//! }) -//! } -//! -//! # #[cfg(feature = "tokio-runtime")] -//! #[pyfunction] -//! fn wrap_sleep(py: Python) -> PyResult<&PyAny> { -//! pyo3_asyncio::tokio::future_into_py(py, async move { -//! let py_sleep = Python::with_gil(|py| { -//! pyo3_asyncio::tokio::into_future(sleep(py)?) -//! })?; -//! -//! py_sleep.await?; -//! -//! Ok(Python::with_gil(|py| py.None())) -//! }) -//! } -//! -//! # #[cfg(feature = "tokio-runtime")] -//! #[pymodule] -//! fn my_mod(py: Python, m: &PyModule) -> PyResult<()> { -//! m.add_function(wrap_pyfunction!(sleep, m)?)?; -//! m.add_function(wrap_pyfunction!(wrap_sleep, m)?)?; -//! Ok(()) -//! } -//! ``` -//! -//! > A special thanks to [@ShadowJonathan](https://github.com/ShadowJonathan) for helping with the -//! design and review of these changes! -//! -//! ## Rust's Event Loop -//! -//! Currently only the Async-Std and Tokio runtimes are supported by this crate. If you need support -//! for another runtime, feel free to make a request on GitHub (or attempt to add support yourself -//! with the [`generic`] module)! -//! -//! > _In the future, we may implement first class support for more Rust runtimes. Contributions are -//! welcome as well!_ -//! -//! ## Features -//! -//! Items marked with -//! <code>attributes</code> -//! are only available when the `attributes` Cargo feature is enabled: -//! -//! ```toml -//! [dependencies.pyo3-asyncio] -//! version = "0.17" -//! features = ["attributes"] -//! ``` -//! -//! Items marked with -//! <code>async-std-runtime</code> -//! are only available when the `async-std-runtime` Cargo feature is enabled: -//! -//! ```toml -//! [dependencies.pyo3-asyncio] -//! version = "0.17" -//! features = ["async-std-runtime"] -//! ``` -//! -//! Items marked with -//! <code>tokio-runtime</code> -//! are only available when the `tokio-runtime` Cargo feature is enabled: -//! -//! ```toml -//! [dependencies.pyo3-asyncio] -//! version = "0.17" -//! features = ["tokio-runtime"] -//! ``` -//! -//! Items marked with -//! <code>testing</code> -//! are only available when the `testing` Cargo feature is enabled: -//! -//! ```toml -//! [dependencies.pyo3-asyncio] -//! version = "0.17" -//! features = ["testing"] -//! ``` - -/// Re-exported for #[test] attributes -#[cfg(all(feature = "attributes", feature = "testing"))] -pub use inventory; - -/// <code>testing</code> Utilities for writing PyO3 Asyncio tests -#[cfg(feature = "testing")] -pub mod testing; - -#[cfg(feature = "async-std")] -pub mod async_std; - -#[cfg(feature = "tokio-runtime")] -pub mod tokio; - -/// Errors and exceptions related to PyO3 Asyncio -pub mod err; - -pub mod generic; - -/// Test README -#[doc(hidden)] -pub mod doc_test { - #[allow(unused)] - macro_rules! doc_comment { - ($x:expr, $module:item) => { - #[doc = $x] - $module - }; - } - - #[allow(unused)] - macro_rules! doctest { - ($x:expr, $y:ident) => { - doc_comment!(include_str!($x), mod $y {}); - }; - } - - #[cfg(all( - feature = "async-std-runtime", - feature = "tokio-runtime", - feature = "attributes" - ))] - doctest!("../README.md", readme_md); -} - -use std::future::Future; - -use futures::channel::oneshot; -use once_cell::sync::OnceCell; -use pyo3::{ - prelude::*, - types::{PyDict, PyTuple}, -}; - -static ASYNCIO: OnceCell<PyObject> = OnceCell::new(); -static CONTEXTVARS: OnceCell<PyObject> = OnceCell::new(); -static ENSURE_FUTURE: OnceCell<PyObject> = OnceCell::new(); -static GET_RUNNING_LOOP: OnceCell<PyObject> = OnceCell::new(); - -fn ensure_future<'p>(py: Python<'p>, awaitable: &'p PyAny) -> PyResult<&'p PyAny> { - ENSURE_FUTURE - .get_or_try_init(|| -> PyResult<PyObject> { - Ok(asyncio(py)?.getattr("ensure_future")?.into()) - })? - .as_ref(py) - .call1((awaitable,)) -} - -pub fn create_future(event_loop: &PyAny) -> PyResult<&PyAny> { - event_loop.call_method0("create_future") -} - -fn close(event_loop: &PyAny) -> PyResult<()> { - event_loop.call_method1( - "run_until_complete", - (event_loop.call_method0("shutdown_asyncgens")?,), - )?; - - // how to do this prior to 3.9? - if event_loop.hasattr("shutdown_default_executor")? { - event_loop.call_method1( - "run_until_complete", - (event_loop.call_method0("shutdown_default_executor")?,), - )?; - } - - event_loop.call_method0("close")?; - - Ok(()) -} - -fn asyncio(py: Python) -> PyResult<&PyAny> { - ASYNCIO - .get_or_try_init(|| Ok(py.import("asyncio")?.into())) - .map(|asyncio| asyncio.as_ref(py)) -} - -/// Get a reference to the Python Event Loop from Rust -/// -/// Equivalent to `asyncio.get_running_loop()` in Python 3.7+. -pub fn get_running_loop(py: Python) -> PyResult<&PyAny> { - // Ideally should call get_running_loop, but calls get_event_loop for compatibility when - // get_running_loop is not available. - GET_RUNNING_LOOP - .get_or_try_init(|| -> PyResult<PyObject> { - let asyncio = asyncio(py)?; - - Ok(asyncio.getattr("get_running_loop")?.into()) - })? - .as_ref(py) - .call0() -} - -fn contextvars(py: Python) -> PyResult<&PyAny> { - Ok(CONTEXTVARS - .get_or_try_init(|| py.import("contextvars").map(|m| m.into()))? - .as_ref(py)) -} - -fn copy_context(py: Python) -> PyResult<&PyAny> { - contextvars(py)?.call_method0("copy_context") -} - -/// Task-local data to store for Python conversions. -#[derive(Debug, Clone)] -pub struct TaskLocals { - /// Track the event loop of the Python task - event_loop: PyObject, - /// Track the contextvars of the Python task - context: PyObject, -} - -impl TaskLocals { - /// At a minimum, TaskLocals must store the event loop. - pub fn new(event_loop: &PyAny) -> Self { - Self { - event_loop: event_loop.into(), - context: event_loop.py().None(), - } - } - - /// Construct TaskLocals with the event loop returned by `get_running_loop` - pub fn with_running_loop(py: Python) -> PyResult<Self> { - Ok(Self::new(get_running_loop(py)?)) - } - - /// Manually provide the contextvars for the current task. - pub fn with_context(self, context: &PyAny) -> Self { - Self { - context: context.into(), - ..self - } - } - - /// Capture the current task's contextvars - pub fn copy_context(self, py: Python) -> PyResult<Self> { - Ok(self.with_context(copy_context(py)?)) - } - - /// Get a reference to the event loop - pub fn event_loop<'p>(&self, py: Python<'p>) -> &'p PyAny { - self.event_loop.clone().into_ref(py) - } - - /// Get a reference to the python context - pub fn context<'p>(&self, py: Python<'p>) -> &'p PyAny { - self.context.clone().into_ref(py) - } -} - -#[pyclass] -struct PyTaskCompleter { - tx: Option<oneshot::Sender<PyResult<PyObject>>>, -} - -#[pymethods] -impl PyTaskCompleter { - #[args(task)] - pub fn __call__(&mut self, task: &PyAny) -> PyResult<()> { - debug_assert!(task.call_method0("done")?.extract()?); - let result = match task.call_method0("result") { - Ok(val) => Ok(val.into()), - Err(e) => Err(e), - }; - - // unclear to me whether or not this should be a panic or silent error. - // - // calling PyTaskCompleter twice should not be possible, but I don't think it really hurts - // anything if it happens. - if let Some(tx) = self.tx.take() { - if tx.send(result).is_err() { - // cancellation is not an error - } - } - - Ok(()) - } -} - -#[pyclass] -struct PyEnsureFuture { - awaitable: PyObject, - tx: Option<oneshot::Sender<PyResult<PyObject>>>, -} - -#[pymethods] -impl PyEnsureFuture { - pub fn __call__(&mut self) -> PyResult<()> { - Python::with_gil(|py| { - let task = ensure_future(py, self.awaitable.as_ref(py))?; - let on_complete = PyTaskCompleter { tx: self.tx.take() }; - task.call_method1("add_done_callback", (on_complete,))?; - - Ok(()) - }) - } -} - -fn call_soon_threadsafe( - event_loop: &PyAny, - context: &PyAny, - args: impl IntoPy<Py<PyTuple>>, -) -> PyResult<()> { - let py = event_loop.py(); - - let kwargs = PyDict::new(py); - kwargs.set_item("context", context)?; - - event_loop.call_method("call_soon_threadsafe", args, Some(kwargs))?; - Ok(()) -} - -/// Convert a Python `awaitable` into a Rust Future -/// -/// This function converts the `awaitable` into a Python Task using `run_coroutine_threadsafe`. A -/// completion handler sends the result of this Task through a -/// `futures::channel::oneshot::Sender<PyResult<PyObject>>` and the future returned by this function -/// simply awaits the result through the `futures::channel::oneshot::Receiver<PyResult<PyObject>>`. -/// -/// # Arguments -/// * `locals` - The Python event loop and context to be used for the provided awaitable -/// * `awaitable` - The Python `awaitable` to be converted -/// -/// # Examples -/// -/// ``` -/// use std::time::Duration; -/// -/// use pyo3::prelude::*; -/// -/// const PYTHON_CODE: &'static str = r#" -/// import asyncio -/// -/// async def py_sleep(duration): -/// await asyncio.sleep(duration) -/// "#; -/// -/// # #[cfg(feature = "tokio-runtime")] -/// async fn py_sleep(seconds: f32) -> PyResult<()> { -/// let test_mod = Python::with_gil(|py| -> PyResult<PyObject> { -/// Ok( -/// PyModule::from_code( -/// py, -/// PYTHON_CODE, -/// "test_into_future/test_mod.py", -/// "test_mod" -/// )? -/// .into() -/// ) -/// })?; -/// -/// Python::with_gil(|py| { -/// pyo3_asyncio::into_future_with_locals( -/// &pyo3_asyncio::tokio::get_current_locals(py)?, -/// test_mod -/// .call_method1(py, "py_sleep", (seconds.into_py(py),))? -/// .as_ref(py), -/// ) -/// })? -/// .await?; -/// Ok(()) -/// } -/// ``` -pub fn into_future_with_locals( - locals: &TaskLocals, - awaitable: &PyAny, -) -> PyResult<impl Future<Output = PyResult<PyObject>> + Send> { - let py = awaitable.py(); - let (tx, rx) = oneshot::channel(); - - call_soon_threadsafe( - locals.event_loop(py), - locals.context(py), - (PyEnsureFuture { - awaitable: awaitable.into(), - tx: Some(tx), - },), - )?; - - Ok(async move { - match rx.await { - Ok(item) => item, - Err(_) => Python::with_gil(|py| { - Err(PyErr::from_value( - asyncio(py)?.call_method0("CancelledError")?, - )) - }), - } - }) -} - -pub fn dump_err(py: Python<'_>) -> impl FnOnce(PyErr) + '_ { - move |e| { - // We can't display Python exceptions via std::fmt::Display, - // so print the error here manually. - e.print_and_set_sys_last_vars(py); - } -} diff --git a/lib/pyo3-asyncio/src/tokio.rs b/lib/pyo3-asyncio/src/tokio.rs deleted file mode 100644 index b56378b6..00000000 --- a/lib/pyo3-asyncio/src/tokio.rs +++ /dev/null @@ -1,787 +0,0 @@ -//! <code>tokio-runtime</code> PyO3 Asyncio functions specific to the tokio runtime -//! -//! Items marked with -//! <code>unstable-streams</code> -//! are only available when the `unstable-streams` Cargo feature is enabled: -//! -//! ```toml -//! [dependencies.pyo3-asyncio] -//! version = "0.17" -//! features = ["unstable-streams"] -//! ``` - -use std::{future::Future, pin::Pin, sync::Mutex}; - -use ::tokio::{ - runtime::{Builder, Runtime}, - task, -}; -use once_cell::{ - sync::{Lazy, OnceCell}, - unsync::OnceCell as UnsyncOnceCell, -}; -use pyo3::prelude::*; - -use crate::{ - generic::{self, ContextExt, LocalContextExt, Runtime as GenericRuntime, SpawnLocalExt}, - TaskLocals, -}; - -/// <code>attributes</code> -/// re-exports for macros -#[cfg(feature = "attributes")] -pub mod re_exports { - /// re-export pending to be used in tokio macros without additional dependency - pub use futures::future::pending; - /// re-export tokio::runtime to build runtimes in tokio macros without additional dependency - pub use tokio::runtime; -} - -/// <code>attributes</code> -#[cfg(feature = "attributes")] -pub use pyo3_asyncio_macros::tokio_main as main; - -/// <code>attributes</code> -/// <code>testing</code> -/// Registers a `tokio` test with the `pyo3-asyncio` test harness -#[cfg(all(feature = "attributes", feature = "testing"))] -pub use pyo3_asyncio_macros::tokio_test as test; - -static TOKIO_BUILDER: Lazy<Mutex<Builder>> = Lazy::new(|| Mutex::new(multi_thread())); -static TOKIO_RUNTIME: OnceCell<Runtime> = OnceCell::new(); - -impl generic::JoinError for task::JoinError { - fn is_panic(&self) -> bool { - task::JoinError::is_panic(self) - } -} - -struct TokioRuntime; - -tokio::task_local! { - static TASK_LOCALS: UnsyncOnceCell<TaskLocals>; -} - -impl GenericRuntime for TokioRuntime { - type JoinError = task::JoinError; - type JoinHandle = task::JoinHandle<()>; - - fn spawn<F>(fut: F) -> Self::JoinHandle - where - F: Future<Output = ()> + Send + 'static, - { - get_runtime().spawn(async move { - fut.await; - }) - } -} - -impl ContextExt for TokioRuntime { - fn scope<F, R>(locals: TaskLocals, fut: F) -> Pin<Box<dyn Future<Output = R> + Send>> - where - F: Future<Output = R> + Send + 'static, - { - let cell = UnsyncOnceCell::new(); - cell.set(locals).unwrap(); - - Box::pin(TASK_LOCALS.scope(cell, fut)) - } - - fn get_task_locals() -> Option<TaskLocals> { - match TASK_LOCALS.try_with(|c| c.get().map(|locals| locals.clone())) { - Ok(locals) => locals, - Err(_) => None, - } - } -} - -impl SpawnLocalExt for TokioRuntime { - fn spawn_local<F>(fut: F) -> Self::JoinHandle - where - F: Future<Output = ()> + 'static, - { - tokio::task::spawn_local(fut) - } -} - -impl LocalContextExt for TokioRuntime { - fn scope_local<F, R>(locals: TaskLocals, fut: F) -> Pin<Box<dyn Future<Output = R>>> - where - F: Future<Output = R> + 'static, - { - let cell = UnsyncOnceCell::new(); - cell.set(locals).unwrap(); - - Box::pin(TASK_LOCALS.scope(cell, fut)) - } -} - -/// Set the task local event loop for the given future -pub async fn scope<F, R>(locals: TaskLocals, fut: F) -> R -where - F: Future<Output = R> + Send + 'static, -{ - TokioRuntime::scope(locals, fut).await -} - -/// Set the task local event loop for the given !Send future -pub async fn scope_local<F, R>(locals: TaskLocals, fut: F) -> R -where - F: Future<Output = R> + 'static, -{ - TokioRuntime::scope_local(locals, fut).await -} - -/// Get the current event loop from either Python or Rust async task local context -/// -/// This function first checks if the runtime has a task-local reference to the Python event loop. -/// If not, it calls [`get_running_loop`](`crate::get_running_loop`) to get the event loop -/// associated with the current OS thread. -pub fn get_current_loop(py: Python) -> PyResult<&PyAny> { - generic::get_current_loop::<TokioRuntime>(py) -} - -/// Either copy the task locals from the current task OR get the current running loop and -/// contextvars from Python. -pub fn get_current_locals(py: Python) -> PyResult<TaskLocals> { - generic::get_current_locals::<TokioRuntime>(py) -} - -/// Initialize the Tokio runtime with a custom build -pub fn init(builder: Builder) { - *TOKIO_BUILDER.lock().unwrap() = builder -} - -/// Get a reference to the current tokio runtime -pub fn get_runtime<'a>() -> &'a Runtime { - TOKIO_RUNTIME.get_or_init(|| { - TOKIO_BUILDER - .lock() - .unwrap() - .build() - .expect("Unable to build Tokio runtime") - }) -} - -fn multi_thread() -> Builder { - let mut builder = Builder::new_multi_thread(); - builder.enable_all(); - builder -} - -/// Run the event loop until the given Future completes -/// -/// The event loop runs until the given future is complete. -/// -/// After this function returns, the event loop can be resumed with [`run_until_complete`] -/// -/// # Arguments -/// * `event_loop` - The Python event loop that should run the future -/// * `fut` - The future to drive to completion -/// -/// # Examples -/// -/// ``` -/// # use std::time::Duration; -/// # -/// # use pyo3::prelude::*; -/// # -/// # pyo3::prepare_freethreaded_python(); -/// # Python::with_gil(|py| -> PyResult<()> { -/// # let event_loop = py.import("asyncio")?.call_method0("new_event_loop")?; -/// pyo3_asyncio::tokio::run_until_complete(event_loop, async move { -/// tokio::time::sleep(Duration::from_secs(1)).await; -/// Ok(()) -/// })?; -/// # Ok(()) -/// # }).unwrap(); -/// ``` -pub fn run_until_complete<F, T>(event_loop: &PyAny, fut: F) -> PyResult<T> -where - F: Future<Output = PyResult<T>> + Send + 'static, - T: Send + Sync + 'static, -{ - generic::run_until_complete::<TokioRuntime, _, T>(event_loop, fut) -} - -/// Run the event loop until the given Future completes -/// -/// # Arguments -/// * `py` - The current PyO3 GIL guard -/// * `fut` - The future to drive to completion -/// -/// # Examples -/// -/// ```no_run -/// # use std::time::Duration; -/// # -/// # use pyo3::prelude::*; -/// # -/// fn main() { -/// Python::with_gil(|py| { -/// pyo3_asyncio::tokio::run(py, async move { -/// tokio::time::sleep(Duration::from_secs(1)).await; -/// Ok(()) -/// }) -/// .map_err(|e| { -/// e.print_and_set_sys_last_vars(py); -/// }) -/// .unwrap(); -/// }) -/// } -/// ``` -pub fn run<F, T>(py: Python, fut: F) -> PyResult<T> -where - F: Future<Output = PyResult<T>> + Send + 'static, - T: Send + Sync + 'static, -{ - generic::run::<TokioRuntime, F, T>(py, fut) -} - -/// Convert a Rust Future into a Python awaitable -/// -/// If the `asyncio.Future` returned by this conversion is cancelled via `asyncio.Future.cancel`, -/// the Rust future will be cancelled as well (new behaviour in `v0.15`). -/// -/// Python `contextvars` are preserved when calling async Python functions within the Rust future -/// via [`into_future`] (new behaviour in `v0.15`). -/// -/// > Although `contextvars` are preserved for async Python functions, synchronous functions will -/// unfortunately fail to resolve them when called within the Rust future. This is because the -/// function is being called from a Rust thread, not inside an actual Python coroutine context. -/// > -/// > As a workaround, you can get the `contextvars` from the current task locals using -/// [`get_current_locals`] and [`TaskLocals::context`](`crate::TaskLocals::context`), then wrap your -/// synchronous function in a call to `contextvars.Context.run`. This will set the context, call the -/// synchronous function, and restore the previous context when it returns or raises an exception. -/// -/// # Arguments -/// * `py` - PyO3 GIL guard -/// * `locals` - The task locals for the given future -/// * `fut` - The Rust future to be converted -/// -/// # Examples -/// -/// ``` -/// use std::time::Duration; -/// -/// use pyo3::prelude::*; -/// -/// /// Awaitable sleep function -/// #[pyfunction] -/// fn sleep_for<'p>(py: Python<'p>, secs: &'p PyAny) -> PyResult<&'p PyAny> { -/// let secs = secs.extract()?; -/// pyo3_asyncio::tokio::future_into_py_with_locals( -/// py, -/// pyo3_asyncio::tokio::get_current_locals(py)?, -/// async move { -/// tokio::time::sleep(Duration::from_secs(secs)).await; -/// Python::with_gil(|py| Ok(py.None())) -/// } -/// ) -/// } -/// ``` -pub fn future_into_py_with_locals<F, T>(py: Python, locals: TaskLocals, fut: F) -> PyResult<&PyAny> -where - F: Future<Output = PyResult<T>> + Send + 'static, - T: IntoPy<PyObject>, -{ - generic::future_into_py_with_locals::<TokioRuntime, F, T>(py, locals, fut) -} - -/// Convert a Rust Future into a Python awaitable -/// -/// If the `asyncio.Future` returned by this conversion is cancelled via `asyncio.Future.cancel`, -/// the Rust future will be cancelled as well (new behaviour in `v0.15`). -/// -/// Python `contextvars` are preserved when calling async Python functions within the Rust future -/// via [`into_future`] (new behaviour in `v0.15`). -/// -/// > Although `contextvars` are preserved for async Python functions, synchronous functions will -/// unfortunately fail to resolve them when called within the Rust future. This is because the -/// function is being called from a Rust thread, not inside an actual Python coroutine context. -/// > -/// > As a workaround, you can get the `contextvars` from the current task locals using -/// [`get_current_locals`] and [`TaskLocals::context`](`crate::TaskLocals::context`), then wrap your -/// synchronous function in a call to `contextvars.Context.run`. This will set the context, call the -/// synchronous function, and restore the previous context when it returns or raises an exception. -/// -/// # Arguments -/// * `py` - The current PyO3 GIL guard -/// * `fut` - The Rust future to be converted -/// -/// # Examples -/// -/// ``` -/// use std::time::Duration; -/// -/// use pyo3::prelude::*; -/// -/// /// Awaitable sleep function -/// #[pyfunction] -/// fn sleep_for<'p>(py: Python<'p>, secs: &'p PyAny) -> PyResult<&'p PyAny> { -/// let secs = secs.extract()?; -/// pyo3_asyncio::tokio::future_into_py(py, async move { -/// tokio::time::sleep(Duration::from_secs(secs)).await; -/// Ok(()) -/// }) -/// } -/// ``` -pub fn future_into_py<F, T>(py: Python, fut: F) -> PyResult<&PyAny> -where - F: Future<Output = PyResult<T>> + Send + 'static, - T: IntoPy<PyObject>, -{ - generic::future_into_py::<TokioRuntime, _, T>(py, fut) -} - -/// Convert a `!Send` Rust Future into a Python awaitable -/// -/// If the `asyncio.Future` returned by this conversion is cancelled via `asyncio.Future.cancel`, -/// the Rust future will be cancelled as well (new behaviour in `v0.15`). -/// -/// Python `contextvars` are preserved when calling async Python functions within the Rust future -/// via [`into_future`] (new behaviour in `v0.15`). -/// -/// > Although `contextvars` are preserved for async Python functions, synchronous functions will -/// unfortunately fail to resolve them when called within the Rust future. This is because the -/// function is being called from a Rust thread, not inside an actual Python coroutine context. -/// > -/// > As a workaround, you can get the `contextvars` from the current task locals using -/// [`get_current_locals`] and [`TaskLocals::context`](`crate::TaskLocals::context`), then wrap your -/// synchronous function in a call to `contextvars.Context.run`. This will set the context, call the -/// synchronous function, and restore the previous context when it returns or raises an exception. -/// -/// # Arguments -/// * `py` - PyO3 GIL guard -/// * `locals` - The task locals for the given future -/// * `fut` - The Rust future to be converted -/// -/// # Examples -/// -/// ``` -/// use std::{rc::Rc, time::Duration}; -/// -/// use pyo3::prelude::*; -/// -/// /// Awaitable non-send sleep function -/// #[pyfunction] -/// fn sleep_for(py: Python, secs: u64) -> PyResult<&PyAny> { -/// // Rc is non-send so it cannot be passed into pyo3_asyncio::tokio::future_into_py -/// let secs = Rc::new(secs); -/// -/// pyo3_asyncio::tokio::local_future_into_py_with_locals( -/// py, -/// pyo3_asyncio::tokio::get_current_locals(py)?, -/// async move { -/// tokio::time::sleep(Duration::from_secs(*secs)).await; -/// Python::with_gil(|py| Ok(py.None())) -/// } -/// ) -/// } -/// -/// # #[cfg(all(feature = "tokio-runtime", feature = "attributes"))] -/// #[pyo3_asyncio::tokio::main] -/// async fn main() -> PyResult<()> { -/// let locals = Python::with_gil(|py| -> PyResult<_> { -/// pyo3_asyncio::tokio::get_current_locals(py) -/// })?; -/// -/// // the main coroutine is running in a Send context, so we cannot use LocalSet here. Instead -/// // we use spawn_blocking in order to use LocalSet::block_on -/// tokio::task::spawn_blocking(move || { -/// // LocalSet allows us to work with !Send futures within tokio. Without it, any calls to -/// // pyo3_asyncio::tokio::local_future_into_py will panic. -/// tokio::task::LocalSet::new().block_on( -/// pyo3_asyncio::tokio::get_runtime(), -/// pyo3_asyncio::tokio::scope_local(locals, async { -/// Python::with_gil(|py| { -/// let py_future = sleep_for(py, 1)?; -/// pyo3_asyncio::tokio::into_future(py_future) -/// })? -/// .await?; -/// -/// Ok(()) -/// }) -/// ) -/// }).await.unwrap() -/// } -/// # #[cfg(not(all(feature = "tokio-runtime", feature = "attributes")))] -/// # fn main() {} -/// ``` -pub fn local_future_into_py_with_locals<F, T>( - py: Python, - locals: TaskLocals, - fut: F, -) -> PyResult<&PyAny> -where - F: Future<Output = PyResult<T>> + 'static, - T: IntoPy<PyObject>, -{ - generic::local_future_into_py_with_locals::<TokioRuntime, _, T>(py, locals, fut) -} - -/// Convert a `!Send` Rust Future into a Python awaitable -/// -/// If the `asyncio.Future` returned by this conversion is cancelled via `asyncio.Future.cancel`, -/// the Rust future will be cancelled as well (new behaviour in `v0.15`). -/// -/// Python `contextvars` are preserved when calling async Python functions within the Rust future -/// via [`into_future`] (new behaviour in `v0.15`). -/// -/// > Although `contextvars` are preserved for async Python functions, synchronous functions will -/// unfortunately fail to resolve them when called within the Rust future. This is because the -/// function is being called from a Rust thread, not inside an actual Python coroutine context. -/// > -/// > As a workaround, you can get the `contextvars` from the current task locals using -/// [`get_current_locals`] and [`TaskLocals::context`](`crate::TaskLocals::context`), then wrap your -/// synchronous function in a call to `contextvars.Context.run`. This will set the context, call the -/// synchronous function, and restore the previous context when it returns or raises an exception. -/// -/// # Arguments -/// * `py` - The current PyO3 GIL guard -/// * `fut` - The Rust future to be converted -/// -/// # Examples -/// -/// ``` -/// use std::{rc::Rc, time::Duration}; -/// -/// use pyo3::prelude::*; -/// -/// /// Awaitable non-send sleep function -/// #[pyfunction] -/// fn sleep_for(py: Python, secs: u64) -> PyResult<&PyAny> { -/// // Rc is non-send so it cannot be passed into pyo3_asyncio::tokio::future_into_py -/// let secs = Rc::new(secs); -/// pyo3_asyncio::tokio::local_future_into_py(py, async move { -/// tokio::time::sleep(Duration::from_secs(*secs)).await; -/// Ok(()) -/// }) -/// } -/// -/// # #[cfg(all(feature = "tokio-runtime", feature = "attributes"))] -/// #[pyo3_asyncio::tokio::main] -/// async fn main() -> PyResult<()> { -/// let locals = Python::with_gil(|py| { -/// pyo3_asyncio::tokio::get_current_locals(py).unwrap() -/// }); -/// -/// // the main coroutine is running in a Send context, so we cannot use LocalSet here. Instead -/// // we use spawn_blocking in order to use LocalSet::block_on -/// tokio::task::spawn_blocking(move || { -/// // LocalSet allows us to work with !Send futures within tokio. Without it, any calls to -/// // pyo3_asyncio::tokio::local_future_into_py will panic. -/// tokio::task::LocalSet::new().block_on( -/// pyo3_asyncio::tokio::get_runtime(), -/// pyo3_asyncio::tokio::scope_local(locals, async { -/// Python::with_gil(|py| { -/// let py_future = sleep_for(py, 1)?; -/// pyo3_asyncio::tokio::into_future(py_future) -/// })? -/// .await?; -/// -/// Ok(()) -/// }) -/// ) -/// }).await.unwrap() -/// } -/// # #[cfg(not(all(feature = "tokio-runtime", feature = "attributes")))] -/// # fn main() {} -/// ``` -pub fn local_future_into_py<F, T>(py: Python, fut: F) -> PyResult<&PyAny> -where - F: Future<Output = PyResult<T>> + 'static, - T: IntoPy<PyObject>, -{ - generic::local_future_into_py::<TokioRuntime, _, T>(py, fut) -} - -/// Convert a Python `awaitable` into a Rust Future -/// -/// This function converts the `awaitable` into a Python Task using `run_coroutine_threadsafe`. A -/// completion handler sends the result of this Task through a -/// `futures::channel::oneshot::Sender<PyResult<PyObject>>` and the future returned by this function -/// simply awaits the result through the `futures::channel::oneshot::Receiver<PyResult<PyObject>>`. -/// -/// # Arguments -/// * `awaitable` - The Python `awaitable` to be converted -/// -/// # Examples -/// -/// ``` -/// use std::time::Duration; -/// -/// use pyo3::prelude::*; -/// -/// const PYTHON_CODE: &'static str = r#" -/// import asyncio -/// -/// async def py_sleep(duration): -/// await asyncio.sleep(duration) -/// "#; -/// -/// async fn py_sleep(seconds: f32) -> PyResult<()> { -/// let test_mod = Python::with_gil(|py| -> PyResult<PyObject> { -/// Ok( -/// PyModule::from_code( -/// py, -/// PYTHON_CODE, -/// "test_into_future/test_mod.py", -/// "test_mod" -/// )? -/// .into() -/// ) -/// })?; -/// -/// Python::with_gil(|py| { -/// pyo3_asyncio::tokio::into_future( -/// test_mod -/// .call_method1(py, "py_sleep", (seconds.into_py(py),))? -/// .as_ref(py), -/// ) -/// })? -/// .await?; -/// Ok(()) -/// } -/// ``` -pub fn into_future(awaitable: &PyAny) -> PyResult<impl Future<Output = PyResult<PyObject>> + Send> { - generic::into_future::<TokioRuntime>(awaitable) -} - -/// <code>unstable-streams</code> Convert an async generator into a stream -/// -/// **This API is marked as unstable** and is only available when the -/// `unstable-streams` crate feature is enabled. This comes with no -/// stability guarantees, and could be changed or removed at any time. -/// -/// # Arguments -/// * `locals` - The current task locals -/// * `gen` - The Python async generator to be converted -/// -/// # Examples -/// ``` -/// use pyo3::prelude::*; -/// use futures::{StreamExt, TryStreamExt}; -/// -/// const TEST_MOD: &str = r#" -/// import asyncio -/// -/// async def gen(): -/// for i in range(10): -/// await asyncio.sleep(0.1) -/// yield i -/// "#; -/// -/// # #[cfg(all(feature = "unstable-streams", feature = "attributes"))] -/// # #[pyo3_asyncio::tokio::main] -/// # async fn main() -> PyResult<()> { -/// let stream = Python::with_gil(|py| { -/// let test_mod = PyModule::from_code( -/// py, -/// TEST_MOD, -/// "test_rust_coroutine/test_mod.py", -/// "test_mod", -/// )?; -/// -/// pyo3_asyncio::tokio::into_stream_with_locals_v1( -/// pyo3_asyncio::tokio::get_current_locals(py)?, -/// test_mod.call_method0("gen")? -/// ) -/// })?; -/// -/// let vals = stream -/// .map(|item| Python::with_gil(|py| -> PyResult<i32> { Ok(item?.as_ref(py).extract()?) })) -/// .try_collect::<Vec<i32>>() -/// .await?; -/// -/// assert_eq!((0..10).collect::<Vec<i32>>(), vals); -/// -/// Ok(()) -/// # } -/// # #[cfg(not(all(feature = "unstable-streams", feature = "attributes")))] -/// # fn main() {} -/// ``` -#[cfg(feature = "unstable-streams")] -pub fn into_stream_with_locals_v1<'p>( - locals: TaskLocals, - gen: &'p PyAny, -) -> PyResult<impl futures::Stream<Item = PyResult<PyObject>> + 'static> { - generic::into_stream_with_locals_v1::<TokioRuntime>(locals, gen) -} - -/// <code>unstable-streams</code> Convert an async generator into a stream -/// -/// **This API is marked as unstable** and is only available when the -/// `unstable-streams` crate feature is enabled. This comes with no -/// stability guarantees, and could be changed or removed at any time. -/// -/// # Arguments -/// * `gen` - The Python async generator to be converted -/// -/// # Examples -/// ``` -/// use pyo3::prelude::*; -/// use futures::{StreamExt, TryStreamExt}; -/// -/// const TEST_MOD: &str = r#" -/// import asyncio -/// -/// async def gen(): -/// for i in range(10): -/// await asyncio.sleep(0.1) -/// yield i -/// "#; -/// -/// # #[cfg(all(feature = "unstable-streams", feature = "attributes"))] -/// # #[pyo3_asyncio::tokio::main] -/// # async fn main() -> PyResult<()> { -/// let stream = Python::with_gil(|py| { -/// let test_mod = PyModule::from_code( -/// py, -/// TEST_MOD, -/// "test_rust_coroutine/test_mod.py", -/// "test_mod", -/// )?; -/// -/// pyo3_asyncio::tokio::into_stream_v1(test_mod.call_method0("gen")?) -/// })?; -/// -/// let vals = stream -/// .map(|item| Python::with_gil(|py| -> PyResult<i32> { Ok(item?.as_ref(py).extract()?) })) -/// .try_collect::<Vec<i32>>() -/// .await?; -/// -/// assert_eq!((0..10).collect::<Vec<i32>>(), vals); -/// -/// Ok(()) -/// # } -/// # #[cfg(not(all(feature = "unstable-streams", feature = "attributes")))] -/// # fn main() {} -/// ``` -#[cfg(feature = "unstable-streams")] -pub fn into_stream_v1<'p>( - gen: &'p PyAny, -) -> PyResult<impl futures::Stream<Item = PyResult<PyObject>> + 'static> { - generic::into_stream_v1::<TokioRuntime>(gen) -} - -/// <code>unstable-streams</code> Convert an async generator into a stream -/// -/// **This API is marked as unstable** and is only available when the -/// `unstable-streams` crate feature is enabled. This comes with no -/// stability guarantees, and could be changed or removed at any time. -/// -/// # Arguments -/// * `locals` - The current task locals -/// * `gen` - The Python async generator to be converted -/// -/// # Examples -/// ``` -/// use pyo3::prelude::*; -/// use futures::{StreamExt, TryStreamExt}; -/// -/// const TEST_MOD: &str = r#" -/// import asyncio -/// -/// async def gen(): -/// for i in range(10): -/// await asyncio.sleep(0.1) -/// yield i -/// "#; -/// -/// # #[cfg(all(feature = "unstable-streams", feature = "attributes"))] -/// # #[pyo3_asyncio::tokio::main] -/// # async fn main() -> PyResult<()> { -/// let stream = Python::with_gil(|py| { -/// let test_mod = PyModule::from_code( -/// py, -/// TEST_MOD, -/// "test_rust_coroutine/test_mod.py", -/// "test_mod", -/// )?; -/// -/// pyo3_asyncio::tokio::into_stream_with_locals_v2( -/// pyo3_asyncio::tokio::get_current_locals(py)?, -/// test_mod.call_method0("gen")? -/// ) -/// })?; -/// -/// let vals = stream -/// .map(|item| Python::with_gil(|py| -> PyResult<i32> { Ok(item.as_ref(py).extract()?) })) -/// .try_collect::<Vec<i32>>() -/// .await?; -/// -/// assert_eq!((0..10).collect::<Vec<i32>>(), vals); -/// -/// Ok(()) -/// # } -/// # #[cfg(not(all(feature = "unstable-streams", feature = "attributes")))] -/// # fn main() {} -/// ``` -#[cfg(feature = "unstable-streams")] -pub fn into_stream_with_locals_v2<'p>( - locals: TaskLocals, - gen: &'p PyAny, -) -> PyResult<impl futures::Stream<Item = PyObject> + 'static> { - generic::into_stream_with_locals_v2::<TokioRuntime>(locals, gen) -} - -/// <code>unstable-streams</code> Convert an async generator into a stream -/// -/// **This API is marked as unstable** and is only available when the -/// `unstable-streams` crate feature is enabled. This comes with no -/// stability guarantees, and could be changed or removed at any time. -/// -/// # Arguments -/// * `gen` - The Python async generator to be converted -/// -/// # Examples -/// ``` -/// use pyo3::prelude::*; -/// use futures::{StreamExt, TryStreamExt}; -/// -/// const TEST_MOD: &str = r#" -/// import asyncio -/// -/// async def gen(): -/// for i in range(10): -/// await asyncio.sleep(0.1) -/// yield i -/// "#; -/// -/// # #[cfg(all(feature = "unstable-streams", feature = "attributes"))] -/// # #[pyo3_asyncio::tokio::main] -/// # async fn main() -> PyResult<()> { -/// let stream = Python::with_gil(|py| { -/// let test_mod = PyModule::from_code( -/// py, -/// TEST_MOD, -/// "test_rust_coroutine/test_mod.py", -/// "test_mod", -/// )?; -/// -/// pyo3_asyncio::tokio::into_stream_v2(test_mod.call_method0("gen")?) -/// })?; -/// -/// let vals = stream -/// .map(|item| Python::with_gil(|py| -> PyResult<i32> { Ok(item.as_ref(py).extract()?) })) -/// .try_collect::<Vec<i32>>() -/// .await?; -/// -/// assert_eq!((0..10).collect::<Vec<i32>>(), vals); -/// -/// Ok(()) -/// # } -/// # #[cfg(not(all(feature = "unstable-streams", feature = "attributes")))] -/// # fn main() {} -/// ``` -#[cfg(feature = "unstable-streams")] -pub fn into_stream_v2<'p>( - gen: &'p PyAny, -) -> PyResult<impl futures::Stream<Item = PyObject> + 'static> { - generic::into_stream_v2::<TokioRuntime>(gen) -} diff --git a/src/asgi/io.rs b/src/asgi/io.rs index 6db89ea5..8396cfda 100644 --- a/src/asgi/io.rs +++ b/src/asgi/io.rs @@ -14,7 +14,7 @@ use tungstenite::Message; use crate::{ http::HV_SERVER, - runtime::{RuntimeRef, future_into_py}, + runtime::{RuntimeRef, future_into_py_iter, future_into_py_futlike}, ws::{HyperWebsocket, UpgradeData} }; use super::{ @@ -81,7 +81,7 @@ impl ASGIHTTPProtocol { impl ASGIHTTPProtocol { fn receive<'p>(&mut self, py: Python<'p>) -> PyResult<&'p PyAny> { let transport = self.request.clone(); - future_into_py(self.rt.clone(), py, async move { + future_into_py_iter(self.rt.clone(), py, async move { let mut req = transport.lock().await; let body = hyper::body::to_bytes(&mut *req).await.unwrap(); Python::with_gil(|py| { @@ -196,7 +196,7 @@ impl ASGIWebsocketProtocol { let accepted = self.accepted.clone(); let tx = self.ws_tx.clone(); let rx = self.ws_rx.clone(); - future_into_py(self.rt.clone(), py, async move { + future_into_py_iter(self.rt.clone(), py, async move { if let Ok(_) = upgrade.send().await { if let Ok(stream) = websocket.await { let mut wtx = tx.lock().await; @@ -222,7 +222,7 @@ impl ASGIWebsocketProtocol { let transport = self.ws_tx.clone(); let closed = self.closed.clone(); let message = ws_message_into_rs(data); - future_into_py(self.rt.clone(), py, async move { + future_into_py_iter(self.rt.clone(), py, async move { if !closed { if let Ok(message) = message { if let Some(ws) = &mut *(transport.lock().await) { @@ -250,7 +250,7 @@ impl ASGIWebsocketProtocol { macro_rules! empty_future { ($rt:expr, $py:expr) => { - future_into_py($rt, $py, async move { + future_into_py_iter($rt, $py, async move { Ok(()) }) }; @@ -262,7 +262,7 @@ impl ASGIWebsocketProtocol { let transport = self.ws_rx.clone(); let accepted = self.accepted.clone(); let closed = self.closed.clone(); - future_into_py(self.rt.clone(), py, async move { + future_into_py_futlike(self.rt.clone(), py, async move { let accepted = accepted.lock().await; match (*accepted, closed) { (false, false) => { diff --git a/src/callbacks.rs b/src/callbacks.rs index fb829f08..b1206eee 100644 --- a/src/callbacks.rs +++ b/src/callbacks.rs @@ -1,4 +1,6 @@ use pyo3::prelude::*; +use pyo3::pyclass::IterNextOutput; + #[derive(Clone)] pub(crate) struct CallbackWrapper { @@ -7,10 +9,172 @@ pub(crate) struct CallbackWrapper { } impl CallbackWrapper { - pub(crate) fn new(callback: PyObject, event_loop: &PyAny, context: &PyAny) -> Self { + pub(crate) fn new( + callback: PyObject, + event_loop: &PyAny, + context: &PyAny + ) -> Self { Self { - callback: callback, + callback, context: pyo3_asyncio::TaskLocals::new(event_loop).with_context(context) } } } + +#[pyclass] +pub(crate) struct PyAwaitableResultYielder { + result: Option<PyResult<PyObject>>, + none: PyObject +} + +impl PyAwaitableResultYielder { + pub(crate) fn set_result(&mut self, result: PyResult<PyObject>) { + self.result = Some(result) + } +} + +#[pymethods] +impl PyAwaitableResultYielder { + fn __iter__(pyself: PyRef<'_, Self>) -> PyRef<'_, Self> { + pyself + } + + fn __next__(mut pyself: PyRefMut<'_, Self>) -> PyResult<IterNextOutput<PyObject, PyObject>> { + match pyself.result.take() { + Some(res) => { + match res { + Ok(v) => Ok(IterNextOutput::Return(v)), + Err(err) => Err(err) + } + }, + _ => Ok(IterNextOutput::Yield(pyself.none.clone())) + } + } +} + +#[pyclass] +pub(crate) struct PyAwaitableResultFutureLike { + py_block: bool, + event_loop: PyObject, + result: Option<PyResult<PyObject>>, + cb: Option<(PyObject, Py<pyo3::types::PyDict>)> +} + +impl PyAwaitableResultFutureLike { + pub(crate) fn new(event_loop: PyObject) -> Self { + Self { + event_loop, + py_block: true, + result: None, + cb: None + } + } + + pub(crate) fn set_result(mut pyself: PyRefMut<'_, Self>, result: PyResult<PyObject>) { + pyself.result = Some(result); + if let Some((cb, ctx)) = pyself.cb.take() { + Python::with_gil(|py| { + let _ = pyself.event_loop.call_method( + py, + "call_soon_threadsafe", + (cb, &pyself), + Some(ctx.as_ref(py)) + ); + }) + } + } +} + +#[pymethods] +impl PyAwaitableResultFutureLike { + #[getter(_asyncio_future_blocking)] + fn get_block(&self) -> bool { + self.py_block + } + + #[setter(_asyncio_future_blocking)] + fn set_block(&mut self, val: bool) { + self.py_block = val + } + + fn get_loop(&mut self) -> PyObject { + self.event_loop.clone() + } + + fn add_done_callback( + mut pyself: PyRefMut<'_, Self>, + py: Python, + cb: PyObject, + context: PyObject + ) -> PyResult<()> { + let kwctx = pyo3::types::PyDict::new(py); + kwctx.set_item("context", context)?; + match pyself.result { + Some(_) => { + pyself.event_loop.call_method(py, "call_soon", (cb, &pyself), Some(kwctx))?; + }, + _ => { + pyself.cb = Some((cb, kwctx.into_py(py))); + } + } + Ok(()) + } + + fn result(&self) {} + + fn __iter__(pyself: PyRef<'_, Self>) -> PyRef<'_, Self> { + pyself + } + + fn __next__( + mut pyself: PyRefMut<'_, Self> + ) -> PyResult<IterNextOutput<PyRefMut<'_, Self>, PyObject>> { + match pyself.result { + Some(_) => { + match pyself.result.take().unwrap() { + Ok(v) => Ok(IterNextOutput::Return(v)), + Err(err) => Err(err) + } + }, + _ => Ok(IterNextOutput::Yield(pyself)) + } + } +} + +#[pyclass] +pub(crate) struct PyIterAwaitableResult { + pub inner: Py<PyAwaitableResultYielder> +} + +impl PyIterAwaitableResult { + pub(crate) fn new(py: Python) -> PyResult<Self> { + let inner = Py::new(py, PyAwaitableResultYielder { result: None, none: py.None() })?; + Ok(Self { inner }) + } +} + +#[pymethods] +impl PyIterAwaitableResult { + fn __await__(&mut self, py: Python) -> PyObject { + self.inner.to_object(py) + } +} + +#[pyclass] +pub(crate) struct PyFutureAwaitableResult { + pub inner: Py<PyAwaitableResultFutureLike> +} + +impl PyFutureAwaitableResult { + pub(crate) fn new(py: Python, event_loop: &PyAny) -> PyResult<Self> { + let inner = Py::new(py, PyAwaitableResultFutureLike::new(event_loop.to_object(py)))?; + Ok(Self { inner }) + } +} + +#[pymethods] +impl PyFutureAwaitableResult { + fn __await__(&mut self, py: Python) -> PyObject { + self.inner.to_object(py) + } +} diff --git a/src/rsgi/io.rs b/src/rsgi/io.rs index 8fcb80cb..bcaf7e21 100644 --- a/src/rsgi/io.rs +++ b/src/rsgi/io.rs @@ -8,7 +8,7 @@ use tokio::sync::{oneshot, Mutex}; use tungstenite::Message; use crate::{ - runtime::{RuntimeRef, future_into_py}, + runtime::{RuntimeRef, future_into_py_iter, future_into_py_futlike}, ws::{HyperWebsocket, UpgradeData} }; use super::{ @@ -31,7 +31,7 @@ impl RSGIHTTPProtocol { request: Request<Body> ) -> Self { Self { - rt: rt, + rt, tx: Some(tx), request: Arc::new(Mutex::new(request)) } @@ -46,7 +46,7 @@ impl RSGIHTTPProtocol { impl RSGIHTTPProtocol { fn __call__<'p>(&self, py: Python<'p>) -> PyResult<&'p PyAny> { let req_ref = self.request.clone(); - future_into_py(self.rt.clone(), py, async move { + future_into_py_iter(self.rt.clone(), py, async move { let mut req = req_ref.lock().await; let body = hyper::body::to_bytes(&mut *req).await.unwrap(); Ok(Python::with_gil(|py| { @@ -113,7 +113,7 @@ impl RSGIWebsocketTransport { impl RSGIWebsocketTransport { fn receive<'p>(&self, py: Python<'p>) -> PyResult<&'p PyAny> { let transport = self.rx.clone(); - future_into_py(self.rt.clone(), py, async move { + future_into_py_futlike(self.rt.clone(), py, async move { if let Ok(mut stream) = transport.try_lock() { loop { match stream.next().await { @@ -143,7 +143,7 @@ impl RSGIWebsocketTransport { fn send_bytes<'p>(&self, py: Python<'p>, data: Vec<u8>) -> PyResult<&'p PyAny> { let transport = self.tx.clone(); - future_into_py(self.rt.clone(), py, async move { + future_into_py_iter(self.rt.clone(), py, async move { if let Ok(mut stream) = transport.try_lock() { return match stream.send(Message::Binary(data)).await { Ok(_) => Ok(()), @@ -156,7 +156,7 @@ impl RSGIWebsocketTransport { fn send_str<'p>(&self, py: Python<'p>, data: String) -> PyResult<&'p PyAny> { let transport = self.tx.clone(); - future_into_py(self.rt.clone(), py, async move { + future_into_py_iter(self.rt.clone(), py, async move { if let Ok(mut stream) = transport.try_lock() { return match stream.send(Message::Text(data)).await { Ok(_) => Ok(()), @@ -266,7 +266,7 @@ impl RSGIWebsocketProtocol { let rth = self.rt.clone(); let mut upgrade = self.upgrade.take().unwrap(); let transport = self.websocket.clone(); - future_into_py(self.rt.clone(), py, async move { + future_into_py_iter(self.rt.clone(), py, async move { let mut ws = transport.lock().await; match upgrade.send().await { Ok(_) => { diff --git a/src/runtime.rs b/src/runtime.rs index 75acd5cc..202919a7 100644 --- a/src/runtime.rs +++ b/src/runtime.rs @@ -1,9 +1,16 @@ use once_cell::{unsync::OnceCell as UnsyncOnceCell}; -use pyo3_asyncio::{TaskLocals, generic as pyrt_generic}; +use pyo3_asyncio::TaskLocals; use pyo3::prelude::*; use std::{future::Future, io, pin::Pin, sync::{Arc, Mutex}}; use tokio::{runtime::Builder, task::{JoinHandle, LocalSet}}; +use super::callbacks::{ + PyIterAwaitableResult, + PyFutureAwaitableResult, + PyAwaitableResultFutureLike +}; + + tokio::task_local! { static TASK_LOCALS: UnsyncOnceCell<TaskLocals>; } @@ -182,160 +189,6 @@ pub(crate) fn into_future( ) } -pub(crate) fn future_into_py_with_locals<R, F, T>( - rt: R, - py: Python, - locals: TaskLocals, - fut: F, -) -> PyResult<&PyAny> -where - R: Runtime + ContextExt + Clone, - F: Future<Output = PyResult<T>> + Send + 'static, - T: IntoPy<PyObject>, -{ - let (cancel_tx, cancel_rx) = futures::channel::oneshot::channel(); - - let py_fut = pyo3_asyncio::create_future(locals.event_loop(py))?; - py_fut.call_method1( - "add_done_callback", - (pyrt_generic::PyDoneCallback { - cancel_tx: Some(cancel_tx), - },), - )?; - - let future_tx1 = PyObject::from(py_fut); - let future_tx2 = future_tx1.clone(); - let rth = rt.handler(); - - rt.spawn(async move { - let rti = rth.handler(); - let locals2 = locals.clone(); - - if let Err(e) = rth.spawn(async move { - let result = rti.scope( - locals2.clone(), - pyrt_generic::Cancellable::new_with_cancel_rx(fut, cancel_rx), - ) - .await; - - Python::with_gil(move |py| { - if pyrt_generic::cancelled(future_tx1.as_ref(py)) - .map_err(pyo3_asyncio::dump_err(py)) - .unwrap_or(false) - { - return; - } - - let _ = pyrt_generic::set_result( - locals2.event_loop(py), - future_tx1.as_ref(py), - result.map(|val| val.into_py(py)), - ) - .map_err(pyo3_asyncio::dump_err(py)); - }); - }) - .await - { - if e.is_panic() { - Python::with_gil(move |py| { - if pyrt_generic::cancelled(future_tx2.as_ref(py)) - .map_err(pyo3_asyncio::dump_err(py)) - .unwrap_or(false) - { - return; - } - - let _ = pyrt_generic::set_result( - locals.event_loop(py), - future_tx2.as_ref(py), - Err(pyo3_asyncio::err::RustPanic::new_err("rust future panicked")), - ) - .map_err(pyo3_asyncio::dump_err(py)); - }); - } - } - }); - - Ok(py_fut) -} - -pub fn local_future_into_py_with_locals<R, F, T>( - rt: R, - py: Python, - locals: TaskLocals, - fut: F, -) -> PyResult<&PyAny> -where - R: Runtime + SpawnLocalExt + LocalContextExt, - F: Future<Output = PyResult<T>> + 'static, - T: IntoPy<PyObject>, -{ - let (cancel_tx, cancel_rx) = futures::channel::oneshot::channel(); - - let py_fut = pyo3_asyncio::create_future(locals.event_loop(py))?; - py_fut.call_method1( - "add_done_callback", - (pyrt_generic::PyDoneCallback { - cancel_tx: Some(cancel_tx), - },), - )?; - - let future_tx1 = PyObject::from(py_fut); - let future_tx2 = future_tx1.clone(); - let rth = rt.handler(); - - rt.spawn_local(async move { - let rti = rth.handler(); - let locals2 = locals.clone(); - - if let Err(e) = rth.spawn_local(async move { - let result = rti.scope_local( - locals2.clone(), - pyrt_generic::Cancellable::new_with_cancel_rx(fut, cancel_rx), - ) - .await; - - Python::with_gil(move |py| { - if pyrt_generic::cancelled(future_tx1.as_ref(py)) - .map_err(pyo3_asyncio::dump_err(py)) - .unwrap_or(false) - { - return; - } - - let _ = pyrt_generic::set_result( - locals2.event_loop(py), - future_tx1.as_ref(py), - result.map(|val| val.into_py(py)), - ) - .map_err(pyo3_asyncio::dump_err(py)); - }); - }) - .await - { - if e.is_panic() { - Python::with_gil(move |py| { - if pyrt_generic::cancelled(future_tx2.as_ref(py)) - .map_err(pyo3_asyncio::dump_err(py)) - .unwrap_or(false) - { - return; - } - - let _ = pyrt_generic::set_result( - locals.event_loop(py), - future_tx2.as_ref(py), - Err(pyo3_asyncio::err::RustPanic::new_err("Rust future panicked")), - ) - .map_err(pyo3_asyncio::dump_err(py)); - }); - } - } - }); - - Ok(py_fut) -} - #[inline] fn get_current_locals<R>(py: Python) -> PyResult<TaskLocals> where @@ -348,23 +201,58 @@ where } } -pub(crate) fn future_into_py<R, F, T>(rt: R, py: Python, fut: F) -> PyResult<&PyAny> +// NOTE: +// `future_into_py_iter` relies on what CPython refers as "bare yield". +// This is generally ~55% faster than `pyo3_asyncio.future_into_py` implementation. +// It consumes more cpu-cycles than `future_into_py_futlike`, +// but for "quick" operations it's something like 12% faster. +pub(crate) fn future_into_py_iter<R, F, T>(rt: R, py: Python, fut: F) -> PyResult<&PyAny> where R: Runtime + ContextExt + Clone, F: Future<Output = PyResult<T>> + Send + 'static, T: IntoPy<PyObject>, { - future_into_py_with_locals::<R, F, T>(rt, py, get_current_locals::<R>(py)?, fut) + let aw = PyIterAwaitableResult::new(py)?; + let fut_res = aw.inner.clone(); + let py_fut = aw.into_py(py).into_ref(py); + + rt.spawn(async move { + let result = fut.await; + Python::with_gil(move |py| { + fut_res.as_ref(py).borrow_mut().set_result(result.map(|v| v.into_py(py))); + }); + }); + + Ok(py_fut) } -#[allow(dead_code)] -pub fn local_future_into_py<R, F, T>(rt: R, py: Python, fut: F) -> PyResult<&PyAny> +// NOTE: +// `future_into_py_futlike` relies on an `asyncio.Future` like implementation. +// This is generally ~38% faster than `pyo3_asyncio.future_into_py` implementation. +// It won't consume more cpu-cycles than standard asyncio implementation, +// and for "long" operations it's something like 6% faster than `future_into_py_iter`. +pub(crate) fn future_into_py_futlike<R, F, T>(rt: R, py: Python, fut: F) -> PyResult<&PyAny> where - R: Runtime + ContextExt + SpawnLocalExt + LocalContextExt, - F: Future<Output = PyResult<T>> + 'static, + R: Runtime + ContextExt + Clone, + F: Future<Output = PyResult<T>> + Send + 'static, T: IntoPy<PyObject>, { - local_future_into_py_with_locals::<R, F, T>(rt, py, get_current_locals::<R>(py)?, fut) + let task_locals = get_current_locals::<R>(py)?; + let aw = PyFutureAwaitableResult::new(py, task_locals.event_loop(py))?; + let fut_res = aw.inner.clone(); + let py_fut = aw.into_py(py).into_ref(py); + + rt.spawn(async move { + let result = fut.await; + Python::with_gil(move |py| { + PyAwaitableResultFutureLike::set_result( + fut_res.as_ref(py).borrow_mut(), + result.map(|v| v.into_py(py)) + ); + }); + }); + + Ok(py_fut) } pub(crate) fn run_until_complete<R, F, T>(rt: R, event_loop: &PyAny, fut: F) -> PyResult<T> @@ -376,20 +264,31 @@ where let py = event_loop.py(); let result_tx = Arc::new(Mutex::new(None)); let result_rx = Arc::clone(&result_tx); - let coro = future_into_py_with_locals::<R, _, ()>( - rt, - py, - TaskLocals::new(event_loop).copy_context(py)?, - async move { - let val = fut.await?; - if let Ok(mut result) = result_tx.lock() { - *result = Some(val); - } - Ok(()) - }, - )?; - - event_loop.call_method1("run_until_complete", (coro,))?; + + let task_locals = TaskLocals::new(event_loop).copy_context(py)?; + let py_fut = event_loop.call_method0("create_future")?; + let loop_tx = event_loop.into_py(py); + let future_tx = py_fut.into_py(py); + + let rth = rt.handler(); + + rt.spawn(async move { + let val = rth.scope( + task_locals.clone(), + fut + ) + .await; + if let Ok(mut result) = result_tx.lock() { + *result = Some(val.unwrap()); + } + + Python::with_gil(move |py| { + let res_method = future_tx.getattr(py, "set_result").unwrap(); + let _ = loop_tx.call_method(py, "call_soon_threadsafe", (res_method, py.None()), None); + }); + }); + + event_loop.call_method1("run_until_complete", (py_fut,))?; let result = result_rx.lock().unwrap().take().unwrap(); Ok(result)

Avoid `pyo3-asyncio` fork Involved steps: - consolidate changes to make a PR into the upstream project - requires #32 - open up a discussion in the upstream project

2023-02-06T17:37:59

0.0

[]

rickh94/ODetaM

rickh94__ODetaM-13

bf29d7583a8d1015ca537c5f0c9f76d5a53cff6a

diff --git a/odetam/async_model.py b/odetam/async_model.py index 45d0816..8e2e3ce 100644 --- a/odetam/async_model.py +++ b/odetam/async_model.py @@ -1,8 +1,7 @@ -from typing import Any, Dict, List, Optional, Union +from typing import Any, Dict, List, Optional, Type, TypeVar, Union from deta import AsyncBase, Deta from deta.base import FetchResponse -from typing_extensions import Self from odetam.exceptions import DetaError, InvalidKey, ItemNotFound from odetam.model import BaseDetaModel, DetaModelMetaClass, handle_db_property @@ -19,9 +18,12 @@ def __db__(cls): return handle_db_property(cls, AsyncBase) +T = TypeVar("T", bound="AsyncDetaModel") + + class AsyncDetaModel(BaseDetaModel, metaclass=AsyncDetaModelMetaClass): @classmethod - async def get(cls, key: str) -> Self: + async def get(cls: Type[T], key: str) -> T: """ Get a single instance :param key: Deta database key @@ -36,7 +38,7 @@ async def get(cls, key: str) -> Self: return cls._return_item_or_raise(item) @classmethod - async def get_or_none(cls, key: str) -> Optional[Self]: + async def get_or_none(cls: Type[T], key: str) -> Optional[T]: """Try to get item by key or return None if item not found""" try: return await cls.get(key) @@ -44,7 +46,7 @@ async def get_or_none(cls, key: str) -> Optional[Self]: return None @classmethod - async def get_all(cls) -> List[Self]: + async def get_all(cls: Type[T]) -> List[T]: """Get all the records from the database""" response: FetchResponse = await cls.__db__.fetch() records: List[Dict[str, Any]] = response.items @@ -56,8 +58,8 @@ async def get_all(cls) -> List[Self]: @classmethod async def query( - cls, query_statement: Union[DetaQuery, DetaQueryStatement, DetaQueryList] - ) -> List[Self]: + cls: Type[T], query_statement: Union[DetaQuery, DetaQueryStatement, DetaQueryList] + ) -> List[T]: """Get items from database based on the query.""" response: FetchResponse = await cls.__db__.fetch(query_statement.as_query()) records: List[Dict[str, Any]] = response.items @@ -70,12 +72,12 @@ async def query( return [cls._deserialize(item) for item in records] @classmethod - async def delete_key(cls, key: str): + async def delete_key(cls, key: str) -> None: """Delete an item based on the key""" await cls.__db__.delete(key) @classmethod - async def put_many(cls, items: List[Self]): + async def put_many(cls: Type[T], items: List[T]) -> List[T]: """Put multiple instances at once :param items: List of pydantic objects to put in the database @@ -103,7 +105,7 @@ async def put_many(cls, items: List[Self]): async def _db_put(cls, data: List[Dict[str, Any]]) -> List[Dict[str, Any]]: return await cls.__db__.put(data) - async def save(self): + async def save(self) -> None: """Saves the record to the database. Behaves as upsert, will create if not present. Database key will then be set on the object.""" # exclude = set() @@ -114,7 +116,7 @@ async def save(self): saved = await self._db_put(self._serialize()) self.key = saved["key"] - async def delete(self): + async def delete(self) -> None: """Delete the open object from the database. The object will still exist in python, but will be deleted from the database and the key attribute will be set to None.""" diff --git a/odetam/model.py b/odetam/model.py index b3a548d..24a99bd 100644 --- a/odetam/model.py +++ b/odetam/model.py @@ -1,13 +1,12 @@ import datetime import re -from typing import Any, Callable, Container, Dict, List, Optional, Union +from typing import Any, Callable, Container, Dict, List, Optional, Type, TypeVar, Union import pydantic import ujson from deta import Base, Deta from deta.base import FetchResponse, _Base from pydantic import BaseModel, Field, ValidationError -from typing_extensions import Self from odetam.exceptions import DetaError, InvalidKey, ItemNotFound from odetam.field import DetaField @@ -53,6 +52,9 @@ def __db__(cls): return handle_db_property(cls, Base) +K = TypeVar("K", bound="BaseDetaModel") + + class BaseDetaModel(BaseModel): __db__ = Optional[_Base] @@ -77,7 +79,8 @@ def _serialize(self, exclude: Optional[Container[str]] = None) -> Dict[str, Any] elif field.type_ == datetime.datetime: as_dict[field_name] = getattr(self, field_name).timestamp() elif field.type_ == datetime.date: - as_dict[field_name] = int(getattr(self, field_name).strftime("%Y%m%d")) + as_dict[field_name] = int( + getattr(self, field_name).strftime("%Y%m%d")) elif field.type_ == datetime.time: as_dict[field_name] = int( getattr(self, field_name).strftime("%H%M%S%f") @@ -90,7 +93,7 @@ def _serialize(self, exclude: Optional[Container[str]] = None) -> Dict[str, Any] return as_dict @classmethod - def _deserialize(cls, data: Dict[str, Any]) -> Self: + def _deserialize(cls: Type[K], data: Dict[str, Any]) -> K: as_dict: Dict[str, Any] = {} for field_name, field in cls.__fields__.items(): if field_name not in data: @@ -100,7 +103,8 @@ def _deserialize(cls, data: Dict[str, Any]) -> Self: elif field.type_ in DETA_TYPES: as_dict[field_name] = data[field_name] elif field.type_ == datetime.datetime: - as_dict[field_name] = datetime.datetime.fromtimestamp(data[field_name]) + as_dict[field_name] = datetime.datetime.fromtimestamp( + data[field_name]) elif field.type_ == datetime.date: as_dict[field_name] = datetime.datetime.strptime( str(data[field_name]), "%Y%m%d" @@ -120,7 +124,7 @@ def _deserialize(cls, data: Dict[str, Any]) -> Self: return cls.parse_obj(as_dict) @classmethod - def _return_item_or_raise(cls, item: Optional[Dict[str, Any]]) -> Self: + def _return_item_or_raise(cls: Type[K], item: Optional[Dict[str, Any]]) -> K: if item is None or item.get("key") == "None": raise ItemNotFound("Could not find item matching that key") try: @@ -129,9 +133,12 @@ def _return_item_or_raise(cls, item: Optional[Dict[str, Any]]) -> Self: raise ItemNotFound("Could not find item matching that key") +T = TypeVar("T", bound="DetaModel") + + class DetaModel(BaseDetaModel, metaclass=DetaModelMetaClass): @classmethod - def get(cls, key: str) -> Self: + def get(cls: Type[T], key: str) -> T: """ Get a single instance :param key: Deta database key @@ -146,7 +153,7 @@ def get(cls, key: str) -> Self: return cls._return_item_or_raise(item) @classmethod - def get_or_none(cls, key: str) -> Optional[Self]: + def get_or_none(cls: Type[T], key: str) -> Optional[T]: """Try to get item by key or return None if item not found""" try: return cls.get(key) @@ -154,7 +161,7 @@ def get_or_none(cls, key: str) -> Optional[Self]: return None @classmethod - def get_all(cls) -> List[Self]: + def get_all(cls: Type[T]) -> List[T]: """Get all the records from the database""" response: FetchResponse = cls.__db__.fetch() records: List[Dict[str, Any]] = response.items @@ -166,24 +173,25 @@ def get_all(cls) -> List[Self]: @classmethod def query( - cls, query_statement: Union[DetaQuery, DetaQueryStatement, DetaQueryList] - ) -> List[Self]: + cls: Type[T], query_statement: Union[DetaQuery, DetaQueryStatement, DetaQueryList] + ) -> List[T]: """Get items from database based on the query.""" response: FetchResponse = cls.__db__.fetch(query_statement.as_query()) records: List[Dict[str, Any]] = response.items while response.last: - response = cls.__db__.fetch(query_statement.as_query(), last=response.last) + response = cls.__db__.fetch( + query_statement.as_query(), last=response.last) records += response.items return [cls._deserialize(item) for item in records] @classmethod - def delete_key(cls, key: str): + def delete_key(cls, key: str) -> None: """Delete an item based on the key""" cls.__db__.delete(key) @classmethod - def put_many(cls, items: List[Self]) -> List[Self]: + def put_many(cls: Type[T], items: List[T]) -> List[T]: """Put multiple instances at once :param items: List of pydantic objects to put in the database @@ -211,7 +219,7 @@ def put_many(cls, items: List[Self]) -> List[Self]: def _db_put(cls, data: List[Dict[str, Any]]) -> List[Dict[str, Any]]: return cls.__db__.put(data) # type: ignore - def save(self): + def save(self) -> None: """Saves the record to the database. Behaves as upsert, will create if not present. Database key will then be set on the object.""" # exclude = set() @@ -222,7 +230,7 @@ def save(self): saved = self._db_put(self._serialize()) self.key = saved["key"] - def delete(self): + def delete(self) -> None: """Delete the open object from the database. The object will still exist in python, but will be deleted from the database and the key attribute will be set to None."""

Async Support An (unofficial) [asynchronous library](https://github.com/leits/aiodeta) for the Deta api has become available. I'd like to support it. Without breaking compatibility, it could be something like `from odetam.async import DetaModel` and then provide async/await versions of the existing api.

2023-05-28T04:38:24

0.0

[]

amjith/fuzzyfinder

amjith__fuzzyfinder-19

43fe7676cad68e269bbace7bb2fd9b77f2e07da9

diff --git a/fuzzyfinder/main.py b/fuzzyfinder/main.py index fa5c14b..ab9c4f4 100755 --- a/fuzzyfinder/main.py +++ b/fuzzyfinder/main.py @@ -3,7 +3,9 @@ from . import export @export -def fuzzyfinder(input, collection, accessor=lambda x: x, sort_results=True): +def fuzzyfinder( + input, collection, accessor=lambda x: x, sort_results=True, ignore_case=True +): """ Args: input (str): A partial string which is typically entered by a user. @@ -12,13 +14,15 @@ def fuzzyfinder(input, collection, accessor=lambda x: x, sort_results=True): accessor (function): If the `collection` is not an iterable of strings, then use the accessor to fetch the string that will be used for fuzzy matching. - sort_results(bool): The suggestions are sorted by considering the - smallest contiguous match, followed by where the - match is found in the full string. If two suggestions - have the same rank, they are then sorted - alpha-numerically. This parameter controls the - *last tie-breaker-alpha-numeric sorting*. The sorting - based on match length and position will be intact. + sort_results (bool): The suggestions are sorted by considering the + smallest contiguous match, followed by where the + match is found in the full string. If two suggestions + have the same rank, they are then sorted + alpha-numerically. This parameter controls the + *last tie-breaker-alpha-numeric sorting*. The sorting + based on match length and position will be intact. + ignore_case (bool): If this parameter is set to False, the filtering + is case-sensitive. Returns: suggestions (generator): A generator object that produces a list of @@ -28,7 +32,7 @@ def fuzzyfinder(input, collection, accessor=lambda x: x, sort_results=True): input = str(input) if not isinstance(input, str) else input pat = '.*?'.join(map(re.escape, input)) pat = '(?=({0}))'.format(pat) # lookahead regex to manage overlapping matches - regex = re.compile(pat, re.IGNORECASE) + regex = re.compile(pat, re.IGNORECASE if ignore_case else 0) for item in collection: r = list(regex.finditer(accessor(item))) if r:

case_insensitive as optional argument? I have some old code that reads: ```python from fuzzyfinder import fuzzyfinder matches = fuzzyfinder(word_before_cursor, fuzzy_words, case_sensitive=True) ``` But I can't see in this project when case_sensitive was ever an option? Am I going crazy? Was case_sensitive the default and the extra parameter I was passing just doing nothing?

I think I just changed my local version :-/ Ah well! I'll leave this here because I think it makes sense and may even create a pull request.

2024-08-28T11:28:46

0.0

[]

lareferencia/dspace-stats-collector

lareferencia__dspace-stats-collector-19

85b02cf6609496d8926739a2980931782067e57b

diff --git a/dspace_stats_collector/collector.py b/dspace_stats_collector/collector.py index 9af9646..ddffa56 100644 --- a/dspace_stats_collector/collector.py +++ b/dspace_stats_collector/collector.py @@ -88,6 +88,9 @@ def run(): loglevel = logging.DEBUG else: loglevel = logging.INFO + + if args.debug: + loglevel = logging.DEBUG logDirName = os.path.expanduser('~') + "/dspace-stats-collector/var/logs" if not os.path.exists(logDirName): @@ -98,10 +101,16 @@ def run(): else: logFileName = "dspace-stats-collector.log" + logging_handlers = [ logging.FileHandler("{}/{}".format(logDirName, logFileName), 'w+') ] + + # if we are in verbose mode, also log to console + if args.verbose: + logging_handlers.append(logging.StreamHandler()) + + logging.basicConfig(level=loglevel, - format="%(levelname)s: %(message)s", - filename="{}/{}".format(logDirName, logFileName), - filemode='a') + format="%(asctime)s %(levelname)s: %(message)s", + handlers=logging_handlers) logger.debug("Verbose: %s" % args.verbose) logger.debug("Repository: %s" % args.repository) @@ -173,6 +182,11 @@ def valid_date_type(arg_date_str): help="increase output verbosity", default=False, action="store_true") + parser.add_argument("--debug", + help="debug mode", + default=False, + action="store_true") + parser.add_argument("-a", "--archived_core", metavar="YYYY", diff --git a/dspace_stats_collector/configcontext.py b/dspace_stats_collector/configcontext.py index 59f86cb..3332e1b 100644 --- a/dspace_stats_collector/configcontext.py +++ b/dspace_stats_collector/configcontext.py @@ -17,18 +17,25 @@ from .dspacedb4 import DSpaceDB4 from .dspacedb5 import DSpaceDB5 from .dspacedb6 import DSpaceDB6 + from .dspacedb7 import DSpaceDB7 + from .dspacedb5cris import DSpaceDB5Cris + from .dspacedb5oracle import DSpaceDB5Oracle + from .dspacedb6oracle import DSpaceDB6Oracle except Exception: #ImportError from dspacedb4 import DSpaceDB4 from dspacedb5 import DSpaceDB5 from dspacedb6 import DSpaceDB6 - + from dspacedb7 import DSpaceDB7 + from dspacedb5cris import DSpaceDB5Cris + from dspacedb5oracle import DSpaceDB5Oracle + from dspacedb6oracle import DSpaceDB6Oracle SAVE_DIR = os.path.expanduser('~') + "/dspace-stats-collector/var/timestamp" DEFAULT_INSTALL_PATH = os.path.expanduser('~') + "/dspace-stats-collector" DEFAULT_COLLECTOR_COMMAND_NAME="dspace-stats-collector" DEFAULT_CONFIG_PATH = DEFAULT_INSTALL_PATH + "/config" -SOLR_STATS_CORE_NAME = "statistics" +DEFAULT_SOLR_STATS_CORE_NAME = "statistics" TIMESTAMP_PATTERN = "%Y-%m-%dT00:00:00.000Z" SOLR_QUERY_ROWS_SIZE = 10 DEFAULT_OUPUT_LIMIT = 100 @@ -94,9 +101,9 @@ def __init__(self, repoName, commandLineArgs): # Solr Context if commandLineArgs.archived_core != None: - self.solrStatsCoreName = SOLR_STATS_CORE_NAME + "-" + commandLineArgs.archived_core + self.solrStatsCoreName = self.getSolrStatsCoreName() + "-" + commandLineArgs.archived_core else: - self.solrStatsCoreName = SOLR_STATS_CORE_NAME + self.solrStatsCoreName = self.getSolrStatsCoreName() self.solrServerURL = self._find_solr_server() self.solrStatsCoreURL = self.solrServerURL + "/" + self.solrStatsCoreName @@ -135,6 +142,14 @@ def __init__(self, repoName, commandLineArgs): self.db = DSpaceDB5(self.dspaceProperties['db.url'],self.dspaceProperties['db.username'],self.dspaceProperties['db.password']) elif self.dspaceMajorVersion == '6': self.db = DSpaceDB6(self.dspaceProperties['db.url'],self.dspaceProperties['db.username'],self.dspaceProperties['db.password']) + elif self.dspaceMajorVersion == '7': + self.db = DSpaceDB7(self.dspaceProperties['db.url'],self.dspaceProperties['db.username'],self.dspaceProperties['db.password']) + elif self.dspaceMajorVersion == '5o': + self.db = DSpaceDB5Oracle(self.dspaceProperties['db.url'],self.dspaceProperties['db.username'],self.dspaceProperties['db.password']) + elif self.dspaceMajorVersion == '6o': + self.db = DSpaceDB6Oracle(self.dspaceProperties['db.url'],self.dspaceProperties['db.username'],self.dspaceProperties['db.password']) + elif self.dspaceMajorVersion == '5c': + self.db = DSpaceDB5Cris(self.dspaceProperties['db.url'],self.dspaceProperties['db.username'],self.dspaceProperties['db.password']) else: logger.error('Only implemented values for dspace.majorVersion are 4, 5 and 6. Received {}'.format(self.dspaceMajorVersion)) raise NotImplementedError @@ -163,7 +178,10 @@ def getSolrLimit(self): return int(self.properties['solr.limit']) def getDspaceMajorVersion(self): - return int(self.properties['dspace.majorVersion']) + return str(self.properties['dspace.majorVersion']) + + def getSolrStatsCoreName(self): + return str( self.properties.get('solr.core', DEFAULT_SOLR_STATS_CORE_NAME) ) ################################################ private methods ########################################## def _read_properties(self): @@ -183,7 +201,9 @@ def _read_properties(self): def _read_dspace_properties(self): javaprops = JavaProperties() - if self.getDspaceMajorVersion() == 6: + if self.getDspaceMajorVersion() == '6': + + ## try to read dspace.cfg try: propertiesFilename = "%s/config/dspace.cfg" % (self.properties["dspace.dir"]) javaprops.load(open(propertiesFilename)) @@ -193,6 +213,7 @@ def _read_dspace_properties(self): logger.exception("Error while trying to read properties file %s" % propertiesFilename) raise + ## try to read local.cfg try: propertiesFilename = "%s/config/local.cfg" % (self.properties["dspace.dir"]) javaprops.load(open(propertiesFilename)) @@ -202,6 +223,16 @@ def _read_dspace_properties(self): logger.debug("Could not read property file %s" % propertiesFilename) pass + elif self.getDspaceMajorVersion() == '5c': + try: + propertiesFilename = "%s/build.properties" % (self.properties["dspace.dir"]) + javaprops.load(open(propertiesFilename)) + property_dict = javaprops.get_property_dict() + logger.debug("Read succesfully property file %s" % propertiesFilename) + except (FileNotFoundError, UnboundLocalError): + logger.exception("Error while trying to read properties file %s" % propertiesFilename) + raise + else: try: propertiesFilename = "%s/config/dspace.cfg" % (self.properties["dspace.dir"]) diff --git a/dspace_stats_collector/dspacedb.py b/dspace_stats_collector/dspacedb.py index 4faf689..500deac 100644 --- a/dspace_stats_collector/dspacedb.py +++ b/dspace_stats_collector/dspacedb.py @@ -17,26 +17,36 @@ def __init__(self, jdbcUrl, username, password): # Parse jdbc url # Postgres template: jdbc:postgresql://localhost:5432/dspace # Oracle template: jdbc:oracle:thin:@//localhost:1521/xe - m = re.match("^jdbc:(postgresql|oracle):[^\/]*\/\/([^:]+):(\d+)/(.*)$", jdbcUrl) + # Oracle template: jdbc:oracle:thin:@localhost:1521:xe + m = re.match("^jdbc:(postgresql|oracle):[^\/|^@]*[@\/\/|\/\/|@]*([^:]+):(\d+)(\/|:)(.*)$", jdbcUrl) + if m is None: logger.error("Could not parse db.url string: %s" % jdbcUrl) raise ValueError - (engine, hostname, port, database) = m.group(1, 2, 3, 4) - - if engine != "postgresql": - logger.error("DB Engine not yet supported: %s" % engine) - raise NotImplementedError - - self.connString = '{engine}://{username}:{password}@{hostname}:{port}/{database}' - self.connString = self.connString.format( + # separate host, port, dbname, and dbschema and separator + (engine, hostname, port, separator, database) = m.group(1, 2, 3, 4, 5) + + # Create connection string depending on engine + if engine == 'postgresql': + connStringTemplate = 'postgresql://{username}:{password}@{hostname}:{port}' + separator + '{database}' + if engine == 'oracle': + if separator == ':': + connStringTemplate = 'oracle+cx_oracle://{username}:{password}@{hostname}:{port}/?service_name={database}' + elif separator == '/': + connStringTemplate = 'oracle+cx_oracle://{username}:{password}@{hostname}:{port}/{database}' + + + # Create connection string based on engine, hostname, port, database. Use separator to determine the kind of oracle connection + self.connString = connStringTemplate.format( engine=engine, username=username, password=password, hostname=hostname, port=port, database=database, - ) + ) + logger.debug('DB Connection String: ' + self.connString) try: self.conn = sqlalchemy.create_engine(self.connString).connect() diff --git a/dspace_stats_collector/dspacedb5cris.py b/dspace_stats_collector/dspacedb5cris.py new file mode 100644 index 0000000..bbc509f --- /dev/null +++ b/dspace_stats_collector/dspacedb5cris.py @@ -0,0 +1,71 @@ +#!/usr/bin/env python +# -*- coding: utf-8 -*- +""" Dspace DB components """ + +import logging +logger = logging.getLogger() + +try: + from .dspacedb import DSpaceDB +except Exception: #ImportError + from dspacedb import DSpaceDB + +class DSpaceDB5Cris(DSpaceDB): + + def __init__(self, jdbcUrl, username, password): + + DSpaceDB.__init__(self,jdbcUrl, username, password) + + self._queryDownloadSQL = """ + SELECT mv.resource_id AS id, + mv2.text_value AS record_title, + h.handle AS handle, + true AS is_download, + i.item_id AS owning_item, + b.sequence_id AS sequence_id, + mv.text_value AS filename + FROM metadatavalue AS mv + RIGHT JOIN bitstream AS b ON mv.resource_id = b.bitstream_id + RIGHT JOIN bundle2bitstream AS bb ON b.bitstream_id = bb.bitstream_id + RIGHT JOIN item2bundle AS i ON i.bundle_id = bb.bundle_id + RIGHT JOIN handle AS h ON h.resource_id = i.item_id + RIGHT JOIN metadatavalue AS mv2 ON mv2.resource_id = i.item_id + WHERE mv.metadata_field_id = {dcTitleId} + AND mv.resource_type_id = 0 + AND b.sequence_id IS NOT NULL + AND b.deleted = FALSE + AND mv2.metadata_field_id = {dcTitleId} + AND mv2.resource_type_id=2 + AND mv.resource_id = {bitstreamId}; + """ + + self._queryItemSQL = """ + SELECT mv.resource_id AS id, + mv.text_value AS record_title, + h.handle AS handle, + false AS is_download, + NULL AS owning_item, + NULL AS sequence_id, + NULL AS filename + FROM metadatavalue AS mv + RIGHT JOIN handle AS h ON h.resource_id = mv.resource_id + WHERE metadata_field_id = {dcTitleId} + AND mv.resource_type_id=2 + AND h.resource_type_id=2 + AND mv.resource_id = {itemId}; + """ + + self._queryTitleSQL = """ + SELECT metadata_field_id AS "dcTitleId" + FROM metadatafieldregistry mfr, + metadataschemaregistry msr + WHERE mfr.metadata_schema_id = msr.metadata_schema_id + AND short_id = 'dc' + AND element = 'title' + AND qualifier IS NULL; + """ + + self._dcTitleId = self.getDcTitleId() + + + diff --git a/dspace_stats_collector/dspacedb5oracle.py b/dspace_stats_collector/dspacedb5oracle.py new file mode 100644 index 0000000..abc602d --- /dev/null +++ b/dspace_stats_collector/dspacedb5oracle.py @@ -0,0 +1,71 @@ +#!/usr/bin/env python +# -*- coding: utf-8 -*- +""" Dspace DB components """ + +import logging +logger = logging.getLogger() + +try: + from .dspacedb import DSpaceDB +except Exception: #ImportError + from dspacedb import DSpaceDB + +class DSpaceDB5Oracle(DSpaceDB): + + def __init__(self, jdbcUrl, username, password): + + DSpaceDB.__init__(self,jdbcUrl, username, password) + + self._queryDownloadSQL = """ + SELECT mv.resource_id AS id, + mv2.text_value AS record_title, + h.handle AS handle, + true AS is_download, + i.item_id AS owning_item, + b.sequence_id AS sequence_id, + mv.text_value AS filename + FROM metadatavalue AS mv + RIGHT JOIN bitstream AS b ON mv.resource_id = b.bitstream_id + RIGHT JOIN bundle2bitstream AS bb ON b.bitstream_id = bb.bitstream_id + RIGHT JOIN item2bundle AS i ON i.bundle_id = bb.bundle_id + RIGHT JOIN handle AS h ON h.resource_id = i.item_id + RIGHT JOIN metadatavalue AS mv2 ON mv2.resource_id = i.item_id + WHERE mv.metadata_field_id = {dcTitleId} + AND mv.resource_type_id = 0 + AND b.sequence_id IS NOT NULL + AND b.deleted = FALSE + AND mv2.metadata_field_id = {dcTitleId} + AND mv2.resource_type_id=2 + AND mv.resource_id = {bitstreamId}; + """ + + self._queryItemSQL = """ + SELECT mv.resource_id AS id, + mv.text_value AS record_title, + h.handle AS handle, + false AS is_download, + NULL AS owning_item, + NULL AS sequence_id, + NULL AS filename + FROM metadatavalue AS mv + RIGHT JOIN handle AS h ON h.resource_id = mv.resource_id + WHERE metadata_field_id = {dcTitleId} + AND mv.resource_type_id=2 + AND h.resource_type_id=2 + AND mv.resource_id = {itemId}; + """ + + self._queryTitleSQL = """ + SELECT metadata_field_id AS "dcTitleId" + FROM metadatafieldregistry mfr, + metadataschemaregistry msr + WHERE mfr.metadata_schema_id = msr.metadata_schema_id + AND short_id = 'dc' + AND element = 'title' + AND qualifier IS NULL; + """ + + self._dcTitleId = self.getDcTitleId() + + + diff --git a/dspace_stats_collector/dspacedb6.py b/dspace_stats_collector/dspacedb6.py index 240d9b1..8ec0b94 100644 --- a/dspace_stats_collector/dspacedb6.py +++ b/dspace_stats_collector/dspacedb6.py @@ -34,7 +34,7 @@ def __init__(self, jdbcUrl, username, password): AND b.sequence_id IS NOT NULL AND b.deleted = FALSE AND mv2.metadata_field_id = {dcTitleId} - AND mv.dspace_object_id::text = '{bitstreamId}'; + AND mv.dspace_object_id = uuid('{bitstreamId}'); """ self._queryItemSQL = """ @@ -49,7 +49,7 @@ def __init__(self, jdbcUrl, username, password): RIGHT JOIN handle AS h ON h.resource_id = mv.dspace_object_id WHERE metadata_field_id = {dcTitleId} AND h.resource_type_id=2 - AND mv.dspace_object_id::text = '{itemId}'; + AND mv.dspace_object_id = uuid('{itemId}'); """ self._queryTitleSQL = """ diff --git a/dspace_stats_collector/dspacedb6oracle.py b/dspace_stats_collector/dspacedb6oracle.py new file mode 100644 index 0000000..7b3098b --- /dev/null +++ b/dspace_stats_collector/dspacedb6oracle.py @@ -0,0 +1,69 @@ +#!/usr/bin/env python +# -*- coding: utf-8 -*- +""" Dspace DB components """ + +import logging +logger = logging.getLogger() + +try: + from .dspacedb import DSpaceDB +except Exception: #ImportError + from dspacedb import DSpaceDB + +class DSpaceDB6Oracle(DSpaceDB): + + def __init__(self, jdbcUrl, username, password): + + DSpaceDB.__init__(self,jdbcUrl, username, password) + + self._queryDownloadSQL = """ + SELECT mv.dspace_object_id::text AS id, + mv2.text_value AS record_title, + h.handle AS handle, + true AS is_download, + i.item_id::text AS owning_item, + b.sequence_id AS sequence_id, + mv.text_value AS filename + FROM metadatavalue AS mv + RIGHT JOIN bitstream AS b ON mv.dspace_object_id = b.uuid + RIGHT JOIN bundle2bitstream AS bb ON b.uuid = bb.bitstream_id + RIGHT JOIN item2bundle AS i ON i.bundle_id = bb.bundle_id + RIGHT JOIN handle AS h ON h.resource_id = i.item_id + RIGHT JOIN metadatavalue AS mv2 ON mv2.dspace_object_id = i.item_id + WHERE mv.metadata_field_id = {dcTitleId} + AND b.sequence_id IS NOT NULL + AND b.deleted = FALSE + AND mv2.metadata_field_id = {dcTitleId} + AND mv.dspace_object_id = uuid('{bitstreamId}'); + """ + + self._queryItemSQL = """ + SELECT mv.dspace_object_id::text AS id, + mv.text_value AS record_title, + h.handle AS handle, + false AS is_download, + NULL AS owning_item, + NULL AS sequence_id, + NULL AS filename + FROM metadatavalue AS mv + RIGHT JOIN handle AS h ON h.resource_id = mv.dspace_object_id + WHERE metadata_field_id = {dcTitleId} + AND h.resource_type_id=2 + AND mv.dspace_object_id = uuid('{itemId}'); + """ + + self._queryTitleSQL = """ + SELECT metadata_field_id AS "dcTitleId" + FROM metadatafieldregistry mfr, + metadataschemaregistry msr + WHERE mfr.metadata_schema_id = msr.metadata_schema_id + AND short_id = 'dc' + AND element = 'title' + AND qualifier IS NULL; + """ + + self._dcTitleId = self.getDcTitleId() + + + + \ No newline at end of file diff --git a/dspace_stats_collector/dspacedb7.py b/dspace_stats_collector/dspacedb7.py new file mode 100644 index 0000000..98c1a3e --- /dev/null +++ b/dspace_stats_collector/dspacedb7.py @@ -0,0 +1,69 @@ +#!/usr/bin/env python +# -*- coding: utf-8 -*- +""" Dspace DB components """ + +import logging +logger = logging.getLogger() + +try: + from .dspacedb import DSpaceDB +except Exception: #ImportError + from dspacedb import DSpaceDB + +class DSpaceDB7(DSpaceDB): + + def __init__(self, jdbcUrl, username, password): + + DSpaceDB.__init__(self,jdbcUrl, username, password) + + self._queryDownloadSQL = """ + SELECT mv.dspace_object_id::text AS id, + mv2.text_value AS record_title, + h.handle AS handle, + true AS is_download, + i.item_id::text AS owning_item, + b.sequence_id AS sequence_id, + mv.text_value AS filename + FROM metadatavalue AS mv + RIGHT JOIN bitstream AS b ON mv.dspace_object_id = b.uuid + RIGHT JOIN bundle2bitstream AS bb ON b.uuid = bb.bitstream_id + RIGHT JOIN item2bundle AS i ON i.bundle_id = bb.bundle_id + RIGHT JOIN handle AS h ON h.resource_id = i.item_id + RIGHT JOIN metadatavalue AS mv2 ON mv2.dspace_object_id = i.item_id + WHERE mv.metadata_field_id = {dcTitleId} + AND b.sequence_id IS NOT NULL + AND b.deleted = FALSE + AND mv2.metadata_field_id = {dcTitleId} + AND mv.dspace_object_id = uuid('{bitstreamId}'); + """ + + self._queryItemSQL = """ + SELECT mv.dspace_object_id::text AS id, + mv.text_value AS record_title, + h.handle AS handle, + false AS is_download, + NULL AS owning_item, + NULL AS sequence_id, + NULL AS filename + FROM metadatavalue AS mv + RIGHT JOIN handle AS h ON h.resource_id = mv.dspace_object_id + WHERE metadata_field_id = {dcTitleId} + AND h.resource_type_id=2 + AND mv.dspace_object_id = uuid('{itemId}'); + """ + + self._queryTitleSQL = """ + SELECT metadata_field_id AS "dcTitleId" + FROM metadatafieldregistry mfr, + metadataschemaregistry msr + WHERE mfr.metadata_schema_id = msr.metadata_schema_id + AND short_id = 'dc' + AND element = 'title' + AND qualifier IS NULL; + """ + + self._dcTitleId = self.getDcTitleId() + + + + \ No newline at end of file diff --git a/dspace_stats_collector/matomooutput.py b/dspace_stats_collector/matomooutput.py index 09bafc4..6cff898 100644 --- a/dspace_stats_collector/matomooutput.py +++ b/dspace_stats_collector/matomooutput.py @@ -37,8 +37,13 @@ def __init__(self, configContext): self._handleCanonicalPrefix = dspaceProperties['handle.canonical.prefix'] else: self._handleCanonicalPrefix = 'http://hdl.handle.net/' - self._dspaceHostname = dspaceProperties['dspace.hostname'] - self._dspaceUrl = dspaceProperties['dspace.url'] + + if configContext.getDspaceMajorVersion() == '7': + self._dspaceHostname = dspaceProperties['dspace.server.url'] + self._dspaceUrl = dspaceProperties['dspace.ui.url'] + else: + self._dspaceHostname = dspaceProperties['dspace.hostname'] + self._dspaceUrl = dspaceProperties['dspace.url'] self._repoProperties = configContext.properties @@ -122,6 +127,8 @@ def getTotalSent(self): def send(self, event): self._buffer.append((event._matomoRequest, event.is_robot, event._src['time'])) + logger.debug( "Event json dump: {}".format( event.toJSON() ) ) + #print(self._buffer) if self.isBufferFull(): logger.debug("Buffer is full") @@ -193,7 +200,9 @@ def flush(self): raise except MatomoInternalServerException as e: # if there is some internal error will discard this event and log the result - logger.error('Matomo internal error occurred: {} with event. This event will be discarded.\n'.format( str(e)) ) + logger.error('Matomo internal error occurred: {} with event. This event will be discarded.\n'.format( str(e)) ) + logger.debug('Request URL was: {}'.format(self._url) ) + logger.debug('Request Payload was: {}'.format(m) ) diff --git a/install-standalone-develop.sh b/install-standalone-develop.sh new file mode 100644 index 0000000..a49b8d2 --- /dev/null +++ b/install-standalone-develop.sh @@ -0,0 +1,56 @@ +#!/bin/bash +# + +INSTALL_PATH=$HOME/dspace-stats-collector +MINICONDA_URL_PREFIX='https://repo.anaconda.com/miniconda/' + +rm -rf $INSTALL_PATH +rm -f delete_this_file.sh + +MACHINE_TYPE=`uname -m` +if [ ${MACHINE_TYPE} == 'x86_64' ]; then + # 64-bit stuff here + MINICONDA_FILE='Miniconda3-py37_4.8.3-Linux-x86_64.sh' +elif [ ${MACHINE_TYPE} == 'x86' ]; then + # 32-bit stuff here + MINICONDA_FILE='Miniconda3-py37_4.8.3-Linux-x86.sh' +elif [ ${MACHINE_TYPE} == 'aarch64' ]; then + # ARM stuff here (EXPERIMENTAL) + MINICONDA_FILE='Miniconda3-py37_4.9.2-Linux-aarch64.sh' +else + echo "Unknown machine type: ${MACHINE_TYPE}" + exit 1 +fi + +MINICONDA_URL=$MINICONDA_URL_PREFIX$MINICONDA_FILE + +echo $MINICONDA_URL + +if [ -x "$(which curl)" ]; then + curl $MINICONDA_URL -o delete_this_file.sh +else + echo "Could not find curl, please install curl." >&2 + exit 1 +fi + +bash delete_this_file.sh -b -f -p $INSTALL_PATH +rm delete_this_file.sh + +cd $INSTALL_PATH + +echo "Installing dspace-stats-collector package dependencies" +curl https://raw.githubusercontent.com/lareferencia/dspace-stats-collector/develop/requirements-p37-dev.txt -o $INSTALL_PATH/requirements.txt +$INSTALL_PATH/bin/pip install -r $INSTALL_PATH/requirements.txt + +$INSTALL_PATH/bin/conda install -y git + +$INSTALL_PATH/bin/git clone --branch=develop https://github.com/lareferencia/dspace-stats-collector.git + +cd $INSTALL_PATH/dspace-stats-collector +$INSTALL_PATH/bin/python setup.py install + +echo "Installing config files" +$INSTALL_PATH/bin/dspace-stats-configure + + + diff --git a/requirements-p37-dev.txt b/requirements-p37-dev.txt new file mode 100644 index 0000000..09887d4 --- /dev/null +++ b/requirements-p37-dev.txt @@ -0,0 +1,25 @@ +bumpversion==0.5.3 +wheel==0.32.1 +watchdog==0.9.0 +flake8==3.5.0 +tox==3.5.2 +coverage==4.5.1 +Sphinx==1.8.1 +twine==1.12.1 + +pytest==3.8.2 +pytest-runner==4.2 + +requests==2.21.0 +pyjavaprops==1.0.2 +SQLAlchemy==1.2.15 +psycopg2-binary==2.8 +cx-Oracle +pysolr==3.8.1 +pandas==1.2 +urllib3==1.24.2 +pytz==2018.7 +python-crontab +anonymizeip +pid + diff --git a/update-standalone-develop.sh b/update-standalone-develop.sh new file mode 100644 index 0000000..839f490 --- /dev/null +++ b/update-standalone-develop.sh @@ -0,0 +1,19 @@ +#!/bin/bash +# + +INSTALL_PATH=$HOME/dspace-stats-collector + +cd $INSTALL_PATH + +echo "Installing dspace-stats-collector package dependencies" +curl https://raw.githubusercontent.com/lareferencia/dspace-stats-collector/develop/requirements-p37.txt -o $INSTALL_PATH/requirements.txt +$INSTALL_PATH/bin/pip install -r $INSTALL_PATH/requirements.txt + +cd $INSTALL_PATH/dspace-stats-collector + +$INSTALL_PATH/bin/git pull +$INSTALL_PATH/bin/python setup.py install + + + +

Cambio consulta SQL de DSpace6 y DSpace7 para comparar por uuid y no texto y mejorar rendimiento Modificación de las consultas SQL para la versión 6 i 7 de DSpace con postgres para que utilize el indice para comprobar la condición dspace_object_id sea igual al uuid buscado. Mejora mucho el rendimiento, una consulta puede pasar de 16 segundos a mili-segundos. Probado para la versión 6 pero también se debería aplicar a la 7 ya que esta parte es común.

2022-09-08T16:27:37

0.0

[]

scikit-hep/hepstats

scikit-hep__hepstats-103

53256cc5ef62e1d781627a9b0df24f136f2a4141

diff --git a/.gitignore b/.gitignore index 785f8ec8..4669666d 100644 --- a/.gitignore +++ b/.gitignore @@ -13,3 +13,5 @@ dist/ /pip-wheel-metadata /docs/_build/* /docs/source/* +/_build/** +/.idea/** diff --git a/CHANGELOG.rst b/CHANGELOG.rst index ecf9d4d5..5e20ba03 100644 --- a/CHANGELOG.rst +++ b/CHANGELOG.rst @@ -4,6 +4,8 @@ Changelog master ****** +* fix toy generation with constraints + Version 0.6.0 ************* diff --git a/src/hepstats/modeling/bayesian_blocks.py b/src/hepstats/modeling/bayesian_blocks.py index 5281e9c8..2a01013d 100644 --- a/src/hepstats/modeling/bayesian_blocks.py +++ b/src/hepstats/modeling/bayesian_blocks.py @@ -14,9 +14,9 @@ from __future__ import annotations from collections.abc import Iterable + import numpy as np import pandas as pd -from typing import Optional, Union class Prior: diff --git a/src/hepstats/splot/sweights.py b/src/hepstats/splot/sweights.py index ac927f6e..65cdb10d 100644 --- a/src/hepstats/splot/sweights.py +++ b/src/hepstats/splot/sweights.py @@ -1,13 +1,14 @@ from __future__ import annotations -import numpy as np -from typing import Any import warnings +from typing import Any + +import numpy as np -from ..utils import eval_pdf -from ..utils.fit.api_check import is_valid_pdf from .exceptions import ModelNotFittedToData from .warnings import AboveToleranceWarning +from ..utils import eval_pdf +from ..utils.fit.api_check import is_valid_pdf def is_sum_of_extended_pdfs(model) -> bool: diff --git a/src/hepstats/utils/fit/__init__.py b/src/hepstats/utils/fit/__init__.py index d67fc4aa..e7a31b74 100644 --- a/src/hepstats/utils/fit/__init__.py +++ b/src/hepstats/utils/fit/__init__.py @@ -1,2 +1,2 @@ -from .sampling import base_sampler, base_sample from .diverse import get_value, eval_pdf, pll, array2dataset, get_nevents, set_values +from .sampling import base_sampler, base_sample diff --git a/src/hepstats/utils/fit/diverse.py b/src/hepstats/utils/fit/diverse.py index 7dc5ffc2..891eeeb7 100644 --- a/src/hepstats/utils/fit/diverse.py +++ b/src/hepstats/utils/fit/diverse.py @@ -1,7 +1,6 @@ from contextlib import ExitStack, contextmanager -import numpy as np -from .api_check import is_valid_pdf +import numpy as np def get_ndims(dataset):

fix of iteration over the param_dict fix of iteration over the param_dict

a fix to this issue https://github.com/scikit-hep/hepstats/issues/101 a fix to this issue https://github.com/scikit-hep/hepstats/issues/101

2023-04-17T15:57:42

0.0

[]

scikit-hep/hepstats

scikit-hep__hepstats-66

c70c31fdd20b330961c2a304092c0796bcb7bca5

diff --git a/.github/workflows/auto-cancellation.yml b/.github/workflows/auto-cancellation.yml deleted file mode 100644 index 8c33f6c7..00000000 --- a/.github/workflows/auto-cancellation.yml +++ /dev/null @@ -1,9 +0,0 @@ -name: auto cancellation running job -on: pull_request - -jobs: - cancel: - name: auto-cancellation-running-action - runs-on: ubuntu-latest - steps: - - uses: fauguste/[email protected] diff --git a/.github/workflows/main.yml b/.github/workflows/main.yml index a7ff0069..ea6afc49 100644 --- a/.github/workflows/main.yml +++ b/.github/workflows/main.yml @@ -6,17 +6,27 @@ on: branches: [ "master"] release: types: - - "published" + - "published" jobs: + cancel: + name: 'Cancel Previous Runs' + runs-on: ubuntu-latest + timeout-minutes: 3 + steps: + - uses: styfle/[email protected] + with: + all_but_latest: true + access_token: ${{ github.token }} + pre-commit: name: Format runs-on: ubuntu-latest steps: - - uses: actions/checkout@v3 - with: - fetch-depth: 0 - - uses: actions/setup-python@v4 + - uses: actions/checkout@v3 + with: + fetch-depth: 0 + - uses: actions/setup-python@v4 checks: runs-on: ${{ matrix.os }} @@ -24,12 +34,12 @@ jobs: fail-fast: false matrix: os: - - ubuntu-latest + - ubuntu-latest python-version: - - "3.7" - - "3.8" - - "3.9" - - "3.10" + - "3.7" + - "3.8" + - "3.9" + - "3.10" include: - os: windows-latest python-version: "3.7" @@ -41,87 +51,87 @@ jobs: python-version: "3.10" name: Check Python ${{ matrix.python-version }} ${{ matrix.os }} steps: - - uses: actions/checkout@v3 - with: - fetch-depth: 0 - - - name: Setup Python ${{ matrix.python-version }} - uses: actions/setup-python@v4 - with: - python-version: ${{ matrix.python-version }} - - - name: Install package - run: python -m pip install -e .[test] - - - name: Test package - run: python -m pytest --doctest-modules --cov=hepstats --cov-report=xml - - - name: Upload coverage to Codecov - if: matrix.python-version == '3.10' && matrix.os == 'ubuntu-latest' - uses: codecov/codecov-action@v3 - with: - file: ./coverage.xml - flags: unittests - name: codecov-umbrella - fail_ci_if_error: true + - uses: actions/checkout@v3 + with: + fetch-depth: 0 + + - name: Setup Python ${{ matrix.python-version }} + uses: actions/setup-python@v4 + with: + python-version: ${{ matrix.python-version }} + + - name: Install package + run: python -m pip install -e .[test] + + - name: Test package + run: python -m pytest --doctest-modules --cov=hepstats --cov-report=xml + + - name: Upload coverage to Codecov + if: matrix.python-version == '3.10' && matrix.os == 'ubuntu-latest' + uses: codecov/codecov-action@v3 + with: + file: ./coverage.xml + flags: unittests + name: codecov-umbrella + fail_ci_if_error: true dist: runs-on: ubuntu-latest steps: - - uses: actions/checkout@v3 - with: - fetch-depth: 0 + - uses: actions/checkout@v3 + with: + fetch-depth: 0 - - name: Build - run: pipx run build + - name: Build + run: pipx run build - - uses: actions/upload-artifact@v3 - with: - path: dist/* + - uses: actions/upload-artifact@v3 + with: + path: dist/* - - name: Check metadata - run: pipx run twine check dist/* + - name: Check metadata + run: pipx run twine check dist/* docs: runs-on: ubuntu-latest steps: - - uses: actions/checkout@v3 - with: - fetch-depth: 0 - - uses: actions/setup-python@v4 - with: - python-version: "3.10" - - name: Install dependencies - run: | - pip install -U -q -e .[docs] - pip list - - name: build docs - run: | - sphinx-build -b html docs docs/_build/html - touch docs/_build/html/.nojekyll - - - name: Deploy docs to GitHub Pages - if: success() && github.event_name == 'push' && github.ref == 'refs/heads/master' - uses: peaceiris/actions-gh-pages@v3 - with: - github_token: ${{ secrets.GITHUB_TOKEN }} - publish_dir: docs/_build/html - force_orphan: true - user_name: 'github-actions[bot]' - user_email: 'github-actions[bot]@users.noreply.github.com' - commit_message: Deploy to GitHub pages + - uses: actions/checkout@v3 + with: + fetch-depth: 0 + - uses: actions/setup-python@v4 + with: + python-version: "3.10" + - name: Install dependencies + run: | + pip install -U -q -e .[docs] + pip list + - name: build docs + run: | + sphinx-build -b html docs docs/_build/html + touch docs/_build/html/.nojekyll + + - name: Deploy docs to GitHub Pages + if: success() && github.event_name == 'push' && github.ref == 'refs/heads/master' + uses: peaceiris/actions-gh-pages@v3 + with: + github_token: ${{ secrets.GITHUB_TOKEN }} + publish_dir: docs/_build/html + force_orphan: true + user_name: 'github-actions[bot]' + user_email: 'github-actions[bot]@users.noreply.github.com' + commit_message: Deploy to GitHub pages publish: - needs: [dist] + needs: [ dist ] runs-on: ubuntu-latest if: github.event_name == 'release' && github.event.action == 'published' steps: - - uses: actions/download-artifact@v3 - with: - name: artifact - path: dist - - - uses: pypa/[email protected] - with: - password: ${{ secrets.pypi_password }} + - uses: actions/download-artifact@v3 + with: + name: artifact + path: dist + + - uses: pypa/[email protected] + with: + password: ${{ secrets.pypi_password }} diff --git a/.pre-commit-config.yaml b/.pre-commit-config.yaml index 29419ede..f1151918 100644 --- a/.pre-commit-config.yaml +++ b/.pre-commit-config.yaml @@ -1,94 +1,99 @@ repos: - - repo: https://github.com/psf/black - rev: 22.8.0 - hooks: - - id: black - - repo: https://github.com/pre-commit/pre-commit-hooks - rev: v4.3.0 - hooks: - - id: check-added-large-files - args: [ '--maxkb=1000' ] - - id: mixed-line-ending - exclude: ^notebooks/ - - id: trailing-whitespace - exclude: ^notebooks/ - - id: check-merge-conflict - - id: check-case-conflict - - id: check-symlinks - - id: check-yaml - exclude: ^notebooks/ - - id: requirements-txt-fixer - - id: debug-statements - - id: end-of-file-fixer - - repo: https://github.com/mgedmin/check-manifest - rev: "0.48" - hooks: - - id: check-manifest - args: [ --update, --no-build-isolation ] - additional_dependencies: [ setuptools-scm ] - - repo: https://github.com/pre-commit/mirrors-mypy - rev: v0.971 - hooks: - - id: mypy - files: src +- repo: https://github.com/psf/black + rev: 22.6.0 + hooks: + - id: black +- repo: https://github.com/pre-commit/pre-commit-hooks + rev: v4.3.0 + hooks: + - id: check-added-large-files + args: ['--maxkb=1000'] + - id: mixed-line-ending + exclude: ^notebooks/ + - id: trailing-whitespace + exclude: ^notebooks/ + - id: check-merge-conflict + - id: check-case-conflict + - id: check-symlinks + - id: check-yaml + exclude: ^notebooks/ + - id: requirements-txt-fixer + - id: debug-statements + - id: end-of-file-fixer +- repo: https://github.com/mgedmin/check-manifest + rev: "0.48" + hooks: + - id: check-manifest + args: [ --update, --no-build-isolation ] + additional_dependencies: [ setuptools-scm ] +- repo: https://github.com/pre-commit/mirrors-mypy + rev: v0.971 + hooks: + - id: mypy + files: src - - repo: https://github.com/pre-commit/pygrep-hooks - rev: v1.9.0 - hooks: - - id: python-use-type-annotations - - id: python-check-mock-methods - - id: python-no-eval - - id: rst-backticks - - id: rst-directive-colons +- repo: https://github.com/roy-ht/pre-commit-jupyter + rev: v1.2.1 + hooks: + - id: jupyter-notebook-cleanup - - repo: https://github.com/asottile/pyupgrade - rev: v2.37.3 - hooks: - - id: pyupgrade - args: [ --py37-plus ] +- repo: https://github.com/pre-commit/pygrep-hooks + rev: v1.9.0 + hooks: + - id: python-use-type-annotations + - id: python-check-mock-methods + - id: python-no-eval + - id: rst-backticks + - id: rst-directive-colons - - repo: https://github.com/asottile/setup-cfg-fmt - rev: v2.0.0 - hooks: - - id: setup-cfg-fmt +- repo: https://github.com/asottile/pyupgrade + rev: v2.37.3 + hooks: + - id: pyupgrade + args: [ --py37-plus ] - # Notebook formatting - - repo: https://github.com/nbQA-dev/nbQA - rev: 1.4.0 - hooks: - - id: nbqa-isort - additional_dependencies: [ isort==5.6.4 ] +- repo: https://github.com/asottile/setup-cfg-fmt + rev: v2.0.0 + hooks: + - id: setup-cfg-fmt - - id: nbqa-pyupgrade - additional_dependencies: [ pyupgrade==2.7.4 ] - args: [ --py37-plus ] +# Notebook formatting +- repo: https://github.com/nbQA-dev/nbQA + rev: 1.4.0 + hooks: + - id: nbqa-isort + additional_dependencies: [ isort==5.6.4 ] + - id: nbqa-pyupgrade + additional_dependencies: [ pyupgrade==2.7.4 ] + args: [ --py37-plus ] - - repo: https://github.com/roy-ht/pre-commit-jupyter - rev: v1.2.1 - hooks: - - id: jupyter-notebook-cleanup - - repo: https://github.com/sondrelg/pep585-upgrade - rev: 'v1.0' - hooks: - - id: upgrade-type-hints - args: [ '--futures=true' ] +- repo: https://github.com/roy-ht/pre-commit-jupyter + rev: v1.2.1 + hooks: + - id: jupyter-notebook-cleanup - - repo: https://github.com/shssoichiro/oxipng - rev: acdd66b4c5a5fda8b08281ad9b3216327b6847b4 - hooks: - - id: oxipng +- repo: https://github.com/sondrelg/pep585-upgrade + rev: 'v1.0' + hooks: + - id: upgrade-type-hints + args: [ '--futures=true' ] - - repo: https://github.com/dannysepler/rm_unneeded_f_str - rev: v0.1.0 - hooks: - - id: rm-unneeded-f-str +- repo: https://github.com/shssoichiro/oxipng + rev: acdd66b4c5a5fda8b08281ad9b3216327b6847b4 + hooks: + - id: oxipng - - repo: https://github.com/python-jsonschema/check-jsonschema - rev: 0.18.2 - hooks: - - id: check-github-workflows - - id: check-github-actions - - id: check-dependabot - - id: check-readthedocs +- repo: https://github.com/dannysepler/rm_unneeded_f_str + rev: v0.1.0 + hooks: + - id: rm-unneeded-f-str + +- repo: https://github.com/python-jsonschema/check-jsonschema + rev: 0.18.2 + hooks: + - id: check-github-workflows + - id: check-github-actions + - id: check-dependabot + - id: check-readthedocs diff --git a/CHANGELOG.rst b/CHANGELOG.rst index 6c983076..0bd52acc 100644 --- a/CHANGELOG.rst +++ b/CHANGELOG.rst @@ -13,7 +13,7 @@ Version 0.5.0 Version 0.4.0 ************* -* loss: upgrade API to use `create_new` to make sure that the losses are comparable. Compatible with zfit 0.6.4+ +* loss: upgrade API to use ``create_new`` to make sure that the losses are comparable. Compatible with zfit 0.6.4+ Version 0.3.1 ************* @@ -24,7 +24,7 @@ Version 0.3.1 Version 0.3.0 ************* * New documentation style -* `hepstats` can now do hypothesis tests, and compute upper limits and confidence intervals for counting analysis +* **hepstats** can now do hypothesis tests, and compute upper limits and confidence intervals for counting analysis * Progess bars are used to see the progression of the generation of the toys Version 0.2.5 @@ -34,8 +34,8 @@ Version 0.2.5 function if the loss is not extended * **hepstats.hypotests** can now be used even if there is no nuisances. The **pll** function in **utils/fit/diverse.py** had to be modified such that if there are no nuisances, the **pll** function returns the value of the loss function. -* add notebooks demos for FC intervals with the `FrequentistCalculator` and `AsymptoticCalculator`. -* add warnings when multiple roots are found in `ConfidenceInterval` +* add notebooks demos for FC intervals with the ``FrequentistCalculator`` and ``AsymptoticCalculator``. +* add warnings when multiple roots are found in ``ConfidenceInterval`` * move toys .yml files from notebook to notebook/toys Version 0.2.4 @@ -59,8 +59,8 @@ Version 0.2.1 * Addition of the **FrequentistCalculator** to performs hypothesis test, upper limit and interval calculations with toys. Toys can be saved and loaded in / from yaml files using the methods: - * `to_yaml` - * `from_yaml` + * ``to_yaml`` + * ``from_yaml`` Version 0.2.0 ************** diff --git a/docs/api/utils.rst b/docs/api/utils.rst index d019d137..1eab1a48 100644 --- a/docs/api/utils.rst +++ b/docs/api/utils.rst @@ -18,10 +18,10 @@ A fitting library should provide six basic objects: * minimizer * fitresult (optional) -A function for each object is defined in this module, all should return `True` to work +A function for each object is defined in this module, all should return ``True`` to work with hepstats. -The `zfit` API is currently the standard fitting API in hepstats. +The **zfit** API is currently the standard fitting API in hepstats. .. currentmodule:: hepstats.utils.fit.api_check diff --git a/docs/getting_started/splot.rst b/docs/getting_started/splot.rst index 19d72b7c..4267922c 100644 --- a/docs/getting_started/splot.rst +++ b/docs/getting_started/splot.rst @@ -1,7 +1,7 @@ splot ##### -A full example using the **sPlot** algorithm can be found `here <https://github.com/scikit-hep/hepstats/tree/master/notebooks/splots/splot_example.ipynb>`_ . **sWeights** for different components in a data sample, modeled with a sum of extended probability density functions, are derived using the `compute_sweights` function: +A full example using the **sPlot** algorithm can be found `here <https://github.com/scikit-hep/hepstats/tree/master/notebooks/splots/splot_example.ipynb>`_ . **sWeights** for different components in a data sample, modeled with a sum of extended probability density functions, are derived using the ``compute_sweights`` function: .. code-block:: python diff --git a/setup.cfg b/setup.cfg index c24b3991..a70c80d0 100644 --- a/setup.cfg +++ b/setup.cfg @@ -38,6 +38,7 @@ install_requires = pandas scipy tqdm + uhi python_requires = >=3.7 package_dir = = src diff --git a/src/hepstats/splot/sweights.py b/src/hepstats/splot/sweights.py index 4a781bde..ac927f6e 100644 --- a/src/hepstats/splot/sweights.py +++ b/src/hepstats/splot/sweights.py @@ -22,7 +22,7 @@ def is_sum_of_extended_pdfs(model) -> bool: if not hasattr(model, "get_models"): return False - return all(m.is_extended for m in model.get_models()) + return all(m.is_extended for m in model.get_models()) and model.is_extended def compute_sweights(model, x: np.ndarray) -> dict[Any, np.ndarray]: diff --git a/src/hepstats/utils/__init__.py b/src/hepstats/utils/__init__.py index 2627ff03..9ae943fe 100644 --- a/src/hepstats/utils/__init__.py +++ b/src/hepstats/utils/__init__.py @@ -1,1 +1,9 @@ -from .fit import eval_pdf, array2dataset, pll, base_sampler, base_sample, get_value +from .fit import ( + eval_pdf, + array2dataset, + pll, + base_sampler, + base_sample, + get_value, + set_values, +) diff --git a/src/hepstats/utils/fit/__init__.py b/src/hepstats/utils/fit/__init__.py index 39912ca8..d67fc4aa 100644 --- a/src/hepstats/utils/fit/__init__.py +++ b/src/hepstats/utils/fit/__init__.py @@ -1,2 +1,2 @@ from .sampling import base_sampler, base_sample -from .diverse import get_value, eval_pdf, pll, array2dataset, get_nevents +from .diverse import get_value, eval_pdf, pll, array2dataset, get_nevents, set_values diff --git a/src/hepstats/utils/fit/api_check.py b/src/hepstats/utils/fit/api_check.py index 96c800a8..53509816 100644 --- a/src/hepstats/utils/fit/api_check.py +++ b/src/hepstats/utils/fit/api_check.py @@ -16,6 +16,9 @@ The `zfit` API is currently the standard fitting API in hepstats. """ +import warnings + +import uhi.typing.plottable def is_valid_parameter(object): @@ -54,7 +57,10 @@ def is_valid_data(object): has_weights = hasattr(object, "weights") has_set_weights = hasattr(object, "set_weights") has_space = hasattr(object, "space") - return has_nevents and has_weights and has_set_weights and has_space + is_histlike = isinstance(object, uhi.typing.plottable.PlottableHistogram) + return ( + has_nevents and has_weights and has_set_weights and has_space + ) or is_histlike def is_valid_pdf(object): diff --git a/src/hepstats/utils/fit/diverse.py b/src/hepstats/utils/fit/diverse.py index 22a4c926..7dc5ffc2 100644 --- a/src/hepstats/utils/fit/diverse.py +++ b/src/hepstats/utils/fit/diverse.py @@ -1,14 +1,24 @@ -from contextlib import ExitStack +from contextlib import ExitStack, contextmanager import numpy as np +from .api_check import is_valid_pdf + + +def get_ndims(dataset): + """Return the number of dimensions in the dataset""" + return len(dataset.obs) + def get_value(value): return np.array(value) -def eval_pdf(model, x, params={}, allow_extended=False): +def eval_pdf(model, x, params=None, allow_extended=False): """Compute pdf of model at a given point x and for given parameters values""" + if params is None: + params = {} + def pdf(model, x): if model.is_extended and allow_extended: ret = model.ext_pdf(x) @@ -46,6 +56,16 @@ def pll(minimizer, loss, pois) -> float: return value +@contextmanager +def set_values(params, values): + old_values = [p.value() for p in params] + for p, v in zip(params, values): + p.set_value(v) + yield + for p, v in zip(params, old_values): + p.set_value(v) + + def array2dataset(dataset_cls, obs, array, weights=None): """ dataset_cls: only used to get the class in which array/weights will be @@ -53,9 +73,9 @@ def array2dataset(dataset_cls, obs, array, weights=None): """ if hasattr(dataset_cls, "from_numpy"): - return dataset_cls.from_numpy(obs=obs, array=array, weights=weights) + return dataset_cls.from_numpy(obs, array=array, weights=weights) else: - return dataset_cls(obs=obs, array=array, weights=weights) + return dataset_cls(obs, array=array, weights=weights) def get_nevents(dataset):

Asymptotic calculator histogram dimensionality When generating the asimov histogram associated with a particular test, why is the generated histogram limited to a one-dimensional fit, implicitly by the call to `space.limit1d` [here](https://github.com/scikit-hep/hepstats/blob/cf650da0408de083e27c52b952fe4ef144014de8/src/hepstats/hypotests/calculators/asymptotic_calculator.py#L34)? This seems to exclude the possibility for hypothesis tests involving a likelihood function evaluated using a PDF with dimensionality greater than 1 which, in my naive understanding, is a desirable use-case for this calculator. Thanks for all your work!

Hey, that's a good catch! Indeed, this should be working for n-dims and is not generalized. Since zfit itself offers now also to directly create binned histograms, this could be the best way to go: https://zfit-tutorials.readthedocs.io/en/latest/tutorials/components/30%20-%20Binned%20models.html#binned-models-from-histograms Or we could simply generalize it as it is now, the problem is maybe that it only takes the midpoint, which is probably often good enough but maybe creates a small bias. I suppose either way this will require some development (and a new release!) before I can use it. There is no workaround, right? Hey, as far as I remember the asymptotic method was only documented for 1D problems (I.e 1D parameter of interest) You should probably check the papers and see if there is an extension to N-dim. It's a 1D poi, but that's a different thing to requiring a 1D space for the asimov dataset. The pdf can be multidimensional while the scan is still over a 1 dimensional yield parameter, for example.

2022-08-30T10:10:21

0.0

[]

etianen/django-s3-storage

etianen__django-s3-storage-161

62c41bbf9696bc62df35590b8eb9b55468e88c4e

diff --git a/CHANGELOG.rst b/CHANGELOG.rst index 5869a4a..5165600 100644 --- a/CHANGELOG.rst +++ b/CHANGELOG.rst @@ -1,6 +1,11 @@ django-s3-storage changelog =========================== +0.15.0 +------ + +- Added support for opening files in text mode on Python 3.11+ (@etianen). + 0.14.0 ------ diff --git a/django_s3_storage/__init__.py b/django_s3_storage/__init__.py index 7f723d0..ce66fdf 100644 --- a/django_s3_storage/__init__.py +++ b/django_s3_storage/__init__.py @@ -3,4 +3,4 @@ """ -__version__ = (0, 14, 0) +__version__ = (0, 15, 0) diff --git a/django_s3_storage/storage.py b/django_s3_storage/storage.py index db7a624..a10f51a 100644 --- a/django_s3_storage/storage.py +++ b/django_s3_storage/storage.py @@ -6,15 +6,15 @@ import shutil from contextlib import closing from datetime import timezone -from functools import wraps, partial -from io import TextIOBase +from functools import partial, wraps +from io import TextIOBase, TextIOWrapper from tempfile import SpooledTemporaryFile from threading import local from urllib.parse import urljoin, urlsplit, urlunsplit import boto3 -from botocore.client import Config from boto3.s3.transfer import TransferConfig +from botocore.client import Config from botocore.exceptions import ClientError from django.conf import settings from django.contrib.staticfiles.storage import ManifestFilesMixin @@ -41,6 +41,7 @@ def _do_wrap_errors(self, name, *args, **kwargs): if code == "NoSuchKey": err_cls = FileNotFoundError raise err_cls(f"S3Storage error at {name!r}: {force_str(ex)}") + return _do_wrap_errors @@ -63,6 +64,7 @@ def do_wrap_path_impl(self, name, *args, **kwargs): # posix paths. We fix this by converting paths to system form, passing them to the default implementation, then # converting them back to posix paths. return _to_posix_path(func(self, _to_sys_path(name), *args, **kwargs)) + return do_wrap_path_impl @@ -108,18 +110,24 @@ def __init__(self, storage): if storage.settings.AWS_ACCESS_KEY_ID: connection_kwargs["aws_access_key_id"] = storage.settings.AWS_ACCESS_KEY_ID if storage.settings.AWS_SECRET_ACCESS_KEY: - connection_kwargs["aws_secret_access_key"] = storage.settings.AWS_SECRET_ACCESS_KEY + connection_kwargs[ + "aws_secret_access_key" + ] = storage.settings.AWS_SECRET_ACCESS_KEY if storage.settings.AWS_SESSION_TOKEN: connection_kwargs["aws_session_token"] = storage.settings.AWS_SESSION_TOKEN if storage.settings.AWS_S3_ENDPOINT_URL: connection_kwargs["endpoint_url"] = storage.settings.AWS_S3_ENDPOINT_URL self.session = boto3.session.Session() - self.s3_connection = self.session.client("s3", config=Config( - s3={"addressing_style": storage.settings.AWS_S3_ADDRESSING_STYLE}, - signature_version=storage.settings.AWS_S3_SIGNATURE_VERSION, - max_pool_connections=storage.settings.AWS_S3_MAX_POOL_CONNECTIONS, - connect_timeout=storage.settings.AWS_S3_CONNECT_TIMEOUT - ), **connection_kwargs) + self.s3_connection = self.session.client( + "s3", + config=Config( + s3={"addressing_style": storage.settings.AWS_S3_ADDRESSING_STYLE}, + signature_version=storage.settings.AWS_S3_SIGNATURE_VERSION, + max_pool_connections=storage.settings.AWS_S3_MAX_POOL_CONNECTIONS, + connect_timeout=storage.settings.AWS_S3_CONNECT_TIMEOUT, + ), + **connection_kwargs, + ) @deconstructible @@ -155,7 +163,7 @@ class S3Storage(Storage): "AWS_S3_FILE_OVERWRITE": False, "AWS_S3_USE_THREADS": True, "AWS_S3_MAX_POOL_CONNECTIONS": 10, - "AWS_S3_CONNECT_TIMEOUT": 60 # 60 seconds + "AWS_S3_CONNECT_TIMEOUT": 60, # 60 seconds } s3_settings_suffix = "" @@ -178,16 +186,24 @@ def _setup(self): setting_key, self._kwargs.get( setting_key.lower(), - getattr(settings, setting_key + self.s3_settings_suffix, setting_default_value), + getattr( + settings, + setting_key + self.s3_settings_suffix, + setting_default_value, + ), ), ) # Validate settings. if not self.settings.AWS_S3_BUCKET_NAME: - raise ImproperlyConfigured(f"Setting AWS_S3_BUCKET_NAME{self.s3_settings_suffix} is required.") + raise ImproperlyConfigured( + f"Setting AWS_S3_BUCKET_NAME{self.s3_settings_suffix} is required." + ) # Create a thread-local connection manager. self._connections = _Local(self) # Set transfer config for S3 operations - self._transfer_config = TransferConfig(use_threads=self.settings.AWS_S3_USE_THREADS) + self._transfer_config = TransferConfig( + use_threads=self.settings.AWS_S3_USE_THREADS + ) @property def s3_connection(self): @@ -201,8 +217,8 @@ def __init__(self, **kwargs): # Check for unknown kwargs. for kwarg_key in kwargs.keys(): if ( - kwarg_key.upper() not in self.default_auth_settings and - kwarg_key.upper() not in self.default_s3_settings + kwarg_key.upper() not in self.default_auth_settings + and kwarg_key.upper() not in self.default_s3_settings ): raise ImproperlyConfigured(f"Unknown S3Storage parameter: {kwarg_key}") # Set up the storage. @@ -211,7 +227,9 @@ def __init__(self, **kwargs): # Re-initialize the storage if an AWS setting changes. setting_changed.connect(self._setting_changed_received) # Register system checks. - checks.register(partial(self.__class__._system_checks, self), checks.Tags.security) + checks.register( + partial(self.__class__._system_checks, self), checks.Tags.security + ) # All done! super().__init__() @@ -237,7 +255,9 @@ def __reduce__(self): def _get_key_name(self, name): if name.startswith("/"): name = name[1:] - return posixpath.normpath(posixpath.join(self.settings.AWS_S3_KEY_PREFIX, _to_posix_path(name))) + return posixpath.normpath( + posixpath.join(self.settings.AWS_S3_KEY_PREFIX, _to_posix_path(name)) + ) def _object_params(self, name): params = { @@ -256,22 +276,29 @@ def _object_put_params(self, name): ), "Metadata": { key: _callable_setting(value, name) - for key, value - in self.settings.AWS_S3_METADATA.items() + for key, value in self.settings.AWS_S3_METADATA.items() }, - "StorageClass": "REDUCED_REDUNDANCY" if self.settings.AWS_S3_REDUCED_REDUNDANCY else "STANDARD", + "StorageClass": "REDUCED_REDUNDANCY" + if self.settings.AWS_S3_REDUCED_REDUNDANCY + else "STANDARD", } params.update(self._object_params(name)) # Set content disposition. - content_disposition = _callable_setting(self.settings.AWS_S3_CONTENT_DISPOSITION, name) + content_disposition = _callable_setting( + self.settings.AWS_S3_CONTENT_DISPOSITION, name + ) if content_disposition: params["ContentDisposition"] = content_disposition # Set content langauge. - content_langauge = _callable_setting(self.settings.AWS_S3_CONTENT_LANGUAGE, name) + content_langauge = _callable_setting( + self.settings.AWS_S3_CONTENT_LANGUAGE, name + ) if content_langauge: params["ContentLanguage"] = content_langauge # Set server-side encryption. - if self.settings.AWS_S3_ENCRYPT_KEY: # If this if False / None / empty then no encryption + if ( + self.settings.AWS_S3_ENCRYPT_KEY + ): # If this if False / None / empty then no encryption if isinstance(self.settings.AWS_S3_ENCRYPT_KEY, str): params["ServerSideEncryption"] = self.settings.AWS_S3_ENCRYPT_KEY if self.settings.AWS_S3_KMS_ENCRYPTION_KEY_ID: @@ -283,11 +310,11 @@ def _object_put_params(self, name): def new_temporary_file(self): """Returns a new file to use when opening from or saving to S3""" - return SpooledTemporaryFile(max_size=1024*1024*10) # 10 MB. + return SpooledTemporaryFile(max_size=1024 * 1024 * 10) # 10 MB. @_wrap_errors def _open(self, name, mode="rb"): - if mode != "rb": + if mode not in ("rb", "rt", "r"): raise ValueError("S3 files can only be opened in read-only mode") # Load the key into a temporary file. It would be nice to stream the # content, but S3 doesn't support seeking, which is sometimes needed. @@ -298,6 +325,9 @@ def _open(self, name, mode="rb"): # Un-gzip if required. if obj.get("ContentEncoding") == "gzip": content = gzip.GzipFile(name, "rb", fileobj=content) + # Decode text if required. + if "b" not in mode: + content = TextIOWrapper(content) # All done! return S3File(content, name, self) @@ -325,7 +355,12 @@ def _save(self, name, content): # Check if the content type is compressible. content_type_family, content_type_subtype = content_type.lower().split("/") content_type_subtype = content_type_subtype.split("+")[-1] - if content_type_family == "text" or content_type_subtype in ("xml", "json", "html", "javascript"): + if content_type_family == "text" or content_type_subtype in ( + "xml", + "json", + "html", + "javascript", + ): # Compress the content. temp_file = self.new_temporary_file() temp_files.append(temp_file) @@ -337,7 +372,9 @@ def _save(self, name, content): temp_file.seek(0) content = temp_file put_params["ContentEncoding"] = "gzip" - put_params["Metadata"][_UNCOMPRESSED_SIZE_META_KEY] = f"{orig_size:d}" + put_params["Metadata"][ + _UNCOMPRESSED_SIZE_META_KEY + ] = f"{orig_size:d}" else: content.seek(0) # Save the file. @@ -346,8 +383,13 @@ def _save(self, name, content): original_close = content.close content.close = lambda: None try: - self.s3_connection.upload_fileobj(content, put_params.pop('Bucket'), put_params.pop('Key'), - ExtraArgs=put_params, Config=self._transfer_config) + self.s3_connection.upload_fileobj( + content, + put_params.pop("Bucket"), + put_params.pop("Key"), + ExtraArgs=put_params, + Config=self._transfer_config, + ) finally: # Restore the original close method. content.close = original_close @@ -385,8 +427,8 @@ def delete(self, name): @_wrap_errors def copy(self, src_name, dst_name): self.s3_connection.copy_object( - CopySource=self._object_params(src_name), - **self._object_params(dst_name)) + CopySource=self._object_params(src_name), **self._object_params(dst_name) + ) @_wrap_errors def rename(self, src_name, dst_name): @@ -402,7 +444,8 @@ def exists(self, name): results = self.s3_connection.list_objects_v2( Bucket=self.settings.AWS_S3_BUCKET_NAME, MaxKeys=1, - Prefix=self._get_key_name(name) + "/", # Add the slash again, since _get_key_name removes it. + Prefix=self._get_key_name(name) + + "/", # Add the slash again, since _get_key_name removes it. ) except ClientError: return False @@ -449,7 +492,9 @@ def url(self, name, extra_params=None, client_method="get_object"): # Use a public URL, if specified. if self.settings.AWS_S3_PUBLIC_URL: if extra_params or client_method != "get_object": - raise ValueError("Use of extra_params or client_method is not allowed with AWS_S3_PUBLIC_URL") + raise ValueError( + "Use of extra_params or client_method is not allowed with AWS_S3_PUBLIC_URL" + ) return urljoin(self.settings.AWS_S3_PUBLIC_URL, filepath_to_uri(name)) # Otherwise, generate the URL. params = extra_params.copy() if extra_params else {} @@ -462,8 +507,16 @@ def url(self, name, extra_params=None, client_method="get_object"): # Strip off the query params if we're not interested in bucket auth. if not self.settings.AWS_S3_BUCKET_AUTH: if extra_params or client_method != "get_object": - raise ValueError("Use of extra_params or client_method is not allowed with AWS_S3_BUCKET_AUTH") - url = urlunsplit(urlsplit(url)[:3] + ("", "",)) + raise ValueError( + "Use of extra_params or client_method is not allowed with AWS_S3_BUCKET_AUTH" + ) + url = urlunsplit( + urlsplit(url)[:3] + + ( + "", + "", + ) + ) # All done! return url @@ -501,8 +554,9 @@ def sync_meta_iter(self): put_params["ContentEncoding"] = content_encoding if content_encoding == "gzip": try: - put_params["Metadata"][_UNCOMPRESSED_SIZE_META_KEY] = \ - obj["Metadata"][_UNCOMPRESSED_SIZE_META_KEY] + put_params["Metadata"][_UNCOMPRESSED_SIZE_META_KEY] = obj[ + "Metadata" + ][_UNCOMPRESSED_SIZE_META_KEY] except KeyError: pass # Update the metadata. @@ -513,7 +567,7 @@ def sync_meta_iter(self): "Key": self._get_key_name(name), }, MetadataDirective="REPLACE", - **put_params + **put_params, ) yield name @@ -529,23 +583,28 @@ class StaticS3Storage(S3Storage): """ default_s3_settings = S3Storage.default_s3_settings.copy() - default_s3_settings.update({ - "AWS_S3_BUCKET_AUTH": False, - }) + default_s3_settings.update( + { + "AWS_S3_BUCKET_AUTH": False, + } + ) s3_settings_suffix = "_STATIC" class ManifestStaticS3Storage(ManifestFilesMixin, StaticS3Storage): - default_s3_settings = StaticS3Storage.default_s3_settings.copy() - default_s3_settings.update({ - "AWS_S3_MAX_AGE_SECONDS_CACHED": 60 * 60 * 24 * 365, # 1 year. - }) + default_s3_settings.update( + { + "AWS_S3_MAX_AGE_SECONDS_CACHED": 60 * 60 * 24 * 365, # 1 year. + } + ) def post_process(self, *args, **kwargs): initial_aws_s3_max_age_seconds = self.settings.AWS_S3_MAX_AGE_SECONDS - self.settings.AWS_S3_MAX_AGE_SECONDS = self.settings.AWS_S3_MAX_AGE_SECONDS_CACHED + self.settings.AWS_S3_MAX_AGE_SECONDS = ( + self.settings.AWS_S3_MAX_AGE_SECONDS_CACHED + ) try: yield from super().post_process(*args, **kwargs) finally: diff --git a/pyproject.toml b/pyproject.toml new file mode 100644 index 0000000..e69de29

Cannot open files in plain text mode I am working with CSV files which need to be opened in plain text mode. With django-storage I had been using: with object.file.open('r') as f: csv_reader = csv.DictReader(f) But that gives me an error and I need to open it in binary mode and do this instead: with object.file.open() as f: csv_reader = csv.DictReader(io.StringIO(f.read().decode('utf-8')) I'm wondering if this is a bug or if this is how my code should change. (I am loving the product, I like that it seeks(0) and handles encoding properly when I am writing CSV files. It is much closer to the the native Django Storage.)

2023-11-04T15:34:56

0.0

[]

frostming/lektor-tailwind

frostming__lektor-tailwind-5

b94a92ec0aa2209dbc4e73f954e3336792376080

diff --git a/lektor_tailwind.py b/lektor_tailwind.py index 968f21d..b3bdf5d 100644 --- a/lektor_tailwind.py +++ b/lektor_tailwind.py @@ -39,17 +39,19 @@ def init_tailwindcss(self): def _run_watcher(self, output_path: str): if not self.input_exists(): return - self.tailwind = subprocess.Popen( - [ - self.tailwind_bin, - "-i", - self.input_css, - "-o", - os.path.join(output_path, self.css_path), - "-w", - ], - cwd=self.env.root_path, - ) + + cmd = [ + self.tailwind_bin, + "--input", + self.input_css, + "--output", + os.path.join(output_path, self.css_path), + "--watch", + ] + if os.environ.get("NODE_ENV") == "production": + cmd.append("--minify") + + self.tailwind = subprocess.Popen(cmd, cwd=self.env.root_path) def input_exists(self) -> bool: return os.path.exists(self.input_css)

Tailwind fails if `lektor build` run from outside the Lektor project directory Since relative paths are used to configure the `content` paths in `tailwind.config.js`, tailwind can not find the source content unless it is run from the Lektor project directory. Lektor-tailwind currently fails to chdir to the Lektor project directory before running tailwind for `lektor build`. (It does chdir before running tailwind in watch mode for `lektor server`.) This fixes that. Tailwind fails if `lektor build` run from outside the Lektor project directory Since relative paths are used to configure the `content` paths in `tailwind.config.js`, tailwind can not find the source content unless it is run from the Lektor project directory. Lektor-tailwind currently fails to chdir to the Lektor project directory before running tailwind for `lektor build`. (It does chdir before running tailwind in watch mode for `lektor server`.) This fixes that.

2023-08-09T16:57:31

0.0

[]

readthedocs/sphinx_rtd_theme

readthedocs__sphinx_rtd_theme-1576

2c19bb149bb87bd14ccc4ad2d30d9e7d35a8d925

diff --git a/.circleci/config.yml b/.circleci/config.yml index d7aa9d7ee..dfe6f3644 100644 --- a/.circleci/config.yml +++ b/.circleci/config.yml @@ -44,54 +44,51 @@ jobs: - image: 'cimg/python:3.9-node' steps: - run-build: {} - py36: - docker: - - image: 'cimg/python:3.6' - steps: - - run-tox: - version: py36 - sphinx-version: "50,51,52,53" - py37: - docker: - - image: 'cimg/python:3.7' - steps: - - run-tox: - version: py37 - sphinx-version: "50,51,52,53" py38: docker: - image: 'cimg/python:3.8' steps: - run-tox: version: py38 - sphinx-version: "50,51,52,53,60,61,62,70,71,latest" + sphinx-version: "60,61,62,70,71" py39: docker: - image: 'cimg/python:3.9' steps: - run-tox: version: py39 - sphinx-version: "50,51,52,53,60,61,62,70,71,72,latest" + sphinx-version: "60,61,62,70,71,72,73,74" py310: docker: - image: 'cimg/python:3.10' steps: - run-tox: version: py310 - sphinx-version: "50,51,52,53,60,61,62,70,71,72,latest" + sphinx-version: "60,61,62,70,71,72,73,74,80,latest,dev" py311: docker: - image: 'cimg/python:3.11' steps: - run-tox: version: py311 - sphinx-version: "53,60,61,62,70,71,72,latest,dev" + sphinx-version: "72,73,74,80,latest,dev" + + py312: + docker: + - image: 'cimg/python:3.12' + steps: + - run-tox: + version: py312 + sphinx-version: "72,73,74,80,latest,dev" workflows: version: 2 tests: jobs: - build + - py312: + requires: + - build - py311: requires: - build @@ -104,9 +101,3 @@ workflows: - py38: requires: - build - - py37: - requires: - - build - - py36: - requires: - - build diff --git a/docs/conf.py b/docs/conf.py index 0c3731fc9..25d002c02 100644 --- a/docs/conf.py +++ b/docs/conf.py @@ -68,10 +68,6 @@ range(1, 100) )) -if 'READTHEDOCS' in os.environ: - html_context['READTHEDOCS'] = True - html_context['READTHEDOCS_PROJECT'] = os.environ.get("READTHEDOCS_PROJECT") - html_logo = "demo/static/logo-wordmark-light.svg" html_show_sourcelink = True html_favicon = "demo/static/favicon.ico" diff --git a/docs/configuring.rst b/docs/configuring.rst index e76c2aeda..3718abc27 100644 --- a/docs/configuring.rst +++ b/docs/configuring.rst @@ -102,7 +102,11 @@ Miscellaneous options :type: string + .. deprecated:: 3.0.0 + The ``analytics_id`` option is deprecated, use the sphinxcontrib-googleanalytics_ extension instead. + .. _gtag.js: https://developers.google.com/gtagjs + .. _sphinxcontrib-googleanalytics: https://pypi.org/project/sphinxcontrib-googleanalytics/ .. confval:: analytics_anonymize_ip @@ -111,6 +115,9 @@ Miscellaneous options :type: boolean :default: ``False`` + .. deprecated:: 3.0.0 + The ``analytics_anonymize_ip`` option is deprecated, use the sphinxcontrib-googleanalytics_ extension instead. + .. confval:: canonical_url This will specify a `canonical URL`_ meta link element to tell search diff --git a/docs/contributing.rst b/docs/contributing.rst index 541a1c264..e8c733d7d 100644 --- a/docs/contributing.rst +++ b/docs/contributing.rst @@ -86,7 +86,7 @@ Use the following steps: # Runs the development webserver $ docker-compose run sphinx_rtd_theme dev - + # If you want to copy stuff out of the Docker environment, run this make # target or read the actual Makefile to see what is going on. # We suggest running this command every time that you want to quickly build @@ -108,13 +108,13 @@ The following cases need to be tested with changes to CSS or JavaScript: We support large, tablet, and mobile viewport sizes * We currently support only HTML5 writer. * Multiple major versions of Sphinx should be tested. - We currently support Sphinx ``>=5.0`` + We currently support Sphinx ``>=6.0`` It's easiest to test combinations of dependency versions using ``tox``: .. code:: console - % tox -e py310-sphinx62 + % tox -e py312-sphinx74 If the tests and build are successful, you can view the built documentation at the directory noted by Sphinx: @@ -123,9 +123,9 @@ the directory noted by Sphinx: build succeeded, 10 warnings. - The HTML pages are in .tox/py310-sphinx62/tmp/html. + The HTML pages are in .tox/py312-sphinx74/tmp/html. ___________________________ summary ___________________________ - py310-sphinx62: commands succeeded + py312-sphinx74: commands succeeded congratulations :) You can then open up this path with a series of browsers to test. @@ -135,12 +135,12 @@ multiple ``tox`` environments, and open both up for comparison: .. code:: console - % tox -e py310-sphinx62 + % tox -e py312-sphinx62 ... - % tox -e py310-sphinx53 + % tox -e py312-sphinx74 ... - % firefox .tox/py310-sphinx62/tmp/html/index.html - % firefox .tox/py310-sphinx53/tmp/html/index.html + % firefox .tox/py312-sphinx62/tmp/html/index.html + % firefox .tox/py312-sphinx74/tmp/html/index.html You can also use a separate ``tox`` environment for building output to compare against. All of the ``tox`` environments have an additional postfix, ``-qa``, to @@ -167,11 +167,11 @@ Currently, the most important environments to QA are: the cases above, we'll be here all day. Python 3, and latest minor of each major Sphinx release should give us enough work. -* ``py310-sphinx53`` -* ``py310-sphinx62`` -* ``py310-sphinx72`` -* ``py310-sphinxlatest`` -* ``py310-sphinxdev`` +* ``py312-sphinx62`` +* ``py312-sphinx74`` +* ``py312-sphinx80`` +* ``py312-sphinxlatest`` +* ``py312-sphinxdev`` Translations ============ diff --git a/docs/development.rst b/docs/development.rst index c4ee9a43c..298fbe2b2 100644 --- a/docs/development.rst +++ b/docs/development.rst @@ -36,13 +36,8 @@ combinations. Unsupported browser/operating system combinations include: -Internet Explorer (any OS, versions <=10) - **Unsupported.** IE11 is the last partially supported version. We do no - testing on prior versions. - -Internet Explorer (any OS, version 11) - We currently only partially support IE11, and only test for major bugs. - Support will be removed in the :ref:`roadmap-release-3.0.0` release. +Internet Explorer (any OS, any version) + Support was removed in the :ref:`roadmap-release-3.0.0` release. Opera (any OS, any version) **Community support only.** We do not receive enough traffic with this @@ -62,11 +57,14 @@ The theme officially supports the following dependencies in your Sphinx project: * - Dependency - Versions * - Python - - 3.6 or greater + - 3.8 or greater * - Sphinx - - 5 or greater + - 6 or greater * - docutils - - >= 0.14, < 0.19 + - > 0.18, < 0.22 + +.. deprecated:: 3.0 + Sphinx < 6 support removed .. deprecated:: 2.0 Sphinx < 5 support removed @@ -179,10 +177,10 @@ HTML4 support will be removed 3.0.0 ~~~~~ -:Planned release date: 2024 Q1 +:Planned release date: 2024 Q3 -This release is not yet fully planned. -However, we already know that we will be dropping support for older version of different dependencies and browsers. +This release will drop support for old Python and Sphinx versions, +and will add support for new ones. Sphinx 5, Python < 3.8 and docutils < 0.18 support will be removed Sphinx 5 is the latest version that supports Python 3.6 and 3.7, diff --git a/setup.cfg b/setup.cfg index b1ef6084e..f7dfdb972 100644 --- a/setup.cfg +++ b/setup.cfg @@ -29,12 +29,11 @@ classifiers = Intended Audience :: Developers Programming Language :: Python :: 3 Programming Language :: Python :: 3 :: Only - Programming Language :: Python :: 3.6 - Programming Language :: Python :: 3.7 Programming Language :: Python :: 3.8 Programming Language :: Python :: 3.9 Programming Language :: Python :: 3.10 Programming Language :: Python :: 3.11 + Programming Language :: Python :: 3.12 Operating System :: OS Independent Topic :: Documentation Topic :: Software Development :: Documentation @@ -43,10 +42,10 @@ classifiers = include_package_data = True zip_safe = False packages = sphinx_rtd_theme -python_requires = >=3.6 +python_requires = >=3.8 install_requires = - sphinx >=5,<8 - docutils <0.21 + sphinx >=6,<9 + docutils >0.18,<0.22 sphinxcontrib-jquery >=4,<5 tests_require = pytest @@ -54,7 +53,6 @@ tests_require = [options.extras_require] dev = transifex-client - sphinxcontrib-httpdomain bump2version wheel twine diff --git a/sphinx_rtd_theme/__init__.py b/sphinx_rtd_theme/__init__.py index 59be0645d..2cb8e6ad5 100644 --- a/sphinx_rtd_theme/__init__.py +++ b/sphinx_rtd_theme/__init__.py @@ -21,6 +21,10 @@ def get_html_theme_path(): """Return list of HTML theme paths.""" + logger.warning( + _('Calling get_html_theme_path is deprecated. If you are calling it to define html_theme_path, you are safe to remove that code.') + ) + cur_dir = path.abspath(path.dirname(path.dirname(__file__))) return cur_dir @@ -32,6 +36,21 @@ def config_initiated(app, config): _('The canonical_url option is deprecated, use the html_baseurl option from Sphinx instead.') ) + if theme_options.get("analytics_id"): + logger.warning( + _('The analytics_id option is deprecated, use the sphinxcontrib-googleanalytics extension instead.') + ) + + if theme_options.get("analytics_anonymize_ip"): + logger.warning( + _('The analytics_anonymize_ip option is deprecated, use the sphinxcontrib-googleanalytics extension instead.') + ) + + if "extra_css_files" in config.html_context: + logger.warning( + _('The extra_css_file option is deprecated, use the html_css_files option from Sphinx instead.') + ) + def extend_html_context(app, pagename, templatename, context, doctree): # Add ``sphinx_version_info`` tuple for use in Jinja templates @@ -52,7 +71,7 @@ def setup(app): if python_version[0] < 3: logger.error("Python 2 is not supported with sphinx_rtd_theme, update to Python 3.") - app.require_sphinx('5.0') + app.require_sphinx('6.0') if app.config.html4_writer: logger.error("'html4_writer' is not supported with sphinx_rtd_theme.") diff --git a/tox.ini b/tox.ini index d601e5073..a010796ac 100644 --- a/tox.ini +++ b/tox.ini @@ -1,28 +1,22 @@ [tox] envlist = - py{36,37,38,39,310}-sphinx{50,51,52,53}{-qa} py{38,39,310}-sphinx{60,61,62,70,71}{-qa} - py{39,310}-sphinx{72,latest,dev}{-qa} - # Python 3.11 working from Sphinx 5.3 and up - py{311}-sphinx{53,60,61,62,70,71,72,latest,dev}{-qa} + py{39,310,311,312}-sphinx{72,73,74,80}{-qa} + py{310,311,312}-sphinx{latest,dev}{-qa} [testenv] setenv = LANG=C deps = . - readthedocs-sphinx-ext pytest - sphinx50: Sphinx>=5.0,<5.1 - sphinx51: Sphinx>=5.1,<5.2 - sphinx52: Sphinx>=5.2,<5.3 - sphinx53: Sphinx>=5.3,<5.4 sphinx60: Sphinx>=6.0,<6.1 sphinx61: Sphinx>=6.1,<6.2 sphinx62: Sphinx>=6.2,<6.3 sphinx70: Sphinx>=7.0,<7.1 sphinx71: Sphinx>=7.1,<7.2 sphinx72: Sphinx>=7.2,<7.3 + sphinx80: Sphinx>=8.0,<8.1 sphinxlatest: Sphinx sphinxdev: https://github.com/sphinx-doc/sphinx/archive/refs/heads/master.zip allowlist_externals =

How is sphinxcontrib-httpdomain used in the project? ### Problem In `setup.cfg` there is a `dev` extra defined with `sphinxcontrib-httpdomain` as one of the dependencies. https://github.com/readthedocs/sphinx_rtd_theme/blob/8ce23cec96f628ac0d29737bf1cf8cc2e750f068/setup.cfg#L57 I can't find it being used anywhere in the project. Is it needed?

2024-07-24T09:37:19

0.0

[]

twopirllc/pandas-ta

twopirllc__pandas-ta-801

23bca1aff2efde5e4fd1215b20416fda616750f7

diff --git a/pandas_ta/utils/_stats.py b/pandas_ta/utils/_stats.py index cffa6ac4..0a85ef32 100644 --- a/pandas_ta/utils/_stats.py +++ b/pandas_ta/utils/_stats.py @@ -1,5 +1,5 @@ # -*- coding: utf-8 -*- -from numpy import array, infty, log, nan, pi, sqrt +from numpy import array, inf, log, nan, pi, sqrt from pandas_ta._typing import Array, IntFloat, Number, Union from pandas_ta.maps import Imports from pandas_ta.utils import hpoly

Pandas-ta doesnt work with numpy 2.0 So pandas-ta doesnt work with the newest numpy 2.0. Reproducible example: ```import pandas_ta as ta``` Log: ``` ImportError: cannot import name 'NaN' from 'numpy' (C:\Users\User\anaconda3\envs\py312Test\Lib\site-packages\numpy\__init__.py) ``` Would be nice to get support for the latest numpy version. Pandas-ta version 0.3.14b Python: 3.12.2

> So pandas-ta doesnt work with the newest numpy 2.0. That's crazy @Chuck321123! How is that possible? 🤷🏼‍♂️ It was released.... yesterday! 😮 Please inform the following packages that they are not up to date with the newest _numpy 2.0_. I don't want them to miss out on the action. <img width="820" alt="Screenshot 2024-06-17 at 12 13 35 PM" src="https://github.com/twopirllc/pandas-ta/assets/18198392/df73773e-3451-42e4-b613-2f41d3bd4df9"> **Note:** _streamlit_ is not necessary for *pandas-ta* > Would be nice to get support for the latest numpy version. Let's do it! ✅ Fork the repo, checkout the _development_ branch, edit (including tests), and submit a PR. 😎 FYI, I'm now running: Python: 3.11.9 & 3.12.4, no longer supporting <=3.9 Pandas TA (_development_) 0.4.17b @twopirllc Unfortunately, I don't know how to do that. However, If anyone knows how to do it I would be very happy I have created the pull request. Have tested with server so it should work. Though haven't touched the package requirements though should also be compatible with 1.22. Kindly check

2024-06-18T19:15:39

0.0

[]

beancount/beancount

beancount__beancount-867

9d2ff2817f803cd3b6efe0755498884e074acda1

diff --git a/beancount/core/account.py b/beancount/core/account.py index 673a1ea6f..f7494ddc6 100644 --- a/beancount/core/account.py +++ b/beancount/core/account.py @@ -44,6 +44,32 @@ TYPE = "<AccountDummy>" +def is_valid_root(string: Account) -> bool: + """Return true if the given string is a valid root account name. + This just check for valid syntax. + + Args: + string: A string, to be checked for account name pattern. + Returns: + A boolean, true if the string has the form of a root account's name. + """ + return isinstance(string, str) and bool(regex.fullmatch(ACC_COMP_TYPE_RE, string)) + + +def is_valid_leaf(string: Account) -> bool: + """Return true if the given string is a valid leaf account name. + This just check for valid syntax. + + Args: + string: A string, to be checked for account name pattern. + Returns: + A boolean, true if the string has the form of a leaf account's name. + """ + return isinstance(string, str) and all( + regex.fullmatch(ACC_COMP_NAME_RE, p) for p in string.split(":") + ) + + def is_valid(string: Account) -> bool: """Return true if the given string is a valid account name. This does not check for the root account types, just the general syntax. @@ -53,7 +79,7 @@ def is_valid(string: Account) -> bool: Returns: A boolean, true if the string has the form of an account's name. """ - return isinstance(string, str) and bool(regex.match("{}$".format(ACCOUNT_RE), string)) + return isinstance(string, str) and bool(regex.fullmatch(ACCOUNT_RE, string)) def join(*components: Tuple[str]) -> Account: diff --git a/beancount/parser/options.py b/beancount/parser/options.py index 33f5609f9..598f40848 100644 --- a/beancount/parser/options.py +++ b/beancount/parser/options.py @@ -114,6 +114,36 @@ def options_validate_booking_method(value): raise ValueError(str(exc)) from exc +def options_validate_root_account(value): + """Validate a root account name. + + Args: + value: A string, the value provided as option. + Returns: + The root account name, if the validation succedes. + Raises: + ValueError: If the value is invalid. + """ + if not account.is_valid_root(value): + raise ValueError(f"Invalid root account name: {value!r}") + return value + + +def options_validate_leaf_account(value): + """Validate a leaf account name. + + Args: + value: A string, the value provided as option. + Returns: + The leaf account name, if the validation succedes. + Raises: + ValueError: If the value is invalid. + """ + if not account.is_valid_leaf(value): + raise ValueError(f"Invalid leaf account name: {value!r}") + return value + + # List of option groups, with their description, option names and default # values. OptGroup = collections.namedtuple("OptGroup", "description options") @@ -269,11 +299,15 @@ def Opt( recognizes account names. """, [ - Opt("name_assets", _TYPES.assets), - Opt("name_liabilities", _TYPES.liabilities), - Opt("name_equity", _TYPES.equity), - Opt("name_income", _TYPES.income), - Opt("name_expenses", _TYPES.expenses), + Opt("name_assets", _TYPES.assets, converter=options_validate_root_account), + Opt( + "name_liabilities", + _TYPES.liabilities, + converter=options_validate_root_account, + ), + Opt("name_equity", _TYPES.equity, converter=options_validate_root_account), + Opt("name_income", _TYPES.income, converter=options_validate_root_account), + Opt("name_expenses", _TYPES.expenses, converter=options_validate_root_account), ], ), OptGroup( @@ -281,7 +315,13 @@ def Opt( Leaf name of the equity account used for summarizing previous transactions into opening balances. """, - [Opt("account_previous_balances", "Opening-Balances")], + [ + Opt( + "account_previous_balances", + "Opening-Balances", + converter=options_validate_leaf_account, + ) + ], ), OptGroup( """ @@ -289,7 +329,13 @@ def Opt( earnings from income and expenses accrued before the beginning of the exercise into the balance sheet. """, - [Opt("account_previous_earnings", "Earnings:Previous")], + [ + Opt( + "account_previous_earnings", + "Earnings:Previous", + converter=options_validate_leaf_account, + ) + ], ), OptGroup( """ @@ -298,7 +344,13 @@ def Opt( will essentially "fixup" the basic accounting equation due to the errors that priced conversions introduce. """, - [Opt("account_previous_conversions", "Conversions:Previous")], + [ + Opt( + "account_previous_conversions", + "Conversions:Previous", + converter=options_validate_leaf_account, + ) + ], ), OptGroup( """ @@ -306,7 +358,13 @@ def Opt( earnings from income and expenses accrued during the current exercise into the balance sheet. This is most often called "Net Income". """, - [Opt("account_current_earnings", "Earnings:Current")], + [ + Opt( + "account_current_earnings", + "Earnings:Current", + converter=options_validate_leaf_account, + ) + ], ), # This common option can be used by reporting systems and is ignored by # Beancount itself. @@ -315,7 +373,13 @@ def Opt( Leaf name of the equity account used for inserting conversions that will zero out remaining amounts due to transfers during the exercise period. """, - [Opt("account_current_conversions", "Conversions:Current")], + [ + Opt( + "account_current_conversions", + "Conversions:Current", + converter=options_validate_leaf_account, + ) + ], ), # This common option can be used by reporting systems and is ignored by # Beancount itself. @@ -328,7 +392,14 @@ def Opt( balance on the sheet. This has no effect on behavior, other than providing a configurable account name for such postings to occur. """, - [Opt("account_unrealized_gains", "Earnings:Unrealized", "Earnings:Unrealized")], + [ + Opt( + "account_unrealized_gains", + "Earnings:Unrealized", + "Earnings:Unrealized", + converter=options_validate_leaf_account, + ) + ], ), OptGroup( """ @@ -337,7 +408,14 @@ def Opt( an account name will automatically enable the addition of postings on all transactions that have a residual amount. """, - [Opt("account_rounding", None, "Rounding")], + [ + Opt( + "account_rounding", + None, + "Rounding", + converter=options_validate_leaf_account, + ) + ], ), OptGroup( """

exception in any query using get_account_sort_key when localizing account names I noticed the "holdings" tab in my fava only shows me the exception quoted below, clicking on "Query" and boiling down a minimal working example blames `get_account_sort_key` in conjunction with renamed (top-level) accounts. exception: ~~~pytb $ bean-query mwe.beancount 'SELECT account, account_sortkey(account) ' Traceback (most recent call last): File "/usr/lib/python3.10/cmd.py", line 214, in onecmd func = getattr(self, 'do_' + cmd) AttributeError: 'BQLShell' object has no attribute 'do_SELECT' During handling of the above exception, another exception occurred: Traceback (most recent call last): File "/usr/lib/python3.10/site-packages/beancount/query/shell.py", line 271, in run_parser self.dispatch(statement) File "/usr/lib/python3.10/site-packages/beancount/query/shell.py", line 251, in dispatch return method(statement) File "/usr/lib/python3.10/site-packages/beancount/query/shell.py", line 416, in on_Select rtypes, rrows = query_execute.execute_query(c_query, File "/usr/lib/python3.10/site-packages/beancount/query/query_execute.py", line 282, in execute_query values = [c_expr(context) for c_expr in c_target_exprs] File "/usr/lib/python3.10/site-packages/beancount/query/query_execute.py", line 282, in <listcomp> values = [c_expr(context) for c_expr in c_target_exprs] File "/usr/lib/python3.10/site-packages/beancount/query/query_env.py", line 421, in __call__ index, name = account_types.get_account_sort_key(context.account_types, args[0]) File "/usr/lib/python3.10/site-packages/beancount/core/account_types.py", line 46, in get_account_sort_key return (account_types.index(get_account_type(account_name)), account_name) ValueError: tuple.index(x): x not in tuple ~~~ MWE: ~~~ledger option "title" "MWE" option "operating_currency" "EUR" 1992-01-01 custom "fava-option" "language" "de" * Bilanzkonten option "name_assets" "Aktiva" option "name_liabilities" "Verbindlichkeiten" option "name_equity" "Passiva:Eigenkapital" option "account_previous_balances" "Passiva:Eigenkapital:Eröffnungssaldo" option "account_current_earnings" "Gewinn" ; net income option "account_previous_earnings" "Rücklagen" ; retained earnings * Erfolgskonten option "name_income" "EK:Einnahmen" option "name_expenses" "EK:Ausgaben" 1992-01-01 open Passiva:Eigenkapital:Eröffnungssaldo 1992-01-01 open Aktiva:Giro 2000-01-01 * "" Aktiva:Giro 1000 EUR Passiva:Eigenkapital:Eröffnungssaldo ~~~

The issue is not the renamed root accounts, but the fact that the account names assigned to the root accounts are leaf accounts. Namely, these options are invalid: ``` option "name_equity" "Passiva:Eigenkapital" option "name_income" "EK:Einnahmen" option "name_expenses" "EK:Ausgaben" ``` There cannot be an account name separator `:` in the names of these accounts. It seems that the validator for these options does not catch this error. +1 We should probably detect that and issue an appropriate error. (Keep this ticket open until we do)

2024-11-09T16:19:42

0.0

[]

beancount/beancount

beancount__beancount-836

c1259dca464cbf15f18b762f5e1933627217f10a

diff --git a/beancount/ops/balance.py b/beancount/ops/balance.py index 5287c9307..7c77ae6d0 100644 --- a/beancount/ops/balance.py +++ b/beancount/ops/balance.py @@ -108,7 +108,6 @@ def check(entries, options_map): # Check that the currency of the balance check is one of the allowed # currencies for that account. expected_amount = entry.amount - open = None try: open, _ = open_close_map[entry.account] except KeyError: @@ -119,6 +118,7 @@ def check(entries, options_map): entry, ) ) + continue if ( expected_amount is not None

Bug fix: keep entries when doing beancount.ops.balance check The check op was removing balance entries in the case where the account does not exist. I think it is better to keep the entry and it to fail validation, this makes it more consistent with other entry types.

2024-07-10T02:16:47

0.0

[]

beancount/beancount

beancount__beancount-835

6bd2e7f37be6b5012634964e6a8093da5bce624f

diff --git a/beancount/ops/balance.py b/beancount/ops/balance.py index 7c77ae6d0..5287c9307 100644 --- a/beancount/ops/balance.py +++ b/beancount/ops/balance.py @@ -108,6 +108,7 @@ def check(entries, options_map): # Check that the currency of the balance check is one of the allowed # currencies for that account. expected_amount = entry.amount + open = None try: open, _ = open_close_map[entry.account] except KeyError: @@ -118,7 +119,6 @@ def check(entries, options_map): entry, ) ) - continue if ( expected_amount is not None

Bug fix: keep entries when doing beancount.ops.balance check The check op was removing balance entries in the case where the account does not exist. I think it is better to keep the entry and it to fail validation, this makes it more consistent with other entry types.

2024-07-05T20:14:23

0.0

[]

SPF-OST/pytrnsys

SPF-OST__pytrnsys-117

677eb7afec5c8481ab49bd7c0160cb06d279a98a

diff --git a/pytrnsys/ddck/replaceVariables.py b/pytrnsys/ddck/replaceVariables.py index 939a0654..a57a340e 100644 --- a/pytrnsys/ddck/replaceVariables.py +++ b/pytrnsys/ddck/replaceVariables.py @@ -57,8 +57,7 @@ def replaceComputedVariablesWithDefaults(inputDdckFilePath: _pl.Path) -> _res.Re return _res.error(result) computedVariables = _res.value(result) - inputDdckContent = inputDdckFilePath.read_text() # pylint: disable=bad-option-value,unspecified-encoding - + inputDdckContent = inputDdckFilePath.read_text(encoding="windows-1252") # pylint: disable=bad-option-value outputDdckContent = inputDdckContent offset = 0 for computedVariable in computedVariables: diff --git a/pytrnsys/rsim/runParallelTrnsys.py b/pytrnsys/rsim/runParallelTrnsys.py index c114b68f..d0edca87 100644 --- a/pytrnsys/rsim/runParallelTrnsys.py +++ b/pytrnsys/rsim/runParallelTrnsys.py @@ -1,7 +1,7 @@ # pylint: skip-file # type: ignore -#!/usr/bin/python +# !/usr/bin/python """ Author : Dani Carbonell Date : 30.09.2016 @@ -12,7 +12,6 @@ import json import os import shutil -import typing as _tp from copy import deepcopy import numpy as num @@ -147,8 +146,6 @@ def readFromFolder(self, pathRun): return returnNames - # self.runFromNames(fileNames) - def readCasesToRun(self, pathRun, nameFileWithCasesToRun): fileToRunWithPath = os.path.join(pathRun, nameFileWithCasesToRun) @@ -172,7 +169,6 @@ def runFromNames(self, path, fileNames): tests.append(exeTrnsys.ExecuteTrnsys(path, fileNames[i])) tests[i].setTrnsysExePath(self.inputs["trnsysExePath"]) - # tests[i].setAddBuildingData(self.inputs["copyBuildingData"]) I comment it until we use again a case where we need this functionality. I guess we dont need this anymore. tests[i].loadDeck(check=self.inputs["checkDeck"]) tests[i].changeParameter(self.parameters) @@ -180,9 +176,6 @@ def runFromNames(self, path, fileNames): tests[i].ignoreOnlinePlotter() tests[i].setRemovePopUpWindow(self.inputs["removePopUpWindow"]) - # tests[i].copyFilesForRunning() - - # tests[i].setTrnsysVersion("TRNSYS17_EXE") self.cmds.append(tests[i].getExecuteTrnsys(self.inputs, useDeckName=tests[i].nameDck)) @@ -259,13 +252,6 @@ def createDecksFromVariant(self, fitParameters={}): parameters = self.parameters parameters.update(fitParameters) - # if (self.inputs["combineAllCases"]): - # sizeUsed = len(variations[0]) - # for var in variations: - # if (len(var) != sizeUsed): - # print "sizeUsed:%d var:%s size:%d" % (sizeUsed, var, len(var)) - # raise ValueError("FATAL ERROR. variations must be of the same size if combineAllCases = True") - myDeckGenerator = createDeck.CreateTrnsysDeck(self.path, self.nameBase, variations) myDeckGenerator.combineAllCases = self.inputs["combineAllCases"] @@ -288,7 +274,7 @@ def createDecksFromVariant(self, fitParameters={}): else: self.logger.info("Master file does not exist, no runs will be excluded") - # creates a list of decks with the appripiate name but nothing changed inside!! + # creates a list of decks with the appropriate name but nothing changed inside!! if self.variationsUsed or (self.changeDDckFilesUsed == True and self.foldersForDDckVariationUsed == False): if successfulCases: fileName = myDeckGenerator.generateDecks(successfulCases=successfulCases) @@ -303,7 +289,6 @@ def createDecksFromVariant(self, fitParameters={}): return tests = [] - cmds = [] variablePath = self.path @@ -311,22 +296,19 @@ def createDecksFromVariant(self, fitParameters={}): self.logger.debug("name to run :%s" % fileName[i]) - # if useLocationStructure: - # variablePath = os.path.join(path,location) #assign subfolder for path + # if useLocationStructure: + # variablePath = os.path.join(path,location) #assign subfolder for path - # # Parameters changed by variation + # Parameters changed by variation localCopyPar = dict.copy(parameters) # if self.variationsUsed: myParameters = myDeckGenerator.getParameters(i) localCopyPar.update(myParameters) - # # We add to the global parameters that also need to be modified - # If we assign like localCopyPar = parameters, then the parameters will change with localCopyPar !! + # We add to the global parameters that also need to be modified + # If we assign like localCopyPar = parameters, then the parameters will change with localCopyPar !! # Otherwise we change the global parameter and some values of last variation will remain. - # newPath = path + fileName[i] - # newFileDck = newPath+"\\"+fileName[i]+".dck" - # print "path:%s fileName:%s newPath:%s newFileDeck:%s"%(path,fileName[i],newPath,newFileDck) tests.append(exeTrnsys.ExecuteTrnsys(variablePath, fileName[i])) @@ -334,8 +316,6 @@ def createDecksFromVariant(self, fitParameters={}): tests[i].setRemovePopUpWindow(self.inputs["removePopUpWindow"]) - # tests[i].setTrnsysVersion("TRNSYS17_EXE") - tests[i].moveFileFromSource() tests[i].loadDeck(useDeckOutputPath=True) @@ -345,23 +325,12 @@ def createDecksFromVariant(self, fitParameters={}): if self.inputs["ignoreOnlinePlotter"] == True: tests[i].ignoreOnlinePlotter() - # ============================================================================== - # RESIZE PARAMETERS PIPE DIAMETER - # ============================================================================== - # tests[i].resizeParameters() tests[i].cleanAndCreateResultsTempFolder() tests[i].moveFileFromSource() if self.inputs["runCases"] == True: - test = os.path.split(tests[i].nameDck)[-1] self.cmds.append(tests[i].getExecuteTrnsys(self.inputs)) - # self.cmds = cmds - - # def checkTempFolderForFinishedSimulation(self,basePath): - # for file in os.listdir(os.path.join(basePath,'temp')): - # if file.endswith('.Prt') - def createLocationFolders(path, locations): for location in locations: if not os.path.exists(location): @@ -409,7 +378,7 @@ def buildTrnsysDeck(self) -> _res.Result[str]: deck.automaticEnegyBalanceStaff() deck.writeDeck() # Deck rewritten with added printer - # deck.addEnergyBalance() + deck.analyseDck() return deck.nameDeck @@ -429,7 +398,6 @@ def addParametricVariations(self, variations): """ if self.inputs["combineAllCases"] == True: - labels = [] values = [] for i, row in enumerate(variations): @@ -441,7 +409,6 @@ def addParametricVariations(self, variations): self.variablesOutput = result.tolist() else: - nVariations = len(variations) sizeOneVariation = len(variations[0]) - 2 for n in range(len(variations)): sizeCase = len(variations[n]) - 2 @@ -499,9 +466,6 @@ def writeRunLogFile(self): logfile.write(line + "\n") logfile.close() - # mainFile = sys.argv[0] - # shutil.copy(mainFile,os.path.join(self.path,'runMainFile.py')) - def readConfig(self, path, name, parseFileCreated=False): """ @@ -563,10 +527,9 @@ def changeDDckFile(self, source, end): nCharacters = len(source) for i in range(len(self.listDdck)): - # self.lines[i].replace(source,end) mySource = self.listDdck[i][ - -nCharacters: - ] # I read only the last characters with the same size as the end file + -nCharacters: + ] # I read only the last characters with the same size as the end file if mySource == source: newDDck = self.listDdck[i][0:-nCharacters] + end self.dictDdckPaths[newDDck] = self.dictDdckPaths[self.listDdck[i]] @@ -608,7 +571,6 @@ def getConfig(self): for line in self.lines: splitLine = line.split() - if splitLine[0] == "variation": variation = [] for i in range(len(splitLine)): @@ -625,7 +587,6 @@ def getConfig(self): self.variation.append(variation) elif splitLine[0] == "deck": - if splitLine[2] == "string": self.parameters[splitLine[1]] = splitLine[3] else: @@ -635,9 +596,7 @@ def getConfig(self): self.parameters[splitLine[1]] = splitLine[2] elif splitLine[0] == "replace": - splitString = line.split('$"') - oldString = splitString[1].split('"')[0] newString = splitString[2].split('"')[0] @@ -657,16 +616,6 @@ def getConfig(self): for i in range(len(splitLine)): if i > 0: self.foldersForDDckVariation.append(splitLine[i]) - - # elif (splitLine[0] == "") - # elif(splitLine[0] == "Relative"): - # - # self.listDdck.append(os.path.join(self.path, splitLine[1])) - # - # elif(splitLine[0] == "Absolute"): - # - # self.listDdck.append(splitLine[1]) - elif splitLine[0] == "fit": self.listFit[splitLine[1]] = [splitLine[2], splitLine[3], splitLine[4]] elif splitLine[0] == "case": @@ -680,7 +629,6 @@ def getConfig(self): self.listDdckPaths.add(self.inputs[splitLine[0]]) self.dictDdckPaths[fullPath] = self.inputs[splitLine[0]] else: - pass if len(self.variation) > 0: @@ -702,8 +650,6 @@ def getConfig(self): def copyConfigFile(self, configPath, configName): configFile = os.path.join(configPath, configName) - # dstPath = os.path.join(self.inputs["pathRef"],self.inputs["addResultsFolder"],self.inputs["nameRef"],configName) - # dstPath = os.path.join(self.inputs["pathRef"],self.inputs["addResultsFolder"],configName) dstPath = os.path.join(configPath, self.inputs["addResultsFolder"], configName) shutil.copyfile(configFile, dstPath) self.logger.debug("copied config file to: %s" % dstPath) @@ -737,17 +683,15 @@ def scaleVariables(self, reference, source, sink): for i in range(2, len(self.variablesOutput[j]), 1): if self.variablesOutput[j][1] == "sizeHpUsed": self.variablesOutput[j][i] = ( - str(round(self.unscaledVariables[j][i], 3)) + "*" + str(round(loadHPsize, 3)) + str(round(self.unscaledVariables[j][i], 3)) + "*" + str(round(loadHPsize, 3)) ) else: self.variablesOutput[j][i] = ( - str(round(self.unscaledVariables[j][i], 3)) + "*" + str(round(loadDemand, 3)) + str(round(self.unscaledVariables[j][i], 3)) + "*" + str(round(loadDemand, 3)) ) def run(): - pathBase = "" - if len(sys.argv) > 1: pathBase, configFile = os.path.split(sys.argv[1]) else: diff --git a/pytrnsys/trnsys_util/buildTrnsysDeck.py b/pytrnsys/trnsys_util/buildTrnsysDeck.py index daa73509..a70c3a8d 100644 --- a/pytrnsys/trnsys_util/buildTrnsysDeck.py +++ b/pytrnsys/trnsys_util/buildTrnsysDeck.py @@ -6,6 +6,7 @@ import logging import os import pathlib as _pl +import re as _re import tkinter as tk import typing as _tp from tkinter import messagebox as tkMessageBox @@ -31,7 +32,8 @@ def __init__(self, _pathDeck, _nameDeck, _nameList, ddckPlaceHolderValuesJsonPat self.pathDeck = _pathDeck self.nameDeck = self.pathDeck + "\%s.dck" % _nameDeck - self._ddckPlaceHolderValuesJsonPath = _pl.Path(ddckPlaceHolderValuesJsonPath) if ddckPlaceHolderValuesJsonPath else None + self._ddckPlaceHolderValuesJsonPath = _pl.Path( + ddckPlaceHolderValuesJsonPath) if ddckPlaceHolderValuesJsonPath else None self.oneSheetList = [] self.nameList = _nameList @@ -77,7 +79,8 @@ def _replacePlaceholdersAndGetContent(self, ddckFilePath: _pl.Path) -> _res.Resu if self._ddckPlaceHolderValuesJsonPath: if not self._ddckPlaceHolderValuesJsonPath.is_file(): - return _res.Error(f"The ddck placeholder values file at {self._ddckPlaceHolderValuesJsonPath} does not exist.") + return _res.Error( + f"The ddck placeholder values file at {self._ddckPlaceHolderValuesJsonPath} does not exist.") placeholderValues = _json.loads(self._ddckPlaceHolderValuesJsonPath.read_text()) @@ -163,7 +166,7 @@ def readDeckList( addedLines = firstThreeLines + self.linesChanged caption = ( - " **********************************************************************\n ** %s.ddck from %s \n **********************************************************************\n" + "**********************************************************************\n** %s.ddck from %s \n**********************************************************************\n" % (nameList, pathList) ) @@ -175,8 +178,11 @@ def readDeckList( unitModifiedLines = [line.replace("£", "") for line in addedLines] addedLines = unitModifiedLines - self.deckText.append(caption) + stream = _io.StringIO(caption) + lines = stream.readlines() + self.deckText.extend(lines) self.deckText = self.deckText + addedLines + self.logger = logging.getLogger("root") # stop propagting to root logger self.logger.propagate = False @@ -239,7 +245,7 @@ def writeDeck(self, addedLines=None): self.overwriteForcedByUser = True if ok: - tempFile = open(tempName, "w") + tempFile = open(tempName, "w", encoding='windows-1252') if addedLines != None: text = addedLines + self.deckText else: @@ -262,20 +268,12 @@ def readTrnsyDeck(self, useDeckName=False): useDeckName=useDeckName, eraseBeginComment=False, eliminateComments=False ) - # self.myDeck.loadDeckWithoutComments() - # self.linesDeckReaded = self.myDeck.linesReadedNoComments - def checkTrnsysDeck(self, nameDck, check=True): - - # self.readTrnsyDeck() - # deckUtils.checkEquationsAndConstants(self.linesDeckReaded) - lines = deckUtils.loadDeck(nameDck, eraseBeginComment=True, eliminateComments=True) if check: deckUtils.checkEquationsAndConstants(lines, self.nameDeck) self.linesDeckReaded = lines - # self.myDeck.checkEquationsAndConstants(self.deckText) #This does not need to read def saveUnitTypeFile(self): @@ -304,7 +302,6 @@ def writeTrnsysTypesUsed(self, name): line = "%s\tNone\t%s\n" % (self.filesUnitUsedInDdck[i], self.filesUsedInDdck[i]) else: line = "%s\t%s\t%s\n" % (self.filesUnitUsedInDdck[i], nameUnitFile, self.filesUsedInDdck[i]) - # line = "%s\t%s\t%s\n" % (self.filesUnitUsedInDdck[i],nameUnitFile[:-1],self.filesUsedInDdck[i]) lines = lines + line @@ -329,8 +326,6 @@ def automaticEnegyBalanceStaff(self): lines = deckUtils.addEnergyBalanceHourlyPrinter(unitId, eBalance) self.deckText = self.deckText[:-4] + lines + self.deckText[-4:] - self.writeDeck() # Deck rewritten with added printer - def replaceLines(self, replaceList): """ Replaces a deck lines with different lines @@ -348,4 +343,40 @@ def replaceLines(self, replaceList): for index, line in enumerate(self.linesChanged): if oldLine in line: self.linesChanged[index] = newLine + "\n" - changedLine = oldLine + + def analyseDck(self): + maxLineWidth = 0 + maxNumberOfConstantsInABlock = 0 + + constantsToCheck = ["UNIT", "EQUATIONS", "CONSTANTS", "PARAMETERS", "INPUTS"] + numOfTrnsysConstants = {} + + for index, line in enumerate(self.deckText): + maxLineWidth = max(maxLineWidth, len(line)) + if maxLineWidth > 1000: + self.logger.warning(f"Line {index + 1} has {maxLineWidth} characters which exceeds the limit.") + maxLineWidth = 0 + + for constant in constantsToCheck: + match = _re.search(fr"^\b{constant}\s*\d+\b", line, _re.MULTILINE) + if match: + if constant == "UNIT": + numOfTrnsysConstants[constant] = numOfTrnsysConstants.get(constant, 0) + 1 + elif constant in ("CONSTANTS", "EQUATIONS"): + split = match.group().split() + numOfTrnsysConstants["EQUATIONS"] = numOfTrnsysConstants.get("EQUATIONS", 0) + int(split[1]) + maxNumberOfConstantsInABlock = max(maxNumberOfConstantsInABlock, int(split[1])) + else: + split = match.group().split() + numOfTrnsysConstants[constant] = numOfTrnsysConstants.get(constant, 0) + int(split[1]) + maxNumberOfConstantsInABlock = max(maxNumberOfConstantsInABlock, int(split[1])) + break + + for constant, number in numOfTrnsysConstants.items(): + if (constant == "UNIT" and number > 1000) or (constant == "EQUATIONS" and number > 500) or ( + constant == "PARAMETERS" and number > 2000) or (constant == "INPUTS" and number > 750): + self.logger.warning(f"There are {number} of {constant} which exceeds the limit") + + if maxNumberOfConstantsInABlock > 250: + self.logger.warning( + f"There are {maxNumberOfConstantsInABlock} of components in one block which exceeds the limit") diff --git a/pytrnsys/trnsys_util/deckTrnsys.py b/pytrnsys/trnsys_util/deckTrnsys.py index e26b4b7a..ab5f548b 100644 --- a/pytrnsys/trnsys_util/deckTrnsys.py +++ b/pytrnsys/trnsys_util/deckTrnsys.py @@ -426,7 +426,7 @@ def changeParameter(self, _parameters): logger.debug("variation deck file at %s" % self.nameDck) - outfile = open(self.nameDck, "w") + outfile = open(self.nameDck, "w", encoding='windows-1252') outfile.writelines(lines) outfile.close() diff --git a/pytrnsys/trnsys_util/deckUtils.py b/pytrnsys/trnsys_util/deckUtils.py index 3401af36..363d1650 100644 --- a/pytrnsys/trnsys_util/deckUtils.py +++ b/pytrnsys/trnsys_util/deckUtils.py @@ -308,7 +308,7 @@ def loadDeck(nameDck, eraseBeginComment=True, eliminateComments=True): list of lines obateined form the deck without the comments """ - infile = open(nameDck, "r") + infile = open(nameDck, "r", encoding='windows-1252') lines = infile.readlines()

Raise warning when default TRNSYS constants limits are exceeded Include pytrnsys_spf/utils/dckAnalyzer.py into the standard dckCheck and raise a warning when the standard TRNSYS constants values (e.g., maxFileWidth or nMaxCardValues) are exceeded.

2022-04-20T07:30:10

0.0

[]

zauberzeug/rosys

zauberzeug__rosys-75

ff212ca21b70951e9e0b1a0ce9cbd799171e8db2

diff --git a/rosys/vision/calibration.py b/rosys/vision/calibration.py index bd1bda692..65ea4ffe4 100644 --- a/rosys/vision/calibration.py +++ b/rosys/vision/calibration.py @@ -2,7 +2,7 @@ import logging from dataclasses import dataclass, field -from typing import Optional +from typing import Optional, overload import cv2 import numpy as np @@ -49,59 +49,59 @@ def rotation(self) -> Rotation: def rotation_array(self) -> np.ndarray: return np.dot(self.extrinsics.rotation.R, self.intrinsics.rotation.R) - def project_to_image(self, world_point: Point3d) -> Point: - world_points = np.array([world_point.tuple], dtype=np.float32) - R = self.rotation_array - Rod = cv2.Rodrigues(R.T)[0] - t = -R.T @ self.extrinsics.translation - K = np.array(self.intrinsics.matrix) - D = np.array(self.intrinsics.distortion, dtype=float) - image_points, _ = cv2.projectPoints(world_points, Rod, t, K, D) - return Point(x=image_points[0, 0, 0], y=image_points[0, 0, 1]) + @overload + def project_to_image(self, world_coordinates: Point3d) -> Optional[Point]: ... + + @overload + def project_to_image(self, world_coordinates: np.ndarray) -> np.ndarray: ... + + def project_to_image(self, world_coordinates: Point3d | np.ndarray) -> Optional[Point] | np.ndarray: + if isinstance(world_coordinates, Point3d): + world_array = np.array([world_coordinates.tuple], dtype=np.float32) + image_array = self.project_to_image(world_array) + if np.isnan(image_array).any(): + return None + return Point(x=image_array[0, 0], y=image_array[0, 1]) # pylint: disable=unsubscriptable-object - def project_array_to_image(self, world_points: np.ndarray) -> np.ndarray: R = self.rotation_array Rod = cv2.Rodrigues(R.T)[0] t = -R.T @ self.extrinsics.translation K = np.array(self.intrinsics.matrix) D = np.array(self.intrinsics.distortion, dtype=float) - image_points, _ = cv2.projectPoints(world_points, Rod, t, K, D) - return image_points.reshape(-1, 2) + image_array, _ = cv2.projectPoints(world_coordinates, Rod, t, K, D) + local_coordinates = (world_coordinates - np.array(self.extrinsics.translation)) @ R + image_array[local_coordinates[:, 2] < 0, :] = np.nan + return image_array.reshape(-1, 2) - def project_from_image(self, image_point: Point, target_height: float = 0) -> Optional[Point3d]: - K = np.array(self.intrinsics.matrix) - D = np.array(self.intrinsics.distortion) - image_points = np.array(image_point.tuple, dtype=np.float32).reshape((1, 1, 2)) - image_points_ = cv2.undistortPoints(image_points, K, D).reshape(-1, 2) - image_points__ = cv2.convertPointsToHomogeneous(image_points_).reshape(-1, 3) - objPoints = image_points__ @ self.rotation_array.T - Z = self.extrinsics.translation[-1] - t = np.array(self.extrinsics.translation) - floor_points = t.T - objPoints * (Z - target_height) / objPoints[:, 2:] + @overload + def project_from_image(self, image_coordinates: Point, target_height: float = 0) -> Optional[Point3d]: ... - reprojection = self.project_to_image(Point3d(x=floor_points[0, 0], y=floor_points[0, 1], z=target_height)) - if objPoints[0, -1] * np.sign(Z) > 0 or reprojection.distance(image_point) > 2: - log.warning(f'reprojection failed with {reprojection.distance(image_point)} px') - return None + @overload + def project_from_image(self, image_coordinates: np.ndarray, target_height: float = 0) -> np.ndarray: ... - return Point3d(x=floor_points[0, 0], y=floor_points[0, 1], z=target_height) + def project_from_image(self, image_coordinates: Point | np.ndarray, target_height: float = 0) -> Optional[Point3d] | np.ndarray: + if isinstance(image_coordinates, Point): + image_points = np.array(image_coordinates.tuple, dtype=np.float32) + world_points = self.project_from_image(image_points, target_height=target_height) + if np.isnan(world_points).any(): + return None + return Point3d(x=world_points[0, 0], y=world_points[0, 1], z=world_points[0, 2]) # pylint: disable=unsubscriptable-object - def project_array_from_image(self, image_points: np.ndarray, target_height: float = 0) -> np.ndarray: K = np.array(self.intrinsics.matrix) D = np.array(self.intrinsics.distortion) - image_points_ = cv2.undistortPoints(image_points.astype(np.float32), K, D).reshape(-1, 2) - image_points__ = cv2.convertPointsToHomogeneous(image_points_).reshape(-1, 3) - objPoints = image_points__ @ self.rotation_array.T + image_coordinates_ = cv2.undistortPoints(image_coordinates.astype(np.float32), K, D).reshape(-1, 2) + image_coordinates__ = cv2.convertPointsToHomogeneous(image_coordinates_).reshape(-1, 3) + objPoints = image_coordinates__ @ self.rotation_array.T Z = self.extrinsics.translation[-1] t = np.array(self.extrinsics.translation) - floor_points = t.T - objPoints * (Z - target_height) / objPoints[:, 2:] + world_points = t.T - objPoints * (Z - target_height) / objPoints[:, 2:] - reprojection = self.project_array_to_image(floor_points) + reprojection = self.project_to_image(world_points) sign = objPoints[:, -1] * np.sign(Z) - distance = np.linalg.norm(reprojection - image_points, axis=1) - floor_points[np.logical_or(sign > 0, distance > 2), :] = np.nan + distance = np.linalg.norm(reprojection - image_coordinates, axis=1) + world_points[np.logical_not(np.logical_and(sign < 0, distance < 2)), :] = np.nan - return floor_points + return world_points @staticmethod def from_points( diff --git a/rosys/vision/camera_objects_.py b/rosys/vision/camera_objects_.py index 2d3d6ae5f..ef0dd1472 100644 --- a/rosys/vision/camera_objects_.py +++ b/rosys/vision/camera_objects_.py @@ -18,7 +18,7 @@ class CameraObjects(Group): With `debug=True` camera IDs are shown (default: `False`). """ - def __init__(self, camera_provider: CameraProvider, camera_projector: CameraProjector, *, + def __init__(self, camera_provider: CameraProvider[CalibratableCamera], camera_projector: CameraProjector, *, px_per_m: float = 10000, debug: bool = False) -> None: super().__init__() diff --git a/rosys/vision/camera_projector.py b/rosys/vision/camera_projector.py index 39e824d55..d60c76472 100644 --- a/rosys/vision/camera_projector.py +++ b/rosys/vision/camera_projector.py @@ -25,7 +25,7 @@ class CameraProjector: It is mainly used for visualization purposes. """ - def __init__(self, camera_provider: CameraProvider) -> None: + def __init__(self, camera_provider: CameraProvider[CalibratableCamera]) -> None: self.camera_provider = camera_provider self.projections: dict[str, Projection] = {} @@ -54,7 +54,7 @@ def project(camera: CalibratableCamera, rows: int = 12, columns: int = 16) -> Pr c, r = np.meshgrid(np.linspace(0, camera.calibration.intrinsics.size.width, columns), np.linspace(0, camera.calibration.intrinsics.size.height, rows)) image_points = np.stack((c.flatten(), r.flatten()), axis=1) - floor_points = camera.calibration.project_array_from_image(image_points).reshape(rows, columns, 3) + floor_points = camera.calibration.project_from_image(image_points).reshape(rows, columns, 3) invalid = np.isnan(floor_points).any(axis=2) return [ [ diff --git a/rosys/vision/detector_simulation.py b/rosys/vision/detector_simulation.py index 487b95802..024c77841 100644 --- a/rosys/vision/detector_simulation.py +++ b/rosys/vision/detector_simulation.py @@ -6,6 +6,7 @@ from .. import rosys from ..geometry import Point3d +from .camera import CalibratableCamera from .camera_provider import CameraProvider from .detections import BoxDetection, Detections, PointDetection from .detector import Detector @@ -34,7 +35,7 @@ class DetectorSimulation(Detector): An optional `noise` parameter controls the spatial accuracy in pixels. """ - def __init__(self, camera_provider: CameraProvider, *, noise: float = 1.0, name: Optional[str] = None) -> None: + def __init__(self, camera_provider: CameraProvider[CalibratableCamera], *, noise: float = 1.0, name: Optional[str] = None) -> None: super().__init__(name=name) self.camera_provider = camera_provider @@ -82,6 +83,8 @@ def detect_from_simulated_objects(self, image: Image) -> None: if np.dot(viewing_direction, object_direction) < 0: continue image_point = camera.calibration.project_to_image(obj.position) + if image_point is None: + continue if not (0 <= image_point.x < image.size.width and 0 <= image_point.y < image.size.height): continue @@ -101,7 +104,9 @@ def detect_from_simulated_objects(self, image: Image) -> None: for dy in [-obj.size[1] / 2, obj.size[1] / 2] for dz in [-obj.size[2] / 2, obj.size[2] / 2] ]) - image_points = camera.calibration.project_array_to_image(world_points) + image_points = camera.calibration.project_to_image(world_points) + if np.any(np.isnan(image_points)): + continue detections.boxes.append(BoxDetection( category_name=obj.category_name, model_name='simulation',

Projections from and to images are inconsistent for points behind camera In one direction the projection yields `None` or `np.nan` if a point is behind the camera. In the other direction the orientation is ignored, but should return `None` or `np.nan` as well. While we're at it, we can also unify the `project_*` methods, which are currently split for `Point` and `np.array`.

2024-01-18T13:24:56

0.0

[]

carsales/pyheif

carsales__pyheif-55

38468c5963f7c7e40f482c71d9662f9c7412895f

diff --git a/pyheif/__init__.py b/pyheif/__init__.py index 3eda86e..108dbdb 100644 --- a/pyheif/__init__.py +++ b/pyheif/__init__.py @@ -1,3 +1,4 @@ +import builtins import os import _libheif_cffi @@ -7,7 +8,7 @@ from .writer import * version_path = os.path.dirname(os.path.abspath(__file__)) + "/data/version.txt" -with open(version_path) as f: +with builtins.open(version_path) as f: __version__ = f.read().strip() diff --git a/pyheif/reader.py b/pyheif/reader.py index 5e2a26f..aa2c1e1 100644 --- a/pyheif/reader.py +++ b/pyheif/reader.py @@ -1,3 +1,5 @@ +import builtins +import functools import pathlib import warnings @@ -8,17 +10,50 @@ class HeifFile: def __init__( - self, *, size, data, metadata, color_profile, has_alpha, bit_depth, stride + self, *, size, has_alpha, bit_depth, metadata, color_profile, data, stride ): self.size = size - self.data = data - self.metadata = metadata - self.color_profile = color_profile self.has_alpha = has_alpha self.mode = "RGBA" if has_alpha else "RGB" self.bit_depth = bit_depth + self.metadata = metadata + self.color_profile = color_profile + self.data = data self.stride = stride + def __repr__(self): + return ( + f"<{self.__class__.__name__} {self.size[0]}x{self.size[1]} {self.mode} " + f"with {str(len(self.data)) + ' bytes' if self.data else 'no'} data>" + ) + + def load(self): + return self # already loaded + + def close(self): + pass # TODO: release self.data here? + + +class UndecodedHeifFile(HeifFile): + def __init__( + self, heif_handle, *, apply_transformations, convert_hdr_to_8bit, **kwargs + ): + self._heif_handle = heif_handle + self.apply_transformations = apply_transformations + self.convert_hdr_to_8bit = convert_hdr_to_8bit + super().__init__(data=None, stride=None, **kwargs) + + def load(self): + self.data, self.stride = _read_heif_image(self._heif_handle, self) + self.close() + self.__class__ = HeifFile + return self + + def close(self): + # Don't call super().close() here, we don't need to free bytes. + if hasattr(self, "_heif_handle"): + del self._heif_handle + def check(fp): d = _get_bytes(fp) @@ -33,14 +68,22 @@ def read_heif(fp, apply_transformations=True): def read(fp, *, apply_transformations=True, convert_hdr_to_8bit=True): + heif_file = open( + fp, + apply_transformations=apply_transformations, + convert_hdr_to_8bit=convert_hdr_to_8bit, + ) + return heif_file.load() + + +def open(fp, *, apply_transformations=True, convert_hdr_to_8bit=True): d = _get_bytes(fp) - result = _read_heif_bytes(d, apply_transformations, convert_hdr_to_8bit) - return result + return _read_heif_bytes(d, apply_transformations, convert_hdr_to_8bit) def _get_bytes(fp): if isinstance(fp, str): - with open(fp, "rb") as f: + with builtins.open(fp, "rb") as f: d = f.read() elif isinstance(fp, bytearray): d = bytes(fp) @@ -59,6 +102,19 @@ def _get_bytes(fp): return d +def _keep_refs(destructor, **refs): + """ + Keep refs to passed arguments until `inner` callback exist. + This prevents collecting parent objects until all children collcted. + """ + + def inner(cdata): + return destructor(cdata) + + inner._refs = refs + return inner + + def _read_heif_bytes(d, apply_transformations, convert_hdr_to_8bit): magic = d[:12] filetype_check = _libheif_cffi.lib.heif_check_filetype(magic, len(magic)) @@ -68,11 +124,9 @@ def _read_heif_bytes(d, apply_transformations, convert_hdr_to_8bit): warnings.warn("Input is an unsupported HEIF/AVIF file type - trying anyway!") ctx = _libheif_cffi.lib.heif_context_alloc() - try: - result = _read_heif_context(ctx, d, apply_transformations, convert_hdr_to_8bit) - finally: - _libheif_cffi.lib.heif_context_free(ctx) - return result + collect = _keep_refs(_libheif_cffi.lib.heif_context_free, data=d) + ctx = _libheif_cffi.ffi.gc(ctx, collect, size=len(d)) + return _read_heif_context(ctx, d, apply_transformations, convert_hdr_to_8bit) def _read_heif_context(ctx, d, apply_transformations, convert_hdr_to_8bit): @@ -94,67 +148,29 @@ def _read_heif_context(ctx, d, apply_transformations, convert_hdr_to_8bit): subcode=error.subcode, message=_libheif_cffi.ffi.string(error.message).decode(), ) - handle = p_handle[0] - - try: - result = _read_heif_handle(handle, apply_transformations, convert_hdr_to_8bit) - finally: - _libheif_cffi.lib.heif_image_handle_release(handle) - return result + collect = _keep_refs(_libheif_cffi.lib.heif_image_handle_release, ctx=ctx) + handle = _libheif_cffi.ffi.gc(p_handle[0], collect) + return _read_heif_handle(handle, apply_transformations, convert_hdr_to_8bit) def _read_heif_handle(handle, apply_transformations, convert_hdr_to_8bit): width = _libheif_cffi.lib.heif_image_handle_get_width(handle) height = _libheif_cffi.lib.heif_image_handle_get_height(handle) - size = (width, height) - has_alpha = bool(_libheif_cffi.lib.heif_image_handle_has_alpha_channel(handle)) bit_depth = _libheif_cffi.lib.heif_image_handle_get_luma_bits_per_pixel(handle) - colorspace = _constants.heif_colorspace_RGB - if convert_hdr_to_8bit or bit_depth <= 8: - if has_alpha: - chroma = _constants.heif_chroma_interleaved_RGBA - else: - chroma = _constants.heif_chroma_interleaved_RGB - else: - if has_alpha: - chroma = _constants.heif_chroma_interleaved_RRGGBBAA_BE - else: - chroma = _constants.heif_chroma_interleaved_RRGGBB_BE - - p_options = _libheif_cffi.lib.heif_decoding_options_alloc() - p_options.ignore_transformations = int(not apply_transformations) - p_options.convert_hdr_to_8bit = int(convert_hdr_to_8bit) - - p_img = _libheif_cffi.ffi.new("struct heif_image **") - error = _libheif_cffi.lib.heif_decode_image( - handle, p_img, colorspace, chroma, p_options, - ) - _libheif_cffi.lib.heif_decoding_options_free(p_options) - if error.code != 0: - raise _error.HeifError( - code=error.code, - subcode=error.subcode, - message=_libheif_cffi.ffi.string(error.message).decode(), - ) - img = p_img[0] - - try: - data, stride = _read_heif_image(img, height) - finally: - _libheif_cffi.lib.heif_image_release(img) metadata = _read_metadata(handle) color_profile = _read_color_profile(handle) - heif_file = HeifFile( - size=size, - data=data, - metadata=metadata, - color_profile=color_profile, + heif_file = UndecodedHeifFile( + handle, + size=(width, height), has_alpha=has_alpha, bit_depth=bit_depth, - stride=stride, + metadata=metadata, + color_profile=color_profile, + apply_transformations=apply_transformations, + convert_hdr_to_8bit=convert_hdr_to_8bit, ) return heif_file @@ -225,15 +241,55 @@ def _read_color_profile(handle): return color_profile -def _read_heif_image(img, height): +def _read_heif_image(handle, heif_file): + colorspace = _constants.heif_colorspace_RGB + if heif_file.convert_hdr_to_8bit or heif_file.bit_depth <= 8: + if heif_file.has_alpha: + chroma = _constants.heif_chroma_interleaved_RGBA + else: + chroma = _constants.heif_chroma_interleaved_RGB + else: + if heif_file.has_alpha: + chroma = _constants.heif_chroma_interleaved_RRGGBBAA_BE + else: + chroma = _constants.heif_chroma_interleaved_RRGGBB_BE + + p_options = _libheif_cffi.lib.heif_decoding_options_alloc() + p_options = _libheif_cffi.ffi.gc( + p_options, _libheif_cffi.lib.heif_decoding_options_free + ) + p_options.ignore_transformations = int(not heif_file.apply_transformations) + p_options.convert_hdr_to_8bit = int(heif_file.convert_hdr_to_8bit) + + p_img = _libheif_cffi.ffi.new("struct heif_image **") + error = _libheif_cffi.lib.heif_decode_image( + handle, p_img, colorspace, chroma, p_options, + ) + if error.code != 0: + raise _error.HeifError( + code=error.code, + subcode=error.subcode, + message=_libheif_cffi.ffi.string(error.message).decode(), + ) + img = p_img[0] + p_stride = _libheif_cffi.ffi.new("int *") p_data = _libheif_cffi.lib.heif_image_get_plane_readonly( img, _constants.heif_channel_interleaved, p_stride ) stride = p_stride[0] - data_length = height * stride + data_length = heif_file.size[1] * stride + + # Release image as soon as no references to p_data left + collect = functools.partial(_release_heif_image, img) + p_data = _libheif_cffi.ffi.gc(p_data, collect, size=data_length) + + # ffi.buffer obligatory keeps a reference to p_data data_buffer = _libheif_cffi.ffi.buffer(p_data, data_length) - data = bytes(data_buffer) - return data, stride + return data_buffer, stride + + +def _release_heif_image(img, p_data=None): + _libheif_cffi.lib.heif_image_release(img)

Return ffi.buffer as data to avoid extra memory copying This saves 3-7% time depending on number of cores. master: ```python $ python -m timeit -n5 -r10 -s "import pyheif; data = open('./full.norot.heic', 'rb').read()" "im = pyheif.read(data)" 5 loops, best of 10: 345 msec per loop ``` avoid-copy: ```python $ python -m timeit -n5 -r10 -s "import pyheif; data = open('./full.norot.heic', 'rb').read()" "im = pyheif.read(data)" 5 loops, best of 10: 323 msec per loop ``` Also this makes `data` writable buffer, which avoid one more copy in such cases: master: ```python In [1]: import pyheif, numpy In [2]: im = pyheif.read('./full.norot.heic') In [3]: buf = numpy.frombuffer(im.data, dtype=numpy.uint8) In [4]: buf[0:3] = b'123' ValueError: assignment destination is read-only In [5]: buf = buf.copy() In [6]: buf[0:3] = b'123' # now this works ``` avoid-copy: ```python In [1]: import pyheif, numpy In [2]: im = pyheif.read('./full.norot.heic') In [3]: buf = numpy.frombuffer(im.data, dtype=numpy.uint8) In [4]: buf[0:3] = b'123' ``` Return ffi.buffer as data to avoid extra memory copying This saves 3-7% time depending on number of cores. master: ```python $ python -m timeit -n5 -r10 -s "import pyheif; data = open('./full.norot.heic', 'rb').read()" "im = pyheif.read(data)" 5 loops, best of 10: 345 msec per loop ``` avoid-copy: ```python $ python -m timeit -n5 -r10 -s "import pyheif; data = open('./full.norot.heic', 'rb').read()" "im = pyheif.read(data)" 5 loops, best of 10: 323 msec per loop ``` Also this makes `data` writable buffer, which avoid one more copy in such cases: master: ```python In [1]: import pyheif, numpy In [2]: im = pyheif.read('./full.norot.heic') In [3]: buf = numpy.frombuffer(im.data, dtype=numpy.uint8) In [4]: buf[0:3] = b'123' ValueError: assignment destination is read-only In [5]: buf = buf.copy() In [6]: buf[0:3] = b'123' # now this works ``` avoid-copy: ```python In [1]: import pyheif, numpy In [2]: im = pyheif.read('./full.norot.heic') In [3]: buf = numpy.frombuffer(im.data, dtype=numpy.uint8) In [4]: buf[0:3] = b'123' ```

2021-10-19T23:52:38

0.0

[]

20tab/bvh-python

20tab__bvh-python-6

ed8c96359c5e8e85fca934808e84ef446cfae444

diff --git a/README.md b/README.md index 0a3324b..eebb0b7 100644 --- a/README.md +++ b/README.md @@ -72,3 +72,8 @@ Python module for parsing BVH (Biovision hierarchical data) mocap files >>> [str(node) for node in mocap.search('JOINT')] ['JOINT mixamorig:Spine', 'JOINT mixamorig:Spine1', 'JOINT mixamorig:Spine2', 'JOINT mixamorig:Neck', 'JOINT mixamorig:Head', 'JOINT mixamorig:HeadTop_End', 'JOINT mixamorig:LeftEye', 'JOINT mixamorig:RightEye', 'JOINT mixamorig:LeftShoulder', 'JOINT mixamorig:LeftArm', 'JOINT mixamorig:LeftForeArm', 'JOINT mixamorig:LeftHand', 'JOINT mixamorig:LeftHandMiddle1', 'JOINT mixamorig:LeftHandMiddle2', 'JOINT mixamorig:LeftHandMiddle3', 'JOINT mixamorig:LeftHandThumb1', 'JOINT mixamorig:LeftHandThumb2', 'JOINT mixamorig:LeftHandThumb3', 'JOINT mixamorig:LeftHandIndex1', 'JOINT mixamorig:LeftHandIndex2', 'JOINT mixamorig:LeftHandIndex3', 'JOINT mixamorig:LeftHandRing1', 'JOINT mixamorig:LeftHandRing2', 'JOINT mixamorig:LeftHandRing3', 'JOINT mixamorig:LeftHandPinky1', 'JOINT mixamorig:LeftHandPinky2', 'JOINT mixamorig:LeftHandPinky3', 'JOINT mixamorig:RightShoulder', 'JOINT mixamorig:RightArm', 'JOINT mixamorig:RightForeArm', 'JOINT mixamorig:RightHand', 'JOINT mixamorig:RightHandMiddle1', 'JOINT mixamorig:RightHandMiddle2', 'JOINT mixamorig:RightHandMiddle3', 'JOINT mixamorig:RightHandThumb1', 'JOINT mixamorig:RightHandThumb2', 'JOINT mixamorig:RightHandThumb3', 'JOINT mixamorig:RightHandIndex1', 'JOINT mixamorig:RightHandIndex2', 'JOINT mixamorig:RightHandIndex3', 'JOINT mixamorig:RightHandRing1', 'JOINT mixamorig:RightHandRing2', 'JOINT mixamorig:RightHandRing3', 'JOINT mixamorig:RightHandPinky1', 'JOINT mixamorig:RightHandPinky2', 'JOINT mixamorig:RightHandPinky3', 'JOINT mixamorig:RightUpLeg', 'JOINT mixamorig:RightLeg', 'JOINT mixamorig:RightFoot', 'JOINT mixamorig:RightToeBase', 'JOINT mixamorig:LeftUpLeg', 'JOINT mixamorig:LeftLeg', 'JOINT mixamorig:LeftFoot', 'JOINT mixamorig:LeftToeBase'] ``` +#### Get joint's direct children +```python +>>> mocap.joint_direct_children('mixamorig:Hips') +[JOINT mixamorig:Spine, JOINT mixamorig:RightUpLeg, JOINT mixamorig:LeftUpLeg] +``` diff --git a/bvh.py b/bvh.py index 08f22bd..40fc5a2 100644 --- a/bvh.py +++ b/bvh.py @@ -113,6 +113,10 @@ def iterate_joints(joint): iterate_joints(next(self.root.filter('ROOT'))) return joints + def joint_direct_children(self, name): + joint = self.get_joint(name) + return [child for child in joint.filter('JOINT')] + def get_joint_index(self, name): return self.get_joints().index(self.get_joint(name))

GetChildren method Currently it's not possible to get the children of a certain joint.

2019-07-23T15:04:08

0.0

[]

ManuelNavarroGarcia/cpsplines

ManuelNavarroGarcia__cpsplines-45

3a956f89a0cc26faf0428ee0e4e388bb89de35f7

diff --git a/cpsplines/fittings/fit_cpsplines.py b/cpsplines/fittings/fit_cpsplines.py index 8361a7e..4089cc9 100644 --- a/cpsplines/fittings/fit_cpsplines.py +++ b/cpsplines/fittings/fit_cpsplines.py @@ -121,6 +121,9 @@ class CPsplines: The fitted decision variables of the B-spline expansion. data_arrangement : str The structure of the data. Must be either "gridded" or "scattered". + cat : Dict[int, str] + The mapping of the label encoder when dealing with binary data. Hence, + it is accesible when the binomial family is considered. References ---------- @@ -594,6 +597,13 @@ def fit( if self.sp_method not in ["grid_search", "optimizer"]: raise ValueError(f"Invalid `sp_method`: {self.sp_method}.") + if self.family.name == "binomial": + self.cat = dict(enumerate(data[y_col].astype("category").cat.categories)) + data = data.assign( + **{y_col: data[y_col].map({v: k for k, v in self.cat.items()})} + ) + if self.cat[1] != 1: + logging.warning(f"{self.cat[1]} is considered as the positive class.") x, y = self._preprocessor(data=data, y_col=y_col) # Construct the B-spline bases diff --git a/data/kyphosis.csv b/data/kyphosis.csv index d961b4a..cfe04d3 100644 --- a/data/kyphosis.csv +++ b/data/kyphosis.csv @@ -1,82 +1,82 @@ Kyphosis,Age,Number,Start -0,71,3,5 -0,158,3,14 -1,128,4,5 -0,2,5,1 -0,1,4,15 -0,1,2,16 -0,61,2,17 -0,37,3,16 -0,113,2,16 -1,59,6,12 -1,82,5,14 -0,148,3,16 -0,18,5,2 -0,1,4,12 -0,168,3,18 -0,1,3,16 -0,78,6,15 -0,175,5,13 -0,80,5,16 -0,27,4,9 -0,22,2,16 -1,105,6,5 -1,96,3,12 -0,131,2,3 -1,15,7,2 -0,9,5,13 -0,8,3,6 -0,100,3,14 -0,4,3,16 -0,151,2,16 -0,31,3,16 -0,125,2,11 -0,130,5,13 -0,112,3,16 -0,140,5,11 -0,93,3,16 -0,1,3,9 -1,52,5,6 -0,20,6,9 -1,91,5,12 -1,73,5,1 -0,35,3,13 -0,143,9,3 -0,61,4,1 -0,97,3,16 -1,139,3,10 -0,136,4,15 -0,131,5,13 -1,121,3,3 -0,177,2,14 -0,68,5,10 -0,9,2,17 -1,139,10,6 -0,2,2,17 -0,140,4,15 -0,72,5,15 -0,2,3,13 -1,120,5,8 -0,51,7,9 -0,102,3,13 -1,130,4,1 -1,114,7,8 -0,81,4,1 -0,118,3,16 -0,118,4,16 -0,17,4,10 -0,195,2,17 -0,159,4,13 -0,18,4,11 -0,15,5,16 -0,158,5,14 -0,127,4,12 -0,87,4,16 -0,206,4,10 -0,11,3,15 -0,178,4,15 -1,157,3,13 -0,26,7,13 -0,120,2,13 -1,42,7,6 -0,36,4,13 +Absent,71,3,5 +Absent,158,3,14 +Present,128,4,5 +Absent,2,5,1 +Absent,1,4,15 +Absent,1,2,16 +Absent,61,2,17 +Absent,37,3,16 +Absent,113,2,16 +Present,59,6,12 +Present,82,5,14 +Absent,148,3,16 +Absent,18,5,2 +Absent,1,4,12 +Absent,168,3,18 +Absent,1,3,16 +Absent,78,6,15 +Absent,175,5,13 +Absent,80,5,16 +Absent,27,4,9 +Absent,22,2,16 +Present,105,6,5 +Present,96,3,12 +Absent,131,2,3 +Present,15,7,2 +Absent,9,5,13 +Absent,8,3,6 +Absent,100,3,14 +Absent,4,3,16 +Absent,151,2,16 +Absent,31,3,16 +Absent,125,2,11 +Absent,130,5,13 +Absent,112,3,16 +Absent,140,5,11 +Absent,93,3,16 +Absent,1,3,9 +Present,52,5,6 +Absent,20,6,9 +Present,91,5,12 +Present,73,5,1 +Absent,35,3,13 +Absent,143,9,3 +Absent,61,4,1 +Absent,97,3,16 +Present,139,3,10 +Absent,136,4,15 +Absent,131,5,13 +Present,121,3,3 +Absent,177,2,14 +Absent,68,5,10 +Absent,9,2,17 +Present,139,10,6 +Absent,2,2,17 +Absent,140,4,15 +Absent,72,5,15 +Absent,2,3,13 +Present,120,5,8 +Absent,51,7,9 +Absent,102,3,13 +Present,130,4,1 +Present,114,7,8 +Absent,81,4,1 +Absent,118,3,16 +Absent,118,4,16 +Absent,17,4,10 +Absent,195,2,17 +Absent,159,4,13 +Absent,18,4,11 +Absent,15,5,16 +Absent,158,5,14 +Absent,127,4,12 +Absent,87,4,16 +Absent,206,4,10 +Absent,11,3,15 +Absent,178,4,15 +Present,157,3,13 +Absent,26,7,13 +Absent,120,2,13 +Present,42,7,6 +Absent,36,4,13 diff --git a/examples/non_gaussian_data.ipynb b/examples/non_gaussian_data.ipynb index af06b72..72a0fa6 100644 --- a/examples/non_gaussian_data.ipynb +++ b/examples/non_gaussian_data.ipynb @@ -64,7 +64,7 @@ "name": "stdout", "output_type": "stream", "text": [ - "[Solve the problem with smoothing parameters (444.5186255273247,): ] Elapsed time (s): 0.069853\n" + "[Solve the problem with smoothing parameters (444.5186255273247,): ] Elapsed time (s): 0.115425\n" ] } ], @@ -125,7 +125,7 @@ "name": "stdout", "output_type": "stream", "text": [ - "[Solve the problem with smoothing parameters (443.3914121205766,): ] Elapsed time (s): 0.126242\n" + "[Solve the problem with smoothing parameters (443.3914121205766,): ] Elapsed time (s): 0.098054\n" ] } ], @@ -203,7 +203,7 @@ "name": "stdout", "output_type": "stream", "text": [ - "[Solve the problem with smoothing parameters (0.1, 0.1): ] Elapsed time (s): 8.514172\n" + "[Solve the problem with smoothing parameters (0.1, 0.1): ] Elapsed time (s): 8.427936\n" ] } ], @@ -295,11 +295,18 @@ "execution_count": 12, "metadata": {}, "outputs": [ + { + "name": "stderr", + "output_type": "stream", + "text": [ + "WARNING:root:Present is considered as the positive class.\n" + ] + }, { "name": "stdout", "output_type": "stream", "text": [ - "[Solve the problem with smoothing parameters (1,): ] Elapsed time (s): 0.021814\n" + "[Solve the problem with smoothing parameters (1,): ] Elapsed time (s): 0.015729\n" ] } ], @@ -320,7 +327,7 @@ "outputs": [ { "data": { - "image/png": "iVBORw0KGgoAAAANSUhEUgAAA4YAAAJXCAYAAADPZkozAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8qNh9FAAAACXBIWXMAAAsTAAALEwEAmpwYAABqF0lEQVR4nO3dd5RU9f3/8ed76R3pilRFVBQsiA0joAQVO/ausf+MaSaa5teoSTQmmphYotGIJfbeFcWOKDYUjIoFpUrvnc/vjxlw3cAyyO7M7s7zcc6c2b1z753XvXtnmBf3zr2RUkKSJEmSVLxKCh1AkiRJklRYFkNJkiRJKnIWQ0mSJEkqchZDSZIkSSpyFkNJkiRJKnIWQ0mSJEkqcrULHSCfWrVqlTp37lzoGN+ybNkyAOrUqVPgJBumpiwHuCxVVU1ZlpqyHOCyVFU1ZVlqynKAy1IV1ZTlAJelqqrKy/LWW29NTym1Lju8qIph586dGTVqVKFjfMuUKVMAaNeuXYGTbJiashzgslRVNWVZaspygMtSVdWUZakpywEuS1VUU5YDXJaqqiovS0SMX9NwDyWVJEmSpCJnMZQkSZKkImcxlCRJkqQiZzGUJEmSpCJnMZQkSZKkImcxlCRJkqQiZzGUJEmSpCJnMZQkSZKkImcxlCRJkqQiZzGUJEmSpCJnMZQkSZKkImcxlCRJkqQiZzGUJEmSpCJnMZQkSZKkImcxlCRJkqQiZzGUJEmSpCJnMZQkSZKkIpf3YhgRm0bE3yNiREQsjIgUEZ1znLYkIn4ZEV9ExOKIeC8ihlRyZEmSJEmq0Qqxx3Bz4AhgFvDyek57CXAR8A9gX+B14N6I2K8iA0qSJElSMaldgOd8KaXUFiAiTgW+n8tEEdEGOA+4LKX05+zg4RGxOXAZ8ERlhJUkSZKkmi7vxTCltPI7TjoIqAvcXmb47cDNEdElpfT5BoXTBpkxA95/H2bOhBYtoGdPaNu20KmK19Sp8PrrMGcObLJJ1fh7TJ0Ko0dX/jaSr+fJl6qwPGvKUAgVtS7KzmeTTaBly4rPu745qsrftiIylJ3vxhvD5MmFf/1XhXX+wQfw+OMwaVJm2xs8GLbZJr8Zqoqq8PeQlFGdTj7TA1gCjCszfEz2fuv8xlFpM2bAyJGwZAm0aZO5HzYs84av/Js6NbP+ly7N/ENbFf4eqzJV9jaSr+fJl6qwPGvLMGNG/jKUl2N918Wa5jNyZPVdnqqYoex8p0yBa6/N3Bfy9V8V1vkHH2TWxYIF0KlT5v7aazPDi01V+HtI+kZ1KoYtgNkppVRm+MxSj6tAPvkEGjfO3CIy982aZf4XUPk3enRm/TdsWHX+HqsyVfY2kq/nyZeqsDxry/DJJ/nLUF6O9V0Xa5pP48bVd3mqYoay8506FVq1ytwX8vVfFdb5449n1kWLFlBSkrlv1SozvNhUhb+HpG9Up2L4nUTE6RExKiJGTZs2rdBxaqw5c6BBg28Pa9Qoc2iI8m/mzMz6L63Qf498ZaqKy74hqsLyrC3DnDn5y1BejvVdF2uaT4MG1Xd5qmKGsvOdOxeaN8/cV+TzrOt5yz5PVVjnkyZl1kVpzZtnhhebqvD3kPSN6lQMZwHNIyLKDF+1p3CNbyMppRtSSr1TSr1bt25dqQGLWbNmsGjRt4ctWJD5n1DlX4sWmfVfWqH/HvnKVBWXfUNUheVZW4ZmzfKXobwc67su1jSfRYuq7/JUxQxl59u0KcyenbmvyOdZ1/OWfZ6qsM432SSzLkqbPTszvNhUhb+HpG9Up2I4BqgHbFZm+KrvFo7NbxyV1q0bzJ+fuaWUuZ8zp3AnqCh2PXtm1v/ChVXn77EqU2VvI/l6nnypCsuztgzduuUvQ3k51nddrGk+8+dX3+WpihnKzrdtW5g+PXNfyNd/VVjngwdn1sXMmbByZeZ++vTM8GJTFf4ekr5RnYrhU8Ay4Ngyw48DPvCMpIXVsiXsvDPUqwdff52533tvzyxWKG3bZtZ/3bqZDx1V4e+xKlNlbyP5ep58qQrLs7YM+T6LZ0WtizXNZ+edq+/yVMUMZefbrh2cfXbmvpCv/6qwzrfZJrMuGjWC8eMz92efXZxnJa0Kfw9J3yjEdQyJiMOyP+6Yvd83IqYB01JKL2bHWQ4MTSn9ACCl9HVEXAn8MiLmAW8DRwIDgAPzugBao5YtoUePQqfQKm3bwi67ZH5u166wWVZp2xYGDqw5z5MvVWF51pRhypSqkaMi5lOIZVlTjpqUYU3zzUf5WdfyVIV1vs02xVkE16Qq/D0kZRSkGAL3lvn92uz9i0C/7M+1srfSfg3MB34EtAM+Ao5IKT1WOTElSZIkqeYrSDFMKZU9gUxO46SUVgCXZm+SJEmSpApQnb5jKEmSJEmqBBZDSZIkSSpyFkNJkiRJKnIWQ0mSJEkqchZDSZIkSSpyFkNJkiRJKnIWQ0mSJEkqchZDSZIkSSpyFkNJkiRJKnIWQ0mSJEkqchZDSZIkSSpyFkNJkiRJKnIWQ0mSJEkqchZDSZIkSSpyFkNJkiRJKnIWQ0mSJEkqchZDSZIkSSpyFkNJkiRJKnIWQ0mSJEkqchZDSZIkSSpyFkNJkiRJKnIWQ0mSJEkqchZDSZIkSSpyFkNJkiRJKnIWQ0mSJEkqchZDSZIkSSpyFkNJkiRJKnIWQ0mSJEkqchZDSZIkSSpyFkNJkiRJKnIWQ0mSJEkqchZDSZIkSSpyFkNJkiRJKnIWQ0mSJEkqchZDSZIkSSpyFkNJkiRJKnIWQ0mSJEkqchZDSZIkSSpyFkNJkiRJKnIWQ0mSJEkqchZDSZIkSSpyFkNJkiRJKnIWQ0mSJEkqchZDSZIkSSpyFkNJkiRJKnIWQ0mSJEkqchZDSZIkSSpyFkNJkiRJKnIWQ0mSJEkqchZDSZIkSSpyFkNJkiRJKnIWQ0mSJEkqchZDSZIkSSpyFkNJkiRJKnIWQ0mSJEkqchZDSZIkSSpyFkNJkiRJKnIWQ0mSJEkqchZDSZIkSSpyFkNJkiRJKnIWQ0mSJEkqchZDSZIkSSpyFkNJkiRJKnIWQ0mSJEkqchZDSZIkSSpyFkNJkiRJKnIWQ0mSJEkqchZDSZIkSSpyFkNJkiRJKnIWQ0mSJEkqchZDSZIkSSpyFkNJkiRJKnIWQ0mSJEkqchZDSZIkSSpyFkNJkiRJKnIWQ0mSJEkqchZDSZIkSSpyFkNJkiRJKnIWQ0mSJEkqchZDSZIkSSpyFkNJkiRJKnIWQ0mSJEkqchZDSZIkSSpyFkNJkiRJKnIWQ0mSJEkqchZDSZIkSSpyeS+GEdEhIu6LiDkRMTciHoiIjjlO2zEihkbElxGxKCI+johLI6JRZeeWJEmSpJqqdj6fLCIaAs8DS4ATgQRcCgyPiJ4ppQXlTNsIGAbUAX4LfAnsBPwO6AYcWbnpJUmSJKlmymsxBE4DugLdU0rjACJiNPAJcAZwZTnT7k6mAA5KKT2THTY8IloA50VEw5TSwsqLLkmSJEk1U74PJT0QeH1VKQRIKX0OvAoctI5p62bv55YZPpvMckQFZZQkSZKkopLvYtgD+GANw8cAW69j2mFk9ixeHhFbR0TjiBgA/Ai4vrzDUCVJkiRJa5fvYtgCmLWG4TOBjcqbMKW0GOhLJvMYYB7wHPAYcE7FxpQkSZKk4pHv7xh+ZxFRH7gbaAMcT+bkM32AC4HlwFlrme504HSAjh1zOvmpJEmSJBWVfBfDWax5z+Da9iSW9gOgH7B5SunT7LCXImIOcENEXJ9Seq/sRCmlG4AbAHr37p2+a3BJkiRJqqnyfSjpGDLfMyxra2DsOqbdFphVqhSu8kb2fqsNzCZJkiRJRSnfxfARYJeI6LpqQER0JnMpikfWMe0UYKOI2LzM8J2z9xMrKqQkSZIkFZN8F8MbgS+AhyPioIg4EHgY+Ar456qRIqJTRCyPiAtLTXsLmRPOPBERJ0ZE/4j4OfBn4C0yl7yQJEmSJK2nvBbD7CUlBgAfA7cBdwCfAwNSSvNLjRpArdL5UkpfALsA7wKXAk8Ap5H5/uDAlNLKyl8CSZIkSap58n5W0pTSl8CQdYzzBWu4YH1KaSxwROUkkyRJkqTilO9DSSVJkiRJVYzFUJIkSZKKnMVQkiRJkoqcxVCSJEmSipzFUJIkSZKKnMVQkiRJkoqcxVCSJEmSipzFUJIkSZKKnMVQkiRJkoqcxVCSJEmSipzFUJIkSZKKnMVQkiRJkoqcxVCSJEmSipzFUJIkSZKKnMVQkiRJkoqcxVCSJEmSipzFUJIkSZKKnMVQkiRJkoqcxVCSJEmSipzFUJIkSZKKnMVQkiRJkoqcxVCSJEmSipzFUJIkSZKKnMVQkiRJkoqcxVCSJEmSipzFUJIkSZKKnMVQkiRJkoqcxVCSJEmSipzFUJIkSZKKnMVQkiRJkoqcxVCSJEmSipzFUJIkSZKKnMVQkiRJkoqcxVCSJEmSipzFUJIkSZKKnMVQkiRJkoqcxVCSJEmSipzFUJIkSZKKnMVQkiRJkoqcxVCSJEmSipzFUJIkSZKKnMVQkiRJkoqcxVCSJEmSipzFUJIkSZKKnMVQkiRJkoqcxVCSJEmSipzFUJIkSZKKnMVQkiRJkoqcxVCSJEmSipzFUJIkSZKKnMVQkiRJkoqcxVCSJEmSilztXEaKiBKgJKW0vNSwQcA2wPMppXcqKZ8kSZIkqZLlVAyBO4ElwAkAEXEmcG32sWURMTilNKwS8kmSJEmSKlmuh5LuAjxR6vefA/8CmgEPAL+u4FySJEmSpDzJtRi2ASYCRMTmQBfgHymlecC/gW0rJ54kSZIkqbLlWgznAi2zP/cDpqeURmd/XwHUr+BckiRJkqQ8yfU7hq8BF0TEcuDHfPuw0s2BCRWcS5IkSZKUJ7nuMfwFmT2Gj5DZO3hRqceOBEZUbCxJkiRJUr7ktMcwpfQJ0C0iWqaUZpR5+EfAlApPJkmSJEnKi1wPJQVgDaWQlNL7FRdHkiRJkpRvay2GEXEh8K+U0qTsz+VJKaVLKjaaJEmSJCkfyttjeBHwFDCJb3+ncE0SYDGUJEmSpGporcUwpVSypp8lSZIkSTWLhU+SJEmSilxOxTAitoiIPqV+bxARf4yIRyPinMqLJ0mSJEmqbLnuMfwHcFip338P/AzYBLgqIv5fRQeTJEmSJOVHrsWwF/AqQESUACcA56eUdgQuBU6vnHiSJEmSpMqWazFsBqy6huH2wEbAfdnfXwC6VmwsSZIkSVK+5FoMpwKbZ3/+PvBpSumr7O+NgeUVHUySJEmSlB/lXcewtEeAP0bENsBJwD9LPbYt8FkF55IkSZIk5UmuxfACoD4wiExJ/EOpxw4EnqngXJIkSZKkPMmpGKaUFgCnreWx3So0kSRJkiQpr3LdYwhARLQAdgVaADOBESmlmZURTJIkSZKUHzkXw4i4lMy1C+uVGrwkIv6cUvpthSeTJEmSJOVFTmcljYgfA78Cbgf6A1tl728HfhUR51ZWQEmSJElS5cp1j+GZwN9SSj8pNewj4MWImA+cDVxd0eEkSZIkSZUv1+sYdgYeX8tjj2cflyRJkiRVQ7kWwxnANmt5rEf2cUmSJElSNZRrMXwQuCQijo+I2gARUTsijgYuBu6vrICSJEmSpMqVazH8JfAuMBRYFBFTgUXAHcB7ZE5MI0mSJEmqhnK9wP28iPgeMBjYg2+uY/gi8GRKKVVeREmSJElSZcr5OobZ8vdY9iZJkiRJqiFyLoarREQboH7Z4SmlLyskkSRJkiQpr3IqhhHRFPgbcCRQby2j1aqoUJIkSZKk/Ml1j+E1wBDgJuB9YEmlJZIkSZIk5VWuxXAf4OcppWs29AkjogNwFTAQCGAY8ONcD0WNiK3IXCKjP9AI+BK4NqX0tw3NJkmSJEnFaH2+Y/jRhj5ZRDQEniezx/FEIAGXAsMjomdKacE6pu+dnf4F4FRgDtANaLyh2SRJkiSpWOVaDO8CDiCzd29DnAZ0BbqnlMYBRMRo4BPgDODKtU0YESXArcBzKaVDSj00fAMzSZIkSVJRW2sxjIgBpX59BvhrRDQBniBzDcNvSSk9n8PzHQi8vqoUZqf7PCJeBQ6inGII9AO2IlMgJUmSJEkVpLw9hsPIHOoZpe67ACeVGqf047mclbQH8PAaho8BDl/HtH2z9/Uj4nVgR2AWmb2Z56eUFuXw/JIkSZKkMsorhv0r4flakClzZc0ENlrHtJtk7+8G/gFcAPQmcyKaDsAha5lOkiRJklSOtRbDlNKL+QySg5Ls/e0ppQuzP78QEbWAyyJiq5TSh2UniojTgdMBOnbsmJ+kkiRJklSNlKx7lG9ERNOI2C0iDs/eN1nP55vFmvcMrm1PYmkzsvfPlhn+TPZ++zVNlFK6IaXUO6XUu3Xr1jkHlSRJkqRikfPlKiLiQuBnZC4NEdnB8yLiipTSpTnOZgyZ7xmWtTUwNodpy7MyxwySJEmSpFJy2mMYEb8DLiLz/b6BwLbA3sA9wO8i4qIcn+8RYJeI6Fpq3p2B3bOPledJMtc/HFRm+D7Z+1E5ZpAkSZIklZLrHsPTgL+klH5eatgY4PmImEPmO3wX5TCfG4FzgIcj4jdkzmZ6CfAV8M9VI0VEJ+BT4OKU0sUAKaUZEfFH4LcRMZfMhe57AxcCQ0tfAkOSJEmSlLtcv2PYDHh6LY89lX18nVJKC4ABwMfAbcAdwOfAgJTS/FKjBpnLX5TNdzHwC+AIMtdTPAu4gkxxlSRJkiR9B7nuMRwJ7ETm2oZl7ZR9PCcppS+BIesY5wu++R5j6eEJuDJ7kyRJkiRVgFyL4bnAgxGxHLgXmAq0JbPn7hTgoIhYvXcvpeSJYCRJkiSpmsi1GI7O3l+WvZUWwPulfk/rMV9JkiRJUoHlWuAuJlP4JEmSJEk1TE7FMKV0USXnkCRJkiQVSK7XMfx+ZQeRJEmSJBVGrpereCoixkXEzyOiVaUmkiRJkiTlVa7FcADwJpmL0U+IiP9ExJ6VF0uSJEmSlC85FcOU0gsppaOBTYHfAr2B4RHxYUT8KCI2qsyQkiRJkqTKk+seQwBSStNTSleklLYABgLTyVxsfkJE3BIR21ZGSEmSJElS5VmvYrhKROxH5qL3uwBfA7cBewJvR8RZFRdPkiRJklTZci6GEdEuIn4dEZ8DjwHNgeOADimlM4HNgX8CF1ZGUEmSJElS5cjpOoYR8QAwGFgM3A5cm1IaU3qclNKKiPgPcHaFp5QkSZIkVZqciiGZvYE/Bm5LKc0vZ7z3gf4bGkqSJEmSlD+5Hkp6WUrpurWVwoj4O0BKaV5K6cUKSydJkiRJqnS5FsObI2LvNT0QEX8DTq24SJIkSZKkfMq1GP4eeCAidig9MCKuBM4EjqroYJIkSZKk/MjpO4YppUsiYhPgyYjYLaX0aUT8GTgHOCql9HClppQkSZIkVZpcTz4DmbONtgGeiYgngTOAY1JKD1RKMkmSJElSXuR8HcOUUgKOAb4CTgeOSyndW1nBJEmSJEn5sdY9hhHx0loeagLMB/5fRPy/7LCUUtqzosNJkiRJkipfeYeSrgTSGobPzt4kSZIkSTXAWothSqlfHnNIkiRJkgok5+8YSpIkSZJqJouhJEmSJBU5i6EkSZIkFTmLoSRJkiQVOYuhJEmSJBW5tRbDiJgZETtkf745IrrkL5YkSZIkKV/K22PYCKiX/fkkoHWlp5EkSZIk5V15F7gfD5wWEavK4fYRUX9tI6eUXqrQZJIkSZKkvCivGF4G/BM4EUjAtWsZL7KP16rYaJIkSZKkfFhrMUwp3RwRTwJbAMOBc4EP8xVMkiRJkpQf5e0xJKU0GZgcEUOBx1NKn+cnliRJkiQpX8othquklE5e9XNENAY2AmallOZXVjBJkiRJUn7kfB3DiBgUEaOA2cAXwOyIeCMiBlZSNkmSJElSHuS0xzAiBgGPA+OAS4ApwMbAkcATEbFfSunZSkspSZIkSao0ORVD4CLgGWD/lNLKVQMj4mLgMeB3gMVQkiRJkqqhXA8l7QVcU7oUAmR/vxbYroJzSZIkSZLyJNdiuARoupbHmmQflyRJkiRVQ7kWwxeASyKiS+mBEdGRzGGmwys2liRJkiQpX3L9juH5wKvARxHxOjAZaAfsQuYspedXSjpJkiRJUqXLaY9hSuljoCdwNVAP2AGoD/wN2C6l9EmlJZQkSZIkVapc9xiSUpoMnFeJWSRJkiRJBZDzBe4lSZIkSTWTxVCSJEmSipzFUJIkSZKKnMVQkiRJkoqcxVCSJEmSilxOxTAiTo+IRpUdRpIkSZKUf7nuMbwOmBQR10REz8oMJEmSJEnKr1yL4WbAtcChwDsRMSIiToyI+pUXTZIkSZKUDzkVw5TSFymlXwIdgKOAhcDNwMSIuCoitqrEjJIkSZKkSrReJ59JKS1PKd2bUtoL6A68D5wLfBARL0bE4MoIKUmSJEmqPOt9VtKIaBIRZwP3A98D3gV+DdQGHomIiys0oSRJkiSpUuVcDCOid0TcCEwC/kKmEO6aUtoxpXRZSml34CLg/1VGUEmSJElS5cj1chVvAyOB/sDFQPuU0okppZFlRn0W2KhiI0qSJEmSKlPtHMebQOZw0adSSqmc8d4GumxwKkmSJElS3uR6KOmfgZfXVAojonFEfA8gpbQ0pTS+IgNKkiRJkipXrsVwOLD1Wh7rnn1ckiRJklQN5VoMo5zH6gErKiCLJEmSJKkA1vodw4joDHQtNah3RDQuM1oD4BTgy4qPJkmSJEnKh/JOPnMi8H9Ayt7+zrf3HKbs78vxEhWSJEmSVG2VVwxvAV4gU/6eJ1P+xpYZZwnwcUppZmWEkyRJkiRVvrUWw+zZRccDRER/4O2U0rx8BZMkSZIk5UdO1zFMKb1Y2UEkSZIkSYVR3slnPgMOSSm9FxGfk/lO4dqklNJmFZ5OkiRJklTpyttj+CIwt9TP5RVDSZIkSVI1Vd53DE8u9fNJeUkjSZIkScq7XC9wL0mSJEmqocr7juEJ6zOjlNKtGx5HkiRJkpRv67qOYa4SYDGUJEmSpGqovGLYJW8pJEmSJEkFs64L3EuSJEmSajhPPiNJkiRJRc4L3EuSJElSkfMC95IkSZJU5LzAvSRJkiQVOb9jKEmSJElFLudiGBHdImJoRHwcEQuy97dExOaVGVCSJEmSVLnK+47hahHRD3gCWAQ8DkwF2gIHAEdGxD4ppRcrKaMkSZIkqRLlVAyBvwDvAINSSvNXDYyIJsAz2cd7V3w8SZIkSVJly/VQ0q2By0uXQoCU0jzgcqBHRQeTJEmSJOVHrsVwAlB3LY/VBSZWTBxJkiRJUr7lWgwvB34XEZuUHhgR7YH/A/5Q0cEkSZIkSfmx1u8YRsStZQY1BT6LiNf55uQzu2R/3hO4ubJCSpIkSZIqT3l7DL8H7FHqthyYDHQC+mTvJwMrs4/nJCI6RMR9ETEnIuZGxAMR0XF9g0fEBRGRIuKV9Z1WkiRJkvSNte4xTCl1rugni4iGwPPAEuBEIAGXAsMjomdKaUGO8+kK/Ab4uqIzSpIkSVKxyfVyFRXlNKAr0D2lNA4gIkYDnwBnAFfmOJ/rgDuA7uR/GSRJkiSpRsn15DOrRUSbiOhY9pbj5AcCr68qhQAppc+BV4GDcnz+Y4AdgF+ub3ZJkiRJ0v/KaW9bRJSQOeTzDKD5WkarlcOsegAPr2H4GODwHHJsBFwF/CKlNDMicnhKSZIkSVJ5ct1j+GPg/wF/AYLM5SkuBT4HPiVziGguWgCz1jB8JrBRDtNfAXwM3JLj80mSJEmS1iHXYngycDGZ6xkCPJhS+j9gKzIXt1/vs4qur4jYAzgBOCullNZjutMjYlREjJo2bVrlBZQkSZKkairXYtgVGJVSWkHmshUNAFJKy4C/AqfkOJ9ZrHnP4Nr2JJb2T+AmYEJENI+I5mQOha2V/b3emiZKKd2QUuqdUurdunXrHGNKkiRJUvHItRjOAepnf55E5mygq9QmU+xyMYbM9wzL2hoYu45ptwLOJFMgV912B3bJ/nxWjhkkSZIkSaXkeqmHd8iUt6ezt99FxCIyew9/D7yd43weAf4cEV1TSp8BRERnMgXvgnVM238Nw/5K5qQ3PwTGreFxSZIkSdI65FoM/0rmcFKA/yNzuYg7sr+PB87JcT43Zsd9OCJ+Q+YC95cAX5E5VBSAiOhE5qQ2F6eULgZIKb1QdmYRMRuovabHJEmSJEm5yakYppSeLfXzlIjoA2wGNAQ+zH7XMJf5LIiIAWQuOXEbmTOcPgf8OKU0v9SoQWZP4HpfZ1GSJEmStH5y3WP4Ldmzgn6nQzdTSl8CQ9YxzhdkyuG65tXvu2SQJEmSJH0j52KYPQvoT4BdgfZkLlPxGvDXlNLsyggnSZIkSap8OR2qGRG9gE+AX5I5O+nY7P2vgI8jYttKSyhJkiRJqlS57jG8GpgB9E4pjV81MHtG0aeAvwP9KjqcJEmSJKny5Xpyl52A35YuhbD6u4D/B/Sp4FySJEmSpDzJtRjOAJas5bHF2cclSZIkSdVQrsXwOuDnEVG/9MCIaACcB1xT0cEkSZIkSfmx1u8YRsTFpX8FOgFfRsQTwFSgLbAfsIjM9QwlSZIkSdVQeSef+c1ahp+whmG/Bi7c8DiSJEmSpHxbazFMKeV6mKkkSSqAZSuWMXPRTGYsmsH0hdOZvnA6n035jJmLZ7K01lKmL5y++rGZi2aybMUyACKCIKhbqy71atejbq26NK7bmHaN29GuUbvMffbWtnFb2jVuR6uGrSgJPxpIUk2V8wXuJUlS5Vm6YikzFs5YXeRmLPym7K0eVuaxOUvmrHV+Des0pGWDlrRq2IqWDVvSqVkn6taqSyIBsDKtZOmKpSxZvoQlK5Ywb8k8Xp/wOpPnTWbR8kX/M79aUYs2jdp8uzQ2ypTGLht1oVfbXnRs1pGIqLR1JEmqPOtVDCNif2BPoAUwE3ghpfR4ZQSTJKmmmbloJv+d/t9v3T6a8RGT501m3tJ5a52uUZ1GtGrYanXJ27zF5t+UvlLlr1XDVqycv5KN6m9El027fKeMKSXmL53PlPlTvnWbumDqt34fPXU0UxdMZfnK5aunbV6/OT3b9qRX216ZW7te9GjdgwZ1GnynLJKk/MmpGEZEE+AxYA9gOZnLU7QEfhoRLwP7p5TmV1pKSZKqiZVpJeNnj+fD6R/+TwmctnDa6vHq1qrLFi23YNs227Lv5vuuLnilS17LBi1p2bAl9WvXL+cZv23KlCkblD8iaFKvCU3qNaFby27ljrsyrWTGwhmMmzmO96a+x7tT3uW9qe9x8zs3s2DZAgBql9Rmh413oG+HvvTt2Jc9Ou1Bq4atNiijJKni5brH8A/ADsDxwF0ppRURUQs4isylLP4AnFs5ESVJqpoWLVvEmGljMoVoynu8OzVzX3rvX+uGrdmy1ZYcvOXBbNlqy9W3Ts06UaukVgHTb7iSKKF1o9a0btSaXTvsunr4yrSST2d+yntT3+PtyW/z6levcs2b13Dl61dSEiXs2WlPhmw1hEO2OoRNmmxSwCWQJK2SazEcAvwmpXTHqgEppRXAHRHRCvgFFkNJUg22ZPkS3pv6Hm9MfIM3J73JW5Pe4r/T/8uKtAKAJnWb0KtdL07odQK92vaiR5sedG/ZnZYNWxY4ef6VRAndWnajW8tuHLb1YUBm/b01+S2eGvcU9429j3OePIcfPvlDduuwG4dtfRiHbnUoHZt1LHBySSpeuRbDlsDYtTw2Nvu4JEk1wsq0ko+mf7S6BL4x8Q3enfIuy1ZmzurZrnE7em/Sm0O3OpTt2m1Hr7a96LJRF8/aWY56teuxW4fd2K3Dblzc/2LGThvL/WPv5/4P7+cnT/+Enzz9E/q078OQrYYwZKshNKJRoSNLUlHJtRh+DuwPPLuGx/bLPi5JUrU0beE0Rk0dxccffMwbk95g1KRRzF0yF8jsCey9SW9+uutP6dO+D33a96F9k/aefXMDbd16a7bec2t+u+dvGTdzHPePvZ/7PryP84edz/nDzmebltswuOtgTupzElu22rLQcSWpxsu1GP4T+EtENAbuACYD7ch8x/BU4KeVE0+SpIq1Mq3kw2kf8sqXr/DqV6/y2lev8emsTwGoU1KHXu16cdy2x7FT+53o074P3Vt2r/bfBazqNm+xOef3PZ/z+57PF7O/4IEPH+Cu9+7i8jcv5/I3L6dH6x4M2WoIx2x7DN1bdS90XEmqkXIqhimlqyKiNZkCeFJ2cABLgctSSn+rnHiSJG2YpSuW8takt3jly1d4+cuXefWrV5m5aCYAbRq1YfcOu3Ns92Pp3bY3A7cZuF5nAFXF69y8Mz/d9acc0+UYJi+YzKszXuW+sfdx6cuXcvFLF7Pv5vvyk11+wt5d93avrSRVoFwvV9EMuBi4AtiFb65j+HpKaVblxZMkaf3MXTKXEV+N4OUvX+aVL19h5MSRLF6+GIBuLbpxUPeD2KPjHvTt2JfNW2xORKy+xIOlsGrZuNHGnLPZOZzT5xymzJ/CDW/dwDVvXsP3b/8+27TZhh/v/GOO7XmsfzdJqgDrLIYRUZvMdQsPSSk9CjxZ6akkScrRtAXTePnLl3lp/Eu8/OXLvDvlXVamlZRECdu3254zdzyTvh0z19Br27htoePqO2rXuB0X7nkh5+9+Pnd+cCdXvX4Vpz56Kr987pec1fsszt7pbP++krQB1lkMU0rLI2IqsCIPeSRJKteEuRN4afxLq28fTv8QyOzt23XTXfnNHr9hj057sHP7nWlSr0mB06qi1atdj5O2O4kTe53I8C+Gc9XrV3HxSxdz+auXc9oOp/HLPX7ptREl6TvI9eQzt5M5ycwTlZhFkqRvSSnx4fQPeXn8y7zy1Su8PP5lxs8ZD0DTek3p27EvJ/Q6ge91+h69N+lN3Vp1C5xY+RIRDOgygAFdBvDR9I+44rUruP6t67nx7Rs5Y8czOL/v+RZESVoPuRbDL4BjIuJN4GEyZyVNpUdIKd1csdEkScVm6YqlvD357W9OFPPlq8xYNAOAto3askenPfjJLj/he52+R8+2PT1bqADo3qo7/zrwX/xqj1/xh5f/wDVvXsONb9/Iebudxy92/wWN6zYudERJqvJyLYbXZO/bAzuu4fEEWAwlSetl3pJ5jJgwYnURHDlhJIuWLwK+OVFM34592aPTHmy20WaehVLl6rpR19UF8dfP/5pLXrqEG9++kUv7X8pJ253kfyRIUjlyLYZdKjWFJKnGW5lW8vGMj3lz4pu8OelNXv3q1f85UczpO56++oyhnkhE31XXjbpy55A7+dHOP+KnT/+UUx89lavfuJo/D/wzAzcbWOh4klQl5VoMFwDzU0qLKzOMJKlmSCnx1dyvVpfANya+wVuT32LukrkANKrTiJ3a78Sv9/g1e3Tcg1023cUTxajC7bLpLrx6yqvcO/ZeLhh2Ad+//fvs120/rhh4BVu33rrQ8SSpSllrMYyIWsBvgR8BTYEVEfEo8IOU0uz8xJMkVQfTF05fXQJXFcGvF3wNQJ2SOvRs25Njtz2WnTbZiZ3a78RWrbbysD7lRURwRI8jOKj7Qfz9jb9z6UuX0vO6npy787n8rt/v/A8JScoqb4/hmcCFwAvAm0BX4BBgLnBypSeTJFVJ85fO561Jb/H8R8/z7tfv8v7M9/l89ucABMGWrbZk3833XV0Ce7XtRb3a9QqcWsWuXu16nLfbeZy03Un85vnf8NfX/8q9Y+/l6n2u5uAtD/b7q5KKXnnF8DTgxpTSGasGRMQZwD8i4oyU0tJKTydJKpiUEhPnTeT9qe/z/tfvM3rqaN6d8i4fTv+QlWklAJs23pRdOu7CWb3PYqf2O7HDxjvQtF7TAieX1q5Vw1Zcv//1nLTdSZz52Jkces+hHNj9QG484EbaNGpT6HiSVDDlFcOuwHllht0NXAd0Aj6prFCSpPxZsXIFX8z+go9mfMRH0z9i7LSxjJk2hrHTxjJnyZzV43Vo2oGebXty2NaH0ad9HzrW7kirBq1o165dAdNL380um+7CqNNH8bfX/8Zvhv+GXtf34rZDbmPvrnsXOpokFUR5xbAxmcNGS5uXvfeAfEmqRpauWMqEuRMYP3s8X8z+go9nfJwpgjM+YtzMcSxd8c1BIC0btGSbNttw7LbH0qNND7Ztsy3btNmGjRps9K15TpkyJd+LIVWo2iW1+dluP2PQ5oM48r4j+f5t3+eCvhfwu36/o06tOoWOJ0l5ta6zkraPiK6lfq9Vavjs0iOmlD6ryGCSpNzNWzKP8XPGM372eMbPGc+Xc7781u+T500mkVaPX7ukNptttBndW3Vn/277071Vd7q37E73Vt1p1bBVAZdEyr9t2mzDm6e9yY+f+jF/fOWPDP9iOHcOuZPOzTsXOpok5c26iuF9axn+0BqGeXo5SaoEy1cu5+sFX6/e41e2+H0550tmLZ71rWnqlNShQ7MOdGrWie9v9n06Nu1Ip+ad6NSsEx2bdaRz887uEZFKaVinITcccAN7d92b0x49je2u344bD7iRw3scXuhokpQX5RVDzzwqSZVg4bKFTJg7gZmLZjJj4QxmLpqZ+XnRDL5e8PXq29QFU5kyfwrTFkz71t4+gCZ1m6wuert32J2Ozb4pfp2ad6Jd43aUREmBllCqvo7ocQQ7bbITR99/NEfcdwSnf3Y6V+1zFQ3rNCx0NEmqVGsthimlofkMIknV1ao9epPnTWby/Mmr76fMn8KU+VNWl76Zi2Yyc+FMFq9YvNZ5Na7bmDaN2tCmURs6N+/MLu13oV3jdmzcZGM2abLJ6uLXvH7z/C2gVGS6bNSFl09+mQuHX8jlr17Oq1+9yl2H3cU2bbYpdDRJqjTrOpRUkorWwmULmTwvU/BKF75VpW/V72vaoweZk7i0bdyWVg1bsXmLzWnZoCX1Vtajeb3mdG7TmRYNWtCyYcvMfYPMfYM6DQqwpJLKqlOrDn/c+48M6DKA4x88np1u3Im/Dvorp+94utc8lFQjWQwlFZ1lK5YxbuY4Js6b+K2Ct6r8rSqCc5eUPTFz5qQtbRu1ZeMmG9OhWQf6tO/Dxo03ZuMmG7Nx441X791r26jtGi/qvupMnl7iQaoeBm42kPfOfI8THjqBMx8/k2c/e5YbD7jxf87SK0nVncVQUo22bMUyxkwbw1uT3uKtyZnbe1PeY8mKJd8ar2GdhqsLXs+2PRm02aDVJW/V8HaN29GqYSu/uycVmbaN2/LksU/yl9f+wq+e/xWjJo3iziF3smuHXQsdTZIqjMVQUo3z2azPuOntm3j2s2cZPXX06hLYtF5Tdth4B87pcw7btduODk07rC5+Tep5eVZJa1cSJfx895+zZ+c9Ofr+o9nj33twcf+LOX/386lV4onZJVV/FkNJNcLSFUt5+L8Pc8PbNzDss2GURAl9O/blnD7n0HuT3uy48Y5s1mIz9/ZJ2iB92vfh7dPf5ozHzuDXz/+a5z9/ntsOuY2Nm2xc6GiStEEshpKqtU9mfMK/3v4X/37330xbOI2OzTpycb+LOXn7k9m06aaFjiepBmpWvxl3DrmTgV0H8sMnf0iv63sx9OCh7Ntt30JHk6TvzGIoqdpZsnwJD/73QW58+0ae//x5akUtDuh+AKfvcDrf3+z7HtYlqdJFBD/Y4Qfs1mE3jrr/KPb7z378dJef8se9/0jdWnULHU+S1pvFUFK1MW72OP48+s/c8u4tzFg0g87NO/P7Ab/n5O1O9jAuSQWxVeutGHnqSM575jyufP1K3pj0Bg8f9TAtGrQodDRJWi8WQ0lV2uLli7l/7P1c8/o1jJg8gtoltTmo+0GcvuPp7N11b78zKKng6teuzz/2+wd9O/blxIdOZLebduOJY5+g60ZdCx1NknJmMZRUJY2dNpYb37qRW0ffysxFM+nUtBO/6vMrfrjHD2nX2GsASqp6jtrmKNo3ac9Bdx3ELv/ahceOeYw+7fsUOpYk5cRiKKnKWLRsEfeNvY8b3r6BV758hToldThkq0M4bYfT2LrB1pREiaVQUpW2R6c9GPGDEex7x770u6Ufdw65k52b71zoWJK0Th6DJangPvj6A8598lw2uXITTnjoBKbMn8Kf9v4TE346gbsPu9tDRiVVK91bdWfED0awbdttOeTuQ7j5g5sLHUmS1sk9hpIKYuGyhdwz5h5ueOsGRkwYQd1adTl0q0M5fYfT6de5HxFR6IiS9J21bdyW4ScO55j7j+HXr/6ar+Z9xTUHX+N/ckmqsiyGkvLqo+kf8fc3/s7to29nzpI5bNFyC/488M+c0OsEWjdqXeh4klRhGtZpyP1H3M8ZD57B9aOvZ/ry6dx68K00qNOg0NEk6X9YDCXlxbwl87j4xYv568i/UitqcdjWh3H6jqezR8c93DsoqcaqVVKLS3a7hA5NOvC7Eb9j4tyJPHL0I7Rq2KrQ0STpWyyGkipVSon/vP8ffv7sz5k8fzI/2P4H/H7A72nbuG2ho0lSXkQEZ/Q8g2023YbjHjyOXW/alSePfZLNW2xe6GiStJoHukuqNPOXzueI+47guAePo33T9ow8dST/OvBflkJJRWnI1kN4/oTnmbVoFrvetCuvT3i90JEkaTWLoaRK8dmsz9jtpt144MMH+NPef2LkqSO9npekordrh10Z8YMRNK/fnL1u3YtnPn2m0JEkCbAYSqoEwz4bRu8bejNh7gSeOvYpfr77zz0TnyRldWvZjVdOfoVuLbqx/3/2576x9xU6kiRZDCVVnJQSV464kkG3D6J90/a8edqbDNxsYKFjSVKV07ZxW1446QX6tO/Dkfcdyc3veK1DSYVlMZRUIRYtW8QJD53Az575GQdveTAjfjCCzVpsVuhYklRlNa/fnKePe5q9u+7NDx75AX97/W+FjiSpiFkMJW2wr+Z8xR7/3oM7Rt/BJf0v4d7D76Vx3caFjiVJVV6juo145KhHOGTLQ/jx0z/mL6/9pdCRJBUpL1chaYO8PP5lDrv3MBYtW8TDRz3MAd0PKHQkSapW6tWux92H3c2xDxzLec+ex/KVyzm/7/mFjiWpyFgMJX0nKSWuH3U95z51Ll036sqLJ73Ilq22LHQsSaqW6tSqw3+G/IdaJbW44LkLWL5yOb/+3q8LHUtSEbEYSlpvS5Yv4YdP/pAb376R/brtxx2H3kHz+s0LHUuSqrXaJbW57ZDbqF1Sm98M/w3LVy7nwj0vJCIKHU1SEbAYSlovk+dN5rB7D+O1r17jV31/xcX9L6ZWSa1Cx5KkGqF2SW1uOegWapfU5qIXL2L5yuVc3P9iy6GkSmcxlJSzNya+wSF3H8LsxbO557B7OLzH4YWOJEk1Tq2SWtx04E3Ujtpc+vKlLF+5nD/s9QfLoaRKZTGUlJNb3r2FMx87k42bbMxrp7xGr3a9Ch1Jkmqskijhnwf8k9oltbns1ctYtnIZVwy8wnIoqdJYDCWVa9mKZZz3zHlc/cbV7NVlL+4+7G5aNmxZ6FiSVOOVRAnXDr6W2iW1+cuIv7B85XKuGnSV5VBSpbAYSlqr6Qunc8S9RzD8i+H8ZJef8KeBf6J2iW8bkpQvEcHV+15NrZJa/G3k31i+cjl/3/fvlkNJFc5PeJLW6N0p73LwXQczZf4Uhh48lBN6nVDoSJJUlCKCqwZdRZ2SOvx5xJ9ZsXIF1wy+hpIoKXQ0STWIxVDS/7j7g7s5+eGTadmwJa+c8gq9N+ld6EiSVNQiYvVRG5e9ehmJxHWDr3PPoaQKYzGUtNqKlSv49fO/5vJXL2f3Drtz/xH307Zx20LHkiSRKYerzk76x1f+SMM6DfnL9/9iOZRUISyGkgCYtWgWxzxwDE+Ne4ozdjyDq/e9mrq16hY6liSplIjg9wN+z4KlC7jq9atoVKcRlwy4pNCxJNUAFkNJjJ02loPuOojxs8dz/eDrOaP3GYWOJElai4jgqn2uYsGyBVz68qU0qtuIC/peUOhYkqo5i6FU5B7+78Mc9+BxNKrTiOdPfJ6+HfsWOpIkaR1KooR/7v9PFi1fxC+f+yUN6zTk3J3PLXQsSdWYxVAqUivTSi558RIuevEidtpkJx448gE2bbppoWNJknJUq6QWtxx0CwuXLeRHT/2IRnUa8YMdflDoWJKqKc9zLBWheUvmMeSeIVz04kWc0OsEXjr5JUuhJFVDdWrV4a4hd7HP5vtw2qOncef7dxY6kqRqymIoFZlxM8exy0278OhHj/LXQX/lloNuoX7t+oWOJUn6jurVrsf9R9zP9zp9j+MfPJ6H/vtQoSNJqoYshlIReWrcU+x0405MmT+Fp497mh/t8iNPcy5JNUDDOg159OhH2an9Thx535E8Pe7pQkeSVM1YDKUikFLiT6/+icH/GUzHZh0Zddoo9uq6V6FjSZIqUJN6TXjy2CfZuvXWHHz3wbz4xYuFjiSpGrEYSjXcwmULOfaBYzl/2PkM2WoIr53yGl026lLoWJKkStC8fnOeOe4ZujTvwv537s/rE14vdCRJ1YTFUKrBvpr3FQc9fBB3fXAXf9zrj9x92N00qtuo0LEkSZWodaPWDDthGG0btWXfO/bl3SnvFjqSpGrAYijVUB/P+JgDHjqAL+d9yWPHPMYFfS/w+4SSVCQ2abIJz53wHE3qNmHgbQMZO21soSNJquIshlIN9MmMT+g/tD/LVy7nkYMeYb9u+xU6kiQpzzo178RzJzxH7ZLa7H3r3nw689NCR5JUhVkMpRpm3Mxx9B/an2UrlnHfAffRvUX3QkeSJBVIt5bdGHb8MJauWMpet+7Fl3O+LHQkSVWUxVCqQT6d+Sn9h/ZnyYolPHfCc2zZYstCR5IkFViPNj145vhnmL14NnvfujdT5k8pdCRJVZDFUKohPpv1Gf2H9mfRskUMO34Y27bdttCRJElVxA4b78CTxz7JpHmT2PvWvZm+cHqhI0mqYiyGUg3w+azP6T+0PwuWLWDYCcPo1a5XoSNJkqqYXTvsyqNHP8qnsz5l0O2DmL14dqEjSapCLIZSNffF7C/oP7Q/85bMY9jxw9iu3XaFjiRJqqL6d+nP/Ufcz/tT32e/O/Zj/tL5hY4kqYqwGErV2PjZ4+k/tD9zlsxh2AnD2H7j7QsdSZJUxe3XbT/uHHInIyeO5MA7D2TRskWFjiSpCsh7MYyIDhFxX0TMiYi5EfFARHTMYbreEXFDRPw3IhZGxJcRcUdEdMlHbqmq+XLOl/Qf2p/Zi2cz7Phh7LDxDoWOJEmqJoZsPYShBw/lhS9e4Mj7jmTZimWFjiSpwPJaDCOiIfA8sCVwInA80A0YHhGN1jH5UUAP4GpgX+ACYAdgVER0qLTQUhU0Ye4E+g/tz8xFM3nmuGfYcZMdCx1JklTNHNfzOK4dfC2PfvwoJz50IitWrih0JEkFVDvPz3ca0BXonlIaBxARo4FPgDOAK8uZ9vKU0rTSAyLiVeDz7HwvrJTEUhUzce5E+t3Sj+kLp/Ps8c+yU/udCh1JklRNndn7TOYsnsMFz11Ak7pNuH7/64mIQseSVAD5LoYHAq+vKoUAKaXPswXvIMophmVLYXbY+IiYBrSvjLBSVTNx7kT6D+3P1wu+5pnjn6FP+z6FjiRJqubO73s+c5bM4Y+v/JFm9Ztx+d6XWw6lIpTvYtgDeHgNw8cAh6/vzCJiK6AN8OEG5pKqvEnzJjHg1gFMnj+ZZ457hl023aXQkSRJNcTvB/yeuUvmcsVrV9CsXjN+/b1fFzqSpDzLdzFsAcxaw/CZwEbrM6OIqA1cD0wDbtrwaFLVNXneZAYMHcCkeZN46tin2LXDroWOJEmqQSKCq/e9mrlL5vKb4b+hab2m/HDnHxY6lqQ8yncxrEj/AHYDBqeU1lQ2AYiI04HTATp2XOfJT6UqZ8r8KQy4dQAT5k7gqeOeYveOuxc6kiSpBiqJEm4+6GbmLZ3HuU+dS9N6TTlxuxMLHUtSnuT7chWzWPOewbXtSVyjiLiMTNk7JaX0THnjppRuSCn1Tin1bt269XqFlQpt6vypDBg6gK/mfMWTxz5J3459Cx1JklSD1S6pzV1D7mLvrntzyiOncP/Y+wsdSVKe5LsYjiHzPcOytgbG5jKDiPg1cD5wbkrptgrMJlUpXy/4mr1u3Yvxc8bz+DGPs0enPQodSZJUBOrVrsdDRz7Ezu135uj7j+bpcU8XOpKkPMh3MXwE2CUiuq4aEBGdgd2zj5UrIs4FLgV+nVL6R2WFlApt2oJp7HXrXnw26zMeP+Zx9uy8Z6EjSZKKSKO6jXji2CfYuvXWHHL3Ibzy5SuFjiSpkuW7GN4IfAE8HBEHRcSBZM5S+hXwz1UjRUSniFgeEReWGnYU8FfgKeD5iNil1G3rfC6EVJmmL5zOXrfuxbiZ43jsmMfo17lfoSNJkopQ8/rNeeb4Z+jQrAOD/zOYtye/XehIkipRXothSmkBMAD4GLgNuIPMBeoHpJTmlxo1gFpl8u2THb4PMKLM7dpKDy/lwcxFMxl420A+mfkJjx79KAO6DCh0JElSEWvTqA3Djh9G8/rNGXT7ID6a9VGhI0mqJPneY0hK6cuU0pCUUtOUUpOU0sEppS/KjPNFSilSSheVGnZSdtiabv3yvBhShZuzeA6Dbh/E2GljeejIh9i7696FjiRJEh2adWDY8cOoXVKbwx89nHGzxxU6kqRKkPdiKOl/zVsyj33v2Jf3przH/Ufcz6DNBxU6kiRJq3Vr2Y3nT3ieRMqUw5mWQ6mmsRhKBbZg6QL2v3N/3pj4Bncfdjf7b7F/oSNJkvQ/tmq9Fffufy9LVy6l/9D+fDbrs0JHklSBLIZSAS1ctpAD7zqQV758hdsPvZ1Dtjqk0JEkSVqrLVtsyT3738PCZQvpP7Q/X8z+otCRJFUQi6FUIIuXL+bguw5m+OfDGXrwUI7a5qhCR5IkaZ16tOzBs8c/y9wlcxkwdABfzfmq0JEkVQCLoVQAS5YvYcg9Q3j2s2e56cCbOK7ncYWOJElSznbYeAeeOe4ZZiyaQf+h/Zk4d2KhI0naQBZDKc+WLF/C4fcezhOfPMEN+9/AydufXOhIkiStt53a78TTxz3N1wu+ZsCtA5g8b3KhI0naABZDKY8WL1/MofccyqMfP8q1+13LaTueVuhIkiR9Z7tsugtPHvskE+dOZK9b92Lq/KmFjiTpO7IYSnmyePliDrn7EJ745An+uf8/OWunswodSZKkDbZ7x9154tgnGD9nPHvduhfTFkwrdCRJ34HFUMqDRcsWceCdB/L0uKf51wH/4vQdTy90JEmSKsz3On2PR49+lE9nfcret+3NjIUzCh1J0nqyGEqVbOGyhex/5/4M+2wYNx90Mz/Y4QeFjiRJUoUb0GUAjxz1CB9N/4iBtw1k1qJZhY4kaT1YDKVKtGDpAgb/ZzAvfPECtx5yKydtd1KhI0mSVGkGbjaQB498kDHTxjDo9kHMWTyn0JEk5chiKFWSeUvmse8d+/LS+Je47ZDbvCSFJKko7NttX+47/D7enfIu+9yxD3OXzC10JEk5sBhKlWDukrnse8e+vPbVa/zn0P9wzLbHFDqSJEl5c0D3A7jn8HsYNWkU+92xH/OXzi90JEnrYDGUKticxXMYdPsgRk4cyd2H3c2R2xxZ6EiSJOXdwVsezJ1D7uT1Ca8z+D+DWbB0QaEjSSqHxVCqQLMXz+b7t3+ftya9xb2H38uQrYcUOpIkSQVz2NaHcfuht/PKl69w4F0HsnDZwkJHkrQWFkOpgsxcNJO9b92bdya/w/1H3M/BWx5c6EiSJBXcUdscxdCDhzL88+EcfNfBLF6+uNCRJK2BxVCqADMWzmDvW/fm/a/f58EjH+SA7gcUOpIkSVXGcT2P4+aDbmbYZ8M49O5DWbJ8SaEjSSrDYihtoOkLp7PXrXsxdtpYHj7qYQZvMbjQkSRJqnJO2u4kbjjgBp4c9ySH33s4S1csLXQkSaVYDKUN8PWCrxkwdAAfzfiIR49+lH0236fQkSRJqrJO3eFUrt3vWh79+FGOuu8olq1YVuhIkrIshtJ3NHX+VPoP7c+4meN4/JjHGbjZwEJHkiSpyjtrp7O4ep+refC/D3LsA8eyfOXyQkeSBNQudACpOvp05qfsc8c+TJo3iSeOfYJ+nfsVOpIkSdXGD3f+IctWLuNnz/yM2iW1ue2Q26hVUqvQsaSiZjGU1tOqi/WuTCsZdvwwdu2wa6EjSZJU7fx015+yfOVyzh92PgBDDx5KnVp1CpxKKl4WQ2k9PPlJ5gvzrRu15qljn6J7q+6FjiRJUrX1i91/AcD5w85n4bKF3H3Y3dSrXa/AqaTi5HcMpRz9+51/c8CdB7BFyy0Y8YMRlkJJkirAL3b/BX/f9+88/NHDHHjXgSxctrDQkaSiZDGU1iGlxCUvXsIpj5zCXl334sWTXqRd43aFjiVJUo1xTp9zuPnAzHUO97l9H+YumVvoSFLRsRhK5Vi+cjlnPnYmF75wISf0OoHHjn6MJvWaFDqWJEk1zsnbn8x/Dv0PIyaMoP/Q/kxbMK3QkaSiYjGU1mLhsoUceveh3PD2Dfyq76+45aBb/FK8JEmV6MhtjuShIx9i7LSx7PHvPfhqzleFjiQVDYuhtAbTF01nwNABPP7J41y737X8fq/fExGFjiVJUo03eIvBPHPcM0yeP5ndb96dj2d8XOhIUlGwGEpljJ87noMePoj3pr7H/Ufcz1k7nVXoSJIkFZU9Ou3BCye+wOLli9ntpt147rPnCh1JqvEshlIpb016i/0f2p9Zi2fx3AnPcfCWBxc6kiRJRWn7jbfn1VNepW3jtnz/9u/zp1f/REqp0LGkGstiKGU9Ne4p9rxlT+rXqs/DBz3Mbh12K3QkSZKKWreW3Rh56kiGbDWE84edz+H3Hs68JfMKHUuqkSyGEnDLu7dwwJ0H0K1lNx47+DG6bdSt0JEkSRLQuG5j7j7sbv488M889N+H6POvPvx3+n8LHUuqcSyGKmopJX7/0u85+eGT6d+5Py+e9CJtG7UtdCxJklRKRPCz3X7Gs8c/y4yFM+hzYx8e+PCBQseSahSLoYrWkuVLOPWRU/nN8N9wXM/jeOyYx2har2mhY0mSpLXo36U/b5/xNlu13ooh9wzhl8N+yYqVKwodS6oRLIYqSlPmT6H/0P7c/O7N/PZ7v+XWg2+lbq26hY4lSZLWYdOmm/LSSS9xxo5ncNmrl7HPHfswfeH0QseSqj2LoYrOqEmj6H1Db96b+h73HHYPF/e/2GsUSpJUjdSrXY/r97+emw68iZfHv8yON+zIqEmjCh1LqtYshioaKSVufOtG9vj3HtQqqcWrp7zK4T0OL3QsSZL0HZ2y/Sm8csorAPS9uS83v3NzgRNJ1ZfFUEVh7pK5HPPAMZz+2On07diXN097k+3abVfoWJIkaQP13qQ3b53+Fnt02oMfPPIDznj0DJYsX1LoWFK1YzFUjffWpLfY4Z87cM+Ye/j9gN/z9HFP06ZRm0LHkiRJFaRVw1Y8dexTXLD7Bdzw9g1875bvMWHuhELHkqoVi6FqrJQSV4+8ml1v2pXFyxfzwokv8Ks9fkVJuNlLklTT1CqpxR/3/iP3H3E/Y6eNZYd/7sDwz4cXOpZUbfgJWTXSzEUzOfSeQ/nRUz/i+5t9n3fPfJc9Ou1R6FiSJKmSHbrVobx52pu0bNiSgbcN5C+v/YWUUqFjSVWexVA1zoivRrD9P7fn8Y8f5y/f/wuPHv0orRq2KnQsSZKUJ1u22pI3Tn2Dg7c8mPOePY+j7j+K+UvnFzqWVKVZDFVjrEwr+dOrf8qcdTRq8copr/DTXX/qpSgkSSpCTeo14d7D7+XyvS/nvrH3sfO/dubjGR8XOpZUZVkMVSNMWzCNwf8ZzPnDzueQrQ7h7TPepk/7PoWOJUmSCigi+MXuv+CZ457h6wVfs9ONO/Hwfx8udCypSrIYqtp78YsX2e6f2zH88+Fcu9+13HPYPTSv37zQsSRJUhWxV9e9eOv0t9ii5RYcfPfBXPbGZaxYuaLQsaQqxWKoamvFyhVc/OLFDLh1AI3rNub1U1/nrJ3O8tBRSZL0Pzo268jLJ7/Mqdufyt/e+RvHPXkcMxbOKHQsqcqwGKpamjxvMgNvG8j/vfB/HL3N0Yw6bZQXrJckSeWqX7s+Nx54I1d87wpem/QavW/szduT3y50LKlKsBiq2rl3zL30ur4XIyeO5OYDb+a2Q26jSb0mhY4lSZKqieO2Oo6HDnqI5SuXs/vNuzP03aGFjiQVnMVQ1caU+VM47J7DOOK+I+jYrCNvnvYmJ29/soeOSpKk9bZ9m+156/S32HXTXTnp4ZM4+/GzWbpiaaFjSQVjMVSVl1Li9tG30+PaHjz28WNcttdlvH7q62zdeutCR5MkSdVYm0ZteOb4Z/j5bj/nulHX0e+WfkycO7HQsaSCsBiqSpswdwIH3HkAxz94PN1bdufdM9/l/L7nU7ukdqGjSZKkGqB2SW3+NPBP3HPYPYyeOpodbtiBl8a/VOhYUt5ZDFUlpZT419v/ose1PXj+8+f566C/8vLJL7Nlqy0LHU2SJNVAh/c4nDdOe4Pm9ZszYOgA/vr6X0kpFTqWlDcWQ1U5Y74eQ/+h/Tnt0dPYYeMdeP+s9/nRLj+iVkmtQkeTJEk12Natt+bN097kgO4H8JOnf8IxDxzDgqULCh1LyguLoaqMeUvmcd4z59Hr+l68//X73LD/DTx3wnNs1mKzQkeTJElFomm9ptx/xP38YcAfuGfMPexy0y6Mmzmu0LGkSmcxVMGllLjrg7vY8potuXLElZyy/Sl8dM5HnLbjaZSEm6gkScqvkijhl3v8kqeOfYpJ8ybR+4beXPfmdSxbsazQ0aRK46duFdTYaWPZ69a9OPr+o9m48caM+MEIbjjgBlo1bFXoaJIkqcgN3Gwgb53+Fr3a9eLsJ86mx7U9uHfMvX73UDWSxVAFMWfxHH7+zM/pdX0v3p3yLtcNvo6Rp45k5013LnQ0SZKk1To378wLJ77Ao0c/Sr3a9TjiviPY+V8788IXLxQ6mlShLIbKq+Url3Pdm9ex+d835y8j/sIJPU/go3M+4szeZ3pyGUmSVCVFBPtvsT/vnvEu/z7o30yeP5n+Q/uz3x37MXrq6ELHkyqExVB5kVLisY8fo+d1PTOHYrTuwajTR3HTQTfRulHrQseTJElap1oltThpu5P4+JyPuWLgFbw+4XW2u347TnjwBL6Y/UWh40kbxGKoSvfiFy/S9999OeDOA1i2chkPHvkgw08czg4b71DoaJIkSeutQZ0GnLfbeXx67qf8YvdfcO/Ye+n+j+789OmfMmPhjELHk74Ti6EqzVuT3mLQ7YPoN7QfX8z+gusHX8/Ys8dy8JYHExGFjidJkrRBNmqwEZftfRkfn/Mxx217HH8b+Te6Xt2VP778RxYuW1joeNJ6sRiqwn047UMOu+cwet/Ym1GTRnHFwCsY98NxnNH7DOrUqlPoeJIkSRWqQ7MO3HTQTYw+czT9OvfjV8//im5/78aNb93I8pXLCx1PyonFUBXmo1kfcdJDJ7HNddvw9KdPc+H3LuTzH33OebudR4M6DQodT5IkqVL1aNODh496mJdPfpnOzTtz+mOns8212/Dghw96iQtVeRZDbZCUEsM/H85xTx5Hv3v6cfeYu/nRzj/is3M/43f9f0fTek0LHVGSJCmv+nbsyysnv8KDRz5IRHDoPYey+8278/L4lwsdTVori6G+k+Url3Pn+3fS+8beDLh1AO9+/S4/7/1zvvzxl1w56ErPNCpJkopaRHDwlgfz/lnvc+MBNzJ+zni+d8v3OPDOAxnz9ZhCx5P+h8VQ62Xeknn89fW/stnVm3HMA8ewYOkCbtj/Bt489k1+uuNPLYSSJEml1C6pzak7nMonP/yEP+71R14a/xI9r+/JKQ+fwldzvip0PGk1i6FyMnHuRC4YdgEdrurAT57+CZ2bd+aRox5h7P8by2k7nkaD2n6HUJIkaW0a1mnIBX0v4NNzP+XHO/+YO96/gy3+sQXnP3s+sxbNKnQ8yWKo8r0/9X1OeugkuvytC1e8dgWDNh/EyFNH8uJJL3JA9wMoCTchSZKkXLVs2JK/DPoLH5/zMUf0OIIrXruCrld35YpXr2DRskWFjqci5qd6/Y+UEsM+G8Y+t+9Dz+t7ct/Y+zir91mM++E47j7sbvq071PoiJIkSdVap+adGHrwUN49811267Abvxj2C7b4xxZc8eoVTF84vdDxVIQshlpt+sLp/O31v9Hz+p4MvG0g7019jz8M+ANf/uRL/rbv3+iyUZdCR5QkSapRerbtyePHPM7wE4fTdaOu/GLYL9j0yk054cETGPHVCC9zobypXegAKqwVK1fw7GfPcvM7N/PwRw+zdMVSdtpkJ2468CaO3fZY6tWuV+iIkiRJNV6/zv148aQX+eDrD7juzeu4bfRt3Db6NrZrtx1n9T6LY7Y9hsZ1Gxc6pmow9xgWqc9mfcZvn/8tnf/WmX3v2JfnP3+es3ufzegzR/PGaW9wyvanWAolSZLybJs223DN4GuY+NOJXDf4OlamlZzx2Bm0v7I9Zz9+Nm9MfMO9iKoU7jEsItMXTufeMffynw/+wytfvkJJlDBos0FcNegqDux+IHVr1S10REmSJAFN6jXhzN5ncsaOZzBiwgiuffNa/v3uv7lu1HVs1WorTux1Ivt225cWK1tQu8SP9NpwbkU13IS5E3js48d4+KOHGfbZMJavXE6P1j34/YDfc0KvE9i06aaFjihJkqS1iAh267Abu3XYjWsWX8M9Y+7hlvdu4YLnLuCC5y6gQe0G9Grdi76d+9KnfR/6tO9Dx2YdiYhCR1c1YzGsYVJKvDPlHR756BEe/fhR3p78NgBdN+rKz3b9GcdsewzbttnWNwtJkqRqpln9Zpy242mctuNpfD7rc1776jWGfzKcd75+h6vfuJqlK5YC0LZR29Ulcef2O9N7k95s1GCjAqdXVWcxrAEWL1/M858/zyMfPcJjHz/GxHkTCTL/u3TZXpdxQPcD2KrVVpZBSZKkGqLLRl3oslEX9mqzFwAtWrdg9NTRjJwwkjcmvcEbE9/g0Y8fXT3+Fi23WF0U+7TvQ6+2vTyfhL7FYlhNTZ0/lcc/eZxHPnqEZz97loXLFtK4bmMGbTaIA7Y4gP267UfrRq0LHVOSJEl5ULdWXXpv0pvem/Tm//H/AJi9eDajJo3ijYmZojjss2HcPvp2AOqU1GG7dtutLop92vehW8tulITnpixWFsNqYmVayeipo3nikyd49ONHGTlhJIlEh6YdOHm7kzlgiwPo17mf//MjSZIkAJrXb87eXfdm7657A5mvHE2YO2F1URw5cST/fvff/OPNf6wef6dNdvrWYahtG7ct5CIojyyGVdSKlSt4b+p7vPjFi7ww/gVeHv8ysxbPAmCnTXbi4v4Xc8AWB9CzbU8PEZUkSdI6RQQdmnWgQ7MODNl6CJD5zPnh9A8zRTF7GOplr1zGirQCgI7NOmaK4iZ92HnTndlh4x28nmINZTGsIpYsX8Lbk9/m1a9e5cXxL/Ly+JeZs2QOAJu32JxDtzqUfp37sVeXvdi4ycYFTitJkqSaoFZJLbZpsw3btNmGU7Y/BYCFyxby9uS3v7Vn8b6x9wFQEiX0aN3jW4eg9mjTw0tm1AD+BQto/tL5XDTiIt6a+hajp49efSapLVpuwRE9jqBf537s2WlP2jdtX+CkkiRJKhYN6zSkb8e+9O3Yd/Wwrxd8zZsT31xdFO//8H7+9c6/Vo+/48Y7ri6Kfdr3oVOzTh7VVs1YDAuoQe0G3P/J/XRp1oVz+5zLbh12Y9cOu9KucbtCR5MkSZJWa9OoDYO3GMzgLQYDme8rfjrr028dgvqPN/7BkhVLVo9f+hDUnTbZyUtmVHEWwwKqVVKLd497l1oltWjXzjIoSZKk6iEi2LzF5mzeYnOO2fYYAJauWMroqaO/dQjqYx8/tnqabi26feuSGdu1284TJ1YhFsMCq1VSq9ARJEmSpA1W+pIZZ+90NgBzFs9ZfcmMkRNH8tznz3HH+3cA31wyY9Xhp13qdmGz5psVchGKmsVQkiRJUqVoVr8Ze3Xdi7267gVkDkGdOG/itw5BHfreUK558xoAmtZtSp9NvzkEtU/7Pn7NKk8shpIkSZLyIiLYtOmmbNp0Uw7d6lDgm0tmPDv2Wd75+h0+mPUBl796+epLZnRo2uFbh6DuuMmOXjKjEuS9GEZEB+AqYCAQwDDgxymlL3OYtj5wCXAc0Bx4Fzg/pfRSZeWVJEmSVHlWXTKj1cpWHL3l0bRr146FyxbyzuR3Mt9XnJTZu3j/h/cDmUtmtGnUhpYNWtKiQQtaNmxJywbZW8PssOzPpYfVrVW3wEtateW1GEZEQ+B5YAlwIpCAS4HhEdEzpbRgHbO4CRgM/Bz4DPh/wNMRsWtK6d1KCy5JkiQpbxrWacjuHXdn9467rx42bcE03pz0Jm9OfJMJcycwc/FMZiycwbiZ4xi5cCQzFs1Yffm3NWlct/H/Fsb6Lb71e9my2ax+M0qiJB+LXHD53mN4GtAV6J5SGgcQEaOBT4AzgCvXNmFE9AKOAU5JKf07O+xFYAxwMXBg5UaXJEmSVCitG7Vmv277sV+3/db4eEqJBcsWMHNRpjDOWDRj9f23hmWHfzH7C2YsmsGsRbNIpDXOsyRKaNGgxf/uhSy7t7LM3snqKN/F8EDg9VWlECCl9HlEvAocRDnFMDvtMuDuUtMuj4i7gAsiol5KaUkl5a4UN98Md9wBM2fC/PnQrh00bgzt28Nxx0G/ft+MO3UqjB6dGbdFC+jZE9q2ze15Sk87Zw58+SXMmwebbAKDB8M221TM88yYAe+/v+Zpy863Th0YORImTfrfHOvKUFHrYtU1V1PasPls6LQbbwyTJ294pqpoQ9ZTZfngA3j88TVvexVpfZe9orbNDclV3rZZkTkqa7vY0PevTz6BJUvWPW1lvkdVlsravsp7PVXkelif7XZ9Xmvr+jexojKuy/q8L1XUev2u623SJGjWDHbdteK367LrYeedYdmywr+W8vWa/q7vz/XqQbdumc+RxaC89RQRNK7bmMZ1G9OxWcec57li5QpmL579vwWydKnM/j5h7gTem/IeMxbNYOGyhWudZ92oz7Etfs/R3X9aJf4dyEWktOZ2XClPFjEFeDildEaZ4dcCh6eUWpcz7V3A9iml7mWGH0GmLG6TUhpT3vP37t07jRo16jvnr0g33wy//S106jSFFSvg/ffbsXJl5k1w441h+nT4zW8y5XDqVBg2LPNG3KgRLFiQ+cds773XvZGVnnbqVHjggUwp2247WLky8zxnn535B2hDnmfMmCmMHAlt2rT7n2nh2/MdMwYeeijzQt50U5g9+5scrVuXn6Gi1sWSJTBiROaD0W67Zd5UV80npSkAa722ZEVlaNQIJkzI5Nh1V2jYcO2ZvuubyZQp5S9LZdqQ9bQmFbEsH3wA114LrVpB8+bf3vYqshyWt+xr2r5y3TY39B+V8nLB2rfNTTdd89/vu/xNKnq7qIj5Tp0Kjz8+hcaNoXXrduVOu67nqazlWx9l/y6VtX2V93pa13v5+ixHRLuct9t1Pc/6/Ju4Ptb1dy/vtbI+70sVtX2t73xKj5/SFBYtgmXL2lXodl12PUyYkPnwf8ghsPXWFf9ayvX9K1+v6Q15f542bQrz58Pgwe2qRfkoz7r+LlXhPba0xcsX/0+R/GLqTN4YM4NF6Uv2aD2YrRoPLmjGNYmIt1JKvcsOz/cBsy2AWWsYPhPYaAOmXfV4tXHddZmNumnTzP9UNmiQuY0dm3lTbNUKbr89M+7o0ZlxGzfO/E9v48aZ30ePXvfzlJ525Eho2TLzD/bUqZn/ZWnVKvO/cxv6PJ98khl/TdOWne/o0ZnnXbQISkq+nWNdGSpqXXz+eWZdtGqV+fm7zmdDMkRk/g6tWmXuNyRTVbQh66myPP54Zv22aPG/215FWt9lr6htc0NylbdtVuTfr7K2iw19XTZunPnPmXVNW5nvUZWlsrav8l5PFbke1me7XZ/X2rr+TayojOuyPu9LFbVeN2S9RWReKxW9XZddD4sWZX5/773Cvpby9ZrekPfnhg0zj1XXzwvro6q9x9avXZ9NmmzCtm23pV/nfgzZegg9l5/G8Z0v4PQuF7JV4x0LnnF95LsY5l1EnB4RoyJi1LRp0wodZ7Wvv86UQsgcJlG7NtStm3kjBNhoI5g4MfPzzJmZ/xUprVGjzPB1KT3tjBmZF1Ddupn/YYHM/8pNmrThzzNnTqbYrmnasvOdMSOzfKsylM6xrgwVtS7mzoX69TO3uXO/+3w2JMOqHM2bZ+43JFNVtCHrqbJMmpRZ36WVfg1UlPVd9oraNjckV3nbZi7LUBEZCjXfmTPX/v61vs9TFbf7ytq+yns9VeR6WJ/tdl3Psz7/JlZUxnVZn/elilqvG7LeNuR5y1N2PSxcmPmsMGNG5T1nLvL1mt7Q9+cGDarv54X1URXfY8uqDhnXJt/FcBZr3jO4tr2BuU4L3+w5/JaU0g0ppd4ppd6tW6/1SNW8a9Pmmxd0nTqwfDksXfrNh5NZszLfNYTM/56VLlGQ+b1FDvtIS0/bsmXmu4xLl36zwc6enTmOf0Ofp1mzb0pt2WnLzrdly8zylX7RrMqxrgwVtS6aNoXFizO3VQX9u8xnQzKsyjF7duZ+QzJVRRuynirLJptk1ndppV8DFWV9l72its0NyVXetpnLMlREhkLNt0WLtb9/re/zVMXtvrK2r/JeTxW5HtZnu13X86zPv4kVlXFd1ud9qaLW64astw153vKUXQ8NG2Y+K7RsWXnPmYt8vaY39P150aLq+3lhfVTF99iyqkPGtcl3MRwD9FjD8K2BsTlM2yV7yYuy0y4Fxv3vJFXXWWdl9rLNnQtNmmRe0IsWZY6jnz49czvuuMy4PXtmxp0/P/O9kPnzM7/37Lnu5yk97c47Z/7nbdq0zDHOM2dmnmfw4A1/nm7dMuOvadqy8+3ZM/O8DRpkvtNROse6MlTUuujSJbMupk/P/Pxd57MhGVLK/B2mT8/cb0imqmhD1lNlGTw4s35nzvzfba8ire+yV9S2uSG5yts2K/LvV1nbxYa+LufPz+yhWNe0lfkeVVkqa/sq7/VUkethfbbb9XmtrevfxIrKuC7r875UUet1Q9ZbSpnXSkVv12XXQ4MGmd979Srsaylfr+kNeX9euDDzWHX9vLA+quJ7bFmrMubyb0pVk++Tz/wY+DOwRUrps+ywzmQuV3FBSukv5Uy7PfA2cFJKaWh2WG3gfWBcSumAdT1/VTr5DKw6K+mU7FlJ21Xrs5JOmTKFGTNg0qR21f6spLl8Ib26nJW0kCefgYo9k1tFLUuhz0q6tuWojmcl/a5/k6p4VtIxY6Zkz0rartqflXRNf5fqeFbS0stR3c9Kuq7XSvU6K+mU7FlJK/5EJ/k8K+n6vH9V/bOSTqFbN+jRo/qflrSyP4Ply9SpMGLEFObMgU02aVflMq7t5DP5LoaNgPeARcBvyFzg/hKgCdAzpTQ/O14n4FPg4pTSxaWmvwsYROYC958DZwH7A7ullN5e1/NXtWIIhf/gXlFqynKAy1JV1ZRlqSnLAS5LVVVTlqWmLAe4LFVRTVkOcFmqqqq8LFXirKQppQXAAOBj4DbgDjIFb8CqUpgVQK015DsZ+DdwKfA40AHYJ5dSKEmSJElas9r5fsKU0pfAkHWM8wWZclh2+CLgp9mbJEmSJKkC5PvkM5IkSZKkKsZiKEmSJElFzmIoSZIkSUXOYihJkiRJRc5iKEmSJElFzmIoSZIkSUXOYihJkiRJRc5iKEmSJElFzmIoSZIkSUXOYihJkiRJRc5iKEmSJElFzmIoSZIkSUXOYihJkiRJRc5iKEmSJElFzmIoSZIkSUXOYihJkiRJRc5iKEmSJElFzmIoSZIkSUUuUkqFzpA3ETENGF/oHEArYHqhQ0iVzO1cNZ3buIqB27lqumLcxjullFqXHVhUxbCqiIhRKaXehc4hVSa3c9V0buMqBm7nquncxr/hoaSSJEmSVOQshpIkSZJU5CyGhXFDoQNIeeB2rprObVzFwO1cNZ3beJbfMZQkSZKkIuceQ0mSJEkqchbDPImIDhFxX0TMiYi5EfFARHQsdC7pu4iIfhGR1nCbXWa8jSLiXxExPSIWRMSwiNi2QLGltYqITSPi7xExIiIWZrfnzmsYr35EXBERkyNiUXb8761hvJKI+GVEfBERiyPivYgYkpeFkdZgPbbxNb23p4jYrsx4buOqUiLisIi4PyLGZ9+fP4qIP0ZEkzLj5fTZJNf3+5rEYpgHEdEQeB7YEjgROB7oBgyPiEaFzCZtoHOBXUvd9l71QEQE8CiwD/BDYAhQh8x2v2n+o0rl2hw4ApgFvFzOeDcBpwEXAvsDk4Gny35oBi4BLgL+AewLvA7cGxH7VWhqKXe5buMAt/Dt9/ZdgY/LjOM2rqrmPGAF8Csynz2uA84Cno2IEljvzya5vt/XGH7HMA8i4kfAlUD3lNK47LAuwCfAL1JKVxYyn7S+IqIfMBwYmFIatpZxDgIeAgaklIZnhzUDPgduTymdm5ewUg4ioiSltDL786nAjUCXlNIXpcbpBbwLnJJS+nd2WG1gDPBRSunA7LA2wFfAZSml/ys1/XNA65RSz7wslFRKLtt49rEE/D6l9Jty5uU2rionIlqnlKaVGXYCMBTYK6X0fK6fTXJ9v69p3GOYHwcCr68qhQAppc+BV4GDCpZKqlwHApNWvfECpJTmkPmfOrd7VSmrPjCvw4HAMuDuUtMtB+4CBkVEvezgQUBd4PYy098ObJv9j0Epr3LcxnPlNq4qp2wpzHoze98+e5/rZ5Nc3+9rFIthfvQAPljD8DHA1nnOIlWkOyJiRUTMiIj/lPnebHnbfceIaJyfiFKF6QF8nlJaWGb4GDIfkjcvNd4SYNwaxgPf91X1nRURS7LfRXw+IvYo87jbuKqLPbP3H2bvc/1skuv7fY1iMcyPFmSO6S9rJrBRnrNIFWEO8BfgVGAAme+a7A2MyB5iBOVv9+C2r+pnXdt0i1L3s9P/flej7HhSVXQ7cDaZ9/TTgZbA89mvEKziNq4qLyLaAxcDw1JKo7KDc/1skuv7fY1Su9ABJFU/KaV3gHdKDXoxIl4C3iBzQpq1fjdFklR1pZSOL/XryxHxMJk9LJcCfQuTSlo/2T1/DwPLgZMLHKfacI9hfsxizXtH1va/EVK1k1J6m8xZ63bKDipvu1/1uFSdrGubnllqvObZs9+VN55U5aWU5gGP8817O7iNqwqLiAZkvjPYFRiUUppQ6uFcP5vk+n5fo1gM82MMmWOVy9oaGJvnLFJlW3VoUXnb/Zcppfn5iyRViDFAl+wliErbGljKN9+3GgPUAzZbw3jg+76qp9KHjbqNq0qKiDrAfUBvYL+U0vtlRsn1s0mu7/c1isUwPx4BdomIrqsGZC8qu3v2Manai4jeQHcyh5NCZttuHxF7lhqnKXAAbveqnh4lc72rw1cNyJ6+/EjgmZTSkuzgp8icze7YMtMfB3yQPSu1VC1k37f355v3dnAbVxWUvVbhHWTOfXBwSun1NYyW62eTXN/vaxS/Y5gfNwLnAA9HxG/I/K/bJWSuAfTPQgaTvouIuIPMNX/eBmYD2wO/BCYCV2dHewQYAdweET8nc1jGL4EA/pTnyNI6RcRh2R93zN7vGxHTgGkppRdTSu9ExN3AX7P/K/05mYsnd6HUB+SU0tcRcSXwy4iYR+Z1ciSZDys18tpXqh7WtY1HxHlk/oNvODAJ6ETmouHtcBtX1XcNmSL3e2BBROxS6rEJ2UNKc/pskuv7fU3jBe7zJHsa/6uAgWQ2vueAH5e9sKxUHUTEL4GjyXxoaAhMAZ4E/i+lNLnUeC2APwMHA/XJvBn/NKX0Xr4zS+uSvbD3mryYUuqXHacBmQ8dxwDNgfeA81NKL5SZVy0yHzZOI/Oh+iPg4pTSfZWRXcrFurbxiDgAuIBMOWwGzCVzzeVLU0ql9xi6javKiYgvyHwuWZPfpZQuyo6X02eTXN/vaxKLoSRJkiQVOb9jKEmSJElFzmIoSZIkSUXOYihJkiRJRc5iKEmSJElFzmIoSZIkSUXOYihJkiRJRc5iKEmq8iLixohIEXFVFchyaERMjYiGhc6yJhHRLyIuioiSMsM7Z9fhqRXwHH+NiCc2dD6SpKrDYihJqtKyFxk+IvvrMRFRu4BZagN/BK5IKS0sVI516Af8H5X7b/zlQP+I6F+JzyFJyiOLoSSpqjsYaAo8AbQB9ilgloOAzsDNBcxQcCmlycCjwM8LnUWSVDEshpKkqu5EYBZwErAo+/v/iIijI+K/EbE4It6PiAMj4oWIeKHMeK0j4vqImBgRS7LTnJ5jllOBp1JKM8vMM0XEpRHxs4gYHxELI+LxiGiTvd0TEXMi4quIOH8N2ftExLCImB8RCyLiuYjoU2acWyJiQkRsHxEvZ5/jk4g4s9Q4F5HZWwiwLJsrlXm6WhFxcURMjojZEfFoRGxa5rmOiYh3snnmZtfnGWXmcxcwKCI65LjuJElVmMVQklRlRcQmwN7A3SmlacBDwAERsVGZ8QYCdwD/BQ4F/gz8FdiizHhNgVeA/YCLgMFk9nxdFxE/XEeWemQO03x5LaMcDwwAzgbOAfYAbgUeBEYDQ8js9bwsIvYrNd+ewIvARmTK7wlk9pC+GBG9yjxHU+A/wO1k9l6+mc2+6pDOfwE3ZX/uC+yavZX2S2Bz4BTgR9nHby+Vp2/29xfJ7K09DLgRaF5mPi+T+RwxcC3rQ5JUjRTsexqSJOXgOKAWmYIFMBQ4GjgSuL7UeL8DxgKHpJQSQER8AIwCPi413o+ATsC2KaVPssOGRURz4P8i4rqU0vK1ZNkOqA+8t5bHlwAHrZo+IrYBfgL8NqV0aXbYC8AhwOFkSiLAhdlp90opzc6O9yzwBZm9f4eWeo4mwNkppeHZ8V4CBmXXyfCU0oSImJAdd+RaluWLlNIxq36JiNbAFRGxSUppErALMDul9ONS0zxTdiYppWnZ59qFIj+0VpJqAvcYSpKqshOBT1JKI7K/DwMmUepw0oioBfQG7l9VCgFSSm8Bn5eZ3z7ASODziKi96gY8DbQEti4nyybZ+2lrefzZMkXsv9n7p0tlWg6MA0offvk94LFVpTA73lzgEWDPMs+xcFUpzI63hEzx7VhO7rLKnk30/ez9qnm8CWwUEbdHxP7Z0rw20/hmvUiSqjGLoSSpSoqI3mSK2gMR0TxbUJoADwC7RMSqw0RbAXWAr9cwm6llfm9DpogtK3O7N/t4y3Ii1c/eL1nL47PK/L60nOH1S/3eApi8hvlNIXN4aXnPsSpP/TUMX5uZZX5ftTz1AVJKL5LZo9mBzGGw07Lff+y5hnktAhqsx3NLkqooi6EkqapatVfwfDKFaNXtnOzwE7L308mUuzZrmEfbMr/PAF4DdlrLbVQ5eWZk78uWtQ01E2i3huHtWHMRrHQppftSSnuSWdZDgI2Bp8peG5FMqZ2e73ySpIpnMZQkVTkRUZfM9+ZGAv3XcHsXOD4iIqW0gkyhGxIRUWoeOwJdysz6KWBL4MuU0qg13OaVE2vVoaFdN3wJv+VFYL+IaLJqQPbnA4AXvsP8Vu0B3OA9eSml+Smlx4B/kimHq/eoZg/h7Qh8tKHPI0kqPE8+I0mqigaTKSE/Sym9UPbBiPgncB2Zs4QOJ3OSlmeAByPiBjKHl15E5nDMlaUmvYrMiWtejoiryJSaRmTK4h4ppYPWFiil9GVEjAf6UOosnhXgEmB/4LmIuBxIZPaSNgQu/g7zG5u9/1lEPAmsSCmVtyf0WyLiYjJ7WoeT+T7npsC5wLvZM8Ousk0240vfIaMkqYpxj6EkqSo6EZjHN9/9K+tOSl3TMKX0LHAssBWZ78WdD/yMTDGcs2qilNIcYDcyJ2A5n8yJYW4mc+mH4azb3WRKXIVJKY0mU3Dnkjnr6m3AfGDPlNLazoBanseAa8lcNmMEmZPJrI+RQGcyJfpZ4HIyezUHlxlvfzLr94XvkFGSVMVEqRO4SZJUY2Qv2j4O+H1K6ZIKmudmZPYy9kspvVIR86yuImIsmTPB/rbQWSRJG85iKEmq9iKiAXAlmctZTCfzPcBfkDkkskdKaU1n/fyuz3UjsHFKqUL3HFYnEXEQmT2tm5W+zIYkqfryO4aSpJpgBZmzeP6DzHcTFwAvA4dXZCnM+i1wRkQ0TCktrOB5VxcNgOMshZJUc7jHUJIkSZKKnCefkSRJkqQiZzGUJEmSpCJnMZQkSZKkImcxlCRJkqQiZzGUJEmSpCJnMZQkSZKkIvf/ATRjp+X82ernAAAAAElFTkSuQmCC", + "image/png": "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", "text/plain": [ "<Figure size 1080x720 with 1 Axes>" ] @@ -357,11 +364,18 @@ "execution_count": 14, "metadata": {}, "outputs": [ + { + "name": "stderr", + "output_type": "stream", + "text": [ + "WARNING:root:Present is considered as the positive class.\n" + ] + }, { "name": "stdout", "output_type": "stream", "text": [ - "[Solve the problem with smoothing parameters (1, 1): ] Elapsed time (s): 0.087073\n" + "[Solve the problem with smoothing parameters (1, 1): ] Elapsed time (s): 0.072660\n" ] } ], @@ -484,7 +498,7 @@ "name": "stdout", "output_type": "stream", "text": [ - "[Solve the problem with smoothing parameters (1, 1): ] Elapsed time (s): 0.166115\n" + "[Solve the problem with smoothing parameters (1, 1): ] Elapsed time (s): 0.113034\n" ] } ], @@ -685,7 +699,7 @@ "name": "stdout", "output_type": "stream", "text": [ - "[Solve the problem with smoothing parameters (0.0001, 10): ] Elapsed time (s): 0.193003\n" + "[Solve the problem with smoothing parameters (0.0001, 10): ] Elapsed time (s): 0.221361\n" ] } ], diff --git a/setup.py b/setup.py index 64692f1..2ec15db 100644 --- a/setup.py +++ b/setup.py @@ -33,8 +33,8 @@ "numpy", "pandas", "scipy", - "tensorly", "statsmodels", + "tensorly", ], extras_require={"dev": ["black", "ipykernel", "mypy", "pip-tools", "pytest"]}, )

`LabelEncoder` for binomial regression When facing binomial regression, the output needs to be 0 or 1 for the fitting procedure to work. This should be done internally, so the user can pass an arbitrary binary category.

2022-12-10T18:59:22

0.0

[]

ManuelNavarroGarcia/cpsplines

ManuelNavarroGarcia__cpsplines-12

46470464da46cb9fd4a9e2cd35907cc94cdf530b

diff --git a/README.md b/README.md index a974ebf..d2a48a5 100644 --- a/README.md +++ b/README.md @@ -1,49 +1,93 @@ # cpsplines -`cpsplines` is a Python module to perform constrained regression under shape constraints on the component functions of the dependent variable. It is assumed that the smooth hypersurface to be estimated is defined through a reduced-rank basis (B−splines) and fitted via a penalized splines approach (P−splines). To embed requirements about the sign of any order partial derivative (including the function itself), the constraints are included in the fitting process as hard constraints, yielding a semidefinite optimization model. In particular, the problem of estimating the component function using this approach is stated as a convex semidefinite optimization problem with a quadratic objective function, which can be easily reformulated as a conic optimization problem. - -Sign related constraints are imposed using a well-known result carried out by Bertsimas and Popescu, 2002. This enables to enforce non-negativity of a univariate polynomial over a finite interval, which can be straightforwardly extended to the sign of any higher order derivative. When only one covariate is related to the response variable, these constraints are successfully fulfilled over the whole domain of the regressor sample. However, when facing multiple regression, this equivalence does not hold, so alternative approaches must be developed. The proposed framework in this repository uses the equivalence relation for univariate polynomials by imposing the constraints over a finite set of curves which belong to the hypersurface. - -At present, `cpsplines` can handle constrained regression problems for data lying on grids. In this setting, the smooth hypersurface is constructed from the tensor products of B-splines basis along each axis, which allows to develop efficient algorithms accelerating the computations (Currie, Durban and Eilers, 2006). On this repository, the fitting procedure is performed using the method `GridCPsplines`, whose main features are the following: +`cpsplines` is a Python module to perform constrained regression under shape +constraints on the component functions of the dependent variable. It is assumed +that the smooth hypersurface to be estimated is defined through a reduced-rank +basis (B−splines) and fitted via a penalized splines approach (P−splines). To +embed requirements about the sign of any order partial derivative (including the +function itself), the constraints are included in the fitting process as hard +constraints, yielding a semidefinite optimization model. In particular, the +problem of estimating the component function using this approach is stated as a +convex semidefinite optimization problem with a quadratic objective function, +which can be easily reformulated as a conic optimization problem. + +Sign related constraints are imposed using a well-known result carried out by +Bertsimas and Popescu, 2002. This enables to enforce non-negativity of a +univariate polynomial over a finite interval, which can be straightforwardly +extended to the sign of any higher order derivative. When only one covariate is +related to the response variable, these constraints are successfully fulfilled +over the whole domain of the regressor sample. However, when facing multiple +regression, this equivalence does not hold, so alternative approaches must be +developed. The proposed framework in this repository uses the equivalence +relation for univariate polynomials by imposing the constraints over a finite +set of curves which belong to the hypersurface. + +At present, `cpsplines` can handle constrained regression problems for data +lying on grids. In this setting, the smooth hypersurface is constructed from the +tensor products of B-splines basis along each axis, which allows to develop +efficient algorithms accelerating the computations (Currie, Durban and Eilers, +2006). On this repository, the fitting procedure is performed using the method +`GridCPsplines`, whose main features are the following: * Arbitrary number of variables. -* Arbitrary knot sequence length to construct the B-spline basis. -* Arbitrary B-spline basis degrees. +* Arbitrary knot sequence length to construct the B-spline basis. +* Arbitrary B-spline basis degrees. * Arbitrary difference orders on the penalty term. -* Out-of-range prediction (backwards and forward) along every dimension (Currie and Durban, 2004), and the constraints are enforced both on the fitting and the prediction region. -* The smoothing parameters are selected as the minimizer of the Generalized Cross Validation criteria, but this routine can be done either by choosing the best parameters out of a set of candidates or by finding them using numerical methods. -* Enforcing sign related constraints over the fitting and prediction range (if prediction is required). Arbitrary number of sign constraints can be imposed along each regressor. -* Enforcing the hypersurface (or any partial derivative) attains a certain value at a certain point. - -Solving the optimization problems is done using [MOSEK](https://www.mosek.com) optimization software. +* Out-of-range prediction (backwards and forward) along every dimension (Currie + and Durban, 2004), and the constraints are enforced both on the fitting and + the prediction region. +* The smoothing parameters are selected as the minimizer of the Generalized + Cross Validation criteria, but this routine can be done either by choosing the + best parameters out of a set of candidates or by finding them using numerical + methods. +* Enforcing sign related constraints over the fitting and prediction range (if + prediction is required). Arbitrary number of sign constraints can be imposed + along each regressor. +* Enforcing the hypersurface (or any partial derivative) attains a certain value + at a certain point. +* Enforcing the hypersurface is a probability density function, i.e., it is + non-negative and it integrates to one. + +Solving the optimization problems is done using [MOSEK](https://www.mosek.com) +optimization software. ## Project structure The current version of the project is structured as follows: * **cpsplines**: the main directory of the project, which consist of: - * **fittings**: contains the smoothing algorithms. + * **fittings**: contains the smoothing algorithms. * **graphics**: constituted by graphic methods to visualize the results. - * **mosek_functions**: contains the functions used to define the optimization problems. - * **psplines**: composed by several methods that build the main objects of P-splines. - * **utils**: integrated by a miscellanea of files used for a variety of purposes (numerical computations, data processing, ...) -* **data**: a folder containing CSV files used in the real data numerical experiments. -* **examples**: a directory containing multiple numerical experiments, using both synthetic and real data sets. -* **tests**: a folder including tests for the main methods of the project. - + * **mosek_functions**: contains the functions used to define the + optimization problems. + * **psplines**: composed by several methods that build the main objects of + P-splines. + * **utils**: integrated by a miscellanea of files used for a variety of + purposes (numerical computations, data processing, ...) +* **data**: a folder containing CSV files used in the real data numerical + experiments. +* **examples**: a directory containing multiple numerical experiments, using + both synthetic and real data sets. +* **tests**: a folder including tests for the main methods of the project. ## Package dependencies `cpsplines` mainly depends on the following packages: * [Joblib](https://joblib.readthedocs.io/). -* [Matplotlib](https://matplotlib.org/). +* [Matplotlib](https://matplotlib.org/). * [MOSEK](https://www.mosek.com). **License Required** * [Numpy](https://numpy.org/). * [Pandas](https://pandas.pydata.org/). * [Scipy](https://www.scipy.org/). * [Tensorly](http://tensorly.org/). -MOSEK requires a license to be used. For research or educational purposes, a free yearly and renewable [academic license](https://www.mosek.com/products/academic-licenses/) is offered by the company. For other cases, a 30-day [trial license](https://www.mosek.com/try/) is available. According to MOSEK indications, the license file (`mosek.lic`) must be located at +MOSEK requires a license to be used. For research or educational purposes, a +free yearly and renewable [academic +license](https://www.mosek.com/products/academic-licenses/) is offered by the +company. For other cases, a 30-day [trial license](https://www.mosek.com/try/) +is available. According to MOSEK indications, the license file (`mosek.lic`) +must be located at + ``` $HOME/mosek/mosek.lic (Linux/OSX) %USERPROFILE%\mosek\mosek.lic (Windows) @@ -51,14 +95,15 @@ $HOME/mosek/mosek.lic (Linux/OSX) ## Installation -1. To clone the repository on your own device, use +1. To clone the repository on your own device, use ``` git clone https://github.com/ManuelNavarroGarcia/cpsplines.git cd cpsplines ``` -2. To install the dependencies, there are two options according to your installation preferences: +2. To install the dependencies, there are two options according to your + installation preferences: * Create and activate a virtual environment with `conda` (recommended) @@ -74,7 +119,8 @@ pip install -r requirements.txt pip install -e .[dev] ``` -3. If neccessary, add version requirements to existing dependencies or add new ones on `setup.py`. Then, update `requirements.txt` file using +3. If neccessary, add version requirements to existing dependencies or add new + ones on `setup.py`. Then, update `requirements.txt` file using ``` pip-compile --extra dev > requirements.txt @@ -82,11 +128,13 @@ pip-compile --extra dev > requirements.txt and update the environment with `pip-sync`. -## Usage +## Usage -To illustrate the usage of the repository, let's see how `GridCPsplines` works with two examples, the first with only one regressor and the second two covariates. +To illustrate the usage of the repository, let's see how `GridCPsplines` works +with two examples, the first with only one regressor and the second two +covariates. -For the univariate case, consider the function +For the univariate case, consider the function <img src="https://latex.codecogs.com/svg.image?f(x)&space;=&space;(2x-1)&space;^3," title="f(x) = (2x-1) ^3," /> @@ -100,7 +148,11 @@ which is a non-decreasing function. We simulated noisy data following the scheme <img src="https://render.githubusercontent.com/render/math?math=$\text{where%20}\:x_l=0,%200.02,%200.04,\ldots,%201.$"> -We fit an unconstrained and a constrained model imposing non-decreasing constraints over the interval [-0.15, 1.12] (forward and backwards prediction). For the basis, cubic B-splines with 10 interior knots are taken with a second-order difference penalty. The smoothing parameter is selected using `scipy.optimize.minimize` with the `"SLSQP"` method. +We fit an unconstrained and a constrained model imposing non-decreasing +constraints over the interval [-0.15, 1.12] (forward and backwards prediction). +For the basis, cubic B-splines with 10 interior knots are taken with a +second-order difference penalty. The smoothing parameter is selected using +`scipy.optimize.minimize` with the `"SLSQP"` method. ```python # Generate the data @@ -146,7 +198,7 @@ curves = plot_curves( ![alt text](./img/example1D.jpg) -For the bivariate case, consider the function +For the bivariate case, consider the function <img src="https://latex.codecogs.com/svg.image?f(x,&space;y)&space;=&space;\sin(x)\sin(y)." title="f(x, y) = \sin(x)\sin(y)." /> @@ -157,13 +209,18 @@ We simulated noisy data following the scheme <img src="https://latex.codecogs.com/svg.image?z_{lm}=&space;f(x_l,&space;y_m)+\varepsilon_{lm},&space;\quad&space;\varepsilon_{lm}\sim&space;\text{N}(0,1)" title="z_{lm}= f(x_l, y_m)+\varepsilon_{lm}, \quad \varepsilon_{lm}\sim \text{N}(0,1)." /> -We fit an unconstrained and a constrained model imposing non-negativity constraints over the interval +We fit an unconstrained and a constrained model imposing non-negativity +constraints over the interval <img src="https://latex.codecogs.com/svg.image?[0,3\pi]\times&space;[0,2\pi]" title="[0,3\pi]\times [0,2\pi]" /> -(no prediction). For the bases, cubic B-splines with 30 and 20 interior knots are taken, respectively, with a second-order difference penalty. The smoothing parameter is selected as the best candidates out of the sets {10, 100} (for the first smoothing parameter) and {10, 50, 100} (for the second smoothing parameter). +(no prediction). For the bases, cubic B-splines with 30 and 20 interior knots +are taken, respectively, with a second-order difference penalty. The smoothing +parameter is selected as the best candidates out of the sets {10, 100} (for the +first smoothing parameter) and {10, 50, 100} (for the second smoothing +parameter). ```python # Generate the data @@ -194,35 +251,58 @@ surface = plot_surfaces( ## Testing -The repository contains a folder with unit tests to guarantee the main methods meets their design and behave as intended. To launch the test suite, it is enough to enter `pytest`. If only one test file wants to be run, the syntax is given by +The repository contains a folder with unit tests to guarantee the main methods +meets their design and behave as intended. To launch the test suite, it is +enough to enter `pytest`. If only one test file wants to be run, the syntax is +given by ``` pytest tests/test_<file_name>.py ``` -Moreover, a GitHub Action runs automatically all the tests but `tests/test_solution.py` (which requires MOSEK license) when any commit is pushed on any Pull Request. - +Moreover, a GitHub Action runs automatically all the tests but +`tests/test_solution.py` (which requires MOSEK license) when any commit is +pushed on any Pull Request. ## Contributing -Contributions to the repository are welcomed! Regardless of whether it is a small fix on the documentation or a notable feature to be included, I encourage you to develop your ideas and make this project greater. Even suggestions about the code structure are highly appreciated. Furthermore, users participating on these submissions will figure as contributors on this main page of the repository. +Contributions to the repository are welcomed! Regardless of whether it is a +small fix on the documentation or a notable feature to be included, I encourage +you to develop your ideas and make this project greater. Even suggestions about +the code structure are highly appreciated. Furthermore, users participating on +these submissions will figure as contributors on this main page of the +repository. There are many ways you can contribute on this repository: -* [Discussions](https://github.com/ManuelNavarroGarcia/cpsplines/discussions). To ask questions you are wondering about or share ideas, you can enter an existing discussion or open a new one. +* [Discussions](https://github.com/ManuelNavarroGarcia/cpsplines/discussions). + To ask questions you are wondering about or share ideas, you can enter an + existing discussion or open a new one. -* [Issues](https://github.com/ManuelNavarroGarcia/cpsplines/issues). If you detect a bug or you want to propose an enhancement of the current version of the code, a issue with reproducible code and/or a detailed description is highly appreciated. +* [Issues](https://github.com/ManuelNavarroGarcia/cpsplines/issues). If you + detect a bug or you want to propose an enhancement of the current version of + the code, a issue with reproducible code and/or a detailed description is + highly appreciated. -* [Pull Requests](https://github.com/ManuelNavarroGarcia/cpsplines/pulls). If you feel I am missing an important feature, either in the code or in the documentation, I encourage you to start a pull request developing this idea. Nevertheless, before starting any major new feature work, I suggest you to open an issue or start a discussion describing what you are planning to do. I note that, before starting a pull request, all unit test must pass on your local repository. +* [Pull Requests](https://github.com/ManuelNavarroGarcia/cpsplines/pulls). If + you feel I am missing an important feature, either in the code or in the + documentation, I encourage you to start a pull request developing this idea. + Nevertheless, before starting any major new feature work, I suggest you to + open an issue or start a discussion describing what you are planning to do. I + note that, before starting a pull request, all unit test must pass on your + local repository. ## Contact Information and Citation -If you have encountered any problem or doubt while using `cpsplines`, please feel free to let me know by sending me an email: +If you have encountered any problem or doubt while using `cpsplines`, please +feel free to let me know by sending me an email: * Name: Manuel Navarro García (he/his) * Email: [email protected] -The formal background of the models used in this project are either published in research paper or under current research. If these techniques are helpful to your own research, consider citing the related papers of the project: +The formal background of the models used in this project are either published in +research paper or under current research. If these techniques are helpful to +your own research, consider citing the related papers of the project: ``` @TECHREPORT{navarro2020, @@ -236,9 +316,19 @@ The formal background of the models used in this project are either published in ## Acknowledgements -Throughout the developing of this project I have received strong support from various individuals. - -I would first like to thank my supervisors, Professor [Vanesa Guerrero](https://github.com/vanesaguerrero) and Professor [María Durbán](https://github.com/MariaDurban), whose insightful comments and invaluable expertise has given way to many of the current functionalities of the repository. - -I would also like to acknowledge the [Komorebi AI](https://komorebi.ai/) team for their assistance and guidance on the technical part of the project. Specially, I would like to thank [Alberto Torres](https://github.com/albertotb), [David Gordo](https://github.com/davidggphy) and [Victor Gallego](https://komorebi.ai/) for their constructive code structure suggestions that have helped notably to improve the computational efficiency and the usage of the algorithms. - +Throughout the developing of this project I have received strong support from +various individuals. + +I would first like to thank my supervisors, Professor [Vanesa +Guerrero](https://github.com/vanesaguerrero) and Professor [María +Durbán](https://github.com/MariaDurban), whose insightful comments and +invaluable expertise has given way to many of the current functionalities of the +repository. + +I would also like to acknowledge the [Komorebi AI](https://komorebi.ai/) team +for their assistance and guidance on the technical part of the project. +Specially, I would like to thank [Alberto Torres](https://github.com/albertotb), +[David Gordo](https://github.com/davidggphy) and [Victor +Gallego](https://komorebi.ai/) for their constructive code structure suggestions +that have helped notably to improve the computational efficiency and the usage +of the algorithms. diff --git a/cpsplines/fittings/grid_cpsplines.py b/cpsplines/fittings/grid_cpsplines.py index b5680e3..aa8d5a6 100644 --- a/cpsplines/fittings/grid_cpsplines.py +++ b/cpsplines/fittings/grid_cpsplines.py @@ -7,6 +7,7 @@ import scipy from cpsplines.mosek_functions.interval_constraints import IntConstraints from cpsplines.mosek_functions.obj_function import ObjectiveFunction +from cpsplines.mosek_functions.pdf_constraints import PDFConstraint from cpsplines.mosek_functions.point_constraints import PointConstraints from cpsplines.psplines.bspline_basis import BsplineBasis from cpsplines.psplines.penalty_matrix import PenaltyMatrix @@ -100,6 +101,10 @@ class GridCPsplines: where the value needs to be fixed. - An array with the values of the derivative to be enforced. - A number corresponding to the tolerancea allowed in the constraint. + pdf_constraint : bool, optional + A boolean indicating whether the fitted hypersurface must satisfy + Probability Density Function (PDF) conditions, i.e., it is non-negative + and it integrates to one. Attributes ---------- @@ -131,6 +136,7 @@ def __init__( Dict[int, Dict[int, Dict[str, Union[int, float]]]] ] = None, pt_constraints: Optional[Dict[Tuple[int], Any]] = None, + pdf_constraint: bool = False, ): self.deg = deg self.ord_d = ord_d @@ -140,6 +146,7 @@ def __init__( self.sp_args = sp_args self.int_constraints = int_constraints self.pt_constraints = pt_constraints + self.pdf_constraint = pdf_constraint def _get_bspline_bases(self, x: Iterable[np.ndarray]) -> List[BsplineBasis]: @@ -180,6 +187,8 @@ def _get_bspline_bases(self, x: Iterable[np.ndarray]) -> List[BsplineBasis]: ) # Generate the design matrix of the B-spline basis bsp.get_matrix_B() + if self.int_constraints is not None or self.pdf_constraint: + bsp.get_matrices_S() bspline_bases.append(bsp) return bspline_bases @@ -296,11 +305,17 @@ def _initialize_model( lin_term=lin_term, ) + if self.pdf_constraint: + pdf_cons = PDFConstraint(bspline=self.bspline_bases) + # Incorporate the condition that the integral over all the space + # must equal to 1 + pdf_cons.integrate_to_one(var_dict=mos_obj_f.var_dict, model=M) + # Enforce the non-negativity constraint if it is not imposed + # explicitly + self.int_constraints = pdf_cons.nonneg_cons(self.int_constraints) + if self.int_constraints is not None: - matrices_S = {} - # If the are interval constraints, construct the matrices S - for i, bsp in enumerate(self.bspline_bases): - matrices_S[i] = bsp.get_matrices_S() + matrices_S = {i: bsp.matrices_S for i, bsp in enumerate(self.bspline_bases)} # Iterate for every variable with constraints and for every # derivative order for var_name in self.int_constraints.keys(): @@ -333,7 +348,6 @@ def _initialize_model( cons2.point_cons(var_dict=mos_obj_f.var_dict, model=M) else: self.pt_constraints = {} - return M def _get_sp_grid_search( diff --git a/cpsplines/mosek_functions/interval_constraints.py b/cpsplines/mosek_functions/interval_constraints.py index f59bd71..a33c792 100644 --- a/cpsplines/mosek_functions/interval_constraints.py +++ b/cpsplines/mosek_functions/interval_constraints.py @@ -12,7 +12,7 @@ class IntConstraints: """ Define the set of constraints that ensures the non-negativity (or - non-positivity) for the derivative order `derivative`along the respective + non-positivity) for the derivative order `derivative` along the respective axis of the variable `var_name`. For the unidimensional case, the requirements are completely fulfilled for every point in the range where the B-spline basis is defined. For the multivariate case, the constraints are diff --git a/cpsplines/mosek_functions/pdf_constraints.py b/cpsplines/mosek_functions/pdf_constraints.py new file mode 100644 index 0000000..7b2843a --- /dev/null +++ b/cpsplines/mosek_functions/pdf_constraints.py @@ -0,0 +1,128 @@ +from typing import Dict, Iterable, Optional, Union + +import mosek.fusion +import numpy as np +from cpsplines.mosek_functions.utils_mosek import matrix_by_tensor_product_mosek +from cpsplines.psplines.bspline_basis import BsplineBasis +from scipy.sparse import diags + + +class PDFConstraint: + + """ + Define the constraints that characterize a Probability Density Function, + i.e., it must be non-negative and must integrate to one. For the first, only + the dictionary of interval constraints is updated with the non-negative + constraints (when they were not already included). + + Parameters + ---------- + bspline : Iterable[BsplineBasis] + An iterable containing the B-spline bases objects used to approximate + the function to estimate. + """ + + def __init__( + self, + bspline: Iterable[BsplineBasis], + ): + self.bspline = bspline + + def integrate_to_one( + self, + var_dict: Dict[str, mosek.fusion.LinearVariable], + model: mosek.fusion.Model, + ) -> mosek.fusion.LinearConstraint: + + """ + Defines the constraint that the integral of the fitted hypersurface + over the fitting and the prediction regions must be equal to one. + + Parameters + ---------- + var_dict : Dict[str, mosek.fusion.LinearVariable] + The dictionary that contains the decision variables used to define + the objective function of the problem. + model : mosek.fusion.Model + The MOSEK model of the problem. + + Returns + ------- + mosek.fusion.LinearConstraint + The constraint that enforces to fitted hypersurface integrates to + one. + """ + + banded_list = [] + for bsp in self.bspline: + # For each B-spline basis, construct a Vandermonde matrix with only + # the inner knots and with powers up to degree `deg` + 1. The first + # column, containing ones, is dropped + vander = np.vander(bsp.knots, N=bsp.deg + 2, increasing=True)[ + bsp.deg : -bsp.deg, 1: + ] + # The vector of constants obtained when integrating monomials + integrand_coef = 1 / np.linspace(1, bsp.deg + 1, bsp.deg + 1) + # Take the difference of a row minus the previous on the Vandermonde + # matrix and multiply each column by the respective integration factor + diff_mat = np.einsum("ij,j->ij", np.diff(vander, axis=0), integrand_coef) + # Multiply each row by the corresponding matrix S and transpose, so + # first term corresponds to first row and so on. Then, use this + # vector to create a banded matrix with inner knots rows nd inner + # rows plus degree columns + banded = diags( + np.array( + [np.dot(diff_mat[i], s) for i, s in enumerate(bsp.matrices_S)] + ).T, + range(bsp.deg + 1), + shape=(bsp.matrixB.shape[1] - bsp.deg, bsp.matrixB.shape[1]), + ).toarray() + banded_list.append(banded) + # Multiply by the coefficient array and sum all the entries + sum_coef = mosek.fusion.Expr.sum( + matrix_by_tensor_product_mosek( + matrices=banded_list, mosek_var=var_dict["theta"] + ) + ) + cons = model.constraint(sum_coef, mosek.fusion.Domain.equalsTo(1.0)) + return cons + + def nonneg_cons( + self, + int_constraints: Optional[Dict[int, Dict[int, Dict[str, Union[int, float]]]]], + ) -> Dict[int, Dict[int, Dict[str, Union[int, float]]]]: + + """ + Includes non-negativity constraints to the problem if they were + already not included. + + Parameters + ---------- + int_constraints : Optional[Dict[int, Dict[int, Dict[str, Union[int, float]]]]] + The nested dictionary containing the interval constraints to be + enforced. + + Returns + ------- + Dict[int, Dict[int, Dict[str, Union[int, float]]]] + A interval constraints dictionary updated with non-negativity + constraints along all axis, if they were not included explicitly. + """ + + # Create a dictionary if no interval constraints already exist + if int_constraints is None: + int_constraints = {} + for i in range(len(self.bspline)): + # Check if any interval constraints exist along each axis + if int_constraints.get(i, None) is not None: + # Check if sign constraints exist for this axis + if int_constraints[i].get(0, None) is None: + # Update the constraint dictionary if non-negativity is + # missing + int_constraints[i].update({0: {"+": 0}}) + else: + int_constraints[i][0].update({"+": 0}) + # If no constraints are imposed on any axis, include non-negativity + else: + int_constraints.update({i: {0: {"+": 0}}}) + return int_constraints diff --git a/cpsplines/psplines/bspline_basis.py b/cpsplines/psplines/bspline_basis.py index 0b15730..6346871 100644 --- a/cpsplines/psplines/bspline_basis.py +++ b/cpsplines/psplines/bspline_basis.py @@ -182,7 +182,7 @@ def get_matrix_B(self) -> None: self.matrixB = self.bspline_basis(x=x_eval) return None - def get_matrices_S(self) -> List[np.ndarray]: + def get_matrices_S(self): """ Generate a list of matrices (one by interval) containing the polynomial @@ -200,11 +200,6 @@ def get_matrices_S(self) -> List[np.ndarray]: and stock prices: a convex optimization approach. Operations Research, 50(2), 358-374. - Returns - ------- - List[np.ndarray] - List of matrices of shape (`deg` + 1, `deg` + 1) with the polynomial - coefficients. """ # Since knots are evenly spaced, the value of the B-spline basis @@ -235,4 +230,5 @@ def get_matrices_S(self) -> List[np.ndarray]: ) S_k = np.linalg.solve(T_k, C) S.append(S_k) - return S + self.matrices_S = S + return None diff --git a/examples/simulated_data.ipynb b/examples/simulated_data.ipynb index c391daf..98dc24b 100644 --- a/examples/simulated_data.ipynb +++ b/examples/simulated_data.ipynb @@ -19,6 +19,7 @@ "import itertools\n", "import matplotlib.pyplot as plt \n", "import numpy as np\n", + "from scipy.stats import norm, multivariate_normal\n", "\n", "from cpsplines.fittings.grid_cpsplines import GridCPsplines\n", "from cpsplines.graphics.plot_one_smoothing import plot_curves, plot_surfaces" @@ -1011,6 +1012,145 @@ " figsize=(10, 6),\n", ")" ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "*Example 14*. Given the gaussian function \n", + "$$f(x) = \\frac{1}{\\sqrt{2\\pi\\sigma^2}}\\exp\\left\\{-\\frac{(x-\\mu)^2}{2\\sigma^2}\\right\\}$$\n", + "with $\\mu=0$ and $\\sigma=2$, we simulated noisy data following the scheme\n", + "$$y_{l}= f(x_l)+\\varepsilon_{l}, \\quad \\varepsilon_{l}\\sim \\text{N}(0,0.025),$$\n", + "where $x_l=\\{\\frac{l -100}{10}\\}_{l=0}^{200}$.\n", + "\n", + "We fit an unconstrained and a constrained model over the interval $[-10,10]$\n", + "imposing that the curve is a probability density function, i.e., it is\n", + "non-negative and it integrates to one over the fitting region. " + ] + }, + { + "cell_type": "code", + "execution_count": 66, + "metadata": {}, + "outputs": [ + { + "data": { + "image/png": "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", + "text/plain": [ + "<Figure size 1080x720 with 1 Axes>" + ] + }, + "metadata": { + "needs_background": "light" + }, + "output_type": "display_data" + } + ], + "source": [ + "np.random.seed(1)\n", + "x14 = np.linspace(-10, 10, 201)\n", + "y14 = norm.pdf(x14,0,2) + np.random.normal(0, 0.025, 201)\n", + "\n", + "example14_1 = GridCPsplines(\n", + " deg=(3,),\n", + " ord_d=(2,),\n", + " n_int=(20,),\n", + " sp_args={\"options\": {\"ftol\": 1e-12}},\n", + " pdf_constraint=False\n", + ")\n", + "example14_1.fit(x=(x14,), y=y14)\n", + "\n", + "example14_2 = GridCPsplines(\n", + " deg=(3,),\n", + " ord_d=(2,),\n", + " n_int=(20,),\n", + " sp_args={\"options\": {\"ftol\": 1e-12}},\n", + " pdf_constraint=True)\n", + "example14_2.fit(x=(x14,), y=y14)\n", + "\n", + "plot14 = plot_curves(\n", + " fittings=(example14_1, example14_2),\n", + " col_curve=(\"g\", \"k\"),\n", + " knot_positions=True,\n", + " constant_constraints=True,\n", + " x=(x14,), \n", + " y=(y14,),\n", + " col_pt=(\"b\",),\n", + " alpha=0.25)" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "*Example 15*. Given the bivariate gaussian function \n", + "$$f(\\mathbf{x}) = \\frac{1}{2\\pi\\sqrt{|\\boldsymbol{\\Sigma}|}}\\exp\\left\\{-\\frac{1}{2}(\\mathbf{x}-\\boldsymbol{\\mu})^\\top\n", + "\\boldsymbol{\\Sigma}^{-1} (\\mathbf{x}-\\boldsymbol{\\mu})\\right\\}$$\n", + "with $\\mathbf{x}=(x,y)$, $\\boldsymbol{\\mu}=(0,0)$ and $\\boldsymbol{\\Sigma}=\\begin{pmatrix} 2 & 1/2 \\\\ 1/2 & 1\n", + "\\end{pmatrix},$ we simulated noisy data following the scheme\n", + "$$z_{lm}= f(x_l, y_m)+\\varepsilon_{lm}, \\quad \\varepsilon_{lm}\\sim \\text{N}(0,0.025),$$\n", + "where $x_l=\\{\\frac{3l -75}{25}\\}_{l=0}^{50}$ and $y_m=\\{\\frac{2m -60}{15}\\}_{m=0}^{60}$.\n", + "\n", + "We fit an unconstrained and a constrained model over the interval $[-3,3]\\times\n", + "[-4,4]$ imposing that the curve is a probability density function, i.e., it is\n", + "non-negative and it integrates to one over the fitting region. " + ] + }, + { + "cell_type": "code", + "execution_count": 70, + "metadata": {}, + "outputs": [ + { + "data": { + "image/png": "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", + "text/plain": [ + "<Figure size 720x432 with 2 Axes>" + ] + }, + "metadata": { + "needs_background": "light" + }, + "output_type": "display_data" + }, + { + "data": { + "image/png": "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", + "text/plain": [ + "<Figure size 720x432 with 2 Axes>" + ] + }, + "metadata": { + "needs_background": "light" + }, + "output_type": "display_data" + } + ], + "source": [ + "x15_0 = np.linspace(-3,3,51)\n", + "x15_1 = np.linspace(-4,4,61)\n", + "y15 = (multivariate_normal.pdf(x = list(itertools.product(x15_0, x15_1)),\n", + " mean=[0,0],\n", + " cov=[[2,0.5],[0.5,1]]) + np.random.normal(0, 0.025, len(x15_0)*len(x15_1))).reshape((len(x15_0),len(x15_1)))\n", + "\n", + "example15 = GridCPsplines(\n", + " deg=(3, 3),\n", + " ord_d=(2, 2),\n", + " n_int=(10, 9),\n", + " sp_method=\"grid_search\",\n", + " int_constraints={0: {0: {\"+\": 0}}, 1: {0: {\"+\": 0}}},\n", + " pdf_constraint=True\n", + " \n", + ")\n", + "example15.fit(x=(x15_0, x15_1), y=y15)\n", + "\n", + "plot15 = plot_surfaces(\n", + " fittings=(example15,),\n", + " col_surface=(\"gist_earth\",),\n", + " orientation=(45, 45),\n", + " figsize=(10, 6),\n", + ")" + ] } ], "metadata": {

Incorporate density functions constraints Beside non-negativity, probability density functions must satisfy that integrate to 1. Incorporate a new type of constraint, `density_constraint`, to ensure that the fitted hypersurface satisfies this requirement.

2022-01-03T19:31:19

0.0

[]

ManuelNavarroGarcia/cpsplines

ManuelNavarroGarcia__cpsplines-10

a9dd210ad5d868629c5c5f679c69d58c3161fea2

diff --git a/cpsplines/utils/box_product.py b/cpsplines/utils/box_product.py new file mode 100644 index 0000000..739a1ef --- /dev/null +++ b/cpsplines/utils/box_product.py @@ -0,0 +1,33 @@ +import numpy as np + + +def box_product(A: np.ndarray, B: np.ndarray) -> np.ndarray: + + """ + Given a m x n matrix A and a m x p matrix B, computes the face-splitting + product or box product using that + A box B = (A kron 1_n^T) circ (1_n^T kron B), + where 1_n is a vector of ones with length n, kron denotes the Kronecker + product of matrices and circ denotes de Hadamard product of matrices. + + Parameters + ---------- + A : np.ndarray with shape (m, n) + The m x n matrix A. + B : np.ndarray with shape (m, p) + The m x p matrix B. + + Returns + ------- + np.ndarray + The resulting m x np array. + + Raises + ------ + ValueError + If any input array is not a matrix. + """ + + if A.ndim != 2 or B.ndim != 2: + raise ValueError("Only two-dimensional arrays are allowed.") + return np.multiply(np.repeat(A, B.shape[1], axis=1), np.tile(B, A.shape[1]))

Incorporate box product (or row-wise products) Row-wise products are needed to smooth scattered data. Implement a fast way to compute them (try not to compute any Kronecker product) and provide tests to evaluate the running time.

2021-11-25T18:57:02

0.0

[]

RWTH-EBC/ebcpy

RWTH-EBC__ebcpy-136

8fad8a2f4c2ec4411247de2a4fc842c6c6c8c219

diff --git a/CHANGELOG.md b/CHANGELOG.md index e84e0a4a..9fcac3f1 100644 --- a/CHANGELOG.md +++ b/CHANGELOG.md @@ -104,3 +104,5 @@ - fix reproduction with modifiers #82 - enable model_name=None upon startup of DymolaAPI #130 - Only create logger handler if not already done by root-logger +- v0.4.2 + - Do not use chars_to_strings from scipy #132 diff --git a/ebcpy/__init__.py b/ebcpy/__init__.py index 9f47111e..d1c2cbd8 100644 --- a/ebcpy/__init__.py +++ b/ebcpy/__init__.py @@ -8,4 +8,4 @@ from .optimization import Optimizer -__version__ = '0.4.1' +__version__ = '0.4.2' diff --git a/ebcpy/modelica/simres.py b/ebcpy/modelica/simres.py index dad8320e..9367c0ad 100644 --- a/ebcpy/modelica/simres.py +++ b/ebcpy/modelica/simres.py @@ -53,7 +53,6 @@ from itertools import count from collections import namedtuple from scipy.io import loadmat -from scipy.io.matlab.mio_utils import chars_to_strings import pandas as pd import numpy as np @@ -229,8 +228,8 @@ def get_strings(str_arr): Strip the whitespace from the right and recode it as utf-8. """ - return [line.rstrip(' \0').encode('latin-1').decode('utf-8') - for line in chars_to_strings(str_arr)] + return ["".join(word_arr).replace(" ", "") + for word_arr in str_arr] class Variable(namedtuple('VariableNamedTuple', ['samples', 'description', 'unit', 'displayUnit'])): diff --git a/requirements.txt b/requirements.txt index c61b2c8d..72b814ec 100644 --- a/requirements.txt +++ b/requirements.txt @@ -1,6 +1,6 @@ -numpy>=1.19.5 +numpy>=1.19.5,<2.0 matplotlib>=3.3.4 -scipy~=1.13.0 +scipy>=1.5.4 pandas>=1.1.5 tables>=3.6.1 scikit-learn>=0.24.2 diff --git a/setup.py b/setup.py index d8099392..9c6cba2a 100644 --- a/setup.py +++ b/setup.py @@ -19,9 +19,9 @@ } INSTALL_REQUIRES = [ - 'numpy>=1.19.5', + 'numpy>=1.19.5,<2.0', 'matplotlib>=3.3.4', - 'scipy~=1.13.0', + 'scipy>=1.5.4', 'pandas>=1.1.5', 'scikit-learn>=0.24.2', 'fmpy>=0.2.27,<0.3.17',

132 mio utils is deprecated Closes #132

2024-07-15T08:33:49

0.0

[]

RWTH-EBC/ebcpy

RWTH-EBC__ebcpy-116

060d10d3a657850343d02c64645b8de7b2d5a526

diff --git a/CHANGELOG.md b/CHANGELOG.md index 19d978b0..e44216e9 100644 --- a/CHANGELOG.md +++ b/CHANGELOG.md @@ -86,3 +86,5 @@ - Fix fmpy version #109 - Fix data format for index frequency of tsd #107 - Fix bug in function clean_and_space_equally_time_series #111 +- v0.3.11 + - Add dymola_exe_path to set of kwargs in DymolaAPI #115 diff --git a/ebcpy/__init__.py b/ebcpy/__init__.py index 5b0f57db..8699cb0e 100644 --- a/ebcpy/__init__.py +++ b/ebcpy/__init__.py @@ -8,4 +8,4 @@ from .optimization import Optimizer -__version__ = '0.3.10' +__version__ = '0.3.11' diff --git a/ebcpy/simulationapi/dymola_api.py b/ebcpy/simulationapi/dymola_api.py index c2672280..1835b7c4 100644 --- a/ebcpy/simulationapi/dymola_api.py +++ b/ebcpy/simulationapi/dymola_api.py @@ -56,10 +56,6 @@ class DymolaAPI(SimulationAPI): :keyword Boolean equidistant_output: If True (Default), Dymola stores variables in an equisdistant output and does not store variables at events. - :keyword str dymola_path: - Path to the dymola installation on the device. Necessary - e.g. on linux, if we can't find the path automatically. - Example: ``dymola_path="C://Program Files//Dymola 2020x"`` :keyword int n_restart: Number of iterations after which Dymola should restart. This is done to free memory. Default value -1. For values @@ -86,7 +82,7 @@ class DymolaAPI(SimulationAPI): If given, the Version needs to be equal to the folder name of your installation. - **Example:** If you have two version installed at + **Example:** If you have two versions installed at - ``C://Program Files//Dymola 2021`` and - ``C://Program Files//Dymola 2020x`` @@ -95,9 +91,18 @@ class DymolaAPI(SimulationAPI): ``dymola_version='Dymola 2020x'``. This parameter is overwritten if ``dymola_path`` is specified. + :keyword str dymola_path: + Path to the dymola installation on the device. Necessary + e.g. on linux, if we can't find the path automatically. + Example: ``dymola_path="C://Program Files//Dymola 2020x"`` :keyword str dymola_interface_path: - Only relevant for the case when the dymola-exe path + Direct path to the .egg-file of the dymola interface. + Only relevant when the dymola_path differs from the interface path. + :keyword str dymola_exe_path: + Direct path to the dymola executable. + Only relevant if the dymola installation do not follow + the official guideline. Example: @@ -129,7 +134,8 @@ class DymolaAPI(SimulationAPI): "mos_script_pre", "mos_script_post", "dymola_version", - "dymola_interface_path" + "dymola_interface_path", + "dymola_exe_path" ] def __init__(self, cd, model_name, packages=None, **kwargs): @@ -146,6 +152,7 @@ def __init__(self, cd, model_name, packages=None, **kwargs): self.mos_script_post = kwargs.pop("mos_script_post", None) self.dymola_version = kwargs.pop("dymola_version", None) self.dymola_interface_path = kwargs.pop("dymola_interface_path", None) + self.dymola_exe_path = kwargs.pop("dymola_exe_path", None) for mos_script in [self.mos_script_pre, self.mos_script_post]: if mos_script is not None: if not os.path.isfile(mos_script): @@ -175,31 +182,30 @@ def __init__(self, cd, model_name, packages=None, **kwargs): if not os.path.exists(dymola_path): raise FileNotFoundError(f"Given path '{dymola_path}' can not be found on " "your machine.") - _dym_install = dymola_path else: # Get the dymola-install-path: _dym_installations = self.get_dymola_install_paths() if _dym_installations: if self.dymola_version: - _found_version = False - for _dym_install in _dym_installations: - if _dym_install.endswith(self.dymola_version): - _found_version = True - break - if not _found_version: - raise ValueError( - f"Given dymola_version '{self.dymola_version}' not found in " - f"the list of dymola installations {_dym_installations}" - ) + dymola_path = _get_dymola_path_of_version( + dymola_installations=_dym_installations, + dymola_version=self.dymola_version + ) else: - _dym_install = _dym_installations[0] # 0 is the newest - self.logger.info("Using dymola installation at %s", _dym_install) + dymola_path = _dym_installations[0] # 0 is the newest + self.logger.info("Using dymola installation at %s", dymola_path) else: - raise FileNotFoundError("Could not find a dymola-interface on your machine.") - self.dymola_exe_path = self.get_dymola_path(_dym_install) + if self.dymola_exe_path is None or self.dymola_interface_path is None: + raise FileNotFoundError( + "Could not find dymola on your machine. " + "Thus, not able to find the `dymola_exe_path` and `dymola_interface_path`. " + "Either specify both or pass an existing `dymola_path`." + ) + if self.dymola_exe_path is None: + self.dymola_exe_path = self.get_dymola_path(dymola_path) self.logger.info("Using dymola.exe: %s", self.dymola_exe_path) if self.dymola_interface_path is None: - self.dymola_interface_path = self.get_dymola_interface_path(_dym_install) + self.dymola_interface_path = self.get_dymola_interface_path(dymola_path) self.logger.info("Using dymola interface: %s", self.dymola_interface_path) self.packages = [] @@ -1022,7 +1028,7 @@ def get_dymola_path(dymola_install_dir, dymola_name=None): """ if dymola_name is None: if "linux" in sys.platform: - dymola_name = "dymola" + dymola_name = "dymola.sh" elif "win" in sys.platform: dymola_name = "Dymola.exe" else: @@ -1166,3 +1172,18 @@ def _check_restart(self): self.sim_counter = 1 else: self.sim_counter += 1 + + +def _get_dymola_path_of_version(dymola_installations: list, dymola_version: str): + """ + Helper function to get the path associated to the dymola_version + from the list of all installations + """ + for dymola_path in dymola_installations: + if dymola_path.endswith(dymola_version): + return dymola_path + # If still here, version was not found + raise ValueError( + f"Given dymola_version '{dymola_version}' not found in " + f"the list of dymola installations {dymola_installations}" + )

calling dymola on linux maschin not working problem: not the right call of dymola via ebcpy - for running the simulation in dymola the python package ebcpy is used - ebcpy calls automaticlly '/opt/dymola-2023x-x86_64/bin64/dymola' - but here the needed libs are not found - it is needed to call '/opt/dymola-2023x-x86_64/bin64/dymola.sh' - solution: edit the file dymola_api.py of the package ebcpy - adapt line 1025: dymola_name = "dymola" > "dymola.sh" System: Ubuntu 22.04 dymola-2023x-x86_64 calling dymola on linux maschin not working problem: not the right call of dymola via ebcpy - for running the simulation in dymola the python package ebcpy is used - ebcpy calls automaticlly '/opt/dymola-2023x-x86_64/bin64/dymola' - but here the needed libs are not found - it is needed to call '/opt/dymola-2023x-x86_64/bin64/dymola.sh' - solution: edit the file dymola_api.py of the package ebcpy - adapt line 1025: dymola_name = "dymola" > "dymola.sh" System: Ubuntu 22.04 dymola-2023x-x86_64

Same as for the `dymola_interface_path`, I would just add the explicit path to the list of kwargs. This solution would not work for our dymola installation, so it seems to depend on the end user. Would that work for you as well? Do you mean I should add the kwarg dymola_interface_path to DymolaAPI()? I tried to use dymola_path here. but it does not work. Now, I tried he kwarg dymola_interface_path as well, but it does not work too. When dymola.sh is called, also the pathes of the libs are added. From my point of view. it is the problem. Sorry, that was not clear: I meant I would add a new option `dymola_exe_path`. So you could specify: ```DymolaAPI(dymola_exe_path='/opt/dymola-2023x-x86_64/bin64/dymola.sh', ...)``` Would that work for you? Another option is to specify a .mos script to run upon starting dymola, so all libs are called. What is the content of the .sh file? Maybe just a .mos file? thx, for your answers. Regarding path kwargs, when a new kwarg is used, the ebcpy respectively dymola_api.py must be adapted, right? This I like to avoid. Is there another possibility? Regarding libraries: the dymola.sh adds paths to load (old) system libraries, like libicui18n.so.60, which are part of the dymola installation. I looked into the dockerfile bim2sim is using. Here I find a symlink of the dymola.sh to /usr/local/bin/dymola. Maybe I will try this. Ah ok, if you use the Dockerfile from bim2sim, it will be similar to the AixLib ci. Here we use: ``` if "linux" in sys.platform: dymola_path = "/usr/local" dymola_interface_path = "/opt/dymola-2022-x86_64/Modelica/Library/python_interface/dymola.egg" ``` for kwargs. @Cudok and I discussed this in recent meeting. From my understanding using dymola.sh should also work with our dymola docker installation. Here is a screenshot from the setup in the dymola container. As you can see the dymola.sh exists and links correctly to the dymola installation. Also running the dymola.sh works ![image](https://github.com/RWTH-EBC/ebcpy/assets/27726960/8ca13512-5004-465a-955b-8e84a7f4dbf5) You can also see in the [Dockerfile](https://git.rwth-aachen.de/EBC/EBC_intern/dymola-docker/-/blame/master/Dockerfile#L62) that the symlink is created: `ln -s /opt/${DYM_FOLDER}/bin64/dymola.sh /usr/local/bin/dymola && \` As different Linux systems use different setups, we should make ebcpy usable "out of the box" for most of them. As @Cudok mentioned, he followed the official guideline installing Dymola under Linux (which we apparently didn't in our dockerfile) and this states to use alien which (could you add some information here @Cudok?) As @DaJansenGit mentioned, it is recommend to use alien to install the dymola rpm package in the full user manual (part of the installation see /documentation/). Following you find the relevant part from the full user manual: 17.2.1 Installing Dymola Dymola for Linux is distributed as an RPM package. The package is installed using the command `rpm –i name-of-distribution.rpm` Optional libraries are installed through separate RPM files. For installation on e.g. Debian or Kubuntu systems conversion to the deb format is required using the alien command: `alien –cki name-of-distribution.rpm` We should definitely follow the official guideline as a default. I will create a fix for this Same as for the `dymola_interface_path`, I would just add the explicit path to the list of kwargs. This solution would not work for our dymola installation, so it seems to depend on the end user. Would that work for you as well? Do you mean I should add the kwarg dymola_interface_path to DymolaAPI()? I tried to use dymola_path here. but it does not work. Now, I tried he kwarg dymola_interface_path as well, but it does not work too. When dymola.sh is called, also the pathes of the libs are added. From my point of view. it is the problem. Sorry, that was not clear: I meant I would add a new option `dymola_exe_path`. So you could specify: ```DymolaAPI(dymola_exe_path='/opt/dymola-2023x-x86_64/bin64/dymola.sh', ...)``` Would that work for you? Another option is to specify a .mos script to run upon starting dymola, so all libs are called. What is the content of the .sh file? Maybe just a .mos file? thx, for your answers. Regarding path kwargs, when a new kwarg is used, the ebcpy respectively dymola_api.py must be adapted, right? This I like to avoid. Is there another possibility? Regarding libraries: the dymola.sh adds paths to load (old) system libraries, like libicui18n.so.60, which are part of the dymola installation. I looked into the dockerfile bim2sim is using. Here I find a symlink of the dymola.sh to /usr/local/bin/dymola. Maybe I will try this. Ah ok, if you use the Dockerfile from bim2sim, it will be similar to the AixLib ci. Here we use: ``` if "linux" in sys.platform: dymola_path = "/usr/local" dymola_interface_path = "/opt/dymola-2022-x86_64/Modelica/Library/python_interface/dymola.egg" ``` for kwargs. @Cudok and I discussed this in recent meeting. From my understanding using dymola.sh should also work with our dymola docker installation. Here is a screenshot from the setup in the dymola container. As you can see the dymola.sh exists and links correctly to the dymola installation. Also running the dymola.sh works ![image](https://github.com/RWTH-EBC/ebcpy/assets/27726960/8ca13512-5004-465a-955b-8e84a7f4dbf5) You can also see in the [Dockerfile](https://git.rwth-aachen.de/EBC/EBC_intern/dymola-docker/-/blame/master/Dockerfile#L62) that the symlink is created: `ln -s /opt/${DYM_FOLDER}/bin64/dymola.sh /usr/local/bin/dymola && \` As different Linux systems use different setups, we should make ebcpy usable "out of the box" for most of them. As @Cudok mentioned, he followed the official guideline installing Dymola under Linux (which we apparently didn't in our dockerfile) and this states to use alien which (could you add some information here @Cudok?) As @DaJansenGit mentioned, it is recommend to use alien to install the dymola rpm package in the full user manual (part of the installation see /documentation/). Following you find the relevant part from the full user manual: 17.2.1 Installing Dymola Dymola for Linux is distributed as an RPM package. The package is installed using the command `rpm –i name-of-distribution.rpm` Optional libraries are installed through separate RPM files. For installation on e.g. Debian or Kubuntu systems conversion to the deb format is required using the alien command: `alien –cki name-of-distribution.rpm` We should definitely follow the official guideline as a default. I will create a fix for this

2023-12-28T10:14:29

0.0

[]

RWTH-EBC/ebcpy

RWTH-EBC__ebcpy-106

5a56a8e8ec6633d2ac5a8c204967b93166b2ec08

diff --git a/ebcpy/modelica/manipulate_ds.py b/ebcpy/modelica/manipulate_ds.py index 0ef51333..922c0595 100644 --- a/ebcpy/modelica/manipulate_ds.py +++ b/ebcpy/modelica/manipulate_ds.py @@ -2,6 +2,8 @@ dsfinal.txt and dsin.txt files created by Modelica.""" from io import StringIO +import re + import pandas as pd @@ -37,17 +39,24 @@ def convert_ds_file_to_dataframe(filename): :return: pd.DataFrame Converted DataFrame """ - # Define relevant parameters - number_line_initial_name = 104 # Line where the string char initialName(,) is always stored - # Open file and get relevant content by splitting the lines. with open(filename, "r") as file: content = file.read().split("\n") # Gets the X out of 'char initialName(X,Y)' - size_initial_names = int(content[number_line_initial_name].split("(")[-1].split(",")[0]) + pattern_size_initial_name = r'^char initialName$(\d+),(\d+)$$' + for number_line_initial_name, line in enumerate(content): + match_size_initial_name = re.match(pattern_size_initial_name, line) + if match_size_initial_name: + size_initial_names = int(match_size_initial_name.string.split("(")[-1].split(",")[0]) + break + else: + raise ValueError("Could not find initial names in file") + # Number of line below line "double initialValue(X,Y)" number_line_initial_value = number_line_initial_name + size_initial_names + 3 + if "double initialValue" in content[number_line_initial_value]: + number_line_initial_value += 1 # Check if two or on-line dsfinal / dsin if "#" in content[number_line_initial_value]:

Fix model translation in Dymola 2023 Dymola 2023 seems to throw errors when loading dsin files. This needs further investigation and fix. Example ``` File "...\ebcpy\ebcpy\modelica\manipulate_ds.py", line 48, in convert_ds_file_to_dataframe size_initial_names = int(content[number_line_initial_name].split("(")[-1].split(",")[0]) ValueError: invalid literal for int() with base 10: 'multizone.zoneParam[1].AExt[3]' ``` Fix model translation in Dymola 2023 Dymola 2023 seems to throw errors when loading dsin files. This needs further investigation and fix. Example ``` File "...\ebcpy\ebcpy\modelica\manipulate_ds.py", line 48, in convert_ds_file_to_dataframe size_initial_names = int(content[number_line_initial_name].split("(")[-1].split(",")[0]) ValueError: invalid literal for int() with base 10: 'multizone.zoneParam[1].AExt[3]' ```

A temporary fix is to disable this feature in the DymolaAPI by using: `extract_variables=False` as a keyword argument. A temporary fix is to disable this feature in the DymolaAPI by using: `extract_variables=False` as a keyword argument.

2023-10-02T11:42:37

0.0

[]

RWTH-EBC/ebcpy

RWTH-EBC__ebcpy-60

50057aa10db14c2636c0ae64ad760f7c30a4f419

diff --git a/CHANGELOG.md b/CHANGELOG.md index e36fd82f..51792076 100644 --- a/CHANGELOG.md +++ b/CHANGELOG.md @@ -58,4 +58,6 @@ - v0.3.0: - Add parallelization option for optimization backends - v0.3.1: - - Avoid DType Warning #53 \ No newline at end of file + - Avoid DType Warning #53 +- v0.3.2: + - Correct bounds for simulation Variables #56 \ No newline at end of file diff --git a/docs/source/conf.py b/docs/source/conf.py index 1bec663a..be7ae0e8 100644 --- a/docs/source/conf.py +++ b/docs/source/conf.py @@ -71,7 +71,7 @@ # The short X.Y version. version = '0.3' # The full version, including alpha/beta/rc tags. -release = '0.3.1' +release = '0.3.2' # The language for content autogenerated by Sphinx. Refer to documentation # for a list of supported languages. diff --git a/ebcpy/__init__.py b/ebcpy/__init__.py index b90ea14b..ee001230 100644 --- a/ebcpy/__init__.py +++ b/ebcpy/__init__.py @@ -8,4 +8,4 @@ from .optimization import Optimizer -__version__ = '0.3.1' +__version__ = '0.3.2' diff --git a/ebcpy/simulationapi/__init__.py b/ebcpy/simulationapi/__init__.py index 3a2c6649..ea17f82e 100644 --- a/ebcpy/simulationapi/__init__.py +++ b/ebcpy/simulationapi/__init__.py @@ -5,6 +5,7 @@ much more user-friendly than the provided APIs by Dymola or fmpy. """ import logging +import warnings import os import itertools from typing import Dict, Union, TypeVar, Any, List @@ -20,25 +21,60 @@ class Variable(BaseModel): Data-Class to store relevant information for a simulation variable (input, parameter, output or local/state). """ + type: Any = Field( + default=None, + title='type', + description='Type of the variable' + ) value: Any = Field( description="Default variable value" ) - max: Union[float, int] = Field( - default=np.inf, + max: Any = Field( + default=None, title='max', - description='Maximal value (upper bound) of the variables value' - ) - min: Union[float, int] = Field( - default=-np.inf, - title='min', - description='Minimal value (lower bound) of the variables value' + description='Maximal value (upper bound) of the variables value. ' + 'Only for ints and floats variables.' ) - type: Any = Field( + min: Any = Field( default=None, - title='type', - description='Type of the variable' + title='min', + description='Minimal value (lower bound) of the variables value. ' + 'Only for ints and floats variables.' ) + @validator("value") + def check_value_type(cls, value, values): + """Check if the given value has correct type""" + _type = values["type"] + if _type is None: + return value # No type -> no conversion + if value is None: + return value # Setting None is allowed. + if not isinstance(value, _type): + return _type(value) + return value + + @validator('max', 'min', always=True) + def check_value(cls, value, values, field): + """Check if the given bounds are correct.""" + # Check if the variable type even allows for min/max bounds + _type = values["type"] + if _type is None: + return value # No type -> no conversion + if _type not in (float, int, bool): + if value is not None: + warnings.warn( + "Setting a min/max for variables " + f"of type {_type} is not supported." + ) + return None + if value is not None: + return _type(value) + if field.name == "min": + return -np.inf if _type != bool else False + # else it is max + return np.inf if _type != bool else True + class SimulationSetup(BaseModel): """ diff --git a/ebcpy/simulationapi/fmu.py b/ebcpy/simulationapi/fmu.py index 773bb24a..243817fb 100644 --- a/ebcpy/simulationapi/fmu.py +++ b/ebcpy/simulationapi/fmu.py @@ -287,11 +287,6 @@ def setup_fmu_instance(self): else: self._fmi_type = 'CoSimulation' - def _to_bound(value): - if value is None or \ - not isinstance(value, (float, int, bool)): - return np.inf - return value self.logger.info("Reading model variables") _types = { @@ -307,8 +302,8 @@ def _to_bound(value): var.start = _types[var.type](var.start) _var_ebcpy = Variable( - min=-_to_bound(var.min), - max=_to_bound(var.max), + min=var.min, + max=var.max, value=var.start, type=_types[var.type] ) diff --git a/setup.py b/setup.py index fb1aec4c..acf2e22e 100644 --- a/setup.py +++ b/setup.py @@ -12,7 +12,7 @@ 'full': [ 'openpyxl>=3.0.5', 'xlrd>=2.0.1', - 'pymoo>=0.4.2', + 'pymoo==0.4.2', ] } @@ -32,7 +32,7 @@ # Add all open-source packages to setup-requires SETUP_REQUIRES = INSTALL_REQUIRES.copy() -VERSION = "0.3.1" +VERSION = "0.3.2" setuptools.setup( name='ebcpy',

Fix parameter bounds in class FMU_API Currently, the bounds of the parameters in an FMU_API are always set to minus infinity and infinity.

2022-08-22T14:01:22

0.0

[]

django/daphne

django__daphne-500

9a282dd6277f4b921e494a9fc30a534a7f0d6ca6

diff --git a/daphne/http_protocol.py b/daphne/http_protocol.py index f0657fdb..b7da1bf7 100755 --- a/daphne/http_protocol.py +++ b/daphne/http_protocol.py @@ -9,7 +9,7 @@ from twisted.web import http from zope.interface import implementer -from .utils import parse_x_forwarded_for +from .utils import HEADER_NAME_RE, parse_x_forwarded_for logger = logging.getLogger(__name__) @@ -69,6 +69,13 @@ def __init__(self, *args, **kwargs): def process(self): try: self.request_start = time.time() + + # Validate header names. + for name, _ in self.requestHeaders.getAllRawHeaders(): + if not HEADER_NAME_RE.fullmatch(name): + self.basic_error(400, b"Bad Request", "Invalid header name") + return + # Get upgrade header upgrade_header = None if self.requestHeaders.hasHeader(b"Upgrade"): diff --git a/daphne/utils.py b/daphne/utils.py index 81f1f9df..06993142 100644 --- a/daphne/utils.py +++ b/daphne/utils.py @@ -1,7 +1,12 @@ import importlib +import re from twisted.web.http_headers import Headers +# Header name regex as per h11. +# https://github.com/python-hyper/h11/blob/a2c68948accadc3876dffcf979d98002e4a4ed27/h11/_abnf.py#L10-L21 +HEADER_NAME_RE = re.compile(rb"[-!#$%&'*+.^_`|~0-9a-zA-Z]+") + def import_by_path(path): """

Daphne allows invalid characters within header names # The bug RFC 9110 defines the header names must consist of a single token, which is defined as follows: ``` token = 1*tchar tchar = "!" / "#" / "$" / "%" / "&" / "'" / "*" / "+" / "-" / "." / "^" / "_" / "`" / "|" / "~" / DIGIT / ALPHA ; any VCHAR, except delimiters ``` When Daphne receives a request with a header name containing any or all of the following characters, it does not reject the message: ``` \x00\x01\x02\x03\x04\x05\x06\x07\x08\t\x0b\x0c\x0e\x0f\x10\x11\x12\x13\x14\x15\x16\x17\x18\x19\x1a\x1b\x1c\x1d\x1e\x1f "(),/;<=>?@[/]{}\xc0\xc1\xc2\xc3\xc4\xc5\xc6\xc7\xc8\xc9\xca\xcb\xcc\xcd\xce\xcf\xd0\xd1\xd2\xd3\xd4\xd5\xd6\xd7\xd8\xd9\xda\xdb\xdc\xdd\xde\xdf\xe0\xe1\xe2\xe3\xe4\xe5\xe6\xe7\xe8\xe9\xea\xeb\xec\xed\xee\xef ``` Thus, when you send the following request: ``` GET / HTTP/1.1\r\n \x00\x01\x02\x03\x04\x05\x06\x07\x08\t\x0b\x0c\x0e\x0f\x10\x11\x12\x13\x14\x15\x16\x17\x18\x19\x1a\x1b\x1c\x1d\x1e\x1f "(),/;<=>?@[/]{}\xc0\xc1\xc2\xc3\xc4\xc5\xc6\xc7\xc8\xc9\xca\xcb\xcc\xcd\xce\xcf\xd0\xd1\xd2\xd3\xd4\xd5\xd6\xd7\xd8\xd9\xda\xdb\xdc\xdd\xde\xdf\xe0\xe1\xe2\xe3\xe4\xe5\xe6\xe7\xe8\xe9\xea\xeb\xec\xed\xee\xef: whatever\r\n \r\n ``` Daphne's interpretation is as follows: ``` [ HTTPRequest( method=b'GET', uri=b'/', version=b'1.1', headers=[ (b'\x00\x01\x02\x03\x04\x05\x06\x07\x08\t\x0b\x0c\x0e\x0f\x10\x11\x12\x13\x14\x15\x16\x17\x18\x19\x1a\x1b\x1c\x1d\x1e\x1f "(),/;<=>?@[/]{}\xc0\xc1\xc2\xc3\xc4\xc5\xc6\xc7\xc8\xc9\xca\xcb\xcc\xcd\xce\xcf\xd0\xd1\xd2\xd3\xd4\xd5\xd6\xd7\xd8\xd9\xda\xdb\xdc\xdd\xde\xdf\xe0\xe1\xe2\xe3\xe4\xe5\xe6\xe7\xe8\xe9\xea\xeb\xec\xed\xee\xef', b'whatever'), ], body=b'', ), ] ``` (i.e. the invalid header name made it into the request unchanged) Nearly all other HTTP implementations reject the above request, including Uvicorn, Hypercorn, AIOHTTP, Apache httpd, Bun, CherryPy, Deno, FastHTTP, Go net/http, Gunicorn, H2O, Hyper, Jetty, libsoup, Lighttpd, Mongoose, Nginx, Node.js LiteSpeed, Passenger, Puma, Tomcat, OpenWrt uhttpd, Uvicorn, Waitress, WEBrick, and OpenBSD httpd. ## Your OS and runtime environment: ``` $ uname -a Linux de8a63fbb2f4 6.7.2-arch1-2 #1 SMP PREEMPT_DYNAMIC Wed, 31 Jan 2024 09:22:15 +0000 x86_64 GNU/Linux ``` ## `pip freeze`: ``` asgiref==3.7.2 attrs==23.2.0 autobahn==23.6.2 Automat==22.10.0 cffi==1.16.0 constantly==23.10.4 cryptography==42.0.2 Cython==0.29.32 daphne @ file:///app/daphne h2==4.1.0 hpack==4.0.0 hyperframe==6.0.1 hyperlink==21.0.0 idna==3.6 incremental==22.10.0 priority==1.3.0 pyasn1==0.5.1 pyasn1-modules==0.3.0 pycparser==2.21 Pygments==2.14.0 pyOpenSSL==24.0.0 python-afl==0.7.4 PyYAML==6.0 service-identity==24.1.0 six==1.16.0 Twisted==23.10.0 txaio==23.1.1 typing_extensions==4.9.0 zope.interface==6.1 ``` (Using Daphne built from main, with the latest commit being https://github.com/django/daphne/commit/993efe62ce9e9c1cd64c81b6ee7dfa7b819482c7 at the time of writing) ## How you're running Channels (runserver? daphne/runworker? Nginx/Apache in front?) Directly invoking Daphne from the command line. ## Console logs and full tracebacks of any errors N/A

Of particular note is that `\x00`, `\t`, and ` ` are permitted, which have historically been sources of exploitable parsing discrepancies between origin servers and gateway servers. @kenballus thanks for the report. Are you up for making a PR here? h11 gives a regex for a header name here: https://github.com/python-hyper/h11/blob/a2c68948accadc3876dffcf979d98002e4a4ed27/h11/_abnf.py#L10-L21C1 That's then validated when parsing the headers here: https://github.com/python-hyper/h11/blob/a2c68948accadc3876dffcf979d98002e4a4ed27/h11/_headers.py#L163 Gunicorn does similar here https://github.com/benoitc/gunicorn/blob/cf55d2cec277f220ebd605989ce78ad1bb553c46/gunicorn/http/message.py#L92-L93 (but matching bad headers, with what looks like a more complex regex.) We'd handle that in the http and ws protocols: https://github.com/django/daphne/blob/993efe62ce9e9c1cd64c81b6ee7dfa7b819482c7/daphne/http_protocol.py#L147-L158 https://github.com/django/daphne/blob/993efe62ce9e9c1cd64c81b6ee7dfa7b819482c7/daphne/ws_protocol.py#L37-L44 Need to investigate whether this should be handled by twisted. https://github.com/twisted/twisted/blob/446ee139189440e890b26a29af256e9b9d0e8eba/src/twisted/web/http_headers.py h11 test examples for `no weird characters in names` are here: https://github.com/python-hyper/h11/blob/a2c68948accadc3876dffcf979d98002e4a4ed27/h11/tests/test_headers.py#L24-L35

2024-02-10T14:23:01

0.0

[]

django/daphne

django__daphne-422

6199d509c200cdd5102b39a3461a62984adebaf7

diff --git a/CHANGELOG.txt b/CHANGELOG.txt index 31e46504..8eb54634 100644 --- a/CHANGELOG.txt +++ b/CHANGELOG.txt @@ -20,6 +20,12 @@ Unreleased * Added ``--log-fmt`` CLI argument. +* Added support for ``ASGI_THREADS`` environment variable, setting the maximum + number of workers used by a ``SyncToAsync`` thread-pool executor. + + Set e.g. ``ASGI_THREADS=4 daphne ...`` when running to limit the number of + workers. + 3.0.2 (2021-04-07) ------------------ diff --git a/daphne/server.py b/daphne/server.py index 43342175..e5728b3c 100755 --- a/daphne/server.py +++ b/daphne/server.py @@ -1,10 +1,18 @@ # This has to be done first as Twisted is import-order-sensitive with reactors import asyncio # isort:skip +import os # isort:skip import sys # isort:skip import warnings # isort:skip +from concurrent.futures import ThreadPoolExecutor # isort:skip from twisted.internet import asyncioreactor # isort:skip + twisted_loop = asyncio.new_event_loop() +if "ASGI_THREADS" in os.environ: + twisted_loop.set_default_executor( + ThreadPoolExecutor(max_workers=int(os.environ["ASGI_THREADS"])) + ) + current_reactor = sys.modules.get("twisted.internet.reactor", None) if current_reactor is not None: if not isinstance(current_reactor, asyncioreactor.AsyncioSelectorReactor):

ASGI_THREADS environment variable has no effect since 2.4.1 Channels documentation mentions that: > By default, the number of threads is set to “the number of CPUs * 5”, so you may see up to this number of threads. If you want to change it, set the ASGI_THREADS environment variable to the maximum number you wish to allow. This behaviour used to work correctly before Daphne v2.4.1. Since 2.4.1 setting ASGI_THREADS envvar has no effect. #### My Setup Python 3.7.5, on Ubuntu 19.10 with following packages: ``` asgiref==3.2.7 channels==2.4.0 channels-redis==2.4.2 daphne==2.4.1 Django==2.2.12 ``` Startup: ``` daphne -b 0.0.0.0 -p 8000 api.config.asgi:application ``` #### How to replicate * Set ASGI_THREADS to 1. * Install packages above, with Daphne v2.40 * Start daphne * Navigate through website and open a few pages * Check number of threads with htop. * Thread count never exceeds 1 * Now install daphne v2.41 * Start server and browse website * Thread count is not limited to 1, keeps increasing. Possibly capped to 5*CPU count, but can't tell for sure.

Nothing related to that was changed in 2.4.1, that stuff is all managed by asgiref. The thread stuff is a lot better in python 3.8 but imo the whole ThreadPoolExecutor design is terrible. @JohnDoee It did, it is this commit https://github.com/django/daphne/commit/27f760a8147efd0c2f4cb842cbe9d50cfdf40d46. Reverting this change resolves the issue. I tested it, sorry, you're right. I assumed the correct loop was installed as global correctly. It seems like the "correct" ThreadPoolExecutor is installed on the wrong loop. Hmmm. Yes. It's because `SyncToAsync` is using `get_event_loop()` which is returning the default event loop for the main thread, rather than the loop we're passing to twisted (for use in the running thread). I assumed that https://github.com/django/daphne/blob/2.4.1/daphne/server.py#L129 would make it all correct but maybe stuff is executed in the wrong order. I've been testing this a bit further, this actually breaks everything that uses asyncio and is initialized at the "wrong" time. The original loop seems to be stopped. Now I have this piece of boilerplate code at the top somewhere imported early to make sure everything else uses the correct loop. ``` import asyncio # isort:skip import daphne.server # isort:skip from twisted.internet import reactor # isort:skip asyncio.set_event_loop(reactor._asyncioEventloop) # isort:skip ``` See https://github.com/django/asgiref/issues/185#issuecomment-667948950: > We need to resolve, probably reverting django/daphne@27f760a — but then that recreates django/channels#1374 — so it's currently pending time to think about it. #299 is related. So I have been looking through asgiref, channels daphne etc… and I came to the following conclusion: Remove `ASGI_THREADS` support from asgiref. Before you say no, hear me out… * `asgiref` is a low level library which basically only requires an event loop to be running, it should not concern itself with more. * It does not know from which thread it is imported and as such it shouldn't make assumption about the event loop there, * `get_event_loop` which is currently used is slated for deprecation and it appears as if only `get_running_loop` will be supported going forward -- which is not something that will exist during import time usually. * Even if the event loop where running and we were able to get it at import time, does it sound safe to change the associated thread pool while the event loop is running? (does python even support that?) Okay, now with `ASGI_THREADS` removed we need to bring back the functionality. How do we do that? As stopgap measure we can simply add it to `daphne/server.py` right after we call `new_event_loop` -- it is our event loop after all, so let's see the pool there. Going further we can deprecate `ASGI_THREADS` completely and implement this in the ASGI servers them self, they are the ones in control (and owners) of the loop after all. So for daphne I'd expect support for `--worker-threads` (or so…), which `manage.py runserver` from channels could also use then. What do you think? > See [django/asgiref#185 (comment)](https://github.com/django/asgiref/issues/185#issuecomment-667948950): > > > We need to resolve, probably reverting [27f760a](https://github.com/django/daphne/commit/27f760a) — but then that recreates [django/channels#1374](https://github.com/django/channels/issues/1374) — so it's currently pending time to think about it. @carltongibson: I started to think about that and ran over your comment in https://github.com/django/channels/issues/1374#issuecomment-566705986 -- when I tried this today on Python 3.10 and up2date asgiref/django I did not get `INCORRECTLY raised SynchronousOnlyOperation` but only: ``` /home/florian/sources/async/test3.py:39: DeprecationWarning: There is no current event loop target=run_in_async_thread, args=(asyncio.get_event_loop(),), kwargs={}, name='async-thread' Async thread starting Will start the loop About to call sync_only from async context Correctly raised SynchronousOnlyOperation About to call sync_only from SYNC context sync_only successfully called. ``` with that in mind it might (?) be possible to actually revert the changes. That said I would recommend against reverting because creating our own event loop and configuring it as needed sounds like the way to go… Nevertheless I'd be interested to know if you have an idea whether python fixed something or we fixed a few bugs in asgiref along the way… Hey @apollo13 — thanks for digging. Current status is _I don't know_. My plan (since late Summer) has been to sit down with Daphne and resolve this issue properly. Life has intervened but, it's my top item (having just got 4.0 compatibility going for django-compressor). So reality is more work (from me) on the Channels Trio in the new year now. However, if you're digging actively I am very happy to work with you on that. > That said I would recommend against reverting because creating our own event loop and configuring it as needed sounds like the way to go… Yes, it seemed like a good idea. The issue was that it turns out that you need to set the event loop in your thread if you're not using the default, and whilst that's possible it creates a bookkeeping overhead, that folks don't want to do. I began a conversation with @orf on the Forum about not using thread for the auto-reloader that was prompted by this issue. https://forum.djangoproject.com/t/feasibility-of-using-subprocess-for-auto-reloader/4970 — That's the way I'd like to go (but that's a lot of Yak...) Hi @carltongibson, while it is true that changes to the autoreloader might be possible; I still think that `ASGI_THREADS` doesn't belong into `asgiref` and we should fix that independent of further improvements in the autoreloader, ie what I outlined in https://github.com/django/daphne/issues/319#issuecomment-991962381 -- is that something we would be okay with or a big nogo (I am asking because this is basically the first time doing anything with asgi and as such I am not very up2date on it's history) > what I outlined in #319 (comment) Certainly happy for Daphne to have that. There needs to be some hard upper limit somewhere, and the protocol server seems the right level for that. (I can't recall without looking if there are other uses of `ASGI_THREADS` in asgiref that would allow dropping it entirely there.) > Nevertheless I'd be interested to know if you have an idea whether python fixed something or we fixed a few bugs in asgiref along the way… OK, so I **finally** had the combination of time/energy/health to look into this (and the related cluster of issues) again. The bottom-line is that I **think** dropping PY36, and `get_event_loop()` in favour of `get_running_loop()` across the board (and particularly in `asgiref.sync`) means that this issue is likely _Gone Away™_ (for some value of that). * Using `get_running_loop()` will get the right event loop, generally * We already call `set_event_loop` in the `Server`, so making sure we did similar in _Worker_ scripts would likely be enough. * Given that, it may be that we **don't** need to worry about the reloader running in a sub-thread (not sure yet) * But I tried **deferring** the twisted reactor setup in server.py, which doesn't look possible. * So, maybe, I can rework the Channels `runserver` to run with a subprocess if that's still needed. That's very preliminary. (I only had the session today to look at it) But maybe the pandemic induced don't look at it for two years has paid off. I'll keep working on this, but it feels like what was intractable now at least has a way of being untangled. OK, so step one is RFC to remove ASGI_THREADS from asgiref: https://github.com/django/asgiref/pull/336 Then we'll bring similar in here to restore this functionality to Daphne.

2022-07-06T10:27:51

0.0

[]

django/daphne

django__daphne-417

eae1ff0df4603843d412feb9f0a27dff5f76c471

diff --git a/.github/workflows/pre-commit.yml b/.github/workflows/pre-commit.yml deleted file mode 100644 index d9394321..00000000 --- a/.github/workflows/pre-commit.yml +++ /dev/null @@ -1,24 +0,0 @@ -name: pre-commit - -on: - push: - branches: - - main - pull_request: - -jobs: - pre-commit: - runs-on: ubuntu-20.04 - - steps: - - uses: actions/checkout@v2 - with: - fetch-depth: 0 - - - uses: actions/setup-python@v2 - with: - python-version: 3.9 - - - uses: pre-commit/[email protected] - with: - token: ${{ secrets.GITHUB_TOKEN }} diff --git a/.pre-commit-config.yaml b/.pre-commit-config.yaml index 5de3ae5a..1367879b 100644 --- a/.pre-commit-config.yaml +++ b/.pre-commit-config.yaml @@ -1,20 +1,20 @@ repos: - repo: https://github.com/asottile/pyupgrade - rev: v2.11.0 + rev: v2.32.1 hooks: - id: pyupgrade args: [--py36-plus] - repo: https://github.com/psf/black - rev: 20.8b1 + rev: 22.3.0 hooks: - id: black language_version: python3 - repo: https://github.com/pycqa/isort - rev: 5.8.0 + rev: 5.10.1 hooks: - id: isort - repo: https://github.com/PyCQA/flake8 - rev: 3.9.0 + rev: 4.0.1 hooks: - id: flake8 additional_dependencies: diff --git a/CHANGELOG.txt b/CHANGELOG.txt index 8463751a..7e6a1207 100644 --- a/CHANGELOG.txt +++ b/CHANGELOG.txt @@ -1,3 +1,16 @@ +Unreleased +---------- + +* Dropped support for Python 3.6. + +* Updated dependencies to the latest versions. + + Previously a range of Twisted versions have been supported. Recent Twisted + releases (22.2, 22.4) have issued security fixes, so those are now the + minimum supported version. Given the stability of Twisted, supporting a + range of versions does not represent a good use of maintainer time. Going + forward the latest Twisted version will be required. + 3.0.2 (2021-04-07) ------------------ diff --git a/README.rst b/README.rst index 7525b27c..70c4ba58 100644 --- a/README.rst +++ b/README.rst @@ -108,7 +108,7 @@ should start with a slash, but not end with one; for example:: Python Support -------------- -Daphne requires Python 3.6 or later. +Daphne requires Python 3.7 or later. Contributing diff --git a/setup.cfg b/setup.cfg index ddb3d5de..e7c19e55 100644 --- a/setup.cfg +++ b/setup.cfg @@ -8,3 +8,4 @@ profile = black [tool:pytest] testpaths = tests +asyncio_mode = strict diff --git a/setup.py b/setup.py index 214df0d0..274e8f89 100755 --- a/setup.py +++ b/setup.py @@ -22,8 +22,8 @@ package_dir={"twisted": "daphne/twisted"}, packages=find_packages() + ["twisted.plugins"], include_package_data=True, - install_requires=["twisted[tls]>=19.7", "autobahn>=0.18", "asgiref>=3.2.10,<4"], - python_requires=">=3.6", + install_requires=["twisted[tls]>=22.4", "autobahn>=22.4.2", "asgiref>=3.5.2,<4"], + python_requires=">=3.7", setup_requires=["pytest-runner"], extras_require={"tests": ["hypothesis", "pytest", "pytest-asyncio"]}, entry_points={ @@ -37,7 +37,6 @@ "Operating System :: OS Independent", "Programming Language :: Python", "Programming Language :: Python :: 3", - "Programming Language :: Python :: 3.6", "Programming Language :: Python :: 3.7", "Programming Language :: Python :: 3.8", "Programming Language :: Python :: 3.9", diff --git a/tox.ini b/tox.ini index da7d4d9f..876ff995 100644 --- a/tox.ini +++ b/tox.ini @@ -1,12 +1,8 @@ [tox] envlist = - py{36,37,38,39,310}-twisted{197,latest} - + py{37,38,39,310} [testenv] usedevelop = true extras = tests commands = pytest -v {posargs} -deps = - twisted187: twisted==19.7.0 - twistedlatest: twisted>=20.3.0

pre-commit failing It looks like the pre-commit GitHub action is failing to install setuptools properly: https://github.com/django/daphne/actions/workflows/pre-commit.yml The action is deprecated: https://github.com/pre-commit/action I suggest we switch to using https://pre-commit.ci

Turns out I have permission to do this - but I don't want to just commit (heh) to it, especially since the issue I opened on Django ( https://github.com/django/django/pull/15178 ) hasn't seen any response. Any thoughts @carltongibson ? > Any thoughts @carltongibson ? 😝 Little bit sceptical about the value-add for django/django **but** if you want to give it a run here, that's fine I guess. It may be a raging success and convert the sceptics. Thanks @adamchainz Hey @adamchainz -- shall we give this a go here, and on Channels and channels_redis? Gives us a fair taste, independently of django/django

2022-05-23T10:17:05

0.0

[]

astronomer/astronomer-cosmos

astronomer__astronomer-cosmos-1097

5b4c5d0b454ea207dea2ce0402d09193afe3930c

diff --git a/cosmos/cache.py b/cosmos/cache.py index d519eaca4..32a21d30d 100644 --- a/cosmos/cache.py +++ b/cosmos/cache.py @@ -230,9 +230,12 @@ def _create_folder_version_hash(dir_path: Path) -> str: filepaths.extend(paths) for filepath in sorted(filepaths): - with open(str(filepath), "rb") as fp: - buf = fp.read() - hasher.update(buf) + try: + with open(str(filepath), "rb") as fp: + buf = fp.read() + hasher.update(buf) + except FileNotFoundError: + logger.warning(f"The dbt project folder contains a symbolic link to a non-existent file: {filepath}") return hasher.hexdigest()

Bug on Cosmos 1.5 `No such file or directory "/usr/local/aiflow/dags/dbt/dbt_venv/bin/python"` Two Astro customers reported that when updating to Cosmos 1.5, they started seeing the error: ``` No such file or directory "/usr/local/aiflow/dags/dbt/dbt_venv/bin/python" ``` As illustrated in the two screenshots below: ![Image](https://github.com/user-attachments/assets/148fc806-c112-42f6-ace3-958e2386e7fa) ![Image](https://github.com/user-attachments/assets/859eb89e-8157-4d93-8492-052f4a7b1500)

After lots of time trying to reproduce and understand this issue, I narrowed it down: - inside the customers dbt project folder, there was a symbolic link to a no longer existent file Cosmos was raising an exception while trying to parse the file and create a version hash to represent the version of that dir. I validated a possible solution with one of the customers earlier today, available in Cosmos 1.5.1a3 - and they confirmed it solved the problem. I'll make a quick PR with the solution and unittests today

2024-07-17T08:40:30

0.0

[]

it-is-me-mario/MARIO

it-is-me-mario__MARIO-121

8897b6963fd484f50c6b7a1ed5b5e1f20493ef46

diff --git a/CHANGELOG.rst b/CHANGELOG.rst index 931f81e..d3cf7b5 100644 --- a/CHANGELOG.rst +++ b/CHANGELOG.rst @@ -2,11 +2,27 @@ Release History **************** +v 0.3.4 +------- + +Parsing functions error fixed +~~~~~~~~ +Recent pandas versions have changed the way they interpret "None" in DataFrames indices and values, which are currently interpreted it as NaN. +This mario update fixes the issue by replacing NaN with the string "None" when parsing excel files. + +Deprecated functions +~~~~~~~~ +Parser for old-fashioned Eurostat SUTs is deprecated. This function relied on peculiarly structured SUTs formats. +In case you need to parse such SUTs, please rearrange them into the standard MARIO format. +You can check the MARIO format from 'SUT.xlsx' file in the mario/test directory in this repository. + + v 0.3.3 ------- Settings ~~~~~~~~ +to_excel function bug in flow mode fixed. v 0.3.0 @@ -54,4 +70,4 @@ Documentation ~~~~~~~~~~~~~ * The tutorials are updated to improve the readiblity and quality of the juputer notebook functionalities. -* New templates for the readthedocs. +* New templates for the readthedocs. \ No newline at end of file diff --git a/doc/source/api_refernces.rst b/doc/source/api_refernces.rst index 6efada1..939b047 100644 --- a/doc/source/api_refernces.rst +++ b/doc/source/api_refernces.rst @@ -100,7 +100,6 @@ Structured Databases mario supports automatic parsing of following database: * Exiobase -* Eurostat Supply and Use * Eora26 * Eora single region @@ -109,7 +108,6 @@ mario supports automatic parsing of following database: mario.parse_exiobase_3 mario.parse_exiobase_sut - mario.parse_eurostat_sut mario.parse_eora mario.parse_from_pymrio mario.parse diff --git a/doc/source/conf.py b/doc/source/conf.py index 6efda1a..fe37a8a 100644 --- a/doc/source/conf.py +++ b/doc/source/conf.py @@ -26,7 +26,7 @@ author = 'Mohammad Amin Tahavori, Lorenzo Rinaldi, Nicolo Golinucci' # The full version, including alpha/beta/rc tags -release = '0.3.1' +release = '0.3.4' # -- General configuration --------------------------------------------------- diff --git a/mario/__init__.py b/mario/__init__.py index 6080439..67d2ba0 100644 --- a/mario/__init__.py +++ b/mario/__init__.py @@ -43,6 +43,9 @@ """ + +import pandas as pd + from mario.version import __version__ from mario.core.AttrData import Database from mario.core.CoreIO import CoreModel @@ -99,4 +102,5 @@ ) from mario.tools.constants import IOT, SUT + __authors__ = " 'Mohammad Amin Tahavori', Lorenzo Rinaldi', 'Nicolò Golinucci' " diff --git a/mario/core/AttrData.py b/mario/core/AttrData.py index a475785..3a19cb3 100644 --- a/mario/core/AttrData.py +++ b/mario/core/AttrData.py @@ -477,6 +477,7 @@ def read_aggregated_index( # Reading the aggregation data from a single Excel file with differetn sheets ''' index = pd.read_excel(io, sheet_name=level, index_col=0) + index.index = index.index.fillna("None") if index.shape[1] > 1: index = index.iloc[:, 0].to_frame() @@ -1159,7 +1160,6 @@ def to_excel( path=None, flows=True, coefficients=False, - units=True, scenario="baseline", include_meta=False, ): @@ -1190,10 +1190,6 @@ def to_excel( if True, in the Excel file, a sheet will be created named coefficients containing the data of the coefficients - units : boolean - if True, in the Excel file, a sheet will be created named units containing - the data of the units - scenario : str defines the scenario to print out the data @@ -1209,17 +1205,23 @@ def to_excel( ) ) + if flows==False and coefficients==False: + raise WrongInput("At least one of the flows or coefficients should be True") + database_excel( self, flows, coefficients, self._getdir(path, "Database", "New_Database.xlsx"), - units, scenario, ) if include_meta: meta = self.meta._to_dict() - with open(self._getdir(path, "Database", "") + "/metadata.json", "w") as fp: + meta_path = self._getdir(path, "Database", "") + meta_path = meta_path.split("/")[:-1] + meta_path = ('/').join(meta_path) + "/metadata.json" + + with open(meta_path, "w") as fp: json.dump(meta, fp) def to_txt( @@ -1227,7 +1229,6 @@ def to_txt( path=None, flows=True, coefficients=False, - units=True, scenario="baseline", _format="txt", include_meta=False, @@ -1270,6 +1271,7 @@ def to_txt( sep : str txt file separator """ + if scenario not in self.scenarios: raise WrongInput( "{} is not a valid scenario. Existing scenarios are {}".format( @@ -1277,12 +1279,14 @@ def to_txt( ) ) + if flows==False and coefficients==False: + raise WrongInput("At least one of the flows or coefficients should be True") + database_txt( self, flows, coefficients, self._getdir(path, "Database", ""), - units, scenario, _format, sep, diff --git a/mario/core/CoreIO.py b/mario/core/CoreIO.py index 9351aef..92931a1 100644 --- a/mario/core/CoreIO.py +++ b/mario/core/CoreIO.py @@ -135,6 +135,10 @@ def __init__( for item in ["matrices", "units", "_indeces"]: setattr(self, item, kwargs["init_by_parsers"][item]) + nan_keys = [key for key, df in renamed_matrices.items() if isinstance(df, pd.DataFrame) and df.isna().any().any()] + if nan_keys: + raise ValueError(f"NaN values found in the following matrices: {nan_keys}") + log_time(logger, "Metadata: initialized.") self.meta._add_attribute(table=table, price=price, source=source, year=year) diff --git a/mario/tools/constants.py b/mario/tools/constants.py index bd6e148..2b58ca8 100644 --- a/mario/tools/constants.py +++ b/mario/tools/constants.py @@ -11,6 +11,12 @@ SUT = "SUT" IOT = "IOT" +FLOWS = "flows" +COEFFICIENTS = "coefficients" + +MONETARY = "Monetary" +HYBRID = "Hybrid" + # represents different levels of aggregation _LEVELS = { SUT: {_MASTER_INDEX[i]: i for i in ["a", "c", "f", "k", "n", "r"]}, @@ -23,6 +29,8 @@ _ACCEPTABLES = { "table": [SUT, IOT], + "mode": [FLOWS, COEFFICIENTS], + 'unit': [MONETARY, HYBRID], } diff --git a/mario/tools/excelhandler.py b/mario/tools/excelhandler.py index 4e8d479..9a5109e 100644 --- a/mario/tools/excelhandler.py +++ b/mario/tools/excelhandler.py @@ -342,7 +342,7 @@ def wrirte_matrices(sheet, Z, V, E, Y, EY, flow_format, header_format): sheet.write(row, col, 0, flow_format) -def database_excel(instance, flows, coefficients, directory, units, scenario): +def database_excel(instance, flows, coefficients, directory, scenario): file = directory workbook = xlsxwriter.Workbook(file) @@ -398,40 +398,39 @@ def database_excel(instance, flows, coefficients, directory, units, scenario): wrirte_matrices(coefficients, Z, V, E, Y, EY, coeff_format, header_format) - if units: - units = workbook.add_worksheet("units") - - data = instance.units - units.write("C1", "unit", header_format) - - counter = 2 - - if instance.table_type == "SUT": - keys = [ - _MASTER_INDEX["a"], - _MASTER_INDEX["c"], - _MASTER_INDEX["f"], - _MASTER_INDEX["k"], - ] - else: - keys = [_MASTER_INDEX["s"], _MASTER_INDEX["f"], _MASTER_INDEX["k"]] - - for key in keys: - item = data[key] - for row in range(item.shape[0]): - units.write("A{}".format(counter), key, header_format) - units.write("B{}".format(counter), item.index[row], header_format) - try: - units.write("C{}".format(counter), item.iloc[row, 0]) - except TypeError: - units.write("C{}".format(counter), "None") + units = workbook.add_worksheet("units") - counter += 1 + data = instance.units + units.write("C1", "unit", header_format) + + counter = 2 + + if instance.table_type == "SUT": + keys = [ + _MASTER_INDEX["a"], + _MASTER_INDEX["c"], + _MASTER_INDEX["f"], + _MASTER_INDEX["k"], + ] + else: + keys = [_MASTER_INDEX["s"], _MASTER_INDEX["f"], _MASTER_INDEX["k"]] + + for key in keys: + item = data[key] + for row in range(item.shape[0]): + units.write("A{}".format(counter), key, header_format) + units.write("B{}".format(counter), item.index[row], header_format) + try: + units.write("C{}".format(counter), item.iloc[row, 0]) + except TypeError: + units.write("C{}".format(counter), "None") + + counter += 1 workbook.close() -def database_txt(instance, flows, coefficients, path, units, scenario, _format, sep): +def database_txt(instance, flows, coefficients, path, scenario, _format, sep): if flows: flows = instance.query( matrices=[_ENUM.V, _ENUM.E, _ENUM.Z, _ENUM.Y, _ENUM.X, _ENUM.EY], @@ -474,38 +473,40 @@ def database_txt(instance, flows, coefficients, path, units, scenario, _format, mode="a", ) - if units: - units = copy.deepcopy(instance.units) - _units = pd.DataFrame() - _index = [] - - if instance.table_type == "SUT": - keys = [ - _MASTER_INDEX["a"], - _MASTER_INDEX["c"], - _MASTER_INDEX["f"], - _MASTER_INDEX["k"], - ] - else: - keys = [_MASTER_INDEX["s"], _MASTER_INDEX["f"], _MASTER_INDEX["k"]] - - for key in keys: - value = units[key] - _index += [key] * value.shape[0] - _units = pd.concat([_units, value]) - - _units.index = [_index, _units.index] - - if (coefficients) and (not flows): - unit_dir = "coefficients" - else: - unit_dir = "flows" + units = copy.deepcopy(instance.units) + _units = pd.DataFrame() + _index = [] + + if instance.table_type == "SUT": + keys = [ + _MASTER_INDEX["a"], + _MASTER_INDEX["c"], + _MASTER_INDEX["f"], + _MASTER_INDEX["k"], + ] + else: + keys = [_MASTER_INDEX["s"], _MASTER_INDEX["f"], _MASTER_INDEX["k"]] + for key in keys: + value = units[key] + _index += [key] * value.shape[0] + _units = pd.concat([_units, value]) + + _units.index = [_index, _units.index] + + unit_dirs = [] + if coefficients: + unit_dirs += ["coefficients"] + if flows: + unit_dirs += ["flows"] + + for unit_dir in unit_dirs: if not os.path.exists(r"{}/{}".format(path, unit_dir)): os.mkdir(r"{}/{}".format(path, unit_dir)) if os.path.exists(r"{}/{}/units.{}".format(path, unit_dir, _format)): os.remove(r"{}/{}/units.{}".format(path, unit_dir, _format)) + _units.to_csv( r"{}/{}/units.{}".format(path, unit_dir, _format), header=True, diff --git a/mario/tools/parsersclass.py b/mario/tools/parsersclass.py index 13a8fc0..c46d93d 100644 --- a/mario/tools/parsersclass.py +++ b/mario/tools/parsersclass.py @@ -4,34 +4,35 @@ """ from mario import Database +import pymrio from mario.tools.tableparser import ( eora_single_region, - txt_praser, + txt_parser, excel_parser, exio3, monetary_sut_exiobase, eora_multi_region, - eurostat_sut, parse_pymrio, hybrid_sut_exiobase_reader, parser_figaro_sut, ) from mario.log_exc.exceptions import WrongInput, LackOfInput +from mario.tools.constants import _ACCEPTABLES, _HMRSUT_EXTENSIONS +import pandas as pd models = {"Database": Database} - def parse_from_txt( - path, - table, - mode, - calc_all=False, - year=None, - name=None, - source=None, - model="Database", - sep=",", + path: str, + table: str, + mode: str, + calc_all: bool = False, + year: int = None, + name:str = None, + source:str =None, + model: str ="Database", + sep: str = ",", **kwargs, ): """Parsing database from text files @@ -74,10 +75,17 @@ def parse_from_txt( ------- mario.Database """ - if model not in models: - raise WrongInput("Available models are {}".format([*models])) - matrices, indeces, units = txt_praser(path, table, mode, sep) + # check key inputs to be correct + errmsg = [] + if table not in _ACCEPTABLES['table']: + errmsg.append(f"Table should be in {_ACCEPTABLES['table']}") + if mode not in _ACCEPTABLES['mode']: + errmsg.append(f"Mode should be in {_ACCEPTABLES['mode']}") + if errmsg: + raise WrongInput(errmsg) + + matrices, indeces, units = txt_parser(path, table, mode, sep) return models[model]( name=name, @@ -91,16 +99,16 @@ def parse_from_txt( def parse_from_excel( - path, - table, - mode, - data_sheet=0, - unit_sheet="units", - calc_all=False, - year=None, - name=None, - source=None, - model="Database", + path: str, + table: str, + mode: str, + data_sheet: str = 0, + unit_sheet: str = "units", + calc_all: bool = False, + year: int = None, + name: str = None, + source: str = None, + model: str ="Database", **kwargs, ): """Parsing database from excel file @@ -147,9 +155,15 @@ def parse_from_excel( mario.Database """ - if model not in models: - raise WrongInput("Available models are {}".format([*models])) - + # check key inputs to be correct + errmsg = [] + if table not in _ACCEPTABLES['table']: + errmsg.append(f"Table should be in {_ACCEPTABLES['table']}") + if mode not in _ACCEPTABLES['mode']: + errmsg.append(f"Mode should be in {_ACCEPTABLES['mode']}") + if errmsg: + raise WrongInput(errmsg) + matrices, indeces, units = excel_parser(path, table, mode, data_sheet, unit_sheet) return models[model]( @@ -164,18 +178,14 @@ def parse_from_excel( def parse_exiobase_sut( - path, - calc_all=False, - name=None, - year=None, - model="Database", + path: str, + calc_all: bool =False, + name: str = None, + year: int = None, + model: str = "Database", **kwargs, ): - """Parsing exiobase mrsut - - .. note:: - - mario v.0.1.0, supports only Monetary Exiobase MRSUT database. + """Parsing Multi-Regional Supply and Use Table from Exiobase Parameters ---------- @@ -215,15 +225,15 @@ def parse_exiobase_sut( def parse_exiobase_3( - path, - calc_all=False, - year=None, - name=None, - model="Database", - version="3.8.2", + path: str, + calc_all: bool =False, + year: int = None, + name: str = None, + model: str = "Database", + version: str = "3.8.2", **kwargs, ): - """Parsing exiobase3 + """Parsing Multi-Regional Input-Output Table from Exiobase .. note:: @@ -255,11 +265,14 @@ def parse_exiobase_3( mario.Database """ - if model not in models: - raise WrongInput("Available models are {}".format([*models])) + # check the inputs to be correct + errmsg = [] if version not in ["3.8.2", "3.8.1"]: - raise WrongInput("Acceptable versions are {}".format(["3.8.2", "3.8.1"])) + errmsg.append("Acceptable versions are {}".format(["3.8.2", "3.8.1"])) + if errmsg: + raise WrongInput(errmsg) + matrices, indeces, units = exio3(path, version) return models[model]( @@ -274,19 +287,20 @@ def parse_exiobase_3( def parse_eora( - path, - multi_region, - table, - indeces=None, - name_convention="full_name", - aggregate_trade=True, - year=None, - name=None, - calc_all=False, - model="Database", + path: str, + multi_region: bool, + table: str, + indeces: str = None, + name_convention: str = "full_name", + aggregate_trade: bool = True, + year: int = None, + name: str = None, + calc_all: bool = False, + model: str = "Database", **kwargs, ) -> object: - """Parsing eora databases + + """Parsing EORA databases .. note:: @@ -327,6 +341,7 @@ def parse_eora( ------- mario.Database """ + if model not in models: raise WrongInput("Available models are {}".format([*models])) @@ -371,9 +386,17 @@ def parse_eora( def parse_exiobase( - table, unit, path, model="Database", name=None, year=None, calc_all=False, **kwargs + table:str, + unit:str, + path:str, + model:str = "Database", + name:str = None, + year:int = None, + calc_all: bool = False, + **kwargs ): - """A unique function for parsing all exiobase databases + + """A unique function for parsing all Exiobase databases Parameters ---------- @@ -403,23 +426,20 @@ def parse_exiobase( if non-valid values are passed to the arguments. """ - if table not in ["IOT", "SUT"]: - raise WrongInput("table only accpets 'IOT' or 'SUT'.") + if table not in _ACCEPTABLES['table']: + raise WrongInput("Table can be only chosen among {}".format(_ACCEPTABLES['table'])) - if unit not in ["Hybrid", "Monetary"]: - raise WrongInput("unit only accpets 'Hybrid' or 'Monetary.'") + if unit not in _ACCEPTABLES['unit']: + raise WrongInput("Unit con be only chosen among {}".format(_ACCEPTABLES['unit'])) if table == "IOT": if unit == "Monetary": parser = parse_exiobase_3 - else: raise WrongInput("Hybrid IOT exiobase is not supported by mario.") - else: if unit == "Monetary": parser = parse_exiobase_sut - else: parser = hybrid_sut_exiobase @@ -433,9 +453,15 @@ def parse_exiobase( def hybrid_sut_exiobase( - path, extensions=[], model="Database", name=None, calc_all=False, **kwargs + path:str, + extensions: list = [], + model: str = "Database", + name: str = None, + calc_all: bool = False, + **kwargs ): - """reads hybrid supply and use exiobase + """ + Parser for hybrid units Exiobase database (v3.3.18) Parameters ---------- @@ -464,8 +490,15 @@ def hybrid_sut_exiobase( For more informatio refer to https://zenodo.org/record/7244919#.Y6hEfi8w2L1 """ - if model not in models: - raise WrongInput("Available models are {}".format([*models])) + + # check the inputs to be correct + errmsg = [] + if extensions != 'all': + if extensions != None: + if any([ext not in _HMRSUT_EXTENSIONS for ext in extensions]): + errmsg.append("Extensions should be chosen among {}".format(_HMRSUT_EXTENSIONS)) + if errmsg: + raise WrongInput(errmsg) matrices, indeces, units = hybrid_sut_exiobase_reader( path=path, @@ -492,11 +525,11 @@ def hybrid_sut_exiobase( def parse_eurostat_sut( - supply_path, - use_path, - model="Database", - name=None, - calc_all=False, + supply_path:str, + use_path:str, + model:str="Database", + name:str=None, + calc_all:bool=False, **kwargs, ) -> object: """Parsing Eurostat databases @@ -509,7 +542,6 @@ def parse_eurostat_sut( * third rule: use only "total" as stock/flow parameter, and only one unit of measure * forth rule: supply must be provided in activity by commodity, use must be provided in commodity by activitiy formats - Parameters ---------- supply_path : str @@ -527,26 +559,15 @@ def parse_eurostat_sut( mario.Database """ - if model not in models: - raise WrongInput("Available models are {}".format([*models])) - - matrices, indeces, units, meta = eurostat_sut( - supply_path, - use_path, - ) - - return models[model]( - name=name, - table="SUT", - source="eurostat", - year=meta["year"], - init_by_parsers={"matrices": matrices, "_indeces": indeces, "units": units}, - calc_all=calc_all, - **kwargs, - ) + raise NotImplemented("This function was deprecated since the parser was too dependent on Eurostat web interface. Downgrade to mariopy==v.3.3.3 in case you need it") -def parse_from_pymrio(io, value_added, satellite_account, include_meta=True): +def parse_from_pymrio( + io, + value_added, + satellite_account, + include_meta=True + ): """Parsing a pymrio database Parameters @@ -585,12 +606,18 @@ def parse_from_pymrio(io, value_added, satellite_account, include_meta=True): ) -def parse_FIGARO_SUT(directory, name=None, calc_all=False, **kwargs): - """reads a FIGARO SUT table +def parse_FIGARO_SUT( + path:str, + name:str = None, + calc_all:bool = False, + **kwargs + ): + + """Download and parse a FIGARO SUT table Parameters ---------- - directory : str + path : str the folder where the files are downloaded name : str, optional a name for the database, by default None @@ -603,7 +630,7 @@ def parse_FIGARO_SUT(directory, name=None, calc_all=False, **kwargs): mario database object """ - matrices, indeces, units, year = parser_figaro_sut(directory) + matrices, indeces, units, year = parser_figaro_sut(path) return models["Database"]( name=name, @@ -614,3 +641,5 @@ def parse_FIGARO_SUT(directory, name=None, calc_all=False, **kwargs): calc_all=calc_all, **kwargs, ) + + diff --git a/mario/tools/tableparser.py b/mario/tools/tableparser.py index 61659b3..ca2e29c 100644 --- a/mario/tools/tableparser.py +++ b/mario/tools/tableparser.py @@ -305,8 +305,46 @@ def get_units(units, table, indeces): # if everything is ok, build a dataframe from the units. return _ +def replace_nan_indices(matrix): + for axis in [0,1]: + if axis == 0: + n_levels = matrix.index.nlevels + else: + n_levels = matrix.columns.nlevels + new_levels = [[] for i in range(n_levels)] + + for level in range(n_levels): + if axis == 0: + new_levels[level] = ["None" if pd.isna(x) else x for x in matrix.index.get_level_values(level)] + else: + new_levels[level] = ["None" if pd.isna(x) else x for x in matrix.columns.get_level_values(level)] -def txt_praser(path, table, mode, sep): + if axis == 0: + if n_levels == 1: + matrix.index = pd.Index(new_levels[0]) + else: + matrix.index = pd.MultiIndex.from_arrays(new_levels) + else: + if n_levels == 1: + matrix.columns = pd.Index(new_levels[0]) + else: + matrix.columns = pd.MultiIndex.from_arrays(new_levels) + + return matrix + +def replace_nan_units_indices(units): + + new_levels = [[],[]] + for level in range(len(units.index.names)): + for i in units.index.get_level_values(level): + if pd.isna(i) or i == '-': + i = "None" + new_levels[level].append(i) + units.index = pd.MultiIndex.from_arrays(new_levels) + units.fillna("None", inplace=True) + return units + +def txt_parser(path, table, mode, sep): if mode == "coefficients": v, e, z = list("vez") else: @@ -323,9 +361,13 @@ def txt_praser(path, table, mode, sep): log_time(logger, "Parser: Reading files finished.") _units = read["units"]["all"] + _units = replace_nan_units_indices(_units) log_time(logger, "Parser: Investigating possible identifiable errors.") + for matrix in read["matrices"].keys(): + read["matrices"][matrix] = replace_nan_indices(read["matrices"][matrix]) + indeces = get_index_txt( Z=read["matrices"][z], V=read["matrices"][v], @@ -364,6 +406,7 @@ def txt_praser(path, table, mode, sep): matrices = {"baseline": {**read["matrices"]}} + return matrices, indeces, units @@ -380,6 +423,8 @@ def excel_parser(path, table, mode, sheet_name, unit_sheet): data = pd.read_excel( path, header=[0, 1, 2], index_col=[0, 1, 2], sheet_name=sheet_name ) + + data = replace_nan_indices(data) indeces = get_index_excel(data, table, mode) if table == "SUT": @@ -424,9 +469,9 @@ def excel_parser(path, table, mode, sheet_name, unit_sheet): (slice(None), _MASTER_INDEX["k"]), (slice(None), _MASTER_INDEX["n"]) ] - V.index = V.index.get_level_values(-1) - E.index = E.index.get_level_values(-1) - EY.index = EY.index.get_level_values(-1) + V.index = indeces['f']['main'] + E.index = indeces['k']['main'] + EY.index = indeces['k']['main'] if mode == "coefficients": matrices = { @@ -452,9 +497,13 @@ def excel_parser(path, table, mode, sheet_name, unit_sheet): } # read the unit sheet from the excel file + _units = pd.read_excel(path, sheet_name=unit_sheet, index_col=[0, 1]) + _units = replace_nan_units_indices(_units) + units = get_units( - pd.read_excel(path, sheet_name=unit_sheet, index_col=[0, 1]), table, indeces + _units, table, indeces ) + rename_index(matrices["baseline"]) sort_frames(matrices["baseline"]) @@ -706,6 +755,15 @@ def eora_single_region(path, table, name_convention="full_name", aggregate_trade elif table == "SUT": Z_index = eora[_MASTER_INDEX["a"]] + eora[_MASTER_INDEX["c"]] + if not all(item in data.index.get_level_values(0) for item in Z_index): + raise WrongFormat( + f"The parsed table does not seem a {table}. Please check the table type." + ) + if 'Commodities' in data.index.get_level_values(0) and table=='IOT': + raise WrongFormat( + f"The parsed table does not seem a {table}. Please check the table type." + ) + Z = data.loc[Z_index, Z_index] Y = data.loc[Z_index, eora[_MASTER_INDEX["n"]]] V = data.loc[eora[_MASTER_INDEX["f"]], Z_index] @@ -1003,159 +1061,6 @@ def eora_multi_region(data_path, index_path, year, price): return matrices, indeces, units -def eurostat_sut( - supply_path, - use_path, -): - supply_file = pd.ExcelFile(supply_path) - use_file = pd.ExcelFile(use_path) - - supply_meta = extract_metadata_from_eurostat(supply_file) - use_meta = extract_metadata_from_eurostat(use_file) - - if "Supply table at basic prices" not in supply_meta["table"]: - raise WrongInput( - "specified supply table dataset is {}. Acceptable dataset is 'Supply table at basic prices'." - "Please refer to the documents for proper download of the dataset".format( - supply_meta["table"] - ) - ) - - if "Use table at basic prices " not in use_meta["table"]: - raise WrongInput( - "specified use table dataset is {}. Acceptable dataset is 'Use table at basic prices '." - "Please refer to the documents for proper download of the dataset".format( - use_meta["table"] - ) - ) - - if ( - supply_meta["country"] != use_meta["country"] - or supply_meta["year"] != use_meta["year"] - ): - raise WrongInput( - "there are mismatched between the country/year of supply and use datasets.\nSupply Dataset: {}\nUse Dataset:{}".format( - supply_meta, use_meta - ) - ) - - supply_data = supply_file.parse( - sheet_name=eurostat_id["supply"]["sheet_name"], - index_col=eurostat_id["supply"]["index_col"], - header=eurostat_id["supply"]["header"], - ) - - use_data = use_file.parse( - sheet_name=eurostat_id["use"]["sheet_name"], - index_col=eurostat_id["use"]["index_col"], - header=eurostat_id["use"]["header"], - ) - - # build Z_matrix - z_index_c = pd.MultiIndex.from_product( - [[supply_meta["country"]], [_MASTER_INDEX["c"]], eurostat_id["c"]] - ) - - z_index_a = pd.MultiIndex.from_product( - [[supply_meta["country"]], [_MASTER_INDEX["a"]], eurostat_id["a"]] - ) - - z_index = z_index_c.append(z_index_a) - - Z = pd.DataFrame(data=0, index=z_index, columns=z_index) - - # fill supply side - Z.loc[ - (supply_meta["country"], _MASTER_INDEX["a"], eurostat_id["a"]), - (supply_meta["country"], _MASTER_INDEX["c"], eurostat_id["c"]), - ] = supply_data.loc[eurostat_id["a"], eurostat_id["c"]].values - - # fill use side - Z.loc[ - (supply_meta["country"], _MASTER_INDEX["c"], eurostat_id["c"]), - (supply_meta["country"], _MASTER_INDEX["a"], eurostat_id["a"]), - ] = use_data.loc[eurostat_id["c"], eurostat_id["a"]].values - - # build V_matrix - V = pd.DataFrame( - data=0, index=eurostat_id["f"] + eurostat_id["c_import"], columns=z_index - ) - - # Activity VA - V.loc[ - eurostat_id["f"], (supply_meta["country"], _MASTER_INDEX["a"], eurostat_id["a"]) - ] = use_data.loc[eurostat_id["f"], eurostat_id["a"]].values - - # Commodity VA - V.loc[ - eurostat_id["c_import"], - (supply_meta["country"], _MASTER_INDEX["c"], eurostat_id["c"]), - ] = supply_data.loc[eurostat_id["c_import"], eurostat_id["c"]].values - - # Building Y matrix - Y_columns = pd.MultiIndex.from_product( - [[supply_meta["country"]], [_MASTER_INDEX["n"]], eurostat_id["n"]] - ) - Y = pd.DataFrame(data=0, index=z_index, columns=Y_columns) - - Y.loc[ - (supply_meta["country"], _MASTER_INDEX["c"], eurostat_id["c"]), - (supply_meta["country"], _MASTER_INDEX["n"], eurostat_id["n"]), - ] = use_data.loc[eurostat_id["c"], eurostat_id["n"]].values - - # Building E and EY - E = pd.DataFrame(data=0, index=["-"], columns=z_index) - EY = pd.DataFrame(data=0, index=["-"], columns=Y_columns) - - # Units - units = { - _MASTER_INDEX["a"]: pd.DataFrame( - use_meta["unit"], index=eurostat_id["a"], columns=["unit"] - ), - _MASTER_INDEX["c"]: pd.DataFrame( - use_meta["unit"], index=eurostat_id["c"], columns=["unit"] - ), - _MASTER_INDEX["f"]: pd.DataFrame( - use_meta["unit"], - index=eurostat_id["c_import"] + eurostat_id["f"], - columns=["unit"], - ), - _MASTER_INDEX["k"]: pd.DataFrame( - "None", - index=E.index, - columns=["unit"], - ), - } - - X = calc_X(Z, Y) - - for matrix in [Z, V, E, EY, Y, X]: - matrix.replace(":", 0, inplace=True) - - matrices = { - "baseline": { - "Z": Z, - "V": V, - "E": E, - "EY": EY, - "Y": Y, - "X": X, - } - } - sort_frames(matrices["baseline"]) - indeces = { - "r": {"main": Z.index.unique(0).tolist()}, - "n": {"main": Y.columns.unique(-1).tolist()}, - "k": {"main": E.index.tolist()}, - "f": {"main": V.index.tolist()}, - "a": {"main": eurostat_id["a"]}, - "c": {"main": eurostat_id["c"]}, - } - rename_index(matrices["baseline"]) - - return matrices, indeces, units, use_meta - - def parse_pymrio(io, value_added, satellite_account): """Extracts the data from pymrio in mario format""" diff --git a/mario/version.py b/mario/version.py index e19434e..334b899 100644 --- a/mario/version.py +++ b/mario/version.py @@ -1,1 +1,1 @@ -__version__ = "0.3.3" +__version__ = "0.3.4" diff --git a/requirements.txt b/requirements.txt index 68fd627..f8fd3dd 100644 --- a/requirements.txt +++ b/requirements.txt @@ -1,10 +1,10 @@ -numpy >= 1.21.2 -plotly >= 4.12.0 -tabulate >= 0.8.9 -openpyxl >= 3.0.6 -IPython >= 7.22.0 +numpy == 2.1.1 +plotly +tabulate +openpyxl == 3.1.0 +IPython >= 8.27.0 pymrio -pytest -pandas +pytest +pandas == 2.2.3 pyyaml -xlsxwriter \ No newline at end of file +xlsxwriter == 3.2.0 \ No newline at end of file diff --git a/setup.py b/setup.py index d821594..0372c30 100644 --- a/setup.py +++ b/setup.py @@ -30,15 +30,15 @@ "mario/test":["*.xlsx"], }, install_requires=[ - "pandas", - "numpy >= 1.21.2", - "xlsxwriter", - "plotly >= 4.12.0", - "tabulate >= 0.8.9", - "openpyxl >= 3.0.6", - "IPython >= 7.22.0", - "pymrio >= 0.4.6", - "pyyaml >= 5.4.1" + "pandas == 2.2.3", + "numpy == 2.1.1", + "xlsxwriter == 3.2.0", + "plotly", + "tabulate", + "openpyxl == 3.1.0", + "IPython >= 8.27.0", + "pymrio", + "pyyaml" ], # classifiers=[

Fixed bugs related to Nans in indices or units New pandas versions interpret None as NaN. This raised issues in parsing databases with labels or units named "None". A check on NaNs was implemented when parsing, replacing NaNs with "None", as MARIO used to deal with since v0.2.2. "-" items where also replaced with "None" in test files (e.g. those loaded with mario.load_test), to align with the approach.

2024-10-29T06:06:57

0.0

[]

it-is-me-mario/MARIO

it-is-me-mario__MARIO-45

66eafdfc1ae8e6fb470c6015895fdbaab0426a3b

diff --git a/mario/core/AttrData.py b/mario/core/AttrData.py index db9ae34..4603b8e 100644 --- a/mario/core/AttrData.py +++ b/mario/core/AttrData.py @@ -1506,6 +1506,91 @@ def add_sectors( for note in notes: self.meta._add_history(f"User note: {note}") + def query( + self, + matrices, + scenarios = ["baseline"], + base_scenario = None, + type="absolute", + ): + """ Requests a specific data from the database + + Parameters + ---------- + matrices : str + list of the matrices to return + + scenarios : str, List[str] + list of scenarios for returing the matrices + + base_scenario : str + str representing the base scenario in case that the data should be returned for the change in data between scenarios + + type: str + #. 'absolute' for absolute difference for scenarios + #. 'relative' for relative difference for scenarios + + + Returns + ------- + dict,pd.DataFrame + If multiple scenarios are passed, it returns a dict where keys are the scenarios and vals are the matrices. matrices itself could be a dict or pd.DataFrame depending if multiple matrices or one matrix is passed + + + Example + ------- + Let's consider the case that multiple scenarios ['sc.1','sc.2'] and multiple matrices ['X','z'] are passed: + + .. code-block:: python + + output = example.query(scenarios= ['sc.1','sc.2'], matrices = ['X','z']) + + the output in this case would be: + + type: Dict[Dict[pd.DataFrame]] + + { + "sc.1" : { + "X": pd.DataFrame, + "z": pd.DataFrame, + }, + "sc.2" : { + "X": pd.DataFrame, + "z": pd.DataFrame, + }, + + } + + if only one scenario ("sc.1") is passed, output will be: + + type Dict[pd.DataFrame] + { + "X": pd.DataFrame, + "Y": pd.DataFrame, + } + + if only one scenario ("sc.1") and one matrix ("X") is passed the output will be a single pd.DataFrame. + """ + data = self.get_data( + matrices=matrices, + units=False, + indeces=False, + format="dict", + scenarios=scenarios, + base_scenario = base_scenario, + type = type + ) + + + if len(matrices) == 1: + for scenario in scenarios: + data[scenario]= data[scenario][matrices[0]] + + if len(scenarios) == 1: + data = data[scenarios[0]] + + return data + def get_data( self, matrices, @@ -1523,7 +1608,7 @@ def get_data( Parameters ---------- - path : str + matrices : str list of the matrices to return units : boolean

Get_data function The function "get_data" is not working if the user provide an integer as "scenarios" parameter. Not even if it is provided within a list. Moreover, the documentation for this function seems missing.

I would add the following: The get_data function was supposed to be used as a back-hand. Probabily for this reason, it looks sub-optimal and may be improved. Best way to use it to extract a specific matrix (e.g. f) of the main object (world) is: world.get_data(matrices=['f'], scenarios=['scenario 2'])['scenario 2'][0] The function may be improved by avoiding the pass the same info twice. Furthermore, a dedicated documentation should be added. > I would add the following: The get_data function was supposed to be used as a back-hand. Probabily for this reason, it looks sub-optimal and may be improved. Best way to use it to extract a specific matrix (e.g. f) of the main object (world) is: > > world.get_data(matrices=['f'], scenarios=['scenario 2'])['scenario 2'][0] > > The function may be improved by avoiding the pass the same info twice. Furthermore, dedicated documentation should be added. @nigolred & @LorenzoRinaldi Probably the best way to solve the problem would be to have a new function with a more user-friendy function that executes the get_data function in a simpler and more intuitive way and leave the get_data function as a private function that is extensively used in the backend. Probably changing this function itself will arise **backward-compatibility** issues. What do you think? For me that can work for sure and may be preferable to avoid backward-compatibility.

2023-04-13T21:46:24

0.0

[]

it-is-me-mario/MARIO

it-is-me-mario__MARIO-28

767ac770cbf78180e3b607247b0ce9f617393989

diff --git a/mario/core/AttrData.py b/mario/core/AttrData.py index 0d405f9..b415a39 100644 --- a/mario/core/AttrData.py +++ b/mario/core/AttrData.py @@ -19,10 +19,7 @@ from mario.tools.tabletransform import SUT_to_IOT import json from mario.tools.utilities import ( - _matrices, _manage_indeces, - _meta_parse_history, - linkages_calculation, check_clusters, run_from_jupyter, filtering, @@ -63,6 +60,7 @@ calc_f, calc_f_dis, calc_X_from_z, + linkages_calculation, ) from mario.tools.sectoradd import adding_new_sector @@ -298,17 +296,14 @@ def to_iot( matrices, indeces, units = SUT_to_IOT(self, method) for scenario in self.scenarios: - _matrices(self, "del", scenario) log_time(logger, f"{scenario} deleted from the database", "warn") self.meta._add_history(f"{scenario} deleted from the database") self.matrices = matrices - _matrices(self, "add", "baseline") self._indeces = indeces self.units = units - _meta_parse_history(self, "parse") self.meta.table = "IOT" self.meta._add_history( @@ -418,20 +413,14 @@ def read_aggregated_index( if index.shape[1] > 1: index = index.iloc[:, 0].to_frame() - if set(index.index.to_list()) != set(self.get_index(level)): - missing_items = [] + difference = set(index.index).difference(set(self.get_index(level))) - for item in set(self.get_index(level)): - if item not in index.index.to_list(): - missing_items.append(item) - - raise WrongExcelFormat( - "Disaggregated indeces of level '{}' in the Excel file " - "does not match with the original indices of the database.\nMissing items are:\n{}".format( - level, missing_items - ) + if difference: + raise WrongInput( + f"Following item are not acceptable for level {level} \n {difference}" ) + index.columns = ["Aggregation"] if index.isnull().values.any(): @@ -552,9 +541,7 @@ def aggregate( __new_matrices, units = _aggregator(self, drop) for scenario in self.scenarios: - _matrices(self, "del", scenario) self.matrices[scenario] = __new_matrices[scenario] - _matrices(self, "add", scenario) self.meta._add_history( "original matrices changed to the aggregated level based on the inputs from {}".format( @@ -569,7 +556,6 @@ def aggregate( _manage_indeces(self, "aggregation") new_index = copy.deepcopy(self._indeces) - _meta_parse_history(self, "aggregation", old_index, new_index) if calc_all: for scenario in self.scenarios: @@ -581,9 +567,7 @@ def reset_to_backup(self): All the matrices and indeces will be updated to the last back-up """ - _matrices(self, function="del") self.matrices = self._backup.matrices - _matrices(self, function="add") self._indeces = self._backup.indeces self.units = self._backup.units self.meta._add_history("Last backup recovered.") @@ -685,7 +669,7 @@ def add_extensions( if not inplace: new = self.copy() new.add_extensions( - io=io, matrix=matrix, backup=backup, inplace=True, units=units, + io=io, matrix=matrix, backup=backup, inplace=True, units=units,calc_all=calc_all,notes=notes,EY=EY ) return new @@ -740,7 +724,7 @@ def add_extensions( "units dataframe should has exactly the same index levels of io" ) - if EY is not None: + if EY is not None and matrix_id == "k": EY = EY.sort_index() if not data.index.equals(EY.index): @@ -776,16 +760,14 @@ def add_extensions( units = info["units"] del info["units"] - info["X"] = calc_X(Z=info["Z"], Y=info["Y"]) + matrices = {"baseline": {**info}} for scenario in self.scenarios: - _matrices(self, "del", scenario) log_time(logger, f"{scenario} deleted from the database", "warn") self.meta._add_history(f"{scenario} deleted from the database") self.matrices = matrices - _matrices(self, "add", "baseline") self.meta._add_history( f"Modification: new '{_MASTER_INDEX[matrix_id]}' added to the database as follow:\n" @@ -946,14 +928,12 @@ def to_single_region(self, region, backup=True, inplace=True): _manage_indeces(self, "single_region", **new_indeces) for scenario in self.scenarios: - _matrices(self, "del", scenario=scenario) log_time(logger, f"Transformation: {scenario} deleted from the database.") self.matrices = {"baseline": {}} for matrix in ["Y", "Z", "E", "EY", "Y", "V", "X"]: self.matrices["baseline"][matrix] = eval(matrix) log_time(logger, "Transformation: New baseline added to the database") - _matrices(self, "add", scenario="baseline") slicer = _MASTER_INDEX["a"] if self.table_type == "SUT" else _MASTER_INDEX["s"] @@ -977,16 +957,7 @@ def to_single_region(self, region, backup=True, inplace=True): "Transformation: The Final Demand emissions are considered only for 'Local Final Demand.'" ) - def backup(self): - """The function creates a backup of the last configuration of database - to be returned in case needed. - """ - self._backup = self._backup_( - copy.deepcopy(self.matrices), - copy.deepcopy(self._indeces), - copy.deepcopy(self.units), - ) def calc_linkages( self, scenario="baseline", normalized=True, cut_diag=True, multi_mode=True, @@ -1048,7 +1019,6 @@ def calc_linkages( } return linkages_calculation( - self, cut_diag=cut_diag, matrices=_matrices, multi_mode=multi_mode, @@ -1515,13 +1485,9 @@ def add_sectors( EY = self.EY # Deleting old values - _matrices(self, "del") for matrix in ["z", "e", "v", "Y", "X", "Z", "E", "V", "EY"]: self.matrices["baseline"][matrix] = eval(matrix) - # Adding new values - _matrices(self, "add") - self.meta._add_history( "Scenarios: all the scenarios deleted from the database." ) @@ -1830,7 +1796,11 @@ def shock_calc( if scenario == "baseline": raise WrongInput("baseline scenario can not be overwritten.") - check_clusters(self, clusters) + check_clusters( + index_dict = self.get_index('all'), + table = self.table_type, + clusters = clusters + ) # have the test for the existence of the database @@ -1888,7 +1858,11 @@ def get_shock_excel( e.g. clusters = {'Region':{'cluster_1':['reg1','reg2']}} """ - check_clusters(self, clusters) + check_clusters( + index_dict = self.get_index('all'), + table = self.table_type, + clusters = clusters + ) _sh_excel(self, num_shock, self._getdir(path, "Excels", "shock.xlsx"), clusters) diff --git a/mario/core/CoreIO.py b/mario/core/CoreIO.py index 391056b..d47d5f2 100644 --- a/mario/core/CoreIO.py +++ b/mario/core/CoreIO.py @@ -15,11 +15,6 @@ from mario.tools.tableparser import dataframe_parser -from mario.tools.utilities import ( - _matrices, - _meta_parse_history, -) - from mario.tools.iomath import ( calc_X, @@ -143,8 +138,6 @@ def __init__( log_time(logger, "Metadata: initialized.") self.meta._add_attribute(table=table, price=price, source=source, year=year) - _matrices(self, function="add") - else: if not all( [matrix is not None for matrix in [Y, E, Z, V, EY, units, table]] @@ -160,15 +153,12 @@ def __init__( log_time(logger, "Metadata: initialized by dataframes.") - _matrices(self, function="del") - _matrices(self, function="add") # Adding notes if passed by user or the parsers if kwargs.get("notes"): for note in kwargs["notes"]: self.meta._add_history(note) - _meta_parse_history(self, "parse") if calc_all: self.calc_all() @@ -996,19 +986,16 @@ def meta_history(self): def __getitem__(self, key): """get item method retuns the data regarding the scenarios""" - self.scenario_exist(key) - - return self.matrices[key] - - def scenario_exist(self, scenario): - """Checks if a scneario exists or not""" - if scenario not in self.scenarios: + if key not in self.scenarios: raise WrongInput( "{} is not a valid scenario. Valid scenarios are {}".format( - scenario, self.scenarios + key, self.scenarios ) ) + return self.matrices[key] + + def __iter__(self): self.__it__ = self.scenarios return self @@ -1065,34 +1052,14 @@ def __eq__(self,other): return True - def _extract_index_from_frames(self,_dict): - ids = { - 'r': {'matrix':['z','X','Y','v','e','Z',],"level":('index',0)}, - 's': {'matrix':['z','X','Y','v','e','Z',],"level":('index',-1)}, - 'a': {'matrix':['z','X','Y','v','e','Z',],"level":('index',-1)}, - 'c': {'matrix':['z','X','Y','v','e','Z',],"level":('index',-1)}, - 'n': {'matrix':['Y','EY'],"level":('columns',1)}, - 'k': {'matrix':['e','E','EY'],"level":('index',0)}, - 'f': {'matrix':['v','V','EY'],"level":('index',0)}, - } + + def backup(self): - indeces = {} - for key,vals in ids.items(): - idx = _MASTER_INDEX[key] - - if idx not in self.sets: - continue - - for matrix in vals['matrix']: - if matrix in _dict: - df = _dict[matrix] - print(idx,matrix) - - if key in ['s','a','c']: - df = df.loc[(slice(None),idx,slice(None)),:] - index = list(getattr(df,vals['level'][0]).unique(vals['level'][1])) - indeces[idx] = index - break - - return indeces - + """The function creates a backup of the last configuration of database + to be returned in case needed. + """ + self._backup = self._backup_( + copy.deepcopy(self.matrices), + copy.deepcopy(self._indeces), + copy.deepcopy(self.units), + ) \ No newline at end of file diff --git a/mario/core/CustomDicts.py b/mario/core/CustomDicts.py deleted file mode 100644 index 3c3547a..0000000 --- a/mario/core/CustomDicts.py +++ /dev/null @@ -1,47 +0,0 @@ - -from collections import UserDict - -class Matrices(UserDict): - """ - A costumized dict for having a better control on matrices property - in mario.Core - - Properties - ----------- - _bs_name : str - This represents the name of the baseline scenario - _dynamic : boolean - Shows if the database is dynamic or not - """ - def __init__(self,baseline_name,dynamic=False,vals={}): - self._bs_name = baseline_name - self._dynamic = dynamic - - super().__init__(vals) - - def __setitem__(self, key, item): - if key == 'baseline': - key = self._bs_name - - if self._dynamic and not isinstance(key,int): - raise ValueError('For a dynamic database, only integer keys are accepted.') - - if not isinstance(item,dict): - raise ValueError('item should be a dict.') - - - super().__setitem__(key, item) - - def __getitem__(self, key): - - if key == 'baseline': - key = self._bs_name - - return super().__getitem__(key) - -if __name__ == '__main__': - test = Matrices(2020,True) - test['baseline'] = {} - #%% - -# %% diff --git a/mario/tools/iomath.py b/mario/tools/iomath.py index e5259f0..d5f678e 100644 --- a/mario/tools/iomath.py +++ b/mario/tools/iomath.py @@ -6,6 +6,11 @@ import numpy as np from copy import deepcopy as dc +from mario.log_exc.logger import log_time +import logging + +logger = logging.getLogger(__name__) + def calc_all_shock(z, e, v, Y): @@ -470,3 +475,97 @@ def X_inverse(X): X_inv[X_inv != 0] = 1 / X_inv[X_inv != 0] return X_inv + + + + +def linkages_calculation(cut_diag, matrices, multi_mode, normalized): + """calculates the linkages""" + if cut_diag: + for key, value in matrices.items(): + np.fill_diagonal(value.values, 0) + + if multi_mode: + + link_types = [ + "Total Forward", + "Total Backward", + "Direct Forward", + "Direct Backward", + ] + geo_types = ["Local", "Foreign"] + links = pd.DataFrame( + 0, + index=matrices["g"].index, + columns=pd.MultiIndex.from_product([link_types, geo_types]), + ) + + for index, values in links.iterrows(): + links.loc[index, ("Total Forward", "Local")] = ( + matrices["g"].loc[index, index[0]].sum().sum() + ) + links.loc[index, ("Total Forward", "Foreign")] = ( + matrices["g"].loc[index].sum().sum() + - matrices["g"].loc[index, index[0]].sum().sum() + ) + + links.loc[index, ("Total Backward", "Local")] = ( + matrices["w"].loc[index, index[0]].sum().sum() + ) + links.loc[index, ("Total Backward", "Foreign")] = ( + matrices["w"].loc[index].sum().sum() + - matrices["w"].loc[index, index[0]].sum().sum() + ) + + links.loc[index, ("Direct Forward", "Local")] = ( + matrices["b"].loc[index, index[0]].sum().sum() + ) + links.loc[index, ("Direct Forward", "Foreign")] = ( + matrices["b"].loc[index].sum().sum() + - matrices["b"].loc[index, index[0]].sum().sum() + ) + + links.loc[index, ("Direct Backward", "Local")] = ( + matrices["z"].loc[index, index[0]].sum().sum() + ) + links.loc[index, ("Direct Backward", "Foreign")] = ( + matrices["z"].loc[index].sum().sum() + - matrices["z"].loc[index, index[0]].sum().sum() + ) + + if normalized: + log_time( + logger, "Normalization not available for multi-regional mode.", "warn" + ) + + # Computing linkages as if there were only one unique region + else: + _forward_t = matrices["g"].sum(axis=1).to_frame() + _backward_t = matrices["w"].sum(axis=0).to_frame() + _forward_d = matrices["b"].sum(axis=1).to_frame() + _backward_d = matrices["z"].sum(axis=0).to_frame() + + _forward_t.columns = ["Total Forward"] + _backward_t.columns = ["Total Backward"] + _forward_d.columns = ["Direct Forward"] + _backward_d.columns = ["Direct Backward"] + + if normalized: + + _forward_t.iloc[:, 0] = _forward_t.iloc[:, 0] / np.average( + _forward_t.values + ) + _backward_t.iloc[:, 0] = _backward_t.iloc[:, 0] / np.average( + _backward_t.values + ) + _forward_d.iloc[:, 0] = _forward_d.iloc[:, 0] / np.average( + _forward_d.values + ) + _backward_d.iloc[:, 0] = _backward_d.iloc[:, 0] / np.average( + _backward_d.values + ) + + links = pd.concat([_forward_t, _backward_t, _forward_d, _backward_d], axis=1) + + return links + diff --git a/mario/tools/plots.py b/mario/tools/plots.py index 09bb2eb..c8cb9a0 100644 --- a/mario/tools/plots.py +++ b/mario/tools/plots.py @@ -9,9 +9,7 @@ from mario.tools.plots_manager import _PLOTS_LAYOUT, Color, _PALETTES from mario.tools.utilities import ( - subplot_grid, run_from_jupyter, - unit_check, ) import plotly.graph_objects as go @@ -20,7 +18,6 @@ import plotly.offline as pltly import pandas as pd import logging -import webbrowser logger = logging.getLogger(__name__) diff --git a/mario/tools/tableparser.py b/mario/tools/tableparser.py index 6f398b6..e226397 100644 --- a/mario/tools/tableparser.py +++ b/mario/tools/tableparser.py @@ -1170,7 +1170,7 @@ def parse_pymrio(io, value_added, satellite_account): """ # be sure that system is calculated - io = io.calc_system() + io = io.calc_all() extensions = {} for value in dir(io): diff --git a/mario/tools/utilities.py b/mario/tools/utilities.py index 04e212e..e66a24d 100644 --- a/mario/tools/utilities.py +++ b/mario/tools/utilities.py @@ -2,6 +2,7 @@ """ contains the utils functions in mario """ +#%% import pandas as pd import numpy as np import logging @@ -85,7 +86,8 @@ def run_from_jupyter(): ipy_str = str(type(get_ipython())) if "zmqshell" in ipy_str: return True - return False + + return False def sort_frames(_dict): @@ -99,88 +101,10 @@ def sort_frames(_dict): _dict[key] = value.sort_index(axis=1, level=1).sort_index(axis=0, level=1) -def validate_path(path: [str, list]): - - if isinstance(path, str): - path = [path] - - for item in path: - if not os.path.exists(item): - raise FileNotFoundError(f"{item} does not exist.") - - -def unique_frmaes(frame: pd.DataFrame,) -> list: - - all_items = [] - for col, values in frame.iteritems(): - given_values = delete_duplicates(values.values) - all_items.extend(given_values) - - return delete_duplicates(all_items) - - def delete_duplicates(_list: [list, tuple]) -> list: return list(dict.fromkeys(_list)) -def index_col_equlity( - original: pd.DataFrame, given: pd.DataFrame, levels: list = ["index", "columns"] -): - - for level in levels: - if not getattr(original, level).equals(getattr(given, level)): - raise WrongExcelFormat(f"{level} does not have the correct format.") - - -def _meta_parse_history(instance, function, old_index=None, new_index=None): - - if function == "parse": - instance.meta._add_history("Database successfully imported.") - for index in instance._indeces.keys(): - instance.meta._add_history( - "Number of {} = {}".format( - _MASTER_INDEX[index], len(instance._indeces[index]["main"]) - ) - ) - - elif function == "aggregation": - for index, value in instance._indeces.items(): - instance.meta._add_history( - "{} aggregated from {} levels to {} levels".format( - _MASTER_INDEX[index], - len(old_index[index]["main"]), - len(new_index[index]["main"]), - ) - ) - - -def _matrices(instance, function, scenario="baseline"): - """ - This function is in charge of deleting or adding the matrices as a callable - attribute to the instance - """ - # acceptable_matrices=_ALL_MATRICES+_SUT_MATRICES - - if function == "del": - - for attribute in _ALL_MATRICES[instance.table_type]: - if scenario != "baseline": - attribute = "{}_c".format(attribute) - try: - delattr(instance, attribute) - except AttributeError: - pass - - if function == "add": - pass - # --TODO-- check if it is necessary to have it - # for attribute,value in instance.matrices[scenario].items(): - # if scenario != 'baseline': - # attribute = '{}_c'.format(attribute) - - # setattr(instance, attribute,value) - - def _manage_indeces(instance, case, **kwargs): if case == "aggregation": @@ -197,89 +121,26 @@ def _manage_indeces(instance, case, **kwargs): for key, value in kwargs.items(): instance._indeces[key] = {"main": value} +def check_clusters(index_dict,table,clusters): -def subplot_grid(subplot_number, orientation="v"): - - if orientation == "v": - j = 0 - n_cols = [] - for i in reversed(range(subplot_number + 1)): - if int(math.sqrt(i) + 0.5) ** 2 == i: - n_cols += [int(math.sqrt(i))] - j += 1 - n_cols = n_cols[0] - - if int(math.sqrt(subplot_number) + 0.5) ** 2 == subplot_number: - n_rows = n_cols - else: - n_rows = n_cols + int(math.ceil((subplot_number - n_cols ** 2) / n_cols)) - - elif orientation == "h": - j = 0 - n_rows = [] - for i in reversed(range(subplot_number + 1)): - if int(math.sqrt(i) + 0.5) ** 2 == i: - n_rows += [int(math.sqrt(i))] - j += 1 - n_rows = n_rows[0] - - if int(math.sqrt(subplot_number) + 0.5) ** 2 == subplot_number: - n_cols = n_rows - else: - n_cols = n_rows + int(math.ceil((subplot_number - n_rows ** 2) / n_rows)) - - grid = [(row + 1, col + 1) for row in range(n_rows) for col in range(n_cols)] - - return (n_rows, n_cols, grid) - + differences = set(clusters).difference(set(_LEVELS[table])) -def str_2_list(*args): - - "this fucntion returns a list in case that the user pass a string" - - output = [] - for arg in args: - if isinstance(arg, str): - output.append([arg]) - else: - output.append(arg) - - return output - - -def master_exist(instance, **kwargs): - """ - 'this function checks if a specific item (any of the items from the MASTER_INDEX are valid or not' - instance = self - for kwargs: - key = the key of the MASTER_INDEX - value = a list of the items that the user is giving - """ - - for key, values in kwargs.items(): - for value in values: - if value not in instance.get_index(_MASTER_INDEX[key]): - raise WrongInput(f"{value} is not a valid {_MASTER_INDEX[key]}.") - - -def check_clusters(instance, clusters): + if differences: + raise WrongInput( + "{} is/are not valid level/s. Valid items are {}".format( + differences, [*_LEVELS[table]] + ) + ) for level, level_cluster in clusters.items(): - if level not in _LEVELS[instance.meta.table]: - raise WrongInput( - "{} is not a valid level. Valid items are {}".format( - level, [*_LEVELS[instance.meta.table]] - ) - ) for cluster, values in level_cluster.items(): - for value in values: - if value not in instance.get_index(level): - raise WrongInput( - "{} in cluster {} for level {} is not a valid item.".format( - value, cluster, level - ) + differences = set(values).difference(set(index_dict[level])) + if differences: + raise WrongInput( + "{} in cluster {} for level {} is/are not a valid item/s.".format( + differences, cluster, level ) - + ) def all_file_reader( path, guide, sub_folder=False, sep="\t", exceptions=[], engine=None @@ -344,15 +205,16 @@ def readers(file_to_read, file): return read -def return_index(df, item, multi_index, del_duplicate, reindex, level=None): +def return_index(df, item, multi_index, del_duplicate, reindex=None, level=None): if multi_index: - index = eval("df.{}.get_level_values({})".format(item, level)) + index = list(getattr(df,item).get_level_values(level)) + else: - index = eval("df.{}".format(item)) + index = list(getattr(df,item)) if del_duplicate: - index = delete_duplicates(list(index)) + index = delete_duplicates(index) return index @@ -368,7 +230,7 @@ def multiindex_contain(inner_index, outer_index, file, check_levels=None): if check_levels is None: if inner_index.nlevels != outer_index.nlevels: raise WrongInput(f"number levels for {file} are not valid.") - levels = [(i, i) for i in range(inner_index.nlevels)] + levels = list(range(inner_index.nlevels))# [i for i in range(inner_index.nlevels)] else: levels = check_levels @@ -377,12 +239,12 @@ def multiindex_contain(inner_index, outer_index, file, check_levels=None): for level in levels: diff = ( - outer_index.levels[level[1]] - .difference(inner_index.levels[level[0]]) + outer_index.levels[level] + .difference(inner_index.levels[level]) .tolist() ) - differences[level[1]] = diff + differences[level] = diff if len(diff): passed = False @@ -399,118 +261,11 @@ def rename_index(_dict): for key, value in _dict.items(): for item in ["index", "columns"]: - if isinstance(eval(f"value.{item}"), pd.MultiIndex): - exec(f"value.{item}.names = _INDEX_NAMES['3levels']") + if isinstance(getattr(value,item), pd.MultiIndex): + getattr(value,item).names= _INDEX_NAMES['3levels'] else: - exec(f"value.{item}.name = _INDEX_NAMES['1level']") - - -def linkages_calculation(instance, cut_diag, matrices, multi_mode, normalized): - """calculates the linkages""" - if cut_diag: - for key, value in matrices.items(): - np.fill_diagonal(value.values, 0) - - if multi_mode: - - link_types = [ - "Total Forward", - "Total Backward", - "Direct Forward", - "Direct Backward", - ] - geo_types = ["Local", "Foreign"] - links = pd.DataFrame( - 0, - index=matrices["g"].index, - columns=pd.MultiIndex.from_product([link_types, geo_types]), - ) - - for index, values in links.iterrows(): - links.loc[index, ("Total Forward", "Local")] = ( - matrices["g"].loc[index, index[0]].sum().sum() - ) - links.loc[index, ("Total Forward", "Foreign")] = ( - matrices["g"].loc[index].sum().sum() - - matrices["g"].loc[index, index[0]].sum().sum() - ) - - links.loc[index, ("Total Backward", "Local")] = ( - matrices["w"].loc[index, index[0]].sum().sum() - ) - links.loc[index, ("Total Backward", "Foreign")] = ( - matrices["w"].loc[index].sum().sum() - - matrices["w"].loc[index, index[0]].sum().sum() - ) - - links.loc[index, ("Direct Forward", "Local")] = ( - matrices["b"].loc[index, index[0]].sum().sum() - ) - links.loc[index, ("Direct Forward", "Foreign")] = ( - matrices["b"].loc[index].sum().sum() - - matrices["b"].loc[index, index[0]].sum().sum() - ) + getattr(value,item).name = _INDEX_NAMES['1level'] - links.loc[index, ("Direct Backward", "Local")] = ( - matrices["z"].loc[index, index[0]].sum().sum() - ) - links.loc[index, ("Direct Backward", "Foreign")] = ( - matrices["z"].loc[index].sum().sum() - - matrices["z"].loc[index, index[0]].sum().sum() - ) - - if normalized: - log_time( - logger, "Normalization not available for multi-regional mode.", "warn" - ) - - # Computing linkages as if there were only one unique region - else: - _forward_t = matrices["g"].sum(axis=1).to_frame() - _backward_t = matrices["w"].sum(axis=0).to_frame() - _forward_d = matrices["b"].sum(axis=1).to_frame() - _backward_d = matrices["z"].sum(axis=0).to_frame() - - _forward_t.columns = ["Total Forward"] - _backward_t.columns = ["Total Backward"] - _forward_d.columns = ["Direct Forward"] - _backward_d.columns = ["Direct Backward"] - - if normalized: - - _forward_t.iloc[:, 0] = _forward_t.iloc[:, 0] / np.average( - _forward_t.values - ) - _backward_t.iloc[:, 0] = _backward_t.iloc[:, 0] / np.average( - _backward_t.values - ) - _forward_d.iloc[:, 0] = _forward_d.iloc[:, 0] / np.average( - _forward_d.values - ) - _backward_d.iloc[:, 0] = _backward_d.iloc[:, 0] / np.average( - _backward_d.values - ) - - links = pd.concat([_forward_t, _backward_t, _forward_d, _backward_d], axis=1) - - return links - - -def unit_check(instance, sets, slicer=None): - - if sets == _MASTER_INDEX["r"]: - raise WrongInput(f"Set '{sets}' do not have any unit") - units = instance.units[sets] - if slicer is not None: - units = units.loc[slicer, :] - uniques = {} - - unique_units = units["unit"].unique() - - for unit in unique_units: - uniques[unit] = units.index[units["unit"] == unit].tolist() - - return uniques def filtering(instance, filters): @@ -572,3 +327,5 @@ def to_single_index(df): df.index = [", ".join(ii) for ii in df.index] return df + +# %%

Depricating the _meta_parse_history Avoinding the meta_parse_history by recording the number of items in a dataset

2022-04-11T16:04:21

0.0

[]

CybersecurityForDemocracy/tiktok-library

CybersecurityForDemocracy__tiktok-library-92

0302e5ac4e1e8f02f8ee2a0304936a183eae603c

diff --git a/src/tiktok_research_api_helper/api_client.py b/src/tiktok_research_api_helper/api_client.py index 5301d9d..13c6b2f 100644 --- a/src/tiktok_research_api_helper/api_client.py +++ b/src/tiktok_research_api_helper/api_client.py @@ -40,6 +40,7 @@ API_ERROR_RETRY_MAX_WAIT = timedelta(minutes=2).total_seconds() DAILY_API_REQUEST_QUOTA = 1000 +NULL_CHARACTERS_TO_STRIP = "\x00\u0000" class ApiRateLimitError(Exception): @@ -629,6 +630,8 @@ def _parse_video_response(response: rq.Response) -> TikTokVideoResponse: error_data = response_json.get("error") response_data_section = response_json.get("data", {}) videos = response_data_section.get("videos", []) + for video in videos: + strip_null_chars_from_json_values(video) return TikTokVideoResponse(data=response_data_section, videos=videos, error=error_data) @@ -637,6 +640,7 @@ def _parse_user_info_response(username: str, response: rq.Response) -> TikTokUse response_json = _extract_response_json_or_raise_error(response) error_data = response_json.get("error") response_data_section = response_json.get("data") + strip_null_chars_from_json_values(response_data_section) # API does not include username in response. for ease of use we add it. response_data_section["username"] = username @@ -652,11 +656,25 @@ def _parse_comments_response(response: rq.Response) -> TikTokCommentsResponse: response_json = _extract_response_json_or_raise_error(response) error_data = response_json.get("error") response_data_section = response_json.get("data", {}) + # TODO(macpd): this might have an issue with null characters in usernames comments = response_data_section.get("comments") + strip_null_chars_from_json_values(comments) return TikTokCommentsResponse(comments=comments, data=response_data_section, error=error_data) +def strip_null_chars_from_json_values(json_response: Mapping[str, Any]): + for key in json_response: + if isinstance(json_response[key], str): + json_response[key] = json_response[key].strip(NULL_CHARACTERS_TO_STRIP) + # Could do this recursively, but don't want a infinite recursion bug + elif isinstance(json_response[key], list): + json_response[key] = [ + v.strip(NULL_CHARACTERS_TO_STRIP) if isinstance(v, str) else v + for v in json_response[key] + ] + + def _extract_response_json_or_raise_error(response: rq.Response | None) -> Mapping[str, Any]: if response is None: raise ValueError("Response is None")

Some values from API have null byte/char, and therefore cannot be inserted to database directly for example video ID `7420068284780875054` has `'username': '٠вяеииа٠ء\x00\x00\x00\x00\x00\x00\x00\x00'` I don't know if we should strip the null character from these strings (`str.replace('\x00', '')`), or unicode escape them (`str.encode('unicode_escape')`)

2024-10-15T16:52:13

0.0

[]

ec-jrc/Thalassa

ec-jrc__Thalassa-91

b9d977cd6999e73f5ad884e9e7b96d4041b60827

diff --git a/README.md b/README.md index 66df9b7..738b674 100644 --- a/README.md +++ b/README.md @@ -14,6 +14,7 @@ Thalassa is currently supporting visualization of the output of the following so - [Schism](https://github.com/schism-dev/schism) - [ADCIRC](https://adcirc.org/) +- [openTELEMAC](https://www.opentelemac.org/) Adding support for new solvers is relatively straight-forward. @@ -30,6 +31,11 @@ Adding support for new solvers is relatively straight-forward. ``` pip install thalassa ``` +for Selafin files (openTELEMAC outputs): +``` +pip install thalassa +pip install xarray-selafin +``` ### Conda diff --git a/docs/index.md b/docs/index.md index b8108e2..366a5b1 100644 --- a/docs/index.md +++ b/docs/index.md @@ -16,6 +16,7 @@ Thalassa is currently supporting visualization of the output of the following so - [Schism](https://github.com/schism-dev/schism) - [ADCIRC](https://adcirc.org/) +- [TELEMAC](https://www.opentelemac.org) Adding support for new solvers is relatively straight-forward. diff --git a/docs/installation.md b/docs/installation.md index c01ac33..d8807f4 100644 --- a/docs/installation.md +++ b/docs/installation.md @@ -12,6 +12,11 @@ ``` pip install thalassa ``` +for Selafin files (openTELEMAC outputs): +``` +pip install thalassa +pip install xarray-selafin +``` ### Conda diff --git a/poetry.lock b/poetry.lock index df8ae78..5f50244 100644 --- a/poetry.lock +++ b/poetry.lock @@ -1,88 +1,88 @@ -# This file is automatically @generated by Poetry 1.7.1 and should not be changed by hand. +# This file is automatically @generated by Poetry 1.8.2 and should not be changed by hand. [[package]] name = "aiohttp" -version = "3.9.0b0" +version = "3.9.5" description = "Async http client/server framework (asyncio)" optional = false python-versions = ">=3.8" files = [ - {file = "aiohttp-3.9.0b0-cp310-cp310-macosx_10_9_universal2.whl", hash = "sha256:50b550b5e317e40a017bab8b25995676af3aa66dd0ef562cd7dce7f1684cd376"}, - {file = "aiohttp-3.9.0b0-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:8f902ad26b9814852e0a17d48f98ba4c879d8136c4fa9b235b5c043dde0a0257"}, - {file = "aiohttp-3.9.0b0-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:c2140de122ecf3eb7947105ceb91fb6632fb21cc1d17f6ff19c3973d2d12730d"}, - {file = "aiohttp-3.9.0b0-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:e014b343225d8d358ee91962b588e863fded12a6e2f9b446bb3be85c678e04ae"}, - {file = "aiohttp-3.9.0b0-cp310-cp310-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:c7016695087e616a2806ccdb1f83609e5fecb3958c270e3e5a42f69d225536f2"}, - {file = "aiohttp-3.9.0b0-cp310-cp310-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:40407d5ec81682225ad5538d9bd68b0f8242caa91e72a6a9a95197fd7d9aebb2"}, - {file = "aiohttp-3.9.0b0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:bd54502e6b4144785f2f14a5f1544ced0a77dbecb1fd422f21dfad95dcb7fcb8"}, - {file = "aiohttp-3.9.0b0-cp310-cp310-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:67f911fd2073621eecfe77b17926460e72980b9b996d0ab7dad5e38805ce2988"}, - {file = "aiohttp-3.9.0b0-cp310-cp310-musllinux_1_1_aarch64.whl", hash = "sha256:638ba28af2c821b70574664a991dfdfaf1a7a7ae1a8068757f7d59cdf2d8361a"}, - {file = "aiohttp-3.9.0b0-cp310-cp310-musllinux_1_1_i686.whl", hash = "sha256:76329f7c1f5f3185d91d61d64615d88fa3dfddf389a83f6cd46a205c5b61e01b"}, - {file = "aiohttp-3.9.0b0-cp310-cp310-musllinux_1_1_ppc64le.whl", hash = "sha256:597b5d44b613dea9c62779592eb0ecae87604628564ecaff8d516457def68184"}, - {file = "aiohttp-3.9.0b0-cp310-cp310-musllinux_1_1_s390x.whl", hash = "sha256:cd5edd7ba2b3f95346e0fc8ba2364bdd93917a1bf8528e7d60ec80cf21dfba7e"}, - {file = "aiohttp-3.9.0b0-cp310-cp310-musllinux_1_1_x86_64.whl", hash = "sha256:72556e0cce47c6e558454316fc5c6a3fb0980344eee8af7aa52b495d82ef12a5"}, - {file = "aiohttp-3.9.0b0-cp310-cp310-win32.whl", hash = "sha256:01a3b241288c4d8171fe5e2434a799d0b82700d2ed2156b43f1d7f4f521ba382"}, - {file = "aiohttp-3.9.0b0-cp310-cp310-win_amd64.whl", hash = "sha256:17962c404788c348ce5b58efaf4969249183c551890e30bfd9c035188d21e3d1"}, - {file = "aiohttp-3.9.0b0-cp311-cp311-macosx_10_9_universal2.whl", hash = "sha256:94197a77859ab1039b9ca6c3c393b8e7b5fc34a9abfbcb58daac38ab89684a99"}, - {file = "aiohttp-3.9.0b0-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:0c78d2cfe1515cfb31ba67edf0518c6677a963ec2039b652b03a886733e72e65"}, - {file = "aiohttp-3.9.0b0-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:28b38a14f564c833e59c99f748b48803e4babeabc6a0307952b01e6c8d642cab"}, - {file = "aiohttp-3.9.0b0-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:e596cfc52380f71e197e7cf0e2d3c4714b4bf66d2d562cdbd5442284bac18909"}, - {file = "aiohttp-3.9.0b0-cp311-cp311-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:6190951b7933c834d9346e21c5a81642caa210d291cda4036daf85fc53162d35"}, - {file = "aiohttp-3.9.0b0-cp311-cp311-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:fb0cb2cbf95cf4cc40307d0d0187f59c4b86b1d7d1a624922a7d0b046deffba7"}, - {file = "aiohttp-3.9.0b0-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:e27c283e21e94fa1582d31b57c514b87ab609882ade413ce43f585d73c8a33fc"}, - {file = "aiohttp-3.9.0b0-cp311-cp311-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:c6826c59b4e99673728bcdaecacbd699b7521f17ca165c63a5e26e23d42aeea5"}, - {file = "aiohttp-3.9.0b0-cp311-cp311-musllinux_1_1_aarch64.whl", hash = "sha256:aa4738f3b1b916b1cc69ed3d1dead9714919dc4d30ae0d5f6d55eadb2c511133"}, - {file = "aiohttp-3.9.0b0-cp311-cp311-musllinux_1_1_i686.whl", hash = "sha256:4b2abd7936f687de3a3ab199b145a9de01ed046eb5640cd66f47da07a9050a78"}, - {file = "aiohttp-3.9.0b0-cp311-cp311-musllinux_1_1_ppc64le.whl", hash = "sha256:652cc00a97bc206c470db06276ce57ff2a53a625795bbce8435ef8b6a4cb0113"}, - {file = "aiohttp-3.9.0b0-cp311-cp311-musllinux_1_1_s390x.whl", hash = "sha256:d54529c1d95d5d200ecb7133a343785e5661a804f3dcee090a7bca3b48189d69"}, - {file = "aiohttp-3.9.0b0-cp311-cp311-musllinux_1_1_x86_64.whl", hash = "sha256:324fe990c97721ea8eb4d439f12b59d1a93cd7e0dd188c7b145bffdfbd327dc3"}, - {file = "aiohttp-3.9.0b0-cp311-cp311-win32.whl", hash = "sha256:3a2ef8318435f40f5906af36fda20b5432e07e6a7e05de3a4d2934c25320b8ff"}, - {file = "aiohttp-3.9.0b0-cp311-cp311-win_amd64.whl", hash = "sha256:887d8757aafc7f6fbda76faaff21fc2aa31b9dca0911ecd6b60b0fe922a2abfc"}, - {file = "aiohttp-3.9.0b0-cp312-cp312-macosx_10_9_universal2.whl", hash = "sha256:9c430c706589a811b38e33e1492d194cbb0f6f2e027877bf038debced703446f"}, - {file = "aiohttp-3.9.0b0-cp312-cp312-macosx_10_9_x86_64.whl", hash = "sha256:9b820981f1c5d6da382e4859318ba78c9b5c583f0920e44a18efb3387b18487e"}, - {file = "aiohttp-3.9.0b0-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:c64677a2df742bcd89b94c35689306663d8246a8534bea5835afc706416f8dd6"}, - {file = "aiohttp-3.9.0b0-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:903155c179cda589d01936953158685747af43d98cdd3673a671c6e7f5c94178"}, - {file = "aiohttp-3.9.0b0-cp312-cp312-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:77cbb6e4a146449f805fa0e725b0b2a06411d21417d8eca699bbee55204201d0"}, - {file = "aiohttp-3.9.0b0-cp312-cp312-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:bc3cc9f5e6e493a2b9c3d241fca870b5a64aa4c247f1192f9e34fae990667df8"}, - {file = "aiohttp-3.9.0b0-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:92071206e570b7da6380f8d376820e2a40230638b8fd8b45b28103b346704c5e"}, - {file = "aiohttp-3.9.0b0-cp312-cp312-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:242e3cb0b2d441a2d20443114eebe3032078d1894ac1d97ab2dd101165ea50e1"}, - {file = "aiohttp-3.9.0b0-cp312-cp312-musllinux_1_1_aarch64.whl", hash = "sha256:044c5a8923bd44a4a0769a2886130c19f7f3a4a1a284f0ff68c2a751920ee39f"}, - {file = "aiohttp-3.9.0b0-cp312-cp312-musllinux_1_1_i686.whl", hash = "sha256:99b1b0d0f63ff48f80aa89be3ff61bc2b980c5b02895c81dbc1e44ce7b6cb5b7"}, - {file = "aiohttp-3.9.0b0-cp312-cp312-musllinux_1_1_ppc64le.whl", hash = "sha256:f737a47b5df97b6da457a0b2739d6d819ffadea2f36336988b53dbdb1796ba89"}, - {file = "aiohttp-3.9.0b0-cp312-cp312-musllinux_1_1_s390x.whl", hash = "sha256:e6d79f8b8347afbecd8047a1f6e74c810eb82497256cc906ee384635174dcaea"}, - {file = "aiohttp-3.9.0b0-cp312-cp312-musllinux_1_1_x86_64.whl", hash = "sha256:2f1b0a821564e315ec5cfa0abaf048355e229995a812380ec7a2200d87a6ed11"}, - {file = "aiohttp-3.9.0b0-cp312-cp312-win32.whl", hash = "sha256:ab2702f281ca504529e82be78dae2b9ca31d51a92ab8b239bd326b74c79d7af4"}, - {file = "aiohttp-3.9.0b0-cp312-cp312-win_amd64.whl", hash = "sha256:b81722b88abd4aab656abfec122646b6171da64340ff92af3bcf1af5f0d1275e"}, - {file = "aiohttp-3.9.0b0-cp38-cp38-macosx_10_9_universal2.whl", hash = "sha256:49e2ca017f506d1a9c60f44301ceff2eb8bbfe24b9cd9b4c4a363d9e5f68e92b"}, - {file = "aiohttp-3.9.0b0-cp38-cp38-macosx_10_9_x86_64.whl", hash = "sha256:06cba5518d8e30b46fcec2a8ed22ec6027fc9864583e0b538da642507f66fe29"}, - {file = "aiohttp-3.9.0b0-cp38-cp38-macosx_11_0_arm64.whl", hash = "sha256:e5201d3f8d0b2748eba5093820861639cac1ea1dfdff537f67152a1c082e1243"}, - {file = "aiohttp-3.9.0b0-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:8c483d0a666f6cbec2e974f760f93499bbcfcb17a7c4035d4c4c653e6a3b21b1"}, - {file = "aiohttp-3.9.0b0-cp38-cp38-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:04f48476ce3e96843b44084fd15139b195781c10ed6eb5ffb706fb9d2ca95ce4"}, - {file = "aiohttp-3.9.0b0-cp38-cp38-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:09fdad08544a4479e5801c777697c155fa9d966c91b6dcf3e1a0d271ad3999f7"}, - {file = "aiohttp-3.9.0b0-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:127aa57415005eb04fb1a3685c9d7b42aef6718be72b8a62b4b30ee00f7d23f4"}, - {file = "aiohttp-3.9.0b0-cp38-cp38-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:aa8f29f0647f10f6bcd9f597f1319d13ce1d6efe2d55169226940093eeadf609"}, - {file = "aiohttp-3.9.0b0-cp38-cp38-musllinux_1_1_aarch64.whl", hash = "sha256:8dc394dea47594825ac2a662c4fac6a8b294acd937396aaec8e41ed03728898b"}, - {file = "aiohttp-3.9.0b0-cp38-cp38-musllinux_1_1_i686.whl", hash = "sha256:c332b343974c6fbfec53e3ac7afebd6ba6cc1777cda67c28fabb3562411a9b5a"}, - {file = "aiohttp-3.9.0b0-cp38-cp38-musllinux_1_1_ppc64le.whl", hash = "sha256:6dfad718b328de3fa30d663393d51feea625322ec723bdecdec3f5f52ba6347f"}, - {file = "aiohttp-3.9.0b0-cp38-cp38-musllinux_1_1_s390x.whl", hash = "sha256:6edaeb63a4657672b04afcc25c253e960125e805f5a8f8cfa7bf682d15115f49"}, - {file = "aiohttp-3.9.0b0-cp38-cp38-musllinux_1_1_x86_64.whl", hash = "sha256:20023087bce5f3adde4872042ea1193d31d98b29682c28a6309d72bce0d9725e"}, - {file = "aiohttp-3.9.0b0-cp38-cp38-win32.whl", hash = "sha256:ad07ee4165a82e646310c152a74997c759d5782aef58bab9d77034b4cc87e153"}, - {file = "aiohttp-3.9.0b0-cp38-cp38-win_amd64.whl", hash = "sha256:494062a8447c6665f5237c47ca8bb5659cd3128ad9b4af5543566a11bb88df5c"}, - {file = "aiohttp-3.9.0b0-cp39-cp39-macosx_10_9_universal2.whl", hash = "sha256:aaff57bd1ab9eb1a205f3b7a00e2dc159d1e7e4373870be0d192358a656d9e60"}, - {file = "aiohttp-3.9.0b0-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:3c212f5066ffe9490856b706a9d9bd457f14716f4db4b1b73939245a1acecc4e"}, - {file = "aiohttp-3.9.0b0-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:d80664b3b82fb9ee2c7b13072651cd68d65fbb3a69721040c08969bab4335628"}, - {file = "aiohttp-3.9.0b0-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:e7cf539fc98297e312308405949ca2f04a347eb021e30d004388cdb5d155a0ec"}, - {file = "aiohttp-3.9.0b0-cp39-cp39-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:6be520717b895508c63df90e48135ba616c702a9229d4be71841dce2ea6a569f"}, - {file = "aiohttp-3.9.0b0-cp39-cp39-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:1b25e926cd16b44aeef29fffbb9fc9f577f52a6230e46926e391545b85cd0ce3"}, - {file = "aiohttp-3.9.0b0-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:35f6cafe361c0323945c13122c282ea22fb0df96e845f34c4d8abd96e2a81995"}, - {file = "aiohttp-3.9.0b0-cp39-cp39-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:5c9851e3d0396686d96a7e3559bf5912ed79c944ff1a6ae3cf7b1da320c3ad2b"}, - {file = "aiohttp-3.9.0b0-cp39-cp39-musllinux_1_1_aarch64.whl", hash = "sha256:0ab413eddeb1a03ba84d06acf7024a646b049d991ed0616bcc1ee40dc8fffa9e"}, - {file = "aiohttp-3.9.0b0-cp39-cp39-musllinux_1_1_i686.whl", hash = "sha256:89b271a8658472a9d400836ee8caee743246bae5c06405a63b6ba366f58df727"}, - {file = "aiohttp-3.9.0b0-cp39-cp39-musllinux_1_1_ppc64le.whl", hash = "sha256:dd941d473b86d0d5a413a1832499e5b80f648d66ca0c8246c26a4ccd66bcf7ec"}, - {file = "aiohttp-3.9.0b0-cp39-cp39-musllinux_1_1_s390x.whl", hash = "sha256:ce4f000279fb85527c017ef429615f2cb5a0cb614c088610849ddc6c2ac8d91b"}, - {file = "aiohttp-3.9.0b0-cp39-cp39-musllinux_1_1_x86_64.whl", hash = "sha256:f50a4f6773a9eedefb24b42c611e31dcd13f6139419a8656f7e525cb8a00687e"}, - {file = "aiohttp-3.9.0b0-cp39-cp39-win32.whl", hash = "sha256:b14dcfcc5ad161d007da71e1c1211909d527d9d7c2795ea9e17191ba25e5d89a"}, - {file = "aiohttp-3.9.0b0-cp39-cp39-win_amd64.whl", hash = "sha256:567245a91a57c41899f5d266814c9da8782d3d949dc1e66469429f08713a3ec6"}, - {file = "aiohttp-3.9.0b0.tar.gz", hash = "sha256:cecc64fd7bae6debdf43437e3c83183c40d4f4d86486946f412c113960598eee"}, + {file = "aiohttp-3.9.5-cp310-cp310-macosx_10_9_universal2.whl", hash = "sha256:fcde4c397f673fdec23e6b05ebf8d4751314fa7c24f93334bf1f1364c1c69ac7"}, + {file = "aiohttp-3.9.5-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:5d6b3f1fabe465e819aed2c421a6743d8debbde79b6a8600739300630a01bf2c"}, + {file = "aiohttp-3.9.5-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:6ae79c1bc12c34082d92bf9422764f799aee4746fd7a392db46b7fd357d4a17a"}, + {file = "aiohttp-3.9.5-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:4d3ebb9e1316ec74277d19c5f482f98cc65a73ccd5430540d6d11682cd857430"}, + {file = "aiohttp-3.9.5-cp310-cp310-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:84dabd95154f43a2ea80deffec9cb44d2e301e38a0c9d331cc4aa0166fe28ae3"}, + {file = "aiohttp-3.9.5-cp310-cp310-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:c8a02fbeca6f63cb1f0475c799679057fc9268b77075ab7cf3f1c600e81dd46b"}, + {file = "aiohttp-3.9.5-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:c26959ca7b75ff768e2776d8055bf9582a6267e24556bb7f7bd29e677932be72"}, + {file = "aiohttp-3.9.5-cp310-cp310-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:714d4e5231fed4ba2762ed489b4aec07b2b9953cf4ee31e9871caac895a839c0"}, + {file = "aiohttp-3.9.5-cp310-cp310-musllinux_1_1_aarch64.whl", hash = "sha256:e7a6a8354f1b62e15d48e04350f13e726fa08b62c3d7b8401c0a1314f02e3558"}, + {file = "aiohttp-3.9.5-cp310-cp310-musllinux_1_1_i686.whl", hash = "sha256:c413016880e03e69d166efb5a1a95d40f83d5a3a648d16486592c49ffb76d0db"}, + {file = "aiohttp-3.9.5-cp310-cp310-musllinux_1_1_ppc64le.whl", hash = "sha256:ff84aeb864e0fac81f676be9f4685f0527b660f1efdc40dcede3c251ef1e867f"}, + {file = "aiohttp-3.9.5-cp310-cp310-musllinux_1_1_s390x.whl", hash = "sha256:ad7f2919d7dac062f24d6f5fe95d401597fbb015a25771f85e692d043c9d7832"}, + {file = "aiohttp-3.9.5-cp310-cp310-musllinux_1_1_x86_64.whl", hash = "sha256:702e2c7c187c1a498a4e2b03155d52658fdd6fda882d3d7fbb891a5cf108bb10"}, + {file = "aiohttp-3.9.5-cp310-cp310-win32.whl", hash = "sha256:67c3119f5ddc7261d47163ed86d760ddf0e625cd6246b4ed852e82159617b5fb"}, + {file = "aiohttp-3.9.5-cp310-cp310-win_amd64.whl", hash = "sha256:471f0ef53ccedec9995287f02caf0c068732f026455f07db3f01a46e49d76bbb"}, + {file = "aiohttp-3.9.5-cp311-cp311-macosx_10_9_universal2.whl", hash = "sha256:e0ae53e33ee7476dd3d1132f932eeb39bf6125083820049d06edcdca4381f342"}, + {file = "aiohttp-3.9.5-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:c088c4d70d21f8ca5c0b8b5403fe84a7bc8e024161febdd4ef04575ef35d474d"}, + {file = "aiohttp-3.9.5-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:639d0042b7670222f33b0028de6b4e2fad6451462ce7df2af8aee37dcac55424"}, + {file = "aiohttp-3.9.5-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:f26383adb94da5e7fb388d441bf09c61e5e35f455a3217bfd790c6b6bc64b2ee"}, + {file = "aiohttp-3.9.5-cp311-cp311-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:66331d00fb28dc90aa606d9a54304af76b335ae204d1836f65797d6fe27f1ca2"}, + {file = "aiohttp-3.9.5-cp311-cp311-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:4ff550491f5492ab5ed3533e76b8567f4b37bd2995e780a1f46bca2024223233"}, + {file = "aiohttp-3.9.5-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:f22eb3a6c1080d862befa0a89c380b4dafce29dc6cd56083f630073d102eb595"}, + {file = "aiohttp-3.9.5-cp311-cp311-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:a81b1143d42b66ffc40a441379387076243ef7b51019204fd3ec36b9f69e77d6"}, + {file = "aiohttp-3.9.5-cp311-cp311-musllinux_1_1_aarch64.whl", hash = "sha256:f64fd07515dad67f24b6ea4a66ae2876c01031de91c93075b8093f07c0a2d93d"}, + {file = "aiohttp-3.9.5-cp311-cp311-musllinux_1_1_i686.whl", hash = "sha256:93e22add827447d2e26d67c9ac0161756007f152fdc5210277d00a85f6c92323"}, + {file = "aiohttp-3.9.5-cp311-cp311-musllinux_1_1_ppc64le.whl", hash = "sha256:55b39c8684a46e56ef8c8d24faf02de4a2b2ac60d26cee93bc595651ff545de9"}, + {file = "aiohttp-3.9.5-cp311-cp311-musllinux_1_1_s390x.whl", hash = "sha256:4715a9b778f4293b9f8ae7a0a7cef9829f02ff8d6277a39d7f40565c737d3771"}, + {file = "aiohttp-3.9.5-cp311-cp311-musllinux_1_1_x86_64.whl", hash = "sha256:afc52b8d969eff14e069a710057d15ab9ac17cd4b6753042c407dcea0e40bf75"}, + {file = "aiohttp-3.9.5-cp311-cp311-win32.whl", hash = "sha256:b3df71da99c98534be076196791adca8819761f0bf6e08e07fd7da25127150d6"}, + {file = "aiohttp-3.9.5-cp311-cp311-win_amd64.whl", hash = "sha256:88e311d98cc0bf45b62fc46c66753a83445f5ab20038bcc1b8a1cc05666f428a"}, + {file = "aiohttp-3.9.5-cp312-cp312-macosx_10_9_universal2.whl", hash = "sha256:c7a4b7a6cf5b6eb11e109a9755fd4fda7d57395f8c575e166d363b9fc3ec4678"}, + {file = "aiohttp-3.9.5-cp312-cp312-macosx_10_9_x86_64.whl", hash = "sha256:0a158704edf0abcac8ac371fbb54044f3270bdbc93e254a82b6c82be1ef08f3c"}, + {file = "aiohttp-3.9.5-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:d153f652a687a8e95ad367a86a61e8d53d528b0530ef382ec5aaf533140ed00f"}, + {file = "aiohttp-3.9.5-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:82a6a97d9771cb48ae16979c3a3a9a18b600a8505b1115cfe354dfb2054468b4"}, + {file = "aiohttp-3.9.5-cp312-cp312-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:60cdbd56f4cad9f69c35eaac0fbbdf1f77b0ff9456cebd4902f3dd1cf096464c"}, + {file = "aiohttp-3.9.5-cp312-cp312-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:8676e8fd73141ded15ea586de0b7cda1542960a7b9ad89b2b06428e97125d4fa"}, + {file = "aiohttp-3.9.5-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:da00da442a0e31f1c69d26d224e1efd3a1ca5bcbf210978a2ca7426dfcae9f58"}, + {file = "aiohttp-3.9.5-cp312-cp312-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:18f634d540dd099c262e9f887c8bbacc959847cfe5da7a0e2e1cf3f14dbf2daf"}, + {file = "aiohttp-3.9.5-cp312-cp312-musllinux_1_1_aarch64.whl", hash = "sha256:320e8618eda64e19d11bdb3bd04ccc0a816c17eaecb7e4945d01deee2a22f95f"}, + {file = "aiohttp-3.9.5-cp312-cp312-musllinux_1_1_i686.whl", hash = "sha256:2faa61a904b83142747fc6a6d7ad8fccff898c849123030f8e75d5d967fd4a81"}, + {file = "aiohttp-3.9.5-cp312-cp312-musllinux_1_1_ppc64le.whl", hash = "sha256:8c64a6dc3fe5db7b1b4d2b5cb84c4f677768bdc340611eca673afb7cf416ef5a"}, + {file = "aiohttp-3.9.5-cp312-cp312-musllinux_1_1_s390x.whl", hash = "sha256:393c7aba2b55559ef7ab791c94b44f7482a07bf7640d17b341b79081f5e5cd1a"}, + {file = "aiohttp-3.9.5-cp312-cp312-musllinux_1_1_x86_64.whl", hash = "sha256:c671dc117c2c21a1ca10c116cfcd6e3e44da7fcde37bf83b2be485ab377b25da"}, + {file = "aiohttp-3.9.5-cp312-cp312-win32.whl", hash = "sha256:5a7ee16aab26e76add4afc45e8f8206c95d1d75540f1039b84a03c3b3800dd59"}, + {file = "aiohttp-3.9.5-cp312-cp312-win_amd64.whl", hash = "sha256:5ca51eadbd67045396bc92a4345d1790b7301c14d1848feaac1d6a6c9289e888"}, + {file = "aiohttp-3.9.5-cp38-cp38-macosx_10_9_universal2.whl", hash = "sha256:694d828b5c41255e54bc2dddb51a9f5150b4eefa9886e38b52605a05d96566e8"}, + {file = "aiohttp-3.9.5-cp38-cp38-macosx_10_9_x86_64.whl", hash = "sha256:0605cc2c0088fcaae79f01c913a38611ad09ba68ff482402d3410bf59039bfb8"}, + {file = "aiohttp-3.9.5-cp38-cp38-macosx_11_0_arm64.whl", hash = "sha256:4558e5012ee03d2638c681e156461d37b7a113fe13970d438d95d10173d25f78"}, + {file = "aiohttp-3.9.5-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:9dbc053ac75ccc63dc3a3cc547b98c7258ec35a215a92bd9f983e0aac95d3d5b"}, + {file = "aiohttp-3.9.5-cp38-cp38-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:4109adee842b90671f1b689901b948f347325045c15f46b39797ae1bf17019de"}, + {file = "aiohttp-3.9.5-cp38-cp38-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:a6ea1a5b409a85477fd8e5ee6ad8f0e40bf2844c270955e09360418cfd09abac"}, + {file = "aiohttp-3.9.5-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:f3c2890ca8c59ee683fd09adf32321a40fe1cf164e3387799efb2acebf090c11"}, + {file = "aiohttp-3.9.5-cp38-cp38-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:3916c8692dbd9d55c523374a3b8213e628424d19116ac4308e434dbf6d95bbdd"}, + {file = "aiohttp-3.9.5-cp38-cp38-musllinux_1_1_aarch64.whl", hash = "sha256:8d1964eb7617907c792ca00b341b5ec3e01ae8c280825deadbbd678447b127e1"}, + {file = "aiohttp-3.9.5-cp38-cp38-musllinux_1_1_i686.whl", hash = "sha256:d5ab8e1f6bee051a4bf6195e38a5c13e5e161cb7bad83d8854524798bd9fcd6e"}, + {file = "aiohttp-3.9.5-cp38-cp38-musllinux_1_1_ppc64le.whl", hash = "sha256:52c27110f3862a1afbcb2af4281fc9fdc40327fa286c4625dfee247c3ba90156"}, + {file = "aiohttp-3.9.5-cp38-cp38-musllinux_1_1_s390x.whl", hash = "sha256:7f64cbd44443e80094309875d4f9c71d0401e966d191c3d469cde4642bc2e031"}, + {file = "aiohttp-3.9.5-cp38-cp38-musllinux_1_1_x86_64.whl", hash = "sha256:8b4f72fbb66279624bfe83fd5eb6aea0022dad8eec62b71e7bf63ee1caadeafe"}, + {file = "aiohttp-3.9.5-cp38-cp38-win32.whl", hash = "sha256:6380c039ec52866c06d69b5c7aad5478b24ed11696f0e72f6b807cfb261453da"}, + {file = "aiohttp-3.9.5-cp38-cp38-win_amd64.whl", hash = "sha256:da22dab31d7180f8c3ac7c7635f3bcd53808f374f6aa333fe0b0b9e14b01f91a"}, + {file = "aiohttp-3.9.5-cp39-cp39-macosx_10_9_universal2.whl", hash = "sha256:1732102949ff6087589408d76cd6dea656b93c896b011ecafff418c9661dc4ed"}, + {file = "aiohttp-3.9.5-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:c6021d296318cb6f9414b48e6a439a7f5d1f665464da507e8ff640848ee2a58a"}, + {file = "aiohttp-3.9.5-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:239f975589a944eeb1bad26b8b140a59a3a320067fb3cd10b75c3092405a1372"}, + {file = "aiohttp-3.9.5-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:3b7b30258348082826d274504fbc7c849959f1989d86c29bc355107accec6cfb"}, + {file = "aiohttp-3.9.5-cp39-cp39-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:cd2adf5c87ff6d8b277814a28a535b59e20bfea40a101db6b3bdca7e9926bc24"}, + {file = "aiohttp-3.9.5-cp39-cp39-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:e9a3d838441bebcf5cf442700e3963f58b5c33f015341f9ea86dcd7d503c07e2"}, + {file = "aiohttp-3.9.5-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:9e3a1ae66e3d0c17cf65c08968a5ee3180c5a95920ec2731f53343fac9bad106"}, + {file = "aiohttp-3.9.5-cp39-cp39-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:9c69e77370cce2d6df5d12b4e12bdcca60c47ba13d1cbbc8645dd005a20b738b"}, + {file = "aiohttp-3.9.5-cp39-cp39-musllinux_1_1_aarch64.whl", hash = "sha256:0cbf56238f4bbf49dab8c2dc2e6b1b68502b1e88d335bea59b3f5b9f4c001475"}, + {file = "aiohttp-3.9.5-cp39-cp39-musllinux_1_1_i686.whl", hash = "sha256:d1469f228cd9ffddd396d9948b8c9cd8022b6d1bf1e40c6f25b0fb90b4f893ed"}, + {file = "aiohttp-3.9.5-cp39-cp39-musllinux_1_1_ppc64le.whl", hash = "sha256:45731330e754f5811c314901cebdf19dd776a44b31927fa4b4dbecab9e457b0c"}, + {file = "aiohttp-3.9.5-cp39-cp39-musllinux_1_1_s390x.whl", hash = "sha256:3fcb4046d2904378e3aeea1df51f697b0467f2aac55d232c87ba162709478c46"}, + {file = "aiohttp-3.9.5-cp39-cp39-musllinux_1_1_x86_64.whl", hash = "sha256:8cf142aa6c1a751fcb364158fd710b8a9be874b81889c2bd13aa8893197455e2"}, + {file = "aiohttp-3.9.5-cp39-cp39-win32.whl", hash = "sha256:7b179eea70833c8dee51ec42f3b4097bd6370892fa93f510f76762105568cf09"}, + {file = "aiohttp-3.9.5-cp39-cp39-win_amd64.whl", hash = "sha256:38d80498e2e169bc61418ff36170e0aad0cd268da8b38a17c4cf29d254a8b3f1"}, + {file = "aiohttp-3.9.5.tar.gz", hash = "sha256:edea7d15772ceeb29db4aff55e482d4bcfb6ae160ce144f2682de02f6d693551"}, ] [package.dependencies] @@ -109,6 +109,17 @@ files = [ [package.dependencies] frozenlist = ">=1.1.0" +[[package]] +name = "ansicolors" +version = "1.1.8" +description = "ANSI colors for Python" +optional = false +python-versions = "*" +files = [ + {file = "ansicolors-1.1.8-py2.py3-none-any.whl", hash = "sha256:00d2dde5a675579325902536738dd27e4fac1fd68f773fe36c21044eb559e187"}, + {file = "ansicolors-1.1.8.zip", hash = "sha256:99f94f5e3348a0bcd43c82e5fc4414013ccc19d70bd939ad71e0133ce9c372e0"}, +] + [[package]] name = "appnope" version = "0.1.4" @@ -169,13 +180,13 @@ tests-no-zope = ["attrs[tests-mypy]", "cloudpickle", "hypothesis", "pympler", "p [[package]] name = "babel" -version = "2.14.0" +version = "2.15.0" description = "Internationalization utilities" optional = false -python-versions = ">=3.7" +python-versions = ">=3.8" files = [ - {file = "Babel-2.14.0-py3-none-any.whl", hash = "sha256:efb1a25b7118e67ce3a259bed20545c29cb68be8ad2c784c83689981b7a57287"}, - {file = "Babel-2.14.0.tar.gz", hash = "sha256:6919867db036398ba21eb5c7a0f6b28ab8cbc3ae7a73a44ebe34ae74a4e7d363"}, + {file = "Babel-2.15.0-py3-none-any.whl", hash = "sha256:08706bdad8d0a3413266ab61bd6c34d0c28d6e1e7badf40a2cebe67644e2e1fb"}, + {file = "babel-2.15.0.tar.gz", hash = "sha256:8daf0e265d05768bc6c7a314cf1321e9a123afc328cc635c18622a2f30a04413"}, ] [package.extras] @@ -204,33 +215,33 @@ lxml = ["lxml"] [[package]] name = "black" -version = "24.1.1" +version = "24.4.2" description = "The uncompromising code formatter." optional = false python-versions = ">=3.8" files = [ - {file = "black-24.1.1-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:2588021038bd5ada078de606f2a804cadd0a3cc6a79cb3e9bb3a8bf581325a4c"}, - {file = "black-24.1.1-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:1a95915c98d6e32ca43809d46d932e2abc5f1f7d582ffbe65a5b4d1588af7445"}, - {file = "black-24.1.1-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:2fa6a0e965779c8f2afb286f9ef798df770ba2b6cee063c650b96adec22c056a"}, - {file = "black-24.1.1-cp310-cp310-win_amd64.whl", hash = "sha256:5242ecd9e990aeb995b6d03dc3b2d112d4a78f2083e5a8e86d566340ae80fec4"}, - {file = "black-24.1.1-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:fc1ec9aa6f4d98d022101e015261c056ddebe3da6a8ccfc2c792cbe0349d48b7"}, - {file = "black-24.1.1-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:0269dfdea12442022e88043d2910429bed717b2d04523867a85dacce535916b8"}, - {file = "black-24.1.1-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:b3d64db762eae4a5ce04b6e3dd745dcca0fb9560eb931a5be97472e38652a161"}, - {file = "black-24.1.1-cp311-cp311-win_amd64.whl", hash = "sha256:5d7b06ea8816cbd4becfe5f70accae953c53c0e53aa98730ceccb0395520ee5d"}, - {file = "black-24.1.1-cp312-cp312-macosx_10_9_x86_64.whl", hash = "sha256:e2c8dfa14677f90d976f68e0c923947ae68fa3961d61ee30976c388adc0b02c8"}, - {file = "black-24.1.1-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:a21725862d0e855ae05da1dd25e3825ed712eaaccef6b03017fe0853a01aa45e"}, - {file = "black-24.1.1-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:07204d078e25327aad9ed2c64790d681238686bce254c910de640c7cc4fc3aa6"}, - {file = "black-24.1.1-cp312-cp312-win_amd64.whl", hash = "sha256:a83fe522d9698d8f9a101b860b1ee154c1d25f8a82ceb807d319f085b2627c5b"}, - {file = "black-24.1.1-cp38-cp38-macosx_10_9_x86_64.whl", hash = "sha256:08b34e85170d368c37ca7bf81cf67ac863c9d1963b2c1780c39102187ec8dd62"}, - {file = "black-24.1.1-cp38-cp38-macosx_11_0_arm64.whl", hash = "sha256:7258c27115c1e3b5de9ac6c4f9957e3ee2c02c0b39222a24dc7aa03ba0e986f5"}, - {file = "black-24.1.1-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:40657e1b78212d582a0edecafef133cf1dd02e6677f539b669db4746150d38f6"}, - {file = "black-24.1.1-cp38-cp38-win_amd64.whl", hash = "sha256:e298d588744efda02379521a19639ebcd314fba7a49be22136204d7ed1782717"}, - {file = "black-24.1.1-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:34afe9da5056aa123b8bfda1664bfe6fb4e9c6f311d8e4a6eb089da9a9173bf9"}, - {file = "black-24.1.1-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:854c06fb86fd854140f37fb24dbf10621f5dab9e3b0c29a690ba595e3d543024"}, - {file = "black-24.1.1-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:3897ae5a21ca132efa219c029cce5e6bfc9c3d34ed7e892113d199c0b1b444a2"}, - {file = "black-24.1.1-cp39-cp39-win_amd64.whl", hash = "sha256:ecba2a15dfb2d97105be74bbfe5128bc5e9fa8477d8c46766505c1dda5883aac"}, - {file = "black-24.1.1-py3-none-any.whl", hash = "sha256:5cdc2e2195212208fbcae579b931407c1fa9997584f0a415421748aeafff1168"}, - {file = "black-24.1.1.tar.gz", hash = "sha256:48b5760dcbfe5cf97fd4fba23946681f3a81514c6ab8a45b50da67ac8fbc6c7b"}, + {file = "black-24.4.2-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:dd1b5a14e417189db4c7b64a6540f31730713d173f0b63e55fabd52d61d8fdce"}, + {file = "black-24.4.2-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:8e537d281831ad0e71007dcdcbe50a71470b978c453fa41ce77186bbe0ed6021"}, + {file = "black-24.4.2-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:eaea3008c281f1038edb473c1aa8ed8143a5535ff18f978a318f10302b254063"}, + {file = "black-24.4.2-cp310-cp310-win_amd64.whl", hash = "sha256:7768a0dbf16a39aa5e9a3ded568bb545c8c2727396d063bbaf847df05b08cd96"}, + {file = "black-24.4.2-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:257d724c2c9b1660f353b36c802ccece186a30accc7742c176d29c146df6e474"}, + {file = "black-24.4.2-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:bdde6f877a18f24844e381d45e9947a49e97933573ac9d4345399be37621e26c"}, + {file = "black-24.4.2-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:e151054aa00bad1f4e1f04919542885f89f5f7d086b8a59e5000e6c616896ffb"}, + {file = "black-24.4.2-cp311-cp311-win_amd64.whl", hash = "sha256:7e122b1c4fb252fd85df3ca93578732b4749d9be076593076ef4d07a0233c3e1"}, + {file = "black-24.4.2-cp312-cp312-macosx_10_9_x86_64.whl", hash = "sha256:accf49e151c8ed2c0cdc528691838afd217c50412534e876a19270fea1e28e2d"}, + {file = "black-24.4.2-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:88c57dc656038f1ab9f92b3eb5335ee9b021412feaa46330d5eba4e51fe49b04"}, + {file = "black-24.4.2-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:be8bef99eb46d5021bf053114442914baeb3649a89dc5f3a555c88737e5e98fc"}, + {file = "black-24.4.2-cp312-cp312-win_amd64.whl", hash = "sha256:415e686e87dbbe6f4cd5ef0fbf764af7b89f9057b97c908742b6008cc554b9c0"}, + {file = "black-24.4.2-cp38-cp38-macosx_10_9_x86_64.whl", hash = "sha256:bf10f7310db693bb62692609b397e8d67257c55f949abde4c67f9cc574492cc7"}, + {file = "black-24.4.2-cp38-cp38-macosx_11_0_arm64.whl", hash = "sha256:98e123f1d5cfd42f886624d84464f7756f60ff6eab89ae845210631714f6db94"}, + {file = "black-24.4.2-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:48a85f2cb5e6799a9ef05347b476cce6c182d6c71ee36925a6c194d074336ef8"}, + {file = "black-24.4.2-cp38-cp38-win_amd64.whl", hash = "sha256:b1530ae42e9d6d5b670a34db49a94115a64596bc77710b1d05e9801e62ca0a7c"}, + {file = "black-24.4.2-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:37aae07b029fa0174d39daf02748b379399b909652a806e5708199bd93899da1"}, + {file = "black-24.4.2-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:da33a1a5e49c4122ccdfd56cd021ff1ebc4a1ec4e2d01594fef9b6f267a9e741"}, + {file = "black-24.4.2-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:ef703f83fc32e131e9bcc0a5094cfe85599e7109f896fe8bc96cc402f3eb4b6e"}, + {file = "black-24.4.2-cp39-cp39-win_amd64.whl", hash = "sha256:b9176b9832e84308818a99a561e90aa479e73c523b3f77afd07913380ae2eab7"}, + {file = "black-24.4.2-py3-none-any.whl", hash = "sha256:d36ed1124bb81b32f8614555b34cc4259c3fbc7eec17870e8ff8ded335b58d8c"}, + {file = "black-24.4.2.tar.gz", hash = "sha256:c872b53057f000085da66a19c55d68f6f8ddcac2642392ad3a355878406fbd4d"}, ] [package.dependencies] @@ -268,76 +279,76 @@ css = ["tinycss2 (>=1.1.0,<1.3)"] [[package]] name = "bokeh" -version = "3.3.4" +version = "3.4.1" description = "Interactive plots and applications in the browser from Python" optional = false python-versions = ">=3.9" files = [ - {file = "bokeh-3.3.4-py3-none-any.whl", hash = "sha256:ad7b6f89d0a7c2be01eff1db0ca24e2755ac41de14539db919a62e791809c309"}, - {file = "bokeh-3.3.4.tar.gz", hash = "sha256:73b7982dc2b8df15bf660cdddc8d3825e829195c438015a5d09824f1a7028368"}, + {file = "bokeh-3.4.1-py3-none-any.whl", hash = "sha256:1e3c502a0a8205338fc74dadbfa321f8a0965441b39501e36796a47b4017b642"}, + {file = "bokeh-3.4.1.tar.gz", hash = "sha256:d824961e4265367b0750ce58b07e564ad0b83ca64b335521cd3421e9b9f10d89"}, ] [package.dependencies] -contourpy = ">=1" +contourpy = ">=1.2" Jinja2 = ">=2.9" numpy = ">=1.16" packaging = ">=16.8" pandas = ">=1.2" pillow = ">=7.1.0" PyYAML = ">=3.10" -tornado = ">=5.1" +tornado = ">=6.2" xyzservices = ">=2021.09.1" [[package]] name = "bottleneck" -version = "1.3.7" +version = "1.3.8" description = "Fast NumPy array functions written in C" optional = false python-versions = "*" files = [ - {file = "Bottleneck-1.3.7-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:ada1a9ff93fd6b1b19f12398a6761940372b00e53d86db98bd4613a751c60043"}, - {file = "Bottleneck-1.3.7-cp310-cp310-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:a477101ee12155a0d2f9a82cd3e2a44b9b1aa53afe5b20acc065c91cf35c3106"}, - {file = "Bottleneck-1.3.7-cp310-cp310-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:cfe5f3010e9ad8ae54871d1e8fd61109c5981ed8d9d14e8496a1c37fe2050a04"}, - {file = "Bottleneck-1.3.7-cp310-cp310-musllinux_1_1_i686.whl", hash = "sha256:955df78713ff58cf815c0a1fa4782b2dc51a8787c0971688472c64b267303855"}, - {file = "Bottleneck-1.3.7-cp310-cp310-musllinux_1_1_x86_64.whl", hash = "sha256:92fc5d3ecfedde3b28c56ca686cff70a8125d1ddc281eb468b5e9d6a61269802"}, - {file = "Bottleneck-1.3.7-cp310-cp310-win32.whl", hash = "sha256:e2a290dcb148c0ddf182052e333892e46730c7d39a1f251af87e3d81a43cdde3"}, - {file = "Bottleneck-1.3.7-cp310-cp310-win_amd64.whl", hash = "sha256:9170ebee4cff423c92b3760afec179bded90eaede7c70dd27cf5f406cc00a1e7"}, - {file = "Bottleneck-1.3.7-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:752c97d64ebebdc10a5568d97b81b4971238fa4b53533248d227c4ea759aee4e"}, - {file = "Bottleneck-1.3.7-cp311-cp311-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:bbda2b27e81a47b7578bce6a8cb2f5eb899279c828d8efb5a154d8ede785093d"}, - {file = "Bottleneck-1.3.7-cp311-cp311-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:79160e49a5a0438468b970967e29addde2d9c6a6ce930144de7ccd8151077603"}, - {file = "Bottleneck-1.3.7-cp311-cp311-musllinux_1_1_i686.whl", hash = "sha256:abfc22762640cc7716c1bd9a409bde0f834167a2584775eb644c6afe7bae3319"}, - {file = "Bottleneck-1.3.7-cp311-cp311-musllinux_1_1_x86_64.whl", hash = "sha256:2a39049aa1cb798538f96150098f50badb37417c41cecfa8a441c0c4e6433c7e"}, - {file = "Bottleneck-1.3.7-cp311-cp311-win32.whl", hash = "sha256:b624023f173332d209ed5bd4134fae43d2432d61c6b17a49a6b7c5591caa7cc4"}, - {file = "Bottleneck-1.3.7-cp311-cp311-win_amd64.whl", hash = "sha256:bd69b1844c90f6bcf1b679a608fb2c0909be5e045f91674d61a0e4c3596644be"}, - {file = "Bottleneck-1.3.7-cp36-cp36m-macosx_10_9_x86_64.whl", hash = "sha256:c8fa2d85298cdd0f88f51fb46604039abe9a5a2b57e00acfe2b64f546754c053"}, - {file = "Bottleneck-1.3.7-cp36-cp36m-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:a09fa97df2f572aed5da487d3fe467829c2212c97b1d63c768ec9be9fd1c57a9"}, - {file = "Bottleneck-1.3.7-cp36-cp36m-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:f239ba0d992c013564979cff4fecf2b251614de3730641e79d05ecc09268c00c"}, - {file = "Bottleneck-1.3.7-cp36-cp36m-musllinux_1_1_i686.whl", hash = "sha256:f5409cf6db5e60256626a7c110b2b4c80d9f6b0d1ec0b66565bbab978f96b7a6"}, - {file = "Bottleneck-1.3.7-cp36-cp36m-musllinux_1_1_x86_64.whl", hash = "sha256:b3c65a9aedada04872cba450676d132d7af8b9da47f58b82f20ffe9b9ff77a46"}, - {file = "Bottleneck-1.3.7-cp36-cp36m-win32.whl", hash = "sha256:b4a47f972e919d22c1725ff7262d209e6ea3a4d9bcfea71fb454a18e166515ec"}, - {file = "Bottleneck-1.3.7-cp36-cp36m-win_amd64.whl", hash = "sha256:8d6865327ebbea5578f99b073538789df2123cf9009e99fe94efc4d25a4b888b"}, - {file = "Bottleneck-1.3.7-cp37-cp37m-macosx_10_9_x86_64.whl", hash = "sha256:685812254238c2059810f8e25c8215b09795b974f5b1a89f0accda3d93cc8734"}, - {file = "Bottleneck-1.3.7-cp37-cp37m-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:5977f1fd02f01035176a99c01b3fb0eefdaff002c30a4710f7279215b0a317f5"}, - {file = "Bottleneck-1.3.7-cp37-cp37m-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:e26c655736701365d66c64056fc272952d0e2a52718c637254d7dd9a7efa97a8"}, - {file = "Bottleneck-1.3.7-cp37-cp37m-musllinux_1_1_i686.whl", hash = "sha256:d982e1e3f72cc7ca4f558e9ef7db7a97ce68a915b1db8249fe088b3e78974b21"}, - {file = "Bottleneck-1.3.7-cp37-cp37m-musllinux_1_1_x86_64.whl", hash = "sha256:b5ec31e0b52cbfd716255695ec5fe87662a5961d2b8b49936f3608e36a04d926"}, - {file = "Bottleneck-1.3.7-cp37-cp37m-win32.whl", hash = "sha256:f2c8b631552908b11905cf87b4a90dd4af332b0726dd9b49d26d6a0fbb38e6d4"}, - {file = "Bottleneck-1.3.7-cp37-cp37m-win_amd64.whl", hash = "sha256:1045f9b199bba2675bd7642c25dec5daa21a03a5f0444eb863f2790af52b00b0"}, - {file = "Bottleneck-1.3.7-cp38-cp38-macosx_10_9_x86_64.whl", hash = "sha256:5af0c679da9dc41d6a3c436785c664fff41b4c1178d46afaa630620ab31e970f"}, - {file = "Bottleneck-1.3.7-cp38-cp38-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:c65d0a9498b2ccda236b28b1b36bc35402e81ebc4575b6cf44033b7e4460f067"}, - {file = "Bottleneck-1.3.7-cp38-cp38-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:752f4c3fc5b7cc49814d41e9ecf6f228f33ea407bdb8c889da53716b5628abc7"}, - {file = "Bottleneck-1.3.7-cp38-cp38-musllinux_1_1_i686.whl", hash = "sha256:54fc579445688c5309a0af7634db3be4e17659c6b664f9332ec404738e847894"}, - {file = "Bottleneck-1.3.7-cp38-cp38-musllinux_1_1_x86_64.whl", hash = "sha256:175ae3d92142769a5acd169ac2d7df22107b72f6d4c932ff56f5afcef39d5877"}, - {file = "Bottleneck-1.3.7-cp38-cp38-win32.whl", hash = "sha256:64ef9ad7187282745205b735e8c33307ec14c7be39f9971bc3f375104355cfb9"}, - {file = "Bottleneck-1.3.7-cp38-cp38-win_amd64.whl", hash = "sha256:fa80c318b5164e39e4f2d5abf95f5ff42744e3d8535a4c85ede94ccc5fca8f9b"}, - {file = "Bottleneck-1.3.7-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:01e61a512d645193cf4cf4fdacf98d3140c26ace0fcf4c4ed8fdfa366c57e0d1"}, - {file = "Bottleneck-1.3.7-cp39-cp39-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:699fb76ef50cf85be8f8d644b533488de3cec4c6ddf00fd770f73cc5caac9938"}, - {file = "Bottleneck-1.3.7-cp39-cp39-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:6311ee47641ed5fbc543d2e49cbfa90bd9ece208a0c1bdc2c2b14b9132982b2a"}, - {file = "Bottleneck-1.3.7-cp39-cp39-musllinux_1_1_i686.whl", hash = "sha256:a1e50fb506693186a4b8f85ac79df1a5af1ab1f774bf20004280e0658a731a89"}, - {file = "Bottleneck-1.3.7-cp39-cp39-musllinux_1_1_x86_64.whl", hash = "sha256:4d2ae3e1ac0626107c6916783befc39473900bb22a9b275e21550c99abe8283c"}, - {file = "Bottleneck-1.3.7-cp39-cp39-win32.whl", hash = "sha256:68d0f9d32d45f62028ab27d2b51d1a2af72a5ca6a7c3b1f86e9115fedb266300"}, - {file = "Bottleneck-1.3.7-cp39-cp39-win_amd64.whl", hash = "sha256:83d71c49dd9d6b99def958b6ccba3c8b5aac7b90849a5a9fe935648436dd46b9"}, - {file = "Bottleneck-1.3.7.tar.gz", hash = "sha256:e1467e373ad469da340ed0ff283214d6531cc08bfdca2083361a3aa6470681f8"}, + {file = "Bottleneck-1.3.8-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:865c8ed5b798c0198b0b80553e09cc0d890c4f5feb3d81d31661517ca7819fa3"}, + {file = "Bottleneck-1.3.8-cp310-cp310-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:d073a31e259d40b25e29dbba80f73abf38afe98fd730c79dad7edd9a0ad6cff5"}, + {file = "Bottleneck-1.3.8-cp310-cp310-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:b806b277ab47495032822f55f43b8d336e4b7e73f8506ed34d3ea3da6d644abc"}, + {file = "Bottleneck-1.3.8-cp310-cp310-musllinux_1_1_i686.whl", hash = "sha256:770b517609916adeb39d3b1a386a29bc316da03dd61e7ee6e8a38325b80cc327"}, + {file = "Bottleneck-1.3.8-cp310-cp310-musllinux_1_1_x86_64.whl", hash = "sha256:2948502b0394ee419945b55b092585222a505c61d41a874c741be49f2cac056f"}, + {file = "Bottleneck-1.3.8-cp310-cp310-win32.whl", hash = "sha256:271b6333522beb8aee32e640ba49a2064491d2c10317baa58a5996be3dd443e4"}, + {file = "Bottleneck-1.3.8-cp310-cp310-win_amd64.whl", hash = "sha256:d41000ea7ca196b5fd39d6fccd34bf0704c8831731cedd2da2dcae3c6ac49c42"}, + {file = "Bottleneck-1.3.8-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:d0a7f454394cd3642498b6e077e70f4a6b9fd46a8eb908c83ac737fdc9f9a98c"}, + {file = "Bottleneck-1.3.8-cp311-cp311-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:2c4ea8b9024dcb4e83b5c118a3c8faa863ace2ad572849da548a74a8ee4e8f2a"}, + {file = "Bottleneck-1.3.8-cp311-cp311-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:f40724b6e965ff5b88b333d4a10097b1629e60c0db21bb3d08c24d7b1a904a16"}, + {file = "Bottleneck-1.3.8-cp311-cp311-musllinux_1_1_i686.whl", hash = "sha256:4bd7183b8dcca89d0e65abe4507c19667dd31dacfbcc8ed705bad642f26a46e1"}, + {file = "Bottleneck-1.3.8-cp311-cp311-musllinux_1_1_x86_64.whl", hash = "sha256:20aa31a7d9d747c499ace1610a6e1f7aba6e3d4a9923e0312f6b4b6d68a59af3"}, + {file = "Bottleneck-1.3.8-cp311-cp311-win32.whl", hash = "sha256:350520105d9449e6565b3f0c4ce1f80a0b3e4d63695ebbf29db41f62e13f6461"}, + {file = "Bottleneck-1.3.8-cp311-cp311-win_amd64.whl", hash = "sha256:167a278902775defde7dfded6e98e3707dfe54971ffd9aec25c43bc74e4e381a"}, + {file = "Bottleneck-1.3.8-cp312-cp312-macosx_10_9_x86_64.whl", hash = "sha256:c6e93ed45c6c83392f73d0333b310b38772df7eb78c120c1447245691bdedaf4"}, + {file = "Bottleneck-1.3.8-cp312-cp312-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:d3400f47dda0196b5af50b0b0678e33cc8c42e52e55ae0a63cdfed60725659bc"}, + {file = "Bottleneck-1.3.8-cp312-cp312-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:fba5fd1805c71b2eeea50bea93d59be449c4af23ebd8da5f75fd74fd0331e314"}, + {file = "Bottleneck-1.3.8-cp312-cp312-musllinux_1_1_i686.whl", hash = "sha256:60139c5c3d2a9c1454a04af5ee981a9f56548d27fa36f264069b149a6e9b01ed"}, + {file = "Bottleneck-1.3.8-cp312-cp312-musllinux_1_1_x86_64.whl", hash = "sha256:99fab17fa26c811ccad63e208314726e718ae6605314329eca09641954550523"}, + {file = "Bottleneck-1.3.8-cp312-cp312-win32.whl", hash = "sha256:d3ae2bb5d4168912e438e377cc1301fa01df949ba59cd86317b3e00404fd4a97"}, + {file = "Bottleneck-1.3.8-cp312-cp312-win_amd64.whl", hash = "sha256:bcba1d5d5328c50f94852ab521fcb26f35d9e0ccd928d120d56455d1a5bb743f"}, + {file = "Bottleneck-1.3.8-cp37-cp37m-macosx_10_9_x86_64.whl", hash = "sha256:8d01fd5389d3160d54619119987ac24b020fa6810b7b398fff4945892237b3da"}, + {file = "Bottleneck-1.3.8-cp37-cp37m-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:7ca25f0003ef65264942f6306d793e0f270ece8b406c5a293dfc7d878146e9f8"}, + {file = "Bottleneck-1.3.8-cp37-cp37m-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:bf7763cf1516fa388c3587d12182fc1bc1c8089eab1a0a1bf09761f4c41af73c"}, + {file = "Bottleneck-1.3.8-cp37-cp37m-musllinux_1_1_i686.whl", hash = "sha256:38837c022350e2a656453f0e448416b7108cf67baccf11d04a0b3b70a48074dd"}, + {file = "Bottleneck-1.3.8-cp37-cp37m-musllinux_1_1_x86_64.whl", hash = "sha256:84ca5e741fae1c1796744dbdd0d2c1789cb74dd79c12ea8ec5834f83430f8520"}, + {file = "Bottleneck-1.3.8-cp37-cp37m-win32.whl", hash = "sha256:f4dfc22a3450227e692ef2ff4657639c33eec88ad04ee3ce29d1a23a4942da24"}, + {file = "Bottleneck-1.3.8-cp37-cp37m-win_amd64.whl", hash = "sha256:90b87eed152bbd760c4eb11473c2cf036abdb26e2f84caeb00787da74fb08c40"}, + {file = "Bottleneck-1.3.8-cp38-cp38-macosx_10_9_x86_64.whl", hash = "sha256:54a1b5d9d63b2d9f2955f8542eea26c418f97873e0abf86ca52beea0208c9306"}, + {file = "Bottleneck-1.3.8-cp38-cp38-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:019dd142d1e870388fb0b649213a0d8e569cce784326e183deba8f17826edd9f"}, + {file = "Bottleneck-1.3.8-cp38-cp38-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:8b5ed34a540eb7df59f45da659af9f792306637de1c69c95f020294f3b9fc4a8"}, + {file = "Bottleneck-1.3.8-cp38-cp38-musllinux_1_1_i686.whl", hash = "sha256:b69fcd4d818bcf9d53497d8accd0d5f852a447728baaa33b9b7168f8c4221d06"}, + {file = "Bottleneck-1.3.8-cp38-cp38-musllinux_1_1_x86_64.whl", hash = "sha256:02616a830bd477f5ba51103396092da4b9d83cea2e88f5b8069e3f4f7b796704"}, + {file = "Bottleneck-1.3.8-cp38-cp38-win32.whl", hash = "sha256:93d359fb83eb3bdd6635ef6e64835c38ffdc211441fc190549f286e6af98b5f6"}, + {file = "Bottleneck-1.3.8-cp38-cp38-win_amd64.whl", hash = "sha256:51c8bb3dffeb72c14f0382b80de76eabac6726d316babbd48f7e4056267d7910"}, + {file = "Bottleneck-1.3.8-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:84453548b0f722c3be912ce3c6b685917fea842bf1252eeb63714a2c1fd1ffc9"}, + {file = "Bottleneck-1.3.8-cp39-cp39-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:92700867504a213cafa9b8d9be529bd6e18dc83366b2ba00e86e80769b93f678"}, + {file = "Bottleneck-1.3.8-cp39-cp39-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:fadfd2f3931fdff42f4b9867eb02ed7c662d01e6099ff6b347b6ced791450651"}, + {file = "Bottleneck-1.3.8-cp39-cp39-musllinux_1_1_i686.whl", hash = "sha256:cfbc4a3a934b677bfbc37ac8757c4e1264a76262b774259bd3fa8a265dbd668b"}, + {file = "Bottleneck-1.3.8-cp39-cp39-musllinux_1_1_x86_64.whl", hash = "sha256:3c74c18f86a1ffac22280b005df8bb8a58505ac6663c4d6807f39873c17dc347"}, + {file = "Bottleneck-1.3.8-cp39-cp39-win32.whl", hash = "sha256:211f881159e8adb3a57df2263028ae6dc89ec4328bfd43f3421e507406c28654"}, + {file = "Bottleneck-1.3.8-cp39-cp39-win_amd64.whl", hash = "sha256:8615eeb75009ba7c0a112a5a6a5154ed3d61fd6b0879631778b3e42e2d9a6d65"}, + {file = "Bottleneck-1.3.8.tar.gz", hash = "sha256:6780d896969ba7f53c8995ba90c87c548beb3db435dc90c60b9a10ed1ab4d868"}, ] [package.dependencies] @@ -348,42 +359,44 @@ doc = ["gitpython", "numpydoc", "sphinx"] [[package]] name = "cartopy" -version = "0.22.0" +version = "0.23.0" description = "A Python library for cartographic visualizations with Matplotlib" optional = false python-versions = ">=3.9" files = [ - {file = "Cartopy-0.22.0-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:b17ad0b056b9a632b12954864c685febb0e1d8a4a45423b83eec119a603fcb8a"}, - {file = "Cartopy-0.22.0-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:bf7ca13a01782810e8b6a9758ed29755b6ce81b5c53721e19cb6636a43b0f575"}, - {file = "Cartopy-0.22.0-cp310-cp310-manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:604c21046dfbe0c9551b28802901d240a7ce25fcaf0c30db0f230cebf93ae2a5"}, - {file = "Cartopy-0.22.0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:4c37d4e2227cb41c971d2bd2555ce081f765d4eaae852def1c059c1d85c8e645"}, - {file = "Cartopy-0.22.0-cp310-cp310-win_amd64.whl", hash = "sha256:3f9c391bd3ba588de397854556dda175edf59f614bbb6dce18c4981154d97d92"}, - {file = "Cartopy-0.22.0-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:3d9d24ef50991269b57a42bdd4f1156426065fbeb41777c7e28d937e2c4a2ae6"}, - {file = "Cartopy-0.22.0-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:472e6713aad729f6fd38bd301666febd23dd3b744b0530cd80e02e242e3f1c74"}, - {file = "Cartopy-0.22.0-cp311-cp311-manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:319b8244f6d06f600be89ad5eaa85c2524e23b240decd203188da26ff14a346b"}, - {file = "Cartopy-0.22.0-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:a401109b88a6bfec1c13dce9b09dc59993c3af426b1879e1e0ee3afd1c293076"}, - {file = "Cartopy-0.22.0-cp311-cp311-win_amd64.whl", hash = "sha256:8ecebbe13d8488c2ec32e7c40222b9f10ba48e386be30c8ec60c867c122399f9"}, - {file = "Cartopy-0.22.0-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:fbc761b80d2fad8cb6ea1c7d3388862e9fbc53c17f14baf6d4038561c7865e6f"}, - {file = "Cartopy-0.22.0-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:3b25dd55844f36115e6c3632a16ed545cbee011068777b5a7a49ed3e1dbcafcb"}, - {file = "Cartopy-0.22.0-cp39-cp39-manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:eec1e9b701272f4cfc43d0123b9f69fa146d915580ee47eda875770c704bf413"}, - {file = "Cartopy-0.22.0-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:8bdc98a5b196488ea824b6e2f07a3ab63c4f1e99a2ab9fdee514a5f38b00407a"}, - {file = "Cartopy-0.22.0-cp39-cp39-win_amd64.whl", hash = "sha256:c5df7fc17705c93a152f3e82d31adb9c2db50448b38b08c076331c522ce7f844"}, - {file = "Cartopy-0.22.0.tar.gz", hash = "sha256:b300f90120931d43f11ef87c064ea1dacec1b59a4940aa76ebf82cf09548bb49"}, + {file = "Cartopy-0.23.0-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:374e66f816c3bafa48ffdbf6abaefa67063b405fac5f425f9be241cdf3498352"}, + {file = "Cartopy-0.23.0-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:2bae450c4c913796cad0b7ce05aa2fa78d1788de47989f0a03183397648e24be"}, + {file = "Cartopy-0.23.0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:a40437596e8ac5e74575eab822c661f4e725bd995cfd9e445069695fe9086b42"}, + {file = "Cartopy-0.23.0-cp310-cp310-win_amd64.whl", hash = "sha256:3292d6d403137eed80d32014c2f28de6282bed8824213f4b4c2170f388b24a1b"}, + {file = "Cartopy-0.23.0-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:86b07b6794b616674e4e485b8574e9197bca54a4467d28dd01ae0bf178f8dc2b"}, + {file = "Cartopy-0.23.0-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:8dece2aa8d5ff7bf989ded6b5f07c980fb5bb772952bc7cdeab469738abdecee"}, + {file = "Cartopy-0.23.0-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:a9dfd28352dc83d6b4e4cf85d84cb50fc4886d4c1510d61f4c7cf22477d1156f"}, + {file = "Cartopy-0.23.0-cp311-cp311-win_amd64.whl", hash = "sha256:b2671b5354e43220f8e1074e7fe30a8b9f71cb38407c78e51db9c97772f0320b"}, + {file = "Cartopy-0.23.0-cp312-cp312-macosx_10_9_x86_64.whl", hash = "sha256:80b9fd666fd47f6370d29f7ad4e352828d54aaf688a03d0b83b51e141cfd77fa"}, + {file = "Cartopy-0.23.0-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:43e36b8b7e7e373a5698757458fd28fafbbbf5f3ebbe2d378f6a5ec3993d6dc0"}, + {file = "Cartopy-0.23.0-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:550173b91155d4d81cd14b4892cb6cabe3dd32bd34feacaa1ec78c0e56287832"}, + {file = "Cartopy-0.23.0-cp312-cp312-win_amd64.whl", hash = "sha256:55219ee0fb069cc3254426e87382cde03546e86c3f7c6759f076823b1e3a44d9"}, + {file = "Cartopy-0.23.0-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:6279af846bf77d9817ab8792a8e38ca561878f048bba1afdae3e3a30c5432bfd"}, + {file = "Cartopy-0.23.0-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:843bf9dc0a18e1a8eed872c49e8092e8a8109e4dce285ad96752841e21e8161e"}, + {file = "Cartopy-0.23.0-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:350ff8802e2bc617c09bd6148aeb46e841775a846bfaa6e635a212d1eaf5ab66"}, + {file = "Cartopy-0.23.0-cp39-cp39-win_amd64.whl", hash = "sha256:b52ab2274ad7504955854ef8d6f603e41f5d7163d02b29d369cecdbd29c2fda1"}, + {file = "Cartopy-0.23.0.tar.gz", hash = "sha256:231f37b35701f2ba31d94959cca75e6da04c2eea3a7f14ce1c75ee3b0eae7676"}, ] [package.dependencies] -matplotlib = ">=3.4" +matplotlib = ">=3.5" numpy = ">=1.21" packaging = ">=20" -pyproj = ">=3.1.0" -pyshp = ">=2.1" +pyproj = ">=3.3.1" +pyshp = ">=2.3" shapely = ">=1.7" [package.extras] -doc = ["beautifulsoup4", "pydata-sphinx-theme", "sphinx", "sphinx-gallery"] +doc = ["pydata-sphinx-theme", "sphinx", "sphinx-gallery"] ows = ["OWSLib (>=0.20.0)", "pillow (>=6.1.0)"] plotting = ["pillow (>=6.1.0)", "scipy (>=1.3.1)"] speedups = ["fiona", "pykdtree"] +srtm = ["beautifulsoup4"] test = ["coveralls", "pytest (>=5.1.2)", "pytest-cov", "pytest-mpl (>=0.11)", "pytest-xdist"] [[package]] @@ -668,35 +681,32 @@ files = [ [[package]] name = "colorcet" -version = "3.0.1" +version = "3.1.0" description = "Collection of perceptually uniform colormaps" optional = false -python-versions = ">=2.7" +python-versions = ">=3.7" files = [ - {file = "colorcet-3.0.1-py2.py3-none-any.whl", hash = "sha256:8daff01824ee9935fdf762d15c444a67d3e361ad4f8b738ad59ac9bf38f30600"}, - {file = "colorcet-3.0.1.tar.gz", hash = "sha256:51455a20353d12fac91f953772d8409f2474e6a0db1af3fa4f7005f405a2480b"}, + {file = "colorcet-3.1.0-py3-none-any.whl", hash = "sha256:2a7d59cc8d0f7938eeedd08aad3152b5319b4ba3bcb7a612398cc17a384cb296"}, + {file = "colorcet-3.1.0.tar.gz", hash = "sha256:2921b3cd81a2288aaf2d63dbc0ce3c26dcd882e8c389cc505d6886bf7aa9a4eb"}, ] -[package.dependencies] -pyct = ">=0.4.4" - [package.extras] -all = ["bokeh", "flake8", "holoviews", "matplotlib", "nbsite (>=0.7.2rc10)", "nbsmoke (>=0.2.6)", "numpy", "pyct (>=0.4.4)", "pydata-sphinx-theme (<0.9.0)", "pytest (>=2.8.5)", "pytest-cov", "pytest-mpl", "setuptools (>=30.3.0)", "sphinx-copybutton", "wheel"] -build = ["pyct (>=0.4.4)", "setuptools (>=30.3.0)", "wheel"] -doc = ["bokeh", "holoviews", "matplotlib", "nbsite (>=0.7.2rc10)", "numpy", "pydata-sphinx-theme (<0.9.0)", "sphinx-copybutton"] +all = ["colorcet[doc]", "colorcet[examples]", "colorcet[tests-extra]", "colorcet[tests]"] +doc = ["colorcet[examples]", "nbsite (>=0.8.4)", "sphinx-copybutton"] examples = ["bokeh", "holoviews", "matplotlib", "numpy"] -tests = ["flake8", "nbsmoke (>=0.2.6)", "pytest (>=2.8.5)", "pytest-cov"] -tests-extra = ["flake8", "nbsmoke (>=0.2.6)", "pytest (>=2.8.5)", "pytest-cov", "pytest-mpl"] +tests = ["packaging", "pre-commit", "pytest (>=2.8.5)", "pytest-cov"] +tests-examples = ["colorcet[examples]", "nbval"] +tests-extra = ["colorcet[tests]", "pytest-mpl"] [[package]] name = "comm" -version = "0.2.1" +version = "0.2.2" description = "Jupyter Python Comm implementation, for usage in ipykernel, xeus-python etc." optional = false python-versions = ">=3.8" files = [ - {file = "comm-0.2.1-py3-none-any.whl", hash = "sha256:87928485c0dfc0e7976fd89fc1e187023cf587e7c353e4a9b417555b44adf021"}, - {file = "comm-0.2.1.tar.gz", hash = "sha256:0bc91edae1344d39d3661dcbc36937181fdaddb304790458f8b044dbc064b89a"}, + {file = "comm-0.2.2-py3-none-any.whl", hash = "sha256:e6fb86cb70ff661ee8c9c14e7d36d6de3b4066f1441be4063df9c5009f0a64d3"}, + {file = "comm-0.2.2.tar.gz", hash = "sha256:3fd7a84065306e07bea1773df6eb8282de51ba82f77c72f9c85716ab11fe980e"}, ] [package.dependencies] @@ -707,64 +717,64 @@ test = ["pytest"] [[package]] name = "contourpy" -version = "1.2.0" +version = "1.2.1" description = "Python library for calculating contours of 2D quadrilateral grids" optional = false python-versions = ">=3.9" files = [ - {file = "contourpy-1.2.0-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:0274c1cb63625972c0c007ab14dd9ba9e199c36ae1a231ce45d725cbcbfd10a8"}, - {file = "contourpy-1.2.0-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:ab459a1cbbf18e8698399c595a01f6dcc5c138220ca3ea9e7e6126232d102bb4"}, - {file = "contourpy-1.2.0-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:6fdd887f17c2f4572ce548461e4f96396681212d858cae7bd52ba3310bc6f00f"}, - {file = "contourpy-1.2.0-cp310-cp310-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:5d16edfc3fc09968e09ddffada434b3bf989bf4911535e04eada58469873e28e"}, - {file = "contourpy-1.2.0-cp310-cp310-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:1c203f617abc0dde5792beb586f827021069fb6d403d7f4d5c2b543d87edceb9"}, - {file = "contourpy-1.2.0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:b69303ceb2e4d4f146bf82fda78891ef7bcd80c41bf16bfca3d0d7eb545448aa"}, - {file = "contourpy-1.2.0-cp310-cp310-musllinux_1_1_aarch64.whl", hash = "sha256:884c3f9d42d7218304bc74a8a7693d172685c84bd7ab2bab1ee567b769696df9"}, - {file = "contourpy-1.2.0-cp310-cp310-musllinux_1_1_x86_64.whl", hash = "sha256:4a1b1208102be6e851f20066bf0e7a96b7d48a07c9b0cfe6d0d4545c2f6cadab"}, - {file = "contourpy-1.2.0-cp310-cp310-win32.whl", hash = "sha256:34b9071c040d6fe45d9826cbbe3727d20d83f1b6110d219b83eb0e2a01d79488"}, - {file = "contourpy-1.2.0-cp310-cp310-win_amd64.whl", hash = "sha256:bd2f1ae63998da104f16a8b788f685e55d65760cd1929518fd94cd682bf03e41"}, - {file = "contourpy-1.2.0-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:dd10c26b4eadae44783c45ad6655220426f971c61d9b239e6f7b16d5cdaaa727"}, - {file = "contourpy-1.2.0-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:5c6b28956b7b232ae801406e529ad7b350d3f09a4fde958dfdf3c0520cdde0dd"}, - {file = "contourpy-1.2.0-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:ebeac59e9e1eb4b84940d076d9f9a6cec0064e241818bcb6e32124cc5c3e377a"}, - {file = "contourpy-1.2.0-cp311-cp311-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:139d8d2e1c1dd52d78682f505e980f592ba53c9f73bd6be102233e358b401063"}, - {file = "contourpy-1.2.0-cp311-cp311-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:1e9dc350fb4c58adc64df3e0703ab076f60aac06e67d48b3848c23647ae4310e"}, - {file = "contourpy-1.2.0-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:18fc2b4ed8e4a8fe849d18dce4bd3c7ea637758c6343a1f2bae1e9bd4c9f4686"}, - {file = "contourpy-1.2.0-cp311-cp311-musllinux_1_1_aarch64.whl", hash = "sha256:16a7380e943a6d52472096cb7ad5264ecee36ed60888e2a3d3814991a0107286"}, - {file = "contourpy-1.2.0-cp311-cp311-musllinux_1_1_x86_64.whl", hash = "sha256:8d8faf05be5ec8e02a4d86f616fc2a0322ff4a4ce26c0f09d9f7fb5330a35c95"}, - {file = "contourpy-1.2.0-cp311-cp311-win32.whl", hash = "sha256:67b7f17679fa62ec82b7e3e611c43a016b887bd64fb933b3ae8638583006c6d6"}, - {file = "contourpy-1.2.0-cp311-cp311-win_amd64.whl", hash = "sha256:99ad97258985328b4f207a5e777c1b44a83bfe7cf1f87b99f9c11d4ee477c4de"}, - {file = "contourpy-1.2.0-cp312-cp312-macosx_10_9_x86_64.whl", hash = "sha256:575bcaf957a25d1194903a10bc9f316c136c19f24e0985a2b9b5608bdf5dbfe0"}, - {file = "contourpy-1.2.0-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:9e6c93b5b2dbcedad20a2f18ec22cae47da0d705d454308063421a3b290d9ea4"}, - {file = "contourpy-1.2.0-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:464b423bc2a009088f19bdf1f232299e8b6917963e2b7e1d277da5041f33a779"}, - {file = "contourpy-1.2.0-cp312-cp312-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:68ce4788b7d93e47f84edd3f1f95acdcd142ae60bc0e5493bfd120683d2d4316"}, - {file = "contourpy-1.2.0-cp312-cp312-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:3d7d1f8871998cdff5d2ff6a087e5e1780139abe2838e85b0b46b7ae6cc25399"}, - {file = "contourpy-1.2.0-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:6e739530c662a8d6d42c37c2ed52a6f0932c2d4a3e8c1f90692ad0ce1274abe0"}, - {file = "contourpy-1.2.0-cp312-cp312-musllinux_1_1_aarch64.whl", hash = "sha256:247b9d16535acaa766d03037d8e8fb20866d054d3c7fbf6fd1f993f11fc60ca0"}, - {file = "contourpy-1.2.0-cp312-cp312-musllinux_1_1_x86_64.whl", hash = "sha256:461e3ae84cd90b30f8d533f07d87c00379644205b1d33a5ea03381edc4b69431"}, - {file = "contourpy-1.2.0-cp312-cp312-win32.whl", hash = "sha256:1c2559d6cffc94890b0529ea7eeecc20d6fadc1539273aa27faf503eb4656d8f"}, - {file = "contourpy-1.2.0-cp312-cp312-win_amd64.whl", hash = "sha256:491b1917afdd8638a05b611a56d46587d5a632cabead889a5440f7c638bc6ed9"}, - {file = "contourpy-1.2.0-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:5fd1810973a375ca0e097dee059c407913ba35723b111df75671a1976efa04bc"}, - {file = "contourpy-1.2.0-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:999c71939aad2780f003979b25ac5b8f2df651dac7b38fb8ce6c46ba5abe6ae9"}, - {file = "contourpy-1.2.0-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:b7caf9b241464c404613512d5594a6e2ff0cc9cb5615c9475cc1d9b514218ae8"}, - {file = "contourpy-1.2.0-cp39-cp39-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:266270c6f6608340f6c9836a0fb9b367be61dde0c9a9a18d5ece97774105ff3e"}, - {file = "contourpy-1.2.0-cp39-cp39-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:dbd50d0a0539ae2e96e537553aff6d02c10ed165ef40c65b0e27e744a0f10af8"}, - {file = "contourpy-1.2.0-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:11f8d2554e52f459918f7b8e6aa20ec2a3bce35ce95c1f0ef4ba36fbda306df5"}, - {file = "contourpy-1.2.0-cp39-cp39-musllinux_1_1_aarch64.whl", hash = "sha256:ce96dd400486e80ac7d195b2d800b03e3e6a787e2a522bfb83755938465a819e"}, - {file = "contourpy-1.2.0-cp39-cp39-musllinux_1_1_x86_64.whl", hash = "sha256:6d3364b999c62f539cd403f8123ae426da946e142312a514162adb2addd8d808"}, - {file = "contourpy-1.2.0-cp39-cp39-win32.whl", hash = "sha256:1c88dfb9e0c77612febebb6ac69d44a8d81e3dc60f993215425b62c1161353f4"}, - {file = "contourpy-1.2.0-cp39-cp39-win_amd64.whl", hash = "sha256:78e6ad33cf2e2e80c5dfaaa0beec3d61face0fb650557100ee36db808bfa6843"}, - {file = "contourpy-1.2.0-pp39-pypy39_pp73-macosx_10_9_x86_64.whl", hash = "sha256:be16975d94c320432657ad2402f6760990cb640c161ae6da1363051805fa8108"}, - {file = "contourpy-1.2.0-pp39-pypy39_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:b95a225d4948b26a28c08307a60ac00fb8671b14f2047fc5476613252a129776"}, - {file = "contourpy-1.2.0-pp39-pypy39_pp73-win_amd64.whl", hash = "sha256:0d7e03c0f9a4f90dc18d4e77e9ef4ec7b7bbb437f7f675be8e530d65ae6ef956"}, - {file = "contourpy-1.2.0.tar.gz", hash = "sha256:171f311cb758de7da13fc53af221ae47a5877be5a0843a9fe150818c51ed276a"}, + {file = "contourpy-1.2.1-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:bd7c23df857d488f418439686d3b10ae2fbf9bc256cd045b37a8c16575ea1040"}, + {file = "contourpy-1.2.1-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:5b9eb0ca724a241683c9685a484da9d35c872fd42756574a7cfbf58af26677fd"}, + {file = "contourpy-1.2.1-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:4c75507d0a55378240f781599c30e7776674dbaf883a46d1c90f37e563453480"}, + {file = "contourpy-1.2.1-cp310-cp310-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:11959f0ce4a6f7b76ec578576a0b61a28bdc0696194b6347ba3f1c53827178b9"}, + {file = "contourpy-1.2.1-cp310-cp310-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:eb3315a8a236ee19b6df481fc5f997436e8ade24a9f03dfdc6bd490fea20c6da"}, + {file = "contourpy-1.2.1-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:39f3ecaf76cd98e802f094e0d4fbc6dc9c45a8d0c4d185f0f6c2234e14e5f75b"}, + {file = "contourpy-1.2.1-cp310-cp310-musllinux_1_1_aarch64.whl", hash = "sha256:94b34f32646ca0414237168d68a9157cb3889f06b096612afdd296003fdd32fd"}, + {file = "contourpy-1.2.1-cp310-cp310-musllinux_1_1_x86_64.whl", hash = "sha256:457499c79fa84593f22454bbd27670227874cd2ff5d6c84e60575c8b50a69619"}, + {file = "contourpy-1.2.1-cp310-cp310-win32.whl", hash = "sha256:ac58bdee53cbeba2ecad824fa8159493f0bf3b8ea4e93feb06c9a465d6c87da8"}, + {file = "contourpy-1.2.1-cp310-cp310-win_amd64.whl", hash = "sha256:9cffe0f850e89d7c0012a1fb8730f75edd4320a0a731ed0c183904fe6ecfc3a9"}, + {file = "contourpy-1.2.1-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:6022cecf8f44e36af10bd9118ca71f371078b4c168b6e0fab43d4a889985dbb5"}, + {file = "contourpy-1.2.1-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:ef5adb9a3b1d0c645ff694f9bca7702ec2c70f4d734f9922ea34de02294fdf72"}, + {file = "contourpy-1.2.1-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:6150ffa5c767bc6332df27157d95442c379b7dce3a38dff89c0f39b63275696f"}, + {file = "contourpy-1.2.1-cp311-cp311-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:4c863140fafc615c14a4bf4efd0f4425c02230eb8ef02784c9a156461e62c965"}, + {file = "contourpy-1.2.1-cp311-cp311-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:00e5388f71c1a0610e6fe56b5c44ab7ba14165cdd6d695429c5cd94021e390b2"}, + {file = "contourpy-1.2.1-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:d4492d82b3bc7fbb7e3610747b159869468079fe149ec5c4d771fa1f614a14df"}, + {file = "contourpy-1.2.1-cp311-cp311-musllinux_1_1_aarch64.whl", hash = "sha256:49e70d111fee47284d9dd867c9bb9a7058a3c617274900780c43e38d90fe1205"}, + {file = "contourpy-1.2.1-cp311-cp311-musllinux_1_1_x86_64.whl", hash = "sha256:b59c0ffceff8d4d3996a45f2bb6f4c207f94684a96bf3d9728dbb77428dd8cb8"}, + {file = "contourpy-1.2.1-cp311-cp311-win32.whl", hash = "sha256:7b4182299f251060996af5249c286bae9361fa8c6a9cda5efc29fe8bfd6062ec"}, + {file = "contourpy-1.2.1-cp311-cp311-win_amd64.whl", hash = "sha256:2855c8b0b55958265e8b5888d6a615ba02883b225f2227461aa9127c578a4922"}, + {file = "contourpy-1.2.1-cp312-cp312-macosx_10_9_x86_64.whl", hash = "sha256:62828cada4a2b850dbef89c81f5a33741898b305db244904de418cc957ff05dc"}, + {file = "contourpy-1.2.1-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:309be79c0a354afff9ff7da4aaed7c3257e77edf6c1b448a779329431ee79d7e"}, + {file = "contourpy-1.2.1-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:2e785e0f2ef0d567099b9ff92cbfb958d71c2d5b9259981cd9bee81bd194c9a4"}, + {file = "contourpy-1.2.1-cp312-cp312-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:1cac0a8f71a041aa587410424ad46dfa6a11f6149ceb219ce7dd48f6b02b87a7"}, + {file = "contourpy-1.2.1-cp312-cp312-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:af3f4485884750dddd9c25cb7e3915d83c2db92488b38ccb77dd594eac84c4a0"}, + {file = "contourpy-1.2.1-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:9ce6889abac9a42afd07a562c2d6d4b2b7134f83f18571d859b25624a331c90b"}, + {file = "contourpy-1.2.1-cp312-cp312-musllinux_1_1_aarch64.whl", hash = "sha256:a1eea9aecf761c661d096d39ed9026574de8adb2ae1c5bd7b33558af884fb2ce"}, + {file = "contourpy-1.2.1-cp312-cp312-musllinux_1_1_x86_64.whl", hash = "sha256:187fa1d4c6acc06adb0fae5544c59898ad781409e61a926ac7e84b8f276dcef4"}, + {file = "contourpy-1.2.1-cp312-cp312-win32.whl", hash = "sha256:c2528d60e398c7c4c799d56f907664673a807635b857df18f7ae64d3e6ce2d9f"}, + {file = "contourpy-1.2.1-cp312-cp312-win_amd64.whl", hash = "sha256:1a07fc092a4088ee952ddae19a2b2a85757b923217b7eed584fdf25f53a6e7ce"}, + {file = "contourpy-1.2.1-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:bb6834cbd983b19f06908b45bfc2dad6ac9479ae04abe923a275b5f48f1a186b"}, + {file = "contourpy-1.2.1-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:1d59e739ab0e3520e62a26c60707cc3ab0365d2f8fecea74bfe4de72dc56388f"}, + {file = "contourpy-1.2.1-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:bd3db01f59fdcbce5b22afad19e390260d6d0222f35a1023d9adc5690a889364"}, + {file = "contourpy-1.2.1-cp39-cp39-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:a12a813949e5066148712a0626895c26b2578874e4cc63160bb007e6df3436fe"}, + {file = "contourpy-1.2.1-cp39-cp39-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:fe0ccca550bb8e5abc22f530ec0466136379c01321fd94f30a22231e8a48d985"}, + {file = "contourpy-1.2.1-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:e1d59258c3c67c865435d8fbeb35f8c59b8bef3d6f46c1f29f6123556af28445"}, + {file = "contourpy-1.2.1-cp39-cp39-musllinux_1_1_aarch64.whl", hash = "sha256:f32c38afb74bd98ce26de7cc74a67b40afb7b05aae7b42924ea990d51e4dac02"}, + {file = "contourpy-1.2.1-cp39-cp39-musllinux_1_1_x86_64.whl", hash = "sha256:d31a63bc6e6d87f77d71e1abbd7387ab817a66733734883d1fc0021ed9bfa083"}, + {file = "contourpy-1.2.1-cp39-cp39-win32.whl", hash = "sha256:ddcb8581510311e13421b1f544403c16e901c4e8f09083c881fab2be80ee31ba"}, + {file = "contourpy-1.2.1-cp39-cp39-win_amd64.whl", hash = "sha256:10a37ae557aabf2509c79715cd20b62e4c7c28b8cd62dd7d99e5ed3ce28c3fd9"}, + {file = "contourpy-1.2.1-pp39-pypy39_pp73-macosx_10_9_x86_64.whl", hash = "sha256:a31f94983fecbac95e58388210427d68cd30fe8a36927980fab9c20062645609"}, + {file = "contourpy-1.2.1-pp39-pypy39_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:ef2b055471c0eb466033760a521efb9d8a32b99ab907fc8358481a1dd29e3bd3"}, + {file = "contourpy-1.2.1-pp39-pypy39_pp73-win_amd64.whl", hash = "sha256:b33d2bc4f69caedcd0a275329eb2198f560b325605810895627be5d4b876bf7f"}, + {file = "contourpy-1.2.1.tar.gz", hash = "sha256:4d8908b3bee1c889e547867ca4cdc54e5ab6be6d3e078556814a22457f49423c"}, ] [package.dependencies] -numpy = ">=1.20,<2.0" +numpy = ">=1.20" [package.extras] bokeh = ["bokeh", "selenium"] docs = ["furo", "sphinx (>=7.2)", "sphinx-copybutton"] -mypy = ["contourpy[bokeh,docs]", "docutils-stubs", "mypy (==1.6.1)", "types-Pillow"] +mypy = ["contourpy[bokeh,docs]", "docutils-stubs", "mypy (==1.8.0)", "types-Pillow"] test = ["Pillow", "contourpy[test-no-images]", "matplotlib"] test-no-images = ["pytest", "pytest-cov", "pytest-xdist", "wurlitzer"] @@ -784,63 +794,63 @@ coverage = ">=6.0.2" [[package]] name = "coverage" -version = "7.4.1" +version = "7.5.1" description = "Code coverage measurement for Python" optional = false python-versions = ">=3.8" files = [ - {file = "coverage-7.4.1-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:077d366e724f24fc02dbfe9d946534357fda71af9764ff99d73c3c596001bbd7"}, - {file = "coverage-7.4.1-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:0193657651f5399d433c92f8ae264aff31fc1d066deee4b831549526433f3f61"}, - {file = "coverage-7.4.1-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:d17bbc946f52ca67adf72a5ee783cd7cd3477f8f8796f59b4974a9b59cacc9ee"}, - {file = "coverage-7.4.1-cp310-cp310-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:a3277f5fa7483c927fe3a7b017b39351610265308f5267ac6d4c2b64cc1d8d25"}, - {file = "coverage-7.4.1-cp310-cp310-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:6dceb61d40cbfcf45f51e59933c784a50846dc03211054bd76b421a713dcdf19"}, - {file = "coverage-7.4.1-cp310-cp310-musllinux_1_1_aarch64.whl", hash = "sha256:6008adeca04a445ea6ef31b2cbaf1d01d02986047606f7da266629afee982630"}, - {file = "coverage-7.4.1-cp310-cp310-musllinux_1_1_i686.whl", hash = "sha256:c61f66d93d712f6e03369b6a7769233bfda880b12f417eefdd4f16d1deb2fc4c"}, - {file = "coverage-7.4.1-cp310-cp310-musllinux_1_1_x86_64.whl", hash = "sha256:b9bb62fac84d5f2ff523304e59e5c439955fb3b7f44e3d7b2085184db74d733b"}, - {file = "coverage-7.4.1-cp310-cp310-win32.whl", hash = "sha256:f86f368e1c7ce897bf2457b9eb61169a44e2ef797099fb5728482b8d69f3f016"}, - {file = "coverage-7.4.1-cp310-cp310-win_amd64.whl", hash = "sha256:869b5046d41abfea3e381dd143407b0d29b8282a904a19cb908fa24d090cc018"}, - {file = "coverage-7.4.1-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:b8ffb498a83d7e0305968289441914154fb0ef5d8b3157df02a90c6695978295"}, - {file = "coverage-7.4.1-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:3cacfaefe6089d477264001f90f55b7881ba615953414999c46cc9713ff93c8c"}, - {file = "coverage-7.4.1-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:5d6850e6e36e332d5511a48a251790ddc545e16e8beaf046c03985c69ccb2676"}, - {file = "coverage-7.4.1-cp311-cp311-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:18e961aa13b6d47f758cc5879383d27b5b3f3dcd9ce8cdbfdc2571fe86feb4dd"}, - {file = "coverage-7.4.1-cp311-cp311-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:dfd1e1b9f0898817babf840b77ce9fe655ecbe8b1b327983df485b30df8cc011"}, - {file = "coverage-7.4.1-cp311-cp311-musllinux_1_1_aarch64.whl", hash = "sha256:6b00e21f86598b6330f0019b40fb397e705135040dbedc2ca9a93c7441178e74"}, - {file = "coverage-7.4.1-cp311-cp311-musllinux_1_1_i686.whl", hash = "sha256:536d609c6963c50055bab766d9951b6c394759190d03311f3e9fcf194ca909e1"}, - {file = "coverage-7.4.1-cp311-cp311-musllinux_1_1_x86_64.whl", hash = "sha256:7ac8f8eb153724f84885a1374999b7e45734bf93a87d8df1e7ce2146860edef6"}, - {file = "coverage-7.4.1-cp311-cp311-win32.whl", hash = "sha256:f3771b23bb3675a06f5d885c3630b1d01ea6cac9e84a01aaf5508706dba546c5"}, - {file = "coverage-7.4.1-cp311-cp311-win_amd64.whl", hash = "sha256:9d2f9d4cc2a53b38cabc2d6d80f7f9b7e3da26b2f53d48f05876fef7956b6968"}, - {file = "coverage-7.4.1-cp312-cp312-macosx_10_9_x86_64.whl", hash = "sha256:f68ef3660677e6624c8cace943e4765545f8191313a07288a53d3da188bd8581"}, - {file = "coverage-7.4.1-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:23b27b8a698e749b61809fb637eb98ebf0e505710ec46a8aa6f1be7dc0dc43a6"}, - {file = "coverage-7.4.1-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:3e3424c554391dc9ef4a92ad28665756566a28fecf47308f91841f6c49288e66"}, - {file = "coverage-7.4.1-cp312-cp312-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:e0860a348bf7004c812c8368d1fc7f77fe8e4c095d661a579196a9533778e156"}, - {file = "coverage-7.4.1-cp312-cp312-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:fe558371c1bdf3b8fa03e097c523fb9645b8730399c14fe7721ee9c9e2a545d3"}, - {file = "coverage-7.4.1-cp312-cp312-musllinux_1_1_aarch64.whl", hash = "sha256:3468cc8720402af37b6c6e7e2a9cdb9f6c16c728638a2ebc768ba1ef6f26c3a1"}, - {file = "coverage-7.4.1-cp312-cp312-musllinux_1_1_i686.whl", hash = "sha256:02f2edb575d62172aa28fe00efe821ae31f25dc3d589055b3fb64d51e52e4ab1"}, - {file = "coverage-7.4.1-cp312-cp312-musllinux_1_1_x86_64.whl", hash = "sha256:ca6e61dc52f601d1d224526360cdeab0d0712ec104a2ce6cc5ccef6ed9a233bc"}, - {file = "coverage-7.4.1-cp312-cp312-win32.whl", hash = "sha256:ca7b26a5e456a843b9b6683eada193fc1f65c761b3a473941efe5a291f604c74"}, - {file = "coverage-7.4.1-cp312-cp312-win_amd64.whl", hash = "sha256:85ccc5fa54c2ed64bd91ed3b4a627b9cce04646a659512a051fa82a92c04a448"}, - {file = "coverage-7.4.1-cp38-cp38-macosx_10_9_x86_64.whl", hash = "sha256:8bdb0285a0202888d19ec6b6d23d5990410decb932b709f2b0dfe216d031d218"}, - {file = "coverage-7.4.1-cp38-cp38-macosx_11_0_arm64.whl", hash = "sha256:918440dea04521f499721c039863ef95433314b1db00ff826a02580c1f503e45"}, - {file = "coverage-7.4.1-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:379d4c7abad5afbe9d88cc31ea8ca262296480a86af945b08214eb1a556a3e4d"}, - {file = "coverage-7.4.1-cp38-cp38-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:b094116f0b6155e36a304ff912f89bbb5067157aff5f94060ff20bbabdc8da06"}, - {file = "coverage-7.4.1-cp38-cp38-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:f2f5968608b1fe2a1d00d01ad1017ee27efd99b3437e08b83ded9b7af3f6f766"}, - {file = "coverage-7.4.1-cp38-cp38-musllinux_1_1_aarch64.whl", hash = "sha256:10e88e7f41e6197ea0429ae18f21ff521d4f4490aa33048f6c6f94c6045a6a75"}, - {file = "coverage-7.4.1-cp38-cp38-musllinux_1_1_i686.whl", hash = "sha256:a4a3907011d39dbc3e37bdc5df0a8c93853c369039b59efa33a7b6669de04c60"}, - {file = "coverage-7.4.1-cp38-cp38-musllinux_1_1_x86_64.whl", hash = "sha256:6d224f0c4c9c98290a6990259073f496fcec1b5cc613eecbd22786d398ded3ad"}, - {file = "coverage-7.4.1-cp38-cp38-win32.whl", hash = "sha256:23f5881362dcb0e1a92b84b3c2809bdc90db892332daab81ad8f642d8ed55042"}, - {file = "coverage-7.4.1-cp38-cp38-win_amd64.whl", hash = "sha256:a07f61fc452c43cd5328b392e52555f7d1952400a1ad09086c4a8addccbd138d"}, - {file = "coverage-7.4.1-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:8e738a492b6221f8dcf281b67129510835461132b03024830ac0e554311a5c54"}, - {file = "coverage-7.4.1-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:46342fed0fff72efcda77040b14728049200cbba1279e0bf1188f1f2078c1d70"}, - {file = "coverage-7.4.1-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:9641e21670c68c7e57d2053ddf6c443e4f0a6e18e547e86af3fad0795414a628"}, - {file = "coverage-7.4.1-cp39-cp39-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:aeb2c2688ed93b027eb0d26aa188ada34acb22dceea256d76390eea135083950"}, - {file = "coverage-7.4.1-cp39-cp39-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:d12c923757de24e4e2110cf8832d83a886a4cf215c6e61ed506006872b43a6d1"}, - {file = "coverage-7.4.1-cp39-cp39-musllinux_1_1_aarch64.whl", hash = "sha256:0491275c3b9971cdbd28a4595c2cb5838f08036bca31765bad5e17edf900b2c7"}, - {file = "coverage-7.4.1-cp39-cp39-musllinux_1_1_i686.whl", hash = "sha256:8dfc5e195bbef80aabd81596ef52a1277ee7143fe419efc3c4d8ba2754671756"}, - {file = "coverage-7.4.1-cp39-cp39-musllinux_1_1_x86_64.whl", hash = "sha256:1a78b656a4d12b0490ca72651fe4d9f5e07e3c6461063a9b6265ee45eb2bdd35"}, - {file = "coverage-7.4.1-cp39-cp39-win32.whl", hash = "sha256:f90515974b39f4dea2f27c0959688621b46d96d5a626cf9c53dbc653a895c05c"}, - {file = "coverage-7.4.1-cp39-cp39-win_amd64.whl", hash = "sha256:64e723ca82a84053dd7bfcc986bdb34af8d9da83c521c19d6b472bc6880e191a"}, - {file = "coverage-7.4.1-pp38.pp39.pp310-none-any.whl", hash = "sha256:32a8d985462e37cfdab611a6f95b09d7c091d07668fdc26e47a725ee575fe166"}, - {file = "coverage-7.4.1.tar.gz", hash = "sha256:1ed4b95480952b1a26d863e546fa5094564aa0065e1e5f0d4d0041f293251d04"}, + {file = "coverage-7.5.1-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:c0884920835a033b78d1c73b6d3bbcda8161a900f38a488829a83982925f6c2e"}, + {file = "coverage-7.5.1-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:39afcd3d4339329c5f58de48a52f6e4e50f6578dd6099961cf22228feb25f38f"}, + {file = "coverage-7.5.1-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:4a7b0ceee8147444347da6a66be737c9d78f3353b0681715b668b72e79203e4a"}, + {file = "coverage-7.5.1-cp310-cp310-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:4a9ca3f2fae0088c3c71d743d85404cec8df9be818a005ea065495bedc33da35"}, + {file = "coverage-7.5.1-cp310-cp310-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:5fd215c0c7d7aab005221608a3c2b46f58c0285a819565887ee0b718c052aa4e"}, + {file = "coverage-7.5.1-cp310-cp310-musllinux_1_1_aarch64.whl", hash = "sha256:4bf0655ab60d754491004a5efd7f9cccefcc1081a74c9ef2da4735d6ee4a6223"}, + {file = "coverage-7.5.1-cp310-cp310-musllinux_1_1_i686.whl", hash = "sha256:61c4bf1ba021817de12b813338c9be9f0ad5b1e781b9b340a6d29fc13e7c1b5e"}, + {file = "coverage-7.5.1-cp310-cp310-musllinux_1_1_x86_64.whl", hash = "sha256:db66fc317a046556a96b453a58eced5024af4582a8dbdc0c23ca4dbc0d5b3146"}, + {file = "coverage-7.5.1-cp310-cp310-win32.whl", hash = "sha256:b016ea6b959d3b9556cb401c55a37547135a587db0115635a443b2ce8f1c7228"}, + {file = "coverage-7.5.1-cp310-cp310-win_amd64.whl", hash = "sha256:df4e745a81c110e7446b1cc8131bf986157770fa405fe90e15e850aaf7619bc8"}, + {file = "coverage-7.5.1-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:796a79f63eca8814ca3317a1ea443645c9ff0d18b188de470ed7ccd45ae79428"}, + {file = "coverage-7.5.1-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:4fc84a37bfd98db31beae3c2748811a3fa72bf2007ff7902f68746d9757f3746"}, + {file = "coverage-7.5.1-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:6175d1a0559986c6ee3f7fccfc4a90ecd12ba0a383dcc2da30c2b9918d67d8a3"}, + {file = "coverage-7.5.1-cp311-cp311-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:1fc81d5878cd6274ce971e0a3a18a8803c3fe25457165314271cf78e3aae3aa2"}, + {file = "coverage-7.5.1-cp311-cp311-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:556cf1a7cbc8028cb60e1ff0be806be2eded2daf8129b8811c63e2b9a6c43bca"}, + {file = "coverage-7.5.1-cp311-cp311-musllinux_1_1_aarch64.whl", hash = "sha256:9981706d300c18d8b220995ad22627647be11a4276721c10911e0e9fa44c83e8"}, + {file = "coverage-7.5.1-cp311-cp311-musllinux_1_1_i686.whl", hash = "sha256:d7fed867ee50edf1a0b4a11e8e5d0895150e572af1cd6d315d557758bfa9c057"}, + {file = "coverage-7.5.1-cp311-cp311-musllinux_1_1_x86_64.whl", hash = "sha256:ef48e2707fb320c8f139424a596f5b69955a85b178f15af261bab871873bb987"}, + {file = "coverage-7.5.1-cp311-cp311-win32.whl", hash = "sha256:9314d5678dcc665330df5b69c1e726a0e49b27df0461c08ca12674bcc19ef136"}, + {file = "coverage-7.5.1-cp311-cp311-win_amd64.whl", hash = "sha256:5fa567e99765fe98f4e7d7394ce623e794d7cabb170f2ca2ac5a4174437e90dd"}, + {file = "coverage-7.5.1-cp312-cp312-macosx_10_9_x86_64.whl", hash = "sha256:b6cf3764c030e5338e7f61f95bd21147963cf6aa16e09d2f74f1fa52013c1206"}, + {file = "coverage-7.5.1-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:2ec92012fefebee89a6b9c79bc39051a6cb3891d562b9270ab10ecfdadbc0c34"}, + {file = "coverage-7.5.1-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:16db7f26000a07efcf6aea00316f6ac57e7d9a96501e990a36f40c965ec7a95d"}, + {file = "coverage-7.5.1-cp312-cp312-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:beccf7b8a10b09c4ae543582c1319c6df47d78fd732f854ac68d518ee1fb97fa"}, + {file = "coverage-7.5.1-cp312-cp312-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:8748731ad392d736cc9ccac03c9845b13bb07d020a33423fa5b3a36521ac6e4e"}, + {file = "coverage-7.5.1-cp312-cp312-musllinux_1_1_aarch64.whl", hash = "sha256:7352b9161b33fd0b643ccd1f21f3a3908daaddf414f1c6cb9d3a2fd618bf2572"}, + {file = "coverage-7.5.1-cp312-cp312-musllinux_1_1_i686.whl", hash = "sha256:7a588d39e0925f6a2bff87154752481273cdb1736270642aeb3635cb9b4cad07"}, + {file = "coverage-7.5.1-cp312-cp312-musllinux_1_1_x86_64.whl", hash = "sha256:68f962d9b72ce69ea8621f57551b2fa9c70509af757ee3b8105d4f51b92b41a7"}, + {file = "coverage-7.5.1-cp312-cp312-win32.whl", hash = "sha256:f152cbf5b88aaeb836127d920dd0f5e7edff5a66f10c079157306c4343d86c19"}, + {file = "coverage-7.5.1-cp312-cp312-win_amd64.whl", hash = "sha256:5a5740d1fb60ddf268a3811bcd353de34eb56dc24e8f52a7f05ee513b2d4f596"}, + {file = "coverage-7.5.1-cp38-cp38-macosx_10_9_x86_64.whl", hash = "sha256:e2213def81a50519d7cc56ed643c9e93e0247f5bbe0d1247d15fa520814a7cd7"}, + {file = "coverage-7.5.1-cp38-cp38-macosx_11_0_arm64.whl", hash = "sha256:5037f8fcc2a95b1f0e80585bd9d1ec31068a9bcb157d9750a172836e98bc7a90"}, + {file = "coverage-7.5.1-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:5c3721c2c9e4c4953a41a26c14f4cef64330392a6d2d675c8b1db3b645e31f0e"}, + {file = "coverage-7.5.1-cp38-cp38-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:ca498687ca46a62ae590253fba634a1fe9836bc56f626852fb2720f334c9e4e5"}, + {file = "coverage-7.5.1-cp38-cp38-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:0cdcbc320b14c3e5877ee79e649677cb7d89ef588852e9583e6b24c2e5072661"}, + {file = "coverage-7.5.1-cp38-cp38-musllinux_1_1_aarch64.whl", hash = "sha256:57e0204b5b745594e5bc14b9b50006da722827f0b8c776949f1135677e88d0b8"}, + {file = "coverage-7.5.1-cp38-cp38-musllinux_1_1_i686.whl", hash = "sha256:8fe7502616b67b234482c3ce276ff26f39ffe88adca2acf0261df4b8454668b4"}, + {file = "coverage-7.5.1-cp38-cp38-musllinux_1_1_x86_64.whl", hash = "sha256:9e78295f4144f9dacfed4f92935fbe1780021247c2fabf73a819b17f0ccfff8d"}, + {file = "coverage-7.5.1-cp38-cp38-win32.whl", hash = "sha256:1434e088b41594baa71188a17533083eabf5609e8e72f16ce8c186001e6b8c41"}, + {file = "coverage-7.5.1-cp38-cp38-win_amd64.whl", hash = "sha256:0646599e9b139988b63704d704af8e8df7fa4cbc4a1f33df69d97f36cb0a38de"}, + {file = "coverage-7.5.1-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:4cc37def103a2725bc672f84bd939a6fe4522310503207aae4d56351644682f1"}, + {file = "coverage-7.5.1-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:fc0b4d8bfeabd25ea75e94632f5b6e047eef8adaed0c2161ada1e922e7f7cece"}, + {file = "coverage-7.5.1-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:0d0a0f5e06881ecedfe6f3dd2f56dcb057b6dbeb3327fd32d4b12854df36bf26"}, + {file = "coverage-7.5.1-cp39-cp39-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:9735317685ba6ec7e3754798c8871c2f49aa5e687cc794a0b1d284b2389d1bd5"}, + {file = "coverage-7.5.1-cp39-cp39-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:d21918e9ef11edf36764b93101e2ae8cc82aa5efdc7c5a4e9c6c35a48496d601"}, + {file = "coverage-7.5.1-cp39-cp39-musllinux_1_1_aarch64.whl", hash = "sha256:c3e757949f268364b96ca894b4c342b41dc6f8f8b66c37878aacef5930db61be"}, + {file = "coverage-7.5.1-cp39-cp39-musllinux_1_1_i686.whl", hash = "sha256:79afb6197e2f7f60c4824dd4b2d4c2ec5801ceb6ba9ce5d2c3080e5660d51a4f"}, + {file = "coverage-7.5.1-cp39-cp39-musllinux_1_1_x86_64.whl", hash = "sha256:d1d0d98d95dd18fe29dc66808e1accf59f037d5716f86a501fc0256455219668"}, + {file = "coverage-7.5.1-cp39-cp39-win32.whl", hash = "sha256:1cc0fe9b0b3a8364093c53b0b4c0c2dd4bb23acbec4c9240b5f284095ccf7981"}, + {file = "coverage-7.5.1-cp39-cp39-win_amd64.whl", hash = "sha256:dde0070c40ea8bb3641e811c1cfbf18e265d024deff6de52c5950677a8fb1e0f"}, + {file = "coverage-7.5.1-pp38.pp39.pp310-none-any.whl", hash = "sha256:6537e7c10cc47c595828b8a8be04c72144725c383c4702703ff4e42e44577312"}, + {file = "coverage-7.5.1.tar.gz", hash = "sha256:54de9ef3a9da981f7af93eafde4ede199e0846cd819eb27c88e2b712aae9708c"}, ] [package.dependencies] @@ -866,22 +876,23 @@ tests = ["pytest", "pytest-cov", "pytest-xdist"] [[package]] name = "dask" -version = "2024.1.1" +version = "2024.5.1" description = "Parallel PyData with Task Scheduling" optional = false python-versions = ">=3.9" files = [ - {file = "dask-2024.1.1-py3-none-any.whl", hash = "sha256:860ce2797905095beff0187c214840b80c77d752dcb9098a8283e3655a762bf5"}, - {file = "dask-2024.1.1.tar.gz", hash = "sha256:d0dc92e81ce68594a0a0ce23ba33f4d648f2c2f4217ab9b79068b7ecfb0416c7"}, + {file = "dask-2024.5.1-py3-none-any.whl", hash = "sha256:af1cadd1fd1d1d44600ff5de43dd029e5668fdf87422131f4e3e3aa2a6a63555"}, + {file = "dask-2024.5.1.tar.gz", hash = "sha256:e071fda67031c314569e37ca70b3e88bb30f1d91ff8ee4122b541845847cc264"}, ] [package.dependencies] bokeh = {version = ">=2.4.2", optional = true, markers = "extra == \"diagnostics\""} click = ">=8.1" cloudpickle = ">=1.5.0" -distributed = {version = "2024.1.1", optional = true, markers = "extra == \"distributed\""} +dask-expr = {version = ">=1.1,<1.2", optional = true, markers = "extra == \"dataframe\""} +distributed = {version = "2024.5.1", optional = true, markers = "extra == \"distributed\""} fsspec = ">=2021.09.0" -importlib-metadata = ">=4.13.0" +importlib-metadata = {version = ">=4.13.0", markers = "python_version < \"3.12\""} jinja2 = {version = ">=2.10.3", optional = true, markers = "extra == \"diagnostics\""} lz4 = {version = ">=4.3.2", optional = true, markers = "extra == \"complete\""} numpy = {version = ">=1.21", optional = true, markers = "extra == \"array\""} @@ -896,20 +907,36 @@ toolz = ">=0.10.0" [package.extras] array = ["numpy (>=1.21)"] complete = ["dask[array,dataframe,diagnostics,distributed]", "lz4 (>=4.3.2)", "pyarrow (>=7.0)", "pyarrow-hotfix"] -dataframe = ["dask[array]", "pandas (>=1.3)"] +dataframe = ["dask-expr (>=1.1,<1.2)", "dask[array]", "pandas (>=1.3)"] diagnostics = ["bokeh (>=2.4.2)", "jinja2 (>=2.10.3)"] -distributed = ["distributed (==2024.1.1)"] +distributed = ["distributed (==2024.5.1)"] test = ["pandas[test]", "pre-commit", "pytest", "pytest-cov", "pytest-rerunfailures", "pytest-timeout", "pytest-xdist"] +[[package]] +name = "dask-expr" +version = "1.1.1" +description = "High Level Expressions for Dask" +optional = false +python-versions = ">=3.9" +files = [ + {file = "dask_expr-1.1.1-py3-none-any.whl", hash = "sha256:455842bff0795309f75f78f98d9be235210e01ad93da9b5a74f8cdaa594ab958"}, + {file = "dask_expr-1.1.1.tar.gz", hash = "sha256:30acb27c19ef3f44681cddd46597395b2da976d426e050e684e2044ebe033f3a"}, +] + +[package.dependencies] +dask = "2024.5.1" +pandas = ">=2" +pyarrow = ">=7.0.0" + [[package]] name = "datashader" -version = "0.16.0" +version = "0.16.1" description = "Data visualization toolchain based on aggregating into a grid" optional = false python-versions = ">=3.9" files = [ - {file = "datashader-0.16.0-py2.py3-none-any.whl", hash = "sha256:a2cb0f839067bf29cf6cc9c07a1dad35f0e4aed3768387056fcbac8748087bfa"}, - {file = "datashader-0.16.0.tar.gz", hash = "sha256:ed4c111957578dcb3fcff972d954f77586dafd71a7345fd5cd069d9fb050d0d1"}, + {file = "datashader-0.16.1-py2.py3-none-any.whl", hash = "sha256:317c16a32a7d1c5208444a83fd6d2cb68d9bfb81802c323489bd24f6b7cf5c5f"}, + {file = "datashader-0.16.1.tar.gz", hash = "sha256:5b2f1dd448defce206ecc16e9960fe9378f32307ea1fcb0c298c22c14f58c881"}, ] [package.dependencies] @@ -928,39 +955,43 @@ toolz = "*" xarray = "*" [package.extras] -all = ["bokeh (>3.1)", "codecov", "dask-geopandas", "fastparquet", "flake8", "geodatasets", "geopandas", "graphviz", "holoviews", "matplotlib (>=3.3)", "nbconvert", "nbformat", "nbsite (>=0.8.2,<0.9.0)", "nbsmoke[verify] (>0.5)", "netcdf4", "networkx", "numpydoc", "panel (>1.1)", "pyarrow", "pytest", "pytest-benchmark", "pytest-cov", "python-graphviz", "python-snappy", "rasterio", "rioxarray", "scikit-image", "shapely (>=2.0.0)", "spatialpandas", "streamz"] -doc = ["bokeh (>3.1)", "dask-geopandas", "fastparquet", "geodatasets", "geopandas", "graphviz", "holoviews", "matplotlib (>=3.3)", "nbsite (>=0.8.2,<0.9.0)", "networkx", "numpydoc", "panel (>1.1)", "python-graphviz", "python-snappy", "rasterio", "scikit-image", "shapely (>=2.0.0)", "spatialpandas", "streamz"] +all = ["bokeh (>3.1)", "codecov", "dask-expr", "dask-geopandas", "fastparquet", "flake8", "geodatasets", "geopandas", "graphviz", "holoviews", "matplotlib (>=3.3)", "nbconvert", "nbformat", "nbsite (>=0.8.4,<0.9.0)", "nbsmoke[verify] (>0.5)", "netcdf4", "networkx", "numpydoc", "panel (>1.1)", "pyarrow", "pytest (<8)", "pytest-benchmark", "pytest-cov", "python-graphviz", "python-snappy", "rasterio", "rioxarray", "scikit-image", "shapely (>=2.0.0)", "spatialpandas", "streamz"] +doc = ["bokeh (>3.1)", "dask-geopandas", "fastparquet", "geodatasets", "geopandas", "graphviz", "holoviews", "matplotlib (>=3.3)", "nbsite (>=0.8.4,<0.9.0)", "networkx", "numpydoc", "panel (>1.1)", "python-graphviz", "python-snappy", "rasterio", "scikit-image", "shapely (>=2.0.0)", "spatialpandas", "streamz"] examples = ["bokeh (>3.1)", "dask-geopandas", "geodatasets", "geopandas", "holoviews", "matplotlib (>=3.3)", "panel (>1.1)", "scikit-image", "shapely (>=2.0.0)", "spatialpandas"] examples-extra = ["bokeh (>3.1)", "dask-geopandas", "fastparquet", "geodatasets", "geopandas", "graphviz", "holoviews", "matplotlib (>=3.3)", "networkx", "panel (>1.1)", "python-graphviz", "python-snappy", "rasterio", "scikit-image", "shapely (>=2.0.0)", "spatialpandas", "streamz"] geopandas = ["dask-geopandas", "geopandas", "shapely (>=2.0.0)"] gpu-tests = ["cudf", "cupy", "dask-cudf"] -tests = ["codecov", "dask-geopandas", "fastparquet", "flake8", "geodatasets", "geopandas", "nbconvert", "nbformat", "nbsmoke[verify] (>0.5)", "netcdf4", "pyarrow", "pytest", "pytest-benchmark", "pytest-cov", "rasterio", "rioxarray", "scikit-image", "shapely (>=2.0.0)", "spatialpandas"] +tests = ["codecov", "dask-expr", "dask-geopandas", "flake8", "geodatasets", "geopandas", "nbconvert", "nbformat", "nbsmoke[verify] (>0.5)", "netcdf4", "pyarrow", "pytest (<8)", "pytest-benchmark", "pytest-cov", "rasterio", "rioxarray", "scikit-image", "shapely (>=2.0.0)", "spatialpandas"] [[package]] name = "debugpy" -version = "1.8.0" +version = "1.8.1" description = "An implementation of the Debug Adapter Protocol for Python" optional = false python-versions = ">=3.8" files = [ - {file = "debugpy-1.8.0-cp310-cp310-macosx_11_0_x86_64.whl", hash = "sha256:7fb95ca78f7ac43393cd0e0f2b6deda438ec7c5e47fa5d38553340897d2fbdfb"}, - {file = "debugpy-1.8.0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:ef9ab7df0b9a42ed9c878afd3eaaff471fce3fa73df96022e1f5c9f8f8c87ada"}, - {file = "debugpy-1.8.0-cp310-cp310-win32.whl", hash = "sha256:a8b7a2fd27cd9f3553ac112f356ad4ca93338feadd8910277aff71ab24d8775f"}, - {file = "debugpy-1.8.0-cp310-cp310-win_amd64.whl", hash = "sha256:5d9de202f5d42e62f932507ee8b21e30d49aae7e46d5b1dd5c908db1d7068637"}, - {file = "debugpy-1.8.0-cp311-cp311-macosx_11_0_universal2.whl", hash = "sha256:ef54404365fae8d45cf450d0544ee40cefbcb9cb85ea7afe89a963c27028261e"}, - {file = "debugpy-1.8.0-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:60009b132c91951354f54363f8ebdf7457aeb150e84abba5ae251b8e9f29a8a6"}, - {file = "debugpy-1.8.0-cp311-cp311-win32.whl", hash = "sha256:8cd0197141eb9e8a4566794550cfdcdb8b3db0818bdf8c49a8e8f8053e56e38b"}, - {file = "debugpy-1.8.0-cp311-cp311-win_amd64.whl", hash = "sha256:a64093656c4c64dc6a438e11d59369875d200bd5abb8f9b26c1f5f723622e153"}, - {file = "debugpy-1.8.0-cp38-cp38-macosx_11_0_x86_64.whl", hash = "sha256:b05a6b503ed520ad58c8dc682749113d2fd9f41ffd45daec16e558ca884008cd"}, - {file = "debugpy-1.8.0-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:3c6fb41c98ec51dd010d7ed650accfd07a87fe5e93eca9d5f584d0578f28f35f"}, - {file = "debugpy-1.8.0-cp38-cp38-win32.whl", hash = "sha256:46ab6780159eeabb43c1495d9c84cf85d62975e48b6ec21ee10c95767c0590aa"}, - {file = "debugpy-1.8.0-cp38-cp38-win_amd64.whl", hash = "sha256:bdc5ef99d14b9c0fcb35351b4fbfc06ac0ee576aeab6b2511702e5a648a2e595"}, - {file = "debugpy-1.8.0-cp39-cp39-macosx_11_0_x86_64.whl", hash = "sha256:61eab4a4c8b6125d41a34bad4e5fe3d2cc145caecd63c3fe953be4cc53e65bf8"}, - {file = "debugpy-1.8.0-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:125b9a637e013f9faac0a3d6a82bd17c8b5d2c875fb6b7e2772c5aba6d082332"}, - {file = "debugpy-1.8.0-cp39-cp39-win32.whl", hash = "sha256:57161629133113c97b387382045649a2b985a348f0c9366e22217c87b68b73c6"}, - {file = "debugpy-1.8.0-cp39-cp39-win_amd64.whl", hash = "sha256:e3412f9faa9ade82aa64a50b602544efcba848c91384e9f93497a458767e6926"}, - {file = "debugpy-1.8.0-py2.py3-none-any.whl", hash = "sha256:9c9b0ac1ce2a42888199df1a1906e45e6f3c9555497643a85e0bf2406e3ffbc4"}, - {file = "debugpy-1.8.0.zip", hash = "sha256:12af2c55b419521e33d5fb21bd022df0b5eb267c3e178f1d374a63a2a6bdccd0"}, + {file = "debugpy-1.8.1-cp310-cp310-macosx_11_0_x86_64.whl", hash = "sha256:3bda0f1e943d386cc7a0e71bfa59f4137909e2ed947fb3946c506e113000f741"}, + {file = "debugpy-1.8.1-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:dda73bf69ea479c8577a0448f8c707691152e6c4de7f0c4dec5a4bc11dee516e"}, + {file = "debugpy-1.8.1-cp310-cp310-win32.whl", hash = "sha256:3a79c6f62adef994b2dbe9fc2cc9cc3864a23575b6e387339ab739873bea53d0"}, + {file = "debugpy-1.8.1-cp310-cp310-win_amd64.whl", hash = "sha256:7eb7bd2b56ea3bedb009616d9e2f64aab8fc7000d481faec3cd26c98a964bcdd"}, + {file = "debugpy-1.8.1-cp311-cp311-macosx_11_0_universal2.whl", hash = "sha256:016a9fcfc2c6b57f939673c874310d8581d51a0fe0858e7fac4e240c5eb743cb"}, + {file = "debugpy-1.8.1-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:fd97ed11a4c7f6d042d320ce03d83b20c3fb40da892f994bc041bbc415d7a099"}, + {file = "debugpy-1.8.1-cp311-cp311-win32.whl", hash = "sha256:0de56aba8249c28a300bdb0672a9b94785074eb82eb672db66c8144fff673146"}, + {file = "debugpy-1.8.1-cp311-cp311-win_amd64.whl", hash = "sha256:1a9fe0829c2b854757b4fd0a338d93bc17249a3bf69ecf765c61d4c522bb92a8"}, + {file = "debugpy-1.8.1-cp312-cp312-macosx_11_0_universal2.whl", hash = "sha256:3ebb70ba1a6524d19fa7bb122f44b74170c447d5746a503e36adc244a20ac539"}, + {file = "debugpy-1.8.1-cp312-cp312-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:a2e658a9630f27534e63922ebf655a6ab60c370f4d2fc5c02a5b19baf4410ace"}, + {file = "debugpy-1.8.1-cp312-cp312-win32.whl", hash = "sha256:caad2846e21188797a1f17fc09c31b84c7c3c23baf2516fed5b40b378515bbf0"}, + {file = "debugpy-1.8.1-cp312-cp312-win_amd64.whl", hash = "sha256:edcc9f58ec0fd121a25bc950d4578df47428d72e1a0d66c07403b04eb93bcf98"}, + {file = "debugpy-1.8.1-cp38-cp38-macosx_11_0_x86_64.whl", hash = "sha256:7a3afa222f6fd3d9dfecd52729bc2e12c93e22a7491405a0ecbf9e1d32d45b39"}, + {file = "debugpy-1.8.1-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:d915a18f0597ef685e88bb35e5d7ab968964b7befefe1aaea1eb5b2640b586c7"}, + {file = "debugpy-1.8.1-cp38-cp38-win32.whl", hash = "sha256:92116039b5500633cc8d44ecc187abe2dfa9b90f7a82bbf81d079fcdd506bae9"}, + {file = "debugpy-1.8.1-cp38-cp38-win_amd64.whl", hash = "sha256:e38beb7992b5afd9d5244e96ad5fa9135e94993b0c551ceebf3fe1a5d9beb234"}, + {file = "debugpy-1.8.1-cp39-cp39-macosx_11_0_x86_64.whl", hash = "sha256:bfb20cb57486c8e4793d41996652e5a6a885b4d9175dd369045dad59eaacea42"}, + {file = "debugpy-1.8.1-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:efd3fdd3f67a7e576dd869c184c5dd71d9aaa36ded271939da352880c012e703"}, + {file = "debugpy-1.8.1-cp39-cp39-win32.whl", hash = "sha256:58911e8521ca0c785ac7a0539f1e77e0ce2df753f786188f382229278b4cdf23"}, + {file = "debugpy-1.8.1-cp39-cp39-win_amd64.whl", hash = "sha256:6df9aa9599eb05ca179fb0b810282255202a66835c6efb1d112d21ecb830ddd3"}, + {file = "debugpy-1.8.1-py2.py3-none-any.whl", hash = "sha256:28acbe2241222b87e255260c76741e1fbf04fdc3b6d094fcf57b6c6f75ce1242"}, + {file = "debugpy-1.8.1.zip", hash = "sha256:f696d6be15be87aef621917585f9bb94b1dc9e8aced570db1b8a6fc14e8f9b42"}, ] [[package]] @@ -976,19 +1007,19 @@ files = [ [[package]] name = "distributed" -version = "2024.1.1" +version = "2024.5.1" description = "Distributed scheduler for Dask" optional = false python-versions = ">=3.9" files = [ - {file = "distributed-2024.1.1-py3-none-any.whl", hash = "sha256:cf05d3b38e1700339b3e36395729ab62110e723efefaecc21a8260fdc7555cf9"}, - {file = "distributed-2024.1.1.tar.gz", hash = "sha256:28cf5e9f4f07197b03ea8e5272e374ce2b9e9dc6742f6c9b525fd81645213c67"}, + {file = "distributed-2024.5.1-py3-none-any.whl", hash = "sha256:cf77583ce67f6696aa716c30f1cba3403ac866241c00d6e9a0bc1500c6818b40"}, + {file = "distributed-2024.5.1.tar.gz", hash = "sha256:c4e641e5fc014de3b43c584c70f703a7d44557b51b1143db812b8bc861aa84e2"}, ] [package.dependencies] click = ">=8.0" cloudpickle = ">=1.5.0" -dask = "2024.1.1" +dask = "2024.5.1" jinja2 = ">=2.10.3" locket = ">=1.0.0" msgpack = ">=1.0.0" @@ -1025,13 +1056,13 @@ files = [ [[package]] name = "exceptiongroup" -version = "1.2.0" +version = "1.2.1" description = "Backport of PEP 654 (exception groups)" optional = false python-versions = ">=3.7" files = [ - {file = "exceptiongroup-1.2.0-py3-none-any.whl", hash = "sha256:4bfd3996ac73b41e9b9628b04e079f193850720ea5945fc96a08633c66912f14"}, - {file = "exceptiongroup-1.2.0.tar.gz", hash = "sha256:91f5c769735f051a4290d52edd0858999b57e5876e9f85937691bd4c9fa3ed68"}, + {file = "exceptiongroup-1.2.1-py3-none-any.whl", hash = "sha256:5258b9ed329c5bbdd31a309f53cbfb0b155341807f6ff7606a1e801a891b29ad"}, + {file = "exceptiongroup-1.2.1.tar.gz", hash = "sha256:a4785e48b045528f5bfe627b6ad554ff32def154f42372786903b7abcfe1aa16"}, ] [package.extras] @@ -1078,35 +1109,35 @@ devel = ["colorama", "json-spec", "jsonschema", "pylint", "pytest", "pytest-benc [[package]] name = "fiona" -version = "1.9.5" +version = "1.9.6" description = "Fiona reads and writes spatial data files" optional = false python-versions = ">=3.7" files = [ - {file = "fiona-1.9.5-cp310-cp310-macosx_10_15_x86_64.whl", hash = "sha256:5f40a40529ecfca5294260316cf987a0420c77a2f0cf0849f529d1afbccd093e"}, - {file = "fiona-1.9.5-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:374efe749143ecb5cfdd79b585d83917d2bf8ecfbfc6953c819586b336ce9c63"}, - {file = "fiona-1.9.5-cp310-cp310-manylinux2014_x86_64.whl", hash = "sha256:35dae4b0308eb44617cdc4461ceb91f891d944fdebbcba5479efe524ec5db8de"}, - {file = "fiona-1.9.5-cp310-cp310-win_amd64.whl", hash = "sha256:5b4c6a3df53bee8f85bb46685562b21b43346be1fe96419f18f70fa1ab8c561c"}, - {file = "fiona-1.9.5-cp311-cp311-macosx_10_15_x86_64.whl", hash = "sha256:6ad04c1877b9fd742871b11965606c6a52f40706f56a48d66a87cc3073943828"}, - {file = "fiona-1.9.5-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:9fb9a24a8046c724787719e20557141b33049466145fc3e665764ac7caf5748c"}, - {file = "fiona-1.9.5-cp311-cp311-manylinux2014_x86_64.whl", hash = "sha256:d722d7f01a66f4ab6cd08d156df3fdb92f0669cf5f8708ddcb209352f416f241"}, - {file = "fiona-1.9.5-cp311-cp311-win_amd64.whl", hash = "sha256:7ede8ddc798f3d447536080c6db9a5fb73733ad8bdb190cb65eed4e289dd4c50"}, - {file = "fiona-1.9.5-cp312-cp312-macosx_10_15_x86_64.whl", hash = "sha256:8b098054a27c12afac4f819f98cb4d4bf2db9853f70b0c588d7d97d26e128c39"}, - {file = "fiona-1.9.5-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:6d9f29e9bcbb33232ff7fa98b4a3c2234db910c1dc6c4147fc36c0b8b930f2e0"}, - {file = "fiona-1.9.5-cp312-cp312-manylinux2014_x86_64.whl", hash = "sha256:f1af08da4ecea5036cb81c9131946be4404245d1b434b5b24fd3871a1d4030d9"}, - {file = "fiona-1.9.5-cp312-cp312-win_amd64.whl", hash = "sha256:c521e1135c78dec0d7774303e5a1b4c62e0efb0e602bb8f167550ef95e0a2691"}, - {file = "fiona-1.9.5-cp37-cp37m-macosx_10_15_x86_64.whl", hash = "sha256:fce4b1dd98810cabccdaa1828430c7402d283295c2ae31bea4f34188ea9e88d7"}, - {file = "fiona-1.9.5-cp37-cp37m-manylinux2014_x86_64.whl", hash = "sha256:3ea04ec2d8c57b5f81a31200fb352cb3242aa106fc3e328963f30ffbdf0ff7c8"}, - {file = "fiona-1.9.5-cp37-cp37m-win_amd64.whl", hash = "sha256:4877cc745d9e82b12b3eafce3719db75759c27bd8a695521202135b36b58c2e7"}, - {file = "fiona-1.9.5-cp38-cp38-macosx_10_15_x86_64.whl", hash = "sha256:ac2c250f509ec19fad7959d75b531984776517ef3c1222d1cc5b4f962825880b"}, - {file = "fiona-1.9.5-cp38-cp38-macosx_11_0_arm64.whl", hash = "sha256:4df21906235928faad856c288cfea0298e9647f09c9a69a230535cbc8eadfa21"}, - {file = "fiona-1.9.5-cp38-cp38-manylinux2014_x86_64.whl", hash = "sha256:81d502369493687746cb8d3cd77e5ada4447fb71d513721c9a1826e4fb32b23a"}, - {file = "fiona-1.9.5-cp38-cp38-win_amd64.whl", hash = "sha256:ce3b29230ef70947ead4e701f3f82be81082b7f37fd4899009b1445cc8fc276a"}, - {file = "fiona-1.9.5-cp39-cp39-macosx_10_15_x86_64.whl", hash = "sha256:8b53ce8de773fcd5e2e102e833c8c58479edd8796a522f3d83ef9e08b62bfeea"}, - {file = "fiona-1.9.5-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:bd2355e859a1cd24a3e485c6dc5003129f27a2051629def70036535ffa7e16a4"}, - {file = "fiona-1.9.5-cp39-cp39-manylinux2014_x86_64.whl", hash = "sha256:9a2da52f865db1aff0eaf41cdd4c87a7c079b3996514e8e7a1ca38457309e825"}, - {file = "fiona-1.9.5-cp39-cp39-win_amd64.whl", hash = "sha256:cfef6db5b779d463298b1113b50daa6c5b55f26f834dc9e37752116fa17277c1"}, - {file = "fiona-1.9.5.tar.gz", hash = "sha256:99e2604332caa7692855c2ae6ed91e1fffdf9b59449aa8032dd18e070e59a2f7"}, + {file = "fiona-1.9.6-cp310-cp310-macosx_10_15_x86_64.whl", hash = "sha256:63e528b5ea3d8b1038d788e7c65117835c787ba7fdc94b1b42f09c2cbc0aaff2"}, + {file = "fiona-1.9.6-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:918bd27d8625416672e834593970f96dff63215108f81efb876fe5c0bc58a3b4"}, + {file = "fiona-1.9.6-cp310-cp310-manylinux2014_x86_64.whl", hash = "sha256:e313210b30d09ed8f829bf625599e248dadd78622728030221f6526580ff26c5"}, + {file = "fiona-1.9.6-cp310-cp310-win_amd64.whl", hash = "sha256:89095c2d542325ee45894b8837e8048cdbb2f22274934e1be3b673ca628010d7"}, + {file = "fiona-1.9.6-cp311-cp311-macosx_10_15_x86_64.whl", hash = "sha256:98cea6f435843b2119731c6b0470e5b7386aa16b6aa7edabbf1ed93aefe029c3"}, + {file = "fiona-1.9.6-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:f4230eccbd896a79d1ebfa551d84bf90f512f7bcbe1ca61e3f82231321f1a532"}, + {file = "fiona-1.9.6-cp311-cp311-manylinux2014_x86_64.whl", hash = "sha256:48b6218224e96de5e36b5eb259f37160092260e5de0dcd82ca200b1887aa9884"}, + {file = "fiona-1.9.6-cp311-cp311-win_amd64.whl", hash = "sha256:c1dd5fbc29b7303bb87eb683455e8451e1a53bb8faf20ef97fdcd843c9e4a7f6"}, + {file = "fiona-1.9.6-cp312-cp312-macosx_10_15_x86_64.whl", hash = "sha256:42d8a0e5570948d3821c493b6141866d9a4d7a64edad2be4ecbb89f81904baac"}, + {file = "fiona-1.9.6-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:39819fb8f5ec6d9971cb01b912b4431615a3d3f50c83798565d8ce41917930db"}, + {file = "fiona-1.9.6-cp312-cp312-manylinux2014_x86_64.whl", hash = "sha256:9b53034efdf93ada9295b081e6a8280af7c75496a20df82d4c2ca46d65b85905"}, + {file = "fiona-1.9.6-cp312-cp312-win_amd64.whl", hash = "sha256:1dcd6eca7524535baf2a39d7981b4a46d33ae28c313934a7c3eae62eecf9dfa5"}, + {file = "fiona-1.9.6-cp37-cp37m-macosx_10_15_x86_64.whl", hash = "sha256:e5404ed08c711489abcb3a50a184816825b8af06eb73ad2a99e18b8e7b47c96a"}, + {file = "fiona-1.9.6-cp37-cp37m-manylinux2014_x86_64.whl", hash = "sha256:53bedd2989e255df1bf3378ae9c06d6d241ec273c280c544bb44ffffebb97fb0"}, + {file = "fiona-1.9.6-cp37-cp37m-win_amd64.whl", hash = "sha256:77653a08564a44e634c44cd74a068d2f55d1d4029edd16d1c8aadcc4d8cc1d2c"}, + {file = "fiona-1.9.6-cp38-cp38-macosx_10_15_x86_64.whl", hash = "sha256:e7617563b36d2be99f048f0d0054b4d765f4aae454398f88f19de9c2c324b7f8"}, + {file = "fiona-1.9.6-cp38-cp38-macosx_11_0_arm64.whl", hash = "sha256:50037c3b7a5f6f434b562b5b1a5b664f1caa7a4383b00af23cdb59bfc6ba852c"}, + {file = "fiona-1.9.6-cp38-cp38-manylinux2014_x86_64.whl", hash = "sha256:bf51846ad602757bf27876f458c5c9f14b09421fac612f64273cc4e3fcabc441"}, + {file = "fiona-1.9.6-cp38-cp38-win_amd64.whl", hash = "sha256:11af1afc1255642a7787fe112c29d01f968f1053e4d4700fc6f3bb879c1622e0"}, + {file = "fiona-1.9.6-cp39-cp39-macosx_10_15_x86_64.whl", hash = "sha256:52e8fec650b72fc5253d8f86b63859acc687182281c29bfacd3930496cf982d1"}, + {file = "fiona-1.9.6-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:c9b92aa1badb2773e7cac19bef3064d73e9d80c67c42f0928db2520a04be6f2f"}, + {file = "fiona-1.9.6-cp39-cp39-manylinux2014_x86_64.whl", hash = "sha256:0eaffbf3bfae9960484c0c08ea461b0c40e111497f04e9475ebf15ac7a22d9dc"}, + {file = "fiona-1.9.6-cp39-cp39-win_amd64.whl", hash = "sha256:f1b49d51a744874608b689f029766aa1e078dd72e94b44cf8eeef6d7bd2e9051"}, + {file = "fiona-1.9.6.tar.gz", hash = "sha256:791b3494f8b218c06ea56f892bd6ba893dfa23525347761d066fb7738acda3b1"}, ] [package.dependencies] @@ -1116,24 +1147,23 @@ click = ">=8.0,<9.0" click-plugins = ">=1.0" cligj = ">=0.5" importlib-metadata = {version = "*", markers = "python_version < \"3.10\""} -setuptools = "*" six = "*" [package.extras] -all = ["Fiona[calc,s3,test]"] +all = ["fiona[calc,s3,test]"] calc = ["shapely"] s3 = ["boto3 (>=1.3.1)"] -test = ["Fiona[s3]", "pytest (>=7)", "pytest-cov", "pytz"] +test = ["fiona[s3]", "pytest (>=7)", "pytest-cov", "pytz"] [[package]] name = "flox" -version = "0.9.0" +version = "0.9.7" description = "GroupBy operations for dask.array" optional = false python-versions = ">=3.9" files = [ - {file = "flox-0.9.0-py3-none-any.whl", hash = "sha256:b775e19c104971e7b869cab45830aadeb182866b59256ce87ea79fc88868e201"}, - {file = "flox-0.9.0.tar.gz", hash = "sha256:a7be517654216d5c531b1fb4ea32508736f9496240a61e3ae2974a0dcfcde89e"}, + {file = "flox-0.9.7-py3-none-any.whl", hash = "sha256:0e15d678c5f3d46fe5c6481519d01ceae40a111133b110e80f3b274881af8497"}, + {file = "flox-0.9.7.tar.gz", hash = "sha256:baa7c0aa9b2836f5cf1b283ce918cf3d61dc9ff0af8bda026a598ba5cc0b7c68"}, ] [package.dependencies] @@ -1150,53 +1180,53 @@ test = ["netCDF4"] [[package]] name = "fonttools" -version = "4.48.1" +version = "4.51.0" description = "Tools to manipulate font files" optional = false python-versions = ">=3.8" files = [ - {file = "fonttools-4.48.1-cp310-cp310-macosx_10_9_universal2.whl", hash = "sha256:702ae93058c81f46461dc4b2c79f11d3c3d8fd7296eaf8f75b4ba5bbf813cd5f"}, - {file = "fonttools-4.48.1-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:97f0a49fa6aa2d6205c6f72f4f98b74ef4b9bfdcb06fd78e6fe6c7af4989b63e"}, - {file = "fonttools-4.48.1-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:d3260db55f1843e57115256e91247ad9f68cb02a434b51262fe0019e95a98738"}, - {file = "fonttools-4.48.1-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:e740a7602c2bb71e1091269b5dbe89549749a8817dc294b34628ffd8b2bf7124"}, - {file = "fonttools-4.48.1-cp310-cp310-musllinux_1_1_aarch64.whl", hash = "sha256:4108b1d247953dd7c90ec8f457a2dec5fceb373485973cc852b14200118a51ee"}, - {file = "fonttools-4.48.1-cp310-cp310-musllinux_1_1_x86_64.whl", hash = "sha256:56339ec557f0c342bddd7c175f5e41c45fc21282bee58a86bd9aa322bec715f2"}, - {file = "fonttools-4.48.1-cp310-cp310-win32.whl", hash = "sha256:bff5b38d0e76eb18e0b8abbf35d384e60b3371be92f7be36128ee3e67483b3ec"}, - {file = "fonttools-4.48.1-cp310-cp310-win_amd64.whl", hash = "sha256:f7449493886da6a17472004d3818cc050ba3f4a0aa03fb47972e4fa5578e6703"}, - {file = "fonttools-4.48.1-cp311-cp311-macosx_10_9_universal2.whl", hash = "sha256:18b35fd1a850ed7233a99bbd6774485271756f717dac8b594958224b54118b61"}, - {file = "fonttools-4.48.1-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:cad5cfd044ea2e306fda44482b3dd32ee47830fa82dfa4679374b41baa294f5f"}, - {file = "fonttools-4.48.1-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:6f30e605c7565d0da6f0aec75a30ec372072d016957cd8fc4469721a36ea59b7"}, - {file = "fonttools-4.48.1-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:aee76fd81a8571c68841d6ef0da750d5ff08ff2c5f025576473016f16ac3bcf7"}, - {file = "fonttools-4.48.1-cp311-cp311-musllinux_1_1_aarch64.whl", hash = "sha256:5057ade278e67923000041e2b195c9ea53e87f227690d499b6a4edd3702f7f01"}, - {file = "fonttools-4.48.1-cp311-cp311-musllinux_1_1_x86_64.whl", hash = "sha256:b10633aafc5932995a391ec07eba5e79f52af0003a1735b2306b3dab8a056d48"}, - {file = "fonttools-4.48.1-cp311-cp311-win32.whl", hash = "sha256:0d533f89819f9b3ee2dbedf0fed3825c425850e32bdda24c558563c71be0064e"}, - {file = "fonttools-4.48.1-cp311-cp311-win_amd64.whl", hash = "sha256:d20588466367f05025bb1efdf4e5d498ca6d14bde07b6928b79199c588800f0a"}, - {file = "fonttools-4.48.1-cp312-cp312-macosx_10_9_universal2.whl", hash = "sha256:0a2417547462e468edf35b32e3dd06a6215ac26aa6316b41e03b8eeaf9f079ea"}, - {file = "fonttools-4.48.1-cp312-cp312-macosx_10_9_x86_64.whl", hash = "sha256:cf5a0cd974f85a80b74785db2d5c3c1fd6cc09a2ba3c837359b2b5da629ee1b0"}, - {file = "fonttools-4.48.1-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:0452fcbfbce752ba596737a7c5ec5cf76bc5f83847ce1781f4f90eab14ece252"}, - {file = "fonttools-4.48.1-cp312-cp312-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:578c00f93868f64a4102ecc5aa600a03b49162c654676c3fadc33de2ddb88a81"}, - {file = "fonttools-4.48.1-cp312-cp312-musllinux_1_1_aarch64.whl", hash = "sha256:63dc592a16cd08388d8c4c7502b59ac74190b23e16dfc863c69fe1ea74605b68"}, - {file = "fonttools-4.48.1-cp312-cp312-musllinux_1_1_x86_64.whl", hash = "sha256:9b58638d8a85e3a1b32ec0a91d9f8171a877b4b81c408d4cb3257d0dee63e092"}, - {file = "fonttools-4.48.1-cp312-cp312-win32.whl", hash = "sha256:d10979ef14a8beaaa32f613bb698743f7241d92f437a3b5e32356dfb9769c65d"}, - {file = "fonttools-4.48.1-cp312-cp312-win_amd64.whl", hash = "sha256:cdfd7557d1bd294a200bd211aa665ca3b02998dcc18f8211a5532da5b8fad5c5"}, - {file = "fonttools-4.48.1-cp38-cp38-macosx_10_9_universal2.whl", hash = "sha256:3cdb9a92521b81bf717ebccf592bd0292e853244d84115bfb4db0c426de58348"}, - {file = "fonttools-4.48.1-cp38-cp38-macosx_10_9_x86_64.whl", hash = "sha256:9b4ec6d42a7555f5ae35f3b805482f0aad0f1baeeef54859492ea3b782959d4a"}, - {file = "fonttools-4.48.1-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:902e9c4e9928301912f34a6638741b8ae0b64824112b42aaf240e06b735774b1"}, - {file = "fonttools-4.48.1-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:a8c8b54bd1420c184a995f980f1a8076f87363e2bb24239ef8c171a369d85a31"}, - {file = "fonttools-4.48.1-cp38-cp38-musllinux_1_1_aarch64.whl", hash = "sha256:12ee86abca46193359ea69216b3a724e90c66ab05ab220d39e3fc068c1eb72ac"}, - {file = "fonttools-4.48.1-cp38-cp38-musllinux_1_1_x86_64.whl", hash = "sha256:6978bade7b6c0335095bdd0bd97f8f3d590d2877b370f17e03e0865241694eb5"}, - {file = "fonttools-4.48.1-cp38-cp38-win32.whl", hash = "sha256:bcd77f89fc1a6b18428e7a55dde8ef56dae95640293bfb8f4e929929eba5e2a2"}, - {file = "fonttools-4.48.1-cp38-cp38-win_amd64.whl", hash = "sha256:f40441437b039930428e04fb05ac3a132e77458fb57666c808d74a556779e784"}, - {file = "fonttools-4.48.1-cp39-cp39-macosx_10_9_universal2.whl", hash = "sha256:0d2b01428f7da26f229a5656defc824427b741e454b4e210ad2b25ed6ea2aed4"}, - {file = "fonttools-4.48.1-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:df48798f9a4fc4c315ab46e17873436c8746f5df6eddd02fad91299b2af7af95"}, - {file = "fonttools-4.48.1-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:2eb4167bde04e172a93cf22c875d8b0cff76a2491f67f5eb069566215302d45d"}, - {file = "fonttools-4.48.1-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:c900508c46274d32d308ae8e82335117f11aaee1f7d369ac16502c9a78930b0a"}, - {file = "fonttools-4.48.1-cp39-cp39-musllinux_1_1_aarch64.whl", hash = "sha256:594206b31c95fcfa65f484385171fabb4ec69f7d2d7f56d27f17db26b7a31814"}, - {file = "fonttools-4.48.1-cp39-cp39-musllinux_1_1_x86_64.whl", hash = "sha256:292922dc356d7f11f5063b4111a8b719efb8faea92a2a88ed296408d449d8c2e"}, - {file = "fonttools-4.48.1-cp39-cp39-win32.whl", hash = "sha256:4709c5bf123ba10eac210d2d5c9027d3f472591d9f1a04262122710fa3d23199"}, - {file = "fonttools-4.48.1-cp39-cp39-win_amd64.whl", hash = "sha256:63c73b9dd56a94a3cbd2f90544b5fca83666948a9e03370888994143b8d7c070"}, - {file = "fonttools-4.48.1-py3-none-any.whl", hash = "sha256:e3e33862fc5261d46d9aae3544acb36203b1a337d00bdb5d3753aae50dac860e"}, - {file = "fonttools-4.48.1.tar.gz", hash = "sha256:8b8a45254218679c7f1127812761e7854ed5c8e34349aebf581e8c9204e7495a"}, + {file = "fonttools-4.51.0-cp310-cp310-macosx_10_9_universal2.whl", hash = "sha256:84d7751f4468dd8cdd03ddada18b8b0857a5beec80bce9f435742abc9a851a74"}, + {file = "fonttools-4.51.0-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:8b4850fa2ef2cfbc1d1f689bc159ef0f45d8d83298c1425838095bf53ef46308"}, + {file = "fonttools-4.51.0-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:b5b48a1121117047d82695d276c2af2ee3a24ffe0f502ed581acc2673ecf1037"}, + {file = "fonttools-4.51.0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:180194c7fe60c989bb627d7ed5011f2bef1c4d36ecf3ec64daec8302f1ae0716"}, + {file = "fonttools-4.51.0-cp310-cp310-musllinux_1_1_aarch64.whl", hash = "sha256:96a48e137c36be55e68845fc4284533bda2980f8d6f835e26bca79d7e2006438"}, + {file = "fonttools-4.51.0-cp310-cp310-musllinux_1_1_x86_64.whl", hash = "sha256:806e7912c32a657fa39d2d6eb1d3012d35f841387c8fc6cf349ed70b7c340039"}, + {file = "fonttools-4.51.0-cp310-cp310-win32.whl", hash = "sha256:32b17504696f605e9e960647c5f64b35704782a502cc26a37b800b4d69ff3c77"}, + {file = "fonttools-4.51.0-cp310-cp310-win_amd64.whl", hash = "sha256:c7e91abdfae1b5c9e3a543f48ce96013f9a08c6c9668f1e6be0beabf0a569c1b"}, + {file = "fonttools-4.51.0-cp311-cp311-macosx_10_9_universal2.whl", hash = "sha256:a8feca65bab31479d795b0d16c9a9852902e3a3c0630678efb0b2b7941ea9c74"}, + {file = "fonttools-4.51.0-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:8ac27f436e8af7779f0bb4d5425aa3535270494d3bc5459ed27de3f03151e4c2"}, + {file = "fonttools-4.51.0-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:0e19bd9e9964a09cd2433a4b100ca7f34e34731e0758e13ba9a1ed6e5468cc0f"}, + {file = "fonttools-4.51.0-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:b2b92381f37b39ba2fc98c3a45a9d6383bfc9916a87d66ccb6553f7bdd129097"}, + {file = "fonttools-4.51.0-cp311-cp311-musllinux_1_1_aarch64.whl", hash = "sha256:5f6bc991d1610f5c3bbe997b0233cbc234b8e82fa99fc0b2932dc1ca5e5afec0"}, + {file = "fonttools-4.51.0-cp311-cp311-musllinux_1_1_x86_64.whl", hash = "sha256:9696fe9f3f0c32e9a321d5268208a7cc9205a52f99b89479d1b035ed54c923f1"}, + {file = "fonttools-4.51.0-cp311-cp311-win32.whl", hash = "sha256:3bee3f3bd9fa1d5ee616ccfd13b27ca605c2b4270e45715bd2883e9504735034"}, + {file = "fonttools-4.51.0-cp311-cp311-win_amd64.whl", hash = "sha256:0f08c901d3866a8905363619e3741c33f0a83a680d92a9f0e575985c2634fcc1"}, + {file = "fonttools-4.51.0-cp312-cp312-macosx_10_9_universal2.whl", hash = "sha256:4060acc2bfa2d8e98117828a238889f13b6f69d59f4f2d5857eece5277b829ba"}, + {file = "fonttools-4.51.0-cp312-cp312-macosx_10_9_x86_64.whl", hash = "sha256:1250e818b5f8a679ad79660855528120a8f0288f8f30ec88b83db51515411fcc"}, + {file = "fonttools-4.51.0-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:76f1777d8b3386479ffb4a282e74318e730014d86ce60f016908d9801af9ca2a"}, + {file = "fonttools-4.51.0-cp312-cp312-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:8b5ad456813d93b9c4b7ee55302208db2b45324315129d85275c01f5cb7e61a2"}, + {file = "fonttools-4.51.0-cp312-cp312-musllinux_1_1_aarch64.whl", hash = "sha256:68b3fb7775a923be73e739f92f7e8a72725fd333eab24834041365d2278c3671"}, + {file = "fonttools-4.51.0-cp312-cp312-musllinux_1_1_x86_64.whl", hash = "sha256:8e2f1a4499e3b5ee82c19b5ee57f0294673125c65b0a1ff3764ea1f9db2f9ef5"}, + {file = "fonttools-4.51.0-cp312-cp312-win32.whl", hash = "sha256:278e50f6b003c6aed19bae2242b364e575bcb16304b53f2b64f6551b9c000e15"}, + {file = "fonttools-4.51.0-cp312-cp312-win_amd64.whl", hash = "sha256:b3c61423f22165541b9403ee39874dcae84cd57a9078b82e1dce8cb06b07fa2e"}, + {file = "fonttools-4.51.0-cp38-cp38-macosx_10_9_universal2.whl", hash = "sha256:1621ee57da887c17312acc4b0e7ac30d3a4fb0fec6174b2e3754a74c26bbed1e"}, + {file = "fonttools-4.51.0-cp38-cp38-macosx_10_9_x86_64.whl", hash = "sha256:e9d9298be7a05bb4801f558522adbe2feea1b0b103d5294ebf24a92dd49b78e5"}, + {file = "fonttools-4.51.0-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:ee1af4be1c5afe4c96ca23badd368d8dc75f611887fb0c0dac9f71ee5d6f110e"}, + {file = "fonttools-4.51.0-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:c18b49adc721a7d0b8dfe7c3130c89b8704baf599fb396396d07d4aa69b824a1"}, + {file = "fonttools-4.51.0-cp38-cp38-musllinux_1_1_aarch64.whl", hash = "sha256:de7c29bdbdd35811f14493ffd2534b88f0ce1b9065316433b22d63ca1cd21f14"}, + {file = "fonttools-4.51.0-cp38-cp38-musllinux_1_1_x86_64.whl", hash = "sha256:cadf4e12a608ef1d13e039864f484c8a968840afa0258b0b843a0556497ea9ed"}, + {file = "fonttools-4.51.0-cp38-cp38-win32.whl", hash = "sha256:aefa011207ed36cd280babfaa8510b8176f1a77261833e895a9d96e57e44802f"}, + {file = "fonttools-4.51.0-cp38-cp38-win_amd64.whl", hash = "sha256:865a58b6e60b0938874af0968cd0553bcd88e0b2cb6e588727117bd099eef836"}, + {file = "fonttools-4.51.0-cp39-cp39-macosx_10_9_universal2.whl", hash = "sha256:60a3409c9112aec02d5fb546f557bca6efa773dcb32ac147c6baf5f742e6258b"}, + {file = "fonttools-4.51.0-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:f7e89853d8bea103c8e3514b9f9dc86b5b4120afb4583b57eb10dfa5afbe0936"}, + {file = "fonttools-4.51.0-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:56fc244f2585d6c00b9bcc59e6593e646cf095a96fe68d62cd4da53dd1287b55"}, + {file = "fonttools-4.51.0-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:0d145976194a5242fdd22df18a1b451481a88071feadf251221af110ca8f00ce"}, + {file = "fonttools-4.51.0-cp39-cp39-musllinux_1_1_aarch64.whl", hash = "sha256:c5b8cab0c137ca229433570151b5c1fc6af212680b58b15abd797dcdd9dd5051"}, + {file = "fonttools-4.51.0-cp39-cp39-musllinux_1_1_x86_64.whl", hash = "sha256:54dcf21a2f2d06ded676e3c3f9f74b2bafded3a8ff12f0983160b13e9f2fb4a7"}, + {file = "fonttools-4.51.0-cp39-cp39-win32.whl", hash = "sha256:0118ef998a0699a96c7b28457f15546815015a2710a1b23a7bf6c1be60c01636"}, + {file = "fonttools-4.51.0-cp39-cp39-win_amd64.whl", hash = "sha256:599bdb75e220241cedc6faebfafedd7670335d2e29620d207dd0378a4e9ccc5a"}, + {file = "fonttools-4.51.0-py3-none-any.whl", hash = "sha256:15c94eeef6b095831067f72c825eb0e2d48bb4cea0647c1b05c981ecba2bf39f"}, + {file = "fonttools-4.51.0.tar.gz", hash = "sha256:dc0673361331566d7a663d7ce0f6fdcbfbdc1f59c6e3ed1165ad7202ca183c68"}, ] [package.extras] @@ -1301,13 +1331,13 @@ files = [ [[package]] name = "fsspec" -version = "2024.2.0" +version = "2024.5.0" description = "File-system specification" optional = false python-versions = ">=3.8" files = [ - {file = "fsspec-2024.2.0-py3-none-any.whl", hash = "sha256:817f969556fa5916bc682e02ca2045f96ff7f586d45110fcb76022063ad2c7d8"}, - {file = "fsspec-2024.2.0.tar.gz", hash = "sha256:b6ad1a679f760dda52b1168c859d01b7b80648ea6f7f7c7f5a8a91dc3f3ecb84"}, + {file = "fsspec-2024.5.0-py3-none-any.whl", hash = "sha256:e0fdbc446d67e182f49a70b82cf7889028a63588fde6b222521f10937b2b670c"}, + {file = "fsspec-2024.5.0.tar.gz", hash = "sha256:1d021b0b0f933e3b3029ed808eb400c08ba101ca2de4b3483fbc9ca23fcee94a"}, ] [package.extras] @@ -1315,7 +1345,7 @@ abfs = ["adlfs"] adl = ["adlfs"] arrow = ["pyarrow (>=1)"] dask = ["dask", "distributed"] -devel = ["pytest", "pytest-cov"] +dev = ["pre-commit", "ruff"] dropbox = ["dropbox", "dropboxdrivefs", "requests"] full = ["adlfs", "aiohttp (!=4.0.0a0,!=4.0.0a1)", "dask", "distributed", "dropbox", "dropboxdrivefs", "fusepy", "gcsfs", "libarchive-c", "ocifs", "panel", "paramiko", "pyarrow (>=1)", "pygit2", "requests", "s3fs", "smbprotocol", "tqdm"] fuse = ["fusepy"] @@ -1332,31 +1362,36 @@ s3 = ["s3fs"] sftp = ["paramiko"] smb = ["smbprotocol"] ssh = ["paramiko"] +test = ["aiohttp (!=4.0.0a0,!=4.0.0a1)", "numpy", "pytest", "pytest-asyncio (!=0.22.0)", "pytest-benchmark", "pytest-cov", "pytest-mock", "pytest-recording", "pytest-rerunfailures", "requests"] +test-downstream = ["aiobotocore (>=2.5.4,<3.0.0)", "dask-expr", "dask[dataframe,test]", "moto[server] (>4,<5)", "pytest-timeout", "xarray"] +test-full = ["adlfs", "aiohttp (!=4.0.0a0,!=4.0.0a1)", "cloudpickle", "dask", "distributed", "dropbox", "dropboxdrivefs", "fastparquet", "fusepy", "gcsfs", "jinja2", "kerchunk", "libarchive-c", "lz4", "notebook", "numpy", "ocifs", "pandas", "panel", "paramiko", "pyarrow", "pyarrow (>=1)", "pyftpdlib", "pygit2", "pytest", "pytest-asyncio (!=0.22.0)", "pytest-benchmark", "pytest-cov", "pytest-mock", "pytest-recording", "pytest-rerunfailures", "python-snappy", "requests", "smbprotocol", "tqdm", "urllib3", "zarr", "zstandard"] tqdm = ["tqdm"] [[package]] name = "future" -version = "0.18.3" +version = "1.0.0" description = "Clean single-source support for Python 3 and 2" optional = false python-versions = ">=2.6, !=3.0.*, !=3.1.*, !=3.2.*" files = [ - {file = "future-0.18.3.tar.gz", hash = "sha256:34a17436ed1e96697a86f9de3d15a3b0be01d8bc8de9c1dffd59fb8234ed5307"}, + {file = "future-1.0.0-py3-none-any.whl", hash = "sha256:929292d34f5872e70396626ef385ec22355a1fae8ad29e1a734c3e43f9fbc216"}, + {file = "future-1.0.0.tar.gz", hash = "sha256:bd2968309307861edae1458a4f8a4f3598c03be43b97521076aebf5d94c07b05"}, ] [[package]] name = "geopandas" -version = "0.14.3" +version = "0.14.4" description = "Geographic pandas extensions" optional = false python-versions = ">=3.9" files = [ - {file = "geopandas-0.14.3-py3-none-any.whl", hash = "sha256:41b31ad39e21bc9e8c4254f78f8dc4ce3d33d144e22e630a00bb336c83160204"}, - {file = "geopandas-0.14.3.tar.gz", hash = "sha256:748af035d4a068a4ae00cab384acb61d387685c833b0022e0729aa45216b23ac"}, + {file = "geopandas-0.14.4-py3-none-any.whl", hash = "sha256:3bb6473cb59d51e1a7fe2dbc24a1a063fb0ebdeddf3ce08ddbf8c7ddc99689aa"}, + {file = "geopandas-0.14.4.tar.gz", hash = "sha256:56765be9d58e2c743078085db3bd07dc6be7719f0dbe1dfdc1d705cb80be7c25"}, ] [package.dependencies] fiona = ">=1.8.21" +numpy = ">=1.22" packaging = "*" pandas = ">=1.4.0" pyproj = ">=3.3.0" @@ -1364,17 +1399,17 @@ shapely = ">=1.8.0" [[package]] name = "geoviews" -version = "1.11.0" +version = "1.12.0" description = "GeoViews is a Python library that makes it easy to explore and visualize geographical, meteorological, and oceanographic datasets, such as those used in weather, climate, and remote sensing research." optional = false python-versions = ">=3.9" files = [ - {file = "geoviews-1.11.0-py2.py3-none-any.whl", hash = "sha256:d54a838708b72936fdcbe3eb9e170f977fa7d3733a45a200c9ae322841d4373a"}, - {file = "geoviews-1.11.0.tar.gz", hash = "sha256:2c4d09554019867adfa4ea655617ce8bb9ffa08b534a7e591f961dc6078cffcc"}, + {file = "geoviews-1.12.0-py3-none-any.whl", hash = "sha256:5e8750d0e9a80dd4f5ce493d26cdde52880677c19d329a980b0a94e966dd3bd3"}, + {file = "geoviews-1.12.0.tar.gz", hash = "sha256:e2cbef0605e8fd1529bc643a31aeb61997f8f93c9b41a5aff8b2b355a76fa789"}, ] [package.dependencies] -bokeh = ">=3.2.0,<3.4.0" +bokeh = ">=3.4.0,<3.5.0" cartopy = ">=0.18.0" holoviews = ">=1.16.0" numpy = "*" @@ -1386,14 +1421,14 @@ shapely = "*" xyzservices = "*" [package.extras] -all = ["cartopy (>=0.20.0)", "codecov", "datashader", "fiona", "geodatasets", "geopandas", "graphviz", "iris (>=3.5)", "lxml", "matplotlib (>2.2)", "mock", "nbsite (>=0.8.2,<0.9.0)", "nbsmoke (>=0.2.0)", "netcdf4", "pandas", "pooch", "pyct", "pytest", "pytest-cov", "pytest-github-actions-annotate-failures", "rioxarray", "scipy", "selenium", "shapely", "xarray", "xesmf"] -build = ["bokeh (==3.3)", "param (>=1.9.2)", "pyct (>=0.4.4)", "pyviz-comms (>=0.6.0)"] -doc = ["cartopy (>=0.20.0)", "datashader", "fiona", "geodatasets", "geopandas", "graphviz", "iris (>=3.5)", "lxml", "matplotlib (>2.2)", "mock", "nbsite (>=0.8.2,<0.9.0)", "netcdf4", "pandas", "pooch", "pyct", "scipy", "selenium", "shapely", "xarray", "xesmf"] +all = ["cartopy (>=0.20.0)", "codecov", "datashader", "fiona", "geodatasets", "geopandas", "graphviz", "iris (>=3.5)", "lxml", "matplotlib (>2.2)", "mock", "nbsite (>=0.8.4,<0.9.0)", "nbval", "netcdf4", "pandas", "pooch", "pyct", "pytest", "pytest-cov", "pytest-github-actions-annotate-failures", "pyviz-comms", "rioxarray", "scipy", "selenium", "shapely", "xarray", "xesmf"] +build = ["bokeh (==3.4)", "param (>=1.9.2)", "pyct (>=0.4.4)", "setuptools"] +doc = ["cartopy (>=0.20.0)", "datashader", "fiona", "geodatasets", "geopandas", "graphviz", "iris (>=3.5)", "lxml", "matplotlib (>2.2)", "mock", "nbsite (>=0.8.4,<0.9.0)", "netcdf4", "pandas", "pooch", "pyct", "scipy", "selenium", "shapely", "xarray", "xesmf"] examples-extra = ["datashader", "fiona", "geodatasets", "geopandas", "iris (>=3.5)", "matplotlib (>2.2)", "mock", "netcdf4", "pandas", "pooch", "pyct", "scipy", "shapely", "xarray", "xesmf"] recommended = ["datashader", "geopandas", "matplotlib (>2.2)", "netcdf4", "pandas", "pooch", "pyct", "scipy", "shapely", "xarray"] -tests = ["fiona", "nbsmoke (>=0.2.0)", "pytest", "rioxarray"] +tests = ["fiona", "nbval", "pytest", "rioxarray"] tests-ci = ["codecov", "pytest-cov", "pytest-github-actions-annotate-failures"] -tests-core = ["geopandas", "matplotlib (>2.2)", "netcdf4", "pandas", "pooch", "pytest", "scipy", "shapely", "xarray"] +tests-core = ["geopandas", "matplotlib (>2.2)", "netcdf4", "pandas", "pooch", "pytest", "pyviz-comms", "scipy", "shapely", "xarray"] [[package]] name = "ghp-import" @@ -1414,13 +1449,13 @@ dev = ["flake8", "markdown", "twine", "wheel"] [[package]] name = "griffe" -version = "0.40.0" +version = "0.45.2" description = "Signatures for entire Python programs. Extract the structure, the frame, the skeleton of your project, to generate API documentation or find breaking changes in your API." optional = false python-versions = ">=3.8" files = [ - {file = "griffe-0.40.0-py3-none-any.whl", hash = "sha256:db1da6d1d8e08cbb20f1a7dee8c09da940540c2d4c1bfa26a9091cf6fc36a9ec"}, - {file = "griffe-0.40.0.tar.gz", hash = "sha256:76c4439eaa2737af46ae003c331ab6ca79c5365b552f7b5aed263a3b4125735b"}, + {file = "griffe-0.45.2-py3-none-any.whl", hash = "sha256:297ec8530d0c68e5b98ff86fb588ebc3aa3559bb5dc21f3caea8d9542a350133"}, + {file = "griffe-0.45.2.tar.gz", hash = "sha256:83ce7dcaafd8cb7f43cbf1a455155015a1eb624b1ffd93249e5e1c4a22b2fdb2"}, ] [package.dependencies] @@ -1446,36 +1481,32 @@ test = ["netCDF4", "pytest"] [[package]] name = "h5py" -version = "3.10.0" +version = "3.11.0" description = "Read and write HDF5 files from Python" optional = false python-versions = ">=3.8" files = [ - {file = "h5py-3.10.0-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:b963fb772964fc1d1563c57e4e2e874022ce11f75ddc6df1a626f42bd49ab99f"}, - {file = "h5py-3.10.0-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:012ab448590e3c4f5a8dd0f3533255bc57f80629bf7c5054cf4c87b30085063c"}, - {file = "h5py-3.10.0-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:781a24263c1270a62cd67be59f293e62b76acfcc207afa6384961762bb88ea03"}, - {file = "h5py-3.10.0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:f42e6c30698b520f0295d70157c4e202a9e402406f50dc08f5a7bc416b24e52d"}, - {file = "h5py-3.10.0-cp310-cp310-win_amd64.whl", hash = "sha256:93dd840bd675787fc0b016f7a05fc6efe37312a08849d9dd4053fd0377b1357f"}, - {file = "h5py-3.10.0-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:2381e98af081b6df7f6db300cd88f88e740649d77736e4b53db522d8874bf2dc"}, - {file = "h5py-3.10.0-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:667fe23ab33d5a8a6b77970b229e14ae3bb84e4ea3382cc08567a02e1499eedd"}, - {file = "h5py-3.10.0-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:90286b79abd085e4e65e07c1bd7ee65a0f15818ea107f44b175d2dfe1a4674b7"}, - {file = "h5py-3.10.0-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:6c013d2e79c00f28ffd0cc24e68665ea03ae9069e167087b2adb5727d2736a52"}, - {file = "h5py-3.10.0-cp311-cp311-win_amd64.whl", hash = "sha256:92273ce69ae4983dadb898fd4d3bea5eb90820df953b401282ee69ad648df684"}, - {file = "h5py-3.10.0-cp312-cp312-macosx_10_9_x86_64.whl", hash = "sha256:3c97d03f87f215e7759a354460fb4b0d0f27001450b18b23e556e7856a0b21c3"}, - {file = "h5py-3.10.0-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:86df4c2de68257b8539a18646ceccdcf2c1ce6b1768ada16c8dcfb489eafae20"}, - {file = "h5py-3.10.0-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:ba9ab36be991119a3ff32d0c7cbe5faf9b8d2375b5278b2aea64effbeba66039"}, - {file = "h5py-3.10.0-cp312-cp312-win_amd64.whl", hash = "sha256:2c8e4fda19eb769e9a678592e67eaec3a2f069f7570c82d2da909c077aa94339"}, - {file = "h5py-3.10.0-cp38-cp38-macosx_10_9_x86_64.whl", hash = "sha256:492305a074327e8d2513011fa9fffeb54ecb28a04ca4c4227d7e1e9616d35641"}, - {file = "h5py-3.10.0-cp38-cp38-macosx_11_0_arm64.whl", hash = "sha256:9450464b458cca2c86252b624279115dcaa7260a40d3cb1594bf2b410a2bd1a3"}, - {file = "h5py-3.10.0-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:fd6f6d1384a9f491732cee233b99cd4bfd6e838a8815cc86722f9d2ee64032af"}, - {file = "h5py-3.10.0-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:3074ec45d3dc6e178c6f96834cf8108bf4a60ccb5ab044e16909580352010a97"}, - {file = "h5py-3.10.0-cp38-cp38-win_amd64.whl", hash = "sha256:212bb997a91e6a895ce5e2f365ba764debeaef5d2dca5c6fb7098d66607adf99"}, - {file = "h5py-3.10.0-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:5dfc65ac21fa2f630323c92453cadbe8d4f504726ec42f6a56cf80c2f90d6c52"}, - {file = "h5py-3.10.0-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:d4682b94fd36ab217352be438abd44c8f357c5449b8995e63886b431d260f3d3"}, - {file = "h5py-3.10.0-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:aece0e2e1ed2aab076c41802e50a0c3e5ef8816d60ece39107d68717d4559824"}, - {file = "h5py-3.10.0-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:43a61b2c2ad65b1fabc28802d133eed34debcc2c8b420cb213d3d4ef4d3e2229"}, - {file = "h5py-3.10.0-cp39-cp39-win_amd64.whl", hash = "sha256:ae2f0201c950059676455daf92700eeb57dcf5caaf71b9e1328e6e6593601770"}, - {file = "h5py-3.10.0.tar.gz", hash = "sha256:d93adc48ceeb33347eb24a634fb787efc7ae4644e6ea4ba733d099605045c049"}, + {file = "h5py-3.11.0-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:1625fd24ad6cfc9c1ccd44a66dac2396e7ee74940776792772819fc69f3a3731"}, + {file = "h5py-3.11.0-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:c072655ad1d5fe9ef462445d3e77a8166cbfa5e599045f8aa3c19b75315f10e5"}, + {file = "h5py-3.11.0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:77b19a40788e3e362b54af4dcf9e6fde59ca016db2c61360aa30b47c7b7cef00"}, + {file = "h5py-3.11.0-cp310-cp310-win_amd64.whl", hash = "sha256:ef4e2f338fc763f50a8113890f455e1a70acd42a4d083370ceb80c463d803972"}, + {file = "h5py-3.11.0-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:bbd732a08187a9e2a6ecf9e8af713f1d68256ee0f7c8b652a32795670fb481ba"}, + {file = "h5py-3.11.0-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:75bd7b3d93fbeee40860fd70cdc88df4464e06b70a5ad9ce1446f5f32eb84007"}, + {file = "h5py-3.11.0-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:52c416f8eb0daae39dabe71415cb531f95dce2d81e1f61a74537a50c63b28ab3"}, + {file = "h5py-3.11.0-cp311-cp311-win_amd64.whl", hash = "sha256:083e0329ae534a264940d6513f47f5ada617da536d8dccbafc3026aefc33c90e"}, + {file = "h5py-3.11.0-cp312-cp312-macosx_10_9_x86_64.whl", hash = "sha256:a76cae64080210389a571c7d13c94a1a6cf8cb75153044fd1f822a962c97aeab"}, + {file = "h5py-3.11.0-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:f3736fe21da2b7d8a13fe8fe415f1272d2a1ccdeff4849c1421d2fb30fd533bc"}, + {file = "h5py-3.11.0-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:aa6ae84a14103e8dc19266ef4c3e5d7c00b68f21d07f2966f0ca7bdb6c2761fb"}, + {file = "h5py-3.11.0-cp312-cp312-win_amd64.whl", hash = "sha256:21dbdc5343f53b2e25404673c4f00a3335aef25521bd5fa8c707ec3833934892"}, + {file = "h5py-3.11.0-cp38-cp38-macosx_10_9_x86_64.whl", hash = "sha256:754c0c2e373d13d6309f408325343b642eb0f40f1a6ad21779cfa9502209e150"}, + {file = "h5py-3.11.0-cp38-cp38-macosx_11_0_arm64.whl", hash = "sha256:731839240c59ba219d4cb3bc5880d438248533366f102402cfa0621b71796b62"}, + {file = "h5py-3.11.0-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:8ec9df3dd2018904c4cc06331951e274f3f3fd091e6d6cc350aaa90fa9b42a76"}, + {file = "h5py-3.11.0-cp38-cp38-win_amd64.whl", hash = "sha256:55106b04e2c83dfb73dc8732e9abad69d83a436b5b82b773481d95d17b9685e1"}, + {file = "h5py-3.11.0-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:f4e025e852754ca833401777c25888acb96889ee2c27e7e629a19aee288833f0"}, + {file = "h5py-3.11.0-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:6c4b760082626120031d7902cd983d8c1f424cdba2809f1067511ef283629d4b"}, + {file = "h5py-3.11.0-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:67462d0669f8f5459529de179f7771bd697389fcb3faab54d63bf788599a48ea"}, + {file = "h5py-3.11.0-cp39-cp39-win_amd64.whl", hash = "sha256:d9c944d364688f827dc889cf83f1fca311caf4fa50b19f009d1f2b525edd33a3"}, + {file = "h5py-3.11.0.tar.gz", hash = "sha256:7b7e8f78072a2edec87c9836f25f34203fd492a4475709a18b417a33cfb21fa9"}, ] [package.dependencies] @@ -1483,13 +1514,13 @@ numpy = ">=1.17.3" [[package]] name = "holoviews" -version = "1.18.1" +version = "1.18.3" description = "Stop plotting your data - annotate your data and let it visualize itself." optional = false python-versions = ">=3.9" files = [ - {file = "holoviews-1.18.1-py2.py3-none-any.whl", hash = "sha256:b1bf6cd9aa72a2e103e8a7e0e57c8c8171a3fc3ebd5b943fe49bf98b6c64367b"}, - {file = "holoviews-1.18.1.tar.gz", hash = "sha256:805c7353ae52e97753e7f0668b051a086410e2d2c13c8fe294ebb4cdd02b2324"}, + {file = "holoviews-1.18.3-py2.py3-none-any.whl", hash = "sha256:b94b96560b64a84c07e89115aaf9b226e6009684800ec84d3c88cbad122c0c46"}, + {file = "holoviews-1.18.3.tar.gz", hash = "sha256:578e30e89d72754f97a83ebe08198fec8e87cc7e49b25b9f31ec393f939ca500"}, ] [package.dependencies] @@ -1502,61 +1533,61 @@ param = ">=1.12.0,<3.0" pyviz-comms = ">=0.7.4" [package.extras] -all = ["bokeh (>=3.1)", "cftime", "codecov", "contourpy", "cudf", "dash (>=1.16)", "dask", "datashader (>=0.11.1)", "ffmpeg", "flaky", "graphviz", "ibis-framework", "ipython (>=5.4.0)", "matplotlib (>=3)", "nbconvert", "nbsite (>=0.8.2,<0.9.0)", "nbval", "netcdf4", "networkx", "notebook", "pillow", "playwright", "plotly (>=4.0)", "pooch", "pre-commit", "pyarrow", "pytest", "pytest-cov", "pytest-github-actions-annotate-failures", "pytest-playwright", "pytest-xdist", "ruff", "scikit-image", "scipy", "selenium", "shapely", "spatialpandas", "streamz (>=0.5.0)", "xarray (>=0.10.4)"] +all = ["bokeh (>=3.1)", "cftime", "codecov", "contourpy", "cudf", "dash (>=1.16)", "dask", "datashader (>=0.11.1)", "ffmpeg", "graphviz", "ibis-framework", "ipython (>=5.4.0)", "matplotlib (>=3)", "myst-nb (<1)", "nbconvert", "nbsite (>=0.8.4,<0.9.0)", "nbval", "netcdf4", "networkx", "notebook", "notebook (>=7.0)", "pillow", "playwright", "plotly (>=4.0)", "pooch", "pre-commit", "pyarrow", "pytest", "pytest-cov", "pytest-github-actions-annotate-failures", "pytest-playwright", "pytest-rerunfailures", "pytest-xdist", "ruff", "scikit-image", "scipy", "scipy (>=1.10)", "selenium", "shapely", "spatialpandas", "streamz (>=0.5.0)", "xarray (>=0.10.4)"] build = ["param (>=1.7.0)", "pyct (>=0.4.4)", "setuptools (>=30.3.0)"] -doc = ["bokeh (>=3.1)", "cftime", "dash (>=1.16)", "dask", "datashader (>=0.11.1)", "ffmpeg", "graphviz", "ipython (>=5.4.0)", "matplotlib (>=3)", "nbsite (>=0.8.2,<0.9.0)", "netcdf4", "networkx", "notebook", "pillow", "plotly (>=4.0)", "pooch", "pyarrow", "scikit-image", "scipy", "selenium", "shapely", "streamz (>=0.5.0)", "xarray (>=0.10.4)"] -examples = ["bokeh (>=3.1)", "cftime", "dash (>=1.16)", "dask", "datashader (>=0.11.1)", "ffmpeg", "ipython (>=5.4.0)", "matplotlib (>=3)", "netcdf4", "networkx", "notebook", "pillow", "plotly (>=4.0)", "pooch", "pyarrow", "scikit-image", "scipy", "shapely", "streamz (>=0.5.0)", "xarray (>=0.10.4)"] -examples-tests = ["bokeh (>=3.1)", "cftime", "dash (>=1.16)", "dask", "datashader (>=0.11.1)", "ffmpeg", "ipython (>=5.4.0)", "matplotlib (>=3)", "nbval", "netcdf4", "networkx", "notebook", "pillow", "plotly (>=4.0)", "pooch", "pyarrow", "scikit-image", "scipy", "shapely", "streamz (>=0.5.0)", "xarray (>=0.10.4)"] +doc = ["bokeh (>=3.1)", "cftime", "dash (>=1.16)", "dask", "datashader (>=0.11.1)", "ffmpeg", "graphviz", "ipython (>=5.4.0)", "matplotlib (>=3)", "myst-nb (<1)", "nbsite (>=0.8.4,<0.9.0)", "netcdf4", "networkx", "notebook", "notebook (>=7.0)", "pillow", "plotly (>=4.0)", "pooch", "pyarrow", "scikit-image", "scipy", "selenium", "shapely", "streamz (>=0.5.0)", "xarray (>=0.10.4)"] +examples = ["bokeh (>=3.1)", "cftime", "dash (>=1.16)", "dask", "datashader (>=0.11.1)", "ffmpeg", "ipython (>=5.4.0)", "matplotlib (>=3)", "netcdf4", "networkx", "notebook", "notebook (>=7.0)", "pillow", "plotly (>=4.0)", "pooch", "pyarrow", "scikit-image", "scipy", "shapely", "streamz (>=0.5.0)", "xarray (>=0.10.4)"] +examples-tests = ["bokeh (>=3.1)", "cftime", "dash (>=1.16)", "dask", "datashader (>=0.11.1)", "ffmpeg", "ipython (>=5.4.0)", "matplotlib (>=3)", "nbval", "netcdf4", "networkx", "notebook", "notebook (>=7.0)", "pillow", "plotly (>=4.0)", "pooch", "pyarrow", "scikit-image", "scipy", "shapely", "streamz (>=0.5.0)", "xarray (>=0.10.4)"] lint = ["pre-commit", "ruff"] notebook = ["ipython (>=5.4.0)", "notebook"] recommended = ["bokeh (>=3.1)", "ipython (>=5.4.0)", "matplotlib (>=3)", "notebook"] -tests = ["bokeh (>=3.1)", "cftime", "contourpy", "dash (>=1.16)", "dask", "datashader (>=0.11.1)", "ffmpeg", "flaky", "ibis-framework", "ipython (>=5.4.0)", "matplotlib (>=3)", "nbconvert", "networkx", "pillow", "plotly (>=4.0)", "pytest", "pytest-cov", "pytest-xdist", "scipy", "selenium", "shapely", "spatialpandas", "xarray (>=0.10.4)"] +tests = ["bokeh (>=3.1)", "cftime", "contourpy", "dash (>=1.16)", "dask", "datashader (>=0.11.1)", "ffmpeg", "ibis-framework", "ipython (>=5.4.0)", "matplotlib (>=3)", "nbconvert", "networkx", "pillow", "plotly (>=4.0)", "pytest", "pytest-cov", "pytest-rerunfailures", "pytest-xdist", "scipy (>=1.10)", "selenium", "shapely", "spatialpandas", "xarray (>=0.10.4)"] tests-ci = ["codecov", "pytest-github-actions-annotate-failures"] -tests-core = ["bokeh (>=3.1)", "contourpy", "flaky", "ipython (>=5.4.0)", "matplotlib (>=3)", "nbconvert", "pillow", "plotly (>=4.0)", "pytest", "pytest-cov", "pytest-xdist"] -tests-gpu = ["bokeh (>=3.1)", "cftime", "contourpy", "cudf", "dash (>=1.16)", "dask", "datashader (>=0.11.1)", "ffmpeg", "flaky", "ibis-framework", "ipython (>=5.4.0)", "matplotlib (>=3)", "nbconvert", "networkx", "pillow", "plotly (>=4.0)", "pytest", "pytest-cov", "pytest-xdist", "scipy", "selenium", "shapely", "spatialpandas", "xarray (>=0.10.4)"] +tests-core = ["bokeh (>=3.1)", "contourpy", "ipython (>=5.4.0)", "matplotlib (>=3)", "nbconvert", "pillow", "plotly (>=4.0)", "pytest", "pytest-cov", "pytest-rerunfailures", "pytest-xdist"] +tests-gpu = ["bokeh (>=3.1)", "cftime", "contourpy", "cudf", "dash (>=1.16)", "dask", "datashader (>=0.11.1)", "ffmpeg", "ibis-framework", "ipython (>=5.4.0)", "matplotlib (>=3)", "nbconvert", "networkx", "pillow", "plotly (>=4.0)", "pytest", "pytest-cov", "pytest-rerunfailures", "pytest-xdist", "scipy (>=1.10)", "selenium", "shapely", "spatialpandas", "xarray (>=0.10.4)"] tests-nb = ["nbval"] ui = ["playwright", "pytest-playwright"] -unit-tests = ["bokeh (>=3.1)", "cftime", "contourpy", "dash (>=1.16)", "dask", "datashader (>=0.11.1)", "ffmpeg", "flaky", "ibis-framework", "ipython (>=5.4.0)", "matplotlib (>=3)", "nbconvert", "netcdf4", "networkx", "notebook", "pillow", "plotly (>=4.0)", "pooch", "pre-commit", "pyarrow", "pytest", "pytest-cov", "pytest-xdist", "ruff", "scikit-image", "scipy", "selenium", "shapely", "spatialpandas", "streamz (>=0.5.0)", "xarray (>=0.10.4)"] +unit-tests = ["bokeh (>=3.1)", "cftime", "contourpy", "dash (>=1.16)", "dask", "datashader (>=0.11.1)", "ffmpeg", "ibis-framework", "ipython (>=5.4.0)", "matplotlib (>=3)", "nbconvert", "netcdf4", "networkx", "notebook", "notebook (>=7.0)", "pillow", "plotly (>=4.0)", "pooch", "pre-commit", "pyarrow", "pytest", "pytest-cov", "pytest-rerunfailures", "pytest-xdist", "ruff", "scikit-image", "scipy", "scipy (>=1.10)", "selenium", "shapely", "spatialpandas", "streamz (>=0.5.0)", "xarray (>=0.10.4)"] [[package]] name = "idna" -version = "3.6" +version = "3.7" description = "Internationalized Domain Names in Applications (IDNA)" optional = false python-versions = ">=3.5" files = [ - {file = "idna-3.6-py3-none-any.whl", hash = "sha256:c05567e9c24a6b9faaa835c4821bad0590fbb9d5779e7caa6e1cc4978e7eb24f"}, - {file = "idna-3.6.tar.gz", hash = "sha256:9ecdbbd083b06798ae1e86adcbfe8ab1479cf864e4ee30fe4e46a003d12491ca"}, + {file = "idna-3.7-py3-none-any.whl", hash = "sha256:82fee1fc78add43492d3a1898bfa6d8a904cc97d8427f683ed8e798d07761aa0"}, + {file = "idna-3.7.tar.gz", hash = "sha256:028ff3aadf0609c1fd278d8ea3089299412a7a8b9bd005dd08b9f8285bcb5cfc"}, ] [[package]] name = "importlib-metadata" -version = "7.0.1" +version = "7.1.0" description = "Read metadata from Python packages" optional = false python-versions = ">=3.8" files = [ - {file = "importlib_metadata-7.0.1-py3-none-any.whl", hash = "sha256:4805911c3a4ec7c3966410053e9ec6a1fecd629117df5adee56dfc9432a1081e"}, - {file = "importlib_metadata-7.0.1.tar.gz", hash = "sha256:f238736bb06590ae52ac1fab06a3a9ef1d8dce2b7a35b5ab329371d6c8f5d2cc"}, + {file = "importlib_metadata-7.1.0-py3-none-any.whl", hash = "sha256:30962b96c0c223483ed6cc7280e7f0199feb01a0e40cfae4d4450fc6fab1f570"}, + {file = "importlib_metadata-7.1.0.tar.gz", hash = "sha256:b78938b926ee8d5f020fc4772d487045805a55ddbad2ecf21c6d60938dc7fcd2"}, ] [package.dependencies] zipp = ">=0.5" [package.extras] -docs = ["furo", "jaraco.packaging (>=9.3)", "jaraco.tidelift (>=1.4)", "rst.linker (>=1.9)", "sphinx (<7.2.5)", "sphinx (>=3.5)", "sphinx-lint"] +docs = ["furo", "jaraco.packaging (>=9.3)", "jaraco.tidelift (>=1.4)", "rst.linker (>=1.9)", "sphinx (>=3.5)", "sphinx-lint"] perf = ["ipython"] -testing = ["flufl.flake8", "importlib-resources (>=1.3)", "packaging", "pyfakefs", "pytest (>=6)", "pytest-black (>=0.3.7)", "pytest-checkdocs (>=2.4)", "pytest-cov", "pytest-enabler (>=2.2)", "pytest-mypy (>=0.9.1)", "pytest-perf (>=0.9.2)", "pytest-ruff"] +testing = ["flufl.flake8", "importlib-resources (>=1.3)", "jaraco.test (>=5.4)", "packaging", "pyfakefs", "pytest (>=6)", "pytest-checkdocs (>=2.4)", "pytest-cov", "pytest-enabler (>=2.2)", "pytest-mypy", "pytest-perf (>=0.9.2)", "pytest-ruff (>=0.2.1)"] [[package]] name = "importlib-resources" -version = "6.1.1" +version = "6.4.0" description = "Read resources from Python packages" optional = false python-versions = ">=3.8" files = [ - {file = "importlib_resources-6.1.1-py3-none-any.whl", hash = "sha256:e8bf90d8213b486f428c9c39714b920041cb02c184686a3dee24905aaa8105d6"}, - {file = "importlib_resources-6.1.1.tar.gz", hash = "sha256:3893a00122eafde6894c59914446a512f728a0c1a45f9bb9b63721b6bacf0b4a"}, + {file = "importlib_resources-6.4.0-py3-none-any.whl", hash = "sha256:50d10f043df931902d4194ea07ec57960f66a80449ff867bfe782b4c486ba78c"}, + {file = "importlib_resources-6.4.0.tar.gz", hash = "sha256:cdb2b453b8046ca4e3798eb1d84f3cce1446a0e8e7b5ef4efb600f19fc398145"}, ] [package.dependencies] @@ -1564,7 +1595,7 @@ zipp = {version = ">=3.1.0", markers = "python_version < \"3.10\""} [package.extras] docs = ["furo", "jaraco.packaging (>=9.3)", "jaraco.tidelift (>=1.4)", "rst.linker (>=1.9)", "sphinx (<7.2.5)", "sphinx (>=3.5)", "sphinx-lint"] -testing = ["pytest (>=6)", "pytest-black (>=0.3.7)", "pytest-checkdocs (>=2.4)", "pytest-cov", "pytest-enabler (>=2.2)", "pytest-mypy (>=0.9.1)", "pytest-ruff", "zipp (>=3.17)"] +testing = ["jaraco.test (>=5.4)", "pytest (>=6)", "pytest-checkdocs (>=2.4)", "pytest-cov", "pytest-enabler (>=2.2)", "pytest-mypy", "pytest-ruff (>=0.2.1)", "zipp (>=3.17)"] [[package]] name = "iniconfig" @@ -1579,13 +1610,13 @@ files = [ [[package]] name = "ipykernel" -version = "6.29.1" +version = "6.29.4" description = "IPython Kernel for Jupyter" optional = false python-versions = ">=3.8" files = [ - {file = "ipykernel-6.29.1-py3-none-any.whl", hash = "sha256:e5dfba210fc9da74a5dae8fa6c41f816e11bd18d10381b2517d9a0d57cc987c4"}, - {file = "ipykernel-6.29.1.tar.gz", hash = "sha256:1547352b32da95a2761011a8dac2af930c26a0703dfa07690d16b7d74dac0ba1"}, + {file = "ipykernel-6.29.4-py3-none-any.whl", hash = "sha256:1181e653d95c6808039c509ef8e67c4126b3b3af7781496c7cbfb5ed938a27da"}, + {file = "ipykernel-6.29.4.tar.gz", hash = "sha256:3d44070060f9475ac2092b760123fadf105d2e2493c24848b6691a7c4f42af5c"}, ] [package.dependencies] @@ -1608,7 +1639,7 @@ cov = ["coverage[toml]", "curio", "matplotlib", "pytest-cov", "trio"] docs = ["myst-parser", "pydata-sphinx-theme", "sphinx", "sphinx-autodoc-typehints", "sphinxcontrib-github-alt", "sphinxcontrib-spelling", "trio"] pyqt5 = ["pyqt5"] pyside6 = ["pyside6"] -test = ["flaky", "ipyparallel", "pre-commit", "pytest (>=7.0)", "pytest-asyncio (==0.23.4)", "pytest-cov", "pytest-timeout"] +test = ["flaky", "ipyparallel", "pre-commit", "pytest (>=7.0)", "pytest-asyncio (>=0.23.5)", "pytest-cov", "pytest-timeout"] [[package]] name = "ipython" @@ -1668,13 +1699,13 @@ testing = ["Django", "attrs", "colorama", "docopt", "pytest (<7.0.0)"] [[package]] name = "jinja2" -version = "3.1.3" +version = "3.1.4" description = "A very fast and expressive template engine." optional = false python-versions = ">=3.7" files = [ - {file = "Jinja2-3.1.3-py3-none-any.whl", hash = "sha256:7d6d50dd97d52cbc355597bd845fabfbac3f551e1f99619e39a35ce8c370b5fa"}, - {file = "Jinja2-3.1.3.tar.gz", hash = "sha256:ac8bd6544d4bb2c9792bf3a159e80bba8fda7f07e81bc3aed565432d5925ba90"}, + {file = "jinja2-3.1.4-py3-none-any.whl", hash = "sha256:bc5dd2abb727a5319567b7a813e6a2e7318c39f4f487cfe6c89c6f9c7d25197d"}, + {file = "jinja2-3.1.4.tar.gz", hash = "sha256:4a3aee7acbbe7303aede8e9648d13b8bf88a429282aa6122a993f0ac800cb369"}, ] [package.dependencies] @@ -1685,13 +1716,13 @@ i18n = ["Babel (>=2.7)"] [[package]] name = "jsonschema" -version = "4.21.1" +version = "4.22.0" description = "An implementation of JSON Schema validation for Python" optional = false python-versions = ">=3.8" files = [ - {file = "jsonschema-4.21.1-py3-none-any.whl", hash = "sha256:7996507afae316306f9e2290407761157c6f78002dcf7419acb99822143d1c6f"}, - {file = "jsonschema-4.21.1.tar.gz", hash = "sha256:85727c00279f5fa6bedbe6238d2aa6403bedd8b4864ab11207d07df3cc1b2ee5"}, + {file = "jsonschema-4.22.0-py3-none-any.whl", hash = "sha256:ff4cfd6b1367a40e7bc6411caec72effadd3db0bbe5017de188f2d6108335802"}, + {file = "jsonschema-4.22.0.tar.gz", hash = "sha256:5b22d434a45935119af990552c862e5d6d564e8f6601206b305a61fdf661a2b7"}, ] [package.dependencies] @@ -1720,13 +1751,13 @@ referencing = ">=0.31.0" [[package]] name = "jupyter-client" -version = "8.6.0" +version = "8.6.2" description = "Jupyter protocol implementation and client libraries" optional = false python-versions = ">=3.8" files = [ - {file = "jupyter_client-8.6.0-py3-none-any.whl", hash = "sha256:909c474dbe62582ae62b758bca86d6518c85234bdee2d908c778db6d72f39d99"}, - {file = "jupyter_client-8.6.0.tar.gz", hash = "sha256:0642244bb83b4764ae60d07e010e15f0e2d275ec4e918a8f7b80fbbef3ca60c7"}, + {file = "jupyter_client-8.6.2-py3-none-any.whl", hash = "sha256:50cbc5c66fd1b8f65ecb66bc490ab73217993632809b6e505687de18e9dea39f"}, + {file = "jupyter_client-8.6.2.tar.gz", hash = "sha256:2bda14d55ee5ba58552a8c53ae43d215ad9868853489213f37da060ced54d8df"}, ] [package.dependencies] @@ -1739,17 +1770,17 @@ traitlets = ">=5.3" [package.extras] docs = ["ipykernel", "myst-parser", "pydata-sphinx-theme", "sphinx (>=4)", "sphinx-autodoc-typehints", "sphinxcontrib-github-alt", "sphinxcontrib-spelling"] -test = ["coverage", "ipykernel (>=6.14)", "mypy", "paramiko", "pre-commit", "pytest", "pytest-cov", "pytest-jupyter[client] (>=0.4.1)", "pytest-timeout"] +test = ["coverage", "ipykernel (>=6.14)", "mypy", "paramiko", "pre-commit", "pytest (<8.2.0)", "pytest-cov", "pytest-jupyter[client] (>=0.4.1)", "pytest-timeout"] [[package]] name = "jupyter-core" -version = "5.7.1" +version = "5.7.2" description = "Jupyter core package. A base package on which Jupyter projects rely." optional = false python-versions = ">=3.8" files = [ - {file = "jupyter_core-5.7.1-py3-none-any.whl", hash = "sha256:c65c82126453a723a2804aa52409930434598fd9d35091d63dfb919d2b765bb7"}, - {file = "jupyter_core-5.7.1.tar.gz", hash = "sha256:de61a9d7fc71240f688b2fb5ab659fbb56979458dc66a71decd098e03c79e218"}, + {file = "jupyter_core-5.7.2-py3-none-any.whl", hash = "sha256:4f7315d2f6b4bcf2e3e7cb6e46772eba760ae459cd1f59d29eb57b0a01bd7409"}, + {file = "jupyter_core-5.7.2.tar.gz", hash = "sha256:aa5f8d32bbf6b431ac830496da7392035d6f61b4f54872f15c4bd2a9c3f536d9"}, ] [package.dependencies] @@ -1759,7 +1790,7 @@ traitlets = ">=5.3" [package.extras] docs = ["myst-parser", "pydata-sphinx-theme", "sphinx-autodoc-typehints", "sphinxcontrib-github-alt", "sphinxcontrib-spelling", "traitlets"] -test = ["ipykernel", "pre-commit", "pytest", "pytest-cov", "pytest-timeout"] +test = ["ipykernel", "pre-commit", "pytest (<8)", "pytest-cov", "pytest-timeout"] [[package]] name = "kiwisolver" @@ -1987,13 +2018,13 @@ tests = ["psutil", "pytest (!=3.3.0)", "pytest-cov"] [[package]] name = "markdown" -version = "3.5.2" +version = "3.6" description = "Python implementation of John Gruber's Markdown." optional = false python-versions = ">=3.8" files = [ - {file = "Markdown-3.5.2-py3-none-any.whl", hash = "sha256:d43323865d89fc0cb9b20c75fc8ad313af307cc087e84b657d9eec768eddeadd"}, - {file = "Markdown-3.5.2.tar.gz", hash = "sha256:e1ac7b3dc550ee80e602e71c1d168002f062e49f1b11e26a36264dafd4df2ef8"}, + {file = "Markdown-3.6-py3-none-any.whl", hash = "sha256:48f276f4d8cfb8ce6527c8f79e2ee29708508bf4d40aa410fbc3b4ee832c850f"}, + {file = "Markdown-3.6.tar.gz", hash = "sha256:ed4f41f6daecbeeb96e576ce414c41d2d876daa9a16cb35fa8ed8c2ddfad0224"}, ] [package.dependencies] @@ -2098,39 +2129,40 @@ files = [ [[package]] name = "matplotlib" -version = "3.8.2" +version = "3.9.0" description = "Python plotting package" optional = false python-versions = ">=3.9" files = [ - {file = "matplotlib-3.8.2-cp310-cp310-macosx_10_12_x86_64.whl", hash = "sha256:09796f89fb71a0c0e1e2f4bdaf63fb2cefc84446bb963ecdeb40dfee7dfa98c7"}, - {file = "matplotlib-3.8.2-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:6f9c6976748a25e8b9be51ea028df49b8e561eed7809146da7a47dbecebab367"}, - {file = "matplotlib-3.8.2-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:b78e4f2cedf303869b782071b55fdde5987fda3038e9d09e58c91cc261b5ad18"}, - {file = "matplotlib-3.8.2-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:4e208f46cf6576a7624195aa047cb344a7f802e113bb1a06cfd4bee431de5e31"}, - {file = "matplotlib-3.8.2-cp310-cp310-musllinux_1_1_x86_64.whl", hash = "sha256:46a569130ff53798ea5f50afce7406e91fdc471ca1e0e26ba976a8c734c9427a"}, - {file = "matplotlib-3.8.2-cp310-cp310-win_amd64.whl", hash = "sha256:830f00640c965c5b7f6bc32f0d4ce0c36dfe0379f7dd65b07a00c801713ec40a"}, - {file = "matplotlib-3.8.2-cp311-cp311-macosx_10_12_x86_64.whl", hash = "sha256:d86593ccf546223eb75a39b44c32788e6f6440d13cfc4750c1c15d0fcb850b63"}, - {file = "matplotlib-3.8.2-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:9a5430836811b7652991939012f43d2808a2db9b64ee240387e8c43e2e5578c8"}, - {file = "matplotlib-3.8.2-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:b9576723858a78751d5aacd2497b8aef29ffea6d1c95981505877f7ac28215c6"}, - {file = "matplotlib-3.8.2-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:5ba9cbd8ac6cf422f3102622b20f8552d601bf8837e49a3afed188d560152788"}, - {file = "matplotlib-3.8.2-cp311-cp311-musllinux_1_1_x86_64.whl", hash = "sha256:03f9d160a29e0b65c0790bb07f4f45d6a181b1ac33eb1bb0dd225986450148f0"}, - {file = "matplotlib-3.8.2-cp311-cp311-win_amd64.whl", hash = "sha256:3773002da767f0a9323ba1a9b9b5d00d6257dbd2a93107233167cfb581f64717"}, - {file = "matplotlib-3.8.2-cp312-cp312-macosx_10_12_x86_64.whl", hash = "sha256:4c318c1e95e2f5926fba326f68177dee364aa791d6df022ceb91b8221bd0a627"}, - {file = "matplotlib-3.8.2-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:091275d18d942cf1ee9609c830a1bc36610607d8223b1b981c37d5c9fc3e46a4"}, - {file = "matplotlib-3.8.2-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:1b0f3b8ea0e99e233a4bcc44590f01604840d833c280ebb8fe5554fd3e6cfe8d"}, - {file = "matplotlib-3.8.2-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:d7b1704a530395aaf73912be741c04d181f82ca78084fbd80bc737be04848331"}, - {file = "matplotlib-3.8.2-cp312-cp312-musllinux_1_1_x86_64.whl", hash = "sha256:533b0e3b0c6768eef8cbe4b583731ce25a91ab54a22f830db2b031e83cca9213"}, - {file = "matplotlib-3.8.2-cp312-cp312-win_amd64.whl", hash = "sha256:0f4fc5d72b75e2c18e55eb32292659cf731d9d5b312a6eb036506304f4675630"}, - {file = "matplotlib-3.8.2-cp39-cp39-macosx_10_12_x86_64.whl", hash = "sha256:deaed9ad4da0b1aea77fe0aa0cebb9ef611c70b3177be936a95e5d01fa05094f"}, - {file = "matplotlib-3.8.2-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:172f4d0fbac3383d39164c6caafd3255ce6fa58f08fc392513a0b1d3b89c4f89"}, - {file = "matplotlib-3.8.2-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:c7d36c2209d9136cd8e02fab1c0ddc185ce79bc914c45054a9f514e44c787917"}, - {file = "matplotlib-3.8.2-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:5864bdd7da445e4e5e011b199bb67168cdad10b501750367c496420f2ad00843"}, - {file = "matplotlib-3.8.2-cp39-cp39-musllinux_1_1_x86_64.whl", hash = "sha256:ef8345b48e95cee45ff25192ed1f4857273117917a4dcd48e3905619bcd9c9b8"}, - {file = "matplotlib-3.8.2-cp39-cp39-win_amd64.whl", hash = "sha256:7c48d9e221b637c017232e3760ed30b4e8d5dfd081daf327e829bf2a72c731b4"}, - {file = "matplotlib-3.8.2-pp39-pypy39_pp73-macosx_10_12_x86_64.whl", hash = "sha256:aa11b3c6928a1e496c1a79917d51d4cd5d04f8a2e75f21df4949eeefdf697f4b"}, - {file = "matplotlib-3.8.2-pp39-pypy39_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:d1095fecf99eeb7384dabad4bf44b965f929a5f6079654b681193edf7169ec20"}, - {file = "matplotlib-3.8.2-pp39-pypy39_pp73-win_amd64.whl", hash = "sha256:bddfb1db89bfaa855912261c805bd0e10218923cc262b9159a49c29a7a1c1afa"}, - {file = "matplotlib-3.8.2.tar.gz", hash = "sha256:01a978b871b881ee76017152f1f1a0cbf6bd5f7b8ff8c96df0df1bd57d8755a1"}, + {file = "matplotlib-3.9.0-cp310-cp310-macosx_10_12_x86_64.whl", hash = "sha256:2bcee1dffaf60fe7656183ac2190bd630842ff87b3153afb3e384d966b57fe56"}, + {file = "matplotlib-3.9.0-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:3f988bafb0fa39d1074ddd5bacd958c853e11def40800c5824556eb630f94d3b"}, + {file = "matplotlib-3.9.0-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:fe428e191ea016bb278758c8ee82a8129c51d81d8c4bc0846c09e7e8e9057241"}, + {file = "matplotlib-3.9.0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:eaf3978060a106fab40c328778b148f590e27f6fa3cd15a19d6892575bce387d"}, + {file = "matplotlib-3.9.0-cp310-cp310-musllinux_1_1_x86_64.whl", hash = "sha256:2e7f03e5cbbfacdd48c8ea394d365d91ee8f3cae7e6ec611409927b5ed997ee4"}, + {file = "matplotlib-3.9.0-cp310-cp310-win_amd64.whl", hash = "sha256:13beb4840317d45ffd4183a778685e215939be7b08616f431c7795276e067463"}, + {file = "matplotlib-3.9.0-cp311-cp311-macosx_10_12_x86_64.whl", hash = "sha256:063af8587fceeac13b0936c42a2b6c732c2ab1c98d38abc3337e430e1ff75e38"}, + {file = "matplotlib-3.9.0-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:9a2fa6d899e17ddca6d6526cf6e7ba677738bf2a6a9590d702c277204a7c6152"}, + {file = "matplotlib-3.9.0-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:550cdda3adbd596078cca7d13ed50b77879104e2e46392dcd7c75259d8f00e85"}, + {file = "matplotlib-3.9.0-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:76cce0f31b351e3551d1f3779420cf8f6ec0d4a8cf9c0237a3b549fd28eb4abb"}, + {file = "matplotlib-3.9.0-cp311-cp311-musllinux_1_1_x86_64.whl", hash = "sha256:c53aeb514ccbbcbab55a27f912d79ea30ab21ee0531ee2c09f13800efb272674"}, + {file = "matplotlib-3.9.0-cp311-cp311-win_amd64.whl", hash = "sha256:a5be985db2596d761cdf0c2eaf52396f26e6a64ab46bd8cd810c48972349d1be"}, + {file = "matplotlib-3.9.0-cp312-cp312-macosx_10_12_x86_64.whl", hash = "sha256:c79f3a585f1368da6049318bdf1f85568d8d04b2e89fc24b7e02cc9b62017382"}, + {file = "matplotlib-3.9.0-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:bdd1ecbe268eb3e7653e04f451635f0fb0f77f07fd070242b44c076c9106da84"}, + {file = "matplotlib-3.9.0-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:d38e85a1a6d732f645f1403ce5e6727fd9418cd4574521d5803d3d94911038e5"}, + {file = "matplotlib-3.9.0-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:0a490715b3b9984fa609116481b22178348c1a220a4499cda79132000a79b4db"}, + {file = "matplotlib-3.9.0-cp312-cp312-musllinux_1_1_x86_64.whl", hash = "sha256:8146ce83cbc5dc71c223a74a1996d446cd35cfb6a04b683e1446b7e6c73603b7"}, + {file = "matplotlib-3.9.0-cp312-cp312-win_amd64.whl", hash = "sha256:d91a4ffc587bacf5c4ce4ecfe4bcd23a4b675e76315f2866e588686cc97fccdf"}, + {file = "matplotlib-3.9.0-cp39-cp39-macosx_10_12_x86_64.whl", hash = "sha256:616fabf4981a3b3c5a15cd95eba359c8489c4e20e03717aea42866d8d0465956"}, + {file = "matplotlib-3.9.0-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:cd53c79fd02f1c1808d2cfc87dd3cf4dbc63c5244a58ee7944497107469c8d8a"}, + {file = "matplotlib-3.9.0-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:06a478f0d67636554fa78558cfbcd7b9dba85b51f5c3b5a0c9be49010cf5f321"}, + {file = "matplotlib-3.9.0-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:81c40af649d19c85f8073e25e5806926986806fa6d54be506fbf02aef47d5a89"}, + {file = "matplotlib-3.9.0-cp39-cp39-musllinux_1_1_x86_64.whl", hash = "sha256:52146fc3bd7813cc784562cb93a15788be0b2875c4655e2cc6ea646bfa30344b"}, + {file = "matplotlib-3.9.0-cp39-cp39-win_amd64.whl", hash = "sha256:0fc51eaa5262553868461c083d9adadb11a6017315f3a757fc45ec6ec5f02888"}, + {file = "matplotlib-3.9.0-pp39-pypy39_pp73-macosx_10_12_x86_64.whl", hash = "sha256:bd4f2831168afac55b881db82a7730992aa41c4f007f1913465fb182d6fb20c0"}, + {file = "matplotlib-3.9.0-pp39-pypy39_pp73-macosx_11_0_arm64.whl", hash = "sha256:290d304e59be2b33ef5c2d768d0237f5bd132986bdcc66f80bc9bcc300066a03"}, + {file = "matplotlib-3.9.0-pp39-pypy39_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:7ff2e239c26be4f24bfa45860c20ffccd118d270c5b5d081fa4ea409b5469fcd"}, + {file = "matplotlib-3.9.0-pp39-pypy39_pp73-win_amd64.whl", hash = "sha256:af4001b7cae70f7eaacfb063db605280058246de590fa7874f00f62259f2df7e"}, + {file = "matplotlib-3.9.0.tar.gz", hash = "sha256:e6d29ea6c19e34b30fb7d88b7081f869a03014f66fe06d62cc77d5a6ea88ed7a"}, ] [package.dependencies] @@ -2139,21 +2171,24 @@ cycler = ">=0.10" fonttools = ">=4.22.0" importlib-resources = {version = ">=3.2.0", markers = "python_version < \"3.10\""} kiwisolver = ">=1.3.1" -numpy = ">=1.21,<2" +numpy = ">=1.23" packaging = ">=20.0" pillow = ">=8" pyparsing = ">=2.3.1" python-dateutil = ">=2.7" +[package.extras] +dev = ["meson-python (>=0.13.1)", "numpy (>=1.25)", "pybind11 (>=2.6)", "setuptools (>=64)", "setuptools_scm (>=7)"] + [[package]] name = "matplotlib-inline" -version = "0.1.6" +version = "0.1.7" description = "Inline Matplotlib backend for Jupyter" optional = false -python-versions = ">=3.5" +python-versions = ">=3.8" files = [ - {file = "matplotlib-inline-0.1.6.tar.gz", hash = "sha256:f887e5f10ba98e8d2b150ddcf4702c1e5f8b3a20005eb0f74bfdbd360ee6f304"}, - {file = "matplotlib_inline-0.1.6-py3-none-any.whl", hash = "sha256:f1f41aab5328aa5aaea9b16d083b128102f8712542f819fe7e6a420ff581b311"}, + {file = "matplotlib_inline-0.1.7-py3-none-any.whl", hash = "sha256:df192d39a4ff8f21b1895d72e6a13f5fcc5099f00fa84384e0ea28c2cc0653ca"}, + {file = "matplotlib_inline-0.1.7.tar.gz", hash = "sha256:8423b23ec666be3d16e16b60bdd8ac4e86e840ebd1dd11a30b9f117f2fa0ab90"}, ] [package.dependencies] @@ -2161,13 +2196,13 @@ traitlets = "*" [[package]] name = "mdit-py-plugins" -version = "0.4.0" +version = "0.4.1" description = "Collection of plugins for markdown-it-py" optional = false python-versions = ">=3.8" files = [ - {file = "mdit_py_plugins-0.4.0-py3-none-any.whl", hash = "sha256:b51b3bb70691f57f974e257e367107857a93b36f322a9e6d44ca5bf28ec2def9"}, - {file = "mdit_py_plugins-0.4.0.tar.gz", hash = "sha256:d8ab27e9aed6c38aa716819fedfde15ca275715955f8a185a8e1cf90fb1d2c1b"}, + {file = "mdit_py_plugins-0.4.1-py3-none-any.whl", hash = "sha256:1020dfe4e6bfc2c79fb49ae4e3f5b297f5ccd20f010187acc52af2921e27dc6a"}, + {file = "mdit_py_plugins-0.4.1.tar.gz", hash = "sha256:834b8ac23d1cd60cec703646ffd22ae97b7955a6d596eb1d304be1e251ae499c"}, ] [package.dependencies] @@ -2202,59 +2237,77 @@ files = [ [[package]] name = "mkdocs" -version = "1.5.3" +version = "1.6.0" description = "Project documentation with Markdown." optional = false -python-versions = ">=3.7" +python-versions = ">=3.8" files = [ - {file = "mkdocs-1.5.3-py3-none-any.whl", hash = "sha256:3b3a78e736b31158d64dbb2f8ba29bd46a379d0c6e324c2246c3bc3d2189cfc1"}, - {file = "mkdocs-1.5.3.tar.gz", hash = "sha256:eb7c99214dcb945313ba30426c2451b735992c73c2e10838f76d09e39ff4d0e2"}, + {file = "mkdocs-1.6.0-py3-none-any.whl", hash = "sha256:1eb5cb7676b7d89323e62b56235010216319217d4af5ddc543a91beb8d125ea7"}, + {file = "mkdocs-1.6.0.tar.gz", hash = "sha256:a73f735824ef83a4f3bcb7a231dcab23f5a838f88b7efc54a0eef5fbdbc3c512"}, ] [package.dependencies] click = ">=7.0" colorama = {version = ">=0.4", markers = "platform_system == \"Windows\""} ghp-import = ">=1.0" -importlib-metadata = {version = ">=4.3", markers = "python_version < \"3.10\""} +importlib-metadata = {version = ">=4.4", markers = "python_version < \"3.10\""} jinja2 = ">=2.11.1" -markdown = ">=3.2.1" +markdown = ">=3.3.6" markupsafe = ">=2.0.1" mergedeep = ">=1.3.4" +mkdocs-get-deps = ">=0.2.0" packaging = ">=20.5" pathspec = ">=0.11.1" -platformdirs = ">=2.2.0" pyyaml = ">=5.1" pyyaml-env-tag = ">=0.1" watchdog = ">=2.0" [package.extras] i18n = ["babel (>=2.9.0)"] -min-versions = ["babel (==2.9.0)", "click (==7.0)", "colorama (==0.4)", "ghp-import (==1.0)", "importlib-metadata (==4.3)", "jinja2 (==2.11.1)", "markdown (==3.2.1)", "markupsafe (==2.0.1)", "mergedeep (==1.3.4)", "packaging (==20.5)", "pathspec (==0.11.1)", "platformdirs (==2.2.0)", "pyyaml (==5.1)", "pyyaml-env-tag (==0.1)", "typing-extensions (==3.10)", "watchdog (==2.0)"] +min-versions = ["babel (==2.9.0)", "click (==7.0)", "colorama (==0.4)", "ghp-import (==1.0)", "importlib-metadata (==4.4)", "jinja2 (==2.11.1)", "markdown (==3.3.6)", "markupsafe (==2.0.1)", "mergedeep (==1.3.4)", "mkdocs-get-deps (==0.2.0)", "packaging (==20.5)", "pathspec (==0.11.1)", "pyyaml (==5.1)", "pyyaml-env-tag (==0.1)", "watchdog (==2.0)"] [[package]] name = "mkdocs-autorefs" -version = "0.5.0" +version = "1.0.1" description = "Automatically link across pages in MkDocs." optional = false python-versions = ">=3.8" files = [ - {file = "mkdocs_autorefs-0.5.0-py3-none-any.whl", hash = "sha256:7930fcb8ac1249f10e683967aeaddc0af49d90702af111a5e390e8b20b3d97ff"}, - {file = "mkdocs_autorefs-0.5.0.tar.gz", hash = "sha256:9a5054a94c08d28855cfab967ada10ed5be76e2bfad642302a610b252c3274c0"}, + {file = "mkdocs_autorefs-1.0.1-py3-none-any.whl", hash = "sha256:aacdfae1ab197780fb7a2dac92ad8a3d8f7ca8049a9cbe56a4218cd52e8da570"}, + {file = "mkdocs_autorefs-1.0.1.tar.gz", hash = "sha256:f684edf847eced40b570b57846b15f0bf57fb93ac2c510450775dcf16accb971"}, ] [package.dependencies] Markdown = ">=3.3" +markupsafe = ">=2.0.1" mkdocs = ">=1.1" +[[package]] +name = "mkdocs-get-deps" +version = "0.2.0" +description = "MkDocs extension that lists all dependencies according to a mkdocs.yml file" +optional = false +python-versions = ">=3.8" +files = [ + {file = "mkdocs_get_deps-0.2.0-py3-none-any.whl", hash = "sha256:2bf11d0b133e77a0dd036abeeb06dec8775e46efa526dc70667d8863eefc6134"}, + {file = "mkdocs_get_deps-0.2.0.tar.gz", hash = "sha256:162b3d129c7fad9b19abfdcb9c1458a651628e4b1dea628ac68790fb3061c60c"}, +] + +[package.dependencies] +importlib-metadata = {version = ">=4.3", markers = "python_version < \"3.10\""} +mergedeep = ">=1.3.4" +platformdirs = ">=2.2.0" +pyyaml = ">=5.1" + [[package]] name = "mkdocs-material" -version = "9.5.7" +version = "9.5.24" description = "Documentation that simply works" optional = false python-versions = ">=3.8" files = [ - {file = "mkdocs_material-9.5.7-py3-none-any.whl", hash = "sha256:0be8ce8bcfebb52bae9b00cf9b851df45b8a92d629afcfd7f2c09b2dfa155ea3"}, - {file = "mkdocs_material-9.5.7.tar.gz", hash = "sha256:16110292575d88a338d2961f3cb665cf12943ff8829e551a9b364f24019e46af"}, + {file = "mkdocs_material-9.5.24-py3-none-any.whl", hash = "sha256:e12cd75954c535b61e716f359cf2a5056bf4514889d17161fdebd5df4b0153c6"}, + {file = "mkdocs_material-9.5.24.tar.gz", hash = "sha256:02d5aaba0ee755e707c3ef6e748f9acb7b3011187c0ea766db31af8905078a34"}, ] [package.dependencies] @@ -2262,7 +2315,7 @@ babel = ">=2.10,<3.0" colorama = ">=0.4,<1.0" jinja2 = ">=3.0,<4.0" markdown = ">=3.2,<4.0" -mkdocs = ">=1.5.3,<1.6.0" +mkdocs = ">=1.6,<2.0" mkdocs-material-extensions = ">=1.3,<2.0" paginate = ">=0.5,<1.0" pygments = ">=2.16,<3.0" @@ -2271,7 +2324,7 @@ regex = ">=2022.4" requests = ">=2.26,<3.0" [package.extras] -git = ["mkdocs-git-committers-plugin-2 (>=1.1,<2.0)", "mkdocs-git-revision-date-localized-plugin (>=1.2,<2.0)"] +git = ["mkdocs-git-committers-plugin-2 (>=1.1,<2.0)", "mkdocs-git-revision-date-localized-plugin (>=1.2.4,<2.0)"] imaging = ["cairosvg (>=2.6,<3.0)", "pillow (>=10.2,<11.0)"] recommended = ["mkdocs-minify-plugin (>=0.7,<1.0)", "mkdocs-redirects (>=1.2,<2.0)", "mkdocs-rss-plugin (>=1.6,<2.0)"] @@ -2288,13 +2341,13 @@ files = [ [[package]] name = "mkdocstrings" -version = "0.24.0" +version = "0.25.1" description = "Automatic documentation from sources, for MkDocs." optional = false python-versions = ">=3.8" files = [ - {file = "mkdocstrings-0.24.0-py3-none-any.whl", hash = "sha256:f4908560c10f587326d8f5165d1908817b2e280bbf707607f601c996366a2264"}, - {file = "mkdocstrings-0.24.0.tar.gz", hash = "sha256:222b1165be41257b494a9d29b14135d2b7ca43f38161d5b10caae03b87bd4f7e"}, + {file = "mkdocstrings-0.25.1-py3-none-any.whl", hash = "sha256:da01fcc2670ad61888e8fe5b60afe9fee5781017d67431996832d63e887c2e51"}, + {file = "mkdocstrings-0.25.1.tar.gz", hash = "sha256:c3a2515f31577f311a9ee58d089e4c51fc6046dbd9e9b4c3de4c3194667fe9bf"}, ] [package.dependencies] @@ -2316,82 +2369,82 @@ python-legacy = ["mkdocstrings-python-legacy (>=0.2.1)"] [[package]] name = "mkdocstrings-python" -version = "1.8.0" +version = "1.10.3" description = "A Python handler for mkdocstrings." optional = false python-versions = ">=3.8" files = [ - {file = "mkdocstrings_python-1.8.0-py3-none-any.whl", hash = "sha256:4209970cc90bec194568682a535848a8d8489516c6ed4adbe58bbc67b699ca9d"}, - {file = "mkdocstrings_python-1.8.0.tar.gz", hash = "sha256:1488bddf50ee42c07d9a488dddc197f8e8999c2899687043ec5dd1643d057192"}, + {file = "mkdocstrings_python-1.10.3-py3-none-any.whl", hash = "sha256:11ff6d21d3818fb03af82c3ea6225b1534837e17f790aa5f09626524171f949b"}, + {file = "mkdocstrings_python-1.10.3.tar.gz", hash = "sha256:321cf9c732907ab2b1fedaafa28765eaa089d89320f35f7206d00ea266889d03"}, ] [package.dependencies] -griffe = ">=0.37" -mkdocstrings = ">=0.20" +griffe = ">=0.44" +mkdocstrings = ">=0.25" [[package]] name = "msgpack" -version = "1.0.7" +version = "1.0.8" description = "MessagePack serializer" optional = false python-versions = ">=3.8" files = [ - {file = "msgpack-1.0.7-cp310-cp310-macosx_10_9_universal2.whl", hash = "sha256:04ad6069c86e531682f9e1e71b71c1c3937d6014a7c3e9edd2aa81ad58842862"}, - {file = "msgpack-1.0.7-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:cca1b62fe70d761a282496b96a5e51c44c213e410a964bdffe0928e611368329"}, - {file = "msgpack-1.0.7-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:e50ebce52f41370707f1e21a59514e3375e3edd6e1832f5e5235237db933c98b"}, - {file = "msgpack-1.0.7-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:4a7b4f35de6a304b5533c238bee86b670b75b03d31b7797929caa7a624b5dda6"}, - {file = "msgpack-1.0.7-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:28efb066cde83c479dfe5a48141a53bc7e5f13f785b92ddde336c716663039ee"}, - {file = "msgpack-1.0.7-cp310-cp310-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:4cb14ce54d9b857be9591ac364cb08dc2d6a5c4318c1182cb1d02274029d590d"}, - {file = "msgpack-1.0.7-cp310-cp310-musllinux_1_1_aarch64.whl", hash = "sha256:b573a43ef7c368ba4ea06050a957c2a7550f729c31f11dd616d2ac4aba99888d"}, - {file = "msgpack-1.0.7-cp310-cp310-musllinux_1_1_i686.whl", hash = "sha256:ccf9a39706b604d884d2cb1e27fe973bc55f2890c52f38df742bc1d79ab9f5e1"}, - {file = "msgpack-1.0.7-cp310-cp310-musllinux_1_1_x86_64.whl", hash = "sha256:cb70766519500281815dfd7a87d3a178acf7ce95390544b8c90587d76b227681"}, - {file = "msgpack-1.0.7-cp310-cp310-win32.whl", hash = "sha256:b610ff0f24e9f11c9ae653c67ff8cc03c075131401b3e5ef4b82570d1728f8a9"}, - {file = "msgpack-1.0.7-cp310-cp310-win_amd64.whl", hash = "sha256:a40821a89dc373d6427e2b44b572efc36a2778d3f543299e2f24eb1a5de65415"}, - {file = "msgpack-1.0.7-cp311-cp311-macosx_10_9_universal2.whl", hash = "sha256:576eb384292b139821c41995523654ad82d1916da6a60cff129c715a6223ea84"}, - {file = "msgpack-1.0.7-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:730076207cb816138cf1af7f7237b208340a2c5e749707457d70705715c93b93"}, - {file = "msgpack-1.0.7-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:85765fdf4b27eb5086f05ac0491090fc76f4f2b28e09d9350c31aac25a5aaff8"}, - {file = "msgpack-1.0.7-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:3476fae43db72bd11f29a5147ae2f3cb22e2f1a91d575ef130d2bf49afd21c46"}, - {file = "msgpack-1.0.7-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:6d4c80667de2e36970ebf74f42d1088cc9ee7ef5f4e8c35eee1b40eafd33ca5b"}, - {file = "msgpack-1.0.7-cp311-cp311-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:5b0bf0effb196ed76b7ad883848143427a73c355ae8e569fa538365064188b8e"}, - {file = "msgpack-1.0.7-cp311-cp311-musllinux_1_1_aarch64.whl", hash = "sha256:f9a7c509542db4eceed3dcf21ee5267ab565a83555c9b88a8109dcecc4709002"}, - {file = "msgpack-1.0.7-cp311-cp311-musllinux_1_1_i686.whl", hash = "sha256:84b0daf226913133f899ea9b30618722d45feffa67e4fe867b0b5ae83a34060c"}, - {file = "msgpack-1.0.7-cp311-cp311-musllinux_1_1_x86_64.whl", hash = "sha256:ec79ff6159dffcc30853b2ad612ed572af86c92b5168aa3fc01a67b0fa40665e"}, - {file = "msgpack-1.0.7-cp311-cp311-win32.whl", hash = "sha256:3e7bf4442b310ff154b7bb9d81eb2c016b7d597e364f97d72b1acc3817a0fdc1"}, - {file = "msgpack-1.0.7-cp311-cp311-win_amd64.whl", hash = "sha256:3f0c8c6dfa6605ab8ff0611995ee30d4f9fcff89966cf562733b4008a3d60d82"}, - {file = "msgpack-1.0.7-cp312-cp312-macosx_10_9_universal2.whl", hash = "sha256:f0936e08e0003f66bfd97e74ee530427707297b0d0361247e9b4f59ab78ddc8b"}, - {file = "msgpack-1.0.7-cp312-cp312-macosx_10_9_x86_64.whl", hash = "sha256:98bbd754a422a0b123c66a4c341de0474cad4a5c10c164ceed6ea090f3563db4"}, - {file = "msgpack-1.0.7-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:b291f0ee7961a597cbbcc77709374087fa2a9afe7bdb6a40dbbd9b127e79afee"}, - {file = "msgpack-1.0.7-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:ebbbba226f0a108a7366bf4b59bf0f30a12fd5e75100c630267d94d7f0ad20e5"}, - {file = "msgpack-1.0.7-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:1e2d69948e4132813b8d1131f29f9101bc2c915f26089a6d632001a5c1349672"}, - {file = "msgpack-1.0.7-cp312-cp312-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:bdf38ba2d393c7911ae989c3bbba510ebbcdf4ecbdbfec36272abe350c454075"}, - {file = "msgpack-1.0.7-cp312-cp312-musllinux_1_1_aarch64.whl", hash = "sha256:993584fc821c58d5993521bfdcd31a4adf025c7d745bbd4d12ccfecf695af5ba"}, - {file = "msgpack-1.0.7-cp312-cp312-musllinux_1_1_i686.whl", hash = "sha256:52700dc63a4676669b341ba33520f4d6e43d3ca58d422e22ba66d1736b0a6e4c"}, - {file = "msgpack-1.0.7-cp312-cp312-musllinux_1_1_x86_64.whl", hash = "sha256:e45ae4927759289c30ccba8d9fdce62bb414977ba158286b5ddaf8df2cddb5c5"}, - {file = "msgpack-1.0.7-cp312-cp312-win32.whl", hash = "sha256:27dcd6f46a21c18fa5e5deed92a43d4554e3df8d8ca5a47bf0615d6a5f39dbc9"}, - {file = "msgpack-1.0.7-cp312-cp312-win_amd64.whl", hash = "sha256:7687e22a31e976a0e7fc99c2f4d11ca45eff652a81eb8c8085e9609298916dcf"}, - {file = "msgpack-1.0.7-cp38-cp38-macosx_10_9_universal2.whl", hash = "sha256:5b6ccc0c85916998d788b295765ea0e9cb9aac7e4a8ed71d12e7d8ac31c23c95"}, - {file = "msgpack-1.0.7-cp38-cp38-macosx_10_9_x86_64.whl", hash = "sha256:235a31ec7db685f5c82233bddf9858748b89b8119bf4538d514536c485c15fe0"}, - {file = "msgpack-1.0.7-cp38-cp38-macosx_11_0_arm64.whl", hash = "sha256:cab3db8bab4b7e635c1c97270d7a4b2a90c070b33cbc00c99ef3f9be03d3e1f7"}, - {file = "msgpack-1.0.7-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:0bfdd914e55e0d2c9e1526de210f6fe8ffe9705f2b1dfcc4aecc92a4cb4b533d"}, - {file = "msgpack-1.0.7-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:36e17c4592231a7dbd2ed09027823ab295d2791b3b1efb2aee874b10548b7524"}, - {file = "msgpack-1.0.7-cp38-cp38-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:38949d30b11ae5f95c3c91917ee7a6b239f5ec276f271f28638dec9156f82cfc"}, - {file = "msgpack-1.0.7-cp38-cp38-musllinux_1_1_aarch64.whl", hash = "sha256:ff1d0899f104f3921d94579a5638847f783c9b04f2d5f229392ca77fba5b82fc"}, - {file = "msgpack-1.0.7-cp38-cp38-musllinux_1_1_i686.whl", hash = "sha256:dc43f1ec66eb8440567186ae2f8c447d91e0372d793dfe8c222aec857b81a8cf"}, - {file = "msgpack-1.0.7-cp38-cp38-musllinux_1_1_x86_64.whl", hash = "sha256:dd632777ff3beaaf629f1ab4396caf7ba0bdd075d948a69460d13d44357aca4c"}, - {file = "msgpack-1.0.7-cp38-cp38-win32.whl", hash = "sha256:4e71bc4416de195d6e9b4ee93ad3f2f6b2ce11d042b4d7a7ee00bbe0358bd0c2"}, - {file = "msgpack-1.0.7-cp38-cp38-win_amd64.whl", hash = "sha256:8f5b234f567cf76ee489502ceb7165c2a5cecec081db2b37e35332b537f8157c"}, - {file = "msgpack-1.0.7-cp39-cp39-macosx_10_9_universal2.whl", hash = "sha256:bfef2bb6ef068827bbd021017a107194956918ab43ce4d6dc945ffa13efbc25f"}, - {file = "msgpack-1.0.7-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:484ae3240666ad34cfa31eea7b8c6cd2f1fdaae21d73ce2974211df099a95d81"}, - {file = "msgpack-1.0.7-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:3967e4ad1aa9da62fd53e346ed17d7b2e922cba5ab93bdd46febcac39be636fc"}, - {file = "msgpack-1.0.7-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:8dd178c4c80706546702c59529ffc005681bd6dc2ea234c450661b205445a34d"}, - {file = "msgpack-1.0.7-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:f6ffbc252eb0d229aeb2f9ad051200668fc3a9aaa8994e49f0cb2ffe2b7867e7"}, - {file = "msgpack-1.0.7-cp39-cp39-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:822ea70dc4018c7e6223f13affd1c5c30c0f5c12ac1f96cd8e9949acddb48a61"}, - {file = "msgpack-1.0.7-cp39-cp39-musllinux_1_1_aarch64.whl", hash = "sha256:384d779f0d6f1b110eae74cb0659d9aa6ff35aaf547b3955abf2ab4c901c4819"}, - {file = "msgpack-1.0.7-cp39-cp39-musllinux_1_1_i686.whl", hash = "sha256:f64e376cd20d3f030190e8c32e1c64582eba56ac6dc7d5b0b49a9d44021b52fd"}, - {file = "msgpack-1.0.7-cp39-cp39-musllinux_1_1_x86_64.whl", hash = "sha256:5ed82f5a7af3697b1c4786053736f24a0efd0a1b8a130d4c7bfee4b9ded0f08f"}, - {file = "msgpack-1.0.7-cp39-cp39-win32.whl", hash = "sha256:f26a07a6e877c76a88e3cecac8531908d980d3d5067ff69213653649ec0f60ad"}, - {file = "msgpack-1.0.7-cp39-cp39-win_amd64.whl", hash = "sha256:1dc93e8e4653bdb5910aed79f11e165c85732067614f180f70534f056da97db3"}, - {file = "msgpack-1.0.7.tar.gz", hash = "sha256:572efc93db7a4d27e404501975ca6d2d9775705c2d922390d878fcf768d92c87"}, + {file = "msgpack-1.0.8-cp310-cp310-macosx_10_9_universal2.whl", hash = "sha256:505fe3d03856ac7d215dbe005414bc28505d26f0c128906037e66d98c4e95868"}, + {file = "msgpack-1.0.8-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:e6b7842518a63a9f17107eb176320960ec095a8ee3b4420b5f688e24bf50c53c"}, + {file = "msgpack-1.0.8-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:376081f471a2ef24828b83a641a02c575d6103a3ad7fd7dade5486cad10ea659"}, + {file = "msgpack-1.0.8-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:5e390971d082dba073c05dbd56322427d3280b7cc8b53484c9377adfbae67dc2"}, + {file = "msgpack-1.0.8-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:00e073efcba9ea99db5acef3959efa45b52bc67b61b00823d2a1a6944bf45982"}, + {file = "msgpack-1.0.8-cp310-cp310-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:82d92c773fbc6942a7a8b520d22c11cfc8fd83bba86116bfcf962c2f5c2ecdaa"}, + {file = "msgpack-1.0.8-cp310-cp310-musllinux_1_1_aarch64.whl", hash = "sha256:9ee32dcb8e531adae1f1ca568822e9b3a738369b3b686d1477cbc643c4a9c128"}, + {file = "msgpack-1.0.8-cp310-cp310-musllinux_1_1_i686.whl", hash = "sha256:e3aa7e51d738e0ec0afbed661261513b38b3014754c9459508399baf14ae0c9d"}, + {file = "msgpack-1.0.8-cp310-cp310-musllinux_1_1_x86_64.whl", hash = "sha256:69284049d07fce531c17404fcba2bb1df472bc2dcdac642ae71a2d079d950653"}, + {file = "msgpack-1.0.8-cp310-cp310-win32.whl", hash = "sha256:13577ec9e247f8741c84d06b9ece5f654920d8365a4b636ce0e44f15e07ec693"}, + {file = "msgpack-1.0.8-cp310-cp310-win_amd64.whl", hash = "sha256:e532dbd6ddfe13946de050d7474e3f5fb6ec774fbb1a188aaf469b08cf04189a"}, + {file = "msgpack-1.0.8-cp311-cp311-macosx_10_9_universal2.whl", hash = "sha256:9517004e21664f2b5a5fd6333b0731b9cf0817403a941b393d89a2f1dc2bd836"}, + {file = "msgpack-1.0.8-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:d16a786905034e7e34098634b184a7d81f91d4c3d246edc6bd7aefb2fd8ea6ad"}, + {file = "msgpack-1.0.8-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:e2872993e209f7ed04d963e4b4fbae72d034844ec66bc4ca403329db2074377b"}, + {file = "msgpack-1.0.8-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:5c330eace3dd100bdb54b5653b966de7f51c26ec4a7d4e87132d9b4f738220ba"}, + {file = "msgpack-1.0.8-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:83b5c044f3eff2a6534768ccfd50425939e7a8b5cf9a7261c385de1e20dcfc85"}, + {file = "msgpack-1.0.8-cp311-cp311-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:1876b0b653a808fcd50123b953af170c535027bf1d053b59790eebb0aeb38950"}, + {file = "msgpack-1.0.8-cp311-cp311-musllinux_1_1_aarch64.whl", hash = "sha256:dfe1f0f0ed5785c187144c46a292b8c34c1295c01da12e10ccddfc16def4448a"}, + {file = "msgpack-1.0.8-cp311-cp311-musllinux_1_1_i686.whl", hash = "sha256:3528807cbbb7f315bb81959d5961855e7ba52aa60a3097151cb21956fbc7502b"}, + {file = "msgpack-1.0.8-cp311-cp311-musllinux_1_1_x86_64.whl", hash = "sha256:e2f879ab92ce502a1e65fce390eab619774dda6a6ff719718069ac94084098ce"}, + {file = "msgpack-1.0.8-cp311-cp311-win32.whl", hash = "sha256:26ee97a8261e6e35885c2ecd2fd4a6d38252246f94a2aec23665a4e66d066305"}, + {file = "msgpack-1.0.8-cp311-cp311-win_amd64.whl", hash = "sha256:eadb9f826c138e6cf3c49d6f8de88225a3c0ab181a9b4ba792e006e5292d150e"}, + {file = "msgpack-1.0.8-cp312-cp312-macosx_10_9_universal2.whl", hash = "sha256:114be227f5213ef8b215c22dde19532f5da9652e56e8ce969bf0a26d7c419fee"}, + {file = "msgpack-1.0.8-cp312-cp312-macosx_10_9_x86_64.whl", hash = "sha256:d661dc4785affa9d0edfdd1e59ec056a58b3dbb9f196fa43587f3ddac654ac7b"}, + {file = "msgpack-1.0.8-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:d56fd9f1f1cdc8227d7b7918f55091349741904d9520c65f0139a9755952c9e8"}, + {file = "msgpack-1.0.8-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:0726c282d188e204281ebd8de31724b7d749adebc086873a59efb8cf7ae27df3"}, + {file = "msgpack-1.0.8-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:8db8e423192303ed77cff4dce3a4b88dbfaf43979d280181558af5e2c3c71afc"}, + {file = "msgpack-1.0.8-cp312-cp312-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:99881222f4a8c2f641f25703963a5cefb076adffd959e0558dc9f803a52d6a58"}, + {file = "msgpack-1.0.8-cp312-cp312-musllinux_1_1_aarch64.whl", hash = "sha256:b5505774ea2a73a86ea176e8a9a4a7c8bf5d521050f0f6f8426afe798689243f"}, + {file = "msgpack-1.0.8-cp312-cp312-musllinux_1_1_i686.whl", hash = "sha256:ef254a06bcea461e65ff0373d8a0dd1ed3aa004af48839f002a0c994a6f72d04"}, + {file = "msgpack-1.0.8-cp312-cp312-musllinux_1_1_x86_64.whl", hash = "sha256:e1dd7839443592d00e96db831eddb4111a2a81a46b028f0facd60a09ebbdd543"}, + {file = "msgpack-1.0.8-cp312-cp312-win32.whl", hash = "sha256:64d0fcd436c5683fdd7c907eeae5e2cbb5eb872fafbc03a43609d7941840995c"}, + {file = "msgpack-1.0.8-cp312-cp312-win_amd64.whl", hash = "sha256:74398a4cf19de42e1498368c36eed45d9528f5fd0155241e82c4082b7e16cffd"}, + {file = "msgpack-1.0.8-cp38-cp38-macosx_10_9_universal2.whl", hash = "sha256:0ceea77719d45c839fd73abcb190b8390412a890df2f83fb8cf49b2a4b5c2f40"}, + {file = "msgpack-1.0.8-cp38-cp38-macosx_10_9_x86_64.whl", hash = "sha256:1ab0bbcd4d1f7b6991ee7c753655b481c50084294218de69365f8f1970d4c151"}, + {file = "msgpack-1.0.8-cp38-cp38-macosx_11_0_arm64.whl", hash = "sha256:1cce488457370ffd1f953846f82323cb6b2ad2190987cd4d70b2713e17268d24"}, + {file = "msgpack-1.0.8-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:3923a1778f7e5ef31865893fdca12a8d7dc03a44b33e2a5f3295416314c09f5d"}, + {file = "msgpack-1.0.8-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:a22e47578b30a3e199ab067a4d43d790249b3c0587d9a771921f86250c8435db"}, + {file = "msgpack-1.0.8-cp38-cp38-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:bd739c9251d01e0279ce729e37b39d49a08c0420d3fee7f2a4968c0576678f77"}, + {file = "msgpack-1.0.8-cp38-cp38-musllinux_1_1_aarch64.whl", hash = "sha256:d3420522057ebab1728b21ad473aa950026d07cb09da41103f8e597dfbfaeb13"}, + {file = "msgpack-1.0.8-cp38-cp38-musllinux_1_1_i686.whl", hash = "sha256:5845fdf5e5d5b78a49b826fcdc0eb2e2aa7191980e3d2cfd2a30303a74f212e2"}, + {file = "msgpack-1.0.8-cp38-cp38-musllinux_1_1_x86_64.whl", hash = "sha256:6a0e76621f6e1f908ae52860bdcb58e1ca85231a9b0545e64509c931dd34275a"}, + {file = "msgpack-1.0.8-cp38-cp38-win32.whl", hash = "sha256:374a8e88ddab84b9ada695d255679fb99c53513c0a51778796fcf0944d6c789c"}, + {file = "msgpack-1.0.8-cp38-cp38-win_amd64.whl", hash = "sha256:f3709997b228685fe53e8c433e2df9f0cdb5f4542bd5114ed17ac3c0129b0480"}, + {file = "msgpack-1.0.8-cp39-cp39-macosx_10_9_universal2.whl", hash = "sha256:f51bab98d52739c50c56658cc303f190785f9a2cd97b823357e7aeae54c8f68a"}, + {file = "msgpack-1.0.8-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:73ee792784d48aa338bba28063e19a27e8d989344f34aad14ea6e1b9bd83f596"}, + {file = "msgpack-1.0.8-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:f9904e24646570539a8950400602d66d2b2c492b9010ea7e965025cb71d0c86d"}, + {file = "msgpack-1.0.8-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:e75753aeda0ddc4c28dce4c32ba2f6ec30b1b02f6c0b14e547841ba5b24f753f"}, + {file = "msgpack-1.0.8-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:5dbf059fb4b7c240c873c1245ee112505be27497e90f7c6591261c7d3c3a8228"}, + {file = "msgpack-1.0.8-cp39-cp39-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:4916727e31c28be8beaf11cf117d6f6f188dcc36daae4e851fee88646f5b6b18"}, + {file = "msgpack-1.0.8-cp39-cp39-musllinux_1_1_aarch64.whl", hash = "sha256:7938111ed1358f536daf311be244f34df7bf3cdedb3ed883787aca97778b28d8"}, + {file = "msgpack-1.0.8-cp39-cp39-musllinux_1_1_i686.whl", hash = "sha256:493c5c5e44b06d6c9268ce21b302c9ca055c1fd3484c25ba41d34476c76ee746"}, + {file = "msgpack-1.0.8-cp39-cp39-musllinux_1_1_x86_64.whl", hash = "sha256:5fbb160554e319f7b22ecf530a80a3ff496d38e8e07ae763b9e82fadfe96f273"}, + {file = "msgpack-1.0.8-cp39-cp39-win32.whl", hash = "sha256:f9af38a89b6a5c04b7d18c492c8ccf2aee7048aff1ce8437c4683bb5a1df893d"}, + {file = "msgpack-1.0.8-cp39-cp39-win_amd64.whl", hash = "sha256:ed59dd52075f8fc91da6053b12e8c89e37aa043f8986efd89e61fae69dc1b011"}, + {file = "msgpack-1.0.8-py3-none-any.whl", hash = "sha256:24f727df1e20b9876fa6e95f840a2a2651e34c0ad147676356f4bf5fbb0206ca"}, ] [[package]] @@ -2506,38 +2559,38 @@ files = [ [[package]] name = "mypy" -version = "1.8.0" +version = "1.10.0" description = "Optional static typing for Python" optional = false python-versions = ">=3.8" files = [ - {file = "mypy-1.8.0-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:485a8942f671120f76afffff70f259e1cd0f0cfe08f81c05d8816d958d4577d3"}, - {file = "mypy-1.8.0-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:df9824ac11deaf007443e7ed2a4a26bebff98d2bc43c6da21b2b64185da011c4"}, - {file = "mypy-1.8.0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:2afecd6354bbfb6e0160f4e4ad9ba6e4e003b767dd80d85516e71f2e955ab50d"}, - {file = "mypy-1.8.0-cp310-cp310-musllinux_1_1_x86_64.whl", hash = "sha256:8963b83d53ee733a6e4196954502b33567ad07dfd74851f32be18eb932fb1cb9"}, - {file = "mypy-1.8.0-cp310-cp310-win_amd64.whl", hash = "sha256:e46f44b54ebddbeedbd3d5b289a893219065ef805d95094d16a0af6630f5d410"}, - {file = "mypy-1.8.0-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:855fe27b80375e5c5878492f0729540db47b186509c98dae341254c8f45f42ae"}, - {file = "mypy-1.8.0-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:4c886c6cce2d070bd7df4ec4a05a13ee20c0aa60cb587e8d1265b6c03cf91da3"}, - {file = "mypy-1.8.0-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:d19c413b3c07cbecf1f991e2221746b0d2a9410b59cb3f4fb9557f0365a1a817"}, - {file = "mypy-1.8.0-cp311-cp311-musllinux_1_1_x86_64.whl", hash = "sha256:9261ed810972061388918c83c3f5cd46079d875026ba97380f3e3978a72f503d"}, - {file = "mypy-1.8.0-cp311-cp311-win_amd64.whl", hash = "sha256:51720c776d148bad2372ca21ca29256ed483aa9a4cdefefcef49006dff2a6835"}, - {file = "mypy-1.8.0-cp312-cp312-macosx_10_9_x86_64.whl", hash = "sha256:52825b01f5c4c1c4eb0db253ec09c7aa17e1a7304d247c48b6f3599ef40db8bd"}, - {file = "mypy-1.8.0-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:f5ac9a4eeb1ec0f1ccdc6f326bcdb464de5f80eb07fb38b5ddd7b0de6bc61e55"}, - {file = "mypy-1.8.0-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:afe3fe972c645b4632c563d3f3eff1cdca2fa058f730df2b93a35e3b0c538218"}, - {file = "mypy-1.8.0-cp312-cp312-musllinux_1_1_x86_64.whl", hash = "sha256:42c6680d256ab35637ef88891c6bd02514ccb7e1122133ac96055ff458f93fc3"}, - {file = "mypy-1.8.0-cp312-cp312-win_amd64.whl", hash = "sha256:720a5ca70e136b675af3af63db533c1c8c9181314d207568bbe79051f122669e"}, - {file = "mypy-1.8.0-cp38-cp38-macosx_10_9_x86_64.whl", hash = "sha256:028cf9f2cae89e202d7b6593cd98db6759379f17a319b5faf4f9978d7084cdc6"}, - {file = "mypy-1.8.0-cp38-cp38-macosx_11_0_arm64.whl", hash = "sha256:4e6d97288757e1ddba10dd9549ac27982e3e74a49d8d0179fc14d4365c7add66"}, - {file = "mypy-1.8.0-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:7f1478736fcebb90f97e40aff11a5f253af890c845ee0c850fe80aa060a267c6"}, - {file = "mypy-1.8.0-cp38-cp38-musllinux_1_1_x86_64.whl", hash = "sha256:42419861b43e6962a649068a61f4a4839205a3ef525b858377a960b9e2de6e0d"}, - {file = "mypy-1.8.0-cp38-cp38-win_amd64.whl", hash = "sha256:2b5b6c721bd4aabaadead3a5e6fa85c11c6c795e0c81a7215776ef8afc66de02"}, - {file = "mypy-1.8.0-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:5c1538c38584029352878a0466f03a8ee7547d7bd9f641f57a0f3017a7c905b8"}, - {file = "mypy-1.8.0-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:4ef4be7baf08a203170f29e89d79064463b7fc7a0908b9d0d5114e8009c3a259"}, - {file = "mypy-1.8.0-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:7178def594014aa6c35a8ff411cf37d682f428b3b5617ca79029d8ae72f5402b"}, - {file = "mypy-1.8.0-cp39-cp39-musllinux_1_1_x86_64.whl", hash = "sha256:ab3c84fa13c04aeeeabb2a7f67a25ef5d77ac9d6486ff33ded762ef353aa5592"}, - {file = "mypy-1.8.0-cp39-cp39-win_amd64.whl", hash = "sha256:99b00bc72855812a60d253420d8a2eae839b0afa4938f09f4d2aa9bb4654263a"}, - {file = "mypy-1.8.0-py3-none-any.whl", hash = "sha256:538fd81bb5e430cc1381a443971c0475582ff9f434c16cd46d2c66763ce85d9d"}, - {file = "mypy-1.8.0.tar.gz", hash = "sha256:6ff8b244d7085a0b425b56d327b480c3b29cafbd2eff27316a004f9a7391ae07"}, + {file = "mypy-1.10.0-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:da1cbf08fb3b851ab3b9523a884c232774008267b1f83371ace57f412fe308c2"}, + {file = "mypy-1.10.0-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:12b6bfc1b1a66095ab413160a6e520e1dc076a28f3e22f7fb25ba3b000b4ef99"}, + {file = "mypy-1.10.0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:9e36fb078cce9904c7989b9693e41cb9711e0600139ce3970c6ef814b6ebc2b2"}, + {file = "mypy-1.10.0-cp310-cp310-musllinux_1_1_x86_64.whl", hash = "sha256:2b0695d605ddcd3eb2f736cd8b4e388288c21e7de85001e9f85df9187f2b50f9"}, + {file = "mypy-1.10.0-cp310-cp310-win_amd64.whl", hash = "sha256:cd777b780312ddb135bceb9bc8722a73ec95e042f911cc279e2ec3c667076051"}, + {file = "mypy-1.10.0-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:3be66771aa5c97602f382230165b856c231d1277c511c9a8dd058be4784472e1"}, + {file = "mypy-1.10.0-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:8b2cbaca148d0754a54d44121b5825ae71868c7592a53b7292eeb0f3fdae95ee"}, + {file = "mypy-1.10.0-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:1ec404a7cbe9fc0e92cb0e67f55ce0c025014e26d33e54d9e506a0f2d07fe5de"}, + {file = "mypy-1.10.0-cp311-cp311-musllinux_1_1_x86_64.whl", hash = "sha256:e22e1527dc3d4aa94311d246b59e47f6455b8729f4968765ac1eacf9a4760bc7"}, + {file = "mypy-1.10.0-cp311-cp311-win_amd64.whl", hash = "sha256:a87dbfa85971e8d59c9cc1fcf534efe664d8949e4c0b6b44e8ca548e746a8d53"}, + {file = "mypy-1.10.0-cp312-cp312-macosx_10_9_x86_64.whl", hash = "sha256:a781f6ad4bab20eef8b65174a57e5203f4be627b46291f4589879bf4e257b97b"}, + {file = "mypy-1.10.0-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:b808e12113505b97d9023b0b5e0c0705a90571c6feefc6f215c1df9381256e30"}, + {file = "mypy-1.10.0-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:8f55583b12156c399dce2df7d16f8a5095291354f1e839c252ec6c0611e86e2e"}, + {file = "mypy-1.10.0-cp312-cp312-musllinux_1_1_x86_64.whl", hash = "sha256:4cf18f9d0efa1b16478c4c129eabec36148032575391095f73cae2e722fcf9d5"}, + {file = "mypy-1.10.0-cp312-cp312-win_amd64.whl", hash = "sha256:bc6ac273b23c6b82da3bb25f4136c4fd42665f17f2cd850771cb600bdd2ebeda"}, + {file = "mypy-1.10.0-cp38-cp38-macosx_10_9_x86_64.whl", hash = "sha256:9fd50226364cd2737351c79807775136b0abe084433b55b2e29181a4c3c878c0"}, + {file = "mypy-1.10.0-cp38-cp38-macosx_11_0_arm64.whl", hash = "sha256:f90cff89eea89273727d8783fef5d4a934be2fdca11b47def50cf5d311aff727"}, + {file = "mypy-1.10.0-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:fcfc70599efde5c67862a07a1aaf50e55bce629ace26bb19dc17cece5dd31ca4"}, + {file = "mypy-1.10.0-cp38-cp38-musllinux_1_1_x86_64.whl", hash = "sha256:075cbf81f3e134eadaf247de187bd604748171d6b79736fa9b6c9685b4083061"}, + {file = "mypy-1.10.0-cp38-cp38-win_amd64.whl", hash = "sha256:3f298531bca95ff615b6e9f2fc0333aae27fa48052903a0ac90215021cdcfa4f"}, + {file = "mypy-1.10.0-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:fa7ef5244615a2523b56c034becde4e9e3f9b034854c93639adb667ec9ec2976"}, + {file = "mypy-1.10.0-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:3236a4c8f535a0631f85f5fcdffba71c7feeef76a6002fcba7c1a8e57c8be1ec"}, + {file = "mypy-1.10.0-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:4a2b5cdbb5dd35aa08ea9114436e0d79aceb2f38e32c21684dcf8e24e1e92821"}, + {file = "mypy-1.10.0-cp39-cp39-musllinux_1_1_x86_64.whl", hash = "sha256:92f93b21c0fe73dc00abf91022234c79d793318b8a96faac147cd579c1671746"}, + {file = "mypy-1.10.0-cp39-cp39-win_amd64.whl", hash = "sha256:28d0e038361b45f099cc086d9dd99c15ff14d0188f44ac883010e172ce86c38a"}, + {file = "mypy-1.10.0-py3-none-any.whl", hash = "sha256:f8c083976eb530019175aabadb60921e73b4f45736760826aa1689dda8208aee"}, + {file = "mypy-1.10.0.tar.gz", hash = "sha256:3d087fcbec056c4ee34974da493a826ce316947485cef3901f511848e687c131"}, ] [package.dependencies] @@ -2564,13 +2617,13 @@ files = [ [[package]] name = "nbclient" -version = "0.9.0" +version = "0.10.0" description = "A client library for executing notebooks. Formerly nbconvert's ExecutePreprocessor." optional = false python-versions = ">=3.8.0" files = [ - {file = "nbclient-0.9.0-py3-none-any.whl", hash = "sha256:a3a1ddfb34d4a9d17fc744d655962714a866639acd30130e9be84191cd97cd15"}, - {file = "nbclient-0.9.0.tar.gz", hash = "sha256:4b28c207877cf33ef3a9838cdc7a54c5ceff981194a82eac59d558f05487295e"}, + {file = "nbclient-0.10.0-py3-none-any.whl", hash = "sha256:f13e3529332a1f1f81d82a53210322476a168bb7090a0289c795fe9cc11c9d3f"}, + {file = "nbclient-0.10.0.tar.gz", hash = "sha256:4b3f1b7dba531e498449c4db4f53da339c91d449dc11e9af3a43b4eb5c5abb09"}, ] [package.dependencies] @@ -2582,23 +2635,23 @@ traitlets = ">=5.4" [package.extras] dev = ["pre-commit"] docs = ["autodoc-traits", "mock", "moto", "myst-parser", "nbclient[test]", "sphinx (>=1.7)", "sphinx-book-theme", "sphinxcontrib-spelling"] -test = ["flaky", "ipykernel (>=6.19.3)", "ipython", "ipywidgets", "nbconvert (>=7.0.0)", "pytest (>=7.0)", "pytest-asyncio", "pytest-cov (>=4.0)", "testpath", "xmltodict"] +test = ["flaky", "ipykernel (>=6.19.3)", "ipython", "ipywidgets", "nbconvert (>=7.0.0)", "pytest (>=7.0,<8)", "pytest-asyncio", "pytest-cov (>=4.0)", "testpath", "xmltodict"] [[package]] name = "nbformat" -version = "5.9.2" +version = "5.10.4" description = "The Jupyter Notebook format" optional = false python-versions = ">=3.8" files = [ - {file = "nbformat-5.9.2-py3-none-any.whl", hash = "sha256:1c5172d786a41b82bcfd0c23f9e6b6f072e8fb49c39250219e4acfff1efe89e9"}, - {file = "nbformat-5.9.2.tar.gz", hash = "sha256:5f98b5ba1997dff175e77e0c17d5c10a96eaed2cbd1de3533d1fc35d5e111192"}, + {file = "nbformat-5.10.4-py3-none-any.whl", hash = "sha256:3b48d6c8fbca4b299bf3982ea7db1af21580e4fec269ad087b9e81588891200b"}, + {file = "nbformat-5.10.4.tar.gz", hash = "sha256:322168b14f937a5d11362988ecac2a4952d3d8e3a2cbeb2319584631226d5b3a"}, ] [package.dependencies] -fastjsonschema = "*" +fastjsonschema = ">=2.15" jsonschema = ">=2.6" -jupyter-core = "*" +jupyter-core = ">=4.12,<5.0.dev0 || >=5.1.dev0" traitlets = ">=5.1" [package.extras] @@ -2662,32 +2715,32 @@ tests = ["Cython", "packaging", "pytest"] [[package]] name = "numba" -version = "0.59.0" +version = "0.59.1" description = "compiling Python code using LLVM" optional = false python-versions = ">=3.9" files = [ - {file = "numba-0.59.0-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:8d061d800473fb8fef76a455221f4ad649a53f5e0f96e3f6c8b8553ee6fa98fa"}, - {file = "numba-0.59.0-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:c086a434e7d3891ce5dfd3d1e7ee8102ac1e733962098578b507864120559ceb"}, - {file = "numba-0.59.0-cp310-cp310-manylinux2014_aarch64.manylinux_2_17_aarch64.whl", hash = "sha256:9e20736bf62e61f8353fb71b0d3a1efba636c7a303d511600fc57648b55823ed"}, - {file = "numba-0.59.0-cp310-cp310-manylinux2014_x86_64.manylinux_2_17_x86_64.whl", hash = "sha256:e86e6786aec31d2002122199486e10bbc0dc40f78d76364cded375912b13614c"}, - {file = "numba-0.59.0-cp310-cp310-win_amd64.whl", hash = "sha256:0307ee91b24500bb7e64d8a109848baf3a3905df48ce142b8ac60aaa406a0400"}, - {file = "numba-0.59.0-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:d540f69a8245fb714419c2209e9af6104e568eb97623adc8943642e61f5d6d8e"}, - {file = "numba-0.59.0-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:1192d6b2906bf3ff72b1d97458724d98860ab86a91abdd4cfd9328432b661e31"}, - {file = "numba-0.59.0-cp311-cp311-manylinux2014_aarch64.manylinux_2_17_aarch64.whl", hash = "sha256:90efb436d3413809fcd15298c6d395cb7d98184350472588356ccf19db9e37c8"}, - {file = "numba-0.59.0-cp311-cp311-manylinux2014_x86_64.manylinux_2_17_x86_64.whl", hash = "sha256:cd3dac45e25d927dcb65d44fb3a973994f5add2b15add13337844afe669dd1ba"}, - {file = "numba-0.59.0-cp311-cp311-win_amd64.whl", hash = "sha256:753dc601a159861808cc3207bad5c17724d3b69552fd22768fddbf302a817a4c"}, - {file = "numba-0.59.0-cp312-cp312-macosx_10_9_x86_64.whl", hash = "sha256:ce62bc0e6dd5264e7ff7f34f41786889fa81a6b860662f824aa7532537a7bee0"}, - {file = "numba-0.59.0-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:8cbef55b73741b5eea2dbaf1b0590b14977ca95a13a07d200b794f8f6833a01c"}, - {file = "numba-0.59.0-cp312-cp312-manylinux2014_aarch64.manylinux_2_17_aarch64.whl", hash = "sha256:70d26ba589f764be45ea8c272caa467dbe882b9676f6749fe6f42678091f5f21"}, - {file = "numba-0.59.0-cp312-cp312-manylinux2014_x86_64.manylinux_2_17_x86_64.whl", hash = "sha256:e125f7d69968118c28ec0eed9fbedd75440e64214b8d2eac033c22c04db48492"}, - {file = "numba-0.59.0-cp312-cp312-win_amd64.whl", hash = "sha256:4981659220b61a03c1e557654027d271f56f3087448967a55c79a0e5f926de62"}, - {file = "numba-0.59.0-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:fe4d7562d1eed754a7511ed7ba962067f198f86909741c5c6e18c4f1819b1f47"}, - {file = "numba-0.59.0-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:6feb1504bb432280f900deaf4b1dadcee68812209500ed3f81c375cbceab24dc"}, - {file = "numba-0.59.0-cp39-cp39-manylinux2014_aarch64.manylinux_2_17_aarch64.whl", hash = "sha256:944faad25ee23ea9dda582bfb0189fb9f4fc232359a80ab2a028b94c14ce2b1d"}, - {file = "numba-0.59.0-cp39-cp39-manylinux2014_x86_64.manylinux_2_17_x86_64.whl", hash = "sha256:5516a469514bfae52a9d7989db4940653a5cbfac106f44cb9c50133b7ad6224b"}, - {file = "numba-0.59.0-cp39-cp39-win_amd64.whl", hash = "sha256:32bd0a41525ec0b1b853da244808f4e5333867df3c43c30c33f89cf20b9c2b63"}, - {file = "numba-0.59.0.tar.gz", hash = "sha256:12b9b064a3e4ad00e2371fc5212ef0396c80f41caec9b5ec391c8b04b6eaf2a8"}, + {file = "numba-0.59.1-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:97385a7f12212c4f4bc28f648720a92514bee79d7063e40ef66c2d30600fd18e"}, + {file = "numba-0.59.1-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:0b77aecf52040de2a1eb1d7e314497b9e56fba17466c80b457b971a25bb1576d"}, + {file = "numba-0.59.1-cp310-cp310-manylinux2014_aarch64.manylinux_2_17_aarch64.whl", hash = "sha256:3476a4f641bfd58f35ead42f4dcaf5f132569c4647c6f1360ccf18ee4cda3990"}, + {file = "numba-0.59.1-cp310-cp310-manylinux2014_x86_64.manylinux_2_17_x86_64.whl", hash = "sha256:525ef3f820931bdae95ee5379c670d5c97289c6520726bc6937a4a7d4230ba24"}, + {file = "numba-0.59.1-cp310-cp310-win_amd64.whl", hash = "sha256:990e395e44d192a12105eca3083b61307db7da10e093972ca285c85bef0963d6"}, + {file = "numba-0.59.1-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:43727e7ad20b3ec23ee4fc642f5b61845c71f75dd2825b3c234390c6d8d64051"}, + {file = "numba-0.59.1-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:411df625372c77959570050e861981e9d196cc1da9aa62c3d6a836b5cc338966"}, + {file = "numba-0.59.1-cp311-cp311-manylinux2014_aarch64.manylinux_2_17_aarch64.whl", hash = "sha256:2801003caa263d1e8497fb84829a7ecfb61738a95f62bc05693fcf1733e978e4"}, + {file = "numba-0.59.1-cp311-cp311-manylinux2014_x86_64.manylinux_2_17_x86_64.whl", hash = "sha256:dd2842fac03be4e5324ebbbd4d2d0c8c0fc6e0df75c09477dd45b288a0777389"}, + {file = "numba-0.59.1-cp311-cp311-win_amd64.whl", hash = "sha256:0594b3dfb369fada1f8bb2e3045cd6c61a564c62e50cf1f86b4666bc721b3450"}, + {file = "numba-0.59.1-cp312-cp312-macosx_10_9_x86_64.whl", hash = "sha256:1cce206a3b92836cdf26ef39d3a3242fec25e07f020cc4feec4c4a865e340569"}, + {file = "numba-0.59.1-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:8c8b4477763cb1fbd86a3be7050500229417bf60867c93e131fd2626edb02238"}, + {file = "numba-0.59.1-cp312-cp312-manylinux2014_aarch64.manylinux_2_17_aarch64.whl", hash = "sha256:7d80bce4ef7e65bf895c29e3889ca75a29ee01da80266a01d34815918e365835"}, + {file = "numba-0.59.1-cp312-cp312-manylinux2014_x86_64.manylinux_2_17_x86_64.whl", hash = "sha256:f7ad1d217773e89a9845886401eaaab0a156a90aa2f179fdc125261fd1105096"}, + {file = "numba-0.59.1-cp312-cp312-win_amd64.whl", hash = "sha256:5bf68f4d69dd3a9f26a9b23548fa23e3bcb9042e2935257b471d2a8d3c424b7f"}, + {file = "numba-0.59.1-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:4e0318ae729de6e5dbe64c75ead1a95eb01fabfe0e2ebed81ebf0344d32db0ae"}, + {file = "numba-0.59.1-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:0f68589740a8c38bb7dc1b938b55d1145244c8353078eea23895d4f82c8b9ec1"}, + {file = "numba-0.59.1-cp39-cp39-manylinux2014_aarch64.manylinux_2_17_aarch64.whl", hash = "sha256:649913a3758891c77c32e2d2a3bcbedf4a69f5fea276d11f9119677c45a422e8"}, + {file = "numba-0.59.1-cp39-cp39-manylinux2014_x86_64.manylinux_2_17_x86_64.whl", hash = "sha256:9712808e4545270291d76b9a264839ac878c5eb7d8b6e02c970dc0ac29bc8187"}, + {file = "numba-0.59.1-cp39-cp39-win_amd64.whl", hash = "sha256:8d51ccd7008a83105ad6a0082b6a2b70f1142dc7cfd76deb8c5a862367eb8c86"}, + {file = "numba-0.59.1.tar.gz", hash = "sha256:76f69132b96028d2774ed20415e8c528a34e3299a40581bae178f0994a2f370b"}, ] [package.dependencies] @@ -2696,13 +2749,13 @@ numpy = ">=1.22,<1.27" [[package]] name = "numbagg" -version = "0.8.0" +version = "0.8.1" description = "Fast N-dimensional aggregation functions with Numba" optional = false python-versions = ">=3.9" files = [ - {file = "numbagg-0.8.0-py3-none-any.whl", hash = "sha256:02bc05396985553ce9e38b5085ebd30bef6abe3a73a838cc255ceb98a86405a0"}, - {file = "numbagg-0.8.0.tar.gz", hash = "sha256:ea44c9f4e914dd614374b64bd5b6b663f67fe65cc29d8bd370e9a565b8c187cc"}, + {file = "numbagg-0.8.1-py3-none-any.whl", hash = "sha256:fc889d7b890bf5e729e64add34852bc70b1447eb53a29fca8b32d345ed792a65"}, + {file = "numbagg-0.8.1.tar.gz", hash = "sha256:3a111361e6062feafd600b1e3c8a8187fd55e955bee735ada843705395fe84d9"}, ] [package.dependencies] @@ -2710,7 +2763,7 @@ numba = "*" numpy = "*" [package.extras] -dev = ["bottleneck", "hypothesis", "jq", "pandas", "pre-commit", "pytest", "pytest-benchmark", "ruff", "setuptools-scm", "tabulate"] +dev = ["bottleneck", "hypothesis", "jq", "mypy", "pandas", "pre-commit", "pytest", "pytest-benchmark", "ruff", "setuptools-scm", "tabulate"] [[package]] name = "numcodecs" @@ -2799,13 +2852,13 @@ files = [ [[package]] name = "numpy-groupies" -version = "0.10.2" +version = "0.11.1" description = "Optimised tools for group-indexing operations: aggregated sum and more." optional = false python-versions = ">=3.9" files = [ - {file = "numpy-groupies-0.10.2.tar.gz", hash = "sha256:f920c4ded899f5975d94fc63d634e7c89622056bbab8cc98a44d4320a0ae8a12"}, - {file = "numpy_groupies-0.10.2-py3-none-any.whl", hash = "sha256:24a574ecdd58c0d669749b67ec31dab971d27e04037b402bbc4f6bedaa2623bc"}, + {file = "numpy_groupies-0.11.1-py3-none-any.whl", hash = "sha256:c9f2471b0ee96e202fb0a591a551192c19fa3bdddad33df8665a7accbaf3263e"}, + {file = "numpy_groupies-0.11.1.tar.gz", hash = "sha256:b0e9cf7a4d015beca969f1e90463fa90972df8380a722eb96d679f8838d2f6fe"}, ] [package.dependencies] @@ -2849,13 +2902,13 @@ tests = ["pytest", "pytest-cov", "pytest-pep8"] [[package]] name = "packaging" -version = "23.2" +version = "24.0" description = "Core utilities for Python packages" optional = false python-versions = ">=3.7" files = [ - {file = "packaging-23.2-py3-none-any.whl", hash = "sha256:8c491190033a9af7e1d931d0b5dacc2ef47509b34dd0de67ed209b5203fc88c7"}, - {file = "packaging-23.2.tar.gz", hash = "sha256:048fb0e9405036518eaaf48a55953c750c11e1a1b68e0dd1a9d62ed0c092cfc5"}, + {file = "packaging-24.0-py3-none-any.whl", hash = "sha256:2ddfb553fdf02fb784c234c7ba6ccc288296ceabec964ad2eae3777778130bc5"}, + {file = "packaging-24.0.tar.gz", hash = "sha256:eb82c5e3e56209074766e6885bb04b8c38a0c015d0a30036ebe7ece34c9989e9"}, ] [[package]] @@ -2870,47 +2923,44 @@ files = [ [[package]] name = "pandas" -version = "2.2.0" +version = "2.2.2" description = "Powerful data structures for data analysis, time series, and statistics" optional = false python-versions = ">=3.9" files = [ - {file = "pandas-2.2.0-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:8108ee1712bb4fa2c16981fba7e68b3f6ea330277f5ca34fa8d557e986a11670"}, - {file = "pandas-2.2.0-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:736da9ad4033aeab51d067fc3bd69a0ba36f5a60f66a527b3d72e2030e63280a"}, - {file = "pandas-2.2.0-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:38e0b4fc3ddceb56ec8a287313bc22abe17ab0eb184069f08fc6a9352a769b18"}, - {file = "pandas-2.2.0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:20404d2adefe92aed3b38da41d0847a143a09be982a31b85bc7dd565bdba0f4e"}, - {file = "pandas-2.2.0-cp310-cp310-musllinux_1_1_aarch64.whl", hash = "sha256:7ea3ee3f125032bfcade3a4cf85131ed064b4f8dd23e5ce6fa16473e48ebcaf5"}, - {file = "pandas-2.2.0-cp310-cp310-musllinux_1_1_x86_64.whl", hash = "sha256:f9670b3ac00a387620489dfc1bca66db47a787f4e55911f1293063a78b108df1"}, - {file = "pandas-2.2.0-cp310-cp310-win_amd64.whl", hash = "sha256:5a946f210383c7e6d16312d30b238fd508d80d927014f3b33fb5b15c2f895430"}, - {file = "pandas-2.2.0-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:a1b438fa26b208005c997e78672f1aa8138f67002e833312e6230f3e57fa87d5"}, - {file = "pandas-2.2.0-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:8ce2fbc8d9bf303ce54a476116165220a1fedf15985b09656b4b4275300e920b"}, - {file = "pandas-2.2.0-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:2707514a7bec41a4ab81f2ccce8b382961a29fbe9492eab1305bb075b2b1ff4f"}, - {file = "pandas-2.2.0-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:85793cbdc2d5bc32620dc8ffa715423f0c680dacacf55056ba13454a5be5de88"}, - {file = "pandas-2.2.0-cp311-cp311-musllinux_1_1_aarch64.whl", hash = "sha256:cfd6c2491dc821b10c716ad6776e7ab311f7df5d16038d0b7458bc0b67dc10f3"}, - {file = "pandas-2.2.0-cp311-cp311-musllinux_1_1_x86_64.whl", hash = "sha256:a146b9dcacc3123aa2b399df1a284de5f46287a4ab4fbfc237eac98a92ebcb71"}, - {file = "pandas-2.2.0-cp311-cp311-win_amd64.whl", hash = "sha256:fbc1b53c0e1fdf16388c33c3cca160f798d38aea2978004dd3f4d3dec56454c9"}, - {file = "pandas-2.2.0-cp312-cp312-macosx_10_9_x86_64.whl", hash = "sha256:a41d06f308a024981dcaa6c41f2f2be46a6b186b902c94c2674e8cb5c42985bc"}, - {file = "pandas-2.2.0-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:159205c99d7a5ce89ecfc37cb08ed179de7783737cea403b295b5eda8e9c56d1"}, - {file = "pandas-2.2.0-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:eb1e1f3861ea9132b32f2133788f3b14911b68102d562715d71bd0013bc45440"}, - {file = "pandas-2.2.0-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:761cb99b42a69005dec2b08854fb1d4888fdf7b05db23a8c5a099e4b886a2106"}, - {file = "pandas-2.2.0-cp312-cp312-musllinux_1_1_aarch64.whl", hash = "sha256:a20628faaf444da122b2a64b1e5360cde100ee6283ae8effa0d8745153809a2e"}, - {file = "pandas-2.2.0-cp312-cp312-musllinux_1_1_x86_64.whl", hash = "sha256:f5be5d03ea2073627e7111f61b9f1f0d9625dc3c4d8dda72cc827b0c58a1d042"}, - {file = "pandas-2.2.0-cp312-cp312-win_amd64.whl", hash = "sha256:a626795722d893ed6aacb64d2401d017ddc8a2341b49e0384ab9bf7112bdec30"}, - {file = "pandas-2.2.0-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:9f66419d4a41132eb7e9a73dcec9486cf5019f52d90dd35547af11bc58f8637d"}, - {file = "pandas-2.2.0-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:57abcaeda83fb80d447f28ab0cc7b32b13978f6f733875ebd1ed14f8fbc0f4ab"}, - {file = "pandas-2.2.0-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:e60f1f7dba3c2d5ca159e18c46a34e7ca7247a73b5dd1a22b6d59707ed6b899a"}, - {file = "pandas-2.2.0-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:eb61dc8567b798b969bcc1fc964788f5a68214d333cade8319c7ab33e2b5d88a"}, - {file = "pandas-2.2.0-cp39-cp39-musllinux_1_1_aarch64.whl", hash = "sha256:52826b5f4ed658fa2b729264d63f6732b8b29949c7fd234510d57c61dbeadfcd"}, - {file = "pandas-2.2.0-cp39-cp39-musllinux_1_1_x86_64.whl", hash = "sha256:bde2bc699dbd80d7bc7f9cab1e23a95c4375de615860ca089f34e7c64f4a8de7"}, - {file = "pandas-2.2.0-cp39-cp39-win_amd64.whl", hash = "sha256:3de918a754bbf2da2381e8a3dcc45eede8cd7775b047b923f9006d5f876802ae"}, - {file = "pandas-2.2.0.tar.gz", hash = "sha256:30b83f7c3eb217fb4d1b494a57a2fda5444f17834f5df2de6b2ffff68dc3c8e2"}, + {file = "pandas-2.2.2-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:90c6fca2acf139569e74e8781709dccb6fe25940488755716d1d354d6bc58bce"}, + {file = "pandas-2.2.2-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:4abfe0be0d7221be4f12552995e58723c7422c80a659da13ca382697de830c08"}, + {file = "pandas-2.2.2-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:8635c16bf3d99040fdf3ca3db669a7250ddf49c55dc4aa8fe0ae0fa8d6dcc1f0"}, + {file = "pandas-2.2.2-cp310-cp310-musllinux_1_1_aarch64.whl", hash = "sha256:40ae1dffb3967a52203105a077415a86044a2bea011b5f321c6aa64b379a3f51"}, + {file = "pandas-2.2.2-cp310-cp310-musllinux_1_1_x86_64.whl", hash = "sha256:8e5a0b00e1e56a842f922e7fae8ae4077aee4af0acb5ae3622bd4b4c30aedf99"}, + {file = "pandas-2.2.2-cp310-cp310-win_amd64.whl", hash = "sha256:ddf818e4e6c7c6f4f7c8a12709696d193976b591cc7dc50588d3d1a6b5dc8772"}, + {file = "pandas-2.2.2-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:696039430f7a562b74fa45f540aca068ea85fa34c244d0deee539cb6d70aa288"}, + {file = "pandas-2.2.2-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:8e90497254aacacbc4ea6ae5e7a8cd75629d6ad2b30025a4a8b09aa4faf55151"}, + {file = "pandas-2.2.2-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:58b84b91b0b9f4bafac2a0ac55002280c094dfc6402402332c0913a59654ab2b"}, + {file = "pandas-2.2.2-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:6d2123dc9ad6a814bcdea0f099885276b31b24f7edf40f6cdbc0912672e22eee"}, + {file = "pandas-2.2.2-cp311-cp311-musllinux_1_1_aarch64.whl", hash = "sha256:2925720037f06e89af896c70bca73459d7e6a4be96f9de79e2d440bd499fe0db"}, + {file = "pandas-2.2.2-cp311-cp311-musllinux_1_1_x86_64.whl", hash = "sha256:0cace394b6ea70c01ca1595f839cf193df35d1575986e484ad35c4aeae7266c1"}, + {file = "pandas-2.2.2-cp311-cp311-win_amd64.whl", hash = "sha256:873d13d177501a28b2756375d59816c365e42ed8417b41665f346289adc68d24"}, + {file = "pandas-2.2.2-cp312-cp312-macosx_10_9_x86_64.whl", hash = "sha256:9dfde2a0ddef507a631dc9dc4af6a9489d5e2e740e226ad426a05cabfbd7c8ef"}, + {file = "pandas-2.2.2-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:1cb51fe389360f3b5a4d57dbd2848a5f033350336ca3b340d1c53a1fad33bcad"}, + {file = "pandas-2.2.2-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:eee3a87076c0756de40b05c5e9a6069c035ba43e8dd71c379e68cab2c20f16ad"}, + {file = "pandas-2.2.2-cp312-cp312-musllinux_1_1_aarch64.whl", hash = "sha256:3e374f59e440d4ab45ca2fffde54b81ac3834cf5ae2cdfa69c90bc03bde04d76"}, + {file = "pandas-2.2.2-cp312-cp312-musllinux_1_1_x86_64.whl", hash = "sha256:43498c0bdb43d55cb162cdc8c06fac328ccb5d2eabe3cadeb3529ae6f0517c32"}, + {file = "pandas-2.2.2-cp312-cp312-win_amd64.whl", hash = "sha256:d187d355ecec3629624fccb01d104da7d7f391db0311145817525281e2804d23"}, + {file = "pandas-2.2.2-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:0ca6377b8fca51815f382bd0b697a0814c8bda55115678cbc94c30aacbb6eff2"}, + {file = "pandas-2.2.2-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:001910ad31abc7bf06f49dcc903755d2f7f3a9186c0c040b827e522e9cef0863"}, + {file = "pandas-2.2.2-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:66b479b0bd07204e37583c191535505410daa8df638fd8e75ae1b383851fe921"}, + {file = "pandas-2.2.2-cp39-cp39-musllinux_1_1_aarch64.whl", hash = "sha256:a77e9d1c386196879aa5eb712e77461aaee433e54c68cf253053a73b7e49c33a"}, + {file = "pandas-2.2.2-cp39-cp39-musllinux_1_1_x86_64.whl", hash = "sha256:92fd6b027924a7e178ac202cfbe25e53368db90d56872d20ffae94b96c7acc57"}, + {file = "pandas-2.2.2-cp39-cp39-win_amd64.whl", hash = "sha256:640cef9aa381b60e296db324337a554aeeb883ead99dc8f6c18e81a93942f5f4"}, + {file = "pandas-2.2.2.tar.gz", hash = "sha256:9e79019aba43cb4fda9e4d983f8e88ca0373adbb697ae9c6c43093218de28b54"}, ] [package.dependencies] numpy = [ - {version = ">=1.22.4,<2", markers = "python_version < \"3.11\""}, - {version = ">=1.23.2,<2", markers = "python_version == \"3.11\""}, - {version = ">=1.26.0,<2", markers = "python_version >= \"3.12\""}, + {version = ">=1.22.4", markers = "python_version < \"3.11\""}, + {version = ">=1.23.2", markers = "python_version == \"3.11\""}, + {version = ">=1.26.0", markers = "python_version >= \"3.12\""}, ] python-dateutil = ">=2.8.2" pytz = ">=2020.1" @@ -2935,6 +2985,7 @@ parquet = ["pyarrow (>=10.0.1)"] performance = ["bottleneck (>=1.3.6)", "numba (>=0.56.4)", "numexpr (>=2.8.4)"] plot = ["matplotlib (>=3.6.3)"] postgresql = ["SQLAlchemy (>=2.0.0)", "adbc-driver-postgresql (>=0.8.0)", "psycopg2 (>=2.9.6)"] +pyarrow = ["pyarrow (>=10.0.1)"] spss = ["pyreadstat (>=1.2.0)"] sql-other = ["SQLAlchemy (>=2.0.0)", "adbc-driver-postgresql (>=0.8.0)", "adbc-driver-sqlite (>=0.8.0)"] test = ["hypothesis (>=6.46.1)", "pytest (>=7.3.2)", "pytest-xdist (>=2.2.0)"] @@ -2942,13 +2993,13 @@ xml = ["lxml (>=4.9.2)"] [[package]] name = "pandas-stubs" -version = "2.1.4.231227" +version = "2.2.2.240514" description = "Type annotations for pandas" optional = false python-versions = ">=3.9" files = [ - {file = "pandas_stubs-2.1.4.231227-py3-none-any.whl", hash = "sha256:211fc23e6ae87073bdf41dbf362c4a4d85e1e3477cb078dbac3da6c7fdaefba8"}, - {file = "pandas_stubs-2.1.4.231227.tar.gz", hash = "sha256:3ea29ef001e9e44985f5ebde02d4413f94891ef6ec7e5056fb07d125be796c23"}, + {file = "pandas_stubs-2.2.2.240514-py3-none-any.whl", hash = "sha256:5d6f64d45a98bc94152a0f76fa648e598cd2b9ba72302fd34602479f0c391a53"}, + {file = "pandas_stubs-2.2.2.240514.tar.gz", hash = "sha256:85b20da44a62c80eb8389bcf4cbfe31cce1cafa8cca4bf1fc75ec45892e72ce8"}, ] [package.dependencies] @@ -2957,24 +3008,24 @@ types-pytz = ">=2022.1.1" [[package]] name = "panel" -version = "1.3.8" +version = "1.4.3" description = "The powerful data exploration & web app framework for Python." optional = false python-versions = ">=3.9" files = [ - {file = "panel-1.3.8-py2.py3-none-any.whl", hash = "sha256:49bf3931986a0ddf3f7b4bda3c65c6a311d5277524acdb4e0bff69cba6bf5775"}, - {file = "panel-1.3.8.tar.gz", hash = "sha256:809afd2b861747a31d6ddaadbbc7c25b8dab392dc78256f68b759214113c5be3"}, + {file = "panel-1.4.3-py3-none-any.whl", hash = "sha256:888050f5411c98eeaaddb4b67da14aeed63898f55aa08592edb33328bce47d6c"}, + {file = "panel-1.4.3.tar.gz", hash = "sha256:f898f28b92551ec5a2c73fae6c8974c7b5bd1f3b610320c149d5e88eb81a0cc5"}, ] [package.dependencies] bleach = "*" -bokeh = ">=3.2.0,<3.4.0" +bokeh = ">=3.4.0,<3.5.0" linkify-it-py = "*" markdown = "*" markdown-it-py = "*" mdit-py-plugins = "*" pandas = ">=1.2" -param = ">=2.0.0,<3.0" +param = ">=2.1.0,<3.0" pyviz-comms = ">=2.0.0" requests = "*" tqdm = ">=4.48.0" @@ -2982,33 +3033,34 @@ typing-extensions = "*" xyzservices = ">=2021.09.1" [package.extras] -all = ["aiohttp", "altair", "anywidget", "channels", "croniter", "datashader", "diskcache", "django (<4)", "fastparquet", "flake8", "folium", "graphviz", "holoviews (>=1.16.0)", "hvplot", "ipyleaflet", "ipympl", "ipython (>=7.0)", "ipyvolume", "ipyvuetify", "ipywidgets", "ipywidgets-bokeh", "jupyter-bokeh (>=3.0.7)", "jupyter-server", "jupyterlab", "lxml", "matplotlib", "nbsite (>=0.8.4)", "nbval", "networkx (>=2.5)", "numba (<0.58)", "numpy", "pandas (<2.1.0)", "pandas (>=1.3)", "parameterized", "pillow", "playwright", "plotly", "plotly (>=4.0)", "pre-commit", "psutil", "pydeck", "pygraphviz", "pyinstrument (>=4.0)", "pytest", "pytest-asyncio (<0.22)", "pytest-cov", "pytest-playwright", "pytest-rerunfailures", "pytest-xdist", "python-graphviz", "pyvista", "reacton", "scikit-image", "scikit-learn", "scipy", "seaborn", "streamz", "twine", "vega-datasets", "vtk", "xarray", "xgboost"] -all-pip = ["aiohttp", "altair", "anywidget", "channels", "croniter", "datashader", "diskcache", "django (<4)", "fastparquet", "flake8", "folium", "graphviz", "holoviews (>=1.16.0)", "hvplot", "ipyleaflet", "ipympl", "ipython (>=7.0)", "ipyvolume", "ipyvuetify", "ipywidgets", "ipywidgets-bokeh", "jupyter-bokeh (>=3.0.7)", "jupyter-server", "jupyterlab", "lxml", "matplotlib", "nbsite (>=0.8.4)", "nbval", "networkx (>=2.5)", "numba (<0.58)", "numpy", "pandas (<2.1.0)", "pandas (>=1.3)", "parameterized", "pillow", "playwright", "plotly", "plotly (>=4.0)", "pre-commit", "psutil", "pydeck", "pyinstrument (>=4.0)", "pytest", "pytest-asyncio (<0.22)", "pytest-cov", "pytest-playwright", "pytest-rerunfailures", "pytest-xdist", "pyvista", "reacton", "scikit-image", "scikit-learn", "scipy", "seaborn", "streamz", "twine", "vega-datasets", "vtk", "xarray", "xgboost"] -build = ["bleach", "bokeh (>=3.3.0,<3.4.0)", "cryptography (<39)", "markdown", "packaging", "param (>=2.0.0)", "pyviz-comms (>=2.0.0)", "requests", "setuptools (>=42)", "tqdm (>=4.48.0)", "urllib3 (<2.0)"] +all = ["aiohttp", "altair", "anywidget", "channels", "croniter", "dask-expr", "datashader", "diskcache", "django (<4)", "fastparquet", "flake8", "folium", "graphviz", "holoviews (>=1.16.0)", "hvplot", "ipyleaflet", "ipympl", "ipython (>=7.0)", "ipyvolume", "ipyvuetify", "ipywidgets", "ipywidgets-bokeh", "jupyter-bokeh (>=3.0.7)", "jupyter-server", "jupyterlab", "lxml", "matplotlib", "nbsite (>=0.8.4)", "nbval", "networkx (>=2.5)", "numba (<0.58)", "numpy", "pandas (<2.1.0)", "pandas (>=1.3)", "parameterized", "pillow", "playwright", "plotly", "plotly (>=4.0)", "pre-commit", "psutil", "pydeck", "pygraphviz", "pyinstrument (>=4.0)", "pytest", "pytest-asyncio", "pytest-cov", "pytest-playwright", "pytest-rerunfailures", "pytest-xdist", "python-graphviz", "pyvista", "reacton", "scikit-image", "scikit-learn", "scipy", "seaborn", "streamz", "textual", "tomli", "twine", "vega-datasets", "vtk", "watchfiles", "xarray", "xgboost"] +all-pip = ["aiohttp", "altair", "anywidget", "channels", "croniter", "dask-expr", "datashader", "diskcache", "django (<4)", "fastparquet", "flake8", "folium", "graphviz", "holoviews (>=1.16.0)", "hvplot", "ipyleaflet", "ipympl", "ipython (>=7.0)", "ipyvolume", "ipyvuetify", "ipywidgets", "ipywidgets-bokeh", "jupyter-bokeh (>=3.0.7)", "jupyter-server", "jupyterlab", "lxml", "matplotlib", "nbsite (>=0.8.4)", "nbval", "networkx (>=2.5)", "numba (<0.58)", "numpy", "pandas (<2.1.0)", "pandas (>=1.3)", "parameterized", "pillow", "playwright", "plotly", "plotly (>=4.0)", "pre-commit", "psutil", "pydeck", "pyinstrument (>=4.0)", "pytest", "pytest-asyncio", "pytest-cov", "pytest-playwright", "pytest-rerunfailures", "pytest-xdist", "pyvista", "reacton", "scikit-image", "scikit-learn", "scipy", "seaborn", "streamz", "textual", "tomli", "twine", "vega-datasets", "vtk", "watchfiles", "xarray", "xgboost"] +build = ["bleach", "bokeh (>=3.4.0,<3.5.0)", "cryptography (<39)", "markdown", "packaging", "param (>=2.0.0)", "pyviz-comms (>=2.0.0)", "requests", "setuptools (>=42)", "tqdm (>=4.48.0)", "urllib3 (<2.0)"] doc = ["holoviews (>=1.16.0)", "jupyterlab", "lxml", "matplotlib", "nbsite (>=0.8.4)", "pandas (<2.1.0)", "pillow", "plotly"] -examples = ["aiohttp", "altair", "channels", "croniter", "datashader", "django (<4)", "fastparquet", "folium", "graphviz", "holoviews (>=1.16.0)", "hvplot", "ipyleaflet", "ipympl", "ipyvolume", "ipyvuetify", "ipywidgets", "ipywidgets-bokeh", "jupyter-bokeh (>=3.0.7)", "networkx (>=2.5)", "plotly (>=4.0)", "pydeck", "pygraphviz", "pyinstrument (>=4.0)", "python-graphviz", "pyvista", "reacton", "scikit-image", "scikit-learn", "seaborn", "streamz", "vega-datasets", "vtk", "xarray", "xgboost"] +examples = ["aiohttp", "altair", "channels", "croniter", "dask-expr", "datashader", "django (<4)", "fastparquet", "folium", "graphviz", "holoviews (>=1.16.0)", "hvplot", "ipyleaflet", "ipympl", "ipyvolume", "ipyvuetify", "ipywidgets", "ipywidgets-bokeh", "jupyter-bokeh (>=3.0.7)", "networkx (>=2.5)", "plotly (>=4.0)", "pydeck", "pygraphviz", "pyinstrument (>=4.0)", "python-graphviz", "pyvista", "reacton", "scikit-image", "scikit-learn", "seaborn", "streamz", "textual", "vega-datasets", "vtk", "xarray", "xgboost"] recommended = ["holoviews (>=1.16.0)", "jupyterlab", "matplotlib", "pillow", "plotly"] -tests = ["altair", "anywidget", "diskcache", "flake8", "folium", "holoviews (>=1.16.0)", "ipympl", "ipython (>=7.0)", "ipyvuetify", "ipywidgets-bokeh", "nbval", "numba (<0.58)", "numpy", "pandas (>=1.3)", "parameterized", "pre-commit", "psutil", "pytest", "pytest-asyncio (<0.22)", "pytest-cov", "pytest-rerunfailures", "pytest-xdist", "reacton", "scipy", "twine"] -tests-core = ["altair", "anywidget", "diskcache", "flake8", "folium", "holoviews (>=1.16.0)", "ipython (>=7.0)", "nbval", "numpy", "pandas (>=1.3)", "parameterized", "pre-commit", "psutil", "pytest", "pytest-asyncio (<0.22)", "pytest-cov", "pytest-rerunfailures", "pytest-xdist", "scipy"] -ui = ["jupyter-server", "playwright", "pytest-playwright"] +tests = ["altair", "anywidget", "diskcache", "flake8", "folium", "holoviews (>=1.16.0)", "ipympl", "ipython (>=7.0)", "ipyvuetify", "ipywidgets-bokeh", "nbval", "numba (<0.58)", "numpy", "pandas (>=1.3)", "parameterized", "pre-commit", "psutil", "pytest", "pytest-asyncio", "pytest-cov", "pytest-rerunfailures", "pytest-xdist", "reacton", "scipy", "textual", "twine", "watchfiles"] +tests-core = ["altair", "anywidget", "diskcache", "flake8", "folium", "holoviews (>=1.16.0)", "ipython (>=7.0)", "nbval", "numpy", "pandas (>=1.3)", "parameterized", "pre-commit", "psutil", "pytest", "pytest-asyncio", "pytest-cov", "pytest-rerunfailures", "pytest-xdist", "scipy", "textual", "watchfiles"] +ui = ["jupyter-server", "playwright", "pytest-playwright", "tomli"] [[package]] name = "papermill" -version = "2.5.0" +version = "2.6.0" description = "Parameterize and run Jupyter and nteract Notebooks" optional = false python-versions = ">=3.8" files = [ - {file = "papermill-2.5.0-py3-none-any.whl", hash = "sha256:c42303afb92e482a60ae1df2577be59a5b7a64c5cd52d37c74c7f74e36085708"}, - {file = "papermill-2.5.0.tar.gz", hash = "sha256:ea7b70c0553f56fe91b0fa9cc5e17012cd699320a8b015373e7870c5e6086c72"}, + {file = "papermill-2.6.0-py3-none-any.whl", hash = "sha256:0f09da6ef709f3f14dde77cb1af052d05b14019189869affff374c9e612f2dd5"}, + {file = "papermill-2.6.0.tar.gz", hash = "sha256:9fe2a91912fd578f391b4cc8d6d105e73124dcd0cde2a43c3c4a1c77ac88ea24"}, ] [package.dependencies] -aiohttp = {version = "3.9.0b0", markers = "python_version == \"3.12\""} +aiohttp = {version = ">=3.9.0", markers = "python_version == \"3.12\""} +ansicolors = "*" click = "*" entrypoints = "*" nbclient = ">=0.2.0" -nbformat = ">=5.1.2" +nbformat = ">=5.2.0" pyyaml = "*" requests = "*" tenacity = ">=5.0.2" @@ -3018,23 +3070,23 @@ tqdm = ">=4.32.2" all = ["PyGithub (>=1.55)", "azure-datalake-store (>=0.0.30)", "azure-identity (>=1.3.1)", "azure-storage-blob (>=12.1.0)", "black (>=19.3b0)", "boto3", "gcsfs (>=0.2.0)", "pyarrow (>=2.0)", "requests (>=2.21.0)"] azure = ["azure-datalake-store (>=0.0.30)", "azure-identity (>=1.3.1)", "azure-storage-blob (>=12.1.0)", "requests (>=2.21.0)"] black = ["black (>=19.3b0)"] -dev = ["attrs (>=17.4.0)", "azure-datalake-store (>=0.0.30)", "azure-identity (>=1.3.1)", "azure-storage-blob (>=12.1.0)", "black (>=19.3b0)", "boto3", "botocore", "bumpversion", "check-manifest", "codecov", "coverage", "flake8", "gcsfs (>=0.2.0)", "google-compute-engine", "ipython (>=5.0)", "ipywidgets", "moto", "notebook", "pip (>=18.1)", "pre-commit", "pytest (>=4.1)", "pytest-cov (>=2.6.1)", "pytest-env (>=0.6.2)", "pytest-mock (>=1.10)", "recommonmark", "requests (>=2.21.0)", "setuptools (>=38.6.0)", "tox", "twine (>=1.11.0)", "wheel (>=0.31.0)"] -docs = ["PyGithub (>=1.55)", "Sphinx (>=3.5.4)", "azure-datalake-store (>=0.0.30)", "azure-identity (>=1.3.1)", "azure-storage-blob (>=12.1.0)", "black (>=19.3b0)", "boto3", "furo (>=2021.4.11b34)", "gcsfs (>=0.2.0)", "moto (>=2.0.5)", "myst-parser (>=0.13.7)", "pyarrow (>=2.0)", "requests (>=2.21.0)", "sphinx-copybutton (>=0.3.1)"] +dev = ["attrs (>=17.4.0)", "azure-datalake-store (>=0.0.30)", "azure-identity (>=1.3.1)", "azure-storage-blob (>=12.1.0)", "black (>=19.3b0)", "boto3", "botocore", "bumpversion", "check-manifest", "codecov", "coverage", "gcsfs (>=0.2.0)", "google-compute-engine", "ipython (>=5.0)", "ipywidgets", "moto (>=5.0.0,<5.1.0)", "notebook", "pip (>=18.1)", "pre-commit", "pytest (>=4.1)", "pytest-cov (>=2.6.1)", "pytest-env (>=0.6.2)", "pytest-mock (>=1.10)", "recommonmark", "requests (>=2.21.0)", "setuptools (>=38.6.0)", "tox", "twine (>=1.11.0)", "wheel (>=0.31.0)"] +docs = ["PyGithub (>=1.55)", "Sphinx (>=7.2.6)", "azure-datalake-store (>=0.0.30)", "azure-identity (>=1.3.1)", "azure-storage-blob (>=12.1.0)", "black (>=19.3b0)", "boto3", "entrypoints", "furo (>=2023.9.10)", "gcsfs (>=0.2.0)", "moto (>=4.2.8)", "myst-parser (>=2.0.0)", "nbformat", "pyarrow (>=2.0)", "requests (>=2.21.0)", "sphinx-copybutton (>=0.5.2)"] gcs = ["gcsfs (>=0.2.0)"] github = ["PyGithub (>=1.55)"] hdfs = ["pyarrow (>=2.0)"] s3 = ["boto3"] -test = ["attrs (>=17.4.0)", "azure-datalake-store (>=0.0.30)", "azure-identity (>=1.3.1)", "azure-storage-blob (>=12.1.0)", "black (>=19.3b0)", "boto3", "botocore", "bumpversion", "check-manifest", "codecov", "coverage", "flake8", "gcsfs (>=0.2.0)", "google-compute-engine", "ipython (>=5.0)", "ipywidgets", "moto", "notebook", "pip (>=18.1)", "pre-commit", "pytest (>=4.1)", "pytest-cov (>=2.6.1)", "pytest-env (>=0.6.2)", "pytest-mock (>=1.10)", "recommonmark", "requests (>=2.21.0)", "setuptools (>=38.6.0)", "tox", "twine (>=1.11.0)", "wheel (>=0.31.0)"] +test = ["attrs (>=17.4.0)", "azure-datalake-store (>=0.0.30)", "azure-identity (>=1.3.1)", "azure-storage-blob (>=12.1.0)", "black (>=19.3b0)", "boto3", "botocore", "bumpversion", "check-manifest", "codecov", "coverage", "gcsfs (>=0.2.0)", "google-compute-engine", "ipython (>=5.0)", "ipywidgets", "moto (>=5.0.0,<5.1.0)", "notebook", "pip (>=18.1)", "pre-commit", "pytest (>=4.1)", "pytest-cov (>=2.6.1)", "pytest-env (>=0.6.2)", "pytest-mock (>=1.10)", "recommonmark", "requests (>=2.21.0)", "setuptools (>=38.6.0)", "tox", "twine (>=1.11.0)", "wheel (>=0.31.0)"] [[package]] name = "param" -version = "2.0.2" +version = "2.1.0" description = "Make your Python code clearer and more reliable by declaring Parameters." optional = false python-versions = ">=3.8" files = [ - {file = "param-2.0.2-py3-none-any.whl", hash = "sha256:b269fd7397886ec609e544f81035fa52e1950da0e76d20080bfeca3d7a0317ca"}, - {file = "param-2.0.2.tar.gz", hash = "sha256:785845a727a588eb94c7666d80551c7e2bb97d4309d3507beab66f95e57f7527"}, + {file = "param-2.1.0-py3-none-any.whl", hash = "sha256:f31d3745d227347d29b5868c4e4e3077df07463889b91d3bb28e634fde211e1c"}, + {file = "param-2.1.0.tar.gz", hash = "sha256:a7b30b08b547e2b78b02aeba6ed34e3c6a638f8e4824a76a96ffa2d7cf57e71f"}, ] [package.extras] @@ -3044,33 +3096,33 @@ examples = ["aiohttp", "pandas", "panel"] lint = ["flake8", "pre-commit"] tests = ["coverage[toml]", "pytest", "pytest-asyncio"] tests-deser = ["odfpy", "openpyxl", "pyarrow", "tables", "xlrd"] -tests-examples = ["nbval", "param[examples]", "pytest", "pytest-asyncio", "pytest-xdist"] +tests-examples = ["nbval", "param[examples]", "pytest (<8.1)", "pytest-asyncio", "pytest-xdist"] tests-full = ["cloudpickle", "gmpy", "ipython", "jsonschema", "nest-asyncio", "numpy", "pandas", "param[tests-deser]", "param[tests-examples]", "param[tests]"] [[package]] name = "parso" -version = "0.8.3" +version = "0.8.4" description = "A Python Parser" optional = false python-versions = ">=3.6" files = [ - {file = "parso-0.8.3-py2.py3-none-any.whl", hash = "sha256:c001d4636cd3aecdaf33cbb40aebb59b094be2a74c556778ef5576c175e19e75"}, - {file = "parso-0.8.3.tar.gz", hash = "sha256:8c07be290bb59f03588915921e29e8a50002acaf2cdc5fa0e0114f91709fafa0"}, + {file = "parso-0.8.4-py2.py3-none-any.whl", hash = "sha256:a418670a20291dacd2dddc80c377c5c3791378ee1e8d12bffc35420643d43f18"}, + {file = "parso-0.8.4.tar.gz", hash = "sha256:eb3a7b58240fb99099a345571deecc0f9540ea5f4dd2fe14c2a99d6b281ab92d"}, ] [package.extras] -qa = ["flake8 (==3.8.3)", "mypy (==0.782)"] -testing = ["docopt", "pytest (<6.0.0)"] +qa = ["flake8 (==5.0.4)", "mypy (==0.971)", "types-setuptools (==67.2.0.1)"] +testing = ["docopt", "pytest"] [[package]] name = "partd" -version = "1.4.1" +version = "1.4.2" description = "Appendable key-value storage" optional = false -python-versions = ">=3.7" +python-versions = ">=3.9" files = [ - {file = "partd-1.4.1-py3-none-any.whl", hash = "sha256:27e766663d36c161e2827aa3e28541c992f0b9527d3cca047e13fb3acdb989e6"}, - {file = "partd-1.4.1.tar.gz", hash = "sha256:56c25dd49e6fea5727e731203c466c6e092f308d8f0024e199d02f6aa2167f67"}, + {file = "partd-1.4.2-py3-none-any.whl", hash = "sha256:978e4ac767ec4ba5b86c6eaa52e5a2a3bc748a2ca839e8cc798f1cc6ce6efb0f"}, + {file = "partd-1.4.2.tar.gz", hash = "sha256:d022c33afbdc8405c226621b015e8067888173d85f7f5ecebb3cafed9a20f02c"}, ] [package.dependencies] @@ -3078,7 +3130,7 @@ locket = "*" toolz = "*" [package.extras] -complete = ["blosc", "numpy (>=1.9.0)", "pandas (>=0.19.0)", "pyzmq"] +complete = ["blosc", "numpy (>=1.20.0)", "pandas (>=1.3)", "pyzmq"] [[package]] name = "pathspec" @@ -3107,79 +3159,80 @@ ptyprocess = ">=0.5" [[package]] name = "pillow" -version = "10.2.0" +version = "10.3.0" description = "Python Imaging Library (Fork)" optional = false python-versions = ">=3.8" files = [ - {file = "pillow-10.2.0-cp310-cp310-macosx_10_10_x86_64.whl", hash = "sha256:7823bdd049099efa16e4246bdf15e5a13dbb18a51b68fa06d6c1d4d8b99a796e"}, - {file = "pillow-10.2.0-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:83b2021f2ade7d1ed556bc50a399127d7fb245e725aa0113ebd05cfe88aaf588"}, - {file = "pillow-10.2.0-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:6fad5ff2f13d69b7e74ce5b4ecd12cc0ec530fcee76356cac6742785ff71c452"}, - {file = "pillow-10.2.0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:da2b52b37dad6d9ec64e653637a096905b258d2fc2b984c41ae7d08b938a67e4"}, - {file = "pillow-10.2.0-cp310-cp310-manylinux_2_28_aarch64.whl", hash = "sha256:47c0995fc4e7f79b5cfcab1fc437ff2890b770440f7696a3ba065ee0fd496563"}, - {file = "pillow-10.2.0-cp310-cp310-manylinux_2_28_x86_64.whl", hash = "sha256:322bdf3c9b556e9ffb18f93462e5f749d3444ce081290352c6070d014c93feb2"}, - {file = "pillow-10.2.0-cp310-cp310-musllinux_1_1_aarch64.whl", hash = "sha256:51f1a1bffc50e2e9492e87d8e09a17c5eea8409cda8d3f277eb6edc82813c17c"}, - {file = "pillow-10.2.0-cp310-cp310-musllinux_1_1_x86_64.whl", hash = "sha256:69ffdd6120a4737710a9eee73e1d2e37db89b620f702754b8f6e62594471dee0"}, - {file = "pillow-10.2.0-cp310-cp310-win32.whl", hash = "sha256:c6dafac9e0f2b3c78df97e79af707cdc5ef8e88208d686a4847bab8266870023"}, - {file = "pillow-10.2.0-cp310-cp310-win_amd64.whl", hash = "sha256:aebb6044806f2e16ecc07b2a2637ee1ef67a11840a66752751714a0d924adf72"}, - {file = "pillow-10.2.0-cp310-cp310-win_arm64.whl", hash = "sha256:7049e301399273a0136ff39b84c3678e314f2158f50f517bc50285fb5ec847ad"}, - {file = "pillow-10.2.0-cp311-cp311-macosx_10_10_x86_64.whl", hash = "sha256:35bb52c37f256f662abdfa49d2dfa6ce5d93281d323a9af377a120e89a9eafb5"}, - {file = "pillow-10.2.0-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:9c23f307202661071d94b5e384e1e1dc7dfb972a28a2310e4ee16103e66ddb67"}, - {file = "pillow-10.2.0-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:773efe0603db30c281521a7c0214cad7836c03b8ccff897beae9b47c0b657d61"}, - {file = "pillow-10.2.0-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:11fa2e5984b949b0dd6d7a94d967743d87c577ff0b83392f17cb3990d0d2fd6e"}, - {file = "pillow-10.2.0-cp311-cp311-manylinux_2_28_aarch64.whl", hash = "sha256:716d30ed977be8b37d3ef185fecb9e5a1d62d110dfbdcd1e2a122ab46fddb03f"}, - {file = "pillow-10.2.0-cp311-cp311-manylinux_2_28_x86_64.whl", hash = "sha256:a086c2af425c5f62a65e12fbf385f7c9fcb8f107d0849dba5839461a129cf311"}, - {file = "pillow-10.2.0-cp311-cp311-musllinux_1_1_aarch64.whl", hash = "sha256:c8de2789052ed501dd829e9cae8d3dcce7acb4777ea4a479c14521c942d395b1"}, - {file = "pillow-10.2.0-cp311-cp311-musllinux_1_1_x86_64.whl", hash = "sha256:609448742444d9290fd687940ac0b57fb35e6fd92bdb65386e08e99af60bf757"}, - {file = "pillow-10.2.0-cp311-cp311-win32.whl", hash = "sha256:823ef7a27cf86df6597fa0671066c1b596f69eba53efa3d1e1cb8b30f3533068"}, - {file = "pillow-10.2.0-cp311-cp311-win_amd64.whl", hash = "sha256:1da3b2703afd040cf65ec97efea81cfba59cdbed9c11d8efc5ab09df9509fc56"}, - {file = "pillow-10.2.0-cp311-cp311-win_arm64.whl", hash = "sha256:edca80cbfb2b68d7b56930b84a0e45ae1694aeba0541f798e908a49d66b837f1"}, - {file = "pillow-10.2.0-cp312-cp312-macosx_10_10_x86_64.whl", hash = "sha256:1b5e1b74d1bd1b78bc3477528919414874748dd363e6272efd5abf7654e68bef"}, - {file = "pillow-10.2.0-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:0eae2073305f451d8ecacb5474997c08569fb4eb4ac231ffa4ad7d342fdc25ac"}, - {file = "pillow-10.2.0-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:b7c2286c23cd350b80d2fc9d424fc797575fb16f854b831d16fd47ceec078f2c"}, - {file = "pillow-10.2.0-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:1e23412b5c41e58cec602f1135c57dfcf15482013ce6e5f093a86db69646a5aa"}, - {file = "pillow-10.2.0-cp312-cp312-manylinux_2_28_aarch64.whl", hash = "sha256:52a50aa3fb3acb9cf7213573ef55d31d6eca37f5709c69e6858fe3bc04a5c2a2"}, - {file = "pillow-10.2.0-cp312-cp312-manylinux_2_28_x86_64.whl", hash = "sha256:127cee571038f252a552760076407f9cff79761c3d436a12af6000cd182a9d04"}, - {file = "pillow-10.2.0-cp312-cp312-musllinux_1_1_aarch64.whl", hash = "sha256:8d12251f02d69d8310b046e82572ed486685c38f02176bd08baf216746eb947f"}, - {file = "pillow-10.2.0-cp312-cp312-musllinux_1_1_x86_64.whl", hash = "sha256:54f1852cd531aa981bc0965b7d609f5f6cc8ce8c41b1139f6ed6b3c54ab82bfb"}, - {file = "pillow-10.2.0-cp312-cp312-win32.whl", hash = "sha256:257d8788df5ca62c980314053197f4d46eefedf4e6175bc9412f14412ec4ea2f"}, - {file = "pillow-10.2.0-cp312-cp312-win_amd64.whl", hash = "sha256:154e939c5f0053a383de4fd3d3da48d9427a7e985f58af8e94d0b3c9fcfcf4f9"}, - {file = "pillow-10.2.0-cp312-cp312-win_arm64.whl", hash = "sha256:f379abd2f1e3dddb2b61bc67977a6b5a0a3f7485538bcc6f39ec76163891ee48"}, - {file = "pillow-10.2.0-cp38-cp38-macosx_10_10_x86_64.whl", hash = "sha256:8373c6c251f7ef8bda6675dd6d2b3a0fcc31edf1201266b5cf608b62a37407f9"}, - {file = "pillow-10.2.0-cp38-cp38-macosx_11_0_arm64.whl", hash = "sha256:870ea1ada0899fd0b79643990809323b389d4d1d46c192f97342eeb6ee0b8483"}, - {file = "pillow-10.2.0-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:b4b6b1e20608493548b1f32bce8cca185bf0480983890403d3b8753e44077129"}, - {file = "pillow-10.2.0-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:3031709084b6e7852d00479fd1d310b07d0ba82765f973b543c8af5061cf990e"}, - {file = "pillow-10.2.0-cp38-cp38-manylinux_2_28_aarch64.whl", hash = "sha256:3ff074fc97dd4e80543a3e91f69d58889baf2002b6be64347ea8cf5533188213"}, - {file = "pillow-10.2.0-cp38-cp38-manylinux_2_28_x86_64.whl", hash = "sha256:cb4c38abeef13c61d6916f264d4845fab99d7b711be96c326b84df9e3e0ff62d"}, - {file = "pillow-10.2.0-cp38-cp38-musllinux_1_1_aarch64.whl", hash = "sha256:b1b3020d90c2d8e1dae29cf3ce54f8094f7938460fb5ce8bc5c01450b01fbaf6"}, - {file = "pillow-10.2.0-cp38-cp38-musllinux_1_1_x86_64.whl", hash = "sha256:170aeb00224ab3dc54230c797f8404507240dd868cf52066f66a41b33169bdbe"}, - {file = "pillow-10.2.0-cp38-cp38-win32.whl", hash = "sha256:c4225f5220f46b2fde568c74fca27ae9771536c2e29d7c04f4fb62c83275ac4e"}, - {file = "pillow-10.2.0-cp38-cp38-win_amd64.whl", hash = "sha256:0689b5a8c5288bc0504d9fcee48f61a6a586b9b98514d7d29b840143d6734f39"}, - {file = "pillow-10.2.0-cp39-cp39-macosx_10_10_x86_64.whl", hash = "sha256:b792a349405fbc0163190fde0dc7b3fef3c9268292586cf5645598b48e63dc67"}, - {file = "pillow-10.2.0-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:c570f24be1e468e3f0ce7ef56a89a60f0e05b30a3669a459e419c6eac2c35364"}, - {file = "pillow-10.2.0-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:d8ecd059fdaf60c1963c58ceb8997b32e9dc1b911f5da5307aab614f1ce5c2fb"}, - {file = "pillow-10.2.0-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:c365fd1703040de1ec284b176d6af5abe21b427cb3a5ff68e0759e1e313a5e7e"}, - {file = "pillow-10.2.0-cp39-cp39-manylinux_2_28_aarch64.whl", hash = "sha256:70c61d4c475835a19b3a5aa42492409878bbca7438554a1f89d20d58a7c75c01"}, - {file = "pillow-10.2.0-cp39-cp39-manylinux_2_28_x86_64.whl", hash = "sha256:b6f491cdf80ae540738859d9766783e3b3c8e5bd37f5dfa0b76abdecc5081f13"}, - {file = "pillow-10.2.0-cp39-cp39-musllinux_1_1_aarch64.whl", hash = "sha256:9d189550615b4948f45252d7f005e53c2040cea1af5b60d6f79491a6e147eef7"}, - {file = "pillow-10.2.0-cp39-cp39-musllinux_1_1_x86_64.whl", hash = "sha256:49d9ba1ed0ef3e061088cd1e7538a0759aab559e2e0a80a36f9fd9d8c0c21591"}, - {file = "pillow-10.2.0-cp39-cp39-win32.whl", hash = "sha256:babf5acfede515f176833ed6028754cbcd0d206f7f614ea3447d67c33be12516"}, - {file = "pillow-10.2.0-cp39-cp39-win_amd64.whl", hash = "sha256:0304004f8067386b477d20a518b50f3fa658a28d44e4116970abfcd94fac34a8"}, - {file = "pillow-10.2.0-cp39-cp39-win_arm64.whl", hash = "sha256:0fb3e7fc88a14eacd303e90481ad983fd5b69c761e9e6ef94c983f91025da869"}, - {file = "pillow-10.2.0-pp310-pypy310_pp73-macosx_10_10_x86_64.whl", hash = "sha256:322209c642aabdd6207517e9739c704dc9f9db943015535783239022002f054a"}, - {file = "pillow-10.2.0-pp310-pypy310_pp73-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:3eedd52442c0a5ff4f887fab0c1c0bb164d8635b32c894bc1faf4c618dd89df2"}, - {file = "pillow-10.2.0-pp310-pypy310_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:cb28c753fd5eb3dd859b4ee95de66cc62af91bcff5db5f2571d32a520baf1f04"}, - {file = "pillow-10.2.0-pp310-pypy310_pp73-manylinux_2_28_aarch64.whl", hash = "sha256:33870dc4653c5017bf4c8873e5488d8f8d5f8935e2f1fb9a2208c47cdd66efd2"}, - {file = "pillow-10.2.0-pp310-pypy310_pp73-manylinux_2_28_x86_64.whl", hash = "sha256:3c31822339516fb3c82d03f30e22b1d038da87ef27b6a78c9549888f8ceda39a"}, - {file = "pillow-10.2.0-pp310-pypy310_pp73-win_amd64.whl", hash = "sha256:a2b56ba36e05f973d450582fb015594aaa78834fefe8dfb8fcd79b93e64ba4c6"}, - {file = "pillow-10.2.0-pp38-pypy38_pp73-win_amd64.whl", hash = "sha256:d8e6aeb9201e655354b3ad049cb77d19813ad4ece0df1249d3c793de3774f8c7"}, - {file = "pillow-10.2.0-pp39-pypy39_pp73-macosx_10_10_x86_64.whl", hash = "sha256:2247178effb34a77c11c0e8ac355c7a741ceca0a732b27bf11e747bbc950722f"}, - {file = "pillow-10.2.0-pp39-pypy39_pp73-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:15587643b9e5eb26c48e49a7b33659790d28f190fc514a322d55da2fb5c2950e"}, - {file = "pillow-10.2.0-pp39-pypy39_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:753cd8f2086b2b80180d9b3010dd4ed147efc167c90d3bf593fe2af21265e5a5"}, - {file = "pillow-10.2.0-pp39-pypy39_pp73-manylinux_2_28_aarch64.whl", hash = "sha256:7c8f97e8e7a9009bcacbe3766a36175056c12f9a44e6e6f2d5caad06dcfbf03b"}, - {file = "pillow-10.2.0-pp39-pypy39_pp73-manylinux_2_28_x86_64.whl", hash = "sha256:d1b35bcd6c5543b9cb547dee3150c93008f8dd0f1fef78fc0cd2b141c5baf58a"}, - {file = "pillow-10.2.0-pp39-pypy39_pp73-win_amd64.whl", hash = "sha256:fe4c15f6c9285dc54ce6553a3ce908ed37c8f3825b5a51a15c91442bb955b868"}, - {file = "pillow-10.2.0.tar.gz", hash = "sha256:e87f0b2c78157e12d7686b27d63c070fd65d994e8ddae6f328e0dcf4a0cd007e"}, + {file = "pillow-10.3.0-cp310-cp310-macosx_10_10_x86_64.whl", hash = "sha256:90b9e29824800e90c84e4022dd5cc16eb2d9605ee13f05d47641eb183cd73d45"}, + {file = "pillow-10.3.0-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:a2c405445c79c3f5a124573a051062300936b0281fee57637e706453e452746c"}, + {file = "pillow-10.3.0-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:78618cdbccaa74d3f88d0ad6cb8ac3007f1a6fa5c6f19af64b55ca170bfa1edf"}, + {file = "pillow-10.3.0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:261ddb7ca91fcf71757979534fb4c128448b5b4c55cb6152d280312062f69599"}, + {file = "pillow-10.3.0-cp310-cp310-manylinux_2_28_aarch64.whl", hash = "sha256:ce49c67f4ea0609933d01c0731b34b8695a7a748d6c8d186f95e7d085d2fe475"}, + {file = "pillow-10.3.0-cp310-cp310-manylinux_2_28_x86_64.whl", hash = "sha256:b14f16f94cbc61215115b9b1236f9c18403c15dd3c52cf629072afa9d54c1cbf"}, + {file = "pillow-10.3.0-cp310-cp310-musllinux_1_1_aarch64.whl", hash = "sha256:d33891be6df59d93df4d846640f0e46f1a807339f09e79a8040bc887bdcd7ed3"}, + {file = "pillow-10.3.0-cp310-cp310-musllinux_1_1_x86_64.whl", hash = "sha256:b50811d664d392f02f7761621303eba9d1b056fb1868c8cdf4231279645c25f5"}, + {file = "pillow-10.3.0-cp310-cp310-win32.whl", hash = "sha256:ca2870d5d10d8726a27396d3ca4cf7976cec0f3cb706debe88e3a5bd4610f7d2"}, + {file = "pillow-10.3.0-cp310-cp310-win_amd64.whl", hash = "sha256:f0d0591a0aeaefdaf9a5e545e7485f89910c977087e7de2b6c388aec32011e9f"}, + {file = "pillow-10.3.0-cp310-cp310-win_arm64.whl", hash = "sha256:ccce24b7ad89adb5a1e34a6ba96ac2530046763912806ad4c247356a8f33a67b"}, + {file = "pillow-10.3.0-cp311-cp311-macosx_10_10_x86_64.whl", hash = "sha256:5f77cf66e96ae734717d341c145c5949c63180842a545c47a0ce7ae52ca83795"}, + {file = "pillow-10.3.0-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:e4b878386c4bf293578b48fc570b84ecfe477d3b77ba39a6e87150af77f40c57"}, + {file = "pillow-10.3.0-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:fdcbb4068117dfd9ce0138d068ac512843c52295ed996ae6dd1faf537b6dbc27"}, + {file = "pillow-10.3.0-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:9797a6c8fe16f25749b371c02e2ade0efb51155e767a971c61734b1bf6293994"}, + {file = "pillow-10.3.0-cp311-cp311-manylinux_2_28_aarch64.whl", hash = "sha256:9e91179a242bbc99be65e139e30690e081fe6cb91a8e77faf4c409653de39451"}, + {file = "pillow-10.3.0-cp311-cp311-manylinux_2_28_x86_64.whl", hash = "sha256:1b87bd9d81d179bd8ab871603bd80d8645729939f90b71e62914e816a76fc6bd"}, + {file = "pillow-10.3.0-cp311-cp311-musllinux_1_1_aarch64.whl", hash = "sha256:81d09caa7b27ef4e61cb7d8fbf1714f5aec1c6b6c5270ee53504981e6e9121ad"}, + {file = "pillow-10.3.0-cp311-cp311-musllinux_1_1_x86_64.whl", hash = "sha256:048ad577748b9fa4a99a0548c64f2cb8d672d5bf2e643a739ac8faff1164238c"}, + {file = "pillow-10.3.0-cp311-cp311-win32.whl", hash = "sha256:7161ec49ef0800947dc5570f86568a7bb36fa97dd09e9827dc02b718c5643f09"}, + {file = "pillow-10.3.0-cp311-cp311-win_amd64.whl", hash = "sha256:8eb0908e954d093b02a543dc963984d6e99ad2b5e36503d8a0aaf040505f747d"}, + {file = "pillow-10.3.0-cp311-cp311-win_arm64.whl", hash = "sha256:4e6f7d1c414191c1199f8996d3f2282b9ebea0945693fb67392c75a3a320941f"}, + {file = "pillow-10.3.0-cp312-cp312-macosx_10_10_x86_64.whl", hash = "sha256:e46f38133e5a060d46bd630faa4d9fa0202377495df1f068a8299fd78c84de84"}, + {file = "pillow-10.3.0-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:50b8eae8f7334ec826d6eeffaeeb00e36b5e24aa0b9df322c247539714c6df19"}, + {file = "pillow-10.3.0-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:9d3bea1c75f8c53ee4d505c3e67d8c158ad4df0d83170605b50b64025917f338"}, + {file = "pillow-10.3.0-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:19aeb96d43902f0a783946a0a87dbdad5c84c936025b8419da0a0cd7724356b1"}, + {file = "pillow-10.3.0-cp312-cp312-manylinux_2_28_aarch64.whl", hash = "sha256:74d28c17412d9caa1066f7a31df8403ec23d5268ba46cd0ad2c50fb82ae40462"}, + {file = "pillow-10.3.0-cp312-cp312-manylinux_2_28_x86_64.whl", hash = "sha256:ff61bfd9253c3915e6d41c651d5f962da23eda633cf02262990094a18a55371a"}, + {file = "pillow-10.3.0-cp312-cp312-musllinux_1_1_aarch64.whl", hash = "sha256:d886f5d353333b4771d21267c7ecc75b710f1a73d72d03ca06df49b09015a9ef"}, + {file = "pillow-10.3.0-cp312-cp312-musllinux_1_1_x86_64.whl", hash = "sha256:4b5ec25d8b17217d635f8935dbc1b9aa5907962fae29dff220f2659487891cd3"}, + {file = "pillow-10.3.0-cp312-cp312-win32.whl", hash = "sha256:51243f1ed5161b9945011a7360e997729776f6e5d7005ba0c6879267d4c5139d"}, + {file = "pillow-10.3.0-cp312-cp312-win_amd64.whl", hash = "sha256:412444afb8c4c7a6cc11a47dade32982439925537e483be7c0ae0cf96c4f6a0b"}, + {file = "pillow-10.3.0-cp312-cp312-win_arm64.whl", hash = "sha256:798232c92e7665fe82ac085f9d8e8ca98826f8e27859d9a96b41d519ecd2e49a"}, + {file = "pillow-10.3.0-cp38-cp38-macosx_10_10_x86_64.whl", hash = "sha256:4eaa22f0d22b1a7e93ff0a596d57fdede2e550aecffb5a1ef1106aaece48e96b"}, + {file = "pillow-10.3.0-cp38-cp38-macosx_11_0_arm64.whl", hash = "sha256:cd5e14fbf22a87321b24c88669aad3a51ec052eb145315b3da3b7e3cc105b9a2"}, + {file = "pillow-10.3.0-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:1530e8f3a4b965eb6a7785cf17a426c779333eb62c9a7d1bbcf3ffd5bf77a4aa"}, + {file = "pillow-10.3.0-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:5d512aafa1d32efa014fa041d38868fda85028e3f930a96f85d49c7d8ddc0383"}, + {file = "pillow-10.3.0-cp38-cp38-manylinux_2_28_aarch64.whl", hash = "sha256:339894035d0ede518b16073bdc2feef4c991ee991a29774b33e515f1d308e08d"}, + {file = "pillow-10.3.0-cp38-cp38-manylinux_2_28_x86_64.whl", hash = "sha256:aa7e402ce11f0885305bfb6afb3434b3cd8f53b563ac065452d9d5654c7b86fd"}, + {file = "pillow-10.3.0-cp38-cp38-musllinux_1_1_aarch64.whl", hash = "sha256:0ea2a783a2bdf2a561808fe4a7a12e9aa3799b701ba305de596bc48b8bdfce9d"}, + {file = "pillow-10.3.0-cp38-cp38-musllinux_1_1_x86_64.whl", hash = "sha256:c78e1b00a87ce43bb37642c0812315b411e856a905d58d597750eb79802aaaa3"}, + {file = "pillow-10.3.0-cp38-cp38-win32.whl", hash = "sha256:72d622d262e463dfb7595202d229f5f3ab4b852289a1cd09650362db23b9eb0b"}, + {file = "pillow-10.3.0-cp38-cp38-win_amd64.whl", hash = "sha256:2034f6759a722da3a3dbd91a81148cf884e91d1b747992ca288ab88c1de15999"}, + {file = "pillow-10.3.0-cp39-cp39-macosx_10_10_x86_64.whl", hash = "sha256:2ed854e716a89b1afcedea551cd85f2eb2a807613752ab997b9974aaa0d56936"}, + {file = "pillow-10.3.0-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:dc1a390a82755a8c26c9964d457d4c9cbec5405896cba94cf51f36ea0d855002"}, + {file = "pillow-10.3.0-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:4203efca580f0dd6f882ca211f923168548f7ba334c189e9eab1178ab840bf60"}, + {file = "pillow-10.3.0-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:3102045a10945173d38336f6e71a8dc71bcaeed55c3123ad4af82c52807b9375"}, + {file = "pillow-10.3.0-cp39-cp39-manylinux_2_28_aarch64.whl", hash = "sha256:6fb1b30043271ec92dc65f6d9f0b7a830c210b8a96423074b15c7bc999975f57"}, + {file = "pillow-10.3.0-cp39-cp39-manylinux_2_28_x86_64.whl", hash = "sha256:1dfc94946bc60ea375cc39cff0b8da6c7e5f8fcdc1d946beb8da5c216156ddd8"}, + {file = "pillow-10.3.0-cp39-cp39-musllinux_1_1_aarch64.whl", hash = "sha256:b09b86b27a064c9624d0a6c54da01c1beaf5b6cadfa609cf63789b1d08a797b9"}, + {file = "pillow-10.3.0-cp39-cp39-musllinux_1_1_x86_64.whl", hash = "sha256:d3b2348a78bc939b4fed6552abfd2e7988e0f81443ef3911a4b8498ca084f6eb"}, + {file = "pillow-10.3.0-cp39-cp39-win32.whl", hash = "sha256:45ebc7b45406febf07fef35d856f0293a92e7417ae7933207e90bf9090b70572"}, + {file = "pillow-10.3.0-cp39-cp39-win_amd64.whl", hash = "sha256:0ba26351b137ca4e0db0342d5d00d2e355eb29372c05afd544ebf47c0956ffeb"}, + {file = "pillow-10.3.0-cp39-cp39-win_arm64.whl", hash = "sha256:50fd3f6b26e3441ae07b7c979309638b72abc1a25da31a81a7fbd9495713ef4f"}, + {file = "pillow-10.3.0-pp310-pypy310_pp73-macosx_10_10_x86_64.whl", hash = "sha256:6b02471b72526ab8a18c39cb7967b72d194ec53c1fd0a70b050565a0f366d355"}, + {file = "pillow-10.3.0-pp310-pypy310_pp73-macosx_11_0_arm64.whl", hash = "sha256:8ab74c06ffdab957d7670c2a5a6e1a70181cd10b727cd788c4dd9005b6a8acd9"}, + {file = "pillow-10.3.0-pp310-pypy310_pp73-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:048eeade4c33fdf7e08da40ef402e748df113fd0b4584e32c4af74fe78baaeb2"}, + {file = "pillow-10.3.0-pp310-pypy310_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:9e2ec1e921fd07c7cda7962bad283acc2f2a9ccc1b971ee4b216b75fad6f0463"}, + {file = "pillow-10.3.0-pp310-pypy310_pp73-manylinux_2_28_aarch64.whl", hash = "sha256:4c8e73e99da7db1b4cad7f8d682cf6abad7844da39834c288fbfa394a47bbced"}, + {file = "pillow-10.3.0-pp310-pypy310_pp73-manylinux_2_28_x86_64.whl", hash = "sha256:16563993329b79513f59142a6b02055e10514c1a8e86dca8b48a893e33cf91e3"}, + {file = "pillow-10.3.0-pp310-pypy310_pp73-win_amd64.whl", hash = "sha256:dd78700f5788ae180b5ee8902c6aea5a5726bac7c364b202b4b3e3ba2d293170"}, + {file = "pillow-10.3.0-pp39-pypy39_pp73-macosx_10_10_x86_64.whl", hash = "sha256:aff76a55a8aa8364d25400a210a65ff59d0168e0b4285ba6bf2bd83cf675ba32"}, + {file = "pillow-10.3.0-pp39-pypy39_pp73-macosx_11_0_arm64.whl", hash = "sha256:b7bc2176354defba3edc2b9a777744462da2f8e921fbaf61e52acb95bafa9828"}, + {file = "pillow-10.3.0-pp39-pypy39_pp73-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:793b4e24db2e8742ca6423d3fde8396db336698c55cd34b660663ee9e45ed37f"}, + {file = "pillow-10.3.0-pp39-pypy39_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:d93480005693d247f8346bc8ee28c72a2191bdf1f6b5db469c096c0c867ac015"}, + {file = "pillow-10.3.0-pp39-pypy39_pp73-manylinux_2_28_aarch64.whl", hash = "sha256:c83341b89884e2b2e55886e8fbbf37c3fa5efd6c8907124aeb72f285ae5696e5"}, + {file = "pillow-10.3.0-pp39-pypy39_pp73-manylinux_2_28_x86_64.whl", hash = "sha256:1a1d1915db1a4fdb2754b9de292642a39a7fb28f1736699527bb649484fb966a"}, + {file = "pillow-10.3.0-pp39-pypy39_pp73-win_amd64.whl", hash = "sha256:a0eaa93d054751ee9964afa21c06247779b90440ca41d184aeb5d410f20ff591"}, + {file = "pillow-10.3.0.tar.gz", hash = "sha256:9d2455fbf44c914840c793e89aa82d0e1763a14253a000743719ae5946814b2d"}, ] [package.extras] @@ -3192,28 +3245,29 @@ xmp = ["defusedxml"] [[package]] name = "platformdirs" -version = "4.2.0" -description = "A small Python package for determining appropriate platform-specific dirs, e.g. a \"user data dir\"." +version = "4.2.2" +description = "A small Python package for determining appropriate platform-specific dirs, e.g. a `user data dir`." optional = false python-versions = ">=3.8" files = [ - {file = "platformdirs-4.2.0-py3-none-any.whl", hash = "sha256:0614df2a2f37e1a662acbd8e2b25b92ccf8632929bc6d43467e17fe89c75e068"}, - {file = "platformdirs-4.2.0.tar.gz", hash = "sha256:ef0cc731df711022c174543cb70a9b5bd22e5a9337c8624ef2c2ceb8ddad8768"}, + {file = "platformdirs-4.2.2-py3-none-any.whl", hash = "sha256:2d7a1657e36a80ea911db832a8a6ece5ee53d8de21edd5cc5879af6530b1bfee"}, + {file = "platformdirs-4.2.2.tar.gz", hash = "sha256:38b7b51f512eed9e84a22788b4bce1de17c0adb134d6becb09836e37d8654cd3"}, ] [package.extras] docs = ["furo (>=2023.9.10)", "proselint (>=0.13)", "sphinx (>=7.2.6)", "sphinx-autodoc-typehints (>=1.25.2)"] test = ["appdirs (==1.4.4)", "covdefaults (>=2.3)", "pytest (>=7.4.3)", "pytest-cov (>=4.1)", "pytest-mock (>=3.12)"] +type = ["mypy (>=1.8)"] [[package]] name = "pluggy" -version = "1.4.0" +version = "1.5.0" description = "plugin and hook calling mechanisms for python" optional = false python-versions = ">=3.8" files = [ - {file = "pluggy-1.4.0-py3-none-any.whl", hash = "sha256:7db9f7b503d67d1c5b95f59773ebb58a8c1c288129a88665838012cfb07b8981"}, - {file = "pluggy-1.4.0.tar.gz", hash = "sha256:8c85c2876142a764e5b7548e7d9a0e0ddb46f5185161049a79b7e974454223be"}, + {file = "pluggy-1.5.0-py3-none-any.whl", hash = "sha256:44e1ad92c8ca002de6377e165f3e0f1be63266ab4d554740532335b9d75ea669"}, + {file = "pluggy-1.5.0.tar.gz", hash = "sha256:2cffa88e94fdc978c4c574f15f9e59b7f4201d439195c3715ca9e2486f1d0cf1"}, ] [package.extras] @@ -3222,13 +3276,13 @@ testing = ["pytest", "pytest-benchmark"] [[package]] name = "pooch" -version = "1.8.0" +version = "1.8.1" description = "\"Pooch manages your Python library's sample data files: it automatically downloads and stores them in a local directory, with support for versioning and corruption checks.\"" optional = false python-versions = ">=3.7" files = [ - {file = "pooch-1.8.0-py3-none-any.whl", hash = "sha256:1bfba436d9e2ad5199ccad3583cca8c241b8736b5bb23fe67c213d52650dbb66"}, - {file = "pooch-1.8.0.tar.gz", hash = "sha256:f59981fd5b9b5d032dcde8f4a11eaa492c2ac6343fae3596a2fdae35fc54b0a0"}, + {file = "pooch-1.8.1-py3-none-any.whl", hash = "sha256:6b56611ac320c239faece1ac51a60b25796792599ce5c0b1bb87bf01df55e0a9"}, + {file = "pooch-1.8.1.tar.gz", hash = "sha256:27ef63097dd9a6e4f9d2694f5cfbf2f0a5defa44fccafec08d601e731d746270"}, ] [package.dependencies] @@ -3310,51 +3364,51 @@ tests = ["pytest"] [[package]] name = "pyarrow" -version = "15.0.0" +version = "16.1.0" description = "Python library for Apache Arrow" optional = false python-versions = ">=3.8" files = [ - {file = "pyarrow-15.0.0-cp310-cp310-macosx_10_15_x86_64.whl", hash = "sha256:0a524532fd6dd482edaa563b686d754c70417c2f72742a8c990b322d4c03a15d"}, - {file = "pyarrow-15.0.0-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:60a6bdb314affa9c2e0d5dddf3d9cbb9ef4a8dddaa68669975287d47ece67642"}, - {file = "pyarrow-15.0.0-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:66958fd1771a4d4b754cd385835e66a3ef6b12611e001d4e5edfcef5f30391e2"}, - {file = "pyarrow-15.0.0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:1f500956a49aadd907eaa21d4fff75f73954605eaa41f61cb94fb008cf2e00c6"}, - {file = "pyarrow-15.0.0-cp310-cp310-manylinux_2_28_aarch64.whl", hash = "sha256:6f87d9c4f09e049c2cade559643424da84c43a35068f2a1c4653dc5b1408a929"}, - {file = "pyarrow-15.0.0-cp310-cp310-manylinux_2_28_x86_64.whl", hash = "sha256:85239b9f93278e130d86c0e6bb455dcb66fc3fd891398b9d45ace8799a871a1e"}, - {file = "pyarrow-15.0.0-cp310-cp310-win_amd64.whl", hash = "sha256:5b8d43e31ca16aa6e12402fcb1e14352d0d809de70edd185c7650fe80e0769e3"}, - {file = "pyarrow-15.0.0-cp311-cp311-macosx_10_15_x86_64.whl", hash = "sha256:fa7cd198280dbd0c988df525e50e35b5d16873e2cdae2aaaa6363cdb64e3eec5"}, - {file = "pyarrow-15.0.0-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:8780b1a29d3c8b21ba6b191305a2a607de2e30dab399776ff0aa09131e266340"}, - {file = "pyarrow-15.0.0-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:fe0ec198ccc680f6c92723fadcb97b74f07c45ff3fdec9dd765deb04955ccf19"}, - {file = "pyarrow-15.0.0-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:036a7209c235588c2f07477fe75c07e6caced9b7b61bb897c8d4e52c4b5f9555"}, - {file = "pyarrow-15.0.0-cp311-cp311-manylinux_2_28_aarch64.whl", hash = "sha256:2bd8a0e5296797faf9a3294e9fa2dc67aa7f10ae2207920dbebb785c77e9dbe5"}, - {file = "pyarrow-15.0.0-cp311-cp311-manylinux_2_28_x86_64.whl", hash = "sha256:e8ebed6053dbe76883a822d4e8da36860f479d55a762bd9e70d8494aed87113e"}, - {file = "pyarrow-15.0.0-cp311-cp311-win_amd64.whl", hash = "sha256:17d53a9d1b2b5bd7d5e4cd84d018e2a45bc9baaa68f7e6e3ebed45649900ba99"}, - {file = "pyarrow-15.0.0-cp312-cp312-macosx_10_15_x86_64.whl", hash = "sha256:9950a9c9df24090d3d558b43b97753b8f5867fb8e521f29876aa021c52fda351"}, - {file = "pyarrow-15.0.0-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:003d680b5e422d0204e7287bb3fa775b332b3fce2996aa69e9adea23f5c8f970"}, - {file = "pyarrow-15.0.0-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:f75fce89dad10c95f4bf590b765e3ae98bcc5ba9f6ce75adb828a334e26a3d40"}, - {file = "pyarrow-15.0.0-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:0ca9cb0039923bec49b4fe23803807e4ef39576a2bec59c32b11296464623dc2"}, - {file = "pyarrow-15.0.0-cp312-cp312-manylinux_2_28_aarch64.whl", hash = "sha256:9ed5a78ed29d171d0acc26a305a4b7f83c122d54ff5270810ac23c75813585e4"}, - {file = "pyarrow-15.0.0-cp312-cp312-manylinux_2_28_x86_64.whl", hash = "sha256:6eda9e117f0402dfcd3cd6ec9bfee89ac5071c48fc83a84f3075b60efa96747f"}, - {file = "pyarrow-15.0.0-cp312-cp312-win_amd64.whl", hash = "sha256:9a3a6180c0e8f2727e6f1b1c87c72d3254cac909e609f35f22532e4115461177"}, - {file = "pyarrow-15.0.0-cp38-cp38-macosx_10_15_x86_64.whl", hash = "sha256:19a8918045993349b207de72d4576af0191beef03ea655d8bdb13762f0cd6eac"}, - {file = "pyarrow-15.0.0-cp38-cp38-macosx_11_0_arm64.whl", hash = "sha256:d0ec076b32bacb6666e8813a22e6e5a7ef1314c8069d4ff345efa6246bc38593"}, - {file = "pyarrow-15.0.0-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:5db1769e5d0a77eb92344c7382d6543bea1164cca3704f84aa44e26c67e320fb"}, - {file = "pyarrow-15.0.0-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:e2617e3bf9df2a00020dd1c1c6dce5cc343d979efe10bc401c0632b0eef6ef5b"}, - {file = "pyarrow-15.0.0-cp38-cp38-manylinux_2_28_aarch64.whl", hash = "sha256:d31c1d45060180131caf10f0f698e3a782db333a422038bf7fe01dace18b3a31"}, - {file = "pyarrow-15.0.0-cp38-cp38-manylinux_2_28_x86_64.whl", hash = "sha256:c8c287d1d479de8269398b34282e206844abb3208224dbdd7166d580804674b7"}, - {file = "pyarrow-15.0.0-cp38-cp38-win_amd64.whl", hash = "sha256:07eb7f07dc9ecbb8dace0f58f009d3a29ee58682fcdc91337dfeb51ea618a75b"}, - {file = "pyarrow-15.0.0-cp39-cp39-macosx_10_15_x86_64.whl", hash = "sha256:47af7036f64fce990bb8a5948c04722e4e3ea3e13b1007ef52dfe0aa8f23cf7f"}, - {file = "pyarrow-15.0.0-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:93768ccfff85cf044c418bfeeafce9a8bb0cee091bd8fd19011aff91e58de540"}, - {file = "pyarrow-15.0.0-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:f6ee87fd6892700960d90abb7b17a72a5abb3b64ee0fe8db6c782bcc2d0dc0b4"}, - {file = "pyarrow-15.0.0-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:001fca027738c5f6be0b7a3159cc7ba16a5c52486db18160909a0831b063c4e4"}, - {file = "pyarrow-15.0.0-cp39-cp39-manylinux_2_28_aarch64.whl", hash = "sha256:d1c48648f64aec09accf44140dccb92f4f94394b8d79976c426a5b79b11d4fa7"}, - {file = "pyarrow-15.0.0-cp39-cp39-manylinux_2_28_x86_64.whl", hash = "sha256:972a0141be402bb18e3201448c8ae62958c9c7923dfaa3b3d4530c835ac81aed"}, - {file = "pyarrow-15.0.0-cp39-cp39-win_amd64.whl", hash = "sha256:f01fc5cf49081426429127aa2d427d9d98e1cb94a32cb961d583a70b7c4504e6"}, - {file = "pyarrow-15.0.0.tar.gz", hash = "sha256:876858f549d540898f927eba4ef77cd549ad8d24baa3207cf1b72e5788b50e83"}, + {file = "pyarrow-16.1.0-cp310-cp310-macosx_10_15_x86_64.whl", hash = "sha256:17e23b9a65a70cc733d8b738baa6ad3722298fa0c81d88f63ff94bf25eaa77b9"}, + {file = "pyarrow-16.1.0-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:4740cc41e2ba5d641071d0ab5e9ef9b5e6e8c7611351a5cb7c1d175eaf43674a"}, + {file = "pyarrow-16.1.0-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:98100e0268d04e0eec47b73f20b39c45b4006f3c4233719c3848aa27a03c1aef"}, + {file = "pyarrow-16.1.0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:f68f409e7b283c085f2da014f9ef81e885d90dcd733bd648cfba3ef265961848"}, + {file = "pyarrow-16.1.0-cp310-cp310-manylinux_2_28_aarch64.whl", hash = "sha256:a8914cd176f448e09746037b0c6b3a9d7688cef451ec5735094055116857580c"}, + {file = "pyarrow-16.1.0-cp310-cp310-manylinux_2_28_x86_64.whl", hash = "sha256:48be160782c0556156d91adbdd5a4a7e719f8d407cb46ae3bb4eaee09b3111bd"}, + {file = "pyarrow-16.1.0-cp310-cp310-win_amd64.whl", hash = "sha256:9cf389d444b0f41d9fe1444b70650fea31e9d52cfcb5f818b7888b91b586efff"}, + {file = "pyarrow-16.1.0-cp311-cp311-macosx_10_15_x86_64.whl", hash = "sha256:d0ebea336b535b37eee9eee31761813086d33ed06de9ab6fc6aaa0bace7b250c"}, + {file = "pyarrow-16.1.0-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:2e73cfc4a99e796727919c5541c65bb88b973377501e39b9842ea71401ca6c1c"}, + {file = "pyarrow-16.1.0-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:bf9251264247ecfe93e5f5a0cd43b8ae834f1e61d1abca22da55b20c788417f6"}, + {file = "pyarrow-16.1.0-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:ddf5aace92d520d3d2a20031d8b0ec27b4395cab9f74e07cc95edf42a5cc0147"}, + {file = "pyarrow-16.1.0-cp311-cp311-manylinux_2_28_aarch64.whl", hash = "sha256:25233642583bf658f629eb230b9bb79d9af4d9f9229890b3c878699c82f7d11e"}, + {file = "pyarrow-16.1.0-cp311-cp311-manylinux_2_28_x86_64.whl", hash = "sha256:a33a64576fddfbec0a44112eaf844c20853647ca833e9a647bfae0582b2ff94b"}, + {file = "pyarrow-16.1.0-cp311-cp311-win_amd64.whl", hash = "sha256:185d121b50836379fe012753cf15c4ba9638bda9645183ab36246923875f8d1b"}, + {file = "pyarrow-16.1.0-cp312-cp312-macosx_10_15_x86_64.whl", hash = "sha256:2e51ca1d6ed7f2e9d5c3c83decf27b0d17bb207a7dea986e8dc3e24f80ff7d6f"}, + {file = "pyarrow-16.1.0-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:06ebccb6f8cb7357de85f60d5da50e83507954af617d7b05f48af1621d331c9a"}, + {file = "pyarrow-16.1.0-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:b04707f1979815f5e49824ce52d1dceb46e2f12909a48a6a753fe7cafbc44a0c"}, + {file = "pyarrow-16.1.0-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:0d32000693deff8dc5df444b032b5985a48592c0697cb6e3071a5d59888714e2"}, + {file = "pyarrow-16.1.0-cp312-cp312-manylinux_2_28_aarch64.whl", hash = "sha256:8785bb10d5d6fd5e15d718ee1d1f914fe768bf8b4d1e5e9bf253de8a26cb1628"}, + {file = "pyarrow-16.1.0-cp312-cp312-manylinux_2_28_x86_64.whl", hash = "sha256:e1369af39587b794873b8a307cc6623a3b1194e69399af0efd05bb202195a5a7"}, + {file = "pyarrow-16.1.0-cp312-cp312-win_amd64.whl", hash = "sha256:febde33305f1498f6df85e8020bca496d0e9ebf2093bab9e0f65e2b4ae2b3444"}, + {file = "pyarrow-16.1.0-cp38-cp38-macosx_10_15_x86_64.whl", hash = "sha256:b5f5705ab977947a43ac83b52ade3b881eb6e95fcc02d76f501d549a210ba77f"}, + {file = "pyarrow-16.1.0-cp38-cp38-macosx_11_0_arm64.whl", hash = "sha256:0d27bf89dfc2576f6206e9cd6cf7a107c9c06dc13d53bbc25b0bd4556f19cf5f"}, + {file = "pyarrow-16.1.0-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:0d07de3ee730647a600037bc1d7b7994067ed64d0eba797ac74b2bc77384f4c2"}, + {file = "pyarrow-16.1.0-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:fbef391b63f708e103df99fbaa3acf9f671d77a183a07546ba2f2c297b361e83"}, + {file = "pyarrow-16.1.0-cp38-cp38-manylinux_2_28_aarch64.whl", hash = "sha256:19741c4dbbbc986d38856ee7ddfdd6a00fc3b0fc2d928795b95410d38bb97d15"}, + {file = "pyarrow-16.1.0-cp38-cp38-manylinux_2_28_x86_64.whl", hash = "sha256:f2c5fb249caa17b94e2b9278b36a05ce03d3180e6da0c4c3b3ce5b2788f30eed"}, + {file = "pyarrow-16.1.0-cp38-cp38-win_amd64.whl", hash = "sha256:e6b6d3cd35fbb93b70ade1336022cc1147b95ec6af7d36906ca7fe432eb09710"}, + {file = "pyarrow-16.1.0-cp39-cp39-macosx_10_15_x86_64.whl", hash = "sha256:18da9b76a36a954665ccca8aa6bd9f46c1145f79c0bb8f4f244f5f8e799bca55"}, + {file = "pyarrow-16.1.0-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:99f7549779b6e434467d2aa43ab2b7224dd9e41bdde486020bae198978c9e05e"}, + {file = "pyarrow-16.1.0-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:f07fdffe4fd5b15f5ec15c8b64584868d063bc22b86b46c9695624ca3505b7b4"}, + {file = "pyarrow-16.1.0-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:ddfe389a08ea374972bd4065d5f25d14e36b43ebc22fc75f7b951f24378bf0b5"}, + {file = "pyarrow-16.1.0-cp39-cp39-manylinux_2_28_aarch64.whl", hash = "sha256:3b20bd67c94b3a2ea0a749d2a5712fc845a69cb5d52e78e6449bbd295611f3aa"}, + {file = "pyarrow-16.1.0-cp39-cp39-manylinux_2_28_x86_64.whl", hash = "sha256:ba8ac20693c0bb0bf4b238751d4409e62852004a8cf031c73b0e0962b03e45e3"}, + {file = "pyarrow-16.1.0-cp39-cp39-win_amd64.whl", hash = "sha256:31a1851751433d89a986616015841977e0a188662fcffd1a5677453f1df2de0a"}, + {file = "pyarrow-16.1.0.tar.gz", hash = "sha256:15fbb22ea96d11f0b5768504a3f961edab25eaf4197c341720c4a387f6c60315"}, ] [package.dependencies] -numpy = ">=1.16.6,<2" +numpy = ">=1.16.6" [[package]] name = "pyarrow-hotfix" @@ -3369,13 +3423,13 @@ files = [ [[package]] name = "pycparser" -version = "2.21" +version = "2.22" description = "C parser in Python" optional = false -python-versions = ">=2.7, !=3.0.*, !=3.1.*, !=3.2.*, !=3.3.*" +python-versions = ">=3.8" files = [ - {file = "pycparser-2.21-py2.py3-none-any.whl", hash = "sha256:8ee45429555515e1f6b185e78100aea234072576aa43ab53aefcae078162fca9"}, - {file = "pycparser-2.21.tar.gz", hash = "sha256:e644fdec12f7872f86c58ff790da456218b10f863970249516d60a5eaca77206"}, + {file = "pycparser-2.22-py3-none-any.whl", hash = "sha256:c3702b6d3dd8c7abc1afa565d7e63d53a1d0bd86cdc24edd75470f4de499cfcc"}, + {file = "pycparser-2.22.tar.gz", hash = "sha256:491c8be9c040f5390f5bf44a5b07752bd07f56edf992381b05c701439eec10f6"}, ] [[package]] @@ -3430,32 +3484,31 @@ tests = ["WebTest", "beautifulsoup4", "flake8", "pyopenssl", "pytest (>=3.6)", " [[package]] name = "pygments" -version = "2.17.2" +version = "2.18.0" description = "Pygments is a syntax highlighting package written in Python." optional = false -python-versions = ">=3.7" +python-versions = ">=3.8" files = [ - {file = "pygments-2.17.2-py3-none-any.whl", hash = "sha256:b27c2826c47d0f3219f29554824c30c5e8945175d888647acd804ddd04af846c"}, - {file = "pygments-2.17.2.tar.gz", hash = "sha256:da46cec9fd2de5be3a8a784f434e4c4ab670b4ff54d605c4c2717e9d49c4c367"}, + {file = "pygments-2.18.0-py3-none-any.whl", hash = "sha256:b8e6aca0523f3ab76fee51799c488e38782ac06eafcf95e7ba832985c8e7b13a"}, + {file = "pygments-2.18.0.tar.gz", hash = "sha256:786ff802f32e91311bff3889f6e9a86e81505fe99f2735bb6d60ae0c5004f199"}, ] [package.extras] -plugins = ["importlib-metadata"] windows-terminal = ["colorama (>=0.4.6)"] [[package]] name = "pymdown-extensions" -version = "10.7" +version = "10.8.1" description = "Extension pack for Python Markdown." optional = false python-versions = ">=3.8" files = [ - {file = "pymdown_extensions-10.7-py3-none-any.whl", hash = "sha256:6ca215bc57bc12bf32b414887a68b810637d039124ed9b2e5bd3325cbb2c050c"}, - {file = "pymdown_extensions-10.7.tar.gz", hash = "sha256:c0d64d5cf62566f59e6b2b690a4095c931107c250a8c8e1351c1de5f6b036deb"}, + {file = "pymdown_extensions-10.8.1-py3-none-any.whl", hash = "sha256:f938326115884f48c6059c67377c46cf631c733ef3629b6eed1349989d1b30cb"}, + {file = "pymdown_extensions-10.8.1.tar.gz", hash = "sha256:3ab1db5c9e21728dabf75192d71471f8e50f216627e9a1fa9535ecb0231b9940"}, ] [package.dependencies] -markdown = ">=3.5" +markdown = ">=3.6" pyyaml = "*" [package.extras] @@ -3463,13 +3516,13 @@ extra = ["pygments (>=2.12)"] [[package]] name = "pyparsing" -version = "3.1.1" +version = "3.1.2" description = "pyparsing module - Classes and methods to define and execute parsing grammars" optional = false python-versions = ">=3.6.8" files = [ - {file = "pyparsing-3.1.1-py3-none-any.whl", hash = "sha256:32c7c0b711493c72ff18a981d24f28aaf9c1fb7ed5e9667c9e84e3db623bdbfb"}, - {file = "pyparsing-3.1.1.tar.gz", hash = "sha256:ede28a1a32462f5a9705e07aea48001a08f7cf81a021585011deba701581a0db"}, + {file = "pyparsing-3.1.2-py3-none-any.whl", hash = "sha256:f9db75911801ed778fe61bb643079ff86601aca99fcae6345aa67292038fb742"}, + {file = "pyparsing-3.1.2.tar.gz", hash = "sha256:a1bac0ce561155ecc3ed78ca94d3c9378656ad4c94c1270de543f621420f94ad"}, ] [package.extras] @@ -3527,13 +3580,13 @@ files = [ [[package]] name = "pytest" -version = "8.0.0" +version = "8.2.1" description = "pytest: simple powerful testing with Python" optional = false python-versions = ">=3.8" files = [ - {file = "pytest-8.0.0-py3-none-any.whl", hash = "sha256:50fb9cbe836c3f20f0dfa99c565201fb75dc54c8d76373cd1bde06b06657bdb6"}, - {file = "pytest-8.0.0.tar.gz", hash = "sha256:249b1b0864530ba251b7438274c4d251c58d868edaaec8762893ad4a0d71c36c"}, + {file = "pytest-8.2.1-py3-none-any.whl", hash = "sha256:faccc5d332b8c3719f40283d0d44aa5cf101cec36f88cde9ed8f2bc0538612b1"}, + {file = "pytest-8.2.1.tar.gz", hash = "sha256:5046e5b46d8e4cac199c373041f26be56fdb81eb4e67dc11d4e10811fc3408fd"}, ] [package.dependencies] @@ -3541,21 +3594,21 @@ colorama = {version = "*", markers = "sys_platform == \"win32\""} exceptiongroup = {version = ">=1.0.0rc8", markers = "python_version < \"3.11\""} iniconfig = "*" packaging = "*" -pluggy = ">=1.3.0,<2.0" -tomli = {version = ">=1.0.0", markers = "python_version < \"3.11\""} +pluggy = ">=1.5,<2.0" +tomli = {version = ">=1", markers = "python_version < \"3.11\""} [package.extras] -testing = ["argcomplete", "attrs (>=19.2.0)", "hypothesis (>=3.56)", "mock", "nose", "pygments (>=2.7.2)", "requests", "setuptools", "xmlschema"] +dev = ["argcomplete", "attrs (>=19.2)", "hypothesis (>=3.56)", "mock", "pygments (>=2.7.2)", "requests", "setuptools", "xmlschema"] [[package]] name = "pytest-cov" -version = "4.1.0" +version = "5.0.0" description = "Pytest plugin for measuring coverage." optional = false -python-versions = ">=3.7" +python-versions = ">=3.8" files = [ - {file = "pytest-cov-4.1.0.tar.gz", hash = "sha256:3904b13dfbfec47f003b8e77fd5b589cd11904a21ddf1ab38a64f204d6a10ef6"}, - {file = "pytest_cov-4.1.0-py3-none-any.whl", hash = "sha256:6ba70b9e97e69fcc3fb45bfeab2d0a138fb65c4d0d6a41ef33983ad114be8c3a"}, + {file = "pytest-cov-5.0.0.tar.gz", hash = "sha256:5837b58e9f6ebd335b0f8060eecce69b662415b16dc503883a02f45dfeb14857"}, + {file = "pytest_cov-5.0.0-py3-none-any.whl", hash = "sha256:4f0764a1219df53214206bf1feea4633c3b558a2925c8b59f144f682861ce652"}, ] [package.dependencies] @@ -3563,17 +3616,17 @@ coverage = {version = ">=5.2.1", extras = ["toml"]} pytest = ">=4.6" [package.extras] -testing = ["fields", "hunter", "process-tests", "pytest-xdist", "six", "virtualenv"] +testing = ["fields", "hunter", "process-tests", "pytest-xdist", "virtualenv"] [[package]] name = "python-dateutil" -version = "2.8.2" +version = "2.9.0.post0" description = "Extensions to the standard Python datetime module" optional = false python-versions = "!=3.0.*,!=3.1.*,!=3.2.*,>=2.7" files = [ - {file = "python-dateutil-2.8.2.tar.gz", hash = "sha256:0123cacc1627ae19ddf3c27a5de5bd67ee4586fbdd6440d9748f8abb483d3e86"}, - {file = "python_dateutil-2.8.2-py2.py3-none-any.whl", hash = "sha256:961d03dc3453ebbc59dbdea9e4e11c5651520a876d0f4db161e8674aae935da9"}, + {file = "python-dateutil-2.9.0.post0.tar.gz", hash = "sha256:37dd54208da7e1cd875388217d5e00ebd4179249f90fb72437e91a35459a0ad3"}, + {file = "python_dateutil-2.9.0.post0-py2.py3-none-any.whl", hash = "sha256:a8b2bc7bffae282281c8140a97d3aa9c14da0b136dfe83f850eea9a5f7470427"}, ] [package.dependencies] @@ -3592,13 +3645,13 @@ files = [ [[package]] name = "pyviz-comms" -version = "3.0.1" +version = "3.0.2" description = "A JupyterLab extension for rendering HoloViz content." optional = false python-versions = ">=3.8" files = [ - {file = "pyviz_comms-3.0.1-py3-none-any.whl", hash = "sha256:0130e952b942906a0eb5fcbcc750262a8e4f565a9b06b3c0d8d631f33b61b78e"}, - {file = "pyviz_comms-3.0.1.tar.gz", hash = "sha256:427c33a5a81780db9b9e757f0675f65ea2292d9a642a2d291cfb5cae6cd46991"}, + {file = "pyviz_comms-3.0.2-py3-none-any.whl", hash = "sha256:31541b976a21b7738557c3ea23bd8e44e94e736b9ed269570dcc28db4449d7e3"}, + {file = "pyviz_comms-3.0.2.tar.gz", hash = "sha256:3167df932656416c4bd711205dad47e986a3ebae1f316258ddc26f9e01513ef7"}, ] [package.dependencies] @@ -3707,104 +3760,99 @@ pyyaml = "*" [[package]] name = "pyzmq" -version = "25.1.2" +version = "26.0.3" description = "Python bindings for 0MQ" optional = false -python-versions = ">=3.6" +python-versions = ">=3.7" files = [ - {file = "pyzmq-25.1.2-cp310-cp310-macosx_10_15_universal2.whl", hash = "sha256:e624c789359f1a16f83f35e2c705d07663ff2b4d4479bad35621178d8f0f6ea4"}, - {file = "pyzmq-25.1.2-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:49151b0efece79f6a79d41a461d78535356136ee70084a1c22532fc6383f4ad0"}, - {file = "pyzmq-25.1.2-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:d9a5f194cf730f2b24d6af1f833c14c10f41023da46a7f736f48b6d35061e76e"}, - {file = "pyzmq-25.1.2-cp310-cp310-manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:faf79a302f834d9e8304fafdc11d0d042266667ac45209afa57e5efc998e3872"}, - {file = "pyzmq-25.1.2-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:7f51a7b4ead28d3fca8dda53216314a553b0f7a91ee8fc46a72b402a78c3e43d"}, - {file = "pyzmq-25.1.2-cp310-cp310-manylinux_2_28_x86_64.whl", hash = "sha256:0ddd6d71d4ef17ba5a87becf7ddf01b371eaba553c603477679ae817a8d84d75"}, - {file = "pyzmq-25.1.2-cp310-cp310-musllinux_1_1_aarch64.whl", hash = "sha256:246747b88917e4867e2367b005fc8eefbb4a54b7db363d6c92f89d69abfff4b6"}, - {file = "pyzmq-25.1.2-cp310-cp310-musllinux_1_1_i686.whl", hash = "sha256:00c48ae2fd81e2a50c3485de1b9d5c7c57cd85dc8ec55683eac16846e57ac979"}, - {file = "pyzmq-25.1.2-cp310-cp310-musllinux_1_1_x86_64.whl", hash = "sha256:5a68d491fc20762b630e5db2191dd07ff89834086740f70e978bb2ef2668be08"}, - {file = "pyzmq-25.1.2-cp310-cp310-win32.whl", hash = "sha256:09dfe949e83087da88c4a76767df04b22304a682d6154de2c572625c62ad6886"}, - {file = "pyzmq-25.1.2-cp310-cp310-win_amd64.whl", hash = "sha256:fa99973d2ed20417744fca0073390ad65ce225b546febb0580358e36aa90dba6"}, - {file = "pyzmq-25.1.2-cp311-cp311-macosx_10_15_universal2.whl", hash = "sha256:82544e0e2d0c1811482d37eef297020a040c32e0687c1f6fc23a75b75db8062c"}, - {file = "pyzmq-25.1.2-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:01171fc48542348cd1a360a4b6c3e7d8f46cdcf53a8d40f84db6707a6768acc1"}, - {file = "pyzmq-25.1.2-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:bc69c96735ab501419c432110016329bf0dea8898ce16fab97c6d9106dc0b348"}, - {file = "pyzmq-25.1.2-cp311-cp311-manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:3e124e6b1dd3dfbeb695435dff0e383256655bb18082e094a8dd1f6293114642"}, - {file = "pyzmq-25.1.2-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:7598d2ba821caa37a0f9d54c25164a4fa351ce019d64d0b44b45540950458840"}, - {file = "pyzmq-25.1.2-cp311-cp311-manylinux_2_28_x86_64.whl", hash = "sha256:d1299d7e964c13607efd148ca1f07dcbf27c3ab9e125d1d0ae1d580a1682399d"}, - {file = "pyzmq-25.1.2-cp311-cp311-musllinux_1_1_aarch64.whl", hash = "sha256:4e6f689880d5ad87918430957297c975203a082d9a036cc426648fcbedae769b"}, - {file = "pyzmq-25.1.2-cp311-cp311-musllinux_1_1_i686.whl", hash = "sha256:cc69949484171cc961e6ecd4a8911b9ce7a0d1f738fcae717177c231bf77437b"}, - {file = "pyzmq-25.1.2-cp311-cp311-musllinux_1_1_x86_64.whl", hash = "sha256:9880078f683466b7f567b8624bfc16cad65077be046b6e8abb53bed4eeb82dd3"}, - {file = "pyzmq-25.1.2-cp311-cp311-win32.whl", hash = "sha256:4e5837af3e5aaa99a091302df5ee001149baff06ad22b722d34e30df5f0d9097"}, - {file = "pyzmq-25.1.2-cp311-cp311-win_amd64.whl", hash = "sha256:25c2dbb97d38b5ac9fd15586e048ec5eb1e38f3d47fe7d92167b0c77bb3584e9"}, - {file = "pyzmq-25.1.2-cp312-cp312-macosx_10_15_universal2.whl", hash = "sha256:11e70516688190e9c2db14fcf93c04192b02d457b582a1f6190b154691b4c93a"}, - {file = "pyzmq-25.1.2-cp312-cp312-macosx_10_9_x86_64.whl", hash = "sha256:313c3794d650d1fccaaab2df942af9f2c01d6217c846177cfcbc693c7410839e"}, - {file = "pyzmq-25.1.2-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:1b3cbba2f47062b85fe0ef9de5b987612140a9ba3a9c6d2543c6dec9f7c2ab27"}, - {file = "pyzmq-25.1.2-cp312-cp312-manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:fc31baa0c32a2ca660784d5af3b9487e13b61b3032cb01a115fce6588e1bed30"}, - {file = "pyzmq-25.1.2-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:02c9087b109070c5ab0b383079fa1b5f797f8d43e9a66c07a4b8b8bdecfd88ee"}, - {file = "pyzmq-25.1.2-cp312-cp312-manylinux_2_28_x86_64.whl", hash = "sha256:f8429b17cbb746c3e043cb986328da023657e79d5ed258b711c06a70c2ea7537"}, - {file = "pyzmq-25.1.2-cp312-cp312-musllinux_1_1_aarch64.whl", hash = "sha256:5074adeacede5f810b7ef39607ee59d94e948b4fd954495bdb072f8c54558181"}, - {file = "pyzmq-25.1.2-cp312-cp312-musllinux_1_1_i686.whl", hash = "sha256:7ae8f354b895cbd85212da245f1a5ad8159e7840e37d78b476bb4f4c3f32a9fe"}, - {file = "pyzmq-25.1.2-cp312-cp312-musllinux_1_1_x86_64.whl", hash = "sha256:b264bf2cc96b5bc43ce0e852be995e400376bd87ceb363822e2cb1964fcdc737"}, - {file = "pyzmq-25.1.2-cp312-cp312-win32.whl", hash = "sha256:02bbc1a87b76e04fd780b45e7f695471ae6de747769e540da909173d50ff8e2d"}, - {file = "pyzmq-25.1.2-cp312-cp312-win_amd64.whl", hash = "sha256:ced111c2e81506abd1dc142e6cd7b68dd53747b3b7ae5edbea4578c5eeff96b7"}, - {file = "pyzmq-25.1.2-cp36-cp36m-macosx_10_9_x86_64.whl", hash = "sha256:7b6d09a8962a91151f0976008eb7b29b433a560fde056ec7a3db9ec8f1075438"}, - {file = "pyzmq-25.1.2-cp36-cp36m-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:967668420f36878a3c9ecb5ab33c9d0ff8d054f9c0233d995a6d25b0e95e1b6b"}, - {file = "pyzmq-25.1.2-cp36-cp36m-manylinux_2_5_i686.manylinux1_i686.whl", hash = "sha256:5edac3f57c7ddaacdb4d40f6ef2f9e299471fc38d112f4bc6d60ab9365445fb0"}, - {file = "pyzmq-25.1.2-cp36-cp36m-manylinux_2_5_x86_64.manylinux1_x86_64.whl", hash = "sha256:0dabfb10ef897f3b7e101cacba1437bd3a5032ee667b7ead32bbcdd1a8422fe7"}, - {file = "pyzmq-25.1.2-cp36-cp36m-musllinux_1_1_aarch64.whl", hash = "sha256:2c6441e0398c2baacfe5ba30c937d274cfc2dc5b55e82e3749e333aabffde561"}, - {file = "pyzmq-25.1.2-cp36-cp36m-musllinux_1_1_i686.whl", hash = "sha256:16b726c1f6c2e7625706549f9dbe9b06004dfbec30dbed4bf50cbdfc73e5b32a"}, - {file = "pyzmq-25.1.2-cp36-cp36m-musllinux_1_1_x86_64.whl", hash = "sha256:a86c2dd76ef71a773e70551a07318b8e52379f58dafa7ae1e0a4be78efd1ff16"}, - {file = "pyzmq-25.1.2-cp36-cp36m-win32.whl", hash = "sha256:359f7f74b5d3c65dae137f33eb2bcfa7ad9ebefd1cab85c935f063f1dbb245cc"}, - {file = "pyzmq-25.1.2-cp36-cp36m-win_amd64.whl", hash = "sha256:55875492f820d0eb3417b51d96fea549cde77893ae3790fd25491c5754ea2f68"}, - {file = "pyzmq-25.1.2-cp37-cp37m-macosx_10_9_x86_64.whl", hash = "sha256:b8c8a419dfb02e91b453615c69568442e897aaf77561ee0064d789705ff37a92"}, - {file = "pyzmq-25.1.2-cp37-cp37m-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:8807c87fa893527ae8a524c15fc505d9950d5e856f03dae5921b5e9aa3b8783b"}, - {file = "pyzmq-25.1.2-cp37-cp37m-manylinux_2_5_i686.manylinux1_i686.whl", hash = "sha256:5e319ed7d6b8f5fad9b76daa0a68497bc6f129858ad956331a5835785761e003"}, - {file = "pyzmq-25.1.2-cp37-cp37m-manylinux_2_5_x86_64.manylinux1_x86_64.whl", hash = "sha256:3c53687dde4d9d473c587ae80cc328e5b102b517447456184b485587ebd18b62"}, - {file = "pyzmq-25.1.2-cp37-cp37m-musllinux_1_1_aarch64.whl", hash = "sha256:9add2e5b33d2cd765ad96d5eb734a5e795a0755f7fc49aa04f76d7ddda73fd70"}, - {file = "pyzmq-25.1.2-cp37-cp37m-musllinux_1_1_i686.whl", hash = "sha256:e690145a8c0c273c28d3b89d6fb32c45e0d9605b2293c10e650265bf5c11cfec"}, - {file = "pyzmq-25.1.2-cp37-cp37m-musllinux_1_1_x86_64.whl", hash = "sha256:00a06faa7165634f0cac1abb27e54d7a0b3b44eb9994530b8ec73cf52e15353b"}, - {file = "pyzmq-25.1.2-cp37-cp37m-win32.whl", hash = "sha256:0f97bc2f1f13cb16905a5f3e1fbdf100e712d841482b2237484360f8bc4cb3d7"}, - {file = "pyzmq-25.1.2-cp37-cp37m-win_amd64.whl", hash = "sha256:6cc0020b74b2e410287e5942e1e10886ff81ac77789eb20bec13f7ae681f0fdd"}, - {file = "pyzmq-25.1.2-cp38-cp38-macosx_10_15_universal2.whl", hash = "sha256:bef02cfcbded83473bdd86dd8d3729cd82b2e569b75844fb4ea08fee3c26ae41"}, - {file = "pyzmq-25.1.2-cp38-cp38-macosx_10_9_x86_64.whl", hash = "sha256:e10a4b5a4b1192d74853cc71a5e9fd022594573926c2a3a4802020360aa719d8"}, - {file = "pyzmq-25.1.2-cp38-cp38-manylinux_2_12_i686.manylinux2010_i686.whl", hash = "sha256:8c5f80e578427d4695adac6fdf4370c14a2feafdc8cb35549c219b90652536ae"}, - {file = "pyzmq-25.1.2-cp38-cp38-manylinux_2_12_x86_64.manylinux2010_x86_64.whl", hash = "sha256:5dde6751e857910c1339890f3524de74007958557593b9e7e8c5f01cd919f8a7"}, - {file = "pyzmq-25.1.2-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:ea1608dd169da230a0ad602d5b1ebd39807ac96cae1845c3ceed39af08a5c6df"}, - {file = "pyzmq-25.1.2-cp38-cp38-musllinux_1_1_aarch64.whl", hash = "sha256:0f513130c4c361201da9bc69df25a086487250e16b5571ead521b31ff6b02220"}, - {file = "pyzmq-25.1.2-cp38-cp38-musllinux_1_1_i686.whl", hash = "sha256:019744b99da30330798bb37df33549d59d380c78e516e3bab9c9b84f87a9592f"}, - {file = "pyzmq-25.1.2-cp38-cp38-musllinux_1_1_x86_64.whl", hash = "sha256:2e2713ef44be5d52dd8b8e2023d706bf66cb22072e97fc71b168e01d25192755"}, - {file = "pyzmq-25.1.2-cp38-cp38-win32.whl", hash = "sha256:07cd61a20a535524906595e09344505a9bd46f1da7a07e504b315d41cd42eb07"}, - {file = "pyzmq-25.1.2-cp38-cp38-win_amd64.whl", hash = "sha256:eb7e49a17fb8c77d3119d41a4523e432eb0c6932187c37deb6fbb00cc3028088"}, - {file = "pyzmq-25.1.2-cp39-cp39-macosx_10_15_universal2.whl", hash = "sha256:94504ff66f278ab4b7e03e4cba7e7e400cb73bfa9d3d71f58d8972a8dc67e7a6"}, - {file = "pyzmq-25.1.2-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:6dd0d50bbf9dca1d0bdea219ae6b40f713a3fb477c06ca3714f208fd69e16fd8"}, - {file = "pyzmq-25.1.2-cp39-cp39-manylinux_2_12_i686.manylinux2010_i686.whl", hash = "sha256:004ff469d21e86f0ef0369717351073e0e577428e514c47c8480770d5e24a565"}, - {file = "pyzmq-25.1.2-cp39-cp39-manylinux_2_12_x86_64.manylinux2010_x86_64.whl", hash = "sha256:c0b5ca88a8928147b7b1e2dfa09f3b6c256bc1135a1338536cbc9ea13d3b7add"}, - {file = "pyzmq-25.1.2-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:2c9a79f1d2495b167119d02be7448bfba57fad2a4207c4f68abc0bab4b92925b"}, - {file = "pyzmq-25.1.2-cp39-cp39-musllinux_1_1_aarch64.whl", hash = "sha256:518efd91c3d8ac9f9b4f7dd0e2b7b8bf1a4fe82a308009016b07eaa48681af82"}, - {file = "pyzmq-25.1.2-cp39-cp39-musllinux_1_1_i686.whl", hash = "sha256:1ec23bd7b3a893ae676d0e54ad47d18064e6c5ae1fadc2f195143fb27373f7f6"}, - {file = "pyzmq-25.1.2-cp39-cp39-musllinux_1_1_x86_64.whl", hash = "sha256:db36c27baed588a5a8346b971477b718fdc66cf5b80cbfbd914b4d6d355e44e2"}, - {file = "pyzmq-25.1.2-cp39-cp39-win32.whl", hash = "sha256:39b1067f13aba39d794a24761e385e2eddc26295826530a8c7b6c6c341584289"}, - {file = "pyzmq-25.1.2-cp39-cp39-win_amd64.whl", hash = "sha256:8e9f3fabc445d0ce320ea2c59a75fe3ea591fdbdeebec5db6de530dd4b09412e"}, - {file = "pyzmq-25.1.2-pp310-pypy310_pp73-macosx_10_9_x86_64.whl", hash = "sha256:a8c1d566344aee826b74e472e16edae0a02e2a044f14f7c24e123002dcff1c05"}, - {file = "pyzmq-25.1.2-pp310-pypy310_pp73-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:759cfd391a0996345ba94b6a5110fca9c557ad4166d86a6e81ea526c376a01e8"}, - {file = "pyzmq-25.1.2-pp310-pypy310_pp73-manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:7c61e346ac34b74028ede1c6b4bcecf649d69b707b3ff9dc0fab453821b04d1e"}, - {file = "pyzmq-25.1.2-pp310-pypy310_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:4cb8fc1f8d69b411b8ec0b5f1ffbcaf14c1db95b6bccea21d83610987435f1a4"}, - {file = "pyzmq-25.1.2-pp310-pypy310_pp73-win_amd64.whl", hash = "sha256:3c00c9b7d1ca8165c610437ca0c92e7b5607b2f9076f4eb4b095c85d6e680a1d"}, - {file = "pyzmq-25.1.2-pp37-pypy37_pp73-macosx_10_9_x86_64.whl", hash = "sha256:df0c7a16ebb94452d2909b9a7b3337940e9a87a824c4fc1c7c36bb4404cb0cde"}, - {file = "pyzmq-25.1.2-pp37-pypy37_pp73-manylinux_2_12_i686.manylinux2010_i686.whl", hash = "sha256:45999e7f7ed5c390f2e87ece7f6c56bf979fb213550229e711e45ecc7d42ccb8"}, - {file = "pyzmq-25.1.2-pp37-pypy37_pp73-manylinux_2_12_x86_64.manylinux2010_x86_64.whl", hash = "sha256:ac170e9e048b40c605358667aca3d94e98f604a18c44bdb4c102e67070f3ac9b"}, - {file = "pyzmq-25.1.2-pp37-pypy37_pp73-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:d1b604734bec94f05f81b360a272fc824334267426ae9905ff32dc2be433ab96"}, - {file = "pyzmq-25.1.2-pp37-pypy37_pp73-win_amd64.whl", hash = "sha256:a793ac733e3d895d96f865f1806f160696422554e46d30105807fdc9841b9f7d"}, - {file = "pyzmq-25.1.2-pp38-pypy38_pp73-macosx_10_9_x86_64.whl", hash = "sha256:0806175f2ae5ad4b835ecd87f5f85583316b69f17e97786f7443baaf54b9bb98"}, - {file = "pyzmq-25.1.2-pp38-pypy38_pp73-manylinux_2_12_i686.manylinux2010_i686.whl", hash = "sha256:ef12e259e7bc317c7597d4f6ef59b97b913e162d83b421dd0db3d6410f17a244"}, - {file = "pyzmq-25.1.2-pp38-pypy38_pp73-manylinux_2_12_x86_64.manylinux2010_x86_64.whl", hash = "sha256:ea253b368eb41116011add00f8d5726762320b1bda892f744c91997b65754d73"}, - {file = "pyzmq-25.1.2-pp38-pypy38_pp73-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:1b9b1f2ad6498445a941d9a4fee096d387fee436e45cc660e72e768d3d8ee611"}, - {file = "pyzmq-25.1.2-pp38-pypy38_pp73-win_amd64.whl", hash = "sha256:8b14c75979ce932c53b79976a395cb2a8cd3aaf14aef75e8c2cb55a330b9b49d"}, - {file = "pyzmq-25.1.2-pp39-pypy39_pp73-macosx_10_9_x86_64.whl", hash = "sha256:889370d5174a741a62566c003ee8ddba4b04c3f09a97b8000092b7ca83ec9c49"}, - {file = "pyzmq-25.1.2-pp39-pypy39_pp73-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:9a18fff090441a40ffda8a7f4f18f03dc56ae73f148f1832e109f9bffa85df15"}, - {file = "pyzmq-25.1.2-pp39-pypy39_pp73-manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:99a6b36f95c98839ad98f8c553d8507644c880cf1e0a57fe5e3a3f3969040882"}, - {file = "pyzmq-25.1.2-pp39-pypy39_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:4345c9a27f4310afbb9c01750e9461ff33d6fb74cd2456b107525bbeebcb5be3"}, - {file = "pyzmq-25.1.2-pp39-pypy39_pp73-manylinux_2_28_x86_64.whl", hash = "sha256:3516e0b6224cf6e43e341d56da15fd33bdc37fa0c06af4f029f7d7dfceceabbc"}, - {file = "pyzmq-25.1.2-pp39-pypy39_pp73-win_amd64.whl", hash = "sha256:146b9b1f29ead41255387fb07be56dc29639262c0f7344f570eecdcd8d683314"}, - {file = "pyzmq-25.1.2.tar.gz", hash = "sha256:93f1aa311e8bb912e34f004cf186407a4e90eec4f0ecc0efd26056bf7eda0226"}, + {file = "pyzmq-26.0.3-cp310-cp310-macosx_10_15_universal2.whl", hash = "sha256:44dd6fc3034f1eaa72ece33588867df9e006a7303725a12d64c3dff92330f625"}, + {file = "pyzmq-26.0.3-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:acb704195a71ac5ea5ecf2811c9ee19ecdc62b91878528302dd0be1b9451cc90"}, + {file = "pyzmq-26.0.3-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:5dbb9c997932473a27afa93954bb77a9f9b786b4ccf718d903f35da3232317de"}, + {file = "pyzmq-26.0.3-cp310-cp310-manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:6bcb34f869d431799c3ee7d516554797f7760cb2198ecaa89c3f176f72d062be"}, + {file = "pyzmq-26.0.3-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:38ece17ec5f20d7d9b442e5174ae9f020365d01ba7c112205a4d59cf19dc38ee"}, + {file = "pyzmq-26.0.3-cp310-cp310-manylinux_2_28_x86_64.whl", hash = "sha256:ba6e5e6588e49139a0979d03a7deb9c734bde647b9a8808f26acf9c547cab1bf"}, + {file = "pyzmq-26.0.3-cp310-cp310-musllinux_1_1_aarch64.whl", hash = "sha256:3bf8b000a4e2967e6dfdd8656cd0757d18c7e5ce3d16339e550bd462f4857e59"}, + {file = "pyzmq-26.0.3-cp310-cp310-musllinux_1_1_i686.whl", hash = "sha256:2136f64fbb86451dbbf70223635a468272dd20075f988a102bf8a3f194a411dc"}, + {file = "pyzmq-26.0.3-cp310-cp310-musllinux_1_1_x86_64.whl", hash = "sha256:e8918973fbd34e7814f59143c5f600ecd38b8038161239fd1a3d33d5817a38b8"}, + {file = "pyzmq-26.0.3-cp310-cp310-win32.whl", hash = "sha256:0aaf982e68a7ac284377d051c742610220fd06d330dcd4c4dbb4cdd77c22a537"}, + {file = "pyzmq-26.0.3-cp310-cp310-win_amd64.whl", hash = "sha256:f1a9b7d00fdf60b4039f4455afd031fe85ee8305b019334b72dcf73c567edc47"}, + {file = "pyzmq-26.0.3-cp310-cp310-win_arm64.whl", hash = "sha256:80b12f25d805a919d53efc0a5ad7c0c0326f13b4eae981a5d7b7cc343318ebb7"}, + {file = "pyzmq-26.0.3-cp311-cp311-macosx_10_15_universal2.whl", hash = "sha256:a72a84570f84c374b4c287183debc776dc319d3e8ce6b6a0041ce2e400de3f32"}, + {file = "pyzmq-26.0.3-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:7ca684ee649b55fd8f378127ac8462fb6c85f251c2fb027eb3c887e8ee347bcd"}, + {file = "pyzmq-26.0.3-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:e222562dc0f38571c8b1ffdae9d7adb866363134299264a1958d077800b193b7"}, + {file = "pyzmq-26.0.3-cp311-cp311-manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:f17cde1db0754c35a91ac00b22b25c11da6eec5746431d6e5092f0cd31a3fea9"}, + {file = "pyzmq-26.0.3-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:4b7c0c0b3244bb2275abe255d4a30c050d541c6cb18b870975553f1fb6f37527"}, + {file = "pyzmq-26.0.3-cp311-cp311-manylinux_2_28_x86_64.whl", hash = "sha256:ac97a21de3712afe6a6c071abfad40a6224fd14fa6ff0ff8d0c6e6cd4e2f807a"}, + {file = "pyzmq-26.0.3-cp311-cp311-musllinux_1_1_aarch64.whl", hash = "sha256:88b88282e55fa39dd556d7fc04160bcf39dea015f78e0cecec8ff4f06c1fc2b5"}, + {file = "pyzmq-26.0.3-cp311-cp311-musllinux_1_1_i686.whl", hash = "sha256:72b67f966b57dbd18dcc7efbc1c7fc9f5f983e572db1877081f075004614fcdd"}, + {file = "pyzmq-26.0.3-cp311-cp311-musllinux_1_1_x86_64.whl", hash = "sha256:f4b6cecbbf3b7380f3b61de3a7b93cb721125dc125c854c14ddc91225ba52f83"}, + {file = "pyzmq-26.0.3-cp311-cp311-win32.whl", hash = "sha256:eed56b6a39216d31ff8cd2f1d048b5bf1700e4b32a01b14379c3b6dde9ce3aa3"}, + {file = "pyzmq-26.0.3-cp311-cp311-win_amd64.whl", hash = "sha256:3191d312c73e3cfd0f0afdf51df8405aafeb0bad71e7ed8f68b24b63c4f36500"}, + {file = "pyzmq-26.0.3-cp311-cp311-win_arm64.whl", hash = "sha256:b6907da3017ef55139cf0e417c5123a84c7332520e73a6902ff1f79046cd3b94"}, + {file = "pyzmq-26.0.3-cp312-cp312-macosx_10_15_universal2.whl", hash = "sha256:068ca17214038ae986d68f4a7021f97e187ed278ab6dccb79f837d765a54d753"}, + {file = "pyzmq-26.0.3-cp312-cp312-macosx_10_9_x86_64.whl", hash = "sha256:7821d44fe07335bea256b9f1f41474a642ca55fa671dfd9f00af8d68a920c2d4"}, + {file = "pyzmq-26.0.3-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:eeb438a26d87c123bb318e5f2b3d86a36060b01f22fbdffd8cf247d52f7c9a2b"}, + {file = "pyzmq-26.0.3-cp312-cp312-manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:69ea9d6d9baa25a4dc9cef5e2b77b8537827b122214f210dd925132e34ae9b12"}, + {file = "pyzmq-26.0.3-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:7daa3e1369355766dea11f1d8ef829905c3b9da886ea3152788dc25ee6079e02"}, + {file = "pyzmq-26.0.3-cp312-cp312-manylinux_2_28_x86_64.whl", hash = "sha256:6ca7a9a06b52d0e38ccf6bca1aeff7be178917893f3883f37b75589d42c4ac20"}, + {file = "pyzmq-26.0.3-cp312-cp312-musllinux_1_1_aarch64.whl", hash = "sha256:1b7d0e124948daa4d9686d421ef5087c0516bc6179fdcf8828b8444f8e461a77"}, + {file = "pyzmq-26.0.3-cp312-cp312-musllinux_1_1_i686.whl", hash = "sha256:e746524418b70f38550f2190eeee834db8850088c834d4c8406fbb9bc1ae10b2"}, + {file = "pyzmq-26.0.3-cp312-cp312-musllinux_1_1_x86_64.whl", hash = "sha256:6b3146f9ae6af82c47a5282ac8803523d381b3b21caeae0327ed2f7ecb718798"}, + {file = "pyzmq-26.0.3-cp312-cp312-win32.whl", hash = "sha256:2b291d1230845871c00c8462c50565a9cd6026fe1228e77ca934470bb7d70ea0"}, + {file = "pyzmq-26.0.3-cp312-cp312-win_amd64.whl", hash = "sha256:926838a535c2c1ea21c903f909a9a54e675c2126728c21381a94ddf37c3cbddf"}, + {file = "pyzmq-26.0.3-cp312-cp312-win_arm64.whl", hash = "sha256:5bf6c237f8c681dfb91b17f8435b2735951f0d1fad10cc5dfd96db110243370b"}, + {file = "pyzmq-26.0.3-cp37-cp37m-macosx_10_9_x86_64.whl", hash = "sha256:0c0991f5a96a8e620f7691e61178cd8f457b49e17b7d9cfa2067e2a0a89fc1d5"}, + {file = "pyzmq-26.0.3-cp37-cp37m-manylinux_2_12_i686.manylinux2010_i686.whl", hash = "sha256:dbf012d8fcb9f2cf0643b65df3b355fdd74fc0035d70bb5c845e9e30a3a4654b"}, + {file = "pyzmq-26.0.3-cp37-cp37m-manylinux_2_12_x86_64.manylinux2010_x86_64.whl", hash = "sha256:01fbfbeb8249a68d257f601deb50c70c929dc2dfe683b754659569e502fbd3aa"}, + {file = "pyzmq-26.0.3-cp37-cp37m-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:1c8eb19abe87029c18f226d42b8a2c9efdd139d08f8bf6e085dd9075446db450"}, + {file = "pyzmq-26.0.3-cp37-cp37m-musllinux_1_1_aarch64.whl", hash = "sha256:5344b896e79800af86ad643408ca9aa303a017f6ebff8cee5a3163c1e9aec987"}, + {file = "pyzmq-26.0.3-cp37-cp37m-musllinux_1_1_i686.whl", hash = "sha256:204e0f176fd1d067671157d049466869b3ae1fc51e354708b0dc41cf94e23a3a"}, + {file = "pyzmq-26.0.3-cp37-cp37m-musllinux_1_1_x86_64.whl", hash = "sha256:a42db008d58530efa3b881eeee4991146de0b790e095f7ae43ba5cc612decbc5"}, + {file = "pyzmq-26.0.3-cp37-cp37m-win32.whl", hash = "sha256:8d7a498671ca87e32b54cb47c82a92b40130a26c5197d392720a1bce1b3c77cf"}, + {file = "pyzmq-26.0.3-cp37-cp37m-win_amd64.whl", hash = "sha256:3b4032a96410bdc760061b14ed6a33613ffb7f702181ba999df5d16fb96ba16a"}, + {file = "pyzmq-26.0.3-cp38-cp38-macosx_10_15_universal2.whl", hash = "sha256:2cc4e280098c1b192c42a849de8de2c8e0f3a84086a76ec5b07bfee29bda7d18"}, + {file = "pyzmq-26.0.3-cp38-cp38-macosx_10_9_x86_64.whl", hash = "sha256:5bde86a2ed3ce587fa2b207424ce15b9a83a9fa14422dcc1c5356a13aed3df9d"}, + {file = "pyzmq-26.0.3-cp38-cp38-manylinux_2_12_i686.manylinux2010_i686.whl", hash = "sha256:34106f68e20e6ff253c9f596ea50397dbd8699828d55e8fa18bd4323d8d966e6"}, + {file = "pyzmq-26.0.3-cp38-cp38-manylinux_2_12_x86_64.manylinux2010_x86_64.whl", hash = "sha256:ebbbd0e728af5db9b04e56389e2299a57ea8b9dd15c9759153ee2455b32be6ad"}, + {file = "pyzmq-26.0.3-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:f6b1d1c631e5940cac5a0b22c5379c86e8df6a4ec277c7a856b714021ab6cfad"}, + {file = "pyzmq-26.0.3-cp38-cp38-musllinux_1_1_aarch64.whl", hash = "sha256:e891ce81edd463b3b4c3b885c5603c00141151dd9c6936d98a680c8c72fe5c67"}, + {file = "pyzmq-26.0.3-cp38-cp38-musllinux_1_1_i686.whl", hash = "sha256:9b273ecfbc590a1b98f014ae41e5cf723932f3b53ba9367cfb676f838038b32c"}, + {file = "pyzmq-26.0.3-cp38-cp38-musllinux_1_1_x86_64.whl", hash = "sha256:b32bff85fb02a75ea0b68f21e2412255b5731f3f389ed9aecc13a6752f58ac97"}, + {file = "pyzmq-26.0.3-cp38-cp38-win32.whl", hash = "sha256:f6c21c00478a7bea93caaaef9e7629145d4153b15a8653e8bb4609d4bc70dbfc"}, + {file = "pyzmq-26.0.3-cp38-cp38-win_amd64.whl", hash = "sha256:3401613148d93ef0fd9aabdbddb212de3db7a4475367f49f590c837355343972"}, + {file = "pyzmq-26.0.3-cp39-cp39-macosx_10_15_universal2.whl", hash = "sha256:2ed8357f4c6e0daa4f3baf31832df8a33334e0fe5b020a61bc8b345a3db7a606"}, + {file = "pyzmq-26.0.3-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:c1c8f2a2ca45292084c75bb6d3a25545cff0ed931ed228d3a1810ae3758f975f"}, + {file = "pyzmq-26.0.3-cp39-cp39-manylinux_2_12_i686.manylinux2010_i686.whl", hash = "sha256:b63731993cdddcc8e087c64e9cf003f909262b359110070183d7f3025d1c56b5"}, + {file = "pyzmq-26.0.3-cp39-cp39-manylinux_2_12_x86_64.manylinux2010_x86_64.whl", hash = "sha256:b3cd31f859b662ac5d7f4226ec7d8bd60384fa037fc02aee6ff0b53ba29a3ba8"}, + {file = "pyzmq-26.0.3-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:115f8359402fa527cf47708d6f8a0f8234f0e9ca0cab7c18c9c189c194dbf620"}, + {file = "pyzmq-26.0.3-cp39-cp39-musllinux_1_1_aarch64.whl", hash = "sha256:715bdf952b9533ba13dfcf1f431a8f49e63cecc31d91d007bc1deb914f47d0e4"}, + {file = "pyzmq-26.0.3-cp39-cp39-musllinux_1_1_i686.whl", hash = "sha256:e1258c639e00bf5e8a522fec6c3eaa3e30cf1c23a2f21a586be7e04d50c9acab"}, + {file = "pyzmq-26.0.3-cp39-cp39-musllinux_1_1_x86_64.whl", hash = "sha256:15c59e780be8f30a60816a9adab900c12a58d79c1ac742b4a8df044ab2a6d920"}, + {file = "pyzmq-26.0.3-cp39-cp39-win32.whl", hash = "sha256:d0cdde3c78d8ab5b46595054e5def32a755fc028685add5ddc7403e9f6de9879"}, + {file = "pyzmq-26.0.3-cp39-cp39-win_amd64.whl", hash = "sha256:ce828058d482ef860746bf532822842e0ff484e27f540ef5c813d516dd8896d2"}, + {file = "pyzmq-26.0.3-cp39-cp39-win_arm64.whl", hash = "sha256:788f15721c64109cf720791714dc14afd0f449d63f3a5487724f024345067381"}, + {file = "pyzmq-26.0.3-pp310-pypy310_pp73-macosx_10_9_x86_64.whl", hash = "sha256:2c18645ef6294d99b256806e34653e86236eb266278c8ec8112622b61db255de"}, + {file = "pyzmq-26.0.3-pp310-pypy310_pp73-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:7e6bc96ebe49604df3ec2c6389cc3876cabe475e6bfc84ced1bf4e630662cb35"}, + {file = "pyzmq-26.0.3-pp310-pypy310_pp73-manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:971e8990c5cc4ddcff26e149398fc7b0f6a042306e82500f5e8db3b10ce69f84"}, + {file = "pyzmq-26.0.3-pp310-pypy310_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:d8416c23161abd94cc7da80c734ad7c9f5dbebdadfdaa77dad78244457448223"}, + {file = "pyzmq-26.0.3-pp310-pypy310_pp73-win_amd64.whl", hash = "sha256:082a2988364b60bb5de809373098361cf1dbb239623e39e46cb18bc035ed9c0c"}, + {file = "pyzmq-26.0.3-pp37-pypy37_pp73-macosx_10_9_x86_64.whl", hash = "sha256:d57dfbf9737763b3a60d26e6800e02e04284926329aee8fb01049635e957fe81"}, + {file = "pyzmq-26.0.3-pp37-pypy37_pp73-manylinux_2_12_i686.manylinux2010_i686.whl", hash = "sha256:77a85dca4c2430ac04dc2a2185c2deb3858a34fe7f403d0a946fa56970cf60a1"}, + {file = "pyzmq-26.0.3-pp37-pypy37_pp73-manylinux_2_12_x86_64.manylinux2010_x86_64.whl", hash = "sha256:4c82a6d952a1d555bf4be42b6532927d2a5686dd3c3e280e5f63225ab47ac1f5"}, + {file = "pyzmq-26.0.3-pp37-pypy37_pp73-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:4496b1282c70c442809fc1b151977c3d967bfb33e4e17cedbf226d97de18f709"}, + {file = "pyzmq-26.0.3-pp37-pypy37_pp73-win_amd64.whl", hash = "sha256:e4946d6bdb7ba972dfda282f9127e5756d4f299028b1566d1245fa0d438847e6"}, + {file = "pyzmq-26.0.3-pp38-pypy38_pp73-macosx_10_9_x86_64.whl", hash = "sha256:03c0ae165e700364b266876d712acb1ac02693acd920afa67da2ebb91a0b3c09"}, + {file = "pyzmq-26.0.3-pp38-pypy38_pp73-manylinux_2_12_i686.manylinux2010_i686.whl", hash = "sha256:3e3070e680f79887d60feeda051a58d0ac36622e1759f305a41059eff62c6da7"}, + {file = "pyzmq-26.0.3-pp38-pypy38_pp73-manylinux_2_12_x86_64.manylinux2010_x86_64.whl", hash = "sha256:6ca08b840fe95d1c2bd9ab92dac5685f949fc6f9ae820ec16193e5ddf603c3b2"}, + {file = "pyzmq-26.0.3-pp38-pypy38_pp73-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:e76654e9dbfb835b3518f9938e565c7806976c07b37c33526b574cc1a1050480"}, + {file = "pyzmq-26.0.3-pp38-pypy38_pp73-win_amd64.whl", hash = "sha256:871587bdadd1075b112e697173e946a07d722459d20716ceb3d1bd6c64bd08ce"}, + {file = "pyzmq-26.0.3-pp39-pypy39_pp73-macosx_10_9_x86_64.whl", hash = "sha256:d0a2d1bd63a4ad79483049b26514e70fa618ce6115220da9efdff63688808b17"}, + {file = "pyzmq-26.0.3-pp39-pypy39_pp73-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:0270b49b6847f0d106d64b5086e9ad5dc8a902413b5dbbb15d12b60f9c1747a4"}, + {file = "pyzmq-26.0.3-pp39-pypy39_pp73-manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:703c60b9910488d3d0954ca585c34f541e506a091a41930e663a098d3b794c67"}, + {file = "pyzmq-26.0.3-pp39-pypy39_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:74423631b6be371edfbf7eabb02ab995c2563fee60a80a30829176842e71722a"}, + {file = "pyzmq-26.0.3-pp39-pypy39_pp73-manylinux_2_28_x86_64.whl", hash = "sha256:4adfbb5451196842a88fda3612e2c0414134874bffb1c2ce83ab4242ec9e027d"}, + {file = "pyzmq-26.0.3-pp39-pypy39_pp73-win_amd64.whl", hash = "sha256:3516119f4f9b8671083a70b6afaa0a070f5683e431ab3dc26e9215620d7ca1ad"}, + {file = "pyzmq-26.0.3.tar.gz", hash = "sha256:dba7d9f2e047dfa2bca3b01f4f84aa5246725203d6284e3790f2ca15fba6b40a"}, ] [package.dependencies] @@ -3812,13 +3860,13 @@ cffi = {version = "*", markers = "implementation_name == \"pypy\""} [[package]] name = "referencing" -version = "0.33.0" +version = "0.35.1" description = "JSON Referencing + Python" optional = false python-versions = ">=3.8" files = [ - {file = "referencing-0.33.0-py3-none-any.whl", hash = "sha256:39240f2ecc770258f28b642dd47fd74bc8b02484de54e1882b74b35ebd779bd5"}, - {file = "referencing-0.33.0.tar.gz", hash = "sha256:c775fedf74bc0f9189c2a3be1c12fd03e8c23f4d371dce795df44e06c5b412f7"}, + {file = "referencing-0.35.1-py3-none-any.whl", hash = "sha256:eda6d3234d62814d1c64e305c1331c9a3a6132da475ab6382eaa997b21ee75de"}, + {file = "referencing-0.35.1.tar.gz", hash = "sha256:25b42124a6c8b632a425174f24087783efb348a6f1e0008e63cd4466fedf703c"}, ] [package.dependencies] @@ -3827,115 +3875,101 @@ rpds-py = ">=0.7.0" [[package]] name = "regex" -version = "2023.12.25" +version = "2024.5.15" description = "Alternative regular expression module, to replace re." optional = false -python-versions = ">=3.7" +python-versions = ">=3.8" files = [ - {file = "regex-2023.12.25-cp310-cp310-macosx_10_9_universal2.whl", hash = "sha256:0694219a1d54336fd0445ea382d49d36882415c0134ee1e8332afd1529f0baa5"}, - {file = "regex-2023.12.25-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:b014333bd0217ad3d54c143de9d4b9a3ca1c5a29a6d0d554952ea071cff0f1f8"}, - {file = "regex-2023.12.25-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:d865984b3f71f6d0af64d0d88f5733521698f6c16f445bb09ce746c92c97c586"}, - {file = "regex-2023.12.25-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:1e0eabac536b4cc7f57a5f3d095bfa557860ab912f25965e08fe1545e2ed8b4c"}, - {file = "regex-2023.12.25-cp310-cp310-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:c25a8ad70e716f96e13a637802813f65d8a6760ef48672aa3502f4c24ea8b400"}, - {file = "regex-2023.12.25-cp310-cp310-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:a9b6d73353f777630626f403b0652055ebfe8ff142a44ec2cf18ae470395766e"}, - {file = "regex-2023.12.25-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:a9cc99d6946d750eb75827cb53c4371b8b0fe89c733a94b1573c9dd16ea6c9e4"}, - {file = "regex-2023.12.25-cp310-cp310-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:88d1f7bef20c721359d8675f7d9f8e414ec5003d8f642fdfd8087777ff7f94b5"}, - {file = "regex-2023.12.25-cp310-cp310-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_12_x86_64.manylinux2010_x86_64.whl", hash = "sha256:cb3fe77aec8f1995611f966d0c656fdce398317f850d0e6e7aebdfe61f40e1cd"}, - {file = "regex-2023.12.25-cp310-cp310-musllinux_1_1_aarch64.whl", hash = "sha256:7aa47c2e9ea33a4a2a05f40fcd3ea36d73853a2aae7b4feab6fc85f8bf2c9704"}, - {file = "regex-2023.12.25-cp310-cp310-musllinux_1_1_i686.whl", hash = "sha256:df26481f0c7a3f8739fecb3e81bc9da3fcfae34d6c094563b9d4670b047312e1"}, - {file = "regex-2023.12.25-cp310-cp310-musllinux_1_1_ppc64le.whl", hash = "sha256:c40281f7d70baf6e0db0c2f7472b31609f5bc2748fe7275ea65a0b4601d9b392"}, - {file = "regex-2023.12.25-cp310-cp310-musllinux_1_1_s390x.whl", hash = "sha256:d94a1db462d5690ebf6ae86d11c5e420042b9898af5dcf278bd97d6bda065423"}, - {file = "regex-2023.12.25-cp310-cp310-musllinux_1_1_x86_64.whl", hash = "sha256:ba1b30765a55acf15dce3f364e4928b80858fa8f979ad41f862358939bdd1f2f"}, - {file = "regex-2023.12.25-cp310-cp310-win32.whl", hash = "sha256:150c39f5b964e4d7dba46a7962a088fbc91f06e606f023ce57bb347a3b2d4630"}, - {file = "regex-2023.12.25-cp310-cp310-win_amd64.whl", hash = "sha256:09da66917262d9481c719599116c7dc0c321ffcec4b1f510c4f8a066f8768105"}, - {file = "regex-2023.12.25-cp311-cp311-macosx_10_9_universal2.whl", hash = "sha256:1b9d811f72210fa9306aeb88385b8f8bcef0dfbf3873410413c00aa94c56c2b6"}, - {file = "regex-2023.12.25-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:d902a43085a308cef32c0d3aea962524b725403fd9373dea18110904003bac97"}, - {file = "regex-2023.12.25-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:d166eafc19f4718df38887b2bbe1467a4f74a9830e8605089ea7a30dd4da8887"}, - {file = "regex-2023.12.25-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:c7ad32824b7f02bb3c9f80306d405a1d9b7bb89362d68b3c5a9be53836caebdb"}, - {file = "regex-2023.12.25-cp311-cp311-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:636ba0a77de609d6510235b7f0e77ec494d2657108f777e8765efc060094c98c"}, - {file = "regex-2023.12.25-cp311-cp311-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:0fda75704357805eb953a3ee15a2b240694a9a514548cd49b3c5124b4e2ad01b"}, - {file = "regex-2023.12.25-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:f72cbae7f6b01591f90814250e636065850c5926751af02bb48da94dfced7baa"}, - {file = "regex-2023.12.25-cp311-cp311-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:db2a0b1857f18b11e3b0e54ddfefc96af46b0896fb678c85f63fb8c37518b3e7"}, - {file = "regex-2023.12.25-cp311-cp311-musllinux_1_1_aarch64.whl", hash = "sha256:7502534e55c7c36c0978c91ba6f61703faf7ce733715ca48f499d3dbbd7657e0"}, - {file = "regex-2023.12.25-cp311-cp311-musllinux_1_1_i686.whl", hash = "sha256:e8c7e08bb566de4faaf11984af13f6bcf6a08f327b13631d41d62592681d24fe"}, - {file = "regex-2023.12.25-cp311-cp311-musllinux_1_1_ppc64le.whl", hash = "sha256:283fc8eed679758de38fe493b7d7d84a198b558942b03f017b1f94dda8efae80"}, - {file = "regex-2023.12.25-cp311-cp311-musllinux_1_1_s390x.whl", hash = "sha256:f44dd4d68697559d007462b0a3a1d9acd61d97072b71f6d1968daef26bc744bd"}, - {file = "regex-2023.12.25-cp311-cp311-musllinux_1_1_x86_64.whl", hash = "sha256:67d3ccfc590e5e7197750fcb3a2915b416a53e2de847a728cfa60141054123d4"}, - {file = "regex-2023.12.25-cp311-cp311-win32.whl", hash = "sha256:68191f80a9bad283432385961d9efe09d783bcd36ed35a60fb1ff3f1ec2efe87"}, - {file = "regex-2023.12.25-cp311-cp311-win_amd64.whl", hash = "sha256:7d2af3f6b8419661a0c421584cfe8aaec1c0e435ce7e47ee2a97e344b98f794f"}, - {file = "regex-2023.12.25-cp312-cp312-macosx_10_9_universal2.whl", hash = "sha256:8a0ccf52bb37d1a700375a6b395bff5dd15c50acb745f7db30415bae3c2b0715"}, - {file = "regex-2023.12.25-cp312-cp312-macosx_10_9_x86_64.whl", hash = "sha256:c3c4a78615b7762740531c27cf46e2f388d8d727d0c0c739e72048beb26c8a9d"}, - {file = "regex-2023.12.25-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:ad83e7545b4ab69216cef4cc47e344d19622e28aabec61574b20257c65466d6a"}, - {file = "regex-2023.12.25-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:b7a635871143661feccce3979e1727c4e094f2bdfd3ec4b90dfd4f16f571a87a"}, - {file = "regex-2023.12.25-cp312-cp312-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:d498eea3f581fbe1b34b59c697512a8baef88212f92e4c7830fcc1499f5b45a5"}, - {file = "regex-2023.12.25-cp312-cp312-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:43f7cd5754d02a56ae4ebb91b33461dc67be8e3e0153f593c509e21d219c5060"}, - {file = "regex-2023.12.25-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:51f4b32f793812714fd5307222a7f77e739b9bc566dc94a18126aba3b92b98a3"}, - {file = "regex-2023.12.25-cp312-cp312-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:ba99d8077424501b9616b43a2d208095746fb1284fc5ba490139651f971d39d9"}, - {file = "regex-2023.12.25-cp312-cp312-musllinux_1_1_aarch64.whl", hash = "sha256:4bfc2b16e3ba8850e0e262467275dd4d62f0d045e0e9eda2bc65078c0110a11f"}, - {file = "regex-2023.12.25-cp312-cp312-musllinux_1_1_i686.whl", hash = "sha256:8c2c19dae8a3eb0ea45a8448356ed561be843b13cbc34b840922ddf565498c1c"}, - {file = "regex-2023.12.25-cp312-cp312-musllinux_1_1_ppc64le.whl", hash = "sha256:60080bb3d8617d96f0fb7e19796384cc2467447ef1c491694850ebd3670bc457"}, - {file = "regex-2023.12.25-cp312-cp312-musllinux_1_1_s390x.whl", hash = "sha256:b77e27b79448e34c2c51c09836033056a0547aa360c45eeeb67803da7b0eedaf"}, - {file = "regex-2023.12.25-cp312-cp312-musllinux_1_1_x86_64.whl", hash = "sha256:518440c991f514331f4850a63560321f833979d145d7d81186dbe2f19e27ae3d"}, - {file = "regex-2023.12.25-cp312-cp312-win32.whl", hash = "sha256:e2610e9406d3b0073636a3a2e80db05a02f0c3169b5632022b4e81c0364bcda5"}, - {file = "regex-2023.12.25-cp312-cp312-win_amd64.whl", hash = "sha256:cc37b9aeebab425f11f27e5e9e6cf580be7206c6582a64467a14dda211abc232"}, - {file = "regex-2023.12.25-cp37-cp37m-macosx_10_9_x86_64.whl", hash = "sha256:da695d75ac97cb1cd725adac136d25ca687da4536154cdc2815f576e4da11c69"}, - {file = "regex-2023.12.25-cp37-cp37m-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:d126361607b33c4eb7b36debc173bf25d7805847346dd4d99b5499e1fef52bc7"}, - {file = "regex-2023.12.25-cp37-cp37m-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:4719bb05094d7d8563a450cf8738d2e1061420f79cfcc1fa7f0a44744c4d8f73"}, - {file = "regex-2023.12.25-cp37-cp37m-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:5dd58946bce44b53b06d94aa95560d0b243eb2fe64227cba50017a8d8b3cd3e2"}, - {file = "regex-2023.12.25-cp37-cp37m-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:22a86d9fff2009302c440b9d799ef2fe322416d2d58fc124b926aa89365ec482"}, - {file = "regex-2023.12.25-cp37-cp37m-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:2aae8101919e8aa05ecfe6322b278f41ce2994c4a430303c4cd163fef746e04f"}, - {file = "regex-2023.12.25-cp37-cp37m-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_12_x86_64.manylinux2010_x86_64.whl", hash = "sha256:e692296c4cc2873967771345a876bcfc1c547e8dd695c6b89342488b0ea55cd8"}, - {file = "regex-2023.12.25-cp37-cp37m-musllinux_1_1_aarch64.whl", hash = "sha256:263ef5cc10979837f243950637fffb06e8daed7f1ac1e39d5910fd29929e489a"}, - {file = "regex-2023.12.25-cp37-cp37m-musllinux_1_1_i686.whl", hash = "sha256:d6f7e255e5fa94642a0724e35406e6cb7001c09d476ab5fce002f652b36d0c39"}, - {file = "regex-2023.12.25-cp37-cp37m-musllinux_1_1_ppc64le.whl", hash = "sha256:88ad44e220e22b63b0f8f81f007e8abbb92874d8ced66f32571ef8beb0643b2b"}, - {file = "regex-2023.12.25-cp37-cp37m-musllinux_1_1_s390x.whl", hash = "sha256:3a17d3ede18f9cedcbe23d2daa8a2cd6f59fe2bf082c567e43083bba3fb00347"}, - {file = "regex-2023.12.25-cp37-cp37m-musllinux_1_1_x86_64.whl", hash = "sha256:d15b274f9e15b1a0b7a45d2ac86d1f634d983ca40d6b886721626c47a400bf39"}, - {file = "regex-2023.12.25-cp37-cp37m-win32.whl", hash = "sha256:ed19b3a05ae0c97dd8f75a5d8f21f7723a8c33bbc555da6bbe1f96c470139d3c"}, - {file = "regex-2023.12.25-cp37-cp37m-win_amd64.whl", hash = "sha256:a6d1047952c0b8104a1d371f88f4ab62e6275567d4458c1e26e9627ad489b445"}, - {file = "regex-2023.12.25-cp38-cp38-macosx_10_9_universal2.whl", hash = "sha256:b43523d7bc2abd757119dbfb38af91b5735eea45537ec6ec3a5ec3f9562a1c53"}, - {file = "regex-2023.12.25-cp38-cp38-macosx_10_9_x86_64.whl", hash = "sha256:efb2d82f33b2212898f1659fb1c2e9ac30493ac41e4d53123da374c3b5541e64"}, - {file = "regex-2023.12.25-cp38-cp38-macosx_11_0_arm64.whl", hash = "sha256:b7fca9205b59c1a3d5031f7e64ed627a1074730a51c2a80e97653e3e9fa0d415"}, - {file = "regex-2023.12.25-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:086dd15e9435b393ae06f96ab69ab2d333f5d65cbe65ca5a3ef0ec9564dfe770"}, - {file = "regex-2023.12.25-cp38-cp38-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:e81469f7d01efed9b53740aedd26085f20d49da65f9c1f41e822a33992cb1590"}, - {file = "regex-2023.12.25-cp38-cp38-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:34e4af5b27232f68042aa40a91c3b9bb4da0eeb31b7632e0091afc4310afe6cb"}, - {file = "regex-2023.12.25-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:9852b76ab558e45b20bf1893b59af64a28bd3820b0c2efc80e0a70a4a3ea51c1"}, - {file = "regex-2023.12.25-cp38-cp38-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:ff100b203092af77d1a5a7abe085b3506b7eaaf9abf65b73b7d6905b6cb76988"}, - {file = "regex-2023.12.25-cp38-cp38-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_12_x86_64.manylinux2010_x86_64.whl", hash = "sha256:cc038b2d8b1470364b1888a98fd22d616fba2b6309c5b5f181ad4483e0017861"}, - {file = "regex-2023.12.25-cp38-cp38-musllinux_1_1_aarch64.whl", hash = "sha256:094ba386bb5c01e54e14434d4caabf6583334090865b23ef58e0424a6286d3dc"}, - {file = "regex-2023.12.25-cp38-cp38-musllinux_1_1_i686.whl", hash = "sha256:5cd05d0f57846d8ba4b71d9c00f6f37d6b97d5e5ef8b3c3840426a475c8f70f4"}, - {file = "regex-2023.12.25-cp38-cp38-musllinux_1_1_ppc64le.whl", hash = "sha256:9aa1a67bbf0f957bbe096375887b2505f5d8ae16bf04488e8b0f334c36e31360"}, - {file = "regex-2023.12.25-cp38-cp38-musllinux_1_1_s390x.whl", hash = "sha256:98a2636994f943b871786c9e82bfe7883ecdaba2ef5df54e1450fa9869d1f756"}, - {file = "regex-2023.12.25-cp38-cp38-musllinux_1_1_x86_64.whl", hash = "sha256:37f8e93a81fc5e5bd8db7e10e62dc64261bcd88f8d7e6640aaebe9bc180d9ce2"}, - {file = "regex-2023.12.25-cp38-cp38-win32.whl", hash = "sha256:d78bd484930c1da2b9679290a41cdb25cc127d783768a0369d6b449e72f88beb"}, - {file = "regex-2023.12.25-cp38-cp38-win_amd64.whl", hash = "sha256:b521dcecebc5b978b447f0f69b5b7f3840eac454862270406a39837ffae4e697"}, - {file = "regex-2023.12.25-cp39-cp39-macosx_10_9_universal2.whl", hash = "sha256:f7bc09bc9c29ebead055bcba136a67378f03d66bf359e87d0f7c759d6d4ffa31"}, - {file = "regex-2023.12.25-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:e14b73607d6231f3cc4622809c196b540a6a44e903bcfad940779c80dffa7be7"}, - {file = "regex-2023.12.25-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:9eda5f7a50141291beda3edd00abc2d4a5b16c29c92daf8d5bd76934150f3edc"}, - {file = "regex-2023.12.25-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:cc6bb9aa69aacf0f6032c307da718f61a40cf970849e471254e0e91c56ffca95"}, - {file = "regex-2023.12.25-cp39-cp39-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:298dc6354d414bc921581be85695d18912bea163a8b23cac9a2562bbcd5088b1"}, - {file = "regex-2023.12.25-cp39-cp39-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:2f4e475a80ecbd15896a976aa0b386c5525d0ed34d5c600b6d3ebac0a67c7ddf"}, - {file = "regex-2023.12.25-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:531ac6cf22b53e0696f8e1d56ce2396311254eb806111ddd3922c9d937151dae"}, - {file = "regex-2023.12.25-cp39-cp39-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:22f3470f7524b6da61e2020672df2f3063676aff444db1daa283c2ea4ed259d6"}, - {file = "regex-2023.12.25-cp39-cp39-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_12_x86_64.manylinux2010_x86_64.whl", hash = "sha256:89723d2112697feaa320c9d351e5f5e7b841e83f8b143dba8e2d2b5f04e10923"}, - {file = "regex-2023.12.25-cp39-cp39-musllinux_1_1_aarch64.whl", hash = "sha256:0ecf44ddf9171cd7566ef1768047f6e66975788258b1c6c6ca78098b95cf9a3d"}, - {file = "regex-2023.12.25-cp39-cp39-musllinux_1_1_i686.whl", hash = "sha256:905466ad1702ed4acfd67a902af50b8db1feeb9781436372261808df7a2a7bca"}, - {file = "regex-2023.12.25-cp39-cp39-musllinux_1_1_ppc64le.whl", hash = "sha256:4558410b7a5607a645e9804a3e9dd509af12fb72b9825b13791a37cd417d73a5"}, - {file = "regex-2023.12.25-cp39-cp39-musllinux_1_1_s390x.whl", hash = "sha256:7e316026cc1095f2a3e8cc012822c99f413b702eaa2ca5408a513609488cb62f"}, - {file = "regex-2023.12.25-cp39-cp39-musllinux_1_1_x86_64.whl", hash = "sha256:3b1de218d5375cd6ac4b5493e0b9f3df2be331e86520f23382f216c137913d20"}, - {file = "regex-2023.12.25-cp39-cp39-win32.whl", hash = "sha256:11a963f8e25ab5c61348d090bf1b07f1953929c13bd2309a0662e9ff680763c9"}, - {file = "regex-2023.12.25-cp39-cp39-win_amd64.whl", hash = "sha256:e693e233ac92ba83a87024e1d32b5f9ab15ca55ddd916d878146f4e3406b5c91"}, - {file = "regex-2023.12.25.tar.gz", hash = "sha256:29171aa128da69afdf4bde412d5bedc335f2ca8fcfe4489038577d05f16181e5"}, + {file = "regex-2024.5.15-cp310-cp310-macosx_10_9_universal2.whl", hash = "sha256:a81e3cfbae20378d75185171587cbf756015ccb14840702944f014e0d93ea09f"}, + {file = "regex-2024.5.15-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:7b59138b219ffa8979013be7bc85bb60c6f7b7575df3d56dc1e403a438c7a3f6"}, + {file = "regex-2024.5.15-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:a0bd000c6e266927cb7a1bc39d55be95c4b4f65c5be53e659537537e019232b1"}, + {file = "regex-2024.5.15-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:5eaa7ddaf517aa095fa8da0b5015c44d03da83f5bd49c87961e3c997daed0de7"}, + {file = "regex-2024.5.15-cp310-cp310-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:ba68168daedb2c0bab7fd7e00ced5ba90aebf91024dea3c88ad5063c2a562cca"}, + {file = "regex-2024.5.15-cp310-cp310-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:6e8d717bca3a6e2064fc3a08df5cbe366369f4b052dcd21b7416e6d71620dca1"}, + {file = "regex-2024.5.15-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:1337b7dbef9b2f71121cdbf1e97e40de33ff114801263b275aafd75303bd62b5"}, + {file = "regex-2024.5.15-cp310-cp310-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:f9ebd0a36102fcad2f03696e8af4ae682793a5d30b46c647eaf280d6cfb32796"}, + {file = "regex-2024.5.15-cp310-cp310-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_12_x86_64.manylinux2010_x86_64.whl", hash = "sha256:9efa1a32ad3a3ea112224897cdaeb6aa00381627f567179c0314f7b65d354c62"}, + {file = "regex-2024.5.15-cp310-cp310-musllinux_1_2_aarch64.whl", hash = "sha256:1595f2d10dff3d805e054ebdc41c124753631b6a471b976963c7b28543cf13b0"}, + {file = "regex-2024.5.15-cp310-cp310-musllinux_1_2_i686.whl", hash = "sha256:b802512f3e1f480f41ab5f2cfc0e2f761f08a1f41092d6718868082fc0d27143"}, + {file = "regex-2024.5.15-cp310-cp310-musllinux_1_2_ppc64le.whl", hash = "sha256:a0981022dccabca811e8171f913de05720590c915b033b7e601f35ce4ea7019f"}, + {file = "regex-2024.5.15-cp310-cp310-musllinux_1_2_s390x.whl", hash = "sha256:19068a6a79cf99a19ccefa44610491e9ca02c2be3305c7760d3831d38a467a6f"}, + {file = "regex-2024.5.15-cp310-cp310-musllinux_1_2_x86_64.whl", hash = "sha256:1b5269484f6126eee5e687785e83c6b60aad7663dafe842b34691157e5083e53"}, + {file = "regex-2024.5.15-cp310-cp310-win32.whl", hash = "sha256:ada150c5adfa8fbcbf321c30c751dc67d2f12f15bd183ffe4ec7cde351d945b3"}, + {file = "regex-2024.5.15-cp310-cp310-win_amd64.whl", hash = "sha256:ac394ff680fc46b97487941f5e6ae49a9f30ea41c6c6804832063f14b2a5a145"}, + {file = "regex-2024.5.15-cp311-cp311-macosx_10_9_universal2.whl", hash = "sha256:f5b1dff3ad008dccf18e652283f5e5339d70bf8ba7c98bf848ac33db10f7bc7a"}, + {file = "regex-2024.5.15-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:c6a2b494a76983df8e3d3feea9b9ffdd558b247e60b92f877f93a1ff43d26656"}, + {file = "regex-2024.5.15-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:a32b96f15c8ab2e7d27655969a23895eb799de3665fa94349f3b2fbfd547236f"}, + {file = "regex-2024.5.15-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:10002e86e6068d9e1c91eae8295ef690f02f913c57db120b58fdd35a6bb1af35"}, + {file = "regex-2024.5.15-cp311-cp311-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:ec54d5afa89c19c6dd8541a133be51ee1017a38b412b1321ccb8d6ddbeb4cf7d"}, + {file = "regex-2024.5.15-cp311-cp311-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:10e4ce0dca9ae7a66e6089bb29355d4432caed736acae36fef0fdd7879f0b0cb"}, + {file = "regex-2024.5.15-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:3e507ff1e74373c4d3038195fdd2af30d297b4f0950eeda6f515ae3d84a1770f"}, + {file = "regex-2024.5.15-cp311-cp311-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:d1f059a4d795e646e1c37665b9d06062c62d0e8cc3c511fe01315973a6542e40"}, + {file = "regex-2024.5.15-cp311-cp311-musllinux_1_2_aarch64.whl", hash = "sha256:0721931ad5fe0dda45d07f9820b90b2148ccdd8e45bb9e9b42a146cb4f695649"}, + {file = "regex-2024.5.15-cp311-cp311-musllinux_1_2_i686.whl", hash = "sha256:833616ddc75ad595dee848ad984d067f2f31be645d603e4d158bba656bbf516c"}, + {file = "regex-2024.5.15-cp311-cp311-musllinux_1_2_ppc64le.whl", hash = "sha256:287eb7f54fc81546346207c533ad3c2c51a8d61075127d7f6d79aaf96cdee890"}, + {file = "regex-2024.5.15-cp311-cp311-musllinux_1_2_s390x.whl", hash = "sha256:19dfb1c504781a136a80ecd1fff9f16dddf5bb43cec6871778c8a907a085bb3d"}, + {file = "regex-2024.5.15-cp311-cp311-musllinux_1_2_x86_64.whl", hash = "sha256:119af6e56dce35e8dfb5222573b50c89e5508d94d55713c75126b753f834de68"}, + {file = "regex-2024.5.15-cp311-cp311-win32.whl", hash = "sha256:1c1c174d6ec38d6c8a7504087358ce9213d4332f6293a94fbf5249992ba54efa"}, + {file = "regex-2024.5.15-cp311-cp311-win_amd64.whl", hash = "sha256:9e717956dcfd656f5055cc70996ee2cc82ac5149517fc8e1b60261b907740201"}, + {file = "regex-2024.5.15-cp312-cp312-macosx_10_9_universal2.whl", hash = "sha256:632b01153e5248c134007209b5c6348a544ce96c46005d8456de1d552455b014"}, + {file = "regex-2024.5.15-cp312-cp312-macosx_10_9_x86_64.whl", hash = "sha256:e64198f6b856d48192bf921421fdd8ad8eb35e179086e99e99f711957ffedd6e"}, + {file = "regex-2024.5.15-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:68811ab14087b2f6e0fc0c2bae9ad689ea3584cad6917fc57be6a48bbd012c49"}, + {file = "regex-2024.5.15-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:f8ec0c2fea1e886a19c3bee0cd19d862b3aa75dcdfb42ebe8ed30708df64687a"}, + {file = "regex-2024.5.15-cp312-cp312-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:d0c0c0003c10f54a591d220997dd27d953cd9ccc1a7294b40a4be5312be8797b"}, + {file = "regex-2024.5.15-cp312-cp312-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:2431b9e263af1953c55abbd3e2efca67ca80a3de8a0437cb58e2421f8184717a"}, + {file = "regex-2024.5.15-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:4a605586358893b483976cffc1723fb0f83e526e8f14c6e6614e75919d9862cf"}, + {file = "regex-2024.5.15-cp312-cp312-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:391d7f7f1e409d192dba8bcd42d3e4cf9e598f3979cdaed6ab11288da88cb9f2"}, + {file = "regex-2024.5.15-cp312-cp312-musllinux_1_2_aarch64.whl", hash = "sha256:9ff11639a8d98969c863d4617595eb5425fd12f7c5ef6621a4b74b71ed8726d5"}, + {file = "regex-2024.5.15-cp312-cp312-musllinux_1_2_i686.whl", hash = "sha256:4eee78a04e6c67e8391edd4dad3279828dd66ac4b79570ec998e2155d2e59fd5"}, + {file = "regex-2024.5.15-cp312-cp312-musllinux_1_2_ppc64le.whl", hash = "sha256:8fe45aa3f4aa57faabbc9cb46a93363edd6197cbc43523daea044e9ff2fea83e"}, + {file = "regex-2024.5.15-cp312-cp312-musllinux_1_2_s390x.whl", hash = "sha256:d0a3d8d6acf0c78a1fff0e210d224b821081330b8524e3e2bc5a68ef6ab5803d"}, + {file = "regex-2024.5.15-cp312-cp312-musllinux_1_2_x86_64.whl", hash = "sha256:c486b4106066d502495b3025a0a7251bf37ea9540433940a23419461ab9f2a80"}, + {file = "regex-2024.5.15-cp312-cp312-win32.whl", hash = "sha256:c49e15eac7c149f3670b3e27f1f28a2c1ddeccd3a2812cba953e01be2ab9b5fe"}, + {file = "regex-2024.5.15-cp312-cp312-win_amd64.whl", hash = "sha256:673b5a6da4557b975c6c90198588181029c60793835ce02f497ea817ff647cb2"}, + {file = "regex-2024.5.15-cp38-cp38-macosx_10_9_universal2.whl", hash = "sha256:87e2a9c29e672fc65523fb47a90d429b70ef72b901b4e4b1bd42387caf0d6835"}, + {file = "regex-2024.5.15-cp38-cp38-macosx_10_9_x86_64.whl", hash = "sha256:c3bea0ba8b73b71b37ac833a7f3fd53825924165da6a924aec78c13032f20850"}, + {file = "regex-2024.5.15-cp38-cp38-macosx_11_0_arm64.whl", hash = "sha256:bfc4f82cabe54f1e7f206fd3d30fda143f84a63fe7d64a81558d6e5f2e5aaba9"}, + {file = "regex-2024.5.15-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:e5bb9425fe881d578aeca0b2b4b3d314ec88738706f66f219c194d67179337cb"}, + {file = "regex-2024.5.15-cp38-cp38-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:64c65783e96e563103d641760664125e91bd85d8e49566ee560ded4da0d3e704"}, + {file = "regex-2024.5.15-cp38-cp38-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:cf2430df4148b08fb4324b848672514b1385ae3807651f3567871f130a728cc3"}, + {file = "regex-2024.5.15-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:5397de3219a8b08ae9540c48f602996aa6b0b65d5a61683e233af8605c42b0f2"}, + {file = "regex-2024.5.15-cp38-cp38-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:455705d34b4154a80ead722f4f185b04c4237e8e8e33f265cd0798d0e44825fa"}, + {file = "regex-2024.5.15-cp38-cp38-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_12_x86_64.manylinux2010_x86_64.whl", hash = "sha256:b2b6f1b3bb6f640c1a92be3bbfbcb18657b125b99ecf141fb3310b5282c7d4ed"}, + {file = "regex-2024.5.15-cp38-cp38-musllinux_1_2_aarch64.whl", hash = "sha256:3ad070b823ca5890cab606c940522d05d3d22395d432f4aaaf9d5b1653e47ced"}, + {file = "regex-2024.5.15-cp38-cp38-musllinux_1_2_i686.whl", hash = "sha256:5b5467acbfc153847d5adb21e21e29847bcb5870e65c94c9206d20eb4e99a384"}, + {file = "regex-2024.5.15-cp38-cp38-musllinux_1_2_ppc64le.whl", hash = "sha256:e6662686aeb633ad65be2a42b4cb00178b3fbf7b91878f9446075c404ada552f"}, + {file = "regex-2024.5.15-cp38-cp38-musllinux_1_2_s390x.whl", hash = "sha256:2b4c884767504c0e2401babe8b5b7aea9148680d2e157fa28f01529d1f7fcf67"}, + {file = "regex-2024.5.15-cp38-cp38-musllinux_1_2_x86_64.whl", hash = "sha256:3cd7874d57f13bf70078f1ff02b8b0aa48d5b9ed25fc48547516c6aba36f5741"}, + {file = "regex-2024.5.15-cp38-cp38-win32.whl", hash = "sha256:e4682f5ba31f475d58884045c1a97a860a007d44938c4c0895f41d64481edbc9"}, + {file = "regex-2024.5.15-cp38-cp38-win_amd64.whl", hash = "sha256:d99ceffa25ac45d150e30bd9ed14ec6039f2aad0ffa6bb87a5936f5782fc1569"}, + {file = "regex-2024.5.15-cp39-cp39-macosx_10_9_universal2.whl", hash = "sha256:13cdaf31bed30a1e1c2453ef6015aa0983e1366fad2667657dbcac7b02f67133"}, + {file = "regex-2024.5.15-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:cac27dcaa821ca271855a32188aa61d12decb6fe45ffe3e722401fe61e323cd1"}, + {file = "regex-2024.5.15-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:7dbe2467273b875ea2de38ded4eba86cbcbc9a1a6d0aa11dcf7bd2e67859c435"}, + {file = "regex-2024.5.15-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:64f18a9a3513a99c4bef0e3efd4c4a5b11228b48aa80743be822b71e132ae4f5"}, + {file = "regex-2024.5.15-cp39-cp39-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:d347a741ea871c2e278fde6c48f85136c96b8659b632fb57a7d1ce1872547600"}, + {file = "regex-2024.5.15-cp39-cp39-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:1878b8301ed011704aea4c806a3cadbd76f84dece1ec09cc9e4dc934cfa5d4da"}, + {file = "regex-2024.5.15-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:4babf07ad476aaf7830d77000874d7611704a7fcf68c9c2ad151f5d94ae4bfc4"}, + {file = "regex-2024.5.15-cp39-cp39-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:35cb514e137cb3488bce23352af3e12fb0dbedd1ee6e60da053c69fb1b29cc6c"}, + {file = "regex-2024.5.15-cp39-cp39-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_12_x86_64.manylinux2010_x86_64.whl", hash = "sha256:cdd09d47c0b2efee9378679f8510ee6955d329424c659ab3c5e3a6edea696294"}, + {file = "regex-2024.5.15-cp39-cp39-musllinux_1_2_aarch64.whl", hash = "sha256:72d7a99cd6b8f958e85fc6ca5b37c4303294954eac1376535b03c2a43eb72629"}, + {file = "regex-2024.5.15-cp39-cp39-musllinux_1_2_i686.whl", hash = "sha256:a094801d379ab20c2135529948cb84d417a2169b9bdceda2a36f5f10977ebc16"}, + {file = "regex-2024.5.15-cp39-cp39-musllinux_1_2_ppc64le.whl", hash = "sha256:c0c18345010870e58238790a6779a1219b4d97bd2e77e1140e8ee5d14df071aa"}, + {file = "regex-2024.5.15-cp39-cp39-musllinux_1_2_s390x.whl", hash = "sha256:16093f563098448ff6b1fa68170e4acbef94e6b6a4e25e10eae8598bb1694b5d"}, + {file = "regex-2024.5.15-cp39-cp39-musllinux_1_2_x86_64.whl", hash = "sha256:e38a7d4e8f633a33b4c7350fbd8bad3b70bf81439ac67ac38916c4a86b465456"}, + {file = "regex-2024.5.15-cp39-cp39-win32.whl", hash = "sha256:71a455a3c584a88f654b64feccc1e25876066c4f5ef26cd6dd711308aa538694"}, + {file = "regex-2024.5.15-cp39-cp39-win_amd64.whl", hash = "sha256:cab12877a9bdafde5500206d1020a584355a97884dfd388af3699e9137bf7388"}, + {file = "regex-2024.5.15.tar.gz", hash = "sha256:d3ee02d9e5f482cc8309134a91eeaacbdd2261ba111b0fef3748eeb4913e6a2c"}, ] [[package]] name = "requests" -version = "2.31.0" +version = "2.32.2" description = "Python HTTP for Humans." optional = false -python-versions = ">=3.7" +python-versions = ">=3.8" files = [ - {file = "requests-2.31.0-py3-none-any.whl", hash = "sha256:58cd2187c01e70e6e26505bca751777aa9f2ee0b7f4300988b709f44e013003f"}, - {file = "requests-2.31.0.tar.gz", hash = "sha256:942c5a758f98d790eaed1a29cb6eefc7ffb0d1cf7af05c3d2791656dbd6ad1e1"}, + {file = "requests-2.32.2-py3-none-any.whl", hash = "sha256:fc06670dd0ed212426dfeb94fc1b983d917c4f9847c863f313c9dfaaffb7c23c"}, + {file = "requests-2.32.2.tar.gz", hash = "sha256:dd951ff5ecf3e3b3aa26b40703ba77495dab41da839ae72ef3c8e5d8e2433289"}, ] [package.dependencies] @@ -3950,222 +3984,206 @@ use-chardet-on-py3 = ["chardet (>=3.0.2,<6)"] [[package]] name = "rpds-py" -version = "0.17.1" +version = "0.18.1" description = "Python bindings to Rust's persistent data structures (rpds)" optional = false python-versions = ">=3.8" files = [ - {file = "rpds_py-0.17.1-cp310-cp310-macosx_10_12_x86_64.whl", hash = "sha256:4128980a14ed805e1b91a7ed551250282a8ddf8201a4e9f8f5b7e6225f54170d"}, - {file = "rpds_py-0.17.1-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:ff1dcb8e8bc2261a088821b2595ef031c91d499a0c1b031c152d43fe0a6ecec8"}, - {file = "rpds_py-0.17.1-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:d65e6b4f1443048eb7e833c2accb4fa7ee67cc7d54f31b4f0555b474758bee55"}, - {file = "rpds_py-0.17.1-cp310-cp310-manylinux_2_17_armv7l.manylinux2014_armv7l.whl", hash = "sha256:a71169d505af63bb4d20d23a8fbd4c6ce272e7bce6cc31f617152aa784436f29"}, - {file = "rpds_py-0.17.1-cp310-cp310-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:436474f17733c7dca0fbf096d36ae65277e8645039df12a0fa52445ca494729d"}, - {file = "rpds_py-0.17.1-cp310-cp310-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:10162fe3f5f47c37ebf6d8ff5a2368508fe22007e3077bf25b9c7d803454d921"}, - {file = "rpds_py-0.17.1-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:720215373a280f78a1814becb1312d4e4d1077b1202a56d2b0815e95ccb99ce9"}, - {file = "rpds_py-0.17.1-cp310-cp310-manylinux_2_5_i686.manylinux1_i686.whl", hash = "sha256:70fcc6c2906cfa5c6a552ba7ae2ce64b6c32f437d8f3f8eea49925b278a61453"}, - {file = "rpds_py-0.17.1-cp310-cp310-musllinux_1_2_aarch64.whl", hash = "sha256:91e5a8200e65aaac342a791272c564dffcf1281abd635d304d6c4e6b495f29dc"}, - {file = "rpds_py-0.17.1-cp310-cp310-musllinux_1_2_i686.whl", hash = "sha256:99f567dae93e10be2daaa896e07513dd4bf9c2ecf0576e0533ac36ba3b1d5394"}, - {file = "rpds_py-0.17.1-cp310-cp310-musllinux_1_2_x86_64.whl", hash = "sha256:24e4900a6643f87058a27320f81336d527ccfe503984528edde4bb660c8c8d59"}, - {file = "rpds_py-0.17.1-cp310-none-win32.whl", hash = "sha256:0bfb09bf41fe7c51413f563373e5f537eaa653d7adc4830399d4e9bdc199959d"}, - {file = "rpds_py-0.17.1-cp310-none-win_amd64.whl", hash = "sha256:20de7b7179e2031a04042e85dc463a93a82bc177eeba5ddd13ff746325558aa6"}, - {file = "rpds_py-0.17.1-cp311-cp311-macosx_10_12_x86_64.whl", hash = "sha256:65dcf105c1943cba45d19207ef51b8bc46d232a381e94dd38719d52d3980015b"}, - {file = "rpds_py-0.17.1-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:01f58a7306b64e0a4fe042047dd2b7d411ee82e54240284bab63e325762c1147"}, - {file = "rpds_py-0.17.1-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:071bc28c589b86bc6351a339114fb7a029f5cddbaca34103aa573eba7b482382"}, - {file = "rpds_py-0.17.1-cp311-cp311-manylinux_2_17_armv7l.manylinux2014_armv7l.whl", hash = "sha256:ae35e8e6801c5ab071b992cb2da958eee76340e6926ec693b5ff7d6381441745"}, - {file = "rpds_py-0.17.1-cp311-cp311-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:149c5cd24f729e3567b56e1795f74577aa3126c14c11e457bec1b1c90d212e38"}, - {file = "rpds_py-0.17.1-cp311-cp311-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:e796051f2070f47230c745d0a77a91088fbee2cc0502e9b796b9c6471983718c"}, - {file = "rpds_py-0.17.1-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:60e820ee1004327609b28db8307acc27f5f2e9a0b185b2064c5f23e815f248f8"}, - {file = "rpds_py-0.17.1-cp311-cp311-manylinux_2_5_i686.manylinux1_i686.whl", hash = "sha256:1957a2ab607f9added64478a6982742eb29f109d89d065fa44e01691a20fc20a"}, - {file = "rpds_py-0.17.1-cp311-cp311-musllinux_1_2_aarch64.whl", hash = "sha256:8587fd64c2a91c33cdc39d0cebdaf30e79491cc029a37fcd458ba863f8815383"}, - {file = "rpds_py-0.17.1-cp311-cp311-musllinux_1_2_i686.whl", hash = "sha256:4dc889a9d8a34758d0fcc9ac86adb97bab3fb7f0c4d29794357eb147536483fd"}, - {file = "rpds_py-0.17.1-cp311-cp311-musllinux_1_2_x86_64.whl", hash = "sha256:2953937f83820376b5979318840f3ee47477d94c17b940fe31d9458d79ae7eea"}, - {file = "rpds_py-0.17.1-cp311-none-win32.whl", hash = "sha256:1bfcad3109c1e5ba3cbe2f421614e70439f72897515a96c462ea657261b96518"}, - {file = "rpds_py-0.17.1-cp311-none-win_amd64.whl", hash = "sha256:99da0a4686ada4ed0f778120a0ea8d066de1a0a92ab0d13ae68492a437db78bf"}, - {file = "rpds_py-0.17.1-cp312-cp312-macosx_10_12_x86_64.whl", hash = "sha256:1dc29db3900cb1bb40353772417800f29c3d078dbc8024fd64655a04ee3c4bdf"}, - {file = "rpds_py-0.17.1-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:82ada4a8ed9e82e443fcef87e22a3eed3654dd3adf6e3b3a0deb70f03e86142a"}, - {file = "rpds_py-0.17.1-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:1d36b2b59e8cc6e576f8f7b671e32f2ff43153f0ad6d0201250a7c07f25d570e"}, - {file = "rpds_py-0.17.1-cp312-cp312-manylinux_2_17_armv7l.manylinux2014_armv7l.whl", hash = "sha256:3677fcca7fb728c86a78660c7fb1b07b69b281964673f486ae72860e13f512ad"}, - {file = "rpds_py-0.17.1-cp312-cp312-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:516fb8c77805159e97a689e2f1c80655c7658f5af601c34ffdb916605598cda2"}, - {file = "rpds_py-0.17.1-cp312-cp312-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:df3b6f45ba4515632c5064e35ca7f31d51d13d1479673185ba8f9fefbbed58b9"}, - {file = "rpds_py-0.17.1-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:a967dd6afda7715d911c25a6ba1517975acd8d1092b2f326718725461a3d33f9"}, - {file = "rpds_py-0.17.1-cp312-cp312-manylinux_2_5_i686.manylinux1_i686.whl", hash = "sha256:dbbb95e6fc91ea3102505d111b327004d1c4ce98d56a4a02e82cd451f9f57140"}, - {file = "rpds_py-0.17.1-cp312-cp312-musllinux_1_2_aarch64.whl", hash = "sha256:02866e060219514940342a1f84303a1ef7a1dad0ac311792fbbe19b521b489d2"}, - {file = "rpds_py-0.17.1-cp312-cp312-musllinux_1_2_i686.whl", hash = "sha256:2528ff96d09f12e638695f3a2e0c609c7b84c6df7c5ae9bfeb9252b6fa686253"}, - {file = "rpds_py-0.17.1-cp312-cp312-musllinux_1_2_x86_64.whl", hash = "sha256:bd345a13ce06e94c753dab52f8e71e5252aec1e4f8022d24d56decd31e1b9b23"}, - {file = "rpds_py-0.17.1-cp312-none-win32.whl", hash = "sha256:2a792b2e1d3038daa83fa474d559acfd6dc1e3650ee93b2662ddc17dbff20ad1"}, - {file = "rpds_py-0.17.1-cp312-none-win_amd64.whl", hash = "sha256:292f7344a3301802e7c25c53792fae7d1593cb0e50964e7bcdcc5cf533d634e3"}, - {file = "rpds_py-0.17.1-cp38-cp38-macosx_10_12_x86_64.whl", hash = "sha256:8ffe53e1d8ef2520ebcf0c9fec15bb721da59e8ef283b6ff3079613b1e30513d"}, - {file = "rpds_py-0.17.1-cp38-cp38-macosx_11_0_arm64.whl", hash = "sha256:4341bd7579611cf50e7b20bb8c2e23512a3dc79de987a1f411cb458ab670eb90"}, - {file = "rpds_py-0.17.1-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:2f4eb548daf4836e3b2c662033bfbfc551db58d30fd8fe660314f86bf8510b93"}, - {file = "rpds_py-0.17.1-cp38-cp38-manylinux_2_17_armv7l.manylinux2014_armv7l.whl", hash = "sha256:b686f25377f9c006acbac63f61614416a6317133ab7fafe5de5f7dc8a06d42eb"}, - {file = "rpds_py-0.17.1-cp38-cp38-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:4e21b76075c01d65d0f0f34302b5a7457d95721d5e0667aea65e5bb3ab415c25"}, - {file = "rpds_py-0.17.1-cp38-cp38-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:b86b21b348f7e5485fae740d845c65a880f5d1eda1e063bc59bef92d1f7d0c55"}, - {file = "rpds_py-0.17.1-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:f175e95a197f6a4059b50757a3dca33b32b61691bdbd22c29e8a8d21d3914cae"}, - {file = "rpds_py-0.17.1-cp38-cp38-manylinux_2_5_i686.manylinux1_i686.whl", hash = "sha256:1701fc54460ae2e5efc1dd6350eafd7a760f516df8dbe51d4a1c79d69472fbd4"}, - {file = "rpds_py-0.17.1-cp38-cp38-musllinux_1_2_aarch64.whl", hash = "sha256:9051e3d2af8f55b42061603e29e744724cb5f65b128a491446cc029b3e2ea896"}, - {file = "rpds_py-0.17.1-cp38-cp38-musllinux_1_2_i686.whl", hash = "sha256:7450dbd659fed6dd41d1a7d47ed767e893ba402af8ae664c157c255ec6067fde"}, - {file = "rpds_py-0.17.1-cp38-cp38-musllinux_1_2_x86_64.whl", hash = "sha256:5a024fa96d541fd7edaa0e9d904601c6445e95a729a2900c5aec6555fe921ed6"}, - {file = "rpds_py-0.17.1-cp38-none-win32.whl", hash = "sha256:da1ead63368c04a9bded7904757dfcae01eba0e0f9bc41d3d7f57ebf1c04015a"}, - {file = "rpds_py-0.17.1-cp38-none-win_amd64.whl", hash = "sha256:841320e1841bb53fada91c9725e766bb25009cfd4144e92298db296fb6c894fb"}, - {file = "rpds_py-0.17.1-cp39-cp39-macosx_10_12_x86_64.whl", hash = "sha256:f6c43b6f97209e370124baf2bf40bb1e8edc25311a158867eb1c3a5d449ebc7a"}, - {file = "rpds_py-0.17.1-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:5e7d63ec01fe7c76c2dbb7e972fece45acbb8836e72682bde138e7e039906e2c"}, - {file = "rpds_py-0.17.1-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:81038ff87a4e04c22e1d81f947c6ac46f122e0c80460b9006e6517c4d842a6ec"}, - {file = "rpds_py-0.17.1-cp39-cp39-manylinux_2_17_armv7l.manylinux2014_armv7l.whl", hash = "sha256:810685321f4a304b2b55577c915bece4c4a06dfe38f6e62d9cc1d6ca8ee86b99"}, - {file = "rpds_py-0.17.1-cp39-cp39-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:25f071737dae674ca8937a73d0f43f5a52e92c2d178330b4c0bb6ab05586ffa6"}, - {file = "rpds_py-0.17.1-cp39-cp39-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:aa5bfb13f1e89151ade0eb812f7b0d7a4d643406caaad65ce1cbabe0a66d695f"}, - {file = "rpds_py-0.17.1-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:dfe07308b311a8293a0d5ef4e61411c5c20f682db6b5e73de6c7c8824272c256"}, - {file = "rpds_py-0.17.1-cp39-cp39-manylinux_2_5_i686.manylinux1_i686.whl", hash = "sha256:a000133a90eea274a6f28adc3084643263b1e7c1a5a66eb0a0a7a36aa757ed74"}, - {file = "rpds_py-0.17.1-cp39-cp39-musllinux_1_2_aarch64.whl", hash = "sha256:5d0e8a6434a3fbf77d11448c9c25b2f25244226cfbec1a5159947cac5b8c5fa4"}, - {file = "rpds_py-0.17.1-cp39-cp39-musllinux_1_2_i686.whl", hash = "sha256:efa767c220d94aa4ac3a6dd3aeb986e9f229eaf5bce92d8b1b3018d06bed3772"}, - {file = "rpds_py-0.17.1-cp39-cp39-musllinux_1_2_x86_64.whl", hash = "sha256:dbc56680ecf585a384fbd93cd42bc82668b77cb525343170a2d86dafaed2a84b"}, - {file = "rpds_py-0.17.1-cp39-none-win32.whl", hash = "sha256:270987bc22e7e5a962b1094953ae901395e8c1e1e83ad016c5cfcfff75a15a3f"}, - {file = "rpds_py-0.17.1-cp39-none-win_amd64.whl", hash = "sha256:2a7b2f2f56a16a6d62e55354dd329d929560442bd92e87397b7a9586a32e3e76"}, - {file = "rpds_py-0.17.1-pp310-pypy310_pp73-macosx_10_12_x86_64.whl", hash = "sha256:a3264e3e858de4fc601741498215835ff324ff2482fd4e4af61b46512dd7fc83"}, - {file = "rpds_py-0.17.1-pp310-pypy310_pp73-macosx_11_0_arm64.whl", hash = "sha256:f2f3b28b40fddcb6c1f1f6c88c6f3769cd933fa493ceb79da45968a21dccc920"}, - {file = "rpds_py-0.17.1-pp310-pypy310_pp73-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:9584f8f52010295a4a417221861df9bea4c72d9632562b6e59b3c7b87a1522b7"}, - {file = "rpds_py-0.17.1-pp310-pypy310_pp73-manylinux_2_17_armv7l.manylinux2014_armv7l.whl", hash = "sha256:c64602e8be701c6cfe42064b71c84ce62ce66ddc6422c15463fd8127db3d8066"}, - {file = "rpds_py-0.17.1-pp310-pypy310_pp73-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:060f412230d5f19fc8c8b75f315931b408d8ebf56aec33ef4168d1b9e54200b1"}, - {file = "rpds_py-0.17.1-pp310-pypy310_pp73-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:b9412abdf0ba70faa6e2ee6c0cc62a8defb772e78860cef419865917d86c7342"}, - {file = "rpds_py-0.17.1-pp310-pypy310_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:9737bdaa0ad33d34c0efc718741abaafce62fadae72c8b251df9b0c823c63b22"}, - {file = "rpds_py-0.17.1-pp310-pypy310_pp73-manylinux_2_5_i686.manylinux1_i686.whl", hash = "sha256:9f0e4dc0f17dcea4ab9d13ac5c666b6b5337042b4d8f27e01b70fae41dd65c57"}, - {file = "rpds_py-0.17.1-pp310-pypy310_pp73-musllinux_1_2_aarch64.whl", hash = "sha256:1db228102ab9d1ff4c64148c96320d0be7044fa28bd865a9ce628ce98da5973d"}, - {file = "rpds_py-0.17.1-pp310-pypy310_pp73-musllinux_1_2_i686.whl", hash = "sha256:d8bbd8e56f3ba25a7d0cf980fc42b34028848a53a0e36c9918550e0280b9d0b6"}, - {file = "rpds_py-0.17.1-pp310-pypy310_pp73-musllinux_1_2_x86_64.whl", hash = "sha256:be22ae34d68544df293152b7e50895ba70d2a833ad9566932d750d3625918b82"}, - {file = "rpds_py-0.17.1-pp38-pypy38_pp73-macosx_10_12_x86_64.whl", hash = "sha256:bf046179d011e6114daf12a534d874958b039342b347348a78b7cdf0dd9d6041"}, - {file = "rpds_py-0.17.1-pp38-pypy38_pp73-macosx_11_0_arm64.whl", hash = "sha256:1a746a6d49665058a5896000e8d9d2f1a6acba8a03b389c1e4c06e11e0b7f40d"}, - {file = "rpds_py-0.17.1-pp38-pypy38_pp73-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:f0b8bf5b8db49d8fd40f54772a1dcf262e8be0ad2ab0206b5a2ec109c176c0a4"}, - {file = "rpds_py-0.17.1-pp38-pypy38_pp73-manylinux_2_17_armv7l.manylinux2014_armv7l.whl", hash = "sha256:f7f4cb1f173385e8a39c29510dd11a78bf44e360fb75610594973f5ea141028b"}, - {file = "rpds_py-0.17.1-pp38-pypy38_pp73-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:7fbd70cb8b54fe745301921b0816c08b6d917593429dfc437fd024b5ba713c58"}, - {file = "rpds_py-0.17.1-pp38-pypy38_pp73-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:9bdf1303df671179eaf2cb41e8515a07fc78d9d00f111eadbe3e14262f59c3d0"}, - {file = "rpds_py-0.17.1-pp38-pypy38_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:fad059a4bd14c45776600d223ec194e77db6c20255578bb5bcdd7c18fd169361"}, - {file = "rpds_py-0.17.1-pp38-pypy38_pp73-manylinux_2_5_i686.manylinux1_i686.whl", hash = "sha256:3664d126d3388a887db44c2e293f87d500c4184ec43d5d14d2d2babdb4c64cad"}, - {file = "rpds_py-0.17.1-pp38-pypy38_pp73-musllinux_1_2_aarch64.whl", hash = "sha256:698ea95a60c8b16b58be9d854c9f993c639f5c214cf9ba782eca53a8789d6b19"}, - {file = "rpds_py-0.17.1-pp38-pypy38_pp73-musllinux_1_2_i686.whl", hash = "sha256:c3d2010656999b63e628a3c694f23020322b4178c450dc478558a2b6ef3cb9bb"}, - {file = "rpds_py-0.17.1-pp38-pypy38_pp73-musllinux_1_2_x86_64.whl", hash = "sha256:938eab7323a736533f015e6069a7d53ef2dcc841e4e533b782c2bfb9fb12d84b"}, - {file = "rpds_py-0.17.1-pp39-pypy39_pp73-macosx_10_12_x86_64.whl", hash = "sha256:1e626b365293a2142a62b9a614e1f8e331b28f3ca57b9f05ebbf4cf2a0f0bdc5"}, - {file = "rpds_py-0.17.1-pp39-pypy39_pp73-macosx_11_0_arm64.whl", hash = "sha256:380e0df2e9d5d5d339803cfc6d183a5442ad7ab3c63c2a0982e8c824566c5ccc"}, - {file = "rpds_py-0.17.1-pp39-pypy39_pp73-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:b760a56e080a826c2e5af09002c1a037382ed21d03134eb6294812dda268c811"}, - {file = "rpds_py-0.17.1-pp39-pypy39_pp73-manylinux_2_17_armv7l.manylinux2014_armv7l.whl", hash = "sha256:5576ee2f3a309d2bb403ec292d5958ce03953b0e57a11d224c1f134feaf8c40f"}, - {file = "rpds_py-0.17.1-pp39-pypy39_pp73-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:1f3c3461ebb4c4f1bbc70b15d20b565759f97a5aaf13af811fcefc892e9197ba"}, - {file = "rpds_py-0.17.1-pp39-pypy39_pp73-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:637b802f3f069a64436d432117a7e58fab414b4e27a7e81049817ae94de45d8d"}, - {file = "rpds_py-0.17.1-pp39-pypy39_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:ffee088ea9b593cc6160518ba9bd319b5475e5f3e578e4552d63818773c6f56a"}, - {file = "rpds_py-0.17.1-pp39-pypy39_pp73-manylinux_2_5_i686.manylinux1_i686.whl", hash = "sha256:3ac732390d529d8469b831949c78085b034bff67f584559340008d0f6041a049"}, - {file = "rpds_py-0.17.1-pp39-pypy39_pp73-musllinux_1_2_aarch64.whl", hash = "sha256:93432e747fb07fa567ad9cc7aaadd6e29710e515aabf939dfbed8046041346c6"}, - {file = "rpds_py-0.17.1-pp39-pypy39_pp73-musllinux_1_2_i686.whl", hash = "sha256:7b7d9ca34542099b4e185b3c2a2b2eda2e318a7dbde0b0d83357a6d4421b5296"}, - {file = "rpds_py-0.17.1-pp39-pypy39_pp73-musllinux_1_2_x86_64.whl", hash = "sha256:0387ce69ba06e43df54e43968090f3626e231e4bc9150e4c3246947567695f68"}, - {file = "rpds_py-0.17.1.tar.gz", hash = "sha256:0210b2668f24c078307260bf88bdac9d6f1093635df5123789bfee4d8d7fc8e7"}, + {file = "rpds_py-0.18.1-cp310-cp310-macosx_10_12_x86_64.whl", hash = "sha256:d31dea506d718693b6b2cffc0648a8929bdc51c70a311b2770f09611caa10d53"}, + {file = "rpds_py-0.18.1-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:732672fbc449bab754e0b15356c077cc31566df874964d4801ab14f71951ea80"}, + {file = "rpds_py-0.18.1-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:4a98a1f0552b5f227a3d6422dbd61bc6f30db170939bd87ed14f3c339aa6c7c9"}, + {file = "rpds_py-0.18.1-cp310-cp310-manylinux_2_17_armv7l.manylinux2014_armv7l.whl", hash = "sha256:7f1944ce16401aad1e3f7d312247b3d5de7981f634dc9dfe90da72b87d37887d"}, + {file = "rpds_py-0.18.1-cp310-cp310-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:38e14fb4e370885c4ecd734f093a2225ee52dc384b86fa55fe3f74638b2cfb09"}, + {file = "rpds_py-0.18.1-cp310-cp310-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:08d74b184f9ab6289b87b19fe6a6d1a97fbfea84b8a3e745e87a5de3029bf944"}, + {file = "rpds_py-0.18.1-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:d70129cef4a8d979caa37e7fe957202e7eee8ea02c5e16455bc9808a59c6b2f0"}, + {file = "rpds_py-0.18.1-cp310-cp310-manylinux_2_5_i686.manylinux1_i686.whl", hash = "sha256:ce0bb20e3a11bd04461324a6a798af34d503f8d6f1aa3d2aa8901ceaf039176d"}, + {file = "rpds_py-0.18.1-cp310-cp310-musllinux_1_2_aarch64.whl", hash = "sha256:81c5196a790032e0fc2464c0b4ab95f8610f96f1f2fa3d4deacce6a79852da60"}, + {file = "rpds_py-0.18.1-cp310-cp310-musllinux_1_2_i686.whl", hash = "sha256:f3027be483868c99b4985fda802a57a67fdf30c5d9a50338d9db646d590198da"}, + {file = "rpds_py-0.18.1-cp310-cp310-musllinux_1_2_x86_64.whl", hash = "sha256:d44607f98caa2961bab4fa3c4309724b185b464cdc3ba6f3d7340bac3ec97cc1"}, + {file = "rpds_py-0.18.1-cp310-none-win32.whl", hash = "sha256:c273e795e7a0f1fddd46e1e3cb8be15634c29ae8ff31c196debb620e1edb9333"}, + {file = "rpds_py-0.18.1-cp310-none-win_amd64.whl", hash = "sha256:8352f48d511de5f973e4f2f9412736d7dea76c69faa6d36bcf885b50c758ab9a"}, + {file = "rpds_py-0.18.1-cp311-cp311-macosx_10_12_x86_64.whl", hash = "sha256:6b5ff7e1d63a8281654b5e2896d7f08799378e594f09cf3674e832ecaf396ce8"}, + {file = "rpds_py-0.18.1-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:8927638a4d4137a289e41d0fd631551e89fa346d6dbcfc31ad627557d03ceb6d"}, + {file = "rpds_py-0.18.1-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:154bf5c93d79558b44e5b50cc354aa0459e518e83677791e6adb0b039b7aa6a7"}, + {file = "rpds_py-0.18.1-cp311-cp311-manylinux_2_17_armv7l.manylinux2014_armv7l.whl", hash = "sha256:07f2139741e5deb2c5154a7b9629bc5aa48c766b643c1a6750d16f865a82c5fc"}, + {file = "rpds_py-0.18.1-cp311-cp311-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:8c7672e9fba7425f79019db9945b16e308ed8bc89348c23d955c8c0540da0a07"}, + {file = "rpds_py-0.18.1-cp311-cp311-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:489bdfe1abd0406eba6b3bb4fdc87c7fa40f1031de073d0cfb744634cc8fa261"}, + {file = "rpds_py-0.18.1-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:3c20f05e8e3d4fc76875fc9cb8cf24b90a63f5a1b4c5b9273f0e8225e169b100"}, + {file = "rpds_py-0.18.1-cp311-cp311-manylinux_2_5_i686.manylinux1_i686.whl", hash = "sha256:967342e045564cef76dfcf1edb700b1e20838d83b1aa02ab313e6a497cf923b8"}, + {file = "rpds_py-0.18.1-cp311-cp311-musllinux_1_2_aarch64.whl", hash = "sha256:2cc7c1a47f3a63282ab0f422d90ddac4aa3034e39fc66a559ab93041e6505da7"}, + {file = "rpds_py-0.18.1-cp311-cp311-musllinux_1_2_i686.whl", hash = "sha256:f7afbfee1157e0f9376c00bb232e80a60e59ed716e3211a80cb8506550671e6e"}, + {file = "rpds_py-0.18.1-cp311-cp311-musllinux_1_2_x86_64.whl", hash = "sha256:9e6934d70dc50f9f8ea47081ceafdec09245fd9f6032669c3b45705dea096b88"}, + {file = "rpds_py-0.18.1-cp311-none-win32.whl", hash = "sha256:c69882964516dc143083d3795cb508e806b09fc3800fd0d4cddc1df6c36e76bb"}, + {file = "rpds_py-0.18.1-cp311-none-win_amd64.whl", hash = "sha256:70a838f7754483bcdc830444952fd89645569e7452e3226de4a613a4c1793fb2"}, + {file = "rpds_py-0.18.1-cp312-cp312-macosx_10_12_x86_64.whl", hash = "sha256:3dd3cd86e1db5aadd334e011eba4e29d37a104b403e8ca24dcd6703c68ca55b3"}, + {file = "rpds_py-0.18.1-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:05f3d615099bd9b13ecf2fc9cf2d839ad3f20239c678f461c753e93755d629ee"}, + {file = "rpds_py-0.18.1-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:35b2b771b13eee8729a5049c976197ff58a27a3829c018a04341bcf1ae409b2b"}, + {file = "rpds_py-0.18.1-cp312-cp312-manylinux_2_17_armv7l.manylinux2014_armv7l.whl", hash = "sha256:ee17cd26b97d537af8f33635ef38be873073d516fd425e80559f4585a7b90c43"}, + {file = "rpds_py-0.18.1-cp312-cp312-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:b646bf655b135ccf4522ed43d6902af37d3f5dbcf0da66c769a2b3938b9d8184"}, + {file = "rpds_py-0.18.1-cp312-cp312-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:19ba472b9606c36716062c023afa2484d1e4220548751bda14f725a7de17b4f6"}, + {file = "rpds_py-0.18.1-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:6e30ac5e329098903262dc5bdd7e2086e0256aa762cc8b744f9e7bf2a427d3f8"}, + {file = "rpds_py-0.18.1-cp312-cp312-manylinux_2_5_i686.manylinux1_i686.whl", hash = "sha256:d58ad6317d188c43750cb76e9deacf6051d0f884d87dc6518e0280438648a9ac"}, + {file = "rpds_py-0.18.1-cp312-cp312-musllinux_1_2_aarch64.whl", hash = "sha256:e1735502458621921cee039c47318cb90b51d532c2766593be6207eec53e5c4c"}, + {file = "rpds_py-0.18.1-cp312-cp312-musllinux_1_2_i686.whl", hash = "sha256:f5bab211605d91db0e2995a17b5c6ee5edec1270e46223e513eaa20da20076ac"}, + {file = "rpds_py-0.18.1-cp312-cp312-musllinux_1_2_x86_64.whl", hash = "sha256:2fc24a329a717f9e2448f8cd1f960f9dac4e45b6224d60734edeb67499bab03a"}, + {file = "rpds_py-0.18.1-cp312-none-win32.whl", hash = "sha256:1805d5901779662d599d0e2e4159d8a82c0b05faa86ef9222bf974572286b2b6"}, + {file = "rpds_py-0.18.1-cp312-none-win_amd64.whl", hash = "sha256:720edcb916df872d80f80a1cc5ea9058300b97721efda8651efcd938a9c70a72"}, + {file = "rpds_py-0.18.1-cp38-cp38-macosx_10_12_x86_64.whl", hash = "sha256:c827576e2fa017a081346dce87d532a5310241648eb3700af9a571a6e9fc7e74"}, + {file = "rpds_py-0.18.1-cp38-cp38-macosx_11_0_arm64.whl", hash = "sha256:aa3679e751408d75a0b4d8d26d6647b6d9326f5e35c00a7ccd82b78ef64f65f8"}, + {file = "rpds_py-0.18.1-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:0abeee75434e2ee2d142d650d1e54ac1f8b01e6e6abdde8ffd6eeac6e9c38e20"}, + {file = "rpds_py-0.18.1-cp38-cp38-manylinux_2_17_armv7l.manylinux2014_armv7l.whl", hash = "sha256:ed402d6153c5d519a0faf1bb69898e97fb31613b49da27a84a13935ea9164dfc"}, + {file = "rpds_py-0.18.1-cp38-cp38-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:338dee44b0cef8b70fd2ef54b4e09bb1b97fc6c3a58fea5db6cc083fd9fc2724"}, + {file = "rpds_py-0.18.1-cp38-cp38-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:7750569d9526199c5b97e5a9f8d96a13300950d910cf04a861d96f4273d5b104"}, + {file = "rpds_py-0.18.1-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:607345bd5912aacc0c5a63d45a1f73fef29e697884f7e861094e443187c02be5"}, + {file = "rpds_py-0.18.1-cp38-cp38-manylinux_2_5_i686.manylinux1_i686.whl", hash = "sha256:207c82978115baa1fd8d706d720b4a4d2b0913df1c78c85ba73fe6c5804505f0"}, + {file = "rpds_py-0.18.1-cp38-cp38-musllinux_1_2_aarch64.whl", hash = "sha256:6d1e42d2735d437e7e80bab4d78eb2e459af48c0a46e686ea35f690b93db792d"}, + {file = "rpds_py-0.18.1-cp38-cp38-musllinux_1_2_i686.whl", hash = "sha256:5463c47c08630007dc0fe99fb480ea4f34a89712410592380425a9b4e1611d8e"}, + {file = "rpds_py-0.18.1-cp38-cp38-musllinux_1_2_x86_64.whl", hash = "sha256:06d218939e1bf2ca50e6b0ec700ffe755e5216a8230ab3e87c059ebb4ea06afc"}, + {file = "rpds_py-0.18.1-cp38-none-win32.whl", hash = "sha256:312fe69b4fe1ffbe76520a7676b1e5ac06ddf7826d764cc10265c3b53f96dbe9"}, + {file = "rpds_py-0.18.1-cp38-none-win_amd64.whl", hash = "sha256:9437ca26784120a279f3137ee080b0e717012c42921eb07861b412340f85bae2"}, + {file = "rpds_py-0.18.1-cp39-cp39-macosx_10_12_x86_64.whl", hash = "sha256:19e515b78c3fc1039dd7da0a33c28c3154458f947f4dc198d3c72db2b6b5dc93"}, + {file = "rpds_py-0.18.1-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:a7b28c5b066bca9a4eb4e2f2663012debe680f097979d880657f00e1c30875a0"}, + {file = "rpds_py-0.18.1-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:673fdbbf668dd958eff750e500495ef3f611e2ecc209464f661bc82e9838991e"}, + {file = "rpds_py-0.18.1-cp39-cp39-manylinux_2_17_armv7l.manylinux2014_armv7l.whl", hash = "sha256:d960de62227635d2e61068f42a6cb6aae91a7fe00fca0e3aeed17667c8a34611"}, + {file = "rpds_py-0.18.1-cp39-cp39-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:352a88dc7892f1da66b6027af06a2e7e5d53fe05924cc2cfc56495b586a10b72"}, + {file = "rpds_py-0.18.1-cp39-cp39-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:4e0ee01ad8260184db21468a6e1c37afa0529acc12c3a697ee498d3c2c4dcaf3"}, + {file = "rpds_py-0.18.1-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:e4c39ad2f512b4041343ea3c7894339e4ca7839ac38ca83d68a832fc8b3748ab"}, + {file = "rpds_py-0.18.1-cp39-cp39-manylinux_2_5_i686.manylinux1_i686.whl", hash = "sha256:aaa71ee43a703c321906813bb252f69524f02aa05bf4eec85f0c41d5d62d0f4c"}, + {file = "rpds_py-0.18.1-cp39-cp39-musllinux_1_2_aarch64.whl", hash = "sha256:6cd8098517c64a85e790657e7b1e509b9fe07487fd358e19431cb120f7d96338"}, + {file = "rpds_py-0.18.1-cp39-cp39-musllinux_1_2_i686.whl", hash = "sha256:4adec039b8e2928983f885c53b7cc4cda8965b62b6596501a0308d2703f8af1b"}, + {file = "rpds_py-0.18.1-cp39-cp39-musllinux_1_2_x86_64.whl", hash = "sha256:32b7daaa3e9389db3695964ce8e566e3413b0c43e3394c05e4b243a4cd7bef26"}, + {file = "rpds_py-0.18.1-cp39-none-win32.whl", hash = "sha256:2625f03b105328729f9450c8badda34d5243231eef6535f80064d57035738360"}, + {file = "rpds_py-0.18.1-cp39-none-win_amd64.whl", hash = "sha256:bf18932d0003c8c4d51a39f244231986ab23ee057d235a12b2684ea26a353590"}, + {file = "rpds_py-0.18.1-pp310-pypy310_pp73-macosx_10_12_x86_64.whl", hash = "sha256:cbfbea39ba64f5e53ae2915de36f130588bba71245b418060ec3330ebf85678e"}, + {file = "rpds_py-0.18.1-pp310-pypy310_pp73-macosx_11_0_arm64.whl", hash = "sha256:a3d456ff2a6a4d2adcdf3c1c960a36f4fd2fec6e3b4902a42a384d17cf4e7a65"}, + {file = "rpds_py-0.18.1-pp310-pypy310_pp73-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:7700936ef9d006b7ef605dc53aa364da2de5a3aa65516a1f3ce73bf82ecfc7ae"}, + {file = "rpds_py-0.18.1-pp310-pypy310_pp73-manylinux_2_17_armv7l.manylinux2014_armv7l.whl", hash = "sha256:51584acc5916212e1bf45edd17f3a6b05fe0cbb40482d25e619f824dccb679de"}, + {file = "rpds_py-0.18.1-pp310-pypy310_pp73-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:942695a206a58d2575033ff1e42b12b2aece98d6003c6bc739fbf33d1773b12f"}, + {file = "rpds_py-0.18.1-pp310-pypy310_pp73-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:b906b5f58892813e5ba5c6056d6a5ad08f358ba49f046d910ad992196ea61397"}, + {file = "rpds_py-0.18.1-pp310-pypy310_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:f6f8e3fecca256fefc91bb6765a693d96692459d7d4c644660a9fff32e517843"}, + {file = "rpds_py-0.18.1-pp310-pypy310_pp73-manylinux_2_5_i686.manylinux1_i686.whl", hash = "sha256:7732770412bab81c5a9f6d20aeb60ae943a9b36dcd990d876a773526468e7163"}, + {file = "rpds_py-0.18.1-pp310-pypy310_pp73-musllinux_1_2_aarch64.whl", hash = "sha256:bd1105b50ede37461c1d51b9698c4f4be6e13e69a908ab7751e3807985fc0346"}, + {file = "rpds_py-0.18.1-pp310-pypy310_pp73-musllinux_1_2_i686.whl", hash = "sha256:618916f5535784960f3ecf8111581f4ad31d347c3de66d02e728de460a46303c"}, + {file = "rpds_py-0.18.1-pp310-pypy310_pp73-musllinux_1_2_x86_64.whl", hash = "sha256:17c6d2155e2423f7e79e3bb18151c686d40db42d8645e7977442170c360194d4"}, + {file = "rpds_py-0.18.1-pp38-pypy38_pp73-macosx_10_12_x86_64.whl", hash = "sha256:6c4c4c3f878df21faf5fac86eda32671c27889e13570645a9eea0a1abdd50922"}, + {file = "rpds_py-0.18.1-pp38-pypy38_pp73-macosx_11_0_arm64.whl", hash = "sha256:fab6ce90574645a0d6c58890e9bcaac8d94dff54fb51c69e5522a7358b80ab64"}, + {file = "rpds_py-0.18.1-pp38-pypy38_pp73-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:531796fb842b53f2695e94dc338929e9f9dbf473b64710c28af5a160b2a8927d"}, + {file = "rpds_py-0.18.1-pp38-pypy38_pp73-manylinux_2_17_armv7l.manylinux2014_armv7l.whl", hash = "sha256:740884bc62a5e2bbb31e584f5d23b32320fd75d79f916f15a788d527a5e83644"}, + {file = "rpds_py-0.18.1-pp38-pypy38_pp73-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:998125738de0158f088aef3cb264a34251908dd2e5d9966774fdab7402edfab7"}, + {file = "rpds_py-0.18.1-pp38-pypy38_pp73-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:e2be6e9dd4111d5b31ba3b74d17da54a8319d8168890fbaea4b9e5c3de630ae5"}, + {file = "rpds_py-0.18.1-pp38-pypy38_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:d0cee71bc618cd93716f3c1bf56653740d2d13ddbd47673efa8bf41435a60daa"}, + {file = "rpds_py-0.18.1-pp38-pypy38_pp73-manylinux_2_5_i686.manylinux1_i686.whl", hash = "sha256:2c3caec4ec5cd1d18e5dd6ae5194d24ed12785212a90b37f5f7f06b8bedd7139"}, + {file = "rpds_py-0.18.1-pp38-pypy38_pp73-musllinux_1_2_aarch64.whl", hash = "sha256:27bba383e8c5231cd559affe169ca0b96ec78d39909ffd817f28b166d7ddd4d8"}, + {file = "rpds_py-0.18.1-pp38-pypy38_pp73-musllinux_1_2_i686.whl", hash = "sha256:a888e8bdb45916234b99da2d859566f1e8a1d2275a801bb8e4a9644e3c7e7909"}, + {file = "rpds_py-0.18.1-pp38-pypy38_pp73-musllinux_1_2_x86_64.whl", hash = "sha256:6031b25fb1b06327b43d841f33842b383beba399884f8228a6bb3df3088485ff"}, + {file = "rpds_py-0.18.1-pp39-pypy39_pp73-macosx_10_12_x86_64.whl", hash = "sha256:48c2faaa8adfacefcbfdb5f2e2e7bdad081e5ace8d182e5f4ade971f128e6bb3"}, + {file = "rpds_py-0.18.1-pp39-pypy39_pp73-macosx_11_0_arm64.whl", hash = "sha256:d85164315bd68c0806768dc6bb0429c6f95c354f87485ee3593c4f6b14def2bd"}, + {file = "rpds_py-0.18.1-pp39-pypy39_pp73-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:6afd80f6c79893cfc0574956f78a0add8c76e3696f2d6a15bca2c66c415cf2d4"}, + {file = "rpds_py-0.18.1-pp39-pypy39_pp73-manylinux_2_17_armv7l.manylinux2014_armv7l.whl", hash = "sha256:fa242ac1ff583e4ec7771141606aafc92b361cd90a05c30d93e343a0c2d82a89"}, + {file = "rpds_py-0.18.1-pp39-pypy39_pp73-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:d21be4770ff4e08698e1e8e0bce06edb6ea0626e7c8f560bc08222880aca6a6f"}, + {file = "rpds_py-0.18.1-pp39-pypy39_pp73-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:5c45a639e93a0c5d4b788b2613bd637468edd62f8f95ebc6fcc303d58ab3f0a8"}, + {file = "rpds_py-0.18.1-pp39-pypy39_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:910e71711d1055b2768181efa0a17537b2622afeb0424116619817007f8a2b10"}, + {file = "rpds_py-0.18.1-pp39-pypy39_pp73-manylinux_2_5_i686.manylinux1_i686.whl", hash = "sha256:b9bb1f182a97880f6078283b3505a707057c42bf55d8fca604f70dedfdc0772a"}, + {file = "rpds_py-0.18.1-pp39-pypy39_pp73-musllinux_1_2_aarch64.whl", hash = "sha256:1d54f74f40b1f7aaa595a02ff42ef38ca654b1469bef7d52867da474243cc633"}, + {file = "rpds_py-0.18.1-pp39-pypy39_pp73-musllinux_1_2_i686.whl", hash = "sha256:8d2e182c9ee01135e11e9676e9a62dfad791a7a467738f06726872374a83db49"}, + {file = "rpds_py-0.18.1-pp39-pypy39_pp73-musllinux_1_2_x86_64.whl", hash = "sha256:636a15acc588f70fda1661234761f9ed9ad79ebed3f2125d44be0862708b666e"}, + {file = "rpds_py-0.18.1.tar.gz", hash = "sha256:dc48b479d540770c811fbd1eb9ba2bb66951863e448efec2e2c102625328e92f"}, ] [[package]] name = "scipy" -version = "1.12.0" +version = "1.13.1" description = "Fundamental algorithms for scientific computing in Python" optional = false python-versions = ">=3.9" files = [ - {file = "scipy-1.12.0-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:78e4402e140879387187f7f25d91cc592b3501a2e51dfb320f48dfb73565f10b"}, - {file = "scipy-1.12.0-cp310-cp310-macosx_12_0_arm64.whl", hash = "sha256:f5f00ebaf8de24d14b8449981a2842d404152774c1a1d880c901bf454cb8e2a1"}, - {file = "scipy-1.12.0-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:e53958531a7c695ff66c2e7bb7b79560ffdc562e2051644c5576c39ff8efb563"}, - {file = "scipy-1.12.0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:5e32847e08da8d895ce09d108a494d9eb78974cf6de23063f93306a3e419960c"}, - {file = "scipy-1.12.0-cp310-cp310-musllinux_1_1_x86_64.whl", hash = "sha256:4c1020cad92772bf44b8e4cdabc1df5d87376cb219742549ef69fc9fd86282dd"}, - {file = "scipy-1.12.0-cp310-cp310-win_amd64.whl", hash = "sha256:75ea2a144096b5e39402e2ff53a36fecfd3b960d786b7efd3c180e29c39e53f2"}, - {file = "scipy-1.12.0-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:408c68423f9de16cb9e602528be4ce0d6312b05001f3de61fe9ec8b1263cad08"}, - {file = "scipy-1.12.0-cp311-cp311-macosx_12_0_arm64.whl", hash = "sha256:5adfad5dbf0163397beb4aca679187d24aec085343755fcdbdeb32b3679f254c"}, - {file = "scipy-1.12.0-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:c3003652496f6e7c387b1cf63f4bb720951cfa18907e998ea551e6de51a04467"}, - {file = "scipy-1.12.0-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:8b8066bce124ee5531d12a74b617d9ac0ea59245246410e19bca549656d9a40a"}, - {file = "scipy-1.12.0-cp311-cp311-musllinux_1_1_x86_64.whl", hash = "sha256:8bee4993817e204d761dba10dbab0774ba5a8612e57e81319ea04d84945375ba"}, - {file = "scipy-1.12.0-cp311-cp311-win_amd64.whl", hash = "sha256:a24024d45ce9a675c1fb8494e8e5244efea1c7a09c60beb1eeb80373d0fecc70"}, - {file = "scipy-1.12.0-cp312-cp312-macosx_10_9_x86_64.whl", hash = "sha256:e7e76cc48638228212c747ada851ef355c2bb5e7f939e10952bc504c11f4e372"}, - {file = "scipy-1.12.0-cp312-cp312-macosx_12_0_arm64.whl", hash = "sha256:f7ce148dffcd64ade37b2df9315541f9adad6efcaa86866ee7dd5db0c8f041c3"}, - {file = "scipy-1.12.0-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:9c39f92041f490422924dfdb782527a4abddf4707616e07b021de33467f917bc"}, - {file = "scipy-1.12.0-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:a7ebda398f86e56178c2fa94cad15bf457a218a54a35c2a7b4490b9f9cb2676c"}, - {file = "scipy-1.12.0-cp312-cp312-musllinux_1_1_x86_64.whl", hash = "sha256:95e5c750d55cf518c398a8240571b0e0782c2d5a703250872f36eaf737751338"}, - {file = "scipy-1.12.0-cp312-cp312-win_amd64.whl", hash = "sha256:e646d8571804a304e1da01040d21577685ce8e2db08ac58e543eaca063453e1c"}, - {file = "scipy-1.12.0-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:913d6e7956c3a671de3b05ccb66b11bc293f56bfdef040583a7221d9e22a2e35"}, - {file = "scipy-1.12.0-cp39-cp39-macosx_12_0_arm64.whl", hash = "sha256:bba1b0c7256ad75401c73e4b3cf09d1f176e9bd4248f0d3112170fb2ec4db067"}, - {file = "scipy-1.12.0-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:730badef9b827b368f351eacae2e82da414e13cf8bd5051b4bdfd720271a5371"}, - {file = "scipy-1.12.0-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:6546dc2c11a9df6926afcbdd8a3edec28566e4e785b915e849348c6dd9f3f490"}, - {file = "scipy-1.12.0-cp39-cp39-musllinux_1_1_x86_64.whl", hash = "sha256:196ebad3a4882081f62a5bf4aeb7326aa34b110e533aab23e4374fcccb0890dc"}, - {file = "scipy-1.12.0-cp39-cp39-win_amd64.whl", hash = "sha256:b360f1b6b2f742781299514e99ff560d1fe9bd1bff2712894b52abe528d1fd1e"}, - {file = "scipy-1.12.0.tar.gz", hash = "sha256:4bf5abab8a36d20193c698b0f1fc282c1d083c94723902c447e5d2f1780936a3"}, + {file = "scipy-1.13.1-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:20335853b85e9a49ff7572ab453794298bcf0354d8068c5f6775a0eabf350aca"}, + {file = "scipy-1.13.1-cp310-cp310-macosx_12_0_arm64.whl", hash = "sha256:d605e9c23906d1994f55ace80e0125c587f96c020037ea6aa98d01b4bd2e222f"}, + {file = "scipy-1.13.1-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:cfa31f1def5c819b19ecc3a8b52d28ffdcc7ed52bb20c9a7589669dd3c250989"}, + {file = "scipy-1.13.1-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:f26264b282b9da0952a024ae34710c2aff7d27480ee91a2e82b7b7073c24722f"}, + {file = "scipy-1.13.1-cp310-cp310-musllinux_1_1_x86_64.whl", hash = "sha256:eccfa1906eacc02de42d70ef4aecea45415f5be17e72b61bafcfd329bdc52e94"}, + {file = "scipy-1.13.1-cp310-cp310-win_amd64.whl", hash = "sha256:2831f0dc9c5ea9edd6e51e6e769b655f08ec6db6e2e10f86ef39bd32eb11da54"}, + {file = "scipy-1.13.1-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:27e52b09c0d3a1d5b63e1105f24177e544a222b43611aaf5bc44d4a0979e32f9"}, + {file = "scipy-1.13.1-cp311-cp311-macosx_12_0_arm64.whl", hash = "sha256:54f430b00f0133e2224c3ba42b805bfd0086fe488835effa33fa291561932326"}, + {file = "scipy-1.13.1-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:e89369d27f9e7b0884ae559a3a956e77c02114cc60a6058b4e5011572eea9299"}, + {file = "scipy-1.13.1-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:a78b4b3345f1b6f68a763c6e25c0c9a23a9fd0f39f5f3d200efe8feda560a5fa"}, + {file = "scipy-1.13.1-cp311-cp311-musllinux_1_1_x86_64.whl", hash = "sha256:45484bee6d65633752c490404513b9ef02475b4284c4cfab0ef946def50b3f59"}, + {file = "scipy-1.13.1-cp311-cp311-win_amd64.whl", hash = "sha256:5713f62f781eebd8d597eb3f88b8bf9274e79eeabf63afb4a737abc6c84ad37b"}, + {file = "scipy-1.13.1-cp312-cp312-macosx_10_9_x86_64.whl", hash = "sha256:5d72782f39716b2b3509cd7c33cdc08c96f2f4d2b06d51e52fb45a19ca0c86a1"}, + {file = "scipy-1.13.1-cp312-cp312-macosx_12_0_arm64.whl", hash = "sha256:017367484ce5498445aade74b1d5ab377acdc65e27095155e448c88497755a5d"}, + {file = "scipy-1.13.1-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:949ae67db5fa78a86e8fa644b9a6b07252f449dcf74247108c50e1d20d2b4627"}, + {file = "scipy-1.13.1-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:de3ade0e53bc1f21358aa74ff4830235d716211d7d077e340c7349bc3542e884"}, + {file = "scipy-1.13.1-cp312-cp312-musllinux_1_1_x86_64.whl", hash = "sha256:2ac65fb503dad64218c228e2dc2d0a0193f7904747db43014645ae139c8fad16"}, + {file = "scipy-1.13.1-cp312-cp312-win_amd64.whl", hash = "sha256:cdd7dacfb95fea358916410ec61bbc20440f7860333aee6d882bb8046264e949"}, + {file = "scipy-1.13.1-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:436bbb42a94a8aeef855d755ce5a465479c721e9d684de76bf61a62e7c2b81d5"}, + {file = "scipy-1.13.1-cp39-cp39-macosx_12_0_arm64.whl", hash = "sha256:8335549ebbca860c52bf3d02f80784e91a004b71b059e3eea9678ba994796a24"}, + {file = "scipy-1.13.1-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:d533654b7d221a6a97304ab63c41c96473ff04459e404b83275b60aa8f4b7004"}, + {file = "scipy-1.13.1-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:637e98dcf185ba7f8e663e122ebf908c4702420477ae52a04f9908707456ba4d"}, + {file = "scipy-1.13.1-cp39-cp39-musllinux_1_1_x86_64.whl", hash = "sha256:a014c2b3697bde71724244f63de2476925596c24285c7a637364761f8710891c"}, + {file = "scipy-1.13.1-cp39-cp39-win_amd64.whl", hash = "sha256:392e4ec766654852c25ebad4f64e4e584cf19820b980bc04960bca0b0cd6eaa2"}, + {file = "scipy-1.13.1.tar.gz", hash = "sha256:095a87a0312b08dfd6a6155cbbd310a8c51800fc931b8c0b84003014b874ed3c"}, ] [package.dependencies] -numpy = ">=1.22.4,<1.29.0" +numpy = ">=1.22.4,<2.3" [package.extras] -dev = ["click", "cython-lint (>=0.12.2)", "doit (>=0.36.0)", "mypy", "pycodestyle", "pydevtool", "rich-click", "ruff", "types-psutil", "typing_extensions"] -doc = ["jupytext", "matplotlib (>2)", "myst-nb", "numpydoc", "pooch", "pydata-sphinx-theme (==0.9.0)", "sphinx (!=4.1.0)", "sphinx-design (>=0.2.0)"] -test = ["asv", "gmpy2", "hypothesis", "mpmath", "pooch", "pytest", "pytest-cov", "pytest-timeout", "pytest-xdist", "scikit-umfpack", "threadpoolctl"] - -[[package]] -name = "setuptools" -version = "69.0.3" -description = "Easily download, build, install, upgrade, and uninstall Python packages" -optional = false -python-versions = ">=3.8" -files = [ - {file = "setuptools-69.0.3-py3-none-any.whl", hash = "sha256:385eb4edd9c9d5c17540511303e39a147ce2fc04bc55289c322b9e5904fe2c05"}, - {file = "setuptools-69.0.3.tar.gz", hash = "sha256:be1af57fc409f93647f2e8e4573a142ed38724b8cdd389706a867bb4efcf1e78"}, -] - -[package.extras] -docs = ["furo", "jaraco.packaging (>=9.3)", "jaraco.tidelift (>=1.4)", "pygments-github-lexers (==0.0.5)", "rst.linker (>=1.9)", "sphinx (<7.2.5)", "sphinx (>=3.5)", "sphinx-favicon", "sphinx-inline-tabs", "sphinx-lint", "sphinx-notfound-page (>=1,<2)", "sphinx-reredirects", "sphinxcontrib-towncrier"] -testing = ["build[virtualenv]", "filelock (>=3.4.0)", "flake8-2020", "ini2toml[lite] (>=0.9)", "jaraco.develop (>=7.21)", "jaraco.envs (>=2.2)", "jaraco.path (>=3.2.0)", "pip (>=19.1)", "pytest (>=6)", "pytest-black (>=0.3.7)", "pytest-checkdocs (>=2.4)", "pytest-cov", "pytest-enabler (>=2.2)", "pytest-mypy (>=0.9.1)", "pytest-perf", "pytest-ruff", "pytest-timeout", "pytest-xdist", "tomli-w (>=1.0.0)", "virtualenv (>=13.0.0)", "wheel"] -testing-integration = ["build[virtualenv] (>=1.0.3)", "filelock (>=3.4.0)", "jaraco.envs (>=2.2)", "jaraco.path (>=3.2.0)", "packaging (>=23.1)", "pytest", "pytest-enabler", "pytest-xdist", "tomli", "virtualenv (>=13.0.0)", "wheel"] +dev = ["cython-lint (>=0.12.2)", "doit (>=0.36.0)", "mypy", "pycodestyle", "pydevtool", "rich-click", "ruff", "types-psutil", "typing_extensions"] +doc = ["jupyterlite-pyodide-kernel", "jupyterlite-sphinx (>=0.12.0)", "jupytext", "matplotlib (>=3.5)", "myst-nb", "numpydoc", "pooch", "pydata-sphinx-theme (>=0.15.2)", "sphinx (>=5.0.0)", "sphinx-design (>=0.4.0)"] +test = ["array-api-strict", "asv", "gmpy2", "hypothesis (>=6.30)", "mpmath", "pooch", "pytest", "pytest-cov", "pytest-timeout", "pytest-xdist", "scikit-umfpack", "threadpoolctl"] [[package]] name = "shapely" -version = "2.0.2" +version = "2.0.4" description = "Manipulation and analysis of geometric objects" optional = false python-versions = ">=3.7" files = [ - {file = "shapely-2.0.2-cp310-cp310-macosx_10_9_universal2.whl", hash = "sha256:6ca8cffbe84ddde8f52b297b53f8e0687bd31141abb2c373fd8a9f032df415d6"}, - {file = "shapely-2.0.2-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:baa14fc27771e180c06b499a0a7ba697c7988c7b2b6cba9a929a19a4d2762de3"}, - {file = "shapely-2.0.2-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:36480e32c434d168cdf2f5e9862c84aaf4d714a43a8465ae3ce8ff327f0affb7"}, - {file = "shapely-2.0.2-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:4ef753200cbffd4f652efb2c528c5474e5a14341a473994d90ad0606522a46a2"}, - {file = "shapely-2.0.2-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:a9a41ff4323fc9d6257759c26eb1cf3a61ebc7e611e024e6091f42977303fd3a"}, - {file = "shapely-2.0.2-cp310-cp310-win32.whl", hash = "sha256:72b5997272ae8c25f0fd5b3b967b3237e87fab7978b8d6cd5fa748770f0c5d68"}, - {file = "shapely-2.0.2-cp310-cp310-win_amd64.whl", hash = "sha256:34eac2337cbd67650248761b140d2535855d21b969d76d76123317882d3a0c1a"}, - {file = "shapely-2.0.2-cp311-cp311-macosx_10_9_universal2.whl", hash = "sha256:5b0c052709c8a257c93b0d4943b0b7a3035f87e2d6a8ac9407b6a992d206422f"}, - {file = "shapely-2.0.2-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:2d217e56ae067e87b4e1731d0dc62eebe887ced729ba5c2d4590e9e3e9fdbd88"}, - {file = "shapely-2.0.2-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:94ac128ae2ab4edd0bffcd4e566411ea7bdc738aeaf92c32a8a836abad725f9f"}, - {file = "shapely-2.0.2-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:fa3ee28f5e63a130ec5af4dc3c4cb9c21c5788bb13c15e89190d163b14f9fb89"}, - {file = "shapely-2.0.2-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:737dba15011e5a9b54a8302f1748b62daa207c9bc06f820cd0ad32a041f1c6f2"}, - {file = "shapely-2.0.2-cp311-cp311-win32.whl", hash = "sha256:45ac6906cff0765455a7b49c1670af6e230c419507c13e2f75db638c8fc6f3bd"}, - {file = "shapely-2.0.2-cp311-cp311-win_amd64.whl", hash = "sha256:dc9342fc82e374130db86a955c3c4525bfbf315a248af8277a913f30911bed9e"}, - {file = "shapely-2.0.2-cp312-cp312-macosx_10_9_universal2.whl", hash = "sha256:06f193091a7c6112fc08dfd195a1e3846a64306f890b151fa8c63b3e3624202c"}, - {file = "shapely-2.0.2-cp312-cp312-macosx_10_9_x86_64.whl", hash = "sha256:eebe544df5c018134f3c23b6515877f7e4cd72851f88a8d0c18464f414d141a2"}, - {file = "shapely-2.0.2-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:7e92e7c255f89f5cdf777690313311f422aa8ada9a3205b187113274e0135cd8"}, - {file = "shapely-2.0.2-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:be46d5509b9251dd9087768eaf35a71360de6afac82ce87c636990a0871aa18b"}, - {file = "shapely-2.0.2-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:a5533a925d8e211d07636ffc2fdd9a7f9f13d54686d00577eeb11d16f00be9c4"}, - {file = "shapely-2.0.2-cp312-cp312-win32.whl", hash = "sha256:084b023dae8ad3d5b98acee9d3bf098fdf688eb0bb9b1401e8b075f6a627b611"}, - {file = "shapely-2.0.2-cp312-cp312-win_amd64.whl", hash = "sha256:ea84d1cdbcf31e619d672b53c4532f06253894185ee7acb8ceb78f5f33cbe033"}, - {file = "shapely-2.0.2-cp37-cp37m-macosx_10_9_x86_64.whl", hash = "sha256:ed1e99702125e7baccf401830a3b94d810d5c70b329b765fe93451fe14cf565b"}, - {file = "shapely-2.0.2-cp37-cp37m-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:e7d897e6bdc6bc64f7f65155dbbb30e49acaabbd0d9266b9b4041f87d6e52b3a"}, - {file = "shapely-2.0.2-cp37-cp37m-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:0521d76d1e8af01e712db71da9096b484f081e539d4f4a8c97342e7971d5e1b4"}, - {file = "shapely-2.0.2-cp37-cp37m-win32.whl", hash = "sha256:5324be299d4c533ecfcfd43424dfd12f9428fd6f12cda38a4316da001d6ef0ea"}, - {file = "shapely-2.0.2-cp37-cp37m-win_amd64.whl", hash = "sha256:78128357a0cee573257a0c2c388d4b7bf13cb7dbe5b3fe5d26d45ebbe2a39e25"}, - {file = "shapely-2.0.2-cp38-cp38-macosx_10_9_universal2.whl", hash = "sha256:87dc2be34ac3a3a4a319b963c507ac06682978a5e6c93d71917618b14f13066e"}, - {file = "shapely-2.0.2-cp38-cp38-macosx_10_9_x86_64.whl", hash = "sha256:42997ac806e4583dad51c80a32d38570fd9a3d4778f5e2c98f9090aa7db0fe91"}, - {file = "shapely-2.0.2-cp38-cp38-macosx_11_0_arm64.whl", hash = "sha256:ccfd5fa10a37e67dbafc601c1ddbcbbfef70d34c3f6b0efc866ddbdb55893a6c"}, - {file = "shapely-2.0.2-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:e7c95d3379ae3abb74058938a9fcbc478c6b2e28d20dace38f8b5c587dde90aa"}, - {file = "shapely-2.0.2-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:6a21353d28209fb0d8cc083e08ca53c52666e0d8a1f9bbe23b6063967d89ed24"}, - {file = "shapely-2.0.2-cp38-cp38-win32.whl", hash = "sha256:03e63a99dfe6bd3beb8d5f41ec2086585bb969991d603f9aeac335ad396a06d4"}, - {file = "shapely-2.0.2-cp38-cp38-win_amd64.whl", hash = "sha256:c6fd29fbd9cd76350bd5cc14c49de394a31770aed02d74203e23b928f3d2f1aa"}, - {file = "shapely-2.0.2-cp39-cp39-macosx_10_9_universal2.whl", hash = "sha256:1f217d28ecb48e593beae20a0082a95bd9898d82d14b8fcb497edf6bff9a44d7"}, - {file = "shapely-2.0.2-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:394e5085b49334fd5b94fa89c086edfb39c3ecab7f669e8b2a4298b9d523b3a5"}, - {file = "shapely-2.0.2-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:fd3ad17b64466a033848c26cb5b509625c87d07dcf39a1541461cacdb8f7e91c"}, - {file = "shapely-2.0.2-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:d41a116fcad58048d7143ddb01285e1a8780df6dc1f56c3b1e1b7f12ed296651"}, - {file = "shapely-2.0.2-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:dea9a0651333cf96ef5bb2035044e3ad6a54f87d90e50fe4c2636debf1b77abc"}, - {file = "shapely-2.0.2-cp39-cp39-win32.whl", hash = "sha256:b8eb0a92f7b8c74f9d8fdd1b40d395113f59bd8132ca1348ebcc1f5aece94b96"}, - {file = "shapely-2.0.2-cp39-cp39-win_amd64.whl", hash = "sha256:794affd80ca0f2c536fc948a3afa90bd8fb61ebe37fe873483ae818e7f21def4"}, - {file = "shapely-2.0.2.tar.gz", hash = "sha256:1713cc04c171baffc5b259ba8531c58acc2a301707b7f021d88a15ed090649e7"}, + {file = "shapely-2.0.4-cp310-cp310-macosx_10_9_universal2.whl", hash = "sha256:011b77153906030b795791f2fdfa2d68f1a8d7e40bce78b029782ade3afe4f2f"}, + {file = "shapely-2.0.4-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:9831816a5d34d5170aa9ed32a64982c3d6f4332e7ecfe62dc97767e163cb0b17"}, + {file = "shapely-2.0.4-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:5c4849916f71dc44e19ed370421518c0d86cf73b26e8656192fcfcda08218fbd"}, + {file = "shapely-2.0.4-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:841f93a0e31e4c64d62ea570d81c35de0f6cea224568b2430d832967536308e6"}, + {file = "shapely-2.0.4-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:d2b4431f522b277c79c34b65da128029a9955e4481462cbf7ebec23aab61fc58"}, + {file = "shapely-2.0.4-cp310-cp310-win32.whl", hash = "sha256:92a41d936f7d6743f343be265ace93b7c57f5b231e21b9605716f5a47c2879e7"}, + {file = "shapely-2.0.4-cp310-cp310-win_amd64.whl", hash = "sha256:30982f79f21bb0ff7d7d4a4e531e3fcaa39b778584c2ce81a147f95be1cd58c9"}, + {file = "shapely-2.0.4-cp311-cp311-macosx_10_9_universal2.whl", hash = "sha256:de0205cb21ad5ddaef607cda9a3191eadd1e7a62a756ea3a356369675230ac35"}, + {file = "shapely-2.0.4-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:7d56ce3e2a6a556b59a288771cf9d091470116867e578bebced8bfc4147fbfd7"}, + {file = "shapely-2.0.4-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:58b0ecc505bbe49a99551eea3f2e8a9b3b24b3edd2a4de1ac0dc17bc75c9ec07"}, + {file = "shapely-2.0.4-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:790a168a808bd00ee42786b8ba883307c0e3684ebb292e0e20009588c426da47"}, + {file = "shapely-2.0.4-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:4310b5494271e18580d61022c0857eb85d30510d88606fa3b8314790df7f367d"}, + {file = "shapely-2.0.4-cp311-cp311-win32.whl", hash = "sha256:63f3a80daf4f867bd80f5c97fbe03314348ac1b3b70fb1c0ad255a69e3749879"}, + {file = "shapely-2.0.4-cp311-cp311-win_amd64.whl", hash = "sha256:c52ed79f683f721b69a10fb9e3d940a468203f5054927215586c5d49a072de8d"}, + {file = "shapely-2.0.4-cp312-cp312-macosx_10_9_universal2.whl", hash = "sha256:5bbd974193e2cc274312da16b189b38f5f128410f3377721cadb76b1e8ca5328"}, + {file = "shapely-2.0.4-cp312-cp312-macosx_10_9_x86_64.whl", hash = "sha256:41388321a73ba1a84edd90d86ecc8bfed55e6a1e51882eafb019f45895ec0f65"}, + {file = "shapely-2.0.4-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:0776c92d584f72f1e584d2e43cfc5542c2f3dd19d53f70df0900fda643f4bae6"}, + {file = "shapely-2.0.4-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:c75c98380b1ede1cae9a252c6dc247e6279403fae38c77060a5e6186c95073ac"}, + {file = "shapely-2.0.4-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:c3e700abf4a37b7b8b90532fa6ed5c38a9bfc777098bc9fbae5ec8e618ac8f30"}, + {file = "shapely-2.0.4-cp312-cp312-win32.whl", hash = "sha256:4f2ab0faf8188b9f99e6a273b24b97662194160cc8ca17cf9d1fb6f18d7fb93f"}, + {file = "shapely-2.0.4-cp312-cp312-win_amd64.whl", hash = "sha256:03152442d311a5e85ac73b39680dd64a9892fa42bb08fd83b3bab4fe6999bfa0"}, + {file = "shapely-2.0.4-cp37-cp37m-macosx_10_9_x86_64.whl", hash = "sha256:994c244e004bc3cfbea96257b883c90a86e8cbd76e069718eb4c6b222a56f78b"}, + {file = "shapely-2.0.4-cp37-cp37m-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:05ffd6491e9e8958b742b0e2e7c346635033d0a5f1a0ea083547fcc854e5d5cf"}, + {file = "shapely-2.0.4-cp37-cp37m-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:2fbdc1140a7d08faa748256438291394967aa54b40009f54e8d9825e75ef6113"}, + {file = "shapely-2.0.4-cp37-cp37m-win32.whl", hash = "sha256:5af4cd0d8cf2912bd95f33586600cac9c4b7c5053a036422b97cfe4728d2eb53"}, + {file = "shapely-2.0.4-cp37-cp37m-win_amd64.whl", hash = "sha256:464157509ce4efa5ff285c646a38b49f8c5ef8d4b340f722685b09bb033c5ccf"}, + {file = "shapely-2.0.4-cp38-cp38-macosx_10_9_universal2.whl", hash = "sha256:489c19152ec1f0e5c5e525356bcbf7e532f311bff630c9b6bc2db6f04da6a8b9"}, + {file = "shapely-2.0.4-cp38-cp38-macosx_10_9_x86_64.whl", hash = "sha256:b79bbd648664aa6f44ef018474ff958b6b296fed5c2d42db60078de3cffbc8aa"}, + {file = "shapely-2.0.4-cp38-cp38-macosx_11_0_arm64.whl", hash = "sha256:674d7baf0015a6037d5758496d550fc1946f34bfc89c1bf247cabdc415d7747e"}, + {file = "shapely-2.0.4-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:6cd4ccecc5ea5abd06deeaab52fcdba372f649728050c6143cc405ee0c166679"}, + {file = "shapely-2.0.4-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:fb5cdcbbe3080181498931b52a91a21a781a35dcb859da741c0345c6402bf00c"}, + {file = "shapely-2.0.4-cp38-cp38-win32.whl", hash = "sha256:55a38dcd1cee2f298d8c2ebc60fc7d39f3b4535684a1e9e2f39a80ae88b0cea7"}, + {file = "shapely-2.0.4-cp38-cp38-win_amd64.whl", hash = "sha256:ec555c9d0db12d7fd777ba3f8b75044c73e576c720a851667432fabb7057da6c"}, + {file = "shapely-2.0.4-cp39-cp39-macosx_10_9_universal2.whl", hash = "sha256:3f9103abd1678cb1b5f7e8e1af565a652e036844166c91ec031eeb25c5ca8af0"}, + {file = "shapely-2.0.4-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:263bcf0c24d7a57c80991e64ab57cba7a3906e31d2e21b455f493d4aab534aaa"}, + {file = "shapely-2.0.4-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:ddf4a9bfaac643e62702ed662afc36f6abed2a88a21270e891038f9a19bc08fc"}, + {file = "shapely-2.0.4-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:485246fcdb93336105c29a5cfbff8a226949db37b7473c89caa26c9bae52a242"}, + {file = "shapely-2.0.4-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:8de4578e838a9409b5b134a18ee820730e507b2d21700c14b71a2b0757396acc"}, + {file = "shapely-2.0.4-cp39-cp39-win32.whl", hash = "sha256:9dab4c98acfb5fb85f5a20548b5c0abe9b163ad3525ee28822ffecb5c40e724c"}, + {file = "shapely-2.0.4-cp39-cp39-win_amd64.whl", hash = "sha256:31c19a668b5a1eadab82ff070b5a260478ac6ddad3a5b62295095174a8d26398"}, + {file = "shapely-2.0.4.tar.gz", hash = "sha256:5dc736127fac70009b8d309a0eeb74f3e08979e530cf7017f2f507ef62e6cfb8"}, ] [package.dependencies] -numpy = ">=1.14" +numpy = ">=1.14,<3" [package.extras] docs = ["matplotlib", "numpydoc (==1.1.*)", "sphinx", "sphinx-book-theme", "sphinx-remove-toctrees"] @@ -4236,17 +4254,18 @@ files = [ [[package]] name = "tenacity" -version = "8.2.3" +version = "8.3.0" description = "Retry code until it succeeds" optional = false -python-versions = ">=3.7" +python-versions = ">=3.8" files = [ - {file = "tenacity-8.2.3-py3-none-any.whl", hash = "sha256:ce510e327a630c9e1beaf17d42e6ffacc88185044ad85cf74c0a8887c6a0f88c"}, - {file = "tenacity-8.2.3.tar.gz", hash = "sha256:5398ef0d78e63f40007c1fb4c0bff96e1911394d2fa8d194f77619c05ff6cc8a"}, + {file = "tenacity-8.3.0-py3-none-any.whl", hash = "sha256:3649f6443dbc0d9b01b9d8020a9c4ec7a1ff5f6f3c6c8a036ef371f573fe9185"}, + {file = "tenacity-8.3.0.tar.gz", hash = "sha256:953d4e6ad24357bceffbc9707bc74349aca9d245f68eb65419cf0c249a1949a2"}, ] [package.extras] -doc = ["reno", "sphinx", "tornado (>=4.5)"] +doc = ["reno", "sphinx"] +test = ["pytest", "tornado (>=4.5)", "typeguard"] [[package]] name = "tomli" @@ -4292,13 +4311,13 @@ files = [ [[package]] name = "tqdm" -version = "4.66.1" +version = "4.66.4" description = "Fast, Extensible Progress Meter" optional = false python-versions = ">=3.7" files = [ - {file = "tqdm-4.66.1-py3-none-any.whl", hash = "sha256:d302b3c5b53d47bce91fea46679d9c3c6508cf6332229aa1e7d8653723793386"}, - {file = "tqdm-4.66.1.tar.gz", hash = "sha256:d88e651f9db8d8551a62556d3cff9e3034274ca5d66e93197cf2490e2dcb69c7"}, + {file = "tqdm-4.66.4-py3-none-any.whl", hash = "sha256:b75ca56b413b030bc3f00af51fd2c1a1a5eac6a0c1cca83cbb37a5c52abce644"}, + {file = "tqdm-4.66.4.tar.gz", hash = "sha256:e4d936c9de8727928f3be6079590e97d9abfe8d39a590be678eb5919ffc186bb"}, ] [package.dependencies] @@ -4312,72 +4331,72 @@ telegram = ["requests"] [[package]] name = "traitlets" -version = "5.14.1" +version = "5.14.3" description = "Traitlets Python configuration system" optional = false python-versions = ">=3.8" files = [ - {file = "traitlets-5.14.1-py3-none-any.whl", hash = "sha256:2e5a030e6eff91737c643231bfcf04a65b0132078dad75e4936700b213652e74"}, - {file = "traitlets-5.14.1.tar.gz", hash = "sha256:8585105b371a04b8316a43d5ce29c098575c2e477850b62b848b964f1444527e"}, + {file = "traitlets-5.14.3-py3-none-any.whl", hash = "sha256:b74e89e397b1ed28cc831db7aea759ba6640cb3de13090ca145426688ff1ac4f"}, + {file = "traitlets-5.14.3.tar.gz", hash = "sha256:9ed0579d3502c94b4b3732ac120375cda96f923114522847de4b3bb98b96b6b7"}, ] [package.extras] docs = ["myst-parser", "pydata-sphinx-theme", "sphinx"] -test = ["argcomplete (>=3.0.3)", "mypy (>=1.7.0)", "pre-commit", "pytest (>=7.0,<7.5)", "pytest-mock", "pytest-mypy-testing"] +test = ["argcomplete (>=3.0.3)", "mypy (>=1.7.0)", "pre-commit", "pytest (>=7.0,<8.2)", "pytest-mock", "pytest-mypy-testing"] [[package]] name = "types-decorator" -version = "5.1.8.20240106" +version = "5.1.8.20240310" description = "Typing stubs for decorator" optional = false python-versions = ">=3.8" files = [ - {file = "types-decorator-5.1.8.20240106.tar.gz", hash = "sha256:32ff92b33615060d23b9d3760124bdb3506c4aa8d9eb50963cf1a3c20b9ecbbf"}, - {file = "types_decorator-5.1.8.20240106-py3-none-any.whl", hash = "sha256:14d21e6a0755dbb8f301f2f532b3eab5148f433c69dad2d98bf5bd2b3a2ef4e7"}, + {file = "types-decorator-5.1.8.20240310.tar.gz", hash = "sha256:52e316b03783886a8a2abdc228f7071680ba65894545cd2085ebe3cf88684a0e"}, + {file = "types_decorator-5.1.8.20240310-py3-none-any.whl", hash = "sha256:3af75dc38f5baf65b9b53ea6661ce2056c5ca7d70d620d0b1f620285c1242757"}, ] [[package]] name = "types-pytz" -version = "2024.1.0.20240203" +version = "2024.1.0.20240417" description = "Typing stubs for pytz" optional = false python-versions = ">=3.8" files = [ - {file = "types-pytz-2024.1.0.20240203.tar.gz", hash = "sha256:c93751ee20dfc6e054a0148f8f5227b9a00b79c90a4d3c9f464711a73179c89e"}, - {file = "types_pytz-2024.1.0.20240203-py3-none-any.whl", hash = "sha256:9679eef0365db3af91ef7722c199dbb75ee5c1b67e3c4dd7bfbeb1b8a71c21a3"}, + {file = "types-pytz-2024.1.0.20240417.tar.gz", hash = "sha256:6810c8a1f68f21fdf0f4f374a432487c77645a0ac0b31de4bf4690cf21ad3981"}, + {file = "types_pytz-2024.1.0.20240417-py3-none-any.whl", hash = "sha256:8335d443310e2db7b74e007414e74c4f53b67452c0cb0d228ca359ccfba59659"}, ] [[package]] name = "typing-extensions" -version = "4.9.0" +version = "4.12.0" description = "Backported and Experimental Type Hints for Python 3.8+" optional = false python-versions = ">=3.8" files = [ - {file = "typing_extensions-4.9.0-py3-none-any.whl", hash = "sha256:af72aea155e91adfc61c3ae9e0e342dbc0cba726d6cba4b6c72c1f34e47291cd"}, - {file = "typing_extensions-4.9.0.tar.gz", hash = "sha256:23478f88c37f27d76ac8aee6c905017a143b0b1b886c3c9f66bc2fd94f9f5783"}, + {file = "typing_extensions-4.12.0-py3-none-any.whl", hash = "sha256:b349c66bea9016ac22978d800cfff206d5f9816951f12a7d0ec5578b0a819594"}, + {file = "typing_extensions-4.12.0.tar.gz", hash = "sha256:8cbcdc8606ebcb0d95453ad7dc5065e6237b6aa230a31e81d0f440c30fed5fd8"}, ] [[package]] name = "tzdata" -version = "2023.4" +version = "2024.1" description = "Provider of IANA time zone data" optional = false python-versions = ">=2" files = [ - {file = "tzdata-2023.4-py2.py3-none-any.whl", hash = "sha256:aa3ace4329eeacda5b7beb7ea08ece826c28d761cda36e747cfbf97996d39bf3"}, - {file = "tzdata-2023.4.tar.gz", hash = "sha256:dd54c94f294765522c77399649b4fefd95522479a664a0cec87f41bebc6148c9"}, + {file = "tzdata-2024.1-py2.py3-none-any.whl", hash = "sha256:9068bc196136463f5245e51efda838afa15aaeca9903f49050dfa2679db4d252"}, + {file = "tzdata-2024.1.tar.gz", hash = "sha256:2674120f8d891909751c38abcdfd386ac0a5a1127954fbc332af6b5ceae07efd"}, ] [[package]] name = "uc-micro-py" -version = "1.0.2" +version = "1.0.3" description = "Micro subset of unicode data files for linkify-it-py projects." optional = false python-versions = ">=3.7" files = [ - {file = "uc-micro-py-1.0.2.tar.gz", hash = "sha256:30ae2ac9c49f39ac6dce743bd187fcd2b574b16ca095fa74cd9396795c954c54"}, - {file = "uc_micro_py-1.0.2-py3-none-any.whl", hash = "sha256:8c9110c309db9d9e87302e2f4ad2c3152770930d88ab385cd544e7a7e75f3de0"}, + {file = "uc-micro-py-1.0.3.tar.gz", hash = "sha256:d321b92cff673ec58027c04015fcaa8bb1e005478643ff4a500882eaab88c48a"}, + {file = "uc_micro_py-1.0.3-py3-none-any.whl", hash = "sha256:db1dffff340817673d7b466ec86114a9dc0e9d4d9b5ba229d9d60e5c12600cd5"}, ] [package.extras] @@ -4385,13 +4404,13 @@ test = ["coverage", "pytest", "pytest-cov"] [[package]] name = "urllib3" -version = "2.2.0" +version = "2.2.1" description = "HTTP library with thread-safe connection pooling, file post, and more." optional = false python-versions = ">=3.8" files = [ - {file = "urllib3-2.2.0-py3-none-any.whl", hash = "sha256:ce3711610ddce217e6d113a2732fafad960a03fd0318c91faa79481e35c11224"}, - {file = "urllib3-2.2.0.tar.gz", hash = "sha256:051d961ad0c62a94e50ecf1af379c3aba230c66c710493493560c0c223c49f20"}, + {file = "urllib3-2.2.1-py3-none-any.whl", hash = "sha256:450b20ec296a467077128bff42b73080516e71b56ff59a60a02bef2232c4fa9d"}, + {file = "urllib3-2.2.1.tar.gz", hash = "sha256:d0570876c61ab9e520d776c38acbbb5b05a776d3f9ff98a5c8fd5162a444cf19"}, ] [package.extras] @@ -4402,38 +4421,43 @@ zstd = ["zstandard (>=0.18.0)"] [[package]] name = "watchdog" -version = "3.0.0" +version = "4.0.1" description = "Filesystem events monitoring" optional = false -python-versions = ">=3.7" +python-versions = ">=3.8" files = [ - {file = "watchdog-3.0.0-cp310-cp310-macosx_10_9_universal2.whl", hash = "sha256:336adfc6f5cc4e037d52db31194f7581ff744b67382eb6021c868322e32eef41"}, - {file = "watchdog-3.0.0-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:a70a8dcde91be523c35b2bf96196edc5730edb347e374c7de7cd20c43ed95397"}, - {file = "watchdog-3.0.0-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:adfdeab2da79ea2f76f87eb42a3ab1966a5313e5a69a0213a3cc06ef692b0e96"}, - {file = "watchdog-3.0.0-cp311-cp311-macosx_10_9_universal2.whl", hash = "sha256:2b57a1e730af3156d13b7fdddfc23dea6487fceca29fc75c5a868beed29177ae"}, - {file = "watchdog-3.0.0-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:7ade88d0d778b1b222adebcc0927428f883db07017618a5e684fd03b83342bd9"}, - {file = "watchdog-3.0.0-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:7e447d172af52ad204d19982739aa2346245cc5ba6f579d16dac4bfec226d2e7"}, - {file = "watchdog-3.0.0-cp37-cp37m-macosx_10_9_x86_64.whl", hash = "sha256:9fac43a7466eb73e64a9940ac9ed6369baa39b3bf221ae23493a9ec4d0022674"}, - {file = "watchdog-3.0.0-cp38-cp38-macosx_10_9_universal2.whl", hash = "sha256:8ae9cda41fa114e28faf86cb137d751a17ffd0316d1c34ccf2235e8a84365c7f"}, - {file = "watchdog-3.0.0-cp38-cp38-macosx_10_9_x86_64.whl", hash = "sha256:25f70b4aa53bd743729c7475d7ec41093a580528b100e9a8c5b5efe8899592fc"}, - {file = "watchdog-3.0.0-cp38-cp38-macosx_11_0_arm64.whl", hash = "sha256:4f94069eb16657d2c6faada4624c39464f65c05606af50bb7902e036e3219be3"}, - {file = "watchdog-3.0.0-cp39-cp39-macosx_10_9_universal2.whl", hash = "sha256:7c5f84b5194c24dd573fa6472685b2a27cc5a17fe5f7b6fd40345378ca6812e3"}, - {file = "watchdog-3.0.0-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:3aa7f6a12e831ddfe78cdd4f8996af9cf334fd6346531b16cec61c3b3c0d8da0"}, - {file = "watchdog-3.0.0-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:233b5817932685d39a7896b1090353fc8efc1ef99c9c054e46c8002561252fb8"}, - {file = "watchdog-3.0.0-pp37-pypy37_pp73-macosx_10_9_x86_64.whl", hash = "sha256:13bbbb462ee42ec3c5723e1205be8ced776f05b100e4737518c67c8325cf6100"}, - {file = "watchdog-3.0.0-pp38-pypy38_pp73-macosx_10_9_x86_64.whl", hash = "sha256:8f3ceecd20d71067c7fd4c9e832d4e22584318983cabc013dbf3f70ea95de346"}, - {file = "watchdog-3.0.0-pp39-pypy39_pp73-macosx_10_9_x86_64.whl", hash = "sha256:c9d8c8ec7efb887333cf71e328e39cffbf771d8f8f95d308ea4125bf5f90ba64"}, - {file = "watchdog-3.0.0-py3-none-manylinux2014_aarch64.whl", hash = "sha256:0e06ab8858a76e1219e68c7573dfeba9dd1c0219476c5a44d5333b01d7e1743a"}, - {file = "watchdog-3.0.0-py3-none-manylinux2014_armv7l.whl", hash = "sha256:d00e6be486affb5781468457b21a6cbe848c33ef43f9ea4a73b4882e5f188a44"}, - {file = "watchdog-3.0.0-py3-none-manylinux2014_i686.whl", hash = "sha256:c07253088265c363d1ddf4b3cdb808d59a0468ecd017770ed716991620b8f77a"}, - {file = "watchdog-3.0.0-py3-none-manylinux2014_ppc64.whl", hash = "sha256:5113334cf8cf0ac8cd45e1f8309a603291b614191c9add34d33075727a967709"}, - {file = "watchdog-3.0.0-py3-none-manylinux2014_ppc64le.whl", hash = "sha256:51f90f73b4697bac9c9a78394c3acbbd331ccd3655c11be1a15ae6fe289a8c83"}, - {file = "watchdog-3.0.0-py3-none-manylinux2014_s390x.whl", hash = "sha256:ba07e92756c97e3aca0912b5cbc4e5ad802f4557212788e72a72a47ff376950d"}, - {file = "watchdog-3.0.0-py3-none-manylinux2014_x86_64.whl", hash = "sha256:d429c2430c93b7903914e4db9a966c7f2b068dd2ebdd2fa9b9ce094c7d459f33"}, - {file = "watchdog-3.0.0-py3-none-win32.whl", hash = "sha256:3ed7c71a9dccfe838c2f0b6314ed0d9b22e77d268c67e015450a29036a81f60f"}, - {file = "watchdog-3.0.0-py3-none-win_amd64.whl", hash = "sha256:4c9956d27be0bb08fc5f30d9d0179a855436e655f046d288e2bcc11adfae893c"}, - {file = "watchdog-3.0.0-py3-none-win_ia64.whl", hash = "sha256:5d9f3a10e02d7371cd929b5d8f11e87d4bad890212ed3901f9b4d68767bee759"}, - {file = "watchdog-3.0.0.tar.gz", hash = "sha256:4d98a320595da7a7c5a18fc48cb633c2e73cda78f93cac2ef42d42bf609a33f9"}, + {file = "watchdog-4.0.1-cp310-cp310-macosx_10_9_universal2.whl", hash = "sha256:da2dfdaa8006eb6a71051795856bedd97e5b03e57da96f98e375682c48850645"}, + {file = "watchdog-4.0.1-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:e93f451f2dfa433d97765ca2634628b789b49ba8b504fdde5837cdcf25fdb53b"}, + {file = "watchdog-4.0.1-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:ef0107bbb6a55f5be727cfc2ef945d5676b97bffb8425650dadbb184be9f9a2b"}, + {file = "watchdog-4.0.1-cp311-cp311-macosx_10_9_universal2.whl", hash = "sha256:17e32f147d8bf9657e0922c0940bcde863b894cd871dbb694beb6704cfbd2fb5"}, + {file = "watchdog-4.0.1-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:03e70d2df2258fb6cb0e95bbdbe06c16e608af94a3ffbd2b90c3f1e83eb10767"}, + {file = "watchdog-4.0.1-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:123587af84260c991dc5f62a6e7ef3d1c57dfddc99faacee508c71d287248459"}, + {file = "watchdog-4.0.1-cp312-cp312-macosx_10_9_universal2.whl", hash = "sha256:093b23e6906a8b97051191a4a0c73a77ecc958121d42346274c6af6520dec175"}, + {file = "watchdog-4.0.1-cp312-cp312-macosx_10_9_x86_64.whl", hash = "sha256:611be3904f9843f0529c35a3ff3fd617449463cb4b73b1633950b3d97fa4bfb7"}, + {file = "watchdog-4.0.1-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:62c613ad689ddcb11707f030e722fa929f322ef7e4f18f5335d2b73c61a85c28"}, + {file = "watchdog-4.0.1-cp38-cp38-macosx_10_9_universal2.whl", hash = "sha256:d4925e4bf7b9bddd1c3de13c9b8a2cdb89a468f640e66fbfabaf735bd85b3e35"}, + {file = "watchdog-4.0.1-cp38-cp38-macosx_10_9_x86_64.whl", hash = "sha256:cad0bbd66cd59fc474b4a4376bc5ac3fc698723510cbb64091c2a793b18654db"}, + {file = "watchdog-4.0.1-cp38-cp38-macosx_11_0_arm64.whl", hash = "sha256:a3c2c317a8fb53e5b3d25790553796105501a235343f5d2bf23bb8649c2c8709"}, + {file = "watchdog-4.0.1-cp39-cp39-macosx_10_9_universal2.whl", hash = "sha256:c9904904b6564d4ee8a1ed820db76185a3c96e05560c776c79a6ce5ab71888ba"}, + {file = "watchdog-4.0.1-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:667f3c579e813fcbad1b784db7a1aaa96524bed53437e119f6a2f5de4db04235"}, + {file = "watchdog-4.0.1-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:d10a681c9a1d5a77e75c48a3b8e1a9f2ae2928eda463e8d33660437705659682"}, + {file = "watchdog-4.0.1-pp310-pypy310_pp73-macosx_10_9_x86_64.whl", hash = "sha256:0144c0ea9997b92615af1d94afc0c217e07ce2c14912c7b1a5731776329fcfc7"}, + {file = "watchdog-4.0.1-pp310-pypy310_pp73-macosx_11_0_arm64.whl", hash = "sha256:998d2be6976a0ee3a81fb8e2777900c28641fb5bfbd0c84717d89bca0addcdc5"}, + {file = "watchdog-4.0.1-pp38-pypy38_pp73-macosx_10_9_x86_64.whl", hash = "sha256:e7921319fe4430b11278d924ef66d4daa469fafb1da679a2e48c935fa27af193"}, + {file = "watchdog-4.0.1-pp38-pypy38_pp73-macosx_11_0_arm64.whl", hash = "sha256:f0de0f284248ab40188f23380b03b59126d1479cd59940f2a34f8852db710625"}, + {file = "watchdog-4.0.1-pp39-pypy39_pp73-macosx_10_9_x86_64.whl", hash = "sha256:bca36be5707e81b9e6ce3208d92d95540d4ca244c006b61511753583c81c70dd"}, + {file = "watchdog-4.0.1-pp39-pypy39_pp73-macosx_11_0_arm64.whl", hash = "sha256:ab998f567ebdf6b1da7dc1e5accfaa7c6992244629c0fdaef062f43249bd8dee"}, + {file = "watchdog-4.0.1-py3-none-manylinux2014_aarch64.whl", hash = "sha256:dddba7ca1c807045323b6af4ff80f5ddc4d654c8bce8317dde1bd96b128ed253"}, + {file = "watchdog-4.0.1-py3-none-manylinux2014_armv7l.whl", hash = "sha256:4513ec234c68b14d4161440e07f995f231be21a09329051e67a2118a7a612d2d"}, + {file = "watchdog-4.0.1-py3-none-manylinux2014_i686.whl", hash = "sha256:4107ac5ab936a63952dea2a46a734a23230aa2f6f9db1291bf171dac3ebd53c6"}, + {file = "watchdog-4.0.1-py3-none-manylinux2014_ppc64.whl", hash = "sha256:6e8c70d2cd745daec2a08734d9f63092b793ad97612470a0ee4cbb8f5f705c57"}, + {file = "watchdog-4.0.1-py3-none-manylinux2014_ppc64le.whl", hash = "sha256:f27279d060e2ab24c0aa98363ff906d2386aa6c4dc2f1a374655d4e02a6c5e5e"}, + {file = "watchdog-4.0.1-py3-none-manylinux2014_s390x.whl", hash = "sha256:f8affdf3c0f0466e69f5b3917cdd042f89c8c63aebdb9f7c078996f607cdb0f5"}, + {file = "watchdog-4.0.1-py3-none-manylinux2014_x86_64.whl", hash = "sha256:ac7041b385f04c047fcc2951dc001671dee1b7e0615cde772e84b01fbf68ee84"}, + {file = "watchdog-4.0.1-py3-none-win32.whl", hash = "sha256:206afc3d964f9a233e6ad34618ec60b9837d0582b500b63687e34011e15bb429"}, + {file = "watchdog-4.0.1-py3-none-win_amd64.whl", hash = "sha256:7577b3c43e5909623149f76b099ac49a1a01ca4e167d1785c76eb52fa585745a"}, + {file = "watchdog-4.0.1-py3-none-win_ia64.whl", hash = "sha256:d7b9f5f3299e8dd230880b6c55504a1f69cf1e4316275d1b215ebdd8187ec88d"}, + {file = "watchdog-4.0.1.tar.gz", hash = "sha256:eebaacf674fa25511e8867028d281e602ee6500045b57f43b08778082f7f8b44"}, ] [package.extras] @@ -4478,13 +4502,13 @@ testing = ["coverage", "pytest (>=3.1.0)", "pytest-cov", "pytest-xdist"] [[package]] name = "xarray" -version = "2024.1.1" +version = "2024.5.0" description = "N-D labeled arrays and datasets in Python" optional = false python-versions = ">=3.9" files = [ - {file = "xarray-2024.1.1-py3-none-any.whl", hash = "sha256:0bec81303b088c8df4f075e1579c00cfd7e5069688e4434007f0b8d7df17fc1c"}, - {file = "xarray-2024.1.1.tar.gz", hash = "sha256:a1ba2d87a74892e213c9c83f4a462dbcdf68212320a4e31b34bd789ba7a64e35"}, + {file = "xarray-2024.5.0-py3-none-any.whl", hash = "sha256:7ddedfe2294a0ab00f02d0fbdcb9c6300ec589f3cf436a9c7b7b577a12cd9bcf"}, + {file = "xarray-2024.5.0.tar.gz", hash = "sha256:e0eb1cb265f265126795f388ed9591f3c752f2aca491f6c0576711fd15b708f2"}, ] [package.dependencies] @@ -4497,8 +4521,8 @@ netCDF4 = {version = "*", optional = true, markers = "extra == \"io\""} numbagg = {version = "*", optional = true, markers = "extra == \"accel\""} numpy = ">=1.23" opt-einsum = {version = "*", optional = true, markers = "extra == \"accel\""} -packaging = ">=22" -pandas = ">=1.5" +packaging = ">=23.1" +pandas = ">=2.0" pooch = {version = "*", optional = true, markers = "extra == \"io\""} pydap = {version = "*", optional = true, markers = "python_version < \"3.10\" and extra == \"io\""} scipy = {version = "*", optional = true, markers = "extra == \"accel\" or extra == \"io\""} @@ -4506,20 +4530,39 @@ zarr = {version = "*", optional = true, markers = "extra == \"io\""} [package.extras] accel = ["bottleneck", "flox", "numbagg", "opt-einsum", "scipy"] -complete = ["xarray[accel,io,parallel,viz]"] +complete = ["xarray[accel,dev,io,parallel,viz]"] +dev = ["hypothesis", "mypy", "pre-commit", "pytest", "pytest-cov", "pytest-env", "pytest-timeout", "pytest-xdist", "ruff", "xarray[complete]"] io = ["cftime", "fsspec", "h5netcdf", "netCDF4", "pooch", "pydap", "scipy", "zarr"] parallel = ["dask[complete]"] viz = ["matplotlib", "nc-time-axis", "seaborn"] +[[package]] +name = "xarray-selafin" +version = "0.1.6" +description = "" +optional = false +python-versions = ">=3.9" +files = [ + {file = "xarray_selafin-0.1.6-py3-none-any.whl", hash = "sha256:e484ae6e26ceb3c91f6a7a4975b171146a94171bf6472dea33a09c90a4e15ad5"}, + {file = "xarray_selafin-0.1.6.tar.gz", hash = "sha256:4fab755eb3089571c1d43b3d84491e4ffa80cdbb0977aa893a63f156be44b493"}, +] + +[package.dependencies] +netcdf4 = "*" +numpy = "*" +scipy = "*" +shapely = "*" +xarray = "*" + [[package]] name = "xyzservices" -version = "2023.10.1" +version = "2024.4.0" description = "Source of XYZ tiles providers" optional = false python-versions = ">=3.8" files = [ - {file = "xyzservices-2023.10.1-py3-none-any.whl", hash = "sha256:6a4c38d3a9f89d3e77153eff9414b36a8ee0850c9e8b85796fd1b2a85b8dfd68"}, - {file = "xyzservices-2023.10.1.tar.gz", hash = "sha256:091229269043bc8258042edbedad4fcb44684b0473ede027b5672ad40dc9fa02"}, + {file = "xyzservices-2024.4.0-py3-none-any.whl", hash = "sha256:b83e48c5b776c9969fffcfff57b03d02b1b1cd6607a9d9c4e7f568b01ef47f4c"}, + {file = "xyzservices-2024.4.0.tar.gz", hash = "sha256:6a04f11487a6fb77d92a98984cd107fbd9157fd5e65f929add9c3d6e604ee88c"}, ] [[package]] @@ -4627,23 +4670,23 @@ multidict = ">=4.0" [[package]] name = "zarr" -version = "2.16.1" +version = "2.18.1" description = "An implementation of chunked, compressed, N-dimensional arrays for Python" optional = false -python-versions = ">=3.8" +python-versions = ">=3.9" files = [ - {file = "zarr-2.16.1-py3-none-any.whl", hash = "sha256:de4882433ccb5b42cc1ec9872b95e64ca3a13581424666b28ed265ad76c7056f"}, - {file = "zarr-2.16.1.tar.gz", hash = "sha256:4276cf4b4a653431042cd53ff2282bc4d292a6842411e88529964504fb073286"}, + {file = "zarr-2.18.1-py3-none-any.whl", hash = "sha256:a1770d194eec4ec0a41a01295a6f724e1c3471d704d3aca906d3b3a7f8830245"}, + {file = "zarr-2.18.1.tar.gz", hash = "sha256:28c360ed123e606c425a694a83300227a907cb86a995fc9eef620ecafbe5f92d"}, ] [package.dependencies] asciitree = "*" -fasteners = "*" +fasteners = {version = "*", markers = "sys_platform != \"emscripten\""} numcodecs = ">=0.10.0" -numpy = ">=1.20,<1.21.0 || >1.21.0" +numpy = ">=1.23" [package.extras] -docs = ["numcodecs[msgpack]", "numpydoc", "pydata-sphinx-theme", "sphinx", "sphinx-copybutton", "sphinx-design", "sphinx-issues", "sphinx-rtd-theme"] +docs = ["numcodecs[msgpack]", "numpydoc", "pydata-sphinx-theme", "sphinx", "sphinx-automodapi", "sphinx-copybutton", "sphinx-design", "sphinx-issues"] jupyter = ["ipytree (>=0.2.2)", "ipywidgets (>=8.0.0)", "notebook"] [[package]] @@ -4659,20 +4702,20 @@ files = [ [[package]] name = "zipp" -version = "3.17.0" +version = "3.18.2" description = "Backport of pathlib-compatible object wrapper for zip files" optional = false python-versions = ">=3.8" files = [ - {file = "zipp-3.17.0-py3-none-any.whl", hash = "sha256:0e923e726174922dce09c53c59ad483ff7bbb8e572e00c7f7c46b88556409f31"}, - {file = "zipp-3.17.0.tar.gz", hash = "sha256:84e64a1c28cf7e91ed2078bb8cc8c259cb19b76942096c8d7b84947690cabaf0"}, + {file = "zipp-3.18.2-py3-none-any.whl", hash = "sha256:dce197b859eb796242b0622af1b8beb0a722d52aa2f57133ead08edd5bf5374e"}, + {file = "zipp-3.18.2.tar.gz", hash = "sha256:6278d9ddbcfb1f1089a88fde84481528b07b0e10474e09dcfe53dad4069fa059"}, ] [package.extras] -docs = ["furo", "jaraco.packaging (>=9.3)", "jaraco.tidelift (>=1.4)", "rst.linker (>=1.9)", "sphinx (<7.2.5)", "sphinx (>=3.5)", "sphinx-lint"] -testing = ["big-O", "jaraco.functools", "jaraco.itertools", "more-itertools", "pytest (>=6)", "pytest-black (>=0.3.7)", "pytest-checkdocs (>=2.4)", "pytest-cov", "pytest-enabler (>=2.2)", "pytest-ignore-flaky", "pytest-mypy (>=0.9.1)", "pytest-ruff"] +docs = ["furo", "jaraco.packaging (>=9.3)", "jaraco.tidelift (>=1.4)", "rst.linker (>=1.9)", "sphinx (>=3.5)", "sphinx-lint"] +testing = ["big-O", "jaraco.functools", "jaraco.itertools", "jaraco.test", "more-itertools", "pytest (>=6,!=8.1.*)", "pytest-checkdocs (>=2.4)", "pytest-cov", "pytest-enabler (>=2.2)", "pytest-ignore-flaky", "pytest-mypy", "pytest-ruff (>=0.2.1)"] [metadata] lock-version = "2.0" python-versions = ">=3.9, <4.0" -content-hash = "606ede5919592db268c2adb911ed22e5f474a51edd4ce54f918521b703e41db5" +content-hash = "34abafe09252a6314c18832a0c53ff4a80cbeb13249ed940c43f3f4a82ea90af" diff --git a/pyproject.toml b/pyproject.toml index e10c9ee..9351c4c 100644 --- a/pyproject.toml +++ b/pyproject.toml @@ -49,6 +49,7 @@ ipykernel = "*" ipython = "*" papermill = "*" pandas-stubs = "*" +xarray-selafin = "*" [tool.poetry.group.docs.dependencies] mkdocs = "*" diff --git a/requirements/requirements-dev.txt b/requirements/requirements-dev.txt index a2da168..0b536c1 100644 --- a/requirements/requirements-dev.txt +++ b/requirements/requirements-dev.txt @@ -1,16 +1,17 @@ -aiohttp==3.9.0b0 ; python_version == "3.12" +aiohttp==3.9.5 ; python_version == "3.12" aiosignal==1.3.1 ; python_version == "3.12" +ansicolors==1.1.8 ; python_version >= "3.9" and python_version < "4.0" appnope==0.1.4 ; python_version >= "3.9" and python_version < "4.0" and platform_system == "Darwin" asciitree==0.3.3 ; python_version >= "3.9" and python_version < "4.0" asttokens==2.4.1 ; python_version >= "3.9" and python_version < "4.0" attrs==23.2.0 ; python_version >= "3.9" and python_version < "4.0" -babel==2.14.0 ; python_version >= "3.9" and python_version < "4.0" +babel==2.15.0 ; python_version >= "3.9" and python_version < "4.0" beautifulsoup4==4.12.3 ; python_version < "3.10" and python_version >= "3.9" -black==24.1.1 ; python_version >= "3.9" and python_version < "4.0" +black==24.4.2 ; python_version >= "3.9" and python_version < "4.0" bleach==6.1.0 ; python_version >= "3.9" and python_version < "4.0" -bokeh==3.3.4 ; python_version >= "3.9" and python_version < "4.0" -bottleneck==1.3.7 ; python_version >= "3.9" and python_version < "4.0" -cartopy==0.22.0 ; python_version >= "3.9" and python_version < "4.0" +bokeh==3.4.1 ; python_version >= "3.9" and python_version < "4.0" +bottleneck==1.3.8 ; python_version >= "3.9" and python_version < "4.0" +cartopy==0.23.0 ; python_version >= "3.9" and python_version < "4.0" certifi==2024.2.2 ; python_version >= "3.9" and python_version < "4.0" cffi==1.16.0 ; python_version >= "3.9" and python_version < "4.0" and implementation_name == "pypy" cftime==1.6.3 ; python_version >= "3.9" and python_version < "4.0" @@ -20,155 +21,157 @@ click==8.1.7 ; python_version >= "3.9" and python_version < "4.0" cligj==0.7.2 ; python_version >= "3.9" and python_version < "4" cloudpickle==3.0.0 ; python_version >= "3.9" and python_version < "4.0" colorama==0.4.6 ; python_version >= "3.9" and python_version < "4.0" -colorcet==3.0.1 ; python_version >= "3.9" and python_version < "4.0" -comm==0.2.1 ; python_version >= "3.9" and python_version < "4.0" -contourpy==1.2.0 ; python_version >= "3.9" and python_version < "4.0" +colorcet==3.1.0 ; python_version >= "3.9" and python_version < "4.0" +comm==0.2.2 ; python_version >= "3.9" and python_version < "4.0" +contourpy==1.2.1 ; python_version >= "3.9" and python_version < "4.0" covdefaults==2.3.0 ; python_version >= "3.9" and python_version < "4.0" -coverage==7.4.1 ; python_version >= "3.9" and python_version < "4.0" -coverage[toml]==7.4.1 ; python_version >= "3.9" and python_version < "4.0" +coverage==7.5.1 ; python_version >= "3.9" and python_version < "4.0" +coverage[toml]==7.5.1 ; python_version >= "3.9" and python_version < "4.0" cycler==0.12.1 ; python_version >= "3.9" and python_version < "4.0" -dask==2024.1.1 ; python_version >= "3.9" and python_version < "4.0" -dask[array,complete,dataframe,diagnostics,distributed]==2024.1.1 ; python_version >= "3.9" and python_version < "4.0" -datashader==0.16.0 ; python_version >= "3.9" and python_version < "4.0" -debugpy==1.8.0 ; python_version >= "3.9" and python_version < "4.0" +dask-expr==1.1.1 ; python_version >= "3.9" and python_version < "4.0" +dask==2024.5.1 ; python_version >= "3.9" and python_version < "4.0" +dask[array,complete,dataframe,diagnostics,distributed]==2024.5.1 ; python_version >= "3.9" and python_version < "4.0" +datashader==0.16.1 ; python_version >= "3.9" and python_version < "4.0" +debugpy==1.8.1 ; python_version >= "3.9" and python_version < "4.0" decorator==5.1.1 ; python_version >= "3.9" and python_version < "4.0" -distributed==2024.1.1 ; python_version >= "3.9" and python_version < "4.0" +distributed==2024.5.1 ; python_version >= "3.9" and python_version < "4.0" docopt==0.6.2 ; python_version < "3.10" and python_version >= "3.9" entrypoints==0.4 ; python_version >= "3.9" and python_version < "4.0" -exceptiongroup==1.2.0 ; python_version >= "3.9" and python_version < "3.11" +exceptiongroup==1.2.1 ; python_version >= "3.9" and python_version < "3.11" executing==2.0.1 ; python_version >= "3.9" and python_version < "4.0" -fasteners==0.19 ; python_version >= "3.9" and python_version < "4.0" +fasteners==0.19 ; python_version >= "3.9" and python_version < "4.0" and sys_platform != "emscripten" fastjsonschema==2.19.1 ; python_version >= "3.9" and python_version < "4.0" -fiona==1.9.5 ; python_version >= "3.9" and python_version < "4.0" -flox==0.9.0 ; python_version >= "3.9" and python_version < "4.0" -fonttools==4.48.1 ; python_version >= "3.9" and python_version < "4.0" +fiona==1.9.6 ; python_version >= "3.9" and python_version < "4.0" +flox==0.9.7 ; python_version >= "3.9" and python_version < "4.0" +fonttools==4.51.0 ; python_version >= "3.9" and python_version < "4.0" frozenlist==1.4.1 ; python_version == "3.12" -fsspec==2024.2.0 ; python_version >= "3.9" and python_version < "4.0" -future==0.18.3 ; python_version >= "3.9" and python_version < "4.0" -geopandas==0.14.3 ; python_version >= "3.9" and python_version < "4.0" -geoviews==1.11.0 ; python_version >= "3.9" and python_version < "4.0" +fsspec==2024.5.0 ; python_version >= "3.9" and python_version < "4.0" +future==1.0.0 ; python_version >= "3.9" and python_version < "4.0" +geopandas==0.14.4 ; python_version >= "3.9" and python_version < "4.0" +geoviews==1.12.0 ; python_version >= "3.9" and python_version < "4.0" ghp-import==2.1.0 ; python_version >= "3.9" and python_version < "4.0" -griffe==0.40.0 ; python_version >= "3.9" and python_version < "4.0" +griffe==0.45.2 ; python_version >= "3.9" and python_version < "4.0" h5netcdf==1.3.0 ; python_version >= "3.9" and python_version < "4.0" -h5py==3.10.0 ; python_version >= "3.9" and python_version < "4.0" -holoviews==1.18.1 ; python_version >= "3.9" and python_version < "4.0" -idna==3.6 ; python_version >= "3.9" and python_version < "4.0" -importlib-metadata==7.0.1 ; python_version >= "3.9" and python_version < "4.0" -importlib-resources==6.1.1 ; python_version >= "3.9" and python_version < "3.10" +h5py==3.11.0 ; python_version >= "3.9" and python_version < "4.0" +holoviews==1.18.3 ; python_version >= "3.9" and python_version < "4.0" +idna==3.7 ; python_version >= "3.9" and python_version < "4.0" +importlib-metadata==7.1.0 ; python_version >= "3.9" and python_version < "3.12" +importlib-resources==6.4.0 ; python_version >= "3.9" and python_version < "3.10" iniconfig==2.0.0 ; python_version >= "3.9" and python_version < "4.0" -ipykernel==6.29.1 ; python_version >= "3.9" and python_version < "4.0" +ipykernel==6.29.4 ; python_version >= "3.9" and python_version < "4.0" ipython==8.18.1 ; python_version >= "3.9" and python_version < "4.0" jedi==0.19.1 ; python_version >= "3.9" and python_version < "4.0" -jinja2==3.1.3 ; python_version >= "3.9" and python_version < "4.0" +jinja2==3.1.4 ; python_version >= "3.9" and python_version < "4.0" jsonschema-specifications==2023.12.1 ; python_version >= "3.9" and python_version < "4.0" -jsonschema==4.21.1 ; python_version >= "3.9" and python_version < "4.0" -jupyter-client==8.6.0 ; python_version >= "3.9" and python_version < "4.0" -jupyter-core==5.7.1 ; python_version >= "3.9" and python_version < "4.0" +jsonschema==4.22.0 ; python_version >= "3.9" and python_version < "4.0" +jupyter-client==8.6.2 ; python_version >= "3.9" and python_version < "4.0" +jupyter-core==5.7.2 ; python_version >= "3.9" and python_version < "4.0" kiwisolver==1.4.5 ; python_version >= "3.9" and python_version < "4.0" linkify-it-py==2.0.3 ; python_version >= "3.9" and python_version < "4.0" llvmlite==0.42.0 ; python_version >= "3.9" and python_version < "4.0" locket==1.0.0 ; python_version >= "3.9" and python_version < "4.0" lz4==4.3.3 ; python_version >= "3.9" and python_version < "4.0" markdown-it-py==3.0.0 ; python_version >= "3.9" and python_version < "4.0" -markdown==3.5.2 ; python_version >= "3.9" and python_version < "4.0" +markdown==3.6 ; python_version >= "3.9" and python_version < "4.0" markupsafe==2.1.5 ; python_version >= "3.9" and python_version < "4.0" -matplotlib-inline==0.1.6 ; python_version >= "3.9" and python_version < "4.0" -matplotlib==3.8.2 ; python_version >= "3.9" and python_version < "4.0" -mdit-py-plugins==0.4.0 ; python_version >= "3.9" and python_version < "4.0" +matplotlib-inline==0.1.7 ; python_version >= "3.9" and python_version < "4.0" +matplotlib==3.9.0 ; python_version >= "3.9" and python_version < "4.0" +mdit-py-plugins==0.4.1 ; python_version >= "3.9" and python_version < "4.0" mdurl==0.1.2 ; python_version >= "3.9" and python_version < "4.0" mergedeep==1.3.4 ; python_version >= "3.9" and python_version < "4.0" -mkdocs-autorefs==0.5.0 ; python_version >= "3.9" and python_version < "4.0" +mkdocs-autorefs==1.0.1 ; python_version >= "3.9" and python_version < "4.0" +mkdocs-get-deps==0.2.0 ; python_version >= "3.9" and python_version < "4.0" mkdocs-material-extensions==1.3.1 ; python_version >= "3.9" and python_version < "4.0" -mkdocs-material==9.5.7 ; python_version >= "3.9" and python_version < "4.0" -mkdocs==1.5.3 ; python_version >= "3.9" and python_version < "4.0" -mkdocstrings-python==1.8.0 ; python_version >= "3.9" and python_version < "4.0" -mkdocstrings==0.24.0 ; python_version >= "3.9" and python_version < "4.0" -msgpack==1.0.7 ; python_version >= "3.9" and python_version < "4.0" +mkdocs-material==9.5.24 ; python_version >= "3.9" and python_version < "4.0" +mkdocs==1.6.0 ; python_version >= "3.9" and python_version < "4.0" +mkdocstrings-python==1.10.3 ; python_version >= "3.9" and python_version < "4.0" +mkdocstrings==0.25.1 ; python_version >= "3.9" and python_version < "4.0" +msgpack==1.0.8 ; python_version >= "3.9" and python_version < "4.0" multidict==6.0.5 ; python_version == "3.12" multipledispatch==1.0.0 ; python_version >= "3.9" and python_version < "4.0" mypy-extensions==1.0.0 ; python_version >= "3.9" and python_version < "4.0" -mypy==1.8.0 ; python_version >= "3.9" and python_version < "4.0" -nbclient==0.9.0 ; python_version >= "3.9" and python_version < "4.0" -nbformat==5.9.2 ; python_version >= "3.9" and python_version < "4.0" +mypy==1.10.0 ; python_version >= "3.9" and python_version < "4.0" +nbclient==0.10.0 ; python_version >= "3.9" and python_version < "4.0" +nbformat==5.10.4 ; python_version >= "3.9" and python_version < "4.0" nest-asyncio==1.6.0 ; python_version >= "3.9" and python_version < "4.0" netcdf4==1.6.5 ; python_version >= "3.9" and python_version < "4.0" -numba==0.59.0 ; python_version >= "3.9" and python_version < "4.0" -numbagg==0.8.0 ; python_version >= "3.9" and python_version < "4.0" +numba==0.59.1 ; python_version >= "3.9" and python_version < "4.0" +numbagg==0.8.1 ; python_version >= "3.9" and python_version < "4.0" numcodecs==0.12.1 ; python_version >= "3.9" and python_version < "4.0" -numpy-groupies==0.10.2 ; python_version >= "3.9" and python_version < "4.0" +numpy-groupies==0.11.1 ; python_version >= "3.9" and python_version < "4.0" numpy-indexed==0.3.7 ; python_version >= "3.9" and python_version < "4.0" numpy==1.26.4 ; python_version >= "3.9" and python_version < "4.0" opt-einsum==3.3.0 ; python_version >= "3.9" and python_version < "4.0" -packaging==23.2 ; python_version >= "3.9" and python_version < "4.0" +packaging==24.0 ; python_version >= "3.9" and python_version < "4.0" paginate==0.5.6 ; python_version >= "3.9" and python_version < "4.0" -pandas-stubs==2.1.4.231227 ; python_version >= "3.9" and python_version < "4.0" -pandas==2.2.0 ; python_version >= "3.9" and python_version < "4.0" -panel==1.3.8 ; python_version >= "3.9" and python_version < "4.0" -papermill==2.5.0 ; python_version >= "3.9" and python_version < "4.0" -param==2.0.2 ; python_version >= "3.9" and python_version < "4.0" -parso==0.8.3 ; python_version >= "3.9" and python_version < "4.0" -partd==1.4.1 ; python_version >= "3.9" and python_version < "4.0" +pandas-stubs==2.2.2.240514 ; python_version >= "3.9" and python_version < "4.0" +pandas==2.2.2 ; python_version >= "3.9" and python_version < "4.0" +panel==1.4.3 ; python_version >= "3.9" and python_version < "4.0" +papermill==2.6.0 ; python_version >= "3.9" and python_version < "4.0" +param==2.1.0 ; python_version >= "3.9" and python_version < "4.0" +parso==0.8.4 ; python_version >= "3.9" and python_version < "4.0" +partd==1.4.2 ; python_version >= "3.9" and python_version < "4.0" pathspec==0.12.1 ; python_version >= "3.9" and python_version < "4.0" pexpect==4.9.0 ; python_version >= "3.9" and python_version < "4.0" and sys_platform != "win32" -pillow==10.2.0 ; python_version >= "3.9" and python_version < "4.0" -platformdirs==4.2.0 ; python_version >= "3.9" and python_version < "4.0" -pluggy==1.4.0 ; python_version >= "3.9" and python_version < "4.0" -pooch==1.8.0 ; python_version >= "3.9" and python_version < "4.0" +pillow==10.3.0 ; python_version >= "3.9" and python_version < "4.0" +platformdirs==4.2.2 ; python_version >= "3.9" and python_version < "4.0" +pluggy==1.5.0 ; python_version >= "3.9" and python_version < "4.0" +pooch==1.8.1 ; python_version >= "3.9" and python_version < "4.0" prompt-toolkit==3.0.43 ; python_version >= "3.9" and python_version < "4.0" psutil==5.9.8 ; python_version >= "3.9" and python_version < "4.0" ptyprocess==0.7.0 ; python_version >= "3.9" and python_version < "4.0" and sys_platform != "win32" pure-eval==0.2.2 ; python_version >= "3.9" and python_version < "4.0" pyarrow-hotfix==0.6 ; python_version >= "3.9" and python_version < "4.0" -pyarrow==15.0.0 ; python_version >= "3.9" and python_version < "4.0" -pycparser==2.21 ; python_version >= "3.9" and python_version < "4.0" and implementation_name == "pypy" +pyarrow==16.1.0 ; python_version >= "3.9" and python_version < "4.0" +pycparser==2.22 ; python_version >= "3.9" and python_version < "4.0" and implementation_name == "pypy" pyct==0.5.0 ; python_version >= "3.9" and python_version < "4.0" pydap==3.4.1 ; python_version < "3.10" and python_version >= "3.9" -pygments==2.17.2 ; python_version >= "3.9" and python_version < "4.0" -pymdown-extensions==10.7 ; python_version >= "3.9" and python_version < "4.0" -pyparsing==3.1.1 ; python_version >= "3.9" and python_version < "4.0" +pygments==2.18.0 ; python_version >= "3.9" and python_version < "4.0" +pymdown-extensions==10.8.1 ; python_version >= "3.9" and python_version < "4.0" +pyparsing==3.1.2 ; python_version >= "3.9" and python_version < "4.0" pyproj==3.6.1 ; python_version >= "3.9" and python_version < "4.0" pyshp==2.3.1 ; python_version >= "3.9" and python_version < "4.0" -pytest-cov==4.1.0 ; python_version >= "3.9" and python_version < "4.0" -pytest==8.0.0 ; python_version >= "3.9" and python_version < "4.0" -python-dateutil==2.8.2 ; python_version >= "3.9" and python_version < "4.0" +pytest-cov==5.0.0 ; python_version >= "3.9" and python_version < "4.0" +pytest==8.2.1 ; python_version >= "3.9" and python_version < "4.0" +python-dateutil==2.9.0.post0 ; python_version >= "3.9" and python_version < "4.0" pytz==2024.1 ; python_version >= "3.9" and python_version < "4.0" -pyviz-comms==3.0.1 ; python_version >= "3.9" and python_version < "4.0" +pyviz-comms==3.0.2 ; python_version >= "3.9" and python_version < "4.0" pywin32==306 ; sys_platform == "win32" and platform_python_implementation != "PyPy" and python_version >= "3.9" and python_version < "4.0" pyyaml-env-tag==0.1 ; python_version >= "3.9" and python_version < "4.0" pyyaml==6.0.1 ; python_version >= "3.9" and python_version < "4.0" -pyzmq==25.1.2 ; python_version >= "3.9" and python_version < "4.0" -referencing==0.33.0 ; python_version >= "3.9" and python_version < "4.0" -regex==2023.12.25 ; python_version >= "3.9" and python_version < "4.0" -requests==2.31.0 ; python_version >= "3.9" and python_version < "4.0" -rpds-py==0.17.1 ; python_version >= "3.9" and python_version < "4.0" -scipy==1.12.0 ; python_version >= "3.9" and python_version < "4.0" -setuptools==69.0.3 ; python_version >= "3.9" and python_version < "4.0" -shapely==2.0.2 ; python_version >= "3.9" and python_version < "4.0" +pyzmq==26.0.3 ; python_version >= "3.9" and python_version < "4.0" +referencing==0.35.1 ; python_version >= "3.9" and python_version < "4.0" +regex==2024.5.15 ; python_version >= "3.9" and python_version < "4.0" +requests==2.32.2 ; python_version >= "3.9" and python_version < "4.0" +rpds-py==0.18.1 ; python_version >= "3.9" and python_version < "4.0" +scipy==1.13.1 ; python_version >= "3.9" and python_version < "4.0" +shapely==2.0.4 ; python_version >= "3.9" and python_version < "4.0" six==1.16.0 ; python_version >= "3.9" and python_version < "4.0" sortedcontainers==2.4.0 ; python_version >= "3.9" and python_version < "4.0" soupsieve==2.5 ; python_version < "3.10" and python_version >= "3.9" stack-data==0.6.3 ; python_version >= "3.9" and python_version < "4.0" tblib==3.0.0 ; python_version >= "3.9" and python_version < "4.0" -tenacity==8.2.3 ; python_version >= "3.9" and python_version < "4.0" +tenacity==8.3.0 ; python_version >= "3.9" and python_version < "4.0" tomli==2.0.1 ; python_version >= "3.9" and python_full_version <= "3.11.0a6" toolz==0.12.1 ; python_version >= "3.9" and python_version < "4.0" tornado==6.4 ; python_version >= "3.9" and python_version < "4.0" -tqdm==4.66.1 ; python_version >= "3.9" and python_version < "4.0" -traitlets==5.14.1 ; python_version >= "3.9" and python_version < "4.0" -types-decorator==5.1.8.20240106 ; python_version >= "3.9" and python_version < "4.0" -types-pytz==2024.1.0.20240203 ; python_version >= "3.9" and python_version < "4.0" -typing-extensions==4.9.0 ; python_version >= "3.9" and python_version < "4.0" -tzdata==2023.4 ; python_version >= "3.9" and python_version < "4.0" -uc-micro-py==1.0.2 ; python_version >= "3.9" and python_version < "4.0" -urllib3==2.2.0 ; python_version >= "3.9" and python_version < "4.0" -watchdog==3.0.0 ; python_version >= "3.9" and python_version < "4.0" +tqdm==4.66.4 ; python_version >= "3.9" and python_version < "4.0" +traitlets==5.14.3 ; python_version >= "3.9" and python_version < "4.0" +types-decorator==5.1.8.20240310 ; python_version >= "3.9" and python_version < "4.0" +types-pytz==2024.1.0.20240417 ; python_version >= "3.9" and python_version < "4.0" +typing-extensions==4.12.0 ; python_version >= "3.9" and python_version < "4.0" +tzdata==2024.1 ; python_version >= "3.9" and python_version < "4.0" +uc-micro-py==1.0.3 ; python_version >= "3.9" and python_version < "4.0" +urllib3==2.2.1 ; python_version >= "3.9" and python_version < "4.0" +watchdog==4.0.1 ; python_version >= "3.9" and python_version < "4.0" wcwidth==0.2.13 ; python_version >= "3.9" and python_version < "4.0" webencodings==0.5.1 ; python_version >= "3.9" and python_version < "4.0" webob==1.8.7 ; python_version >= "3.9" and python_version < "3.10" -xarray==2024.1.1 ; python_version >= "3.9" and python_version < "4.0" -xarray[accel,io]==2024.1.1 ; python_version >= "3.9" and python_version < "4.0" -xyzservices==2023.10.1 ; python_version >= "3.9" and python_version < "4.0" +xarray-selafin==0.1.6 ; python_version >= "3.9" and python_version < "4.0" +xarray==2024.5.0 ; python_version >= "3.9" and python_version < "4.0" +xarray[accel,io]==2024.5.0 ; python_version >= "3.9" and python_version < "4.0" +xyzservices==2024.4.0 ; python_version >= "3.9" and python_version < "4.0" yarl==1.9.4 ; python_version == "3.12" -zarr==2.16.1 ; python_version >= "3.9" and python_version < "4.0" +zarr==2.18.1 ; python_version >= "3.9" and python_version < "4.0" zict==3.0.0 ; python_version >= "3.9" and python_version < "4.0" -zipp==3.17.0 ; python_version >= "3.9" and python_version < "4.0" +zipp==3.18.2 ; python_version >= "3.9" and python_version < "3.12" diff --git a/requirements/requirements.txt b/requirements/requirements.txt index ce1db99..77eac06 100644 --- a/requirements/requirements.txt +++ b/requirements/requirements.txt @@ -2,9 +2,9 @@ asciitree==0.3.3 ; python_version >= "3.9" and python_version < "4.0" attrs==23.2.0 ; python_version >= "3.9" and python_version < "4.0" beautifulsoup4==4.12.3 ; python_version < "3.10" and python_version >= "3.9" bleach==6.1.0 ; python_version >= "3.9" and python_version < "4.0" -bokeh==3.3.4 ; python_version >= "3.9" and python_version < "4.0" -bottleneck==1.3.7 ; python_version >= "3.9" and python_version < "4.0" -cartopy==0.22.0 ; python_version >= "3.9" and python_version < "4.0" +bokeh==3.4.1 ; python_version >= "3.9" and python_version < "4.0" +bottleneck==1.3.8 ; python_version >= "3.9" and python_version < "4.0" +cartopy==0.23.0 ; python_version >= "3.9" and python_version < "4.0" certifi==2024.2.2 ; python_version >= "3.9" and python_version < "4.0" cftime==1.6.3 ; python_version >= "3.9" and python_version < "4.0" charset-normalizer==3.3.2 ; python_version >= "3.9" and python_version < "4.0" @@ -13,90 +13,90 @@ click==8.1.7 ; python_version >= "3.9" and python_version < "4.0" cligj==0.7.2 ; python_version >= "3.9" and python_version < "4" cloudpickle==3.0.0 ; python_version >= "3.9" and python_version < "4.0" colorama==0.4.6 ; python_version >= "3.9" and python_version < "4.0" and platform_system == "Windows" -colorcet==3.0.1 ; python_version >= "3.9" and python_version < "4.0" -contourpy==1.2.0 ; python_version >= "3.9" and python_version < "4.0" +colorcet==3.1.0 ; python_version >= "3.9" and python_version < "4.0" +contourpy==1.2.1 ; python_version >= "3.9" and python_version < "4.0" cycler==0.12.1 ; python_version >= "3.9" and python_version < "4.0" -dask==2024.1.1 ; python_version >= "3.9" and python_version < "4.0" -dask[array,complete,dataframe,diagnostics,distributed]==2024.1.1 ; python_version >= "3.9" and python_version < "4.0" -datashader==0.16.0 ; python_version >= "3.9" and python_version < "4.0" -distributed==2024.1.1 ; python_version >= "3.9" and python_version < "4.0" +dask-expr==1.1.1 ; python_version >= "3.9" and python_version < "4.0" +dask==2024.5.1 ; python_version >= "3.9" and python_version < "4.0" +dask[array,complete,dataframe,diagnostics,distributed]==2024.5.1 ; python_version >= "3.9" and python_version < "4.0" +datashader==0.16.1 ; python_version >= "3.9" and python_version < "4.0" +distributed==2024.5.1 ; python_version >= "3.9" and python_version < "4.0" docopt==0.6.2 ; python_version < "3.10" and python_version >= "3.9" -fasteners==0.19 ; python_version >= "3.9" and python_version < "4.0" -fiona==1.9.5 ; python_version >= "3.9" and python_version < "4.0" -flox==0.9.0 ; python_version >= "3.9" and python_version < "4.0" -fonttools==4.48.1 ; python_version >= "3.9" and python_version < "4.0" -fsspec==2024.2.0 ; python_version >= "3.9" and python_version < "4.0" -future==0.18.3 ; python_version >= "3.9" and python_version < "4.0" -geopandas==0.14.3 ; python_version >= "3.9" and python_version < "4.0" -geoviews==1.11.0 ; python_version >= "3.9" and python_version < "4.0" +fasteners==0.19 ; python_version >= "3.9" and python_version < "4.0" and sys_platform != "emscripten" +fiona==1.9.6 ; python_version >= "3.9" and python_version < "4.0" +flox==0.9.7 ; python_version >= "3.9" and python_version < "4.0" +fonttools==4.51.0 ; python_version >= "3.9" and python_version < "4.0" +fsspec==2024.5.0 ; python_version >= "3.9" and python_version < "4.0" +future==1.0.0 ; python_version >= "3.9" and python_version < "4.0" +geopandas==0.14.4 ; python_version >= "3.9" and python_version < "4.0" +geoviews==1.12.0 ; python_version >= "3.9" and python_version < "4.0" h5netcdf==1.3.0 ; python_version >= "3.9" and python_version < "4.0" -h5py==3.10.0 ; python_version >= "3.9" and python_version < "4.0" -holoviews==1.18.1 ; python_version >= "3.9" and python_version < "4.0" -idna==3.6 ; python_version >= "3.9" and python_version < "4.0" -importlib-metadata==7.0.1 ; python_version >= "3.9" and python_version < "4.0" -importlib-resources==6.1.1 ; python_version >= "3.9" and python_version < "3.10" -jinja2==3.1.3 ; python_version >= "3.9" and python_version < "4.0" +h5py==3.11.0 ; python_version >= "3.9" and python_version < "4.0" +holoviews==1.18.3 ; python_version >= "3.9" and python_version < "4.0" +idna==3.7 ; python_version >= "3.9" and python_version < "4.0" +importlib-metadata==7.1.0 ; python_version >= "3.9" and python_version < "3.12" +importlib-resources==6.4.0 ; python_version >= "3.9" and python_version < "3.10" +jinja2==3.1.4 ; python_version >= "3.9" and python_version < "4.0" kiwisolver==1.4.5 ; python_version >= "3.9" and python_version < "4.0" linkify-it-py==2.0.3 ; python_version >= "3.9" and python_version < "4.0" llvmlite==0.42.0 ; python_version >= "3.9" and python_version < "4.0" locket==1.0.0 ; python_version >= "3.9" and python_version < "4.0" lz4==4.3.3 ; python_version >= "3.9" and python_version < "4.0" markdown-it-py==3.0.0 ; python_version >= "3.9" and python_version < "4.0" -markdown==3.5.2 ; python_version >= "3.9" and python_version < "4.0" +markdown==3.6 ; python_version >= "3.9" and python_version < "4.0" markupsafe==2.1.5 ; python_version >= "3.9" and python_version < "4.0" -matplotlib==3.8.2 ; python_version >= "3.9" and python_version < "4.0" -mdit-py-plugins==0.4.0 ; python_version >= "3.9" and python_version < "4.0" +matplotlib==3.9.0 ; python_version >= "3.9" and python_version < "4.0" +mdit-py-plugins==0.4.1 ; python_version >= "3.9" and python_version < "4.0" mdurl==0.1.2 ; python_version >= "3.9" and python_version < "4.0" -msgpack==1.0.7 ; python_version >= "3.9" and python_version < "4.0" +msgpack==1.0.8 ; python_version >= "3.9" and python_version < "4.0" multipledispatch==1.0.0 ; python_version >= "3.9" and python_version < "4.0" netcdf4==1.6.5 ; python_version >= "3.9" and python_version < "4.0" -numba==0.59.0 ; python_version >= "3.9" and python_version < "4.0" -numbagg==0.8.0 ; python_version >= "3.9" and python_version < "4.0" +numba==0.59.1 ; python_version >= "3.9" and python_version < "4.0" +numbagg==0.8.1 ; python_version >= "3.9" and python_version < "4.0" numcodecs==0.12.1 ; python_version >= "3.9" and python_version < "4.0" -numpy-groupies==0.10.2 ; python_version >= "3.9" and python_version < "4.0" +numpy-groupies==0.11.1 ; python_version >= "3.9" and python_version < "4.0" numpy-indexed==0.3.7 ; python_version >= "3.9" and python_version < "4.0" numpy==1.26.4 ; python_version >= "3.9" and python_version < "4.0" opt-einsum==3.3.0 ; python_version >= "3.9" and python_version < "4.0" -packaging==23.2 ; python_version >= "3.9" and python_version < "4.0" -pandas==2.2.0 ; python_version >= "3.9" and python_version < "4.0" -panel==1.3.8 ; python_version >= "3.9" and python_version < "4.0" -param==2.0.2 ; python_version >= "3.9" and python_version < "4.0" -partd==1.4.1 ; python_version >= "3.9" and python_version < "4.0" -pillow==10.2.0 ; python_version >= "3.9" and python_version < "4.0" -platformdirs==4.2.0 ; python_version >= "3.9" and python_version < "4.0" -pooch==1.8.0 ; python_version >= "3.9" and python_version < "4.0" +packaging==24.0 ; python_version >= "3.9" and python_version < "4.0" +pandas==2.2.2 ; python_version >= "3.9" and python_version < "4.0" +panel==1.4.3 ; python_version >= "3.9" and python_version < "4.0" +param==2.1.0 ; python_version >= "3.9" and python_version < "4.0" +partd==1.4.2 ; python_version >= "3.9" and python_version < "4.0" +pillow==10.3.0 ; python_version >= "3.9" and python_version < "4.0" +platformdirs==4.2.2 ; python_version >= "3.9" and python_version < "4.0" +pooch==1.8.1 ; python_version >= "3.9" and python_version < "4.0" psutil==5.9.8 ; python_version >= "3.9" and python_version < "4.0" pyarrow-hotfix==0.6 ; python_version >= "3.9" and python_version < "4.0" -pyarrow==15.0.0 ; python_version >= "3.9" and python_version < "4.0" +pyarrow==16.1.0 ; python_version >= "3.9" and python_version < "4.0" pyct==0.5.0 ; python_version >= "3.9" and python_version < "4.0" pydap==3.4.1 ; python_version < "3.10" and python_version >= "3.9" -pyparsing==3.1.1 ; python_version >= "3.9" and python_version < "4.0" +pyparsing==3.1.2 ; python_version >= "3.9" and python_version < "4.0" pyproj==3.6.1 ; python_version >= "3.9" and python_version < "4.0" pyshp==2.3.1 ; python_version >= "3.9" and python_version < "4.0" -python-dateutil==2.8.2 ; python_version >= "3.9" and python_version < "4.0" +python-dateutil==2.9.0.post0 ; python_version >= "3.9" and python_version < "4.0" pytz==2024.1 ; python_version >= "3.9" and python_version < "4.0" -pyviz-comms==3.0.1 ; python_version >= "3.9" and python_version < "4.0" +pyviz-comms==3.0.2 ; python_version >= "3.9" and python_version < "4.0" pyyaml==6.0.1 ; python_version >= "3.9" and python_version < "4.0" -requests==2.31.0 ; python_version >= "3.9" and python_version < "4.0" -scipy==1.12.0 ; python_version >= "3.9" and python_version < "4.0" -setuptools==69.0.3 ; python_version >= "3.9" and python_version < "4.0" -shapely==2.0.2 ; python_version >= "3.9" and python_version < "4.0" +requests==2.32.2 ; python_version >= "3.9" and python_version < "4.0" +scipy==1.13.1 ; python_version >= "3.9" and python_version < "4.0" +shapely==2.0.4 ; python_version >= "3.9" and python_version < "4.0" six==1.16.0 ; python_version >= "3.9" and python_version < "4.0" sortedcontainers==2.4.0 ; python_version >= "3.9" and python_version < "4.0" soupsieve==2.5 ; python_version < "3.10" and python_version >= "3.9" tblib==3.0.0 ; python_version >= "3.9" and python_version < "4.0" toolz==0.12.1 ; python_version >= "3.9" and python_version < "4.0" tornado==6.4 ; python_version >= "3.9" and python_version < "4.0" -tqdm==4.66.1 ; python_version >= "3.9" and python_version < "4.0" -typing-extensions==4.9.0 ; python_version >= "3.9" and python_version < "4.0" -tzdata==2023.4 ; python_version >= "3.9" and python_version < "4.0" -uc-micro-py==1.0.2 ; python_version >= "3.9" and python_version < "4.0" -urllib3==2.2.0 ; python_version >= "3.9" and python_version < "4.0" +tqdm==4.66.4 ; python_version >= "3.9" and python_version < "4.0" +typing-extensions==4.12.0 ; python_version >= "3.9" and python_version < "4.0" +tzdata==2024.1 ; python_version >= "3.9" and python_version < "4.0" +uc-micro-py==1.0.3 ; python_version >= "3.9" and python_version < "4.0" +urllib3==2.2.1 ; python_version >= "3.9" and python_version < "4.0" webencodings==0.5.1 ; python_version >= "3.9" and python_version < "4.0" webob==1.8.7 ; python_version >= "3.9" and python_version < "3.10" -xarray==2024.1.1 ; python_version >= "3.9" and python_version < "4.0" -xarray[accel,io]==2024.1.1 ; python_version >= "3.9" and python_version < "4.0" -xyzservices==2023.10.1 ; python_version >= "3.9" and python_version < "4.0" -zarr==2.16.1 ; python_version >= "3.9" and python_version < "4.0" +xarray==2024.5.0 ; python_version >= "3.9" and python_version < "4.0" +xarray[accel,io]==2024.5.0 ; python_version >= "3.9" and python_version < "4.0" +xyzservices==2024.4.0 ; python_version >= "3.9" and python_version < "4.0" +zarr==2.18.1 ; python_version >= "3.9" and python_version < "4.0" zict==3.0.0 ; python_version >= "3.9" and python_version < "4.0" -zipp==3.17.0 ; python_version >= "3.9" and python_version < "4.0" +zipp==3.18.2 ; python_version >= "3.9" and python_version < "3.12" diff --git a/thalassa/api.py b/thalassa/api.py index f25d6bf..564a03b 100644 --- a/thalassa/api.py +++ b/thalassa/api.py @@ -67,14 +67,14 @@ def open_dataset( Parameters: path: The path to the dataset file (netCDF, zarr, grib) normalize: Boolean flag indicating whether the dataset should be converted/normalized to the "Thalassa schema". - Normalization is currently only supported for ``SCHISM`` and ``ADCIRC`` netcdf files. + Normalization is currently only supported for ``SCHISM``, ``TELEMAC``, and ``ADCIRC`` netcdf files. kwargs: The ``kwargs`` are being passed through to ``xarray.open_dataset``. """ import xarray as xr default_kwargs: dict[str, T.Any] = dict( - mask_and_scale=True, + # mask_and_scale=True, cache=False, drop_variables=ADCIRC_VARIABLES_TO_BE_DROPPED, ) diff --git a/thalassa/normalization.py b/thalassa/normalization.py index cd4855f..df2d306 100644 --- a/thalassa/normalization.py +++ b/thalassa/normalization.py @@ -19,19 +19,29 @@ class THALASSA_FORMATS(enum.Enum): UNKNOWN = "UNKNOWN" ADCIRC = "ADCIRC" SCHISM = "SCHISM" + TELEMAC = "TELEMAC" GENERIC = "GENERIC" PYPOSEIDON = "PYPOSEIDON" # fmt: off +EDGE_DIM = "edge" +FACE_DIM = "face" +NODE_DIM = "node" +VERTICAL_DIM = "layer" +CONNECTIVITY = "face_nodes" +VERTICE_DIM = "max_no_vertices" +X_DIM = "lon" +Y_DIM = "lat" + _GENERIC_DIMS = { - "node", + NODE_DIM, "triface", "three", } _GENERIC_VARS = { - "lon", - "lat", + X_DIM, + Y_DIM, "triface_nodes", } _SCHISM_DIMS = { @@ -45,6 +55,14 @@ class THALASSA_FORMATS(enum.Enum): "SCHISM_hgrid_node_y", "SCHISM_hgrid_face_nodes", } +_TELEMAC_DIMS = { + "node", +} +_TELEMAC_VARS = { + "x", + "y", + "ikle2", +} _PYPOSEIDON_DIMS = { "nSCHISM_hgrid_face", "nSCHISM_hgrid_node", @@ -76,6 +94,11 @@ def is_schism(ds: xarray.Dataset) -> bool: return _SCHISM_DIMS.issubset(ds.dims) and _SCHISM_VARS.issubset(total_vars) +def is_telemac(ds: xarray.Dataset) -> bool: + total_vars = list(ds.data_vars.keys()) + list(ds.coords.keys()) + list(ds.attrs.keys()) + return _TELEMAC_DIMS.issubset(ds.dims) and _TELEMAC_VARS.issubset(total_vars) + + def is_pyposeidon(ds: xarray.Dataset) -> bool: return _PYPOSEIDON_DIMS.issubset(ds.dims) and _PYPOSEIDON_VARS.issubset(ds.data_vars) @@ -87,6 +110,8 @@ def is_adcirc(ds: xarray.Dataset) -> bool: def infer_format(ds: xarray.Dataset) -> THALASSA_FORMATS: if is_schism(ds): fmt = THALASSA_FORMATS.SCHISM + elif is_telemac(ds): + fmt = THALASSA_FORMATS.TELEMAC elif is_pyposeidon(ds): fmt = THALASSA_FORMATS.PYPOSEIDON elif is_generic(ds): @@ -120,37 +145,58 @@ def normalize_generic(ds: xarray.Dataset) -> xarray.Dataset: def normalize_schism(ds: xarray.Dataset) -> xarray.Dataset: ds = ds.rename( { - "nSCHISM_hgrid_edge": "edge", - "nSCHISM_hgrid_face": "face", - "nSCHISM_hgrid_node": "node", - "SCHISM_hgrid_face_nodes": "face_nodes", - "nMaxSCHISM_hgrid_face_nodes": "max_no_vertices", - "SCHISM_hgrid_node_x": "lon", - "SCHISM_hgrid_node_y": "lat", + "nSCHISM_hgrid_edge": EDGE_DIM, + "nSCHISM_hgrid_face": FACE_DIM, + "nSCHISM_hgrid_node": NODE_DIM, + "SCHISM_hgrid_face_nodes": CONNECTIVITY, + "nMaxSCHISM_hgrid_face_nodes": VERTICE_DIM, + "SCHISM_hgrid_node_x": X_DIM, + "SCHISM_hgrid_node_y": Y_DIM, }, ) if "nSCHISM_vgrid_layers" in ds.dims: # I.e. OLD Schism IO or "merged" new IO ds = ds.rename( { - "nSCHISM_vgrid_layers": "layer", + "nSCHISM_vgrid_layers": VERTICAL_DIM, }, ) # SCHISM output uses one-based indices for `face_nodes` # Let's ensure that we use zero-based indices everywhere. - ds["face_nodes"] -= 1 + ds[CONNECTIVITY] -= 1 + return ds + + +def normalize_telemac(ds: xarray.Dataset) -> xarray.Dataset: + ds = ds.rename( + { + "node": NODE_DIM, + "x": X_DIM, + "y": Y_DIM, + }, + ) + if "plan" in ds.dims: + ds = ds.rename( + { + "plan": VERTICAL_DIM, + }, + ) + + # TELEMAC output uses one-based indices for `face_nodes` + # Let's ensure that we use zero-based indices everywhere. + ds[CONNECTIVITY] = ((FACE_DIM, VERTICE_DIM), ds.attrs['ikle2'] - 1) return ds def normalize_pyposeidon(ds: xarray.Dataset) -> xarray.Dataset: ds = ds.rename( { - "nSCHISM_hgrid_face": "face", - "nSCHISM_hgrid_node": "node", - "SCHISM_hgrid_face_nodes": "face_nodes", - "nMaxSCHISM_hgrid_face_nodes": "max_no_vertices", - "SCHISM_hgrid_node_x": "lon", - "SCHISM_hgrid_node_y": "lat", + "nSCHISM_hgrid_face": FACE_DIM, + "nSCHISM_hgrid_node": NODE_DIM, + "SCHISM_hgrid_face_nodes": CONNECTIVITY, + "nMaxSCHISM_hgrid_face_nodes": VERTICE_DIM, + "SCHISM_hgrid_node_x": X_DIM, + "SCHISM_hgrid_node_y": Y_DIM, }, ) return ds @@ -159,16 +205,16 @@ def normalize_pyposeidon(ds: xarray.Dataset) -> xarray.Dataset: def normalize_adcirc(ds: xarray.Dataset) -> xarray.Dataset: ds = ds.rename( { - "x": "lon", - "y": "lat", - "element": "face_nodes", - "nvertex": "max_no_vertices", - "nele": "face", + "x": X_DIM, + "y": Y_DIM, + "element": CONNECTIVITY, + "nvertex": VERTICE_DIM, + "nele": FACE_DIM, }, ) # ADCIRC output uses one-based indices for `face_nodes` # Let's ensure that we use zero-based indices everywhere. - ds["face_nodes"] -= 1 + ds[CONNECTIVITY] -= 1 return ds @@ -176,6 +222,7 @@ def normalize_adcirc(ds: xarray.Dataset) -> xarray.Dataset: THALASSA_FORMATS.ADCIRC: normalize_adcirc, THALASSA_FORMATS.GENERIC: normalize_generic, THALASSA_FORMATS.SCHISM: normalize_schism, + THALASSA_FORMATS.TELEMAC: normalize_telemac, THALASSA_FORMATS.PYPOSEIDON: normalize_pyposeidon, }

normalisation for GENERIC formats -- naming conventions This issue is more to open a discussion on naming conventions. My Dataset has the following format: ``` Dimensions: time: 7272 node: 301572 face: 588314 max_no_vertices: 3 Coordinates: longitude (node) float32 ... latitude (node) float32 ... time (time) datetime64[ns] 2023-01-01 ... 2023-10-30T23:00:00 Data variables: u (time, node) float32 ... v (time, node) float32 ... elev (time, node) float32 ... face_nodes (face, max_no_vertices) int32 ... ``` When I try to plot: ```python plot_map = thalassa.plot( ds=ds.isel(time=-1), variable="max_elev", clim_min=0, clim_max=1, clabel="meters", cmap = 'jet' ) ``` I get the following error: ``` KeyError Traceback (most recent call last) Cell In[22], [line 1](vscode-notebook-cell:?execution_count=22&line=1) ----> [1](vscode-notebook-cell:?execution_count=22&line=1) plot_map = thalassa.plot( [2](vscode-notebook-cell:?execution_count=22&line=2) ds=ds.isel(time=-1), [3](vscode-notebook-cell:?execution_count=22&line=3) variable="max_elev", [4](vscode-notebook-cell:?execution_count=22&line=4) clim_min=0, [5](vscode-notebook-cell:?execution_count=22&line=5) clim_max=1, [6](vscode-notebook-cell:?execution_count=22&line=6) clabel="meters", [7](vscode-notebook-cell:?execution_count=22&line=7) cmap = 'jet' [8](vscode-notebook-cell:?execution_count=22&line=8) ) File ~/work/gist/Validation_Protocol/.venv/lib/python3.11/site-packages/thalassa/plotting.py:153, in plot(ds, variable, bbox, title, cmap, colorbar, clabel, clim_min, clim_max, x_range, y_range, show_mesh) [93].venv/lib/python3.11/site-packages/thalassa/plotting.py:93) """ [94].venv/lib/python3.11/site-packages/thalassa/plotting.py:94) Return the plot of the specified `variable`. [95].venv/lib/python3.11/site-packages/thalassa/plotting.py:95) (...) [149].venv/lib/python3.11/site-packages/thalassa/plotting.py:149) [150].venv/lib/python3.11/site-packages/thalassa/plotting.py:150) """ [151].venv/lib/python3.11/site-packages/thalassa/plotting.py:151) import holoviews as hv --> [153].venv/lib/python3.11/site-packages/thalassa/plotting.py:153) ds = normalization.normalize(ds) [154].venv/lib/python3.11/site-packages/thalassa/plotting.py:154) _sanity_check(ds=ds, variable=variable) [155].venv/lib/python3.11/site-packages/thalassa/plotting.py:155) if bbox: File ~/work/gist/Validation_Protocol/.venv/lib/python3.11/site-packages/thalassa/normalization.py:203, in normalize(ds) [201].venv/lib/python3.11/site-packages/thalassa/normalization.py:201) logger.debug("Dataset normalization: Started") [202].venv/lib/python3.11/site-packages/thalassa/normalization.py:202) fmt = infer_format(ds) --> [203].venv/lib/python3.11/site-packages/thalassa/normalization.py:203) normalizer_func = NORMALIZE_DISPATCHER[fmt] [204].venv/lib/python3.11/site-packages/thalassa/normalization.py:204) normalized_ds = normalizer_func(ds) [205].venv/lib/python3.11/site-packages/thalassa/normalization.py:205) # Handle quad elements [206].venv/lib/python3.11/site-packages/thalassa/normalization.py:206) # Splitting quad elements to triangles, means that the number of faces increases [207].venv/lib/python3.11/site-packages/thalassa/normalization.py:207) # There are two options: (...) [210].venv/lib/python3.11/site-packages/thalassa/normalization.py:210) # I'd rather avoid altering the values of the provided netcdf file therefore we go for option #2, [211].venv/lib/python3.11/site-packages/thalassa/normalization.py:211) # i.e. we create the `triface_nodes` variable. KeyError: <THALASSA_FORMATS.UNKNOWN: 'UNKNOWN'> ``` That is caused by: ```python ds = thalassa.normalization.normalize(ds) ``` I found the solution by making my dataset compliant with the `GENERIC` [format](https://github.com/ec-jrc/Thalassa/blob/b9d977cd6999e73f5ad884e9e7b96d4041b60827/thalassa/normalization.py#L27): ```python ds = ds.rename({ "longitude": "lon", "latitude": "lat", "face_nodes": "triface_nodes", "max_no_vertices": "three", "face": "triface", }) # make variable compliant with thalassa ``` I also understood the convention `face_nodes` / `triface_nodes` is before / after the Quad elements [test](https://github.com/ec-jrc/Thalassa/blob/b9d977cd6999e73f5ad884e9e7b96d4041b60827/thalassa/normalization.py#L212). ## concerning the connectivity of the mesh: Would it be possible to skip the quad element test if you know already that the mesh is an unstructured triangle grid? Also, I think ideally the connectivity of mesh should be renamed to `mesh_topology`, according to [UGRID](https://ugrid-conventions.github.io/ugrid-conventions/#2d-triangular-mesh-topology) and [cf_xarray](https://cf-xarray.readthedocs.io/en/latest/sgrid_ugrid.html#ugrid) conventions ## concerning the coordinates Would it acceptable to consider other names for coordinates ? like `Lon`, `x` or `longitude`, or should this aspect be dealt by the user?

Is this output from Telemac? If yes then the easiest way to add support would be to add an extra "schema" in `normalization.py`. I think that this entails defining: - `is_telemac()` - `normalize_telemac` and then updating: - `infer_format()` - `NORMALIZE_DISPATCHER` If you can create a small file, then add a test too: https://github.com/ec-jrc/Thalassa/blob/b9d977cd6999e73f5ad884e9e7b96d4041b60827/tests/normalization_test.py The reason we have `triface_nodes` etc is because: 1. when it comes to unstructured meshes, Holoviews only supports triangles 2. we want to be able to support STOFS-3D output too, which uses quads. So the idea is that we keep the original mesh connectivity as is, and we create a new variable that holds the triangle-only connectivity. If there are no quads, then the new variable is just a copy of the original. The thalassa code internally uses the new variable everywhere. WRT CF naming-conventions etc, we can discuss it at https://github.com/ec-jrc/Thalassa/issues/80 Just to make it clear, we could add support for UGRID compliant netcdfs too. I am for sure not against that, but it is probably going to be more work compared to just updating the current setup. But feel free to investigate it if you want. It is for sure going to be useful in the context of seamesh etc. > Is this output from Telemac? From TELEMAC yes, but through [pyPoseidon](https://github.com/ec-jrc/pyPoseidon/blob/92a797eb396bc5263fd95d94f93c088450056a65/pyposeidon/telemac.py#L1153) where I convert from Selafin file to xarray Dataset. All details from one format to other is in the [`norm.py`](https://github.com/ec-jrc/pyPoseidon/blob/telemac/pyposeidon/utils/norm.py#L110) function of the `telemac` branch > the easiest way to add support would be to add an extra "schema" in normalization.py. So there are 2 ways to solve this: 1. use `xarray-selafin` in `thalassa` to read directly from Selafins. But it might be too much abstraction for the user and also if `xarray-selafin` changes the backend, we'd have to adapt the TELEMAC bindings here too. 2. adapt the conversion convention in pyPoseidon A third option would be a combination of both. The quickest win is clearly option 2 though. The idea is to try to keep thalassa decoupled from pyposeidon. If you add the functionality in pyposeidon, then TELEMAC output will be visualizable by thalassa iff you post-process it with pyposeidon. If you add the functionality in thalassa, then TELEMAC output will be visualizable by thalassa regardless of the way you run/post-processed the results. That being said, if you truly want to make the TELEMAC output compatible with pyPoseidon, you will probably need to rename the TELEMAC variables/dims anyhow in order to make them compatible with the existing pyposeidon functionality (which uses SCHISM names, pretty much ubiquitusly).

2024-05-19T11:59:10

0.0

[]

GeoStat-Framework/GSTools

GeoStat-Framework__GSTools-340

90003fbbe193cfa835a891b403b4024bb21867f9

diff --git a/.github/workflows/main.yml b/.github/workflows/main.yml index e116c483..f8dc74cd 100644 --- a/.github/workflows/main.yml +++ b/.github/workflows/main.yml @@ -16,16 +16,14 @@ jobs: source_check: name: source check runs-on: ubuntu-latest - strategy: - fail-fast: false steps: - - uses: actions/checkout@v2 + - uses: actions/checkout@v4 with: fetch-depth: '0' - name: Set up Python 3.9 - uses: actions/setup-python@v2 + uses: actions/setup-python@v5 with: python-version: 3.9 @@ -38,6 +36,10 @@ jobs: run: | python -m black --check --diff --color . + - name: black preview + run: | + python -m black --preview --diff --color . + - name: isort check run: | python -m isort --check --diff --color . @@ -58,60 +60,68 @@ jobs: matrix: cfg: - { os: ubuntu-latest, arch: x86_64 } - - { os: ubuntu-latest, arch: i686 } - { os: windows-latest, arch: AMD64 } - - { os: windows-latest, arch: x86 } - { os: macos-latest, arch: x86_64 } - { os: macos-latest, arch: arm64 } - { os: macos-latest, arch: universal2 } steps: - - uses: actions/checkout@v2 + - uses: actions/checkout@v4 with: fetch-depth: '0' - name: Build wheels - uses: pypa/[email protected] + uses: pypa/[email protected] env: CIBW_ARCHS: ${{ matrix.cfg.arch }} with: output-dir: dist - - uses: actions/upload-artifact@v2 + - uses: actions/upload-artifact@v3 with: path: ./dist/*.whl build_sdist: - name: sdist on ${{ matrix.os }} with py ${{ matrix.python-version }} + name: sdist on ${{ matrix.os }} with py ${{ matrix.ver.py }} numpy${{ matrix.ver.np }} scipy${{ matrix.ver.sp }} runs-on: ${{ matrix.os }} strategy: fail-fast: false matrix: os: [ubuntu-latest, windows-latest, macos-latest] - python-version: ['3.8', '3.9', '3.10', '3.11', '3.12'] - + # https://github.com/scipy/oldest-supported-numpy/blob/main/setup.cfg + ver: + - {py: '3.8', np: '==1.20.0', sp: '==1.5.4'} + - {py: '3.9', np: '==1.20.0', sp: '==1.5.4'} + - {py: '3.10', np: '==1.21.6', sp: '==1.7.2'} + - {py: '3.11', np: '==1.23.2', sp: '==1.9.2'} + - {py: '3.12', np: '==1.26.2', sp: '==1.11.2'} + - {py: '3.12', np: '>=2.0.0rc1', sp: '>=1.13.0'} steps: - - uses: actions/checkout@v2 + - uses: actions/checkout@v4 with: fetch-depth: '0' - - name: Set up Python ${{ matrix.python-version }} - uses: actions/setup-python@v2 + - name: Set up Python ${{ matrix.ver.py }} + uses: actions/setup-python@v5 with: - python-version: ${{ matrix.python-version }} + python-version: ${{ matrix.ver.py }} - name: Install dependencies + run: | + python -m pip install --upgrade pip + pip install build "coveralls>=3.0.0" + + - name: Install GSTools env: GSTOOLS_BUILD_PARALLEL: 1 run: | - python -m pip install --upgrade pip - pip install build coveralls>=3.0.0 pip install -v --editable .[test] - name: Run tests env: GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }} run: | + pip install "numpy${{ matrix.ver.np }}" "scipy${{ matrix.ver.sp }}" python -m pytest --cov gstools --cov-report term-missing -v tests/ python -m coveralls --service=github @@ -120,8 +130,8 @@ jobs: # PEP 517 package builder from pypa python -m build --sdist --outdir dist . - - uses: actions/upload-artifact@v2 - if: matrix.os == 'ubuntu-latest' && matrix.python-version == '3.9' + - uses: actions/upload-artifact@v3 + if: matrix.os == 'ubuntu-latest' && matrix.ver.py == '3.9' with: path: dist/*.tar.gz @@ -130,7 +140,7 @@ jobs: runs-on: ubuntu-latest steps: - - uses: actions/download-artifact@v2 + - uses: actions/download-artifact@v3 with: name: artifact path: dist diff --git a/.gitignore b/.gitignore index 872d3f40..bcdc980b 100644 --- a/.gitignore +++ b/.gitignore @@ -117,6 +117,7 @@ src/gstools/_version.py # generated docs docs/source/examples/ docs/source/api/ +docs/source/sg_execution_times.rst # other settings .vscode/ diff --git a/CHANGELOG.md b/CHANGELOG.md index 00513d23..61601c72 100755 --- a/CHANGELOG.md +++ b/CHANGELOG.md @@ -2,6 +2,25 @@ All notable changes to **GSTools** will be documented in this file. +## [Unreleased] - ? + +### Enhancements + +- added global variable `config.NUM_THREADS` to select number of threads for parallel computation ([#336](https://github.com/GeoStat-Framework/GSTools/pull/336)) +- prepare numpy 2 support ([#340](https://github.com/GeoStat-Framework/GSTools/pull/340)) + - numpy 2.0.0rc1 for building extensions (for Python 3.9 and above) + - check multiple numpy and scipy versions in CI + - fixed minimal versions for numpy + - use `np.asarray` everywhere with `np.atleast_(n)d` + - fix long/longlong integer issue in cython on windows by always using 64bit integers + +### Bugfixes +- build docs with latest sphinx version ([#340](https://github.com/GeoStat-Framework/GSTools/pull/340)) + +### Changes +- require pyvista 0.40 at least ([#340](https://github.com/GeoStat-Framework/GSTools/pull/340)) + + ## [1.5.1] - Nifty Neon - 2023-11 ### Enhancements diff --git a/README.md b/README.md index 69f1f3ec..6cb69901 100644 --- a/README.md +++ b/README.md @@ -345,7 +345,7 @@ in memory for immediate 3D plotting in Python. ## Requirements: -- [NumPy >= 1.14.5](https://www.numpy.org) +- [NumPy >= 1.20.0](https://www.numpy.org) - [SciPy >= 1.1.0](https://www.scipy.org/scipylib) - [hankel >= 1.0.0](https://github.com/steven-murray/hankel) - [emcee >= 3.0.0](https://github.com/dfm/emcee) @@ -366,7 +366,7 @@ You can contact us via <[email protected]>. ## License -[LGPLv3][license_link] © 2018-2021 +[LGPLv3][license_link] © 2018-2024 [pip_link]: https://pypi.org/project/gstools [conda_link]: https://docs.conda.io/en/latest/miniconda.html diff --git a/docs/source/conf.py b/docs/source/conf.py index eb5e5fae..e89928fc 100644 --- a/docs/source/conf.py +++ b/docs/source/conf.py @@ -259,10 +259,9 @@ def setup(app): # Example configuration for intersphinx: refer to the Python standard library. intersphinx_mapping = { "Python": ("https://docs.python.org/", None), - "NumPy": ("http://docs.scipy.org/doc/numpy/", None), - "SciPy": ("http://docs.scipy.org/doc/scipy/reference", None), - "matplotlib": ("http://matplotlib.org", None), - "Sphinx": ("http://www.sphinx-doc.org/en/stable/", None), + "NumPy": ("https://numpy.org/doc/stable/", None), + "SciPy": ("https://docs.scipy.org/doc/scipy/", None), + "matplotlib": ("https://matplotlib.org/stable/", None), "hankel": ("https://hankel.readthedocs.io/en/latest/", None), "emcee": ("https://emcee.readthedocs.io/en/latest/", None), } diff --git a/docs/source/index.rst b/docs/source/index.rst index d71494ac..ecad0583 100644 --- a/docs/source/index.rst +++ b/docs/source/index.rst @@ -414,7 +414,7 @@ in memory for immediate 3D plotting in Python. Requirements ============ -- `Numpy >= 1.14.5 <http://www.numpy.org>`_ +- `NumPy >= 1.20.0 <http://www.numpy.org>`_ - `SciPy >= 1.1.0 <http://www.scipy.org>`_ - `hankel >= 1.0.0 <https://github.com/steven-murray/hankel>`_ - `emcee >= 3.0.0 <https://github.com/dfm/emcee>`_ diff --git a/examples/00_misc/00_tpl_stable.py b/examples/00_misc/00_tpl_stable.py index 4d1a89d8..474b0f55 100644 --- a/examples/00_misc/00_tpl_stable.py +++ b/examples/00_misc/00_tpl_stable.py @@ -40,6 +40,7 @@ \sigma^2_{\ell_{\mathrm{up}}} &= C\cdot\frac{\ell_{\mathrm{up}}^{2H}}{2H} """ + import numpy as np import gstools as gs diff --git a/examples/00_misc/01_export.py b/examples/00_misc/01_export.py index ad754abf..e38294fe 100644 --- a/examples/00_misc/01_export.py +++ b/examples/00_misc/01_export.py @@ -7,6 +7,7 @@ These can be viewed for example with `Paraview <https://www.paraview.org/>`__. """ + # sphinx_gallery_thumbnail_path = 'pics/paraview.png' import gstools as gs diff --git a/examples/00_misc/02_check_rand_meth_sampling.py b/examples/00_misc/02_check_rand_meth_sampling.py index 0b23a929..58d998b4 100644 --- a/examples/00_misc/02_check_rand_meth_sampling.py +++ b/examples/00_misc/02_check_rand_meth_sampling.py @@ -2,6 +2,7 @@ Check Random Sampling --------------------- """ + import numpy as np from matplotlib import pyplot as plt from mpl_toolkits.mplot3d import Axes3D diff --git a/examples/00_misc/04_herten.py b/examples/00_misc/04_herten.py index d9e1b16e..1e1b8a23 100644 --- a/examples/00_misc/04_herten.py +++ b/examples/00_misc/04_herten.py @@ -30,6 +30,7 @@ functions, since the only produce the ``herten_transmissivity.gz`` and ``grid_dim_origin_spacing.txt``, which are already present. """ + import os import matplotlib.pyplot as plt diff --git a/examples/00_misc/05_standalone_field.py b/examples/00_misc/05_standalone_field.py index fb0c2082..e467f043 100644 --- a/examples/00_misc/05_standalone_field.py +++ b/examples/00_misc/05_standalone_field.py @@ -7,6 +7,7 @@ In the following example we will produce 10000 random points in 4D with random values and plot them. """ + import numpy as np import gstools as gs diff --git a/examples/01_random_field/03_unstr_srf_export.py b/examples/01_random_field/03_unstr_srf_export.py index e86e6f3e..94d00952 100644 --- a/examples/01_random_field/03_unstr_srf_export.py +++ b/examples/01_random_field/03_unstr_srf_export.py @@ -6,6 +6,7 @@ Normally, such a grid would be read in, but we can simply generate one and then create a random field at those coordinates. """ + import numpy as np import gstools as gs diff --git a/examples/01_random_field/04_srf_merge.py b/examples/01_random_field/04_srf_merge.py index 00aa6511..241ed079 100644 --- a/examples/01_random_field/04_srf_merge.py +++ b/examples/01_random_field/04_srf_merge.py @@ -6,6 +6,7 @@ to merge two unstructured rectangular fields. """ + # sphinx_gallery_thumbnail_number = 2 import numpy as np diff --git a/examples/01_random_field/06_pyvista_support.py b/examples/01_random_field/06_pyvista_support.py index 4cc371dd..29de8dd7 100644 --- a/examples/01_random_field/06_pyvista_support.py +++ b/examples/01_random_field/06_pyvista_support.py @@ -12,6 +12,7 @@ The :any:`Field.mesh` method enables easy field creation on PyVista meshes used by the :any:`SRF` or :any:`Krige` class. """ + # sphinx_gallery_thumbnail_path = 'pics/GS_pyvista_cut.png' import pyvista as pv @@ -21,8 +22,8 @@ # We create a structured grid with PyVista containing 50 segments on all three # axes each with a length of 2 (whatever unit). -dim, spacing = (50, 50, 50), (2, 2, 2) -grid = pv.UniformGrid(dim, spacing) +dims, spacing = (50, 50, 50), (2, 2, 2) +grid = pv.ImageData(dimensions=dims, spacing=spacing) ############################################################################### # Now we set up the SRF class as always. We'll use an anisotropic model. diff --git a/examples/02_cov_model/01_basic_methods.py b/examples/02_cov_model/01_basic_methods.py index fb9bfe7a..4c97fba8 100755 --- a/examples/02_cov_model/01_basic_methods.py +++ b/examples/02_cov_model/01_basic_methods.py @@ -36,6 +36,7 @@ correlation function as demonstrated in the introductory example. If one of the above functions is given, the others will be determined: """ + import gstools as gs model = gs.Exponential(dim=3, var=2.0, len_scale=10, nugget=0.5) diff --git a/examples/02_cov_model/02_aniso_rotation.py b/examples/02_cov_model/02_aniso_rotation.py index 2a8bac78..b7459e39 100755 --- a/examples/02_cov_model/02_aniso_rotation.py +++ b/examples/02_cov_model/02_aniso_rotation.py @@ -6,6 +6,7 @@ represents the isotropic case for the model. Nevertheless, you can provide anisotropy ratios by: """ + import gstools as gs model = gs.Gaussian(dim=3, var=2.0, len_scale=10, anis=0.5) diff --git a/examples/02_cov_model/03_spectral_methods.py b/examples/02_cov_model/03_spectral_methods.py index 677811a9..61c7e49b 100755 --- a/examples/02_cov_model/03_spectral_methods.py +++ b/examples/02_cov_model/03_spectral_methods.py @@ -23,6 +23,7 @@ You can access these methods by: """ + import gstools as gs model = gs.Gaussian(dim=3, var=2.0, len_scale=10) diff --git a/examples/02_cov_model/04_different_scales.py b/examples/02_cov_model/04_different_scales.py index cd6f4dee..0e2e1991 100755 --- a/examples/02_cov_model/04_different_scales.py +++ b/examples/02_cov_model/04_different_scales.py @@ -16,6 +16,7 @@ You can access it by: """ + import gstools as gs model = gs.Stable(dim=3, var=2.0, len_scale=10) diff --git a/examples/02_cov_model/05_additional_para.py b/examples/02_cov_model/05_additional_para.py index 02507d96..3264cec4 100755 --- a/examples/02_cov_model/05_additional_para.py +++ b/examples/02_cov_model/05_additional_para.py @@ -10,6 +10,7 @@ This leads to the so called **stable** covariance model and we can define it by """ + import numpy as np import gstools as gs diff --git a/examples/02_cov_model/06_fitting_para_ranges.py b/examples/02_cov_model/06_fitting_para_ranges.py index bd3d1196..7c8f083e 100755 --- a/examples/02_cov_model/06_fitting_para_ranges.py +++ b/examples/02_cov_model/06_fitting_para_ranges.py @@ -6,6 +6,7 @@ variogram data. In the following we will use the self defined stable model from a previous example. """ + import numpy as np import gstools as gs diff --git a/examples/03_variogram/00_fit_variogram.py b/examples/03_variogram/00_fit_variogram.py index 9dc1b305..7334ed2c 100644 --- a/examples/03_variogram/00_fit_variogram.py +++ b/examples/03_variogram/00_fit_variogram.py @@ -2,6 +2,7 @@ Fit Variogram ------------- """ + import numpy as np import gstools as gs diff --git a/examples/03_variogram/01_find_best_model.py b/examples/03_variogram/01_find_best_model.py index c0fc383f..eab031cf 100755 --- a/examples/03_variogram/01_find_best_model.py +++ b/examples/03_variogram/01_find_best_model.py @@ -2,6 +2,7 @@ Finding the best fitting variogram model ---------------------------------------- """ + import numpy as np from matplotlib import pyplot as plt diff --git a/examples/03_variogram/02_multi_vario.py b/examples/03_variogram/02_multi_vario.py index f893e452..71048849 100755 --- a/examples/03_variogram/02_multi_vario.py +++ b/examples/03_variogram/02_multi_vario.py @@ -5,6 +5,7 @@ In this example, we demonstrate how to estimate a variogram from multiple fields on the same point-set that should have the same statistical properties. """ + import matplotlib.pyplot as plt import numpy as np diff --git a/examples/03_variogram/03_directional_2d.py b/examples/03_variogram/03_directional_2d.py index e5744cb2..460c7513 100755 --- a/examples/03_variogram/03_directional_2d.py +++ b/examples/03_variogram/03_directional_2d.py @@ -7,6 +7,7 @@ Afterwards we will fit a model to this estimated variogram and show the result. """ + import numpy as np from matplotlib import pyplot as plt diff --git a/examples/03_variogram/04_directional_3d.py b/examples/03_variogram/04_directional_3d.py index 22ee7e9f..6a8b6ddf 100755 --- a/examples/03_variogram/04_directional_3d.py +++ b/examples/03_variogram/04_directional_3d.py @@ -7,6 +7,7 @@ Afterwards we will fit a model to this estimated variogram and show the result. """ + import matplotlib.pyplot as plt import numpy as np from mpl_toolkits.mplot3d import Axes3D diff --git a/examples/03_variogram/05_auto_fit_variogram.py b/examples/03_variogram/05_auto_fit_variogram.py index 9c1cf2f9..2fcc7fbd 100644 --- a/examples/03_variogram/05_auto_fit_variogram.py +++ b/examples/03_variogram/05_auto_fit_variogram.py @@ -2,6 +2,7 @@ Fit Variogram with automatic binning ------------------------------------ """ + import numpy as np import gstools as gs diff --git a/examples/03_variogram/06_auto_bin_latlon.py b/examples/03_variogram/06_auto_bin_latlon.py index 20f6f9bc..cc248ea1 100644 --- a/examples/03_variogram/06_auto_bin_latlon.py +++ b/examples/03_variogram/06_auto_bin_latlon.py @@ -8,6 +8,7 @@ We use a data set from 20 meteo-stations choosen randomly. """ + import numpy as np import gstools as gs diff --git a/examples/04_vector_field/00_2d_vector_field.py b/examples/04_vector_field/00_2d_vector_field.py index 6a6c71ac..2e722764 100644 --- a/examples/04_vector_field/00_2d_vector_field.py +++ b/examples/04_vector_field/00_2d_vector_field.py @@ -5,6 +5,7 @@ As a first example we are going to generate a 2d vector field with a Gaussian covariance model on a structured grid: """ + import numpy as np import gstools as gs diff --git a/examples/04_vector_field/01_3d_vector_field.py b/examples/04_vector_field/01_3d_vector_field.py index 13d76e87..5b1872bd 100755 --- a/examples/04_vector_field/01_3d_vector_field.py +++ b/examples/04_vector_field/01_3d_vector_field.py @@ -6,6 +6,7 @@ Gaussian covariance model. The mesh on which we generate the field will be externally defined and it will be generated by PyVista. """ + # sphinx_gallery_thumbnail_path = 'pics/GS_3d_vector_field.png' import pyvista as pv @@ -17,7 +18,7 @@ ############################################################################### # create a uniform grid with PyVista dims, spacing, origin = (40, 30, 10), (1, 1, 1), (-10, 0, 0) -mesh = pv.UniformGrid(dims=dims, spacing=spacing, origin=origin) +mesh = pv.ImageData(dimensions=dims, spacing=spacing, origin=origin) ############################################################################### # create an incompressible random 3d velocity field on the given mesh diff --git a/examples/05_kriging/00_simple_kriging.py b/examples/05_kriging/00_simple_kriging.py index bb83a32c..1a245b6c 100755 --- a/examples/05_kriging/00_simple_kriging.py +++ b/examples/05_kriging/00_simple_kriging.py @@ -27,6 +27,7 @@ The mean of the field has to be given beforehand. """ + import numpy as np from gstools import Gaussian, krige diff --git a/examples/05_kriging/01_ordinary_kriging.py b/examples/05_kriging/01_ordinary_kriging.py index 3bc54bd9..d26254ef 100644 --- a/examples/05_kriging/01_ordinary_kriging.py +++ b/examples/05_kriging/01_ordinary_kriging.py @@ -27,6 +27,7 @@ Here we use ordinary kriging in 1D (for plotting reasons) with 5 given observations/conditions. The estimated mean can be accessed by ``krig.mean``. """ + import numpy as np from gstools import Gaussian, krige diff --git a/examples/05_kriging/02_pykrige_interface.py b/examples/05_kriging/02_pykrige_interface.py index bb166d29..a6fbf03e 100755 --- a/examples/05_kriging/02_pykrige_interface.py +++ b/examples/05_kriging/02_pykrige_interface.py @@ -11,6 +11,7 @@ To demonstrate the general workflow, we compare ordinary kriging of PyKrige with the corresponding GSTools routine in 2D: """ + import numpy as np from matplotlib import pyplot as plt from pykrige.ok import OrdinaryKriging diff --git a/examples/05_kriging/03_compare_kriging.py b/examples/05_kriging/03_compare_kriging.py index 12f5c43b..463faa0a 100755 --- a/examples/05_kriging/03_compare_kriging.py +++ b/examples/05_kriging/03_compare_kriging.py @@ -2,6 +2,7 @@ Compare Kriging --------------- """ + import matplotlib.pyplot as plt import numpy as np diff --git a/examples/05_kriging/04_extdrift_kriging.py b/examples/05_kriging/04_extdrift_kriging.py index b9bbb54a..2e6a168a 100755 --- a/examples/05_kriging/04_extdrift_kriging.py +++ b/examples/05_kriging/04_extdrift_kriging.py @@ -2,6 +2,7 @@ External Drift Kriging ---------------------- """ + import numpy as np from gstools import SRF, Gaussian, krige diff --git a/examples/05_kriging/05_universal_kriging.py b/examples/05_kriging/05_universal_kriging.py index d3cf2880..5501694a 100755 --- a/examples/05_kriging/05_universal_kriging.py +++ b/examples/05_kriging/05_universal_kriging.py @@ -13,6 +13,7 @@ To access only the estimated mean/drift, we provide a switch `only_mean` in the call routine. """ + import numpy as np from gstools import SRF, Gaussian, krige diff --git a/examples/05_kriging/06_detrended_kriging.py b/examples/05_kriging/06_detrended_kriging.py index b68e6502..6d20cf1d 100755 --- a/examples/05_kriging/06_detrended_kriging.py +++ b/examples/05_kriging/06_detrended_kriging.py @@ -2,6 +2,7 @@ Detrended Kriging ----------------- """ + import numpy as np from gstools import SRF, Gaussian, krige diff --git a/examples/05_kriging/07_detrended_ordinary_kriging.py b/examples/05_kriging/07_detrended_ordinary_kriging.py index 9a59fc33..81d01744 100755 --- a/examples/05_kriging/07_detrended_ordinary_kriging.py +++ b/examples/05_kriging/07_detrended_ordinary_kriging.py @@ -2,6 +2,7 @@ Detrended Ordinary Kriging -------------------------- """ + import numpy as np from gstools import SRF, Gaussian, krige diff --git a/examples/05_kriging/09_pseudo_inverse.py b/examples/05_kriging/09_pseudo_inverse.py index 9920a956..7615d888 100755 --- a/examples/05_kriging/09_pseudo_inverse.py +++ b/examples/05_kriging/09_pseudo_inverse.py @@ -17,6 +17,7 @@ In the following we have two different values at the same location. The resulting kriging field will hold the average at this point. """ + import numpy as np from gstools import Gaussian, krige diff --git a/examples/06_conditioned_fields/00_condition_ensemble.py b/examples/06_conditioned_fields/00_condition_ensemble.py index 716edd20..5cc07eed 100644 --- a/examples/06_conditioned_fields/00_condition_ensemble.py +++ b/examples/06_conditioned_fields/00_condition_ensemble.py @@ -6,6 +6,7 @@ with 5 given observations/conditions, to generate an ensemble of conditioned random fields. """ + import matplotlib.pyplot as plt import numpy as np diff --git a/examples/06_conditioned_fields/01_2D_condition_ensemble.py b/examples/06_conditioned_fields/01_2D_condition_ensemble.py index 4a03e66b..81f51464 100644 --- a/examples/06_conditioned_fields/01_2D_condition_ensemble.py +++ b/examples/06_conditioned_fields/01_2D_condition_ensemble.py @@ -4,6 +4,7 @@ Let's create an ensemble of conditioned random fields in 2D. """ + import matplotlib.pyplot as plt import numpy as np diff --git a/examples/07_transformations/00_log_normal.py b/examples/07_transformations/00_log_normal.py index 3d2f4532..d44c1627 100755 --- a/examples/07_transformations/00_log_normal.py +++ b/examples/07_transformations/00_log_normal.py @@ -6,6 +6,7 @@ See :any:`transform.normal_to_lognormal` """ + import gstools as gs # structured field with a size of 100x100 and a grid-size of 1x1 diff --git a/examples/07_transformations/01_binary.py b/examples/07_transformations/01_binary.py index a403abb7..125e29d0 100755 --- a/examples/07_transformations/01_binary.py +++ b/examples/07_transformations/01_binary.py @@ -8,6 +8,7 @@ See :any:`transform.binary` """ + import gstools as gs # structured field with a size of 100x100 and a grid-size of 1x1 diff --git a/examples/07_transformations/02_discrete.py b/examples/07_transformations/02_discrete.py index 268cc960..48f67a2d 100755 --- a/examples/07_transformations/02_discrete.py +++ b/examples/07_transformations/02_discrete.py @@ -9,6 +9,7 @@ See :any:`transform.discrete` """ + import numpy as np import gstools as gs diff --git a/examples/07_transformations/03_zinn_harvey.py b/examples/07_transformations/03_zinn_harvey.py index 01e65790..fad1fb64 100755 --- a/examples/07_transformations/03_zinn_harvey.py +++ b/examples/07_transformations/03_zinn_harvey.py @@ -9,6 +9,7 @@ See :any:`transform.zinnharvey` """ + import gstools as gs # structured field with a size of 100x100 and a grid-size of 1x1 diff --git a/examples/07_transformations/04_bimodal.py b/examples/07_transformations/04_bimodal.py index 8234a486..4dd6fb29 100755 --- a/examples/07_transformations/04_bimodal.py +++ b/examples/07_transformations/04_bimodal.py @@ -11,6 +11,7 @@ See: :any:`transform.normal_to_arcsin` and :any:`transform.normal_to_uquad` """ + import gstools as gs # structured field with a size of 100x100 and a grid-size of 1x1 diff --git a/examples/07_transformations/05_combinations.py b/examples/07_transformations/05_combinations.py index fb27c5e6..1fbe367e 100755 --- a/examples/07_transformations/05_combinations.py +++ b/examples/07_transformations/05_combinations.py @@ -15,6 +15,7 @@ If you don't specify `field` and `store` everything happens inplace. """ + # sphinx_gallery_thumbnail_number = 1 import gstools as gs diff --git a/examples/08_geo_coordinates/00_field_generation.py b/examples/08_geo_coordinates/00_field_generation.py index b7b3748a..5b1a6fca 100755 --- a/examples/08_geo_coordinates/00_field_generation.py +++ b/examples/08_geo_coordinates/00_field_generation.py @@ -16,6 +16,7 @@ To generate the field, we simply pass ``(lat, lon)`` as the position tuple to the :any:`SRF` class. """ + import numpy as np import gstools as gs diff --git a/examples/08_geo_coordinates/01_dwd_krige.py b/examples/08_geo_coordinates/01_dwd_krige.py index 44b9f28c..3c17fb7e 100755 --- a/examples/08_geo_coordinates/01_dwd_krige.py +++ b/examples/08_geo_coordinates/01_dwd_krige.py @@ -15,6 +15,7 @@ In order to keep the number of dependecies low, the calls of both functions shown beneath are commented out. """ + # sphinx_gallery_thumbnail_number = 2 import matplotlib.pyplot as plt import numpy as np diff --git a/examples/09_spatio_temporal/03_geographic_coordinates.py b/examples/09_spatio_temporal/03_geographic_coordinates.py index 56684a94..b1cfbff6 100644 --- a/examples/09_spatio_temporal/03_geographic_coordinates.py +++ b/examples/09_spatio_temporal/03_geographic_coordinates.py @@ -18,6 +18,7 @@ We will set a spatial length-scale of `1000` and a time length-scale of `100` days. """ + import numpy as np import gstools as gs diff --git a/examples/10_normalizer/00_lognormal_kriging.py b/examples/10_normalizer/00_lognormal_kriging.py index b7bf4a8a..9880bc37 100644 --- a/examples/10_normalizer/00_lognormal_kriging.py +++ b/examples/10_normalizer/00_lognormal_kriging.py @@ -14,6 +14,7 @@ In this example we will use ordinary kriging. """ + import numpy as np import gstools as gs diff --git a/examples/10_normalizer/01_auto_fit.py b/examples/10_normalizer/01_auto_fit.py index 80a49569..71ad1385 100644 --- a/examples/10_normalizer/01_auto_fit.py +++ b/examples/10_normalizer/01_auto_fit.py @@ -18,6 +18,7 @@ We will generate the "original" field on a 60x60 mesh, from which we will take samples in order to pretend a situation of data-scarcity. """ + import matplotlib.pyplot as plt import numpy as np diff --git a/examples/10_normalizer/02_compare.py b/examples/10_normalizer/02_compare.py index 24f7c9f8..2dd74488 100644 --- a/examples/10_normalizer/02_compare.py +++ b/examples/10_normalizer/02_compare.py @@ -6,6 +6,7 @@ But first, we define a convenience routine and make some imports as always. """ + import matplotlib.pyplot as plt import numpy as np diff --git a/pyproject.toml b/pyproject.toml index 70b7d0a6..17bcb246 100644 --- a/pyproject.toml +++ b/pyproject.toml @@ -2,8 +2,9 @@ requires = [ "setuptools>=64", "setuptools_scm>=7", - "oldest-supported-numpy", - "Cython>=3.0", + "numpy>=2.0.0rc1,<2.3; python_version >= '3.9'", + "oldest-supported-numpy; python_version < '3.9'", + "Cython>=3.0.10,<3.1.0", "extension-helpers>=1", ] build-backend = "setuptools.build_meta" @@ -48,7 +49,7 @@ dependencies = [ "emcee>=3.0.0", "hankel>=1.0.0", "meshio>=5.1.0", - "numpy>=1.14.5", + "numpy>=1.20.0", "pyevtk>=1.1.1", "scipy>=1.1.0", ] @@ -60,20 +61,20 @@ doc = [ "meshzoo>=0.7", "numpydoc>=1.1", "pykrige>=1.5,<2", - "pyvista>=0.29", - "sphinx>=4", + "pyvista>=0.40", + "sphinx>=7", "sphinx-gallery>=0.8", - "sphinx-rtd-theme>=1,<1.1", + "sphinx-rtd-theme>=2", "sphinxcontrib-youtube>=1.1", ] plotting = [ "matplotlib>=3", - "pyvista>=0.29", + "pyvista>=0.40", ] rust = ["gstools_core>=0.2.0,<1"] test = ["pytest-cov>=3"] lint = [ - "black", + "black>=24", "pylint", "isort[colors]", "cython-lint", @@ -161,11 +162,9 @@ target-version = [ # Switch to using build build-frontend = "build" # Disable building PyPy wheels on all platforms, 32bit for py3.10/11/12, musllinux builds, py3.6/7 -skip = ["cp36-*", "cp37-*", "pp*", "cp31*-win32", "cp31*-manylinux_i686", "*-musllinux_*"] +skip = ["cp36-*", "cp37-*", "pp*", "*-win32", "*-manylinux_i686", "*-musllinux_*"] # Run the package tests using `pytest` test-extras = "test" test-command = "pytest -v {package}/tests" # Skip trying to test arm64 builds on Intel Macs test-skip = "*-macosx_arm64 *-macosx_universal2:arm64" -# no wheels for linux-32bit anymore for numpy>=1.22 -environment = "PIP_PREFER_BINARY=1" diff --git a/setup.py b/setup.py index f465ffd4..b27548a9 100644 --- a/setup.py +++ b/setup.py @@ -1,4 +1,5 @@ """GSTools: A geostatistical toolbox.""" + import os import numpy as np diff --git a/src/gstools/__init__.py b/src/gstools/__init__.py index 6d62d558..11e63a2b 100644 --- a/src/gstools/__init__.py +++ b/src/gstools/__init__.py @@ -129,6 +129,7 @@ DEGREE_SCALE RADIAN_SCALE """ + # Hooray! from gstools import ( config, diff --git a/src/gstools/config.py b/src/gstools/config.py index f0a87bb6..24ce20c7 100644 --- a/src/gstools/config.py +++ b/src/gstools/config.py @@ -4,6 +4,7 @@ .. currentmodule:: gstools.config """ + NUM_THREADS = None # pylint: disable=W0611 diff --git a/src/gstools/covmodel/base.py b/src/gstools/covmodel/base.py index be530075..23e19881 100644 --- a/src/gstools/covmodel/base.py +++ b/src/gstools/covmodel/base.py @@ -8,6 +8,7 @@ .. autosummary:: CovModel """ + # pylint: disable=C0103, R0201, E1101, C0302, W0613 import copy diff --git a/src/gstools/covmodel/fit.py b/src/gstools/covmodel/fit.py index dc2d5a3a..b420be70 100755 --- a/src/gstools/covmodel/fit.py +++ b/src/gstools/covmodel/fit.py @@ -8,6 +8,7 @@ .. autosummary:: fit_variogram """ + # pylint: disable=C0103, W0632 import numpy as np from scipy.optimize import curve_fit @@ -308,9 +309,11 @@ def _pre_init_guess(model, init_guess, mean_x=1.0, mean_y=1.0): for opt in model.opt_arg: init_guess.setdefault( opt, - default_arg_from_bounds(bnd[opt]) - if default - else getattr(model, opt), + ( + default_arg_from_bounds(bnd[opt]) + if default + else getattr(model, opt) + ), ) # convert all init guesses to float (except "anis") for arg in model.iso_arg: diff --git a/src/gstools/covmodel/models.py b/src/gstools/covmodel/models.py index a2aaf6a3..ab94e279 100644 --- a/src/gstools/covmodel/models.py +++ b/src/gstools/covmodel/models.py @@ -20,6 +20,7 @@ SuperSpherical JBessel """ + # pylint: disable=C0103, E1101, R0201 import warnings diff --git a/src/gstools/covmodel/plot.py b/src/gstools/covmodel/plot.py index 334ccbe2..32148c14 100644 --- a/src/gstools/covmodel/plot.py +++ b/src/gstools/covmodel/plot.py @@ -24,6 +24,7 @@ plot_spectral_density plot_spectral_rad_pdf """ + # pylint: disable=C0103, C0415, E1130 import numpy as np diff --git a/src/gstools/covmodel/tpl_models.py b/src/gstools/covmodel/tpl_models.py index 5e980760..b728e7b9 100644 --- a/src/gstools/covmodel/tpl_models.py +++ b/src/gstools/covmodel/tpl_models.py @@ -11,6 +11,7 @@ TPLStable TPLSimple """ + # pylint: disable=C0103, E1101 import warnings diff --git a/src/gstools/field/base.py b/src/gstools/field/base.py index bb514141..2006b858 100755 --- a/src/gstools/field/base.py +++ b/src/gstools/field/base.py @@ -8,6 +8,7 @@ .. autosummary:: Field """ + # pylint: disable=C0103, C0415 from collections.abc import Iterable from copy import copy @@ -362,7 +363,7 @@ def transform( ---------- method : :class:`str` Method to use. - See :any:`gstools.transform` for available transformations. + See :py:mod:`gstools.transform` for available transformations. field : :class:`str`, optional Name of field to be transformed. The default is "field". store : :class:`str` or :class:`bool`, optional diff --git a/src/gstools/field/cond_srf.py b/src/gstools/field/cond_srf.py index 385216ba..c3e03fe2 100644 --- a/src/gstools/field/cond_srf.py +++ b/src/gstools/field/cond_srf.py @@ -8,6 +8,7 @@ .. autosummary:: CondSRF """ + # pylint: disable=C0103, W0231, W0221, W0222, E1102 import numpy as np diff --git a/src/gstools/field/generator.py b/src/gstools/field/generator.py index 571c9733..5beab10d 100644 --- a/src/gstools/field/generator.py +++ b/src/gstools/field/generator.py @@ -12,6 +12,7 @@ RandMeth IncomprRandMeth """ + # pylint: disable=C0103, W0222, C0412, W0231 import warnings from abc import ABC, abstractmethod diff --git a/src/gstools/field/plot.py b/src/gstools/field/plot.py index d06a22a1..ab28c974 100644 --- a/src/gstools/field/plot.py +++ b/src/gstools/field/plot.py @@ -9,6 +9,7 @@ plot_field plot_vec_field """ + # pylint: disable=C0103, W0613, E1101 import numpy as np from scipy import interpolate as inter diff --git a/src/gstools/field/srf.py b/src/gstools/field/srf.py index a8a1e575..d88e46c0 100644 --- a/src/gstools/field/srf.py +++ b/src/gstools/field/srf.py @@ -8,6 +8,7 @@ .. autosummary:: SRF """ + # pylint: disable=C0103, W0221, E1102 import numpy as np diff --git a/src/gstools/field/tools.py b/src/gstools/field/tools.py index 4b128d8a..dfa2e3c6 100644 --- a/src/gstools/field/tools.py +++ b/src/gstools/field/tools.py @@ -10,6 +10,7 @@ to_vtk_helper generate_on_mesh """ + # pylint: disable=W0212, C0415 import meshio import numpy as np diff --git a/src/gstools/field/upscaling.py b/src/gstools/field/upscaling.py index 6dcbaa7b..857bfc45 100644 --- a/src/gstools/field/upscaling.py +++ b/src/gstools/field/upscaling.py @@ -11,6 +11,7 @@ var_coarse_graining var_no_scaling """ + # pylint: disable=W0613 import warnings diff --git a/src/gstools/krige/__init__.py b/src/gstools/krige/__init__.py index bb3ef699..66d03246 100644 --- a/src/gstools/krige/__init__.py +++ b/src/gstools/krige/__init__.py @@ -16,6 +16,7 @@ ExtDrift Detrended """ + from gstools.krige.base import Krige from gstools.krige.methods import ( Detrended, diff --git a/src/gstools/krige/base.py b/src/gstools/krige/base.py index 336f4cc0..49a4f62f 100755 --- a/src/gstools/krige/base.py +++ b/src/gstools/krige/base.py @@ -8,6 +8,7 @@ .. autosummary:: Krige """ + # pylint: disable=C0103, W0221, E1102, R0201, C0412 import collections @@ -383,9 +384,7 @@ def _pre_ext_drift(self, pnt_cnt, ext_drift=None, set_cond=False): the drift values at the given positions """ if ext_drift is not None: - ext_drift = np.array( - ext_drift, dtype=np.double, ndmin=2, copy=False - ) + ext_drift = np.atleast_2d(np.asarray(ext_drift, dtype=np.double)) if ext_drift.size == 0: # treat empty array as no ext_drift return np.array([]) if set_cond: diff --git a/src/gstools/krige/methods.py b/src/gstools/krige/methods.py index b258a02d..19ffed56 100644 --- a/src/gstools/krige/methods.py +++ b/src/gstools/krige/methods.py @@ -12,6 +12,7 @@ ExtDrift Detrended """ + # pylint: disable=C0103 from gstools.krige.base import Krige diff --git a/src/gstools/krige/tools.py b/src/gstools/krige/tools.py index e3112ae4..62926595 100644 --- a/src/gstools/krige/tools.py +++ b/src/gstools/krige/tools.py @@ -9,6 +9,7 @@ set_condition get_drift_functions """ + # pylint: disable=C0103 from itertools import combinations_with_replacement diff --git a/src/gstools/normalizer/base.py b/src/gstools/normalizer/base.py index 0072d6a9..4a8477c6 100644 --- a/src/gstools/normalizer/base.py +++ b/src/gstools/normalizer/base.py @@ -8,6 +8,7 @@ .. autosummary:: Normalizer """ + # pylint: disable=R0201 import warnings diff --git a/src/gstools/normalizer/methods.py b/src/gstools/normalizer/methods.py index f66cfd19..a46dc230 100644 --- a/src/gstools/normalizer/methods.py +++ b/src/gstools/normalizer/methods.py @@ -13,6 +13,7 @@ Modulus Manly """ + # pylint: disable=E1101 import numpy as np diff --git a/src/gstools/normalizer/tools.py b/src/gstools/normalizer/tools.py index 7a5237b7..3e395d29 100644 --- a/src/gstools/normalizer/tools.py +++ b/src/gstools/normalizer/tools.py @@ -9,6 +9,7 @@ apply_mean_norm_trend remove_trend_norm_mean """ + import numpy as np from gstools.normalizer.base import Normalizer diff --git a/src/gstools/random/rng.py b/src/gstools/random/rng.py index 4f698557..ad78a0c6 100644 --- a/src/gstools/random/rng.py +++ b/src/gstools/random/rng.py @@ -8,6 +8,7 @@ .. autosummary:: RNG """ + # pylint: disable=E1101 import emcee as mc import numpy as np diff --git a/src/gstools/tools/export.py b/src/gstools/tools/export.py index 3e522c92..38254ceb 100644 --- a/src/gstools/tools/export.py +++ b/src/gstools/tools/export.py @@ -13,6 +13,7 @@ to_vtk_structured to_vtk_unstructured """ + # pylint: disable=C0103, E1101 import numpy as np from pyevtk.hl import gridToVTK, pointsToVTK diff --git a/src/gstools/tools/geometric.py b/src/gstools/tools/geometric.py index 7f2ea10e..55408965 100644 --- a/src/gstools/tools/geometric.py +++ b/src/gstools/tools/geometric.py @@ -29,6 +29,7 @@ chordal_to_great_circle great_circle_to_chordal """ + # pylint: disable=C0103 import numpy as np @@ -383,7 +384,7 @@ def generate_st_grid(pos, time, mesh_type="unstructured"): if mesh_type != "unstructured": pos = generate_grid(pos) else: - pos = np.array(pos, dtype=np.double, ndmin=2, copy=False) + pos = np.atleast_2d(np.asarray(pos, dtype=np.double)) out = [np.repeat(p.reshape(-1), np.size(time)) for p in pos] out.append(np.tile(time, np.size(pos[0]))) return np.asarray(out, dtype=np.double) @@ -552,7 +553,7 @@ def format_unstruct_pos_shape(pos, shape, check_stacked_shape=False): # now we try to be smart pre_len = len(np.atleast_1d(pos)) # care about 1D: pos can be given as 1D array here -> convert to 2D array - pos = np.array(pos, dtype=np.double, ndmin=2, copy=False) + pos = np.atleast_2d(np.asarray(pos, dtype=np.double)) post_len = len(pos) # first array dimension should be spatial dimension (1D is special case) dim = post_len if pre_len == post_len else 1 @@ -606,7 +607,7 @@ def ang2dir(angles, dtype=np.double, dim=None): the array of direction vectors """ pre_dim = np.asanyarray(angles).ndim - angles = np.array(angles, ndmin=2, dtype=dtype, copy=False) + angles = np.atleast_2d(np.asarray(angles, dtype=dtype)) if len(angles.shape) > 2: raise ValueError(f"Can't interpret angles array {angles}") dim = angles.shape[1] + 1 if dim is None else dim diff --git a/src/gstools/tools/misc.py b/src/gstools/tools/misc.py index 068200b7..aaba1501 100755 --- a/src/gstools/tools/misc.py +++ b/src/gstools/tools/misc.py @@ -10,6 +10,7 @@ list_format eval_func """ + # pylint: disable=C0103, C0415 import numpy as np diff --git a/src/gstools/tools/special.py b/src/gstools/tools/special.py index c48887ea..1457b736 100644 --- a/src/gstools/tools/special.py +++ b/src/gstools/tools/special.py @@ -14,6 +14,7 @@ tpl_exp_spec_dens tpl_gau_spec_dens """ + # pylint: disable=C0103, E1101 import numpy as np from scipy import special as sps diff --git a/src/gstools/transform/array.py b/src/gstools/transform/array.py index 30c92bb6..87564edf 100644 --- a/src/gstools/transform/array.py +++ b/src/gstools/transform/array.py @@ -18,6 +18,7 @@ array_to_arcsin array_to_uquad """ + # pylint: disable=C0103, C0123, R0911 from warnings import warn diff --git a/src/gstools/transform/field.py b/src/gstools/transform/field.py index 5d204407..4a281564 100644 --- a/src/gstools/transform/field.py +++ b/src/gstools/transform/field.py @@ -26,6 +26,7 @@ normal_to_arcsin normal_to_uquad """ + # pylint: disable=C0103, C0123, R0911, R1735 import numpy as np @@ -111,7 +112,7 @@ def apply(fld, method, field="field", store=True, process=False, **kwargs): Field class containing a generated field. method : :class:`str` Method to use. - See :any:`gstools.transform` for available transformations. + See :py:mod:`gstools.transform` for available transformations. field : :class:`str`, optional Name of field to be transformed. The default is "field". store : :class:`str` or :class:`bool`, optional diff --git a/src/gstools/variogram/binning.py b/src/gstools/variogram/binning.py index e8e42f38..86d4fdc2 100644 --- a/src/gstools/variogram/binning.py +++ b/src/gstools/variogram/binning.py @@ -8,6 +8,7 @@ .. autosummary:: standard_bins """ + import numpy as np from gstools.tools import RADIAN_SCALE diff --git a/src/gstools/variogram/estimator.pyx b/src/gstools/variogram/estimator.pyx index ab418e32..e00824be 100644 --- a/src/gstools/variogram/estimator.pyx +++ b/src/gstools/variogram/estimator.pyx @@ -116,7 +116,7 @@ ctypedef double (*_estimator_func)(const double) nogil cdef inline void normalization_matheron( double[:] variogram, - long[:] counts, + np.int64_t[:] counts, ): cdef int i for i in range(variogram.shape[0]): @@ -125,10 +125,10 @@ cdef inline void normalization_matheron( cdef inline void normalization_cressie( double[:] variogram, - long[:] counts, + np.int64_t[:] counts, ): cdef int i - cdef long cnt + cdef np.int64_t cnt for i in range(variogram.shape[0]): # avoid division by zero cnt = max(counts[i], 1) @@ -139,12 +139,12 @@ cdef inline void normalization_cressie( ctypedef void (*_normalization_func)( double[:], - long[:], + np.int64_t[:], ) cdef inline void normalization_matheron_vec( double[:, :] variogram, - long[:, :] counts, + np.int64_t[:, :] counts, ): cdef int d for d in range(variogram.shape[0]): @@ -152,7 +152,7 @@ cdef inline void normalization_matheron_vec( cdef inline void normalization_cressie_vec( double[:, :] variogram, - long[:, :] counts, + np.int64_t[:, :] counts, ): cdef int d for d in range(variogram.shape[0]): @@ -160,7 +160,7 @@ cdef inline void normalization_cressie_vec( ctypedef void (*_normalization_func_vec)( double[:, :], - long[:, :], + np.int64_t[:, :], ) cdef _estimator_func choose_estimator_func(str estimator_type): @@ -221,7 +221,7 @@ def directional( cdef int f_max = f.shape[0] cdef double[:, :] variogram = np.zeros((d_max, len(bin_edges)-1)) - cdef long[:, :] counts = np.zeros((d_max, len(bin_edges)-1), dtype=long) + cdef np.int64_t[:, :] counts = np.zeros((d_max, len(bin_edges)-1), dtype=np.int64) cdef int i, j, k, m, d cdef double dist @@ -287,7 +287,7 @@ def unstructured( cdef int f_max = f.shape[0] cdef double[:] variogram = np.zeros(len(bin_edges)-1) - cdef long[:] counts = np.zeros(len(bin_edges)-1, dtype=long) + cdef np.int64_t[:] counts = np.zeros(len(bin_edges)-1, dtype=np.int64) cdef int i, j, k, m cdef double dist @@ -324,7 +324,7 @@ def structured( cdef int k_max = i_max + 1 cdef double[:] variogram = np.zeros(k_max) - cdef long[:] counts = np.zeros(k_max, dtype=long) + cdef np.int64_t[:] counts = np.zeros(k_max, dtype=np.int64) cdef int i, j, k cdef int num_threads_c = set_num_threads(num_threads) @@ -356,7 +356,7 @@ def ma_structured( cdef int k_max = i_max + 1 cdef double[:] variogram = np.zeros(k_max) - cdef long[:] counts = np.zeros(k_max, dtype=long) + cdef np.int64_t[:] counts = np.zeros(k_max, dtype=np.int64) cdef int i, j, k cdef int num_threads_c = set_num_threads(num_threads) diff --git a/src/gstools/variogram/variogram.py b/src/gstools/variogram/variogram.py index 31e29191..afcf336f 100644 --- a/src/gstools/variogram/variogram.py +++ b/src/gstools/variogram/variogram.py @@ -9,6 +9,7 @@ vario_estimate vario_estimate_axis """ + # pylint: disable=C0412 import numpy as np @@ -262,7 +263,7 @@ def vario_estimate( John Wiley & Sons. (2007) """ if bin_edges is not None: - bin_edges = np.array(bin_edges, ndmin=1, dtype=np.double, copy=False) + bin_edges = np.atleast_1d(np.asarray(bin_edges, dtype=np.double)) bin_centers = (bin_edges[:-1] + bin_edges[1:]) / 2.0 # allow multiple fields at same positions (ndmin=2: first axis -> field ID) # need to convert to ma.array, since list of ma.array is not recognised @@ -315,7 +316,7 @@ def vario_estimate( # set directions dir_no = 0 if direction is not None and dim > 1: - direction = np.array(direction, ndmin=2, dtype=np.double, copy=False) + direction = np.atleast_2d(np.asarray(direction, dtype=np.double)) if len(direction.shape) > 2: raise ValueError(f"Can't interpret directions: {direction}") if direction.shape[1] != dim: @@ -473,7 +474,7 @@ def vario_estimate_axis( if not config.USE_RUST: mask = np.asarray(mask, dtype=np.int32) else: - field = np.array(field, ndmin=1, dtype=np.double, copy=False) + field = np.atleast_1d(np.asarray(field, dtype=np.double)) missing_mask = None # free space axis_to_swap = AXIS_DIR[direction] if direction in AXIS else int(direction)

Numpy 2 compatibility Numpy 2.0.0rc1 was released: https://pypi.org/project/numpy/2.0.0c1 and is officially usable to build extensions. We should be prepared to be compatible with it. - Building with numpy 2.0.0rc1 works :tada: - Running tests with numpy 2 fails :sob: A lot of the tests fail because of emcee and this issue: https://github.com/dfm/emcee/issues/509 Other errors come from these lines in `estimator.pyx`: ```cython cdef long[:, :] counts = np.zeros((d_max, len(bin_edges)-1), dtype=long) cdef long[:] counts = np.zeros(len(bin_edges)-1, dtype=long) cdef long[:] counts = np.zeros(k_max, dtype=long) cdef long[:] counts = np.zeros(k_max, dtype=long) ``` and raise ``` E ValueError: Buffer dtype mismatch, expected 'long' but got 'long long' ``` This only occurs on windows and is related to this: https://numpy.org/devdocs/numpy_2_0_migration_guide.html#windows-default-integer One other category of errors is: ``` If using `np.array(obj, copy=False)` replace it with `np.asarray(obj)` to allow a copy when needed (no behavior change in NumPy 1.x). ``` `np.array(..., copy=False)` is used in [krige/base.py](https://github.com/GeoStat-Framework/GSTools/blob/90003fbbe193cfa835a891b403b4024bb21867f9/src/gstools/krige/base.py#L387), 3 times in [variogram/variogram.py](https://github.com/GeoStat-Framework/GSTools/blob/90003fbbe193cfa835a891b403b4024bb21867f9/src/gstools/variogram/variogram.py#L265) and 3 times in [tools/geometric.py](https://github.com/GeoStat-Framework/GSTools/blob/90003fbbe193cfa835a891b403b4024bb21867f9/src/gstools/tools/geometric.py#L386) In order to mimic the `ndmin=...` feature, we could replace these with e.g.: ```python pos = np.atleast_2d(np.asarray(pos, dtype=np.double)) ```

2024-03-14T16:12:28

0.0

[]

GeoStat-Framework/GSTools

GeoStat-Framework__GSTools-145

d66613bebe69ad1fbeb8cdac86e2db4bee999b2f

diff --git a/gstools/covmodel/base.py b/gstools/covmodel/base.py index ec34f999..e121acb2 100644 --- a/gstools/covmodel/base.py +++ b/gstools/covmodel/base.py @@ -36,6 +36,7 @@ set_dim, compare, model_repr, + default_arg_from_bounds, ) from gstools.covmodel import plot from gstools.covmodel.fit import fit_variogram @@ -414,7 +415,10 @@ def default_opt_arg(self): Should be given as a dictionary when overridden. """ - return {} + return { + opt: default_arg_from_bounds(bnd) + for (opt, bnd) in self.default_opt_arg_bounds().items() + } def default_opt_arg_bounds(self): """Provide default boundaries for optional arguments.""" @@ -607,11 +611,19 @@ def fit_variogram( and set to the current sill of the model. Then, the procedure above is applied. Default: None - init_guess : :class:`str`, optional + init_guess : :class:`str` or :class:`dict`, optional Initial guess for the estimation. Either: * "default": using the default values of the covariance model + ("len_scale" will be mean of given bin centers; + "var" and "nugget" will be mean of given variogram values + (if in given bounds)) * "current": using the current values of the covariance model + * dict: dictionary with parameter names and given value + (separate "default" can bet set to "default" or "current" for + unspecified values to get same behavior as given above + ("default" by default)) + Example: ``{"len_scale": 10, "default": "current"}`` Default: "default" weights : :class:`str`, :class:`numpy.ndarray`, :class:`callable`, optional diff --git a/gstools/covmodel/fit.py b/gstools/covmodel/fit.py index 0fa17075..96462d9e 100755 --- a/gstools/covmodel/fit.py +++ b/gstools/covmodel/fit.py @@ -77,11 +77,19 @@ def fit_variogram( and set to the current sill of the model. Then, the procedure above is applied. Default: None - init_guess : :class:`str`, optional + init_guess : :class:`str` or :class:`dict`, optional Initial guess for the estimation. Either: * "default": using the default values of the covariance model + ("len_scale" will be mean of given bin centers; + "var" and "nugget" will be mean of given variogram values + (if in given bounds)) * "current": using the current values of the covariance model + * dict: dictionary with parameter names and given value + (separate "default" can bet set to "default" or "current" for + unspecified values to get same behavior as given above + ("default" by default)) + Example: ``{"len_scale": 10, "default": "current"}`` Default: "default" weights : :class:`str`, :class:`numpy.ndarray`, :class:`callable`optional @@ -170,6 +178,10 @@ def fit_variogram( # prepare variogram data # => concatenate directional variograms to have a 1D array for x and y x_data, y_data, is_dir_vario = _check_vario(model, x_data, y_data) + # prepare init guess dictionary + init_guess = _pre_init_guess( + model, init_guess, np.mean(x_data), np.mean(y_data) + ) # only fit anisotropy if a directional variogram was given anis &= is_dir_vario # set weights @@ -239,7 +251,7 @@ def _pre_para(model, para_select, sill, anis): if model.var > sill: raise ValueError( "fit: if sill is fixed and variance deselected, " - + "the set variance should be less than the given sill." + "the set variance should be less than the given sill." ) para_select["nugget"] = False model.nugget = sill - model.var @@ -247,7 +259,7 @@ def _pre_para(model, para_select, sill, anis): if model.nugget > sill: raise ValueError( "fit: if sill is fixed and nugget deselected, " - + "the set nugget should be less than the given sill." + "the set nugget should be less than the given sill." ) para_select["var"] = False model.var = sill - model.nugget @@ -269,6 +281,51 @@ def _pre_para(model, para_select, sill, anis): return para, sill, constrain_sill, anis +def _pre_init_guess(model, init_guess, mean_x=1.0, mean_y=1.0): + # init guess should be a dict + if not isinstance(init_guess, dict): + init_guess = {"default": init_guess} + # "default" init guess is the respective default value + default_guess = init_guess.pop("default", "default") + if default_guess not in ["default", "current"]: + raise ValueError( + "fit_variogram: unknown def. guess: {}".format(default_guess) + ) + default = default_guess == "default" + # check invalid names for given init guesses + invalid_para = set(init_guess) - set(model.iso_arg + ["anis"]) + if invalid_para: + raise ValueError( + "fit_variogram: unknown init guess: {}".format(invalid_para) + ) + bnd = model.arg_bounds + # default length scale is mean of given bin centers (respecting "rescale") + init_guess.setdefault( + "len_scale", mean_x * model.rescale if default else model.len_scale + ) + # init guess for variance and nugget is mean of given variogram + for par in ["var", "nugget"]: + init_guess.setdefault(par, mean_y if default else getattr(model, par)) + # anis setting + init_guess.setdefault( + "anis", default_arg_from_bounds(bnd["anis"]) if default else model.anis + ) + # correctly handle given values for anis (need a list of values) + init_guess["anis"] = list(set_anis(model.dim, init_guess["anis"])) + # set optional arguments + for opt in model.opt_arg: + init_guess.setdefault( + opt, + default_arg_from_bounds(bnd[opt]) + if default + else getattr(model, opt), + ) + # convert all init guesses to float (except "anis") + for arg in model.iso_arg: + init_guess[arg] = float(init_guess[arg]) + return init_guess + + def _check_vario(model, x_data, y_data): # prepare variogram data x_data = np.array(x_data).reshape(-1) @@ -283,8 +340,8 @@ def _check_vario(model, x_data, y_data): elif x_data.size != y_data.size: raise ValueError( "CovModel.fit_variogram: Wrong number of empirical variograms! " - + "Either provide only one variogram to fit an isotropic model, " - + "or directional ones for all main axes to fit anisotropy." + "Either provide only one variogram to fit an isotropic model, " + "or directional ones for all main axes to fit anisotropy." ) if is_dir_vario and model.latlon: raise ValueError( @@ -327,10 +384,7 @@ def _init_curve_fit_para(model, para, init_guess, constrain_sill, sill, anis): init_guess_list.append( _init_guess( bounds=[low_bounds[-1], top_bounds[-1]], - current=getattr(model, par), - default=model.rescale if par == "len_scale" else 1.0, - typ=init_guess, - para_name=par, + default=init_guess[par], ) ) for opt in model.opt_arg: @@ -340,37 +394,27 @@ def _init_curve_fit_para(model, para, init_guess, constrain_sill, sill, anis): init_guess_list.append( _init_guess( bounds=[low_bounds[-1], top_bounds[-1]], - current=getattr(model, opt), - default=model.default_opt_arg()[opt], - typ=init_guess, - para_name=opt, + default=init_guess[opt], ) ) if anis: - low_bounds += [model.anis_bounds[0]] * (model.dim - 1) - top_bounds += [model.anis_bounds[1]] * (model.dim - 1) - if init_guess == "default": - def_arg = default_arg_from_bounds(model.anis_bounds) - init_guess_list += [def_arg] * (model.dim - 1) - elif init_guess == "current": - init_guess_list += list(model.anis) - else: - raise ValueError( - "CovModel.fit: unknown init_guess: '{}'".format(init_guess) + for i in range(model.dim - 1): + low_bounds.append(model.anis_bounds[0]) + top_bounds.append(model.anis_bounds[1]) + init_guess_list.append( + _init_guess( + bounds=[low_bounds[-1], top_bounds[-1]], + default=init_guess["anis"][i], + ) ) - return (low_bounds, top_bounds), init_guess_list -def _init_guess(bounds, current, default, typ, para_name): +def _init_guess(bounds, default): """Proper determination of initial guess.""" - if typ == "default": - if bounds[0] < default < bounds[1]: - return default - return default_arg_from_bounds(bounds) - if typ == "current": - return current - raise ValueError("CovModel.fit: unknown init_guess: '{}'".format(typ)) + if bounds[0] < default < bounds[1]: + return default + return default_arg_from_bounds(bounds) def _get_curve(model, para, constrain_sill, sill, anis, is_dir_vario):

Variogram fitting does not work for 'large' coordinates? Hello, I have encountered an issue in fitting a variogram model to an estimated variogram and it seems to be connected to the used coordinates. Along the tutorial on variogram fitting I tried the following code: ``` bin_max = 300000 bins = np.linspace(0, bin_max, 10) bin_center, gamma = gstools.vario_estimate_unstructured((y_coord, x_coord), vals, bins) fit_model = gstools.Exponential(dim=2) fit_param, _ = fit_model.fit_variogram(bin_center, gamma) plt.plot(bin_center, gamma) gs.covmodel.plot.plot_variogram(fit_model, ax=plt.gca(), x_max=bins[-1]) ``` The arrays `x_coord` and `y_coord` contain the coordinates for data points in Germany in the Gauss-Krüger projection (EPSG 31467). Hence, they are in the range 3280000 to 3935000 for the x direction and 5235000 to 6105000 for the y direction. This is the resulting variogram plot, which does not look right: ![gstools_issue1](https://user-images.githubusercontent.com/64479100/111665034-abea9b00-8812-11eb-8389-460012c18e53.png) I tried to locate my error and found that when doing the following conversions beforehand, the result looks very different: ``` x_coord = x_coord / 10 ** 6 y_coord = y_coord / 10 ** 6 bin_max = bin_max / 10 ** 6 ``` This is the resulting variogram plot: ![gstools_issue2](https://user-images.githubusercontent.com/64479100/111665562-26b3b600-8813-11eb-8c71-2db8d401e49d.png) Is this expected behavior? -- I use Python 3.7.9 and gstools 1.2.1

Hey @krihabu , this is a problem with scipy's [curve_fit](https://docs.scipy.org/doc/scipy/reference/generated/scipy.optimize.curve_fit.html). The routine gives up to fast and doesn't search the higher regions for length-scales. There were some major changes since 1.2.1 in `develop` and v1.3 will be released soon. There you can provide initial guesses for the variogram fitting: ```python import numpy as np import gstools as gs ### SYNTHETIC DATA x = np.random.RandomState(19970221).rand(1000) * 10000. y = np.random.RandomState(20011012).rand(1000) * 10000. model = gs.Exponential(dim=2, var=2, len_scale=2000) srf = gs.SRF(model, mean=0, seed=19970221) field = srf((x, y)) ### SYNTHETIC DATA # estimate the variogram of the field (auto binning) bin_center, gamma = gs.vario_estimate((x, y), field) # fit the variogram with a stable model. (no nugget fitted) fit_model = gs.Stable(dim=2) fit_model.len_scale = 1000 # init guess for len_scale fit_model.fit_variogram(bin_center, gamma, nugget=False, init_guess="current") # output ax = fit_model.plot(x_max=max(bin_center)) ax.scatter(bin_center, gamma) print(fit_model) ``` ![vario_fit](https://user-images.githubusercontent.com/19690642/111751798-31169400-8895-11eb-9485-db70325f74d1.png) ```bash Stable(dim=2, var=1.7, len_scale=1.68e+03, nugget=0.0, anis=[1.0], angles=[0.0], alpha=0.932) ``` So providing an initial guess for len-scale could be a good tutorial (like taking 1/10 of the domain size as often reported as a good guess in literature). But you are right, it should work out of the box. Another idea I was thinking about for some time was to fit lenght scale via its order of magnitude, i.e. log-value. This could help finding the correct size. Hope this helps for now! Will keep this open to keep track of it. Cheers, Sebastian Just in case you didn't see the [instructions](https://geostat-framework.readthedocs.io/projects/gstools/en/latest/#installation) and you need a solution urgently, it's pretty easy to install the latest and not yet released version of GSTools via pip. Thank you for the quick response! I will go with my dirty fix for the moment and will try the `init_guess` option when the new version is released. Another option could be to change the scale of the model with the new `rescale` parameter, so that the `len_scale` is given in kilometer and not meter: ```python # estimate the variogram of the field bin_center, gamma = gs.vario_estimate((x, y), field) # fit the variogram with a rescale factor of 1/1000 fit_model = gs.Exponential(dim=2, rescale=1 / 1000) fit_model.fit_variogram(bin_center, gamma) # output ax = fit_model.plot(x_max=max(bin_center)) ax.scatter(bin_center, gamma) print(fit_model) ``` ![vario_fit_rescale](https://user-images.githubusercontent.com/19690642/111821844-6944c380-88e3-11eb-95c6-7d56de2be48f.png) ```bash Exponential(dim=2, var=1.61, len_scale=1.68, nugget=0.0509, anis=[1.0], angles=[0.0]) ``` As you see, `len_scale=1.68` could be estimated with the correct order of magnitude. Note, that the default init guess is always `1`. The rescaled length can be accessed by: ```python fit_model.len_rescaled ``` ```bash 1678.510771993176 ```

2021-03-21T13:49:55

0.0

[]

meshpro/meshplex

meshpro__meshplex-147

3b8ed501b70ef6101e45e3aab9604b1add8cbe3f

diff --git a/meshplex/_mesh.py b/meshplex/_mesh.py index 67e159d..d2b5890 100644 --- a/meshplex/_mesh.py +++ b/meshplex/_mesh.py @@ -188,7 +188,7 @@ def cell_circumcenters(self): return self._circumcenters[-1] @property - def cell_circumradius(self): + def cell_circumradius(self) -> ArrayLike: """Get the circumradii of all cells""" if self._cell_circumradii is None: self._compute_cell_values() @@ -295,7 +295,7 @@ def signed_circumcenter_distances(self): return self._signed_circumcenter_distances - def _compute_cell_values(self, mask=slice(None)): + def _compute_cell_values(self, mask=None): """Computes the volumes of all edges, facets, cells etc. in the mesh. It starts off by computing the (squared) edge lengths, then complements the edge with one vertex to form face. It computes an orthogonal basis of the face (with modified @@ -303,6 +303,9 @@ def _compute_cell_values(self, mask=slice(None)): of the face is computed. Then, it complements again to form the 3-simplex, again forms an orthogonal basis with Gram-Schmidt, and so on. """ + if mask is None: + mask = slice(None) + e = self.points[self.idx[-1][..., mask]] e0 = e[0] diff = e[1] - e[0] @@ -316,11 +319,11 @@ def _compute_cell_values(self, mask=slice(None)): vv = _dot(diff, self.n - 1) circumradii2 = 0.25 * vv sqrt_vv = np.sqrt(vv) - lmbda = 0.5 * np.sqrt(vv) + lmbda = 0.5 * sqrt_vv sumx = np.array(e + circumcenters[-1]) - partitions = 0.5 * np.sqrt(np.array([vv, vv])) + partitions = 0.5 * np.array([sqrt_vv, sqrt_vv]) norms2 = np.array(volumes2) for kk, idx in enumerate(self.idx[:-1][::-1]): @@ -329,8 +332,12 @@ def _compute_cell_values(self, mask=slice(None)): p0 = self.points[idx][:, mask] v = p0 - e0 # modified gram-schmidt - for w, ww in zip(orthogonal_basis, norms2): - alpha = np.einsum("...k,...k->...", w, v) / ww + for w, w_dot_w in zip(orthogonal_basis, norms2): + w_dot_v = np.einsum("...k,...k->...", w, v) + # Compute <w, v> / <w, w>, but don't set the output value where w==0. + # The value remains uninitialized and gets canceled out in the next + # iteration when multiplied by w. + alpha = np.divide(w_dot_v, w_dot_w, where=w_dot_w > 0.0) v -= _multiply(w, alpha, self.n - 1 - kk) vv = np.einsum("...k,...k->...", v, v) @@ -353,22 +360,34 @@ def _compute_cell_values(self, mask=slice(None)): c = circumcenters[-1] p0c2 = _dot(p0 - c, self.n - 1 - kk) - # - sigma = 0.5 * (p0c2 - circumradii2) / vv - lmbda2 = sigma ** 2 * vv + # Be a bit careful here. sigma and lmbda can be negative. Also make sure + # that the values aren't nan when they should be inf (for degenerate + # simplices, i.e., vv == 0). + a = 0.5 * (p0c2 - circumradii2) + sqrt_vv = np.sqrt(vv) + with warnings.catch_warnings(): + # silence division-by-0 warnings + # Happens for degenerate cells (sqrt(vv) == 0), and this case is + # supported by meshplex. The values lmbda and sigma will just be +-inf. + warnings.simplefilter("ignore", category=RuntimeWarning) + lmbda = a / sqrt_vv + sigma_k0 = a[k0] / vv[k0] # circumcenter, squared circumradius # <https://math.stackexchange.com/a/4064749/36678> - # - circumradii2 = lmbda2[k0] + circumradii2[k0] - circumcenters.append(c[k0] + _multiply(v[k0], sigma[k0], self.n - 2 - kk)) + lmbda2_k0 = sigma_k0 * a[k0] + circumradii2 = lmbda2_k0 + circumradii2[k0] + with warnings.catch_warnings(): + # Similar as above: The multiplicattion `v * sigma` correctly produces + # nans for degenerate cells. + warnings.simplefilter("ignore", category=RuntimeWarning) + circumcenters.append( + c[k0] + _multiply(v[k0], sigma_k0, self.n - 2 - kk) + ) sumx += circumcenters[-1] # cell partitions - # don't use sqrt(lmbda2) here; lmbda, just like sigma, can be negative - sqrt_vv = np.sqrt(vv) - lmbda = sigma * sqrt_vv partitions *= lmbda / (kk + 2) # The integral of x, @@ -806,6 +825,8 @@ def remove_dangling_points(self): if self._is_point_used is not None: self._is_point_used = self._is_point_used[is_part_of_cell] + return np.sum(~is_part_of_cell) + @property def q_radius_ratio(self): """Ratio of incircle and circumcircle ratios times (n-1). ("Normalized shape diff --git a/meshplex/_mesh_tri.py b/meshplex/_mesh_tri.py index 4a6558f..6639985 100644 --- a/meshplex/_mesh_tri.py +++ b/meshplex/_mesh_tri.py @@ -1,5 +1,6 @@ import warnings +import npx import numpy as np from ._mesh import Mesh @@ -263,11 +264,15 @@ def flip_until_delaunay(self, tol=0.0, max_steps=100): if step > max_steps: m = np.min(self.signed_circumcenter_distances) - warnings.warn( + msg = ( f"Maximum number of edge flips reached ({max_steps}). " - f"Smallest signed circumcenter distance: {m:.3e}." + + f"Smallest signed circumcenter distance: {m:.3e}. " + + f"Try increasing the tolerance (currently {tol}) " + + f"or max_steps (currentframe {max_steps})." ) + warnings.warn(msg) break + # raise MeshplexError(msg) interior_facets_cells = self.facets_cells["interior"][1:3].T adj_cells = interior_facets_cells[is_flip_interior_facet].T @@ -291,8 +296,8 @@ def flip_until_delaunay(self, tol=0.0, max_steps=100): multiflip_cell_gids = cell_gids[num_flips_per_cell > 1] # actually perform the flips - self.flip_interior_facets(is_flip_interior_facet) - num_flips += np.sum(is_flip_interior_facet) + num = self.flip_interior_facets(is_flip_interior_facet) + num_flips += num # check the new signed_circumcenter_distances new_scd = self.signed_circumcenter_distances[is_flip_interior_facet] @@ -311,16 +316,47 @@ def flip_until_delaunay(self, tol=0.0, max_steps=100): ) message += "Leaving those facets as they are." warnings.warn(message) - # exit(1) # TODO remove + # Check which edges need to be flipped next. Don't flip edges which have + # just been flipped though. (This can happen due to round-off errors.) is_flip_interior_facet_old = is_flip_interior_facet.copy() is_flip_interior_facet = self.signed_circumcenter_distances < -tol - # Simply don't flip edges which have just been flipped is_flip_interior_facet[is_flip_interior_facet_old] = False return num_flips def flip_interior_facets(self, is_flip_interior_facet): + facets_cells_flip = self.facets_cells["interior"][:, is_flip_interior_facet] + # facet_gids = facets_cells_flip[0] + adj_cells = facets_cells_flip[1:3] + lids = facets_cells_flip[3:5] + + new_edges = np.array( + [ + self.cells("points")[adj_cells[0], lids[0]], + self.cells("points")[adj_cells[1], lids[1]], + ] + ).T + new_edges = np.sort(new_edges, axis=1) + do_actually_flip = np.ones(len(new_edges), dtype=bool) + + # Check if some flips would lead to the same flipped edge. This can happen, for + # example, in triangular shell meshes in 3D, and leads to weird behavior down + # the line. See <https://github.com/nschloe/meshplex/issues/130>. + _, inv = npx.unique_rows(new_edges, return_inverse=True) + is_unique = np.zeros(len(new_edges), dtype=bool) + is_unique[inv] = True + do_actually_flip &= is_unique + + # Check if expected new edges are already present in the mesh for the same + # reasons as above + already_exists = npx.isin_rows(new_edges, self.facets["points"]) + do_actually_flip &= ~already_exists + + # finally apply the filter + is_flip_interior_facet[is_flip_interior_facet] = do_actually_flip + + # now actuall perform the flip facets_cells_flip = self.facets_cells["interior"][:, is_flip_interior_facet] facet_gids = facets_cells_flip[0] adj_cells = facets_cells_flip[1:3] @@ -457,6 +493,8 @@ def flip_interior_facets(self, is_flip_interior_facet): self._update_cell_values(update_cell_ids, update_interior_facet_ids) + return np.sum(is_flip_interior_facet) + def _update_cell_values(self, cell_ids, interior_facet_ids): """Updates all sorts of cell information for the given cell IDs.""" # update idx diff --git a/setup.cfg b/setup.cfg index ca16de7..9ef63af 100644 --- a/setup.cfg +++ b/setup.cfg @@ -1,6 +1,6 @@ [metadata] name = meshplex -version = 0.16.0 +version = 0.16.1 author = Nico Schlömer author_email = [email protected] description = Fast tools for simplex meshes

mwe fails on 0.15.13 but works on 0.15.12 import meshplex import numpy as np cells = np.array( [ [4, 5, 2], [5, 6, 2], [8, 7, 6], [6, 7, 2], [4, 2, 1], [1, 2, 0], [2, 7, 3], [0, 2, 3], ] ) points = np.array( [ [4.00540750e00, 8.41882994e00, 1.87847398e-01], [4.15720337e00, 8.14281236e00, 9.12266346e-10], [4.30002833e00, 8.57424769e00, 1.04310192e-02], [4.05235481e00, 8.89348539e00, 2.06854700e-01], [4.25687935e00, 8.41702649e00, -1.80354934e-10], [4.28647665e00, 8.56681008e00, -3.27591539e-10], [4.28974575e00, 8.58639357e00, -1.84739328e-10], [4.36607782e00, 9.03478181e00, 2.94211691e-02], [4.33143378e00, 9.01150991e00, 2.88600116e-10], ] ) meshplex.MeshTri mesh = meshplex.MeshTri(points, cells) mesh.flip_until_delaunay() mesh = meshplex.MeshTri(mesh.points, mesh.cells["points"]) mesh.flip_until_delaunay() mesh.create_facets() flip_until_delaunay failure [trimesh_example.zip](https://github.com/nschloe/meshplex/files/6241541/trimesh_example.zip) Hi Nico, I get a problem again on the following mesh. Version 0.15.14 Output: ``` control_uniqueness faces are unique control_uniqueness [trimesh_example.zip](https://github.com/nschloe/meshplex/files/6241537/trimesh_example.zip) face doubled (181419, 182934, 182936) 2 face doubled (181419, 182933, 182936) 2 finished with errors ```

`flip_until_delaunay()` may (and does, in this case) create duplicate cells. You can remove them with `mesh.remove_duplicate_cells()` after the flip, then it should all work. It's not done automatically because there are cases where duplicates are appropriate. Right, `flip_until_delaunay()` might produce duplicate cells sometimes. Just call ```python mesh.remove_duplicate_cells() ``` This will take care of it. True I simplified the example too much. I wanted to say the following, iff I use flip_until_delaunay followed by remove_duplicate_cells I get a mesh with non-manifold edges: ```python import meshplex import numpy as np def unique_rows(a): # The numpy alternative `np.unique(a, axis=0)` is slow; cf. # <https://github.com/numpy/numpy/issues/11136>. b = np.ascontiguousarray(a).view(np.dtype((np.void, a.dtype.itemsize * a.shape[1]))) a_unique, inv, cts = np.unique(b, return_inverse=True, return_counts=True) a_unique = a_unique.view(a.dtype).reshape(-1, a.shape[1]) return a_unique, inv, cts def control_mesh(mesh): s = mesh.idx_hierarchy.shape a = np.sort(mesh.idx_hierarchy.reshape(s[0], -1).T) a_unique, inv, cts = unique_rows(a) if cts.max() > 2: print("control_mesh found non-manifold edges",cts.max()) for j in range(len(cts)): if cts[j] > 2: print(j, cts[j]) return False print("good_mesh") return True mesh = meshplex._reader.read("contour.obj") control_mesh(mesh) mesh.flip_until_delaunay() print("removing duplicate cells", mesh.remove_duplicate_cells()) #print("removing duplicate cells", mesh.remove_duplicate_cells()) control_mesh(mesh) ``` outputs: ``` good_mesh removing duplicate cells 2 control_mesh found non-manifold edges 4 433587 3 494068 3 641571 4 ``` Dear nico, were you able to look into this issue. Shall I try to reduce the mesh to a minimal example? Not yet. I'm hoping to get to it next week. dear nico, any news is it a hard problem? Not yet, no. Still busy with other things in meshplex, so not too far away from it. hi nico, is there any progress or information about this issue. I just need some information for my team. best I'm looking at it right now. You never know how deep the rabbit hole goes, but I'm hoping to get it done this afternoon. Let's kick it off with an MWE: ```python import numpy as np import npx import meshplex def control_mesh(mesh): s = mesh.idx_hierarchy.shape a = np.sort(mesh.idx_hierarchy.reshape(s[0], -1).T) _, cts = npx.unique_rows(a, return_counts=True) if cts.max() > 2: print("control_mesh found non-manifold edges", cts.max()) for idx in np.where(cts > 2)[0]: print(idx, cts[idx]) else: print("good_mesh") points = [ [0.011, 0.21, 0.69], [0.664, 0.70, 0.69], [0.664, 0.96, 1.43], [0.619, 0.96, 1.44], [0.664, 0.93, 1.44], [1.374, 0.21, 0.69], [0.670, 0.96, 1.44], ] cells = [[2, 0, 1], [2, 3, 0], [4, 0, 3], [4, 6, 5], [2, 5, 6]] mesh = meshplex.MeshTri(points, cells) control_mesh(mesh) print("flip") mesh.flip_until_delaunay() control_mesh(mesh) ``` ``` good_mesh flip control_mesh found non-manifold edges 4 5 4 ``` Isn’t this a round off error when evaluating the flip predicate? Flips are notorious for this... yes I remember robustness issues in geometric algorithms. offsetting 2d-curves in CAD (like rhino-cad) fails and with a random small translation they work again. with gpc I had issues with a probability of about 1e-4. When it failed I had to do some random translation. With polygon_clipper no robustness issues so far. Do you think random translation could help? import random def random_trans(points): eps = .0001 sx = eps*random.random() sy = eps * random.random() sz = eps * random.random() points = [(x + sx, y+ sy, z + sz) for x,y,z in points] return points Nope tried it does not help. I'm not sure how the code is compiled on the back end by your hardware, but in my experience IFort compiler enables unsafe math ops by default while GNU based compilers do not. Generally if we want to repeat work, we thus use GNU with safe math ops Here's an even simpler MWE: ```python import numpy as np import npx import meshplex def control_mesh(mesh): s = mesh.idx_hierarchy.shape a = np.sort(mesh.idx_hierarchy.reshape(s[0], -1).T) _, cts = npx.unique_rows(a, return_counts=True) if cts.max() > 2: print("control_mesh found non-manifold edges", cts.max()) for idx in np.where(cts > 2)[0]: print(idx, cts[idx]) else: print("good_mesh") points = [ [0.0, 0.0, 0.0], [1.0, 0.5, 0.0], [1.0, 0.76, 1.0], [1.0, 0.8, 1.0], [1.0, 0.7, 1.0], [1.0, 0.7, 0.0], ] cells = [[2, 0, 1], [2, 3, 0], [4, 0, 3], [4, 3, 5], [2, 5, 3]] mesh = meshplex.MeshTri(points, cells) mesh.save("out0.vtu") control_mesh(mesh) print("flip 1...") mesh.flip_until_delaunay(max_steps=1) mesh.save("out1.vtu") n = mesh.remove_duplicate_cells() print(f"{n = }") control_mesh(mesh) control_mesh(mesh) mesh = meshplex.MeshTri(mesh.points, mesh.cells("points")) mesh.save("out4.vtu") ``` ![screenshot](https://user-images.githubusercontent.com/181628/115889644-672ad300-a454-11eb-8dc9-e37431a7c1f1.png) So, this situation is a little crazy to start out with. It's not too strange that the edge flipper produces non-manifold edges. what are the options. Do you think it is better to find some preprocessing step or a postprocessor to solve the non-manifold edges?

2021-04-24T23:30:53

0.0

[]

jquast/blessed

jquast__blessed-260

a34c6b1869b4dd467c6d1ab6895872bb72db7e0f

diff --git a/blessed/keyboard.py b/blessed/keyboard.py index 31cc98c..401ec45 100644 --- a/blessed/keyboard.py +++ b/blessed/keyboard.py @@ -1,6 +1,7 @@ """Sub-module providing 'keyboard awareness'.""" # std imports +import os import re import time import platform @@ -448,4 +449,28 @@ def _read_until(term, pattern, timeout): ('KEY_BEGIN', curses.KEY_BEG), ) +#: Default delay, in seconds, of Escape key detection in +#: :meth:`Terminal.inkey`.` curses has a default delay of 1000ms (1 second) for +#: escape sequences. This is too long for modern applications, so we set it to +#: 350ms, or 0.35 seconds. It is still a bit conservative, for remote telnet or +#: ssh servers, for example. +DEFAULT_ESCDELAY = 0.35 + + +def _reinit_escdelay(): + # pylint: disable=W0603 + # Using the global statement: this is necessary to + # allow test coverage without complex module reload + global DEFAULT_ESCDELAY + if os.environ.get('ESCDELAY'): + try: + DEFAULT_ESCDELAY = int(os.environ['ESCDELAY']) / 1000.0 + except ValueError: + # invalid values of 'ESCDELAY' are ignored + pass + + +_reinit_escdelay() + + __all__ = ('Keystroke', 'get_keyboard_codes', 'get_keyboard_sequences',) diff --git a/blessed/terminal.py b/blessed/terminal.py index 76214e9..fc0b0c4 100644 --- a/blessed/terminal.py +++ b/blessed/terminal.py @@ -18,7 +18,8 @@ # local from .color import COLOR_DISTANCE_ALGORITHMS -from .keyboard import (_time_left, +from .keyboard import (DEFAULT_ESCDELAY, + _time_left, _read_until, resolve_sequence, get_keyboard_codes, @@ -1425,8 +1426,8 @@ def keypad(self): self.stream.write(self.rmkx) self.stream.flush() - def inkey(self, timeout=None, esc_delay=0.35): - """ + def inkey(self, timeout=None, esc_delay=DEFAULT_ESCDELAY): + r""" Read and return the next keyboard event within given timeout. Generally, this should be used inside the :meth:`raw` context manager. @@ -1434,12 +1435,16 @@ def inkey(self, timeout=None, esc_delay=0.35): :arg float timeout: Number of seconds to wait for a keystroke before returning. When ``None`` (default), this method may block indefinitely. - :arg float esc_delay: To distinguish between the keystroke of - ``KEY_ESCAPE``, and sequences beginning with escape, the parameter - ``esc_delay`` specifies the amount of time after receiving escape - (``chr(27)``) to seek for the completion of an application key - before returning a :class:`~.Keystroke` instance for - ``KEY_ESCAPE``. + :arg float esc_delay: Time in seconds to block after Escape key + is received to await another key sequence beginning with + escape such as *KEY_LEFT*, sequence ``'\x1b[D'``], before returning a + :class:`~.Keystroke` instance for ``KEY_ESCAPE``. + + Users may override the default value of ``esc_delay`` in seconds, + using environment value of ``ESCDELAY`` as milliseconds, see + `ncurses(3)`_ section labeled *ESCDELAY* for details. Setting + the value as an argument to this function will override any + such preference. :rtype: :class:`~.Keystroke`. :returns: :class:`~.Keystroke`, which may be empty (``u''``) if ``timeout`` is specified and keystroke is not received. @@ -1454,11 +1459,12 @@ def inkey(self, timeout=None, esc_delay=0.35): <https://docs.microsoft.com/en-us/windows/win32/api/timeapi/nf-timeapi-timebeginperiod>`_. Decreasing the time resolution will reduce this to 10 ms, while increasing it, which is rarely done, will have a perceptable impact on the behavior. + + _`ncurses(3)`: https://www.man7.org/linux/man-pages/man3/ncurses.3x.html """ resolve = functools.partial(resolve_sequence, mapper=self._keymap, codes=self._keycodes) - stime = time.time() # re-buffer previously received keystrokes, diff --git a/tox.ini b/tox.ini index 017f354..d3a4405 100644 --- a/tox.ini +++ b/tox.ini @@ -67,8 +67,9 @@ commands = autopep8 --in-place --recursive --aggressive --aggressive blessed/ bin/ setup.py [testenv:docformatter] +# docformatter pinned due to https://github.com/PyCQA/docformatter/issues/264 deps = - docformatter + docformatter<1.7.4 untokenize commands = docformatter \ @@ -83,8 +84,9 @@ commands = {toxinidir}/docs/conf.py [testenv:docformatter_check] +# docformatter pinned due to https://github.com/PyCQA/docformatter/issues/264 deps = - docformatter + docformatter<1.7.4 untokenize commands = docformatter \

Conditionally set default esc_delay of inkey() from environment from https://github.com/jquast/blessed/issues/156#issuecomment-600790544 - We should honor this lesser known `ESCDELAY` environment value as the default value of `Terminal.inkey(esc_delay)` if set, https://stackoverflow.com/a/28020568 and https://invisible-island.net/ncurses/man/ncurses.3x.html > Specifies the total time, in milliseconds, for which ncurses will await a character sequence, e.g., a function key. The default value, 1000

2023-10-31T00:01:54

0.0

[]

dftbplus/skpar

dftbplus__skpar-7

8d6ac606fd8ddeb16e0bc8ca0a0a742d8d829ecd

diff --git a/skpar/core/input.py b/skpar/core/input.py index 16887d3..e72cd42 100644 --- a/skpar/core/input.py +++ b/skpar/core/input.py @@ -29,7 +29,7 @@ def get_input(filename): else: #if intype == 'yaml': try: - spec = yaml.load(infile) + spec = yaml.safe_load(infile) except yaml.YAMLError: LOGGER.warning('Input not a valid YAML') raise

Fix yaml loader error by using safe_load Invoking skpar currently results in a yaml parsing error, that is already fixed by f0a11be3600eb06320cc4325b4ddd840059d5299 and could be merged.

2022-09-20T09:06:59

0.0

[]

oasis-open/cti-taxii-server

oasis-open__cti-taxii-server-142

6fb6f9ddb50fe7fd5348a8ba84da8f421555d56a

diff --git a/medallion/__init__.py b/medallion/__init__.py index 215fb1bd..6d016880 100644 --- a/medallion/__init__.py +++ b/medallion/__init__.py @@ -18,7 +18,9 @@ log = logging.getLogger(__name__) log.addHandler(ch) + application_instance = Flask(__name__) +application_instance.app_context().push() auth = HTTPBasicAuth() @@ -30,7 +32,6 @@ def load_app(config_file): set_config(application_instance, "taxii", configuration) set_config(application_instance, "backend", configuration) register_blueprints(application_instance) - return application_instance @@ -46,11 +47,14 @@ def set_config(flask_application_instance, prop_name, config): try: flask_application_instance.users_backend = config[prop_name] except KeyError: - log.warning("You did not give user information in your config.") - log.warning("We are giving you the default user information of:") - log.warning("User = user") - log.warning("Pass = pass") - flask_application_instance.users_backend = {"user": "pass"} + if config.get('no_auth'): + flask_application_instance.config['no_auth'] = True + else: + log.warning("You did not give user information in your config.") + log.warning("We are giving you the default user information of:") + log.warning("User = user") + log.warning("Pass = pass") + flask_application_instance.users_backend = {"user": "pass"} elif prop_name == "backend": try: flask_application_instance.medallion_backend = connect_to_backend(config[prop_name]) diff --git a/medallion/scripts/run.py b/medallion/scripts/run.py index 93b42d9c..29358e47 100644 --- a/medallion/scripts/run.py +++ b/medallion/scripts/run.py @@ -57,6 +57,13 @@ def _get_argparser(): choices=["DEBUG", "INFO", "WARN", "ERROR", "CRITICAL"], ) + parser.add_argument( + "--no-auth", + default=False, + action="store_true", + help="Turn off credential checking for all endpoints.", + ) + parser.add_argument( "CONFIG_PATH", metavar="CONFIG_PATH", @@ -75,6 +82,9 @@ def main(): with open(medallion_args.CONFIG_PATH, "r") as f: configuration = json.load(f) + if medallion_args.no_auth: + configuration['no_auth'] = True + set_config(application_instance, "users", configuration) set_config(application_instance, "taxii", configuration) set_config(application_instance, "backend", configuration) diff --git a/medallion/views/__init__.py b/medallion/views/__init__.py index aee3adfb..47aeac15 100644 --- a/medallion/views/__init__.py +++ b/medallion/views/__init__.py @@ -1,6 +1,7 @@ import re from flask import request +from functools import wraps from ..exceptions import ProcessingError @@ -25,3 +26,15 @@ def validate_version_parameter_in_accept_header(): if found is False: raise ProcessingError("Media type in the Accept header is invalid or not found", 406) + + +def conditional_auth(func): + from .. import auth, current_app + + @wraps(func) + def wrapper(): + if current_app.config.get('no_auth'): + return func + else: + return auth.login_required(func) + return wrapper() diff --git a/medallion/views/collections.py b/medallion/views/collections.py index 702428a4..b1625c2a 100644 --- a/medallion/views/collections.py +++ b/medallion/views/collections.py @@ -1,7 +1,6 @@ from flask import Blueprint, Response, current_app, json -from . import MEDIA_TYPE_TAXII_V21, validate_version_parameter_in_accept_header -from .. import auth +from . import MEDIA_TYPE_TAXII_V21, validate_version_parameter_in_accept_header, conditional_auth from .discovery import api_root_exists from .objects import collection_exists @@ -9,7 +8,7 @@ @collections_bp.route("/<string:api_root>/collections/", methods=["GET"]) [email protected]_required +@conditional_auth def get_collections(api_root): """ Defines TAXII API - Collections: @@ -36,7 +35,7 @@ def get_collections(api_root): @collections_bp.route("/<string:api_root>/collections/<string:collection_id>/", methods=["GET"]) [email protected]_required +@conditional_auth def get_collection(api_root, collection_id): """ Defines TAXII API - Collections: diff --git a/medallion/views/discovery.py b/medallion/views/discovery.py index 2fea0c25..08d4d21e 100644 --- a/medallion/views/discovery.py +++ b/medallion/views/discovery.py @@ -1,7 +1,6 @@ from flask import Blueprint, Response, current_app, json -from . import MEDIA_TYPE_TAXII_V21, validate_version_parameter_in_accept_header -from .. import auth +from . import MEDIA_TYPE_TAXII_V21, validate_version_parameter_in_accept_header, conditional_auth from ..exceptions import ProcessingError discovery_bp = Blueprint("discovery", __name__) @@ -14,7 +13,7 @@ def api_root_exists(api_root): @discovery_bp.route("/taxii2/", methods=["GET"]) [email protected]_required +@conditional_auth def get_server_discovery(): """ Defines TAXII API - Server Information: @@ -30,7 +29,6 @@ def get_server_discovery(): # depending upon the credentials. validate_version_parameter_in_accept_header() server_discovery = current_app.medallion_backend.server_discovery() - if server_discovery: return Response( response=json.dumps(server_discovery), @@ -41,7 +39,7 @@ def get_server_discovery(): @discovery_bp.route("/<string:api_root>/", methods=["GET"]) [email protected]_required +@conditional_auth def get_api_root_information(api_root): """ Defines TAXII API - Server Information: @@ -67,7 +65,7 @@ def get_api_root_information(api_root): @discovery_bp.route("/<string:api_root>/status/<string:status_id>/", methods=["GET"]) [email protected]_required +@conditional_auth def get_status(api_root, status_id): """ Defines TAXII API - Server Information: diff --git a/medallion/views/manifest.py b/medallion/views/manifest.py index 35367904..cb869e4e 100644 --- a/medallion/views/manifest.py +++ b/medallion/views/manifest.py @@ -1,7 +1,6 @@ from flask import Blueprint, Response, current_app, json, request -from . import MEDIA_TYPE_TAXII_V21, validate_version_parameter_in_accept_header -from .. import auth +from . import MEDIA_TYPE_TAXII_V21, validate_version_parameter_in_accept_header, conditional_auth from .discovery import api_root_exists from .objects import ( collection_exists, permission_to_read, validate_limit_parameter @@ -11,7 +10,7 @@ @manifest_bp.route("/<string:api_root>/collections/<string:collection_id>/manifest/", methods=["GET"]) [email protected]_required +@conditional_auth def get_object_manifest(api_root, collection_id): """ Defines TAXII API - Collections: diff --git a/medallion/views/objects.py b/medallion/views/objects.py index 1d30c7c8..f796b423 100644 --- a/medallion/views/objects.py +++ b/medallion/views/objects.py @@ -3,8 +3,7 @@ from flask import Blueprint, Response, current_app, json, request -from . import MEDIA_TYPE_TAXII_V21, validate_version_parameter_in_accept_header -from .. import auth +from . import MEDIA_TYPE_TAXII_V21, validate_version_parameter_in_accept_header, conditional_auth from ..common import get_timestamp from ..exceptions import ProcessingError from .discovery import api_root_exists @@ -86,7 +85,7 @@ def validate_limit_parameter(): @objects_bp.route("/<string:api_root>/collections/<string:collection_id>/objects/", methods=["GET", "POST"]) [email protected]_required +@conditional_auth def get_or_add_objects(api_root, collection_id): """ Defines TAXII API - Collections: @@ -138,7 +137,7 @@ def get_or_add_objects(api_root, collection_id): @objects_bp.route("/<string:api_root>/collections/<string:collection_id>/objects/<string:object_id>/", methods=["GET", "DELETE"]) [email protected]_required +@conditional_auth def get_or_delete_object(api_root, collection_id, object_id): """ Defines TAXII API - Collections: @@ -187,7 +186,7 @@ def get_or_delete_object(api_root, collection_id, object_id): @objects_bp.route("/<string:api_root>/collections/<string:collection_id>/objects/<string:object_id>/versions/", methods=["GET"]) [email protected]_required +@conditional_auth def get_object_versions(api_root, collection_id, object_id): """ Defines TAXII API - Collections: Get Object Versions section

bug fix del _id key

# [Codecov](https://codecov.io/gh/oasis-open/cti-taxii-server/pull/4?src=pr&el=h1) Report > Merging [#4](https://codecov.io/gh/oasis-open/cti-taxii-server/pull/4?src=pr&el=desc) into [master](https://codecov.io/gh/oasis-open/cti-taxii-server/commit/eaa4eb69feafa2492eccc26d6d3743a4e87f94de?src=pr&el=desc) will **not change** coverage. > The diff coverage is `n/a`. [![Impacted file tree graph](https://codecov.io/gh/oasis-open/cti-taxii-server/pull/4/graphs/tree.svg?height=150&width=650&token=eu7JYJfW12&src=pr)](https://codecov.io/gh/oasis-open/cti-taxii-server/pull/4?src=pr&el=tree) ```diff @@ Coverage Diff @@ ## master #4 +/- ## ======================================= Coverage 14.99% 14.99% ======================================= Files 14 14 Lines 587 587 ======================================= Hits 88 88 Misses 499 499 ``` ------ [Continue to review full report at Codecov](https://codecov.io/gh/oasis-open/cti-taxii-server/pull/4?src=pr&el=continue). > **Legend** - [Click here to learn more](https://docs.codecov.io/docs/codecov-delta) > `Δ = absolute <relative> (impact)`, `ø = not affected`, `? = missing data` > Powered by [Codecov](https://codecov.io/gh/oasis-open/cti-taxii-server/pull/4?src=pr&el=footer). Last update [eaa4eb6...2bbbc9e](https://codecov.io/gh/oasis-open/cti-taxii-server/pull/4?src=pr&el=lastupdated). Read the [comment docs](https://docs.codecov.io/docs/pull-request-comments). This is a good fix - @gtback, should we merge in?? Thanks, @zhangymJLU . Before we can merge this, you'll need to sign the Individual Contributor License Agreement (CLA) located [here](https://www.oasis-open.org/resources/open-repositories/cla). If this is a problem, let us know. OK, I have accepted the license on this page and replay the confirm email now, @gtback. @zhangymJLU (Yangming Zhang) - Many thanks, and your CLA is now on file. -rcc

2021-03-15T13:50:26

0.0

[]

oasis-open/cti-taxii-server

oasis-open__cti-taxii-server-136

d1367f579eaac1b7b311a63da9df599afe6a9c3e

diff --git a/README.rst b/README.rst index 9eb60c60..c1d73dcc 100644 --- a/README.rst +++ b/README.rst @@ -52,17 +52,17 @@ Medallion provides a command-line interface to start the TAXII Server .. code-block:: text - usage: medallion [-h] - [--host HOST] - [--port PORT] - [--debug-mode] - [--log-level {DEBUG,INFO,WARN,ERROR,CRITICAL}] - CONFIG_PATH + usage: medallion [-h] [--host HOST] [--port PORT] [--debug-mode] + [--log-level {DEBUG,INFO,WARN,ERROR,CRITICAL}] + [-c CONF_FILE] [--conf-dir CONF_DIR | --no-conf-dir] + [--conf-check] [CONFIG_PATH] medallion v3.0.0 positional arguments: - CONFIG_PATH The location of the JSON configuration file to use. + CONFIG_PATH Deprecated argument for specifying a single + JSON configuration file. Do not specify this + and `--conf-file`. optional arguments: -h, --help show this help message and exit @@ -76,17 +76,35 @@ Medallion provides a command-line interface to start the TAXII Server --log-level {DEBUG,INFO,WARN,ERROR,CRITICAL} The logging output level for medallion. + -c CONF_FILE, --conf-file CONF_FILE + Path to a single configuration file. Defaults to the + value of the MEDALLION_CONFFILE environment variable + or /etc/xdg/medallion/3/medallion.conf. + + --conf-dir CONF_DIR Path to a directory containing JSON configuration + files with names ending in .json or .conf. Defaults to + the value of the MEDALLION_CONFDIR environment + variable or /etc/xdg/medallion/3/config.d. + + --no-conf-dir Disable the use of any configuration directory as + described for --conf-dir. This may be used to ensure + that the default or some value from the environment is + not used. + + --conf-check Evaluate medallion configuration without running the + server. + To run *medallion* .. code-block:: bash - $ python medallion/scripts/run.py <config-file> + $ python medallion/scripts/run.py --conf-file <config-file> Make sure medallion is using the same port that your TAXII client will be connecting on. You can specify which port medallion runs on using the `--port` option, for example .. code-block:: bash - $ medallion --port 80 config_file.json + $ medallion --port 80 --conf-file config_file.json The <config_file> contains: @@ -163,7 +181,7 @@ We also provide a Docker image to make it easier to run *medallion* .. code-block:: bash - $ docker build . -t medallion + $ docker build . -t medallion -f docker_utils/Dockerfile If operating behind a proxy, add the following option (replacing `<proxy>` with your proxy location and port): ``--build-arg https_proxy=<proxy>``. diff --git a/docker-compose.yml b/docker-compose.yml index 64c0fa65..d4139d18 100644 --- a/docker-compose.yml +++ b/docker-compose.yml @@ -7,7 +7,10 @@ services: build: context: . dockerfile: docker_utils/Dockerfile - command: [sh, -c, "medallion --host 0.0.0.0 docker_config.json"] + command: [sh, -c, "medallion --host 0.0.0.0"] + env: + MEDALLION_BACKEND_MODULE_CLASS: "MongoBackend" + MEDALLION_BACKEND_MONGO_URI: "mongodb://root:example@mongo:27017/" ports: - 5000:5000 mongo: diff --git a/docker_utils/Dockerfile b/docker_utils/Dockerfile index 62300d70..d748b001 100644 --- a/docker_utils/Dockerfile +++ b/docker_utils/Dockerfile @@ -1,8 +1,15 @@ FROM python:alpine COPY . /opt/taxii -COPY ./docker_utils/*.json /opt/taxii WORKDIR /opt/taxii RUN pip install --upgrade pip setuptools \ && pip install pymongo \ && pip install . + +# Set up the default configuration files +ARG MEDALLION_CONFFILE=/opt/taxii/medallion.conf +ENV MEDALLION_CONFFILE "${MEDALLION_CONFFILE}" +COPY ./docker_utils/base_config.json "${MEDALLION_CONFFILE}" +ARG MEDALLION_CONFDIR=/opt/taxii/medallion.d/ +ENV MEDALLION_CONFDIR "${MEDALLION_CONFDIR}" +COPY ./docker_utils/config.d/ "${MEDALLION_CONFDIR}/" diff --git a/docker_utils/base_config.json b/docker_utils/base_config.json new file mode 100644 index 00000000..e0fee00d --- /dev/null +++ b/docker_utils/base_config.json @@ -0,0 +1,5 @@ +{ + "taxii": { + "max_page_size": 100 + } +} diff --git a/docker_utils/config.d/backend.json b/docker_utils/config.d/backend.json new file mode 100644 index 00000000..52962eac --- /dev/null +++ b/docker_utils/config.d/backend.json @@ -0,0 +1,6 @@ +{ + "backend": { + "module_class": "MongoBackend", + "uri": "mongodb://DUMMYUSER:DUMMYPASSWORD@mongo:27017/" + } +} diff --git a/docker_utils/config.d/users.json b/docker_utils/config.d/users.json new file mode 100644 index 00000000..c75ae2f5 --- /dev/null +++ b/docker_utils/config.d/users.json @@ -0,0 +1,7 @@ +{ + "users": { + "admin": "Password0", + "user1": "Password1", + "user2": "Password2" + } +} diff --git a/docker_utils/docker_config.json b/docker_utils/docker_config.json deleted file mode 100644 index 80d47013..00000000 --- a/docker_utils/docker_config.json +++ /dev/null @@ -1,14 +0,0 @@ -{ - "backend": { - "module_class": "MongoBackend", - "uri": "mongodb://root:example@mongo:27017/" - }, - "users": { - "admin": "Password0", - "user1": "Password1", - "user2": "Password2" - }, - "taxii": { - "max_page_size": 100 - } -} diff --git a/medallion/backends/__init__.py b/medallion/backends/__init__.py index 10141298..1f393832 100644 --- a/medallion/backends/__init__.py +++ b/medallion/backends/__init__.py @@ -4,6 +4,7 @@ import pkgutil import sys +import environ import pkg_resources from . import base @@ -27,3 +28,17 @@ ep_obj = ep.load() if inspect.isclass(ep_obj): base.BackendRegistry.register(ep.name, ep_obj) + + +# Finally we define a top-level environ config class which includes and configs +# defined in the registered backends [email protected](prefix="BACKEND") +class BackendConfig(object): + for name, clsobj in base.BackendRegistry.iter_(): + # We have to use a magic attribute name here since `config`s don't have + # a specific mixin or type we can check for + try: + locals()[name] = environ.group(clsobj.Config, optional=True) + except AttributeError: + pass + module_class = environ.var(None) diff --git a/medallion/backends/base.py b/medallion/backends/base.py index 49ed4662..2cfae6d6 100644 --- a/medallion/backends/base.py +++ b/medallion/backends/base.py @@ -19,6 +19,10 @@ def register(mcls, kind, clsobj): def get(mcls, kind): return mcls.__SUBCLASS_MAP[kind] + @classmethod + def iter_(mcls): + yield from mcls.__SUBCLASS_MAP.items() + class Backend(object, metaclass=BackendRegistry): diff --git a/medallion/backends/memory_backend.py b/medallion/backends/memory_backend.py index 9fef670b..c844f47a 100644 --- a/medallion/backends/memory_backend.py +++ b/medallion/backends/memory_backend.py @@ -4,6 +4,7 @@ import os import uuid +import environ from six import string_types from ..common import ( @@ -36,6 +37,10 @@ class MemoryBackend(Backend): # access control is handled at the views level + @environ.config(prefix="MEMORY") + class Config(object): + filename = environ.var(None) + def __init__(self, **kwargs): # Refuse to run under a WSGI server since this is an internal backend if ( diff --git a/medallion/backends/mongodb_backend.py b/medallion/backends/mongodb_backend.py index 3413c7a8..00d44c88 100644 --- a/medallion/backends/mongodb_backend.py +++ b/medallion/backends/mongodb_backend.py @@ -1,6 +1,7 @@ import logging import uuid +import environ from pymongo import MongoClient from pymongo.errors import ConnectionFailure, ServerSelectionTimeoutError @@ -35,6 +36,10 @@ class MongoBackend(Backend): # access control is handled at the views level + @environ.config(prefix="MONGO") + class Config(object): + uri = environ.var() + def __init__(self, **kwargs): try: self.client = MongoClient(kwargs.get("uri")) diff --git a/medallion/config.py b/medallion/config.py new file mode 100644 index 00000000..af2c2ab0 --- /dev/null +++ b/medallion/config.py @@ -0,0 +1,172 @@ +import collections.abc +import json +import logging +import pathlib + +import appdirs +import attr +import environ +import jsonmerge +import packaging.version + +from . import backends +from . import version as _medallion_version + +# Find a config directory for the closest compatible version including the +# current one, or fall back to the current version's one if none are extant +_curr_version = packaging.version.Version(_medallion_version.__version__) +_compat_versions = {_curr_version.major, 3, 2} +for _cand_version in sorted(_compat_versions, reverse=True): + _appdirs = appdirs.AppDirs( + "medallion", "oasis-open", version=str(_cand_version) + ) + _default_conf_p = pathlib.Path(_appdirs.site_config_dir).resolve() + if _default_conf_p.is_dir(): + break +else: + _appdirs = appdirs.AppDirs( + "medallion", "oasis-open", version=str(_curr_version.major), + ) + _default_conf_p = pathlib.Path(_appdirs.site_config_dir).resolve() +# Set the default config file and config.d directory paths +DEFAULT_CONFFILE = _default_conf_p / "medallion.conf" +DEFAULT_CONFDIR = _default_conf_p / "config.d" +# Configuration directories are scanned for files matching these suffixes +CONFDIR_SUFFIXES = {".json", ".conf"} + +LOGGER = logging.getLogger(__name__) + + +class _LazyJSONDumper(object): + """An lazy stringifier for dumping objects as JSON, e.g. for logging.""" + def __init__(self, obj, cls=json.JSONEncoder, indent=None): + super(_LazyJSONDumper, self).__init__() + # Beware: If `obj` is mutable then we'll get future mutations. This + # isn't really an issue for logging format evaluation, but it might be + # in other contexts/multi-thread. We concretise on the first call to + # `__str__()` so the output's immutable/repeatable after that at least. + self._obj = obj + self._dumpcls = cls + self._indent = indent + + def __str__(self): + # The first stringification will concretise `obj` + if not isinstance(self._obj, str): + self._obj = json.dumps( + self._obj, cls=self._dumpcls, indent=self._indent, + ) + return self._obj + + def __repr__(self): # pragma: no cover + return repr(str(self)) + + [email protected](prefix="TAXII") +class TAXIIConfig(object): + max_page_size = environ.var(None, converter=lambda i: int(i) if i else i) + + [email protected](prefix="MEDALLION") +class MedallionConfig(object): + backend = environ.group(backends.BackendConfig) + taxii = environ.group(TAXIIConfig) + + @classmethod + def __strip(cls, dict_): + for k, v in tuple(dict_.items()): + if isinstance(v, dict): + cls.__strip(v) + if v is None or v == {}: + del dict_[k] + + def as_dict(self): + dictified = attr.asdict(self) + self.__strip(dictified) + return dictified + + +def _load_config_file(conf_file_p): + try: + config_data = json.load(conf_file_p.open()) + except json.decoder.JSONDecodeError as exc: + raise ValueError(f"Invalid JSON data in {conf_file_p}") from exc + if not isinstance(config_data, collections.abc.Mapping): + raise TypeError(f"{conf_file_p} must contain a JSON object") + return config_data + + +def load_config(conf_file=DEFAULT_CONFFILE, conf_dir=DEFAULT_CONFDIR): + LOGGER.debug( + "Attempting to load configuration from '%s' and '%s'", + conf_file, conf_dir, + ) + conf_files = list() + # Sanity check that both the `conf_file` and `conf_dir` exist by resolving + # them with strictness determined by whether they are the defaults or not + if conf_file is not None: + conf_file_p = pathlib.Path(conf_file).resolve( + strict=conf_file is not DEFAULT_CONFFILE + ) + conf_files.append(conf_file_p) + # Build a lexicographically sorted list of children of the `conf_dir` which + # are both files and have an acceptable file suffix. We allow the potential + # `NotADirectoryError` here to bubble up since `DEFAULT_CONFDIR` existing + # but not being a directory seems like a problem. There's an option for + # disabling the use of a config dir so we won't feel bad about doing this. + if conf_dir is not None: + conf_dir_p = pathlib.Path(conf_dir).resolve( + strict=conf_dir is not DEFAULT_CONFDIR + ) + conf_dir_files = ( + p for p in conf_dir_p.iterdir() if p.suffix in CONFDIR_SUFFIXES + ) + try: + conf_files.extend(sorted(conf_dir_files, key=lambda p: p.name)) + except FileNotFoundError: + pass + LOGGER.debug( + "Configuration files to load in order: %s", + ", ".join(repr(str(p)) for p in conf_files), + ) + # Start with an empty config and progressively merge in data from files + config_data = dict() + for conf_file_p in conf_files: + try: + new_data = _load_config_file(conf_file_p) + except (FileNotFoundError, IsADirectoryError): + # Skip missing files since we haven't already exploded which means + # they're probably not important enough to care about (or TOCTOU?). + # We also don't care about sub-directories of the config dir. + pass + else: + LOGGER.debug( + "Configuration data from '%s' to be merged: %s", + conf_file_p, _LazyJSONDumper(new_data, indent=2), + ) + config_data = jsonmerge.merge(config_data, new_data) + # Load extra configuration from the environment + env_config = MedallionConfig.from_environ().as_dict() + LOGGER.debug( + "Configuration data from environment to be merged: %s", + _LazyJSONDumper(env_config, indent=2), + ) + config_data = jsonmerge.merge(config_data, env_config) + # We promote config pulled from the environment for the specified backend + # `module_class` since the `environ-config` variable layout will nest it. + # Any of the following statements could `KeyError` if the configuration is + # incomplete or no config was sourced from the environment, but we're + # lenient about this here in favour of validating in the main app logic. + try: + backend_config = config_data["backend"] + backend_kind = backend_config["module_class"] + backend_kind_config = backend_config.pop(backend_kind) + except KeyError: + pass + else: + backend_config.update(backend_kind_config) + # Return the finalised configuration dictionary + LOGGER.debug( + "Merged configuration: %s", + _LazyJSONDumper(config_data, indent=2), + ) + return config_data diff --git a/medallion/scripts/run.py b/medallion/scripts/run.py index 93b42d9c..152e5027 100644 --- a/medallion/scripts/run.py +++ b/medallion/scripts/run.py @@ -1,11 +1,13 @@ import argparse -import json +import inspect import logging +import os import textwrap from medallion import ( __version__, application_instance, register_blueprints, set_config ) +import medallion.config log = logging.getLogger("medallion") @@ -57,11 +59,53 @@ def _get_argparser(): choices=["DEBUG", "INFO", "WARN", "ERROR", "CRITICAL"], ) - parser.add_argument( + config_path_group = parser.add_mutually_exclusive_group() + config_path_group.add_argument( "CONFIG_PATH", metavar="CONFIG_PATH", + nargs="?", type=str, - help="The location of the JSON configuration file to use.", + help=inspect.cleandoc(""" + Deprecated argument for specifying a single JSON configuration + file. Do not specify this and `--conf-file`. + """), + ) + config_path_group.add_argument( + "-c", "--conf-file", + default=os.environ.get( + "MEDALLION_CONFFILE", medallion.config.DEFAULT_CONFFILE + ), + help=inspect.cleandoc(f""" + Path to a single configuration file. Defaults to the value of + the MEDALLION_CONFFILE environment variable or + {medallion.config.DEFAULT_CONFFILE}. + """), + ) + config_dir_group = parser.add_mutually_exclusive_group() + config_dir_group.add_argument( + "--conf-dir", + default=os.environ.get( + "MEDALLION_CONFDIR", medallion.config.DEFAULT_CONFDIR + ), + help=inspect.cleandoc(f""" + Path to a directory containing JSON configuration files with names + ending in .json or .conf. Defaults to the value of the + MEDALLION_CONFDIR environment variable or + {medallion.config.DEFAULT_CONFDIR}. + """), + ) + config_dir_group.add_argument( + "--no-conf-dir", + action="store_true", + help=inspect.cleandoc(""" + Disable the use of any configuration directory as described for + --conf-dir. This may be used to ensure that the default or some + value from the environment is not used. + """), + ) + parser.add_argument( + "--conf-check", action="store_true", + help="Evaluate medallion configuration without running the server.", ) return parser @@ -70,21 +114,27 @@ def _get_argparser(): def main(): medallion_parser = _get_argparser() medallion_args = medallion_parser.parse_args() + # Configuration checking sets up debug logging and does not run the app + if medallion_args.conf_check: + medallion_args.log_level = logging.DEBUG log.setLevel(medallion_args.log_level) - with open(medallion_args.CONFIG_PATH, "r") as f: - configuration = json.load(f) + configuration = medallion.config.load_config( + medallion_args.CONFIG_PATH or medallion_args.conf_file, + medallion_args.conf_dir if not medallion_args.no_conf_dir else None, + ) set_config(application_instance, "users", configuration) set_config(application_instance, "taxii", configuration) set_config(application_instance, "backend", configuration) register_blueprints(application_instance) - application_instance.run( - host=medallion_args.host, - port=medallion_args.port, - debug=medallion_args.debug_mode, - ) + if not medallion_args.conf_check: + application_instance.run( + host=medallion_args.host, + port=medallion_args.port, + debug=medallion_args.debug_mode, + ) if __name__ == "__main__": diff --git a/setup.py b/setup.py index 78c28604..2bceede0 100644 --- a/setup.py +++ b/setup.py @@ -48,8 +48,12 @@ def get_long_description(): keywords="taxii taxii2 server json cti cyber threat intelligence", packages=find_packages(exclude=["*.test", "*.test.data"]), install_requires=[ + "appdirs>=1.4.4", + "environ-config>=21.1", "flask>=0.12.1", "Flask-HTTPAuth", + "jsonmerge", + "packaging", "pytz", "six", ], @@ -63,6 +67,7 @@ def get_long_description(): "coverage", "pytest", "pytest-cov", + "pytest-subtests", "tox", ], "docs": [ diff --git a/tox.ini b/tox.ini index e25a3608..fb9be89e 100644 --- a/tox.ini +++ b/tox.ini @@ -7,6 +7,7 @@ deps = tox pytest pytest-cov + pytest-subtests coverage responses pymongo

[RFC] Rework medallion configuration At the moment we pass a config file path to `medallion.load_app()` which is a bit of a pain and makes it a bit annoying to use a WSGI server. It would also be nice to move away from having a single config file to supporting something more like a conf.d structure with multiple JSON documents, as well as sourcing some configuration from the environment (e.g. 12 factor app style secrets). I've been looking into `environ-config` [0] which looks reasonable but might make the arbitrary kwarg passing currently done for backends painful. I've also been looking at `jsonmerge` [1] which should, in theory be capable of doing an ordered merge of documents in a medallion config.d directory. Would the following configuration hunting flow be sensible? Are there any other cases which should be handled? ![config-flow](https://user-images.githubusercontent.com/58414429/93841428-e7c6fc80-fcd6-11ea-97bc-3e44f2193054.png) [0] https://github.com/hynek/environ-config [1] https://github.com/avian2/jsonmerge

2020-09-23T02:04:00

0.0

[]

tomtalp/pyERGM

tomtalp__pyERGM-11

ffc1c8263b46f672534fe59266ca36374292bf96

diff --git a/ergm.py b/ergm.py index e7b335b..12e4d1f 100644 --- a/ergm.py +++ b/ergm.py @@ -96,8 +96,7 @@ def _get_random_thetas(self, sampling_method="uniform"): raise ValueError(f"Sampling method {sampling_method} not supported. See docs for supported samplers.") def generate_networks_for_sample(self, replace=True): - sampler = sampling.NaiveMetropolisHastings(self._thetas, self._network_statistics, - is_directed=self._is_directed) + sampler = sampling.NaiveMetropolisHastings(self._thetas, self._network_statistics) G = nx.erdos_renyi_graph(self._n_nodes, self._seed_MCMC_proba, directed=self._is_directed) seed_network = nx.to_numpy_array(G) @@ -164,7 +163,8 @@ def fit(self, observed_network, """ def nll_grad(thetas): - model = ERGM(self._n_nodes, self._network_statistics.metrics, initial_thetas=thetas, is_directed=self._is_directed) + model = ERGM(self._n_nodes, self._network_statistics.metrics, initial_thetas=thetas, + is_directed=self._is_directed) observed_features = model._network_statistics.calculate_statistics(observed_network) @@ -287,8 +287,7 @@ def sample_network(self, sampling_method="NaiveMetropolisHastings", seed_network The sampled connectivity matrix. """ if sampling_method == "NaiveMetropolisHastings": - sampler = sampling.NaiveMetropolisHastings(self._thetas, self._network_statistics, - is_directed=self._is_directed) + sampler = sampling.NaiveMetropolisHastings(self._thetas, self._network_statistics) else: raise ValueError(f"Sampling method {sampling_method} not supported. See docs for supported samplers.") diff --git a/metrics.py b/metrics.py index f865eeb..d77ee26 100644 --- a/metrics.py +++ b/metrics.py @@ -7,7 +7,6 @@ from utils import * -## TODO - Metrics needs to implement a diff() function class Metric(ABC): def __init__(self, metric_name, requires_graph=False): self.metric_name = metric_name @@ -26,7 +25,23 @@ def get_effective_feature_count(self, n): """ return 1 + def calc_change_score(self, net_1: np.ndarray | nx.Graph, net_2: np.ndarray | nx.Graph, is_turned_on: bool, + indices: tuple): + """ + The default naive way to calculate the change score (namely, the difference in statistics) of a pair of + networks. + Returns + ------- + statistic of net_2 minus statistic of net_1. + """ + net_2_stat = self.calculate(net_2) + net_1_stat = self.calculate(net_1) + change_score = net_2_stat - net_1_stat + return change_score + + +# TODO: override the change_score function with a more efficient calculation when possible. class NumberOfEdgesUndirected(Metric): def __init__(self): super().__init__(metric_name="num_edges_undirected", requires_graph=False) @@ -35,6 +50,9 @@ def __init__(self): def calculate(self, W: np.ndarray): return np.sum(W) // 2 + def calc_change_score(self, net_1: np.ndarray, net_2: np.ndarray, is_turned_on: bool, indices: tuple): + return 1 if is_turned_on else -1 + class NumberOfEdgesDirected(Metric): def __init__(self): @@ -44,17 +62,33 @@ def __init__(self): def calculate(self, W: np.ndarray): return np.sum(W) + def calc_change_score(self, net_1: np.ndarray, net_2: np.ndarray, is_turned_on: bool, indices: tuple): + return 1 if is_turned_on else -1 + +# TODO: change the name of this one to undirected and implement also a directed version? class NumberOfTriangles(Metric): def __init__(self): - super().__init__(metric_name="num_triangles", requires_graph=True) + super().__init__(metric_name="num_triangles", requires_graph=False) self._is_directed = False - def calculate(self, G: nx.Graph): - if isinstance(G, nx.DiGraph): + def calculate(self, W: np.ndarray): + if not np.all(W.T == W): raise ValueError("NumOfTriangles not implemented for directed graphs") - return sum( - nx.triangles(G).values()) // 3 # nx.triangles counts each triangle 3 times, for every one of its nodes + # the (i,j)-th entry of W^3 counts the number of 3-length paths from node i to node j. Thus, the i-th element on + # the diagonal counts the number of triangles that node 1 is part of (3-length paths from i to itself). As the + # graph is undirected, we get that each path is counted twice ("forward and backwards"), thus the division by 2. + # Additionally, each triangle is counted 3 times by diagonal elements (once for each node that takes part in + # forming it), thus the division by 3. + return (np.linalg.matrix_power(W, 3)).diagonal().sum() // (3 * 2) + + def calc_change_score(self, net_1: np.ndarray, net_2: np.ndarray, is_turned_on: bool, indices: tuple): + # The triangles that are affected by the edge toggling are those that involve it, namely, if the (i,j)-th edge + # is toggled, the change in absolute value equals to the number of nodes k for which the edges (i,k) and (j,k) + # exist. This is equivalent to the number of 2-length paths from i to j, which is the (i,j)-th entry of W^2. + # If the edge is turned on, the change is positive, and otherwise negative. + sign = 1 if is_turned_on else -1 + return sign * ((net_1 @ net_1)[indices[0], indices[1]]) class Reciprocity(Metric): @@ -75,6 +109,17 @@ def get_effective_feature_count(self, n): # n choose 2 return n * (n - 1) // 2 + def calc_change_score(self, net_1: np.ndarray, net_2: np.ndarray, is_turned_on: bool, indices: tuple): + # Note: we intentionally initialize the whole matrix and return np.triu_indices() by the end (rather than + # initializing an array of zeros of size n choose 2) to ensure compliance with the indexing returned by + # the calculate method. + all_changes = np.zeros(net_1.shape) + if net_1[indices[1], indices[0]] and is_turned_on: + all_changes[indices[0], indices[1]] = 1 + elif net_1[indices[1], indices[0]] and not is_turned_on: + all_changes[indices[0], indices[1]] = -1 + return all_changes[np.triu_indices(all_changes.shape[0], 1)] + class TotalReciprocity(Metric): """ @@ -88,6 +133,14 @@ def __init__(self): def calculate(self, W: np.ndarray): return (W * W.T).sum() / 2 + def calc_change_score(self, net_1: np.ndarray, net_2: np.ndarray, is_turned_on: bool, indices: tuple): + if net_1[indices[1], indices[0]] and is_turned_on: + return 1 + elif net_1[indices[1], indices[0]] and not is_turned_on: + return -1 + else: + return 0 + class MetricsCollection: def __init__(self, metrics: Collection[Metric], is_directed: bool): @@ -153,3 +206,32 @@ def calculate_statistics(self, W: np.ndarray): feature_idx += n_features_from_metric return statistics + + def calc_change_scores(self, W1: np.ndarray, W2: np.ndarray, is_turned_on: bool, indices: tuple): + """ + Calculates the vector of change scores, namely g(net_2) - g(net_1) + """ + if W1.shape != W2.shape: + raise ValueError(f"The dimensions of the given networks do not match! {W1.shape}, {W2.shape}") + if self.requires_graph: + G1 = connectivity_matrix_to_G(W1, directed=self.is_directed) + G2 = connectivity_matrix_to_G(W2, directed=self.is_directed) + + n_nodes = W1.shape[0] + change_scores = np.zeros(self.get_num_of_features(n_nodes)) + + feature_idx = 0 + for metric in self.metrics: + if metric.requires_graph: + inputs = (G1, G2) + else: + inputs = (W1, W2) + + n_features_from_metric = metric.get_effective_feature_count(n_nodes) + change_scores[feature_idx:feature_idx + n_features_from_metric] = metric.calc_change_score(inputs[0], + inputs[1], + is_turned_on, + indices) + feature_idx += n_features_from_metric + + return change_scores diff --git a/sampling.py b/sampling.py index 0b06e8c..863f833 100644 --- a/sampling.py +++ b/sampling.py @@ -5,16 +5,16 @@ class Sampler(): - def __init__(self, thetas, network_stats_calculator, is_directed=False): + def __init__(self, thetas, network_stats_calculator): self.thetas = deepcopy(thetas) self.network_stats_calculator = deepcopy(network_stats_calculator) - self.is_directed = is_directed def sample(self, initial_state, n_iterations): pass + class NaiveMetropolisHastings(Sampler): - def __init__(self, thetas, network_stats_calculator: MetricsCollection, is_directed=False, burn_in=1000, steps_per_sample=10): + def __init__(self, thetas, network_stats_calculator: MetricsCollection, burn_in=1000, steps_per_sample=10): """ An implementation for the symmetric proposal Metropolis-Hastings algorithm for ERGMS, using the logit of the acceptance rate. See docs for more details. @@ -27,12 +27,8 @@ def __init__(self, thetas, network_stats_calculator: MetricsCollection, is_direc network_stats_calculator : MetricsCollection A MetricsCollection object that can calculate statistics of a network. - - is_directed : bool - A boolean flag indicating whether the network is directed or not. - """ - super().__init__(thetas, network_stats_calculator, is_directed) + super().__init__(thetas, network_stats_calculator) ## TODO - these two params need to be dependent on the network size self.burn_in = burn_in @@ -43,22 +39,20 @@ def flip_network_edge(self, current_network, i, j): Flip the edge between nodes i, j. If it's an undirected network, we flip entries W_i,j and W_j,i. """ proposed_network = current_network.copy() + is_turned_on = not proposed_network[i, j] proposed_network[i, j] = 1 - proposed_network[i, j] - - if not self.is_directed: + + if not self.network_stats_calculator.is_directed: proposed_network[j, i] = 1 - proposed_network[j, i] - - return proposed_network - - # TODO - this needs to go to Metric() - def _calculate_weighted_change_score(self, proposed_network, current_network): + + return proposed_network, is_turned_on + + def _calculate_weighted_change_score(self, proposed_network, current_network, is_turned_on, indices: tuple): """ Calculate g(proposed_network)-g(current_network) and then inner product with thetas. """ - g_proposed = self.network_stats_calculator.calculate_statistics(proposed_network) - g_current = self.network_stats_calculator.calculate_statistics(current_network) - change_score = g_proposed - g_current - + change_score = self.network_stats_calculator.calc_change_scores(current_network, proposed_network, is_turned_on, + indices) return np.dot(self.thetas, change_score) def sample(self, initial_state, num_of_nets, replace=True): @@ -87,10 +81,11 @@ def sample(self, initial_state, num_of_nets, replace=True): while networks_count != num_of_nets: random_entry = get_random_nondiagonal_matrix_entry(current_network.shape[0]) - - proposed_network = self.flip_network_edge(current_network, random_entry[0], random_entry[1]) - change_score = self._calculate_weighted_change_score(proposed_network, current_network) + proposed_network, is_turned_on = self.flip_network_edge(current_network, random_entry[0], random_entry[1]) + + change_score = self._calculate_weighted_change_score(proposed_network, current_network, is_turned_on, + random_entry) if change_score >= 1: current_network = proposed_network.copy() @@ -98,18 +93,18 @@ def sample(self, initial_state, num_of_nets, replace=True): acceptance_proba = min(1, np.exp(change_score)) if np.random.rand() <= acceptance_proba: current_network = proposed_network.copy() - + if (mcmc_iter_count - self.burn_in) % self.steps_per_sample == 0: sampled_networks[:, :, networks_count] = current_network if not replace: - if np.unique(sampled_networks[:, :, :networks_count+1], axis=2).shape[2] == networks_count + 1: + if np.unique(sampled_networks[:, :, :networks_count + 1], axis=2).shape[2] == networks_count + 1: networks_count += 1 else: - sampled_networks[:, :, networks_count] = np.zeros((net_size, net_size)) + sampled_networks[:, :, networks_count] = np.zeros((net_size, net_size)) else: networks_count += 1 mcmc_iter_count += 1 - - return sampled_networks \ No newline at end of file + + return sampled_networks

Optim enhance Exported the `change_score` function from the sampler to be a method of `Metric`, such that it can be overridden by custom efficient calculations for each `Metric`. Implemented such efficient calculations for the existing `Metric`s.

2024-08-04T13:45:58

0.0

[]

tomtalp/pyERGM

tomtalp__pyERGM-3

9ccbc82abb92c35e797c4fab2655f29222caf0a0

diff --git a/.gitignore b/.gitignore index 82f9275..7b6caf3 100644 --- a/.gitignore +++ b/.gitignore @@ -159,4 +159,4 @@ cython_debug/ # be found at https://github.com/github/gitignore/blob/main/Global/JetBrains.gitignore # and can be added to the global gitignore or merged into this file. For a more nuclear # option (not recommended) you can uncomment the following to ignore the entire idea folder. -#.idea/ +.idea/ diff --git a/.idea/.gitignore b/.idea/.gitignore new file mode 100644 index 0000000..13566b8 --- /dev/null +++ b/.idea/.gitignore @@ -0,0 +1,8 @@ +# Default ignored files +/shelf/ +/workspace.xml +# Editor-based HTTP Client requests +/httpRequests/ +# Datasource local storage ignored files +/dataSources/ +/dataSources.local.xml diff --git a/.idea/inspectionProfiles/profiles_settings.xml b/.idea/inspectionProfiles/profiles_settings.xml new file mode 100644 index 0000000..105ce2d --- /dev/null +++ b/.idea/inspectionProfiles/profiles_settings.xml @@ -0,0 +1,6 @@ +<component name="InspectionProjectProfileManager"> + <settings> + <option name="USE_PROJECT_PROFILE" value="false" /> + <version value="1.0" /> + </settings> +</component> \ No newline at end of file diff --git a/.idea/misc.xml b/.idea/misc.xml new file mode 100644 index 0000000..e7c17b2 --- /dev/null +++ b/.idea/misc.xml @@ -0,0 +1,4 @@ +<?xml version="1.0" encoding="UTF-8"?> +<project version="4"> + <component name="ProjectRootManager" version="2" project-jdk-name="Python 3.11 (pyERGM)" project-jdk-type="Python SDK" /> +</project> \ No newline at end of file diff --git a/.idea/modules.xml b/.idea/modules.xml new file mode 100644 index 0000000..5ba93c2 --- /dev/null +++ b/.idea/modules.xml @@ -0,0 +1,8 @@ +<?xml version="1.0" encoding="UTF-8"?> +<project version="4"> + <component name="ProjectModuleManager"> + <modules> + <module fileurl="file://$PROJECT_DIR$/.idea/pyERGM.iml" filepath="$PROJECT_DIR$/.idea/pyERGM.iml" /> + </modules> + </component> +</project> \ No newline at end of file diff --git a/.idea/pyERGM.iml b/.idea/pyERGM.iml new file mode 100644 index 0000000..54c4c0a --- /dev/null +++ b/.idea/pyERGM.iml @@ -0,0 +1,14 @@ +<?xml version="1.0" encoding="UTF-8"?> +<module type="PYTHON_MODULE" version="4"> + <component name="NewModuleRootManager"> + <content url="file://$MODULE_DIR$"> + <excludeFolder url="file://$MODULE_DIR$/venv" /> + </content> + <orderEntry type="jdk" jdkName="Python 3.11 (pyERGM)" jdkType="Python SDK" /> + <orderEntry type="sourceFolder" forTests="false" /> + </component> + <component name="PyDocumentationSettings"> + <option name="format" value="NUMPY" /> + <option name="myDocStringFormat" value="NumPy" /> + </component> +</module> \ No newline at end of file diff --git a/.idea/vcs.xml b/.idea/vcs.xml new file mode 100644 index 0000000..35eb1dd --- /dev/null +++ b/.idea/vcs.xml @@ -0,0 +1,6 @@ +<?xml version="1.0" encoding="UTF-8"?> +<project version="4"> + <component name="VcsDirectoryMappings"> + <mapping directory="" vcs="Git" /> + </component> +</project> \ No newline at end of file diff --git a/ergm.py b/ergm.py index ecc55ad..e0c32ac 100644 --- a/ergm.py +++ b/ergm.py @@ -1,6 +1,7 @@ import numpy as np import networkx as nx -from scipy.optimize import minimize +from scipy.optimize import minimize, OptimizeResult +import time import sampling @@ -8,13 +9,17 @@ class ERGM(): - def __init__(self, - n_nodes, - network_statistics: NetworkStatistics, - is_directed=False, - initial_thetas=None, + def __init__(self, + n_nodes, + network_statistics: NetworkStatistics, + is_directed=False, + initial_thetas=None, initial_normalization_factor=None, - seed_MCMC_proba=0.25): + seed_MCMC_proba=0.25, + n_networks_for_norm=100, + n_mcmc_steps=500, + verbose=True, + optimization_options={}): """ An ERGM model object. @@ -40,21 +45,28 @@ def __init__(self, """ self._n_nodes = n_nodes self._network_statistics = network_statistics - + if initial_thetas is not None: self._thetas = initial_thetas else: self._thetas = self._get_random_thetas(sampling_method="uniform") - + if initial_normalization_factor is not None: self._normalization_factor = initial_normalization_factor else: self._normalization_factor = np.random.normal(50, 10) - self._is_directed = is_directed self._seed_MCMC_proba = seed_MCMC_proba + self.optimization_iter = 0 + self.optimization_start_time = None + + self.n_networks_for_norm = n_networks_for_norm + self.n_mcmc_steps = n_mcmc_steps + self.verbose = verbose + self.optimization_options = optimization_options + def print_model_parameters(self): print(f"Number of nodes: {self._n_nodes}") print(f"Thetas: {self._thetas}") @@ -63,11 +75,14 @@ def print_model_parameters(self): # print(f"Network statistics: {self._network_statistics}") def calculate_weight(self, W: np.ndarray): + if len(W.shape) != 2 or W.shape[0] != self._n_nodes or W.shape[1] != self._n_nodes: + raise ValueError(f"The dimensions of the given adjacency matrix, {W.shape}, don't comply with the number of" + f" nodes in the network: {self._n_nodes}") features = self._network_statistics.calculate_statistics(W) weight = np.exp(np.dot(self._thetas, features)) return weight - + def _get_random_thetas(self, sampling_method="uniform"): if sampling_method == "uniform": return np.random.uniform(-1, 1, self._network_statistics.get_num_of_statistics()) @@ -79,19 +94,19 @@ def _generate_networks_for_sample(self, n_networks, n_mcmc_steps): for _ in range(n_networks): net = self.sample_network(steps=n_mcmc_steps, sampling_method="NaiveMetropolisHastings") networks.append(net) - + return networks - def _approximate_normalization_factor(self, n_networks, n_mcmc_steps): - networks_for_sample = self._generate_networks_for_sample(n_networks, n_mcmc_steps) - + def _approximate_normalization_factor(self): + networks_for_sample = self._generate_networks_for_sample(self.n_networks, self.n_mcmc_steps) + self._normalization_factor = 0 for network in networks_for_sample: weight = self.calculate_weight(network) self._normalization_factor += weight - def fit(self, observed_network, n_networks_for_norm=100, n_mcmc_steps=500, verbose=True, optimization_options={}): + def fit(self, observed_network): """ Initial version, a simple MLE calculated by minimizing negative log likelihood. Normalizaiton factor is approximated via MCMC. @@ -100,12 +115,11 @@ def fit(self, observed_network, n_networks_for_norm=100, n_mcmc_steps=500, verbo """ self._thetas = self._get_random_thetas(sampling_method="uniform") - if verbose: + if self.verbose: print(f"\tStarting fit with initial normalization factor: {self._normalization_factor}") - if optimization_options != {}: - print(f"\tOptimization options: {optimization_options}") - + if self.optimization_options != {}: + print(f"\tOptimization options: {self.optimization_options}") def negative_log_likelihood(thetas): """ @@ -115,8 +129,8 @@ def negative_log_likelihood(thetas): """ # print(f"""Calculating negative log likelihood for thetas: {thetas}""") self._thetas = thetas - - self._approximate_normalization_factor(n_networks_for_norm, n_mcmc_steps) + + self._approximate_normalization_factor() Z = self._normalization_factor # print(f"Approximated Z - {Z}") @@ -125,8 +139,9 @@ def negative_log_likelihood(thetas): log_likelihood = np.log(y_observed_weight) - np.log(Z) return -log_likelihood - - result = minimize(negative_log_likelihood, self._thetas, method='Nelder-Mead', options=optimization_options) + + result = minimize(negative_log_likelihood, self._thetas, method='Nelder-Mead', + options=self.optimization_options) self._thetas = result.x print("\tOptimization result:") print(f"\tTheta: {self._thetas}") @@ -153,7 +168,7 @@ def calculate_probability(self, W: np.ndarray): weight = self.calculate_weight(W) prob = weight / self._normalization_factor - + return prob def sample_network(self, sampling_method="NaiveMetropolisHastings", seed_network=None, steps=500): @@ -175,10 +190,11 @@ def sample_network(self, sampling_method="NaiveMetropolisHastings", seed_network The sampled connectivity matrix. """ if sampling_method == "NaiveMetropolisHastings": - sampler = sampling.NaiveMetropolisHastings(self._thetas, self._network_statistics, is_directed=self._is_directed) + sampler = sampling.NaiveMetropolisHastings(self._thetas, self._network_statistics, + is_directed=self._is_directed) else: raise ValueError(f"Sampling method {sampling_method} not supported. See docs for supported samplers.") - + if seed_network is None: G = nx.erdos_renyi_graph(self._n_nodes, self._seed_MCMC_proba, directed=self._is_directed) seed_network = nx.to_numpy_array(G) @@ -187,6 +203,101 @@ def sample_network(self, sampling_method="NaiveMetropolisHastings", seed_network return network + +class BruteForceERGM(ERGM): + """ + A class that implements ERGM by iterating over the entire space of networks and calculating stuff exactly (rather + than using statistical methods for approximating and sampling). + This is mainly for tests. + """ + # The maximum number of nodes that is allowed for carrying brute force calculations (i.e. iterating the whole space + # of networks and calculating stuff exactly). This becomes not tractable above this limit. + MAX_NODES_BRUTE_FORCE_DIRECTED = 5 + MAX_NODES_BRUTE_FORCE_NOT_DIRECTED = 7 + + def __init__(self, + n_nodes, + network_statistics: NetworkStatistics, + is_directed=False, + initial_thetas=None): + super().__init__(n_nodes, + network_statistics, + is_directed, + initial_thetas) + self._all_weights = self._calc_all_weights() + self._normalization_factor = self._all_weights.sum() + + def _validate_net_size(self): + return ( + (self._is_directed and self._n_nodes <= BruteForceERGM.MAX_NODES_BRUTE_FORCE_DIRECTED) + or + (not self._is_directed and self._n_nodes <= BruteForceERGM.MAX_NODES_BRUTE_FORCE_NOT_DIRECTED) + ) + + def _calc_all_weights(self): + if not self._validate_net_size(): + directed_str = 'directed' if self._is_directed else 'not directed' + size_limit = BruteForceERGM.MAX_NODES_BRUTE_FORCE_DIRECTED if self._is_directed else ( + BruteForceERGM.MAX_NODES_BRUTE_FORCE_NOT_DIRECTED) + raise ValueError( + f"The number of nodes {self._n_nodes} is larger than the maximum allowed for brute force of " + f"{directed_str} graphs calculations {size_limit}") + num_pos_connects = self._n_nodes * (self._n_nodes - 1) + if not self._is_directed: + num_pos_connects //= 2 + space_size = 2 ** num_pos_connects + all_weights = np.zeros(space_size) + for i in range(space_size): + cur_adj_mat = construct_adj_mat_from_int(i, self._n_nodes, self._is_directed) + all_weights[i] = super().calculate_weight(cur_adj_mat) + return all_weights + + def calculate_weight(self, W: np.ndarray): + adj_mat_idx = construct_int_from_adj_mat(W, self._is_directed) + return self._all_weights[adj_mat_idx] + + def sample_network(self, sampling_method="Exact", seed_network=None, steps=0): + if sampling_method != "Exact": + raise ValueError("BruteForceERGM supports only exact sampling (this is its whole purpose)") + + all_nets_probs = self._all_weights / self._normalization_factor + sampled_idx = np.random.choice(all_nets_probs.size, p=all_nets_probs) + return construct_adj_mat_from_int(sampled_idx, self._n_nodes, self._is_directed) + + def fit(self, observed_network): + def nll(thetas): + model = BruteForceERGM(self._n_nodes, self._network_statistics, initial_thetas=thetas, + is_directed=self._is_directed) + return np.log(model._normalization_factor) - np.log(model.calculate_weight(observed_network)) + + def nll_grad(thetas): + model = BruteForceERGM(self._n_nodes, self._network_statistics, initial_thetas=thetas, + is_directed=self._is_directed) + observed_features = model._network_statistics.calculate_statistics(observed_network) + all_probs = model._all_weights / model._normalization_factor + num_features = model._network_statistics.get_num_of_statistics() + num_nets = all_probs.size + all_features_by_all_nets = np.zeros((num_features, num_nets)) + for i in range(num_nets): + all_features_by_all_nets[:, i] = model._network_statistics.calculate_statistics( + construct_adj_mat_from_int(i, self._n_nodes, self._is_directed)) + expected_features = all_features_by_all_nets @ all_probs + return expected_features - observed_features + + def after_iteration_callback(intermediate_result: OptimizeResult): + self.optimization_iter += 1 + cur_time = time.time() + print(f'iteration: {self.optimization_iter}, time from start ' + f'training: {cur_time - self.optimization_start_time} ' + f'log likelihood: {-intermediate_result.fun}') + + self.optimization_iter = 0 + print("optimization started") + self.optimization_start_time = time.time() + res = minimize(nll, self._thetas, jac=nll_grad, callback=after_iteration_callback) + self._thetas = res.x + print(res) + # n_nodes = 5 # stats_calculator = NetworkStatistics(metric_names=["num_edges"]) # ergm = ERGM(n_nodes, stats_calculator, is_directed=False) @@ -250,6 +361,3 @@ def sample_network(self, sampling_method="NaiveMetropolisHastings", seed_network # print(samples_before_fit) # print("Samples after fit:") # print(samples_after_fit) - - - \ No newline at end of file diff --git a/examples/first_example.ipynb b/examples/first_example.ipynb index c7d6b3a..da9ba5e 100644 --- a/examples/first_example.ipynb +++ b/examples/first_example.ipynb @@ -2,25 +2,45 @@ "cells": [ { "cell_type": "code", - "execution_count": 1, - "metadata": {}, + "execution_count": 6, + "metadata": { + "ExecuteTime": { + "end_time": "2024-07-21T12:28:56.044446900Z", + "start_time": "2024-07-21T12:28:55.917504Z" + } + }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ - "/Users/tomtalpir/Random/pyERGM\n" + "D:\\OrenRichter\\Research\\pyERGM\n" + ] + }, + { + "name": "stderr", + "output_type": "stream", + "text": [ + "D:\\OrenRichter\\Research\\pyERGM\\venv\\Lib\\site-packages\\IPython\\core\\magics\\osm.py:417: UserWarning: This is now an optional IPython functionality, setting dhist requires you to install the `pickleshare` library.\n", + " self.shell.db['dhist'] = compress_dhist(dhist)[-100:]\n" ] } ], "source": [ - "%cd /Users/tomtalpir/Random/pyERGM" + "from pathlib import Path\n", + "parent_dir = str(Path.cwd().parent)\n", + "%cd $parent_dir" ] }, { "cell_type": "code", - "execution_count": 2, - "metadata": {}, + "execution_count": 8, + "metadata": { + "ExecuteTime": { + "end_time": "2024-07-21T12:30:25.054064700Z", + "start_time": "2024-07-21T12:30:23.471407200Z" + } + }, "outputs": [], "source": [ "from utils import *\n", @@ -39,16 +59,21 @@ }, { "cell_type": "code", - "execution_count": 3, - "metadata": {}, + "execution_count": 9, + "metadata": { + "ExecuteTime": { + "end_time": "2024-07-21T13:32:42.547782400Z", + "start_time": "2024-07-21T13:32:42.531930200Z" + } + }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Number of nodes: 10\n", - "Thetas: [-0.29766759]\n", - "Normalization factor approx: 50.779457576278446\n", + "Thetas: [0.33527386]\n", + "Normalization factor approx: 36.49400038899532\n", "Is directed: False\n" ] } @@ -70,104 +95,20 @@ }, { "cell_type": "code", - "execution_count": 4, - "metadata": {}, + "execution_count": 10, + "metadata": { + "ExecuteTime": { + "end_time": "2024-07-21T13:33:29.562627900Z", + "start_time": "2024-07-21T13:33:26.542060800Z" + } + }, "outputs": [ { "data": { - "text/html": [ - "<div>\n", - "<style scoped>\n", - " .dataframe tbody tr th:only-of-type {\n", - " vertical-align: middle;\n", - " }\n", - "\n", - " .dataframe tbody tr th {\n", - " vertical-align: top;\n", - " }\n", - "\n", - " .dataframe thead th {\n", - " text-align: right;\n", - " }\n", - "</style>\n", - "<table border=\"1\" class=\"dataframe\">\n", - " <thead>\n", - " <tr style=\"text-align: right;\">\n", - " <th></th>\n", - " <th>edge_count</th>\n", - " <th>triangle_count</th>\n", - " </tr>\n", - " </thead>\n", - " <tbody>\n", - " <tr>\n", - " <th>0</th>\n", - " <td>33</td>\n", - " <td>47</td>\n", - " </tr>\n", - " <tr>\n", - " <th>1</th>\n", - " <td>35</td>\n", - " <td>57</td>\n", - " </tr>\n", - " <tr>\n", - " <th>2</th>\n", - " <td>36</td>\n", - " <td>62</td>\n", - " </tr>\n", - " <tr>\n", - " <th>3</th>\n", - " <td>36</td>\n", - " <td>57</td>\n", - " </tr>\n", - " <tr>\n", - " <th>4</th>\n", - " <td>39</td>\n", - " <td>79</td>\n", - " </tr>\n", - " <tr>\n", - " <th>5</th>\n", - " <td>36</td>\n", - " <td>63</td>\n", - " </tr>\n", - " <tr>\n", - " <th>6</th>\n", - " <td>34</td>\n", - " <td>53</td>\n", - " </tr>\n", - " <tr>\n", - " <th>7</th>\n", - " <td>38</td>\n", - " <td>76</td>\n", - " </tr>\n", - " <tr>\n", - " <th>8</th>\n", - " <td>38</td>\n", - " <td>72</td>\n", - " </tr>\n", - " <tr>\n", - " <th>9</th>\n", - " <td>32</td>\n", - " <td>46</td>\n", - " </tr>\n", - " </tbody>\n", - "</table>\n", - "</div>" - ], - "text/plain": [ - " edge_count triangle_count\n", - "0 33 47\n", - "1 35 57\n", - "2 36 62\n", - "3 36 57\n", - "4 39 79\n", - "5 36 63\n", - "6 34 53\n", - "7 38 76\n", - "8 38 72\n", - "9 32 46" - ] + "text/plain": " edge_count triangle_count\n0 45 120\n1 45 120\n2 45 120\n3 45 120\n4 45 120\n5 45 120\n6 45 120\n7 45 120\n8 45 120\n9 45 120", + "text/html": "<div>\n<style scoped>\n .dataframe tbody tr th:only-of-type {\n vertical-align: middle;\n }\n\n .dataframe tbody tr th {\n vertical-align: top;\n }\n\n .dataframe thead th {\n text-align: right;\n }\n</style>\n<table border=\"1\" class=\"dataframe\">\n <thead>\n <tr style=\"text-align: right;\">\n <th></th>\n <th>edge_count</th>\n <th>triangle_count</th>\n </tr>\n </thead>\n <tbody>\n <tr>\n <th>0</th>\n <td>45</td>\n <td>120</td>\n </tr>\n <tr>\n <th>1</th>\n <td>45</td>\n <td>120</td>\n </tr>\n <tr>\n <th>2</th>\n <td>45</td>\n <td>120</td>\n </tr>\n <tr>\n <th>3</th>\n <td>45</td>\n <td>120</td>\n </tr>\n <tr>\n <th>4</th>\n <td>45</td>\n <td>120</td>\n </tr>\n <tr>\n <th>5</th>\n <td>45</td>\n <td>120</td>\n </tr>\n <tr>\n <th>6</th>\n <td>45</td>\n <td>120</td>\n </tr>\n <tr>\n <th>7</th>\n <td>45</td>\n <td>120</td>\n </tr>\n <tr>\n <th>8</th>\n <td>45</td>\n <td>120</td>\n </tr>\n <tr>\n <th>9</th>\n <td>45</td>\n <td>120</td>\n </tr>\n </tbody>\n</table>\n</div>" }, - "execution_count": 4, + "execution_count": 10, "metadata": {}, "output_type": "execute_result" } @@ -202,8 +143,13 @@ }, { "cell_type": "code", - "execution_count": 42, - "metadata": {}, + "execution_count": 11, + "metadata": { + "ExecuteTime": { + "end_time": "2024-07-21T13:33:52.394915900Z", + "start_time": "2024-07-21T13:33:52.379092600Z" + } + }, "outputs": [ { "name": "stdout", @@ -211,8 +157,8 @@ "text": [ "Baseline ERGM parameters - \n", "Number of nodes: 5\n", - "Thetas: [-1.0986122886681098]\n", - "Normalization factor approx: 19.45775812919028\n", + "Thetas: [np.float64(-1.0986122886681098)]\n", + "Normalization factor approx: 43.441946503045074\n", "Is directed: False\n" ] } @@ -237,20 +183,25 @@ }, { "cell_type": "code", - "execution_count": 43, - "metadata": {}, + "execution_count": 12, + "metadata": { + "ExecuteTime": { + "end_time": "2024-07-21T13:34:07.998379900Z", + "start_time": "2024-07-21T13:34:07.904510500Z" + } + }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Sampled a random network from a model with theta = -1.0986122886681098\n", - "[[0. 0. 0. 1. 1.]\n", - " [0. 0. 1. 1. 1.]\n", - " [0. 1. 0. 0. 1.]\n", - " [1. 1. 0. 0. 1.]\n", - " [1. 1. 1. 1. 0.]]\n", - "Network has statistics - edge count: 7, triangle count: 3\n" + "[[0. 0. 1. 1. 1.]\n", + " [0. 0. 0. 1. 0.]\n", + " [1. 0. 0. 0. 0.]\n", + " [1. 1. 0. 0. 0.]\n", + " [1. 0. 0. 0. 0.]]\n", + "Network has statistics - edge count: 4, triangle count: 0\n" ] } ], @@ -276,8 +227,13 @@ }, { "cell_type": "code", - "execution_count": 46, - "metadata": {}, + "execution_count": 13, + "metadata": { + "ExecuteTime": { + "end_time": "2024-07-21T13:35:20.570400200Z", + "start_time": "2024-07-21T13:35:19.572400900Z" + } + }, "outputs": [ { "name": "stdout", @@ -285,13 +241,13 @@ "text": [ "Initial ERGM parameters:\n", "Number of nodes: 5\n", - "Thetas: [-0.33698419]\n", - "Normalization factor approx: 43.1253267687915\n", + "Thetas: [0.38415461]\n", + "Normalization factor approx: 50.33647025496875\n", "Is directed: False\n", "\n", "Mean of samples before fitting, calculated on 10 samples- \n", - "edge_count 8.0\n", - "triangle_count 5.0\n", + "edge_count 10.0\n", + "triangle_count 10.0\n", "dtype: float64\n" ] } @@ -331,28 +287,33 @@ }, { "cell_type": "code", - "execution_count": 47, - "metadata": {}, + "execution_count": 14, + "metadata": { + "ExecuteTime": { + "end_time": "2024-07-21T14:09:49.796634200Z", + "start_time": "2024-07-21T13:35:32.539389400Z" + } + }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Fitting ERGM...\n", - "\tStarting fit with initial normalization factor: 43.1253267687915\n", + "\tStarting fit with initial normalization factor: 50.33647025496875\n", "\tOptimization options: {'disp': True, 'maxiter': 500, 'maxfev': 500}\n", "\tOptimization result:\n", - "\tTheta: [-0.15381744]\n", - "\tNormalization factor: 135.44816303424523\n", + "\tTheta: [-0.39642108]\n", + "\tNormalization factor: 41.88198286288824\n", " message: Maximum number of function evaluations has been exceeded.\n", " success: False\n", " status: 1\n", - " fun: 5.963890498779282\n", - " x: [-1.538e-01]\n", - " nit: 186\n", + " fun: 5.243735736850018\n", + " x: [-3.964e-01]\n", + " nit: 189\n", " nfev: 500\n", - " final_simplex: (array([[-1.538e-01],\n", - " [-1.538e-01]]), array([ 5.964e+00, 5.984e+00]))\n", + " final_simplex: (array([[-3.964e-01],\n", + " [-3.964e-01]]), array([ 5.244e+00, 5.271e+00]))\n", "Done fitting!\n" ] }, @@ -360,8 +321,8 @@ "name": "stderr", "output_type": "stream", "text": [ - "/Users/tomtalpir/Random/pyERGM/ergm.py:130: RuntimeWarning: Maximum number of function evaluations has been exceeded.\n", - " result = minimize(negative_log_likelihood, self._thetas, method='Nelder-Mead', options=optimization_options)\n" + "D:\\OrenRichter\\Research\\pyERGM\\ergm.py:129: RuntimeWarning: Maximum number of function evaluations has been exceeded.\n", + " # OR: Providing the Scipy optimizer with a gradient function can streamline the optimization substantially. The\n" ] } ], diff --git a/sampling.py b/sampling.py index ab29c64..a1a07e3 100644 --- a/sampling.py +++ b/sampling.py @@ -55,7 +55,7 @@ def sample(self, initial_state, n_iterations): for i in range(n_iterations): random_entry = get_random_nondiagonal_matrix_entry(current_network.shape[0]) - + y_plus = self.override_network_edge(current_network, random_entry[0], random_entry[1], 1) y_minus = self.override_network_edge(current_network, random_entry[0], random_entry[1], 0) diff --git a/utils.py b/utils.py index d2d906c..9b5c430 100644 --- a/utils.py +++ b/utils.py @@ -1,6 +1,7 @@ import numpy as np import networkx as nx + def perturb_network_by_overriding_edge(network, value, i, j, is_directed): perturbed_net = network.copy() perturbed_net[i, j] = value @@ -9,6 +10,7 @@ def perturb_network_by_overriding_edge(network, value, i, j, is_directed): return perturbed_net + def connectivity_matrix_to_G(W: np.ndarray, directed=False): """ Convert a connectivity matrix to a graph object. @@ -30,6 +32,7 @@ def connectivity_matrix_to_G(W: np.ndarray, directed=False): G = nx.from_numpy_array(W) return G + def get_random_nondiagonal_matrix_entry(n: int): """ Get a random entry for a non-diagonal entry of a matrix. @@ -46,13 +49,14 @@ def get_random_nondiagonal_matrix_entry(n: int): """ i = np.random.randint(n) j = np.random.randint(n) - + while i == j: i = np.random.randint(n) j = np.random.randint(n) - + return (i, j) + class NetworkStatistics(): def __init__(self, metric_names=[], custom_metrics={}, directed=False): """ @@ -87,10 +91,10 @@ def __init__(self, metric_names=[], custom_metrics={}, directed=False): self._validate_predefined_metrics() self.custom_metrics = custom_metrics - self._validate_custom_metrics() + self._validate_custom_metrics() + + self._register_metrics() - self._register_metrics() - def _validate_predefined_metrics(self): if len(self.metric_names) == 0: raise ValueError("At least one metric must be specified.") @@ -105,7 +109,7 @@ def _validate_custom_metrics(self): for metric_name, stat_func in self.custom_metrics: if not callable(stat_func): raise ValueError(f"Custom Metric {metric_name} is not a function.") - + def _register_metrics(self): self.statistics_functions = {} @@ -113,10 +117,11 @@ def _register_metrics(self): if metric_name == "num_edges": func = lambda G: len(G.edges()) elif metric_name == "num_triangles": - func = lambda G: sum(nx.triangles(G).values()) // 3 # nx.triangles counts each triangle 3 times, for every node in it + func = lambda G: sum( + nx.triangles(G).values()) // 3 # nx.triangles counts each triangle 3 times, for every node in it self.statistics_functions[metric_name] = func - + def get_num_of_statistics(self): """ Get the number of statistics that are registered. @@ -127,7 +132,7 @@ def get_num_of_statistics(self): The number of statistics. """ return len(self.statistics_functions) - + def calculate_statistics(self, W: np.ndarray, verbose=False): """ Calculate the statistics of a graph. @@ -144,12 +149,37 @@ def calculate_statistics(self, W: np.ndarray, verbose=False): An array of statistics, represented as a numpy array of values, or a dict if verbose=True """ G = connectivity_matrix_to_G(W, directed=self._is_directed) - + stats = {name: func(G) for name, func in self.statistics_functions.items()} if verbose: return stats else: return np.array(list(stats.values())) - - \ No newline at end of file + +def construct_adj_mat_from_int(int_code: int, num_nodes: int, is_directed: bool) -> np.ndarray: + num_pos_connects = num_nodes * (num_nodes - 1) + if not is_directed: + num_pos_connects //= 2 + adj_mat_str = f'{int_code:0{num_pos_connects}b}' + cur_adj_mat = np.zeros((num_nodes, num_nodes)) + mat_entries_arr = np.array(list(adj_mat_str), 'uint8') + if is_directed: + cur_adj_mat[~np.eye(num_nodes, dtype=bool)] = mat_entries_arr + else: + upper_triangle_indices = np.triu_indices(num_nodes, k=1) + cur_adj_mat[upper_triangle_indices] = mat_entries_arr + lower_triangle_indices_aligned = (upper_triangle_indices[1], upper_triangle_indices[0]) + cur_adj_mat[lower_triangle_indices_aligned] = mat_entries_arr + return cur_adj_mat + + +def construct_int_from_adj_mat(adj_mat: np.ndarray, is_directed: bool) -> int: + if len(adj_mat.shape) != 2 or adj_mat.shape[0] != adj_mat.shape[1]: + raise ValueError(f"The dimensions of the given adjacency matrix {adj_mat.shape} are not valid for an " + f"adjacency matrix (should be a 2D squared matrix)") + num_nodes = adj_mat.shape[0] + adj_mat_no_diag = adj_mat[~np.eye(num_nodes, dtype=bool)] + if not is_directed: + adj_mat_no_diag = adj_mat_no_diag[:adj_mat_no_diag.size // 2] + return round((adj_mat_no_diag * 2 ** np.arange(adj_mat_no_diag.size - 1, -1, -1).astype(np.ulonglong)).sum())

Brute force ERGM Implemented an ERGM that iterates the entire space of networks, and calculates stuff exactly, to be used as a benchmark. This is obviously tractable only for small networks.

2024-07-23T09:15:15

0.0

[]

ninoseki/bukowski

ninoseki__bukowski-6

ba4f10ac6c5816001fda686ab4576a1af4a62dda

diff --git a/pyproject.toml b/pyproject.toml index bb95b18..d80ab1c 100644 --- a/pyproject.toml +++ b/pyproject.toml @@ -50,6 +50,7 @@ select = [ "F", # pyflakes "I", # isort "N", # pep8-naming + "PT", # flake8-pytest-style "RET", # flake8-return "RUF", # Ruff-specific rules "SIM", # flake8-simplify diff --git a/src/bukowski/main.py b/src/bukowski/main.py index bd90900..a6119d3 100644 --- a/src/bukowski/main.py +++ b/src/bukowski/main.py @@ -1,9 +1,11 @@ +import re from collections.abc import Iterable from dataclasses import dataclass from functools import partial from typing import Any, cast import tomlkit +from packaging.utils import canonicalize_name from poetry.core.constraints.version import ( Version, VersionConstraint, @@ -124,13 +126,20 @@ def init() -> TOMLDocument: return pyproject +def normalize_name(name: str): + # NOTE: > names must start and end with a letter or digit and may only contain -, _, ., and alphanumeric characters. # noqa: E501 + # (ref, https://github.com/astral-sh/uv/blob/main/crates/uv-cli/src/lib.rs#L46-L53) + subbed = re.sub("[^a-zA-Z0-9._-]+", "-", name) + return canonicalize_name(subbed) + + @safe def set_project(pyproject: TOMLDocument, *, poetry: Poetry) -> TOMLDocument: package = poetry.package content = cast(dict[str, Any], pyproject["project"]) - content["name"] = package.name + content["name"] = normalize_name(package.name) content["version"] = package.version.text content["description"] = package.description

Project names should be hyphenated to avoid TOML parser error Currently Bukowski turns `name = "Bingo to Bongo"` into `name = "bingo to bongo"`, but it should return `name = "bingo-to-bongo"`.

2024-09-17T09:35:02

0.0

[]

pydata/pydata-sphinx-theme

pydata__pydata-sphinx-theme-2007

5135b8f2b26e470a8aa49bdcda284b3968d3ffda

diff --git a/src/pydata_sphinx_theme/assets/styles/content/_code.scss b/src/pydata_sphinx_theme/assets/styles/content/_code.scss index 6371097ea..d7dd169b2 100644 --- a/src/pydata_sphinx_theme/assets/styles/content/_code.scss +++ b/src/pydata_sphinx_theme/assets/styles/content/_code.scss @@ -59,6 +59,10 @@ code.literal { a > code { color: var(--pst-color-inline-code-links); + + &:hover { + color: var(--pst-color-link-hover); + } } // Minimum opacity needed for linenos to be WCAG AA conformant

hovering on monospaced links changes underline color but not text color Not sure if that is intentional or not? You can see this by hovering a function name on https://scikit-learn.org/stable/api/index.html for example.

2024-10-07T21:10:16

0.0

[]

nsat/pypredict

nsat__pypredict-63

2084f535b9a04caa315391e580e5a725271b29d9

diff --git a/predict.c b/predict.c index 703deb7..26a3e2e 100644 --- a/predict.c +++ b/predict.c @@ -119,14 +119,6 @@ static PyObject *PredictException; static PyObject *NoTransitException; -/* - TODO: This is a refactoring hack to ensure we get consistent before/after - tests. Freezes the clock to a specified time, making calculations - deterministic. (remove when finished along with (2) call sites) -*/ -char debug_freeze_time = 0; -struct tm debug_frozen_tm = { tm_year: 114, tm_mon: 10, tm_mday: 2 }; - // This struct represents an observation of a particular satellite // from a particular reference point (on earth) at a particular time // and is used primarily in the PyPredict code. @@ -175,8 +167,8 @@ struct { long catnum; // Catalog Number (a.k.a NORAD id) long setnum; // Element Set No. char designator[10]; // Designator - int year; // Reference Epoch (year?) (?) - double refepoch; // Reference Epoch (?) + int year; // Reference Epoch from TLE (2-digit year) + double refepoch; // Reference Epoch from TLE (day of the year and fractional portion of the day) double incl; // Inclination double raan; // RAAN double eccn; // Eccentricity @@ -2083,13 +2075,6 @@ void Calculate_RADec(double time, vector_t *pos, vector_t *vel, geodetic_t *geod time_t CurrentTime() { - //TODO: Refactoring Hack (remove) - if (debug_freeze_time) - { - time_t debug_frozen_time = mktime(&debug_frozen_tm); - return debug_frozen_time; - } - return time(NULL); } @@ -2610,14 +2595,6 @@ double CurrentDaynum() { /* Read the system clock and return the number of days since 31Dec79 00:00:00 UTC (daynum 0) */ - - //TODO: Refactoring Hack (remove) - if (debug_freeze_time) - { - time_t debug_frozen_time = mktime(&debug_frozen_tm); - return ((((double)debug_frozen_time)/86400.0) - 3651.0); - } - int x; struct timeval tptr; double usecs, seconds; @@ -2630,44 +2607,6 @@ double CurrentDaynum() return ((seconds/86400.0)-3651.0); } -char *Daynum2String(daynum) -double daynum; -{ - /* This function takes the given epoch as a fractional number of - days since 31Dec79 00:00:00 UTC and returns the corresponding - date as a string of the form "Tue 12Oct99 17:22:37". */ - - char timestr[26]; - time_t t; - int x; - - /* Convert daynum to Unix time (seconds since 01-Jan-70) */ - t=(time_t)(86400.0*(daynum+3651.0)); - - sprintf(timestr,"%s",asctime(gmtime(&t))); - - if (timestr[8]==' ') - { - timestr[8]='0'; - } - - for (x=0; x<=3; output[x]=timestr[x], x++); - - output[4]=timestr[8]; - output[5]=timestr[9]; - output[6]=timestr[4]; - output[7]=timestr[5]; - output[8]=timestr[6]; - output[9]=timestr[22]; - output[10]=timestr[23]; - output[11]=' '; - - for (x=12; x<=19; output[x]=timestr[x-1], x++); - - output[20]=0; - return output; -} - void FindMoon(double daynum) { /* This function determines the position of the moon, including @@ -3543,7 +3482,7 @@ static char quick_find_docs[] = static PyObject* quick_predict(PyObject* self, PyObject *args) { - double now; + double tle_epoch; int lastel=0; char errbuff[100]; observation obs = { 0 }; @@ -3554,18 +3493,16 @@ static PyObject* quick_predict(PyObject* self, PyObject *args) goto cleanup_and_raise_exception; } - now=CurrentDaynum(); - if (load(args) != 0) { // load will set the appropriate exception string if it fails. goto cleanup_and_raise_exception; } - //TODO: Seems like this should be based on the freshness of the TLE, not wall clock. - if ((daynum<now-365.0) || (daynum>now+365.0)) + tle_epoch=(double)DayNum(1,0,sat.year)+sat.refepoch; + if ((daynum<tle_epoch-365.0) || (daynum>tle_epoch+365.0)) { - sprintf(errbuff, "time %s too far from present\n", Daynum2String(daynum)); + sprintf(errbuff, "day number %.0f too far from tle epoch day %.0f\n", daynum, tle_epoch); PyErr_SetString(PredictException, errbuff); goto cleanup_and_raise_exception; } diff --git a/setup.py b/setup.py index d0b07cc..10aa204 100755 --- a/setup.py +++ b/setup.py @@ -7,7 +7,7 @@ setup( name="pypredict", - version="1.7.1", + version="1.7.2", author="Jesse Trutna", author_email="[email protected]", maintainer="Spire Global Inc",

Base staleness check on TLE Epoch instead of wall clock Fixes #45 The staleness check within `predict.c` was based on being within a year of the current time. We change it so that it has to be within a year of the Epoch date on the TLE. PredictException: time XXX too far from present should use TLE Epoch The [code that determines staleness](https://github.com/nsat/pypredict/blob/15f32d057d2c6a325c27c477a67675858136f183/predict.c#L3565-L3571) works off of current time instead of TLE Epoch. That makes it hard to write unit tests, because they will fail a year later. As indicated by the comment on the code, this should measure against the TLE Epoch instead.

Hi, thanks for your pull request! This looks like it will be a helpful fix. Before I merge your PR I'd like to make sure it's tested. Would you mind adding some test cases demonstrating the desired behavior and edge cases?

2024-04-23T19:21:01

0.0

[]

BigDataBiology/SemiBin

BigDataBiology__SemiBin-78

c065098d63a46fc684bf9db8b18717150aca7ea1

diff --git a/ChangeLog b/ChangeLog index 82b4f3e..69f5383 100644 --- a/ChangeLog +++ b/ChangeLog @@ -1,6 +1,7 @@ Version 0.7.0 * Reuse markers.hmmout to make the training from several samples faster * Add option `--tmpdir` to set temporary directory + * Substitute FragGeneScan with Prodigal (FragGeneScan can still be used with `--predictor` parameter) Version 0.6.0 Mon 7 Feb 2022 by BigDataBiology * Provide pretrained models from soil, cat gut, human oral, diff --git a/README.md b/README.md index bd22968..066619e 100644 --- a/README.md +++ b/README.md @@ -47,12 +47,13 @@ You can download the source code from github and install. Install dependence packages using conda: [MMseqs2](https://github.com/soedinglab/MMseqs2), -[Bedtools](http://bedtools.readthedocs.org/]), [Hmmer](http://hmmer.org/), and -[Fraggenescan](https://sourceforge.net/projects/fraggenescan/). +[Bedtools](http://bedtools.readthedocs.org/]), [Hmmer](http://hmmer.org/), +[Fraggenescan](https://sourceforge.net/projects/fraggenescan/) and [Prodigal](https://github.com/hyattpd/Prodigal) ```bash conda install -c conda-forge -c bioconda mmseqs2=13.45111 conda install -c bioconda bedtools hmmer fraggenescan==1.30 +conda install -c bioconda prodigal ``` Once the dependencies are installed, you can install by running: diff --git a/SemiBin/cluster.py b/SemiBin/cluster.py index 429d0d2..8c24701 100644 --- a/SemiBin/cluster.py +++ b/SemiBin/cluster.py @@ -83,7 +83,7 @@ def cal_kl(m, v, use_ne='auto'): def cluster(model, data, device, max_edges, max_node, is_combined, - logger, n_sample, out, contig_dict, binned_length, num_process, minfasta, recluster, random_seed): + logger, n_sample, out, contig_dict, binned_length, num_process, minfasta, recluster, random_seed, predictor = 'prodigal'): """ Cluster contigs into bins max_edges: max edges of one contig considered in binning @@ -219,7 +219,8 @@ def cluster(model, data, device, max_edges, max_node, is_combined, cfasta, binned_length, num_process, - multi_mode=True) + multi_mode=True, + predictor=predictor) for ix,bin_path in enumerate(bin_files): # if there are no hits, the output will be naturally empty diff --git a/SemiBin/main.py b/SemiBin/main.py index 10ef704..cad0a92 100644 --- a/SemiBin/main.py +++ b/SemiBin/main.py @@ -166,6 +166,14 @@ def parse_args(args): default=None, help='Path to the trained semi-supervised deep learning model.') + for p in [single_easy_bin, multi_easy_bin, training, binning]: + p.add_argument('--predictor', + required=False, + type=str, + help='gene predictor used to estimate the number of bins(prodigal/fraggenescan)', + dest='predictor', + default='prodigal') + for p in [single_easy_bin, multi_easy_bin, generate_cannot_links, training, binning]: p.add_argument('--tmpdir', required=False, @@ -611,7 +619,7 @@ def fasta_sample_iter(fn): def training(logger, contig_fasta, num_process, data, data_split, cannot_link, batchsize, - epoches, output, device, ratio, min_length, mode): + epoches, output, device, ratio, min_length, mode, predictor = 'prodigal'): """ Training the model @@ -666,13 +674,14 @@ def training(logger, contig_fasta, num_process, epoches, device, num_process, - mode = mode) + mode = mode, + predictor=predictor) def binning(logger, num_process, data, max_edges, max_node, minfasta, binned_length, contig_dict, recluster,model_path, - random_seed,output, device, environment): + random_seed,output, device, environment, predictor = 'prodigal'): """ Clustering the contigs to get the final bins. @@ -711,12 +720,13 @@ def binning(logger, num_process, data, num_process, minfasta, recluster, - random_seed) + random_seed, + predictor=predictor) def single_easy_binning(args, logger, binned_length, must_link_threshold, - contig_dict, recluster,random_seed, output, device, environment): + contig_dict, recluster,random_seed, output, device, environment, predictor = 'prodigal'): """ contain `generate_cannot_links`, `generate_sequence_features_single`, `train`, `bin` in one command for single-sample and co-assembly binning """ @@ -748,17 +758,17 @@ def single_easy_binning(args, logger, binned_length, training(logger, [args.contig_fasta], args.num_process, [data_path], [data_split_path], [os.path.join(output, 'cannot', 'cannot.txt')], - args.batchsize, args.epoches, output, device, args.ratio, args.min_len, mode='single') + args.batchsize, args.epoches, output, device, args.ratio, args.min_len, mode='single', predictor=predictor) binning(logger, args.num_process, data_path, args.max_edges, args.max_node, args.minfasta_kb * 1000, binned_length, contig_dict, recluster, os.path.join(output, 'model.h5') if environment is None else None, - random_seed, output, device, environment if environment is not None else None) + random_seed, output, device, environment if environment is not None else None, predictor=predictor) def multi_easy_binning(args, logger, recluster, - random_seed, output, device): + random_seed, output, device, predictor = 'prodigal'): """ contain `generate_cannot_links`, `generate_sequence_features_multi`, `train`, `bin` in one command for multi-sample binning """ @@ -774,7 +784,7 @@ def multi_easy_binning(args, logger, recluster, args.ratio, args.min_len, args.ml_threshold, - output, ) + output,) for sample in sample_list: logger.info( @@ -810,13 +820,13 @@ def multi_easy_binning(args, logger, recluster, training(logger, [sample_fasta], args.num_process, [sample_data], [sample_data_split], [sample_cannot], args.batchsize, args.epoches, os.path.join(output, 'samples', sample), - device, args.ratio, args.min_len, mode='single') + device, args.ratio, args.min_len, mode='single', predictor=predictor) binning(logger, args.num_process, sample_data, args.max_edges, args.max_node, args.minfasta_kb * 1000, binned_length, contig_dict, recluster, os.path.join(output, 'samples', sample, 'model.h5'), random_seed , - os.path.join(output, 'samples', sample), device, None) + os.path.join(output, 'samples', sample), device, None, predictor=predictor) os.makedirs(os.path.join(output, 'bins'), exist_ok=True) for sample in sample_list: @@ -910,7 +920,7 @@ def main(): set_random_seed(args.random_seed) training(logger, args.contig_fasta, args.num_process, args.data, args.data_split, args.cannot_link, - args.batchsize, args.epoches, out, device, args.ratio, args.min_len, args.mode) + args.batchsize, args.epoches, out, device, args.ratio, args.min_len, args.mode, predictor=args.predictor) if args.cmd == 'bin': @@ -918,7 +928,7 @@ def main(): set_random_seed(args.random_seed) binning(logger, args.num_process, args.data, args.max_edges, args.max_node, args.minfasta_kb * 1000, binned_length, - contig_dict, args.recluster, args.model_path, args.random_seed,out, device, args.environment) + contig_dict, args.recluster, args.model_path, args.random_seed,out, device, args.environment, predictor=args.predictor) if args.cmd == 'single_easy_bin': @@ -935,7 +945,8 @@ def main(): args.random_seed, out, device, - args.environment) + args.environment, + predictor=args.predictor) if args.cmd == 'multi_easy_bin': check_install() @@ -947,7 +958,8 @@ def main(): args.recluster, args.random_seed, out, - device) + device, + predictor=args.predictor) if __name__ == '__main__': diff --git a/SemiBin/semi_supervised_model.py b/SemiBin/semi_supervised_model.py index ab26b2c..e408080 100644 --- a/SemiBin/semi_supervised_model.py +++ b/SemiBin/semi_supervised_model.py @@ -139,7 +139,7 @@ def __len__(self): def train(out, contig_fastas, binned_lengths, logger, datas, data_splits, cannot_links, is_combined=True, - batchsize=2048, epoches=20, device=None, num_process = 8, mode = 'single'): + batchsize=2048, epoches=20, device=None, num_process = 8, mode = 'single', predictor = 'prodigal'): """ Train model from one sample(--mode single) or several samples(--mode several) """ @@ -171,7 +171,8 @@ def train(out, contig_fastas, binned_lengths, logger, datas, data_splits, cannot contig_fastas[data_index], binned_length=binned_lengths[data_index], num_process=num_process, - output=out) + output=out, + predictor=predictor) if epoch == 0: logger.info('Generate training data of {}:'.format(data_index)) diff --git a/SemiBin/utils.py b/SemiBin/utils.py index 59e8dde..7682731 100644 --- a/SemiBin/utils.py +++ b/SemiBin/utils.py @@ -28,6 +28,15 @@ def expect_file_list(fs): args.num_process = multiprocessing.cpu_count() logger.info(f'Setting number of CPUs to {args.num_process}') + if args.num_process > multiprocessing.cpu_count(): + args.num_process = multiprocessing.cpu_count() + + if args.cmd in ['single_easy_bin', 'multi_easy_bin', 'train', 'bin']: + if args.predictor not in ['prodigal', 'fraggenescan']: + sys.stderr.write( + f"Error: SemiBin support prodigal or fraggenescan gene predictor.\n") + sys.exit(1) + if args.cmd == 'generate_cannot_links': if args.GTDB_reference is not None: expect_file(args.GTDB_reference) @@ -199,7 +208,7 @@ def generate_cannot_link(mmseqs_path, namelist, num_threshold, output,sample): 'TIGR02386': 'TIGR02387', } -def get_marker(hmmout, fasta_path=None, min_contig_len=None, multi_mode=False): +def get_marker(hmmout, fasta_path=None, min_contig_len=None, multi_mode=False, predictor = None): import pandas as pd data = pd.read_table(hmmout, sep=r'\s+', comment='#', header=None, usecols=(0,3,5,15,16), names=['orf', 'gene', 'qlen', 'qstart', 'qend']) @@ -209,7 +218,10 @@ def get_marker(hmmout, fasta_path=None, min_contig_len=None, multi_mode=False): qlen = data[['gene','qlen']].drop_duplicates().set_index('gene')['qlen'] def contig_name(ell): - contig,_,_,_ = ell.rsplit( '_', 3) + if predictor == 'prodigal': + contig,_ = ell.rsplit( '_', 1) + else: + contig,_,_,_ = ell.rsplit( '_', 3) return contig data = data.query('(qend - qstart) / qlen > 0.4').copy() @@ -244,24 +256,69 @@ def extract_seeds(vs, sel): counts = data.groupby('gene')['orf'].count() return extract_seeds(counts, data) -def cal_num_bins(fasta_path, binned_length, num_process, multi_mode=False, output = None): - '''Estimate number of bins from a FASTA file +def prodigal(contig_file, contig_output): + with open(contig_output + '.out', 'w') as prodigal_out_log: + subprocess.check_call( + ['prodigal', + '-i', contig_file, + '-p', 'meta', + '-q', + '-a', contig_output + ], + stdout=prodigal_out_log, + ) - Parameters - fasta_path: path - binned_length: int (minimal contig length) - num_process: int (number of CPUs to use) - multi_mode: bool, optional (if True, treat input as resulting from concatenating multiple files) - ''' - with tempfile.TemporaryDirectory() as tdir: - if output is not None: - if os.path.exists(os.path.join(output, 'markers.hmmout')): - return get_marker(os.path.join(output, 'markers.hmmout'), fasta_path, binned_length, multi_mode) - else: - target_dir = output - else: - target_dir = tdir - contig_output = os.path.join(tdir, 'contigs.faa') +def run_prodigal(fasta_path, num_process, output): + from .error import LoggingPool + + contig_dict = {} + for h, seq in fasta_iter(fasta_path): + contig_dict[h] = seq + + num_contig = len(contig_dict) + + num_split = num_contig // num_process + split_contig_dict = {} + split_index = 0 + for h in contig_dict: + if split_index not in split_contig_dict: + split_contig_dict[split_index] = [] + split_contig_dict[split_index].append(h) + if split_index < num_process - 1: + if len(split_contig_dict[split_index]) == num_split: + split_index += 1 + + ### split_fasta + for index in range(num_process): + with open(os.path.join(output, 'contig_{}.fa'.format(index)), 'wt') as concat_out: + for h in split_contig_dict[index]: + concat_out.write(f'>{h}\n{contig_dict[h]}\n') + + with LoggingPool(num_process) if num_process != 0 else LoggingPool() as pool: + try: + for index in range(num_process): + pool.apply_async( + prodigal, + args=( + os.path.join(output, 'contig_{}.fa'.format(index)), + os.path.join(output, 'contig_{}.faa'.format(index)), + )) + pool.close() + pool.join() + except: + sys.stderr.write( + f"Error: Running prodigal fail\n") + sys.exit(1) + + contig_output = os.path.join(output, 'contigs.faa') + with open(contig_output, 'w') as f: + for index in range(num_process): + f.write(open(os.path.join(output, 'contig_{}.faa'.format(index)), 'r').read()) + return contig_output + +def run_fraggengscan(fasta_path, num_process, output): + try: + contig_output = os.path.join(output, 'contigs.faa') with open(contig_output + '.out', 'w') as frag_out_log: # We need to call FragGeneScan instead of the Perl wrapper because the # Perl wrapper does not handle filepaths correctly if they contain spaces @@ -277,22 +334,57 @@ def cal_num_bins(fasta_path, binned_length, num_process, multi_mode=False, outpu ], stdout=frag_out_log, ) + except: + sys.stderr.write( + f"Error: Running fraggenescan fail\n") + sys.exit(1) + return contig_output + +def cal_num_bins(fasta_path, binned_length, num_process, multi_mode=False, output = None, predictor = 'prodigal'): + '''Estimate number of bins from a FASTA file + + Parameters + fasta_path: path + binned_length: int (minimal contig length) + num_process: int (number of CPUs to use) + multi_mode: bool, optional (if True, treat input as resulting from concatenating multiple files) + ''' + with tempfile.TemporaryDirectory() as tdir: + if output is not None: + if os.path.exists(os.path.join(output, 'markers.hmmout')): + return get_marker(os.path.join(output, 'markers.hmmout'), fasta_path, binned_length, multi_mode, predictor=predictor) + else: + target_dir = output + else: + target_dir = tdir + + if predictor == 'prodigal': + contig_output = run_prodigal(fasta_path, num_process, tdir) + else: + contig_output = run_fraggengscan(fasta_path, num_process, tdir) hmm_output = os.path.join(target_dir, 'markers.hmmout') - with open(os.path.join(tdir, 'markers.hmmout.out'), 'w') as hmm_out_log: - subprocess.check_call( - ['hmmsearch', - '--domtblout', - hmm_output, - '--cut_tc', - '--cpu', str(num_process), - os.path.split(__file__)[0] + '/marker.hmm', - contig_output + '.faa', - ], - stdout=hmm_out_log, - ) + try: + with open(os.path.join(tdir, 'markers.hmmout.out'), 'w') as hmm_out_log: + subprocess.check_call( + ['hmmsearch', + '--domtblout', + hmm_output, + '--cut_tc', + '--cpu', str(num_process), + os.path.split(__file__)[0] + '/marker.hmm', + contig_output if predictor == 'prodigal' else contig_output + '.faa', + ], + stdout=hmm_out_log, + ) + except: + sys.stderr.write( + f"Error: Running hmmsearch fail\n") + sys.exit(1) + + marker = get_marker(hmm_output, fasta_path, binned_length, multi_mode, predictor=predictor) - return get_marker(hmm_output, fasta_path, binned_length, multi_mode) + return marker def write_bins(namelist, contig_labels, output, contig_dict, diff --git a/docs/index.md b/docs/index.md index 2a93721..f3e3e10 100644 --- a/docs/index.md +++ b/docs/index.md @@ -57,6 +57,7 @@ Reconstruct bins with single or co-assembly binning using one line command. * `--ml-threshold` : Length threshold for generating must-link constraints.(By default, the threshold is calculated from the contig, and the default minimum value is 4,000 bp) * `--no-recluster` : Do not recluster bins. * `--taxonomy-annotation-table` : TAXONOMY_TSV, Pre-computed mmseqs2 format taxonomy TSV file to bypass mmseqs2 GTDB annotation [advanced] +* `--predictor` : gene predictor used to estimate the number of bins(prodigal/fraggenescan) #### multi_easy_bin @@ -73,7 +74,7 @@ The following options (including synonyms) are the same as for Run the contig annotations using mmseqs with GTDB and generate cannot-link file used in the semi-supervised deep learning model training. -The following options are the same as for `single_easy_bin`: `-i/--input-fasta`, `-o/--output`, `--cannot-name`, `-r/--reference-db-data-dir`, `--ratio`, `--min-len`, `--ml-threshold` and `--taxonomy-annotation-table`. +The following options are the same as for `single_easy_bin`: `-i/--input-fasta`, `-o/--output`, `--cannot-name`, `-r/--reference-db-data-dir`, `--ratio`, `--min-len`, `--ml-threshold`, `--taxonomy-annotation-table` and `--predictor`. #### generate_sequence_features_single @@ -98,7 +99,7 @@ Training the model. * `-c/--cannot-link` : Path to the input cannot link file generated from other additional biological information, one row for each cannot link constraint. The file format: contig_1,contig_2. * `--mode`: [single/several] Train models from one sample or several samples(train model across several samples can get better pre-trained model for single-sample binning.) In several mode, must input data, data_split, cannot, fasta files for corresponding sample with same order. *Note:* You can just set `several` with this option when single-sample binning. Training from several samples with multi-sample binning is not support. -The following options are the same as for `single_easy_bin`: `-i/--input-fasta`, `-o/--output`, `--epoches`, `--batch-size`, `-p/--processes/-t/--threads`, `--random-seed`, `--ratio`, `--min-len`. +The following options are the same as for `single_easy_bin`: `-i/--input-fasta`, `-o/--output`, `--epoches`, `--batch-size`, `-p/--processes/-t/--threads`, `--random-seed`, `--ratio`, `--min-len` and `--predictor`. #### bin @@ -106,7 +107,7 @@ Cluster contigs into bins. * `--model`: Path to the trained model. -The following options are the same as for `single_easy_bin`: `--data`,`-i/--input-fasta`, `-o/--output`, `--minfasta-kbs`, `--recluster`, `--max-node`, `--max-edges`, `-p/--processes/-t/--threads`, `--random-seed`, `--environment`, `--ratio`, `--min-len` and `--no-recluster`. +The following options are the same as for `single_easy_bin`: `--data`,`-i/--input-fasta`, `-o/--output`, `--minfasta-kbs`, `--recluster`, `--max-node`, `--max-edges`, `-p/--processes/-t/--threads`, `--random-seed`, `--environment`, `--ratio`, `--min-len`, `--no-recluster` and `--predictor`. #### download_GTDB

Substituting the ORF finder Substitute the ORF finder (FragGeneScan) with a better one (Prodigal).

Background: https://github.com/BigDataBiology/SemiBin/issues/73

2022-02-21T05:18:01

0.0

[]

emdgroup/tnmf

emdgroup__tnmf-51

0d1bbcc0a4e7763779f1d412cd7e0721806c9823

diff --git a/CHANGELOG.md b/CHANGELOG.md index 43b664c3..8d1ce64e 100644 --- a/CHANGELOG.md +++ b/CHANGELOG.md @@ -9,6 +9,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0 - Unit test Code Coverage Reporting - Demo: Racoon image - Include examples in pytest bench +- Include demos in pytest bench ### Changed - Fixed internal TODOs to improve code quality diff --git a/demos/demo_image.py b/demos/demo_image.py index cd79dc72..359b62e2 100644 --- a/demos/demo_image.py +++ b/demos/demo_image.py @@ -24,7 +24,7 @@ def st_define_sample_params(verbose: bool = True) -> Tuple[dict, float]: help_scale = \ '''Downscaling the image allows for quicker interactive experimentation. Note that this does not change atom size.''' - scale = st.sidebar.number_input('Image Scale', min_value=0.05, value=0.25, step=0.05, help=help_scale) + scale = st.sidebar.number_input('Image scale', min_value=0.05, value=0.25, step=0.05, help=help_scale) if verbose: st.sidebar.caption(help_scale) diff --git a/demos/demo_selector.py b/demos/demo_selector.py index c552f010..d1aeface 100644 --- a/demos/demo_selector.py +++ b/demos/demo_selector.py @@ -1,13 +1,11 @@ +import sys + import importlib.resources from importlib import import_module import numpy as np import streamlit as st -# show TNMF header -with importlib.resources.path('logos', 'tnmf_header.png') as img_file: - st.image(str(img_file), use_column_width='always') - # mapping of streamlit demo names to the corresponding demo file and optional parameters DEMO_NAME_DICT = { '1-D Synthetic Signals': ('synthetic_signals', {'n_dims': 1}), @@ -15,43 +13,57 @@ 'Racoon Image': ('demo_image', {}), } -# create progress bar on the top that is shown for all demos -progress_bar = st.sidebar.progress(1.) - -# toggle verbose mode -help_verbose = 'Displays / hides **detailed explanations**.' -verbose = st.sidebar.checkbox('Verbose', True, help=help_verbose) -if verbose: - st.sidebar.caption(help_verbose) - -# select the demo -help_select_demo = 'The specific **demo example** that gets executed.' -selected_demo = st.sidebar.selectbox('Demo example', list(DEMO_NAME_DICT.keys()), 1, help=help_select_demo) -if verbose: - st.sidebar.caption(help_select_demo) - -# select the random seed -help_seed = 'The fixed **random seed** that is used for the simulation.' -seed = st.sidebar.number_input('Random seed', value=42, help=help_seed) -np.random.seed(seed) -if verbose: - st.sidebar.caption(help_seed) - -# show demo info text -if verbose: - st.markdown(''' - ## Demo guide - This dashboard demonstrates the use of the *Transform-Invariant Non-Negative Matrix Factorization (TNMF) package* - in the specific context of learning **shift-invariant representations**. - Via the **drop-down menu** on the left, you can choose between different demo examples. - - ## Usage - Detailed explanations of the available options and the shown output (including this text) can be displayed or hidden - using the **'Verbose' checkbox**. When the checkbox is unticked, the description of the control widgets can still - be accessed via the **tooltips** next to them. - ''') - -# extract the demo name and parameters, import the corresponding module, and execute the demo -demo_name, demo_args = DEMO_NAME_DICT[selected_demo] -demo_module = import_module(demo_name) -demo_module.main(progress_bar, verbose=verbose, **demo_args) + +def main(demo_name: str): + # show TNMF header + with importlib.resources.path('logos', 'tnmf_header.png') as img_file: + st.image(str(img_file), use_column_width='always') + + # create progress bar on the top that is shown for all demos + progress_bar = st.sidebar.progress(1.) + + # toggle verbose mode + help_verbose = 'Displays / hides **detailed explanations**.' + verbose = st.sidebar.checkbox('Verbose', True, help=help_verbose) + if verbose: + st.sidebar.caption(help_verbose) + + # select the demo + help_select_demo = 'The specific **demo example** that gets executed.' + default_demo = list(DEMO_NAME_DICT.keys()).index(demo_name) + selected_demo = st.sidebar.selectbox('Demo example', list(DEMO_NAME_DICT.keys()), default_demo, help=help_select_demo) + if verbose: + st.sidebar.caption(help_select_demo) + + # select the random seed + help_seed = 'The fixed **random seed** that is used for the simulation.' + seed = st.sidebar.number_input('Random seed', value=42, help=help_seed) + np.random.seed(seed) + if verbose: + st.sidebar.caption(help_seed) + + # show demo info text + if verbose: + st.markdown(''' + ## Demo guide + This dashboard demonstrates the use of the *Transform-Invariant Non-Negative Matrix Factorization (TNMF) package* + in the specific context of learning **shift-invariant representations**. + Via the **drop-down menu** on the left, you can choose between different demo examples. + + ## Usage + Detailed explanations of the available options and the shown output (including this text) can be displayed or hidden + using the **'Verbose' checkbox**. When the checkbox is unticked, the description of the control widgets can still + be accessed via the **tooltips** next to them. + ''') + + # extract the demo name and parameters, import the corresponding module, and execute the demo + demo_name, demo_args = DEMO_NAME_DICT[selected_demo] + demo_module = import_module(demo_name) + demo_module.main(progress_bar, verbose=verbose, **demo_args) + + +if __name__ == '__main__': + if len(sys.argv) > 1: + main(sys.argv[1]) + else: + main('2-D Synthetic Signals') diff --git a/scripts/tnmf.py b/scripts/tnmf.py index 4ddbea64..d0fdfc3c 100644 --- a/scripts/tnmf.py +++ b/scripts/tnmf.py @@ -3,6 +3,7 @@ from glob import glob from pathlib import Path from typing import List +from demos.demo_selector import DEMO_NAME_DICT DEMO_FILE = Path(__file__).absolute().parent.parent / 'demos' / 'demo_selector.py' EXAMPLES_FOLDER = Path(__file__).absolute().parent.parent / 'examples' @@ -19,6 +20,11 @@ def get_examples() -> List[Path]: return examples +def get_demos() -> List[str]: + """Returns a list containing the names of all demos.""" + return list(DEMO_NAME_DICT.keys()) + + def main(): """Entry point for the `tnmf` command.""" @@ -27,10 +33,12 @@ def main(): # create subparsers for the `example` and `demo` commands subparsers = parser.add_subparsers(dest='command') - demo_parser = subparsers.add_parser("demo", help='Runs the streamlit demo.') # noqa: F841 + demo_parser = subparsers.add_parser("demo", help='Runs the streamlit demo. Type `tnmf demo -h` to show the list ' + 'of available demos.') + demo_parser.add_argument("demo_name", choices=get_demos(), help='Name of the demo.', default=None, nargs='?') + example_parser = subparsers.add_parser("example", help='Runs a specific example. Type `tnmf example -h` to show the list ' 'of available examples.') - examples = get_examples() example_parser.add_argument("example_name", choices=[e.stem for e in examples], help='Name of the example.') @@ -41,8 +49,12 @@ def main(): if args.command == 'example': args_example = parser.parse_args() example = [e for e in examples if e.stem == args_example.example_name][0] - os.system(f'python {example}') # noqa: S605 + os.system(f'python {example}') # noqa: S605, DUO106 - # run the demo + # parse the (optional) demo name and run the demo elif args.command == 'demo': - os.system(f'streamlit run {DEMO_FILE}') # noqa: S605 + args_demo = parser.parse_args() + if args_demo.demo_name is not None: + os.system(f'streamlit run {DEMO_FILE} "{args_demo.demo_name}"') # noqa: S605, DUO106 + else: + os.system(f'streamlit run {DEMO_FILE}') # noqa: S605 diff --git a/setup.cfg b/setup.cfg index fde5a601..b0445eaa 100644 --- a/setup.cfg +++ b/setup.cfg @@ -105,7 +105,7 @@ generated-members=numpy.*, torch.* source = tnmf examples - # TODO: add 'demos' here + demos branch = True command_line = -m pytest # create individual files per process that need to be combined afterwards @@ -115,7 +115,7 @@ parallel = True [coverage:report] # Target code coverage in integer percent. If overall coverage is below, coverage.py reports failure. # TODO: should gradually increase this -fail_under = 67 +fail_under = 90 # Do not report files with coverage at 100% skip_covered = True # Do not report empty files diff --git a/tnmf/utils/demo.py b/tnmf/utils/demo.py index 6e083ef0..bf75e096 100644 --- a/tnmf/utils/demo.py +++ b/tnmf/utils/demo.py @@ -58,8 +58,7 @@ def st_define_nmf_params(default_params: dict, have_ground_truth: bool = True, v n_atoms_default = default_params['n_atoms'] n_atoms = st.sidebar.number_input('# Atoms', value=n_atoms_default, min_value=1, help=help_n_atoms) if not use_n_atoms else n_atoms_default - if verbose and not use_n_atoms: - st.sidebar.caption(help_n_atoms) + explanation(help_n_atoms, verbose and not use_n_atoms) help_atom_shape = 'The **size of each atom** dimension.' if have_ground_truth: @@ -78,14 +77,12 @@ def st_define_nmf_params(default_params: dict, have_ground_truth: bool = True, v value=default_atom_shape[0], min_value=1, help=help_atom_shape)] * len(default_params['atom_shape']) ) if not use_atom_shape else default_atom_shape - if not use_atom_shape: - explanation(help_atom_shape, verbose) + explanation(help_atom_shape, verbose and not use_atom_shape) help_sparsity_H = 'The strength of the **L1 activation sparsity regularization** imposed on the optimization problem.' sparsity_H = st.sidebar.number_input('Activation sparsity', min_value=0.0, value=0.0, step=0.01, help=help_sparsity_H) - if verbose: - st.sidebar.caption(help_sparsity_H) + explanation(help_sparsity_H, verbose) help_inhibition_strength = \ '''The strength of the **same-atom lateral activation sparsity regularization** imposed on the optimization problem. @@ -107,8 +104,7 @@ def st_define_nmf_params(default_params: dict, have_ground_truth: bool = True, v help_n_iterations = '''The **number of multiplicative updates** to the atom dictionary and activation tensors.''' n_iterations = st.sidebar.number_input('# Iterations', value=100, min_value=1, help=help_n_iterations) - if verbose: - st.sidebar.caption(help_n_iterations) + explanation(help_n_iterations, verbose) help_backend = \ '''The **optimization backend** for computing the multiplicative gradients.

Need to unit-test examples and demo Currently, we would not recognize if these fail, e.g. due to interface incompatibilities. Easiest might be to create dummy unit tests that set matplotlib into non-interactive mode and run the examples just to check that nothing fails horribly (no result verification).

2021-07-28T12:59:42

0.0

[]

emdgroup/tnmf

emdgroup__tnmf-49

0315335e225f2b84cc98404b04163bd3c1e48d20

diff --git a/CHANGELOG.md b/CHANGELOG.md index 9fd1267b..409ca3e9 100644 --- a/CHANGELOG.md +++ b/CHANGELOG.md @@ -8,6 +8,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0 ### Added - Unit test Code Coverage Reporting - Demo: Racoon image +- Include examples in pytest bench ## [0.1.1] - 2021-07-22 ### Added

Need to unit-test examples and demo Currently, we would not recognize if these fail, e.g. due to interface incompatibilities. Easiest might be to create dummy unit tests that set matplotlib into non-interactive mode and run the examples just to check that nothing fails horribly (no result verification).

2021-07-20T22:11:25

0.0

[]

emdgroup/tnmf

emdgroup__tnmf-43

be5db34d6271280d5645d4323ec019bf4456438a

diff --git a/.gitignore b/.gitignore index 25694cc9..3f25472e 100644 --- a/.gitignore +++ b/.gitignore @@ -20,6 +20,7 @@ _build *.egg-info build dist +htmlcov # Package Version tnmf/_version.py diff --git a/CHANGELOG.md b/CHANGELOG.md index 368ee224..43f79f6a 100644 --- a/CHANGELOG.md +++ b/CHANGELOG.md @@ -4,6 +4,10 @@ All notable changes to this project will be documented in this file. The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/), and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html). +## Unreleased +### Added +- Unit test Code Coverage Reporting + ## [0.1.1] - 2021-07-22 ### Added - Bugfix: added logo to package content diff --git a/README.md b/README.md index e68a1cdc..579e8951 100644 --- a/README.md +++ b/README.md @@ -1,6 +1,6 @@ [![Flake8 Linter](https://github.com/emdgroup/tnmf/actions/workflows/flake8.yml/badge.svg)](https://github.com/emdgroup/tnmf/actions/workflows/flake8.yml) [![Pylint Linter](https://github.com/emdgroup/tnmf/actions/workflows/pylint.yml/badge.svg)](https://github.com/emdgroup/tnmf/actions/workflows/pylint.yml) -[![Pytest](https://github.com/emdgroup/tnmf/actions/workflows/pytest.yml/badge.svg)](https://github.com/emdgroup/tnmf/actions/workflows/pytest.yml) +[![Pytest and Coverage](https://github.com/emdgroup/tnmf/actions/workflows/pytest.yml/badge.svg)](https://github.com/emdgroup/tnmf/actions/workflows/pytest.yml) [![Build Documentation](https://github.com/emdgroup/tnmf/actions/workflows/sphinx.yml/badge.svg)](https://github.com/emdgroup/tnmf/actions/workflows/sphinx.yml) [![Publish to PyPI](https://github.com/emdgroup/tnmf/actions/workflows/publish-to-pypi.yml/badge.svg)](https://github.com/emdgroup/tnmf/actions/workflows/publish-to-pypi.yml) [![Open in Streamlit](https://static.streamlit.io/badges/streamlit_badge_black_white.svg)](https://share.streamlit.io/adriansosic/tnmf/main/demos/demo_selector.py) @@ -76,16 +76,51 @@ Now, you should be able to execute the unit tests by calling `pytest` to verify ## Pull Requests Before creating a pull request, you should always try to ensure that the automated code quality and unit tests do not fail. -To execute them, change into the repository root directory and run the following commands: +This section explains how to run them locally to understand and fix potential issues. + +### Code Style and Quality +Code style and quality are checked using [flake8](https://flake8.pycqa.org/) and [pylint](http://pylint.pycqa.org/). +To execute them, change into the [repository root directory](.), run the following commands and inspect their output: ``` flake8 pylint tnmf +``` + +In order for a pull request to be accaptable, no errors may be reported here. + +### Unit Tests +Automated unit tests reside inside the folder [tnmf/tests](tnmf/tests). They can be executed via +[pytest](https://docs.pytest.org/) by changing into the [repository root directory](.) and running + +``` pytest ``` -The output of the individual commands should be pretty instructive, so fixing potential issues is usually rather straightforward. +Debugging potential failures from the command line might be cumbersome. +Most Python IDEs, however, also support `pytest` natively in their debugger. +Again, for a pull request to be acceptable, no failures may be reported here. + +### Code Coverage +Code coverage in the unit tests is measured using [coverage](https://coverage.readthedocs.io). +A coverage report can be created locally from the [repository root directory](.) via + +``` +coverage run +coverage report +``` + +This will output a concise table with an overview of python files that are not fully covered with unit tests along with the line numbers of code that has not been executed. +A more detailed, interactive report can be created using + +``` +coverage html +``` + +Then, you can open the file `htmlcov/index.html` in a web browser of your choice to navigate through code annotated with coverage data. +Required overall coverage to is configured in [setup.cfg](setup.cfg), under the key `fail_under` in section `[coverage:report]`. + ## Building the Documentation -To build the documentation locally, change into the `doc` subdirectory and run `make html`. +To build the documentation locally, change into the [doc subdirectory](doc) and run `make html`. Then, the documentation resides at `doc\_build\html\index.html`. diff --git a/requirements.txt b/requirements.txt index b6016bb6..f5e3e4ed 100644 --- a/requirements.txt +++ b/requirements.txt @@ -8,8 +8,8 @@ streamlit==0.84.0 torch==1.8.1 # Development and Tests +coverage==5.5 pytest==6.2.3 -pytest-cov==2.11.1 flake8==3.9.1 flake8-bandit==2.1.2 flake8-bugbear==21.4.3 diff --git a/setup.cfg b/setup.cfg index 2ad7f607..e133821d 100644 --- a/setup.cfg +++ b/setup.cfg @@ -46,11 +46,8 @@ test = pytest [tool:pytest] log_cli = True -# TODO: temporarily disable coverage because it breaks PyCharm's debugger -# https://www.jetbrains.com/help/pycharm/run-debug-configuration-py-test.html addopts = --ignore doc -# --cov=tnmf [flake8] exclude = .git,__pycache__,doc,env,venv,.venv @@ -103,3 +100,30 @@ bad-names-rgxs=_?num_.* # List of members which are set dynamically and missed by Pylint inference # system, and so shouldn't trigger E1101 when accessed. generated-members=numpy.*, torch.* + +[coverage:run] +source = tnmf +branch = True +command_line = -m pytest + +[coverage:report] +# Target code coverage in integer percent. If overall coverage is below, coverage.py reports failure. +# TODO: should gradually increase this +fail_under = 64 +# Do not report files with coverage at 100% +skip_covered = True +# Do not report empty files +skip_empty = True +# List line numbers in the report table +show_missing = True +# exclusion list (regular expressions) +exclude_lines = + pragma: no cover + raise NotImplementedError + raise AssertionError + raise ValueError + except ImportError: + except PackageNotFoundError: +# file patterns to ignore +omit = + tnmf/_version.py \ No newline at end of file diff --git a/tnmf/backends/NumPy_CachingFFT.py b/tnmf/backends/NumPy_CachingFFT.py index ba15552c..d0a1401a 100644 --- a/tnmf/backends/NumPy_CachingFFT.py +++ b/tnmf/backends/NumPy_CachingFFT.py @@ -58,36 +58,26 @@ def __itruediv__(self, other): self.c /= other.c if isinstance(other, CachingFFT) else other return self - def set_fft_params(self, fft_axes: Tuple[int, ...], fft_shape: Tuple[int, ...]): - self._fft_axes = fft_axes - self._fft_shape = fft_shape - def _invalidate(self): self._c = None self._f = None self._f_padded = None self._f_reversed = None + @property def has_c(self) -> bool: """Check if the field in coordinate space has already been computed""" return self._c is not None + @property def has_f(self) -> bool: """Check if the field in fourier space has already been computed""" return self._f is not None - @property - def shape(self) -> Tuple[int]: - return self._c.shape - - @property - def ndim(self) -> int: - return self._c.ndim - @property def c(self) -> np.ndarray: """Getter for field in coordinate space""" - if self._c is None: + if not self.has_c: self._logger.debug(f'Computing {self._field_name}(x) = FFT^-1[ {self._field_name}(f) ]', ) assert self.has_f self._c = irfftn(self._f, axes=self._fft_axes, s=self._fft_shape, workers=self._fft_workers) @@ -103,9 +93,9 @@ def c(self, c: np.ndarray): @property def f(self) -> np.ndarray: """Getter for field in fourier space""" - if self._f is None: + if not self.has_f: self._logger.debug(f'Computing {self._field_name}(f) = FFT[ {self._field_name}(x) ]') - assert self.has_c() + assert self.has_c self._f = rfftn(self._c, axes=self._fft_axes, s=self._fft_shape, workers=self._fft_workers) return self._f @@ -113,7 +103,7 @@ def f_padded(self, pad_mode: Dict = None, pad_width: Tuple[Tuple[int, int], ...] """Getter for padded field in fourier space""" if self._f_padded is None: self._logger.debug(f'Computing {self._field_name}_padded(f) = FFT[ {self._field_name}_padded(x) ]') - assert self.has_c() + assert self.has_c c = np.pad(self._c, pad_width, **pad_mode) # TODO: we should actually make sure that the padding does not change between calls @@ -132,7 +122,7 @@ def f_reversed(self) -> np.ndarray: """Getter for time-reversed field in fourier space, intentionally no setter for now""" if self._f_reversed is None: self._logger.debug(f'Computing {self._field_name}_rev(f) = FFT[ {self._field_name}(-x) ]') - assert self.has_c() + assert self.has_c c_reversed = np.flip(self._c, axis=self._fft_axes) self._f_reversed = rfftn(c_reversed, axes=self._fft_axes, s=self._fft_shape, workers=self._fft_workers) return self._f_reversed diff --git a/tnmf/backends/NumPy_FFT.py b/tnmf/backends/NumPy_FFT.py index 5bd6fc61..87fed939 100644 --- a/tnmf/backends/NumPy_FFT.py +++ b/tnmf/backends/NumPy_FFT.py @@ -3,7 +3,7 @@ Shift-invariance is implemented via fast convolution in the Fourier domain using :func:`scipy.fft.rfftn` and :func:`scipy.fft.irfftn`. """ -from typing import Tuple, Dict, Union +from typing import Tuple, Dict import numpy as np from scipy.fft import next_fast_len, rfftn, irfftn @@ -12,7 +12,7 @@ def fftconvolve_sum( - in1: Union[np.ndarray, Tuple[np.ndarray, ...]], + in1: Tuple[np.ndarray, ...], in2: np.ndarray, fft_axes: Tuple[int, ...], slices: Tuple[slice, ...], @@ -43,10 +43,6 @@ def padded_rfftn( # Fourier transform (for real data) return rfftn(array_padded, axes=axes) - if not isinstance(in1, tuple): - assert isinstance(in1, np.ndarray) - in1 = (in1, ) - assert isinstance(in2, np.ndarray) ndim = in2.ndim

Code Coverage of unit tests Should be analyzed and monitored.

2021-07-12T17:10:05

0.0

[]

Yikai-Liao/symusic

Yikai-Liao__symusic-11

1bb8873e3077b80051be59afbaf1dc4c52ce574a

diff --git a/setup.py b/setup.py index bf13886..4e36a1c 100644 --- a/setup.py +++ b/setup.py @@ -139,8 +139,8 @@ def build_extension(self, ext: CMakeExtension) -> None: long_description=open("README.md", encoding="utf-8").read(), long_description_content_type="text/markdown", license=open("LICENSE", encoding="utf-8").read(), - package_dir={"symusic": "py-symusic"}, - packages=["py-symusic"], + package_dir={"symusic": "symusic"}, + packages=["symusic"], ext_modules=[CMakeExtension("symusic.core", ".")], cmdclass={"build_ext": CMakeBuild}, package_data={"symusic": ["**/*.pyi"]}, diff --git a/py-symusic/__init__.py b/symusic/__init__.py similarity index 100% rename from py-symusic/__init__.py rename to symusic/__init__.py diff --git a/py-symusic/core.pyi b/symusic/core.pyi similarity index 100% rename from py-symusic/core.pyi rename to symusic/core.pyi diff --git a/py-symusic/factory.py b/symusic/factory.py similarity index 100% rename from py-symusic/factory.py rename to symusic/factory.py diff --git a/py-symusic/types.py b/symusic/types.py similarity index 100% rename from py-symusic/types.py rename to symusic/types.py

Tests CI workflows Following #4 , this PR set up GitHub actions to run the tests on pushes and pull requests. **Not ready for merge yet. Right now the action only perform builds as done in the "publish" action**

Also, the tests pass for 4 MIDIs out of 26, I'll try to investigate why There is a time limitation of github actions. So we'd better reduce some platform (like linux arm64 based on qemu and macos) and some python versions for test. <img width="1072" alt="image" src="https://github.com/Yikai-Liao/symusic/assets/110762732/f0234b30-62d1-4e88-b456-635f6656ccba"> Absolutely, apologies for this unexpected situation 😅 In the next commit I'll temporarily disable the action anyway @Yikai-Liao @lzqlzzq I updated the `tests.yml` workflow, removing all the builds parts and reducing the number of platforms and python versions. It is possible to run the tests on macOS arm runners, but as the [pricing / hours multiplicator](https://github.com/actions/runner-images/issues/8439) is high it might be a better idea to not test it. Should we however use QEMU with linux to test on arm64? Hopefully the tests should run very quick, they only take ~10min for miditoolkit, here the install should be a bit longer but the rest should be fast. It might also be a good idea to rename the symusic directory, as it prevents to import the installed symusic module when running tests locally (from the root project directory). What do you think? Well, Linux containers using qemu was about `8 times slower` than x64 containers (in our previous pratice). Thankfully, we don't have that many bugs to fix right now, which means the pace of releases has slowed down considerably. (I'm currently focusing on design `libsymusic`, which is extracted from the single header file in `symusic` with more friendly c++ interface and will be easier to maintain and extend) Therefore, the test can be done without worrying too much about exceeding the total time limit of github actions. starting next month. And by the way, my high-performance arm development board (8 cores, 16g RAM and 256g eMMC) has arrived 🎉! And I'll test symusic on it frequently. > It might also be a good idea to rename the symusic directory, as it prevents to import the installed symusic module when running tests locally (from the root project directory). What do you think? I agree with it. The problem is that I'm not familiar with how `setup.py` is actually working. And the reason why I named that folder symusic is that it can be copied to installation path directly(without renaming it). I just don't know how to achieve more complex operations. That's great news! So indeed I think we can safely reduce the range of platforms / python versions of tests. I can take a look for the setup.py behaviour, and if it's not complicated include the changes in this PR. BTW, in the future switching to a `pyproject.toml` package description could be considered (setup.py is legacy), if this is possible as I'm not sure the bindings are supported. BTW, I just updated the test, with the right sorting only 8 are failing now! 🎉 There seem to be two sorts of problems that should be easy to fix: 1. A mismatch between note velocity and duration. This surely comes from a FIFO error, I encountered a similar issue with miditoolkit. And that's the same error causing the last miditok test to fail; 2. Tempo assertion errors, with tempos being equal but the `.tempo` value is a fraction different, likely to come from a float type conversion error somewhere between the original MIDI format and the one written. I don't know if it can be fixed, but if it cannot I suggest to rounds `.tempo` when parsing a MIDI, or to do the rounding in `Tempo.__equal__`. <img width="477" alt="Capture d’écran 2023-12-19 à 17 32 54" src="https://github.com/Yikai-Liao/symusic/assets/56734983/684773b1-1b53-4ac5-a611-fa37179677e5"> For reference on the FIFO issue when writing a MIDI: https://github.com/YatingMusic/miditoolkit/pull/28 And how the FIFO loading was added: https://github.com/YatingMusic/miditoolkit/pull/14 For `setup.py`, we just need to change the root directory name for the `packages` and `package_dir` entries. Would you like it to be done in this PR? And if so, which new name should we give to the dir? * First, for the new folder name of `symusic`, I have checked some pupular repository: | repository | folder name | | -- | -- | | [numpy](https://github.com/numpy/numpy)| numpy | | [pytorch](https://github.com/pytorch/pytorch) | torch | | [mido](https://github.com/mido/mido) | mido | | [polars](https://github.com/pola-rs/polars) | py-polars | | [matplotlib](https://github.com/matplotlib/matplotlib) | lib/matplotlib | So, what about `py-symusic`? (BTW, why do many libraries just use the same name as the folder name? ) * Second, I have understood what's wrong with currently `midi writing` strategy. However, I'm still confused, becuase midi also support real time applications. If we send two `note-on` events in the same channel, with the same pitch (the only difference lies their velocity), then, how can we decide their order based on the given information to bring it into line with the `FIFO` rule? And is there any standard describing this rule in midi? Actually, it feels more like an undefined behavior to me. * Third, for the `__eq__` of `tempo`, introduing an epsilon in compartion might cause lots of unexpected problems according to this [anwser](https://stackoverflow.com/questions/17333/how-do-you-compare-float-and-double-while-accounting-for-precision-loss) in stack overflow. At the same time, I don't find standard for bpm quantization. In short I think this issue should be handled with care. * Ok for `py-symusic`. I don't why they keep the same name when the package has to be built before being tested, but they surely have some way to handle it; * Indeed there is no specific recommendation in the MIDI specifications when to it comes to the order of incoming events. mido is not FIFO if I remember well. However I think it would be good practice to at least keep a consistency between how the MIDI information is parsed and how it is written, either FIFO or LIFO, in order to keep the same information when dumping back; * Noted. I suggest to maybe merge this branch as it comes with a lot of file changes, with minimal test platforms activated, before solving the above issues? I fail to merge. > refusing to allow an OAuth App to create or update workflow `.github/workflows/lint.yml` without workflow`scope I'm not sure what's the cause of this issue, if it's not solves yet I'll look into it after making sure the pytest action is run properly (8 tests should fail, right now I don't know why it cannot find the pytest command). It works on browser, but not on my phone Ok, there will still be one or two things to fix for the action to run properly.

2023-12-21T16:45:09

0.0

[]

go-python/gopy

go-python__gopy-321

d342bb85d1c2f1abdf90781820bcfe726e3f8a8c

diff --git a/.github/workflows/ci.yml b/.github/workflows/ci.yml index 5f1dbd4..5be3246 100644 --- a/.github/workflows/ci.yml +++ b/.github/workflows/ci.yml @@ -22,7 +22,7 @@ jobs: name: Build strategy: matrix: - go-version: [1.19.x, 1.18.x] + go-version: [1.20.x, 1.19.x] platform: [ubuntu-latest] #platform: [ubuntu-latest, macos-latest, windows-latest] runs-on: ${{ matrix.platform }} @@ -52,26 +52,17 @@ jobs: if: matrix.platform == 'ubuntu-latest' run: | sudo apt-get update - sudo apt-get install curl libffi-dev python-cffi python3-cffi python3-pip + sudo apt-get install curl libffi-dev python3-cffi python3-pip # pypy3 isn't packaged in ubuntu yet. TEMPDIR=$(mktemp -d) - curl -L https://downloads.python.org/pypy/pypy2.7-${PYPYVERSION}-linux64.tar.bz2 --output $TEMPDIR/pypy2.tar.bz2 curl -L https://downloads.python.org/pypy/pypy3.6-${PYPYVERSION}-linux64.tar.bz2 --output $TEMPDIR/pypy3.tar.bz2 - tar xf $TEMPDIR/pypy2.tar.bz2 -C $TEMPDIR tar xf $TEMPDIR/pypy3.tar.bz2 -C $TEMPDIR - sudo ln -s $TEMPDIR/pypy2.7-$PYPYVERSION-linux64/bin/pypy /usr/local/bin/pypy sudo ln -s $TEMPDIR/pypy3.6-$PYPYVERSION-linux64/bin/pypy3 /usr/local/bin/pypy3 - # install pip (for pypy, python2) - curl -L https://bootstrap.pypa.io/pip/2.7/get-pip.py --output ${TEMPDIR}/get-pip2.py - python2 ${TEMPDIR}/get-pip2.py # curl -L https://bootstrap.pypa.io/get-pip.py --output ${TEMPDIR}/get-pip.py - # pypy ${TEMPDIR}/get-pip.py # pypy3 ${TEMPDIR}/get-pip.py # install pybindgen - python2 -m pip install --user -U pybindgen python3 -m pip install --user -U pybindgen - # pypy -m pip install --user -U pybindgen # pypy3 -m pip install --user -U pybindgen # install goimports diff --git a/SUPPORT_MATRIX.md b/SUPPORT_MATRIX.md index f23c8f6..9153644 100644 --- a/SUPPORT_MATRIX.md +++ b/SUPPORT_MATRIX.md @@ -3,31 +3,31 @@ NOTE: File auto-generated by TestCheckSupportMatrix in main_test.go. Please don't modify manually. -Feature |py2 | py3 ---- | --- | --- -_examples/arrays | yes | yes -_examples/cgo | yes | yes -_examples/consts | yes | yes -_examples/cstrings | yes | yes -_examples/empty | yes | yes -_examples/funcs | yes | yes -_examples/gopygc | yes | yes -_examples/gostrings | yes | yes -_examples/hi | no | yes -_examples/iface | no | yes -_examples/lot | yes | yes -_examples/maps | yes | yes -_examples/named | yes | yes -_examples/osfile | yes | yes -_examples/pkgconflict | yes | yes -_examples/pointers | yes | yes -_examples/pyerrors | yes | yes -_examples/rename | yes | yes -_examples/seqs | yes | yes -_examples/simple | yes | yes -_examples/sliceptr | yes | yes -_examples/slices | yes | yes -_examples/structs | yes | yes -_examples/unicode | no | yes -_examples/variadic | no | yes -_examples/vars | yes | yes +Feature |py3 +--- | --- +_examples/arrays | yes +_examples/cgo | yes +_examples/consts | yes +_examples/cstrings | yes +_examples/empty | yes +_examples/funcs | yes +_examples/gopygc | yes +_examples/gostrings | yes +_examples/hi | yes +_examples/iface | yes +_examples/lot | yes +_examples/maps | yes +_examples/named | yes +_examples/osfile | yes +_examples/pkgconflict | yes +_examples/pointers | yes +_examples/pyerrors | yes +_examples/rename | yes +_examples/seqs | yes +_examples/simple | yes +_examples/sliceptr | yes +_examples/slices | yes +_examples/structs | yes +_examples/unicode | yes +_examples/variadic | yes +_examples/vars | yes diff --git a/appveyor.yml b/appveyor.yml index acded50..03aeeb6 100644 --- a/appveyor.yml +++ b/appveyor.yml @@ -14,7 +14,7 @@ branches: environment: GOPATH: C:\gopath - GOROOT: C:\go115 + GOROOT: C:\go119 GOPY_APPVEYOR_CI: '1' GOTRACEBACK: 'crash' #CPYTHON2DIR: "C:\\Python27-x64" @@ -22,17 +22,13 @@ environment: #PATH: '%GOPATH%\bin;%CPYTHON2DIR%;%CPYTHON2DIR%\\Scripts;%CPYTHON3DIR%;%CPYTHON3DIR%\\Scripts;C:\msys64\mingw64\bin;C:\msys64\usr\bin\;%PATH%' PATH: '%GOPATH%\bin;%GOROOT%\bin;%CPYTHON3DIR%;%CPYTHON3DIR%\\Scripts;C:\msys64\mingw64\bin;C:\msys64\usr\bin\;%PATH%' -stack: go 1.15 +stack: go 1.19 build_script: - python --version - #- "%CPYTHON2DIR%\\python --version" - "%CPYTHON3DIR%\\python --version" - #- "%CPYTHON2DIR%\\python -m pip install --upgrade pip" - "%CPYTHON3DIR%\\python -m pip install --upgrade pip" - #- "%CPYTHON2DIR%\\python -m pip install cffi" - "%CPYTHON3DIR%\\python -m pip install cffi" - #- "%CPYTHON2DIR%\\python -m pip install pybindgen" - "%CPYTHON3DIR%\\python -m pip install pybindgen" - go version - go env diff --git a/bind/utils.go b/bind/utils.go index 42ef5bc..02620d6 100644 --- a/bind/utils.go +++ b/bind/utils.go @@ -103,7 +103,13 @@ func (pc *PyConfig) AllFlags() string { // python VM (python, python2, python3, pypy, etc...) func GetPythonConfig(vm string) (PyConfig, error) { code := `import sys -import distutils.sysconfig as ds +try: + import sysconfig as ds + def _get_python_inc(): + return ds.get_path('include') +except ImportError: + import distutils.sysconfig as ds + _get_python_inc = ds.get_config_var import json import os version=sys.version_info.major @@ -133,7 +139,7 @@ else: print(json.dumps({ "version": sys.version_info.major, "minor": sys.version_info.minor, - "incdir": ds.get_python_inc(), + "incdir": _get_python_inc(), "libdir": ds.get_config_var("LIBDIR"), "libpy": ds.get_config_var("LIBRARY"), "shlibs": ds.get_config_var("SHLIBS"), diff --git a/cmd_build.go b/cmd_build.go index 0722629..7b57a7e 100644 --- a/cmd_build.go +++ b/cmd_build.go @@ -5,6 +5,7 @@ package main import ( + "bytes" "fmt" "log" "os" @@ -215,11 +216,17 @@ func runBuild(mode bind.BuildMode, cfg *BuildCfg) error { if bind.WindowsOS { fmt.Printf("Doing windows sed hack to fix declspec for PyInit\n") - cmd = exec.Command("sed", "-i", "s/ PyInit_/ __declspec(dllexport) PyInit_/g", cfg.Name+".c") - cmdout, err = cmd.CombinedOutput() + fname := cfg.Name + ".c" + raw, err := os.ReadFile(fname) if err != nil { - fmt.Printf("cmd had error: %v output:\no%v\n", err, string(cmdout)) - return err + fmt.Printf("could not read %s: %+v", fname, err) + return fmt.Errorf("could not read %s: %w", fname, err) + } + raw = bytes.ReplaceAll(raw, []byte(" PyInit_"), []byte(" __declspec(dllexport) PyInit_")) + err = os.WriteFile(fname, raw, 0644) + if err != nil { + fmt.Printf("could not apply sed hack to fix declspec for PyInit: %+v", err) + return fmt.Errorf("could not apply sed hack to fix PyInit: %w", err) } } diff --git a/go.mod b/go.mod index d0e31cc..512c22f 100644 --- a/go.mod +++ b/go.mod @@ -1,6 +1,6 @@ module github.com/go-python/gopy -go 1.18 +go 1.19 require ( github.com/gonuts/commander v0.1.0

'sed' is not recognized as an internal or external command... sed is not a windows utility and shouldn't be used. I had to build in git bash after handling a handful of other issues to get a package to build. Why not just use python to do it?

I also have this error. C:\python\python.exe build.py Doing windows sed hack to fix declspec for PyInit cmd had error: exec: "sed": executable file not found in %PATH%

2023-03-31T16:08:16

0.0

[]

glotzerlab/rowan

glotzerlab__rowan-56

3dc965380ff1aec428e9fe88ba87a59e7acf7db1

diff --git a/.circleci/config.yml b/.circleci/config.yml deleted file mode 100644 index efc4e69..0000000 --- a/.circleci/config.yml +++ /dev/null @@ -1,130 +0,0 @@ -version: 2.1 - -orbs: - codecov: codecov/[email protected] - -references: - install_dependencies: &install_dependencies - name: Install dependencies - command: | - python -m pip install --progress-bar off -U virtualenv --user - mkdir -p ./venv - python -m virtualenv ./venv --clear - . venv/bin/activate - python -m pip install --progress-bar off -U -r requirements/requirements-testing.txt - - run_tests: &run_tests - name: run tests - command: | - # Run with coverage - . venv/bin/activate - python -m pytest -v --cov=rowan/ --cov-report=xml - -jobs: - linux-python-39: - docker: - - image: cimg/python:3.9 - working_directory: ~/repo - steps: - - checkout - - run: - <<: *install_dependencies - - run: - <<: *run_tests - - linux-python-310: - docker: - - image: cimg/python:3.10 - working_directory: ~/repo - steps: - - checkout - - run: - <<: *install_dependencies - - run: - <<: *run_tests - - linux-python-311: - docker: - - image: cimg/python:3.11 - working_directory: ~/repo - steps: - - checkout - - run: - <<: *install_dependencies - - run: - <<: *run_tests - - linux-python-312: - docker: - - image: cimg/python:3.12 - working_directory: ~/repo - steps: - - checkout - - run: - <<: *install_dependencies - - run: - <<: *run_tests - - linux-oldest: - # Run tests against a set of the oldest dependencies we support. - docker: - - image: cimg/python:3.8 - working_directory: ~/repo - steps: - - checkout - - run: - name: Install oldest dependencies - command: | - python -m pip install --progress-bar off -U virtualenv --user - mkdir -p ./venv - python -m virtualenv ./venv --clear - . venv/bin/activate - python -m pip install --progress-bar off -U -r .circleci/oldest-test-reqs.txt - python -m pip install --progress-bar off -U pytest - - run: - name: run tests - command: | - . venv/bin/activate - python -m pytest -v tests - - build_and_deploy: - docker: - - image: cimg/python:3.12 - working_directory: ~/repo - steps: - - checkout - - run: - <<: *install_dependencies - - run: - <<: *run_tests - - run: - name: Deploy dist and wheels - command: | - . venv/bin/activate - python -m pytest -v tests - python --version - python -m pip --version - python -m pip install --progress-bar off --user -U twine build - python -m twine --version - python -m build . - twine upload --username ${PYPI_USERNAME} --password ${PYPI_PASSWORD} dist/* - -workflows: - version: 2 - testing: - jobs: - - linux-python-39 - - linux-python-310 - - linux-python-311 - - linux-python-312: - post-steps: - - codecov/upload - - linux-oldest - deploy: - jobs: - - build_and_deploy: - filters: - tags: - only: /^v.*/ - branches: - ignore: /.*/ diff --git a/.github/workflows/release.yaml b/.github/workflows/release.yaml new file mode 100644 index 0000000..0a020e4 --- /dev/null +++ b/.github/workflows/release.yaml @@ -0,0 +1,61 @@ +name: PyPI + +concurrency: + group: ${{ github.workflow }}-${{ github.ref }} + cancel-in-progress: true + +on: + pull_request: + + push: + branches: + - "trunk-*" + tags: + - "v*" + + workflow_dispatch: + +jobs: + build_wheel: + name: Build wheel + runs-on: ubuntu-24.04 + steps: + - name: Checkout + uses: actions/checkout@a5ac7e51b41094c92402da3b24376905380afc29 # v4.1.6 + - name: Set up Python + uses: actions/setup-python@82c7e631bb3cdc910f68e0081d67478d79c6982d # v5.1.0 + with: + python-version: 3.12 + - name: Set up Python environment + uses: glotzerlab/workflows/setup-uv@5cfac9da9cb78e16ae97a9119b6fd13c1c2d6f5e # 0.1.0 + with: + lockfile: ".github/workflows/requirements-build.txt" + + - name: Build wheel + run: python3 -m build --outdir dist/ . + + - uses: actions/upload-artifact@65462800fd760344b1a7b4382951275a0abb4808 # v4.3.3 + with: + name: dist + path: dist + + upload_pypi: + name: Publish [PyPI] + needs: [build_wheel] + runs-on: ubuntu-latest + permissions: + id-token: write + steps: + - name: Download artifacts + uses: actions/download-artifact@65a9edc5881444af0b9093a5e628f2fe47ea3b2e # v4.1.7 + with: + merge-multiple: 'true' + pattern: dist + path: dist + + - name: Check files + run: ls -lR dist + + - name: Upload to PyPI + if: startsWith(github.ref, 'refs/tags/v') + uses: pypa/gh-action-pypi-publish@ec4db0b4ddc65acdf4bff5fa45ac92d78b56bdf0 # v1.9.0 diff --git a/.github/workflows/requirements-build.in b/.github/workflows/requirements-build.in new file mode 100644 index 0000000..f9a7807 --- /dev/null +++ b/.github/workflows/requirements-build.in @@ -0,0 +1,1 @@ +build == 1.1.1 diff --git a/.github/workflows/requirements-build.txt b/.github/workflows/requirements-build.txt new file mode 100644 index 0000000..2b16d8c --- /dev/null +++ b/.github/workflows/requirements-build.txt @@ -0,0 +1,8 @@ +# This file was autogenerated by uv via the following command: +# uv pip compile requirements-build.in +build==1.1.1 + # via -r requirements-build.in +packaging==24.1 + # via build +pyproject-hooks==1.1.0 + # via build

Migrate to Github Actions Migration to GitHub actions will streamline the CI and bring Rowan up to date with the rest of the Glotzerlab software suite. This will also help in testing for Numpy 2.x compatibility.

2024-06-20T16:47:16

0.0

[]

trevismd/statannotations

trevismd__statannotations-35

c4704fbbff565c1d3eea0e93f4417b0ec67bd9b0

diff --git a/CHANGELOG.md b/CHANGELOG.md index c23571f..ecb14f2 100644 --- a/CHANGELOG.md +++ b/CHANGELOG.md @@ -1,4 +1,13 @@ ## v0.4 +### v0.4.3 +- The `correction_format` parameter allows changing how the multiple +comparisons correction method adjusts the annotation when changing a result +to non-significant. +- Fix the `show_test_name` configuration. +- Fix the `verbose` parameter + ([#37](https://github.com/trevismd/statannotations/pull/37) by + [mxposed](https://github.com/mxposed)) + ### v0.4.2 - Support of `FacetGrid` with - Support empty initialization only defining pairs diff --git a/docs/build/doctrees/environment.pickle b/docs/build/doctrees/environment.pickle deleted file mode 100644 index 880be64..0000000 Binary files a/docs/build/doctrees/environment.pickle and /dev/null differ diff --git a/docs/build/doctrees/statannotations.doctree b/docs/build/doctrees/statannotations.doctree index c28d580..1f119a8 100644 Binary files a/docs/build/doctrees/statannotations.doctree and b/docs/build/doctrees/statannotations.doctree differ diff --git a/docs/build/doctrees/statannotations.stats.doctree b/docs/build/doctrees/statannotations.stats.doctree index a4c0dfe..747056f 100644 Binary files a/docs/build/doctrees/statannotations.stats.doctree and b/docs/build/doctrees/statannotations.stats.doctree differ diff --git a/docs/build/html/genindex.html b/docs/build/html/genindex.html index 6ff36ff..8c420f7 100644 --- a/docs/build/html/genindex.html +++ b/docs/build/html/genindex.html @@ -180,6 +180,8 @@ <h1 id="index">Index</h1> <h2 id="A">A</h2> <table style="width: 100%" class="indextable genindextable"><tr> <td style="width: 33%; vertical-align: top;"><ul> + <li><a href="statannotations.stats.html#statannotations.stats.StatResult.StatResult.adjust">adjust() (statannotations.stats.StatResult.StatResult method)</a> +</li> <li><a href="statannotations.html#statannotations.Annotator.Annotator.alpha">alpha (statannotations.Annotator.Annotator property)</a> </li> <li><a href="statannotations.html#statannotations.Annotator.Annotator.annotate">annotate() (statannotations.Annotator.Annotator method)</a> @@ -241,6 +243,8 @@ <h2 id="C">C</h2> <li><a href="statannotations.html#statannotations.Annotator.Annotator.configure">configure() (statannotations.Annotator.Annotator method)</a> </li> <li><a href="statannotations.stats.html#statannotations.stats.StatResult.StatResult.corrected_significance">corrected_significance (statannotations.stats.StatResult.StatResult property)</a> +</li> + <li><a href="statannotations.html#statannotations.PValueFormat.PValueFormat.correction_format">correction_format (statannotations.PValueFormat.PValueFormat property)</a> </li> <li><a href="statannotations.stats.html#statannotations.stats.StatResult.StatResult.correction_method">correction_method (statannotations.stats.StatResult.StatResult property)</a> </li> @@ -302,10 +306,12 @@ <h2 id="G">G</h2> <li><a href="statannotations.html#statannotations.PValueFormat.PValueFormat.get_configuration">(statannotations.PValueFormat.PValueFormat method)</a> </li> </ul></li> - <li><a href="statannotations.stats.html#statannotations.stats.ComparisonsCorrection.get_correction_parameters">get_correction_parameters() (in module statannotations.stats.ComparisonsCorrection)</a> + <li><a href="statannotations.html#statannotations.PValueFormat.get_corrected_star">get_corrected_star() (in module statannotations.PValueFormat)</a> </li> </ul></td> <td style="width: 33%; vertical-align: top;"><ul> + <li><a href="statannotations.stats.html#statannotations.stats.ComparisonsCorrection.get_correction_parameters">get_correction_parameters() (in module statannotations.stats.ComparisonsCorrection)</a> +</li> <li><a href="statannotations.html#statannotations.Annotator.Annotator.get_empty_annotator">get_empty_annotator() (statannotations.Annotator.Annotator static method)</a> </li> <li><a href="statannotations.stats.html#statannotations.stats.utils.get_num_comparisons">get_num_comparisons() (in module statannotations.stats.utils)</a> diff --git a/docs/build/html/objects.inv b/docs/build/html/objects.inv index faca044..ca638c7 100644 Binary files a/docs/build/html/objects.inv and b/docs/build/html/objects.inv differ diff --git a/docs/build/html/searchindex.js b/docs/build/html/searchindex.js index 6baea8a..6750b03 100644 --- a/docs/build/html/searchindex.js +++ b/docs/build/html/searchindex.js @@ -1,1 +1,1 @@ -Search.setIndex({docnames:["index","modules","setup","statannotations","statannotations.stats"],envversion:{"sphinx.domains.c":2,"sphinx.domains.changeset":1,"sphinx.domains.citation":1,"sphinx.domains.cpp":3,"sphinx.domains.index":1,"sphinx.domains.javascript":2,"sphinx.domains.math":2,"sphinx.domains.python":3,"sphinx.domains.rst":2,"sphinx.domains.std":2,sphinx:56},filenames:["index.rst","modules.rst","setup.rst","statannotations.rst","statannotations.stats.rst"],objects:{"":{statannotations:[3,0,0,"-"]},"statannotations.Annotation":{Annotation:[3,1,1,""]},"statannotations.Annotation.Annotation":{formatted_output:[3,2,1,""],print_labels_and_content:[3,3,1,""],text:[3,2,1,""]},"statannotations.Annotator":{Annotator:[3,1,1,""]},"statannotations.Annotator.Annotator":{alpha:[3,2,1,""],annotate:[3,3,1,""],annotate_custom_annotations:[3,3,1,""],apply_and_annotate:[3,3,1,""],apply_test:[3,3,1,""],comparisons_correction:[3,2,1,""],configure:[3,3,1,""],fig:[3,2,1,""],get_annotations_text:[3,3,1,""],get_configuration:[3,3,1,""],get_empty_annotator:[3,3,1,""],get_offset_func:[3,3,1,""],has_type0_comparisons_correction:[3,3,1,""],loc:[3,2,1,""],new_plot:[3,3,1,""],orient:[3,2,1,""],plot_and_annotate:[3,3,1,""],plot_and_annotate_facets:[3,3,1,""],print_pvalue_legend:[3,3,1,""],pvalue_format:[3,2,1,""],reset_configuration:[3,3,1,""],set_custom_annotations:[3,3,1,""],set_pvalues:[3,3,1,""],set_pvalues_and_annotate:[3,3,1,""],test:[3,2,1,""],validate_test_short_name:[3,3,1,""],verbose:[3,2,1,""]},"statannotations.PValueFormat":{Formatter:[3,1,1,""],PValueFormat:[3,1,1,""],sort_pvalue_thresholds:[3,4,1,""]},"statannotations.PValueFormat.Formatter":{config:[3,3,1,""],format_data:[3,3,1,""]},"statannotations.PValueFormat.PValueFormat":{config:[3,3,1,""],format_data:[3,3,1,""],get_configuration:[3,3,1,""],print_legend_if_used:[3,3,1,""],pvalue_format_string:[3,2,1,""],pvalue_thresholds:[3,2,1,""],simple_format_string:[3,2,1,""],text_format:[3,2,1,""]},"statannotations.format_annotations":{pval_annotation_text:[3,4,1,""],simple_text:[3,4,1,""]},"statannotations.stats":{ComparisonsCorrection:[4,0,0,"-"],StatResult:[4,0,0,"-"],StatTest:[4,0,0,"-"],test:[4,0,0,"-"],utils:[4,0,0,"-"]},"statannotations.stats.ComparisonsCorrection":{ComparisonsCorrection:[4,1,1,""],check_valid_correction_name:[4,4,1,""],get_correction_parameters:[4,4,1,""],get_validated_comparisons_correction:[4,4,1,""]},"statannotations.stats.ComparisonsCorrection.ComparisonsCorrection":{apply:[4,3,1,""],document:[4,3,1,""]},"statannotations.stats.StatResult":{StatResult:[4,1,1,""]},"statannotations.stats.StatResult.StatResult":{corrected_significance:[4,2,1,""],correction_method:[4,2,1,""],formatted_output:[4,2,1,""],significance_suffix:[4,2,1,""]},"statannotations.stats.StatTest":{StatTest:[4,1,1,""],wilcoxon:[4,4,1,""]},"statannotations.stats.StatTest.StatTest":{from_library:[4,3,1,""],short_name:[4,2,1,""]},"statannotations.stats.test":{apply_test:[4,4,1,""]},"statannotations.stats.utils":{check_alpha:[4,4,1,""],check_num_comparisons:[4,4,1,""],check_pvalues:[4,4,1,""],get_num_comparisons:[4,4,1,""],return_results:[4,4,1,""]},"statannotations.utils":{InvalidParametersError:[3,5,1,""],check_is_in:[3,4,1,""],check_not_none:[3,4,1,""],check_order_in_data:[3,4,1,""],check_pairs_in_data:[3,4,1,""],check_valid_text_format:[3,4,1,""],empty_dict_if_none:[3,4,1,""],get_closest:[3,4,1,""],get_x_values:[3,4,1,""],raise_expected_got:[3,4,1,""],remove_null:[3,4,1,""],render_collection:[3,4,1,""]},statannotations:{Annotation:[3,0,0,"-"],Annotator:[3,0,0,"-"],PValueFormat:[3,0,0,"-"],format_annotations:[3,0,0,"-"],stats:[4,0,0,"-"],utils:[3,0,0,"-"]}},objnames:{"0":["py","module","Python module"],"1":["py","class","Python class"],"2":["py","property","Python property"],"3":["py","method","Python method"],"4":["py","function","Python function"],"5":["py","exception","Python exception"]},objtypes:{"0":"py:module","1":"py:class","2":"py:property","3":"py:method","4":"py:function","5":"py:exception"},terms:{"0":[3,4],"001":3,"01":3,"05":[3,4],"1":[3,4],"12141511":3,"1e":3,"2":3,"3":3,"4":3,"5":3,"5e":3,"9981846":3,"case":4,"class":[3,4],"default":[3,4],"float":4,"function":[3,4],"int":4,"return":3,"short":3,"static":[3,4],"true":[3,4],A:3,For:3,If:3,One:3,Or:3,The:3,a_list:3,accept:3,add:3,addit:[3,4],after:3,alpha:[3,4],an:3,annot:1,annotate_custom_annot:3,annotate_param:3,annotation_func:3,annotation_param:3,appli:[3,4],apply_and_annot:3,apply_test:[3,4],ar:3,arg:[3,4],argument:[3,4],arrai:3,assum:3,auto:3,ax:3,ax_op_aft:3,ax_op_befor:3,bar:3,base:[3,4],befor:3,behavior:[3,4],being:3,benjamini:4,between:3,bonferroni:4,bool:4,box:3,boxplot:3,calcul:3,call:3,callabl:4,can:3,check:3,check_alpha:4,check_is_in:3,check_not_non:3,check_num_comparison:4,check_order_in_data:3,check_pairs_in_data:3,check_pvalu:4,check_valid_correction_nam:4,check_valid_text_format:3,classmethod:3,close:3,closest:3,collect:3,color:3,com:3,compar:3,comparison:[3,4],comparisons_correct:[3,4],comparisonscorrect:[1,3],comput:3,config:3,configur:3,constructor:3,content:1,coord:3,coordin:3,core:3,corr_kwarg:4,correct:[3,4],corrected_signific:4,correction_method:4,creat:3,custom:3,data:[3,4],datafram:3,defin:3,dict:[3,4],displai:3,document:4,each:3,earlier:4,either:4,empty_dict_if_non:3,engin:3,equal:3,equival:4,error:3,error_typ:3,exact:3,except:3,exist:3,expect:3,facetgrid:3,fals:3,fig:3,first:3,follow:3,for_:3,format:[3,4],format_annot:1,format_data:3,formatt:3,formatted_output:[3,4],fraction:3,frame:3,from:[3,4],from_librari:4,func:4,func_nam:3,fwer:3,gener:3,get:4,get_annotations_text:3,get_closest:3,get_configur:3,get_correction_paramet:4,get_empty_annot:3,get_num_comparison:4,get_offset_func:3,get_validated_comparisons_correct:4,get_x_valu:3,got:3,group:3,group_data1:4,group_data2:4,gt:4,has_type0_comparisons_correct:3,have:3,hochberg:4,hold:3,holm:4,http:3,hue:3,hue_ord:3,index:0,initi:3,instanc:[3,4],instead:3,interfac:4,interpret:4,invalidparameterserror:3,keyword:4,kruskal:4,kwarg:[3,4],label:3,leven:4,line_height:3,line_offset:3,line_offset_to_group:3,line_width:3,link:3,list:3,loc:3,ls:4,mann:4,map:3,messag:3,method:[3,4],method_typ:4,mode:3,modul:[0,1],multipl:[3,4],multipletest:3,must:3,mylist:3,mynumb:3,name:[3,4],new_plot:3,none:[3,4],ns:3,num_comparison:[3,4],number:[3,4],object:[3,4],one:[3,4],onto:3,option:[3,4],order:3,orient:3,otherwis:3,overrid:3,p:3,p_valu:4,packag:[0,1],page:0,pair:3,panda:3,paramet:[3,4],pass:[3,4],per:3,perform:3,plot:3,plot_and_annot:3,plot_and_annotate_facet:3,plot_param:3,point:3,posit:3,print_labels_and_cont:3,print_legend_if_us:3,print_pvalue_legend:3,properti:[3,4],provid:[3,4],pval:4,pval_annotation_text:3,pvalu:[3,4],pvalue_format:3,pvalue_format_str:3,pvalue_threshold:3,pvalueformat:1,rais:3,raise_expected_got:3,rang:3,refer:3,remove_nul:3,render:3,render_collect:3,reset_configur:3,result:[3,4],results_arrai:4,return_result:4,run:4,same:3,sampl:4,seaborn:3,search:0,see:3,sep:3,seri:3,set:3,set_custom_annot:3,set_pvalu:3,set_pvalues_and_annot:3,short_nam:4,show:3,show_test_nam:3,significance_suffix:4,simpl:3,simple_format_str:3,simple_text:3,singl:3,smallest:3,sort:3,sort_pvalue_threshold:3,stackoverflow:3,standard:3,star:3,stat:[1,3],stat_nam:4,stat_str:4,statist:[3,4],statresult:[1,3],stats_param:[3,4],statsmodel:[3,4],statsmodels_api:4,stattest:[1,3],str:[3,4],string:3,struct:3,submodul:1,subpackag:1,t:4,test:[1,3],test_descript:4,test_ind:4,test_long_nam:4,test_nam:4,test_pair:4,test_result_list:4,test_short_nam:[3,4],test_welch:4,text:3,text_annot_custom:3,text_format:3,text_offset:3,them:3,thereof:3,thi:[3,4],three:3,threshold:3,top:3,tupl:3,two:[3,4],type:3,union:[3,4],us:[3,4],use_fixed_offset:3,util:1,valid_valu:3,validate_test_short_nam:3,valu:[3,4],valueerror:3,verbos:[3,4],version:4,vs:3,w:3,wa:3,whitnei:4,wilcoxon:4,within:3,work:3,x:3,y:3,yekuti:4},titles:["Welcome to statannotations\u2019s documentation!","statannotations","setup module","statannotations package","statannotations.stats package"],titleterms:{annot:3,comparisonscorrect:4,content:[0,3,4],document:0,format_annot:3,indic:0,modul:[2,3,4],packag:[3,4],pvalueformat:3,s:0,setup:2,stat:4,statannot:[0,1,3,4],statresult:4,stattest:4,submodul:[3,4],subpackag:3,tabl:0,test:4,util:[3,4],welcom:0}}) \ No newline at end of file +Search.setIndex({docnames:["index","modules","setup","statannotations","statannotations.stats"],envversion:{"sphinx.domains.c":2,"sphinx.domains.changeset":1,"sphinx.domains.citation":1,"sphinx.domains.cpp":3,"sphinx.domains.index":1,"sphinx.domains.javascript":2,"sphinx.domains.math":2,"sphinx.domains.python":3,"sphinx.domains.rst":2,"sphinx.domains.std":2,sphinx:56},filenames:["index.rst","modules.rst","setup.rst","statannotations.rst","statannotations.stats.rst"],objects:{"":{statannotations:[3,0,0,"-"]},"statannotations.Annotation":{Annotation:[3,1,1,""]},"statannotations.Annotation.Annotation":{formatted_output:[3,2,1,""],print_labels_and_content:[3,3,1,""],text:[3,2,1,""]},"statannotations.Annotator":{Annotator:[3,1,1,""]},"statannotations.Annotator.Annotator":{alpha:[3,2,1,""],annotate:[3,3,1,""],annotate_custom_annotations:[3,3,1,""],apply_and_annotate:[3,3,1,""],apply_test:[3,3,1,""],comparisons_correction:[3,2,1,""],configure:[3,3,1,""],fig:[3,2,1,""],get_annotations_text:[3,3,1,""],get_configuration:[3,3,1,""],get_empty_annotator:[3,3,1,""],get_offset_func:[3,3,1,""],has_type0_comparisons_correction:[3,3,1,""],loc:[3,2,1,""],new_plot:[3,3,1,""],orient:[3,2,1,""],plot_and_annotate:[3,3,1,""],plot_and_annotate_facets:[3,3,1,""],print_pvalue_legend:[3,3,1,""],pvalue_format:[3,2,1,""],reset_configuration:[3,3,1,""],set_custom_annotations:[3,3,1,""],set_pvalues:[3,3,1,""],set_pvalues_and_annotate:[3,3,1,""],test:[3,2,1,""],validate_test_short_name:[3,3,1,""],verbose:[3,2,1,""]},"statannotations.PValueFormat":{Formatter:[3,1,1,""],PValueFormat:[3,1,1,""],get_corrected_star:[3,4,1,""],sort_pvalue_thresholds:[3,4,1,""]},"statannotations.PValueFormat.Formatter":{config:[3,3,1,""],format_data:[3,3,1,""]},"statannotations.PValueFormat.PValueFormat":{config:[3,3,1,""],correction_format:[3,2,1,""],format_data:[3,3,1,""],get_configuration:[3,3,1,""],print_legend_if_used:[3,3,1,""],pvalue_format_string:[3,2,1,""],pvalue_thresholds:[3,2,1,""],simple_format_string:[3,2,1,""],text_format:[3,2,1,""]},"statannotations.format_annotations":{pval_annotation_text:[3,4,1,""],simple_text:[3,4,1,""]},"statannotations.stats":{ComparisonsCorrection:[4,0,0,"-"],StatResult:[4,0,0,"-"],StatTest:[4,0,0,"-"],test:[4,0,0,"-"],utils:[4,0,0,"-"]},"statannotations.stats.ComparisonsCorrection":{ComparisonsCorrection:[4,1,1,""],check_valid_correction_name:[4,4,1,""],get_correction_parameters:[4,4,1,""],get_validated_comparisons_correction:[4,4,1,""]},"statannotations.stats.ComparisonsCorrection.ComparisonsCorrection":{apply:[4,3,1,""],document:[4,3,1,""]},"statannotations.stats.StatResult":{StatResult:[4,1,1,""]},"statannotations.stats.StatResult.StatResult":{adjust:[4,3,1,""],corrected_significance:[4,2,1,""],correction_method:[4,2,1,""],formatted_output:[4,2,1,""],significance_suffix:[4,2,1,""]},"statannotations.stats.StatTest":{StatTest:[4,1,1,""],wilcoxon:[4,4,1,""]},"statannotations.stats.StatTest.StatTest":{from_library:[4,3,1,""],short_name:[4,2,1,""]},"statannotations.stats.test":{apply_test:[4,4,1,""]},"statannotations.stats.utils":{check_alpha:[4,4,1,""],check_num_comparisons:[4,4,1,""],check_pvalues:[4,4,1,""],get_num_comparisons:[4,4,1,""],return_results:[4,4,1,""]},"statannotations.utils":{InvalidParametersError:[3,5,1,""],check_is_in:[3,4,1,""],check_not_none:[3,4,1,""],check_order_in_data:[3,4,1,""],check_pairs_in_data:[3,4,1,""],check_valid_text_format:[3,4,1,""],empty_dict_if_none:[3,4,1,""],get_closest:[3,4,1,""],get_x_values:[3,4,1,""],raise_expected_got:[3,4,1,""],remove_null:[3,4,1,""],render_collection:[3,4,1,""]},statannotations:{Annotation:[3,0,0,"-"],Annotator:[3,0,0,"-"],PValueFormat:[3,0,0,"-"],format_annotations:[3,0,0,"-"],stats:[4,0,0,"-"],utils:[3,0,0,"-"]}},objnames:{"0":["py","module","Python module"],"1":["py","class","Python class"],"2":["py","property","Python property"],"3":["py","method","Python method"],"4":["py","function","Python function"],"5":["py","exception","Python exception"]},objtypes:{"0":"py:module","1":"py:class","2":"py:property","3":"py:method","4":"py:function","5":"py:exception"},terms:{"0":[3,4],"001":3,"01":3,"05":[3,4],"1":[3,4],"12141511":3,"1e":3,"2":3,"3":3,"4":3,"5":3,"5e":3,"9981846":3,"case":4,"class":[3,4],"default":[3,4],"float":4,"function":[3,4],"int":4,"return":3,"short":3,"static":[3,4],"true":[3,4],A:3,For:3,If:3,One:3,Or:3,The:3,a_list:3,accept:3,ad:3,add:3,addit:[3,4],adjust:4,after:3,alpha:[3,4],an:3,annot:1,annotate_custom_annot:3,annotate_param:3,annotation_func:3,annotation_param:3,appli:[3,4],apply_and_annot:3,apply_test:[3,4],ar:3,arg:[3,4],argument:[3,4],arrai:3,assum:3,auto:3,ax:3,ax_op_aft:3,ax_op_befor:3,bar:3,base:[3,4],befor:3,behavior:[3,4],being:3,benjamini:4,between:3,bonferroni:4,bool:4,box:3,boxplot:3,calcul:3,call:3,callabl:4,can:3,check:3,check_alpha:4,check_is_in:3,check_not_non:3,check_num_comparison:4,check_order_in_data:3,check_pairs_in_data:3,check_pvalu:4,check_valid_correction_nam:4,check_valid_text_format:3,classmethod:3,close:3,closest:3,collect:3,color:3,com:3,compar:3,comparison:[3,4],comparisons_correct:[3,4],comparisonscorrect:[1,3],comput:3,config:3,configur:3,constructor:3,content:1,coord:3,coordin:3,core:3,corr_kwarg:4,correct:[3,4],corrected_signific:4,correction_format:3,correction_method:4,correspond:3,creat:3,custom:3,data:[3,4],datafram:3,defin:3,dict:[3,4],displai:3,document:4,each:3,earlier:4,either:4,empty_dict_if_non:3,engin:3,equal:3,equival:4,error:3,error_typ:3,exact:3,except:3,exist:3,expect:3,facetgrid:3,fals:3,fig:3,first:3,follow:3,for_:3,format:[3,4],format_annot:1,format_data:3,formatt:3,formatted_output:[3,4],fraction:3,frame:3,from:[3,4],from_librari:4,func:4,func_nam:3,fwer:3,gener:3,get:4,get_annotations_text:3,get_closest:3,get_configur:3,get_corrected_star:3,get_correction_paramet:4,get_empty_annot:3,get_num_comparison:4,get_offset_func:3,get_validated_comparisons_correct:4,get_x_valu:3,got:3,group:3,group_data1:4,group_data2:4,gt:4,has_type0_comparisons_correct:3,have:3,hochberg:4,hold:3,holm:4,how:3,http:3,hue:3,hue_ord:3,index:0,initi:3,instanc:[3,4],instead:3,interfac:4,interpret:4,invalidparameterserror:3,keyword:4,kruskal:4,kwarg:[3,4],label:3,leven:4,line_height:3,line_offset:3,line_offset_to_group:3,line_width:3,link:3,list:3,loc:3,ls:4,mann:4,map:3,messag:3,method:[3,4],method_typ:4,mode:3,modul:[0,1],multipl:[3,4],multipletest:3,must:3,mylist:3,mynumb:3,name:[3,4],new_plot:3,none:[3,4],notat:3,ns:3,num_comparison:[3,4],number:[3,4],object:[3,4],one:[3,4],onto:3,option:[3,4],order:3,orient:3,origin:3,otherwis:3,output:3,overrid:3,p:3,p_valu:4,packag:[0,1],page:0,pair:3,panda:3,paramet:[3,4],pass:[3,4],per:3,perform:3,plot:3,plot_and_annot:3,plot_and_annotate_facet:3,plot_param:3,point:3,posit:3,print:3,print_labels_and_cont:3,print_legend_if_us:3,print_pvalue_legend:3,properti:[3,4],provid:[3,4],pval:4,pval_annotation_text:3,pvalu:[3,4],pvalue_format:3,pvalue_format_str:3,pvalue_threshold:3,pvalueformat:1,rais:3,raise_expected_got:3,rang:3,re:3,refer:3,remov:3,remove_nul:3,render:3,render_collect:3,replac:3,reset_configur:3,result:[3,4],results_arrai:4,return_result:4,run:4,same:3,sampl:4,seaborn:3,search:0,see:3,sep:3,seri:3,set:3,set_custom_annot:3,set_pvalu:3,set_pvalues_and_annot:3,short_nam:4,short_test_nam:3,show:3,show_test_nam:3,signific:3,significance_suffix:4,simpl:3,simple_format_str:3,simple_text:3,singl:3,smallest:3,sort:3,sort_pvalue_threshold:3,stackoverflow:3,standard:3,star:3,stat:[1,3],stat_nam:4,stat_str:4,stat_summari:4,statist:[3,4],statresult:[1,3],stats_param:[3,4],statsmodel:[3,4],statsmodels_api:4,stattest:[1,3],str:[3,4],string:3,struct:3,submodul:1,subpackag:1,suffix:3,t:4,test:[1,3],test_descript:4,test_ind:4,test_long_nam:4,test_nam:4,test_pair:4,test_result_list:4,test_short_nam:[3,4],test_welch:4,text:3,text_annot_custom:3,text_format:3,text_offset:3,them:3,thereof:3,thi:[3,4],three:3,threshold:3,top:3,tupl:3,two:[3,4],type:3,union:[3,4],us:[3,4],use_fixed_offset:3,util:1,valid_valu:3,validate_test_short_nam:3,valu:[3,4],valueerror:3,verbos:[3,4],version:4,vs:3,w:3,wa:3,when:3,whether:3,whitnei:4,wilcoxon:4,within:3,work:3,x:3,y:3,yekuti:4},titles:["Welcome to statannotations\u2019s documentation!","statannotations","setup module","statannotations package","statannotations.stats package"],titleterms:{annot:3,comparisonscorrect:4,content:[0,3,4],document:0,format_annot:3,indic:0,modul:[2,3,4],packag:[3,4],pvalueformat:3,s:0,setup:2,stat:4,statannot:[0,1,3,4],statresult:4,stattest:4,submodul:[3,4],subpackag:3,tabl:0,test:4,util:[3,4],welcom:0}}) \ No newline at end of file diff --git a/docs/build/html/statannotations.html b/docs/build/html/statannotations.html index 170abcf..57267ef 100644 --- a/docs/build/html/statannotations.html +++ b/docs/build/html/statannotations.html @@ -303,6 +303,27 @@ <h2>Submodules<a class="headerlink" href="#submodules" title="Permalink to this </dd> </dl> </li> +<li><dl> +<dt><cite>correction_format</cite>: How to format the star notation on the plot when</dt><dd>the multiple comparisons correction method removes the significance +* <cite>default</cite>: a ‘ (ns)’ suffix is added, such as in printed output, +<blockquote> +<div>corresponds to “{star} ({suffix})” +</div></blockquote> +<ul class="simple"> +<li><dl class="simple"> +<dt><cite>replace</cite>: the original star value is replaced with ‘ns’</dt><dd>corresponds to “{suffix}” +</dd> +</dl> +</li> +<li><dl class="simple"> +<dt>a custom formatting string using “{star}” for the original</dt><dd>pvalue and ‘{suffix}’ for ‘ns’ +</dd> +</dl> +</li> +</ul> +</dd> +</dl> +</li> <li><cite>line_height</cite>: in axes fraction coordinates</li> <li><cite>line_offset</cite></li> <li><cite>line_offset_to_group</cite></li> @@ -533,6 +554,11 @@ <h2>Submodules<a class="headerlink" href="#submodules" title="Permalink to this config(**parameters)<a class="headerlink" href="#statannotations.PValueFormat.PValueFormat.config" title="Permalink to this definition">¶</a></dt> <dd></dd></dl> +<dl class="py property"> +<dt class="sig sig-object py" id="statannotations.PValueFormat.PValueFormat.correction_format"> +property correction_format<a class="headerlink" href="#statannotations.PValueFormat.PValueFormat.correction_format" title="Permalink to this definition">¶</a></dt> +<dd></dd></dl> + <dl class="py method"> <dt class="sig sig-object py" id="statannotations.PValueFormat.PValueFormat.format_data"> format_data(result)<a class="headerlink" href="#statannotations.PValueFormat.PValueFormat.format_data" title="Permalink to this definition">¶</a></dt> @@ -570,6 +596,11 @@ <h2>Submodules<a class="headerlink" href="#submodules" title="Permalink to this </dd></dl> +<dl class="py function"> +<dt class="sig sig-object py" id="statannotations.PValueFormat.get_corrected_star"> +statannotations.PValueFormat.get_corrected_star(star: str, res: <a class="reference internal" href="statannotations.stats.html#statannotations.stats.StatResult.StatResult" title="statannotations.stats.StatResult.StatResult">statannotations.stats.StatResult.StatResult</a>, correction_format='{star} ({suffix})') → str<a class="headerlink" href="#statannotations.PValueFormat.get_corrected_star" title="Permalink to this definition">¶</a></dt> +<dd></dd></dl> + <dl class="py function"> <dt class="sig sig-object py" id="statannotations.PValueFormat.sort_pvalue_thresholds"> statannotations.PValueFormat.sort_pvalue_thresholds(pvalue_thresholds)<a class="headerlink" href="#statannotations.PValueFormat.sort_pvalue_thresholds" title="Permalink to this definition">¶</a></dt> @@ -580,7 +611,7 @@ <h2>Submodules<a class="headerlink" href="#submodules" title="Permalink to this <h2>statannotations.format_annotations module<a class="headerlink" href="#module-statannotations.format_annotations" title="Permalink to this headline">¶</a></h2> <dl class="py function"> <dt class="sig sig-object py" id="statannotations.format_annotations.pval_annotation_text"> -statannotations.format_annotations.pval_annotation_text(result: Union[List[<a class="reference internal" href="statannotations.stats.html#statannotations.stats.StatResult.StatResult" title="statannotations.stats.StatResult.StatResult">statannotations.stats.StatResult.StatResult</a>], <a class="reference internal" href="statannotations.stats.html#statannotations.stats.StatResult.StatResult" title="statannotations.stats.StatResult.StatResult">statannotations.stats.StatResult.StatResult</a>], pvalue_thresholds: List) → Union[List[str], str]<a class="headerlink" href="#statannotations.format_annotations.pval_annotation_text" title="Permalink to this definition">¶</a></dt> +statannotations.format_annotations.pval_annotation_text(result: List[<a class="reference internal" href="statannotations.stats.html#statannotations.stats.StatResult.StatResult" title="statannotations.stats.StatResult.StatResult">statannotations.stats.StatResult.StatResult</a>], pvalue_thresholds: List) → List[tuple]<a class="headerlink" href="#statannotations.format_annotations.pval_annotation_text" title="Permalink to this definition">¶</a></dt> <dd><dl class="field-list simple"> <dt class="field-odd">Parameters</dt> <dd class="field-odd"><ul class="simple"> @@ -597,7 +628,7 @@ <h2>Submodules<a class="headerlink" href="#submodules" title="Permalink to this <dl class="py function"> <dt class="sig sig-object py" id="statannotations.format_annotations.simple_text"> -statannotations.format_annotations.simple_text(result: <a class="reference internal" href="statannotations.stats.html#statannotations.stats.StatResult.StatResult" title="statannotations.stats.StatResult.StatResult">statannotations.stats.StatResult.StatResult</a>, pvalue_format, pvalue_thresholds) → str<a class="headerlink" href="#statannotations.format_annotations.simple_text" title="Permalink to this definition">¶</a></dt> +statannotations.format_annotations.simple_text(result: <a class="reference internal" href="statannotations.stats.html#statannotations.stats.StatResult.StatResult" title="statannotations.stats.StatResult.StatResult">statannotations.stats.StatResult.StatResult</a>, pvalue_format, pvalue_thresholds, short_test_name=True) → str<a class="headerlink" href="#statannotations.format_annotations.simple_text" title="Permalink to this definition">¶</a></dt> <dd>Generates simple text for test name and pvalue. <dl class="field-list simple"> <dt class="field-odd">Parameters</dt> @@ -605,7 +636,7 @@ <h2>Submodules<a class="headerlink" href="#submodules" title="Permalink to this <li>result – StatResult instance</li> <li>pvalue_format – format string for pvalue</li> <li>pvalue_thresholds – String to display per pvalue range</li> -<li>result – StatResult</li> +<li>short_test_name – whether to display the test (short) name</li> </ul> </dd> <dt class="field-even">Returns</dt> diff --git a/docs/build/html/statannotations.stats.html b/docs/build/html/statannotations.stats.html index 099ec99..2b11749 100644 --- a/docs/build/html/statannotations.stats.html +++ b/docs/build/html/statannotations.stats.html @@ -221,6 +221,11 @@ <h2>Submodules<a class="headerlink" href="#submodules" title="Permalink to this <dt class="sig sig-object py" id="statannotations.stats.StatResult.StatResult"> class statannotations.stats.StatResult.StatResult(test_description, test_short_name, stat_str, stat, pval, alpha=0.05)<a class="headerlink" href="#statannotations.stats.StatResult.StatResult" title="Permalink to this definition">¶</a></dt> <dd>Bases: <code class="xref py py-class docutils literal notranslate">object</code> +<dl class="py method"> +<dt class="sig sig-object py" id="statannotations.stats.StatResult.StatResult.adjust"> +adjust(stat_summary)<a class="headerlink" href="#statannotations.stats.StatResult.StatResult.adjust" title="Permalink to this definition">¶</a></dt> +<dd></dd></dl> + <dl class="py property"> <dt class="sig sig-object py" id="statannotations.stats.StatResult.StatResult.corrected_significance"> property corrected_significance<a class="headerlink" href="#statannotations.stats.StatResult.StatResult.corrected_significance" title="Permalink to this definition">¶</a></dt> diff --git a/statannotations/Annotator.py b/statannotations/Annotator.py index a8e5511..372dcbf 100644 --- a/statannotations/Annotator.py +++ b/statannotations/Annotator.py @@ -117,7 +117,7 @@ def __init__(self, ax, pairs, plot='boxplot', data=None, x=None, self._comparisons_correction = None self._loc = "inside" - self._verbose = 1 + self._verbose = verbose self._just_configured = True self.show_test_name = True self.use_fixed_offset = False @@ -253,6 +253,14 @@ def configure(self, **parameters): * `comparisons_correction`: Method for multiple comparisons correction. One of `statsmodels` `multipletests` methods (w/ default FWER), or a `ComparisonsCorrection` instance. + * `correction_format`: How to format the star notation on the plot when + the multiple comparisons correction method removes the significance + * `default`: a ' (ns)' suffix is added, such as in printed output, + corresponds to "{star} ({suffix})" + * `replace`: the original star value is replaced with 'ns' + corresponds to "{suffix}" + * a custom formatting string using "{star}" for the original + pvalue and '{suffix}' for 'ns' * `line_height`: in axes fraction coordinates * `line_offset` * `line_offset_to_group` diff --git a/statannotations/PValueFormat.py b/statannotations/PValueFormat.py index 598984b..2f35b4d 100644 --- a/statannotations/PValueFormat.py +++ b/statannotations/PValueFormat.py @@ -1,14 +1,18 @@ +from statannotations.stats.StatResult import StatResult + from statannotations.format_annotations import pval_annotation_text, \ simple_text from statannotations.utils import DEFAULT, check_valid_text_format, \ InvalidParametersError CONFIGURABLE_PARAMETERS = [ + 'correction_format', 'fontsize', 'pvalue_format_string', 'simple_format_string', 'text_format', - 'pvalue_thresholds' + 'pvalue_thresholds', + 'show_test_name' ] @@ -32,6 +36,8 @@ def __init__(self): self.fontsize = 'medium' self._default_pvalue_thresholds = True self._pvalue_thresholds = self._get_pvalue_thresholds(DEFAULT) + self._correction_format = "{star} ({suffix})" + self.show_test_name = True def config(self, **parameters): @@ -91,6 +97,17 @@ def pvalue_format_string(self, pvalue_format_string): self._get_pvalue_and_simple_formats(pvalue_format_string) ) + @property + def correction_format(self): + return self._correction_format + + @correction_format.setter + def correction_format(self, correction_format: str): + self._correction_format = { + "replace": "{suffix}", + "default": "{star} ({suffix})", + }.get(correction_format, correction_format) + @property def pvalue_thresholds(self): return self._pvalue_thresholds @@ -156,17 +173,32 @@ def _update_pvalue_thresholds(self): def format_data(self, result): if self.text_format == 'full': - return ("{} p = {}{}" + text = (f"{result.test_short_name} " if self.show_test_name + else "") + + return ("{}p = {}{}" .format('{}', self.pvalue_format_string, '{}') - .format(result.test_short_name, result.pvalue, - result.significance_suffix)) + .format(text, result.pvalue, result.significance_suffix)) elif self.text_format == 'star': - return pval_annotation_text(result, self.pvalue_thresholds) + was_list = False + + if not isinstance(result, list): + result = [result] + + annotations = [ + get_corrected_star(star, res, self._correction_format) + for star, res + in pval_annotation_text(result, self.pvalue_thresholds)] + + if was_list: + return annotations + + return annotations[0] elif self.text_format == 'simple': return simple_text(result, self.simple_format_string, - self.pvalue_thresholds) + self.pvalue_thresholds, self.show_test_name) def get_configuration(self): return {key: getattr(self, key) for key in CONFIGURABLE_PARAMETERS} @@ -175,3 +207,11 @@ def get_configuration(self): def sort_pvalue_thresholds(pvalue_thresholds): return sorted(pvalue_thresholds, key=lambda threshold_notation: threshold_notation[0]) + + +def get_corrected_star(star: str, res: StatResult, + correction_format="{star} ({suffix})") -> str: + if res.significance_suffix: + return correction_format.format(star=star, + suffix=res.significance_suffix) + return star diff --git a/statannotations/_version.py b/statannotations/_version.py index df12433..f6b7e26 100644 --- a/statannotations/_version.py +++ b/statannotations/_version.py @@ -1,1 +1,1 @@ -__version__ = "0.4.2" +__version__ = "0.4.3" diff --git a/statannotations/format_annotations.py b/statannotations/format_annotations.py index 81bb6e8..ef87581 100644 --- a/statannotations/format_annotations.py +++ b/statannotations/format_annotations.py @@ -5,8 +5,8 @@ from operator import itemgetter -def pval_annotation_text(result: Union[List[StatResult], StatResult], - pvalue_thresholds: List) -> Union[List[str], str]: +def pval_annotation_text(result: List[StatResult], + pvalue_thresholds: List) -> List[tuple]: """ :param result: StatResult instance or list thereof @@ -14,11 +14,6 @@ def pval_annotation_text(result: Union[List[StatResult], StatResult], :returns: A List of rendered annotations if a list of StatResults was provided, a string otherwise. """ - was_list = True - - if not isinstance(result, list): - was_list = False - result = [result] x1_pval = np.array([res.pvalue for res in result]) @@ -34,27 +29,27 @@ def pval_annotation_text(result: Union[List[StatResult], StatResult], x_annot[x1_pval <= pvalue_thresholds[-1][0]] = pvalue_thresholds[-1][1] - x_annot = [f"{star}{res.significance_suffix}" - for star, res in zip(x_annot, result)] - - return x_annot if was_list else x_annot[0] + return [(star, res) for star, res in zip(x_annot, result)] -def simple_text(result: StatResult, pvalue_format, pvalue_thresholds) -> str: +def simple_text(result: StatResult, pvalue_format, pvalue_thresholds, + short_test_name=True) -> str: """ Generates simple text for test name and pvalue. :param result: StatResult instance :param pvalue_format: format string for pvalue :param pvalue_thresholds: String to display per pvalue range - :param result: StatResult + :param short_test_name: whether to display the test (short) name :returns: simple annotation """ # Sort thresholds thresholds = sorted(pvalue_thresholds, key=lambda x: x[0]) - text = result.test_short_name and f"{result.test_short_name} " or "" + text = (f"{result.test_short_name} " + if short_test_name and result.test_short_name + else "") for threshold in thresholds: if result.pvalue < threshold[0]: @@ -63,4 +58,4 @@ def simple_text(result: StatResult, pvalue_format, pvalue_thresholds) -> str: else: pval_text = "p = {}".format(pvalue_format).format(result.pvalue) - return text + pval_text + result.significance_suffix + return result.adjust(text + pval_text) diff --git a/statannotations/stats/StatResult.py b/statannotations/stats/StatResult.py index 45f3c2f..a71c0e6 100644 --- a/statannotations/stats/StatResult.py +++ b/statannotations/stats/StatResult.py @@ -44,7 +44,7 @@ def formatted_output(self): stat_summary = '{}, P_val:{:.3e}'.format(description, self.pvalue) - stat_summary += self.significance_suffix + stat_summary = self.adjust(stat_summary) if self.stat_str is not None or self.stat_value is not None: stat_summary += ' {}={:.3e}'.format(self.stat_str, self.stat_value) @@ -55,8 +55,13 @@ def formatted_output(self): def significance_suffix(self): # will add this only if a correction method is specified if self._corrected_significance is False and self.pvalue <= self.alpha: - return ' (ns)' + return 'ns' return "" def __str__(self): return self.formatted_output + + def adjust(self, stat_summary): + if self.significance_suffix: + return f"{stat_summary} ({self.significance_suffix})" + return stat_summary diff --git a/usage/example.ipynb b/usage/example.ipynb index 91a3e66..07fe9f3 100644 --- a/usage/example.ipynb +++ b/usage/example.ipynb @@ -231,7 +231,7 @@ }, { "data": { - "text/plain": "(<matplotlib.axes._subplots.AxesSubplot at 0x7fc28184f6a0>,\n [<statannotations.Annotation.Annotation at 0x7fc28164aeb8>,\n <statannotations.Annotation.Annotation at 0x7fc28183d6d8>,\n <statannotations.Annotation.Annotation at 0x7fc282ada5c0>])" + "text/plain": "(<matplotlib.axes._subplots.AxesSubplot at 0x7fc8f09eb6d8>,\n [<statannotations.Annotation.Annotation at 0x7fc8ef839c50>,\n <statannotations.Annotation.Annotation at 0x7fc8ef84fe80>,\n <statannotations.Annotation.Annotation at 0x7fc8ef86d940>])" }, "execution_count": 6, "metadata": {}, @@ -247,6 +247,68 @@ } ] }, + { + "cell_type": "markdown", + "source": [ + "To have `\"ns\"` instead of `\"* (ns)\"`, configure with `correction_format=\"replace\"`." + ], + "metadata": { + "collapsed": false, + "pycharm": { + "name": "#%% md\n" + } + } + }, + { + "cell_type": "code", + "execution_count": 7, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "p-value annotation legend:\n", + " ns: p <= 1.00e+00\n", + " *: 1.00e-02 < p <= 5.00e-02\n", + " **: 1.00e-03 < p <= 1.00e-02\n", + " ***: 1.00e-04 < p <= 1.00e-03\n", + " ****: p <= 1.00e-04\n", + "\n", + "Thur vs. Fri: Mann-Whitney-Wilcoxon test two-sided with Benjamini-Hochberg correction, P_val:6.477e-01 U_stat=6.305e+02\n", + "Thur vs. Sat: Mann-Whitney-Wilcoxon test two-sided with Benjamini-Hochberg correction, P_val:4.690e-02 (ns) U_stat=2.180e+03\n", + "Sun vs. Fri: Mann-Whitney-Wilcoxon test two-sided with Benjamini-Hochberg correction, P_val:2.680e-02 (ns) U_stat=9.605e+02\n" + ] + }, + { + "data": { + "text/plain": "(<matplotlib.axes._subplots.AxesSubplot at 0x7fc8f0c539b0>,\n [<statannotations.Annotation.Annotation at 0x7fc8f0ad47f0>,\n <statannotations.Annotation.Annotation at 0x7fc8f0ad4978>,\n <statannotations.Annotation.Annotation at 0x7fc8f0ad4be0>])" + }, + "execution_count": 7, + "metadata": {}, + "output_type": "execute_result" + }, + { + "data": { + "text/plain": "<Figure size 432x288 with 1 Axes>", + "image/png": "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\n" + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "ax = sns.boxplot(data=df, x=x, y=y, order=order)\n", + "annot.new_plot(ax, data=df, x=x, y=y, order=order)\n", + "annot.configure(comparisons_correction=\"BH\", correction_format=\"replace\")\n", + "annot.apply_and_annotate()" + ], + "metadata": { + "collapsed": false, + "pycharm": { + "name": "#%%\n" + } + } + }, { "cell_type": "markdown", "metadata": {}, @@ -263,7 +325,7 @@ }, { "cell_type": "code", - "execution_count": 7, + "execution_count": 8, "metadata": {}, "outputs": [ { @@ -320,7 +382,7 @@ }, { "cell_type": "code", - "execution_count": 8, + "execution_count": 9, "metadata": {}, "outputs": [ { @@ -328,7 +390,7 @@ "text/plain": " carat cut color clarity depth table price x y z\n0 0.23 Ideal E SI2 61.5 55.0 326 3.95 3.98 2.43\n1 0.21 Premium E SI1 59.8 61.0 326 3.89 3.84 2.31\n2 0.23 Good E VS1 56.9 65.0 327 4.05 4.07 2.31\n3 0.29 Premium I VS2 62.4 58.0 334 4.20 4.23 2.63\n4 0.31 Good J SI2 63.3 58.0 335 4.34 4.35 2.75", "text/html": "<div>\n<style scoped>\n .dataframe tbody tr th:only-of-type {\n vertical-align: middle;\n }\n\n .dataframe tbody tr th {\n vertical-align: top;\n }\n\n .dataframe thead th {\n text-align: right;\n }\n</style>\n<table border=\"1\" class=\"dataframe\">\n <thead>\n <tr style=\"text-align: right;\">\n <th></th>\n <th>carat</th>\n <th>cut</th>\n <th>color</th>\n <th>clarity</th>\n <th>depth</th>\n <th>table</th>\n <th>price</th>\n <th>x</th>\n <th>y</th>\n <th>z</th>\n </tr>\n </thead>\n <tbody>\n <tr>\n <th>0</th>\n <td>0.23</td>\n <td>Ideal</td>\n <td>E</td>\n <td>SI2</td>\n <td>61.5</td>\n <td>55.0</td>\n <td>326</td>\n <td>3.95</td>\n <td>3.98</td>\n <td>2.43</td>\n </tr>\n <tr>\n <th>1</th>\n <td>0.21</td>\n <td>Premium</td>\n <td>E</td>\n <td>SI1</td>\n <td>59.8</td>\n <td>61.0</td>\n <td>326</td>\n <td>3.89</td>\n <td>3.84</td>\n <td>2.31</td>\n </tr>\n <tr>\n <th>2</th>\n <td>0.23</td>\n <td>Good</td>\n <td>E</td>\n <td>VS1</td>\n <td>56.9</td>\n <td>65.0</td>\n <td>327</td>\n <td>4.05</td>\n <td>4.07</td>\n <td>2.31</td>\n </tr>\n <tr>\n <th>3</th>\n <td>0.29</td>\n <td>Premium</td>\n <td>I</td>\n <td>VS2</td>\n <td>62.4</td>\n <td>58.0</td>\n <td>334</td>\n <td>4.20</td>\n <td>4.23</td>\n <td>2.63</td>\n </tr>\n <tr>\n <th>4</th>\n <td>0.31</td>\n <td>Good</td>\n <td>J</td>\n <td>SI2</td>\n <td>63.3</td>\n <td>58.0</td>\n <td>335</td>\n <td>4.34</td>\n <td>4.35</td>\n <td>2.75</td>\n </tr>\n </tbody>\n</table>\n</div>" }, - "execution_count": 8, + "execution_count": 9, "metadata": {}, "output_type": "execute_result" } @@ -346,7 +408,7 @@ }, { "cell_type": "code", - "execution_count": 9, + "execution_count": 10, "outputs": [ { "name": "stdout", @@ -420,7 +482,7 @@ }, { "cell_type": "code", - "execution_count": 10, + "execution_count": 11, "metadata": {}, "outputs": [ { @@ -428,7 +490,7 @@ "text/plain": " total_bill tip sex smoker day time size tip_bucket\n0 16.99 1.01 Female No Sun Dinner 2 (0.991, 4.0]\n1 10.34 1.66 Male No Sun Dinner 3 (0.991, 4.0]\n2 21.01 3.50 Male No Sun Dinner 3 (0.991, 4.0]\n3 23.68 3.31 Male No Sun Dinner 2 (0.991, 4.0]\n4 24.59 3.61 Female No Sun Dinner 4 (0.991, 4.0]", "text/html": "<div>\n<style scoped>\n .dataframe tbody tr th:only-of-type {\n vertical-align: middle;\n }\n\n .dataframe tbody tr th {\n vertical-align: top;\n }\n\n .dataframe thead th {\n text-align: right;\n }\n</style>\n<table border=\"1\" class=\"dataframe\">\n <thead>\n <tr style=\"text-align: right;\">\n <th></th>\n <th>total_bill</th>\n <th>tip</th>\n <th>sex</th>\n <th>smoker</th>\n <th>day</th>\n <th>time</th>\n <th>size</th>\n <th>tip_bucket</th>\n </tr>\n </thead>\n <tbody>\n <tr>\n <th>0</th>\n <td>16.99</td>\n <td>1.01</td>\n <td>Female</td>\n <td>No</td>\n <td>Sun</td>\n <td>Dinner</td>\n <td>2</td>\n <td>(0.991, 4.0]</td>\n </tr>\n <tr>\n <th>1</th>\n <td>10.34</td>\n <td>1.66</td>\n <td>Male</td>\n <td>No</td>\n <td>Sun</td>\n <td>Dinner</td>\n <td>3</td>\n <td>(0.991, 4.0]</td>\n </tr>\n <tr>\n <th>2</th>\n <td>21.01</td>\n <td>3.50</td>\n <td>Male</td>\n <td>No</td>\n <td>Sun</td>\n <td>Dinner</td>\n <td>3</td>\n <td>(0.991, 4.0]</td>\n </tr>\n <tr>\n <th>3</th>\n <td>23.68</td>\n <td>3.31</td>\n <td>Male</td>\n <td>No</td>\n <td>Sun</td>\n <td>Dinner</td>\n <td>2</td>\n <td>(0.991, 4.0]</td>\n </tr>\n <tr>\n <th>4</th>\n <td>24.59</td>\n <td>3.61</td>\n <td>Female</td>\n <td>No</td>\n <td>Sun</td>\n <td>Dinner</td>\n <td>4</td>\n <td>(0.991, 4.0]</td>\n </tr>\n </tbody>\n</table>\n</div>" }, - "execution_count": 10, + "execution_count": 11, "metadata": {}, "output_type": "execute_result" } @@ -441,7 +503,7 @@ }, { "cell_type": "code", - "execution_count": 11, + "execution_count": 12, "metadata": { "scrolled": true }, @@ -450,7 +512,7 @@ "data": { "text/plain": "[(0.991, 4.0], (4.0, 7.0], (7.0, 10.0]]\nCategories (3, interval[float64]): [(0.991, 4.0] < (4.0, 7.0] < (7.0, 10.0]]" }, - "execution_count": 11, + "execution_count": 12, "metadata": {}, "output_type": "execute_result" } @@ -463,7 +525,7 @@ }, { "cell_type": "code", - "execution_count": 12, + "execution_count": 13, "metadata": { "scrolled": false }, @@ -526,7 +588,7 @@ }, { "cell_type": "code", - "execution_count": 13, + "execution_count": 14, "metadata": {}, "outputs": [ { @@ -599,7 +661,7 @@ }, { "cell_type": "code", - "execution_count": 14, + "execution_count": 15, "metadata": {}, "outputs": [ { @@ -652,7 +714,7 @@ }, { "cell_type": "code", - "execution_count": 15, + "execution_count": 16, "outputs": [ { "name": "stdout", @@ -672,9 +734,9 @@ }, { "data": { - "text/plain": "(<matplotlib.axes._subplots.AxesSubplot at 0x7fc284d10f60>,\n [<statannotations.Annotation.Annotation at 0x7fc281c837b8>,\n <statannotations.Annotation.Annotation at 0x7fc2817bfc18>,\n <statannotations.Annotation.Annotation at 0x7fc284d8e518>])" + "text/plain": "(<matplotlib.axes._subplots.AxesSubplot at 0x7fc8f230e160>,\n [<statannotations.Annotation.Annotation at 0x7fc8f0f07e10>,\n <statannotations.Annotation.Annotation at 0x7fc8f234cfd0>,\n <statannotations.Annotation.Annotation at 0x7fc8efa6e048>])" }, - "execution_count": 15, + "execution_count": 16, "metadata": {}, "output_type": "execute_result" }, @@ -729,7 +791,7 @@ }, { "cell_type": "code", - "execution_count": 16, + "execution_count": 17, "metadata": { "scrolled": true }, @@ -762,7 +824,7 @@ }, { "cell_type": "code", - "execution_count": 17, + "execution_count": 18, "metadata": { "scrolled": false }, @@ -814,7 +876,7 @@ }, { "cell_type": "code", - "execution_count": 18, + "execution_count": 19, "outputs": [ { "name": "stdout", @@ -867,7 +929,7 @@ }, { "cell_type": "code", - "execution_count": 19, + "execution_count": 20, "metadata": { "scrolled": false, "pycharm": { @@ -926,7 +988,7 @@ }, { "cell_type": "code", - "execution_count": 20, + "execution_count": 21, "outputs": [ { "name": "stdout", @@ -947,7 +1009,7 @@ { "data": { "text/plain": "<Figure size 864x432 with 1 Axes>", - "image/png": "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\n" + "image/png": "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\n" }, "metadata": {}, "output_type": "display_data" @@ -1009,7 +1071,7 @@ }, { "cell_type": "code", - "execution_count": 21, + "execution_count": 22, "outputs": [ { "name": "stdout", @@ -1029,16 +1091,16 @@ }, { "data": { - "text/plain": "(<matplotlib.axes._subplots.AxesSubplot at 0x7fc283af2668>,\n [<statannotations.Annotation.Annotation at 0x7fc283bf3780>,\n <statannotations.Annotation.Annotation at 0x7fc283be1dd8>,\n <statannotations.Annotation.Annotation at 0x7fc283bc3ef0>])" + "text/plain": "(<matplotlib.axes._subplots.AxesSubplot at 0x7fc8f23c5518>,\n [<statannotations.Annotation.Annotation at 0x7fc8efb92748>,\n <statannotations.Annotation.Annotation at 0x7fc8efb92ac8>,\n <statannotations.Annotation.Annotation at 0x7fc8efe8ef28>])" }, - "execution_count": 21, + "execution_count": 22, "metadata": {}, "output_type": "execute_result" }, { "data": { "text/plain": "<Figure size 864x432 with 1 Axes>", - "image/png": "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\n" + "image/png": "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\n" }, "metadata": {}, "output_type": "display_data" @@ -1077,7 +1139,7 @@ }, { "cell_type": "code", - "execution_count": 22, + "execution_count": 23, "outputs": [ { "name": "stdout", @@ -1128,7 +1190,7 @@ }, { "cell_type": "code", - "execution_count": 23, + "execution_count": 24, "outputs": [ { "name": "stdout", @@ -1169,7 +1231,7 @@ }, { "cell_type": "code", - "execution_count": 23, + "execution_count": 24, "outputs": [], "source": [], "metadata": {

Set verbose flag from parameters

2021-10-17T15:44:28

0.0

[]

trevismd/statannotations

trevismd__statannotations-28

284ba0a621bde694a85320d976df93a8d15d899d

diff --git a/CHANGELOG.md b/CHANGELOG.md index 80407ff..9c2425b 100644 --- a/CHANGELOG.md +++ b/CHANGELOG.md @@ -1,4 +1,7 @@ ## v0.4 +### v0.4.1 + - Support for horizontal orientation + ### v0.4.0 - Major refactoring, change to an Annotator `class` to prepare and add annotations in separate function (now method) calls. diff --git a/README.md b/README.md index 7e9b432..fff5153 100644 --- a/README.md +++ b/README.md @@ -42,6 +42,7 @@ corresponding branch). - **Interface to use any other function from any source with minimal extra code** - Smart layout of multiple annotations with correct y offsets. +- **Support for vertical and horizontal orientation** - Annotations can be located inside or outside the plot. - **Corrections for multiple testing can be applied (binding to `statsmodels.stats.multitest.multipletests` methods):** @@ -51,7 +52,7 @@ corresponding branch). - Benjamini-Yekutieli - **And any other function from any source with minimal extra code** - Format of the statistical test annotation can be customized: - star annotation, simplified p-value, or explicit p-value. + star annotation, simplified p-value format, or explicit p-value. - Optionally, custom p-values can be given as input. In this case, no statistical test is performed, but **corrections for multiple testing can be applied.** diff --git a/codecov.yml b/codecov.yml index 0d54c1d..4cc71ba 100644 --- a/codecov.yml +++ b/codecov.yml @@ -6,4 +6,4 @@ coverage: patch: default: - threshold: 0.5% \ No newline at end of file + threshold: 2% diff --git a/docs/build/doctrees/environment.pickle b/docs/build/doctrees/environment.pickle index 771aba2..169ad27 100644 Binary files a/docs/build/doctrees/environment.pickle and b/docs/build/doctrees/environment.pickle differ diff --git a/docs/build/doctrees/statannotations.doctree b/docs/build/doctrees/statannotations.doctree index ee0691c..759af7f 100644 Binary files a/docs/build/doctrees/statannotations.doctree and b/docs/build/doctrees/statannotations.doctree differ diff --git a/docs/build/html/genindex.html b/docs/build/html/genindex.html index 935f030..e1e1951 100644 --- a/docs/build/html/genindex.html +++ b/docs/build/html/genindex.html @@ -166,6 +166,7 @@ <h1 id="index">Index</h1> | <a href="#L">L</a> | <a href="#M">M</a> | <a href="#N">N</a> + | <a href="#O">O</a> | <a href="#P">P</a> | <a href="#R">R</a> | <a href="#S">S</a> @@ -365,6 +366,14 @@ <h2 id="N">N</h2> </ul></td> </tr></table> +<h2 id="O">O</h2> +<table style="width: 100%" class="indextable genindextable"><tr> + <td style="width: 33%; vertical-align: top;"><ul> + <li><a href="statannotations.html#statannotations.Annotator.Annotator.orient">orient (statannotations.Annotator.Annotator property)</a> +</li> + </ul></td> +</tr></table> + <h2 id="P">P</h2> <table style="width: 100%" class="indextable genindextable"><tr> <td style="width: 33%; vertical-align: top;"><ul> diff --git a/docs/build/html/objects.inv b/docs/build/html/objects.inv index 0d82dd7..b77fb8b 100644 Binary files a/docs/build/html/objects.inv and b/docs/build/html/objects.inv differ diff --git a/docs/build/html/searchindex.js b/docs/build/html/searchindex.js index 38da9ed..2e1b48e 100644 --- a/docs/build/html/searchindex.js +++ b/docs/build/html/searchindex.js @@ -1,1 +1,1 @@ -Search.setIndex({docnames:["index","modules","setup","statannotations","statannotations.stats"],envversion:{"sphinx.domains.c":2,"sphinx.domains.changeset":1,"sphinx.domains.citation":1,"sphinx.domains.cpp":3,"sphinx.domains.index":1,"sphinx.domains.javascript":2,"sphinx.domains.math":2,"sphinx.domains.python":3,"sphinx.domains.rst":2,"sphinx.domains.std":2,sphinx:56},filenames:["index.rst","modules.rst","setup.rst","statannotations.rst","statannotations.stats.rst"],objects:{"":{statannotations:[3,0,0,"-"]},"statannotations.Annotation":{Annotation:[3,1,1,""]},"statannotations.Annotation.Annotation":{formatted_output:[3,2,1,""],print_labels_and_content:[3,3,1,""],text:[3,2,1,""]},"statannotations.Annotator":{Annotator:[3,1,1,""]},"statannotations.Annotator.Annotator":{alpha:[3,2,1,""],annotate:[3,3,1,""],annotate_custom_annotations:[3,3,1,""],apply_and_annotate:[3,3,1,""],apply_test:[3,3,1,""],comparisons_correction:[3,2,1,""],configure:[3,3,1,""],fig:[3,2,1,""],get_annotations_text:[3,3,1,""],get_configuration:[3,3,1,""],get_offset_func:[3,3,1,""],has_type0_comparisons_correction:[3,3,1,""],loc:[3,2,1,""],new_plot:[3,3,1,""],print_pvalue_legend:[3,3,1,""],pvalue_format:[3,2,1,""],reset_configuration:[3,3,1,""],set_custom_annotations:[3,3,1,""],set_pvalues:[3,3,1,""],set_pvalues_and_annotate:[3,3,1,""],test:[3,2,1,""],validate_test_short_name:[3,3,1,""],verbose:[3,2,1,""]},"statannotations.PValueFormat":{Formatter:[3,1,1,""],PValueFormat:[3,1,1,""],sort_pvalue_thresholds:[3,4,1,""]},"statannotations.PValueFormat.Formatter":{config:[3,3,1,""],format_data:[3,3,1,""]},"statannotations.PValueFormat.PValueFormat":{config:[3,3,1,""],format_data:[3,3,1,""],get_configuration:[3,3,1,""],print_legend_if_used:[3,3,1,""],pvalue_format_string:[3,2,1,""],pvalue_thresholds:[3,2,1,""],simple_format_string:[3,2,1,""],text_format:[3,2,1,""]},"statannotations.format_annotations":{pval_annotation_text:[3,4,1,""],simple_text:[3,4,1,""]},"statannotations.stats":{ComparisonsCorrection:[4,0,0,"-"],StatResult:[4,0,0,"-"],StatTest:[4,0,0,"-"],test:[4,0,0,"-"],utils:[4,0,0,"-"]},"statannotations.stats.ComparisonsCorrection":{ComparisonsCorrection:[4,1,1,""],check_valid_correction_name:[4,4,1,""],get_correction_parameters:[4,4,1,""],get_validated_comparisons_correction:[4,4,1,""]},"statannotations.stats.ComparisonsCorrection.ComparisonsCorrection":{apply:[4,3,1,""],document:[4,3,1,""]},"statannotations.stats.StatResult":{StatResult:[4,1,1,""]},"statannotations.stats.StatResult.StatResult":{corrected_significance:[4,2,1,""],correction_method:[4,2,1,""],formatted_output:[4,2,1,""],significance_suffix:[4,2,1,""]},"statannotations.stats.StatTest":{StatTest:[4,1,1,""],wilcoxon:[4,4,1,""]},"statannotations.stats.StatTest.StatTest":{from_library:[4,3,1,""],short_name:[4,2,1,""]},"statannotations.stats.test":{apply_test:[4,4,1,""]},"statannotations.stats.utils":{check_alpha:[4,4,1,""],check_num_comparisons:[4,4,1,""],check_pvalues:[4,4,1,""],get_num_comparisons:[4,4,1,""],return_results:[4,4,1,""]},"statannotations.utils":{check_is_in:[3,4,1,""],check_not_none:[3,4,1,""],check_order_in_data:[3,4,1,""],check_pairs_in_data:[3,4,1,""],check_valid_text_format:[3,4,1,""],get_closest:[3,4,1,""],get_x_values:[3,4,1,""],raise_expected_got:[3,4,1,""],remove_null:[3,4,1,""],render_collection:[3,4,1,""]},statannotations:{Annotation:[3,0,0,"-"],Annotator:[3,0,0,"-"],PValueFormat:[3,0,0,"-"],format_annotations:[3,0,0,"-"],stats:[4,0,0,"-"],utils:[3,0,0,"-"]}},objnames:{"0":["py","module","Python module"],"1":["py","class","Python class"],"2":["py","property","Python property"],"3":["py","method","Python method"],"4":["py","function","Python function"]},objtypes:{"0":"py:module","1":"py:class","2":"py:property","3":"py:method","4":"py:function"},terms:{"0":[3,4],"05":[3,4],"1":[3,4],"12141511":3,"9981846":3,"case":4,"class":[3,4],"default":[3,4],"float":4,"function":[3,4],"int":4,"return":3,"short":3,"static":4,"true":[3,4],A:3,If:3,Or:3,The:3,a_list:3,accept:3,add:3,addit:4,alpha:[3,4],an:3,annot:1,annotate_custom_annot:3,appli:[3,4],apply_and_annot:3,apply_test:[3,4],ar:3,arg:[3,4],argument:[3,4],arrai:3,assum:3,auto:3,ax:3,bar:3,base:[3,4],behavior:4,benjamini:4,between:3,bonferroni:4,bool:4,box:3,boxplot:3,calcul:3,callabl:4,check:3,check_alpha:4,check_is_in:3,check_not_non:3,check_num_comparison:4,check_order_in_data:3,check_pairs_in_data:3,check_pvalu:4,check_valid_correction_nam:4,check_valid_text_format:3,classmethod:3,close:3,closest:3,collect:3,color:3,com:3,comparison:[3,4],comparisons_correct:[3,4],comparisonscorrect:[1,3],comput:3,config:3,configur:3,constructor:3,content:1,coordin:3,core:3,corr_kwarg:4,correct:4,corrected_signific:4,correction_method:4,custom:3,data:[3,4],datafram:3,dict:4,displai:3,document:4,each:3,earlier:4,either:4,engin:3,equal:3,equival:4,error:3,error_typ:3,exact:3,exist:3,expect:3,fals:3,fig:3,follow:3,for_:3,format:[3,4],format_annot:1,format_data:3,formatt:3,formatted_output:[3,4],fraction:3,frame:3,from:[3,4],from_librari:4,func:4,gener:3,get:4,get_annotations_text:3,get_closest:3,get_configur:3,get_correction_paramet:4,get_num_comparison:4,get_offset_func:3,get_validated_comparisons_correct:4,get_x_valu:3,got:3,group:3,group_data1:4,group_data2:4,gt:4,has_type0_comparisons_correct:3,hochberg:4,hold:3,holm:4,http:3,hue:3,hue_ord:3,index:0,instanc:[3,4],instead:3,interfac:4,interpret:4,keyword:4,kruskal:4,kwarg:[3,4],label:3,leven:4,line_height:3,line_offset:3,line_offset_to_group:3,line_width:3,link:3,list:3,loc:3,ls:4,mann:4,messag:3,method:4,method_typ:4,mode:3,modul:[0,1],multipl:4,must:3,mylist:3,mynumb:3,name:[3,4],new_plot:3,none:[3,4],num_comparison:[3,4],number:[3,4],object:[3,4],one:[3,4],option:[3,4],order:3,otherwis:3,overrid:3,p:3,p_valu:4,packag:[0,1],page:0,pair:3,panda:3,paramet:[3,4],pass:[3,4],per:3,perform:3,plot:3,plot_param:3,point:3,posit:3,print_labels_and_cont:3,print_legend_if_us:3,print_pvalue_legend:3,properti:[3,4],provid:[3,4],pval:4,pval_annotation_text:3,pvalu:[3,4],pvalue_format:3,pvalue_format_str:3,pvalue_threshold:3,pvalueformat:1,rais:3,raise_expected_got:3,rang:3,refer:3,remove_nul:3,render:3,render_collect:3,reset_configur:3,result:[3,4],results_arrai:4,return_result:4,run:4,same:3,sampl:4,seaborn:3,search:0,sep:3,seri:3,set:3,set_custom_annot:3,set_pvalu:3,set_pvalues_and_annot:3,short_nam:4,show:3,show_test_nam:3,significance_suffix:4,simpl:3,simple_format_str:3,simple_text:3,smallest:3,sort:3,sort_pvalue_threshold:3,stackoverflow:3,standard:3,stat:[1,3],stat_nam:4,stat_str:4,statist:[3,4],statresult:[1,3],stats_param:[3,4],statsmodel:4,statsmodels_api:4,stattest:[1,3],str:[3,4],string:3,struct:3,submodul:1,subpackag:1,t:4,test:[1,3],test_descript:4,test_ind:4,test_long_nam:4,test_nam:4,test_pair:4,test_result_list:4,test_short_nam:[3,4],test_welch:4,text:3,text_annot_custom:3,text_format:3,text_offset:3,them:3,thereof:3,thi:[3,4],three:3,threshold:3,top:3,tupl:3,two:[3,4],type:3,union:[3,4],us:[3,4],use_fixed_offset:3,util:1,valid_valu:3,validate_test_short_nam:3,valu:[3,4],valueerror:3,verbos:[3,4],version:4,vs:3,wa:3,whitnei:4,wilcoxon:4,work:3,x:3,y:3,yekuti:4},titles:["Welcome to statannotations\u2019s documentation!","statannotations","setup module","statannotations package","statannotations.stats package"],titleterms:{annot:3,comparisonscorrect:4,content:[0,3,4],document:0,format_annot:3,indic:0,modul:[2,3,4],packag:[3,4],pvalueformat:3,s:0,setup:2,stat:4,statannot:[0,1,3,4],statresult:4,stattest:4,submodul:[3,4],subpackag:3,tabl:0,test:4,util:[3,4],welcom:0}}) \ No newline at end of file +Search.setIndex({docnames:["index","modules","setup","statannotations","statannotations.stats"],envversion:{"sphinx.domains.c":2,"sphinx.domains.changeset":1,"sphinx.domains.citation":1,"sphinx.domains.cpp":3,"sphinx.domains.index":1,"sphinx.domains.javascript":2,"sphinx.domains.math":2,"sphinx.domains.python":3,"sphinx.domains.rst":2,"sphinx.domains.std":2,sphinx:56},filenames:["index.rst","modules.rst","setup.rst","statannotations.rst","statannotations.stats.rst"],objects:{"":{statannotations:[3,0,0,"-"]},"statannotations.Annotation":{Annotation:[3,1,1,""]},"statannotations.Annotation.Annotation":{formatted_output:[3,2,1,""],print_labels_and_content:[3,3,1,""],text:[3,2,1,""]},"statannotations.Annotator":{Annotator:[3,1,1,""]},"statannotations.Annotator.Annotator":{alpha:[3,2,1,""],annotate:[3,3,1,""],annotate_custom_annotations:[3,3,1,""],apply_and_annotate:[3,3,1,""],apply_test:[3,3,1,""],comparisons_correction:[3,2,1,""],configure:[3,3,1,""],fig:[3,2,1,""],get_annotations_text:[3,3,1,""],get_configuration:[3,3,1,""],get_offset_func:[3,3,1,""],has_type0_comparisons_correction:[3,3,1,""],loc:[3,2,1,""],new_plot:[3,3,1,""],orient:[3,2,1,""],print_pvalue_legend:[3,3,1,""],pvalue_format:[3,2,1,""],reset_configuration:[3,3,1,""],set_custom_annotations:[3,3,1,""],set_pvalues:[3,3,1,""],set_pvalues_and_annotate:[3,3,1,""],test:[3,2,1,""],validate_test_short_name:[3,3,1,""],verbose:[3,2,1,""]},"statannotations.PValueFormat":{Formatter:[3,1,1,""],PValueFormat:[3,1,1,""],sort_pvalue_thresholds:[3,4,1,""]},"statannotations.PValueFormat.Formatter":{config:[3,3,1,""],format_data:[3,3,1,""]},"statannotations.PValueFormat.PValueFormat":{config:[3,3,1,""],format_data:[3,3,1,""],get_configuration:[3,3,1,""],print_legend_if_used:[3,3,1,""],pvalue_format_string:[3,2,1,""],pvalue_thresholds:[3,2,1,""],simple_format_string:[3,2,1,""],text_format:[3,2,1,""]},"statannotations.format_annotations":{pval_annotation_text:[3,4,1,""],simple_text:[3,4,1,""]},"statannotations.stats":{ComparisonsCorrection:[4,0,0,"-"],StatResult:[4,0,0,"-"],StatTest:[4,0,0,"-"],test:[4,0,0,"-"],utils:[4,0,0,"-"]},"statannotations.stats.ComparisonsCorrection":{ComparisonsCorrection:[4,1,1,""],check_valid_correction_name:[4,4,1,""],get_correction_parameters:[4,4,1,""],get_validated_comparisons_correction:[4,4,1,""]},"statannotations.stats.ComparisonsCorrection.ComparisonsCorrection":{apply:[4,3,1,""],document:[4,3,1,""]},"statannotations.stats.StatResult":{StatResult:[4,1,1,""]},"statannotations.stats.StatResult.StatResult":{corrected_significance:[4,2,1,""],correction_method:[4,2,1,""],formatted_output:[4,2,1,""],significance_suffix:[4,2,1,""]},"statannotations.stats.StatTest":{StatTest:[4,1,1,""],wilcoxon:[4,4,1,""]},"statannotations.stats.StatTest.StatTest":{from_library:[4,3,1,""],short_name:[4,2,1,""]},"statannotations.stats.test":{apply_test:[4,4,1,""]},"statannotations.stats.utils":{check_alpha:[4,4,1,""],check_num_comparisons:[4,4,1,""],check_pvalues:[4,4,1,""],get_num_comparisons:[4,4,1,""],return_results:[4,4,1,""]},"statannotations.utils":{check_is_in:[3,4,1,""],check_not_none:[3,4,1,""],check_order_in_data:[3,4,1,""],check_pairs_in_data:[3,4,1,""],check_valid_text_format:[3,4,1,""],get_closest:[3,4,1,""],get_x_values:[3,4,1,""],raise_expected_got:[3,4,1,""],remove_null:[3,4,1,""],render_collection:[3,4,1,""]},statannotations:{Annotation:[3,0,0,"-"],Annotator:[3,0,0,"-"],PValueFormat:[3,0,0,"-"],format_annotations:[3,0,0,"-"],stats:[4,0,0,"-"],utils:[3,0,0,"-"]}},objnames:{"0":["py","module","Python module"],"1":["py","class","Python class"],"2":["py","property","Python property"],"3":["py","method","Python method"],"4":["py","function","Python function"]},objtypes:{"0":"py:module","1":"py:class","2":"py:property","3":"py:method","4":"py:function"},terms:{"0":[3,4],"05":[3,4],"1":[3,4],"12141511":3,"9981846":3,"case":4,"class":[3,4],"default":[3,4],"float":4,"function":[3,4],"int":4,"return":3,"short":3,"static":4,"true":[3,4],A:3,If:3,Or:3,The:3,a_list:3,accept:3,add:3,addit:4,alpha:[3,4],an:3,annot:1,annotate_custom_annot:3,appli:[3,4],apply_and_annot:3,apply_test:[3,4],ar:3,arg:[3,4],argument:[3,4],arrai:3,assum:3,auto:3,ax:3,bar:3,base:[3,4],behavior:4,benjamini:4,between:3,bonferroni:4,bool:4,box:3,boxplot:3,calcul:3,callabl:4,check:3,check_alpha:4,check_is_in:3,check_not_non:3,check_num_comparison:4,check_order_in_data:3,check_pairs_in_data:3,check_pvalu:4,check_valid_correction_nam:4,check_valid_text_format:3,classmethod:3,close:3,closest:3,collect:3,color:3,com:3,comparison:[3,4],comparisons_correct:[3,4],comparisonscorrect:[1,3],comput:3,config:3,configur:3,constructor:3,content:1,coord:3,coordin:3,core:3,corr_kwarg:4,correct:4,corrected_signific:4,correction_method:4,custom:3,data:[3,4],datafram:3,dict:4,displai:3,document:4,each:3,earlier:4,either:4,engin:3,equal:3,equival:4,error:3,error_typ:3,exact:3,exist:3,expect:3,fals:3,fig:3,follow:3,for_:3,format:[3,4],format_annot:1,format_data:3,formatt:3,formatted_output:[3,4],fraction:3,frame:3,from:[3,4],from_librari:4,func:4,gener:3,get:4,get_annotations_text:3,get_closest:3,get_configur:3,get_correction_paramet:4,get_num_comparison:4,get_offset_func:3,get_validated_comparisons_correct:4,get_x_valu:3,got:3,group:3,group_data1:4,group_data2:4,gt:4,has_type0_comparisons_correct:3,hochberg:4,hold:3,holm:4,http:3,hue:3,hue_ord:3,index:0,instanc:[3,4],instead:3,interfac:4,interpret:4,keyword:4,kruskal:4,kwarg:[3,4],label:3,leven:4,line_height:3,line_offset:3,line_offset_to_group:3,line_width:3,link:3,list:3,loc:3,ls:4,mann:4,messag:3,method:4,method_typ:4,mode:3,modul:[0,1],multipl:4,must:3,mylist:3,mynumb:3,name:[3,4],new_plot:3,none:[3,4],num_comparison:[3,4],number:[3,4],object:[3,4],one:[3,4],option:[3,4],order:3,orient:3,otherwis:3,overrid:3,p:3,p_valu:4,packag:[0,1],page:0,pair:3,panda:3,paramet:[3,4],pass:[3,4],per:3,perform:3,plot:3,plot_param:3,point:3,posit:3,print_labels_and_cont:3,print_legend_if_us:3,print_pvalue_legend:3,properti:[3,4],provid:[3,4],pval:4,pval_annotation_text:3,pvalu:[3,4],pvalue_format:3,pvalue_format_str:3,pvalue_threshold:3,pvalueformat:1,rais:3,raise_expected_got:3,rang:3,refer:3,remove_nul:3,render:3,render_collect:3,reset_configur:3,result:[3,4],results_arrai:4,return_result:4,run:4,same:3,sampl:4,seaborn:3,search:0,sep:3,seri:3,set:3,set_custom_annot:3,set_pvalu:3,set_pvalues_and_annot:3,short_nam:4,show:3,show_test_nam:3,significance_suffix:4,simpl:3,simple_format_str:3,simple_text:3,smallest:3,sort:3,sort_pvalue_threshold:3,stackoverflow:3,standard:3,stat:[1,3],stat_nam:4,stat_str:4,statist:[3,4],statresult:[1,3],stats_param:[3,4],statsmodel:4,statsmodels_api:4,stattest:[1,3],str:[3,4],string:3,struct:3,submodul:1,subpackag:1,t:4,test:[1,3],test_descript:4,test_ind:4,test_long_nam:4,test_nam:4,test_pair:4,test_result_list:4,test_short_nam:[3,4],test_welch:4,text:3,text_annot_custom:3,text_format:3,text_offset:3,them:3,thereof:3,thi:[3,4],three:3,threshold:3,top:3,tupl:3,two:[3,4],type:3,union:[3,4],us:[3,4],use_fixed_offset:3,util:1,valid_valu:3,validate_test_short_nam:3,valu:[3,4],valueerror:3,verbos:[3,4],version:4,vs:3,wa:3,whitnei:4,wilcoxon:4,work:3,x:3,y:3,yekuti:4},titles:["Welcome to statannotations\u2019s documentation!","statannotations","setup module","statannotations package","statannotations.stats package"],titleterms:{annot:3,comparisonscorrect:4,content:[0,3,4],document:0,format_annot:3,indic:0,modul:[2,3,4],packag:[3,4],pvalueformat:3,s:0,setup:2,stat:4,statannot:[0,1,3,4],statresult:4,stattest:4,submodul:[3,4],subpackag:3,tabl:0,test:4,util:[3,4],welcom:0}}) \ No newline at end of file diff --git a/docs/build/html/statannotations.html b/docs/build/html/statannotations.html index aaacef6..7363f25 100644 --- a/docs/build/html/statannotations.html +++ b/docs/build/html/statannotations.html @@ -352,6 +352,11 @@ <h2>Submodules<a class="headerlink" href="#submodules" title="Permalink to this new_plot(ax, pairs=None, plot='boxplot', data=None, x=None, y=None, hue=None, order=None, hue_order=None, engine: str = 'seaborn', **plot_params)<a class="headerlink" href="#statannotations.Annotator.Annotator.new_plot" title="Permalink to this definition">¶</a></dt> <dd></dd></dl> +<dl class="py property"> +<dt class="sig sig-object py" id="statannotations.Annotator.Annotator.orient"> +property orient<a class="headerlink" href="#statannotations.Annotator.Annotator.orient" title="Permalink to this definition">¶</a></dt> +<dd></dd></dl> + <dl class="py method"> <dt class="sig sig-object py" id="statannotations.Annotator.Annotator.print_pvalue_legend"> print_pvalue_legend()<a class="headerlink" href="#statannotations.Annotator.Annotator.print_pvalue_legend" title="Permalink to this definition">¶</a></dt> @@ -544,7 +549,7 @@ <h2>Submodules<a class="headerlink" href="#submodules" title="Permalink to this <dl class="py function"> <dt class="sig sig-object py" id="statannotations.utils.check_pairs_in_data"> -statannotations.utils.check_pairs_in_data(pairs: Union[list, tuple], data: Optional[Union[List[list], pandas.core.frame.DataFrame]] = None, x: Optional[Union[str, list]] = None, hue: Optional[str] = None, hue_order: Optional[List[str]] = None)<a class="headerlink" href="#statannotations.utils.check_pairs_in_data" title="Permalink to this definition">¶</a></dt> +statannotations.utils.check_pairs_in_data(pairs: Union[list, tuple], data: Optional[Union[List[list], pandas.core.frame.DataFrame]] = None, coord: Optional[Union[str, list]] = None, hue: Optional[str] = None, hue_order: Optional[List[str]] = None)<a class="headerlink" href="#statannotations.utils.check_pairs_in_data" title="Permalink to this definition">¶</a></dt> <dd>Checks that values referred to in <cite>order</cite> and <cite>pairs</cite> exist in data. </dd></dl> diff --git a/statannotations/Annotator.py b/statannotations/Annotator.py index 390bd83..00dcee6 100644 --- a/statannotations/Annotator.py +++ b/statannotations/Annotator.py @@ -123,7 +123,7 @@ def __init__(self, ax, pairs, plot='boxplot', data=None, x=None, self.text_offset_impact_above = 0 self.color = '0.2' self.line_width = 1.5 - self.y_offset = None + self.value_offset = None self.custom_annotations = None def new_plot(self, ax, pairs=None, plot='boxplot', data=None, x=None, @@ -144,6 +144,10 @@ def new_plot(self, ax, pairs=None, plot='boxplot', data=None, x=None, return self + @property + def orient(self): + return self._plotter.orient + @property def verbose(self): return self._verbose @@ -154,8 +158,8 @@ def verbose(self, verbose): self._plotter.verbose = verbose @property - def _y_stack_arr(self): - return self._plotter.y_stack_arr + def _value_stack_arr(self): + return self._plotter.value_stack_arr @property def fig(self): @@ -179,13 +183,13 @@ def annotate(self, line_offset=None, line_offset_to_group=None): "test (again) which will probably lead to " "unexpected results") - self._update_y_for_loc() + self._update_value_for_loc() ann_list = [] - orig_ylim = self.ax.get_ylim() + orig_value_lim = self._plotter.get_value_lim() offset_func = self.get_offset_func(self.loc) - self.y_offset, self.line_offset_to_group = offset_func( + self.value_offset, self.line_offset_to_group = offset_func( line_offset, line_offset_to_group) if self._verbose: @@ -199,16 +203,24 @@ def annotate(self, line_offset=None, line_offset_to_group=None): self._annotate_pair(annotation, ax_to_data=ax_to_data, ann_list=ann_list, - orig_ylim=orig_ylim) + orig_value_lim=orig_value_lim) # reset transformation - y_stack_max = max(self._y_stack_arr[1, :]) + y_stack_max = max(self._value_stack_arr[1, :]) ax_to_data = self._plotter.get_transform_func('ax_to_data') - ylims = ([(0, 0), (0, max(1.04 * y_stack_max, 1))] - if self.loc == 'inside' - else [(0, 0), (0, 1)]) - - self.ax.set_ylim(ax_to_data.transform(ylims)[:, 1]) + value_lims = ( + ([(0, 0), (0, max(1.04 * y_stack_max, 1))] + if self.loc == 'inside' + else [(0, 0), (0, 1)]) + if self.orient == 'v' + else + ([(0, 0), (max(1.04 * y_stack_max, 1), 0)] + if self.loc == 'inside' + else [(0, 0), (1, 0)]) + ) + set_lims = self.ax.set_ylim if self.orient == 'v' else self.ax.set_xlim + transformed = ax_to_data.transform(value_lims) + set_lims(transformed[:, 1 if self.orient == 'v' else 0]) return self.ax, self.annotations @@ -444,61 +456,75 @@ def _get_results(self, num_comparisons, pvalues=None, return test_result_list - def _annotate_pair(self, annotation, ax_to_data, ann_list, orig_ylim): + def _annotate_pair(self, annotation, ax_to_data, ann_list, orig_value_lim): if self._verbose >= 1: annotation.print_labels_and_content() - x1 = annotation.structs[0]['x'] - x2 = annotation.structs[1]['x'] - xi1 = annotation.structs[0]['xi'] - xi2 = annotation.structs[1]['xi'] + group_coord_1 = annotation.structs[0]['group_coord'] + group_coord_2 = annotation.structs[1]['group_coord'] + group_i1 = annotation.structs[0]['group_i'] + group_i2 = annotation.structs[1]['group_i'] # Find y maximum for all the y_stacks *in between* group1 and group2 - i_ymax_in_range_x1_x2 = xi1 + np.nanargmax( - self._y_stack_arr[1, np.where((x1 <= self._y_stack_arr[0, :]) - & (self._y_stack_arr[0, :] <= x2))]) + i_value_max_in_range_g1_g2 = group_i1 + np.nanargmax( + self._value_stack_arr[1, + np.where((group_coord_1 + <= self._value_stack_arr[0, :]) + & (self._value_stack_arr[0, :] + <= group_coord_2))]) - y = self._get_y_for_pair(i_ymax_in_range_x1_x2) + value = self._get_value_for_pair(i_value_max_in_range_g1_g2) # Determine lines in axes coordinates - ax_line_x = [x1, x1, x2, x2] - ax_line_y = [y, y + self.line_height, y + self.line_height, y] + ax_line_group = [group_coord_1, group_coord_1, + group_coord_2, group_coord_2] + ax_line_value = [value, value + self.line_height, + value + self.line_height, value] + + lists = ((ax_line_group, ax_line_value) if self.orient == 'v' + else (ax_line_value, ax_line_group)) points = [ax_to_data.transform((x, y)) for x, y - in zip(ax_line_x, ax_line_y)] + in zip(*lists)] line_x, line_y = zip(*points) self._plot_line(line_x, line_y) + xy_params = self._get_xy_params(group_coord_1, group_coord_2, line_x, + line_y) + ann = self.ax.annotate( - annotation.text, xy=(np.mean([x1, x2]), line_y[2]), - xytext=(0, self.text_offset), textcoords='offset points', + annotation.text, textcoords='offset points', xycoords='data', ha='center', va='bottom', fontsize=self._pvalue_format.fontsize, clip_on=False, - annotation_clip=False) + annotation_clip=False, **xy_params) + if annotation.text is not None: ann_list.append(ann) plt.draw() - self.ax.set_ylim(orig_ylim) + set_lim = {'v': 'set_ylim', + 'h': 'set_xlim'}[self.orient] + + getattr(self.ax, set_lim)(orig_value_lim) - y_top_annot = self._annotate_pair_text(ann, y) + value_top_annot = self._annotate_pair_text(ann, value) else: - y_top_annot = y + self.line_height + value_top_annot = value + self.line_height - # Fill the highest y position of the annotation into the y_stack array - # for all positions in the range x1 to x2 - self._y_stack_arr[1, - (x1 <= self._y_stack_arr[0, :]) - & (self._y_stack_arr[0, :] <= x2)] = y_top_annot + # Fill the highest value position of the annotation into the value_stack + # array for all positions in the range group_coord_1 to group_coord_2 + self._value_stack_arr[ + 1, (group_coord_1 <= self._value_stack_arr[0, :]) + & (self._value_stack_arr[0, :] <= group_coord_2)] = value_top_annot - # Increment the counter of annotations in the y_stack array - self._y_stack_arr[2, xi1:xi2 + 1] += 1 + # Increment the counter of annotations in the value_stack array + self._value_stack_arr[2, group_i1:group_i2 + 1] += 1 - def _update_y_for_loc(self): + def _update_value_for_loc(self): if self._loc == 'outside': - self._y_stack_arr[1, :] = 1 + self._value_stack_arr[1, :] = 1 def _check_test_pvalues_perform(self): if self.test is None: @@ -582,8 +608,7 @@ def _get_offsets_inside(line_offset, line_offset_to_group): else: if line_offset_to_group is None: line_offset_to_group = 0.06 - y_offset = line_offset - return y_offset, line_offset_to_group + return line_offset, line_offset_to_group @staticmethod def _get_offsets_outside(line_offset, line_offset_to_group): @@ -623,9 +648,9 @@ def _plot_line(self, line_x, line_y): line.set_clip_on(False) self.ax.add_line(line) - def _annotate_pair_text(self, ann, y): + def _annotate_pair_text(self, ann, value): - y_top_annot = None + value_top_annot = None got_mpl_error = False if not self.use_fixed_offset: @@ -633,7 +658,8 @@ def _annotate_pair_text(self, ann, y): bbox = ann.get_window_extent() pix_to_ax = self._plotter.get_transform_func('pix_to_ax') bbox_ax = bbox.transformed(pix_to_ax) - y_top_annot = bbox_ax.ymax + value_coord_max = {'v': 'ymax', 'h': 'xmax'}[self.orient] + value_top_annot = getattr(bbox_ax, value_coord_max) except RuntimeError: got_mpl_error = True @@ -644,40 +670,45 @@ def _annotate_pair_text(self, ann, y): "back to a fixed y offset. Layout may be not " "optimal.") - # We will apply a fixed offset in points, - # based on the font size of the annotation. fontsize_points = FontProperties( size='medium').get_size_in_points() - offset_trans = mtransforms.offset_copy( - self.ax.transAxes, fig=self.fig, x=0, - y=fontsize_points + self.text_offset, units='points') - y_top_display = offset_trans.transform( - (0, y + self.line_height)) - y_top_annot = (self.ax.transAxes.inverted() - .transform(y_top_display)[1]) + direction = {'h': -1, 'v': 1}[self.orient] + x, y = [0, fontsize_points + self.text_offset][::direction] + offset_trans = mtransforms.offset_copy(trans=self.ax.transAxes, + fig=self.fig, + units='points', x=x, y=y) + + value_top_display = offset_trans.transform( + (value + self.line_height, value + self.line_height)) + + value_coord = {'h': 0, 'v': 1}[self.orient] + + value_top_annot = (self.ax.transAxes.inverted() + .transform(value_top_display)[value_coord]) self.text_offset_impact_above = ( - y_top_annot - y - self.y_offset - self.line_height) - return y_top_annot + value_top_annot - value - self.value_offset - self.line_height) + + return value_top_annot def _reset_default_values(self): for attribute, default_value in _DEFAULT_VALUES.items(): setattr(self, attribute, default_value) - def _get_y_for_pair(self, i_ymax_in_range_x1_x2): + def _get_value_for_pair(self, i_ymax_in_range_x1_x2): - ymax_in_range_x1_x2 = self._y_stack_arr[1, i_ymax_in_range_x1_x2] + ymax_in_range_x1_x2 = self._value_stack_arr[1, i_ymax_in_range_x1_x2] # Choose the best offset depending on whether there is an annotation # below at the x position in the range [x1, x2] where the stack is the # highest - if self._y_stack_arr[2, i_ymax_in_range_x1_x2] == 0: + if self._value_stack_arr[2, i_ymax_in_range_x1_x2] == 0: # there is only a group below offset = self.line_offset_to_group else: # there is an annotation below - offset = self.y_offset + self.text_offset_impact_above + offset = self.value_offset + self.text_offset_impact_above return ymax_in_range_x1_x2 + offset @@ -697,3 +728,28 @@ def _get_plotter(engine, *args, **kwargs): if engine_plotter is None: raise NotImplementedError(f"{engine} engine not implemented.") return engine_plotter(*args, **kwargs) + + def _get_xy_params_horizontal(self, group_coord_1, group_coord_2, + line_x: np.ndarray): + return { + 'xy': (line_x[2], np.mean([group_coord_1, group_coord_2])), + 'xytext': (self.text_offset, 0), + 'rotation': 270, + 'rotation_mode': 'anchor' + } + + def _get_xy_params_vertical(self, group_coord_1, group_coord_2, + line_y: np.ndarray): + return { + 'xy': (np.mean([group_coord_1, group_coord_2]), line_y[2]), + 'xytext': (0, self.text_offset), + } + + def _get_xy_params(self, group_coord_1, group_coord_2, line_x: np.ndarray, + line_y: np.ndarray): + if self.orient == 'h': + return self._get_xy_params_horizontal(group_coord_1, group_coord_2, + line_x) + + return self._get_xy_params_vertical(group_coord_1, group_coord_2, + line_y) diff --git a/statannotations/_Xpositions.py b/statannotations/_GroupsPositions.py similarity index 51% rename from statannotations/_Xpositions.py rename to statannotations/_GroupsPositions.py index 2520f3c..046d2b9 100644 --- a/statannotations/_Xpositions.py +++ b/statannotations/_GroupsPositions.py @@ -3,7 +3,7 @@ from statannotations.utils import get_closest -class _XPositions: +class _GroupsPositions: def __init__(self, plotter, group_names): self._plotter = plotter self._hue_names = self._plotter.hue_names @@ -15,40 +15,42 @@ def __init__(self, plotter, group_names): "Using hues with only one hue is not supported.") self.hue_offsets = self._plotter.hue_offsets - self._xunits = self.hue_offsets[1] - self.hue_offsets[0] + self._axis_units = self.hue_offsets[1] - self.hue_offsets[0] - self._xpositions = { - np.round(self.get_group_x_position(group_name), 1): group_name + self._groups_positions = { + np.round(self.get_group_axis_position(group_name), 1): group_name for group_name in group_names } - self._xpositions_list = sorted(self._xpositions.keys()) + self._groups_positions_list = sorted(self._groups_positions.keys()) if self._hue_names is None: - self._xunits = ((max(list(self._xpositions.keys())) + 1) - / len(self._xpositions)) + self._axis_units = ((max(list(self._groups_positions.keys())) + 1) + / len(self._groups_positions)) - self._xranges = { - (pos - self._xunits / 2, pos + self._xunits / 2, pos): group_name - for pos, group_name in self._xpositions.items()} + self._axis_ranges = { + (pos - self._axis_units / 2, + pos + self._axis_units / 2, + pos): group_name + for pos, group_name in self._groups_positions.items()} @property - def xpositions(self): - return self._xpositions + def axis_positions(self): + return self._groups_positions @property - def xunits(self): - return self._xunits + def axis_units(self): + return self._axis_units - def get_xpos_location(self, pos): + def get_axis_pos_location(self, pos): """ Finds the x-axis location of a categorical variable """ - for xrange in self._xranges: - if (pos >= xrange[0]) & (pos <= xrange[1]): - return xrange[2] + for axis_range in self._axis_ranges: + if (pos >= axis_range[0]) & (pos <= axis_range[1]): + return axis_range[2] - def get_group_x_position(self, group): + def get_group_axis_position(self, group): """ group_name can be either a name "cat" or a tuple ("cat", "hue") """ @@ -64,5 +66,5 @@ def get_group_x_position(self, group): group_pos = self._plotter.group_names.index(cat) + hue_offset return group_pos - def find_closest(self, xpos): - return get_closest(list(self._xpositions_list), xpos) + def find_closest(self, pos): + return get_closest(list(self._groups_positions_list), pos) diff --git a/statannotations/_Plotter.py b/statannotations/_Plotter.py index 9dac681..5223865 100644 --- a/statannotations/_Plotter.py +++ b/statannotations/_Plotter.py @@ -8,7 +8,7 @@ from matplotlib.collections import PathCollection from matplotlib.patches import Rectangle -from statannotations._Xpositions import _XPositions +from statannotations._GroupsPositions import _GroupsPositions from statannotations.utils import check_not_none, check_order_in_data, \ check_pairs_in_data, render_collection, check_is_in, remove_null @@ -18,16 +18,18 @@ class _Plotter: - def __init__(self, ax, pairs, data=None, x=None, hue=None, - order=None, hue_order=None, verbose=False): + def __init__(self, ax, pairs, data=None, x=None, y=None, hue=None, + order=None, hue_order=None, verbose=False, **plot_params): self.ax = ax self._fig = plt.gcf() check_not_none("pairs", pairs) - check_order_in_data(data, x, order) - check_pairs_in_data(pairs, data, x, hue, hue_order) + group_coord = y if plot_params.get("orient") == "h" else x + check_order_in_data(data, group_coord, order) + check_pairs_in_data(pairs, data, group_coord, hue, hue_order) self.pairs = pairs self._struct_pairs = None self.verbose = verbose + self.orient = plot_params.get("orient", "v") def get_transform_func(self, kind: str): """ @@ -48,9 +50,11 @@ def get_transform_func(self, kind: str): if kind == 'pix_to_ax': return self.ax.transAxes.inverted() - data_to_ax = \ - self.ax.transData + self.ax.get_xaxis_transform().inverted() + transform = {'v': self.ax.get_xaxis_transform, + 'h': self.ax.get_yaxis_transform}[self.orient] + data_to_ax = \ + self.ax.transData + transform().inverted() if kind == 'data_to_ax': return data_to_ax @@ -75,34 +79,35 @@ def __init__(self, ax, pairs, plot='boxplot', data=None, x=None, y=None, hue=None, order=None, hue_order=None, verbose=False, **plot_params): - _Plotter.__init__(self, ax, pairs, data, x, hue, order, hue_order, - verbose) + _Plotter.__init__(self, ax, pairs, data, x, y, hue, order, hue_order, + verbose, **plot_params) self.check_plot_is_implemented(plot) - self.plot = plot self.plotter = self._get_plotter(plot, x, y, hue, data, order, hue_order, **plot_params) self.group_names, self.labels = self._get_group_names_and_labels() - self.xpositions = _XPositions(self.plotter, self.group_names) + self.groups_positions = _GroupsPositions(self.plotter, + self.group_names) self.reordering = None - self.ymaxes = self._generate_ymaxes() + self.value_maxes = self._generate_value_maxes() self.structs = self._get_structs() self.pairs = pairs self._struct_pairs = self._get_group_struct_pairs() - self._y_stack_arr = np.array( - [[struct['x'], struct['ymax'], 0] for struct in self.structs] + self._value_stack_arr = np.array( + [[struct['group_coord'], struct['value_max'], 0] + for struct in self.structs] ).T self.reordering = None self._sort_group_struct_pairs() @property - def y_stack_arr(self): - return self._y_stack_arr + def value_stack_arr(self): + return self._value_stack_arr # noinspection PyProtectedMember def _get_plotter(self, plot, x, y, hue, data, order, hue_order, @@ -187,7 +192,7 @@ def _get_group_names_and_labels(self): return group_names, labels - def _generate_ymaxes(self): + def _generate_value_maxes(self): """ given plotter and the names of two categorical variables, returns highest y point drawn between those two variables before @@ -197,38 +202,40 @@ def _generate_ymaxes(self): (eg, error bars and/or bar charts). """ - ymaxes = {name: 0 for name in self.group_names} + value_maxes = {name: 0 for name in self.group_names} data_to_ax = self.get_transform_func('data_to_ax') if self.plot == 'violinplot': - ymaxes = self._get_ymaxes_violin(ymaxes, data_to_ax) + value_maxes = self._get_value_maxes_violin(value_maxes, data_to_ax) else: for child in self.ax.get_children(): - xname, ypos = self._get_child_ypos(child, data_to_ax) + group_name, value_pos = self._get_value_pos(child, data_to_ax) - if ypos is not None and ypos > ymaxes[xname]: - ymaxes[xname] = ypos + if value_pos is not None and value_pos > value_maxes[group_name]: + value_maxes[group_name] = value_pos - return ymaxes + return value_maxes def _get_structs(self): structs = [ { 'group': group_name, 'label': self.labels[b_idx], - 'x': self.xpositions.get_group_x_position(group_name), + 'group_coord': (self.groups_positions + .get_group_axis_position(group_name)), 'group_data': self._get_group_data(group_name), - 'ymax': self.ymaxes[group_name] - } for b_idx, group_name in enumerate(self.group_names)] + 'value_max': self.value_maxes[group_name] + } + for b_idx, group_name in enumerate(self.group_names)] - # Sort the group data structures by position along the x axis - structs = sorted(structs, key=lambda struct: struct['x']) + # Sort the group data structures by position along the groups axis + structs = sorted(structs, key=lambda struct: struct['group_coord']) - # Add the index position in the list of groups along the x axis - structs = [dict(struct, xi=i) + # Add the index position in the list of groups along the groups axis + structs = [dict(struct, group_i=i) for i, struct in enumerate(structs)] return structs @@ -246,7 +253,7 @@ def _get_group_struct_pairs(self): group_struct2 = dict(group_structs_dic[group2], i_group_pair=i_group_pair) - if group_struct1['x'] <= group_struct2['x']: + if group_struct1['group_coord'] <= group_struct2['group_coord']: group_struct_pairs.append((group_struct1, group_struct2)) else: @@ -289,62 +296,79 @@ def _get_group_data(self, group_name): return group_data - def _get_ymaxes_violin(self, ymaxes, data_to_ax): + def _get_value_maxes_violin(self, value_maxes, data_to_ax): for group_idx, group_name in enumerate(self.plotter.group_names): if self.plotter.hue_names: for hue_idx, hue_name in enumerate(self.plotter.hue_names): - ypos = max(self.plotter.support[group_idx][hue_idx]) - ymaxes[(group_name, hue_name)] = \ - data_to_ax.transform((0, ypos))[1] + value_pos = max(self.plotter.support[group_idx][hue_idx]) + value_maxes[(group_name, hue_name)] = \ + data_to_ax.transform((0, value_pos))[1] else: - ypos = max(self.plotter.support[group_idx]) - ymaxes[group_name] = data_to_ax.transform((0, ypos))[1] - return ymaxes + value_pos = max(self.plotter.support[group_idx]) + value_maxes[group_name] = data_to_ax.transform((0, + value_pos))[1] + return value_maxes - def _get_child_ypos(self, child, data_to_ax): + def _get_value_pos(self, child, data_to_ax): if (type(child) == PathCollection and len(child.properties()['offsets'])): - return self._get_ypos_for_path_collection( + return self._get_value_pos_for_path_collection( child, data_to_ax) elif type(child) in (lines.Line2D, Rectangle): - return self._get_ypos_for_line2d_or_rectangle( + return self._get_value_pos_for_line2d_or_rectangle( child, data_to_ax) return None, None - def _get_ypos_for_path_collection(self, child, data_to_ax): - ymax = child.properties()['offsets'][:, 1].max() - xpos = float(np.round(np.nanmean( - child.properties()['offsets'][:, 0]), 1)) - if xpos not in self.xpositions.xpositions: + def _get_value_pos_for_path_collection(self, child, data_to_ax): + group_coord = {"v": 0, "h": 1}[self.orient] + value_coord = (group_coord + 1) % 2 + + direction = {"v": 1, "h": -1}[self.orient] + + value_max = child.properties()['offsets'][:, value_coord].max() + group_pos = float(np.round(np.nanmean( + child.properties()['offsets'][:, group_coord]), 1)) + + if group_pos not in self.groups_positions.axis_positions: if self.verbose: warnings.warn( "Invalid x-position found. Are the same parameters passed " "to seaborn and statannotations calls? or are there few " "data points?") - xpos = self.xpositions.find_closest(xpos) + group_pos = self.groups_positions.find_closest(group_pos) + + group_name = self.groups_positions.axis_positions[group_pos] - xname = self.xpositions.xpositions[xpos] - ypos = data_to_ax.transform((0, ymax))[1] + value_pos = data_to_ax.transform((0, value_max)[::direction]) - return xname, ypos + return group_name, value_pos[value_coord] - def _get_ypos_for_line2d_or_rectangle(self, child, data_to_ax): + def _get_value_pos_for_line2d_or_rectangle(self, child, data_to_ax): bbox = self.ax.transData.inverted().transform( child.get_window_extent(self.fig.canvas.get_renderer())) - if (bbox[:, 0].max() - bbox[:, 0].min()) > 1.1*self.xpositions.xunits: + group_coord = {"v": 0, "h": 1}[self.orient] + direction = {"v": 1, "h": -1}[self.orient] + + value_coord = (group_coord + 1) % 2 + + if ((bbox[:, group_coord].max() - bbox[:, group_coord].min()) + > 1.1 * self.groups_positions.axis_units): return None, None - raw_xpos = np.round(bbox[:, 0].mean(), 1) - xpos = self.xpositions.get_xpos_location(raw_xpos) - if xpos not in self.xpositions.xpositions: + + raw_group_pos = np.round(bbox[:, group_coord].mean(), 1) + group_pos = self.groups_positions.get_axis_pos_location(raw_group_pos) + + if group_pos not in self.groups_positions.axis_positions: return None, None - xname = self.xpositions.xpositions[xpos] - ypos = bbox[:, 1].max() - ypos = data_to_ax.transform((0, ypos))[1] - return xname, ypos + group_name = self.groups_positions.axis_positions[group_pos] + value_pos = bbox[:, value_coord].max() + + value_pos = data_to_ax.transform((0, value_pos)[::direction]) + return group_name, value_pos[value_coord] def _sort_group_struct_pairs(self): # Draw first the annotations with the shortest between-groups distance, @@ -357,7 +381,8 @@ def _sort_group_struct_pairs(self): @staticmethod def _absolute_group_struct_pair_in_tuple_x_diff(group_struct_pair): - return abs(group_struct_pair[1][1]['x'] - group_struct_pair[1][0]['x']) + return abs(group_struct_pair[1][1]['group_coord'] + - group_struct_pair[1][0]['group_coord']) @staticmethod def check_plot_is_implemented(plot, engine="seaborn"): @@ -377,3 +402,8 @@ def fix_and_warn(dodge, hue, plot): "Implicitly setting dodge to True as it is necessary in " "statannotations. It must have been True for the seaborn " "call to yield consistent results when using `hue`.") + + def get_value_lim(self): + if self.orient == 'v': + return self.ax.get_ylim() + return self.ax.get_xlim() diff --git a/statannotations/_version.py b/statannotations/_version.py index 6a9beea..3d26edf 100644 --- a/statannotations/_version.py +++ b/statannotations/_version.py @@ -1,1 +1,1 @@ -__version__ = "0.4.0" +__version__ = "0.4.1" diff --git a/statannotations/utils.py b/statannotations/utils.py index c7d44fc..934239b 100644 --- a/statannotations/utils.py +++ b/statannotations/utils.py @@ -75,19 +75,19 @@ def _check_pairs_in_data_no_hue(pairs: Union[list, tuple], def _check_pairs_in_data_with_hue(pairs: Union[list, tuple], data: Union[List[list], pd.DataFrame] = None, - x: Union[str, list] = None, + group_coord: Union[str, list] = None, hue: str = None) -> set: - x_values = get_x_values(data, x) - seen_x_values = set() + x_values = get_x_values(data, group_coord) + seen_group_values = set() hue_values = set(data[hue].unique()) - for x_value, hue_value in itertools.chain(itertools.chain(*pairs)): - if x_value not in seen_x_values and x_value not in x_values: - raise ValueError(f"Missing x value `{x_value}` in {x}" + for group_value, hue_value in itertools.chain(itertools.chain(*pairs)): + if group_value not in seen_group_values and group_value not in x_values: + raise ValueError(f"Missing group value `{group_value}` in {group_coord}" f" (specified in `pairs`)") - seen_x_values.add(x_value) + seen_group_values.add(group_value) if hue_value not in hue_values: raise ValueError(f"Missing hue value `{hue_value}` in {hue}" @@ -108,7 +108,7 @@ def _check_hue_order_in_data(hue, hue_values: set, def check_pairs_in_data(pairs: Union[list, tuple], data: Union[List[list], pd.DataFrame] = None, - x: Union[str, list] = None, + coord: Union[str, list] = None, hue: str = None, hue_order: List[str] = None): """ @@ -116,9 +116,9 @@ def check_pairs_in_data(pairs: Union[list, tuple], """ if hue is None and hue_order is None: - _check_pairs_in_data_no_hue(pairs, data, x) + _check_pairs_in_data_no_hue(pairs, data, coord) else: - hue_values = _check_pairs_in_data_with_hue(pairs, data, x, hue) + hue_values = _check_pairs_in_data_with_hue(pairs, data, coord, hue) _check_hue_order_in_data(hue, hue_values, hue_order) diff --git a/usage/example.ipynb b/usage/example.ipynb index 10b91f9..91a3e66 100644 --- a/usage/example.ipynb +++ b/usage/example.ipynb @@ -231,7 +231,7 @@ }, { "data": { - "text/plain": "(<matplotlib.axes._subplots.AxesSubplot at 0x7fedb54c7748>,\n [<statannotations.Annotation.Annotation at 0x7fedb4233ac8>,\n <statannotations.Annotation.Annotation at 0x7fedb4233d30>,\n <statannotations.Annotation.Annotation at 0x7fedb4233320>])" + "text/plain": "(<matplotlib.axes._subplots.AxesSubplot at 0x7fc28184f6a0>,\n [<statannotations.Annotation.Annotation at 0x7fc28164aeb8>,\n <statannotations.Annotation.Annotation at 0x7fc28183d6d8>,\n <statannotations.Annotation.Annotation at 0x7fc282ada5c0>])" }, "execution_count": 6, "metadata": {}, @@ -672,7 +672,7 @@ }, { "data": { - "text/plain": "(<matplotlib.axes._subplots.AxesSubplot at 0x7fedb4460898>,\n [<statannotations.Annotation.Annotation at 0x7fedb40b4400>,\n <statannotations.Annotation.Annotation at 0x7fedb715ccf8>,\n <statannotations.Annotation.Annotation at 0x7fedb6c8ef28>])" + "text/plain": "(<matplotlib.axes._subplots.AxesSubplot at 0x7fc284d10f60>,\n [<statannotations.Annotation.Annotation at 0x7fc281c837b8>,\n <statannotations.Annotation.Annotation at 0x7fc2817bfc18>,\n <statannotations.Annotation.Annotation at 0x7fc284d8e518>])" }, "execution_count": 15, "metadata": {}, @@ -947,7 +947,7 @@ { "data": { "text/plain": "<Figure size 864x432 with 1 Axes>", - "image/png": "iVBORw0KGgoAAAANSUhEUgAAAtoAAAF2CAYAAABQ2D87AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjEsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+j8jraAAAgAElEQVR4nOzdd3xUVfr48c+dnt5DEkLvvUsJoSu9s4KoKAqi32WL7lrWHy4iuuruKtZV0NVdsSAKKMVFQJRepEoPKJCEkN4nmX5/f0QGhgkQIMmE5Hm/Xr5e3nPbMxkmeebcc56jqKqqIoQQQgghhKhUGl8HIIQQQgghRG0kibYQQgghhBBVQBJtIYQQQgghqoAk2kIIIYQQQlQBSbSFEEIIIYSoAjpfB1AVXC4XZrMZvV6Poii+DkcIIYQQQtRCqqpit9sJCAhAo/Huv66VibbZbCYpKcnXYQghhBBCiDqgZcuWBAUFebXXykRbr9cDZS/aYDD4OBohhBBCCFEb2Ww2kpKS3Lnn5Wplon1huIjBYMBoNPo4GiGEEEIIUZtdaaiyTIYUQgghhBCiCkiiLYQQQgghRBWQRFsIIYQQQogqUCvHaAshhBBCVCe73U5qaioWi8XXoYgqYjKZiI+Pv+LEx/JIoi2EEEIIcZNSU1MJCgqicePGsoZHLaSqKjk5OaSmptKkSZMKn1djE+0ff/yRL774AlVVue222/jNb37j65CEEEIIIcplsVgkya7FFEUhIiKCrKys6zqvxo7RLiws5LnnnuPll1/mu+++83U4QgghhBBXJUl27XYj72+N6dF+//332bp1q3v7gw8+QFVV/vnPfzJt2jQfRiaEEEIIUXUOHDjAK6+8Qn5+PqqqEhMTw5NPPkmLFi18HRpPPfUUBw4cYPny5fj7+7vbu3TpwqpVq4iPj/dhdDVfjUm0Z8yYwYwZM9zbhYWFvPjii0ydOpUOHTr4MDIhhBBCiKphs9mYNWsWH3zwAe3atQPg66+/ZubMmXz33XdotVofRwjnzp3jhRde4IUXXvB1KLecGpNoX+75558nPT2d//73v8TGxvKnP/3J1yGJX1ksFsxmM3PnzuWVV15BVVUWL15M06ZNSUhIwGQy+TpEIYQQ4pZQWlpKUVERJSUl7rYxY8YQGBjInDlziI6O5tFHHwXKEvB169Yxbdo0FixYQIMGDTh58iQOh4N58+bRrVs3ioqKmDdvHsePH0dRFBITE3nsscfQ6XR06NCBhx56iG3btpGZmcmMGTOYOnXqNWOcNm0aX3/9Nd9++y1Dhw712r9hwwbeeustXC4XAQEB/OUvf6Fjx468+eabnDt3jqysLM6dO0e9evX4xz/+QXR0NJ9++ilLlixBr9djNBp57rnnKCgo4E9/+hMbN25Eo9FQWlrKoEGDWLNmDZMmTWL8+PHs2LGD8+fPM3bsWP74xz8C8Pnnn7N48WI0Gg2RkZE888wzNGnShKeeeorAwEBOnDhBeno6rVq14uWXXyYgIKCS3r0KUKtYUVGROnLkSDUlJcXdtnLlSnX48OHq7bffrn788ceVfk+LxaLu2bNHtVgslX5toaorVqxQH3zwQXXQoEHqAw88oN57773quHHj1KlTp6p/+9vffB2eEEIIUe2OHj16w+d+8MEHaseOHdVBgwapf/7zn9UvvvhCLSkpUY8ePaomJCSodrtdVVVVnTp1qrp582Z1586daps2bdz3/Pe//63efffdqqqq6hNPPKHOnz9fdblcqtVqVR944AF14cKFqqqqasuWLdXFixerqqqqhw4dUtu3b3/NXOnJJ59U33//fXXLli3qbbfdpqalpamqqqqdO3dWU1JS1FOnTql9+vRRk5OTVVVV1e3bt6sJCQlqUVGR+sYbb6iDBw9Wi4qKVFVV1VmzZqmvv/666nA41Hbt2qkZGRmqqpblFUuWLFFVVVXHjBmj/vDDD6qqquoXX3yhPvroo6qqqurAgQPVl156SVVVVU1PT1c7dOigJicnq9u3b1eHDBmi5uTkqKqqqsuWLVOHDx+uulwu9cknn1QnT56sWq1W1WazqePGjVO//PLLG36fVNX7fb5WzlmlPdoHDx5kzpw5nDlzxt2WkZHBggULWL58OQaDgSlTptCzZ0+aN29e6fc/fPhwpV9TQFxcHCUlJcTFxdGxY0caN27MJ598gtlspkePHuzdu9fXIQohhBDVSqfTYTabb+jcO++8k5EjR7J371727dvHokWLWLRoER999BFxcXF8++23NGzYkPT0dLp06cLevXuJjY2lYcOGmM1mmjZtyrJlyzCbzWzatIkPP/zQ3UM+btw4Pv30U+6++24AevfujdlspnHjxthsNrKzswkNDb1ibA6HA5vNRpcuXRg1ahSPPfYYixYtQlVVSktL2bp1Kz169CA8PByz2UzHjh0JDQ1lz5492Gw2unbtiqIomM1mmjdvTnZ2NhaLhSFDhjB58mT69u1L7969GTRoEGazmUmTJvHZZ5/RvXt3PvvsM/7whz9gNptxuVz06dMHs9lMYGAgYWFhpKens3HjRoYMGYLRaMRsNjN06FBeeOEFd09/r169sNvtADRt2pSsrKwbfp+gbKjP9eQ5VZpoL126lLlz5/LEE0+427Zv306vXr3cb+rQoUNZu3Yts2fPrvT7t2/fHqPRWOnXretUVeXJJ5+kVatWnD59mujoaLp27UpJSQn169cnMDDQ1yEKIYQQ1erYsWM3NCRh79697N+/nxkzZjB8+HCGDx/Ok08+yahRozhw4AD33nsvq1evpnHjxkyZMoXAwEBMJhN+fn7u+/n5+aEoCgEBAaiqir+/v3ufwWBAVVX3dlhYmEecl16nPDqdDoPBQEBAAE8++SSTJ09m8eLFKIqCn58fOp0OnU7ncQ1FUdznBQYGuvcZjUb3sa+99hpJSUls376djz76iG+//ZbXX3+dSZMm8fbbb3Po0CEsFgv9+vUDQKPREBoa6r6WVqvFZDKh1Wrd8V2gqip6vR6dTkdQUJB7n16vR6/X39TQEYPBQKdOndzbVqv1qh27VVre74UXXqB79+4ebZmZmURFRbm3o6OjycjIqMowRCVTFIVOnTphMplo06YNERERNGjQgFatWkmSLYQQQlyH8PBw3nnnHfbs2eNuy8rKori4mJYtWzJ06FCOHTvGt99+y8SJE695vb59+/Lxxx+jqio2m42lS5fSp0+fSonVYDDwyiuv8MEHH7hXwOzduzdbt24lJSUFwD2G+tJk9HK5ubn079+f0NBQ7r//fv74xz9y6NAhoCzxHzNmDE8//TRTpky5ZkyJiYl888035ObmArBs2TJCQ0Np1KjRzb7cSlHtkyFdLpdHHUJVVaXupBBCCCHqpCZNmvD222+zYMEC0tPTMRqNBAUF8be//Y2mTZsCZU//s7OzCQ8Pv+b15syZw/PPP8/o0aOx2+0kJiby8MMPV1q8TZs25cknn2TOnDkANG/enLlz5zJ79mycTicmk4l3332XoKCgK14jPDycRx55hPvvv9/dK/3888+790+YMIGlS5cybty4a8aTkJDA/fffz3333YfL5SI8PJyFCxei0dSMpWIUVVXVqr7JoEGD+Oijj4iPj2fFihXs2bPHXSLm7bffRlXVSh06cqEbX4aOCCGEEKI6HDt2jDZt2lT6dUtKSrjnnnv461//SufOnSv9+jWNqqq89957nDt3jnnz5vk6HC+Xv8/XyjmrvUe7T58+vPnmm+Tm5uLn58e6deuYP39+dYchhBBCCFGjbdmyhT/96U/cddddVZZk79y5kxdffLHcfT179uTpp5+ukvteyeDBg4mOjuZf//pXtd63qlR7ol2vXj0effRRpk2bht1uZ9KkSXTs2LG6wxBCCCGEqNESExPZvXt3ld6jV69efP3111V6j+uxceNGX4dQqaol0b78hzZ69GhGjx5dHbcWQgghhBDCJ2rsypB1wfLly1m7dq2vwxDAsGHDmDBhgq/DEEIIIUQtUjOmZNZRa9euJSkpyddh1HlJSUnyhUcIIYQQlU56tH2sZcuWLFq0yNdh1GkPPfSQr0MQQgghRC0kPdpCCCGEEEJUAUm0hRBCCCFqmdTUVFq1asW2bds82gcNGkRqaqqPoqp7ZOiIEEIIIYQPuFwqm/en8vXmn8nOtxAZamJsv2b06xKPRnPzq2br9XqeeeYZVq5cSWBgYCVELK6XJNo+NGbMGF+HIJD3QQghRPVzuVRe/O9uDiRlYbE5AcgvtvL2lwfZ9lMaf7nvtptOtqOjo+nTpw8vv/yy1+KA7777LitXrkSr1ZKQkMDjjz/O+fPnmT17Ni1atODYsWNERETw+uuvExAQwNNPP83JkycBmDp1KiNGjGDw4MF89913BAYGkpqaykMPPcSiRYvKvUZoaCjff/89r732Gi6XiwYNGvDcc88RGRnJoEGDGDNmDFu3bqW0tJSXX36ZoKAg7rvvPjZu3IhGo2HXrl289957zJw5k3fffRe9Xk9qaiqDBg3C39+fDRs2ALBo0SIiIyOveq8Lq5Xv2rWLt956i8WLF/Phhx+yYsUKNBoNHTt25Lnnnrupn/0FMnTEh0aNGsWoUaN8HUadJ++DEEKI6rZ5f6pHkn2BxebkQFIWmw+cq5T7PPXUU2zdutVjCMnmzZvZuHEjy5YtY8WKFZw9e5YlS5YAcPz4caZPn87q1asJDg5m1apV7N+/n4KCAr766isWLlzInj17CAwMZMCAAe6qXV999RXjxo274jVycnL461//yttvv82qVavo2rWrRzIbGhrKl19+yZQpU1i4cCGNGjVyJ8MXrn+hDO/BgweZN28ey5Yt45NPPiE8PJzly5fTqlUr1qxZc817Xc7pdLJw4UKWLVvG8uXLsdvtZGRkVMrPXxJtIYQQQohq9vXmn72S7AssNidfbzpVKfcJDAxk/vz5PPPMMxQXFwNly66PHDkSPz8/dDodEydOZMeOHQBERETQtm1bAFq0aEFBQQEtWrTg9OnTPPjgg6xdu5YnnngCgIkTJ7pXlVy9ejVjx4694jV++uknOnbsSHx8PACTJ09m586d7jgTExPdx+fn57uvv3LlSkpLS9m5cyeDBw8Gyiq2xcbG4ufnR1hYGL179wYgLi6OwsLCa97rclqtli5dujBp0iTeeustpk+fTr169W7q536BJNpCCCGEENUsO99yU/uvR9++fd1DSABcLpfXMQ6HAwCj0ehuUxQFVVUJCwtjzZo13HPPPZw+fZrx48dTWFhIjx49yMzMZN26dcTHx7uT0/Kucfk9VVV13/PScxTl4nCZYcOGsW3bNr799lv69evnPkav13tcS6vVemxf616qqnq8ZoB//etfPPvss6iqyowZM9i9e7fXz+hGSKIthBBCCFHNIkNNN7X/el0YQpKZmUmvXr1Ys2YNFosFh8PBsmXL6NWr1xXP/e6773j88ccZMGAAc+bMwd/fn/Pnz6MoCuPGjeP555+/5urKnTp14uDBg+6KJ59//jk9e/a86jl+fn7069ePV1999bpWb77avcLCwjh16pT7dQHk5uYyYsQIWrZsyR/+8AcSEhI4ceJEhe93NZJoCyGEEEJUs7H9mmEyaMvdZzJoGdu/eaXe78IQErvdzoABAxgwYAATJ05k5MiRxMXFcc8991zx3H79+mEymRg5ciS/+c1vGDNmDK1atQJg5MiRlJaWMmTIkKvePzIykueee47Zs2czcuRIdu/ezbx5864Z98iRIwkMDKRTp04Vfq1Xu9fvf/97XnjhBSZOnEhQUBAA4eHhTJ48mUmTJjFhwgRsNhsTJ06s8P2uRlEv9J/XIlarlcOHD9O+fXuPxxdCCCGEEFXh2LFjtGnTpsLHl1d1BMqS7M4toyql6khVc7lcfPbZZ5w+fZo5c+ZU+vWdTicLFiwgIiKC6dOnV/r1b8Tl7/O1ck4p7yeEEEIIUc00GoW/3Hcbmw+c4+tNpy7W0e7fnH6d69f4JBtg9uzZnD9/nn//+99Vcv2JEycSFhbGO++8UyXXrw6SaAshhBBC+IBGozCgazwDusb7OpQb8q9//atKr//VV19V6fWrg4zRFkIIIYQQogpIoi2EEEIIIUQVkERbCCGEEEKIKiCJthBCCCGEEFVAJkMKIYQQQtRCa9euZdGiRTgcDlRVZezYscyYMcPXYdUpkmgLIYQQQviAqrooPrKVgl2rcBTloAuKIKTnaALb9UVRbm7QQUZGBi+//DLLly8nLCwMs9nMvffeS5MmTRg8eHAlvQJxLZJoizrNnpuG+eQe9KH18G/RHUVT/ipdQgghRGVSVRcZX/6D0tMHUe1WAGzmArK/eRfzsR3Um/T4TSXbeXl52O12LBYLAAEBAbz00kvs27ePKVOmsGTJEgCWL1/OwYMH6dSpE1u2bKGgoICUlBQSEhJ49tlnAXj33XdZuXIlWq2WhIQEHn/8cc6fP8/s2bNp0aIFx44dIyIigtdff53169ezc+dOXnnlFQDefPNNjEYjVquVtLQ0zpw5Q25uLo888gg7duzg4MGDtG7dmgULFqAoyhXvNW3aNDZu3Oi+JsDDDz/M008/zcmTJwGYOnUqd9555w3/zKqCjNEWdVbp2cOkLHqU3A3/JePLv5O58g1fhySEEKKOKD6y1SPJvkC1Wyk9fRDzkW03df3WrVszePBghgwZwqRJk/jHP/6By+Vi8uTJZGVlkZycDJTVqp4wYQIA+/fv54033mDlypV8//33nDhxgk2bNrFx40aWLVvGihUrOHv2rDtJP378ONOnT2f16tUEBwezatUqRowYwY4dOyguLgZg9erVjB07FoCkpCQWL17M/Pnz+ctf/sLMmTNZvXo1R48evea9yrN//34KCgr46quvWLhwIXv27Lmpn1lVkERb1FkFO1eC0+HeNh/Zij0/w4cRCSGEqCsKdq3ySrIvUO1W8netuul7zJs3j40bN3LXXXeRlpbGnXfeyfr16xk/fjwrV64kLS2NnJwcOnXqBECXLl0IDAzEz8+PBg0aUFBQwM6dOxk5ciR+fn7odDomTpzIjh07AIiIiKBt27YAtGjRgoKCAgICAujfvz/r169nz549NGjQgHr16gGQkJCATqcjLi6OqKgomjdvjk6no169ete8V3latGjB6dOnefDBB1m7di1PPPHETf/MKpsMHRF1lupyldOoVn8gQggh6hxHUc5V9zuLsm/q+j/88AMlJSWMGDGCiRMnMnHiRJYuXcqXX37J3LlzmTFjBgaDwd3bDGA0Gt3/rygKqqriKudvpcPhuOLxULZ0+jvvvEN8fLy7txxAr9e7/1+n805Br3SvS699oU2n0xEWFsaaNWvYtm0bmzZtYvz48axZs4bg4OAK/Yyqg/Roizor5LaRcMn4N/+WPdCHxfgwIiGEEHWFLijiqvu1QZE3dX2TycQrr7xCamoqAKqqcuzYMdq0aUP9+vWJiYlhyZIlHol2eXr16sWaNWuwWCw4HA6WLVtGr169rnpO9+7dSU9PZ9euXQwZMqTCMV/pXsHBweTn55Obm4vNZmPLli0AfPfddzz++OMMGDCAOXPm4O/vz/nz5yt8v+ogPdqizvJv1oX6D/6DkqTd6ELrEdg2wdchCSGEqCNCeo4m+5t3yx0+ouiNhPYcfVPX79WrF7Nnz+bhhx/GbrcDkJiYyG9/+1sARowYwbp169zDOq5k4MCBHDt2jIkTJ+JwOOjbty/33HMP6enpVz3v9ttvJz8/H4PBUOGYr3QvnU7HjBkzmDRpEjExMXTo0AGAfv36sW7dOkaOHInRaGTMmDG0atWqwverDoqq1r5n5VarlcOHD9O+fXuPxxpCCCGEEFXhQm9xRZVXdQTKkmy/Jp1uuurI1TgcDp544gmGDRvGHXfcUanXVlUVu93O9OnTefrpp2nXrl2lXt/XLn+fr5VzytARIYQQQohqpiga6k16nKgRj2CIaYY2IARDTDOiRjxSpUm2qqokJiaiKMp1DeuoqKysLBISEujUqVOtS7JvhAwdEUKIOsZisWA2m5k7dy6vvPIKqqqyePFimjZt6q4KAHDvvffy6aefYrfbr+vxrxCiYhRFQ2D7RALbJ1bjPZWrVvK4WdHR0fz4449Vdv1bjSTaQghRx6xdu5bVq1dz+vRp/u///g+73U5RURH+/v7s3r2b8ePH8/e//52MjAxmzpzJ9OnTSUiQOQxCCHG9ZOiIEELUMaNGjcJgMNCuXTvGjh3LX//6VyIiItBoNDzwwAO0bt2aXr160b59e+Li4iTJFqKCauG0N3GJG3l/pUdbCCHqGK1Wy6xZs2jVqhWnT58mLCyMuXPnUlJSQkBAAAA9e/Zk5syZ7N+/38fRCnFrMJlM5OTkEBERgaIovg5HVDJVVcnJycFkMl3XeVJ1RAghhBDiJtntdlJTU7FYLL4ORVQRk8lEfHy8x8I718o5pUdbCCGEEOIm6fV6mjRp4uswRA0jY7SF+JWqqjhLinwdhhBCCCFqCenRFgKwnDtJ5tev4chLRx/VkHoT/oQhMt7XYQkhhBDiFiY92kIAWavfwpFXtpysPSuZ7P8t8nFEQgghhLjVSY+2qPNUpx17dqpHmy3zjG+CEbes5cuXs3btWl+HIYBhw4YxYcIEX4chhBDSoy2EotVjatjWo82vSUcfRSNuVWvXriUpKcnXYdR5SUlJ8oVHCFFjSI+2EED02D+Sve7fWNNO4deoHRF3PODrkMQtqGXLlixaJMOOfOmhhx7ydQhCCOEmibYQgC44gphJT/g6DCGEEELUIjJ0RAghhBBCiCogibYQQgghhBBVQIaOiDrLUZxHzvoPsZ47ialhWyJuvx+tX5CvwxJCCCGuy65du1iwYAENGjTg5MmTOBwO5s2bh6qqvPTSS7hcLgBmzZrF0KFDfRxt3SKJtqizsla+SenpgwAUH8pEtVupN/HPPo5K3KrGjBnj6xAE8j6Iuuunn35i7ty5tGnThg8++IAFCxag1WqZPn06I0eO5Pjx43z++eeSaFczSbRFnaS6nO4k+4KSX/b7KBpRG4waNcrXIQjkfRB1V1xcHG3atAGgbdu2rFixgrvvvpvnnnuOjRs30qdPHx577DEfR1n3yBhtUee4HDYsKcfQhcd6tBuiG/koIiGEEOLmmEwm9/8rioKqqkyZMoWVK1eSkJDA1q1bGTNmDFar1YdR1j2SaIs6xZaZTMpbj3D+47k48jLQmAIB0EfEETV8lo+jE0IIISrPlClTOHbsGBMmTGD+/PkUFhaSlZXl67DqFBk6IuqU3M1LcJrzyzZUF6rdSvzDb6IPj0VRFN8GJ4QQQlSiP//5z/ztb3/jtddeQ1EUZs+eTXx8vK/DqlMk0RZ1irMo12NbddpRFEWSbCGEELesnj17snr16nK3ly9f7quwBDJ0RNQxge0TPbaNcS3QXzZWWwghhBCiMkiPtqhTQnqMRNEZKUnajT4ijtDe430dkhBCiBpm+fLlrF271tdhCGDYsGFMmDDB12HcMEm0RZ0T3GUIwV2G+DoMIYQQNdTatWtJSkqiZcuWvg6lTktKSgKQRFsIIYQQojZp2bIlixYt8nUYddpDDz3k6xBumozRFkIIIYQQogpIoi2EEEIIIUQVkERbCCGEEEKIKiCJthBCCCGEEFVAJkMKIYQQQlxizJgxvg5BUDveB0m0q1BxiY3NB86hulQSu8QTHGDwdUhCCHHLcpYWUbB7NY78TAJa9yag1W2+DknUUqNGjfJ1CILa8T5Iol1Fikvt/GHBJjJzSwD4YuNJXn9sACGBRh9HVjdZUo+Tu/FjHEW5BLZPJKzfZBRFRk4JcStJ/2w+1vM/A1B8eDPa0HooigbVbkUXHEFYvymUnvmJ4iNb0QVHEDH4PkwNWvs4aiFEXSaZRhXZcuCcO8kGyCmwsGlfqg8jqrtctlLSP/8blpRjOPIzyN/6JYV7Kr7il8taivnELqznf6nCKIUQV2PLPOtOsi9w5mfgyDuPszgXa9pJ0pf+jYKdX+MsysF6Lon0pS/isll8FLEQQkiPdtVRVe8mH4QhwJp2CpfF7NFW+ssBQnqMuOa5tuxU0hY/g6ukEIDgbsOIHDazSuIUQlyZxhQIigZU15UPcjk9Ny3FWNN/wa9h2yqOTgghyic92lUksXN9osP83NvhwSYGdI33YUR1lz4yHhStZ6OiVOjc/B0r3Ek2QOHetdjz0iszPCFEBeiCIwjpOfq6zlG0egyR8ntXCOE70qNdRQL9Dbz22AA27UvFpar07xIv47N9RBcYhiG6IbaM0+620uSjuOxWNPqrvyeukiLvttJiCKv0MMUt5PP1J1jxwykURWHSoBZMHNTC1yHVCRGDpxHUYQC27BQK96/HcuYQoAAqGqM/oYm/wXL2CCUn96LxDyLyjgfQ+gf7OmwhRB0miXYVCvI3MKpvU1+HIQBUz0fKqrUEZ0kBmpDoq54W1HkwJaf2urcN9ZpgiG1WJSGKW8O+E5l8vPa4e/s/a47SslEYHZpF+jCqusMQ3RBDdEMC2ybgKMhC0ZtQtFoUnR5Fq8fZvj/5O77CUZyHxuDv63CFEHWcJNqiVrNmnMF6/mdMDdphy0x2txtimqG/RpKtqmW9ZCE9x+AoykUfEUdA694U7fsWbVAE/s27omi0V72GqH1OnMn1ajt+JrfCiXZmbgnBAQZMRvn1e7N0IVEAqE4HRQc3Ys04Q+nP+3AUZAFgPrKF6HGPEdguwZdhCiHqMPlNL2qtgt2ryVn/4a9bCgG/9oAZIuMJ6zflmudnrXyD4sObAdAY/Qlo3Yu0/z6NaisFwL9FD2LufKqqwhc1VNumEV5t7cppu1xekYXnP9hFUnI+fkYtD47pwNBejaoixDona/Xb7s+q175v3sG/eVc0Rr9y9wshRFWSyZDVIK/IwqnUfFyuitUdsVgdbNyTzLc7z1Jcaq/i6Gon1eUkb/Pnl7ZgyzhD/ftfJGrUb9EFXz0xsmWnevzhdllLyP3uI3eSDVBy8kesGWcqOXJR03VqEcX0Ue0IDTQSFmRk5rj2tG1y7UR7yboTJCXnA1BqdbJwxU8UFFurOtxaz1laTPGRrVfcr9pKr5iECyFEVZPyEfIAACAASURBVJMe7Sq2bONJFv/vGE6XSv2oAJ6b1YfosCuPG7TYHDz2+mZSMsom4S1Zd5wFjw4gNEgmUl4X1YXLYfNoup56umo5x154HO1xnEO+CNVFEwY2Z8LA5td1Tmpmsce23eEiI7dEJknfrAoM33JZzdc8RgghqoL0aFeh3EKLO8kGOJdlZumGpKues+PQeXeSDZBdYGHjnuSrnCHKo2j1BHce4tEW3G1Yhc83xDbDGHt5IuX5RMJYvxXGuOtLtkTd1aNtjMd2ZIiJpvVDfBRN7WFNPe5dW/uS5Fsx+BHQtm81RyWEEGWkR7sKZeeXupPsCzJyLq4WmZlXwuGfc2gSF0xokJE9RzNIzvAuJ2d3XmWBBnFFEXc8gDG2WdlkyMbtCWzdu8LnKopCzF3PcPbV+7k0wVb0RkJuG4U2MJygjgNQKliPW4gxiU2x2Z1sPXiO6DB/7h3RBp1W+jpulj0vw6stqNMgNAYTqstFcNc70IdefeKzEEJUFUm0q1Cz+FBiIwI4n3PxsWVCpzgAfjyazt/+sxuHsyyJ02kV9/8b9Vqs9rJydEH+BgZ1a1jNkdcOikZLUKdBBHUadN3nWlKOk/XNO2UL21yyymdAmz6ED5hamWGKW1Ch2caG3clYbQ4GdGtAbGRAucdZ7U5+SS2gfnQgwQEG7hzSkjuHtKzmaG99Jb8cJGfd+zjyszA2bFM2ZMthI7jbMPxbdCX3OwPqhaFiioagToMx1Zfa5kII31NUtZy1wm9xVquVw4cP0759e4zG6h3/6HC6PHqpMnJL+Hz9CTJyS+jbuT7DezcG4NEFP3AqteCK15kwoDmB/noGdmtAZKjMlq9OztJikt+YefEP96+Cuw0jfPC0ay5yI2o3i83B7//5g/sLtJ9Rx4JH+1M/KtDjuJMpecx7fycFxTb0Og2/u7MzA7s18EXItzSXrZSzbzyEai0pd3/MXc+g0RvJ3/EVqtNOcLfhBLTsUc1RitooLbuYz749QWZeCX071WdU3ybyFFN4uVbOKT3alSQtq5hXPt1LUnI+zeNDeGxqNxrUC6JeuD+/n9zF6/hrVRNJ7Fyf5g1CqypccRWW5KNeSTaAIaapJNmCPccyPJ5SlVodrN91lvtHtfM47j+rj1JQXPbvyO5w8d5Xh+jbqT56nQwXuR62rJQrJtkAJaf2EnnHg8Q0aFONUYnazul08czCHWTmlv3bO3o6F61WYUSfJj6OTNxq5Dd+JXlj6QF36a5TqQW8vmT/VY9P6Bh3xX2tG4V5JNmnUvLZcSiNEotUuLheLosZW1Yy6uWTpS4/zmHDlpOG6nKij4wv/yCNBtdV/uCLuqG8cdV6nXfli8w8z38rRSV2LDZHlcVVWykGE4rBdMX9hogrfF4BZ0khBXv+R2ny0aoITdRiP58rcCfZF2z/Kc1H0YhbmfRoV5Kk5DzP7ZS8KxxZ5r6Rbdm8/xxZ+RfrMrdvFkHv9rHc3vPiIhbvLDvIN9vPAGXjtV/6bQINY4IrL/BarHD/BnLW/RvVYUMfEUfMlDnoQ+t5HVdyah+ZK1/HVVqMLiSamDufImzAVPI2LXFXM1D0RrJXvUXO2veJGv1bAtv0qe6XI2qI7m3q0bxBKKdSyr5YhwUZy114JrFzfb747qR7u3OLKIL8DdUW561OdTnJ/GoB5mM7gLLqIarTgT48BntOGrhcBLTuecU5GOaf95Px+QvuORamBm2Im/Z8tcUvbm1RoX5oNYpHQYPYyMCrnCFE+bTPPvvss74OorI5nU4yMzOJjo5Gp6ue7xKHf8kh/ZKKIh2aRTK4x5UnMSqKQt/OcVhsTvxNOsb1b8YjEzrSunG4+9Fyeo6ZVz/d5z6nrGJBGgEmvQwruQaXxcz5T+e5h4C4SotwlRYR0LqXx3Gq6uL8J3NxmcvGy7usZmy5aUSP/D9CbhtNQNsEbFkpOPLSf72wg9KzRwjpORpFkQdCdZFGozC4ewMaxQbTtXU0D0/oSGiQd49r+6YR+Bl1qKpK7w6xzBrfAYP+2jWfRRnzsR3kb/niYoPTQfT4x4gaNpPg7sMJuW0UwZ0Ho1yhjnb6Z/NxWS7WLncUZmNq1A59aDQlP++ncP96XJZi9JHxMu5WePEz6jAZdRw6lY1LVWkYE8Ts33TG36T3dWiihrlWzik92pXkD5O78PaXBzl6OofWjcL57aRO1zwnIsSPqUNbkZFTQrP4UK9f9sUl3kNFCs023v7yIAF+ehI716+0+GsbR3Ge1zhrW6Z3PXLVZsFZlOt5XMZpVKcdjdEPY73GOAuzPfa7SgpxWUvR+knvRl1w9nwhOp3GY7KjXqe95udPq9UwYWALJgyU6hc3wn7hy+0lbJlnoU3vCn32XBbvRWrs2eewZZwhZ/2H7rbg1BFE3vHgzQUraqVx/ZsxqHsD8gotNIwJki9k4oZIl1wliQz1Y+6MXnz+wkjmPdSb6HB/7A4XS9af4Km3t7JwxU8UlXgmft9sP80D89fx+JtbmPHCes6mF3rsbxYfcsUFLXYd9v4jJC7SR9T3GmttyzxD0cGNHm0aoz+mhp6T2FylxaQuegzHr73cAZfV3zY17iBJdh1gtTt55t3tzP7n9zz80ne8/NGPuFy1rkhTjRXQoofXqo8FP67Blp3q0eayWTAf34Ul9Tguhw3Hr1+MA9pcVjdf0RDQuhcFP67xaC7at15WeBVXFBxgoFFssCTZ4obJ0JEq9J/VR/jiu5Nk5ZWSlJzPz+cKGNS9rLxXicXOXxftwO4oGwNcanWQW2ihX5eLyaGiKCR0jOWbbafdNbYvSOgYS4fmkdX3Ym4xiqLg37wbhXu/9Vg1zpJ6gtDe4zyO9WvWBUvqcZxFOe42V2kRhXv+hyGqAcHdhpY9nna58G/RnchhD0n1kTrgux+TWbnlF/d2ckYR2QWldGgWKUNAqoE2IARtQAglp/ZebHTaUZ0OdEEROEsKUB02zr3/J4oOfkfRwY3k7/iagl0rMSf9SNTwh1CdDuwFWeiCI4me8CeMUQ0p3LcOV+nFhcEUvYHQPuNRNNLvJIS4frfs0JGTJ0/y5ptv4u/vz+jRo0lISPB1SNdt22UzlA8kZWEutRPgp6fQbMNqc3rsv3Ri5AUmg869eM0FWo3CmH7NKj/gWkZj8AOXZ5UHV6n342RdYCgBrXtjPZfk0a46bGT/byENf7eIsMQ7CUu8s0rjFTVLZp7353HD7mTOnC/k1T/0kx6uamCI9p7nYk7aTdGBDQDoQqJwFl8y8fzXz7st4zS5339C9Ng/EDlspsf5YX1/Q+bXr3NhxdfQ3uNRtDX2T6EQ4gpKTu4l5/uPcZUUEtRpEGED7qqRc6dq7G+XkpISnn76abRaLa+++uotmWjXC/cn65I/1qGBRnYdPs/XW35Bo0BMhL/HBEqDTovN7vToLTPotdzWNoZdRy4OFRnaqxEBfjIh41ocl/RQX6ANLH8oTmDbBPK3fek1rtNpLsBlLUHrL5Ve6preHWL5cuNJr+Eip1LySUrOo1Wj8Apf6/u9Ke5rjR/QnDt6elcpEd6M9VtiqNcEW8bpX1sUXCUXh9g5CrKueK4tK6Xc9sD2iRiiG1F69jDG2KaY4ltXZshCiGrgKM4nY9k/UJ1lw77yty9HF1qP4C5DfByZtxqT+r///vvcf//97v86dOiAxWLhd7/7HYmJib4O74Y8OLo9oUFlQwxMBi2jE5uyYMl+fjlXwKnUAjJySzAZLybVx87k8vHa417XeWxqV0b1bULbJuFMvaMVM8a2r7bXcCszRDdEH+5Zr9xlKaFw/wavY3XBEdR/4O8YYj2fFBjjW0mSXUc1jw9l3sxe5a7MajJWvI/iVGo+Cz7bR3J6EamZxby59ABHfvH+Eii8KYqG2HvmET7wHoJ7jCCoc3ml/Mp/suDfvNsVr2uIbkhIjxGSZAtxi7KmnnAn2RdYzh72UTRXV2N6tGfMmMGMGTPc24cPH6Zx48YsWbKEBx54gBEjRvgwuhvTvEEoH8y5g7PnC4mLCuDrTT977FdVsFg9h4XsPpLOA6MvTs5zOF28u/wnNu0/h16noXubeuUujiG8KYqGmLueIfPr17CmngBAtVvI/uYdCvasxRjTmPD+U9AFl4111/qH4NegLU5zAarTgV+TTkQMuhco69nO27IUW+ZZ/Jp2JrT3OHncXAd0bhnNszN78eRbWzH/upprv871aXQdtewPJmVdKOV8se1kFu2aRlRmqLWW1hRAaJ/xQFkvddGhTeAsGyKi6I1EjZ5N6emfQNHgKMjCUZBJQMsehCVOwuWwkb/1SyzJRzHGNScs8U40Rn9fvhwhRCUwxDYBReMxB+vyjrKaosZmClarlf/3//4fgYGB9O/f39fhVIjLpfLLuQJCg4zuXjC9TuOued0o9tp/nEutnmOKN+xO5vu9ZbPsrTYnH31zjK6tomkWL3W0K0IfGo0hsoE70b7Annkae+ZpbOm/ED/zVQCyVr+F+fhO9zEavQFdUBgAGcv/ieXX1eUsKcdQ7RbCB95TTa9C+FKjmGAW/WUIe46lExHiR8frnIRcXuWgJnHlD2ESV2eIakDc3c9S8OM3KBotwbeNwhTXnMA2fXBazBTuXoPdPwhjXEsUrZ7sb96laP96oOxza8/PJGbSEz5+FUKIm6UPiSZyxMPkfv8xLouZwPaJhHQf5uuwylXliXZxcTFTpkzh3XffJT6+rKLGqlWreOedd3A4HNx3333cfffdXud169aNbt2u/OivpskttDDn3e2kZBShUWDS4JbcO7yNxzG92scyvHdj1u06i6LAkNsasnbHWY9jVDy7vk6nFXjd63RaoSTa18EQ2wwOeA8XgbK6vPbcNBzF+R5JNoD52HaiRjyM01zgTrIvKD62QxLtOiQ4wMCg7ldegOpqurSKZtKgFqzc/DMuFUYkNKZX+5hKjrDuMDVog6lBG6/29CXPuyc0Fx/aREjPMe5VJS8oSfoR1emQp1FC1ALBnQcT1GkguJwo2po7b61Kf9scPHiQOXPmcObMGXdbRkYGCxYsYPny5RgMBqZMmULPnj1p3rx5pd//8OHqG6+zdm8+KRllq5C5VFi6IYkYv0LCArWk5drRaRUy8+2cz7DQv0MQ3ZoFEGBy8tMJPWm5F8cZNY7SsndvWTmr87k2bGaLx300GlAs59m798qTgMRF+owTBBz8GoVfawxodCiXVCJRtXoOHzxI8M7/eI30tBkCy94Ll5MQvR8a+8WJraUaP/f7JGo/VVVJzbah0SjUj7j+ZdTbx0CrCbGoqBh0Nvbt23ftk0SFaYqzCbmsalDBrpW4tAaPiUhOYxD7Dhys3uCEEBXjsGJIO4LicmKLbYtqDPB1RJWiQom20+lkyZIlbN26Fa1Wy8CBA5k4ceI1z1u6dClz587liScuPqrbvn07vXr1IjS0rEd26NChrF27ltmzZ9/gS7iy9u3bYzRWT73jbw/tBoo92kLrNWLJ+iROnM3zOt5GIE9O68ZzTUp476tDnE4roEuraB4Y3Q6jQceL/9nNriOZAIQFGdHpNIQEGLhraGtuayu9YRWhqirJb76L89fEWgH04bGoTjuOvHQUnYHIoTOItJWSc1kZQEVnoMHY3+L362I2xf6zyFrzDqrdijYwnPrjZ2Os17iaX5HwBYvNwV8X7uDYmbIVRDu3jGLujF7otDVmLnmd5yjKI3nb+x7jNQE0ThuK0Q/VWgp6E8FNOxBwZAWGmKaE9ZuM1lQ7/pALcatz2Syc+/efseeeByAoZTf1H/wnuqCKV3fyFavVetWO3Qol2s8//zynTp1i7NixqKrKsmXLSE5O5tFHH73qeS+88IJXW2ZmJlFRUe7t6Ohofvrpp4qEUaP17RTHjkPn3dsRISZSM4rKTbIBtv+URnGpnXrh/sx5oKfHvt1H0z3K+eUVWQFo0yicbq3rVUH0tUtp8lHyty3DUZzntby6s6SARn/8N/asFLTBkWhNAZhP7PK6RsTt091JNkBgu0T8m3XFnpeBIbqhPHquI0qtDj5ff8KdZENZPfydh8/Tt9PVl2AX1UcXFEZIrzEU7PjKe19wFBq9AY0pCPPRrUDZeG1HXjoxk5+u7lCFEIAl7RQlp/ZiiKhPQJvemE/scifZUFaAoOin7wlLmIgl5Rglp/aij4wnsF1i2QJyt5AKZQvbtm1jzZo16PVlY2DGjBnDmDFjrplol8flcnks9KCqaq1Y+CEy1A+TQYvF5kSjUZg0uAVZud4LXlxg0Gsx6MrvEcsrtJTbvvnAORK71KdX+9hKibk2chTlkv7ZfFSHrdz9gW36oCgaDNEX6xj7t+hOQOte7jHafk07YajXhJwN/0FjDCC46x1oA0LQmAIwxjatltchfG//iUxe+uhHSiwOr325V/iMQtmqr0s3JPHLuQI6tYhiXP9maKX3u8qF9pmAszCX4iNb4JK5Lvas5HKPLzm1D5fdKqu8ClHNio/vIHPZK1z4nAae2otf087eB6oqxUe2kPnVa+6m0l8OEj32D9UUaeWoUKIdHh6O0+l0J9qKohAcfGO1hWNiYtizZ497Oysri+jo6Bu6Vk3y75WHsfy60qPLpfLFhiT+3/SefLX5Z68FLwCmDm3ttYyz1e7E6XRxW9sY/E1Hyv0Df+kCN8Jb6emDV0yydeFxRNx+v1e7otFSb+Lj2LJTweVCdTk495+/uEuIFR7cSOw98zCE3vr/TkXFvff1oXI/gzqtQrP6oThdKlqNdyfBPz/Zy49HMwDYn5RFbpGFKbe3wqjXYrbYySmw0DQuBE0554obozodpP33aezZZRWaUDQYYppiO3/qiudog8JQdDV3ApUQtVXBrtVc+mW4+PAWwvpNRh8e6+7V1gaEENRxIBnL/uFxbvHhLUQMuR9twK1TualCiXbr1q2ZOnUqEyZMQKvV8s033xAWFsaHH34IwPTp0yt8wz59+vDmm2+Sm5uLn58f69atY/78+TcWfQ1y+XLNeUVWGsUGM39Wb77ZfgajXsugbvEUmu00jgumQb0gj+O/3HiSJetPYLc76dclnhceSWDJuhMeQ0gA9hxLZ2ivRvhdx4IZdcnlC9RcKqBl96vOTDZEllXFyV77njvJBnAWZJL69iMY41sTM+mJW+oDLm5ceUuwAzicKk+9vZWIEBOP39Pdox52qdXBnmMZHsev2vwLKzf/gl6nweF0oaoQFxnA/Fl9iA6Xms6VofT0TxeTbADVhdbk/bPVGP1xWUtQDH5EDp1ZI5drFqK28xrFoChoDH7Un/4yxUe2ojrtBLTtiy4w1HuYpkZTVj/7FlKhbM1qtdKqVSuOHDkC4C7Tl5SUdLXTylWvXj0effRRpk2bht1uZ9KkSXTs2PG6r1PTJHauz6otv7i32zaJwKjX0rF5FB2bR13lzLISfv9dc7F83A/7UmnTJJw5D/Tkh30pvPLJxQoFB09m89UPp7hrqKxoVh5TfCuCe4yk8Mc1Hu0avyBCe0+o0DU0Ru+VAAGsqcfJ3byEqOGzbjpOUfMldqrPhh/LH3YAkFNg4S//2kr/rvE8MqEj/iY9Br0Wo0HrsRDVhX4bu+PiRL20bDNL1p/g95O7VFX4dYqi964Eo49qiMYvCPPRbYBCYIf+RA5/CHt2KvrwuCt+zoUQVSu093jSU0+4Jy8HdR7i7sAK7jbU89g+40n/IglcTvd+rb9nR2VNV6FE+8UXX7ypm2zcuNFje/To0YwePfqmrlnT3HV7K9bvOusePnL0lxyOn82ldaOyGbPns828/NGPnMsqJirMjz9P7UbTX2thn04r9LrehbbgAO/xg7+UU1tbXBTac7RXog3gsprL/YAWH9lKyck9uJx2NKYgNAYTiikA1WL2OtaelVIlMYua5+GJHYkM9ePo6Ry0WoX9J7xLaqoq/LA3FX+jjkcmdkKrUejZNpZN+1PLuaKnjFwZBna9nCWFuCzFXk+uTA3bYWrUDsvZss4gjX8wId2How+LwTF4GqCgCy578mCsoavHCVFX+LfoRvxDCyg5tQ9DZH38mnW98rHNu9Fg1muU/HwAQ2Q8fk1uvY7ZCiXau3btYtGiRRQUeCZ4X375ZZUEdSs6eDLLnWRDWS/WeysO8cof++NyqTz+5mYKisvGDqdkFPPk21tZPG8YJoOODs0i0WnLHitf0KVlWS94q4Zh+Bl1HitGdmklY4WvRhcShT4y3uNRsqu0iHP/fpz6M19BH3Lx55e1+m2KDm4s7zLl8msmPZB1hVGv5e5hZU+O8ouszP7nRvdn+HI/7Etl3a6z1Av35647WrP3eDrFpd7juy/Vt9OVhzkJb3lblpK3dRm4HBjrtyJm8tNo/QKBskfRsVPnUnJyD87SIgJa9kTrH4Q99zyWlGNlyXWwLHkvRE1hiIx3D9e8Fn14HCFXGRZa01Uo0Z4zZw733nsvDRve2MpodYHN7vRqy8wvG+OZnFHk9QfaYnNy4kwebZtGkJFr5v8mduR/O85QYnEwvE9j+nQs+0cV4Kdn7oxe/HfNUfKKLAzq1oBhvRpX9cu55ZkadfAcswm4rCUUH95KWELZEJLsdR9cV5JtatiO0N7jKjVOUfOdTMnDoNPy3Kw+/GfVEVIzi8kuKEW9ZI7zhUmT57LMLPhsH7PGdyCnoKwyid3hYteR8/ib9IQEGrDZXfTtFMfwPk188XJuSbacNPI2f+7etp47QcGuVYQPuMvdpmi0BLS6WCq1+Oi2smoFvz6ejrjjAUJ6jKy+oIUQggom2hEREUybNq2qY7ml9e4Qy+tLD3hUGGndKAyAQ6fKX8VRp9Mw66UNZOWVoigwaVALpo1o63Vcu6YR/P13iVUTeC2l2sufyKYxlA3FcRTnUbjnf9d1zYDWvW65+p3ixlmsDv666OJCNXqdxj3OOibcH0WB9NwS/E16zKUXV3d1ulTeWfYT//h9Iq1+HTo2fXQ77xuICrNlnvVqs+edL+dIcBRmU5p8lLwfPvVYwCZv8+cEdxsmn2EhaihVddXKCcoVekWDBg3ik08+ITk5mbS0NPd/4iI/k54/TumC8deSffHRgTw4pj0AKzb97HX8pEEt2LgnhaxfKxuoKizbeJLMPBm3WRkCWvb0atOFxRDYYQAAtuwUr1XkrkYXWo+gDv0rKzxxC9jwY7LHQjWXTmZMzy3h3uFtWf7yaIb3bux1rgps/6n8RFBcv7zNS7za9OHe6wmYT+4h+e3fkvX16zgKPDs4VLsNj8cQQogaQXU6yPpmIWdensrZ1x6k6NAPvg6pUlWoRzsvL49XX30VP7+Ls7QVRWHfvn1XOavuGditAb3ax5JbaCEuMsBdwsZu90zojAYt00a0Yd77Oz3aXWpZJYMNu5PZtC+VyFA/7hvZlpYNw6rtNdQGJT/vJ2v122UbGh2G6IYEtu9HcNc7QFU5/+k8Sk97r0aqC4vBFN8aVBfWjNM4CnNRdHoCO/QnvP8UNDrvygai9srOv/KCUwAOlwudVsPk21tyKiWfAyc9Ezsp3Vc5LGmnvIaBAehCvOeq5G1aAq7yx8YHdR4sK7oKUQMV7ltH0f51ADjN+WStehtTw7Ye86luZRX6rfP999+zdetWIiMjqzqeW56fUUf9qECPtlGJTfj4f8fd26P7NkVRFAZ0a8De45nu9vpRASSdzeWzdSeAshJgz763kw+euR2TQf5AVISqqmT/byEuS3FZg8uBotES2rOsyk3B7tVeSbZ/616EdB16S85mFlWnb+f6fLXpZ5zlLDhVL9zfvUKryaDjuVm9WfDZPr7fW5YQdmweyeAeDdzHnzlfyKmUPNo1jSQ2MqB6XkAtodrK+cKj0eBfTqUCl/WyJ4KKQlCnwZjiWxEoT6SEqJGsaSc9G1QXtvO/1K1EOyIigvDw8KqOpdaaPKQVjWOCOfxLDq0ahdGnQywpGUX0aFOPx+/pxub954gK82PSoBa8sfSAx7lFJTZOpeTTvpl8yakQpwNHQbZH06VjOUtTjl9+Bhq9idxNn6FsWwaA6rAR2L4fId2HV22sokZrHh/Kc7N68822Mxj0Gvp1ieeXcwUY9FoGdW/gsWiUoig8NrUbd93RGpvDSaOYiyvnrtz8M+99fRgAjUbh8Xu60bdT/Wp/PbcqXWg0ijEA1Xqx3Gbk0Jnogryf9AV3vYPcjYsvnhsUiep0YIxtJmOzhaihTA3aUHx488UGjQ5j/Za+C6iSVSjRbtmyJVOnTmXgwIEYDBcfn1/PipB1Xc/2sfRsH0tGbgmz//k9KRnFGA1aZo7twJwHLo4nbhwTzL5Lerl1WoX46FurOLsvKTo9/s27UnJqr7vtwnhtVXVRkrTb65zicsaDWc8loTGYCOo4sMpiFTXf5QtOdW9T76rHX95b7XSpfPrtxS93rl+3JdGuuIwvXvJIsoO63FE2DKwcob3HoQuJKquNn/QjjsIsig/9gDlpNw0feUtWdRWiBgrqMgR7XjpFB79D6xdE+MB70AXVns7dCiXaFouFJk2acObMmSoOp/b7ZO0xUjLKhjVYbU4WfXWIvp3iCPArWxr8N0Na8ktaAQeSsgjw0/Pg6HaEBnkvWiOuLGrM78nb9BnWtFOYGrUjrN9knOZCsjf8x726VEUUHviOwHaJoCjSG1YH5RSUYjLo3J/NG+FyqVgvm6NRarl6fW1xkT33PLZMz9U5S88exllahNYvCEdxPvb8DFSnDWNkQ7QBIQS2TcCadoqLa3KCai2h+NgOQroPq+ZXIIS4FkXREDF4GhGDa2d1u2pZGVJAUnIe63edZcchz0oENruT7IJS9x/zQD8982f1oaDYir9Jh15XluCdSsnH6XLRsmGYe5KlKJ/WL5DIYTPd24X7N5D9zbtc+oe3Iqwpxzj996ngcuLfvBtRo2ej9Q++9onillZqdTD/g10cOlU2BCkuKoD5s/oQHVY2ufH4mVy+35tCaKCREQlNCAm88hdhvU7D7T0bsFLFCAAAIABJREFU8r/tZ9xtoUFGNu9PpV+Xii3WUJdpA0NRDCZUm8Xd5shNI/mNh/Bv3RPzkW0XqwdptESNfISgjgPRltMblrPhPwS07OFeIVIIIaqDoqrXrne0f/9+Fi1aRElJCaqq4nK5SE1N5YcffqiGEK+f1Wrl8OHDtG/fHqPR973Bp1Ly+fMbm3CWU00uJsKfRX8ZcsXk2eF0Mf+DXe7hJG0ah/PcrN4yOfIabDlp5G35HEdBNtbzP4PTfu2TFM2v5b/K/0gEdRpE1KjfVm6gosb54rskPvrmmEdb8/hQFjzan0M/ZzPn3e3uevkBfnrmTL/tqnMonC6V7/ck8+XGk5zLujgEYtqINvxmcO0Zh1hVsta+R9Heb6nIF2WNXyCN/vA+qsPB2dcfRLVbPfYH9xhJ5B0PVFGkQoi66Fo5Z4XqaM+ZM4cuXbpQXFzM6NGjCQwM5I47yh8jJ7x9s/10uUk2wJjEZlftod51JN1jzPaxM7ls2udd6kpc5Cwt4vzHz2A+shVr6vFrJtkavyD04bGE9p1EzN3PXvG40uSjOEuLKzlaUdOkZBR5tZ1KzcfpUlm/66zHolTmUjtPv7ONA0kXP6OHf87mw1VHWL/rLHaHC61GoW+n+pzP8ayI8e1O70VYhCd7fiZF+9ZR0adRrtJiXDbr/2fvvMOjKtP/fZ/pJZPeG0kIIfQeOgKKCFgQe+99dS37XcuurvtbXVbX3nZ1dW1rbwgqKoj03gKBUAKB9N5mMn3m/P4YmGQyk2QQQkhy7uvyupz3vOfkmZCZ8znv+zyfB5lai9zgv3Lt50oiIdEBbreIs72bt4REkAQltAVB4PbbbycnJ4eMjAxeeukl1q1b19Wx9RoC3bgBBGBkVkzAY8epa7QGNSbhwbR3HUdfuR2XqaGDWQKKiATkodEgyHBbjDjqymlY8zmO+nKU0YG39J31FRS9chtN237qmuAlzgjGDY73G+sXb0AuE9Br/PO1RbFFNL+7ZA+PvrGOr1cW8MrnO3n+Y09RrlwuQ6PyzfM/mdzvvoK1OL/dxlKC0n/lSJsxCrnWY68aMXmB33GZVioslwiOZZuOcv1ff+TSR77j5U93+DSsaosoigSRHCDRRwlKaOv1nkr61NRUDh48iEajQSbrfW0yu4KXP93BvqP1AY+JwGP/WsfewloKyxoDdoWcMDQBrbrlBq2Qy5gyUnIsCITbaadm6ZvgtPsdU4TFINMaUCVkEj7tCpwNFbiaavxu4qa8NcRd9gihY85DnZSFIjwWWhVCik47tcvfk1a2ezFTRyZx5ayBqJSe77i4SB0PXTMGgIvO6o9B5y+Q9Volb3+bx9crC3zG1+WWUdtoQamQcc3sbO+4Qi5wzXnZbS8j0QZ1Qn88SxItKKNTCJ+0gMQbFxIybDqKiHiU0SmETZxP3MUPeOfpsnL8rmcp2IoouqXPr0SHVNaZee2LnTSa7LjcIsu3FLF0faHfPFEUee+7PVz22Pdc/fhSvl3t3wVaQiKoRN/hw4dz//338/vf/5477riDI0eOoFBIOcKd8fa3eSzfUuQ3LpcJ3iYYDUYbT7y53utM0C/ewB+vG0vqMR/emAgt/7hnKovXHMLlEpk3OZ2UOGlVJhBusxG3tdlvXBXbj/grH0dhiMDttHP0xZvbbcVsK9pDyZu/J3ruHd6CytL3/+RJQTmG6LTjMtZ6V84keh/XnJfNNedlY7U50bTyy46P0vP2n2bx5H82etuzh+pVnDehH//36hq/68hkAgq5R7BfOK0/I7NiOFzWxNCMKKLDtX7zJXxRRScTfd5t1K36BLfdQuiIs4mafYvXBSj63Jsx5W/AZWlCP2AcMk2LvaKt7KDnIbm105Ago/j1u3E2VqNOyCR2wUMow3tHUwyJU8ehkgba9qk6WOy/S7p+Vzlf/ep5uLbh4u1v8xiSHkVmSvjpCFOihxCUWn7sscfIzc0lPT2dP/3pT6xbt47nn3++q2Pr0ZitDr5fd9hvfEh6JHuP3aCP09r+62iFkYdeWc1/Hp3ltfXLSArj/iv9u6BJ+KIIjUIV3x97RcuqgmHkOUTPvdObBy86bIE7zbXG7aJm6VuEZE9EptahHzjeR2grIxNRxqR0cAGJ3kJrkX0cnUbJs/dOZc/hWuoarYzOjkUQAj+7zWvjSpIaH+p9iJYIjtAxszGMngWi6GOzaa86Sun7j3kdSep//YjImdcRPnE+1uJ8Kj592nfHSq7EZW7CbW4CwFZeQO2y/xJ/2SOn9f1InPlkp0WikMt88rMDFTzvO1rnN7b/aJ0ktNvB7bBhPrgVQa5AlzkGQd65BLUW76Nu5Ue4TA2EDDuL8MmX9DjntaCEtiAIREV5CktEUSQsLIyYmI5zi/syuw/V8OEP+Thd/nfe6+YOZvGaQ6zfVR7gTA9Wm+fJ+A/XjunKMHsl8Zc9TN2qj3FUF6PNHENEmw+lXGtAlznGp6FNQI51mFTFphI2/nxApHnfRpQR8UScdSWCIKVO9XWGZPgW2503MY3v17VsL182cwDXzxsMeB683/t+L3mHahmYGsGN5w/u0BZQwhdBkLXNIKF+3Vc+tn8Adas/I3TsHIy7VvqlhUXNvI7aZf/1GbNXHumCaCV6OpGhGh69YRzv/7CXpmY7s3JSmZWT6j1eWWdmz+EaYtvsSgkCDM6Q7CMD4TIbKX3vEZz1FQCo4vuTeMNTyBSqds9xW5sp//Qp7+JY/apPkOvDCB0167TEfKoISmg/8cQTANxwww38+c9/ZurUqTz22GO8+uqrXRpcT6S20cKTb23AHqBwQqmQMSAlnPsuH0WkQcO+onqiwjRsyqvwm7tqRwljB8cxfbTktXsiKEKjiL3g3g7nxF78AJVfPY/l8I5258j0EShjPL97QZARPuEiwidcdEpjlehd3HHxMEYMiOFIeROjBsaQ3a/Fy/mNL3exaofHLai40kiDycZfbp3QXaH2CtzWAHnWTjui0xHQ716dmIkqNg171RHvmDZteBdGKNGTyRkST84Q/8LojXnl/OP9Ld70zzHZsRwqaUSllHHlrIGkJ0rdRwNh3PWrV2QD2CsOYd6/mZAhU9o9x1p6wG8H2nJ4Z+8U2nl5eXz55Ze89dZbXHzxxTz00EMsWOBf0d3XMVkcLF59OKDIBnA43ewvqmdY/2juWDAcs9WBw+lm4rBK3vgi1++8zXsqJKHdBchUWuKvfIyG9d/QuGkx7uOFUYIAMjmqmFRi598vrVpLnBCCIDBxWAIThyX4Hdu81/dhemt+JV+vLGDB9MzTFV6vwzDyHCyHc33G9IMmIteGEDp2Lqa9a3E2VHnHNckDibvkD9T8/A72yqNoM0YSNevGbohcoqfQ1GxnU145ITol4wbHo5DL+OjHfV6RDR7L3Y/+3xxvLYZEYESHv1uaO8BYa1TRycf6W7RoI1VMv1MeW1cTlNAWRRGZTMa6deu48847AU9bdokWtu+rYuH7m7Ha22/xLQiQGO0p1vnop3y+WlGAw+lmwtB4nvv9NH7//Eoft9ikGE+xXVOzHb1GgVz6IJ8yBEFGxORLiJh8CbaKQpzGWlRxaVgO7USmUqMIk1KjJPxxOF0crTCSHBMSMH+7PVLjDOwv8nUfenfJHuwOF1fOGniqw+wThAyahOxKLQ0bvsFtM6PPnkjY+AsAUBgiSLnjFSxHdiPTGtAkDQBAGZlAwpV/7s6wJXoIFbXNPPTyapqaPS5WQ/tH8fSdkzHbnD7zbHYXLreIQh7oKhLHCRk6jYaNi70r1HJ9OPqB4zs8RxEWQ/TsW6hd8T9EuwVd5hjCJlxwOsI9pQR1p0hNTeW2226jpKSEnJwcHnroIbKzJWuq1ryzJK9DkQ0wYkAMUWFalqw9zKc/H/COb8yrYEBqBNfOGcQnP+/H6XIzJCOKs0Yn8cjra9lzuJbwEDV3XzqcicMSu/qt9Aoc9RXIdWHI1O07O4huF46aUpQR8ch1oZS88wdvoZQqLp2kmxYiyFus3Bz1FQgKFYoA7Z0lej/5hXU8/d4mGk12dBoF/3ftWMYOigvq3DsvGc6jr6/1+45YvPqQJLRPAl3/Uej6j8LRUAWi2yffU1Ao0SQPxHI0D3vVUVSxPW8lTKL7+H5doVdkA+QdqiXvcA1zJ6bx3vd7veMzx6agVkoquzOUEfEk3/IsxtwVIFMQOmoW8iB87UPHnEfIiJmIdhtyXc90XAtKaC9cuJBly5YxZswYlEolY8eOZf78+QAcOXKEtLS0royxR1DT4O9kIdDSz0wQ4IZ5g7E5XLz/3R6/ucs3HeWtx2Yxd1IaJouD+Cg9r36+kz2HawFoMNl46dMdjMqKPaGVtL6G01hPxed/x15xGEGpIWrWjQHzuey1pVR8+jTOhkoEpRqZWucV2QD2ykLMB7ehz56A22Gj8stnjm1TC4SOPpfoObefxnclcSbw1re7aTR5brxmq5N/fZXLO39uv0Pu7kM1bNhdTnykjuGZ0Tic/g/idofUde5kEN0uqhe/immPx1pRl5VD3IIHEeRKbBWFlH/0pDeXO3T8BeizxqEMj0cRKhWsSXSM3RH483rJzAHER+nJPVhNRlKYT5GkRMcoIxOJnHHtCZ8nU6igg6LJM52gFJtOp+Oii1oKwa666irv/z/wwAN88803pz6yHoTLLSIP0MBHBGLCtWQkhXH+lHQyk8N59sOtPnZ+x6lptLJyWzFZ/SJIjPakjBSWNfrMMVudVNab6SfZg7VL/ZrPsVd4bBVFh5Xan95BP3CC35Nw3YoPcTZUHptnw+Ww+V1LdHm2CI07f2mVCyrStP0n9EOmoE0d3HVvROKMo7LW16O9psGC0+UOmJu5YXcZf39vi/e1QackUCdn6SZ9cpgPbvOKbADzgc2Y9q7HMOwsGtZ96VMw2bRpCU2bloAgI3LGNYRPnN8dIUv0EM6bmMbyzUXe2qmUuBBvJ+fJIxKZPELaXe5K3NZm6td+ia3iENp+wwifNN9nh7kncdJLo1LbUcg7VIPR7N+NEMDudPGnm3LYdbCG79cVsnZnacB5Dqeb5z/ejkyAey8fxTk5qYwaGOtjkh8boSU5tmdunZwuHLW+v1/R5aD4X/cQPuVSNMnZVH//L5xNNeBytnOFYwgy6jd+i7V0v2d+G5z1FSAJ7T7FpOGJ3lbrgLc4KhA/rD/i89podvjNmTIikVsvGnpKY+xNiC4nDRsWYcpbjdtmRhWTSsS0y9Ekt6QtOhr8HZscNZ7vAJfF2M6F3dSt+gTDiLN77Fa0xMnjdLn5acMRCkoaGZYZzYwxyT5WsOmJYbz04HRWbi8hRKtkVk6qVPB4Gqn69mWvDa/16B5cliaiz72lm6P6bZy00O5pxuFdQUf5WWkJoTz/0TZW7QgssAGfZhduET5cupdzclKZNCyBqnozewvrSIzWc8uFQ5HLpN93R+iyxmIt8k3NcVubqVv+vl/1coeIbhwVh3FUHPb8A7VGrkSbMfIURSzRU7h9/jDCQtTsLqhhQGo4V5/rW6dSUNzAD+sLkcmEgGkiGpUcp8uNyy0ybWQy9181Sipw7oC6VZ/QuGGR97XFVI+1ZB8pd7+OIiQCAP2AsdQt/wBalZHbqz0PQ4YRZ2M96p+mB4DLicvcKAntPoLV7uTNr3ezYXcZ8dF67pg/nGWbj7Jss6dz8/ItRVTWNnPVbN/PdEqcgevmDOqOkPs0bocNc8F2n7Hmvev7rtCW8HSRGj0wlu37q/yO1dRbyD3ovyLamrabAs0WJ0+/u4mNx/y1R2bF8MQt41FKZc2dEpZzPqLTScO6r/zthIIV2W1p8w+kHzheKojsg6iU8nZvuqXVJh5+bY13m1ml8BfQoXoVLz80A5fLLTWrCYLmfRv9xkSHDcvhXAzDpwOgCI/zLYYBLEWeQjVd/9Fo0oZhKy9AptLhMtZ656jiMzzWYRJ9gk9/3s/yLR5Rfaikkaff24Sx2XcX+vNfDnLJzAGopMLGbkdQKJGHhOMytTg1KcJjuzGik0NaTjlFPHHrBB6/ZTxTRvh66JbWNLdzRvsMz4z2imyAnQeqWbW9/RVxiRY8tn0LCBl+Vpf9DE2KtMLR1yiqaCL3YLVPS+bWrN1Z6uODb3e60Wl81zHqjTZUCpkksoNE2c6NVRnZ8h0ryOQesd36eISnyUj1d69hPbIb0WbBZaxFFZ+BLnMMoTnnE3/Fn7oucIkzjrzDtT6vG012vzQQp8vN6mNNpSS6F0GQET37VgSl57tSpjUQdc4N3RzVb0da0T5FrNxWzLvf7fG6EpwoWpWcIf2jmDA0EYvNwZb8Sp/jb36zi/wjdVw/d5B0ow6CqJnXY96/2eeJOCCt83Z8xj1pJoJajyZxAJbCnQBo04ZhGDGjCyKWOFP511e53pzruEgd/7hnCtFtWi8H+kyqlXLM1pZaAFEUcbs9/63cXsK+o3UMTo/irFFJUgpeACJnXk/Z+48hOlu+UwW5ElVcms+86Dm3U/XNC7gtJuT6cKJn3wrgt/XsqCkh+ZZ/dnncEmceWakR7D/aci8I0SoZkhHFpj2+Of51Tf5F8RLdgz57Av3ShmGvLUUV2w+ZsufqnpMW2pK1n8fY/pXPduA+ibpQi93F1vwq5k5KJyUumncW++YWWu0uft50lJpGC3+9beJJRty7cRrrqVvxATKNHnl4HDis2CuPBJyriss4ZgVWAC4nyqhEoufehTqhP466cpSRCciUahwNVYhOu7Td3Mc4WtHkU9hYWWfmm5UF3DZ/mM+86WOSWb65yNuUZnhmNDlD4nj725bP8Tk5/dCoFbyzOI9Fqw4BsHT9EYoqmrh+rlRY2xZ1fDrK6GSvixB4ipub923EmPsLruZGDMNnED5xPqn3voWzvgJlVKLXmUAVk4y9qsh7rjI6BVtFIXW/foTTWEPI4CmET14gdYDtA1wzO5vqejOb91QQG6njrktGkBQTwvYDVTgcLeleUyQnkTMKmUaPJimru8M4aYIS2s3NzTz33HMcPnyYl19+mRdeeIGHH34YvV7Piy++2NUxnvEcLG7oVGS3t3DalqUbjvDELRPQqOQBG+Ds2F+Fw+mS8rU7oGrRi34FkQGRK7BXHPIZctSW4WpuRKZUoz62cuYyG7FXFqKKSemCaCXOZBoCrHDVG/3HNCoF/7xvKnsL65DLBLLTPDn8KbGh7DhQRf+kMKaN8jyk/bjhiM+5S9cfkYR2O8hDwn1eC3Il1UvfgmP1F3UrPkSuC8UwYqZfQ5rouXdR+fXzuJpqUITFEHXuTVR8+hSuZo+TU/2qT5CptYSNm3d63oxEt6HXKvnTTeNxudw+Bcj//N1Ulqw9jCjCvMnpJB7rxhyIRpMNi81JfJT+dIQs0YsISmg/9dRTxMbGUltbi1qtxmQy8cQTT/D88893dXw9gux+kcgEOhTbKbEGiirbsZtqhValYN+Runa7TMZF6iSR3QFuu6VTka2Kz8AwYga1P70T8Lh5/yYcNSU46stRRiTQsGHRscJKgcizryd8woVdELnEmcjgjChiI3VU1Zm9YzPGBN7VEASBIRm+jVBGZ8cyOts311ijVvh8vrUaKYOvPZxNvrm1ossBLl+rRPOhHRhGzMTtsGHcuRxHbRm6rHFokrMJn3QxzvpKDKPOxm02eUW299yCbZLQ7kO0dfnpnxzO/VeO7vS8D37Yy9e/FuByiwzPjObPN49HKzWOkwiSoP5S8vPzWbhwIatWrUKr1fLcc89x/vnnd3VsPYaYCC3/d91YXvtiJ82WwP7M00YlEapX8fnyA1jtLqx2J06XrzKXCXDhtAyefndzwGuE6lXcc+mIUx5/b0JQalCExeJs9HeAOY7bbgF3+w4k5sJc3K2aYLQgUr/6U0JHn4tMpTkF0Uqc6SgVMv5x9xS+WVVAfZOVGWNTGDc43nu8ttHClr2VRIdrGT0wFlkQ9pvXnjeI17/ciSh6drquPS+703P6KsGkdahiPE1/qr5+3uu727TtR5/vgaZtPxJ/+aMgU4C75TtaFS3tUkl0TGFZI1/8ctD7eleBpyfGpTMHdGNUEj2JoIS2rE3XQ5fL5TfW15kyIokpI5J45bMdXm/O48RH6bhoWn80agVzJqXzwEurKCj2XVnRqRXMHJfC61/mBtyavuPiYcye0E9aze4EQRCIOf9uqha95Ld6dRy3w07tsnfbvUbrVuxtER02RIcNJKHdZ4iJ0HJ7m5xsgIKSBh57Yy0Wm2d1esqIRB6+fhz7jtTx1a8HsTvczJuSTk4rYQ4we0I/BqdHsv9oPYPSI0nqYLu6rxM+cT5ViwKkJx4TzNr+owkbfz7OphqvyD5O64dt0WnHtHcd0efdSu3y9xHtFjQpgwiffGlXvwWJHk5ptclvLPdgNfMmp0ur2t2Ay2JCEARkmp6TwhPUX8m4ceP45z//idVqZc2aNXz00Ufk5OR0dWw9kuy0SD+hXVFr5ta/L+O6OYM4d3w/yttY/gnA1bMH8vbiwCkPQzOiOH9KRleF3OvQpg0j9d43cRrrEEWRys+ewlFb5j3ubtMxTm6IJO6yxyh772FwB07ZOY5uwFjk+rAuiVuiZ7Fo5SGvyAZYm1vGnIJq/vr2JuwOz/iOA1U8e+9Usvv5+q6nxBlIiZOapXRGyJApCAollV8/7/1sCgoViTf9A4U+3PtZFJ0OkMk7/vwKMkJHzSJk6DTc1mbJC18iKIZnxqBVK7DYWnZCdh6o5panlvH3uyeTlhDajdH1HUS3i5of3sS461cQZISNm0PUOTd2d1hBIX/yySef7GzShAkT2LVrF+Xl5axfv57Ro0fz4IMPIpefmaurLpeLqqoqYmNjUShO7xNncpyBvEM1VDdYfMZtdheb91by08ajNLUxyp84LIGiCiMVrfJAWzMgJYIpI5O6LObeiCCTIdfokWtDMAybgUytRR4aRdjEizHv2+hTmSrXGggdey5Nm5b4XUem1hE+5RIUhkhChkwh8uwbEOTSKoYErNxeQkmV72pXVJiG3QW+DaoMWhWjBvbcZgvdjSo6GW3GCNwOG6rYVKLn3I4mob9P+pZMqcZtN2Mr2Q94xLgqOhlXc6PntUpLzNw7kevDEOQKZGptwJ8lIdEWtUrOiAExVNaZqWx1j7Y7XBib7UwZId2buxKnsQ7TnjU079tI09algAiiG1vpAdQp2V7f/O6kM80ZlGJYtWoV99xzD/fcc493bNGiRcyfP//URdpLUCvl/O2OSVzyyHcBj9c1tXQrlMkEzp+czjXnZfvZ+bWmqt7MHQuX0y8hlJsvGCJVPXeC6HbRuPl7zAXbcNvMyHWhhAyZSuykBdjKDiFTaXFbWwSSKiaV2p//iyI0BmdTtXdc0OhxOx3Ur/4cBAGZWo/bbiNy+lXd8bYkzjDmTUpn054K3MeqoAenRzI8M9onnxOg3mjlP4t2c6S8iTHZcVx0Vn/kQeRyS7SgScpq1+bLaayndtk7WEsPos0YiW7AWPQDJyDXhtC8fyMucxOCXEXdr/9DpjUQPuliFKHRNKz/Glt5AZrUoYRPuFB6gJZol6zUCG6bP5Tf/fNXn/Hf2jdDIjhs5Yco+/AJ/y7Px3BUF0P6mV+31uE3y4oVK3A6nTz77LOIooh4bBXQ6XTy6quvSkK7HVRKORGhauo7Mb93u0UunNYfnUbJFbOy2HO4htJq/06SB4/lc5fVNFNZa+blh6Z3Rdi9hrpfP6Jx47c+Y5bDO3Hbmqld/oFPMRSAuWCr9//l+ghUiZlYDm5BtLb6txDBbWmiYd2XCAoVEVMu6dL3IHHmMyIrhufvm8ba3FJiwrWcPS4VtUpOhEHtU2exakeJt/Z2V0ENFpuTa6QCyN+M5egeLEd2oY7vjy5rHNVLXsFSuMtzrKkGQakmbOwcAEKGTMVcsI2Kz/7uPd98aDualGzM+z1F55bDubhM9UTPvuX0vxmJHkN4iNrvvn5OjlRM+1tx28yY9qxFdNrRD56MIiTCb07DpsXtimwEGdoeILKhE6Gdn5/Pxo0bqa2t5YMPPmg5SaHgxhtv7OrYejT/d81YnvlgC42t0kTiInU+W09pCaHEReoAiI3Q8cYfz+a1L3b65Xi35nBZI3VNViJDpWK89jAFdAyBpu3L/ER2W1zN9cg7KbIw5v4iCW0JADJTwslM8fV6bmsh1tbgZs3OUklo/0Zqf/mQxo2LvK9Dx1/gFdnHsRzO9Xlt2rve57Xb3OQV2d45e9ZKQluiXURR5Ik3N/iI7IumZTBzbGo3RtVzcTtslL77sLd2qn7dVyTf8k8UodE+80SH/46BIjIRuUZP+KSLe0xviw6F9vF0kY8++ohrrrnmdMXUKxiWGc3f7pxEdb2FrNRwbHY3DSYr63LL2HWohpRYA9fNHUSjyUZ1g4UIg5pt+ZXMmZjG1vwK6o2ePzC5TMDVyqBbrZKjUZ2ZufFnCgpDFC5jnd+4o66885NlcpSRCR1OadtEQ6L3cqConte+2MnRCiNjs+O474qR3nbr9UYr//56F9v3VSGKngfn2+YPZdKwBBavaelmKJMJ3vQSgNgIKT/4t1C1+BVMu1f5jBm3/YQyJsWzhXwMVZxv45q2N+9AuO0WbydYib7Bjv1VHK0wMnpgDKnxHRc0FpY1cbis0WesqKLzvhgSgTEf2OxrUGBuwpi7goipl/vMCx17HuaDW0H0rFZoUgaReP1TpzXWU0FQSWmXXXYZy5Yto7nZs5XucrkoKirigQce6NLgehp7C2upbbAyfEA0r32xk415FYAnd7OuyUpFrRmFXOCGeUOYf1Z/Fq85xLtL9vj5aQ/sF+EV2q42XXBsdhe3PLWMtx47B4NOdXreWA8j8uzrqfx8IW5bm+JStxOZPgx3s+8XpqDSItotgIAiLBbzoR2gVIMjQOqPXEnM3Du6LniJMwa3W+SZD7ZQVe8pbN68t4Kn393MdXMHMax/NK98tpOt+ZXe+fuL6vnASe/+AAAgAElEQVR/72zizUfPRqdRsuNAFRmJYaTEG3h3yR4cTjfhBjU3zJO6QJ4otsojfiIbAEEgZs6dVC159VgL9iRi5tzpMyUs53zMB7dirzrS/g9wOahd8SHxl/7x1AYucUbyn0W7vQ/D78oEHr1hHBOGtv+QZdCp/JrShRnUXR1mn6Jx03fINCGEjZvrHdOljyDppn9gyl+PIjQGw4gZ3Rjhbycoof3AAw9QXFxMdXU1gwcPJjc3V7L3a8NLn27nly2eVZW27dP3FrasrjpdIu9/v5exg2L57+I9fkIaYP/R+g5/lsni4KsVB7nx/CGnKPrehTZ1MKn3vUXJu4/irCn2OSZX64m75I/Yyg6ijEpCHdsPuS6U2hUf0LTlB5z15TjrW1a+FeHx6IedhYCIwhCJYfgMqWiqj1DbaPWK7OPkH6njsTfWMSsnlZ0H/JsiNTXbKa4wcc152T7pIdNGJlFe20z/pDDJC/830Lp4uTVh4y9Ak5JNyl2v4TY3BbTelOsMJN36HPbyQzRuXYpp98qA13LUlJzKkCXOUEwWB9+vK/S+drtFvlxxsEOhHROh5eLpmXz1awHgyde+/OzAxbkSnaPLykEZlei7qm1rpvbndzwuQ+nDveOq+Ay0zQ3Ya0pxNlT1mHSR1gTdGfLnn3/mySef5KabbsLtdhOEK2CfoaTK6BXZQLvt04/jdLk5Wt4UUGQHS01jOwUCEgDIVFq0aUMxthHaMm0I2pRsRIcV05612Er3EzZuHuZDOwNex9lQQeOaz0CuIHy85EzQl4gK0/i1Xz/Oss1F9EswcLTcd/tYpZCRHOffgCYsRO1NOZE4cTQpg/xuzBEzriNikqcgXxAEP5HtNNUjU2mRqTQIgoA6MZOwsXOOrYz7f/fqBozt0vcgcWbgdrcYOxwnmHvxjecPYebYFKrqLQztH4VGJd0LfisypZqkm56h+rs3aN63weeY5Wiej9CuWfoWxh0/A1C34kPiLnsYfQ/7rAb1l3LcGzAtLY0DBw4wZ84cjEYpP+k4Zqt/gZ0gtFg1t91ySo4NYcLQBNRKOTZHYFHetnCyLdNHJ59UzH2B0OEzMW5d6jMmqLQ0bvmB2p/f8Y6ZD2xBdHbsEIPLScP6r9FljWvXZkyidyGTCTxy/Vhe/zKXQyWNfscFfC369Foldy0YLqV0dQGCTE7idU/RuOUHXKZ6QoZNQ9tvaMC5bpuZyq+ew1KYi6DUEDnjasLGzfMeayuy5aFRGIZOI2LaFV39NiTOAEL1KmaMTfFZHLtoanAN4VLjQzvN55YIDplaR+i4OX5CW53Q3/v/ruZGjDuXtxwU3TRuWNQ7hbZOp2PJkiVkZ2fz+eefk5GRgdncvgjsawxICSc6XEtNqyY1l5+TRUmlCbcocsGUDIqrjKzLLSMuUseV5w5ELpdx4/mDefOb3QGvOTIrhuvmDGLngWpWbCvG5XLT1GxHrZJz0bT+jB0Ud7reXo9FnZCBNnMMllatma2FuVgLfV0J7FVH0WXlYG6q7fSa9upiSWj3AbbsreC97/fSZLJz9rgUymtMmK2+D8VHypu8/69VK3jnT+eg10oiu6uQ68OC8rBv2LgYy7HPuOiwUrvsPfRZOSjCYrBX+zs6GYZNJ3L61ac8Xokzl3svH8XogbHeIudB6VKX0O5AmzqEiGlX0rBxEbjdhI6dgy6rJS1ZFEWf5nKeMXfby5zxBCW0n3jiCT7//HP+7//+jy+//JLrrrtOKoRsRbPFQaPJd0W0rtHKIzeM874elhnN3EnpPnPOn5JBcmwIq7aXsmp7MY5WRZEThyUQFqLmrNHJnCWtXv9m5GpdUPPCJy3A2ViNvbIQBBkhQ6eizRhF9bcvtbqYAl1Gz/DtlPjtNBhtLHx/Cw6n5wv9q18LGJwe6VNroVMrMLdqyWyxOamqt5AuCe1ux09Mi27sNSUowmI8W9KCzOtiAKDrP+o0RyjR3chlAtNGSffVM4GIqZcRPnkBiKJfaqYiJJyQYdNaFUILhOVccPqDPEmCEtpfffUVf/yjpxr7pZde6mR236O6weK9KR+ntNq/eGfltmJWbC0m3KDm8nOySI41MDIrlpFZscyZlMbnyw9gsjg4b2IackHgrmd+ocFkY3B6JPddPkrK8fwNyA3+JvhtMYw+F03SAJJvfQ571VFk2lAUx84TnTaatv2ETKUhfPKlQVmFSfRsDhTV+32eQ3RKMpPDKK02kRRjYFBaJEvWtlj4RYZqcDjdPPPBFsxWJ7Mn9GPS8MSA1y+uNPLJz/upa7IyY0wysyekdeXb6XPoMkdj3r/J+1qm0aNJ9hSmqmJSiVvwEA3rv0F0uwjLmYcmZVB3hSpxmrA7XGzbV4lSIWfUwFipM+sZhiBrv0A85vx70GWOwV5Tgq7/aDRJA05jZKcGQWxbFRCACy64gCVLlpyOeE4JNpuNvLw8hg4dilrd9eLU5Ra5Y+Fyn5zqm84fzIIZLX8Q63LL+McHW7yvI0PVvPXYLNRK/z+w4kojv/vnCp+87szkMF58YHqXxN+bcTTVUPHRX3HUlfmMh4yYiX7AOBSh0agTgsvPk+gb1DRYuPXpZT4FUsmxIZRUtTw8z52UhiAIbNhdRnyUnqvPzebv72/2qdd46s5JjBgQ43Ntu8PFrU8v8+kc+eDVo5kxpudV0p+piKJI48ZvMe5eiUIfTsT0q6V0rz6M0WznDy+vpqzGY0+c3S+Cv989BaVC1smZEhLB0ZnmDGpFOzk5mZtvvpnRo0ej17d0zbvppptOXaQ9GLlM4K+3T+TDpflU1DYzaVgi88/K9JmzJrfU53Vdk409h2sZPTDW73ob88ppWwRdUNJIvdFKhEHqCBkM5oJt1Pz4Ns6mGnRZOUSdewv2mmIctWWoE/pjGDEDQSbHbW3G0VCJMrzjnHe3w4bLVI8iPA5BkFZDejPR4Vruu2IU7y7Zg9FsZ8qIRFbv9P38bswr5/2/nMedCzzV8au2l/gVRa/LLfMT2vlH6nxENsD6XWWS0O4Et91KzdI3ad63EUVEHNGzb223GFIQBMInzid84vzTHKXEmciyTUVekQ2w72g9m/dWMLmdHScJiVNNUEI7PNzTCa+0tLSTmX2XpJgQHrl+XLvHY8P9u8FFtGN4Hx/p3wJcrZQTolX+9gD7EG67hcpvXjzWhAbM+zciN4QTMflSmrb9hKu5HrfVjHH3SupXfozotKNOHkj8ZY8i1xn8rmfau46aH/6N22ZGGZ1M/OWPooyIP91vS+I0MnNsCjPGJONyi8gEgd2Haqhr1X45rs1nNC7SvxYg0Njughq/sYRofztACV/q136BKW81AI7qYiq/+iep976FTCml00l0jNnm8B+z+I9JdD1uu5X6NZ9hLcpHnTSAyGlXItP4653eRlBCe+HChe0ee/DBB3nhhRdOWUC9lfQk/0YKB4oaSE/0H580PIHhmdHsOnZTFgS465LhUqOLILHXlHpF9nGM237CuO1nbxFU/ZovwO3muNWXrWQ/dSs/InzKpbjtFlQRCYhuFy5zE9Xf/8t7PUdNCXW//o+4BX84re9J4vQjCAIKuWf34s4Fw3nh4+1Y7S7CQlTccqFvs6jstEjmTEzjx41HEEUYlBbJnElpftf8dbt/U5QLp6b7jUn4Yi3O93nttphw1JR6076cxnrsVUdQJ2Yi1/o/LLvtVlyWJuRaA4JC1WFOqETvYsaYFBatOoTtWH+LcIOaCcPab07THo0mGyqlHK1a8s/+rdT8+B9vwyhb2UGcTbV9ohvrSf/FFBYWdj5JguYAT9D/W5rPtn2VXDdnEClxLTcHuVzG03dNpqiiidIqE8MGxEir2SeAp3GFgI9frij6vnb7+5cbdyzDuGPZsYsoPV7oTv9/N0ettLPT15g4LJH3/xJDSZWJtIRQVAFqK+6+dASXzhyAxeakX0Jgr12l3DftSKuWEy6lg3WKJnkgtpL93tcyTQjK6CQAjLtXUf3d6+B2ISjVxF36sI87kDF3BTU/v4No9zT5kmkNRJ93GyGDJ5/eNyHRLSTFhPD876exbFMRKqWM8yamnZDXvd3h4vmPt7FhdzlKuYzLZ2VxxTkDuzDi3ktzqyJlONbDQnQjCL07X753v7sziJzB8X435waTjQ27y3nyPxtwufy9IVPjQ5k4PFES2SdI46YlBOr85kdHH26XI6DIBtBl9iyzfIlTg06jJCs1IqDIPk5spK5dkQ1wxayBtE7xv2TmABRy6Wu4MyKmXo5+8GSQKVBGJRK34CFkSjWi6Kbul/e9D86iw0bdig+957ksJmp+/I9XZAO4LUaql7yGyyI1Xesr9IsP5daLhnL93MHERgRn+XqcHzccYf2uckQR7E43/1u6j8Iy/wZWEp3TNuVSER6LIMho2LSEkv88QNlHT2It3tdN0XUd0h7IaSI2UsfTd03im5UF7C6owWhuEXFV9RYKy5vITA5HFEVqG61EGNTIpRvwbyLoFec2frrBoIpPlzrISfxmZoxJIT0xjF0F1QxIjpAaZQSJTKUl7uIH/Q+4XLjMvoLZUVtG5aIXUUWnoOk3BNFp9ztNdNpx1JYiP2b7JyHRHkcr/B/IiiqMAdM+JTomevatVH71LK7mRmTaEKLn3I5x9yrqlr/nnVNeVkDq7/6NXNt7alckoX2KcTjduFxuNAHyuLL7RfLoDTm89sVOftp41DuuUsiIi9RRXGnk7+9tpqTKRGSomgevHuPnWiDRObrMMVgO7+x8otvZ+Zw2REy70s9UX0KiLYVljfzrq10UVTQxZlAcd18yAv2xnam0hFDSOlj1lggeQaFEnZyFrdUqmOi00bxnLc2AJn0kirAYnI3VPufJtCGo4qTceInOUbWxAVQpZAzPlPop/BY0Kdmk/u5N7LWlKCMTkCnVVH7jW+Mn2i1Yi/PRZ7VvLtHTkJZMTyGL1xzi2r8s5Yo//8DzH2/za3pxnKtnZ9M/2fM0rFXLuWPBcAw6FW99s9vr1VvXZOOVz3bgbuvzJ9EpoWPnEDnzOlRx6WhSBqFKyOz8pHaQ6cNQhMeiiksnes4d6AdIaSMSHeN2iyx8fwv5R+potjpZvaOU/y7Z091h9VpUUUntHrMW7iTmgnvRDRiLTBeKoNKiSswi/vJHJccSiU75edNRvlvXUoemVsr4883jiQiV6ip+K4JCiTouzfv5U8Wktp2BKrp3de086aW5IPrd9AlKqoz8Z1Ge9/XKbSVkpURwwVT/ZiiRoRpeemA6FbXNhIWovVXMRyuafOZV1Vuw2p3oNFKO9okQyEe36LU7/Va1grgSSTf8XbLykzgh6pqslLfy7QXIO+Rv6ydxalDHZ2Dkl4DHBIUKdXwG8Zc/epqjkugNrN7h6xJkc7gx6IMvpJTonLCceVhL9mE5tANBqSZi2pUoI0/cFeZMJmihXVpaSmNjo4+wHjJkCC+++GKXBNbTOFzqXxwRaKw18VG+/pFjB8WxbHOR9/WgtEhJZJ8ios69hapFLyI6bCBXgitwoaOgUB3L6RQInzRfEtkSJ0x1gwWFXMDpavmuzEqN6MaIejeGkWdjLdmPac9akMkBEVxOQCBi2hXI1P49DCQkgiG6Tf8LmUwgUlrNPqXIVFoSrvwzTlMDMpUamar3fV6DEtovv/wy//3vf4mKivKOCYLAL7/8Qnq6lOcGMCQjCoVchrOVe8iIrPbzq6vqzHz160HqjTamj05m0vBEbps/DKVCRu7BajKSwv28eiV+O/qscaTe+xb26qOo49JxuxyYdv4Kai2qsBjcDhvKsBhU8enYyg8h14ejDPfv2ikhEYiCkga+XX0Iu8PFzgPVPiI7IUrHzdJn+aQRXU5qf3kfU94aFIYIIs++EV3GCAS5ktiLfk/07FtBrkB0ObEW7UUVnYQyUur+J/HbuXLWQHYX1FBVb0EmwNXnDpSEdhehCAnv7hC6DEEMIvdj5syZfPLJJ8TFddym+kyhs77zXcXW/Er+92M+JrOD2RP6cdnZWQHnOZxu7nzmF6rqzN6xx27MYeJvMNGX8EV0u7CW7EeuC0WmCcFRXYSzqZb6NZ/hMjdhGDGTqFk3ddqwwlbhyctTx0sPkhIdU9Ng4a5nfsFq9/dmBxieGc3Td0mezSdLw8ZvqfvlA+9rQakh9b63kJ9AZzlHQyWiy4UqShLgEsHhdLk5UFRPTLiOmIjet9oqcfJ0pjmDWtFOSEjoMSK7Oxk7KI6xg1p+T0fKG9m4u4LiKiPDM6M5d3w/BEEg/0itj8gGWLW9RBLaJ4nT1ED5/55osfcThGONalpo2roUZWQiYePmBryG6HJQ8flCLIdzAdBmjCD+8kcR5FIKj4Q/jSYb7yzOa1dkA2QE6AorceJYj/oWlIoOK/byQ2jTh3d6rii6qV78qreNu7b/aOIv/SOCQvpcS3SMQi5jcHpU5xMlJNohKKE9ceJEnn32Wc4++2w0mpZtkyFDpO3QQNQbrfz17Y0cKmnJ0V69o5SqegvXzRkUcOspMkzajjpZmrZ85+uh3c5mjbV0P6GjZmE5shuZNgRNUsvOgyl/g1dkA1gO59Kcv5GQoVNxO2xYj+QhDwlHndC/y96HRM+g3mjl/hdWUtdka3fO6IExXDFL6iJ3KlAnDsBcsM37WpArUcX2C+rc5v2bvSIbwHJoO8a81YSOPPuUxykhISHRmqCE9tdffw3Ajz/+6B07nqMt4c8Xvxz0EdnHWb65iOvmDCI51sBF0/rz7epDACRE6Vkw/bdb0El4cBrrg5qnik6l+M37cDZUAaAfNJG4BX8AwNVU6ze/+od/U7P8PXA6cNs8bhKKiAQSrn4cZbi009NX+XVrSYciGzyt26XOrqeGsAkX4qgtxbR3HXJ9OFGzbkSu73y3wF5bSvWS1/zGnQ2VXRGmhIREB7jtVk/bdbkCmaJvOLgEJbRXrFjR1XH0KsqqTQHHVUoZf317I25RJCJETVZKODGROu5aMJywEE9ez9b8Sr5bexiFXMYlMwZIneNOgJChUzHtXhnwmKDWgduNYcQM3PZmr8gGaM7fgLVkP5rkgeizx1O/+jPEVq4kosOK6LD6XM9ZX07ZB4+T+rt/dZrvLdFzqW208Pa3eew+VINWrWDB9EzmTDqet9+5tem+o3WcNzGtS2PsK8iUamLn30/MhfeCIENo3cu+AxrWf41ot/gOCjL0A8d3QZQSvRGXW2R3QTVyuYyhGVEIgoDLLbLzQBVWm4sxg2LRqKRGZh3hdtioXvIqzfkbPAMyOeGTFhB51pXdG9hpIKi/DLPZzLPPPsvq1atxOp1MnjyZP/3pT4SE9J4WmaeSEQNi2LavymdMJkBVvZmKWt/c7APFDdjsLv5y6wQOFNXzt3c2crxHzbb8SiaNSCQ9MYy5k9Ikq79O0GWMJP6KxzDm/oqg1iJTqnE2VKFNH07o2DleQVz9/b/8znWZPR7myshEEq79K41bvsdeeRRHbYnfXO85xlps5YfRJA3omjck0e08+Z8NHCn3tGBuNNl546tdKBVyzslJZcaYFBatOkS90bOqHR2moc5o82kyNUTK7TzlnOiDrau5yW8s6pwbpPQviaCw2Jw88vpar13vkIwonrxtIv/v7Y3sPuaPHxuh5bn7pkmNbDqgacv3LSIbwO2iYe0X6DJGoEkZ1H2BnQaCEtoLFy7E5XLx+uuv43K5+Pjjj/nb3/7GM88809Xx9UjCDf5Vp+lJYQHTScCzim2xOdmYV07rRpBOt8jqHaWs3lHKtn2VLLx7SleF3GvQZY5BlzmmwzmG4dMx5q4A0WPFKDdE+RRUaZIHokkeSPOBLVR+8Y/2LyTIUIRKrXh7K8WVTV6R3Zr1u8s4JyeViFANr/5hBqu2lyCXyzhrVBLb91fxwQ/5mCwOZo/vx9nj2nY9kzjdGEbMwHJou/e1KjaV0HHzujEiiZ7EL1uKfHpi7Dlcy6c/7/OKbPA0l/tx41GuOleqx2gPW8XhdsYLJaENkJuby+LFi72vn3rqKebNk76o2iM+0t9uKjXO0K7QjjCoUSnlxEXq2r1m3qFaSqtNJMVIuwgniyZlEAnX/hVj7q/ItSGE5cwL2I5ZnzWO8MmX0rjlOwSZHG3/UZgLtiHaLCDIiJxxDQqD1IikN7Ixr5zXvtgZ8FjrRlNhIWounNayMjptVDLTRvWu9sE9nZBBkxAuU9K8dx2KsGjCci4IOu1Eom/jcossWnXIb7zRZPcbM1sDN0GT8KBNG+67og2AgDZtaLfEczoJSmi7XC7cbjcymQwAt9uNXC7lpbbHoPRI5k5KY+mGI4giDO0fxe3zh2F3uFm3q8xv/vmTM5DLBKaOTOLfX+/yaXZxHJmAt1W7RHCYD+/03FxDYwgZOROZXOktntKkZKMMj0MeEh5wK9plbsLtchIx5VIMY2Z7fLnlCkTRjb3yCPKQyF5tsN+XsdqdvPTJdpqtTr9jSTEhXDpTShXqTtxOO9bifBShMUH7YeuzxqHPGtfFkUn0Nnbsr6KyjRWvUiHjsrMHsPNAFTWNntodlUIm7V51gmHUOTiNtTRtXYrosCM3RBA541pUMb3/9xa0vd/999/PVVddBcAnn3zC+PFSIUkgNuaV89WKg7hFkdvnD2PEgBhS4gwAPHLDOP7x/hY/sT08y5N+YDI7AopsgAum9pc6Up0Azfs3Ufnls97X9Ws/B1FElzmGsEkLqF7yKs76ChRhMcRe/KDX4s/tsFH59XNYCrb7XE9uiCL2ot+j7TcEdXzGaX0vEqeXqjqzn8hOjNHzx+vGkZEY2uFqaG2jhVXbS1Ap5UwfkyI5jpxiHHVllH34F1ymOgDCJ11M5Ixruzkqid5Ks8V/lXraqCQSY0J47vfTWLrhCBabk1k5/UhLCD39AfYgBEFG5FlXEXnWVd0dymknKKH9yCOP8MYbb/DCCy/gcrmYOnUqd999d1fH1qPYXVDDq1/spLym2Tt2sLiBf9471WfeJTMz2bSn3Cuoh2ZEkd3P4ywSE6ElKUZPaXXLNUZnxXDt3EGkJYSxY38VBr2KzGRpJbUzjDvbWE8e89Q2F2zDVnkEl9Fj4+dsrKb6+3+RcvuLgKehTVuRDZ7Cx+olr5Fyz+sIgqxrg5foVpJiDcREaKmub3GqmDw8kf6dNJ6pqjdz/wsrMZo9N+fFaw7zyoPT0Ug7UaeM+nVfeUU2QMOGbwkdPRtFWEw3RiXRWxk3OI6oMA21rVau55/lseKNCtNy7Xm9O7f4dNCwYRHG3BXIdaFEnHUl2n69L5UkqDuAQqHgvvvu47777uvqeHokNoeLhe9v9t5gjyOKsDW/ioH9Wiz6Vu8o9Vm1DtG1+EgKgsBjN+bwn0V5FFU2MXZQPLdeNBSLzck9z66gvNYjwM8alcwfru244K+vI9O2n8vuMvn6bTuqixFFEUEQsFcXtXues7GKis8Wos8eT+jIc05ZrBJnFnKZwOM3j+edxXmU1TQzYWhCUEVOv2wu8vkOKK9pZuOeCqaPlnK2TxUuU4PvgOjG1dwoCW2JU0JVnZl3luRxpKyJ0dmx3DBvMM/dN40f1hditbuYlZMqrVyfQoy7V1K34kMAHLWlVHz2d1J/9yZynaGbIzu1dCi0r7rqKj755BNGjRoVcLt0+3b/lb++SHGF0U9kHyc13vcPZtV2X7u4TXnl2Bwu1Er5sfmh/O3OST5zPl9+wCuyAVbtKOGCqek+Al7Cl/CJF2M+tAO32d/aSxERh7OuvGVAJkO0WxDUOnT9R2Havard61oObcdyaDtum5nw8Rd2RegSZwDpiWE8defkEzpHkPl/RwYYkjgJDMOmYzncUqSqjElBlSClckmcGp5+b7PXYaRsbSGiCHcuGM71cwd3c2S9k9ZdmAFEhw1r8d5e53HfodB++eWXAfjuu+/8jonttLfuiyTHhaDXKv3yuc4Zl8qk4b7FOlFhGq/vLnhaXnyzsoArO2jT3Gjy7z4XqOpZogVVTAqpd7+BpXAXLosRU95qXOZGDMNnYK8qwtRaaLtdmAtzCcmeSMiQqThN9TSsX4Tb1oxMqQFBhtviK9hNu1dLQlvCh1k5qXy/rpCGY5/vlDgD44cmdHNUvYuQoVNBrjjmIBJD2IQLpVQuiVNCvdHqY+MHsG2f1D20K/EvhBR6ZXFkh0I7NjYWgL/85S+8/fbbPscuv/xyPv/8866LrAehUSl4+Lqx/OvrXVTVmckZ7En5iA1g13fzBUP585vrfZpafLH8ABdOzQjYkEYURbJSIli+peh4mjFRYRpGZElbpZ0hU2vRZ3uejENHtaR61K/x/7tVhLb8PsPHX0j4+AsRXQ6qFr1M8762lkQglxxHJNoQFabltT/MYG1uGWqljMkjkrw7VRKnjpBBEwkZNLG7w5DoZYTqVEQY1D4LYf3ipTSRriR03FyspQcwH9iCoFITMe1KlJG9b3GiQ6F93333UVhYSHFxMRdccIF33Ol0olL1jR71wTJqYCxvPdp53u6wzGhSYkM4WtHSCMPudGNzuPyEttFs58//Xu99yk6K0TNmUBwXTe0v3cBPgtCxc2k+sAX7MQN9w+hz0SRm+s0z7loVUGTLtAYiz7q6y+OU6Dm43SKiKBIWombe5PTOT5CQkDijkMtl/P7KUbz06Q4ajDb6xRu45cLeV5h3JiFTqom/7GFcFiOCQhWwn0VvoEOh/cc//pHS0lIef/xxHn/8ce+4XC4nM9NfmEh0TlmNCa3G99eelhDKq5/vJCXWwKVnD8BwrEByyZrDPltZpdXNRBgaOVjSEHC1XCI45NoQkm5+FnvFYWQaPcqI+IDzHLWlfmOh4+YROf1qZCrJalHCw/frCvnox3ysdhezx/dj2uhkvl9bCCDVUkhInKFU1DZTUdtMdlokGpXnnjwmO453Hz+XRpONqDCt3zlOl5uCYs/9t77JisstkpUqNS07WeTa3lX82BZBDCLZunWzmuOYzWZ0ujNT7F1iTcQAACAASURBVNlsNvLy8hg6dChq9ZnzhGS1O7n978t9tqayUiI4UNzigjEkI4p/3ONptf7ypztYviWwC8ZjN45j4rDgmjVIgKVoD8advyBTaQkbf0G74trnnKN5lP/vLy0DcgUpd76KMjy2CyOV6EkUVTRxzz9/9RmTywVcx5yFlAoZr/1hBolSR1cJiTOGL345wIdL8xFFCAtR8dSdkzt1EymtNvH4m+t9bD8BBqdH8tfbJ3rFukTfozPNGdRfxooVK3jllVcwm82Ioojb7aahoYEdO3ac8oB7OuU1zYSFqNBplJTVmAjVq71NK3YV1PiIbPCscLdmz+Fayqqb2ZhXTkUrp5G2rNxeIgntdhDdLoy7fsVWehBNvyEowuMo/9+TILoBaN63gZS7XkOm7vhBUdtvKLEX3U/j1h8QFCrCJy/AVl5A46bFqBMHEDJ0qlSI1UuprDOzLreUUL2aqaPaz7U+WNzgN+ZqZd/pcHq6wV52dlaXxSohIRE8RrOdj3/a7615ajTZ+finfTx2Y06H53380z4/kQ2wt7COldtKOG9iWhdEK9EbCEpoP/vss9x///188skn3HbbbSxfvhy9Xt/VsfUo6pqs/O2djRSUNKJWygkLUVFVb0GpkHH93MHMP6s/UQE6O8oC+H898dZ6n7avWrUci83lMyc6wLaWhIfan/9L07YfATDuXI4ytp9XZAO4mhsxH9pByGBf+za33YL16F4UEXGoopO9eWMxF/wOmVJNzbJ3Me/beGz2UuwVh4maddPpelsSp4nCskb++OoarHbPZ+6njUd49t6pAS1OB6dHIRPA3cG+YKAtaIkTw20zYynchSIsFnUbOz9r8T7cNjPa9GEIcqkTp0THGJvtOF1un7H6Jmun5wUS2cepC+J8ib5LUEJbq9Uyd+5c8vPzUavVPPnkk8ybN4+HH364q+PrMXy6bD8FJZ58apvDRdWxD6XD6eadxXnkH6nlinMGkhitp6xV90iz1eEnpFuLbMBPZBt0ShbMkHLkAyG6XX5dIR1VR/3myXW+24T2qiLKPvqL13c7ZOg0mg9sRrQf/wIV8JgxttC0/Wciz7lBWtXuZfyw/ohXZAPsO1pP3uFahvWP9pnXaLLx44YjZCSHU1NvBkHgnHEp/LD+COZjLdzVSjmjBkoOQSeDvaqIsv89jtvi2f0LHTuX6Nm3IIoilZ8vxFywDQBFRDxJN/wdub7jDp4SfZvEmBAGpkawv6glZXP6mJROz5s2Kon8I3V+4wq5jMkjpN3lMwG3zYKgUCLIz6w0nqCiUavV2O12UlNTyc/PZ/z48QFXd/oypVWmDo+v31XOxt3lRIX5rmo7XSIzxybx86b2OxK2ZdzgeGmVrD0EGYJai9hOAyEAbVYOmjZtXuvXfenT3MaUt7rNWf5LloJCiUeAS/R2/v3VLpJiQ7h6drY3l/PJtzdS0Cp15O5LRxCiUXpFNngeutfsKOXCaf1Pe8y9hYb1X3tFNkDT1qWETbgAZ32lV2QDOOsraNr2ExHTLu+OMCV6EI/fMp5vVhYc6/waz8yxnXs3nz8lA4Vcxobd5eg0CkQ8XWTnTU6XbAC7GbfdQtW3L2M+sBWZVk/U2TdgGDGzu8PyEpTQnjlzJrfffjvPPPMMV1xxBdu2bSMiQqq0bc34IfHsKqjpcI5bhOoG3y2mpBg9545PY/nmIu/2s0wmIIoioghKuYAgk2F3tKywTRjaeSFfX0UQBCLPuoqapW8RSBwjkxO/4CHvg6LbYcO48xesJftO+GcZRp4jPXD2QuZOSmPltmKfVe2iSiNFlUbyC+t4+8+zqKoz+4hsgF+3FjNzrP/KWNtGVhInhsvSdhFDxG1pxhWg66vLYvQbk5BoS1iImhvPH3LC5503MU3KxT4Dadi4GPOBLQC4LSaqf/g32oxRKAxnhk4NSmjfeeedXHjhhcTFxfH666+zdetWH19tCc/TrsPpZm1uKbGROhKidCzdcNRndast0eFanrxtIvFReh69MYfFqw8jlwlcPCOTlFgDh0sbyE6LpLy2mU9/3o/J7GDW+H5SEWQnhI4+F1EUMe36FaexFpexZbtP0Oip+ekdwnLmoYpOpuKzp7Ee3XPCP0PQ6AmffMmpDFviDOH/t3ffgVFV6fvAn+klk14hhQCBhBYgQSCh9w4W1CCCuhbsu7ri6qpY2VXXXVdZ1/LTZXW/lo3ICigdVAihhBpCCyUVQnqbTDL1/v4YmTBMKsxkMsnz+Yt75tybd0Ju8s657zmnd09f/GP5ZOw5dhG7jlzE+auW2KzS6nE6pwLRPX0glYhgumriY4CPEklDeuA/m06hps66c6tSLsGExIgOfw9difewKai/0DjxXh7WF4qw3pD5h0Gi8YdZ+2sJgFgC7yETHM7XnTuE2uO/QOLlB7/R8yFWeaPm8BYYyy5CHZMIr9iWJ8ERUedmuJxj32Axw1CW32kS7TYt7/fAAw941M6QnWF5v58PFeCvXx1usY9IBHz52izbutnkHLqzh3A59U8t9hEr1Ai9848o+uLFli8mkQLmpj8sKaIGInzJ69cbJnmA1RtOYO3P52zHYhHwyR+nITRAjdTt2fhy8ylYBMDPW4E3Hk5GrzAfFFfosCk9B0azBTNG9UIUHyvfMN35I9CeTIfMLxg+I2ZDorIul2isLkFNxiZY9Dp4D5sCZbj96i66c4dw+b+NvwukviGQBYWj/nxj4h40axl8EqZ3zBsh6sYsJgN0Zw7AYqiHV+xoSNStr58tCBbU5xyHWVcNdd8E273fcDEb9TmZkIdGw1hZjIpt/7KdI5Kr0OvJT1pdWcxZbmh5P3fvDGk2m3Hvvffi2WefxZAhQ1z+9Zwp60J5q30UMglkUk6kc7barF9a7WPR66DLPtD6xZpJsgFAn38SxqpiyPxC2xMeeZBbJ8Ug60IZsvOrIJWIcdeMWIT+ulnUHVP7Y/zwcBSX6xDXO8C2BGBogPq6HktT89R9h0Pdd7hDu8w3BIFT72n2vNrj9r8LTNUlMFWX2LXVHNnORJvIxQSzCUVfvAh90XkAQOUv3yD8N29B6hPU4nnFa962lYWI1T7ouXQl9IWnUfrDB7Y+PiPnwi/5VmizdkHiHYCAyXd3WJLdFp16Z8iPPvoIISGeuTlIv0h/bNnXuNqFCMBdM+Pw7Y6ztnrrRdPjuMi9C0g0bXtcJAvuBYjEdkv/XT2CLfULhamquMVrWPTNL/lEns9Xo8BffzsBl0q18PaSOzx9Cgv0QlgglzrtrCRefo6NYglgaay/lyg7zx9koq5Kd/6ILckGAHNdFWqObEPAhEXNntNw6ZwtyQYAi64GNQd+QH2+fbln7aEt6PX7zxEwabHzA3eCFrO8iIgIREREYMuWLS6f9PXpp58iLS3Ndrxo0SL069cPFoulhbM6J4tFwKXSWsikYhhNFqgUUtw3dyBmJffGkD6BWJ92AZEhGsxKjnZ3qF2S36j50GVntJgkK6MGwnvQWFi0laj46f8AWB83hS78A8w1pYBYAq+40Sjf/m/UHt7a7HW6+taxZMWdHT2T3+hffxf8OortPXwaJBp/VO22lj2KZAr4jbvdnSESdQtCE0+Hm2qze93ouD65pYm2zq5Nw6nz589vsn3Dhg1OC+SBBx7AAw88YDt++umnodFokJWVhfz8fPzlL39x2tdytR0Z+Vj7c+MntwaDCYlxocjOr8RLn+yF0WT98LD/RDGG9w9GdZ0Bk0dEYmg/x/V2zRYB5VX1CPRTQdLE5jbUSDCbUH3gB9TnHofXgGQoI+MgUfvCVFMKY2UxxEovmGrKoT2yDQ35J3E59U0ETFkC2ck0mCovQxHeD8qI/hDLGsuUgmctg++IWTDVlEOXm4ma/RtwZUsxzeDxkPoEuuvtElErpD5BiHzkfdQc3Iz6vCyIJFJ4D5kAr9hRMJYVQhU9hOtuE3UAdUwCpP5hMFVeBmAd2PIeOqXFc5SRAyALjoKx9Nflj8US+AyfBmN0PEo3rLL187lpFsQy98zHa4s2TYY8cKCxltVoNOLHH39EZGQkHnnkEZcGBwCrVq3CxIkT21Wj7e7JkH//5jB2ZBTYtS2/OxFHzpRie0bT62WLRMBrDyVhWP/GUpns/Eq8+UUGSivrEeSnwnNLRyC2V4BLY/dk5dtWo/rAD7ZjdewohC181nZsMeqR//5DsDQ0LhcmksohmAy2Y3mPvoj4zdvNfg1DaT50Zw9CFtAT6v43QSRuemtuopYUldXhTF4F4qIDWHriYg0Xz+LS53+0lYhJvHwR+cg/OlUNJ1F3YK6vRW3mTxD0DdAMGQ+Zf+tLFZvra1F7ZBtMddXwHjQOip7WsuWGi2dRn3scitBoqGMSXB16i25oMuQVI0faL3+UnJyMlJSUNiXaWq0WKSkp+OijjxARYV3masOGDfjwww9hMplwzz33YPHi5utqnnjiibaE2KnE9gqwS7TFIqB/lD+On29+gqQgADsyCuwS7Q/WHLNt+1pWVY/3U4/ig+WdZxH2zkZ7Is3uWJedAYtRb/uka6oqtkuyAdgl2QBguHyhxa8hD46CPLj1zQ2oa6jVGaDVGXD0bBn8NAoMjPYHRCL4alr/AH+uoAqnciswIDoAMZGNtcI7MvLx/n+PwCJYfzf8NiWhyfW3yTm0Wbvs5mGY66qhO38EmoFj3BgVdUYGoxmfrs/CvuNF6BHkhQcXDLG7d+nGSFTe8BvVdIVEi+ck3+rQrgzvB2V4P2eF5lLXNROvsrISJSUlrfY7duwYXnzxReTm5traiouL8e6772Lt2rWQy+VISUnBqFGjXDK5Misry+nXbIsgqYCR/b1w+HwdlDIxpgzzxcXc0+gTYIRCJoLe2PRDhPq6Shw61LjTWV5Rtd3r+Zdr8eTbm5EyPhAKGVcruZa3VAUpGjcRschVOHLsuPVxAQBYzPBVaCDWNybbAuz3dhTEUrv/A+qezBYB3++rRFauzmHbI5EIiI9WY/4o/2bLuQ5ka7HxYOPP4uwRfhjZ31rn/dm6ItvmVBYB+GzdMfiKWv99StdHWa3Ftfvoni+8DFM973Oyt/1oNdJOWjc9qqzVY8XHafjdgjCWbdINaVOiffXSfoIgoKioCHfeeWer56WmpuLll1/Gs882Pr5PT0/H6NGj4edn/ZQ4Y8YMbN68GY8//nh7Y2+VO9fRvukm6/fq2kmko0c0YF9WES6V1eH7X87bvZaU0B+JiY2jpSOOmbD/xGW7PjnFeuRUeWPxzDjXBe+h6oOVKE59Exa9DiKpHGGzlyFm4Ai7PvrwAJRt+xeMFUXw6j8SEIvtJjsGTb0XfRMTOzp06mS27MvD8dyLTb4mCMCxHB2mJsVhYmLTI9F/37DZ7jj9TD0eWWTdTMW45ge714xmERL5M+cy5gH9cek/F2AsKwRgLSnrPe1W7upKDr7cbb8cZG29GaER/dGrB9fCp+ZdKR1pTpsS7ZdeegklJSWorq5GbGwsvL29IZG0Xpu6cuVKh7aSkhIEBzdO+gsJCUFmZmZbwujUzhdW4ftfzsNgNMNgNONsYRUgAD1DNJiUEIGZSdEQiUTw91FiVnJvfLnZccvvgyeLsW1/Pqq1evQI8oJGLWvya+Vddtx6mABV1CBEPfkJ9JcvQB4U1eRi+IqeMQi/x34zG99R81B/4Ri8BiRB2tRyYNTttOUe+/s3R/DVljO4d+4AVGsNOHS6BFFh3rhtUj+YTParJRmvOp6ZFG33IZtbOruWRO2NiAf/hoa8ExAp1FD2dP3StNR5lVfXo7BEi9gofygV9ilQTIQfzhY0PonyUskQGshafroxbUq0d+zYgS+//BIajQYikcg2Urt37952f0GLxWI3ktDUqK+nKa+ux/P/TEO93uzwWnVOBU7lVKBeb8atkxp/wSfGheCbbWdsxyIAaccu2Y4LS+xria82PNYz1xbvCGK5Cqoox81CBIsZdaf3wVhRBHW/EZAF9ED5lk9Rd3ofpP5hCJpxP6RefqjPyUTDpbNQRQ2EMnKAG94BdQaJcSHYsLvlen2zRUBReR3+/PlBW9v+E5dxtqAK88f1wVdbG+/v+eP62P5939xB6BXmjVO5lRgQHcD67A4gEkug6h3v7jDIzTbvzcVHazNhtgjQqGR45cHRdgsM3D1rAC6X1+FIdimCfJV4dOFQ7nVBN6xNP0Hbtm3D7t274e9/4/vGh4WF4eDBxj9MpaWlHrspzRUHTlxuMsm+2i9HCu0S7bjoAPz+rgR8v+s8xCIR6vWmFpPrK0ID1Jg5utcNx9zdlHz/d9SdSgcAVO76L7xiR6HutPWDouHyBVz+9i34JMxAVdq31j4AfG6ajaDp97srZHKjxLhQPLZwKH7ckwMAiAzV4PDpUtQ1GFs992h2KX5/VyL6hPviVK51ZZHRg3vYXheLRZg6shemjuR9TNRR9EYzVv9wAuZfJ0ho6434YuMprHykcVKsj5ccry1LRoPeBLlMAjFrs8kJ2pRoR0dHw8fHOTVKycnJWLVqFSoqKqBSqbB161a8/vrrTrm2uwT5XTvVpok+vo59JiZG2mo83/oio02J9qA+gR7/BKCjGatLbEk2AECwoD7nmF0fi64GVfvW27XVZGyEus9wty8dRO4xMynarqzjz58fQHpmUavnqRRSqJRSjBrcA6OuSrCJyH0a9CboGuw3SCmvbnrzk2tLSohuRJt+mpYsWYK7774bo0aNglTaeMr1TGAMDQ3FU089haVLl8JoNGLhwoWIj/fsR3qJcaFIGtIDe483/UfYT6PA3bNanrx414w4nMgpR2WN3tYmEYtsn74BINBXiTun9ndO0N2ICI4fTEQKNaDX2Y7FSi9YjAaHfjVHtzPRJgDAQzcPwckLFajS6u3ag3yVqNebUNdgglgswr1zB0Ih4/rqziRYzLA01EGidu6kNIu+HobSfMhDoiCWtz5gQp7LV6PA8P7BOJJdamublBjhxoiou2hTov3JJ59Ao9Ggtrb2ur7Izp077Y7nzZtnt5KJpxOLRfjjvSORV1QDo8mCsEAv/HSoAMF+Snip5YiN8oe8hT+8eqMZa3aeRW2dEQE+SsxMisaIASGIDPXGmdxKSCUimAUBA6IDIJPyD3h7SX2Doeo91G4UW6zUQBEaDd25w5D6BEIkU8JSVuBwLje1oCsCfVX4/OUZOFdYBY1ahuz8yl//eIegXm/CmbwKRIZ6I7CJp1d0/XQXjqJ0wwcwayug6BGD0NuegdTXcRfddl/3/BEUr/0rBEM9xAo1Qm57BureQ50QMXVWf1h6E7776SzyimqREBeC2cnR7g6JuoE2Jdr19fX4+uuvXR2Lx+vVwweCIOBsQRUS4kIQHqxx6HO2oBJKuRSRoY0rYqzdeRY7D1qTvIqaBqz96SzmjesDpVyKof2tf1AqaxpwOq8S/aP8OVrWToIgwFBuv1SbsSQXmrjRCF34LEp//BDazJ+aPFck4SNEapR5rgyXy+swfng4JiY0TmJUKaR2m021xGwRcCqnHF4qGXr35PbfLRHMJpSuXwVznXUlCH3ROZTv+Byhtz5zw9cu2/IpBIN1QzCLXofyrf+Cetl7N3xd6ry8VDIsnT3Q3WFQEwSzCZW7U6E9mQ7B2ACx0hvewybDd+Rch3LZhotnUbXnO1gMOvgMnw7NoLFuirpt2pRF9O7dG6dPn0ZcHNdubolWZ8CLH6fjfKF1o5mpN0XhtynDAQC6BiNWfLIXZ/IqAQDjhoVj+d2JEIlEOP1r2xUNBjNyL1VjcN8gAMCWfbn48DvrTGlvtRyvPjQa/SJvfGJqd2Fp0MJcU+bQri/OgUgsgaE4t9lzr6y9S92b0WTGb//2MwqKrfMoPlqbideWJSE+pn0jq9VaPZ7/5x4UFFufDk5MiMDvF3MN7eaY66ptSfYVhuI8p1zbVF12zXFpMz2JyNUq09agas93tmOzthIV2/8NsVwJn+HTGtvrqlH01SsQDNb6+oa8ExCrNFD3GdbhMbdVm7YXLCoqwsKFCzFjxgxb2UdXKv1wlo3pubYkGwC2Z+TjVE4FzBYBn/940pZkA8Duoxdx9NdasQG9A+yuo1JIbCNdeqMZ/+/7LFutdq3OgC82nnL1W+lSBKMBYrXjyKG6dzwsDXUQqxzX277CK260K0MjD/HTwQJbkg1YR6U/XNP+9f837smxJdkA8PPhQpzMKXdKjF2R1CcQsiD7OlpVH+eUd2gGJtsdew3gluxE7qI7e7Dp9uwMu+P6nExbkn1F3en9LovLGdo0ov3000+7Oo4uoayq3qHtYmktPlqbiQuXqpvtf9ukGJRW1mPXkUIE+qrw4M2D4aWyblazesMJ6I32SweWVzt+HWqaSVuJi/9aDouu8fsvkingkzgTXkMm4OK/noWxvHH9coglgMUMQASvAUnwuWlOxwdNnU5xhc6hrbpO30TPllXUOp5z9QRochR2+x9QtnU1jGUFUPdNQMCku51y3aDZD0PqE4SGi9lQRsbBL/lWp1yXOq/8yzVY/cNJXC6vQ3J8T9w1PRYSSZvGG8nF5EERMBTnOLTLAsPtjqX+YY59AhzbOpM2JdojR450dRxdwrhh4di8LxfCrwuFeKtlqKhpaDLJViulGDEwFAAgk0rwxB3D8MQd9o8+yqvrsSnd8QdvwnDOlG4r7YndMNfZf/8DJi+B74hZ0J7aa59kA/AbvQABkxZ3ZIjkASYkRODbHWchXNU2MaH99+HEhAhs3ZeLK4sJ+WkUGB574xP7ujJZQE/0SHnBob3h0jlY6muhih4MkaTpXXRbIpYpeK93I2azBa98ug+lldaBqtTt2ZBLxbhzWqybIyMA8J90Fwyl+TCUNJaGKcL7wy/5Frt+yvB+8Bk5FzUZGwHBAmWvwfBJmN7R4bYLZ3o50ZCYIKy4fzS27MuFWinDzKRe+PO/Mxz6xUb547Hbh8LfWwnAur7nzkMFqKhpwLih4ejVw7qEVa3OCItgf254sBdun8Il/tpKJHb8Eb8ywbHJ9chFHN0gR1FhPljx4Gh8ti4LugYTJiSE4945jjuQtmZQn0C8+lAStu7Ph0Ylw80T+kKtbH+S2N0Vr/2rbW18qV8oet6zElIN561Q8/KLa21J9hUHTxUz0e4kZL4hiHjwbzBWlUAkkUEwGyDzC22yb9C0++A3+mYIxnrIAnp2cKTtx0TbyUYMCMWIAdYfjjU7z6LymkfFSrkEL/xmpC3JFgQBL36cbqvf/m7nOfzpkTEY0DsA0T18EBPhi3NX1X3fOS2Wu1W1g2bweFQf+AGmqmIA1tExzUBrLaY6JhHykF62T9BitQ98hk91W6zUuY2IC8WIuKZ/8bfHsP4hbV6hhBw1FJ6x24DKVFWMmoyNHJ2mFoUGqKGUS9BgaCzFvDKoRZ2HzK9tvxul3v4APOPDNRNtFypvomZ70fRYW5INAGfyKu0mSZrMFmxMz7FNkHz1oWSs23UexeU6jBnaE0lDuNNce0hUGkQ88FfrdusiMbziRtk2phBJZeh5759RdyodFoMemgFJkHhxuTWizsysq2lTG9HV1EoZHrt9GD5am4m6eiNio/yxeAZXUiPXY6LtQuOGh2Njeo6t/EOjkmHaqF4or66H2SLA31sJwzUTHQHrjpAllTp4q+UQBAFLZg3o4Mi7FrFCBe+hk5t+TaaAd/ykDo6IiK6Xqnc8JN6BMNf+ulqLSAzv+IlujYk8w8SECCQP6YFanYEbS1GHEQmCILTezbPo9XpkZWVh8ODBUCgUbo3l8JkSbN6bC7VSirFDw/HZ+iwUlliXCZOIRbAIAq7+H5CIRfDxkqOyVg8RAAHWcpTldyeylvM6mXU1MJTkQdGjL3d6JOoCTNWlqM74Eeb6WnjHT4aqV9P18iZtJYwVRdZ7X+bevwXUeZzKqcCXW06hpLIeA3sHYMmsAbbE22IRsCk9B8fOlaFvuC8WTOgLpZxjktS81nJOJtod6PXP9uPAycvXde6i6bG4i4+52k17cg9K16+CYDZCJFch7PY/QBU9xN1hEZGL1RzeirItnwEWE8RqH/RIeQmKHn3cHRa5Wc6lavzu3Z9hsTS2hQSo8fFzUyCViPHFxpP4dsdZ22tjhvbEc0tvckOk5Clayzn5Mc3JsvMrsW7XeQgCMHdsb0glYqzfdQFmiwXnCitbv0AzcotYg9hegmBB+dZ/QTAbrceGepRv/xwRD7zj5siIyJUsRj3Kd3wBWEzWY10NKn7+Ej0WveTmyMjdfjlcaJdkA0BJhQ4nc8oRHxOMnw4W2L22N/MSGgwmjmrTdeNPjhNdLq/D8//cY6u73nv8EkQiEYwm613d1Gpy1xKJgKaeMSTGcZWCdjObHSZJmWqb3oVPECyoO7MfxpICqPoOgzKcSygSeSpLgw6CwX4yenP3PnUvflctRnA1X411JNLPR4my6sadB71Ucsi4qQ3dAP70ONG+rCK7yY0ms2BLsgFrAh0T6QcvlQxKuQQRIRpEhGgQ4q9CiL8Kg/oE4rmlN2HyiEgE+ang761AzyAvLJk1ANNH9XLHW/JoIqkMXnGj7No0g8c32bds48co+e4dVO7+Ly79+4/QntjdESESkQtIvf2hjLKv29YMHOumaKgzmTYyCpEh3g5tvcKsS/3dO3sglHIJAOucqd/MG8TdI+mGcETbidoyi3l2UjSmtZI0J8d3/gXYPUXwvCcgC4yAvug8VNGD4TtyrkMfc70Wtcd2XtUioGrfBmgGjeu4QInIqUIXLkdV+loYSgqgjkmAz4iZ7g6JOgEvlQz/WD4JR8+WorC4FkNigtC7Z+OyrkP7B2P1S9NxJr8S0T18uDoJ3TAm2k6UNKQHEuNCcOh0CQDrTpEyiRiHz1iP42OCMOE6tm2m6yeWKRAwIaXlTiLRrzU7VzdxUyAiTyZReSNwyj3uDoM6IbFYhITYECTENl2SqVHLkeiEzamIACbaTiWViPHKg0nIuVQNNj/xGQAAHEVJREFUi0VA3wg/ANZZzmaLgJhfj8l96k7vQ82hzRDJFPBLvhXKiFhIlF7wSZiOmoObrJ1EYvgm3+zeQImIiMjjMdF2gasfQzV1fPh0CT7feBI1dQZMGxmFRdNjOYLaAerzT6D4u3dwZei6Pvc4Ih/5AFJvfwROvx/qvgkwlOZD1WcYFKHRbo2ViIiIPB8T7Q6irTdiU3oOikrr8NPhQpjM1kmSX289g2A/Vat123Tj6s4cwNX1IYJRj5qMHxEw+W6IRCKoYxKgjklwX4BERETUpTDR7gDnC6vwyqf7UFWrb/L1Y2fLWk20C4prUVRehyF9g6BS8L/tesj8HGvuqvath9fAMVCE9XZDROTJ9EYzjp8rQ6Cv0uGpFRERtczcUAdddgbEChXUMYkQSbpmbtM131Un8um641i360KLffpGtPxH+j+bTiF1ezYAwFstx58eHYPoHj5Oi7E7MFZeBgQLJL7BMFeXNr4gmKE9mcZEm9rlcnkd/vCPNFTUWNfbnZUUjUcXDnVzVEREnsFUU4aLq5+DWWvdyE8REYeeS16DSCxxc2TOx8UhXchsEbAhLafJ16QSMcRiESYMj8Dcsc0neZU1DVizs3E72FqdAd9sO+P0WLsy3fkjKPjotyjftto+yf6VxIuTVKl9vvvpnC3JBoBNe3NRUFzrvoCIiDxIzeGttiQbAPSFp1F/4ZgbI3Idjmi7kNlsgcXiuM2jRiXDnx5NRo8gjW1b14LiWuw6chH+PgpMSoy0lYdo640O16jWNl2CQk2rSl9r24r5WvLQ3vAZNqWDIyJP19Q9WFNncEMkRESeRzA6/g4VTF3zdygTbReSyySICPFCYUmdrU0mFWHZLUPQu6cfMk5exldbTqO4QodandHWZ/uBfLzz5HiIxSJEhnqjf5QfsvOrbK9PvSmqQ9+HpxNMxmtaRAhLeQEimQLKyDiIRHywQ+0z9aYo7D1eZDsOD9YgLjqgyb6XyrTYsPsCDEYLpo2MwrnCKhw8VQxtvRHhQRrMG9cHMZF8qkLU2QmCgNyiGvh4ybmRzQ3yHjYFNUe22RJuqX8YVH2Huzkq1xAJguA45Orh9Ho9srKyMHjwYCgUig792roGI345chG6eiPGDQ9H6vZsbNmX59AvxF+Fksr6Zq/z5mNjMahPIABruci6XedRVFaH5PieGMOdI9ul9vgvKF3/vu3Ya+AYhN7ytBsjoq7g0Oli/Hy4EIE+SiwY3xf+PkqHPjV1Bjz85g7U6qwjNSIRcO1vXLlMgn88Mwk9grw6Imwiug7VWj1WfLIXFy5WQywCbpvcD0tnD3R3WB7NUFYI7fFfIFao4T1sCiRqz5x71lrOyRFtJzKaLHh21W7kXbbWan67IxsGk6XJvi0l2QAglTSuq+2tluPumQOcF2g34z1kAqQ+QdCdOwh5UCQ0g8e7OyTqAhLjQlvdPe7AiSJbkg04JtkAYDCakXbsIm6f0t/ZIRKRk3z/y3lcuFgNALAIwLc7zmJSYiQiQ73dHJnnkgdFIGDSYneH4XJMtJ3o8OliW5INAHUNJshl7S9LiI8JQmyvph9D0/VR9RoEVa9B7g6DuhkfTdueqAU0MRpORJ1HcYXOoa2kUsdEm1rF4lQnEokdd3c0GJse0W7OlBGRePWhJGeFRE5grqtGxc9fofSHD1Cfe9zd4ZAHSYwLxfD+wbbjQF/HhHpQn0CMGxbekWERUTuNGWpfsumnUdjKO4lawhFtJ0qIDUGfcF/b46Vr3T6lH4ordNh15KKtTSmXYMH4PogI9UZcrwCEBbJOszMRLGZc+s9LMJZb/89qj/2EsJQXoO6ikzbIuSRiEV5bloyTOeUwGM0Y0jcI5TUNOFdQBalEBB+NAnF8ekXU6Y2J74mn70rAjox8+GmUuHNaf9uqYUQt4U+JE0klYrz1+FikHb2EPZkXcfBUid3rft4KLJ09EJMSI7Ftfx56Bmtw+5R+UCtlboqYWtNQeNqWZFsJqD22k4k2tcvA3o0jXyH+aoT4q90YDRFdj0mJkZiUGOnuMMjDMNF2MqVciqkjozCwdwAyz/0Mg9EMAFAppEgeYn30NGJAKEYMaHkSFXUOEpVj/V1TbURERETXYqLtIj2DNXj78bHYmJ4LsViEuWN6I8iP6252NMFsRN2pfTBpK+EVNwoyv/Z9wJEHR0ETPwnazJ8AWHeR9B093xWhEhERURfDdbSpS7v05Sto+HUCo0imQM8lb0DRo0+7r9Nw6RzM2kqooodALOcKEURERNR6zslVR6jLarh41pZkA9YtX6sP/nhd11L2jIFX/5uYZBMREVGbsXSEurAmHtZ0uec35Crl1fXYffQiFHIpJgwP56RlIiJqNyba1GUpw/tDGTUQDfknAQAiqRw+I2a5OSryBMUVOjz17s+o1RkBABt2n8czixOxPaMAIgAzk6K5UQUREbWKiTZ1aWGLXkLdiTSYtFXQDBgNWUDP1k+ibm/b/jxbkg0ABcVaLH9/Nwwm6wZU2zPy8cHyyZzgTERELWKi7SJmswUWQYBMKrFrv7Lcn8lsafZRtN5ohiAIkIjFqKptgEohhUYtd3nMXZFYKof30MnuDoM8jeMmr7YkGwB0DSakHbuEmyf07cCgiIjI0zDRdoF1u87j6y2noTeaMXVkLzx8azxEAP7fuuPYlJ4Li0WAAKBfpB+eXTLCthuk2WzBP7/LxLYDebh2LZheYd5496kJDok7ETnfxIQIfLfzHExma3KtVkqhazDZ9dGoWLNN1F00GEzYsi8PhSVajBoUxr0wqM246oiT5RbV4NN1WahrMMFkFrB5by52ZuRjT+Yl/JCWA/OvSTYAnC2owsf/a1wVY3tGAbbud0yyASDvci3W7DzbIe+BqLvbe7zIlmQD1hHsqLDGmuyYSD+MGx7ujtCIyA3e+uIgPl2Xhc17c/Hqp/uw/UC+u0MiD8ERbSc7V1Dl0Ha2sApqRdPf6gsXG/ufv+h47tUyz5Zh0fQbi4+IWnfhYrVD290z46CQSQERMLRfMCTiJupLiKjLKa2sx8FTxXZtm/fmYurIKPcERB6FI9pONrhvIMTX/AEeGhOMof2Cm+wff1V7c32umDCcE/naw2LUozZrF2qP7YSloc7d4ZAHGdY/xO5YIZdgcN8gJMSFICE2hEk2UTeikEsc7nm1kuOU1DaSV1555RV3B+FsZrMZJSUlCAkJgVTasTeDRi1HVJg38i7XQi6TYOHkfpiZFI0eQV7w8ZIjv7gWFosAuUyCMfE9seyWIZDLrHXXUaHeUMolOH+xGiaT2TYhSwQgeUhP3Dt3EEQi/oFvC4tRj4ur/4Daw1uhy86A9sRuaAaP54Yz1CZ9w30hEolQUqlDRLAGT9wxDFFhPu4Oi4jcQCGXQG8042ROhe348duHIcRf7ebIqDNoLefkFuzUJdUe/wWl69+3awuYshR+oxe4KSIiIvJk5wqqUFhSi6H9g+HvzUEbsmot5+SzD+qSBLPRsc3k2EZERNQWMZF+iIn0c3cY5GFYo01dkldcEiQ+QbZjsUoD7/iJ7guIiIiIuh2OaFOXJFF6IeI3b6M28ycIZhO8h0yA9KrEm4iIiMjVmGhTlyXx8oVf0s3uDoOIiIi6KSbanYDFIiC7oBK+Xgr0CLLuEllaWY/LFXWI6+XP3SCJuphqrR5b9uVB12DE5BGRXNGEiKiLYqLtZpW1DXjhw3QUFNcCAOaN64MAHyX+s/EkLAIQ4KPEGw8nIzLUu5UrEZEn0BvN+P17u1BcoQMAbEjLwd9+Ox69ejDZJiLqajgZ0s3W/XLelmQDwIbdF/B/m6xJNgBU1DTg661n3BQdETnbwVPFtiQbAAxGM7ZxO2cioi6JibablVbVO7SZLfbHZU30ISLPpJA5loIp5SwPIyLqiphou9mE4RGt9pmY2HofIvIMw2NDMLB3gO04wEeJWcnR7guIiIhchjXabjZyUBiev+cmrNl5FmcLquxe6xHkhTum9MPUkb3cFB0ROZtELMKfHhmDQ6dLoGswYuSgMKiVMneHRURELsBEuxNIju8JAPjz5xl27WOH9mSSTdQFSSRijBwU5u4wiIjIxVg60kncNDAUUWGNK4toVDJMH8Ukm4iIiMhTcUS7k5BJJfjLE+Ow68hFNBhMGDcsHIG+KneHRURERETXiYm2E10ur8O6XeehazBh+qheGNQnsF3nq5UyqBRSZJ4rg7beiJsnxECjYu0mERERkSdiou0kugYjlq/ajapaPQDg58OFePvxsYjtFdDKmY027c3FP9ccsx0fP1eGtx4f5+xQiYiIiKgDsEbbSQ6dLrEl2YB1W/WfDxW26xo7Muw3rTiZU4GisjqnxEdEREREHYuJtpP4auSObd6Kdl3DT2PfXyoRw4ulI0REREQeiYm2kwzpG4TRgxuX6woP1mBWUnS7rrFoeiy81dbEWiQCUqb1h4+XYwJPRERERJ2fSBAEwd1BOJter0dWVhYGDx4MhaJ9o8o3Kju/EroGI4b0DYJE0v7PMboGI07mVKBnkBd6BmtcECEREREROUNrOScnQzpZ/yj/GzpfrZRhxIBQJ0VDRERERO7C0hEiIiLq0koqdNDqDO4Og7ohjmgTERFRl1RXb8TK1Qdw/HwZpBIxFk2PxR1T+7s7LOpGOKLtIsfPl+GDNcfwzbYzqOWnaCIiog63btd5HD9fBgAwmS34v82ncKlU6+aoqDvhiLYLHDpdjFc/3Ycr00zTMy/h709NhFgscm9gRERE3cjFEvukWhCAwlItFxugDsMRbRfYuj8PV6/lknOpBtn5le4LiIiIqBsaOSjM7thLJcPgPoFuioa6I45ou4CX0nGTGW48Q0RE1LEmJERAW2/Ejox8+GoUuGtGLNRN/I0mchUm2i5wy8QY7D9xGTV11trsySMiERnq7eaoiIiIup85Y3pjzpjebe5vsQgs9SSnYaLtApGh3vjk+ak4kl2CQB8VBvQOcHdIRERE1AJdgxHv//co9mYVIdhPhYdvjee+FnTDWKPtIl4qGcYODWeSTURE5AG+2ZaNPZmXYLEIKK7Q4S//dxD1epO7wyIPx0SbiIiIur1rFy3QNZhQWFLrpmioq2CiTURERN3eoGtWI/FWyxAV5uOmaKirYI02ERERdXt3TO2PypoGpGdeQmiAFx68eTAUMom7wyIPx0SbiIiIuj2FTIIn7xyOJ+8c7u5QqAth6QgRERERkQsw0SYiIiIicgEm2kRERERELsBEm4iIiIjIBZhoExERERG5QKdddeTChQt45pln0KdPHwwePBj33nuvu0MiIiIiImqzTjuifejQIYSFhUGpVGL4cC61Q0RERESepdOMaH/66adIS0uzHa9YsQJTpkyBRqPBI488gs8++8yN0RERERERtU+nSbQfeOABPPDAA7bj77//HklJSZDL5ZBKO02YRERERERt0mkz2D59+uDNN9+ERqPBHXfc4e5wiIiIiIjaxeWJtlarRUpKCj766CNEREQAADZs2IAPP/wQJpMJ99xzDxYvXuxwXnx8PN59911Xh0dERERE5BIuTbSPHTuGF198Ebm5uba24uJivPvuu1i7di3kcjlSUlIwatQoxMTEOP3rZ2VlOf2aRERERERt4dJEOzU1FS+//DKeffZZW1t6ejpGjx4NPz8/AMCMGTOwefNmPP74407/+oMHD4ZCoXD6dYmIiIiI9Hp9iwO7Lk20V65c6dBWUlKC4OBg23FISAgyMzNdGQYRERERUYfr8HW0LRYLRCKR7VgQBLtjIiIiIqKuoMMT7bCwMJSWltqOS0tLERIS0tFhEBERERG5VIcn2snJydi7dy8qKipQX1+PrVu3Yvz48R0dBhERERGRS3X4OtqhoaF46qmnsHTpUhiNRixcuBDx8fEdHQYRERERkUuJBEEQ3B2Es12ZAerpq44UFNdiR0Y+lAopZozuBX9vpbtDIiIiIqJftZZzdtqdIbu7/Ms1ePq9XdAbzACArfvz8MHyyVAp+F9GRERE5Ak6vEab2mZHRoEtyQaA0sp6HDxZ7MaIiIiIiKg9mGh3Ukq5xKFNoXBsIyIiIqLOiYl2JzUjKRpBvo012QOiA5AYy2UQiYiIiDwFC347qQAfJT54djIOnCyGSi7BiAGhkEj4uYiIiIjIUzDR7sTUShkmJkS4OwwiIiIiug4cIiUiIiIicgEm2kRERERELsBEm4iIiIjIBZhoExERERG5ABNtIiIiIiIXYKJNREREROQCTLSJiIiIiFyAiTYRERERkQt0yQ1rBEEAABgMBjdHQkRERERd1ZVc80ruea0umWgbjUYAQHZ2tpsjISIiIqKuzmg0QqlUOrSLhOZScA9msVhQV1cHmUwGkUjk7nCIiIiIqAsSBAFGoxFeXl4Qix0rsrtkok1ERERE5G6cDElERERE5AJMtImIiIiIXICJNhERERGRCzDRJiIiIiJyASbaREREREQuwESbiIiIiMgFmGgTEREREbkAE22yU1hYiMmTJzu0x8bGwmw2Y8WKFZg7dy7mzZuHDRs22M6JjY3Fnj177M6ZPHkyCgsLAQD/+Mc/MGfOHMyZMwdvv/22698IEbV4P19RXFyMsWPH2p3T2v0MAFqtFnPnzrVrI6LWXX3/Nef999/HxIkTsXr16jb17yirVq3CmDFjsGDBAsyfPx/z5s3Dvn373B1Wp8ZEm9ps/fr10Gq1+OGHH/D555/jjTfegFarBQDIZDK89NJLtuOrpaenIy0tDf/73//w/fff48SJE9i2bVtHh09E1/jll1+wdOlSlJaW2rW3dD8DwLFjx7Bo0SLk5uZ2QJRE3c+6deuwevVq3Hfffe4OxUFKSgrWrVuH9evX4+2338bTTz/t7pA6NSba1Ga33HKLbTS6pKQEMpkMMpkMABASEoLk5GS89dZbDucFBwfjueeeg1wuh0wmQ9++fXHp0qUOjZ2IHK1ZswarVq1yaG/pfgaA1NRUvPzyywgJCXF1iERd1v79+/Gb3/wGjz76KGbMmIEnn3wSBoMBK1asQHFxMR577DGcOnXK1n/VqlV29+uVp0xmsxl//vOfccstt2D+/Pn497//3eL1N2/ejAULFmDBggWYN28eYmNjkZmZiezsbCxZsgS33XYbJk2ahK+//rrV91BbW4vAwECnf2+6Eqm7A6DOp6SkBAsWLGjyNalUihdeeAHr1q3DQw89BIVCYXvtueeew7x587Bnzx6MGTPG1t6vXz/bv3Nzc7Fp06Y23cBEdONaup+bSrKvaO5+BoCVK1c6NUai7urIkSPYtGkTQkJCcMcddyAtLQ2vvfYa0tLS8MknnyAiIqLVa6SmpgIA/ve//8FgMOD+++/H4MGDm73+zJkzMXPmTADAG2+8gREjRiA+Ph4rV67Eo48+iqSkJBQUFGD+/PlYtGiRw9f75ptvsH37dhgMBuTl5eG1115z4nek62GiTQ5CQkKwbt06u7ara8RWrlyJZ555BkuWLEFCQgKio6MBABqNBq+//jpeeuklrF+/3uG6Z8+exbJly/Dss8/aziEi12rtfm5Oa/czEd24fv36ISwsDADQt29fVFdXt/sae/fuxalTp2y10jqdDmfOnEFMTEyL11+zZg1OnjyJzz//HID1w/Xu3bvx8ccfIzs7Gzqdrsmvl5KSgieeeAIAcOHCBSxevBi9e/dGYmJiu2PvDphoU5tlZWVBo9EgOjoa/v7+GDduHM6cOWOXNI8dO7bJR86HDh3Ck08+iT/+8Y+YM2dOB0dORNejufuZiJzj6qfCIpEIgiA021ckEsFisdiOjUYjAMBsNmP58uWYPn06AKCiogJeXl44evRos9c/fPgwPvroI3zzzTe2EtDf/e538PHxwaRJkzB79mz88MMPrcbfp08fJCQk4OjRo0y0m8EabWqzY8eO4S9/+QssFgu0Wi3S0tKQkJDg0O+5555DWloaSkpKAABFRUV47LHH8M477zDJJvIw197PROQe/v7+OHfuHAAgMzPTNol59OjRSE1NhdFoRF1dHe666y4cPXq02esUFRXhmWeewd/+9jcEBQXZ2vfs2YMnn3wSU6dOxa5duwBYk/iW1NTU4OTJkxg4cOCNvr0uiyPa1GYpKSk4c+YM5s2bB7FYjMWLF2P48OEOy3tdeeR8//33AwA+++wz6PV6vPnmm3bXaqr2i4g6l2vvZyJyj9mzZ2PLli2YPXs2Bg0aZEtuU1JSkJeXh1tuuQUmkwm33norRo0ahf379zd5nX/+85+oq6vDK6+8Ykukly1bhieeeAJ33XUXFAoF4uLiEB4ejsLCQvTq1cvu/Cs12mKxGHq9HrfffjuSkpJc++Y9mEho6TkFERERERFdF5aOEBERERG5ABNtIiIiIiIXYKJNREREROQCTLSJiIiIiFyAiTYRERERkQsw0SYi6iZWrVrV7HbJ3377Lb788ssOjoiIqGtjok1ERDh06BAaGhrcHQYRUZfCDWuIiDxUXV0dnn/+eeTl5UEsFmPQoEGYM2cOVq5cads+ef/+/Xj99ddtx+fPn8fixYtRXV2NAQMG4OWXX8bevXuxc+dO7NmzB0qlEl988QVWrFiBMWPGAABeeOEF9O/fHzU1NcjLy8Ply5dRWlqKuLg4rFy5EhqNBsXFxXjttddQVFQEo9GIOXPm4OGHH3bb94aIqDPgiDYRkYfatm0b6urqsG7dOqxZswYAHHZqvVZ+fj5WrVqFDRs2QBAEfPjhh5g2bRomT56Me++9F4sXL8aiRYuQmpoKANBqtdi5cyduueUWAEBGRgb+/ve/Y9OmTZBKpfjggw8AAMuXL8dtt92GtWvXYs2aNUhPT8fGjRtd+O6JiDo/JtpERB4qMTER586dw5IlS/DJJ5/gnnvuQVRUVIvnTJs2DQEBARCJRLjtttuQnp7u0OfWW29Feno6KioqsH79ekycOBE+Pj4AgJkzZyIoKAhisRgLFy5EWloadDodMjIy8N5772HBggW44447UFRUhNOnT7vkfRMReQqWjhAReajIyEhs27YN+/fvx759+3DfffchJSUFgiDY+hiNRrtzJBKJ7d8WiwVSqeOfAR8fH8ycORPr16/Hhg0b8PLLLzd7vlgshsVigSAI+Oabb6BSqQAAFRUVUCgUTnuvRESeiCPaREQe6quvvsLzzz+PsWPHYvny5Rg7diwA4NKlSygvL4cgCPjxxx/tztm5cyeqq6thNpuRmpqK8ePHA7Am0CaTydZv8eLF+OKLLyAIAuLj423tO3bsQG1tLSwWC1JTUzFp0iRoNBoMGzYMq1evBgDU1NRg0aJF2LFjh6u/BUREnRpHtImIPNTNN9+MAwcOYPbs2VCpVOjRoweWLFmCuro63HbbbQgODsbEiRNx/Phx2zl9+/bFsmXLUFNTg8TERDz00EMAgPHjx+PNN98EACxbtgxxcXHw9fVFSkqK3dcMCgrCgw8+iMrKStx00022CY/vvPMOXn/9dcybNw8GgwFz587F/PnzO+g7QUTUOYmEq58xEhERwTppcsmSJdi8ebOtHGTVqlWorKzEihUr3BwdEZFn4Ig2ERHZee+995CamopXX33VlmQTEVH7cUSbiIiIiMgFOBmSiIiIiMgFmGgTEREREbkAE20iIiIiIhdgok1ERERE5AJMtImIiIiIXICJNhERERGRC/x/7c+/A8T7MZIAAAAASUVORK5CYII=\n" + "image/png": "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\n" }, "metadata": {}, "output_type": "display_data" @@ -1001,7 +1001,7 @@ { "cell_type": "markdown", "source": [ - "And a similar swarm plot, with a different format for pvalues" + "With a horizontal orientation" ], "metadata": { "collapsed": false @@ -1010,6 +1010,74 @@ { "cell_type": "code", "execution_count": 21, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "p-value annotation legend:\n", + " ns: p <= 1.00e+00\n", + " *: 1.00e-02 < p <= 5.00e-02\n", + " **: 1.00e-03 < p <= 1.00e-02\n", + " ***: 1.00e-04 < p <= 1.00e-03\n", + " ****: p <= 1.00e-04\n", + "\n", + "H1N1_Nonsynonymous vs. H1N1_Synonymous: Mann-Whitney-Wilcoxon test two-sided, P_val:5.014e-04 U_stat=2.624e+03\n", + "H3N2_Nonsynonymous vs. H3N2_Synonymous: Mann-Whitney-Wilcoxon test two-sided, P_val:1.294e-03 U_stat=1.535e+04\n", + "Influenza B_Nonsynonymous vs. Influenza B_Synonymous: Mann-Whitney-Wilcoxon test two-sided, P_val:2.026e-01 U_stat=3.340e+02\n" + ] + }, + { + "data": { + "text/plain": "(<matplotlib.axes._subplots.AxesSubplot at 0x7fc283af2668>,\n [<statannotations.Annotation.Annotation at 0x7fc283bf3780>,\n <statannotations.Annotation.Annotation at 0x7fc283be1dd8>,\n <statannotations.Annotation.Annotation at 0x7fc283bc3ef0>])" + }, + "execution_count": 21, + "metadata": {}, + "output_type": "execute_result" + }, + { + "data": { + "text/plain": "<Figure size 864x432 with 1 Axes>", + "image/png": "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\n" + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "fig, ax = plt.subplots(1,1, figsize=(12,6))\n", + "fig_args_horiz = {\n", + " **fig_args,\n", + " 'y': 'subtype',\n", + " 'x': 'mutation_rate_samp'\n", + "}\n", + "ax.set_xscale('log')\n", + "sns.stripplot(ax=ax, orient='h', **fig_args_horiz)\n", + "annotator.new_plot(ax, plot='swarmplot', orient='h', **fig_args_horiz)\n", + "annotator.apply_and_annotate()" + ], + "metadata": { + "collapsed": false, + "pycharm": { + "name": "#%%\n" + } + } + }, + { + "cell_type": "markdown", + "source": [ + "Back to vertical, with a different format for pvalues" + ], + "metadata": { + "collapsed": false, + "pycharm": { + "name": "#%% md\n" + } + } + }, + { + "cell_type": "code", + "execution_count": 22, "outputs": [ { "name": "stdout", @@ -1060,7 +1128,7 @@ }, { "cell_type": "code", - "execution_count": 22, + "execution_count": 23, "outputs": [ { "name": "stdout", @@ -1101,7 +1169,7 @@ }, { "cell_type": "code", - "execution_count": 22, + "execution_count": 23, "outputs": [], "source": [], "metadata": {

Horizontal Support for horizontal orientation, at the moment tested in box plots and strip plots. Mentioned in #26 , https://github.com/webermarcolivier/statannot/issues/54

2021-07-25T15:16:18

0.0

[]

trevismd/statannotations

trevismd__statannotations-27

4beac149bf62a9a9f3a2e9d7857b76bd62889b9a

diff --git a/CHANGELOG.md b/CHANGELOG.md index 80407ff..9c2425b 100644 --- a/CHANGELOG.md +++ b/CHANGELOG.md @@ -1,4 +1,7 @@ ## v0.4 +### v0.4.1 + - Support for horizontal orientation + ### v0.4.0 - Major refactoring, change to an Annotator `class` to prepare and add annotations in separate function (now method) calls. diff --git a/README.md b/README.md index 7e9b432..fff5153 100644 --- a/README.md +++ b/README.md @@ -42,6 +42,7 @@ corresponding branch). - **Interface to use any other function from any source with minimal extra code** - Smart layout of multiple annotations with correct y offsets. +- **Support for vertical and horizontal orientation** - Annotations can be located inside or outside the plot. - **Corrections for multiple testing can be applied (binding to `statsmodels.stats.multitest.multipletests` methods):** @@ -51,7 +52,7 @@ corresponding branch). - Benjamini-Yekutieli - **And any other function from any source with minimal extra code** - Format of the statistical test annotation can be customized: - star annotation, simplified p-value, or explicit p-value. + star annotation, simplified p-value format, or explicit p-value. - Optionally, custom p-values can be given as input. In this case, no statistical test is performed, but **corrections for multiple testing can be applied.** diff --git a/docs/build/doctrees/environment.pickle b/docs/build/doctrees/environment.pickle index 771aba2..169ad27 100644 Binary files a/docs/build/doctrees/environment.pickle and b/docs/build/doctrees/environment.pickle differ diff --git a/docs/build/doctrees/statannotations.doctree b/docs/build/doctrees/statannotations.doctree index ee0691c..759af7f 100644 Binary files a/docs/build/doctrees/statannotations.doctree and b/docs/build/doctrees/statannotations.doctree differ diff --git a/docs/build/html/genindex.html b/docs/build/html/genindex.html index 935f030..e1e1951 100644 --- a/docs/build/html/genindex.html +++ b/docs/build/html/genindex.html @@ -166,6 +166,7 @@ <h1 id="index">Index</h1> | <a href="#L">L</a> | <a href="#M">M</a> | <a href="#N">N</a> + | <a href="#O">O</a> | <a href="#P">P</a> | <a href="#R">R</a> | <a href="#S">S</a> @@ -365,6 +366,14 @@ <h2 id="N">N</h2> </ul></td> </tr></table> +<h2 id="O">O</h2> +<table style="width: 100%" class="indextable genindextable"><tr> + <td style="width: 33%; vertical-align: top;"><ul> + <li><a href="statannotations.html#statannotations.Annotator.Annotator.orient">orient (statannotations.Annotator.Annotator property)</a> +</li> + </ul></td> +</tr></table> + <h2 id="P">P</h2> <table style="width: 100%" class="indextable genindextable"><tr> <td style="width: 33%; vertical-align: top;"><ul> diff --git a/docs/build/html/objects.inv b/docs/build/html/objects.inv index 0d82dd7..b77fb8b 100644 Binary files a/docs/build/html/objects.inv and b/docs/build/html/objects.inv differ diff --git a/docs/build/html/searchindex.js b/docs/build/html/searchindex.js index 38da9ed..2e1b48e 100644 --- a/docs/build/html/searchindex.js +++ b/docs/build/html/searchindex.js @@ -1,1 +1,1 @@ -Search.setIndex({docnames:["index","modules","setup","statannotations","statannotations.stats"],envversion:{"sphinx.domains.c":2,"sphinx.domains.changeset":1,"sphinx.domains.citation":1,"sphinx.domains.cpp":3,"sphinx.domains.index":1,"sphinx.domains.javascript":2,"sphinx.domains.math":2,"sphinx.domains.python":3,"sphinx.domains.rst":2,"sphinx.domains.std":2,sphinx:56},filenames:["index.rst","modules.rst","setup.rst","statannotations.rst","statannotations.stats.rst"],objects:{"":{statannotations:[3,0,0,"-"]},"statannotations.Annotation":{Annotation:[3,1,1,""]},"statannotations.Annotation.Annotation":{formatted_output:[3,2,1,""],print_labels_and_content:[3,3,1,""],text:[3,2,1,""]},"statannotations.Annotator":{Annotator:[3,1,1,""]},"statannotations.Annotator.Annotator":{alpha:[3,2,1,""],annotate:[3,3,1,""],annotate_custom_annotations:[3,3,1,""],apply_and_annotate:[3,3,1,""],apply_test:[3,3,1,""],comparisons_correction:[3,2,1,""],configure:[3,3,1,""],fig:[3,2,1,""],get_annotations_text:[3,3,1,""],get_configuration:[3,3,1,""],get_offset_func:[3,3,1,""],has_type0_comparisons_correction:[3,3,1,""],loc:[3,2,1,""],new_plot:[3,3,1,""],print_pvalue_legend:[3,3,1,""],pvalue_format:[3,2,1,""],reset_configuration:[3,3,1,""],set_custom_annotations:[3,3,1,""],set_pvalues:[3,3,1,""],set_pvalues_and_annotate:[3,3,1,""],test:[3,2,1,""],validate_test_short_name:[3,3,1,""],verbose:[3,2,1,""]},"statannotations.PValueFormat":{Formatter:[3,1,1,""],PValueFormat:[3,1,1,""],sort_pvalue_thresholds:[3,4,1,""]},"statannotations.PValueFormat.Formatter":{config:[3,3,1,""],format_data:[3,3,1,""]},"statannotations.PValueFormat.PValueFormat":{config:[3,3,1,""],format_data:[3,3,1,""],get_configuration:[3,3,1,""],print_legend_if_used:[3,3,1,""],pvalue_format_string:[3,2,1,""],pvalue_thresholds:[3,2,1,""],simple_format_string:[3,2,1,""],text_format:[3,2,1,""]},"statannotations.format_annotations":{pval_annotation_text:[3,4,1,""],simple_text:[3,4,1,""]},"statannotations.stats":{ComparisonsCorrection:[4,0,0,"-"],StatResult:[4,0,0,"-"],StatTest:[4,0,0,"-"],test:[4,0,0,"-"],utils:[4,0,0,"-"]},"statannotations.stats.ComparisonsCorrection":{ComparisonsCorrection:[4,1,1,""],check_valid_correction_name:[4,4,1,""],get_correction_parameters:[4,4,1,""],get_validated_comparisons_correction:[4,4,1,""]},"statannotations.stats.ComparisonsCorrection.ComparisonsCorrection":{apply:[4,3,1,""],document:[4,3,1,""]},"statannotations.stats.StatResult":{StatResult:[4,1,1,""]},"statannotations.stats.StatResult.StatResult":{corrected_significance:[4,2,1,""],correction_method:[4,2,1,""],formatted_output:[4,2,1,""],significance_suffix:[4,2,1,""]},"statannotations.stats.StatTest":{StatTest:[4,1,1,""],wilcoxon:[4,4,1,""]},"statannotations.stats.StatTest.StatTest":{from_library:[4,3,1,""],short_name:[4,2,1,""]},"statannotations.stats.test":{apply_test:[4,4,1,""]},"statannotations.stats.utils":{check_alpha:[4,4,1,""],check_num_comparisons:[4,4,1,""],check_pvalues:[4,4,1,""],get_num_comparisons:[4,4,1,""],return_results:[4,4,1,""]},"statannotations.utils":{check_is_in:[3,4,1,""],check_not_none:[3,4,1,""],check_order_in_data:[3,4,1,""],check_pairs_in_data:[3,4,1,""],check_valid_text_format:[3,4,1,""],get_closest:[3,4,1,""],get_x_values:[3,4,1,""],raise_expected_got:[3,4,1,""],remove_null:[3,4,1,""],render_collection:[3,4,1,""]},statannotations:{Annotation:[3,0,0,"-"],Annotator:[3,0,0,"-"],PValueFormat:[3,0,0,"-"],format_annotations:[3,0,0,"-"],stats:[4,0,0,"-"],utils:[3,0,0,"-"]}},objnames:{"0":["py","module","Python module"],"1":["py","class","Python class"],"2":["py","property","Python property"],"3":["py","method","Python method"],"4":["py","function","Python function"]},objtypes:{"0":"py:module","1":"py:class","2":"py:property","3":"py:method","4":"py:function"},terms:{"0":[3,4],"05":[3,4],"1":[3,4],"12141511":3,"9981846":3,"case":4,"class":[3,4],"default":[3,4],"float":4,"function":[3,4],"int":4,"return":3,"short":3,"static":4,"true":[3,4],A:3,If:3,Or:3,The:3,a_list:3,accept:3,add:3,addit:4,alpha:[3,4],an:3,annot:1,annotate_custom_annot:3,appli:[3,4],apply_and_annot:3,apply_test:[3,4],ar:3,arg:[3,4],argument:[3,4],arrai:3,assum:3,auto:3,ax:3,bar:3,base:[3,4],behavior:4,benjamini:4,between:3,bonferroni:4,bool:4,box:3,boxplot:3,calcul:3,callabl:4,check:3,check_alpha:4,check_is_in:3,check_not_non:3,check_num_comparison:4,check_order_in_data:3,check_pairs_in_data:3,check_pvalu:4,check_valid_correction_nam:4,check_valid_text_format:3,classmethod:3,close:3,closest:3,collect:3,color:3,com:3,comparison:[3,4],comparisons_correct:[3,4],comparisonscorrect:[1,3],comput:3,config:3,configur:3,constructor:3,content:1,coordin:3,core:3,corr_kwarg:4,correct:4,corrected_signific:4,correction_method:4,custom:3,data:[3,4],datafram:3,dict:4,displai:3,document:4,each:3,earlier:4,either:4,engin:3,equal:3,equival:4,error:3,error_typ:3,exact:3,exist:3,expect:3,fals:3,fig:3,follow:3,for_:3,format:[3,4],format_annot:1,format_data:3,formatt:3,formatted_output:[3,4],fraction:3,frame:3,from:[3,4],from_librari:4,func:4,gener:3,get:4,get_annotations_text:3,get_closest:3,get_configur:3,get_correction_paramet:4,get_num_comparison:4,get_offset_func:3,get_validated_comparisons_correct:4,get_x_valu:3,got:3,group:3,group_data1:4,group_data2:4,gt:4,has_type0_comparisons_correct:3,hochberg:4,hold:3,holm:4,http:3,hue:3,hue_ord:3,index:0,instanc:[3,4],instead:3,interfac:4,interpret:4,keyword:4,kruskal:4,kwarg:[3,4],label:3,leven:4,line_height:3,line_offset:3,line_offset_to_group:3,line_width:3,link:3,list:3,loc:3,ls:4,mann:4,messag:3,method:4,method_typ:4,mode:3,modul:[0,1],multipl:4,must:3,mylist:3,mynumb:3,name:[3,4],new_plot:3,none:[3,4],num_comparison:[3,4],number:[3,4],object:[3,4],one:[3,4],option:[3,4],order:3,otherwis:3,overrid:3,p:3,p_valu:4,packag:[0,1],page:0,pair:3,panda:3,paramet:[3,4],pass:[3,4],per:3,perform:3,plot:3,plot_param:3,point:3,posit:3,print_labels_and_cont:3,print_legend_if_us:3,print_pvalue_legend:3,properti:[3,4],provid:[3,4],pval:4,pval_annotation_text:3,pvalu:[3,4],pvalue_format:3,pvalue_format_str:3,pvalue_threshold:3,pvalueformat:1,rais:3,raise_expected_got:3,rang:3,refer:3,remove_nul:3,render:3,render_collect:3,reset_configur:3,result:[3,4],results_arrai:4,return_result:4,run:4,same:3,sampl:4,seaborn:3,search:0,sep:3,seri:3,set:3,set_custom_annot:3,set_pvalu:3,set_pvalues_and_annot:3,short_nam:4,show:3,show_test_nam:3,significance_suffix:4,simpl:3,simple_format_str:3,simple_text:3,smallest:3,sort:3,sort_pvalue_threshold:3,stackoverflow:3,standard:3,stat:[1,3],stat_nam:4,stat_str:4,statist:[3,4],statresult:[1,3],stats_param:[3,4],statsmodel:4,statsmodels_api:4,stattest:[1,3],str:[3,4],string:3,struct:3,submodul:1,subpackag:1,t:4,test:[1,3],test_descript:4,test_ind:4,test_long_nam:4,test_nam:4,test_pair:4,test_result_list:4,test_short_nam:[3,4],test_welch:4,text:3,text_annot_custom:3,text_format:3,text_offset:3,them:3,thereof:3,thi:[3,4],three:3,threshold:3,top:3,tupl:3,two:[3,4],type:3,union:[3,4],us:[3,4],use_fixed_offset:3,util:1,valid_valu:3,validate_test_short_nam:3,valu:[3,4],valueerror:3,verbos:[3,4],version:4,vs:3,wa:3,whitnei:4,wilcoxon:4,work:3,x:3,y:3,yekuti:4},titles:["Welcome to statannotations\u2019s documentation!","statannotations","setup module","statannotations package","statannotations.stats package"],titleterms:{annot:3,comparisonscorrect:4,content:[0,3,4],document:0,format_annot:3,indic:0,modul:[2,3,4],packag:[3,4],pvalueformat:3,s:0,setup:2,stat:4,statannot:[0,1,3,4],statresult:4,stattest:4,submodul:[3,4],subpackag:3,tabl:0,test:4,util:[3,4],welcom:0}}) \ No newline at end of file +Search.setIndex({docnames:["index","modules","setup","statannotations","statannotations.stats"],envversion:{"sphinx.domains.c":2,"sphinx.domains.changeset":1,"sphinx.domains.citation":1,"sphinx.domains.cpp":3,"sphinx.domains.index":1,"sphinx.domains.javascript":2,"sphinx.domains.math":2,"sphinx.domains.python":3,"sphinx.domains.rst":2,"sphinx.domains.std":2,sphinx:56},filenames:["index.rst","modules.rst","setup.rst","statannotations.rst","statannotations.stats.rst"],objects:{"":{statannotations:[3,0,0,"-"]},"statannotations.Annotation":{Annotation:[3,1,1,""]},"statannotations.Annotation.Annotation":{formatted_output:[3,2,1,""],print_labels_and_content:[3,3,1,""],text:[3,2,1,""]},"statannotations.Annotator":{Annotator:[3,1,1,""]},"statannotations.Annotator.Annotator":{alpha:[3,2,1,""],annotate:[3,3,1,""],annotate_custom_annotations:[3,3,1,""],apply_and_annotate:[3,3,1,""],apply_test:[3,3,1,""],comparisons_correction:[3,2,1,""],configure:[3,3,1,""],fig:[3,2,1,""],get_annotations_text:[3,3,1,""],get_configuration:[3,3,1,""],get_offset_func:[3,3,1,""],has_type0_comparisons_correction:[3,3,1,""],loc:[3,2,1,""],new_plot:[3,3,1,""],orient:[3,2,1,""],print_pvalue_legend:[3,3,1,""],pvalue_format:[3,2,1,""],reset_configuration:[3,3,1,""],set_custom_annotations:[3,3,1,""],set_pvalues:[3,3,1,""],set_pvalues_and_annotate:[3,3,1,""],test:[3,2,1,""],validate_test_short_name:[3,3,1,""],verbose:[3,2,1,""]},"statannotations.PValueFormat":{Formatter:[3,1,1,""],PValueFormat:[3,1,1,""],sort_pvalue_thresholds:[3,4,1,""]},"statannotations.PValueFormat.Formatter":{config:[3,3,1,""],format_data:[3,3,1,""]},"statannotations.PValueFormat.PValueFormat":{config:[3,3,1,""],format_data:[3,3,1,""],get_configuration:[3,3,1,""],print_legend_if_used:[3,3,1,""],pvalue_format_string:[3,2,1,""],pvalue_thresholds:[3,2,1,""],simple_format_string:[3,2,1,""],text_format:[3,2,1,""]},"statannotations.format_annotations":{pval_annotation_text:[3,4,1,""],simple_text:[3,4,1,""]},"statannotations.stats":{ComparisonsCorrection:[4,0,0,"-"],StatResult:[4,0,0,"-"],StatTest:[4,0,0,"-"],test:[4,0,0,"-"],utils:[4,0,0,"-"]},"statannotations.stats.ComparisonsCorrection":{ComparisonsCorrection:[4,1,1,""],check_valid_correction_name:[4,4,1,""],get_correction_parameters:[4,4,1,""],get_validated_comparisons_correction:[4,4,1,""]},"statannotations.stats.ComparisonsCorrection.ComparisonsCorrection":{apply:[4,3,1,""],document:[4,3,1,""]},"statannotations.stats.StatResult":{StatResult:[4,1,1,""]},"statannotations.stats.StatResult.StatResult":{corrected_significance:[4,2,1,""],correction_method:[4,2,1,""],formatted_output:[4,2,1,""],significance_suffix:[4,2,1,""]},"statannotations.stats.StatTest":{StatTest:[4,1,1,""],wilcoxon:[4,4,1,""]},"statannotations.stats.StatTest.StatTest":{from_library:[4,3,1,""],short_name:[4,2,1,""]},"statannotations.stats.test":{apply_test:[4,4,1,""]},"statannotations.stats.utils":{check_alpha:[4,4,1,""],check_num_comparisons:[4,4,1,""],check_pvalues:[4,4,1,""],get_num_comparisons:[4,4,1,""],return_results:[4,4,1,""]},"statannotations.utils":{check_is_in:[3,4,1,""],check_not_none:[3,4,1,""],check_order_in_data:[3,4,1,""],check_pairs_in_data:[3,4,1,""],check_valid_text_format:[3,4,1,""],get_closest:[3,4,1,""],get_x_values:[3,4,1,""],raise_expected_got:[3,4,1,""],remove_null:[3,4,1,""],render_collection:[3,4,1,""]},statannotations:{Annotation:[3,0,0,"-"],Annotator:[3,0,0,"-"],PValueFormat:[3,0,0,"-"],format_annotations:[3,0,0,"-"],stats:[4,0,0,"-"],utils:[3,0,0,"-"]}},objnames:{"0":["py","module","Python module"],"1":["py","class","Python class"],"2":["py","property","Python property"],"3":["py","method","Python method"],"4":["py","function","Python function"]},objtypes:{"0":"py:module","1":"py:class","2":"py:property","3":"py:method","4":"py:function"},terms:{"0":[3,4],"05":[3,4],"1":[3,4],"12141511":3,"9981846":3,"case":4,"class":[3,4],"default":[3,4],"float":4,"function":[3,4],"int":4,"return":3,"short":3,"static":4,"true":[3,4],A:3,If:3,Or:3,The:3,a_list:3,accept:3,add:3,addit:4,alpha:[3,4],an:3,annot:1,annotate_custom_annot:3,appli:[3,4],apply_and_annot:3,apply_test:[3,4],ar:3,arg:[3,4],argument:[3,4],arrai:3,assum:3,auto:3,ax:3,bar:3,base:[3,4],behavior:4,benjamini:4,between:3,bonferroni:4,bool:4,box:3,boxplot:3,calcul:3,callabl:4,check:3,check_alpha:4,check_is_in:3,check_not_non:3,check_num_comparison:4,check_order_in_data:3,check_pairs_in_data:3,check_pvalu:4,check_valid_correction_nam:4,check_valid_text_format:3,classmethod:3,close:3,closest:3,collect:3,color:3,com:3,comparison:[3,4],comparisons_correct:[3,4],comparisonscorrect:[1,3],comput:3,config:3,configur:3,constructor:3,content:1,coord:3,coordin:3,core:3,corr_kwarg:4,correct:4,corrected_signific:4,correction_method:4,custom:3,data:[3,4],datafram:3,dict:4,displai:3,document:4,each:3,earlier:4,either:4,engin:3,equal:3,equival:4,error:3,error_typ:3,exact:3,exist:3,expect:3,fals:3,fig:3,follow:3,for_:3,format:[3,4],format_annot:1,format_data:3,formatt:3,formatted_output:[3,4],fraction:3,frame:3,from:[3,4],from_librari:4,func:4,gener:3,get:4,get_annotations_text:3,get_closest:3,get_configur:3,get_correction_paramet:4,get_num_comparison:4,get_offset_func:3,get_validated_comparisons_correct:4,get_x_valu:3,got:3,group:3,group_data1:4,group_data2:4,gt:4,has_type0_comparisons_correct:3,hochberg:4,hold:3,holm:4,http:3,hue:3,hue_ord:3,index:0,instanc:[3,4],instead:3,interfac:4,interpret:4,keyword:4,kruskal:4,kwarg:[3,4],label:3,leven:4,line_height:3,line_offset:3,line_offset_to_group:3,line_width:3,link:3,list:3,loc:3,ls:4,mann:4,messag:3,method:4,method_typ:4,mode:3,modul:[0,1],multipl:4,must:3,mylist:3,mynumb:3,name:[3,4],new_plot:3,none:[3,4],num_comparison:[3,4],number:[3,4],object:[3,4],one:[3,4],option:[3,4],order:3,orient:3,otherwis:3,overrid:3,p:3,p_valu:4,packag:[0,1],page:0,pair:3,panda:3,paramet:[3,4],pass:[3,4],per:3,perform:3,plot:3,plot_param:3,point:3,posit:3,print_labels_and_cont:3,print_legend_if_us:3,print_pvalue_legend:3,properti:[3,4],provid:[3,4],pval:4,pval_annotation_text:3,pvalu:[3,4],pvalue_format:3,pvalue_format_str:3,pvalue_threshold:3,pvalueformat:1,rais:3,raise_expected_got:3,rang:3,refer:3,remove_nul:3,render:3,render_collect:3,reset_configur:3,result:[3,4],results_arrai:4,return_result:4,run:4,same:3,sampl:4,seaborn:3,search:0,sep:3,seri:3,set:3,set_custom_annot:3,set_pvalu:3,set_pvalues_and_annot:3,short_nam:4,show:3,show_test_nam:3,significance_suffix:4,simpl:3,simple_format_str:3,simple_text:3,smallest:3,sort:3,sort_pvalue_threshold:3,stackoverflow:3,standard:3,stat:[1,3],stat_nam:4,stat_str:4,statist:[3,4],statresult:[1,3],stats_param:[3,4],statsmodel:4,statsmodels_api:4,stattest:[1,3],str:[3,4],string:3,struct:3,submodul:1,subpackag:1,t:4,test:[1,3],test_descript:4,test_ind:4,test_long_nam:4,test_nam:4,test_pair:4,test_result_list:4,test_short_nam:[3,4],test_welch:4,text:3,text_annot_custom:3,text_format:3,text_offset:3,them:3,thereof:3,thi:[3,4],three:3,threshold:3,top:3,tupl:3,two:[3,4],type:3,union:[3,4],us:[3,4],use_fixed_offset:3,util:1,valid_valu:3,validate_test_short_nam:3,valu:[3,4],valueerror:3,verbos:[3,4],version:4,vs:3,wa:3,whitnei:4,wilcoxon:4,work:3,x:3,y:3,yekuti:4},titles:["Welcome to statannotations\u2019s documentation!","statannotations","setup module","statannotations package","statannotations.stats package"],titleterms:{annot:3,comparisonscorrect:4,content:[0,3,4],document:0,format_annot:3,indic:0,modul:[2,3,4],packag:[3,4],pvalueformat:3,s:0,setup:2,stat:4,statannot:[0,1,3,4],statresult:4,stattest:4,submodul:[3,4],subpackag:3,tabl:0,test:4,util:[3,4],welcom:0}}) \ No newline at end of file diff --git a/docs/build/html/statannotations.html b/docs/build/html/statannotations.html index aaacef6..7363f25 100644 --- a/docs/build/html/statannotations.html +++ b/docs/build/html/statannotations.html @@ -352,6 +352,11 @@ <h2>Submodules<a class="headerlink" href="#submodules" title="Permalink to this new_plot(ax, pairs=None, plot='boxplot', data=None, x=None, y=None, hue=None, order=None, hue_order=None, engine: str = 'seaborn', **plot_params)<a class="headerlink" href="#statannotations.Annotator.Annotator.new_plot" title="Permalink to this definition">¶</a></dt> <dd></dd></dl> +<dl class="py property"> +<dt class="sig sig-object py" id="statannotations.Annotator.Annotator.orient"> +property orient<a class="headerlink" href="#statannotations.Annotator.Annotator.orient" title="Permalink to this definition">¶</a></dt> +<dd></dd></dl> + <dl class="py method"> <dt class="sig sig-object py" id="statannotations.Annotator.Annotator.print_pvalue_legend"> print_pvalue_legend()<a class="headerlink" href="#statannotations.Annotator.Annotator.print_pvalue_legend" title="Permalink to this definition">¶</a></dt> @@ -544,7 +549,7 @@ <h2>Submodules<a class="headerlink" href="#submodules" title="Permalink to this <dl class="py function"> <dt class="sig sig-object py" id="statannotations.utils.check_pairs_in_data"> -statannotations.utils.check_pairs_in_data(pairs: Union[list, tuple], data: Optional[Union[List[list], pandas.core.frame.DataFrame]] = None, x: Optional[Union[str, list]] = None, hue: Optional[str] = None, hue_order: Optional[List[str]] = None)<a class="headerlink" href="#statannotations.utils.check_pairs_in_data" title="Permalink to this definition">¶</a></dt> +statannotations.utils.check_pairs_in_data(pairs: Union[list, tuple], data: Optional[Union[List[list], pandas.core.frame.DataFrame]] = None, coord: Optional[Union[str, list]] = None, hue: Optional[str] = None, hue_order: Optional[List[str]] = None)<a class="headerlink" href="#statannotations.utils.check_pairs_in_data" title="Permalink to this definition">¶</a></dt> <dd>Checks that values referred to in <cite>order</cite> and <cite>pairs</cite> exist in data. </dd></dl> diff --git a/statannotations/Annotator.py b/statannotations/Annotator.py index 390bd83..00dcee6 100644 --- a/statannotations/Annotator.py +++ b/statannotations/Annotator.py @@ -123,7 +123,7 @@ def __init__(self, ax, pairs, plot='boxplot', data=None, x=None, self.text_offset_impact_above = 0 self.color = '0.2' self.line_width = 1.5 - self.y_offset = None + self.value_offset = None self.custom_annotations = None def new_plot(self, ax, pairs=None, plot='boxplot', data=None, x=None, @@ -144,6 +144,10 @@ def new_plot(self, ax, pairs=None, plot='boxplot', data=None, x=None, return self + @property + def orient(self): + return self._plotter.orient + @property def verbose(self): return self._verbose @@ -154,8 +158,8 @@ def verbose(self, verbose): self._plotter.verbose = verbose @property - def _y_stack_arr(self): - return self._plotter.y_stack_arr + def _value_stack_arr(self): + return self._plotter.value_stack_arr @property def fig(self): @@ -179,13 +183,13 @@ def annotate(self, line_offset=None, line_offset_to_group=None): "test (again) which will probably lead to " "unexpected results") - self._update_y_for_loc() + self._update_value_for_loc() ann_list = [] - orig_ylim = self.ax.get_ylim() + orig_value_lim = self._plotter.get_value_lim() offset_func = self.get_offset_func(self.loc) - self.y_offset, self.line_offset_to_group = offset_func( + self.value_offset, self.line_offset_to_group = offset_func( line_offset, line_offset_to_group) if self._verbose: @@ -199,16 +203,24 @@ def annotate(self, line_offset=None, line_offset_to_group=None): self._annotate_pair(annotation, ax_to_data=ax_to_data, ann_list=ann_list, - orig_ylim=orig_ylim) + orig_value_lim=orig_value_lim) # reset transformation - y_stack_max = max(self._y_stack_arr[1, :]) + y_stack_max = max(self._value_stack_arr[1, :]) ax_to_data = self._plotter.get_transform_func('ax_to_data') - ylims = ([(0, 0), (0, max(1.04 * y_stack_max, 1))] - if self.loc == 'inside' - else [(0, 0), (0, 1)]) - - self.ax.set_ylim(ax_to_data.transform(ylims)[:, 1]) + value_lims = ( + ([(0, 0), (0, max(1.04 * y_stack_max, 1))] + if self.loc == 'inside' + else [(0, 0), (0, 1)]) + if self.orient == 'v' + else + ([(0, 0), (max(1.04 * y_stack_max, 1), 0)] + if self.loc == 'inside' + else [(0, 0), (1, 0)]) + ) + set_lims = self.ax.set_ylim if self.orient == 'v' else self.ax.set_xlim + transformed = ax_to_data.transform(value_lims) + set_lims(transformed[:, 1 if self.orient == 'v' else 0]) return self.ax, self.annotations @@ -444,61 +456,75 @@ def _get_results(self, num_comparisons, pvalues=None, return test_result_list - def _annotate_pair(self, annotation, ax_to_data, ann_list, orig_ylim): + def _annotate_pair(self, annotation, ax_to_data, ann_list, orig_value_lim): if self._verbose >= 1: annotation.print_labels_and_content() - x1 = annotation.structs[0]['x'] - x2 = annotation.structs[1]['x'] - xi1 = annotation.structs[0]['xi'] - xi2 = annotation.structs[1]['xi'] + group_coord_1 = annotation.structs[0]['group_coord'] + group_coord_2 = annotation.structs[1]['group_coord'] + group_i1 = annotation.structs[0]['group_i'] + group_i2 = annotation.structs[1]['group_i'] # Find y maximum for all the y_stacks *in between* group1 and group2 - i_ymax_in_range_x1_x2 = xi1 + np.nanargmax( - self._y_stack_arr[1, np.where((x1 <= self._y_stack_arr[0, :]) - & (self._y_stack_arr[0, :] <= x2))]) + i_value_max_in_range_g1_g2 = group_i1 + np.nanargmax( + self._value_stack_arr[1, + np.where((group_coord_1 + <= self._value_stack_arr[0, :]) + & (self._value_stack_arr[0, :] + <= group_coord_2))]) - y = self._get_y_for_pair(i_ymax_in_range_x1_x2) + value = self._get_value_for_pair(i_value_max_in_range_g1_g2) # Determine lines in axes coordinates - ax_line_x = [x1, x1, x2, x2] - ax_line_y = [y, y + self.line_height, y + self.line_height, y] + ax_line_group = [group_coord_1, group_coord_1, + group_coord_2, group_coord_2] + ax_line_value = [value, value + self.line_height, + value + self.line_height, value] + + lists = ((ax_line_group, ax_line_value) if self.orient == 'v' + else (ax_line_value, ax_line_group)) points = [ax_to_data.transform((x, y)) for x, y - in zip(ax_line_x, ax_line_y)] + in zip(*lists)] line_x, line_y = zip(*points) self._plot_line(line_x, line_y) + xy_params = self._get_xy_params(group_coord_1, group_coord_2, line_x, + line_y) + ann = self.ax.annotate( - annotation.text, xy=(np.mean([x1, x2]), line_y[2]), - xytext=(0, self.text_offset), textcoords='offset points', + annotation.text, textcoords='offset points', xycoords='data', ha='center', va='bottom', fontsize=self._pvalue_format.fontsize, clip_on=False, - annotation_clip=False) + annotation_clip=False, **xy_params) + if annotation.text is not None: ann_list.append(ann) plt.draw() - self.ax.set_ylim(orig_ylim) + set_lim = {'v': 'set_ylim', + 'h': 'set_xlim'}[self.orient] + + getattr(self.ax, set_lim)(orig_value_lim) - y_top_annot = self._annotate_pair_text(ann, y) + value_top_annot = self._annotate_pair_text(ann, value) else: - y_top_annot = y + self.line_height + value_top_annot = value + self.line_height - # Fill the highest y position of the annotation into the y_stack array - # for all positions in the range x1 to x2 - self._y_stack_arr[1, - (x1 <= self._y_stack_arr[0, :]) - & (self._y_stack_arr[0, :] <= x2)] = y_top_annot + # Fill the highest value position of the annotation into the value_stack + # array for all positions in the range group_coord_1 to group_coord_2 + self._value_stack_arr[ + 1, (group_coord_1 <= self._value_stack_arr[0, :]) + & (self._value_stack_arr[0, :] <= group_coord_2)] = value_top_annot - # Increment the counter of annotations in the y_stack array - self._y_stack_arr[2, xi1:xi2 + 1] += 1 + # Increment the counter of annotations in the value_stack array + self._value_stack_arr[2, group_i1:group_i2 + 1] += 1 - def _update_y_for_loc(self): + def _update_value_for_loc(self): if self._loc == 'outside': - self._y_stack_arr[1, :] = 1 + self._value_stack_arr[1, :] = 1 def _check_test_pvalues_perform(self): if self.test is None: @@ -582,8 +608,7 @@ def _get_offsets_inside(line_offset, line_offset_to_group): else: if line_offset_to_group is None: line_offset_to_group = 0.06 - y_offset = line_offset - return y_offset, line_offset_to_group + return line_offset, line_offset_to_group @staticmethod def _get_offsets_outside(line_offset, line_offset_to_group): @@ -623,9 +648,9 @@ def _plot_line(self, line_x, line_y): line.set_clip_on(False) self.ax.add_line(line) - def _annotate_pair_text(self, ann, y): + def _annotate_pair_text(self, ann, value): - y_top_annot = None + value_top_annot = None got_mpl_error = False if not self.use_fixed_offset: @@ -633,7 +658,8 @@ def _annotate_pair_text(self, ann, y): bbox = ann.get_window_extent() pix_to_ax = self._plotter.get_transform_func('pix_to_ax') bbox_ax = bbox.transformed(pix_to_ax) - y_top_annot = bbox_ax.ymax + value_coord_max = {'v': 'ymax', 'h': 'xmax'}[self.orient] + value_top_annot = getattr(bbox_ax, value_coord_max) except RuntimeError: got_mpl_error = True @@ -644,40 +670,45 @@ def _annotate_pair_text(self, ann, y): "back to a fixed y offset. Layout may be not " "optimal.") - # We will apply a fixed offset in points, - # based on the font size of the annotation. fontsize_points = FontProperties( size='medium').get_size_in_points() - offset_trans = mtransforms.offset_copy( - self.ax.transAxes, fig=self.fig, x=0, - y=fontsize_points + self.text_offset, units='points') - y_top_display = offset_trans.transform( - (0, y + self.line_height)) - y_top_annot = (self.ax.transAxes.inverted() - .transform(y_top_display)[1]) + direction = {'h': -1, 'v': 1}[self.orient] + x, y = [0, fontsize_points + self.text_offset][::direction] + offset_trans = mtransforms.offset_copy(trans=self.ax.transAxes, + fig=self.fig, + units='points', x=x, y=y) + + value_top_display = offset_trans.transform( + (value + self.line_height, value + self.line_height)) + + value_coord = {'h': 0, 'v': 1}[self.orient] + + value_top_annot = (self.ax.transAxes.inverted() + .transform(value_top_display)[value_coord]) self.text_offset_impact_above = ( - y_top_annot - y - self.y_offset - self.line_height) - return y_top_annot + value_top_annot - value - self.value_offset - self.line_height) + + return value_top_annot def _reset_default_values(self): for attribute, default_value in _DEFAULT_VALUES.items(): setattr(self, attribute, default_value) - def _get_y_for_pair(self, i_ymax_in_range_x1_x2): + def _get_value_for_pair(self, i_ymax_in_range_x1_x2): - ymax_in_range_x1_x2 = self._y_stack_arr[1, i_ymax_in_range_x1_x2] + ymax_in_range_x1_x2 = self._value_stack_arr[1, i_ymax_in_range_x1_x2] # Choose the best offset depending on whether there is an annotation # below at the x position in the range [x1, x2] where the stack is the # highest - if self._y_stack_arr[2, i_ymax_in_range_x1_x2] == 0: + if self._value_stack_arr[2, i_ymax_in_range_x1_x2] == 0: # there is only a group below offset = self.line_offset_to_group else: # there is an annotation below - offset = self.y_offset + self.text_offset_impact_above + offset = self.value_offset + self.text_offset_impact_above return ymax_in_range_x1_x2 + offset @@ -697,3 +728,28 @@ def _get_plotter(engine, *args, **kwargs): if engine_plotter is None: raise NotImplementedError(f"{engine} engine not implemented.") return engine_plotter(*args, **kwargs) + + def _get_xy_params_horizontal(self, group_coord_1, group_coord_2, + line_x: np.ndarray): + return { + 'xy': (line_x[2], np.mean([group_coord_1, group_coord_2])), + 'xytext': (self.text_offset, 0), + 'rotation': 270, + 'rotation_mode': 'anchor' + } + + def _get_xy_params_vertical(self, group_coord_1, group_coord_2, + line_y: np.ndarray): + return { + 'xy': (np.mean([group_coord_1, group_coord_2]), line_y[2]), + 'xytext': (0, self.text_offset), + } + + def _get_xy_params(self, group_coord_1, group_coord_2, line_x: np.ndarray, + line_y: np.ndarray): + if self.orient == 'h': + return self._get_xy_params_horizontal(group_coord_1, group_coord_2, + line_x) + + return self._get_xy_params_vertical(group_coord_1, group_coord_2, + line_y) diff --git a/statannotations/_Xpositions.py b/statannotations/_GroupsPositions.py similarity index 51% rename from statannotations/_Xpositions.py rename to statannotations/_GroupsPositions.py index 2520f3c..046d2b9 100644 --- a/statannotations/_Xpositions.py +++ b/statannotations/_GroupsPositions.py @@ -3,7 +3,7 @@ from statannotations.utils import get_closest -class _XPositions: +class _GroupsPositions: def __init__(self, plotter, group_names): self._plotter = plotter self._hue_names = self._plotter.hue_names @@ -15,40 +15,42 @@ def __init__(self, plotter, group_names): "Using hues with only one hue is not supported.") self.hue_offsets = self._plotter.hue_offsets - self._xunits = self.hue_offsets[1] - self.hue_offsets[0] + self._axis_units = self.hue_offsets[1] - self.hue_offsets[0] - self._xpositions = { - np.round(self.get_group_x_position(group_name), 1): group_name + self._groups_positions = { + np.round(self.get_group_axis_position(group_name), 1): group_name for group_name in group_names } - self._xpositions_list = sorted(self._xpositions.keys()) + self._groups_positions_list = sorted(self._groups_positions.keys()) if self._hue_names is None: - self._xunits = ((max(list(self._xpositions.keys())) + 1) - / len(self._xpositions)) + self._axis_units = ((max(list(self._groups_positions.keys())) + 1) + / len(self._groups_positions)) - self._xranges = { - (pos - self._xunits / 2, pos + self._xunits / 2, pos): group_name - for pos, group_name in self._xpositions.items()} + self._axis_ranges = { + (pos - self._axis_units / 2, + pos + self._axis_units / 2, + pos): group_name + for pos, group_name in self._groups_positions.items()} @property - def xpositions(self): - return self._xpositions + def axis_positions(self): + return self._groups_positions @property - def xunits(self): - return self._xunits + def axis_units(self): + return self._axis_units - def get_xpos_location(self, pos): + def get_axis_pos_location(self, pos): """ Finds the x-axis location of a categorical variable """ - for xrange in self._xranges: - if (pos >= xrange[0]) & (pos <= xrange[1]): - return xrange[2] + for axis_range in self._axis_ranges: + if (pos >= axis_range[0]) & (pos <= axis_range[1]): + return axis_range[2] - def get_group_x_position(self, group): + def get_group_axis_position(self, group): """ group_name can be either a name "cat" or a tuple ("cat", "hue") """ @@ -64,5 +66,5 @@ def get_group_x_position(self, group): group_pos = self._plotter.group_names.index(cat) + hue_offset return group_pos - def find_closest(self, xpos): - return get_closest(list(self._xpositions_list), xpos) + def find_closest(self, pos): + return get_closest(list(self._groups_positions_list), pos) diff --git a/statannotations/_Plotter.py b/statannotations/_Plotter.py index 9dac681..5223865 100644 --- a/statannotations/_Plotter.py +++ b/statannotations/_Plotter.py @@ -8,7 +8,7 @@ from matplotlib.collections import PathCollection from matplotlib.patches import Rectangle -from statannotations._Xpositions import _XPositions +from statannotations._GroupsPositions import _GroupsPositions from statannotations.utils import check_not_none, check_order_in_data, \ check_pairs_in_data, render_collection, check_is_in, remove_null @@ -18,16 +18,18 @@ class _Plotter: - def __init__(self, ax, pairs, data=None, x=None, hue=None, - order=None, hue_order=None, verbose=False): + def __init__(self, ax, pairs, data=None, x=None, y=None, hue=None, + order=None, hue_order=None, verbose=False, **plot_params): self.ax = ax self._fig = plt.gcf() check_not_none("pairs", pairs) - check_order_in_data(data, x, order) - check_pairs_in_data(pairs, data, x, hue, hue_order) + group_coord = y if plot_params.get("orient") == "h" else x + check_order_in_data(data, group_coord, order) + check_pairs_in_data(pairs, data, group_coord, hue, hue_order) self.pairs = pairs self._struct_pairs = None self.verbose = verbose + self.orient = plot_params.get("orient", "v") def get_transform_func(self, kind: str): """ @@ -48,9 +50,11 @@ def get_transform_func(self, kind: str): if kind == 'pix_to_ax': return self.ax.transAxes.inverted() - data_to_ax = \ - self.ax.transData + self.ax.get_xaxis_transform().inverted() + transform = {'v': self.ax.get_xaxis_transform, + 'h': self.ax.get_yaxis_transform}[self.orient] + data_to_ax = \ + self.ax.transData + transform().inverted() if kind == 'data_to_ax': return data_to_ax @@ -75,34 +79,35 @@ def __init__(self, ax, pairs, plot='boxplot', data=None, x=None, y=None, hue=None, order=None, hue_order=None, verbose=False, **plot_params): - _Plotter.__init__(self, ax, pairs, data, x, hue, order, hue_order, - verbose) + _Plotter.__init__(self, ax, pairs, data, x, y, hue, order, hue_order, + verbose, **plot_params) self.check_plot_is_implemented(plot) - self.plot = plot self.plotter = self._get_plotter(plot, x, y, hue, data, order, hue_order, **plot_params) self.group_names, self.labels = self._get_group_names_and_labels() - self.xpositions = _XPositions(self.plotter, self.group_names) + self.groups_positions = _GroupsPositions(self.plotter, + self.group_names) self.reordering = None - self.ymaxes = self._generate_ymaxes() + self.value_maxes = self._generate_value_maxes() self.structs = self._get_structs() self.pairs = pairs self._struct_pairs = self._get_group_struct_pairs() - self._y_stack_arr = np.array( - [[struct['x'], struct['ymax'], 0] for struct in self.structs] + self._value_stack_arr = np.array( + [[struct['group_coord'], struct['value_max'], 0] + for struct in self.structs] ).T self.reordering = None self._sort_group_struct_pairs() @property - def y_stack_arr(self): - return self._y_stack_arr + def value_stack_arr(self): + return self._value_stack_arr # noinspection PyProtectedMember def _get_plotter(self, plot, x, y, hue, data, order, hue_order, @@ -187,7 +192,7 @@ def _get_group_names_and_labels(self): return group_names, labels - def _generate_ymaxes(self): + def _generate_value_maxes(self): """ given plotter and the names of two categorical variables, returns highest y point drawn between those two variables before @@ -197,38 +202,40 @@ def _generate_ymaxes(self): (eg, error bars and/or bar charts). """ - ymaxes = {name: 0 for name in self.group_names} + value_maxes = {name: 0 for name in self.group_names} data_to_ax = self.get_transform_func('data_to_ax') if self.plot == 'violinplot': - ymaxes = self._get_ymaxes_violin(ymaxes, data_to_ax) + value_maxes = self._get_value_maxes_violin(value_maxes, data_to_ax) else: for child in self.ax.get_children(): - xname, ypos = self._get_child_ypos(child, data_to_ax) + group_name, value_pos = self._get_value_pos(child, data_to_ax) - if ypos is not None and ypos > ymaxes[xname]: - ymaxes[xname] = ypos + if value_pos is not None and value_pos > value_maxes[group_name]: + value_maxes[group_name] = value_pos - return ymaxes + return value_maxes def _get_structs(self): structs = [ { 'group': group_name, 'label': self.labels[b_idx], - 'x': self.xpositions.get_group_x_position(group_name), + 'group_coord': (self.groups_positions + .get_group_axis_position(group_name)), 'group_data': self._get_group_data(group_name), - 'ymax': self.ymaxes[group_name] - } for b_idx, group_name in enumerate(self.group_names)] + 'value_max': self.value_maxes[group_name] + } + for b_idx, group_name in enumerate(self.group_names)] - # Sort the group data structures by position along the x axis - structs = sorted(structs, key=lambda struct: struct['x']) + # Sort the group data structures by position along the groups axis + structs = sorted(structs, key=lambda struct: struct['group_coord']) - # Add the index position in the list of groups along the x axis - structs = [dict(struct, xi=i) + # Add the index position in the list of groups along the groups axis + structs = [dict(struct, group_i=i) for i, struct in enumerate(structs)] return structs @@ -246,7 +253,7 @@ def _get_group_struct_pairs(self): group_struct2 = dict(group_structs_dic[group2], i_group_pair=i_group_pair) - if group_struct1['x'] <= group_struct2['x']: + if group_struct1['group_coord'] <= group_struct2['group_coord']: group_struct_pairs.append((group_struct1, group_struct2)) else: @@ -289,62 +296,79 @@ def _get_group_data(self, group_name): return group_data - def _get_ymaxes_violin(self, ymaxes, data_to_ax): + def _get_value_maxes_violin(self, value_maxes, data_to_ax): for group_idx, group_name in enumerate(self.plotter.group_names): if self.plotter.hue_names: for hue_idx, hue_name in enumerate(self.plotter.hue_names): - ypos = max(self.plotter.support[group_idx][hue_idx]) - ymaxes[(group_name, hue_name)] = \ - data_to_ax.transform((0, ypos))[1] + value_pos = max(self.plotter.support[group_idx][hue_idx]) + value_maxes[(group_name, hue_name)] = \ + data_to_ax.transform((0, value_pos))[1] else: - ypos = max(self.plotter.support[group_idx]) - ymaxes[group_name] = data_to_ax.transform((0, ypos))[1] - return ymaxes + value_pos = max(self.plotter.support[group_idx]) + value_maxes[group_name] = data_to_ax.transform((0, + value_pos))[1] + return value_maxes - def _get_child_ypos(self, child, data_to_ax): + def _get_value_pos(self, child, data_to_ax): if (type(child) == PathCollection and len(child.properties()['offsets'])): - return self._get_ypos_for_path_collection( + return self._get_value_pos_for_path_collection( child, data_to_ax) elif type(child) in (lines.Line2D, Rectangle): - return self._get_ypos_for_line2d_or_rectangle( + return self._get_value_pos_for_line2d_or_rectangle( child, data_to_ax) return None, None - def _get_ypos_for_path_collection(self, child, data_to_ax): - ymax = child.properties()['offsets'][:, 1].max() - xpos = float(np.round(np.nanmean( - child.properties()['offsets'][:, 0]), 1)) - if xpos not in self.xpositions.xpositions: + def _get_value_pos_for_path_collection(self, child, data_to_ax): + group_coord = {"v": 0, "h": 1}[self.orient] + value_coord = (group_coord + 1) % 2 + + direction = {"v": 1, "h": -1}[self.orient] + + value_max = child.properties()['offsets'][:, value_coord].max() + group_pos = float(np.round(np.nanmean( + child.properties()['offsets'][:, group_coord]), 1)) + + if group_pos not in self.groups_positions.axis_positions: if self.verbose: warnings.warn( "Invalid x-position found. Are the same parameters passed " "to seaborn and statannotations calls? or are there few " "data points?") - xpos = self.xpositions.find_closest(xpos) + group_pos = self.groups_positions.find_closest(group_pos) + + group_name = self.groups_positions.axis_positions[group_pos] - xname = self.xpositions.xpositions[xpos] - ypos = data_to_ax.transform((0, ymax))[1] + value_pos = data_to_ax.transform((0, value_max)[::direction]) - return xname, ypos + return group_name, value_pos[value_coord] - def _get_ypos_for_line2d_or_rectangle(self, child, data_to_ax): + def _get_value_pos_for_line2d_or_rectangle(self, child, data_to_ax): bbox = self.ax.transData.inverted().transform( child.get_window_extent(self.fig.canvas.get_renderer())) - if (bbox[:, 0].max() - bbox[:, 0].min()) > 1.1*self.xpositions.xunits: + group_coord = {"v": 0, "h": 1}[self.orient] + direction = {"v": 1, "h": -1}[self.orient] + + value_coord = (group_coord + 1) % 2 + + if ((bbox[:, group_coord].max() - bbox[:, group_coord].min()) + > 1.1 * self.groups_positions.axis_units): return None, None - raw_xpos = np.round(bbox[:, 0].mean(), 1) - xpos = self.xpositions.get_xpos_location(raw_xpos) - if xpos not in self.xpositions.xpositions: + + raw_group_pos = np.round(bbox[:, group_coord].mean(), 1) + group_pos = self.groups_positions.get_axis_pos_location(raw_group_pos) + + if group_pos not in self.groups_positions.axis_positions: return None, None - xname = self.xpositions.xpositions[xpos] - ypos = bbox[:, 1].max() - ypos = data_to_ax.transform((0, ypos))[1] - return xname, ypos + group_name = self.groups_positions.axis_positions[group_pos] + value_pos = bbox[:, value_coord].max() + + value_pos = data_to_ax.transform((0, value_pos)[::direction]) + return group_name, value_pos[value_coord] def _sort_group_struct_pairs(self): # Draw first the annotations with the shortest between-groups distance, @@ -357,7 +381,8 @@ def _sort_group_struct_pairs(self): @staticmethod def _absolute_group_struct_pair_in_tuple_x_diff(group_struct_pair): - return abs(group_struct_pair[1][1]['x'] - group_struct_pair[1][0]['x']) + return abs(group_struct_pair[1][1]['group_coord'] + - group_struct_pair[1][0]['group_coord']) @staticmethod def check_plot_is_implemented(plot, engine="seaborn"): @@ -377,3 +402,8 @@ def fix_and_warn(dodge, hue, plot): "Implicitly setting dodge to True as it is necessary in " "statannotations. It must have been True for the seaborn " "call to yield consistent results when using `hue`.") + + def get_value_lim(self): + if self.orient == 'v': + return self.ax.get_ylim() + return self.ax.get_xlim() diff --git a/statannotations/_version.py b/statannotations/_version.py index 6a9beea..3d26edf 100644 --- a/statannotations/_version.py +++ b/statannotations/_version.py @@ -1,1 +1,1 @@ -__version__ = "0.4.0" +__version__ = "0.4.1" diff --git a/statannotations/utils.py b/statannotations/utils.py index c7d44fc..934239b 100644 --- a/statannotations/utils.py +++ b/statannotations/utils.py @@ -75,19 +75,19 @@ def _check_pairs_in_data_no_hue(pairs: Union[list, tuple], def _check_pairs_in_data_with_hue(pairs: Union[list, tuple], data: Union[List[list], pd.DataFrame] = None, - x: Union[str, list] = None, + group_coord: Union[str, list] = None, hue: str = None) -> set: - x_values = get_x_values(data, x) - seen_x_values = set() + x_values = get_x_values(data, group_coord) + seen_group_values = set() hue_values = set(data[hue].unique()) - for x_value, hue_value in itertools.chain(itertools.chain(*pairs)): - if x_value not in seen_x_values and x_value not in x_values: - raise ValueError(f"Missing x value `{x_value}` in {x}" + for group_value, hue_value in itertools.chain(itertools.chain(*pairs)): + if group_value not in seen_group_values and group_value not in x_values: + raise ValueError(f"Missing group value `{group_value}` in {group_coord}" f" (specified in `pairs`)") - seen_x_values.add(x_value) + seen_group_values.add(group_value) if hue_value not in hue_values: raise ValueError(f"Missing hue value `{hue_value}` in {hue}" @@ -108,7 +108,7 @@ def _check_hue_order_in_data(hue, hue_values: set, def check_pairs_in_data(pairs: Union[list, tuple], data: Union[List[list], pd.DataFrame] = None, - x: Union[str, list] = None, + coord: Union[str, list] = None, hue: str = None, hue_order: List[str] = None): """ @@ -116,9 +116,9 @@ def check_pairs_in_data(pairs: Union[list, tuple], """ if hue is None and hue_order is None: - _check_pairs_in_data_no_hue(pairs, data, x) + _check_pairs_in_data_no_hue(pairs, data, coord) else: - hue_values = _check_pairs_in_data_with_hue(pairs, data, x, hue) + hue_values = _check_pairs_in_data_with_hue(pairs, data, coord, hue) _check_hue_order_in_data(hue, hue_values, hue_order) diff --git a/usage/example.ipynb b/usage/example.ipynb index 10b91f9..91a3e66 100644 --- a/usage/example.ipynb +++ b/usage/example.ipynb @@ -231,7 +231,7 @@ }, { "data": { - "text/plain": "(<matplotlib.axes._subplots.AxesSubplot at 0x7fedb54c7748>,\n [<statannotations.Annotation.Annotation at 0x7fedb4233ac8>,\n <statannotations.Annotation.Annotation at 0x7fedb4233d30>,\n <statannotations.Annotation.Annotation at 0x7fedb4233320>])" + "text/plain": "(<matplotlib.axes._subplots.AxesSubplot at 0x7fc28184f6a0>,\n [<statannotations.Annotation.Annotation at 0x7fc28164aeb8>,\n <statannotations.Annotation.Annotation at 0x7fc28183d6d8>,\n <statannotations.Annotation.Annotation at 0x7fc282ada5c0>])" }, "execution_count": 6, "metadata": {}, @@ -672,7 +672,7 @@ }, { "data": { - "text/plain": "(<matplotlib.axes._subplots.AxesSubplot at 0x7fedb4460898>,\n [<statannotations.Annotation.Annotation at 0x7fedb40b4400>,\n <statannotations.Annotation.Annotation at 0x7fedb715ccf8>,\n <statannotations.Annotation.Annotation at 0x7fedb6c8ef28>])" + "text/plain": "(<matplotlib.axes._subplots.AxesSubplot at 0x7fc284d10f60>,\n [<statannotations.Annotation.Annotation at 0x7fc281c837b8>,\n <statannotations.Annotation.Annotation at 0x7fc2817bfc18>,\n <statannotations.Annotation.Annotation at 0x7fc284d8e518>])" }, "execution_count": 15, "metadata": {}, @@ -947,7 +947,7 @@ { "data": { "text/plain": "<Figure size 864x432 with 1 Axes>", - "image/png": "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\n" + "image/png": "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\n" }, "metadata": {}, "output_type": "display_data" @@ -1001,7 +1001,7 @@ { "cell_type": "markdown", "source": [ - "And a similar swarm plot, with a different format for pvalues" + "With a horizontal orientation" ], "metadata": { "collapsed": false @@ -1010,6 +1010,74 @@ { "cell_type": "code", "execution_count": 21, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "p-value annotation legend:\n", + " ns: p <= 1.00e+00\n", + " *: 1.00e-02 < p <= 5.00e-02\n", + " **: 1.00e-03 < p <= 1.00e-02\n", + " ***: 1.00e-04 < p <= 1.00e-03\n", + " ****: p <= 1.00e-04\n", + "\n", + "H1N1_Nonsynonymous vs. H1N1_Synonymous: Mann-Whitney-Wilcoxon test two-sided, P_val:5.014e-04 U_stat=2.624e+03\n", + "H3N2_Nonsynonymous vs. H3N2_Synonymous: Mann-Whitney-Wilcoxon test two-sided, P_val:1.294e-03 U_stat=1.535e+04\n", + "Influenza B_Nonsynonymous vs. Influenza B_Synonymous: Mann-Whitney-Wilcoxon test two-sided, P_val:2.026e-01 U_stat=3.340e+02\n" + ] + }, + { + "data": { + "text/plain": "(<matplotlib.axes._subplots.AxesSubplot at 0x7fc283af2668>,\n [<statannotations.Annotation.Annotation at 0x7fc283bf3780>,\n <statannotations.Annotation.Annotation at 0x7fc283be1dd8>,\n <statannotations.Annotation.Annotation at 0x7fc283bc3ef0>])" + }, + "execution_count": 21, + "metadata": {}, + "output_type": "execute_result" + }, + { + "data": { + "text/plain": "<Figure size 864x432 with 1 Axes>", + "image/png": "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\n" + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "fig, ax = plt.subplots(1,1, figsize=(12,6))\n", + "fig_args_horiz = {\n", + " **fig_args,\n", + " 'y': 'subtype',\n", + " 'x': 'mutation_rate_samp'\n", + "}\n", + "ax.set_xscale('log')\n", + "sns.stripplot(ax=ax, orient='h', **fig_args_horiz)\n", + "annotator.new_plot(ax, plot='swarmplot', orient='h', **fig_args_horiz)\n", + "annotator.apply_and_annotate()" + ], + "metadata": { + "collapsed": false, + "pycharm": { + "name": "#%%\n" + } + } + }, + { + "cell_type": "markdown", + "source": [ + "Back to vertical, with a different format for pvalues" + ], + "metadata": { + "collapsed": false, + "pycharm": { + "name": "#%% md\n" + } + } + }, + { + "cell_type": "code", + "execution_count": 22, "outputs": [ { "name": "stdout", @@ -1060,7 +1128,7 @@ }, { "cell_type": "code", - "execution_count": 22, + "execution_count": 23, "outputs": [ { "name": "stdout", @@ -1101,7 +1169,7 @@ }, { "cell_type": "code", - "execution_count": 22, + "execution_count": 23, "outputs": [], "source": [], "metadata": {

Horizontal orientation There seems to have a great demand for it (https://github.com/webermarcolivier/statannot/issues/54). Let's implement it. Maybe this PR for statannot will still be useful (https://github.com/webermarcolivier/statannot/pull/53)?

2021-07-25T13:04:36

0.0

[]

trevismd/statannotations

trevismd__statannotations-25

4728ce3b8c2475a73b7819e55dca36584234a7dc

diff --git a/.github/workflows/python-package.yml b/.github/workflows/python-package.yml index c581d62..8504874 100644 --- a/.github/workflows/python-package.yml +++ b/.github/workflows/python-package.yml @@ -5,9 +5,9 @@ name: Python package on: push: - branches: [ master, dev, v0.2, v0.3, refactoring] + branches: [ master, dev, v0.2, v0.3, v0.4] pull_request: - branches: [ master, dev, v0.2, v0.3 ] + branches: [ master, dev, v0.2, v0.3, v0.4] jobs: build: diff --git a/CHANGELOG.md b/CHANGELOG.md index 54f959b..80407ff 100644 --- a/CHANGELOG.md +++ b/CHANGELOG.md @@ -1,20 +1,30 @@ ## v0.4 ### v0.4.0 - - Major refactoring, change to an Annotator `class` to prepare and add annotations distinctly + - Major refactoring, change to an Annotator `class` to prepare and add + annotations in separate function (now method) calls. - Support for `violinplot` - - Fixes in rendering of non significant tests results after multiple comparisons correction + - Fixes in rendering of non-significant tests results after multiple + comparisons correction - Fix the printout of the star annotation legend - - Fix the bug when providing a dataframe with categories as different columns. + - Fix the failure when providing a dataframe with categories as different + columns. + - Fix a bug where an incorrect xunits calculation resulted in wrong + association of points within a box, leading to erroneous max y position for + that box. + - Reduced code complexity, more SOLID. - Many unit and integration tests ## v0.3 ### v0.3.2 - Fix `simple` format outputs - - Fix `ImportError` when applying a multiple comparison correction without statsmodels. - - Multiple comparison correction is `None` by default, as `statsmodels` is an additional dependency. + - Fix `ImportError` when applying a multiple comparison correction without + statsmodels. + - Multiple comparison correction is `None` by default, as `statsmodels` is an + additional dependency. ### v0.3.1 - - Added support of functions returning more than the two expected values when used in `StatTest` + - Added support of functions returning more than the two expected values when + used in `StatTest` - Fix version numbers in CHANGELOG ### v0.3.0 @@ -23,10 +33,17 @@ - Refactoring with implementation of `StatTest` ([#7](https://github.com/trevismd/statannotations/pull/5)), removing the `stat_func` parameter, and `test_long_name`. - - Annotations y-positions based on plot coordinates instead of data coordinates - ([#5](https://github.com/trevismd/statannotations/pull/5) by [JosephLalli](https://github.com/JosephLalli), fixes https://github.com/webermarcolivier/statannot/issues/21). + - Annotations y-positions based on plot coordinates instead of data + coordinates + ([#5](https://github.com/trevismd/statannotations/pull/5) by + [JosephLalli](https://github.com/JosephLalli), + fixes https://github.com/webermarcolivier/statannot/issues/21). - Add this CHANGELOG ## v0.2 ### v0.2.8 - - Fix bug on group/box named 0, fixes https://github.com/trevismd/statannotations/issues/10, originally in https://github.com/webermarcolivier/statannot/issues/78. Independently fixed in https://github.com/webermarcolivier/statannot/pull/73 before this issue. + - Fix bug on group/box named 0, fixes + https://github.com/trevismd/statannotations/issues/10, originally in + https://github.com/webermarcolivier/statannot/issues/78. Independently + fixed in https://github.com/webermarcolivier/statannot/pull/73 before this + issue. diff --git a/CONTRIBUTING.md b/CONTRIBUTING.md new file mode 100644 index 0000000..8d04ec8 --- /dev/null +++ b/CONTRIBUTING.md @@ -0,0 +1,21 @@ +## Contributing to statannotations +Thank you for your interest! This package is truly the work of multiple +developers and participating users and this mustn't change. + +Contributions are appreciated in all their forms, such as the opening of new +issues when discovering a (possible) bug, suggestion/upvote new features, +participation to discussions, improvements of the documentation, etc. + +If you wish specifically to submit code changes, please use the following +guide: +- [ ] If it is not an answer to an open issue, please consider opening one + first, even more if the changes are major and/or possibly controversial +- [ ] Make sure to have the dev dependencies installed: `packaging` and + `statsmodels` +- Fork the repository and make a new branch from the latest `dev` commit +- [ ] Make your code modifications, including unit testing if possible. + Please follow PEP8 guidelines. +- [ ] Run the test suite and make sure previous (and new) tests pass +- [ ] Update CHANGELOG.md to describe your changes +- [ ] Open a PR to the `dev` branch, containing at least what was added to the + changelog, as well as a reference to the issue if any diff --git a/README.md b/README.md index 545bfd1..e8f3582 100644 --- a/README.md +++ b/README.md @@ -3,22 +3,23 @@ ## What is it Python package to optionally compute statistical test and add statistical -annotations on plots generated with seaborn. -This branch is now at version 0.4.0-alpha, while the latest release is v0.3.2. +annotations on plots generated with seaborn. ## Derived work -This repository is evolving independently from +This repository is based on [webermarcolivier/statannot](https://github.com/webermarcolivier/statannot) -by Marc Weber. It is based on commit 1835078 of Feb 21, 2020, tagged "v0.2.3". + (commit 1835078 of Feb 21, 2020, tagged "v0.2.3"). Additions/modifications since that version are below represented **in bold** (previous fixes are not listed). - From version 0.4 onwards (introduction of `Annotator`), `statannot`'s API - is no longer usable in `statannotations` +**! From version 0.4.0 onwards (introduction of `Annotator`), `statannot`'s API +is no longer usable in `statannotations`**. +Please use the latest v0.3.2 release if you must keep `statannot``s API in your +code, but are looking for bug fixes we have covered. -The statannot interface, at least until its version 0.2.3, is usable in +Indeed, the statannot interface, at least until its version 0.2.3, is usable in statannotations until v.0.3.x, which already provides additional features (see corresponding branch). @@ -54,7 +55,9 @@ corresponding branch). - Optionally, custom p-values can be given as input. In this case, no statistical test is performed, but **corrections for multiple testing can be applied.** -- And various fixes (see [CHANGELOG.md](https://github.com/trevismd/statannotations/blob/master/CHANGELOG.md)). +- Any text can be used as annotation +- And various fixes (see + [CHANGELOG.md](https://github.com/trevismd/statannotations/blob/master/CHANGELOG.md)). ## Installation @@ -68,7 +71,7 @@ or, after cloning the repository, ```bash pip install . -# OR, to have optional dependencies too (multiple comparisons) +# OR, to have optional dependencies too (multiple comparisons & testing) pip install -r requirements.txt . ``` @@ -89,11 +92,12 @@ df = sns.load_dataset("tips") x = "day" y = "total_bill" order = ['Sun', 'Thur', 'Fri', 'Sat'] -box_pairs=[("Thur", "Fri"), ("Thur", "Sat"), ("Fri", "Sun")] + ax = sns.boxplot(data=df, x=x, y=y, order=order) +pairs=[("Thur", "Fri"), ("Thur", "Sat"), ("Fri", "Sun")] -annotator = Annotator(ax, box_pairs, data=df, x=x, y=y, order=order) +annotator = Annotator(ax, pairs, data=df, x=x, y=y, order=order) annotator.configure(test='Mann-Whitney', text_format='star', loc='outside') annotator.apply_and_annotate() ``` @@ -122,5 +126,5 @@ annotator.apply_and_annotate() In addition to git's history, contributions to statannotations are logged in the changelog. If you don't know where to start, there may be a few ideas in opened issues or -discussion, or something to work for the documentation. -NB: to run the test suite, the `statsmodels` and `packaging` packages are required. +discussion, or something to work for the documentation. +NB: More on [CONTRIBUTING.md](CONTRIBUTING.md) diff --git a/codecov.yml b/codecov.yml index 8b13789..0d54c1d 100644 --- a/codecov.yml +++ b/codecov.yml @@ -1,1 +1,9 @@ +coverage: + status: + project: + default: + threshold: 0.5 # Allow coverage to drop by up to 0.5% in a PR before marking it failed + patch: + default: + threshold: 0.5% \ No newline at end of file diff --git a/coverage.svg b/coverage.svg index b6c4e36..ee07d4c 100644 --- a/coverage.svg +++ b/coverage.svg @@ -15,7 +15,7 @@ <g fill="#fff" text-anchor="middle" font-family="DejaVu Sans,Verdana,Geneva,sans-serif" font-size="11"> <text x="31.5" y="15" fill="#010101" fill-opacity=".3">coverage</text> <text x="31.5" y="14">coverage</text> - <text x="80" y="15" fill="#010101" fill-opacity=".3">95%</text> - <text x="80" y="14">95%</text> + <text x="80" y="15" fill="#010101" fill-opacity=".3">96%</text> + <text x="80" y="14">96%</text> </g> </svg> diff --git a/docs/build/doctrees/environment.pickle b/docs/build/doctrees/environment.pickle index 9e93656..771aba2 100644 Binary files a/docs/build/doctrees/environment.pickle and b/docs/build/doctrees/environment.pickle differ diff --git a/docs/build/doctrees/statannotations.doctree b/docs/build/doctrees/statannotations.doctree index 9b18037..ee0691c 100644 Binary files a/docs/build/doctrees/statannotations.doctree and b/docs/build/doctrees/statannotations.doctree differ diff --git a/docs/build/doctrees/statannotations.stats.doctree b/docs/build/doctrees/statannotations.stats.doctree index 2ab9a38..a4c0dfe 100644 Binary files a/docs/build/doctrees/statannotations.stats.doctree and b/docs/build/doctrees/statannotations.stats.doctree differ diff --git a/docs/build/html/_sources/statannotations.rst.txt b/docs/build/html/_sources/statannotations.rst.txt index d3f9b4e..3a24356 100644 --- a/docs/build/html/_sources/statannotations.rst.txt +++ b/docs/build/html/_sources/statannotations.rst.txt @@ -12,6 +12,14 @@ Subpackages Submodules ---------- +statannotations.Annotation module +--------------------------------- + +.. automodule:: statannotations.Annotation + :members: + :undoc-members: + :show-inheritance: + statannotations.Annotator module -------------------------------- diff --git a/docs/build/html/_static/alabaster.css b/docs/build/html/_static/alabaster.css index 0eddaeb..cac2609 100644 --- a/docs/build/html/_static/alabaster.css +++ b/docs/build/html/_static/alabaster.css @@ -638,7 +638,7 @@ a:hover tt, a:hover code { /* Make nested-list/multi-paragraph items look better in Releases changelog * pages. Without this, docutils' magical list fuckery causes inconsistent - * formatting between different release sub-lists. + * formatter between different release sub-lists. */ div#changelog > div.section > ul > li > p:only-child { margin-bottom: 0; diff --git a/docs/build/html/genindex.html b/docs/build/html/genindex.html index 9db3161..935f030 100644 --- a/docs/build/html/genindex.html +++ b/docs/build/html/genindex.html @@ -177,15 +177,19 @@ <h1 id="index">Index</h1> <h2 id="A">A</h2> <table style="width: 100%" class="indextable genindextable"><tr> <td style="width: 33%; vertical-align: top;"><ul> - <li><a href="statannotations.html#statannotations.Annotator.Annotator.absolute_box_struct_pair_in_tuple_x_diff">absolute_box_struct_pair_in_tuple_x_diff() (statannotations.Annotator.Annotator static method)</a> -</li> <li><a href="statannotations.html#statannotations.Annotator.Annotator.alpha">alpha (statannotations.Annotator.Annotator property)</a> </li> <li><a href="statannotations.html#statannotations.Annotator.Annotator.annotate">annotate() (statannotations.Annotator.Annotator method)</a> +</li> + <li><a href="statannotations.html#statannotations.Annotator.Annotator.annotate_custom_annotations">annotate_custom_annotations() (statannotations.Annotator.Annotator method)</a> +</li> + <li><a href="statannotations.html#statannotations.Annotation.Annotation">Annotation (class in statannotations.Annotation)</a> </li> </ul></td> <td style="width: 33%; vertical-align: top;"><ul> <li><a href="statannotations.html#statannotations.Annotator.Annotator">Annotator (class in statannotations.Annotator)</a> +</li> + <li><a href="statannotations.stats.html#statannotations.stats.ComparisonsCorrection.ComparisonsCorrection.apply">apply() (statannotations.stats.ComparisonsCorrection.ComparisonsCorrection method)</a> </li> <li><a href="statannotations.html#statannotations.Annotator.Annotator.apply_and_annotate">apply_and_annotate() (statannotations.Annotator.Annotator method)</a> </li> @@ -202,10 +206,6 @@ <h2 id="C">C</h2> <table style="width: 100%" class="indextable genindextable"><tr> <td style="width: 33%; vertical-align: top;"><ul> <li><a href="statannotations.stats.html#statannotations.stats.utils.check_alpha">check_alpha() (in module statannotations.stats.utils)</a> -</li> - <li><a href="statannotations.html#statannotations.utils.check_box_pairs_in_data">check_box_pairs_in_data() (in module statannotations.utils)</a> -</li> - <li><a href="statannotations.html#statannotations.Annotator.check_implemented_plotters">check_implemented_plotters() (in module statannotations.Annotator)</a> </li> <li><a href="statannotations.html#statannotations.utils.check_is_in">check_is_in() (in module statannotations.utils)</a> </li> @@ -214,21 +214,27 @@ <h2 id="C">C</h2> <li><a href="statannotations.stats.html#statannotations.stats.utils.check_num_comparisons">check_num_comparisons() (in module statannotations.stats.utils)</a> </li> <li><a href="statannotations.html#statannotations.utils.check_order_in_data">check_order_in_data() (in module statannotations.utils)</a> +</li> + <li><a href="statannotations.html#statannotations.utils.check_pairs_in_data">check_pairs_in_data() (in module statannotations.utils)</a> </li> <li><a href="statannotations.stats.html#statannotations.stats.utils.check_pvalues">check_pvalues() (in module statannotations.stats.utils)</a> </li> - </ul></td> - <td style="width: 33%; vertical-align: top;"><ul> <li><a href="statannotations.stats.html#statannotations.stats.ComparisonsCorrection.check_valid_correction_name">check_valid_correction_name() (in module statannotations.stats.ComparisonsCorrection)</a> </li> + </ul></td> + <td style="width: 33%; vertical-align: top;"><ul> <li><a href="statannotations.html#statannotations.utils.check_valid_text_format">check_valid_text_format() (in module statannotations.utils)</a> </li> <li><a href="statannotations.html#statannotations.Annotator.Annotator.comparisons_correction">comparisons_correction (statannotations.Annotator.Annotator property)</a> </li> <li><a href="statannotations.stats.html#statannotations.stats.ComparisonsCorrection.ComparisonsCorrection">ComparisonsCorrection (class in statannotations.stats.ComparisonsCorrection)</a> </li> - <li><a href="statannotations.html#statannotations.PValueFormat.PValueFormat.config">config() (statannotations.PValueFormat.PValueFormat method)</a> + <li><a href="statannotations.html#statannotations.PValueFormat.Formatter.config">config() (statannotations.PValueFormat.Formatter method)</a> + + <ul> + <li><a href="statannotations.html#statannotations.PValueFormat.PValueFormat.config">(statannotations.PValueFormat.PValueFormat method)</a> </li> + </ul></li> <li><a href="statannotations.html#statannotations.Annotator.Annotator.configure">configure() (statannotations.Annotator.Annotator method)</a> </li> <li><a href="statannotations.stats.html#statannotations.stats.StatResult.StatResult.corrected_significance">corrected_significance (statannotations.stats.StatResult.StatResult property)</a> @@ -249,11 +255,23 @@ <h2 id="D">D</h2> <h2 id="F">F</h2> <table style="width: 100%" class="indextable genindextable"><tr> <td style="width: 33%; vertical-align: top;"><ul> - <li><a href="statannotations.html#statannotations.PValueFormat.PValueFormat.format_result">format_result() (statannotations.PValueFormat.PValueFormat method)</a> + <li><a href="statannotations.html#statannotations.Annotator.Annotator.fig">fig (statannotations.Annotator.Annotator property)</a> +</li> + <li><a href="statannotations.html#statannotations.PValueFormat.Formatter.format_data">format_data() (statannotations.PValueFormat.Formatter method)</a> + + <ul> + <li><a href="statannotations.html#statannotations.PValueFormat.PValueFormat.format_data">(statannotations.PValueFormat.PValueFormat method)</a> </li> + </ul></li> </ul></td> <td style="width: 33%; vertical-align: top;"><ul> - <li><a href="statannotations.stats.html#statannotations.stats.StatResult.StatResult.formatted_output">formatted_output (statannotations.stats.StatResult.StatResult property)</a> + <li><a href="statannotations.html#statannotations.Annotation.Annotation.formatted_output">formatted_output (statannotations.Annotation.Annotation property)</a> + + <ul> + <li><a href="statannotations.stats.html#statannotations.stats.StatResult.StatResult.formatted_output">(statannotations.stats.StatResult.StatResult property)</a> +</li> + </ul></li> + <li><a href="statannotations.html#statannotations.PValueFormat.Formatter">Formatter (class in statannotations.PValueFormat)</a> </li> <li><a href="statannotations.stats.html#statannotations.stats.StatTest.StatTest.from_library">from_library() (statannotations.stats.StatTest.StatTest static method)</a> </li> @@ -265,7 +283,7 @@ <h2 id="G">G</h2> <td style="width: 33%; vertical-align: top;"><ul> <li><a href="statannotations.html#statannotations.Annotator.Annotator.get_annotations_text">get_annotations_text() (statannotations.Annotator.Annotator method)</a> </li> - <li><a href="statannotations.html#statannotations.Annotator.Annotator.get_basic_plot">get_basic_plot() (statannotations.Annotator.Annotator static method)</a> + <li><a href="statannotations.html#statannotations.utils.get_closest">get_closest() (in module statannotations.utils)</a> </li> <li><a href="statannotations.html#statannotations.Annotator.Annotator.get_configuration">get_configuration() (statannotations.Annotator.Annotator method)</a> @@ -278,6 +296,8 @@ <h2 id="G">G</h2> <li><a href="statannotations.stats.html#statannotations.stats.ComparisonsCorrection.get_correction_parameters">get_correction_parameters() (in module statannotations.stats.ComparisonsCorrection)</a> </li> <li><a href="statannotations.stats.html#statannotations.stats.utils.get_num_comparisons">get_num_comparisons() (in module statannotations.stats.utils)</a> +</li> + <li><a href="statannotations.html#statannotations.Annotator.Annotator.get_offset_func">get_offset_func() (statannotations.Annotator.Annotator class method)</a> </li> <li><a href="statannotations.stats.html#statannotations.stats.ComparisonsCorrection.get_validated_comparisons_correction">get_validated_comparisons_correction() (in module statannotations.stats.ComparisonsCorrection)</a> </li> @@ -310,6 +330,8 @@ <h2 id="M">M</h2> <ul> <li><a href="statannotations.html#module-statannotations">statannotations</a> +</li> + <li><a href="statannotations.html#module-statannotations.Annotation">statannotations.Annotation</a> </li> <li><a href="statannotations.html#module-statannotations.Annotator">statannotations.Annotator</a> </li> @@ -346,6 +368,8 @@ <h2 id="N">N</h2> <h2 id="P">P</h2> <table style="width: 100%" class="indextable genindextable"><tr> <td style="width: 33%; vertical-align: top;"><ul> + <li><a href="statannotations.html#statannotations.Annotation.Annotation.print_labels_and_content">print_labels_and_content() (statannotations.Annotation.Annotation method)</a> +</li> <li><a href="statannotations.html#statannotations.PValueFormat.PValueFormat.print_legend_if_used">print_legend_if_used() (statannotations.PValueFormat.PValueFormat method)</a> </li> <li><a href="statannotations.html#statannotations.Annotator.Annotator.print_pvalue_legend">print_pvalue_legend() (statannotations.Annotator.Annotator method)</a> @@ -407,6 +431,13 @@ <h2 id="S">S</h2> <ul> <li><a href="statannotations.html#module-statannotations">module</a> +</li> + </ul></li> + <li> + statannotations.Annotation + + <ul> + <li><a href="statannotations.html#module-statannotations.Annotation">module</a> </li> </ul></li> <li> @@ -495,6 +526,8 @@ <h2 id="T">T</h2> </li> </ul></td> <td style="width: 33%; vertical-align: top;"><ul> + <li><a href="statannotations.html#statannotations.Annotation.Annotation.text">text (statannotations.Annotation.Annotation property)</a> +</li> <li><a href="statannotations.html#statannotations.PValueFormat.PValueFormat.text_format">text_format (statannotations.PValueFormat.PValueFormat property)</a> </li> </ul></td> @@ -504,6 +537,10 @@ <h2 id="V">V</h2> <table style="width: 100%" class="indextable genindextable"><tr> <td style="width: 33%; vertical-align: top;"><ul> <li><a href="statannotations.html#statannotations.Annotator.Annotator.validate_test_short_name">validate_test_short_name() (statannotations.Annotator.Annotator method)</a> +</li> + </ul></td> + <td style="width: 33%; vertical-align: top;"><ul> + <li><a href="statannotations.html#statannotations.Annotator.Annotator.verbose">verbose (statannotations.Annotator.Annotator property)</a> </li> </ul></td> </tr></table> diff --git a/docs/build/html/modules.html b/docs/build/html/modules.html index aff2f2a..b54367e 100644 --- a/docs/build/html/modules.html +++ b/docs/build/html/modules.html @@ -179,6 +179,7 @@ <h1>statannotations<a class="headerlink" href="#statannotations" title="Permalin </ul> </li> <li class="toctree-l2"><a class="reference internal" href="statannotations.html#submodules">Submodules</a></li> +<li class="toctree-l2"><a class="reference internal" href="statannotations.html#module-statannotations.Annotation">statannotations.Annotation module</a></li> <li class="toctree-l2"><a class="reference internal" href="statannotations.html#module-statannotations.Annotator">statannotations.Annotator module</a></li> <li class="toctree-l2"><a class="reference internal" href="statannotations.html#module-statannotations.PValueFormat">statannotations.PValueFormat module</a></li> <li class="toctree-l2"><a class="reference internal" href="statannotations.html#module-statannotations.format_annotations">statannotations.format_annotations module</a></li> diff --git a/docs/build/html/objects.inv b/docs/build/html/objects.inv index 8a3e2e4..0d82dd7 100644 Binary files a/docs/build/html/objects.inv and b/docs/build/html/objects.inv differ diff --git a/docs/build/html/py-modindex.html b/docs/build/html/py-modindex.html index 78c0a07..b4c7f07 100644 --- a/docs/build/html/py-modindex.html +++ b/docs/build/html/py-modindex.html @@ -171,6 +171,11 @@ <h1>Python Module Index</h1> <td> <a href="statannotations.html#module-statannotations"><code class="xref">statannotations</code></a></td><td> </td></tr> + <tr class="cg-1"> + <td></td> + <td>    + <a href="statannotations.html#module-statannotations.Annotation"><code class="xref">statannotations.Annotation</code></a></td><td> + </td></tr> <tr class="cg-1"> <td></td> <td>    diff --git a/docs/build/html/searchindex.js b/docs/build/html/searchindex.js index 1aa74be..38da9ed 100644 --- a/docs/build/html/searchindex.js +++ b/docs/build/html/searchindex.js @@ -1,1 +1,1 @@ -Search.setIndex({docnames:["index","modules","setup","statannotations","statannotations.stats"],envversion:{"sphinx.domains.c":2,"sphinx.domains.changeset":1,"sphinx.domains.citation":1,"sphinx.domains.cpp":3,"sphinx.domains.index":1,"sphinx.domains.javascript":2,"sphinx.domains.math":2,"sphinx.domains.python":3,"sphinx.domains.rst":2,"sphinx.domains.std":2,sphinx:56},filenames:["index.rst","modules.rst","setup.rst","statannotations.rst","statannotations.stats.rst"],objects:{"":{statannotations:[3,0,0,"-"]},"statannotations.Annotator":{Annotator:[3,1,1,""],check_implemented_plotters:[3,4,1,""]},"statannotations.Annotator.Annotator":{absolute_box_struct_pair_in_tuple_x_diff:[3,2,1,""],alpha:[3,3,1,""],annotate:[3,2,1,""],apply_and_annotate:[3,2,1,""],apply_test:[3,2,1,""],comparisons_correction:[3,3,1,""],configure:[3,2,1,""],get_annotations_text:[3,2,1,""],get_basic_plot:[3,2,1,""],get_configuration:[3,2,1,""],has_type0_comparisons_correction:[3,2,1,""],loc:[3,3,1,""],new_plot:[3,2,1,""],print_pvalue_legend:[3,2,1,""],pvalue_format:[3,3,1,""],reset_configuration:[3,2,1,""],set_custom_annotations:[3,2,1,""],set_pvalues:[3,2,1,""],set_pvalues_and_annotate:[3,2,1,""],test:[3,3,1,""],validate_test_short_name:[3,2,1,""]},"statannotations.PValueFormat":{PValueFormat:[3,1,1,""],sort_pvalue_thresholds:[3,4,1,""]},"statannotations.PValueFormat.PValueFormat":{config:[3,2,1,""],format_result:[3,2,1,""],get_configuration:[3,2,1,""],print_legend_if_used:[3,2,1,""],pvalue_format_string:[3,3,1,""],pvalue_thresholds:[3,3,1,""],simple_format_string:[3,3,1,""],text_format:[3,3,1,""]},"statannotations.format_annotations":{pval_annotation_text:[3,4,1,""],simple_text:[3,4,1,""]},"statannotations.stats":{ComparisonsCorrection:[4,0,0,"-"],StatResult:[4,0,0,"-"],StatTest:[4,0,0,"-"],test:[4,0,0,"-"],utils:[4,0,0,"-"]},"statannotations.stats.ComparisonsCorrection":{ComparisonsCorrection:[4,1,1,""],check_valid_correction_name:[4,4,1,""],get_correction_parameters:[4,4,1,""],get_validated_comparisons_correction:[4,4,1,""]},"statannotations.stats.ComparisonsCorrection.ComparisonsCorrection":{document:[4,2,1,""]},"statannotations.stats.StatResult":{StatResult:[4,1,1,""]},"statannotations.stats.StatResult.StatResult":{corrected_significance:[4,3,1,""],correction_method:[4,3,1,""],formatted_output:[4,3,1,""],significance_suffix:[4,3,1,""]},"statannotations.stats.StatTest":{StatTest:[4,1,1,""],wilcoxon:[4,4,1,""]},"statannotations.stats.StatTest.StatTest":{from_library:[4,2,1,""],short_name:[4,3,1,""]},"statannotations.stats.test":{apply_test:[4,4,1,""]},"statannotations.stats.utils":{check_alpha:[4,4,1,""],check_num_comparisons:[4,4,1,""],check_pvalues:[4,4,1,""],get_num_comparisons:[4,4,1,""],return_results:[4,4,1,""]},"statannotations.utils":{check_box_pairs_in_data:[3,4,1,""],check_is_in:[3,4,1,""],check_not_none:[3,4,1,""],check_order_in_data:[3,4,1,""],check_valid_text_format:[3,4,1,""],get_x_values:[3,4,1,""],raise_expected_got:[3,4,1,""],remove_null:[3,4,1,""],render_collection:[3,4,1,""]},statannotations:{Annotator:[3,0,0,"-"],PValueFormat:[3,0,0,"-"],format_annotations:[3,0,0,"-"],stats:[4,0,0,"-"],utils:[3,0,0,"-"]}},objnames:{"0":["py","module","Python module"],"1":["py","class","Python class"],"2":["py","method","Python method"],"3":["py","property","Python property"],"4":["py","function","Python function"]},objtypes:{"0":"py:module","1":"py:class","2":"py:method","3":"py:property","4":"py:function"},terms:{"0":[3,4],"05":[3,4],"1":[3,4],"case":4,"class":[3,4],"default":[3,4],"float":4,"function":[3,4],"int":4,"return":3,"short":3,"static":[3,4],"true":[3,4],A:3,Or:3,The:3,a_set:[],absolute_box_struct_pair_in_tuple_x_diff:3,accept:3,activate_configured_warn:[],add:3,addit:4,alpha:[3,4],an:3,annot:1,appear:3,appli:3,apply_and_annot:3,apply_test:[3,4],ar:3,arg:4,argument:[3,4],arrai:3,auto:3,ax:3,b:3,bar:3,barplot:3,base:[3,4],befor:3,behavior:4,benjamini:4,between:3,bonferroni:4,bool:4,box:3,box_data1:4,box_data2:4,box_pair:3,box_struct_pair:3,boxplot:3,calcul:3,callabl:4,can:3,check:3,check_alpha:4,check_box_pairs_in_data:3,check_correct_number_custom_annot:[],check_implemented_plott:3,check_is_in:3,check_not_non:3,check_num_comparison:4,check_order_box_pairs_in_data:[],check_order_in_data:3,check_pval_correction_input_valu:[],check_pvalu:4,check_valid_correction_nam:4,check_valid_text_format:3,collect:3,color:3,comparison:[3,4],comparisons_correct:[3,4],comparisonscorrect:[1,3],comput:3,config:3,configur:3,content:1,coordin:3,core:3,corr_kwarg:4,correct:4,corrected_signific:4,correction_method:4,custom:3,data:[3,4],datafram:3,deactivate_configured_warn:[],dict:4,displai:3,document:4,each:3,earlier:4,either:4,equival:4,error:3,error_typ:3,exact:3,exist:3,expect:3,fals:3,follow:3,for_:3,format:[3,4],format_annot:1,format_result:3,formatted_output:4,fraction:3,frame:3,from:4,from_librari:4,func:4,gener:3,get:4,get_annotations_text:3,get_basic_plot:3,get_configur:3,get_correction_paramet:4,get_num_comparison:4,get_validated_comparisons_correct:4,get_x_valu:3,given:3,got:3,gt:4,has_type0_comparisons_correct:3,hochberg:4,holm:4,hue:3,hue_ord:3,index:0,instanc:[3,4],instead:3,interfac:4,interpret:4,keyword:4,kruskal:4,kwarg:4,label:3,leven:4,line_height:3,line_offset:3,line_offset_to_box:3,line_width:3,list:3,loc:3,ls:4,mann:4,messag:3,method:4,method_typ:4,mode:3,modul:[0,1],multipl:4,must:3,name:[3,4],new_plot:3,none:[3,4],num_comparison:[3,4],number:[3,4],object:[3,4],one:[3,4],option:[3,4],order:3,otherwis:3,overrid:3,p:3,p_valu:4,p_values_arrai:[],packag:[0,1],page:0,pair:3,panda:3,param:[],paramet:[3,4],pass:[3,4],per:3,perform:3,perform_stat_test:3,plot:3,plot_param:3,point:3,print_legend_if_us:3,print_pvalue_legend:3,properti:[3,4],provid:[3,4],pval:4,pval_annotation_text:3,pvalu:[3,4],pvalue_format:3,pvalue_format_str:3,pvalue_threshold:3,pvalueformat:1,rais:3,raise_expected_got:3,rang:3,refer:3,remove_nul:3,render:3,render_collect:3,render_set:[],reset_configur:3,result:[3,4],results_arrai:4,return_result:4,run:4,same:3,sampl:4,seaborn:3,search:0,seri:3,set:3,set_custom_annot:3,set_pvalu:3,set_pvalues_and_annot:3,short_nam:4,should_warn_about_configur:[],show:3,show_test_nam:3,significance_suffix:4,simpl:3,simple_format_str:3,simple_text:3,sort_pvalue_threshold:3,standard:3,stat:[1,3],stat_nam:4,stat_str:4,stat_test:[],statist:[3,4],statresult:[1,3],stats_param:[3,4],statsmodel:4,statsmodels_api:4,stattest:[1,3],str:[3,4],string:3,submodul:1,subpackag:1,t:4,test:[1,3],test_descript:4,test_ind:4,test_long_nam:4,test_nam:4,test_pair:4,test_short_nam:[3,4],test_welch:4,text:3,text_annot_custom:3,text_format:3,text_offset:3,thereof:3,thi:[3,4],threshold:3,top:3,tupl:3,two:[3,4],type:3,union:[3,4],unit:3,update_y_for_loc:[],us:[3,4],use_fixed_offset:3,util:1,valid_valu:3,validate_test_short_nam:3,valu:[3,4],valueerror:3,verbos:[3,4],version:4,wa:3,what:3,whitnei:4,width:3,wilcoxon:4,work:3,x:3,y:3,yekuti:4},titles:["Welcome to statannotations\u2019s documentation!","statannotations","setup module","statannotations package","statannotations.stats package"],titleterms:{annot:3,comparisonscorrect:4,content:[0,3,4],document:0,format_annot:3,indic:0,modul:[2,3,4],packag:[3,4],pvalueformat:3,s:0,setup:2,stat:4,statannot:[0,1,3,4],statresult:4,stattest:4,submodul:[3,4],subpackag:3,tabl:0,test:4,util:[3,4],welcom:0}}) \ No newline at end of file +Search.setIndex({docnames:["index","modules","setup","statannotations","statannotations.stats"],envversion:{"sphinx.domains.c":2,"sphinx.domains.changeset":1,"sphinx.domains.citation":1,"sphinx.domains.cpp":3,"sphinx.domains.index":1,"sphinx.domains.javascript":2,"sphinx.domains.math":2,"sphinx.domains.python":3,"sphinx.domains.rst":2,"sphinx.domains.std":2,sphinx:56},filenames:["index.rst","modules.rst","setup.rst","statannotations.rst","statannotations.stats.rst"],objects:{"":{statannotations:[3,0,0,"-"]},"statannotations.Annotation":{Annotation:[3,1,1,""]},"statannotations.Annotation.Annotation":{formatted_output:[3,2,1,""],print_labels_and_content:[3,3,1,""],text:[3,2,1,""]},"statannotations.Annotator":{Annotator:[3,1,1,""]},"statannotations.Annotator.Annotator":{alpha:[3,2,1,""],annotate:[3,3,1,""],annotate_custom_annotations:[3,3,1,""],apply_and_annotate:[3,3,1,""],apply_test:[3,3,1,""],comparisons_correction:[3,2,1,""],configure:[3,3,1,""],fig:[3,2,1,""],get_annotations_text:[3,3,1,""],get_configuration:[3,3,1,""],get_offset_func:[3,3,1,""],has_type0_comparisons_correction:[3,3,1,""],loc:[3,2,1,""],new_plot:[3,3,1,""],print_pvalue_legend:[3,3,1,""],pvalue_format:[3,2,1,""],reset_configuration:[3,3,1,""],set_custom_annotations:[3,3,1,""],set_pvalues:[3,3,1,""],set_pvalues_and_annotate:[3,3,1,""],test:[3,2,1,""],validate_test_short_name:[3,3,1,""],verbose:[3,2,1,""]},"statannotations.PValueFormat":{Formatter:[3,1,1,""],PValueFormat:[3,1,1,""],sort_pvalue_thresholds:[3,4,1,""]},"statannotations.PValueFormat.Formatter":{config:[3,3,1,""],format_data:[3,3,1,""]},"statannotations.PValueFormat.PValueFormat":{config:[3,3,1,""],format_data:[3,3,1,""],get_configuration:[3,3,1,""],print_legend_if_used:[3,3,1,""],pvalue_format_string:[3,2,1,""],pvalue_thresholds:[3,2,1,""],simple_format_string:[3,2,1,""],text_format:[3,2,1,""]},"statannotations.format_annotations":{pval_annotation_text:[3,4,1,""],simple_text:[3,4,1,""]},"statannotations.stats":{ComparisonsCorrection:[4,0,0,"-"],StatResult:[4,0,0,"-"],StatTest:[4,0,0,"-"],test:[4,0,0,"-"],utils:[4,0,0,"-"]},"statannotations.stats.ComparisonsCorrection":{ComparisonsCorrection:[4,1,1,""],check_valid_correction_name:[4,4,1,""],get_correction_parameters:[4,4,1,""],get_validated_comparisons_correction:[4,4,1,""]},"statannotations.stats.ComparisonsCorrection.ComparisonsCorrection":{apply:[4,3,1,""],document:[4,3,1,""]},"statannotations.stats.StatResult":{StatResult:[4,1,1,""]},"statannotations.stats.StatResult.StatResult":{corrected_significance:[4,2,1,""],correction_method:[4,2,1,""],formatted_output:[4,2,1,""],significance_suffix:[4,2,1,""]},"statannotations.stats.StatTest":{StatTest:[4,1,1,""],wilcoxon:[4,4,1,""]},"statannotations.stats.StatTest.StatTest":{from_library:[4,3,1,""],short_name:[4,2,1,""]},"statannotations.stats.test":{apply_test:[4,4,1,""]},"statannotations.stats.utils":{check_alpha:[4,4,1,""],check_num_comparisons:[4,4,1,""],check_pvalues:[4,4,1,""],get_num_comparisons:[4,4,1,""],return_results:[4,4,1,""]},"statannotations.utils":{check_is_in:[3,4,1,""],check_not_none:[3,4,1,""],check_order_in_data:[3,4,1,""],check_pairs_in_data:[3,4,1,""],check_valid_text_format:[3,4,1,""],get_closest:[3,4,1,""],get_x_values:[3,4,1,""],raise_expected_got:[3,4,1,""],remove_null:[3,4,1,""],render_collection:[3,4,1,""]},statannotations:{Annotation:[3,0,0,"-"],Annotator:[3,0,0,"-"],PValueFormat:[3,0,0,"-"],format_annotations:[3,0,0,"-"],stats:[4,0,0,"-"],utils:[3,0,0,"-"]}},objnames:{"0":["py","module","Python module"],"1":["py","class","Python class"],"2":["py","property","Python property"],"3":["py","method","Python method"],"4":["py","function","Python function"]},objtypes:{"0":"py:module","1":"py:class","2":"py:property","3":"py:method","4":"py:function"},terms:{"0":[3,4],"05":[3,4],"1":[3,4],"12141511":3,"9981846":3,"case":4,"class":[3,4],"default":[3,4],"float":4,"function":[3,4],"int":4,"return":3,"short":3,"static":4,"true":[3,4],A:3,If:3,Or:3,The:3,a_list:3,accept:3,add:3,addit:4,alpha:[3,4],an:3,annot:1,annotate_custom_annot:3,appli:[3,4],apply_and_annot:3,apply_test:[3,4],ar:3,arg:[3,4],argument:[3,4],arrai:3,assum:3,auto:3,ax:3,bar:3,base:[3,4],behavior:4,benjamini:4,between:3,bonferroni:4,bool:4,box:3,boxplot:3,calcul:3,callabl:4,check:3,check_alpha:4,check_is_in:3,check_not_non:3,check_num_comparison:4,check_order_in_data:3,check_pairs_in_data:3,check_pvalu:4,check_valid_correction_nam:4,check_valid_text_format:3,classmethod:3,close:3,closest:3,collect:3,color:3,com:3,comparison:[3,4],comparisons_correct:[3,4],comparisonscorrect:[1,3],comput:3,config:3,configur:3,constructor:3,content:1,coordin:3,core:3,corr_kwarg:4,correct:4,corrected_signific:4,correction_method:4,custom:3,data:[3,4],datafram:3,dict:4,displai:3,document:4,each:3,earlier:4,either:4,engin:3,equal:3,equival:4,error:3,error_typ:3,exact:3,exist:3,expect:3,fals:3,fig:3,follow:3,for_:3,format:[3,4],format_annot:1,format_data:3,formatt:3,formatted_output:[3,4],fraction:3,frame:3,from:[3,4],from_librari:4,func:4,gener:3,get:4,get_annotations_text:3,get_closest:3,get_configur:3,get_correction_paramet:4,get_num_comparison:4,get_offset_func:3,get_validated_comparisons_correct:4,get_x_valu:3,got:3,group:3,group_data1:4,group_data2:4,gt:4,has_type0_comparisons_correct:3,hochberg:4,hold:3,holm:4,http:3,hue:3,hue_ord:3,index:0,instanc:[3,4],instead:3,interfac:4,interpret:4,keyword:4,kruskal:4,kwarg:[3,4],label:3,leven:4,line_height:3,line_offset:3,line_offset_to_group:3,line_width:3,link:3,list:3,loc:3,ls:4,mann:4,messag:3,method:4,method_typ:4,mode:3,modul:[0,1],multipl:4,must:3,mylist:3,mynumb:3,name:[3,4],new_plot:3,none:[3,4],num_comparison:[3,4],number:[3,4],object:[3,4],one:[3,4],option:[3,4],order:3,otherwis:3,overrid:3,p:3,p_valu:4,packag:[0,1],page:0,pair:3,panda:3,paramet:[3,4],pass:[3,4],per:3,perform:3,plot:3,plot_param:3,point:3,posit:3,print_labels_and_cont:3,print_legend_if_us:3,print_pvalue_legend:3,properti:[3,4],provid:[3,4],pval:4,pval_annotation_text:3,pvalu:[3,4],pvalue_format:3,pvalue_format_str:3,pvalue_threshold:3,pvalueformat:1,rais:3,raise_expected_got:3,rang:3,refer:3,remove_nul:3,render:3,render_collect:3,reset_configur:3,result:[3,4],results_arrai:4,return_result:4,run:4,same:3,sampl:4,seaborn:3,search:0,sep:3,seri:3,set:3,set_custom_annot:3,set_pvalu:3,set_pvalues_and_annot:3,short_nam:4,show:3,show_test_nam:3,significance_suffix:4,simpl:3,simple_format_str:3,simple_text:3,smallest:3,sort:3,sort_pvalue_threshold:3,stackoverflow:3,standard:3,stat:[1,3],stat_nam:4,stat_str:4,statist:[3,4],statresult:[1,3],stats_param:[3,4],statsmodel:4,statsmodels_api:4,stattest:[1,3],str:[3,4],string:3,struct:3,submodul:1,subpackag:1,t:4,test:[1,3],test_descript:4,test_ind:4,test_long_nam:4,test_nam:4,test_pair:4,test_result_list:4,test_short_nam:[3,4],test_welch:4,text:3,text_annot_custom:3,text_format:3,text_offset:3,them:3,thereof:3,thi:[3,4],three:3,threshold:3,top:3,tupl:3,two:[3,4],type:3,union:[3,4],us:[3,4],use_fixed_offset:3,util:1,valid_valu:3,validate_test_short_nam:3,valu:[3,4],valueerror:3,verbos:[3,4],version:4,vs:3,wa:3,whitnei:4,wilcoxon:4,work:3,x:3,y:3,yekuti:4},titles:["Welcome to statannotations\u2019s documentation!","statannotations","setup module","statannotations package","statannotations.stats package"],titleterms:{annot:3,comparisonscorrect:4,content:[0,3,4],document:0,format_annot:3,indic:0,modul:[2,3,4],packag:[3,4],pvalueformat:3,s:0,setup:2,stat:4,statannot:[0,1,3,4],statresult:4,stattest:4,submodul:[3,4],subpackag:3,tabl:0,test:4,util:[3,4],welcom:0}}) \ No newline at end of file diff --git a/docs/build/html/statannotations.html b/docs/build/html/statannotations.html index 13780ed..aaacef6 100644 --- a/docs/build/html/statannotations.html +++ b/docs/build/html/statannotations.html @@ -96,6 +96,7 @@ </ul> </li> <li class="toctree-l3"><a class="reference internal" href="#submodules">Submodules</a></li> +<li class="toctree-l3"><a class="reference internal" href="#module-statannotations.Annotation">statannotations.Annotation module</a></li> <li class="toctree-l3"><a class="reference internal" href="#module-statannotations.Annotator">statannotations.Annotator module</a></li> <li class="toctree-l3"><a class="reference internal" href="#module-statannotations.PValueFormat">statannotations.PValueFormat module</a></li> <li class="toctree-l3"><a class="reference internal" href="#module-statannotations.format_annotations">statannotations.format_annotations module</a></li> @@ -195,29 +196,54 @@ <h2>Subpackages<a class="headerlink" href="#subpackages" title="Permalink to thi </div> <div class="section" id="submodules"> <h2>Submodules<a class="headerlink" href="#submodules" title="Permalink to this headline">¶</a></h2> +</div> +<div class="section" id="module-statannotations.Annotation"> +<h2>statannotations.Annotation module<a class="headerlink" href="#module-statannotations.Annotation" title="Permalink to this headline">¶</a></h2> +<dl class="py class"> +<dt class="sig sig-object py" id="statannotations.Annotation.Annotation"> +class statannotations.Annotation.Annotation(structs, data: Union[str, <a class="reference internal" href="statannotations.stats.html#statannotations.stats.StatResult.StatResult" title="statannotations.stats.StatResult.StatResult">statannotations.stats.StatResult.StatResult</a>], formatter: Optional[<a class="reference internal" href="#statannotations.PValueFormat.Formatter" title="statannotations.PValueFormat.Formatter">statannotations.PValueFormat.Formatter</a>] = None)<a class="headerlink" href="#statannotations.Annotation.Annotation" title="Permalink to this definition">¶</a></dt> +<dd>Bases: <code class="xref py py-class docutils literal notranslate">object</code> +Holds data, linked structs and an optional Formatter. +<dl class="py property"> +<dt class="sig sig-object py" id="statannotations.Annotation.Annotation.formatted_output"> +property formatted_output<a class="headerlink" href="#statannotations.Annotation.Annotation.formatted_output" title="Permalink to this definition">¶</a></dt> +<dd></dd></dl> + +<dl class="py method"> +<dt class="sig sig-object py" id="statannotations.Annotation.Annotation.print_labels_and_content"> +print_labels_and_content(sep=' vs. ')<a class="headerlink" href="#statannotations.Annotation.Annotation.print_labels_and_content" title="Permalink to this definition">¶</a></dt> +<dd></dd></dl> + +<dl class="py property"> +<dt class="sig sig-object py" id="statannotations.Annotation.Annotation.text"> +property text<a class="headerlink" href="#statannotations.Annotation.Annotation.text" title="Permalink to this definition">¶</a></dt> +<dd></dd></dl> + +</dd></dl> + </div> <div class="section" id="module-statannotations.Annotator"> <h2>statannotations.Annotator module<a class="headerlink" href="#module-statannotations.Annotator" title="Permalink to this headline">¶</a></h2> <dl class="py class"> <dt class="sig sig-object py" id="statannotations.Annotator.Annotator"> -class statannotations.Annotator.Annotator(ax, box_pairs, plot='boxplot', data=None, x=None, y=None, hue=None, order=None, hue_order=None, **plot_params)<a class="headerlink" href="#statannotations.Annotator.Annotator" title="Permalink to this definition">¶</a></dt> +class statannotations.Annotator.Annotator(ax, pairs, plot='boxplot', data=None, x=None, y=None, hue=None, order=None, hue_order=None, engine='seaborn', verbose=True, **plot_params)<a class="headerlink" href="#statannotations.Annotator.Annotator" title="Permalink to this definition">¶</a></dt> <dd>Bases: <code class="xref py py-class docutils literal notranslate">object</code> Optionally computes statistical test between pairs of data series, and add -statistical annotation on top of the boxes/bars. The same exact arguments -<cite>data</cite>, <cite>x</cite>, <cite>y</cite>, <cite>hue</cite>, <cite>order</cite>, <cite>width</cite>, <cite>hue_order</cite> (and <cite>units</cite>) as in -the seaborn boxplot/barplot function must be passed to this function. -This function works in one of the two following modes: -a) <cite>perform_stat_test</cite> is True (default): -* statistical test as given by argument <cite>test</cite> is performed. -* The <cite>test_short_name</cite> argument can be used to customize what appears -before the pvalues if test is a string. -b) <cite>perform_stat_test</cite> is False: no statistical test is performed, list of -custom p-values <cite>pvalues</cite> are used for each pair of boxes. -<dl class="py method"> -<dt class="sig sig-object py" id="statannotations.Annotator.Annotator.absolute_box_struct_pair_in_tuple_x_diff"> -static absolute_box_struct_pair_in_tuple_x_diff(box_struct_pair)<a class="headerlink" href="#statannotations.Annotator.Annotator.absolute_box_struct_pair_in_tuple_x_diff" title="Permalink to this definition">¶</a></dt> -<dd></dd></dl> - +statistical annotation on top of the groups (boxes, bars…). The same +exact arguments provided to the seaborn plotting function must be passed to +the constructor. +<dl class="simple"> +<dt>This Annotator works in one of the three following modes:</dt><dd><ul class="simple"> +<li>Add custom text annotations (<cite>set_custom_annotations</cite>)</li> +<li>Format pvalues and add them to the plot (<cite>set_pvalues</cite>)</li> +<li><dl class="simple"> +<dt>Perform a statistical test and then add the results to the plot</dt><dd>(<cite>apply_test</cite>) +</dd> +</dl> +</li> +</ul> +</dd> +</dl> <dl class="py property"> <dt class="sig sig-object py" id="statannotations.Annotator.Annotator.alpha"> property alpha<a class="headerlink" href="#statannotations.Annotator.Annotator.alpha" title="Permalink to this definition">¶</a></dt> @@ -225,10 +251,21 @@ <h2>Submodules<a class="headerlink" href="#submodules" title="Permalink to this <dl class="py method"> <dt class="sig sig-object py" id="statannotations.Annotator.Annotator.annotate"> -annotate(line_offset=None, line_offset_to_box=None)<a class="headerlink" href="#statannotations.Annotator.Annotator.annotate" title="Permalink to this definition">¶</a></dt> +annotate(line_offset=None, line_offset_to_group=None)<a class="headerlink" href="#statannotations.Annotator.Annotator.annotate" title="Permalink to this definition">¶</a></dt> <dd>Add configured annotations to the plot. </dd></dl> +<dl class="py method"> +<dt class="sig sig-object py" id="statannotations.Annotator.Annotator.annotate_custom_annotations"> +annotate_custom_annotations(text_annot_custom)<a class="headerlink" href="#statannotations.Annotator.Annotator.annotate_custom_annotations" title="Permalink to this definition">¶</a></dt> +<dd><dl class="field-list simple"> +<dt class="field-odd">Parameters</dt> +<dd class="field-odd">text_annot_custom – List of strings to annotate for each +<cite>pair</cite> +</dd> +</dl> +</dd></dl> + <dl class="py method"> <dt class="sig sig-object py" id="statannotations.Annotator.Annotator.apply_and_annotate"> apply_and_annotate()<a class="headerlink" href="#statannotations.Annotator.Annotator.apply_and_annotate" title="Permalink to this definition">¶</a></dt> @@ -243,7 +280,7 @@ <h2>Submodules<a class="headerlink" href="#submodules" title="Permalink to this <dd class="field-odd"><ul class="simple"> <li>stats_params – Parameters for statistical test functions.</li> <li>num_comparisons – Override number of comparisons otherwise -calculated with number of box_pairs</li> +calculated with number of pairs</li> </ul> </dd> </dl> @@ -263,7 +300,7 @@ <h2>Submodules<a class="headerlink" href="#submodules" title="Permalink to this <li><cite>comparisons_correction</cite></li> <li><cite>line_height</cite>: in axes fraction coordinates</li> <li><cite>line_offset</cite></li> -<li><cite>line_offset_to_box</cite></li> +<li><cite>line_offset_to_group</cite></li> <li><cite>line_width</cite></li> <li><cite>loc</cite></li> <li><cite>pvalue_format</cite></li> @@ -280,19 +317,24 @@ <h2>Submodules<a class="headerlink" href="#submodules" title="Permalink to this </ul> </dd></dl> +<dl class="py property"> +<dt class="sig sig-object py" id="statannotations.Annotator.Annotator.fig"> +property fig<a class="headerlink" href="#statannotations.Annotator.Annotator.fig" title="Permalink to this definition">¶</a></dt> +<dd></dd></dl> + <dl class="py method"> <dt class="sig sig-object py" id="statannotations.Annotator.Annotator.get_annotations_text"> get_annotations_text()<a class="headerlink" href="#statannotations.Annotator.Annotator.get_annotations_text" title="Permalink to this definition">¶</a></dt> <dd></dd></dl> <dl class="py method"> -<dt class="sig sig-object py" id="statannotations.Annotator.Annotator.get_basic_plot"> -static get_basic_plot(ax, box_pairs, plot='boxplot', data=None, x=None, y=None, hue=None, order=None, hue_order=None, **plot_params)<a class="headerlink" href="#statannotations.Annotator.Annotator.get_basic_plot" title="Permalink to this definition">¶</a></dt> +<dt class="sig sig-object py" id="statannotations.Annotator.Annotator.get_configuration"> +get_configuration()<a class="headerlink" href="#statannotations.Annotator.Annotator.get_configuration" title="Permalink to this definition">¶</a></dt> <dd></dd></dl> <dl class="py method"> -<dt class="sig sig-object py" id="statannotations.Annotator.Annotator.get_configuration"> -get_configuration()<a class="headerlink" href="#statannotations.Annotator.Annotator.get_configuration" title="Permalink to this definition">¶</a></dt> +<dt class="sig sig-object py" id="statannotations.Annotator.Annotator.get_offset_func"> +classmethod get_offset_func(position)<a class="headerlink" href="#statannotations.Annotator.Annotator.get_offset_func" title="Permalink to this definition">¶</a></dt> <dd></dd></dl> <dl class="py method"> @@ -307,7 +349,7 @@ <h2>Submodules<a class="headerlink" href="#submodules" title="Permalink to this <dl class="py method"> <dt class="sig sig-object py" id="statannotations.Annotator.Annotator.new_plot"> -new_plot(ax, box_pairs=None, plot='boxplot', data=None, x=None, y=None, hue=None, order=None, hue_order=None, **plot_params)<a class="headerlink" href="#statannotations.Annotator.Annotator.new_plot" title="Permalink to this definition">¶</a></dt> +new_plot(ax, pairs=None, plot='boxplot', data=None, x=None, y=None, hue=None, order=None, hue_order=None, engine: str = 'seaborn', **plot_params)<a class="headerlink" href="#statannotations.Annotator.Annotator.new_plot" title="Permalink to this definition">¶</a></dt> <dd></dd></dl> <dl class="py method"> @@ -331,7 +373,7 @@ <h2>Submodules<a class="headerlink" href="#submodules" title="Permalink to this <dd><dl class="field-list simple"> <dt class="field-odd">Parameters</dt> <dd class="field-odd">text_annot_custom – List of strings to annotate for each -<cite>box_pair</cite> +<cite>pair</cite> </dd> </dl> </dd></dl> @@ -342,9 +384,9 @@ <h2>Submodules<a class="headerlink" href="#submodules" title="Permalink to this <dd><dl class="field-list simple"> <dt class="field-odd">Parameters</dt> <dd class="field-odd"><ul class="simple"> -<li>pvalues – list or array of p-values for each box pair comparison.</li> +<li>pvalues – list or array of p-values for each pair comparison.</li> <li>num_comparisons – Override number of comparisons otherwise -calculated with number of box_pairs</li> +calculated with number of pairs</li> </ul> </dd> </dl> @@ -365,28 +407,44 @@ <h2>Submodules<a class="headerlink" href="#submodules" title="Permalink to this validate_test_short_name()<a class="headerlink" href="#statannotations.Annotator.Annotator.validate_test_short_name" title="Permalink to this definition">¶</a></dt> <dd></dd></dl> -</dd></dl> - -<dl class="py function"> -<dt class="sig sig-object py" id="statannotations.Annotator.check_implemented_plotters"> -statannotations.Annotator.check_implemented_plotters(plot)<a class="headerlink" href="#statannotations.Annotator.check_implemented_plotters" title="Permalink to this definition">¶</a></dt> +<dl class="py property"> +<dt class="sig sig-object py" id="statannotations.Annotator.Annotator.verbose"> +property verbose<a class="headerlink" href="#statannotations.Annotator.Annotator.verbose" title="Permalink to this definition">¶</a></dt> <dd></dd></dl> +</dd></dl> + </div> <div class="section" id="module-statannotations.PValueFormat"> <h2>statannotations.PValueFormat module<a class="headerlink" href="#module-statannotations.PValueFormat" title="Permalink to this headline">¶</a></h2> +<dl class="py class"> +<dt class="sig sig-object py" id="statannotations.PValueFormat.Formatter"> +class statannotations.PValueFormat.Formatter<a class="headerlink" href="#statannotations.PValueFormat.Formatter" title="Permalink to this definition">¶</a></dt> +<dd>Bases: <code class="xref py py-class docutils literal notranslate">object</code> +<dl class="py method"> +<dt class="sig sig-object py" id="statannotations.PValueFormat.Formatter.config"> +config(*args, **kwargs)<a class="headerlink" href="#statannotations.PValueFormat.Formatter.config" title="Permalink to this definition">¶</a></dt> +<dd></dd></dl> + +<dl class="py method"> +<dt class="sig sig-object py" id="statannotations.PValueFormat.Formatter.format_data"> +format_data(data)<a class="headerlink" href="#statannotations.PValueFormat.Formatter.format_data" title="Permalink to this definition">¶</a></dt> +<dd></dd></dl> + +</dd></dl> + <dl class="py class"> <dt class="sig sig-object py" id="statannotations.PValueFormat.PValueFormat"> class statannotations.PValueFormat.PValueFormat<a class="headerlink" href="#statannotations.PValueFormat.PValueFormat" title="Permalink to this definition">¶</a></dt> -<dd>Bases: <code class="xref py py-class docutils literal notranslate">object</code> +<dd>Bases: <a class="reference internal" href="#statannotations.PValueFormat.Formatter" title="statannotations.PValueFormat.Formatter"><code class="xref py py-class docutils literal notranslate">statannotations.PValueFormat.Formatter</code></a> <dl class="py method"> <dt class="sig sig-object py" id="statannotations.PValueFormat.PValueFormat.config"> config(**parameters)<a class="headerlink" href="#statannotations.PValueFormat.PValueFormat.config" title="Permalink to this definition">¶</a></dt> <dd></dd></dl> <dl class="py method"> -<dt class="sig sig-object py" id="statannotations.PValueFormat.PValueFormat.format_result"> -format_result(result)<a class="headerlink" href="#statannotations.PValueFormat.PValueFormat.format_result" title="Permalink to this definition">¶</a></dt> +<dt class="sig sig-object py" id="statannotations.PValueFormat.PValueFormat.format_data"> +format_data(result)<a class="headerlink" href="#statannotations.PValueFormat.PValueFormat.format_data" title="Permalink to this definition">¶</a></dt> <dd></dd></dl> <dl class="py method"> @@ -431,16 +489,16 @@ <h2>Submodules<a class="headerlink" href="#submodules" title="Permalink to this <h2>statannotations.format_annotations module<a class="headerlink" href="#module-statannotations.format_annotations" title="Permalink to this headline">¶</a></h2> <dl class="py function"> <dt class="sig sig-object py" id="statannotations.format_annotations.pval_annotation_text"> -statannotations.format_annotations.pval_annotation_text(result: Union[List[<a class="reference internal" href="statannotations.stats.html#statannotations.stats.StatResult.StatResult" title="statannotations.stats.StatResult.StatResult">statannotations.stats.StatResult.StatResult</a>], <a class="reference internal" href="statannotations.stats.html#statannotations.stats.StatResult.StatResult" title="statannotations.stats.StatResult.StatResult">statannotations.stats.StatResult.StatResult</a>], pvalue_thresholds: List)<a class="headerlink" href="#statannotations.format_annotations.pval_annotation_text" title="Permalink to this definition">¶</a></dt> +statannotations.format_annotations.pval_annotation_text(result: Union[List[<a class="reference internal" href="statannotations.stats.html#statannotations.stats.StatResult.StatResult" title="statannotations.stats.StatResult.StatResult">statannotations.stats.StatResult.StatResult</a>], <a class="reference internal" href="statannotations.stats.html#statannotations.stats.StatResult.StatResult" title="statannotations.stats.StatResult.StatResult">statannotations.stats.StatResult.StatResult</a>], pvalue_thresholds: List) → Union[List[str], str]<a class="headerlink" href="#statannotations.format_annotations.pval_annotation_text" title="Permalink to this definition">¶</a></dt> <dd><dl class="field-list simple"> <dt class="field-odd">Parameters</dt> <dd class="field-odd"><ul class="simple"> <li>result – StatResult instance or list thereof</li> -<li>pvalue_thresholds – thresholds to use for formatting</li> +<li>pvalue_thresholds – thresholds to use for formatter</li> </ul> </dd> <dt class="field-even">Returns</dt> -<dd class="field-even">A Series of rendered annotations if a list of StatResults was +<dd class="field-even">A List of rendered annotations if a list of StatResults was provided, a string otherwise. </dd> </dl> @@ -468,12 +526,6 @@ <h2>Submodules<a class="headerlink" href="#submodules" title="Permalink to this </div> <div class="section" id="module-statannotations.utils"> <h2>statannotations.utils module<a class="headerlink" href="#module-statannotations.utils" title="Permalink to this headline">¶</a></h2> -<dl class="py function"> -<dt class="sig sig-object py" id="statannotations.utils.check_box_pairs_in_data"> -statannotations.utils.check_box_pairs_in_data(box_pairs: Union[list, tuple], data: Optional[Union[List[list], pandas.core.frame.DataFrame]] = None, x: Optional[Union[str, list]] = None, hue: Optional[str] = None, hue_order: Optional[List[str]] = None)<a class="headerlink" href="#statannotations.utils.check_box_pairs_in_data" title="Permalink to this definition">¶</a></dt> -<dd>Checks that values referred to in <cite>order</cite> and <cite>box_pairs</cite> exist in data. -</dd></dl> - <dl class="py function"> <dt class="sig sig-object py" id="statannotations.utils.check_is_in"> statannotations.utils.check_is_in(x, valid_values, error_type=<class 'ValueError'>, label=None)<a class="headerlink" href="#statannotations.utils.check_is_in" title="Permalink to this definition">¶</a></dt> @@ -490,11 +542,25 @@ <h2>Submodules<a class="headerlink" href="#submodules" title="Permalink to this statannotations.utils.check_order_in_data(data, x, order) → None<a class="headerlink" href="#statannotations.utils.check_order_in_data" title="Permalink to this definition">¶</a></dt> <dd></dd></dl> +<dl class="py function"> +<dt class="sig sig-object py" id="statannotations.utils.check_pairs_in_data"> +statannotations.utils.check_pairs_in_data(pairs: Union[list, tuple], data: Optional[Union[List[list], pandas.core.frame.DataFrame]] = None, x: Optional[Union[str, list]] = None, hue: Optional[str] = None, hue_order: Optional[List[str]] = None)<a class="headerlink" href="#statannotations.utils.check_pairs_in_data" title="Permalink to this definition">¶</a></dt> +<dd>Checks that values referred to in <cite>order</cite> and <cite>pairs</cite> exist in data. +</dd></dl> + <dl class="py function"> <dt class="sig sig-object py" id="statannotations.utils.check_valid_text_format"> statannotations.utils.check_valid_text_format(text_format)<a class="headerlink" href="#statannotations.utils.check_valid_text_format" title="Permalink to this definition">¶</a></dt> <dd></dd></dl> +<dl class="py function"> +<dt class="sig sig-object py" id="statannotations.utils.get_closest"> +statannotations.utils.get_closest(a_list, value)<a class="headerlink" href="#statannotations.utils.get_closest" title="Permalink to this definition">¶</a></dt> +<dd>Assumes myList is sorted. Returns closest value to myNumber. +If two numbers are equally close, return the smallest number. +from <a class="reference external" href="https://stackoverflow.com/a/12141511/9981846">https://stackoverflow.com/a/12141511/9981846</a> +</dd></dl> + <dl class="py function"> <dt class="sig sig-object py" id="statannotations.utils.get_x_values"> statannotations.utils.get_x_values(data, x) → set<a class="headerlink" href="#statannotations.utils.get_x_values" title="Permalink to this definition">¶</a></dt> diff --git a/docs/build/html/statannotations.stats.html b/docs/build/html/statannotations.stats.html index 4a93cad..099ec99 100644 --- a/docs/build/html/statannotations.stats.html +++ b/docs/build/html/statannotations.stats.html @@ -95,6 +95,7 @@ </ul> </li> <li class="toctree-l3"><a class="reference internal" href="statannotations.html#submodules">Submodules</a></li> +<li class="toctree-l3"><a class="reference internal" href="statannotations.html#module-statannotations.Annotation">statannotations.Annotation module</a></li> <li class="toctree-l3"><a class="reference internal" href="statannotations.html#module-statannotations.Annotator">statannotations.Annotator module</a></li> <li class="toctree-l3"><a class="reference internal" href="statannotations.html#module-statannotations.PValueFormat">statannotations.PValueFormat module</a></li> <li class="toctree-l3"><a class="reference internal" href="statannotations.html#module-statannotations.format_annotations">statannotations.format_annotations module</a></li> @@ -186,6 +187,11 @@ <h2>Submodules<a class="headerlink" href="#submodules" title="Permalink to this <dt class="sig sig-object py" id="statannotations.stats.ComparisonsCorrection.ComparisonsCorrection"> class statannotations.stats.ComparisonsCorrection.ComparisonsCorrection(method: Union[str, callable], alpha: float = 0.05, name: Optional[str] = None, method_type: Optional[int] = None, statsmodels_api: bool = True, corr_kwargs: Optional[dict] = None)<a class="headerlink" href="#statannotations.stats.ComparisonsCorrection.ComparisonsCorrection" title="Permalink to this definition">¶</a></dt> <dd>Bases: <code class="xref py py-class docutils literal notranslate">object</code> +<dl class="py method"> +<dt class="sig sig-object py" id="statannotations.stats.ComparisonsCorrection.ComparisonsCorrection.apply"> +apply(test_result_list)<a class="headerlink" href="#statannotations.stats.ComparisonsCorrection.ComparisonsCorrection.apply" title="Permalink to this definition">¶</a></dt> +<dd></dd></dl> + <dl class="py method"> <dt class="sig sig-object py" id="statannotations.stats.ComparisonsCorrection.ComparisonsCorrection.document"> document(func)<a class="headerlink" href="#statannotations.stats.ComparisonsCorrection.ComparisonsCorrection.document" title="Permalink to this definition">¶</a></dt> @@ -258,7 +264,7 @@ <h2>Submodules<a class="headerlink" href="#submodules" title="Permalink to this <dl class="py function"> <dt class="sig sig-object py" id="statannotations.stats.StatTest.wilcoxon"> -statannotations.stats.StatTest.wilcoxon(box_data1, box_data2, verbose=1, **stats_params)<a class="headerlink" href="#statannotations.stats.StatTest.wilcoxon" title="Permalink to this definition">¶</a></dt> +statannotations.stats.StatTest.wilcoxon(group_data1, group_data2, verbose=1, **stats_params)<a class="headerlink" href="#statannotations.stats.StatTest.wilcoxon" title="Permalink to this definition">¶</a></dt> <dd>This function provides the equivalent behavior from earlier versions of statannot/statannotations. </dd></dl> @@ -268,13 +274,13 @@ <h2>Submodules<a class="headerlink" href="#submodules" title="Permalink to this <h2>statannotations.stats.test module<a class="headerlink" href="#module-statannotations.stats.test" title="Permalink to this headline">¶</a></h2> <dl class="py function"> <dt class="sig sig-object py" id="statannotations.stats.test.apply_test"> -statannotations.stats.test.apply_test(box_data1, box_data2, test: Optional[Union[<a class="reference internal" href="#statannotations.stats.StatTest.StatTest" title="statannotations.stats.StatTest.StatTest">statannotations.stats.StatTest.StatTest</a>, str]] = None, comparisons_correction: Optional[Union[<a class="reference internal" href="#statannotations.stats.ComparisonsCorrection.ComparisonsCorrection" title="statannotations.stats.ComparisonsCorrection.ComparisonsCorrection">statannotations.stats.ComparisonsCorrection.ComparisonsCorrection</a>, str]] = None, num_comparisons: int = 1, alpha: float = 0.05, **stats_params)<a class="headerlink" href="#statannotations.stats.test.apply_test" title="Permalink to this definition">¶</a></dt> +statannotations.stats.test.apply_test(group_data1, group_data2, test: Optional[Union[<a class="reference internal" href="#statannotations.stats.StatTest.StatTest" title="statannotations.stats.StatTest.StatTest">statannotations.stats.StatTest.StatTest</a>, str]] = None, comparisons_correction: Optional[Union[<a class="reference internal" href="#statannotations.stats.ComparisonsCorrection.ComparisonsCorrection" title="statannotations.stats.ComparisonsCorrection.ComparisonsCorrection">statannotations.stats.ComparisonsCorrection.ComparisonsCorrection</a>, str]] = None, num_comparisons: int = 1, alpha: float = 0.05, **stats_params)<a class="headerlink" href="#statannotations.stats.test.apply_test" title="Permalink to this definition">¶</a></dt> <dd>Get formatted result of two-sample statistical test. <dl class="field-list simple"> <dt class="field-odd">Parameters</dt> <dd class="field-odd"><ul class="simple"> -<li>box_data1 – data</li> -<li>box_data2 – data</li> +<li>group_data1 – data</li> +<li>group_data2 – data</li> <li>test – Union[StatTest, str]: Statistical test to run. Either a <cite>StatTest</cite> instance or one of: - <cite>Levene</cite> diff --git a/docs/source/statannotations.rst b/docs/source/statannotations.rst index d3f9b4e..3a24356 100644 --- a/docs/source/statannotations.rst +++ b/docs/source/statannotations.rst @@ -12,6 +12,14 @@ Subpackages Submodules ---------- +statannotations.Annotation module +--------------------------------- + +.. automodule:: statannotations.Annotation + :members: + :undoc-members: + :show-inheritance: + statannotations.Annotator module -------------------------------- diff --git a/statannotations/Annotation.py b/statannotations/Annotation.py new file mode 100644 index 0000000..3434ddd --- /dev/null +++ b/statannotations/Annotation.py @@ -0,0 +1,45 @@ +from typing import Union + +from statannotations.PValueFormat import Formatter +from statannotations.stats.StatResult import StatResult + + +class Annotation: + """ + Holds data, linked structs and an optional Formatter. + """ + def __init__(self, structs, data: Union[str, StatResult], + formatter: Formatter = None): + """ + :param structs: plot structures concerned by the data + :param data: a string or StatResult to be formatted by a formatter + :param formatter: A Formatter object. Statannotations provides a + PValueFormatter for StatResult objects. + """ + self.structs = structs + self.data = data + self.is_custom = isinstance(data, str) + self.formatter = formatter + + @property + def text(self): + if self.is_custom: + return self.data + else: + if self.formatter is None: + raise ValueError("Missing a PValueFormat object to " + "format the statistical result.") + return self.formatter.format_data(self.data) + + @property + def formatted_output(self): + if isinstance(self.data, str): + return self.data + else: + return self.data.formatted_output + + def print_labels_and_content(self, sep=" vs. "): + labels_string = sep.join([struct["label"] + for struct in self.structs]) + + print(f"{labels_string}: {self.formatted_output}") diff --git a/statannotations/Annotator.py b/statannotations/Annotator.py index 8cf3d14..390bd83 100644 --- a/statannotations/Annotator.py +++ b/statannotations/Annotator.py @@ -4,14 +4,13 @@ import matplotlib.pyplot as plt import matplotlib.transforms as mtransforms import numpy as np -import seaborn as sns from matplotlib import lines -from matplotlib.collections import PathCollection from matplotlib.font_manager import FontProperties -from matplotlib.patches import Rectangle +from statannotations.Annotation import Annotation from statannotations.PValueFormat import PValueFormat, \ CONFIGURABLE_PARAMETERS as PVALUE_CONFIGURABLE_PARAMETERS +from statannotations._Plotter import _Plotter, _SeabornPlotter from statannotations.stats.ComparisonsCorrection import \ get_validated_comparisons_correction from statannotations.stats.StatResult import StatResult @@ -19,8 +18,7 @@ from statannotations.stats.test import apply_test, IMPLEMENTED_TESTS from statannotations.stats.utils import check_alpha, check_pvalues, \ check_num_comparisons, get_num_comparisons -from statannotations.utils import check_is_in, remove_null, check_not_none, \ - check_box_pairs_in_data, check_order_in_data, render_collection +from statannotations.utils import check_is_in, render_collection CONFIGURABLE_PARAMETERS = [ 'alpha', @@ -28,7 +26,7 @@ 'comparisons_correction', 'line_height', 'line_offset', - 'line_offset_to_box', + 'line_offset_to_group', 'line_width', 'loc', 'pvalue_format', @@ -40,8 +38,6 @@ 'verbose', ] -IMPLEMENTED_PLOTTERS = ['barplot', 'boxplot', 'stripplot', 'swarmplot', - 'violinplot'] _DEFAULT_VALUES = { "alpha": 0.05, @@ -61,57 +57,50 @@ "custom_annotations": None } - -def check_implemented_plotters(plot): - if plot not in IMPLEMENTED_PLOTTERS: - raise NotImplementedError( - f"Only {render_collection(IMPLEMENTED_PLOTTERS)} are supported.") +ENGINE_PLOTTERS = {"seaborn": _SeabornPlotter} class Annotator: """ Optionally computes statistical test between pairs of data series, and add - statistical annotation on top of the boxes/bars. The same exact arguments - `data`, `x`, `y`, `hue`, `order`, `width`, `hue_order` (and `units`) as in - the seaborn boxplot/barplot function must be passed to this function. - - This function works in one of the two following modes: - a) `perform_stat_test` is True (default): - * statistical test as given by argument `test` is performed. - * The `test_short_name` argument can be used to customize what appears - before the pvalues if test is a string. - b) `perform_stat_test` is False: no statistical test is performed, list of - custom p-values `pvalues` are used for each pair of boxes. + statistical annotation on top of the groups (boxes, bars...). The same + exact arguments provided to the seaborn plotting function must be passed to + the constructor. + + This Annotator works in one of the three following modes: + - Add custom text annotations (`set_custom_annotations`) + - Format pvalues and add them to the plot (`set_pvalues`) + - Perform a statistical test and then add the results to the plot + (`apply_test`) """ - def __init__(self, ax, box_pairs, plot='boxplot', data=None, x=None, - y=None, hue=None, order=None, hue_order=None, **plot_params): + def __init__(self, ax, pairs, plot='boxplot', data=None, x=None, + y=None, hue=None, order=None, hue_order=None, + engine="seaborn", verbose=True, **plot_params): """ :param ax: Ax of existing plot - :param box_pairs: can be of either form: - For non-grouped boxplot: `[(cat1, cat2), (cat3, cat4)]`. - For boxplot grouped by hue: `[ + :param pairs: can be of either form: + For non-grouped plots: `[(cat1, cat2), (cat3, cat4)]`. + For plots grouped by hue: `[ ((cat1, hue1), (cat2, hue2)), ((cat3, hue3), (cat4, hue4)) ]` - :param plot: type of the plot, one of 'boxplot' or 'barplot'. + :param plot: type of the plot (default 'boxplot') :param data: seaborn plot's data :param x: seaborn plot's x :param y: seaborn plot's y :param hue: seaborn plot's hue :param order: seaborn plot's order :param hue_order: seaborn plot's hue_order - :param plot_params: Other parameters for seaborn plotter + :param engine: currently only "seaborn" is implemented + :param verbose: verbosity flag + :param plot_params: Other parameters for plotter engine """ - self.box_pairs = box_pairs + self.pairs = pairs self.ax = ax - basic_plot = self.get_basic_plot(ax, box_pairs, plot, data, x, y, hue, - order, hue_order, **plot_params) - - self.y_stack_arr, self.box_struct_pairs, self.fig = basic_plot - - self._reordering = None - self._sort_box_struct_pairs() + self._plotter = self._get_plotter(engine, ax, pairs, plot, data, x, y, + hue, order, hue_order, + verbose=verbose, **plot_params) self._test = None self.perform_stat_test = None @@ -123,11 +112,11 @@ def __init__(self, ax, box_pairs, plot='boxplot', data=None, x=None, self._comparisons_correction = None self._loc = "inside" - self.verbose = 1 + self._verbose = 1 self._just_configured = True self.show_test_name = True self.use_fixed_offset = False - self.line_offset_to_box = None + self.line_offset_to_group = None self.line_offset = None self.line_height = _DEFAULT_VALUES["line_height"] self.text_offset = 1 @@ -137,57 +126,44 @@ def __init__(self, ax, box_pairs, plot='boxplot', data=None, x=None, self.y_offset = None self.custom_annotations = None - self.OFFSET_FUNCS = {"inside": Annotator._get_offsets_inside, - "outside": Annotator._get_offsets_outside} - - @staticmethod - def get_basic_plot( - ax, box_pairs, plot='boxplot', data=None, x=None, y=None, hue=None, - order=None, hue_order=None, **plot_params): - check_implemented_plotters(plot) - check_not_none("box_pairs", box_pairs) - check_order_in_data(data, x, order) - check_box_pairs_in_data(box_pairs, data, x, hue, hue_order) - - plotter = Annotator._get_plotter(plot, x, y, hue, data, order, - hue_order, **plot_params) - - box_names, labels = Annotator._get_box_names_and_labels(plotter) - - fig = plt.gcf() - ymaxes = Annotator._generate_ymaxes(ax, fig, plotter, box_names) - box_structs = Annotator._get_box_structs( - plotter, box_names, labels, ymaxes) - - box_struct_pairs = Annotator._get_box_struct_pairs( - box_pairs, box_structs) + def new_plot(self, ax, pairs=None, plot='boxplot', data=None, x=None, + y=None, hue=None, order=None, hue_order=None, + engine: str = "seaborn", **plot_params): + self.ax = ax - y_stack_arr = np.array( - [[box_struct['x'], box_struct['ymax'], 0] - for box_struct in box_structs] - ).T + if pairs is None: + pairs = self.pairs - return y_stack_arr, box_struct_pairs, fig + self._plotter: _Plotter = self._get_plotter( + engine, ax, pairs, plot, data, x, y, hue, order, hue_order, + **plot_params) - def new_plot(self, ax, box_pairs=None, plot='boxplot', data=None, x=None, - y=None, hue=None, order=None, hue_order=None, **plot_params): - self.ax = ax + self.line_offset = None + self.line_offset_to_group = None + self.perform_stat_test = None - if box_pairs is None: - box_pairs = self.box_pairs + return self - basic_plot = self.get_basic_plot(ax, box_pairs, plot, data, x, y, hue, - order, hue_order, **plot_params) + @property + def verbose(self): + return self._verbose - self.y_stack_arr, self.box_struct_pairs, self.fig = basic_plot + @verbose.setter + def verbose(self, verbose): + self._verbose = verbose + self._plotter.verbose = verbose - self._sort_box_struct_pairs() + @property + def _y_stack_arr(self): + return self._plotter.y_stack_arr - self.line_offset = None - self.line_offset_to_box = None - self.perform_stat_test = None + @property + def fig(self): + return self._plotter.fig - return self + @property + def _struct_pairs(self): + return self._plotter.struct_pairs def reset_configuration(self): @@ -196,7 +172,7 @@ def reset_configuration(self): return self - def annotate(self, line_offset=None, line_offset_to_box=None): + def annotate(self, line_offset=None, line_offset_to_group=None): """Add configured annotations to the plot.""" if self._should_warn_about_configuration: warnings.warn("Annotator was reconfigured without applying the " @@ -208,28 +184,26 @@ def annotate(self, line_offset=None, line_offset_to_box=None): ann_list = [] orig_ylim = self.ax.get_ylim() - offset_func = self.OFFSET_FUNCS[self.loc] - self.y_offset, self.line_offset_to_box = offset_func( - line_offset, line_offset_to_box) + offset_func = self.get_offset_func(self.loc) + self.y_offset, self.line_offset_to_group = offset_func( + line_offset, line_offset_to_group) - if self.verbose: + if self._verbose: self.print_pvalue_legend() - ax_to_data = Annotator._get_transform_func(self.ax, 'ax_to_data') + ax_to_data = self._plotter.get_transform_func('ax_to_data') self.validate_test_short_name() - annotations = self.custom_annotations or self.annotations - for box_structs, result in zip(self.box_struct_pairs, - annotations): - self._annotate_pair(box_structs, result, + for annotation in self.annotations: + self._annotate_pair(annotation, ax_to_data=ax_to_data, ann_list=ann_list, orig_ylim=orig_ylim) # reset transformation - y_stack_max = max(self.y_stack_arr[1, :]) - ax_to_data = Annotator._get_transform_func(self.ax, 'ax_to_data') + y_stack_max = max(self._y_stack_arr[1, :]) + ax_to_data = self._plotter.get_transform_func('ax_to_data') ylims = ([(0, 0), (0, max(1.04 * y_stack_max, 1))] if self.loc == 'inside' else [(0, 0), (0, 1)]) @@ -250,7 +224,7 @@ def configure(self, **parameters): * `comparisons_correction` * `line_height`: in axes fraction coordinates * `line_offset` - * `line_offset_to_box` + * `line_offset_to_group` * `line_width` * `loc` * `pvalue_format` @@ -283,13 +257,7 @@ def configure(self, **parameters): self._activate_configured_warning() - if parameters.get("alpha"): - pvalue_format = parameters.get("pvalue_format") - if (pvalue_format is None - or pvalue_format.get("pvalue_thresholds") is None): - warnings.warn("Changing alpha without updating " - "pvalue_thresholds can result in inconsistent " - "plotting results") + self._warn_alpha_thresholds_if_necessary(parameters) return self def apply_test(self, num_comparisons='auto', **stats_params): @@ -297,7 +265,7 @@ def apply_test(self, num_comparisons='auto', **stats_params): :param stats_params: Parameters for statistical test functions. :param num_comparisons: Override number of comparisons otherwise - calculated with number of box_pairs + calculated with number of pairs """ self._check_test_pvalues_perform() @@ -319,9 +287,9 @@ def set_pvalues_and_annotate(self, pvalues, num_comparisons='auto'): def set_pvalues(self, pvalues, num_comparisons='auto'): """ - :param pvalues: list or array of p-values for each box pair comparison. + :param pvalues: list or array of p-values for each pair comparison. :param num_comparisons: Override number of comparisons otherwise - calculated with number of box_pairs + calculated with number of pairs """ self.perform_stat_test = False @@ -340,14 +308,24 @@ def set_pvalues(self, pvalues, num_comparisons='auto'): def set_custom_annotations(self, text_annot_custom): """ :param text_annot_custom: List of strings to annotate for each - `box_pair` + `pair` """ self._check_correct_number_custom_annotations(text_annot_custom) - self.custom_annotations = text_annot_custom + self.annotations = [Annotation(struct, text) for + struct, text in zip(self._struct_pairs, + text_annot_custom)] self.show_test_name = False self._deactivate_configured_warning() return self + def annotate_custom_annotations(self, text_annot_custom): + """ + :param text_annot_custom: List of strings to annotate for each + `pair` + """ + self.set_custom_annotations(text_annot_custom) + return self.annotate() + def get_configuration(self): configuration = {key: getattr(self, key) for key in CONFIGURABLE_PARAMETERS} @@ -356,15 +334,14 @@ def get_configuration(self): def get_annotations_text(self): if self.annotations is not None: - return [self._get_text(self.annotations[new_idx]) + return [self.annotations[new_idx].text for new_idx in self._reordering] - - if self.custom_annotations is not None: - return [self.custom_annotations[new_idx] - for new_idx in self._reordering] - return None + @property + def _reordering(self): + return self._plotter.reordering + @property def alpha(self): return self._alpha @@ -421,6 +398,11 @@ def pvalue_format(self, pvalue_format_config): def _should_warn_about_configuration(self): return self._just_configured + @classmethod + def get_offset_func(cls, position): + return {"inside": cls._get_offsets_inside, + "outside": cls._get_offsets_outside}[position] + def _activate_configured_warning(self): self._just_configured = True @@ -432,59 +414,50 @@ def print_pvalue_legend(self): self._pvalue_format.print_legend_if_used() def _get_results(self, num_comparisons, pvalues=None, - **stats_params) -> List[StatResult]: + **stats_params) -> List[Annotation]: test_result_list = [] if num_comparisons == "auto": - num_comparisons = len(self.box_struct_pairs) + num_comparisons = len(self._struct_pairs) - for box_struct1, box_struct2 in self.box_struct_pairs: - box1 = box_struct1['box'] - box2 = box_struct2['box'] + for group_struct1, group_struct2 in self._struct_pairs: + group1 = group_struct1['group'] + group2 = group_struct2['group'] if self.perform_stat_test: result = self._get_stat_result_from_test( - box_struct1, box_struct2, num_comparisons, **stats_params) + group_struct1, group_struct2, num_comparisons, + **stats_params) else: - result = self._get_custom_results(box_struct1, pvalues) + result = self._get_custom_results(group_struct1, pvalues) - result.box1 = box1 - result.box2 = box2 + result.group1 = group1 + result.group2 = group2 - test_result_list.append(result) + test_result_list.append(Annotation((group_struct1, group_struct2), + result, + formatter=self.pvalue_format)) # Perform other types of correction methods for multiple testing - Annotator._apply_comparisons_correction(self.comparisons_correction, - test_result_list) + self._apply_comparisons_correction(test_result_list) return test_result_list - def _get_text(self, result): - return self._pvalue_format.format_result(result) - - def _annotate_pair(self, box_structs, result, ax_to_data, - ann_list, orig_ylim): - - if self.custom_annotations is not None: - i_box_pair = box_structs[0]['i_box_pair'] - text = self.custom_annotations[i_box_pair] - - else: - if self.verbose >= 1: - _print_result_line(box_structs, result.formatted_output) + def _annotate_pair(self, annotation, ax_to_data, ann_list, orig_ylim): - text = self._get_text(result) + if self._verbose >= 1: + annotation.print_labels_and_content() - x1 = box_structs[0]['x'] - x2 = box_structs[1]['x'] - xi1 = box_structs[0]['xi'] - xi2 = box_structs[1]['xi'] + x1 = annotation.structs[0]['x'] + x2 = annotation.structs[1]['x'] + xi1 = annotation.structs[0]['xi'] + xi2 = annotation.structs[1]['xi'] - # Find y maximum for all the y_stacks *in between* box1 and box2 + # Find y maximum for all the y_stacks *in between* group1 and group2 i_ymax_in_range_x1_x2 = xi1 + np.nanargmax( - self.y_stack_arr[1, np.where((x1 <= self.y_stack_arr[0, :]) - & (self.y_stack_arr[0, :] <= x2))]) + self._y_stack_arr[1, np.where((x1 <= self._y_stack_arr[0, :]) + & (self._y_stack_arr[0, :] <= x2))]) y = self._get_y_for_pair(i_ymax_in_range_x1_x2) @@ -499,13 +472,13 @@ def _annotate_pair(self, box_structs, result, ax_to_data, line_x, line_y = zip(*points) self._plot_line(line_x, line_y) - if text is not None: - ann = self.ax.annotate( - text, xy=(np.mean([x1, x2]), line_y[2]), - xytext=(0, self.text_offset), textcoords='offset points', - xycoords='data', ha='center', va='bottom', - fontsize=self._pvalue_format.fontsize, clip_on=False, - annotation_clip=False) + ann = self.ax.annotate( + annotation.text, xy=(np.mean([x1, x2]), line_y[2]), + xytext=(0, self.text_offset), textcoords='offset points', + xycoords='data', ha='center', va='bottom', + fontsize=self._pvalue_format.fontsize, clip_on=False, + annotation_clip=False) + if annotation.text is not None: ann_list.append(ann) plt.draw() self.ax.set_ylim(orig_ylim) @@ -516,205 +489,16 @@ def _annotate_pair(self, box_structs, result, ax_to_data, # Fill the highest y position of the annotation into the y_stack array # for all positions in the range x1 to x2 - self.y_stack_arr[1, - (x1 <= self.y_stack_arr[0, :]) - & (self.y_stack_arr[0, :] <= x2)] = y_top_annot + self._y_stack_arr[1, + (x1 <= self._y_stack_arr[0, :]) + & (self._y_stack_arr[0, :] <= x2)] = y_top_annot # Increment the counter of annotations in the y_stack array - self.y_stack_arr[2, xi1:xi2 + 1] += 1 + self._y_stack_arr[2, xi1:xi2 + 1] += 1 def _update_y_for_loc(self): if self._loc == 'outside': - self.y_stack_arr[1, :] = 1 - - @staticmethod - def _get_box_x_position(b_plotter, box_name): - """ - box_name can be either a name "cat" or a tuple ("cat", "hue") - """ - if b_plotter.plot_hues is None: - cat = box_name - hue_offset = 0 - else: - cat = box_name[0] - hue_level = box_name[1] - hue_offset = b_plotter.hue_offsets[ - b_plotter.hue_names.index(hue_level)] - - group_pos = b_plotter.group_names.index(cat) - box_pos = group_pos + hue_offset - return box_pos - - @staticmethod - def _get_xpos_location(pos, xranges): - """ - Finds the x-axis location of a categorical variable - """ - for xrange in xranges: - if (pos >= xrange[0]) & (pos <= xrange[1]): - return xrange[2] - - @staticmethod - def _get_ypos_for_line2d_or_rectangle( - xpositions, data_to_ax, ax, fig, child): - xunits = (max(list(xpositions.keys())) + 1) / len(xpositions) - xranges = {(pos - xunits / 2, pos + xunits / 2, pos): box_name - for pos, box_name in xpositions.items()} - box = ax.transData.inverted().transform( - child.get_window_extent(fig.canvas.get_renderer())) - - if (box[:, 0].max() - box[:, 0].min()) > 1.1 * xunits: - return None, None - raw_xpos = np.round(box[:, 0].mean(), 1) - xpos = Annotator._get_xpos_location(raw_xpos, xranges) - if xpos not in xpositions: - return None, None - xname = xpositions[xpos] - ypos = box[:, 1].max() - ypos = data_to_ax.transform((0, ypos))[1] - - return xname, ypos - - @staticmethod - def _get_ypos_for_path_collection(xpositions, data_to_ax, child): - ymax = child.properties()['offsets'][:, 1].max() - xpos = float(np.round(np.nanmean( - child.properties()['offsets'][:, 0]), 1)) - - xname = xpositions[xpos] - ypos = data_to_ax.transform((0, ymax))[1] - - return xname, ypos - - @staticmethod - def _generate_ymaxes(ax, fig, plotter, box_names): - """ - given box plotter and the names of two categorical variables, - returns highest y point drawn between those two variables before - annotations. - - The highest data point is often not the highest item drawn - (eg, error bars and/or bar charts). - """ - xpositions = { - np.round(Annotator._get_box_x_position(plotter, box_name), - 1): box_name - for box_name in box_names} - - ymaxes = {name: 0 for name in box_names} - - data_to_ax = Annotator._get_transform_func(ax, 'data_to_ax') - # noinspection PyProtectedMember - if isinstance(plotter, sns.categorical._ViolinPlotter): - ymaxes = Annotator._get_ymaxes_violin(plotter, ymaxes, data_to_ax) - - else: - for child in ax.get_children(): - - xname, ypos = Annotator._get_child_ypos( - child, ax, fig, xpositions, data_to_ax) - - if ypos is not None and ypos > ymaxes[xname]: - ymaxes[xname] = ypos - - return ymaxes - - @staticmethod - def _get_child_ypos(child, ax, fig, xpositions, data_to_ax): - if ((type(child) == PathCollection) - and len(child.properties()['offsets'])): - return Annotator._get_ypos_for_path_collection( - xpositions, data_to_ax, child) - - elif (type(child) == lines.Line2D) or (type(child) == Rectangle): - return Annotator._get_ypos_for_line2d_or_rectangle( - xpositions, data_to_ax, ax, fig, child) - - return None, None - - @staticmethod - def _get_ymaxes_violin(plotter, ymaxes, data_to_ax): - for group_idx, group_name in enumerate(plotter.group_names): - if plotter.hue_names: - for hue_idx, hue_name in enumerate(plotter.hue_names): - ypos = max(plotter.support[group_idx][hue_idx]) - ymaxes[(group_name, hue_name)] = \ - data_to_ax.transform((0, ypos))[1] - else: - ypos = max(plotter.support[group_idx]) - ymaxes[group_name] = data_to_ax.transform((0, ypos))[1] - return ymaxes - - @staticmethod - def _get_group_data(b_plotter, cat): - index = b_plotter.group_names.index(cat) - - return b_plotter.plot_data[index], index - - @staticmethod - def _get_box_data_with_hue(b_plotter, box_name): - cat = box_name[0] - group_data, index = Annotator._get_group_data(b_plotter, cat) - - hue_level = box_name[1] - hue_mask = b_plotter.plot_hues[index] == hue_level - - return group_data[hue_mask] - - @staticmethod - def _get_box_data_without_hue(b_plotter, box_name): - cat = box_name - group_data, _ = Annotator._get_group_data(b_plotter, cat) - - return group_data - - @staticmethod - def _get_box_data(b_plotter, box_name): - """ - box_name can be either a name "cat" or a tuple ("cat", "hue") - - Almost a duplicate of seaborn's code, because there is not - direct access to the box_data in the BoxPlotter class. - """ - if b_plotter.plot_hues is None: - data = Annotator._get_box_data_without_hue(b_plotter, box_name) - else: - data = Annotator._get_box_data_with_hue(b_plotter, box_name) - - box_data = remove_null(data) - - return box_data - - @staticmethod - def _get_transform_func(ax, kind): - """ - Given an axis object, returns one of three possible transformation - functions to move between coordinate systems, depending on the value of - kind: - 'data_to_ax': converts data coordinates to axes coordinates - 'ax_to_data': converts axes coordinates to data coordinates - 'pix_to_ax': converts pixel coordinates to axes coordinates - 'all': return tuple of all three - - This function should be called whenever axes limits are altered. - """ - - check_is_in(kind, ['data_to_ax', 'ax_to_data', 'pix_to_ax', 'all'], - 'kind') - - if kind == 'pix_to_ax': - return ax.transAxes.inverted() - - data_to_ax = ax.transData + ax.get_xaxis_transform().inverted() - - if kind == 'data_to_ax': - return data_to_ax - - elif kind == 'ax_to_data': - return data_to_ax.inverted() - - else: - return data_to_ax, data_to_ax.inverted(), ax.transAxes.inverted() + self._y_stack_arr[1, :] = 1 def _check_test_pvalues_perform(self): if self.test is None: @@ -732,9 +516,9 @@ def _check_pvalues_no_perform(self, pvalues): raise ValueError("If `perform_stat_test` is False, " "custom `pvalues` must be specified.") check_pvalues(pvalues) - if len(pvalues) != len(self.box_pairs): + if len(pvalues) != len(self.pairs): raise ValueError("`pvalues` should be of the same length as " - "`box_pairs`.") + "`pairs`.") def _check_test_no_perform(self): if self.test is not None: @@ -745,53 +529,24 @@ def _check_correct_number_custom_annotations(self, text_annot_custom): if text_annot_custom is None: return - if len(text_annot_custom) != len(self.box_struct_pairs): + if len(text_annot_custom) != len(self._struct_pairs): raise ValueError( - "`text_annot_custom` should be of same length as `box_pairs`.") - - @staticmethod - def _apply_type1_comparisons_correction(comparisons_correction, - test_result_list): - original_pvalues = [result.pvalue for result in test_result_list] + "`text_annot_custom` should be of same length as `pairs`.") - significant_pvalues = comparisons_correction(original_pvalues) - correction_name = comparisons_correction.name - for is_significant, result in zip(significant_pvalues, - test_result_list): - result.correction_method = correction_name - result.corrected_significance = is_significant - - @staticmethod - def _apply_type0_comparisons_correction(comparisons_correction, - test_result_list): - alpha = comparisons_correction.alpha - for result in test_result_list: - result.correction_method = comparisons_correction.name - result.corrected_significance = ( - result.pvalue < alpha - or np.isclose(result.pvalue, alpha)) - - @staticmethod - def _apply_comparisons_correction(comparisons_correction, - test_result_list): - if comparisons_correction is None: + def _apply_comparisons_correction(self, annotations): + if self.comparisons_correction is None: return - - if comparisons_correction.type == 1: - Annotator._apply_type1_comparisons_correction( - comparisons_correction, test_result_list) - else: - Annotator._apply_type0_comparisons_correction( - comparisons_correction, test_result_list) + self.comparisons_correction.apply( + [annotation.data for annotation in annotations]) - def _get_stat_result_from_test(self, box_struct1, box_struct2, + def _get_stat_result_from_test(self, group_struct1, group_struct2, num_comparisons, **stats_params) -> StatResult: result = apply_test( - box_struct1['box_data'], - box_struct2['box_data'], + group_struct1['group_data'], + group_struct2['group_data'], self.test, comparisons_correction=self.comparisons_correction, num_comparisons=num_comparisons, @@ -801,8 +556,8 @@ def _get_stat_result_from_test(self, box_struct1, box_struct2, return result - def _get_custom_results(self, box_struct1, pvalues) -> StatResult: - pvalue = pvalues[box_struct1['i_box_pair']] + def _get_custom_results(self, group_struct1, pvalues) -> StatResult: + pvalue = pvalues[group_struct1['i_group_pair']] if self.has_type0_comparisons_correction(): pvalue = self.comparisons_correction(pvalue, len(pvalues)) @@ -819,162 +574,28 @@ def _get_custom_results(self, box_struct1, pvalues) -> StatResult: return result @staticmethod - def _get_box_struct_pairs(box_pairs, box_structs): - box_structs_dic = {box_struct['box']: box_struct - for box_struct in box_structs} - - box_struct_pairs = [] - - for i_box_pair, (box1, box2) in enumerate(box_pairs): - box_struct1 = dict(box_structs_dic[box1], i_box_pair=i_box_pair) - box_struct2 = dict(box_structs_dic[box2], i_box_pair=i_box_pair) - - if box_struct1['x'] <= box_struct2['x']: - box_struct_pairs.append((box_struct1, box_struct2)) - else: - box_struct_pairs.append((box_struct2, box_struct1)) - - return box_struct_pairs - - @staticmethod - def absolute_box_struct_pair_in_tuple_x_diff(box_struct_pair): - return abs(box_struct_pair[1][1]['x'] - box_struct_pair[1][0]['x']) - - def _sort_box_struct_pairs(self): - # Draw first the annotations with the shortest between-boxes distance, - # in order to reduce overlapping between annotations. - new_box_struct_pairs = sorted( - enumerate(self.box_struct_pairs), - key=self.absolute_box_struct_pair_in_tuple_x_diff) - - self._reordering, self.box_struct_pairs = zip(*new_box_struct_pairs) - - # noinspection PyProtectedMember - @staticmethod - def _get_plotter(plot, x, y, hue, data, order, hue_order, - **plot_params): - if not plot_params.pop("dodge", True): - raise ValueError("`dodge` must be True in statannotations") - - if plot == 'boxplot': - plotter = sns.categorical._BoxPlotter( - - x, y, hue, data, order, hue_order, - orient=plot_params.get("orient"), - width=plot_params.get("width", 0.8), - dodge=True, - fliersize=plot_params.get("fliersize", 5), - linewidth=plot_params.get("linewidth"), - saturation=.75, color=None, palette=None) - - elif plot == 'swarmplot': - plotter = sns.categorical._SwarmPlotter( - x, y, hue, data, order, hue_order, - orient=plot_params.get("orient"), - dodge=True, color=None, palette=None) - - elif plot == 'stripplot': - plotter = sns.categorical._StripPlotter( - x, y, hue, data, order, hue_order, - jitter=plot_params.get("jitter", True), - orient=plot_params.get("orient"), - dodge=True, color=None, palette=None) - - elif plot == 'barplot': - plotter = sns.categorical._BarPlotter( - x, y, hue, data, order, hue_order, - estimator=plot_params.get("estimator", np.mean), - ci=plot_params.get("ci", 95), - n_boot=plot_params.get("nboot", 1000), - units=plot_params.get("units"), - orient=plot_params.get("orient"), - seed=plot_params.get("seed"), - color=None, palette=None, saturation=.75, - errcolor=".26", errwidth=plot_params.get("errwidth"), - capsize=None, - dodge=True) - - elif plot == "violinplot": - plotter = sns.categorical._ViolinPlotter( - x, y, hue, data, order, hue_order, - bw=plot_params.get("bw", "scott"), - cut=plot_params.get("cut", 2), - scale=plot_params.get("scale", "area"), - scale_hue=plot_params.get("scale_hue", True), - gridsize=plot_params.get("gridsize", 100), - width=plot_params.get("width", 0.8), - inner=plot_params.get("inner", None), - split=plot_params.get("split", False), - dodge=True, orient=plot_params.get("orient"), - linewidth=plot_params.get("linewidth"), color=None, - palette=None, saturation=.75) - - else: - raise NotImplementedError( - f"Only {render_collection(IMPLEMENTED_PLOTTERS)} are " - f"supported.") - - return plotter - - @staticmethod - def _get_offsets_inside(line_offset, line_offset_to_box): + def _get_offsets_inside(line_offset, line_offset_to_group): if line_offset is None: line_offset = 0.05 - if line_offset_to_box is None: - line_offset_to_box = 0.06 + if line_offset_to_group is None: + line_offset_to_group = 0.06 else: - if line_offset_to_box is None: - line_offset_to_box = 0.06 + if line_offset_to_group is None: + line_offset_to_group = 0.06 y_offset = line_offset - return y_offset, line_offset_to_box + return y_offset, line_offset_to_group @staticmethod - def _get_offsets_outside(line_offset, line_offset_to_box): + def _get_offsets_outside(line_offset, line_offset_to_group): if line_offset is None: line_offset = 0.03 - if line_offset_to_box is None: - line_offset_to_box = line_offset + if line_offset_to_group is None: + line_offset_to_group = line_offset else: - line_offset_to_box = line_offset + line_offset_to_group = line_offset y_offset = line_offset - return y_offset, line_offset_to_box - - @staticmethod - def _get_box_structs(box_plotter, box_names, labels, ymaxes): - box_structs = [ - { - 'box': box_names[i], - 'label': labels[i], - 'x': Annotator._get_box_x_position(box_plotter, box_names[i]), - 'box_data': Annotator._get_box_data(box_plotter, box_names[i]), - 'ymax': ymaxes[box_names[i]] - } for i in range(len(box_names))] - - # Sort the box data structures by position along the x axis - box_structs = sorted(box_structs, key=lambda a: a['x']) - - # Add the index position in the list of boxes along the x axis - box_structs = [dict(box_struct, xi=i) - for i, box_struct in enumerate(box_structs)] - - return box_structs - - @staticmethod - def _get_box_names_and_labels(box_plotter): - group_names = box_plotter.group_names - hue_names = box_plotter.hue_names - - if box_plotter.plot_hues is None: - box_names = group_names - labels = box_names - else: - box_names = [(group_name, hue_name) for group_name in group_names - for hue_name in hue_names] - labels = ['{}_{}'.format(group_name, hue_name) - for (group_name, hue_name) in box_names] - - return box_names, labels + return y_offset, line_offset_to_group def validate_test_short_name(self): if self.test_short_name is not None: @@ -1010,7 +631,7 @@ def _annotate_pair_text(self, ann, y): if not self.use_fixed_offset: try: bbox = ann.get_window_extent() - pix_to_ax = Annotator._get_transform_func(self.ax, 'pix_to_ax') + pix_to_ax = self._plotter.get_transform_func('pix_to_ax') bbox_ax = bbox.transformed(pix_to_ax) y_top_annot = bbox_ax.ymax @@ -1018,7 +639,7 @@ def _annotate_pair_text(self, ann, y): got_mpl_error = True if self.use_fixed_offset or got_mpl_error: - if self.verbose >= 1: + if self._verbose >= 1: print("Warning: cannot get the text bounding box. Falling " "back to a fixed y offset. Layout may be not " "optimal.") @@ -1046,22 +667,33 @@ def _reset_default_values(self): def _get_y_for_pair(self, i_ymax_in_range_x1_x2): - ymax_in_range_x1_x2 = self.y_stack_arr[1, i_ymax_in_range_x1_x2] + ymax_in_range_x1_x2 = self._y_stack_arr[1, i_ymax_in_range_x1_x2] # Choose the best offset depending on whether there is an annotation # below at the x position in the range [x1, x2] where the stack is the # highest - if self.y_stack_arr[2, i_ymax_in_range_x1_x2] == 0: - # there is only a box below - offset = self.line_offset_to_box + if self._y_stack_arr[2, i_ymax_in_range_x1_x2] == 0: + # there is only a group below + offset = self.line_offset_to_group else: # there is an annotation below offset = self.y_offset + self.text_offset_impact_above return ymax_in_range_x1_x2 + offset + @staticmethod + def _warn_alpha_thresholds_if_necessary(parameters): + if parameters.get("alpha"): + pvalue_format = parameters.get("pvalue_format") + if (pvalue_format is None + or pvalue_format.get("pvalue_thresholds") is None): + warnings.warn("Changing alpha without updating " + "pvalue_thresholds can result in inconsistent " + "plotting results") -def _print_result_line(box_structs, formatted_output): - label1 = box_structs[0]['label'] - label2 = box_structs[1]['label'] - print(f"{label1} v.s. {label2}: {formatted_output}") + @staticmethod + def _get_plotter(engine, *args, **kwargs): + engine_plotter = ENGINE_PLOTTERS.get(engine) + if engine_plotter is None: + raise NotImplementedError(f"{engine} engine not implemented.") + return engine_plotter(*args, **kwargs) diff --git a/statannotations/PValueFormat.py b/statannotations/PValueFormat.py index be31722..e002992 100644 --- a/statannotations/PValueFormat.py +++ b/statannotations/PValueFormat.py @@ -12,8 +12,20 @@ ] -class PValueFormat: +class Formatter: def __init__(self): + pass + + def config(self, *args, **kwargs): + pass + + def format_data(self, data): + pass + + +class PValueFormat(Formatter): + def __init__(self): + Formatter.__init__(self) self._pvalue_format_string = '{:.3e}' self._simple_format_string = '{:.2f}' self._text_format = "star" @@ -144,7 +156,7 @@ def _update_pvalue_thresholds(self): self._pvalue_thresholds = self._get_pvalue_thresholds( self._pvalue_thresholds) - def format_result(self, result): + def format_data(self, result): if self.text_format == 'full': return ("{} p = {}{}" .format('{}', self.pvalue_format_string, '{}') diff --git a/statannotations/_Plotter.py b/statannotations/_Plotter.py new file mode 100644 index 0000000..9dac681 --- /dev/null +++ b/statannotations/_Plotter.py @@ -0,0 +1,379 @@ +import itertools +import warnings + +import matplotlib.pyplot as plt +import numpy as np +import seaborn as sns +from matplotlib import lines +from matplotlib.collections import PathCollection +from matplotlib.patches import Rectangle + +from statannotations._Xpositions import _XPositions +from statannotations.utils import check_not_none, check_order_in_data, \ + check_pairs_in_data, render_collection, check_is_in, remove_null + +IMPLEMENTED_PLOTTERS = { + 'seaborn': ['barplot', 'boxplot', 'stripplot', 'swarmplot', 'violinplot'] +} + + +class _Plotter: + def __init__(self, ax, pairs, data=None, x=None, hue=None, + order=None, hue_order=None, verbose=False): + self.ax = ax + self._fig = plt.gcf() + check_not_none("pairs", pairs) + check_order_in_data(data, x, order) + check_pairs_in_data(pairs, data, x, hue, hue_order) + self.pairs = pairs + self._struct_pairs = None + self.verbose = verbose + + def get_transform_func(self, kind: str): + """ + Given an axis object, returns one of three possible transformation + functions to move between coordinate systems, depending on the value of + kind: + 'data_to_ax': converts data coordinates to axes coordinates + 'ax_to_data': converts axes coordinates to data coordinates + 'pix_to_ax': converts pixel coordinates to axes coordinates + 'all': return tuple of all three + + This function should be called whenever axes limits are altered. + """ + + check_is_in(kind, ['data_to_ax', 'ax_to_data', 'pix_to_ax', 'all'], + 'kind') + + if kind == 'pix_to_ax': + return self.ax.transAxes.inverted() + + data_to_ax = \ + self.ax.transData + self.ax.get_xaxis_transform().inverted() + + if kind == 'data_to_ax': + return data_to_ax + + elif kind == 'ax_to_data': + return data_to_ax.inverted() + + else: + return (data_to_ax, data_to_ax.inverted(), + self.ax.transAxes.inverted()) + + @property + def fig(self): + return self._fig + + @property + def struct_pairs(self): + return self._struct_pairs + + +class _SeabornPlotter(_Plotter): + def __init__(self, ax, pairs, plot='boxplot', data=None, x=None, + y=None, hue=None, order=None, hue_order=None, verbose=False, + **plot_params): + + _Plotter.__init__(self, ax, pairs, data, x, hue, order, hue_order, + verbose) + + self.check_plot_is_implemented(plot) + + self.plot = plot + self.plotter = self._get_plotter(plot, x, y, hue, data, order, + hue_order, **plot_params) + + self.group_names, self.labels = self._get_group_names_and_labels() + self.xpositions = _XPositions(self.plotter, self.group_names) + self.reordering = None + self.ymaxes = self._generate_ymaxes() + + self.structs = self._get_structs() + self.pairs = pairs + self._struct_pairs = self._get_group_struct_pairs() + + self._y_stack_arr = np.array( + [[struct['x'], struct['ymax'], 0] for struct in self.structs] + ).T + + self.reordering = None + self._sort_group_struct_pairs() + + @property + def y_stack_arr(self): + return self._y_stack_arr + + # noinspection PyProtectedMember + def _get_plotter(self, plot, x, y, hue, data, order, hue_order, + **plot_params): + dodge = plot_params.pop("dodge", None) + self.fix_and_warn(dodge, hue, plot) + + if plot == 'boxplot': + plotter = sns.categorical._BoxPlotter( + + x, y, hue, data, order, hue_order, + orient=plot_params.get("orient"), + width=plot_params.get("width", 0.8), + dodge=True, + fliersize=plot_params.get("fliersize", 5), + linewidth=plot_params.get("linewidth"), + saturation=.75, color=None, palette=None) + + elif plot == 'swarmplot': + plotter = sns.categorical._SwarmPlotter( + x, y, hue, data, order, hue_order, + orient=plot_params.get("orient"), + dodge=True, color=None, palette=None) + + elif plot == 'stripplot': + plotter = sns.categorical._StripPlotter( + x, y, hue, data, order, hue_order, + jitter=plot_params.get("jitter", True), + orient=plot_params.get("orient"), + dodge=True, color=None, palette=None) + + elif plot == 'barplot': + plotter = sns.categorical._BarPlotter( + x, y, hue, data, order, hue_order, + estimator=plot_params.get("estimator", np.mean), + ci=plot_params.get("ci", 95), + n_boot=plot_params.get("nboot", 1000), + units=plot_params.get("units"), + orient=plot_params.get("orient"), + seed=plot_params.get("seed"), + color=None, palette=None, saturation=.75, + errcolor=".26", errwidth=plot_params.get("errwidth"), + capsize=None, + dodge=True) + + elif plot == "violinplot": + plotter = sns.categorical._ViolinPlotter( + x, y, hue, data, order, hue_order, + bw=plot_params.get("bw", "scott"), + cut=plot_params.get("cut", 2), + scale=plot_params.get("scale", "area"), + scale_hue=plot_params.get("scale_hue", True), + gridsize=plot_params.get("gridsize", 100), + width=plot_params.get("width", 0.8), + inner=plot_params.get("inner", None), + split=plot_params.get("split", False), + dodge=True, orient=plot_params.get("orient"), + linewidth=plot_params.get("linewidth"), color=None, + palette=None, saturation=.75) + + else: + raise NotImplementedError( + f"Only {render_collection(IMPLEMENTED_PLOTTERS)} are " + f"supported.") + + return plotter + + def _get_group_names_and_labels(self): + + if self.plotter.plot_hues is None: + group_names = self.plotter.group_names + labels = group_names + + else: + labels = [] + group_names = [] + for group_name, hue_name in \ + itertools.product(self.plotter.group_names, + self.plotter.hue_names): + group_names.append((group_name, hue_name)) + labels.append(f'{group_name}_{hue_name}') + + return group_names, labels + + def _generate_ymaxes(self): + """ + given plotter and the names of two categorical variables, + returns highest y point drawn between those two variables before + annotations. + + The highest data point is often not the highest item drawn + (eg, error bars and/or bar charts). + """ + + ymaxes = {name: 0 for name in self.group_names} + + data_to_ax = self.get_transform_func('data_to_ax') + + if self.plot == 'violinplot': + ymaxes = self._get_ymaxes_violin(ymaxes, data_to_ax) + + else: + for child in self.ax.get_children(): + + xname, ypos = self._get_child_ypos(child, data_to_ax) + + if ypos is not None and ypos > ymaxes[xname]: + ymaxes[xname] = ypos + + return ymaxes + + def _get_structs(self): + structs = [ + { + 'group': group_name, + 'label': self.labels[b_idx], + 'x': self.xpositions.get_group_x_position(group_name), + 'group_data': self._get_group_data(group_name), + 'ymax': self.ymaxes[group_name] + } for b_idx, group_name in enumerate(self.group_names)] + + # Sort the group data structures by position along the x axis + structs = sorted(structs, key=lambda struct: struct['x']) + + # Add the index position in the list of groups along the x axis + structs = [dict(struct, xi=i) + for i, struct in enumerate(structs)] + + return structs + + def _get_group_struct_pairs(self): + group_structs_dic = {struct['group']: struct + for struct in self.structs} + + group_struct_pairs = [] + + for i_group_pair, (group1, group2) in enumerate(self.pairs): + group_struct1 = dict(group_structs_dic[group1], + i_group_pair=i_group_pair) + + group_struct2 = dict(group_structs_dic[group2], + i_group_pair=i_group_pair) + + if group_struct1['x'] <= group_struct2['x']: + group_struct_pairs.append((group_struct1, group_struct2)) + + else: + group_struct_pairs.append((group_struct2, group_struct1)) + + return group_struct_pairs + + def _get_group_data_with_hue(self, group_name): + cat = group_name[0] + group_data, index = self._get_group_data_from_plotter(cat) + + hue_level = group_name[1] + hue_mask = self.plotter.plot_hues[index] == hue_level + + return group_data[hue_mask] + + def _get_group_data_from_plotter(self, cat): + index = self.plotter.group_names.index(cat) + + return self.plotter.plot_data[index], index + + def _get_group_data_without_hue(self, group_name): + group_data, _ = self._get_group_data_from_plotter(group_name) + + return group_data + + def _get_group_data(self, group_name): + """ + group_name can be either a name "cat" or a tuple ("cat", "hue") + + Almost a duplicate of seaborn's code, because there is not + direct access to the group_data in the respective Plotter class. + """ + if self.plotter.plot_hues is None: + data = self._get_group_data_without_hue(group_name) + else: + data = self._get_group_data_with_hue(group_name) + + group_data = remove_null(data) + + return group_data + + def _get_ymaxes_violin(self, ymaxes, data_to_ax): + for group_idx, group_name in enumerate(self.plotter.group_names): + if self.plotter.hue_names: + for hue_idx, hue_name in enumerate(self.plotter.hue_names): + ypos = max(self.plotter.support[group_idx][hue_idx]) + ymaxes[(group_name, hue_name)] = \ + data_to_ax.transform((0, ypos))[1] + else: + ypos = max(self.plotter.support[group_idx]) + ymaxes[group_name] = data_to_ax.transform((0, ypos))[1] + return ymaxes + + def _get_child_ypos(self, child, data_to_ax): + if (type(child) == PathCollection + and len(child.properties()['offsets'])): + return self._get_ypos_for_path_collection( + child, data_to_ax) + + elif type(child) in (lines.Line2D, Rectangle): + return self._get_ypos_for_line2d_or_rectangle( + child, data_to_ax) + + return None, None + + def _get_ypos_for_path_collection(self, child, data_to_ax): + ymax = child.properties()['offsets'][:, 1].max() + xpos = float(np.round(np.nanmean( + child.properties()['offsets'][:, 0]), 1)) + if xpos not in self.xpositions.xpositions: + if self.verbose: + warnings.warn( + "Invalid x-position found. Are the same parameters passed " + "to seaborn and statannotations calls? or are there few " + "data points?") + xpos = self.xpositions.find_closest(xpos) + + xname = self.xpositions.xpositions[xpos] + ypos = data_to_ax.transform((0, ymax))[1] + + return xname, ypos + + def _get_ypos_for_line2d_or_rectangle(self, child, data_to_ax): + bbox = self.ax.transData.inverted().transform( + child.get_window_extent(self.fig.canvas.get_renderer())) + + if (bbox[:, 0].max() - bbox[:, 0].min()) > 1.1*self.xpositions.xunits: + return None, None + raw_xpos = np.round(bbox[:, 0].mean(), 1) + xpos = self.xpositions.get_xpos_location(raw_xpos) + if xpos not in self.xpositions.xpositions: + return None, None + xname = self.xpositions.xpositions[xpos] + ypos = bbox[:, 1].max() + ypos = data_to_ax.transform((0, ypos))[1] + + return xname, ypos + + def _sort_group_struct_pairs(self): + # Draw first the annotations with the shortest between-groups distance, + # in order to reduce overlapping between annotations. + new_group_struct_pairs = sorted( + enumerate(self._struct_pairs), + key=self._absolute_group_struct_pair_in_tuple_x_diff) + + self.reordering, self._struct_pairs = zip(*new_group_struct_pairs) + + @staticmethod + def _absolute_group_struct_pair_in_tuple_x_diff(group_struct_pair): + return abs(group_struct_pair[1][1]['x'] - group_struct_pair[1][0]['x']) + + @staticmethod + def check_plot_is_implemented(plot, engine="seaborn"): + if plot not in IMPLEMENTED_PLOTTERS[engine]: + raise NotImplementedError( + f"Only {render_collection(IMPLEMENTED_PLOTTERS[engine])} are " + f"supported with {engine} engine.") + + @staticmethod + def fix_and_warn(dodge, hue, plot): + if dodge is False and hue: + raise ValueError("`dodge` cannot be False in statannotations.") + + if plot in ("swarmplot", 'stripplot') and hue: + if dodge is None: + warnings.warn( + "Implicitly setting dodge to True as it is necessary in " + "statannotations. It must have been True for the seaborn " + "call to yield consistent results when using `hue`.") diff --git a/statannotations/_Xpositions.py b/statannotations/_Xpositions.py new file mode 100644 index 0000000..2520f3c --- /dev/null +++ b/statannotations/_Xpositions.py @@ -0,0 +1,68 @@ +import numpy as np + +from statannotations.utils import get_closest + + +class _XPositions: + def __init__(self, plotter, group_names): + self._plotter = plotter + self._hue_names = self._plotter.hue_names + + if self._hue_names is not None: + nb_hues = len(self._hue_names) + if nb_hues == 1: + raise ValueError( + "Using hues with only one hue is not supported.") + + self.hue_offsets = self._plotter.hue_offsets + self._xunits = self.hue_offsets[1] - self.hue_offsets[0] + + self._xpositions = { + np.round(self.get_group_x_position(group_name), 1): group_name + for group_name in group_names + } + + self._xpositions_list = sorted(self._xpositions.keys()) + + if self._hue_names is None: + self._xunits = ((max(list(self._xpositions.keys())) + 1) + / len(self._xpositions)) + + self._xranges = { + (pos - self._xunits / 2, pos + self._xunits / 2, pos): group_name + for pos, group_name in self._xpositions.items()} + + @property + def xpositions(self): + return self._xpositions + + @property + def xunits(self): + return self._xunits + + def get_xpos_location(self, pos): + """ + Finds the x-axis location of a categorical variable + """ + for xrange in self._xranges: + if (pos >= xrange[0]) & (pos <= xrange[1]): + return xrange[2] + + def get_group_x_position(self, group): + """ + group_name can be either a name "cat" or a tuple ("cat", "hue") + """ + if self._plotter.plot_hues is None: + cat = group + hue_offset = 0 + else: + cat = group[0] + hue_level = group[1] + hue_offset = self._plotter.hue_offsets[ + self._plotter.hue_names.index(hue_level)] + + group_pos = self._plotter.group_names.index(cat) + hue_offset + return group_pos + + def find_closest(self, xpos): + return get_closest(list(self._xpositions_list), xpos) diff --git a/statannotations/_version.py b/statannotations/_version.py index 8c2ecfb..6a9beea 100644 --- a/statannotations/_version.py +++ b/statannotations/_version.py @@ -1,1 +1,1 @@ -__version__ = "0.4.0a" +__version__ = "0.4.0" diff --git a/statannotations/format_annotations.py b/statannotations/format_annotations.py index 29ba57b..81bb6e8 100644 --- a/statannotations/format_annotations.py +++ b/statannotations/format_annotations.py @@ -2,48 +2,42 @@ from typing import Union, List import numpy as np import pandas as pd +from operator import itemgetter def pval_annotation_text(result: Union[List[StatResult], StatResult], - pvalue_thresholds: List): + pvalue_thresholds: List) -> Union[List[str], str]: """ :param result: StatResult instance or list thereof - :param pvalue_thresholds: thresholds to use for formatting - :returns: A Series of rendered annotations if a list of StatResults was + :param pvalue_thresholds: thresholds to use for formatter + :returns: A List of rendered annotations if a list of StatResults was provided, a string otherwise. """ - single_value = False + was_list = True - if isinstance(result, list): - x1_pval = np.array([res.pvalue for res in result]) - x1_signif_suff = [res.significance_suffix for res in result] - else: - x1_pval = np.array([result.pvalue]) - x1_signif_suff = [result.significance_suffix] - single_value = True + if not isinstance(result, list): + was_list = False + result = [result] + + x1_pval = np.array([res.pvalue for res in result]) - # Sort the threshold array - pvalue_thresholds = (pd.DataFrame(pvalue_thresholds) - .sort_values(by=0, ascending=False) - .values) - x_annot = pd.Series(["" for _ in range(len(x1_pval))]) + pvalue_thresholds = sorted(pvalue_thresholds, key=itemgetter(0), + reverse=True) - for i in range(0, len(pvalue_thresholds)): + x_annot = pd.Series(["" for _ in x1_pval]) - if i < len(pvalue_thresholds) - 1: - condition = (x1_pval <= pvalue_thresholds[i][0])\ - & (pvalue_thresholds[i + 1][0] < x1_pval) - x_annot[condition] = pvalue_thresholds[i][1] + for i, threshold in enumerate(pvalue_thresholds[:-1]): + condition = ((x1_pval <= threshold[0]) + & (pvalue_thresholds[i + 1][0] < x1_pval)) + x_annot[condition] = threshold[1] - else: - condition = x1_pval < pvalue_thresholds[i][0] - x_annot[condition] = pvalue_thresholds[i][1] + x_annot[x1_pval <= pvalue_thresholds[-1][0]] = pvalue_thresholds[-1][1] - x_annot = pd.Series([f"{star}{signif}" - for star, signif in zip(x_annot, x1_signif_suff)]) + x_annot = [f"{star}{res.significance_suffix}" + for star, res in zip(x_annot, result)] - return x_annot if not single_value else x_annot.iloc[0] + return x_annot if was_list else x_annot[0] def simple_text(result: StatResult, pvalue_format, pvalue_thresholds) -> str: diff --git a/statannotations/stats/ComparisonsCorrection.py b/statannotations/stats/ComparisonsCorrection.py index ab0b016..33688b9 100644 --- a/statannotations/stats/ComparisonsCorrection.py +++ b/statannotations/stats/ComparisonsCorrection.py @@ -197,3 +197,27 @@ def __call__(self, pvalues: Union[List[float], float], result = result[:n_vals] return result if got_pvalues_in_array else result[0] + + def _apply_type1_comparisons_correction(self, test_result_list): + original_pvalues = [result.pvalue for result in test_result_list] + + significant_pvalues = self(original_pvalues) + for is_significant, result in zip(significant_pvalues, + test_result_list): + result.correction_method = self.name + result.corrected_significance = is_significant + + def _apply_type0_comparisons_correction(self, test_result_list): + for result in test_result_list: + result.correction_method = self.name + result.corrected_significance = ( + result.pvalue < self.alpha + or np.isclose(result.pvalue, self.alpha)) + + def apply(self, test_result_list): + + if self.type == 1: + self._apply_type1_comparisons_correction(test_result_list) + + else: + self._apply_type0_comparisons_correction(test_result_list) diff --git a/statannotations/stats/StatTest.py b/statannotations/stats/StatTest.py index 0fddcc5..40e776b 100644 --- a/statannotations/stats/StatTest.py +++ b/statannotations/stats/StatTest.py @@ -6,20 +6,20 @@ # Also example for how to add other functions -def wilcoxon(box_data1, box_data2, verbose=1, **stats_params): +def wilcoxon(group_data1, group_data2, verbose=1, **stats_params): """ This function provides the equivalent behavior from earlier versions of statannot/statannotations. """ zero_method = stats_params.pop('zero_method', None) if zero_method is None: - return stats.wilcoxon(box_data1, box_data2, **stats_params) + return stats.wilcoxon(group_data1, group_data2, **stats_params) if zero_method == "AUTO": - zero_method = len(box_data1) <= 20 and "pratt" or "wilcox" + zero_method = len(group_data1) <= 20 and "pratt" or "wilcox" if verbose >= 1: print("Using zero_method ", zero_method) - return stats.wilcoxon(box_data1, box_data2, + return stats.wilcoxon(group_data1, group_data2, zero_method=zero_method, **stats_params) @@ -71,10 +71,10 @@ def __init__(self, func: Callable, test_long_name: str, f"------ Original function documentation ------ \n" f"{func.__doc__}") - def __call__(self, box_data1, box_data2, alpha=0.05, + def __call__(self, group_data1, group_data2, alpha=0.05, **stat_params): - stat, pval = self._func(box_data1, box_data2, *self.args, + stat, pval = self._func(group_data1, group_data2, *self.args, **{**self.kwargs, **stat_params})[:2] return StatResult(self._test_long_name, self._test_short_name, diff --git a/statannotations/utils.py b/statannotations/utils.py index 6aa6396..c7d44fc 100644 --- a/statannotations/utils.py +++ b/statannotations/utils.py @@ -1,4 +1,5 @@ import itertools +from bisect import bisect_left from typing import List, Union import pandas as pd @@ -57,40 +58,40 @@ def check_order_in_data(data, x, order) -> None: f"(specified in `order`)") -def _check_box_pairs_in_data_no_hue(box_pairs: Union[list, tuple], - data: Union[List[list], - pd.DataFrame] = None, - x: Union[str, list] = None) -> None: +def _check_pairs_in_data_no_hue(pairs: Union[list, tuple], + data: Union[List[list], + pd.DataFrame] = None, + x: Union[str, list] = None) -> None: x_values = get_x_values(data, x) - pairs_x_values = set(itertools.chain(*box_pairs)) + pairs_x_values = set(itertools.chain(*pairs)) unmatched_x_values = pairs_x_values - x_values if unmatched_x_values: raise ValueError(f"Missing x value(s) " f"`{render_collection(unmatched_x_values)}` in {x}" - f" (specified in `box_pairs`) in data") + f" (specified in `pairs`) in data") -def _check_box_pairs_in_data_with_hue(box_pairs: Union[list, tuple], - data: Union[List[list], - pd.DataFrame] = None, - x: Union[str, list] = None, - hue: str = None) -> set: +def _check_pairs_in_data_with_hue(pairs: Union[list, tuple], + data: Union[List[list], + pd.DataFrame] = None, + x: Union[str, list] = None, + hue: str = None) -> set: x_values = get_x_values(data, x) seen_x_values = set() hue_values = set(data[hue].unique()) - for x_value, hue_value in itertools.chain(itertools.chain(*box_pairs)): + for x_value, hue_value in itertools.chain(itertools.chain(*pairs)): if x_value not in seen_x_values and x_value not in x_values: raise ValueError(f"Missing x value `{x_value}` in {x}" - f" (specified in `box_pairs`)") + f" (specified in `pairs`)") seen_x_values.add(x_value) if hue_value not in hue_values: raise ValueError(f"Missing hue value `{hue_value}` in {hue}" - f" (specified in `box_pairs`)") + f" (specified in `pairs`)") return hue_values @@ -105,19 +106,19 @@ def _check_hue_order_in_data(hue, hue_values: set, f" in {hue} (specified in `hue_order`)") -def check_box_pairs_in_data(box_pairs: Union[list, tuple], - data: Union[List[list], pd.DataFrame] = None, - x: Union[str, list] = None, - hue: str = None, - hue_order: List[str] = None): +def check_pairs_in_data(pairs: Union[list, tuple], + data: Union[List[list], pd.DataFrame] = None, + x: Union[str, list] = None, + hue: str = None, + hue_order: List[str] = None): """ - Checks that values referred to in `order` and `box_pairs` exist in data. + Checks that values referred to in `order` and `pairs` exist in data. """ if hue is None and hue_order is None: - _check_box_pairs_in_data_no_hue(box_pairs, data, x) + _check_pairs_in_data_no_hue(pairs, data, x) else: - hue_values = _check_box_pairs_in_data_with_hue(box_pairs, data, x, hue) + hue_values = _check_pairs_in_data_with_hue(pairs, data, x, hue) _check_hue_order_in_data(hue, hue_values, hue_order) @@ -136,3 +137,20 @@ def check_valid_text_format(text_format): def render_collection(collection): return '","'.join(map(str, collection)) + + +def get_closest(a_list, value): + """ + Assumes myList is sorted. Returns closest value to myNumber. + If two numbers are equally close, return the smallest number. + from https://stackoverflow.com/a/12141511/9981846 + """ + pos = bisect_left(a_list, value) + if pos == 0: + return a_list[0] + if pos == len(a_list): + return a_list[-1] + before, after = a_list[pos - 1: pos + 1] + if after - value < value - before: + return after + return before diff --git a/usage/example.ipynb b/usage/example.ipynb index 7d17821..10b91f9 100644 --- a/usage/example.ipynb +++ b/usage/example.ipynb @@ -62,9 +62,9 @@ " ***: 1.00e-04 < p <= 1.00e-03\n", " ****: p <= 1.00e-04\n", "\n", - "Thur v.s. Fri: Mann-Whitney-Wilcoxon test two-sided, P_val:6.477e-01 U_stat=6.305e+02\n", - "Thur v.s. Sat: Mann-Whitney-Wilcoxon test two-sided, P_val:4.690e-02 U_stat=2.180e+03\n", - "Sun v.s. Fri: Mann-Whitney-Wilcoxon test two-sided, P_val:2.680e-02 U_stat=9.605e+02\n" + "Thur vs. Fri: Mann-Whitney-Wilcoxon test two-sided, P_val:6.477e-01 U_stat=6.305e+02\n", + "Thur vs. Sat: Mann-Whitney-Wilcoxon test two-sided, P_val:4.690e-02 U_stat=2.180e+03\n", + "Sun vs. Fri: Mann-Whitney-Wilcoxon test two-sided, P_val:2.680e-02 U_stat=9.605e+02\n" ] }, { @@ -99,7 +99,9 @@ "cell_type": "markdown", "metadata": {}, "source": [ - "The `annotate` method returns the tuple `ax, test_results`, where `test_results` is a list of `StatResult` objects, containing both the original data of the boxes and the statistical test results (p-value, etc)." + "The `annotate` method returns the tuple `ax, annotations`, where `annotations` is a list of `Annotation` objects.\n", + "The `data` attribute of an Annotation is a `StatResult` object, which containing both the original data of the groups\n", + "(boxes for boxplots), and the statistical test results (p-value, etc)." ] }, { @@ -115,15 +117,15 @@ "Mann-Whitney-Wilcoxon test two-sided, P_val:4.690e-02 U_stat=2.180e+03\n", "Mann-Whitney-Wilcoxon test two-sided, P_val:2.680e-02 U_stat=9.605e+02\n", "\n", - "StatResult attributes: dict_keys(['test_description', 'test_short_name', 'stat_str', 'stat_value', 'pvalue', '_corrected_significance', '_correction_method', 'alpha', 'box1', 'box2'])\n" + "StatResult attributes: dict_keys(['test_description', 'test_short_name', 'stat_str', 'stat_value', 'pvalue', '_corrected_significance', '_correction_method', 'alpha', 'group1', 'group2'])\n" ] } ], "source": [ "for res in test_results:\n", - " print(res)\n", + " print(res.data)\n", "\n", - "print(\"\\nStatResult attributes:\", test_results[0].__dict__.keys())" + "print(\"\\nStatResult attributes:\", test_results[0].data.__dict__.keys())" ] }, { @@ -132,7 +134,7 @@ "source": [ "#### Applying a multiple comparisons correction (requires `statsmodels`)\n", "In this example, note that previously configured parameters are remembered if not changed.\n", - "This is the case for `box_pairs`, `test`, `text_format`, and `loc`.\n", + "This is the case for `pairs`, `test`, `text_format`, and `loc`.\n", "\n", "To avoid this, you can call `annot.reset_configuration()` to get back to all default/unset values." ] @@ -153,9 +155,9 @@ " ***: 1.00e-04 < p <= 1.00e-03\n", " ****: p <= 1.00e-04\n", "\n", - "Thur v.s. Fri: Mann-Whitney-Wilcoxon test two-sided with Bonferroni correction, P_val:1.000e+00 U_stat=6.305e+02\n", - "Thur v.s. Sat: Mann-Whitney-Wilcoxon test two-sided with Bonferroni correction, P_val:1.407e-01 U_stat=2.180e+03\n", - "Sun v.s. Fri: Mann-Whitney-Wilcoxon test two-sided with Bonferroni correction, P_val:8.041e-02 U_stat=9.605e+02\n" + "Thur vs. Fri: Mann-Whitney-Wilcoxon test two-sided with Bonferroni correction, P_val:1.000e+00 U_stat=6.305e+02\n", + "Thur vs. Sat: Mann-Whitney-Wilcoxon test two-sided with Bonferroni correction, P_val:1.407e-01 U_stat=2.180e+03\n", + "Sun vs. Fri: Mann-Whitney-Wilcoxon test two-sided with Bonferroni correction, P_val:8.041e-02 U_stat=9.605e+02\n" ] }, { @@ -222,14 +224,14 @@ " ***: 1.00e-04 < p <= 1.00e-03\n", " ****: p <= 1.00e-04\n", "\n", - "Thur v.s. Fri: Mann-Whitney-Wilcoxon test two-sided with Benjamini-Hochberg correction, P_val:6.477e-01 U_stat=6.305e+02\n", - "Thur v.s. Sat: Mann-Whitney-Wilcoxon test two-sided with Benjamini-Hochberg correction, P_val:4.690e-02 (ns) U_stat=2.180e+03\n", - "Sun v.s. Fri: Mann-Whitney-Wilcoxon test two-sided with Benjamini-Hochberg correction, P_val:2.680e-02 (ns) U_stat=9.605e+02\n" + "Thur vs. Fri: Mann-Whitney-Wilcoxon test two-sided with Benjamini-Hochberg correction, P_val:6.477e-01 U_stat=6.305e+02\n", + "Thur vs. Sat: Mann-Whitney-Wilcoxon test two-sided with Benjamini-Hochberg correction, P_val:4.690e-02 (ns) U_stat=2.180e+03\n", + "Sun vs. Fri: Mann-Whitney-Wilcoxon test two-sided with Benjamini-Hochberg correction, P_val:2.680e-02 (ns) U_stat=9.605e+02\n" ] }, { "data": { - "text/plain": "(<matplotlib.axes._subplots.AxesSubplot at 0x7fb3e83f8128>,\n [<statannotations.stats.StatResult.StatResult at 0x7fb3e83ef908>,\n <statannotations.stats.StatResult.StatResult at 0x7fb3e6e8cf60>,\n <statannotations.stats.StatResult.StatResult at 0x7fb3e83efcc0>])" + "text/plain": "(<matplotlib.axes._subplots.AxesSubplot at 0x7fedb54c7748>,\n [<statannotations.Annotation.Annotation at 0x7fedb4233ac8>,\n <statannotations.Annotation.Annotation at 0x7fedb4233d30>,\n <statannotations.Annotation.Annotation at 0x7fedb4233320>])" }, "execution_count": 6, "metadata": {}, @@ -275,9 +277,9 @@ " ***: 1.00e-04 < p <= 1.00e-03\n", " ****: p <= 1.00e-04\n", "\n", - "Sun v.s. Thur: Custom statistical test with Benjamini-Hochberg correction, P_val:1.000e-01\n", - "Sun v.s. Fri: Custom statistical test with Benjamini-Hochberg correction, P_val:1.000e-03\n", - "Sun v.s. Sat: Custom statistical test with Benjamini-Hochberg correction, P_val:1.000e-01\n" + "Sun vs. Thur: Custom statistical test with Benjamini-Hochberg correction, P_val:1.000e-01\n", + "Sun vs. Fri: Custom statistical test with Benjamini-Hochberg correction, P_val:1.000e-03\n", + "Sun vs. Sat: Custom statistical test with Benjamini-Hochberg correction, P_val:1.000e-01\n" ] }, { @@ -293,9 +295,9 @@ "x = \"day\"\n", "y = \"total_bill\"\n", "order = ['Sun', 'Thur', 'Fri', 'Sat']\n", - "box_pairs=[(\"Sun\", \"Thur\"), (\"Sun\", \"Sat\"), (\"Fri\", \"Sun\")]\n", + "pairs = [(\"Sun\", \"Thur\"), (\"Sun\", \"Sat\"), (\"Fri\", \"Sun\")]\n", "ax = sns.boxplot(data=df, x=x, y=y, order=order)\n", - "annot.new_plot(ax, box_pairs=box_pairs, data=df, x=x, y=y, order=order)\n", + "annot.new_plot(ax, pairs=pairs, data=df, x=x, y=y, order=order)\n", "annot.configure(test=None, loc='inside')\n", "annot.set_pvalues([0.1, 0.1, 0.001])\n", "annot.annotate()\n", @@ -357,15 +359,15 @@ " ***: 1.00e-04 < p <= 1.00e-03\n", " ****: p <= 1.00e-04\n", "\n", - "E_Ideal v.s. E_Premium: Mann-Whitney-Wilcoxon test two-sided with Benjamini-Hochberg correction, P_val:1.560e-31 U_stat=3.756e+06\n", - "I_Ideal v.s. I_Premium: Mann-Whitney-Wilcoxon test two-sided with Benjamini-Hochberg correction, P_val:5.141e-61 U_stat=1.009e+06\n", - "J_Ideal v.s. J_Premium: Mann-Whitney-Wilcoxon test two-sided with Benjamini-Hochberg correction, P_val:4.018e-37 U_stat=2.337e+05\n", - "E_Ideal v.s. E_Good: Mann-Whitney-Wilcoxon test two-sided with Benjamini-Hochberg correction, P_val:5.201e-19 U_stat=1.480e+06\n", - "I_Ideal v.s. I_Good: Mann-Whitney-Wilcoxon test two-sided with Benjamini-Hochberg correction, P_val:5.008e-13 U_stat=4.359e+05\n", - "J_Ideal v.s. J_Good: Mann-Whitney-Wilcoxon test two-sided with Benjamini-Hochberg correction, P_val:1.006e-04 U_stat=1.174e+05\n", - "E_Ideal v.s. E_Very Good: Mann-Whitney-Wilcoxon test two-sided with Benjamini-Hochberg correction, P_val:1.736e-02 U_stat=4.850e+06\n", - "E_Good v.s. I_Ideal: Mann-Whitney-Wilcoxon test two-sided with Benjamini-Hochberg correction, P_val:5.906e-01 U_stat=9.882e+05\n", - "I_Premium v.s. J_Ideal: Mann-Whitney-Wilcoxon test two-sided with Benjamini-Hochberg correction, P_val:5.159e-27 U_stat=8.084e+05\n" + "E_Ideal vs. E_Premium: Mann-Whitney-Wilcoxon test two-sided with Benjamini-Hochberg correction, P_val:1.560e-31 U_stat=3.756e+06\n", + "I_Ideal vs. I_Premium: Mann-Whitney-Wilcoxon test two-sided with Benjamini-Hochberg correction, P_val:5.141e-61 U_stat=1.009e+06\n", + "J_Ideal vs. J_Premium: Mann-Whitney-Wilcoxon test two-sided with Benjamini-Hochberg correction, P_val:4.018e-37 U_stat=2.337e+05\n", + "E_Ideal vs. E_Good: Mann-Whitney-Wilcoxon test two-sided with Benjamini-Hochberg correction, P_val:5.201e-19 U_stat=1.480e+06\n", + "I_Ideal vs. I_Good: Mann-Whitney-Wilcoxon test two-sided with Benjamini-Hochberg correction, P_val:5.008e-13 U_stat=4.359e+05\n", + "J_Ideal vs. J_Good: Mann-Whitney-Wilcoxon test two-sided with Benjamini-Hochberg correction, P_val:1.006e-04 U_stat=1.174e+05\n", + "E_Ideal vs. E_Very Good: Mann-Whitney-Wilcoxon test two-sided with Benjamini-Hochberg correction, P_val:1.736e-02 U_stat=4.850e+06\n", + "E_Good vs. I_Ideal: Mann-Whitney-Wilcoxon test two-sided with Benjamini-Hochberg correction, P_val:5.906e-01 U_stat=9.882e+05\n", + "I_Premium vs. J_Ideal: Mann-Whitney-Wilcoxon test two-sided with Benjamini-Hochberg correction, P_val:5.159e-27 U_stat=8.084e+05\n" ] }, { @@ -383,7 +385,7 @@ "hue = \"cut\"\n", "hue_order=['Ideal', 'Premium', 'Good', 'Very Good', 'Fair']\n", "order = [\"E\", \"I\", \"J\"]\n", - "box_pairs=[\n", + "pairs=[\n", " ((\"E\", \"Ideal\"), (\"E\", \"Very Good\")),\n", " ((\"E\", \"Ideal\"), (\"E\", \"Premium\")),\n", " ((\"E\", \"Ideal\"), (\"E\", \"Good\")),\n", @@ -395,7 +397,7 @@ " ((\"I\", \"Premium\"), (\"J\", \"Ideal\")),\n", " ]\n", "ax = sns.boxplot(data=df, x=x, y=y, order=order, hue=hue, hue_order=hue_order)\n", - "annot.new_plot(ax, box_pairs, data=df, x=x, y=y, order=order, hue=hue, hue_order=hue_order)\n", + "annot.new_plot(ax, pairs, data=df, x=x, y=y, order=order, hue=hue, hue_order=hue_order)\n", "annot.configure(test='Mann-Whitney', verbose=2)\n", "annot.apply_test()\n", "annot.annotate()\n", @@ -477,7 +479,7 @@ " ***: 1.00e-04 < p <= 1.00e-03\n", " ****: p <= 1.00e-04\n", "\n", - "Fri_(0.991, 4.0] v.s. Sat_(7.0, 10.0]: t-test independent samples, P_val:6.176e-07 t=-7.490e+00\n" + "Fri_(0.991, 4.0] vs. Sat_(7.0, 10.0]: t-test independent samples, P_val:6.176e-07 t=-7.490e+00\n" ] }, { @@ -538,15 +540,15 @@ " ***: 1.00e-04 < p <= 1.00e-03\n", " ****: p <= 1.00e-04\n", "\n", - "SI2_I v.s. SI2_J: t-test independent samples, P_val:3.710e-01 t=-8.949e-01\n", - "SI1_I v.s. SI1_J: t-test independent samples, P_val:8.513e-05 t=-3.937e+00\n", - "VS1_I v.s. VS1_J: t-test independent samples, P_val:9.140e-05 t=-3.923e+00\n", - "VS2_I v.s. VS2_J: t-test independent samples, P_val:6.806e-03 t=-2.709e+00\n", - "VVS2_I v.s. VVS2_J: t-test independent samples, P_val:9.076e-14 t=-7.672e+00\n", - "VVS1_I v.s. VVS1_J: t-test independent samples, P_val:6.536e-10 t=-6.321e+00\n", - "I1_I v.s. I1_J: t-test independent samples, P_val:2.905e-02 t=-2.206e+00\n", - "IF_I v.s. IF_J: t-test independent samples, P_val:4.412e-03 t=-2.881e+00\n", - "SI1_E v.s. SI2_E: t-test independent samples, P_val:4.178e-66 t=-1.749e+01\n" + "SI2_I vs. SI2_J: t-test independent samples, P_val:3.710e-01 t=-8.949e-01\n", + "SI1_I vs. SI1_J: t-test independent samples, P_val:8.513e-05 t=-3.937e+00\n", + "VS1_I vs. VS1_J: t-test independent samples, P_val:9.140e-05 t=-3.923e+00\n", + "VS2_I vs. VS2_J: t-test independent samples, P_val:6.806e-03 t=-2.709e+00\n", + "VVS2_I vs. VVS2_J: t-test independent samples, P_val:9.076e-14 t=-7.672e+00\n", + "VVS1_I vs. VVS1_J: t-test independent samples, P_val:6.536e-10 t=-6.321e+00\n", + "I1_I vs. I1_J: t-test independent samples, P_val:2.905e-02 t=-2.206e+00\n", + "IF_I vs. IF_J: t-test independent samples, P_val:4.412e-03 t=-2.881e+00\n", + "SI1_E vs. SI2_E: t-test independent samples, P_val:4.178e-66 t=-1.749e+01\n" ] }, { @@ -566,14 +568,14 @@ "y = \"carat\"\n", "hue = \"color\"\n", "hue_order=[\"E\", \"I\", \"J\"]\n", - "box_pairs = [\n", + "pairs = [\n", " ((\"SI2\", \"E\"), (\"SI1\", \"E\"))\n", " ]\n", - "box_pairs = box_pairs + [((clar, 'I'), (clar, 'J'))\n", + "pairs = pairs + [((clar, 'I'), (clar, 'J'))\n", " for clar in df['clarity'].unique()]\n", "width = 0.4\n", "ax = sns.barplot(data=df, x=x, y=y, hue=hue, hue_order=hue_order, seed=2021)\n", - "annot.new_plot(ax, box_pairs, plot='barplot',\n", + "annot.new_plot(ax, pairs, plot='barplot',\n", " data=df, x=x, y=y, hue=hue, hue_order=hue_order, seed=2021)\n", "annot.apply_test().annotate()\n", "\n", @@ -604,9 +606,9 @@ "name": "stdout", "output_type": "stream", "text": [ - "Sat_Yes v.s. Sat_No: t-test independent samples, P_val:4.304e-01 t=7.922e-01\n", - "Thur_No v.s. Fri_No: t-test independent samples, P_val:7.425e-01 t=-3.305e-01\n", - "Thur_Yes v.s. Sun_No: t-test independent samples, P_val:5.623e-01 t=-5.822e-01\n" + "Sat_Yes vs. Sat_No: t-test independent samples, P_val:4.304e-01 t=7.922e-01\n", + "Thur_No vs. Fri_No: t-test independent samples, P_val:7.425e-01 t=-3.305e-01\n", + "Thur_Yes vs. Sun_No: t-test independent samples, P_val:5.623e-01 t=-5.822e-01\n" ] }, { @@ -623,16 +625,16 @@ "x = \"day\"\n", "y = \"total_bill\"\n", "hue = \"smoker\"\n", - "box_pairs = [((\"Thur\", \"No\"), (\"Fri\", \"No\")),\n", - " ((\"Sat\", \"Yes\"), (\"Sat\", \"No\")),\n", - " ((\"Sun\", \"No\"), (\"Thur\", \"Yes\"))]\n", + "pairs = [((\"Thur\", \"No\"), (\"Fri\", \"No\")),\n", + " ((\"Sat\", \"Yes\"), (\"Sat\", \"No\")),\n", + " ((\"Sun\", \"No\"), (\"Thur\", \"Yes\"))]\n", "ax = sns.boxplot(data=df, x=x, y=y, hue=hue)\n", - "annot.new_plot(ax, box_pairs,\n", + "annot.new_plot(ax, pairs,\n", " data=df, x=x, y=y, hue=hue)\n", "annot.configure(test='t-test_ind', text_format='full', loc='inside',\n", " comparisons_correction=None, line_height=0.05, text_offset=8)\n", "\n", - "annot.apply_test().annotate(line_offset_to_box=0.2, line_offset=0.1)\n", + "annot.apply_test().annotate(line_offset_to_group=0.2, line_offset=0.1)\n", "\n", "plt.legend(loc='upper left', bbox_to_anchor=(1.03, 1))\n", "plt.savefig('example_tuning_y_offsets.svg')\n", @@ -663,14 +665,14 @@ " ***: 1.00e-04 < p <= 1.00e-03\n", " ****: p <= 1.00e-04\n", "\n", - "setosa v.s. versicolor: Bartlett statistical test for equal variances, P_val:8.660e-03 Stat=6.892e+00\n", - "versicolor v.s. virginica: Bartlett statistical test for equal variances, P_val:1.478e-01 Stat=2.095e+00\n", - "setosa v.s. virginica: Bartlett statistical test for equal variances, P_val:6.379e-05 Stat=1.599e+01\n" + "setosa vs. versicolor: Bartlett statistical test for equal variances, P_val:8.660e-03 Stat=6.892e+00\n", + "versicolor vs. virginica: Bartlett statistical test for equal variances, P_val:1.478e-01 Stat=2.095e+00\n", + "setosa vs. virginica: Bartlett statistical test for equal variances, P_val:6.379e-05 Stat=1.599e+01\n" ] }, { "data": { - "text/plain": "(<matplotlib.axes._subplots.AxesSubplot at 0x7fb3e9c366a0>,\n [<statannotations.stats.StatResult.StatResult at 0x7fb3e9b0fcf8>,\n <statannotations.stats.StatResult.StatResult at 0x7fb3e74f5b38>,\n <statannotations.stats.StatResult.StatResult at 0x7fb3e9c2b5c0>])" + "text/plain": "(<matplotlib.axes._subplots.AxesSubplot at 0x7fedb4460898>,\n [<statannotations.Annotation.Annotation at 0x7fedb40b4400>,\n <statannotations.Annotation.Annotation at 0x7fedb715ccf8>,\n <statannotations.Annotation.Annotation at 0x7fedb6c8ef28>])" }, "execution_count": 15, "metadata": {}, @@ -695,7 +697,7 @@ "x = \"species\"\n", "y = \"sepal_length\"\n", "\n", - "box_pairs = [(\"setosa\", \"versicolor\"), (\"setosa\", \"virginica\"), (\"versicolor\", \"virginica\")]\n", + "pairs = [(\"setosa\", \"versicolor\"), (\"setosa\", \"virginica\"), (\"versicolor\", \"virginica\")]\n", "\n", "# Required descriptors for annotate\n", "custom_long_name = 'Bartlett statistical test for equal variances'\n", @@ -706,7 +708,7 @@ "# Then, same as usual\n", "ax = sns.boxplot(data=df, x=x, y=y)\n", "annot.reset_configuration()\n", - "annot.new_plot(ax, box_pairs, data=df, x=x, y=y)\n", + "annot.new_plot(ax, pairs, data=df, x=x, y=y)\n", "annot.configure(test=custom_test, comparisons_correction=None,\n", " text_format='star').apply_test().annotate()" ], @@ -748,7 +750,7 @@ "\n", "test_short_name = 'Bartlett'\n", "pvalues = []\n", - "for pair in box_pairs:\n", + "for pair in pairs:\n", " data1 = df.groupby(x)[y].get_group(pair[0])\n", " data2 = df.groupby(x)[y].get_group(pair[1])\n", " stat, p = bartlett(data1, data2)\n", @@ -776,9 +778,9 @@ " ***: 1.00e-04 < p <= 1.00e-03\n", " ****: p <= 1.00e-04\n", "\n", - "setosa v.s. versicolor: Custom statistical test, P_val:8.660e-03\n", - "versicolor v.s. virginica: Custom statistical test, P_val:1.478e-01\n", - "setosa v.s. virginica: Custom statistical test, P_val:6.379e-05\n" + "setosa vs. versicolor: Custom statistical test, P_val:8.660e-03\n", + "versicolor vs. virginica: Custom statistical test, P_val:1.478e-01\n", + "setosa vs. virginica: Custom statistical test, P_val:6.379e-05\n" ] }, { @@ -792,7 +794,7 @@ ], "source": [ "ax = sns.boxplot(data=df, x=x, y=y)\n", - "annot.new_plot(ax=ax, box_pairs=box_pairs,\n", + "annot.new_plot(ax=ax, pairs=pairs,\n", " data=df, x=x, y=y)\n", "(annot\n", " .configure(test=None, test_short_name=test_short_name)\n", @@ -825,9 +827,9 @@ " ***: 1.00e-04 < p <= 1.00e-03\n", " ****: p <= 1.00e-04\n", "\n", - "setosa v.s. versicolor: Custom statistical test, P_val:8.660e-03\n", - "versicolor v.s. virginica: Custom statistical test, P_val:1.478e-01\n", - "setosa v.s. virginica: Custom statistical test, P_val:6.379e-05\n" + "setosa vs. versicolor: Custom statistical test, P_val:8.660e-03\n", + "versicolor vs. virginica: Custom statistical test, P_val:1.478e-01\n", + "setosa vs. virginica: Custom statistical test, P_val:6.379e-05\n" ] }, { @@ -841,7 +843,7 @@ ], "source": [ "ax = sns.violinplot(data=df, x=x, y=y)\n", - "annot.new_plot(ax=ax, box_pairs=box_pairs, plot=\"violinplot\",\n", + "annot.new_plot(ax=ax, pairs=pairs, plot=\"violinplot\",\n", " data=df, x=x, y=y)\n", "(annot\n", " .configure(test=None, test_short_name=test_short_name)\n", @@ -873,6 +875,22 @@ } }, "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "p-value annotation legend:\n", + " ns: p <= 1.00e+00\n", + " *: 1.00e-02 < p <= 5.00e-02\n", + " **: 1.00e-03 < p <= 1.00e-02\n", + " ***: 1.00e-04 < p <= 1.00e-03\n", + " ****: p <= 1.00e-04\n", + "\n", + "Thur vs. Fri: first pair\n", + "Thur vs. Sat: second pair\n", + "Sun vs. Fri: third pair\n" + ] + }, { "data": { "text/plain": "<Figure size 432x288 with 1 Axes>", @@ -888,7 +906,7 @@ "y = \"total_bill\"\n", "order = ['Sun', 'Thur', 'Fri', 'Sat']\n", "ax = sns.boxplot(data=df, x=x, y=y, order=order)\n", - "annot.new_plot(ax, box_pairs=[(\"Thur\", \"Fri\"),(\"Thur\", \"Sat\"), (\"Fri\", \"Sun\")],\n", + "annot.new_plot(ax, pairs=[(\"Thur\", \"Fri\"),(\"Thur\", \"Sat\"), (\"Fri\", \"Sun\")],\n", " data=df, x=x, y=y, order=order)\n", "annot.configure(loc='outside')\n", "annot.set_custom_annotations([\"first pair\", \"second pair\", \"third pair\"])\n", @@ -921,15 +939,15 @@ " ***: 1.00e-04 < p <= 1.00e-03\n", " ****: p <= 1.00e-04\n", "\n", - "H1N1_Nonsynonymous v.s. H1N1_Synonymous: Mann-Whitney-Wilcoxon test two-sided, P_val:5.014e-04 U_stat=2.624e+03\n", - "H3N2_Nonsynonymous v.s. H3N2_Synonymous: Mann-Whitney-Wilcoxon test two-sided, P_val:1.294e-03 U_stat=1.535e+04\n", - "Influenza B_Nonsynonymous v.s. Influenza B_Synonymous: Mann-Whitney-Wilcoxon test two-sided, P_val:2.026e-01 U_stat=3.340e+02\n" + "H1N1_Nonsynonymous vs. H1N1_Synonymous: Mann-Whitney-Wilcoxon test two-sided, P_val:5.014e-04 U_stat=2.624e+03\n", + "H3N2_Nonsynonymous vs. H3N2_Synonymous: Mann-Whitney-Wilcoxon test two-sided, P_val:1.294e-03 U_stat=1.535e+04\n", + "Influenza B_Nonsynonymous vs. Influenza B_Synonymous: Mann-Whitney-Wilcoxon test two-sided, P_val:2.026e-01 U_stat=3.340e+02\n" ] }, { "data": { "text/plain": "<Figure size 864x432 with 1 Axes>", - "image/png": "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\n" + "image/png": "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\n" }, "metadata": {}, "output_type": "display_data" @@ -939,7 +957,13 @@ "samples = pd.read_csv('samples.tsv', sep = '\\t')\n", "samples = samples.loc[samples.mutation_rate_samp < 1]\n", "samples = samples.drop(columns=['mutation_rate_samp'])\n", - "samples = samples.melt(id_vars=['sampleID', 'subtype'], value_vars=['nonsynon_mutation_rate_samp','synon_mutation_rate_samp'], var_name='Synon_Nonsynon', value_name='mutation_rate_samp').replace('nonsynon_mutation_rate_samp', 'Nonsynonymous').replace('synon_mutation_rate_samp', 'Synonymous')\n", + "samples = (\n", + " samples.melt(id_vars=['sampleID', 'subtype'],\n", + " value_vars=['nonsynon_mutation_rate_samp','synon_mutation_rate_samp'],\n", + " var_name='Synon_Nonsynon', value_name='mutation_rate_samp')\n", + " .replace('nonsynon_mutation_rate_samp', 'Nonsynonymous')\n", + " .replace('synon_mutation_rate_samp', 'Synonymous'))\n", + "\n", "samples = samples.loc[samples.mutation_rate_samp > 0]\n", "\n", "significanceComparisons = [(('H3N2','Synonymous'), ('H3N2','Nonsynonymous')),\n", @@ -962,7 +986,7 @@ "ax.set_yscale('log')\n", "\n", "sns.stripplot(ax=ax, **fig_args)\n", - "annotator = Annotator(ax=ax, box_pairs=significanceComparisons,\n", + "annotator = Annotator(ax=ax, pairs=significanceComparisons,\n", " **fig_args, plot='stripplot')\n", "annotator.configure(**configuration).apply_test().annotate()\n", "fig.savefig(f'flu_dataset_log_scale_in_axes_strip.svg', format='svg')" @@ -991,9 +1015,9 @@ "name": "stdout", "output_type": "stream", "text": [ - "H1N1_Nonsynonymous v.s. H1N1_Synonymous: Mann-Whitney-Wilcoxon test two-sided, P_val:5.014e-04 U_stat=2.624e+03\n", - "H3N2_Nonsynonymous v.s. H3N2_Synonymous: Mann-Whitney-Wilcoxon test two-sided, P_val:1.294e-03 U_stat=1.535e+04\n", - "Influenza B_Nonsynonymous v.s. Influenza B_Synonymous: Mann-Whitney-Wilcoxon test two-sided, P_val:2.026e-01 U_stat=3.340e+02\n" + "H1N1_Nonsynonymous vs. H1N1_Synonymous: Mann-Whitney-Wilcoxon test two-sided, P_val:5.014e-04 U_stat=2.624e+03\n", + "H3N2_Nonsynonymous vs. H3N2_Synonymous: Mann-Whitney-Wilcoxon test two-sided, P_val:1.294e-03 U_stat=1.535e+04\n", + "Influenza B_Nonsynonymous vs. Influenza B_Synonymous: Mann-Whitney-Wilcoxon test two-sided, P_val:2.026e-01 U_stat=3.340e+02\n" ] }, { @@ -1042,9 +1066,9 @@ "name": "stdout", "output_type": "stream", "text": [ - "H1N1_Nonsynonymous v.s. H1N1_Synonymous: Mann-Whitney-Wilcoxon test two-sided, P_val:5.014e-04 U_stat=2.624e+03\n", - "H3N2_Nonsynonymous v.s. H3N2_Synonymous: Mann-Whitney-Wilcoxon test two-sided, P_val:1.294e-03 U_stat=1.535e+04\n", - "Influenza B_Nonsynonymous v.s. Influenza B_Synonymous: Mann-Whitney-Wilcoxon test two-sided, P_val:2.026e-01 U_stat=3.340e+02\n" + "H1N1_Nonsynonymous vs. H1N1_Synonymous: Mann-Whitney-Wilcoxon test two-sided, P_val:5.014e-04 U_stat=2.624e+03\n", + "H3N2_Nonsynonymous vs. H3N2_Synonymous: Mann-Whitney-Wilcoxon test two-sided, P_val:1.294e-03 U_stat=1.535e+04\n", + "Influenza B_Nonsynonymous vs. Influenza B_Synonymous: Mann-Whitney-Wilcoxon test two-sided, P_val:2.026e-01 U_stat=3.340e+02\n" ] }, { diff --git a/usage/example_tuning_y_offsets.svg b/usage/example_tuning_y_offsets.svg index 56e1850..42a7655 100644 --- a/usage/example_tuning_y_offsets.svg +++ b/usage/example_tuning_y_offsets.svg @@ -347,7 +347,7 @@ z <g id="matplotlib.axis_2"> <g id="ytick_1"> <g id="line2d_1"> - <path clip-path="url(#p51652a52ce)" d="M 54 227.515867 + <path clip-path="url(#p5b84badbdb)" d="M 54 227.515867 L 388.8 227.515867 " style="fill:none;stroke:#cccccc;stroke-linecap:round;"/> </g> @@ -399,7 +399,7 @@ z </g> <g id="ytick_2"> <g id="line2d_2"> - <path clip-path="url(#p51652a52ce)" d="M 54 201.23688 + <path clip-path="url(#p5b84badbdb)" d="M 54 201.23688 L 388.8 201.23688 " style="fill:none;stroke:#cccccc;stroke-linecap:round;"/> </g> @@ -439,7 +439,7 @@ z </g> <g id="ytick_3"> <g id="line2d_3"> - <path clip-path="url(#p51652a52ce)" d="M 54 174.957893 + <path clip-path="url(#p5b84badbdb)" d="M 54 174.957893 L 388.8 174.957893 " style="fill:none;stroke:#cccccc;stroke-linecap:round;"/> </g> @@ -488,7 +488,7 @@ z </g> <g id="ytick_4"> <g id="line2d_4"> - <path clip-path="url(#p51652a52ce)" d="M 54 148.678905 + <path clip-path="url(#p5b84badbdb)" d="M 54 148.678905 L 388.8 148.678905 " style="fill:none;stroke:#cccccc;stroke-linecap:round;"/> </g> @@ -521,7 +521,7 @@ z </g> <g id="ytick_5"> <g id="line2d_5"> - <path clip-path="url(#p51652a52ce)" d="M 54 122.399918 + <path clip-path="url(#p5b84badbdb)" d="M 54 122.399918 L 388.8 122.399918 " style="fill:none;stroke:#cccccc;stroke-linecap:round;"/> </g> @@ -562,7 +562,7 @@ z </g> <g id="ytick_6"> <g id="line2d_6"> - <path clip-path="url(#p51652a52ce)" d="M 54 96.12093 + <path clip-path="url(#p5b84badbdb)" d="M 54 96.12093 L 388.8 96.12093 " style="fill:none;stroke:#cccccc;stroke-linecap:round;"/> </g> @@ -611,7 +611,7 @@ z </g> <g id="ytick_7"> <g id="line2d_7"> - <path clip-path="url(#p51652a52ce)" d="M 54 69.841943 + <path clip-path="url(#p5b84badbdb)" d="M 54 69.841943 L 388.8 69.841943 " style="fill:none;stroke:#cccccc;stroke-linecap:round;"/> </g> @@ -640,7 +640,7 @@ z </g> <g id="ytick_8"> <g id="line2d_8"> - <path clip-path="url(#p51652a52ce)" d="M 54 43.562956 + <path clip-path="url(#p5b84badbdb)" d="M 54 43.562956 L 388.8 43.562956 " style="fill:none;stroke:#cccccc;stroke-linecap:round;"/> </g> @@ -772,7 +772,7 @@ z </g> </g> <g id="patch_3"> - <path clip-path="url(#p51652a52ce)" d="M 77.131636 218.291943 + <path clip-path="url(#p5b84badbdb)" d="M 77.131636 218.291943 L 106.959273 218.291943 L 106.959273 201.736181 L 77.131636 201.736181 @@ -781,7 +781,7 @@ z " style="fill:#5875a4;stroke:#4c4c4c;stroke-linejoin:miter;stroke-width:1.5;"/> </g> <g id="patch_4"> - <path clip-path="url(#p51652a52ce)" d="M 107.568 223.074719 + <path clip-path="url(#p5b84badbdb)" d="M 107.568 223.074719 L 137.395636 223.074719 L 137.395636 200.527347 L 107.568 200.527347 @@ -790,7 +790,7 @@ z " style="fill:#cc8963;stroke:#4c4c4c;stroke-linejoin:miter;stroke-width:1.5;"/> </g> <g id="patch_5"> - <path clip-path="url(#p51652a52ce)" d="M 153.222545 223.074719 + <path clip-path="url(#p5b84badbdb)" d="M 153.222545 223.074719 L 183.050182 223.074719 L 183.050182 204.745125 L 153.222545 204.745125 @@ -799,7 +799,7 @@ z " style="fill:#5875a4;stroke:#4c4c4c;stroke-linejoin:miter;stroke-width:1.5;"/> </g> <g id="patch_6"> - <path clip-path="url(#p51652a52ce)" d="M 183.658909 214.113584 + <path clip-path="url(#p5b84badbdb)" d="M 183.658909 214.113584 L 213.486545 214.113584 L 213.486545 194.522599 L 183.658909 194.522599 @@ -808,7 +808,7 @@ z " style="fill:#cc8963;stroke:#4c4c4c;stroke-linejoin:miter;stroke-width:1.5;"/> </g> <g id="patch_7"> - <path clip-path="url(#p51652a52ce)" d="M 229.313455 218.567872 + <path clip-path="url(#p5b84badbdb)" d="M 229.313455 218.567872 L 259.141091 218.567872 L 259.141091 183.386878 L 229.313455 183.386878 @@ -817,7 +817,7 @@ z " style="fill:#5875a4;stroke:#4c4c4c;stroke-linejoin:miter;stroke-width:1.5;"/> </g> <g id="patch_8"> - <path clip-path="url(#p51652a52ce)" d="M 259.749818 215.085906 + <path clip-path="url(#p5b84badbdb)" d="M 259.749818 215.085906 L 289.577455 215.085906 L 289.577455 199.528746 L 259.749818 199.528746 @@ -826,7 +826,7 @@ z " style="fill:#cc8963;stroke:#4c4c4c;stroke-linejoin:miter;stroke-width:1.5;"/> </g> <g id="patch_9"> - <path clip-path="url(#p51652a52ce)" d="M 305.404364 208.686973 + <path clip-path="url(#p5b84badbdb)" d="M 305.404364 208.686973 L 335.232 208.686973 L 335.232 168.716633 L 305.404364 168.716633 @@ -835,7 +835,7 @@ z " style="fill:#5875a4;stroke:#4c4c4c;stroke-linejoin:miter;stroke-width:1.5;"/> </g> <g id="patch_10"> - <path clip-path="url(#p51652a52ce)" d="M 335.840727 214.954511 + <path clip-path="url(#p5b84badbdb)" d="M 335.840727 214.954511 L 365.668364 214.954511 L 365.668364 188.097386 L 335.840727 188.097386 @@ -844,7 +844,7 @@ z " style="fill:#cc8963;stroke:#4c4c4c;stroke-linejoin:miter;stroke-width:1.5;"/> </g> <g id="patch_11"> - <path clip-path="url(#p51652a52ce)" d="M 107.263636 253.794855 + <path clip-path="url(#p5b84badbdb)" d="M 107.263636 253.794855 L 107.263636 253.794855 L 107.263636 253.794855 L 107.263636 253.794855 @@ -852,7 +852,7 @@ z " style="fill:#5875a4;stroke:#4c4c4c;stroke-linejoin:miter;stroke-width:0.75;"/> </g> <g id="patch_12"> - <path clip-path="url(#p51652a52ce)" d="M 107.263636 253.794855 + <path clip-path="url(#p5b84badbdb)" d="M 107.263636 253.794855 L 107.263636 253.794855 L 107.263636 253.794855 L 107.263636 253.794855 @@ -860,22 +860,22 @@ z " style="fill:#cc8963;stroke:#4c4c4c;stroke-linejoin:miter;stroke-width:0.75;"/> </g> <g id="line2d_9"> - <path clip-path="url(#p51652a52ce)" d="M 92.045455 218.291943 + <path clip-path="url(#p5b84badbdb)" d="M 92.045455 218.291943 L 92.045455 226.622382 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_10"> - <path clip-path="url(#p51652a52ce)" d="M 92.045455 201.736181 + <path clip-path="url(#p5b84badbdb)" d="M 92.045455 201.736181 L 92.045455 179.057415 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_11"> - <path clip-path="url(#p51652a52ce)" d="M 84.588545 226.622382 + <path clip-path="url(#p5b84badbdb)" d="M 84.588545 226.622382 L 99.502364 226.622382 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_12"> - <path clip-path="url(#p51652a52ce)" d="M 84.588545 179.057415 + <path clip-path="url(#p5b84badbdb)" d="M 84.588545 179.057415 L 99.502364 179.057415 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> @@ -886,244 +886,244 @@ L 2.12132 0 L -0 -3.535534 L -2.12132 -0 z -" id="m93c95c2b91" style="stroke:#4c4c4c;stroke-linejoin:miter;"/> +" id="m81d9bdafc0" style="stroke:#4c4c4c;stroke-linejoin:miter;"/> </defs> - <g clip-path="url(#p51652a52ce)"> - <use style="fill:#4c4c4c;stroke:#4c4c4c;stroke-linejoin:miter;" x="92.045455" xlink:href="#m93c95c2b91" y="167.915124"/> - <use style="fill:#4c4c4c;stroke:#4c4c4c;stroke-linejoin:miter;" x="92.045455" xlink:href="#m93c95c2b91" y="140.50614"/> + <g clip-path="url(#p5b84badbdb)"> + <use style="fill:#4c4c4c;stroke:#4c4c4c;stroke-linejoin:miter;" x="92.045455" xlink:href="#m81d9bdafc0" y="167.915124"/> + <use style="fill:#4c4c4c;stroke:#4c4c4c;stroke-linejoin:miter;" x="92.045455" xlink:href="#m81d9bdafc0" y="140.50614"/> </g> </g> <g id="line2d_14"> - <path clip-path="url(#p51652a52ce)" d="M 122.481818 223.074719 + <path clip-path="url(#p5b84badbdb)" d="M 122.481818 223.074719 L 122.481818 234.059335 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_15"> - <path clip-path="url(#p51652a52ce)" d="M 122.481818 200.527347 + <path clip-path="url(#p5b84badbdb)" d="M 122.481818 200.527347 L 122.481818 175.483472 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_16"> - <path clip-path="url(#p51652a52ce)" d="M 115.024909 234.059335 + <path clip-path="url(#p5b84badbdb)" d="M 115.024909 234.059335 L 129.938727 234.059335 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_17"> - <path clip-path="url(#p51652a52ce)" d="M 115.024909 175.483472 + <path clip-path="url(#p5b84badbdb)" d="M 115.024909 175.483472 L 129.938727 175.483472 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_18"> - <g clip-path="url(#p51652a52ce)"> - <use style="fill:#4c4c4c;stroke:#4c4c4c;stroke-linejoin:miter;" x="122.481818" xlink:href="#m93c95c2b91" y="162.265142"/> - <use style="fill:#4c4c4c;stroke:#4c4c4c;stroke-linejoin:miter;" x="122.481818" xlink:href="#m93c95c2b91" y="163.657928"/> - <use style="fill:#4c4c4c;stroke:#4c4c4c;stroke-linejoin:miter;" x="122.481818" xlink:href="#m93c95c2b91" y="145.551706"/> + <g clip-path="url(#p5b84badbdb)"> + <use style="fill:#4c4c4c;stroke:#4c4c4c;stroke-linejoin:miter;" x="122.481818" xlink:href="#m81d9bdafc0" y="162.265142"/> + <use style="fill:#4c4c4c;stroke:#4c4c4c;stroke-linejoin:miter;" x="122.481818" xlink:href="#m81d9bdafc0" y="163.657928"/> + <use style="fill:#4c4c4c;stroke:#4c4c4c;stroke-linejoin:miter;" x="122.481818" xlink:href="#m81d9bdafc0" y="145.551706"/> </g> </g> <g id="line2d_19"> - <path clip-path="url(#p51652a52ce)" d="M 168.136364 223.074719 + <path clip-path="url(#p5b84badbdb)" d="M 168.136364 223.074719 L 168.136364 238.684437 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_20"> - <path clip-path="url(#p51652a52ce)" d="M 168.136364 204.745125 + <path clip-path="url(#p5b84badbdb)" d="M 168.136364 204.745125 L 168.136364 177.664628 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_21"> - <path clip-path="url(#p51652a52ce)" d="M 160.679455 238.684437 + <path clip-path="url(#p5b84badbdb)" d="M 160.679455 238.684437 L 175.593273 238.684437 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_22"> - <path clip-path="url(#p51652a52ce)" d="M 160.679455 177.664628 + <path clip-path="url(#p5b84badbdb)" d="M 160.679455 177.664628 L 175.593273 177.664628 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_23"> - <g clip-path="url(#p51652a52ce)"> - <use style="fill:#4c4c4c;stroke:#4c4c4c;stroke-linejoin:miter;" x="168.136364" xlink:href="#m93c95c2b91" y="148.232162"/> + <g clip-path="url(#p5b84badbdb)"> + <use style="fill:#4c4c4c;stroke:#4c4c4c;stroke-linejoin:miter;" x="168.136364" xlink:href="#m81d9bdafc0" y="148.232162"/> </g> </g> <g id="line2d_24"> - <path clip-path="url(#p51652a52ce)" d="M 198.572727 214.113584 + <path clip-path="url(#p5b84badbdb)" d="M 198.572727 214.113584 L 198.572727 221.051237 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_25"> - <path clip-path="url(#p51652a52ce)" d="M 198.572727 194.522599 + <path clip-path="url(#p5b84badbdb)" d="M 198.572727 194.522599 L 198.572727 194.010159 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_26"> - <path clip-path="url(#p51652a52ce)" d="M 191.115818 221.051237 + <path clip-path="url(#p5b84badbdb)" d="M 191.115818 221.051237 L 206.029636 221.051237 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_27"> - <path clip-path="url(#p51652a52ce)" d="M 191.115818 194.010159 + <path clip-path="url(#p5b84badbdb)" d="M 191.115818 194.010159 L 206.029636 194.010159 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_28"/> <g id="line2d_29"> - <path clip-path="url(#p51652a52ce)" d="M 244.227273 218.567872 + <path clip-path="url(#p5b84badbdb)" d="M 244.227273 218.567872 L 244.227273 245.727206 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_30"> - <path clip-path="url(#p51652a52ce)" d="M 244.227273 183.386878 + <path clip-path="url(#p5b84badbdb)" d="M 244.227273 183.386878 L 244.227273 137.378941 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_31"> - <path clip-path="url(#p51652a52ce)" d="M 236.770364 245.727206 + <path clip-path="url(#p5b84badbdb)" d="M 236.770364 245.727206 L 251.684182 245.727206 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_32"> - <path clip-path="url(#p51652a52ce)" d="M 236.770364 137.378941 + <path clip-path="url(#p5b84badbdb)" d="M 236.770364 137.378941 L 251.684182 137.378941 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_33"> - <g clip-path="url(#p51652a52ce)"> - <use style="fill:#4c4c4c;stroke:#4c4c4c;stroke-linejoin:miter;" x="244.227273" xlink:href="#m93c95c2b91" y="120.27132"/> + <g clip-path="url(#p5b84badbdb)"> + <use style="fill:#4c4c4c;stroke:#4c4c4c;stroke-linejoin:miter;" x="244.227273" xlink:href="#m81d9bdafc0" y="120.27132"/> </g> </g> <g id="line2d_34"> - <path clip-path="url(#p51652a52ce)" d="M 274.663636 215.085906 + <path clip-path="url(#p5b84badbdb)" d="M 274.663636 215.085906 L 274.663636 234.742589 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_35"> - <path clip-path="url(#p51652a52ce)" d="M 274.663636 199.528746 + <path clip-path="url(#p5b84badbdb)" d="M 274.663636 199.528746 L 274.663636 177.506954 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_36"> - <path clip-path="url(#p51652a52ce)" d="M 267.206727 234.742589 + <path clip-path="url(#p5b84badbdb)" d="M 267.206727 234.742589 L 282.120545 234.742589 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_37"> - <path clip-path="url(#p51652a52ce)" d="M 267.206727 177.506954 + <path clip-path="url(#p5b84badbdb)" d="M 267.206727 177.506954 L 282.120545 177.506954 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_38"> - <g clip-path="url(#p51652a52ce)"> - <use style="fill:#4c4c4c;stroke:#4c4c4c;stroke-linejoin:miter;" x="274.663636" xlink:href="#m93c95c2b91" y="150.203087"/> - <use style="fill:#4c4c4c;stroke:#4c4c4c;stroke-linejoin:miter;" x="274.663636" xlink:href="#m93c95c2b91" y="171.620461"/> - <use style="fill:#4c4c4c;stroke:#4c4c4c;stroke-linejoin:miter;" x="274.663636" xlink:href="#m93c95c2b91" y="126.946183"/> - <use style="fill:#4c4c4c;stroke:#4c4c4c;stroke-linejoin:miter;" x="274.663636" xlink:href="#m93c95c2b91" y="126.788509"/> - <use style="fill:#4c4c4c;stroke:#4c4c4c;stroke-linejoin:miter;" x="274.663636" xlink:href="#m93c95c2b91" y="159.637243"/> + <g clip-path="url(#p5b84badbdb)"> + <use style="fill:#4c4c4c;stroke:#4c4c4c;stroke-linejoin:miter;" x="274.663636" xlink:href="#m81d9bdafc0" y="150.203087"/> + <use style="fill:#4c4c4c;stroke:#4c4c4c;stroke-linejoin:miter;" x="274.663636" xlink:href="#m81d9bdafc0" y="171.620461"/> + <use style="fill:#4c4c4c;stroke:#4c4c4c;stroke-linejoin:miter;" x="274.663636" xlink:href="#m81d9bdafc0" y="126.946183"/> + <use style="fill:#4c4c4c;stroke:#4c4c4c;stroke-linejoin:miter;" x="274.663636" xlink:href="#m81d9bdafc0" y="126.788509"/> + <use style="fill:#4c4c4c;stroke:#4c4c4c;stroke-linejoin:miter;" x="274.663636" xlink:href="#m81d9bdafc0" y="159.637243"/> </g> </g> <g id="line2d_39"> - <path clip-path="url(#p51652a52ce)" d="M 320.318182 208.686973 + <path clip-path="url(#p5b84badbdb)" d="M 320.318182 208.686973 L 320.318182 234.742589 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_40"> - <path clip-path="url(#p51652a52ce)" d="M 320.318182 168.716633 + <path clip-path="url(#p5b84badbdb)" d="M 320.318182 168.716633 L 320.318182 134.619647 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_41"> - <path clip-path="url(#p51652a52ce)" d="M 312.861273 234.742589 + <path clip-path="url(#p5b84badbdb)" d="M 312.861273 234.742589 L 327.775091 234.742589 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_42"> - <path clip-path="url(#p51652a52ce)" d="M 312.861273 134.619647 + <path clip-path="url(#p5b84badbdb)" d="M 312.861273 134.619647 L 327.775091 134.619647 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_43"/> <g id="line2d_44"> - <path clip-path="url(#p51652a52ce)" d="M 350.754545 214.954511 + <path clip-path="url(#p5b84badbdb)" d="M 350.754545 214.954511 L 350.754545 230.748183 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_45"> - <path clip-path="url(#p51652a52ce)" d="M 350.754545 188.097386 + <path clip-path="url(#p5b84badbdb)" d="M 350.754545 188.097386 L 350.754545 153.75075 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_46"> - <path clip-path="url(#p51652a52ce)" d="M 343.297636 230.748183 + <path clip-path="url(#p5b84badbdb)" d="M 343.297636 230.748183 L 358.211455 230.748183 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_47"> - <path clip-path="url(#p51652a52ce)" d="M 343.297636 153.75075 + <path clip-path="url(#p5b84badbdb)" d="M 343.297636 153.75075 L 358.211455 153.75075 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_48"> - <g clip-path="url(#p51652a52ce)"> - <use style="fill:#4c4c4c;stroke:#4c4c4c;stroke-linejoin:miter;" x="350.754545" xlink:href="#m93c95c2b91" y="127.208973"/> + <g clip-path="url(#p5b84badbdb)"> + <use style="fill:#4c4c4c;stroke:#4c4c4c;stroke-linejoin:miter;" x="350.754545" xlink:href="#m81d9bdafc0" y="127.208973"/> </g> </g> <g id="line2d_49"> - <path clip-path="url(#p51652a52ce)" d="M 244.227273 91.084363 + <path clip-path="url(#p5b84badbdb)" d="M 244.227273 91.084363 L 244.227273 84.18429 L 274.663636 84.18429 L 274.663636 91.084363 " style="fill:none;stroke:#333333;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_50"> - <path clip-path="url(#p51652a52ce)" d="M 122.481818 116.364749 + <path clip-path="url(#p5b84badbdb)" d="M 122.481818 116.364749 L 122.481818 109.464676 L 198.572727 109.464676 L 198.572727 116.364749 " style="fill:none;stroke:#333333;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_51"> - <path clip-path="url(#p51652a52ce)" d="M 92.045455 61.276864 + <path clip-path="url(#p5b84badbdb)" d="M 92.045455 61.276864 L 92.045455 54.37679 L 350.754545 54.37679 L 350.754545 61.276864 " style="fill:none;stroke:#333333;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_52"> - <path clip-path="url(#p51652a52ce)" d="M 77.131636 210.513363 + <path clip-path="url(#p5b84badbdb)" d="M 77.131636 210.513363 L 106.959273 210.513363 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_53"> - <path clip-path="url(#p51652a52ce)" d="M 107.568 211.87987 + <path clip-path="url(#p5b84badbdb)" d="M 107.568 211.87987 L 137.395636 211.87987 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_54"> - <path clip-path="url(#p51652a52ce)" d="M 153.222545 218.528454 + <path clip-path="url(#p5b84badbdb)" d="M 153.222545 218.528454 L 183.050182 218.528454 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_55"> - <path clip-path="url(#p51652a52ce)" d="M 183.658909 203.247223 + <path clip-path="url(#p5b84badbdb)" d="M 183.658909 203.247223 L 213.486545 203.247223 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_56"> - <path clip-path="url(#p51652a52ce)" d="M 229.313455 200.212 + <path clip-path="url(#p5b84badbdb)" d="M 229.313455 200.212 L 259.141091 200.212 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_57"> - <path clip-path="url(#p51652a52ce)" d="M 259.749818 206.965699 + <path clip-path="url(#p5b84badbdb)" d="M 259.749818 206.965699 L 289.577455 206.965699 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_58"> - <path clip-path="url(#p51652a52ce)" d="M 305.404364 193.090394 + <path clip-path="url(#p5b84badbdb)" d="M 305.404364 193.090394 L 335.232 193.090394 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_59"> - <path clip-path="url(#p51652a52ce)" d="M 335.840727 205.362681 + <path clip-path="url(#p5b84badbdb)" d="M 335.840727 205.362681 L 365.668364 205.362681 " style="fill:none;stroke:#4c4c4c;stroke-linecap:round;stroke-width:1.5;"/> </g> @@ -1426,7 +1426,7 @@ z </g> </g> <defs> - <clipPath id="p51652a52ce"> + <clipPath id="p5b84badbdb"> <rect height="217.44" width="334.8" x="54" y="34.56"/> </clipPath> </defs> diff --git a/usage/flu_dataset_log_scale_in_axes.svg b/usage/flu_dataset_log_scale_in_axes.svg index f5e13a3..3cb7293 100644 --- a/usage/flu_dataset_log_scale_in_axes.svg +++ b/usage/flu_dataset_log_scale_in_axes.svg @@ -510,7 +510,7 @@ z <g id="matplotlib.axis_2"> <g id="ytick_1"> <g id="line2d_1"> - <path clip-path="url(#p37bf755c76)" d="M 108 355.511801 + <path clip-path="url(#p19cc01c258)" d="M 108 355.511801 L 777.6 355.511801 " style="fill:none;stroke:#cccccc;stroke-linecap:round;"/> </g> @@ -584,7 +584,7 @@ z </g> <g id="ytick_2"> <g id="line2d_2"> - <path clip-path="url(#p37bf755c76)" d="M 108 260.424716 + <path clip-path="url(#p19cc01c258)" d="M 108 260.424716 L 777.6 260.424716 " style="fill:none;stroke:#cccccc;stroke-linecap:round;"/> </g> @@ -619,7 +619,7 @@ z </g> <g id="ytick_3"> <g id="line2d_3"> - <path clip-path="url(#p37bf755c76)" d="M 108 165.337631 + <path clip-path="url(#p19cc01c258)" d="M 108 165.337631 L 777.6 165.337631 " style="fill:none;stroke:#cccccc;stroke-linecap:round;"/> </g> @@ -635,7 +635,7 @@ L 777.6 165.337631 </g> <g id="ytick_4"> <g id="line2d_4"> - <path clip-path="url(#p37bf755c76)" d="M 108 70.250546 + <path clip-path="url(#p19cc01c258)" d="M 108 70.250546 L 777.6 70.250546 " style="fill:none;stroke:#cccccc;stroke-linecap:round;"/> </g> @@ -790,21 +790,21 @@ z <g id="PathCollection_1"/> <g id="PathCollection_2"/> <g id="line2d_5"> - <path clip-path="url(#p37bf755c76)" d="M 398.16 122.800151 + <path clip-path="url(#p19cc01c258)" d="M 398.16 122.800151 L 398.16 117.015595 L 487.44 117.015595 L 487.44 122.800151 " style="fill:none;stroke:#333333;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_6"> - <path clip-path="url(#p37bf755c76)" d="M 174.96 84.565272 + <path clip-path="url(#p19cc01c258)" d="M 174.96 84.565272 L 174.96 78.780716 L 264.24 78.780716 L 264.24 84.565272 " style="fill:none;stroke:#333333;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_7"> - <path clip-path="url(#p37bf755c76)" d="M 621.36 136.464802 + <path clip-path="url(#p19cc01c258)" d="M 621.36 136.464802 L 621.36 130.680247 L 710.64 130.680247 L 710.64 136.464802 @@ -842,589 +842,589 @@ C -2.236584 -1.29895 -2.5 -0.663008 -2.5 0 C -2.5 0.663008 -2.236584 1.29895 -1.767767 1.767767 C -1.29895 2.236584 -0.663008 2.5 0 2.5 z -" id="C0_0_eaabaca787"/> +" id="C0_0_e64610c732"/> </defs> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_eaabaca787" y="364.853283"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_e64610c732" y="364.853283"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_eaabaca787" y="336.244734"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_e64610c732" y="336.244734"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_eaabaca787" y="327.330163"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_e64610c732" y="327.330163"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="169.710003" xlink:href="#C0_0_eaabaca787" y="327.325352"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="169.710003" xlink:href="#C0_0_e64610c732" y="327.325352"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_eaabaca787" y="321.598017"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_e64610c732" y="321.598017"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="170.818467" xlink:href="#C0_0_eaabaca787" y="318.873123"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="170.818467" xlink:href="#C0_0_e64610c732" y="318.873123"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="179.092496" xlink:href="#C0_0_eaabaca787" y="318.863354"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="179.092496" xlink:href="#C0_0_e64610c732" y="318.863354"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_eaabaca787" y="308.233739"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_e64610c732" y="308.233739"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="169.71" xlink:href="#C0_0_eaabaca787" y="308.23363"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="169.71" xlink:href="#C0_0_e64610c732" y="308.23363"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="180.20999" xlink:href="#C0_0_eaabaca787" y="308.225156"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="180.20999" xlink:href="#C0_0_e64610c732" y="308.225156"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="164.460017" xlink:href="#C0_0_eaabaca787" y="308.222402"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="164.460017" xlink:href="#C0_0_e64610c732" y="308.222402"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="185.459988" xlink:href="#C0_0_eaabaca787" y="308.22121"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="185.459988" xlink:href="#C0_0_e64610c732" y="308.22121"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="159.210017" xlink:href="#C0_0_eaabaca787" y="308.220715"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="159.210017" xlink:href="#C0_0_e64610c732" y="308.220715"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="190.709988" xlink:href="#C0_0_eaabaca787" y="308.220501"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="190.709988" xlink:href="#C0_0_e64610c732" y="308.220501"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="153.960017" xlink:href="#C0_0_eaabaca787" y="308.219795"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="153.960017" xlink:href="#C0_0_e64610c732" y="308.219795"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_eaabaca787" y="297.14141"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_e64610c732" y="297.14141"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="170.375862" xlink:href="#C0_0_eaabaca787" y="294.980284"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="170.375862" xlink:href="#C0_0_e64610c732" y="294.980284"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="179.537229" xlink:href="#C0_0_eaabaca787" y="294.969863"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="179.537229" xlink:href="#C0_0_e64610c732" y="294.969863"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="184.787228" xlink:href="#C0_0_eaabaca787" y="294.967678"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="184.787228" xlink:href="#C0_0_e64610c732" y="294.967678"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="165.299683" xlink:href="#C0_0_eaabaca787" y="293.848819"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="165.299683" xlink:href="#C0_0_e64610c732" y="293.848819"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="189.865013" xlink:href="#C0_0_eaabaca787" y="293.841365"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="189.865013" xlink:href="#C0_0_e64610c732" y="293.841365"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="160.049695" xlink:href="#C0_0_eaabaca787" y="293.8393"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="160.049695" xlink:href="#C0_0_e64610c732" y="293.8393"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="195.115012" xlink:href="#C0_0_eaabaca787" y="293.839283"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="195.115012" xlink:href="#C0_0_e64610c732" y="293.839283"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="154.79974" xlink:href="#C0_0_eaabaca787" y="293.820936"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="154.79974" xlink:href="#C0_0_e64610c732" y="293.820936"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="200.26563" xlink:href="#C0_0_eaabaca787" y="292.980658"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="200.26563" xlink:href="#C0_0_e64610c732" y="292.980658"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_eaabaca787" y="290.253898"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_e64610c732" y="290.253898"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="169.710014" xlink:href="#C0_0_eaabaca787" y="290.24361"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="169.710014" xlink:href="#C0_0_e64610c732" y="290.24361"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="180.209979" xlink:href="#C0_0_eaabaca787" y="290.241316"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="180.209979" xlink:href="#C0_0_e64610c732" y="290.241316"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="164.68005" xlink:href="#C0_0_eaabaca787" y="288.973439"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="164.68005" xlink:href="#C0_0_e64610c732" y="288.973439"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="184.492216" xlink:href="#C0_0_eaabaca787" y="287.67623"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="184.492216" xlink:href="#C0_0_e64610c732" y="287.67623"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_eaabaca787" y="281.957343"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_e64610c732" y="281.957343"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="169.71" xlink:href="#C0_0_eaabaca787" y="281.957326"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="169.71" xlink:href="#C0_0_e64610c732" y="281.957326"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="179.412672" xlink:href="#C0_0_eaabaca787" y="279.608336"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="179.412672" xlink:href="#C0_0_e64610c732" y="279.608336"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="165.260883" xlink:href="#C0_0_eaabaca787" y="279.603519"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="165.260883" xlink:href="#C0_0_e64610c732" y="279.603519"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="184.662667" xlink:href="#C0_0_eaabaca787" y="279.602641"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="184.662667" xlink:href="#C0_0_e64610c732" y="279.602641"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="160.010884" xlink:href="#C0_0_eaabaca787" y="279.602216"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="160.010884" xlink:href="#C0_0_e64610c732" y="279.602216"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="189.912663" xlink:href="#C0_0_eaabaca787" y="279.596838"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="189.912663" xlink:href="#C0_0_e64610c732" y="279.596838"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="154.760903" xlink:href="#C0_0_eaabaca787" y="279.590004"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="154.760903" xlink:href="#C0_0_e64610c732" y="279.590004"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_eaabaca787" y="276.227583"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_e64610c732" y="276.227583"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="169.777725" xlink:href="#C0_0_eaabaca787" y="275.517702"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="169.777725" xlink:href="#C0_0_e64610c732" y="275.517702"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="179.097481" xlink:href="#C0_0_eaabaca787" y="273.498302"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="179.097481" xlink:href="#C0_0_e64610c732" y="273.498302"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="184.34748" xlink:href="#C0_0_eaabaca787" y="273.494529"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="184.34748" xlink:href="#C0_0_e64610c732" y="273.494529"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="174.361283" xlink:href="#C0_0_eaabaca787" y="271.58529"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="174.361283" xlink:href="#C0_0_e64610c732" y="271.58529"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="167.35685" xlink:href="#C0_0_eaabaca787" y="271.583392"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="167.35685" xlink:href="#C0_0_e64610c732" y="271.583392"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="162.106856" xlink:href="#C0_0_eaabaca787" y="271.577054"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="162.106856" xlink:href="#C0_0_e64610c732" y="271.577054"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="187.358661" xlink:href="#C0_0_eaabaca787" y="269.862489"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="187.358661" xlink:href="#C0_0_e64610c732" y="269.862489"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_eaabaca787" y="265.231804"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_e64610c732" y="265.231804"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="169.710005" xlink:href="#C0_0_eaabaca787" y="265.225982"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="169.710005" xlink:href="#C0_0_e64610c732" y="265.225982"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="180.209976" xlink:href="#C0_0_eaabaca787" y="265.218325"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="180.209976" xlink:href="#C0_0_e64610c732" y="265.218325"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="185.459976" xlink:href="#C0_0_eaabaca787" y="265.217742"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="185.459976" xlink:href="#C0_0_e64610c732" y="265.217742"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="164.460021" xlink:href="#C0_0_eaabaca787" y="265.214756"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="164.460021" xlink:href="#C0_0_e64610c732" y="265.214756"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="190.709972" xlink:href="#C0_0_eaabaca787" y="265.212449"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="190.709972" xlink:href="#C0_0_e64610c732" y="265.212449"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="159.210023" xlink:href="#C0_0_eaabaca787" y="265.211275"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="159.210023" xlink:href="#C0_0_e64610c732" y="265.211275"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="195.959969" xlink:href="#C0_0_eaabaca787" y="265.207677"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="195.959969" xlink:href="#C0_0_e64610c732" y="265.207677"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="154.755889" xlink:href="#C0_0_eaabaca787" y="262.864246"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="154.755889" xlink:href="#C0_0_e64610c732" y="262.864246"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="200.409433" xlink:href="#C0_0_eaabaca787" y="262.854339"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="200.409433" xlink:href="#C0_0_e64610c732" y="262.854339"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="149.505906" xlink:href="#C0_0_eaabaca787" y="262.852999"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="149.505906" xlink:href="#C0_0_e64610c732" y="262.852999"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="174.033579" xlink:href="#C0_0_eaabaca787" y="260.867548"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="174.033579" xlink:href="#C0_0_e64610c732" y="260.867548"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="179.247123" xlink:href="#C0_0_eaabaca787" y="260.345939"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="179.247123" xlink:href="#C0_0_e64610c732" y="260.345939"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="169.047919" xlink:href="#C0_0_eaabaca787" y="259.478366"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="169.047919" xlink:href="#C0_0_e64610c732" y="259.478366"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="184.270248" xlink:href="#C0_0_eaabaca787" y="259.056609"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="184.270248" xlink:href="#C0_0_e64610c732" y="259.056609"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_eaabaca787" y="255.133958"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_e64610c732" y="255.133958"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="169.771727" xlink:href="#C0_0_eaabaca787" y="254.45605"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="169.771727" xlink:href="#C0_0_e64610c732" y="254.45605"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="179.754578" xlink:href="#C0_0_eaabaca787" y="253.327647"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="179.754578" xlink:href="#C0_0_e64610c732" y="253.327647"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="165.113941" xlink:href="#C0_0_eaabaca787" y="252.410323"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="165.113941" xlink:href="#C0_0_e64610c732" y="252.410323"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="173.106762" xlink:href="#C0_0_eaabaca787" y="250.985557"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="173.106762" xlink:href="#C0_0_e64610c732" y="250.985557"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="184.212226" xlink:href="#C0_0_eaabaca787" y="250.985381"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="184.212226" xlink:href="#C0_0_e64610c732" y="250.985381"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="189.462222" xlink:href="#C0_0_eaabaca787" y="250.979552"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="189.462222" xlink:href="#C0_0_e64610c732" y="250.979552"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="160.147123" xlink:href="#C0_0_eaabaca787" y="250.973806"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="160.147123" xlink:href="#C0_0_e64610c732" y="250.973806"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="194.712198" xlink:href="#C0_0_eaabaca787" y="250.966056"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="194.712198" xlink:href="#C0_0_e64610c732" y="250.966056"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="154.928967" xlink:href="#C0_0_eaabaca787" y="250.486198"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="154.928967" xlink:href="#C0_0_e64610c732" y="250.486198"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="177.429427" xlink:href="#C0_0_eaabaca787" y="248.469301"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="177.429427" xlink:href="#C0_0_e64610c732" y="248.469301"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="168.785553" xlink:href="#C0_0_eaabaca787" y="248.467519"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="168.785553" xlink:href="#C0_0_e64610c732" y="248.467519"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="199.049167" xlink:href="#C0_0_eaabaca787" y="248.467419"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="199.049167" xlink:href="#C0_0_e64610c732" y="248.467419"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="163.845692" xlink:href="#C0_0_eaabaca787" y="246.966166"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="163.845692" xlink:href="#C0_0_e64610c732" y="246.966166"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="174.333039" xlink:href="#C0_0_eaabaca787" y="244.888724"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="174.333039" xlink:href="#C0_0_e64610c732" y="244.888724"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="179.059296" xlink:href="#C0_0_eaabaca787" y="242.958253"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="179.059296" xlink:href="#C0_0_e64610c732" y="242.958253"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="169.870691" xlink:href="#C0_0_eaabaca787" y="242.552852"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="169.870691" xlink:href="#C0_0_e64610c732" y="242.552852"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="184.251342" xlink:href="#C0_0_eaabaca787" y="242.301269"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="184.251342" xlink:href="#C0_0_e64610c732" y="242.301269"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="164.789456" xlink:href="#C0_0_eaabaca787" y="241.437691"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="164.789456" xlink:href="#C0_0_e64610c732" y="241.437691"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_eaabaca787" y="239.253126"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_e64610c732" y="239.253126"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="180.95004" xlink:href="#C0_0_eaabaca787" y="238.821941"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="180.95004" xlink:href="#C0_0_e64610c732" y="238.821941"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="170.032046" xlink:href="#C0_0_eaabaca787" y="237.72412"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="170.032046" xlink:href="#C0_0_e64610c732" y="237.72412"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="164.953372" xlink:href="#C0_0_eaabaca787" y="236.60067"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="164.953372" xlink:href="#C0_0_e64610c732" y="236.60067"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="185.488893" xlink:href="#C0_0_eaabaca787" y="236.593682"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="185.488893" xlink:href="#C0_0_e64610c732" y="236.593682"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="159.703389" xlink:href="#C0_0_eaabaca787" y="236.58931"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="159.703389" xlink:href="#C0_0_e64610c732" y="236.58931"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="190.738888" xlink:href="#C0_0_eaabaca787" y="236.587651"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="190.738888" xlink:href="#C0_0_e64610c732" y="236.587651"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_eaabaca787" y="234.220606"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_e64610c732" y="234.220606"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="180.179646" xlink:href="#C0_0_eaabaca787" y="233.744512"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="180.179646" xlink:href="#C0_0_e64610c732" y="233.744512"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="170.620825" xlink:href="#C0_0_eaabaca787" y="231.7247"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="170.620825" xlink:href="#C0_0_e64610c732" y="231.7247"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_eaabaca787" y="228.41365"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_e64610c732" y="228.41365"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="180.065213" xlink:href="#C0_0_eaabaca787" y="227.379546"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="180.065213" xlink:href="#C0_0_e64610c732" y="227.379546"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="172.073166" xlink:href="#C0_0_eaabaca787" y="224.710299"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="172.073166" xlink:href="#C0_0_e64610c732" y="224.710299"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="176.517526" xlink:href="#C0_0_eaabaca787" y="222.350092"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="176.517526" xlink:href="#C0_0_e64610c732" y="222.350092"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="172.175421" xlink:href="#C0_0_eaabaca787" y="219.857825"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="172.175421" xlink:href="#C0_0_e64610c732" y="219.857825"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="180.849524" xlink:href="#C0_0_eaabaca787" y="219.845311"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="180.849524" xlink:href="#C0_0_e64610c732" y="219.845311"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="166.925442" xlink:href="#C0_0_eaabaca787" y="219.845244"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="166.925442" xlink:href="#C0_0_e64610c732" y="219.845244"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="176.402495" xlink:href="#C0_0_eaabaca787" y="217.488692"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="176.402495" xlink:href="#C0_0_e64610c732" y="217.488692"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="171.245614" xlink:href="#C0_0_eaabaca787" y="215.494083"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="171.245614" xlink:href="#C0_0_e64610c732" y="215.494083"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="180.086055" xlink:href="#C0_0_eaabaca787" y="214.329404"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="180.086055" xlink:href="#C0_0_e64610c732" y="214.329404"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="166.184811" xlink:href="#C0_0_eaabaca787" y="214.314513"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="166.184811" xlink:href="#C0_0_e64610c732" y="214.314513"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="185.241932" xlink:href="#C0_0_eaabaca787" y="213.493596"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="185.241932" xlink:href="#C0_0_e64610c732" y="213.493596"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="161.026619" xlink:href="#C0_0_eaabaca787" y="213.488957"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="161.026619" xlink:href="#C0_0_e64610c732" y="213.488957"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="190.491918" xlink:href="#C0_0_eaabaca787" y="213.483422"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="190.491918" xlink:href="#C0_0_e64610c732" y="213.483422"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="155.776623" xlink:href="#C0_0_eaabaca787" y="213.483098"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="155.776623" xlink:href="#C0_0_e64610c732" y="213.483098"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="195.741916" xlink:href="#C0_0_eaabaca787" y="213.479362"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="195.741916" xlink:href="#C0_0_e64610c732" y="213.479362"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="150.526626" xlink:href="#C0_0_eaabaca787" y="213.478317"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="150.526626" xlink:href="#C0_0_e64610c732" y="213.478317"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="200.991912" xlink:href="#C0_0_eaabaca787" y="213.473832"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="200.991912" xlink:href="#C0_0_e64610c732" y="213.473832"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="145.294046" xlink:href="#C0_0_eaabaca787" y="213.117429"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="145.294046" xlink:href="#C0_0_e64610c732" y="213.117429"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="206.18685" xlink:href="#C0_0_eaabaca787" y="212.833363"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="206.18685" xlink:href="#C0_0_e64610c732" y="212.833363"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_eaabaca787" y="209.897757"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_e64610c732" y="209.897757"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="180.14391" xlink:href="#C0_0_eaabaca787" y="209.196444"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="180.14391" xlink:href="#C0_0_e64610c732" y="209.196444"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="169.933723" xlink:href="#C0_0_eaabaca787" y="208.617217"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="169.933723" xlink:href="#C0_0_e64610c732" y="208.617217"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="164.740422" xlink:href="#C0_0_eaabaca787" y="207.967349"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="164.740422" xlink:href="#C0_0_e64610c732" y="207.967349"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="185.187474" xlink:href="#C0_0_eaabaca787" y="207.965338"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="185.187474" xlink:href="#C0_0_e64610c732" y="207.965338"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="159.606412" xlink:href="#C0_0_eaabaca787" y="207.040491"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="159.606412" xlink:href="#C0_0_e64610c732" y="207.040491"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="190.321587" xlink:href="#C0_0_eaabaca787" y="207.038886"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="190.321587" xlink:href="#C0_0_e64610c732" y="207.038886"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="154.356418" xlink:href="#C0_0_eaabaca787" y="207.033452"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="154.356418" xlink:href="#C0_0_e64610c732" y="207.033452"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="173.646709" xlink:href="#C0_0_eaabaca787" y="205.604889"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="173.646709" xlink:href="#C0_0_e64610c732" y="205.604889"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="195.05219" xlink:href="#C0_0_eaabaca787" y="205.116025"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="195.05219" xlink:href="#C0_0_e64610c732" y="205.116025"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="149.623563" xlink:href="#C0_0_eaabaca787" y="205.114548"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="149.623563" xlink:href="#C0_0_e64610c732" y="205.114548"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="178.206759" xlink:href="#C0_0_eaabaca787" y="203.407715"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="178.206759" xlink:href="#C0_0_e64610c732" y="203.407715"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="169.31113" xlink:href="#C0_0_eaabaca787" y="203.104531"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="169.31113" xlink:href="#C0_0_e64610c732" y="203.104531"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="183.395232" xlink:href="#C0_0_eaabaca787" y="202.730895"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="183.395232" xlink:href="#C0_0_e64610c732" y="202.730895"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="164.147683" xlink:href="#C0_0_eaabaca787" y="202.302745"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="164.147683" xlink:href="#C0_0_e64610c732" y="202.302745"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="188.618949" xlink:href="#C0_0_eaabaca787" y="202.287795"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="188.618949" xlink:href="#C0_0_e64610c732" y="202.287795"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="158.905458" xlink:href="#C0_0_eaabaca787" y="202.061532"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="158.905458" xlink:href="#C0_0_e64610c732" y="202.061532"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="153.660371" xlink:href="#C0_0_eaabaca787" y="201.869769"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="153.660371" xlink:href="#C0_0_e64610c732" y="201.869769"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="198.626342" xlink:href="#C0_0_eaabaca787" y="201.86834"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="198.626342" xlink:href="#C0_0_e64610c732" y="201.86834"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="146.047922" xlink:href="#C0_0_eaabaca787" y="201.868033"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="146.047922" xlink:href="#C0_0_e64610c732" y="201.868033"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="203.848882" xlink:href="#C0_0_eaabaca787" y="201.41545"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="203.848882" xlink:href="#C0_0_e64610c732" y="201.41545"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="174.677654" xlink:href="#C0_0_eaabaca787" y="200.125082"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="174.677654" xlink:href="#C0_0_e64610c732" y="200.125082"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="193.150493" xlink:href="#C0_0_eaabaca787" y="200.048951"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="193.150493" xlink:href="#C0_0_e64610c732" y="200.048951"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="141.407376" xlink:href="#C0_0_eaabaca787" y="199.794549"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="141.407376" xlink:href="#C0_0_e64610c732" y="199.794549"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="208.734953" xlink:href="#C0_0_eaabaca787" y="199.793404"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="208.734953" xlink:href="#C0_0_e64610c732" y="199.793404"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="136.159252" xlink:href="#C0_0_eaabaca787" y="199.676046"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="136.159252" xlink:href="#C0_0_e64610c732" y="199.676046"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="213.983102" xlink:href="#C0_0_eaabaca787" y="199.675669"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="213.983102" xlink:href="#C0_0_e64610c732" y="199.675669"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="131.2128" xlink:href="#C0_0_eaabaca787" y="199.52131"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="131.2128" xlink:href="#C0_0_e64610c732" y="199.52131"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="179.993462" xlink:href="#C0_0_eaabaca787" y="199.238546"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="179.993462" xlink:href="#C0_0_e64610c732" y="199.238546"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="169.776832" xlink:href="#C0_0_eaabaca787" y="198.535095"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="169.776832" xlink:href="#C0_0_e64610c732" y="198.535095"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="184.865302" xlink:href="#C0_0_eaabaca787" y="197.586251"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="184.865302" xlink:href="#C0_0_e64610c732" y="197.586251"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="165.173428" xlink:href="#C0_0_eaabaca787" y="196.403376"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="165.173428" xlink:href="#C0_0_e64610c732" y="196.403376"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="189.920554" xlink:href="#C0_0_eaabaca787" y="196.389823"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="189.920554" xlink:href="#C0_0_e64610c732" y="196.389823"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="159.937239" xlink:href="#C0_0_eaabaca787" y="196.081986"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="159.937239" xlink:href="#C0_0_e64610c732" y="196.081986"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_eaabaca787" y="195.166563"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_e64610c732" y="195.166563"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="194.864027" xlink:href="#C0_0_eaabaca787" y="194.896978"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="194.864027" xlink:href="#C0_0_e64610c732" y="194.896978"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="180.333223" xlink:href="#C0_0_eaabaca787" y="194.814008"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="180.333223" xlink:href="#C0_0_e64610c732" y="194.814008"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="169.935692" xlink:href="#C0_0_eaabaca787" y="193.880522"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="169.935692" xlink:href="#C0_0_e64610c732" y="193.880522"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="155.480531" xlink:href="#C0_0_eaabaca787" y="193.738446"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="155.480531" xlink:href="#C0_0_e64610c732" y="193.738446"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="185.000983" xlink:href="#C0_0_eaabaca787" y="192.784561"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="185.000983" xlink:href="#C0_0_e64610c732" y="192.784561"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="177.252999" xlink:href="#C0_0_eaabaca787" y="191.223501"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="177.252999" xlink:href="#C0_0_e64610c732" y="191.223501"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="165.737133" xlink:href="#C0_0_eaabaca787" y="191.218602"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="165.737133" xlink:href="#C0_0_e64610c732" y="191.218602"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="160.487134" xlink:href="#C0_0_eaabaca787" y="191.215615"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="160.487134" xlink:href="#C0_0_e64610c732" y="191.215615"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="172.478476" xlink:href="#C0_0_eaabaca787" y="189.379688"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="172.478476" xlink:href="#C0_0_e64610c732" y="189.379688"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="181.695518" xlink:href="#C0_0_eaabaca787" y="188.860824"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="181.695518" xlink:href="#C0_0_e64610c732" y="188.860824"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="187.445791" xlink:href="#C0_0_eaabaca787" y="188.860821"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="187.445791" xlink:href="#C0_0_e64610c732" y="188.860821"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="192.576512" xlink:href="#C0_0_eaabaca787" y="187.921065"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="192.576512" xlink:href="#C0_0_e64610c732" y="187.921065"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="156.980381" xlink:href="#C0_0_eaabaca787" y="187.915941"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="156.980381" xlink:href="#C0_0_e64610c732" y="187.915941"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="175.410374" xlink:href="#C0_0_eaabaca787" y="185.701675"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="175.410374" xlink:href="#C0_0_e64610c732" y="185.701675"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="170.255553" xlink:href="#C0_0_eaabaca787" y="184.861231"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="170.255553" xlink:href="#C0_0_e64610c732" y="184.861231"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="180.46705" xlink:href="#C0_0_eaabaca787" y="184.509548"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="180.46705" xlink:href="#C0_0_e64610c732" y="184.509548"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="165.063223" xlink:href="#C0_0_eaabaca787" y="184.205848"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="165.063223" xlink:href="#C0_0_e64610c732" y="184.205848"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="184.798294" xlink:href="#C0_0_eaabaca787" y="182.003838"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="184.798294" xlink:href="#C0_0_e64610c732" y="182.003838"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_eaabaca787" y="179.346155"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_e64610c732" y="179.346155"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="170.513744" xlink:href="#C0_0_eaabaca787" y="176.988497"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="170.513744" xlink:href="#C0_0_e64610c732" y="176.988497"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="179.286465" xlink:href="#C0_0_eaabaca787" y="176.834562"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="179.286465" xlink:href="#C0_0_e64610c732" y="176.834562"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_eaabaca787" y="174.484336"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_e64610c732" y="174.484336"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="183.736877" xlink:href="#C0_0_eaabaca787" y="174.482503"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="183.736877" xlink:href="#C0_0_e64610c732" y="174.482503"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="166.187614" xlink:href="#C0_0_eaabaca787" y="174.476492"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="166.187614" xlink:href="#C0_0_e64610c732" y="174.476492"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="161.024387" xlink:href="#C0_0_eaabaca787" y="173.673696"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="161.024387" xlink:href="#C0_0_e64610c732" y="173.673696"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="188.266949" xlink:href="#C0_0_eaabaca787" y="172.241536"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="188.266949" xlink:href="#C0_0_e64610c732" y="172.241536"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="193.516947" xlink:href="#C0_0_eaabaca787" y="172.237632"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="193.516947" xlink:href="#C0_0_e64610c732" y="172.237632"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="173.978286" xlink:href="#C0_0_eaabaca787" y="170.128711"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="173.978286" xlink:href="#C0_0_e64610c732" y="170.128711"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="179.103612" xlink:href="#C0_0_eaabaca787" y="169.168184"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="179.103612" xlink:href="#C0_0_e64610c732" y="169.168184"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="169.292124" xlink:href="#C0_0_eaabaca787" y="168.12974"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="169.292124" xlink:href="#C0_0_e64610c732" y="168.12974"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="184.203128" xlink:href="#C0_0_eaabaca787" y="168.114221"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="184.203128" xlink:href="#C0_0_e64610c732" y="168.114221"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="164.059802" xlink:href="#C0_0_eaabaca787" y="167.766194"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="164.059802" xlink:href="#C0_0_e64610c732" y="167.766194"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="188.933503" xlink:href="#C0_0_eaabaca787" y="166.190959"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="188.933503" xlink:href="#C0_0_e64610c732" y="166.190959"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_eaabaca787" y="164.358533"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_e64610c732" y="164.358533"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="180.209994" xlink:href="#C0_0_eaabaca787" y="164.351989"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="180.209994" xlink:href="#C0_0_e64610c732" y="164.351989"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="170.139319" xlink:href="#C0_0_eaabaca787" y="162.602472"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="170.139319" xlink:href="#C0_0_e64610c732" y="162.602472"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="164.889323" xlink:href="#C0_0_eaabaca787" y="162.597059"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="164.889323" xlink:href="#C0_0_e64610c732" y="162.597059"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_eaabaca787" y="159.747356"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_e64610c732" y="159.747356"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="180.183647" xlink:href="#C0_0_eaabaca787" y="159.303658"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="180.183647" xlink:href="#C0_0_e64610c732" y="159.303658"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="185.433605" xlink:href="#C0_0_eaabaca787" y="159.286053"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="185.433605" xlink:href="#C0_0_e64610c732" y="159.286053"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="170.283045" xlink:href="#C0_0_eaabaca787" y="157.733064"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="170.283045" xlink:href="#C0_0_e64610c732" y="157.733064"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="165.344636" xlink:href="#C0_0_eaabaca787" y="156.228311"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="165.344636" xlink:href="#C0_0_e64610c732" y="156.228311"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_eaabaca787" y="154.565069"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_e64610c732" y="154.565069"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="180.021095" xlink:href="#C0_0_eaabaca787" y="153.386395"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="180.021095" xlink:href="#C0_0_e64610c732" y="153.386395"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="170.480743" xlink:href="#C0_0_eaabaca787" y="152.252394"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="170.480743" xlink:href="#C0_0_e64610c732" y="152.252394"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="185.086447" xlink:href="#C0_0_eaabaca787" y="152.220836"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="185.086447" xlink:href="#C0_0_e64610c732" y="152.220836"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_eaabaca787" y="149.452862"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_e64610c732" y="149.452862"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="166.64423" xlink:href="#C0_0_eaabaca787" y="149.225728"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="166.64423" xlink:href="#C0_0_e64610c732" y="149.225728"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_eaabaca787" y="140.321933"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="174.96" xlink:href="#C0_0_e64610c732" y="140.321933"/> </g> </g> <g id="PathCollection_4"> @@ -1439,592 +1439,592 @@ C -2.236584 -1.29895 -2.5 -0.663008 -2.5 0 C -2.5 0.663008 -2.236584 1.29895 -1.767767 1.767767 C -1.29895 2.236584 -0.663008 2.5 0 2.5 z -" id="C1_0_d7e70ab93f"/> +" id="C1_0_92e804cb43"/> </defs> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_d7e70ab93f" y="311.794598"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_92e804cb43" y="311.794598"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_d7e70ab93f" y="301.901588"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_92e804cb43" y="301.901588"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="262.252355" xlink:href="#C1_0_d7e70ab93f" y="297.797806"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="262.252355" xlink:href="#C1_0_92e804cb43" y="297.797806"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_d7e70ab93f" y="291.118288"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_92e804cb43" y="291.118288"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="259.701964" xlink:href="#C1_0_d7e70ab93f" y="288.888843"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="259.701964" xlink:href="#C1_0_92e804cb43" y="288.888843"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="265.532194" xlink:href="#C1_0_d7e70ab93f" y="286.820857"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="265.532194" xlink:href="#C1_0_92e804cb43" y="286.820857"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="262.5816" xlink:href="#C1_0_d7e70ab93f" y="283.153522"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="262.5816" xlink:href="#C1_0_92e804cb43" y="283.153522"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="266.73428" xlink:href="#C1_0_d7e70ab93f" y="280.440755"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="266.73428" xlink:href="#C1_0_92e804cb43" y="280.440755"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_d7e70ab93f" y="273.294834"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_92e804cb43" y="273.294834"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="258.99008" xlink:href="#C1_0_d7e70ab93f" y="273.270379"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="258.99008" xlink:href="#C1_0_92e804cb43" y="273.270379"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="267.497457" xlink:href="#C1_0_d7e70ab93f" y="269.817681"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="267.497457" xlink:href="#C1_0_92e804cb43" y="269.817681"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="272.747451" xlink:href="#C1_0_d7e70ab93f" y="269.811058"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="272.747451" xlink:href="#C1_0_92e804cb43" y="269.811058"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="255.743569" xlink:href="#C1_0_d7e70ab93f" y="269.785933"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="255.743569" xlink:href="#C1_0_92e804cb43" y="269.785933"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="250.493571" xlink:href="#C1_0_d7e70ab93f" y="269.782482"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="250.493571" xlink:href="#C1_0_92e804cb43" y="269.782482"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="277.997326" xlink:href="#C1_0_d7e70ab93f" y="269.780527"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="277.997326" xlink:href="#C1_0_92e804cb43" y="269.780527"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="245.243572" xlink:href="#C1_0_d7e70ab93f" y="269.779563"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="245.243572" xlink:href="#C1_0_92e804cb43" y="269.779563"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="263.194506" xlink:href="#C1_0_d7e70ab93f" y="267.27744"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="263.194506" xlink:href="#C1_0_92e804cb43" y="267.27744"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="281.434117" xlink:href="#C1_0_d7e70ab93f" y="266.428761"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="281.434117" xlink:href="#C1_0_92e804cb43" y="266.428761"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="241.828264" xlink:href="#C1_0_d7e70ab93f" y="266.412171"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="241.828264" xlink:href="#C1_0_92e804cb43" y="266.412171"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="258.406583" xlink:href="#C1_0_d7e70ab93f" y="265.964836"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="258.406583" xlink:href="#C1_0_92e804cb43" y="265.964836"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_d7e70ab93f" y="260.282911"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_92e804cb43" y="260.282911"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="258.990007" xlink:href="#C1_0_d7e70ab93f" y="260.275668"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="258.990007" xlink:href="#C1_0_92e804cb43" y="260.275668"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="269.489939" xlink:href="#C1_0_d7e70ab93f" y="260.26154"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="269.489939" xlink:href="#C1_0_92e804cb43" y="260.26154"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="253.752941" xlink:href="#C1_0_d7e70ab93f" y="259.964629"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="253.752941" xlink:href="#C1_0_92e804cb43" y="259.964629"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="272.38985" xlink:href="#C1_0_d7e70ab93f" y="256.565483"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="272.38985" xlink:href="#C1_0_92e804cb43" y="256.565483"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_d7e70ab93f" y="255.427417"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_92e804cb43" y="255.427417"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="258.990008" xlink:href="#C1_0_d7e70ab93f" y="255.419612"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="258.990008" xlink:href="#C1_0_92e804cb43" y="255.419612"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="253.740008" xlink:href="#C1_0_d7e70ab93f" y="255.41812"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="253.740008" xlink:href="#C1_0_92e804cb43" y="255.41812"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="277.459012" xlink:href="#C1_0_d7e70ab93f" y="255.411801"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="277.459012" xlink:href="#C1_0_92e804cb43" y="255.411801"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="248.490014" xlink:href="#C1_0_d7e70ab93f" y="255.411799"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="248.490014" xlink:href="#C1_0_92e804cb43" y="255.411799"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="282.709" xlink:href="#C1_0_d7e70ab93f" y="255.402063"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="282.709" xlink:href="#C1_0_92e804cb43" y="255.402063"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="243.240036" xlink:href="#C1_0_d7e70ab93f" y="255.398946"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="243.240036" xlink:href="#C1_0_92e804cb43" y="255.398946"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="287.958998" xlink:href="#C1_0_d7e70ab93f" y="255.39888"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="287.958998" xlink:href="#C1_0_92e804cb43" y="255.39888"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="237.990041" xlink:href="#C1_0_d7e70ab93f" y="255.392631"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="237.990041" xlink:href="#C1_0_92e804cb43" y="255.392631"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="293.208977" xlink:href="#C1_0_d7e70ab93f" y="255.386258"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="293.208977" xlink:href="#C1_0_92e804cb43" y="255.386258"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="232.835552" xlink:href="#C1_0_d7e70ab93f" y="254.550737"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="232.835552" xlink:href="#C1_0_92e804cb43" y="254.550737"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="298.35754" xlink:href="#C1_0_d7e70ab93f" y="254.518889"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="298.35754" xlink:href="#C1_0_92e804cb43" y="254.518889"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="267.290088" xlink:href="#C1_0_d7e70ab93f" y="251.818608"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="267.290088" xlink:href="#C1_0_92e804cb43" y="251.818608"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="272.540085" xlink:href="#C1_0_d7e70ab93f" y="251.814102"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="272.540085" xlink:href="#C1_0_92e804cb43" y="251.814102"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="228.707259" xlink:href="#C1_0_d7e70ab93f" y="251.811551"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="228.707259" xlink:href="#C1_0_92e804cb43" y="251.811551"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="302.512566" xlink:href="#C1_0_d7e70ab93f" y="251.808685"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="302.512566" xlink:href="#C1_0_92e804cb43" y="251.808685"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="223.457266" xlink:href="#C1_0_d7e70ab93f" y="251.804133"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="223.457266" xlink:href="#C1_0_92e804cb43" y="251.804133"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="262.253025" xlink:href="#C1_0_d7e70ab93f" y="250.568662"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="262.253025" xlink:href="#C1_0_92e804cb43" y="250.568662"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="257.003118" xlink:href="#C1_0_d7e70ab93f" y="250.542262"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="257.003118" xlink:href="#C1_0_92e804cb43" y="250.542262"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="251.902653" xlink:href="#C1_0_d7e70ab93f" y="249.491581"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="251.902653" xlink:href="#C1_0_92e804cb43" y="249.491581"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_d7e70ab93f" y="246.069025"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_92e804cb43" y="246.069025"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="259.914095" xlink:href="#C1_0_d7e70ab93f" y="243.556743"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="259.914095" xlink:href="#C1_0_92e804cb43" y="243.556743"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="268.53352" xlink:href="#C1_0_d7e70ab93f" y="243.517426"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="268.53352" xlink:href="#C1_0_92e804cb43" y="243.517426"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_d7e70ab93f" y="241.198577"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_92e804cb43" y="241.198577"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="272.999865" xlink:href="#C1_0_d7e70ab93f" y="241.187009"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="272.999865" xlink:href="#C1_0_92e804cb43" y="241.187009"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="255.482088" xlink:href="#C1_0_d7e70ab93f" y="241.180026"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="255.482088" xlink:href="#C1_0_92e804cb43" y="241.180026"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="250.232089" xlink:href="#C1_0_d7e70ab93f" y="241.177939"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="250.232089" xlink:href="#C1_0_92e804cb43" y="241.177939"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="278.249853" xlink:href="#C1_0_d7e70ab93f" y="241.177174"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="278.249853" xlink:href="#C1_0_92e804cb43" y="241.177174"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="245.017601" xlink:href="#C1_0_d7e70ab93f" y="240.663103"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="245.017601" xlink:href="#C1_0_92e804cb43" y="240.663103"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_d7e70ab93f" y="235.610721"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_92e804cb43" y="235.610721"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="258.995115" xlink:href="#C1_0_d7e70ab93f" y="235.415056"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="258.995115" xlink:href="#C1_0_92e804cb43" y="235.415056"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="269.44962" xlink:href="#C1_0_d7e70ab93f" y="235.061864"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="269.44962" xlink:href="#C1_0_92e804cb43" y="235.061864"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="274.69962" xlink:href="#C1_0_d7e70ab93f" y="235.061728"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="274.69962" xlink:href="#C1_0_92e804cb43" y="235.061728"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="254.476913" xlink:href="#C1_0_d7e70ab93f" y="233.157061"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="254.476913" xlink:href="#C1_0_92e804cb43" y="233.157061"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="249.226939" xlink:href="#C1_0_d7e70ab93f" y="233.143013"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="249.226939" xlink:href="#C1_0_92e804cb43" y="233.143013"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="279.430393" xlink:href="#C1_0_d7e70ab93f" y="233.139163"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="279.430393" xlink:href="#C1_0_92e804cb43" y="233.139163"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="244.044893" xlink:href="#C1_0_d7e70ab93f" y="232.431942"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="244.044893" xlink:href="#C1_0_92e804cb43" y="232.431942"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_d7e70ab93f" y="226.81512"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_92e804cb43" y="226.81512"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="258.990008" xlink:href="#C1_0_d7e70ab93f" y="226.80728"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="258.990008" xlink:href="#C1_0_92e804cb43" y="226.80728"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="269.48994" xlink:href="#C1_0_d7e70ab93f" y="226.793858"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="269.48994" xlink:href="#C1_0_92e804cb43" y="226.793858"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="253.740118" xlink:href="#C1_0_d7e70ab93f" y="226.778611"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="253.740118" xlink:href="#C1_0_92e804cb43" y="226.778611"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="274.739907" xlink:href="#C1_0_d7e70ab93f" y="226.778235"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="274.739907" xlink:href="#C1_0_92e804cb43" y="226.778235"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="248.49012" xlink:href="#C1_0_d7e70ab93f" y="226.775105"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="248.49012" xlink:href="#C1_0_92e804cb43" y="226.775105"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="279.989881" xlink:href="#C1_0_d7e70ab93f" y="226.764233"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="279.989881" xlink:href="#C1_0_92e804cb43" y="226.764233"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="244.028666" xlink:href="#C1_0_d7e70ab93f" y="224.438013"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="244.028666" xlink:href="#C1_0_92e804cb43" y="224.438013"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="284.454966" xlink:href="#C1_0_d7e70ab93f" y="224.432093"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="284.454966" xlink:href="#C1_0_92e804cb43" y="224.432093"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="238.778674" xlink:href="#C1_0_d7e70ab93f" y="224.430566"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="238.778674" xlink:href="#C1_0_92e804cb43" y="224.430566"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="289.704966" xlink:href="#C1_0_d7e70ab93f" y="224.430269"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="289.704966" xlink:href="#C1_0_92e804cb43" y="224.430269"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="233.528696" xlink:href="#C1_0_d7e70ab93f" y="224.417648"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="233.528696" xlink:href="#C1_0_92e804cb43" y="224.417648"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="294.954929" xlink:href="#C1_0_d7e70ab93f" y="224.413712"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="294.954929" xlink:href="#C1_0_92e804cb43" y="224.413712"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="228.278699" xlink:href="#C1_0_d7e70ab93f" y="224.412765"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="228.278699" xlink:href="#C1_0_92e804cb43" y="224.412765"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="300.204887" xlink:href="#C1_0_d7e70ab93f" y="224.396036"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="300.204887" xlink:href="#C1_0_92e804cb43" y="224.396036"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="223.22843" xlink:href="#C1_0_d7e70ab93f" y="223.201424"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="223.22843" xlink:href="#C1_0_92e804cb43" y="223.201424"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="305.25821" xlink:href="#C1_0_d7e70ab93f" y="223.19381"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="305.25821" xlink:href="#C1_0_92e804cb43" y="223.19381"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="220.4928" xlink:href="#C1_0_d7e70ab93f" y="223.189148"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="220.4928" xlink:href="#C1_0_92e804cb43" y="223.189148"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="263.526439" xlink:href="#C1_0_d7e70ab93f" y="222.422431"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="263.526439" xlink:href="#C1_0_92e804cb43" y="222.422431"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="268.631655" xlink:href="#C1_0_d7e70ab93f" y="222.419677"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="268.631655" xlink:href="#C1_0_92e804cb43" y="222.419677"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="258.375763" xlink:href="#C1_0_d7e70ab93f" y="221.564055"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="258.375763" xlink:href="#C1_0_92e804cb43" y="221.564055"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_d7e70ab93f" y="217.275239"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_92e804cb43" y="217.275239"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="259.004415" xlink:href="#C1_0_d7e70ab93f" y="216.946904"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="259.004415" xlink:href="#C1_0_92e804cb43" y="216.946904"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="269.27673" xlink:href="#C1_0_d7e70ab93f" y="216.024336"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="269.27673" xlink:href="#C1_0_92e804cb43" y="216.024336"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="254.023025" xlink:href="#C1_0_d7e70ab93f" y="215.546836"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="254.023025" xlink:href="#C1_0_92e804cb43" y="215.546836"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="274.359389" xlink:href="#C1_0_d7e70ab93f" y="214.913814"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="274.359389" xlink:href="#C1_0_92e804cb43" y="214.913814"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="249.116212" xlink:href="#C1_0_d7e70ab93f" y="213.970083"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="249.116212" xlink:href="#C1_0_92e804cb43" y="213.970083"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="279.488977" xlink:href="#C1_0_d7e70ab93f" y="213.969654"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="279.488977" xlink:href="#C1_0_92e804cb43" y="213.969654"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="243.86622" xlink:href="#C1_0_d7e70ab93f" y="213.961969"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="243.86622" xlink:href="#C1_0_92e804cb43" y="213.961969"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_d7e70ab93f" y="212.533043"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_92e804cb43" y="212.533043"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="258.51585" xlink:href="#C1_0_d7e70ab93f" y="212.532311"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="258.51585" xlink:href="#C1_0_92e804cb43" y="212.532311"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="268.602196" xlink:href="#C1_0_d7e70ab93f" y="210.065925"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="268.602196" xlink:href="#C1_0_92e804cb43" y="210.065925"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="273.852183" xlink:href="#C1_0_d7e70ab93f" y="210.056094"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="273.852183" xlink:href="#C1_0_92e804cb43" y="210.056094"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="254.163461" xlink:href="#C1_0_d7e70ab93f" y="210.05287"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="254.163461" xlink:href="#C1_0_92e804cb43" y="210.05287"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="246.688119" xlink:href="#C1_0_d7e70ab93f" y="210.038947"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="246.688119" xlink:href="#C1_0_92e804cb43" y="210.038947"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="281.913812" xlink:href="#C1_0_d7e70ab93f" y="210.037084"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="281.913812" xlink:href="#C1_0_92e804cb43" y="210.037084"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="241.438122" xlink:href="#C1_0_d7e70ab93f" y="210.034042"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="241.438122" xlink:href="#C1_0_92e804cb43" y="210.034042"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="287.163804" xlink:href="#C1_0_d7e70ab93f" y="210.029456"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="287.163804" xlink:href="#C1_0_92e804cb43" y="210.029456"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="236.188125" xlink:href="#C1_0_d7e70ab93f" y="210.029435"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="236.188125" xlink:href="#C1_0_92e804cb43" y="210.029435"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="292.413804" xlink:href="#C1_0_d7e70ab93f" y="210.029332"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="292.413804" xlink:href="#C1_0_92e804cb43" y="210.029332"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="230.938125" xlink:href="#C1_0_d7e70ab93f" y="210.029281"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="230.938125" xlink:href="#C1_0_92e804cb43" y="210.029281"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="297.663801" xlink:href="#C1_0_d7e70ab93f" y="210.024879"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="297.663801" xlink:href="#C1_0_92e804cb43" y="210.024879"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="225.688128" xlink:href="#C1_0_d7e70ab93f" y="210.024603"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="225.688128" xlink:href="#C1_0_92e804cb43" y="210.024603"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_d7e70ab93f" y="206.456861"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_92e804cb43" y="206.456861"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="258.990055" xlink:href="#C1_0_d7e70ab93f" y="206.43648"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="258.990055" xlink:href="#C1_0_92e804cb43" y="206.43648"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="269.268046" xlink:href="#C1_0_d7e70ab93f" y="205.181284"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="269.268046" xlink:href="#C1_0_92e804cb43" y="205.181284"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="254.257564" xlink:href="#C1_0_d7e70ab93f" y="204.516936"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="254.257564" xlink:href="#C1_0_92e804cb43" y="204.516936"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="274.456441" xlink:href="#C1_0_d7e70ab93f" y="204.50404"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="274.456441" xlink:href="#C1_0_92e804cb43" y="204.50404"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="249.203059" xlink:href="#C1_0_d7e70ab93f" y="203.318259"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="249.203059" xlink:href="#C1_0_92e804cb43" y="203.318259"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="279.514606" xlink:href="#C1_0_d7e70ab93f" y="203.316429"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="279.514606" xlink:href="#C1_0_92e804cb43" y="203.316429"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="243.964867" xlink:href="#C1_0_d7e70ab93f" y="203.021057"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="243.964867" xlink:href="#C1_0_92e804cb43" y="203.021057"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="284.752771" xlink:href="#C1_0_d7e70ab93f" y="203.018884"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="284.752771" xlink:href="#C1_0_92e804cb43" y="203.018884"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="238.72024" xlink:href="#C1_0_d7e70ab93f" y="202.820506"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="238.72024" xlink:href="#C1_0_92e804cb43" y="202.820506"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_d7e70ab93f" y="200.071485"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_92e804cb43" y="200.071485"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="269.489988" xlink:href="#C1_0_d7e70ab93f" y="200.061978"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="269.489988" xlink:href="#C1_0_92e804cb43" y="200.061978"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="259.156659" xlink:href="#C1_0_d7e70ab93f" y="198.963192"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="259.156659" xlink:href="#C1_0_92e804cb43" y="198.963192"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="274.243379" xlink:href="#C1_0_d7e70ab93f" y="198.179627"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="274.243379" xlink:href="#C1_0_92e804cb43" y="198.179627"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="253.989996" xlink:href="#C1_0_d7e70ab93f" y="198.17632"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="253.989996" xlink:href="#C1_0_92e804cb43" y="198.17632"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="279.493367" xlink:href="#C1_0_d7e70ab93f" y="198.170092"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="279.493367" xlink:href="#C1_0_92e804cb43" y="198.170092"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="248.740002" xlink:href="#C1_0_d7e70ab93f" y="198.170038"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="248.740002" xlink:href="#C1_0_92e804cb43" y="198.170038"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="284.743313" xlink:href="#C1_0_d7e70ab93f" y="198.14992"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="284.743313" xlink:href="#C1_0_92e804cb43" y="198.14992"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="243.490061" xlink:href="#C1_0_d7e70ab93f" y="198.148897"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="243.490061" xlink:href="#C1_0_92e804cb43" y="198.148897"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="289.993312" xlink:href="#C1_0_d7e70ab93f" y="198.148057"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="289.993312" xlink:href="#C1_0_92e804cb43" y="198.148057"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="238.24008" xlink:href="#C1_0_d7e70ab93f" y="198.137035"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="238.24008" xlink:href="#C1_0_92e804cb43" y="198.137035"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="295.243035" xlink:href="#C1_0_d7e70ab93f" y="198.102436"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="295.243035" xlink:href="#C1_0_92e804cb43" y="198.102436"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="262.756765" xlink:href="#C1_0_d7e70ab93f" y="195.818282"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="262.756765" xlink:href="#C1_0_92e804cb43" y="195.818282"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="268.006663" xlink:href="#C1_0_d7e70ab93f" y="195.790641"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="268.006663" xlink:href="#C1_0_92e804cb43" y="195.790641"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="233.786179" xlink:href="#C1_0_d7e70ab93f" y="195.789691"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="233.786179" xlink:href="#C1_0_92e804cb43" y="195.789691"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="299.71932" xlink:href="#C1_0_d7e70ab93f" y="195.785661"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="299.71932" xlink:href="#C1_0_92e804cb43" y="195.785661"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="257.725508" xlink:href="#C1_0_d7e70ab93f" y="194.551767"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="257.725508" xlink:href="#C1_0_92e804cb43" y="194.551767"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_d7e70ab93f" y="189.413314"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_92e804cb43" y="189.413314"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="259.018616" xlink:href="#C1_0_d7e70ab93f" y="188.951013"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="259.018616" xlink:href="#C1_0_92e804cb43" y="188.951013"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="269.459935" xlink:href="#C1_0_d7e70ab93f" y="188.939481"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="269.459935" xlink:href="#C1_0_92e804cb43" y="188.939481"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="274.709933" xlink:href="#C1_0_d7e70ab93f" y="188.936081"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="274.709933" xlink:href="#C1_0_92e804cb43" y="188.936081"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="253.949709" xlink:href="#C1_0_d7e70ab93f" y="187.796529"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="253.949709" xlink:href="#C1_0_92e804cb43" y="187.796529"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="279.777146" xlink:href="#C1_0_d7e70ab93f" y="187.776306"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="279.777146" xlink:href="#C1_0_92e804cb43" y="187.776306"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="248.800065" xlink:href="#C1_0_d7e70ab93f" y="186.933748"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="248.800065" xlink:href="#C1_0_92e804cb43" y="186.933748"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="284.928831" xlink:href="#C1_0_d7e70ab93f" y="186.922261"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="284.928831" xlink:href="#C1_0_92e804cb43" y="186.922261"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="243.609746" xlink:href="#C1_0_d7e70ab93f" y="186.2671"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="243.609746" xlink:href="#C1_0_92e804cb43" y="186.2671"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="262.16309" xlink:href="#C1_0_d7e70ab93f" y="185.341185"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="262.16309" xlink:href="#C1_0_92e804cb43" y="185.341185"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="267.213582" xlink:href="#C1_0_d7e70ab93f" y="184.130507"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="267.213582" xlink:href="#C1_0_92e804cb43" y="184.130507"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="257.194757" xlink:href="#C1_0_d7e70ab93f" y="183.908409"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="257.194757" xlink:href="#C1_0_92e804cb43" y="183.908409"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="272.456984" xlink:href="#C1_0_d7e70ab93f" y="183.908283"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="272.456984" xlink:href="#C1_0_92e804cb43" y="183.908283"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="288.777358" xlink:href="#C1_0_d7e70ab93f" y="183.906492"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="288.777358" xlink:href="#C1_0_92e804cb43" y="183.906492"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="252.059442" xlink:href="#C1_0_d7e70ab93f" y="182.986723"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="252.059442" xlink:href="#C1_0_92e804cb43" y="182.986723"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="277.588098" xlink:href="#C1_0_d7e70ab93f" y="182.970059"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="277.588098" xlink:href="#C1_0_92e804cb43" y="182.970059"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="247.148419" xlink:href="#C1_0_d7e70ab93f" y="181.419343"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="247.148419" xlink:href="#C1_0_92e804cb43" y="181.419343"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="282.504498" xlink:href="#C1_0_d7e70ab93f" y="181.414748"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="282.504498" xlink:href="#C1_0_92e804cb43" y="181.414748"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="241.898557" xlink:href="#C1_0_d7e70ab93f" y="181.3872"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="241.898557" xlink:href="#C1_0_92e804cb43" y="181.3872"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="236.648563" xlink:href="#C1_0_d7e70ab93f" y="181.380799"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="236.648563" xlink:href="#C1_0_92e804cb43" y="181.380799"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_d7e70ab93f" y="179.043565"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_92e804cb43" y="179.043565"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="258.990003" xlink:href="#C1_0_d7e70ab93f" y="179.038821"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="258.990003" xlink:href="#C1_0_92e804cb43" y="179.038821"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="267.925797" xlink:href="#C1_0_d7e70ab93f" y="175.886138"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="267.925797" xlink:href="#C1_0_92e804cb43" y="175.886138"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="261.97366" xlink:href="#C1_0_d7e70ab93f" y="175.044137"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="261.97366" xlink:href="#C1_0_92e804cb43" y="175.044137"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="256.723661" xlink:href="#C1_0_d7e70ab93f" y="175.040548"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="256.723661" xlink:href="#C1_0_92e804cb43" y="175.040548"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="273.078775" xlink:href="#C1_0_d7e70ab93f" y="175.037672"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="273.078775" xlink:href="#C1_0_92e804cb43" y="175.037672"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="251.473694" xlink:href="#C1_0_d7e70ab93f" y="175.024813"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="251.473694" xlink:href="#C1_0_92e804cb43" y="175.024813"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="278.328506" xlink:href="#C1_0_d7e70ab93f" y="174.992799"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="278.328506" xlink:href="#C1_0_92e804cb43" y="174.992799"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="246.237125" xlink:href="#C1_0_d7e70ab93f" y="174.707864"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="246.237125" xlink:href="#C1_0_92e804cb43" y="174.707864"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="283.566852" xlink:href="#C1_0_d7e70ab93f" y="174.69753"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="283.566852" xlink:href="#C1_0_92e804cb43" y="174.69753"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="240.987151" xlink:href="#C1_0_d7e70ab93f" y="174.693904"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="240.987151" xlink:href="#C1_0_92e804cb43" y="174.693904"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="288.816849" xlink:href="#C1_0_d7e70ab93f" y="174.692786"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="288.816849" xlink:href="#C1_0_92e804cb43" y="174.692786"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="236.634487" xlink:href="#C1_0_d7e70ab93f" y="172.214809"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="236.634487" xlink:href="#C1_0_92e804cb43" y="172.214809"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="293.15579" xlink:href="#C1_0_d7e70ab93f" y="172.19659"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="293.15579" xlink:href="#C1_0_92e804cb43" y="172.19659"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_d7e70ab93f" y="170.758267"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_92e804cb43" y="170.758267"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="259.187695" xlink:href="#C1_0_d7e70ab93f" y="169.552995"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="259.187695" xlink:href="#C1_0_92e804cb43" y="169.552995"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="269.284939" xlink:href="#C1_0_d7e70ab93f" y="169.531185"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="269.284939" xlink:href="#C1_0_92e804cb43" y="169.531185"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="254.763381" xlink:href="#C1_0_d7e70ab93f" y="167.166077"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="254.763381" xlink:href="#C1_0_92e804cb43" y="167.166077"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="273.290803" xlink:href="#C1_0_d7e70ab93f" y="166.665277"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="273.290803" xlink:href="#C1_0_92e804cb43" y="166.665277"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_d7e70ab93f" y="164.681466"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_92e804cb43" y="164.681466"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="259.101496" xlink:href="#C1_0_d7e70ab93f" y="164.668857"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="259.101496" xlink:href="#C1_0_92e804cb43" y="164.668857"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="277.969594" xlink:href="#C1_0_d7e70ab93f" y="164.654027"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="277.969594" xlink:href="#C1_0_92e804cb43" y="164.654027"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="269.281149" xlink:href="#C1_0_d7e70ab93f" y="163.443325"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="269.281149" xlink:href="#C1_0_92e804cb43" y="163.443325"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="263.189367" xlink:href="#C1_0_d7e70ab93f" y="160.337324"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="263.189367" xlink:href="#C1_0_92e804cb43" y="160.337324"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="258.151866" xlink:href="#C1_0_d7e70ab93f" y="160.308161"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="258.151866" xlink:href="#C1_0_92e804cb43" y="160.308161"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="270.596639" xlink:href="#C1_0_d7e70ab93f" y="159.150938"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="270.596639" xlink:href="#C1_0_92e804cb43" y="159.150938"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="253.085036" xlink:href="#C1_0_d7e70ab93f" y="159.147192"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="253.085036" xlink:href="#C1_0_92e804cb43" y="159.147192"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="265.620942" xlink:href="#C1_0_d7e70ab93f" y="156.407724"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="265.620942" xlink:href="#C1_0_92e804cb43" y="156.407724"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="260.59086" xlink:href="#C1_0_d7e70ab93f" y="156.381842"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="260.59086" xlink:href="#C1_0_92e804cb43" y="156.381842"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="274.676997" xlink:href="#C1_0_d7e70ab93f" y="156.360992"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="274.676997" xlink:href="#C1_0_92e804cb43" y="156.360992"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="268.641423" xlink:href="#C1_0_d7e70ab93f" y="152.781196"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="268.641423" xlink:href="#C1_0_92e804cb43" y="152.781196"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="257.54481" xlink:href="#C1_0_d7e70ab93f" y="152.7706"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="257.54481" xlink:href="#C1_0_92e804cb43" y="152.7706"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_d7e70ab93f" y="149.465622"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_92e804cb43" y="149.465622"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="259.649718" xlink:href="#C1_0_d7e70ab93f" y="147.313817"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="259.649718" xlink:href="#C1_0_92e804cb43" y="147.313817"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="268.826854" xlink:href="#C1_0_d7e70ab93f" y="147.308609"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="268.826854" xlink:href="#C1_0_92e804cb43" y="147.308609"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="274.076634" xlink:href="#C1_0_d7e70ab93f" y="147.268016"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="274.076634" xlink:href="#C1_0_92e804cb43" y="147.268016"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_d7e70ab93f" y="143.569627"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_92e804cb43" y="143.569627"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="259.163522" xlink:href="#C1_0_d7e70ab93f" y="142.439121"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="259.163522" xlink:href="#C1_0_92e804cb43" y="142.439121"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="268.438812" xlink:href="#C1_0_d7e70ab93f" y="140.90799"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="268.438812" xlink:href="#C1_0_92e804cb43" y="140.90799"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_d7e70ab93f" y="127.733366"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_92e804cb43" y="127.733366"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="261.113693" xlink:href="#C1_0_d7e70ab93f" y="124.171381"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="261.113693" xlink:href="#C1_0_92e804cb43" y="124.171381"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_d7e70ab93f" y="116.087931"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_92e804cb43" y="116.087931"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_d7e70ab93f" y="101.918938"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="264.24" xlink:href="#C1_0_92e804cb43" y="101.918938"/> </g> </g> <g id="PathCollection_5"> @@ -2039,277 +2039,277 @@ C -2.236584 -1.29895 -2.5 -0.663008 -2.5 0 C -2.5 0.663008 -2.236584 1.29895 -1.767767 1.767767 C -1.29895 2.236584 -0.663008 2.5 0 2.5 z -" id="C2_0_e0b3868871"/> +" id="C2_0_f2b1f24b3b"/> </defs> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_e0b3868871" y="318.553243"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_f2b1f24b3b" y="318.553243"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_e0b3868871" y="308.206677"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_f2b1f24b3b" y="308.206677"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="393.837939" xlink:href="#C2_0_e0b3868871" y="305.689681"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="393.837939" xlink:href="#C2_0_f2b1f24b3b" y="305.689681"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="401.57735" xlink:href="#C2_0_e0b3868871" y="304.840763"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="401.57735" xlink:href="#C2_0_f2b1f24b3b" y="304.840763"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_e0b3868871" y="293.851516"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_f2b1f24b3b" y="293.851516"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="392.910014" xlink:href="#C2_0_e0b3868871" y="293.841403"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="392.910014" xlink:href="#C2_0_f2b1f24b3b" y="293.841403"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="403.409756" xlink:href="#C2_0_e0b3868871" y="293.808801"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="403.409756" xlink:href="#C2_0_f2b1f24b3b" y="293.808801"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="389.848692" xlink:href="#C2_0_e0b3868871" y="290.239384"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="389.848692" xlink:href="#C2_0_f2b1f24b3b" y="290.239384"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_e0b3868871" y="271.595185"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_f2b1f24b3b" y="271.595185"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="392.910235" xlink:href="#C2_0_e0b3868871" y="271.553233"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="392.910235" xlink:href="#C2_0_f2b1f24b3b" y="271.553233"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="403.409764" xlink:href="#C2_0_e0b3868871" y="271.553172"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="403.409764" xlink:href="#C2_0_f2b1f24b3b" y="271.553172"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_e0b3868871" y="265.228305"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_f2b1f24b3b" y="265.228305"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="392.910066" xlink:href="#C2_0_e0b3868871" y="265.206087"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="392.910066" xlink:href="#C2_0_f2b1f24b3b" y="265.206087"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="403.409803" xlink:href="#C2_0_e0b3868871" y="265.189848"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="403.409803" xlink:href="#C2_0_f2b1f24b3b" y="265.189848"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="408.659802" xlink:href="#C2_0_e0b3868871" y="265.187392"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="408.659802" xlink:href="#C2_0_f2b1f24b3b" y="265.187392"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="387.672579" xlink:href="#C2_0_e0b3868871" y="264.900143"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="387.672579" xlink:href="#C2_0_f2b1f24b3b" y="264.900143"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="413.126275" xlink:href="#C2_0_e0b3868871" y="262.85715"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="413.126275" xlink:href="#C2_0_f2b1f24b3b" y="262.85715"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="383.025001" xlink:href="#C2_0_e0b3868871" y="262.837923"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="383.025001" xlink:href="#C2_0_f2b1f24b3b" y="262.837923"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="418.36266" xlink:href="#C2_0_e0b3868871" y="262.538035"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="418.36266" xlink:href="#C2_0_f2b1f24b3b" y="262.538035"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="377.787033" xlink:href="#C2_0_e0b3868871" y="262.537915"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="377.787033" xlink:href="#C2_0_f2b1f24b3b" y="262.537915"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="423.499105" xlink:href="#C2_0_e0b3868871" y="261.62085"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="423.499105" xlink:href="#C2_0_f2b1f24b3b" y="261.62085"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_e0b3868871" y="253.305883"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_f2b1f24b3b" y="253.305883"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="393.015157" xlink:href="#C2_0_e0b3868871" y="252.422909"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="393.015157" xlink:href="#C2_0_f2b1f24b3b" y="252.422909"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="403.084825" xlink:href="#C2_0_e0b3868871" y="251.769702"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="403.084825" xlink:href="#C2_0_f2b1f24b3b" y="251.769702"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="388.19808" xlink:href="#C2_0_e0b3868871" y="250.659809"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="388.19808" xlink:href="#C2_0_f2b1f24b3b" y="250.659809"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_e0b3868871" y="248.469722"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_f2b1f24b3b" y="248.469722"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="406.131089" xlink:href="#C2_0_e0b3868871" y="248.158589"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="406.131089" xlink:href="#C2_0_f2b1f24b3b" y="248.158589"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_e0b3868871" y="241.773311"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_f2b1f24b3b" y="241.773311"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="393.852899" xlink:href="#C2_0_e0b3868871" y="239.23809"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="393.852899" xlink:href="#C2_0_f2b1f24b3b" y="239.23809"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_e0b3868871" y="236.593294"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_f2b1f24b3b" y="236.593294"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="403.40999" xlink:href="#C2_0_e0b3868871" y="236.584609"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="403.40999" xlink:href="#C2_0_f2b1f24b3b" y="236.584609"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="393.604555" xlink:href="#C2_0_e0b3868871" y="234.389318"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="393.604555" xlink:href="#C2_0_f2b1f24b3b" y="234.389318"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="407.602358" xlink:href="#C2_0_e0b3868871" y="233.915738"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="407.602358" xlink:href="#C2_0_f2b1f24b3b" y="233.915738"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_e0b3868871" y="231.715173"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_f2b1f24b3b" y="231.715173"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="403.223143" xlink:href="#C2_0_e0b3868871" y="230.542788"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="403.223143" xlink:href="#C2_0_f2b1f24b3b" y="230.542788"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="396.367135" xlink:href="#C2_0_e0b3868871" y="227.547891"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="396.367135" xlink:href="#C2_0_f2b1f24b3b" y="227.547891"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_e0b3868871" y="222.038458"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_f2b1f24b3b" y="222.038458"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="393.593566" xlink:href="#C2_0_e0b3868871" y="219.850762"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="393.593566" xlink:href="#C2_0_f2b1f24b3b" y="219.850762"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="402.72461" xlink:href="#C2_0_e0b3868871" y="219.848048"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="402.72461" xlink:href="#C2_0_f2b1f24b3b" y="219.848048"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="407.974561" xlink:href="#C2_0_e0b3868871" y="219.8288"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="407.974561" xlink:href="#C2_0_f2b1f24b3b" y="219.8288"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="389.003287" xlink:href="#C2_0_e0b3868871" y="217.698952"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="389.003287" xlink:href="#C2_0_f2b1f24b3b" y="217.698952"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_e0b3868871" y="216.600018"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_f2b1f24b3b" y="216.600018"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="403.744095" xlink:href="#C2_0_e0b3868871" y="215.498616"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="403.744095" xlink:href="#C2_0_f2b1f24b3b" y="215.498616"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="393.668006" xlink:href="#C2_0_e0b3868871" y="214.305031"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="393.668006" xlink:href="#C2_0_f2b1f24b3b" y="214.305031"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="388.72135" xlink:href="#C2_0_e0b3868871" y="212.819723"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="388.72135" xlink:href="#C2_0_f2b1f24b3b" y="212.819723"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="407.926716" xlink:href="#C2_0_e0b3868871" y="212.818859"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="407.926716" xlink:href="#C2_0_f2b1f24b3b" y="212.818859"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="383.472614" xlink:href="#C2_0_e0b3868871" y="212.722431"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="383.472614" xlink:href="#C2_0_f2b1f24b3b" y="212.722431"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_e0b3868871" y="210.012585"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_f2b1f24b3b" y="210.012585"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="403.384385" xlink:href="#C2_0_e0b3868871" y="209.575135"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="403.384385" xlink:href="#C2_0_f2b1f24b3b" y="209.575135"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="394.27286" xlink:href="#C2_0_e0b3868871" y="207.032348"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="394.27286" xlink:href="#C2_0_f2b1f24b3b" y="207.032348"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="407.039123" xlink:href="#C2_0_e0b3868871" y="206.392062"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="407.039123" xlink:href="#C2_0_f2b1f24b3b" y="206.392062"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_e0b3868871" y="202.28269"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_f2b1f24b3b" y="202.28269"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="392.972654" xlink:href="#C2_0_e0b3868871" y="201.599742"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="392.972654" xlink:href="#C2_0_f2b1f24b3b" y="201.599742"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="402.243918" xlink:href="#C2_0_e0b3868871" y="199.496462"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="402.243918" xlink:href="#C2_0_f2b1f24b3b" y="199.496462"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="407.485358" xlink:href="#C2_0_e0b3868871" y="199.243376"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="407.485358" xlink:href="#C2_0_f2b1f24b3b" y="199.243376"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="397.764241" xlink:href="#C2_0_e0b3868871" y="197.184366"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="397.764241" xlink:href="#C2_0_f2b1f24b3b" y="197.184366"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="392.601705" xlink:href="#C2_0_e0b3868871" y="196.378407"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="392.601705" xlink:href="#C2_0_f2b1f24b3b" y="196.378407"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="403.164253" xlink:href="#C2_0_e0b3868871" y="195.131288"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="403.164253" xlink:href="#C2_0_f2b1f24b3b" y="195.131288"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="408.398208" xlink:href="#C2_0_e0b3868871" y="194.784913"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="408.398208" xlink:href="#C2_0_f2b1f24b3b" y="194.784913"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="388.392795" xlink:href="#C2_0_e0b3868871" y="193.72817"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="388.392795" xlink:href="#C2_0_f2b1f24b3b" y="193.72817"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="383.188077" xlink:href="#C2_0_e0b3868871" y="193.14709"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="383.188077" xlink:href="#C2_0_f2b1f24b3b" y="193.14709"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_e0b3868871" y="191.204515"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_f2b1f24b3b" y="191.204515"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="395.06723" xlink:href="#C2_0_e0b3868871" y="187.621708"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="395.06723" xlink:href="#C2_0_f2b1f24b3b" y="187.621708"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="400.163611" xlink:href="#C2_0_e0b3868871" y="186.556984"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="400.163611" xlink:href="#C2_0_f2b1f24b3b" y="186.556984"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="405.267767" xlink:href="#C2_0_e0b3868871" y="185.519166"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="405.267767" xlink:href="#C2_0_f2b1f24b3b" y="185.519166"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="391.306287" xlink:href="#C2_0_e0b3868871" y="184.528132"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="391.306287" xlink:href="#C2_0_f2b1f24b3b" y="184.528132"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="386.056289" xlink:href="#C2_0_e0b3868871" y="184.52442"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="386.056289" xlink:href="#C2_0_f2b1f24b3b" y="184.52442"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="410.287316" xlink:href="#C2_0_e0b3868871" y="184.21994"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="410.287316" xlink:href="#C2_0_f2b1f24b3b" y="184.21994"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="380.841427" xlink:href="#C2_0_e0b3868871" y="184.012297"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="380.841427" xlink:href="#C2_0_f2b1f24b3b" y="184.012297"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_e0b3868871" y="180.811369"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_f2b1f24b3b" y="180.811369"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="395.481505" xlink:href="#C2_0_e0b3868871" y="176.997998"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="395.481505" xlink:href="#C2_0_f2b1f24b3b" y="176.997998"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="400.726839" xlink:href="#C2_0_e0b3868871" y="176.811114"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="400.726839" xlink:href="#C2_0_f2b1f24b3b" y="176.811114"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="405.964536" xlink:href="#C2_0_e0b3868871" y="176.507744"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="405.964536" xlink:href="#C2_0_f2b1f24b3b" y="176.507744"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="390.641179" xlink:href="#C2_0_e0b3868871" y="175.280913"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="390.641179" xlink:href="#C2_0_f2b1f24b3b" y="175.280913"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="409.664307" xlink:href="#C2_0_e0b3868871" y="173.361997"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="409.664307" xlink:href="#C2_0_f2b1f24b3b" y="173.361997"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_e0b3868871" y="172.229668"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_f2b1f24b3b" y="172.229668"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="403.409983" xlink:href="#C2_0_e0b3868871" y="172.218341"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="403.409983" xlink:href="#C2_0_f2b1f24b3b" y="172.218341"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="386.969213" xlink:href="#C2_0_e0b3868871" y="172.112013"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="386.969213" xlink:href="#C2_0_f2b1f24b3b" y="172.112013"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="381.723573" xlink:href="#C2_0_e0b3868871" y="171.931355"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="381.723573" xlink:href="#C2_0_f2b1f24b3b" y="171.931355"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="414.632729" xlink:href="#C2_0_e0b3868871" y="171.929442"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="414.632729" xlink:href="#C2_0_f2b1f24b3b" y="171.929442"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="393.268716" xlink:href="#C2_0_e0b3868871" y="170.618866"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="393.268716" xlink:href="#C2_0_f2b1f24b3b" y="170.618866"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="376.927678" xlink:href="#C2_0_e0b3868871" y="170.127538"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="376.927678" xlink:href="#C2_0_f2b1f24b3b" y="170.127538"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="400.090722" xlink:href="#C2_0_e0b3868871" y="168.106549"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="400.090722" xlink:href="#C2_0_f2b1f24b3b" y="168.106549"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_e0b3868871" y="162.570934"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_f2b1f24b3b" y="162.570934"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="393.291767" xlink:href="#C2_0_e0b3868871" y="160.911074"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="393.291767" xlink:href="#C2_0_f2b1f24b3b" y="160.911074"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="401.694271" xlink:href="#C2_0_e0b3868871" y="159.292268"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="401.694271" xlink:href="#C2_0_f2b1f24b3b" y="159.292268"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_e0b3868871" y="153.37535"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_f2b1f24b3b" y="153.37535"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="392.922518" xlink:href="#C2_0_e0b3868871" y="153.069347"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="392.922518" xlink:href="#C2_0_f2b1f24b3b" y="153.069347"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_e0b3868871" y="147.554881"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="398.16" xlink:href="#C2_0_f2b1f24b3b" y="147.554881"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="394.107509" xlink:href="#C2_0_e0b3868871" y="144.736114"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="394.107509" xlink:href="#C2_0_f2b1f24b3b" y="144.736114"/> </g> </g> <g id="PathCollection_6"> @@ -2324,259 +2324,259 @@ C -2.236584 -1.29895 -2.5 -0.663008 -2.5 0 C -2.5 0.663008 -2.236584 1.29895 -1.767767 1.767767 C -1.29895 2.236584 -0.663008 2.5 0 2.5 z -" id="C3_0_acbc77e91b"/> +" id="C3_0_761689fab7"/> </defs> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_acbc77e91b" y="280.449976"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_761689fab7" y="280.449976"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_acbc77e91b" y="275.459547"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_761689fab7" y="275.459547"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_acbc77e91b" y="269.772241"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_761689fab7" y="269.772241"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="482.192459" xlink:href="#C3_0_acbc77e91b" y="269.636549"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="482.192459" xlink:href="#C3_0_761689fab7" y="269.636549"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_acbc77e91b" y="260.330797"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_761689fab7" y="260.330797"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_acbc77e91b" y="255.45697"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_761689fab7" y="255.45697"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="482.190509" xlink:href="#C3_0_acbc77e91b" y="255.395219"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="482.190509" xlink:href="#C3_0_761689fab7" y="255.395219"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="490.412756" xlink:href="#C3_0_acbc77e91b" y="251.802423"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="490.412756" xlink:href="#C3_0_761689fab7" y="251.802423"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="495.662744" xlink:href="#C3_0_acbc77e91b" y="251.792999"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="495.662744" xlink:href="#C3_0_761689fab7" y="251.792999"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_acbc77e91b" y="241.007548"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_761689fab7" y="241.007548"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_acbc77e91b" y="235.072473"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_761689fab7" y="235.072473"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="482.678536" xlink:href="#C3_0_acbc77e91b" y="233.20473"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="482.678536" xlink:href="#C3_0_761689fab7" y="233.20473"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="490.529516" xlink:href="#C3_0_acbc77e91b" y="231.487664"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="490.529516" xlink:href="#C3_0_761689fab7" y="231.487664"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_acbc77e91b" y="226.845842"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_761689fab7" y="226.845842"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="482.196559" xlink:href="#C3_0_acbc77e91b" y="226.624271"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="482.196559" xlink:href="#C3_0_761689fab7" y="226.624271"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="491.879044" xlink:href="#C3_0_acbc77e91b" y="224.478511"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="491.879044" xlink:href="#C3_0_761689fab7" y="224.478511"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="497.128268" xlink:href="#C3_0_acbc77e91b" y="224.402285"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="497.128268" xlink:href="#C3_0_761689fab7" y="224.402285"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="477.757705" xlink:href="#C3_0_acbc77e91b" y="224.256686"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="477.757705" xlink:href="#C3_0_761689fab7" y="224.256686"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_acbc77e91b" y="221.604922"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_761689fab7" y="221.604922"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_acbc77e91b" y="216.539783"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_761689fab7" y="216.539783"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="482.317312" xlink:href="#C3_0_acbc77e91b" y="215.569272"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="482.317312" xlink:href="#C3_0_761689fab7" y="215.569272"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="489.790658" xlink:href="#C3_0_acbc77e91b" y="212.57522"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="489.790658" xlink:href="#C3_0_761689fab7" y="212.57522"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="495.035982" xlink:href="#C3_0_acbc77e91b" y="212.388117"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="495.035982" xlink:href="#C3_0_761689fab7" y="212.388117"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="478.665687" xlink:href="#C3_0_acbc77e91b" y="212.383652"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="478.665687" xlink:href="#C3_0_761689fab7" y="212.383652"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="484.789587" xlink:href="#C3_0_acbc77e91b" y="211.226127"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="484.789587" xlink:href="#C3_0_761689fab7" y="211.226127"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="499.480697" xlink:href="#C3_0_acbc77e91b" y="210.028388"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="499.480697" xlink:href="#C3_0_761689fab7" y="210.028388"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="474.224512" xlink:href="#C3_0_acbc77e91b" y="210.019172"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="474.224512" xlink:href="#C3_0_761689fab7" y="210.019172"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="504.727517" xlink:href="#C3_0_acbc77e91b" y="209.874073"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="504.727517" xlink:href="#C3_0_761689fab7" y="209.874073"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_acbc77e91b" y="205.316997"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_761689fab7" y="205.316997"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="482.19088" xlink:href="#C3_0_acbc77e91b" y="205.235818"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="482.19088" xlink:href="#C3_0_761689fab7" y="205.235818"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="492.617134" xlink:href="#C3_0_acbc77e91b" y="204.580849"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="492.617134" xlink:href="#C3_0_761689fab7" y="204.580849"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="477.433755" xlink:href="#C3_0_acbc77e91b" y="203.360206"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="477.433755" xlink:href="#C3_0_761689fab7" y="203.360206"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="497.527712" xlink:href="#C3_0_acbc77e91b" y="203.012476"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="497.527712" xlink:href="#C3_0_761689fab7" y="203.012476"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="472.200882" xlink:href="#C3_0_acbc77e91b" y="203.002364"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="472.200882" xlink:href="#C3_0_761689fab7" y="203.002364"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_acbc77e91b" y="200.855936"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_761689fab7" y="200.855936"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="492.604299" xlink:href="#C3_0_acbc77e91b" y="200.058073"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="492.604299" xlink:href="#C3_0_761689fab7" y="200.058073"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="483.265011" xlink:href="#C3_0_acbc77e91b" y="198.167702"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="483.265011" xlink:href="#C3_0_761689fab7" y="198.167702"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="478.015081" xlink:href="#C3_0_acbc77e91b" y="198.144923"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="478.015081" xlink:href="#C3_0_761689fab7" y="198.144923"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="497.337152" xlink:href="#C3_0_acbc77e91b" y="198.139164"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="497.337152" xlink:href="#C3_0_761689fab7" y="198.139164"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="489.274817" xlink:href="#C3_0_acbc77e91b" y="196.701698"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="489.274817" xlink:href="#C3_0_761689fab7" y="196.701698"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="473.577138" xlink:href="#C3_0_acbc77e91b" y="195.776119"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="473.577138" xlink:href="#C3_0_761689fab7" y="195.776119"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="501.677634" xlink:href="#C3_0_acbc77e91b" y="195.644881"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="501.677634" xlink:href="#C3_0_761689fab7" y="195.644881"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_acbc77e91b" y="189.721571"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_761689fab7" y="189.721571"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="482.19001" xlink:href="#C3_0_acbc77e91b" y="189.713094"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="482.19001" xlink:href="#C3_0_761689fab7" y="189.713094"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="492.677182" xlink:href="#C3_0_acbc77e91b" y="189.411929"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="492.677182" xlink:href="#C3_0_761689fab7" y="189.411929"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="477.008309" xlink:href="#C3_0_acbc77e91b" y="189.000231"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="477.008309" xlink:href="#C3_0_761689fab7" y="189.000231"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="496.621844" xlink:href="#C3_0_acbc77e91b" y="186.486097"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="496.621844" xlink:href="#C3_0_761689fab7" y="186.486097"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_acbc77e91b" y="185.186955"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_761689fab7" y="185.186955"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="492.325384" xlink:href="#C3_0_acbc77e91b" y="183.938031"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="492.325384" xlink:href="#C3_0_761689fab7" y="183.938031"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="482.414748" xlink:href="#C3_0_acbc77e91b" y="183.903549"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="482.414748" xlink:href="#C3_0_761689fab7" y="183.903549"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="477.164752" xlink:href="#C3_0_acbc77e91b" y="183.897578"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="477.164752" xlink:href="#C3_0_761689fab7" y="183.897578"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="500.766196" xlink:href="#C3_0_acbc77e91b" y="183.764262"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="500.766196" xlink:href="#C3_0_761689fab7" y="183.764262"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="471.917198" xlink:href="#C3_0_acbc77e91b" y="183.762274"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="471.917198" xlink:href="#C3_0_761689fab7" y="183.762274"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="505.259488" xlink:href="#C3_0_acbc77e91b" y="181.471088"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="505.259488" xlink:href="#C3_0_761689fab7" y="181.471088"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="467.475677" xlink:href="#C3_0_acbc77e91b" y="181.398257"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="467.475677" xlink:href="#C3_0_761689fab7" y="181.398257"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_acbc77e91b" y="177.877542"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_761689fab7" y="177.877542"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="482.190076" xlink:href="#C3_0_acbc77e91b" y="177.853726"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="482.190076" xlink:href="#C3_0_761689fab7" y="177.853726"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="492.596932" xlink:href="#C3_0_acbc77e91b" y="177.046387"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="492.596932" xlink:href="#C3_0_761689fab7" y="177.046387"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="477.486685" xlink:href="#C3_0_acbc77e91b" y="175.883828"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="477.486685" xlink:href="#C3_0_761689fab7" y="175.883828"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="497.317281" xlink:href="#C3_0_acbc77e91b" y="175.105635"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="497.317281" xlink:href="#C3_0_761689fab7" y="175.105635"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_acbc77e91b" y="170.79334"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_761689fab7" y="170.79334"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="482.396269" xlink:href="#C3_0_acbc77e91b" y="169.562723"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="482.396269" xlink:href="#C3_0_761689fab7" y="169.562723"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="492.011179" xlink:href="#C3_0_acbc77e91b" y="168.612723"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="492.011179" xlink:href="#C3_0_761689fab7" y="168.612723"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="478.098653" xlink:href="#C3_0_acbc77e91b" y="167.016047"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="478.098653" xlink:href="#C3_0_761689fab7" y="167.016047"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_acbc77e91b" y="165.774909"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_761689fab7" y="165.774909"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="482.556218" xlink:href="#C3_0_acbc77e91b" y="164.673669"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="482.556218" xlink:href="#C3_0_761689fab7" y="164.673669"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_acbc77e91b" y="161.234091"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_761689fab7" y="161.234091"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="492.591086" xlink:href="#C3_0_acbc77e91b" y="160.377472"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="492.591086" xlink:href="#C3_0_761689fab7" y="160.377472"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="482.346137" xlink:href="#C3_0_acbc77e91b" y="160.160809"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="482.346137" xlink:href="#C3_0_761689fab7" y="160.160809"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="497.632716" xlink:href="#C3_0_acbc77e91b" y="159.140727"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="497.632716" xlink:href="#C3_0_761689fab7" y="159.140727"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="477.764945" xlink:href="#C3_0_acbc77e91b" y="157.99523"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="477.764945" xlink:href="#C3_0_761689fab7" y="157.99523"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_acbc77e91b" y="156.391807"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_761689fab7" y="156.391807"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="501.736161" xlink:href="#C3_0_acbc77e91b" y="156.375038"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="501.736161" xlink:href="#C3_0_761689fab7" y="156.375038"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="472.949087" xlink:href="#C3_0_acbc77e91b" y="156.229756"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="472.949087" xlink:href="#C3_0_761689fab7" y="156.229756"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="484.398426" xlink:href="#C3_0_acbc77e91b" y="152.777876"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="484.398426" xlink:href="#C3_0_761689fab7" y="152.777876"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="490.220912" xlink:href="#C3_0_acbc77e91b" y="152.631089"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="490.220912" xlink:href="#C3_0_761689fab7" y="152.631089"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="493.904978" xlink:href="#C3_0_acbc77e91b" y="149.472223"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="493.904978" xlink:href="#C3_0_761689fab7" y="149.472223"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_acbc77e91b" y="147.347022"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_761689fab7" y="147.347022"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="482.197045" xlink:href="#C3_0_acbc77e91b" y="147.117396"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="482.197045" xlink:href="#C3_0_761689fab7" y="147.117396"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="497.257321" xlink:href="#C3_0_acbc77e91b" y="146.060001"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="497.257321" xlink:href="#C3_0_761689fab7" y="146.060001"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_acbc77e91b" y="140.975908"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="487.44" xlink:href="#C3_0_761689fab7" y="140.975908"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="482.190784" xlink:href="#C3_0_acbc77e91b" y="140.899286"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="482.190784" xlink:href="#C3_0_761689fab7" y="140.899286"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="492.68895" xlink:href="#C3_0_acbc77e91b" y="140.887246"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="492.68895" xlink:href="#C3_0_761689fab7" y="140.887246"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="477.01552" xlink:href="#C3_0_acbc77e91b" y="140.153818"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="477.01552" xlink:href="#C3_0_761689fab7" y="140.153818"/> </g> </g> <g id="PathCollection_7"> @@ -2591,103 +2591,103 @@ C -2.236584 -1.29895 -2.5 -0.663008 -2.5 0 C -2.5 0.663008 -2.236584 1.29895 -1.767767 1.767767 C -1.29895 2.236584 -0.663008 2.5 0 2.5 z -" id="C4_0_07995dbc83"/> +" id="C4_0_d190ea9174"/> </defs> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_07995dbc83" y="340.777665"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_d190ea9174" y="340.777665"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_07995dbc83" y="334.732129"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_d190ea9174" y="334.732129"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_07995dbc83" y="308.670776"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_d190ea9174" y="308.670776"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_07995dbc83" y="294.28087"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_d190ea9174" y="294.28087"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_07995dbc83" y="280.039095"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_d190ea9174" y="280.039095"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="616.110019" xlink:href="#C4_0_07995dbc83" y="280.027168"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="616.110019" xlink:href="#C4_0_d190ea9174" y="280.027168"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="626.609957" xlink:href="#C4_0_07995dbc83" y="280.021079"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="626.609957" xlink:href="#C4_0_d190ea9174" y="280.021079"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_07995dbc83" y="273.941421"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_d190ea9174" y="273.941421"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_07995dbc83" y="265.654057"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_d190ea9174" y="265.654057"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="618.288247" xlink:href="#C4_0_07995dbc83" y="262.058378"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="618.288247" xlink:href="#C4_0_d190ea9174" y="262.058378"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_07995dbc83" y="248.935194"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_d190ea9174" y="248.935194"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="616.110107" xlink:href="#C4_0_07995dbc83" y="248.9069"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="616.110107" xlink:href="#C4_0_d190ea9174" y="248.9069"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="626.609767" xlink:href="#C4_0_07995dbc83" y="248.893435"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="626.609767" xlink:href="#C4_0_d190ea9174" y="248.893435"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_07995dbc83" y="243.396163"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_d190ea9174" y="243.396163"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_07995dbc83" y="238.959995"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_d190ea9174" y="238.959995"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="616.633467" xlink:href="#C4_0_07995dbc83" y="237.030008"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="616.633467" xlink:href="#C4_0_d190ea9174" y="237.030008"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="626.057533" xlink:href="#C4_0_07995dbc83" y="236.980151"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="626.057533" xlink:href="#C4_0_d190ea9174" y="236.980151"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_07995dbc83" y="227.765285"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_d190ea9174" y="227.765285"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="617.903032" xlink:href="#C4_0_07995dbc83" y="224.428352"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="617.903032" xlink:href="#C4_0_d190ea9174" y="224.428352"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_07995dbc83" y="220.236148"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_d190ea9174" y="220.236148"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_07995dbc83" y="213.258957"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_d190ea9174" y="213.258957"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="616.791913" xlink:href="#C4_0_07995dbc83" y="211.073723"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="616.791913" xlink:href="#C4_0_d190ea9174" y="211.073723"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_07995dbc83" y="207.483215"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_d190ea9174" y="207.483215"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="617.301235" xlink:href="#C4_0_07995dbc83" y="204.670895"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="617.301235" xlink:href="#C4_0_d190ea9174" y="204.670895"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_07995dbc83" y="199.632282"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_d190ea9174" y="199.632282"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="616.168703" xlink:href="#C4_0_07995dbc83" y="198.971089"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="616.168703" xlink:href="#C4_0_d190ea9174" y="198.971089"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="626.370515" xlink:href="#C4_0_07995dbc83" y="198.308418"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="626.370515" xlink:href="#C4_0_d190ea9174" y="198.308418"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_07995dbc83" y="193.584019"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_d190ea9174" y="193.584019"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="616.630728" xlink:href="#C4_0_07995dbc83" y="191.658824"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="616.630728" xlink:href="#C4_0_d190ea9174" y="191.658824"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_07995dbc83" y="188.072881"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_d190ea9174" y="188.072881"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="616.246801" xlink:href="#C4_0_07995dbc83" y="187.067312"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="616.246801" xlink:href="#C4_0_d190ea9174" y="187.067312"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_07995dbc83" y="153.818469"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#4c72b0;" x="621.36" xlink:href="#C4_0_d190ea9174" y="153.818469"/> </g> </g> <g id="PathCollection_8"> @@ -2702,85 +2702,85 @@ C -2.236584 -1.29895 -2.5 -0.663008 -2.5 0 C -2.5 0.663008 -2.236584 1.29895 -1.767767 1.767767 C -1.29895 2.236584 -0.663008 2.5 0 2.5 z -" id="C5_0_0f975b7987"/> +" id="C5_0_a3035d72c9"/> </defs> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="710.64" xlink:href="#C5_0_0f975b7987" y="290.143163"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="710.64" xlink:href="#C5_0_a3035d72c9" y="290.143163"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="710.64" xlink:href="#C5_0_0f975b7987" y="271.05296"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="710.64" xlink:href="#C5_0_a3035d72c9" y="271.05296"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="705.390139" xlink:href="#C5_0_0f975b7987" y="271.020684"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="705.390139" xlink:href="#C5_0_a3035d72c9" y="271.020684"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="715.889732" xlink:href="#C5_0_0f975b7987" y="271.008152"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="715.889732" xlink:href="#C5_0_a3035d72c9" y="271.008152"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="710.64" xlink:href="#C5_0_0f975b7987" y="256.669448"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="710.64" xlink:href="#C5_0_a3035d72c9" y="256.669448"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="705.39031" xlink:href="#C5_0_0f975b7987" y="256.621285"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="705.39031" xlink:href="#C5_0_a3035d72c9" y="256.621285"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="715.888557" xlink:href="#C5_0_0f975b7987" y="256.565491"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="715.888557" xlink:href="#C5_0_a3035d72c9" y="256.565491"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="702.293321" xlink:href="#C5_0_0f975b7987" y="253.041079"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="702.293321" xlink:href="#C5_0_a3035d72c9" y="253.041079"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="710.64" xlink:href="#C5_0_0f975b7987" y="242.397727"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="710.64" xlink:href="#C5_0_a3035d72c9" y="242.397727"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="710.64" xlink:href="#C5_0_0f975b7987" y="236.301304"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="710.64" xlink:href="#C5_0_a3035d72c9" y="236.301304"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="705.390018" xlink:href="#C5_0_0f975b7987" y="236.289651"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="705.390018" xlink:href="#C5_0_a3035d72c9" y="236.289651"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="710.64" xlink:href="#C5_0_0f975b7987" y="225.669755"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="710.64" xlink:href="#C5_0_a3035d72c9" y="225.669755"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="706.662207" xlink:href="#C5_0_0f975b7987" y="222.776093"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="706.662207" xlink:href="#C5_0_a3035d72c9" y="222.776093"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="710.64" xlink:href="#C5_0_0f975b7987" y="215.218066"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="710.64" xlink:href="#C5_0_a3035d72c9" y="215.218066"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="705.677351" xlink:href="#C5_0_0f975b7987" y="213.771309"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="705.677351" xlink:href="#C5_0_a3035d72c9" y="213.771309"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="712.988203" xlink:href="#C5_0_0f975b7987" y="211.252465"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="712.988203" xlink:href="#C5_0_a3035d72c9" y="211.252465"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="708.247768" xlink:href="#C5_0_0f975b7987" y="209.346952"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="708.247768" xlink:href="#C5_0_a3035d72c9" y="209.346952"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="716.10116" xlink:href="#C5_0_0f975b7987" y="207.68215"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="716.10116" xlink:href="#C5_0_a3035d72c9" y="207.68215"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="710.64" xlink:href="#C5_0_0f975b7987" y="197.198782"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="710.64" xlink:href="#C5_0_a3035d72c9" y="197.198782"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="710.64" xlink:href="#C5_0_0f975b7987" y="190.657145"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="710.64" xlink:href="#C5_0_a3035d72c9" y="190.657145"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="705.420436" xlink:href="#C5_0_0f975b7987" y="190.180405"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="705.420436" xlink:href="#C5_0_a3035d72c9" y="190.180405"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="710.64" xlink:href="#C5_0_0f975b7987" y="178.222524"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="710.64" xlink:href="#C5_0_a3035d72c9" y="178.222524"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="710.64" xlink:href="#C5_0_0f975b7987" y="169.847016"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="710.64" xlink:href="#C5_0_a3035d72c9" y="169.847016"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="705.390063" xlink:href="#C5_0_0f975b7987" y="169.82537"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="705.390063" xlink:href="#C5_0_a3035d72c9" y="169.82537"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="715.360099" xlink:href="#C5_0_0f975b7987" y="167.90583"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="715.360099" xlink:href="#C5_0_a3035d72c9" y="167.90583"/> </g> - <g clip-path="url(#p37bf755c76)"> - <use style="fill:#dd8452;" x="710.64" xlink:href="#C5_0_0f975b7987" y="161.551986"/> + <g clip-path="url(#p19cc01c258)"> + <use style="fill:#dd8452;" x="710.64" xlink:href="#C5_0_a3035d72c9" y="161.551986"/> </g> </g> <g id="text_10"> @@ -3002,10 +3002,10 @@ C -3.464901 -2.012324 -3.872983 -1.027127 -3.872983 0 C -3.872983 1.027127 -3.464901 2.012324 -2.738613 2.738613 C -2.012324 3.464901 -1.027127 3.872983 0 3.872983 z -" id="ma1779d3c2f" style="stroke:#4c72b0;"/> +" id="m9b16eda43b" style="stroke:#4c72b0;"/> </defs> <g> - <use style="fill:#4c72b0;stroke:#4c72b0;" x="665.974063" xlink:href="#ma1779d3c2f" y="83.487344"/> + <use style="fill:#4c72b0;stroke:#4c72b0;" x="665.974063" xlink:href="#m9b16eda43b" y="83.487344"/> </g> </g> <g id="text_14"> @@ -3038,10 +3038,10 @@ C -3.464901 -2.012324 -3.872983 -1.027127 -3.872983 0 C -3.872983 1.027127 -3.464901 2.012324 -2.738613 2.738613 C -2.012324 3.464901 -1.027127 3.872983 0 3.872983 z -" id="mf97074e5f0" style="stroke:#dd8452;"/> +" id="mc9e39b2edf" style="stroke:#dd8452;"/> </defs> <g> - <use style="fill:#dd8452;stroke:#dd8452;" x="665.974063" xlink:href="#mf97074e5f0" y="99.310156"/> + <use style="fill:#dd8452;stroke:#dd8452;" x="665.974063" xlink:href="#mc9e39b2edf" y="99.310156"/> </g> </g> <g id="text_15"> @@ -3063,7 +3063,7 @@ z </g> </g> <defs> - <clipPath id="p37bf755c76"> + <clipPath id="p19cc01c258"> <rect height="326.16" width="669.6" x="108" y="51.84"/> </clipPath> </defs> diff --git a/usage/flu_dataset_log_scale_in_axes_strip.svg b/usage/flu_dataset_log_scale_in_axes_strip.svg index c2160f1..9c37087 100644 --- a/usage/flu_dataset_log_scale_in_axes_strip.svg +++ b/usage/flu_dataset_log_scale_in_axes_strip.svg @@ -115,7 +115,7 @@ Q 16.703125 10.84375 15.234375 8.453125 z " id="ArialMT-50"/> </defs> - <g style="fill:#262626;" transform="translate(185.204637 395.373594)scale(0.11 -0.11)"> + <g style="fill:#262626;" transform="translate(185.158394 395.373594)scale(0.11 -0.11)"> <use xlink:href="#ArialMT-72"/> <use x="72.216797" xlink:href="#ArialMT-51"/> <use x="127.832031" xlink:href="#ArialMT-78"/> @@ -139,7 +139,7 @@ L 37.25 71.875 z " id="ArialMT-49"/> </defs> - <g style="fill:#262626;" transform="translate(429.472379 395.373594)scale(0.11 -0.11)"> + <g style="fill:#262626;" transform="translate(429.107034 395.373594)scale(0.11 -0.11)"> <use xlink:href="#ArialMT-72"/> <use x="72.216797" xlink:href="#ArialMT-49"/> <use x="127.832031" xlink:href="#ArialMT-78"/> @@ -345,7 +345,7 @@ L 16.796875 33.0625 z " id="ArialMT-66"/> </defs> - <g style="fill:#262626;" transform="translate(660.284028 395.373594)scale(0.11 -0.11)"> + <g style="fill:#262626;" transform="translate(659.59958 395.373594)scale(0.11 -0.11)"> <use xlink:href="#ArialMT-73"/> <use x="27.783203" xlink:href="#ArialMT-110"/> <use x="83.398438" xlink:href="#ArialMT-102"/> @@ -510,7 +510,7 @@ z <g id="matplotlib.axis_2"> <g id="ytick_1"> <g id="line2d_1"> - <path clip-path="url(#p24d2ef447f)" d="M 108 355.173964 + <path clip-path="url(#p078fcafce0)" d="M 108 355.173964 L 777.6 355.173964 " style="fill:none;stroke:#cccccc;stroke-linecap:round;"/> </g> @@ -584,7 +584,7 @@ z </g> <g id="ytick_2"> <g id="line2d_2"> - <path clip-path="url(#p24d2ef447f)" d="M 108 258.658399 + <path clip-path="url(#p078fcafce0)" d="M 108 258.658399 L 777.6 258.658399 " style="fill:none;stroke:#cccccc;stroke-linecap:round;"/> </g> @@ -619,7 +619,7 @@ z </g> <g id="ytick_3"> <g id="line2d_3"> - <path clip-path="url(#p24d2ef447f)" d="M 108 162.142834 + <path clip-path="url(#p078fcafce0)" d="M 108 162.142834 L 777.6 162.142834 " style="fill:none;stroke:#cccccc;stroke-linecap:round;"/> </g> @@ -635,7 +635,7 @@ L 777.6 162.142834 </g> <g id="ytick_4"> <g id="line2d_4"> - <path clip-path="url(#p24d2ef447f)" d="M 108 65.627269 + <path clip-path="url(#p078fcafce0)" d="M 108 65.627269 L 777.6 65.627269 " style="fill:none;stroke:#cccccc;stroke-linecap:round;"/> </g> @@ -790,24 +790,24 @@ z <g id="PathCollection_1"/> <g id="PathCollection_2"/> <g id="line2d_5"> - <path clip-path="url(#p24d2ef447f)" d="M 394.679924 118.96632 -L 394.679924 113.094863 -L 492.387021 113.094863 -L 492.387021 118.96632 + <path clip-path="url(#p078fcafce0)" d="M 394.3784 118.96632 +L 394.3784 113.094863 +L 491.957856 113.094863 +L 491.957856 118.96632 " style="fill:none;stroke:#333333;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_6"> - <path clip-path="url(#p24d2ef447f)" d="M 150.412182 80.157043 -L 150.412182 74.285587 -L 248.119279 74.285587 -L 248.119279 80.157043 + <path clip-path="url(#p078fcafce0)" d="M 150.429759 80.157043 +L 150.429759 74.285587 +L 248.009215 74.285587 +L 248.009215 80.157043 " style="fill:none;stroke:#333333;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="line2d_7"> - <path clip-path="url(#p24d2ef447f)" d="M 638.947667 132.836253 -L 638.947667 126.964797 -L 736.654764 126.964797 -L 736.654764 132.836253 + <path clip-path="url(#p078fcafce0)" d="M 638.32704 132.836253 +L 638.32704 126.964797 +L 735.906496 126.964797 +L 735.906496 132.836253 " style="fill:none;stroke:#333333;stroke-linecap:round;stroke-width:1.5;"/> </g> <g id="patch_3"> @@ -842,589 +842,589 @@ C -2.236584 -1.29895 -2.5 -0.663008 -2.5 0 C -2.5 0.663008 -2.236584 1.29895 -1.767767 1.767767 C -1.29895 2.236584 -0.663008 2.5 0 2.5 z -" id="C0_0_2a88d1c6b0"/> +" id="C0_0_af5b03a438"/> </defs> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="147.060601" xlink:href="#C0_0_2a88d1c6b0" y="234.464917"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="147.384047" xlink:href="#C0_0_af5b03a438" y="234.464917"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="153.728993" xlink:href="#C0_0_2a88d1c6b0" y="234.469355"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="156.319934" xlink:href="#C0_0_af5b03a438" y="234.469355"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="142.975993" xlink:href="#C0_0_2a88d1c6b0" y="186.01941"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="143.707055" xlink:href="#C0_0_af5b03a438" y="186.01941"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="145.058984" xlink:href="#C0_0_2a88d1c6b0" y="217.469375"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="158.997726" xlink:href="#C0_0_af5b03a438" y="217.469375"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="157.886778" xlink:href="#C0_0_2a88d1c6b0" y="181.602768"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="146.26079" xlink:href="#C0_0_af5b03a438" y="181.602768"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="141.682206" xlink:href="#C0_0_2a88d1c6b0" y="206.660532"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="157.257495" xlink:href="#C0_0_af5b03a438" y="206.660532"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="155.013194" xlink:href="#C0_0_2a88d1c6b0" y="217.469308"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="158.3782" xlink:href="#C0_0_af5b03a438" y="217.469308"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="156.076163" xlink:href="#C0_0_2a88d1c6b0" y="159.361091"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="152.719874" xlink:href="#C0_0_af5b03a438" y="159.361091"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="152.644414" xlink:href="#C0_0_2a88d1c6b0" y="200.097852"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="142.68083" xlink:href="#C0_0_af5b03a438" y="200.097852"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="162.204833" xlink:href="#C0_0_2a88d1c6b0" y="237.168751"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="156.736524" xlink:href="#C0_0_af5b03a438" y="237.168751"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="144.539814" xlink:href="#C0_0_2a88d1c6b0" y="204.470561"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="142.793119" xlink:href="#C0_0_af5b03a438" y="204.470561"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="145.589841" xlink:href="#C0_0_2a88d1c6b0" y="170.604131"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="139.50113" xlink:href="#C0_0_af5b03a438" y="170.604131"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="161.521618" xlink:href="#C0_0_2a88d1c6b0" y="148.828987"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="154.13707" xlink:href="#C0_0_af5b03a438" y="148.828987"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="160.930571" xlink:href="#C0_0_2a88d1c6b0" y="161.142385"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="152.786131" xlink:href="#C0_0_af5b03a438" y="161.142385"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="142.043267" xlink:href="#C0_0_2a88d1c6b0" y="261.123162"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="151.352819" xlink:href="#C0_0_af5b03a438" y="261.123162"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="162.581739" xlink:href="#C0_0_2a88d1c6b0" y="213.05278"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="154.503085" xlink:href="#C0_0_af5b03a438" y="213.05278"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="156.507004" xlink:href="#C0_0_2a88d1c6b0" y="246.52338"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="153.029995" xlink:href="#C0_0_af5b03a438" y="246.52338"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="139.967855" xlink:href="#C0_0_2a88d1c6b0" y="202.518813"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="141.345687" xlink:href="#C0_0_af5b03a438" y="202.518813"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="149.960264" xlink:href="#C0_0_2a88d1c6b0" y="263.518057"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="142.718516" xlink:href="#C0_0_af5b03a438" y="263.518057"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="139.778789" xlink:href="#C0_0_2a88d1c6b0" y="145.788884"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="157.425812" xlink:href="#C0_0_af5b03a438" y="145.788884"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="144.746448" xlink:href="#C0_0_2a88d1c6b0" y="199.222027"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="138.52142" xlink:href="#C0_0_af5b03a438" y="199.222027"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="139.087382" xlink:href="#C0_0_2a88d1c6b0" y="198.762645"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="158.958474" xlink:href="#C0_0_af5b03a438" y="198.762645"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="150.118409" xlink:href="#C0_0_2a88d1c6b0" y="211.870604"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="138.610994" xlink:href="#C0_0_af5b03a438" y="211.870604"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="147.871271" xlink:href="#C0_0_2a88d1c6b0" y="203.015022"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="149.926606" xlink:href="#C0_0_af5b03a438" y="203.015022"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="158.517174" xlink:href="#C0_0_2a88d1c6b0" y="246.521469"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="162.583694" xlink:href="#C0_0_af5b03a438" y="246.521469"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="147.357868" xlink:href="#C0_0_2a88d1c6b0" y="185.060335"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="149.861194" xlink:href="#C0_0_af5b03a438" y="185.060335"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="148.132868" xlink:href="#C0_0_2a88d1c6b0" y="199.222339"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="150.875252" xlink:href="#C0_0_af5b03a438" y="199.222339"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="159.911231" xlink:href="#C0_0_2a88d1c6b0" y="292.574965"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="151.180541" xlink:href="#C0_0_af5b03a438" y="292.574965"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="152.844267" xlink:href="#C0_0_2a88d1c6b0" y="278.118541"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="153.250301" xlink:href="#C0_0_af5b03a438" y="278.118541"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="162.550598" xlink:href="#C0_0_2a88d1c6b0" y="288.925258"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="144.932173" xlink:href="#C0_0_af5b03a438" y="288.925258"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="144.998518" xlink:href="#C0_0_2a88d1c6b0" y="197.117394"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="141.24368" xlink:href="#C0_0_af5b03a438" y="197.117394"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="158.082028" xlink:href="#C0_0_2a88d1c6b0" y="293.722513"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="148.861059" xlink:href="#C0_0_af5b03a438" y="293.722513"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="151.318422" xlink:href="#C0_0_2a88d1c6b0" y="244.997663"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="160.2273" xlink:href="#C0_0_af5b03a438" y="244.997663"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="158.55584" xlink:href="#C0_0_2a88d1c6b0" y="192.062031"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="149.797892" xlink:href="#C0_0_af5b03a438" y="192.062031"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="140.910831" xlink:href="#C0_0_2a88d1c6b0" y="236.731089"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="149.177982" xlink:href="#C0_0_af5b03a438" y="236.731089"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="140.983768" xlink:href="#C0_0_2a88d1c6b0" y="326.564074"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="156.055941" xlink:href="#C0_0_af5b03a438" y="326.564074"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="140.314076" xlink:href="#C0_0_2a88d1c6b0" y="206.072603"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="148.029634" xlink:href="#C0_0_af5b03a438" y="206.072603"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="162.242098" xlink:href="#C0_0_2a88d1c6b0" y="287.636006"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="146.603045" xlink:href="#C0_0_af5b03a438" y="287.636006"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="148.970788" xlink:href="#C0_0_2a88d1c6b0" y="186.019413"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="153.228295" xlink:href="#C0_0_af5b03a438" y="186.019413"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="160.718673" xlink:href="#C0_0_2a88d1c6b0" y="234.463233"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="142.488332" xlink:href="#C0_0_af5b03a438" y="234.463233"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="148.875736" xlink:href="#C0_0_2a88d1c6b0" y="204.465046"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="157.532766" xlink:href="#C0_0_af5b03a438" y="204.465046"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="147.786354" xlink:href="#C0_0_2a88d1c6b0" y="249.057643"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="149.075558" xlink:href="#C0_0_af5b03a438" y="249.057643"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="141.687814" xlink:href="#C0_0_2a88d1c6b0" y="164.607881"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="146.980851" xlink:href="#C0_0_af5b03a438" y="164.607881"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="157.625102" xlink:href="#C0_0_2a88d1c6b0" y="196.553038"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="150.781267" xlink:href="#C0_0_af5b03a438" y="196.553038"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="141.588186" xlink:href="#C0_0_2a88d1c6b0" y="156.468577"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="143.965871" xlink:href="#C0_0_af5b03a438" y="156.468577"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="161.339998" xlink:href="#C0_0_2a88d1c6b0" y="222.407449"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="150.09534" xlink:href="#C0_0_af5b03a438" y="222.407449"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="146.873728" xlink:href="#C0_0_2a88d1c6b0" y="280.51449"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="144.404388" xlink:href="#C0_0_af5b03a438" y="280.51449"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="138.436364" xlink:href="#C0_0_2a88d1c6b0" y="193.661519"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="148.761673" xlink:href="#C0_0_af5b03a438" y="193.661519"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="153.227196" xlink:href="#C0_0_2a88d1c6b0" y="280.514507"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="140.397152" xlink:href="#C0_0_af5b03a438" y="280.514507"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="149.526654" xlink:href="#C0_0_2a88d1c6b0" y="163.008982"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="139.1638" xlink:href="#C0_0_af5b03a438" y="163.008982"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="144.694992" xlink:href="#C0_0_2a88d1c6b0" y="307.172431"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="148.797431" xlink:href="#C0_0_af5b03a438" y="307.172431"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="160.631469" xlink:href="#C0_0_2a88d1c6b0" y="240.518048"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="148.936914" xlink:href="#C0_0_af5b03a438" y="240.518048"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="154.109667" xlink:href="#C0_0_2a88d1c6b0" y="226.166435"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="139.513321" xlink:href="#C0_0_af5b03a438" y="226.166435"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="157.974153" xlink:href="#C0_0_2a88d1c6b0" y="199.223789"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="147.412371" xlink:href="#C0_0_af5b03a438" y="199.223789"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="138.949906" xlink:href="#C0_0_2a88d1c6b0" y="307.172214"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="159.539624" xlink:href="#C0_0_af5b03a438" y="307.172214"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="150.25069" xlink:href="#C0_0_2a88d1c6b0" y="271.928388"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="145.317523" xlink:href="#C0_0_af5b03a438" y="271.928388"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="152.664616" xlink:href="#C0_0_2a88d1c6b0" y="196.99711"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="158.404104" xlink:href="#C0_0_af5b03a438" y="196.99711"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="143.546974" xlink:href="#C0_0_2a88d1c6b0" y="295.926683"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="143.351895" xlink:href="#C0_0_af5b03a438" y="295.926683"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="157.664136" xlink:href="#C0_0_2a88d1c6b0" y="191.114521"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="150.510906" xlink:href="#C0_0_af5b03a438" y="191.114521"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="155.430434" xlink:href="#C0_0_2a88d1c6b0" y="194.87592"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="153.889938" xlink:href="#C0_0_af5b03a438" y="194.87592"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="147.302341" xlink:href="#C0_0_2a88d1c6b0" y="364.655781"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="159.276902" xlink:href="#C0_0_af5b03a438" y="364.655781"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="149.465308" xlink:href="#C0_0_2a88d1c6b0" y="197.452892"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="141.670173" xlink:href="#C0_0_af5b03a438" y="197.452892"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="147.121025" xlink:href="#C0_0_2a88d1c6b0" y="181.294506"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="148.233078" xlink:href="#C0_0_af5b03a438" y="181.294506"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="141.270149" xlink:href="#C0_0_2a88d1c6b0" y="317.974952"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="157.678147" xlink:href="#C0_0_af5b03a438" y="317.974952"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="154.585335" xlink:href="#C0_0_2a88d1c6b0" y="269.984711"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="147.294209" xlink:href="#C0_0_af5b03a438" y="269.984711"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="147.61082" xlink:href="#C0_0_2a88d1c6b0" y="164.961136"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="158.442183" xlink:href="#C0_0_af5b03a438" y="164.961136"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="143.649969" xlink:href="#C0_0_2a88d1c6b0" y="202.517314"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="157.570921" xlink:href="#C0_0_af5b03a438" y="202.517314"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="142.022999" xlink:href="#C0_0_2a88d1c6b0" y="292.574949"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="151.838554" xlink:href="#C0_0_af5b03a438" y="292.574949"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="154.283824" xlink:href="#C0_0_2a88d1c6b0" y="171.426949"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="146.673253" xlink:href="#C0_0_af5b03a438" y="171.426949"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="149.079039" xlink:href="#C0_0_2a88d1c6b0" y="192.419882"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="150.847812" xlink:href="#C0_0_af5b03a438" y="192.419882"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="144.040599" xlink:href="#C0_0_2a88d1c6b0" y="195.839019"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="145.619372" xlink:href="#C0_0_af5b03a438" y="195.839019"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="151.537642" xlink:href="#C0_0_2a88d1c6b0" y="258.578438"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="151.563431" xlink:href="#C0_0_af5b03a438" y="258.578438"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="155.912469" xlink:href="#C0_0_2a88d1c6b0" y="249.071341"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="142.147195" xlink:href="#C0_0_af5b03a438" y="249.071341"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="161.424916" xlink:href="#C0_0_2a88d1c6b0" y="231.577382"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="143.816134" xlink:href="#C0_0_af5b03a438" y="231.577382"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="159.442792" xlink:href="#C0_0_2a88d1c6b0" y="288.92293"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="161.670412" xlink:href="#C0_0_af5b03a438" y="288.92293"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="153.839016" xlink:href="#C0_0_2a88d1c6b0" y="196.996728"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="149.388348" xlink:href="#C0_0_af5b03a438" y="196.996728"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="142.41653" xlink:href="#C0_0_2a88d1c6b0" y="248.570576"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="157.82596" xlink:href="#C0_0_af5b03a438" y="248.570576"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="152.904397" xlink:href="#C0_0_2a88d1c6b0" y="161.149028"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="161.27347" xlink:href="#C0_0_af5b03a438" y="161.149028"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="142.754295" xlink:href="#C0_0_2a88d1c6b0" y="150.012056"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="151.687053" xlink:href="#C0_0_af5b03a438" y="150.012056"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="146.252225" xlink:href="#C0_0_2a88d1c6b0" y="211.011584"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="153.792147" xlink:href="#C0_0_af5b03a438" y="211.011584"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="146.916709" xlink:href="#C0_0_2a88d1c6b0" y="263.516866"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="156.430205" xlink:href="#C0_0_af5b03a438" y="263.516866"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="147.354924" xlink:href="#C0_0_2a88d1c6b0" y="269.978278"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="161.750727" xlink:href="#C0_0_af5b03a438" y="269.978278"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="159.038186" xlink:href="#C0_0_2a88d1c6b0" y="274.69867"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="153.586616" xlink:href="#C0_0_af5b03a438" y="274.69867"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="146.835508" xlink:href="#C0_0_2a88d1c6b0" y="171.425088"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="147.166331" xlink:href="#C0_0_af5b03a438" y="171.425088"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="146.92867" xlink:href="#C0_0_2a88d1c6b0" y="181.959734"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="161.412731" xlink:href="#C0_0_af5b03a438" y="181.959734"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="157.273868" xlink:href="#C0_0_2a88d1c6b0" y="166.030933"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="160.349209" xlink:href="#C0_0_af5b03a438" y="166.030933"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="157.569593" xlink:href="#C0_0_2a88d1c6b0" y="307.172933"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="156.271312" xlink:href="#C0_0_af5b03a438" y="307.172933"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="161.53809" xlink:href="#C0_0_2a88d1c6b0" y="204.472191"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="155.112727" xlink:href="#C0_0_af5b03a438" y="204.472191"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="154.943557" xlink:href="#C0_0_2a88d1c6b0" y="188.412611"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="141.475733" xlink:href="#C0_0_af5b03a438" y="188.412611"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="144.79596" xlink:href="#C0_0_2a88d1c6b0" y="169.150455"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="161.580107" xlink:href="#C0_0_af5b03a438" y="169.150455"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="138.787223" xlink:href="#C0_0_2a88d1c6b0" y="292.556326"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="153.016583" xlink:href="#C0_0_af5b03a438" y="292.556326"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="140.933693" xlink:href="#C0_0_2a88d1c6b0" y="190.002095"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="143.433511" xlink:href="#C0_0_af5b03a438" y="190.002095"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="152.878676" xlink:href="#C0_0_2a88d1c6b0" y="152.896666"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="160.626923" xlink:href="#C0_0_af5b03a438" y="152.896666"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="146.383385" xlink:href="#C0_0_2a88d1c6b0" y="220.011786"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="147.424477" xlink:href="#C0_0_af5b03a438" y="220.011786"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="145.250027" xlink:href="#C0_0_2a88d1c6b0" y="171.418987"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="139.708714" xlink:href="#C0_0_af5b03a438" y="171.418987"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="148.825657" xlink:href="#C0_0_2a88d1c6b0" y="257.697832"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="148.189933" xlink:href="#C0_0_af5b03a438" y="257.697832"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="141.757968" xlink:href="#C0_0_2a88d1c6b0" y="193.349058"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="140.811951" xlink:href="#C0_0_af5b03a438" y="193.349058"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="148.174666" xlink:href="#C0_0_2a88d1c6b0" y="292.577062"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="146.650705" xlink:href="#C0_0_af5b03a438" y="292.577062"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="146.108432" xlink:href="#C0_0_2a88d1c6b0" y="136.751329"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="143.393213" xlink:href="#C0_0_af5b03a438" y="136.751329"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="155.283425" xlink:href="#C0_0_2a88d1c6b0" y="205.410931"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="144.519781" xlink:href="#C0_0_af5b03a438" y="205.410931"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="159.384261" xlink:href="#C0_0_2a88d1c6b0" y="173.812482"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="145.901263" xlink:href="#C0_0_af5b03a438" y="173.812482"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="155.429041" xlink:href="#C0_0_2a88d1c6b0" y="215.077353"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="138.66636" xlink:href="#C0_0_af5b03a438" y="215.077353"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="159.555641" xlink:href="#C0_0_2a88d1c6b0" y="263.513214"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="138.436364" xlink:href="#C0_0_af5b03a438" y="263.513214"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="145.199378" xlink:href="#C0_0_2a88d1c6b0" y="239.386134"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="140.403664" xlink:href="#C0_0_af5b03a438" y="239.386134"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="144.782713" xlink:href="#C0_0_2a88d1c6b0" y="186.546072"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="154.308511" xlink:href="#C0_0_af5b03a438" y="186.546072"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="150.03091" xlink:href="#C0_0_2a88d1c6b0" y="210.640421"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="161.992634" xlink:href="#C0_0_af5b03a438" y="210.640421"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="138.812507" xlink:href="#C0_0_2a88d1c6b0" y="197.116231"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="144.834949" xlink:href="#C0_0_af5b03a438" y="197.116231"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="148.290301" xlink:href="#C0_0_2a88d1c6b0" y="251.454711"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="145.577076" xlink:href="#C0_0_af5b03a438" y="251.454711"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="141.310817" xlink:href="#C0_0_2a88d1c6b0" y="169.146493"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="154.492914" xlink:href="#C0_0_af5b03a438" y="169.146493"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="155.253548" xlink:href="#C0_0_2a88d1c6b0" y="156.000344"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="151.712636" xlink:href="#C0_0_af5b03a438" y="156.000344"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="138.571849" xlink:href="#C0_0_2a88d1c6b0" y="278.111605"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="156.859975" xlink:href="#C0_0_af5b03a438" y="278.111605"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="161.91792" xlink:href="#C0_0_2a88d1c6b0" y="199.648095"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="152.633655" xlink:href="#C0_0_af5b03a438" y="199.648095"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="153.636897" xlink:href="#C0_0_2a88d1c6b0" y="246.521571"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="148.540912" xlink:href="#C0_0_af5b03a438" y="246.521571"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="154.861214" xlink:href="#C0_0_2a88d1c6b0" y="188.40958"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="144.146144" xlink:href="#C0_0_af5b03a438" y="188.40958"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="142.491003" xlink:href="#C0_0_2a88d1c6b0" y="211.002179"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="142.401129" xlink:href="#C0_0_af5b03a438" y="211.002179"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="158.600038" xlink:href="#C0_0_2a88d1c6b0" y="232.060628"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="142.137987" xlink:href="#C0_0_af5b03a438" y="232.060628"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="140.747655" xlink:href="#C0_0_2a88d1c6b0" y="211.855489"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="152.06608" xlink:href="#C0_0_af5b03a438" y="211.855489"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="159.948904" xlink:href="#C0_0_2a88d1c6b0" y="151.208438"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="138.88828" xlink:href="#C0_0_af5b03a438" y="151.208438"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="158.672272" xlink:href="#C0_0_2a88d1c6b0" y="252.600066"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="147.99149" xlink:href="#C0_0_af5b03a438" y="252.600066"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="153.011734" xlink:href="#C0_0_2a88d1c6b0" y="271.924558"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="160.654869" xlink:href="#C0_0_af5b03a438" y="271.924558"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="148.842531" xlink:href="#C0_0_2a88d1c6b0" y="179.059415"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="149.069864" xlink:href="#C0_0_af5b03a438" y="179.059415"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="151.3317" xlink:href="#C0_0_2a88d1c6b0" y="249.065509"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="154.346603" xlink:href="#C0_0_af5b03a438" y="249.065509"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="149.740745" xlink:href="#C0_0_2a88d1c6b0" y="263.520399"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="141.220871" xlink:href="#C0_0_af5b03a438" y="263.520399"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="144.611396" xlink:href="#C0_0_2a88d1c6b0" y="263.52343"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="140.956356" xlink:href="#C0_0_af5b03a438" y="263.52343"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="143.521923" xlink:href="#C0_0_2a88d1c6b0" y="261.124522"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="160.717068" xlink:href="#C0_0_af5b03a438" y="261.124522"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="142.982425" xlink:href="#C0_0_2a88d1c6b0" y="307.174143"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="147.323679" xlink:href="#C0_0_af5b03a438" y="307.174143"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="161.033468" xlink:href="#C0_0_2a88d1c6b0" y="200.477101"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="151.793433" xlink:href="#C0_0_af5b03a438" y="200.477101"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="147.722977" xlink:href="#C0_0_2a88d1c6b0" y="269.986637"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="140.252609" xlink:href="#C0_0_af5b03a438" y="269.986637"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="150.856048" xlink:href="#C0_0_2a88d1c6b0" y="320.750698"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="155.47779" xlink:href="#C0_0_af5b03a438" y="320.750698"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="146.556013" xlink:href="#C0_0_2a88d1c6b0" y="278.124431"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="149.053455" xlink:href="#C0_0_af5b03a438" y="278.124431"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="161.673047" xlink:href="#C0_0_2a88d1c6b0" y="273.978125"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="144.265826" xlink:href="#C0_0_af5b03a438" y="273.978125"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="152.082481" xlink:href="#C0_0_2a88d1c6b0" y="235.616775"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="159.079075" xlink:href="#C0_0_af5b03a438" y="235.616775"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="145.902122" xlink:href="#C0_0_2a88d1c6b0" y="210.352087"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="158.068443" xlink:href="#C0_0_af5b03a438" y="210.352087"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="138.679104" xlink:href="#C0_0_2a88d1c6b0" y="199.663269"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="139.72174" xlink:href="#C0_0_af5b03a438" y="199.663269"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="148.26969" xlink:href="#C0_0_2a88d1c6b0" y="164.976889"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="144.302111" xlink:href="#C0_0_af5b03a438" y="164.976889"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="152.836815" xlink:href="#C0_0_2a88d1c6b0" y="146.019431"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="144.638777" xlink:href="#C0_0_af5b03a438" y="146.019431"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="152.426918" xlink:href="#C0_0_2a88d1c6b0" y="263.537703"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="159.055351" xlink:href="#C0_0_af5b03a438" y="263.537703"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="142.416817" xlink:href="#C0_0_2a88d1c6b0" y="211.006731"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="160.205048" xlink:href="#C0_0_af5b03a438" y="211.006731"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="153.3067" xlink:href="#C0_0_2a88d1c6b0" y="148.861019"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="150.968999" xlink:href="#C0_0_af5b03a438" y="148.861019"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="152.086506" xlink:href="#C0_0_2a88d1c6b0" y="242.889012"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="150.013456" xlink:href="#C0_0_af5b03a438" y="242.889012"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="143.634083" xlink:href="#C0_0_2a88d1c6b0" y="240.262686"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="138.659046" xlink:href="#C0_0_af5b03a438" y="240.262686"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="152.888113" xlink:href="#C0_0_2a88d1c6b0" y="229.527227"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="150.585288" xlink:href="#C0_0_af5b03a438" y="229.527227"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="153.931949" xlink:href="#C0_0_2a88d1c6b0" y="263.524022"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="155.072047" xlink:href="#C0_0_af5b03a438" y="263.524022"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="156.698164" xlink:href="#C0_0_2a88d1c6b0" y="199.418433"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="160.92629" xlink:href="#C0_0_af5b03a438" y="199.418433"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="161.473311" xlink:href="#C0_0_2a88d1c6b0" y="257.269739"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="142.238979" xlink:href="#C0_0_af5b03a438" y="257.269739"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="157.028195" xlink:href="#C0_0_2a88d1c6b0" y="288.935701"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="144.811952" xlink:href="#C0_0_af5b03a438" y="288.935701"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="159.767095" xlink:href="#C0_0_2a88d1c6b0" y="249.077259"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="139.111581" xlink:href="#C0_0_af5b03a438" y="249.077259"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="142.34957" xlink:href="#C0_0_2a88d1c6b0" y="196.84005"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="145.95522" xlink:href="#C0_0_af5b03a438" y="196.84005"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="157.140446" xlink:href="#C0_0_2a88d1c6b0" y="263.531794"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="160.646619" xlink:href="#C0_0_af5b03a438" y="263.531794"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="161.055627" xlink:href="#C0_0_2a88d1c6b0" y="211.011912"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="162.535656" xlink:href="#C0_0_af5b03a438" y="211.011912"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="156.812933" xlink:href="#C0_0_2a88d1c6b0" y="253.288159"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="143.256002" xlink:href="#C0_0_af5b03a438" y="253.288159"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="144.04369" xlink:href="#C0_0_2a88d1c6b0" y="217.482077"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="147.182743" xlink:href="#C0_0_af5b03a438" y="217.482077"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="156.572331" xlink:href="#C0_0_2a88d1c6b0" y="249.077437"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="161.523331" xlink:href="#C0_0_af5b03a438" y="249.077437"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="158.332349" xlink:href="#C0_0_2a88d1c6b0" y="307.185539"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="157.423385" xlink:href="#C0_0_af5b03a438" y="307.185539"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="159.330613" xlink:href="#C0_0_2a88d1c6b0" y="154.424025"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="162.548681" xlink:href="#C0_0_af5b03a438" y="154.424025"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="140.208615" xlink:href="#C0_0_2a88d1c6b0" y="192.146247"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="160.065648" xlink:href="#C0_0_af5b03a438" y="192.146247"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="150.967225" xlink:href="#C0_0_2a88d1c6b0" y="200.78484"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="158.695069" xlink:href="#C0_0_af5b03a438" y="200.78484"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="138.535192" xlink:href="#C0_0_2a88d1c6b0" y="307.171497"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="159.030584" xlink:href="#C0_0_af5b03a438" y="307.171497"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="142.703322" xlink:href="#C0_0_2a88d1c6b0" y="317.984868"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="153.194575" xlink:href="#C0_0_af5b03a438" y="317.984868"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="140.435589" xlink:href="#C0_0_2a88d1c6b0" y="182.812804"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="146.287527" xlink:href="#C0_0_af5b03a438" y="182.812804"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="161.393653" xlink:href="#C0_0_2a88d1c6b0" y="292.584627"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="144.868187" xlink:href="#C0_0_af5b03a438" y="292.584627"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="154.259142" xlink:href="#C0_0_2a88d1c6b0" y="225.116796"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="155.954442" xlink:href="#C0_0_af5b03a438" y="225.116796"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="161.836345" xlink:href="#C0_0_2a88d1c6b0" y="159.366585"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="148.357881" xlink:href="#C0_0_af5b03a438" y="159.366585"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="153.144964" xlink:href="#C0_0_2a88d1c6b0" y="156.018214"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="161.821635" xlink:href="#C0_0_af5b03a438" y="156.018214"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="157.287056" xlink:href="#C0_0_2a88d1c6b0" y="278.130212"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="158.574791" xlink:href="#C0_0_af5b03a438" y="278.130212"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="153.362297" xlink:href="#C0_0_2a88d1c6b0" y="207.372381"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="160.998279" xlink:href="#C0_0_af5b03a438" y="207.372381"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="159.552743" xlink:href="#C0_0_2a88d1c6b0" y="291.703424"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="154.184456" xlink:href="#C0_0_af5b03a438" y="291.703424"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="155.765416" xlink:href="#C0_0_2a88d1c6b0" y="293.73309"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="140.496466" xlink:href="#C0_0_af5b03a438" y="293.73309"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="144.456313" xlink:href="#C0_0_2a88d1c6b0" y="268.237954"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="154.139575" xlink:href="#C0_0_af5b03a438" y="268.237954"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="159.571159" xlink:href="#C0_0_2a88d1c6b0" y="261.134578"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="161.580194" xlink:href="#C0_0_af5b03a438" y="261.134578"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="150.166371" xlink:href="#C0_0_2a88d1c6b0" y="176.361807"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="157.141028" xlink:href="#C0_0_af5b03a438" y="176.361807"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="139.233875" xlink:href="#C0_0_2a88d1c6b0" y="188.417584"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="155.561754" xlink:href="#C0_0_af5b03a438" y="188.417584"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="143.021597" xlink:href="#C0_0_2a88d1c6b0" y="293.720295"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="149.363576" xlink:href="#C0_0_af5b03a438" y="293.720295"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="143.890492" xlink:href="#C0_0_2a88d1c6b0" y="197.375617"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="146.512782" xlink:href="#C0_0_af5b03a438" y="197.375617"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="160.474152" xlink:href="#C0_0_2a88d1c6b0" y="234.476448"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="149.952187" xlink:href="#C0_0_af5b03a438" y="234.476448"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="154.369352" xlink:href="#C0_0_2a88d1c6b0" y="190.97031"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="138.905186" xlink:href="#C0_0_af5b03a438" y="190.97031"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="141.943243" xlink:href="#C0_0_2a88d1c6b0" y="307.176938"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="159.088023" xlink:href="#C0_0_af5b03a438" y="307.176938"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="151.113188" xlink:href="#C0_0_2a88d1c6b0" y="193.675276"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="138.557088" xlink:href="#C0_0_af5b03a438" y="193.675276"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="155.950648" xlink:href="#C0_0_2a88d1c6b0" y="326.568957"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="139.493833" xlink:href="#C0_0_af5b03a438" y="326.568957"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="159.531119" xlink:href="#C0_0_2a88d1c6b0" y="211.017531"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="146.092482" xlink:href="#C0_0_af5b03a438" y="211.017531"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="152.040567" xlink:href="#C0_0_2a88d1c6b0" y="167.00589"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="138.937364" xlink:href="#C0_0_af5b03a438" y="167.00589"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="139.968931" xlink:href="#C0_0_2a88d1c6b0" y="335.61745"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="142.151369" xlink:href="#C0_0_af5b03a438" y="335.61745"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="161.062189" xlink:href="#C0_0_2a88d1c6b0" y="240.92954"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="162.220076" xlink:href="#C0_0_af5b03a438" y="240.92954"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="143.707986" xlink:href="#C0_0_2a88d1c6b0" y="250.523607"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="144.323115" xlink:href="#C0_0_af5b03a438" y="250.523607"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="141.303205" xlink:href="#C0_0_2a88d1c6b0" y="185.065536"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="144.448734" xlink:href="#C0_0_af5b03a438" y="185.065536"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="139.400884" xlink:href="#C0_0_2a88d1c6b0" y="211.02224"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="151.66963" xlink:href="#C0_0_af5b03a438" y="211.02224"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="153.628071" xlink:href="#C0_0_2a88d1c6b0" y="307.18565"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="141.480561" xlink:href="#C0_0_af5b03a438" y="307.18565"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="150.95575" xlink:href="#C0_0_2a88d1c6b0" y="278.124"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="147.987779" xlink:href="#C0_0_af5b03a438" y="278.124"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="154.442719" xlink:href="#C0_0_2a88d1c6b0" y="259.107884"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="159.881403" xlink:href="#C0_0_af5b03a438" y="259.107884"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="151.572956" xlink:href="#C0_0_2a88d1c6b0" y="173.96873"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="152.128647" xlink:href="#C0_0_af5b03a438" y="173.96873"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="145.647245" xlink:href="#C0_0_2a88d1c6b0" y="286.319309"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="149.540781" xlink:href="#C0_0_af5b03a438" y="286.319309"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="157.132116" xlink:href="#C0_0_2a88d1c6b0" y="278.125323"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="151.313386" xlink:href="#C0_0_af5b03a438" y="278.125323"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="148.712155" xlink:href="#C0_0_2a88d1c6b0" y="211.007792"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="154.006386" xlink:href="#C0_0_af5b03a438" y="211.007792"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="145.900809" xlink:href="#C0_0_2a88d1c6b0" y="205.412973"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="142.844745" xlink:href="#C0_0_af5b03a438" y="205.412973"/> </g> </g> <g id="PathCollection_4"> @@ -1439,592 +1439,592 @@ C -2.236584 -1.29895 -2.5 -0.663008 -2.5 0 C -2.5 0.663008 -2.236584 1.29895 -1.767767 1.767767 C -1.29895 2.236584 -0.663008 2.5 0 2.5 z -" id="C1_0_82f4c9fa18"/> +" id="C1_0_f6794d661a"/> </defs> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="258.379252" xlink:href="#C1_0_82f4c9fa18" y="207.50588"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="246.094122" xlink:href="#C1_0_f6794d661a" y="207.50588"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="247.850998" xlink:href="#C1_0_82f4c9fa18" y="253.544249"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="238.182833" xlink:href="#C1_0_f6794d661a" y="253.544249"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="253.281851" xlink:href="#C1_0_82f4c9fa18" y="149.397766"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="255.283893" xlink:href="#C1_0_f6794d661a" y="149.397766"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="240.56049" xlink:href="#C1_0_82f4c9fa18" y="249.90831"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="255.015429" xlink:href="#C1_0_f6794d661a" y="249.90831"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="241.408416" xlink:href="#C1_0_82f4c9fa18" y="193.052371"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="250.339336" xlink:href="#C1_0_f6794d661a" y="193.052371"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="259.934517" xlink:href="#C1_0_82f4c9fa18" y="200.692182"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="240.075278" xlink:href="#C1_0_f6794d661a" y="200.692182"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="239.254965" xlink:href="#C1_0_82f4c9fa18" y="230.96294"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="236.212083" xlink:href="#C1_0_f6794d661a" y="230.96294"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="236.756359" xlink:href="#C1_0_82f4c9fa18" y="207.506037"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="256.243371" xlink:href="#C1_0_f6794d661a" y="207.506037"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="254.206824" xlink:href="#C1_0_82f4c9fa18" y="180.992598"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="241.909136" xlink:href="#C1_0_f6794d661a" y="180.992598"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="242.571711" xlink:href="#C1_0_82f4c9fa18" y="213.590998"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="246.136914" xlink:href="#C1_0_f6794d661a" y="213.590998"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="240.011511" xlink:href="#C1_0_82f4c9fa18" y="213.106325"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="256.281903" xlink:href="#C1_0_f6794d661a" y="213.106325"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="252.595156" xlink:href="#C1_0_82f4c9fa18" y="191.79585"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="251.289133" xlink:href="#C1_0_f6794d661a" y="191.79585"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="250.314214" xlink:href="#C1_0_82f4c9fa18" y="171.643345"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="250.939533" xlink:href="#C1_0_f6794d661a" y="171.643345"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="242.541666" xlink:href="#C1_0_82f4c9fa18" y="172.84981"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="247.688815" xlink:href="#C1_0_f6794d661a" y="172.84981"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="249.449495" xlink:href="#C1_0_82f4c9fa18" y="137.346191"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="242.59829" xlink:href="#C1_0_f6794d661a" y="137.346191"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="259.575594" xlink:href="#C1_0_82f4c9fa18" y="210.046514"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="240.168943" xlink:href="#C1_0_f6794d661a" y="210.046514"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="250.967417" xlink:href="#C1_0_82f4c9fa18" y="195.448058"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="244.072482" xlink:href="#C1_0_f6794d661a" y="195.448058"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="245.311474" xlink:href="#C1_0_82f4c9fa18" y="207.506058"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="256.037815" xlink:href="#C1_0_f6794d661a" y="207.506058"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="257.977537" xlink:href="#C1_0_82f4c9fa18" y="155.859397"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="251.371631" xlink:href="#C1_0_f6794d661a" y="155.859397"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="249.741509" xlink:href="#C1_0_82f4c9fa18" y="167.644903"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="254.242134" xlink:href="#C1_0_f6794d661a" y="167.644903"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="251.885075" xlink:href="#C1_0_82f4c9fa18" y="197.39849"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="244.677001" xlink:href="#C1_0_f6794d661a" y="197.39849"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="245.411996" xlink:href="#C1_0_82f4c9fa18" y="207.505933"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="244.494458" xlink:href="#C1_0_f6794d661a" y="207.505933"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="249.876478" xlink:href="#C1_0_82f4c9fa18" y="230.966846"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="238.804304" xlink:href="#C1_0_f6794d661a" y="230.966846"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="249.267529" xlink:href="#C1_0_82f4c9fa18" y="222.105446"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="251.949688" xlink:href="#C1_0_f6794d661a" y="222.105446"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="250.386981" xlink:href="#C1_0_82f4c9fa18" y="224.506452"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="252.156096" xlink:href="#C1_0_f6794d661a" y="224.506452"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="257.542629" xlink:href="#C1_0_82f4c9fa18" y="195.44702"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="253.905468" xlink:href="#C1_0_f6794d661a" y="195.44702"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="241.847823" xlink:href="#C1_0_82f4c9fa18" y="203.859105"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="245.805328" xlink:href="#C1_0_f6794d661a" y="203.859105"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="251.643648" xlink:href="#C1_0_82f4c9fa18" y="253.550718"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="256.576606" xlink:href="#C1_0_f6794d661a" y="253.550718"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="253.504963" xlink:href="#C1_0_82f4c9fa18" y="210.047256"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="240.61208" xlink:href="#C1_0_f6794d661a" y="210.047256"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="247.624951" xlink:href="#C1_0_82f4c9fa18" y="232.914387"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="246.226675" xlink:href="#C1_0_f6794d661a" y="232.914387"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="243.410223" xlink:href="#C1_0_82f4c9fa18" y="184.051726"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="258.081928" xlink:href="#C1_0_f6794d661a" y="184.051726"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="259.218639" xlink:href="#C1_0_82f4c9fa18" y="271.697041"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="241.362785" xlink:href="#C1_0_f6794d661a" y="271.697041"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="255.752941" xlink:href="#C1_0_82f4c9fa18" y="287.550138"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="245.454037" xlink:href="#C1_0_f6794d661a" y="287.550138"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="246.691099" xlink:href="#C1_0_82f4c9fa18" y="171.991517"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="239.216934" xlink:href="#C1_0_f6794d661a" y="171.991517"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="245.677676" xlink:href="#C1_0_82f4c9fa18" y="289.813077"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="242.0264" xlink:href="#C1_0_f6794d661a" y="289.813077"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="238.009157" xlink:href="#C1_0_82f4c9fa18" y="258.492772"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="248.758797" xlink:href="#C1_0_f6794d661a" y="258.492772"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="255.409151" xlink:href="#C1_0_82f4c9fa18" y="207.501413"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="254.971773" xlink:href="#C1_0_f6794d661a" y="207.501413"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="256.322911" xlink:href="#C1_0_82f4c9fa18" y="265.61407"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="252.603318" xlink:href="#C1_0_f6794d661a" y="265.61407"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="253.535011" xlink:href="#C1_0_82f4c9fa18" y="171.639666"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="248.915537" xlink:href="#C1_0_f6794d661a" y="171.639666"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="248.705316" xlink:href="#C1_0_82f4c9fa18" y="178.461475"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="244.662664" xlink:href="#C1_0_f6794d661a" y="178.461475"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="243.204981" xlink:href="#C1_0_82f4c9fa18" y="222.123213"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="240.596198" xlink:href="#C1_0_f6794d661a" y="222.123213"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="254.869873" xlink:href="#C1_0_82f4c9fa18" y="176.05467"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="237.829611" xlink:href="#C1_0_f6794d661a" y="176.05467"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="242.275766" xlink:href="#C1_0_82f4c9fa18" y="163.998748"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="240.829997" xlink:href="#C1_0_f6794d661a" y="163.998748"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="243.833897" xlink:href="#C1_0_82f4c9fa18" y="143.801762"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="260.156396" xlink:href="#C1_0_f6794d661a" y="143.801762"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="242.344419" xlink:href="#C1_0_82f4c9fa18" y="220.082416"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="247.596949" xlink:href="#C1_0_f6794d661a" y="220.082416"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="246.956265" xlink:href="#C1_0_82f4c9fa18" y="241.497113"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="246.234432" xlink:href="#C1_0_f6794d661a" y="241.497113"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="243.153733" xlink:href="#C1_0_82f4c9fa18" y="200.69404"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="247.471406" xlink:href="#C1_0_f6794d661a" y="200.69404"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="259.925785" xlink:href="#C1_0_82f4c9fa18" y="220.085211"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="257.72277" xlink:href="#C1_0_f6794d661a" y="220.085211"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="258.022917" xlink:href="#C1_0_82f4c9fa18" y="268.153783"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="252.2602" xlink:href="#C1_0_f6794d661a" y="268.153783"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="241.535526" xlink:href="#C1_0_82f4c9fa18" y="171.988598"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="249.550905" xlink:href="#C1_0_f6794d661a" y="171.988598"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="246.972551" xlink:href="#C1_0_82f4c9fa18" y="222.106407"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="251.150873" xlink:href="#C1_0_f6794d661a" y="222.106407"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="248.354133" xlink:href="#C1_0_82f4c9fa18" y="233.471627"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="260.118044" xlink:href="#C1_0_f6794d661a" y="233.471627"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="241.121248" xlink:href="#C1_0_82f4c9fa18" y="202.585053"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="253.691718" xlink:href="#C1_0_f6794d661a" y="202.585053"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="246.162138" xlink:href="#C1_0_82f4c9fa18" y="249.918429"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="259.128552" xlink:href="#C1_0_f6794d661a" y="249.918429"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="251.462552" xlink:href="#C1_0_82f4c9fa18" y="211.497648"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="240.52906" xlink:href="#C1_0_f6794d661a" y="211.497648"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="242.691366" xlink:href="#C1_0_82f4c9fa18" y="211.505885"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="253.030406" xlink:href="#C1_0_f6794d661a" y="211.505885"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="240.791723" xlink:href="#C1_0_82f4c9fa18" y="200.390168"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="237.810066" xlink:href="#C1_0_f6794d661a" y="200.390168"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="254.64912" xlink:href="#C1_0_82f4c9fa18" y="247.561017"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="242.260588" xlink:href="#C1_0_f6794d661a" y="247.561017"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="240.04065" xlink:href="#C1_0_82f4c9fa18" y="184.063387"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="236.739356" xlink:href="#C1_0_f6794d661a" y="184.063387"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="253.187488" xlink:href="#C1_0_82f4c9fa18" y="201.910724"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="257.001252" xlink:href="#C1_0_f6794d661a" y="201.910724"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="239.892017" xlink:href="#C1_0_82f4c9fa18" y="258.1914"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="244.68886" xlink:href="#C1_0_f6794d661a" y="258.1914"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="253.716625" xlink:href="#C1_0_82f4c9fa18" y="244.087044"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="256.303264" xlink:href="#C1_0_f6794d661a" y="244.087044"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="244.541493" xlink:href="#C1_0_82f4c9fa18" y="183.386724"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="248.583029" xlink:href="#C1_0_f6794d661a" y="183.386724"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="253.082695" xlink:href="#C1_0_82f4c9fa18" y="278.975137"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="240.825724" xlink:href="#C1_0_f6794d661a" y="278.975137"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="258.763335" xlink:href="#C1_0_82f4c9fa18" y="163.490425"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="250.410914" xlink:href="#C1_0_f6794d661a" y="163.490425"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="253.603683" xlink:href="#C1_0_82f4c9fa18" y="161.44896"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="245.048643" xlink:href="#C1_0_f6794d661a" y="161.44896"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="239.982136" xlink:href="#C1_0_82f4c9fa18" y="155.863199"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="240.884275" xlink:href="#C1_0_f6794d661a" y="155.863199"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="251.264943" xlink:href="#C1_0_82f4c9fa18" y="207.501133"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="254.297786" xlink:href="#C1_0_f6794d661a" y="207.501133"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="250.365584" xlink:href="#C1_0_82f4c9fa18" y="178.433513"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="251.591911" xlink:href="#C1_0_f6794d661a" y="178.433513"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="251.590567" xlink:href="#C1_0_82f4c9fa18" y="178.427016"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="249.190229" xlink:href="#C1_0_f6794d661a" y="178.427016"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="251.802738" xlink:href="#C1_0_82f4c9fa18" y="233.273023"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="256.337985" xlink:href="#C1_0_f6794d661a" y="233.273023"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="257.942809" xlink:href="#C1_0_82f4c9fa18" y="248.627482"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="237.841219" xlink:href="#C1_0_f6794d661a" y="248.627482"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="250.527294" xlink:href="#C1_0_82f4c9fa18" y="222.088465"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="238.131314" xlink:href="#C1_0_f6794d661a" y="222.088465"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="255.099916" xlink:href="#C1_0_82f4c9fa18" y="201.897635"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="241.706771" xlink:href="#C1_0_f6794d661a" y="201.897635"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="252.681877" xlink:href="#C1_0_82f4c9fa18" y="249.915839"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="240.919088" xlink:href="#C1_0_f6794d661a" y="249.915839"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="235.970132" xlink:href="#C1_0_82f4c9fa18" y="180.040151"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="244.096836" xlink:href="#C1_0_f6794d661a" y="180.040151"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="248.750732" xlink:href="#C1_0_82f4c9fa18" y="238.59991"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="239.128033" xlink:href="#C1_0_f6794d661a" y="238.59991"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="256.992603" xlink:href="#C1_0_82f4c9fa18" y="146.032382"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="256.853443" xlink:href="#C1_0_f6794d661a" y="146.032382"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="244.922454" xlink:href="#C1_0_82f4c9fa18" y="149.38701"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="254.542932" xlink:href="#C1_0_f6794d661a" y="149.38701"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="244.432822" xlink:href="#C1_0_82f4c9fa18" y="224.492239"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="246.875296" xlink:href="#C1_0_f6794d661a" y="224.492239"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="260.006577" xlink:href="#C1_0_82f4c9fa18" y="224.506833"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="242.851065" xlink:href="#C1_0_f6794d661a" y="224.506833"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="247.460189" xlink:href="#C1_0_82f4c9fa18" y="184.918602"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="247.258302" xlink:href="#C1_0_f6794d661a" y="184.918602"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="246.05976" xlink:href="#C1_0_82f4c9fa18" y="252.663849"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="248.662857" xlink:href="#C1_0_f6794d661a" y="252.663849"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="240.869412" xlink:href="#C1_0_82f4c9fa18" y="153.031341"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="257.97259" xlink:href="#C1_0_f6794d661a" y="153.031341"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="248.055513" xlink:href="#C1_0_82f4c9fa18" y="171.975546"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="254.173867" xlink:href="#C1_0_f6794d661a" y="171.975546"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="251.274615" xlink:href="#C1_0_82f4c9fa18" y="176.049856"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="245.272048" xlink:href="#C1_0_f6794d661a" y="176.049856"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="240.338251" xlink:href="#C1_0_82f4c9fa18" y="268.154761"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="250.048883" xlink:href="#C1_0_f6794d661a" y="268.154761"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="249.971481" xlink:href="#C1_0_82f4c9fa18" y="197.38884"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="242.563136" xlink:href="#C1_0_f6794d661a" y="197.38884"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="254.302549" xlink:href="#C1_0_82f4c9fa18" y="186.099251"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="247.123948" xlink:href="#C1_0_f6794d661a" y="186.099251"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="237.610395" xlink:href="#C1_0_82f4c9fa18" y="153.052503"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="239.687678" xlink:href="#C1_0_f6794d661a" y="153.052503"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="242.188487" xlink:href="#C1_0_82f4c9fa18" y="224.54389"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="245.92367" xlink:href="#C1_0_f6794d661a" y="224.54389"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="242.588526" xlink:href="#C1_0_82f4c9fa18" y="195.434979"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="251.74373" xlink:href="#C1_0_f6794d661a" y="195.434979"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="256.111274" xlink:href="#C1_0_82f4c9fa18" y="161.464013"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="258.764094" xlink:href="#C1_0_f6794d661a" y="161.464013"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="249.208031" xlink:href="#C1_0_82f4c9fa18" y="193.04828"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="258.285678" xlink:href="#C1_0_f6794d661a" y="193.04828"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="258.558145" xlink:href="#C1_0_82f4c9fa18" y="186.0958"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="251.073876" xlink:href="#C1_0_f6794d661a" y="186.0958"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="250.710288" xlink:href="#C1_0_82f4c9fa18" y="264.735803"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="242.682328" xlink:href="#C1_0_f6794d661a" y="264.735803"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="241.517366" xlink:href="#C1_0_82f4c9fa18" y="200.392373"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="249.524376" xlink:href="#C1_0_f6794d661a" y="200.392373"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="256.88586" xlink:href="#C1_0_82f4c9fa18" y="143.842965"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="239.822537" xlink:href="#C1_0_f6794d661a" y="143.842965"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="237.862394" xlink:href="#C1_0_82f4c9fa18" y="195.446167"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="239.374956" xlink:href="#C1_0_f6794d661a" y="195.446167"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="236.958222" xlink:href="#C1_0_82f4c9fa18" y="207.513801"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="257.654684" xlink:href="#C1_0_f6794d661a" y="207.513801"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="253.008381" xlink:href="#C1_0_82f4c9fa18" y="186.580203"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="248.611767" xlink:href="#C1_0_f6794d661a" y="186.580203"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="250.085054" xlink:href="#C1_0_82f4c9fa18" y="253.570174"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="243.333949" xlink:href="#C1_0_f6794d661a" y="253.570174"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="259.836948" xlink:href="#C1_0_82f4c9fa18" y="241.537021"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="258.724275" xlink:href="#C1_0_f6794d661a" y="241.537021"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="253.924423" xlink:href="#C1_0_82f4c9fa18" y="160.22007"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="242.255379" xlink:href="#C1_0_f6794d661a" y="160.22007"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="256.382545" xlink:href="#C1_0_82f4c9fa18" y="239.143428"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="254.189263" xlink:href="#C1_0_f6794d661a" y="239.143428"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="251.707675" xlink:href="#C1_0_82f4c9fa18" y="169.104835"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="238.735356" xlink:href="#C1_0_f6794d661a" y="169.104835"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="241.675864" xlink:href="#C1_0_82f4c9fa18" y="258.507112"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="235.940039" xlink:href="#C1_0_f6794d661a" y="258.507112"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="244.466881" xlink:href="#C1_0_82f4c9fa18" y="186.110956"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="250.443807" xlink:href="#C1_0_f6794d661a" y="186.110956"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="252.121767" xlink:href="#C1_0_82f4c9fa18" y="120.358149"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="247.848001" xlink:href="#C1_0_f6794d661a" y="120.358149"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="244.132684" xlink:href="#C1_0_82f4c9fa18" y="222.123514"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="254.562908" xlink:href="#C1_0_f6794d661a" y="222.123514"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="257.174049" xlink:href="#C1_0_82f4c9fa18" y="193.081391"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="237.101102" xlink:href="#C1_0_f6794d661a" y="193.081391"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="258.748591" xlink:href="#C1_0_82f4c9fa18" y="171.995159"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="254.342073" xlink:href="#C1_0_f6794d661a" y="171.995159"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="247.30306" xlink:href="#C1_0_82f4c9fa18" y="178.46614"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="244.727938" xlink:href="#C1_0_f6794d661a" y="178.46614"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="248.181746" xlink:href="#C1_0_82f4c9fa18" y="253.570175"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="248.722821" xlink:href="#C1_0_f6794d661a" y="253.570175"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="257.373042" xlink:href="#C1_0_82f4c9fa18" y="239.122481"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="253.220497" xlink:href="#C1_0_f6794d661a" y="239.122481"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="252.161833" xlink:href="#C1_0_82f4c9fa18" y="138.900323"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="257.12959" xlink:href="#C1_0_f6794d661a" y="138.900323"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="243.413596" xlink:href="#C1_0_82f4c9fa18" y="271.721863"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="248.472292" xlink:href="#C1_0_f6794d661a" y="271.721863"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="247.154204" xlink:href="#C1_0_82f4c9fa18" y="249.923002"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="251.094287" xlink:href="#C1_0_f6794d661a" y="249.923002"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="242.700389" xlink:href="#C1_0_82f4c9fa18" y="166.399388"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="239.608628" xlink:href="#C1_0_f6794d661a" y="166.399388"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="251.882565" xlink:href="#C1_0_82f4c9fa18" y="180.990652"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="242.041104" xlink:href="#C1_0_f6794d661a" y="180.990652"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="253.812388" xlink:href="#C1_0_82f4c9fa18" y="249.91293"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="257.851049" xlink:href="#C1_0_f6794d661a" y="249.91293"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="237.048773" xlink:href="#C1_0_82f4c9fa18" y="253.586027"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="260.052025" xlink:href="#C1_0_f6794d661a" y="253.586027"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="254.488895" xlink:href="#C1_0_82f4c9fa18" y="253.578104"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="241.787346" xlink:href="#C1_0_f6794d661a" y="253.578104"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="244.277306" xlink:href="#C1_0_82f4c9fa18" y="268.18575"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="242.785547" xlink:href="#C1_0_f6794d661a" y="268.18575"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="253.029259" xlink:href="#C1_0_82f4c9fa18" y="207.533096"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="257.383617" xlink:href="#C1_0_f6794d661a" y="207.533096"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="237.05314" xlink:href="#C1_0_82f4c9fa18" y="239.131686"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="255.954924" xlink:href="#C1_0_f6794d661a" y="239.131686"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="243.583859" xlink:href="#C1_0_82f4c9fa18" y="169.123327"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="256.660227" xlink:href="#C1_0_f6794d661a" y="169.123327"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="237.199561" xlink:href="#C1_0_82f4c9fa18" y="230.981106"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="253.375578" xlink:href="#C1_0_f6794d661a" y="230.981106"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="238.832207" xlink:href="#C1_0_82f4c9fa18" y="310.800003"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="239.424911" xlink:href="#C1_0_f6794d661a" y="310.800003"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="240.281005" xlink:href="#C1_0_82f4c9fa18" y="222.125064"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="249.343377" xlink:href="#C1_0_f6794d661a" y="222.125064"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="259.358059" xlink:href="#C1_0_82f4c9fa18" y="285.451085"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="244.287032" xlink:href="#C1_0_f6794d661a" y="285.451085"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="256.467924" xlink:href="#C1_0_82f4c9fa18" y="219.21394"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="238.650188" xlink:href="#C1_0_f6794d661a" y="219.21394"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="250.187038" xlink:href="#C1_0_82f4c9fa18" y="180.057066"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="245.933532" xlink:href="#C1_0_f6794d661a" y="180.057066"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="240.165863" xlink:href="#C1_0_82f4c9fa18" y="222.131073"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="249.387768" xlink:href="#C1_0_f6794d661a" y="222.131073"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="244.30431" xlink:href="#C1_0_82f4c9fa18" y="220.868179"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="243.650988" xlink:href="#C1_0_f6794d661a" y="220.868179"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="258.499304" xlink:href="#C1_0_82f4c9fa18" y="171.94305"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="252.576129" xlink:href="#C1_0_f6794d661a" y="171.94305"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="259.50988" xlink:href="#C1_0_82f4c9fa18" y="123.973646"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="245.055722" xlink:href="#C1_0_f6794d661a" y="123.973646"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="251.818792" xlink:href="#C1_0_82f4c9fa18" y="253.57659"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="258.271907" xlink:href="#C1_0_f6794d661a" y="253.57659"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="245.586634" xlink:href="#C1_0_82f4c9fa18" y="195.39986"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="247.881776" xlink:href="#C1_0_f6794d661a" y="195.39986"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="244.260205" xlink:href="#C1_0_82f4c9fa18" y="207.543075"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="244.830856" xlink:href="#C1_0_f6794d661a" y="207.543075"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="249.581326" xlink:href="#C1_0_82f4c9fa18" y="112.153263"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="250.832433" xlink:href="#C1_0_f6794d661a" y="112.153263"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="259.664889" xlink:href="#C1_0_82f4c9fa18" y="232.914524"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="259.171605" xlink:href="#C1_0_f6794d661a" y="232.914524"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="254.113023" xlink:href="#C1_0_82f4c9fa18" y="214.860694"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="242.932513" xlink:href="#C1_0_f6794d661a" y="214.860694"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="237.050506" xlink:href="#C1_0_82f4c9fa18" y="248.654279"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="259.907996" xlink:href="#C1_0_f6794d661a" y="248.654279"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="255.774791" xlink:href="#C1_0_82f4c9fa18" y="195.468533"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="257.452508" xlink:href="#C1_0_f6794d661a" y="195.468533"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="256.928251" xlink:href="#C1_0_82f4c9fa18" y="214.527426"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="241.193256" xlink:href="#C1_0_f6794d661a" y="214.527426"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="242.931765" xlink:href="#C1_0_82f4c9fa18" y="230.245093"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="241.544543" xlink:href="#C1_0_f6794d661a" y="230.245093"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="251.561055" xlink:href="#C1_0_82f4c9fa18" y="220.863447"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="243.046826" xlink:href="#C1_0_f6794d661a" y="220.863447"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="258.702734" xlink:href="#C1_0_82f4c9fa18" y="193.053335"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="247.020449" xlink:href="#C1_0_f6794d661a" y="193.053335"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="255.558219" xlink:href="#C1_0_82f4c9fa18" y="182.446898"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="238.62372" xlink:href="#C1_0_f6794d661a" y="182.446898"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="257.124961" xlink:href="#C1_0_82f4c9fa18" y="253.557128"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="258.845369" xlink:href="#C1_0_f6794d661a" y="253.557128"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="252.780081" xlink:href="#C1_0_82f4c9fa18" y="207.529824"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="249.089013" xlink:href="#C1_0_f6794d661a" y="207.529824"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="258.133333" xlink:href="#C1_0_82f4c9fa18" y="264.752642"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="251.840956" xlink:href="#C1_0_f6794d661a" y="264.752642"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="250.348547" xlink:href="#C1_0_82f4c9fa18" y="224.503275"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="246.209373" xlink:href="#C1_0_f6794d661a" y="224.503275"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="248.947053" xlink:href="#C1_0_82f4c9fa18" y="253.55706"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="241.004185" xlink:href="#C1_0_f6794d661a" y="253.55706"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="241.86021" xlink:href="#C1_0_82f4c9fa18" y="268.156745"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="251.733384" xlink:href="#C1_0_f6794d661a" y="268.156745"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="259.504468" xlink:href="#C1_0_82f4c9fa18" y="153.078774"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="249.795698" xlink:href="#C1_0_f6794d661a" y="153.078774"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="257.110999" xlink:href="#C1_0_82f4c9fa18" y="200.188809"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="241.631853" xlink:href="#C1_0_f6794d661a" y="200.188809"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="246.258515" xlink:href="#C1_0_82f4c9fa18" y="181.218033"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="255.722131" xlink:href="#C1_0_f6794d661a" y="181.218033"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="252.720844" xlink:href="#C1_0_82f4c9fa18" y="268.192474"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="252.511781" xlink:href="#C1_0_f6794d661a" y="268.192474"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="237.477009" xlink:href="#C1_0_82f4c9fa18" y="220.875907"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="239.655477" xlink:href="#C1_0_f6794d661a" y="220.875907"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="236.124965" xlink:href="#C1_0_82f4c9fa18" y="143.848251"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="255.613208" xlink:href="#C1_0_f6794d661a" y="143.848251"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="258.487344" xlink:href="#C1_0_82f4c9fa18" y="253.560292"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="240.031487" xlink:href="#C1_0_f6794d661a" y="253.560292"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="241.038911" xlink:href="#C1_0_82f4c9fa18" y="224.52231"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="249.407197" xlink:href="#C1_0_f6794d661a" y="224.52231"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="252.643726" xlink:href="#C1_0_82f4c9fa18" y="97.771411"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="248.087805" xlink:href="#C1_0_f6794d661a" y="97.771411"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="251.967058" xlink:href="#C1_0_82f4c9fa18" y="140.047813"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="243.026387" xlink:href="#C1_0_f6794d661a" y="140.047813"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="253.680507" xlink:href="#C1_0_82f4c9fa18" y="222.110402"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="256.489812" xlink:href="#C1_0_f6794d661a" y="222.110402"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="254.282556" xlink:href="#C1_0_82f4c9fa18" y="211.505449"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="251.073362" xlink:href="#C1_0_f6794d661a" y="211.505449"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="249.843732" xlink:href="#C1_0_82f4c9fa18" y="281.728657"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="250.747585" xlink:href="#C1_0_f6794d661a" y="281.728657"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="240.221786" xlink:href="#C1_0_82f4c9fa18" y="300.758372"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="240.584159" xlink:href="#C1_0_f6794d661a" y="300.758372"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="248.869938" xlink:href="#C1_0_82f4c9fa18" y="252.696176"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="252.337546" xlink:href="#C1_0_f6794d661a" y="252.696176"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="256.705209" xlink:href="#C1_0_82f4c9fa18" y="268.160248"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="243.575031" xlink:href="#C1_0_f6794d661a" y="268.160248"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="248.336187" xlink:href="#C1_0_82f4c9fa18" y="195.468478"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="243.470738" xlink:href="#C1_0_f6794d661a" y="195.468478"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="251.424509" xlink:href="#C1_0_82f4c9fa18" y="166.421525"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="237.478428" xlink:href="#C1_0_f6794d661a" y="166.421525"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="249.762694" xlink:href="#C1_0_82f4c9fa18" y="254.74119"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="260.063666" xlink:href="#C1_0_f6794d661a" y="254.74119"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="250.402907" xlink:href="#C1_0_82f4c9fa18" y="196.273547"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="249.749826" xlink:href="#C1_0_f6794d661a" y="196.273547"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="252.583235" xlink:href="#C1_0_82f4c9fa18" y="207.510713"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="239.645129" xlink:href="#C1_0_f6794d661a" y="207.510713"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="239.201727" xlink:href="#C1_0_82f4c9fa18" y="171.63853"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="255.408884" xlink:href="#C1_0_f6794d661a" y="171.63853"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="253.584778" xlink:href="#C1_0_82f4c9fa18" y="239.124599"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="241.551166" xlink:href="#C1_0_f6794d661a" y="239.124599"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="239.746061" xlink:href="#C1_0_82f4c9fa18" y="180.99247"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="253.657239" xlink:href="#C1_0_f6794d661a" y="180.99247"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="256.518422" xlink:href="#C1_0_82f4c9fa18" y="258.514464"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="248.847057" xlink:href="#C1_0_f6794d661a" y="258.514464"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="238.00373" xlink:href="#C1_0_82f4c9fa18" y="207.515692"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="237.949501" xlink:href="#C1_0_f6794d661a" y="207.515692"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="245.202313" xlink:href="#C1_0_82f4c9fa18" y="157.067408"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="252.983725" xlink:href="#C1_0_f6794d661a" y="157.067408"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="252.639654" xlink:href="#C1_0_82f4c9fa18" y="296.59294"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="257.586564" xlink:href="#C1_0_f6794d661a" y="296.59294"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="244.352236" xlink:href="#C1_0_82f4c9fa18" y="184.939129"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="245.766564" xlink:href="#C1_0_f6794d661a" y="184.939129"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="252.431817" xlink:href="#C1_0_82f4c9fa18" y="203.879793"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="251.702575" xlink:href="#C1_0_f6794d661a" y="203.879793"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="239.319924" xlink:href="#C1_0_82f4c9fa18" y="161.476812"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="251.839272" xlink:href="#C1_0_f6794d661a" y="161.476812"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="256.393049" xlink:href="#C1_0_82f4c9fa18" y="157.037807"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="240.664193" xlink:href="#C1_0_f6794d661a" y="157.037807"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="253.633429" xlink:href="#C1_0_82f4c9fa18" y="239.121704"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="245.897539" xlink:href="#C1_0_f6794d661a" y="239.121704"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="251.37943" xlink:href="#C1_0_82f4c9fa18" y="195.478211"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="248.704589" xlink:href="#C1_0_f6794d661a" y="195.478211"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="238.811921" xlink:href="#C1_0_82f4c9fa18" y="212.463793"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="251.190932" xlink:href="#C1_0_f6794d661a" y="212.463793"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="236.718956" xlink:href="#C1_0_82f4c9fa18" y="171.653835"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="240.410848" xlink:href="#C1_0_f6794d661a" y="171.653835"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="257.770529" xlink:href="#C1_0_82f4c9fa18" y="264.281748"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="247.082909" xlink:href="#C1_0_f6794d661a" y="264.281748"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="245.479865" xlink:href="#C1_0_82f4c9fa18" y="224.535933"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="248.19913" xlink:href="#C1_0_f6794d661a" y="224.535933"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="245.82339" xlink:href="#C1_0_82f4c9fa18" y="195.474854"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="239.345103" xlink:href="#C1_0_f6794d661a" y="195.474854"/> </g> </g> <g id="PathCollection_5"> @@ -2039,277 +2039,277 @@ C -2.236584 -1.29895 -2.5 -0.663008 -2.5 0 C -2.5 0.663008 -2.236584 1.29895 -1.767767 1.767767 C -1.29895 2.236584 -0.663008 2.5 0 2.5 z -" id="C2_0_e29aafa51a"/> +" id="C2_0_832c39148d"/> </defs> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="390.84124" xlink:href="#C2_0_e29aafa51a" y="149.690246"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="390.475999" xlink:href="#C2_0_832c39148d" y="149.690246"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="405.346413" xlink:href="#C2_0_e29aafa51a" y="207.044912"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="397.965421" xlink:href="#C2_0_832c39148d" y="207.044912"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="393.88288" xlink:href="#C2_0_e29aafa51a" y="246.208"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="393.484872" xlink:href="#C2_0_832c39148d" y="246.208"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="406.578652" xlink:href="#C2_0_e29aafa51a" y="170.287749"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="390.724552" xlink:href="#C2_0_832c39148d" y="170.287749"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="389.621541" xlink:href="#C2_0_e29aafa51a" y="248.746795"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="382.201112" xlink:href="#C2_0_832c39148d" y="248.746795"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="387.448552" xlink:href="#C2_0_e29aafa51a" y="210.338243"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="391.878186" xlink:href="#C2_0_832c39148d" y="210.338243"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="405.604467" xlink:href="#C2_0_e29aafa51a" y="207.488934"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="401.75448" xlink:href="#C2_0_832c39148d" y="207.488934"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="389.459367" xlink:href="#C2_0_e29aafa51a" y="231.75118"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="392.341688" xlink:href="#C2_0_832c39148d" y="231.75118"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="396.200572" xlink:href="#C2_0_e29aafa51a" y="183.680963"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="397.036616" xlink:href="#C2_0_832c39148d" y="183.680963"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="385.15736" xlink:href="#C2_0_e29aafa51a" y="168.833673"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="398.36239" xlink:href="#C2_0_832c39148d" y="168.833673"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="406.122402" xlink:href="#C2_0_e29aafa51a" y="219.69547"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="404.938844" xlink:href="#C2_0_832c39148d" y="219.69547"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="383.552829" xlink:href="#C2_0_e29aafa51a" y="225.28767"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="398.746544" xlink:href="#C2_0_832c39148d" y="225.28767"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="405.77538" xlink:href="#C2_0_e29aafa51a" y="228.327559"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="397.640305" xlink:href="#C2_0_832c39148d" y="228.327559"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="388.489197" xlink:href="#C2_0_e29aafa51a" y="168.835615"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="387.393832" xlink:href="#C2_0_832c39148d" y="168.835615"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="399.468616" xlink:href="#C2_0_e29aafa51a" y="263.201059"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="388.367034" xlink:href="#C2_0_832c39148d" y="263.201059"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="387.503315" xlink:href="#C2_0_e29aafa51a" y="210.337366"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="393.12258" xlink:href="#C2_0_832c39148d" y="210.337366"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="386.225279" xlink:href="#C2_0_e29aafa51a" y="260.803466"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="385.188585" xlink:href="#C2_0_832c39148d" y="260.803466"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="390.749613" xlink:href="#C2_0_e29aafa51a" y="181.098047"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="382.219913" xlink:href="#C2_0_832c39148d" y="181.098047"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="392.403273" xlink:href="#C2_0_e29aafa51a" y="203.81402"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="392.134443" xlink:href="#C2_0_832c39148d" y="203.81402"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="386.084369" xlink:href="#C2_0_e29aafa51a" y="141.231824"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="390.362719" xlink:href="#C2_0_832c39148d" y="141.231824"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="389.526552" xlink:href="#C2_0_e29aafa51a" y="144.092936"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="404.109521" xlink:href="#C2_0_832c39148d" y="144.092936"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="383.975172" xlink:href="#C2_0_e29aafa51a" y="317.660183"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="406.491078" xlink:href="#C2_0_832c39148d" y="317.660183"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="393.41676" xlink:href="#C2_0_e29aafa51a" y="260.803345"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="398.879198" xlink:href="#C2_0_832c39148d" y="260.803345"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="398.153593" xlink:href="#C2_0_e29aafa51a" y="169.018987"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="403.645853" xlink:href="#C2_0_832c39148d" y="169.018987"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="395.7849" xlink:href="#C2_0_e29aafa51a" y="261.107859"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="387.422005" xlink:href="#C2_0_832c39148d" y="261.107859"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="393.088997" xlink:href="#C2_0_e29aafa51a" y="188.398313"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="402.909115" xlink:href="#C2_0_832c39148d" y="188.398313"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="404.482785" xlink:href="#C2_0_e29aafa51a" y="196.814829"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="399.962593" xlink:href="#C2_0_832c39148d" y="196.814829"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="394.651677" xlink:href="#C2_0_e29aafa51a" y="169.138409"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="401.18723" xlink:href="#C2_0_832c39148d" y="169.138409"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="395.221958" xlink:href="#C2_0_e29aafa51a" y="217.452616"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="405.58835" xlink:href="#C2_0_832c39148d" y="217.452616"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="395.822051" xlink:href="#C2_0_e29aafa51a" y="150.000846"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="399.980208" xlink:href="#C2_0_832c39148d" y="150.000846"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="387.132602" xlink:href="#C2_0_e29aafa51a" y="164.953349"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="386.345534" xlink:href="#C2_0_832c39148d" y="164.953349"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="402.280259" xlink:href="#C2_0_e29aafa51a" y="204.463925"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="397.447235" xlink:href="#C2_0_832c39148d" y="204.463925"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="387.228369" xlink:href="#C2_0_e29aafa51a" y="239.726796"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="396.305339" xlink:href="#C2_0_832c39148d" y="239.726796"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="385.624783" xlink:href="#C2_0_e29aafa51a" y="181.308809"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="382.908606" xlink:href="#C2_0_832c39148d" y="181.308809"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="398.622557" xlink:href="#C2_0_e29aafa51a" y="263.495117"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="388.163306" xlink:href="#C2_0_832c39148d" y="263.495117"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="389.655695" xlink:href="#C2_0_e29aafa51a" y="250.536382"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="391.681741" xlink:href="#C2_0_832c39148d" y="250.536382"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="394.237926" xlink:href="#C2_0_e29aafa51a" y="173.978374"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="384.30987" xlink:href="#C2_0_832c39148d" y="173.978374"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="389.015028" xlink:href="#C2_0_e29aafa51a" y="249.873362"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="397.423924" xlink:href="#C2_0_832c39148d" y="249.873362"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="390.194567" xlink:href="#C2_0_e29aafa51a" y="190.95988"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="404.836804" xlink:href="#C2_0_832c39148d" y="190.95988"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="404.660691" xlink:href="#C2_0_e29aafa51a" y="181.621632"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="389.189746" xlink:href="#C2_0_832c39148d" y="181.621632"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="397.577362" xlink:href="#C2_0_e29aafa51a" y="288.920969"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="391.157691" xlink:href="#C2_0_832c39148d" y="288.920969"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="386.870298" xlink:href="#C2_0_e29aafa51a" y="269.996681"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="387.354986" xlink:href="#C2_0_832c39148d" y="269.996681"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="406.646185" xlink:href="#C2_0_e29aafa51a" y="177.849032"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="401.82034" xlink:href="#C2_0_832c39148d" y="177.849032"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="385.814516" xlink:href="#C2_0_e29aafa51a" y="263.534152"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="394.00015" xlink:href="#C2_0_832c39148d" y="263.534152"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="397.667787" xlink:href="#C2_0_e29aafa51a" y="181.617863"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="403.797191" xlink:href="#C2_0_832c39148d" y="181.617863"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="397.07881" xlink:href="#C2_0_e29aafa51a" y="173.480755"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="395.449402" xlink:href="#C2_0_832c39148d" y="173.480755"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="400.978708" xlink:href="#C2_0_e29aafa51a" y="213.057381"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="386.941419" xlink:href="#C2_0_832c39148d" y="213.057381"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="403.436157" xlink:href="#C2_0_e29aafa51a" y="167.503409"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="397.79978" xlink:href="#C2_0_832c39148d" y="167.503409"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="397.997283" xlink:href="#C2_0_e29aafa51a" y="269.954036"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="387.778748" xlink:href="#C2_0_832c39148d" y="269.954036"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="393.433251" xlink:href="#C2_0_e29aafa51a" y="173.788682"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="387.307693" xlink:href="#C2_0_832c39148d" y="173.788682"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="393.550647" xlink:href="#C2_0_e29aafa51a" y="292.587365"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="404.853782" xlink:href="#C2_0_832c39148d" y="292.587365"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="386.15666" xlink:href="#C2_0_e29aafa51a" y="211.845864"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="387.110946" xlink:href="#C2_0_832c39148d" y="211.845864"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="395.829875" xlink:href="#C2_0_e29aafa51a" y="159.334573"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="402.396676" xlink:href="#C2_0_832c39148d" y="159.334573"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="391.076298" xlink:href="#C2_0_e29aafa51a" y="234.468961"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="384.700241" xlink:href="#C2_0_832c39148d" y="234.468961"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="392.892075" xlink:href="#C2_0_e29aafa51a" y="292.577101"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="401.454188" xlink:href="#C2_0_832c39148d" y="292.577101"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="403.328802" xlink:href="#C2_0_e29aafa51a" y="169.126912"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="388.92847" xlink:href="#C2_0_832c39148d" y="169.126912"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="384.642931" xlink:href="#C2_0_e29aafa51a" y="184.761682"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="395.76762" xlink:href="#C2_0_832c39148d" y="184.761682"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="406.646808" xlink:href="#C2_0_e29aafa51a" y="192.032499"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="387.919868" xlink:href="#C2_0_832c39148d" y="192.032499"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="387.428216" xlink:href="#C2_0_e29aafa51a" y="182.627554"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="390.590157" xlink:href="#C2_0_832c39148d" y="182.627554"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="388.185244" xlink:href="#C2_0_e29aafa51a" y="193.649931"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="399.092373" xlink:href="#C2_0_832c39148d" y="193.649931"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="387.248514" xlink:href="#C2_0_e29aafa51a" y="199.642913"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="405.863955" xlink:href="#C2_0_832c39148d" y="199.642913"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="396.628793" xlink:href="#C2_0_e29aafa51a" y="198.949705"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="385.575148" xlink:href="#C2_0_832c39148d" y="198.949705"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="384.284636" xlink:href="#C2_0_e29aafa51a" y="251.432621"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="385.166954" xlink:href="#C2_0_832c39148d" y="251.432621"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="392.275744" xlink:href="#C2_0_e29aafa51a" y="214.175328"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="391.077134" xlink:href="#C2_0_832c39148d" y="214.175328"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="406.65758" xlink:href="#C2_0_e29aafa51a" y="210.239489"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="384.744497" xlink:href="#C2_0_832c39148d" y="210.239489"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="393.260896" xlink:href="#C2_0_e29aafa51a" y="196.55794"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="395.349085" xlink:href="#C2_0_832c39148d" y="196.55794"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="389.712695" xlink:href="#C2_0_e29aafa51a" y="156.006652"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="384.327311" xlink:href="#C2_0_832c39148d" y="156.006652"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="406.594033" xlink:href="#C2_0_e29aafa51a" y="263.492625"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="390.050536" xlink:href="#C2_0_832c39148d" y="263.492625"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="389.42298" xlink:href="#C2_0_e29aafa51a" y="172.235492"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="394.096075" xlink:href="#C2_0_832c39148d" y="172.235492"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="385.888469" xlink:href="#C2_0_e29aafa51a" y="215.290772"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="382.555439" xlink:href="#C2_0_832c39148d" y="215.290772"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="405.930654" xlink:href="#C2_0_e29aafa51a" y="292.544009"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="406.385103" xlink:href="#C2_0_832c39148d" y="292.544009"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="395.917735" xlink:href="#C2_0_e29aafa51a" y="304.603374"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="387.370058" xlink:href="#C2_0_832c39148d" y="304.603374"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="398.618192" xlink:href="#C2_0_e29aafa51a" y="167.004699"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="383.888157" xlink:href="#C2_0_832c39148d" y="167.004699"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="391.897436" xlink:href="#C2_0_e29aafa51a" y="259.872502"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="405.872309" xlink:href="#C2_0_832c39148d" y="259.872502"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="394.545991" xlink:href="#C2_0_e29aafa51a" y="269.954098"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="405.243677" xlink:href="#C2_0_832c39148d" y="269.954098"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="395.492362" xlink:href="#C2_0_e29aafa51a" y="307.158181"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="383.197247" xlink:href="#C2_0_832c39148d" y="307.158181"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="395.793623" xlink:href="#C2_0_e29aafa51a" y="192.384078"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="403.445135" xlink:href="#C2_0_832c39148d" y="192.384078"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="387.666471" xlink:href="#C2_0_e29aafa51a" y="237.153489"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="382.690067" xlink:href="#C2_0_832c39148d" y="237.153489"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="404.107674" xlink:href="#C2_0_e29aafa51a" y="232.231875"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="396.516601" xlink:href="#C2_0_832c39148d" y="232.231875"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="406.10931" xlink:href="#C2_0_e29aafa51a" y="234.460146"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="385.60786" xlink:href="#C2_0_832c39148d" y="234.460146"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="382.755137" xlink:href="#C2_0_e29aafa51a" y="217.472154"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="386.807757" xlink:href="#C2_0_832c39148d" y="217.472154"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="392.761503" xlink:href="#C2_0_e29aafa51a" y="229.517557"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="394.427127" xlink:href="#C2_0_832c39148d" y="229.517557"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="387.366921" xlink:href="#C2_0_e29aafa51a" y="246.523807"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="397.382376" xlink:href="#C2_0_832c39148d" y="246.523807"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="383.203214" xlink:href="#C2_0_e29aafa51a" y="190.370071"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="404.475197" xlink:href="#C2_0_832c39148d" y="190.370071"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="390.607014" xlink:href="#C2_0_e29aafa51a" y="157.649777"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="386.150771" xlink:href="#C2_0_832c39148d" y="157.649777"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="393.449033" xlink:href="#C2_0_e29aafa51a" y="263.511599"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="390.416439" xlink:href="#C2_0_832c39148d" y="263.511599"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="399.232117" xlink:href="#C2_0_e29aafa51a" y="217.474908"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="395.497037" xlink:href="#C2_0_832c39148d" y="217.474908"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="399.286681" xlink:href="#C2_0_e29aafa51a" y="194.467998"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="399.792777" xlink:href="#C2_0_832c39148d" y="194.467998"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="385.211454" xlink:href="#C2_0_e29aafa51a" y="303.741702"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="383.287595" xlink:href="#C2_0_832c39148d" y="303.741702"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="393.815193" xlink:href="#C2_0_e29aafa51a" y="261.127375"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="383.289701" xlink:href="#C2_0_832c39148d" y="261.127375"/> </g> </g> <g id="PathCollection_6"> @@ -2324,259 +2324,259 @@ C -2.236584 -1.29895 -2.5 -0.663008 -2.5 0 C -2.5 0.663008 -2.236584 1.29895 -1.767767 1.767767 C -1.29895 2.236584 -0.663008 2.5 0 2.5 z -" id="C3_0_8e4775d9e6"/> +" id="C3_0_947d114575"/> </defs> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="484.468514" xlink:href="#C3_0_8e4775d9e6" y="149.245404"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="482.642807" xlink:href="#C3_0_947d114575" y="149.245404"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="482.754636" xlink:href="#C3_0_8e4775d9e6" y="182.290351"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="481.498165" xlink:href="#C3_0_947d114575" y="182.290351"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="480.648449" xlink:href="#C3_0_8e4775d9e6" y="207.348341"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="489.850049" xlink:href="#C3_0_947d114575" y="207.348341"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="493.587132" xlink:href="#C3_0_8e4775d9e6" y="163.846465"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="493.24087" xlink:href="#C3_0_947d114575" y="163.846465"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="482.651815" xlink:href="#C3_0_8e4775d9e6" y="238.949529"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="485.118408" xlink:href="#C3_0_947d114575" y="238.949529"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="503.345121" xlink:href="#C3_0_8e4775d9e6" y="180.844268"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="496.060057" xlink:href="#C3_0_947d114575" y="180.844268"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="503.643901" xlink:href="#C3_0_8e4775d9e6" y="229.28663"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="484.159309" xlink:href="#C3_0_947d114575" y="229.28663"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="498.201577" xlink:href="#C3_0_8e4775d9e6" y="209.895621"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="483.579613" xlink:href="#C3_0_947d114575" y="209.895621"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="500.462082" xlink:href="#C3_0_8e4775d9e6" y="186.884487"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="488.988825" xlink:href="#C3_0_947d114575" y="186.884487"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="498.614771" xlink:href="#C3_0_8e4775d9e6" y="152.898133"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="496.006158" xlink:href="#C3_0_947d114575" y="152.898133"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="493.207548" xlink:href="#C3_0_8e4775d9e6" y="221.947021"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="489.82018" xlink:href="#C3_0_947d114575" y="221.947021"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="498.883095" xlink:href="#C3_0_8e4775d9e6" y="192.905385"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="496.084735" xlink:href="#C3_0_947d114575" y="192.905385"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="495.87073" xlink:href="#C3_0_8e4775d9e6" y="214.114189"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="494.178335" xlink:href="#C3_0_947d114575" y="214.114189"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="502.743843" xlink:href="#C3_0_8e4775d9e6" y="156.888241"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="489.193843" xlink:href="#C3_0_947d114575" y="156.888241"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="480.564122" xlink:href="#C3_0_8e4775d9e6" y="224.350175"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="490.264074" xlink:href="#C3_0_947d114575" y="224.350175"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="494.082118" xlink:href="#C3_0_8e4775d9e6" y="180.846286"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="496.495753" xlink:href="#C3_0_947d114575" y="180.846286"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="490.401903" xlink:href="#C3_0_8e4775d9e6" y="209.900153"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="501.257731" xlink:href="#C3_0_947d114575" y="209.900153"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="504.435815" xlink:href="#C3_0_8e4775d9e6" y="162.586682"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="491.668964" xlink:href="#C3_0_947d114575" y="162.586682"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="484.804459" xlink:href="#C3_0_8e4775d9e6" y="193.978079"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="499.069394" xlink:href="#C3_0_947d114575" y="193.978079"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="489.856168" xlink:href="#C3_0_8e4775d9e6" y="136.580688"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="484.655298" xlink:href="#C3_0_947d114575" y="136.580688"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="503.12363" xlink:href="#C3_0_8e4775d9e6" y="143.648879"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="490.762288" xlink:href="#C3_0_947d114575" y="143.648879"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="503.746342" xlink:href="#C3_0_8e4775d9e6" y="268.008621"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="503.865569" xlink:href="#C3_0_947d114575" y="268.008621"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="496.022779" xlink:href="#C3_0_8e4775d9e6" y="154.690129"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="501.94414" xlink:href="#C3_0_947d114575" y="154.690129"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="489.531087" xlink:href="#C3_0_8e4775d9e6" y="200.736617"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="480.990598" xlink:href="#C3_0_947d114575" y="200.736617"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="490.414802" xlink:href="#C3_0_8e4775d9e6" y="186.893091"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="488.106325" xlink:href="#C3_0_947d114575" y="186.893091"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="500.874879" xlink:href="#C3_0_8e4775d9e6" y="166.431399"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="496.360536" xlink:href="#C3_0_947d114575" y="166.431399"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="489.759249" xlink:href="#C3_0_8e4775d9e6" y="174.846957"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="482.564612" xlink:href="#C3_0_947d114575" y="174.846957"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="488.083892" xlink:href="#C3_0_8e4775d9e6" y="153.062619"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="496.25633" xlink:href="#C3_0_947d114575" y="153.062619"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="481.327631" xlink:href="#C3_0_8e4775d9e6" y="161.468898"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="496.917684" xlink:href="#C3_0_947d114575" y="161.468898"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="499.433094" xlink:href="#C3_0_8e4775d9e6" y="232.925293"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="498.124995" xlink:href="#C3_0_947d114575" y="232.925293"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="501.056995" xlink:href="#C3_0_8e4775d9e6" y="180.987666"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="483.935009" xlink:href="#C3_0_947d114575" y="180.987666"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="485.418008" xlink:href="#C3_0_8e4775d9e6" y="157.977647"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="485.348265" xlink:href="#C3_0_947d114575" y="157.977647"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="480.63167" xlink:href="#C3_0_8e4775d9e6" y="253.616023"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="501.726635" xlink:href="#C3_0_947d114575" y="253.616023"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="502.891607" xlink:href="#C3_0_8e4775d9e6" y="249.897009"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="482.113209" xlink:href="#C3_0_947d114575" y="249.897009"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="493.907014" xlink:href="#C3_0_8e4775d9e6" y="180.981605"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="488.976452" xlink:href="#C3_0_947d114575" y="180.981605"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="503.613412" xlink:href="#C3_0_8e4775d9e6" y="273.919096"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="486.52526" xlink:href="#C3_0_947d114575" y="273.919096"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="490.099284" xlink:href="#C3_0_8e4775d9e6" y="268.146351"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="481.746383" xlink:href="#C3_0_947d114575" y="268.146351"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="496.27247" xlink:href="#C3_0_8e4775d9e6" y="193.038595"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="489.402808" xlink:href="#C3_0_947d114575" y="193.038595"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="482.787592" xlink:href="#C3_0_8e4775d9e6" y="278.984496"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="495.742904" xlink:href="#C3_0_947d114575" y="278.984496"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="497.515676" xlink:href="#C3_0_8e4775d9e6" y="172.847465"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="495.92895" xlink:href="#C3_0_947d114575" y="172.847465"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="486.253493" xlink:href="#C3_0_8e4775d9e6" y="143.881955"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="494.399192" xlink:href="#C3_0_947d114575" y="143.881955"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="490.992131" xlink:href="#C3_0_8e4775d9e6" y="195.442985"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="489.845999" xlink:href="#C3_0_947d114575" y="195.442985"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="484.906943" xlink:href="#C3_0_8e4775d9e6" y="149.394395"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="485.468309" xlink:href="#C3_0_947d114575" y="149.394395"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="487.853034" xlink:href="#C3_0_8e4775d9e6" y="155.852835"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="481.592737" xlink:href="#C3_0_947d114575" y="155.852835"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="503.389649" xlink:href="#C3_0_8e4775d9e6" y="200.373399"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="485.56097" xlink:href="#C3_0_947d114575" y="200.373399"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="490.272519" xlink:href="#C3_0_8e4775d9e6" y="142.575599"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="493.624116" xlink:href="#C3_0_947d114575" y="142.575599"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="492.579434" xlink:href="#C3_0_8e4775d9e6" y="157.108159"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="491.704913" xlink:href="#C3_0_947d114575" y="157.108159"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="499.437824" xlink:href="#C3_0_8e4775d9e6" y="253.553345"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="497.182347" xlink:href="#C3_0_947d114575" y="253.553345"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="499.669222" xlink:href="#C3_0_8e4775d9e6" y="201.975598"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="482.376175" xlink:href="#C3_0_947d114575" y="201.975598"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="495.295186" xlink:href="#C3_0_8e4775d9e6" y="172.057581"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="483.5617" xlink:href="#C3_0_947d114575" y="172.057581"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="499.19217" xlink:href="#C3_0_8e4775d9e6" y="207.49562"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="498.21515" xlink:href="#C3_0_947d114575" y="207.49562"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="502.676055" xlink:href="#C3_0_8e4775d9e6" y="195.43714"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="481.272222" xlink:href="#C3_0_947d114575" y="195.43714"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="496.997314" xlink:href="#C3_0_8e4775d9e6" y="186.160914"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="500.774194" xlink:href="#C3_0_947d114575" y="186.160914"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="486.169542" xlink:href="#C3_0_8e4775d9e6" y="165.467127"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="499.939626" xlink:href="#C3_0_947d114575" y="165.467127"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="492.143969" xlink:href="#C3_0_8e4775d9e6" y="137.415129"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="487.242776" xlink:href="#C3_0_947d114575" y="137.415129"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="503.078155" xlink:href="#C3_0_8e4775d9e6" y="222.172179"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="499.125875" xlink:href="#C3_0_947d114575" y="222.172179"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="490.374768" xlink:href="#C3_0_8e4775d9e6" y="174.027489"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="483.291365" xlink:href="#C3_0_947d114575" y="174.027489"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="502.677992" xlink:href="#C3_0_8e4775d9e6" y="181.022666"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="487.635458" xlink:href="#C3_0_947d114575" y="181.022666"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="502.918885" xlink:href="#C3_0_8e4775d9e6" y="210.090067"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="503.739559" xlink:href="#C3_0_947d114575" y="210.090067"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="483.206813" xlink:href="#C3_0_8e4775d9e6" y="200.383663"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="500.62283" xlink:href="#C3_0_947d114575" y="200.383663"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="491.788259" xlink:href="#C3_0_8e4775d9e6" y="258.563069"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="498.261694" xlink:href="#C3_0_947d114575" y="258.563069"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="483.996538" xlink:href="#C3_0_8e4775d9e6" y="208.720706"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="484.946676" xlink:href="#C3_0_947d114575" y="208.720706"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="483.991053" xlink:href="#C3_0_8e4775d9e6" y="178.518662"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="495.292717" xlink:href="#C3_0_947d114575" y="178.518662"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="494.442289" xlink:href="#C3_0_8e4775d9e6" y="186.578797"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="495.226955" xlink:href="#C3_0_947d114575" y="186.578797"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="494.090316" xlink:href="#C3_0_8e4775d9e6" y="137.337355"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="489.204008" xlink:href="#C3_0_947d114575" y="137.337355"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="500.209518" xlink:href="#C3_0_8e4775d9e6" y="167.680504"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="491.528342" xlink:href="#C3_0_947d114575" y="167.680504"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="491.545529" xlink:href="#C3_0_8e4775d9e6" y="202.722805"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="495.554183" xlink:href="#C3_0_947d114575" y="202.722805"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="502.646344" xlink:href="#C3_0_8e4775d9e6" y="224.575075"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="493.016723" xlink:href="#C3_0_947d114575" y="224.575075"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="498.392112" xlink:href="#C3_0_8e4775d9e6" y="202.640406"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="491.705303" xlink:href="#C3_0_947d114575" y="202.640406"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="492.124383" xlink:href="#C3_0_8e4775d9e6" y="146.039082"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="490.434564" xlink:href="#C3_0_947d114575" y="146.039082"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="504.285576" xlink:href="#C3_0_8e4775d9e6" y="249.906575"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="488.495378" xlink:href="#C3_0_947d114575" y="249.906575"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="498.112434" xlink:href="#C3_0_8e4775d9e6" y="231.029491"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="499.422787" xlink:href="#C3_0_947d114575" y="231.029491"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="483.606158" xlink:href="#C3_0_8e4775d9e6" y="174.871131"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="492.959149" xlink:href="#C3_0_947d114575" y="174.871131"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="503.110589" xlink:href="#C3_0_8e4775d9e6" y="198.194725"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="481.061928" xlink:href="#C3_0_947d114575" y="198.194725"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="502.134113" xlink:href="#C3_0_8e4775d9e6" y="219.255421"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="483.090312" xlink:href="#C3_0_947d114575" y="219.255421"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="495.491828" xlink:href="#C3_0_8e4775d9e6" y="178.444737"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="481.707434" xlink:href="#C3_0_947d114575" y="178.444737"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="502.256086" xlink:href="#C3_0_8e4775d9e6" y="213.129098"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="501.416026" xlink:href="#C3_0_947d114575" y="213.129098"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="486.322971" xlink:href="#C3_0_8e4775d9e6" y="207.504974"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="502.828976" xlink:href="#C3_0_947d114575" y="207.504974"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="486.730442" xlink:href="#C3_0_8e4775d9e6" y="197.384877"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="495.866076" xlink:href="#C3_0_947d114575" y="197.384877"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="491.023196" xlink:href="#C3_0_8e4775d9e6" y="153.045598"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="480.372062" xlink:href="#C3_0_947d114575" y="153.045598"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="496.067059" xlink:href="#C3_0_8e4775d9e6" y="195.466107"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="479.983803" xlink:href="#C3_0_947d114575" y="195.466107"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="487.096181" xlink:href="#C3_0_8e4775d9e6" y="137.325134"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="493.137643" xlink:href="#C3_0_947d114575" y="137.325134"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="487.523493" xlink:href="#C3_0_8e4775d9e6" y="183.609011"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="496.77052" xlink:href="#C3_0_947d114575" y="183.609011"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="496.161036" xlink:href="#C3_0_8e4775d9e6" y="222.094808"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="497.659624" xlink:href="#C3_0_947d114575" y="222.094808"/> </g> </g> <g id="PathCollection_7"> @@ -2591,103 +2591,103 @@ C -2.236584 -1.29895 -2.5 -0.663008 -2.5 0 C -2.5 0.663008 -2.236584 1.29895 -1.767767 1.767767 C -1.29895 2.236584 -0.663008 2.5 0 2.5 z -" id="C4_0_cbaeea66ea"/> +" id="C4_0_56c013b1a9"/> </defs> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="639.27349" xlink:href="#C4_0_cbaeea66ea" y="217.866084"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="626.312302" xlink:href="#C4_0_56c013b1a9" y="217.866084"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="639.176222" xlink:href="#C4_0_cbaeea66ea" y="196.952689"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="632.392078" xlink:href="#C4_0_56c013b1a9" y="196.952689"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="627.715185" xlink:href="#C4_0_cbaeea66ea" y="234.86163"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="647.480004" xlink:href="#C4_0_56c013b1a9" y="234.86163"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="631.41688" xlink:href="#C4_0_cbaeea66ea" y="225.50833"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="641.107262" xlink:href="#C4_0_56c013b1a9" y="225.50833"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="648.179464" xlink:href="#C4_0_cbaeea66ea" y="334.082122"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="626.198008" xlink:href="#C4_0_56c013b1a9" y="334.082122"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="637.633266" xlink:href="#C4_0_cbaeea66ea" y="184.198957"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="639.538056" xlink:href="#C4_0_56c013b1a9" y="184.198957"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="631.069641" xlink:href="#C4_0_cbaeea66ea" y="272.378164"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="628.672093" xlink:href="#C4_0_56c013b1a9" y="272.378164"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="639.080733" xlink:href="#C4_0_cbaeea66ea" y="260.316603"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="637.068338" xlink:href="#C4_0_56c013b1a9" y="260.316603"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="637.450074" xlink:href="#C4_0_cbaeea66ea" y="241.374028"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="636.764606" xlink:href="#C4_0_56c013b1a9" y="241.374028"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="638.629363" xlink:href="#C4_0_cbaeea66ea" y="246.996271"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="644.076285" xlink:href="#C4_0_56c013b1a9" y="246.996271"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="649.503452" xlink:href="#C4_0_cbaeea66ea" y="190.813563"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="640.23112" xlink:href="#C4_0_56c013b1a9" y="190.813563"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="646.603658" xlink:href="#C4_0_cbaeea66ea" y="195.608936"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="631.816717" xlink:href="#C4_0_56c013b1a9" y="195.608936"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="642.346318" xlink:href="#C4_0_cbaeea66ea" y="307.629253"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="643.711389" xlink:href="#C4_0_56c013b1a9" y="307.629253"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="640.499877" xlink:href="#C4_0_cbaeea66ea" y="208.566013"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="643.644886" xlink:href="#C4_0_56c013b1a9" y="208.566013"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="640.924142" xlink:href="#C4_0_cbaeea66ea" y="236.871217"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="641.476195" xlink:href="#C4_0_56c013b1a9" y="236.871217"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="647.927393" xlink:href="#C4_0_cbaeea66ea" y="204.921565"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="644.665343" xlink:href="#C4_0_56c013b1a9" y="204.921565"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="635.139978" xlink:href="#C4_0_cbaeea66ea" y="222.121266"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="642.248136" xlink:href="#C4_0_56c013b1a9" y="222.121266"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="640.864215" xlink:href="#C4_0_cbaeea66ea" y="185.219633"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="649.311864" xlink:href="#C4_0_56c013b1a9" y="185.219633"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="649.849965" xlink:href="#C4_0_cbaeea66ea" y="150.450621"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="640.638034" xlink:href="#C4_0_56c013b1a9" y="150.450621"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="644.243909" xlink:href="#C4_0_cbaeea66ea" y="278.555335"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="630.300969" xlink:href="#C4_0_56c013b1a9" y="278.555335"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="631.351261" xlink:href="#C4_0_cbaeea66ea" y="202.066996"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="642.517633" xlink:href="#C4_0_56c013b1a9" y="202.066996"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="627.130877" xlink:href="#C4_0_cbaeea66ea" y="210.784076"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="630.742367" xlink:href="#C4_0_56c013b1a9" y="210.784076"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="631.054337" xlink:href="#C4_0_cbaeea66ea" y="278.567443"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="631.175821" xlink:href="#C4_0_56c013b1a9" y="278.567443"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="646.860781" xlink:href="#C4_0_cbaeea66ea" y="246.967553"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="646.416406" xlink:href="#C4_0_56c013b1a9" y="246.967553"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="640.558154" xlink:href="#C4_0_cbaeea66ea" y="293.023169"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="644.189262" xlink:href="#C4_0_56c013b1a9" y="293.023169"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="648.292345" xlink:href="#C4_0_cbaeea66ea" y="188.859446"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="645.952417" xlink:href="#C4_0_56c013b1a9" y="188.859446"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="641.973182" xlink:href="#C4_0_cbaeea66ea" y="196.281563"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="627.07897" xlink:href="#C4_0_56c013b1a9" y="196.281563"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="643.628242" xlink:href="#C4_0_cbaeea66ea" y="263.9663"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="628.758756" xlink:href="#C4_0_56c013b1a9" y="263.9663"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="640.521946" xlink:href="#C4_0_cbaeea66ea" y="234.912236"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="642.339311" xlink:href="#C4_0_56c013b1a9" y="234.912236"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="640.988051" xlink:href="#C4_0_cbaeea66ea" y="340.218479"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="637.099455" xlink:href="#C4_0_56c013b1a9" y="340.218479"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="635.31085" xlink:href="#C4_0_cbaeea66ea" y="246.953886"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="630.691432" xlink:href="#C4_0_56c013b1a9" y="246.953886"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#4c72b0;" x="627.729206" xlink:href="#C4_0_cbaeea66ea" y="278.549155"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#4c72b0;" x="637.948529" xlink:href="#C4_0_56c013b1a9" y="278.549155"/> </g> </g> <g id="PathCollection_8"> @@ -2702,85 +2702,85 @@ C -2.236584 -1.29895 -2.5 -0.663008 -2.5 0 C -2.5 0.663008 -2.236584 1.29895 -1.767767 1.767767 C -1.29895 2.236584 -0.663008 2.5 0 2.5 z -" id="C5_0_25efba35e2"/> +" id="C5_0_210b45a4e1"/> </defs> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="733.84608" xlink:href="#C5_0_25efba35e2" y="205.123489"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="740.338326" xlink:href="#C5_0_210b45a4e1" y="205.123489"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="741.282928" xlink:href="#C5_0_25efba35e2" y="269.413549"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="733.60375" xlink:href="#C5_0_210b45a4e1" y="269.413549"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="741.367481" xlink:href="#C5_0_25efba35e2" y="187.358818"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="737.572804" xlink:href="#C5_0_210b45a4e1" y="187.358818"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="729.180363" xlink:href="#C5_0_25efba35e2" y="187.84272"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="724.324087" xlink:href="#C5_0_210b45a4e1" y="187.84272"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="725.711059" xlink:href="#C5_0_25efba35e2" y="234.160756"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="725.322845" xlink:href="#C5_0_210b45a4e1" y="234.160756"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="729.710243" xlink:href="#C5_0_25efba35e2" y="251.163838"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="747.163636" xlink:href="#C5_0_210b45a4e1" y="251.163838"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="747.163636" xlink:href="#C5_0_25efba35e2" y="164.749615"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="732.435562" xlink:href="#C5_0_210b45a4e1" y="164.749615"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="745.76086" xlink:href="#C5_0_25efba35e2" y="254.741197"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="742.964507" xlink:href="#C5_0_210b45a4e1" y="254.741197"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="730.516978" xlink:href="#C5_0_25efba35e2" y="166.697991"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="738.025184" xlink:href="#C5_0_210b45a4e1" y="166.697991"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="741.103652" xlink:href="#C5_0_25efba35e2" y="175.221295"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="731.217025" xlink:href="#C5_0_210b45a4e1" y="175.221295"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="731.263442" xlink:href="#C5_0_25efba35e2" y="240.360593"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="729.212931" xlink:href="#C5_0_210b45a4e1" y="240.360593"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="729.14624" xlink:href="#C5_0_25efba35e2" y="208.74744"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="727.449826" xlink:href="#C5_0_210b45a4e1" y="208.74744"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="743.312936" xlink:href="#C5_0_25efba35e2" y="234.172585"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="724.192658" xlink:href="#C5_0_210b45a4e1" y="234.172585"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="741.709241" xlink:href="#C5_0_25efba35e2" y="166.719963"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="727.035455" xlink:href="#C5_0_210b45a4e1" y="166.719963"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="729.359299" xlink:href="#C5_0_25efba35e2" y="206.813301"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="737.130465" xlink:href="#C5_0_210b45a4e1" y="206.813301"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="732.050251" xlink:href="#C5_0_25efba35e2" y="212.772616"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="737.721393" xlink:href="#C5_0_210b45a4e1" y="212.772616"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="724.854083" xlink:href="#C5_0_25efba35e2" y="158.300318"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="744.54167" xlink:href="#C5_0_210b45a4e1" y="158.300318"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="744.956943" xlink:href="#C5_0_25efba35e2" y="223.381319"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="730.20411" xlink:href="#C5_0_210b45a4e1" y="223.381319"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="739.807474" xlink:href="#C5_0_25efba35e2" y="220.444186"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="724.902241" xlink:href="#C5_0_210b45a4e1" y="220.444186"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="728.761014" xlink:href="#C5_0_25efba35e2" y="288.823302"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="728.259834" xlink:href="#C5_0_210b45a4e1" y="288.823302"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="743.595589" xlink:href="#C5_0_25efba35e2" y="269.400828"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="732.049746" xlink:href="#C5_0_210b45a4e1" y="269.400828"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="727.620926" xlink:href="#C5_0_25efba35e2" y="254.797829"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="724.479442" xlink:href="#C5_0_210b45a4e1" y="254.797829"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="740.025616" xlink:href="#C5_0_25efba35e2" y="211.304124"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="729.409629" xlink:href="#C5_0_210b45a4e1" y="211.304124"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="732.648138" xlink:href="#C5_0_25efba35e2" y="194.48263"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="735.811378" xlink:href="#C5_0_210b45a4e1" y="194.48263"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="731.495627" xlink:href="#C5_0_25efba35e2" y="254.846716"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="737.565829" xlink:href="#C5_0_210b45a4e1" y="254.846716"/> </g> - <g clip-path="url(#p24d2ef447f)"> - <use style="fill:#dd8452;" x="725.983054" xlink:href="#C5_0_25efba35e2" y="269.446309"/> + <g clip-path="url(#p078fcafce0)"> + <use style="fill:#dd8452;" x="742.497576" xlink:href="#C5_0_210b45a4e1" y="269.446309"/> </g> </g> <g id="text_10"> @@ -2805,7 +2805,7 @@ Q 8.796875 56.9375 3.125 58.453125 z " id="ArialMT-42"/> </defs> - <g style="fill:#262626;" transform="translate(436.527535 109.709863)scale(0.12 -0.12)"> + <g style="fill:#262626;" transform="translate(436.16219 109.709863)scale(0.12 -0.12)"> <use xlink:href="#ArialMT-42"/> <use x="38.916016" xlink:href="#ArialMT-42"/> <use x="77.832031" xlink:href="#ArialMT-42"/> @@ -2813,14 +2813,14 @@ z </g> <g id="text_11">  - <g style="fill:#262626;" transform="translate(194.595105 70.900587)scale(0.12 -0.12)"> + <g style="fill:#262626;" transform="translate(194.548862 70.900587)scale(0.12 -0.12)"> <use xlink:href="#ArialMT-42"/> <use x="38.916016" xlink:href="#ArialMT-42"/> </g> </g> <g id="text_12">  - <g style="fill:#262626;" transform="translate(681.464653 123.579797)scale(0.12 -0.12)"> + <g style="fill:#262626;" transform="translate(680.780205 123.579797)scale(0.12 -0.12)"> <use xlink:href="#ArialMT-110"/> <use x="55.615234" xlink:href="#ArialMT-115"/> </g> @@ -2910,10 +2910,10 @@ C -3.464901 -2.012324 -3.872983 -1.027127 -3.872983 0 C -3.872983 1.027127 -3.464901 2.012324 -2.738613 2.738613 C -2.012324 3.464901 -1.027127 3.872983 0 3.872983 z -" id="m332170e5fd" style="stroke:#4c72b0;"/> +" id="md5fa658d96" style="stroke:#4c72b0;"/> </defs> <g> - <use style="fill:#4c72b0;stroke:#4c72b0;" x="665.974063" xlink:href="#m332170e5fd" y="83.487344"/> + <use style="fill:#4c72b0;stroke:#4c72b0;" x="665.974063" xlink:href="#md5fa658d96" y="83.487344"/> </g> </g> <g id="text_14"> @@ -2946,10 +2946,10 @@ C -3.464901 -2.012324 -3.872983 -1.027127 -3.872983 0 C -3.872983 1.027127 -3.464901 2.012324 -2.738613 2.738613 C -2.012324 3.464901 -1.027127 3.872983 0 3.872983 z -" id="m8557ba0fcf" style="stroke:#dd8452;"/> +" id="mcda2f6a9ba" style="stroke:#dd8452;"/> </defs> <g> - <use style="fill:#dd8452;stroke:#dd8452;" x="665.974063" xlink:href="#m8557ba0fcf" y="99.310156"/> + <use style="fill:#dd8452;stroke:#dd8452;" x="665.974063" xlink:href="#mcda2f6a9ba" y="99.310156"/> </g> </g> <g id="text_15"> @@ -2971,7 +2971,7 @@ z </g> </g> <defs> - <clipPath id="p24d2ef447f"> + <clipPath id="p078fcafce0"> <rect height="326.16" width="669.6" x="108" y="51.84"/> </clipPath> </defs>

Create Annotation, rename box_pairs and box_ to group In this PR, more responsibility is extracted from `Annotator` towards an `Annotation` class. `Annotation` objects now link the plot structures (boxes for boxplots) to the statistical result or custom annotation. Also, it was time to rename the variables from `box_` to `group_` as the package is definitely not only about boxplots.

2021-07-21T18:29:52

0.0

[]

trevismd/statannotations

trevismd__statannotations-13

0005f989961a57d712ccecb671499d5ab4068834

diff --git a/CHANGELOG.md b/CHANGELOG.md index b3c5b18..dbf1858 100644 --- a/CHANGELOG.md +++ b/CHANGELOG.md @@ -1,5 +1,14 @@ -## v3 -### v3.0 +## v0.3 +### v0.3.2 + - Fix `simple` format outputs + - Fix `ImportError` when applying a multiple comparison correction without statsmodels. + - Multiple comparison correction is `None` by default, as `statsmodels` is an additional dependency. + +### v0.3.1 + - Added support of functions returning more than the two expected values when used in `StatTest` + - Fix version numbers in CHANGELOG + +### v0.3.0 (Explicitly Requires Python 3.6) - Refactoring with implementation of `StatTest` @@ -9,6 +18,6 @@ ([#5](https://github.com/trevismd/statannotations/pull/5) by [JosephLalli](https://github.com/JosephLalli), fixes https://github.com/webermarcolivier/statannot/issues/21). - Add this CHANGELOG -## v2 -### v2.8 - - Fix bug on group/box named 0, fixes https://github.com/trevismd/statannotations/issues/10, originally in https://github.com/webermarcolivier/statannot/issues/78. Independently fixed in https://github.com/webermarcolivier/statannot/pull/73 before this issue. \ No newline at end of file +## v0.2 +### v0.2.8 + - Fix bug on group/box named 0, fixes https://github.com/trevismd/statannotations/issues/10, originally in https://github.com/webermarcolivier/statannot/issues/78. Independently fixed in https://github.com/webermarcolivier/statannot/pull/73 before this issue. diff --git a/coverage.svg b/coverage.svg index eb83f3f..dd6df12 100644 --- a/coverage.svg +++ b/coverage.svg @@ -9,13 +9,13 @@ </mask> <g mask="url(#a)"> <path fill="#555" d="M0 0h63v20H0z"/> - <path fill="#fe7d37" d="M63 0h36v20H63z"/> + <path fill="#dfb317" d="M63 0h36v20H63z"/> <path fill="url(#b)" d="M0 0h99v20H0z"/> </g> <g fill="#fff" text-anchor="middle" font-family="DejaVu Sans,Verdana,Geneva,sans-serif" font-size="11"> <text x="31.5" y="15" fill="#010101" fill-opacity=".3">coverage</text> <text x="31.5" y="14">coverage</text> - <text x="80" y="15" fill="#010101" fill-opacity=".3">59%</text> - <text x="80" y="14">59%</text> + <text x="80" y="15" fill="#010101" fill-opacity=".3">61%</text> + <text x="80" y="14">61%</text> </g> </svg> diff --git a/statannotations/__init__.py b/statannotations/__init__.py index 924768d..e099ae2 100644 --- a/statannotations/__init__.py +++ b/statannotations/__init__.py @@ -1,4 +1,4 @@ -__all__ = ["__version__", "stats", "statannotations"] +__all__ = ["__version__", "stats", "statannotations", "add_stat_annotation"] from .statannotations import add_stat_annotation from ._version import __version__ diff --git a/statannotations/_version.py b/statannotations/_version.py index 493f741..f9aa3e1 100644 --- a/statannotations/_version.py +++ b/statannotations/_version.py @@ -1,1 +1,1 @@ -__version__ = "0.3.0" +__version__ = "0.3.2" diff --git a/statannotations/statannotations.py b/statannotations/statannotations.py index b46745d..34f64f0 100644 --- a/statannotations/statannotations.py +++ b/statannotations/statannotations.py @@ -29,7 +29,7 @@ def add_stat_annotation(ax, plot='boxplot', data=None, x=None, y=None, text_annot_custom=None, loc='inside', show_test_name=True, pvalue_thresholds=DEFAULT, stats_params: dict = None, - comparisons_correction='bonferroni', + comparisons_correction=None, num_comparisons='auto', use_fixed_offset=False, line_offset_to_box=None, line_offset=None, line_height=0.02, text_offset=1, color='0.2', @@ -52,7 +52,7 @@ def add_stat_annotation(ax, plot='boxplot', data=None, x=None, y=None, :param data: seaborn plot's data :param x: seaborn plot's x :param y: seaborn plot's y - :param test: StatResult or name from list of scipy functions supported + :param test: `StatTest` instance or name from list of scipy functions supported :param hue: seaborn plot's hue :param order: seaborn plot's order :param hue_order: seaborn plot's hue_order @@ -65,7 +65,7 @@ def add_stat_annotation(ax, plot='boxplot', data=None, x=None, y=None, :param test_short_name: How the test name should show on the plot, if show_test_name is True (default). Default is the full test name - :param pvalue_format_string: defaults to `"{.3e}"` + :param pvalue_format_string: defaults to `"{:.3e}"` or `"{:.2f}"` for `"simple"` format :param pvalue_thresholds: list of lists, or tuples. Default is: For "star" text_format: `[[1e-4, "****"], [1e-3, "***"], [1e-2, "**"], [0.05, "*"], [1, "ns"]]`. @@ -186,7 +186,7 @@ def get_box_data(b_plotter, box_name): else: pvalue_thresholds = [[1e-5, "1e-5"], [1e-4, "1e-4"], [1e-3, "0.001"], [1e-2, "0.01"], - [5e-2, "0.05"], [1, "ns"]] + [5e-2, "0.05"]] if stats_params is None: stats_params = dict() @@ -233,7 +233,12 @@ def get_box_data(b_plotter, box_name): pass elif isinstance(comparisons_correction, str): check_valid_correction_name(comparisons_correction) - comparisons_correction = ComparisonsCorrection(comparisons_correction) + try: + comparisons_correction = ComparisonsCorrection(comparisons_correction) + except ImportError as e: + raise ImportError(f"{e} Please install statsmodels or pass " + f"`comparisons_correction=None`.") + elif not(isinstance(comparisons_correction, ComparisonsCorrection)): raise ValueError("comparisons_correction must be a statmodels " "method name or a ComparisonCorrection instance") @@ -465,7 +470,10 @@ def get_transform_func(ax, kind): text = pval_annotation_text(result, pvalue_thresholds) elif text_format == 'simple': if show_test_name: - test_short_name = show_test_name and test_short_name or test + if test_short_name is None: + test_short_name = (test + if isinstance(test, str) + else test.short_name) else: test_short_name = "" text = simple_text(result, simple_format_string, diff --git a/statannotations/stats/ComparisonsCorrection.py b/statannotations/stats/ComparisonsCorrection.py index 3664649..c23a62c 100644 --- a/statannotations/stats/ComparisonsCorrection.py +++ b/statannotations/stats/ComparisonsCorrection.py @@ -6,12 +6,9 @@ try: from statsmodels.stats.multitest import multipletests -except ImportError as statsmodel_import_error: +except ImportError: multipletests = None -else: - statsmodel_import_error = None - # For user convenience methods_names = {'bonferroni': 'Bonferroni', 'bonf': 'Bonferroni', @@ -84,7 +81,9 @@ def __init__(self, method: Union[str, callable], alpha: float = 0.05, if isinstance(method, str): if multipletests is None: - raise ImportError(statsmodel_import_error) + raise ImportError("The statsmodels package is required to use " + "one of the multiple comparisons correction " + "methods proposed in statannotations.") self.name = name if name is not None else method self.method, self.type = get_correction_parameters(method) diff --git a/statannotations/stats/StatTest.py b/statannotations/stats/StatTest.py index 2898680..ef97e12 100644 --- a/statannotations/stats/StatTest.py +++ b/statannotations/stats/StatTest.py @@ -83,7 +83,9 @@ def __init__(self, func: Callable, test_long_name: str, Wrapper for any statistical function to use with statannotations. The function must be callable with two 1D collections of data points as first and second arguments, can have any number of additional arguments - , and must return a) a statistical value, and b) a pvalue. + , and must at least return, in order, + a) a statistical value, and + b) a pvalue. :param func: Function to be called @@ -117,7 +119,11 @@ def __call__(self, box_data1, box_data2, alpha=0.05, **stat_params): stat, pval = self._func(box_data1, box_data2, *self.args, - **{**self.kwargs, **stat_params}) + **{**self.kwargs, **stat_params})[:2] return StatResult(self._test_long_name, self._test_short_name, self._stat_name, stat, pval, alpha=alpha) + + @property + def short_name(self): + return self._test_short_name

Support functions with more returned values in `StatTest` This fixes the bug in the situation where the function passed to StatTest would return more than two_values. And includes minor changes from master branch

2021-06-17T07:15:42

0.0

[]

RalphTro/epcis-event-hash-generator

RalphTro__epcis-event-hash-generator-61

e38f16a390d5caa71bdfe34addabdc2f92831896

diff --git a/epcis_event_hash_generator/hash_generator.py b/epcis_event_hash_generator/hash_generator.py index 050de39..14ac917 100644 --- a/epcis_event_hash_generator/hash_generator.py +++ b/epcis_event_hash_generator/hash_generator.py @@ -58,23 +58,25 @@ def _fix_time_stamp_format(timestamp): def _child_to_pre_hash_string(child, sub_child_order): + logging.debug("Processing '%s'", child) text = "" grand_child_text = "" if sub_child_order: grand_child_text = _recurse_through_children_in_order(child[2], sub_child_order) if child[1]: text = child[1].strip() - if child[0].lower().find("time") > 0 and child[0].lower().find("offset") < 0: + if child[0].lower().find("time") >= 0 and child[0].lower().find("offset") < 0: text = _fix_time_stamp_format(text) else: text = _canonize_value(text) if text: text = "=" + text - logging.debug("Adding text '%s'", text) if text or grand_child_text: - return child[0] + text + grand_child_text + re = child[0] + text + grand_child_text + logging.debug("pre hash string element: '%s'", text) + return re return "" @@ -129,6 +131,7 @@ def _canonize_value(text): if converted: logging.debug("Converted %s to %s", text, converted) return converted + logging.debug("No canonical form for '%s'", text) return text

Date format with pre-hash string Hi, In rule 8 of the algorithm, you say: ``` All timestamps SHALL be expressed in UTC; the zero UTC offset SHALL be expressed with the capital letter 'Z'. ``` which means that timestamps can have a time offset, right? This is indeed the case of the `time` field of a `sensorMetadata`. If you look at the example `epcisDocWithSensorDataObjectEvent.jsonld`, the pre-hash string is: ``` sensorMetadatatime=2019-04-02T14:05:00.000+01:00 ``` (It keeps the offset so everything is working as expected) However, when you look at the pre-hash string of the last event in `SensorDataExamples.xml`, we have this field: ``` <declarationTime>2020-09-29T15:00:00.000+02:00</declarationTime> ``` But in the pre-hash string, it converts the date like this: ``` declarationTime=2020-09-29T13:00:00.000 ``` Is it the expected output? If it is the case, do I miss something in the algorithm? Thanks a lot.

Dear @clementh59 , Good catch! You are correct. *All* timestamps should be converted to zero UTC. I hope you do not need a fix urgently? ;-) Kind regards and thanks again; Ralph Dear @RalphTro , I see, thanks! No, that's ok, I just want to be sure I understand the algorithm correctly :) Regards, Clément

2021-10-18T21:18:38

0.0

[]

RalphTro/epcis-event-hash-generator

RalphTro__epcis-event-hash-generator-45

ce6ed045f5ec5147bcb39a9a120002996b06a8f2

diff --git a/.flake8 b/.flake8 index fe83fbf..8318a8f 100644 --- a/.flake8 +++ b/.flake8 @@ -5,3 +5,6 @@ exclude = context.py max-complexity = 10 max-line-length = 120 +per-file-ignores = + tests/dl_normaliser_test.py:E501 + epcis_event_hash_generator/dl_normaliser.py:E501,C901 \ No newline at end of file diff --git a/epcis_event_hash_generator/dl_normaliser.py b/epcis_event_hash_generator/dl_normaliser.py new file mode 100644 index 0000000..75efdcd --- /dev/null +++ b/epcis_event_hash_generator/dl_normaliser.py @@ -0,0 +1,445 @@ +""" +GS1 Digital Link Normaliser. + +This script accepts any valid URI scheme accommodating a GS1 ID, +i.e. EPC URIs, EPC Class URIs, EPC ID Pattern URIs, or GS1 Digital Link URIs. +It converts all of them into one normalised form, meaning that it +(a) converts it into a canonical GS1 DL URI, +(b) ensures that it only contains the most fine-granular ID level, +(c) strips off any further attributes. + +.. module:: dl_normaliser + :synopsis: Normalises the gs1 id formats to digital link for https://github.com/RalphTro/epcis-event-hash-generator/ + +.. moduleauthor:: Ralph Troeger <[email protected]> + +Copyright 2019-2021 Ralph Troeger + +This program is free software: you can redistribute it and/or modify +it under the terms given in the LICENSE file. + +This program is distributed in the hope that it will be useful, but +WITHOUT ANY WARRANTY; without even the implied warranty of +MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the LICENSE +file for details. + +""" + +from re import match +from gtin import GTIN +import logging + + +def __web_uri_percent_encoder(input): + """ + Function percent-encodes URL-unsafe characters in GS1 Digital Link URIs. + + Table 7-1 in the GS1 Digital Link Standard requires + the following symbols to be percent-encoded: + '!', space, '#', '%', '&', '(', ')', '*', '+', ',', '/', ':' + EPC URIs already prohibit some of them (e.g. '#') + Function 'webURIPercentEncoder' is called to ensure that + data elements accommodating these symbols are percent-encoded. + + Parameters + ---------- + input : str + Character requiring percent-encoding. + + Returns + ------- + str + Percent-encoded equivalent of character. + """ + return ( + input.replace('!', '%21') + .replace('(', '%28') + .replace(')', '%29') + .replace('*', '%2A') + .replace('+', '%2B') + .replace(',', '%2C') + .replace(':', '%3A') + ) + + +def normaliser(uri): + """ + Function converts any standard URI conveying a GS1 Key in Canonical GS1 DL URI. + + Function 'normaliser' expects any URI to be used in EPCIS events + that convey a GS1 key, i.e. EPC URIs, EPC Class URIs, + EPC ID Pattern URIs, or GS1 Digital Link URIs. + It returns a corresponding, constrained version of a + canonical GS1 Digital Link URI, i.e. with + the lowest level of identification and without CPV/query string. + + Parameters + ---------- + uri : str + Valid EPC URI, EPC Pattern URI, EPC Class URI, GS1 Digital Link URI. + + Returns + ------- + str + Constrained, canonicalised GS1 Digital Link URI equivalent. + None + """ + if not isinstance(uri, str): + logging.warning("dl normaliser called with non-string argument") + return None + + try: + partition = uri.index('.') + except ValueError: + logging.debug("No '.' in %s. Not a normalisable uri.", uri) + return None + + # EPC URIs + if match( + r'^urn:epc:id:sgtin:((\d{6}\.\d{7})|(\d{7}\.\d{6})|(\d{8}\.\d{5})|(\d{9}\.\d{4})|(\d{10}\.\d{3})|(\d{11}\.\d{2})|(\d{12}\.\d{1}))\.(\%2[125-9A-Fa-f]|\%3[0-9A-Fa-f]|\%4[1-9A-Fa-f]|\%5[0-9AaFf]|\%6[1-9A-Fa-f]|\%7[0-9Aa]|[!\')(*+,.0-9:;=A-Za-z_-]){1,20}$', + uri) is not None: + gs1companyprefix = uri[17:partition] + itemref = uri[(partition + 1):(partition + 1 + (13 - len(gs1companyprefix)))] + rawGTIN = itemref[0:1] + gs1companyprefix + itemref[1:] + serial = uri[32:] + return ('https://id.gs1.org/01/' + str(GTIN(raw=rawGTIN)) + '/21/' + __web_uri_percent_encoder(serial)) + + if match( + r'^urn:epc:id:sscc:((\d{6}\.\d{11}$)|(\d{7}\.\d{10}$)|(\d{8}\.\d{9}$)|(\d{9}\.\d{8}$)|(\d{10}\.\d{7}$)|(\d{11}\.\d{6}$)|(\d{12}\.\d{5}$))', + uri) is not None: + gs1companyprefix = uri[16:partition] + serialref = uri[(partition + 1):] + rawSSCC = uri[(partition + 1):(partition + 2)] + gs1companyprefix + serialref + return ('https://id.gs1.org/00/' + str(GTIN(raw=rawSSCC))) + + if match( + r'^urn:epc:id:sgln:((\d{6}\.\d{6})|(\d{7}\.\d{5})|(\d{8}\.\d{4})|(\d{9}\.\d{3})|(\d{10}\.\d{2})|(\d{11}\.\d{1})|(\d{12}\.))\.(\%2[125-9A-Fa-f]|\%3[0-9A-Fa-f]|\%4[1-9A-Fa-f]|\%5[0-9AaFf]|\%6[1-9A-Fa-f]|\%7[0-9Aa]|[!\')(*+,.0-9:;=A-Za-z_-]){1,20}$', + uri) is not None: + gs1companyprefix = uri[16:partition] + locationref = uri[(partition + 1):(partition + 1 + + (12 - len(gs1companyprefix)))] + rawGLN = gs1companyprefix + locationref + extension = uri[30:] + if extension == '0': + return ('https://id.gs1.org/414/' + str(GTIN(raw=rawGLN))) + else: + return ('https://id.gs1.org/414/' + str(GTIN(raw=rawGLN)) + '/254/' + __web_uri_percent_encoder(extension)) + + if match( + r'^urn:epc:id:grai:(([\d]{6}\.[\d]{6})|([\d]{7}\.[\d]{5})|([\d]{8}\.[\d]{4})|([\d]{9}\.[\d]{3})|([\d]{10}\.[\d]{2})|([\d]{11}\.[\d]{1})|([\d]{12}\.\.))\.(\%2[125-9A-Fa-f]|\%3[0-9A-Fa-f]|\%4[1-9A-Fa-f]|\%5[0-9AaFf]|\%6[1-9A-Fa-f]|\%7[0-9Aa]|[!\')(*+,.0-9:;=A-Za-z_-]){1,16}$', + uri) is not None: + gs1companyprefix = uri[16:partition] + assetref = uri[(partition + 1):(partition + 1 + (12 - len(gs1companyprefix)))] + rawGRAI = '0' + gs1companyprefix + assetref + serial = uri[30:] + return ('https://id.gs1.org/8003/' + str(GTIN(raw=rawGRAI)) + __web_uri_percent_encoder(serial)) + + if match( + r'^urn:epc:id:giai:(([\d]{6}\.(\%2[125-9A-Fa-f]|\%3[0-9A-Fa-f]|\%4[1-9A-Fa-f]|\%5[0-9AaFf]|\%6[1-9A-Fa-f]|\%7[0-9Aa]|[!\')(*+,.0-9:;=A-Za-z_-]){1,24})|([\d]{7}\.(\%2[125-9A-Fa-f]|\%3[0-9A-Fa-f]|\%4[1-9A-Fa-f]|\%5[0-9AaFf]|\%6[1-9A-Fa-f]|\%7[0-9Aa]|[!\')(*+,.0-9:;=A-Za-z_-]){1,23})|([\d]{8}\.(\%2[125-9A-Fa-f]|\%3[0-9A-Fa-f]|\%4[1-9A-Fa-f]|\%5[0-9AaFf]|\%6[1-9A-Fa-f]|\%7[0-9Aa]|[!\')(*+,.0-9:;=A-Za-z_-]){1,22})|([\d]{9}\.(\%2[125-9A-Fa-f]|\%3[0-9A-Fa-f]|\%4[1-9A-Fa-f]|\%5[0-9AaFf]|\%6[1-9A-Fa-f]|\%7[0-9Aa]|[!\')(*+,.0-9:;=A-Za-z_-]){1,21})|([\d]{10}\.(\%2[125-9A-Fa-f]|\%3[0-9A-Fa-f]|\%4[1-9A-Fa-f]|\%5[0-9AaFf]|\%6[1-9A-Fa-f]|\%7[0-9Aa]|[!\')(*+,.0-9:;=A-Za-z_-]){1,20})|([\d]{11}\.(\%2[125-9A-Fa-f]|\%3[0-9A-Fa-f]|\%4[1-9A-Fa-f]|\%5[0-9AaFf]|\%6[1-9A-Fa-f]|\%7[0-9Aa]|[!\')(*+,.0-9:;=A-Za-z_-]){1,19})|([\d]{12}\.(\%2[125-9A-Fa-f]|\%3[0-9A-Fa-f]|\%4[1-9A-Fa-f]|\%5[0-9AaFf]|\%6[1-9A-Fa-f]|\%7[0-9Aa]|[!\')(*+,.0-9:;=A-Za-z_-]){1,18}))$', + uri) is not None: + gs1companyprefix = uri[16:partition] + assetref = uri[(partition + 1):] + return ('https://id.gs1.org/8004/' + gs1companyprefix + __web_uri_percent_encoder(assetref)) + + if match( + r'^urn:epc:id:gsrn:(([\d]{6}\.[\d]{11}$)|([\d]{7}\.[\d]{10}$)|([\d]{8}\.[\d]{9}$)|([\d]{9}\.[\d]{8}$)|([\d]{10}\.[\d]{7}$)|([\d]{11}\.[\d]{6}$)|([\d]{12}\.[\d]{5}$))', + uri) is not None: + gs1companyprefix = uri[16:partition] + serviceref = uri[(partition + 1):] + rawGSRN = gs1companyprefix + serviceref + return ('https://id.gs1.org/8018/' + str(GTIN(raw=rawGSRN))) + + if match( + r'^urn:epc:id:gsrnp:(([\d]{6}\.[\d]{11}$)|([\d]{7}\.[\d]{10}$)|([\d]{8}\.[\d]{9}$)|([\d]{9}\.[\d]{8}$)|([\d]{10}\.[\d]{7}$)|([\d]{11}\.[\d]{6}$)|([\d]{12}\.[\d]{5}$))', + uri) is not None: + gs1companyprefix = uri[17:partition] + serviceref = uri[(partition + 1):] + rawGSRNP = gs1companyprefix + serviceref + return ('https://id.gs1.org/8017/' + str(GTIN(raw=rawGSRNP))) + + if match( + r'^urn:epc:id:gdti:(([\d]{6}\.[\d]{6})|([\d]{7}\.[\d]{5})|([\d]{8}\.[\d]{4})|([\d]{9}\.[\d]{3})|([\d]{10}\.[\d]{2})|([\d]{11}\.[\d]{1})|([\d]{12}\.\.))(\%2[125-9A-Fa-f]|\%3[0-9A-Fa-f]|\%4[1-9A-Fa-f]|\%5[0-9AaFf]|\%6[1-9A-Fa-f]|\%7[0-9Aa]|[!\')(*+,.0-9:;=A-Za-z_-]){1,20}$', + uri) is not None: + gs1companyprefix = uri[16:partition] + documenttype = uri[(partition + 1):(partition + 1 + + (12 - len(gs1companyprefix)))] + rawGDTI = gs1companyprefix + documenttype + serial = uri[30:] + return ('https://id.gs1.org/253/' + str(GTIN(raw=rawGDTI)) + __web_uri_percent_encoder(serial)) + + if match( + r'^urn:epc:id:cpi:((\d{6}\.(\%2[3dfDF]|\%3[0-9]|\%4[1-9A-Fa-f]|\%5[0-9Aa]|[0-9A-Z-]){1,24})|(\d{7}\.(\%2[3dfDF]|\%3[0-9]|\%4[1-9A-Fa-f]|\%5[0-9Aa]|[0-9A-Z-]){1,23})|(\d{8}\.(\%2[3dfDF]|\%3[0-9]|\%4[1-9A-Fa-f]|\%5[0-9Aa]|[0-9A-Z-]){1,22})|(\d{9}\.(\%2[3dfDF]|\%3[0-9]|\%4[1-9A-Fa-f]|\%5[0-9Aa]|[0-9A-Z-]){1,21})|(\d{10}\.(\%2[3dfDF]|\%3[0-9]|\%4[1-9A-Fa-f]|\%5[0-9Aa]|[0-9A-Z-]){1,20})|(\d{11}\.(\%2[3dfDF]|\%3[0-9]|\%4[1-9A-Fa-f]|\%5[0-9Aa]|[0-9A-Z-]){1,19})|(\d{12}\.(\%2[3dfDF]|\%3[0-9]|\%4[1-9A-Fa-f]|\%5[0-9Aa]|[0-9A-Z-]){1,18}))\.[\d]{1,12}$', + uri) is not None: + gs1companyprefix = uri[15:partition] + separator = uri.rfind('.') + cpref = uri[(partition + 1):(separator)] + rawCPI = gs1companyprefix + cpref + serial = uri[(separator + 1):] + return ('https://id.gs1.org/8010/' + __web_uri_percent_encoder(rawCPI) + '/8011/' + serial) + + if match( + r'^urn:epc:id:sgcn:(([\d]{6}\.[\d]{6})|([\d]{7}\.[\d]{5})|([\d]{8}\.[\d]{4})|([\d]{9}\.[\d]{3})|([\d]{10}\.[\d]{2})|([\d]{11}\.[\d]{1})|([\d]{12}\.))\.[\d]{1,12}$', + uri) is not None: + gs1companyprefix = uri[16:partition] + couponref = uri[(partition + 1):(partition + 1 + (12 - len(gs1companyprefix)))] + rawSGCN = gs1companyprefix + couponref + serial = uri[30:] + return ('https://id.gs1.org/255/' + str(GTIN(raw=rawSGCN)) + serial) + + if match( + r'^urn:epc:id:ginc:([\d]{6}\.(\%2[125-9A-Fa-f]|\%3[0-9A-Fa-f]|\%4[1-9A-Fa-f]|\%5[0-9AaFf]|\%6[1-9A-Fa-f]|\%7[0-9Aa]|[!\')(*+,.0-9:;=A-Za-z_-]){1,24}|[\d]{7}\.(\%2[125-9A-Fa-f]|\%3[0-9A-Fa-f]|\%4[1-9A-Fa-f]|\%5[0-9AaFf]|\%6[1-9A-Fa-f]|\%7[0-9Aa]|[!\')(*+,.0-9:;=A-Za-z_-]){1,23}|[\d]{8}\.(\%2[125-9A-Fa-f]|\%3[0-9A-Fa-f]|\%4[1-9A-Fa-f]|\%5[0-9AaFf]|\%6[1-9A-Fa-f]|\%7[0-9Aa]|[!\')(*+,.0-9:;=A-Za-z_-]){1,22}|[\d]{9}\.(\%2[125-9A-Fa-f]|\%3[0-9A-Fa-f]|\%4[1-9A-Fa-f]|\%5[0-9AaFf]|\%6[1-9A-Fa-f]|\%7[0-9Aa]|[!\')(*+,.0-9:;=A-Za-z_-]){1,21}|[\d]{10}\.(\%2[125-9A-Fa-f]|\%3[0-9A-Fa-f]|\%4[1-9A-Fa-f]|\%5[0-9AaFf]|\%6[1-9A-Fa-f]|\%7[0-9Aa]|[!\')(*+,.0-9:;=A-Za-z_-]){1,20}|[\d]{11}\.(\%2[125-9A-Fa-f]|\%3[0-9A-Fa-f]|\%4[1-9A-Fa-f]|\%5[0-9AaFf]|\%6[1-9A-Fa-f]|\%7[0-9Aa]|[!\')(*+,.0-9:;=A-Za-z_-]){1,19}|[\d]{12}\.(\%2[125-9A-Fa-f]|\%3[0-9A-Fa-f]|\%4[1-9A-Fa-f]|\%5[0-9AaFf]|\%6[1-9A-Fa-f]|\%7[0-9Aa]|[!\')(*+,.0-9:;=A-Za-z_-]){1,18})$', + uri) is not None: + gs1companyprefix = uri[16:partition] + consignmentref = uri[(partition + 1):] + return ('https://id.gs1.org/401/' + gs1companyprefix + __web_uri_percent_encoder(consignmentref)) + + if match( + r'^urn:epc:id:gsin:(([\d]{6}\.[\d]{10}$)|([\d]{7}\.[\d]{9}$)|([\d]{8}\.[\d]{8}$)|([\d]{9}\.[\d]{7}$)|([\d]{10}\.[\d]{6}$)|([\d]{11}\.[\d]{5}$)|([\d]{12}\.[\d]{4}$))', + uri) is not None: + gs1companyprefix = uri[16:partition] + shipperref = uri[(partition + 1):] + rawGSIN = gs1companyprefix + shipperref + return ('https://id.gs1.org/402/' + str(GTIN(raw=rawGSIN))) + + if match( + r'^urn:epc:id:itip:(([\d]{6}\.[\d]{7})|([\d]{7}\.[\d]{6})|([\d]{8}\.[\d]{5})|([\d]{9}\.[\d]{4})|([\d]{10}\.[\d]{3})|([\d]{11}\.[\d]{2})|([\d]{12}\.[\d]{1}))\.[\d]{2}\.[\d]{2}\.(\%2[125-9A-Fa-f]|\%3[0-9A-Fa-f]|\%4[1-9A-Fa-f]|\%5[0-9AaFf]|\%6[1-9A-Fa-f]|\%7[0-9Aa]|[!\')(*+,.0-9:;=A-Za-z_-]){1,20}$', + uri) is not None: + gs1companyprefix = uri[16:partition] + itemref = uri[(partition + 1):(partition + 1 + (13 - len(gs1companyprefix)))] + rawGTIN = itemref[0:1] + gs1companyprefix + itemref[1:] + piece = uri[31:33] + total = uri[34:36] + serial = uri[37:] + return ('https://id.gs1.org/8006/' + str(GTIN(raw=rawGTIN)) + + piece + total + '/21/' + __web_uri_percent_encoder(serial)) + + if match( + r'^urn:epc:id:upui:((\d{6}\.\d{7})|(\d{7}\.\d{6})|(\d{8}\.\d{5})|(\d{9}\.\d{4})|(\d{10}\.\d{3})|(\d{11}\.\d{2})|(\d{12}\.\d{1}))\.(\%2[125-9A-Fa-f]|\%3[0-9A-Fa-f]|\%4[1-9A-Fa-f]|\%5[0-9AaFf]|\%6[1-9A-Fa-f]|\%7[0-9Aa]|[!\')(*+,.0-9:;=A-Za-z_-]){1,28}$', + uri) is not None: + gs1companyprefix = uri[16:partition] + itemref = uri[(partition + 1):(partition + 1 + (13 - len(gs1companyprefix)))] + rawGTIN = itemref[0:1] + gs1companyprefix + itemref[1:] + serial = uri[31:] + return ('https://id.gs1.org/01/' + str(GTIN(raw=rawGTIN)) + '/235/' + __web_uri_percent_encoder(serial)) + + if match( + r'^urn:epc:id:pgln:(([\d]{6}\.[\d]{6})|([\d]{7}\.[\d]{5})|([\d]{8}\.[\d]{4})|([\d]{9}\.[\d]{3})|([\d]{10}\.[\d]{2})|([\d]{11}\.[\d]{1})|([\d]{12}\.))$', + uri) is not None: + gs1companyprefix = uri[16:partition] + partyref = uri[(partition + 1):(partition + 1 + (12 - len(gs1companyprefix)))] + rawGLN = gs1companyprefix + partyref + return ('https://id.gs1.org/417/' + str(GTIN(raw=rawGLN))) + + # EPC Class URIs + if match( + r'^urn:epc:class:lgtin:(([\d]{6}\.[\d]{7})|([\d]{7}\.[\d]{6})|([\d]{8}\.[\d]{5})|([\d]{9}\.[\d]{4})|([\d]{10}\.[\d]{3})|([\d]{11}\.[\d]{2})|([\d]{12}\.[\d]{1}))\.(\%2[125-9A-Fa-f]|\%3[0-9A-Fa-f]|\%4[1-9A-Fa-f]|\%5[0-9AaFf]|\%6[1-9A-Fa-f]|\%7[0-9Aa]|[!\')(*+,.0-9:;=A-Za-z_-]){1,20}$', + uri) is not None: + gs1companyprefix = uri[20:partition] + itemref = uri[(partition + 1):(partition + 1 + (13 - len(gs1companyprefix)))] + rawGTIN = itemref[0:1] + gs1companyprefix + itemref[1:] + lot = uri[35:] + return ('https://id.gs1.org/01/' + str(GTIN(raw=rawGTIN)) + '/10/' + __web_uri_percent_encoder(lot)) + + # EPC ID Pattern URIs + if match( + r'^urn:epc:idpat:sgtin:((\d{6}\.\d{7})|(\d{7}\.\d{6})|(\d{8}\.\d{5})|(\d{9}\.\d{4})|(\d{10}\.\d{3})|(\d{11}\.\d{2})|(\d{12}\.\d{1}))\.\*$', + uri) is not None: + gs1companyprefix = uri[20:partition] + itemref = uri[(partition + 1):(partition + 1 + (13 - len(gs1companyprefix)))] + rawGTIN = itemref[0:1] + gs1companyprefix + itemref[1:] + return ('https://id.gs1.org/01/' + str(GTIN(raw=rawGTIN))) + + if match( + r'^urn:epc:idpat:grai:(([\d]{6}\.[\d]{6})|([\d]{7}\.[\d]{5})|([\d]{8}\.[\d]{4})|([\d]{9}\.[\d]{3})|([\d]{10}\.[\d]{2})|([\d]{11}\.[\d]{1})|([\d]{12}\.\.))\.\*$', + uri) is not None: + gs1companyprefix = uri[19:partition] + assetref = uri[(partition + 1):(partition + 1 + (12 - len(gs1companyprefix)))] + rawGRAI = '0' + gs1companyprefix + assetref + return ('https://id.gs1.org/8003/' + str(GTIN(raw=rawGRAI))) + + if match( + r'^urn:epc:idpat:gdti:(([\d]{6}\.[\d]{6})|([\d]{7}\.[\d]{5})|([\d]{8}\.[\d]{4})|([\d]{9}\.[\d]{3})|([\d]{10}\.[\d]{2})|([\d]{11}\.[\d]{1})|([\d]{12}\.\.))\.\*$', + uri) is not None: + gs1companyprefix = uri[19:partition] + documenttype = uri[(partition + 1):(partition + 1 + + (12 - len(gs1companyprefix)))] + rawGDTI = gs1companyprefix + documenttype + return ('https://id.gs1.org/253/' + str(GTIN(raw=rawGDTI))) + + if match( + r'^urn:epc:idpat:sgcn:(([\d]{6}\.[\d]{6})|([\d]{7}\.[\d]{5})|([\d]{8}\.[\d]{4})|([\d]{9}\.[\d]{3})|([\d]{10}\.[\d]{2})|([\d]{11}\.[\d]{1})|([\d]{12}\.\.))\.\*$', + uri) is not None: + gs1companyprefix = uri[19:partition] + couponref = uri[(partition + 1):(partition + 1 + (12 - len(gs1companyprefix)))] + rawSGCN = gs1companyprefix + couponref + return ('https://id.gs1.org/255/' + str(GTIN(raw=rawSGCN))) + + if match( + r'^urn:epc:idpat:cpi:((\d{6}\.(\%2[3dfDF]|\%3[0-9]|\%4[1-9A-Fa-f]|\%5[0-9Aa]|[0-9A-Z-]){1,24})|(\d{7}\.(\%2[3dfDF]|\%3[0-9]|\%4[1-9A-Fa-f]|\%5[0-9Aa]|[0-9A-Z-]){1,23})|(\d{8}\.(\%2[3dfDF]|\%3[0-9]|\%4[1-9A-Fa-f]|\%5[0-9Aa]|[0-9A-Z-]){1,22})|(\d{9}\.(\%2[3dfDF]|\%3[0-9]|\%4[1-9A-Fa-f]|\%5[0-9Aa]|[0-9A-Z-]){1,21})|(\d{10}\.(\%2[3dfDF]|\%3[0-9]|\%4[1-9A-Fa-f]|\%5[0-9Aa]|[0-9A-Z-]){1,20})|(\d{11}\.(\%2[3dfDF]|\%3[0-9]|\%4[1-9A-Fa-f]|\%5[0-9Aa]|[0-9A-Z-]){1,19})|(\d{12}\.(\%2[3dfDF]|\%3[0-9]|\%4[1-9A-Fa-f]|\%5[0-9Aa]|[0-9A-Z-]){1,18}))\.\*$', + uri) is not None: + gs1companyprefix = uri[18:partition] + separator = uri.rfind('.') + cpref = uri[(partition + 1):(separator)] + rawCPI = gs1companyprefix + cpref + return ('https://id.gs1.org/8010/' + __web_uri_percent_encoder(rawCPI)) + + if match( + r'^urn:epc:idpat:itip:(([\d]{6}\.[\d]{7})|([\d]{7}\.[\d]{6})|([\d]{8}\.[\d]{5})|([\d]{9}\.[\d]{4})|([\d]{10}\.[\d]{3})|([\d]{11}\.[\d]{2})|([\d]{12}\.[\d]{1}))\.[\d]{2}\.[\d]{2}\.\*$', + uri) is not None: + gs1companyprefix = uri[19:partition] + itemref = uri[(partition + 1):(partition + 1 + (13 - len(gs1companyprefix)))] + rawGTIN = itemref[0:1] + gs1companyprefix + itemref[1:] + piece = uri[34:36] + total = uri[37:39] + return ('https://id.gs1.org/8006/' + str(GTIN(raw=rawGTIN)) + piece + total) + + if match( + r'^urn:epc:idpat:upui:((\d{6}\.\d{7})|(\d{7}\.\d{6})|(\d{8}\.\d{5})|(\d{9}\.\d{4})|(\d{10}\.\d{3})|(\d{11}\.\d{2})|(\d{12}\.\d{1}))\.\*$', + uri) is not None: + gs1companyprefix = uri[19:partition] + itemref = uri[(partition + 1):(partition + 1 + (13 - len(gs1companyprefix)))] + rawGTIN = itemref[0:1] + gs1companyprefix + itemref[1:] + return ('https://id.gs1.org/01/' + str(GTIN(raw=rawGTIN))) + + # GS1 DL URIs + if match( + r'^https?:(\/\/((([^\/?#]*)@)?([^\/?#:]*)(:([^\/?#]*))?))?((([^?#]*)(\/(01|gtin|8006|itip|8010|cpid|414|gln|417|party|8017|gsrnp|8018|gsrn|255|gcn|00|sscc|253|gdti|401|ginc|402|gsin|8003|grai|8004|giai)\/)(\d{4}[^\/]+)(\/[^/]+\/[^/]+)?[/]?(\?([^?\n]*))?(#([^\n]*))?)|(\/[A-Za-z_-]{10}$))', + uri) is None: + return None + + # remove query string + if uri.find('?') >= 0: + uri = uri[:uri.index('?')] + + # replace short names for keys/key extensions with AIs + uri = (uri.replace('/gtin/', '/01/') + .replace('/itip/', '/8006/') + .replace('/cpid/', '/8010/') + .replace('/gln/', '/414/') + .replace('/party/', '/417/') + .replace('/gsrnp/', '/8017/') + .replace('/gsrn/', '/8018/') + .replace('/gcn/', '/255/') + .replace('/sscc/', '/00/') + .replace('/gdti/', '/253/') + .replace('/ginc/', '/401/') + .replace('/gsin/', '/402/') + .replace('/grai/', '/8003/') + .replace('/giai/', '/8004/') + .replace('/cpv/', '/22/') + .replace('/lot/', '/10/') + .replace('/ser/', '/21/')) + + # prefix with canonical domain name + if match( + r'^https:\/\/id.gs1.org\/(01|8006|8010|414|417|8017|8018|255|00|253|401|402|8003|8004)\/(\d{4}[^\/]+)(\/[^\/]+\/[^\/]+)?[\/]?(\?([^?\n]*))?(#([^\n]*))?|(\/[A-Za-z_-]{10}$)', + uri) is None: + if uri.find('/00/') != -1: + uri = 'https://id.gs1.org' + uri[(uri.find('/00/')):] + elif uri.find('/01/') != -1: + uri = 'https://id.gs1.org' + uri[(uri.find('/01/')):] + elif uri.find('/253/') != -1: + uri = 'https://id.gs1.org' + uri[(uri.find('/253/')):] + elif uri.find('/255/') != -1: + uri = 'https://id.gs1.org' + uri[(uri.find('/255/')):] + elif uri.find('/401/') != -1: + uri = 'https://id.gs1.org' + uri[(uri.find('/401/')):] + elif uri.find('/402/') != -1: + uri = 'https://id.gs1.org' + uri[(uri.find('/402/')):] + elif uri.find('/414/') != -1: + uri = 'https://id.gs1.org' + uri[(uri.find('/414/')):] + elif uri.find('/417/') != -1: + uri = 'https://id.gs1.org' + uri[(uri.find('/417/')):] + elif uri.find('/8003/') != -1: + uri = 'https://id.gs1.org' + uri[(uri.find('/8003/')):] + elif uri.find('/8004/') != -1: + uri = 'https://id.gs1.org' + uri[(uri.find('/8004/')):] + elif uri.find('/8006/') != -1: + uri = 'https://id.gs1.org' + uri[(uri.find('/8006/')):] + elif uri.find('/8010/') != -1: + uri = 'https://id.gs1.org' + uri[(uri.find('/8010/')):] + elif uri.find('/8017/') != -1: + uri = 'https://id.gs1.org' + uri[(uri.find('/8017/')):] + elif uri.find('/8018/') != -1: + uri = 'https://id.gs1.org' + uri[(uri.find('/8018/')):] + + # ensure that all GTIN formats are padded to 14 digits + if match(r'^https:\/\/id.gs1.org\/01\/\d{14}', uri) is None: + if match(r'^https:\/\/id.gs1.org\/01\/\d{13}', uri) is not None: + uri = uri.replace('/01/', '/01/0') + elif match(r'^https:\/\/id.gs1.org\/01\/\d{12}', uri) is not None: + uri = uri.replace('/01/', '/01/00') + elif match(r'^https:\/\/id.gs1.org\/01\/\d{8}', uri) is not None: + uri = uri.replace('/01/', '/01/000000') + + # remove cpv + x = uri[(uri.find('/22/') + 4):] + + # for 01/8006 only: + if ( + match( + r'https:\/\/id.gs1.org\/8006\/\d{18}\/22\/([\x22\x27\x2D\x2E\x30-\x39\x3B-\x3F\x41-\x5A\x5F\x61-\x7A]{0,20})$', + uri) or match( + r'https:\/\/id.gs1.org\/01\/\d{14}\/22\/([\x22\x27\x2D\x2E\x30-\x39\x3B-\x3F\x41-\x5A\x5F\x61-\x7A]{0,20})$', + uri)) is not None: + uri = (uri[:(uri.find('/22/'))]) + (x[x.find('/'):-1]) + + # for 01/8006 followed by other key qualifiers: + if ( + match( + r'https:\/\/id.gs1.org\/8006\/\d{18}\/22\/([\x2F\x22\x27\x2D\x2E\x30-\x39\x3B-\x3F\x41-\x5A\x5F\x61-\x7A]{0,20})$', + uri) or match( + r'https:\/\/id.gs1.org\/01\/\d{14}\/22\/([\x2F\x22\x27\x2D\x2E\x30-\x39\x3B-\x3F\x41-\x5A\x5F\x61-\x7A]{0,20})$', + uri)) is not None: + uri = (uri[:(uri.find('/22/'))]) + (x[x.find('/'):]) + + # take only lowest ID granularity level (i.e. if serial is present, omit lot) + if (match( + r'https:\/\/id.gs1.org\/8006\/\d{18}\/10\/([\x22\x27\x2D\x2E\x30-\x39\x3B-\x3F\x41-\x5A\x5F\x61-\x7A]{0,20})\/21\/([\x22\x27\x2D\x2E\x30-\x39\x3B-\x3F\x41-\x5A\x5F\x61-\x7A]{0,20})$', + uri) or match( + r'https:\/\/id.gs1.org\/01\/(\d{14})\/10\/([\x22\x27\x2D\x2E\x30-\x39\x3B-\x3F\x41-\x5A\x5F\x61-\x7A]{0,20})\/21\/([\x22\x27\x2D\x2E\x30-\x39\x3B-\x3F\x41-\x5A\x5F\x61-\x7A]{0,20})$', + uri)) is not None: + y = uri[(uri.find('/10/') + 4):] + uri = (uri[:(uri.find('/10/'))]) + (y[y.find('/'):]) + + # ensure that output has a valid syntax + if (match(r'https:\/\/id.gs1.org\/00\/(\d{18})$', uri) or + match( + r'https:\/\/id.gs1.org\/01\/(\d{14})\/21\/([\x22\x27\x2D\x2E\x30-\x39\x3B-\x3F\x41-\x5A\x5F\x61-\x7A]{0,20})$', + uri) or + match( + r'https:\/\/id.gs1.org\/01\/(\d{14})\/10\/([\x22\x27\x2D\x2E\x30-\x39\x3B-\x3F\x41-\x5A\x5F\x61-\x7A]{0,20})$', + uri) or + match(r'https:\/\/id.gs1.org\/01\/(\d{14})$', uri) or + match( + r'https:\/\/id.gs1.org\/01\/(\d{14})\/235\/([\x22\x27\x2D\x2E\x30-\x39\x3B-\x3F\x41-\x5A\x5F\x61-\x7A]{0,28})$', + uri) or + match(r'https:\/\/id.gs1.org\/253\/(\d{13})([\x22\x27\x2D\x2E\x30-\x39\x3B-\x3F\x41-\x5A\x5F\x61-\x7A]{0,17})$', + uri) or + match(r'https:\/\/id.gs1.org\/255\/(\d{13})(\d{0,12})$', uri) or + match(r'https:\/\/id.gs1.org\/401\/([\x22\x27\x2D\x2E\x30-\x39\x3B-\x3F\x41-\x5A\x5F\x61-\x7A]{0,30})$', uri) or + match(r'https:\/\/id.gs1.org\/402\/(\d{17})$', uri) or + match(r'https:\/\/id.gs1.org\/414\/(\d{13})$', uri) or + match( + r'https:\/\/id.gs1.org\/414\/(\d{13})\/254\/([\x22\x27\x2D\x2E\x30-\x39\x3B-\x3F\x41-\x5A\x5F\x61-\x7A]{0,20})$', + uri) or + match(r'https:\/\/id.gs1.org\/417\/(\d{13})$', uri) or + match( + r'https:\/\/id.gs1.org\/8003\/(\d{14})([\x22\x27\x2D\x2E\x30-\x39\x3B-\x3F\x41-\x5A\x5F\x61-\x7A]{0,16})$', + uri) or + match(r'https:\/\/id.gs1.org\/8004\/([\x22\x27\x2D\x2E\x30-\x39\x3B-\x3F\x41-\x5A\x5F\x61-\x7A]{0,30})$', + uri) or + match( + r'https:\/\/id.gs1.org\/8006\/(\d{18})\/21\/([\x22\x27\x2D\x2E\x30-\x39\x3B-\x3F\x41-\x5A\x5F\x61-\x7A]{0,20})$', + uri) or + match( + r'https:\/\/id.gs1.org\/8006\/(\d{18})\/10\/([\x22\x27\x2D\x2E\x30-\x39\x3B-\x3F\x41-\x5A\x5F\x61-\x7A]{0,20})$', + uri) or + match(r'https:\/\/id.gs1.org\/8006\/(\d{18})$', uri) or + match(r'https:\/\/id.gs1.org\/8010\/([\x23\x2D\x2F\x30-\x39\x41-\x5A]{0,30})\/8011/(\d{0,12})$', uri) or + match(r'https:\/\/id.gs1.org\/8010\/([\x23\x2D\x2F\x30-\x39\x41-\x5A]{0,30})$', uri) or + match(r'https:\/\/id.gs1.org\/8017\/(\d{18})$', uri) or + match(r'https:\/\/id.gs1.org\/8018\/(\d{18})$', uri) + ) is not None: # noqa E124 + return (uri) + return None diff --git a/epcis_event_hash_generator/hash_generator.py b/epcis_event_hash_generator/hash_generator.py index c172135..0516090 100644 --- a/epcis_event_hash_generator/hash_generator.py +++ b/epcis_event_hash_generator/hash_generator.py @@ -29,6 +29,7 @@ except ImportError: from context import epcis_event_hash_generator # noqa: F401 +from epcis_event_hash_generator.dl_normaliser import normaliser as dl_normaliser from epcis_event_hash_generator.xml_to_py import event_list_from_epcis_document_xml as read_xml from epcis_event_hash_generator.xml_to_py import event_list_from_epcis_document_xml_str as read_xml_str from epcis_event_hash_generator.json_to_py import event_list_from_epcis_document_json as read_json @@ -127,6 +128,10 @@ def canonize_value(text): """Run a value through all format canonizations""" text = try_format_web_vocabulary(text) text = try_format_numeric(text) + converted = dl_normaliser(text) + if converted: + logging.debug("Converted %s to %s", text, converted) + return converted return text @@ -163,6 +168,7 @@ def generic_child_list_to_prehash_string(children): for child in children: text = child[1].strip() if text: + text = canonize_value(text) text = "=" + text list_of_values.append(child[0] + text + generic_child_list_to_prehash_string(child[2])) @@ -197,6 +203,7 @@ def compute_prehash_from_file(path, enforce=None): events = read_json(path) else: logging.error("Filename '%s' ending not recognized.", path) + return None return compute_prehash_from_events(events) @@ -287,4 +294,4 @@ def calculate_hash(prehash_string_list, hashalg="sha256"): hashValueList.append(hash_string) - return (hashValueList, prehash_string_list) + return hashValueList, prehash_string_list diff --git a/requirements.txt b/requirements.txt index dd70786..13ca11d 100644 --- a/requirements.txt +++ b/requirements.txt @@ -1,2 +1,3 @@ -python_dateutil==2.8.1 -Flask==1.1.2 +python_dateutil>=2.8 +Flask>=1.1 +gtin>=0.1.13 \ No newline at end of file diff --git a/setup.py b/setup.py index bdd42ec..78472a5 100644 --- a/setup.py +++ b/setup.py @@ -29,6 +29,7 @@ }, install_requires=[ 'python_dateutil>=2.8', - 'Flask>=1.1' + 'Flask>=1.1', + 'gtin>=0.1.13' ], )

Digital Link Syntax for EPCs Ralph wrote a separate Normalizer for this

2021-03-02T10:33:55

0.0

[]

RalphTro/epcis-event-hash-generator

RalphTro__epcis-event-hash-generator-35

621877b5ad49d0260cedf4c19b45fa2a06f17f58

diff --git a/epcis_event_hash_generator/context.py b/epcis_event_hash_generator/context.py index df8b0df..913be06 100644 --- a/epcis_event_hash_generator/context.py +++ b/epcis_event_hash_generator/context.py @@ -1,5 +1,6 @@ -import epcis_event_hash_generator import os import sys sys.path.insert(0, os.path.abspath(os.path.join(os.path.dirname(__file__), '..'))) + +import epcis_event_hash_generator diff --git a/epcis_event_hash_generator/hash_generator.py b/epcis_event_hash_generator/hash_generator.py index 2686da0..7e7084d 100644 --- a/epcis_event_hash_generator/hash_generator.py +++ b/epcis_event_hash_generator/hash_generator.py @@ -69,7 +69,7 @@ def child_to_pre_hash_string(child, sub_child_order): if child[0].lower().find("time") > 0 and child[0].lower().find("offset") < 0: text = fix_time_stamp_format(text) else: - text = format_if_numeric(text) + text = canonize_value(text) if text: text = "=" + text @@ -123,8 +123,24 @@ def recurse_through_children_in_order(child_list, child_order): return pre_hash -def format_if_numeric(text): - """remove leading/trailing zeros, leading "+", etc. from numbers""" +def canonize_value(text): + """Run a value through all format canonizations""" + text = try_format_web_vocabulary(text) + text = try_format_numeric(text) + return text + + +def try_format_web_vocabulary(text): + """Replace old CBV URNs by new web vocabulary equivalents.""" + return text.replace('urn:epcglobal:cbv:bizstep:', 'https://ns.gs1.org/voc/Bizstep-' + ).replace('urn:epcglobal:cbv:disp:', 'https://ns.gs1.org/voc/Disp-' + ).replace('urn:epcglobal:cbv:btt:', 'https://ns.gs1.org/voc/BTT-' + ).replace('urn:epcglobal:cbv:sdt:', 'https://ns.gs1.org/voc/SDT-' + ).replace('urn:epcglobal:cbv:er:', 'https://ns.gs1.org/voc/ER-') + + +def try_format_numeric(text): + """remove leading/trailing zeros, leading "+", etc. from numbers. Non numeric values are left untouched.""" try: numeric = float(text) if int(numeric) == numeric: # remove trailing .0

GS1 Web Voc Canonicalisation of standard voc elements See # 14: If present, any URN-based standard vocabulary value (starting with ‘urn:epcglobal:cbv’) SHALL be expressed in its corresponding GS1 Web Vocabulary URI equivalent (starting with ‘https://ns.gs1.org’). Example: ‘urn:epcglobal:cbv:bizstep:receiving’ --> ‘https://ns.gs1.org/cbv/BizStep-receiving’

The following need to be replaced: ``` return (inputStr.replace('urn:epcglobal:cbv:bizstep:', 'https://gs1.org/voc/Bizstep-').replace('urn:epcglobal:cbv:disp:', 'https://gs1.org/voc/Disp-').replace('urn:epcglobal:cbv:btt:', 'https://gs1.org/voc/BTT-').replace('urn:epcglobal:cbv:sdt:', 'https://gs1.org/voc/SDT-').replace('urn:epcglobal:cbv:er:', 'https://gs1.org/voc/ER-')) ``` where the url needs to be *ns*.gs1....

2020-12-08T10:53:54

0.0

[]

drasyl/drasyl

drasyl__drasyl-611

89fc8c3527425cb4ae8357f7386a3b9ec2cea1e1

diff --git a/drasyl-extras/src/main/java/org/drasyl/handler/connection/ConnectionHandler.java b/drasyl-extras/src/main/java/org/drasyl/handler/connection/ConnectionHandler.java index 4ca783291a..1a6b0f609d 100644 --- a/drasyl-extras/src/main/java/org/drasyl/handler/connection/ConnectionHandler.java +++ b/drasyl-extras/src/main/java/org/drasyl/handler/connection/ConnectionHandler.java @@ -1054,7 +1054,7 @@ private void segmentArrivesOnListenState(final ChannelHandlerContext ctx, } // RFC 9293: Set RCV.NXT to SEG.SEQ+1, IRS is set to SEG.SEQ, - tcb.rcvNxt(advanceSeq(seg.seq(), seg.len())); + tcb.rcvNxt(advanceSeq(seg.seq(), 1)); tcb.irs(seg.seq()); // RFC 9293: and any other control or text should be queued for processing later. @@ -1206,7 +1206,7 @@ private void segmentArrivesOnSynSentState(final ChannelHandlerContext ctx, if (seg.isSyn()) { // RFC 9293: If the SYN bit is on and the security/compartment is acceptable, // RFC 9293: then RCV.NXT is set to SEG.SEQ+1, IRS is set to SEG.SEQ. - tcb.rcvNxt(advanceSeq(seg.seq(), seg.len())); + tcb.rcvNxt(advanceSeq(seg.seq(), 1)); tcb.irs(seg.seq()); if (seg.isAck()) { @@ -2056,7 +2056,8 @@ private boolean establishedStateProcessing(final ChannelHandlerContext ctx, // RFC 5681: given connection (TCP.UNA from [RFC793]) and (e) the advertised window in the // RFC 5681: incoming acknowledgment equals the advertised window in the last incoming acknowledgment. final boolean isRfc5681Duplicate = !tcb.retransmissionQueue().isEmpty() && - seg.len() == 0 && // SYN and FIN have length >0 by definition + seg.len() == 0 && + !seg.isAck() && !seg.isFin() && seg.ack() == tcb.sndUna() && seg.wnd() == tcb.lastAdvertisedWindow(); diff --git a/drasyl-extras/src/main/java/org/drasyl/handler/connection/ReceiveBuffer.java b/drasyl-extras/src/main/java/org/drasyl/handler/connection/ReceiveBuffer.java index 26ada7dbd8..dea3c9874f 100644 --- a/drasyl-extras/src/main/java/org/drasyl/handler/connection/ReceiveBuffer.java +++ b/drasyl-extras/src/main/java/org/drasyl/handler/connection/ReceiveBuffer.java @@ -131,6 +131,9 @@ public void receive(final ChannelHandlerContext ctx, ReferenceCountUtil.touch(seg, "ReceiveBuffer receive " + seg.toString()); final ByteBuf content = seg.content(); if (content.isReadable()) { + // (T/TCP or TCP Fast Open not implemented; SYN/FIN flag might require special attention) + assert !seg.isSyn() && !seg.isFin() : "not supported (yet)"; + if (head == null) { // first SEG to be added to RCV.WND? // SEG is located at the left edge of our RCV.WND? @@ -197,15 +200,15 @@ else if (greaterThan(seg.seq(), tcb.rcvNxt()) && lessThan(seg.seq(), add(tcb.rcv head::len, () -> head.next ); - tcb.advanceRcvNxt(ctx, head.len()); + tcb.rcvNxt(ctx, add(tcb.rcvNxt(), head.len())); addToHeadBuf(ctx, head.content()); head = head.next; size--; assert head == null || lessThanOrEqualTo(tcb.rcvNxt(), head.seq()) : tcb.rcvNxt() + " must be less than or equal to " + head; } } - else if (seg.len() > 0) { - tcb.advanceRcvNxt(ctx, seg.len()); + else { + tcb.rcvNxt(ctx, seg.nxtSeq()); } } diff --git a/drasyl-extras/src/main/java/org/drasyl/handler/connection/Segment.java b/drasyl-extras/src/main/java/org/drasyl/handler/connection/Segment.java index b7a299506c..7bd161db86 100644 --- a/drasyl-extras/src/main/java/org/drasyl/handler/connection/Segment.java +++ b/drasyl-extras/src/main/java/org/drasyl/handler/connection/Segment.java @@ -316,13 +316,17 @@ public Map<SegmentOption, Object> options() { * @return the length (in segments) of this segment */ public int len() { + return content().readableBytes(); + } + + public long nxtSeq() { if (isSyn() || isFin()) { // the SYN and FIN flags each count as one segment (SYN and FIN are never set at the // same time) - return 1 + content().readableBytes(); + return add(seq(), 1 + len()); } else { - return content().readableBytes(); + return add(seq(), len()); } } diff --git a/drasyl-extras/src/main/java/org/drasyl/handler/connection/TransmissionControlBlock.java b/drasyl-extras/src/main/java/org/drasyl/handler/connection/TransmissionControlBlock.java index 3c3e22c6eb..334e149d73 100644 --- a/drasyl-extras/src/main/java/org/drasyl/handler/connection/TransmissionControlBlock.java +++ b/drasyl-extras/src/main/java/org/drasyl/handler/connection/TransmissionControlBlock.java @@ -38,7 +38,6 @@ import static org.drasyl.handler.connection.Segment.MIN_SEQ_NO; import static org.drasyl.handler.connection.Segment.SEG_HDR_SIZE; import static org.drasyl.handler.connection.Segment.add; -import static org.drasyl.handler.connection.Segment.advanceSeq; import static org.drasyl.handler.connection.Segment.sub; import static org.drasyl.util.NumberUtil.max; import static org.drasyl.util.NumberUtil.min; @@ -453,8 +452,8 @@ public ReceiveBuffer receiveBuffer() { * Advances SND.NXT and places the {@code seg} on the outgoing segment queue. */ void send(final ChannelHandlerContext ctx, final Segment seg, final ChannelPromise promise) { - if (sndNxt == seg.seq() && seg.len() > 0) { - final long newSndNxt = add(seg.lastSeq(), 1); + if (sndNxt == seg.seq()) { + final long newSndNxt = seg.nxtSeq(); if (LOG.isTraceEnabled()) { LOG.trace("{} Send data [{},{}]. Advance SND.NXT from {} to {} (+{}).", ctx.channel(), seg.seq(), seg.lastSeq(), sndNxt, newSndNxt, Segment.sub(newSndNxt, sndNxt)); } @@ -692,9 +691,8 @@ public long ssthresh() { return ssthresh; } - public void advanceRcvNxt(final ChannelHandlerContext ctx, final int advancement) { - final long newRcvNxt = advanceSeq(rcvNxt, advancement); - if (LOG.isTraceEnabled()) { + public void rcvNxt(final ChannelHandlerContext ctx, final long newRcvNxt) { + if (LOG.isTraceEnabled() && newRcvNxt != rcvNxt) { LOG.trace("{} Advance RCV.NXT from {} to {} (+{}).", ctx.channel(), rcvNxt, newRcvNxt, Segment.sub(newRcvNxt, rcvNxt)); } rcvNxt = newRcvNxt;

Segment#len should not return +1 if SYN or FIN is set

2024-03-16T12:33:22

0.0

[]

apriha/snps

apriha__snps-162

7d2abc2423e12dc97c770205079cf0fee689311e

diff --git a/.gitignore b/.gitignore index 4d9ef8d..038d98c 100644 --- a/.gitignore +++ b/.gitignore @@ -109,7 +109,7 @@ output*/ resources/* # snps tests -tests/output/ +tests/output/* tests/resources/* !tests/resources/gsa_chrpos_map.txt !tests/resources/gsa_rsid_map.txt @@ -119,3 +119,7 @@ tests/input/discrepant_snps[12].csv tests/input/ftdna.csv.gz tests/input/generic.fa.gz tests/input/testvcf.vcf.gz +!tests/output/ +!tests/output/vcf_qc/ +!tests/output/vcf_chrom_prefix_chr.vcf +!tests/output/vcf_generic.vcf diff --git a/README.rst b/README.rst index 0e219cf..9a4f214 100644 --- a/README.rst +++ b/README.rst @@ -29,6 +29,7 @@ Build / Assembly Detection and Remapping Data Cleaning ````````````` +- Perform quality control (QC) / filter low quality SNPs based on `chip clusters <https://doi.org/10.1016/j.csbj.2021.06.040>`_ - Fix several common issues when loading SNPs - Sort SNPs based on chromosome and position - Deduplicate RSIDs @@ -39,6 +40,7 @@ Data Cleaning Analysis ```````` - Derive sex from SNPs +- Detect deduced genotype / chip array and chip version based on `chip clusters <https://doi.org/10.1016/j.csbj.2021.06.040>`_ - Predict ancestry from SNPs (when installed with `ezancestry <https://github.com/arvkevi/ezancestry>`_) Supported Genotype Files @@ -213,14 +215,14 @@ Ok, so far we've merged the SNPs from two files (ensuring the same build in the identifying discrepancies along the way). Then, we remapped the SNPs to Build 38. Now, let's save the merged and remapped dataset consisting of 1M+ SNPs to a tab-separated values (TSV) file: ->>> saved_snps = s.save("out.txt") +>>> saved_snps = s.to_tsv("out.txt") Saving output/out.txt >>> print(saved_snps) output/out.txt Moreover, let's get the reference sequences for this assembly and save the SNPs as a VCF file: ->>> saved_snps = s.save("out.vcf", vcf=True) +>>> saved_snps = s.to_vcf("out.vcf") Downloading resources/fasta/GRCh38/Homo_sapiens.GRCh38.dna.chromosome.1.fa.gz Downloading resources/fasta/GRCh38/Homo_sapiens.GRCh38.dna.chromosome.2.fa.gz Downloading resources/fasta/GRCh38/Homo_sapiens.GRCh38.dna.chromosome.3.fa.gz diff --git a/docs/snps.rst b/docs/snps.rst index 28ba6b1..d9110f2 100644 --- a/docs/snps.rst +++ b/docs/snps.rst @@ -11,7 +11,7 @@ SNPs :exclude-members: sort_snps, remap_snps, save_snps, snp_count, get_snp_count, not_null_snps, get_summary, get_assembly, get_chromosomes, get_chromosomes_summary, duplicate_snps, discrepant_XY_snps, heterozygous_MT_snps, heterozygous_snps, homozygous_snps, - discrepant_positions, discrepant_genotypes, discrepant_snps, is_valid + discrepant_positions, discrepant_genotypes, discrepant_snps, is_valid, save snps.ensembl ------------ @@ -36,6 +36,7 @@ snps.io.reader :members: :undoc-members: :show-inheritance: + :exclude-members: read_file snps.io.writer ~~~~~~~~~~~~~~ @@ -44,6 +45,7 @@ snps.io.writer :members: :undoc-members: :show-inheritance: + :exclude-members: write_file snps.resources -------------- diff --git a/src/snps/io/reader.py b/src/snps/io/reader.py index 3ea8765..9912f8e 100644 --- a/src/snps/io/reader.py +++ b/src/snps/io/reader.py @@ -42,6 +42,7 @@ import logging import os import re +import warnings import zipfile import zlib @@ -211,31 +212,9 @@ def read(self): @classmethod def read_file(cls, file, only_detect_source, resources, rsids): - """Read `file`. - - Parameters - ---------- - file : str or bytes - path to file to load or bytes to load - only_detect_source : bool - only detect the source of the data - resources : Resources - instance of Resources - rsids : tuple - rsids to extract if loading a VCF file - - Returns - ------- - dict - dict with the following items: - - snps (pandas.DataFrame) - dataframe of parsed SNPs - source (str) - detected source of SNPs - phased (bool) - flag indicating if SNPs are phased - """ + warnings.warn( + "This method will be removed in a future release.", DeprecationWarning + ) r = cls(file, only_detect_source, resources, rsids) return r.read() diff --git a/src/snps/io/writer.py b/src/snps/io/writer.py index 324c7b6..34c7938 100644 --- a/src/snps/io/writer.py +++ b/src/snps/io/writer.py @@ -37,11 +37,13 @@ import datetime import logging +import warnings import numpy as np import pandas as pd import snps +from snps.io import get_empty_snps_dataframe from snps.utils import save_df_as_csv, clean_str logger = logging.getLogger(__name__) @@ -57,6 +59,9 @@ def __init__( vcf=False, atomic=True, vcf_alt_unavailable=".", + vcf_chrom_prefix="", + vcf_qc_only=False, + vcf_qc_filter=False, **kwargs, ): """Initialize a `Writer`. @@ -73,6 +78,12 @@ def __init__( atomically write output to a file on local filesystem vcf_alt_unavailable : str representation of VCF ALT allele when ALT is not able to be determined + vcf_chrom_prefix : str + prefix for chromosomes in VCF CHROM column + vcf_qc_only : bool + for VCF, output only SNPs that pass quality control + vcf_qc_filter : bool + for VCF, populate VCF FILTER column based on quality control results **kwargs additional parameters to `pandas.DataFrame.to_csv` """ @@ -81,9 +92,21 @@ def __init__( self._vcf = vcf self._atomic = atomic self._vcf_alt_unavailable = vcf_alt_unavailable + self._vcf_chrom_prefix = vcf_chrom_prefix + self._vcf_qc_only = vcf_qc_only + self._vcf_qc_filter = vcf_qc_filter self._kwargs = kwargs def write(self): + """Write SNPs to file or buffer. + + Returns + ------- + str + path to file in output directory if SNPs were saved, else empty str + discrepant_vcf_position : pd.DataFrame + SNPs with discrepant positions discovered while saving VCF + """ if self._vcf: return self._write_vcf() else: @@ -97,38 +120,21 @@ def write_file( vcf=False, atomic=True, vcf_alt_unavailable=".", + vcf_qc_only=False, + vcf_qc_filter=False, **kwargs, ): - """Save SNPs to file. - - Parameters - ---------- - snps : SNPs - SNPs to save to file or write to buffer - filename : str or buffer - filename for file to save or buffer to write to - vcf : bool - flag to save file as VCF - atomic : bool - atomically write output to a file on local filesystem - vcf_alt_unavailable : str - representation of VCF ALT allele when ALT is not able to be determined - **kwargs - additional parameters to `pandas.DataFrame.to_csv` - - Returns - ------- - str - path to file in output directory if SNPs were saved, else empty str - discrepant_vcf_position : pd.DataFrame - SNPs with discrepant positions discovered while saving VCF - """ + warnings.warn( + "This method will be removed in a future release.", DeprecationWarning + ) w = cls( snps=snps, filename=filename, vcf=vcf, atomic=atomic, vcf_alt_unavailable=vcf_alt_unavailable, + vcf_qc_only=vcf_qc_only, + vcf_qc_filter=vcf_qc_filter, **kwargs, ) return w.write() @@ -257,6 +263,12 @@ def _write_vcf(self): "assembly": self._snps.assembly, "chrom": chrom, "snps": pd.DataFrame(df.loc[(df["chrom"] == chrom)]), + "cluster": self._snps.cluster + if self._vcf_qc_only or self._vcf_qc_filter + else "", + "low_quality_snps": self._snps.low_quality + if self._vcf_qc_only or self._vcf_qc_filter + else get_empty_snps_dataframe(), } ) @@ -279,6 +291,10 @@ def _write_vcf(self): discrepant_vcf_position = pd.concat(discrepant_vcf_position) comment += "".join(contigs) + + if self._vcf_qc_filter and self._snps.cluster: + comment += '##FILTER=<ID=lq,Description="Low quality SNP per Lu et al.: https://doi.org/10.1016/j.csbj.2021.06.040">\n' + comment += '##FORMAT=<ID=GT,Number=1,Type=String,Description="Genotype">\n' comment += "#CHROM\tPOS\tID\tREF\tALT\tQUAL\tFILTER\tINFO\tFORMAT\tSAMPLE\n" @@ -302,6 +318,8 @@ def _create_vcf_representation(self, task): assembly = task["assembly"] chrom = task["chrom"] snps = task["snps"] + cluster = task["cluster"] + low_quality_snps = task["low_quality_snps"] if len(snps.loc[snps["genotype"].notnull()]) == 0: return { @@ -315,6 +333,16 @@ def _create_vcf_representation(self, task): contig = f'##contig=<ID={seq.ID},URL={seq.url},length={seq.length},assembly={seq.build},md5={seq.md5},species="{seq.species}">\n' + if self._vcf_qc_only and cluster: + # drop low quality SNPs if SNPs object maps to a cluster + snps = snps.drop(snps.index.intersection(low_quality_snps.index)) + + if self._vcf_qc_filter and cluster: + # initialize filter for all SNPs if SNPs object maps to a cluster, + snps["filter"] = "PASS" + # then indicate SNPs that were identified as low quality + snps.loc[snps.index.intersection(low_quality_snps.index), "filter"] = "lq" + snps = snps.reset_index() df = pd.DataFrame( @@ -346,10 +374,13 @@ def _create_vcf_representation(self, task): } ) - df["CHROM"] = snps["chrom"] + df["CHROM"] = self._vcf_chrom_prefix + snps["chrom"] df["POS"] = snps["pos"] df["ID"] = snps["rsid"] + if self._vcf_qc_filter and cluster: + df["FILTER"] = snps["filter"] + # drop SNPs with discrepant positions (outside reference sequence) discrepant_vcf_position = snps.loc[ (snps.pos - seq.start < 0) | (snps.pos - seq.start > seq.length - 1) diff --git a/src/snps/resources.py b/src/snps/resources.py index d7ee9a1..0f7c89a 100644 --- a/src/snps/resources.py +++ b/src/snps/resources.py @@ -90,6 +90,8 @@ def _init_resource_attributes(self): self._gsa_chrpos_map = None self._dbsnp_151_37_reverse = None self._opensnp_datadump_filenames = [] + self._chip_clusters = None + self._low_quality_snps = None def get_reference_sequences( self, @@ -235,6 +237,8 @@ def get_all_resources(self): source, target ) resources["gsa_resources"] = self.get_gsa_resources() + resources["chip_clusters"] = self.get_chip_clusters() + resources["low_quality_snps"] = self.get_low_quality_snps() return resources def get_all_reference_sequences(self, **kwargs): @@ -269,6 +273,90 @@ def get_gsa_resources(self): "dbsnp_151_37_reverse": self.get_dbsnp_151_37_reverse(), } + def get_chip_clusters(self): + """Get resource for identifying deduced genotype / chip array based on chip clusters. + + Returns + ------- + pandas.DataFrame + + References + ---------- + 1. Chang Lu, Bastian Greshake Tzovaras, Julian Gough, A survey of + direct-to-consumer genotype data, and quality control tool + (GenomePrep) for research, Computational and Structural + Biotechnology Journal, Volume 19, 2021, Pages 3747-3754, ISSN + 2001-0370, https://doi.org/10.1016/j.csbj.2021.06.040. + """ + if self._chip_clusters is None: + chip_clusters_path = self._download_file( + "https://supfam.mrc-lmb.cam.ac.uk/GenomePrep/datadir/the_list.tsv.gz", + "chip_clusters.tsv.gz", + ) + + df = pd.read_csv( + chip_clusters_path, + sep="\t", + names=["locus", "clusters"], + dtype={"locus": object, "clusters": pd.CategoricalDtype(ordered=False)}, + ) + clusters = df.clusters + df = df.locus.str.split(":", expand=True) + df.rename({0: "chrom", 1: "pos"}, axis=1, inplace=True) + df.pos = df.pos.astype(np.uint32) + df.chrom = df.chrom.astype(pd.CategoricalDtype(ordered=False)) + df["clusters"] = clusters + + self._chip_clusters = df + + return self._chip_clusters + + def get_low_quality_snps(self): + """Get listing of low quality SNPs for quality control based on chip clusters. + + Returns + ------- + pandas.DataFrame + + References + ---------- + 1. Chang Lu, Bastian Greshake Tzovaras, Julian Gough, A survey of + direct-to-consumer genotype data, and quality control tool + (GenomePrep) for research, Computational and Structural + Biotechnology Journal, Volume 19, 2021, Pages 3747-3754, ISSN + 2001-0370, https://doi.org/10.1016/j.csbj.2021.06.040. + """ + if self._low_quality_snps is None: + low_quality_snps_path = self._download_file( + "https://supfam.mrc-lmb.cam.ac.uk/GenomePrep/datadir/badalleles.tsv.gz", + "low_quality_snps.tsv.gz", + ) + + df = pd.read_csv( + low_quality_snps_path, + sep="\t", + names=["cluster", "loci"], + ) + + cluster_dfs = [] + for row in df.itertuples(): + loci = row.loci.split(",") + cluster_dfs.append( + pd.DataFrame({"cluster": [row.cluster] * len(loci), "locus": loci}) + ) + + df = pd.concat(cluster_dfs) + df.reset_index(inplace=True, drop=True) + cluster = df.cluster.astype(pd.CategoricalDtype(ordered=False)) + df = df.locus.str.split(":", expand=True) + df.rename({0: "chrom", 1: "pos"}, axis=1, inplace=True) + df.pos = df.pos.astype(np.uint32) + df.chrom = df.chrom.astype(pd.CategoricalDtype(ordered=False)) + df["cluster"] = cluster + self._low_quality_snps = df + + return self._low_quality_snps + def get_dbsnp_151_37_reverse(self): """Get and load RSIDs that are on the reference reverse (-) strand in dbSNP 151 and lower. diff --git a/src/snps/snps.py b/src/snps/snps.py index aa4250f..89e8008 100644 --- a/src/snps/snps.py +++ b/src/snps/snps.py @@ -35,6 +35,7 @@ """ +import copy from itertools import groupby, count import logging import os @@ -103,6 +104,7 @@ def __init__( self._discrepant_XY = get_empty_snps_dataframe() self._heterozygous_MT = get_empty_snps_dataframe() self._discrepant_vcf_position = get_empty_snps_dataframe() + self._low_quality = get_empty_snps_dataframe().index self._discrepant_merge_positions = pd.DataFrame() self._discrepant_merge_genotypes = pd.DataFrame() self._source = [] @@ -112,6 +114,9 @@ def __init__( self._output_dir = output_dir self._resources = Resources(resources_dir=resources_dir) self._parallelizer = Parallelizer(parallelize=parallelize, processes=processes) + self._cluster = "" + self._chip = "" + self._chip_version = "" if file: @@ -218,6 +223,30 @@ def snps(self): """ return self._snps + @property + def snps_qc(self): + """Normalized SNPs, after quality control. + + Any low quality SNPs, identified per + :meth:`identify_low_quality_snps() <snps.snps.SNPs.identify_low_quality_snps>`, + are not included in the result. + + Returns + ------- + pandas.DataFrame + normalized ``snps`` dataframe + """ + if len(self._low_quality) == 0: + # ensure low quality SNPs, if any, are identified + self.identify_low_quality_snps() + + if len(self._low_quality) > 0: + # filter out low quality SNPs + return self._snps.drop(self._low_quality) + else: + # no low quality SNPs to filter + return self._snps + @property def duplicate(self): """Duplicate SNPs. @@ -268,6 +297,22 @@ def discrepant_vcf_position(self): """ return self._discrepant_vcf_position + @property + def low_quality(self): + """SNPs identified as low quality, if any, per + :meth:`identify_low_quality_snps() <snps.snps.SNPs.identify_low_quality_snps>`. + + Returns + ------- + pandas.DataFrame + normalized ``snps`` dataframe + """ + if len(self._low_quality) == 0: + # ensure low quality SNPs, if any, are identified + self.identify_low_quality_snps() + + return self._snps.loc[self._low_quality] + @property def discrepant_merge_positions(self): """SNPs with discrepant positions discovered while merging SNPs. @@ -487,6 +532,59 @@ def phased(self): """ return self._phased + @property + def cluster(self): + """Detected chip cluster, if any, per + :meth:`compute_cluster_overlap <snps.snps.SNPs.compute_cluster_overlap>`. + + Notes + ----- + Refer to :meth:`compute_cluster_overlap <snps.snps.SNPs.compute_cluster_overlap>` + for more details about chip clusters. + + Returns + ------- + str + detected chip cluster, e.g., 'c1', else empty str + """ + if not self._cluster: + self.compute_cluster_overlap() + return self._cluster + + @property + def chip(self): + """Detected deduced genotype / chip array, if any, per + :meth:`compute_cluster_overlap <snps.snps.SNPs.compute_cluster_overlap>`. + + Returns + ------- + str + detected chip array, else empty str + """ + if not self._chip: + self.compute_cluster_overlap() + return self._chip + + @property + def chip_version(self): + """Detected genotype / chip array version, if any, per + :meth:`compute_cluster_overlap <snps.snps.SNPs.compute_cluster_overlap>`. + + Notes + ----- + Chip array version is only applicable to 23andMe (v3, v4, v5) and AncestryDNA + (v1, v2) files. + + Returns + ------- + str + detected chip array version, e.g., 'v4', else empty str + """ + + if not self._chip_version: + self.compute_cluster_overlap() + return self._chip_version + def heterozygous(self, chrom=""): """Get heterozygous SNPs. @@ -585,27 +683,41 @@ def valid(self): return True def save( - self, filename="", vcf=False, atomic=True, vcf_alt_unavailable=".", **kwargs + self, + filename="", + vcf=False, + atomic=True, + vcf_alt_unavailable=".", + vcf_qc_only=False, + vcf_qc_filter=False, + **kwargs, ): - """Save SNPs to file. - - Parameters - ---------- - filename : str or buffer - filename for file to save or buffer to write to - vcf : bool - flag to save file as VCF - atomic : bool - atomically write output to a file on local filesystem - vcf_alt_unavailable : str - representation of VCF ALT allele when ALT is not able to be determined - **kwargs - additional parameters to `pandas.DataFrame.to_csv` + warnings.warn( + "Method `save` has been replaced by `to_csv`, `to_tsv`, and `to_vcf`.", + DeprecationWarning, + ) + return self._save( + filename, + vcf, + atomic, + vcf_alt_unavailable, + vcf_qc_only, + vcf_qc_filter, + **kwargs, + ) - Returns - ------- - str - path to file in output directory if SNPs were saved, else empty str + def _save( + self, + filename="", + vcf=False, + atomic=True, + vcf_alt_unavailable=".", + vcf_chrom_prefix="", + vcf_qc_only=False, + vcf_qc_filter=False, + **kwargs, + ): + """Save SNPs to file. References ---------- @@ -615,14 +727,18 @@ def save( if "sep" not in kwargs: kwargs["sep"] = "\t" - path, *extra = Writer.write_file( + w = Writer( snps=self, filename=filename, vcf=vcf, atomic=atomic, vcf_alt_unavailable=vcf_alt_unavailable, + vcf_chrom_prefix=vcf_chrom_prefix, + vcf_qc_only=vcf_qc_only, + vcf_qc_filter=vcf_qc_filter, **kwargs, ) + path, *extra = w.write() if len(extra) == 1 and not extra[0].empty: self._discrepant_vcf_position = extra[0] @@ -633,11 +749,112 @@ def save( return path + def to_csv(self, filename="", atomic=True, **kwargs): + """Output SNPs as comma-separated values. + + Parameters + ---------- + filename : str or buffer + filename for file to save or buffer to write to + atomic : bool + atomically write output to a file on local filesystem + **kwargs + additional parameters to `pandas.DataFrame.to_csv` + + Returns + ------- + str + path to file in output directory if SNPs were saved, else empty str + """ + kwargs["sep"] = "," + return self._save(filename=filename, atomic=atomic, **kwargs) + + def to_tsv(self, filename="", atomic=True, **kwargs): + """Output SNPs as tab-separated values. + + Note that this results in the same default output as `save`. + + Parameters + ---------- + filename : str or buffer + filename for file to save or buffer to write to + atomic : bool + atomically write output to a file on local filesystem + **kwargs + additional parameters to `pandas.DataFrame.to_csv` + + Returns + ------- + str + path to file in output directory if SNPs were saved, else empty str + """ + kwargs["sep"] = "\t" + return self._save(filename=filename, atomic=atomic, **kwargs) + + def to_vcf( + self, + filename="", + atomic=True, + alt_unavailable=".", + chrom_prefix="", + qc_only=False, + qc_filter=False, + **kwargs, + ): + """Output SNPs as Variant Call Format. + + Parameters + ---------- + filename : str or buffer + filename for file to save or buffer to write to + atomic : bool + atomically write output to a file on local filesystem + alt_unavailable : str + representation of ALT allele when ALT is not able to be determined + chrom_prefix : str + prefix for chromosomes in VCF CHROM column + qc_only : bool + output only SNPs that pass quality control + qc_filter : bool + populate FILTER column based on quality control results + **kwargs + additional parameters to `pandas.DataFrame.to_csv` + + Returns + ------- + str + path to file in output directory if SNPs were saved, else empty str + + Notes + ----- + Parameters `qc_only` and `qc_filter`, if true, will identify low quality SNPs per + :meth:`identify_low_quality_snps() <snps.snps.SNPs.identify_low_quality_snps>`, + if not done already. Moreover, these parameters have no effect if this SNPs + object does not map to a cluster per + :meth:`compute_cluster_overlap() <snps.snps.SNPs.compute_cluster_overlap>`. + + References + ---------- + 1. The Variant Call Format (VCF) Version 4.2 Specification, 8 Mar 2019, + https://samtools.github.io/hts-specs/VCFv4.2.pdf + """ + return self._save( + filename=filename, + vcf=True, + atomic=atomic, + vcf_alt_unavailable=alt_unavailable, + vcf_chrom_prefix=chrom_prefix, + vcf_qc_only=qc_only, + vcf_qc_filter=qc_filter, + **kwargs, + ) + def _filter(self, chrom=""): return self.snps.loc[self.snps.chrom == chrom] if chrom else self.snps def _read_raw_data(self, file, only_detect_source, rsids): - return Reader.read_file(file, only_detect_source, self._resources, rsids) + r = Reader(file, only_detect_source, self._resources, rsids) + return r.read() def _assign_par_snps(self): """Assign PAR SNPs to the X or Y chromosome using SNP position. @@ -1523,8 +1740,11 @@ def remap_snps(self, target_assembly, complement_bases=True): return self.remap(target_assembly, complement_bases) def save_snps(self, filename="", vcf=False, atomic=True, **kwargs): - warnings.warn("This method has been renamed to `save`.", DeprecationWarning) - return self.save(filename, vcf, atomic, **kwargs) + warnings.warn( + "Method `save_snps` has been replaced by `to_csv`, `to_tsv`, and `to_vcf`.", + DeprecationWarning, + ) + return self._save(filename, vcf, atomic, **kwargs) @property def snp_count(self): @@ -1730,3 +1950,149 @@ def max_pop(row): d["ezancestry_df"] = predictions return d + + def compute_cluster_overlap(self, cluster_overlap_threshold=0.95): + """Compute overlap with chip clusters. + + Chip clusters, which are defined in [1]_, are associated with deduced genotype / + chip arrays and DTC companies. + + This method also sets the values returned by the `cluster`, `chip`, and + `chip_version` properties, based on max overlap, if the specified threshold is + satisfied. + + Parameters + ---------- + cluster_overlap_threshold : float + threshold for cluster to overlap this SNPs object, and vice versa, to set + values returned by the `cluster`, `chip`, and `chip_version` properties + + Returns + ------- + pandas.DataFrame + pandas.DataFrame with the following columns: + + `company_composition` + DTC company composition of associated cluster from [1]_ + `chip_base_deduced` + deduced genotype / chip array of associated cluster from [1]_ + `snps_in_cluster` + count of SNPs in cluster + `snps_in_common` + count of SNPs in common with cluster (inner merge with cluster) + `overlap_with_cluster` + percentage overlap of `snps_in_common` with cluster + `overlap_with_self` + percentage overlap of `snps_in_common` with this SNPs object + + References + ---------- + .. [1] Chang Lu, Bastian Greshake Tzovaras, Julian Gough, A survey of + direct-to-consumer genotype data, and quality control tool + (GenomePrep) for research, Computational and Structural + Biotechnology Journal, Volume 19, 2021, Pages 3747-3754, ISSN + 2001-0370, https://doi.org/10.1016/j.csbj.2021.06.040. + """ + + # information from Lu et. al (Ref. [1]_), Table 2 and Fig. 2 + df = pd.DataFrame( + data={ + "cluster_id": ["c1", "c3", "c4", "c5", "v5"], + "company_composition": [ + "23andMe-v4", + "AncestryDNA-v1, FTDNA, MyHeritage", + "23andMe-v3", + "AncestryDNA-v2", + "23andMe-v5, LivingDNA", + ], + "chip_base_deduced": [ + "HTS iSelect HD", + "OmniExpress", + "OmniExpress plus", + "OmniExpress plus", + "Illumina GSAs", + ], + "snps_in_cluster": [0] * 5, + "snps_in_common": [0] * 5, + } + ) + df.set_index("cluster_id", inplace=True) + + if self.build != 37: + to_remap = copy.deepcopy(self) + to_remap.remap(37) # clusters are relative to Build 37 + self_snps = to_remap.snps[["chrom", "pos"]].drop_duplicates() + else: + self_snps = self.snps[["chrom", "pos"]].drop_duplicates() + + chip_clusters = self._resources.get_chip_clusters() + + for cluster in df.index.values: + cluster_snps = chip_clusters.loc[ + chip_clusters.clusters.str.contains(cluster) + ][["chrom", "pos"]] + df.loc[cluster, "snps_in_cluster"] = len(cluster_snps) + df.loc[cluster, "snps_in_common"] = len( + self_snps.merge(cluster_snps, how="inner") + ) + + df["overlap_with_cluster"] = df.snps_in_common / df.snps_in_cluster + df["overlap_with_self"] = df.snps_in_common / len(self_snps) + + max_overlap = df.overlap_with_cluster.idxmax() + + if ( + df.overlap_with_cluster.loc[max_overlap] > cluster_overlap_threshold + and df.overlap_with_self.loc[max_overlap] > cluster_overlap_threshold + ): + self._cluster = max_overlap + self._chip = df.chip_base_deduced.loc[max_overlap] + + company_composition = df.company_composition.loc[max_overlap] + + if self.source in company_composition: + if self.source == "23andMe" or self.source == "AncestryDNA": + i = company_composition.find("v") + self._chip_version = company_composition[i : i + 2] + else: + logger.warning( + "Detected SNPs data source not found in cluster's company composition" + ) + + return df + + def identify_low_quality_snps(self): + """Identify low quality SNPs based on chip clusters. + + Any low quality SNPs are removed from the + :meth:`snps_qc <snps.snps.SNPs.snps_qc>` dataframe and are made + available as :meth:`low_quality <snps.snps.SNPs.low_quality>`. + + Notes + ----- + Chip clusters, which are defined in [1]_, are associated with low quality SNPs. + As such, low quality SNPs will only be identified when this SNPs object corresponds + to a cluster per + :meth:`compute_cluster_overlap() <snps.snps.SNPs.compute_cluster_overlap>`. + """ + if self.build != 37: + to_remap = copy.deepcopy(self) + to_remap.remap(37) # clusters are relative to Build 37 + self_snps = to_remap._snps[["chrom", "pos"]] + else: + self_snps = self._snps[["chrom", "pos"]] + + low_quality_snps = self._resources.get_low_quality_snps() + + if self.cluster: + cluster_snps = low_quality_snps.loc[ + low_quality_snps.cluster.str.contains(self.cluster) + ][["chrom", "pos"]] + # keep index after merge; https://stackoverflow.com/a/11982843 + merged = ( + self_snps.reset_index() + .merge(cluster_snps, how="inner") + .set_index("rsid") + ) + + self._low_quality = merged.index diff --git a/src/snps/utils.py b/src/snps/utils.py index cd36801..5c50bb1 100644 --- a/src/snps/utils.py +++ b/src/snps/utils.py @@ -43,6 +43,7 @@ import os import re import shutil +import tempfile import zipfile from atomicwrites import atomic_write @@ -182,14 +183,19 @@ def save_df_as_csv( df.to_csv(destination, **kwargs) destination.seek(0) elif atomic: - with atomic_write(destination, mode="w", overwrite=True) as f: + fd, tmp_path = tempfile.mkstemp(dir=path) + + with open(fd, mode="w") as f: f.write(s) - # https://stackoverflow.com/a/29233924 - df.to_csv(f, **kwargs) + + # https://stackoverflow.com/a/29233924 + df.to_csv(tmp_path, mode="a", **kwargs) + + os.rename(tmp_path, destination) else: with open(destination, mode="w") as f: f.write(s) - df.to_csv(f, **kwargs) + df.to_csv(destination, mode="a", **kwargs) return destination else:

Truncated position values in output files With `pandas` 1.0.0, using `na_rep` with `to_csv` causes position values in output files to be truncated (e.g., files created via `save_df_as_csv`). See https://github.com/pandas-dev/pandas/issues/31447 In the meantime, a workaround would be to pin the version of `pandas` to <=0.25.3.

2022-11-04T03:53:14

0.0

[]

apriha/snps

apriha__snps-148

b3ebe4c5985c881314495d8040c28faafd9bb6d9

diff --git a/.github/workflows/ci.yml b/.github/workflows/ci.yml index 203553e..93b2819 100644 --- a/.github/workflows/ci.yml +++ b/.github/workflows/ci.yml @@ -23,7 +23,7 @@ jobs: - uses: actions/setup-python@v2 - name: Install Black run: | - pip install black==19.10b0 + pip install black - name: Lint with Black run: | black --check --diff . @@ -111,3 +111,29 @@ jobs: uses: codecov/codecov-action@v1 with: fail_ci_if_error: true + + test-extras: + needs: [test] + runs-on: ${{ matrix.os }} + strategy: + matrix: + os: [ubuntu-latest] + python-version: [3.7, 3.8, 3.9] + + steps: + - uses: actions/checkout@v2 + with: + fetch-depth: 0 + persist-credentials: false + - name: Setup Python ${{ matrix.python-version }} + uses: actions/setup-python@v2 + with: + python-version: ${{ matrix.python-version }} + - name: Install dependencies + run: | + pip install --upgrade pip setuptools wheel + pip install pytest-cov + pip install .[ezancestry] + - name: Test with pytest + run: | + pytest --cov=snps tests diff --git a/Pipfile b/Pipfile index 2f151b3..5f0f1c1 100644 --- a/Pipfile +++ b/Pipfile @@ -9,7 +9,7 @@ pytest-cov = "*" pytest-watch = "*" sphinx = "*" sphinx-rtd-theme = "*" -black = "==19.10b0" +black = "==21.10b0" matplotlib = "*" [packages] diff --git a/Pipfile.lock b/Pipfile.lock index 38791ea..f16991e 100644 --- a/Pipfile.lock +++ b/Pipfile.lock @@ -1,7 +1,7 @@ { "_meta": { "hash": { - "sha256": "11f3af56d80f81e99d56db7de554f1cf85aa21768fb2f867d122e29efdd99f0e" + "sha256": "1cbbd1515d1a00b1feb0abee9bbec0adffdd45f8a1ad655653a67c8de53742ad" }, "pipfile-spec": 6, "requires": { @@ -26,70 +26,87 @@ }, "numpy": { "hashes": [ - "sha256:1676b0a292dd3c99e49305a16d7a9f42a4ab60ec522eac0d3dd20cdf362ac010", - "sha256:16f221035e8bd19b9dc9a57159e38d2dd060b48e93e1d843c49cb370b0f415fd", - "sha256:43909c8bb289c382170e0282158a38cf306a8ad2ff6dfadc447e90f9961bef43", - "sha256:4e465afc3b96dbc80cf4a5273e5e2b1e3451286361b4af70ce1adb2984d392f9", - "sha256:55b745fca0a5ab738647d0e4db099bd0a23279c32b31a783ad2ccea729e632df", - "sha256:5d050e1e4bc9ddb8656d7b4f414557720ddcca23a5b88dd7cff65e847864c400", - "sha256:637d827248f447e63585ca3f4a7d2dfaa882e094df6cfa177cc9cf9cd6cdf6d2", - "sha256:6690080810f77485667bfbff4f69d717c3be25e5b11bb2073e76bb3f578d99b4", - "sha256:66fbc6fed94a13b9801fb70b96ff30605ab0a123e775a5e7a26938b717c5d71a", - "sha256:67d44acb72c31a97a3d5d33d103ab06d8ac20770e1c5ad81bdb3f0c086a56cf6", - "sha256:6ca2b85a5997dabc38301a22ee43c82adcb53ff660b89ee88dded6b33687e1d8", - "sha256:6e51534e78d14b4a009a062641f465cfaba4fdcb046c3ac0b1f61dd97c861b1b", - "sha256:70eb5808127284c4e5c9e836208e09d685a7978b6a216db85960b1a112eeace8", - "sha256:830b044f4e64a76ba71448fce6e604c0fc47a0e54d8f6467be23749ac2cbd2fb", - "sha256:8b7bb4b9280da3b2856cb1fc425932f46fba609819ee1c62256f61799e6a51d2", - "sha256:a9c65473ebc342715cb2d7926ff1e202c26376c0dcaaee85a1fd4b8d8c1d3b2f", - "sha256:c1c09247ccea742525bdb5f4b5ceeacb34f95731647fe55774aa36557dbb5fa4", - "sha256:c5bf0e132acf7557fc9bb8ded8b53bbbbea8892f3c9a1738205878ca9434206a", - "sha256:db250fd3e90117e0312b611574cd1b3f78bec046783195075cbd7ba9c3d73f16", - "sha256:e515c9a93aebe27166ec9593411c58494fa98e5fcc219e47260d9ab8a1cc7f9f", - "sha256:e55185e51b18d788e49fe8305fd73ef4470596b33fc2c1ceb304566b99c71a69", - "sha256:ea9cff01e75a956dbee133fa8e5b68f2f92175233de2f88de3a682dd94deda65", - "sha256:f1452578d0516283c87608a5a5548b0cdde15b99650efdfd85182102ef7a7c17", - "sha256:f39a995e47cb8649673cfa0579fbdd1cdd33ea497d1728a6cb194d6252268e48" - ], - "markers": "python_version >= '3.7'", - "version": "==1.20.3" + "sha256:043e83bfc274649c82a6f09836943e4a4aebe5e33656271c7dbf9621dd58b8ec", + "sha256:160ccc1bed3a8371bf0d760971f09bfe80a3e18646620e9ded0ad159d9749baa", + "sha256:188031f833bbb623637e66006cf75e933e00e7231f67e2b45cf8189612bb5dc3", + "sha256:28f15209fb535dd4c504a7762d3bc440779b0e37d50ed810ced209e5cea60d96", + "sha256:29fb3dcd0468b7715f8ce2c0c2d9bbbaf5ae686334951343a41bd8d155c6ea27", + "sha256:2a6ee9620061b2a722749b391c0d80a0e2ae97290f1b32e28d5a362e21941ee4", + "sha256:300321e3985c968e3ae7fbda187237b225f3ffe6528395a5b7a5407f73cf093e", + "sha256:32437f0b275c1d09d9c3add782516413e98cd7c09e6baf4715cbce781fc29912", + "sha256:3c09418a14471c7ae69ba682e2428cae5b4420a766659605566c0fa6987f6b7e", + "sha256:49c6249260890e05b8111ebfc391ed58b3cb4b33e63197b2ec7f776e45330721", + "sha256:4cc9b512e9fb590797474f58b7f6d1f1b654b3a94f4fa8558b48ca8b3cfc97cf", + "sha256:508b0b513fa1266875524ba8a9ecc27b02ad771fe1704a16314dc1a816a68737", + "sha256:50cd26b0cf6664cb3b3dd161ba0a09c9c1343db064e7c69f9f8b551f5104d654", + "sha256:5c4193f70f8069550a1788bd0cd3268ab7d3a2b70583dfe3b2e7f421e9aace06", + "sha256:5dfe9d6a4c39b8b6edd7990091fea4f852888e41919d0e6722fe78dd421db0eb", + "sha256:63571bb7897a584ca3249c86dd01c10bcb5fe4296e3568b2e9c1a55356b6410e", + "sha256:75621882d2230ab77fb6a03d4cbccd2038511491076e7964ef87306623aa5272", + "sha256:75eb7cadc8da49302f5b659d40ba4f6d94d5045fbd9569c9d058e77b0514c9e4", + "sha256:88a5d6b268e9ad18f3533e184744acdaa2e913b13148160b1152300c949bbb5f", + "sha256:8a10968963640e75cc0193e1847616ab4c718e83b6938ae74dea44953950f6b7", + "sha256:90bec6a86b348b4559b6482e2b684db4a9a7eed1fa054b86115a48d58fbbf62a", + "sha256:98339aa9911853f131de11010f6dd94c8cec254d3d1f7261528c3b3e3219f139", + "sha256:a99a6b067e5190ac6d12005a4d85aa6227c5606fa93211f86b1dafb16233e57d", + "sha256:bffa2eee3b87376cc6b31eee36d05349571c236d1de1175b804b348dc0941e3f", + "sha256:c6c2d535a7beb1f8790aaa98fd089ceab2e3dd7ca48aca0af7dc60e6ef93ffe1", + "sha256:cc14e7519fab2a4ed87d31f99c31a3796e4e1fe63a86ebdd1c5a1ea78ebd5896", + "sha256:dd0482f3fc547f1b1b5d6a8b8e08f63fdc250c58ce688dedd8851e6e26cff0f3", + "sha256:dde972a1e11bb7b702ed0e447953e7617723760f420decb97305e66fb4afc54f", + "sha256:e54af82d68ef8255535a6cdb353f55d6b8cf418a83e2be3569243787a4f4866f", + "sha256:e606e6316911471c8d9b4618e082635cfe98876007556e89ce03d52ff5e8fcf0", + "sha256:f41b018f126aac18583956c54544db437f25c7ee4794bcb23eb38bef8e5e192a", + "sha256:f8f4625536926a155b80ad2bbff44f8cc59e9f2ad14cdda7acf4c135b4dc8ff2", + "sha256:fe52dbe47d9deb69b05084abd4b0df7abb39a3c51957c09f635520abd49b29dd" + ], + "markers": "python_version < '3.10' and platform_machine != 'aarch64' and platform_machine != 'arm64'", + "version": "==1.21.3" }, "pandas": { "hashes": [ - "sha256:167693a80abc8eb28051fbd184c1b7afd13ce2c727a5af47b048f1ea3afefff4", - "sha256:2111c25e69fa9365ba80bbf4f959400054b2771ac5d041ed19415a8b488dc70a", - "sha256:298f0553fd3ba8e002c4070a723a59cdb28eda579f3e243bc2ee397773f5398b", - "sha256:2b063d41803b6a19703b845609c0b700913593de067b552a8b24dd8eeb8c9895", - "sha256:2cb7e8f4f152f27dc93f30b5c7a98f6c748601ea65da359af734dd0cf3fa733f", - "sha256:52d2472acbb8a56819a87aafdb8b5b6d2b3386e15c95bde56b281882529a7ded", - "sha256:612add929bf3ba9d27b436cc8853f5acc337242d6b584203f207e364bb46cb12", - "sha256:649ecab692fade3cbfcf967ff936496b0cfba0af00a55dfaacd82bdda5cb2279", - "sha256:68d7baa80c74aaacbed597265ca2308f017859123231542ff8a5266d489e1858", - "sha256:8d4c74177c26aadcfb4fd1de6c1c43c2bf822b3e0fc7a9b409eeaf84b3e92aaa", - "sha256:971e2a414fce20cc5331fe791153513d076814d30a60cd7348466943e6e909e4", - "sha256:9db70ffa8b280bb4de83f9739d514cd0735825e79eef3a61d312420b9f16b758", - "sha256:b730add5267f873b3383c18cac4df2527ac4f0f0eed1c6cf37fcb437e25cf558", - "sha256:bd659c11a4578af740782288cac141a322057a2e36920016e0fc7b25c5a4b686", - "sha256:c601c6fdebc729df4438ec1f62275d6136a0dd14d332fc0e8ce3f7d2aadb4dd6", - "sha256:d0877407359811f7b853b548a614aacd7dea83b0c0c84620a9a643f180060950" + "sha256:003ba92db58b71a5f8add604a17a059f3068ef4e8c0c365b088468d0d64935fd", + "sha256:10e10a2527db79af6e830c3d5842a4d60383b162885270f8cffc15abca4ba4a9", + "sha256:22808afb8f96e2269dcc5b846decacb2f526dd0b47baebc63d913bf847317c8f", + "sha256:2d1dc09c0013d8faa7474574d61b575f9af6257ab95c93dcf33a14fd8d2c1bab", + "sha256:35c77609acd2e4d517da41bae0c11c70d31c87aae8dd1aabd2670906c6d2c143", + "sha256:372d72a3d8a5f2dbaf566a5fa5fa7f230842ac80f29a931fb4b071502cf86b9a", + "sha256:42493f8ae67918bf129869abea8204df899902287a7f5eaf596c8e54e0ac7ff4", + "sha256:4acc28364863127bca1029fb72228e6f473bb50c32e77155e80b410e2068eeac", + "sha256:5298a733e5bfbb761181fd4672c36d0c627320eb999c59c65156c6a90c7e1b4f", + "sha256:5ba0aac1397e1d7b654fccf263a4798a9e84ef749866060d19e577e927d66e1b", + "sha256:9707bdc1ea9639c886b4d3be6e2a45812c1ac0c2080f94c31b71c9fa35556f9b", + "sha256:a2aa18d3f0b7d538e21932f637fbfe8518d085238b429e4790a35e1e44a96ffc", + "sha256:a388960f979665b447f0847626e40f99af8cf191bce9dc571d716433130cb3a7", + "sha256:a51528192755f7429c5bcc9e80832c517340317c861318fea9cea081b57c9afd", + "sha256:b528e126c13816a4374e56b7b18bfe91f7a7f6576d1aadba5dee6a87a7f479ae", + "sha256:c1aa4de4919358c5ef119f6377bc5964b3a7023c23e845d9db7d9016fa0c5b1c", + "sha256:c2646458e1dce44df9f71a01dc65f7e8fa4307f29e5c0f2f92c97f47a5bf22f5", + "sha256:c2f44425594ae85e119459bb5abb0748d76ef01d9c08583a667e3339e134218e", + "sha256:d47750cf07dee6b55d8423471be70d627314277976ff2edd1381f02d52dbadf9", + "sha256:d99d2350adb7b6c3f7f8f0e5dfb7d34ff8dd4bc0a53e62c445b7e43e163fce63", + "sha256:dd324f8ee05925ee85de0ea3f0d66e1362e8c80799eb4eb04927d32335a3e44a", + "sha256:eaca36a80acaacb8183930e2e5ad7f71539a66805d6204ea88736570b2876a7b", + "sha256:f567e972dce3bbc3a8076e0b675273b4a9e8576ac629149cf8286ee13c259ae5", + "sha256:fe48e4925455c964db914b958f6e7032d285848b7538a5e1b19aeb26ffaea3ec" ], "markers": "python_full_version >= '3.7.1'", - "version": "==1.2.4" + "version": "==1.3.4" }, "python-dateutil": { "hashes": [ - "sha256:73ebfe9dbf22e832286dafa60473e4cd239f8592f699aa5adaf10050e6e1823c", - "sha256:75bb3f31ea686f1197762692a9ee6a7550b59fc6ca3a1f4b5d7e32fb98e2da2a" + "sha256:0123cacc1627ae19ddf3c27a5de5bd67ee4586fbdd6440d9748f8abb483d3e86", + "sha256:961d03dc3453ebbc59dbdea9e4e11c5651520a876d0f4db161e8674aae935da9" ], "markers": "python_version >= '2.7' and python_version not in '3.0, 3.1, 3.2, 3.3'", - "version": "==2.8.1" + "version": "==2.8.2" }, "pytz": { "hashes": [ - "sha256:83a4a90894bf38e243cf052c8b58f381bfe9a7a483f6a9cab140bc7f702ac4da", - "sha256:eb10ce3e7736052ed3623d49975ce333bcd712c7bb19a58b9e2089d4057d0798" + "sha256:3672058bc3453457b622aab7a1c3bfd5ab0bdae451512f6cf25f64ed37f5b87c", + "sha256:acad2d8b20a1af07d4e4c9d2e9285c5ed9104354062f275f3fcd88dcef4f1326" ], - "version": "==2021.1" + "version": "==2021.3" }, "six": { "hashes": [ @@ -112,13 +129,6 @@ ], "version": "==0.7.12" }, - "appdirs": { - "hashes": [ - "sha256:7d5d0167b2b1ba821647616af46a749d1c653740dd0d2415100fe26e27afdf41", - "sha256:a841dacd6b99318a741b166adb07e19ee71a274450e68237b4650ca1055ab128" - ], - "version": "==1.4.4" - }, "attrs": { "hashes": [ "sha256:149e90d6d8ac20db7a955ad60cf0e6881a3f20d37096140088356da6c716b0b1", @@ -137,34 +147,34 @@ }, "black": { "hashes": [ - "sha256:1b30e59be925fafc1ee4565e5e08abef6b03fe455102883820fe5ee2e4734e0b", - "sha256:c2edb73a08e9e0e6f65a0e6af18b059b8b1cdd5bef997d7a0b181df93dc81539" + "sha256:6eb7448da9143ee65b856a5f3676b7dda98ad9abe0f87fce8c59291f15e82a5b", + "sha256:a9952229092e325fe5f3dae56d81f639b23f7131eb840781947e4b2886030f33" ], "index": "pypi", - "version": "==19.10b0" + "version": "==21.10b0" }, "certifi": { "hashes": [ - "sha256:2bbf76fd432960138b3ef6dda3dde0544f27cbf8546c458e60baf371917ba9ee", - "sha256:50b1e4f8446b06f41be7dd6338db18e0990601dce795c2b1686458aa7e8fa7d8" + "sha256:78884e7c1d4b00ce3cea67b44566851c4343c120abd683433ce934a68ea58872", + "sha256:d62a0163eb4c2344ac042ab2bdf75399a71a2d8c7d47eac2e2ee91b9d6339569" ], - "version": "==2021.5.30" + "version": "==2021.10.8" }, - "chardet": { + "charset-normalizer": { "hashes": [ - "sha256:0d6f53a15db4120f2b08c94f11e7d93d2c911ee118b6b30a04ec3ee8310179fa", - "sha256:f864054d66fd9118f2e67044ac8981a54775ec5b67aed0441892edb553d21da5" + "sha256:e019de665e2bcf9c2b64e2e5aa025fa991da8720daa3c1138cadd2fd1856aed0", + "sha256:f7af805c321bfa1ce6714c51f254e0d5bb5e5834039bc17db7ebe3a4cec9492b" ], - "markers": "python_version >= '2.7' and python_version not in '3.0, 3.1, 3.2, 3.3, 3.4'", - "version": "==4.0.0" + "markers": "python_version >= '3'", + "version": "==2.0.7" }, "click": { "hashes": [ - "sha256:8c04c11192119b1ef78ea049e0a6f0463e4c48ef00a30160c704337586f3ad7a", - "sha256:fba402a4a47334742d782209a7c79bc448911afe1149d07bdabdf480b3e2f4b6" + "sha256:353f466495adaeb40b6b5f592f9f91cb22372351c84caeb068132442a4518ef3", + "sha256:410e932b050f5eed773c4cda94de75971c89cdb3155a72a0831139a79e5ecb5b" ], "markers": "python_version >= '3.6'", - "version": "==8.0.1" + "version": "==8.0.3" }, "colorama": { "hashes": [ @@ -175,69 +185,70 @@ "version": "==0.4.4" }, "coverage": { - "hashes": [ - "sha256:004d1880bed2d97151facef49f08e255a20ceb6f9432df75f4eef018fdd5a78c", - "sha256:01d84219b5cdbfc8122223b39a954820929497a1cb1422824bb86b07b74594b6", - "sha256:040af6c32813fa3eae5305d53f18875bedd079960822ef8ec067a66dd8afcd45", - "sha256:06191eb60f8d8a5bc046f3799f8a07a2d7aefb9504b0209aff0b47298333302a", - "sha256:13034c4409db851670bc9acd836243aeee299949bd5673e11844befcb0149f03", - "sha256:13c4ee887eca0f4c5a247b75398d4114c37882658300e153113dafb1d76de529", - "sha256:184a47bbe0aa6400ed2d41d8e9ed868b8205046518c52464fde713ea06e3a74a", - "sha256:18ba8bbede96a2c3dde7b868de9dcbd55670690af0988713f0603f037848418a", - "sha256:1aa846f56c3d49205c952d8318e76ccc2ae23303351d9270ab220004c580cfe2", - "sha256:217658ec7187497e3f3ebd901afdca1af062b42cfe3e0dafea4cced3983739f6", - "sha256:24d4a7de75446be83244eabbff746d66b9240ae020ced65d060815fac3423759", - "sha256:2910f4d36a6a9b4214bb7038d537f015346f413a975d57ca6b43bf23d6563b53", - "sha256:2949cad1c5208b8298d5686d5a85b66aae46d73eec2c3e08c817dd3513e5848a", - "sha256:2a3859cb82dcbda1cfd3e6f71c27081d18aa251d20a17d87d26d4cd216fb0af4", - "sha256:2cafbbb3af0733db200c9b5f798d18953b1a304d3f86a938367de1567f4b5bff", - "sha256:2e0d881ad471768bf6e6c2bf905d183543f10098e3b3640fc029509530091502", - "sha256:30c77c1dc9f253283e34c27935fded5015f7d1abe83bc7821680ac444eaf7793", - "sha256:3487286bc29a5aa4b93a072e9592f22254291ce96a9fbc5251f566b6b7343cdb", - "sha256:372da284cfd642d8e08ef606917846fa2ee350f64994bebfbd3afb0040436905", - "sha256:41179b8a845742d1eb60449bdb2992196e211341818565abded11cfa90efb821", - "sha256:44d654437b8ddd9eee7d1eaee28b7219bec228520ff809af170488fd2fed3e2b", - "sha256:4a7697d8cb0f27399b0e393c0b90f0f1e40c82023ea4d45d22bce7032a5d7b81", - "sha256:51cb9476a3987c8967ebab3f0fe144819781fca264f57f89760037a2ea191cb0", - "sha256:52596d3d0e8bdf3af43db3e9ba8dcdaac724ba7b5ca3f6358529d56f7a166f8b", - "sha256:53194af30d5bad77fcba80e23a1441c71abfb3e01192034f8246e0d8f99528f3", - "sha256:5fec2d43a2cc6965edc0bb9e83e1e4b557f76f843a77a2496cbe719583ce8184", - "sha256:6c90e11318f0d3c436a42409f2749ee1a115cd8b067d7f14c148f1ce5574d701", - "sha256:74d881fc777ebb11c63736622b60cb9e4aee5cace591ce274fb69e582a12a61a", - "sha256:7501140f755b725495941b43347ba8a2777407fc7f250d4f5a7d2a1050ba8e82", - "sha256:796c9c3c79747146ebd278dbe1e5c5c05dd6b10cc3bcb8389dfdf844f3ead638", - "sha256:869a64f53488f40fa5b5b9dcb9e9b2962a66a87dab37790f3fcfb5144b996ef5", - "sha256:8963a499849a1fc54b35b1c9f162f4108017b2e6db2c46c1bed93a72262ed083", - "sha256:8d0a0725ad7c1a0bcd8d1b437e191107d457e2ec1084b9f190630a4fb1af78e6", - "sha256:900fbf7759501bc7807fd6638c947d7a831fc9fdf742dc10f02956ff7220fa90", - "sha256:92b017ce34b68a7d67bd6d117e6d443a9bf63a2ecf8567bb3d8c6c7bc5014465", - "sha256:970284a88b99673ccb2e4e334cfb38a10aab7cd44f7457564d11898a74b62d0a", - "sha256:972c85d205b51e30e59525694670de6a8a89691186012535f9d7dbaa230e42c3", - "sha256:9a1ef3b66e38ef8618ce5fdc7bea3d9f45f3624e2a66295eea5e57966c85909e", - "sha256:af0e781009aaf59e25c5a678122391cb0f345ac0ec272c7961dc5455e1c40066", - "sha256:b6d534e4b2ab35c9f93f46229363e17f63c53ad01330df9f2d6bd1187e5eaacf", - "sha256:b7895207b4c843c76a25ab8c1e866261bcfe27bfaa20c192de5190121770672b", - "sha256:c0891a6a97b09c1f3e073a890514d5012eb256845c451bd48f7968ef939bf4ae", - "sha256:c2723d347ab06e7ddad1a58b2a821218239249a9e4365eaff6649d31180c1669", - "sha256:d1f8bf7b90ba55699b3a5e44930e93ff0189aa27186e96071fac7dd0d06a1873", - "sha256:d1f9ce122f83b2305592c11d64f181b87153fc2c2bbd3bb4a3dde8303cfb1a6b", - "sha256:d314ed732c25d29775e84a960c3c60808b682c08d86602ec2c3008e1202e3bb6", - "sha256:d636598c8305e1f90b439dbf4f66437de4a5e3c31fdf47ad29542478c8508bbb", - "sha256:deee1077aae10d8fa88cb02c845cfba9b62c55e1183f52f6ae6a2df6a2187160", - "sha256:ebe78fe9a0e874362175b02371bdfbee64d8edc42a044253ddf4ee7d3c15212c", - "sha256:f030f8873312a16414c0d8e1a1ddff2d3235655a2174e3648b4fa66b3f2f1079", - "sha256:f0b278ce10936db1a37e6954e15a3730bea96a0997c26d7fee88e6c396c2086d", - "sha256:f11642dddbb0253cc8853254301b51390ba0081750a8ac03f20ea8103f0c56b6" + "extras": [ + "toml" + ], + "hashes": [ + "sha256:0147f7833c41927d84f5af9219d9b32f875c0689e5e74ac8ca3cb61e73a698f9", + "sha256:04a92a6cf9afd99f9979c61348ec79725a9f9342fb45e63c889e33c04610d97b", + "sha256:10ab138b153e4cc408b43792cb7f518f9ee02f4ff55cd1ab67ad6fd7e9905c7e", + "sha256:2e5b9c17a56b8bf0c0a9477fcd30d357deb486e4e1b389ed154f608f18556c8a", + "sha256:326d944aad0189603733d646e8d4a7d952f7145684da973c463ec2eefe1387c2", + "sha256:359a32515e94e398a5c0fa057e5887a42e647a9502d8e41165cf5cb8d3d1ca67", + "sha256:35cd2230e1ed76df7d0081a997f0fe705be1f7d8696264eb508076e0d0b5a685", + "sha256:3b270c6b48d3ff5a35deb3648028ba2643ad8434b07836782b1139cf9c66313f", + "sha256:42a1fb5dee3355df90b635906bb99126faa7936d87dfc97eacc5293397618cb7", + "sha256:479228e1b798d3c246ac89b09897ee706c51b3e5f8f8d778067f38db73ccc717", + "sha256:4cd919057636f63ab299ccb86ea0e78b87812400c76abab245ca385f17d19fb5", + "sha256:4d8b453764b9b26b0dd2afb83086a7c3f9379134e340288d2a52f8a91592394b", + "sha256:51a441011a30d693e71dea198b2a6f53ba029afc39f8e2aeb5b77245c1b282ef", + "sha256:557594a50bfe3fb0b1b57460f6789affe8850ad19c1acf2d14a3e12b2757d489", + "sha256:572f917267f363101eec375c109c9c1118037c7cc98041440b5eabda3185ac7b", + "sha256:62512c0ec5d307f56d86504c58eace11c1bc2afcdf44e3ff20de8ca427ca1d0e", + "sha256:65ad3ff837c89a229d626b8004f0ee32110f9bfdb6a88b76a80df36ccc60d926", + "sha256:666c6b32b69e56221ad1551d377f718ed00e6167c7a1b9257f780b105a101271", + "sha256:6e994003e719458420e14ffb43c08f4c14990e20d9e077cb5cad7a3e419bbb54", + "sha256:72bf437d54186d104388cbae73c9f2b0f8a3e11b6e8d7deb593bd14625c96026", + "sha256:738e823a746841248b56f0f3bd6abf3b73af191d1fd65e4c723b9c456216f0ad", + "sha256:78287731e3601ea5ce9d6468c82d88a12ef8fe625d6b7bdec9b45d96c1ad6533", + "sha256:7833c872718dc913f18e51ee97ea0dece61d9930893a58b20b3daf09bb1af6b6", + "sha256:7e083d32965d2eb6638a77e65b622be32a094fdc0250f28ce6039b0732fbcaa8", + "sha256:8186b5a4730c896cbe1e4b645bdc524e62d874351ae50e1db7c3e9f5dc81dc26", + "sha256:8605add58e6a960729aa40c0fd9a20a55909dd9b586d3e8104cc7f45869e4c6b", + "sha256:977ce557d79577a3dd510844904d5d968bfef9489f512be65e2882e1c6eed7d8", + "sha256:994ce5a7b3d20981b81d83618aa4882f955bfa573efdbef033d5632b58597ba9", + "sha256:9ad5895938a894c368d49d8470fe9f519909e5ebc6b8f8ea5190bd0df6aa4271", + "sha256:9eb0a1923354e0fdd1c8a6f53f5db2e6180d670e2b587914bf2e79fa8acfd003", + "sha256:a00284dbfb53b42e35c7dd99fc0e26ef89b4a34efff68078ed29d03ccb28402a", + "sha256:a11a2c019324fc111485e79d55907e7289e53d0031275a6c8daed30690bc50c0", + "sha256:ab6a0fe4c96f8058d41948ddf134420d3ef8c42d5508b5a341a440cce7a37a1d", + "sha256:ae6de0e41f44794e68d23644636544ed8003ce24845f213b24de097cbf44997f", + "sha256:b1d0a1bce919de0dd8da5cff4e616b2d9e6ebf3bd1410ff645318c3dd615010a", + "sha256:b8e4f15b672c9156c1154249a9c5746e86ac9ae9edc3799ee3afebc323d9d9e0", + "sha256:bbca34dca5a2d60f81326d908d77313816fad23d11b6069031a3d6b8c97a54f9", + "sha256:bf656cd74ff7b4ed7006cdb2a6728150aaad69c7242b42a2a532f77b63ea233f", + "sha256:c40966b683d92869b72ea3c11fd6b99a091fd30e12652727eca117273fc97366", + "sha256:c95257aa2ccf75d3d91d772060538d5fea7f625e48157f8ca44594f94d41cb33", + "sha256:db2797ed7a7e883b9ab76e8e778bb4c859fc2037d6fd0644d8675e64d58d1653", + "sha256:dc5023be1c2a8b0a0ab5e31389e62c28b2453eb31dd069f4b8d1a0f9814d951a", + "sha256:e14bceb1f3ae8a14374be2b2d7bc12a59226872285f91d66d301e5f41705d4d6", + "sha256:e3c4f5211394cd0bf6874ac5d29684a495f9c374919833dcfff0bd6d37f96201", + "sha256:e76f017b6d4140a038c5ff12be1581183d7874e41f1c0af58ecf07748d36a336", + "sha256:e7d5606b9240ed4def9cbdf35be4308047d11e858b9c88a6c26974758d6225ce", + "sha256:f0f80e323a17af63eac6a9db0c9188c10f1fd815c3ab299727150cc0eb92c7a4", + "sha256:fb2fa2f6506c03c48ca42e3fe5a692d7470d290c047ee6de7c0f3e5fa7639ac9", + "sha256:ffa8fee2b1b9e60b531c4c27cf528d6b5d5da46b1730db1f4d6eee56ff282e07" ], - "markers": "python_version >= '2.7' and python_version not in '3.0, 3.1, 3.2, 3.3, 3.4' and python_version < '4'", - "version": "==5.5" + "markers": "python_version >= '3.6'", + "version": "==6.1.1" }, "cycler": { "hashes": [ - "sha256:1d8a5ae1ff6c5cf9b93e8811e581232ad8920aeec647c37316ceac982b08cb2d", - "sha256:cd7b2d1018258d7247a71425e9f26463dfb444d411c39569972f4ce586b0c9d8" + "sha256:3a27e95f763a428a739d2add979fa7494c912a32c17c4c38c4d5f082cad165a3", + "sha256:9c87405839a19696e837b3b818fed3f5f69f16f1eec1a1ad77e043dcea9c772f" ], - "version": "==0.10.0" + "markers": "python_version >= '3.6'", + "version": "==0.11.0" }, "docopt": { "hashes": [ @@ -247,19 +258,19 @@ }, "docutils": { "hashes": [ - "sha256:0c5b78adfbf7762415433f5515cd5c9e762339e23369dbe8000d84a4bf4ab3af", - "sha256:c2de3a60e9e7d07be26b7f2b00ca0309c207e06c100f9cc2a94931fc75a478fc" + "sha256:686577d2e4c32380bb50cbb22f575ed742d58168cee37e99117a854bcd88f125", + "sha256:cf316c8370a737a022b72b56874f6602acf974a37a9fba42ec2876387549fc61" ], "markers": "python_version >= '2.7' and python_version not in '3.0, 3.1, 3.2, 3.3, 3.4'", - "version": "==0.16" + "version": "==0.17.1" }, "idna": { "hashes": [ - "sha256:b307872f855b18632ce0c21c5e45be78c0ea7ae4c15c828c20788b26921eb3f6", - "sha256:b97d804b1e9b523befed77c48dacec60e6dcb0b5391d57af6a65a312a90648c0" + "sha256:84d9dd047ffa80596e0f246e2eab0b391788b0503584e8945f2368256d2735ff", + "sha256:9d643ff0a55b762d5cdb124b8eaa99c66322e2157b69160bc32796e824360e6d" ], - "markers": "python_version >= '2.7' and python_version not in '3.0, 3.1, 3.2, 3.3'", - "version": "==2.10" + "markers": "python_version >= '3'", + "version": "==3.3" }, "imagesize": { "hashes": [ @@ -271,11 +282,11 @@ }, "importlib-metadata": { "hashes": [ - "sha256:833b26fb89d5de469b24a390e9df088d4e52e4ba33b01dc5e0e4f41b81a16c00", - "sha256:b142cc1dd1342f31ff04bb7d022492b09920cb64fed867cd3ea6f80fe3ebd139" + "sha256:b618b6d2d5ffa2f16add5697cf57a46c76a56229b0ed1c438322e4e95645bd15", + "sha256:f284b3e11256ad1e5d03ab86bb2ccd6f5339688ff17a4d797a0fe7df326f23b1" ], "markers": "python_version < '3.8'", - "version": "==4.5.0" + "version": "==4.8.1" }, "iniconfig": { "hashes": [ @@ -286,81 +297,128 @@ }, "jinja2": { "hashes": [ - "sha256:1f06f2da51e7b56b8f238affdd6b4e2c61e39598a378cc49345bc1bd42a978a4", - "sha256:703f484b47a6af502e743c9122595cc812b0271f661722403114f71a79d0f5a4" + "sha256:827a0e32839ab1600d4eb1c4c33ec5a8edfbc5cb42dafa13b81f182f97784b45", + "sha256:8569982d3f0889eed11dd620c706d39b60c36d6d25843961f33f77fb6bc6b20c" ], "markers": "python_version >= '3.6'", - "version": "==3.0.1" + "version": "==3.0.2" }, "kiwisolver": { "hashes": [ - "sha256:0cd53f403202159b44528498de18f9285b04482bab2a6fc3f5dd8dbb9352e30d", - "sha256:1e1bc12fb773a7b2ffdeb8380609f4f8064777877b2225dec3da711b421fda31", - "sha256:225e2e18f271e0ed8157d7f4518ffbf99b9450fca398d561eb5c4a87d0986dd9", - "sha256:232c9e11fd7ac3a470d65cd67e4359eee155ec57e822e5220322d7b2ac84fbf0", - "sha256:31dfd2ac56edc0ff9ac295193eeaea1c0c923c0355bf948fbd99ed6018010b72", - "sha256:33449715e0101e4d34f64990352bce4095c8bf13bed1b390773fc0a7295967b3", - "sha256:401a2e9afa8588589775fe34fc22d918ae839aaaf0c0e96441c0fdbce6d8ebe6", - "sha256:44a62e24d9b01ba94ae7a4a6c3fb215dc4af1dde817e7498d901e229aaf50e4e", - "sha256:50af681a36b2a1dee1d3c169ade9fdc59207d3c31e522519181e12f1b3ba7000", - "sha256:563c649cfdef27d081c84e72a03b48ea9408c16657500c312575ae9d9f7bc1c3", - "sha256:5989db3b3b34b76c09253deeaf7fbc2707616f130e166996606c284395da3f18", - "sha256:5a7a7dbff17e66fac9142ae2ecafb719393aaee6a3768c9de2fd425c63b53e21", - "sha256:5c3e6455341008a054cccee8c5d24481bcfe1acdbc9add30aa95798e95c65621", - "sha256:5f6ccd3dd0b9739edcf407514016108e2280769c73a85b9e59aa390046dbf08b", - "sha256:72c99e39d005b793fb7d3d4e660aed6b6281b502e8c1eaf8ee8346023c8e03bc", - "sha256:78751b33595f7f9511952e7e60ce858c6d64db2e062afb325985ddbd34b5c131", - "sha256:834ee27348c4aefc20b479335fd422a2c69db55f7d9ab61721ac8cd83eb78882", - "sha256:8be8d84b7d4f2ba4ffff3665bcd0211318aa632395a1a41553250484a871d454", - "sha256:950a199911a8d94683a6b10321f9345d5a3a8433ec58b217ace979e18f16e248", - "sha256:a357fd4f15ee49b4a98b44ec23a34a95f1e00292a139d6015c11f55774ef10de", - "sha256:a53d27d0c2a0ebd07e395e56a1fbdf75ffedc4a05943daf472af163413ce9598", - "sha256:acef3d59d47dd85ecf909c359d0fd2c81ed33bdff70216d3956b463e12c38a54", - "sha256:b38694dcdac990a743aa654037ff1188c7a9801ac3ccc548d3341014bc5ca278", - "sha256:b9edd0110a77fc321ab090aaa1cfcaba1d8499850a12848b81be2222eab648f6", - "sha256:c08e95114951dc2090c4a630c2385bef681cacf12636fb0241accdc6b303fd81", - "sha256:c5518d51a0735b1e6cee1fdce66359f8d2b59c3ca85dc2b0813a8aa86818a030", - "sha256:c8fd0f1ae9d92b42854b2979024d7597685ce4ada367172ed7c09edf2cef9cb8", - "sha256:ca3820eb7f7faf7f0aa88de0e54681bddcb46e485beb844fcecbcd1c8bd01689", - "sha256:cf8b574c7b9aa060c62116d4181f3a1a4e821b2ec5cbfe3775809474113748d4", - "sha256:d3155d828dec1d43283bd24d3d3e0d9c7c350cdfcc0bd06c0ad1209c1bbc36d0", - "sha256:f8d6f8db88049a699817fd9178782867bf22283e3813064302ac59f61d95be05", - "sha256:fd34fbbfbc40628200730bc1febe30631347103fc8d3d4fa012c21ab9c11eca9" + "sha256:0007840186bacfaa0aba4466d5890334ea5938e0bb7e28078a0eb0e63b5b59d5", + "sha256:19554bd8d54cf41139f376753af1a644b63c9ca93f8f72009d50a2080f870f77", + "sha256:1d45d1c74f88b9f41062716c727f78f2a59a5476ecbe74956fafb423c5c87a76", + "sha256:1d819553730d3c2724582124aee8a03c846ec4362ded1034c16fb3ef309264e6", + "sha256:2210f28778c7d2ee13f3c2a20a3a22db889e75f4ec13a21072eabb5693801e84", + "sha256:22521219ca739654a296eea6d4367703558fba16f98688bd8ce65abff36eaa84", + "sha256:25405f88a37c5f5bcba01c6e350086d65e7465fd1caaf986333d2a045045a223", + "sha256:2b65bd35f3e06a47b5c30ea99e0c2b88f72c6476eedaf8cfbc8e66adb5479dcf", + "sha256:2ddb500a2808c100e72c075cbb00bf32e62763c82b6a882d403f01a119e3f402", + "sha256:2f8f6c8f4f1cff93ca5058d6ec5f0efda922ecb3f4c5fb76181f327decff98b8", + "sha256:30fa008c172355c7768159983a7270cb23838c4d7db73d6c0f6b60dde0d432c6", + "sha256:3dbb3cea20b4af4f49f84cffaf45dd5f88e8594d18568e0225e6ad9dec0e7967", + "sha256:4116ba9a58109ed5e4cb315bdcbff9838f3159d099ba5259c7c7fb77f8537492", + "sha256:44e6adf67577dbdfa2d9f06db9fbc5639afefdb5bf2b4dfec25c3a7fbc619536", + "sha256:5326ddfacbe51abf9469fe668944bc2e399181a2158cb5d45e1d40856b2a0589", + "sha256:70adc3658138bc77a36ce769f5f183169bc0a2906a4f61f09673f7181255ac9b", + "sha256:72be6ebb4e92520b9726d7146bc9c9b277513a57a38efcf66db0620aec0097e0", + "sha256:7843b1624d6ccca403a610d1277f7c28ad184c5aa88a1750c1a999754e65b439", + "sha256:7ba5a1041480c6e0a8b11a9544d53562abc2d19220bfa14133e0cdd9967e97af", + "sha256:80efd202108c3a4150e042b269f7c78643420cc232a0a771743bb96b742f838f", + "sha256:82f49c5a79d3839bc8f38cb5f4bfc87e15f04cbafa5fbd12fb32c941cb529cfb", + "sha256:83d2c9db5dfc537d0171e32de160461230eb14663299b7e6d18ca6dca21e4977", + "sha256:8d93a1095f83e908fc253f2fb569c2711414c0bfd451cab580466465b235b470", + "sha256:8dc3d842fa41a33fe83d9f5c66c0cc1f28756530cd89944b63b072281e852031", + "sha256:9661a04ca3c950a8ac8c47f53cbc0b530bce1b52f516a1e87b7736fec24bfff0", + "sha256:a498bcd005e8a3fedd0022bb30ee0ad92728154a8798b703f394484452550507", + "sha256:a7a4cf5bbdc861987a7745aed7a536c6405256853c94abc9f3287c3fa401b174", + "sha256:b5074fb09429f2b7bc82b6fb4be8645dcbac14e592128beeff5461dcde0af09f", + "sha256:b6a5431940f28b6de123de42f0eb47b84a073ee3c3345dc109ad550a3307dd28", + "sha256:ba677bcaff9429fd1bf01648ad0901cea56c0d068df383d5f5856d88221fe75b", + "sha256:bcadb05c3d4794eb9eee1dddf1c24215c92fb7b55a80beae7a60530a91060560", + "sha256:bf7eb45d14fc036514c09554bf983f2a72323254912ed0c3c8e697b62c4c158f", + "sha256:c358721aebd40c243894298f685a19eb0491a5c3e0b923b9f887ef1193ddf829", + "sha256:c4550a359c5157aaf8507e6820d98682872b9100ce7607f8aa070b4b8af6c298", + "sha256:c6572c2dab23c86a14e82c245473d45b4c515314f1f859e92608dcafbd2f19b8", + "sha256:cba430db673c29376135e695c6e2501c44c256a81495da849e85d1793ee975ad", + "sha256:dedc71c8eb9c5096037766390172c34fb86ef048b8e8958b4e484b9e505d66bc", + "sha256:e6f5eb2f53fac7d408a45fbcdeda7224b1cfff64919d0f95473420a931347ae9", + "sha256:ec2eba188c1906b05b9b49ae55aae4efd8150c61ba450e6721f64620c50b59eb", + "sha256:ee040a7de8d295dbd261ef2d6d3192f13e2b08ec4a954de34a6fb8ff6422e24c", + "sha256:eedd3b59190885d1ebdf6c5e0ca56828beb1949b4dfe6e5d0256a461429ac386", + "sha256:f441422bb313ab25de7b3dbfd388e790eceb76ce01a18199ec4944b369017009", + "sha256:f8eb7b6716f5b50e9c06207a14172cf2de201e41912ebe732846c02c830455b9", + "sha256:fc4453705b81d03568d5b808ad8f09c77c47534f6ac2e72e733f9ca4714aa75c" ], - "markers": "python_version >= '3.6'", - "version": "==1.3.1" + "markers": "python_version >= '3.7'", + "version": "==1.3.2" }, "markupsafe": { "hashes": [ "sha256:01a9b8ea66f1658938f65b93a85ebe8bc016e6769611be228d797c9d998dd298", "sha256:023cb26ec21ece8dc3907c0e8320058b2e0cb3c55cf9564da612bc325bed5e64", "sha256:0446679737af14f45767963a1a9ef7620189912317d095f2d9ffa183a4d25d2b", + "sha256:04635854b943835a6ea959e948d19dcd311762c5c0c6e1f0e16ee57022669194", "sha256:0717a7390a68be14b8c793ba258e075c6f4ca819f15edfc2a3a027c823718567", "sha256:0955295dd5eec6cb6cc2fe1698f4c6d84af2e92de33fbcac4111913cd100a6ff", + "sha256:0d4b31cc67ab36e3392bbf3862cfbadac3db12bdd8b02a2731f509ed5b829724", "sha256:10f82115e21dc0dfec9ab5c0223652f7197feb168c940f3ef61563fc2d6beb74", + "sha256:168cd0a3642de83558a5153c8bd34f175a9a6e7f6dc6384b9655d2697312a646", "sha256:1d609f577dc6e1aa17d746f8bd3c31aa4d258f4070d61b2aa5c4166c1539de35", + "sha256:1f2ade76b9903f39aa442b4aadd2177decb66525062db244b35d71d0ee8599b6", + "sha256:20dca64a3ef2d6e4d5d615a3fd418ad3bde77a47ec8a23d984a12b5b4c74491a", + "sha256:2a7d351cbd8cfeb19ca00de495e224dea7e7d919659c2841bbb7f420ad03e2d6", + "sha256:2d7d807855b419fc2ed3e631034685db6079889a1f01d5d9dac950f764da3dad", "sha256:2ef54abee730b502252bcdf31b10dacb0a416229b72c18b19e24a4509f273d26", + "sha256:36bc903cbb393720fad60fc28c10de6acf10dc6cc883f3e24ee4012371399a38", + "sha256:37205cac2a79194e3750b0af2a5720d95f786a55ce7df90c3af697bfa100eaac", "sha256:3c112550557578c26af18a1ccc9e090bfe03832ae994343cfdacd287db6a6ae7", + "sha256:3dd007d54ee88b46be476e293f48c85048603f5f516008bee124ddd891398ed6", + "sha256:4296f2b1ce8c86a6aea78613c34bb1a672ea0e3de9c6ba08a960efe0b0a09047", "sha256:47ab1e7b91c098ab893b828deafa1203de86d0bc6ab587b160f78fe6c4011f75", "sha256:49e3ceeabbfb9d66c3aef5af3a60cc43b85c33df25ce03d0031a608b0a8b2e3f", + "sha256:4dc8f9fb58f7364b63fd9f85013b780ef83c11857ae79f2feda41e270468dd9b", "sha256:4efca8f86c54b22348a5467704e3fec767b2db12fc39c6d963168ab1d3fc9135", "sha256:53edb4da6925ad13c07b6d26c2a852bd81e364f95301c66e930ab2aef5b5ddd8", + "sha256:5855f8438a7d1d458206a2466bf82b0f104a3724bf96a1c781ab731e4201731a", "sha256:594c67807fb16238b30c44bdf74f36c02cdf22d1c8cda91ef8a0ed8dabf5620a", + "sha256:5b6d930f030f8ed98e3e6c98ffa0652bdb82601e7a016ec2ab5d7ff23baa78d1", + "sha256:5bb28c636d87e840583ee3adeb78172efc47c8b26127267f54a9c0ec251d41a9", + "sha256:60bf42e36abfaf9aff1f50f52644b336d4f0a3fd6d8a60ca0d054ac9f713a864", "sha256:611d1ad9a4288cf3e3c16014564df047fe08410e628f89805e475368bd304914", + "sha256:6300b8454aa6930a24b9618fbb54b5a68135092bc666f7b06901f897fa5c2fee", + "sha256:63f3268ba69ace99cab4e3e3b5840b03340efed0948ab8f78d2fd87ee5442a4f", "sha256:6557b31b5e2c9ddf0de32a691f2312a32f77cd7681d8af66c2692efdbef84c18", "sha256:693ce3f9e70a6cf7d2fb9e6c9d8b204b6b39897a2c4a1aa65728d5ac97dcc1d8", "sha256:6a7fae0dd14cf60ad5ff42baa2e95727c3d81ded453457771d02b7d2b3f9c0c2", "sha256:6c4ca60fa24e85fe25b912b01e62cb969d69a23a5d5867682dd3e80b5b02581d", + "sha256:6fcf051089389abe060c9cd7caa212c707e58153afa2c649f00346ce6d260f1b", "sha256:7d91275b0245b1da4d4cfa07e0faedd5b0812efc15b702576d103293e252af1b", + "sha256:89c687013cb1cd489a0f0ac24febe8c7a666e6e221b783e53ac50ebf68e45d86", + "sha256:8d206346619592c6200148b01a2142798c989edcb9c896f9ac9722a99d4e77e6", "sha256:905fec760bd2fa1388bb5b489ee8ee5f7291d692638ea5f67982d968366bef9f", "sha256:97383d78eb34da7e1fa37dd273c20ad4320929af65d156e35a5e2d89566d9dfb", "sha256:984d76483eb32f1bcb536dc27e4ad56bba4baa70be32fa87152832cdd9db0833", + "sha256:99df47edb6bda1249d3e80fdabb1dab8c08ef3975f69aed437cb69d0a5de1e28", + "sha256:9f02365d4e99430a12647f09b6cc8bab61a6564363f313126f775eb4f6ef798e", "sha256:a30e67a65b53ea0a5e62fe23682cfe22712e01f453b95233b25502f7c61cb415", "sha256:ab3ef638ace319fa26553db0624c4699e31a28bb2a835c5faca8f8acf6a5a902", + "sha256:aca6377c0cb8a8253e493c6b451565ac77e98c2951c45f913e0b52facdcff83f", + "sha256:add36cb2dbb8b736611303cd3bfcee00afd96471b09cda130da3581cbdc56a6d", "sha256:b2f4bf27480f5e5e8ce285a8c8fd176c0b03e93dcc6646477d4630e83440c6a9", "sha256:b7f2d075102dc8c794cbde1947378051c4e5180d52d276987b8d28a3bd58c17d", + "sha256:baa1a4e8f868845af802979fcdbf0bb11f94f1cb7ced4c4b8a351bb60d108145", "sha256:be98f628055368795d818ebf93da628541e10b75b41c559fdf36d104c5787066", + "sha256:bf5d821ffabf0ef3533c39c518f3357b171a1651c1ff6827325e4489b0e46c3c", + "sha256:c47adbc92fc1bb2b3274c4b3a43ae0e4573d9fbff4f54cd484555edbf030baf1", + "sha256:cdfba22ea2f0029c9261a4bd07e830a8da012291fbe44dc794e488b6c9bb353a", + "sha256:d6c7ebd4e944c85e2c3421e612a7057a2f48d478d79e61800d81468a8d842207", "sha256:d7f9850398e85aba693bb640262d3611788b1f29a79f0c93c565694658f4071f", + "sha256:d8446c54dc28c01e5a2dbac5a25f071f6653e6e40f3a8818e8b45d790fe6ef53", + "sha256:deb993cacb280823246a026e3b2d81c493c53de6acfd5e6bfe31ab3402bb37dd", + "sha256:e0f138900af21926a02425cf736db95be9f4af72ba1bb21453432a07f6082134", + "sha256:e9936f0b261d4df76ad22f8fee3ae83b60d7c3e871292cd42f40b81b70afae85", + "sha256:f0567c4dc99f264f49fe27da5f735f414c4e7e7dd850cfd8e69f0862d7c74ea9", "sha256:f5653a225f31e113b152e56f154ccbe59eeb1c7487b39b9d9f9cdb58e6c79dc5", "sha256:f826e31d18b516f653fe296d967d700fddad5901ae07c622bb3705955e1faa94", "sha256:f8ba0e8349a38d3001fae7eadded3f6606f0da5d748ee53cc1dab1d6527b9509", @@ -372,121 +430,154 @@ }, "matplotlib": { "hashes": [ - "sha256:0bea5ec5c28d49020e5d7923c2725b837e60bc8be99d3164af410eb4b4c827da", - "sha256:1c1779f7ab7d8bdb7d4c605e6ffaa0614b3e80f1e3c8ccf7b9269a22dbc5986b", - "sha256:21b31057bbc5e75b08e70a43cefc4c0b2c2f1b1a850f4a0f7af044eb4163086c", - "sha256:32fa638cc10886885d1ca3d409d4473d6a22f7ceecd11322150961a70fab66dd", - "sha256:3a5c18dbd2c7c366da26a4ad1462fe3e03a577b39e3b503bbcf482b9cdac093c", - "sha256:5826f56055b9b1c80fef82e326097e34dc4af8c7249226b7dd63095a686177d1", - "sha256:6382bc6e2d7e481bcd977eb131c31dee96e0fb4f9177d15ec6fb976d3b9ace1a", - "sha256:6475d0209024a77f869163ec3657c47fed35d9b6ed8bccba8aa0f0099fbbdaa8", - "sha256:6a6a44f27aabe720ec4fd485061e8a35784c2b9ffa6363ad546316dfc9cea04e", - "sha256:7a58f3d8fe8fac3be522c79d921c9b86e090a59637cb88e3bc51298d7a2c862a", - "sha256:7ad19f3fb6145b9eb41c08e7cbb9f8e10b91291396bee21e9ce761bb78df63ec", - "sha256:85f191bb03cb1a7b04b5c2cca4792bef94df06ef473bc49e2818105671766fee", - "sha256:956c8849b134b4a343598305a3ca1bdd3094f01f5efc8afccdebeffe6b315247", - "sha256:a9d8cb5329df13e0cdaa14b3b43f47b5e593ec637f13f14db75bb16e46178b05", - "sha256:b1d5a2cedf5de05567c441b3a8c2651fbde56df08b82640e7f06c8cd91e201f6", - "sha256:b26535b9de85326e6958cdef720ecd10bcf74a3f4371bf9a7e5b2e659c17e153", - "sha256:c541ee5a3287efe066bbe358320853cf4916bc14c00c38f8f3d8d75275a405a9", - "sha256:d8d994cefdff9aaba45166eb3de4f5211adb4accac85cbf97137e98f26ea0219", - "sha256:df815378a754a7edd4559f8c51fc7064f779a74013644a7f5ac7a0c31f875866" + "sha256:01c9de93a2ca0d128c9064f23709362e7fefb34910c7c9e0b8ab0de8258d5eda", + "sha256:41b6e307458988891fcdea2d8ecf84a8c92d53f84190aa32da65f9505546e684", + "sha256:48e1e0859b54d5f2e29bb78ca179fd59b971c6ceb29977fb52735bfd280eb0f5", + "sha256:54a026055d5f8614f184e588f6e29064019a0aa8448450214c0b60926d62d919", + "sha256:556965514b259204637c360d213de28d43a1f4aed1eca15596ce83f768c5a56f", + "sha256:5c988bb43414c7c2b0a31bd5187b4d27fd625c080371b463a6d422047df78913", + "sha256:6a724e3a48a54b8b6e7c4ae38cd3d07084508fa47c410c8757e9db9791421838", + "sha256:6be8df61b1626e1a142c57e065405e869e9429b4a6dab4a324757d0dc4d42235", + "sha256:844a7b0233e4ff7fba57e90b8799edaa40b9e31e300b8d5efc350937fa8b1bea", + "sha256:85f0c9cf724715e75243a7b3087cf4a3de056b55e05d4d76cc58d610d62894f3", + "sha256:a78a3b51f29448c7f4d4575e561f6b0dbb8d01c13c2046ab6c5220eb25c06506", + "sha256:b884715a59fec9ad3b6048ecf3860f3b2ce965e676ef52593d6fa29abcf7d330", + "sha256:b8b53f336a4688cfce615887505d7e41fd79b3594bf21dd300531a4f5b4f746a", + "sha256:c70b6311dda3e27672f1bf48851a0de816d1ca6aaf3d49365fbdd8e959b33d2b", + "sha256:ebfb01a65c3f5d53a8c2a8133fec2b5221281c053d944ae81ff5822a68266617", + "sha256:eeb1859efe7754b1460e1d4991bbd4a60a56f366bc422ef3a9c5ae05f0bc70b5", + "sha256:f15edcb0629a0801738925fe27070480f446fcaa15de65946ff946ad99a59a40", + "sha256:f1c5efc278d996af8a251b2ce0b07bbeccb821f25c8c9846bdcb00ffc7f158aa", + "sha256:f72657f1596199dc1e4e7a10f52a4784ead8a711f4e5b59bea95bdb97cf0e4fd", + "sha256:fc4f526dfdb31c9bd6b8ca06bf9fab663ca12f3ec9cdf4496fb44bc680140318", + "sha256:fcd6f1954943c0c192bfbebbac263f839d7055409f1173f80d8b11a224d236da" ], "index": "pypi", - "version": "==3.4.2" + "version": "==3.4.3" }, - "numpy": { + "mypy-extensions": { "hashes": [ - "sha256:1676b0a292dd3c99e49305a16d7a9f42a4ab60ec522eac0d3dd20cdf362ac010", - "sha256:16f221035e8bd19b9dc9a57159e38d2dd060b48e93e1d843c49cb370b0f415fd", - "sha256:43909c8bb289c382170e0282158a38cf306a8ad2ff6dfadc447e90f9961bef43", - "sha256:4e465afc3b96dbc80cf4a5273e5e2b1e3451286361b4af70ce1adb2984d392f9", - "sha256:55b745fca0a5ab738647d0e4db099bd0a23279c32b31a783ad2ccea729e632df", - "sha256:5d050e1e4bc9ddb8656d7b4f414557720ddcca23a5b88dd7cff65e847864c400", - "sha256:637d827248f447e63585ca3f4a7d2dfaa882e094df6cfa177cc9cf9cd6cdf6d2", - "sha256:6690080810f77485667bfbff4f69d717c3be25e5b11bb2073e76bb3f578d99b4", - "sha256:66fbc6fed94a13b9801fb70b96ff30605ab0a123e775a5e7a26938b717c5d71a", - "sha256:67d44acb72c31a97a3d5d33d103ab06d8ac20770e1c5ad81bdb3f0c086a56cf6", - "sha256:6ca2b85a5997dabc38301a22ee43c82adcb53ff660b89ee88dded6b33687e1d8", - "sha256:6e51534e78d14b4a009a062641f465cfaba4fdcb046c3ac0b1f61dd97c861b1b", - "sha256:70eb5808127284c4e5c9e836208e09d685a7978b6a216db85960b1a112eeace8", - "sha256:830b044f4e64a76ba71448fce6e604c0fc47a0e54d8f6467be23749ac2cbd2fb", - "sha256:8b7bb4b9280da3b2856cb1fc425932f46fba609819ee1c62256f61799e6a51d2", - "sha256:a9c65473ebc342715cb2d7926ff1e202c26376c0dcaaee85a1fd4b8d8c1d3b2f", - "sha256:c1c09247ccea742525bdb5f4b5ceeacb34f95731647fe55774aa36557dbb5fa4", - "sha256:c5bf0e132acf7557fc9bb8ded8b53bbbbea8892f3c9a1738205878ca9434206a", - "sha256:db250fd3e90117e0312b611574cd1b3f78bec046783195075cbd7ba9c3d73f16", - "sha256:e515c9a93aebe27166ec9593411c58494fa98e5fcc219e47260d9ab8a1cc7f9f", - "sha256:e55185e51b18d788e49fe8305fd73ef4470596b33fc2c1ceb304566b99c71a69", - "sha256:ea9cff01e75a956dbee133fa8e5b68f2f92175233de2f88de3a682dd94deda65", - "sha256:f1452578d0516283c87608a5a5548b0cdde15b99650efdfd85182102ef7a7c17", - "sha256:f39a995e47cb8649673cfa0579fbdd1cdd33ea497d1728a6cb194d6252268e48" + "sha256:090fedd75945a69ae91ce1303b5824f428daf5a028d2f6ab8a299250a846f15d", + "sha256:2d82818f5bb3e369420cb3c4060a7970edba416647068eb4c5343488a6c604a8" ], - "markers": "python_version >= '3.7'", - "version": "==1.20.3" + "version": "==0.4.3" + }, + "numpy": { + "hashes": [ + "sha256:043e83bfc274649c82a6f09836943e4a4aebe5e33656271c7dbf9621dd58b8ec", + "sha256:160ccc1bed3a8371bf0d760971f09bfe80a3e18646620e9ded0ad159d9749baa", + "sha256:188031f833bbb623637e66006cf75e933e00e7231f67e2b45cf8189612bb5dc3", + "sha256:28f15209fb535dd4c504a7762d3bc440779b0e37d50ed810ced209e5cea60d96", + "sha256:29fb3dcd0468b7715f8ce2c0c2d9bbbaf5ae686334951343a41bd8d155c6ea27", + "sha256:2a6ee9620061b2a722749b391c0d80a0e2ae97290f1b32e28d5a362e21941ee4", + "sha256:300321e3985c968e3ae7fbda187237b225f3ffe6528395a5b7a5407f73cf093e", + "sha256:32437f0b275c1d09d9c3add782516413e98cd7c09e6baf4715cbce781fc29912", + "sha256:3c09418a14471c7ae69ba682e2428cae5b4420a766659605566c0fa6987f6b7e", + "sha256:49c6249260890e05b8111ebfc391ed58b3cb4b33e63197b2ec7f776e45330721", + "sha256:4cc9b512e9fb590797474f58b7f6d1f1b654b3a94f4fa8558b48ca8b3cfc97cf", + "sha256:508b0b513fa1266875524ba8a9ecc27b02ad771fe1704a16314dc1a816a68737", + "sha256:50cd26b0cf6664cb3b3dd161ba0a09c9c1343db064e7c69f9f8b551f5104d654", + "sha256:5c4193f70f8069550a1788bd0cd3268ab7d3a2b70583dfe3b2e7f421e9aace06", + "sha256:5dfe9d6a4c39b8b6edd7990091fea4f852888e41919d0e6722fe78dd421db0eb", + "sha256:63571bb7897a584ca3249c86dd01c10bcb5fe4296e3568b2e9c1a55356b6410e", + "sha256:75621882d2230ab77fb6a03d4cbccd2038511491076e7964ef87306623aa5272", + "sha256:75eb7cadc8da49302f5b659d40ba4f6d94d5045fbd9569c9d058e77b0514c9e4", + "sha256:88a5d6b268e9ad18f3533e184744acdaa2e913b13148160b1152300c949bbb5f", + "sha256:8a10968963640e75cc0193e1847616ab4c718e83b6938ae74dea44953950f6b7", + "sha256:90bec6a86b348b4559b6482e2b684db4a9a7eed1fa054b86115a48d58fbbf62a", + "sha256:98339aa9911853f131de11010f6dd94c8cec254d3d1f7261528c3b3e3219f139", + "sha256:a99a6b067e5190ac6d12005a4d85aa6227c5606fa93211f86b1dafb16233e57d", + "sha256:bffa2eee3b87376cc6b31eee36d05349571c236d1de1175b804b348dc0941e3f", + "sha256:c6c2d535a7beb1f8790aaa98fd089ceab2e3dd7ca48aca0af7dc60e6ef93ffe1", + "sha256:cc14e7519fab2a4ed87d31f99c31a3796e4e1fe63a86ebdd1c5a1ea78ebd5896", + "sha256:dd0482f3fc547f1b1b5d6a8b8e08f63fdc250c58ce688dedd8851e6e26cff0f3", + "sha256:dde972a1e11bb7b702ed0e447953e7617723760f420decb97305e66fb4afc54f", + "sha256:e54af82d68ef8255535a6cdb353f55d6b8cf418a83e2be3569243787a4f4866f", + "sha256:e606e6316911471c8d9b4618e082635cfe98876007556e89ce03d52ff5e8fcf0", + "sha256:f41b018f126aac18583956c54544db437f25c7ee4794bcb23eb38bef8e5e192a", + "sha256:f8f4625536926a155b80ad2bbff44f8cc59e9f2ad14cdda7acf4c135b4dc8ff2", + "sha256:fe52dbe47d9deb69b05084abd4b0df7abb39a3c51957c09f635520abd49b29dd" + ], + "markers": "python_version < '3.10' and platform_machine != 'aarch64' and platform_machine != 'arm64'", + "version": "==1.21.3" }, "packaging": { "hashes": [ - "sha256:5b327ac1320dc863dca72f4514ecc086f31186744b84a230374cc1fd776feae5", - "sha256:67714da7f7bc052e064859c05c595155bd1ee9f69f76557e21f051443c20947a" + "sha256:096d689d78ca690e4cd8a89568ba06d07ca097e3306a4381635073ca91479966", + "sha256:14317396d1e8cdb122989b916fa2c7e9ca8e2be9e8060a6eff75b6b7b4d8a7e0" ], - "markers": "python_version >= '2.7' and python_version not in '3.0, 3.1, 3.2, 3.3'", - "version": "==20.9" + "markers": "python_version >= '3.6'", + "version": "==21.2" }, "pathspec": { "hashes": [ - "sha256:86379d6b86d75816baba717e64b1a3a3469deb93bb76d613c9ce79edc5cb68fd", - "sha256:aa0cb481c4041bf52ffa7b0d8fa6cd3e88a2ca4879c533c9153882ee2556790d" + "sha256:7d15c4ddb0b5c802d161efc417ec1a2558ea2653c2e8ad9c19098201dc1c993a", + "sha256:e564499435a2673d586f6b2130bb5b95f04a3ba06f81b8f895b651a3c76aabb1" ], - "version": "==0.8.1" + "version": "==0.9.0" }, "pillow": { "hashes": [ - "sha256:01425106e4e8cee195a411f729cff2a7d61813b0b11737c12bd5991f5f14bcd5", - "sha256:031a6c88c77d08aab84fecc05c3cde8414cd6f8406f4d2b16fed1e97634cc8a4", - "sha256:083781abd261bdabf090ad07bb69f8f5599943ddb539d64497ed021b2a67e5a9", - "sha256:0d19d70ee7c2ba97631bae1e7d4725cdb2ecf238178096e8c82ee481e189168a", - "sha256:0e04d61f0064b545b989126197930807c86bcbd4534d39168f4aa5fda39bb8f9", - "sha256:12e5e7471f9b637762453da74e390e56cc43e486a88289995c1f4c1dc0bfe727", - "sha256:22fd0f42ad15dfdde6c581347eaa4adb9a6fc4b865f90b23378aa7914895e120", - "sha256:238c197fc275b475e87c1453b05b467d2d02c2915fdfdd4af126145ff2e4610c", - "sha256:3b570f84a6161cf8865c4e08adf629441f56e32f180f7aa4ccbd2e0a5a02cba2", - "sha256:463822e2f0d81459e113372a168f2ff59723e78528f91f0bd25680ac185cf797", - "sha256:4d98abdd6b1e3bf1a1cbb14c3895226816e666749ac040c4e2554231068c639b", - "sha256:5afe6b237a0b81bd54b53f835a153770802f164c5570bab5e005aad693dab87f", - "sha256:5b70110acb39f3aff6b74cf09bb4169b167e2660dabc304c1e25b6555fa781ef", - "sha256:5cbf3e3b1014dddc45496e8cf38b9f099c95a326275885199f427825c6522232", - "sha256:624b977355cde8b065f6d51b98497d6cd5fbdd4f36405f7a8790e3376125e2bb", - "sha256:63728564c1410d99e6d1ae8e3b810fe012bc440952168af0a2877e8ff5ab96b9", - "sha256:66cc56579fd91f517290ab02c51e3a80f581aba45fd924fcdee01fa06e635812", - "sha256:6c32cc3145928c4305d142ebec682419a6c0a8ce9e33db900027ddca1ec39178", - "sha256:8b56553c0345ad6dcb2e9b433ae47d67f95fc23fe28a0bde15a120f25257e291", - "sha256:8bb1e155a74e1bfbacd84555ea62fa21c58e0b4e7e6b20e4447b8d07990ac78b", - "sha256:95d5ef984eff897850f3a83883363da64aae1000e79cb3c321915468e8c6add5", - "sha256:a013cbe25d20c2e0c4e85a9daf438f85121a4d0344ddc76e33fd7e3965d9af4b", - "sha256:a787ab10d7bb5494e5f76536ac460741788f1fbce851068d73a87ca7c35fc3e1", - "sha256:a7d5e9fad90eff8f6f6106d3b98b553a88b6f976e51fce287192a5d2d5363713", - "sha256:aac00e4bc94d1b7813fe882c28990c1bc2f9d0e1aa765a5f2b516e8a6a16a9e4", - "sha256:b91c36492a4bbb1ee855b7d16fe51379e5f96b85692dc8210831fbb24c43e484", - "sha256:c03c07ed32c5324939b19e36ae5f75c660c81461e312a41aea30acdd46f93a7c", - "sha256:c5236606e8570542ed424849f7852a0ff0bce2c4c8d0ba05cc202a5a9c97dee9", - "sha256:c6b39294464b03457f9064e98c124e09008b35a62e3189d3513e5148611c9388", - "sha256:cb7a09e173903541fa888ba010c345893cd9fc1b5891aaf060f6ca77b6a3722d", - "sha256:d68cb92c408261f806b15923834203f024110a2e2872ecb0bd2a110f89d3c602", - "sha256:dc38f57d8f20f06dd7c3161c59ca2c86893632623f33a42d592f097b00f720a9", - "sha256:e98eca29a05913e82177b3ba3d198b1728e164869c613d76d0de4bde6768a50e", - "sha256:f217c3954ce5fd88303fc0c317af55d5e0204106d86dea17eb8205700d47dec2" + "sha256:066f3999cb3b070a95c3652712cffa1a748cd02d60ad7b4e485c3748a04d9d76", + "sha256:0a0956fdc5defc34462bb1c765ee88d933239f9a94bc37d132004775241a7585", + "sha256:0b052a619a8bfcf26bd8b3f48f45283f9e977890263e4571f2393ed8898d331b", + "sha256:1394a6ad5abc838c5cd8a92c5a07535648cdf6d09e8e2d6df916dfa9ea86ead8", + "sha256:1bc723b434fbc4ab50bb68e11e93ce5fb69866ad621e3c2c9bdb0cd70e345f55", + "sha256:244cf3b97802c34c41905d22810846802a3329ddcb93ccc432870243211c79fc", + "sha256:25a49dc2e2f74e65efaa32b153527fc5ac98508d502fa46e74fa4fd678ed6645", + "sha256:2e4440b8f00f504ee4b53fe30f4e381aae30b0568193be305256b1462216feff", + "sha256:3862b7256046fcd950618ed22d1d60b842e3a40a48236a5498746f21189afbbc", + "sha256:3eb1ce5f65908556c2d8685a8f0a6e989d887ec4057326f6c22b24e8a172c66b", + "sha256:3f97cfb1e5a392d75dd8b9fd274d205404729923840ca94ca45a0af57e13dbe6", + "sha256:493cb4e415f44cd601fcec11c99836f707bb714ab03f5ed46ac25713baf0ff20", + "sha256:4acc0985ddf39d1bc969a9220b51d94ed51695d455c228d8ac29fcdb25810e6e", + "sha256:5503c86916d27c2e101b7f71c2ae2cddba01a2cf55b8395b0255fd33fa4d1f1a", + "sha256:5b7bb9de00197fb4261825c15551adf7605cf14a80badf1761d61e59da347779", + "sha256:5e9ac5f66616b87d4da618a20ab0a38324dbe88d8a39b55be8964eb520021e02", + "sha256:620582db2a85b2df5f8a82ddeb52116560d7e5e6b055095f04ad828d1b0baa39", + "sha256:62cc1afda735a8d109007164714e73771b499768b9bb5afcbbee9d0ff374b43f", + "sha256:70ad9e5c6cb9b8487280a02c0ad8a51581dcbbe8484ce058477692a27c151c0a", + "sha256:72b9e656e340447f827885b8d7a15fc8c4e68d410dc2297ef6787eec0f0ea409", + "sha256:72cbcfd54df6caf85cc35264c77ede902452d6df41166010262374155947460c", + "sha256:792e5c12376594bfcb986ebf3855aa4b7c225754e9a9521298e460e92fb4a488", + "sha256:7b7017b61bbcdd7f6363aeceb881e23c46583739cb69a3ab39cb384f6ec82e5b", + "sha256:81f8d5c81e483a9442d72d182e1fb6dcb9723f289a57e8030811bac9ea3fef8d", + "sha256:82aafa8d5eb68c8463b6e9baeb4f19043bb31fefc03eb7b216b51e6a9981ae09", + "sha256:84c471a734240653a0ec91dec0996696eea227eafe72a33bd06c92697728046b", + "sha256:8c803ac3c28bbc53763e6825746f05cc407b20e4a69d0122e526a582e3b5e153", + "sha256:93ce9e955cc95959df98505e4608ad98281fff037350d8c2671c9aa86bcf10a9", + "sha256:9a3e5ddc44c14042f0844b8cf7d2cd455f6cc80fd7f5eefbe657292cf601d9ad", + "sha256:a4901622493f88b1a29bd30ec1a2f683782e57c3c16a2dbc7f2595ba01f639df", + "sha256:a5a4532a12314149d8b4e4ad8ff09dde7427731fcfa5917ff16d0291f13609df", + "sha256:b8831cb7332eda5dc89b21a7bce7ef6ad305548820595033a4b03cf3091235ed", + "sha256:b8e2f83c56e141920c39464b852de3719dfbfb6e3c99a2d8da0edf4fb33176ed", + "sha256:c70e94281588ef053ae8998039610dbd71bc509e4acbc77ab59d7d2937b10698", + "sha256:c8a17b5d948f4ceeceb66384727dde11b240736fddeda54ca740b9b8b1556b29", + "sha256:d82cdb63100ef5eedb8391732375e6d05993b765f72cb34311fab92103314649", + "sha256:d89363f02658e253dbd171f7c3716a5d340a24ee82d38aab9183f7fdf0cdca49", + "sha256:d99ec152570e4196772e7a8e4ba5320d2d27bf22fdf11743dd882936ed64305b", + "sha256:ddc4d832a0f0b4c52fff973a0d44b6c99839a9d016fe4e6a1cb8f3eea96479c2", + "sha256:e3dacecfbeec9a33e932f00c6cd7996e62f53ad46fbe677577394aaa90ee419a", + "sha256:eb9fc393f3c61f9054e1ed26e6fe912c7321af2f41ff49d3f83d05bacf22cc78" + ], + "markers": "python_version >= '3.6'", + "version": "==8.4.0" + }, + "platformdirs": { + "hashes": [ + "sha256:367a5e80b3d04d2428ffa76d33f124cf11e8fff2acdaa9b43d545f5c7d661ef2", + "sha256:8868bbe3c3c80d42f20156f22e7131d2fb321f5bc86a2a345375c6481a67021d" ], "markers": "python_version >= '3.6'", - "version": "==8.2.0" + "version": "==2.4.0" }, "pluggy": { "hashes": [ - "sha256:15b2acde666561e1298d71b523007ed7364de07029219b604cf808bfa1c765b0", - "sha256:966c145cd83c96502c3c3868f50408687b38434af77734af1e9ca461a4081d2d" + "sha256:4224373bacce55f955a878bf9cfa763c1e360858e330072059e10bad68531159", + "sha256:74134bbf457f031a36d68416e1509f34bd5ccc019f0bcc952c7b909d06b37bd3" ], - "markers": "python_version >= '2.7' and python_version not in '3.0, 3.1, 3.2, 3.3'", - "version": "==0.13.1" + "markers": "python_version >= '3.6'", + "version": "==1.0.0" }, "py": { "hashes": [ @@ -498,11 +589,11 @@ }, "pygments": { "hashes": [ - "sha256:a18f47b506a429f6f4b9df81bb02beab9ca21d0a5fee38ed15aef65f0545519f", - "sha256:d66e804411278594d764fc69ec36ec13d9ae9147193a1740cd34d272ca383b8e" + "sha256:b8e67fe6af78f492b3c4b3e2970c0624cbf08beb1e493b2c99b9fa1b67a20380", + "sha256:f398865f7eb6874156579fdf36bc840a03cab64d1cde9e93d68f46a425ec52c6" ], "markers": "python_version >= '3.5'", - "version": "==2.9.0" + "version": "==2.10.0" }, "pyparsing": { "hashes": [ @@ -514,19 +605,19 @@ }, "pytest": { "hashes": [ - "sha256:50bcad0a0b9c5a72c8e4e7c9855a3ad496ca6a881a3641b4260605450772c54b", - "sha256:91ef2131a9bd6be8f76f1f08eac5c5317221d6ad1e143ae03894b862e8976890" + "sha256:131b36680866a76e6781d13f101efb86cf674ebb9762eb70d3082b6f29889e89", + "sha256:7310f8d27bc79ced999e760ca304d69f6ba6c6649c0b60fb0e04a4a77cacc134" ], "index": "pypi", - "version": "==6.2.4" + "version": "==6.2.5" }, "pytest-cov": { "hashes": [ - "sha256:261bb9e47e65bd099c89c3edf92972865210c36813f80ede5277dceb77a4a62a", - "sha256:261ceeb8c227b726249b376b8526b600f38667ee314f910353fa318caa01f4d7" + "sha256:578d5d15ac4a25e5f961c938b85a05b09fdaae9deef3bb6de9a6e766622ca7a6", + "sha256:e7f0f5b1617d2210a2cabc266dfe2f4c75a8d32fb89eafb7ad9d06f6d076d470" ], "index": "pypi", - "version": "==2.12.1" + "version": "==3.0.0" }, "pytest-watch": { "hashes": [ @@ -537,72 +628,67 @@ }, "python-dateutil": { "hashes": [ - "sha256:73ebfe9dbf22e832286dafa60473e4cd239f8592f699aa5adaf10050e6e1823c", - "sha256:75bb3f31ea686f1197762692a9ee6a7550b59fc6ca3a1f4b5d7e32fb98e2da2a" + "sha256:0123cacc1627ae19ddf3c27a5de5bd67ee4586fbdd6440d9748f8abb483d3e86", + "sha256:961d03dc3453ebbc59dbdea9e4e11c5651520a876d0f4db161e8674aae935da9" ], "markers": "python_version >= '2.7' and python_version not in '3.0, 3.1, 3.2, 3.3'", - "version": "==2.8.1" + "version": "==2.8.2" }, "pytz": { "hashes": [ - "sha256:83a4a90894bf38e243cf052c8b58f381bfe9a7a483f6a9cab140bc7f702ac4da", - "sha256:eb10ce3e7736052ed3623d49975ce333bcd712c7bb19a58b9e2089d4057d0798" + "sha256:3672058bc3453457b622aab7a1c3bfd5ab0bdae451512f6cf25f64ed37f5b87c", + "sha256:acad2d8b20a1af07d4e4c9d2e9285c5ed9104354062f275f3fcd88dcef4f1326" ], - "version": "==2021.1" + "version": "==2021.3" }, "regex": { "hashes": [ - "sha256:01afaf2ec48e196ba91b37451aa353cb7eda77efe518e481707e0515025f0cd5", - "sha256:11d773d75fa650cd36f68d7ca936e3c7afaae41b863b8c387a22aaa78d3c5c79", - "sha256:18c071c3eb09c30a264879f0d310d37fe5d3a3111662438889ae2eb6fc570c31", - "sha256:1e1c20e29358165242928c2de1482fb2cf4ea54a6a6dea2bd7a0e0d8ee321500", - "sha256:281d2fd05555079448537fe108d79eb031b403dac622621c78944c235f3fcf11", - "sha256:314d66636c494ed9c148a42731b3834496cc9a2c4251b1661e40936814542b14", - "sha256:32e65442138b7b76dd8173ffa2cf67356b7bc1768851dded39a7a13bf9223da3", - "sha256:339456e7d8c06dd36a22e451d58ef72cef293112b559010db3d054d5560ef439", - "sha256:3916d08be28a1149fb97f7728fca1f7c15d309a9f9682d89d79db75d5e52091c", - "sha256:3a9cd17e6e5c7eb328517969e0cb0c3d31fd329298dd0c04af99ebf42e904f82", - "sha256:47bf5bf60cf04d72bf6055ae5927a0bd9016096bf3d742fa50d9bf9f45aa0711", - "sha256:4c46e22a0933dd783467cf32b3516299fb98cfebd895817d685130cc50cd1093", - "sha256:4c557a7b470908b1712fe27fb1ef20772b78079808c87d20a90d051660b1d69a", - "sha256:52ba3d3f9b942c49d7e4bc105bb28551c44065f139a65062ab7912bef10c9afb", - "sha256:563085e55b0d4fb8f746f6a335893bda5c2cef43b2f0258fe1020ab1dd874df8", - "sha256:598585c9f0af8374c28edd609eb291b5726d7cbce16be6a8b95aa074d252ee17", - "sha256:619d71c59a78b84d7f18891fe914446d07edd48dc8328c8e149cbe0929b4e000", - "sha256:67bdb9702427ceddc6ef3dc382455e90f785af4c13d495f9626861763ee13f9d", - "sha256:6d1b01031dedf2503631d0903cb563743f397ccaf6607a5e3b19a3d76fc10480", - "sha256:741a9647fcf2e45f3a1cf0e24f5e17febf3efe8d4ba1281dcc3aa0459ef424dc", - "sha256:7c2a1af393fcc09e898beba5dd59196edaa3116191cc7257f9224beaed3e1aa0", - "sha256:7d9884d86dd4dd489e981d94a65cd30d6f07203d90e98f6f657f05170f6324c9", - "sha256:90f11ff637fe8798933fb29f5ae1148c978cccb0452005bf4c69e13db951e765", - "sha256:919859aa909429fb5aa9cf8807f6045592c85ef56fdd30a9a3747e513db2536e", - "sha256:96fcd1888ab4d03adfc9303a7b3c0bd78c5412b2bfbe76db5b56d9eae004907a", - "sha256:97f29f57d5b84e73fbaf99ab3e26134e6687348e95ef6b48cfd2c06807005a07", - "sha256:980d7be47c84979d9136328d882f67ec5e50008681d94ecc8afa8a65ed1f4a6f", - "sha256:a91aa8619b23b79bcbeb37abe286f2f408d2f2d6f29a17237afda55bb54e7aac", - "sha256:ade17eb5d643b7fead300a1641e9f45401c98eee23763e9ed66a43f92f20b4a7", - "sha256:b9c3db21af35e3b3c05764461b262d6f05bbca08a71a7849fd79d47ba7bc33ed", - "sha256:bd28bc2e3a772acbb07787c6308e00d9626ff89e3bfcdebe87fa5afbfdedf968", - "sha256:bf5824bfac591ddb2c1f0a5f4ab72da28994548c708d2191e3b87dd207eb3ad7", - "sha256:c0502c0fadef0d23b128605d69b58edb2c681c25d44574fc673b0e52dce71ee2", - "sha256:c38c71df845e2aabb7fb0b920d11a1b5ac8526005e533a8920aea97efb8ec6a4", - "sha256:ce15b6d103daff8e9fee13cf7f0add05245a05d866e73926c358e871221eae87", - "sha256:d3029c340cfbb3ac0a71798100ccc13b97dddf373a4ae56b6a72cf70dfd53bc8", - "sha256:e512d8ef5ad7b898cdb2d8ee1cb09a8339e4f8be706d27eaa180c2f177248a10", - "sha256:e8e5b509d5c2ff12f8418006d5a90e9436766133b564db0abaec92fd27fcee29", - "sha256:ee54ff27bf0afaf4c3b3a62bcd016c12c3fdb4ec4f413391a90bd38bc3624605", - "sha256:fa4537fb4a98fe8fde99626e4681cc644bdcf2a795038533f9f711513a862ae6", - "sha256:fd45ff9293d9274c5008a2054ecef86a9bfe819a67c7be1afb65e69b405b3042" - ], - "version": "==2021.4.4" + "sha256:0c186691a7995ef1db61205e00545bf161fb7b59cdb8c1201c89b333141c438a", + "sha256:0dcc0e71118be8c69252c207630faf13ca5e1b8583d57012aae191e7d6d28b84", + "sha256:0f7552429dd39f70057ac5d0e897e5bfe211629652399a21671e53f2a9693a4e", + "sha256:129472cd06062fb13e7b4670a102951a3e655e9b91634432cfbdb7810af9d710", + "sha256:13ec99df95003f56edcd307db44f06fbeb708c4ccdcf940478067dd62353181e", + "sha256:1f2b59c28afc53973d22e7bc18428721ee8ca6079becf1b36571c42627321c65", + "sha256:2b20f544cbbeffe171911f6ce90388ad36fe3fad26b7c7a35d4762817e9ea69c", + "sha256:2fb698037c35109d3c2e30f2beb499e5ebae6e4bb8ff2e60c50b9a805a716f79", + "sha256:34d870f9f27f2161709054d73646fc9aca49480617a65533fc2b4611c518e455", + "sha256:391703a2abf8013d95bae39145d26b4e21531ab82e22f26cd3a181ee2644c234", + "sha256:450dc27483548214314640c89a0f275dbc557968ed088da40bde7ef8fb52829e", + "sha256:45b65d6a275a478ac2cbd7fdbf7cc93c1982d613de4574b56fd6972ceadb8395", + "sha256:5095a411c8479e715784a0c9236568ae72509450ee2226b649083730f3fadfc6", + "sha256:530fc2bbb3dc1ebb17f70f7b234f90a1dd43b1b489ea38cea7be95fb21cdb5c7", + "sha256:56f0c81c44638dfd0e2367df1a331b4ddf2e771366c4b9c5d9a473de75e3e1c7", + "sha256:5e9c9e0ce92f27cef79e28e877c6b6988c48b16942258f3bc55d39b5f911df4f", + "sha256:6d7722136c6ed75caf84e1788df36397efdc5dbadab95e59c2bba82d4d808a4c", + "sha256:74d071dbe4b53c602edd87a7476ab23015a991374ddb228d941929ad7c8c922e", + "sha256:7b568809dca44cb75c8ebb260844ea98252c8c88396f9d203f5094e50a70355f", + "sha256:80bb5d2e92b2258188e7dcae5b188c7bf868eafdf800ea6edd0fbfc029984a88", + "sha256:8d1cdcda6bd16268316d5db1038965acf948f2a6f43acc2e0b1641ceab443623", + "sha256:9f665677e46c5a4d288ece12fdedf4f4204a422bb28ff05f0e6b08b7447796d1", + "sha256:a30513828180264294953cecd942202dfda64e85195ae36c265daf4052af0464", + "sha256:a7a986c45d1099a5de766a15de7bee3840b1e0e1a344430926af08e5297cf666", + "sha256:a940ca7e7189d23da2bfbb38973832813eab6bd83f3bf89a977668c2f813deae", + "sha256:ab7c5684ff3538b67df3f93d66bd3369b749087871ae3786e70ef39e601345b0", + "sha256:be04739a27be55631069b348dda0c81d8ea9822b5da10b8019b789e42d1fe452", + "sha256:c0938ddd60cc04e8f1faf7a14a166ac939aac703745bfcd8e8f20322a7373019", + "sha256:cb46b542133999580ffb691baf67410306833ee1e4f58ed06b6a7aaf4e046952", + "sha256:d134757a37d8640f3c0abb41f5e68b7cf66c644f54ef1cb0573b7ea1c63e1509", + "sha256:de557502c3bec8e634246588a94e82f1ee1b9dfcfdc453267c4fb652ff531570", + "sha256:ded0c4a3eee56b57fcb2315e40812b173cafe79d2f992d50015f4387445737fa", + "sha256:e1dae12321b31059a1a72aaa0e6ba30156fe7e633355e445451e4021b8e122b6", + "sha256:eb672217f7bd640411cfc69756ce721d00ae600814708d35c930930f18e8029f", + "sha256:ee684f139c91e69fe09b8e83d18b4d63bf87d9440c1eb2eeb52ee851883b1b29", + "sha256:f3f9a91d3cc5e5b0ddf1043c0ae5fa4852f18a1c0050318baf5fc7930ecc1f9c" + ], + "version": "==2021.10.23" }, "requests": { "hashes": [ - "sha256:27973dd4a904a4f13b263a19c866c13b92a39ed1c964655f025f3f8d3d75b804", - "sha256:c210084e36a42ae6b9219e00e48287def368a26d03a048ddad7bfee44f75871e" + "sha256:6c1246513ecd5ecd4528a0906f910e8f0f9c6b8ec72030dc9fd154dc1a6efd24", + "sha256:b8aa58f8cf793ffd8782d3d8cb19e66ef36f7aba4353eec859e74678b01b07a7" ], - "markers": "python_version >= '2.7' and python_version not in '3.0, 3.1, 3.2, 3.3, 3.4'", - "version": "==2.25.1" + "markers": "python_version >= '2.7' and python_version not in '3.0, 3.1, 3.2, 3.3, 3.4, 3.5'", + "version": "==2.26.0" }, "six": { "hashes": [ @@ -621,19 +707,19 @@ }, "sphinx": { "hashes": [ - "sha256:b5c2ae4120bf00c799ba9b3699bc895816d272d120080fbc967292f29b52b48c", - "sha256:d1cb10bee9c4231f1700ec2e24a91be3f3a3aba066ea4ca9f3bbe47e59d5a1d4" + "sha256:94078db9184491e15bce0a56d9186e0aec95f16ac20b12d00e06d4e36f1058a6", + "sha256:98a535c62a4fcfcc362528592f69b26f7caec587d32cd55688db580be0287ae0" ], "index": "pypi", - "version": "==4.0.2" + "version": "==4.2.0" }, "sphinx-rtd-theme": { "hashes": [ - "sha256:32bd3b5d13dc8186d7a42fc816a23d32e83a4827d7d9882948e7b837c232da5a", - "sha256:4a05bdbe8b1446d77a01e20a23ebc6777c74f43237035e76be89699308987d6f" + "sha256:4d35a56f4508cfee4c4fb604373ede6feae2a306731d533f409ef5c3496fdbd8", + "sha256:eec6d497e4c2195fa0e8b2016b337532b8a699a68bcb22a512870e16925c6a5c" ], "index": "pypi", - "version": "==0.5.2" + "version": "==1.0.0" }, "sphinxcontrib-applehelp": { "hashes": [ @@ -691,6 +777,13 @@ "markers": "python_version >= '2.6' and python_version not in '3.0, 3.1, 3.2, 3.3'", "version": "==0.10.2" }, + "tomli": { + "hashes": [ + "sha256:c6ce0015eb38820eaf32b5db832dbc26deb3dd427bd5f6556cf0acac2c214fee", + "sha256:f04066f68f5554911363063a30b108d2b5a5b1a010aa8b6132af78489fe3aade" + ], + "version": "==1.2.2" + }, "typed-ast": { "hashes": [ "sha256:01ae5f73431d21eead5015997ab41afa53aa1fbe252f9da060be5dad2c730ace", @@ -724,55 +817,62 @@ "sha256:f8afcf15cc511ada719a88e013cec87c11aff7b91f019295eb4530f96fe5ef2f", "sha256:fb1bbeac803adea29cedd70781399c99138358c26d05fcbd23c13016b7f5ec65" ], + "markers": "python_version < '3.8'", "version": "==1.4.3" }, "typing-extensions": { "hashes": [ - "sha256:0ac0f89795dd19de6b97debb0c6af1c70987fd80a2d62d1958f7e56fcc31b497", - "sha256:50b6f157849174217d0656f99dc82fe932884fb250826c18350e159ec6cdf342", - "sha256:779383f6086d90c99ae41cf0ff39aac8a7937a9283ce0a414e5dd782f4c94a84" + "sha256:49f75d16ff11f1cd258e1b988ccff82a3ca5570217d7ad8c5f48205dd99a677e", + "sha256:d8226d10bc02a29bcc81df19a26e56a9647f8b0a6d4a83924139f4a8b01f17b7", + "sha256:f1d25edafde516b146ecd0613dabcc61409817af4766fbbcfb8d1ad4ec441a34" ], "markers": "python_version < '3.8'", - "version": "==3.10.0.0" + "version": "==3.10.0.2" }, "urllib3": { "hashes": [ - "sha256:753a0374df26658f99d826cfe40394a686d05985786d946fbe4165b5148f5a7c", - "sha256:a7acd0977125325f516bda9735fa7142b909a8d01e8b2e4c8108d0984e6e0098" + "sha256:4987c65554f7a2dbf30c18fd48778ef124af6fab771a377103da0585e2336ece", + "sha256:c4fdf4019605b6e5423637e01bc9fe4daef873709a7973e195ceba0a62bbc844" ], "markers": "python_version >= '2.7' and python_version not in '3.0, 3.1, 3.2, 3.3, 3.4' and python_version < '4'", - "version": "==1.26.5" + "version": "==1.26.7" }, "watchdog": { "hashes": [ - "sha256:0237db4d9024859bea27d0efb59fe75eef290833fd988b8ead7a879b0308c2db", - "sha256:104266a778906ae0e971368d368a65c4cd032a490a9fca5ba0b78c6c7ae11720", - "sha256:188145185c08c73c56f1478ccf1f0f0f85101191439679b35b6b100886ce0b39", - "sha256:1a62a4671796dc93d1a7262286217d9e75823c63d4c42782912d39a506d30046", - "sha256:255a32d44bbbe62e52874ff755e2eefe271b150e0ec240ad7718a62a7a7a73c4", - "sha256:3d6405681471ebe0beb3aa083998c4870e48b57f8afdb45ea1b5957cc5cf1014", - "sha256:4b219d46d89cfa49af1d73175487c14a318a74cb8c5442603fd13c6a5b418c86", - "sha256:581e3548159fe7d2a9f377a1fbcb41bdcee46849cca8ab803c7ac2e5e04ec77c", - "sha256:58ebb1095ee493008a7789d47dd62e4999505d82be89fc884d473086fccc6ebd", - "sha256:598d772beeaf9c98d0df946fbabf0c8365dd95ea46a250c224c725fe0c4730bc", - "sha256:668391e6c32742d76e5be5db6bf95c455fa4b3d11e76a77c13b39bccb3a47a72", - "sha256:6ef9fe57162c4c361692620e1d9167574ba1975ee468b24051ca11c9bba6438e", - "sha256:91387ee2421f30b75f7ff632c9d48f76648e56bf346a7c805c0a34187a93aab4", - "sha256:a42e6d652f820b2b94cd03156c62559a2ea68d476476dfcd77d931e7f1012d4a", - "sha256:a6471517315a8541a943c00b45f1d252e36898a3ae963d2d52509b89a50cb2b9", - "sha256:d34ce2261f118ecd57eedeef95fc2a495fc4a40b3ed7b3bf0bd7a8ccc1ab4f8f", - "sha256:edcd9ef3fd460bb8a98eb1fcf99941e9fd9f275f45f1a82cb1359ec92975d647" + "sha256:25fb5240b195d17de949588628fdf93032ebf163524ef08933db0ea1f99bd685", + "sha256:3386b367e950a11b0568062b70cc026c6f645428a698d33d39e013aaeda4cc04", + "sha256:3becdb380d8916c873ad512f1701f8a92ce79ec6978ffde92919fd18d41da7fb", + "sha256:4ae38bf8ba6f39d5b83f78661273216e7db5b00f08be7592062cb1fc8b8ba542", + "sha256:8047da932432aa32c515ec1447ea79ce578d0559362ca3605f8e9568f844e3c6", + "sha256:8f1c00aa35f504197561060ca4c21d3cc079ba29cf6dd2fe61024c70160c990b", + "sha256:922a69fa533cb0c793b483becaaa0845f655151e7256ec73630a1b2e9ebcb660", + "sha256:9693f35162dc6208d10b10ddf0458cc09ad70c30ba689d9206e02cd836ce28a3", + "sha256:a0f1c7edf116a12f7245be06120b1852275f9506a7d90227648b250755a03923", + "sha256:a36e75df6c767cbf46f61a91c70b3ba71811dfa0aca4a324d9407a06a8b7a2e7", + "sha256:aba5c812f8ee8a3ff3be51887ca2d55fb8e268439ed44110d3846e4229eb0e8b", + "sha256:ad6f1796e37db2223d2a3f302f586f74c72c630b48a9872c1e7ae8e92e0ab669", + "sha256:ae67501c95606072aafa865b6ed47343ac6484472a2f95490ba151f6347acfc2", + "sha256:b2fcf9402fde2672545b139694284dc3b665fd1be660d73eca6805197ef776a3", + "sha256:b52b88021b9541a60531142b0a451baca08d28b74a723d0c99b13c8c8d48d604", + "sha256:b7d336912853d7b77f9b2c24eeed6a5065d0a0cc0d3b6a5a45ad6d1d05fb8cd8", + "sha256:bd9ba4f332cf57b2c1f698be0728c020399ef3040577cde2939f2e045b39c1e5", + "sha256:be9be735f827820a06340dff2ddea1fb7234561fa5e6300a62fe7f54d40546a0", + "sha256:cca7741c0fcc765568350cb139e92b7f9f3c9a08c4f32591d18ab0a6ac9e71b6", + "sha256:d0d19fb2441947b58fbf91336638c2b9f4cc98e05e1045404d7a4cb7cddc7a65", + "sha256:e02794ac791662a5eafc6ffeaf9bcc149035a0e48eb0a9d40a8feb4622605a3d", + "sha256:e0f30db709c939cabf64a6dc5babb276e6d823fd84464ab916f9b9ba5623ca15", + "sha256:e92c2d33858c8f560671b448205a268096e17870dcf60a9bb3ac7bfbafb7f5f9" ], "markers": "python_version >= '3.6'", - "version": "==2.1.2" + "version": "==2.1.6" }, "zipp": { "hashes": [ - "sha256:3607921face881ba3e026887d8150cca609d517579abe052ac81fc5aeffdbd76", - "sha256:51cb66cc54621609dd593d1787f286ee42a5c0adbb4b29abea5a63edc3e03098" + "sha256:71c644c5369f4a6e07636f0aa966270449561fcea2e3d6747b8d23efaa9d7832", + "sha256:9fe5ea21568a0a70e50f273397638d39b03353731e6cbbb3fd8502a33fec40bc" ], "markers": "python_version >= '3.6'", - "version": "==3.4.1" + "version": "==3.6.0" } } } diff --git a/README.rst b/README.rst index 7b18ec0..2ec4c67 100644 --- a/README.rst +++ b/README.rst @@ -36,6 +36,11 @@ Data Cleaning - Deduplicate alleles on MT - Assign PAR SNPs to the X or Y chromosome +Analysis +```````` +- Derive sex from SNPs +- Predict ancestry from SNPs (when installed with `ezancestry <https://github.com/arvkevi/ezancestry>`_) + Supported Genotype Files ------------------------ ``snps`` supports `VCF <https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3137218/>`_ files and @@ -72,6 +77,11 @@ Python dependencies) via ``pip``:: $ pip install snps +For `ancestry prediction <https://snps.readthedocs.io/en/stable/snps.html#snps.snps.SNPs.predict_ancestry>`_ +capability, ``snps`` can be installed with `ezancestry <https://github.com/arvkevi/ezancestry>`_:: + + $ pip install snps[ezancestry] + Examples -------- Download Example Data diff --git a/setup.py b/setup.py index b1d49d8..dd5d674 100644 --- a/setup.py +++ b/setup.py @@ -119,6 +119,7 @@ }, keywords="snps dna chromosomes bioinformatics vcf", install_requires=["numpy", "pandas!=1.0.0,!=1.1.0", "atomicwrites"], + extras_require={"ezancestry": ["ezancestry"]}, python_requires=">=3.7.1", platforms=["any"], ) diff --git a/src/snps/io/reader.py b/src/snps/io/reader.py index a43d64e..c869439 100644 --- a/src/snps/io/reader.py +++ b/src/snps/io/reader.py @@ -68,7 +68,7 @@ def get_empty_snps_dataframe(): - """ Get empty dataframe normalized for usage with ``snps``. + """Get empty dataframe normalized for usage with ``snps``. Returns ------- @@ -81,10 +81,10 @@ def get_empty_snps_dataframe(): class Reader: - """ Class for reading and parsing raw data / genotype files. """ + """Class for reading and parsing raw data / genotype files.""" def __init__(self, file="", only_detect_source=False, resources=None, rsids=()): - """ Initialize a `Reader`. + """Initialize a `Reader`. Parameters ---------- @@ -104,7 +104,7 @@ def __init__(self, file="", only_detect_source=False, resources=None, rsids=()): self._rsids = frozenset(rsids) def read(self): - """ Read and parse a raw data / genotype file. + """Read and parse a raw data / genotype file. Returns ------- @@ -195,7 +195,7 @@ def read(self): @classmethod def read_file(cls, file, only_detect_source, resources, rsids): - """ Read `file`. + """Read `file`. Parameters ---------- @@ -362,12 +362,12 @@ def _handle_bytes_data(self, file, include_data=False): @staticmethod def is_zip(bytes_data): - """ Check whether or not a bytes_data file is a valid Zip file.""" + """Check whether or not a bytes_data file is a valid Zip file.""" return zipfile.is_zipfile(io.BytesIO(bytes_data)) @staticmethod def is_gzip(bytes_data): - """ Check whether or not a bytes_data file is a valid gzip file.""" + """Check whether or not a bytes_data file is a valid gzip file.""" return binascii.hexlify(bytes_data[:2]) == b"1f8b" @staticmethod @@ -379,7 +379,7 @@ def _read_line(f, decode): return f.readline() def read_helper(self, source, parser): - """ Generic method to help read files. + """Generic method to help read files. Parameters ---------- @@ -431,7 +431,7 @@ def read_helper(self, source, parser): return {"snps": df, "source": source, "phased": phased, "build": build} def read_23andme(self, file, compression): - """ Read and parse 23andMe file. + """Read and parse 23andMe file. https://www.23andme.com @@ -515,7 +515,7 @@ def parser(): return self.read_helper("23andMe", parser) def read_ftdna(self, file, compression): - """ Read and parse Family Tree DNA (FTDNA) file. + """Read and parse Family Tree DNA (FTDNA) file. https://www.familytreedna.com @@ -602,7 +602,7 @@ def parser(): return self.read_helper("FTDNA", parser) def read_ftdna_famfinder(self, file, compression): - """ Read and parse Family Tree DNA (FTDNA) "famfinder" file. + """Read and parse Family Tree DNA (FTDNA) "famfinder" file. https://www.familytreedna.com @@ -641,7 +641,7 @@ def parser(): return self.read_helper("FTDNA", parser) def read_ancestry(self, file, compression): - """ Read and parse Ancestry.com file. + """Read and parse Ancestry.com file. http://www.ancestry.com @@ -690,7 +690,7 @@ def parser(): return self.read_helper("AncestryDNA", parser) def read_myheritage(self, file, compression): - """ Read and parse MyHeritage file. + """Read and parse MyHeritage file. https://www.myheritage.com @@ -745,7 +745,7 @@ def parser(): return self.read_helper("MyHeritage", parser) def read_livingdna(self, file, compression): - """ Read and parse LivingDNA file. + """Read and parse LivingDNA file. https://livingdna.com/ @@ -777,7 +777,7 @@ def parser(): return self.read_helper("LivingDNA", parser) def read_mapmygenome(self, file, compression, header): - """ Read and parse Mapmygenome file. + """Read and parse Mapmygenome file. https://mapmygenome.in @@ -827,7 +827,7 @@ def parse(rsid_col_name, rsid_col_idx): return self.read_helper("Mapmygenome", parser) def read_genes_for_good(self, file, compression): - """ Read and parse Genes For Good file. + """Read and parse Genes For Good file. https://genesforgood.sph.umich.edu/readme/readme1.2.txt @@ -1008,7 +1008,7 @@ def parser(): return self.read_helper(source, parser) def read_tellmegen(self, file, compression): - """ Read and parse tellmeGen files. + """Read and parse tellmeGen files. https://www.tellmegen.com/ @@ -1058,7 +1058,7 @@ def parser(): return self.read_helper("tellmeGen", parser) def read_gsa(self, data_or_filename, compresion, comments): - """ Read and parse Illumina Global Screening Array files + """Read and parse Illumina Global Screening Array files Parameters @@ -1085,7 +1085,7 @@ def read_gsa(self, data_or_filename, compresion, comments): return self._read_gsa_helper(data_or_filename, source) def read_dnaland(self, file, compression): - """ Read and parse DNA.land files. + """Read and parse DNA.land files. https://dna.land/ @@ -1117,7 +1117,7 @@ def parser(): return self.read_helper("DNA.Land", parser) def read_snps_csv(self, file, comments, compression): - """ Read and parse CSV file generated by ``snps``. + """Read and parse CSV file generated by ``snps``. https://pypi.org/project/snps/ @@ -1178,7 +1178,7 @@ def parse_csv(sep): return self.read_helper(source, parser) def read_generic(self, file, compression, skip=1): - """ Read and parse generic CSV or TSV file. + """Read and parse generic CSV or TSV file. Notes ----- @@ -1244,7 +1244,7 @@ def parse(sep): return self.read_helper("generic", parser) def read_vcf(self, file, compression, provider, rsids=()): - """ Read and parse VCF file. + """Read and parse VCF file. Notes ----- diff --git a/src/snps/io/writer.py b/src/snps/io/writer.py index f4a2fcf..7975899 100644 --- a/src/snps/io/writer.py +++ b/src/snps/io/writer.py @@ -48,10 +48,10 @@ class Writer: - """ Class for writing SNPs to files. """ + """Class for writing SNPs to files.""" def __init__(self, snps=None, filename="", vcf=False, atomic=True, **kwargs): - """ Initialize a `Writer`. + """Initialize a `Writer`. Parameters ---------- @@ -80,7 +80,7 @@ def write(self): @classmethod def write_file(cls, snps=None, filename="", vcf=False, atomic=True, **kwargs): - """ Save SNPs to file. + """Save SNPs to file. Parameters ---------- @@ -106,7 +106,7 @@ def write_file(cls, snps=None, filename="", vcf=False, atomic=True, **kwargs): return w.write() def _write_csv(self): - """ Write SNPs to a CSV file. + """Write SNPs to a CSV file. Returns ------- @@ -147,7 +147,7 @@ def _write_csv(self): ) def _write_vcf(self): - """ Write SNPs to a VCF file. + """Write SNPs to a VCF file. References ---------- diff --git a/src/snps/resources.py b/src/snps/resources.py index a040003..d7ee9a1 100644 --- a/src/snps/resources.py +++ b/src/snps/resources.py @@ -70,10 +70,10 @@ class Resources(metaclass=Singleton): - """ Object used to manage resources required by `snps`. """ + """Object used to manage resources required by `snps`.""" def __init__(self, resources_dir="resources"): - """ Initialize a ``Resources`` object. + """Initialize a ``Resources`` object. Parameters ---------- @@ -122,7 +122,7 @@ def get_reference_sequences( "MT", ), ): - """ Get Homo sapiens reference sequences for `chroms` of `assembly`. + """Get Homo sapiens reference sequences for `chroms` of `assembly`. Notes ----- @@ -163,7 +163,7 @@ def _reference_chroms_available(self, assembly, chroms): return False def get_assembly_mapping_data(self, source_assembly, target_assembly): - """ Get assembly mapping data. + """Get assembly mapping data. Parameters ---------- @@ -182,7 +182,7 @@ def get_assembly_mapping_data(self, source_assembly, target_assembly): ) def download_example_datasets(self): - """ Download example datasets from `openSNP <https://opensnp.org>`_. + """Download example datasets from `openSNP <https://opensnp.org>`_. Per openSNP, "the data is donated into the public domain using `CC0 1.0 <http://creativecommons.org/publicdomain/zero/1.0/>`_." @@ -217,7 +217,7 @@ def download_example_datasets(self): return paths def get_all_resources(self): - """ Get / download all resources used throughout `snps`. + """Get / download all resources used throughout `snps`. Notes ----- @@ -238,7 +238,7 @@ def get_all_resources(self): return resources def get_all_reference_sequences(self, **kwargs): - """ Get Homo sapiens reference sequences for Builds 36, 37, and 38 from Ensembl. + """Get Homo sapiens reference sequences for Builds 36, 37, and 38 from Ensembl. Notes ----- @@ -254,7 +254,7 @@ def get_all_reference_sequences(self, **kwargs): return self._reference_sequences def get_gsa_resources(self): - """ Get resources for reading Global Screening Array files. + """Get resources for reading Global Screening Array files. https://support.illumina.com/downloads/infinium-global-screening-array-v2-0-product-files.html @@ -270,7 +270,7 @@ def get_gsa_resources(self): } def get_dbsnp_151_37_reverse(self): - """ Get and load RSIDs that are on the reference reverse (-) strand in dbSNP 151 and lower. + """Get and load RSIDs that are on the reference reverse (-) strand in dbSNP 151 and lower. Returns ------- @@ -321,7 +321,7 @@ def get_dbsnp_151_37_reverse(self): return self._dbsnp_151_37_reverse def get_opensnp_datadump_filenames(self): - """ Get filenames internal to the `openSNP <https://opensnp.org>`_ datadump zip. + """Get filenames internal to the `openSNP <https://opensnp.org>`_ datadump zip. Per openSNP, "the data is donated into the public domain using `CC0 1.0 <http://creativecommons.org/publicdomain/zero/1.0/>`_." @@ -350,7 +350,7 @@ def get_opensnp_datadump_filenames(self): return self._opensnp_datadump_filenames def load_opensnp_datadump_file(self, filename): - """ Load the specified file from the openSNP datadump. + """Load the specified file from the openSNP datadump. Per openSNP, "the data is donated into the public domain using `CC0 1.0 <http://creativecommons.org/publicdomain/zero/1.0/>`_." @@ -380,7 +380,7 @@ def load_opensnp_datadump_file(self, filename): @staticmethod def _get_opensnp_datadump_filenames(filename): - """ Get list of filenames internal to the openSNP datadump zip. + """Get list of filenames internal to the openSNP datadump zip. Parameters ---------- @@ -400,7 +400,7 @@ def _get_opensnp_datadump_filenames(filename): @staticmethod def _write_data_to_gzip(f, data): - """ Write `data` to `f` in `gzip` format. + """Write `data` to `f` in `gzip` format. Parameters ---------- @@ -412,7 +412,7 @@ def _write_data_to_gzip(f, data): @staticmethod def _load_assembly_mapping_data(filename): - """ Load assembly mapping data. + """Load assembly mapping data. Parameters ---------- @@ -446,7 +446,7 @@ def _load_assembly_mapping_data(filename): def _get_paths_reference_sequences( self, sub_dir="fasta", assembly="GRCh37", chroms=() ): - """ Get local paths to Homo sapiens reference sequences from Ensembl. + """Get local paths to Homo sapiens reference sequences from Ensembl. Notes ----- @@ -550,7 +550,7 @@ def _create_reference_sequences(self, assembly, chroms, urls, paths): def _get_path_assembly_mapping_data( self, source_assembly, target_assembly, retries=10 ): - """ Get local path to assembly mapping data, downloading if necessary. + """Get local path to assembly mapping data, downloading if necessary. Parameters ---------- @@ -630,7 +630,7 @@ def _download_assembly_mapping_data( os.remove(f_tmp.name) def get_gsa_rsid(self): - """ Get and load GSA RSID map. + """Get and load GSA RSID map. https://support.illumina.com/downloads/infinium-global-screening-array-v2-0-product-files.html @@ -657,7 +657,7 @@ def get_gsa_rsid(self): return self._gsa_rsid_map def get_gsa_chrpos(self): - """ Get and load GSA chromosome position map. + """Get and load GSA chromosome position map. https://support.illumina.com/downloads/infinium-global-screening-array-v2-0-product-files.html @@ -697,7 +697,7 @@ def _get_path_opensnp_datadump(self): ) def _download_file(self, url, filename, compress=False, timeout=30): - """ Download a file to the resources folder. + """Download a file to the resources folder. Download data from `url`, save as `filename`, and optionally compress with gzip. @@ -765,7 +765,7 @@ def _download_file(self, url, filename, compress=False, timeout=30): @staticmethod def _print_download_msg(path): - """ Print download message. + """Print download message. Parameters ---------- @@ -776,10 +776,10 @@ def _print_download_msg(path): class ReferenceSequence: - """ Object used to represent and interact with a reference sequence. """ + """Object used to represent and interact with a reference sequence.""" def __init__(self, ID="", url="", path="", assembly="", species="", taxonomy=""): - """ Initialize a ``ReferenceSequence`` object. + """Initialize a ``ReferenceSequence`` object. Parameters ---------- @@ -818,7 +818,7 @@ def __repr__(self): @property def ID(self): - """ Get reference sequence chromosome. + """Get reference sequence chromosome. Returns ------- @@ -828,7 +828,7 @@ def ID(self): @property def chrom(self): - """ Get reference sequence chromosome. + """Get reference sequence chromosome. Returns ------- @@ -838,7 +838,7 @@ def chrom(self): @property def url(self): - """ Get URL to Ensembl reference sequence. + """Get URL to Ensembl reference sequence. Returns ------- @@ -848,7 +848,7 @@ def url(self): @property def path(self): - """ Get path to local reference sequence. + """Get path to local reference sequence. Returns ------- @@ -858,7 +858,7 @@ def path(self): @property def assembly(self): - """ Get reference sequence assembly. + """Get reference sequence assembly. Returns ------- @@ -868,7 +868,7 @@ def assembly(self): @property def build(self): - """ Get reference sequence build. + """Get reference sequence build. Returns ------- @@ -879,7 +879,7 @@ def build(self): @property def species(self): - """ Get reference sequence species. + """Get reference sequence species. Returns ------- @@ -889,7 +889,7 @@ def species(self): @property def taxonomy(self): - """ Get reference sequence taxonomy. + """Get reference sequence taxonomy. Returns ------- @@ -899,7 +899,7 @@ def taxonomy(self): @property def sequence(self): - """ Get reference sequence. + """Get reference sequence. Returns ------- @@ -910,7 +910,7 @@ def sequence(self): @property def md5(self): - """ Get reference sequence MD5 hash. + """Get reference sequence MD5 hash. Returns ------- @@ -921,7 +921,7 @@ def md5(self): @property def start(self): - """ Get reference sequence start position (1-based). + """Get reference sequence start position (1-based). Returns ------- @@ -932,7 +932,7 @@ def start(self): @property def end(self): - """ Get reference sequence end position (1-based). + """Get reference sequence end position (1-based). Returns ------- @@ -943,7 +943,7 @@ def end(self): @property def length(self): - """ Get reference sequence length. + """Get reference sequence length. Returns ------- @@ -953,7 +953,7 @@ def length(self): return self._sequence.size def clear(self): - """ Clear reference sequence. """ + """Clear reference sequence.""" self._sequence = np.array([], dtype=np.uint8) self._md5 = "" self._start = 0 diff --git a/src/snps/snps.py b/src/snps/snps.py index fa22528..2e73e2b 100644 --- a/src/snps/snps.py +++ b/src/snps/snps.py @@ -68,7 +68,7 @@ def __init__( processes=os.cpu_count(), rsids=(), ): - """ Object used to read, write, and remap genotype / raw data files. + """Object used to read, write, and remap genotype / raw data files. Parameters ---------- @@ -163,6 +163,7 @@ def __init__( if deduplicate_MT_chrom: self._deduplicate_MT_chrom() + else: logger.warning("no SNPs loaded...") @@ -177,7 +178,7 @@ def __repr__(self): @property def source(self): - """ Summary of the SNP data source(s). + """Summary of the SNP data source(s). Returns ------- @@ -188,7 +189,7 @@ def source(self): @property def snps(self): - """ Normalized SNPs. + """Normalized SNPs. Notes ----- @@ -219,7 +220,7 @@ def snps(self): @property def duplicate(self): - """ Duplicate SNPs. + """Duplicate SNPs. A duplicate SNP has the same RSID as another SNP. The first occurrence of the RSID is not considered a duplicate SNP. @@ -233,7 +234,7 @@ def duplicate(self): @property def discrepant_XY(self): - """ Discrepant XY SNPs. + """Discrepant XY SNPs. A discrepant XY SNP is a heterozygous SNP in the non-PAR region of the X or Y chromosome found during deduplication for a detected male genotype. @@ -247,7 +248,7 @@ def discrepant_XY(self): @property def heterozygous_MT(self): - """ Heterozygous SNPs on the MT chromosome found during deduplication. + """Heterozygous SNPs on the MT chromosome found during deduplication. Returns ------- @@ -258,7 +259,7 @@ def heterozygous_MT(self): @property def discrepant_vcf_position(self): - """ SNPs with discrepant positions discovered while saving VCF. + """SNPs with discrepant positions discovered while saving VCF. Returns ------- @@ -269,7 +270,7 @@ def discrepant_vcf_position(self): @property def discrepant_merge_positions(self): - """ SNPs with discrepant positions discovered while merging SNPs. + """SNPs with discrepant positions discovered while merging SNPs. Notes ----- @@ -295,7 +296,7 @@ def discrepant_merge_positions(self): @property def discrepant_merge_genotypes(self): - """ SNPs with discrepant genotypes discovered while merging SNPs. + """SNPs with discrepant genotypes discovered while merging SNPs. Notes ----- @@ -322,7 +323,7 @@ def discrepant_merge_genotypes(self): @property def discrepant_merge_positions_genotypes(self): - """ SNPs with discrepant positions and / or genotypes discovered while merging SNPs. + """SNPs with discrepant positions and / or genotypes discovered while merging SNPs. Notes ----- @@ -351,7 +352,7 @@ def discrepant_merge_positions_genotypes(self): @property def build(self): - """ Build of SNPs. + """Build of SNPs. Returns ------- @@ -361,7 +362,7 @@ def build(self): @property def build_detected(self): - """ Status indicating if build of SNPs was detected. + """Status indicating if build of SNPs was detected. Returns ------- @@ -371,7 +372,7 @@ def build_detected(self): @property def assembly(self): - """ Assembly of SNPs. + """Assembly of SNPs. Returns ------- @@ -388,7 +389,7 @@ def assembly(self): @property def count(self): - """ Count of SNPs. + """Count of SNPs. Returns ------- @@ -398,7 +399,7 @@ def count(self): @property def chromosomes(self): - """ Chromosomes of SNPs. + """Chromosomes of SNPs. Returns ------- @@ -412,7 +413,7 @@ def chromosomes(self): @property def chromosomes_summary(self): - """ Summary of the chromosomes of SNPs. + """Summary of the chromosomes of SNPs. Returns ------- @@ -449,7 +450,7 @@ def as_range(iterable): @property def sex(self): - """ Sex derived from SNPs. + """Sex derived from SNPs. Returns ------- @@ -463,7 +464,7 @@ def sex(self): @property def unannotated_vcf(self): - """ Indicates if VCF file is unannotated. + """Indicates if VCF file is unannotated. Returns ------- @@ -476,7 +477,7 @@ def unannotated_vcf(self): @property def phased(self): - """ Indicates if genotype is phased. + """Indicates if genotype is phased. Returns ------- @@ -485,7 +486,7 @@ def phased(self): return self._phased def heterozygous(self, chrom=""): - """ Get heterozygous SNPs. + """Get heterozygous SNPs. Parameters ---------- @@ -506,7 +507,7 @@ def heterozygous(self, chrom=""): ] def homozygous(self, chrom=""): - """ Get homozygous SNPs. + """Get homozygous SNPs. Parameters ---------- @@ -527,7 +528,7 @@ def homozygous(self, chrom=""): ] def notnull(self, chrom=""): - """ Get not null genotype SNPs. + """Get not null genotype SNPs. Parameters ---------- @@ -545,7 +546,7 @@ def notnull(self, chrom=""): @property def summary(self): - """ Summary of SNPs. + """Summary of SNPs. Returns ------- @@ -567,7 +568,7 @@ def summary(self): @property def valid(self): - """ Determine if ``SNPs`` is valid. + """Determine if ``SNPs`` is valid. ``SNPs`` is valid when the input file has been successfully parsed. @@ -582,7 +583,7 @@ def valid(self): return True def save(self, filename="", vcf=False, atomic=True, **kwargs): - """ Save SNPs to file. + """Save SNPs to file. Parameters ---------- @@ -628,7 +629,7 @@ def _read_raw_data(self, file, only_detect_source, rsids): return Reader.read_file(file, only_detect_source, self._resources, rsids) def _assign_par_snps(self): - """ Assign PAR SNPs to the X or Y chromosome using SNP position. + """Assign PAR SNPs to the X or Y chromosome using SNP position. References ----- @@ -702,7 +703,7 @@ def _extract_build(self, item): return int(assembly_name[-2:]) def detect_build(self): - """ Detect build of SNPs. + """Detect build of SNPs. Use the coordinates of common SNPs to identify the build / assembly of a genotype file that is being loaded. @@ -798,7 +799,7 @@ def lookup_build_with_snp_pos(pos, s): return build def get_count(self, chrom=""): - """ Count of SNPs. + """Count of SNPs. Parameters ---------- @@ -817,7 +818,7 @@ def determine_sex( y_snps_not_null_threshold=0.3, chrom="X", ): - """ Determine sex from SNPs using thresholds. + """Determine sex from SNPs using thresholds. Parameters ---------- @@ -907,7 +908,7 @@ def _deduplicate_alleles(self, rsids): ) def _deduplicate_sex_chrom(self, chrom): - """ Deduplicate a chromosome in the non-PAR region. """ + """Deduplicate a chromosome in the non-PAR region.""" discrepant_XY_snps = self._get_non_par_snps(chrom) @@ -925,12 +926,12 @@ def _deduplicate_sex_chrom(self, chrom): self._deduplicate_alleles(non_par_snps) def _deduplicate_XY_chrom(self): - """ Fix chromosome issue where some data providers duplicate male X and Y chromosomes""" + """Fix chromosome issue where some data providers duplicate male X and Y chromosomes""" self._deduplicate_sex_chrom("X") self._deduplicate_sex_chrom("Y") def _deduplicate_MT_chrom(self): - """ Deduplicate MT chromosome. """ + """Deduplicate MT chromosome.""" heterozygous_MT_snps = self._snps.loc[self.heterozygous("MT").index].index # save heterozygous MT SNPs @@ -945,7 +946,7 @@ def _deduplicate_MT_chrom(self): @staticmethod def get_par_regions(build): - """ Get PAR regions for the X and Y chromosomes. + """Get PAR regions for the X and Y chromosomes. Parameters ---------- @@ -998,7 +999,7 @@ def get_par_regions(build): return pd.DataFrame() def sort(self): - """ Sort SNPs based on ordered chromosome list and position. """ + """Sort SNPs based on ordered chromosome list and position.""" sorted_list = sorted( (str(x) for x in self._snps["chrom"].unique()), key=self._natural_sort_key ) @@ -1027,7 +1028,7 @@ def sort(self): self._snps = snps def remap(self, target_assembly, complement_bases=True): - """ Remap SNP coordinates from one assembly to another. + """Remap SNP coordinates from one assembly to another. This method uses the assembly map endpoint of the Ensembl REST API service (via ``Resources``'s ``EnsemblRestClient``) to convert SNP coordinates / positions from one @@ -1146,7 +1147,7 @@ def remap(self, target_assembly, complement_bases=True): return chromosomes_remapped, chromosomes_not_remapped def _remapper(self, task): - """ Remap SNPs for a chromosome. + """Remap SNPs for a chromosome. Parameters ---------- @@ -1269,7 +1270,7 @@ def merge( remap=True, chrom="", ): - """ Merge other ``SNPs`` objects into this ``SNPs`` object. + """Merge other ``SNPs`` objects into this ``SNPs`` object. Parameters ---------- @@ -1611,3 +1612,110 @@ def is_valid(self): DeprecationWarning, ) return self.valid + + def predict_ancestry( + self, + output_directory=None, + write_predictions=False, + models_directory=None, + aisnps_directory=None, + n_components=None, + k=None, + thousand_genomes_directory=None, + samples_directory=None, + algorithm=None, + aisnps_set=None, + ): + """Predict genetic ancestry for SNPs. + + Predictions by `ezancestry <https://github.com/arvkevi/ezancestry>`_. + + Notes + ----- + Populations below are described `here <https://www.internationalgenome.org/faq/what-do-the-population-codes-mean/>`_. + + Parameters + ---------- + various : optional + See the available settings for `predict` at `ezancestry <https://github.com/arvkevi/ezancestry>`_. + + Returns + ------- + dict + dict with the following keys: + + `population_code` (str) + max predicted population for the sample + `population_description` (str) + descriptive name of the population + `population_percent` (float) + predicted probability for the max predicted population + `superpopulation_code` (str) + max predicted super population (continental) for the sample + `superpopulation_description` (str) + descriptive name of the super population + `superpopulation_percent` (float) + predicted probability for the max predicted super population + `ezancestry_df` (pandas.DataFrame) + pandas.DataFrame with the following columns: + + `component1`, `component2`, `component3` + The coordinates of the sample in the dimensionality-reduced component space. Can be + used as (x, y, z,) coordinates for plotting in a 3d scatter plot. + `predicted_population_population` + The max predicted population for the sample. + `ACB`, `ASW`, `BEB`, `CDX`, `CEU`, `CHB`, `CHS`, `CLM`, `ESN`, `FIN`, `GBR`, `GIH`, `GWD`, `IBS`, `ITU`, `JPT`, `KHV`, `LWK`, `MSL`, `MXL`, `PEL`, `PJL`, `PUR`, `STU`, `TSI`, `YRI` + Predicted probabilities for each of the populations. These sum to 1.0. + `predicted_population_superpopulation` + The max predicted super population (continental) for the sample. + `AFR`, `AMR`, `EAS`, `EUR`, `SAS` + Predicted probabilities for each of the super populations. These sum to 1.0. + `population_description`, `superpopulation_name` + Descriptive names of the population and super population. + + """ + if not self.valid: + return {} + + try: + from ezancestry.commands import predict + except ModuleNotFoundError: + raise ModuleNotFoundError( + "Ancestry prediction requires the ezancestry package; please install it using pip install ezancestry" + ) + + def max_pop(row): + popcode = row["predicted_population_population"] + popdesc = row["population_description"] + poppct = row[popcode] + superpopcode = row["predicted_population_superpopulation"] + superpopdesc = row["superpopulation_name"] + superpoppct = row[superpopcode] + + return { + "population_code": popcode, + "population_description": popdesc, + "population_percent": poppct, + "superpopulation_code": superpopcode, + "superpopulation_description": superpopdesc, + "superpopulation_percent": superpoppct, + } + + predictions = predict( + self.snps, + output_directory, + write_predictions, + models_directory, + aisnps_directory, + n_components, + k, + thousand_genomes_directory, + samples_directory, + algorithm, + aisnps_set, + ) + + d = dict(predictions.apply(max_pop, axis=1).iloc[0]) + d["ezancestry_df"] = predictions + + return d diff --git a/src/snps/utils.py b/src/snps/utils.py index a157da6..cd36801 100644 --- a/src/snps/utils.py +++ b/src/snps/utils.py @@ -55,7 +55,7 @@ class Parallelizer: def __init__(self, parallelize=False, processes=os.cpu_count()): - """ Initialize a `Parallelizer`. + """Initialize a `Parallelizer`. Parameters ---------- @@ -68,7 +68,7 @@ def __init__(self, parallelize=False, processes=os.cpu_count()): self._processes = processes def __call__(self, f, tasks): - """ Optionally parallelize execution of a function. + """Optionally parallelize execution of a function. Parameters ---------- @@ -100,7 +100,7 @@ def __call__(cls, *args, **kwargs): def create_dir(path): - """ Create directory specified by `path` if it doesn't already exist. + """Create directory specified by `path` if it doesn't already exist. Parameters ---------- @@ -124,7 +124,7 @@ def create_dir(path): def save_df_as_csv( df, path, filename, comment="", prepend_info=True, atomic=True, **kwargs ): - """ Save dataframe to a CSV file. + """Save dataframe to a CSV file. Parameters ---------- @@ -198,7 +198,7 @@ def save_df_as_csv( def clean_str(s): - """ Clean a string so that it can be used as a Python variable name. + """Clean a string so that it can be used as a Python variable name. Parameters ---------- @@ -216,7 +216,7 @@ def clean_str(s): def zip_file(src, dest, arcname): - """ Zip a file. + """Zip a file. Parameters ---------- @@ -240,7 +240,7 @@ def zip_file(src, dest, arcname): def gzip_file(src, dest): - """ Gzip a file. + """Gzip a file. Parameters ----------

Support current version of `black` Fixes #144.

2021-11-06T03:37:59

0.0

[]

apriha/snps

apriha__snps-137

28de7e13dc0f48ef3234d3be7be35d3ff54a6db3

diff --git a/.circleci/config.yml b/.circleci/config.yml new file mode 100644 index 00000000..32c52d1a --- /dev/null +++ b/.circleci/config.yml @@ -0,0 +1,54 @@ +# Python CircleCI 2.1 configuration file +# +# Check https://circleci.com/docs/2.0/language-python/ for more details +# +version: 2.1 +jobs: + unittests: + docker: + - image: circleci/python:3.8.3 + environment: + PIPENV_VENV_IN_PROJECT: "1" + + steps: + - checkout + # increment vX to manually purge cache + - restore_cache: + name: Restore dependencies from cache, if possible + keys: + - v1-venv-{{ checksum "Pipfile.lock" }} + - run: + name: Dependency installation + # pipenv is itself include in circleci docker image + # prefer the sync command to remove unused cached deps + command: pipenv sync --dev + - save_cache: + name: Save server dependencies + paths: + - .venv + key: v1-venv-{{ checksum "Pipfile.lock" }} + - run: + name: Dependency output for debugging + command: pipenv graph + #- run: + # name: Dependency security check + # command: pipenv check + - run: + name: run tests + command: | + mkdir -p server-test-reports/pytest + pipenv run pytest --cov=snps --cov-report xml:server-test-reports/coverage.xml --junitxml server-test-reports/pytest/tests.xml tests + bash <(curl -s https://codecov.io/bash) + - store_test_results: + # must be a directory with named subdirectories + path: server-test-reports + - store_artifacts: + path: server-test-reports + - run: + name: check linting + command: pipenv run black --check --diff . + +workflows: + test: + jobs: + - "unittests" \ No newline at end of file diff --git a/.gitignore b/.gitignore index 5e4fd224..4d9ef8d9 100644 --- a/.gitignore +++ b/.gitignore @@ -113,6 +113,7 @@ tests/output/ tests/resources/* !tests/resources/gsa_chrpos_map.txt !tests/resources/gsa_rsid_map.txt +!tests/resources/dbsnp_151_37_reverse.txt tests/input/23andme.txt.zip tests/input/discrepant_snps[12].csv tests/input/ftdna.csv.gz diff --git a/src/snps/io/reader.py b/src/snps/io/reader.py index 8581351c..a43d64ee 100644 --- a/src/snps/io/reader.py +++ b/src/snps/io/reader.py @@ -171,7 +171,7 @@ def read(self): d = self.read_mapmygenome(file, compression, first_line) elif "lineage" in first_line or "snps" in first_line: d = self.read_snps_csv(file, comments, compression) - elif "Chromosome" in first_line: + elif "rsid\tChromosome\tposition\tgenotype" == first_line.strip(): d = self.read_tellmegen(file, compression) elif re.match("^#*[ \t]*rsid[, \t]*chr", first_line): d = self.read_generic(file, compression) @@ -183,10 +183,9 @@ def read(self): d = self.read_genes_for_good(file, compression) elif "DNA.Land" in comments: d = self.read_dnaland(file, compression) - elif "CODIGO46" in comments: - d = self.read_codigo46(file) - elif "SANO" in comments: - d = self.read_sano(file) + elif first_line.startswith("[Header]"): + # Global Screening Array, includes SANO and CODIGO46 + d = self.read_gsa(file, compression, comments) # detect build from comments if build was not already detected from `read` method if not d["build"]: @@ -259,31 +258,71 @@ def _extract_comments(self, f, decode=False, include_data=False): return first_line, comments, data def _detect_build_from_comments(self, comments, source): - comments = comments.lower() - if "build 37" in comments: - return 37 - elif "build 36" in comments: - return 36 - - # allow more variations for VCF + # if its a VCF parse these properly if source == "vcf": - if "https://pypi.org/project/snps/" in comments: # remove `snps` version - comments = f"{comments[: comments.find('snps v')]}{comments[comments.find('https://pypi.org/project/snps/'):]}" - if "hg19" in comments: - return 37 - elif "ncbi36" in comments: + for line in comments.split("\n"): + line = line.strip() + if not line: + # skip blanks + continue + assert line.startswith("#"), line + if not line.startswith("##"): + # skip comments but not preamble + continue + if "=" not in line: + # skip lines without key/value in + continue + line = line[2:].strip() + key = line[: line.index("=")] + value = line[line.index("=") + 1 :] + if key.lower() == "contig": + assert value.startswith("<"), value + assert value.endswith(">"), value + parts = value[1:-1].split(",") + for part in parts: + part_key = part[: part.index("=")] + part_value = part[part.index("=") + 1 :] + if part_key.lower() == "assembly": + if "36" in part_value: + return 36 + elif "37" in part_value or "hg19" in part_value: + return 37 + elif "38" in part_value: + return 38 + elif part_key.lower() == "length": + if "249250621" == part_value: + return 37 # length of chromosome 1 + elif "248956422" == part_value: + return 38 # length of chromosome 1 + # couldn't find anything + return 0 + else: + # not a vcf + if "build 36" in comments.lower(): return 36 - elif "grch38" in comments: - return 38 - elif "build 38" in comments: - return 38 - elif "b37" in comments: + elif "build 37" in comments.lower(): return 37 - elif "hg38" in comments: + elif "build 38" in comments.lower(): return 38 - elif "b38" in comments: + # these can cause false positives + # elif "b37" in comments.lower(): + # return 37 + # elif "b38" in comments.lower(): + # return 38 + # elif "hg19" in comments.lower(): + # return 37 + # elif "hg38" in comments.lower(): + # return 38 + elif "grch38" in comments.lower(): return 38 - return 0 + elif "grch37" in comments.lower(): + return 37 + elif "249250621" in comments.lower(): + return 37 # length of chromosome 1 + elif "248956422" in comments.lower(): + return 38 # length of chromosome 1 + else: + return 0 def _handle_bytes_data(self, file, include_data=False): compression = "infer" @@ -323,12 +362,12 @@ def _handle_bytes_data(self, file, include_data=False): @staticmethod def is_zip(bytes_data): - """Check whether or not a bytes_data file is a valid Zip file.""" + """ Check whether or not a bytes_data file is a valid Zip file.""" return zipfile.is_zipfile(io.BytesIO(bytes_data)) @staticmethod def is_gzip(bytes_data): - """Check whether or not a bytes_data file is a valid gzip file.""" + """ Check whether or not a bytes_data file is a valid gzip file.""" return binascii.hexlify(bytes_data[:2]) == b"1f8b" @staticmethod @@ -476,7 +515,7 @@ def parser(): return self.read_helper("23andMe", parser) def read_ftdna(self, file, compression): - """Read and parse Family Tree DNA (FTDNA) file. + """ Read and parse Family Tree DNA (FTDNA) file. https://www.familytreedna.com @@ -623,7 +662,7 @@ def parser(): comment="#", header=0, engine="c", - sep="\s+", + sep=r"\s+", # delim_whitespace=True, # https://stackoverflow.com/a/15026839 na_values=0, names=["rsid", "chrom", "pos", "allele1", "allele2"], @@ -777,7 +816,8 @@ def parse(rsid_col_name, rsid_col_idx): else: df = parse("SNP.Name", 0) - df["genotype"] = df["Allele1...Top"] + df["Allele2...Top"] + # uses Illumina definition of "Plus" from https://emea.support.illumina.com/bulletins/2017/06/how-to-interpret-dna-strand-and-allele-information-for-infinium-.html + df["genotype"] = df["Allele1...Plus"] + df["Allele2...Plus"] df.rename(columns={"Chr": "chrom", "Position": "pos"}, inplace=True) df.index.name = "rsid" df = df[["chrom", "pos", "genotype"]] @@ -818,52 +858,149 @@ def parser(): return self.read_helper("GenesForGood", parser) - def _read_gsa_helper(self, file, source, strand, dtypes, na_values="--"): + def _read_gsa_helper(self, file, source): def parser(): - gsa_resources = self._resources.get_gsa_resources() + # read the comments so we get to the actual data if isinstance(file, str): try: with open(file, "rb") as f: - first_line, comments, data = self._extract_comments( + _, _, data = self._extract_comments( f, decode=True, include_data=True ) except UnicodeDecodeError: # compressed file on filesystem with open(file, "rb") as f: - ( - first_line, - comments, - data, - compression, - ) = self._handle_bytes_data(f.read(), include_data=True) + _, _, data, _ = self._handle_bytes_data( + f.read(), include_data=True + ) else: - first_line, comments, data, compression = self._handle_bytes_data( - file.read(), include_data=True + _, _, data, _ = self._handle_bytes_data(file.read(), include_data=True) + + # turn the data into a pandas dataframe for manipulation + df = pd.read_csv( + io.StringIO(data), + sep="\t", + engine="c", + dtype={ + "Position": NORMALIZED_DTYPES["pos"], + "Chr": NORMALIZED_DTYPES["chrom"], + }, + ) + + # reserve columns we want out + assert "rsid" not in df.columns + assert "chrom" not in df.columns + assert "pos" not in df.columns + assert "genotype" not in df.columns + + # prefer the specified chromosome and position, in prsent + # this uses the gsa names, so needs to be done before rsid + if "Chr" in df.columns and "Position" in df.columns: + # put the chromosome in the right column with the right type + df["chrom"] = df["Chr"] + # put the position in the right column with the right type + df["pos"] = df["Position"] + + else: + # use an external source to map snp names to chromosome and position + df = df.merge( + self._resources.get_gsa_chrpos(), + how="inner", # inner join, everything must have a position + left_on="SNP Name", + right_on="gsaname_chrpos", + suffixes=(None, "_gsa"), ) + # make sure its the right types + df["chrom"] = df["gsachr"].apply(str).astype(NORMALIZED_DTYPES["chrom"]) + df["pos"] = df["gsapos"].astype(NORMALIZED_DTYPES["pos"]) + + # use the given rsid when avaliable, SNP Name when unavaliable + df["rsid"] = df["SNP Name"].astype(NORMALIZED_DTYPES["rsid"]) + if "RsID" in df.columns: + df.loc[df["RsID"] != ".", "rsid"] = df.loc[df["rsid"] != ".", "RsID"] + else: + # if given RSIDs are not avaliable, then use the external mapping to turn + # SNP names into rsids where possible + df = df.merge( + self._resources.get_gsa_rsid(), + how="left", # left-hand join, gsa rsids may be NA + left_on="SNP Name", + right_on="gsaname_rsid", + suffixes=(None, "_gsa_rsid"), + ) + df.loc[~pd.isna(df["gsaname_rsid"]), "rsid"] = df.loc[ + ~pd.isna(df["gsaname_rsid"]), "gsarsid" + ] - data_io = io.StringIO(data) - df = pd.read_csv(data_io, sep="\t", dtype=dtypes, na_values=na_values) + # combine the alleles into genotype + # prefer Plus strand as that is forward reference + if "Allele1 - Plus" in df.columns and "Allele2 - Plus" in df.columns: + df["genotype"] = (df["Allele1 - Plus"] + df["Allele2 - Plus"]).astype( + NORMALIZED_DTYPES["genotype"] + ) + elif ( + "Allele1 - Forward" in df.columns and "Allele2 - Forward" in df.columns + ): + # if strand is forward, need to take reverse complement of *some* rsids + # this is because it is Illumina forward, which is dbSNP strand, which + # is reverse reference for some RSIDs before dbSNP 151. + + # load list of reversable rsids + dbsnp151 = self._resources.get_dbsnp_151_37_reverse() + # keep only the rsids + dbsnp151 = dbsnp151.filter(items=("dbsnp151revrsid",), axis=1) + + # add it as an extra column + df = df.merge( + dbsnp151, + how="left", + left_on="rsid", + right_on="dbsnp151revrsid", + suffixes=(None, "_dbsnp151rev"), + ) - def map_rsids(x): - return gsa_resources["rsid_map"].get(x) + # create plus strand columns from the forward alleles and flip them if appropriate + for i in (1, 2): + df[f"Allele{i} - Plus"] = df[f"Allele{i} - Forward"] + df.loc[ + (df[f"Allele{i} - Forward"] == "A") + & (~pd.isna(df["dbsnp151revrsid"])), + f"Allele{i} - Plus", + ] = "T" + df.loc[ + (df[f"Allele{i} - Forward"] == "T") + & (~pd.isna(df["dbsnp151revrsid"])), + f"Allele{i} - Plus", + ] = "A" + df.loc[ + (df[f"Allele{i} - Forward"] == "C") + & (~pd.isna(df["dbsnp151revrsid"])), + f"Allele{i} - Plus", + ] = "G" + df.loc[ + (df[f"Allele{i} - Forward"] == "G") + & (~pd.isna(df["dbsnp151revrsid"])), + f"Allele{i} - Plus", + ] = "C" + + # create a genotype by combining the new plus columns + df["genotype"] = (df["Allele1 - Plus"] + df["Allele2 - Plus"]).astype( + NORMALIZED_DTYPES["genotype"] + ) + else: + raise ValueError("No supported allele column") - def map_chr(x): - chrpos = gsa_resources["chrpos_map"].get(x) - return chrpos.split(":")[0] if chrpos else None + # mark -- genotype as na + df.loc[df["genotype"] == "--", "genotype"] = np.nan - def map_pos(x): - chrpos = gsa_resources["chrpos_map"].get(x) - return chrpos.split(":")[1] if chrpos else None + # keep only the columns we want + df = df.filter(items=("rsid", "chrom", "pos", "genotype"), axis=1) - df["rsid"] = df["SNP Name"].apply(map_rsids) - df["chrom"] = df["SNP Name"].apply(map_chr) - df["pos"] = df["SNP Name"].apply(map_pos) - df["genotype"] = df[f"Allele1 - {strand}"] + df[f"Allele2 - {strand}"] + # discard rows without values df.dropna(subset=["rsid", "chrom", "pos"], inplace=True) - df = df.astype(NORMALIZED_DTYPES) - df = df[["rsid", "chrom", "pos", "genotype"]] + # reindex for the new identifiers df.set_index(["rsid"], inplace=True) return (df,) @@ -887,57 +1024,65 @@ def read_tellmegen(self, file, compression): """ def parser(): - gsa_resources = self._resources.get_gsa_resources() - df = pd.read_csv( file, sep="\t", skiprows=1, na_values="--", names=["rsid", "chrom", "pos", "genotype"], - index_col=0, dtype=NORMALIZED_DTYPES, compression=compression, ) - df.rename(index=gsa_resources["rsid_map"], inplace=True) - return (df,) - return self.read_helper("tellmeGen", parser) + # use the external mapping to turn + # SNP names into rsids where possible + df = df.merge( + self._resources.get_gsa_rsid(), + how="left", # left-hand join, gsa rsids may be NA + left_on="rsid", + right_on="gsaname_rsid", + suffixes=(None, "_gsa_rsid"), + ) + df.loc[~pd.isna(df["gsaname_rsid"]), "rsid"] = df.loc[ + ~pd.isna(df["gsaname_rsid"]), "gsarsid" + ] - def read_codigo46(self, file): - """ Read and parse Codigo46 files. + # keep only the columns we want + df = df.filter(items=("rsid", "chrom", "pos", "genotype"), axis=1) - https://codigo46.com.mx + # reindex for the new identifiers + df.set_index(["rsid"], inplace=True) - Parameters - ---------- - data : str - data string + return (df,) - Returns - ------- - dict - result of `read_helper` - """ - return self._read_gsa_helper(file, "Codigo46", "Plus", {}) + return self.read_helper("tellmeGen", parser) - def read_sano(self, file): - """ Read and parse Sano Genetics files. + def read_gsa(self, data_or_filename, compresion, comments): + """ Read and parse Illumina Global Screening Array files - https://sanogenetics.com Parameters ---------- - data : str - data string + data_or_filename : str or bytes + either the filename to read from or the bytes data itself Returns ------- dict result of `read_helper` """ - dtype = {"Chr": object, "Position": np.uint32} - return self._read_gsa_helper(file, "Sano", "Forward", dtype, na_values="-") + + # pick the source + # ideally we want something more specific than GSA + if "SANO" in comments: + source = "Sano" + elif "CODIGO46" in comments: + source = "Codigo46" + else: + # default to generic global screening array + source = "GSA" + + return self._read_gsa_helper(data_or_filename, source) def read_dnaland(self, file, compression): """ Read and parse DNA.land files. diff --git a/src/snps/resources.py b/src/snps/resources.py index 42bb319b..a0400034 100644 --- a/src/snps/resources.py +++ b/src/snps/resources.py @@ -61,6 +61,7 @@ from atomicwrites import atomic_write import numpy as np +import pandas as pd from snps.ensembl import EnsemblRestClient from snps.utils import create_dir, Singleton @@ -81,8 +82,13 @@ def __init__(self, resources_dir="resources"): """ self._resources_dir = os.path.abspath(resources_dir) self._ensembl_rest_client = EnsemblRestClient() + self._init_resource_attributes() + + def _init_resource_attributes(self): self._reference_sequences = {} - self._gsa_resources = {} + self._gsa_rsid_map = None + self._gsa_chrpos_map = None + self._dbsnp_151_37_reverse = None self._opensnp_datadump_filenames = [] def get_reference_sequences( @@ -256,11 +262,63 @@ def get_gsa_resources(self): ------- dict """ - if not self._gsa_resources: - self._gsa_resources = self._load_gsa_resources( - self._get_path_gsa_rsid_map(), self._get_path_gsa_chrpos_map() + + return { + "rsid_map": self.get_gsa_rsid(), + "chrpos_map": self.get_gsa_chrpos(), + "dbsnp_151_37_reverse": self.get_dbsnp_151_37_reverse(), + } + + def get_dbsnp_151_37_reverse(self): + """ Get and load RSIDs that are on the reference reverse (-) strand in dbSNP 151 and lower. + + Returns + ------- + pandas.DataFrame + + References + ---------- + 1. Sherry ST, Ward MH, Kholodov M, Baker J, Phan L, Smigielski EM, Sirotkin K. + dbSNP: the NCBI database of genetic variation. Nucleic Acids Res. 2001 Jan 1; + 29(1):308-11. + 2. Database of Single Nucleotide Polymorphisms (dbSNP). Bethesda (MD): National Center + for Biotechnology Information, National Library of Medicine. (dbSNP Build ID: 151). + Available from: http://www.ncbi.nlm.nih.gov/SNP/ + """ + if self._dbsnp_151_37_reverse is None: + # download the file from the cloud, if not done already + dbsnp_rev_path = self._download_file( + "https://sano-public.s3.eu-west-2.amazonaws.com/dbsnp151.b37.snps_reverse.txt.gz", + "dbsnp_151_37_reverse.txt.gz", + ) + + # load into pandas + rsids = pd.read_csv( + dbsnp_rev_path, + sep=" ", + header=None, # dont infer header as there isn't one + names=( + "dbsnp151revrsid", + "dbsnp151freqa", + "dbsnp151freqt", + "dbsnp151freqc", + "dbsnp151freqg", + ), + dtype={ + "dbsnp151revrsid": "string", + "dbsnp151freqa": "double", + "dbsnp151freqt": "double", + "dbsnp151freqc": "double", + "dbsnp151freqg": "double", + }, + engine="c", # force c engine for performance + comment="#", # skip the first row ) - return self._gsa_resources + + # store in memory so we don't have to load again + self._dbsnp_151_37_reverse = rsids + + return self._dbsnp_151_37_reverse def get_opensnp_datadump_filenames(self): """ Get filenames internal to the `openSNP <https://opensnp.org>`_ datadump zip. @@ -571,38 +629,66 @@ def _download_assembly_mapping_data( # remove temp file os.remove(f_tmp.name) - def _load_gsa_resources(self, rsid_map, chrpos_map): - d = {} - - with gzip.open(rsid_map, "rb") as f: - gsa_rsid_map = f.read().decode("utf-8") + def get_gsa_rsid(self): + """ Get and load GSA RSID map. - d["rsid_map"] = dict( - (x.split("\t")[0].strip(), x.split("\t")[1].strip()) - for x in gsa_rsid_map.split("\n")[:-1] - ) + https://support.illumina.com/downloads/infinium-global-screening-array-v2-0-product-files.html - with gzip.open(chrpos_map, "rb") as f: - gsa_chrpos_map = f.read().decode("utf-8") + Returns + ------- + pandas.DataFrame + """ + if self._gsa_rsid_map is None: + # download the file from the cloud, if not done already + rsid_path = self._download_file( + "https://sano-public.s3.eu-west-2.amazonaws.com/gsa_rsid_map.txt.gz", + "gsa_rsid_map.txt.gz", + ) + # load into pandas + rsids = pd.read_csv( + rsid_path, + sep=r"\s+", # whitespace separators + header=0, # dont infer header as there isn't one + names=("gsaname_rsid", "gsarsid"), + dtype={"gsaname_rsid": "string", "gsarsid": "string"}, + engine="c", # force c engine for performance + ) + self._gsa_rsid_map = rsids + return self._gsa_rsid_map - d["chrpos_map"] = dict( - (x.split("\t")[0], x.split("\t")[1] + ":" + x.split("\t")[2]) - for x in gsa_chrpos_map.split("\n")[:-1] - ) + def get_gsa_chrpos(self): + """ Get and load GSA chromosome position map. - return d + https://support.illumina.com/downloads/infinium-global-screening-array-v2-0-product-files.html - def _get_path_gsa_rsid_map(self): - return self._download_file( - "https://sano-public.s3.eu-west-2.amazonaws.com/gsa_rsid_map.txt.gz", - "gsa_rsid_map.txt.gz", - ) + Returns + ------- + pandas.DataFrame + """ + if self._gsa_chrpos_map is None: + # download the file from the cloud, if not done already + chrpos_path = self._download_file( + "https://sano-public.s3.eu-west-2.amazonaws.com/gsa_chrpos_map.txt.gz", + "gsa_chrpos_map.txt.gz", + ) + # load into pandas + + chrpos = pd.read_csv( + chrpos_path, + sep=r"\s+", # whitespace separators + header=0, # dont infer header as there isn't one + names=("gsaname_chrpos", "gsachr", "gsapos", "gsacm"), + dtype={ + "gsaname_chrpos": "string", + "gsachr": "category", + "gsapos": "uint32", + "gsacm": "double", + }, + engine="c", # force c engine for performance + ) - def _get_path_gsa_chrpos_map(self): - return self._download_file( - "https://sano-public.s3.eu-west-2.amazonaws.com/gsa_chrpos_map.txt.gz", - "gsa_chrpos_map.txt.gz", - ) + self._gsa_chrpos_map = chrpos + return self._gsa_chrpos_map def _get_path_opensnp_datadump(self): return self._download_file( @@ -645,7 +731,9 @@ def _download_file(self, url, filename, compress=False, timeout=30): # http://stackoverflow.com/a/7244263 with urllib.request.urlopen( url, timeout=timeout - ) as response, atomic_write(destination, mode="wb") as f: + ) as response, atomic_write( + destination, mode="wb", overwrite=True + ) as f: self._print_download_msg(destination) data = response.read() # a `bytes` object @@ -667,6 +755,11 @@ def _download_file(self, url, filename, compress=False, timeout=30): except socket.timeout: logger.warning(f"Timeout downloading {url}") destination = "" + except FileExistsError: + # if the file exists, another process has created it while it was + # being downloaded + # in such a case, the other copy is identical, so ignore this error + pass return destination diff --git a/src/snps/snps.py b/src/snps/snps.py index bddcc43e..ab68670d 100644 --- a/src/snps/snps.py +++ b/src/snps/snps.py @@ -84,11 +84,11 @@ def __init__( name / path of resources directory deduplicate : bool deduplicate RSIDs and make SNPs available as `SNPs.duplicate` - deduplicate_XY_chrom : bool - deduplicate alleles in the non-PAR regions of X and Y for males; see - `SNPs.discrepant_XY` deduplicate_MT_chrom : bool deduplicate alleles on MT; see `SNPs.heterozygous_MT` + deduplicate_XY_chrom : bool or str + deduplicate alleles in the non-PAR regions of X and Y for males; see `SNPs.discrepant_XY` + if a `str` then this is the sex determination method to use X Y or XY parallelize : bool utilize multiprocessing to speedup calculations processes : int @@ -131,6 +131,8 @@ def __init__( d["source"].split(", ") if ", " in d["source"] else [d["source"]] ) self._phased = d["phased"] + self._build = d["build"] + self._build_detected = True if d["build"] else False if not self._snps.empty: self.sort() @@ -138,13 +140,11 @@ def __init__( if deduplicate: self._deduplicate_rsids() - # prefer to use SNP positions to detect build - self._build = self.detect_build() - self._build_detected = True if self._build else False - + # use build detected from `read` method or comments, if any + # otherwise use SNP positions to detect build if not self._build_detected: - # use build detected from `read` method or comments, if any - self._build = d["build"] + self._build = self.detect_build() + self._build_detected = True if self._build else False if not self._build: self._build = 37 # assume Build 37 / GRCh37 if not detected @@ -156,7 +156,9 @@ def __init__( self.sort() if deduplicate_XY_chrom: - if self.determine_sex() == "Male": + if ( + deduplicate_XY_chrom is True and self.determine_sex() == "Male" + ) or self.determine_sex(chrom=deduplicate_XY_chrom) == "Male": self._deduplicate_XY_chrom() if deduplicate_MT_chrom: @@ -1163,58 +1165,68 @@ def _remapper(self, task): pos_end = int(temp["pos"].describe()["max"]) for mapping in mappings["mappings"]: - # skip if mapping is outside of range of SNP positions - if ( - mapping["original"]["end"] < pos_start - or mapping["original"]["start"] > pos_end - ): - continue - orig_range_len = mapping["original"]["end"] - mapping["original"]["start"] - mapped_range_len = mapping["mapped"]["end"] - mapping["mapped"]["start"] + orig_start = mapping["original"]["start"] + orig_end = mapping["original"]["end"] + mapped_start = mapping["mapped"]["start"] + mapped_end = mapping["mapped"]["end"] orig_region = mapping["original"]["seq_region_name"] mapped_region = mapping["mapped"]["seq_region_name"] - if orig_region != mapped_region: - logger.warning("discrepant chroms") - continue - - if orig_range_len != mapped_range_len: - logger.warning( - "discrepant coords" - ) # observed when mapping NCBI36 -> GRCh38 + # skip if mapping is outside of range of SNP positions + if orig_end < pos_start or orig_start > pos_end: continue # find the SNPs that are being remapped for this mapping snp_indices = temp.loc[ ~temp["remapped"] - & (temp["pos"] >= mapping["original"]["start"]) - & (temp["pos"] <= mapping["original"]["end"]) + & (temp["pos"] >= orig_start) + & (temp["pos"] <= orig_end) ].index - if len(snp_indices) > 0: - # remap the SNPs - if mapping["mapped"]["strand"] == -1: - # flip and (optionally) complement since we're mapping to minus strand - diff_from_start = ( - temp.loc[snp_indices, "pos"] - mapping["original"]["start"] - ) - temp.loc[snp_indices, "pos"] = ( - mapping["mapped"]["end"] - diff_from_start - ) + # if there are no snp here, skip + if not len(snp_indices): + continue - if complement_bases: - temp.loc[snp_indices, "genotype"] = temp.loc[ - snp_indices, "genotype" - ].apply(self._complement_bases) - else: - # mapping is on same (plus) strand, so just remap based on offset - offset = mapping["mapped"]["start"] - mapping["original"]["start"] - temp.loc[snp_indices, "pos"] = temp["pos"] + offset + orig_range_len = orig_end - orig_start + mapped_range_len = mapped_end - mapped_start + + # if this would change chromosome, skip + # TODO allow within normal chromosomes + # TODO flatten patches + if orig_region != mapped_region: + logger.warning( + f"discrepant chroms for {len(snp_indices)} SNPs from {orig_region} to {mapped_region}" + ) + continue + + # if there is any stretching or squashing of the region + # observed when mapping NCBI36 -> GRCh38 + # TODO disallow skipping a version when remapping + if orig_range_len != mapped_range_len: + logger.warning( + f"discrepant coords for {len(snp_indices)} SNPs from {orig_region}:{orig_start}-{orig_end} to {mapped_region}:{mapped_start}-{mapped_end}" + ) + continue + + # remap the SNPs + if mapping["mapped"]["strand"] == -1: + # flip and (optionally) complement since we're mapping to minus strand + diff_from_start = temp.loc[snp_indices, "pos"] - orig_start + temp.loc[snp_indices, "pos"] = mapped_end - diff_from_start + + if complement_bases: + temp.loc[snp_indices, "genotype"] = temp.loc[ + snp_indices, "genotype" + ].apply(self._complement_bases) + else: + # mapping is on same (plus) strand, so just remap based on offset + offset = mapped_start - orig_start + temp.loc[snp_indices, "pos"] = temp["pos"] + offset - # mark these SNPs as remapped - temp.loc[snp_indices, "remapped"] = True + # mark these SNPs as remapped + temp.loc[snp_indices, "remapped"] = True return temp

more specific tellmegen detection We've had a few files wrongly identified as TellMeGen when they weren't, which means they failed to parse. This makes the TellMeGen detection more strict, which should mean these files fail to parse as a "unrecognised" type. more specific tellmegen detection We've had a few files wrongly identified as TellMeGen when they weren't, which means they failed to parse. This makes the TellMeGen detection more strict, which should mean these files fail to parse as a "unrecognised" type.

2021-06-27T04:59:07

0.0

[]

tira-io/tira

tira-io__tira-646

ac75ad53d6a94230713616253ec468545bfdbe44

diff --git a/.devcontainer/devcontainer.json b/.devcontainer/devcontainer.json index 985c4822..1ec1e985 100644 --- a/.devcontainer/devcontainer.json +++ b/.devcontainer/devcontainer.json @@ -56,6 +56,8 @@ "customizations": { "vscode": { "settings": { + "editor.formatOnSave": true, + "json.format.keepLines": true, "livePreview.portNumber": 3080, "remote.autoForwardPorts": false, "launch": { @@ -93,12 +95,21 @@ "stopAll": true } ] - } + }, + "[python]": { + "editor.codeActionsOnSave": { "source.organizeImports": "explicit" } + }, + "python.formatting.provider": "black", + "isort.args": [ "--profile", "black" ] }, "extensions": [ "Vue.volar", "ms-vscode.live-server", - "ms-python.python" + "ms-python.python", + "ms-python.isort", + "ms-python.black-formatter", + "ms-python.flake8", + "ms-python.mypy-type-checker" ] } }, @@ -108,5 +119,5 @@ // Needed for parts of the python-client that build and run docker containers "privileged": true, - "mounts": ["source=/var/run/docker.sock,target=/var/run/docker.sock,type=bind"] + "mounts": [ "source=/var/run/docker.sock,target=/var/run/docker.sock,type=bind" ] } diff --git a/.github/workflows/linters.yml b/.github/workflows/linters.yml new file mode 100644 index 00000000..a9d31af7 --- /dev/null +++ b/.github/workflows/linters.yml @@ -0,0 +1,83 @@ +name: Linters + +on: + push: {} + workflow_dispatch: {} + +jobs: + mypy: + runs-on: ubuntu-latest + timeout-minutes: 15 + + steps: + - uses: actions/checkout@v4 + - name: Set up Python 3.9 + uses: actions/setup-python@v4 + with: + python-version: 3.9 + - name: Install mypy + run: pip3 install mypy + - name: Run mypy on backend + working-directory: ${{github.workspace}}/application + run: | + mkdir .mypy_cache + mypy . --disallow-untyped-calls --explicit-package-bases --ignore-missing-imports --install-types --non-interactive --cache-dir=.mypy_cache/ + - name: Run mypy on python-client + working-directory: ${{github.workspace}}/python-client + run: | + mkdir .mypy_cache + mypy . --disallow-untyped-calls --explicit-package-bases --ignore-missing-imports --install-types --non-interactive --cache-dir=.mypy_cache/ + + flake8: + runs-on: ubuntu-latest + timeout-minutes: 15 + steps: + - uses: actions/checkout@v4 + - name: Set up Python 3.9 + uses: actions/setup-python@v4 + with: + python-version: 3.9 + - name: Install flake8 + run: pip3 install flake8 + - name: Run flake8 on backend + working-directory: ${{github.workspace}}/application + run: flake8 . + - name: Run flake8 on python-client + working-directory: ${{github.workspace}}/python-client + run: flake8 . + + black: + runs-on: ubuntu-latest + timeout-minutes: 15 + steps: + - uses: actions/checkout@v4 + - name: Set up Python 3.9 + uses: actions/setup-python@v4 + with: + python-version: 3.9 + - name: Install black + run: pip3 install black + - name: Run black on backend + working-directory: ${{github.workspace}}/application + run: black --check . + - name: Run black on python-client + working-directory: ${{github.workspace}}/python-client + run: black --check . + + isort: + runs-on: ubuntu-latest + timeout-minutes: 15 + steps: + - uses: actions/checkout@v4 + - name: Set up Python 3.9 + uses: actions/setup-python@v4 + with: + python-version: 3.9 + - name: Install isort + run: pip3 install isort + - name: Run isort on backend + working-directory: ${{github.workspace}}/application + run: isort --check . + - name: Run isort on python-client + working-directory: ${{github.workspace}}/python-client + run: isort --check . \ No newline at end of file diff --git a/Dockerfile.dev b/Dockerfile.dev index 92cc778e..4d7143a1 100644 --- a/Dockerfile.dev +++ b/Dockerfile.dev @@ -38,8 +38,11 @@ RUN apt-get update && apt-get install -y npm yarn ######################################################################################################################## ENV PIP_BREAK_SYSTEM_PACKAGES 1 USER root -RUN apt-get update && apt-get install -y python3 python3-pip python3-dev pkg-config default-libmysqlclient-dev \ +RUN <<EOF +apt-get update && apt-get install -y python3 python3-pip python3-dev pkg-config default-libmysqlclient-dev \ libpcre3-dev +pip3 install black flake8 isort mypy +EOF COPY application/requirements.txt /tmp/application/requirements.txt RUN <<EOF # Create a dummy secret diff --git a/README.md b/README.md index d3e2f3ad..a1b42c73 100644 --- a/README.md +++ b/README.md @@ -1,7 +1,7 @@ <h1 align="center">TIRA</h1> - <img src="https://github.com/tira-io/tira-branding/raw/master/tira-icons/logo-tira-120x120-transparent.png"> + <img width=120px src="https://assets.tira.io/tira-icons/tira-logo.svg">  @@ -14,6 +14,12 @@ <a href="https://tira.io"> <img alt="Deployment" src="https://img.shields.io/badge/dynamic/json?url=https%3A%2F%2Fwww.tira.io%2Fapi%2Fv1%2F&query=%24.version&prefix=v.&label=tira.io"/> </a> + <a href="https://github.com/tira-io/tira/actions/workflows/run-all-tests.yml"> + <img alt="Tests" src="https://github.com/tira-io/tira/actions/workflows/run-all-tests.yml/badge.svg"/> + </a> + <a href="https://github.com/tira-io/tira/actions/workflows/linters.yml"> + <img alt="Linters" src="https://github.com/tira-io/tira/actions/workflows/linters.yml/badge.svg"/> + </a> <a href="https://tira-io.github.io/tira/">Documentation</a>  |  <a href="./application/">Backend</a>  |  diff --git a/application/pyproject.toml b/application/pyproject.toml new file mode 100644 index 00000000..a5adcfac --- /dev/null +++ b/application/pyproject.toml @@ -0,0 +1,8 @@ +[tool.black] +line-length = 120 + +[tool.isort] +profile = "black" +multi_line_output = 3 +line_length = 120 +include_trailing_comma = true diff --git a/application/requirements.txt b/application/requirements.txt index ab27b754..29e27b2c 100644 --- a/application/requirements.txt +++ b/application/requirements.txt @@ -1,3 +1,4 @@ +setuptools==71 grpcio>=1.53.2 # grpcio-tools==1.36.1 # still needed? protobuf<4.0dev diff --git a/application/setup.cfg b/application/setup.cfg new file mode 100644 index 00000000..8332de34 --- /dev/null +++ b/application/setup.cfg @@ -0,0 +1,3 @@ +[flake8] +max-line-length = 120 +extend-ignore = E203 diff --git a/python-client/pyproject.toml b/python-client/pyproject.toml index e34796ec..a5adcfac 100644 --- a/python-client/pyproject.toml +++ b/python-client/pyproject.toml @@ -1,2 +1,8 @@ [tool.black] -line-length = 120 \ No newline at end of file +line-length = 120 + +[tool.isort] +profile = "black" +multi_line_output = 3 +line_length = 120 +include_trailing_comma = true diff --git a/python-client/setup.cfg b/python-client/setup.cfg index ad5c0c57..7a6763d4 100644 --- a/python-client/setup.cfg +++ b/python-client/setup.cfg @@ -56,6 +56,3 @@ tira.static_redirects = *.json [flake8] max-line-length = 120 extend-ignore = E203 - -[isort] -profile=black

Disentangle the Frontend and the Backend

2024-05-28T07:10:02

0.0

[]

tira-io/tira

tira-io__tira-632

3f5fe190b6fc2851fd3da1bdd5ac639b7379b63c

diff --git a/application/Makefile b/application/Makefile index 775fdfaba..ced022a55 100644 --- a/application/Makefile +++ b/application/Makefile @@ -1,7 +1,7 @@ .PHONY: help setup run-develop build-docker clean -VERSION_APPLICATION=0.0.103 -VERSION_GRPC=0.0.103 +VERSION_APPLICATION=0.0.104 +VERSION_GRPC=0.0.104 .DEFAULT: help help: diff --git a/application/src/django_admin/settings.py b/application/src/django_admin/settings.py index ba6d51988..e7661fccf 100644 --- a/application/src/django_admin/settings.py +++ b/application/src/django_admin/settings.py @@ -350,12 +350,14 @@ def logger_config(log_dir: Path): 'webpage-classification': 'OpenWebSearch/irixys23-tira-submission-template', 'valueeval-2024-human-value-detection': 'touche-webis-de/valueeval24-tira-software-submission-template', 'workshop-on-open-web-search': 'tira-io/wows24-submission-template', + 'nlpbuw-fsu-sose-24': 'webis-de/natural-language-processing-exercises', } REFERENCE_DATASETS = { 'ir-lab-padua-2024': 'ir-lab-padua-2024/longeval-tiny-train-20240315-training', 'ir-benchmarks': 'ir-benchmarks/cranfield-20230107-training', 'workshop-on-open-web-search': 'workshop-on-open-web-search/retrieval-20231027-training', + 'generative-ai-authorship-verification-panclef-2024': 'generative-ai-authorship-verification-panclef-2024/pan24-generative-authorship-tiny-smoke-20240417-training', } CODE_SUBMISSION_REPOSITORY_NAMESPACE = 'tira-io' diff --git a/application/src/tira/frontend-vuetify/src/components/EditTask.vue b/application/src/tira/frontend-vuetify/src/components/EditTask.vue index 8ed2eea41..63499430a 100644 --- a/application/src/tira/frontend-vuetify/src/components/EditTask.vue +++ b/application/src/tira/frontend-vuetify/src/components/EditTask.vue @@ -33,13 +33,13 @@ A very small dataset (e.g., of 5 input instances) that can be used that participants get very fast feedback that their software is compatible with the data format for the task. This is intended to be fully public so that participants can use the smoke test dataset to test their software on their machine <v-divider/> <h3 class="my-1">Component 2 (optional): An Evaluator</h3> - A docker image that uses the ground-truth data and the outputs of some software as input to test if the outputs are valid and to measure their effectiveness. We already have many evaluators for many different scenarious available. Please look <a href="https://github.com/tira-io/tira/wiki/Organizing-Tasks#create-a-docker-image-for-your-evaluator">at the documentation</a> for existing evaluators and a step-by-step guide to build a new evaluator. + A docker image that uses the ground-truth data and the outputs of some software as input to test if the outputs are valid and to measure their effectiveness. We already have many evaluators for many different scenarious available. Please look <a href="https://tira-io.github.io/tira/organizers/organizing-tasks.html#create-a-docker-image-for-your-evaluator">at the documentation</a> for existing evaluators and a step-by-step guide to build a new evaluator. <v-divider/> <h3 class="my-1">Component 3 (optional): Datasets</h3> You can upload as many datasets (consisting of the inputs for software submissions and the ground truth for evaluation) as you like in whatever format you want. TIRA distinguishes between public datasets for which participants can see the outputs to verify that their software submission is correct and test datasets for which all outputs and evaluations are blinded and not visible to participants. <v-divider/> <h3 class="my-1">Component 4 (optional): Baselines</h3> - In the best case, you provide the code, a published docker image, and instructions on how to compile the code into a docker image to simplify participation in your shared tasks. We have some examples on baselines that you can adopt for your shared task <a href="https://github.com/tira-io/tira/wiki/Organizing-Tasks#provide-public-baselines-to-simplify-participation">in the documentation</a>. + In the best case, you provide the code, a published docker image, and instructions on how to compile the code into a docker image to simplify participation in your shared tasks. We have some examples on baselines that you can adopt for your shared task <a href="https://tira-io.github.io/tira/organizers/organizing-tasks.html#provide-public-baselines-to-simplify-participation">in the documentation</a>. </v-card-text> </v-window-item> diff --git a/application/src/tira/frontend-vuetify/src/components/SoftwareDetails.vue b/application/src/tira/frontend-vuetify/src/components/SoftwareDetails.vue index 21ddf8a45..e88d31efc 100644 --- a/application/src/tira/frontend-vuetify/src/components/SoftwareDetails.vue +++ b/application/src/tira/frontend-vuetify/src/components/SoftwareDetails.vue @@ -141,7 +141,7 @@ export default { get('/task/' + this.task_id + '/vm/' + this.run.vm_id + '/run_details/' + this.run.run_id) .then(inject_response(this, {'loading': false})) .catch(() => {this.details_not_visible = true; this.loading = false}) - }, + }, updateTab() { this.component_tab = this.selectedComponentTab } diff --git a/application/src/tira/frontend-vuetify/src/submission-components/CodeSubmission.vue b/application/src/tira/frontend-vuetify/src/submission-components/CodeSubmission.vue index 549e24974..7377dbc07 100644 --- a/application/src/tira/frontend-vuetify/src/submission-components/CodeSubmission.vue +++ b/application/src/tira/frontend-vuetify/src/submission-components/CodeSubmission.vue @@ -3,7 +3,41 @@ <v-card-item v-if="loading"><loading :loading="loading"/></v-card-item> <v-card-item v-if="!loading && disabled"> <v-card-text> - Code submissions are not enabled for this task. Please contact your organizers <a :href="organizer_link">{{organizer}}</a> to enable code submissions. + You can submit to TIRA from your source code repository via GitHub actions that dockerize your code, test the docker image, and upload it to TIRA. This way, you don't need to install Docker on your machine, and our prepared Github action sends the metadata (i.e., the commit hash, the branch, the repository, etc.) to TIRA to ensure that the submission can be linked to the code that produced it. Please do not hesitate to contact your organizers <a :href="organizer_link">{{organizer}}</a> or the <a href="https://www.tira.io/categories">forum</a> in case of questions or problems. + + + You can setup a software submission in three steps: + + + <v-stepper :items="['Setup Your Git Repository', 'Add the Github Action', 'Submit']" flat :border=false> + <template v-slot:item.1> + <v-card title="" flat> + <h3>Create your Git repository</h3> + Please create your git repository with your code. The repository can be private or public. + + <h3>Add the TIRA_CLIENT_TOKEN secret to your Git repository</h3> + In your repository, go to "Settings" -> "Secrets and variables" -> "Actions" and add a new repository secret with the name TIRA_CLIENT_TOKEN and the value: + + <code-snippet title="TIRA_CLIENT_TOKEN" :code="token" expand_message=""/> + </v-card> + </template> + + <template v-slot:item.2> + <v-card title="" flat> + Please download the Github action <a :href="'/data-download/git-repo-template/' + user_id + '/' + task_id + '.zip'" target="_blank">here</a>. + + This zip directory contains the github action in the file ".github/workflows/upload-software-to-tira.yml" that you can add to your git repository. The rest of the zip directory shows exemplary how to connect the code of your repository to your repository, see the README.md in the zip for more details. + </v-card> + </template> + + <template v-slot:item.3> + <v-card title="" flat> + After you have added the Github action (e.g., ".github/workflows/upload-software-to-tira.yml") to your Github repository, please run it by clicking on "Actions" -> "Upload Software to TIRA" -> "Run workflow". + + After successful completion of the Github action, your new submission appears under the "Docker" tab above where you can select and run your new submission. + </v-card> + </template> + </v-stepper> </v-card-text> </v-card-item> <v-card-item v-if="!loading && !disabled"> @@ -83,7 +117,8 @@ export default { ssh_repo_url: '', owner_url: '', http_owner_url: '', - disabled: false + disabled: false, + token: '<ADD-YOUR-TOKEN-HERE>' } }, methods: { @@ -106,6 +141,9 @@ export default { get(`/api/get_software_submission_git_repository/${this.task_id}/${this.user_id}`) .then(inject_response(this, {'loading': false})) .catch(reportError("Problem loading the data of the task.", "This might be a short-term hiccup, please try again. We got the following error: ")) + + get('/api/token/' + this.user_id) + .then(inject_response(this)) }, computed: { organizer_link() {return get_link_to_organizer(this.organizer_id)}, diff --git a/application/src/tira/frontend-vuetify/src/submission-components/DockerSubmission.vue b/application/src/tira/frontend-vuetify/src/submission-components/DockerSubmission.vue index 6f6e951a6..d9ecd4019 100644 --- a/application/src/tira/frontend-vuetify/src/submission-components/DockerSubmission.vue +++ b/application/src/tira/frontend-vuetify/src/submission-components/DockerSubmission.vue @@ -114,7 +114,7 @@ export default { .then(inject_response(this, {'loading': false}, true)) .catch(reportError("Problem While Loading the Docker Details of the Task " + this.task_id, "This might be a short-term hiccup, please try again. We got the following error: ")) this.load_re_ranking_datasets() - this.step === '' ? this.tab = docker.docker_softwares[0].display_name : this.tab = "newDockerImage" + this.step === '' ? this.tab = this.docker.docker_softwares[0].display_name : this.tab = "newDockerImage" }, watch: { step(old_value, new_value) { diff --git a/application/src/tira/frontend-vuetify/src/submission-components/UploadSubmission.vue b/application/src/tira/frontend-vuetify/src/submission-components/UploadSubmission.vue index b1d483461..86bfc53fb 100644 --- a/application/src/tira/frontend-vuetify/src/submission-components/UploadSubmission.vue +++ b/application/src/tira/frontend-vuetify/src/submission-components/UploadSubmission.vue @@ -33,34 +33,87 @@ </v-col> </v-row> <v-window v-model="tab" v-if="!loading && role !== 'guest'" :touch="{left: null, right: null}"> - <v-window-item value="newUploadGroup"> - <h2>Create Run Upload Group</h2> - Please click on "Add Upload Group" below to create a new run upload group. - Please use one upload group (you can edit the metadata of an upload group later) per approach. I.e., in TIRA, you can usually run software submissions on different datasets. For manually uploaded runs, we employ the same methodology: Please create one run upload group per approach, so that you can upload "executions" of the same approach on different datasets while maintaining the documentation. - + <v-window-item value="newUploadGroup"> + <h2>New Upload</h2> + <v-stepper :items="['About', 'Specify Metadata', 'Upload']" v-model="stepperModel" hide-actions="true" flat :border=false> + <template v-slot:item.1> + <v-card title="Specify what you want to upload" flat> + <v-radio-group v-model="upload_configuration"> + <v-radio label="I want to upload runs" value="upload-config-1"/> + <v-radio label="I want to upload a model from Hugging Face" value="upload-config-2"/> + </v-radio-group> + </v-card> + </template> - <v-form> - <v-radio-group v-model="upload_type"> - <v-radio label="Rename all uploaded files as configured by the organizers" value="upload-1"/> - <v-radio label="I want to configure the name of my uploaded file manually (Only for expert users, e.g., to normalize all uploaded files of this group.)" value="upload-2"/> - </v-radio-group> + <template v-slot:item.2> + <v-card flat> + <div v-if="upload_configuration === 'upload-config-1'"> + In TIRA, approaches are often evaluated on multiple datasets. Please ensure that you correctly group runs created by the same approach. You have no upload groups yet. + 0">You already have uploads for the following approaches: + <v-list density="compact"> + <v-list-item v-for="i in userUploadGroups"> + <v-list-item-title>Upload run(s) for approach <a href="javascript:void(0);" @click="stepperModel=1; tab=i.id">{{i.display_name}}</a></v-list-item-title> + </v-list-item> + </v-list> + - <v-text-field v-if="upload_type === 'upload-2'" v-model="rename_file_to" label="Rename a file after upload to this name " /> - </v-form> + <div> + <v-checkbox label="I want to upload runs for a new approach" v-model="new_approach"></v-checkbox> - <v-btn variant="outlined" @click="addUpload()"> - Add Upload Group - </v-btn> + <div v-if="new_approach"> + You will create a new run upload group. Please ensure that you only create only one run upload group per approach, so that you can upload "executions" of the same approach on different datasets while maintaining the documentation. + <v-form> + <v-radio-group v-model="upload_type"> + <v-radio label="Rename all uploaded files as configured by the organizers" value="upload-1"/> + <v-radio label="I want to configure the name of my uploaded file manually (Only for expert users, e.g., to normalize all uploaded files of this group.)" value="upload-2"/> + </v-radio-group> - <div v-if="!showImportSubmission"> - - If you participate in multiple tasks, you can also <a href="javascript:void(0);" @click="showImportSubmission=true">import your upload group</a>. + <v-text-field v-if="upload_type === 'upload-2'" v-model="rename_file_to" label="Rename a file after upload to this name " /> + </v-form> + + Please click on "Add Upload Group" to create a new run upload group (i.e., if the run that you want to upload does not belong to the upload groups listed above). + + <v-btn variant="outlined" @click="addUpload()">Add Upload Group</v-btn> + + <div v-if="!showImportSubmission"> + + If you participate in multiple tasks, you can also <a href="javascript:void(0);" @click="showImportSubmission=true">import your upload group</a>. + </div> + <div v-if="showImportSubmission"> + + <import-submission submission_type="upload"/> + </div> + </div> </div> - <div v-if="showImportSubmission"> - - <import-submission submission_type="upload"/> + </div> + + <div v-if="upload_configuration === 'upload-config-2'"> + Please specify the model that will be downloaded via huggingface-cli + <v-text-field v-model="hugging_face_model" label="Hugging Face Model " /> + <div v-if="hugging_face_model"> + The following command will be used to download your model: + + <code-snippet title="TIRA will download and mount the model via" :code="'huggingface-cli download ' + hugging_face_model" expand_message=""/> + You can later mount the downloaded model into your software submission. Please click next to upload/import this Hugging Face model to TIRA. </div> + </div> + </v-card> + </template> + + <template v-slot:item.3> + <v-card title="Step Three" flat> + The Huggingface upload is disabled for this task. Please <a :href="contact_organizer">contact</a> the organizer <a :href="link_organizer"> {{ organizer }}</a> to enable this. + </v-card> + </template> + + + <v-stepper-actions @click:prev="stepperModel=1" @click:next="nextStep" :disabled='disableUploadStepper'></v-stepper-actions> + + </v-stepper> + + + </v-window-item> <v-window-item v-for="us in this.all_uploadgroups" :value="us.id"> <loading :loading="description === 'no-description'"/> @@ -80,26 +133,29 @@ </div> <v-form v-if="description !== 'no-description'"> - <v-file-input v-model="fileHandle" - :rules="[v => !!v || 'File is required']" - label="Click to add run file" - ></v-file-input> - <v-autocomplete label="Input Dataset" - :items="datasets" - item-title="display_name" - item-value="dataset_id" - prepend-icon="mdi-file-document-multiple-outline" - v-model="selectedDataset" - variant="underlined" - clearable/> - <v-text-field v-if="'' + rename_to !== 'null' && '' + rename_to !== '' && '' + rename_to !== 'undefined'" v-model="rename_to" label="Uploaded Files are renamed to (immutable for reproducibility)" disabled="true"/> + <h2>Upload new Submissions for Approach {{us.display_name}}</h2> + <v-radio-group v-model="upload_type_next_upload"> + <v-radio :label="'I want to upload a new run for approach ' + us.display_name + ' via the web UI'" value="upload-via-ui"/> + <v-radio :label="'I want to batch upload runs for approach ' + us.display_name + ' via a script'" value="upload-via-script"/> + </v-radio-group> + <div v-if="upload_type_next_upload == 'upload-via-ui'"> + <v-file-input v-model="fileHandle" :rules="[v => !!v || 'File is required']" label="Click to add run file" /> + <v-autocomplete label="Input Dataset" :items="datasets" item-title="display_name" item-value="dataset_id" prepend-icon="mdi-file-document-multiple-outline" v-model="selectedDataset" variant="underlined" clearable/> + <v-text-field v-if="'' + rename_to !== 'null' && '' + rename_to !== '' && '' + rename_to !== 'undefined'" v-model="rename_to" label="Uploaded Files are renamed to (immutable for reproducibility)" disabled="true"/> + </div> </v-form> - <v-btn v-if="description !== 'no-description'" color="primary" :loading="uploading" :disabled="uploading || fileHandle === null || selectedDataset === ''" + <v-btn v-if="description !== 'no-description' && upload_type_next_upload == 'upload-via-ui'" color="primary" :loading="uploading" :disabled="uploading || fileHandle === null || selectedDataset === ''" @click="fileUpload(us.id)">Upload Run</v-btn> + + <div v-if="upload_type_next_upload == 'upload-via-script'"> + <code-snippet title="(1) Make sure that your TIRA client is up to date and authenticated" :code="tira_setup_code" expand_message="(1) Make sure that your TIRA client is up to date and authenticated"/> + + <code-snippet title="Batch upload runs via python" :code="batch_upload_code(us.display_name)" expand_message=""/> + </div> - <h2>Submissions</h2> + <h2>Your Existing Submissions for Approach {{us.display_name}}</h2> <run-list :task_id="task_id" :organizer="organizer" :organizer_id="organizer_id" :vm_id="user_id_for_task" :upload_id="us.id" ref="upload-run-list" /> </v-window-item> </v-window> @@ -108,14 +164,15 @@ <script> import { VAutocomplete } from 'vuetify/components' -import { extractTaskFromCurrentUrl, extractUserFromCurrentUrl, get, inject_response, reportError, extractRole, post_file, reportSuccess, handleModifiedSubmission } from "@/utils"; +import { extractTaskFromCurrentUrl, extractUserFromCurrentUrl, get, inject_response, reportError, extractRole, post_file, reportSuccess, handleModifiedSubmission, get_link_to_organizer, get_contact_link_to_organizer } from "@/utils"; import { Loading, LoginToSubmit, RunList } from "@/components"; import EditSubmissionDetails from "@/submission-components/EditSubmissionDetails.vue"; import ImportSubmission from "./ImportSubmission.vue"; +import CodeSnippet from "../components/CodeSnippet.vue" export default { name: "upload-submission", - components: {EditSubmissionDetails, Loading, VAutocomplete, LoginToSubmit, RunList, ImportSubmission}, + components: {EditSubmissionDetails, Loading, VAutocomplete, LoginToSubmit, RunList, ImportSubmission, CodeSnippet}, props: ['organizer', 'organizer_id'], emits: ['refresh_running_submissions'], data () { @@ -130,7 +187,12 @@ export default { uploadDataset: '', uploadFormError: '', upload_type: 'upload-1', + new_approach: false, + upload_configuration: '', + upload_type_next_upload: '', rename_file_to: '', + hugging_face_model: '', + stepperModel: '', fileHandle: null, description: 'no-description', rename_to: '', @@ -138,6 +200,7 @@ export default { all_uploadgroups: [{"id": null, "display_name": 'loading...'}], selectedDataset: '', showImportSubmission: false, + token: 'YOUR-TOKEN-HERE', datasets: [{"dataset_id": "loading...", "display_name": "loading...",}] } }, @@ -150,6 +213,32 @@ export default { } return ret.concat(this.all_uploadgroups) + }, + userUploadGroups() { + let ret = [] + + for (let i of this.all_uploadgroups) { + if (i.display_name != 'default upload') { + ret.push(i) + } + } + + return ret + }, + link_organizer() {return get_link_to_organizer(this.organizer_id);}, + contact_organizer() {return get_contact_link_to_organizer(this.organizer_id);}, + tira_setup_code() { + return 'pip3 uninstall -y tira\npip3 install tira\ntira-cli login --token ' + this.token + }, + disableUploadStepper() { + if(this.stepperModel == '1' && !this.upload_configuration) { + return 'next'; + } + if (this.stepperModel == '2' && (this.upload_configuration == 'upload-config-1' || !this.hugging_face_model)) { + return 'next' + } + + return false; } }, methods: { @@ -172,6 +261,26 @@ export default { this.description = editedDetails.description handleModifiedSubmission(editedDetails, this.all_uploadgroups) }, + batch_upload_code(display_name) { + return 'from tira.rest_api_client import Client\n' + + 'from pathlib import Path\n' + + 'from tqdm import tqdm\n\n' + + 'tira = Client()\n' + + 'approach = \'' + this.task_id + '/' + this.user_id_for_task + '/' + display_name + '\'\n' + + 'dataset_ids = [' + this.datasets.map((i) => '\'' + i.dataset_id + '\'') + ']\n\n' + + 'for dataset_id in tqdm(dataset_ids):\n' + + ' # assume run to upload is located in a file dataset_id/run\n' + + ' run_file = Path(dataset_id) / \'run\'\n' + + ' tira.upload_run(approach=approach, dataset_id=dataset_id, file_path=run_file);\n' + }, + nextStep() { + if (this.stepperModel == 1) { + this.stepperModel = 2; + } + else if (this.stepperModel == 2 && this.upload_configuration === 'upload-config-2') { + this.stepperModel = 3; + } + }, deleteUpload(id_to_delete) { get(`/task/${this.task_id}/vm/${this.user_id_for_task}/upload-delete/${this.tab}`) .then(message => { @@ -204,6 +313,10 @@ export default { get('/api/submissions-for-task/' + this.task_id + '/' + this.user_id_for_task + '/upload') .then(inject_response(this, {'loading': false}, true)) .catch(reportError("Problem While Loading The Submissions of the Task " + this.task_id, "This might be a short-term hiccup, please try again. We got the following error: ")) + + get('/api/token/' + this.user_id_for_task) + .then(inject_response(this)) + this.tab = this.all_uploadgroups[0].display_name }, watch: { diff --git a/application/src/tira/management/commands/archive_runs_to_zenodo.py b/application/src/tira/management/commands/archive_runs_to_zenodo.py index 5eb8bcc6b..965a40cc5 100644 --- a/application/src/tira/management/commands/archive_runs_to_zenodo.py +++ b/application/src/tira/management/commands/archive_runs_to_zenodo.py @@ -33,11 +33,14 @@ def handle(self, *args, **options): ], 'trec-core': [ 'wapo-v2-trec-core-2018-20230107-training', 'disks45-nocr-trec8-20230209-training', 'disks45-nocr-trec7-20230209-training', 'disks45-nocr-trec-robust-2004-20230209-training' + ], + 'ir-lab': [ + 'anthology-20240411-training' ] } #we publish document processors only for fully public datasets, query processors can be published on all groups - fully_public_datasets = dataset_groups['trec-recent'] + dataset_groups['tiny-test-collections'] + dataset_groups['trec-medical'] + dataset_groups['clef-labs'] + fully_public_datasets = dataset_groups['trec-recent'] + dataset_groups['tiny-test-collections'] + dataset_groups['trec-medical'] + dataset_groups['clef-labs'] + dataset_groups['ir-lab'] systems = { 'ir-benchmarks': { @@ -47,6 +50,24 @@ def handle(self, *args, **options): 'seanmacavaney': { 'DocT5Query': 'doc-t5-query', 'corpus-graph': 'corpus-graph', + }, + 'ows': { + 'pyterrier-anceindex': 'pyterrier-anceindex' + }, + 'ir-lab-sose-2024': { + 'tira-ir-starter': { + 'Index (tira-ir-starter-pyterrier)': 'ir-lab-sose-2024', + 'Index (pyterrier-stanford-lemmatizer)': 'ir-lab-sose-2024', + }, + 'seanmacavaney': { + 'DocT5Query': 'ir-lab-sose-2024', + 'corpus-graph': 'ir-lab-sose-2024', + }, + 'ows': { + 'pyterrier-anceindex': 'ir-lab-sose-2024' + }, + 'naverlabseurope': { + 'Splade (Index)': 'ir-lab-sose-2024' } }, } @@ -105,7 +126,7 @@ def handle(self, *args, **options): ret[task_id][user_id][display_name] = {} output_dir = systems[task_id][user_id][display_name] for i in tqdm(fully_public_datasets): - run_id = model.runs(task_id, i, user_id, display_name)[0] + run_id = model.runs(task_id, i, user_id, display_name)[0]['run_id'] target_file = f'{output_dir}/{run_id}.zip' zip_file = zip_run(i, user_id, run_id) @@ -147,4 +168,4 @@ def handle(self, *args, **options): print(json.dumps(ret)) - \ No newline at end of file + diff --git a/application/src/tira/templates/tira/git-repo-template/Dockerfile b/application/src/tira/templates/tira/git-repo-template/Dockerfile new file mode 100644 index 000000000..a947e0de5 --- /dev/null +++ b/application/src/tira/templates/tira/git-repo-template/Dockerfile @@ -0,0 +1,9 @@ +FROM python:3 + +ADD script.py /script.py +ADD requirements.txt /requirements.txt + +RUN pip3 install -r /requirements.txt + +ENTRYPOINT [ "python3", "/script.py", "-i", "$inputDataset", "-o", "$outputDir" ] + diff --git a/application/src/tira/templates/tira/git-repo-template/README.md b/application/src/tira/templates/tira/git-repo-template/README.md new file mode 100644 index 000000000..9397db9c8 --- /dev/null +++ b/application/src/tira/templates/tira/git-repo-template/README.md @@ -0,0 +1,27 @@ +# Code Submission Github Repository for Task {{ task_id }} + +This is a git repository template for code submissions to the shared task {{ task_id }} in TIRA. + +This repository explains how to configure a Github action for your code submission. +The corresponding Github action is [.github/workflows/upload-software-to-tira.yml](.github/workflows/upload-software-to-tira.yml). + +You must copy it to the .github/workflows of your own repository and adjust it to your code. + +The rest of this repository explains how the Github action works with a minimal example, so that you can test it on your machine but also directly by running the action in your own repository. + +The github action makes the following steps: + +(1) Install dependencies, i.e., install python, tira (`pip3 install tira`) and Docker. + +(2) Build the Docker image: (you might have to modify the Dockerfile, e.g., add your own code, install other dependencies, etc. + +``` +docker build -t {{ image }} -f Dockerfile . +``` + +(3) Test that the docker image works on a small example dataset and push it to TIRA: + +``` +tira-run --input-dataset {{ input_dataset }} --image {{ image }} +``` + diff --git a/application/src/tira/templates/tira/git-repo-template/github-action.yml b/application/src/tira/templates/tira/git-repo-template/github-action.yml new file mode 100644 index 000000000..f689b6d41 --- /dev/null +++ b/application/src/tira/templates/tira/git-repo-template/github-action.yml @@ -0,0 +1,25 @@ +name: Upload Software to TIRA +on: workflow_dispatch +jobs: + docker-build: + runs-on: ubuntu-latest + timeout-minutes: 45 + steps: + - name: Checkout + uses: actions/checkout@v3 + - name: Set up QEMU + uses: docker/setup-qemu-action@v2 + - name: Set up Docker Buildx + uses: docker/setup-buildx-action@v2 + - name: Set up Dependencies + uses: actions/setup-python@v4 + with: + python-version: 3.8 + - name: Install required python packages + run: | + python -m pip install --upgrade pip + pip3 install tira + - name: Build, test, and upload image + run: | + docker build -t {{ image }} -f {% verbatim %}${{ github.workspace }}{% endverbatim %}/Dockerfile {% verbatim %}${{ github.workspace }}{% endverbatim %} + tira-run --tira-client-token {% verbatim %}${{ secrets.TIRA_CLIENT_TOKEN }}{% endverbatim %} --input-dataset {{ input_dataset }} --image {{ image }} --push true --fail-if-output-is-empty diff --git a/application/src/tira/templates/tira/git-repo-template/script.py b/application/src/tira/templates/tira/git-repo-template/script.py new file mode 100644 index 000000000..af43c5d21 --- /dev/null +++ b/application/src/tira/templates/tira/git-repo-template/script.py @@ -0,0 +1,16 @@ +import argparse + +def parse_args(): + parser = argparse.ArgumentParser(prog='script') + parser.add_argument('-i', '--input', required=True, help='the input to the script.') + parser.add_argument('-o', '--output', required=True, help='the output of the script.') + + return parser.parse_args() + +if __name__ == '__main__': + args = parse_args() + + print(f'This is a demo, I ignore the passed input {args.input} and write some content into the output file {args.output}.') + with open(args.output + '/predictions.jsonl', 'w') as f: + f.write('hello world') + print('Done. I wrote "hello world" to {args.output}/predictions.jsonl.') diff --git a/application/src/tira/tira_model.py b/application/src/tira/tira_model.py index 41c17adec..aacf76dd1 100644 --- a/application/src/tira/tira_model.py +++ b/application/src/tira/tira_model.py @@ -129,12 +129,20 @@ def discourse_api_client(): def tira_run_command(image, command, task_id): - input_dataset = f'{task_id}/ADD-DATASET-ID-HERE' - if task_id in settings.REFERENCE_DATASETS: - input_dataset = settings.REFERENCE_DATASETS[task_id] + input_dataset = reference_dataset(task_id) return f'tira-run \\\n --input-dataset {input_dataset} \\\n --image {image} \\\n --command \'{command}\'' +def reference_dataset(task_id): + if task_id in settings.REFERENCE_DATASETS: + return settings.REFERENCE_DATASETS[task_id] + else: + available_datasets = get_datasets_by_task(task_id) + available_datasets = [i['dataset_id'] for i in available_datasets if i['dataset_id'].endswith('-training') and not i['is_confidential'] and not i['is_deprecated']] + if len(available_datasets) > 0: + return f'{task_id}/{available_datasets[0]}' + else: + return f'{task_id}/ADD-DATASET-ID-HERE' def tira_docker_registry_token(docker_software_help): ret = docker_software_help.split('docker login -u ')[1].split(' -p') diff --git a/application/src/tira/urls.py b/application/src/tira/urls.py index f4976ba1f..d4a9e8440 100644 --- a/application/src/tira/urls.py +++ b/application/src/tira/urls.py @@ -12,6 +12,7 @@ path('background_jobs/<str:task_id>/<str:job_id>', views.background_jobs, name='background_jobs'), path('task/<str:task_id>/user/<str:vm_id>/dataset/<str:dataset_id>/download/<str:run_id>.zip', views.download_rundir, name='download_rundir'), + path('data-download/git-repo-template/<str:vm_id>/<str:task_id>.zip', views.download_repo_template, name='download_repo_template'), path('data-download/<str:dataset_type>/<str:input_type>/<str:dataset_id>.zip', views.download_datadir, name='download_datadir'), re_path(r'^frontend-vuetify/.*', views.veutify_page, name='vuetify_page'), diff --git a/application/src/tira/views.py b/application/src/tira/views.py index 7d3c106e9..5df2dd840 100644 --- a/application/src/tira/views.py +++ b/application/src/tira/views.py @@ -1,4 +1,5 @@ from django.shortcuts import render, redirect +from django.template.loader import render_to_string from django.http import JsonResponse, FileResponse from django.conf import settings from django.core.cache import cache @@ -17,6 +18,7 @@ import zipfile import json from http import HTTPStatus +import tempfile logger = logging.getLogger("tira") logger.info("Views: Logger active") @@ -188,6 +190,31 @@ def download_rundir(request, task_id, dataset_id, vm_id, run_id): def download_input_rundir(request, task_id, dataset_id, vm_id, run_id): return download_rundir(request, task_id, dataset_id, vm_id, input_run_id) +def download_repo_template(request, task_id, vm_id): + with tempfile.TemporaryDirectory() as tmpdirname: + directory = Path(tmpdirname) / f'git-repo-template-{task_id}' + os.makedirs(directory, exist_ok=True) + os.makedirs(directory / '.github' / 'workflows', exist_ok=True) + context = { + 'task_id': task_id, + 'image': f'registry.webis.de/code-research/tira/tira-user-{vm_id}/github-action-submission:0.0.1', + 'input_dataset': model.reference_dataset(task_id), + } + + with (directory / 'README.md').open('w') as readme, (directory / 'script.py').open('w') as script, (directory / 'requirements.txt').open('w') as requirements, (directory / 'Dockerfile').open('w') as dockerfile, (directory / '.github' / 'workflows' / 'upload-software-to-tira.yml').open('w') as ci: + readme.write(render_to_string('tira/git-repo-template/README.md', context=context)) + dockerfile.write(render_to_string('tira/git-repo-template/Dockerfile', context=context)) + requirements.write('argparse') + script.write(render_to_string('tira/git-repo-template/script.py', context=context)) + ci.write(render_to_string('tira/git-repo-template/github-action.yml', context=context)) + + zipped = Path(tmpdirname) / f'{task_id}.zip' + with zipfile.ZipFile(zipped, "w") as zipf: + for f in directory.rglob('*'): + zipf.write(f, arcname=f.relative_to(directory)) + + return FileResponse(open(zipped, "rb"), as_attachment=True, filename=f"git-repo-template-{task_id}.zip") + @check_permissions def download_datadir(request, dataset_type, input_type, dataset_id): input_type = input_type.lower().replace('input', '') diff --git a/python-client/Makefile b/python-client/Makefile index 35e487907..6cc0ca7b6 100644 --- a/python-client/Makefile +++ b/python-client/Makefile @@ -3,7 +3,7 @@ build-pypi-package: run-tests python3 -m build --sdist . python3 -m build --wheel . - twine upload dist/tira-0.0.113-py3-none-any.whl dist/tira-0.0.113.tar.gz + twine upload dist/tira-0.0.125-py3-none-any.whl dist/tira-0.0.125.tar.gz run-tests: docker run -u root --rm -v /var/run/docker.sock:/var/run/docker.sock -v ${PWD}:/app -w /app --entrypoint pytest webis/tira:python-client-dev-0.0.5 diff --git a/python-client/tira/__init__.py b/python-client/tira/__init__.py index 15e9d7b5b..58f610684 100644 --- a/python-client/tira/__init__.py +++ b/python-client/tira/__init__.py @@ -1,1 +1,1 @@ -__version__ = "0.0.113" +__version__ = "0.0.125" diff --git a/python-client/tira/io_utils.py b/python-client/tira/io_utils.py index 3c60bd380..d647b4c17 100644 --- a/python-client/tira/io_utils.py +++ b/python-client/tira/io_utils.py @@ -140,6 +140,35 @@ def parse_prototext_key_values(file_name): yield ret +def to_prototext(m: List[Dict[str, Any]], upper_k: str = "") -> str: + ret = "" + + def _to_prototext(d: Dict[str, Any], upper_k: str = "") -> str: + ret = "" + for k, v in d.items(): + new_k = upper_k + if not new_k: + new_k = k + elif not new_k.endswith(k): + new_k = upper_k + "_" + k + if isinstance(v, dict): + ret += _to_prototext(v, upper_k=new_k) + else: + ret += ( + 'measure{\n key: "' + + str(new_k.replace("_", " ").title()) + + '"\n value: "' + + str(v) + + '"\n}\n' + ) + return ret + + for d in m: + ret += _to_prototext(d, upper_k=upper_k) + + return ret + + def all_environment_variables_for_github_action_or_fail(params): ret = {} diff --git a/python-client/tira/ir_datasets_util.py b/python-client/tira/ir_datasets_util.py index 1f189c49b..28eab76c9 100644 --- a/python-client/tira/ir_datasets_util.py +++ b/python-client/tira/ir_datasets_util.py @@ -31,6 +31,26 @@ def default_text(self): """ return self.title +class DictDocsstore(): + def __init__(self, docs): + self.docs = docs + + def __getitem__(self, item): + return self.docs[item] + + def get(self, item): + return self.docs.get(item) + + def __iter__(self): + return self.docs.__iter__() + + def __len__(self): + return len(self.docs) + + def get_many_iter(self, docids): + for docid in docids: + yield self.get(docid) + def register_dataset_from_re_rank_file(ir_dataset_id, df_re_rank, original_ir_datasets_id=None): """ @@ -80,7 +100,7 @@ def docs_count(self): return sum(1 for _ in self.stream_docs()) def docs_store(self): - return self.__parsed_docs() + return DictDocsstore(self.__parsed_docs()) def __parsed_docs(self): if self.docs is None: @@ -253,7 +273,9 @@ def load(self, dataset_id): docs = self.lazy_docs(lambda: get_download_dir_from_tira(dataset_id, False), None, True) queries = self.lazy_queries(lambda: get_download_dir_from_tira(dataset_id, True), None) qrels = self.lazy_qrels(lambda: get_download_dir_from_tira(dataset_id, True), None) - return Dataset(docs, queries, qrels) + ret = Dataset(docs, queries, qrels) + ret.dataset_id = lambda: dataset_id + return ret def topics_file(self, tira_path): return get_download_dir_from_tira(tira_path, True) + '/queries.xml' diff --git a/python-client/tira/license_agreements.py b/python-client/tira/license_agreements.py new file mode 100644 index 000000000..c97c14f1e --- /dev/null +++ b/python-client/tira/license_agreements.py @@ -0,0 +1,5 @@ +def print_license_agreement(url): + if not url: + return + if url and 'longeval' in url or 'leipzig-topics-small-20240119' in url or 'jena-topics-20231026-test' in url or 'jena-topics-small-20240119-training' in url or 'leipzig-topics-20231025-test' in url or 'training-20231104-training' in url or 'validation-20231104-training' in url: + print('The download is derived from The LongEval Dataset under the "Qwant LongEval Attribution-NonCommercial-ShareAlike License". Hence, the download is also under this License. By using it, you agree to the terms of this license. Please find details at: https://lindat.mff.cuni.cz/repository/xmlui/page/Qwant_LongEval_BY-NC-SA_License') \ No newline at end of file diff --git a/python-client/tira/local_execution_integration.py b/python-client/tira/local_execution_integration.py index a760347fc..e5576e186 100644 --- a/python-client/tira/local_execution_integration.py +++ b/python-client/tira/local_execution_integration.py @@ -91,7 +91,6 @@ def ensure_image_available_locally(self, image, client=None): def extract_entrypoint(self, image): image_name = image - from tira.third_party_integrations import extract_to_be_executed_notebook_from_command_or_none self.ensure_image_available_locally(image_name) image = self.__docker_client().images.get(image_name) ret = image.attrs['Config']['Entrypoint'] @@ -104,13 +103,16 @@ def extract_entrypoint(self, image): ret = json.loads(i) break - ret = ' '.join(ret) - executable = extract_to_be_executed_notebook_from_command_or_none(ret) + return self.make_command_absolute(image_name, ' '.join(ret)) + + def make_command_absolute(self, image_name, command): + from tira.third_party_integrations import extract_to_be_executed_notebook_from_command_or_none + executable = extract_to_be_executed_notebook_from_command_or_none(command) - if not executable or executable.startswith('/'): - return ret + if not executable or executable.startswith('/') or executable.startswith("'/") or executable.startswith('"/'): + return command else: - return ret.replace(executable, (self.docker_image_work_dir(image_name) + '/' + executable).replace('//', '/')) + return command.replace(executable, (self.docker_image_work_dir(image_name) + '/' + executable).replace('//', '/').replace('/./', '/')) def __docker_client(self): try: @@ -165,7 +167,7 @@ def login_docker_client(self, task_name, team_name, client=None): return True - def run(self, identifier=None, image=None, command=None, input_dir=None, output_dir=None, evaluate=False, dry_run=False, docker_software_id_to_output=None, software_id=None, allow_network=False, input_run=None, additional_volumes=None, eval_dir='tira-evaluation'): + def run(self, identifier=None, image=None, command=None, input_dir=None, output_dir=None, evaluate=False, dry_run=False, docker_software_id_to_output=None, software_id=None, allow_network=False, input_run=None, additional_volumes=None, eval_dir='tira-evaluation', gpu_count=0): previous_stages = [] original_args = {'identifier': identifier, 'image': image, 'command': command, 'input_dir': input_dir, 'output_dir': output_dir, 'evaluate': evaluate, 'dry_run': dry_run, 'docker_software_id_to_output': docker_software_id_to_output, 'software_id': software_id} s_id = 'unknown-software-id' @@ -228,7 +230,11 @@ def run(self, identifier=None, image=None, command=None, input_dir=None, output_ if input_run: environment['inputRun'] = '/tira-data/input-run' - container = client.containers.run(image, entrypoint='sh', command=f'-c "{command}; sleep .1"', environment=environment, volumes=volumes, detach=True, remove=True, network_disabled = not allow_network) + device_requests = [] + if gpu_count != 0: + device_requests=[docker.types.DeviceRequest(count=gpu_count, capabilities=[['gpu']])] + + container = client.containers.run(image, entrypoint='sh', command=f'-c "{command}; sleep .1"', environment=environment, volumes=volumes, detach=True, remove=True, network_disabled = not allow_network, device_requests=device_requests) for line in container.attach(stdout=True, stream=True, logs=True): print(line.decode('utf-8')) @@ -307,19 +313,27 @@ def export_submission_from_jupyter_notebook(self, notebook): return None ret = [] - + ret += ['TIRA_COMMAND=/workspace/run-pyterrier-notebook.py --input ${TIRA_INPUT_DIRECTORY} --output ${TIRA_OUTPUT_DIRECTORY} --notebook /workspace/' + notebook.split("/")[-1]] notebook_content = json.load(open(notebook, 'r')) return '\n'.join(ret) + def normalize_image_name(self, image, required_prefix): + if required_prefix and not image.startswith(required_prefix): + ret = (required_prefix + '/' + image[:10]).replace('//', '/') + ':' + (image.split(':')[-1] if ':' in image else 'latest') + return ret.replace('-:', ':').replace('::', ':').replace('/:', ':') + else: + return image + def push_image(self, image, required_prefix=None, task_name=None, team_name=None): client = self.__docker_client() if not self.docker_client_is_authenticated(client): self.login_docker_client(task_name, team_name, client) - + new_image = image if required_prefix and not image.startswith(required_prefix): - new_image = (required_prefix + '/' + image[:10]).replace('//', '/') + ':' + (image.split(':')[-1] if ':' in image else 'latest') + new_image = self.normalize_image_name(image, required_prefix) + print(f'I tag the image "{image}" as "{new_image}" for upload to TIRA (only internal).') client.images.get(image).tag(new_image) image = new_image diff --git a/python-client/tira/pandas_integration.py b/python-client/tira/pandas_integration.py index 5cc596528..da285607f 100644 --- a/python-client/tira/pandas_integration.py +++ b/python-client/tira/pandas_integration.py @@ -1,8 +1,24 @@ +from typing import List, Tuple +PANDAS_DTYPES = {'docno': str, 'doc_id': str, 'id': str, 'qid': str, 'query_id': str, 'queryid': str, 'docno': str, 'doc_id': str, 'docid': str} + class PandasIntegration(): + """Handling of inputs/outputs in TIRA with pandas. All methods here work in the TIRA sandbox without internet connection (when the data is mounted read-only). + """ def __init__(self, tira_client): self.tira_client = tira_client - def from_retriever_submission(self, approach, dataset, previous_stage=None, datasets=None): + def from_retriever_submission(self, approach: str, dataset: str, previous_stage: str=None, datasets: List[str]=None): + """Load a run file as pandas dataframe from tira. Compatible with PyTerrier. + + Args: + approach (str): the approach for which the run should be loaded, in the format 'task/team/software'. + dataset (str): the dataset id, either an tira or ir_datasets id. + previous_stage (str, optional): The previous stage, in case the approach itself is ambigious. Defaults to None. + datasets (List[str], optional): _description_. Defaults to None. + + Returns: + pd.DataFrame: The run file parsed to a pandas DataFrame. + """ import pandas as pd from tira.ir_datasets_util import translate_irds_id_to_tirex task, team, software = approach.split('/') @@ -42,7 +58,20 @@ def __matching_files(self, approach, dataset, file_selection): return sorted(list(ret)) - def transform_queries(self, approach, dataset, file_selection=('/*.jsonl', '/*.jsonl.gz')): + def transform_queries(self, approach:str, dataset:str, file_selection: Tuple[str,...]=('/*.jsonl', '/*.jsonl.gz')): + """Load and transform the query processing outputs specified by the approach on the dataset for direct re-use as a PyTerrier query transformation. + + Args: + approach str: the approach for which the run should be loaded, in the format 'task/team/software'. + dataset (str):the dataset id, either an tira or ir_datasets id. + file_selection (Tuple[str,...], optional): The search glob to outputs specified by the approach on the dataset. Defaults to ('/*.jsonl', '/*.jsonl.gz'). + + Raises: + ValueError: If no approach with the identifier 'approach' was found or if there was an error parsing the outputs. + + Returns: + pd.DataFrame: a DataFrame with the parsed query processing outputs compatible with PyTerrier query transformations. + """ import pandas as pd matching_files = self.__matching_files(approach, dataset, file_selection) @@ -57,6 +86,19 @@ def transform_queries(self, approach, dataset, file_selection=('/*.jsonl', '/*.j return ret def transform_documents(self, approach, dataset, file_selection=('/*.jsonl', '/*.jsonl.gz')): + """Load and transform the document processing outputs specified by the approach on the dataset for direct re-use as a PyTerrier document transformation. + + Args: + approach str: the approach for which the run should be loaded, in the format 'task/team/software'. + dataset (str):the dataset id, either an tira or ir_datasets id. + file_selection (Tuple[str,...], optional): The search glob to outputs specified by the approach on the dataset. Defaults to ('/*.jsonl', '/*.jsonl.gz'). + + Raises: + ValueError: If no approach with the identifier 'approach' was found or if there was an error parsing the outputs. + + Returns: + pd.DataFrame: a DataFrame with the parsed document processing outputs compatible with PyTerrier document transformations. + """ import pandas as pd matching_files = self.__matching_files(approach, dataset, file_selection) if len(matching_files) == 0: @@ -68,3 +110,60 @@ def transform_documents(self, approach, dataset, file_selection=('/*.jsonl', '/* ret['docno'] = ret['doc_id'] del ret['doc_id'] return ret + + def __matching_dataset_files(self, task, dataset, truth_dataset, file_selection): + from glob import glob + ret = [] + local_dir = self.tira_client.download_dataset(task, dataset, truth_dataset) + + for glob_entry in file_selection: + for i in glob(local_dir + glob_entry): + ret += [i] + + return ret + + def inputs(self, task, dataset, file_selection=('/*.jsonl', '/*.jsonl.gz'), dtype=PANDAS_DTYPES): + """Load the inputs to systems for a task from tira. + + Args: + approach str: the approach for which the run should be loaded, in the format 'task/team/software'. + dataset (str):the dataset id, either an tira or ir_datasets id. + file_selection (Tuple[str,...], optional): The search glob to outputs specified by the approach on the dataset. Defaults to ('/*.jsonl', '/*.jsonl.gz'). + dtype (Tuple[str,...], optional): Transformations of the data types while loading with pandas. Defaults to PANDAS_DTYPES. + + Raises: + ValueError: If the dataset is not public or does not exist. + + Returns: + pd.DataFrame: A DataFrame with all inputs to systems. + """ + import pandas as pd + matching_files = self.__matching_dataset_files(task, dataset, False, file_selection) + + if len(matching_files) == 0: + raise ValueError('Could not find a dataset output. Used file_selection: ' + str(file_selection) + '. Please specify the file_selection to resolve this.') + + return pd.read_json(matching_files[0], lines=True, dtype=dtype) + + def truths(self, task, dataset, file_selection=('/*.jsonl', '/*.jsonl.gz'), dtype=PANDAS_DTYPES): + """Load the truths, i.e., ground truth labels, for a task from tira. + + Args: + approach str: the approach for which the run should be loaded, in the format 'task/team/software'. + dataset (str):the dataset id, either an tira or ir_datasets id. + file_selection (Tuple[str,...], optional): The search glob to outputs specified by the approach on the dataset. Defaults to ('/*.jsonl', '/*.jsonl.gz'). + dtype (Tuple[str,...], optional): Transformations of the data types while loading with pandas. Defaults to PANDAS_DTYPES. + + Raises: + ValueError: If the truth is not public or does not exist. + + Returns: + pd.DataFrame: A DataFrame with all the ground truth labels. + """ + import pandas as pd + matching_files = self.__matching_dataset_files(task, dataset, True, file_selection) + + if len(matching_files) == 0: + raise ValueError('Could not find a dataset output. Used file_selection: ' + str(file_selection) + '. Please specify the file_selection to resolve this.') + + return pd.read_json(matching_files[0], lines=True, dtype=dtype) diff --git a/python-client/tira/pyterrier_integration.py b/python-client/tira/pyterrier_integration.py index 4b495d76f..ef2686ed1 100644 --- a/python-client/tira/pyterrier_integration.py +++ b/python-client/tira/pyterrier_integration.py @@ -1,4 +1,6 @@ import os +from pathlib import Path + class PyTerrierIntegration(): def __init__(self, tira_client): @@ -99,7 +101,8 @@ def index(self, approach, dataset): Load an PyTerrier index from TIRA. """ import pyterrier as pt - ret = self.tira_client.get_run_output(approach, dataset) + '/index' + from tira.ir_datasets_util import translate_irds_id_to_tirex + ret = self.tira_client.get_run_output(approach, translate_irds_id_to_tirex(dataset)) + '/index' return pt.IndexFactory.of(os.path.abspath(ret)) def from_submission(self, approach, dataset=None, datasets=None): @@ -111,10 +114,9 @@ def from_submission(self, approach, dataset=None, datasets=None): return TiraRerankingTransformer(approach, self.tira_client, dataset, datasets) def from_retriever_submission(self, approach, dataset, previous_stage=None, datasets=None): - import pyterrier as pt + from tira.pyterrier_util import TiraSourceTransformer ret = self.pd.from_retriever_submission(approach, dataset, previous_stage, datasets) - - return pt.Transformer.from_df(ret) + return TiraSourceTransformer(ret) def transform_queries(self, approach, dataset, file_selection=('/*.jsonl', '/*.jsonl.gz'), prefix=''): from pyterrier.apply import generic @@ -143,7 +145,85 @@ def __transform_df(df): return generic(__transform_df) + def doc_features(self, approach, dataset, file_selection=('/*.jsonl', '/*.jsonl.gz')): + import numpy as np + from pyterrier.apply import doc_features + ret = self.pd.transform_documents(approach, dataset, file_selection) + cols = [i for i in ret.columns if i not in ['docno']] + ret = {str(i['docno']): np.array([i[c] for c in cols]) for _, i in ret.iterrows()} + + return doc_features(lambda row: ret[str(row['docno'])]) + + def query_features(self, approach, dataset, file_selection=('/*.jsonl', '/*.jsonl.gz')): + import numpy as np + from pyterrier.transformer import Transformer + ret = self.pd.transform_queries(approach, dataset, file_selection) + cols = [i for i in ret.columns if i not in ['qid']] + ret = {str(i['qid']): np.array([i[c] for c in cols]) for _, i in ret.iterrows()} + + class ApplyQueryFeatureTransformer(Transformer): + def __repr__(self): + return "tira.pt.query_features()" + + def transform(self, inputRes): + outputRes = inputRes.copy() + + outputRes["features"] = outputRes.apply(lambda i: ret[str(i['qid'])], axis=1) + return outputRes + + return ApplyQueryFeatureTransformer() + def reranker(self, approach, irds_id=None): from tira.pyterrier_util import TiraLocalExecutionRerankingTransformer return TiraLocalExecutionRerankingTransformer(approach, self.tira_client, irds_id=irds_id) +class PyTerrierAnceIntegration(): + """The pyterrier_ance integration to re-use cached ANCE indices. Wraps https://github.com/terrierteam/pyterrier_ance + """ + def __init__(self, tira_client): + self.tira_client = tira_client + + def ance_retrieval(self, dataset:str): + """Load a cached pyterrier_ance.ANCEIndexer submitted as workshop-on-open-web-search/ows/pyterrier-anceindex from tira. + + References (for citation): + https://arxiv.org/pdf/2007.00808.pdf + https://github.com/microsoft/ANCE/ + + Args: + dataset (str): the dataset id, either an tira or ir_datasets id. + + Returns: + pyterrier_ance.ANCERetrieval: the ANCE index. + """ + from tira.ir_datasets_util import translate_irds_id_to_tirex + ance_index = Path(self.tira_client.get_run_output('ir-lab-sose-2024/ows/pyterrier-anceindex', translate_irds_id_to_tirex(dataset))) / 'anceindex' + ance_checkpoint = self.tira_client.load_resource('Passage_ANCE_FirstP_Checkpoint.zip') + import pyterrier_ance + return pyterrier_ance.ANCERetrieval(ance_checkpoint, ance_index) + + +class PyTerrierSpladeIntegration(): + """The pyt_splade integration to re-use cached Splade indices. Wraps https://github.com/cmacdonald/pyt_splade + """ + def __init__(self, tira_client): + self.tira_client = tira_client + + def splade_index(self, dataset:str, approach: str='workshop-on-open-web-search/naverlabseurope/Splade (Index)'): + """Load a cached pyt_splade index submitted as the passed approach (default 'workshop-on-open-web-search/naverlabseurope/Splade (Index)') from tira. + + References (for citation): + https://github.com/naver/splade?tab=readme-ov-file#cite-scroll + ToDo: Ask Thibault what to cite. + + Args: + dataset (str): the dataset id, either an tira or ir_datasets id. + approach (str, optional): the approach id, defaults 'workshop-on-open-web-search/naverlabseurope/Splade (Index)'. + + Returns: + The PyTerrier index suitable for retrieval. + """ + from tira.ir_datasets_util import translate_irds_id_to_tirex + import pyterrier as pt + ret = Path(self.tira_client.get_run_output('ir-lab-sose-2024/naverlabseurope/Splade (Index)', translate_irds_id_to_tirex(dataset))) / 'spladeindex' + return pt.IndexFactory.of(os.path.abspath(ret)) diff --git a/python-client/tira/pyterrier_util.py b/python-client/tira/pyterrier_util.py index 03b512a90..fa46ca3bf 100644 --- a/python-client/tira/pyterrier_util.py +++ b/python-client/tira/pyterrier_util.py @@ -1,4 +1,4 @@ -from pyterrier.transformer import Transformer +from pyterrier.transformer import Transformer, SourceTransformer import json import pandas as pd import tempfile @@ -25,6 +25,29 @@ def merge_runs(topics, run_file): return topics[keeping].merge(df, how='left', left_on=join_on, right_on=join_on) +class TiraSourceTransformer(SourceTransformer): + def __init__(self, rtr, **kwargs): + super().__init__(rtr, **kwargs) + + + def transform(self, topics): + import numpy as np + if 'docno' not in topics.columns: + return super().transform(topics) + elif 'qid' not in topics.columns: + raise ValueError('The dataframe needs to have a column "qid"') + + keeping = topics.columns + + common_columns = np.intersect1d(topics.columns, self.df.columns) + + # we drop columns in topics that exist in the self.df + drop_columns = [i for i in common_columns if i not in ("qid", "docno")] + if len(drop_columns) > 0: + keeping = topics.columns[~ topics.columns.isin(drop_columns)] + + return topics[keeping].merge(self.df, on=["qid", "docno"]) + class TiraFullRankTransformer(Transformer): """ A Transformer that re-executes some full-rank approach submitted to a shared task in TIRA. @@ -69,6 +92,7 @@ def __init__(self, approach, tira_client, dataset=None, datasets=None, **kwargs) def transform(self, topics): import numpy as np assert "qid" in topics.columns + if 'tira_task' not in topics.columns or 'tira_dataset' not in topics.columns or 'tira_first_stage_run_id' not in topics.columns: if self.datasets: tira_configurations = [{'tira_task': self.task, 'tira_dataset': i, 'tira_first_stage_run_id': None} for i in self.datasets] diff --git a/python-client/tira/rest_api_client.py b/python-client/tira/rest_api_client.py index 9b87b2c6f..a822450ad 100644 --- a/python-client/tira/rest_api_client.py +++ b/python-client/tira/rest_api_client.py @@ -7,14 +7,14 @@ import io import time from random import randint -from tira.pyterrier_integration import PyTerrierIntegration +from tira.pyterrier_integration import PyTerrierIntegration, PyTerrierAnceIntegration, PyTerrierSpladeIntegration from tira.pandas_integration import PandasIntegration from tira.local_execution_integration import LocalExecutionIntegration import logging from .tira_client import TiraClient from typing import Optional, List, Dict, Union from functools import lru_cache -from tira.tira_redirects import redirects, mirror_url, dataset_ir_redirects +from tira.tira_redirects import redirects, mirror_url, dataset_ir_redirects, RESOURCE_REDIRECTS from tqdm import tqdm import hashlib @@ -38,6 +38,8 @@ def __init__(self, base_url: Optional[str]=None, api_key: str=None, failsave_ret self.fail_if_api_key_is_invalid() self.pd = PandasIntegration(self) self.pt = PyTerrierIntegration(self) + self.pt_ance = PyTerrierAnceIntegration(self) + self.pt_splade = PyTerrierSpladeIntegration(self) self.local_execution = LocalExecutionIntegration(self) self.failsave_retries = failsave_retries @@ -45,7 +47,12 @@ def __init__(self, base_url: Optional[str]=None, api_key: str=None, failsave_ret def load_settings(self): try: - return json.load(open(self.tira_cache_dir + '/.tira-settings.json', 'r')) + ret = json.load(open(self.tira_cache_dir + '/.tira-settings.json', 'r')) + if 'api_key' not in ret: + ret['api_key'] = 'no-api-key' + if 'api_user_name' not in ret: + ret['api_user_name'] = 'no-api-key-user' + return ret except Exception: logging.info(f'No settings given in {self.tira_cache_dir}/.tira-settings.json. I will use defaults.') return {'api_key': 'no-api-key', 'api_user_name': 'no-api-key-user'} @@ -53,6 +60,7 @@ def load_settings(self): def update_settings(self, k, v): settings = self.load_settings() settings[k] = v + os.makedirs(self.tira_cache_dir, exist_ok=True) json.dump(settings, open(Path(self.tira_cache_dir) / '.tira-settings.json', 'w+')) if k == 'api_key': @@ -61,10 +69,10 @@ def update_settings(self, k, v): def api_key_is_valid(self): role = self.json_response('/api/role') - if (self.api_user_name is None or self.api_user_name == 'no-api-key-user') and role and 'context' in role and 'user_id' in role['context']: + if (self.api_user_name is None or self.api_user_name == 'no-api-key-user') and role and 'context' in role and 'user_id' in role['context'] and role['context']['user_id']: self.api_user_name = role['context']['user_id'] - return role and 'status' in role and 'role' in role and 0 == role['status'] + return role and 'status' in role and 'role' in role and 0 == role['status'] and 'user_id' in role['context'] and role['context']['user_id'] and 'role' in role['context'] and role['context']['role'] and 'guest' != role['context']['role'] def fail_if_api_key_is_invalid(self): if not self.api_key_is_valid(): @@ -123,6 +131,8 @@ def add_docker_software(self, image, command, tira_vm_id, tira_task_id, code_rep ret = ret.content.decode('utf8') ret = json.loads(ret) assert ret['status'] == 0 + + print(f'Software with name {ret["context"]["display_name"]} was created.') logging.info(f'Software with name {ret["context"]["display_name"]} was created.') logging.info(f'Please visit {self.base_url}/submit/{tira_task_id}/user/{tira_vm_id}/docker-submission to run your software.') @@ -211,6 +221,29 @@ def evaluations(self, task, dataset, join_submissions=True): def run_was_already_executed_on_dataset(self, approach, dataset): return self.get_run_execution_or_none(approach, dataset) is not None + def load_resource(self, resource:str): + """Load a resource (usually a zip) from TIRA/Zenodo. Serves as utikity function in case some additional resources must be loaded. + + Args: + resource (str): The resource identifier + + Raises: + ValueError: If the resource is not known. + + Returns: + str: The path to the downloaded resource. + """ + target_file = f'{self.tira_cache_dir}/raw_resources/{resource}' + if os.path.exists(target_file): + return target_file + + if resource not in RESOURCE_REDIRECTS: + raise ValueError(f'Resource {resource} not supported.') + + os.makedirs(f'{self.tira_cache_dir}/raw_resources', exist_ok=True) + self.download_and_extract_zip(RESOURCE_REDIRECTS[resource], target_file, extract=False) + return target_file + def get_run_output(self, approach, dataset, allow_without_evaluation=False): """ Downloads the run (or uses the cached version) of the specified approach on the specified dataset. @@ -220,7 +253,10 @@ def get_run_output(self, approach, dataset, allow_without_evaluation=False): if mounted_output_in_sandbox: return mounted_output_in_sandbox - task, team, software = approach.split('/') + task, team, software = approach.split('/') + if '/' in dataset: + dataset = dataset.split('/')[-1] + run_execution = self.get_run_execution_or_none(approach, dataset) if run_execution: @@ -276,6 +312,8 @@ def get_run_execution_or_none(self, approach, dataset, previous_stage_run_id=Non return ret[['task', 'dataset', 'team', 'run_id']].iloc[0].to_dict() def download_run(self, task, dataset, software, team=None, previous_stage=None, return_metadata=False): + if '/' in dataset: + dataset = dataset.split('/')[-1] ret = self.get_run_execution_or_none(f'{task}/{team}/{software}', dataset, previous_stage) if not ret: raise ValueError(f'I could not find a run for the filter criteria task="{task}", dataset="{dataset}", software="{software}", team={team}, previous_stage={previous_stage}') @@ -308,6 +346,8 @@ def download_dataset(self, task, dataset, truth_dataset=False): """ if 'TIRA_INPUT_DATASET' in os.environ: return os.environ['TIRA_INPUT_DATASET'] + if '/' in dataset: + dataset = dataset.split('/')[-1] dataset = dataset_ir_redirects(dataset) @@ -383,7 +423,7 @@ def evaluate_run(self, team, dataset, run_id): return ret - def download_and_extract_zip(self, url, target_dir): + def download_and_extract_zip(self, url, target_dir, extract=True): url = redirects(url=url)['urls'][0] if url.split('://')[1].startswith('files.webis.de'): print(f'Download from the Incubator: {url}') @@ -413,10 +453,16 @@ def download_and_extract_zip(self, url, target_dir): for data in r.iter_content(chunk_size=1024): size = response_content.write(data) bar.update(size) - print('Download finished. Extract...') - z = zipfile.ZipFile(response_content) - z.extractall(target_dir) - print('Extraction finished: ', target_dir) + if extract: + print('Download finished. Extract...') + z = zipfile.ZipFile(response_content) + z.extractall(target_dir) + print('Extraction finished: ', target_dir) + else: + print('Download finished. Persist...') + with open(target_dir, 'wb') as file_out: + file_out.write(response_content.getbuffer()) + print('Download finished: ', target_dir) return except Exception as e: @@ -517,6 +563,10 @@ def upload_run(self, file_path: Path, dataset_id: str, approach: str=None, task_ self.fail_if_api_key_is_invalid() + if file_path is None or not file_path.is_file(): + logging.warn(f'The passed file {file_path} does not exist.') + raise ValueError(f'The passed file {file_path} does not exist.') + # TODO: check that task_id and vm_id don't contain illegal characters (e.g., '/') url = f"/task/{task_id}/vm/{vm_id}/upload/{dataset_id}/{upload_id}" logging.info(f"Submitting the runfile at {url}") diff --git a/python-client/tira/third_party_integrations.py b/python-client/tira/third_party_integrations.py index a1a074264..4f5fcf1a8 100644 --- a/python-client/tira/third_party_integrations.py +++ b/python-client/tira/third_party_integrations.py @@ -3,6 +3,7 @@ from tira.io_utils import all_lines_to_pandas import logging from pathlib import Path +import shlex def ensure_pyterrier_is_loaded(boot_packages=("com.github.terrierteam:terrier-prf:-SNAPSHOT", ), packages=(), patch_ir_datasets=True): @@ -150,13 +151,16 @@ def extract_to_be_executed_notebook_from_command_or_none(command:str): return command.split('--notebook')[1].strip().split(' ')[0].strip() if command is not None and '.py' in command: - for arg in command.split(' '): - if arg.endswith('.py'): + for arg in shlex.split(command, posix=False): + if arg.endswith('.py') or arg.endswith('.py"') or arg.endswith('.py\''): return arg if command is not None: - command = command.split(' ')[0] - if command.endswith('.sh') or command.endswith('.bash'): + command = shlex.split(command, posix=False) + if len(command) > 0: + command = command[0] + + if command and (command.endswith('.sh') or command.endswith('.bash') or command.endswith('.sh"') or command.endswith('.sh\'') or command.endswith('.bash"') or command.endswith('.bash\'')): return command return None @@ -263,6 +267,7 @@ def extract_previous_stages_from_docker_image(image:str, command:str = None): if notebook is None: return [] + notebook = shlex.split(notebook)[0] tira = Client() @@ -276,6 +281,11 @@ def extract_previous_stages_from_docker_image(image:str, command:str = None): return extract_previous_stages_from_notebook(Path(local_file)) def load_ir_datasets(): + try: + from ir_datasets.datasets.base import Dataset + except: + return None + # Detect if we are in the TIRA sandbox if 'TIRA_INPUT_DATASET' in os.environ: from tira.ir_datasets_util import static_ir_dataset diff --git a/python-client/tira/tira_cli.py b/python-client/tira/tira_cli.py index ebb3668df..a582866c7 100644 --- a/python-client/tira/tira_cli.py +++ b/python-client/tira/tira_cli.py @@ -23,6 +23,7 @@ def parse_args(): parser_login = subparsers.add_parser('login', help='Login your TIRA client to the tira server.') parser_login.add_argument('--token', required=True, default=None, help='The token to login to the server.') + parser_login.add_argument('--print-docker-auth', required=False, default=False, action='store_true', help='Print the docker credentials as TIRA_DOCKER_REGISTRY_USER and TIRA_DOCKER_REGISTRY_TOKEN to stdout.') args = parser.parse_args() @@ -45,3 +46,5 @@ def main(): print(client.download_dataset(None, args.dataset)) if args.command == 'login': client.login(args.token) + if args.print_docker_auth: + print(client.local_execution.docker_client_is_authenticated()) diff --git a/python-client/tira/tira_redirects.py b/python-client/tira/tira_redirects.py index 761c8a84d..cf0bd06bc 100644 --- a/python-client/tira/tira_redirects.py +++ b/python-client/tira/tira_redirects.py @@ -1,5 +1,4 @@ - - +from tira.license_agreements import print_license_agreement STATIC_REDIRECTS = { 'ir-benchmarks': { @@ -201,6 +200,66 @@ } } } + }, + + "ir-lab-sose-2024": { + "tira-ir-starter": { + "Index (tira-ir-starter-pyterrier)": { + "ir-acl-anthology-20240411-training": { + "run_id": "2024-04-11-19-43-23", + "md5": "ebb5b8f1d8c7ad36612f408da1203ff2", + "urls": ["https://files.webis.de/data-in-production/data-research/tira-zenodo-dump-preparation/ir-lab-sose2024/2024-04-11-19-43-23.zip"] + } + }, + "Index (pyterrier-stanford-lemmatizer)": { + "ir-acl-anthology-20240411-training": { + "run_id": "2024-04-16-11-05-06", + "md5": "70aa7ce8a437ba47ebf9e17d75871a5f", + "urls": ["https://files.webis.de/data-in-production/data-research/tira-zenodo-dump-preparation/ir-lab-sose2024/2024-04-16-11-05-06.zip"] + } + } + }, + "seanmacavaney": { + "DocT5Query": { + "anthology-20240411-training": { + "run_id": "2024-04-09-22-03-26", + "md5": "1a0f65a8f47051db435ec6031106db7b", + "urls": ["https://files.webis.de/data-in-production/data-research/tira-zenodo-dump-preparation/ir-lab-sose2024/2024-04-09-22-03-26.zip"], + } + }, + "corpus-graph": { + "anthology-20240411-training": { + "run_id": "2024-04-09-16-35-50", + "md5": "75f743914d44d7252425844136ea9722", + "urls": ["https://files.webis.de/data-in-production/data-research/tira-zenodo-dump-preparation/ir-lab-sose2024/2024-04-09-16-35-50.zip"], + } + }, + #"some-interface": { + #copy of corpus-graph + # "anthology-20240411-training": { + # "run_id": "2024-04-09-16-35-50", + # "md5": "75f743914d44d7252425844136ea9722" + # } + #} + }, + "ows": { + "pyterrier-anceindex": { + "ir-acl-anthology-20240411-training": { + "run_id": "2024-04-11-19-47-18", + "md5": "1112dc9e60b6dfa4ac80571ad3425200", + "urls": ["https://files.webis.de/data-in-production/data-research/tira-zenodo-dump-preparation/ir-lab-sose2024/2024-04-11-19-47-18.zip"], + } + } + }, + "naverlabseurope": { + "Splade (Index)": { + "ir-acl-anthology-20240411-training": { + "run_id": "2024-04-14-08-40-58", + "md5": "03f73b47664cc1c843529b948094bf81", + "urls": ["https://files.webis.de/data-in-production/data-research/tira-zenodo-dump-preparation/ir-lab-sose2024/2024-04-14-08-40-58.zip"], + } + } + } } } @@ -302,10 +361,20 @@ 'training-20231104-training': 'https://zenodo.org/records/10628882/files/', 'jena-topics-small-20240119-training': 'https://zenodo.org/records/10628882/files/', 'leipzig-topics-small-20240119-training': 'https://zenodo.org/records/10628882/files/', + 'ir-acl-anthology-20240411-training': 'https://files.webis.de/data-in-production/data-research/tira-zenodo-dump-preparation/ir-lab-sose2024/', + 'longeval-2023-06-20240418-training': 'https://files.webis.de/data-in-production/data-research/tira-zenodo-dump-preparation/ir-lab-padua2024/', + 'longeval-2023-08-20240418-training': 'https://files.webis.de/data-in-production/data-research/tira-zenodo-dump-preparation/ir-lab-padua2024/', +} + +RESOURCE_REDIRECTS = { + 'Passage_ANCE_FirstP_Checkpoint.zip': 'https://files.webis.de/data-in-production/data-research/tira-zenodo-dump-preparation/ir-lab-sose2024/Passage_ANCE_FirstP_Checkpoint.zip', + 'custom-terrier-token-processing-1.0-SNAPSHOT-jar-with-dependencies.jar': 'https://files.webis.de/data-in-production/data-research/tira-zenodo-dump-preparation/ir-lab-sose2024/custom-terrier-token-processing-1.0-SNAPSHOT-jar-with-dependencies.jar' } DATASET_ID_REDIRECTS = { - 'longeval-tiny-train-20240315-training': 'training-20231104-training' + 'longeval-tiny-train-20240315-training': 'training-20231104-training', + 'longeval-2023-06-20240422-training': 'longeval-2023-06-20240418-training', + 'longeval-2023-08-20240422-training': 'longeval-2023-08-20240418-training' } @@ -351,6 +420,8 @@ def mirror_url(url): def redirects(approach=None, dataset=None, url=None): default_ret = {'urls': [url]} if url is not None: + print_license_agreement(url) + if '/task/' in url and '/user/' in url and '/dataset/' in url and '/download/' in url and '.zip' in url: #/task/{task}/user/{team}/dataset/{dataset}/download/{run_id}.zip ret = url.split('/task/')[1] diff --git a/python-client/tira/tira_run.py b/python-client/tira/tira_run.py index 1353feb96..ab6002d40 100755 --- a/python-client/tira/tira_run.py +++ b/python-client/tira/tira_run.py @@ -51,7 +51,8 @@ def parse_args(): group.add_argument('--export-submission-from-jupyter-notebook', required=False, default=None, type=str) group.add_argument('--export-submission-environment', nargs='*', required=False, default=None, type=str) parser.add_argument('--command', required=False) - parser.add_argument('--verbose', required=False, default=False, type=bool) + parser.add_argument('--verbose', required=False, default=False, action='store_true') + parser.add_argument('--gpus', required=False, default=False, type=str, help='GPU devices to add to the container ("all" or -1 to pass all GPUs, or the number of devices).') parser.add_argument('--evaluate', required=False, default=False, type=bool) parser.add_argument('--evaluation-directory', required=False, default=str(os.path.abspath("tira-evaluation"))) parser.add_argument('--dry-run', required=False, default=False, type=bool) @@ -98,6 +99,7 @@ def parse_args(): args.previous_stages = [] if not args.input_run else [args.input_run] if args.input_run is None and args.input_run_directory is None: args.input_run = extract_previous_stages_from_docker_image(args.image, args.command) + args.command = LocalExecutionIntegration().make_command_absolute(args.image, args.command) args.previous_stages = args.input_run if args.input_run and len(args.input_run) == 1: args.input_run = args.input_run[0] @@ -112,8 +114,17 @@ def parse_args(): if args.tira_client_token: if api_key_valid: args.tira_client_token = rest_client.api_key + elif 'TIRA_CLIENT_TOKEN' in os.environ and os.environ.get('TIRA_CLIENT_TOKEN'): + rest_client = RestClient(api_key=os.environ.get('TIRA_CLIENT_TOKEN')) + api_key_valid = rest_client.api_key_is_valid() + if api_key_valid: + args.tira_client_token = rest_client.api_key + else: + parser.error('The option --tira-client-token (or environment variable TIRA_CLIENT_TOKEN) is required when --push is active.') else: - parser.error('The option --tira-client-token (or environment variable TIRA_CLIENT_TOKEN) is required when --push is active.') + rest_client = RestClient(api_key=args.tira_client_token) + if not rest_client.api_key_is_valid(): + parser.error('The option --tira-client-token (or environment variable TIRA_CLIENT_TOKEN) is required when --push is active.') if not args.tira_task_id and args.input_dataset: args.tira_task_id = args.input_dataset.split('/')[0] @@ -253,7 +264,11 @@ def main(args=None): # command={args.command} ############################################################################################################################ ''') - client.local_execution.run(identifier=args.approach, image=args.image, command=args.command, input_dir=input_dir, output_dir=args.output_directory, dry_run=args.dry_run, allow_network=args.allow_network, input_run=args.input_run, additional_volumes=args.v, evaluate=evaluate, eval_dir=args.evaluation_directory) + gpus = 0 + if args.gpus: + gpus = -1 if args.gpus == 'all' else int(args.gpus) + + client.local_execution.run(identifier=args.approach, image=args.image, command=args.command, input_dir=input_dir, output_dir=args.output_directory, dry_run=args.dry_run, allow_network=args.allow_network, input_run=args.input_run, additional_volumes=args.v, evaluate=evaluate, eval_dir=args.evaluation_directory, gpu_count=gpus) if args.fail_if_output_is_empty and (not os.path.exists(args.output_directory) or not os.listdir(args.output_directory)): raise ValueError('The software produced an empty output directory, it likely failed?') @@ -271,11 +286,12 @@ def main(args=None): for f in os.listdir(args.output_directory): print(f' - {f}: {os.path.getsize(args.output_directory + "/" + f)} bytes') - continue_user = input("Are the outputs correct and should I push the software to TIRA? (y/N) ").lower() + if not args.fail_if_output_is_empty: + continue_user = input("Are the outputs correct and should I push the software to TIRA? (y/N) ").lower() - if continue_user and continue_user.lower() not in ['y', 'yes']: - print('You did not specify yes, I will not push the software.') - return + if continue_user and continue_user.lower() not in ['y', 'yes']: + print('You did not specify yes, I will not push the software.') + return registry_prefix = client.docker_registry() + '/code-research/tira/tira-user-' + args.tira_vm_id + '/' print('Push Docker image')

No flag for --gpu in tira-run for local tests **Current behavior** **Expected behavior**

2024-04-22T22:35:02

0.0

[]

tira-io/tira

tira-io__tira-584

3971c30c3d3aaa741e01387d5ced757754bc71c4

diff --git a/.devcontainer.json b/.devcontainer.json new file mode 100644 index 000000000..38b68dece --- /dev/null +++ b/.devcontainer.json @@ -0,0 +1,8 @@ +{ + "image": "webis/tira-application:basis-0.0.95", + "customizations": { + "vscode": { + "extensions": ["ms-python.python", "ms-python.vscode-pylance", "ms-toolsai.jupyter"] + } + } +} diff --git a/README.md b/README.md index 535654164..c1d1bbb73 100644 --- a/README.md +++ b/README.md @@ -12,7 +12,7 @@ Components: ## Setup Your Local Development Environment -We use [dev containers](https://code.visualstudio.com/docs/devcontainers/containers) to simplify development. Please install Docker and an IDE with support for dev containers on your machine (we usually use VS Code). +We use [dev containers](https://code.visualstudio.com/docs/devcontainers/containers) to simplify development. You can [open this repository directly in codespaces](https://github.com/codespaces/new/webis-de/tira-io/tree/tira). If you want to develop locally, please install Docker and an IDE with support for dev containers on your machine (we usually use VS Code). First, please clone the repository: ``` @@ -21,7 +21,7 @@ git clone [email protected]:tira-io/tira.git Please open the directory `application` in VS Code, and confirm to use the provided dev container. -If you want to work on production data, please ensure that you can login to ssh.webis.de, and then do the following: +If you want to work on production data (not necessary in most cases, you usually can skip this step), please ensure that you can login to ssh.webis.de, and then do the following: ``` make import-data-from-dump diff --git a/application/.devcontainer.json b/application/.devcontainer.json index b290a47f2..38b68dece 100644 --- a/application/.devcontainer.json +++ b/application/.devcontainer.json @@ -1,5 +1,5 @@ { - "image": "webis/tira-application:basis-0.0.94", + "image": "webis/tira-application:basis-0.0.95", "customizations": { "vscode": { "extensions": ["ms-python.python", "ms-python.vscode-pylance", "ms-toolsai.jupyter"] diff --git a/application/Dockerfile.application b/application/Dockerfile.application index e83bce614..c4e859015 100644 --- a/application/Dockerfile.application +++ b/application/Dockerfile.application @@ -9,7 +9,7 @@ RUN cd /src/tira/frontend-vuetify \ # Only change in case of new / updated dependencies -FROM webis/tira-application:basis-0.0.94 +FROM webis/tira-application:basis-0.0.95 # This Dockerfile ensures that all dependencies do rarely change by starting from a basis image # that contains already all dependencies (so that the minor versions do rarely change, but we diff --git a/application/Dockerfile.application-dev b/application/Dockerfile.application-dev index 9b1916730..f7850169b 100644 --- a/application/Dockerfile.application-dev +++ b/application/Dockerfile.application-dev @@ -1,4 +1,4 @@ -# docker build -t webis/tira-application:basis-0.0.94 -f Dockerfile.application-dev . +# docker build -t webis/tira-application:basis-0.0.95 -f Dockerfile.application-dev . FROM ubuntu:22.04 RUN apt-get update \ @@ -53,3 +53,24 @@ RUN chown -R tira:tira /usr/local && \ RUN pip3 install tira==0.0.97 ir-measures==0.3.1 +RUN apt-get install -y wget curl + +RUN wget 'https://nodejs.org/dist/v20.10.0/node-v20.10.0-linux-x64.tar.xz' \ + && tar -xf node-v20.10.0-linux-x64.tar.xz \ + && rm node-v20.10.0-linux-x64.tar.xz \ + && mv node-v20.10.0-linux-x64/bin/* /usr/local/bin/ \ + && mv node-v20.10.0-linux-x64/lib/node_modules/ /usr/local/lib/ \ + && npm install --global yarn + +ADD src/tira/frontend-vuetify/package.json src/tira/frontend-vuetify/yarn.lock src/tira/frontend-vuetify/vite.config.ts src/tira/frontend-vuetify/jest.config.js src/tira/frontend-vuetify/babel.config.js /tmp-del/ + +RUN cd /tmp-del \ + && yarn create vuetify \ + && yarn install \ + && yarn install --dev \ + && mv node_modules/* /usr/local/lib/node_modules \ + && mv node_modules/.bin /usr/local/lib/node_modules/.bin \ + && mv node_modules/.yarn-integrity /usr/local/lib/node_modules/.yarn-integrity \ + && cd / \ + && rm -R /tmp-del/ + diff --git a/application/Makefile b/application/Makefile index cbdff5cf5..13972707f 100644 --- a/application/Makefile +++ b/application/Makefile @@ -2,8 +2,8 @@ VENV_NAME?=venv PYTHON=${PWD}/${VENV_NAME}/bin/python3 -VERSION_APPLICATION=0.0.94 -VERSION_GRPC=0.0.94 +VERSION_APPLICATION=0.0.96 +VERSION_GRPC=0.0.96 .DEFAULT: help help: @@ -71,10 +71,18 @@ run-develop: && python src/manage.py run_develop vite-build: - docker run -v ${PWD}:/app --platform linux/amd64 --rm -ti -w /app/src/tira/frontend-vuetify --entrypoint yarn webis/tira:vuetify-dev-0.0.1 build + @cd src/tira/frontend-vuetify \ + && yarn build vite-build-light: - docker run -v ${PWD}:/app --platform linux/amd64 --rm -ti -w /app/src/tira/frontend-vuetify --entrypoint yarn webis/tira:vuetify-dev-0.0.1 build-light + @cd src/tira/frontend-vuetify \ + && yarn build-light + +vite-build-docker: + docker run -v ${PWD}:/app --platform linux/amd64 --rm -ti -w /app/src/tira/frontend-vuetify --entrypoint yarn webis/tira-application:basis-0.0.95 build + +vite-build-light-docker: + docker run -v ${PWD}:/app --platform linux/amd64 --rm -ti -w /app/src/tira/frontend-vuetify --entrypoint yarn webis/tira-application:basis-0.0.95 build-light run-docker: diff --git a/application/src/django_admin/settings.py b/application/src/django_admin/settings.py index 4fa6107a6..f88e7bd0f 100644 --- a/application/src/django_admin/settings.py +++ b/application/src/django_admin/settings.py @@ -339,7 +339,11 @@ def logger_config(log_dir: Path): DISCOURSE_API_URL = 'https://www.tira.io' -CODE_SUBMISSION_REFERENCE_REPOSITORY = 'mam10eks/tira-software-submission-template' +CODE_SUBMISSION_REFERENCE_REPOSITORIES = { + 'ir-lab-jena-leipzig-wise-2023': 'tira-io/tira-ir-lab-wise-submission-template', + 'webpage-classification': 'OpenWebSearch/irixys23-tira-submission-template', + 'valueeval-2024-human-value-detection': 'touche-webis-de/valueeval24-tira-software-submission-template', +} CODE_SUBMISSION_REPOSITORY_NAMESPACE = 'tira-io' try: DISRAPTOR_API_KEY = open(DISRAPTOR_SECRET_FILE, "r").read().strip() diff --git a/application/src/tira/data/HybridDatabase.py b/application/src/tira/data/HybridDatabase.py index 3f37d8c5e..992884a72 100644 --- a/application/src/tira/data/HybridDatabase.py +++ b/application/src/tira/data/HybridDatabase.py @@ -19,6 +19,7 @@ logger = logging.getLogger("tira_db") +#SELECT tira_dockersoftware.vm_id, COUNT(DISTINCT(tira_dockersoftware.docker_software_id)) AS 'Software Count', SUM(tira_run.run_id IS NOT NULL) AS 'Executed Runs' FROM tira_dockersoftware LEFT JOIN tira_run ON tira_dockersoftware.docker_software_id = tira_run.docker_software_id where tira_dockersoftware.task_id = 'webpage-classification' GROUP BY tira_dockersoftware.vm_id; class HybridDatabase(object): """ @@ -1048,10 +1049,19 @@ def __link_to_code(build_environment): except: return None - if 'TIRA_JUPYTER_NOTEBOOK' not in build_environment or 'GITHUB_REPOSITORY' not in build_environment or 'GITHUB_WORKFLOW' not in build_environment or 'GITHUB_SHA' not in build_environment: + if 'GITHUB_REPOSITORY' not in build_environment or 'GITHUB_WORKFLOW' not in build_environment or 'GITHUB_SHA' not in build_environment: return None - if build_environment['GITHUB_WORKFLOW'] == ".github/workflows/upload-notebook-submission.yml": + if build_environment['GITHUB_WORKFLOW'] == "Upload Docker Software to TIRA": + if 'TIRA_DOCKER_PATH' not in build_environment: + return None + + return f'https://github.com/{build_environment["GITHUB_REPOSITORY"]}/tree/{build_environment["GITHUB_SHA"]}/{build_environment["TIRA_DOCKER_PATH"]}' + + if build_environment['GITHUB_WORKFLOW'] == ".github/workflows/upload-notebook-submission.yml" or build_environment['GITHUB_WORKFLOW'] == 'Upload Notebook to TIRA': + if 'TIRA_JUPYTER_NOTEBOOK' not in build_environment: + return None + return f'https://github.com/{build_environment["GITHUB_REPOSITORY"]}/tree/{build_environment["GITHUB_SHA"]}/jupyter-notebook-submissions/{build_environment["TIRA_JUPYTER_NOTEBOOK"]}' return None diff --git a/application/src/tira/endpoints/data_api.py b/application/src/tira/endpoints/data_api.py index fa4892bbe..c8fdf6d89 100644 --- a/application/src/tira/endpoints/data_api.py +++ b/application/src/tira/endpoints/data_api.py @@ -439,6 +439,34 @@ def tirex_components(request, context): context['tirex_components'] = settings.TIREX_COMPONENTS return JsonResponse({'status': 0, 'context': context}) +def get_snippet_to_run_components(request): + all_components = settings.TIREX_COMPONENTS + component_ids = request.GET.get('components', 'false') + + # All links with display_name == "Submission in TIREx" have the ID of the component in their link, its a bit ugly, but at the moment we need to extract the ID from there. + # E.g., the ID from the URL "/submissions/ir-benchmarks/ows/query-segmentation-hyp-a" would be ir-benchmarks/ows/query-segmentation-hyp-a + + # Also Ugly: we need to determine which type of processor (query processor, document processor, etc) something is by using its top-level category, e.g., "Query Processing". + + # I think it makes sense to build a small method that uses the settings.TIREX_COMPONENTS as input and produces a mapping form component ID (the thing below "Submission in TIREx") to the properties, e.g., query processor true or false, etc. + + # I think we can hard code everything against ROBUST04, we can switch this later. + dataset_initialization = 'dataset = pt.get_dataset("irds:disks45/nocr/trec-robust-2004")\n' + + additional_variables = '' + + # If we have a query processor, we need to add an additional variable "topics" + # just for this hard coded example: + current_component_is_query_processor = True + if current_component_is_query_processor: + additional_variables += "topics = dataset.get_topics(variant='title')\n" + + component_definitions = "tira.pt.transform_queries('ir-benchmarks/ows/query-segmentation-hyb-i', dataset)\n" + + snippet = (dataset_initialization + additional_variables + component_definitions).strip() + + return JsonResponse({'status': 0, 'context': {'snippet': snippet}}) + @add_context def reranking_datasets(request, context, task_id): diff --git a/application/src/tira/endpoints/vm_api.py b/application/src/tira/endpoints/vm_api.py index 2243e372b..157749a10 100644 --- a/application/src/tira/endpoints/vm_api.py +++ b/application/src/tira/endpoints/vm_api.py @@ -521,6 +521,7 @@ def add_software_submission_git_repository(request, task_id, vm_id): try: data = json.loads(request.body) external_owner = data['external_owner'] + private = not data.get('allow_public_repo', False) disraptor_user = get_disraptor_user(request, allow_unauthenticated_user=False) if not disraptor_user or not type(disraptor_user) == str: @@ -529,7 +530,8 @@ def add_software_submission_git_repository(request, task_id, vm_id): if not model.github_user_exists(external_owner): return JsonResponse({'status': 1, 'message': f"The user '{external_owner}' does not exist on Github, maybe a typo?"}) - software_submission_git_repo = model.get_submission_git_repo(vm_id, task_id, disraptor_user, external_owner) + software_submission_git_repo = model.get_submission_git_repo(vm_id, task_id, disraptor_user, external_owner, + private) return JsonResponse({'status': 0, "context": software_submission_git_repo}) except Exception as e: @@ -541,7 +543,7 @@ def add_software_submission_git_repository(request, task_id, vm_id): @check_permissions def get_software_submission_git_repository(request, task_id, vm_id): try: - if not model.load_docker_data(task_id, vm_id, cache, force_cache_refresh=False): + if task_id not in settings.CODE_SUBMISSION_REFERENCE_REPOSITORIES or not model.load_docker_data(task_id, vm_id, cache, force_cache_refresh=False): return JsonResponse({'status': 0, "context": {'disabled': True}}) return JsonResponse({'status': 0, "context": model.get_submission_git_repo(vm_id, task_id)}) diff --git a/application/src/tira/frontend-vuetify/src/components/RunReviewForm.vue b/application/src/tira/frontend-vuetify/src/components/RunReviewForm.vue index e506df51c..15d634d65 100644 --- a/application/src/tira/frontend-vuetify/src/components/RunReviewForm.vue +++ b/application/src/tira/frontend-vuetify/src/components/RunReviewForm.vue @@ -48,13 +48,13 @@ <v-divider class="py-6" v-if="!loading"/> <h2 v-if="!loading">Output</h2> - <v-tabs v-model="tab" align-tabs="title" v-if="!loading"> + <v-tabs v-model="tab" v-if="!loading" class="my-2"> <v-tab v-for="item in detailed_tabs" :key="item.key" :value="item.key" >{{ item.key }}</v-tab> </v-tabs> <v-window v-if="!loading" v-model="tab" :touch="{left: () => {}, right: () => {}}"> <v-window-item v-for="item in detailed_tabs" :key="item.key" :value="item.key"> <v-card flat> - <v-code tag="pre">{{item.content }}</v-code> + <code-snippet title="Click the icon to copy code" :code="item.content" expand_message=""/> </v-card> </v-window-item> </v-window> @@ -62,12 +62,13 @@ <script lang="ts"> import Loading from './Loading.vue' +import CodeSnippet from "@/components/CodeSnippet.vue"; import MetadataItems from './MetadataItems.vue' import { get, post, reportError, reportSuccess, inject_response, extractDatasetFromCurrentUrl } from '../utils' export default { name: "run-review-form", - components: { Loading, MetadataItems }, + components: { Loading, CodeSnippet, MetadataItems }, emits: ['review-run'], props: ['run_id', 'vm_id', 'dataset_id_from_props', ], data() { diff --git a/application/src/tira/frontend-vuetify/src/submission-components/CodeSubmission.vue b/application/src/tira/frontend-vuetify/src/submission-components/CodeSubmission.vue index cf3e3fc5e..549e24974 100644 --- a/application/src/tira/frontend-vuetify/src/submission-components/CodeSubmission.vue +++ b/application/src/tira/frontend-vuetify/src/submission-components/CodeSubmission.vue @@ -9,8 +9,12 @@ <v-card-item v-if="!loading && !disabled"> <v-card-text v-if="!repo_url"> <v-form ref="form" v-model="valid"> + + <v-row class="d-flex align-center justify-center"> + <v-col cols="6"><v-checkbox v-model="allow_public_repo" label="Allow Public Github Repository (You can change this later. Using a public Github repository has the advantage that Github Actions are free, otherwise, they count towards the bill of organizers.)" /></v-col> + </v-row> <v-row class="d-flex align-center justify-center"> - <v-col cols="6"><v-text-field v-model="new_git_account" label="Your GitHub Account" :rules="[v => !!(v && v.length) || 'Please select your GitHub account.']"/></v-col> + <v-col cols="6"><v-text-field @keydown.enter.prevent="connectToCodeRepo()" v-model="new_git_account" label="Your GitHub Account" :rules="[v => !!(v && v.length) || 'Please select your GitHub account.']"/></v-col> </v-row> <v-row class="d-flex align-center justify-center"> <v-col cols="6"><v-btn block color="primary" :loading="submit_in_progress" @click="connectToCodeRepo()" text="Add Code Repository"/></v-col> @@ -71,6 +75,7 @@ export default { return { loading: true, new_git_account: '', + allow_public_repo: true, valid: false, submit_in_progress: false, repo_url: '', @@ -90,12 +95,9 @@ export default { } this.submit_in_progress = true - post(`/api/add_software_submission_git_repository/${this.task_id}/${this.user_id}`, {"external_owner": this.new_git_account}) + post(`/api/add_software_submission_git_repository/${this.task_id}/${this.user_id}`, {"external_owner": this.new_git_account, "allow_public_repo": this.allow_public_repo}) .then(reportSuccess('Your git repository was created.')) - .then(message => { - this.repo_url = message.context.repo_url - this.owner_url = message.context.owner_url - }) + .then(inject_response(this)) .catch(reportError("Problem while adding your git repository.", "This might be a short-term hiccup, please try again. We got the following error: ")) .then(() => {this.submit_in_progress = false}) }, diff --git a/application/src/tira/git_runner_integration.py b/application/src/tira/git_runner_integration.py index 60c3f2d56..1c946ccb7 100644 --- a/application/src/tira/git_runner_integration.py +++ b/application/src/tira/git_runner_integration.py @@ -28,8 +28,16 @@ def normalize_file(file_content, tira_user_name, task_id): - return file_content.replace('TIRA_USER_FOR_AUTOMATIC_REPLACEMENT', tira_user_name)\ - .replace('TIRA_TASK_ID_FOR_AUTOMATIC_REPLACEMENT', task_id) + default_datasets = {'webpage-classification': 'webpage-classification/tiny-sample-20231023-training', + 'ir-lab-jena-leipzig-wise-2023': 'workshop-on-open-web-search/retrieval-20231027-training', + 'ir-lab-jena-leipzig-sose-2023': 'workshop-on-open-web-search/retrieval-20231027-training', + 'workshop-on-open-web-search': 'workshop-on-open-web-search/retrieval-20231027-training', + 'ir-benchmarks': 'workshop-on-open-web-search/retrieval-20231027-training', + } + + return file_content.replace('TIRA_USER_FOR_AUTOMATIC_REPLACEMENT', tira_user_name) \ + .replace('TIRA_TASK_ID_FOR_AUTOMATIC_REPLACEMENT', task_id) \ + .replace('TIRA_DATASET_FOR_AUTOMATIC_REPLACEMENT', default_datasets.get(task_id, '<TODO-ADD-DATASET>')) def convert_size(size_bytes): @@ -185,7 +193,7 @@ def initialize_user_repository(self, git_repository_id, repo_name, token): 'user_name': repo_name.replace('tira-user-', ''), 'repo_name': repo_name, 'token': token, - 'image_prefix': self.image_registry_prefix +'/' + repo_name +'/' + 'image_prefix': self.image_registry_prefix + '/' + repo_name + '/' }) with tempfile.TemporaryDirectory() as tmp_dir: @@ -1184,7 +1192,7 @@ def get_git_runner_for_software_integration(self, reference_repository_name, use user_repository_namespace, github_user, tira_user_name, dockerhub_token, dockerhub_user, tira_client_token, repository_search_prefix, tira_task_id, tira_code_repository_id, - tira_client_user): + tira_client_user, private): user = self.gitHoster_client.get_user() try: user_repo = user.get_repo(f'{user_repository_namespace}/{user_repository_name}') @@ -1199,18 +1207,18 @@ def get_git_runner_for_software_integration(self, reference_repository_name, use tira_user_name, dockerhub_token, dockerhub_user, tira_client_token, repository_search_prefix, tira_task_id, tira_code_repository_id, - tira_client_user) + tira_client_user, private) def create_software_submission_repository_for_user(self, reference_repository_name, user_repository_name, user_repository_namespace, github_user, tira_user_name, dockerhub_token, dockerhub_user, tira_client_token, repository_search_prefix, tira_task_id, tira_code_repository_id, - tira_client_user): + tira_client_user, private): reference_repo = self.gitHoster_client.get_repo(reference_repository_name) org = self.gitHoster_client.get_organization(user_repository_namespace) repo = org.create_repo(user_repository_name, - f'The repository of user {tira_user_name} for code submissions in TIRA.', private=True) + f'The repository of user {tira_user_name} for code submissions in TIRA.', private=private) repo.add_to_collaborators(github_user, 'admin') repo.create_secret('TIRA_DOCKER_REGISTRY_TOKEN', dockerhub_token) diff --git a/application/src/tira/tira_model.py b/application/src/tira/tira_model.py index 296c3ddd9..08ed6084c 100644 --- a/application/src/tira/tira_model.py +++ b/application/src/tira/tira_model.py @@ -185,7 +185,7 @@ def github_user_exists(user_name): return g.git_user_exists(user_name) -def get_submission_git_repo(vm_id, task_id, disraptor_user=None, external_owner=None): +def get_submission_git_repo(vm_id, task_id, disraptor_user=None, external_owner=None, private=True): user_repository_name = slugify(task_id) + '-' + slugify(vm_id) repository_url = settings.CODE_SUBMISSION_REPOSITORY_NAMESPACE + '/' + user_repository_name ret = model.get_submission_git_repo_or_none(repository_url, vm_id) @@ -196,7 +196,7 @@ def get_submission_git_repo(vm_id, task_id, disraptor_user=None, external_owner= docker_data = load_docker_data(task_id, vm_id, cache, force_cache_refresh=False) docker_registry_user = docker_data['docker_registry_user'] docker_registry_token = docker_data['docker_registry_token'] - reference_repository = settings.CODE_SUBMISSION_REFERENCE_REPOSITORY + reference_repository = settings.CODE_SUBMISSION_REFERENCE_REPOSITORIES[task_id] disraptor_description = disraptor_user + '-repo-' + task_id + '-' + vm_id discourse_api_key = discourse_api_client().generate_api_key(disraptor_user, disraptor_description) @@ -219,6 +219,7 @@ def get_submission_git_repo(vm_id, task_id, disraptor_user=None, external_owner= tira_task_id=task_id, tira_code_repository_id=repository_url, tira_client_user=disraptor_user, + private=private ) ret.confirmed = True @@ -230,7 +231,7 @@ def get_submission_git_repo(vm_id, task_id, disraptor_user=None, external_owner= def git_pipeline_is_enabled_for_task(task_id, cache, force_cache_refresh=False): evaluators_for_task = get_evaluators_for_task(task_id, cache, force_cache_refresh) git_runners_for_task = [i['is_git_runner'] for i in evaluators_for_task] - + # We enable the docker part only if all evaluators use the docker variant. return len(git_runners_for_task) > 0 and all(i for i in git_runners_for_task) diff --git a/application/src/tira/urls.py b/application/src/tira/urls.py index 0dade4af0..b37ec8bc3 100644 --- a/application/src/tira/urls.py +++ b/application/src/tira/urls.py @@ -107,6 +107,7 @@ path('api/registration/add_registration/<str:vm_id>/<str:task_id>', data_api.add_registration, name='add_registration'), path('api/submissions-for-task/<str:task_id>/<str:user_id>/<str:submission_type>', data_api.submissions_for_task, name="submissions_for_task"), path('api/tirex-components', data_api.tirex_components, name='tirex_components'), + path('api/snippets-for-tirex-components', data_api.get_snippet_to_run_components, name='get_snippet_to_run_components'), path('api/re-ranking-datasets/<str:task_id>', data_api.reranking_datasets, name='reranking_datasets'), path('api/submissions-of-user/<str:vm_id>', data_api.submissions_of_user, name='submissions_of_user'), path('api/add_software_submission_git_repository/<str:task_id>/<str:vm_id>', vm_api.add_software_submission_git_repository, name='add_software_submission_git_repository'), diff --git a/application/src/tirex-components.yml b/application/src/tirex-components.yml index cd32d7130..d3d8fd281 100644 --- a/application/src/tirex-components.yml +++ b/application/src/tirex-components.yml @@ -207,6 +207,8 @@ - display_name: Tutorial href: https://github.com/OpenWebSearch/wows-code/blob/main/ecir24/post-hoc-notebooks/keyphrase-extraction.ipynb target: _blank + - display_name: Submission in TIREx + href: /submissions/ir-benchmarks/ows/bce-fo - display_name: Health Classification component_type: [TIREx, Tutorial] @@ -287,7 +289,9 @@ - display_name: Tutorial href: https://github.com/webis-de/ir-pad/tree/public/tutorials#tutorial-8-query-segmentation-in-progress-by-maik target: _blank - + - display_name: Submission in TIREx + href: /submissions/ir-benchmarks/ows/query-segmentation-hyp-a + - display_name: hyp-b focus_type: [Precision] component_type: [TIREx, Tutorial] @@ -295,6 +299,8 @@ - display_name: Tutorial href: https://github.com/webis-de/ir-pad/tree/public/tutorials#tutorial-8-query-segmentation-in-progress-by-maik target: _blank + - display_name: Submission in TIREx + href: /submissions/ir-benchmarks/ows/query-segmentation-hyp-b - display_name: hyp-i focus_type: [Precision] diff --git a/python-client/Makefile b/python-client/Makefile index 308537015..cd986a83a 100644 --- a/python-client/Makefile +++ b/python-client/Makefile @@ -3,7 +3,7 @@ build-pypi-package: run-tests python3 -m build --sdist . python3 -m build --wheel . - twine upload dist/tira-0.0.97-py3-none-any.whl dist/tira-0.0.97.tar.gz + twine upload dist/tira-0.0.102-py3-none-any.whl dist/tira-0.0.102.tar.gz run-tests: docker run -u root --rm -v /var/run/docker.sock:/var/run/docker.sock -v ${PWD}:/app -w /app --entrypoint pytest webis/tira:python-client-dev-0.0.4 diff --git a/python-client/sphinx/development/python-client/workflow.rst b/python-client/sphinx/development/python-client/workflow.rst index 49b14e313..b581dc665 100644 --- a/python-client/sphinx/development/python-client/workflow.rst +++ b/python-client/sphinx/development/python-client/workflow.rst @@ -35,6 +35,13 @@ Unit Tests pytest +Code Coverage +~~~~~~~~~~~~~ +.. code-block:: bash + + coverage run -m pytest tests + coverage xml --omit=tests/** + Generating documentation ~~~~~~~~~~~~~~~~~~~~~~~~ From within the `python-client/sphynx` directory run diff --git a/python-client/tira/__init__.py b/python-client/tira/__init__.py index 625f000df..945bce51f 100644 --- a/python-client/tira/__init__.py +++ b/python-client/tira/__init__.py @@ -1,1 +1,1 @@ -__version__ = "0.0.97" +__version__ = "0.0.102" diff --git a/python-client/tira/io_utils.py b/python-client/tira/io_utils.py index 8091a76ff..3c60bd380 100644 --- a/python-client/tira/io_utils.py +++ b/python-client/tira/io_utils.py @@ -3,18 +3,19 @@ import gzip import json import os -from typing import Any, Iterable, Dict, Union +from typing import Any, List, Iterable, Dict, Union, Generator +from pathlib import Path +import logging def parse_jsonl_line(input: Union[str, bytearray, bytes], load_default_text: bool) -> Dict: - """ Deseralizes the line using JSON deserialization. Optionally strips the 'original_query' and 'original_document' fields from the resulting object and converts the qid and docno fields to strings. :param str | bytearray | bytes input: A json-serialized string. - :param bool load_default_text: If true, the origianl_query and original_document fields are removed and the qid and docno - values are converted to strings. + :param bool load_default_text: If true, the original_query and original_document fields are removed and the qid and + docno values are converted to strings. :return: The deserialized and (optionally) processed object. :rtype: dict @@ -43,7 +44,11 @@ def parse_jsonl_line(input: Union[str, bytearray, bytes], load_default_text: boo return obj -def stream_all_lines(input_file, load_default_text): +def stream_all_lines(input_file: Union[str, Iterable[bytes]], load_default_text: bool) -> Generator[Dict, Any, Any]: + """ + .. todo:: add documentation + .. todo:: this function has two semantics: handling a file and handling file-contents + """ if type(input_file) is str: if not os.path.isfile(input_file): return @@ -61,7 +66,11 @@ def stream_all_lines(input_file, load_default_text): yield parse_jsonl_line(line, load_default_text) -def all_lines_to_pandas(input_file: str | Iterable[str], load_default_text): +def all_lines_to_pandas(input_file: Union[str, Iterable[str]], load_default_text: bool) -> pd.DataFrame: + """ + .. todo:: add documentation + .. todo:: this function has two semantics: handling a file and handling file-contents + """ if type(input_file) is str: if input_file.endswith('.gz'): with gzip.open(input_file, 'rt', encoding='utf-8') as f: @@ -79,17 +88,36 @@ def all_lines_to_pandas(input_file: str | Iterable[str], load_default_text): return pd.DataFrame(ret) -def __num(s): +def __num(input: str) -> Union[str, int, float]: + """ + Converts the input to an int or float if possible. Returns the inputted string otherwise. + + :param str input: The string that should be converted to a float or int if possible. + :return: The intrepteted input. + :rtype: str | int | float + + :Example: + >>> __num("hello world") + "hello world" + >>> __num("-42") + -42 + >>> __num("3.5") + 3.5 + >>> __num("2e-6") + 2e-6 + >>> __num(" -42") + " -42" + """ try: - return int(s) + return int(input) except ValueError: try: - return float(s) + return float(input) except ValueError: - return s + return input -def run_cmd(cmd, ignore_failure=False): +def run_cmd(cmd: List[str], ignore_failure=False): import subprocess exit_code = subprocess.call(cmd) @@ -148,7 +176,7 @@ def all_environment_variables_for_github_action_or_fail(params): return [k + '=' + v for k, v in ret.items()] -def load_output_of_directory(directory, evaluation=False, verbose=False): +def load_output_of_directory(directory: Path, evaluation: bool=False) -> Union[Dict, pd.DataFrame]: files = glob(str(directory) + '/*') if evaluation: @@ -159,8 +187,7 @@ def load_output_of_directory(directory, evaluation=False, verbose=False): files = files[0] - if verbose: - print(f'Read file from {files}') + logging.debug(f'Read file from {files}') if evaluation: ret = {} diff --git a/python-client/tira/local_execution_integration.py b/python-client/tira/local_execution_integration.py index 3db0db4bf..3261bed74 100644 --- a/python-client/tira/local_execution_integration.py +++ b/python-client/tira/local_execution_integration.py @@ -7,6 +7,8 @@ import subprocess import tarfile import logging +from pathlib import Path +import pandas as pd class LocalExecutionIntegration(): @@ -174,7 +176,7 @@ def run(self, identifier=None, image=None, command=None, input_dir=None, output_ container = client.containers.run(image, entrypoint='sh', command=f'-c "{command}; sleep .1"', environment=environment, volumes=volumes, detach=True, remove=True, network_disabled = not allow_network) for line in container.attach(stdout=True, stream=True, logs=True): - logging.info(line.decode('utf-8')) + print(line.decode('utf-8')) if evaluate: evaluation_volumes = {str(eval_dir): {'bind': '/tira-data/eval_output', 'mode': 'rw'}} @@ -196,7 +198,7 @@ def run(self, identifier=None, image=None, command=None, input_dir=None, output_ container = client.containers.run(image, entrypoint='sh', command=f'-c "{command}; sleep .1"', volumes=evaluation_volumes, detach=True, remove=True, network_disabled = not allow_network) for line in container.attach(stdout=True, stream=True, logs=True): - logging.info(line.decode('utf-8'), flush=True) + print(line.decode('utf-8'), flush=True) if evaluate: approach_name = identifier if identifier else f'"{command}"@{image}' diff --git a/python-client/tira/tira_client.py b/python-client/tira/tira_client.py index 34d3cc7f2..11e46778f 100644 --- a/python-client/tira/tira_client.py +++ b/python-client/tira/tira_client.py @@ -1,5 +1,5 @@ from abc import ABC -from typing import TYPE_CHECKING, Literal, overload +from typing import TYPE_CHECKING, Literal, Union, overload if TYPE_CHECKING: from typing import Any, Dict, Optional @@ -75,6 +75,6 @@ def download_run( team=None, previous_stage=None, return_metadata: bool = False, - ) -> "pd.DataFrame | tuple[pd.DataFrame, str]": + ) -> "Union[pd.DataFrame, tuple[pd.DataFrame, str]]": # .. todo:: typehint pass diff --git a/python-client/tira/tira_run.py b/python-client/tira/tira_run.py index c755d9c11..23b247e81 100755 --- a/python-client/tira/tira_run.py +++ b/python-client/tira/tira_run.py @@ -43,6 +43,7 @@ def parse_args(): parser.add_argument('--tira-vm-id', required=False, default=os.environ.get('TIRA_VM_ID')) parser.add_argument('--tira-task-id', required=False, default=os.environ.get('TIRA_TASK_ID')) parser.add_argument('--tira-code-repository-id', required=False, default=os.environ.get('TIRA_CODE_REPOSITORY_ID')) + parser.add_argument('--fail-if-output-is-empty', required=False, default=False, action='store_true', help='After the execution of the software, fail if the output directory is empty.') args = parser.parse_args() if args.export_submission_from_jupyter_notebook: @@ -165,8 +166,17 @@ def main(): print(f'Done: Evaluation truth for dataset {dataset} is available.') logging.basicConfig(level=logging.DEBUG if args.verbose else logging.INFO) + print(f''' +########################################### TIRA RUN CONFIGURATION ########################################### +# image=${args.image} +# command=${args.command} +############################################################################################################## +''') client.local_execution.run(identifier=args.approach, image=args.image, command=args.command, input_dir=input_dir, output_dir=args.output_directory, dry_run=args.dry_run, allow_network=args.allow_network, input_run=args.input_run, additional_volumes=args.v, evaluate=evaluate, eval_dir=args.evaluation_directory) - + + if args.fail_if_output_is_empty and (not os.path.exists(args.output_directory) or not os.listdir(args.output_directory)): + raise ValueError('The software produced an empty output directory, it likely failed?') + if args.push.lower() == 'true': print('Push Docker image') client.local_execution.push_image(args.image, args.tira_docker_registry_user, args.tira_docker_registry_token)

Changed code display in run-review to code-snippet component. Changed code display in run review to CodeSnippter component. Also checked if we display code in other place without using CodeSnippet component, but didn't find anything. <img width="764" alt="image" src="https://github.com/tira-io/tira/assets/73996300/1f27aef4-29af-4e7a-8a4f-f61517a7416e"> tira-run seems to not print stderr and stdout to the console within Github Actions Learned this during the IRIXYS hackathon

2023-12-01T15:31:37

0.0

[]

GoogleCloudPlatform/snapshot-debugger

GoogleCloudPlatform__snapshot-debugger-143

70f8def8d2daf0d209e0620f7fd19a38a283d44e

diff --git a/README.md b/README.md index 8e65f70..51a99d6 100644 --- a/README.md +++ b/README.md @@ -249,6 +249,42 @@ Note: The information printed by the `init` command can be accessed from within your Firebase project. It’s safe to run the `snapshot-dbg-cli init --use-default-rtdb` command multiple times to view this information. +#### Setting up Firebase RTDB in other regions + +By default, `snapshot-dbg-cli init` will create a Firebase Realtime Database in +`us-central1`. It is possible to create and use a database in any region +supported by Firebase Realtime Database. See +[supported RTDB locations][rtdb_locations]. + +Setting up your database in a non-default region comes with some trade-offs: +* As a positive, you get to control where your snapshot data will be stored. + This may be important for compliance reasons. +* As a negative, the vsCode extension and agents will be unable to + automatically find the database. The database URL will need to be provided + explicitly via configuration, see the following for details: + * [Configuring the Java Agent][java_agent_config] + * [Configuring the Python Agent][python_agent_config] + * [Configuring the Node.js Agent][nodejs_agent_config] + * [Configuring the VsCode Extension][extension_config] + +You can set up your database in a non-default location as follows: +``` +snapshot-dbg-cli init --location={YOUR_LOCATION} +``` + +For example, you may want to set up your database in Belgium, and so would run +```snapshot-dbg-cli init --location=europe-west1``` + +Make note of the database URL provided in the command output; you will need to +provide this to your debug agent(s) and the vsCode plugin. + +[rtdb_locations]: https://firebase.google.com/docs/projects/locations#rtdb-locations +[java_agent_config]: https://github.com/GoogleCloudPlatform/cloud-debug-java#snapshot-debugger---firebase-realtime-database-backend +[python_agent_config]: https://github.com/GoogleCloudPlatform/cloud-debug-python/blob/main/README.md#snapshot-debugger---firebase-realtime-database-backend +[nodejs_agent_config]: https://github.com/googleapis/cloud-debug-nodejs/blob/main/README.md#snapshot-debugger---firebase-realtime-database-backend +[extension_config]: https://github.com/GoogleCloudPlatform/snapshot-debugger/blob/main/snapshot_dbg_extension/README.md#configuration + + ## Set up Snapshot Debugger in your Google Cloud project ### Working Samples diff --git a/snapshot_dbg_cli/firebase_management_rest_service.py b/snapshot_dbg_cli/firebase_management_rest_service.py index 90e1c03..4063100 100644 --- a/snapshot_dbg_cli/firebase_management_rest_service.py +++ b/snapshot_dbg_cli/firebase_management_rest_service.py @@ -58,6 +58,9 @@ {response} """ +VALID_LOCATIONS_URL = ( + "https://firebase.google.com/docs/projects/locations#rtdb-locations") + class FirebaseManagementRestService: """Implements a service for making Firebase RTDB management REST requests. @@ -282,6 +285,20 @@ def rtdb_instance_create(self, database_id, location): self._user_output.debug("Got 400:", parsed_error) return DatabaseCreateResponse( DatabaseCreateStatus.FAILED_PRECONDITION) + + if parsed_error["error"]["status"] == "INVALID_ARGUMENT": + print_http_error( + self._user_output, request, err, error_message=error_message) + + self._user_output.error( + "This was attempting to create the database instance. One " + "potential reason for this is if an invalid location was " + "specified, valid locations for RTDBs can be found at " + f"{VALID_LOCATIONS_URL}. To note, valid RTDB locations are a " + "subset of valid Google Cloud regions.") + + raise SilentlyExitError from err + except (TypeError, KeyError, ValueError): pass diff --git a/snapshot_dbg_cli/firebase_types.py b/snapshot_dbg_cli/firebase_types.py index 83f00fc..a5b71cb 100644 --- a/snapshot_dbg_cli/firebase_types.py +++ b/snapshot_dbg_cli/firebase_types.py @@ -17,6 +17,7 @@ """ from enum import Enum +import re FIREBASE_MANAGMENT_API_SERVICE = 'firebase.googleapis.com' FIREBASE_RTDB_MANAGMENT_API_SERVICE = 'firebasedatabase.googleapis.com' @@ -123,12 +124,26 @@ def __init__(self, database_instance): self.database_url = database_instance['databaseUrl'] self.type = database_instance['type'] self.state = database_instance['state'] + self.location = self.extract_location(self.name) + + if self.location is None: + raise ValueError( + f"Failed to extract location from project name '{self.name}'") + except KeyError as e: missing_key = e.args[0] error_message = ('DatabaseInstance is missing expected field ' f"'{missing_key}' instance: {database_instance}") raise ValueError(error_message) from e + @staticmethod + def extract_location(name): + location_search = re.search('/locations/([^/]+)/', name) + if not location_search or len(location_search.groups()) != 1: + return None + + return location_search.groups()[0] + class DatabaseCreateResponse: """Represents the response of a database create request. diff --git a/snapshot_dbg_cli/init_command.py b/snapshot_dbg_cli/init_command.py index 8421f52..b0a99cf 100644 --- a/snapshot_dbg_cli/init_command.py +++ b/snapshot_dbg_cli/init_command.py @@ -192,6 +192,21 @@ state: {db_state} """ +LOCATION_MISMATCH_ERROR_MSG = ( + "ERROR the following database already exists: '{full_database_name}', " + "however its location '{existing_location}' does not match the requested " + "location '{requested_location}'.\n\n" + "The database ID ('{database_id}' in this case) must be unique across " + 'locations and projects.\n\n' + "If you meant to finalize the initialization of '{database_id}' in " + "'{requested_location}' (or verify it is already correctly initialized) " + "rerun the init command and specify '--location={existing_location}'.\n\n" + "If you meant to create a new database in '{requested_location}', given " + "'{database_id}' is already in use, you'll need to use a new name by " + "providing the '--database-id' argument to the init command. Note, even " + "if you delete '{database_id}', the name will remain reserved and will not " + 'be available to be reused in a new location, a new name must be chosen.') + class InitCommand: """This class implements the init command. @@ -222,14 +237,6 @@ def register(self, args_subparsers, required_parsers, common_parsers): # Only some locations are supported, see: # https://firebase.google.com/docs/projects/locations#rtdb-locations - # - # If unsupported location is used, this error occurs - # "error": { - # "code": 400, - # "message": "Request contains an invalid argument.", - # "status": "INVALID_ARGUMENT" - # } - # For now however we only support 'us-central1' parser.add_argument( '-l', '--location', help=LOCATION_HELP, default=DEFAULT_LOCATION) self.args_parser = parser @@ -243,11 +250,6 @@ def cmd(self, args, cli_services): self.permissions_service = cli_services.permissions_service self.project_id = cli_services.project_id - if args.location != DEFAULT_LOCATION: - self.user_output.error('ERROR: Currently the only supported location is ' - f"'{DEFAULT_LOCATION}'") - raise SilentlyExitError - # If the user does not have the required permissions this will emit an error # message and exit. self.permissions_service.check_required_permissions(REQUIRED_PERMISSIONS) @@ -347,6 +349,16 @@ def check_and_handle_database_instance(self, args, firebase_project): if status == DatabaseGetStatus.EXISTS: database_instance = instance_response.database_instance + if args.location != database_instance.location: + self.user_output.error( + LOCATION_MISMATCH_ERROR_MSG.format( + full_database_name=database_instance.name, + existing_location=database_instance.location, + requested_location=args.location, + database_id=database_id)) + + raise SilentlyExitError + elif status == DatabaseGetStatus.DOES_NOT_EXIST: create_response = self.firebase_management_service.rtdb_instance_create( database_id=database_id, location=args.location)

CLI general `location` support Currently the CLI `init` command restricts the `location` to `us-central1`. However, it is possible to create a Firebase RTDB instances in other regions, which can be found at https://firebase.google.com/docs/projects/locations#rtdb-locations. The CLI should be modified to support these other locations as well. This issue was motivated by issue #134 .

2023-05-17T14:49:19

0.0

[]

GoogleCloudPlatform/snapshot-debugger

GoogleCloudPlatform__snapshot-debugger-64

59c525e9a2877be7882e2648cdb668c1913d864f

diff --git a/snapshot_dbg_cli/breakpoint_utils.py b/snapshot_dbg_cli/breakpoint_utils.py index 8f39031..f1767a8 100644 --- a/snapshot_dbg_cli/breakpoint_utils.py +++ b/snapshot_dbg_cli/breakpoint_utils.py @@ -131,7 +131,7 @@ def normalize_breakpoint(bp, bpid=None): Returns: The normalized breakpoint on success, None on failure. """ - if bp is None: + if type(bp) is not dict: return None if 'id' not in bp and bpid is not None:

Harden `normalize_breakpoint` utility function The `normalize_breakpoint` utility function should ensure the breakpoint passed in is a dict as expected. This issue came about when using a DB that had some incorrectly formatted data, and the CLI had the following error: ``` python3 -m snapshot_dbg_cli list_snapshots --include-inactive Traceback (most recent call last): File "/usr/lib/python3.10/runpy.py", line 196, in _run_module_as_main return _run_code(code, main_globals, None, File "/usr/lib/python3.10/runpy.py", line 86, in _run_code exec(code, run_globals) File "/snapshot-debugger/snapshot_dbg_cli/__main__.py", line 18, in <module> snapshot_dbg_cli.run_main() File "/snapshot-debugger/snapshot_dbg_cli/__init__.py", line 33, in run_main main() File "/snapshot-debugger/snapshot_dbg_cli/__init__.py", line 28, in main snapshot_dbg_cli.cli_run.run() File "/snapshot-debugger/snapshot_dbg_cli/cli_run.py", line 69, in run args.func(args=args, cli_services=cli_services) File "/snapshot-debugger/snapshot_dbg_cli/list_snapshots_command.py", line 96, in cmd snapshots = debugger_rtdb_service.get_snapshots( File "/snapshot-debugger/snapshot_dbg_cli/snapshot_debugger_rtdb_service.py", line 212, in get_snapshots return self._get_breakpoints(debuggee_id, include_inactive, 'CAPTURE', File "/snapshot-debugger/snapshot_dbg_cli/snapshot_debugger_rtdb_service.py", line 225, in _get_breakpoints breakpoints += self._get_breakpoints_by_path_and_filter( File "/snapshot-debugger/snapshot_dbg_cli/snapshot_debugger_rtdb_service.py", line 236, in _get_breakpoints_by_path_and_filter breakpoints = [ File "/snapshot-debugger/snapshot_dbg_cli/snapshot_debugger_rtdb_service.py", line 238, in <listcomp> if normalize_breakpoint(bp, bpid) and bp['action'] == action and File "/snapshot-debugger/snapshot_dbg_cli/breakpoint_utils.py", line 138, in normalize_breakpoint bp['id'] = bpid TypeError: 'str' object does not support item assignment ```

2022-07-28T16:17:52

0.0

[]

GoogleCloudPlatform/snapshot-debugger

GoogleCloudPlatform__snapshot-debugger-62

a73dd5990b3132fb3b57b737af0548dc7544b5ea

diff --git a/snapshot_dbg_cli/snapshot_parser.py b/snapshot_dbg_cli/snapshot_parser.py index 8fe9d4a..e576a6e 100644 --- a/snapshot_dbg_cli/snapshot_parser.py +++ b/snapshot_dbg_cli/snapshot_parser.py @@ -73,9 +73,10 @@ def parse_expressions(self): def parse_locals(self, stack_frame_index): local_variables = [] - if stack_frame_index < len( - self.stack_frames) and 'locals' in self.stack_frames[stack_frame_index]: - local_variables = self.stack_frames[stack_frame_index]['locals'] + if stack_frame_index < len(self.stack_frames): + stack_frame = self.stack_frames[stack_frame_index] + for p in ['arguments', 'locals']: + local_variables.extend(stack_frame.get(p, [])) return self._resolve_variables(local_variables)

The get_snapshot command does not list function arguments The captured snapshot data for functions in the call stack contains captured variable data for both `locals` and `arguments`. Currently the `get_snapshot` command is only displaying the variables classified as `locals`, the `arguments` variables are missing.

2022-07-28T14:43:37

0.0

[]

GoogleCloudPlatform/snapshot-debugger

GoogleCloudPlatform__snapshot-debugger-33

9038f87a96813e32f8fdf00dd433ed7ca3558aaf

diff --git a/cli/data_formatter.py b/cli/data_formatter.py index 8546c1a..b1b4860 100644 --- a/cli/data_formatter.py +++ b/cli/data_formatter.py @@ -39,7 +39,7 @@ def build_table(self, headers, values): """ widths = [ max(len(headers[i]), max([len(v[i]) - for v in values])) + for v in values], default=0)) for i in range(len(headers)) ]

List snapshots failure when there are no snapshots to display When the `list_snapshots` command has no snapshots to display, it results in the following failure: ``` File "[...snip]/snapshot-debugger/cli/data_formatter.py", line 41, in <listcomp> max(len(headers[i]), max([len(v[i]) ValueError: max() arg is an empty sequence ```

2022-06-06T18:35:57

0.0

[]

ansys/aedt-testing

ansys__aedt-testing-211

5f12d6d10c6a0a50654aa84f7a728a87af85cf94

diff --git a/docs/CONTRIBUTE.md b/docs/CONTRIBUTE.md index 8ce14dc..6235ce2 100644 --- a/docs/CONTRIBUTE.md +++ b/docs/CONTRIBUTE.md @@ -2,13 +2,13 @@  -- [Configure your enviornment](#configure-your-enviornment) +- [Configure your environment](#configure-your-environment) - [Run all tests and validations via tox](#run-all-tests-and-validations-via-tox) - [Build package](#build-package)  -## Configure your enviornment +## Configure your environment Install all dependencies ```bash pip install .[test]

Project report for [Project] does not exist Correlation plot is generated when comparing 2022R1 and 2022R2 result. However, the correlation plot is not generated when comparing 2022R1/R2 with 2023R1. Getting "error_exception" with the error message "Project report for [myProject] does not exist".

2022-08-17T13:53:38

0.0

[]

ansys/aedt-testing

ansys__aedt-testing-101

2e611333d948ee9dacd37a20660bef7efc8db429

diff --git a/pyproject.toml b/pyproject.toml index f7362c2..cabf824 100644 --- a/pyproject.toml +++ b/pyproject.toml @@ -18,14 +18,13 @@ classifiers = ["License :: OSI Approved :: MIT License"] dynamic = ["version", "description"] dependencies = [ - "pyaedt==0.4.12", + "pyaedt==0.4.16", "Django==3.2.8", ] # development dependencies [project.optional-dependencies] test = [ - "black==21.9b0", "pre-commit==2.15.0", "mypy==0.910", "pytest==6.2.5",

hide pyaedt logs under --debug flag

2021-12-06T13:42:21

0.0

[]

ansys/aedt-testing

ansys__aedt-testing-10

347912f0a9634aedcf3880fe58930c96e9a481f6

diff --git a/requirements.txt b/requirements.txt index 53af3bb..f44d0db 100644 --- a/requirements.txt +++ b/requirements.txt @@ -1,1 +1,2 @@ -pyaedt==0.3.27 \ No newline at end of file +pyaedt==0.3.27 +Django==3.2.8 \ No newline at end of file

report progress of the run: queued, running, status (success/fail) consider using HTML report see https://stackoverflow.com/questions/6748559/generating-html-documents-in-python https://github.com/nedbat/coveragepy/blob/2d2d6088b94845ffc4062853aa7230879ee66a44/coverage/html.py#L125 https://stackoverflow.com/questions/37912260/plotting-graph-using-python-and-dispaying-it-using-html

2021-10-26T12:13:03

0.0

[]

QEF/qeschema

QEF__qeschema-28

d10f0b240eb12b18c479a03857448d1a3d7bed6a

diff --git a/.travis.yml b/.travis.yml index ad8b443..2207d4e 100644 --- a/.travis.yml +++ b/.travis.yml @@ -4,7 +4,6 @@ python: - "3.7" - "3.8" - "3.9" - - "pypy3" install: - pip install -r requirements-dev.txt script: diff --git a/docs/conf.py b/docs/conf.py index c56dc27..16ff381 100644 --- a/docs/conf.py +++ b/docs/conf.py @@ -22,7 +22,7 @@ author = 'Davide Brunato, Pietro Delugas' # The full version, including alpha/beta/rc tags -release = 'v1.1' +release = '1.2.0' # -- General configuration --------------------------------------------------- diff --git a/qeschema/__init__.py b/qeschema/__init__.py index 31ab244..de06f4c 100644 --- a/qeschema/__init__.py +++ b/qeschema/__init__.py @@ -1,6 +1,5 @@ -# -*- coding: utf-8 -*- # -# Copyright (c), 2015-2020, Quantum Espresso Foundation and SISSA (Scuola +# Copyright (c), 2015-2021, Quantum Espresso Foundation and SISSA (Scuola # Internazionale Superiore di Studi Avanzati). All rights reserved. # This file is distributed under the terms of the MIT License. See the # file 'LICENSE' in the root directory of the present distribution, or @@ -15,11 +14,11 @@ from .exceptions import QESchemaError, XmlDocumentError from .utils import set_logger -__version__ = '1.1.0' +__version__ = '1.2.0' __all__ = [ 'XmlDocument', 'QeDocument', 'PwDocument', 'PhononDocument', 'NebDocument', 'TdDocument', 'TdSpectrumDocument', 'RawInputConverter', 'PwInputConverter', 'PhononInputConverter', 'TdInputConverter', 'TdSpectrumInputConverter', - 'QESchemaError', 'XmlDocumentError', 'set_logger' + 'NebInputConverter', 'QESchemaError', 'XmlDocumentError', 'set_logger' ] diff --git a/qeschema/cards.py b/qeschema/cards.py index 82ef609..2233bd1 100644 --- a/qeschema/cards.py +++ b/qeschema/cards.py @@ -47,6 +47,23 @@ def get_atomic_species_card(name, **kwargs): return lines +def get_positions_units(item): + """ + Read different type of positions and return them together with units. + + :param item: Data item + :return: Atomic positions dict and unit, empty dict and None if not found + """ + + ptypes = ('atomic_positions', 'crystal_positions', 'wyckoff_positions') + units = ('bohr', 'crystal', 'crystal_sg') + for ptype, unit in zip(ptypes, units): + if ptype in item: + return item[ptype], unit + + return {}, None + + def get_atomic_positions_cell_card(name, **kwargs): """ Convert XML data to ATOMIC_POSITIONS card @@ -62,23 +79,15 @@ def get_atomic_positions_cell_card(name, **kwargs): return [] # Find atoms - atomic_positions = atomic_structure.get('atomic_positions', {}) - wyckoff_positions = atomic_structure.get('wyckoff_positions', {}) + positions, units = get_positions_units(atomic_structure) try: - is_wyckoff = False - if atomic_positions: - atoms = atomic_positions['atom'] - elif wyckoff_positions: - atoms = wyckoff_positions['atom'] - is_wyckoff = True - else: - atoms = [] + atoms = positions['atom'] except KeyError: logger.error("Cannot find any atoms for building ATOMIC_POSITIONS!") return [] - else: - if not isinstance(atoms, list): - atoms = [atoms] + + if not isinstance(atoms, list): + atoms = [atoms] # Check atoms with position constraints free_positions = kwargs.get('free_positions') @@ -89,7 +98,7 @@ def get_atomic_positions_cell_card(name, **kwargs): logger.error("ATOMIC_POSITIONS: incorrect number of position constraints!") # Add atomic positions - lines = ['%s %s' % (name, 'crystal_sg' if is_wyckoff else 'bohr')] + lines = ['%s %s' % (name, units)] for k in range(len(atoms)): try: line = '{:4}'.format(atoms[k].get('@name')) @@ -233,7 +242,6 @@ def get_cell_parameters_card(name, **kwargs): except KeyError: logger.error("Missing required arguments when building CELL_PARAMETERS card!") return [] - # Add cell parameters card cells = atomic_structure.get('cell', {}) if cells: @@ -262,7 +270,7 @@ def get_qpoints_card(name, **kwargs): qplot = kwargs.get('qplot', False) if not qplot and not ldisp: try: - xq = kwargs['xq_dir'] + xq = kwargs['xq'] except KeyError: xq = [0.e0, 0.e0, 0.e0] line = "{:6.4f} {:8.4f} {:8.4f}".format(xq[0], xq[1], xq[2]) @@ -284,6 +292,19 @@ def get_qpoints_card(name, **kwargs): return lines +def get_nat_todo_card(name, **kwargs): + + natom = kwargs['nat_todo']['@natom'] + if natom == 0: + assert 'atom' not in kwargs['nat_todo'] + return [] + + atoms = kwargs['nat_todo']['atom'] + + assert natom == len(atoms) + return [' '.join(map(str, atoms))] + + def get_climbing_images(name, **kwargs): assert isinstance(name, str) try: @@ -332,9 +353,7 @@ def get_neb_images_positions_card(name, **kwargs): lines = ['BEGIN_POSITIONS ', 'FIRST_IMAGE '] for pos, item in enumerate(images): - positions = item.get('atomic_positions', - item.get('crystal_positions', - item.get('wyckoff_positions', {}))) + positions, units = get_positions_units(item) atoms = positions.get('atom', []) if pos == 0: @@ -345,30 +364,33 @@ def get_neb_images_positions_card(name, **kwargs): if first_nat != int(item.get('@nat', 0)): logger.error("nat provided in first image differs from number " "of atoms in atomic_positions!!!") + return '' if free_positions and len(free_positions) != 3 * first_nat: logger.error("ATOMIC_POSITIONS: incorrect number of position constraints!") + return '' else: if len(atoms) != first_nat: logger.error('found images with differing number of atoms !!!') + return '' if pos < len(images) - 1: lines.append('INTERMEDIATE_IMAGE ') else: lines.append('LAST_IMAGE ') - lines.append('ATOMIC_POSITIONS { bohr }') - - for k in range(len(atoms)): - sp_name = '{:4}'.format(atoms[k]['@name']) - coords = '{:12.8f} {:12.8f} {:12.8f}'.format(*atoms[k]['$']) - if k < len(free_positions): - try: - free_pos = '{:4d}{:4d}{:4d}'.format(*free_positions[3 * k:3 * k + 3]) - except IndexError: - pass - else: - if sum(free_positions[3 * k + n] for n in range(3)) < 3: - lines.append('%s %s %s' % (sp_name, coords, free_pos)) + lines.append('ATOMIC_POSITIONS { %s }' % units) + + # reshape to 3x3 + free_positions_sq = [free_positions[i:i + 3] + for i in range(0, len(free_positions), 3)] + + for k, atom in enumerate(atoms): + sp_name = '{:4}'.format(atom['@name']) + coords = '{:12.8f} {:12.8f} {:12.8f}'.format(*atom['$']) + + if free_positions_sq and sum(free_positions_sq[k]) < 3: + free_pos = '{:4d}{:4d}{:4d}'.format(*free_positions_sq[k]) + lines.append('%s %s %s' % (sp_name, coords, free_pos)) else: lines.append('%s %s' % (sp_name, coords)) diff --git a/qeschema/converters.py b/qeschema/converters.py index f669a11..709a1c6 100644 --- a/qeschema/converters.py +++ b/qeschema/converters.py @@ -21,6 +21,10 @@ logger = logging.getLogger('qeschema') +FCP_NAMELIST = 'FCP' +PH_NAT_TODO_CARD = 'ph_nat_todo_card' +OPTIONAL_CARDS = {PH_NAT_TODO_CARD} + def conversion_maps_builder(template_map): """ @@ -79,7 +83,8 @@ def _build_maps(dict_element, path): if isinstance(item, tuple) or isinstance(item, list): try: variant_map[path_key] = _check_variant(*item) - logger.debug("Added single-variant mapping: '{}'={}".format(path_key, variant_map[path_key])) + logger.debug("Added single-variant mapping: %r=%r", + path_key, variant_map[path_key]) except TypeError: variants = [] for variant in item: @@ -90,7 +95,8 @@ def _build_maps(dict_element, path): else: raise TypeError("Expect a tuple, list or string! {0}".format(variant)) variant_map[path_key] = tuple(variants) - logger.debug("Added multi-variant mapping: '{}'={}".format(path_key, variant_map[path_key])) + logger.debug("Added multi-variant mapping: %r=%r", + path_key, variant_map[path_key]) invariant_map = BiunivocalMap() variant_map = dict() @@ -116,7 +122,7 @@ class RawInputConverter(Container): """ A Fortran's namelist builder. """ - target_pattern = re.compile(r'(\w+)(?:\[((?:\w+)(?:%\w+)*)\]|)') + target_pattern = re.compile(r'(\w+)(?:\[((?:\w+)(?:%\w+)*)]|)') """RE pattern to extract Fortran input's namelist/card and name of a parameter""" def __init__(self, invariant_map, variant_map, input_namelists=None, input_cards=None): @@ -191,7 +197,7 @@ def set_parameter(self, path, value): else: namelist = name = None if name is None: - raise ValueError("Wrong value for invariant parameter '{0}'! '{1}'".format(path, target)) + raise ValueError("Wrong value {!r} for invariant parameter {!r}".format(target, path)) self._input[namelist][name] = to_fortran(value) logger.debug("Set {0}[{1}]={2}".format(namelist, name, self._input[namelist][name])) @@ -238,16 +244,21 @@ def get_qe_input(self): _input = self._input lines = [] for namelist in self.input_namelists: + + if namelist == FCP_NAMELIST and not len(_input[namelist]): + # Skip empty &FCP/ section + continue + lines.append('&%s' % namelist) for name, value in sorted(_input[namelist].items(), key=lambda x: x[0].lower()): - logger.debug("Add input for parameter {}[{}] with value {}".format(namelist, name, value)) + logger.debug("Add input for parameter %s[%r] with value %r", namelist, name, value) if isinstance(value, dict): # Variant conversion: apply to_fortran_input function with saved arguments try: to_fortran_input = value['_get_qe_input'] except KeyError: logger.debug( - 'No conversion function for parameter %s[%s], skip ... ' % (namelist, name) + 'No conversion function for parameter %s[%r], skip ... ', namelist, name ) continue @@ -255,7 +266,7 @@ def get_qe_input(self): lines.extend(to_fortran_input(name, **value)) else: logger.error( - 'Parameter %s[%s] conversion function is not callable!' % (namelist, name) + 'Parameter %s[%r] conversion function is not callable!', namelist, name ) else: # Simple invariant conversion @@ -265,14 +276,19 @@ def get_qe_input(self): logger.debug("Add card: %s" % card) card_args = _input[card] logger.debug("Card arguments: {0}".format(card_args)) - if '_get_qe_input' not in card_args or \ - not callable(card_args['_get_qe_input']): + + if card not in OPTIONAL_CARDS and \ + ('_get_qe_input' not in card_args or + not callable(card_args['_get_qe_input'])): logger.error("Missing conversion function for card '%s'" % card) + _get_qe_input = card_args.get('_get_qe_input', None) + if callable(_get_qe_input): lines.extend(_get_qe_input(card, **card_args)) - else: + elif card not in OPTIONAL_CARDS: logger.error('Card conversion function not found!') + return '\n'.join(lines) def clear_input(self): @@ -325,9 +341,11 @@ class PwInputConverter(RawInputConverter): '$': [ ('SYSTEM[ibrav]', options.set_ibrav_to_zero, None), ("ATOMIC_POSITIONS", cards.get_atomic_positions_cell_card, None), - ("CELL_PARAMETERS", cards.get_cell_parameters_card, None) + ("CELL_PARAMETERS", cards.get_cell_parameters_card, None) ], 'atomic_positions': ('ATOMIC_FORCES', cards.get_atomic_forces_card, None), + 'crystal_positions': ('ATOMIC_FORCES', cards.get_atomic_forces_card, None), + 'wyckoff_positions': ('ATOMIC_FORCES', cards.get_atomic_forces_card, None) }, 'dft': { 'functional': "SYSTEM[input_dft]", @@ -347,7 +365,8 @@ class PwInputConverter(RawInputConverter): 'dftU': { 'lda_plus_u_kind': 'SYSTEM[lda_plus_u_kind]', 'Hubbard_U': { - '$': ('SYSTEM[Hubbard_U]', options.get_specie_related_values, None), + '$': [('SYSTEM[Hubbard_U]', options.get_specie_related_values, None), + ('SYSTEM[lda_plus_u]', options.set_lda_plus_u_flag, None)] }, 'Hubbard_J0': { '$': ('SYSTEM[Hubbard_J0]', options.get_specie_related_values, None), @@ -362,7 +381,7 @@ class PwInputConverter(RawInputConverter): '$': ('SYSTEM[Hubbard_J]', options.get_specie_related_values, None), }, 'starting_ns': { - '$': ('SYSTEM[starting_ns_eigenvalue]', options.get_specie_related_values, None), + '$': ('SYSTEM[starting_ns_eigenvalue]', options.get_specie_related_values, None) }, 'U_projection_type': 'SYSTEM[U_projection_type]', }, @@ -374,6 +393,8 @@ class PwInputConverter(RawInputConverter): 'london_rcut': 'SYSTEM[london_rcut]', 'xdm_a1': 'SYSTEM[xdm_a1]', 'xdm_a2': 'SYSTEM[xdm_a2]', + 'dftd3_version': 'SYSTEM[dftd3_version]', + 'dftd3_threebody': 'SYSTEM[dftd3_threebody]', 'london_c6': { '$': ('SYSTEM[london_c6]', options.get_specie_related_values, None), } @@ -417,7 +438,8 @@ class PwInputConverter(RawInputConverter): 'nr1': "SYSTEM[nr1b]", 'nr2': "SYSTEM[nr2b]", 'nr3': "SYSTEM[nr3b]", - } + }, + 'spline_ps': "SYSTEM[spline_ps]" }, 'electron_control': { 'diagonalization': "ELECTRONS[diagonalization]", @@ -466,6 +488,7 @@ class PwInputConverter(RawInputConverter): 'free_cell': ("CELL_PARAMETERS", cards.get_cell_parameters_card, None), 'fix_volume': ("CELL[cell_dofree]", options.get_cell_dofree, None), 'fix_area': ("CELL[cell_dofree]", options.get_cell_dofree, None), + 'fix_xy': ("CELL[cell_dofree]", options.get_cell_dofree, None), 'isotropic': ("CELL[cell_dofree]", options.get_cell_dofree, None), }, 'symmetry_flags': { @@ -484,8 +507,8 @@ class PwInputConverter(RawInputConverter): 'w': "SYSTEM[esm_w]", 'efield': "SYSTEM[esm_efield]" }, - 'fcp_opt': "CONTROL[lfcpopt]", - 'fcp_mu': "SYSTEM[fcp_mu]" + 'fcp_opt': "CONTROL[lfcp]", + 'fcp_mu': "FCP[fcp_mu]" }, 'ekin_functional': { 'ecfixed': "SYSTEM[ecfixed]", @@ -496,7 +519,8 @@ class PwInputConverter(RawInputConverter): '$': ('ATOMIC_FORCES', cards.get_atomic_forces_card, None) }, 'free_positions': { - '$': [("ATOMIC_POSITIONS", cards.get_atomic_positions_cell_card, None), ("CELL_PARAMETERS",)]}, + '$': [("ATOMIC_POSITIONS", cards.get_atomic_positions_cell_card, None), + ("CELL_PARAMETERS",)]}, 'electric_field': { 'electric_potential': [ ("CONTROL[tefield]", options.get_electric_potential_related), @@ -533,12 +557,15 @@ class PwInputConverter(RawInputConverter): def __init__(self, **kwargs): super(PwInputConverter, self).__init__( *conversion_maps_builder(self.PW_TEMPLATE_MAP), - input_namelists=('CONTROL', 'SYSTEM', 'ELECTRONS', 'IONS', 'CELL'), + input_namelists=('CONTROL', 'SYSTEM', 'ELECTRONS', 'IONS', 'CELL', + FCP_NAMELIST), input_cards=('ATOMIC_SPECIES', 'ATOMIC_POSITIONS', 'K_POINTS', 'CELL_PARAMETERS', 'ATOMIC_FORCES') ) if 'xml_file' in kwargs: - self._input['CONTROL']['input_xml_schema_file'] = u'\'{}\''.format(os.path.basename(kwargs['xml_file'])) + self._input['CONTROL']['input_xml_schema_file'] = "{!r}".format( + os.path.basename(kwargs['xml_file']) + ) def clear_input(self): super(PwInputConverter, self).clear_input() @@ -549,7 +576,9 @@ class PhononInputConverter(RawInputConverter): Convert to/from Fortran input for Phonon. """ PHONON_TEMPLATE_MAP = { - 'xq': ('qPointsSpecs', cards.get_qpoints_card, None), + 'xq': { + '$': ('qPointsSpecs', cards.get_qpoints_card, None), + }, 'scf_ph': { 'tr2_ph': "INPUTPH[tr2_ph]", 'niter_ph': "INPUTPH[niter_ph]", @@ -608,7 +637,11 @@ class PhononInputConverter(RawInputConverter): 'last_q': "INPUTPH[last_q]", 'start_irr': "INPUTPH[start_irr]", 'last_irr': "INPUTPH[last_irr]", - 'nat_todo': "INPUTPH[nat_todo]", + 'nat_todo': + { + '@natom': 'INPUTPH[nat_todo]', + '$': [(PH_NAT_TODO_CARD, cards.get_nat_todo_card, None)] + }, 'modenum': "INPUTPH[modenum]", 'only_init': "INPUTPH[only_init]", 'ldiag': "INPUTPH[ldiag]", @@ -637,9 +670,9 @@ class PhononInputConverter(RawInputConverter): }, 'q_points': { 'grid': { - 'nq1': "INPUTPH[nq1]", - 'nq2': "INPUTPH[nq2]", - 'nq3': "INPUTPH[nq3]" + '@nq1': "INPUTPH[nq1]", + '@nq2': "INPUTPH[nq2]", + '@nq3': "INPUTPH[nq3]" }, 'q_points_list': ('qPointsSpecs', cards.get_qpoints_card, None), 'nqs': ('qPointsSpecs', cards.get_qpoints_card, None) @@ -650,7 +683,7 @@ def __init__(self, **_kwargs): super(PhononInputConverter, self).__init__( *conversion_maps_builder(self.PHONON_TEMPLATE_MAP), input_namelists=('INPUTPH',), - input_cards=('qPointsSpecs',) + input_cards=('qPointsSpecs', PH_NAT_TODO_CARD) ) @@ -670,6 +703,7 @@ class NebInputConverter(RawInputConverter): 'elasticConstMin': "PATH[k_min]", 'pathThreshold': "PATH[path_thr]", 'endImagesOptimizationFlag': "PATH[first_last_opt]", + 'minimumImageFlag': "PATH[minimum_image]", 'temperature': "PATH[temp_req]", 'climbingImage': [ "PATH[CI_scheme]", @@ -677,7 +711,7 @@ class NebInputConverter(RawInputConverter): ], 'useMassesFlag': "PATH[use_masses]", 'useFreezingFlag': "PATH[use_freezing]", - 'constantBiasFlag': "PATH[lfcpopt]", + 'constantBiasFlag': "PATH[lfcp]", 'targetFermiEnergy': "PATH[fcp_mu]", 'totChargeFirst': "PATH[fcp_tot_charge_first]", 'totChargeLast': "PATH[fcp_tot_charge_last]", @@ -711,8 +745,9 @@ def __init__(self, **_kwargs): def get_qe_input(self): """ - Overrides method in RawInputConverter because few lines in between the namelists are requested for - the NEB input. + Overrides method in RawInputConverter because few lines + in between the namelists are requested for the NEB input. + :return: a string containing the text input for NEB calculations """ qe_input = super(NebInputConverter, self).get_qe_input().split('\n') diff --git a/qeschema/documents.py b/qeschema/documents.py index 83b7e46..cf683c8 100644 --- a/qeschema/documents.py +++ b/qeschema/documents.py @@ -21,7 +21,7 @@ except ImportError: yaml = None -from .namespaces import * +from .namespaces import XSD_NAMESPACE from .converters import RawInputConverter, PwInputConverter, PhononInputConverter, \ NebInputConverter, TdInputConverter, TdSpectrumInputConverter from .exceptions import XmlDocumentError diff --git a/qeschema/hdf5/__init__.py b/qeschema/hdf5/__init__.py new file mode 100644 index 0000000..1e7bc0f --- /dev/null +++ b/qeschema/hdf5/__init__.py @@ -0,0 +1,18 @@ +# +# Copyright (c), 2021, Quantum Espresso Foundation and SISSA (Scuola +# Internazionale Superiore di Studi Avanzati). All rights reserved. +# This file is distributed under the terms of the MIT License. See the +# file 'LICENSE' in the root directory of the present distribution, or +# http://opensource.org/licenses/MIT. +# +from .charge import read_charge_file_hdf5, get_minus_indexes, \ + get_charge_r, write_charge +from .readutils import get_wf_attributes, get_wavefunctions, \ + get_wfc_miller_indices, read_pseudo_file, create_header, \ + read_postqe_output_file, read_etotv + +__all__ = [ + 'read_charge_file_hdf5', 'get_minus_indexes', 'get_charge_r', 'write_charge', + 'get_wf_attributes', 'get_wavefunctions', 'get_wfc_miller_indices', + 'read_pseudo_file', 'create_header', 'read_postqe_output_file', 'read_etotv' +] diff --git a/qeschema/hdf5/charge.py b/qeschema/hdf5/charge.py new file mode 100644 index 0000000..42246c2 --- /dev/null +++ b/qeschema/hdf5/charge.py @@ -0,0 +1,162 @@ +# +# Copyright (c), 2021, Quantum Espresso Foundation and SISSA (Scuola +# Internazionale Superiore di Studi Avanzati). All rights reserved. +# This file is distributed under the terms of the MIT License. See the +# file 'LICENSE' in the root directory of the present distribution, or +# http://opensource.org/licenses/MIT. +# +import numpy as np + +try: + import h5py +except ImportError: + h5py = None + H5PY_ERR = 'h5py module is missing' + + +def read_charge_file_hdf5(filename): + """ + Reads a PW charge file in HDF5 format. + + :param filename: the name of the HDF5 file to read. + :return: a dictionary describing the content of file \ + keys=[nr, ngm_g, gamma_only, rhog_, MillerIndexes] + """ + + if not h5py: + raise ImportError(H5PY_ERR) + + with h5py.File(filename, "r") as h5f: + MI = h5f.get('MillerIndices')[:] + nr1 = 2 * max(abs(MI[:, 0])) + 1 + nr2 = 2 * max(abs(MI[:, 1])) + 1 + nr3 = 2 * max(abs(MI[:, 2])) + 1 + nr = np.array([nr1, nr2, nr3]) + res = dict(h5f.attrs.items()) + res.update({'MillInd': MI, 'nr_min': nr}) + rhog = h5f['rhotot_g'][:].reshape(res['ngm_g'], 2).dot([1.e0, 1.e0j]) + res['rhotot_g'] = rhog + if 'rhodiff_g' in h5f.keys(): + rhog = h5f['rhodiff_g'][:].reshape(res['ngm_g'], 2).dot([1.e0, 1.e0j]) + res['rhodiff_g'] = rhog + return res + + +def get_minus_indexes(g1, g2, g3): + """ + Used for getting the corresponding minus Miller Indexes. It is meant + to be used for converting Gamma Trick grids and is defined only for + the for i >=0, in the i =0 plan is defined only for j >=0 and when + i=0 j=0 k must be >=0. Out of this domain returns None. + + :param g1: rank 1 array containing first Miller Index + :param g2: rank 1 array containing second Miller Index + :param g3: rank 1 array containing third Miller Index + :return: a rank 2 array with dimension (ngm/2,3) containing mirrored Miller indexes + """ + + def scalar_func(i, j, k): + """ + scalar function to be vectorized + :param i: 1st Miller Index + :param j: 2nd + :param k: 3rd + :return: the mirrored mirror indexes + """ + if i > 0: + return -i, j, k + elif i == 0 and j > 0: + return 0, -j, k + elif i == 0 and j == 0 and k > 0: + return 0, 0, -k + else: + return i, j, k + + vector_func = np.vectorize(scalar_func) + + res = np.array(vector_func(g1, g2, g3)) + return res.transpose() + + +def get_charge_r(filename, nr=None): + """ + Reads a charge file written with QE in HDF5 format. *nr = [nr1,nr2,nr3]* (the dimensions of + the charge k-points grid) are given as parameter (taken for the xml output file by the caller). + + Notes: In the new format, the values of the charge in the reciprocal space are stored. + Besides, only the values of the charge > cutoff are stored, together with the Miller indexes. + Hence + """ + cdata = read_charge_file_hdf5(filename) + if nr is None: + nr1, nr2, nr3 = cdata['nrmin'] + else: + nr1, nr2, nr3 = nr + gamma_only = 'TRUE' in str(cdata['gamma_only']).upper() + + # Load the total charge + rho_temp = np.zeros([nr1, nr2, nr3], dtype=np.complex128) + for (i, j, k), rho in zip(cdata['MillInd'], cdata['rhotot_g']): + try: + rho_temp[i, j, k] = rho + except IndexError: + pass + + if gamma_only: + rhotot_g = cdata['rhotot_g'].conjugate() + MI = get_minus_indexes( + cdata['MillInd'][:, 0], cdata['MillInd'][:, 1], cdata['MillInd'][:, 2] + ) + print("MI", MI) + for (i, j, k), rho in zip(MI, rhotot_g): + try: + rho_temp[i, j, k] = rho + except IndexError: + pass + + rhotot_r = np.fft.ifftn(rho_temp) * nr1 * nr2 * nr3 + + # Read the charge difference spin up - spin down if present (for magnetic calculations) + if 'rhodiff_g' in cdata.keys(): + rho_temp = np.zeros([nr1, nr2, nr3], dtype=np.complex128) + for (i, j, k), rho in zip(cdata['MillInd'], cdata['rhodiff_g']): + try: + rho_temp[i, j, k] = rho + except IndexError: + pass + + if gamma_only: + rhodiff_g = cdata['rhodiff_g'].conjugate() + for (i, j, k), rho in zip(MI, rhodiff_g): + try: + rho_temp[i, j, k] = rho + except IndexError: + pass + + rhodiff_r = np.fft.ifftn(rho_temp) * nr1 * nr2 * nr3 + return rhotot_r.real, rhodiff_r.real + else: + return rhotot_r.real, None + + +def write_charge(filename, charge, header): + """ + Write the charge or another quantity calculated by postqe into a text file *filename*. + """ + + fout = open(filename, "w") + + # The header contains some information on the system, the grid nr, etc. + fout.write(header) + nr = charge.shape + count = 0 + # Loop with the order as in QE files + for z in range(0, nr[2]): + for y in range(0, nr[1]): + for x in range(0, nr[0]): + fout.write(" {:.9E}".format(charge[x, y, z])) + count += 1 + if count % 5 == 0: + fout.write("\n") + + fout.close() diff --git a/qeschema/hdf5/readutils.py b/qeschema/hdf5/readutils.py new file mode 100644 index 0000000..dc92a46 --- /dev/null +++ b/qeschema/hdf5/readutils.py @@ -0,0 +1,288 @@ +# +# Copyright (c), 2021, Quantum Espresso Foundation and SISSA (Scuola +# Internazionale Superiore di Studi Avanzati). All rights reserved. +# This file is distributed under the terms of the MIT License. See the +# file 'LICENSE' in the root directory of the present distribution, or +# http://opensource.org/licenses/MIT. +# +""" +A collection of functions for reading different files and quantities. +""" +import numpy as np +from xml.etree import ElementTree + +try: + import h5py +except ImportError: + h5py = None + H5PY_ERR = 'h5py module is missing' + + +# TODO update to the new format +def get_wf_attributes(filename): + """ + Read attributes from a wfc file. + + :param filename: the path to the wfc file + :return: a dictionary with all attributes included reciprocal vectors + """ + + if not h5py: + raise ImportError(H5PY_ERR) + + with h5py.File(filename, "r") as f: + res = dict(f.attrs) + mi_attrs = f.get('MillerIndices').attrs + bg = np.array(mi_attrs.get(x) for x in ['bg1', 'bg2', 'bg3']) + res.update({'bg': bg}) + return res + + +def get_wavefunctions(filename, start_band=None, stop_band=None): + """ + Returns a numpy array with the wave functions for bands from start_band to + stop_band. If not specified starts from 1st band and ends with last one. + Band numbering is Python style starts from 0.abs + + :param filename: path to the wfc file + :param start_band: first band to read, default first band in the file + :param stop_band: last band to read, default last band in the file + :return: a numpy array with shape [nbnd,npw] + """ + + if not h5py: + raise ImportError(H5PY_ERR) + + with h5py.File(filename, "r") as f: + igwx = f.attrs.get('igwx') + if start_band is None: + start_band = 0 + if stop_band is None: + stop_band = f.attrs.get('nbnd') + if stop_band == start_band: + stop_band = start_band + 1 + res = f.get('evc')[start_band:stop_band, :] + + res = np.asarray(x.reshape([igwx, 2]).dot([1.e0, 1.e0j]) for x in res[:]) + return res + + +def get_wfc_miller_indices(filename): + """ + Reads miller indices from the wfc file + + :param filename: path to the wfc file + :return: a np.array of integers with shape [igwx,3] + """ + + if not h5py: + raise ImportError(H5PY_ERR) + + with h5py.File(filename, "r") as f: + res = f.get("MillerIndices")[:, :] + return res + + +def read_pseudo_file(xmlfile): + """ + This function reads a pseudopotential XML-like file in the QE UPF format (text), + returning the content of each tag in a dictionary. The file is read in strings + and completed with a root UPF tag when it lacks, to avoids an XML syntax error. + + TODO: add support for UPF-schema files + """ + def iter_upf_file(): + """ + Creates an iterator over the lines of an UPF file, + inserting the root <UPF> tag when missing. + """ + with open(xmlfile, 'r') as f: + fake_root = None + for line in f: + if fake_root is not None: + line = line.replace('&input', '&input') + line = line.replace('&inputp', '&inputp') + yield line + else: + line = line.strip() + if line.startswith("<UPF") and line[4] in ('>', ' '): + yield line + fake_root = False + elif line: + yield "<UPF>" + yield line + fake_root = True + if fake_root is True: + yield "</UPF>" + + pseudo = {} + psroot = ElementTree.fromstringlist(list(iter_upf_file())) + + # PP_INFO + try: + pp_info = psroot.find('PP_INFO').text + except AttributeError: + pp_info = "" + try: + pp_input = psroot.find('PP_INFO/PP_INPUTFILE').text + except AttributeError: + pp_input = "" + pseudo.update(PP_INFO=dict(INFO=pp_info, PP_INPUT=pp_input)) + + # PP_HEADER + pseudo.update(PP_HEADER=dict(psroot.find('PP_HEADER').items())) + + # PP_MESH + pp_mesh = dict(psroot.find('PP_MESH').items()) + pp_r = np.array([float(x) for x in psroot.find('PP_MESH/PP_R').text.split()]) + pp_rab = np.array([float(x) for x in psroot.find('PP_MESH/PP_RAB').text.split()]) + pp_mesh.update(PP_R=pp_r, PP_RAB=pp_rab) + pseudo.update(PP_MESH=pp_mesh) + + # PP_LOCAL + node = psroot.find('PP_LOCAL') + if node is not None: + pp_local = np.array([x for x in map(float, node.text.split())]) + else: + pp_local = None + pseudo.update(PP_LOCAL=pp_local) + + # PP_RHOATOM + node = psroot.find('PP_RHOATOM') + if node is not None: + pp_rhoatom = np.array([v for v in map(float, node.text.split())]) + else: + pp_rhoatom = None + pseudo.update(PP_RHOATOM=pp_rhoatom) + + # PP_NONLOCAL + node = psroot.find('PP_NONLOCAL') + if node is not None: + betas = list() + dij = None + pp_aug = None + pp_q = None + for el in node: + if 'PP_BETA' in el.tag: + beta = dict(el.items()) + val = np.array(x for x in map(float, el.text.split())) + beta.update(beta=val) + betas.append(beta) + elif 'PP_DIJ' in el.tag: + text = '\n'.join(el.text.strip().split('\n')[1:]) + dij = np.array([x for x in map(float, text.split())]) + elif 'PP_AUGMENTATION' in el.tag: + pp_aug = dict(el.items()) + pp_qijl = list() + pp_qij = list() + for q in el: + if 'PP_QIJL' in q.tag: + qijl = dict(q.items()) + val = np.array(x for x in map(float, q.text.split())) + qijl.update(qijl=val) + pp_qijl.append(qijl) + elif 'PP_QIJ' in q.tag: + qij = dict(q.items()) + val = np.array(x for x in map(float, q.text.split())) + qij.update(qij=val) + pp_qij.append(qij) + elif q.tag == 'PP_Q': + pp_q = np.array(x for x in map(float, q.text.split())) + pp_aug.update(PP_QIJL=pp_qijl, PP_QIJ=pp_qij, PP_Q=pp_q) + pp_nonlocal = dict(PP_BETA=betas, PP_DIJ=dij, PP_AUGMENTATION=pp_aug) + else: + pp_nonlocal = None + + pseudo.update(PP_NONLOCAL=pp_nonlocal) + return pseudo + + +def create_header(prefix, nr, nr_smooth, ibrav, celldms, nat, ntyp, + atomic_species, atomic_positions): + """ + Creates the header lines for the output charge (or potential) text file as in pp.x. + The format is: + + system_prefix + fft_grid (nr1,nr2,nr3) fft_smooth (nr1,nr2,nr3) nat ntyp + ibrav celldms (6 real as in QE) + missing line with respect to pp.x + List of atoms + List of positions for each atom + """ + text = "# " + prefix + "\n" + text += "# {:8d} {:8d} {:8d} {:8d} {:8d} {:8d} {:8d} {:8d}\n".format( + nr[0], nr[1], nr[2], nr_smooth[0], nr_smooth[1], nr_smooth[2], nat, ntyp + ) + text += "# {:6d} {:8E} {:8E} {:8E} {:8E} {:8E} {:8E}\n".format(ibrav, *celldms) + + # TODO This line is to be implemented + text += "# " + 4 * "XXXX " + "\n" + + ityp = 1 + for typ in atomic_species: + text += "# {:4d} ".format(ityp) + typ["@name"] + "\n" + ityp += 1 + + ipos = 1 + for pos in atomic_positions: + text += "# {:4d} ".format(ipos) + coords = [float(x) for x in pos['$']] + text += " {:9E} {:9E} {:9E} ".format(*coords) + text += pos["@name"] + "\n" + ipos += 1 + + return text + + +def read_postqe_output_file(filename): + """ + This function reads the output charge (or other quantity) as the output + format of postqe. + """ + tempcharge = [] + count = 0 + nr = np.zeros(3, dtype=int) + with open(filename, "r") as lines: + for line in lines: + words = line.split() + if count == 1: + nr[0] = int(words[1]) + nr[1] = int(words[2]) + nr[2] = int(words[3]) + if words[0] != '#': # skip the first lines beginning with # + for j in range(len(words)): # len(words)=5 except maybe for the last line + tempcharge.append(float(words[j])) + count += 1 + + charge = np.zeros((nr[0], nr[1], nr[2])) + count = 0 + + # Loops according to the order it is done in QE + for z in range(nr[2]): + for y in range(nr[1]): + for x in range(nr[0]): + charge[x, y, z] = tempcharge[count] + count += 1 + + return charge + + +def read_etotv(fname): + """ + Read cell volumes and the corresponding energies from input file *fname* + (1st col, volumes, 2nd col energies). Units must be :math:`a.u.^3` and + :math:`Ryd/cell` + """ + Vx = [] + Ex = [] + + with open(fname, "r") as lines: + for line in lines: + linesplit = line.split() + V = float(linesplit[0]) + E = float(linesplit[1]) + Vx.append(V) + Ex.append(E) + + return np.array(Vx), np.array(Ex) diff --git a/qeschema/options.py b/qeschema/options.py index b6040fb..3ad75a8 100644 --- a/qeschema/options.py +++ b/qeschema/options.py @@ -209,31 +209,26 @@ def get_control_gdir(name, **kwargs): def get_cell_dofree(name, **kwargs): assert isinstance(name, str) - try: - fix_volume = kwargs['fix_volume'] - except KeyError: - fix_volume = False - try: - fix_area = kwargs['fix_area'] - except KeyError: - fix_area = False - try: - isotropic = kwargs['isotropic'] - except KeyError: - isotropic = False + cell_dofree_str = "cell_dofree = '%s'" + cell_dofree_all = 'all' + ret = [cell_dofree_str % cell_dofree_all] + + map_data = { + 'fix_volume': 'shape', + 'fix_area': '2Dshape', + 'fix_xy': '2Dxy', + 'isotropic': 'volume' + } + + if len(set(kwargs).intersection(map_data)) > 1: + logger.error("only one of %s can be true" % ', '.join(map_data)) + return ret - cell_dofree = "cell_dofree = 'all'" - if (fix_volume and fix_area) or (fix_volume and isotropic) or (fix_area and isotropic): - logger.error("only one of fix_volume fix_area and isotropic can be true") - return [cell_dofree] + for key, val in map_data.items(): + if kwargs.get(key): + return [cell_dofree_str % val] - if fix_volume: - cell_dofree = "cell_dofree = 'shape'" - if fix_area: - cell_dofree = "cell_dofree = '2Dshape'" - if isotropic: - cell_dofree = "cell_dofree = 'volume'" - return [cell_dofree] + return ret def neb_set_system_nat(name, **kwargs): @@ -289,7 +284,7 @@ def set_lda_plus_u_flag(name, **kwargs): for value in iter(related_data if isinstance(related_data, list) else [related_data]): if value.get('@label') != 'no Hubbard' and value['$'] > 0: - lines.append('lda_plus_u = .t.') + lines.append(f" {name} = .true.") break return lines diff --git a/qeschema/schemas/ph_xmlschema.xsd b/qeschema/schemas/ph_xmlschema.xsd index f8d0f39..21db2bd 100644 --- a/qeschema/schemas/ph_xmlschema.xsd +++ b/qeschema/schemas/ph_xmlschema.xsd @@ -89,14 +89,14 @@ Created by Mauro Palumbo   <complexType name="inputPHType"> - <sequence> + <all> <element type="qes_ph:scf_phType" name="scf_ph" minOccurs="0" /> <element type="qes_ph:filesType" name="files" minOccurs="0" /> <element type="qes_ph:control_phType" name="control_ph" minOccurs="0" /> <element type="qes_ph:control_jobType" name="control_job" minOccurs="0" /> <element type="qes_ph:control_dielType" name="control_diel" minOccurs="0"/> <element type="qes_ph:control_qplotType" name="control_qplot" - minOccurs="0" /> + minOccurs="0" /> <element type="qes_ph:miscellaneaType" name="miscellanea" minOccurs="0" />  <element type="qes_ph:irr_reprType" name="irr_repr" minOccurs="0" /> @@ -107,8 +107,9 @@ Created by Mauro Palumbo <element type="qes_ph:lraman_optionsType" name="lraman_options" minOccurs="0" />  + <element name="xq" type="qes_ph:k_pointType" minOccurs="0"/> <element type="qes_ph:q_pointsType" name="q_points" minOccurs="0" /> - </sequence> + </all> </complexType>  <complexType name="scf_phType"> @@ -173,7 +174,7 @@ Created by Mauro Palumbo <complexType name="miscellaneaType"> <sequence>  - <element type="qes_ph:amassType" name="amass" minOccurs="0" /> + <element type="qes_ph:amassType" name="amass" minOccurs="0" maxOccurs="unbounded"/> <element type="double" name="start_magnetization" minOccurs="0" /> <element type="nonNegativeInteger" name="verbosity" default="0" minOccurs="0" /> @@ -181,6 +182,7 @@ Created by Mauro Palumbo <element type="boolean" name="low_directory_check" default="false" minOccurs="0" /> <element type="boolean" name="nogg" default="false" minOccurs="0" /> + <element type="qes_ph:monkhorst_packType" name="nscf_MPgrid" minOccurs="0"/> </sequence> </complexType>  @@ -202,9 +204,8 @@ Created by Mauro Palumbo <element type="positiveInteger" name="start_irr" default="1" minOccurs="0" /> <element type="positiveInteger" name="last_irr" default="3" - minOccurs="0" /> - <element type="nonNegativeInteger" name="nat_todo" default="0" - minOccurs="0" /> + minOccurs="0" /> + <element type="qes_ph:nat_todoType" name="nat_todo" minOccurs="0" maxOccurs="1"/> <element type="nonNegativeInteger" name="modenum" default="0" minOccurs="0" /> <element type="boolean" name="only_init" default="false" @@ -217,7 +218,9 @@ Created by Mauro Palumbo <complexType name="electron_phonon_optionsType"> <sequence> <element type="qes_ph:electron_phononType" name="electron_phonon" - minOccurs="0" /> + minOccurs="0" /> + <element type="double" name="sigma" default="0.02" minOccurs="0"/> + <element type="integer" name="nsigma" default="10" minOccurs="0"/> <element type="qes_ph:dvscf_starType" name="dvscf_star" minOccurs="0" /> <element type="qes_ph:drho_starType" name="drho_star" minOccurs="0" /> </sequence> @@ -270,10 +273,8 @@ Created by Mauro Palumbo <complexType name="q_pointsType"> <sequence>  - <element type="qes_ph:q_monkhorst_packType" name="monkhorst_pack" - minOccurs="0" maxOccurs="1"/> + <element type="qes_ph:q_monkhorst_packType" name="grid" minOccurs="0" maxOccurs="1"/>  - <element type="positiveInteger" name="nq"/>  <element type="nonNegativeInteger" name="nqs" default="1" minOccurs="0" /> <element type="qes_ph:q_points_listType" name="q_points_list" minOccurs="0" /> @@ -295,6 +296,14 @@ Created by Mauro Palumbo </extension> </simpleContent> </complexType> +  + <complexType name="nat_todoType"> + <sequence> + <element type="positiveInteger" name="atom" minOccurs="0" maxOccurs="unbounded"/> + </sequence> + <attribute type="nonNegativeInteger" name="natom"/> + </complexType> +    <complexType name="k_points_IBZType"> diff --git a/qeschema/schemas/qes.xsd b/qeschema/schemas/qes.xsd index 176ecc1..3b7dd7d 100644 --- a/qeschema/schemas/qes.xsd +++ b/qeschema/schemas/qes.xsd @@ -287,6 +287,9 @@ Authors: Antonio Zambon, Paolo Giannozzi, Pietro Delugas <enumeration value="VDW-DF2-C09"/> <enumeration value="VDW-DF2-B86R"/> <enumeration value="RVV10"/> + <enumeration value="BEEF"/> + <enumeration value="SCAN"/> + <enumeration value="R2SCAN"/> </restriction> </simpleType> @@ -456,6 +459,7 @@ Authors: Antonio Zambon, Paolo Giannozzi, Pietro Delugas <element type="qes:basisSetItemType" name="fft_grid" minOccurs="0" /> <element type="qes:basisSetItemType" name="fft_smooth" minOccurs="0" /> <element type="qes:basisSetItemType" name="fft_box" minOccurs="0" /> + <element type="boolean" name="spline_ps" minOccurs="0" maxOccurs="1"/> </sequence> </complexType> <complexType name="basis_setType"> @@ -519,6 +523,7 @@ Authors: Antonio Zambon, Paolo Giannozzi, Pietro Delugas <enumeration value="davidson"/> <enumeration value="cg"/> <enumeration value="ppcg"/> + <enumeration value="paro"/> </restriction> </simpleType> <complexType name="k_points_IBZType"> @@ -588,9 +593,12 @@ Authors: Antonio Zambon, Paolo Giannozzi, Pietro Delugas <element type="double" name="pressure" default="0.0" /> <element type="double" name="wmass" minOccurs="0" /> <element type="double" name="cell_factor" minOccurs="0" /> - <element type="boolean" name="fix_volume" minOccurs="0" /> - <element type="boolean" name="fix_area" minOccurs="0" /> - <element type="boolean" name="isotropic" minOccurs="0" /> + <choice minOccurs="0" maxOccurs="1"> + <element type="boolean" name="fix_volume" /> + <element type="boolean" name="fix_area" /> + <element type="boolean" name="fix_xy" /> + <element type="boolean" name="isotropic" /> + </choice> <element type="qes:integerMatrixType" name="free_cell" minOccurs="0" /> </sequence> </complexType> diff --git a/qeschema/schemas/releases/ph-20210706.xsd b/qeschema/schemas/releases/ph-20210706.xsd new file mode 100644 index 0000000..3311196 --- /dev/null +++ b/qeschema/schemas/releases/ph-20210706.xsd @@ -0,0 +1,629 @@ +<?xml version="1.0"?> + + + +<schema xmlns="http://www.w3.org/2001/XMLSchema" + xmlns:qes_ph="http://www.quantum-espresso.org/ns/qes/qes_ph_1.0" + targetNamespace="http://www.quantum-espresso.org/ns/qes/qes_ph_1.0"> + +  + <element name="espressoph" type="qes_ph:espressophType" /> + <complexType name="espressophType"> + <sequence> + <element type="qes_ph:general_infoType" name="general_info" minOccurs="0"/> + <element type="qes_ph:parallel_infoType" name="parallel_info" minOccurs="0"/> + <element type="qes_ph:inputPHType" name="inputPH" /> + <element type="qes_ph:outputPHType" name="outputPH" minOccurs="0" /> + <element type="qes_ph:statusType" name="status" minOccurs="0" /> + <element type="nonNegativeInteger" name="cputime" minOccurs="0" /> + <element type="qes_ph:closedType" name="closed" minOccurs="0" /> + </sequence> + </complexType> + + +  + <complexType name="general_infoType"> + <sequence> + <element type="qes_ph:formatType" name="format"/> + <element type="qes_ph:creatorType" name="creator"/> + <element type="qes_ph:createdType" name="created"/> + <element type="string" name="job"/> + </sequence> + </complexType> + + <complexType name="formatType"> + <simpleContent> + <extension base="string"> + <attribute type="string" name="NAME"/> + <attribute type="string" name="VERSION"/> + </extension> + </simpleContent> + </complexType> + + <complexType name="creatorType"> + <simpleContent> + <extension base="string"> + <attribute type="string" name="NAME"/> + <attribute type="string" name="VERSION"/> + </extension> + </simpleContent> + </complexType> + + <complexType name="createdType"> + <simpleContent> + <extension base="string"> + <attribute type="string" name="DATE"/> + <attribute type="string" name="TIME"/> + </extension> + </simpleContent> + </complexType> + + +  + <complexType name="parallel_infoType"> + <sequence> + <element type="positiveInteger" name="nprocs"/> + <element type="positiveInteger" name="nthreads"/> + <element type="positiveInteger" name="ntasks"/> + <element type="positiveInteger" name="nbgrp"/> + <element type="positiveInteger" name="npool"/> + <element type="positiveInteger" name="ndiag"/> + <element type="positiveInteger" name="nimages" minOccurs="0"/> + </sequence> + </complexType> + + +  + <complexType name="inputPHType"> + <all> + <element type="qes_ph:scf_phType" name="scf_ph" minOccurs="0" /> + <element type="qes_ph:filesType" name="files" minOccurs="0" /> + <element type="qes_ph:control_phType" name="control_ph" minOccurs="0" /> + <element type="qes_ph:control_jobType" name="control_job" minOccurs="0" /> + <element type="qes_ph:control_dielType" name="control_diel" minOccurs="0"/> + <element type="qes_ph:control_qplotType" name="control_qplot" + minOccurs="0" /> + <element type="qes_ph:miscellaneaType" name="miscellanea" minOccurs="0" /> +  + <element type="qes_ph:irr_reprType" name="irr_repr" minOccurs="0" /> +  + <element type="qes_ph:electron_phonon_optionsType" + name="electron_phonon_options" minOccurs="0" /> +  + <element type="qes_ph:lraman_optionsType" name="lraman_options" + minOccurs="0" /> +  + <element name="xq" type="qes_ph:k_pointType" minOccurs="0"/> + <element type="qes_ph:q_pointsType" name="q_points" minOccurs="0" /> + </all> + </complexType> + + <complexType name="scf_phType"> + <sequence> + <element type="double" name="tr2_ph" default="1E-12" minOccurs="0" /> + <element type="positiveInteger" name="niter_ph" default="100" + minOccurs="0" /> + <element type="double" name="alpha_mix" default="0.7" minOccurs="0" /> + <element type="positiveInteger" name="nmix_ph" default="4" + minOccurs="0" /> + </sequence> + </complexType> + + <complexType name="filesType"> + <sequence> + <element type="string" name="prefix" default="pwscf" minOccurs="0" /> + <element type="string" name="outdir" default="./" minOccurs="0" /> + <element type="string" name="fildyn" default="matdyn" minOccurs="0" /> + <element type="string" name="fildrho" default="" minOccurs="0" /> + <element type="string" name="fildvscf" default="" minOccurs="0" /> + <element type="boolean" name="lqdir" default="false" minOccurs="0" /> + </sequence> + </complexType> + + <complexType name="control_phType"> + <sequence> + <element type="boolean" name="ldisp" default="false" minOccurs="0" /> + <element type="boolean" name="epsil" default="false" minOccurs="0" /> + <element type="boolean" name="trans" default="true" minOccurs="0" /> + <element type="boolean" name="zeu" default="false" minOccurs="0" /> + <element type="boolean" name="zue" default="false" minOccurs="0" /> + <element type="boolean" name="elop" default="false" minOccurs="0" /> + <element type="boolean" name="fpol" default="false" minOccurs="0" /> + <element type="boolean" name="lraman" default="false" minOccurs="0" /> + <element type="boolean" name="search_sym" default="true" minOccurs="0"/> + </sequence> + </complexType> + + <complexType name="control_jobType"> + <sequence> + <element type="boolean" name="recover" default="false" minOccurs="0" /> + <element type="double" name="max_seconds" default="1E7" minOccurs="0" /> + </sequence> + </complexType> + + <complexType name="control_dielType"> + <sequence> + <element type="boolean" name="lrpa" default="false" minOccurs="0" /> + <element type="boolean" name="lnoloc" default="false" minOccurs="0" /> + </sequence> + </complexType> + + <complexType name="control_qplotType"> + <sequence> + <element type="boolean" name="qplot" default="false" minOccurs="0" /> + <element type="boolean" name="q2d" default="false" minOccurs="0" /> + <element type="boolean" name="q_in_band_form" default="false" + minOccurs="0" /> + </sequence> + </complexType> + + <complexType name="miscellaneaType"> + <sequence> +  + <element type="qes_ph:amassType" name="amass" minOccurs="0" maxOccurs="unbounded"/> + <element type="double" name="start_magnetization" minOccurs="0" /> + <element type="nonNegativeInteger" name="verbosity" default="0" + minOccurs="0" /> + <element type="boolean" name="reduce_io" default="false" minOccurs="0"/> + <element type="boolean" name="low_directory_check" default="false" + minOccurs="0" /> + <element type="boolean" name="nogg" default="false" minOccurs="0" /> + <element type="qes_ph:monkhorst_packType" name="nscf_MPgrid" minOccurs="0"/> + </sequence> + </complexType> + + <complexType name="amassType"> + <simpleContent> + <extension base="double"> + <attribute type="positiveInteger" name="atom"/> + </extension> + </simpleContent> + </complexType> + + <complexType name="irr_reprType"> + <sequence> + <sequence> + <element type="positiveInteger" name="start_q" default="1" + minOccurs="0" /> + <element type="positiveInteger" name="last_q" default="1" + minOccurs="0" /> + <element type="positiveInteger" name="start_irr" default="1" + minOccurs="0" /> + <element type="positiveInteger" name="last_irr" default="3" + minOccurs="0" /> + <element type="nonNegativeInteger" name="nat_todo" default="0" + minOccurs="0" /> + <element type="nonNegativeInteger" name="modenum" default="0" + minOccurs="0" /> + <element type="boolean" name="only_init" default="false" + minOccurs="0" /> + <element type="boolean" name="ldiag" default="false" minOccurs="0" /> + </sequence> + </sequence> + </complexType> + + <complexType name="electron_phonon_optionsType"> + <sequence> + <element type="qes_ph:electron_phononType" name="electron_phonon" + minOccurs="0" /> + <element type="double" name="sigma" default="0.02" minOccurs="0"/> + <element type="integer" name="nsigma" default="10" minOccurs="0"/> + <element type="qes_ph:dvscf_starType" name="dvscf_star" minOccurs="0" /> + <element type="qes_ph:drho_starType" name="drho_star" minOccurs="0" /> + </sequence> + </complexType> + + <simpleType name="electron_phononType"> + <restriction base="string"> + <enumeration value="" /> + <enumeration value="simple" /> + <enumeration value="interpolated" /> + </restriction> + </simpleType> + + <complexType name="dvscf_starType"> + <sequence> + <element type="boolean" name="open" default="false" + minOccurs="0" /> + <element type="string" name="dir" minOccurs="0" /> + <element type="string" name="ext" default="dvscf" + minOccurs="0" /> + <element type="string" name="basis" default="cartesian" + minOccurs="0" /> + <element type="boolean" name="pat" default="true" + minOccurs="0" /> + </sequence> + </complexType> +  + <complexType name="drho_starType"> + <sequence> + <element type="boolean" name="open" default="false" + minOccurs="0" /> + <element type="string" name="dir" minOccurs="0" /> + <element type="string" name="ext" default="drho" + minOccurs="0" /> + <element type="string" name="basis" default="modes" + minOccurs="0" /> + <element type="boolean" name="pat" default="true" + minOccurs="0" /> + </sequence> + </complexType> + + <complexType name="lraman_optionsType"> + <sequence> + <element type="double" name="eth_rps" default="1E-9" minOccurs="0" /> + <element type="double" name="eth_ns" default="1E-12" minOccurs="0" /> + <element type="double" name="dek" default="1E-3" minOccurs="0" /> + </sequence> + </complexType> + + <complexType name="q_pointsType"> + <sequence> +  + <element type="qes_ph:q_monkhorst_packType" name="grid" minOccurs="0" maxOccurs="1"/> +  +  + <element type="nonNegativeInteger" name="nqs" default="1" minOccurs="0" /> + <element type="qes_ph:q_points_listType" name="q_points_list" minOccurs="0" /> + </sequence> + </complexType> +  + <complexType name="q_points_listType"> + <sequence> + <element type="qes_ph:k_pointType" name="q_point" maxOccurs="unbounded" + minOccurs="0"/> + </sequence> + </complexType> + <complexType name="q_monkhorst_packType"> + <simpleContent> + <extension base="string"> + <attribute type="positiveInteger" name="nq1"/> + <attribute type="positiveInteger" name="nq2"/> + <attribute type="positiveInteger" name="nq3"/> + </extension> + </simpleContent> + </complexType> + +  + <complexType name="k_points_IBZType"> + <sequence> + <element type="qes_ph:monkhorst_packType" name="monkhorst_pack" minOccurs="0" + maxOccurs="1"/> + <element type="positiveInteger" name="nk"/> + <element type="qes_ph:k_pointType" name="k_point" maxOccurs="unbounded" + minOccurs="0"/> + </sequence> + </complexType> + <complexType name="monkhorst_packType"> + <simpleContent> + <extension base="string"> + <attribute type="positiveInteger" name="nk1"/> + <attribute type="positiveInteger" name="nk2"/> + <attribute type="positiveInteger" name="nk3"/> + <attribute type="boolean" name="k1"/> + <attribute type="boolean" name="k2"/> + <attribute type="boolean" name="k3"/> + </extension> + </simpleContent> + </complexType> + <complexType name="k_pointType"> + <simpleContent> + <extension base="qes_ph:d3vectorType"> + <attribute type="double" name="weight" use="optional"/> + </extension> + </simpleContent> + </complexType> +  +  + <complexType name="closedType"> + <simpleContent> + <extension base="string"> + <attribute type="string" name="DATE"/> + <attribute type="string" name="TIME"/> + </extension> + </simpleContent> + </complexType> + +  + <simpleType name="statusType"> + <restriction base="unsignedByte"> + <enumeration value="0" /> + <enumeration value="1" /> + <enumeration value="2" /> + <enumeration value="3" /> + <enumeration value="255"/> + </restriction> + </simpleType> + +  + <complexType name="outputPHType"> + <sequence> +  + <element type="qes_ph:atomic_speciesType" name="atomic_species"/> + <element type="qes_ph:atomic_structureType" name="atomic_structure"/> +  + <element type="qes_ph:unit_cell_volumeType" name="unit_cell_volume"/> + <element type="nonNegativeInteger" name="ibrav" + minOccurs="0" /> + <element type="qes_ph:cell_dimensionsType" name="cell_dimensions" + minOccurs="0" /> + <element type="double" name="start_magnetization" minOccurs="0" /> + <element type="nonNegativeInteger" name="number_of_q" minOccurs="0"/> + <element type="nonNegativeInteger" name="number_of_irr_q" + minOccurs="0"/> + <element type="qes_ph:irr_repr_listType" name="irr_repr_list" + minOccurs="0" /> + <element type="qes_ph:dynamical_matType" name="dynamical_mat" + minOccurs="0" /> + <element type="qes_ph:force_constantsType" name="force_constants" + minOccurs="0" /> + <element type="qes_ph:frequenciesType" name="frequencies" + minOccurs="0"/> + </sequence> + </complexType> + + +  + <complexType name="unit_cell_volumeType"> + <simpleContent> + <extension base="double"> + <attribute type="string" name="unit" use="required"/> + </extension> + </simpleContent> + </complexType> + +  + <simpleType name="cell_dimensionsType"> + <restriction> + <simpleType> + <list itemType="double"/> + </simpleType> + <length value="6"/> + </restriction> + </simpleType> + +  + <complexType name="irr_repr_listType"> + <sequence> + <element type="qes_ph:repr_at_qType" name="repr_at_q" maxOccurs="unbounded" /> + </sequence> + </complexType> + + <complexType name="repr_at_qType"> + <sequence> + <element type="qes_ph:d3vectorType" name="q_point" /> + <element type="qes_ph:reprType" name="repr" maxOccurs="unbounded" /> + </sequence> + </complexType> + + <complexType name="reprType"> + <sequence> + <element type="qes_ph:modeType" name="mode" maxOccurs="unbounded" /> + </sequence> + <attribute type="nonNegativeInteger" name="number" /> + <attribute type="nonNegativeInteger" name="n_modes" /> + <attribute type="string" name="descr" /> + </complexType> + + <complexType name="modeType"> + <sequence> + <element type="qes_ph:displacementType" name="displacement" maxOccurs="unbounded"/> + </sequence> + <attribute type="nonNegativeInteger" name="number" /> + </complexType> + + +  + <complexType name="dynamical_matType"> + <sequence> + <element type="qes_ph:D_elements_symmequivType" + name="D_elements_symmequiv" maxOccurs="unbounded" /> + </sequence> + </complexType> + +  + <complexType name="D_elements_symmequivType"> + <sequence> + <element type="qes_ph:D_qType" name="D_q" maxOccurs="unbounded" /> + </sequence> + <attribute type="nonNegativeInteger" name="number_of_equiv_q" /> + </complexType> + +  + <complexType name="D_qType"> + <sequence> + <element type="qes_ph:d3vectorType" name="q_point" /> + <element type="qes_ph:phi_qType" name="phi_q" maxOccurs="unbounded" /> + </sequence> + </complexType> + +  + <complexType name="phi_qType"> + <simpleContent> + <extension base="qes_ph:d3complexDType"> + <attribute type="positiveInteger" name="at1" /> + <attribute type="positiveInteger" name="at2" /> + </extension> + </simpleContent> + </complexType> + + +  + <complexType name="frequenciesType"> + <sequence> + <element type="qes_ph:frequencyType" name="frequency" minOccurs="0" + maxOccurs="unbounded"/> + </sequence> + </complexType> + +  + <complexType name="frequencyType"> + <sequence> +  + <element type="qes_ph:omegaType" name="omega" maxOccurs="2"/> + <element type="qes_ph:displacementType" name="displacement" maxOccurs="unbounded"/> + </sequence> + <attribute type="integer" name="index" use="required"/> + </complexType> + +  + <complexType name="omegaType"> + <simpleContent> + <extension base="double"> + <attribute type="string" name="unit" use="required"/> + </extension> + </simpleContent> + </complexType> + + <complexType name="displacementType"> + <sequence> + <element type="qes_ph:disp_x_y_zType" name="x" maxOccurs="1"/> + <element type="qes_ph:disp_x_y_zType" name="y" maxOccurs="1"/> + <element type="qes_ph:disp_x_y_zType" name="z" maxOccurs="1"/> + </sequence> + <attribute type="integer" name="atom" use="required"/> + </complexType> + +  + <simpleType name="disp_x_y_zType"> + <restriction> + <simpleType> + <list itemType="double"/> + </simpleType> + <length value="2"/> + </restriction> + </simpleType> + + +  + <complexType name="force_constantsType"> + <sequence> + <element type="qes_ph:phi_rType" name="phi_r" maxOccurs="unbounded" /> + </sequence> + <attribute type="positiveInteger" name="nq1" /> + <attribute type="positiveInteger" name="nq2" /> + <attribute type="positiveInteger" name="nq3" /> + </complexType> + +  + <complexType name="phi_rType"> + <sequence> + <element type="qes_ph:phi_r_elType" name="phi_r_el" + maxOccurs="unbounded" /> + </sequence> + <attribute type="positiveInteger" name="at1" /> + <attribute type="positiveInteger" name="at2" /> + </complexType> + +  + <complexType name="phi_r_elType"> + <simpleContent> + <extension base="double"> + <attribute type="positiveInteger" name="m1" /> + <attribute type="positiveInteger" name="m2" /> + <attribute type="positiveInteger" name="m3" /> + </extension> + </simpleContent> + </complexType> + +  + <complexType name="atomic_speciesType"> + <sequence> + <element type="qes_ph:specieType" name="specie" maxOccurs="unbounded"/> + </sequence> + <attribute type="positiveInteger" name="ntyp"/> + </complexType> + <complexType name="specieType"> + <sequence> + <element type="double" name="mass" minOccurs="0" /> + <element type="string" name="pseudo_file"/> + <element type="double" name="starting_magnetization" minOccurs="0"/> + </sequence> + <attribute type="string" name="name"/> + </complexType> + <complexType name="atomic_structureType"> + <sequence> + <element type="qes_ph:atomic_positionsType" name="atomic_positions" + minOccurs="0" /> + <element type="qes_ph:wyckoff_positionsType" name="wyckoff_positions" + minOccurs="0" /> + <element type="qes_ph:cellType" name="cell"/> + </sequence> + <attribute type="positiveInteger" name="nat" use="optional"/> + <attribute type="double" name="alat" use="optional"/> + </complexType> + <complexType name="atomic_positionsType"> + <sequence> + <element type="qes_ph:atomType" name="atom" maxOccurs="unbounded" /> + </sequence> + </complexType> + <complexType name="atomType"> + <simpleContent> + <extension base="qes_ph:d3vectorType"> + <attribute type="string" name="name" /> + <attribute type="string" name="position" use="optional" /> + </extension> + </simpleContent> + </complexType> + <complexType name="wyckoff_positionsType"> + <sequence> + <element type="qes_ph:atomType" name="atom"/> + </sequence> + <attribute type="unsignedByte" name="space_group" use="optional"/> + <attribute type="string" name="more_options" use="optional"/> + </complexType> + <complexType name="cellType"> + <sequence> + <element type="qes_ph:d3vectorType" name="a1"/> + <element type="qes_ph:d3vectorType" name="a2"/> + <element type="qes_ph:d3vectorType" name="a3"/> + </sequence> + </complexType> + +  + <simpleType name="d3realDType"> + <restriction> + <simpleType> + <list itemType="double"/> + </simpleType> + <length value="9"/> + </restriction> + </simpleType> + +  + <simpleType name="d3complexDType"> + <restriction> + <simpleType> + <list itemType="double"/> + </simpleType> + <length value="18"/> + </restriction> + </simpleType> + + +  + <simpleType name="d3vectorType"> + <restriction> + <simpleType> + <list itemType="double"/> + </simpleType> + <length value="3"/> + </restriction> + </simpleType> + +</schema> diff --git a/requirements-dev.txt b/requirements-dev.txt index 52d8698..8b6d1b9 100644 --- a/requirements-dev.txt +++ b/requirements-dev.txt @@ -1,9 +1,12 @@ # Requirements for setup a development environment for the qeschema package. setuptools tox +flake8 coverage xmlschema>=1.3.0 pyyaml +numpy +h5py Sphinx sphinx_rtd_theme -e . diff --git a/setup.py b/setup.py index 216c1b5..86d6604 100644 --- a/setup.py +++ b/setup.py @@ -15,8 +15,8 @@ setup( name='qeschema', - version='1.1.0', - install_requires=['xmlschema>=1.3.0', 'pyyaml'], + version='1.2.0', + install_requires=['xmlschema>=1.3.0', 'pyyaml', 'numpy'], packages=['qeschema'], package_data={'qeschema': ['schemas/*.xsd']}, scripts = ['scripts/xml2qeinput.py', 'scripts/yaml2qeinput.py'], @@ -37,8 +37,8 @@ 'Programming Language :: Python :: 3.7', 'Programming Language :: Python :: 3.8', 'Programming Language :: Python :: 3.9', + 'Programming Language :: Python :: 3.10', 'Programming Language :: Python :: Implementation :: CPython', - 'Programming Language :: Python :: Implementation :: PyPy', 'Topic :: Scientific/Engineering :: Physics', 'Topic :: Utilities', ] diff --git a/tox.ini b/tox.ini index 07297d1..ae9347a 100644 --- a/tox.ini +++ b/tox.ini @@ -4,7 +4,7 @@ # and then run "tox" from this directory. [tox] -envlist = py{36,37,38,39}, pypy3, docs, flake8, coverage +envlist = py{36,37,38,39,310}, docs, flake8, coverage skip_missing_interpreters = true toxworkdir = {homedir}/.tox/qeschema @@ -12,6 +12,8 @@ toxworkdir = {homedir}/.tox/qeschema deps = xmlschema>=1.3.0 pyyaml + numpy + h5py docs: Sphinx docs: sphinx_rtd_theme flake8: flake8 @@ -30,7 +32,7 @@ max-line-length = 119 [testenv:flake8] commands = - flake8 --ignore=F401,F403,F405,F811,F821 qeschema + flake8 qeschema [testenv:coverage] commands =

Propagate dftd3_version, dftd3_threebody from XML to .in Also fix a test.

2021-11-30T14:39:50

0.0

[]

QEF/qeschema

QEF__qeschema-12

3185274c496dd4a5ef657b6718cb90fe08999765

diff --git a/qeschema/cards.py b/qeschema/cards.py index 82ef609..cc055e9 100644 --- a/qeschema/cards.py +++ b/qeschema/cards.py @@ -233,7 +233,6 @@ def get_cell_parameters_card(name, **kwargs): except KeyError: logger.error("Missing required arguments when building CELL_PARAMETERS card!") return [] - # Add cell parameters card cells = atomic_structure.get('cell', {}) if cells: @@ -262,7 +261,7 @@ def get_qpoints_card(name, **kwargs): qplot = kwargs.get('qplot', False) if not qplot and not ldisp: try: - xq = kwargs['xq_dir'] + xq = kwargs['xq'] except KeyError: xq = [0.e0, 0.e0, 0.e0] line = "{:6.4f} {:8.4f} {:8.4f}".format(xq[0], xq[1], xq[2]) diff --git a/qeschema/converters.py b/qeschema/converters.py index 4e7fae4..73691a6 100644 --- a/qeschema/converters.py +++ b/qeschema/converters.py @@ -327,7 +327,7 @@ class PwInputConverter(RawInputConverter): '$': [ ('SYSTEM[ibrav]', options.set_ibrav_to_zero, None), ("ATOMIC_POSITIONS", cards.get_atomic_positions_cell_card, None), - ("CELL_PARAMETERS", cards.get_cell_parameters_card, None) + ("CELL_PARAMETERS", cards.get_cell_parameters_card, None) ], 'atomic_positions': ('ATOMIC_FORCES', cards.get_atomic_forces_card, None), }, @@ -349,7 +349,8 @@ class PwInputConverter(RawInputConverter): 'dftU': { 'lda_plus_u_kind': 'SYSTEM[lda_plus_u_kind]', 'Hubbard_U': { - '$': ('SYSTEM[Hubbard_U]', options.get_specie_related_values, None), + '$': [('SYSTEM[Hubbard_U]', options.get_specie_related_values, None), + ('SYSTEM[lda_plus_u]',options.set_lda_plus_u_flag,None)] }, 'Hubbard_J0': { '$': ('SYSTEM[Hubbard_J0]', options.get_specie_related_values, None), @@ -554,7 +555,9 @@ class PhononInputConverter(RawInputConverter): Convert to/from Fortran input for Phonon. """ PHONON_TEMPLATE_MAP = { - 'xq': ('qPointsSpecs', cards.get_qpoints_card, None), + 'xq': { + '$': ('qPointsSpecs', cards.get_qpoints_card, None), + }, 'scf_ph': { 'tr2_ph': "INPUTPH[tr2_ph]", 'niter_ph': "INPUTPH[niter_ph]", @@ -642,9 +645,9 @@ class PhononInputConverter(RawInputConverter): }, 'q_points': { 'grid': { - 'nq1': "INPUTPH[nq1]", - 'nq2': "INPUTPH[nq2]", - 'nq3': "INPUTPH[nq3]" + '@nq1': "INPUTPH[nq1]", + '@nq2': "INPUTPH[nq2]", + '@nq3': "INPUTPH[nq3]" }, 'q_points_list': ('qPointsSpecs', cards.get_qpoints_card, None), 'nqs': ('qPointsSpecs', cards.get_qpoints_card, None) diff --git a/qeschema/options.py b/qeschema/options.py index b6040fb..e27f45b 100644 --- a/qeschema/options.py +++ b/qeschema/options.py @@ -126,6 +126,7 @@ def set_ibrav_to_zero(name, **_kwargs): return [line] + def get_system_eamp(name, **kwargs): try: electric_potential = kwargs['electric_potential'] @@ -289,7 +290,7 @@ def set_lda_plus_u_flag(name, **kwargs): for value in iter(related_data if isinstance(related_data, list) else [related_data]): if value.get('@label') != 'no Hubbard' and value['$'] > 0: - lines.append('lda_plus_u = .t.') + lines.append(f" {name} = .true.") break return lines diff --git a/qeschema/schemas/ph_xmlschema.xsd b/qeschema/schemas/ph_xmlschema.xsd index f8d0f39..3311196 100644 --- a/qeschema/schemas/ph_xmlschema.xsd +++ b/qeschema/schemas/ph_xmlschema.xsd @@ -89,14 +89,14 @@ Created by Mauro Palumbo   <complexType name="inputPHType"> - <sequence> + <all> <element type="qes_ph:scf_phType" name="scf_ph" minOccurs="0" /> <element type="qes_ph:filesType" name="files" minOccurs="0" /> <element type="qes_ph:control_phType" name="control_ph" minOccurs="0" /> <element type="qes_ph:control_jobType" name="control_job" minOccurs="0" /> <element type="qes_ph:control_dielType" name="control_diel" minOccurs="0"/> <element type="qes_ph:control_qplotType" name="control_qplot" - minOccurs="0" /> + minOccurs="0" /> <element type="qes_ph:miscellaneaType" name="miscellanea" minOccurs="0" />  <element type="qes_ph:irr_reprType" name="irr_repr" minOccurs="0" /> @@ -107,8 +107,9 @@ Created by Mauro Palumbo <element type="qes_ph:lraman_optionsType" name="lraman_options" minOccurs="0" />  + <element name="xq" type="qes_ph:k_pointType" minOccurs="0"/> <element type="qes_ph:q_pointsType" name="q_points" minOccurs="0" /> - </sequence> + </all> </complexType>  <complexType name="scf_phType"> @@ -173,7 +174,7 @@ Created by Mauro Palumbo <complexType name="miscellaneaType"> <sequence>  - <element type="qes_ph:amassType" name="amass" minOccurs="0" /> + <element type="qes_ph:amassType" name="amass" minOccurs="0" maxOccurs="unbounded"/> <element type="double" name="start_magnetization" minOccurs="0" /> <element type="nonNegativeInteger" name="verbosity" default="0" minOccurs="0" /> @@ -181,6 +182,7 @@ Created by Mauro Palumbo <element type="boolean" name="low_directory_check" default="false" minOccurs="0" /> <element type="boolean" name="nogg" default="false" minOccurs="0" /> + <element type="qes_ph:monkhorst_packType" name="nscf_MPgrid" minOccurs="0"/> </sequence> </complexType>  @@ -217,7 +219,9 @@ Created by Mauro Palumbo <complexType name="electron_phonon_optionsType"> <sequence> <element type="qes_ph:electron_phononType" name="electron_phonon" - minOccurs="0" /> + minOccurs="0" /> + <element type="double" name="sigma" default="0.02" minOccurs="0"/> + <element type="integer" name="nsigma" default="10" minOccurs="0"/> <element type="qes_ph:dvscf_starType" name="dvscf_star" minOccurs="0" /> <element type="qes_ph:drho_starType" name="drho_star" minOccurs="0" /> </sequence> @@ -270,10 +274,8 @@ Created by Mauro Palumbo <complexType name="q_pointsType"> <sequence>  - <element type="qes_ph:q_monkhorst_packType" name="monkhorst_pack" - minOccurs="0" maxOccurs="1"/> + <element type="qes_ph:q_monkhorst_packType" name="grid" minOccurs="0" maxOccurs="1"/>  - <element type="positiveInteger" name="nq"/>  <element type="nonNegativeInteger" name="nqs" default="1" minOccurs="0" /> <element type="qes_ph:q_points_listType" name="q_points_list" minOccurs="0" /> diff --git a/qeschema/schemas/releases/ph-20210706.xsd b/qeschema/schemas/releases/ph-20210706.xsd new file mode 100644 index 0000000..3311196 --- /dev/null +++ b/qeschema/schemas/releases/ph-20210706.xsd @@ -0,0 +1,629 @@ +<?xml version="1.0"?> + + + +<schema xmlns="http://www.w3.org/2001/XMLSchema" + xmlns:qes_ph="http://www.quantum-espresso.org/ns/qes/qes_ph_1.0" + targetNamespace="http://www.quantum-espresso.org/ns/qes/qes_ph_1.0"> + +  + <element name="espressoph" type="qes_ph:espressophType" /> + <complexType name="espressophType"> + <sequence> + <element type="qes_ph:general_infoType" name="general_info" minOccurs="0"/> + <element type="qes_ph:parallel_infoType" name="parallel_info" minOccurs="0"/> + <element type="qes_ph:inputPHType" name="inputPH" /> + <element type="qes_ph:outputPHType" name="outputPH" minOccurs="0" /> + <element type="qes_ph:statusType" name="status" minOccurs="0" /> + <element type="nonNegativeInteger" name="cputime" minOccurs="0" /> + <element type="qes_ph:closedType" name="closed" minOccurs="0" /> + </sequence> + </complexType> + + +  + <complexType name="general_infoType"> + <sequence> + <element type="qes_ph:formatType" name="format"/> + <element type="qes_ph:creatorType" name="creator"/> + <element type="qes_ph:createdType" name="created"/> + <element type="string" name="job"/> + </sequence> + </complexType> + + <complexType name="formatType"> + <simpleContent> + <extension base="string"> + <attribute type="string" name="NAME"/> + <attribute type="string" name="VERSION"/> + </extension> + </simpleContent> + </complexType> + + <complexType name="creatorType"> + <simpleContent> + <extension base="string"> + <attribute type="string" name="NAME"/> + <attribute type="string" name="VERSION"/> + </extension> + </simpleContent> + </complexType> + + <complexType name="createdType"> + <simpleContent> + <extension base="string"> + <attribute type="string" name="DATE"/> + <attribute type="string" name="TIME"/> + </extension> + </simpleContent> + </complexType> + + +  + <complexType name="parallel_infoType"> + <sequence> + <element type="positiveInteger" name="nprocs"/> + <element type="positiveInteger" name="nthreads"/> + <element type="positiveInteger" name="ntasks"/> + <element type="positiveInteger" name="nbgrp"/> + <element type="positiveInteger" name="npool"/> + <element type="positiveInteger" name="ndiag"/> + <element type="positiveInteger" name="nimages" minOccurs="0"/> + </sequence> + </complexType> + + +  + <complexType name="inputPHType"> + <all> + <element type="qes_ph:scf_phType" name="scf_ph" minOccurs="0" /> + <element type="qes_ph:filesType" name="files" minOccurs="0" /> + <element type="qes_ph:control_phType" name="control_ph" minOccurs="0" /> + <element type="qes_ph:control_jobType" name="control_job" minOccurs="0" /> + <element type="qes_ph:control_dielType" name="control_diel" minOccurs="0"/> + <element type="qes_ph:control_qplotType" name="control_qplot" + minOccurs="0" /> + <element type="qes_ph:miscellaneaType" name="miscellanea" minOccurs="0" /> +  + <element type="qes_ph:irr_reprType" name="irr_repr" minOccurs="0" /> +  + <element type="qes_ph:electron_phonon_optionsType" + name="electron_phonon_options" minOccurs="0" /> +  + <element type="qes_ph:lraman_optionsType" name="lraman_options" + minOccurs="0" /> +  + <element name="xq" type="qes_ph:k_pointType" minOccurs="0"/> + <element type="qes_ph:q_pointsType" name="q_points" minOccurs="0" /> + </all> + </complexType> + + <complexType name="scf_phType"> + <sequence> + <element type="double" name="tr2_ph" default="1E-12" minOccurs="0" /> + <element type="positiveInteger" name="niter_ph" default="100" + minOccurs="0" /> + <element type="double" name="alpha_mix" default="0.7" minOccurs="0" /> + <element type="positiveInteger" name="nmix_ph" default="4" + minOccurs="0" /> + </sequence> + </complexType> + + <complexType name="filesType"> + <sequence> + <element type="string" name="prefix" default="pwscf" minOccurs="0" /> + <element type="string" name="outdir" default="./" minOccurs="0" /> + <element type="string" name="fildyn" default="matdyn" minOccurs="0" /> + <element type="string" name="fildrho" default="" minOccurs="0" /> + <element type="string" name="fildvscf" default="" minOccurs="0" /> + <element type="boolean" name="lqdir" default="false" minOccurs="0" /> + </sequence> + </complexType> + + <complexType name="control_phType"> + <sequence> + <element type="boolean" name="ldisp" default="false" minOccurs="0" /> + <element type="boolean" name="epsil" default="false" minOccurs="0" /> + <element type="boolean" name="trans" default="true" minOccurs="0" /> + <element type="boolean" name="zeu" default="false" minOccurs="0" /> + <element type="boolean" name="zue" default="false" minOccurs="0" /> + <element type="boolean" name="elop" default="false" minOccurs="0" /> + <element type="boolean" name="fpol" default="false" minOccurs="0" /> + <element type="boolean" name="lraman" default="false" minOccurs="0" /> + <element type="boolean" name="search_sym" default="true" minOccurs="0"/> + </sequence> + </complexType> + + <complexType name="control_jobType"> + <sequence> + <element type="boolean" name="recover" default="false" minOccurs="0" /> + <element type="double" name="max_seconds" default="1E7" minOccurs="0" /> + </sequence> + </complexType> + + <complexType name="control_dielType"> + <sequence> + <element type="boolean" name="lrpa" default="false" minOccurs="0" /> + <element type="boolean" name="lnoloc" default="false" minOccurs="0" /> + </sequence> + </complexType> + + <complexType name="control_qplotType"> + <sequence> + <element type="boolean" name="qplot" default="false" minOccurs="0" /> + <element type="boolean" name="q2d" default="false" minOccurs="0" /> + <element type="boolean" name="q_in_band_form" default="false" + minOccurs="0" /> + </sequence> + </complexType> + + <complexType name="miscellaneaType"> + <sequence> +  + <element type="qes_ph:amassType" name="amass" minOccurs="0" maxOccurs="unbounded"/> + <element type="double" name="start_magnetization" minOccurs="0" /> + <element type="nonNegativeInteger" name="verbosity" default="0" + minOccurs="0" /> + <element type="boolean" name="reduce_io" default="false" minOccurs="0"/> + <element type="boolean" name="low_directory_check" default="false" + minOccurs="0" /> + <element type="boolean" name="nogg" default="false" minOccurs="0" /> + <element type="qes_ph:monkhorst_packType" name="nscf_MPgrid" minOccurs="0"/> + </sequence> + </complexType> + + <complexType name="amassType"> + <simpleContent> + <extension base="double"> + <attribute type="positiveInteger" name="atom"/> + </extension> + </simpleContent> + </complexType> + + <complexType name="irr_reprType"> + <sequence> + <sequence> + <element type="positiveInteger" name="start_q" default="1" + minOccurs="0" /> + <element type="positiveInteger" name="last_q" default="1" + minOccurs="0" /> + <element type="positiveInteger" name="start_irr" default="1" + minOccurs="0" /> + <element type="positiveInteger" name="last_irr" default="3" + minOccurs="0" /> + <element type="nonNegativeInteger" name="nat_todo" default="0" + minOccurs="0" /> + <element type="nonNegativeInteger" name="modenum" default="0" + minOccurs="0" /> + <element type="boolean" name="only_init" default="false" + minOccurs="0" /> + <element type="boolean" name="ldiag" default="false" minOccurs="0" /> + </sequence> + </sequence> + </complexType> + + <complexType name="electron_phonon_optionsType"> + <sequence> + <element type="qes_ph:electron_phononType" name="electron_phonon" + minOccurs="0" /> + <element type="double" name="sigma" default="0.02" minOccurs="0"/> + <element type="integer" name="nsigma" default="10" minOccurs="0"/> + <element type="qes_ph:dvscf_starType" name="dvscf_star" minOccurs="0" /> + <element type="qes_ph:drho_starType" name="drho_star" minOccurs="0" /> + </sequence> + </complexType> + + <simpleType name="electron_phononType"> + <restriction base="string"> + <enumeration value="" /> + <enumeration value="simple" /> + <enumeration value="interpolated" /> + </restriction> + </simpleType> + + <complexType name="dvscf_starType"> + <sequence> + <element type="boolean" name="open" default="false" + minOccurs="0" /> + <element type="string" name="dir" minOccurs="0" /> + <element type="string" name="ext" default="dvscf" + minOccurs="0" /> + <element type="string" name="basis" default="cartesian" + minOccurs="0" /> + <element type="boolean" name="pat" default="true" + minOccurs="0" /> + </sequence> + </complexType> +  + <complexType name="drho_starType"> + <sequence> + <element type="boolean" name="open" default="false" + minOccurs="0" /> + <element type="string" name="dir" minOccurs="0" /> + <element type="string" name="ext" default="drho" + minOccurs="0" /> + <element type="string" name="basis" default="modes" + minOccurs="0" /> + <element type="boolean" name="pat" default="true" + minOccurs="0" /> + </sequence> + </complexType> + + <complexType name="lraman_optionsType"> + <sequence> + <element type="double" name="eth_rps" default="1E-9" minOccurs="0" /> + <element type="double" name="eth_ns" default="1E-12" minOccurs="0" /> + <element type="double" name="dek" default="1E-3" minOccurs="0" /> + </sequence> + </complexType> + + <complexType name="q_pointsType"> + <sequence> +  + <element type="qes_ph:q_monkhorst_packType" name="grid" minOccurs="0" maxOccurs="1"/> +  +  + <element type="nonNegativeInteger" name="nqs" default="1" minOccurs="0" /> + <element type="qes_ph:q_points_listType" name="q_points_list" minOccurs="0" /> + </sequence> + </complexType> +  + <complexType name="q_points_listType"> + <sequence> + <element type="qes_ph:k_pointType" name="q_point" maxOccurs="unbounded" + minOccurs="0"/> + </sequence> + </complexType> + <complexType name="q_monkhorst_packType"> + <simpleContent> + <extension base="string"> + <attribute type="positiveInteger" name="nq1"/> + <attribute type="positiveInteger" name="nq2"/> + <attribute type="positiveInteger" name="nq3"/> + </extension> + </simpleContent> + </complexType> + +  + <complexType name="k_points_IBZType"> + <sequence> + <element type="qes_ph:monkhorst_packType" name="monkhorst_pack" minOccurs="0" + maxOccurs="1"/> + <element type="positiveInteger" name="nk"/> + <element type="qes_ph:k_pointType" name="k_point" maxOccurs="unbounded" + minOccurs="0"/> + </sequence> + </complexType> + <complexType name="monkhorst_packType"> + <simpleContent> + <extension base="string"> + <attribute type="positiveInteger" name="nk1"/> + <attribute type="positiveInteger" name="nk2"/> + <attribute type="positiveInteger" name="nk3"/> + <attribute type="boolean" name="k1"/> + <attribute type="boolean" name="k2"/> + <attribute type="boolean" name="k3"/> + </extension> + </simpleContent> + </complexType> + <complexType name="k_pointType"> + <simpleContent> + <extension base="qes_ph:d3vectorType"> + <attribute type="double" name="weight" use="optional"/> + </extension> + </simpleContent> + </complexType> +  +  + <complexType name="closedType"> + <simpleContent> + <extension base="string"> + <attribute type="string" name="DATE"/> + <attribute type="string" name="TIME"/> + </extension> + </simpleContent> + </complexType> + +  + <simpleType name="statusType"> + <restriction base="unsignedByte"> + <enumeration value="0" /> + <enumeration value="1" /> + <enumeration value="2" /> + <enumeration value="3" /> + <enumeration value="255"/> + </restriction> + </simpleType> + +  + <complexType name="outputPHType"> + <sequence> +  + <element type="qes_ph:atomic_speciesType" name="atomic_species"/> + <element type="qes_ph:atomic_structureType" name="atomic_structure"/> +  + <element type="qes_ph:unit_cell_volumeType" name="unit_cell_volume"/> + <element type="nonNegativeInteger" name="ibrav" + minOccurs="0" /> + <element type="qes_ph:cell_dimensionsType" name="cell_dimensions" + minOccurs="0" /> + <element type="double" name="start_magnetization" minOccurs="0" /> + <element type="nonNegativeInteger" name="number_of_q" minOccurs="0"/> + <element type="nonNegativeInteger" name="number_of_irr_q" + minOccurs="0"/> + <element type="qes_ph:irr_repr_listType" name="irr_repr_list" + minOccurs="0" /> + <element type="qes_ph:dynamical_matType" name="dynamical_mat" + minOccurs="0" /> + <element type="qes_ph:force_constantsType" name="force_constants" + minOccurs="0" /> + <element type="qes_ph:frequenciesType" name="frequencies" + minOccurs="0"/> + </sequence> + </complexType> + + +  + <complexType name="unit_cell_volumeType"> + <simpleContent> + <extension base="double"> + <attribute type="string" name="unit" use="required"/> + </extension> + </simpleContent> + </complexType> + +  + <simpleType name="cell_dimensionsType"> + <restriction> + <simpleType> + <list itemType="double"/> + </simpleType> + <length value="6"/> + </restriction> + </simpleType> + +  + <complexType name="irr_repr_listType"> + <sequence> + <element type="qes_ph:repr_at_qType" name="repr_at_q" maxOccurs="unbounded" /> + </sequence> + </complexType> + + <complexType name="repr_at_qType"> + <sequence> + <element type="qes_ph:d3vectorType" name="q_point" /> + <element type="qes_ph:reprType" name="repr" maxOccurs="unbounded" /> + </sequence> + </complexType> + + <complexType name="reprType"> + <sequence> + <element type="qes_ph:modeType" name="mode" maxOccurs="unbounded" /> + </sequence> + <attribute type="nonNegativeInteger" name="number" /> + <attribute type="nonNegativeInteger" name="n_modes" /> + <attribute type="string" name="descr" /> + </complexType> + + <complexType name="modeType"> + <sequence> + <element type="qes_ph:displacementType" name="displacement" maxOccurs="unbounded"/> + </sequence> + <attribute type="nonNegativeInteger" name="number" /> + </complexType> + + +  + <complexType name="dynamical_matType"> + <sequence> + <element type="qes_ph:D_elements_symmequivType" + name="D_elements_symmequiv" maxOccurs="unbounded" /> + </sequence> + </complexType> + +  + <complexType name="D_elements_symmequivType"> + <sequence> + <element type="qes_ph:D_qType" name="D_q" maxOccurs="unbounded" /> + </sequence> + <attribute type="nonNegativeInteger" name="number_of_equiv_q" /> + </complexType> + +  + <complexType name="D_qType"> + <sequence> + <element type="qes_ph:d3vectorType" name="q_point" /> + <element type="qes_ph:phi_qType" name="phi_q" maxOccurs="unbounded" /> + </sequence> + </complexType> + +  + <complexType name="phi_qType"> + <simpleContent> + <extension base="qes_ph:d3complexDType"> + <attribute type="positiveInteger" name="at1" /> + <attribute type="positiveInteger" name="at2" /> + </extension> + </simpleContent> + </complexType> + + +  + <complexType name="frequenciesType"> + <sequence> + <element type="qes_ph:frequencyType" name="frequency" minOccurs="0" + maxOccurs="unbounded"/> + </sequence> + </complexType> + +  + <complexType name="frequencyType"> + <sequence> +  + <element type="qes_ph:omegaType" name="omega" maxOccurs="2"/> + <element type="qes_ph:displacementType" name="displacement" maxOccurs="unbounded"/> + </sequence> + <attribute type="integer" name="index" use="required"/> + </complexType> + +  + <complexType name="omegaType"> + <simpleContent> + <extension base="double"> + <attribute type="string" name="unit" use="required"/> + </extension> + </simpleContent> + </complexType> + + <complexType name="displacementType"> + <sequence> + <element type="qes_ph:disp_x_y_zType" name="x" maxOccurs="1"/> + <element type="qes_ph:disp_x_y_zType" name="y" maxOccurs="1"/> + <element type="qes_ph:disp_x_y_zType" name="z" maxOccurs="1"/> + </sequence> + <attribute type="integer" name="atom" use="required"/> + </complexType> + +  + <simpleType name="disp_x_y_zType"> + <restriction> + <simpleType> + <list itemType="double"/> + </simpleType> + <length value="2"/> + </restriction> + </simpleType> + + +  + <complexType name="force_constantsType"> + <sequence> + <element type="qes_ph:phi_rType" name="phi_r" maxOccurs="unbounded" /> + </sequence> + <attribute type="positiveInteger" name="nq1" /> + <attribute type="positiveInteger" name="nq2" /> + <attribute type="positiveInteger" name="nq3" /> + </complexType> + +  + <complexType name="phi_rType"> + <sequence> + <element type="qes_ph:phi_r_elType" name="phi_r_el" + maxOccurs="unbounded" /> + </sequence> + <attribute type="positiveInteger" name="at1" /> + <attribute type="positiveInteger" name="at2" /> + </complexType> + +  + <complexType name="phi_r_elType"> + <simpleContent> + <extension base="double"> + <attribute type="positiveInteger" name="m1" /> + <attribute type="positiveInteger" name="m2" /> + <attribute type="positiveInteger" name="m3" /> + </extension> + </simpleContent> + </complexType> + +  + <complexType name="atomic_speciesType"> + <sequence> + <element type="qes_ph:specieType" name="specie" maxOccurs="unbounded"/> + </sequence> + <attribute type="positiveInteger" name="ntyp"/> + </complexType> + <complexType name="specieType"> + <sequence> + <element type="double" name="mass" minOccurs="0" /> + <element type="string" name="pseudo_file"/> + <element type="double" name="starting_magnetization" minOccurs="0"/> + </sequence> + <attribute type="string" name="name"/> + </complexType> + <complexType name="atomic_structureType"> + <sequence> + <element type="qes_ph:atomic_positionsType" name="atomic_positions" + minOccurs="0" /> + <element type="qes_ph:wyckoff_positionsType" name="wyckoff_positions" + minOccurs="0" /> + <element type="qes_ph:cellType" name="cell"/> + </sequence> + <attribute type="positiveInteger" name="nat" use="optional"/> + <attribute type="double" name="alat" use="optional"/> + </complexType> + <complexType name="atomic_positionsType"> + <sequence> + <element type="qes_ph:atomType" name="atom" maxOccurs="unbounded" /> + </sequence> + </complexType> + <complexType name="atomType"> + <simpleContent> + <extension base="qes_ph:d3vectorType"> + <attribute type="string" name="name" /> + <attribute type="string" name="position" use="optional" /> + </extension> + </simpleContent> + </complexType> + <complexType name="wyckoff_positionsType"> + <sequence> + <element type="qes_ph:atomType" name="atom"/> + </sequence> + <attribute type="unsignedByte" name="space_group" use="optional"/> + <attribute type="string" name="more_options" use="optional"/> + </complexType> + <complexType name="cellType"> + <sequence> + <element type="qes_ph:d3vectorType" name="a1"/> + <element type="qes_ph:d3vectorType" name="a2"/> + <element type="qes_ph:d3vectorType" name="a3"/> + </sequence> + </complexType> + +  + <simpleType name="d3realDType"> + <restriction> + <simpleType> + <list itemType="double"/> + </simpleType> + <length value="9"/> + </restriction> + </simpleType> + +  + <simpleType name="d3complexDType"> + <restriction> + <simpleType> + <list itemType="double"/> + </simpleType> + <length value="18"/> + </restriction> + </simpleType> + + +  + <simpleType name="d3vectorType"> + <restriction> + <simpleType> + <list itemType="double"/> + </simpleType> + <length value="3"/> + </restriction> + </simpleType> + +</schema>

lda_plus_u = .t. is missing from input Try attached xml, it is expected that lda_plus_u = .t. should be present in the .in [FeO_LDAU_standard.txt](https://github.com/QEF/qeschema/files/6468051/FeO_LDAU_standard.txt)

The presence of tag "dftU" implies that lda_plus_u = .t. Related qexsd commit: https://github.com/QEF/qexsd/commit/a6d365433641f2b41d6155bd49b4628d3adcf51d yes I forgot to add the fix here. Thanks Sasha. This is a blocker to switch to qeschema (everything else seems to be working). Hopefully it is resolved soon.

2021-06-29T21:25:13

0.0

[]

eclipse-zenoh/zenoh-python

eclipse-zenoh__zenoh-python-78

91ab2c2c837e7cfdeadbda8f365feb44921b126c

diff --git a/.github/workflows/release.yml b/.github/workflows/release.yml index 54861ec8..485465b0 100644 --- a/.github/workflows/release.yml +++ b/.github/workflows/release.yml @@ -144,3 +144,21 @@ jobs: with: name: wheels path: dist + + deploy-wheels: + needs: [macos, windows, linux, linux-cross, linux-armv6] + name: deploy wheels to pypi + runs-on: ubuntu-latest + steps: + - uses: actions/download-artifact@v2 + with: + name: wheels + path: dist + - name: Check dist + run: ls -al ./dist/* + - name: publish + if: github.event_name == 'release' && github.event.action == 'released' + uses: pypa/gh-action-pypi-publish@master + with: + user: __token__ + password: ${{ secrets.PYPI_API_TOKEN }} diff --git a/.gitignore b/.gitignore index 0b9c26bf..f01d859d 100644 --- a/.gitignore +++ b/.gitignore @@ -35,3 +35,6 @@ docs/_* *.bak .vscode/ + +**/pycache +**/__pycache__ diff --git a/.gitmodules b/.gitmodules deleted file mode 100644 index e69de29b..00000000 diff --git a/Cargo.lock b/Cargo.lock index f773e926..46edc929 100644 --- a/Cargo.lock +++ b/Cargo.lock @@ -4,39 +4,35 @@ version = 3 [[package]] name = "aes" -version = "0.7.5" +version = "0.8.1" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "9e8b47f52ea9bae42228d07ec09eb676433d7c4ed1ebdf0f1d1c29ed446f1ab8" +checksum = "bfe0133578c0986e1fe3dfcd4af1cc5b2dd6c3dbf534d69916ce16a2701d40ba" dependencies = [ "cfg-if", "cipher", "cpufeatures", - "opaque-debug 0.3.0", ] [[package]] name = "aho-corasick" -version = "0.7.18" +version = "0.7.19" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "1e37cfd5e7657ada45f742d6e99ca5788580b5c529dc78faf11ece6dc702656f" +checksum = "b4f55bd91a0978cbfd91c457a164bab8b4001c833b7f323132c0a4e1922dd44e" dependencies = [ "memchr", ] [[package]] -name = "ansi_term" -version = "0.12.1" +name = "anyhow" +version = "1.0.65" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "d52a9bb7ec0cf484c551830a7ce27bd20d67eac647e1befb56b0be4ee39a55d2" -dependencies = [ - "winapi", -] +checksum = "98161a4e3e2184da77bb14f02184cdd111e83bbbcc9979dfee3c44b9a85f5602" [[package]] -name = "anyhow" -version = "1.0.56" +name = "array-init" +version = "2.0.1" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "4361135be9122e0870de935d7c439aef945b9f9ddd4199a553b5270b49c82a27" +checksum = "bfb6d71005dc22a708c7496eee5c8dc0300ee47355de6256c3b35b12b5fef596" [[package]] name = "async-attributes" @@ -50,9 +46,9 @@ dependencies = [ [[package]] name = "async-channel" -version = "1.6.1" +version = "1.7.1" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "2114d64672151c0c5eaa5e131ec84a74f06e1e559830dabba01ca30605d66319" +checksum = "e14485364214912d3b19cc3435dde4df66065127f05fa0d75c712f36f12c2f28" dependencies = [ "concurrent-queue", "event-listener", @@ -75,27 +71,27 @@ dependencies = [ [[package]] name = "async-global-executor" -version = "2.0.3" +version = "2.3.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "c026b7e44f1316b567ee750fea85103f87fcb80792b860e979f221259796ca0a" +checksum = "0da5b41ee986eed3f524c380e6d64965aea573882a8907682ad100f7859305ca" dependencies = [ "async-channel", "async-executor", "async-io", - "async-mutex", + "async-lock", "blocking", "futures-lite", - "num_cpus", "once_cell", "tokio", ] [[package]] name = "async-io" -version = "1.6.0" +version = "1.9.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "a811e6a479f2439f0c04038796b5cfb3d2ad56c230e0f2d3f7b04d68cfee607b" +checksum = "83e21f3a490c72b3b0cf44962180e60045de2925d8dff97918f7ee43c8f637c7" dependencies = [ + "autocfg", "concurrent-queue", "futures-lite", "libc", @@ -118,22 +114,14 @@ dependencies = [ "event-listener", ] -[[package]] -name = "async-mutex" -version = "1.4.0" -source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "479db852db25d9dbf6204e6cb6253698f175c15726470f78af0d918e99d6156e" -dependencies = [ - "event-listener", -] - [[package]] name = "async-process" -version = "1.3.0" +version = "1.5.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "83137067e3a2a6a06d67168e49e68a0957d215410473a740cea95a2425c0b7c6" +checksum = "02111fd8655a613c25069ea89fc8d9bb89331fa77486eb3bc059ee757cfa481c" dependencies = [ "async-io", + "autocfg", "blocking", "cfg-if", "event-listener", @@ -157,9 +145,9 @@ dependencies = [ [[package]] name = "async-std" -version = "1.11.0" +version = "1.12.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "52580991739c5cdb36cde8b2a516371c0a3b70dda36d916cc08b82372916808c" +checksum = "62565bb4402e926b29953c785397c6dc0391b7b446e45008b0049eb43cec6f5d" dependencies = [ "async-attributes", "async-channel", @@ -176,7 +164,6 @@ dependencies = [ "kv-log-macro", "log", "memchr", - "num_cpus", "once_cell", "pin-project-lite", "pin-utils", @@ -186,15 +173,15 @@ dependencies = [ [[package]] name = "async-task" -version = "4.2.0" +version = "4.3.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "30696a84d817107fc028e049980e09d5e140e8da8f1caeb17e8e950658a3cea9" +checksum = "7a40729d2133846d9ed0ea60a8b9541bccddab49cd30f0715a1da672fe9a2524" [[package]] name = "async-trait" -version = "0.1.52" +version = "0.1.57" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "061a7acccaa286c011ddc30970520b98fa40e00c9d644633fb26b5fc63a265e3" +checksum = "76464446b8bc32758d7e88ee1a804d9914cd9b1cb264c029899680b0be29826f" dependencies = [ "proc-macro2", "quote", @@ -218,15 +205,6 @@ dependencies = [ "winapi", ] -[[package]] -name = "autocfg" -version = "0.1.8" -source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "0dde43e75fd43e8a1bf86103336bc699aa8d17ad1be60c76c0bdfd4828e19b78" -dependencies = [ - "autocfg 1.1.0", -] - [[package]] name = "autocfg" version = "1.1.0" @@ -241,9 +219,9 @@ checksum = "904dfeac50f3cdaba28fc6f57fdcddb75f49ed61346676a78c4ffe55877802fd" [[package]] name = "base64ct" -version = "1.1.1" +version = "1.5.2" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "e6b4d9b1225d28d360ec6a231d65af1fd99a2a095154c8040689617290569c5c" +checksum = "ea2b2456fd614d856680dcd9fcc660a51a820fa09daef2e49772b56a193c8474" [[package]] name = "bincode" @@ -262,41 +240,13 @@ checksum = "bef38d45163c2f1dde094a7dfd33ccf595c92905c8f8f4fdc18d06fb1037718a" [[package]] name = "block-buffer" -version = "0.7.3" -source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "c0940dc441f31689269e10ac70eb1002a3a1d3ad1390e030043662eb7fe4688b" -dependencies = [ - "block-padding 0.1.5", - "byte-tools", - "byteorder", - "generic-array 0.12.4", -] - -[[package]] -name = "block-buffer" -version = "0.9.0" -source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "4152116fd6e9dadb291ae18fc1ec3575ed6d84c29642d97890f4b4a3417297e4" -dependencies = [ - "block-padding 0.2.1", - "generic-array 0.14.5", -] - -[[package]] -name = "block-padding" -version = "0.1.5" +version = "0.10.3" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "fa79dedbb091f449f1f39e53edf88d5dbe95f895dae6135a8d7b881fb5af73f5" +checksum = "69cce20737498f97b993470a6e536b8523f0af7892a4f928cceb1ac5e52ebe7e" dependencies = [ - "byte-tools", + "generic-array", ] -[[package]] -name = "block-padding" -version = "0.2.1" -source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "8d696c370c750c948ada61c69a0ee2cbbb9c50b1019ddb86d9317157a99c2cae" - [[package]] name = "blocking" version = "1.2.0" @@ -313,15 +263,9 @@ dependencies = [ [[package]] name = "bumpalo" -version = "3.9.1" +version = "3.11.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "a4a45a46ab1f2412e53d3a0ade76ffad2025804294569aae387231a0cd6e0899" - -[[package]] -name = "byte-tools" -version = "0.3.1" -source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "e3b5ca7a04898ad4bcd41c90c5285445ff5b791899bb1b0abdd2a2aa791211d7" +checksum = "c1ad822118d20d2c234f427000d5acc36eabe1e29a348c89b63dd60b13f28e5d" [[package]] name = "byteorder" @@ -331,9 +275,9 @@ checksum = "14c189c53d098945499cdfa7ecc63567cf3886b3332b312a5b4585d8d3a6a610" [[package]] name = "bytes" -version = "1.1.0" +version = "1.2.1" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "c4872d67bab6358e59559027aa3b9157c53d9358c51423c17554809a8858e0f8" +checksum = "ec8a7b6a70fde80372154c65702f00a0f56f3e1c36abbc6c440484be248856db" [[package]] name = "cache-padded" @@ -355,42 +299,52 @@ checksum = "baf1de4339761588bc0619e3cbc0120ee582ebb74b53b4efbf79117bd2da40fd" [[package]] name = "cipher" -version = "0.3.0" +version = "0.4.3" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "7ee52072ec15386f770805afd189a01c8841be8696bed250fa2f13c4c0d6dfb7" +checksum = "d1873270f8f7942c191139cb8a40fd228da6c3fd2fc376d7e92d47aa14aeb59e" dependencies = [ - "generic-array 0.14.5", + "crypto-common", + "inout", ] [[package]] name = "clap" -version = "2.34.0" +version = "3.2.22" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "a0610544180c38b88101fecf2dd634b174a62eef6946f84dfc6a7127512b381c" +checksum = "86447ad904c7fb335a790c9d7fe3d0d971dc523b8ccd1561a520de9a85302750" dependencies = [ - "ansi_term", "atty", "bitflags", + "clap_lex", + "indexmap", "strsim", + "termcolor", "textwrap", - "unicode-width", - "vec_map", +] + +[[package]] +name = "clap_lex" +version = "0.2.4" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "2850f2f5a82cbf437dd5af4d49848fbdfc27c157c3d010345776f952765261c5" +dependencies = [ + "os_str_bytes", ] [[package]] name = "concurrent-queue" -version = "1.2.2" +version = "1.2.4" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "30ed07550be01594c6026cff2a1d7fe9c8f683caa798e12b68694ac9e88286a3" +checksum = "af4780a44ab5696ea9e28294517f1fffb421a83a25af521333c838635509db9c" dependencies = [ "cache-padded", ] [[package]] name = "const-oid" -version = "0.6.2" +version = "0.7.1" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "9d6f2aa4d0537bcc1c74df8755072bd31c1ef1a3a1b85a68e8404a8c353b7b8b" +checksum = "e4c78c047431fee22c1a7bb92e00ad095a02a983affe4d8a72e2a2c62c1b94f3" [[package]] name = "core-foundation" @@ -410,49 +364,48 @@ checksum = "5827cebf4670468b8772dd191856768aedcb1b0278a04f989f7766351917b9dc" [[package]] name = "cpufeatures" -version = "0.2.1" +version = "0.2.5" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "95059428f66df56b63431fdb4e1947ed2190586af5c5a8a8b71122bdf5a7f469" +checksum = "28d997bd5e24a5928dd43e46dc529867e207907fe0b239c3477d924f7f2ca320" dependencies = [ "libc", ] [[package]] name = "crossbeam-utils" -version = "0.8.8" +version = "0.8.11" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "0bf124c720b7686e3c2663cf54062ab0f68a88af2fb6a030e87e30bf721fcb38" +checksum = "51887d4adc7b564537b15adcfb307936f8075dfcd5f00dde9a9f1d29383682bc" dependencies = [ "cfg-if", - "lazy_static", + "once_cell", ] [[package]] name = "crypto-bigint" -version = "0.2.11" +version = "0.3.2" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "f83bd3bb4314701c568e340cd8cf78c975aa0ca79e03d3f6d1677d5b0c9c0c03" +checksum = "03c6a1d5fa1de37e071642dfa44ec552ca5b299adb128fab16138e24b548fd21" dependencies = [ - "generic-array 0.14.5", - "rand_core", + "generic-array", "subtle", ] [[package]] -name = "crypto-mac" -version = "0.11.1" +name = "crypto-common" +version = "0.1.6" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "b1d1a86f49236c215f271d40892d5fc950490551400b02ef360692c29815c714" +checksum = "1bfb12502f3fc46cca1bb51ac28df9d618d813cdc3d2f25b9fe775a34af26bb3" dependencies = [ - "generic-array 0.14.5", - "subtle", + "generic-array", + "typenum", ] [[package]] name = "ctor" -version = "0.1.22" +version = "0.1.23" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "f877be4f7c9f246b183111634f75baa039715e3f46ce860677d3b19a69fb229c" +checksum = "cdffe87e1d521a10f9696f833fe502293ea446d7f256c06128293a4119bdf4cb" dependencies = [ "quote", "syn", @@ -460,47 +413,40 @@ dependencies = [ [[package]] name = "der" -version = "0.4.5" +version = "0.5.1" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "79b71cca7d95d7681a4b3b9cdf63c8dbc3730d0584c2c74e31416d64a90493f4" +checksum = "6919815d73839e7ad218de758883aae3a257ba6759ce7a9992501efbb53d705c" dependencies = [ "const-oid", "crypto-bigint", + "pem-rfc7468", ] [[package]] name = "digest" -version = "0.8.1" -source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "f3d0c8c8752312f9713efd397ff63acb9f85585afbf179282e720e7704954dd5" -dependencies = [ - "generic-array 0.12.4", -] - -[[package]] -name = "digest" -version = "0.9.0" +version = "0.10.5" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "d3dd60d1080a57a05ab032377049e0591415d2b31afd7028356dbf3cc6dcb066" +checksum = "adfbc57365a37acbd2ebf2b64d7e69bb766e2fea813521ed536f5d0520dcf86c" dependencies = [ - "generic-array 0.14.5", + "block-buffer", + "crypto-common", + "subtle", ] [[package]] -name = "dirs-next" -version = "2.0.0" +name = "dirs" +version = "4.0.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "b98cf8ebf19c3d1b223e151f99a4f9f0690dca41414773390fc824184ac833e1" +checksum = "ca3aa72a6f96ea37bbc5aa912f6788242832f75369bdfdadcb0e38423f100059" dependencies = [ - "cfg-if", - "dirs-sys-next", + "dirs-sys", ] [[package]] -name = "dirs-sys-next" -version = "0.1.2" +name = "dirs-sys" +version = "0.3.7" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "4ebda144c4fe02d1f7ea1a7d9641b6fc6b580adcfa024ae48797ecdeb6825b4d" +checksum = "1b1d1d91c932ef41c0f2663aa8b0ca0342d444d842c06914aa0a7e352d0bada6" dependencies = [ "libc", "redox_users", @@ -513,25 +459,31 @@ version = "0.6.0-dev" dependencies = [ "async-std", "env_logger", + "form_urlencoded", "futures", "json5", "log", "pyo3", - "pyo3-asyncio", "serde_json", "uhlc", - "validated_struct 0.1.11", + "validated_struct", "zenoh", "zenoh-buffers", "zenoh-cfg-properties", "zenoh-core", ] +[[package]] +name = "either" +version = "1.8.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "90e5c1c8368803113bf0c9584fc495a58b86dc8a29edbf8fe877d21d9507e797" + [[package]] name = "env_logger" -version = "0.9.0" +version = "0.9.1" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "0b2cf0344971ee6c64c31be0d530793fba457d322dfec2810c453d0ef228f9c3" +checksum = "c90bf5f19754d10198ccb95b70664fc925bd1fc090a0fd9a6ebc54acc8cd6272" dependencies = [ "atty", "humantime", @@ -542,49 +494,52 @@ dependencies = [ [[package]] name = "event-listener" -version = "2.5.2" +version = "2.5.3" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "77f3309417938f28bf8228fcff79a4a37103981e3e186d2ccd19c74b38f4eb71" - -[[package]] -name = "fake-simd" -version = "0.1.2" -source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "e88a8acf291dafb59c2d96e8f59828f3838bb1a70398823ade51a84de6a6deed" +checksum = "0206175f82b8d6bf6652ff7d71a1e27fd2e4efde587fd368662814d6ec1d9ce0" [[package]] name = "fastrand" -version = "1.7.0" +version = "1.8.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "c3fcf0cee53519c866c09b5de1f6c56ff9d647101f81c1964fa632e148896cdf" +checksum = "a7a407cfaa3385c4ae6b23e84623d48c2798d06e3e6a1878f7f59f17b3f86499" dependencies = [ "instant", ] [[package]] name = "fixedbitset" -version = "0.4.1" +version = "0.4.2" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "279fb028e20b3c4c320317955b77c5e0c9701f05a1d309905d6fc702cdc5053e" +checksum = "0ce7134b9999ecaf8bcd65542e436736ef32ddca1b3e06094cb6ec5755203b80" [[package]] name = "flume" -version = "0.10.12" +version = "0.10.14" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "843c03199d0c0ca54bc1ea90ac0d507274c28abcc4f691ae8b4eaa375087c76a" +checksum = "1657b4441c3403d9f7b3409e47575237dac27b1b5726df654a6ecbf92f0f7577" dependencies = [ "futures-core", "futures-sink", "nanorand", "pin-project", - "spin 0.9.2", + "spin 0.9.4", +] + +[[package]] +name = "form_urlencoded" +version = "1.1.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "a9c384f161156f5260c24a097c56119f9be8c798586aecc13afbcbe7b7e26bf8" +dependencies = [ + "percent-encoding", ] [[package]] name = "futures" -version = "0.3.21" +version = "0.3.24" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "f73fe65f54d1e12b726f517d3e2135ca3125a437b6d998caf1962961f7172d9e" +checksum = "7f21eda599937fba36daeb58a22e8f5cee2d14c4a17b5b7739c7c8e5e3b8230c" dependencies = [ "futures-channel", "futures-core", @@ -597,9 +552,9 @@ dependencies = [ [[package]] name = "futures-channel" -version = "0.3.21" +version = "0.3.24" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "c3083ce4b914124575708913bca19bfe887522d6e2e6d0952943f5eac4a74010" +checksum = "30bdd20c28fadd505d0fd6712cdfcb0d4b5648baf45faef7f852afb2399bb050" dependencies = [ "futures-core", "futures-sink", @@ -607,15 +562,15 @@ dependencies = [ [[package]] name = "futures-core" -version = "0.3.21" +version = "0.3.24" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "0c09fd04b7e4073ac7156a9539b57a484a8ea920f79c7c675d05d289ab6110d3" +checksum = "4e5aa3de05362c3fb88de6531e6296e85cde7739cccad4b9dfeeb7f6ebce56bf" [[package]] name = "futures-executor" -version = "0.3.21" +version = "0.3.24" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "9420b90cfa29e327d0429f19be13e7ddb68fa1cccb09d65e5706b8c7a749b8a6" +checksum = "9ff63c23854bee61b6e9cd331d523909f238fc7636290b96826e9cfa5faa00ab" dependencies = [ "futures-core", "futures-task", @@ -624,9 +579,9 @@ dependencies = [ [[package]] name = "futures-io" -version = "0.3.21" +version = "0.3.24" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "fc4045962a5a5e935ee2fdedaa4e08284547402885ab326734432bed5d12966b" +checksum = "bbf4d2a7a308fd4578637c0b17c7e1c7ba127b8f6ba00b29f717e9655d85eb68" [[package]] name = "futures-lite" @@ -645,9 +600,9 @@ dependencies = [ [[package]] name = "futures-macro" -version = "0.3.21" +version = "0.3.24" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "33c1e13800337f4d4d7a316bf45a567dbcb6ffe087f16424852d97e97a91f512" +checksum = "42cd15d1c7456c04dbdf7e88bcd69760d74f3a798d6444e16974b505b0e62f17" dependencies = [ "proc-macro2", "quote", @@ -656,21 +611,21 @@ dependencies = [ [[package]] name = "futures-sink" -version = "0.3.21" +version = "0.3.24" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "21163e139fa306126e6eedaf49ecdb4588f939600f0b1e770f4205ee4b7fa868" +checksum = "21b20ba5a92e727ba30e72834706623d94ac93a725410b6a6b6fbc1b07f7ba56" [[package]] name = "futures-task" -version = "0.3.21" +version = "0.3.24" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "57c66a976bf5909d801bbef33416c41372779507e7a6b3a5e25e4749c58f776a" +checksum = "a6508c467c73851293f390476d4491cf4d227dbabcd4170f3bb6044959b294f1" [[package]] name = "futures-util" -version = "0.3.21" +version = "0.3.24" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "d8b7abd5d659d9b90c8cba917f6ec750a74e2dc23902ef9cd4cc8c8b22e6036a" +checksum = "44fb6cb1be61cc1d2e43b262516aafcf63b241cffdb1d3fa115f91d9c7b09c90" dependencies = [ "futures-channel", "futures-core", @@ -695,18 +650,9 @@ dependencies = [ [[package]] name = "generic-array" -version = "0.12.4" +version = "0.14.6" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "ffdf9f34f1447443d37393cc6c2b8313aebddcd96906caf34e54c68d8e57d7bd" -dependencies = [ - "typenum", -] - -[[package]] -name = "generic-array" -version = "0.14.5" -source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "fd48d33ec7f05fbfa152300fdad764757cbded343c1aa1cff2fbaf4134851803" +checksum = "bff49e947297f3312447abdca79f45f4738097cc82b06e72054d2223f601f1b9" dependencies = [ "typenum", "version_check", @@ -714,28 +660,17 @@ dependencies = [ [[package]] name = "getrandom" -version = "0.2.5" +version = "0.2.7" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "d39cd93900197114fa1fcb7ae84ca742095eed9442088988ae74fa744e930e77" +checksum = "4eb1a864a501629691edf6c15a593b7a51eebaa1e8468e9ddc623de7c9b58ec6" dependencies = [ "cfg-if", "js-sys", "libc", - "wasi 0.10.2+wasi-snapshot-preview1", + "wasi", "wasm-bindgen", ] -[[package]] -name = "ghost" -version = "0.1.2" -source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "1a5bcf1bbeab73aa4cf2fde60a846858dc036163c7c33bec309f8d17de785479" -dependencies = [ - "proc-macro2", - "quote", - "syn", -] - [[package]] name = "git-version" version = "0.3.5" @@ -766,9 +701,9 @@ checksum = "9b919933a397b79c37e33b77bb2aa3dc8eb6e165ad809e58ff75bc7db2e34574" [[package]] name = "gloo-timers" -version = "0.2.3" +version = "0.2.4" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "4d12a7f4e95cfe710f1d624fb1210b7d961a5fb05c4fd942f4feab06e61f590e" +checksum = "5fb7d06c1c8cc2a29bee7ec961009a0b2caa0793ee4900c2ffb348734ba1c8f9" dependencies = [ "futures-channel", "futures-core", @@ -778,9 +713,9 @@ dependencies = [ [[package]] name = "hashbrown" -version = "0.11.2" +version = "0.12.3" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "ab5ef0d4909ef3724cc8cce6ccc8572c5c817592e9285f5464f8e86f8bd3726e" +checksum = "8a9ee70c43aaf417c914396645a0fa852624801b24ebb7ae78fe8272889ac888" [[package]] name = "hermit-abi" @@ -799,12 +734,11 @@ checksum = "7f24254aa9a54b5c858eaee2f5bccdb46aaf0e486a595ed5fd8f86ba55232a70" [[package]] name = "hmac" -version = "0.11.0" +version = "0.12.1" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "2a2a2320eb7ec0ebe8da8f744d7812d9fc4cb4d09344ac01898dbcb6a20ae69b" +checksum = "6c49c37c09c17a53d937dfbb742eb3a961d65a994e6bcdcf37e7399d0cc8ab5e" dependencies = [ - "crypto-mac", - "digest 0.9.0", + "digest", ] [[package]] @@ -824,35 +758,27 @@ checksum = "9a3a5bfb195931eeb336b2a7b4d761daec841b97f947d34394601737a7bba5e4" [[package]] name = "indexmap" -version = "1.8.0" +version = "1.9.1" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "282a6247722caba404c065016bbfa522806e51714c34f5dfc3e4a3a46fcb4223" +checksum = "10a35a97730320ffe8e2d410b5d3b69279b98d2c14bdb8b70ea89ecf7888d41e" dependencies = [ - "autocfg 1.1.0", + "autocfg", "hashbrown", ] [[package]] name = "indoc" -version = "0.3.6" +version = "1.0.7" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "47741a8bc60fb26eb8d6e0238bbb26d8575ff623fdc97b1a2c00c050b9684ed8" -dependencies = [ - "indoc-impl", - "proc-macro-hack", -] +checksum = "adab1eaa3408fb7f0c777a73e7465fd5656136fc93b670eb6df3c88c2c1344e3" [[package]] -name = "indoc-impl" -version = "0.3.6" +name = "inout" +version = "0.1.3" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "ce046d161f000fffde5f432a0d034d0341dc152643b2598ed5bfce44c4f3a8f0" +checksum = "a0c10553d664a4d0bcff9f4215d0aac67a639cc68ef660840afe309b807bc9f5" dependencies = [ - "proc-macro-hack", - "proc-macro2", - "quote", - "syn", - "unindent", + "generic-array", ] [[package]] @@ -865,47 +791,34 @@ dependencies = [ ] [[package]] -name = "inventory" -version = "0.1.11" -source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "f0eb5160c60ba1e809707918ee329adb99d222888155835c6feedba19f6c3fd4" -dependencies = [ - "ctor", - "ghost", - "inventory-impl", -] - -[[package]] -name = "inventory-impl" -version = "0.1.11" +name = "ipnetwork" +version = "0.19.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "7e41b53715c6f0c4be49510bb82dee2c1e51c8586d885abe65396e82ed518548" +checksum = "1f84f1612606f3753f205a4e9a2efd6fe5b4c573a6269b2cc6c3003d44a0d127" dependencies = [ - "proc-macro2", - "quote", - "syn", + "serde", ] [[package]] -name = "ipnetwork" -version = "0.18.0" +name = "itertools" +version = "0.10.5" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "4088d739b183546b239688ddbc79891831df421773df95e236daf7867866d355" +checksum = "b0fd2260e829bddf4cb6ea802289de2f86d6a7a690192fbe91b3f46e0f2c8473" dependencies = [ - "serde", + "either", ] [[package]] name = "itoa" -version = "1.0.1" +version = "1.0.3" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "1aab8fc367588b89dcee83ab0fd66b72b50b72fa1904d7095045ace2b0c81c35" +checksum = "6c8af84674fe1f223a982c933a0ee1086ac4d4052aa0fb8060c12c6ad838e754" [[package]] name = "js-sys" -version = "0.3.56" +version = "0.3.60" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "a38fc24e30fd564ce974c02bf1d337caddff65be6cc4735a1f7eab22a7440f04" +checksum = "49409df3e3bf0856b916e2ceaca09ee28e6871cf7d9ce97a692cacfdb2a25a47" dependencies = [ "wasm-bindgen", ] @@ -923,9 +836,9 @@ dependencies = [ [[package]] name = "keccak" -version = "0.1.0" +version = "0.1.2" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "67c21572b4949434e4fc1e1978b99c5f77064153c59d998bf13ecd96fb5ecba7" +checksum = "f9b7d56ba4a8344d6be9729995e6b06f928af29998cdf79fe390cbf6b1fee838" [[package]] name = "kv-log-macro" @@ -947,9 +860,9 @@ dependencies = [ [[package]] name = "libc" -version = "0.2.120" +version = "0.2.133" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "ad5c14e80759d0939d013e6ca49930e59fc53dd8e5009132f76240c179380c09" +checksum = "c0f80d65747a3e43d1596c7c5492d95d5edddaabd45a7fcdb02b95f644164966" [[package]] name = "libloading" @@ -963,46 +876,35 @@ dependencies = [ [[package]] name = "libm" -version = "0.2.2" -source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "33a33a362ce288760ec6a508b94caaec573ae7d3bbbd91b87aa0bad4456839db" - -[[package]] -name = "linked-hash-map" -version = "0.5.4" +version = "0.2.5" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "7fb9b38af92608140b86b693604b9ffcc5824240a484d1ecd4795bacb2fe88f3" +checksum = "292a948cd991e376cf75541fe5b97a1081d713c618b4f1b9500f8844e49eb565" [[package]] name = "lock_api" -version = "0.4.6" +version = "0.4.9" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "88943dd7ef4a2e5a4bfa2753aaab3013e34ce2533d1996fb18ef591e315e2b3b" +checksum = "435011366fe56583b16cf956f9df0095b405b82d76425bc8981c0e22e60ec4df" dependencies = [ + "autocfg", "scopeguard", ] [[package]] name = "log" -version = "0.4.14" +version = "0.4.17" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "51b9bbe6c47d51fc3e1a9b945965946b4c44142ab8792c50835a980d362c2710" +checksum = "abb12e687cfb44aa40f41fc3978ef76448f9b6038cad6aef4259d3c095a2382e" dependencies = [ "cfg-if", "value-bag", ] -[[package]] -name = "maplit" -version = "1.0.2" -source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "3e2e65a1a2e43cfcb47a895c4c8b10d1f4a61097f9f254f183aee60cad9c651d" - [[package]] name = "memchr" -version = "2.4.1" +version = "2.5.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "308cc39be01b73d0d18f82a0e7b2a3df85245f84af96fdddc5d202d27e47b86a" +checksum = "2dffe52ecf27772e601905b7522cb4ef790d2cc203488bbd0e2fe85fcb74566d" [[package]] name = "memoffset" @@ -1010,43 +912,19 @@ version = "0.6.5" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "5aa361d4faea93603064a027415f07bd8e1d5c88c9fbf68bf56a285428fd79ce" dependencies = [ - "autocfg 1.1.0", -] - -[[package]] -name = "mio" -version = "0.7.14" -source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "8067b404fe97c70829f082dec8bcf4f71225d7eaea1d8645349cb76fa06205cc" -dependencies = [ - "libc", - "log", - "miow", - "ntapi", - "winapi", + "autocfg", ] [[package]] name = "mio" -version = "0.8.2" +version = "0.8.4" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "52da4364ffb0e4fe33a9841a98a3f3014fb964045ce4f7a45a398243c8d6b0c9" +checksum = "57ee1c23c7c63b0c9250c339ffdc69255f110b298b901b9f6c82547b7b87caaf" dependencies = [ "libc", "log", - "miow", - "ntapi", - "wasi 0.11.0+wasi-snapshot-preview1", - "winapi", -] - -[[package]] -name = "miow" -version = "0.3.7" -source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "b9f1c5b025cda876f66ef43a113f91ebc9f4ccef34843000e0adf6ebbab84e21" -dependencies = [ - "winapi", + "wasi", + "windows-sys", ] [[package]] @@ -1060,9 +938,9 @@ dependencies = [ [[package]] name = "nix" -version = "0.22.3" +version = "0.23.1" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "e4916f159ed8e5de0082076562152a76b7a1f64a01fd9d1e0fea002c37624faf" +checksum = "9f866317acbd3a240710c63f065ffb1e4fd466259045ccb504130b7f668f35c6" dependencies = [ "bitflags", "cc", @@ -1073,33 +951,30 @@ dependencies = [ [[package]] name = "nix" -version = "0.23.1" +version = "0.25.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "9f866317acbd3a240710c63f065ffb1e4fd466259045ccb504130b7f668f35c6" +checksum = "e322c04a9e3440c327fca7b6c8a63e6890a32fa2ad689db972425f07e0d22abb" dependencies = [ + "autocfg", "bitflags", - "cc", "cfg-if", "libc", "memoffset", + "pin-utils", ] [[package]] -name = "ntapi" -version = "0.3.7" +name = "no-std-net" +version = "0.6.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "c28774a7fd2fbb4f0babd8237ce554b73af68021b5f695a3cebd6c59bac0980f" -dependencies = [ - "winapi", -] +checksum = "43794a0ace135be66a25d3ae77d41b91615fb68ae937f904090203e81f755b65" [[package]] name = "num-bigint-dig" -version = "0.7.0" +version = "0.8.1" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "4547ee5541c18742396ae2c895d0717d0f886d8823b8399cdaf7b07d63ad0480" +checksum = "566d173b2f9406afbc5510a90925d5a2cd80cae4605631f1212303df265de011" dependencies = [ - "autocfg 0.1.8", "byteorder", "lazy_static", "libm", @@ -1113,32 +988,32 @@ dependencies = [ [[package]] name = "num-integer" -version = "0.1.44" +version = "0.1.45" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "d2cc698a63b549a70bc047073d2949cce27cd1c7b0a4a862d08a8031bc2801db" +checksum = "225d3389fb3509a24c93f5c29eb6bde2586b98d9f016636dff58d7c6f7569cd9" dependencies = [ - "autocfg 1.1.0", + "autocfg", "num-traits", ] [[package]] name = "num-iter" -version = "0.1.42" +version = "0.1.43" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "b2021c8337a54d21aca0d59a92577a029af9431cb59b909b03252b9c164fad59" +checksum = "7d03e6c028c5dc5cac6e2dec0efda81fc887605bb3d884578bb6d6bf7514e252" dependencies = [ - "autocfg 1.1.0", + "autocfg", "num-integer", "num-traits", ] [[package]] name = "num-traits" -version = "0.2.14" +version = "0.2.15" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "9a64b1ec5cda2586e284722486d802acf1f7dbdc623e2bfc57e65ca1cd099290" +checksum = "578ede34cf02f8924ab9447f50c28075b4d3e5b269972345e7e0372b38c6cdcd" dependencies = [ - "autocfg 1.1.0", + "autocfg", "libm", ] @@ -1154,21 +1029,9 @@ dependencies = [ [[package]] name = "once_cell" -version = "1.10.0" -source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "87f3e037eac156d1775da914196f0f37741a274155e34a0b7e427c35d2a2ecb9" - -[[package]] -name = "opaque-debug" -version = "0.2.3" -source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "2839e79665f131bdb5782e51f2c6c9599c133c6098982a54c794358bf432529c" - -[[package]] -name = "opaque-debug" -version = "0.3.0" +version = "1.15.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "624a8340c38c1b80fd549087862da4ba43e08858af025b236e509b6649fc13d5" +checksum = "e82dad04139b71a90c080c8463fe0dc7902db5192d939bd0950f074d014339e1" [[package]] name = "openssl-probe" @@ -1178,13 +1041,19 @@ checksum = "ff011a302c396a5197692431fc1948019154afc178baf7d8e37367442a4601cf" [[package]] name = "ordered-float" -version = "2.10.0" +version = "3.1.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "7940cf2ca942593318d07fcf2596cdca60a85c9e7fab408a5e21a4f9dcd40d87" +checksum = "98ffdb14730ed2ef599c65810c15b000896e21e8776b512de0db0c3d7335cc2a" dependencies = [ "num-traits", ] +[[package]] +name = "os_str_bytes" +version = "6.3.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "9ff7415e9ae3fff1225851df9e0d9e4e5479f947619774677a63572e55e80eff" + [[package]] name = "parking" version = "2.0.0" @@ -1193,77 +1062,63 @@ checksum = "427c3892f9e783d91cc128285287e70a59e206ca452770ece88a76f7a3eddd72" [[package]] name = "parking_lot" -version = "0.11.2" +version = "0.12.1" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "7d17b78036a60663b797adeaee46f5c9dfebb86948d1255007a1d6be0271ff99" +checksum = "3742b2c103b9f06bc9fff0a37ff4912935851bee6d36f3c02bcc755bcfec228f" dependencies = [ - "instant", "lock_api", "parking_lot_core", ] [[package]] name = "parking_lot_core" -version = "0.8.5" +version = "0.9.3" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "d76e8e1493bcac0d2766c42737f34458f1c8c50c0d23bcb24ea953affb273216" +checksum = "09a279cbf25cb0757810394fbc1e359949b59e348145c643a939a525692e6929" dependencies = [ "cfg-if", - "instant", "libc", "redox_syscall", "smallvec", - "winapi", + "windows-sys", ] [[package]] name = "paste" -version = "0.1.18" -source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "45ca20c77d80be666aef2b45486da86238fabe33e38306bd3118fe4af33fa880" -dependencies = [ - "paste-impl", - "proc-macro-hack", -] - -[[package]] -name = "paste" -version = "1.0.6" +version = "1.0.9" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "0744126afe1a6dd7f394cb50a716dbe086cb06e255e53d8d0185d82828358fb5" +checksum = "b1de2e551fb905ac83f73f7aedf2f0cb4a0da7e35efa24a202a936269f1f18e1" [[package]] -name = "paste-impl" -version = "0.1.18" +name = "pem-rfc7468" +version = "0.3.1" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "d95a7db200b97ef370c8e6de0088252f7e0dfff7d047a28528e47456c0fc98b6" +checksum = "01de5d978f34aa4b2296576379fcc416034702fd94117c56ffd8a1a767cefb30" dependencies = [ - "proc-macro-hack", + "base64ct", ] [[package]] -name = "pem-rfc7468" -version = "0.2.4" +name = "percent-encoding" +version = "2.2.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "84e93a3b1cc0510b03020f33f21e62acdde3dcaef432edc95bea377fbd4c2cd4" -dependencies = [ - "base64ct", -] +checksum = "478c572c3d73181ff3c2539045f6eb99e5491218eae919370993b890cdbdd98e" [[package]] name = "pest" -version = "2.1.3" +version = "2.3.1" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "10f4872ae94d7b90ae48754df22fd42ad52ce740b8f370b03da4835417403e53" +checksum = "cb779fcf4bb850fbbb0edc96ff6cf34fd90c4b1a112ce042653280d9a7364048" dependencies = [ + "thiserror", "ucd-trie", ] [[package]] name = "pest_derive" -version = "2.1.0" +version = "2.3.1" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "833d1ae558dc601e9a60366421196a8d94bc0ac980476d0b67e1d0988d72b2d0" +checksum = "502b62a6d0245378b04ffe0a7fb4f4419a4815fce813bd8a0ec89a56e07d67b1" dependencies = [ "pest", "pest_generator", @@ -1271,9 +1126,9 @@ dependencies = [ [[package]] name = "pest_generator" -version = "2.1.3" +version = "2.3.1" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "99b8db626e31e5b81787b9783425769681b347011cc59471e33ea46d2ea0cf55" +checksum = "451e629bf49b750254da26132f1a5a9d11fd8a95a3df51d15c4abd1ba154cb6c" dependencies = [ "pest", "pest_meta", @@ -1284,20 +1139,20 @@ dependencies = [ [[package]] name = "pest_meta" -version = "2.1.3" +version = "2.3.1" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "54be6e404f5317079812fc8f9f5279de376d8856929e21c184ecf6bbd692a11d" +checksum = "bcec162c71c45e269dfc3fc2916eaeb97feab22993a21bcce4721d08cd7801a6" dependencies = [ - "maplit", + "once_cell", "pest", - "sha-1", + "sha1", ] [[package]] name = "petgraph" -version = "0.6.0" +version = "0.6.2" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "4a13a2fa9d0b63e5f22328828741e523766fff0ee9e779316902290dff3f824f" +checksum = "e6d5014253a1331579ce62aa67443b4a658c5e7dd03d4bc6d302b94474888143" dependencies = [ "fixedbitset", "indexmap", @@ -1305,18 +1160,18 @@ dependencies = [ [[package]] name = "pin-project" -version = "1.0.10" +version = "1.0.12" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "58ad3879ad3baf4e44784bc6a718a8698867bb991f8ce24d1bcbe2cfb4c3a75e" +checksum = "ad29a609b6bcd67fee905812e544992d216af9d755757c05ed2d0e15a74c6ecc" dependencies = [ "pin-project-internal", ] [[package]] name = "pin-project-internal" -version = "1.0.10" +version = "1.0.12" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "744b6f092ba29c3650faf274db506afd39944f48420f6c86b17cfe0ee1cb36bb" +checksum = "069bdb1e05adc7a8990dce9cc75370895fbe4e3d58b9b73bf1aee56359344a55" dependencies = [ "proc-macro2", "quote", @@ -1325,9 +1180,9 @@ dependencies = [ [[package]] name = "pin-project-lite" -version = "0.2.8" +version = "0.2.9" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "e280fbe77cc62c91527259e9442153f4688736748d24660126286329742b4c6c" +checksum = "e0a7ae3ac2f1173085d398531c705756c94a4c56843785df85a60c1a0afac116" [[package]] name = "pin-utils" @@ -1337,53 +1192,52 @@ checksum = "8b870d8c151b6f2fb93e84a13146138f05d02ed11c7e7c54f8826aaaf7c9f184" [[package]] name = "pkcs1" -version = "0.2.4" +version = "0.3.3" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "116bee8279d783c0cf370efa1a94632f2108e5ef0bb32df31f051647810a4e2c" +checksum = "a78f66c04ccc83dd4486fd46c33896f4e17b24a7a3a6400dedc48ed0ddd72320" dependencies = [ "der", - "pem-rfc7468", + "pkcs8", "zeroize", ] [[package]] name = "pkcs8" -version = "0.7.6" +version = "0.8.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "ee3ef9b64d26bad0536099c816c6734379e45bbd5f14798def6809e5cc350447" +checksum = "7cabda3fb821068a9a4fab19a683eac3af12edf0f34b94a8be53c4972b8149d0" dependencies = [ "der", - "pem-rfc7468", - "pkcs1", "spki", "zeroize", ] [[package]] name = "pnet" -version = "0.28.0" +version = "0.31.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "4b6d2a0409666964722368ef5fb74b9f93fac11c18bef3308693c16c6733f103" +checksum = "0caaf5b11fd907ff15cf14a4477bfabca4b37ab9e447a4f8dead969a59cdafad" dependencies = [ - "ipnetwork", "pnet_base", "pnet_datalink", "pnet_packet", - "pnet_sys", "pnet_transport", ] [[package]] name = "pnet_base" -version = "0.28.0" +version = "0.31.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "25488cd551a753dcaaa6fffc9f69a7610a412dd8954425bf7ffad5f7d1156fb8" +checksum = "f9d3a993d49e5fd5d4d854d6999d4addca1f72d86c65adf224a36757161c02b6" +dependencies = [ + "no-std-net", +] [[package]] name = "pnet_datalink" -version = "0.28.0" +version = "0.31.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "d4d1f8ab1ef6c914cf51dc5dfe0be64088ea5f3b08bbf5a31abc70356d271198" +checksum = "e466faf03a98ad27f6e15cd27a2b7cc89e73e640a43527742977bc503c37f8aa" dependencies = [ "ipnetwork", "libc", @@ -1394,9 +1248,9 @@ dependencies = [ [[package]] name = "pnet_macros" -version = "0.28.0" +version = "0.31.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "30490e0852e58402b8fae0d39897b08a24f493023a4d6cf56b2e30f31ed57548" +checksum = "48dd52a5211fac27e7acb14cfc9f30ae16ae0e956b7b779c8214c74559cef4c3" dependencies = [ "proc-macro2", "quote", @@ -1406,18 +1260,18 @@ dependencies = [ [[package]] name = "pnet_macros_support" -version = "0.28.0" +version = "0.31.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "d4714e10f30cab023005adce048f2d30dd4ac4f093662abf2220855655ef8f90" +checksum = "89de095dc7739349559913aed1ef6a11e73ceade4897dadc77c5e09de6740750" dependencies = [ "pnet_base", ] [[package]] name = "pnet_packet" -version = "0.28.0" +version = "0.31.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "8588067671d03c9f4254b2e66fecb4d8b93b5d3e703195b84f311cd137e32130" +checksum = "bc3b5111e697c39c8b9795b9fdccbc301ab696699e88b9ea5a4e4628978f495f" dependencies = [ "glob", "pnet_base", @@ -1427,9 +1281,9 @@ dependencies = [ [[package]] name = "pnet_sys" -version = "0.28.0" +version = "0.31.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "d9a3f32b0df45515befd19eed04616f6b56a488da92afc61164ef455e955f07f" +checksum = "328e231f0add6d247d82421bf3790b4b33b39c8930637f428eef24c4c6a90805" dependencies = [ "libc", "winapi", @@ -1437,9 +1291,9 @@ dependencies = [ [[package]] name = "pnet_transport" -version = "0.28.0" +version = "0.31.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "932b2916d693bcc5fa18443dc99142e0a6fd31a6ce75a511868f7174c17e2bce" +checksum = "ff597185e6f1f5671b3122e4dba892a1c73e17c17e723d7669bd9299cbe7f124" dependencies = [ "libc", "pnet_base", @@ -1449,10 +1303,11 @@ dependencies = [ [[package]] name = "polling" -version = "2.2.0" +version = "2.3.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "685404d509889fade3e86fe3a5803bca2ec09b0c0778d5ada6ec8bf7a8de5259" +checksum = "899b00b9c8ab553c743b3e11e87c5c7d423b2a2de229ba95b24a756344748011" dependencies = [ + "autocfg", "cfg-if", "libc", "log", @@ -1474,70 +1329,57 @@ checksum = "dbf0c48bc1d91375ae5c3cd81e3722dff1abcf81a30960240640d223f59fe0e5" [[package]] name = "proc-macro2" -version = "1.0.36" +version = "1.0.43" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "c7342d5883fbccae1cc37a2353b09c87c9b0f3afd73f5fb9bba687a1f733b029" +checksum = "0a2ca2c61bc9f3d74d2886294ab7b9853abd9c1ad903a3ac7815c58989bb7bab" dependencies = [ - "unicode-xid", + "unicode-ident", ] [[package]] name = "pyo3" -version = "0.15.1" +version = "0.17.1" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "7cf01dbf1c05af0a14c7779ed6f3aa9deac9c3419606ac9de537a2d649005720" +checksum = "12f72538a0230791398a0986a6518ebd88abc3fded89007b506ed072acc831e1" dependencies = [ "cfg-if", "indoc", "libc", + "memoffset", "parking_lot", - "paste 0.1.18", "pyo3-build-config", + "pyo3-ffi", "pyo3-macros", "unindent", ] [[package]] -name = "pyo3-asyncio" -version = "0.15.0" +name = "pyo3-build-config" +version = "0.17.1" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "0897c7e36110a32b726b975359b2bbe90c37fcf1266046d3b1c08c616a47a886" +checksum = "fc4cf18c20f4f09995f3554e6bcf9b09bd5e4d6b67c562fdfaafa644526ba479" dependencies = [ - "async-std", - "futures", - "inventory", "once_cell", - "pin-project-lite", - "pyo3", - "pyo3-asyncio-macros", -] - -[[package]] -name = "pyo3-asyncio-macros" -version = "0.15.0" -source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "4daada7260db6d6dbc1f9b50f3e3d21f9eb28c034722c9a61ac05ff56ffd62db" -dependencies = [ - "proc-macro2", - "quote", - "syn", + "target-lexicon", ] [[package]] -name = "pyo3-build-config" -version = "0.15.1" +name = "pyo3-ffi" +version = "0.17.1" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "dbf9e4d128bfbddc898ad3409900080d8d5095c379632fbbfbb9c8cfb1fb852b" +checksum = "a41877f28d8ebd600b6aa21a17b40c3b0fc4dfe73a27b6e81ab3d895e401b0e9" dependencies = [ - "once_cell", + "libc", + "pyo3-build-config", ] [[package]] name = "pyo3-macros" -version = "0.15.1" +version = "0.17.1" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "67701eb32b1f9a9722b4bc54b548ff9d7ebfded011c12daece7b9063be1fd755" +checksum = "2e81c8d4bcc2f216dc1b665412df35e46d12ee8d3d046b381aad05f1fcf30547" dependencies = [ + "proc-macro2", "pyo3-macros-backend", "quote", "syn", @@ -1545,21 +1387,20 @@ dependencies = [ [[package]] name = "pyo3-macros-backend" -version = "0.15.1" +version = "0.17.1" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "f44f09e825ee49a105f2c7b23ebee50886a9aee0746f4dd5a704138a64b0218a" +checksum = "85752a767ee19399a78272cc2ab625cd7d373b2e112b4b13db28de71fa892784" dependencies = [ "proc-macro2", - "pyo3-build-config", "quote", "syn", ] [[package]] name = "quinn" -version = "0.8.1" +version = "0.8.5" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "584865613896a1f644d757e52c45c573441c8b04cac38ac13990b0235203db66" +checksum = "5b435e71d9bfa0d8889927231970c51fb89c58fa63bffcab117c9c7a41e5ef8f" dependencies = [ "bytes", "futures-channel", @@ -1567,7 +1408,7 @@ dependencies = [ "fxhash", "quinn-proto", "quinn-udp", - "rustls 0.20.4", + "rustls 0.20.6", "thiserror", "tokio", "tracing", @@ -1576,15 +1417,15 @@ dependencies = [ [[package]] name = "quinn-proto" -version = "0.8.1" +version = "0.8.4" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "d2b1562bf4998b0c6d1841a4742b7103bb82cdde61374833de826bab9e8ad498" +checksum = "3fce546b9688f767a57530652488420d419a8b1f44a478b451c3d1ab6d992a55" dependencies = [ "bytes", "fxhash", "rand", "ring", - "rustls 0.20.4", + "rustls 0.20.6", "rustls-native-certs", "rustls-pemfile 0.2.1", "slab", @@ -1596,13 +1437,12 @@ dependencies = [ [[package]] name = "quinn-udp" -version = "0.1.1" +version = "0.1.3" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "df185e5e5f7611fa6e628ed8f9633df10114b03bbaecab186ec55822c44ac727" +checksum = "9f832d8958db3e84d2ec93b5eb2272b45aa23cf7f8fe6e79f578896f4e6c231b" dependencies = [ "futures-util", "libc", - "mio 0.7.14", "quinn-proto", "socket2", "tokio", @@ -1611,9 +1451,9 @@ dependencies = [ [[package]] name = "quote" -version = "1.0.16" +version = "1.0.21" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "b4af2ec4714533fcdf07e886f17025ace8b997b9ce51204ee69b6da831c3da57" +checksum = "bbe448f377a7d6961e30f5955f9b8d106c3f5e449d493ee1b125c1d43c2b5179" dependencies = [ "proc-macro2", ] @@ -1641,27 +1481,27 @@ dependencies = [ [[package]] name = "rand_core" -version = "0.6.3" +version = "0.6.4" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "d34f1408f55294453790c48b2f1ebbb1c5b4b7563eb1f418bcfcfdbb06ebb4e7" +checksum = "ec0be4795e2f6a28069bec0b5ff3e2ac9bafc99e6a9a7dc3547996c5c816922c" dependencies = [ "getrandom", ] [[package]] name = "redox_syscall" -version = "0.2.11" +version = "0.2.16" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "8380fe0152551244f0747b1bf41737e0f8a74f97a14ccefd1148187271634f3c" +checksum = "fb5a58c1855b4b6819d59012155603f0b22ad30cad752600aadfcb695265519a" dependencies = [ "bitflags", ] [[package]] name = "redox_users" -version = "0.4.2" +version = "0.4.3" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "7776223e2696f1aa4c6b0170e83212f47296a00424305117d013dfe86fb0fe55" +checksum = "b033d837a7cf162d7993aded9304e30a83213c648b6e389db233191f891e5c2b" dependencies = [ "getrandom", "redox_syscall", @@ -1670,9 +1510,9 @@ dependencies = [ [[package]] name = "regex" -version = "1.5.5" +version = "1.6.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "1a11647b6b25ff05a515cb92c365cec08801e83423a235b51e231e1808747286" +checksum = "4c4eb3267174b8c6c2f654116623910a0fef09c4753f8dd83db29c48a0df988b" dependencies = [ "aho-corasick", "memchr", @@ -1681,9 +1521,9 @@ dependencies = [ [[package]] name = "regex-syntax" -version = "0.6.25" +version = "0.6.27" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "f497285884f3fcff424ffc933e56d7cbca511def0c9831a7f9b5f6153e3cc89b" +checksum = "a3f87b73ce11b1619a3c6332f45341e0047173771e8b8b73f87bfeefb7b56244" [[package]] name = "ring" @@ -1700,22 +1540,32 @@ dependencies = [ "winapi", ] +[[package]] +name = "ringbuffer-spsc" +version = "0.1.6" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "3c98a13710286c8fa1b6b549078247332183534f31910d4c929aa45c03a98e50" +dependencies = [ + "array-init", + "cache-padded", +] + [[package]] name = "rsa" -version = "0.5.0" +version = "0.6.1" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "e05c2603e2823634ab331437001b411b9ed11660fbc4066f3908c84a9439260d" +checksum = "4cf22754c49613d2b3b119f0e5d46e34a2c628a937e3024b8762de4e7d8c710b" dependencies = [ "byteorder", - "digest 0.9.0", - "lazy_static", + "digest", "num-bigint-dig", "num-integer", "num-iter", "num-traits", "pkcs1", "pkcs8", - "rand", + "rand_core", + "smallvec", "subtle", "zeroize", ] @@ -1744,9 +1594,9 @@ dependencies = [ [[package]] name = "rustls" -version = "0.20.4" +version = "0.20.6" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "4fbfeb8d0ddb84706bc597a5574ab8912817c52a397f819e5b614e2265206921" +checksum = "5aab8ee6c7097ed6057f43c187a62418d0c05a4bd5f18b3571db50ee0f9ce033" dependencies = [ "log", "ring", @@ -1756,12 +1606,12 @@ dependencies = [ [[package]] name = "rustls-native-certs" -version = "0.6.1" +version = "0.6.2" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "5ca9ebdfa27d3fc180e42879037b5338ab1c040c06affd00d8338598e7800943" +checksum = "0167bac7a9f490495f3c33013e7722b53cb087ecbe082fb0c6387c96f634ea50" dependencies = [ "openssl-probe", - "rustls-pemfile 0.2.1", + "rustls-pemfile 1.0.1", "schannel", "security-framework", ] @@ -1777,27 +1627,27 @@ dependencies = [ [[package]] name = "rustls-pemfile" -version = "0.3.0" +version = "1.0.1" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "1ee86d63972a7c661d1536fefe8c3c8407321c3df668891286de28abcd087360" +checksum = "0864aeff53f8c05aa08d86e5ef839d3dfcf07aeba2db32f12db0ef716e87bd55" dependencies = [ "base64", ] [[package]] name = "ryu" -version = "1.0.9" +version = "1.0.11" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "73b4b750c782965c211b42f022f59af1fbceabdd026623714f104152f1ec149f" +checksum = "4501abdff3ae82a1c1b477a17252eb69cee9e66eb915c1abaa4f44d873df9f09" [[package]] name = "schannel" -version = "0.1.19" +version = "0.1.20" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "8f05ba609c234e60bee0d547fe94a4c7e9da733d1c962cf6e59efa4cd9c8bc75" +checksum = "88d6731146462ea25d9244b2ed5fd1d716d25c52e4d54aa4fb0f3c4e9854dbe2" dependencies = [ "lazy_static", - "winapi", + "windows-sys", ] [[package]] @@ -1828,9 +1678,9 @@ dependencies = [ [[package]] name = "security-framework" -version = "2.6.1" +version = "2.7.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "2dc14f172faf8a0194a3aded622712b0de276821addc574fa54fc0a1167e10dc" +checksum = "2bc1bb97804af6631813c55739f771071e0f2ed33ee20b68c86ec505d906356c" dependencies = [ "bitflags", "core-foundation", @@ -1851,24 +1701,24 @@ dependencies = [ [[package]] name = "semver" -version = "1.0.6" +version = "1.0.14" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "a4a3381e03edd24287172047536f20cabde766e2cd3e65e6b00fb3af51c4f38d" +checksum = "e25dfac463d778e353db5be2449d1cce89bd6fd23c9f1ea21310ce6e5a1b29c4" [[package]] name = "serde" -version = "1.0.136" +version = "1.0.144" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "ce31e24b01e1e524df96f1c2fdd054405f8d7376249a5110886fb4b658484789" +checksum = "0f747710de3dcd43b88c9168773254e809d8ddbdf9653b84e2554ab219f17860" dependencies = [ "serde_derive", ] [[package]] name = "serde_derive" -version = "1.0.136" +version = "1.0.144" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "08597e7152fcd306f41838ed3e37be9eaeed2b61c42e2117266a554fab4662f9" +checksum = "94ed3a816fb1d101812f83e789f888322c34e291f894f19590dc310963e87a00" dependencies = [ "proc-macro2", "quote", @@ -1877,9 +1727,9 @@ dependencies = [ [[package]] name = "serde_json" -version = "1.0.79" +version = "1.0.85" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "8e8d9fa5c3b304765ce1fd9c4c8a3de2c8db365a5b91be52f186efc675681d95" +checksum = "e55a28e3aaef9d5ce0506d0a14dbba8054ddc7e499ef522dd8b26859ec9d4a44" dependencies = [ "itoa", "ryu", @@ -1888,67 +1738,65 @@ dependencies = [ [[package]] name = "serde_yaml" -version = "0.8.23" +version = "0.9.13" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "a4a521f2940385c165a24ee286aa8599633d162077a54bdcae2a6fd5a7bfa7a0" +checksum = "8613d593412a0deb7bbd8de9d908efff5a0cb9ccd8f62c641e7b2ed2f57291d1" dependencies = [ "indexmap", + "itoa", "ryu", "serde", - "yaml-rust", + "unsafe-libyaml", ] [[package]] -name = "sha-1" -version = "0.8.2" +name = "sha1" +version = "0.10.5" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "f7d94d0bede923b3cea61f3f1ff57ff8cdfd77b400fb8f9998949e0cf04163df" +checksum = "f04293dc80c3993519f2d7f6f511707ee7094fe0c6d3406feb330cdb3540eba3" dependencies = [ - "block-buffer 0.7.3", - "digest 0.8.1", - "fake-simd", - "opaque-debug 0.2.3", + "cfg-if", + "cpufeatures", + "digest", ] [[package]] name = "sha3" -version = "0.9.1" +version = "0.10.5" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "f81199417d4e5de3f04b1e871023acea7389672c4135918f05aa9cbf2f2fa809" +checksum = "e2904bea16a1ae962b483322a1c7b81d976029203aea1f461e51cd7705db7ba9" dependencies = [ - "block-buffer 0.9.0", - "digest 0.9.0", + "digest", "keccak", - "opaque-debug 0.3.0", ] [[package]] name = "shared_memory" -version = "0.12.0" +version = "0.12.4" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "681a9e90340f748af3a1cc52eb2c040eee29f976b763e99ad90fc0c5df6f9791" +checksum = "ba8593196da75d9dc4f69349682bd4c2099f8cde114257d1ef7ef1b33d1aba54" dependencies = [ "cfg-if", "libc", - "nix 0.22.3", + "nix 0.23.1", "rand", - "winapi", + "win-sys", ] [[package]] name = "shellexpand" -version = "2.1.0" +version = "2.1.2" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "83bdb7831b2d85ddf4a7b148aa19d0587eddbe8671a436b7bd1182eaad0f2829" +checksum = "7ccc8076840c4da029af4f87e4e8daeb0fca6b87bbb02e10cb60b791450e11e4" dependencies = [ - "dirs-next", + "dirs", ] [[package]] name = "signal-hook" -version = "0.3.13" +version = "0.3.14" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "647c97df271007dcea485bb74ffdb57f2e683f1306c854f468a0c244badabf2d" +checksum = "a253b5e89e2698464fc26b545c9edceb338e18a89effeeecfea192c3025be29d" dependencies = [ "libc", "signal-hook-registry", @@ -1965,21 +1813,24 @@ dependencies = [ [[package]] name = "slab" -version = "0.4.5" +version = "0.4.7" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "9def91fd1e018fe007022791f865d0ccc9b3a0d5001e01aabb8b40e46000afb5" +checksum = "4614a76b2a8be0058caa9dbbaf66d988527d86d003c11a94fbd335d7661edcef" +dependencies = [ + "autocfg", +] [[package]] name = "smallvec" -version = "1.8.0" +version = "1.9.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "f2dd574626839106c320a323308629dcb1acfc96e32a8cba364ddc61ac23ee83" +checksum = "2fd0db749597d91ff862fd1d55ea87f7855a744a8425a64695b6fca237d1dad1" [[package]] name = "socket2" -version = "0.4.4" +version = "0.4.7" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "66d72b759436ae32898a2af0a14218dbf55efde3feeb170eb623637db85ee1e0" +checksum = "02e2d2db9033d13a1567121ddd7a095ee144db4e1ca1b1bda3419bc0da294ebd" dependencies = [ "libc", "winapi", @@ -1993,19 +1844,20 @@ checksum = "6e63cff320ae2c57904679ba7cb63280a3dc4613885beafb148ee7bf9aa9042d" [[package]] name = "spin" -version = "0.9.2" +version = "0.9.4" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "511254be0c5bcf062b019a6c89c01a664aa359ded62f78aa72c6fc137c0590e5" +checksum = "7f6002a767bff9e83f8eeecf883ecb8011875a21ae8da43bffb817a57e78cc09" dependencies = [ "lock_api", ] [[package]] name = "spki" -version = "0.4.1" +version = "0.5.4" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "5c01a0c15da1b0b0e1494112e7af814a678fec9bd157881b49beac661e9b6f32" +checksum = "44d01ac02a6ccf3e07db148d2be087da624fea0221a16152ed01f0496a6b0a27" dependencies = [ + "base64ct", "der", ] @@ -2023,9 +1875,9 @@ dependencies = [ [[package]] name = "strsim" -version = "0.8.0" +version = "0.10.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "8ea5119cdb4c55b55d432abb513a0429384878c15dde60cc77b1c99de1a95a6a" +checksum = "73473c0e59e6d5812c5dfe2a064a6444949f089e20eec9a2e5506596494e4623" [[package]] name = "subtle" @@ -2035,26 +1887,20 @@ checksum = "6bdef32e8150c2a081110b42772ffe7d7c9032b606bc226c8260fd97e0976601" [[package]] name = "syn" -version = "1.0.89" +version = "1.0.100" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "ea297be220d52398dcc07ce15a209fce436d361735ac1db700cab3b6cdfb9f54" +checksum = "52205623b1b0f064a4e71182c3b18ae902267282930c6d5462c91b859668426e" dependencies = [ "proc-macro2", "quote", - "unicode-xid", + "unicode-ident", ] [[package]] -name = "synstructure" -version = "0.12.6" +name = "target-lexicon" +version = "0.12.4" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "f36bdaa60a83aca3921b5259d5400cbf5e90fc51931376a9bd4a0eb79aa7210f" -dependencies = [ - "proc-macro2", - "quote", - "syn", - "unicode-xid", -] +checksum = "c02424087780c9b71cc96799eaeddff35af2bc513278cda5c99fc1f5d026d3c1" [[package]] name = "termcolor" @@ -2067,27 +1913,24 @@ dependencies = [ [[package]] name = "textwrap" -version = "0.11.0" +version = "0.15.1" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "d326610f408c7a4eb6f51c37c330e496b08506c9457c9d34287ecc38809fb060" -dependencies = [ - "unicode-width", -] +checksum = "949517c0cf1bf4ee812e2e07e08ab448e3ae0d23472aee8a06c985f0c8815b16" [[package]] name = "thiserror" -version = "1.0.30" +version = "1.0.35" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "854babe52e4df1653706b98fcfc05843010039b406875930a70e4d9644e5c417" +checksum = "c53f98874615aea268107765aa1ed8f6116782501d18e53d08b471733bea6c85" dependencies = [ "thiserror-impl", ] [[package]] name = "thiserror-impl" -version = "1.0.30" +version = "1.0.35" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "aa32fd3f627f367fe16f893e2597ae3c05020f8bba2666a4e6ea73d377e5714b" +checksum = "f8b463991b4eab2d801e724172285ec4195c650e8ec79b149e6c2a8e6dd3f783" dependencies = [ "proc-macro2", "quote", @@ -2096,9 +1939,9 @@ dependencies = [ [[package]] name = "tinyvec" -version = "1.5.1" +version = "1.6.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "2c1c1d5a42b6245520c249549ec267180beaffcc0615401ac8e31853d4b6d8d2" +checksum = "87cc5ceb3875bb20c2890005a4e226a4651264a5c75edb2421b52861a0a0cb50" dependencies = [ "tinyvec_macros", ] @@ -2111,13 +1954,15 @@ checksum = "cda74da7e1a664f795bb1f8a87ec406fb89a02522cf6e50620d016add6dbbf5c" [[package]] name = "tokio" -version = "1.17.0" +version = "1.21.1" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "2af73ac49756f3f7c01172e34a23e5d0216f6c32333757c2c61feb2bbff5a5ee" +checksum = "0020c875007ad96677dcc890298f4b942882c5d4eb7cc8f439fc3bf813dc9c95" dependencies = [ + "autocfg", "libc", - "mio 0.8.2", + "mio", "num_cpus", + "once_cell", "pin-project-lite", "socket2", "winapi", @@ -2125,9 +1970,9 @@ dependencies = [ [[package]] name = "tracing" -version = "0.1.32" +version = "0.1.36" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "4a1bdf54a7c28a2bbf701e1d2233f6c77f473486b94bee4f9678da5a148dca7f" +checksum = "2fce9567bd60a67d08a16488756721ba392f24f29006402881e43b19aac64307" dependencies = [ "cfg-if", "pin-project-lite", @@ -2137,9 +1982,9 @@ dependencies = [ [[package]] name = "tracing-attributes" -version = "0.1.20" +version = "0.1.22" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "2e65ce065b4b5c53e73bb28912318cb8c9e9ad3921f1d669eb0e68b4c8143a2b" +checksum = "11c75893af559bc8e10716548bdef5cb2b983f8e637db9d0e15126b61b484ee2" dependencies = [ "proc-macro2", "quote", @@ -2148,11 +1993,11 @@ dependencies = [ [[package]] name = "tracing-core" -version = "0.1.23" +version = "0.1.29" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "aa31669fa42c09c34d94d8165dd2012e8ff3c66aca50f3bb226b68f216f2706c" +checksum = "5aeea4303076558a00714b823f9ad67d58a3bbda1df83d8827d21193156e22f7" dependencies = [ - "lazy_static", + "once_cell", ] [[package]] @@ -2163,15 +2008,15 @@ checksum = "dcf81ac59edc17cc8697ff311e8f5ef2d99fcbd9817b34cec66f90b6c3dfd987" [[package]] name = "ucd-trie" -version = "0.1.3" +version = "0.1.5" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "56dee185309b50d1f11bfedef0fe6d036842e3fb77413abef29f8f8d1c5d4c1c" +checksum = "9e79c4d996edb816c91e4308506774452e55e95c3c9de07b6729e17e15a5ef81" [[package]] name = "uhlc" -version = "0.4.1" +version = "0.5.1" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "9d74cc14aac0f650dae365e42250073bc0f89602bad5751d6159642be37690d6" +checksum = "7908438f98a5824af02b34c2b31fb369c5764ef835d26df0badbb9897fb28245" dependencies = [ "hex", "humantime", @@ -2182,22 +2027,22 @@ dependencies = [ ] [[package]] -name = "unicode-width" -version = "0.1.9" +name = "unicode-ident" +version = "1.0.4" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "3ed742d4ea2bd1176e236172c8429aaf54486e7ac098db29ffe6529e0ce50973" +checksum = "dcc811dc4066ac62f84f11307873c4850cb653bfa9b1719cee2bd2204a4bc5dd" [[package]] -name = "unicode-xid" -version = "0.2.2" +name = "unindent" +version = "0.1.10" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "8ccb82d61f80a663efe1f787a51b16b5a51e3314d6ac365b08639f52387b33f3" +checksum = "58ee9362deb4a96cef4d437d1ad49cffc9b9e92d202b6995674e928ce684f112" [[package]] -name = "unindent" -version = "0.1.8" +name = "unsafe-libyaml" +version = "0.2.4" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "514672a55d7380da379785a4d70ca8386c8883ff7eaae877be4d2081cebe73d8" +checksum = "c1e5fa573d8ac5f1a856f8d7be41d390ee973daf97c806b2c1a465e4e1406e68" [[package]] name = "untrusted" @@ -2218,23 +2063,13 @@ dependencies = [ [[package]] name = "uuid" -version = "0.8.2" +version = "1.1.2" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "bc5cf98d8186244414c848017f0e2676b3fcb46807f6668a97dfe67359a3c4b7" +checksum = "dd6469f4314d5f1ffec476e05f17cc9a78bc7a27a6a857842170bdf8d6f98d2f" dependencies = [ "getrandom", ] -[[package]] -name = "validated_struct" -version = "0.1.11" -source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "85affc53ebb88700a36aefff74b6a4db38dbccf26ec777d40c9e9b471be261f8" -dependencies = [ - "serde", - "validated_struct_macros 0.1.10", -] - [[package]] name = "validated_struct" version = "2.1.0" @@ -2244,19 +2079,7 @@ dependencies = [ "json5", "serde", "serde_json", - "validated_struct_macros 2.1.0", -] - -[[package]] -name = "validated_struct_macros" -version = "0.1.10" -source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "ca689f3addec28c7d67ed3139498d0325253c71bf58ef57c2ba76141f7ce576f" -dependencies = [ - "proc-macro2", - "quote", - "syn", - "unzip-n", + "validated_struct_macros", ] [[package]] @@ -2273,9 +2096,9 @@ dependencies = [ [[package]] name = "value-bag" -version = "1.0.0-alpha.8" +version = "1.0.0-alpha.9" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "79923f7731dc61ebfba3633098bf3ac533bbd35ccd8c57e7088d9a5eebe0263f" +checksum = "2209b78d1249f7e6f3293657c9779fe31ced465df091bbd433a1cf88e916ec55" dependencies = [ "ctor", "version_check", @@ -2299,12 +2122,6 @@ version = "1.1.0" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "9d5b2c62b4012a3e1eca5a7e077d13b3bf498c4073e33ccd58626607748ceeca" -[[package]] -name = "wasi" -version = "0.10.2+wasi-snapshot-preview1" -source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "fd6fbd9a79829dd1ad0cc20627bf1ed606756a7f77edff7b66b7064f9cb327c6" - [[package]] name = "wasi" version = "0.11.0+wasi-snapshot-preview1" @@ -2313,9 +2130,9 @@ checksum = "9c8d87e72b64a3b4db28d11ce29237c246188f4f51057d65a7eab63b7987e423" [[package]] name = "wasm-bindgen" -version = "0.2.79" +version = "0.2.83" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "25f1af7423d8588a3d840681122e72e6a24ddbcb3f0ec385cac0d12d24256c06" +checksum = "eaf9f5aceeec8be17c128b2e93e031fb8a4d469bb9c4ae2d7dc1888b26887268" dependencies = [ "cfg-if", "wasm-bindgen-macro", @@ -2323,13 +2140,13 @@ dependencies = [ [[package]] name = "wasm-bindgen-backend" -version = "0.2.79" +version = "0.2.83" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "8b21c0df030f5a177f3cba22e9bc4322695ec43e7257d865302900290bcdedca" +checksum = "4c8ffb332579b0557b52d268b91feab8df3615f265d5270fec2a8c95b17c1142" dependencies = [ "bumpalo", - "lazy_static", "log", + "once_cell", "proc-macro2", "quote", "syn", @@ -2338,9 +2155,9 @@ dependencies = [ [[package]] name = "wasm-bindgen-futures" -version = "0.4.29" +version = "0.4.33" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "2eb6ec270a31b1d3c7e266b999739109abce8b6c87e4b31fcfcd788b65267395" +checksum = "23639446165ca5a5de86ae1d8896b737ae80319560fbaa4c2887b7da6e7ebd7d" dependencies = [ "cfg-if", "js-sys", @@ -2350,9 +2167,9 @@ dependencies = [ [[package]] name = "wasm-bindgen-macro" -version = "0.2.79" +version = "0.2.83" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "2f4203d69e40a52ee523b2529a773d5ffc1dc0071801c87b3d270b471b80ed01" +checksum = "052be0f94026e6cbc75cdefc9bae13fd6052cdcaf532fa6c45e7ae33a1e6c810" dependencies = [ "quote", "wasm-bindgen-macro-support", @@ -2360,9 +2177,9 @@ dependencies = [ [[package]] name = "wasm-bindgen-macro-support" -version = "0.2.79" +version = "0.2.83" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "bfa8a30d46208db204854cadbb5d4baf5fcf8071ba5bf48190c3e59937962ebc" +checksum = "07bc0c051dc5f23e307b13285f9d75df86bfdf816c5721e573dec1f9b8aa193c" dependencies = [ "proc-macro2", "quote", @@ -2373,15 +2190,15 @@ dependencies = [ [[package]] name = "wasm-bindgen-shared" -version = "0.2.79" +version = "0.2.83" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "3d958d035c4438e28c70e4321a2911302f10135ce78a9c7834c0cab4123d06a2" +checksum = "1c38c045535d93ec4f0b4defec448e4291638ee608530863b1e2ba115d4fff7f" [[package]] name = "web-sys" -version = "0.3.56" +version = "0.3.60" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "c060b319f29dd25724f09a2ba1418f142f539b2be99fbf4d2d5a8f7330afb8eb" +checksum = "bcda906d8be16e728fd5adc5b729afad4e444e106ab28cd1c7256e54fa61510f" dependencies = [ "js-sys", "wasm-bindgen", @@ -2416,6 +2233,15 @@ dependencies = [ "cc", ] +[[package]] +name = "win-sys" +version = "0.3.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "5b7b128a98c1cfa201b09eb49ba285887deb3cbe7466a98850eb1adabb452be5" +dependencies = [ + "windows", +] + [[package]] name = "winapi" version = "0.3.9" @@ -2448,18 +2274,95 @@ source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "712e227841d057c1ee1cd2fb22fa7e5a5461ae8e48fa2ca79ec42cfc1931183f" [[package]] -name = "yaml-rust" -version = "0.4.5" +name = "windows" +version = "0.34.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "56c1936c4cc7a1c9ab21a1ebb602eb942ba868cbd44a99cb7cdc5892335e1c85" +checksum = "45296b64204227616fdbf2614cefa4c236b98ee64dfaaaa435207ed99fe7829f" dependencies = [ - "linked-hash-map", + "windows_aarch64_msvc 0.34.0", + "windows_i686_gnu 0.34.0", + "windows_i686_msvc 0.34.0", + "windows_x86_64_gnu 0.34.0", + "windows_x86_64_msvc 0.34.0", ] +[[package]] +name = "windows-sys" +version = "0.36.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "ea04155a16a59f9eab786fe12a4a450e75cdb175f9e0d80da1e17db09f55b8d2" +dependencies = [ + "windows_aarch64_msvc 0.36.1", + "windows_i686_gnu 0.36.1", + "windows_i686_msvc 0.36.1", + "windows_x86_64_gnu 0.36.1", + "windows_x86_64_msvc 0.36.1", +] + +[[package]] +name = "windows_aarch64_msvc" +version = "0.34.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "17cffbe740121affb56fad0fc0e421804adf0ae00891205213b5cecd30db881d" + +[[package]] +name = "windows_aarch64_msvc" +version = "0.36.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "9bb8c3fd39ade2d67e9874ac4f3db21f0d710bee00fe7cab16949ec184eeaa47" + +[[package]] +name = "windows_i686_gnu" +version = "0.34.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "2564fde759adb79129d9b4f54be42b32c89970c18ebf93124ca8870a498688ed" + +[[package]] +name = "windows_i686_gnu" +version = "0.36.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "180e6ccf01daf4c426b846dfc66db1fc518f074baa793aa7d9b9aaeffad6a3b6" + +[[package]] +name = "windows_i686_msvc" +version = "0.34.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "9cd9d32ba70453522332c14d38814bceeb747d80b3958676007acadd7e166956" + +[[package]] +name = "windows_i686_msvc" +version = "0.36.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "e2e7917148b2812d1eeafaeb22a97e4813dfa60a3f8f78ebe204bcc88f12f024" + +[[package]] +name = "windows_x86_64_gnu" +version = "0.34.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "cfce6deae227ee8d356d19effc141a509cc503dfd1f850622ec4b0f84428e1f4" + +[[package]] +name = "windows_x86_64_gnu" +version = "0.36.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "4dcd171b8776c41b97521e5da127a2d86ad280114807d0b2ab1e462bc764d9e1" + +[[package]] +name = "windows_x86_64_msvc" +version = "0.34.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "d19538ccc21819d01deaf88d6a17eae6596a12e9aafdbb97916fb49896d89de9" + +[[package]] +name = "windows_x86_64_msvc" +version = "0.36.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "c811ca4a8c853ef420abd8592ba53ddbbac90410fab6903b3e79972a631f7680" + [[package]] name = "zenoh" version = "0.6.0-dev.0" -source = "git+https://github.com/eclipse-zenoh/zenoh#a3fecd9909ff0d7bccedbf35fc0f1c7e57c0d34b" +source = "git+https://github.com/eclipse-zenoh/zenoh?branch=master#4d8f680f746c57f16cf9cc15e3b44bf32d3de83e" dependencies = [ "async-global-executor", "async-std", @@ -2468,8 +2371,8 @@ dependencies = [ "env_logger", "event-listener", "flume", + "form_urlencoded", "futures", - "futures-lite", "git-version", "hex", "lazy_static", @@ -2504,7 +2407,7 @@ dependencies = [ [[package]] name = "zenoh-buffers" version = "0.6.0-dev.0" -source = "git+https://github.com/eclipse-zenoh/zenoh#a3fecd9909ff0d7bccedbf35fc0f1c7e57c0d34b" +source = "git+https://github.com/eclipse-zenoh/zenoh?branch=master#4d8f680f746c57f16cf9cc15e3b44bf32d3de83e" dependencies = [ "async-std", "bincode", @@ -2519,7 +2422,7 @@ dependencies = [ [[package]] name = "zenoh-cfg-properties" version = "0.6.0-dev.0" -source = "git+https://github.com/eclipse-zenoh/zenoh#a3fecd9909ff0d7bccedbf35fc0f1c7e57c0d34b" +source = "git+https://github.com/eclipse-zenoh/zenoh?branch=master#4d8f680f746c57f16cf9cc15e3b44bf32d3de83e" dependencies = [ "zenoh-core", "zenoh-macros", @@ -2528,7 +2431,7 @@ dependencies = [ [[package]] name = "zenoh-collections" version = "0.6.0-dev.0" -source = "git+https://github.com/eclipse-zenoh/zenoh#a3fecd9909ff0d7bccedbf35fc0f1c7e57c0d34b" +source = "git+https://github.com/eclipse-zenoh/zenoh?branch=master#4d8f680f746c57f16cf9cc15e3b44bf32d3de83e" dependencies = [ "async-std", "async-trait", @@ -2541,14 +2444,15 @@ dependencies = [ [[package]] name = "zenoh-config" version = "0.6.0-dev.0" -source = "git+https://github.com/eclipse-zenoh/zenoh#a3fecd9909ff0d7bccedbf35fc0f1c7e57c0d34b" +source = "git+https://github.com/eclipse-zenoh/zenoh?branch=master#4d8f680f746c57f16cf9cc15e3b44bf32d3de83e" dependencies = [ "flume", "json5", + "num_cpus", "serde", "serde_json", "serde_yaml", - "validated_struct 2.1.0", + "validated_struct", "zenoh-cfg-properties", "zenoh-core", "zenoh-protocol-core", @@ -2558,16 +2462,17 @@ dependencies = [ [[package]] name = "zenoh-core" version = "0.6.0-dev.0" -source = "git+https://github.com/eclipse-zenoh/zenoh#a3fecd9909ff0d7bccedbf35fc0f1c7e57c0d34b" +source = "git+https://github.com/eclipse-zenoh/zenoh?branch=master#4d8f680f746c57f16cf9cc15e3b44bf32d3de83e" dependencies = [ "anyhow", "lazy_static", + "zenoh-macros", ] [[package]] name = "zenoh-crypto" version = "0.6.0-dev.0" -source = "git+https://github.com/eclipse-zenoh/zenoh#a3fecd9909ff0d7bccedbf35fc0f1c7e57c0d34b" +source = "git+https://github.com/eclipse-zenoh/zenoh?branch=master#4d8f680f746c57f16cf9cc15e3b44bf32d3de83e" dependencies = [ "aes", "hmac", @@ -2580,7 +2485,7 @@ dependencies = [ [[package]] name = "zenoh-link" version = "0.6.0-dev.0" -source = "git+https://github.com/eclipse-zenoh/zenoh#a3fecd9909ff0d7bccedbf35fc0f1c7e57c0d34b" +source = "git+https://github.com/eclipse-zenoh/zenoh?branch=master#4d8f680f746c57f16cf9cc15e3b44bf32d3de83e" dependencies = [ "async-std", "async-trait", @@ -2599,7 +2504,7 @@ dependencies = [ [[package]] name = "zenoh-link-commons" version = "0.6.0-dev.0" -source = "git+https://github.com/eclipse-zenoh/zenoh#a3fecd9909ff0d7bccedbf35fc0f1c7e57c0d34b" +source = "git+https://github.com/eclipse-zenoh/zenoh?branch=master#4d8f680f746c57f16cf9cc15e3b44bf32d3de83e" dependencies = [ "async-std", "async-trait", @@ -2614,15 +2519,16 @@ dependencies = [ [[package]] name = "zenoh-link-quic" version = "0.6.0-dev.0" -source = "git+https://github.com/eclipse-zenoh/zenoh#a3fecd9909ff0d7bccedbf35fc0f1c7e57c0d34b" +source = "git+https://github.com/eclipse-zenoh/zenoh?branch=master#4d8f680f746c57f16cf9cc15e3b44bf32d3de83e" dependencies = [ "async-std", "async-trait", + "futures", "log", "quinn", - "rustls 0.20.4", + "rustls 0.20.6", "rustls-native-certs", - "rustls-pemfile 0.3.0", + "rustls-pemfile 1.0.1", "webpki 0.22.0", "zenoh-cfg-properties", "zenoh-config", @@ -2636,7 +2542,7 @@ dependencies = [ [[package]] name = "zenoh-link-tcp" version = "0.6.0-dev.0" -source = "git+https://github.com/eclipse-zenoh/zenoh#a3fecd9909ff0d7bccedbf35fc0f1c7e57c0d34b" +source = "git+https://github.com/eclipse-zenoh/zenoh?branch=master#4d8f680f746c57f16cf9cc15e3b44bf32d3de83e" dependencies = [ "async-std", "async-trait", @@ -2651,11 +2557,12 @@ dependencies = [ [[package]] name = "zenoh-link-tls" version = "0.6.0-dev.0" -source = "git+https://github.com/eclipse-zenoh/zenoh#a3fecd9909ff0d7bccedbf35fc0f1c7e57c0d34b" +source = "git+https://github.com/eclipse-zenoh/zenoh?branch=master#4d8f680f746c57f16cf9cc15e3b44bf32d3de83e" dependencies = [ "async-rustls", "async-std", "async-trait", + "futures", "log", "zenoh-cfg-properties", "zenoh-config", @@ -2669,7 +2576,7 @@ dependencies = [ [[package]] name = "zenoh-link-udp" version = "0.6.0-dev.0" -source = "git+https://github.com/eclipse-zenoh/zenoh#a3fecd9909ff0d7bccedbf35fc0f1c7e57c0d34b" +source = "git+https://github.com/eclipse-zenoh/zenoh?branch=master#4d8f680f746c57f16cf9cc15e3b44bf32d3de83e" dependencies = [ "async-std", "async-trait", @@ -2686,12 +2593,13 @@ dependencies = [ [[package]] name = "zenoh-link-unixsock_stream" version = "0.6.0-dev.0" -source = "git+https://github.com/eclipse-zenoh/zenoh#a3fecd9909ff0d7bccedbf35fc0f1c7e57c0d34b" +source = "git+https://github.com/eclipse-zenoh/zenoh?branch=master#4d8f680f746c57f16cf9cc15e3b44bf32d3de83e" dependencies = [ "async-std", "async-trait", + "futures", "log", - "nix 0.23.1", + "nix 0.25.0", "uuid", "zenoh-core", "zenoh-link-commons", @@ -2702,7 +2610,7 @@ dependencies = [ [[package]] name = "zenoh-macros" version = "0.6.0-dev.0" -source = "git+https://github.com/eclipse-zenoh/zenoh#a3fecd9909ff0d7bccedbf35fc0f1c7e57c0d34b" +source = "git+https://github.com/eclipse-zenoh/zenoh?branch=master#4d8f680f746c57f16cf9cc15e3b44bf32d3de83e" dependencies = [ "proc-macro2", "quote", @@ -2714,7 +2622,7 @@ dependencies = [ [[package]] name = "zenoh-plugin-trait" version = "0.6.0-dev.0" -source = "git+https://github.com/eclipse-zenoh/zenoh#a3fecd9909ff0d7bccedbf35fc0f1c7e57c0d34b" +source = "git+https://github.com/eclipse-zenoh/zenoh?branch=master#4d8f680f746c57f16cf9cc15e3b44bf32d3de83e" dependencies = [ "libloading", "log", @@ -2727,7 +2635,7 @@ dependencies = [ [[package]] name = "zenoh-protocol" version = "0.6.0-dev.0" -source = "git+https://github.com/eclipse-zenoh/zenoh#a3fecd9909ff0d7bccedbf35fc0f1c7e57c0d34b" +source = "git+https://github.com/eclipse-zenoh/zenoh?branch=master#4d8f680f746c57f16cf9cc15e3b44bf32d3de83e" dependencies = [ "log", "uhlc", @@ -2739,10 +2647,12 @@ dependencies = [ [[package]] name = "zenoh-protocol-core" version = "0.6.0-dev.0" -source = "git+https://github.com/eclipse-zenoh/zenoh#a3fecd9909ff0d7bccedbf35fc0f1c7e57c0d34b" +source = "git+https://github.com/eclipse-zenoh/zenoh?branch=master#4d8f680f746c57f16cf9cc15e3b44bf32d3de83e" dependencies = [ "hex", + "itertools", "lazy_static", + "rand", "serde", "uhlc", "uuid", @@ -2752,12 +2662,12 @@ dependencies = [ [[package]] name = "zenoh-sync" version = "0.6.0-dev.0" -source = "git+https://github.com/eclipse-zenoh/zenoh#a3fecd9909ff0d7bccedbf35fc0f1c7e57c0d34b" +source = "git+https://github.com/eclipse-zenoh/zenoh?branch=master#4d8f680f746c57f16cf9cc15e3b44bf32d3de83e" dependencies = [ "async-std", "event-listener", "flume", - "futures-lite", + "futures", "tokio", "zenoh-core", ] @@ -2765,15 +2675,17 @@ dependencies = [ [[package]] name = "zenoh-transport" version = "0.6.0-dev.0" -source = "git+https://github.com/eclipse-zenoh/zenoh#a3fecd9909ff0d7bccedbf35fc0f1c7e57c0d34b" +source = "git+https://github.com/eclipse-zenoh/zenoh?branch=master#4d8f680f746c57f16cf9cc15e3b44bf32d3de83e" dependencies = [ + "async-executor", "async-global-executor", "async-std", "async-trait", "flume", "log", - "paste 1.0.6", + "paste", "rand", + "ringbuffer-spsc", "rsa", "serde", "zenoh-buffers", @@ -2791,7 +2703,7 @@ dependencies = [ [[package]] name = "zenoh-util" version = "0.6.0-dev.0" -source = "git+https://github.com/eclipse-zenoh/zenoh#a3fecd9909ff0d7bccedbf35fc0f1c7e57c0d34b" +source = "git+https://github.com/eclipse-zenoh/zenoh?branch=master#4d8f680f746c57f16cf9cc15e3b44bf32d3de83e" dependencies = [ "async-std", "clap", @@ -2804,6 +2716,7 @@ dependencies = [ "libloading", "log", "pnet", + "pnet_datalink", "shellexpand", "winapi", "zenoh-cfg-properties", @@ -2815,21 +2728,6 @@ dependencies = [ [[package]] name = "zeroize" -version = "1.4.3" +version = "1.5.7" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "d68d9dcec5f9b43a30d38c49f91dfedfaac384cb8f085faca366c26207dd1619" -dependencies = [ - "zeroize_derive", -] - -[[package]] -name = "zeroize_derive" -version = "1.3.2" -source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "3f8f187641dad4f680d25c4bfc4225b418165984179f26ca76ec4fb6441d3a17" -dependencies = [ - "proc-macro2", - "quote", - "syn", - "synstructure", -] +checksum = "c394b5bd0c6f669e7275d9c20aa90ae064cb22e75a1cad54e1b34088034b149f" diff --git a/Cargo.toml b/Cargo.toml index 9737bfc1..007f47c3 100644 --- a/Cargo.toml +++ b/Cargo.toml @@ -34,17 +34,17 @@ crate-type = ["cdylib"] complete_n = ["zenoh/complete_n"] [dependencies] -zenoh = { git = "https://github.com/eclipse-zenoh/zenoh" } -zenoh-cfg-properties = { git = "https://github.com/eclipse-zenoh/zenoh" } -zenoh-core = { git = "https://github.com/eclipse-zenoh/zenoh" } -zenoh-buffers = { git = "https://github.com/eclipse-zenoh/zenoh" } -validated_struct = "0.1" +async-std = "=1.12.0" +env_logger = "0.9.1" +form_urlencoded = "1.1.0" +futures = "0.3.24" json5 = "0.4.1" -serde_json = "1.0" -async-std = "=1.11.0" -uhlc = "0.4.0" -futures = "0.3.12" -log = "0.4" -env_logger = "0.9.0" -pyo3 = { version = "0.16", features = ["extension-module", "abi3-py37"] } -pyo3-asyncio = { version = "0.16", features = ["attributes", "async-std-runtime"] } +log = "0.4.17" +pyo3 = { version = "0.17.1", features = ["extension-module", "abi3-py37"] } +serde_json = "1.0.85" +uhlc = "0.5.1" +validated_struct = "2.1.0" +zenoh = { git = "https://github.com/eclipse-zenoh/zenoh", branch = "master" } +zenoh-buffers = { git = "https://github.com/eclipse-zenoh/zenoh", branch = "master" } +zenoh-cfg-properties = { git = "https://github.com/eclipse-zenoh/zenoh", branch = "master" } +zenoh-core = { git = "https://github.com/eclipse-zenoh/zenoh", branch = "master" } diff --git a/Jenkinsfile b/Jenkinsfile deleted file mode 100644 index 9322d9e1..00000000 --- a/Jenkinsfile +++ /dev/null @@ -1,119 +0,0 @@ -pipeline { - agent { label 'MacMini' } - options { skipDefaultCheckout() } - parameters { - gitParameter(name: 'GIT_TAG', - type: 'PT_BRANCH_TAG', - description: 'The Git tag to checkout. If not specified "master" will be checkout.', - defaultValue: 'master') - booleanParam(name: 'PUBLISH_RESULTS', - description: 'Publish the resulting wheels to Eclipse download and to pypi.org (if not a branch)', - defaultValue: false) - booleanParam(name: 'BUILD_MACOSX', - description: 'Build wheels for macosx_10_7_x86_64.', - defaultValue: true) - booleanParam(name: 'BUILD_LINUX64', - description: 'Build wheels with manylinux2010-x86-64.', - defaultValue: true) - booleanParam(name: 'BUILD_LINUX32', - description: 'Build wheels with manylinux2010-i686.', - defaultValue: true) - booleanParam(name: 'BUILD_AARCH64', - description: 'Build wheels with manylinux2014-aarch64.', - defaultValue: true) - } - environment { - LABEL = get_label() - DOWNLOAD_DIR="/home/data/httpd/download.eclipse.org/zenoh/zenoh-python/${LABEL}" - } - - stages { - stage('Checkout Git TAG') { - steps { - deleteDir() - checkout([$class: 'GitSCM', - branches: [[name: "${params.GIT_TAG}"]], - doGenerateSubmoduleConfigurations: false, - extensions: [], - gitTool: 'Default', - submoduleCfg: [], - userRemoteConfigs: [[url: 'https://github.com/eclipse-zenoh/zenoh-python.git']] - ]) - } - } - - stage('MacOS wheels') { - when { expression { return params.BUILD_MACOSX }} - steps { - sh ''' - set +x - . ~/.zshenv - env - maturin build --release $MATURIN_PYTHONS_OPT - ''' - } - } - - stage('Manylinux2010-x64 wheels') { - when { expression { return params.BUILD_LINUX64 }} - steps { - sh ''' - docker run --init --rm -v $(pwd):/workdir -w /workdir adlinktech/zenoh-dev-manylinux2010-x86_64-gnu maturin build --release --manylinux 2010 - ''' - } - } - - stage('Manylinux2010-i686 wheels') { - when { expression { return params.BUILD_LINUX32 }} - steps { - sh ''' - docker run --init --rm -v $(pwd):/workdir -w /workdir adlinktech/zenoh-dev-manylinux2010-i686-gnu maturin build --release --manylinux 2010 - ''' - } - } - - stage('Manylinux2014-aarch64 wheels') { - when { expression { return params.BUILD_AARCH64 }} - steps { - sh ''' - docker run --init --rm -v $(pwd):/workdir -w /workdir adlinktech/zenoh-dev-manylinux2014-aarch64-gnu maturin build --release --manylinux 2014 - ''' - } - } - - stage('Deploy to download.eclipse.org') { - when { expression { return params.PUBLISH_RESULTS }} - steps { - sshagent ( ['projects-storage.eclipse.org-bot-ssh']) { - sh ''' - if [[ ${GIT_TAG} == origin/* ]]; then - ssh [email protected] rm -fr ${DOWNLOAD_DIR} - fi - ssh [email protected] mkdir -p ${DOWNLOAD_DIR} - scp target/wheels/*.tar.gz [email protected]:${DOWNLOAD_DIR}/ - find target -name "*.whl" | xargs -J FILES scp FILES [email protected]:${DOWNLOAD_DIR}/ - ''' - } - } - } - - stage('Deploy on pypi.org') { - when { expression { return params.PUBLISH_RESULTS && !env.GIT_TAG.startsWith('origin/') }} - steps { - sh ''' - python3 -m twine upload --repository eclipse-zenoh `find target -name "*.whl"` target/wheels/*.tar.gz - ''' - } - } - } - - post { - success { - archiveArtifacts artifacts: 'target/**/*.whl, target/wheels/*.tar.gz', fingerprint: true - } - } -} - -def get_label() { - return env.GIT_TAG.startsWith('origin/') ? env.GIT_TAG.minus('origin/') : env.GIT_TAG -} diff --git a/README.md b/README.md index f163413f..eb5c3b14 100644 --- a/README.md +++ b/README.md @@ -48,6 +48,7 @@ It relies on the [zenoh](https://github.com/eclipse-zenoh/zenoh/tree/master/zeno Requirements: * Python >= 3.7 * pip >= 19.3.1 + * (Optional) A Python virtual environment (for instance [virtualenv](docs.python.org/3.10/tutorial/venv.html) or [miniconda](https://docs.conda.io/en/latest/miniconda.html)) * [Rust and Cargo](https://doc.rust-lang.org/cargo/getting-started/installation.html). Steps: @@ -56,19 +57,24 @@ Steps: pip install -r requirements-dev.txt ``` - * Ensure your system can find the building tool maturin. - For example, it is placed at _$HOME/.local/bin/maturin_ by default on Ubuntu 20.04. - + * Ensure your system can find the building tool `maturin` (installed by previous step). + For example, it is placed at _$HOME/.local/bin/maturin_ by default on Ubuntu 20.04. ```bash export PATH="$HOME/.local/bin:$PATH" ``` - * Build zenoh-python - ```bash - maturin build --release - ``` + * Build and install zenoh-python: + * With a virtual environment active: + ```bash + maturin develop --release + ``` + * Without one: + ```bash + maturin build --release + pip install ./target/wheels/<there should only be one .whl file here> + ``` + -This will automatically build the zenoh Rust API, as well as the zenoh-python API and install it in your Python environement. ------------------------------- ## Running the Examples diff --git a/docs/conf.py b/docs/conf.py index 9589cd83..e50d3c81 100644 --- a/docs/conf.py +++ b/docs/conf.py @@ -72,7 +72,7 @@ # List of patterns, relative to source directory, that match files and # directories to ignore when looking for source files. # This pattern also affects html_static_path and html_extra_path. -exclude_patterns = ['_build', 'Thumbs.db', '.DS_Store'] +exclude_patterns = ['_build', 'Thumbs.db', '.DS_Store', '**/*.rs'] # The name of the Pygments (syntax highlighting) style to use. pygments_style = None diff --git a/docs/index.rst b/docs/index.rst index f9b32f30..6e4b189b 100644 --- a/docs/index.rst +++ b/docs/index.rst @@ -19,23 +19,7 @@ Zenoh API Reference module zenoh ============ .. automodule:: zenoh - :members: init_logger, config_from_file, open, async_open, scout, async_scout - - -AsyncQueryable --------------- -.. autoclass:: zenoh.AsyncQueryable - :members: - -AsyncSession ------------- -.. autoclass:: zenoh.AsyncSession - :members: - -AsyncSubscriber ---------------- -.. autoclass:: zenoh.AsyncSubscriber - :members: + :members: init_logger, open, scout Config ------ @@ -48,11 +32,11 @@ CongestionControl :members: :undoc-members: -ConsolidationMode ------------------ -.. autoclass:: zenoh.ConsolidationMode - :members: - :undoc-members: +.. ConsolidationMode +.. ----------------- +.. .. autoclass:: zenoh.ConsolidationMode +.. :members: +.. :undoc-members: Encoding -------- @@ -70,22 +54,10 @@ KeyExpr .. autoclass:: zenoh.KeyExpr :members: -KnownEncoding -------------- -.. autoclass:: zenoh.KnownEncoding - :members: - :undoc-members: - -PeerId ------- -.. autoclass:: zenoh.PeerId - :members: - -Period ------- -.. autoclass:: zenoh.Period +ZenohId +------- +.. autoclass:: zenoh.ZenohId :members: - :special-members: __init__ Priority -------- @@ -145,27 +117,20 @@ Session .. autoclass:: zenoh.Session :members: -SourceInfo +Subscriber ---------- -.. autoclass:: zenoh.SourceInfo +.. autoclass:: zenoh.Subscriber :members: -SubMode -------- -.. autoclass:: zenoh.SubMode +PullSubscriber +-------------- +.. autoclass:: zenoh.PullSubscriber :members: - :undoc-members: -Subscriber +Publisher ---------- -.. autoclass:: zenoh.Subscriber - :members: - -Target ------- -.. autoclass:: zenoh.Target +.. autoclass:: zenoh.Publisher :members: - :undoc-members: Timestamp --------- @@ -177,41 +142,7 @@ Value .. autoclass:: zenoh.Value :members: -ValueSelector -------------- -.. autoclass:: zenoh.ValueSelector - :members: - -WhatAmI -------- -.. autoclass:: zenoh.WhatAmI - :members: - :undoc-members: - -ZError ------- -.. autoexception:: zenoh.ZError - :members: - - -*********** -submodules -*********** - -module zenoh.config -=================== -.. autoclass:: zenoh.config - :members: - :undoc-members: - -module zenoh.info -================= -.. autoclass:: zenoh.info - :members: - :undoc-members: - -module zenoh.queryable -====================== -.. autoclass:: zenoh.queryable +Info +----- +.. autoclass:: zenoh.Info :members: - :undoc-members: diff --git a/examples/README.md b/examples/README.md index d93b6876..7a5a2481 100644 --- a/examples/README.md +++ b/examples/README.md @@ -97,7 +97,7 @@ ### z_get Sends a query message for a selector. - The queryables with a matching path or selector (for instance [z_eval](#z_eval) and [z_storage](#z_storage)) + The queryables with a matching path or selector (for instance [z_queryable](#z_queryable) and [z_storage](#z_storage)) will receive this query and reply with paths/values that will be received by the query callback. Typical usage: @@ -109,7 +109,7 @@ python3 z_get.py -s '/demo/**' ``` -### z_eval +### z_queryable Creates a queryable function with a key expression. This queryable function will be triggered by each call to a get operation on zenoh @@ -117,11 +117,11 @@ Typical usage: ```bash - python3 z_eval.py + python3 z_queryable.py ``` or ```bash - python3 z_eval.py -k /demo/example/eval -v 'This is the result' + python3 z_queryable.py -k /demo/example/queryable -v 'This is the result' ``` ### z_storage @@ -161,4 +161,4 @@ In `asyncio` directory there are similar examples than thse described above, but leveraging Python's [asyncio](https://docs.python.org/fr/3/library/asyncio.html). -Especially, the `asyncio/z_get_parallel.py` and `asyncio/z_eval.py` examples show how a zenoh application can issue concurent queries and how another zenoh application can concurrently compute replies to those queries. +Especially, the `asyncio/z_get_parallel.py` and `asyncio/z_queryable.py` examples show how a zenoh application can issue concurent queries and how another zenoh application can concurrently compute replies to those queries. \ No newline at end of file diff --git a/examples/asyncio/z_delete.py b/examples/asyncio/z_delete.py deleted file mode 100644 index 46ca5cd5..00000000 --- a/examples/asyncio/z_delete.py +++ /dev/null @@ -1,76 +0,0 @@ -# -# Copyright (c) 2022 ZettaScale Technology -# -# This program and the accompanying materials are made available under the -# terms of the Eclipse Public License 2.0 which is available at -# http://www.eclipse.org/legal/epl-2.0, or the Apache License, Version 2.0 -# which is available at https://www.apache.org/licenses/LICENSE-2.0. -# -# SPDX-License-Identifier: EPL-2.0 OR Apache-2.0 -# -# Contributors: -# ZettaScale Zenoh Team, <[email protected]> -# - -import asyncio -import sys -import time -import argparse -import json -import zenoh -from zenoh import config - - -async def main(): - # --- Command line argument parsing --- --- --- --- --- --- - parser = argparse.ArgumentParser( - prog='z_put', - description='zenoh put example') - parser.add_argument('--mode', '-m', dest='mode', - choices=['peer', 'client'], - type=str, - help='The zenoh session mode.') - parser.add_argument('--connect', '-e', dest='connect', - metavar='ENDPOINT', - action='append', - type=str, - help='Endpoints to connect to.') - parser.add_argument('--listen', '-l', dest='listen', - metavar='ENDPOINT', - action='append', - type=str, - help='Endpoints to listen on.') - parser.add_argument('--key', '-k', dest='key', - default='/demo/example/zenoh-python-put', - type=str, - help='The key expression matching resources to delete.') - parser.add_argument('--config', '-c', dest='config', - metavar='FILE', - type=str, - help='A configuration file.') - - args = parser.parse_args() - conf = zenoh.config_from_file( - args.config) if args.config is not None else zenoh.Config() - if args.mode is not None: - conf.insert_json5(zenoh.config.MODE_KEY, json.dumps(args.mode)) - if args.connect is not None: - conf.insert_json5(zenoh.config.CONNECT_KEY, json.dumps(args.connect)) - if args.listen is not None: - conf.insert_json5(zenoh.config.LISTEN_KEY, json.dumps(args.listen)) - key = args.key - - # zenoh-net code --- --- --- --- --- --- --- --- --- --- --- - - # initiate logging - zenoh.init_logger() - - print("Openning session...") - session = await zenoh.async_open(conf) - - print("Deleting resources matching '{}'...".format(key)) - await session.delete(key) - - await session.close() - -asyncio.run(main()) diff --git a/examples/asyncio/z_eval.py b/examples/asyncio/z_eval.py deleted file mode 100644 index 7d7b8d3d..00000000 --- a/examples/asyncio/z_eval.py +++ /dev/null @@ -1,112 +0,0 @@ -# -# Copyright (c) 2022 ZettaScale Technology -# -# This program and the accompanying materials are made available under the -# terms of the Eclipse Public License 2.0 which is available at -# http://www.eclipse.org/legal/epl-2.0, or the Apache License, Version 2.0 -# which is available at https://www.apache.org/licenses/LICENSE-2.0. -# -# SPDX-License-Identifier: EPL-2.0 OR Apache-2.0 -# -# Contributors: -# ZettaScale Zenoh Team, <[email protected]> -# - -import asyncio -import sys -import time -import argparse -import json -import zenoh -from zenoh import config, Sample -from zenoh.queryable import EVAL - - -async def main(): - # --- Command line argument parsing --- --- --- --- --- --- - parser = argparse.ArgumentParser( - prog='z_eval', - description='zenoh eval example') - parser.add_argument('--mode', '-m', dest='mode', - choices=['peer', 'client'], - type=str, - help='The zenoh session mode.') - parser.add_argument('--connect', '-e', dest='connect', - metavar='ENDPOINT', - action='append', - type=str, - help='Endpoints to connect to.') - parser.add_argument('--listen', '-l', dest='listen', - metavar='ENDPOINT', - action='append', - type=str, - help='Endpoints to listen on.') - parser.add_argument('--key', '-k', dest='key', - default='/demo/example/zenoh-python-eval', - type=str, - help='The key expression matching queries to evaluate.') - parser.add_argument('--value', '-v', dest='value', - default='Eval from Python!', - type=str, - help='The value to reply to queries.') - parser.add_argument('--config', '-c', dest='config', - metavar='FILE', - type=str, - help='A configuration file.') - - args = parser.parse_args() - conf = zenoh.config_from_file( - args.config) if args.config is not None else zenoh.Config() - if args.mode is not None: - conf.insert_json5(zenoh.config.MODE_KEY, json.dumps(args.mode)) - if args.connect is not None: - conf.insert_json5(zenoh.config.CONNECT_KEY, json.dumps(args.connect)) - if args.listen is not None: - conf.insert_json5(zenoh.config.LISTEN_KEY, json.dumps(args.listen)) - key = args.key - value = args.value - - # zenoh-net code --- --- --- --- --- --- --- --- --- --- --- - - # Note: As an example the concrete implementation of the eval callback is implemented here as a coroutine. - # It checks if the query's value_selector (the substring after '?') is a float, and if yes, sleeps for this number of seconds. - # Run example/asyncio/z_get_parallel.py example to see how 3 concurrent get() are executed in parallel in this z_eval.py - async def eval_corouting(query): - selector = query.selector - try: - sleep_time = selector.parse_value_selector().properties.get('sleep') - if sleep_time is not None: - print(" Sleeping {} secs before replying".format( - float(sleep_time))) - await asyncio.sleep(float(sleep_time)) - except Exception as e: - print(" WARN: error in value selector: {}. Ignore it.".format(e)) - print(" Replying to query on {}".format(selector)) - reply = "{} (this is the reply to query on {})".format(value, selector) - query.reply(Sample(key_expr=key, payload=reply.encode())) - - async def eval_callback(query): - print(">> [Queryable ] Received Query '{}'".format(query.selector)) - # schedule a task that will call eval_corouting(query) - asyncio.create_task(eval_corouting(query)) - - # initiate logging - zenoh.init_logger() - - print("Openning session...") - session = await zenoh.async_open(conf) - - print("Creating Queryable on '{}'...".format(key)) - queryable = await session.queryable(key, eval_callback, kind=EVAL) - - print("Enter 'q' to quit......") - c = '\0' - while c != 'q': - c = sys.stdin.read(1) - if c == '': - time.sleep(1) - - await queryable.close() - await session.close() - -asyncio.run(main()) diff --git a/examples/asyncio/z_get.py b/examples/asyncio/z_get.py deleted file mode 100644 index 412f9b52..00000000 --- a/examples/asyncio/z_get.py +++ /dev/null @@ -1,99 +0,0 @@ -# -# Copyright (c) 2022 ZettaScale Technology -# -# This program and the accompanying materials are made available under the -# terms of the Eclipse Public License 2.0 which is available at -# http://www.eclipse.org/legal/epl-2.0, or the Apache License, Version 2.0 -# which is available at https://www.apache.org/licenses/LICENSE-2.0. -# -# SPDX-License-Identifier: EPL-2.0 OR Apache-2.0 -# -# Contributors: -# ZettaScale Zenoh Team, <[email protected]> -# - -import asyncio -import sys -import time -import argparse -import json -import zenoh -from zenoh import config, queryable, QueryTarget, Target - - -async def main(): - # --- Command line argument parsing --- --- --- --- --- --- - parser = argparse.ArgumentParser( - prog='z_get', - description='zenoh get example') - parser.add_argument('--mode', '-m', dest='mode', - choices=['peer', 'client'], - type=str, - help='The zenoh session mode.') - parser.add_argument('--connect', '-e', dest='connect', - metavar='ENDPOINT', - action='append', - type=str, - help='Endpoints to connect to.') - parser.add_argument('--listen', '-l', dest='listen', - metavar='ENDPOINT', - action='append', - type=str, - help='Endpoints to listen on.') - parser.add_argument('--selector', '-s', dest='selector', - default='/demo/example/**', - type=str, - help='The selection of resources to query.') - parser.add_argument('--kind', '-k', dest='kind', - choices=['ALL_KINDS', 'STORAGE', 'EVAL'], - default='ALL_KINDS', - type=str, - help='The KIND of queryables to query.') - parser.add_argument('--target', '-t', dest='target', - choices=['ALL', 'BEST_MATCHING', - 'ALL_COMPLETE', 'NONE'], - default='ALL', - type=str, - help='The target queryables of the query.') - parser.add_argument('--config', '-c', dest='config', - metavar='FILE', - type=str, - help='A configuration file.') - - args = parser.parse_args() - conf = zenoh.config_from_file( - args.config) if args.config is not None else zenoh.Config() - if args.mode is not None: - conf.insert_json5(zenoh.config.MODE_KEY, json.dumps(args.mode)) - if args.connect is not None: - conf.insert_json5(zenoh.config.CONNECT_KEY, json.dumps(args.connect)) - if args.listen is not None: - conf.insert_json5(zenoh.config.LISTEN_KEY, json.dumps(args.listen)) - selector = args.selector - kind = { - 'ALL_KINDS': queryable.ALL_KINDS, - 'STORAGE': queryable.STORAGE, - 'EVAL': queryable.EVAL}.get(args.kind) - target = { - 'ALL': Target.All(), - 'BEST_MATCHING': Target.BestMatching(), - 'ALL_COMPLETE': Target.AllComplete(), - 'NONE': Target.No()}.get(args.target) - - # zenoh-net code --- --- --- --- --- --- --- --- --- --- --- - - # initiate logging - zenoh.init_logger() - - print("Openning session...") - session = await zenoh.async_open(conf) - - print("Sending Query '{}'...".format(selector)) - replies = await session.get(selector, target=QueryTarget(kind, target)) - for reply in replies: - print(">> Received ('{}': '{}')" - .format(reply.sample.key_expr, reply.sample.payload.decode("utf-8"))) - - await session.close() - -asyncio.run(main()) diff --git a/examples/asyncio/z_get_parallel.py b/examples/asyncio/z_get_parallel.py deleted file mode 100644 index 8b0b33e9..00000000 --- a/examples/asyncio/z_get_parallel.py +++ /dev/null @@ -1,109 +0,0 @@ -# -# Copyright (c) 2022 ZettaScale Technology -# -# This program and the accompanying materials are made available under the -# terms of the Eclipse Public License 2.0 which is available at -# http://www.eclipse.org/legal/epl-2.0, or the Apache License, Version 2.0 -# which is available at https://www.apache.org/licenses/LICENSE-2.0. -# -# SPDX-License-Identifier: EPL-2.0 OR Apache-2.0 -# -# Contributors: -# ZettaScale Zenoh Team, <[email protected]> -# - -import asyncio -import sys -import time -import argparse -import json -import zenoh -from zenoh import config, queryable, QueryTarget, Target - - -async def main(): - # --- Command line argument parsing --- --- --- --- --- --- - parser = argparse.ArgumentParser( - prog='z_get_parallel', - description='zenoh parallel get example') - parser.add_argument('--mode', '-m', dest='mode', - choices=['peer', 'client'], - type=str, - help='The zenoh session mode.') - parser.add_argument('--connect', '-e', dest='connect', - metavar='ENDPOINT', - action='append', - type=str, - help='Endpoints to connect to.') - parser.add_argument('--listen', '-l', dest='listen', - metavar='ENDPOINT', - action='append', - type=str, - help='Endpoints to listen on.') - parser.add_argument('--selector', '-s', dest='selector', - default='/demo/example/**', - type=str, - help='The selection of resources to query.') - parser.add_argument('--kind', '-k', dest='kind', - choices=['ALL_KINDS', 'STORAGE', 'EVAL'], - default='ALL_KINDS', - type=str, - help='The KIND of queryables to query.') - parser.add_argument('--target', '-t', dest='target', - choices=['ALL', 'BEST_MATCHING', - 'ALL_COMPLETE', 'NONE'], - default='ALL', - type=str, - help='The target queryables of the query.') - parser.add_argument('--config', '-c', dest='config', - metavar='FILE', - type=str, - help='A configuration file.') - - args = parser.parse_args() - conf = zenoh.config_from_file( - args.config) if args.config is not None else zenoh.Config() - if args.mode is not None: - conf.insert_json5(zenoh.config.MODE_KEY, json.dumps(args.mode)) - if args.connect is not None: - conf.insert_json5(zenoh.config.CONNECT_KEY, json.dumps(args.connect)) - if args.listen is not None: - conf.insert_json5(zenoh.config.LISTEN_KEY, json.dumps(args.listen)) - selector = args.selector - kind = { - 'ALL_KINDS': queryable.ALL_KINDS, - 'STORAGE': queryable.STORAGE, - 'EVAL': queryable.EVAL}.get(args.kind) - target = { - 'ALL': Target.All(), - 'BEST_MATCHING': Target.BestMatching(), - 'ALL_COMPLETE': Target.AllComplete(), - 'NONE': Target.No()}.get(args.target) - - # zenoh-net code --- --- --- --- --- --- --- --- --- --- --- - - # initiate logging - zenoh.init_logger() - - print("Openning session...") - session = await zenoh.async_open(conf) - - async def do_query(sleep_time): - print("Sending Query '{}?(sleep={})'...".format(selector, sleep_time)) - replies = await session.get("{}?(sleep={})".format(selector, sleep_time), target=QueryTarget(kind, target)) - for reply in replies: - print(">> Received ('{}': '{}')" - .format(reply.sample.key_expr, reply.sample.payload.decode("utf-8"))) - - start = time.time() - await asyncio.gather( - asyncio.create_task(do_query(1)), - asyncio.create_task(do_query(2)), - asyncio.create_task(do_query(3)), - ) - end = time.time() - print(f'Time: {end-start:.2f} sec') - - await session.close() - -asyncio.run(main()) diff --git a/examples/asyncio/z_info.py b/examples/asyncio/z_info.py deleted file mode 100644 index c828e7a4..00000000 --- a/examples/asyncio/z_info.py +++ /dev/null @@ -1,71 +0,0 @@ -# -# Copyright (c) 2022 ZettaScale Technology -# -# This program and the accompanying materials are made available under the -# terms of the Eclipse Public License 2.0 which is available at -# http://www.eclipse.org/legal/epl-2.0, or the Apache License, Version 2.0 -# which is available at https://www.apache.org/licenses/LICENSE-2.0. -# -# SPDX-License-Identifier: EPL-2.0 OR Apache-2.0 -# -# Contributors: -# ZettaScale Zenoh Team, <[email protected]> -# - -import asyncio -import sys -import time -import argparse -import json -import zenoh -from zenoh import config - - -async def main(): - # --- Command line argument parsing --- --- --- --- --- --- - parser = argparse.ArgumentParser( - prog='z_info', - description='zenoh info example') - parser.add_argument('--mode', '-m', dest='mode', - choices=['peer', 'client'], - type=str, - help='The zenoh session mode.') - parser.add_argument('--connect', '-e', dest='connect', - metavar='ENDPOINT', - action='append', - type=str, - help='Endpoints to connect to.') - parser.add_argument('--listen', '-l', dest='listen', - metavar='ENDPOINT', - action='append', - type=str, - help='Endpoints to listen on.') - parser.add_argument('--config', '-c', dest='config', - metavar='FILE', - type=str, - help='A configuration file.') - - args = parser.parse_args() - conf = zenoh.config_from_file( - args.config) if args.config is not None else zenoh.Config() - if args.mode is not None: - conf.insert_json5(zenoh.config.MODE_KEY, json.dumps(args.mode)) - if args.connect is not None: - conf.insert_json5(zenoh.config.CONNECT_KEY, json.dumps(args.connect)) - if args.listen is not None: - conf.insert_json5(zenoh.config.LISTEN_KEY, json.dumps(args.listen)) - # zenoh-net code --- --- --- --- --- --- --- --- --- --- --- - - # initiate logging - zenoh.init_logger() - - print("Openning session...") - session = await zenoh.async_open(conf) - - info = await session.info() - for key in info: - print("{} : {}".format(key, info[key])) - - await session.close() - -asyncio.run(main()) diff --git a/examples/asyncio/z_pub.py b/examples/asyncio/z_pub.py deleted file mode 100644 index 071d1738..00000000 --- a/examples/asyncio/z_pub.py +++ /dev/null @@ -1,96 +0,0 @@ -# -# Copyright (c) 2022 ZettaScale Technology -# -# This program and the accompanying materials are made available under the -# terms of the Eclipse Public License 2.0 which is available at -# http://www.eclipse.org/legal/epl-2.0, or the Apache License, Version 2.0 -# which is available at https://www.apache.org/licenses/LICENSE-2.0. -# -# SPDX-License-Identifier: EPL-2.0 OR Apache-2.0 -# -# Contributors: -# ZettaScale Zenoh Team, <[email protected]> -# - -import asyncio -import sys -import time -import argparse -import itertools -import json -import zenoh -from zenoh import config - - -async def main(): - # --- Command line argument parsing --- --- --- --- --- --- - parser = argparse.ArgumentParser( - prog='z_pub', - description='zenoh pub example') - parser.add_argument('--mode', '-m', dest='mode', - choices=['peer', 'client'], - type=str, - help='The zenoh session mode.') - parser.add_argument('--connect', '-e', dest='connect', - metavar='ENDPOINT', - action='append', - type=str, - help='Endpoints to connect to.') - parser.add_argument('--listen', '-l', dest='listen', - metavar='ENDPOINT', - action='append', - type=str, - help='Endpoints to listen on.') - parser.add_argument('--key', '-k', dest='key', - default='/demo/example/zenoh-python-pub', - type=str, - help='The key expression to publish onto.') - parser.add_argument('--value', '-v', dest='value', - default='Pub from Python!', - type=str, - help='The value to publish.') - parser.add_argument("--iter", dest="iter", type=int, - help="How many puts to perform") - parser.add_argument('--config', '-c', dest='config', - metavar='FILE', - type=str, - help='A configuration file.') - - args = parser.parse_args() - conf = zenoh.config_from_file( - args.config) if args.config is not None else zenoh.Config() - if args.mode is not None: - conf.insert_json5(zenoh.config.MODE_KEY, json.dumps(args.mode)) - if args.connect is not None: - conf.insert_json5(zenoh.config.CONNECT_KEY, json.dumps(args.connect)) - if args.listen is not None: - conf.insert_json5(zenoh.config.LISTEN_KEY, json.dumps(args.listen)) - key = args.key - value = args.value - - # zenoh-net code --- --- --- --- --- --- --- --- --- --- --- - - # initiate logging - zenoh.init_logger() - - print("Openning session...") - session = await zenoh.async_open(conf) - - print("Declaring key expression '{}'...".format(key), end='') - rid = await session.declare_expr(key) - print(" => RId {}".format(rid)) - - print("Declaring publication on '{}'...".format(rid)) - await session.declare_publication(rid) - - for idx in itertools.count() if args.iter is None else range(args.iter): - time.sleep(1) - buf = "[{:4d}] {}".format(idx, value) - print("Putting Data ('{}': '{}')...".format(rid, buf)) - await session.put(rid, bytes(buf, encoding='utf8')) - - await session.undeclare_publication(rid) - await session.undeclare_expr(rid) - await session.close() - -asyncio.run(main()) diff --git a/examples/asyncio/z_put.py b/examples/asyncio/z_put.py deleted file mode 100644 index d31e8f0d..00000000 --- a/examples/asyncio/z_put.py +++ /dev/null @@ -1,81 +0,0 @@ -# -# Copyright (c) 2022 ZettaScale Technology -# -# This program and the accompanying materials are made available under the -# terms of the Eclipse Public License 2.0 which is available at -# http://www.eclipse.org/legal/epl-2.0, or the Apache License, Version 2.0 -# which is available at https://www.apache.org/licenses/LICENSE-2.0. -# -# SPDX-License-Identifier: EPL-2.0 OR Apache-2.0 -# -# Contributors: -# ZettaScale Zenoh Team, <[email protected]> -# - -import asyncio -import sys -import time -import argparse -import json -import zenoh -from zenoh import config - - -async def main(): - # --- Command line argument parsing --- --- --- --- --- --- - parser = argparse.ArgumentParser( - prog='z_put', - description='zenoh put example') - parser.add_argument('--mode', '-m', dest='mode', - choices=['peer', 'client'], - type=str, - help='The zenoh session mode.') - parser.add_argument('--connect', '-e', dest='connect', - metavar='ENDPOINT', - action='append', - type=str, - help='Endpoints to connect to.') - parser.add_argument('--listen', '-l', dest='listen', - metavar='ENDPOINT', - action='append', - type=str, - help='Endpoints to listen on.') - parser.add_argument('--key', '-k', dest='key', - default='/demo/example/zenoh-python-put', - type=str, - help='The key expression to write.') - parser.add_argument('--value', '-v', dest='value', - default='Put from Python!', - type=str, - help='The value to write.') - parser.add_argument('--config', '-c', dest='config', - metavar='FILE', - type=str, - help='A configuration file.') - - args = parser.parse_args() - conf = zenoh.config_from_file( - args.config) if args.config is not None else zenoh.Config() - if args.mode is not None: - conf.insert_json5(zenoh.config.MODE_KEY, json.dumps(args.mode)) - if args.connect is not None: - conf.insert_json5(zenoh.config.CONNECT_KEY, json.dumps(args.connect)) - if args.listen is not None: - conf.insert_json5(zenoh.config.LISTEN_KEY, json.dumps(args.listen)) - key = args.key - value = args.value - - # zenoh-net code --- --- --- --- --- --- --- --- --- --- --- - - # initiate logging - zenoh.init_logger() - - print("Openning session...") - session = await zenoh.async_open(conf) - - print("Putting Data ('{}': '{}')...".format(key, value)) - await session.put(key, value) - - await session.close() - -asyncio.run(main()) diff --git a/examples/asyncio/z_scout.py b/examples/asyncio/z_scout.py deleted file mode 100644 index 4def468b..00000000 --- a/examples/asyncio/z_scout.py +++ /dev/null @@ -1,33 +0,0 @@ -# -# Copyright (c) 2022 ZettaScale Technology -# -# This program and the accompanying materials are made available under the -# terms of the Eclipse Public License 2.0 which is available at -# http://www.eclipse.org/legal/epl-2.0, or the Apache License, Version 2.0 -# which is available at https://www.apache.org/licenses/LICENSE-2.0. -# -# SPDX-License-Identifier: EPL-2.0 OR Apache-2.0 -# -# Contributors: -# ZettaScale Zenoh Team, <[email protected]> -# - -import asyncio -import sys -import time -import argparse -import zenoh -from zenoh import WhatAmI - - -async def main(): - # initiate logging - zenoh.init_logger() - - print("Scouting...") - hellos = await zenoh.async_scout(WhatAmI.Peer | WhatAmI.Router, 1.0) - - for hello in hellos: - print(hello) - -asyncio.run(main()) diff --git a/examples/asyncio/z_sub.py b/examples/asyncio/z_sub.py deleted file mode 100644 index 006c136c..00000000 --- a/examples/asyncio/z_sub.py +++ /dev/null @@ -1,91 +0,0 @@ -# -# Copyright (c) 2022 ZettaScale Technology -# -# This program and the accompanying materials are made available under the -# terms of the Eclipse Public License 2.0 which is available at -# http://www.eclipse.org/legal/epl-2.0, or the Apache License, Version 2.0 -# which is available at https://www.apache.org/licenses/LICENSE-2.0. -# -# SPDX-License-Identifier: EPL-2.0 OR Apache-2.0 -# -# Contributors: -# ZettaScale Zenoh Team, <[email protected]> -# - -import asyncio -import sys -import time -from datetime import datetime -import argparse -import json -import zenoh -from zenoh import Reliability, SubMode - - -async def main(): - # --- Command line argument parsing --- --- --- --- --- --- - parser = argparse.ArgumentParser( - prog='z_sub', - description='zenoh sub example') - parser.add_argument('--mode', '-m', dest='mode', - choices=['peer', 'client'], - type=str, - help='The zenoh session mode.') - parser.add_argument('--connect', '-e', dest='connect', - metavar='ENDPOINT', - action='append', - type=str, - help='Endpoints to connect to.') - parser.add_argument('--listen', '-l', dest='listen', - metavar='ENDPOINT', - action='append', - type=str, - help='Endpoints to listen on.') - parser.add_argument('--key', '-k', dest='key', - default='/demo/example/**', - type=str, - help='The key expression to subscribe to.') - parser.add_argument('--config', '-c', dest='config', - metavar='FILE', - type=str, - help='A configuration file.') - - args = parser.parse_args() - conf = zenoh.config_from_file( - args.config) if args.config is not None else zenoh.Config() - if args.mode is not None: - conf.insert_json5(zenoh.config.MODE_KEY, json.dumps(args.mode)) - if args.connect is not None: - conf.insert_json5(zenoh.config.CONNECT_KEY, json.dumps(args.connect)) - if args.listen is not None: - conf.insert_json5(zenoh.config.LISTEN_KEY, json.dumps(args.listen)) - key = args.key - - # zenoh-net code --- --- --- --- --- --- --- --- --- --- --- - - async def listener(sample): - print(">> [Subscriber] Received {} ('{}': '{}')" - .format(sample.kind, sample.key_expr, sample.payload.decode("utf-8"))) - - # initiate logging - zenoh.init_logger() - - print("Openning session...") - session = await zenoh.async_open(conf) - - print("Creating Subscriber on '{}'...".format(key)) - - sub = await session.subscribe(key, listener, reliability=Reliability.Reliable, mode=SubMode.Push) - - print("Enter 'q' to quit...") - c = '\0' - while c != 'q': - c = sys.stdin.read(1) - if c == '': - time.sleep(1) - - await sub.close() - await session.close() - - -asyncio.run(main()) diff --git a/examples/z_delete.py b/examples/z_delete.py index 9839123c..1be4e87a 100644 --- a/examples/z_delete.py +++ b/examples/z_delete.py @@ -38,7 +38,7 @@ type=str, help='Endpoints to listen on.') parser.add_argument('--key', '-k', dest='key', - default='/demo/example/zenoh-python-put', + default='demo/example/zenoh-python-put', type=str, help='The key expression matching resources to delete.') parser.add_argument('--config', '-c', dest='config', @@ -47,8 +47,7 @@ help='A configuration file.') args = parser.parse_args() -conf = zenoh.config_from_file( - args.config) if args.config is not None else zenoh.Config() +conf = zenoh.Config.from_file(args.config) if args.config is not None else zenoh.Config() if args.mode is not None: conf.insert_json5(zenoh.config.MODE_KEY, json.dumps(args.mode)) if args.connect is not None: diff --git a/examples/z_get.py b/examples/z_get.py index d62cf3af..c38a9413 100644 --- a/examples/z_get.py +++ b/examples/z_get.py @@ -17,7 +17,7 @@ import argparse import json import zenoh -from zenoh import config, queryable, QueryTarget, Target +from zenoh import config, QueryTarget # --- Command line argument parsing --- --- --- --- --- --- parser = argparse.ArgumentParser( @@ -38,14 +38,9 @@ type=str, help='Endpoints to listen on.') parser.add_argument('--selector', '-s', dest='selector', - default='/demo/example/**', + default='demo/example/**', type=str, help='The selection of resources to query.') -parser.add_argument('--kind', '-k', dest='kind', - choices=['ALL_KINDS', 'STORAGE', 'EVAL'], - default='ALL_KINDS', - type=str, - help='The KIND of queryables to query.') parser.add_argument('--target', '-t', dest='target', choices=['ALL', 'BEST_MATCHING', 'ALL_COMPLETE', 'NONE'], default='ALL', @@ -66,15 +61,10 @@ if args.listen is not None: conf.insert_json5(zenoh.config.LISTEN_KEY, json.dumps(args.listen)) selector = args.selector -kind = { - 'ALL_KINDS': queryable.ALL_KINDS, - 'STORAGE': queryable.STORAGE, - 'EVAL': queryable.EVAL}.get(args.kind) target = { - 'ALL': Target.All(), - 'BEST_MATCHING': Target.BestMatching(), - 'ALL_COMPLETE': Target.AllComplete(), - 'NONE': Target.No()}.get(args.target) + 'ALL': QueryTarget.ALL(), + 'BEST_MATCHING': QueryTarget.BEST_MATCHING(), + 'ALL_COMPLETE': QueryTarget.ALL_COMPLETE(),}.get(args.target) # zenoh-net code --- --- --- --- --- --- --- --- --- --- --- @@ -85,9 +75,14 @@ session = zenoh.open(conf) print("Sending Query '{}'...".format(selector)) -replies = session.get(selector, target=QueryTarget(kind, target)) -for reply in replies: - print(">> Received ('{}': '{}')" - .format(reply.sample.key_expr, reply.sample.payload.decode("utf-8"))) +replies = session.get(selector, zenoh.ListCollector(), target=target) +for reply in replies(): + try: + print(">> Received ('{}': '{}')" + .format(reply.ok.key_expr, reply.ok.payload.decode("utf-8"))) + except: + print(">> Received (ERROR: '{}')" + .format(reply.err.payload.decode("utf-8"))) + session.close() diff --git a/examples/z_info.py b/examples/z_info.py index 9083b1b1..58f4b300 100644 --- a/examples/z_info.py +++ b/examples/z_info.py @@ -17,7 +17,6 @@ import argparse import json import zenoh -from zenoh import config # --- Command line argument parsing --- --- --- --- --- --- parser = argparse.ArgumentParser( @@ -60,7 +59,7 @@ session = zenoh.open(conf) info = session.info() -for key in info: - print("{} : {}".format(key, info[key])) - +print(f"zid: {info.zid()}") +print(f"routers: {info.routers_zid()}") +print(f"peers: {info.peers_zid()}") session.close() diff --git a/examples/z_pub.py b/examples/z_pub.py index d26c790c..377ffd66 100644 --- a/examples/z_pub.py +++ b/examples/z_pub.py @@ -39,7 +39,7 @@ type=str, help='Endpoints to listen on.') parser.add_argument('--key', '-k', dest='key', - default='/demo/example/zenoh-python-pub', + default='demo/example/zenoh-python-pub', type=str, help='The key expression to publish onto.') parser.add_argument('--value', '-v', dest='value', @@ -54,8 +54,7 @@ help='A configuration file.') args = parser.parse_args() -conf = zenoh.config_from_file( - args.config) if args.config is not None else zenoh.Config() +conf = zenoh.Config.from_file(args.config) if args.config is not None else zenoh.Config() if args.mode is not None: conf.insert_json5(zenoh.config.MODE_KEY, json.dumps(args.mode)) if args.connect is not None: @@ -65,27 +64,20 @@ key = args.key value = args.value -# zenoh-net code --- --- --- --- --- --- --- --- --- --- --- - # initiate logging zenoh.init_logger() print("Openning session...") session = zenoh.open(conf) -print("Declaring key expression '{}'...".format(key), end='') -rid = session.declare_expr(key) -print(" => RId {}".format(rid)) - -print("Declaring publication on '{}'...".format(rid)) -session.declare_publication(rid) +print(f"Declaring publication on '{key}'...") +pub = session.declare_publisher(key) for idx in itertools.count() if args.iter is None else range(args.iter): time.sleep(1) - buf = "[{:4d}] {}".format(idx, value) - print("Putting Data ('{}': '{}')...".format(rid, buf)) - session.put(rid, buf) + buf = f"[{idx:4d}] {value}" + print(f"Putting Data ('{key}': '{buf}')...") + pub.put(buf) -session.undeclare_publication(rid) -session.undeclare_expr(rid) +pub.undeclare() session.close() diff --git a/examples/z_pub_thr.py b/examples/z_pub_thr.py index a40d6fb9..f29008e0 100644 --- a/examples/z_pub_thr.py +++ b/examples/z_pub_thr.py @@ -17,7 +17,7 @@ import argparse import json import zenoh -from zenoh import config, CongestionControl +from zenoh import config, CongestionControl, Value # --- Command line argument parsing --- --- --- --- --- --- parser = argparse.ArgumentParser( @@ -46,8 +46,7 @@ help='A configuration file.') args = parser.parse_args() -conf = zenoh.config_from_file( - args.config) if args.config is not None else zenoh.Config() +conf = zenoh.Config.from_file(args.config) if args.config is not None else zenoh.Config() if args.mode is not None: conf.insert_json5(zenoh.config.MODE_KEY, json.dumps(args.mode)) if args.connect is not None: @@ -64,14 +63,11 @@ data = bytearray() for i in range(0, size): data.append(i % 10) -data = bytes(data) -congestion_control = CongestionControl.Block +data = Value(bytes(data)) +congestion_control = CongestionControl.BLOCK() session = zenoh.open(conf) - -rid = session.declare_expr('/test/thr') - -pub = session.declare_publication(rid) +pub = session.declare_publisher('test/thr', congestion_control=congestion_control) while True: - session.put(rid, data, congestion_control=congestion_control) + pub.put(data) diff --git a/examples/z_pull.py b/examples/z_pull.py index 856bbdda..48161b06 100644 --- a/examples/z_pull.py +++ b/examples/z_pull.py @@ -18,7 +18,7 @@ import argparse import json import zenoh -from zenoh import Reliability, SubMode +from zenoh import Reliability # --- Command line argument parsing --- --- --- --- --- --- parser = argparse.ArgumentParser( @@ -39,7 +39,7 @@ type=str, help='Endpoints to listen on.') parser.add_argument('--key', '-k', dest='key', - default='/demo/example/**', + default='demo/example/**', type=str, help='The key expression matching resources to pull.') parser.add_argument('--config', '-c', dest='config', @@ -62,10 +62,7 @@ def listen(sample): - time = '(not specified)' if sample.source_info is None or sample.timestamp is None else datetime.fromtimestamp( - sample.timestamp.time) - print(">> [Subscriber] Received {} ('{}': '{}')" - .format(sample.kind, sample.key_expr, sample.payload.decode("utf-8"), time)) + print(f">> [Subscriber] Received {sample.kind} ('{sample.key_expr}': '{sample.payload.decode('utf-8')}')") # initiate logging @@ -76,8 +73,7 @@ def listen(sample): print("Creating Subscriber on '{}'...".format(key)) -sub = session.subscribe( - key, listen, reliability=Reliability.Reliable, mode=SubMode.Pull) +sub = session.declare_pull_subscriber(key, listen, reliability=Reliability.RELIABLE()) print("Press <enter> to pull data...") c = '\0' @@ -88,5 +84,5 @@ def listen(sample): else: sub.pull() -sub.close() +sub.undeclare() session.close() diff --git a/examples/z_put.py b/examples/z_put.py index de83001c..68c24f38 100644 --- a/examples/z_put.py +++ b/examples/z_put.py @@ -38,7 +38,7 @@ type=str, help='Endpoints to listen on.') parser.add_argument('--key', '-k', dest='key', - default='/demo/example/zenoh-python-put', + default='demo/example/zenoh-python-put', type=str, help='The key expression to write.') parser.add_argument('--value', '-v', dest='value', diff --git a/examples/z_eval.py b/examples/z_queryable.py similarity index 87% rename from examples/z_eval.py rename to examples/z_queryable.py index cdfe1ce5..62bf85ca 100644 --- a/examples/z_eval.py +++ b/examples/z_queryable.py @@ -18,12 +18,11 @@ import json import zenoh from zenoh import config, Sample -from zenoh.queryable import EVAL # --- Command line argument parsing --- --- --- --- --- --- parser = argparse.ArgumentParser( - prog='z_eval', - description='zenoh eval example') + prog='z_queryable', + description='zenoh queryable example') parser.add_argument('--mode', '-m', dest='mode', choices=['peer', 'client'], type=str, @@ -39,11 +38,11 @@ type=str, help='Endpoints to listen on.') parser.add_argument('--key', '-k', dest='key', - default='/demo/example/zenoh-python-eval', + default='demo/example/zenoh-python-queryable', type=str, - help='The key expression matching queries to evaluate.') + help='The key expression matching queries to reply to.') parser.add_argument('--value', '-v', dest='value', - default='Eval from Python!', + default='Queryable from Python!', type=str, help='The value to reply to queries.') parser.add_argument('--config', '-c', dest='config', @@ -66,9 +65,9 @@ # zenoh-net code --- --- --- --- --- --- --- --- --- --- --- -def eval_callback(query): +def queryable_callback(query): print(">> [Queryable ] Received Query '{}'".format(query.selector)) - query.reply(Sample(key_expr=key, payload=value.encode())) + query.reply(Sample(key, value)) # initiate logging @@ -78,7 +77,7 @@ def eval_callback(query): session = zenoh.open(conf) print("Creating Queryable on '{}'...".format(key)) -queryable = session.queryable(key, eval_callback, kind=EVAL) +queryable = session.declare_queryable(key, queryable_callback) print("Enter 'q' to quit......") c = '\0' @@ -87,5 +86,5 @@ def eval_callback(query): if c == '': time.sleep(1) -queryable.close() +queryable.undeclare() session.close() diff --git a/examples/z_scout.py b/examples/z_scout.py index b910f880..fc2da860 100644 --- a/examples/z_scout.py +++ b/examples/z_scout.py @@ -12,17 +12,16 @@ # ZettaScale Zenoh Team, <[email protected]> # -import sys -import time -import argparse import zenoh -from zenoh import WhatAmI # initiate logging zenoh.init_logger() print("Scouting...") -hellos = zenoh.scout(WhatAmI.Peer | WhatAmI.Router, 1.0) +scout = zenoh.scout(what = "peer|router", timeout=1.0) -for hello in hellos: +def dbg(x): + print(x) + return x +for hello in dbg(scout.receiver()): print(hello) diff --git a/examples/z_storage.py b/examples/z_storage.py index 15ad0a3d..0baa4c9b 100644 --- a/examples/z_storage.py +++ b/examples/z_storage.py @@ -17,8 +17,7 @@ import argparse import json import zenoh -from zenoh import Reliability, SampleKind, SubMode, Sample, KeyExpr -from zenoh.queryable import STORAGE +from zenoh import Reliability, SampleKind, Query, Sample, KeyExpr # --- Command line argument parsing --- --- --- --- --- --- parser = argparse.ArgumentParser( @@ -39,7 +38,7 @@ type=str, help='Endpoints to listen on.') parser.add_argument('--key', '-k', dest='key', - default='/demo/example/**', + default='demo/example/**', type=str, help='The key expression matching resources to store.') parser.add_argument('--config', '-c', dest='config', @@ -63,22 +62,20 @@ store = {} -def listener(sample): +def listener(sample: Sample): print(">> [Subscriber] Received {} ('{}': '{}')" .format(sample.kind, sample.key_expr, sample.payload.decode("utf-8"))) if sample.kind == SampleKind.DELETE: - store.pop(str(sample.key_expr), None) + store.pop(sample.key_expr, None) else: - store[str(sample.key_expr)] = (sample.value, sample.source_info) + store[sample.key_expr] = sample -def query_handler(query): +def query_handler(query: Query): print(">> [Queryable ] Received Query '{}'".format(query.selector)) replies = [] - for stored_name, (data, source_info) in store.items(): - if KeyExpr.intersect(query.key_selector, stored_name): - sample = Sample(stored_name, data) - sample.with_source_info(source_info) + for stored_name, sample in store.items(): + if query.key_expr.intersects(stored_name): query.reply(sample) @@ -89,11 +86,10 @@ def query_handler(query): session = zenoh.open(conf) print("Creating Subscriber on '{}'...".format(key)) -sub = session.subscribe( - key, listener, reliability=Reliability.Reliable, mode=SubMode.Push) +sub = session.declare_subscriber(key, listener, reliability=Reliability.RELIABLE()) print("Creating Queryable on '{}'...".format(key)) -queryable = session.queryable(key, query_handler, kind=STORAGE) +queryable = session.declare_queryable(key, query_handler) print("Enter 'q' to quit......") c = '\0' @@ -102,6 +98,6 @@ def query_handler(query): if c == '': time.sleep(1) -sub.close() -queryable.close() +sub.undeclare() +queryable.undeclare() session.close() diff --git a/examples/z_sub.py b/examples/z_sub.py index 0b87a1cb..d126ea27 100644 --- a/examples/z_sub.py +++ b/examples/z_sub.py @@ -18,7 +18,7 @@ import argparse import json import zenoh -from zenoh import Reliability, SubMode +from zenoh import Reliability, Sample # --- Command line argument parsing --- --- --- --- --- --- parser = argparse.ArgumentParser( @@ -39,7 +39,7 @@ type=str, help='Endpoints to listen on.') parser.add_argument('--key', '-k', dest='key', - default='/demo/example/**', + default='demo/example/**', type=str, help='The key expression to subscribe to.') parser.add_argument('--config', '-c', dest='config', @@ -61,10 +61,6 @@ # zenoh-net code --- --- --- --- --- --- --- --- --- --- --- -def listener(sample): - print(">> [Subscriber] Received {} ('{}': '{}')" - .format(sample.kind, sample.key_expr, sample.payload.decode("utf-8"))) - # initiate logging zenoh.init_logger() @@ -74,8 +70,15 @@ def listener(sample): print("Creating Subscriber on '{}'...".format(key)) -sub = session.subscribe( - key, listener, reliability=Reliability.Reliable, mode=SubMode.Push) + +def listener(sample: Sample): + print(f">> [Subscriber] Received {sample.kind} ('{sample.key_expr}': '{sample.payload.decode('utf-8')}')") + + +# WARNING, you MUST store the return value in order for the subscription to work!! +# This is because if you don't, the reference counter will reach 0 and the subscription +# will be immediately undeclared. +sub = session.declare_subscriber(key, listener, reliability=Reliability.RELIABLE()) print("Enter 'q' to quit...") c = '\0' @@ -84,5 +87,7 @@ def listener(sample): if c == '': time.sleep(1) -sub.close() +# Cleanup: note that even if you forget it, cleanup will happen automatically when +# the reference counter reaches 0 +sub.undeclare() session.close() diff --git a/examples/z_sub_queued.py b/examples/z_sub_queued.py new file mode 100644 index 00000000..98489643 --- /dev/null +++ b/examples/z_sub_queued.py @@ -0,0 +1,97 @@ +# +# Copyright (c) 2022 ZettaScale Technology +# +# This program and the accompanying materials are made available under the +# terms of the Eclipse Public License 2.0 which is available at +# http://www.eclipse.org/legal/epl-2.0, or the Apache License, Version 2.0 +# which is available at https://www.apache.org/licenses/LICENSE-2.0. +# +# SPDX-License-Identifier: EPL-2.0 OR Apache-2.0 +# +# Contributors: +# ZettaScale Zenoh Team, <[email protected]> +# + +import sys +import time +from datetime import datetime +import argparse +import json +import zenoh +from threading import Thread +from zenoh import Reliability, Sample + +# --- Command line argument parsing --- --- --- --- --- --- +parser = argparse.ArgumentParser( + prog='z_sub', + description='zenoh sub example') +parser.add_argument('--mode', '-m', dest='mode', + choices=['peer', 'client'], + type=str, + help='The zenoh session mode.') +parser.add_argument('--connect', '-e', dest='connect', + metavar='ENDPOINT', + action='append', + type=str, + help='Endpoints to connect to.') +parser.add_argument('--listen', '-l', dest='listen', + metavar='ENDPOINT', + action='append', + type=str, + help='Endpoints to listen on.') +parser.add_argument('--key', '-k', dest='key', + default='demo/example/**', + type=str, + help='The key expression to subscribe to.') +parser.add_argument('--config', '-c', dest='config', + metavar='FILE', + type=str, + help='A configuration file.') + +args = parser.parse_args() +conf = zenoh.config_from_file( + args.config) if args.config is not None else zenoh.Config() +if args.mode is not None: + conf.insert_json5(zenoh.config.MODE_KEY, json.dumps(args.mode)) +if args.connect is not None: + conf.insert_json5(zenoh.config.CONNECT_KEY, json.dumps(args.connect)) +if args.listen is not None: + conf.insert_json5(zenoh.config.LISTEN_KEY, json.dumps(args.listen)) +key = args.key + +# zenoh-net code --- --- --- --- --- --- --- --- --- --- --- + + + +# initiate logging +zenoh.init_logger() + +print("Openning session...") +session = zenoh.open(conf) + +print("Creating Subscriber on '{}'...".format(key)) + + +# WARNING, you MUST store the return value in order for the subscription to work!! +# This is because if you don't, the reference counter will reach 0 and the subscription +# will be immediately undeclared. +sub = session.declare_subscriber(key, zenoh.Queue(), reliability=Reliability.RELIABLE()) + +def consumer(): + for sample in sub.receiver: # zenoh.Queue's receiver (the queue itself) is an iterator + print(f">> [Subscriber] Received {sample.kind} ('{sample.key_expr}': '{sample.payload.decode('utf-8')}')") + +t = Thread(target=consumer) +t.start() +print("Enter 'q' to quit...") +c = '\0' +while c != 'q': + c = sys.stdin.read(1) + if c == '': + time.sleep(1) + +# Cleanup: note that even if you forget it, cleanup will happen automatically when +# the reference counter reaches 0 +sub.undeclare() +t.join() +session.close() diff --git a/examples/z_sub_thr.py b/examples/z_sub_thr.py index 86f88050..18a592ef 100644 --- a/examples/z_sub_thr.py +++ b/examples/z_sub_thr.py @@ -14,11 +14,10 @@ import sys import time -import datetime import argparse import json import zenoh -from zenoh import Reliability, SubMode +from zenoh import Reliability # --- Command line argument parsing --- --- --- --- --- --- parser = argparse.ArgumentParser( @@ -38,12 +37,6 @@ action='append', type=str, help='Endpoints to listen on.') -parser.add_argument('--samples', '-s', dest='samples', - default=10, - metavar='NUMBER', - action='append', - type=int, - help='Number of throughput measurements.') parser.add_argument('--number', '-n', dest='number', default=50000, metavar='NUMBER', @@ -56,56 +49,58 @@ help='A configuration file.') args = parser.parse_args() -conf = zenoh.config_from_file( - args.config) if args.config is not None else zenoh.Config() +conf = zenoh.Config.from_file(args.config) if args.config is not None else zenoh.Config() if args.mode is not None: conf.insert_json5(zenoh.config.MODE_KEY, json.dumps(args.mode)) if args.connect is not None: conf.insert_json5(zenoh.config.CONNECT_KEY, json.dumps(args.connect)) if args.listen is not None: conf.insert_json5(zenoh.config.LISTEN_KEY, json.dumps(args.listen)) -m = args.samples n = args.number -# zenoh-net code --- --- --- --- --- --- --- --- --- --- --- - - -def print_stats(start): - stop = datetime.datetime.now() - print("{:.6f} msgs/sec".format(n / (stop - start).total_seconds())) +batch_count = 0 count = 0 start = None -nm = 0 - +global_start = None def listener(sample): - global n, m, count, start, nm + global n, count, batch_count, start, global_start if count == 0: - start = datetime.datetime.now() + start = time.time() + if global_start is None: + global_start = start count += 1 elif count < n: count += 1 else: - print_stats(start) - nm += 1 + stop = time.time() + print(f"{n / (stop - start):.6f} msgs/sec") + batch_count += 1 count = 0 - if nm >= m: - sys.exit(0) +def report(): + global n, m, count, batch_count, global_start + end = time.time() + total = batch_count * n + count + print(f"Received {total} messages in {end - global_start}: averaged {total / (end - global_start):.6f} msgs/sec") # initiate logging zenoh.init_logger() session = zenoh.open(conf) -rid = session.declare_expr('/test/thr') - -sub = session.subscribe( - rid, listener, reliablity=Reliability.Reliable, mode=SubMode.Push) +sub = session.declare_subscriber("test/thr", (listener, report), reliability=Reliability.RELIABLE()) -time.sleep(600) +print("Enter 'q' to quit...") +c = '\0' +while c != 'q': + c = sys.stdin.read(1) + if c == '': + time.sleep(1) -session.undeclare_expr(rid) +sub.undeclare() session.close() +# while `sub.undeclare()` only returns once the unsubscription is done (no more callbacks will be queued from that instant), already queued callbacks may still be running in threads that Python can't see. +time.sleep(0.1) \ No newline at end of file diff --git a/rustup-install.sh b/rustup-install.sh deleted file mode 100644 index a4a79a29..00000000 --- a/rustup-install.sh +++ /dev/null @@ -1,662 +0,0 @@ -#!/bin/sh -# shellcheck shell=dash - -# This is just a little script that can be downloaded from the internet to -# install rustup. It just does platform detection, downloads the installer -# and runs it. - -# It runs on Unix shells like {a,ba,da,k,z}sh. It uses the common `local` -# extension. Note: Most shells limit `local` to 1 var per line, contra bash. - -if [ "$KSH_VERSION" = 'Version JM 93t+ 2010-03-05' ]; then - # The version of ksh93 that ships with many illumos systems does not - # support the "local" extension. Print a message rather than fail in - # subtle ways later on: - echo 'rustup does not work with this ksh93 version; please try bash!' >&2 - exit 1 -fi - - -set -u - -# If RUSTUP_UPDATE_ROOT is unset or empty, default it. -RUSTUP_UPDATE_ROOT="${RUSTUP_UPDATE_ROOT:-https://static.rust-lang.org/rustup}" - -#XXX: If you change anything here, please make the same changes in setup_mode.rs -usage() { - cat 1>&2 <<EOF -rustup-init 1.24.3 (c1c769109 2021-05-31) -The installer for rustup - -USAGE: - rustup-init [FLAGS] [OPTIONS] - -FLAGS: - -v, --verbose Enable verbose output - -q, --quiet Disable progress output - -y Disable confirmation prompt. - --no-modify-path Don't configure the PATH environment variable - -h, --help Prints help information - -V, --version Prints version information - -OPTIONS: - --default-host <default-host> Choose a default host triple - --default-toolchain <default-toolchain> Choose a default toolchain to install - --default-toolchain none Do not install any toolchains - --profile [minimal|default|complete] Choose a profile - -c, --component <components>... Component name to also install - -t, --target <targets>... Target name to also install -EOF -} - -main() { - downloader --check - need_cmd uname - need_cmd mktemp - need_cmd chmod - need_cmd mkdir - need_cmd rm - need_cmd rmdir - - get_architecture || return 1 - local _arch="$RETVAL" - assert_nz "$_arch" "arch" - - local _ext="" - case "$_arch" in - *windows*) - _ext=".exe" - ;; - esac - - local _url="${RUSTUP_UPDATE_ROOT}/dist/${_arch}/rustup-init${_ext}" - - local _dir - _dir="$(ensure mktemp -d)" - local _file="${_dir}/rustup-init${_ext}" - - local _ansi_escapes_are_valid=false - if [ -t 2 ]; then - if [ "${TERM+set}" = 'set' ]; then - case "$TERM" in - xterm*|rxvt*|urxvt*|linux*|vt*) - _ansi_escapes_are_valid=true - ;; - esac - fi - fi - - # check if we have to use /dev/tty to prompt the user - local need_tty=yes - for arg in "$@"; do - case "$arg" in - -h|--help) - usage - exit 0 - ;; - -y) - # user wants to skip the prompt -- we don't need /dev/tty - need_tty=no - ;; - *) - ;; - esac - done - - if $_ansi_escapes_are_valid; then - printf "\33[1minfo:\33[0m downloading installer\n" 1>&2 - else - printf '%s\n' 'info: downloading installer' 1>&2 - fi - - ensure mkdir -p "$_dir" - ensure downloader "$_url" "$_file" "$_arch" - ensure chmod u+x "$_file" - if [ ! -x "$_file" ]; then - printf '%s\n' "Cannot execute $_file (likely because of mounting /tmp as noexec)." 1>&2 - printf '%s\n' "Please copy the file to a location where you can execute binaries and run ./rustup-init${_ext}." 1>&2 - exit 1 - fi - - if [ "$need_tty" = "yes" ]; then - # The installer is going to want to ask for confirmation by - # reading stdin. This script was piped into `sh` though and - # doesn't have stdin to pass to its children. Instead we're going - # to explicitly connect /dev/tty to the installer's stdin. - if [ ! -t 1 ]; then - err "Unable to run interactively. Run with -y to accept defaults, --help for additional options" - fi - - ignore "$_file" "$@" < /dev/tty - else - ignore "$_file" "$@" - fi - - local _retval=$? - - ignore rm "$_file" - ignore rmdir "$_dir" - - return "$_retval" -} - -check_proc() { - # Check for /proc by looking for the /proc/self/exe link - # This is only run on Linux - if ! test -L /proc/self/exe ; then - err "fatal: Unable to find /proc/self/exe. Is /proc mounted? Installation cannot proceed without /proc." - fi -} - -get_bitness() { - need_cmd head - # Architecture detection without dependencies beyond coreutils. - # ELF files start out "\x7fELF", and the following byte is - # 0x01 for 32-bit and - # 0x02 for 64-bit. - # The printf builtin on some shells like dash only supports octal - # escape sequences, so we use those. - local _current_exe_head - _current_exe_head=$(head -c 5 /proc/self/exe ) - if [ "$_current_exe_head" = "$(printf '\177ELF\001')" ]; then - echo 32 - elif [ "$_current_exe_head" = "$(printf '\177ELF\002')" ]; then - echo 64 - else - err "unknown platform bitness" - fi -} - -is_host_amd64_elf() { - need_cmd head - need_cmd tail - # ELF e_machine detection without dependencies beyond coreutils. - # Two-byte field at offset 0x12 indicates the CPU, - # but we're interested in it being 0x3E to indicate amd64, or not that. - local _current_exe_machine - _current_exe_machine=$(head -c 19 /proc/self/exe | tail -c 1) - [ "$_current_exe_machine" = "$(printf '\076')" ] -} - -get_endianness() { - local cputype=$1 - local suffix_eb=$2 - local suffix_el=$3 - - # detect endianness without od/hexdump, like get_bitness() does. - need_cmd head - need_cmd tail - - local _current_exe_endianness - _current_exe_endianness="$(head -c 6 /proc/self/exe | tail -c 1)" - if [ "$_current_exe_endianness" = "$(printf '\001')" ]; then - echo "${cputype}${suffix_el}" - elif [ "$_current_exe_endianness" = "$(printf '\002')" ]; then - echo "${cputype}${suffix_eb}" - else - err "unknown platform endianness" - fi -} - -get_architecture() { - local _ostype _cputype _bitness _arch _clibtype - _ostype="$(uname -s)" - _cputype="$(uname -m)" - _clibtype="gnu" - - if [ "$_ostype" = Linux ]; then - if [ "$(uname -o)" = Android ]; then - _ostype=Android - fi - if ldd --version 2>&1 | grep -q 'musl'; then - _clibtype="musl" - fi - fi - - if [ "$_ostype" = Darwin ] && [ "$_cputype" = i386 ]; then - # Darwin `uname -m` lies - if sysctl hw.optional.x86_64 | grep -q ': 1'; then - _cputype=x86_64 - fi - fi - - if [ "$_ostype" = SunOS ]; then - # Both Solaris and illumos presently announce as "SunOS" in "uname -s" - # so use "uname -o" to disambiguate. We use the full path to the - # system uname in case the user has coreutils uname first in PATH, - # which has historically sometimes printed the wrong value here. - if [ "$(/usr/bin/uname -o)" = illumos ]; then - _ostype=illumos - fi - - # illumos systems have multi-arch userlands, and "uname -m" reports the - # machine hardware name; e.g., "i86pc" on both 32- and 64-bit x86 - # systems. Check for the native (widest) instruction set on the - # running kernel: - if [ "$_cputype" = i86pc ]; then - _cputype="$(isainfo -n)" - fi - fi - - case "$_ostype" in - - Android) - _ostype=linux-android - ;; - - Linux) - check_proc - _ostype=unknown-linux-$_clibtype - _bitness=$(get_bitness) - ;; - - FreeBSD) - _ostype=unknown-freebsd - ;; - - NetBSD) - _ostype=unknown-netbsd - ;; - - DragonFly) - _ostype=unknown-dragonfly - ;; - - Darwin) - _ostype=apple-darwin - ;; - - illumos) - _ostype=unknown-illumos - ;; - - MINGW* | MSYS* | CYGWIN*) - _ostype=pc-windows-gnu - ;; - - *) - err "unrecognized OS type: $_ostype" - ;; - - esac - - case "$_cputype" in - - i386 | i486 | i686 | i786 | x86) - _cputype=i686 - ;; - - xscale | arm) - _cputype=arm - if [ "$_ostype" = "linux-android" ]; then - _ostype=linux-androideabi - fi - ;; - - armv6l) - _cputype=arm - if [ "$_ostype" = "linux-android" ]; then - _ostype=linux-androideabi - else - _ostype="${_ostype}eabihf" - fi - ;; - - armv7l | armv8l) - _cputype=armv7 - if [ "$_ostype" = "linux-android" ]; then - _ostype=linux-androideabi - else - _ostype="${_ostype}eabihf" - fi - ;; - - aarch64 | arm64) - _cputype=aarch64 - ;; - - x86_64 | x86-64 | x64 | amd64) - _cputype=x86_64 - ;; - - mips) - _cputype=$(get_endianness mips '' el) - ;; - - mips64) - if [ "$_bitness" -eq 64 ]; then - # only n64 ABI is supported for now - _ostype="${_ostype}abi64" - _cputype=$(get_endianness mips64 '' el) - fi - ;; - - ppc) - _cputype=powerpc - ;; - - ppc64) - _cputype=powerpc64 - ;; - - ppc64le) - _cputype=powerpc64le - ;; - - s390x) - _cputype=s390x - ;; - riscv64) - _cputype=riscv64gc - ;; - *) - err "unknown CPU type: $_cputype" - - esac - - # Detect 64-bit linux with 32-bit userland - if [ "${_ostype}" = unknown-linux-gnu ] && [ "${_bitness}" -eq 32 ]; then - case $_cputype in - x86_64) - if [ -n "${RUSTUP_CPUTYPE:-}" ]; then - _cputype="$RUSTUP_CPUTYPE" - else { - # 32-bit executable for amd64 = x32 - if is_host_amd64_elf; then { - echo "This host is running an x32 userland; as it stands, x32 support is poor," 1>&2 - echo "and there isn't a native toolchain -- you will have to install" 1>&2 - echo "multiarch compatibility with i686 and/or amd64, then select one" 1>&2 - echo "by re-running this script with the RUSTUP_CPUTYPE environment variable" 1>&2 - echo "set to i686 or x86_64, respectively." 1>&2 - echo 1>&2 - echo "You will be able to add an x32 target after installation by running" 1>&2 - echo " rustup target add x86_64-unknown-linux-gnux32" 1>&2 - exit 1 - }; else - _cputype=i686 - fi - }; fi - ;; - mips64) - _cputype=$(get_endianness mips '' el) - ;; - powerpc64) - _cputype=powerpc - ;; - aarch64) - _cputype=armv7 - if [ "$_ostype" = "linux-android" ]; then - _ostype=linux-androideabi - else - _ostype="${_ostype}eabihf" - fi - ;; - riscv64gc) - err "riscv64 with 32-bit userland unsupported" - ;; - esac - fi - - # Detect armv7 but without the CPU features Rust needs in that build, - # and fall back to arm. - # See https://github.com/rust-lang/rustup.rs/issues/587. - if [ "$_ostype" = "unknown-linux-gnueabihf" ] && [ "$_cputype" = armv7 ]; then - if ensure grep '^Features' /proc/cpuinfo | grep -q -v neon; then - # At least one processor does not have NEON. - _cputype=arm - fi - fi - - _arch="${_cputype}-${_ostype}" - - RETVAL="$_arch" -} - -say() { - printf 'rustup: %s\n' "$1" -} - -err() { - say "$1" >&2 - exit 1 -} - -need_cmd() { - if ! check_cmd "$1"; then - err "need '$1' (command not found)" - fi -} - -check_cmd() { - command -v "$1" > /dev/null 2>&1 -} - -assert_nz() { - if [ -z "$1" ]; then err "assert_nz $2"; fi -} - -# Run a command that should never fail. If the command fails execution -# will immediately terminate with an error showing the failing -# command. -ensure() { - if ! "$@"; then err "command failed: $*"; fi -} - -# This is just for indicating that commands' results are being -# intentionally ignored. Usually, because it's being executed -# as part of error handling. -ignore() { - "$@" -} - -# This wraps curl or wget. Try curl first, if not installed, -# use wget instead. -downloader() { - local _dld - local _ciphersuites - local _err - local _status - if check_cmd curl; then - _dld=curl - elif check_cmd wget; then - _dld=wget - else - _dld='curl or wget' # to be used in error message of need_cmd - fi - - if [ "$1" = --check ]; then - need_cmd "$_dld" - elif [ "$_dld" = curl ]; then - get_ciphersuites_for_curl - _ciphersuites="$RETVAL" - if [ -n "$_ciphersuites" ]; then - _err=$(curl --proto '=https' --tlsv1.2 --ciphers "$_ciphersuites" --silent --show-error --fail --location "$1" --output "$2" 2>&1) - _status=$? - else - echo "Warning: Not enforcing strong cipher suites for TLS, this is potentially less secure" - if ! check_help_for "$3" curl --proto --tlsv1.2; then - echo "Warning: Not enforcing TLS v1.2, this is potentially less secure" - _err=$(curl --silent --show-error --fail --location "$1" --output "$2" 2>&1) - _status=$? - else - _err=$(curl --proto '=https' --tlsv1.2 --silent --show-error --fail --location "$1" --output "$2" 2>&1) - _status=$? - fi - fi - if [ -n "$_err" ]; then - echo "$_err" >&2 - if echo "$_err" | grep -q 404$; then - err "installer for platform '$3' not found, this may be unsupported" - fi - fi - return $_status - elif [ "$_dld" = wget ]; then - get_ciphersuites_for_wget - _ciphersuites="$RETVAL" - if [ -n "$_ciphersuites" ]; then - _err=$(wget --https-only --secure-protocol=TLSv1_2 --ciphers "$_ciphersuites" "$1" -O "$2" 2>&1) - _status=$? - else - echo "Warning: Not enforcing strong cipher suites for TLS, this is potentially less secure" - if ! check_help_for "$3" wget --https-only --secure-protocol; then - echo "Warning: Not enforcing TLS v1.2, this is potentially less secure" - _err=$(wget "$1" -O "$2" 2>&1) - _status=$? - else - _err=$(wget --https-only --secure-protocol=TLSv1_2 "$1" -O "$2" 2>&1) - _status=$? - fi - fi - if [ -n "$_err" ]; then - echo "$_err" >&2 - if echo "$_err" | grep -q ' 404 Not Found$'; then - err "installer for platform '$3' not found, this may be unsupported" - fi - fi - return $_status - else - err "Unknown downloader" # should not reach here - fi -} - -check_help_for() { - local _arch - local _cmd - local _arg - _arch="$1" - shift - _cmd="$1" - shift - - local _category - if "$_cmd" --help | grep -q 'For all options use the manual or "--help all".'; then - _category="all" - else - _category="" - fi - - case "$_arch" in - - *darwin*) - if check_cmd sw_vers; then - case $(sw_vers -productVersion) in - 10.*) - # If we're running on macOS, older than 10.13, then we always - # fail to find these options to force fallback - if [ "$(sw_vers -productVersion | cut -d. -f2)" -lt 13 ]; then - # Older than 10.13 - echo "Warning: Detected macOS platform older than 10.13" - return 1 - fi - ;; - 11.*) - # We assume Big Sur will be OK for now - ;; - *) - # Unknown product version, warn and continue - echo "Warning: Detected unknown macOS major version: $(sw_vers -productVersion)" - echo "Warning TLS capabilities detection may fail" - ;; - esac - fi - ;; - - esac - - for _arg in "$@"; do - if ! "$_cmd" --help $_category | grep -q -- "$_arg"; then - return 1 - fi - done - - true # not strictly needed -} - -# Return cipher suite string specified by user, otherwise return strong TLS 1.2-1.3 cipher suites -# if support by local tools is detected. Detection currently supports these curl backends: -# GnuTLS and OpenSSL (possibly also LibreSSL and BoringSSL). Return value can be empty. -get_ciphersuites_for_curl() { - if [ -n "${RUSTUP_TLS_CIPHERSUITES-}" ]; then - # user specified custom cipher suites, assume they know what they're doing - RETVAL="$RUSTUP_TLS_CIPHERSUITES" - return - fi - - local _openssl_syntax="no" - local _gnutls_syntax="no" - local _backend_supported="yes" - if curl -V | grep -q ' OpenSSL/'; then - _openssl_syntax="yes" - elif curl -V | grep -iq ' LibreSSL/'; then - _openssl_syntax="yes" - elif curl -V | grep -iq ' BoringSSL/'; then - _openssl_syntax="yes" - elif curl -V | grep -iq ' GnuTLS/'; then - _gnutls_syntax="yes" - else - _backend_supported="no" - fi - - local _args_supported="no" - if [ "$_backend_supported" = "yes" ]; then - # "unspecified" is for arch, allows for possibility old OS using macports, homebrew, etc. - if check_help_for "notspecified" "curl" "--tlsv1.2" "--ciphers" "--proto"; then - _args_supported="yes" - fi - fi - - local _cs="" - if [ "$_args_supported" = "yes" ]; then - if [ "$_openssl_syntax" = "yes" ]; then - _cs=$(get_strong_ciphersuites_for "openssl") - elif [ "$_gnutls_syntax" = "yes" ]; then - _cs=$(get_strong_ciphersuites_for "gnutls") - fi - fi - - RETVAL="$_cs" -} - -# Return cipher suite string specified by user, otherwise return strong TLS 1.2-1.3 cipher suites -# if support by local tools is detected. Detection currently supports these wget backends: -# GnuTLS and OpenSSL (possibly also LibreSSL and BoringSSL). Return value can be empty. -get_ciphersuites_for_wget() { - if [ -n "${RUSTUP_TLS_CIPHERSUITES-}" ]; then - # user specified custom cipher suites, assume they know what they're doing - RETVAL="$RUSTUP_TLS_CIPHERSUITES" - return - fi - - local _cs="" - if wget -V | grep -q '\-DHAVE_LIBSSL'; then - # "unspecified" is for arch, allows for possibility old OS using macports, homebrew, etc. - if check_help_for "notspecified" "wget" "TLSv1_2" "--ciphers" "--https-only" "--secure-protocol"; then - _cs=$(get_strong_ciphersuites_for "openssl") - fi - elif wget -V | grep -q '\-DHAVE_LIBGNUTLS'; then - # "unspecified" is for arch, allows for possibility old OS using macports, homebrew, etc. - if check_help_for "notspecified" "wget" "TLSv1_2" "--ciphers" "--https-only" "--secure-protocol"; then - _cs=$(get_strong_ciphersuites_for "gnutls") - fi - fi - - RETVAL="$_cs" -} - -# Return strong TLS 1.2-1.3 cipher suites in OpenSSL or GnuTLS syntax. TLS 1.2 -# excludes non-ECDHE and non-AEAD cipher suites. DHE is excluded due to bad -# DH params often found on servers (see RFC 7919). Sequence matches or is -# similar to Firefox 68 ESR with weak cipher suites disabled via about:config. -# $1 must be openssl or gnutls. -get_strong_ciphersuites_for() { - if [ "$1" = "openssl" ]; then - # OpenSSL is forgiving of unknown values, no problems with TLS 1.3 values on versions that don't support it yet. - echo "TLS_AES_128_GCM_SHA256:TLS_CHACHA20_POLY1305_SHA256:TLS_AES_256_GCM_SHA384:ECDHE-ECDSA-AES128-GCM-SHA256:ECDHE-RSA-AES128-GCM-SHA256:ECDHE-ECDSA-CHACHA20-POLY1305:ECDHE-RSA-CHACHA20-POLY1305:ECDHE-ECDSA-AES256-GCM-SHA384:ECDHE-RSA-AES256-GCM-SHA384" - elif [ "$1" = "gnutls" ]; then - # GnuTLS isn't forgiving of unknown values, so this may require a GnuTLS version that supports TLS 1.3 even if wget doesn't. - # Begin with SECURE128 (and higher) then remove/add to build cipher suites. Produces same 9 cipher suites as OpenSSL but in slightly different order. - echo "SECURE128:-VERS-SSL3.0:-VERS-TLS1.0:-VERS-TLS1.1:-VERS-DTLS-ALL:-CIPHER-ALL:-MAC-ALL:-KX-ALL:+AEAD:+ECDHE-ECDSA:+ECDHE-RSA:+AES-128-GCM:+CHACHA20-POLY1305:+AES-256-GCM" - fi -} - -main "$@" || exit 1 diff --git a/src/async_session.rs b/src/async_session.rs deleted file mode 100644 index bcd1d3f5..00000000 --- a/src/async_session.rs +++ /dev/null @@ -1,772 +0,0 @@ -// -// Copyright (c) 2017, 2022 ZettaScale Technology -// -// This program and the accompanying materials are made available under the -// terms of the Eclipse Public License 2.0 which is available at -// http://www.eclipse.org/legal/epl-2.0, or the Apache License, Version 2.0 -// which is available at https://www.apache.org/licenses/LICENSE-2.0. -// -// SPDX-License-Identifier: EPL-2.0 OR Apache-2.0 -// -// Contributors: -// ZettaScale Zenoh team, <[email protected]> -// -use super::async_types::{AsyncQueryable, AsyncSubscriber}; -use super::encoding::Encoding; -use super::sample_kind::SampleKind; -use super::types::{ - zkey_expr_of_pyany, zvalue_of_pyany, CongestionControl, KeyExpr, Period, Priority, Query, - QueryConsolidation, QueryTarget, Reliability, Reply, Sample, SubMode, ZnSubOps, -}; -use super::{to_pyerr, ZError}; -use async_std::channel::bounded; -use async_std::sync::Arc; -use async_std::task; -use futures::prelude::*; -use futures::select; -use log::warn; -use pyo3::exceptions; -use pyo3::prelude::*; -use pyo3::types::{IntoPyDict, PyDict, PyTuple}; -use pyo3_asyncio::async_std::future_into_py; -use std::collections::HashMap; -use zenoh::prelude::{ExprId, KeyExpr as ZKeyExpr, Selector, ZFuture, ZInt}; - -/// A zenoh session to be used with asyncio. -#[pyclass] -pub struct AsyncSession { - s: Option<Arc<zenoh::Session>>, -} - -#[pymethods] -impl AsyncSession { - // NOTE: See https://github.com/awestlake87/pyo3-asyncio/issues/50 for the options - // to implement asyncronous methods with async-pyo3. - // Here we choosed to wrap the Session in an Arc, and move a clone of it in each `future_into_py()` call. - // Similarly, each argument coming from Python is converted to Rust and cloned (or Arc-wrapped) before - // moving into the `future_into_py()` call. - - /// Close the zenoh Session. - pub fn close<'p>(&mut self, py: Python<'p>) -> PyResult<&'p PyAny> { - // NOTE: should be sufficient to take the Arc<Session>. Once all arcs are dropped, Session will close. - // Still, we should provide a wait to await for the actual closure... - let _s = self.try_take()?; - future_into_py(py, async move { Ok(()) }) - } - - /// Get informations about the zenoh Session. - /// - /// This method is a **coroutine**. - /// - /// :rtype: **dict[str, str]** - /// - /// :Example: - /// - /// >>> import asyncio, zenoh - /// >>> async def main(): - /// >>> s = await zenoh.async_open() - /// >>> info = await s.info() - /// >>> for key in info: - /// >>> print("{} : {}".format(key, info[key])) - /// >>> - /// >>> asyncio.run(main()) - pub fn info<'p>(&self, py: Python<'p>) -> PyResult<&'p PyAny> { - let s = self.try_clone()?; - future_into_py(py, async move { - use zenoh_cfg_properties::KeyTranscoder; - let result = s - .info() - .map(|props| { - let hashmap: HashMap<String, String> = props - .0 - .into_iter() - .filter_map(|(k, v)| zenoh::info::InfoTranscoder::decode(k).map(|k| (k, v))) - .collect(); - Python::with_gil(|py| hashmap.into_py_dict(py).to_object(py)) - }) - .await; - Ok(result) - }) - } - - /// Put data. - /// - /// The *key_expr* parameter also accepts the following types that can be converted to a :class:`KeyExpr`: - /// - /// * **int** for a mapped key expression - /// * **str** for a literal key expression - /// * **(int, str)** for a mapped key expression with suffix - /// - /// This method is a **coroutine**. - /// - /// :param key_expr: The key expression matching resources to write - /// :type key_expr: :class:`KeyExpr` - /// :param value: The value to write - /// :type value: any type convertible to a :class:`Value` - /// :param \**kwargs: - /// See below - /// - /// :Keyword Arguments: - /// * **encoding** (:class:`Encoding`) -- - /// Set the encoding of the written data - /// * **kind** ( **int** ) -- - /// Set the kind of the written data - /// * **congestion_control** (:class:`CongestionControl`) -- - /// Set the congestion control to apply when routing the data - /// * **priority** (:class:`Priority`) -- - /// Set the priority of the written data - /// * **local_routing** ( **bool** ) -- - /// Enable or disable local routing - /// - /// :raise: :class:`ZError` - /// - /// :Examples: - /// - /// >>> import asyncio, zenoh - /// >>> async def main(): - /// >>> s = await zenoh.async_open() - /// >>> await s.put('/key/expression', bytes('value', encoding='utf8')) - /// >>> - /// >>> asyncio.run(main()) - #[pyo3(text_signature = "(self, key_expr, value, **kwargs)")] - #[args(kwargs = "**")] - pub fn put<'p>( - &self, - key_expr: &PyAny, - value: &PyAny, - kwargs: Option<&PyDict>, - py: Python<'p>, - ) -> PyResult<&'p PyAny> { - let s = self.try_clone()?; - let k = zkey_expr_of_pyany(key_expr)?.to_owned(); - let mut v = zvalue_of_pyany(value)?; - let mut encoding: Option<Encoding> = None; - let mut kind: Option<SampleKind> = None; - let mut congestion_control: Option<CongestionControl> = None; - let mut priority: Option<Priority> = None; - let mut local_routing: Option<bool> = None; - if let Some(kwargs) = kwargs { - if let Some(e) = kwargs.get_item("encoding") { - encoding = e.extract().ok() - } - if let Some(k) = kwargs.get_item("kind") { - kind = k.extract().ok() - } - if let Some(cc) = kwargs.get_item("congestion_control") { - congestion_control = cc.extract().ok() - } - if let Some(p) = kwargs.get_item("priority") { - priority = p.extract().ok() - } - if let Some(lr) = kwargs.get_item("local_routing") { - local_routing = lr.extract().ok() - } - } - if let Some(encoding) = encoding { - v.encoding = encoding.into(); - } - future_into_py(py, async move { - let mut writer = s - .put(k, v) - .kind(kind.unwrap_or_default().kind) - .congestion_control(congestion_control.unwrap_or_default().cc) - .priority(priority.unwrap_or_default().p); - if let Some(local_routing) = local_routing { - writer = writer.local_routing(local_routing); - } - writer.await.map_err(to_pyerr) - }) - } - - /// Delete data. - /// - /// The *key_expr* parameter also accepts the following types that can be converted to a :class:`KeyExpr`: - /// - /// * **int** for a mapped key expression - /// * **str** for a literal key expression - /// * **(int, str)** for a mapped key expression with suffix - /// - /// This method is a **coroutine**. - /// - /// :param key_expr: The key expression matching resources to delete - /// :type key_expr: :class:`KeyExpr` - /// :param \**kwargs: - /// See below - /// - /// :Keyword Arguments: - /// * **congestion_control** (:class:`CongestionControl`) -- - /// Set the congestion control to apply when routing the data - /// * **priority** (:class:`Priority`) -- - /// Set the priority of the written data - /// * **local_routing** ( **bool** ) -- - /// Enable or disable local routing - /// - /// :raise: :class:`ZError` - /// - /// :Examples: - /// - /// >>> import asyncio, zenoh - /// >>> async def main(): - /// >>> s = await zenoh.async_open() - /// >>> await s.delete('/key/expression') - /// >>> - /// >>> asyncio.run(main()) - #[pyo3(text_signature = "(self, key_expr, **kwargs)")] - #[args(kwargs = "**")] - pub fn delete<'p>( - &self, - key_expr: &PyAny, - kwargs: Option<&PyDict>, - py: Python<'p>, - ) -> PyResult<&'p PyAny> { - let s = self.try_clone()?; - let k = zkey_expr_of_pyany(key_expr)?.to_owned(); - let mut congestion_control: Option<CongestionControl> = None; - let mut priority: Option<Priority> = None; - let mut local_routing: Option<bool> = None; - if let Some(kwargs) = kwargs { - if let Some(cc) = kwargs.get_item("congestion_control") { - congestion_control = cc.extract().ok() - } - if let Some(p) = kwargs.get_item("priority") { - priority = p.extract().ok() - } - if let Some(lr) = kwargs.get_item("local_routing") { - local_routing = lr.extract().ok() - } - } - future_into_py(py, async move { - let mut writer = s - .delete(k) - .congestion_control(congestion_control.unwrap_or_default().cc) - .priority(priority.unwrap_or_default().p); - if let Some(local_routing) = local_routing { - writer = writer.local_routing(local_routing); - } - writer.await.map_err(to_pyerr) - }) - } - - /// Associate a numerical Id with the given key expression. - /// - /// This numerical Id will be used on the network to save bandwidth and - /// ease the retrieval of the concerned resource in the routing tables. - /// - /// The *key_expr* parameter also accepts the following types that can be converted to a :class:`KeyExpr`: - /// - /// * **int** for a mapped key expression - /// * **str** for a literal key expression - /// * **(int, str)** for a mapped key expression with suffix - /// - /// This method is a **coroutine**. - /// - /// :param key_expr: The key expression to map to a numerical Id - /// :type key_expr: :class:`KeyExpr` - /// :rtype: **int** - /// :raise: :class:`ZError` - /// - /// :Examples: - /// - /// >>> import asyncio, zenoh - /// >>> async def main(): - /// >>> s = await zenoh.async_open() - /// >>> rid = await s.declare_expr('/key/expression') - /// >>> - /// >>> asyncio.run(main()) - #[pyo3(text_signature = "(self, key_expr)")] - pub fn declare_expr<'p>(&self, key_expr: &PyAny, py: Python<'p>) -> PyResult<&'p PyAny> { - let s = self.try_clone()?; - let k = zkey_expr_of_pyany(key_expr)?.to_owned(); - future_into_py(py, async move { s.declare_expr(k).await.map_err(to_pyerr) }) - } - - /// Undeclare the *numerical Id/key expression* association previously declared - /// with :meth:`declare_expr`. - /// - /// This method is a **coroutine**. - /// - /// :param rid: The numerical Id to unmap - /// :type rid: ExprId - /// :raise: :class:`ZError` - /// - /// :Examples: - /// - /// >>> import asyncio, zenoh - /// >>> async def main(): - /// >>> s = await zenoh.async_open() - /// >>> rid = await s.declare_expr('/key/expression') - /// >>> await s.undeclare_expr(rid) - /// >>> - /// >>> asyncio.run(main()) - #[pyo3(text_signature = "(self, rid)")] - pub fn undeclare_expr<'p>(&self, rid: ExprId, py: Python<'p>) -> PyResult<&'p PyAny> { - let s = self.try_clone()?; - future_into_py( - py, - async move { s.undeclare_expr(rid).wait().map_err(to_pyerr) }, - ) - } - - /// Declare a publication for the given key expression. - /// - /// Written expressions that match the given key expression will only be sent on the network - /// if matching subscribers exist in the system. - /// - /// The *key_expr* parameter also accepts the following types that can be converted to a :class:`KeyExpr`: - /// - /// * **int** for a mapped key expression - /// * **str** for a literal key expression - /// * **(int, str)** for a mapped key expression with suffix - /// - /// This method is a **coroutine**. - /// - /// :param key_expr: The key expression to publish - /// :type key_expr: :class:`KeyExpr` - /// :raise: :class:`ZError` - /// - /// :Examples: - /// - /// >>> import asyncio, zenoh - /// >>> async def main(): - /// >>> s = await zenoh.async_open() - /// >>> rid = await s.declare_publication('/key/expression') - /// >>> await s.put('/key/expression', bytes('value', encoding='utf8')) - /// >>> - /// >>> asyncio.run(main()) - #[pyo3(text_signature = "(self, key_expr)")] - fn declare_publication<'p>(&self, key_expr: &PyAny, py: Python<'p>) -> PyResult<&'p PyAny> { - let s = self.try_clone()?; - let k = zkey_expr_of_pyany(key_expr)?.to_owned(); - future_into_py(py, async move { - s.declare_publication(k).await.map_err(to_pyerr) - }) - } - - #[pyo3(text_signature = "(self, key_expr)")] - fn undeclare_publication<'p>(&self, key_expr: &PyAny, py: Python<'p>) -> PyResult<&'p PyAny> { - let s = self.try_clone()?; - let k = zkey_expr_of_pyany(key_expr)?.to_owned(); - future_into_py(py, async move { - s.undeclare_publication(k).await.map_err(to_pyerr) - }) - } - - /// Create an AsyncSubscriber for the given key expression. - /// - /// The *key_expr* parameter also accepts the following types that can be converted to a :class:`KeyExpr`: - /// - /// * **int** for a mapped key expression - /// * **str** for a literal key expression - /// * **(int, str)** for a mapped key expression with suffix - /// - /// This method is a **coroutine**. - /// - /// :param key_expr: The key expression to subscribe - /// :type key_expr: :class:`KeyExpr` - /// :param callback: the subscription callback (must be a **coroutine**) - /// :type callback: async function(:class:`Sample`) - /// :param \**kwargs: - /// See below - /// - /// :Keyword Arguments: - /// * **reliability** (:class:`Reliability`) -- - /// Set the subscription reliability (BestEffort by default) - /// * **mode** (:class:`SubMode`) -- - /// Set the subscription mode (Push by default) - /// * **period** (:class:`Period`) -- - /// Set the subscription period - /// * **local** ( **bool** ) -- - /// If true make the subscription local only (false by default) - /// - /// :rtype: :class:`Subscriber` - /// :raise: :class:`ZError` - /// - /// :Examples: - /// - /// >>> import asyncio, zenoh - /// >>> from zenoh import Reliability, SubMode - /// >>> - /// >>> async def callback(sample): - /// >>> print("Received : {}".format(sample)) - /// >>> - /// >>> async def main(): - /// >>> s = await zenoh.async_open() - /// >>> sub = await s.subscribe('/key/expression', callback, - /// ... reliability=Reliability.Reliable, - /// ... mode=SubMode.Push) - /// >>> await asycio.sleep(60) - /// >>> - /// >>> asyncio.run(main()) - #[pyo3(text_signature = "(self, key_expr, callback, **kwargs)")] - #[args(kwargs = "**")] - fn subscribe<'p>( - &self, - key_expr: &PyAny, - callback: &PyAny, - kwargs: Option<&PyDict>, - py: Python<'p>, - ) -> PyResult<&'p PyAny> { - let s = self.try_clone()?; - let k = zkey_expr_of_pyany(key_expr)?.to_owned(); - // Note: callback cannot be passed as such in task below because it's not Send - let cb_obj: Py<PyAny> = callback.into(); - // note: extract from kwargs here because it's not Send and cannot be moved into future_into_py(py, F) - let mut reliability: Option<Reliability> = None; - let mut mode: Option<SubMode> = None; - let mut period: Option<Period> = None; - let mut local = false; - if let Some(kwargs) = kwargs { - if let Some(r) = kwargs.get_item("reliability") { - reliability = Some(r.extract()?); - } - if let Some(m) = kwargs.get_item("mode") { - mode = Some(m.extract()?); - } - if let Some(p) = kwargs.get_item("period") { - period = Some(p.extract()?) - } - if let Some(p) = kwargs.get_item("local") { - local = p.extract::<bool>()?; - } - } - - future_into_py(py, async move { - // note: create SubscriberBuilder in this async block since its lifetime is bound to s which is moved here. - let mut sub_builder = s.subscribe(&k); - if let Some(r) = reliability { - sub_builder = sub_builder.reliability(r.r); - } - if let Some(m) = mode { - sub_builder = sub_builder.mode(m.m); - } - if let Some(p) = period { - sub_builder = sub_builder.period(Some(p.p)); - } - if local { - sub_builder = sub_builder.local(); - } - - let zn_sub = sub_builder.await.map_err(to_pyerr)?; - // Note: workaround to allow moving of zn_sub into the task below. - // Otherwise, s is moved also, but can't because it doesn't have 'static lifetime. - let mut static_zn_sub = unsafe { - std::mem::transmute::< - zenoh::subscriber::Subscriber<'_>, - zenoh::subscriber::Subscriber<'static>, - >(zn_sub) - }; - - let (unregister_tx, unregister_rx) = bounded::<ZnSubOps>(8); - // Note: This is done to ensure that even if the call-back into Python - // does any blocking call we do not incour the risk of blocking - // any of the task resolving futures. - let loop_handle = task::spawn_blocking(move || { - Python::with_gil(|py| { - // Run a Python event loop in this task, to allow coroutines execution within the callback - match pyo3_asyncio::async_std::run(py, async move { - loop { - select!( - s = static_zn_sub.receiver().next().fuse() => { - // call the async callback and transform the resulting Python awaitable into a Rust future - let future = match Python::with_gil(|py| { - let cb_args = PyTuple::new(py, &[Sample { s: s.unwrap() }]); - cb_obj.as_ref(py).call1(cb_args).and_then(pyo3_asyncio::async_std::into_future) - }) { - Ok(f) => f, - Err(e) => { warn!("Error calling async queryable callback: {}", e); continue } - }; - // await the future (by default callbacks are executed in sequence) - if let Err(e) = future.await { - warn!("Error suring axecution of async queryable callback: {}", e); - } - }, - op = unregister_rx.recv().fuse() => { - match op { - Ok(ZnSubOps::Pull) => { - if let Err(e) = static_zn_sub.pull().await { - warn!("Error pulling the subscriber: {}", e); - } - }, - Ok(ZnSubOps::Unregister) => { - if let Err(e) = static_zn_sub.close().await { - warn!("Error undeclaring subscriber: {}", e); - } - return Ok(()) - }, - _ => return Ok(()) - } - } - ) - } - }) { - Ok(()) => warn!("Queryable loop running"), - Err(e) => warn!("Failed to start Queryable loop: {}", e), - } - }) - }); - Ok(AsyncSubscriber { - unregister_tx, - loop_handle: Some(loop_handle), - }) - }) - } - - /// Create an AsyncQueryable for the given key expression. - /// - /// The *key_expr* parameter also accepts the following types that can be converted to a :class:`KeyExpr`: - /// - /// * **int** for a mapped key expression - /// * **str** for a literal key expression - /// * **(int, str)** for a mapped key expression with suffix - /// - /// This method is a **coroutine**. - /// - /// :param key_expr: The key expression the Queryable will reply to - /// :type key_expr: :class:`KeyExpr` - /// :param callback: the queryable callback (must be a **coroutine**) - /// :type callback: async function(:class:`Query`) - /// :param \**kwargs: - /// See below - /// - /// :Keyword Arguments: - /// * **kind** ( **int** ) -- - /// Set the queryable kind. This must be a mask of constants defined in :mod:`zenoh.queryable`) - /// (`queryable.EVAL` by default) - /// * **complete** ( **bool** ) -- - /// Set the queryable completeness (true by default) - /// - /// :rtype: :class:`Queryable` - /// :raise: :class:`ZError` - /// - /// :Examples: - /// - /// >>> import asyncio, zenoh - /// >>> from zenoh import Sample, queryable - /// >>> - /// >>> async def callback(query): - /// ... print("Received : {}".format(query)) - /// ... query.reply(Sample('/key/expression', bytes('value', encoding='utf8'))) - /// >>> - /// >>> async def main(): - /// >>> s = await zenoh.async_open() - /// >>> q = await s.queryable('/key/expression', queryable.EVAL, callback) - /// >>> await asycio.sleep(60) - /// >>> - /// >>> asyncio.run(main()) - #[pyo3(text_signature = "(self, key_expr, callback, **kwargs)")] - #[args(kwargs = "**")] - fn queryable<'p>( - &self, - key_expr: &PyAny, - callback: &PyAny, - kwargs: Option<&PyDict>, - py: Python<'p>, - ) -> PyResult<&'p PyAny> { - let s = self.try_clone()?; - let k = zkey_expr_of_pyany(key_expr)?.to_owned(); - // Note: callback cannot be passed as such in task below because it's not Send - let cb_obj: Py<PyAny> = callback.into(); - // note: extract from kwargs here because it's not Send and cannot be moved into future_into_py(py, F) - let mut kind: Option<ZInt> = None; - let mut complete: Option<bool> = None; - if let Some(kwargs) = kwargs { - if let Some(k) = kwargs.get_item("kind") { - kind = Some(k.extract()?); - } - if let Some(p) = kwargs.get_item("local") { - complete = Some(p.extract::<bool>()?); - } - } - - future_into_py(py, async move { - let mut builder = s.queryable(k); - if let Some(k) = kind { - builder = builder.kind(k); - } - if let Some(c) = complete { - builder = builder.complete(c); - } - let zn_quer = builder.await.map_err(to_pyerr)?; - // Note: workaround to allow moving of zn_quer into the task below. - // Otherwise, s is moved also, but can't because it doesn't have 'static lifetime. - let mut zn_quer = unsafe { - std::mem::transmute::< - zenoh::queryable::Queryable<'_>, - zenoh::queryable::Queryable<'static>, - >(zn_quer) - }; - - let (unregister_tx, unregister_rx) = bounded::<bool>(1); - // Note: This is done to ensure that even if the call-back into Python - // does any blocking call we do not incour the risk of blocking - // any of the task resolving futures. - let loop_handle = task::spawn_blocking(move || { - Python::with_gil(|py| { - // Run a Python event loop in this task, to allow coroutines execution within the callback - match pyo3_asyncio::async_std::run(py, async move { - loop { - select!( - q = zn_quer.receiver().next().fuse() => { - // call the async callback and transform the resulting Python awaitable into a Rust future - let future = match Python::with_gil(|py| { - let cb_args = PyTuple::new(py, &[Query { q: async_std::sync::Arc::new(q.unwrap()) }]); - cb_obj.as_ref(py).call1(cb_args).and_then(pyo3_asyncio::async_std::into_future) - }) { - Ok(f) => f, - Err(e) => { warn!("Error calling async queryable callback: {}", e); continue } - }; - // await the future (by default callbacks are executed in sequence) - if let Err(e) = future.await { - warn!("Error suring axecution of async queryable callback: {}", e); - } - }, - _ = unregister_rx.recv().fuse() => { - if let Err(e) = zn_quer.close().await { - warn!("Error undeclaring queryable: {}", e); - } - return Ok(()) - } - ) - } - }) { - Ok(()) => warn!("Queryable loop running"), - Err(e) => warn!("Failed to start Queryable loop: {}", e), - } - }) - }); - Ok(AsyncQueryable { - unregister_tx, - loop_handle: Some(loop_handle), - }) - }) - } - - /// Query data from the matching queryables in the system. - /// - /// Replies are collected in a list. - /// - /// The *selector* parameter also accepts the following types that can be converted to a :class:`Selector`: - /// - /// * **KeyExpr** for a key expression with no value selector - /// * **int** for a key expression id with no value selector - /// * **str** for a litteral selector - /// - /// This method is a **coroutine**. - /// - /// :param selector: The selection of resources to query - /// :type selector: :class:`Selector` - /// :param \**kwargs: - /// See below - /// - /// :Keyword Arguments: - /// * **target** (:class:`QueryTarget`) -- - /// Set the kind of queryables that should be target of this query - /// * **consolidation** (:class:`QueryConsolidation`) -- - /// Set the consolidation mode of the query - /// * **local_routing** ( **bool** ) -- - /// Enable or disable local routing - /// - /// :rtype: [:class:`Reply`] - /// :raise: :class:`ZError` - /// - /// :Examples: - /// - /// >>> import asyncio, zenoh - /// >>> - /// >>> async def main(): - /// >>> s = await zenoh.async_open() - /// >>> replies = await s.get('/key/selector?value_selector') - /// >>> for reply in replies: - /// ... print("Received : {}".format(reply.sample)) - /// >>> - /// >>> asyncio.run(main()) - #[pyo3(text_signature = "(self, selector, **kwargs)")] - #[args(kwargs = "**")] - fn get<'p>( - &self, - selector: &PyAny, - kwargs: Option<&PyDict>, - py: Python<'p>, - ) -> PyResult<&'p PyAny> { - let s = self.try_clone()?; - - let selector: Selector = match selector.get_type().name()? { - "KeyExpr" => { - let key_expr: PyRef<KeyExpr> = selector.extract()?; - key_expr.inner.clone().into() - } - "int" => { - let id: u64 = selector.extract()?; - ZKeyExpr::from(id).into() - } - "str" => { - let name: &str = selector.extract()?; - Selector::from(name) - } - x => { - return Err(PyErr::new::<exceptions::PyValueError, _>(format!( - "Cannot convert type '{}' to a zenoh Selector", - x - ))) - } - } - .to_owned(); - // note: extract from kwargs here because it's not Send and cannot be moved into future_into_py(py, F) - let mut target: Option<QueryTarget> = None; - let mut consolidation: Option<QueryConsolidation> = None; - let mut local_routing: Option<bool> = None; - if let Some(kwargs) = kwargs { - if let Some(arg) = kwargs.get_item("target") { - target = Some(arg.extract::<QueryTarget>()?); - } - if let Some(arg) = kwargs.get_item("consolidation") { - consolidation = Some(arg.extract::<QueryConsolidation>()?); - } - if let Some(arg) = kwargs.get_item("local_routing") { - local_routing = Some(arg.extract::<bool>()?); - } - } - - future_into_py(py, async move { - let mut getter = s.get(selector); - if let Some(t) = target { - getter = getter.target(t.t); - } - if let Some(c) = consolidation { - getter = getter.consolidation(c.c); - } - if let Some(lr) = local_routing { - getter = getter.local_routing(lr); - } - let mut reply_rcv = getter.await.map_err(to_pyerr)?; - let mut replies: Vec<Reply> = Vec::new(); - - while let Some(reply) = reply_rcv.next().await { - replies.push(Reply { r: reply }); - } - Ok(replies) - }) - } -} - -impl AsyncSession { - pub(crate) fn new(s: zenoh::Session) -> Self { - AsyncSession { - s: Some(s.into_arc()), - } - } - - #[inline] - fn try_clone(&self) -> PyResult<Arc<zenoh::Session>> { - self.s - .clone() - .ok_or_else(|| PyErr::new::<ZError, _>("zenoh session was closed")) - } - - #[inline] - fn try_take(&mut self) -> PyResult<Arc<zenoh::Session>> { - self.s - .take() - .ok_or_else(|| PyErr::new::<ZError, _>("zenoh session was closed")) - } -} diff --git a/src/async_types.rs b/src/async_types.rs deleted file mode 100644 index d9bec576..00000000 --- a/src/async_types.rs +++ /dev/null @@ -1,87 +0,0 @@ -// -// Copyright (c) 2017, 2022 ZettaScale Technology Inc. -// -// This program and the accompanying materials are made available under the -// terms of the Eclipse Public License 2.0 which is available at -// http://www.eclipse.org/legal/epl-2.0, or the Apache License, Version 2.0 -// which is available at https://www.apache.org/licenses/LICENSE-2.0. -// -// SPDX-License-Identifier: EPL-2.0 OR Apache-2.0 -// -// Contributors: -// ZettaScale Zenoh team, <[email protected]> -// -use crate::types::*; -use async_std::channel::Sender; -use log::warn; -use pyo3::prelude::*; -use pyo3_asyncio::async_std::future_into_py; - -/// A subscriber to be used with asyncio. -#[pyclass] -pub(crate) struct AsyncSubscriber { - pub(crate) unregister_tx: Sender<ZnSubOps>, - pub(crate) loop_handle: Option<async_std::task::JoinHandle<()>>, -} - -#[pymethods] -impl AsyncSubscriber { - /// Pull available data for a pull-mode subscriber. - /// - /// This method is a **coroutine**. - fn pull<'p>(&self, py: Python<'p>) -> PyResult<&'p PyAny> { - let s = self.unregister_tx.clone(); - future_into_py(py, async move { - if let Err(e) = s.send(ZnSubOps::Pull).await { - warn!("Error in Subscriber::pull() : {}", e); - } - Ok(()) - }) - } - - /// Close the subscriber. - /// - /// This method is a **coroutine**. - fn close<'p>(&mut self, py: Python<'p>) -> PyResult<&'p PyAny> { - if let Some(handle) = self.loop_handle.take() { - let s = self.unregister_tx.clone(); - future_into_py(py, async move { - if let Err(e) = s.send(ZnSubOps::Unregister).await { - warn!("Error in Subscriber::close() : {}", e); - } - handle.await; - Ok(()) - }) - } else { - future_into_py(py, async move { Ok(()) }) - } - } -} - -/// A queryable to be used with asyncio. -#[pyclass] -pub(crate) struct AsyncQueryable { - pub(crate) unregister_tx: Sender<bool>, - pub(crate) loop_handle: Option<async_std::task::JoinHandle<()>>, -} - -#[pymethods] -impl AsyncQueryable { - /// Close the queryable. - /// - /// This method is a **coroutine**. - fn close<'p>(&mut self, py: Python<'p>) -> PyResult<&'p PyAny> { - if let Some(handle) = self.loop_handle.take() { - let s = self.unregister_tx.clone(); - future_into_py(py, async move { - if let Err(e) = s.send(true).await { - warn!("Error in Queryable::close() : {}", e); - } - handle.await; - Ok(()) - }) - } else { - future_into_py(py, async move { Ok(()) }) - } - } -} diff --git a/src/closures.rs b/src/closures.rs new file mode 100644 index 00000000..43e9eda6 --- /dev/null +++ b/src/closures.rs @@ -0,0 +1,98 @@ +// +// Copyright (c) 2017, 2022 ZettaScale Technology Inc. +// +// This program and the accompanying materials are made available under the +// terms of the Eclipse Public License 2.0 which is available at +// http://www.eclipse.org/legal/epl-2.0, or the Apache License, Version 2.0 +// which is available at https://www.apache.org/licenses/LICENSE-2.0. +// +// SPDX-License-Identifier: EPL-2.0 OR Apache-2.0 +// +// Contributors: +// ZettaScale Zenoh team, <[email protected]> +// +use std::{convert::TryFrom, sync::Arc}; + +use pyo3::{prelude::*, types::PyTuple}; +use zenoh::prelude::IntoCallbackReceiverPair; + +trait CallbackUnwrap { + type Output; + fn cb_unwrap(self) -> Self::Output; +} +impl<T> CallbackUnwrap for PyResult<T> { + type Output = T; + fn cb_unwrap(self) -> Self::Output { + match self { + Ok(o) => o, + Err(e) => Python::with_gil(|py| { + if let Some(trace) = e.traceback(py).and_then(|trace| trace.format().ok()) { + panic!("Exception thrown in callback: {}.\n{}", e, trace) + } else { + panic!("Exception thrown in callback: {}.", e,) + } + }), + } + } +} + +pub(crate) struct PyClosure { + pub(crate) pycall: Py<PyAny>, + pub(crate) drop: Option<Py<PyAny>>, + _marker: std::marker::PhantomData, +} +impl TryFrom<&PyAny> for PyClosure { + type Error = PyErr; + fn try_from(value: &PyAny) -> Result<Self, Self::Error> { + Python::with_gil(|py| { + let pycall = match value.getattr("call") { + Ok(value) => value.into_py(py), + Err(e) => return Err(e), + }; + let drop = match value.getattr("drop") { + Ok(value) => { + if value.is_none() { + None + } else { + Some(value.into_py(py)) + } + } + Err(e) => return Err(e), + }; + Ok(PyClosure { + pycall, + drop, + _marker: std::marker::PhantomData, + }) + }) + } +} +impl<I: IntoPy<Py<PyTuple>>> PyClosure { + pub fn call(&self, args: I) -> PyResult<PyObject> { + Python::with_gil(|py| self.pycall.call1(py, args)) + } +} +impl Drop for PyClosure { + fn drop(&mut self) { + if let Some(drop) = self.drop.take() { + Python::with_gil(|py| drop.call0(py)).unwrap(); + } + } +} +impl<T, I> IntoCallbackReceiverPair<'static, T> for PyClosure<(I,)> +where + T: Into, + I: Send + Sync + 'static, + (I,): IntoPy<Py<PyTuple>>, +{ + type Receiver = (); + + fn into_cb_receiver_pair(self) -> (zenoh::handlers::Callback<'static, T>, Self::Receiver) { + ( + Arc::new(move |reply| { + self.call((reply.into(),)).cb_unwrap(); + }), + (), + ) + } +} diff --git a/src/config.rs b/src/config.rs new file mode 100644 index 00000000..7d8f0e8d --- /dev/null +++ b/src/config.rs @@ -0,0 +1,69 @@ +// +// Copyright (c) 2017, 2022 ZettaScale Technology Inc. +// +// This program and the accompanying materials are made available under the +// terms of the Eclipse Public License 2.0 which is available at +// http://www.eclipse.org/legal/epl-2.0, or the Apache License, Version 2.0 +// which is available at https://www.apache.org/licenses/LICENSE-2.0. +// +// SPDX-License-Identifier: EPL-2.0 OR Apache-2.0 +// +// Contributors: +// ZettaScale Zenoh team, <[email protected]> +// + +#![allow(clippy::borrow_deref_ref)] // false positives with pyo3 macros + +use pyo3::prelude::*; +use validated_struct::ValidatedMap; +use zenoh::config::Config; +use zenoh_core::zerror; + +use crate::{ToPyErr, ToPyResult}; + +#[pyclass(subclass)] +pub struct _Config(pub(crate) Option<Config>); + +#[pymethods] +impl _Config { + #[allow(clippy::new_without_default)] + #[new] + pub fn new() -> Self { + _Config(Some(Default::default())) + } + #[staticmethod] + pub fn from_file(expr: &str) -> PyResult<Self> { + match Config::from_file(expr) { + Ok(k) => Ok(Self(Some(k))), + Err(e) => Err(e.to_pyerr()), + } + } + #[staticmethod] + pub fn from_json5(expr: &str) -> PyResult<Self> { + match Config::from_deserializer(&mut json5::Deserializer::from_str(expr).to_pyres()?) { + Ok(k) => Ok(Self(Some(k))), + Err(Ok(_)) => Err(zenoh_core::zerror!( + "{} did parse into a config, but invalid values were found", + expr, + ) + .to_pyerr()), + Err(Err(e)) => Err(e.to_pyerr()), + } + } + + pub fn get_json(&self, path: &str) -> PyResult<String> { + self.0 + .as_ref() + .ok_or_else(|| zerror!("Attempted to use a moved configuration").to_pyerr())? + .get_json(path) + .to_pyres() + } + + pub fn insert_json5(&mut self, path: &str, value: &str) -> PyResult<()> { + self.0 + .as_mut() + .ok_or_else(|| zerror!("Attempted to use a moved configuration").to_pyerr())? + .insert_json5(path, value) + .to_pyres() + } +} diff --git a/src/encoding.rs b/src/encoding.rs deleted file mode 100644 index 53d58542..00000000 --- a/src/encoding.rs +++ /dev/null @@ -1,359 +0,0 @@ -// -// Copyright (c) 2017, 2022 ZettaScale Technology Inc. -// -// This program and the accompanying materials are made available under the -// terms of the Eclipse Public License 2.0 which is available at -// http://www.eclipse.org/legal/epl-2.0, or the Apache License, Version 2.0 -// which is available at https://www.apache.org/licenses/LICENSE-2.0. -// -// SPDX-License-Identifier: EPL-2.0 OR Apache-2.0 -// -// Contributors: -// ZettaScale Zenoh team, <[email protected]> -// -use pyo3::prelude::*; -use pyo3::PyObjectProtocol; -use zenoh::prelude::{Encoding as ZEncoding, KnownEncoding as ZKnownEncoding}; - -/// An encoding known by zenoh and which maps to an integer for wire-efficiency. -#[pyclass] -#[derive(Clone)] -pub struct KnownEncoding { - inner: ZKnownEncoding, -} - -#[allow(non_snake_case)] -#[pymethods] -impl KnownEncoding { - #[classattr] - fn Empty() -> Self { - KnownEncoding { - inner: ZKnownEncoding::Empty, - } - } - - #[classattr] - fn AppOctetStream() -> Self { - KnownEncoding { - inner: ZKnownEncoding::AppOctetStream, - } - } - - #[classattr] - fn AppCustom() -> Self { - KnownEncoding { - inner: ZKnownEncoding::AppCustom, - } - } - - #[classattr] - fn TextPlain() -> Self { - KnownEncoding { - inner: ZKnownEncoding::TextPlain, - } - } - - #[classattr] - fn AppProperties() -> Self { - KnownEncoding { - inner: ZKnownEncoding::AppProperties, - } - } - - #[classattr] - fn AppJson() -> Self { - KnownEncoding { - inner: ZKnownEncoding::AppJson, - } - } - - #[classattr] - fn AppSql() -> Self { - KnownEncoding { - inner: ZKnownEncoding::AppSql, - } - } - - #[classattr] - fn AppInteger() -> Self { - KnownEncoding { - inner: ZKnownEncoding::AppInteger, - } - } - - #[classattr] - fn AppFloat() -> Self { - KnownEncoding { - inner: ZKnownEncoding::AppFloat, - } - } - - #[classattr] - fn AppXml() -> Self { - KnownEncoding { - inner: ZKnownEncoding::AppXml, - } - } - - #[classattr] - fn AppXhtmlXml() -> Self { - KnownEncoding { - inner: ZKnownEncoding::AppXhtmlXml, - } - } - - #[classattr] - fn AppXWwwFormUrlencoded() -> Self { - KnownEncoding { - inner: ZKnownEncoding::AppXWwwFormUrlencoded, - } - } - - #[classattr] - fn TextJson() -> Self { - KnownEncoding { - inner: ZKnownEncoding::TextJson, - } - } - - #[classattr] - fn TextHtml() -> Self { - KnownEncoding { - inner: ZKnownEncoding::TextHtml, - } - } - - #[classattr] - fn TextXml() -> Self { - KnownEncoding { - inner: ZKnownEncoding::TextXml, - } - } - - #[classattr] - fn TextCss() -> Self { - KnownEncoding { - inner: ZKnownEncoding::TextCss, - } - } - - #[classattr] - fn TextCsv() -> Self { - KnownEncoding { - inner: ZKnownEncoding::TextCsv, - } - } - - #[classattr] - fn TextJavascript() -> Self { - KnownEncoding { - inner: ZKnownEncoding::TextJavascript, - } - } - - #[classattr] - fn ImageJpeg() -> Self { - KnownEncoding { - inner: ZKnownEncoding::ImageJpeg, - } - } - - #[classattr] - fn ImagePng() -> Self { - KnownEncoding { - inner: ZKnownEncoding::ImagePng, - } - } - - #[classattr] - fn ImageGif() -> Self { - KnownEncoding { - inner: ZKnownEncoding::ImageGif, - } - } -} -impl From<ZKnownEncoding> for KnownEncoding { - fn from(e: ZKnownEncoding) -> Self { - KnownEncoding { inner: e } - } -} -impl From<KnownEncoding> for ZKnownEncoding { - fn from(e: KnownEncoding) -> Self { - e.inner - } -} - -// zenoh.encoding (simulate the package as a class, and consts as class attributes) -/// A zenoh encoding is a HTTP Mime type represented, for wire efficiency, -/// as an integer prefix (that maps to a string) and a string suffix. -#[allow(non_camel_case_types)] -#[pyclass] -#[derive(Debug, Default, Clone, PartialEq)] -pub struct Encoding { - pub(crate) e: ZEncoding, -} -impl From<ZEncoding> for Encoding { - fn from(e: ZEncoding) -> Self { - Encoding { e } - } -} -impl From<Encoding> for ZEncoding { - fn from(e: Encoding) -> Self { - e.e - } -} - -#[allow(non_snake_case)] -#[pymethods] -impl Encoding { - #[new] - fn new(prefix: KnownEncoding, suffix: Option<String>) -> Self { - if let Some(suffix) = suffix { - Encoding { - e: ZEncoding::WithSuffix(prefix.inner, suffix.into()), - } - } else { - Encoding { - e: ZEncoding::Exact(prefix.inner), - } - } - } - - /// the Encoding's prefix. - /// - /// :type: :class:`KnownEncoding` - #[getter] - fn prefix(&self) -> PyResult<KnownEncoding> { - Ok((*self.e.prefix()).into()) - } - - /// the Encoding's suffix. - /// - /// :type: **str** - #[getter] - fn get_suffix(&self) -> PyResult<&str> { - Ok(self.e.suffix()) - } - - /// Returns a copy of this Encoding, but changing its suffix. - /// - /// :param suffix: The new suffix - /// :type suffix: **str** - /// :rtype: :class:`Encoding` - fn with_suffix(&self, suffix: String) -> Self { - self.e.clone().with_suffix(suffix).into() - } - - #[classattr] - pub fn EMPTY() -> Self { - ZEncoding::EMPTY.into() - } - - #[classattr] - pub fn APP_OCTET_STREAM() -> Self { - ZEncoding::APP_OCTET_STREAM.into() - } - - #[classattr] - pub fn APP_CUSTOM() -> Self { - ZEncoding::APP_CUSTOM.into() - } - - #[classattr] - pub fn TEXT_PLAIN() -> Self { - ZEncoding::TEXT_PLAIN.into() - } - - #[classattr] - pub fn STRING() -> Self { - ZEncoding::STRING.into() - } - - #[classattr] - pub fn APP_PROPERTIES() -> Self { - ZEncoding::APP_PROPERTIES.into() - } - #[classattr] - pub fn APP_JSON() -> Self { - ZEncoding::APP_JSON.into() - } - #[classattr] - pub fn APP_SQL() -> Self { - ZEncoding::APP_SQL.into() - } - #[classattr] - pub fn APP_INTEGER() -> Self { - ZEncoding::APP_INTEGER.into() - } - #[classattr] - pub fn APP_FLOAT() -> Self { - ZEncoding::APP_FLOAT.into() - } - #[classattr] - pub fn APP_XML() -> Self { - ZEncoding::APP_XML.into() - } - #[classattr] - pub fn APP_XHTML_XML() -> Self { - ZEncoding::APP_XHTML_XML.into() - } - #[classattr] - pub fn APP_X_WWW_FORM_URLENCODED() -> Self { - ZEncoding::APP_X_WWW_FORM_URLENCODED.into() - } - #[classattr] - pub fn TEXT_JSON() -> Self { - ZEncoding::TEXT_JSON.into() - } - #[classattr] - pub fn TEXT_HTML() -> Self { - ZEncoding::TEXT_HTML.into() - } - #[classattr] - pub fn TEXT_XML() -> Self { - ZEncoding::TEXT_XML.into() - } - #[classattr] - pub fn TEXT_CSS() -> Self { - ZEncoding::TEXT_CSS.into() - } - #[classattr] - pub fn TEXT_CSV() -> Self { - ZEncoding::TEXT_CSV.into() - } - #[classattr] - pub fn TEXT_JAVASCRIPT() -> Self { - ZEncoding::TEXT_JAVASCRIPT.into() - } - #[classattr] - pub fn IMG_JPG() -> Self { - ZEncoding::IMG_JPG.into() - } - #[classattr] - pub fn IMG_PNG() -> Self { - ZEncoding::IMG_PNG.into() - } - #[classattr] - pub fn IMG_GIF() -> Self { - ZEncoding::IMG_GIF.into() - } - #[classattr] - pub fn DEFAULT() -> Self { - Self::default() - } - #[staticmethod] - fn from_str(s: String) -> Self { - ZEncoding::from(s).into() - } -} -#[pyproto] -impl PyObjectProtocol for Encoding { - fn __str__(&self) -> PyResult<String> { - Ok(self.e.to_string()) - } -} -impl std::fmt::Display for Encoding { - fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result { - write!(f, "{}", self.e) - } -} diff --git a/src/enums.rs b/src/enums.rs new file mode 100644 index 00000000..3209d696 --- /dev/null +++ b/src/enums.rs @@ -0,0 +1,272 @@ +// +// Copyright (c) 2017, 2022 ZettaScale Technology Inc. +// +// This program and the accompanying materials are made available under the +// terms of the Eclipse Public License 2.0 which is available at +// http://www.eclipse.org/legal/epl-2.0, or the Apache License, Version 2.0 +// which is available at https://www.apache.org/licenses/LICENSE-2.0. +// +// SPDX-License-Identifier: EPL-2.0 OR Apache-2.0 +// +// Contributors: +// ZettaScale Zenoh team, <[email protected]> +// +use crate::ToPyErr; +use pyo3::prelude::*; +use zenoh::prelude::{Encoding, KnownEncoding, Priority, SampleKind}; +use zenoh::publication::CongestionControl; +use zenoh::query::{ConsolidationMode, QueryTarget}; +use zenoh::subscriber::Reliability; + +#[macro_export] +macro_rules! derive_richcmp { + ($tyname: expr) => { + fn __richcmp__(&self, other: &Self, op: pyo3::pyclass::CompareOp) -> PyResult<bool> { + match op { + pyo3::pyclass::CompareOp::Eq => Ok(self == other), + pyo3::pyclass::CompareOp::Ne => Ok(self != other), + _ => Err(zenoh_core::zerror!("{} does not support comparison", $tyname).to_pyerr()), + } + } + }; + () => { + fn __richcmp__(&self, other: &Self, op: pyo3::pyclass::CompareOp) -> bool { + match op { + pyo3::pyclass::CompareOp::Lt => self < other, + pyo3::pyclass::CompareOp::Le => self <= other, + pyo3::pyclass::CompareOp::Eq => self == other, + pyo3::pyclass::CompareOp::Ne => self != other, + pyo3::pyclass::CompareOp::Gt => self > other, + pyo3::pyclass::CompareOp::Ge => self >= other, + } + } + }; +} + +#[pyclass(subclass)] +#[repr(transparent)] +#[derive(Clone, Debug, PartialEq, Eq)] +pub struct _Encoding(pub(crate) Encoding); +#[pymethods] +impl _Encoding { + #[new] + pub fn new(this: Self) -> Self { + this + } + derive_richcmp!("Encoding"); + #[classattr] + pub const EMPTY: Self = Self(Encoding::Exact(KnownEncoding::Empty)); + #[classattr] + pub const APP_OCTET_STREAM: Self = Self(Encoding::Exact(KnownEncoding::AppOctetStream)); + #[classattr] + pub const APP_CUSTOM: Self = Self(Encoding::Exact(KnownEncoding::AppCustom)); + #[classattr] + pub const TEXT_PLAIN: Self = Self(Encoding::Exact(KnownEncoding::TextPlain)); + #[classattr] + pub const APP_PROPERTIES: Self = Self(Encoding::Exact(KnownEncoding::AppProperties)); + #[classattr] + pub const APP_JSON: Self = Self(Encoding::Exact(KnownEncoding::AppJson)); + #[classattr] + pub const APP_SQL: Self = Self(Encoding::Exact(KnownEncoding::AppSql)); + #[classattr] + pub const APP_INTEGER: Self = Self(Encoding::Exact(KnownEncoding::AppInteger)); + #[classattr] + pub const APP_FLOAT: Self = Self(Encoding::Exact(KnownEncoding::AppFloat)); + #[classattr] + pub const APP_XML: Self = Self(Encoding::Exact(KnownEncoding::AppXml)); + #[classattr] + pub const APP_XHTML_XML: Self = Self(Encoding::Exact(KnownEncoding::AppXhtmlXml)); + #[classattr] + pub const APP_X_WWW_FORM_URLENCODED: Self = + Self(Encoding::Exact(KnownEncoding::AppXWwwFormUrlencoded)); + #[classattr] + pub const TEXT_JSON: Self = Self(Encoding::Exact(KnownEncoding::TextJson)); + #[classattr] + pub const TEXT_HTML: Self = Self(Encoding::Exact(KnownEncoding::TextHtml)); + #[classattr] + pub const TEXT_XML: Self = Self(Encoding::Exact(KnownEncoding::TextXml)); + #[classattr] + pub const TEXT_CSS: Self = Self(Encoding::Exact(KnownEncoding::TextCss)); + #[classattr] + pub const TEXT_CSV: Self = Self(Encoding::Exact(KnownEncoding::TextCsv)); + #[classattr] + pub const TEXT_JAVASCRIPT: Self = Self(Encoding::Exact(KnownEncoding::TextJavascript)); + #[classattr] + pub const IMAGE_JPEG: Self = Self(Encoding::Exact(KnownEncoding::ImageJpeg)); + #[classattr] + pub const IMAGE_PNG: Self = Self(Encoding::Exact(KnownEncoding::ImagePng)); + #[classattr] + pub const IMAGE_GIF: Self = Self(Encoding::Exact(KnownEncoding::ImageGif)); + #[staticmethod] + pub fn from_str(s: String) -> Self { + Self(s.into()) + } + pub fn __str__(&self) -> String { + self.0.to_string() + } + pub fn append(&mut self, suffix: String) { + unsafe { + let mut tmp = std::ptr::read(&self.0); + tmp = tmp.with_suffix(suffix); + std::ptr::write(&mut self.0, tmp); + } + } + pub fn equals(&self, other: &Self) -> bool { + self == other + } +} + +#[pyclass(subclass)] +#[derive(Clone, PartialEq, Eq)] +pub struct _Priority(pub(crate) Priority); +#[pymethods] +impl _Priority { + #[new] + pub fn new(this: Self) -> Self { + this + } + derive_richcmp!(); + #[classattr] + pub const REAL_TIME: Self = Self(Priority::RealTime); + #[classattr] + pub const INTERACTIVE_HIGH: Self = Self(Priority::InteractiveHigh); + #[classattr] + pub const INTERACTIVE_LOW: Self = Self(Priority::InteractiveLow); + #[classattr] + pub const DATA_HIGH: Self = Self(Priority::DataHigh); + #[classattr] + pub const DATA: Self = Self(Priority::Data); + #[classattr] + pub const DATA_LOW: Self = Self(Priority::DataLow); + #[classattr] + pub const BACKGROUND: Self = Self(Priority::Background); + pub fn __str__(&self) -> &'static str { + match self.0 { + Priority::RealTime => "REAL_TIME", + Priority::InteractiveHigh => "INTERACTIVE_HIGH", + Priority::InteractiveLow => "INTERACTIVE_LOW", + Priority::DataHigh => "DATA_HIGH", + Priority::Data => "DATA", + Priority::DataLow => "DATA_LOW", + Priority::Background => "BACKGROUND", + } + } +} +impl std::cmp::PartialOrd for _Priority { + fn partial_cmp(&self, other: &Self) -> Option<std::cmp::Ordering> { + (self.0 as u8).partial_cmp(&(other.0 as u8)) + } +} + +#[pyclass(subclass)] +#[derive(Clone, PartialEq, Eq)] +pub struct _SampleKind(pub(crate) SampleKind); +#[pymethods] +impl _SampleKind { + #[new] + pub fn new(this: Self) -> Self { + this + } + derive_richcmp!("SampleKind"); + #[classattr] + pub const PUT: Self = Self(SampleKind::Put); + #[classattr] + pub const DELETE: Self = Self(SampleKind::Delete); + pub fn __str__(&self) -> &'static str { + match self.0 { + SampleKind::Put => "PUT", + SampleKind::Delete => "DELETE", + } + } +} + +#[pyclass(subclass)] +#[derive(Clone, PartialEq, Eq)] +pub struct _CongestionControl(pub(crate) CongestionControl); +#[pymethods] +impl _CongestionControl { + #[new] + pub fn new(this: Self) -> Self { + this + } + derive_richcmp!("CongestionControl"); + #[classattr] + pub const BLOCK: Self = Self(CongestionControl::Block); + #[classattr] + pub const DROP: Self = Self(CongestionControl::Drop); + pub fn __str__(&self) -> &'static str { + match self.0 { + CongestionControl::Block => "BLOCK", + CongestionControl::Drop => "DROP", + } + } +} + +#[pyclass(subclass)] +#[derive(Clone, PartialEq, Eq)] +pub struct _Reliability(pub(crate) Reliability); +#[pymethods] +impl _Reliability { + #[new] + pub fn new(this: Self) -> Self { + this + } + derive_richcmp!("Reliability"); + #[classattr] + pub const BEST_EFFORT: Self = Self(Reliability::BestEffort); + #[classattr] + pub const RELIABLE: Self = Self(Reliability::Reliable); + pub fn __str__(&self) -> &'static str { + match self.0 { + Reliability::BestEffort => "BEST_EFFORT", + Reliability::Reliable => "RELIABLE", + } + } +} + +#[pyclass(subclass)] +#[derive(Clone, PartialEq, Eq)] +pub struct _QueryTarget(pub(crate) QueryTarget); +#[pymethods] +impl _QueryTarget { + #[new] + pub fn new(this: Self) -> Self { + this + } + derive_richcmp!("QueryTarget"); + #[classattr] + pub const BEST_MATCHING: Self = Self(QueryTarget::BestMatching); + #[classattr] + pub const ALL: Self = Self(QueryTarget::All); + #[classattr] + pub const ALL_COMPLETE: Self = Self(QueryTarget::AllComplete); + pub fn __str__(&self) -> &'static str { + match self.0 { + QueryTarget::BestMatching => "BEST_MATCHING", + QueryTarget::All => "ALL", + QueryTarget::AllComplete => "ALL_COMPLETE", + #[cfg(feature = "complete_n")] + QueryTarget::Complete(_) => "COMPLETE_N", + } + } +} + +#[pyclass(subclass)] +#[derive(Clone, PartialEq, Eq)] +pub struct _QueryConsolidation(pub(crate) Option<ConsolidationMode>); +#[pymethods] +impl _QueryConsolidation { + #[new] + pub fn new(this: Self) -> Self { + this + } + derive_richcmp!("QueryConsolidation"); + #[classattr] + pub const AUTO: Self = Self(None); + #[classattr] + pub const NONE: Self = Self(Some(ConsolidationMode::None)); + #[classattr] + pub const MONOTONIC: Self = Self(Some(ConsolidationMode::Monotonic)); + #[classattr] + pub const LATEST: Self = Self(Some(ConsolidationMode::Latest)); +} diff --git a/src/keyexpr.rs b/src/keyexpr.rs new file mode 100644 index 00000000..a53932c4 --- /dev/null +++ b/src/keyexpr.rs @@ -0,0 +1,119 @@ +// +// Copyright (c) 2017, 2022 ZettaScale Technology Inc. +// +// This program and the accompanying materials are made available under the +// terms of the Eclipse Public License 2.0 which is available at +// http://www.eclipse.org/legal/epl-2.0, or the Apache License, Version 2.0 +// which is available at https://www.apache.org/licenses/LICENSE-2.0. +// +// SPDX-License-Identifier: EPL-2.0 OR Apache-2.0 +// +// Contributors: +// ZettaScale Zenoh team, <[email protected]> +// + +use pyo3::prelude::*; +use std::{ + collections::{hash_map::DefaultHasher, HashMap}, + convert::{TryFrom, TryInto}, +}; +use zenoh::prelude::{sync::SyncResolve, KeyExpr, Selector}; + +use crate::{session::_Session, ToPyErr}; + +#[pyclass(subclass)] +#[derive(Clone)] +pub struct _KeyExpr(pub(crate) KeyExpr<'static>); + +#[pymethods] +impl _KeyExpr { + #[new] + pub fn pynew(this: Self) -> Self { + this + } + #[staticmethod] + pub fn new(expr: String) -> PyResult<Self> { + match expr.try_into() { + Ok(k) => Ok(Self(k)), + Err(e) => Err(e.to_pyerr()), + } + } + #[staticmethod] + pub fn autocanonize(expr: String) -> PyResult<Self> { + match KeyExpr::autocanonize(expr) { + Ok(k) => Ok(Self(k)), + Err(e) => Err(e.to_pyerr()), + } + } + + pub fn intersects(&self, other: &Self) -> bool { + self.0.intersects(&other.0) + } + + pub fn includes(&self, other: &Self) -> bool { + self.0.includes(&other.0) + } + + pub fn equals(&self, other: &Self) -> bool { + self.0 == other.0 + } + + pub fn undeclare(&self, session: &_Session) -> PyResult<()> { + session + .0 + .undeclare(self.0.clone()) + .res_sync() + .map_err(|e| e.to_pyerr()) + } + + pub fn __str__(&self) -> &str { + self.0.as_str() + } + + pub fn __hash__(&self) -> isize { + use std::hash::*; + let mut hasher: DefaultHasher = BuildHasherDefault::default().build_hasher(); + self.0.hash(&mut hasher); + hasher.finish() as isize + } + + pub fn __eq__(&self, other: &Self) -> bool { + self.0 == other.0 + } +} + +#[pyclass(subclass)] +#[derive(Clone)] +pub struct _Selector(pub(crate) Selector<'static>); +#[pymethods] +impl _Selector { + #[new] + pub fn pynew(this: Self) -> Self { + this + } + #[staticmethod] + pub fn new(expr: String) -> PyResult<Self> { + match Selector::try_from(expr) { + Ok(o) => Ok(_Selector(o)), + Err(e) => Err(e.to_pyerr()), + } + } + #[getter] + pub fn key_expr(&self) -> _KeyExpr { + _KeyExpr(self.0.key_expr.clone()) + } + #[getter] + pub fn get_parameters(&self) -> &str { + self.0.parameters() + } + #[setter] + pub fn set_parameters(&mut self, parameters: String) { + self.0.set_parameters(parameters) + } + pub fn decode_parameters(&self) -> PyResult<HashMap<String, String>> { + self.0.parameters_map().map_err(|e| e.to_pyerr()) + } + pub fn __str__(&self) -> String { + self.0.to_string() + } +} diff --git a/src/lib.rs b/src/lib.rs index 7625fde2..d9abd5d7 100644 --- a/src/lib.rs +++ b/src/lib.rs @@ -11,399 +11,105 @@ // Contributors: // ZettaScale Zenoh team, <[email protected]> // -use async_std::prelude::FutureExt; -use async_std::task; -use encoding::KnownEncoding; -use futures::prelude::*; -use pyo3::prelude::*; -use pyo3::{create_exception, wrap_pyfunction}; -use pyo3_asyncio::async_std::future_into_py; -use zenoh::config::{Config as ZConfig, Notifier}; -use zenoh::prelude::ValidatedMap; -use zenoh_core::zerror; - -pub(crate) mod types; -pub(crate) use types::*; +use pyo3::{prelude::*, types::PyDict, ToPyObject}; +mod closures; +mod config; +mod enums; +mod keyexpr; +mod queryable; mod session; -use session::*; -mod async_types; -mod encoding; -mod sample_kind; -use async_types::*; -mod async_session; -use async_session::*; +mod value; -create_exception!(zenoh, ZError, pyo3::exceptions::PyException); +pyo3::create_exception!(zenoh, ZError, pyo3::exceptions::PyException); -fn to_pyerr(err: zenoh_core::Error) -> PyErr { - PyErr::new::<ZError, _>(err.to_string()) -} -/// The zenoh API. -/// -/// Examples of use: -/// ^^^^^^^^^^^^^^^^ -/// -/// Publish -/// """"""" -/// -/// >>> import zenoh -/// >>> s = zenoh.open() -/// >>> s.put('/resource/name', bytes('value', 'utf8')) -/// -/// Subscribe -/// """"""""" -/// -/// >>> import zenoh -/// >>> def listener(sample): -/// ... print("Received: {} = {}".format(sample.key_expr, sample.payload.decode("utf-8"))) -/// >>> -/// >>> s = zenoh.open() -/// >>> sub = s.subscribe('/resource/*', listener) -/// -/// Get -/// """ -/// -/// >>> import zenoh -/// >>> s = zenoh.open() -/// >>> for reply in s.get('/resource/name'): -/// ... print("Received: {} = {}".format(reply.sample.key_expr, reply.sample.payload.decode('utf-8')) -#[pymodule] -pub fn zenoh(py: Python, m: &PyModule) -> PyResult<()> { - m.add_class::<config>()?; - // force addition of "zenoh.config" module - // (see https://github.com/PyO3/pyo3/issues/759#issuecomment-653964601) - py.run( - "\ -import sys -sys.modules['zenoh.config'] = config - ", - None, - Some(m.dict()), - )?; - - m.add_class::<info>()?; - // force addition of "zenoh.info" module - // (see https://github.com/PyO3/pyo3/issues/759#issuecomment-653964601) - py.run( - "\ -import sys -sys.modules['zenoh.info'] = info - ", - None, - Some(m.dict()), - )?; - - m.add_class::<queryable>()?; - // force addition of "zenoh.queryable" module - // (see https://github.com/PyO3/pyo3/issues/759#issuecomment-653964601) - py.run( - "\ -import sys -sys.modules['zenoh.queryable'] = queryable - ", - None, - Some(m.dict()), - )?; - - m.add_class::<AsyncQueryable>()?; - m.add_class::<AsyncSession>()?; - m.add_class::<AsyncSubscriber>()?; - m.add_class::<Config>()?; - m.add_class::<CongestionControl>()?; - m.add_class::<ConsolidationMode>()?; - m.add_class::<encoding::Encoding>()?; - m.add_class::<Hello>()?; - m.add_class::<KeyExpr>()?; - m.add_class::<KnownEncoding>()?; - m.add_class::<PeerId>()?; - m.add_class::<Period>()?; - m.add_class::<Priority>()?; - m.add_class::<Query>()?; - m.add_class::<Queryable>()?; - m.add_class::<QueryConsolidation>()?; - m.add_class::<QueryTarget>()?; - m.add_class::<Reliability>()?; - m.add_class::<Reply>()?; - m.add_class::<Sample>()?; - m.add_class::<sample_kind::SampleKind>()?; - m.add_class::<Selector>()?; - m.add_class::<Session>()?; - m.add_class::<SourceInfo>()?; - m.add_class::<SubMode>()?; - m.add_class::<Subscriber>()?; - m.add_class::<Target>()?; - m.add_class::<Timestamp>()?; - m.add_class::<Value>()?; - m.add_class::<ValueSelector>()?; - m.add_class::<WhatAmI>()?; - m.add("ZError", py.get_type::<ZError>())?; - m.add_wrapped(wrap_pyfunction!(init_logger))?; - m.add_wrapped(wrap_pyfunction!(config_from_file))?; - m.add_wrapped(wrap_pyfunction!(open))?; - m.add_wrapped(wrap_pyfunction!(async_open))?; - m.add_wrapped(wrap_pyfunction!(scout))?; - m.add_wrapped(wrap_pyfunction!(async_scout))?; - Ok(()) -} -/// Initialize the logger used by the Rust implementation of this API. -/// -/// Once initialized, you can configure the logs displayed by the API using the ``RUST_LOG`` environment variable. -/// For instance, start python with the *debug* logs available:: -/// -/// $ RUST_LOG=debug python -/// -/// More details on the RUST_LOG configuration on https://docs.rs/env_logger/latest/env_logger -/// -#[pyfunction] -fn init_logger() { - env_logger::init(); +pub(crate) trait ToPyErr { + fn to_pyerr(self) -> PyErr; } - -/// Parse a configuration file for zenoh, returning a Config object. -/// -/// :param path: The path to the config file. -/// :rtype: :class:`Config` -#[pyfunction] -fn config_from_file(path: &str) -> PyResult<Config> { - Config::from_file(path) +impl<E: std::error::Error> ToPyErr for E { + fn to_pyerr(self) -> PyErr { + PyErr::new::<ZError, _>(self.to_string()) + } } - -/// The main configuration structure for Zenoh. -/// -/// To construct a configuration, we advise that you use a configuration file -/// (JSON, JSON5 and YAML are currently supported, please use the proper extension for your format as the deserializer will be picked according to it). -/// A Config object can then be amended calling :func:`Config.insert_json5`. -/// -/// :Example: -/// -/// >>> import zenoh, json -/// >>> conf = zenoh.Config.from_file('zenoh-config.json5') -/// >>> conf.insert_json5(zenoh.config.MODE_KEY, json.dumps('client')) -/// >>> print(conf.json()) -#[pyclass] -#[derive(Debug, Clone)] -pub struct Config { - inner: ZConfig, +pub(crate) trait ToPyResult<T> { + fn to_pyres(self) -> Result<T, PyErr>; } -#[pymethods] -impl Config { - /// Constructor of a default configuration. - #[new] - pub fn new() -> Self { - Config { - inner: ZConfig::default(), - } - } - - /// Parse a configuration file for zenoh, returning a Config object. - /// - /// :param path: The path to the config file. - /// :rtype: :class:`Config` - /// :raise: :class:`ZError` - pub fn insert_json5(&mut self, key: &str, value: &str) -> PyResult<()> { - match self.inner.insert_json5(key, value) { - Ok(()) => Ok(()), - Err(e) => Err(PyErr::new::<pyo3::exceptions::PyException, _>( - e.to_string(), - )), - } - } - - /// Returns the config as a JSON string - /// - /// :rtype: str - pub fn json(&self) -> String { - serde_json::to_string(&self.inner).unwrap() - } - - /// Parse a JSON5 string configuration for zenoh, returning a Config object. - /// - /// :param input: The configuration as a JSON5 string. - /// :rtype: :class:`Config` - /// :raise: :class:`ZError` - #[staticmethod] - pub fn from_json5(input: &str) -> PyResult<Self> { - let mut d = match json5::Deserializer::from_str(input) { - Ok(d) => d, - Err(e) => return Err(to_pyerr(zerror!(e).into())), - }; - match ZConfig::from_deserializer(&mut d) { - Ok(inner) => Ok(Config { inner }), - Err(e) => Err(to_pyerr( - match e { - Ok(c) => zerror!("invalid configuration: {:?}", c), - Err(e) => zerror!(e), - } - .into(), - )), - } - } - - /// Parse a configuration file for zenoh, returning a Config object. - /// - /// :param path: The path to the config file. - /// :rtype: :class:`Config` - /// :raise: :class:`ZError` - #[staticmethod] - pub fn from_file(path: &str) -> PyResult<Self> { - match ZConfig::from_file(path) { - Ok(inner) => Ok(Config { inner }), - Err(e) => Err(to_pyerr(e)), - } +impl<T, E: ToPyErr> ToPyResult<T> for Result<T, E> { + fn to_pyres(self) -> Result<T, PyErr> { + self.map_err(ToPyErr::to_pyerr) } } -impl Default for Config { - fn default() -> Self { - Self::new() +enum ExtractError { + Unavailable(Option<PyErr>), + Other(PyErr), +} +impl From<PyErr> for ExtractError { + fn from(e: PyErr) -> Self { + Self::Other(e) } } - -#[pyclass] -#[derive(Clone)] -pub struct ConfigNotifier { - inner: Notifier<ZConfig>, +pub(crate) trait PyExtract<K> { + fn extract_item<'a, V: FromPyObject<'a>>(&'a self, key: K) -> Result<V, ExtractError>; } -#[pymethods] -impl ConfigNotifier { - pub fn insert_json5(&mut self, key: &str, value: &str) -> PyResult<()> { - match self.inner.insert_json5(key, value) { - Ok(()) => Ok(()), - Err(e) => Err(PyErr::new::<pyo3::exceptions::PyException, _>( - e.to_string(), - )), +impl<K: ToPyObject> PyExtract<K> for PyAny { + fn extract_item<'a, V: FromPyObject<'a>>(&'a self, key: K) -> Result<V, ExtractError> { + match self.get_item(key) { + Ok(item) => Ok(item.extract::<V>()?), + Err(e) => Err(ExtractError::Unavailable(Some(e))), } } - pub fn json(&self) -> String { - serde_json::to_string(&*self.inner.lock()).unwrap() - } } - -/// Open a zenoh Session. -/// -/// :param config: The configuration of the zenoh session -/// :type config: :class:`Config`, optional -/// :rtype: :class:`Session` -/// :raise: :class:`ZError` -/// -/// :Example: -/// -/// >>> import zenoh -/// >>> z = zenoh.open(zenoh.config::peer()) -#[pyfunction] -#[pyo3(text_signature = "(config)")] -fn open(config: Option<Config>) -> PyResult<Session> { - let s = task::block_on(zenoh::open(config.unwrap_or_default().inner)).map_err(to_pyerr)?; - Ok(Session::new(s)) +impl<K: ToPyObject> PyExtract<K> for PyDict { + fn extract_item<'a, V: FromPyObject<'a>>(&'a self, key: K) -> Result<V, ExtractError> { + match self.get_item(key) { + Some(item) => Ok(item.extract::<V>()?), + None => Err(ExtractError::Unavailable(None)), + } + } } -/// Coroutine to open a zenoh AsyncSession (similar to a Session, but for asyncio usage). -/// -/// :param config: The configuration of the zenoh session -/// :type config: :class:`Config`, optional -/// :rtype: :class:`AsyncSession` -/// :raise: :class:`ZError` -/// -/// :Example: -/// -/// >>> import asyncio, zenoh -/// >>> async def main(): -/// >>> z = await zenoh.async_open() -/// >>> -/// >>> asyncio.run(main()) -#[pyfunction] -#[pyo3(text_signature = "(config)")] -fn async_open(py: Python, config: Option<Config>) -> PyResult<&PyAny> { - future_into_py(py, async { - let s = zenoh::open(config.unwrap_or_default().inner) - .await - .map_err(to_pyerr)?; - Ok(AsyncSession::new(s)) - }) +#[pymodule] +fn zenoh(_py: Python, m: &PyModule) -> PyResult<()> { + m.add_class::<config::_Config>()?; + m.add_class::<keyexpr::_KeyExpr>()?; + m.add_class::<keyexpr::_Selector>()?; + m.add_class::<session::_Session>()?; + m.add_class::<session::_Publisher>()?; + m.add_class::<session::_Subscriber>()?; + m.add_class::<session::_PullSubscriber>()?; + m.add_class::<session::_Scout>()?; + m.add_class::<queryable::_Query>()?; + m.add_class::<queryable::_Queryable>()?; + m.add_class::<value::_Value>()?; + m.add_class::<value::_Sample>()?; + m.add_class::<value::_Reply>()?; + m.add_class::<value::_Timestamp>()?; + m.add_class::<value::_Hello>()?; + m.add_class::<value::_ZenohId>()?; + m.add_class::<enums::_CongestionControl>()?; + m.add_class::<enums::_Encoding>()?; + m.add_class::<enums::_Priority>()?; + m.add_class::<enums::_SampleKind>()?; + m.add_class::<enums::_Reliability>()?; + m.add_class::<enums::_QueryConsolidation>()?; + m.add_class::<enums::_QueryTarget>()?; + m.add_wrapped(wrap_pyfunction!(init_logger))?; + m.add_wrapped(wrap_pyfunction!(session::scout))?; + Ok(()) } -/// Scout for routers and/or peers. +/// Initialize the logger used by the Rust implementation of this API. /// -/// Sends scout messages for a specified duration and returns -/// a list of received :class:`Hello` messages. +/// Once initialized, you can configure the logs displayed by the API using the ``RUST_LOG`` environment variable. +/// For instance, start python with the *debug* logs available:: /// -/// :param whatami: The kind of zenoh process to scout for -/// :type whatami: **int** -/// :param scout_duration: the duration of scout (in seconds) -/// :type scout_duration: **float** -/// :param config: The configuration to use for scouting -/// :type config: :class:`Config`, optional -/// :rtype: list of :class:`Hello` -/// :raise: :class:`ZError` +/// $ RUST_LOG=debug python /// -/// :Example: +/// More details on the RUST_LOG configuration on https://docs.rs/env_logger/latest/env_logger /// -/// >>> import zenoh -/// >>> hellos = zenoh.scout(WhatAmI.Peer | WhatAmI.Router, 1.0) -/// >>> for hello in hellos: -/// >>> print(hello) #[pyfunction] -#[pyo3(text_signature = "(whatami, scout_duration, config)")] -fn scout(whatami: WhatAmI, scout_duration: f64, config: Option<Config>) -> PyResult<Vec<Hello>> { - task::block_on(async move { - let mut result = Vec::<Hello>::new(); - let mut receiver = zenoh::scout(whatami, config.unwrap_or_default().inner) - .await - .unwrap(); - let scout = async { - while let Some(h) = receiver.next().await { - result.push(Hello { h }) - } - }; - let timeout = async_std::task::sleep(std::time::Duration::from_secs_f64(scout_duration)); - FutureExt::race(scout, timeout).await; - Ok(result) - }) +fn init_logger() { + let _ = env_logger::try_init(); } -/// Coroutine to scout for routers and/or peers. -/// -/// Sends scout messages for a specified duration and returns -/// a list of received :class:`Hello` messages. -/// -/// :param whatami: The kind of zenoh process to scout for -/// :type whatami: **int** -/// :param scout_duration: the duration of scout (in seconds) -/// :type scout_duration: **float** -/// :param config: The configuration to use for scouting -/// :type config: :class:`Config`, optional -/// :rtype: list of :class:`Hello` -/// :raise: :class:`ZError` -/// -/// :Example: -/// -/// >>> import asyncio, zenoh -/// >>> async def main(): -/// >>> hellos = await zenoh.async_scout(WhatAmI.Peer | WhatAmI.Router, 1.0) -/// >>> for hello in hellos: -/// >>> print(hello) -/// >>> -/// >>> asyncio.run(main()) -#[pyfunction] -#[pyo3(text_signature = "(whatami, scout_duration, config)")] -fn async_scout( - whatami: WhatAmI, - scout_duration: f64, - config: Option<Config>, - py: Python, -) -> PyResult<&PyAny> { - future_into_py(py, async move { - let mut result = Vec::<Hello>::new(); - let mut receiver = zenoh::scout(whatami, config.unwrap_or_default().inner) - .await - .unwrap(); - let scout = async { - while let Some(h) = receiver.next().await { - result.push(Hello { h }) - } - }; - let timeout = async_std::task::sleep(std::time::Duration::from_secs_f64(scout_duration)); - FutureExt::race(scout, timeout).await; - Ok(result) - }) -} +pub(crate) use value::PyAnyToValue; diff --git a/src/queryable.rs b/src/queryable.rs new file mode 100644 index 00000000..c0f6aa28 --- /dev/null +++ b/src/queryable.rs @@ -0,0 +1,84 @@ +// +// Copyright (c) 2017, 2022 ZettaScale Technology Inc. +// +// This program and the accompanying materials are made available under the +// terms of the Eclipse Public License 2.0 which is available at +// http://www.eclipse.org/legal/epl-2.0, or the Apache License, Version 2.0 +// which is available at https://www.apache.org/licenses/LICENSE-2.0. +// +// SPDX-License-Identifier: EPL-2.0 OR Apache-2.0 +// +// Contributors: +// ZettaScale Zenoh team, <[email protected]> +// +use std::{collections::HashMap, sync::Arc}; + +use pyo3::prelude::*; +use zenoh::{ + queryable::{Query, Queryable}, + selector::Parameters, +}; +use zenoh_core::SyncResolve; + +use crate::{ + keyexpr::{_KeyExpr, _Selector}, + value::_Sample, + ToPyErr, +}; + +#[pyclass(subclass)] +#[derive(Clone)] +pub struct _Query(pub(crate) Arc<Query>); +#[pymethods] +impl _Query { + #[new] + pub fn pynew(this: Self) -> Self { + this + } + #[getter] + pub fn key_expr(&self) -> _KeyExpr { + _KeyExpr(self.0.key_expr().clone()) + } + #[getter] + pub fn parameters(&self) -> &str { + self.0.parameters() + } + pub fn decode_parameters(&self) -> PyResult<HashMap<String, String>> { + let mut res = HashMap::new(); + for (k, v) in self.0.parameters().decode() { + let k = k.into_owned(); + match res.entry(k) { + std::collections::hash_map::Entry::Occupied(e) => { + return Err(zenoh_core::zerror!( + "Detected duplicate key {} in value selector {}", + e.key(), + self.0.parameters() + ) + .to_pyerr()) + } + std::collections::hash_map::Entry::Vacant(e) => { + e.insert(v.into_owned()); + } + } + } + Ok(res) + } + #[getter] + pub fn selector(&self) -> _Selector { + _Selector(self.0.selector().clone().into_owned()) + } + pub fn reply(&self, sample: _Sample) -> PyResult<()> { + self.0 + .reply(Ok(sample.into())) + .res_sync() + .map_err(|e| e.to_pyerr()) + } +} +impl From<Query> for _Query { + fn from(q: Query) -> Self { + Self(Arc::new(q)) + } +} + +#[pyclass(subclass)] +pub struct _Queryable(pub(crate) Queryable<'static, ()>); diff --git a/src/sample_kind.rs b/src/sample_kind.rs deleted file mode 100644 index fd3dcdb7..00000000 --- a/src/sample_kind.rs +++ /dev/null @@ -1,81 +0,0 @@ -// -// Copyright (c) 2017, 2022 ZettaScale Technology Inc. -// -// This program and the accompanying materials are made available under the -// terms of the Eclipse Public License 2.0 which is available at -// http://www.eclipse.org/legal/epl-2.0, or the Apache License, Version 2.0 -// which is available at https://www.apache.org/licenses/LICENSE-2.0. -// -// SPDX-License-Identifier: EPL-2.0 OR Apache-2.0 -// -// Contributors: -// ZettaScale Zenoh team, <[email protected]> -// -use pyo3::class::basic::CompareOp; -use pyo3::prelude::*; -use pyo3::PyObjectProtocol; -use zenoh::prelude::SampleKind as ZSampleKind; - -// zenoh.sample_kind (simulate the package as a class, and consts as class attributes) -/// Constants defining the different data kinds. -#[allow(non_camel_case_types)] -#[pyclass] -#[derive(Debug, Clone, PartialEq, Default)] -pub struct SampleKind { - pub kind: ZSampleKind, -} -impl From<ZSampleKind> for SampleKind { - fn from(kind: ZSampleKind) -> Self { - SampleKind { kind } - } -} -impl From<SampleKind> for ZSampleKind { - fn from(kind: SampleKind) -> Self { - kind.kind - } -} - -#[allow(non_snake_case)] -#[pymethods] -impl SampleKind { - #[classattr] - pub fn PUT() -> Self { - ZSampleKind::Put.into() - } - - #[classattr] - pub fn PATCH() -> Self { - ZSampleKind::Patch.into() - } - - #[classattr] - pub fn DELETE() -> Self { - ZSampleKind::Delete.into() - } - - #[classattr] - pub fn DEFAULT() -> Self { - ZSampleKind::default().into() - } -} - -#[pyproto] -impl PyObjectProtocol for SampleKind { - fn __str__(&self) -> PyResult<String> { - Ok(self.to_string()) - } - - fn __richcmp__(&'p self, other: PyRef<'p, SampleKind>, op: CompareOp) -> PyResult<PyObject> { - match op { - CompareOp::Eq => Ok(self.eq(&*other).into_py(other.py())), - CompareOp::Ne => Ok((!self.eq(&*other)).into_py(other.py())), - _ => Ok(other.py().NotImplemented()), - } - } -} - -impl std::fmt::Display for SampleKind { - fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result { - write!(f, "{}", self.kind) - } -} diff --git a/src/session.rs b/src/session.rs index 049d032c..6c9ecacf 100644 --- a/src/session.rs +++ b/src/session.rs @@ -1,5 +1,3 @@ -use crate::ConfigNotifier; - // // Copyright (c) 2017, 2022 ZettaScale Technology Inc. // @@ -13,637 +11,294 @@ use crate::ConfigNotifier; // Contributors: // ZettaScale Zenoh team, <[email protected]> // -use super::encoding::Encoding; -use super::sample_kind::SampleKind; -use super::types::{ - zkey_expr_of_pyany, zvalue_of_pyany, CongestionControl, KeyExpr, Period, Priority, Query, - QueryConsolidation, QueryTarget, Queryable, Reliability, Reply, Sample, SubMode, Subscriber, - ZnSubOps, -}; -use super::{to_pyerr, ZError}; -use async_std::channel::bounded; -use async_std::task; -use futures::prelude::*; -use futures::select; -use log::warn; -use pyo3::exceptions; -use pyo3::prelude::*; -use pyo3::types::{IntoPyDict, PyDict, PyList, PyTuple}; -use std::collections::HashMap; -use zenoh::prelude::{ExprId, KeyExpr as ZKeyExpr, Selector, ZFuture, ZInt}; -/// A zenoh session. -#[pyclass] -pub struct Session { - s: Option<zenoh::Session>, -} +#![allow(clippy::borrow_deref_ref)] // false positives with pyo3 macros -#[pymethods] -impl Session { - /// Close the zenoh Session. - pub fn close(&mut self) -> PyResult<()> { - let s = self.take()?; - s.close().wait().map_err(to_pyerr) - } +use std::convert::TryInto; +use std::sync::Arc; - /// Get informations about the zenoh Session. - /// - /// :rtype: **dict[str, str]** - /// - /// :Example: - /// - /// >>> import zenoh - /// >>> s = zenoh.open() - /// >>> info = s.info() - /// >>> for key in info: - /// >>> print("{} : {}".format(key, info[key])) - pub fn info(&self, py: Python) -> PyResult<PyObject> { - use zenoh_cfg_properties::KeyTranscoder; - let s = self.as_ref()?; - let props = s.info().wait(); - let pydict: HashMap<String, String> = props - .0 - .into_iter() - .filter_map(|(k, v)| zenoh::info::InfoTranscoder::decode(k).map(|k| (k, v))) - .collect(); - Ok(pydict.into_py_dict(py).to_object(py)) - } +use pyo3::{prelude::*, types::PyDict}; +use zenoh::config::whatami::{WhatAmI, WhatAmIMatcher}; +use zenoh::prelude::SessionDeclarations; +use zenoh::publication::Publisher; +use zenoh::scouting::Scout; +use zenoh::subscriber::{PullSubscriber, Subscriber}; +use zenoh::Session; +use zenoh_core::SyncResolve; - /// Get informations about the zenoh Session. - /// - /// :rtype: dict {str: str} - /// - /// :Example: - /// - /// >>> import zenoh - /// >>> s = zenoh.open() - /// >>> info = s.info() - /// >>> for key in info: - /// >>> print("{} : {}".format(key, info[key])) - /// - /// - /// Get the current configuration of the zenoh Session. - /// - /// The returned ConfigNotifier can be used to read the current - /// zenoh configuration through the json function or - /// modify the zenoh configuration through the insert_json5 funtion. - /// - /// :rtype: dict {str: str} - /// - /// :Example: - /// - /// >>> import zenoh - /// >>> s = zenoh.open() - /// >>> config = s.config() - /// >>> config.insert_json5("connect/endpoints", "[\"tcp/10.10.10.10:7448\"]") - /// - pub fn config(&self) -> PyResult<ConfigNotifier> { - Ok(ConfigNotifier { - inner: self.as_ref()?.config().clone(), - }) - } +use crate::closures::PyClosure; +use crate::config::_Config; +use crate::enums::{ + _CongestionControl, _Priority, _QueryConsolidation, _QueryTarget, _Reliability, _SampleKind, +}; +use crate::keyexpr::{_KeyExpr, _Selector}; +use crate::queryable::{_Query, _Queryable}; +use crate::value::{_Hello, _Reply, _Sample, _Value, _ZenohId}; +use crate::{PyAnyToValue, PyExtract, ToPyErr}; - /// Put data. - /// - /// The *key_expr* parameter also accepts the following types that can be converted to a :class:`KeyExpr`: - /// - /// * **int** for a mapped key expression - /// * **str** for a literal key expression - /// * **(int, str)** for a mapped key expression with suffix - /// - /// :param key_expr: The key expression matching resources to write - /// :type key_expr: :class:`KeyExpr` - /// :param value: The value to write - /// :type value: any type convertible to a :class:`Value` - /// :param \**kwargs: - /// See below - /// - /// :Keyword Arguments: - /// * **encoding** (:class:`Encoding`) -- - /// Set the encoding of the written data - /// * **kind** ( **int** ) -- - /// Set the kind of the written data - /// * **congestion_control** (:class:`CongestionControl`) -- - /// Set the congestion control to apply when routing the data - /// * **priority** (:class:`Priority`) -- - /// Set the priority of the written data - /// * **local_routing** ( **bool** ) -- - /// Enable or disable local routing - /// - /// :raise: :class:`ZError` - /// - /// :Examples: - /// - /// >>> import zenoh - /// >>> s = zenoh.open() - /// >>> s.put('/key/expression', 'value') - #[pyo3(text_signature = "(self, key_expr, value, **kwargs)")] +#[pyclass(subclass)] +#[derive(Clone)] +pub struct _Session(pub(crate) Arc<Session>); + +#[pymethods] +impl _Session { + #[new] + pub fn new(config: Option<&mut crate::config::_Config>) -> PyResult<Self> { + let config = config.and_then(|c| c.0.take()).unwrap_or_default(); + let session = zenoh::open(config).res_sync().map_err(|e| e.to_pyerr())?; + Ok(_Session(Arc::new(session))) + } #[args(kwargs = "**")] - pub fn put(&self, key_expr: &PyAny, value: &PyAny, kwargs: Option<&PyDict>) -> PyResult<()> { - let s = self.as_ref()?; - let k = zkey_expr_of_pyany(key_expr)?; - let mut v = zvalue_of_pyany(value)?; - let mut encoding: Option<Encoding> = None; - let mut kind: Option<SampleKind> = None; - let mut congestion_control: Option<CongestionControl> = None; - let mut priority: Option<Priority> = None; - let mut local_routing: Option<bool> = None; + pub fn put( + &self, + key_expr: &crate::keyexpr::_KeyExpr, + value: &PyAny, + kwargs: Option<&PyDict>, + ) -> PyResult<()> { + let s = &self.0; + let k = &key_expr.0; + let v = value.to_value()?; + let mut builder = s.put(k, v); if let Some(kwargs) = kwargs { - if let Some(e) = kwargs.get_item("encoding") { - encoding = e.extract().ok() - } - if let Some(k) = kwargs.get_item("kind") { - kind = k.extract().ok() - } - if let Some(cc) = kwargs.get_item("congestion_control") { - congestion_control = cc.extract().ok() + match kwargs.extract_item::<_SampleKind>("kind") { + Ok(kind) => builder = builder.kind(kind.0), + Err(crate::ExtractError::Other(e)) => return Err(e), + _ => {} } - if let Some(p) = kwargs.get_item("priority") { - priority = p.extract().ok() + match kwargs.extract_item::<_CongestionControl>("congestion_control") { + Ok(congestion_control) => { + builder = builder.congestion_control(congestion_control.0) + } + Err(crate::ExtractError::Other(e)) => return Err(e), + _ => {} } - if let Some(lr) = kwargs.get_item("local_routing") { - local_routing = lr.extract().ok() + match kwargs.extract_item::<_Priority>("priority") { + Ok(priority) => builder = builder.priority(priority.0), + Err(crate::ExtractError::Other(e)) => return Err(e), + _ => {} } } - if let Some(encoding) = encoding { - v.encoding = encoding.into(); - } - let mut writer = s - .put(k, v) - .kind(kind.unwrap_or_default().kind) - .congestion_control(congestion_control.unwrap_or_default().cc) - .priority(priority.unwrap_or_default().p); - if let Some(local_routing) = local_routing { - writer = writer.local_routing(local_routing); - } - writer.wait().map_err(to_pyerr) + builder.res_sync().map_err(|e| e.to_pyerr()) } - - /// Delete data. - /// - /// The *key_expr* parameter also accepts the following types that can be converted to a :class:`KeyExpr`: - /// - /// * **int** for a mapped key expression - /// * **str** for a literal key expression - /// * **(int, str)** for a mapped key expression with suffix - /// - /// :param key_expr: The key expression matching resources to delete - /// :type key_expr: :class:`KeyExpr` - /// :param \**kwargs: - /// See below - /// - /// :Keyword Arguments: - /// * **congestion_control** (:class:`CongestionControl`) -- - /// Set the congestion control to apply when routing the data - /// * **priority** (:class:`Priority`) -- - /// Set the priority of the written data - /// * **local_routing** ( **bool** ) -- - /// Enable or disable local routing - /// - /// :raise: :class:`ZError` - /// - /// :Examples: - /// - /// >>> import zenoh - /// >>> s = zenoh.open() - /// >>> s.delete('/key/expression') - #[pyo3(text_signature = "(self, key_expr, **kwargs)")] #[args(kwargs = "**")] - pub fn delete(&self, key_expr: &PyAny, kwargs: Option<&PyDict>) -> PyResult<()> { - let s = self.as_ref()?; - let k = zkey_expr_of_pyany(key_expr)?; - let mut congestion_control: Option<CongestionControl> = None; - let mut priority: Option<Priority> = None; - let mut local_routing: Option<bool> = None; + pub fn delete( + &self, + key_expr: &crate::keyexpr::_KeyExpr, + kwargs: Option<&PyDict>, + ) -> PyResult<()> { + let s = &self.0; + let k = &key_expr.0; + let mut builder = s.delete(k); if let Some(kwargs) = kwargs { - if let Some(cc) = kwargs.get_item("congestion_control") { - congestion_control = cc.extract().ok() + match kwargs.extract_item::<_SampleKind>("kind") { + Ok(kind) => builder = builder.kind(kind.0), + Err(crate::ExtractError::Other(e)) => return Err(e), + _ => {} } - if let Some(p) = kwargs.get_item("priority") { - priority = p.extract().ok() + match kwargs.extract_item::<_CongestionControl>("congestion_control") { + Ok(congestion_control) => { + builder = builder.congestion_control(congestion_control.0) + } + Err(crate::ExtractError::Other(e)) => return Err(e), + _ => {} } - if let Some(lr) = kwargs.get_item("local_routing") { - local_routing = lr.extract().ok() + match kwargs.extract_item::<_Priority>("priority") { + Ok(priority) => builder = builder.priority(priority.0), + Err(crate::ExtractError::Other(e)) => return Err(e), + _ => {} } } - let mut writer = s - .delete(k) - .congestion_control(congestion_control.unwrap_or_default().cc) - .priority(priority.unwrap_or_default().p); - if let Some(local_routing) = local_routing { - writer = writer.local_routing(local_routing); - } - writer.wait().map_err(to_pyerr) - } - - /// Associate a numerical Id with the given key expression. - /// - /// This numerical Id will be used on the network to save bandwidth and - /// ease the retrieval of the concerned resource in the routing tables. - /// - /// The *key_expr* parameter also accepts the following types that can be converted to a :class:`KeyExpr`: - /// - /// * **int** for a mapped key expression - /// * **str** for a literal key expression - /// * **(int, str)** for a mapped key expression with suffix - /// - /// :param key_expr: The key expression to map to a numerical Id - /// :type key_expr: :class:`KeyExpr` - /// :rtype: **int** - /// :raise: :class:`ZError` - /// - /// :Examples: - /// - /// >>> import zenoh - /// >>> s = zenoh.open() - /// >>> rid = s.declare_expr('/key/expression') - #[pyo3(text_signature = "(self, key_expr)")] - pub fn declare_expr(&self, key_expr: &PyAny) -> PyResult<ExprId> { - let s = self.as_ref()?; - let k = zkey_expr_of_pyany(key_expr)?; - s.declare_expr(&k).wait().map_err(to_pyerr) + builder.res_sync().map_err(|e| e.to_pyerr()) } - /// Undeclare the *numerical Id/key expression* association previously declared - /// with :meth:`declare_expr`. - /// - /// :param rid: The numerical Id to unmap - /// :type rid: :class:`ExprId` - /// :raise: :class:`ZError` - /// - /// :Examples: - /// - /// >>> import zenoh - /// >>> s = zenoh.open() - /// >>> rid = s.declare_expr('/key/expression') - /// >>> s.undeclare_expr(rid) - #[pyo3(text_signature = "(self, rid)")] - pub fn undeclare_expr(&self, rid: ExprId) -> PyResult<()> { - let s = self.as_ref()?; - s.undeclare_expr(rid).wait().map_err(to_pyerr) + #[args(kwargs = "**")] + pub fn get( + &self, + selector: &_Selector, + callback: &PyAny, + kwargs: Option<&PyDict>, + ) -> PyResult<()> { + let callback: PyClosure<(_Reply,)> = <_ as TryInto<_>>::try_into(callback)?; + let mut builder = self.0.get(&selector.0).with(callback); + if let Some(kwargs) = kwargs { + match kwargs.extract_item::<_QueryConsolidation>("consolidation") { + Ok(_QueryConsolidation(Some(value))) => builder = builder.consolidation(value), + Err(crate::ExtractError::Other(e)) => return Err(e), + _ => {} + } + match kwargs.extract_item::<_QueryTarget>("target") { + Ok(value) => builder = builder.target(value.0), + Err(crate::ExtractError::Other(e)) => return Err(e), + _ => {} + } + } + builder.res_sync().map_err(|e| e.to_pyerr()) } - /// Declare a publication for the given key expression. - /// - /// Written expressions that match the given key expression will only be sent on the network - /// if matching subscribers exist in the system. - /// - /// The *key_expr* parameter also accepts the following types that can be converted to a :class:`KeyExpr`: - /// - /// * **int** for a mapped key expression - /// * **str** for a literal key expression - /// * **(int, str)** for a mapped key expression with suffix - /// - /// :param key_expr: The key expression to publish - /// :type key_expr: :class:`KeyExpr` - /// :raise: :class:`ZError` - /// - /// :Examples: - /// - /// >>> import zenoh - /// >>> s = zenoh.open() - /// >>> rid = s.declare_publication('/key/expression') - /// >>> s.put('/key/expression', bytes('value', encoding='utf8')) - #[pyo3(text_signature = "(self, key_expr)")] - fn declare_publication(&self, key_expr: &PyAny) -> PyResult<()> { - let s = self.as_ref()?; - let k = zkey_expr_of_pyany(key_expr)?; - s.declare_publication(&k).wait().map_err(to_pyerr)?; - Ok(()) - } - #[pyo3(text_signature = "(self, key_expr)")] - fn undeclare_publication(&self, key_expr: &PyAny) -> PyResult<()> { - let s = self.as_ref()?; - let k = zkey_expr_of_pyany(key_expr)?; - s.undeclare_publication(&k).wait().map_err(to_pyerr)?; - Ok(()) + pub fn declare_keyexpr(&self, key_expr: &_KeyExpr) -> PyResult<_KeyExpr> { + match self.0.declare_keyexpr(&key_expr.0).res_sync() { + Ok(k) => Ok(_KeyExpr(k.into_owned())), + Err(e) => Err(e.to_pyerr()), + } } - /// Create a Subscriber for the given key expression. - /// - /// The *key_expr* parameter also accepts the following types that can be converted to a :class:`KeyExpr`: - /// - /// * **int** for a mapped key expression - /// * **str** for a literal key expression - /// * **(int, str)** for a mapped key expression with suffix - /// - /// :param key_expr: The key expression to subscribe - /// :type key_expr: :class:`KeyExpr` - /// :param callback: the subscription callback - /// :type callback: function(:class:`Sample`) - /// :param \**kwargs: - /// See below - /// - /// :Keyword Arguments: - /// * **reliability** (:class:`Reliability`) -- - /// Set the subscription reliability (BestEffort by default) - /// * **mode** (:class:`SubMode`) -- - /// Set the subscription mode (Push by default) - /// * **period** (:class:`Period`) -- - /// Set the subscription period - /// * **local** ( **bool** ) -- - /// If true make the subscription local only (false by default) - /// - /// :rtype: :class:`Subscriber` - /// :raise: :class:`ZError` - /// - /// :Examples: - /// - /// >>> import zenoh, time - /// >>> from zenoh import Reliability, SubMode - /// >>> - /// >>> s = zenoh.open() - /// >>> sub = s.subscribe('/key/expression', - /// ... lambda sample: print("Received : {}".format(sample)), - /// ... reliability=Reliability.Reliable, - /// ... mode=SubMode.Push) - /// >>> time.sleep(60) - #[pyo3(text_signature = "(self, key_expr, callback, **kwargs)")] #[args(kwargs = "**")] - fn subscribe( + pub fn declare_queryable( &self, - key_expr: &PyAny, + key_expr: _KeyExpr, callback: &PyAny, kwargs: Option<&PyDict>, - ) -> PyResult<Subscriber> { - let s = self.as_ref()?; - let k = zkey_expr_of_pyany(key_expr)?; - let mut sub_builder = s.subscribe(&k); + ) -> PyResult<_Queryable> { + let callback: PyClosure<(_Query,)> = <_ as TryInto<_>>::try_into(callback)?; + let mut builder = self.0.declare_queryable(key_expr.0).with(callback); if let Some(kwargs) = kwargs { - if let Some(arg) = kwargs.get_item("reliability") { - sub_builder = sub_builder.reliability(arg.extract::<Reliability>()?.r); + match kwargs.extract_item::<bool>("complete") { + Ok(value) => builder = builder.complete(value), + Err(crate::ExtractError::Other(e)) => return Err(e), + _ => {} } - if let Some(arg) = kwargs.get_item("mode") { - sub_builder = sub_builder.mode(arg.extract::<SubMode>()?.m); - } - if let Some(arg) = kwargs.get_item("period") { - sub_builder = sub_builder.period(Some(arg.extract::<Period>()?.p)); + } + match builder.res_sync() { + Ok(o) => Ok(_Queryable(o)), + Err(e) => Err(e.to_pyerr()), + } + } + + #[args(kwargs = "**")] + pub fn declare_publisher( + &self, + key_expr: _KeyExpr, + kwargs: Option<&PyDict>, + ) -> PyResult<_Publisher> { + let mut builder = self.0.declare_publisher(key_expr.0); + if let Some(kwargs) = kwargs { + match kwargs.extract_item::<_Priority>("priority") { + Ok(value) => builder = builder.priority(value.0), + Err(crate::ExtractError::Other(e)) => return Err(e), + _ => {} } - if let Some(arg) = kwargs.get_item("local") { - if arg.extract::<bool>()? { - sub_builder = sub_builder.local(); - } + match kwargs.extract_item::<_CongestionControl>("congestion_control") { + Ok(value) => builder = builder.congestion_control(value.0), + Err(crate::ExtractError::Other(e)) => return Err(e), + _ => {} } } - let zn_sub = sub_builder.wait().map_err(to_pyerr)?; - // Note: workaround to allow moving of zn_sub into the task below. - // Otherwise, s is moved also, but can't because it doesn't have 'static lifetime. - let mut static_zn_sub = unsafe { - std::mem::transmute::< - zenoh::subscriber::Subscriber<'_>, - zenoh::subscriber::Subscriber<'static>, - >(zn_sub) - }; - - // Note: callback cannot be passed as such in task below because it's not Send - let cb_obj: Py<PyAny> = callback.into(); - - let (unregister_tx, unregister_rx) = bounded::<ZnSubOps>(8); - // Note: This is done to ensure that even if the call-back into Python - // does any blocking call we do not incour the risk of blocking - // any of the task resolving futures. - let loop_handle = task::spawn_blocking(move || { - task::block_on(async move { - loop { - select!( - s = static_zn_sub.receiver().next().fuse() => { - // Acquire Python GIL to call the callback - let gil = Python::acquire_gil(); - let py = gil.python(); - let cb_args = PyTuple::new(py, &[Sample { s: s.unwrap() }]); - if let Err(e) = cb_obj.as_ref(py).call1(cb_args) { - warn!("Error calling subscriber callback:"); - e.print(py); - } - }, - op = unregister_rx.recv().fuse() => { - match op { - Ok(ZnSubOps::Pull) => { - if let Err(e) = static_zn_sub.pull().await { - warn!("Error pulling the subscriber: {}", e); - } - }, - Ok(ZnSubOps::Unregister) => { - if let Err(e) = static_zn_sub.close().await { - warn!("Error undeclaring subscriber: {}", e); - } - return - }, - _ => return - } - } - ) - } - }) - }); - Ok(Subscriber { - unregister_tx, - loop_handle: Some(loop_handle), - }) + match builder.res_sync() { + Ok(o) => Ok(_Publisher(o)), + Err(e) => Err(e.to_pyerr()), + } } - /// Create a Queryable for the given key expression. - /// - /// The *key_expr* parameter also accepts the following types that can be converted to a :class:`KeyExpr`: - /// - /// * **int** for a mapped key expression - /// * **str** for a literal key expression - /// * **(int, str)** for a mapped key expression with suffix - /// - /// :param key_expr: The key expression the Queryable will reply to - /// :type key_expr: :class:`KeyExpr` - /// :param callback: the queryable callback - /// :type callback: function(:class:`Query`) - /// :param \**kwargs: - /// See below - /// - /// :Keyword Arguments: - /// * **kind** ( **int** ) -- - /// Set the queryable kind. This must be a mask of constants defined in :mod:`zenoh.queryable`) - /// (`queryable.EVAL` by default) - /// * **complete** ( **bool** ) -- - /// Set the queryable completeness (true by default) - /// - /// :rtype: :class:`Queryable` - /// :raise: :class:`ZError` - /// - /// :Examples: - /// - /// >>> import zenoh, time - /// >>> from zenoh import Sample, queryable - /// >>> def callback(query): - /// ... print("Received : {}".format(query)) - /// ... query.reply(Sample('/key/expression', bytes('value', encoding='utf8'))) - /// >>> - /// >>> s = zenoh.open() - /// >>> q = s.queryable('/key/expression', callback, kind=queryable.EVAL) - /// >>> time.sleep(60) - #[pyo3(text_signature = "(self, key_expr, callback, **kwargs)")] #[args(kwargs = "**")] - fn queryable( + pub fn declare_subscriber( &self, - key_expr: &PyAny, + key_expr: &_KeyExpr, callback: &PyAny, kwargs: Option<&PyDict>, - ) -> PyResult<Queryable> { - let s = self.as_ref()?; - let k = zkey_expr_of_pyany(key_expr)?; - let mut builder = s.queryable(k); + ) -> PyResult<_Subscriber> { + let callback: PyClosure<(_Sample,)> = <_ as TryInto<_>>::try_into(callback)?; + let mut builder = self.0.declare_subscriber(&key_expr.0).with(callback); if let Some(kwargs) = kwargs { - if let Some(arg) = kwargs.get_item("kind") { - builder = builder.kind(arg.extract::<ZInt>()?); - } - if let Some(arg) = kwargs.get_item("complete") { - builder = builder.complete(arg.extract::<bool>()?); + match kwargs.extract_item::<_Reliability>("reliability") { + Ok(reliabilty) => builder = builder.reliability(reliabilty.0), + Err(crate::ExtractError::Other(e)) => return Err(e), + _ => {} } } - let zn_quer = builder.wait().map_err(to_pyerr)?; - - // Note: workaround to allow moving of zn_quer into the task below. - // Otherwise, s is moved also, but can't because it doesn't have 'static lifetime. - let mut zn_quer = unsafe { - std::mem::transmute::< - zenoh::queryable::Queryable<'_>, - zenoh::queryable::Queryable<'static>, - >(zn_quer) - }; - - // Note: callback cannot be passed as such in task below because it's not Send - let cb_obj: Py<PyAny> = callback.into(); - - let (unregister_tx, unregister_rx) = bounded::<bool>(1); - // Note: This is done to ensure that even if the call-back into Python - // does any blocking call we do not incour the risk of blocking - // any of the task resolving futures. - let loop_handle = task::spawn_blocking(move || { - task::block_on(async move { - loop { - select!( - q = zn_quer.receiver().next().fuse() => { - // Acquire Python GIL to call the callback - let gil = Python::acquire_gil(); - let py = gil.python(); - let cb_args = PyTuple::new(py, &[Query { q: async_std::sync::Arc::new(q.unwrap()) }]); - if let Err(e) = cb_obj.as_ref(py).call1(cb_args) { - warn!("Error calling queryable callback:"); - e.print(py); - } - }, - _ = unregister_rx.recv().fuse() => { - if let Err(e) = zn_quer.close().await { - warn!("Error undeclaring queryable: {}", e); - } - return - } - ) - } - }) - }); - Ok(Queryable { - unregister_tx, - loop_handle: Some(loop_handle), - }) + let subscriber = builder.res().map_err(|e| e.to_pyerr())?; + Ok(_Subscriber(subscriber)) } - /// Query data from the matching queryables in the system. - /// - /// Replies are collected in a list. - /// - /// The *selector* parameter also accepts the following types that can be converted to a :class:`Selector`: - /// - /// * **KeyExpr** for a key expression with no value selector - /// * **int** for a key expression id with no value selector - /// * **str** for a litteral selector - /// - /// :param selector: The selection of resources to query - /// :type selector: :class:`Selector` - /// :param \**kwargs: - /// See below - /// - /// :Keyword Arguments: - /// * **target** (:class:`QueryTarget`) -- - /// Set the kind of queryables that should be target of this query - /// * **consolidation** (:class:`QueryConsolidation`) -- - /// Set the consolidation mode of the query - /// * **local_routing** ( **bool** ) -- - /// Enable or disable local routing - /// - /// :rtype: [:class:`Reply`] - /// :raise: :class:`ZError` - /// - /// :Examples: - /// - /// >>> import zenoh, time - /// >>> - /// >>> s = zenoh.open() - /// >>> replies = s.get('/key/selector?value_selector') - /// >>> for reply in replies: - /// ... print("Received : {}".format(reply.sample)) - #[pyo3(text_signature = "(self, selector, **kwargs)")] #[args(kwargs = "**")] - fn get(&self, selector: &PyAny, kwargs: Option<&PyDict>) -> PyResult<Py<PyList>> { - let s = self.as_ref()?; - task::block_on(async { - let selector: Selector = match selector.get_type().name()? { - "KeyExpr" => { - let key_expr: PyRef<KeyExpr> = selector.extract()?; - key_expr.inner.clone().into() - } - "int" => { - let id: u64 = selector.extract()?; - ZKeyExpr::from(id).into() - } - "str" => { - let name: &str = selector.extract()?; - Selector::from(name) - } - x => { - return Err(PyErr::new::<exceptions::PyValueError, _>(format!( - "Cannot convert type '{}' to a zenoh Selector", - x - ))) - } - }; - let mut getter = s.get(selector); - if let Some(kwargs) = kwargs { - if let Some(arg) = kwargs.get_item("target") { - getter = getter.target(arg.extract::<QueryTarget>()?.t); - } - if let Some(arg) = kwargs.get_item("consolidation") { - getter = getter.consolidation(arg.extract::<QueryConsolidation>()?.c); - } - if let Some(arg) = kwargs.get_item("local_routing") { - getter = getter.local_routing(arg.extract::<bool>()?); - } - } - let mut replies = getter.wait().map_err(to_pyerr)?; - let gil = Python::acquire_gil(); - let py = gil.python(); - let result = PyList::empty(py); - while let Some(reply) = replies.next().await { - result.append(Reply { r: reply })?; + pub fn declare_pull_subscriber( + &self, + key_expr: &_KeyExpr, + callback: &PyAny, + kwargs: Option<&PyDict>, + ) -> PyResult<_PullSubscriber> { + let callback: PyClosure<(_Sample,)> = <_ as TryInto<_>>::try_into(callback)?; + let mut builder = self + .0 + .declare_subscriber(&key_expr.0) + .pull_mode() + .with(callback); + if let Some(kwargs) = kwargs { + match kwargs.extract_item::<_Reliability>("reliability") { + Ok(reliabilty) => builder = builder.reliability(reliabilty.0), + Err(crate::ExtractError::Other(e)) => return Err(e), + _ => {} } - Ok(result.into()) - }) + } + let subscriber = builder.res().map_err(|e| e.to_pyerr())?; + Ok(_PullSubscriber(subscriber)) + } + + pub fn zid(&self) -> _ZenohId { + _ZenohId(self.0.zid()) + } + pub fn routers_zid(&self) -> Vec<_ZenohId> { + self.0 + .info() + .routers_zid() + .res_sync() + .map(_ZenohId) + .collect() + } + pub fn peers_zid(&self) -> Vec<_ZenohId> { + self.0.info().peers_zid().res_sync().map(_ZenohId).collect() } } -impl Session { - pub(crate) fn new(s: zenoh::Session) -> Self { - Session { s: Some(s) } +#[pyclass(subclass)] +#[derive(Clone)] +pub struct _Publisher(Publisher<'static>); +#[pymethods] +impl _Publisher { + #[new] + pub fn pynew(this: Self) -> Self { + this + } + #[getter] + pub fn key_expr(&self) -> _KeyExpr { + _KeyExpr(self.0.key_expr().clone()) + } + pub fn put(&self, value: _Value) -> PyResult<()> { + self.0.put(value).res_sync().map_err(|e| e.to_pyerr()) + } + pub fn delete(&self) -> PyResult<()> { + self.0.delete().res_sync().map_err(|e| e.to_pyerr()) } +} + +#[pyclass(subclass)] +pub struct _Subscriber(Subscriber<'static, ()>); - #[inline] - fn as_ref(&self) -> PyResult<&zenoh::Session> { - self.s - .as_ref() - .ok_or_else(|| PyErr::new::<ZError, _>("zenoh session was closed")) +#[pyclass(subclass)] +pub struct _PullSubscriber(PullSubscriber<'static, ()>); +#[pymethods] +impl _PullSubscriber { + fn pull(&self) -> PyResult<()> { + self.0.pull().res_sync().map_err(|e| e.to_pyerr()) } +} + +#[pyclass(subclass)] +pub struct _Scout(Scout<()>); - #[inline] - fn take(&mut self) -> PyResult<zenoh::Session> { - self.s - .take() - .ok_or_else(|| PyErr::new::<ZError, _>("zenoh session was closed")) +#[pyfunction] +pub fn scout(callback: &PyAny, config: Option<&_Config>, what: Option<&str>) -> PyResult<_Scout> { + let callback: PyClosure<(_Hello,)> = <_ as TryInto<_>>::try_into(callback)?; + let what: WhatAmIMatcher = match what { + None => WhatAmI::Client | WhatAmI::Peer | WhatAmI::Router, + Some(s) => match s.parse() { + Ok(w) => w, + Err(_) => return Err(zenoh_core::zerror!("Couldn't parse `{}` into a WhatAmiMatcher: must be a `|`-separated list of `peer`, `client` or `router`", s).to_pyerr()) + }, + }; + let config = config.and_then(|c| c.0.clone()).unwrap_or_default(); + let scout = zenoh::scout(what, config).with(callback).res_sync(); + match scout { + Ok(scout) => Ok(_Scout(scout)), + Err(e) => Err(e.to_pyerr()), } } diff --git a/src/types.rs b/src/types.rs deleted file mode 100644 index 664d547c..00000000 --- a/src/types.rs +++ /dev/null @@ -1,1565 +0,0 @@ -use std::collections::HashMap; -use std::ops::BitOr; - -// -// Copyright (c) 2017, 2022 ZettaScale Technology Inc. -// -// This program and the accompanying materials are made available under the -// terms of the Eclipse Public License 2.0 which is available at -// http://www.eclipse.org/legal/epl-2.0, or the Apache License, Version 2.0 -// which is available at https://www.apache.org/licenses/LICENSE-2.0. -// -// SPDX-License-Identifier: EPL-2.0 OR Apache-2.0 -// -// Contributors: -// ZettaScale Zenoh team, <[email protected]> -// -use crate::encoding::Encoding; -use crate::sample_kind::SampleKind; -use crate::to_pyerr; -use async_std::channel::Sender; -use async_std::task; -use log::warn; -use pyo3::exceptions; -use pyo3::number::PyNumberOrProtocol; -use pyo3::prelude::*; -use pyo3::types::{PyBytes, PyTuple}; -use pyo3::PyObjectProtocol; -use zenoh::config::whatami::WhatAmIMatcher; -use zenoh::config::WhatAmI as ZWhatAmI; -use zenoh::prelude::{ - Encoding as ZEncoding, KeyExpr as ZKeyExpr, KnownEncoding as ZKnownEncoding, - Selector as ZSelector, Value as ZValue, ZInt, -}; -use zenoh_buffers::traits::SplitBuffer; - -// zenoh.config (simulate the package as a class, and consts as class attributes) -// -/// The following constants define the several configuration keys accepted for a zenoh -/// session configuration and the associated accepted values. -#[allow(non_camel_case_types)] -#[pyclass] -pub(crate) struct config {} - -#[allow(non_snake_case)] -#[pymethods] -impl config { - /// The library mode. - /// - /// - Accepted values : `"peer"`, `"client"`. - /// - Default value : `"peer"`. - #[classattr] - pub fn MODE_KEY() -> &'static str { - "mode" - } - - /// The locator of a peer to connect to. - /// - /// - Accepted values : `<locator>` (ex: `"tcp/10.10.10.10:7447"`). - /// - Default value : None. - /// - Multiple values accepted. - #[classattr] - pub fn CONNECT_KEY() -> &'static str { - "connect/endpoints" - } - - /// A locator to listen on. - /// - /// - Accepted values : `<locator>` (ex: `"tcp/10.10.10.10:7447"`). - /// - Default value : None. - /// - Multiple values accepted. - #[classattr] - pub fn LISTEN_KEY() -> &'static str { - "listen/endpoints" - } - - /// The user name to use for authentication. - /// - /// - Accepted values : `<string>`. - /// - Default value : None. - #[classattr] - pub fn USER_KEY() -> &'static str { - "transport/auth/usrpwd/user" - } - - /// The password to use for authentication. - /// - /// - Accepted values : `<string>`. - /// - Default value : None. - #[classattr] - fn PASSWORD_KEY() -> &'static str { - "transport/auth/usrpwd/password" - } - - /// Activates/Desactivates multicast scouting. - /// - /// - Accepted values : `"true"`, `"false"`. - /// - Default value : `"true"`. - #[classattr] - pub fn MULTICAST_SCOUTING_KEY() -> &'static str { - "scouting/multicast/enabled" - } - - /// The network interface to use for multicast scouting. - /// - /// - Accepted values : `"auto"`, `<ip address>`, `<interface name>`. - /// - Default value : `"auto"`. - #[classattr] - pub fn MULTICAST_INTERFACE_KEY() -> &'static str { - "scouting/multicast/interface" - } - - /// The multicast address and ports to use for multicast scouting. - /// - /// - Accepted values : `<ip address>:<port>`. - /// - Default value : `"224.0.0.224:7447"`. - #[classattr] - pub fn MULTICAST_IPV4_ADDRESS_KEY() -> &'static str { - "scouting/multicast/address" - } - - /// In client mode, the period dedicated to scouting a router before failing. - /// - /// - Accepted values : `<float in seconds>`. - /// - Default value : `"3.0"`. - #[classattr] - pub fn SCOUTING_TIMEOUT_KEY() -> &'static str { - "scouting/timeout" - } - - /// In peer mode, the period dedicated to scouting first remote peers before doing anything else. - /// - /// - Accepted values : `<float in seconds>`. - /// - Default value : `"0.2"`. - #[classattr] - pub fn SCOUTING_DELAY_KEY() -> &'static str { - "scouting/delay" - } - - /// Indicates if data messages should be timestamped. - /// - /// - Accepted values : `"true"`, `"false"`. - /// - Default value : `"false"`. - #[classattr] - pub fn ADD_TIMESTAMP_KEY() -> &'static str { - "add_timestamp" - } - - /// Indicates if local writes/queries should reach local subscribers/queryables. - #[classattr] - pub fn LOCAL_ROUTING_KEY() -> &'static str { - "local_routing" - } -} - -// zenoh.info (simulate the package as a class, and consts as class attributes) -/// Constants and helpers to interpret the properties returned by :func:`zenoh.Session.info`. -#[allow(non_camel_case_types)] -#[pyclass] -pub(crate) struct info {} - -#[allow(non_snake_case)] -#[pymethods] -impl info { - #[classattr] - fn ZN_INFO_PID_KEY() -> ZInt { - zenoh::info::ZN_INFO_PID_KEY - } - - #[classattr] - fn ZN_INFO_PEER_PID_KEY() -> ZInt { - zenoh::info::ZN_INFO_PEER_PID_KEY - } - - #[classattr] - fn ZN_INFO_ROUTER_PID_KEY() -> ZInt { - zenoh::info::ZN_INFO_ROUTER_PID_KEY - } -} - -// zenoh.whatami (simulate the package as a class, and consts as class attributes) -/// Constants defining the different zenoh process to look for with :func:`zenoh.scout`. -#[allow(non_camel_case_types)] -#[pyclass] -#[derive(Debug, Clone, Copy, PartialEq, Eq, Hash)] -pub struct WhatAmI { - inner: WhatAmIMatcher, -} -impl From<WhatAmI> for WhatAmIMatcher { - fn from(w: WhatAmI) -> Self { - w.inner - } -} -impl BitOr for WhatAmI { - type Output = Self; - fn bitor(self, rhs: Self) -> Self::Output { - WhatAmI { - inner: self.inner | rhs.inner, - } - } -} -#[pyproto] -impl pyo3::PyNumberProtocol for WhatAmI -where - <Self as PyNumberOrProtocol<'p>>::Left: BitOr< - <Self as PyNumberOrProtocol<'p>>::Right, - Output = <Self as PyNumberOrProtocol<'p>>::Result, - >, -{ - fn __or__(lhs: Self, rhs: Self) -> Self - where - Self: PyNumberOrProtocol<'p>, - { - lhs | rhs - } -} - -#[allow(non_snake_case)] -#[pymethods] -impl WhatAmI { - #[classattr] - fn Router() -> Self { - WhatAmI { - inner: ZWhatAmI::Router.into(), - } - } - - #[classattr] - fn Peer() -> Self { - WhatAmI { - inner: ZWhatAmI::Peer.into(), - } - } - - #[classattr] - fn Client() -> Self { - WhatAmI { - inner: ZWhatAmI::Client.into(), - } - } -} - -impl std::fmt::Display for WhatAmI { - fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result { - write!(f, "{}", self.inner.to_str()) - } -} - -/// A Hello message received as a response to a :meth:`scout` -#[pyclass] -#[derive(Clone)] -pub(crate) struct Hello { - pub(crate) h: zenoh::scouting::Hello, -} - -#[pymethods] -impl Hello { - /// The PeerId of the Hello message sender - /// - /// :type: :class:`PeerId` or `None` - #[getter] - fn pid(&self) -> Option<PeerId> { - self.h.pid.as_ref().map(|p| PeerId { p: *p }) - } - - /// The mode of the Hello message sender (bitmask of constants from :class:`whatami`) - /// - /// :type: :class:`whatami` or `None` - #[getter] - fn whatami(&self) -> Option<WhatAmI> { - self.h.whatami.map(|w| WhatAmI { inner: w.into() }) - } - - /// The locators list of the Hello message sender - /// - /// :type: list of str or `None` - #[getter] - fn locators(&self) -> Option<Vec<String>> { - self.h - .locators - .as_ref() - .map(|v| v.iter().map(|l| l.to_string()).collect()) - } -} - -#[pyproto] -impl PyObjectProtocol for Hello { - fn __str__(&self) -> String { - self.h.to_string() - } -} - -// zenoh.resource_name (simulate the package as a class with static methodss) -#[allow(non_camel_case_types)] -#[pyclass] -pub struct KeyExpr { - pub(crate) inner: ZKeyExpr<'static>, -} - -#[allow(non_snake_case)] -#[pymethods] -impl KeyExpr { - /// Return true if both resource names intersect. - /// - /// :param s1: the 1st resource name - /// :type s1: str - /// :param s2: the 2nd resource name - /// :type s2: str - #[staticmethod] - #[pyo3(text_signature = "(s1, s2)")] - fn intersect(s1: &PyAny, s2: &PyAny) -> bool { - let s1 = zkey_expr_of_pyany(s1).unwrap(); - let s2 = zkey_expr_of_pyany(s2).unwrap(); - match (s1.as_id_and_suffix(), s2.as_id_and_suffix()) { - ((s1, _), (s2, _)) if s1 != s2 => false, - ((_, s1), (_, s2)) => zenoh::utils::key_expr::intersect(s1, s2), - } - } -} - -#[pyproto] -impl PyObjectProtocol for KeyExpr { - fn __str__(&self) -> PyResult<String> { - Ok(self.inner.to_string()) - } -} - -impl From<KeyExpr> for ZKeyExpr<'static> { - fn from(r: KeyExpr) -> ZKeyExpr<'static> { - r.inner - } -} - -impl From<ZKeyExpr<'static>> for KeyExpr { - fn from(inner: ZKeyExpr<'static>) -> KeyExpr { - KeyExpr { inner } - } -} - -pub(crate) fn zkey_expr_of_pyany(obj: &PyAny) -> PyResult<ZKeyExpr> { - match obj.get_type().name()? { - "KeyExpr" => { - let rk: PyRef<KeyExpr> = obj.extract()?; - Ok(rk.inner.clone()) - } - "int" => { - let id: u64 = obj.extract()?; - Ok(id.into()) - } - "str" => { - let name: String = obj.extract()?; - Ok(name.into()) - } - "tuple" => { - let tuple: &PyTuple = obj.downcast()?; - if tuple.len() == 2 - && tuple.get_item(0)?.get_type().name()? == "int" - && tuple.get_item(1)?.get_type().name()? == "str" - { - let id: u64 = tuple.get_item(0)?.extract()?; - let suffix: String = tuple.get_item(1)?.extract()?; - Ok(ZKeyExpr::from(id).with_suffix(&suffix).to_owned()) - } else { - Err(PyErr::new::<exceptions::PyValueError, _>(format!( - "Cannot convert type '{:?}' to a zenoh-net KeyExpr", - tuple - ))) - } - } - x => Err(PyErr::new::<exceptions::PyValueError, _>(format!( - "Cannot convert type '{}' to a zenoh-net KeyExpr", - x - ))), - } -} - -/// An expression identifying a selection of resources. -/// -/// A selector is the conjunction of an key expression identifying a set -/// of resource keys and a value selector filtering out the resource values. -/// -/// Structure of a selector::: -/// -/// /s1/s2/..../sn?x>1&y<2&...&z=4(p1=v1;p2=v2;...;pn=vn)[a;b;x;y;...;z] -/// |key_selector||---------------- value_selector --------------------| -/// |--- filter --| |---- properties ---| |--fragment-| -/// -/// where: -/// * **key_selector**: an expression identifying a set of Resources. -/// * **filter**: a list of `value_selectors` separated by `'&'` allowing to perform filtering on the values -/// associated with the matching keys. Each `value_selector` has the form "`field`-`operator`-`value`" value where: -/// -/// * *field* is the name of a field in the value (is applicable and is existing. otherwise the `value_selector` is false) -/// * *operator* is one of a comparison operators: `<` , `>` , `<=` , `>=` , `=` , `!=` -/// * *value* is the the value to compare the field’s value with -/// -/// * **fragment**: a list of fields names allowing to return a sub-part of each value. -/// This feature only applies to structured values using a “self-describing” encoding, such as JSON or XML. -/// It allows to select only some fields within the structure. A new structure with only the selected fields -/// will be used in place of the original value. -/// -/// *NOTE: the filters and fragments are not yet supported in current zenoh version.* -#[allow(non_camel_case_types)] -#[pyclass] -pub struct Selector { - pub(crate) s: ZSelector<'static>, -} - -#[allow(non_snake_case)] -#[pymethods] -impl Selector { - /// The part of this selector identifying which keys should be part of the selection. - /// I.e. all characters before `?`. - /// - /// :type: :class:`KeyExpr` - #[getter] - fn key_selector(&self) -> KeyExpr { - KeyExpr { - inner: self.s.key_selector.to_owned(), - } - } - - /// the part of this selector identifying which values should be part of the selection. - /// I.e. all characters starting from `?`. - /// - /// :type: str - #[getter] - fn value_selector(&self) -> &str { - self.s.value_selector.as_ref() - } - - /// Parses the `value_selector` part of this `Selector`. - /// - /// :rtype: :class:`ValueSelector` - fn parse_value_selector(&self) -> PyResult<ValueSelector> { - let zvs = self.s.parse_value_selector().map_err(to_pyerr)?; - Ok(ValueSelector { - filter: zvs.filter.to_owned(), - properties: zvs.properties.0, - fragment: zvs.fragment.map(|cow| cow.to_owned()), - }) - } -} - -#[pyproto] -impl PyObjectProtocol for Selector { - fn __str__(&self) -> String { - self.s.to_string() - } -} - -impl From<Selector> for ZSelector<'static> { - fn from(s: Selector) -> ZSelector<'static> { - s.s - } -} - -impl From<ZSelector<'static>> for Selector { - fn from(s: ZSelector<'static>) -> Selector { - Selector { s } - } -} - -/// A class that can be used to help decoding or encoding the `value_selector` part of a :class:`Selector`. -#[allow(non_camel_case_types)] -#[pyclass] -pub struct ValueSelector { - pub(crate) filter: String, - pub(crate) properties: HashMap<String, String>, - pub(crate) fragment: Option<String>, -} - -#[allow(non_snake_case)] -#[pymethods] -impl ValueSelector { - /// the filter part of this `ValueSelector`, if any (all characters after `?` and before `(` or `[`) - /// - /// :type: str - #[getter] - fn filter(&self) -> &str { - self.filter.as_ref() - } - - /// the properties part of this `ValueSelector`) (all characters between ``( )`` and after `?`) - /// - /// :type: str - #[getter] - fn properties(&self) -> HashMap<String, String> { - self.properties.clone() - } - - /// the filter part of this `ValueSelector`, if any (all characters after `?` and before `(` or `[`) - /// - /// :type: str - #[getter] - fn fragment(&self) -> Option<&str> { - self.fragment.as_ref().map(|s| s.as_ref()) - } -} - -/// A Peer id -#[pyclass] -pub(crate) struct PeerId { - pub(crate) p: zenoh::prelude::PeerId, -} - -#[pyproto] -impl PyObjectProtocol for PeerId { - fn __str__(&self) -> String { - self.p.to_string() - } -} - -/// A zenoh Value, consisting of a payload (bytes) and an :class:`Encoding`. -/// -/// It can be created directly from the supported primitive types. -/// The value is automatically encoded in the payload and the Encoding is set accordingly. -/// -/// Or it can be created from a tuple **(payload, encoding)**, where: -/// -/// - payload has type **bytes** or **str** (the string is automatically converted into bytes) -/// - encoding has type :class:`Encoding` -/// -/// :Examples: -/// -/// >>> import json, zenoh -/// >>> from zenoh import Encoding, Value -/// >>> -/// >>> string_value = Value('Hello World!') -/// >>> int_value = Value(42) -/// >>> float_value = Value(3.14) -/// >>> bytes_value = Value(b'\x48\x69\x21') -/// >>> properties_value = Value({'p1': 'v1', 'p2': 'v2'}) -/// >>> -/// >>> json_value = Value((json.dumps(['foo', {'bar': ('baz', None, 1.0, 2)}]), Encoding.TEXT_JSON)) -/// >>> xml_value = Value(('<foo>bar</foo>', Encoding.TEXT_XML)) -/// >>> custom_value = Value((b'\x48\x69\x21', Encoding.APP_CUSTOM.with_suffix('my_encoding'))) -#[pyclass] -#[derive(Clone)] -pub struct Value { - pub(crate) v: ZValue, -} -impl From<Value> for ZValue { - fn from(v: Value) -> Self { - v.v - } -} -impl From<ZValue> for Value { - fn from(v: ZValue) -> Self { - Value { v } - } -} - -trait IntoPyAlt { - fn into_py_alt(self, py: Python) -> U; -} - -impl IntoPyAlt<PyObject> for serde_json::Value { - fn into_py_alt(self, py: Python) -> PyObject { - match self { - serde_json::Value::Null => py.None(), - serde_json::Value::Bool(v) => v.into_py(py), - serde_json::Value::Number(v) => v.into_py_alt(py), - serde_json::Value::String(v) => v.into_py(py), - serde_json::Value::Array(a) => a - .into_iter() - .map(|v| v.into_py_alt(py)) - .collect::<Vec<_>>() - .into_py(py), - serde_json::Value::Object(m) => m - .into_iter() - .map(|(k, v)| (k, v.into_py_alt(py))) - .collect::<std::collections::HashMap<_, _>>() - .into_py(py), - } - } -} -impl IntoPyAlt<PyObject> for serde_json::Number { - fn into_py_alt(self, py: Python) -> PyObject { - if let Some(v) = self.as_u64() { - return v.into_py(py); - } - if let Some(v) = self.as_i64() { - return v.into_py(py); - } - if let Some(v) = self.as_f64() { - return v.into_py(py); - } - unreachable!() - } -} - -#[allow(non_snake_case)] -#[pymethods] -impl Value { - #[new] - fn new(any: &PyAny) -> PyResult<Self> { - Ok(Value { - v: zvalue_of_pyany(any)?, - }) - } - - /// the payload the Value. - /// - /// :type: **bytes** - #[getter] - fn payload<'a>(&self, py: Python<'a>) -> &'a PyBytes { - PyBytes::new(py, self.v.payload.contiguous().as_ref()) - } - - /// the encoding of the Value. - /// - /// :type: :class:`Encoding` - #[getter] - fn encoding(&self) -> PyResult<Encoding> { - Ok(self.v.encoding.clone().into()) - } - - /// Try to decode the value's payload according to it's encoding, and return a typed object or primitive. - /// - /// :rtype: depend on the encoding flag (e.g. str for a StringUtf8 Value, int for an Integer Value ...) - fn decode(&self, py: Python) -> PyResult<PyObject> { - match self.v.encoding.prefix() { - ZKnownEncoding::Empty | ZKnownEncoding::AppOctetStream => { - Ok(self.v.payload.contiguous().into_py(py)) - } - ZKnownEncoding::TextPlain => { - Ok(String::from_utf8_lossy(&self.v.payload.contiguous()).into_py(py)) - } - ZKnownEncoding::AppProperties => self - .v - .as_properties() - .map(|v| v.0.into_py(py)) - .ok_or_else(|| { - exceptions::PyTypeError::new_err( - "Failed to decode Value's payload as Properties", - ) - }), - ZKnownEncoding::AppJson | ZKnownEncoding::TextJson => self - .v - .as_json() - .map(|v: serde_json::Value| v.into_py_alt(py)) - .ok_or_else(|| { - exceptions::PyTypeError::new_err("Failed to decode Value's payload as JSON") - }), - ZKnownEncoding::AppInteger => self - .v - .as_integer() - .map(|v: i64| v.into_py(py)) - .ok_or_else(|| { - exceptions::PyTypeError::new_err("Failed to decode Value's payload as Integer") - }), - ZKnownEncoding::AppFloat => { - self.v - .as_float() - .map(|v: f64| v.into_py(py)) - .ok_or_else(|| { - exceptions::PyTypeError::new_err( - "Failed to decode Value's payload as Float", - ) - }) - } - _ => Err(exceptions::PyTypeError::new_err(format!( - "Don't know how to decode Value's payload with encoding: {}", - self.v.encoding - ))), - } - } -} - -#[pyproto] -impl PyObjectProtocol for Value { - fn __str__(&self) -> PyResult<String> { - Ok(format!("{:?}", self.v)) - } - - fn __repr__(&self) -> PyResult<String> { - self.__str__() - } -} - -pub(crate) fn zvalue_of_pyany(obj: &PyAny) -> PyResult<ZValue> { - match obj.get_type().name()? { - "Value" => { - let v: Value = obj.extract()?; - Ok(v.v) - } - "bytes" => { - let buf: &[u8] = obj.extract()?; - Ok(Vec::from(buf).into()) - } - "str" => { - let s: String = obj.extract()?; - Ok(s.into()) - } - "dict" => { - let props: HashMap<String, String> = obj.extract()?; - Ok(zenoh::prelude::Properties::from(props).into()) - } - "int" => { - let i: i64 = obj.extract()?; - Ok(i.into()) - } - "float" => { - let f: f64 = obj.extract()?; - Ok(f.into()) - } - "tuple" => { - let tuple: &PyTuple = obj.downcast()?; - if tuple.len() == 2 - && (tuple.get_item(0)?.get_type().name()? == "bytes" - || tuple.get_item(0)?.get_type().name()? == "str") - && (tuple.get_item(1)?.get_type().name()? == "str" - || tuple.get_item(1)?.get_type().name()? == "Encoding") - { - let buf: &[u8] = if tuple.get_item(0)?.get_type().name()? == "bytes" { - tuple.get_item(0)?.extract()? - } else { - tuple.get_item(0)?.extract::<&str>()?.as_bytes() - }; - let encoding_descr: ZEncoding = if tuple.get_item(1)?.get_type().name()? == "str" { - tuple.get_item(1)?.extract::<String>()?.into() - } else { - tuple.get_item(1)?.extract::<Encoding>()?.e - }; - Ok(ZValue::new(Vec::from(buf).into()).encoding(encoding_descr)) - } else { - Err(PyErr::new::<exceptions::PyValueError, _>(format!( - "Cannot convert type '{:?}' to a zenoh Value", - tuple - ))) - } - } - x => Err(PyErr::new::<exceptions::PyValueError, _>(format!( - "Cannot convert type '{}' to a zenoh Value", - x - ))), - } -} - -/// A Timestamp composed of a time and the identifier of the timestamp source. -#[pyclass] -#[derive(Debug, Clone, Copy)] -pub(crate) struct Timestamp { - pub(crate) t: zenoh::time::Timestamp, -} - -#[pymethods] -impl Timestamp { - /// The time in seconds since the UNIX EPOCH (January 1, 1970, 00:00:00 (UTC)) - /// as a floating point number. - /// - /// :type: **float** - #[getter] - fn time(&self) -> f64 { - self.t.get_time().to_duration().as_secs_f64() - } - - /// The identifier of the timestamp source - /// - /// :type: **bytes** - #[getter] - fn id(&self) -> &[u8] { - self.t.get_id().as_slice() - } -} - -#[pyproto] -impl PyObjectProtocol for Timestamp { - fn __str__(&self) -> String { - self.t.to_string() - } -} - -/// Some informations about the associated data -#[pyclass] -#[derive(Clone)] -pub(crate) struct SourceInfo { - pub(crate) i: zenoh::prelude::SourceInfo, -} - -#[pymethods] -impl SourceInfo { - /// The :class:`PeerId` of the data source. - /// - /// :type: :class:`PeerId` or `None` - #[getter] - fn source_id(&self) -> Option<PeerId> { - self.i.source_id.as_ref().map(|p| PeerId { p: *p }) - } - - /// The source sequence number of the data. - /// - /// :type: int or `None` - #[getter] - fn source_sn(&self) -> Option<ZInt> { - self.i.source_sn - } - - /// The :class:`PeerId` of the 1st router that routed the data. - /// - /// :type: :class:`PeerId` or `None` - #[getter] - fn first_router_id(&self) -> Option<PeerId> { - self.i.first_router_id.as_ref().map(|p| PeerId { p: *p }) - } - - /// The first router sequence number of the data. - /// - /// :type: int or `None` - #[getter] - fn first_router_sn(&self) -> Option<ZInt> { - self.i.first_router_sn - } -} - -/// A zenoh sample. -/// -/// :param key_expr: the resource name -/// :type key_expr: str -/// :param payload: the data payload -/// :type payload: bytes -/// :param source_info: some information about the data -/// :type source_info: SourceInfo, optional -#[pyclass] -#[pyo3(text_signature = "(key_expr, payload, source_info=None)")] -#[derive(Clone)] -pub(crate) struct Sample { - pub(crate) s: zenoh::prelude::Sample, -} - -impl pyo3::conversion::ToPyObject for Sample { - fn to_object(&self, py: Python) -> pyo3::PyObject { - pyo3::IntoPy::into_py(pyo3::Py::new(py, self.clone()).unwrap(), py) - } -} - -#[pymethods] -impl Sample { - #[new] - fn new(key_expr: &PyAny, payload: &PyAny) -> Self { - let key_expr = zkey_expr_of_pyany(key_expr).unwrap(); - let payload = zvalue_of_pyany(payload).unwrap(); - Sample { - s: zenoh::prelude::Sample::new(key_expr.to_owned(), payload), - } - } - pub fn with_timestamp(&mut self, timestamp: Timestamp) { - unsafe { - let s = std::ptr::read(self); - let s = s.s.with_timestamp(timestamp.t); - std::ptr::write(self, Sample { s }); - } - } - pub fn with_source_info(&mut self, info: SourceInfo) { - unsafe { - let s = std::ptr::read(self); - let s = s.s.with_source_info(info.i); - std::ptr::write(self, Sample { s }); - } - } - - /// The resource name - /// - /// :type: str - #[getter] - fn key_expr(&self) -> KeyExpr { - self.s.key_expr.to_owned().into() - } - - /// The data payload - /// - /// DEPRECATED: use the strictly equivalent code: `sample.value.payload` - /// - /// :type: bytes - #[getter] - fn payload<'a>(&self, py: Python<'a>) -> &'a PyBytes { - PyBytes::new(py, self.s.value.payload.contiguous().as_ref()) - } - - /// The data payload - /// - /// :type: bytes - #[getter] - fn value(&self) -> Value { - Value { - v: self.s.value.clone(), - } - } - - /// The data payload - /// - /// :type: bytes - #[getter] - fn kind(&self) -> SampleKind { - self.s.kind.into() - } - - /// Some information about the data - /// - /// :type: :class:`SourceInfo` or `None` - #[getter] - fn source_info(&self) -> Option<SourceInfo> { - Some(SourceInfo { - i: self.s.source_info.clone(), - }) - } - - /// The timestamp - /// - /// :type: :class:`Timestamp` or `None` - #[getter] - fn timestamp(&self) -> Option<Timestamp> { - self.s.timestamp.map(|t| Timestamp { t }) - } -} - -#[pyproto] -impl PyObjectProtocol for Sample { - fn __str__(&self) -> String { - format!("{:?}", self.s) - } - - fn __repr__(&self) -> String { - self.__str__() - } -} - -// zenoh.Reliability (simulate the enum as a class with static methods for the cases, -// waiting for https://github.com/PyO3/pyo3/issues/834 to be fixed) -// -/// The kind of reliability -#[pyclass] -#[derive(Clone, Copy, PartialEq, Default)] -pub(crate) struct Reliability { - pub(crate) r: zenoh::subscriber::Reliability, -} - -#[allow(non_snake_case)] -#[pymethods] -impl Reliability { - #[classattr] - fn BestEffort() -> Reliability { - Reliability { - r: zenoh::subscriber::Reliability::BestEffort, - } - } - - #[classattr] - fn Reliable() -> Reliability { - Reliability { - r: zenoh::subscriber::Reliability::Reliable, - } - } -} - -// zenoh.SubMode (simulate the enum as a class with static methods for the cases, -// waiting for https://github.com/PyO3/pyo3/issues/834 to be fixed) -// -/// The subscription mode. -#[pyclass] -#[derive(Clone, Default)] -pub(crate) struct SubMode { - pub(crate) m: zenoh::subscriber::SubMode, -} - -#[allow(non_snake_case)] -#[pymethods] -impl SubMode { - #[classattr] - fn Push() -> SubMode { - SubMode { - m: zenoh::subscriber::SubMode::Push, - } - } - - #[classattr] - fn Pull() -> SubMode { - SubMode { - m: zenoh::subscriber::SubMode::Pull, - } - } -} - -/// A time period. -#[pyclass] -#[pyo3(text_signature = "(origin, period, duration)")] -#[derive(Clone)] -pub(crate) struct Period { - pub(crate) p: zenoh::time::Period, -} - -#[pymethods] -impl Period { - #[new] - fn new(origin: ZInt, period: ZInt, duration: ZInt) -> Period { - Period { - p: zenoh::time::Period { - origin, - period, - duration, - }, - } - } -} - -pub(crate) enum ZnSubOps { - Pull, - Unregister, -} - -/// A subscriber -#[pyclass] -pub(crate) struct Subscriber { - pub(crate) unregister_tx: Sender<ZnSubOps>, - pub(crate) loop_handle: Option<async_std::task::JoinHandle<()>>, -} - -#[pymethods] -impl Subscriber { - /// Pull available data for a pull-mode subscriber. - fn pull(&self) { - task::block_on(async { - if let Err(e) = self.unregister_tx.send(ZnSubOps::Pull).await { - warn!("Error in Subscriber::pull() : {}", e); - } - }); - } - - /// Close the subscriber. - fn close(&mut self) { - if let Some(handle) = self.loop_handle.take() { - task::block_on(async { - if let Err(e) = self.unregister_tx.send(ZnSubOps::Unregister).await { - warn!("Error in Subscriber::close() : {}", e); - } - handle.await; - }); - } - } -} - -// zenoh.queryable (simulate the package as a class, and consts as class attributes) -// -/// Constants defining the different modes of a zenoh :class:`zenoh.Queryable`. -#[allow(non_camel_case_types)] -#[pyclass] -pub(crate) struct queryable {} - -#[allow(non_snake_case)] -#[pymethods] -impl queryable { - #[classattr] - fn ALL_KINDS() -> ZInt { - zenoh::queryable::ALL_KINDS - } - - #[classattr] - fn STORAGE() -> ZInt { - zenoh::queryable::STORAGE - } - - #[classattr] - fn EVAL() -> ZInt { - zenoh::queryable::EVAL - } -} - -/// Type received by a queryable callback. See :meth:`Session.register_queryable`. -#[pyclass] -#[derive(Clone)] -pub(crate) struct Query { - pub(crate) q: async_std::sync::Arc<zenoh::queryable::Query>, -} - -impl pyo3::conversion::ToPyObject for Query { - fn to_object(&self, py: Python) -> pyo3::PyObject { - pyo3::IntoPy::into_py(pyo3::Py::new(py, self.clone()).unwrap(), py) - } -} - -#[pymethods] -impl Query { - /// The key_selector of the query - /// - /// :type: :class:`Selector` - #[getter] - fn selector(&self) -> Selector { - self.q.selector().to_owned().into() - } - - /// The key_selector of the query - /// - /// :type: :class:`KeyExpr` - #[getter] - fn key_selector(&self) -> KeyExpr { - self.q.key_selector().to_owned().into() - } - - /// The value_selector of the query - /// - /// :type: str - #[getter] - fn value_selector(&self) -> String { - self.q.value_selector().to_string() - } - - /// Send a reply to the query - /// - /// :param sample: the reply sample - /// :type: Sample - #[pyo3(text_signature = "(self, sample)")] - fn reply(&self, sample: Sample) { - self.q.reply(sample.s); - } -} - -/// An entity able to reply to queries. -#[pyclass] -pub(crate) struct Queryable { - pub(crate) unregister_tx: Sender<bool>, - pub(crate) loop_handle: Option<async_std::task::JoinHandle<()>>, -} - -#[pymethods] -impl Queryable { - /// Close the queryable. - fn close(&mut self) { - if let Some(handle) = self.loop_handle.take() { - task::block_on(async { - if let Err(e) = self.unregister_tx.send(true).await { - warn!("Error in Queryable::close() : {}", e); - } - handle.await; - }); - } - } -} - -// zenoh.Target (simulate the enum as a class with static methods for the cases, -// waiting for https://github.com/PyO3/pyo3/issues/834 to be fixed) -// -/// The queryables that should be target of a :class:`Query` -#[pyclass] -#[derive(Clone)] -pub(crate) struct Target { - pub(crate) t: zenoh::query::Target, -} - -#[allow(non_snake_case)] -#[pymethods] -impl Target { - #[staticmethod] - fn BestMatching() -> Target { - Target { - t: zenoh::query::Target::BestMatching, - } - } - - #[cfg(features = "complete_n")] - #[staticmethod] - #[pyo3(text_signature = "(n)")] - fn Complete(n: ZInt) -> Target { - Target { - t: zenoh::query::Target::Complete { n }, - } - } - - #[staticmethod] - fn All() -> Target { - Target { - t: zenoh::query::Target::All, - } - } - - #[staticmethod] - fn AllComplete() -> Target { - Target { - t: zenoh::query::Target::AllComplete, - } - } - - #[staticmethod] - fn No() -> Target { - Target { - t: zenoh::query::Target::None, - } - } -} - -/// The queryables that should be target of a :class:`Query`. -/// -/// :param kind: the kind of queryable (one constant from :class:`queryable`) -/// :type kind: int, optional -/// :param target: a characteristic of the queryable. -/// :type target: Target, optional -#[pyclass] -#[pyo3(text_signature = "(kind=None, target=None)")] -#[derive(Clone, Default)] -pub(crate) struct QueryTarget { - pub(crate) t: zenoh::query::QueryTarget, -} - -#[pymethods] -impl QueryTarget { - #[new] - fn new(kind: Option<ZInt>, target: Option<Target>) -> QueryTarget { - let mut t = zenoh::query::QueryTarget::default(); - if let Some(k) = kind { - t.kind = k; - } - if let Some(target) = target { - t.target = target.t; - } - QueryTarget { t } - } -} - -// zenoh.QueryConsolidation (simulate the enum as a class with static methods for the cases, -// waiting for https://github.com/PyO3/pyo3/issues/834 to be fixed) -// -/// The kind of consolidation that should be applied on replies to a :meth:`Session.get`. -#[pyclass] -#[derive(Clone)] -pub(crate) struct ConsolidationMode { - pub(crate) c: zenoh::query::ConsolidationMode, -} - -#[allow(non_snake_case)] -#[pymethods] -impl ConsolidationMode { - #[classattr] - fn No() -> ConsolidationMode { - ConsolidationMode { - c: zenoh::query::ConsolidationMode::None, - } - } - - #[classattr] - fn Lazy() -> ConsolidationMode { - ConsolidationMode { - c: zenoh::query::ConsolidationMode::Lazy, - } - } - - #[classattr] - fn Full() -> ConsolidationMode { - ConsolidationMode { - c: zenoh::query::ConsolidationMode::Full, - } - } -} - -// zenoh.ConsolidationStrategy -// -/// The kind of consolidation that should be applied on replies to a :meth:`Session.get` -/// at the different stages of the reply process. -/// -/// :param first_routers: the consolidation mode to apply on first routers of the replies routing path (default: :attr:`ConsolidationMode.Lazy`) -/// :type first_routers: ConsolidationMode, optional -/// :param last_router: the consolidation mode to apply on last router of the replies routing path (default: :attr:`ConsolidationMode.Lazy`) -/// :type last_router: ConsolidationMode, optional -/// :param reception: the consolidation mode to apply at reception of the replies (default: :attr:`ConsolidationMode.Full`) -/// :type reception: ConsolidationMode, optional -#[pyclass] -#[pyo3(text_signature = "(first_routers=None, last_router=None, reception=None)")] -#[derive(Clone, Default)] -pub(crate) struct ConsolidationStrategy { - pub(crate) c: zenoh::query::ConsolidationStrategy, -} - -#[pymethods] -impl ConsolidationStrategy { - #[new] - fn new( - first_routers: Option<ConsolidationMode>, - last_router: Option<ConsolidationMode>, - reception: Option<ConsolidationMode>, - ) -> ConsolidationStrategy { - let mut c = zenoh::query::ConsolidationStrategy::default(); - if let Some(f) = first_routers { - c.first_routers = f.c; - } - if let Some(l) = last_router { - c.last_router = l.c; - } - if let Some(r) = reception { - c.reception = r.c; - } - ConsolidationStrategy { c } - } - - /// No consolidation performed. - /// - /// This is usefull when querying timeseries data bases or - /// when using quorums. - #[staticmethod] - fn none() -> Self { - Self { - c: zenoh::query::ConsolidationStrategy::none(), - } - } - - /// Lazy consolidation performed at all stages. - /// - /// This strategy offers the best latency. Replies are directly - /// transmitted to the application when received without needing - /// to wait for all replies. - /// - /// This mode does not garantie that there will be no duplicates. - #[staticmethod] - pub fn lazy() -> Self { - Self { - c: zenoh::query::ConsolidationStrategy::lazy(), - } - } - - /// Full consolidation performed at reception. - /// - /// This is the default strategy. It offers the best latency while - /// garantying that there will be no duplicates. - #[staticmethod] - pub fn reception() -> Self { - Self { - c: zenoh::query::ConsolidationStrategy::reception(), - } - } - - /// Full consolidation performed on last router and at reception. - /// - /// This mode offers a good latency while optimizing bandwidth on - /// the last transport link between the router and the application. - #[staticmethod] - pub fn last_router() -> Self { - Self { - c: zenoh::query::ConsolidationStrategy::last_router(), - } - } - - /// Full consolidation performed everywhere. - /// - /// This mode optimizes bandwidth on all links in the system - /// but will provide a very poor latency. - #[staticmethod] - pub fn full() -> Self { - Self { - c: zenoh::query::ConsolidationStrategy::full(), - } - } -} - -// zenoh.QueryConsolidation (simulate the enum as a class with static methods for the cases, -// waiting for https://github.com/PyO3/pyo3/issues/834 to be fixed) -// -/// The replies consolidation strategy to apply on replies to a :meth:`Session.get`. -#[pyclass] -#[pyo3(text_signature = "(first_routers=None, last_router=None, reception=None)")] -#[derive(Clone, Default)] -pub(crate) struct QueryConsolidation { - pub(crate) c: zenoh::query::QueryConsolidation, -} - -#[allow(non_snake_case)] -#[pymethods] -impl QueryConsolidation { - /// Automatic query consolidation strategy selection. - /// - /// A query consolidation strategy will automatically be selected depending - /// the query selector. If the selector contains time range properties, - /// no consolidation is performed. Otherwise the reception strategy is used. - #[staticmethod] - fn Auto() -> Self { - QueryConsolidation { - c: zenoh::query::QueryConsolidation::Auto, - } - } - - /// User defined query consolidation strategy. - #[staticmethod] - fn Manual(strategy: ConsolidationStrategy) -> Self { - QueryConsolidation { - c: zenoh::query::QueryConsolidation::Manual(strategy.c), - } - } - - /// No consolidation performed. - /// - /// This is usefull when querying timeseries data bases or - /// when using quorums. - #[staticmethod] - fn none() -> Self { - Self { - c: zenoh::query::QueryConsolidation::none(), - } - } - - /// Lazy consolidation performed at all stages. - /// - /// This strategy offers the best latency. Replies are directly - /// transmitted to the application when received without needing - /// to wait for all replies. - /// - /// This mode does not garantie that there will be no duplicates. - #[staticmethod] - pub fn lazy() -> Self { - Self { - c: zenoh::query::QueryConsolidation::lazy(), - } - } - - /// Full consolidation performed at reception. - /// - /// This is the default strategy. It offers the best latency while - /// garantying that there will be no duplicates. - #[staticmethod] - pub fn reception() -> Self { - Self { - c: zenoh::query::QueryConsolidation::reception(), - } - } - - /// Full consolidation performed on last router and at reception. - /// - /// This mode offers a good latency while optimizing bandwidth on - /// the last transport link between the router and the application. - #[staticmethod] - pub fn last_router() -> Self { - Self { - c: zenoh::query::QueryConsolidation::last_router(), - } - } - - /// Full consolidation performed everywhere. - /// - /// This mode optimizes bandwidth on all links in the system - /// but will provide a very poor latency. - #[staticmethod] - pub fn full() -> Self { - Self { - c: zenoh::query::QueryConsolidation::full(), - } - } -} - -/// Type received by a get callback. See :meth:`Session.get`. -#[pyclass] -#[derive(Clone)] -pub(crate) struct Reply { - pub(crate) r: zenoh::query::Reply, -} - -impl pyo3::conversion::ToPyObject for Reply { - fn to_object(&self, py: Python) -> pyo3::PyObject { - pyo3::IntoPy::into_py(pyo3::Py::new(py, self.clone()).unwrap(), py) - } -} - -#[pymethods] -impl Reply { - /// The sample - /// - /// :type: Sample - #[getter] - fn sample(&self) -> Sample { - Sample { - s: self.r.sample.clone(), - } - } - - /// The kind of reply source - /// - /// :type: int - #[getter] - fn replier_kind(&self) -> ZInt { - self.r.replier_kind - } - - /// The identifier of reply source - /// - /// :type: PeerId - fn replier_id(&self) -> PeerId { - PeerId { - p: self.r.replier_id, - } - } -} - -// zenoh.CongestionControl (simulate the enum as a class with static methods for the cases, -// waiting for https://github.com/PyO3/pyo3/issues/834 to be fixed) -// -/// The kind of congestion control. -#[pyclass] -#[derive(Clone, Default)] -pub struct CongestionControl { - pub(crate) cc: zenoh::publication::CongestionControl, -} - -#[allow(non_snake_case)] -#[pymethods] -impl CongestionControl { - #[classattr] - fn Drop() -> CongestionControl { - CongestionControl { - cc: zenoh::publication::CongestionControl::Drop, - } - } - #[classattr] - fn Block() -> CongestionControl { - CongestionControl { - cc: zenoh::publication::CongestionControl::Block, - } - } -} - -#[pyclass] -#[derive(Clone, Default)] -pub struct Priority { - pub(crate) p: zenoh::prelude::Priority, -} - -#[allow(non_snake_case)] -#[pymethods] -impl Priority { - #[classattr] - fn Background() -> Self { - Priority { - p: zenoh::prelude::Priority::Background, - } - } - #[classattr] - fn Data() -> Self { - Priority { - p: zenoh::prelude::Priority::Data, - } - } - #[classattr] - fn DataHigh() -> Self { - Priority { - p: zenoh::prelude::Priority::DataHigh, - } - } - #[classattr] - fn DataLow() -> Self { - Priority { - p: zenoh::prelude::Priority::DataLow, - } - } - #[classattr] - fn InteractiveHigh() -> Self { - Priority { - p: zenoh::prelude::Priority::InteractiveHigh, - } - } - #[classattr] - fn InteractiveLow() -> Self { - Priority { - p: zenoh::prelude::Priority::InteractiveLow, - } - } - #[classattr] - fn RealTime() -> Self { - Priority { - p: zenoh::prelude::Priority::RealTime, - } - } -} diff --git a/src/value.rs b/src/value.rs new file mode 100644 index 00000000..2fa29c44 --- /dev/null +++ b/src/value.rs @@ -0,0 +1,332 @@ +// Copyright (c) 2017, 2022 ZettaScale Technology Inc. + +// This program and the accompanying materials are made available under the +// terms of the Eclipse Public License 2.0 which is available at +// http://www.eclipse.org/legal/epl-2.0, or the Apache License, Version 2.0 +// which is available at https://www.apache.org/licenses/LICENSE-2.0. + +// SPDX-License-Identifier: EPL-2.0 OR Apache-2.0 + +// Contributors: +// ZettaScale Zenoh team, <[email protected]> + +use pyo3::{prelude::*, types::PyBytes}; +use uhlc::Timestamp; +use zenoh::{ + prelude::{Encoding, KeyExpr, Sample, Value, ZenohId}, + query::Reply, + scouting::Hello, +}; +use zenoh_buffers::{SplitBuffer, ZBuf}; + +use crate::{ + enums::{_Encoding, _SampleKind}, + keyexpr::_KeyExpr, + ToPyErr, +}; + +#[derive(Clone)] +pub(crate) enum Payload { + Zenoh(ZBuf), + Python(Py<PyBytes>), +} +impl Payload { + pub(crate) fn into_zbuf(self) -> ZBuf { + match self { + Payload::Zenoh(buf) => buf, + Payload::Python(buf) => Python::with_gil(|py| ZBuf::from(buf.as_bytes(py).to_owned())), + } + } + pub(crate) fn into_pybytes(self) -> Py<PyBytes> { + match self { + Payload::Zenoh(buf) => { + Python::with_gil(|py| Py::from(PyBytes::new(py, buf.contiguous().as_ref()))) + } + Payload::Python(buf) => buf, + } + } +} +impl From<ZBuf> for Payload { + fn from(buf: ZBuf) -> Self { + Payload::Zenoh(buf) + } +} +impl From<Py<PyBytes>> for Payload { + fn from(buf: Py<PyBytes>) -> Self { + Payload::Python(buf) + } +} +#[pyclass(subclass)] +#[derive(Clone)] +pub struct _Value { + pub(crate) payload: Payload, + pub(crate) encoding: Encoding, +} +#[pymethods] +impl _Value { + #[new] + pub fn pynew(this: Self) -> Self { + this + } + #[staticmethod] + pub fn new(payload: Py<PyBytes>, encoding: Option<_Encoding>) -> Self { + Self { + payload: payload.into(), + encoding: encoding.map(|e| e.0).unwrap_or(Encoding::EMPTY), + } + } + #[getter] + pub fn payload(&mut self) -> Py<PyBytes> { + if let Payload::Python(buf) = &self.payload { + return buf.clone(); + } + let payload = unsafe { std::ptr::read(&self.payload) }; + let buf = payload.into_pybytes(); + unsafe { std::ptr::write(&mut self.payload, Payload::Python(buf.clone())) }; + buf + } + pub fn with_payload(&mut self, payload: Py<PyBytes>) { + self.payload = Payload::Python(payload) + } + #[getter] + pub fn encoding(&self) -> _Encoding { + _Encoding(self.encoding.clone()) + } + pub fn with_encoding(&mut self, encoding: _Encoding) { + self.encoding = encoding.0; + } +} +impl From<Value> for _Value { + fn from(value: Value) -> Self { + _Value { + payload: value.payload.into(), + encoding: value.encoding, + } + } +} +impl From<_Value> for Value { + fn from(value: _Value) -> Self { + Value::new(value.payload.into_zbuf()).encoding(value.encoding) + } +} + +pub(crate) trait PyAnyToValue { + fn to_value(self) -> PyResult<Value>; +} +impl PyAnyToValue for &PyAny { + fn to_value(self) -> PyResult<Value> { + let encoding: _Encoding = self.getattr("encoding")?.extract()?; + let payload: &PyBytes = self.getattr("payload")?.extract()?; + Ok(Value::new(ZBuf::from(payload.as_bytes().to_owned())).encoding(encoding.0)) + } +} + +#[pyclass(subclass)] +#[derive(Clone)] +pub struct _Sample { + key_expr: KeyExpr<'static>, + value: _Value, + kind: _SampleKind, + timestamp: Option<_Timestamp>, +} +impl From<Sample> for _Sample { + fn from(sample: Sample) -> Self { + let Sample { + key_expr, + value, + kind, + timestamp, + .. + } = sample; + _Sample { + key_expr, + value: value.into(), + kind: _SampleKind(kind), + timestamp: timestamp.map(_Timestamp), + } + } +} + +#[pyclass(subclass)] +#[derive(Clone, Copy, PartialEq, Eq)] +pub struct _ZenohId(pub(crate) ZenohId); +#[pymethods] +impl _ZenohId { + #[new] + pub fn pynew(this: Self) -> Self { + this + } + pub fn __str__(&self) -> String { + self.0.to_string() + } +} + +#[pyclass(subclass)] +#[derive(Clone, Copy, PartialEq, Eq, PartialOrd, Ord)] +pub struct _Timestamp(Timestamp); +#[pymethods] +impl _Timestamp { + #[new] + pub fn pynew(this: Self) -> Self { + this + } + crate::derive_richcmp!(); + #[getter] + pub fn seconds_since_unix_epoch(&self) -> PyResult<f64> { + match self + .0 + .get_time() + .to_system_time() + .duration_since(std::time::UNIX_EPOCH) + { + Ok(o) => Ok(o.as_secs_f64()), + Err(e) => Err(e.to_pyerr()), + } + } +} + +#[pymethods] +impl _Sample { + #[new] + pub fn pynew(this: Self) -> Self { + this + } + #[getter] + pub fn value(&self) -> _Value { + self.value.clone() + } + #[getter] + pub fn key_expr(&self) -> _KeyExpr { + _KeyExpr(self.key_expr.clone()) + } + #[getter] + pub fn payload(&mut self) -> Py<PyBytes> { + if let Payload::Python(buf) = &self.value.payload { + return buf.clone(); + } + let payload = unsafe { std::ptr::read(&self.value.payload) }; + let buf = payload.into_pybytes(); + unsafe { std::ptr::write(&mut self.value.payload, Payload::Python(buf.clone())) }; + buf + } + #[getter] + pub fn encoding(&self) -> _Encoding { + _Encoding(self.value.encoding.clone()) + } + #[getter] + pub fn kind(&self) -> _SampleKind { + self.kind.clone() + } + #[getter] + pub fn timestamp(&self) -> Option<_Timestamp> { + self.timestamp + } + #[staticmethod] + pub fn new( + key_expr: _KeyExpr, + value: _Value, + kind: _SampleKind, + timestamp: Option<_Timestamp>, + ) -> Self { + _Sample { + key_expr: key_expr.0, + value, + kind, + timestamp, + } + } +} + +impl From<_Sample> for Sample { + fn from(sample: _Sample) -> Self { + let _Sample { + key_expr, + value, + kind, + timestamp, + } = sample; + let mut sample = Sample::new(key_expr, value); + sample.kind = kind.0; + sample.timestamp = timestamp.map(|t| t.0); + sample + } +} + +#[pyclass(subclass)] +#[derive(Clone)] +pub struct _Reply { + #[pyo3(get)] + pub replier_id: _ZenohId, + pub reply: Result<_Sample, _Value>, +} +#[pymethods] +impl _Reply { + #[new] + pub fn pynew(this: Self) -> Self { + this + } + #[getter] + pub fn ok(&self) -> PyResult<_Sample> { + match &self.reply { + Ok(o) => Ok(o.clone()), + Err(_) => Err(zenoh_core::zerror!("Called `Reply.ok` on a non-ok reply.").to_pyerr()), + } + } + #[getter] + pub fn err(&self) -> PyResult<_Value> { + match &self.reply { + Err(o) => Ok(o.clone()), + Ok(_) => Err(zenoh_core::zerror!("Called `Reply.err` on a non-err reply.").to_pyerr()), + } + } +} +impl From<Reply> for _Reply { + fn from(reply: Reply) -> Self { + _Reply { + replier_id: _ZenohId(reply.replier_id), + reply: match reply.sample { + Ok(o) => Ok(o.into()), + Err(e) => Err(e.into()), + }, + } + } +} + +#[pyclass(subclass)] +#[derive(Clone)] +pub struct _Hello(pub(crate) Hello); +#[pymethods] +impl _Hello { + #[new] + pub fn pynew(this: Self) -> Self { + this + } + #[getter] + pub fn zid(&self) -> Option<_ZenohId> { + self.0.zid.map(_ZenohId) + } + #[getter] + pub fn whatami(&self) -> Option<&'static str> { + match self.0.whatami { + Some(zenoh::config::whatami::WhatAmI::Client) => Some("client"), + Some(zenoh::config::whatami::WhatAmI::Peer) => Some("peer"), + Some(zenoh::config::whatami::WhatAmI::Router) => Some("router"), + None => None, + } + } + #[getter] + pub fn locators(&self) -> Vec<String> { + match &self.0.locators { + Some(locators) => locators.iter().map(|l| l.to_string()).collect(), + None => Vec::new(), + } + } + pub fn __str__(&self) -> String { + self.0.to_string() + } +} +impl From<Hello> for _Hello { + fn from(h: Hello) -> Self { + _Hello(h) + } +} diff --git a/zenoh/__init__.py b/zenoh/__init__.py new file mode 100644 index 00000000..266d4c60 --- /dev/null +++ b/zenoh/__init__.py @@ -0,0 +1,44 @@ +# +# Copyright (c) 2022 ZettaScale Technology +# +# This program and the accompanying materials are made available under the +# terms of the Eclipse Public License 2.0 which is available at +# http://www.eclipse.org/legal/epl-2.0, or the Apache License, Version 2.0 +# which is available at https://www.apache.org/licenses/LICENSE-2.0. +# +# SPDX-License-Identifier: EPL-2.0 OR Apache-2.0 +# +# Contributors: +# ZettaScale Zenoh Team, <[email protected]> +# +from .zenoh import init_logger, scout as _scout +from .keyexpr import IntoKeyExpr, IntoSelector, KeyExpr, Selector +from .config import Config +from .session import Session, Publisher, Subscriber, PullSubscriber, Info +from .enums import CongestionControl, Encoding, Priority, QueryConsolidation, QueryTarget, Reliability, SampleKind +from .value import Hello, Value, IntoValue, IValue, Sample, IntoSample, ZenohId, Timestamp, Reply +from .closures import Closure, IClosure, IntoClosure, Handler, IHandler, IntoHandler, ListCollector, Queue +from .queryable import Queryable, Query +from typing import Any + +def open(*args, **kwargs): + return Session(*args, **kwargs) + +class Scout: + def __init__(self, inner, receiver): + self._inner_ = inner + self.receiver = receiver + + def stop(self): + self._inner_ = None + +def scout(handler: IntoHandler[Hello, Any, Any] = None, what: str = None, config: Config = None, timeout=None): + from threading import Timer + if handler is None: + handler = ListCollector() + handler = Handler(handler, lambda x: Hello._upgrade_) + scout = _scout(handler.closure, config, what) + scout = Scout(scout, handler.receiver) + if timeout: + Timer(timeout, lambda: scout.stop()).start() + return scout \ No newline at end of file diff --git a/zenoh/closures.py b/zenoh/closures.py new file mode 100644 index 00000000..db2ccbb7 --- /dev/null +++ b/zenoh/closures.py @@ -0,0 +1,268 @@ +# +# Copyright (c) 2022 ZettaScale Technology +# +# This program and the accompanying materials are made available under the +# terms of the Eclipse Public License 2.0 which is available at +# http://www.eclipse.org/legal/epl-2.0, or the Apache License, Version 2.0 +# which is available at https://www.apache.org/licenses/LICENSE-2.0. +# +# SPDX-License-Identifier: EPL-2.0 OR Apache-2.0 +# +# Contributors: +# ZettaScale Zenoh Team, <[email protected]> +# +import abc +from typing import Generic, Callable, Union, Any, TypeVar, Tuple, List +from threading import Condition, Thread +from collections import deque +import time + +In = TypeVar("In") +Out = TypeVar("Out") +Receiver = TypeVar("Receiver") +CallbackCall = Callable[[In], Out] +CallbackDrop = Callable[[], None] + +class IClosure(Generic[In, Out]): + """ + A Closure is a pair of a `call` function that will be used as a callback, + and a `drop` function that will be called when the closure is destroyed. + """ + @property + @abc.abstractmethod + def call(self) -> Callable[[In], Out]: + """ + Returns the closure's call function as a lambda. + """ + ... + @property + @abc.abstractmethod + def drop(self) -> Callable[[], None]: + """ + Returns the closure's destructor as a lambda. + """ + ... + def __enter__(self): + drop = self.drop + if drop is not None: + drop() + def __exit__(self, *args): + drop = self.drop + if drop is not None: + drop() + +class IHandler(Generic[In, Out, Receiver]): + """ + A Handler is a value that may be converted into a callback closure for zenoh to use on one side, while possibly providing a receiver for the data that zenoh would provide through that callback. + """ + @property + @abc.abstractmethod + def closure(self) -> IClosure[In, Out]: + """ + The part of the handler that should be passed as a callback to a zenoh function. + """ + ... + @property + @abc.abstractmethod + def receiver(self) -> Receiver: + "The part of the handler that should be used as the receiver when the handler is channel-like." + ... + +IntoClosure = Union[IHandler[In, Out, Any], IClosure[In, Out], Tuple[CallbackCall, CallbackDrop], CallbackCall] +class Closure(IClosure, Generic[In, Out]): + """ + A Closure is a pair of a `call` function that will be used as a callback, + and a `drop` function that will be called when the closure is destroyed. + """ + def __init__(self, closure: IntoClosure[In, Out], type_adaptor: Callable[[Any], In] = None): + _call_ = None + self._drop_ = lambda: None + if isinstance(closure, IHandler): + closure = closure.closure + # dev-note: do not elif here, the next if will catch the obtained closure. + if isinstance(closure, IClosure): + _call_ = closure.call + self._drop_ = closure.drop + elif isinstance(closure, tuple): + _call_, self._drop_ = closure + elif callable(closure): + _call_ = closure + else: + raise TypeError("Unexpected type as input for zenoh.Closure") + if type_adaptor is not None: + self._call_ = lambda *args: _call_(type_adaptor(*args)) + else: + self._call_ = _call_ + @property + def call(self) -> Callable[[In], Out]: + return self._call_ + + @property + def drop(self) -> Callable[[], None]: + return self._drop_ + +IntoHandler = Union[IHandler[In, Out, Receiver], IClosure[In, Out], Tuple[IClosure, Receiver], Tuple[CallbackCall,CallbackDrop, Receiver], Tuple[CallbackCall,CallbackDrop], CallbackCall] +class Handler(IHandler, Generic[In, Out, Receiver]): + """ + A Handler is a value that may be converted into a callback closure for zenoh to use on one side, while possibly providing a receiver for the data that zenoh would provide through that callback. + """ + def __init__(self, input: IntoHandler[In, Out, Receiver], type_adaptor: Callable[[Any], In] = None): + self._receiver_ = None + if isinstance(input, IHandler): + self._receiver_ = input.receiver + self._closure_ = input.closure + elif isinstance(input, IClosure): + self._closure_ = input + elif isinstance(input, tuple): + if isinstance(input[0], IClosure): + self._closure_, self._receiver_ = input + elif len(input) == 3: + call, drop, self._receiver_ = input + self._closure_ = (call, drop) + else: + self._closure_ = input + else: + self._closure_ = input + self._closure_ = Closure(self._closure_, type_adaptor) + + @property + def closure(self) -> IClosure[In, Out]: + return self._closure_ + @property + def receiver(self) -> Receiver: + return self._receiver_ + + + +class ListCollector(IHandler[In, None, Callable[[],List[In]]], Generic[In]): + """ + A simple collector that aggregates values into a list. + + When used as a handler, it provides a callback that appends elements to a list, + and provides a function that will await the closing of the callback before returning said list. + """ + def __init__(self, timeout=None): + self._vec_ = [] + self._cv_ = Condition() + self._done_ = False + self.timeout = timeout + + @property + def closure(self): + def call(x): + self._vec_.append(x) + def drop(): + with self._cv_: + self._done_ = True + self._cv_.notify() + return Closure((call, drop)) + + @property + def receiver(self): + def wait(): + with self._cv_: + if not self._done_: + self._cv_.wait(timeout=self.timeout) + return self._vec_ + return wait + +class Queue(IHandler[In, None, 'Queue'], Generic[In]): + """ + A simple single-producer, single-consumer queue implementation. + + When used as a handler, it provides itself as the receiver, and will provide a + callback that appends elements to the queue. + """ + def __init__(self, timeout=None): + self._vec_ = deque() + self._cv_ = Condition() + self._done_ = False + + @property + def closure(self) -> IClosure[In, None]: + def call(x): self.put(x) + def drop(): self.close() + return Closure((call, drop)) + + @property + def receiver(self) -> 'Queue': + return self + + def put(self, value): + """ + Puts one element on the queue. + """ + with self._cv_: + self._vec_.append(value) + self._cv_.notify() + + + def get(self, timeout=None): + """ + Gets one element from the queue. + + Raises a `StopIteration` exception if the queue was closed before the timeout ran out. + Raises a `TimeoutError` if the timeout ran out. + """ + try: + return self._vec_.pop() + except IndexError: + pass + if self._done_: + raise StopIteration() + with self._cv_: + self._cv_.wait(timeout=timeout) + try: + return self._vec_.pop() + except IndexError: + pass + if self._done_: + raise StopIteration() + else: + raise TimeoutError() + + def close(self): + with self._cv_: + self._done_ = True + self._cv_.notify() + + def get_remaining(self, timeout=None) -> List[In]: + """ + Awaits the closing of the queue, returning the remaining queued values in a list. + The values inserted into the queue up until this happens will be available through `get`. + + Raises a `TimeoutError` if the timeout in seconds provided was exceeded before closing. + """ + end = (time.time() + timeout) if timeout is not None else None + while not self._done_: + with self._cv_: + self._cv_.wait(timeout=(timeout - time.time()) if timeout else None) + if self._done_: + return + elif time.time() >= end: + raise TimeoutError() + return list(self._vec_) + + def __iter__(self): + return self + def __next__(self): + return self.get() + +if __name__ == "__main__": + def get(collector): + import time + def target(): + with Closure(collector) as closure: + closure = Closure(collector) + closure.call('hi') + closure.call('there') + Thread(target=target).start() + + collector = ListCollector() + get(collector) + print(collector.receiver()) + assert collector.receiver() == ["hi", "there"] + + print("done") + + diff --git a/zenoh/config.py b/zenoh/config.py new file mode 100644 index 00000000..9f3007e1 --- /dev/null +++ b/zenoh/config.py @@ -0,0 +1,61 @@ +# +# Copyright (c) 2022 ZettaScale Technology +# +# This program and the accompanying materials are made available under the +# terms of the Eclipse Public License 2.0 which is available at +# http://www.eclipse.org/legal/epl-2.0, or the Apache License, Version 2.0 +# which is available at https://www.apache.org/licenses/LICENSE-2.0. +# +# SPDX-License-Identifier: EPL-2.0 OR Apache-2.0 +# +# Contributors: +# ZettaScale Zenoh Team, <[email protected]> +# +from typing import Union +from .zenoh import _Config +import json + +class Config(_Config): + def __init__(self): + super().__init__() + @staticmethod + def from_file(filename: str): + """ + Reads the configuration from a file. + The file's extension must be json, json5 or yaml. + """ + c = super(Config).from_file(filename) + return c + @staticmethod + def from_obj(obj): + """ + Reads the configuration from `obj` as if it was a JSON file. + """ + c = Config.from_json5(json.dumps(obj)) + return c + @staticmethod + def from_json5(json: str): + """ + Reads the configuration from a JSON5 string. + + JSON5 is a superset of JSON, so any JSON string is a valid input for this function. + """ + c = super(Config).from_json5(json) + return c + + def get_json(self, path: str) -> str: + """ + Returns the part of the configuration at `path`, + in a JSON-serialized form. + """ + return super().get_json(path) + + def insert_json5(self, path: str, value: str) -> str: + """ + Inserts the provided value (read from JSON) at the given path in the configuration. + """ + return super().insert_json5(path, value) + +MODE_KEY = "mode" +CONNECT_KEY = "connect/endpoints" +LISTEN_KEY = "listen/endpoints" \ No newline at end of file diff --git a/zenoh/enums.py b/zenoh/enums.py new file mode 100644 index 00000000..48753a68 --- /dev/null +++ b/zenoh/enums.py @@ -0,0 +1,231 @@ +# +# Copyright (c) 2022 ZettaScale Technology +# +# This program and the accompanying materials are made available under the +# terms of the Eclipse Public License 2.0 which is available at +# http://www.eclipse.org/legal/epl-2.0, or the Apache License, Version 2.0 +# which is available at https://www.apache.org/licenses/LICENSE-2.0. +# +# SPDX-License-Identifier: EPL-2.0 OR Apache-2.0 +# +# Contributors: +# ZettaScale Zenoh Team, <[email protected]> +# +from .zenoh import _Encoding, _SampleKind, _CongestionControl, _Priority, _Reliability, _QueryTarget, _QueryConsolidation + +class Priority(_Priority): + """ + The priority of a sending operation. + + They are ordered à la Linux priority: + `Priority.REAL_TIME() < Priority.INTERACTIVE_HIGH() < Priority.INTERACTIVE_LOW() < Priority.DATA() < Priority.BACKGROUND()` + """ + def __new__(cls, inner: _SampleKind): + return super().__new__(cls, inner) + @staticmethod + def REAL_TIME() -> 'Priority': + return Priority(_Priority.REAL_TIME) + @staticmethod + def INTERACTIVE_HIGH() -> 'Priority': + return Priority(_Priority.INTERACTIVE_HIGH) + @staticmethod + def INTERACTIVE_LOW() -> 'Priority': + return Priority(_Priority.INTERACTIVE_LOW) + @staticmethod + def DATA_HIGH() -> 'Priority': + return Priority(_Priority.DATA_HIGH) + @staticmethod + def DATA() -> 'Priority': + return Priority(_Priority.DATA) + @staticmethod + def DATA_LOW() -> 'Priority': + return Priority(_Priority.DATA_LOW) + @staticmethod + def BACKGROUND() -> 'Priority': + return Priority(_Priority.BACKGROUND) + def __eq__(self, other) -> bool: + return super().__eq__(other) + def __ne__(self, other) -> bool: + return not self.__eq__(other) + def __eq__(self, other) -> bool: + return super().__eq__(other) + def __lt__(self, other) -> bool: + return super().__lt__(other) + def __le__(self, other) -> bool: + return super().__le__(other) + def __gt__(self, other) -> bool: + return super().__gt__(other) + def __ge__(self, other) -> bool: + return super().__ge__(other) + +class SampleKind(_SampleKind): + "Similar to an HTTP METHOD: only PUT and DELETE are currently supported." + def __new__(cls, inner: _SampleKind): + return super().__new__(cls, inner) + @staticmethod + def PUT() -> 'SampleKind': + return SampleKind(_SampleKind.PUT) + @staticmethod + def DELETE() -> 'SampleKind': + return SampleKind(_SampleKind.DELETE) + def __eq__(self, other) -> bool: + return super().__eq__(other) + def __ne__(self, other) -> bool: + return not self.__eq__(other) + +class CongestionControl(_CongestionControl): + """ + Defines the network's behaviour regarding a message when heavily congested. + """ + def __new__(cls, inner: _CongestionControl): + return super().__new__(cls, inner) + @staticmethod + def DROP() -> 'CongestionControl': + "Allows the message to be dropped if all buffers are full." + return CongestionControl(_CongestionControl.DROP) + @staticmethod + def BLOCK() -> 'CongestionControl': + """ + Prevents the message from being dropped at all cost. + In the face of heavy congestion on a part of the network, this could result in your publisher node blocking. + """ + return CongestionControl(_CongestionControl.BLOCK) + def __eq__(self, other) -> bool: + return super().__eq__(other) + def __ne__(self, other) -> bool: + return not self.__eq__(other) + +class Encoding(_Encoding): + def __new__(cls, inner: _Encoding): + return super().__new__(cls, inner) + @staticmethod + def from_str(s: str) -> 'Encoding': + return super(Encoding, Encoding).from_str(s) + def append(self, s: str): + super().append(s) + @staticmethod + def EMPTY() -> 'Encoding': + return Encoding(_Encoding.EMPTY ) + @staticmethod + def APP_OCTET_STREAM() -> 'Encoding': + return Encoding(_Encoding.APP_OCTET_STREAM) + @staticmethod + def APP_CUSTOM() -> 'Encoding': + return Encoding(_Encoding.APP_CUSTOM) + @staticmethod + def TEXT_PLAIN() -> 'Encoding': + return Encoding(_Encoding.TEXT_PLAIN) + @staticmethod + def APP_PROPERTIES() -> 'Encoding': + return Encoding(_Encoding.APP_PROPERTIES) + @staticmethod + def APP_JSON() -> 'Encoding': + return Encoding(_Encoding.APP_JSON) + @staticmethod + def APP_SQL() -> 'Encoding': + return Encoding(_Encoding.APP_SQL) + @staticmethod + def APP_INTEGER() -> 'Encoding': + return Encoding(_Encoding.APP_INTEGER) + @staticmethod + def APP_FLOAT() -> 'Encoding': + return Encoding(_Encoding.APP_FLOAT) + @staticmethod + def APP_XML() -> 'Encoding': + return Encoding(_Encoding.APP_XML) + @staticmethod + def APP_XHTML_XML() -> 'Encoding': + return Encoding(_Encoding.APP_XHTML_XML) + @staticmethod + def APP_X_WWW_FORM_URLENCODED() -> 'Encoding': + return Encoding(_Encoding.APP_X_WWW_FORM_URLENCODED) + @staticmethod + def TEXT_JSON() -> 'Encoding': + return Encoding(_Encoding.TEXT_JSON) + @staticmethod + def TEXT_HTML() -> 'Encoding': + return Encoding(_Encoding.TEXT_HTML) + @staticmethod + def TEXT_XML() -> 'Encoding': + return Encoding(_Encoding.TEXT_XML) + @staticmethod + def TEXT_CSS() -> 'Encoding': + return Encoding(_Encoding.TEXT_CSS) + @staticmethod + def TEXT_CSV() -> 'Encoding': + return Encoding(_Encoding.TEXT_CSV) + @staticmethod + def TEXT_JAVASCRIPT() -> 'Encoding': + return Encoding(_Encoding.TEXT_JAVASCRIPT) + @staticmethod + def IMAGE_JPEG() -> 'Encoding': + return Encoding(_Encoding.IMAGE_JPEG) + @staticmethod + def IMAGE_PNG() -> 'Encoding': + return Encoding(_Encoding.IMAGE_PNG) + @staticmethod + def IMAGE_GIF() -> 'Encoding': + return Encoding(_Encoding.IMAGE_GIF) + def __eq__(self, other) -> bool: + return super().__eq__(other) + def __ne__(self, other) -> bool: + return not self.__eq__(other) + +class Reliability(_Reliability): + "Used by subscribers to inform the network of the reliability it wishes to obtain." + def __new__(cls, inner: _Reliability): + return super().__new__(cls, inner) + @staticmethod + def BEST_EFFORT() -> 'CongestionControl': + "Informs the network that dropping some messages is acceptable" + return Reliability(_Reliability.BEST_EFFORT) + @staticmethod + def RELIABLE() -> 'CongestionControl': + """ + Informs the network that this subscriber wishes for all publications to reliably reach it. + + Note that if a publisher puts a sample with the `CongestionControl.DROP()` option, this reliability + requirement may be infringed to prevent slow readers from blocking the network. + """ + return Reliability(_Reliability.RELIABLE) + def __eq__(self, other) -> bool: + return super().__eq__(other) + def __ne__(self, other) -> bool: + return not self.__eq__(other) + +class QueryTarget(_QueryTarget): + def __new__(cls, inner: _QueryTarget): + return super().__new__(cls, inner) + @staticmethod + def BEST_MATCHING() -> 'QueryTarget': + return QueryTarget(_QueryTarget.BEST_MATCHING) + @staticmethod + def ALL() -> 'QueryTarget': + return QueryTarget(_QueryTarget.ALL) + @staticmethod + def ALL_COMPLETE() -> 'QueryTarget': + return QueryTarget(_QueryTarget.ALL_COMPLETE) + def __eq__(self, other) -> bool: + return super().__eq__(other) + def __ne__(self, other) -> bool: + return not self.__eq__(other) + +class QueryConsolidation(_QueryConsolidation): + def __new__(cls, inner: _QueryConsolidation): + return super().__new__(cls, inner) + @staticmethod + def AUTO() -> 'QueryConsolidation': + return QueryConsolidation(_QueryConsolidation.AUTO) + @staticmethod + def NONE() -> 'QueryConsolidation': + return QueryConsolidation(_QueryConsolidation.NONE) + @staticmethod + def MONOTONIC() -> 'QueryConsolidation': + return QueryConsolidation(_QueryConsolidation.MONOTONIC) + @staticmethod + def LATEST() -> 'QueryConsolidation': + return QueryConsolidation(_QueryConsolidation.LATEST) + def __eq__(self, other) -> bool: + return super().__eq__(other) + def __ne__(self, other) -> bool: + return not self.__eq__(other) \ No newline at end of file diff --git a/zenoh/keyexpr.py b/zenoh/keyexpr.py new file mode 100644 index 00000000..b308e232 --- /dev/null +++ b/zenoh/keyexpr.py @@ -0,0 +1,182 @@ +# +# Copyright (c) 2022 ZettaScale Technology +# +# This program and the accompanying materials are made available under the +# terms of the Eclipse Public License 2.0 which is available at +# http://www.eclipse.org/legal/epl-2.0, or the Apache License, Version 2.0 +# which is available at https://www.apache.org/licenses/LICENSE-2.0. +# +# SPDX-License-Identifier: EPL-2.0 OR Apache-2.0 +# +# Contributors: +# ZettaScale Zenoh Team, <[email protected]> +# +from typing import Union, Dict +from .zenoh import _KeyExpr, _Selector + +IntoKeyExpr = Union['KeyExpr', _KeyExpr, str] + +class KeyExpr(_KeyExpr): + """ + Zenoh's address space is designed around keys which serve as the names of ressources. + + Keys are slash-separated lists of non-empty UTF8 strings. They may not contain the following characters: `$*#?`. + + Zenoh's operations are executed on key expressions, a small language that allows the definition + of sets of keys via the use of wildcards: + - `*` is the single-chunk wildcard, and will match any chunk: "a/*/c" will match "a/b/c", "a/hello/c", etc... + - `**` is the 0 or more chunks wildcard: "a/**/c" matches "a/c", "a/b/c", "a/b/hello/c"... + - `$*` is the subchunk wildcard, it will match any amount of non-/ characters: "a/b$*" matches "a/b", "a/because", "a/blue"... but not "a/c" nor "a/blue/c" + + To allow for better performance and gain the property that two key expressions define the same + set if and only if they are the same string, the rules of canon form are mandatory for a key + expression to be propagated by a Zenoh network: + - `**/**` may not exist, as it could always be replaced by the shorter `**`, + - `**/*` may not exist, and must be written as its equivalent `*/**` instead, + - `$*` may not exist alone in a chunk, as it must be written `*` instead. + + The `KeyExpr.autocanonize` constructor exists to correct eventual infrigements of the canonization rules. + + A KeyExpr is a string that has been validated to be a valid Key Expression. + """ + def __new__(cls, expr: IntoKeyExpr): + """ + The default constructor for KeyExpr will ensure that the passed expression is valid. + It won't however try to correct expressions that aren't canon. + + You may use `KeyExpr.autocanonize(expr)` instead if you are unsure if the expression + you will use for construction will be canon. + + Raises a zenoh.ZError exception if `expr` is not a valid key expression. + """ + if isinstance(expr, KeyExpr): + return expr + elif isinstance(expr, _KeyExpr): + return _KeyExpr.__new__(cls, expr) + else: + return _KeyExpr.__new__(cls, _KeyExpr.new(expr)) + + def _upgrade_(this: _KeyExpr) -> 'KeyExpr': + return _KeyExpr.__new__(KeyExpr, expr) + + @staticmethod + def autocanonize(expr: str) -> 'KeyExpr': + """ + This alternative constructor for key expressions will attempt to canonize the passed + expression before checking if it is valid. + + Raises a zenoh.ZError exception if `expr` is not a valid key expression. + """ + if isinstance(expr, KeyExpr): + return expr + else: + e = _KeyExpr.autocanonize(expr) + return KeyExpr(e.as_str()) + + def intersects(self, other: 'KeyExpr') -> bool: + """ + This method returns `True` if there exists at least one key that belongs to both sets + defined by `self` and `other`. + """ + return super().intersects(other) + + def includes(self, other: 'KeyExpr') -> bool: + """ + This method returns `True` if all of the keys defined by `other` also belong to the set + defined by `self`. + """ + return super().includes(other) + + def undeclare(self, session: 'Session'): + """ + Undeclares a key expression previously declared on the session. + """ + super().undeclare(session) + + def __eq__(self, other: 'KeyExpr') -> bool: + """ + Corresponds to set equality. + """ + return super().__eq__(other) + + def __truediv__(self, other: IntoKeyExpr) -> 'KeyExpr': + """ + Joins two key expressions with a `/`. + + Raises a zenoh.ZError exception if `other` is not a valid key expression. + """ + return KeyExpr.autocanonize(f"{self}/{other}") + + def __str__(self): + return super().__str__() + + def __hash__(self): + return super().__hash__() + +IntoSelector = Union['Selector', _Selector, IntoKeyExpr] +class Selector(_Selector): + """ + A selector is the combination of a [Key Expression](crate::prelude::KeyExpr), which defines the + set of keys that are relevant to an operation, and a `parameters`, a set of key-value pairs + with a few uses: + * specifying arguments to a queryable, allowing the passing of Remote Procedure Call parameters + * filtering by value, + * filtering by metadata, such as the timestamp of a value, + + When in string form, selectors look a lot like a URI, with similar semantics: + * the `key_expr` before the first `?` must be a valid key expression. + * the `parameters` after the first `?` should be encoded like the query section of a URL: + * key-value pairs are separated by `&`, + * the key and value are separated by the first `=`, + * in the absence of `=`, the value is considered to be the empty string, + * both key and value should use percent-encoding to escape characters, + * defining a value for the same key twice is considered undefined behavior. + + Zenoh intends to standardize the usage of a set of keys. To avoid conflicting with RPC parameters, + the Zenoh team has settled on reserving the set of keys that start with non-alphanumeric characters. + + This document will summarize the standardized keys for which Zenoh provides helpers to facilitate + coherent behavior for some operations. + + Queryable implementers are encouraged to prefer these standardized keys when implementing their + associated features, and to prefix their own keys to avoid having conflicting keys with other + queryables. + + Here are the currently standardized keys for Zenoh: + * `_time`: used to express interest in only values dated within a certain time range, values for + this key must be readable by the [Zenoh Time DSL](zenoh_util::time_range::TimeRange) for the value to be considered valid. + * `_filter`: *TBD* Zenoh intends to provide helper tools to allow the value associated with + this key to be treated as a predicate that the value should fulfill before being returned. + A DSL will be designed by the Zenoh team to express these predicates. + """ + def __new__(cls, selector: IntoSelector): + if isinstance(selector, Selector): + return selector + if isinstance(selector, _Selector): + return Selector._upgrade_(selector) + return Selector._upgrade_(super().new(str(selector))) + @staticmethod + def _upgrade_(this: _Selector) -> 'Selector': + return _Selector.__new__(Selector, this) + @property + def key_expr(self) -> KeyExpr: + "The key expression part of the selector." + return KeyExpr(super().key_expr) + @property + def parameters(self): + "The value selector part of the selector." + return super().parameters + @parameters.setter + def set_parameters(self, parameters: str): + super().parameters = parameters + def decode_parameters(self) -> Dict[str, str]: + """ + Decodes the value selector part of the selector. + + Raises a ZError if some keys were duplicated: duplicated keys are considered undefined behaviour, + but we encourage you to refuse to process incoming messages with duplicated keys, as they might be + attempting to use HTTP Parameter Pollution like exploits. + """ + return super().decode_parameters() + def __str__(self): + return super().__str__() \ No newline at end of file diff --git a/zenoh/queryable.py b/zenoh/queryable.py new file mode 100644 index 00000000..774ad582 --- /dev/null +++ b/zenoh/queryable.py @@ -0,0 +1,70 @@ +# +# Copyright (c) 2022 ZettaScale Technology +# +# This program and the accompanying materials are made available under the +# terms of the Eclipse Public License 2.0 which is available at +# http://www.eclipse.org/legal/epl-2.0, or the Apache License, Version 2.0 +# which is available at https://www.apache.org/licenses/LICENSE-2.0. +# +# SPDX-License-Identifier: EPL-2.0 OR Apache-2.0 +# +# Contributors: +# ZettaScale Zenoh Team, <[email protected]> +# +from typing import Dict + +from .zenoh import _Query, _Queryable +from .keyexpr import KeyExpr, Selector +from .value import Sample + +class Queryable: + """ + A handle to a queryable. + + Its main purpose is to keep the queryable active as long as it exists. + + When constructed through `Session.declare_queryable(session, keyexpr, handler)`, it exposes `handler`'s receiver + through `self.receiver`. + """ + def __init__(self, inner: _Queryable, receiver): + self._inner_ = inner + self.receiver = receiver + + def undeclare(self): + "Stops the queryable." + self._inner_ = None + +class Query(_Query): + def __new__(cls, inner: _Query): + return super().__new__(cls, inner) + @property + def key_expr(self) -> KeyExpr: + "The query's targeted key expression" + return KeyExpr(super().key_expr) + @property + def parameters(self) -> str: + """ + The query's value selector. + If you'd rather not bother with parsing it yourself, use `self.decode_parameters()` instead. + """ + return super().parameters + + def decode_parameters(self) -> Dict[str, str]: + """ + Decodes the value selector into a dictionary. + + Raises a ZError if duplicate keys are found, as they might otherwise be used for HTTP Parameter Pollution like attacks. + """ + return super().decode_parameters + @property + def selector(self) -> Selector: + """ + The query's selector as a whole. + """ + return Selector._upgrade_(super().selector) + def reply(self, sample: Sample): + """ + Allows you to reply to a query. + You may send any amount of replies to a single query, including 0. + """ + super().reply(sample) \ No newline at end of file diff --git a/zenoh/session.py b/zenoh/session.py new file mode 100644 index 00000000..17271784 --- /dev/null +++ b/zenoh/session.py @@ -0,0 +1,245 @@ +# +# Copyright (c) 2022 ZettaScale Technology +# +# This program and the accompanying materials are made available under the +# terms of the Eclipse Public License 2.0 which is available at +# http://www.eclipse.org/legal/epl-2.0, or the Apache License, Version 2.0 +# which is available at https://www.apache.org/licenses/LICENSE-2.0. +# +# SPDX-License-Identifier: EPL-2.0 OR Apache-2.0 +# +# Contributors: +# ZettaScale Zenoh Team, <[email protected]> +# +from typing import Union, Any, List + +from .zenoh import _Session, _Config, _Publisher, _Subscriber, _PullSubscriber + +from .keyexpr import KeyExpr, IntoKeyExpr, Selector, IntoSelector +from .config import Config +from .closures import IntoHandler, Handler, Receiver +from .enums import * +from .value import IntoValue, Value, Sample, Reply, ZenohId +from .queryable import Queryable, Query + + +class Publisher: + "Use `Publisher`s (constructed with `Session.declare_publisher`) when you want to send values often for the same key expression, as declaring them informs Zenoh that this is you intent, and optimizations will be set up to do so." + + def __init__(self, p: _Publisher): + self._inner_ = p + + def put(self, value: IntoValue, encoding: Encoding = None): + "An optimised version of `session.put(self.key_expr, value, encoding=encoding)" + self._inner_.put(Value(value, encoding)) + + def delete(self): + "An optimised version of `session.delete(self.key_expr)" + self._inner_.delete() + + @property + def key_expr(self) -> KeyExpr: + "This `Publisher`'s key expression" + return KeyExpr(self._inner_.key_expr) + + def undeclare(self): + "Stops the publisher." + self._inner_ = None + + +class Subscriber: + """ + A handle to a subscription. + + Its main purpose is to keep the subscription active as long as it exists. + + When constructed through `Session.declare_subscriber(session, keyexpr, handler)`, it exposes `handler`'s receiver + through `self.receiver`. + """ + + def __init__(self, s: _Subscriber, receiver=None): + self._subscriber_ = s + self.receiver = receiver + + def undeclare(self): + "Undeclares the subscription" + self._subscriber_ = None + + +class PullSubscriber: + """ + A handle to a pull subscription. + + Its main purpose is to keep the subscription active as long as it exists. + + When constructed through `Session.declare_pull_subscriber(session, keyexpr, handler)`, it exposes `handler`'s receiver + through `self.receiver`. + + Calling `self.pull()` will prompt the Zenoh network to send a new sample when available. + """ + + def __init__(self, s: _PullSubscriber, receiver=None): + self._subscriber_ = s + self.receiver = receiver + + def pull(self): + """ + Prompts the Zenoh network to send a new sample if available. + Note that this sample will not be returned by this function, but provided to the handler's callback. + """ + self._subscriber_.pull() + + def undeclare(self): + "Undeclares the subscription" + self._subscriber_ = None + + +class Session(_Session): + """ + A Zenoh Session, the core interraction point with a Zenoh network. + """ + def __new__(cls, config: Union[Config, Any] = None): + if config is None: + return super().__new__(cls) + elif isinstance(config, _Config): + return super().__new__(cls, config) + else: + return super().__new__(cls, Config.from_obj(config)) + + def put(self, keyexpr: IntoKeyExpr, value: IntoValue, encoding=None, + priority: Priority = None, congestion_control: CongestionControl = None, + sample_kind: SampleKind = None): + """ + Sends a value over Zenoh. + """ + value = Value(value, encoding) + keyexpr = KeyExpr(keyexpr) + kwargs = dict() + if priority is not None: + kwargs['priority'] = priority + if congestion_control is not None: + kwargs['congestion_control'] = congestion_control + if sample_kind is not None: + kwargs['sample_kind'] = sample_kind + return super().put(keyexpr, value, **kwargs) + + def delete(self, keyexpr: IntoKeyExpr, + priority: Priority = None, congestion_control: CongestionControl = None): + """ + Deletes a value. + """ + keyexpr = KeyExpr(keyexpr) + kwargs = dict() + if priority is not None: + kwargs['priority'] = priority + if congestion_control is not None: + kwargs['congestion_control'] = congestion_control + return super().delete(keyexpr, **kwargs) + + def get(self, selector: IntoSelector, handler: IntoHandler[Reply, Any, Receiver], consolidation: QueryConsolidation = None, target: QueryTarget = None) -> Receiver: + """ + Emits a query. + """ + handler = Handler(handler, lambda x: Reply(x)) + kwargs = dict() + if consolidation is not None: + kwargs["conconsolidation"] = consolidation + if target is not None: + kwargs["target"] = target + super().get(Selector(selector), handler.closure, **kwargs) + return handler.receiver + + def declare_keyexpr(self, keyexpr: IntoKeyExpr) -> KeyExpr: + """ + Informs Zenoh that you intend to use the provided Key Expression repeatedly. + + This function returns an optimized representation of the passed `keyexpr`. + """ + return KeyExpr(super().declare_keyexpr(KeyExpr(keyexpr))) + + def declare_queryable(self, keyexpr: IntoKeyExpr, handler: IntoHandler[Query, Any, Any], complete: bool = None): + """ + Declares a queryable, which will receive queries intersecting with `keyexpr`. + + These queries are passed to the `handler` as instances of the `Query`class, which lets you respond when applicatble. + + The `handler`'s receiver is returned as the `receiver` field of the return value. + + IMPORTANT: due to how RAII and Python work, you MUST bind this function's return value to a variable in order for it to function as expected. + This is because as soon as a value is no longer referenced in Python, that value's destructor will run, which will undeclare your queryable, stopping it immediately. + """ + handler = Handler(handler, lambda x: Query(x)) + kwargs = dict() + if complete is not None: + kwargs['complete'] = complete + inner = super().declare_queryable(KeyExpr(keyexpr), handler.closure, **kwargs) + return Queryable(inner, handler.receiver) + + def declare_publisher(self, keyexpr: IntoKeyExpr, priority: Priority = None, congestion_control: CongestionControl = None): + """ + Declares a publisher, which you may use to send values repeatedly onto a same key expression. + """ + kwargs = dict() + if priority is not None: + kwargs['priority'] = priority + if congestion_control is not None: + kwargs['congestion_control'] = congestion_control + return Publisher(super().declare_publisher(KeyExpr(keyexpr), **kwargs)) + + def declare_subscriber(self, keyexpr: IntoKeyExpr, handler: IntoHandler[Sample, Any, Any], reliability: Reliability = None) -> Subscriber: + """ + Declares a subscriber, which will receive any published sample with a key expression intersecting `keyexpr`. + + These samples are passed to the `handler`'s closure as instances of the `Sample` class. + + The `handler`'s receiver is returned as the `receiver` field of the return value. + + IMPORTANT: due to how RAII and Python work, you MUST bind this function's return value to a variable in order for it to function as expected. + This is because as soon as a value is no longer referenced in Python, that value's destructor will run, which will undeclare your subscriber, deactivating the subscription immediately. + """ + handler = Handler(handler, lambda x: Sample._upgrade_(x)) + kwargs = dict() + if reliability is not None: + kwargs['reliability'] = reliability + s = super().declare_subscriber(KeyExpr(keyexpr), handler.closure, **kwargs) + return Subscriber(s, handler.receiver) + + def declare_pull_subscriber(self, keyexpr: IntoKeyExpr, handler: IntoHandler[Sample, Any, Any], reliability: Reliability = None) -> PullSubscriber: + """ + Declares a pull-mode subscriber, which will receive a single published sample with a key expression intersecting `keyexpr` any time its `pull` method is called. + + These samples are passed to the `handler`'s closure as instances of the `Sample` class. + + The `handler`'s receiver is returned as the `receiver` field of the return value. + """ + handler = Handler(handler, lambda x: Sample._upgrade_(x)) + kwargs = dict() + if reliability is not None: + kwargs['reliability'] = reliability + s = super().declare_pull_subscriber(KeyExpr(keyexpr), handler.closure, **kwargs) + return PullSubscriber(s, handler.receiver) + + def close(self): + "Closes the Session" + pass + + def info(self): + "Returns an accessor for informations about this Session" + return Info(self) + + +class Info: + def __init__(self, session: _Session): + self.session = session + + def zid(self) -> ZenohId: + "Returns this Zenoh Session's identifier" + return ZenohId._upgrade_(self.session.zid()) + + def routers_zid(self) -> List[ZenohId]: + "Returns the neighbooring routers' identifiers" + return [ZenohId._upgrade_(id) for id in self.session.routers_zid()] + + def peers_zid(self) -> List[ZenohId]: + "Returns the neighbooring peers' identifiers" + return [ZenohId._upgrade_(id) for id in self.session.peers_zid()] diff --git a/zenoh/value.py b/zenoh/value.py new file mode 100644 index 00000000..1bc6c44a --- /dev/null +++ b/zenoh/value.py @@ -0,0 +1,205 @@ +# +# Copyright (c) 2022 ZettaScale Technology +# +# This program and the accompanying materials are made available under the +# terms of the Eclipse Public License 2.0 which is available at +# http://www.eclipse.org/legal/epl-2.0, or the Apache License, Version 2.0 +# which is available at https://www.apache.org/licenses/LICENSE-2.0. +# +# SPDX-License-Identifier: EPL-2.0 OR Apache-2.0 +# +# Contributors: +# ZettaScale Zenoh Team, <[email protected]> +# +import abc +from typing import Union, Tuple, Optional, List +import json + +from .enums import Encoding, SampleKind +from .zenoh import _Value, _Encoding, _Sample, _SampleKind, _Reply, _ZenohId, _Timestamp, _Hello +from .keyexpr import KeyExpr, IntoKeyExpr + +class IValue: + "The IValue interface exposes how to recover a value's payload in a binary-serialized format, as well as that format's encoding." + @property + @abc.abstractmethod + def payload(self) -> bytes: + ... + + @property + @abc.abstractmethod + def encoding(self) -> Encoding: + ... + +IntoValue = Union[IValue, bytes, str, int, float, object] + +class Value(_Value, IValue): + """ + A Value is a pair of a binary payload, and a mime-type-like encoding string. + + When constructed with `encoding==None`, the encoding will be selected depending on the payload's type. + """ + def __new__(cls, payload: IntoValue, encoding: Encoding=None): + if encoding is None: + if isinstance(payload, Value): + return payload + return Value.autoencode(payload) + else: + if not isinstance(payload, bytes): + raise TypeError("`encoding` was passed, but `payload` is not of type `bytes`") + return Value.new(payload, encoding) + + @staticmethod + def autoencode(value: IntoValue) -> 'Value': + if isinstance(value, IValue): + return Value.new(value.payload, value.encoding) + if isinstance(value, bytes): + return Value.new(value, Encoding.APP_OCTET_STREAM()) + if isinstance(value, str): + return Value.new(value.encode(), Encoding.TEXT_PLAIN()) + if isinstance(value, int): + return Value.new(f"{value}".encode(), Encoding.APP_INTEGER()) + if isinstance(value, float): + return Value.new(f"{value}".encode(), Encoding.APP_FLOAT()) + return Value.new(json.dumps(value).encode(), Encoding.APP_JSON()) + + @staticmethod + def new(payload: bytes, encoding: Encoding = None) -> 'Value': + return Value._upgrade_(_Value.new(payload, encoding)) + + @property + def payload(self) -> bytes: + return super().payload + + @payload.setter + def payload(self, payload: bytes): + super().with_payload(payload) + + @property + def encoding(self) -> Encoding: + return Encoding(super().encoding) + + @encoding.setter + def encoding(self, encoding: Encoding): + super().with_encoding(encoding) + + @staticmethod + def _upgrade_(inner: _Value) -> 'Value': + if isinstance(inner, Value): + return inner + return _Value.__new__(Value, inner) + +class ZenohId(_ZenohId): + """A Zenoh UUID""" + @staticmethod + def _upgrade_(this: _ZenohId) -> 'ZenohId': + return _ZenohId.__new__(ZenohId, this) + def __str__(self) -> str: + return super().__str__() + def __repr__(self) -> str: + return str(self) + +class Timestamp(_Timestamp): + """ + A timestamp taken from the Zenoh HLC (Hybrid Logical Clock). + + These timestamps are guaranteed to be unique, as each machine annotates its perceived time with a UUID, which is used as the least significant part of the comparison operation. + """ + @staticmethod + def _upgrade_(this: _Timestamp) -> 'Timestamp': + return _Timestamp.__new__(Timestamp, this) + @property + def seconds_since_unix_epoch(self) -> float: + """ + Returns the number of seconds since the Unix Epoch. + + You shouldn't use this for comparison though, and rely on comparison operators between members of this class. + """ + return super().seconds_since_unix_epoch + + +IntoSample = Union[_Sample, Tuple[IntoKeyExpr, IntoValue, SampleKind], Tuple[KeyExpr, IntoValue]] +class Sample(_Sample): + """ + A KeyExpr-Value pair, annotated with the kind (PUT or DELETE) of publication used to emit it and a timestamp. + """ + def __new__(cls, key: IntoKeyExpr, value: IntoValue, kind: SampleKind = None, timestamp: Timestamp = None): + kind = _SampleKind.PUT if kind is None else kind + return Sample._upgrade_(super().new(KeyExpr(key), Value(value), kind, timestamp)) + @property + def key_expr(self) -> KeyExpr: + "The sample's key expression" + return KeyExpr(super().key_expr) + @property + def value(self) -> Value: + "The sample's value" + return Value._upgrade_(super().value) + @property + def payload(self) -> bytes: + "A shortcut to `self.value.payload`" + return super().payload + @property + def encoding(self) -> Encoding: + "A shortcut to `self.value.encoding`" + return Encoding(super().encoding) + @property + def kind(self) -> SampleKind: + "The sample's kind" + return SampleKind(super().kind) + @property + def timestamp(self) -> Optional[Timestamp]: + "The sample's timestamp. May be None." + ts = super().timestamp + return None if ts is None else Timestamp._upgrade_(ts) + @staticmethod + def _upgrade_(inner: _Sample) -> 'Sample': + if isinstance(inner, Sample): + return inner + return _Sample.__new__(Sample, inner) + +class Reply(_Reply): + def __new__(cls, inner: _Reply): + return super().__new__(cls, inner) + @property + def replier_id(self) -> ZenohId: + "The reply's sender's id." + return ZenohId._upgrade_(super().replier_id) + @property + def ok(self) -> Sample: + """ + The reply's inner data sample. + + Raises a ZError if the `self` is actually an `err` reply. + """ + return Sample._upgrade_(super().ok) + @property + def err(self) -> Value: + """ + The reply's error value. + + Raises a ZError if the `self` is actually an `ok` reply. + """ + return Value._upgrade_(super().err) + +class Hello(_Hello): + "Represents a single Zenoh node discovered through scouting." + @property + def zid(self) -> ZenohId: + "The node's Zenoh UUID." + zid = super().zid + return None if zid is None else ZenohId._upgrade_(zid) + @property + def whatami(self) -> str: + "The node's type, returning either None, 'peer', 'router', or 'client'." + return super().whatami + @property + def locators(self) -> List[str]: + "The locators through which this node may be adressed." + return super().locators + @staticmethod + def _upgrade_(inner: _Hello) -> 'Sample': + if isinstance(inner, Hello): + return inner + return _Hello.__new__(Hello, inner) + def __str__(self): + return super().__str__() \ No newline at end of file

Point to master branch This is a preparatory PR. It will be merged once https://github.com/eclipse-zenoh/zenoh/pull/347 is merged and `Cargo.lock` is synced.

2022-09-22T14:18:12

0.0

[]

jefflester/minitrino

jefflester__minitrino-36

a71ce5a2cb0d596e0fdf3d0b419c7ee2be6616c7

diff --git a/cli/minitrino/cmd/cmd_provision.py b/cli/minitrino/cmd/cmd_provision.py index 3b57e653..f47e60d9 100644 --- a/cli/minitrino/cmd/cmd_provision.py +++ b/cli/minitrino/cmd/cmd_provision.py @@ -77,6 +77,7 @@ def cli(ctx, modules, no_rollback, docker_native): utils.check_daemon(ctx.docker_client) utils.check_lib(ctx) + utils.check_starburst_ver(ctx) modules = append_running_modules(modules) check_compatibility(modules) check_enterprise(modules) @@ -391,12 +392,14 @@ def check_dup_configs(ctx): for i, config in enumerate(configs): if config.startswith("#"): continue - config = utils.parse_key_value_pair(config, err_type=err.UserError) + config = utils.parse_key_value_pair( + config, err_type=err.UserError, key_to_upper=False + ) if config is None: continue if i + 1 != len(configs): next_config = utils.parse_key_value_pair( - configs[i + 1], err_type=err.UserError + configs[i + 1], err_type=err.UserError, key_to_upper=False ) if config[0] == next_config[0]: duplicates.extend(["=".join(config), "=".join(next_config)]) @@ -506,7 +509,7 @@ def append_configs(user_configs, current_configs, filename): if filename == TRINO_CONFIG: for user_config in user_configs: user_config = utils.parse_key_value_pair( - user_config, err_type=err.UserError + user_config, err_type=err.UserError, key_to_upper=False ) if user_config is None: continue @@ -514,7 +517,7 @@ def append_configs(user_configs, current_configs, filename): if current_config.startswith("#"): continue current_config = utils.parse_key_value_pair( - current_config, err_type=err.UserError + current_config, err_type=err.UserError, key_to_upper=False ) if current_config is None: continue diff --git a/cli/minitrino/cmd/cmd_snapshot.py b/cli/minitrino/cmd/cmd_snapshot.py index bcecfbbf..5ea62976 100644 --- a/cli/minitrino/cmd/cmd_snapshot.py +++ b/cli/minitrino/cmd/cmd_snapshot.py @@ -205,8 +205,8 @@ def build_command_string(ctx, modules=[]): if modules: options = "" for module in modules: - options += f" {module}" - option_string = f"--module {options}" + options += f"--module {module} " + option_string = f"{options}" bash_source = '"${BASH_SOURCE%/*}"' command_string = ( diff --git a/cli/minitrino/utils.py b/cli/minitrino/utils.py index 6bbcbcc8..0f952a24 100644 --- a/cli/minitrino/utils.py +++ b/cli/minitrino/utils.py @@ -240,6 +240,23 @@ def check_lib(ctx): ctx.minitrino_lib_dir +def check_starburst_ver(ctx): + """Checks if a proper Starburst version is provided.""" + + starburst_ver = ctx.env.get_var("STARBURST_VER", "") + error_msg = ( + f"Provided Starburst version '{starburst_ver}' is invalid. " + f"The provided version must be 354-e or higher." + ) + + try: + starburst_ver_int = int(starburst_ver[0:3]) + if starburst_ver_int < 354 or "-e" not in starburst_ver: + raise err.UserError(error_msg) + except: + raise err.UserError(error_msg) + + def generate_identifier(identifiers=None): """Returns an 'object identifier' string used for creating log messages, e.g. '[ID: 12345] [Name: trino]'. @@ -263,7 +280,9 @@ def generate_identifier(identifiers=None): return " ".join(identifier) -def parse_key_value_pair(key_value_pair, err_type=err.MinitrinoError): +def parse_key_value_pair( + key_value_pair, err_type=err.MinitrinoError, key_to_upper=True +): """Parses a key-value pair in string form and returns the resulting pair as both a 2-element list. If the string cannot be split by "=", a MinitrinoError is raised. @@ -273,6 +292,7 @@ def parse_key_value_pair(key_value_pair, err_type=err.MinitrinoError): `"TRINO=354-e"`. - `err_type`: The exception to raise if an "=" delimiter is not in the key-value pair. Defaults to `MinitrinoError`. + - `key_to_upper`: If `True`, the key will be forced to uppercase. ### Return Values - A list `[k, v]`, but will return `None` if the stripped input is an empty @@ -298,6 +318,8 @@ def parse_key_value_pair(key_value_pair, err_type=err.MinitrinoError): key_value_pair[i] = key_value_pair[i].strip() if not key_value_pair[0]: raise err_type(err_msg) + elif key_to_upper: + key_value_pair[0] = key_value_pair[0].upper() if not len(key_value_pair) == 2: raise err_type(err_msg) diff --git a/cli/setup.py b/cli/setup.py index 758a31e9..673d7dac 100644 --- a/cli/setup.py +++ b/cli/setup.py @@ -10,7 +10,7 @@ setup( name="minitrino", - version="2.0.0", + version="2.0.1", description="A command line tool that makes it easy to run modular Trino environments locally.", long_description=README, long_description_content_type="text/markdown", diff --git a/lib/minitrino.env b/lib/minitrino.env index 18716b81..28d63dbd 100644 --- a/lib/minitrino.env +++ b/lib/minitrino.env @@ -2,9 +2,11 @@ COMPOSE_PROJECT_NAME=minitrino STARBURST_VER=354-e +DB2_VER=11.5.4.0 ELASTICSEARCH_VER=7.6.2 MYSQL_VER=5 OPEN_LDAP_VER=1.3.0 +ORACLE_VER=12.2.0.1 POSTGRES_VER=11 POSTGRES_DELTA_LAKE_VER=11 POSTGRES_HIVE_MINIO_VER=11 @@ -13,3 +15,7 @@ POSTGRES_EVENT_LOGGER_VER=11 POSTGRES_RANGER_VER=12.2 SEP_RANGER_VER=2.0.49 SQLSERVER_VER=2017-latest + +METASTORE_HIVE_S3_PORT=9085 +METASTORE_DELTA_LAKE_PORT=9084 +MINIO_DELTA_LAKE_PORT=9001 diff --git a/lib/modules/catalog/db2/db2.yml b/lib/modules/catalog/db2/db2.yml new file mode 100644 index 00000000..77fc2651 --- /dev/null +++ b/lib/modules/catalog/db2/db2.yml @@ -0,0 +1,22 @@ +version: "3.8" +services: + + trino: + volumes: + - "./modules/catalog/db2/resources/trino/db2.properties:/etc/starburst/catalog/db2.properties" + + db2: + image: "ibmcom/db2:${DB2_VER}" + container_name: "db2" + labels: + - "com.starburst.tests=minitrino" + - "com.starburst.tests.module.db2=catalog-db2" + env_file: + - "./modules/catalog/db2/resources/db2/db2.env" + ports: + - "50000:50000" + volumes: + - ./modules/catalog/db2/resources/db2/db2-storage/:/database + privileged: true + tty: true + stdin_open: true diff --git a/lib/modules/catalog/db2/metadata.json b/lib/modules/catalog/db2/metadata.json new file mode 100644 index 00000000..65f17414 --- /dev/null +++ b/lib/modules/catalog/db2/metadata.json @@ -0,0 +1,5 @@ +{ + "description": "Creates a Db2 catalog using the Db2 connector.", + "incompatibleModules": [], + "enterprise": true +} \ No newline at end of file diff --git a/lib/modules/catalog/db2/readme.md b/lib/modules/catalog/db2/readme.md new file mode 100644 index 00000000..020948d2 --- /dev/null +++ b/lib/modules/catalog/db2/readme.md @@ -0,0 +1,10 @@ +# Db2 Connector Module + +This module provisions a standalone Db2 service. + +## Usage + + minitrino --env STARBURST_VER=<ver> provision --module db2 + docker exec -it trino bash + trino-cli + trino> show schemas from db2; diff --git a/lib/modules/catalog/db2/resources/db2/db2-storage/.gitkeep b/lib/modules/catalog/db2/resources/db2/db2-storage/.gitkeep new file mode 100644 index 00000000..e69de29b diff --git a/lib/modules/catalog/db2/resources/db2/db2.env b/lib/modules/catalog/db2/resources/db2/db2.env new file mode 100644 index 00000000..7e304ad5 --- /dev/null +++ b/lib/modules/catalog/db2/resources/db2/db2.env @@ -0,0 +1,3 @@ +DBNAME=minitrino +DB2INST1_PASSWORD=trinoRocks15 +LICENSE=accept diff --git a/lib/modules/catalog/db2/resources/trino/db2.properties b/lib/modules/catalog/db2/resources/trino/db2.properties new file mode 100644 index 00000000..983cd676 --- /dev/null +++ b/lib/modules/catalog/db2/resources/trino/db2.properties @@ -0,0 +1,4 @@ +connector.name=db2 +connection-url=jdbc:db2://db2:50000/minitrino +connection-user=DB2INST1 +connection-password=trinoRocks15 diff --git a/lib/modules/catalog/delta-lake/delta-lake.yml b/lib/modules/catalog/delta-lake/delta-lake.yml index e8dd24b4..e37860d5 100644 --- a/lib/modules/catalog/delta-lake/delta-lake.yml +++ b/lib/modules/catalog/delta-lake/delta-lake.yml @@ -37,13 +37,13 @@ services: "./opt/bin/start-hive-metastore.sh", ] ports: - - "9083:9084" + - "${METASTORE_DELTA_LAKE_PORT}:9083" minio-delta-lake: image: "minio/minio" container_name: "minio-delta-lake" ports: - - "9000:9001" + - "${MINIO_DELTA_LAKE_PORT}:9000" volumes: - "./modules/catalog/delta-lake/resources/minio/data/:/data" labels: diff --git a/lib/modules/catalog/delta-lake/readme.md b/lib/modules/catalog/delta-lake/readme.md index 47b36490..eb6eedbc 100644 --- a/lib/modules/catalog/delta-lake/readme.md +++ b/lib/modules/catalog/delta-lake/readme.md @@ -1,14 +1,20 @@ # Delta-Lake Module + This module uses the Delta Lake connector. There is no Spark backend, so tables -need to be created via CTAS queries from Trino. Example: +need to be created via `CREATE TABLE AS ...` queries from Trino. Example: -``` -CREATE TABLE delta.default.customer -WITH ( - location = 's3a://sample-bucket/default/' -) -AS SELECT * FROM tpch.tiny.customer; -``` + CREATE TABLE delta.default.customer + WITH ( + location = 's3a://sample-bucket/default/' + ) + AS SELECT * FROM tpch.tiny.customer; This will create the table `delta.default.customer` and a corresponding `_delta_log` directory in the backing MinIO object storage. + +## Usage + + minitrino --env STARBURST_VER=<ver> provision --module delta-lake + docker exec -it trino bash + trino-cli + trino> show schemas from delta; diff --git a/lib/modules/catalog/elasticsearch/readme.md b/lib/modules/catalog/elasticsearch/readme.md index 4758bc50..6c275a6f 100644 --- a/lib/modules/catalog/elasticsearch/readme.md +++ b/lib/modules/catalog/elasticsearch/readme.md @@ -1,41 +1,49 @@ # Elasticsearch Connector Module -This module contains an ES container with some preloaded data. It contains: a schema (ES mapping), a table (ES doc mapping), and data (ES docs). + +This module contains an ES container with some preloaded data. It contains: a +schema (ES mapping), a table (ES doc mapping), and data (ES docs). ## Loading your own data + Since port 9200 is exposed on localhost you can add your own data like this: -```bash -# Create user index -curl -XPUT http://localhost:9200/user?pretty=true; - -# Create user mapping -curl -XPUT http://localhost:9200/user/_mapping/profile?include_type_name=true -H 'Content-Type: application/json' -d ' -{ - "profile" : { - "properties" : { - "full_name" : { "type" : "text", "store" : true }, - "bio" : { "type" : "text", "store" : true }, - "age" : { "type" : "integer" }, - "location" : { "type" : "geo_point" }, - "enjoys_coffee" : { "type" : "boolean" }, - "created_on" : { - "type" : "date", - "format" : "date_time" + # Create user index + curl -XPUT http://localhost:9200/user?pretty=true; + + # Create user mapping + curl -XPUT http://localhost:9200/user/_mapping/profile?include_type_name=true -H 'Content-Type: application/json' -d ' + { + "profile" : { + "properties" : { + "full_name" : { "type" : "text", "store" : true }, + "bio" : { "type" : "text", "store" : true }, + "age" : { "type" : "integer" }, + "location" : { "type" : "geo_point" }, + "enjoys_coffee" : { "type" : "boolean" }, + "created_on" : { + "type" : "date", + "format" : "date_time" + } } } } -} -'; - -# Create user profile records -curl -XPOST http://localhost:9200/user/profile/1?pretty=true -H 'Content-Type: application/json' -d ' -{ - "full_name" : "Andrew Puch", - "bio" : "My name is Andrew. I have a short bio.", - "age" : 26, - "location" : "41.1246110,-73.4232880", - "enjoys_coffee" : true, - "created_on" : "2015-05-02T14:45:10.000-04:00" -} -'; -``` \ No newline at end of file + '; + + # Create user profile records + curl -XPOST http://localhost:9200/user/profile/1?pretty=true -H 'Content-Type: application/json' -d ' + { + "full_name" : "Andrew Puch", + "bio" : "My name is Andrew. I have a short bio.", + "age" : 26, + "location" : "41.1246110,-73.4232880", + "enjoys_coffee" : true, + "created_on" : "2015-05-02T14:45:10.000-04:00" + } + '; + +## Usage + + minitrino --env STARBURST_VER=<ver> provision --module elasticsearch + docker exec -it trino bash + trino-cli + trino> show schemas from elasticsearch; diff --git a/lib/modules/catalog/hive-minio/readme.md b/lib/modules/catalog/hive-minio/readme.md index 066fb8a1..b189ed58 100644 --- a/lib/modules/catalog/hive-minio/readme.md +++ b/lib/modules/catalog/hive-minio/readme.md @@ -1,18 +1,32 @@ # Hive-Minio Module -This module uses Minio as a local implementation of S3 object storage. You can write data to this service, and the files will be written to your machine. You can read more about Minio [here](https://docs.min.io/docs/minio-docker-quickstart-guide.html). This module also uses a Hive metastore container along with a Postgres container for the metastore's backend storage. -You can access the Minio UI at `http://localhost:9000` with `access-key` and `secret-key` for credentials. +This module uses Minio as a local implementation of S3 server. You can write +data to this service and the files will be written to your machine. You can read +more about Minio +[here](https://docs.min.io/docs/minio-docker-quickstart-guide.html). This module +also uses a Hive metastore container along with a Postgres container for the +metastore's backend storage. + +You can access the Minio UI at `http://localhost:9000` with `access-key` and +`secret-key` for credentials. You can create a table with ORC data with Trino very quickly: -``` -trino> create schema hive_minio.tiny with (location='s3a://sample-bucket/tiny/'); -CREATE SCHEMA + trino> create schema hive_minio.tiny with (location='s3a://sample-bucket/tiny/'); + CREATE SCHEMA -trino> create table hive_minio.tiny.customer as select * from tpch.tiny.customer; -CREATE TABLE: 1500 rows -``` + trino> create table hive_minio.tiny.customer as select * from tpch.tiny.customer; + CREATE TABLE: 1500 rows You will see the ORC data stored in your local Minio bucket. -Note the [relevant commit](https://github.com/starburstdata/docker-images/commit/6b29c2359a173ca6971267fa05191258b1964c8b#diff-8961ce993089ebecf98d0457b676e626) for the property `S3_PATH_STYLE_ACCESS` in `hms.env`. +Note the [relevant +commit](https://github.com/starburstdata/docker-images/commit/6b29c2359a173ca6971267fa05191258b1964c8b#diff-8961ce993089ebecf98d0457b676e626) +for the property `S3_PATH_STYLE_ACCESS` in `hms.env`. + +## Usage + + minitrino --env STARBURST_VER=<ver> provision --module hive-minio + docker exec -it trino bash + trino-cli + trino> show schemas from hive_minio; diff --git a/lib/modules/catalog/hive-s3/hive-s3.yml b/lib/modules/catalog/hive-s3/hive-s3.yml index 930643b6..c3cd19b4 100644 --- a/lib/modules/catalog/hive-s3/hive-s3.yml +++ b/lib/modules/catalog/hive-s3/hive-s3.yml @@ -42,7 +42,7 @@ services: "./opt/bin/start-hive-metastore.sh", ] ports: - - "9083:9085" + - "${METASTORE_HIVE_S3_PORT}:9083" volumes: postgres-hive-s3-data: diff --git a/lib/modules/catalog/hive-s3/readme.md b/lib/modules/catalog/hive-s3/readme.md index b4765a7b..2b182d42 100644 --- a/lib/modules/catalog/hive-s3/readme.md +++ b/lib/modules/catalog/hive-s3/readme.md @@ -1,2 +1,15 @@ # Hive-S3 Module -This module uses a Hive metastore container along with a Postgres container for the metastore's backend storage. Using Minitrino's configuration, you can input AWS credentials to link the module to a an S3 bucket for data storage, reading, and writing. Removing the volume associated with the Postgres container will effectively remove your metastore's database, so only remove the volume if you are prepared to lose all of your metadata. \ No newline at end of file + +This module uses a Hive metastore container along with a Postgres container for +the metastore's backend storage. Using Minitrino's config file, you can input +AWS credentials to link the module to a an S3 bucket for data storage, reading, +and writing. Removing the volume associated with the Postgres container will +effectively remove your metastore's database, so only remove the volume if you +are prepared to lose all of your metadata. + +## Usage + + minitrino --env STARBURST_VER=<ver> provision --module hive-s3 + docker exec -it trino bash + trino-cli + trino> show schemas from hive_s3; diff --git a/lib/modules/catalog/mysql/mysql.yml b/lib/modules/catalog/mysql/mysql.yml index 0968f4c4..37e41f43 100644 --- a/lib/modules/catalog/mysql/mysql.yml +++ b/lib/modules/catalog/mysql/mysql.yml @@ -15,3 +15,5 @@ services: - "./modules/catalog/mysql/resources/mysql/mysql.env" environment: MINITRINO_BOOTSTRAP: "bootstrap.sh" + ports: + - "3306:3306" diff --git a/lib/modules/catalog/mysql/readme.md b/lib/modules/catalog/mysql/readme.md index b71de207..9a5cf1f7 100644 --- a/lib/modules/catalog/mysql/readme.md +++ b/lib/modules/catalog/mysql/readme.md @@ -1,2 +1,11 @@ -# mysql Connector Module -This module provisions a standalone MySQL service. Other modules that uses MySQL as a backend will need a more unique name to avoid conflicts with this one. +# MySQL Connector Module + +This module provisions a standalone MySQL service. Other modules that uses MySQL +as a backend will need a more unique name to avoid conflicts with this one. + +## Usage + + minitrino --env STARBURST_VER=<ver> provision --module mysql + docker exec -it trino bash + trino-cli + trino> show schemas from mysql; diff --git a/lib/modules/catalog/oracle/metadata.json b/lib/modules/catalog/oracle/metadata.json new file mode 100644 index 00000000..0149edc2 --- /dev/null +++ b/lib/modules/catalog/oracle/metadata.json @@ -0,0 +1,5 @@ +{ + "description": "Creates an Oracle catalog using the standard Oracle connector.", + "incompatibleModules": [], + "enterprise": false +} \ No newline at end of file diff --git a/lib/modules/catalog/oracle/oracle.yml b/lib/modules/catalog/oracle/oracle.yml new file mode 100644 index 00000000..2362e40a --- /dev/null +++ b/lib/modules/catalog/oracle/oracle.yml @@ -0,0 +1,19 @@ +version: "3.8" +services: + + trino: + volumes: + - "./modules/catalog/oracle/resources/trino/oracle.properties:/etc/starburst/catalog/oracle.properties" + + oracle: + image: "store/oracle/database-enterprise:${ORACLE_VER}" + container_name: "oracle" + labels: + - "com.starburst.tests=minitrino" + - "com.starburst.tests.module.oracle=catalog-oracle" + environment: + MINITRINO_BOOTSTRAP: "bootstrap-oracle.sh" + env_file: + - "./modules/catalog/oracle/resources/oracle/ora.conf" + ports: + - "1521:1521" diff --git a/lib/modules/catalog/oracle/readme.md b/lib/modules/catalog/oracle/readme.md new file mode 100644 index 00000000..5d2b5ef5 --- /dev/null +++ b/lib/modules/catalog/oracle/readme.md @@ -0,0 +1,27 @@ +# Oracle Connector Module + +This module provisions a standalone Oracle service. + +## Usage + +You will want to use the `trino` schema within the `oracle` catalog. +Additionally, the Oracle server takes some time to become available after the +container boots up. If you see an `Unable to start the Universal Connection +Pool` error, wait 1-2 minutes and try querying Oracle again. + + # Login with the Docker Hub account used to "purchase" the image + echo "your-password" | docker login -u <user> --password-stdindocker + minitrino --env STARBURST_VER=<ver> provision --module oracle + docker exec -it trino bash + trino-cli + trino> show schemas from oracle; + trino> create table oracle.trino.test (a int); + +## Image Pull Notes + +To pull this image, you will need to: + +- Create a Docker Hub account and then "purchase" the [developer-tier + image](https://hub.docker.com/_/oracle-database-enterprise-edition). +- Authenticate to your Docker Hub account via the Docker CLI (see `docker login + --help` for more information) diff --git a/lib/modules/catalog/oracle/resources/bootstrap/bootstrap-oracle.sh b/lib/modules/catalog/oracle/resources/bootstrap/bootstrap-oracle.sh new file mode 100755 index 00000000..8c039d20 --- /dev/null +++ b/lib/modules/catalog/oracle/resources/bootstrap/bootstrap-oracle.sh @@ -0,0 +1,80 @@ +#!/usr/bin/env bash + +set -ex + +function health_check() { + su - oracle -c "source /home/oracle/.bashrc; sqlplus sys/Oradoc_db1 as sysdba @/tmp/health-check.sql" +} + +function create_user() { + su - oracle -c "source /home/oracle/.bashrc; sqlplus sys/Oradoc_db1 as sysdba @/tmp/create-user.sql" +} + +function grant_privileges() { + su - oracle -c "source /home/oracle/.bashrc; sqlplus sys/Oradoc_db1 as sysdba @/tmp/grant-privileges.sql" +} + +echo "Creating health check SQL script..." +cat <<EOT >> /tmp/health-check.sql +SELECT INSTANCE_NAME, STATUS, DATABASE_STATUS FROM V\$INSTANCE; +exit; +EOT + +echo "Creating user-creation SQL script..." +cat <<EOT >> /tmp/create-user.sql +ALTER SESSION SET "_ORACLE_SCRIPT"=true; +CREATE USER trino IDENTIFIED BY trinoRocks15; +exit; +EOT + +echo "Creating user-privilege SQL script..." +cat <<EOT >> /tmp/grant-privileges.sql +ALTER SESSION SET "_ORACLE_SCRIPT"=true; +GRANT CONNECT, RESOURCE, DBA TO trino; +GRANT CREATE SESSION, CREATE TABLE TO trino; +GRANT UNLIMITED TABLESPACE TO trino; +GRANT ALL PRIVILEGES TO trino; +exit; +EOT + +echo "Performing health checks..." +COUNTER=0 && set +e +while [[ "${COUNTER}" -lt 121 ]]; do + if health_check | grep -q "trinosid\|OPEN\|ACTIVE"; then + break + elif [[ "${COUNTER}" == 121 ]]; then + echo "Database is not up and running and timed out after approx. 2 minutes. Exiting" + exit 1 + else + sleep 1 + ((COUNTER++)) + fi +done + +echo "Creating user..." +COUNTER=0 && set +e +while [[ "${COUNTER}" -lt 61 ]]; do + if create_user | grep -q "ERROR"; then + sleep 1 + ((COUNTER++)) + elif [[ "${COUNTER}" == 61 ]]; then + echo "User creation failed after approx. 1 minute. Exiting" + exit 1 + else + break + fi +done + +echo "Granting user privileges..." +COUNTER=0 && set +e +while [[ "${COUNTER}" -lt 61 ]]; do + if grant_privileges | grep -q "ERROR"; then + sleep 1 + ((COUNTER++)) + elif [[ "${COUNTER}" == 61 ]]; then + echo "Privilege grant failed after approx. 1 minute. Exiting" + exit 1 + else + break + fi +done diff --git a/lib/modules/catalog/oracle/resources/oracle/ora.conf b/lib/modules/catalog/oracle/resources/oracle/ora.conf new file mode 100644 index 00000000..64e7c22a --- /dev/null +++ b/lib/modules/catalog/oracle/resources/oracle/ora.conf @@ -0,0 +1,3 @@ +DB_PDB=trino +DB_SID=trinosid +DB_MEMORY=2GB diff --git a/lib/modules/catalog/oracle/resources/trino/oracle.properties b/lib/modules/catalog/oracle/resources/trino/oracle.properties new file mode 100644 index 00000000..0c81e14a --- /dev/null +++ b/lib/modules/catalog/oracle/resources/trino/oracle.properties @@ -0,0 +1,4 @@ +connector.name=oracle +connection-url=jdbc:oracle:thin:@(DESCRIPTION=(ADDRESS=(PROTOCOL=TCP)(HOST=oracle)(PORT=1521))(CONNECT_DATA=(SERVER=DEDICATED)(SERVICE_NAME=trinosid.localdomain))) +connection-user=trino +connection-password=trinoRocks15 diff --git a/lib/modules/catalog/postgres/postgres.yml b/lib/modules/catalog/postgres/postgres.yml index 229ab71e..cb0d2117 100644 --- a/lib/modules/catalog/postgres/postgres.yml +++ b/lib/modules/catalog/postgres/postgres.yml @@ -13,3 +13,5 @@ services: - "com.starburst.tests.module.postgres=catalog-postgres" env_file: - "./modules/catalog/postgres/resources/postgres/postgres.env" + ports: + - "5432:5432" diff --git a/lib/modules/catalog/postgres/readme.md b/lib/modules/catalog/postgres/readme.md index 66b108fb..d3ea8c18 100644 --- a/lib/modules/catalog/postgres/readme.md +++ b/lib/modules/catalog/postgres/readme.md @@ -1,2 +1,12 @@ # Postgres Connector Module -This module provisions a standalone Postgres service. It is named uniquely to avoid conflicts with other modules that may use Trino as a backend, such as the `hive-s3` module. \ No newline at end of file + +This module provisions a standalone Postgres service. It is named uniquely to +avoid conflicts with other modules that may use Trino as a backend, such as the +`hive-s3` module. + +## Usage + + minitrino --env STARBURST_VER=<ver> provision --module postgres + docker exec -it trino bash + trino-cli + trino> show schemas from postgres; diff --git a/lib/modules/catalog/snowflake-distributed/readme.md b/lib/modules/catalog/snowflake-distributed/readme.md index 2e18ff4f..bc50901d 100644 --- a/lib/modules/catalog/snowflake-distributed/readme.md +++ b/lib/modules/catalog/snowflake-distributed/readme.md @@ -1,2 +1,12 @@ # Snowflake Distributed Connector Module -Placeholder readme \ No newline at end of file + +This module hooks up to an externally-hosted Snowflake service and leverages the +[distributed Snowflake +connector](https://docs.starburst.io/latest/connector/starburst-snowflake.html). + +## Usage + + minitrino --env STARBURST_VER=<ver> provision --module snowflake-distributed + docker exec -it trino bash + trino-cli + trino> show schemas from snowflake_distributed; diff --git a/lib/modules/catalog/snowflake-jdbc/readme.md b/lib/modules/catalog/snowflake-jdbc/readme.md index 35b2507d..aa777d58 100644 --- a/lib/modules/catalog/snowflake-jdbc/readme.md +++ b/lib/modules/catalog/snowflake-jdbc/readme.md @@ -1,2 +1,12 @@ # Snowflake JDBC Connector Module -Placeholder readme \ No newline at end of file + +This module hooks up to an externally-hosted Snowflake service and leverages the +[JDBC Snowflake +connector](https://docs.starburst.io/latest/connector/starburst-snowflake.html). + +## Usage + + minitrino --env STARBURST_VER=<ver> provision --module snowflake-jdbc + docker exec -it trino bash + trino-cli + trino> show schemas from snowflake_jdbc; diff --git a/lib/modules/catalog/sqlserver/readme.md b/lib/modules/catalog/sqlserver/readme.md index 5906cca6..cdd757fa 100644 --- a/lib/modules/catalog/sqlserver/readme.md +++ b/lib/modules/catalog/sqlserver/readme.md @@ -1,6 +1,15 @@ # SQL Server Connector Module + This module provisions a standalone SQL Server service. Default database created is `master`. -Note that the 2017 version of SQL Server is used by default, as previous versions were only available on Windows and do not have Docker containers. +Note that the 2017 version of SQL Server is used by default, as previous +versions were only available on Windows and do not have Docker containers. + +## Usage + + minitrino --env STARBURST_VER=<ver> provision --module sqlserver + docker exec -it trino bash + trino-cli + trino> show schemas from sqlserver; diff --git a/lib/modules/security/event-logger/readme.md b/lib/modules/security/event-logger/readme.md index 2e169591..81a3e983 100644 --- a/lib/modules/security/event-logger/readme.md +++ b/lib/modules/security/event-logger/readme.md @@ -1,2 +1,13 @@ # Event Logger Module -Placeholder readme \ No newline at end of file + +This module configures and deploys the necessary components for the [Starburst +event logger](https://docs.starburst.io/latest/security/event-logger.html). This +feature logs query history and persists the data to an external database. This +is a prerequisite for Starburst Insights query history. + +## Usage + + minitrino --env STARBURST_VER=<ver> provision --module event-logger + docker exec -it trino bash + trino-cli + trino> show schemas from postgres_event_logger; diff --git a/lib/modules/security/file-access-control/readme.md b/lib/modules/security/file-access-control/readme.md index a4f896ae..3c87f945 100644 --- a/lib/modules/security/file-access-control/readme.md +++ b/lib/modules/security/file-access-control/readme.md @@ -1,31 +1,18 @@ # File Access Control Module + A module which utilizes Trino's [file-based system access control plugin](https://docs.starburst.io/latest/security/file-system-access-control.html). This also makes used of the [file-based group provider](https://docs.starburst.io/latest/security/group-file.html). -## Sample Usage -To provision this module, run: - -```shell -minitrino provision --module file-access-control -``` - -You will need to supply a username to the Trino CLI in order to map to a group -(see `group.txt` for which users belong to which groups). Example: - -```shell -trino-cli --user admin-2 -trino-cli --user metadata-1 -trino-cli --user platform -``` - ## Policies + The access policy is located in the `rules.json` file which defines groups of users that map to a certain access control permission. The users for the groups are defined in the `group.txt` file. -- Users in the `platform-admins` group have full access to all objects within Trino +- Users in the `platform-admins` group have full access to all objects within + Trino - Users in the `metadata-users` group only have access to the tables within the `system.metadata` schema - Users in the `platform-users` group only have access to the tables within the @@ -33,3 +20,18 @@ are defined in the `group.txt` file. You can modify this module to further specify access control permissions to other catalogs provisioned with other Minitrino modules. + +## Usage + + minitrino --env STARBURST_VER=<ver> provision --module file-access-control + docker exec -it trino bash + trino-cli --user admin + trino> show schemas from tpch; + +You will need to supply a username to the Trino CLI in order to map to a group +(see `lib/modules/security/file-access-control/resources/trino/group.txt` for +which users belong to which groups). Example: + + trino-cli --user admin-2 + trino-cli --user metadata-1 + trino-cli --user platform diff --git a/lib/modules/security/ldap/readme.md b/lib/modules/security/ldap/readme.md index d22547f0..76034daf 100644 --- a/lib/modules/security/ldap/readme.md +++ b/lib/modules/security/ldap/readme.md @@ -1,24 +1,29 @@ # LDAP Module + This module provisions an LDAP server for authenticating users in Trino. This -also enables SSL / TLS between the LDAP server and Trino, and between Trino -and clients. It is compatible with other security modules like **system-ranger** -and **event-logger**, but is mutually-exclusive of the **password-file** module. +also enables SSL / TLS between the LDAP server and Trino, and between Trino and +clients. It is compatible with other security modules like **system-ranger** and +**event-logger**, but is mutually-exclusive of the **password-file** module. -## Requirements -- N/A +## Usage -## Sample Usage -To provision this module, run: + minitrino --env STARBURST_VER=<ver> provision --module ldap + docker exec -it trino bash + + trino-cli --server https://trino:8443 \ + --truststore-path /etc/starburst/ssl/truststore.jks --truststore-password trinoRocks15 \ + --keystore-path /etc/starburst/ssl/keystore.jks --keystore-password trinoRocks15 \ + --user bob --password -```shell -minitrino provision --module ldap -``` + trino> show schemas from tpch; ## Default Usernames and Passwords + - alice / trinoRocks15 - bob / trinoRocks15 ## Client Keystore and Truststore + The Java keystore and truststore needed for clients and drivers to securely connect to Trino are located in a volume mount `~/.minitrino/ssl`. These two files are transient and will be automatically replaced whenever Minitrino is @@ -28,27 +33,24 @@ provisioned with a security module that enables SSL. Via Docker: -``` -docker exec -it trino trino-cli --server https://trino:8443 \ - --truststore-path /etc/starburst/ssl/truststore.jks --truststore-password trinoRocks15 \ - --keystore-path /etc/starburst/ssl/keystore.jks --keystore-password trinoRocks15 \ - --user bob --password -``` + docker exec -it trino trino-cli --server https://trino:8443 \ + --truststore-path /etc/starburst/ssl/truststore.jks --truststore-password trinoRocks15 \ + --keystore-path /etc/starburst/ssl/keystore.jks --keystore-password trinoRocks15 \ + --user bob --password -Via Host Machine: +Via host machine: -``` -trino-cli-xxx-executable.jar --server https://localhost:8443 \ - --truststore-path ~/.minitrino/ssl/truststore.jks --truststore-password trinoRocks15 \ - --keystore-path ~/.minitrino/ssl/keystore.jks --keystore-password trinoRocks15 \ - --user bob --password -``` + trino-cli-xxx-executable.jar --server https://localhost:8443 \ + --truststore-path ~/.minitrino/ssl/truststore.jks --truststore-password trinoRocks15 \ + --keystore-path ~/.minitrino/ssl/keystore.jks --keystore-password trinoRocks15 \ + --user bob --password Note that the CLI will prompt you for the password. ## Accessing the Trino Web UI -Open a web browser and go to https://localhost:8443 and log in with a valid LDAP -username and password. + +Open a web browser and go to <https://localhost:8443> and log in with a valid +LDAP username and password. To have the browser accept the self-signed certificate, do the following: @@ -61,32 +63,28 @@ Risk and Continue**. this website**. ## Adding a New User to LDAP + 1. Open a shell to the LDAP container -``` -docker exec -it ldap bash -``` + docker exec -it ldap bash 2. Create an LDIF file with the new user information: -``` -cat << EOF > jeff.ldif -dn: uid=jeff,dc=example,dc=com -changetype: add -uid: jeff -objectClass: inetOrgPerson -objectClass: organizationalPerson -objectClass: person -objectClass: top -cn: jeff -sn: jeff -mail: [email protected] -userPassword: trinoRocks15 -EOF -``` + cat << EOF > jeff.ldif + dn: uid=jeff,dc=example,dc=com + changetype: add + uid: jeff + objectClass: inetOrgPerson + objectClass: organizationalPerson + objectClass: person + objectClass: top + cn: jeff + sn: jeff + mail: [email protected] + userPassword: trinoRocks15 + EOF 3. Use the **ldapmodify** tool to add the new user -``` -ldapmodify -x -D "cn=admin,dc=example,dc=com" -w trinoRocks15 -H ldaps://ldap:636 -f jeff.ldif -``` + ldapmodify -x -D "cn=admin,dc=example,dc=com" \ + -w trinoRocks15 -H ldaps://ldap:636 -f jeff.ldif diff --git a/lib/modules/security/password-file/readme.md b/lib/modules/security/password-file/readme.md index 486f4517..b4f84c98 100644 --- a/lib/modules/security/password-file/readme.md +++ b/lib/modules/security/password-file/readme.md @@ -1,24 +1,29 @@ # Password File Authentication Module -This module configures Trino to authenticate users with a password file. It -also enables SSL / TLS between Trino and clients. It is compatible with other + +This module configures Trino to authenticate users with a password file. It also +enables SSL / TLS between Trino and clients. It is compatible with other security modules like **system-ranger** and **event-logger**, but is mutually-exclusive of the **ldap** module. -## Requirements -- N/A +## Usage -## Sample Usage -To provision this module, run: + minitrino --env STARBURST_VER=<ver> provision --module password-file + docker exec -it trino bash -```shell -minitrino provision --module ldap -``` + trino-cli --server https://trino:8443 \ + --truststore-path /etc/starburst/ssl/truststore.jks --truststore-password trinoRocks15 \ + --keystore-path /etc/starburst/ssl/keystore.jks --keystore-password trinoRocks15 \ + --user bob --password + + trino> show schemas from tpch; ## Default Usernames and Passwords + - alice / trinoRocks15 - bob / trinoRocks15 ## Client Keystore and Truststore + The Java keystore and truststore needed for clients and drivers to securely connect to Trino are located in a volume mount `~/.minitrino/ssl`. These two files are transient and will be automatically replaced whenever Minitrino is @@ -28,26 +33,23 @@ provisioned with a security module that enables SSL. Via Docker: -``` -docker exec -it trino trino-cli --server https://trino:8443 \ - --truststore-path /etc/starburst/ssl/truststore.jks --truststore-password trinoRocks15 \ - --keystore-path /etc/starburst/ssl/keystore.jks --keystore-password trinoRocks15 \ - --user bob --password -``` + docker exec -it trino trino-cli --server https://trino:8443 \ + --truststore-path /etc/starburst/ssl/truststore.jks --truststore-password trinoRocks15 \ + --keystore-path /etc/starburst/ssl/keystore.jks --keystore-password trinoRocks15 \ + --user bob --password Via Host Machine: -``` -trino-cli-xxx-executable.jar --server https://localhost:8443 \ - --truststore-path ~/.minitrino/ssl/truststore.jks --truststore-password trinoRocks15 \ - --keystore-path ~/.minitrino/ssl/keystore.jks --keystore-password trinoRocks15 \ - --user bob --password -``` + trino-cli-xxx-executable.jar --server https://localhost:8443 \ + --truststore-path ~/.minitrino/ssl/truststore.jks --truststore-password trinoRocks15 \ + --keystore-path ~/.minitrino/ssl/keystore.jks --keystore-password trinoRocks15 \ + --user bob --password Note that the CLI will prompt you for the password. ## Accessing the Trino Web UI -Open a web browser and go to https://localhost:8443 and log in with a valid + +Open a web browser and go to <https://localhost:8443> and log in with a valid username and password. To have the browser accept the self-signed certificate, do the following: @@ -64,6 +66,4 @@ this website**. Example with username `jeff` and password `trinoRocks15` -``` -docker exec trino htpasswd -bB -C 10 /etc/starburst/password.db jeff trinoRocks15 -``` + docker exec trino htpasswd -bB -C 10 /etc/starburst/password.db jeff trinoRocks15 diff --git a/lib/modules/security/system-ranger/readme.md b/lib/modules/security/system-ranger/readme.md index d3a7eb86..8da2d428 100644 --- a/lib/modules/security/system-ranger/readme.md +++ b/lib/modules/security/system-ranger/readme.md @@ -1,21 +1,26 @@ # System-Level Ranger Access Control Module -This module provisions a Ranger server and Postgres database for Ranger storage. It is an unsecured deployment without SSL or an authentication mechanism. -## Requirements -- Starburst Data license -- Access to Starburst Data Harbor repository for Docker images +This module provisions a Ranger server and Postgres database for Ranger storage. +It is an unsecured deployment without SSL or an authentication mechanism. + +## Usage + + minitrino --env STARBURST_VER=<ver> provision --module system-ranger + docker exec -it trino bash + trino-cli --user bob + trino> select * from tpch.sf100.customer limit 1; -## Sample Usage -To provision this module, run: +## Requirements -```shell -minitrino provision --module system-ranger -``` +- Starburst Data license +- Access to Starburst Data Harbor repository for Docker images ## Policies + - Bob: admin access to TPCH `sf100` schema - Alice: admin access to TPCH `sf10` schema ## Accessing Ranger + - Go to `localhost:6080` -- Sign in with user: admin and pass: trinoRocks15 +- Sign in with user: `admin` and password: `trinoRocks15` diff --git a/lib/version b/lib/version index 227cea21..38f77a65 100644 --- a/lib/version +++ b/lib/version @@ -1,1 +1,1 @@ -2.0.0 +2.0.1 diff --git a/readme.md b/readme.md index 26f4b278..bed4fcdc 100644 --- a/readme.md +++ b/readme.md @@ -1,40 +1,100 @@ # Minitrino + A command line tool that makes it easy to run modular Trino environments locally. Compatible with Staburst version 354-e and later. -[![PyPI version](https://badge.fury.io/py/minitrino.svg)](https://badge.fury.io/py/minitrino) -[![Build Status](https://travis-ci.org/jefflester/minitrino.svg?branch=master)](https://travis-ci.org/jefflester/minitrino) -[![Trino Slack](https://img.shields.io/static/v1?logo=slack&logoColor=959DA5&label=Slack&labelColor=333a41&message=join%20conversation&color=3AC358)](https://trinodb.io/slack.html) +[![PyPI +version](https://badge.fury.io/py/minitrino.svg)](https://badge.fury.io/py/minitrino) +[![Build +Status](https://travis-ci.org/jefflester/minitrino.svg?branch=master)](https://app.travis-ci.com/jefflester/minitrino.svg?branch=master) +[![Trino +Slack](https://img.shields.io/static/v1?logo=slack&logoColor=959DA5&label=Slack&labelColor=333a41&message=join%20conversation&color=3AC358)](https://trinodb.io/slack.html) ----- -**Latest Stable Release**: 2.0.0 +**Latest Stable Release**: 2.0.1 ----- ## Overview -- [Requirements](#requirements) -- [Installation](#installation) -- [CLI](#cli) - - [Top-Level CLI Options](#top-level-cli-options) - - [Provisioning Environments](#provisioning-environments) - - [Removing Resources](#removing-resources) - - [Shutting Down Environments](#shutting-down-environments) - - [Taking Environment Snapshots](#taking-environment-snapshots) - - [Manage User Configuration](#manage-user-configuration) - - [Install the Library](#install-the-library) - - [Display Module Metadata](#display-module-metadata) - - [Display CLI Version](#display-cli-version) - - [Pointing the CLI to the Minitrino Library](#pointing-the-cli-to-the-minitrino-library) -- [Minitrino Configuration File](#minitrino-configuration-file) -- [Project Structure](#project-structure) -- [Adding New Modules (Tutorial)](#adding-new-modules-tutorial) -- [Troubleshooting](#troubleshooting) -- [Reporting Bugs and Contributing](#reporting-bugs-and-contributing) + +- [Minitrino](#minitrino) + - [Overview](#overview) + - [Requirements](#requirements) + - [Installation](#installation) + - [End Users](#end-users) + - [Developers](#developers) + - [CLI](#cli) + - [Top-Level CLI Options](#top-level-cli-options) + - [Provisioning Environments](#provisioning-environments) + - [Environment Variables](#environment-variables) + - [Using Licensed Starburst Features](#using-licensed-starburst-features) + - [Removing Resources](#removing-resources) + - [Shutting Down Environments](#shutting-down-environments) + - [Taking Environment Snapshots](#taking-environment-snapshots) + - [Manage User Configuration](#manage-user-configuration) + - [Install the Library](#install-the-library) + - [Display Module Metadata](#display-module-metadata) + - [Display Minitrino Versions](#display-minitrino-versions) + - [Pointing the CLI to the Minitrino Library](#pointing-the-cli-to-the-minitrino-library) + - [Minitrino Configuration File](#minitrino-configuration-file) + - [[CLI] Section](#cli-section) + - [[DOCKER] Section](#docker-section) + - [[TRINO] Section](#trino-section) + - [[MODULES] Section](#modules-section) + - [Project Structure](#project-structure) + - [Trino Dockerfile](#trino-dockerfile) + - [Adding New Modules (Tutorial)](#adding-new-modules-tutorial) + - [Create the Module Directory](#create-the-module-directory) + - [Add Trino Resources](#add-trino-resources) + - [Add the Docker Compose YAML](#add-the-docker-compose-yaml) + - [Add a Metadata File](#add-a-metadata-file) + - [Add a Readme File](#add-a-readme-file) + - [Review Progress](#review-progress) + - [Configure the Docker Compose YAML File](#configure-the-docker-compose-yaml-file) + - [Important Implementation Details: Paths and Labels](#important-implementation-details-paths-and-labels) + - [Path References for Volumes and Build Contexts](#path-references-for-volumes-and-build-contexts) + - [Minitrino Docker Labels](#minitrino-docker-labels) + - [Test the New Catalog](#test-the-new-catalog) + - [Customizing Images](#customizing-images) + - [Bootstrap Scripts](#bootstrap-scripts) + - [Managing Trino's `config.properties` File](#managing-trinos-configproperties-file) + - [Troubleshooting](#troubleshooting) + - [Reporting Bugs and Contributing](#reporting-bugs-and-contributing) + Library](#pointing-the-cli-to-the-minitrino-library) + - [Minitrino Configuration File](#minitrino-configuration-file) + - [[CLI] Section](#cli-section) + - [[DOCKER] Section](#docker-section) + - [[TRINO] Section](#trino-section) + - [[MODULES] Section](#modules-section) + - [Project Structure](#project-structure) + - [Trino Dockerfile](#trino-dockerfile) + - [Adding New Modules (Tutorial)](#adding-new-modules-tutorial) + - [Create the Module Directory](#create-the-module-directory) + - [Add Trino Resources](#add-trino-resources) + - [Add the Docker Compose YAML](#add-the-docker-compose-yaml) + - [Add a Metadata File](#add-a-metadata-file) + - [Add a Readme File](#add-a-readme-file) + - [Review Progress](#review-progress) + - [Configure the Docker Compose YAML + File](#configure-the-docker-compose-yaml-file) + - [Important Implementation Details: Paths and + Labels](#important-implementation-details-paths-and-labels) + - [Path References for Volumes and Build + Contexts](#path-references-for-volumes-and-build-contexts) + - [Minitrino Docker Labels](#minitrino-docker-labels) + - [Test the New Catalog](#test-the-new-catalog) + - [Customizing Images](#customizing-images) + - [Bootstrap Scripts](#bootstrap-scripts) + - [Managing Trino's `config.properties` + File](#managing-trinos-configproperties-file) + - [Troubleshooting](#troubleshooting) + - [Reporting Bugs and Contributing](#reporting-bugs-and-contributing) ----- ## Requirements + - Docker 19.03.0+ - Docker Compose (1.29.0+) - Python 3.8+ @@ -46,17 +106,20 @@ locally. Compatible with Staburst version 354-e and later. ## Installation ### End Users + Minitrino is available on PyPI and the library is available for public download -on GitHub. To install the Minitrino CLI, run `pip install minitrino`. To -install the library, run `minitrino lib_install`. +on GitHub. To install the Minitrino CLI, run `pip install minitrino`. To install +the library, run `minitrino lib_install`. ### Developers + In the project's root, run `./install.sh` to install the Minitrino CLI. If you encounter errors during installation, try running `sudo -H ./install.sh -v`. ----- ## CLI + Minitrino is built with [Click](https://click.palletsprojects.com/en/7.x/), a popular, open-source toolkit used to build Python-based CLIs. @@ -65,8 +128,9 @@ options can be specified with a shorthand alternative, which is the first letter of each option, i.e. `--module` can be `-m`. ### Top-Level CLI Options + You can get help, enable verbose output, and change the runtime library -directory for any command. +directory for any command. ``` Usage: minitrino [OPTIONS] COMMAND [ARGS]... @@ -87,6 +151,7 @@ Options: ``` ### Provisioning Environments + You can provision an environment via the `provision` command. ``` @@ -152,6 +217,7 @@ Using the structure of the Minitrino library, it is able to merge multiple Docker Compose files together. #### Environment Variables + Environment variables passed to Docker containers are sourced through two locations. The first is from the `minitrino.env` file in the library root. These variables define the versions of the provisioned Docker services. The second is @@ -163,10 +229,11 @@ Any existing environment variable can be overridden with the top-level `--env` option, and any unset variable can be set with it. #### Using Licensed Starburst Features + If you are using licensed features, you will need to provide a path to a valid Starburst license. This can be set via `minitrino config` or provided via the `--env` option at command runtime. The variable for this is -`STARBURST_LIC_PATH`. +`STARBURST_LIC_PATH`. Additionally, you need to uncomment the volume mount in the library's root `docker-compose.yml` file: @@ -179,6 +246,7 @@ Additionally, you need to uncomment the volume mount in the library's root ``` ### Removing Resources + You can remove resources with the `remove` command. ``` @@ -203,8 +271,8 @@ Notes: - Named volumes tied to any *existing* container cannot be forcibly removed, neither by Minitrino nor by the Docker CLI/SDK. - Images tied to stopped containers can be forcibly removed, but any image tied - to a running container cannot be forcibly removed, neither by Minitrino nor - by the Docker CLI. + to a running container cannot be forcibly removed, neither by Minitrino nor by + the Docker CLI. - You can find a module's label key by looking at the module's `docker-compose.yml` file in the Minitrino library. @@ -220,6 +288,7 @@ minitrino -v remove \ This will only remove volumes associated to the Postgres catalog module. ### Shutting Down Environments + You can shut down an active environment with the `down` command. ``` @@ -244,8 +313,9 @@ minitrino -v down ``` ### Taking Environment Snapshots + You can capture snapshots for both active and inactive environments with the -`snapshot` command. +`snapshot` command. ``` Usage: minitrino snapshot [OPTIONS] @@ -280,7 +350,7 @@ Options: --help Show this message and exit. ``` -Notes: +Notes: - Minitrino records the original `provision` command and places it in the snapshot file as `provision-snapshot.sh`; this can be directly executed. This @@ -299,7 +369,8 @@ minitrino snapshot -n super-cool-env -m hive-s3 -m elasticsearch -m ldap ``` ### Manage User Configuration -You can manage Minitrino configuration with the `config` command. + +You can manage Minitrino configuration with the `config` command. ``` Usage: minitrino config [OPTIONS] @@ -317,6 +388,7 @@ Options: ``` ### Install the Library + You can install the Minitrino library with the `lib_install` command. Note that it is best practice to have the library version match the CLI version. You can check these versions with `minitrino version`. @@ -332,7 +404,8 @@ Options: ``` ### Display Module Metadata -You can see Minitrino module metadata with the `modules` command. + +You can see Minitrino module metadata with the `modules` command. ``` Usage: minitrino modules [OPTIONS] @@ -349,8 +422,9 @@ Options: ``` ### Display Minitrino Versions + You can display the Minitrino CLI and library versions with the `version` -command. +command. ``` Usage: minitrino version [OPTIONS] @@ -362,6 +436,7 @@ Options: ``` ### Pointing the CLI to the Minitrino Library + The Minitrino CLI should always point to a compatible library with the expected structure. The library directory can be set one of four ways, listed below in the order of precedence: @@ -382,11 +457,13 @@ pointer to the library in Minitrino's configuration via the `LIB_PATH` config. ----- ## Minitrino Configuration File + Sticky configuration is set in `~/.minitrino/minitrino.cfg`. The sections in this file each serve a separate purpose. ### [CLI] Section -These configs allow the user to customize the behavior of Minitrino. + +These configs allow the user to customize the behavior of Minitrino. - LIB_PATH: The filesystem path of the Minitrino library (specifically to the `lib/` directory). @@ -394,17 +471,19 @@ These configs allow the user to customize the behavior of Minitrino. "nano", etc. Defaults to the shell's default editor. ### [DOCKER] Section + These configs allow the user to customize how Minitrino uses Docker. - DOCKER_HOST: A URL pointing to an accessible Docker host. This is automatically detected by Docker otherwise. ### [TRINO] Section + These configs allow the user to propagate config to the Trino container. Since many modules can append to Trino's core files, the supported way to make propagate changes to these Trino files is with these configs. -- CONFIG: Configuration for Trino's `config.properties` file. +- CONFIG: Configuration for Trino's `config.properties` file. - JVM_CONFIG: Configuration for Trino's `jvm.config` file. A multiline example of this section (note the indentation): @@ -419,6 +498,7 @@ JVM_CONFIG= ``` ### [MODULES] Section + This section sets environment variables passed to containers provisioned by Minitrino. Environment variables are only passed to a container if the variable is specified in the module's `docker-compose.yml` file. @@ -446,9 +526,11 @@ Variables propagated to the Trino container are supported by Trino secrets. ----- -## Project Structure +## Project Structure + The library is built around Docker Compose files and utilizes Docker's ability -to [extend Compose files](https://docs.docker.com/compose/extends/#multiple-compose-files). +to [extend Compose +files](https://docs.docker.com/compose/extends/#multiple-compose-files). The Starburst Trino service is defined in a Compose file at the library root, and all other services look up in the directory tree to reference the parent @@ -514,19 +596,21 @@ services: Notice that the volume mount is not relative to the `lib/modules/catalog/postgres/` directory––it is relative to the parent directory which houses the top-level `docker-compose.yml` file. Also, notice the -labels––these labels will be used to identify Docker resources tied to -Minitrino modules so that the CLI commands actually work. +labels––these labels will be used to identify Docker resources tied to Minitrino +modules so that the CLI commands actually work. ### Trino Dockerfile -Minitrino modifies the Starburst Trino Docker image by adding the Trino CLI -to the image as well as by providing `sudo` to the `trino` user. This is -required for certain bootstrap scripts (i.e. using `yum` to install packages in -a Trino container for a module). This image is compatible with Starburst Trino -images back to Starburst Trino version `332-e.0`. + +Minitrino modifies the Starburst Trino Docker image by adding the Trino CLI to +the image as well as by providing `sudo` to the `trino` user. This is required +for certain bootstrap scripts (i.e. using `yum` to install packages in a Trino +container for a module). This image is compatible with Starburst Trino images +back to Starburst Trino version `332-e.0`. ----- ## Adding New Modules (Tutorial) + Adding new modules is relatively simple, but there are a few important guidelines to follow to ensure compatibility with the Minitrino CLI. The design rules are the same for both catalogs and security modules. The example below @@ -534,6 +618,7 @@ demonstrates the process of creating a new catalog module for a Postgres service. ### Create the Module Directory + Create the module's directory in the `lib/modules/catalog/` directory: ```sh @@ -541,11 +626,12 @@ mkdir lib/modules/catalog/postgres/ cd lib/modules/catalog/postgres/ ``` -### Add Trino Resources +### Add Trino Resources + All resources for a module go inside of a `resources/` directory within the module. Inside this directory, place Trino-specific resources into a `trino/` directory, then mount the resources to the Trino service defined in the root -`docker-compose.yml` file. +`docker-compose.yml` file. ```sh mkdir -p resources/trino/ @@ -565,11 +651,12 @@ EOF" ----- **Note**: Passwords should always be `trinoRocks15` for consistency throughout -modules. +modules. ----- ### Add the Docker Compose YAML + In `lib/modules/catalog/postgres/`, add a Docker Compose file: ```sh @@ -598,10 +685,11 @@ POSTGRES_DB=minitrino EOF" ``` -This file will initialize Postgres with a database `minitrino`, a user -`trino`, and a password `trinoRocks15`. +This file will initialize Postgres with a database `minitrino`, a user `trino`, +and a password `trinoRocks15`. ### Add a Metadata File + The `metadata.json` file allows Minitrino to obtain key information for the module. It is required for a module to work with the CLI. @@ -622,6 +710,7 @@ alongside the given module. The `*` wildcard is a supported convention if the module is incompatible with all other modules. ### Add a Readme File + This step is not required for personal development, but it is required to commit a module to the Minitrino repository. @@ -633,7 +722,8 @@ touch readme.md This file should contain an overview of the module. -### Review Progress +### Review Progress + The resulting directory tree should look like this (from the `/modules/catalog/` directory): @@ -650,6 +740,7 @@ postgres ``` ### Configure the Docker Compose YAML File + We will now define the `postgres.yml` Docker Compose file. Set it up as follows, and **read the important notes after**: @@ -674,9 +765,11 @@ services: ``` ### Important Implementation Details: Paths and Labels + We can observe a few things about the Compose file we just defined. #### Path References for Volumes and Build Contexts + First, the volumes we mount *are not relative to the Compose file itself*, they are relative to the base `docker-compose.yml` file in the library root. This is because the CLI extends Compose files, meaning that all path references in child @@ -685,19 +778,20 @@ Compose files need to be relative to the positioning of the parent Compose file. The base Compose file is determined when you execute a Docker Compose command––the first Compose file referenced in the command becomes the base file, and that happens to be the `docker-compose.yml` file in the library root. This -is how Minitrino constructs these commands. +is how Minitrino constructs these commands. If this is confusing, you can read more about extending Compose files on the -[Docker docs](https://docs.docker.com/compose/extends/#multiple-compose-files). +[Docker docs](https://docs.docker.com/compose/extends/#multiple-compose-files). #### Minitrino Docker Labels + Secondly, notice how we applied sets of labels to the Postgres service. These labels tell the CLI which resources to target when executing commands. -In general, there is no need to apply labels to the Trino service since they -are already applied in the parent Compose file **unless** the module is an -extension of the Trino service itself (i.e. the Snowflake modules). Labels -should always be applied to: +In general, there is no need to apply labels to the Trino service since they are +already applied in the parent Compose file **unless** the module is an extension +of the Trino service itself (i.e. the Snowflake modules). Labels should always +be applied to: - Docker services (AKA the resulting container) - Named volumes @@ -753,11 +847,12 @@ volumes: actually define any new services/containers. See the Snowflake catalog modules for an example of this. For these modules, the only label requirement is to add the module-specific label to the Trino service in the relevant -`docker-compose.yml` file +`docker-compose.yml` file ----- ### Test the New Catalog + We are all finished up. We can test our new catalog through the Minitrino CLI: ```sh @@ -773,6 +868,7 @@ trino> show catalogs; ``` ### Customizing Images + If you need to build an image from a local Dockerfile, you can do so and structure the Compose file accordingly. See the library's root `docker-compose.yml` file for an example of this. Path references for volumes @@ -780,12 +876,12 @@ and the image build context will follow the same convention as volume mount paths described earlier. ### Bootstrap Scripts -Minitrino supports container bootstrap scripts. These scripts **do not -replace** the entrypoint (or default command) for a given container. The script -is copied from the Minitrino library to the container, executed, and then -removed from the container. Containers are restarted after each bootstrap script -execution, **so the bootstrap scripts themselves should not restart the -container's service**. + +Minitrino supports container bootstrap scripts. These scripts **do not replace** +the entrypoint (or default command) for a given container. The script is copied +from the Minitrino library to the container, executed, and then removed from the +container. Containers are restarted after each bootstrap script execution, **so +the bootstrap scripts themselves should not restart the container's service**. If a bootstrap script has already executed in a container *and* the volume associated with the container still exists, Minitrino will not re-execute the @@ -813,6 +909,7 @@ services: The `elasticsearch` module is a good example of this. ### Managing Trino's `config.properties` File + Many modules can change the Trino `config.properties` and `jvm.config` files. Because of this, there are two supported ways to modify these files with Minitrino. @@ -823,8 +920,8 @@ This will propagate the config to the Trino container when it is provisioned. Generally speaking, this can be used for any type of configuration (i.e. memory configuration) that is unlikely to be modified by any module. This also applies to the `jvm.config` file, which has identical support via the `JVM_CONFIG` -variable. If there are duplicate configs in either file, Minitrino will warn -the user. +variable. If there are duplicate configs in either file, Minitrino will warn the +user. To set these configs, your configuration file should look like: @@ -884,12 +981,13 @@ as possible. ----- ## Reporting Bugs and Contributing + To report bugs, please file a GitHub issue on the [Minitrino repository](https://github.com/jefflester/minitrino). Bug reports should: - Contain any relevant log messages (if the bug is tied to a command, running with the `-v` flag will make debugging easier) -- Describe what the expected outcome is +- Describe what the expected outcome is - Describe the proposed code fix (optional) Contributors have two options: @@ -897,7 +995,7 @@ Contributors have two options: 1. Fork the repository, then make a PR to merge your changes 2. If you have been added as a contributor, you can go with the method above or you can create a feature branch, then submit a PR for that feature branch - when it is ready to be merged. + when it is ready to be merged. In either case, please provide a comprehensive description of your changes with the PR. diff --git a/release-notes/2.0.1.md b/release-notes/2.0.1.md new file mode 100644 index 00000000..6f15ee47 --- /dev/null +++ b/release-notes/2.0.1.md @@ -0,0 +1,27 @@ +# Minitrino Release Notes: 2.0.1 + +## Release Overview + +- [Release Overview](#release-overview) +- [CLI Changes and Additions](#cli-changes-and-additions) +- [Library Changes and Additions](#library-changes-and-additions) +- [Other](#other) + +## CLI Changes and Additions + +- Check for valid SEP version before provisioning modules + +## Library Changes and Additions + +- Add Db2 catalog module +- Add Oracle catalog module (address part of issue #3) +- Exposed all catalog modules on a host port for alternative client access + +## Other + +- Fix conflicting port issue with modules that use overlapping services + (resolves issue #35) +- Standardize module readmes and complete all incomplete readmes (resolves issue + #8) +- Fix issue with snapshots where the incorrect CLI syntax is written to the + provision command

Populate module readmes The readme documents for each module should be accompanied with a high-level description of what the module does and any important instructions for interacting with the module. Will close this issue once all modules have this and it becomes a standard process for new module PRs.

2021-08-20T07:03:08

0.0

[]

securefederatedai/openfl

securefederatedai__openfl-991

72312932edb198b2de04aee395ebd688ba9895c4

diff --git a/openfl/experimental/component/aggregator/__init__.py b/openfl/experimental/component/aggregator/__init__.py index 6686ce37b8..c2af4cc2ac 100644 --- a/openfl/experimental/component/aggregator/__init__.py +++ b/openfl/experimental/component/aggregator/__init__.py @@ -3,4 +3,4 @@ """openfl.experimental.component.aggregator package.""" # FIXME: Too much recursion. -from openfl.experimental.component.aggregator import Aggregator +from openfl.experimental.component.aggregator.aggregator import Aggregator diff --git a/openfl/experimental/interface/cli/collaborator.py b/openfl/experimental/interface/cli/collaborator.py index c5c6d06ac3..c5f8c924ee 100644 --- a/openfl/experimental/interface/cli/collaborator.py +++ b/openfl/experimental/interface/cli/collaborator.py @@ -75,7 +75,6 @@ def start_(plan, collaborator_name, secure, data_config="plan/data.yaml"): import yaml from yaml.loader import SafeLoader - collaborator_name = collaborator_name.lower() with open(data_config, "r") as f: data = yaml.load(f, Loader=SafeLoader) if data.get(collaborator_name, None) is None: @@ -118,7 +117,7 @@ def generate_cert_request(collaborator_name, silent, skip_package): from openfl.cryptography.participant import generate_csr from openfl.experimental.interface.cli.cli_helper import CERT_DIR - common_name = f"{collaborator_name}".lower() + common_name = f"{collaborator_name}" subject_alternative_name = f"DNS:{common_name}" file_name = f"col_{common_name}" @@ -282,7 +281,7 @@ def certify(collaborator_name, silent, request_pkg=None, import_=False): from openfl.experimental.interface.cli.cli_helper import CERT_DIR from openfl.utilities.utils import rmtree - common_name = f"{collaborator_name}".lower() + common_name = f"{collaborator_name}" if not import_: if request_pkg: diff --git a/openfl/experimental/interface/participants.py b/openfl/experimental/interface/participants.py index 4f48ae61fd..49d678a426 100644 --- a/openfl/experimental/interface/participants.py +++ b/openfl/experimental/interface/participants.py @@ -9,7 +9,7 @@ class Participant: def __init__(self, name: str = ""): self.private_attributes = {} - self._name = name.lower() + self._name = name @property def name(self): @@ -17,7 +17,7 @@ def name(self): @name.setter def name(self, name: str): - self._name = name.lower() + self._name = name def private_attributes(self, attrs: Dict[str, Any]) -> None: """

Issue observed in Workflow API Global DP & Vertical FL tutorials **Describe the bug** Running Workflow interface Global DP tutorials leads to unexpected **KeyError** exception. Similar issue is observed with Workflow interface Vertical_FL tutorial: Workflow_Interface_Vertical_FL.ipynb **To Reproduce** Steps to reproduce the error with Global DP tutorial: 1. Go to: openfl-tutorials/experimental/Global_DP/Workflow_Interface_Mnist_Implementation_1.py 2. Run tutorial with `python Workflow_Interface_Mnist_Implementation_1.py --config_path test_config.yml` 3. See error: **KeyError: 'portland'** **Expected behavior** Tutorial should run successfully without any error. **Desktop (please complete the following information):** - OS: Ubuntu (WSL) - OS Version: 22.04.3 LTS - Python: 3.8.19 [KeyError_logs.txt](https://github.com/user-attachments/files/15513351/KeyError_logs.txt)

2024-06-12T09:42:47

0.0

[]

securefederatedai/openfl

securefederatedai__openfl-550

88c437bc27bc2eafa68fb025fe6f07329dcfbe26

diff --git a/.github/workflows/taskrunner.yml b/.github/workflows/taskrunner.yml index 3545fd4c0b..510fb21b94 100644 --- a/.github/workflows/taskrunner.yml +++ b/.github/workflows/taskrunner.yml @@ -29,4 +29,4 @@ jobs: pip install . - name: Test TaskRunner API run: | - bash tests/github/test_hello_federation.sh keras_cnn_mnist aggregator col1 col2 $(hostname --all-fqdns | awk '{print $1}') --rounds-to-train 3 + bash tests/github/test_hello_federation.sh keras_cnn_mnist aggregator col1 col2 $(hostname --all-fqdns | awk '{print $1}') --rounds-to-train 3 --save-model output_model diff --git a/docs/running_the_federation.rst b/docs/running_the_federation.rst index ebe818cdec..49ea72b796 100644 --- a/docs/running_the_federation.rst +++ b/docs/running_the_federation.rst @@ -782,10 +782,10 @@ However, continue with the following procedure for details in creating a federat .. _creating_workspaces: -STEP 1: Create a Workspace on the Aggregator +STEP 1: Create a Workspace ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ -1. Start a Python 3.8 (>=3.6, <3.9) virtual environment and confirm |productName| is available. +1. Start a Python 3.8 (>=3.6, <3.11) virtual environment and confirm |productName| is available. .. code-block:: python @@ -1112,6 +1112,27 @@ STEP 3: Start the Federation When the last round of training is completed, the Aggregator stores the final weights in the protobuf file that was specified in the YAML file, which in this example is located at **save/${WORKSPACE_TEMPLATE}_latest.pbuf**. +Post Experiment +^^^^^^^^^^^^^^^ + +Experiment owners may access the final model in its native format. +Among other training artifacts, the aggregator creates the last and best aggregated (highest validation score) model snapshots. One may convert a snapshot to the native format and save the model to disk by calling the following command from the workspace: + +.. code-block:: console + + fx model save -i model_protobuf_path.pth -o save_model_path + +In order for this command to succeed, the **TaskRunner** used in the experiment must implement a :code:`save_native()` method. + +Another way to access the trained model is by calling the API command directly from a Python script: + +.. code-block:: python + + from openfl import get_model + model = get_model(plan_config, cols_config, data_config, model_protobuf_path) + +In fact, the :code:`get_model()` method returns a **TaskRunner** object loaded with the chosen model snapshot. Users may utilize the linked model as a regular Python object. + .. _running_the_federation_docker: diff --git a/openfl/__init__.py b/openfl/__init__.py index e29f0839b7..91cfdfc81f 100644 --- a/openfl/__init__.py +++ b/openfl/__init__.py @@ -1,5 +1,6 @@ # Copyright (C) 2020-2021 Intel Corporation # SPDX-License-Identifier: Apache-2.0 """openfl base package.""" -# flake8: noqa from .__version__ import __version__ +# flake8: noqa +from .interface.model import get_model diff --git a/openfl/interface/model.py b/openfl/interface/model.py new file mode 100644 index 0000000000..94468b797e --- /dev/null +++ b/openfl/interface/model.py @@ -0,0 +1,109 @@ +# Copyright (C) 2020-2022 Intel Corporation +# SPDX-License-Identifier: Apache-2.0 +"""Model CLI module.""" + +from click import confirm +from click import group +from click import option +from click import pass_context +from click import style +from click import Path as ClickPath +from logging import getLogger +from pathlib import Path + +from openfl.federated import Plan +from openfl.federated import TaskRunner +from openfl.protocols import utils +from openfl.pipelines import NoCompressionPipeline +from openfl.utilities.workspace import set_directory + +logger = getLogger(__name__) + + +@group() +@pass_context +def model(context): + """Manage Federated Learning Models.""" + context.obj['group'] = 'model' + + [email protected](name='save') +@pass_context +@option('-i', '--input', 'model_protobuf_path', required=True, + help='The model protobuf to convert', + type=ClickPath(exists=True)) +@option('-o', '--output', 'output_filepath', required=False, + help='Filename the model will be saved to in native format', + default='output_model', type=ClickPath(writable=True)) +@option('-p', '--plan-config', required=False, + help='Federated learning plan [plan/plan.yaml]', + default='plan/plan.yaml', type=ClickPath(exists=True)) +@option('-c', '--cols-config', required=False, + help='Authorized collaborator list [plan/cols.yaml]', + default='plan/cols.yaml', type=ClickPath(exists=True)) +@option('-d', '--data-config', required=False, + help='The data set/shard configuration file [plan/data.yaml]', + default='plan/data.yaml', type=ClickPath(exists=True)) +def save_(context, plan_config, cols_config, data_config, model_protobuf_path, output_filepath): + """ + Save the model in native format (PyTorch / Keras). + """ + output_filepath = Path(output_filepath).absolute() + if output_filepath.exists(): + if not confirm(style( + 'Do you want to overwrite the {}?'.format(output_filepath), fg='red', bold=True + )): + logger.info('Exiting') + context.obj['fail'] = True + return + + task_runner = get_model(plan_config, cols_config, data_config, model_protobuf_path) + + task_runner.save_native(output_filepath) + logger.info(f'Saved model in native format: 🠆 {output_filepath}') + + +def get_model( + plan_config: str, + cols_config: str, + data_config: str, + model_protobuf_path: str +) -> TaskRunner: + """ + Initialize TaskRunner and load it with provided model.pbuf. + + Contrary to its name, this function returns a TaskRunner instance. + The reason for this behavior is the flexibility of the TaskRunner interface and + the diversity of the ways we store models in our template workspaces. + """ + + # Here we change cwd to the experiment workspace folder + # because plan.yaml usually contains relative paths to components. + workspace_path = Path(plan_config).resolve().parent.parent + plan_config = Path(plan_config).resolve().relative_to(workspace_path) + cols_config = Path(cols_config).resolve().relative_to(workspace_path) + data_config = Path(data_config).resolve().relative_to(workspace_path) + + with set_directory(workspace_path): + plan = Plan.parse( + plan_config_path=plan_config, + cols_config_path=cols_config, + data_config_path=data_config + ) + collaborator_name = list(plan.cols_data_paths)[0] + data_loader = plan.get_data_loader(collaborator_name) + task_runner = plan.get_task_runner(data_loader=data_loader) + + model_protobuf_path = Path(model_protobuf_path).resolve() + logger.info(f'Loading OpenFL model protobuf: 🠆 {model_protobuf_path}') + + model_protobuf = utils.load_proto(model_protobuf_path) + + tensor_dict, _ = utils.deconstruct_model_proto(model_protobuf, NoCompressionPipeline()) + + # This may break for multiple models. + # task_runner.set_tensor_dict will need to handle multiple models + task_runner.set_tensor_dict(tensor_dict, with_opt_vars=False) + + del task_runner.data_loader + return task_runner diff --git a/openfl/utilities/workspace.py b/openfl/utilities/workspace.py index 08847f76aa..9afaa748d5 100644 --- a/openfl/utilities/workspace.py +++ b/openfl/utilities/workspace.py @@ -8,6 +8,7 @@ import shutil import sys import time +from contextlib import contextmanager from pathlib import Path from subprocess import check_call from sys import executable @@ -126,3 +127,23 @@ def _is_package_versioned(package: str) -> bool: and package not in ['pkg-resources==0.0.0', 'pkg_resources==0.0.0'] and '-e ' not in package ) + + +@contextmanager +def set_directory(path: Path): + """ + Sets provided path as the cwd within the context. + + Args: + path (Path): The path to the cwd + + Yields: + None + """ + + origin = Path().absolute() + try: + os.chdir(path) + yield + finally: + os.chdir(origin)

Load a trained pbuf model into general pytorch env **Is your feature request related to a problem? Please describe.** I would like to test the trained model externally, using an independent pytorch-based script that compares several models. The pbuf format is serialized differently that the regular .pth torch snapshots and can't be loaded using model.load_state_dict. **Describe the solution you'd like** It would be great if you can provide either a tool for converting pbuf to pth, or a flag enabling to save the model in pth format during the learning rounds. Thanks a lot.

2022-10-27T21:59:45

0.0

[]